it;

THE

METALLOGRAPHY OF IRON
AND STEEL

ALBERT SAUVEUR

i <

Professor of Metallurgy and Metallography in Harvard University

FIRST EDITION FIRST THOUSAND

McG RAW-HILL BOOK COMPANY

239 WEST 39TH STREET, NEW YORK

6 BOUVERIE STREET, LONDON, B.C.

1912

. ,3

THE UNIVERSITY PRESS, CAMBRIDGE, D. 8.A.

THE MEMORY OF

4Hp Jfatfter

I REVERENTLY AND LOVINGLY
DEDICATE THIS BOOK

PREFACE

WHILE several excellent books on metallography have been published and while
numerous papers on the metallography of iron and steel have appeared in the
scientific and technical press, a well-balanced, specific, and comprehensive treatise
on the subject has not heretofore been written. In the belief that there is a real and
urgent need of such a treatise the author has endeavored to supply it, craving for his
effort the indulgent criticism of his readers. He offers his book to those seeking self-
instruction in the metallography of iron and steel, their special needs having been
carefully considered in the arrangement of the lessons; he offers it to teachers and
students trusting that they will find it valuable and suggestive as a text-book; he
offers it to manufacturers and users of iron and steel in the belief that he has given
due weight to the practical side of the subject and has avoided discussions of ill-
founded or purely speculative theories; he offers it to the general reader interested
in the scientific or practical features of the metallography of iron and steel, as the
language used should be readily understood by those lacking specialized knowledge
of the subject; he offers it to experts in the hope that they will find it not entirely
devoid of original thought, original treatment, and suggestiveness.

In the matter of illustrations and especially of photomicrographs the author's aim
has been to utilize the best available, using his own or those taken in his laboratory
only when no better ones have, to his knowledge, been published by others. The
original source of every illustration has been indicated and the author desires to ex-
press his indebtedness to the following writers, the figures in parenthesis showing the
number of illustrations from each: Andrews (3), Arnold (7), Bayley (1), Belaiew (5),
Brearley (2), Carpenter and Keeling (l),Sherard Cowper-Coles (1), Desch (4), Edwards
(2), Ewing and Rosenhain (2), Guillet (18), Gcerens (9), Gulliver (2), Hall (1),
Houghton (1), Kroll (1), Law (8), Levy (1), Longmuir (2), Matweieff (1), Maurer (1),
Mellor (1), Osmond (17), Roberts-Austen (1), Robin (1), Roland-Gosselin (1), Rosen-
hain (2), Saladin (2), Sorby (1), Stead (13), Tschermak (3), Tschernoff (1), Wiist (5),
Ziegler (1). All illustrations not otherwise inscribed are the author's.

The author cannot refrain from expressing here the sorrow and sense of personal
loss he experienced when the news was received, while this book was passing through
the press, of the death of Floris Osmond, for to the author, as no doubt to many others,
Osmond's work and Osmond's life have been an inspiration. Osmond belonged to
that admirable class of French scientists, who, like Pasteur and Berthelot, have so
lofty a conception of the duty of the scientist that they give to the world the fruit of

VI PREFACE

their genius and of their untiring labors with no thought of monetary return or even
of honorary recognition. If Sorby was the pioneer of metallography and Tschernoff
its father, Osmond has been its torch-bearer for he, more than any other, has been our
guide. While he is no longer with us, his light will long continue to burn and to show
the way to promising and productive fields of research.

The author desires to place on record his warm appreciation of the assistance he
received from Mr. H. M. Boylston in passing this book through the press, and also

for many valuable suggestions.

ALBERT SAUVEUR.

HARVARD UNIVERSITY,

CAMBRIDGE, MASSACHUSETTS,
August 19, 1912.

TABLE OF CONTENTS

INTRODUCTION

PAGE

THE INDUSTRIAL IMPORTANCE OF METALLOGRAPHY 1

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

THE MICROSCOPE 1

The stage 1

Plain stages 3

Mechanical stages 3

Objectives 3

Eye-pieces 3

Iris diaphragms 7

Specimen holders 7 j

UNIVERSAL METALLOSCOPE 10

Electromagnetic stage 11

Templets for the examination of small specimens 11

Support of non-magnetic specimens 11

Leveling-devices of stand and stage 12

Motion of the stage 12

Mechanical stage 12

Examination of transparent objects 13

ILLUMINATION OF THE SAMPLES 14

SOURCES OF LIGHT AND CONDENSERS 18

Monochromatic light 22

PHOTOMICROGRAPHIC CAMERAS 22

INVERTED MICROSCOPES 28

POLISHING APPARATUS 28

Hand polishing 28

Polishing by power 30

PYROMETERS AND ELECTRIC FURNACES 30

Pyrometers 30

Electric Furnaces 35

LESSON I PURE METALS

Microstructure 1

Crystallization 1

Idiomorphic crystals 2

Allot rimorphic crystals 2

Crystallization of metals 2

Grains of metals

Crystalline orientation of the grains 3

Cubic crystallization of metals. Etching pits

Summary

Impurities "

Influence of thermal treatment 7

Influence of mechanical treatment

Examination 8

vii

viii TABLE OF CONTEXTS

LESSON II PURE IRON

PAGE

Microstructure 1

Cubic crystallization of iron 2

Ferrite 4

Allotropy of iron 4

Influence of impurities 10

Influence of heat treatment - 10

Influence of mechanical treatment 11

Straining of iron. Slip bands 11

Examination 12

LESSON III WROUGHT IRON

Chemical composition 1

Microstructure of longitudinal section 1

Microstructure of transverse section 2

Chemical composition of slag 3

Microstructure of slag 3

Influence of thermal and mechanical treatments 4

Experiments 4

Polishing by hand 4

Polishing by power 6

Etching 6

Etching with picric acid 6

Examination 7

Etching with diluted nitric acid 7

Etching with concentrated nitric acid 8

Examination 8

LESSON IV LOW CARBON STEEL

Normal structure 1

Grading of steel vs. carbon content 1

Low carbon steel vs. wrought iron 1

The structure of low carbon steel 2

Pearlite 3

Free ferrite 4

Cementite 5

Experiments 5

Polishing 5

Etching 6

Photomicrography 6

Exposure 7

Diaphragms and shutters 7

Monochromatic light 7

Photographic plates 8

Development 8

Printing 8

Mounting 8

Examination . 8

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

Medium high carbon steel 1

High carbon steel 4

,/^Eutectoid steel 4

TABLE OF CONTENTS ix

FACE

Hyper-eutectoid steel 4

Free cementite 5

Hypo- vs. hyper-eutectoid steel 6

Etching of cementite 7

Carbon content of pearlite 8

Structural composition of steel 8

Chemical vs. structural composition 11

Micro-test for determination of carbon in steel 12

Physical properties of the constituents of steel 14

Tenacity of steel vs. its structural composition 15

Steel of maximum strength 17

Ductility of steel vs. its structural composition 17

Diagram showing the relation between the tenacity and ductility of steel and its carbon content 19

Experiments 19

Etching 19

Etching with sodium picrate 19

Photomicrography 19

Examination 20

LESSON VI IMPURITIES IN STEEL

Metallic impurities 1

Xon-metallic or oxidized impurities 1

Metallic vs. non-metallic impurities 1

Gaseous impurities 1

Impurities vs. physical properties of steel 1

Silicon in steel . 1

Phosphorus in steel 2

Sulphur in steel 2

Manganese in steel 5

Chemical vs. structural composition 6

Xon-metallic or oxidized impurities 8

Gaseous impurities 9

Segregation of impurities. Ghosts 10

Experiments 11

High vs. low phosphorus steel 11

High sulphur steel 12

Oxidized Bessemer metal 12

Segregated steel -. . 12

Examination 12

LESSON VII THE THERMAL CRITICAL POINTS OF IRON AND STEEL

THEIR OCCURRENCE

Point of recalescence 1

Notation 2

Critical range. Transformation range 2

Position of Ari and Aci 2

Speed of cooling and heating vs. position of AI 4

Chemical composition vs. position of Ai 5

Upper critical points 5

Thermal critical points in pure iron 5

Thermal critical points in very low carbon steel 6

Peculiarities of the point A 2 6

Thermal critical points of medium high carbon steel 6

Merging of A 3 and A 2 .6

Thermal critical points in eutectoid steel 7

x TABLE OF CONTENTS

PAGE

Merging of A 3 . 2 and AI 7

Thermal critical points in hyper-eutectoid steel 7

Merging of A 3 . 2 .i and A cm 8

Minor critical points 8

Data showing the position of the critical points 8

Relative quantities of heat evolved or absorbed at the critical points 8

Graphical representation of the position and magnitude of the critical points 10

Determination of the thermal critical points 10

Cooling and heating curves 10

Use of neutral bodies 14

Additional illustrations of cooling curves 18

Self-recording pyrometers 18

Historical 19

Experiments If

Examination 20

LESSON VIII THE THERMAL CRITICAL POINTS OF IRON AND STEEL

THEIR CAUSES

Causes of the upper points A 3 and A 2 in carbonless iron 1

Causes of the upper critical points A 3 and A 2 in low carbon steel 3

Cause of the point A 3 . 2 4

Cause of the point AI 4

The point At an allotropic point 6

Pearlite formation 7

Cause of the point A cm 7

Allotropy of cementite 9

Cause of the point A 3 . 2 .i in eutectoid steel 9

Cause of the point A 3 . 2 .i in hyper-eutectoid steel 11

Formation of beta iron 11

Summary 11

Another view of the allotropic changes 14

Examination 10

LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL

THEIR EFFECTS

Changes at A 3 1

Dilatation 1

Electrical conductivity 1

Crystallization 2

Hardness, ductility, strength 2

Dissolving power for carbon 2

Structural properties 3

Other properties 3

Changes at A 2 3

Dilatation 3

Magnetic properties 3

Crystallization 5

Hardness, ductility, strength 5

Dissolving power for carbon 5

Structural properties 5

Other properties 6

Changes at A 3 .z 6

Changes at AI 6

Changes at A 3 . 2 .i 6

Changes at A cm 7

TABLE OF CONTENTS xi

PAGE

Structural change at Ai and A 3 . 2 .i 7

Prevailing conditions above and below the critical range 7

Properties of gamma, beta, and alpha iron 8

Examination 8

LESSON X CAST STEEL

Structure of cast eutectoid steel 1

Structure of cast hypo-eutectoid steel 2

Structure of cast eutectoid vs. cast hypo-eutectoid steel 3

Structure of cast hyper-eutectoid steel 4

Ingotism 5

Structure of cast steel vs. structure of meteorites 6

Octahedric crystallization of austenite 7

Experiments 10

Examination 10

LESSON XI THE MECHANICAL TREATMENT OF STEEL

Hot working 1

Finishing temperatures 3

Structure of hot worked eutectoid steel 4

Structure of hot worked hypo-eutectoid steel 4

Structure of hot worked hyper-eutectoid steel 6

Sorbite 6

Hot working of steel vs. its critical range 7

Cold working 8

Mechanical refining 9

Experiments 10

Examination 10

LESSON XII THE ANNEALING OF STEEL

Purpose of annealing 1

Nature of the annealing operation 1

Heating for annealing 1

Time at annealing temperature 2

Cooling from annealing temperature 2

Rate of cooling vs. carbon content 3

Rate of cooling vs. size of object 3

Furnace cooling from annealing temperature 4

Air cooling from annealing temperature 4

Properties of sorbite 5

Influence of maximum temperature 5

Influence of time at maximum temperature ( >

Oil and water quenching from annealing temperature (>

Double annealing treatment 8

Annealing eutectoid steel 10

Annealing hypo-eutectoid steel

Annealing hyper-eutectoid steel 12

Annealing steel castings 13

Spheroidizing of pearlite-cementite

Varieties of pearlite 15

Graphitizing of cementite

Burnt steel 17

Crystalline growth of austenite above the critical range

Crystalline growth of ferrite below the critical range 23

xii TABLE OF CONTENTS

PAGE

Brittleness of low carbon steel 26

Experiments 29

Examination 29

LESSON XIII THE HARDENING OF STEEL

Heating for hardening 1

Cooling for hardening 1

Structural changes on hardening 2

Austenite 3

Nature of austenite 3

Occurrence of austenite 4

Etching of austenite 6

Structure of austenite 7

Properties of austenite 7

Martensite 10

Nature of martensite 10

Occurrence of martensite 10

Etching of martensite 10

Structure of martensite 11

Properties of martensite 11

Troostite 11

Nature of troostite 11

Occurrence of troostite 12

Properties of troostite 13

Etching of troostite 13

Structure of troostite 13

Sorbite 13

Troosto-sorbite 15

Hardenite 15

Rate of cooling through critical range vs. structure of steel 15

Are the transition stages distinct constituents? 17

Metarals and aggregates 18

Hardening eutectoid steel 18

Hardening hyper-eutectoid steel 18

Hardening hypo-eutectoid steel 18

Steel of maximum hardening power 20

Hardening large pieces 20

Hardening and tempering in one operation 20

Experiments 20

Etching 21

Examination 21

LESSON XIV THE TEMPERING OF HARDENED STEEL

Tempering temperatures

Tempering colors

Time at tempering temperature 1

Rate of cooling from tempering temperature 2

Hardening and tempering combined

Explanation of the tempering of steel 2

Tempering austenitic steels 3

Tempering martensitic steel 5

Tempering t roost it ic steel 5

Tempering troosto-martensitic steel 5

Tempering troosto-sorbitic steel 5

Osmondite 5

Structural changes on slow cooling, quick cooling, and reheating 7

TABLE OF CONTENTS xiii

PAGE

Microstructure of hardened and tempered steel . . 7

Carbon condition in tempered steel 8

Decrease of hardness on tempering 9

Heat liberated on tempering 9

Experiments 10

Examination 10

LESSON XV THEORIES OF THE HARDENING OF STEEL

Retention theories 1

Solution theories 2

Gamma iron theory 2

Beta iron or allotropic theory 2

Alpha iron theory _ 4

Carbon theories 4

The hardening carbon theory 4

The subcarbide theory 4

The stress theory 5

Tempering and the retention theories 6

Tempering and the stress theory 6

Summary 6

Examination 7

LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL

Composition of the iron and steel subjected to carburizing 1

Carburizing temperature 1

Time at carburizing temperature 2

Dist ribution of the carbon 2

Carburizing materials 4

Mechanism of cementation 5

Cooling from carburizing temperature 5

Heat treatment of case hardened articles 5

Tempering case hardened steel 6

Experiments 6

Examination 6

LESSON XVII SPECIAL STEELS
GENERAL CONSIDERATIONS

Ternary steels 1

Influence of the special element upon the location of the critical range 3

Pearlitic steels 6

Martensitic steels 7

Austenitic (polyhedric) steels 7

Cementitic (carbide) steels 8

Treatments of special steels 8

Treatment of pearlitic steels 8

Treatment of martensitic steels 9

Treatment of austenitic steels 9

Treatment of cementitic steels 9

Quaternary steels 10

Examination 10

LESSON XVIII SPECIAL STEELS
CONSTITUTION, PROPERTIES, TREATMENT, AND USES OF MOST IMPORTANT TYPES

Nickel steel 1

Manganese steel 5

Tungsten steels 12

xiv TABLE OF CONTENTS

PAGE

Chrome steels 13

Vanadium steels 15

Silicon steels 15

Chrome-nickel steels 16

Quaternary vanadium steels 17

Chrome-tungsten or high-speed steels 17

Experiments 20

Examination 20

LESSON XIX CAST IRON

Formation 01 combined and graphitic carbon 1

Cast iron containing only graphitic carbon 1

Cast iron containing only combined carbon . 3

Cast iron containing both combined and graphitic carbon 8

Mottled cast iron 10

Structural composition of cast iron 10

Physical properties of cast iron vs. its structural composition 11

Chilled cast iron castings 13

Examination 13

LESSON XX IMPURITIES IN CAST IRON

Silicon in cast iron 1

Sulphur in cast iron 1

Manganese in cast iron 2

Phosphorus in cast iron 2

Structural composition of phosphoretic cast iron 7

Chemical vs. structural composition 8

Experiments 9

Examination 9

LESSON XXI MALLEABLE CAST IRON

Graphitizing of cementite 1

Malleable cast-iron castings 1

Original castings 2

Annealing operation 3

Packing materials 3

Annealing for malleablizing 4

Annealing for " white heart " castings 4

Annealing for " black heart " castings 5

Gray cast iron vs. malleable cast iron 7

Experiments 8

Examination 8

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

Solidification of pure metals 1

Solidification of binary alloys the constituents of which form solid solutions 3

Fusibility curves of binary alloys whose component metals are completely soluble in each other

when solid 5

Binary alloys forming definite compounds and solid solutions 8

Binary alloys whose component metals are insoluble in each other in the solid state 9

Binary alloys whose component metals are partially soluble in each other when solid 17

Examination 21

TABLE OF CONTENTS xv
LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

PAGE

Fusibility curve of iron-carbon alloys .......................... \

Structural composition of iron-carbon alloys immediately after solidification ......... 3

Iron-graphite fusibility curve ............................. 7

Combined graphite-cement ite diagram ......................... 7

Graphitizing of cementite ............................... 7

Structure of iron-carbon alloys immediately after solidification .............. 10

Complete equilibrium diagram ............................. 12

Historical ...................................... 16

Examination .................................... 21

LESSON XXIV THE PHASE RULE

Enunciation of the phase rule ............................. 1

Equilibrium ..................................... 1

Degrees of freedom ................................. 2

Phases ....................................... 3

Components .................................... 3

The phase rule applied to alloys ............................ 3

The phase rule applied to pure metals ......................... 4

The phase rule applied to binary alloys ......................... 4

The phase rule applied to iron-carbon alloys ....................... 6

Examination .................................... 8

APPENDIX I MANIPULATIONS AND APPARATUS

POLISHING AND POLISHING MACHINES ......................... 1

DEVELOPMENT OF THE STRUCTURES .......................... 9

Polishing in relief ................................ 9

Polish-attack .................................. 9

Etching .................................... 10

Electrolytic etching ............................... 11

Heat tinting .................................. 11

Hot etching ................................... 11

Washing and drying ............................... 11

Preserving .................................... 12

MOUNTING AND MOUNTING DEVICES .......................... 12

Plastic mountings ................................ 12

Leveling stages ................................. 14

METALLURGICAL MICROSCOPES ............................ 16

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

I. GENERAL PLAN .................................. 1

II. LIST OF MICROSCOPIC SUBSTANCES ......................... 2

III. DEFINITIONS AND DESCRIPTIONS .......................... 4

Austenite .................................... 4

Cementite ................................... 6

Martensite ................................... 7

Ferrite ..................................... 7

Osmondite ................................... 8

Ferronite .................................... 9

Hardenite ...................................

Pearlite .................................... 9

Graphite .................................... 1( J

XVI TABLE OF CONTENTS

III. DEFINITIONS AND DESCRIPTIONS continued PAGE

Troostite 11

Sorbite 11

Manganese sulphide 12

Ferrous sulphide 12

MISCELLANEOUS 12

INDEX 1-15

ERRATA

Lesson IV, page 5, fifth paragraph, line 3, for "an iron carbide Mn 3 C" read "a carbide MnjC."
Lesson V, page 4, second paragraph, line 3, for "euctectic" read "eutectic."

Lesson V, page 9, the first two equations should read
"F = per cent free ferrite = 41.18
P = per cent pearlite = 58.82"

Lesson V, page 16, instead of "T = 1250 P + 100 (50 - P)"
the third equation should read

"T = 1250 P + 50 (100 - P)"

instead of "T = 5000 + 1150 P"
the fourth equation should read
"T = 5000 + 1200 P"

instead of "T = 5000 + 1150 80 - 120C "
the sixth equation should read

"T - 5000 + 1200 8 - 120C "

instead of "T

7
955,000 - 138,000 C"

7
the seventh equation should read

_ 995,000 - 144,000 C"

instead of "T = 136,000 - 20,000 C."
the eighth equation should read

"T = 142,000 - 20,600 C."

last line, instead of "111,000" read "116,250"
instead of "106,000" read "111,100"

Lesson V, page 17, last line of footnote, for "allow heat treatment" read "allow for heat treatment."
Lesson X, page 6, eighth line from the bottom, for "peroid" read "period."
Lesson XV, page 6, next to last line, last word, for "carbon" read "iron."

INTRODUCTION

THE INDUSTRIAL IMPORTANCE OF METALLOGRAPHY 1

Twenty years ago the science of metallography was practically unknown and it
is only within the last fifteen years that it has been seriously considered by metal
manufacturers and consumers as a valuable method of testing and investigating.
That so much has been accomplished in so short a time is highly gratifying to the
many workers, practical or scientific, who have contributed by their efforts to the
progress of metallography.

To realize the practical importance of metallography it should be borne in mind
that the physical properties of metals and alloys that is, those properties to which
these substances owe their exceptional industrial importance are much more closely
related to their proximate than to their ultimate composition, and that microscopical
examination reveals, in part at least, the proximate composition of metals and alloys,
whereas chemical analysis seldom does more than reveal their ultimate composition.

It will bear repeating that from the knowledge of the proximate composition of a
certain industrial metal or alloy we are able to infer its properties and, therefore,
predict its adaptability with a much greater degree of accuracy than if we knew only
its ultimate composition.

The analytical chemist may tell us, for instance, that a steel which he has analyzed
contains 0.50 per cent of carbon, without our being able to form any idea as to its
properties, for such steel may have a tenacity of some 75,000 Ibs. per square inch or
of some 200,000 Ibs., a ductility represented by an elongation of some 25 per cent, or
practically no ductility at all; it may be so hard that it cannot be filed or so soft as
to be easily machined, etc.

The metal microscopist, on the contrary, on examining the same steel will report
its structural, i.e. its proximate, composition, informing us that it contains, for in-
stance, approximately 50 per cent of ferrite and 50 per cent of pearlite, and we know
at once that the steel is fairly soft, ductile, and tenacious; or he may report the
presence of 100 per cent of martensite, and we know that the steel is extremely hard,
very tenacious, and deprived of ductility.

Which of the two reports is of more immediate practical value, the chemist's or
the metallographist's? Surely, that of the metallographist.

Nor is it only in the domain of metals that we find such close relationship between
properties and proximate composition, for, on the contrary, it is quite true of all
substances. How many organic bodies, for instance, have practically the same ulti-
mate composition and still are totally unlike in properties because of their different
proximate composition, i.e. different grouping and association of their ultimate con-

1 Abstracted from a paper presented at the Congress of Technology at the fiftieth anniversary
of the granting of the charter of the Massachusetts Institute of Technology.

W 'THE INDUSTRIAL IMPORTANCE OF METALLOGRAPHY

stituents. If we were better acquainted with the proximate composition of substances
many unexplained facts would become clear to us.

Unfortunately the chemist too often is able to give us positive information in
regard to the proportion of the ultimate constituents only, his reference to proximate
analysis being of the nature of speculation. Ultimate analysis has reached a high
degree of perfection in regard to accuracy as well as to speed of methods and analy-
tical chemists have built up a marvelous structure calling for the greatest admiration,
their searching methods never failing to lay bare the ultimate composition of sub-
stances. But how much darkness still surrounds the proximate composition of
bodies and how great the reward awaiting the lifting of the veil!

The forceful and prophetic writing in 1890 of Prof. Henry M. Howe naturally
comes to mind. Speaking of the properties and constitution of steel, Professor Howe
wrote :

"If these views be correct, then, no matter how accurate and extended our knowl-
edge of ultimate composition, and how vast the statistics on which our inferences are
based, if we attempt to predict mechanical properties from them accurately we be-
come metallurgical Wigginses . . .

"Ultimate analysis never will, proximate analysis may, but by methods which
are not yet even guessed at, and in the face of fearful obstacles.

"How often do we look for the coming of the master mind which can decipher our
undecipherable results and solve our insoluble equations, while if we will but rub our
own dull eyes and glance from the petty details of our phenomena to their great out-
lines their meaning stands forth unmistakably; they tell us that we have followed
false clues and paths which lead but to terminal morasses. In vain we flounder in
the sloughs and quagmires at the foot of the rugged mountain of knowledge seeking
a royal road to its summit. If we are to climb, it must be by the precipitous paths
of proximate analysis, and the sooner we are armed and shod for the ascent, the sooner
we devise weapons for this arduous task, the better.

"By what methods ultimate composition is to be determined is for the chemist
rather than the metallurgist to discover. But, if we may take a leaf from lithology,
if we can sufficiently comminute our metal (ay, there's the rub!) by observing dif-
ferences in specific gravity (as in ore dressing), in rate of solubility under rigidly fixed
conditions, in degree of attraction by the magnet, in cleavage, luster, and crystalline
form under the microscope, in readiness of oxidation by mixtures of gases in rigidly
fixed proportions, we may learn much.

"Will the game be worth the candle? Given the proximate composition, will not
the mechanical properties of the metal be so greatly influenced by slight and unde-
terminable changes in the crystalline form, size, and arrangement of the component
minerals, so dependent on trifling variations in manufacture as to be still only roughly
deducible?"

The above was written before the days of metallography, or at least when metal-
lography had barely appeared in the metallurgical sky and when no one yet had fan-
cied what would be the brilliant career of the newcomer. Metallography has done
much to supply the need so vividly and timely depicted by Professor Howe, precisely
because by lifting a corner of the veil hiding from our view the proximate composi-
tion of metals and alloys it has thrown a flood of light upon the real constitution of
these important products. Has the game been worth the candle? Will any one
hesitate to answer in the affirmative Professor Howe's question?

INTRODUCTION THE INDUSTRIAL IMPORTANCE OF METALLOGRAPHY 3

Professor Howe with his usual acumen was conscious of the fact that proximate
analysis, while likely to reveal a great deal more of the constitution of metals than
ultimate analysis ever could, might still leave us in such ignorance of their physical
structure as to throw but little additional light upon the subject. His fear was cer-
tainly well founded and surely if the proximate composition had been obtained by
chemical analysis it would indeed have told us little of the structure or anatomy of
the metals. In the domain of proximate composition chemistry cannot do more for
the metallurgist than it does for the physician.

Invaluable information chemistry does give, without which both the physician and
the metallurgist would be in utter darkness, but it throws little or no light upon the
anatomy of living or inanimate matter. Its very methods which call for the destruc-
tion of the physical structure of matter show how incapable it is to render assistance
in this, our great need.

The parallel drawn here between metals and living matter is not fantastic. It
has been aptly made by Osmond, who said rightly that modern science was treating
the industrial metal like a living organism and that we were led to study its anatomy,
i.e. its physical and chemical constitution; its biology, i.e. the influence exerted upon
its constitution by the various treatments, thermal and mechanical, to which the
metal is lawfully subjected; and its pathology, i.e. the action of impurities and
defective treatments upon its normal constitution.

Fortunately metallography does more than reveal the proximate composition of
metals. It is a true dissecting method which lays bare their anatomy that is, the
physical grouping of the proximate constituents, their distribution, relative dimen-
sions, etc., all of which necessarily affect the properties. For two pieces of steel, for
instance, might have exactly the same proximate composition that is, might con-
tain, let us say, the same proportion of pearlite and ferrite and still differ quite a
little as to strength, ductility, etc., and that because of a different structural arrange-
ment of the two proximate constituents; in other words, because of unlike anatomy.

It is not to be supposed that the path trodden during the last score of years was
at all times smooth and free from obstacles. Indeed, the truth of the proverb that
there is no royal road to knowledge was constantly and forcibly impressed on the
mind of those engaged in the arduous task of lifting metallography to a higher level.

Its short history resembles the history of the development of all sciences. At the
outset a mist so thick surrounds the goal that only the most courageous and better
equipped attempt to pierce it and perchance they may be rewarded by a gleam of
light. This gives courage to others and the new recruits add strength to the besieg-
ing party. Then follows the well-known attacking methods of scientific tactics and
strategy, and after many defeats and now and then a victorious battle the goal is in
sight, but only in sight and never to be actually reached, for in our way stands the
great universal mystery of nature : what is matter? what is life?

Nevertheless there is reward enough for the scientist in the feeling that he has
approached the goal, that he has secured a better point of vantage from which to
contemplate it. The game was worth the candle. And if scientific workers must
necessarily fail in their efforts to arrive at the true definition of matter, whatever be
the field of their labor, they at least learn a great deal concernirig the ways of matter,
and it is with the ways of matter that the material world is chiefly concerned. Hence
the usefulness of scientific investigation, hence the usefulness of metallography.

Like any other science with any claim to commercial recognition, metallography

4 INTRODUCTION THE INDUSTRIAL IMPORTANCE OF METALLOGRAPHY

has had first to withstand the attack and later to overcome the ill-will and reluctance
of the so-called "practical man" with a decided contempt for anything scientific.
He represents the industrial philistine clumsily standing in the way of scientific ap-
plications to industrial operations. Fortunately, while his interference may retard
progress, it cannot prevent it. Had he had his own way neither the testing machine,
nor the chemical laboratory, nor the metallographical laboratory, nor the pyrometer
would ever have been introduced in iron and steel works.

Speaking in 1904 of the practical value of metallography in iron and steel making,
the author wrote the following, which it may not be out of place to reproduce here :
"History, however, must repeat itself, and the evolution of the metallographist bids
fair to be an exact duplicate of the evolution of the iron chemist ; the same landmarks
indicate his course; distrust, reluctant acceptance, unreasonable and foolish expecta-
tion from his work, disappointment because these expectations were not fulfilled and
finally the finding of his proper sphere and recognition of his worth. The metal-
lographist has passed through the first three stages of this evolution, is emerging
from the fourth, and entering into the last. For so young a candidate to recognition
in iron and steel making this record is on the whole very creditable."

We may say to-day that he has definitely entered the last stage and that the ad-
verse criticisms still heard from time to time, generally from the pen or mouth of
ignorant persons, are like the desultory firing of a defeated and retreating enemy.

In the United States alone the microscope is in daily use for the examination of
metals and alloys in more than two hundred laboratories of large industrial firms,
while metallography is taught in practically every scientific or technical school.

A. S.

HARVARD UNIVERSITY,
February, 1912.

APPARATUS FOR THE METALLO GRAPHIC LABORATORY. 1

Only those apparatus which the author has found most satisfactory are here men-
tioned other instruments will be found described in an Appendix.

THE MICROSCOPE

While any good microscope of the ordinary type, substantially built and provided
with a satisfactory fine adjustment, may be used with a certain degree of success for
the examination of metals and alloys, those who are restricted to its use will soon
find themselves seriously handicapped in several directions and unable to obtain the
desired results. The following considerations will make this clear.

The Stage. Ordinary microscope stands being constructed for the examination
of objects by transmitted light, i.e. by light proceeding from below the stage and
passing through the object on its way to the eye, are provided with fixed stages. This,
however, is a serious objection when the instrument is applied to the examination of
metals and other opaque objects which must necessarily be illuminated by light di-
rected upon them from above the stage, and which therefore require the use of an
" illuminator " attached to the objective and consequently moving with it. It will
be readily understood that it is of considerable importance that the position of this
illuminator, and therefore of the objective to which it is attached, be kept constant,
once the necessary adjustments are effected, since any change in its position would
require a readjustment of the source of light, the condensing lenses, diaphragm, etc.
To that effect the stage should be provided with a rack and pinion motion by means
of which the coarse focusing at least may be done (Fig. 1).

This rack and pinion motion of the stage, moreover, permits of a much greater
working distance, allowing plenty of room for the insertion of the illuminator and
nose-piece, the use of specimen holders, and the examination of bulky specimens with
low-power objectives. In the microscope illustrated in Figure 1, the working distance
measures over 5 inches as against 4 inches or less in ordinary stands.

By not departing more than necessary from the usual construction of microscopes,
none of the essential features required for the examination of transparent prepara-
tions need be sacrificed, and the full efficiency of the microscope is retained for such
examination, sub-stage condensers, polarizing prisms, etc., being readily attached
when needed. The possibility of applying his instrument to all kinds of microscopical
work with equally satisfactory results should appeal strongly to the metallographist,
for there is no laboratory where, occasionally at least, examination of transparent
objects is not desirable or even imperative.

1 Abstracted in part from papers by the author on "Apparatus for the Microscopical Examina-
tion of Metals," American Society for Testing Materials, Vol. X, 1910; and "The Universal Metal-
loscope," American Institute of Mining Engineers, June, 1911.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Fig. 1. Metallurgical microscope, eye-piece, vertical illuminator, objective,
magnetic specimen holder, and mechanical stage.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY 3

A less expensive but satisfactory microscope is shown in Figures 2 and 3. The
latter illustration includes an auxiliary tube inserted between the objective and illu-
minator for the examination of large samples which may be placed on the base of the
microscope or supported in some other suitable way below the stage.

(a) Plain Stages. While a mechanical stage adds greatly to the convenience of
the manipulations, a plain stage may be used with satisfactory results. It should be
provided with strong clips to hold in place the specimen holders soon to be described,
and should preferably be circular and revolving (Fig. 4). When provided with cen-
tering screws like the stage of the stand illustrated in Figure 1, the object may be moved
gently while under examination, a very desirable feature especially when using high-
power objectives, in which case the moving of the object entirely by hand is very jerky.
In order to derive the full benefit of the use of the magnetic holder described later,
the central opening of the stage should not be less than 1J4 inches in diameter.

(b) Mechanical Stages. The great superiority of a mechanical stage permitting,
as it does, a systematic examination of the object over its entire surface, need not be
insisted upon. In connection with the magnetic holder it makes it possible, moreover,
to examine repeatedly and at any time the same spot of any specimen, as will soon be
explained. The mechanical stage illustrated in Figure 5 has been especially designed
to fit the metallurgical microscope (Fig. 1), and is very readily substituted for the plain
stage. The central opening measures 1^ inches in diameter, permitting the convenient
use of the magnetic holder.

Objectives. Ordinary achromatic objectives give satisfactory results. They
should, however, be corrected for uncovered objects, as the placing of cover glasses
over bright metallic surfaces is accompanied by light reflection causing loss of clearness
and definition.

Some believe that the objectives should be provided with short mounts so as to
bring the reflector of the vertical illuminator as near the back lens of the objective as
possible, and thus, in their opinion, materially decreasing the amount of glare caused
by the reflection of light by the lenses of the objectives. Three objectives, one of low,
one of medium, and one of high power, will generally suffice for metallographic work.
The following focal lengths are recommended: 32-mm. or 1^-in., 16-mm. or %-m..,
and 4-mm. or Y-\\\. These objectives are shown, in short mounts, in Figure 6. The
32-mm. objective is provided with a society screw at its lower end in order that the
vertical illuminator may be inserted between the objective and the object, this being
desirable with very low-power lenses. In case a higher power is needed, a 2-mm. or
lV~ m - oil immersion objective will be found very satisfactory. When a very low-
power lens is required, as for instance in the examination of fractures or of very coarse
structures, a 48-mm. or 2-in. objective will give good results. It is suggested that it
be provided at its lower end with a society screw to permit the attachment of the ver-
tical illuminator, which in the case of such low-power lenses should be placed
between the object and the objective, as explained later.

Eye-Pieces. With achromatic objectives ordinary Huygenian eye-pieces are
used. Two eye-pieces, respectively of 1-in. and 2-in. focal length, will generally cover
the range of magnification needed.

For the taking of photomicrographs, projection eye-pieces are said to possess
some superiority, especially when high-power objectives are used, as they then yield
flatter and more sharply defined images. The Zeiss projection eye-piece No. 2 is very
satisfactory.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Fig. 2. Student microscope (.6 actual size).

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Fig. 3. Student microscope fitted with auxiliary tube.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Fig- 4. Plain revolving stage, magnetic specimen
holder, and specimen.

Fig. 5. Mechanical stage to fit metallurgical microscope,
magnetic specimen holder, and specimen.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Iris Diaphragms. Iris diaphragms are sometimes inserted between the objec-
tives and the illuminator so as to control the size of the pencil of light proceeding from
the object, with a view of securing sharper definition. Their use in that position,
however, is of doubtful value, as it may cause some distortion of the image. It seems
preferable to place the iris diaphragm between the source of light and the illuminator,
thus regulating the amount of light entering the latter. When placed between the
objective and the illuminator it increases, moreover, their distance apart, which we
have seen to be objectionable. If a diaphragm must be attached to the microscope,
it is better to place it between the tube nose and the illuminator. When using low-

32 mm. 1(5 mm.

Fig. 6. Short mounted achromatic objectives.

4 mm.

power lenses it might also be screwed to the lower end of the objective, thus control-
ling the light returned by the object before entering the objective.

Specimen Holders. In order to examine a piece of metal under the microscope,
it is of course necessary that the polished and otherwise prepared surface be held in
a plane accurately perpendicular to the optical axis of the instrument. This may be
accomplished by so shaping the sample that it will have two sides exactly parallel,
and preparing one of them for microscopical examination. This operation, however,
is at best tedious and laborious, and metallographists have endeavored to replace it
by the use of more or less ingenious devices for holding the specimens in the proper

Fig. 7. Specimen holder.

position. Some embed their samples in wax or in some other plastic material, while
others have recourse to stages provided with special leveling devices.

The simple holder shown in Figure 7 gives better satisfaction, requiring no mount-
ing whatever of the samples. The specimen, no matter how irregular in shape, is
held firmly in place by a rubber band and the holder placed on the stage like an
ordinary slide. If the correction of the objective demands it, a cover glass may be
inserted between the sample and the holder. It will be apparent that the required
manipulations are very simple and quickly performed.

In the case of specimens smaller than the opening of the holder, however, the use
of a cover glass is necessary to hold them in place. This is objectionable, at least

8 APPARATUS FOR THE METALLOGRAPHIO LABORATORY

when using high-power objectives, which should be corrected for uncovered objects.
To overcome this difficulty, a little templet may be used having a triangular opening
and inserted between the specimen and the holder (Fig. 8). This templet is made
very thin so as to permit the use of high-power objectives, which must be brought
very close indeed to the object. It will also be noticed that one side of the upper part
of the holder has been removed, exposing to .view a larger portion of the sample and
permitting a more ready approach of high-power objectives. Large samples are, of
course, placed in the holder without any templet.

A still simpler and more effective device can be used to hold in place samples of
iron and steel and other magnetic substances. The device consists of a V-shaped

(a)

(6)

Fig. 8. (a) Specimen holder and large sample.

(6) Specimen holder, templet, and small
sample.

permanent magnet of special steel about 1 inch wide and 2J^ inches long (Fig. 9).
This little magnet is placed on the stage of the microscope like an ordinary glass slide
(Figs. 4 and 5) and tne sample to be examined suspended to it from below, being held
in place by the attraction of both poles. Small samples are suspended near the small
end of the V-shaped opening, while larger ones are placed nearer the wider end of the
opening. This holder, therefore, is universal in its application within the limits of
samples of suitable size for microscopical examination. If the opening of the stage
be sufficiently large, say \ l / inches or more in diameter, the magnet may be kept per-
manently on the stage, as the samples may then be readily removed or attached to
the magnet with the fingers from below the stage. This adds so much to the conve-
nience of this little device that it is strongly urged, in case the central aperture of the
stage is too small, to have it suitably enlarged. The magnet is kept in place, like any
glass slide, by the clips of the microscope and, also like any glass slide, may be moved
about for the inspection of the different parts of the specimen. The side of the magnet

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

resting on the stage having been ground perfectly flat, it will be evident that the sur-
face of the sample under examination will always be accurately in the proper position,
permitting the use of high-power objectives without fear of difficulty arising from ever
so slight an inclination of the sample.

When used in connection with a mechanical stage (Fig. 5) the convenience of this
little holder becomes still more apparent and its usefulness is further increased. It
then affords, moreover, a ready means for the repeated examination of the same spot
of any sample at any time. To that effect the holder is laid upon the prepared surface
and two scratches made by drawing a needle across the specimen along the sides of

Fig. 9.

(a)

(6)

\" / v,

- (a) Magnetic specimen holder
(b) Scratched specimen.

the V-shaped opening, as shown in Figure 9. When it is desired to examine the sam-
ple, the latter is suspended to the magnet so that the needle markings coincide closely
with the sides of the magnet opening, in this way securing a permanent position for
the sample. The position of the magnet itself is controlled, in the usual way, by means
of the graduating devices of the mechanical stage.

Finally, by placing the sample below the stage and bringing the prepared surface
on a level with the stage, considerably greater working distance is secured, a gain
which has its importance.

The only limitation of this holder is due to the fact that with very small speci-

Fig. 10. (a) Magnetic holder.

(b) Steel templet.

mens it is impossible to use high-power objectives (^ inch or less in focal length),
because the mounting of the objective comes in contact with the sides of the magnet
and prevents the focusing of the object. For the examination by high-power objec-
tives, the use of a very thin steel templet (not over 0.01 inch thick) of the same dimen-
sions as the magnetic holder, but with a V-shaped opening considerably narrower
(Fig. 10) is recommended. This templet is placed over the magnetic holder so as to
exactly cover it, thereby becoming magnetized. The small iron and steel samples
are suspended to this thin steel plate in the usual way and may then be examined
with the highest power objectives.

10 APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Universal Metalloscope. The instrument shown in Figures 11 to 15 was devised
especially for the ready examination of large iron and steel samples, but it will be ap-
parent that it can be used with equally satisfactory results for small samples both
opaque and transparent.

The microsope stand proper consists of a microscope-tube, provided with both
coarse and fine adjustments, and with a draw-tube, rigidly mounted on a bar supported
at both ends on substantial and firm cast-iron legs. 1 The height between the table
and the under side of the supporting bar is 5 inches and the distance between the sup-
porting legs 12 inches.

This arrangement affords free space below the objective for the examination of

Fig. 11. Universal metalloscope: stand, eye-piece, vertical illuminator, objec-
tive, electromagnetic stage, and rail section.

large specimens of metals, such as full rail sections, without detracting in the least
from the value of the instrument when applied to the examination of the usual small
specimen, as explained later. Many metal microscopists frequently have to examine
bulky specimens, and this is altogether impossible with the ordinary microscopes as
well as with the special metallurgical microscopes which have been designed and de-
scribed from time to time.

Recourse must be had to all sorts of makeshifts for the proper support of large
specimens, or, more often, the microscopist gives up the attempt altogether, or else
resigns himself to the cutting of the bulky samples into small pieces to be laboriously
polished and separately examined.

It is believed that an instrument permitting the examination of large as well as

1 In a more recent model the supporting bar is mounted on three legs, permitting the ready lev-
eling of the instrument.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

of small specimens with equal ease and accuracy will be welcomed by metallographists,
and that it will lead to more frequent examinations of full sections of metal imple-
ments, a departure which should bring fruitful results.

Electromagnetic Stage. The perplexing question of the proper support, for mi-
croscopical examination, of iron and steel specimens of all sizes and shapes has been
effectively solved by the use of the electromagnetic stage illustrated in Figure 11.
This stage consists of a steel plate 7 by 14 inches having a V-shaped opening, and con-
verted into a powerful electromagnet by means of two bobbins with solenoids sur-
rounding the arms of the steel plate, as clearly shown in the illustration. Electrical
connection is readily made with any suitable current, and the use of an incandescent
lamp in series provides in a simple way the necessary outside resistance to prevent
heating of the solenoids. Large specimens of iron and steel, such as rail sections,
A, Figure 12, are firmly held in an accurate position by the attraction of the mag-
netic stage, the extremities of the flange only and a narrow space on each side of the

"ih

IA1

ABC
Fig. 12. (A) Electromagnetic stage and rail section. (B) Electromag-
netic stage, templet, and medium-size specimen. (C) Electromag-
netic stage, two templets, and small specimen.

head being hidden from view. The size and shape of the stage-opening make pos-
sible the ready support of specimens measuring from 2 to 6 inches in their greatest
dimension.

Templets for the Examination of Small Specimens. For the examination of iron
and steel samples from 2 inches in length down to the smallest dimensions, a steel
templet, also with a V-shaped opening, is placed on the stage, shown at B, Figure 12.
This templet through its contact with the stage becomes strongly magnetized and the
specimens to be examined are suspended to it.

For the examination of very small specimens with high-power lenses the thickness
of this templet would prevent the necessary close approach of the objective. To
make this approach possible a very thin steel templet (not exceeding 0.01 inch thick)
is used, shown at C, Figure 12, which makes possible actual contact between a high-
power objective and the smallest specimen.

Support of Non-Magnetic Specimens. For the support of non-magnetic speci-
mens, such as non-ferrous metals, rocks, cement, etc., a very simple device is provided,
consisting of two crossbars and rubber bands, which is readily attached to the stage
and by means of which the non-magnetic specimens, as well as the templets when
needed, are firmly held in place regardless of their size or shape.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Leveling- Devices of Stand and Stage. It is, of course, essential, especially when
using high-power objectives, that the optical axis of the microscope be accurately
perpendicular to the surface under examination. To secure this result both the stand
and the stage are provided with leveling-screws, as shown in Figure 11. For leveling
the stage a small spirit-level may be placed upon it, or better, upon the sample under
examination, and the necessary adjustment quickly made. For leveling the micro-
scope stand the eye-piece should be removed, the small level placed on top of the tube,
and the leveling-screws adjusted. By placing the instruments on a table or desk
having a smooth and flat top, it is evident that, barring accidents, the stand and
stage will remain indefinitely accurately leveled.

Motion of the Stage. In order to examine the entire surface of a large specimen
it is necessary to bring in turn within the field of the microscope the different portions

Fig. 13. Back leg of electromagnetic stage and sliding plate.

of the specimen, and this necessitates the moving of the stage in various directions.
The weight of the stage, however, would create considerable friction between the legs
and the supporting table, making the sliding motion jerky and otherwise unsteady.
To overcome this difficulty the back leg of the stage is provided with a small wheel
running in a groove cut in a small brass plate fastened to the table or desk, shown in
Figure 13. The mounting of the wheel is provided with a pivot fitting snugly into a
hole in the leg. This construction makes possible the ready back-and-forth motion
of the stage, as well as its free circular displacement around the axis of the back leg
thus permitting to bring quickly any desired portion of the object under the objec-
tive. As the bulk of the weight is supported by the back leg, the arrangement makes
possible a very steady and smooth motion of the stage.

Mechanical Stage. The use of a mechanical stage is often highly desirable. This
is taken care of in the present instrument in two different ways: (1) by the use of a
mechanical stage suitably attached to the electromagnetic stage, and (2) by the use
of a mechanical stage independently mounted on a separate base of the usual horseshoe
pattern.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

The first method is illustrated in Figure 14. A mechanical stage of usual construc-
tion is screwed on a brass plate provided with two small pins fitting two correspond-
ing holes in the magnetic stage, thus securing a firm and constant position for the
mechanical stage. When using a mechanical stage, however, a rigid and constant
position should also be secured between it and the microscope stand. To that effect
a brass plate is provided, with recesses to receive the back legs of the stand as well
as the front legs of the stage, shown in Figure 14. It is then possible at any time to
place the microscope stand and the stage in exactly the same relative positions.

The second method consists in the use of a mechanical stage separately mounted
on an ordinary horseshoe base, shown in Figure 15. To secure a constant relative

Fig. 14. Universal metalloscope: electromagnetic stage with mechanical
stage, magnetic specimen holder, small specimen, and base-plate.

position between stand and stage, the foot of the latter fits into recesses provided for
that purpose in the base-plate.

The use of this independently mounted mechanical stage offers the additional
advantage resulting from the vertical up-and-down racking of the stage, rendering
unnecessary any vertical adjustment of the light and condenser, as well understood
by metallographists.

Examination of Transparent Objects. To adapt the universal metalloscope to
the examination of transparent objects, thereby converting it into an ordinary micro-
scope, or, if desired, into a petrographical microscope, a separate stage on horseshoe
base should be used, as shown in Figure 15, when the necessary Abbe condenser,
analyzer, polarizer, etc., can readily be attached. The instrument is then in no
way inferior to high-class microscopes for examination by transmitted or polarized
light.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

ILLUMINATION OF THE SAMPLES

Opaque objects such as metals and alloys must necessarily be examined by re-
flected light, i.e. by light thrown upon them from above the stage, their treatment
differing in this respect from that of other microscopic preparations, which are gen-
erally examined by transmitted light, i.e. by light sent through them and proceed-
ing from below the stage.

With the low-power objectives there are two possible ways of illuminating opaque
specimens: (1) by directing the light obliquely upon the object, and (2) by causing
the light to fall perpendicularly upon it by means of so-called "vertical illuminators."

Fig. 15. Universal mctalloseope : mechanical stage on
horseshoe base, magnetic specimen holder, small speci-
men, and base plate.

With medium-high and high-power objectives the second method only is possible,
because the distance between the specimen and the front lens of the objective is now
so small that obliquely reflected light cannot reach the surface under examination.
With very low-power objectives i.e. having a focal length of one inch or more
the vertical illuminator may be placed between the lens and the object; but with
higher power objectives it must of course be inserted between the objective and the
microscope tube, the objective then acting as a light condenser and increasing the
intensity of the illumination.

Oblique illumination may be obtained (a) by allowing daylight or artificial light
to fall freely upon the object; (6) by directing the light upon the object by means of
mirrors, reflectors, or condensers; (c) by the use of a " lieberkiihn " ; and (d) by the
use of a "parabolic reflector."

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Vertical illumination may be produced (a) by means of an opaque reflector con-
sisting of a totally reflecting prism or of a mirror covering only a portion of the ob-
jective, the light returned by the object reaching the eye by passing through the
uncovered portion; and (6) by means of a transparent reflector, generally a plain
glass disk or glass square, reflecting upon the object a portion of the incident light
and permitting the passage of a portion of the light returned by the object, which
thus reaches the eye.

When a highly polished surface is examined by obliquely reflected light, since the
angle of reflection is equal to the angle of incidence, the totality of the light is reflected
outside the objective (Fig. 16) and the object appears uniformly dark. In case the
metallic specimen contains some portions duller in appearance,~~these will scatter a
certain amount of light a part of which will enter the objective (Fig. 16), and those
portions will therefore appear brighter. A similar effect is produced when the speci-

() (ft) (c)

Fig. 16. (a) Oblique and vertical illuminations of bright surface.
(6) Oblique and vertical illuminations of dull surface.
(c) Oblique and vertical illuminations of hills and valleys.

men, instead of being perfectly flat, contains microscopic hills and valleys, the sides
of which may be so inclined as to reflect some light into the microscope (Fig. 16),
consequently appearing bright. Viewed by oblique light, therefore, the relative dark-
ness or brightness of a constituent will vary inversely with its true appearance and
will also depend upon its orientation, since this will affect the angle of incidence of
the light striking it. Generally speaking, the darker a constituent the brighter will
it seem to be when illuminated by oblique light, the latter yielding, so to speak, a
negative image. Oblique illumination, moreover, cannot be made as intense as ver-
tical illumination and, as already explained, is possible only with low-power objectives.
For these and other reasons, while it is not without value, it is only used occasionally
by metallographists.

To increase the intensity of oblique illumination and to make its use possible with
somewhat higher powers, such appliances as the "lieberkiihn" and the parabolic
reflector have been used. The "lieberkuhn," so called from the name of its inventor,
consists of a small concave mirror attached to the objective and reflecting upon the
object some light proceeding from below the stage and passing around the object.
It will be evident that only small size objects can be thus illuminated.

JO APPARATUS FOR THE METALLOGRAPHIC LABORATORY

The parabolic reflector (Fig. 17), first constructed by Messrs. Beck of London for
Dr. Sorby, consists of a parabolic mirror placed on one side between the objective and
the object and condensing the incident light upon the latter. It should be attached
to the objective. Dr. Sorby later added a silver mirror in the shape of a half disk to
the same mount, so as to be able to obtain at will vertical and oblique illumination
when using low-power objectives (Fig. 17). When vertical illumination is desired, the
small mirror is swung over the objective, covering only a portion of it, and directing
vertical rays of light upon the object. This combination is known as the Sorby-Beck
reflector.

The effects of a vertical illumination are precisely opposite to those of an oblique
illumination, as clearly shown in Figure 16, highly polished surfaces reflecting the to-
tality of the light into the objective, while dull ones appear dull because they reflect
most of the light outside.

To produce a vertical illumination we have the choice between an opaque or a

() (6)

Fig. 17. (a) Parabolic reflector.

(6) Sorby-Beck parabolic reflector.

transparent (glass) reflector. The opaque reflector consists of a totally reflecting
right-angled prism, or of a mirror placed between the microscope tube and the objec-
tive and covering only a portion (generally about one half) of its aperture. The beam
of light enters the illuminator through a side opening provided for that purpose and
is reflected downwards by the reflector, being condensed upon the object by the lenses
of the objective itself. The light sent back by the object reaches the eye through the
uncovered part of the objective.

The first vertical illuminator was designed by Professor Hamilton L. Smith of
Hobart College, Geneva, N. Y., and consisted of a small annular silver mirror
(Fig. 18), forming an angle of 45 with the axis of the microscope, the light reflected by
the object passing through the central opening on its way to the eye. Semi-circular mir-
rors, similarly mounted and partially covering the objective (Fig. 18), have been used
with equal satisfaction, and the author has obtained good results with a very small
central mirror (Fig. 18) suitably mounted, reflecting the light upon the central por-
tion of the objective lenses, and permitting the returned light to reach the eye
through the free space surrounding the mirror.

Instead of a mirror, a totally reflecting right-angled prism may be used as shown

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

in Figures 18 and 19, covering half of the aperture of the objective. The prism is
so mounted that it can he rotated around its horizontal axis, this being needed in order
to secure the best illumination of the sample. Nachet, of Paris, provides his prism
with an additional motion permitting it to cover a greater or smaller portion of the
objective. These reflecting prisms are now used much more than the reflecting mirrors.

(6) (c) (rf)

Fig. 18. (a) Annular mirror.

(6) Semi-circular mirror.

(d) Totally reflecting prism.

(e) Plain glass disk.

In 1874 Nachet constructed for the International Commission of the Meter some
objectives provided with totally reflecting prisms as permanent parts of their mount-
ings. In low-power objectives a prism was placed above the first lens (Fig. 20), while
with higher power objectives it was necessarily inserted above the double or triple
lens system. These objectives are called illuminating objectives. This arrangement,

A B

Fig. 19. Vertical illuminator. Totally reflecting prism. Zeiss.

however, has not been found very satisfactory and with one notable exception is sel-
dom used by metallographists.

In vertical illuminators having a transparent reflector, the latter consists of a plain
glass disk covering the whole of the aperture of the objective (Figs. 18 and 21). The
incident light is in part reflected upon the object, while another portion passes freely
through the glass reflector. A part of the light returned by the object is again re-
flected by the glass illuminator, while another portion passes through it and thus reaches

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

the eye. The glass reflector is so mounted that it can be rotated around its horizontal
axis (Fig. 21). The amount of light permitted to enter the illuminator may be regu-
lated by an iris diaphragm attached to the side opening or independently mounted
and placed between it and the source of light, or by a revolving sleeve attached to the
illuminator and provided with different size openings. The first plain glass illumina-
tor was constructed by Mr. Beck of London.

With very low-power objectives it is preferable to place the vertical illuminator

Fig. 20. Nachct illuminating objectives.

between the objective and the object, attaching it to the former in some suitable way,
as for instance by providing the lower end of the objective with a society screw (see
Fig. 6).

While the author is well aware that some metallographists of note prefer the prism
to the plain glass type of vertical illuminator, in his opinion the plain glass reflector is
greatly superior. While the illumination obtained by its use is not quite as intense,
it is certainly more uniform and less liable to produce a distortion of the image.

An improved construction of the plain glass vertical illuminator is illustrated in

Fig. 21. Bausch and Lomb plain glass disk vertical illuminator.

Figure 21. The glass reflector is inserted into a brass ring which on the side opposite
the milled head is screwed into the wall of the brass mounting, practically doing away
with the frequent breaking of the glass and greatly facilitating its cleaning. The milled
head is large, which makes it possible to impart a more delicate motion to the glass
reflector.

SOURCES OF LIGHT AND CONDENSERS

The illumination of opaque objects such as metals and alloys requires an intense
source of light, especially for their photography. Daylight and ordinary gas or oil
flames should be discarded as not suitable for the purpose, the sources of light which

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

have been found most satisfactory being, in the order of their excellence, intensity,
and decreasing cost: (1) the electric arc lamp, (2) the Nernst lamp, and (3) the Wels-
bach gas lamp.

The Welsbach lamp outfits (Figs. 22 and 23) are very inexpensive and quite satis-

Fig. 22. Welsbach lamp and double-convex condensing lens.

factory for visual examination by low- and medium high-power objectives. In
taking photomicrographs, however, their lack of intensity necessitates very long
exposures, while with high-power objectives the light received upon the camera screen

Fig. 23. Welsbach lamp and bull's eye condenser.

is so faint as to render proper focusing of the object a very difficult, if not impossible,
operation.

Two kinds of electric arc lamps are now supplied, one with large carbons (Fig. 24)
and a smaller one with carbons measuring only J/ inch in diameter (Fig. 27). The
carbons should be placed at right angles, as this arrangement directs the maximum
amount of light into the condensers. The positive or horizontal carbon should be
cored and larger than the vertical or negative carbon. While automatic feeding of

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

the carbons (Fig. 26) is a valuable feature, it is not by any means essential, as
remarkably effective hand-feed lamps are now constructed by which a very steady
light can be maintained (Fig. 25). Automatic mechanisms, moreover, are liable to
get out of order and occasional sudden shiftings of the light are difficult to entirely
eliminate.

The large carbon lamp yields, of course, by far the most intense illumination and

Fig. 24. Large arc lamp outfit.

is the only one suitable for direct projection of metallic samples upon a screen for public
exhibition. In taking photomicrographs with the large arc lamp the needed exposures
are often instantaneous and seldom exceed five or, at the most, ten seconds. The
lamp consumes from fifteen to twenty amperes.

The small arc lamp (Fig. 27) is very satisfactory for visual examination and is,
of course, much less expensive. It, however, requires longer exposures when photo-

Fig. 25. Hand-feed arc lamp.

Fig. 26. Automatic feed arc lamp.

graphing. The position of the carbons can be regulated with great nicety by 'inde-
pendent adjustments, thus securing a very uniform light. The lamp consumes about
five amperes.

The Nernst lamp (Fig. 28) is used successfully by many microscopists and undoubt-
edly affords a very satisfactory illumination both for visual examination and for pho-
tomicrography. In taking photographs, exposures of ten seconds or more are needed,
according to the magnification and the character of the specimen.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Summing up, if we desire a cheap and convenient form of illumination for visual
examination with objectives of low and medium high power, the Welsbach lamp will
be found in every way satisfactory; while for the taking of photomicrographs and for
examination by high-power objectives the electric arc lamp and the Nernst lamp
should be recommended, bearing in mind that the large arc lamp will yield light of

Fig. 27. Small electric arc lamp, bull's eye condenser, and rheostat.

greatest intensity but will, on the other hand, be much more costly. When neither
gas nor suitable electric current are available, an acetylene lamp should be used,
provided tanks of acetylene gas can readily be obtained.

Condensers. Some kind of condensing attachment must be placed between the
source of light and the vertical illuminator so that a large portion of the light may be
utilized and a beam of suitable size directed into the illuminator. In the case of light

Fig. 28. Nernst lamp and special hull's eye condenser on adjustable

supports.

proceeding from a luminous point or at least from a small luminous area, as for instance
with the electric arc, at least two lenses or systems of lenses are needed, one system,
PL and ML (Fig. 29), placed near the so'urce of light, to collect the divergent rays
and convert them into a parallel beam, and a second lens CL placed at some distance
from the first, to convert the parallel beam into a converging one. The vertical illu-
minator should be located at such a distance from the condensing lens that the beam

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

of light will cover a little more than the opening through which it enters the illumi-
nator. A glass cooling cell CC, filled with distilled water or some other suitable liquid,
should be placed between the two lenses in order to absorb a large amount of heat and
thereby prevent injury to the objective. An iris diaphragm, /, should also be used to
control the amount of light entering the vertical illuminator. This diaphragm should
be placed in front of the converging lens and should be provided with clips for holding
ground and colored glasses. These various parts should be mounted on a so-called
"optical bench" B upon which they can slide.

With a large luminous body such as the Welsbach mantle, a single double-convex
lens (Fig. 22) or a bull's eye condenser (planoconvex) (Fig. 23) is sufficient to collect
and condense the necessary amount of light. It should, of course, be placed at the
proper distance both from the vertical illuminator and from the source of light. The
use of an iris diaphragm attached to the lens or on a separate mount is advisable,
since it affords a ready means of controlling the amount of light admitted into the
illuminator.

Monochromatic Light. The different lamps described above all yield, of course,

CL CC PL ML

Fig. 29. Condensing lenses, cooling cell, iris diaphragm, automatic shutter,

and optical bench.

white light, and since the correction even of apochromatic objectives for chromatic
aberration is never perfect, it is evident that the use of monochromatic light i.e.
light of one wave length is preferable, especially for photographing. Monochromatic
light may be obtained in two ways: (a) by using a source of light actually monochro-
matic, and (b) by causing white light to pass through colored glass screens or colored
solutions (light filters), preventing the passage of some undesirable rays. The mercury
arc lamp yields a nearly monochromatic light and has been tried by Le Chatelier with
satisfactory results. It seems more convenient, however, when monochromatic light
is wanted, to use light filters of suitable colors, in which cat,e colored glass screens will
be found easier to manipulate than glass cells containing colored solutions.

PHOTOMICROGRAPHIC CAMERAS

For taking photomicrographs, a light-tight connection should be established be-
tween the microscope and camera, both instruments being placed vertically or hori-
zontally (Figs. 30 and 31). Whether to use the camera in a vertical or horizontal
position is a debatable and debated question. Among well-known metallographists,
Le Chatelier and Martens prefer the horizontal while Osmond and Stead have a strong

APPARATUS FOR THE METALLOGRAPHIC LABORATORY 23

Fig. 30. Photomicrographic camera (vertical position), showing metal-
lurgical microscope, mechanical stage, automatic shutter, etc.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

I
I

-a

C o

8'S

-4_J

bi.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

preference for the vertical position. The latter certainly affords greater stability and
eliminates all danger of heavy specimens slipping while being photographed. It is
often said that the vertical position is inconvenient because of the necessity of mount-
ing on a stool for focusing the image on the screen, but the objection appears trifling.
It is also argued by some that the microscope should be used in a horizontal position
because this makes it possible for the operator to sit while performing his work. This
objection also appears to have little weight.

By the use of a large, totally reflecting prism, however, connecting the microscope

Fig. 33. Plan view of metallurgical microscope and photomicrographic camera joined by

totally reflecting prism; large arc lamp outfit and Welsbach outfit No. 5.
Legend: C = camera.

B = camera base in reversed position.

F = Fine adjustment of microscope.

F'= Pulley for controlling F from the position, 1, of the observer at the camera.

L = Arc lamp outfit.

1 = Iris diaphragm.

P = Horizontal vertical attachment.
S = Automatic shutter with iris diaphragm.

W = Welsbach lamp arranged for visual examination at same microscope.
D = Condenser for Welsbach lamp.

2 = Position of observer at microscope.
A. = Adjustment for arc lamp carbons.

and camera (Figs. 32 and 33), it is possible to maintain the microscope vertical, un-
doubtedly the best position, while using the camera horizontally. A thread belt con-
nects the fine adjustment of the microscope with a pulley mounted on a standard
at the end of the camera bed bar. This enables the operator to control the fine ad-
justment easily from his position at the camera screen, while the placing of the arc
lamp on the same side of the microscope as the camera makes it possible for him to
reach with his right hand the various adjustments of the lamp, thus securing max-
imum intensity and uniformity of illumination, two points so essential in taking
photomicrographs.

For visual examination, the front standard of the camera to which the special prism

APPAKATUS FOR THE METALLOGRAPHIC LABORATORY

Fig. 34. Universal metalloscope, Nernst lamp outfit, and vertical camera.

Fig. 35. Universal metalloscope, arc lamp outfit, large totally reflecting
prism, and horizontal camera.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

is attached is moved along the bed bar to a sufficient distance. To bring the camera
into position for photography the front standard is brought towards the microscope
until the collar beneath the prism rests upon the eye-piece of the microscope, the focus
of the microscope not being changed in the least. The bottom of this collar which is
fitted with a felt rim rests directly upon the eye-piece without exerting any pressure,
but tightly enough to make a light-tight connection. The image is then focused upon
the camera screen by the fine adjustment alone and the exposure quickly made.

The universal metalloscope already described is seen in Figure 34 with vertical
camera and Nernst lamp and in Figure 35 with horizontal camera, totally reflecting
prism and large arc lamp.

INVERTED MICROSCOPES

Le Chatelier was the first to suggest the use of an inverted microscope for the ex-
amination of metallic surfaces (see Appendix). In this style of microsope the stage
is placed horizontally above the objective, the latter being necessarily pointed up-
wards (Figs. 36 to 38).

In the inverted type of microscope and photographic attachment illustrated in
these pages, it has been attempted to simplify the construction with corresponding
material decrease in price.

The microscope is permanently connected with the camera by a totally reflecting
prism P (Fig. 37) set rigidly below the vertical illuminator. A separate tube set at
right angles to the first is provided for visual examination, another totally reflecting
prism P' (Figs. 37 and 38) being fastened to the inner end and serving to reflect the
image from the body tube through the eye-tube to the eye. When a photograph is
to be taken this prism P' is simply withdrawn from the field by means of the draw
tube. The eye-tube is fitted with pin and slot which mark the limits to which the
small prism P' may be pushed in and withdrawn, so that the vertical illuminator being
once set, the only adjustment necessary is at the arc lamp. With the Nernst and Wels-
bach lamps, after the light and the vertical illuminator are once set, no more adjust-
ments are necessary. The two totally reflecting prisms need never be rotated and in
fact cannot be moved, except for the sliding motion of the prism P' as already described.

The stage, which is revolving and provided with centering screws, is of course
equipped with both coarse and fine adjustment, and a mechanical stage may readily
be substituted for the plain stage.

With this inverted microscope the use of a magnetic holder will also be found very
convenient, for the sample is then held firmly in place instead of resting loosely on
the stage, thereby increasing the usefulness of the mechanical stage.

The placing of the light on the same side of the microscope as the camera makes
it possible for the operator to regulate his illumination while focusing the object on
the camera's screen.

POLISHING APPARATUS

Hand Polishing. When, in spite of the length and laboriousness of the operation,
iron and steel samples are to be polished by hand, four smooth blocks of wood should
be prepared, some 6 by 10 inches and 1 inch thick. One of these should lie covered
with canvas or linen duck and the others with fine broadcloth. The blocks are to be
used as described in Lesson III, page 5.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

2!)

3
O

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

Polishing by Power. The power polishing machine shown in Figure 39 has been
found very satisfactory. It consists of a heavy iron pedestal upon which is mounted
a grinder having emery-wheel and cast-iron disks revolving in a vertical plane, thus
giving four polishing surfaces of graduated fineness. The polishing powders mixed
with water are applied to the various disks by means of brushes, and shields are pro-
vided to catch any surplus water that may be thrown off during the operation. Should
a cloth become worn or torn it is readily and quickly replaced. This machine very
much shortens the time necessary for the preparation of samples and is far superior
to those where only one block is made to rotate at a time. A speed of 1200 revolutions
per minute has been found most satisfactory for polishing iron and steel samples, but

Fig. 37. Fig. 38.

Fig. 37. Inverted metalloscope. Vertical section, front view.
Fig. 38. Inverted metalloscope. Vertical section, end view on AB (Fig. 37).

L = Source of Light.

R = Vertical illuminator reflector. 1

P' = Totally reflecting prism which reflects image into the eye-tube when latter is pushed
in.

P = Totally reflecting prism which reflects image into camera when eye-tube is pulled out.

S = Metalloscope stage.

O = Specimen.

by the use of a variable speed electric motor to run the polishing machine various
speeds may be readily obtained.

The polishing motor shown in Figure 40 possesses the advantages of directly driven
over belt driven machinery. It is provided with the same polishing disks as the pol-
ishing machine and can be built both for constant and for variable speed.

The operation of polishing with these machines is described in Lesson III, page 6.

PYROMETERS AND ELECTRIC FURNACES

Pyrometers. The Le Chatelier thermo-electric pyrometer is undoubtedly the
instrument best adapted to the measurement of temperatures needed to control such
heat treatments as are likely to be performed in a metallographical laboratory. The
thermo-couple consists of a wire of pure platinum and of a wire of platinum alloyed
with 10 per cent of rhodium or iridium. To measure the electromotive force created

1 A reflecting prism may of course be used if preferred.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY 31

Fig. 39. Power polishing machine.

Fig. 40. Polishing motor.

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

by heating the couple the author does not know of any more accurate and reliable
instrument than the galvanometer constructed by Siemens and Halske (Fig. 41).

Fig. 41. Siemens and Halske galvanometer.

Pt. Rh.
Fig. 42. Saladin self-recording thermo-electric pyrometer.

In its latest form it has a resistence of 400 ohms, has a scale of 180 divisions, each
one corresponding to 10 microvolts and a second graduation giving the temperature

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

din>ctly for the couple sold with the apparatus. The use of cheaper couples and
cheaper galvanometers is not to be commended for they are unsuitable for accu-
rate scientific work and uncertain even for controlling industrial operations.

An autographic recording pyrometer is very useful and quite indispensable for
the detection of faint evolutions or absorptions of heat. Indeed without its use there

Fig. 43. Le Chatelier-Saladin self-re-
cording thermo-electric pyrometer.

Fig. 44. Le Chatelier-Saladin self-recording thermo-electric pyrometer.

are many delicate thermal treatments that could not be performed. Several auto-
graphic instruments are now constructed. To meet the needs of the metallographist
the author believes that the pyrometric recorder designed by Le Chatelier and Sala-
din and constructed by Pellin of Paris (Figs 42 to 45) will be found most satisfactory.
In an early form the different parts were arranged as shown in Figure 42. The light
proceeding from the source S after passing through a lens is received by the mirror
of a sensitive galvanometer Gi the deflections of which measure the difference in

APPARATUS FOR THE METALLOGRAPHIC LABORATORY

temperature between the sample under examination and the neutral body. This
horizontal deflection of the beam of light is converted into a vertical deflection by
passing through a totally reflecting prism M placed at an angle of 45 deg.

This vertically moving beam of light is received by the mirror of a second gal-
vanometer Gi whose deflections are a measure of the temperature of the sample. The

\ For ordituzrv

G. ) *

2~ -- . . .'. . ''////' /, ' ' ' '/ '/ -'" ' '' '

^J Laourup Curves

7 .^w^w.i r ,.i 0% ,X M^SM^I

IV10%lr. <

7 : ; . -, ' -\ N 1 " ^SS *' ' x ^l%

.^ ^ Pt

C H

''- , / ',;/ -/, , ' >, ;/

Sensitive
Galvanometer

Fig. 45. Roberts-Austen method of connecting the sample, neutral body,
and galvanometers.

Fig. 46. Herseus electric muffle furnace.

beam then passes through a lens and reaches the screen or plate P. L is a lens at t he-
conjugate foci of which are placed the mirrors of the two galvanometers. Two
motions are in this way imparted to the spot of light, (1) a horizontal motion pro-
portional to the temperature of the sample and (2) a vertical motion proportional to
the difference in temperature between the sample and the neutral body. The result-
ing curves are known as differential curves (see Lesson VII, pages 10 and seg.)

In recent models the apparatus has been simplified and made more compact

APPARATUS FOR THE METALLOGRAPHIC LABORATORY 35

(Figs. 43 and 44) by placing the galvanometers so near each other that the lens L,
Figure 42, could be omitted and the entire instrument placed in a metallic or wooden
case (Fig. 44.)

The connections between the sample and the neutral body and between these
and ftie galvanometers are made, as first suggested by Roberts-Austen, and as clearly
shown in Figure 45. In this illustration galvanometer GI corresponds to galvanom-
eter G-2 of Figure 42 and galvanometer Gz to galvanometer GI.

ELECTRIC FURNACES

For the experimental heat treatment of small iron and steel samples an electric
resistance furnace is extremely useful. For such purposes the Heraus platinum
wound muffle furnaces (Fig. 46) are very satisfactory. The furnaces are provided
with a safety device to take care of any overload and with an internal adjustable
rheostat. They are made in different sizes, the inside measurements of the largest
type being nine inches in length, six inches in width, and three and one half inches in
height. Using 110 volt current they consume from five to fifteen amperes according
to size and temperature. If care be taken never to use temperatures exceeding 1000
deg. C. for any great length of time, and, preferably, never to exceed 1000 deg., the
furnace will be found very durable. The use of cheaper furnaces wound with base
metals is a doubtful economy as frequent rewindings are required while the maxi-
mum temperature that can be safely used with such furnaces is fully 200 deg. C.
lower than with platinum resistance furnaces.

LESSON I

PURE METALS

Microstructure. When a properly prepared sample of a pure metal is examined
under the microscope, the revealed structure generally presents the appearance of a
polygonal 1 network (Figs. 1 and 2), an indication that the metal itself is composed of

Fig. 1. Pure gold. Cast.
Magnified 50 diameters. (Andrews.)

Fig. 2. Pure copper.

Magnified 8 diameters.

(Houghton.)

irregular polyhedral 2 grains, each mesh or polygon of the network representing a
section through a polyhedron. 2

Crystallization. When a substance passes from the liquid to the solid state, the
process of solidification is generally accompanied by crystallization, i.e. the molecules
of the substance so arrange themselves as to form small solids of regular geometrical
outlines, such as cubes, octahedra, 3 etc. Each of these spontaneously formed sym-
metrical solids is called a crystal and any substance made up of crystals is said to be
crystalline.

Crystals possess the property of breaking more readily in one or more directions.
This property is called "cleavage." The planes of cleavage or direction of easy rup-
ture are generally parallel to the faces of the crystals. A cubic crystal, for instance,

1 A polygon is a closed geometrical figure with straight sides (necessarily three or more).

2 A polyhedron (plural, polyhedra or polyhedrons) is a closed geometrical solid bounded by
plane (smooth) faces (necessarily four or more).

3 An octahedron (plural, octahedra or octahedrons) is a geometrical solid (a polyhedron) bounded
by eight plane faces.

2 LESSON I PURE METALS

splits readily in three planes parallel to the three sets of faces of the solid. In Figure 3,
ABC, DEF, and GHI indicate the cleavage planes of a cubic crystal. The direction of
its cleavage planes constitutes the orientation of the crystal.

Solid substances which are not crystalline are said to be "amorphous." They are
characterized by the absence of any symmetrical grouping of their molecules. Glass
is a good example of an amorphous substance.

Idiomorphic Crystals. When the fluidity of a substance and other conditions are
such that the formation and growth of the crystals are given free play, perfect (and

Fig. 3. Cleavage planes.
(Mellor.)

Fig. 4. Crystallization from
centers. (Desch.)

sometimes very large) crystals are produced. These perfect crystals, with faultless
geometrical outlines, perfect cubes for instance, are called "idiomorphic" crystals.

Allotrimorphic Crystals. When the free development of crystals is hindered by
less favorable crystallizing conditions, such for instance as collision or contact with
other crystals, likewise in process of formation, the regular external form is not pre-
served and the resulting imperfect crystals are called " allotrimorphic " crystals, also,
but more seldom, "anhedrons" or faceless crystals. Such crystals are said to have
taken their shape from their surroundings. It should be noted, however, that allotri-
morphic crystals, like idiomorphic crystals, are composed of crystalline matter. An
allotrimorphic crystal may be regarded as resulting from the mutilation of an idio-
morphic crystal, the mutilation affecting the external shape only, and not the crystal-
line character of the substance. 1

Crystallization of Metals. Metals when they solidify generally give rise to the
formation of allotrimorphic crystals. The explanation for this is to be found in the
fact that crystallization sets in simultaneously at many different centers (see Fig. 4).
From each center a crystal grows through successive addition of crystalline matter
similarly oriented, until meeting with other surrounding crystalline growths, radiating
from other centers, the free development of its external form is arrested. The polyg-
onal networks shown in Figures 1 and 2, representing the structure respectively of
pure gold and pure copper, do not indicate, therefore, cleavage planes, i.e. outlines of
true crystals, but merely boundaries or junction lines between adjacent crystalline
growths or "grains." They mark the regions where neighboring crystalline growths
collided with resulting distortion of their external forms.

1 Crystalline groups or aggregates of allotrimorphic crystals are sometimes called "crystallites"
while if they assume some distinct form they may be further described as "dendrites" or "tree-
like," "fern leaves," "star-like" crystallites, "crystalline grains," etc.

LESSON I PURE METALS 3

Grains of Metals. If crystallographers were interested in the constitution and
structure of metals they would undoubtedly refer to the small polyhedral grains of
which they are composed as allotrimorphic crystals. This expression, however,
although the only scientifically correct one, does not appeal to metallurgists, who
generally call these imperfect crystals "crystal grains," "crystalline grains," or even
simply "grains." The expression crystalline grains which is quite common appears
satisfactory since it suggests the two main facts to be remembered as to the nature
of the grains, (1) that they are not perfect crystals and (2) that they are nevertheless
crystalline. Dana himself writes that cast iron is made up of crystalline grains.
He says further: "Crystallization produces masses made of crystalline grains when
it cannot make distinct crystals." Nor is there very great objection if, for the sake
of brevity, the word grain alone be used, bearing in mind once for all the crystalline
character of metals.

Fig. 5. Pure platinum.

Cast. Magnified 120 diameters.

(Andrews.)

To render the polygonal boundaries of crystalline grains visible under the micro-
scope the highly polished metallic surface must generally be treated by an acid or
some other chemical reagent capable of dissolving or corroding the metal, with or
without deposition of a film of some precipitated matter. According to Ewing these
boundaries are made evident by the differential action of the acid which produces
differences of level by attacking one grain more energetically than its neighbors.
Each of the short sloping surfaces which connect one grain with another appears
black under vertical illumination because it does not reflect the light back into the
tube of the microscope.

Crystalline Orientation of the Grains. A slight chemical attack or etching of
the polished samples brings out merely the polygonal structure described and illus-
trated in the preceding pages. If the etching be somewhat deeper, however (through
the use of a stronger reagent or because of a longer attack), it is observed that the
polygons or meshes of the network are differently colored (Fig. 5), some appearing
very dark, others less dark, others, still, brilliant. This heterogeneousness in the
appearance of sections through adjacent grains is due to the fact elucidated above
that each grain, i.e. each allotrimorphic crystal, is made up of crystalline matter similarly

4 LESSON I PURE METALS

oriented in the same grain but differently in different grains. Bearing this in mind
the dissimilarity of coloration between contiguous grains is readily explained, for if
the crystalline matter of any individual grain is so oriented that it reflects the in-
cident light into the microscope tube, that grain will appear bright, while, on the
contrary, if its orientation is such that the light is reflected outside the microscope,
the corresponding grain will appear dark. By slightly inclining the sample in various
directions, or by rotating it, some of the grains that were bright become dark, being,
so to speak, extinguished, while some of the dark grains become brightly illuminated,
because in so doing we change the direction of the light reflected by each individual
grain section. Similar results are obtained by changing the direction of the incident
light. The kaleidoscopic effect just described affords a conclusive proof of the crys-
talline nature of metals and of the correctness of the explanation offered to account
for the dissimilar appearance, as to color, of contiguous grains.

Cubic Crystallization of Metals Etching Pits. By still deeper etching of the
polished surfaces of pure metals, it is sometimes possible to bring out clearly the
crystalline character of the individual grains (see Fig. 6). The figures thus outlined,

Fig. 6. Typical etching

figures of pure metals.

(Gulliver.)

in reality small cavities, are often called "etching pits" or "etching figures." These
figures frequently correspond to sections of cubes or of geometrical solids derived
from the cube, indicating that most metals crystallize in the cubic system (also called
regular, or isometric, or monometric system). 1

Summary. Summing up the indications obtained by the microscopical exam-
ination of polished and etched surfaces of pure metals, it has been shown (1) that a
slight etching outlines the polygonal boundaries of adjacent crystalline grains, (2)
that a deeper etching imparts different colorations to the various polygons or grain
sections, a phenomenon which is due to the constancy of crystalline orientation in
any individual grain and to the change of orientation as we pass from one grain to
the next, and (3) that a still deeper attack often brings out clearly pits of distinct
geometrical forms, often cubic, indicating that the majority of metals crystallize in
the cubic system. 2

1 The other crystallographic systems are the hexagonal, the tetragonal, the orthorhombic, the
monoclinic, and the triclinic.

2 We have other indications of the cubic crystallization of metals such for instance as the hex-
agonal character of many of the polygons which crystallographers consider to be due to interfering
cubes and octahedra (the octahedron is a form belonging to the cubic system). Again, under favor-
able crystallizing conditions nearly perfect cubes have been obtained in the case of several metals.

LESSON I PURE METALS 5

Impurities. It is well known that the addition of surprisingly small amounts of
impurities or foreign substances often affect very greatly some of the most important
properties of metals, such as their strength, ductility, fusibility, electrical conduc-
tivity, etc., and we naturally look for correspondingly marked changes of structure.
In order to understand this important influence of impurities upon the properties of
metals it will be necessary to consider at some length the nature of the union which
exists between the metal and the impurity. Let us first note that by impurity we
mean a very small proportion of some foreign substance which may be any other
metal, a metalloid, or a definite compound.

The metal or metalloid contaminating^ the metal may (1) remain uncombined or
(2) it may combine with some (generally a small amount) of UTe^metal to form a
definite chemical compound. The uncombined contaminating metal or metalloid
or resulting chemical compound may then (a) be soluble in the solid metal forming
with it a "solid solution" or (6) be insoluble in the solid metal in which case it is
rejected by the crystalline grains, in the form of an eutectic alloy.

The meaning of the expressions "solid solutions" and "eutectic alloys" should
now be explained. As Professor Howe has well expressed it the essential features of
an ordinary liquid solution are (1) a complete merging of the constituents and (2) in
indefinite proportions. By complete merging is meant a union so intimate that the
separate existence of the constituents cannot be detected by any physical means,
such for instance as microscopic examination under the highest magnification. The
homogeneity of the substance is such that it resists any physical attempt at breaking
it. The merging moreover must remain complete and absolute for varying propor-
tions of the constituents, for it is evident that if it existed only for certain well-defined
proportions of the component parts, the resulting substance would be of the nature
of a definite chemical compound and not of a solution.

Bearing in mind these characteristics of ordinary solutions, we find that in some
substances, while passing from the liquid to the solid state, the constituents remain
completely merged and in indefinite proportions. The essential characteristics of
liquid solutions are retained in the solid state. Hence the name of solid solutions given
to such substances. A common and excellent example of solid solutions is found in
the case of glass in which the three usual constituents, silica, lime, and alkali, are so
completely merged that their existence cannot be detected by physical means; the
microscopical examination of glass under the highest magnification fails to reveal the
presence of its component parts. Glass on solidifying passes from the condition of a
liquid solution to that of a solid solution. Many metals likewise form on solidifying
solid solutions; i.e. they solidify into a mass so absolutely homogeneous that the
identity of the component metals is entirely lost. The union between some metals
and metalloids also frequently forms solid solutions.

It is held by some crystallographers that in order to form solid solutions the uni-
ting substances must be "isomorphous," that is, must be capable of yielding crystals
of the same form, hence the name of "isomorphous mixtures" frequently given to
solid solutions. 1 The homogeneous crystals formed by solid solutions are often called

1 If isomorphism favors the formation of solid solutions, as it undoubtedly does, seeing that
most metals are isomorphous, we naturally infer that they will readily form solid solutions. We
now know that such is the case, for if metals are not generally soluble in each other (when solid) in
all proportions there are few instances of metals entirely insoluble in each other in the solid state.
The formation of solid solutions between metals is therefore very frequent.

6 LESSON I PURE METALS

"mixed crystals" and that expression frequently used as an equivalent for solid
solution. There are some crystallographers, however, who believe that isomorphism
of the constituents is not essential to the formation between them of solid solutions,
and that the use, therefore, of isomorphous mixtures as synonymous of solid solu-
tion is not warranted. The use of the expression mixed crystals is likewise to be
discouraged because it suggests a mixture, and, therefore, heterogeneity, which is
precisely contrary to the nature of solid solutions.

Considering now those impurities, whether metals, metalloids, or definite com-
pounds, which form solid solutions with the metal they contaminate, it is found
as might have been expected that their presence has no great influence upon the
character of the structure. Suitably prepared surfaces of such impure metals still
exhibit the polygonal network structure characteristic of pure metals. The small
polyhedra of which the impure metal is composed, however, are now allotrimorphic

Fig. 7. Gold containing 0.20 per cent
lead. Magnified 100 diameters.

(Andrews.)

crystals of a solid solution instead of a pure metal. While the character of the struc-
ture remains the same, the dimension of the grains may be markedly affected, being
often increased by the presence of a small amount of impurity forming a solid solution
with the metal; such for instance is the action of phosphorus on iron, to be considered
later. This enlargement of the grain is generally accompanied by decreased ductility
or even by brittleness.

The second group of impurities, namely those foreign substances, whether they
remain or not uncombined, which do not form solid solutions with the contaminated
metal may usually be readily detected under the microscope as they are generally
rejected to the grain boundaries during the process of solidification (or afterwards)
as shown in Figure 7. These insoluble impurities are not, however, rejected as such
by the crystalline grains, but on the contrary unite mechanically with a small amount
of the metal to form what is known as an "eutectic alloy," that is an alloy of lowest
melting-point, and it is this alloy which is expelled by the solidifying grains. The
formation and nature of eutectic alloys will be considered at greater length in a
subsequent lesson.

It will be apparent that those contaminatir substances which fail to be dis-
solved by the metal, may form actual membram surrounding each grain, the mem-

LESSON I PURE METALS 7

branes being of the nature of an eutectic alloy. As might be anticipated, the presence
of such membranes, whether continuous or not, have generally a very marked influence
upon the properties of the metals, frequently 'decreasing their ductility, weldability,
electrical conductivity, etc., and often increasing their fusibility, hardness, etc.

The rejection during solidification and subsequent cooling of those impurities
which fail to be retained in solid solution by the metal, to the grain boundaries or
other crystallographic planes, reveals the crystalline forms of the grains themselves.
The location of these impurities affords additional evidence of the cubic crystallization
of metals. It will be shown later that the cubic crystallization of iron is in this way
clearly revealed.

The above remarks are of a very general character and refer more especially to
the behavior of impurities while the metal solidifies. In the majority of cases no
further changes take place in the nature of the constituents as the metal cools to
atmospheric temperature, i.e. the constituents formed on solidification are those
found in the metal after complete and slow cooling. In some instances, however,
and notably in the case of iron and its usual impurities, carbon, silicon, phosphorus,
manganese, and sulphur, some important changes take place at temperatures con-
siderably below the solidification point of the metal which will be duly described at
the proper time.

Influence of Thermal Treatment. The size of the crystalline grains of pure
metals varies in different metals even when cast and cooled under identical condi-
tions. Their dimension is generally affected also by the rate of cooling during solidi-
fication, and, therefore, by the size of the casting, since a large casting will cool
more slowly than a smaller one.

The common belief is that the prolonged exposure of pure metals to a high tem-
perature (annealing) tends to enlarge the grains, the enlargement being the greater
the higher the temperature, the longer the time of exposure, and the slower the
cooling. While such growth undoubtedly takes place in the case of commercial and,
therefore, impure metals, at least after straining, it is held by some metallurgists
that in absolutely pure metals the grain will not grow on annealing even after
straining.

This view is based upon a theory brilliantly conceived by Ewing and Rosenhain and supported
by the results of skilfully conducted experiments. These scientists argue that even so-called pure
metals always contain a certain amount of impurities, and that even a very minute amount of im-
purity would suffice to form a thin but practically continuous film of eutectic in the crystalline
boundaries. They contend "that there is constant diffusion from the surface of the crystal into
the eutectic and equally constant re-deposition of metal upon the crystal from the eutectic. If there
are several crystals in contact with the same eutectic, there will be, under some conditions, a state
of dynamic equilibrium between them, the amount dissolved from each being exactly counter-
balanced by the amount deposited upon it; if, however, there is any difference in their solution pres-
sure in respect to the eutectic, then the less soluble will grow at the expense of the more soluble.
The metal constituting the eutectic films being much nearer its melting-point than the rest of the
mass, would thus be favorable to comparatively rapid diffusion, but the rate of such diffusion, and,
consequently, the rate of growth of crystals, would be enormously increased by heating the metal
to a temperature above the melting-point of the eutectic in question."

The theory proposed depends upon the existence of a difference in the solubility of the two crys-
stal faces in contact with the eutectic film. The only difference between these two faces is, appar-
ently, in the orientation of the crystalline elements, but this difference is sufficient, in the authors'
opinion, to produce a difference in their .rate of solution in the eutectic film, seeing that it results in
such marked difference in their solubilit in the etching acid, which, as is well known, attacks some

8 LESSON I PURE METALS

grains much more readily and deeply than neighboring ones whose elements have another orienta-
tion. To account for the influence of the orientation of the elements upon the solubility of the crys-
tals, the authors suggest to extend to alloys the electrolytic theory of solution. "Such differential
actions," they say, "may most probably be attributed to differences of electrical potential in the
surfaces involved. If we accept this view of the matter, then the diffusion across films of eutectic
becomes a case of electrolysis."

This theory explains why only strained crystals of the metals examined will grow, while un-
strained crystals show no tendency to change, even at high temperatures. "The explanation, on the
electrolytic theory, is that in the unstrained state the crystals are surrounded by practically con-
tinuous films of eutectic, and that electrolysis only becomes possible when severe distortion has
broken through these films in places, allowing the actual crystals to come into contact; the electro-
lytic circuit would then be for each pair of crystals, from one crystal to the other by direct contact
and back through the eutectic film."

If the authors' conception be true, recrystallization by annealing in a perfectly pure metal
would not occur but, as they rightly say, it is almost hopeless to obtain a sample of metal suffi-
ciently pure to prove or disprove the theory. They argue, however, that if the growth of crystals
is due to the presence of a eutectic film between them, the absence of such film would be a barrier
to all such growth, and that a weld between two clean-cut surfaces should show no growth of
crystal across the weld. This they actually proved to be so in the case of lead.

If they are right, it likewise follows that in metals contaminated by impurities with which
they form solid solutions there should be no growth of grains on annealing, because of the absence
of the needed eutectic film.

The remarkable crystalline growth of very low carbon steel after severe straining
followed by annealing at suitable temperatures, described in another lesson, is a
striking instance of the action of straining (cold .working) in promoting grain growth
in subsequent annealing.

Notwithstanding Ewing's and Rosenhain's strong argument, satisfactory evi-
dences are still lacking in support of the contention that both straining and the
presence of impurities are necessary conditions to grain growth on annealing. So
far as the matter has been investigated it does seem that the grain of a pure metal
will not grow unless it has been previously strained, but that it must also contain
some eutectic forming impurities has not been satisfactorily demonstrated.

Influence of Mechanical Treatment. Metals are frequently subjected to power-
ful pressure exerted by rolls, presses, hammers, etc., with a view of producing metallic
objects of desired shapes. This treatment affects the structure and, therefore, the
properties of the metal. Roughly speaking such vigorous kneading has a tendency
to reduce the size of the final grains, either through preventing the formation of
large grains or by breaking up or distorting preexisting grains. A smaller grain in
turn generally implies greater ductility (provided it be not distorted) and often greater
strength. The effect of work upon the structure and properties of commercial iron
and steel will be duly considered in these lessons.

Examination

I. Explain the formation and nature of the polyhedral grains of which pure metals

are composed.
II. Explain the meaning of (1) solid solutions, (2) isomorphous mixtures, and (3)

mixed crystals.

III. Describe the changes of structure, if any, produced in metals by the presence of
small quantities of impurities.

LESSON H

PURE IRON

Chemically pure iron is not a commercial product. It can only be obtained in
small quantities by carefully conducted laboratory manipulations when, even with
the most refined care, it is quite impossible to produce it absolutely pure. Until
quite recently the purest commercial iron was of Swedish origin and contained as
much as 99.8 per cent of iron. A commercial product known as "American Ingot
Iron" 1 is now manufactured which the makers claim to contain 99.94 per cent of

' .'.

: :,->>

Fig. 1. Electrolytic iron. Magnified 7.5 diameters. Slightly etched.
(Sherard Cowper-Coles.)

iron. Iron, or rather very low carbon steel, of a high degree of purity is also pro-
duced at the present time through refining in electric furnaces. Finally iron has been
obtained in relatively large quantity, by electrolytic deposition, of a degree of purity
exceeding 99.9 per cent iron.

Microstructure. When a sample of nearly pure iron is suitably prepared and
examined under the microscope some regions can readily be found absolutely free

1 The expression "ingot" iron is applied to iron containing very little carbon, obtained molten
and, therefore, cast, after removal from the furnace, whereas by wrought iron is meant iron (also
generally low in carbon) obtained pasty and, consequently, always containing a certain amount of
slag. Barring the presence of slag in wrought iron, both ingot iron and wrought iron may have
identical chemical compositions. The expression ingot iron is seldom used in the United States,
where iron obtained in a liquid state, and containing little carbon, is called low or very low carbon
steel, or mild, very mild, extra mild, steel, or again soft, very soft, dead soft steel.

LESSON II PURE IRON

from carbon and slag and exhibiting, therefore, the structure of the pure metal.
Such structure is illustrated in Figure 1 after a slight etching of the polished surface.
It will be noted that it is similar to the structure of pure gold and of pure copper
described in the preceding lesson : like gold and copper and, indeed, like most metals,

Fig. 2. Ferrite grains. Natural size.

Etched 10 minutes in nitric acid (1 to 10 water).

(Stead.)

iron is made up of polyhedral crystalline grains (allotrimorphic crystals). Upon
deeper etching the dissimilarity, as to coloration, of adjacent iron grains is clearly
brought out as shown in Figure 2. As explained in Lesson I, this appearance is due
to the fact that the grains of iron are composed of crystalline elements having the
same orientation in the same grain but different ones in different grains.

Cubic Crystallization of Iron. A still deeper etching indicates clearly the cubic
character of the crystallization of pure iron. This is illustrated diagrammatically in

LESSON II PURE IRON 3

Figure 3 and by means of a photomicrograph in Figure 4. It will be noted that the
etching pits are similarly oriented in the same grain but that the orientation in ad-
jacent grains differs. As seen in Figure 3, the etching figures may appear as triangular
wedges. This occurs when the section cuts the small cubes of a grain at a certain

Fig. 3. Etching pits in iron.
(Desch.)

Fig. 5. Silicon steel, 4.5 per
cent silicon. Magnified 60
diameters. Part of a single
grain. Etched three hours in
nitric acid (1 to 10 water).
(Stead.)

Fig. 4. Etching pits in ferrite.
(J. F. Hoyland.)

Fig. 6. Cubic crystals of iron. Magnified
250 diameters. Obtained through the reduc-
tion of ferrous chloride. (Osmond.)

angle, i.e. when it cuts obliquely a corner of each cube. This cubic structure is fur-
ther illustrated in a remarkable manner in Figure 5, in the case of iron containing
\Yi per cent of silicon after etching three hours in dilute nitric acid. The photo-
graph shows a portion of a single grain, hence the constancy of orientation noted.
The presence of so much silicon apparently favors the development of a coarse cubic
crystallization.

Osmond, through the reduction of ferrous chloride and the crystallization of the

LESSON II PURE IRON

resulting metallic iron, obtained perfect isolated cubic crystals (Fig. 6). Finally
almost perfect cubes have been separated by Stead from a large granule of phos-
phoretic iron (Fig. 7). Another indication of the cubic crystallization of iron is found
in the occurrence of large crystallites (Fig. 8), generally resembling pine trees, in
cavities of large castings of iron and steel, under conditions, therefore, favorable to
the free development of crystals, these crystallites being composed of small octa-
hedra, a crystalline form of the cubic system. Finally it will be shown later that
the structural location of some of the impurities generally present in commercial
iron affords further proof of the cubic crystallization of iron.

Ferrite. Mineralogical names have been given to the constituents of iron and
steel, and pure iron, or rather carbonless iron, considered as a microscopical con-

Fig. 7. Cubic crystals of phosphoretic iron.
Magnified 5 diameters. Phosphorous 0.75
per cent, carbon per cent. (Stead.)

stituent has been called "ferrite," a name suggested by Professor H. M. Howe and
universally adopted. 1 Pure iron, therefore, is composed of polyhedral crystalline
grains of ferrite. It will be seen in subsequent lessons that the ferrite of commercial
grades of iron and steel is not pure iron, but rather a solid solution of iron holding
small amounts of silicon, phosphorous, and possibly other impurities.

Allotropy of Iron. The study of the crystallization of iron is complicated by
the existence of several allotropic forms of that metal.

Allotropy suggests marked and sudden changes in some of the properties of a

1 This constituent was called "free iron" by Sorby, who was the first scientist to describe the
microscopical structure of iron and steel.

In the last report (1909) of the Committe on the Uniform Nomenclature of Iron and Steel, ferrite
is described as follows: "Alpha iron holding in solution in the case of commercial grades of iron and
steel, small and varying amounts of silicon, manganese, phosphorus, and of some other rarer ele-
ments, but not more than 0.05 per cent of carbon, if any." There is no conclusive evidence, how-
ever, of the presence of carbon in ferrite, nor is it likely that iron retains any manganese in solid
solution when only a small amount of that impurity is present. (See Lesson VI on Impurities in
Iron and Steel.) Stead has suggested that when ferrite consists only of pure iron it should be called
"ferro-ferrite."

LESSON II PURE IROX

. g. iron crystallite about half
natural size. (Tschernoff.)

6 LESSON II PURE IRON

substance occurring at certain critical temperatures, without any change of state or
of chemical composition. 1 Polymorphism, sometimes used as the equivalent of allo-
tropy, refers more specifically to the property of some substances of crystallizing in
more than one form, while allotropy does not necessarily imply such property.

Many substances undergo allotropic changes. It is a matter of common knowl-
edge, for instance, that sulphur exists under two distinct conditions, namely, as pris-
matic sulphur and as octahedral sulphur, the prismatic form being the one stable
above 95.6 deg. C. and the octahedral, the stable form below that critical tempera-
ture. On heating octahedral sulphur it begins to change into the prismatic form at
the temperature of 95.6 deg., and, likewise, on cooling, prismatic sulphur begins to
pass to the octahedral form at that temperature. Many of the physical properties of
sulphur (crystalline form, specific heat, heat of combustion, etc.) undergo sudden
changes as the substance passes from one allotropic form to another.

At those critical temperatures which mark the passage of one allotropic form into
another, spontaneous evolutions of heat take place on cooling and spontaneous ab-
sorptions of heat on heating. These thermal disturbances indicate a change of in-
ternal energy which when not accompanied by changes of state or by chemical
changes are evidences of allotropy. The usual method of detecting the existence of
such thermal critical points will be described in another lesson.

Osmond's momentous discovery of the existence of two thermal critical points
in pure iron proves the existence of iron in at least three allotropic forms. The two
critical temperatures of pure iron correspond respectively to about 875 and 750 deg.
C. Above 875 deg., iron exists in a certain allotropic condition known as y (gamma)
iron. On cooling, when 875 deg. is reached iron passes from the / form to another
allotropic condition called ft (beta) iron and at 750 deg. ft iron in turn changes into a
third allotropic form, (alpha) iron, which remains the stable form at atmospheric
temperature. The reverse changes occur on heating, that is iron passes from the a,
to the ft form and then to the 7 form on passing through the two critical tempera-
tures. 2

As should be expected the passage of one allotropic form into another implies
corresponding and, generally, sudden changes in many of the physical properties of
iron. Gamma, beta, and alpha iron differ widely in regard to many of their physical
characteristics. It is only desired in this lesson, however, to inquire into the possible
differences of crystallization which may exist between the three allotropic conditions

1 In the Nomenclature of Metallography published by the Iron and Steel Institute in 1901
allotropy is described as "the capacity to undergo without change of composition a change of chem-
ical and physical properties."

Roberts-Austen defined allotropy as "a change of internal energy occurring in an element at
a critical temperature unaccompanied by a change of state." In the above definition the word
substance should be used, however, in place of element, for allotropic changes are not confined to
elements; chemical compounds also, and possibly solid solutions, may undergo allotropic trans-
formations.

Tiemann ("Iron and Steel") thus defines allotropy "a change in the properties of an element
without change of state. It is habitually accompanied by a change of internal energy. It is due in
some, and perhaps in all, cases to a change in the number or in the arrangement of the atoms of
the molecule (Howe). Allotropic varieties are sometimes termed 'isomerides.'"

2 Le Chatelier and some other writers, while admitting the existence of a lower critical point in
pure iron, doubt that it indicates an allotropic change, being consequently inclined to deny the
existence of beta iron. This opinion will be considered and the allotropy of iron treated at greater
length in another lesson.

LESSON II PURE IRON 7

of iron, postponing until a later lesson a description of the modification of the other
properties.

Osmond and, later, Osmond and Cartaud have carefully investigated the difficult
problem of the crystallization of the different allotropic forms of iron. Their con-
clusions were (1) that the three allotropic forms of iron crystallize in the cubic sys-
tem, 1 (2) that octahedra are the prevailing crystalline forms of gamma iron, (3) that
the cube is the prevailing form of beta and of alpha iron, (4) that beta and alpha iron
are capable of forming isomorphous mixtures (solid solutions), (5) that gamma iron
docs not form isomorphous mixtures with beta or alpha iron. .

We also have the statement of Osmond that the transformation of gamma iron
into beta iron appears to include a change in the planes of symmetry, at least in car-
burized iron.

Again it has been shown by Osmond and confirmed by other investigators that
the occurrence of twinnings is frequent in gamma iron, while beta and alpha iron are
free from it. By twinning is meant the grouping of two or more crystals or parts of
crystals in such a way that they are symmetrical to each other with respect to a
plane between them (the twinning plane) which plane, however, is not a plane of
symmetry for either crystal.

Twinnings are also produced through pressure or other stress tending to strain
the crystals especially when the straining is followed by annealing. In Figures 9, 10,
and 11 are shown twinnings respectively in marble (due to pressure), in brass, and in
gamma iron. The straight parallel lines and bands running through the polyhedral
grains indicate the twinning planes.

From these conclusions which have been confirmed by the results obtained by
other investigators, it follows that the allotropy of iron could not be proved by its
crystallography, since the thermal critical points are not accompanied by changes in
the crystalline form of iron. While, however, in the instances of allotropy which have
been noted and studied allotropic changes are generally accompanied by changes of
crystalline forms, it does not by any means follow that any allotropic transformation
must necessarily imply a crystalline change.

Bearing in mind the existence of three allotropic conditions of iron, let us follow
in our imagination the crystallization of iron during solidification and its subsequent
cooling to atmospheric temperature. On solidifying, polyhedra or allotrimorphic
crystals of gamma iron are formed, which according to Osmond are chiefly made up
of octahedral crystals. Were it possible to examine a section of the solidified metal
at such a high temperature the usual polygonal network structure characteristic of
pure metals would be revealed. Upon further cooling below the solidification point,
no change of crystalline form should take place until the first critical temperature
(875 deg. C.) is reached when the iron changes from the gamma to the beta condition.

Does this allotropic change affect the preexisting crystallization of gamma iron
or does it consist merely in a transformation in situ of each crystalline grain of gamma
iron into a grain of beta iron, retaining the original external form of the gamma
grain, and leaving undisturbed, therefore, the polygonal structure observed under the
microscope? It may reasonably be supposed that the allotropic transformation takes
place without affecting the external form of the crystalline grains, but in view of

1 In Le Chatelier's opinion there is no proof of the cubic form of gamma iron. He thinks that
the facts observed are contrary to that hypothesis and that it is more probable that gamma iron is
rhomboh'edral or orthorhombic.

LESSON II PURE IRON

Osmond's statement that the octahedron is the prevailing crystalline form of gamma
iron and bearing in mind that the small crystals revealed by suitable etching of pure
iron are generally cubic, we naturally infer that the octahedral character of each
grain of gamma iron has been obliterated, the small octahedral elements of gamma

Fig. 9. Twinnings in marble (caused by

pressure). Magnified about 5 diameters.

(Bayley.)

Fig. 11. Twinnings in gamma iron.
Magnified 200 diameters. (Osmond.)

Fig. 10. Twinnings in brass. Magnified 100 diameters. (Law.)

iron having been replaced by small cubic elements of beta (and later of alpha) iron.
These conclusions, however, should be accepted with reserve, as we lack evidences of
a very conclusive character. It has been shown that the change of gamma iron into
beta iron is accompanied by an abrupt expansion of the cooling metal which expan-
sion is followed by the normal contraction of cooling substances. We infer from

LESSON II PURE IRON

this that, momentarily at least, each grain of beta iron occupies more space than its
gamma iron progenitor.

At 750 deg. C. or thereabout, the iron passes from the beta to the third or alpha
form. We may here ask the same questions as to the probable effect of this change
upon (1) the outward form of each beta grain and (2) upon the internal crystalline
structure of each grain. Since both beta and alpha iron crystallize in the cubic sys-
tem, the cube being the prevailing crystalline form of both, and since, according to
Osmond, they are isomorphous, that is capable of forming solid solutions, it seems
probable that the change of beta to alpha iron affects neither the external form nor

Fig. 12. Polished sample of iron, etched, then
reheated in hydrogen to 950 deg. C. and slowly
cooled. Magnified 600 diameters. (Kroll.)

the internal crystalline arrangement of the beta grains; in other words that each
small cubic element of beta iron is converted bodily (although probably gradually
and not abruptly) into a cubic element of alpha iron.

If the above represents the real mechanism of the allotropic changes, it will be
evident that the polyhedral grains revealed through the etching of polished sections
of pure iron were formed during solidification and consisted originally of gamma
iron, these grains having retained their external shape while undergoing allotropic
transformation, but having probably undergone at the upper critical point an internal
change, the octahedral form of the crystalline elements having been replaced by the
cubic form.

Some experimental evidences, however, have recently been brought forward
which would lead to somewhat different conclusions. Kroll in a valuable and sug-
gestive contribution expresses his belief that he has been able to develop in a sample
of nearly pure iron three distinct polygonal networks, corresponding respectively to
the crystallization of gamma, beta, and alpha iron. This interesting composite struc-
ture is illustrated in Figure 12. It was obtained in heating a polished section of pure

10 LESSON II PURE IRON

iron in a current of hydrogen gas, when it was found that the three polygonal struc-
tures were revealed. If the meaning of this structure is correctly interpreted by
Kroll, it, of course, points to marked crystallographic changes closely related to the
thermal points of Osmond. As Professor Le Chatelier rightly says, however, the
accuracy of Mr. KrolPs conclusions might be questioned, because changes of struc-
ture must take place on cooling as well as on heating so that five and not three net-
work structures should be observed.

Rosenhain and Humfrey by straining a bar of iron heated to a high temperature
at the center (the temperature, therefore, decreasing towards both ends) have made
evident the existence of three distinct kinds of distortions with sharp lines of de-
marcation between them, corresponding in all probability to the distortion respect-
ively of gamma, beta, and alpha iron. The estimation of the temperature of different
portions of the bar by means of fusible salts appears to sustain the authors' con-
tention thus furnishing an additional support, and a substantial one, to Osmond's
brilliant theory of the allotropy of iron.

Nevertheless information of a more positive and concordant nature is still needed
to settle to the satisfaction of all the question of the relation between the thermal
critical points of pure iron and possible crystallographic changes.

While the allotropy of iron is of little industrial importance in the case of iron
and of low carbon steel it becomes, on the contrary, of very great moment in high
carbon steel, for it is probably the cause of that invaluable property possessed by
such steels of becoming intensely hard upon rapid cooling from a sufficiently high
temperature. The question of the hardening of steel will be duly considered in these
lessons.

Influence of Impurities. Commercial iron is always contaminated by the
presence of at least five elements, namely, manganese, silicon, phosphorus, sulphur,
and carbon, generally referred to, although often wrongly, as impurities. The im-
portant question of the influence of these substances upon the structure, and, there-
fore, upon the properties, of iron and steel will be fully considered in another lesson,
when it will be shown that these elements form definite compounds with iron, FeSi,
Fe 3 C, Fe 3 P, or with each other, Mn 3 C, MnS, and that some of these compounds,
FeSi, Fe 3 P, are retained by the iron in solid solution, while others, Fe 3 C, Mn 3 C, MnS,
are rejected to the boundaries of the crystalline grains or along other crystallographic
planes, the former two giving rise to the formation of eutectoid mixtures. This
behavior of the impurities of iron and steel conforms with the general behavior of
impurities described in Lesson I.

Influence of Heat Treatment. The size of the grains of iron is affected, like
that of other metals, by the temperature from which it cools, the length of time it is
kept at that temperature, the rate of cooling, etc., in other words by what is known
as its heat or thermal treatment. Generally speaking it may be said that the higher
the temperature the larger the grains and also that the slower the cooling the larger
the grains. These results might have been anticipated if it be considered that slow
cooling from a high temperature is a condition favorable to the formation and growth
of crystals. As stated in Lesson I, however, it is not certain that pure metals under-
go any crystalline growth on reheating (annealing) unless they have previously been
strained and, indeed, unless they contain also a trace at least of impurities. For
similar reasons we may doubt the existence of any crystalline growth in annealing
chemically pure iron or indeed impure iron unless it has been previously strained.

LESSON II PURE IRON

The influence of heat treatment upon the structure and physical properties of
commercial irons and steels will be dealt with at length in these lessons.

Influence of Mechanical Treatment. Like that of other metals the structure of
iron is affected by mechanical work. Undisturbed cooling being a necessary condi-
tion to the free development of crystals, it will be evident that if the metal be vigor-
ously worked, that is subjected to powerful mechanical pressure, while cooling from
a high temperature, the formation of crystalline grains will be greatly hindered or
preexisting crystals broken or distorted. The important influence of work upon the
structure (and therefore upon the properties) of iron and steel will be duly considered
in these lessons.

Straining of Iron. Slip Bands. Ewing and, later, Ewmg -and Rosenhain
through some skilfully conducted experiments and convincing reasoning have re-
vealed the character of the strain produced in a pure metal by the action of a stress

Fig. 13. Slip bands in Swedish iron strained

by tension. Magnified 400 diameters.

(Ewing and Rosenhain.)

which may eventually cause its rupture. This will be briefly described here in the
case of pure iron.

Polished sheets of metals were strained very gradually while being examined
under the microscope. When the yield point is reached, i.e. when plastic deforma-
tion begins to occur, black lines are seen to cross the crystalline grains of which
the metal is made up. These lines are generally quite straight and are parallel in the
same grains but have different directions in different grains. Figure 13 shows the
appearance under vertical light of Swedish iron, strained by tension, magnified 400
diameters.

As the strain increases other systems of lines, inclined to the first, make their
appearance and eventually two, three, and even four systems of intersecting lines
may be seen in each grain.

These lines are not cracks but steps in the surface, which steps are due to slips
along the cleavage or gliding planes of the crystals. In Figure 14, AB represents the

12 LESSON II PURE IRON

polished surface of two grains, C the junction line between these two grains. The
pull takes place in the direction of the arrows.

Yielding occurs by finite amounts of slips at a limited number of places, a, b, c,
d, e, resulting in short portions of inclined cleavage or sliding surface appearing black
under the microscope, because they do not send back any light into the tube. By
oblique light some of these slip bands appear black while others are bright. When
the surface is rotated some of the bands which were black become bright and vice
versa owing to the change of their position in regard to the incidence of the light,

Before straining.
a, o ~_ *

After straining

Fig. 14. Diagram illustrating the effect of strain

upon the structure of metals and alloys.

(Ewing and Rosenhain.)

which is conclusive proof that the black lines are not cracks, but inclined surfaces as
described above.

It is seen then that, contrary to the general belief, metals remain crystalline
after the severest strain, and that the flow or plastic strain in metals is not a homo-
geneous shear such as occurs in the flow of viscous fluids, but is the result of a limited
number of separate slips, the crystalline elements themselves undergoing no deforma-
tion.

Examination

I. Describe the microstructure of pure iron as revealed by the microscope (1)
after a slight etching of the polished surface, (2) after a deeper etching, and
(3) after a still deeper etching.

II. Give some evidences of the cubic crystallization of iron.
III. Describe the allotropy of iron.

LESSON III

WROUGHT IRON

Wrought iron is the name given to commercial iron free enough Trom carbon and
other impurities to be malleable when such metal is manufactured through the re-
duction of iron ores or the refining of cast iron at a temperature so low that it is
obtained in a pasty condition and, therefore, mechanically mixed with a considerable
amount of the slag formed during the operation.

When the refining treatment is conducted at a temperature sufficiently high to
deliver the resulting products in a molten condition, the refined metal which is then
free from slag is called steel. Wrought iron and steel may otherwise have identical
chemical composition, although usually steel contains more manganese and less
silicon than wrought iron, often more carbon and less phosphorus.

Commercial iron which is not malleable is called cast (pig) iron. The modern
method of producing wrought iron consists in refining cast iron in a non-regenera-
tive reverberatory furnace (the puddling furnace), while the refining of cast iron for
the production of steel is conducted (1) in the Bessemer converter or (2) in a regen-
erative reverberatory furnace (the Siemens open-hearth furnace). Cast (pig) iron is
the result of smelting iron ore in blast-furnaces.

Chemical Composition. Wrought iron contains, besides an appreciable amount
of slag, a small proportion of carbon and small quantities of the usual impurities,
manganese, silicon, phosphorus, and sulphur.

Microstructure of Longitudinal Section. Upon being withdrawn from the
puddling furnace, the white hot, pasty balls of wrought iron are subjected to vigorous
forging or squeezing, thus expelling a large amount of slag and firmly welding to-
gether the particles of iron. Through additional heating and forging or rolling the
metallic mass is converted into such elongated shapes as blooms, billets, bars, etc.
These operations so affect the structure as to impart unlike appearances to sections
cut longitudinally, i.e. in the direction of forging or rolling, and sections cut trans-
versally, i.e. at right angles to that direction. The microstructure of the longitudinal
section of a wrought-iron bar is shown in Figures 1 and 2. From our knowledge of
the chemical composition of wrought iron we should be able to anticipate its micro-
structure. The ground mass or matrix of the metal consists of polyhedral crystalline
grains of iron, that is of ferrite, similar in every respect to the crystalline grains of
pure iron and of pure metals in general described in Lessons I and II. The ferrite of
wrought iron, however, as explained in Lesson II, is not pure iron but rather a solid
solution of iron in which are dissolved small quantities of silicon, phosphorus, and other
minor impurities. This true character of commercial ferrite is too often lost sight of
and the constituent considered as pure iron. The difference in coloration between
adjacent grains of this commercial ferrite should be noted, and will of course be readily
understood in view of the explanation given in Lesson I to account for this phenom-

2 LESSON III WROUGHT IRON

enon. Many irregular black lines, varying much in thickness and length, but all
running in the same direction are clearly seen. These lines indicate the location of
the slag which has assumed the shape of fibers or streaks running in the direction of
the rolling or forging, thus imparting a fibrous appearance to the metal.

The presence of a small amount of carbon in wrought iron results in the occurrence
of a new constituent in the shape of small dark particles located between some of the
grains. Under low magnification these carbon-holding particles are not readily dis-
tinguishable from slag particles and as this carburized constituent is not a very im-
portant one in the case of wrought iron it seems advisable to postpone its description.

Fig. 1 . Wrought iron. Longitudinal

section. Magnification not stated. Fig. 2. Wrought iron. Longitudinal section.

(Longmuir.) Magnified 100 diameters. (Boynton.)

Summing up, wrought iron consists essentially of a mass of ferrite containing many
elongated particles of slag.

Microstructure of Transverse Section. The microstructure of the transverse
section of a wrought-iron bar is illustrated in Figures 3 and 4. Like the structure of
the longitudinal section, it consists of a polygonal network, indicating that the metal
is made up of polyhedral crystalline grains of ferrite. The slag, however, which in
the longitudinal section occurred as fibers running in a direction parallel to the rolling
or forging, here assumes the shape of irregular, dark areas, corresponding to the cross
sections of the slag fibers. It will be noted that in both the longitudinal and trans-
verse sections the ferrite grains are equiaxed, that is, they show no sign of having
been elongated in the direction of the rolling. It was thought for many years that
wrought iron actually had a fibrous structure and, indeed, the number of persons
still holding this view is surprisingly large. Many valuable properties were attrib-
uted to puddled iron on account of its "fibrous structure" which were denied to
steel because of its "crystalline structure." The microscope has summarily dis-
posed of this erroneous belief in showing that the ferrite constituting the bulk of
wrought iron is in no way different from the ferrite forming the bulk of low carbon
steel. Both are equally crystalline.

LESSON III WROUGHT IRON 3

Chemical Composition of Slag. The essential chemical constituents of the slag
produced in the puddling furnace and retained in part by the iron are iron oxides,
both ferric (Fe 2 O 3 ) and ferrous (FeO), oxide of manganese (MnO), silica (Si0 2 ), and
phosphoric acid (P 2 O 5 ). Of these the oxides of iron and manganese are basic in their
chemical affinity while silica and phosphoric acid are acid. These bases and acids
combine with each other to form neutral compounds: silicates and phosphates of
iron and manganese.

Microstructure of Slag. It will be seen in Figures 1 and 2 that the slag fibers
are really made up of at least two constituents, a dark and a lighter one, the light

Fig. 3. Wrought iron. Transverse

section. Magnification not stated.

(Longmuir.)

Fig. 4. Wrought iron. Transverse section.
Magnification not stated. (Guillet.)

constituent moreover often occurring in the form of small rounded areas. This
structure of the slag is shown more clearly and on a larger scale in Figure 5. We
are naturally led to speculate as to the nature of these two distinct constituents of
the slag, and in view of our knowledge of the chemical composition of slag as stated
in the preceding paragraph we are tempted to conclude that one of the constituents
is a silicate of iron and manganese while the other is a phosphate of the same bases.
The accuracy of this deduction, however, remains to be proven.

According to Matweieff the rounded light areas consist of iron oxide mixed or
not with manganese oxide, and the darker background of silicate of iron and man-
ganese.

Matweieff recommends the following method to distinguish between the different constituents
of slag. The polished sample placed in a tube is heated and treated by a current of pure hydrogen
which causes the reduction of the metallic oxides while the silicates are unaffected. To detect the
presence of ferrous oxide (FeO) the sample heated to a red heat is acted upon by steam, a treatment
resulting in oxidizing the ferrous oxide into magnetic oxide (Fe 3 O 4 ), while the silicates again remain
unaltered. To detect the presence of manganese in the particles of oxides revealed by the hydrogen
treatment the previously treated sample is repolished and etched with a dilute solution of ferric
chloride in alcohol: if the white metallic grains resulting from the hydrogen treatment are colored
darker than the surrounding iron they contain some manganese. Finally to detect the presence of
iron and manganese sulphide the polished sample is etched with a dilute solution of tartaric acid
which colors sulphide of manganese lightly and iron sulphide decidedly.

4 LESSON III WROUGHT IRON

Rosenhain considers it probable that the two distinct constituents of wrought-
iron slag are two different silicates or, possibly, oxides of iron.

It is seen that writers generally ignore the presence of phosphoric acid in the slag
from the puddling furnace, and still it generally contains from 3 to 5 per cent of it,
and occasionally considerably more. If we assume that this phosphoric acid forms
with iron a phosphate of the formula 3FeO.P 2 O 6 , a simple calculation, according to
atomic weights of the elements involved, will show that the presence of 5 per cent of
phosphoric acid would mean the formation of over 12.50 per cent of this phosphate
of iron. It is hardly to be supposed that this phosphate is absorbed by some other

Fig. 5. Particle of slag in wrought iron.
Magnified 200 diameters. (Guillet.)

constituent of the slag. On the contrary it seems highly probable that it must be
present as a distinct constituent.

Influence of Thermal and Mechanical Treatments. The dimensions of the fer-
rite grains of wrought iron are affected by the treatments, both thermal and mechan-
ical, received by the metal. The effect of these treatments upon the structure of
iron and steel will be considered in another lesson.

Experiments

A small piece of wrought-iron bar should be procured measuring preferably Yi in.
square or round and % in. long. This piece should be sawed in two longitudinally,
conveniently with a hack-saw (preferably a power hack-saw), and one of the freshly
cut surfaces prepared for microscopical examination.

The various methods which have been used or recommended for the polishing of
iron and steel specimens preliminary to their microscopical examination will be
found duly described in an appendix to these lessons. In connection with the experi-
ments described in this book only those methods will be mentioned which in the
author's opinion have been found most satisfactory.

Polishing by Hand. The sharp edges of the sample should be filed or ground
in order to avoid tearing the polishing cloths in the following operations. The sur-

LESSON III WROUGHT IRON 5

face to be prepared should be filed first with a coarse and then with a smooth file so
as to obtain a perfectly flat, smooth surface. This filing can be advantageously re-
placed by grinding on a fine emery-wheel. It is recommended that both filing or
grinding be conducted with a very gentle pressure. In case the polishing is to be
done by hand the hand polishing outfit described at the beginning of these lessons is
recommended.

A small amount of No. 80 emery powder 1 mixed with sufficient water to form a
thick paste should be placed on one of the polishing blocks covered with cotton
cloth. This paste should be spread over the block, conveniently by means of a
spatula, and with the addition of a little more water if necessary. The sample of
metal should now be rubbed back and forth over this block, "Being careful to rub
always in the same direction until the marks left by the file or emery-wheel have all
been removed and replaced by finer markings due to the action of the emery powder.
After this treatment the sample should be carefully washed, as well as the fingers of
the operator, preferably in running water, and the sample rubbed over the second
polishing block covered with cotton cloth and a little flour emery and water, pre-
cisely as before. On passing from the first to the second polishing block, the sample
should be turned at right angles and kept in that position, in order that the new
marks may be perpendicular to the old ones, and the polishing should be continued
until the marks left by the coarse emery have been entirely effaced and replaced by
finer ones. The sample after being carefully washed is ready for the next block.
Some of the tripoli powder should be spread, with the addition of water, over one of
the blocks covered with broadcloth, and the sample polished upon this block until
the markings left by the previous polishing have been completely removed. After
careful washing the sample should now be rubbed over the last polishing block, cov-
ered with broadcloth, rouge, 2 and water, holding as usual the sample so as to rub it
at right angles to the markings left by the tripoli. After these markings have been
removed the sample should have a very bright surface and be free from even micro-
scopical scratches. At this stage a magnifying-glass is very useful for inspecting
the specimen in order to ascertain whether it is ready for the etching treatment. For
this purpose the vertical magnifier described and illustrated under "Apparatus" at
the beginning of these lessons will be found very satisfactory.

On examining the polished sample of wrought iron with the naked eye many
small, elongated cavities will be detected which will be more apparent still if viewed
through a magnifying-glass. These marks correspond to the location of the slag
fibers and will be readily distinguished from scratches.

The specimen should now be carefully washed and dried with a soft cloth, pref-
erably a fine piece of old linen. Where an air-blast is at hand, as is generally the
case in chemical laboratories, it is advisable to dry the specimen with this blast (a
hot blast is more effective than a cold one) instead of rubbing it with a cloth. The
sample may then be passed gently once or twice on a piece of chamois leather
stretched over a smooth piece of wood and carefully protected from dust when not in
use, when it will be ready for the next or etching operation. When polishing, the
sample should be pressed lightly upon the blocks and great care taken not to carry
any coarse powder over a polishing block upon which a finer powder is used as the
presence of but a few coarser grains will greatly lengthen the operation. It is of

1 The polishing powders should be of the very best quality obtainable.

2 It is essential to use the best commercial grade of jeweler's rouge.

6 LESSON III WROUGHT IRON

much importance, therefore, to keep all the blocks carefully covered when not in
use as well as the bottles containing the powders.

Polishing by Power. Polishing by hand is at best a tedious and laborious
operation and whenever possible it is highly advisable to replace it by the use of a
power polishing machine. Very satisfactory and effective polishing machines and
polishing motors are illustrated and described in the introductory chapter on ap-
paratus.

When using these polishing machines or polishing motors the manipulations are
as follows:

The metal surface to be prepared is pressed lightly upon the emery-wheel until a
perfectly flat surface is obtained, when it should be washed with the usual precau-
tions and pressed upon the cloth-covered cast-iron disk placed next to the emery-
wheel and upon which flour emery and water have been applied. Care should be
taken to hold the specimen so that the new marks will cross the old ones at right
angles and the grinding should be continued until the emery-wheel marks have been
completely erased. After washing the specimen it is ready for treatment on the
next surface covered with broadcloth upon which has been spread tripoli powder
and water, here again turning the sample 90 degrees. When the marks left by the
preceding operation have been removed, the specimen is washed and given the final
polishing treatment by pressing it lightly upon the other side of the cast-iron disk
upon which rouge and water is used. The various polishing powders mixed with
water may be conveniently applied to their respective disks by means of flat and
rather stiff brushes. The surface of a properly polished sample should be highly
specular and free from scratches.

Time will be saved by exerting a slight pressure only while polishing, especially
on the emery-wheel and emery disk, because deep marks due to these abrasers will
be troublesome to remove with the finer powders. With these machines a sample of
steel measuring % in. square or J^ in. in diameter is readily polished in 10 minutes.

Etching. If the polished sample of wrought iron were now placed under the
microscope, it would be possible to detect some of the slag particles but the structure
of the iron itself could not be seen, because all parts of the sample being uniformly
bright would reflect the light to the same extent. To make the structure apparent
under the microscope it is necessary to impart unlike appearances to the constituents.
This is generally accomplished by producing a slight corrosion or etching of the pol-
ished surface. For this purpose acid solutions are generally used which attack some
constituents more deeply than others or to the exclusion of others, which action may
or may not be accompanied by the deposition of some precipitated matter.

Arnold considers the operation of etching with dilute acids to be of an electro-
lytic nature, some of the constituents being electro-negative to others, hence the
attack of some of these (electro-positive constituents) to the exclusion of others
(electro-negative constituents) and the darker coloration of the former.

The various methods which have been used or recommended for the development
of the structure of iron and steel samples will be found duly described in an appendix
to these lessons but only those methods which in the author's opinion are most satis-
factory will be mentioned in connection with the experiments of this book.

Etching with Picric Acid. (Igevsky.) An etching solution should be prepared
containing 5 grams of picric acid, chemically pure, and 95 cubic centimeters of
absolute alcohol. This should be kept in a well-stopped glass bottle.

LESSON III WROUGHT IRON 7

A small amount of this solution should be poured in a glass or porcelain dish,
preferably a small crystallizing glass dish with cover, and the sample immersed in it
for 30 seconds, when it should be removed, conveniently with a pair of pincers (pref-
erably with platinum tips), and washed in alcohol. The sample should now be dried,
preferably by means of a blast for which a foot-blower will answer very well. After
rubbing the sample very gently once or twice upon a smooth piece of chamois leather
stretched on a wooden block and carefully kept free from dust, it will be ready for
examination.

Examination. The prepared sample should be suspended to the magnetic
specimen holder described under "Apparatus" in such a way as tc^ expose to view as
much as possible of its surface. The source of light and condensers should be ad-
justed so that a beam of light of suitable size enters the vertical illuminator ; the
light beam should cover a little more than the aperture of the illuminator. A 2 in.
(5X) eyepiece and a 16 mm. (% in.) objective will be a satisfactory combination for
the examination. The image of the specimen should be focussed roughly by the rack
and pinion motion of the stage, and the milled head of the vertical illuminator turned
tentatively and gently right and left until the sample appears brightly and uniformly
lighted. The object should now be brought to a sharp focus by means of the fine
adjustment.

The etching treatment should have outlined the joints between the ferrite grains
clearly and sharply. If the structure lacks clearness it is safe to infer that the etch-
ing was not properly done. In that case the sample should be rubbed a few times on
the chamois leather block and again examined without repeating the etching. If
the structure remains ill-defined, rub the specimen a minute or two on the rouge
block or disk, wash, dry, and repeat the etching treatment until satisfactory results
are obtained.

Should the boundaries of the ferrite grains appear too faint, the etching treat-
ment should be repeated without repolishing, so as to etch these lines more deeply.

As the usual purpose of the microscopical examination of samples of wrought
iron is to ascertain the quantity and mode of occurrence of the slag and the dimen-
sions of the ferrite grains, it is not generally desired to etch the sample so deeply
that some of the grains become deeply colored, still less that etching pits begin to
appear.

In this experiment, however, the student is advised to etch his sample gradually
so that the different stages of the structure may be clearly seen: (1) before etching:
slag fibers and a brilliant structureless matrix, (2) after a slight etching: ferrite grains
sharply defined but remaining uncolored or but slightly colored, (3) after a deeper
etching: some of the ferrite grains deeply colored, and (4) after a still deeper etching:
small cubic etching pits beginning to appear.

The production of these etching pits, however, is often a troublesome and uncer-
tain operation. Heyn recommends for that purpose etching with double chloride of
copper and ammonium, others (Stead) "a sufficiently long immersion in lukewarm
20 per cent sulphuric acid, followed by cleaning in nitric acid."

Etching with Diluted Nitric Acid. The sample used in the above experiment
should be rubbed a short while on the rouge block so as to remove the effect of the
etching, washed and dried in the usual way and etched with a solution containing
10 c.c. of concentrated, chemically pure nitric acid and 90 c.c. of absolute alcohol.
The etching should be conducted in the same way but as this reagent acts more

8 LESSON III WROUGHT IRON

quickly the sample should not be left in the solution more than 10 or 15 seconds,
when it should be washed in alcohol, carefully dried and passed gently once or twice
over the chamois leather block.

If the boundaries of the grains are too faintly developed the etching should be
repeated without repolishing.

The appearance of the sample after this treatment should be identical to that
resulting from etching with picric acid, the only essential difference between these
two reagents being the slower action of the latter.

Etching with Concentrated Nitric Acid. (Sauveur.) The corrosion due to the
last treatment should be removed by rubbing the sample on the rouge block which
after careful washing and drying should now be etched with concentrated nitric acid
as follows: The polished specimen conveniently held with a pair of pincers (prefer-
ably with platinum tips) should be dipped in a beaker or other vessel containing
concentrated nitric acid (1.42 specific gravity) and immediately afterwards held
under an abundant stream of running water. When iron is immersed in concen-
trated nitric it assumes the passive state, that is, it is not affected by the acid. As
soon, however, as the layer of concentrated acid which covers the polished surface
is diluted by the running water, the steel is vigorously attacked but for so short a
time (since the water soon removes all traces of acid) that there is little danger of
etching too deeply. One such treatment is generally sufficient to bring out the struc-
ture sharply and clearly but if the specimen is found insufficiently etched, the etch-
ing should be repeated in exactly the same manner. The author believes that the
simplicity of this etching treatment and the excellent results generally obtained
have been overlooked by metallographists.

Transverse Section of Wrought-Iron Bar. The student should prepare a trans-
verse section (preferably not over y% in. thick) of the same wrought-iron bar, following
exactly the manipulations described for the polishing and etching of the longitudinal
section. He should compare the structure of the two sections and notice (1) their
similarity as to the appearance of the ferrite grains and (2) the unlike occurrence of
the slag which in the transverse section is present as small irregular areas correspond-
ing to cross sections of the slag fibers of the longitudinal section.

Examination

I. Describe briefly the structure of commercial wrought iron, explaining the dif-
ference between the appearances of longitudinal and transverse sections.

II. Describe the structure of your samples and mention any difficulty which you
may have encountered in your manipulations.

III. If you have any preference for one of the etching methods described give your
reasons in support of it.

LESSON IV

LOW CARBON STEEL

In this and the following lessons steel will be considered as_a^pAire alloy of iron
and carbon, i.e free from the impurities (silicon, manganese, sulphur, and phosphorus)
always present in commercial products. The influence of these elements upon the
structure of steel will form the subject of another lesson.

Normal Structure. The structures described in this and the next lesson refer to
the condition of steel after forging followed by heating to, and slowly cooling from,
a high temperature (900 to 1000 deg. C.). Such treatments, for reasons that will
be understood later, promote soundness, remove internal strains, prevent excessive
coarseness of structure (as in castings), and permit a state of stable equilibrium to be
assumed by the constituents. The resulting structure may be conveniently called
the "normal" structure and it will be so called in these lessons.

Grading of Steel vs. Carbon Content. Steel is generally graded according to
the amount of carbon it contains. The following terms are those most commonly
used :

Very low carbon steel, very mild or extra mild steel,

very soft or dead soft steel carbon not over 0.10 per cent

Low carbon steel, mild steel, soft steel carbon not over 0.25 per cent

Medium high carbon steel, half hard steel .... carbon 0.26 to 0.60 per cent

High carbon steel, hard steel carbon over 0.60 per cent

Very high carbon steel, very hard or extra hard steel carbon over 1.25 per cent

This classification is somewhat arbitrary as there are no sharp lines of demarcations
universally recognized between the various grades.

It will be seen in another lesson that steel containing about 0.85 per cent carbon
is also known as eutectoid steel, steel containing less carbon as hypo-eutectoid steel
and more highly carburized metal as hyper-eutectoid steel.

Low Carbon Steel vs. Wrought Iron. As already mentioned the distinction
between low carbon steel and wrought iron is based upon the difference between the
methods employed for their respective manufacture rather than upon unlike chemical
or physical properties, for these metals may indeed be quite identical both physically
and chemically. The mere melting of wrought iron would undoubtedly, in accord-
ance with the universally accepted definition of steel, convert it into steel since we
would now have a malleable metal initially cast. Such treatment would of course
result in the elimination of the slag mechanically retained by the wrought iron: the
melted metal would be slagless, barring cemented steel, another essential property of
steel. Since wrought iron generally contains but a small amount of carbon, melting
it would convert it into low carbon steel.

2 LESSON IV LOW CARBON STEEL

The Structure of Low Carbon Steel. From the above considerations regarding
the resemblance between wrought iron and low carbon steel, the structure of the
latter may fairly be anticipated. Seeing that low carbon steel may be considered as
wrought iron from which the mechanically held slag has been expelled through melt-

Fig. 1. Steel. Carbon 0.08 per cent.
Magnification not stated. (Arnold.)

Fig. 2. Steel. Carbon about 0.20 per cent.
Magnified 200 diameters. (Guillet.)

ing, we should expect the absence of slag to be the only marked difference between
the structure of low carbon steel and that of wrought iron.

The microstructure of low carbon steel in the case of samples containing respec-
tively about 0.10 and 0.20 per cent carbon is illustrated in Figures 1 to 5. It will be

ill "&^i&

-M.;y,- '.,V:-Y

Fig. 3. Steel. Carbon 0.10 per cent. Magnified 100
diameters. (Boynton.)

seen (Figs. 1, 2, and 3) to consist chiefly of a mass of ferrite (carbonless iron) exhibiting
the usual polyhedral crystalline grains described in preceding lessons. The ferrite pres-
ent in low carbon steel is similar in every respect to the ferrite of wrought iron. At
the junctions of many ferrite grains, however, some dark areas will be noted, an evi-

LESSON IV LOW CARBON STEEL 3

dence of the existence in the metal of another constituent. Since ferrite is practically
free from carbon, it is evident that the carbon present in the steel must have segre-
gated into these small dark masses. As to the exact nature of this dark constituent
it will be apparent that it cannot consist of pure carbon for it is well known that the
carbon present in steel does not exist in the free state but on the contrary is combined
with some of the iron forming a definite chemical compound or carbide of iron whose
formula is FesC. 1 This iron carbide must necessarily be located in the dark areas,
but are these" made up exclusively of this carbide? To find an answer to this question
let us examine the structure of steel under a higher magnification (Figs. 1 and 5).
This reveals the existence of two components in each dark particle occurring as small
wavy or curved parallel plates or lamellae alternately dark and" white. As to the
nature of these two components, it is evident that one of them must be the carbide
FesC and the other necessarily iron or ferrite, since according to the proximate analy-
sis of steel, these are the only two constituents which, to the best of our knowledge,
are present in pure unhardened carbon steel.

Pearlite. Howe named the microscopical constituent just described pearlite
(originally written pearlyte) following in this Dr. Sorby who was the first observer to
describe it and who had proposed for it the name of "pearly constituent" because it
frequently exhibits a display of colors very suggestive of mother-of-pearl, especially
when viewed by oblique illumination. This appearance is due to the fact that these
plates are extremely thin, seldom measuring over 25000 of an inch in thickness, and
that the plates of carbide, being much harder than the ferrite plates, stand in relief
after polishing, resulting in an arrangement very similar to the refraction gratings of
physicists. Mother-of-pearl likewise is made up of very thin alternate plates of dif-
ferent colors and possibly of different hardness. The carbide plates remain bright
not being affected by the usual etching reagents, while the ferrite plates appear dark
because of their being somewhat tarnished by the etching and also because, being
depressed owing to their greater softness, they stand in the shadow of the carbide
plates. It will be shown in another lesson that in many series of alloys of two metals
the alloy of lowest melting-point called the "eutectic" alloy, nearly always exhibits
a composite structure like that of pearlite, i.e. made up of parallel plates alternately
of one and the other constituents. It will also be shown that in spite of this very
great structural resemblance pearlite is not a true eutectic alloy. Howe proposed
to call "eutectoid" the kind of mechanical mixture found in pearlite and this most
appropriate term has been universally adopted.

Because of the minute dimensions of the lamellae of pearlite a high magnifica-
tion, generally not less than 250 or 300 diameters, is required for its resolution.

It should be stated here that pearlite does not always assume such a distinctly
laminated structure. In many instances its structure remains ill defined or has a
granular rather than a lamellar appearance, while its behavior towards the etching
reagents likewise varies. It will be shown at the proper time that this is due to the
treatments to which steel may be subjected and that the exact nature of these ill-de-
fined forms of pearlite (often called transition constituents) has given rise to a large
amount of discussion and has been the object of many investigations. It may be

1 The existence of the carbide Fe 3 C in unhardened steel was first shown in 1885 by Abel and
Muller, working independently, and has since been confirmed by many other investigators. Its
existence is proved by dissolving unhardened steel in a suitable solvent and analyzing the carbona-
ceous residue.

4 LESSON IV LOW CARBON STEEL

assumed for the present that any pearlite which is not distinctly lamellar is not true
pearlite.

It will be noted that in Figure 2 the pearlite occupies about twice the area covered
by the same constituent in Figure 1. We infer from this that the amount of pearlite
in low carbon steel at least increases progressively with the carbon content. Doub-
ling the amount of carbon doubles of course the proportion of the iron carbide in the
steel, and since the amount of pearlite is apparently also doubled it follows that iron
carbide and ferrite must unite with each other in fixed ratio to form pearlite, in other
words that pearlite always contains the same proportion of carbide and hence also
of carbon. The accuracy of this conclusion will soon be shown.

Free Ferrite. To distinguish between the ferrite included in pearlite and the
ferrite forming the balance of low carbon steel, the latter is sometimes called "free"

Fig. 4. Steel. Structure of pearlite. Magnified
1000 diameters. (Osmond.)

Fig. 5. Steel. Hypo-eutectoid.
Magnified 750 diameters. Pear-
lite particles and surrounding
ferrite. (Goerens.)

ferrite, "structurally free" ferrite, "excess" ferrite, "massive" ferrite, "non-eutec-
toid" or "pro-eutectoid" ferrite, "surplus" ferrite. In these lessons it will be referred
to as free ferrite while the ferrite forming part of the pearlite will be called pearlite-
ferrite. Some writers refer to the latter as eutectoid-ferrite.

In the absence of any conclusive evidence to the contrary, it is natural to infer
that free ferrite and pearlite-ferrite are identical, that is, pure iron in pure steel and
iron holding in solution small quantities of silicon and phosphorus, and possibly of
other impurities, in impure (commercial) steel.

This has been doubted by some writers, however, who have noted that the ferrite of some pear-
lites was more readily colored on etching than free ferrite and they saw in this and in some other
evidences an indication that pearlite-ferrite may be less pure than free ferrite. Benedicks, for in-
stance, believes, or at least believed at one time, that the pearlite-ferrite of some steels could con-
tain as much as 0.27 per cent of carbon dissolved in beta iron, whereas free ferrite is in the alpha
condition. This carburized and allotropic ferrite Benedicks called "ferronite."

LESSON IV LOW CARBON STEEL 5

Cementite. The name of cementite has been given by Howe to the carbide
Fe 3 C and universally adopted. The term is derived from "cement" steel (cementa-
tion steel, blister steel, converted steel) which being generally a high carbon steel
contains a great deal of this carbide, that is, of cementite.

According to the atomic weights of iron (56) and of carbon (12) cementite must
contain

12 x 100

- = 6.67 per cent carbon
3 x 56 + 12

The carbon present in cementite is frequently referred to as "cement" carbon,
occasionally as carbide carbon, to distinguish it from other forms^pf carbon found in
iron and steel and to be described later (hardening carbon, graphitic carbon, temper
carbon, etc.).

Cementite is an extremely hard substance, being in fact the hardest of all the
constituents occurring in iron and steel, harder even than hardened, high carbon steel.
Howe states that it is harder than glass and nearly as brittle. As it scratches feldspar
but not quartz it is generally assigned to rank 6 or 6.5 in the Mohs scale of hardness. 1

It will be shown later that when steel contains an appreciable amount of manga-
nese, as is nearly always the case in commercial products, a portion at least of this
manganese also forms anapaa carbide Mn 3 C and that this carbide unites with the iron
carbide Fe 3 C to form cementite. It is well to bear in mind, therefore, that in com-
mercial steel cementite generally contains besides Fe 3 C varying amounts of this car-
bide of manganese. As the atomic weight of manganese is nearly the same as that of
iron, 55 compared to 56, it so happens that the presence of manganese in cementite
affects but very little its carbon content, which for all practical purposes may be
taken as 6.67 regardless of the amount of manganese it may contain. Cementite
containing much manganese has been called manganiferous cementite by some writers.

Whether any portion of the other impurities present in iron and steel (sulphur,
silicon, phosphorus) is ever included in cementite is not positively known but in the
absence of indications to the contrary it is generally assumed that cementite is free,
practically at least, from these metalloids.

Cementite generally remains bright and brilliant after the ordinary etching treat-
ments employed to reveal the structure of steel. It will be shown later, however,
that some special reagents may be used which color it deeply.

Experiments

The student should procure samples of forged steel containing respectively about
0.10 and 0.20 per cent carbon. These should be heated to 1000 deg. C. and slowly
cooled, preferably with the furnace in which they were heated.

Polishing. Specimens should be cut from these samples of suitable size for
microscopical examinations (preferably not over }/ in. square or round and J/ in.
thick). These specimens should be polished for examination in accordance with the
instruction given in Lesson III, taking care to prepare a freshly cut surface, that is a
portion of the sample which did not suffer from decarburization in the furnace.

1 The Mohs scale is as follows, beginning with the softest and ending with the hardest mineral
and each mineral being capable of scratching the preceding ones: (1) Talc, (2) Gypsum, (3) Calcite,
(4) Fluorite, (5) Apatite, (6) Feldspar, (7) Quartz, (8) Topaz, (9) Corundum, and (10) Diamond.

6 LESSON IV LOW CARBON STEEL

Etching. These samples should be etched successively with (1) a solution of
picric acid in absolute alcohol, (2) a solution of nitric acid in absolute alcohol, and
(3) concentrated nitric acid, following the instructions for etching, washing, drying,
etc., given in Lesson III. They should be carefully examined after each etching and
the treatment repeated in case the structure does not appear sharply and clearly
defined.

Examination. The prepared specimens, suspended to the magnetic holder,
should first be examined with a low power objective (^ in. or 16 mm.) and eyepiece
(2 in. or 5X) a combination which will yield a magnification of about 50 diameters.
The central portion of the specimens should be observed and their structure com-
pared with the illustrations of similar steels reproduced in this lesson. It will be in-
structive to examine also the edges of the samples and to note that the outside of
the bars have been somewhat decarburized through the heating operation, unless
indeed the bars had been effectively protected against oxidation. This decarburiza-
tion will be apparent from a decrease in the proportion of pearlite.

The contrast between the bright ferrite and the dark areas of pearlite should be
very marked, and the junction lines between the ferrite grains should appear like a
delicate but distinct network. If these appearances lack intensity the etching treat-
ment should be repeated without repolishing. While a deeper etching, however, will
bring out more distinctly the junction lines between the ferrite grains, it will some-
what blur the structure of pearlite. If the structure is ill defined the sample should
be rubbed a short while on the rouge block or disk and the etching repeated. The
most satisfactory etching is the one which will show great contrast between the ferrite
and pearlite, while bringing out somewhat faintly the ferrite grains.

To reveal the composite structure of pearlite a higher magnification is needed.
To that end a 4 mm. or }/ in. objective and a 1 in. eyepiece will be found satisfac-
tory, as this will yield a magnification of about 430 diameters.

Examination under high power requires careful adjustment of the light and ver-
tical illufninator and careful focusing. The parallel plates of pearlite should be
clearly seen, although it is not always possible to resolve satisfactorily every particle
of that constituent.

Photomicrography. The student should proceed to take low power photomicro-
graphs of the two samples of wrought iron and two samples of steel which he has
so far prepared and examined.

For the taking of photomicrographs the appliances described and illustrated under
"Apparatus" are recommended.

After having selected the spot to be photographed, light tight connection should
be established between the microscope and the camera (in the case of the Metalloscope
the camera and microscope are permanently connected) and the light carefully ad-
justed so as to obtain on the screen of the camera as bright and even an illumination
as possible. The image should now be focused as sharply as possible, using a focus-
ing cloth if necessary and gently turning the fine adjustment screw of the stand. A
focusing glass may be used with great advantage for this operation and is of special
importance when photographing with high power objectives. It should be placed on
the plain glass circle which occupies the center of the screen of the camera, and the
image focused while being viewed through this lens. By this means we magnify the
image formed upon the camera screen, and are therefore able to focuss it more sharply
in its finer details. Considerable light, however, is lost and the object will often ap-

LESSON IV LOW CARBON STEEL 7

pear but dimly lighted. The rule is to secure the clearest possible image while work-
ing tentatively the fine adjustment in both directions, bearing in mind that, at its
best, the image may. appear blurred and dimly lighted.

An ordinary eyepiece may be used in place of a focusing glass with, in many
cases, satisfactory results.

Exposure. After the image has been properly illuminated and focused the sen-
sitive plate should be introduced and exposed, with the ordinary precautions, for a
suitable length of time. The required time of exposure will vary according to (a)
the kind of photographic plate used and (6) the amount and nature of light reaching
the plate, which in turn will depend upon (1) the nature of the prepared surface,
especially its power to reflect light, (2) the kind of illumination-used, (3) the position
of the diaphragm or diaphragms controlling the amount of light allowed to reach the
plate, (4) the kind of light filters used, if any, (5) the resolving and magnifying powers
of the combination of objective and eyepiece used, and (6) the distance between the
screen of the camera and the object.

Specimens which after etching remain quite bright naturally reflect more light
and consequently require for their photography a shorter time than duller specimens.
Generally speaking the higher the magnification the less light, hence the longer the
exposure. The use of colored screens or solutions (light filters) as a rule lengthens the
exposure considerably. By placing the screen of the camera at a greater distance
from the object (i.e. by extending the bellows of the camera) the magnification is
increased but with accompanying loss of light, and, therefore, increased length of
exposure.

Using a rather slow plate, no screens, a combination of objective and eyepiece
yielding a magnification of 100 diameters at a distance of 24 to 30 inches from the
object, the diaphragm being wide open, the exposure for most iron and steel samples
would vary between a fraction of a second with a powerful arc lamp and some 10 to
20 minutes with a welsbach lamp. The use of higher magnifications, of screens, and
the closing of the diaphragm may lengthen the exposure to such an extent as to re-
quire one minute or more with an arc lamp and one hour or more with a welsbach
lamp.

The use of sources of light of greater intensity than the welsbach mantle but
less intense than the electric arc, such as the Nernst lamp, the acetylene lamp, or
the oxy-hydrogen lamp, calls for exposures of intermediate lengths between the two
extremes considered in the preceding paragraph.

Diaphragms and Shutters. It is sometimes advantageous to be able to control
the pencil of light entering the illuminator with a view of securing sharper definition.
To that effect an iris diaphragm suitably mounted should be placed between the
condensing lens or lenses and the vertical illuminator. Some sort of an automatic
shutter is convenient to control the exposure of the plates. This shutter may ad-
vantageously be combined with the iris diaphragm. Instead of being placed between
the source of light and the vertical illuminator, diaphragms and shutters are some-
times inserted between the camera and the microscope. The best disposition con-
sists in placing an iris diaphragm between the condensing lenses and the vertical
illuminator, thus controlling the amount of light entering the latter, and another,
diaphragm combined with automatic shutter between the camera and microscope tube.

Monochromatic Light. The different sources of light used for microscopical
work yield white light and since the correction, even of apoehromatic objectives, for

LESSON IV LOW CARBON STEEL

chromatic aberration is never perfect, it is evident that the use of monochromatic
light, i.e. light of one wave length, is theoretically preferable, especially for photo-
graphing.

Monochromatic light may be obtained in two ways : (a) by using a source of light
actually monochromatic and (6) by causing white light to pass through colored glass
screens or colored solutions (light filters), preventing the passage of some undesirable
rays. The mercury arc lamp yields a nearly monochromatic light and has been tried
by Le Chatelier with satisfactory results. It seems more convenient, however, when
monochromatic light is wanted, to use light filters of suitable colors, in which case
colored glass screens will be found easier to manipulate than glass cells containing
colored solutions.

The beginner is advised to dispense with the use of colored screens or other ray
filters until he has acquired experience and skill in taking photomicrographs, when he
will be better qualified to judge of their merits and to employ them intelligently.

Photographic Plates. The use of so-called "Process" or "Contrast" plates is
recommended. These plates are slow but generally yield negatives with sharp con-
trasts. Orthochromatic plates may also be used with, in some instances, excellent
results. These plates are much more rapid but as they call for the use of a colored
screen, the time of exposure may be even longer than with the slower kind.

Development. Formula and directions accompany each box of plates and the
student could not do better than to follow them faithfully.

Printing. Any printing out or developing paper may be used, the printing, de-
veloping, or toning being conducted in the usual way. Drying on ferrotype plates
affords a quick means of finishing the prints and giving them a satisfactory luster.
It is recommended to trim the prints round, 2 to 2^ inches in diameter, by means
of a margin trimmer and suitable circular forms, as this will give them a very neat
appearance.

Mounting. It is well to paste the prints on suitable cardboard mounts afford-
ing room for the recording of useful data. A very satisfactory mount has been
illustrated under "Apparatus."

Examination

I. Describe the structure of low carbon steel and more especially of pearlite.
II. What is free ferrite?

III. Describe the structure of your samples and mention any difficulty encountered
in polishing, etching, or photographing them.

LESSON V

MEDIUM HIGH AND HIGH CARBON STEEL

Medium High Carbon Steel. The normal structure of steel (i.e. its structure
after forging, reheating to a high temperature and slow cooling) containing about
0.30 per cent carbon is illustrated by a drawing in Figure 1 and by a photomicro-
graph in Figure 2. It will be noted, on comparing this structure to that of lower
carbon steels (Lesson IV), that the introduction of more carbon in the iron has re-
sulted, as would be expected, in the occurrence of a greater amount of pearlite and of
a correspondingly smaller proportion of ferrite. The pearlite occupies now roughly
about one third of the total area. The junction lines between the grains of ferrite

Fig. 1. Steel. Carbon 0.38 per cent.

Magnification not stated.

(Arnold.)

Fig. 2. Steel. Carbon 0.33 per cent. Mag-
nified 100 diameters. Heated to 1000 deg.
C. and slowly cooled in furnace. (Hall.)

should be noted. Under sufficiently high power the pearlite areas exhibit the char-
acteristic lamellar structure described in Lesson IV.

On further addition of carbon, the amount of pearlite, which is evidently propor-
tional to the percentage of carbon, increases correspondingly, as shown in Figures 3
and 4 illustrating the microstructure of steel containing about 0.50 per cent carbon.
The pearlite occupies here over one half of the total area. It will be noticed that
the ferrite areas are only occasiqnally resolved into polyhedral grains, apparently
because the ferrite now occurs in particles often too small to be made up of several
crystalline grains. These small masses of ferrite, however, are still made up of crys-
talline matter as described and illustrated in Lesson II. A high power photomicro-

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

graph of 0.45 per cent carbon steel is shown in Figure 5. The laminations of pearlite
are clearly seen.

Fig. 3. Steel. Carbon 0.59 per cent.

Magnification not stated.

(Arnold.)

Fig. 4. Steel. Carbon 0.50 per cent. Mag-
nified 100 diameters. Heated to 1000 dcg.
C. and slowly cooled in furnace. (Burger.)

Fig. 5. Steel. Carbon 0.45 per cent. Magnified
1000 diameters. (Osmond.)

When steel contains but a small, although appreciable, amount of ferrite, as is
the case with carbon contents between 0.50 and 0.70 per cent the ferrite frequently
forms envelopes or membranes surrounding the pearlite grains, an arrangement

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 3

generally described as a network structure the pearlite forming the meshes and the
free ferrite the net proper. These pearlite meshes are also described as "cells" or
"kernels" and the ferrite membranes as "cell walls" or "shells."

Fig. 6. Steel. Carbon 0.50 per cent. Magnified
100 diameters. Heated to 1000 deg. C. and
cooled in air. (Burger.)

Fig. 7. Steel. Hypo-euteetoid. (Sorby.)

It will be shown later that this network structure is promoted by rather rapid
cooling from a high temperature, as for instance by cooling small pieces in air.

Structures of this type are illustrated in Figures 6 and 7. The latter illustration
is of special interest being a reproduction of one of Sorby's original drawings and
therefore, the first drawing of pearlite ever published.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

High Carbon Steel. Since the introduction of increasing amounts of carbon in
steel results in the formation of a correspondingly^ihcreasing proportion of pearlite
and decreasing proportion of ferrite, a degree of carburization must necessarily be
reached, when the whole mass will be mlde up of pearlite, the ferrite having finally
disappeared. This critical point in the structure of steel is attained when the metal
contains somewhere between 0.80 and 0.90 per cent of carbon, exceptionally pure
steel requiring the larger proportion of carbon and impure steel the smaller for the
complete disappearance of ferrite.

Eutectoid Steel. Steel made up exclusively of pearlite is now quite universally
called " ejotectoid " steel, after Howe, the name suggesting the great resemblance
between pearTTEe and euctectic alloys, while, at the same time, clearly indicating
that pearlite is not a real eutectic alloy. Previous to Howe's happy suggestion

Fig. 8. Steel. Carbon 0.89 per cent.
Magnification not stated. (Arnold.)

Fig. 9. Steel. Eutectoid. Magnified
750 diameters. (Goerens.)

this steel was commonly described as "eutectic" or "saturated" steel. It has also
been termed "aeolic" or "benmutic" steel but these names have now been aban-
doned. The structure of eutectoid steel is illustrated in Figures 8 and 9.

Steel containing less than 0.85 per cent carbon or thereabout, and in which,
therefore, some free ferrite is present, is called "hypo-eutectoid," while steel more
highly carburized than eutectoid steel is called "hyper-eutectoid." It will be shown
presently that hyper-eutectoid steel contains free cementite.

Hyper-Eutectoid Steel. The normal structure of steel containing from 1.10 to
1.50 per cent carbon is illustrated in Figures 10 to 13 both under low and high mag-
nification. These steels will be seen to consist, like hypo-eutectoid steel, of two con-
stituents, one of which being pearlite as clearly shown when examined under high
power. The other constituent remain's bright after etching and might at first be
taken for ferrite. Upon reflection, however, it will be evident that such cannot be its
nature. The light constituent of hyper-eutectoid steel consists of cementite which
is now present in excess over the amount required to form pearlite, just as in hypo-

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 5

eutectoid steel, ferrite is the excess constituent. It will be evident that the ferrite
and cementite which constitute all grades of carbon steels combine with each other
in suitable proportions to form pearlite, leaving, as the case may be, an excess of

Fig. 10. Stool. Carbon 1.20 per cent.
Magnification not stated. (Arnold.)

Fig. 11. Steel. Carbon 1.10 per cent. Magnified 100 diameters.
(Boynton.)

ferrite (in hypo-eutectoid steel) or of cementite (in hyper-eutectoid steel). A more
scientific explanation of the formation of the normal structure of steel will be offered
in a subsequent lesson.

Free Cementite. To distinguish between the cementite forming part of the
pearlite (the bright plates of that constituent) and the cementite constituting the

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

balance of hyper-eutectoid steel, the latter is generally called "free" cementite,
"structurally free" cementite, "excess" cementite, "massive" cementite, "non-
eutectoid" cementite, "surplus" cementite, while the cementite included in the
pearlite is sometimes referred to as pearlite-cementite or eutectoid cementite. In
these lessons the cementite in excess over the eutectoid ratio will be called free cemen-
tite.

In the absence of any conclusive evidences to the contrary, and in conformity
with the nature of eutectic alloys in general, it is assumed that free cementite and
pearlite-cementite are identical in composition and properties.

As already stated in Lesson IV the cementite of commercial steel is not pure
Fe 3 C but contains small and varying amounts of Mn 3 C.

Fig. 12. Steel. Carbon 1.43 per cent.

(Boynton.)

Magnified 50 diameters.

As shown in Figures 10 and 11 hyper-eutectoid steel like hypo-eutectoid steel
may assume a network structure. In both cases the meshes consist of pearlite but
the net proper which in hypo-eutectoid steel represents membranes of free ferrite
indicate now the occurrence of membranes of free cementite.

Hypo- vs. Hyper-Eutectoid Steel. While there is considerable similarity be-
tween the structure of steel containing but a slight excess of ferrite and the structure
of steel containing but a slight excess of cementite, a little experience and careful
examination will reveal differences in their appearances and properties which will
make it possible, generally, to distinguish between them. Cementite has a more
metallic luster than ferrite and remains bright and structureless, even after prolonged
etching with the ordinary reagents, while ferrite is colored and, if present in suf-
ficiently large masses, resolved into grains by such treatment. Cementite is extremely
hard, standing in relief, while ferrite being soft is depressed by the polishing opera-
tion. Ferrite is readily scratched by a needle drawn across the polished surface while

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 7

cementite remains unmarked. Again it will be noted that in Figures 10 and 12 some
of the pearlite grains are cut by plates or needles of cementite, independent of the net-
work of cementite, while when the network consists of ferrite, needles of ferrite are
frequently observed penetrating the pearlite grains, but for the most part connected
with the network itself (Fig. 6).

When cementite is in excess the pearlite grains are generally smaller and the net-
work finer than is the case with excess of ferrite. Free cementite does not, of course,
always assume the shape of a fine network. It will be shown in subsequent lessons
that its mode of occurrence depends upon the treatment to which the steel was sub-
jected.

Etching of Cementite. It has been seen that cementite is not acted upon by the
usual reagents employed in the etching of steel sections (picric acid, nitric acid, tine-

Fig. 13. Steel. Carbon 1.43 per cent.
(Boynton.)

Magnified 500 diameters.

ture of iodine, etc.) but that on the contrary it remains brilliant and structureless.
Kourbatoff, however, discovered a reagent which deeply colors cementite while leav-
ing the ferrite unaffected (Fig. 14), thus affording a sure means of distinguishing
between the two. The treatment consists in immersing the polished sample in a
boiling solution of sodium picrate in an excess of sodium hydroxide for some 5 to 10
minutes, when the cementite assumes a brown to blackish coloration. The etching
solution may be prepared by adding 2 parts of picric acid to 98 parts of a solution
containing 25 per cent of caustic soda, for instance 2 grams of picric acid in 98 cubic
centimeters of a solution made up of 24.5 grams of caustic soda and 73.5 cubic centi-
meters of water.

More recently Matweieff has recommended the use of a 2 per cent solution of
oxalate of ammonium, used cold for 30 minutes, which colors the cementite red
(Fig. 15).

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

Carbon Content of Pearlite. The percentage of carbon in pearlite, and, there-
fore, in eutectoid steel, has been stated to be somewhere between 0.80 and 0.90 in
commercial steel of ordinary quality, because when steel of that degree of carburiza-
tion is examined under the microscope it is found to be free from any appreciable
amount of free ferrite or of free cementite. As the presence of a very small amount
of any of these two constituents in the free state, however, is very difficult to ascer-
tain, it will be evident that it is quite impossible to speak positively as to the exact
amount of carbon needed to exclude both free ferrite and free cementite from the
structure. Moreover this carbon content of pearlite varies somewhat with the com-

Fig. 14. Steel. Hyper-eutectoid. Free
cementite colored dark by sodium picrate.
Magnified 500 diameters. (Guillet.)

Fig. 15. Steel. Hyper-eutectoid. Free cemen-
tite colored dark by ammonium oxalate. Mag-
nified 142 diameters. (Matweieff.)

position of the steel and with the treatment it has received. In steel of ordinary
commercial purity the eutectoid point appears to be in the vicinity of 0.85 per cent
carbon.

Structural Composition of Steel. Bearing in mind that hypo-eutectoid steel is
composed of free ferrite and pearlite and that hyper-eutectoid steel consists of free
cementite and pearlite, and knowing the proportion of carbon in pearlite (0.85 per
cent?) and in cementite (6.67 per cent), the structural composition of any steel may
be readily calculated, provided we know the percentage of carbon it contains.

In case of hypo-eutectoid steel we have the two following equations :

(1) F + P = 100

(2)

100

in which F represents the percentage of free ferrite in the steel, P the percentage of
pearlite, E the percentage of carbon in pearlite, and C the percentage of carbon in
the steel. The first equation expresses the fact that the steel is composed of ferrite
and pearlite and the second equation the fact that all the carbon in the steel is in-

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 9

eluded in the pearlite. Assuming, for instance, that pearlite contains 0.85 per cent
carbon and the steel 0.50 per cent carbon, the resolution of these two equations indi-
cates that steel of that grade has the following structural composition:

F = per cent free ferrite = 41.8
P = per cent pearlite = 58.2

In case of hyper-eutectoid steel the following two equations may be written :

(1) P + Cm = 100

100 100

in which P represents the percentage of pearlite, Cm the percentage of free cemen-
tite, E the percentage of carbon in pearlite, C the percentage of carbon in the steel.
The first equation expresses the fact that hyper-eutectoid steel is composed of pear-
lite and free cementite and the second the fact that the carbon in the steel is dis-
tributed between the pearlite and the free cementite, forming E per cent of the
pearlite and 6.67 per cent of the cementite. Assuming the value of E to be 0.85 and
the steel to contain 1.25 per cent carbon, these equations give for a steel of that grade

P = per cent pearlite = 93
Cm = per cent free cementite = 7

Supposing that pearlite or eutectoid steel contains 0.85 per cent carbon, since the
whole of that carbon is present in the cementite plates of pearlite and since cemen-
tite contains 6.67 per cent carbon (as called for by its chemical formula FesC), the
percentage of cementite in pearlite may be readily calculated, as follows:

' X per cent cementite = 0.85
100

i on

hence, per cent cementite = 0.85 x - = 12.74

6.67

and per cent ferrite = 100 - 12.74 = 87.26

or roughly 1 part by weight of cementite to 6.6 parts by weight of ferrite.

If it be considered, however, (1) that the exact carbon content of pearlite is not,
and, hardly can be, known, (2) that it varies somewhat both with composition and
treatment, and (3) that in commercial steel it is probably not far from 0.85 per cent,
we are fully warranted to assume, for the sake of the great simplicity it introduces
in the calculations, that pearlite contains exactly 1 part by weight of cementite to
7 parts by weight of ferrite, which would be the case if the eutectoid point, corre-
sponded to 0.834 per cent carbon, as indicated below:

1 part cementite + 7 parts ferrite yields 8 parts pearlite
or 12.50 per cent cementite + 87.50 per cent ferrite = 100 per cent pearlite

and since cementite contains 6.67 per cent carbon, 12.50 per cent cementite will con-
tain 6.67 X .1250 = 0.834 per cent carbon. Assuming then that such is the carbon
content of eutectoid steel, so that 1 part of cementite gives exactly 8 parts by weight
of pearlite and, noting that the carbon in the steel produces exactly 15 times its own

10 LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

weight of cementite, 1 the calculation of the structural composition of any steel be-
comes extremely simple.

In case of hypo-eutectoid steel (steel containing less than 0.834 per cent carbon)
we have

per cent total cementite = per cent total carbon x 15
and per cent pearlite = per cent total cementite x 8

or, more simply,

per cent pearlite = per cent carbon x 120
i.e. P = 120 C

and, of course, per cent ferrite = F = 100 - P.

With hyper-eutectoid steel (steel containing more than 0.834 per cent carbon) the
figuring is as follows:

Since 8 parts of pearlite contains 7 parts of ferrite and since in hyper-eutectoid
steel the totality of the ferrite (total ferrite) is included in the pearlite (there being
no free ferrite) we have

per cent pearlite = P = f total ferrite
or, since total ferrite = 100 total cementite,
P = f (100 - total cementite)

But total cementite = carbon x 15, hence

P = f (100 - 15 C)

800 - 120 C
or P = -

and, of course, free cementite = Cm = 100 P.

Summing up, in order to find the percentage of pearlite in hypo-eutectoid steel it
will suffice to multiply its carbon content by 120 (P = 120 C), the balance of the steel,
consisting, of course, of free ferrite (F = 100 P) ; to find the percentage of pearlite
in hyper-eutectoid steel, the percentage of carbon in the steel should be substituted

800 120 C
for C in the formula: P = - - and the balance of the steel will be made -up

of free cementite (Cm = 100 - P).

Taking, for instance, a steel containing 0.50 per cent carbon. Its structural com-
position will be:

120 x 0.50 = 60 per cent pearlite, and

100 - 60 = 40 per cent ferrite

If a steel contains 1.25 per cent carbon the resulting percentage of pearlite will

son _ 190 v i 9^
be - -or nearly 93 per cent and the free cementite (Cm), 100 - 93 = 7

per cent.

1 This follows from the composition of FesC indicated by the atomic weights of iron and carbon:
(3 x 56) Iron + 12 Carbon = 180 Fe 3 C

180

hence one part carbon produces = 15 parts FesC or cementite.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 11

In these lessons it will be assumed for the sake of the simplicity it introduces that
pearlite contains 0.834 per cent carbon, -that is exactly 1 part by weight of cementite
to 7 parts of ferrite.

Chemical vs. Structural Composition. Disregarding for the present the existence
of impurities, the ultimate analysis of steel reveals the presence of so much carbon
and so much iron. The proximate chemical analysis of steel reveals (in steel slowly
cooled from a high temperature) the presence of so much iron and so much carbide
of iron, Fe 3 C. In a similar way we may consider two different structural composi-
tions, an ultimate and a proximate one. The ultimate structural composition reveals
the presence of so much total ferrite and so much total cementite, while the proxi-
mate structural composition informs us of the percentages of pearlite, free ferrite,
and free cementite in the steel. It will be evident that the chemical proximate com-
position is identical to the ultimate structural composition, the names of the con-
stituents only being different, iron and carbide in the first case, ferrite and cementite
in the latter.

These various compositions are tabulated below:

Constituents

ultimate Fe C

Chemical Composition . _, _,

proximate J? e b e 3 C

ultimate total ferrite total cementite
Structural Composition . ... .. . , ...

proximate pearlite tree territe free cementite

It is apparent that the proximate structural composition affords more valuable
information than is obtainable through the other three kinds of analysis, for not only
does it indicate the chemical nature of the proximate constituents but also their
structural association and occurrence, upon which depend, to a very great extent, the
physical properties of steel. In the following table (page 12) the ultimate chemical
composition as well as the structural composition, both ultimate and proximate, of
steel containing from 0.1 to 2.0 per cent of carbon, have been calculated for each
increase of carbon of 0.1 per cent. Corrections for variations of 0.01 per cent carbon
can readily be obtained by interpolation and are indicated in a second table. The
values given for the proximate compositions are based upon the assumption that
pearlite contains 0.834 per cent carbon.

These compositions are shown also diagrammatically in Figure 16 which will be
readily understood. ABC represents the free ferrite in hypo-eutectoid steel, ACD the
pearlite in hypo-eutectoid steel, DCEF the pearlite in hyper-eutectoid steel, DFG the
free cementite in hyper-eutectoid steel, ABEH the total ferrite in any steel, AHG
the total cementite in any steel, ACEH the pearlite-ferrite in any steel, and AHFD
the pearlite-cementite in any steel.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

CHEMICAL COMPOSITION

STRUCTURAL COMPOSITION

ULTIMATE

PROXIMATE

Total Cementite

Total Ferrite

Pearlite

Free Ferrite

Free Cementite

99.9

1.5

98.5

12.0

88.0

99.8

3.0

97.0

24.0

76.0

99.7

4.5

95.5

36.0

64.0

99.6

6.0

94.0

48.0

52.0

99.5

7.5

92.5

60.0

40.0

99.4

9.0

91.0

72.0

28.0

99.3

10.5

89.5

84.0

16.0

99.2

12.0

88.0

96.0

4.0

99.1

13.5

86.5

98.7

1.3

1.0

99.0

15.0

85.0

97.0

3.0

1.1

98.9

16.5

83.5

95.3

4.7

1.2

98.8

18.0

82.0

93.6

6.4

1.3

98.7

19.5

80.5

91.9

8.1

1.4

98.6

21.0

79.0

90.2

9.8

1.5

98.5

22.5

77.5

88.5

11.5

1.6

98.4

24.0

76.0

86.8

13.2

1.7

98.3

25.5

74.5

85.1

14.9

1.8

98.2

27.0

73.0

83.4

16.6

1.9

98.1

28.5

71.5

81.7

18.3

2.0

98.0

30.0

70.0

80.0

20.0

CARBON

HYPO-EUTECTOID STEEL

HYPER-EUTECTOID STEEL

Values to be added to % of pearlite and
subtracted from % cementite

Values to be subtracted from % pearlite and
added to % cementite

0.01

1.2

0.17

0.02

2.4

0.34

0.03

3.6

0.51

0.04

4.8

0.68

0.05

6.0

0.85

0.06

7.2

1.02

0.07

8.4

1.19

0.08

9.6

1.36

0.09

10.8

1.53

Micro-Test for Determination of Carbon in Steel. Since the amount of pearlite
in steel is proportional to the percentage of carbon it contains, it should be possible to
estimate the latter with a fair degree of accuracy from the area occupied by the
pearlite. After a little experience and by taking the necessary precautions it will be
found that, in the case of decidedly hypo-eutectoid steels at least (steels containing
say less than 0.60 per cent carbon), results are obtained fully as accurate as those of
the colorimetric method and, on the whole, more reliable, since the possibility of
serious errors is practically eliminated. By the micro-test, for instance, a steel with
0.25 per cent carbon might be reported as containing 0.20 or 0.30 per cent of that

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

o o

ftj n

N cy

14 LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

element, but it could hardly be reported as containing 0.15 or 0.35 per cent. With
chemical methods on the contrary, even with the combustion method, such errors are
possible, and occasionally occur, through mechanical loss, faulty manipulations, im-
pure reagents, mistakes in weighing or figuring, etc. Chemical analysis calls for the
complete destruction of the anatomy of the metal, destroying at the same time evi-
dences of serious error; micrographic analysis, on the contrary, is based upon the
anatomy itself and therefore very serious errors are quite impossible.

In order to yield results at all satisfactory, however, care should be taken that all
samples be first annealed, that is reheated to 900 or 1000 deg. C. and cooled slowly,
so that the normal amounts of pearlite may be formed. To attempt to apply the
micro-test to forged samples for instance is certain to lead to failure. Nor can the
test be applied to hypo-eutectoid steel containing but a slight amount of free ferrite,
for instance to steel with from 0.60 to 0.80 per cent carbon, because of the difficulty
of estimating accurately the area occupied by so small a proportion of the constit-
uent in excess, and therefore by the pearlite itself. In the case of hyper-eutectoid
steel, the differences between the contents of free cementite in steels of materially
different carbon contents is so small as to resist accurate determination. For instance,
steels respectively with 1.10 and 1.40 per cent carbon will contain 95.3 and 90.2 per
cent pearlite, a difference of less than 5 per cent in their contents of pearlite, a quan-
tity too small to be estimated with satisfactory accuracy under the microscope.

To sum up, the micro-test for the determination of carbon in steel, if it is to replace
chemical determinations, should be applied only to steels containing less than some
0.60 per cent carbon which have been annealed as above stated.

The author has found the following method to yield in some instances satisfactory
results: the sample after annealing and quick polishing and etching (a few small
polishing scratches will not matter) is placed under the microscope, using a 16 mm.
objective and a 5X eyepiece, and its image thrown on the screen of the camera. In
place of the ordinary camera screen, however, another screen is substituted of ground
glass, ruled into 81 squares (9x9), so that every square covered by pearlite evidently
means very nearly 0.01 per cent of carbon in the steel (exactly 0.01 per cent carbon
if we assume pearlite to contain 0.81 per cent carbon). It is then sufficient to esti-
mate the number of squares occupied by pearlite to arrive at the carbon content of
the steel. The results may be checked by estimating the carbon in two or more dif-
ferent spots and reporting the average if the agreement is sufficiently close.

Physical Properties of the Constituents of Steel. It will now be timely and
profitable to inquire into the physical properties of the three constituents, ferrite,
cementite, and pearlite of which steel in its normal condition is composed.

It will be evident that the physical properties of commercial ferrite must resemble
closely those of wrought iron and of very low carbon steel. Ferrite, therefore, is very
soft, very ductile and relatively weak, having a ductility corresponding to an elonga-
tion of at least 40 per cent and a tensile strength of some 50,000 pounds per square
inch. It is magnetic, has a high electric conductivity, and is deprived of hardening
power, industrially speaking at least, since carbonless iron cannot be materially
hardened by rapid cooling from a high temperature.

The properties of pearlite are evidently those of eutectoid steel in its normal, i.e.
pearlitic condition, from which we may infer that pearlite has a tenacity of some
125,000 pounds per square inch, an elongation of some 10 per cent, that it is hard,
and for reasons later to be explained, that it possesses maximum hardening power.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

With the exception of its very great hardness little is positively known as to the
physical properties of cementite. It may be assumed, however, that so hard and
brittle a substance must greatly lack tenacity. Its tensile strength probably does
not exceed 5000 pounds per square inch and may be considerably less, while its
ductility must be practically nil. It possesses no hardening power.

These properties of the constituents of steel in its normal condition are tabulated
below :

CONSTITUENTS

TENSILE STRENGTH
LBS. PER SQ. IN.

ELONGATION

% IN 2 IN.

HARDNESS

HARDENING POWER

Ferrite

50,000 *

Soft

None

Pearlite

125,000

10 *

Hard

Maximum

Cementite

5000 (?)

Very hard

None

Tenacity of Steel vs. its Structural Composition. Knowing the physical prop-
erties of the three constituents of steel, it should be possible to foretell with some
degree of accuracy the physical properties of any steel of known structural com-
position, on the reasonable assumption that these constituents impart to the steel
their own physical properties in a degree proportional to the amounts in which they
are present. The properties of steel made up for instance of 50 per cent ferrite and
50 per cent pearlite should be the means of the properties of ferrite and of pearlite.
Let us assume such reasoning to be correct and let us apply it to the tensile strength
first of hypo-eutectoid steel and then of hyper-eutectoid steel.

The tensile strength (T) of any hypo-eutectoid steel will be expressed by the fol-
lowing formula in function of its structural composition, that is in function of the
percentages of ferrite (F) and pearlite (P) which it contains:

50,000 F + 125,000 P
1 =

100

in which 50,000 represents the tensile strength of ferrite and 125,000 the strength of
pearlite.

Or simplifying:

T = 500 F + 1250 P

or again in terms of pearlite alone, since F = 100 - P

T = 500 (100 - P) + 1250 P
or T = 50,000 + 750 P

or finally in terms of carbon since P = 120 C

T = 50,000 + 90,000 C.

On applying this simple formula to steels containing respectively 0.10, 0.25, and
0.50 per cent carbon we find for these metals tensile strengths respectively of 59,000,
72,500, and 95,000 pounds per square inch. These values agree closely with our
knowledge of the average tenacity of such steels when in a pearlitic condition, and

16 LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

prove the value of the formula derived from the considerations outlined above as to
the relation existing between the physical properties of steel and its structural com-
position. It should be borne in mind that in working out this formula it has been
assumed that pearlite contains 0.834 per cent carbon.

The values obtained for various hypo-eutectoid steel should be accurate only for
steel in what has been termed in these lessons its normal condition, that is steel
which has been forged, reheated to a high temperature, and slowly cooled. It should
be noted, however, as later explained, that steel forged and finished at a fairly high
temperature are practically in this so-called normal condition, so that the formula
may be used, and fair results expected, to calculate the tensile strength of such hot
forged steel. If the steel be forged until its temperature is quite low and, especially,
if it be cold worked, it is well known that its tensile strength is generally increased.
Neither can the formula be used, of course, in the case of hardened steel or of steel
castings. It may, however, be applied to steel castings which have been properly
annealed, when the tensile strength may be brought up to the level of steel forgings
finished fairly hot as explained in another lesson.

Again the formula is of value only in case of commercial steels containing the
usual proportions of impurities especially of manganese. It applies only to steels in
which the percentage of manganese varies roughly with the carbon content from
some 0.20 to 0.80 per cent. The presence of a larger proportion of manganese would
increase the tenacity materially.

Passing to the tensile strength of hyper-eutectoid steel, our ignorance as to the
tenacity of cementite does not permit the writing of a formula with the same degree
of confidence. Let us assume, tentatively, however, that cementite has a tensile
strength of 5000 pounds per square inch and then proceed as we did in the case of
hypo-eutectoid steel.

The tensile strength of any hyper-eutectoid steel may be expressed by the follow-
ing formula in terms of the percentages of pearlite (P) and cementite (Cm) which it
contains :

125,000 P + 5000 Cm

100
or simplifying

T = 1250 P + 50 Cm

or in terms of pearlite only, since Cm = 100 - P,

T = 1250 P + 100 (50 - P)
T = 5000+ 1150 P

800 - 120 C
or since, as previously shown, P = -

T= 5000+ 1150

or simplifying

955,000 - 138,000 C

or approximately T = 136,000 - 20,000 C.

Applying this formula to steels containing respectively 1.25 and 1.50 per cent
carbon, we find for their respective strength 111,000 and 106,000 per square inch,

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 17

which are fair values for the average tenacity of pearlitic steels of those degrees of
carburization. 1

Steel of Maximum Strength. From the preceding considerations it seems evi-
dent that eutectoid steel must possess maximum tensile strength since the influence
of the presence of ever so small an amount of free ferrite in hypo-eutectoid steel or
of free cementite in hyper-eutectoid steel must necessarily be a weakening one, be-
cause of the relative weakness of free ferrite and free cementite as compared to the
strength of pearlite. By most writers, on the other hand, steel of maximum tenacity
is often stated to contain in the vicinity of 1 per cent carbon, that is to be slightly
hyper-eutectoid.

It is not clear, however, that the results upon which the statement is based were
obtained in testing steel in its pearlitic condition. On the contrary it seems probable
that a large number of the steels tested were in a sorbitic rather than in a pearlitic
condition because of relatively quick cooling through the critical range as explained
in a subsequent lesson. And while it appears that pearlitic steel must have its maxi-
mum tenacity when composed entirely of pearlite, it may well be that when in a sor-
bitic condition maximum strength corresponds to a higher degree of carburization,
i.e. 1 per cent, because sorbite may contain and indeed often does contain more carbon
than pearlite. Indeed the cases on record show that when the steels were made pear-
litic through very slow cooling maxium tenacity corresponds closely to the eutectoid
composition. Arnold, for instance, tested a series of very pure carbon steel and after
slow cooling in the furnace from 1000 deg. C he found a very sharp maximum in the
tenacity corresponding to 0.89 per cent carbon. On cooling these same steels in air,
on the contrary, and therefore making them sorbitic, maximum tenacity corresponded
to 1.20 per cent carbon. Harbord likewise ascertained the tenacity of very pure
steels and found after slow cooling (in the furnace) from 900 deg. C that the maxi-
mum tenacity corresponded to 0.947 per cent carbon.

Ductility of Steel vs. Its Structural Composition. From the known ductility,
as expressed by its elongation under tension, of ferrite and the known elongation of

1 Empirical formulas have often been suggested to express the relation between the tenacity of
steel and its carbon content. Deshayes proposed for unannealed steel

T = 30.09 + 18.05 C + 36.11 C 2
Thurston (minimum values) for unannealed steel

T = 42.32 + 49.37
and for annealed steel

T = 35.27 + 42.32
Bauschinger for Bessemer steel

T = 43.64 (1 + C 2 )
Weyrauch (minimum values)

T = 44.17 (1 +C)
Salom (average values)

T = 31.74 + 70.53 C.

The above formulas express the tenacity in kilograms per square millimeter. Campbell, for acid
open hearth steel, gives

T = 40.000 + 1000 C + 1000 P + xMn + R
and for basic open hearth steel

T = 41.500 + 770 C + 1000 P + yMn + R

in which x and y are values given in a table and dependent upon the percentage of manganese and
of carbon present. R is a variable to allow heat treatment.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

pearlite, respectively 40 and 10 per cent in two inches, it should be possible to work
out a formula expressing the ductility of any hypo-eutectoid steel in the annealed
(pearlitic) condition. In terms of ferrite and pearlite the ductility should be

40 F + 10 P
100

or simplifying

D = .4 F + .1 P

or in terms of pearlite alone since F = 100 - P

D = .4 (100 - P) + .1 P = 40 - .3 P

and since P = 120 C, the ductility in terms of carbon will be

D = 40 - 36 C

Pearlitic steels, for instance, containing 0.25 and 0.50 per cent carbon should have
elongations respectively of 31 and 22 per cent. 1

SO tgs.ooo

it -w> /octooo
<b

< s
s *
N \

.^ jo **- 73000

* 4
1

g) tO ^ SQOOO

Ci *

/O 25.000

X Co^-Aon
- />eo/-///

Fig. 17. D

100

^ <=

; =^ =C; . ; .

''tv

\ x

/X N

"<.

-'/

.' 20 -*tj L> .ao /o /^ /. ^ /a /.a 2~o
J e 3 / IS /8 2f 4 27 JO
2* -*a 7 96 57 9J o. oo.a eos aj -* 80

iagram showing the relation between the tenacity and ductility of annealed (pearlitic
steels and the carbon content.

1 It is interesting to compare this formula with some others that have been proposed. Howe
gives for the elongation in 8 inches of steel under 0.50 per cent carbon

D = 33 - 60 (C 2 + 0.1)
and for steel between 0.50 and 1.00 per cent carbon:

D = 12 - 11.9 V^C- 0.5
Deshayes for the elongation in 8 inches, gives

D = 42 - 56 C
and for the elongation in 4 inches

D = 35 - 30 C

These formulas give lower values for the elongation of steel than the author's formula, but all indi-
cations point to the fact that they refer to steel in a rather sorbitic condition and, therefore, more
tenacious and less ductile, whereas the formula here suggested is for truly pearlitic steel only.

LESSON V MEDIUM HIGH AND HIGH CARBON STEEL 19

Diagram Showing the Relation Between the Tenacity and Ductility of Steel and
Its Carbon Content. By plotting the formulas suggested in this lesson to express
the relation between the carbon content of steel and its tenacity and ductility the
curves of Figure 17 are obtained. To the tenacity and ductility curves a third curve
has been added showing the variation of the amount of pearlite with the carbon
content.

Experiments

The student should procure five samples of forged steel of good commercial qual-
ity, containing respectively about 0.30, 0.50, 0.85, 1.25, and 1.50 per cent carbon.
These should be heated to 900 or 1000 deg. C. and slowly cooled from that tempera-
ture. Specimens should be cut from the treated samples and prepared for micro-
scopical examination in the usual way.

Etching. Any one of the three methods described in Lessons III and IV may
be applied with satisfactory results. The student is advised to use the method which,
in connection with the experiments of the two previous lessons, he has found to give
the best results.

Examinations. All samples should be carefully examined, firstly with low power
objectives and secondly with high power objectives. Observation under low power
should in all cases reveal clearly and sharply the pearlite and ferrite areas in hypo-
eutectoid steel, the pearlite and cementite areas in hyper-eutectoid steel. Examina-
tion under high power, 300 diameters or more, should satisfactorily reveal the
structure of the pearlite. In the case of the eutectoid steel, low power observation
will reveal but an indistinct structure, because of the absence of any free ferrite or
cementite and of the fineness of the pearlite structure.

Etching with Sodium Picrate. The hyper-eutectoid steels should be rubbed on
the rouge block or disk so as to efface the pattern produced by the etching and treated
with a boiling solution of sodium picrate in an excess of sodium hydroxide in order to
color darkly the free cementite. The instructions given in the lesson for this opera-
tion should be followed.

Photomicrography. All samples should be photographed both with low and
high power objectives. In taking low power photomicrographs the directions given
in Lesson IV should be followed.

For the taking of high power photographs a 4 mm. (% in.) objective and a 5X
(2 in.) or if needed a 10X (1 in.) eyepiece are recommended. With the camera screen
some 15 inches from the eyepiece a magnification of 325 diameters will be obtained
in case of the 5X eyepiece and of 650 diameters with the 10X eyepiece.

The needed manipulations for the taking of high power photomicrographs are the
same as to their nature as those required for taking low power photographs but they
call for much greater accuracy. The adjustment of the source of light and of all
parts used in condensing the light, including the vertical illuminator should be done
with the greatest possible care and delicacy while the focusing of the image on the
camera screen could hardly receive too much attention. The student is urged not to
be discouraged if his first attempts at taking high power photomicrographs are fail-
ures. Patience, perseverance, and experience will eventually lead to the mastery of
the manipulations required for successful high power photomicrography of metal
sections.

20 LESSON V MEDIUM HIGH AND HIGH CARBON STEEL

Examination

I. Describe the structure of hypo-eutectoid, of eutectoid, and of hyper-eutectoid
steel.

II. Assuming pearlite to contain 0.834 per cent carbon, what will be the proximate
structural composition of steels containing respectively 0.12, 0.27, 0.56, and
1.15 per cent carbon? What will be their ultimate structural composition?

III. Accepting as correct the formulas given in these lessons what will be the tensile

strength of the steels mentioned in Question II.

IV. Mention any difficulty encountered in conducting the experiments of this lesson.

LESSON VI

IMPURITIES IN STEEL

Metallic Impurities. Commercial grades of steel always contain, besides carbon,
varying amounts of silicon, phosphorus, sulphur, and manganese, often an appreciable
proportion of copper and traces at least of many other metals and metalloids. These
may be called the metallic impurities.

Non-metallic or Oxidized Impurities. Non-metallic or oxidized impurities,
chiefly oxides and silicates of iron and manganese, are also frequently found in steel,
principally through the retention by the metal of some of the slag produced during
the refining operation. Hibbard has recently suggested the name of "sonims" for
this class of impurities.

Metallic vs. Non-metallic Impurities. There is a sharp distinction between the
behavior of metallic and non-metallic impurities, the former forming true alloys with
the contaminated metal, the latter being merely inclusions, their union with the
metal being purely mechanical.

Gaseous Impurities. Steel always contains some gases, apparently held in solu-
tion and called "occluded" gases, chiefly hydrogen, nitrogen, and carbon monoxide
(CO).

Impurities vs. Physical Properties of Steel. It is well known that surprisingly
small proportions of some of the metallic impurities just mentioned have a very
marked influence upon the physical properties of steel. Some 0.2 per cent phosphorus,
for instance, renders many grades of steel so brittle as to unfit them for most com-
mercial uses. And as it is logical to suppose that there exists a very close relation
between the structure of a metal and its physical characteristics, we naturally ex-
pect to find important structural changes corresponding to marked alterations of
physical properties. We should expect, for instance, the structure of a high phos-
phorus, brittle steel to be quite different from the structure of a low phosphorus,
tough steel of otherwise identical composition. In the present state of metallography
the microscope does not always reveal such differences of structures as we are led to
look for. We may reasonably anticipate, however, that, as the science progresses,
structural differences will be detected of a magnitude fairly in keeping with the deep
changes of physical properties brought about by slight changes of chemical composi-
tion. Indeed in recent years material advance has been made in this direction and
the influence of the usual impurities upon the properties of steel has been on the
whole satisfactorily accounted for by metallographic methods as will be apparent
from the description which follows.

Silicon in Steel. All grades of steel contain a trace at least of silicon (Si) and
occasionally as much as 0.5 per cent, and even more, most grades containing between
0.05 and 0.3 per cent.

2 LESSON VI IMPURITIES IN STEEL

When present in such small proportion silicon is entirely dissolved in the iron
with which it forms a solid solution. It is probable, however, that it is not held in
solution by the iron in its elementary condition, Si, but rather as a silicide of iron,
FeSi. Since the atomic weight of iron is 56 and that of silicon 28 it will be evident
that 28 parts by weight or silicon produces 56 + 28 or 84 parts by weight of FeSi, or

/84 \

that silicon produces exactly 3 times its own weight of FeSi ( = 3 I. For instance

\28 /

0.1 per cent silicon in the steel will give rise to the formation of 0.3 per cent of FeSi
and this small amount of iron silicide will be held in solid solution by the iron. The
ferrite of commercial steel, therefore, always contains a small amount of silicon in
the form of an iron silicide, and let it be borne in mind that this applies to the ferrite
forming part of the pearlite of all slowly cooled steels as well as to the free ferrite of
hypo-eutectoid steel.

It has been stated in another lesson that when an impurity forms a solid solution
with the contaminated metal, changes of crystalline forms are not generally ob-
served. This is true in the present case for there is apparently no structural difference
between a steel with some 0.3 or 0.4 per cent silicon and a steel nearly free from that
element but otherwise of identical composition. The presence of silicon in steel can-
not as yet be satisfactorily detected, even qualitatively, by metallographic methods,
although we have the unquestionably accurate statement of Le Chatelier that silicon
causes ferrite to etch more slowly.

In view of the similarity of structure between steel containing much silicon (i.e.
several tenths of 1 per cent) and steel practically free from it, we should expect that
the presence of a small amount of silicon cannot affect materially the properties of
steel, and this we know to be the case.

Phosphorus in Steel. Steel of satisfactory quality contains from a trace to 0.1
per cent of phosphorus (P). As in the case of silicon this small amount of phosphorus
is held in solid solution by the iron, not, however, in the elementary state, P, but as
the phosphide of iron FesP. The atomic weight of iron being 56, that of phosphorus
31, and the phosphide containing three atoms of iron for each atom of phosphorus, it
will be obvious that 31 parts by weight of phosphorus will form 3 x 56 + 31 or 199
parts of the phosphide Fe 3 P, or roughly, 1 part by weight of phosphorus will give
rise to the formation of 6 parts of phosphide. For instance, the presence in steel of
0.05 per cent phosphorus results in the formation of 0.3 per cent of Fe 3 P held in solid
solution by the ferrite, this being true of the ferrite included in the pearlite of all
slowly cooled steel as well as of the free ferrite of hypo-eutectoid steel.

While phosphorus in common with other metallic impurities forming solid solu-
tions does not alter the crystalline form of steel, it is generally believed to have a
marked tendency to enlarge the grains of the metal, which tendency would account
for the well-known brittleness imparted to steel by phosphorus when present in ex-
cess of 0.1 per cent. The brittleness caused by a large grain will be considered fur-
ther in another lesson.

Except for this enlargement of the grains, microscopical examination does not re-
veal the presence of the usually small percentages of phosphorus occurring in steel,
although it is said by some writers that phosphorus as well as manganese causes ferrite
to etch darker.

Sulphur in Steel. Steel of satisfactory commercial quality may contain from a
mere trace to some 0.1 per cent sulphur, generally between 0.01 and 0.05 per cent.

LESSON VI IMPURITIES IN STEEL

It is universally known that manganese and sulphur have very great reciprocal af-
finity so that when brought together at a high temperature they combine chemically
with each other to form the sulphide of manganese, MnS. This is what happens in
steel which always contains manganese as well as sulphur. From the atomic weight
of manganese, 55, and that of sulphur, 32, it will be seen that 32 parts by weight of
sulphur produces 87 parts of MnS, or approximately 2^ parts of sulphide for each
part of sulphur. Steel with 0.05 per cent sulphur, for instance, will contain about
0.125 per cent of MnS, provided, of course, there is enough manganese present to
satisfy the sulphur which must necessarily be so in properly made steel.

The existence of the sulphide of manganese, MnS, in steel has been conclusively
proven. In steel castings it occurs as rounded areas the color ef- which is generally
described as pale or dove gray or slate color. In forgings it occurs in elongated
particles, bands, or strings of the same tint, running parallel to the direction of the
forging or rolling (Figs. 1 and 2). -^ ^^

Fig. 1. Steel. Carbon 0.46 per cent. Manganese Fig. 2. Steel. Hypo-eutectoid. Manganese
1.07 per cent. Sulphur 0.54 per cent. Forged, sulphide in ferrite areas. Magnified 300
reheated to 1200 deg. C. and cooled in air. diameters. (Levy.)
Magnified 460 diameters. Right half section is
longitudinal (direction of rolling), the left half is
transverse. (Arnold.)

According to Le Chatelier MnS has a melting-point superior even to that of iron,
solidifying, therefore, first, and the bulk of it rising to the top of the bath or ingot,
the manganese is in this way helpful in removing sulphur from steel. Some writers,
however, question this higher melting-point of MnS. Levy reports that the melting-
point of pure MnS is probably not far from 1400 deg. C., and, therefore, below the
melting-point of hypo-eutectoid steel at least, while the presence of some FeS would
lower materially its melting-point. It appears probable that the solidification of the
sulphide must follow and not precede that of the iron.

This view seems to be supported by the location of the sulphide particles at the
boundaries of the pearlite grains of eutectoid steel, in the free ferrite of hypo-eutec-
toid steel or in the free cementite of hyper-eutectoid steel. What MnS is retained by
the solid steel, since it occurs as shown in the shape of small individual grains or

4 LESSON VI IMPURITIES IN STEEL

elongated particles, can only injure the metal through breaking up its continuity
and, in view of the very small amount of sulphur, and, therefore, of MnS, present in
steel of good quality, it is evident that this breaking up and its action upon the prop-
erties must be very slight. This is in agreement with the known fact that a small
amount of sulphur in steel containing also the proper amount of manganese has no
appreciably injurious effect.

Seeing that steel seldom contains much more than some 0.05 per cent sulphur,
hence more than 0.125 per cent MnS, it is not to be expected that this compound
will generally be detected in polished and etched steel sections. Indeed whenever
detected it points to a segregation of the sulphide together with other impurities
(ghost lines) as described later.

In case sulphur occurs in excess over the amount needed to form the sulphide
MnS with the manganese present in the steel, the excess sulphur, that is the sulphur
left over after satisfying the manganese, combines with some of the iron, forming the
iron sulphide FeS. It should be noted at once, however, that it requires less than
2 parts by weight of manganese (atomic weight 55) to combine with 1 part of sulphur
(atomic weight 32). In other words if the steel contains twice as much manganese
as it does sulphur, this should theoretically be enough to convert the whole of the
sulphur into the sulphide MnS. As it is very seldom indeed that steel does not con-
tain a much larger proportion of manganese than that compared to its sulphur con-
tent, the occurrence of free FeS in steel should be very rare. It is not to be expected
in metal of good quality, its presence pointing to a very abnormal composition,
namely high sulphur content and very low percentage of manganese. According to
Levy, however, MnS and FeS are readily soluble in each other in the solid state, MnS
being capable of holding as much as 50 per cent of FeS in solid solution. According
to this writer MnS is seldom free from FeS even when the steel contains considerable
manganese, the mass action exerted by the presence of so large a proportion of iron
preventing the manganese from taking hold of the totality of the sulphur in spite of
its greater affinity for it. MnS nearly free from FeS appears quite dark, while its
color becomes lighter and more yellowish as the proportion of FeS increases. Levy
notes also that in high carbon steel the MnS areas are generally colored darker than
in low carbon steel, indicating greater freedom from FeS, apparently owing to the
fact that in high carbon steel the mass action exerted by iron is not so great since it
contains less iron.

The sulphide FeS exhibits a marked tendency to form continuous envelopes or
membranes surrounding each grain of pearlite (Fig. 3), and probably consisting of a
eutectic alloy of iron and iron sulphide (the composition of the eutectic is apparently :
FeS 85 per cent, Fe 15 per cent). These membranes being weak and brittle impart
weakness and brittleness to the steel. The well-known red-shortness caused by sul-
phur in the absence of a sufficient amount of manganese (to form MnS) is probably
due to the low melting-point (950 deg. C. according to some writers) of this iron-
iron sulphide eutectic. At a high temperature the melting of this eutectic destroys
the cohesion between the grains of the metal resulting in cracks being developed dur-
ing the process of forging or rolling, and in extreme cases in the metal actually break-
ing into several pieces. The presence of a large amount of FeS in some Bessemer steel
at the end of the blow, before the addition of manganese, is undoubtedly largely re-
sponsible for the marked red-shortness of the metal at this stage of the operation.

Under the microscope FeS appears yellow or pale brown. Tests showing the

LESSON VI IMPURITIES IN STEEL 5

presence of sulphur in the constituents described above as sulphides may be con-
ducted as follows, provided they occur in sufficiently large particles : A sheet of silver
bromide (photographic) paper should be pressed upon the polished section and
moistened with sulphuric acid when the sulphur present will be evolved as H 2 S (sul-
phuretted hydrogen) and will darken the paper. Another method (Law) consists in
covering the section with a coating of gelatine containing an acid solution of lead or
cadmium salt; the acid decomposes the sulphide forming H 2 S which produces a deep
brown or yellow stain of lead or cadmium sulphide, PbS or CdS.

Manganese in Steel. It has been seen that manganese combines readily with
sulphur and that the resulting manganese sulphide, MnS, can be detected in polished
steel sections as a pale or dove gray constituent assuming the shape of rounded areas
in castings and of bands or threads in forgings. Manganese silicate is also occasionally

Fig. 3. Red short steel. Magnified 300 diameters. Sulphur
0.54 per cent. Unetched. Network of FeS. (Ziegler.)

found in steel as later explained and may sometimes be mistaken for MnS. Satis-
factory tests for the distinction of these two constituents will be described.

When manganese occurs in excess over the amount required to form MnS with
the totality of the sulphur present, as is almost universally the case, the manganese
in excess combines with some of the carbon to form the carbide of manganese, Mn 3 C,
and this carbide is found associated with the iron carbide, Fe 3 C, in cementite. The
comentite of commercial steel, therefore, is seldom a pure iron carbide, containing,
on the contrary varying amounts of Mn 3 C. Since iron and manganese have practi-
cally the same atomic weight, however (55 and 56 respectively), it remains practically
true that carbon forms 15 times its own weight of cementite, even when the latter
contains a large proportion of MnsC.

There is no metallographic test by which cementite free from manganese can be
distinguished from cementite rich in MnaC.

Some authors mention the possible presence of the manganese silicide, MnSi, in
steel, while solid solution between manganese and iron is frequently referred to.
While manganese and iron (ferrite) undoubtedly form solid solutions, it does not
seem likely that these are produced when manganese is present in small proportion,

6 LESSON VI IMPURITIES IN STEEL

say not over 1 per cent. In that case it seems more probable that manganese is found
in the two forms described above, (1) as a manganese sulphide MnS, containing prac-
tically the totality of the sulphur in steel of good quality, that is, containing not over
0.05 per cent sulphur and not less than 0.25 per cent manganese, and (2) as the man-
ganese carbide Mn 3 C, associated with Fe 3 C in cementite.

Chemical vs. Structural Composition. Knowing the probable chemical forms of
the five metallic impurities always present in steel, carbon, silicon, phosphorus, sul-
phur, and manganese, as well as their structural associations, it will be interesting
and profitable to consider accordingly the proximate chemical composition as well as
the ultimate and proximate structural compositions of a steel of known ultimate
chemical composition. Let us assume a steel of the following ultimate chemical
composition:

C 0.50 per cent

Mn 0.80 " "

S 0.05 " "

P 0.04 " "

Si 0.10 " "
Fe (by diff.) 98.51 " "

100.00

Bearing in mind the atomic weights of these elements (Fe, 56; C, 12; Mn, 55; S, 32;
P, 31; Si, 28) and the formulas of the chemical compounds formed (MnS, FeSi, Fe 3 P,
Mn 3 C, Fe 3 C), it will be readily seen that:

(1) 0.05 per cent S will give rise to the formation of 0.13 per cent MnS.

(2) 0.13 per cent MnS contains about 0.08 per cent Mn.

(3) This leaves 0.80 - 0.08 = 0.72 per cent manganese in excess to combine with C.

(4) 0.72 per cent Mn will form 0.77 per cent MnsC.

(5) 0.77 per cent Mn 3 C contains about 0.05 per cent carbon.

(6) This leaves 0.50 0.05 = 0.45 carbon to combine with iron.

(7) 0.45 per cent carbon results in the formation of 6.75 per cent of Fe 3 C.

(8) 0.04 per cent of P corresponds to about 0.25 per cent of Fe 3 P.

(9) 0.10 per cent Si gives 0.30 per cent FeSi.

The proximate chemical composition of the steel considered will be

Fe 3 C

6.75 per

cent

Mn 3 C

0.77 "

Fe 3 P

0.25 "

FeSi

0.30 "

MnS

0.13 "

Fe (by diff.)

91.80 "

100.00

As to the ultimate structural composition of the steel, we know that the cemen-
tite contains the Fe 3 C and the Mn 3 C hence we have 6.75 + 0.77 = 7.52 per cent
cementite. In pure steel the percentage of cementite would have been 0.50 X 15 =
7.50 per cent. The slight difference between the two numbers is due to the presence
of manganese in the commercial steel, and to a slight difference between the atomic

LESSON VI IMPURITIES IN STEEL 7

weights of manganese and that of iron (55 compared to 56), a difference so slight
that for all practical purposes we may assume that in commercial steels as well as in
pure steel the percentage of carbon multiplied by 15 gives the amount of cementite
formed. The total ferrite present in this steel contains all the free iron, as well as the
small proportions of Fe 3 P and FeSi present, hence this steel contains 91.80 + 0.25 +
0.30 = 92.35 total ferrite. In pure steel the proportion of total ferrite would have
been 100 .7.50 or 92.50 per cent. The difference between the two values is evi-
dently due to the presence of a trifle greater amount of cementite, and to the pres-
ence of 0.13 per cent MnS.

The ultimate structural composition of the steel under consideration is, therefore:

Cementite 7.52

Total ferrite 92.35

MnS 0.13

100.00

Finally its proximate structural composition will be, since the pearlite of hypo-
eutectoid steel contains 8 times the weight of total cementite (assuming the eutec-
toid carbon point to be 0.834 per cent) :

Pearlite 7.52 x 8 = 60.16

Free ferrite (by diff.) 39.71

MnS .13

100.00

Ignoring the presence of impurities the quick method described in Lesson V
would have given pearlite 60 per cent, ferrite 40 per cent, i.e. values which may be
considered identical for any practical purposes. It follows from this that in calcu-
lating the structural composition of any carbon steel of ordinary commercial quality
the presence of the impurities need not be considered; the steel may be treated as if
it was made exclusively of iron and carbon.

The relation between chemical and structural compositions, both ultimate and
proximate, is further shown in the following table.

CHEMICAL COMPOSITION

STRUCTURAL COMPOSITION

ULTIMATE

PROXIMATE

ULTIMATE PROXIMATE

Fe(by diff
Si

) 98.51
0.10

Fe(by diff.) 91.80 '
FeSi 0.30

[ Free Ferrite 39.71 Free
Total Ferrite 92.35 < Ferrite

P
C
Mn

0.04
0.50
0.80
0.05

Fe s P 0.25
Fe 3 C 6.75 '
MnsC 0.77 j
MnS 0.13

( Pearlite Ferrite 52.64 j 60-16% p ear .
Cementite 7 52 j nte

MnS 0.13 0.13

100.00

100.00 100.00

LESSON VI IMPURITIES IN STEEL

Non-Metallic or Oxidized Impurities. As already mentioned steel not infre-
quently contains small amounts of non-metallic or oxidized impurities, chiefly iron
and manganese oxides and silicates, derived mainly from the retention by the metal

Fig. 4. Manganese sulphide (light constituent) and manganese silicate
in steel. Magnified 1000 diameters. (Law.)

Fig. 5. Manganese sulphide (light constituent) and iron silicate in
mild steel. Unetched. Magnified 1000 diameters. (Law.)

in the shape of minute particles of some of the slag formed during the process of
manufacture. Their mode of occurrence is very different from that of the metallic
impurities just examined (with the exception of MnS which behaves more like a non-
metallic than like a metallic impurity). These oxidized impurities do not alloy with

LESSON VI IMPURITIES IN STEEL 9

the metal; their association with it remains a purely mechanical one, like small peb-
bles in a mass of clay. These oxides and silicates commonly occur as rounded or
slightly elongated particles and can generally be detected in the polished section
before etching.

Manganese silicate, probably 2MnO.3SiC>2, and manganese sulphide, MnS, which
occasionally occur together, have a somewhat similar appearance, care being re-
quired in order to differentiate between them, although the former is as a rule de-
cidedly darker (Fig. 4). Stead recommends the placing of a drop of sulphuric acid on
the polished specimen, when H 2 S gas will be evolved where MnS is present, particles of
silicates of manganese, on the contrary, evolving no gas. The dissolving of MnS
also leaves pits. Stead also advises heat tinting as the best means of distinguishing
between the sulphide and the silicate, the heating to be continued until the specimen
has assumed a light brown coloration, when the MnS remaining bright can be sharply
differentiated from the silicate.

According to Levy, sulphide and silicate of manganese are readily soluble when

^H ' ' :

Fig. 6. Ghost lines in low carbon steel. Magnified
95 diameters. (Boylston.)

molten but on solidifying the sulphide crystallizes in well-marked dendritic forms,
the resulting mixture of sulphide and silicate (Fig. 4) resembling slag inclusions (see
Lesson III, Fig. 5).

Manganese sulphide and iron silicate may also occur in close vicinity, the latter
constituent being darker and frequently broken in many irregular fragments by the
working of the metal (see Fig. 5).

At the end of the refining operation by which steel is produced, especially towards
the latter part of the Bessemer blow, a considerable amount of iron oxide is formed
and in spite of the steps taken for removing it from the bath (addition of manganese,
etc.) some of it, occasionally quite a little, is retained by the metal, when it is a source
of red-shortness besides having other detrimental effects. This iron oxide generally
occurs as small dark points readily detected in the polished section before etching.

For further treatment of polished sections with a view of identifying oxidized im-
purities, the student is referred to Lesson III where the constitution of slag in iron
has been treated at some length.

Gaseous Impurities. It has not been possible so far to detect the presence of
occluded gases in steel by means of metallographic methods. While the problem

LESSON VI IMPURITIES IN STEEL

seems a very difficult one to solve, the statement that it can never be solved would
not be justified for the discovery of some metallographic treatment by which a metal
rich in certain gases may be distinguished from a similar metal free from them is well
within the limits of reasonable expectation.

Fig. 7. Ghost lines in low carbon steel. Magnified 10 diameters. (Law.)

Fig. 8. Ghost line.s in low carbon steel. Magnified 200 diameters. (Law.)

Segregation of Impurities. Ghosts. The very small proportions of impurities
generally found in steel of good quality have little, if any, injurious effect upon its
most important and useful physical properties, so long as they remain uniformly
distributed throughout the metallic mass, i.e. so long as the steel is chemically homo-
geneous. These impurities, on the contrary, may become extremely injurious when

LESSON VI IMPURITIES IN STEEL 11

they show a tendency to "segregate," i.e. to collect in certain portion or portions of
steel castings and forgings, when the segregated portions may contain so large an
amount of impurities as to have their useful properties utterly destroyed. Segre-
gated metal is generally brittle, weak, and hard.

Under the microscope a metal suffering from this segregation of impurities gen-
erally is found to contain bands of varying widths and lengths, technically known as
"ghosts" or "ghost lines," in which the presence of abnormally large proportions of
MnS, phosphorus, and carbon can generally be detected by the ordinary metallo-
graphic tests, these (S, P, and C) being the three impurities .showing the greatest ten-
dency to segregate. Photomicrographs of ghost lines are shown in Figures 6, 7, 8, and 9.
Ghost lines etch more rapidly than the surrounding metal therefore appearing darker

Fig. 9. Ghost lines in low carbon steel. Magnified 2000 diameters.
Manganese sulphid and pearlite particles. (Law.)

after etching even to the naked eye. These lines can generally be detected before
etching because of the manganese sulphide which they contain.

Bannister mentions two kinds of ghost lines, (1) those showing marked segrega-
gation of C, S, and P, and considerable Si and Mn and (2) those containing little Si
and Mn. Houghton, on the other hand, refers to ghost lines containing S and P but
no carbon.

Experiments

High vs. Low Phosphorus Steel. The student should procure two samples of
forged steel containing preferably from 0.30 to 0.50 per cent carbon and of nearly
identical composition, except as to phosphorus content which should be high in one
sample (if possible considerably more than 0.1 per cent), and low in the other (not
over 0.05 per cent). These samples should be heat treated in the usual way so that
they may assume their normal structure and a specimen prepared from each sample
for microscopical examination etching them with picric or nitric acid in alcohol or
with concentrated nitric acid according to individual preference. Upon being ex-

12 LESSON VI IMPURITIES IN STEEL

amined under the microscope the larger grain of the high phosphorus steel should be
apparent.

Photograph each sample, using such magnification as will best bring out the fea-
ture to be illustrated, namely difference in grain size.

High Sulphur Steel. A sample of steel casting and a sample of steel forging
both containing if possible considerably more than 0.10 per cent sulphur should be
obtained. These need not be heat treated but may at once be polished and exam-
ined, first before etching and then after etching, for the detection of sulphide areas,
as described in the lesson. If the sulphide areas are of sufficient size the chemical
test described should be applied for the detection of sulphur.

The samples should be photographed with a view of bringing out sharply the
sulphide flaws.

Oxidized Bessemer Metal. A sample of Bessemer metal preferably high in sul-
phur should be procured if possible, taken at the end of the blow and before the ad-
dition of manganese or any other recarburizer. A specimen of suitable size should
be polished and examined under the microscope, both before and after etching. The
metal should contain both sulphide of iron, FeS, and iron oxide the appearance of
which has been described in this lesson. The specimen should be photographed so
as to show these two impurities.

Segregated Steel. If a sample of segregated steel can be obtained it should be
prepared, examined, and photographed in order to reveal the presence of ghost lines.

Examination

I. Describe briefly the appearance under the microscope of the following impuri-
ties: Si, S, Mn, P, iron oxide, and manganese silicate.

II. Explain the meaning of "ghost" lines.
III. A steel has the following ultimate, chemical compostion:

C 0.60 per cent
Mn 0.75 " "
S 0.04 " "
P 0.06 " "
Si 0.15 " "

Fe(bydiff.) 98.40 " "

100.00

What will be (a) its proximate chemical composition, (6) its ultimate structural
composition, and (c) its proximate structural composition?

LESSON VII

THE THERMAL CRITICAL POINTS OF STEEL
THEIR OCCURRENCE

The structure of steel described in the preceding lessons, i.e. its normal structure,
is greatly affected by the treatment or treatments, both mechanical and thermal, to
which the metal may be subjected during the process of manufacture of finished
objects. It is to the close relation existing between the treatment and the structure
on the one hand, and between the structure and the physical properties of the metal
on the other, that metallography owes its industrial importance. It is essential,
therefore, that the student should have a clear understanding of these relations. As
a preparation to this important study, however, it will be necessary to describe a
phenomenon of the greatest moment in the treatment of steel, namely, the occurrence
of spontaneous absorptions or evolutions of heat during the heating or cooling of the
metal. These are generally termed the "thermal" critical points or simply "critical
points," also "retardations," "transformation" points, and "critical temperatures."

Point of Recalescence. If a piece of steel containing some 0.60 per cent carbon
be heated to a high temperature, say to 1000 deg. C., and allowed to cool slowly
from that temperature, and if its rate of cooling be carefully ascertained, conveniently
by means of a Le Chatelier pyrometer, it is found that the cooling proceeds at first at
a nearly uniformly retarded rate. If, for instance, it requires 10 seconds for the metal
to cool through the first five degrees (from 1000 to 995 deg.), and 12 seconds to cool
through the next five degrees (995 to 990 deg.), it will require some 14 seconds for
the next five degrees, 16 seconds for the following five, and so on, the cooling through
each range of five degrees being a little slower than the preceding cooling of five
degrees. All cooling bodies, whatever their nature, generally follow this law. The
plotting of time and temperature as coordinates yields smooth curves, sometimes
approaching straight lines (see curve B, Fig. 9).

In the case of the steel we are now considering, when a certain temperature is
reached, in the majority of cases some 650 to 700 deg. C., a most interesting and
significant phenomenon takes place; the cooling of the metal is momentarily arrested,
the pyrometer, for a certain length of time, failing to record any further fall of tem-
perature. Indeed, when the circumstances are favorable, the temperature of the
cooling mass actually rises; the metal becomes visibly hotter; it "recalesces," hence
the name of " recalcscence " given to this thermal critical point. If the experiment
be conducted in a dark room, this recalescence or spontaneous glow of the steel is
plainly visible. After a while the metal resumes its normal rate of cooling which is
then continued down to atmospheric temperature.

It is evident that at this critical point the surrounding atmosphere does not cease
to abstract heat from the piece of steel, and, since its temperature nevertheless

2 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

remains stationary or even rises, it must be that heat is here spontaneously gene-
rated within the metal in amount sufficient to make up, or more than make up, for
the heat lost by radiation and conductivity.

In heating, as might be expected, the reverse phenomenon takes place : an absorp-
tion of heat causing a retardation in the rise of the temperature, or even a momentary
stop, the pyrometer failing for a few moments to record any further increase of
temperature or recording only an abnormally low increase, although heat continues
to be applied to the steel at the same speed. Actual lowering of the temperature of
the steel is not generally observed at this critical point on heating, i.e. the steel does
not grow perceptibly colder.

Notation. Osmond, who was the first to determine accurately the position and
magnitude of the point of recalescence and who is the discoverer of the upper critical
points soon to be described, adopted Tschernoff's previous notations, and designated
the critical points by the letter A. 1 To distinguish critical points on cooling from
those occurring on heating the former are called Ar (from the French refroidissement,
meaning cooling) and the latter Ac (from the French chauffage, heating). To dis-
tinguish further between the point of recalescence and its reversal on heating on the
one hand, and critical points occurring at higher temperatures on the other, the nota-
tions Ari and Aci are used for the recalescence point and its reversal, and Ar 2 , Ac 2 ,
Ar 3 , Acs for the two upper reversible critical points soon to be described. The nota-
tions AI, A 2 , A 3 are frequently used when the points and their reversals are consid-
ered collectively. By the notation AI, for instance, is meant the point of recalescence
Ari and its reversal Aci. These notations will be used in these lessons.

Brinell, in his important work on the heat treatment of steel, used the letter V for
the point of recalescence and W for its reversal on heating. These symbols, however,
are now very seldom used.

The expression "point of recalescence" is frequently used indifferently for the
point on cooling, where heat is evolved causing a recalescence of the metal, and for
the reverse phenomenon on heating, at Aci, where, of course instead of a recalescence
taking place, an absorption of heat occurs causing the metal to lose heat. It is ob-
vious that the term recalescence should not be applied to the point Aci. The point
of recalescence is also called sometimes recalescent point and, seldom, Gore's phe-
nomenon (see Historical Sketch at end of lesson). The point Aci has been called
point of " decalescence " by some writers and one of them at least refers to it as
the "calescence" point.

Critical Range. Transformation Range. When the various critical points oc-
curring in steel are considered collectively the range of temperature they cover is
frequently called the critical range, or, more seldom, but very appropriately, the
transformation range. It will soon be shown that the critical range may include one,
two, or three critical points. The meaning of the expressions "critical range on heat-
ing" and "critical range on cooling" is obvious.

Positions of Ari and Aci. The critical points Ar t and Aci do not occur at exactly
the same temperature, Aci being generally situated some 25 to 50 deg. higher
than Ari. When the point Ari, for instance, is found at 690 deg. C., the point API
will generally occur somewhere between 715 and 740 deg. It does not follow, how-
ever, that these two points are not the opposite phases of the same phenomenon.

1 The point A of Tschernoff indicated the temperature at which steel suddenly acquires harden-
ing properties on heating or loses them on cooling.

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL 3

The fact that the critical point on cooling lags behind the point on heating and vice
versa, is evidently a case of hysteresis so often observed in physical phenomena and
which implies a resistance of certain bodies to undergo a certain transformation, when
theoretically the transformation is due, the delayed transformation finally taking
place with added violence. This was vividly depicted by Howe some twenty years
ago in the case of iron. He wrote: "Just as we can cool water below its freezing-
point without completely freezing it, thereby rapidly increasing the strength with
which the water tends to freeze, so by a relatively rapid cooling we can carry the
metal considerably below Ari without giving the An change time to proceed far,
strengthening the while the tendency toward this change, which keeps kindling more

gap due fo
Hysferes/s

Fig. 1. Diagram showing reversible critical point.

and more till it bursts into a blaze, with such evolution of heat as actually to reca-
lesce, to raise the temperature of the metal by some 10 deg., in spite of the continued
abstraction of heat by the continued cooling of the furnace."

The slower the heating and cooling the nearer will the two points approach each
other, so that with infinitely slow cooling and heating they would undoubtedly occur
at exactly the same temperature. If there remained any doubt as to the points Aci
and Ari representing the opposite phases of the same phenomenon, i.e. of A being
a reversible point, it would suffice to dispel it to consider the fact that in order to
induce the retardation Ar t the steel must first be heated past the point Aci; and re-
ciprocally the retardation Aci cannot take place unless the metal has first been cooled
to a point below Ari. To illustrate : the melting of ice and the freezing of water are
undoubtedly the opposite phases of the same phenomenon, each one undoes the
work of the other, and in order to freeze the water we must first melt the ice and
likewise to melt the ice the water must first be frozen; one change cannot be induced
unless the opposite one has last taken place. Indeed it is possible through very slow

4 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

and undisturbed cooling to lower the temperature of water below its freezing-point
before it starts freezing, a clear instance of hysteresis, although in this case called
"surfusion," and when freezing takes place the temperature of the water rises to its
normal freezing-point, a clear case of recalescence although deprived of glow.

The diagram shown in Figure 1 illustrates further this reversibility of the point
AI. Let two parallel lines represent the phases Ac and Ar of the critical point. Let
condition A represent the state of the metal stable above Ac and condition B the
state of the metal stable below Ar. The gap between Ac and Ar is due to hysteresis.
MN is the temperature at which both the Ac point and the Ar point would occur if
there was no hysteresis as, for instance, if the metal could be heated and cooled in-
finitely slowly. Assuming the metal to be in condition B at a, below Ar, on heating
it from a to b above Ac on reaching the Ac point at x it passes from the condition B
to the condition A with absorption of heat causing a retardation in the heating; on
cooling from b to c, that is to a temperature below Ac but above Ar, the condition A
is retained so that upon heating from c to d no transformation can take place at x'
on passing through the Ac point and therefore no critical point observed. If the
metal be cooled from d to e, however, on passing through Ar at y it changes from
condition A to condition B with evolution of heat, causing a retardation in the rate
of cooling; if it now be heated again from e to / above Ac, a critical point will be ob-
served at x" since the metal now in condition B will pass to condition A.

It will be evident that between Ar and Ac the metal may be in condition A or
condition B depending upon whether it was last cooled from above Ac or heated
from below Ar.

Speed of Cooling and Heating vs. Position of AI. It has been seen that the
faster the cooling the lower is the position of the point Ari and the faster the heating
the higher the point Aci, that is, the faster the cooling and heating the greater the
gap between the opposite phases Ari and Aci of the reversible point AI.

The cooling of a piece of steel may be so rapid, as in quenching, as to prevent
altogether the retardation Ari from taking place, because a low temperature is so
quickly reached that the rigidity of the metal prevents the transformation of which
An is a manifestation. In other words time and a certain amount of plasticity are
required for the transformation Ari to occur, and in quenching, time is denied when
the metal is sufficiently plastic (i.e. at a red heat), while when time is given (i.e. after
quenching) the metal has lost its plasticity. It remains untransformed or but par-
tially transformed. It will be shown in another lesson that this suppression of the
point Ari is probably the cause of the hardening of carbon steel by sudden cooling. 1

Le Chatelier rightly reminds us that the speed of the transformations occurring
at the critical points of steel follows the general laws which govern the speed of all
chemical phenomena. In other words that the speed of the transformation is the
greater (1) the higher the absolute temperature and (2) the wider the range between
the actual temperature and the temperature of equilibrium, that is the temperature
at which the transformation is due. Above the critical temperature both influences
act in the same direction and the speed of transformation increases without limit.
Below the critical temperature these influences act in opposite directions necessarily
giving rise to the existence of a maximum speed. According to Le Chatelier this
notion of variable speeds of transformation accounts for all the peculiarities of the

1 It will be explained later that some writers have doubted the suppression of the transforma-
tions on rapid cooling and have suggested another explanation of the burdening of steel.

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL 5

hardening treatment. On heating it is hardly possible to raise the temperature of
transformation more than 100 deg. C. through very rapid heating, while during cool-
ing the speed reaches its maximum at about 600 deg. C., is very feeble below 200, and
nearly null at atmospheric temperature.

Chemical Composition vs. Position of AI. Generally speaking impurities have a
tendency to lower the position of Aci and Ari, some of them decidedly. Osmond, for
instance, indicates the position of An in a steel containing 1 per cent of Mn as 685
deg. whereas with 4 per cent of manganese the same point was lowered to 590 deg.
It is conceivable that further increase of that element must lower still more the crit-
ical point, so that finally it may be lowered below atmospheric^ temperature, being
apparently eliminated. It will be shown in another lesson that this is precisely what
occurs in the cases of manganese steel and high nickel steel, containing respectively
some 13 per cent of manganese or some 25 per cent of nickel. These steels exhibit no
retardation on cooling from a high temperature to atmospheric temperature. When
cooled to lower temperatures, however, by immersing them in freezing mixtures, or,
if need be, in liquid air, the retardations may again occur. In the case of commercial
steel of good quality the proportion of impurities, with the possible exception of man-
ganese, varies within relatively very narrow limits, so that no great variation should
be expected in the position of the critical point AI. Neither is it clear that the amount
of carbon present in steel has a marked effect upon the position of the point AI, al-
though some writers state that the point is lifted as the carbon increases. The point
An almost invariably occurs somewhere between 650 and 700 deg. C. and its reversal
Aci 25 to 50 deg. higher. Both Ari and Aci would probably occur at about 710 deg.
could the cooling and heating be infinitely slow.

Upper Critical Points. The existence of upper critical points, that is of thermal
retardations occurring at temperatures higher than that of the recalescence point, has
already been alluded to. These points were discovered by Osmond and their dis-
covery ushered in a new epoch in the scientific study of iron and steel. To describe
these points it is advisable to consider first the thermal retardations occurring in
cooling and heating carbonless iron and then similar retardations exhibited by steel
containing increasing amounts of carbon.

Thermal Critical Points in Pure Iron. On cooling from a high temperature, say
1000 deg. C., a piece of the purest iron obtainable and ascertaining its rate of cooling
as previously explained, the metal is found to cool normally, i.e. at a uniformly re-
tarded rate, until a temperature of some 900 to 850 deg. C. is reached when a marked
retardation is observed in the rate of cooling, indicating a spontaneous evolution of
heat, in this case, however, insufficient to cause an actual rise of temperature, i.e. a
recalescence of the metal. The cooling then resumes, or nearly resumes, a normal
rate of cooling, until at about 750 deg. C. a second evolution of heat takes place
causing another retardation in the rate of cooling, not so marked, however, nor so
sharply defined as the first one. The metal then cools normally or quite so to atmos-
pheric temperature. We have thus detected two unmistakeable spontaneous evolu-
tions of heat in the cooling of pure iron. The corresponding critical points are called
Ar 3 and Ar 2 , the latter symbol indicating the lower point. It should be noted that
the recalescence point which should occur at some 675 deg. is here absent. Carbon-
less iron has no point of recalescence.

These two upper points like the point of recalescence are reversible critical points,
i.e. on heating the opposite phases of the transformations (whatever those trans-

6 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

formations may be) take place with absorption of heat, causing a retardation in the
rate of heating and the corresponding points being designated by the symbols Ac 3
and Ac 2 . The point Ac 3 occurs at a temperature some 20 to 30 deg. higher than its
reversal Ar 3 , while Ac 2 occurs at nearly the same temperature as Ar 2 .

Thermal Critical Points in Very Low Carbon Steel. Let us now take a sample
of steel containing some 0.10 per cent carbon, and let us ascertain its rate of cooling
from a high temperature precisely as before. Three thermal retardations will be de-
tected, Ar 3 at about 850 deg., Ar 2 near 750 deg., and Ari (point of recalescence) near
675 deg. Of these three spontaneous evolutions of heat the upper one at A.TS will be
the most marked, while at Ar 2 and at Ari they will be quite faint, their satisfactory
detection calling for the use of delicate instruments and careful manipulations. On
heating corresponding retardations will occur, due to spontaneous absorptions of
heat, the resulting critical points being designated as Ac 3 , Ac 2 , and Aci. Of these Ac 3
and Aci will occur at temperatures some 25 deg. or more higher than Ar 3 and An,
while Ac 2 will occupy nearly the same position as Ar 2 on the temperature scale, that
is about 750 deg.

Peculiarities of the Point A 2 . The point A 2 is generally less marked than the
points A 3 and AI. Unlike A 3 its position is little affected by the carbon content, and
unlike A 3 and AI the point on heating, Ac 2 , occurs at nearly the same temperature as
the point on cooling, Ar 2 . To these peculiarities must be added another one, namely,
the fact that A 2 appears to cover a wide range of temperature. While its intensity
decreases with fall of temperature its lower limit probably extends to considerably
below 700 deg. In other words the transformation of which A 2 is a manifestation is
not completed by the time the point AI is reached. Indeed Osmond mentions 550
deg. C. as the probable lower limit of the point A 2 . Some explanations of these pecu-
liarities of the point A 2 will soon be offered.

Thermal Critical Points of Medium High Carbon Steel. The determination of
the rate of cooling of a steel containing some 0.45 per cent carbon reveals the exis-
tence of two critical points, one, evidently the point of recalescence, Ari, at the usual
temperature (650 to 700 deg.) and one upper point in the vicinity of 725 deg. Does
the presence in this steel of only one upper point mean that one of the two upper
points detected in carbonless iron and in very low carbon 'steel has disappeared,
because of the presence of more carbon, or does it mean that the two upper points
have now united into a single one? The latter view is generally assumed to be the
correct one and this single upper point of medium high carbon steel is designated
accordingly by Ar 3 . 2 . This notation clearly implies that the two distinct evolutions
of heat which in carbonless iron and in very soft steel occur separately at Ar 3 and
Ar 2 here occur at one and the same temperature. Increasing the carbon content
decreases the interval of temperature between the two upper points until, finally, for
a certain carbon content the points meet to form the double point Ar 3 . 2 .

Merging of A 3 and A 2 . It has been seen that as the carbon increases the point
A 3 is gradually lowered until finally it merges with A 2 , whose position is not greatly
affected by the presence of carbon, to form the point A 3 . 2 . It would be interesting to
know the exact proportion of carbon required to cause this merging. This, however,
is difficult to ascertain because of the experimental difficulty of separating two crit-
ical points situated very near each other as they must be in the vicinity of the mer-
ging point, and also because this merging will be shifted somewhat by speed of heating
and cooling and by slight changes of chemical composition. From the mass of experi-

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL 7

mental evidences which has been published it seems probable that the merging takes
place at about 0.30 per cent carbon.

Thermal Critical Point in Eutectoid Steel. Eutectoid steel, that is steel con-
taining some 0.85 per cent of carbon, exhibits but one critical point, the point of re-
calescence, very marked at about 675 deg. C. on cooling. Shall it be inferred from
the occurrence of this single point that in eutectoid steel the transformations of which
the upper points A 3 and A 2 or the double point A 3 .2 are manifestations do not take
place? Or shall it be assumed that these transformations now take place at the same
temperature as the transformation corresponding to the critical point A[? In other
words that increasing the amount of carbon has so depressed the position of the two
upper points as to cause them to unite with the lower point, "forming now a triple
point to be designated as Ar 3 . 2 .i? This is the view generally held. The critical point
on heating is designated by the notation Ac 3 . 2 .i. It will be explained in another
lesson why the points A 3 , A 2 , or A 3 . 2 cannot exist in eutectoid or hyper-eutectoid steel,
when it will also be shown that the single point of eutectoid steel is not in fact a
merging of A 3 , A 2 , and AI, but merely the point Ai, the upper points having disap-
peared.

Merging of A 3 . 2 and AI. As the carbon content of the steel increases still more
after the merging of A 3 and A 2 has been effected, the interval between the points
A 3 . 2 and AI gradually diminishes until these two points, in turn, appear to merge to
form the triple point A 3 . 2 .i- Theoretically this apparent merging should occur when
the steel is composed entirely of pearlite, that is, when it contains in the vicinity of
0.85 per cent carbon, for reasons that will later be made clear. As a matter of fact,
however, the merging seems to take place long before so large a proportion of carbon
is present, for the point A 3 . 2 is seldom detected in steel containing more than some
0.50 or 0.60 per cent of carbon; this is probably due, as already explained, to the
difficulty of separating, experimentally, two critical points so close to each other.

Thermal Critical Points in Hyper-Eutectoid Steel. Carefully conducted obser-
vations reveal the existence of an upper critical point in hyper-eutectoid steel, at
least in steel containing a decided amount of free cementite, and, of course, of the
point of recalescence. It seems proper to designate this upper point by the symbol
A cm (Ar cm on cooling, Ac cm on heating) for reasons later to be given, cm stand-
ing for cementite. At least one writer, however, has designated this point on cooling
by the notation Ar mc , me standing for massive cementite. Other writers have called
it an A 3 point, a notation from which one would naturally infer that this upper point
of hyper-eutectoid steel is similar to the upper point of iron and of very low
carbon steel, which is not the case.

Purely theoretical considerations lead us to infer that the position of the point
Acm is lowered as the proportion of carbon decreases, finally merging with the point
A 3 .2.i at the eutectoid point. It would follow from this that the single point of eu-
tectoid steel is really a merging of four points A 3 , A 2 , AI, and Ac m and that it should
accordingly be designated by A 3 . 2 .i cm . It is, however, the universal custom to ignore
this contribution of Ac m to the single point of eutectoid steel and to use for the latter
the notation A 3 . 2 .i.

The amount of heat evolved at Ar cm is very slight, hence the difficulty of detect-
ing this point. Carpenter and Keeling ascertained its existence in steels containing
respectively 1.31, 1.51, 1.69, 1.85, and 1.97 per cent carbon at the following corre-
sponding temperatures: 883, 911, 985, 1030, and 1042 deg. C. With lower carbon

8 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

contents, that is nearer the eutectoid composition, the heat evolved is so slight that
the detection of Ar cm as a separate point is quite impossible. In theoretical dia-
grams, however, the existence of this point is always indicated in all hyper-eutectoid
steels with a sharp merging with A 3 . 2 .i at the eutectoid point.

The point Ar cm then should occur in all hyper-eutectoid steels at temperatures
increasing from some 700 to 1050 deg. C. as the carbon increases from 0.85 to 2.00
per cent.

Merging of A 3 . 2 .i and A cm . As already explained theoretically the merging of
the points A 3 . 2 .i and A cm should take place at the eutectoid composition, that is, for
steel containing in the vicinity of 0.85 per cent carbon. Experimentally, however,
the point A cm cannot be detected in steel containing less than some 1.20 per cent
carbon. Bearing in mind that hypo-eutectoid steels containing more than 0.60 per
cent carbon or thereabout have likewise but one critical point so far as experimental
evidences are concerned, it will be seen that for all practical purposes we may con-
sider all grades of steel containing from 0.60 to 1.20 per cent carbon as having but
one critical point, namely, the point of recalescence, at some 675 deg. C. on cooling,
although theoretically eutectoid steel only should have but one such point.

Minor Critical Points. Some experimenters believe to have discovered some critical points
other than those so far described. These points, which may be referred to as minor critical points,
correspond to very faint evolutions or absorptions of heat, and produce, therefore, but very slight
jogs in the thermal curves. Their existence is not fully established and they appear to have but
little if any influence upon the practical side of our subject. They should, however, be mentioned in
these lessons so that the student may at least have some idea of their nature and claims to recogni-
tion. Roberts-Austen in 1898 detected a slight evolution of heat between 550 and 600 on cooling in
iron and hypo-eutectoid steel, and this point was again detected by Carpenter and Keeling in 1904.
The latter observers named it the Aro point, following in this Roberts-Austen.

Roberts-Austen detected another evolution of heat in pure iron between 450 and 500 deg. C.
the existence of which he ascribed to the presence of hydrogen resulting in a separation of hydroxide
of iron taking place at this critical point. Finally the same observer described one more slight evo-
lution of heat in pure iron at about 270 deg. C. which he tentatively ascribed to the formation of an
iron-iron hydroxide eutectic.

Arnold believes in the existence of a critical point between A 3 and A 2 , of maximum intensity
when the steel contains some 50 per cent of pearlite (about 0.45 per cent carbon) which he thinks is
due to the formation or segregation of pearlite and hardenite, a constituent later to be described.

Data Showing the Position of the Critical Points. By far the most comprehen-
sive set of determinations of the critical points of iron and steel was made by Car-
penter and Keeling. Their results are shown in the table on the following page.
The table includes the critical points occurring during the solidification period of the
various steels and irons investigated. These will be considered in another lesson.

Relative Quantities of Heat Evolved or Absorbed at the Critical Points. The
various critical points that have been considered in the preceding pages do not indi-
cate evolutions or absorptions of equal quantities of heat; they are not of equal in-
tensity. The point A 3 is very marked and sharply denned in carbonless iron but
decreases rapidly in intensity as the carbon increases. The point A 2 is relatively
feeble and not very sharply denned and as already mentioned shows a tendency to
cover a considerable range of temperature. Its intensity moreover is little affected
by the carbon content of the steel. The point A 3 . 2 , being a merging of A 3 and A 2 , is
more intense than A 2 but less intense than A 3 in carbonless iron owing to the fact
that when the merging takes place the A 3 point has lost much of its intensity. The

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

Begin

Maxim

" p '" ~

"" "

-^
K

?
s

s sags

Maxim

fioov^-t ; oot*-c*o

e2?S5trP : : r-^^s

i c-. t-- 1- i C>>> Cc

Ar.m.

Solidi

Maximum.

S S

-tr-ti-t i-Hr

0000000 0000000 Ortr-lrtrH^lr^

10 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

point Ai is feeble in very low carbon steel but its intensity increases rapidly with the
carbon content, becoming so great as to cause the metal to glow or recalesce as pre-
viously described and being maximum for steel of eutectoid composition. These
differences in the thermal values of the critical points will be explained in another
lesson.

Graphical Representation of the Position and Magnitude of the Critical Points.
The position of the critical points corresponding to various percentages of carbon is
illustrated graphically in Figure 2. The diagram refers to the critical points of cool-
ing, i.e. the Ar points, and it should be borne in mind that the corresponding points
on heating, the Ac points, occur some 25 to 50 deg. higher, with the exception of
the point Ac 2 which seems to occupy nearly the same position as the point Ar 2 . An
attempt has been made in this diagram to indicate the relative intensities of the
various points by shaded areas of proportional thickness on both sides of the lines
indicating their position. This is based chiefly on theoretical considerations and is
in accordance with the generally accepted views regarding the causes of the critical
points as explained in the next lesson. An examination of the diagram shows (1)
that the point Ar 3 intense in carbonless iron decreases gradually in intensity as the
carbon increases, (2) that the intensity of Ar 2 is not greatly affected by the carbon
content, (3) that Ar 3 . 2 fairly intense at first becomes rapidly feebler and finally dis-
appears just as it meets Ar!, (4) that Ari at first very faint becomes more marked
with increased carbon, being maximum for a carbon content of some 0.85 per cent
(the eutectoid point), (5) that the point Ar 3 . 2 .i very intense at the eutectoid point
gradually loses some of its intensity, although always remaining pronounced, and
(6) that the point Ar cm very faint near the eutectoid composition increases in in-
tensity with the carbon content.

These theoretical inferences are well supported by experimental evidences in the
case of the magnitude of the points Ari and Ar 3 . 2 .i and quite satisfactorily in regard
to Ar 3 and Ar cm . They ascribed to the points Ar 2 and Ar 3 . 2 , however, a magnitude
and a sharpness which is not borne out by experiments as later explained when it
will also be seen that some writers doubt the accuracy of the explanation generally
offered to account for the point A 2 .

The diagram, therefore, while undoubtedly useful, is probably but approximately
accurate and likely to be modified with increased knowledge of the facts it aims to
depict.

Determination of the Thermal Critical Points. The thermal critical points are
universally determined by means of the Le Chatelier thermo-electric pyrometer.
Indeed it is the invention of this invaluable instrument that made the detection of
the upper critical points possible. Had it not been invented we probably would still
be in ignorance of the existence of the upper points, while we would have but little
knowledge of the exact position of the point of recalescence. The necessary experi-
mental manipulations will be found described in the instructions given to carry on
the experiments appended to this lesson.

Cooling and Heating Curves. The determination of the thermal critical points
calls for the construction of heating and cooling curves. In these curves successive
falls (or rises) of temperature, say of 10 deg. C., 6 - 10, 6 - 20, - 30 . . . are plotted
as ordinates, while as abscissae are plotted (a) the corresponding time intervals in sec-
onds, t, t', t", t'" . . . elapsed since the beginning of the observation, or (6) the
actual intervals of time t' - t, t" - t', t"' - t" . . . required for each noted fall of

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

temperature. In other words the coordinates are 6 and t in the first instance, 6 and

in the second. The curve obtained by the first method is known as a time-
d0

temperature curve while the second method yields an inverse rate curve. 1

Time-temperature curves representing the heating and cooling of pure iron are
shown in Figure 3. While in these curves the evolutions or absorptions of heat cor-

1500
HOC
1800
1200

1100

1000
MO

800
TOO
600
600

Solid! 11

nation To

Iron

V^SW

800

Ac 2

^ /3 Iron

srtr

<r Iron
\

10 90 80 W CO .70 80 SO 101

Wumtes

Fig. 3. Time-temperature curves. Heating and cooling of pure iron.

(Goerens.)

responding to the points A 3 and A 2 can be detected, they do not stand out very
conspicuously, and it may well be feared that slight thermal retardations might
escape detection in curves of this kind since they would cause but very slight jogs
in the curve. These considerations led Osmond to adopt the inverse rate method
for the plotting of thermal curves. Curves of this type are shown in Figures 4 and 5.
The thermal points correspond to sharp peaks in the curves, the lengths of which
are roughly proportional to the amount of heat evolved on cooling or absorbed on

1 It is evident that similar curves would result from reversing the observations, i.e. noting the
successive falls of temperatures 9, e' , e" . . . corresponding to equal intervals of time, say of 15
seconds, t + 15, t + 30, t + 45 . . . and plotting the former as ordinates and the latter as abscissa;.

The coordinates in this case would be and t.

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

Fig. 4. Inverse rate curves. Cooling of steels containing respectively
0.02, 0.14, 0.45, and 1.24 per cent carbon. (Osmond.)

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

heating. This method is quite universally applied unless, as later explained, a neutral
body is used for the determination of the critical points.

Use of Neutral Bodies. The method described above for the detection of the
thermal critical points is open to the objection that the rate of cooling or heating
of the steel under observation is necessarily affected by any irregularity in the cooling
or heating of the furnace itself and by other outside agencies. These disturbing fac-
tors introduce irregularities in the thermal curves which may render their interpreta-
tion difficult and may indeed altogether hide the existence of critical points where
but a very small amount of heat is evolved or absorbed. Again, on cooling for in-
stance, when a critical point is reached the temperature of the metal is affected in
opposite directions (1) by the cooling furnace which has a tendency to lower its tern-

TIM
IN SIC

NOS
70

70
60
50
40
30

LU
CO

a
a

o
i

V5$

^Q-

sfc^-

: rrrr--

^"***~-

^. u _

MANG^.^-

Jllfi

_ _ m

'-.*!

00*1 150 1100* 1050 1000* 950 000 850 800* 750* 700* 650* 600 550* 50045046o3'50

TEMPERATURE.

Fig. 5. Inverse rate curves. Cooling of pure iron, low carbon steel, high carbon
steel, and manganese steel. (Roberts-Austen.)

perature and (2) by the evolution of heat occurring at the critical point, the effect of
which is to raise its temperature. It is evident that the cooling influence of the fur-
nace has a tendency to decrease the apparent magnitude of the critical point and
therefore to mask it.

The elimination of these objectionable influences should result in sharper thermal
curves and in the detection of faint evolutions or absorptions of heat. This was ac-
complished by Roberts-Austen through the use of a neutral body and double thermo-
couple so connected that the difference of temperature between the metal under
investigation and the neutral body is recorded, as well as the actual temperature of
the metal. 1 If the heat capacities and emissivities of the metal and of the neutral
piece were identical their temperature would be exactly the same except at the crit-
ical points, when heat is evolved or absorbed by the metal while the neutral body is, of

1 For the arrangement of the galvanometers, connections, etc., see the description of the Saladin-
Le Chatelier-Pellin instrument described under "Apparatus" and the description of other instru-
ments using neutral bodies in an appendix to these lessons.

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

900 gOO 700 600 500 900 800 700 600 500

C = 0,06

Mn=o,3i

00,10

C=O.I5

= 0.22

Mn=0,53

C= 0,36
Mn^O, 52

C- 0,4-8
Mn--0.45

~=0, 66
In = 0.7 1

900 V00 700 600' soo Temperatures 300' foo' 700' 6OO'

Fig. g, Difference curves. Cooling and heating of various steels. (Saladin.)

course, free from any such thermal disturbance. Any difference of temperature
between the two pieces, therefore, would indicate a critical point. Since it is generally
impossible, however, to use a neutral body having exactly the same heat capacity as
that of the metal under observation, it will be apparent that there will always be a
difference between the temperatures of the two pieces, one always lagging behind the

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

other, and that the critical points will correspond to sudden increase in the difference
between their respective temperatures. Since, however, the critical points are now
caused solely by abrupt differences between the temperatures of the two bodies, they
are freed from the irregularities mentioned above as well as from the masking in-
fluence of the falling or rising temperature of the furnace, seeing that both pieces are
now equally affected, and the curves obtained should indicate more sharply and con-
spicuously the existence of even faint absorptions or evolutions of heat.

The neutral body should, of course, be free from any thermal transformation
within the range of temperature covered by the experiments. Platinum, porcelain,
clay, 25 per cent nickel steel, and (by the author) austenitic manganese steel have
been used. The plotting of the thermal curves when a neutral body is used may be
done in two different ways, (1) successive falls (or rises) - 10, & - 20, - 30 ... of
the temperature of the metal as indicated by one of the galvanometers may be plotted

tooo 900

600

too

Fig. 7. Difference curves. Cooling and heating curves taken on same
photographic plate. (Saladin.)

as ordinates against the corresponding differences in temperature - 0\, 6' - 6'\,
0" - B'\ ... of the two cooling bodies, as indicated by the second galvanometer,
the coordinates in this case being 6 and 6 - 61, and the curve known as a "difference"
curve, or (2) according to Rosenhain, successive falls of temperature may be plotted
as ordinates against the corresponding rate of cooling for each degree of temperature

Q - Qi & - 6 i & - 6 i ag a k sc j ssa;) the coordinates being in this method 6 and
6 6' 6"

d (0 ~ QV and the curve known as a "derived differential" curve.

Difference curves are shown in Figures 6, 7, and 8. The curves of Figures 6 and 7
were taken with a Saladin-Le Chatelier-Pellin instrument, Figure 6 shows the crit-
ical points on heating and cooling of a series of carbon steels containing from 0.06 to
0.66 per cent carbon, the cooling and heating curves having been taken on separate
photographic plates. In Figure 7 is shown the heating and cooling curves taken on
the same plate of two steels containing respectively 0.36 and 0.46 per cent carbon.

The curves of Figure 8 are difference curves of a series of very pure carbon steels
taken by Carpenter and Keeling.

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

Fig. 8. Difference curves. Cooling of a series of very pure
carbon steels. (Carpenter and Keeling.)

The purpose of Rosenhain's derived differential method of plotting is to eliminate
the irregularities from which difference curves still suffer and which are due chiefly to
differences between the heat capacities and emissivities of the sample and neutral
body resulting in differences in their rates of cooling and heating. The resulting

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

curves cannot, of course, be autographically recorded. They call for the replotting
of the data afforded by the difference (0 vs. 6 - 61) curves.

Additional Illustrations of Cooling Curves. The different types of cooling curves
described in the preceding pages are well illustrated in Figure 9. These curves were
constructed from the data given in the following table, in which each unit in t repre-
sents intervals of time of 15 seconds, the corresponding temperatures of the sample,
and - 0\ corresponding differences of temperature between the sample and a neutral
body cooling under identical conditions. The curve B, representing the cooling curve
of a neutral body free from critical transformations, has been added for comparison.

BSD

64O

B30

O20

BIO

Temperature-time
curve.

At:

Inverse rale
curve.

e-e,

Difference
curve.

Derived differen-
tial curve.

Fig. 9. Different types of cooling curves. (Desch.)

0-0 t

850.0

8.5

829.0

11.0

848.0

8.5

825.0

8.2

844.7

7.5

19.5

823.3

7.3

842.0

7.0

822.2

6.7

839.5

6.3

821.7

7.7

- '838.5

7.0

821.5

8.5

838.2

8.8

821.3

9.8

838.1

10.2

821.1

10.1

838.0

12.0

24.5

819.0

9.5

837.9

13.6

815.0

6.0

837.5

15.5

813.0

5.0

836.0

14.6

811.6

4.7

833.0

13.0

Self-Recording Pyrometers. With the use of neutral bodies self-recording in-
struments are generally employed. The Saladin-Le Chatelier-Pellin autographic
pyrometer has been described under "Apparatus" and other types of self-recording
instruments will be found described and illustrated in an appendix to these lessons.
The self-recording may be by means of photographic plates or by some other mechan-
ical devices. The former method calls for the use of mirror galvanometers sending a

LESSON VII THE THERMAL CRITICAL POINTS OF STEEL 19

beam of light upon the photographic plate while in other autographic recorders
needle galvanometers are used. The relative merits between photographic recorders
and other types are summed up by Burgess as follows :

"It is evidently of great advantage to use self-recording apparatus when possible,
and it then becomes necessary to choose between the photographic type and the
autographic. The latter possesses the advantage that the experimenter may watch
any part of the record, and can therefore control the operation and at any moment
vary the conditions affecting the experiment; whereas with a photographic recording
apparatus, as usually constructed, the observer does not know whether or not the
experiment is progressing properly until it is finished and he has developed the sen-
sitive plate. The manipulation by the photographic method 4s usually also more
delicate and time consuming and the adjustment less sure, and the record often re-
quires further graphical interpretation. The autographic method is in general not
adapted for interpreting phenomena taking place within an interval of a few seconds,
so that for very rapid cooling it is necessary to employ the photographic method. It
is possible to construct the photographic recorder so as to obtain a very considerable
range of speeds with the same apparatus, while it is difficult and costly to construct
an autographic recorder having more than two speeds."

Historical. A brief historical sketch of the discovery of the critical points of
iron and steel will not be without interest. In 1868 Tschernoff in studying the har-
dening of steel used the notation A for the temperature at which hardening by rapid
cooling becomes suddenly possible in high carbon steel. This was the point A 3 .2.i.

In 1869 Gore noted that at a dark red heat steel exhibited on cooling a spon-
taneous dilatation of short duration followed by normal contraction. Evidently the
point of recalescence An or Ar 3 . 2 .i.

In 1873 Barrett repeated Gore's experiments and discovered, on heating, a momen-
tary contraction at nearly the same temperature as the dilatation on cooling. This
was the point Aci or ACS.S.I. He further noted that this dilatation or contraction was
very feeble in iron (the AI point in very low carbon steel) and very marked in hard
steel (the point AS.J.I). Barrett also discovered the spontaneous glow taking place
on cooling a wire and gave it the name of recalescence.

In 1879 Barrus showed that the increase of hardness resulting from quenching
was not gradual but sudden, thus pointing to the existence of a thermal critical point
(the AI point).

In 1885 Osmond published his discovery of the upper critical points As and A 2 in
iron and low carbon steel and gave the first accurate determination of the position of
the point AI.

Experiments

Small samples of steel should be obtained containing respectively some 0.10, 0.25,
0.50, 0.85, and 1.50 per cent of carbon, preferably J/6 in. round or square and 1 in.
long. A small hole, about % in. in diameter, should be drilled in the end of each
sample and extended to the center or even right through the sample. The end of
the thermo-couple of a Le Chatelier pyrometer should be inserted in these small
holes and firmly packed with loose asbestos. The sample thus attached to the thermo-
couple should now bo introduced in a suitable furnace, preferably an electric resis-
tance furnace (see "Apparatus") and gradually heated to a temperature of some

20 LESSON VII THE THERMAL CRITICAL POINTS OF STEEL

1000 deg. C. While the metal is thus being heated its temperature should be observed
and the time intervals required for each rise of temperature of say 10 deg. care-
fully recorded. In a similar way the rate of cooling should be noted while the metal
cools from 1000 to 500 deg. Inverse rate curves should be constructed which will
bring out the evolutions or absorptions of heat having taken place during the
thermal treatment. The accuracy of the method greatly depends upon the care
and skill exercised. The experiment calls for the services of two observers in order
that one may watch the galvanometer while the other notes and records the corre-
sponding times.

There is no difficulty by this method in detecting the points Ai of medium carbon
steel and A 3 . 2 .i of high carbon steel because very marked retardations occur at these
points. The detection of the points A 3 and A 2 and especially A cm , where but small
evolutions of heat are involved, on the contrary is not always possible by this
method, their satisfactory detection calling for the use of neutral bodies and self-
recording instruments. The same samples may be used.

Examination

I. Describe the occurrence of the thermal critical points in a steel containing 0.25

per cent carbon.

II. Construct the inverse rate cooling curve of a piece of steel whose cooling through
successive ranges of 10 deg. C. has required the time intervals indicated be-
low:

TEMPERATURES TIME

DEC. C. SECONDS

750740 10

4030 11

3020 12

2010 13

710700 14

700690 16

9080 20

8070 45

7060 20

6050 17

5040 15

4030 16

3020 18

2010 20

610600 23

III. Describe the influence of the rate of heating and cooling upon the position of

the critical points.

IV. Explain why the points Ac 3 and Ar 3 do not correspond to the same temperature. <

V. Explain the use of neutral bodies in the determination of critical points by the
pyrometric method.

LESSON VIII

THE THERMAL CRITICAL POINTS OF STEEL

THEIR CAUSES

The thermal critical points described in the preceding lesson are evidently out-
ward manifestations of internal transformations taking place spontaneously at certain
critical temperatures. We should now inquire into the nature of these transforma-
tions.

Let us remember that there are but three well-known causes of spontaneous evo-
lutions of heat in cooling bodies and of spontaneous absorptions on heating. These
are: (1) formation of chemical compounds, a phenomenon which is almost always
accompanied by a spontaneous evolution of heat (the heat of formation), and the
reverse phase, the dissociation of the compound with absorption of heat (the heat of
dissociation), (2) changes of state, that is, the passage of a substance from the solid
to the liquid or from the liquid to the gaseous state, or directly from the solid to the
gaseous state, which changes are always accompanied by spontaneous absorptions of
heat, and the opposite phases of the same phenomena, the passage of a body from
the gaseous to the liquid or from the liquid to the solid state, when heat is evolved
(the latent heat of solidification in case of a substance passing from the liquid to the
solid state, etc.) these evolutions or absorptions of heat, as the case may be, main-
taining the temperature of the substance constant while a change of state is in progress
as, for instance, during solidification or melting, (3) allotropic or polymorphic trans-
formations which are always accompanied by an evolution of heat when the body
passes from one allotropic condition to another, and by an absorption of heat when
it returns to its first allotropic form or vice versa. The meaning of allotropy has been
explained in Lesson II. .

Causes of the Upper Points A 3 and A 2 in Carbonless Iron. It has been seen that
at these points spontaneous evolutions or absorptions of heat occur in chemically
pure iron which, since they are not accompanied by any change of state (the metal
being considerably below its solidification point) must, it seems, necessarily indicate
the existence of iron under three allotropic forms. It has been mentioned in Lesson
II that the allotropic form stable above A 3 is known as y (gamma) iron, that stable
between A 3 and A 2 as ft (beta) iron, and the form stable below A 2 as a (alpha) iron.
The following then takes place during the cooling and heating of pure iron: as the
metal cools from a high temperature when the point Ar 3 is reached, it passes from
the gamma to the beta condition with evolution of heat, while at Ar 2 it passes from
the beta to the alpha form also with evolution of heat. On heating the reversals
take place, the iron passing with absorptions of heat from the alpha to the beta con-
dition at Ac2 and from the beta to the gamma condition at Acs. That the point AS
indicates an allotropic transformation is universally admitted, no one doubting the

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

existence of iron in at least two allotropic conditions. Most authoritative writers
believe with Osmond that the point A2 also indicates an allotropic transformation,
and that iron, therefore, assumes three distinct allotropic forms, as explained above.

free

free Alpha_
ferr-ite

r~err/te

_ /- /qu/d Solufron of
/fon crncj Car~hon

Solid/ f /'cat ion poitrt

-5o//iy -So/uf/on o/
ofrrj mcf /f~<on and
Ccrr-ibon CAc/stsnite)

Solution
CA usfert/te >

5o//'cf So/of /on

-A,

P&ar/jfe

jB /-/ C Temperature

Fig. 1. Diagram depicting structural changes in 0.20 per cent carbon steel
as it cools slowly from the molten condition to atmospheric temperature.

As eminent an observer as Le Chatelier, however, has expressed doubts as to the
allotropic character of the point A 2 . His reasons will be considered later. In these
lessons iron will be assumed to exist in three allotropic conditions, of which A 3 and

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL 3

A 2 are the transformation points, this being the generally accepted theory and, in
the author's opinion, the most probable one.

Causes of the Upper Critical Points A., and A 2 in Low Carbon Steel. As might
be expected the points A 3 and A 2 occurring in very low carbon steel also indicate
allotropic changes in the metal. According to most writers, however, only that por-
tion of the metal which at ordinary temperature exists as free ferrite is here affected.
To make this matter clear it will be necessary to anticipate somewhat our subject,
\vhile the diagram shown in Figure 1 will be useful.

In this diagram the three critical points of steel containing 0.20 per cent carbon,
A 3 , AS, and AI, as well as its solidification point and atmospheric temperature are
represented by parallel lines drawn at suitable intervals in the scale "of temperature.
The metal is represented by the rectangular area ABCD. The diagram illustrates
the following facts later to be discussed at greater length: (1) in the molten condi-
tion steel is considered to be a liquid solution of iron and carbon, (2) on reaching its
solidification point the metal is converted into a solid solution of gamma iron and
carbon known as austenite, (3) upon reaching Ar 3 some ferrite begins to be set free,
(4) the ferrite as it is set free assumes the beta state, this liberation of ferrite and its
allotropic transformation being probably one and the same phenomenon, (5) the
formation of free ferrite continues as the steel cools from Ar 3 to Ari, EFG represent-
ing the ferrite thus liberated, (6) on reaching the point Ar 2 the ferrite liberated be-
tween Ar 3 and Ar 2 , ML in the diagram, passes from the beta to the alpha condition,
(7) the ferrite liberated between Ar 2 and Ari, LNF in the diagram, assumes the alpha
condition (according to some writers) without passing by the beta condition, while in
the opinion of others the beta condition is assumed but is immediately followed by
the alpha state, (8) while ferrite is being set free, the balance of the steel, EKIF,
(according to most writers) preserves its condition of solid solution, gamma iron
plus carbon, (9) upon reaching the point Ar t the residual solid solution, FIi, is con-
verted bodily into pearlite, (10) from Ar t down to atmospheric temperature no fur-
ther structural change takes place, the steel being finally made up of BH = GF per
cent ferrite and of HC = FI per cent pearlite.

On heating the opposite changes take place: (1) at Aci transformation of FI
pearlite into FI solid solution (gamma iron + carbon = austenite), while this solid
solution begins immediately to assimilate some of the free ferrite, which as it is as-
similated passes to the gamma condition, (2) between AI and A 3 absorption of free
ferrite continues, being completed at Acs, (3) on reaching the point Ac 2 the ferrite,
ML in the diagram, which has not been absorbed between Aci and Ac 2 now passes to
the beta condition. This diagram depicts accurately the generally accepted views in
regard tc the meaning of the critical points. If these views are correct several inter-
esting inferences may be drawn as to the relative intensities of the critical points.
The point A 3 in low carbon steel does not indicate a complete transformation, as too
often loosely stated, but merely the beginning, at Ar 3 , or the end, at Ac 3 , of a trans-
formation extending over a considerable range of temperature, i.e. from AI to A 3 .
Theoretically, therefore, it would seem as if the point A 3 must correspond to a mere
change of direction in cooling and heating curves rather than to well-marked jogs.
The fact that a decided jog marks the point Ar 3 in very low carbon steel might be
ascribed to hysteresis, the metal cooling to a temperature below that at which the
AS change is due so that when the transformation begins to take place it does so with
added intensity, hence the jog. The jog corresponding to the Ac 3 point of very low

4 LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

carbon steel is not so readily explained. But is not this jog much less pronounced
than the one corresponding to Ar 3 ? The point Ar marks (1) a complete transforma-
tion, namely, the passage from the beta to the alpha state of the free ferrite liberated
between Ar 3 and Ar 2 and (2) the beginning of a transformation, namely, the passage
to the alpha condition of the ferrite which continues to be liberated between Ar 2 and
Ari. Because of the complete transformation implied by the point A 2 we readily
understand that it should correspond to a jog both in the heating and cooling curves,
and since Ar 2 is due chiefly to the allotropic transformation of the ferrite liberated
between Ar 3 and Ar 2 , we readily understand why it should occur at nearly the same
temperature regardless of the carbon content. In the light of what precedes, how-
ever, the point A 3 in steel instead of being regarded as the manifestation of trans-
formations occurring and completing themselves at a certain temperature, in reality
indicates the beginning or end of transformations extending over considerable range
of temperature, namely, from A 3 to AI.

Cause of the Point A 3 .2. It has been seen that the point A 3 . 2 is apparently a
merging of the points A 3 and A 2 of lower carbon steel and it seems natural to infer
that the transformations which these points indicate, namely, the two allotropic
changes, are here likewise merged, that is, that they now take place at the same tem-
perature. In other words that when the point Ar 3 . 2 is reached on cooling the iron
passes from the gamma to the beta and then immediately to the alpha state, the
heat evolved being due to this double allotropic transformation. Some writers have
claimed, however, that at the point Ar 3 . 2 the iron passes directly from the gamma to
the alpha condition, the change of gamma to beta being suppressed in steel contain-
ing over 0.35 per cent carbon or thereabout. If such hypothesis were true it would
have some important bearing upon the probable theory of the hardening of steel as
explained in another lesson. In the author's opinion the more generally accepted
view is better supported by experimental facts and other evidences and in these
lessons the point A 3 . 2 will be considered as implying a double allotropic change. Most
metallographists believe that like the independent points A 3 and A 2 the double point
A 3 . 2 is the result of allotropic changes affecting the free ferrite only.

This setting free and allotropic transformation of ferrite is depicted diagrammati-
cally in Figure 2 in the case of steel containing 0.60 per cent carbon and having, there-
fore, the two critical points A 3 . 2 and AL EOF indicates the gradual liberation of
ferrite and its conversion to the alpha state as the metal cools from Ar 3 . 2 to Ari, the
steel, after complete cooling, being made up of BH = GF per cent ferrite and CH = FI
per cent pearlite.

If, as generally stated, the allotropic transformation of which A 3 . 2 is a manifesta-
tion affects only the free (pro-eutectoid) ferrite, the intensity of the point A 3 . 2 must
decrease rapidly with decreasing pro-eutectoid ferrite, i.e. as the eutectoid composi-
tion is approached, and this point must vanish altogether as it meets the point AI
(see Lesson VII, Fig. 2), from which it further follows that the single point of eutec-
toid steel is not in reality a triple point as the notation A 3 . 2 .i would imply, resulting
from the merging of AI and A 3 . 2 , but that on the contrary it remains a single point,
being merely the continuation of the AI point of hypo-eutectoid steel.

Cause of the Point AI. It has been seen that the point AI does not occur in
carbonless iron, only feebly in iron containing little carbon, and with increased in-
tensity as the carbon increases to the eutectoid point. The conclusion seems irre-
sistible that the point AI must be closely related to the carbon, that it must indicate

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL 5

a sudden change in its condition. If steel be rapidly cooled from above the point Aci
and then treated with certain dilute acids, practically all the carbon escapes as hy-
drocarbons, whereas the same steel after slow cooling through Ari when similarly

5o//c/ solution
(Austen/ fe)

Tempera-furs

Fig. 2. Diagram depicting structural changes in 0.60 per cent carbon steel
as it cools slowly from the molten condition to atmospheric temperature.

treated yields a carbonaceous residue, which upon being analyzed is found to consist
of the carbide Fe 3 C. It is assumed that upon quick cooling we retain the carbon,
partially at least, in the form in which it normally exists above AI, and seeing that

6 LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

when subjected to a similar treatment this carbon behaves so very differently from
the carbon of slowly cooled steel, the conclusion is very logical that carbon exists
above AI in a different condition from that normal below AI. Above A: it is called
"hardening" carbon, below AI "cement" carbon. On heating steel past the point
A.CI the carbon changes from the cement to the hardening condition, and vice versa
on cooling at Ari from the hardening to the cement condition. It is, moreover, gen-
erally believed that this hardening carbon is carbon in solid solution in the iron. If
it be so the heat evolved at Ari is clearly in part at least the heat of formation of
the carbide Fe 3 C and the heat absorbed at Aci clearly the heat of dissociation of that
carbide which now is resolved again into its elements according to the reversible

rcaction

The intensity of the AI point should then increase as the carbon increases or,
rather, as the amount of pearlite increases, and should be maximum, therefore, at the
eutectoid point as indicated in Figure 2, Lesson VII. With higher carbon content it
diminishes slightly because the free (pro-eutectoid) cementite which is now present
takes no part in the transformation occurring at Ari, having been formed at^a higher
temperature, namely, at Ar cm , as later explained.

It would seem that the cause of the AI point, i.e. the point of recalescence, is in
this way explained in a perfectly satisfactory manner. The correctness of this theory
appears to be further supported by microscopical evidences which reveal the presence
of Fe 3 C in slowly cooled steel while pointing to the probable absence of it in suddenly
cooled steel.

In recent years, however, it has seemed more and more probable to students of
metallography that it is not carbon in its elementary state which is dissolved in
iron at a high temperature, but rather the carbide Fe 3 C itself and that the difference
between the behavior of the carbon in hardened steel and in slowly cooled steel might
well be satisfactorily accounted for on the ground that in hardened steel Fe 3 C is dis-
solved in iron and in that form is much more readily acted upon by acids, being there-
by converted into hydrocarbons, whereas Fe 3 C when in the free crystallized condition.
as in slowly cooled steel, resists the action of the acids and remains undissolved. If it
is Fe 3 C and not C which is dissolved in iron above the critical range, it is evident
that the point Ari cannot be caused by the formation of Fe 3 C. But it may well be
due to the crystallization or falling out of solution of Fe 3 C. To be sure this is a fall-
ing out of a solid solution, but cannot we conceive that the falling out of a constit-
uent of a solid solution is accompanied by an evolution of heat even if it does not
imply a change of state? In other words is it not possible, or even probable, that
crystallization in the solid state is accompanied by an evolution of heat? Surely this
crystallization implies an allotropic or at least a polymorphic transformation and
are not such transformations always accompanied by heat evolutions?

The author offers these thoughts as possibly worthy of attention and as a possible
explanation of the evolution of heat at Ar t if we assume that Fe 3 C and not C, as it
now seems so probable, is dissolved in iron above that point.

The Point A! an Allotropic Point. Most writers describe the point AI as purely
a carbon point, that is, a manifestation of a change affecting the condition of the
carbon only as explained in the foregoing pages. These same writers, however, as-
sert that the upper critical points, A 3 and A 2 in low carbon steel or A 3 .2 in higher car-
bon steel, affect only the condition of the free (pro-eutectoid) ferrite. In this they

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL 7

are inconsistent, for if the upper point or points indicate allotropic transformation of
the free ferrite only then the lower point A t is decidedly an allotropic point seeing
that it corresponds to allotropic transformations of the pearlite-ferrite and that in
steel containing more than some 0.40 per cent carbon there is more pearlite-ferrite
than free ferrite. In other words the point AI is always an allotropic point indicating
an allotropic transformation of the pearlite-ferrite similar to the allotropic transfor-
mation of the free ferrite occurring at the upper points, and in case of steel with
more than 0.40 per cent carbon the allotropic change taking place at AI affects a larger
bulk of iron than the change at A 3 . 2 . To make the matter clear let us consider (Fig. 2)
a steel containing some 0.60 per cent of carbon and, therefore^made up after slow
cooling of 72 per cent of pearlite and 28 per cent of free ferrite. This steel will con-
tain about 72 x | = 63 per cent of pearlite-ferrite represented by FO in Figure 2.
When the point AI is reached this 63 per cent of iron is still in the gamma condi-
tion (according to the general belief) and now passes to the alpha condition either
directly or first assuming the beta state. The allotropic character of the point AI is
therefore evident. Indeed it is sufficient to account for the heat evolved at Ari or
absorbed at Aci without the assistance of any change occurring in the carbon condi-
tion, for it is in perfect agreement with the increased intensity of the point AI as the
carbon increases and with its maximum at the eutectoid composition, since as the
carbon increases the amount of pearlite and therefore of pearlite-ferrite likewise in-
creases.

Summing up, three reasons may be given for the evolution of heat at Ari: (1) for-
mation of the carbide Fe 3 C based on the assumption that carbon as such is dissolved
in iron, (2) crystallization of the carbide FesC based on the assumption that this car-
bide is dissolved in iron and that crystallization not implying a change of state may
produce heat, and (3) allotropic transformation of the iron present in austenite of
eutectoid composition. It seems probable that both (2) and (3) contribute to the
heat developed at Ar^

Pearlite Formation. Whatever differences of opinion may exist as to the exact
cause or causes of the evolution of heat corresponding to the point Ari all agree that
it is due to the transformation of austenite of eutectoid composition (sometimes
called hardenite) into pearlite, i.e. the conversion of a solid solution into an aggre-
gate (ferrite plus cementite). It is well to bear in mind the changes in the condition
of the iron and carbon which this transformation seems to imply: (1) passage of the
iron from the gamma to the beta condition, (2) immediately followed by its conver-
sion into alpha iron or, according to some writers, (1 and 2) the conversion of
gamma iron directly into alpha iron, skipping the beta state, (3) the crystallizing of
alpha iron into parallel plates or lamellae, and (4a) the formation and crystallizing
or (46) the crystallizing only of Fe 3 C into parallel plates alternating with the ferrite
plates.

Cause of the Point A cm . The point Ar cm undoubtedly indicates the beginning
of the liberation of free cementite in hyper-eutectoid steel as it cools from Ar cm to
Ar 3 . 2 .i. This gradual formation of free cementite is well shown in Figure 3 where it
is represented by the triangle EFG. When the point A 3 . 2 .t is reached the residual
austenite, now of eutectoid composition, is converted bodily into pearlite, the steel
consisting finally of BH = GF per cent free cementite and DC = FI per cent pearl-
ite. It will be seen that this upper point of hyper-eutectoid steel, like the points A 3
and A 2 of hypo-eutectoid steel, does not indicate a complete transformation but

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

merely the beginning of a transformation covering a wide range of temperature,
namely, from A cm to AS.Z.I. If it corresponds to a jog rather than to a mere change
of direction in cooling curves this must be ascribed to hysteresis and its tendency to

L/qt//c/ so/uf/on
iron and carbon

jotid'ficcrfron

point

So//c/

gamm iron anc/

ca/-bon fAustenifel

So//c/ so/ut/on
(A ustenife)

3-L-l

Fig. 3. Diagram depicting structural changes in 1.25 per cent carbon steel
as it cools slowly from the molten condition to atmospheric temperature.

accentuate the beginning of a transformation as previously explained. The evolu-
tion of heat corresponding to the liberation of free cementite may be explained in
two ways: (1) actual formation of Fe 3 C, carbon and not the carbide being in solution

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL 9

above Ar cm , or (2) crystallization or falling out of solution of Fe 3 C based on the more
probable assumption that Fe 3 C and not C is dissolved by the iron above Ar cm , and
on the further assumption that this crystallization in the solid state, since it evidently
implies an allotropic or, at least, polymorphic change, must be accompanied by an
evolution of heat.

Since in hyper-eutectoid steel containing even as much as 1.5 per cent carbon
there is but a small proportion of free cementite (some 11 per cent) the point Ar cm is
caused by the evolution of but a small amount of heat and must therefore be faint.
Its intensity, moreover, must decrease as the carbon decreases and the point must
disappear altogether as the eutectoid composition is reached, that is, just as it meets
the single point of eutectoid steel. This is indicated in Figure 2 of Lesson VII.

Allotropy of Cementite. If we believe, as most metallographists now do, that
FesC and not C forms a solid solution with carbon above the point AI or A 3 .2.i, it
follows that this dissolved Fe 3 C crystallizes or falls out of solution at certain critical
temperatures, namely, Ar cm for the free cementite of hyper-eutectoid steel and Ari
(or Ar 3 . 2 .i) for the pearlite-cementite of all steels, and that this crystallization is ac-
companied by an evolution of heat. This falling out of solution really implies a
spontaneous change of crystalline form and is therefore an evidence of polymorphism,
hence of allotropy, for if allotropy does not necessarily imply polymorphism, poly-
morphism implies allotropy. Are we not then justified in believing that Fe 3 C may
exist under two allotropic forms: (1) an allotropic variety soluble in iron, which we
may call gamma cementite and (2) an allotropic variety insoluble in iron, which we
may call alpha cementite, constituting the free cementite of hyper-eutectoid steel
and the hard plates of pearlite?

The fact that the crystallizing or falling out of solution of free ferrite in hypo-
eutectoid steel implies an allotropic transformation of the liberated ferrite, points
with force to a similar transformation forming part of the liberation of free cementite
in hyper-eutectoid steel. In the case of iron we are able to actually prove this allo-
tropy through the cooling of very pure iron and the testing of its properties while in
the case of cementite such direct proof is not yet at hand because of the difficulty of
obtaining pure cementite and of testing it after producing it.

To generalize, it would seem as if the crystallizing or falling out of solution of a
substance at certain critical temperatures always implies a spontaneous change of
crystalline form and, therefore, an allotropic transformation of the substance separa-
ting from the solution, whether that solution be liquid or solid. In the former case
the falling out of solution implies a change of state, the separating substance passing
from the liquid to the solid state, but does it make it less of an allotropic change?
Allotropic transformations which also imply changes of state might be called in-
stances of major allotropy to distinguish them from those instances in which changes
of internal energy are not accompanied by changes of state.

Cause of the Point A 3 . 2 .i in Eutectoid Steel. In the case of eutectoid steel the
solid solution (austenite) is originally of eutectoid composition and, therefore, on
cooling reaches the point Ar 3 . 2 .i without rejecting either ferrite or cementite, hence
the absence of upper points in eutectoid steel. On passing through the point Ar 3 . 2 .i
(Fig. 4) this austenite is converted into pearlite. Pearlite contains 87.50 per cent of
ferrite which undergoes allotropic transformation at Ar 3 .s.i, hence the intense allo-
tropic character of this point. The crystallizing of cementite probably contributes
also to the heat evolved at Ar 3 . 2 .i whether it implies the formation of Fe 3 C or

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

not as explained before. It has also been argued that this crystallizing of cementite
probably implies an allotropic transformation, in which case the point A 3 . 2 .i (and AI

3-i-l

_
iron and

so/idtf/cof/on

/OO//7/

5o//cf
-gamma iron and
carbon (Ausfen/fe)

-Pear/if e

A //T? o.s/1 herjc

Fig. 4. Diagram depicting structural changes in eutectoid steel as it cools
slowly from the molten condition to atmospheric temperature.

as well) would be solely an allotropic point, resulting from the simultaneous allo-
tropic transformation of both the iron and the FesC of austenite of eutectoid
composition.

Finally let us bear in mind that notwithstanding its notation the single point of

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL 11

eutectoid steel does not in fact result from the merging of the several points of
hypo-eutectoid steel, being merely a continuation of the point AI. This point Ai
is essentially a "pearlite" point while the upper points of hypo-eutectoid steel
are "ferrite" points and the upper point of hyper-eutectoid steel is a "cementite"
point.

Cause of the Point A 3 . 2 .i in Hyper-Eutectoid Steel. The point A 3 . 2 .i in hyper-
eutectoid steel is of exactly the same nature as the point A 3 .i of eutectoid steel and
the point A! of hypo-eutectoid steel. It marks the formation of pearlite, bearing in
mind the various changes in the condition of the carbon and iron implied by that
formation. As the proportion of pearlite now decreases with "increase of carbon the
intensity of the point A 3 .2.i likewise diminishes.

Formation of Beta Iron. Allusion has been made in these pages on several
occasions to the different opinions entertained as to the formation of beta iron in steel
of various carbon contents. The matter should receive additional attention. The
following views are held: (1) iron in carbonless iron and in all grades of steel passes
from the gamma to the beta condition before assuming the alpha state, the absence
of an independent A 2 point in medium high and in high carbon steel notwithstand-
ing, (2) iron passes from the gamma to the beta condition only in those grades of iron
and steel which exhibit the point A 2 , and, therefore, only in carbonless iron and in
steel containing less than some 0.30 per cent carbon, in higher carbon steel the iron
passing directly from the gamma to the alpha state, (3) the beta condition does not
exist even in carbonless iron, the point A 2 not being an allotropic point.

Of these three different views the first is the one most generally accepted and, in
the author's opinion, best supported by evidence. It will be seen in another lesson
to afford the most acceptable theory of the hardening of steel. The second view is
based entirely upon the absence of the A 2 point in steel containing more than 0.30
per cent carbon, a very weak foundation, for there is no reason why the points A 3 . 2
and A 3 . 2 .i cannot include the gamma-to-beta transformation. The third view ad-
vanced by Le Chatelier will require more convincing arguments in order to uproot
the belief that the point A 2 , occurring as it does in the purest iron and, as will be
shown later, marking a sudden and pronounced change in its magnetic properties, is
an allotropic point.

Summary. The apparent causes of the thermal critical points of iron and steel
may be briefly summed up as follows:

The point Ar 3 of carbonless iron and of steel containing less than some 0.35 per
cent carbon marks the beginning of the liberation of ferrite (which liberation con-
tinues down to the point Ari) and the transformation of that ferrite from the gamma
to the beta condition, this setting free of ferrite and its allotropic transformation
being probably simultaneous.

The point Ar 2 of carbonless iron and of steel containing less than some 0.35 per
cent carbon indicates the transformation from the beta to the alpha condition of the
ferrite liberated between Ar 3 and Ar 2 and the beginning of the passage of the ferrite,
which continues to be liberated as the metal comes from Ar 2 to Arj, from the gamma
to the beta and then to the alpha condition or, as some writers claim, directly from
the gamma to the alpha state.

The point Ar 3 . 2 of steel containing somewhere between 0.35 and 0.85 per cent
carbon marks the beginning of the liberation of ferrite which takes place between
Ar 3 . 2 and Ar! and the passage of that ferrite from the gamma to the beta condition

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

and immediately to the alpha state or, according to some writers, directly from the
gamma to the alpha condition.

Since only the formation and transformation of free ferrite is involved at the points
A 3 , A 2 , and As. 2 these points may properly be called "ferrite" points.

The point Ai of hypo-eutectoid steel, as well as the point A 3 . 2 .i of eutectoid and
hyper-eutectoid steel, marks the rather sudden transformation of the residual solid
solution (austenite), now of eutectoid composition, into pearlite, bearing in mind the
changes in the conditions of the iron and carbon which such transformation implies.
The points Ai and A 3 . 2 .i may properly be called "pearlite" points.

The point Ar cm of hyper-eutectoid steel marks the beginning of the setting free
of cementite as the metal cools from Ar cm to Ar 3 . 2 .i, the liberation of cementite
probably involving an allotropic change of that constituent as previously explained.

//OO'

9 OO

Cr/f/ccf/
Range

700'

f~err/fe

Solid So/.uf/on
/ron and Carbon
(A usfen/fe)

c
f^errtfe

-/-Austen/fa

7oC O .25

Pear//fe

.35 /O

Cemenf/fe

Fig. 5. Diagram showing the relation between the critical points and the structural
composition of slowly cooled steel.

In Figure 5 an attempt has been made at showing diagrammatically the relation
between the critical points of steel and its structural composition after slow cooling.
It will be readily understood. The upper part of the diagram shows the location of
the critical points, the lower part the structural composition in percentages of ferrite,
pearlite, and cementite. Taking, for instance, a steel containing 0.25 per cent carbon :
above A 3 at A it is a solid solution of carbon and iron (austenite) ; on cooling through
Ar 3 at B some beta ferrite is set free, this liberation continuing from B to C; on cool-
ing through Ar 2 at C the free beta ferrite is converted into alpha ferrite while addi-
tional alpha ferrite forms as the metal cools to An, that is from C to D. The ferrite
liberated on cooling from B to D, that is on cooling through the critical range, is repre-
sented in the lower part of the diagram by DE which is also the final percentage of
free ferrite in the steel. As the metal cools through its An point at D, the residual

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

austenite, at present of eutectoid composition, is converted into pearlite, EF repre-
senting the pearlite here formed, i.e. the percentage of pearlite in the steel. The

A.-

f~ree- Beta
ferrtte

A,-

A/phct _

*7-<?e A/pno_
ferr/fe

_ So/u^'on of

/ron and Cor

Go/Tima /ron one/
Corfoon {Aust&ntte

So//c/ 5o/uf/'on of
~ 3efo /ron one/ $.

Car Aon (Mortensife ')

So//c/ v5o////o/? of
- A Ipho /r~on and
Carbon f'

A t

O / / x lerna&ro f urc

D ri C

Fig. 6. Diagram depicting structural changes in 0.20 per cent carbon steel slowly
cooled, assuming that the iron remaining in solution as well as the free iron (ferrite)
undergoes allo tropic changes. To be compared with Figure 1.

structural formation of any steel can be followed in the same way in this diagram.
It will be obvious that the vertical distances representing the percentages of ferrite,
pearlite, or cementite in any steel may also be regarded as proportional to the intensities

14 LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

of the corresponding critical points. For instance, the distance ED may be assumed
to be proportional to the combined intensities of the points Ar 3 and Ar 2 of a 0.25 per
cent carbon steel and the distance EF proportional to the intensity of the AI point.
Interpreted in this way the diagram indicates what has already been pointed out: (1)
that the intensities of A 3 and Aa decrease as the carbon increases, these points vanishing
on' reaching the eutectoid composition, (2) that the point AI, very faint at first, in-
creases rapidly with increased carbon, becoming maximum at the eutectoid point and
then decreasing, and (3) that the point A cm , always faint, increases slightly as the
carbon increases above 0.85 per cent.

Another View of the Allotropic Changes. It will be obvious from the description of the
underlying causes of the critical points given in the foregoing pages that, according to the general
belief, iron must first be freed from solid solution before it can undergo any allotropic changes or, at
least, its liberation from solution and its allotropic transformation take place simultaneously, the
latter never preceding the former. In 1906 the author expressed the opinion that it was far from
certain that its liberation from solution must precede, or at least be simultaneous with, the allotropic
changes affecting iron at certain critical temperatures. He ventured to put forward, in a tentative
way, the hypothesis that iron in solution might first undergo an allotropic transformation and then
be expelled in its new allotropic form. This view was not favorably received but, as it has not been
shown to be by any means untenable, the author still believes it worthy of record in these pages. It
is evident that if the allotropic transformation of iron from the gamma to the beta and then to the
alpha state precedes its liberation from solution, three solid solutions of carbon in iron are formed
during the slow cooling of steel, namely, carbon (or the carbide Fe 3 C) dissolved (1) in gamma iron,
(2) in beta iron, and (3) in alpha iron. The first solid solution is universally called austenite while
the hypothesis leads almost irresistibly to regarding the solid solution in beta iron as martensite and
the solid solution in alpha iron as troostite, two constituents to be described in another lesson.

With the assistance of the diagram, Figure 6, and by comparing it with Figure 1 the working of the
present hypothesis will be readily understood. In Figure 6 are depicted the structural changes taking
place during the slow cooling of steel containing 0.20 per cent carbon and therefore exhibiting the
three critical points As, A 2 , and AI. On cooling through the point A 3 the solid solution of carbon
and gamma iron (austenite) existing above As is converted into a solid solution of carbon in beta
iron (martensite?). In the beta condition, however, iron cannot be retained in solution and begins
immediately to be liberated, and its liberation continues as the metal cools down to A-,. Between
As and A 2 we have a sr>lid solution of beta iron decreasing in amount and increasing free beta ferrite.
On cooling through Ar. both the free beta ferrite and the dissolved beta ferrite pass to the alpha si ,-i te,
giving rise to the formation of free alpha ferrite and of a solution of carbon in alpha iron (troostite?).
On cooling from Ar 2 to Ari additional alpha ferrite is liberated while the proportion of carbon-alpha
iron solution decreases correspondingly. At An the remaining solution has become of eutectoid
composition and is converted bodily into pearlite, the mechanism of this transformation being well
understood. It will be evident that in the case of hypo-eutectoid steel having but one upper critical
point, Ars.2, in cooling through that point the metal would pass from the condition of a solid solution
of carbon in gamma iron to that of a solid solution in beta iron and then immediately to that of a
solid solution in alpha iron, the steel between Ar 3 . 2 and Arj being composed of this solution in alpha
iron (troostite?) and of free alpha ferrite. With eutectoid steel the following changes would take
place as it cools through its single critical point Ar 3 . 2 .i: (1) transformation of gamma iron solid solu-
tion (austenite) into beta iron solid solution (martensite?); (2) immediately followed by the forma-
tion of alpha iron solid solution (troostite?) ; (3) immediately followed by formation of pearlite.

The author thinks that the decreasing intensities of the points A 3 and A 2 as the carbon content
increases is the fact most difficult to reconcile with the hypothesis just outlined, for if these points
are due to allotropic transformations affecting the entire bulk of the steel their intensities should be
quite independent of the amount of carbon present. To be sure, this gradual diminution of the
magnitude of the points A 3 and A 2 as the carbon increases is likewise difficult of explanation in the
light of the universally accepted hypothesis that free iron only can be allotropically transformed, for
it has been made clear in the foregoing pages that these points must indicate then the beginnings of
transformations and not transformations carried to completion at those critical points, so that the

LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

in tensities of the points should be little affected by the magnitude of the transformations themselves,
that is, by the amount of free ferrite undergoing allotropio transformation or, which is the same thing,
by the percentage of carbon in the steel.

The view just outlined as to the mechanisms of the allotropic changes is further depicted dia-

ii^aa^^

Sa:i::v::i-!::L;:; : :j

grammatically in Figure 7, in which the critical points are represented as covering certain ranges
of temperature making it possible to show, graphically, the changes taking place within these
ranges. Taking an iron carbon alloy having, for instance, the composition a (some 0.20 per cent
carbon), the diagram shows that above Ar 3 it is made up of aa', i.e. of 100 per cent austenite; on cool-
ing through Ar 3 it is gradually converted into martensite; between Ar 3 and Ar 2 beta ferrite is

16 LESSON VIII THE THERMAL CRITICAL POINTS OF STEEL

liberatedj in passing through Ar 2 the remaining martensite is gradually converted into a solid solu-
tion of carbon and alpha iron (troostite?) while the free beta ferrite is converted into free alpha
ferrite; between Ar 2 and An additional alpha ferrite is liberated ; in cooling through An the remain-
ing solid solution (troostite?), now of eutectoid composition, is converted into pearlite. The struc-
tural changes occurring in steel having but one upper critical point and in steel of eutectoid
composition are similarly depicted. This diagram is reproduced from the "Journal of the Iron
and Steel Institute," No. IV for 1906, Plate LII.

Examination

Describe and discuss the theories that have been suggested to account for the
critical points of steel A 3 , A 2 , Ai, A 3 . 2 , A 3 .2.i, and Ac m .

LESSON IX

THE THERMAL CRITICAL POINTS OF IRON AND STEEL

THEIR EFFECTS

o
'It has been shown in preceding lessons that the thermal critical points of iron

and steel are due chiefly if not wholly to allotropic transformations of the iron. It is
a well-known fact that when a substance undergoes an allotropic transformation
many of its properties undergo likewise deep and sudden changes at the critical
temperatures. Color, crystallization, dilatation, conductivity both for heat and elec-
tricity, strength, ductility, hardness, specific gravity are properties frequently affected
as a body passes from one allotropic form to another. We should expect, therefore,
such changes to take place, as iron undergoes its allotropic transformations, if not in
all, at least in some of the above properties. And it is because such changes do take
place that a clear understanding of the occurrence and significance of the critical
points is of much practical importance to the iron and steel metallurgist.

CHANGES AT A 3

It has been shown that the point A 3 occurs in carbonless iron and in steel con-
taining less than some 0.35 per cent of carbon and that it is universally believed that
this point indicates an allotropic change, the iron passing from the gamma to the
beta condition on cooling at Ar 3 , and vice versa, from the beta to the gamma condi-
tion on heating at Acs. It should be borne in mind, however, as fully explained in
Lesson VIII that the general belief is that free ferrite only undergoes this change.
As the metal cools past its point Ar 3 the following abrupt changes in some of its
properties have been noted.

Dilatation. The metal, which above the point Ar 3 was contracting, as is the
general rule with all cooling bodies, on passing through the point Ar 3 undergoes sud-
denly a marked dilatation, amounting to over jcW of its length, immediately fol-
lowed again by a normal contraction. Such dilatation implies that the change of
gamma into beta iron takes place with augmentation of volume, or in other words
that gamma iron is denser, has a higher specific gravity than beta iron. The dilata-
tions occurring at Ar 3 in the case of steels containing respectively 0.05 and 0.15 per
cent carbon are shown graphically in Figure 1. On heating, at Ac 3 a spontaneous
contraction occurs of the same magnitude as the dilatation on cooling. Had we no
other evidence of an allotropic transformation of the iron 'at this critical tempera-
ture, this sudden dilatation taking place as it does in pure iron would justify our
belief in its existence.

Electrical Conductivity. Above the point A 3 the metal has an electrical resis-
tance some ten times greater than its resistance at ordinary temperature. As it cools
from a high temperature to the point Ar 3 there is but a feeble decrease of its electrical

2 LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL

resistance, but as soon as Ars is reached it begins abruptly and sharply to decrease
and keeps on decreasing at a normal rate to atmospheric temperature. At A 3 , there-
fore, we have a sudden and marked change in the variation of the electrical conduc-
tivity corresponding to a sharp break in the curve expressing the relation between
temperature and electrical resistance as shown in Figure 2. On heating, at Ac 3 the
opposite change takes place, that is, the electrical resistance quite suddenly ceases to
increase. So marked and sudden a change in a physical property is in itself a proof
of an allotropic transformation.

Crystallization. It has been shown in Lesson II that while gamma and beta
iron both crystallize in the cubic system (Osmond) octahedra are the prevailing form
of gamma iron while the cube is the crystalline form of beta iron, and that the trans-

^OO 4OO' 6OO <3OO /OOO" //OO'

Fig. 1. Dilatation curves of various carbon steels.

formation of gamma into beta iron includes a change in the planes of symmetry, at
least of carburized iron (Osmond). As already mentioned, however, the crystallo-
graphic differences between gamma -and beta iron are not such as to prove the exis-
tence of these two allotropic varieties of iron.

Hardness, Ductility, Strength. Evidences will be offered later to show that as
the metal passes through the point Ar 3 the iron becomes harder, stronger, and less
ductile; in other words that gamma iron is softer, more ductile, and weaker than
beta iron.

Dissolving Power for Carbon. Above the point A 3 iron possesses dissolving
power for carbon, while according to some writers it loses that power on passing
through Ars ; in other words gamma iron can dissolve carbon but beta iron is deprived
of that power. It does not, however, seem, by any means, proven that beta iron
cannot dissolve carbon, many authoritative workers holding the opposite view. This
question will be discussed at greater length in another lesson.

LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL 3

Structural Properties. It has been explained at length in Lesson VIII that
the point Ar 3 corresponds to an abrupt structural change, namely, the beginning of
the setting free of ferrite (Fig. 1, Lesson VIII).

Other Properties. Le Chatelier mentions a change in the variation of the
thermo-electric force and a sudden but slight variation in magnetic properties as
taking place at As.

CHANGES AT A 2

As a separate point A 2 occurs in carbonless iron and in steel containing less than
some 0.35 per cent of carbon. The changes of properties taking place at A 2 are not
generally as abrupt nor are they as marked as those occurring at the other critical
points. It is precisely because of this lack of sharpness and suddenness that some
metallographists, notably Le Chatelier, have questioned the accuracy of the gen-
erally accepted view that this point like A 3 indicates an allotropic transformation.
Careful consideration of the evidences at hand appear to show, however, that changes

O <?OO 4OO <SOO QOO /OOO
Fig. 2. Electrical resistance curves of iron and high carbon steel.

of properties do occur at A 2 sufficiently marked and sudden to warrant the classifica-
tion of this point as an allotropic one. The fact that these changes are more gradual
than at the other critical points is logically explained by Osmond on the ground that
beta and alpha iron are isomorphous, that is, capable of forming solid solutions and
that therefore the passage of one variety into the other must necessarily be gradual
as well as the variations of the properties of iron which the transformation implies.

Dilatation. According to Le Chatelier, to Charpy and Grenet, and to some
others, no dilatation takes place as the metal cools past the point Ar 2 and they see in
this an indication that A 2 is not an allotropic point. Osmond's reply is that the
curves obtained by Charpy and Grenet, for instance, do indicate a dilatation at Ar 2
which, however, the authors fail to notice because the transformation not being
sudden the dilatation likewise is gradual, whereas the authors were looking for sud-
den dilatations only. If according to Osmond iron does expand on passing through
the point Ar 2 the specific gravity of beta iron must be greater than that of alpha iron.

Magnetic Properties. Above the point A 2 steel is non-magnetic, it is not at-
tracted by a magnet, but in passing through Ar 2 it suddenly becomes strongly mag-
netic. Is not this abrupt and momentous change alone, in the magnetic properties of

LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL

a metal, sufficient proof of an allotropic transformation? It is true that magnetism
is not fully regained until a considerably lower temperature is reached, probably
some 550 deg. C., according to Osmond, but the fact remains that the greatest part of
the final magnetism of the metal is abruptly acquired as it cools past the point A2.
What transformation other than an allotropic one can satisfactorily account for this?

The relation between the carbon content of steel and the temperatures at which
the metal loses its magnetism on heating and regains it on cooling is shown graphically
in Figure 3. The points plotted in this diagram represent the average values of a great
number of determinations made by Madame Sklodowska Curie with a series of very
pure carbon steels. It will be noticed that the points of magnetic changes correspond
closely to the thermal critical points A 2 , A 3 . 2 , or A 3 . 2 .i. With little carbon there is but
a small gap between the appearance of magnetism on cooling and its disappearance on
heating, because the points Ar 2 and Ac 2 occur at nearly the same temperature; with
0.50 per cent carbon the magnetic points are lowered and so, likewise, the point A 3 . 2
while the gap increases, this being consistent with the greater gap between Ar 3 . 2 and
Ac 3 . 2 ; with 0.84 per cent carbon the magnetic points are further lowered and the gap
between them increased still more, this being in harmony with the location of the
point A 3 . 2 .i which is lower than A 3 . 2 and with the greater gap between Ar 3 . 2 .i and Ac 3 . 2 .i.

Further experimental evidences that the points of magnetic transformations coin-
cide with the thermal critical points are given in the following tables showing the re-
sults of several hundred determinations. All the steels used in connection with the
results given in Table II contained in the vicinity of one per cent carbon.

TABLE I. COMPARISON OF THE MAGNETIC METHOD WITH THE
ORDINARY OR COOLING-CURVE METHOD. (BOYLSTON.)

Ac s .j. t

APPROXIMATE
CARBON CONTENT
OF STEEL

METHOD

NUMBER OF
TESTS

MEAN

MAX.
MIN.

MKAX

MAX.

Mix.

Magnetic ....

773

{788
1759

708

{711
I 706

1.25%

Ordinary ....

764

{781
1754

718

{736

hoi

Ac,,..,

Ar 3 ,

Magnetic ....

780

{790
1770

741

{750

1734

0.40%

Ordinary ....

827

{830
1822

754

{756
1753

Ac 2

Ar,

Magnetic ....

764

{772
1755

764

(771
1758

0.15%

Ordinary ....

768

{772
1764

767

[783
1748

LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL 5
TABLE II. RESULTS OBTAINED BY STUDENTS AT HARVARD UNIVERSITY

STEEL
NUMBER

METHOD

Acj.,.,

NUMBER OF
TESTS

Ar 3 . 2 .,

NUMBER OF
TESTS

Magnetic ....

753

679

Ordinary ....

739

688

Magnetic ....

752

131

695

136

Ordinary ....

750

695

Magnetic ....

761

704

Ordinary ....

757

70S

Magnetic ....

762

7QQ_

Ordinary ....

751

701

Magnetic ....

776

720

Ordinary ....

760

727

Crystallization. The cube being the crystalline form both of beta and of alpha
iron and these two allotropic varieties being capable of dissolving each other in all

Fig. 3. Temperatures of magnetic transformations of various carbon steels.

proportions (Osmond) a crystalline transformation at the point A 2 is not to be ex-
pected. The crystallography of iron so far as it has been investigated does not re-
veal the existence of the point A 2 .

Hardness, Ductility, Strength. It will be shown in another lesson that as the
metal passes through the point Ar 2 it becomes softer, more ductile, and less tenacious,
in other words that beta iron is harder, stronger, and less ductile than alpha iron.

Dissolving Power for Carbon. Most metallographists believe that alpha iron
does not possess any dissolving power for carbon, or at least that it is only capable of
dissolving a very small amount of that element. On the other hand, as already men-
tioned, it is far from certain that beta iron is incapable of dissolving carbon, state-
ments to the contrary notwithstanding. It is possible, therefore, that on cooling
through Ar 2 , iron loses its dissolving power for carbon.

Structural Properties. By referring to Figure 1 of Lesson VIII it will be seen
that there is no apparent structural change connected with the point A 2 . As the steel
cools past Ar 2 the liberation of ferrite started at Ar 3 merely continues, to end only at
Ari. Of course the ferrite liberated above Ar 2 now passes from the beta to the alpha

6 LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL

condition but this allotropic transformation does not appear to include any struc-
tural change.

Other Properties. Goerens and Cavalier both mention a sudden decrease in the
specific heat of iron as taking place at Ac 2 , and vice versa a sudden increase at Ar 2 .

CHANGES AT A 3 . 2

It has been shown that the point A 3 . 2 resulting from the merging of A 3 and A 2
occurs, theoretically at least, in steels containing from some 0.35 to 0.85 per cent
carbon. As might be expected the changes of properties corresponding to the point
A 3 . 2 are the same as those taking place in lower carbon steel at A 3 and A 2 . As the
metal cools through Ar 3 . 2 the following variations of properties are, therefore, noted:
(1) a marked dilatation, (2) a sudden decrease of electrical resistance, (3) a gain of
magnetism, (4) a probable loss of dissolving power for carbon, (5) the beginning of
the liberation of alpha ferrite (see Lesson VIII, Fig. 2).

CHANGES AT AI

The point AI occurs in steel containing from a mere trace to 0.85 per cent carbon.
It corresponds, as explained in Lesson VIII, to the transformation of the residual
austenite (now of eutectoid composition) into pearlite. This formation of pearlite
implies that the iron contained in this residual austenite (and forming about 85 per
cent of its bulk) undergoes on cooling through Ari, the same allotropic changes as
those affecting the free ferrite on cooling through Ar 3 and Ar 2 (or Ar 3 . 2 ). It follows
from this that, theoretically at least, the following sudden changes of properties
should be noted on cooling through Ar^ (1) a dilatation increasing with the carbon
content and being maximum with 0.85 per cent carbon caused by the allotropic
transformation of the iron, (2) increased magnetism because of the transformation of
additional non-magnetic gamma iron into magnetic alpha iron, (3) decreased electrical
resistance because of additional transformation of high resistance gamma iron into
low resistance alpha iron, and (4) additional loss of dissolving power for carbon be-
cause of the formation of additional alpha iron.

Of the above changes the dilatation only has been conclusively shown to occur
and to increase with the carbon content (Fig. 1). The point AI has never, to the
author's knowledge, been connected with critical variations of the electrical and
magnetic properties of steel, but on purely theoretical ground he does not see how the
conclusion can be avoided that such critical variations must exist provided of course
that we are right in assuming that in austenite of eutectoid composition the iron is
still in the gamma condition, that is, non-magnetic and of high electrical resistance.
It is possible that if experiments were conducted with a view of detecting these vari-
ations, the results would confirm these theoretical deductions.

CHANGES AT A 3 . 2 .i

The point A 3 . 2 .i occurs in eutectoid and in hyper-eutectoid steel and marks the
transformation of the austenite of eutectoid steel or of the residual austenite of hyper-
eutectoid steel, into pearlite, as shown in Figures 3 and 4, Lesson VIII. In these
steels, however, no liberation of free ferrite occurs above the point Ar 3 . 2 .i from which
it follows that the totality of the iron undergoes its allotropic change or changes on
passing through Ar 3 . 2 .i. The variations of the properties which in hypo-eutectoid

LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL 7

steel occur at A 3 , A 2 , and AI must therefore, in the case of eutectoid and hyper-
eutectoid steel all take place at the point A 3 . 2 .i. These sudden changes of properties
are, on cooling: (1) a marked dilatation, maximum in eutectoid steel (Fig. 1), (2) a
sudden decrease of electrical resistance, (3) a sudden gain of magnetism, (4) a loss
of dissolving power for carbon.

Le Chatelier mentions the fact that on cooling through Ari or Ar 3 . 2 .i steel ac-
quires a temporary malleability. If a steel bar, for instance, of sufficient length be
held horizontally by one extremity while cooling, it at first remains rigid, but on
passing through its point of recalescence it quite suddenly bends.

CHANGES AT A cm

The point Ar cm occurs in hyper-eutectoid steel and marks the beginning of the
liberation of free cementite as the metal cools from Ar cm to Ar 3 . 2 .i (see Fig. 3, Les-
son VIII). Except for this structural change no other marked changes of properties
have so far been connected with this point.

Structural Change at A! and A 3 . 2 .i. The spontaneous transformation of aus-
tenite of eutectoid composition into pearlite, that is, of a solid solution into an aggre-
gate, at Ari or Ar 3 . 2 .i and the reverse transformation, from aggregate to solid solution,
at Aci or Ac 3 . 2 .i, imply structural changes of momentous importance to the steel
metallurgist. While these will be dealt with at length in another lesson it seems
proper to record here their significance. These structural changes give the key to the
rational treatment of steel. They make possible the refining of steel by heat treatment
seeing that on heating steel through its critical range we may change it from the
condition of a coarse aggregate (a coarse structure) to the condition of a fine, nearly
amorphous, solid solution. They also make possible the hardening of steel through
sudden cooling from above the critical range as will be fully explained in another lesson.

Prevailing Conditions Above and Below the Critical Range. The following
condensed statement of some of the most significant conditions prevailing above
and below the critical range of iron-carbon alloys may be useful in keeping these
fundamental facts in mind. By critical range is, of course, meant here the critical
points, both on heating and cooling, considered collectively and it will be understood
that the -conditions described as existing above the range change quite abruptly to
the conditions prevailing below the range as the metal cools through the range, or
vice versa as it is heated above it. The references made to the crystallizing of the
metal and to the influence of work both above and below the range will be under-
stood after reading the following lessons dealing with the treatment of steel.

CONDITIONS AND PROPERTIES OF IRON-CARBON ALLOYS AND OF THEIR CONSTITUENTS

Above Critical Range Below Critical Range

Solid solution (austenite). Aggregate (ferrite + cementite).

Hardening (dissolved) carbon. Cement carbon (Fe 3 C).

Gamma iron. Alpha iron.
Alloys containing a sufficient amount of Same alloys deprived of hardening power.

carbon possess hardening power.

Alloys are non-magnetic. Alloys are magnetic.

Metal crystallizes on slow cooling. Metal does not crystallize on slow cooling.

AVork prevents crystallization. Work distorts structure.

8 LESSON IX THE THERMAL CRITICAL POINTS OF IRON AND STEEL

Properties of Gamma, Beta, and Alpha Iron. The various properties of gamma,
beta, and alpha iron described in the preceding pages, have been tabulated below as
well as some other data of interest. These are in accordance with the views most
generally held, but the author is well aware that those entertaining different views
may take exception to some of the entries.

GAMMA IRON

BETA IRON

ALPHA IRON

Metallurgical name

Austenite

Beta ferrite

Ferrite, alpha ferrite,

pearlite ferrite

Solvent power for C (or

dissolves carbon up to

probably some but opin-

probably none, but

Fe 3 C)

1.7 per cent or Fe 3 C up

ions differ

opinions differ

to 25.5 per cent

Range of temperature

above A 3 , A 3 . 2 , or A 3 . 2 .i in

as free beta ferrite be-

below A 2 , Aa.2, or A 3 . 2 .i

in which stable

case of pro-eutectoid

tween As and A 2

ferrite, above Ai or

A 3 . 2 .i in case of eutec-

toid ferrite

System of crystalliza-

cubic (orthorhombic ac-

cubic

tion

cording to Le Chate-

lier)

Prevailing crystalline

octahedra

cubes

forms

Other crystalline char-

frequent twinnings

no twinnings

acteristics

Specific gravity

greater than beta and

greater than alpha iron

smaller than gamma

alpha iron (dilatation

(gradual dilatation at

and beta iron

at Ar 3 )

Ar 2 , Osmond)

Electric conductivity

ten times smaller than

greater than that of gam-

greater than that of

that of alpha iron at

ma iron and increasing

beta iron and increas-

ordinary temperature

with falling tempera-

ing with falling tem-

ture

perature

Magnetic properties

non-magnetic

feebly magnetic

strongly magnetic

Hardness

softer than beta iron,

very hard

soft

harder than alpha iron

Examination

I. Describe and discuss the change of properties occurring in steel containing
0.25 per cent carbon as it cools from a temperature exceeding its critical
range to ordinary temperature.

II. Describe the changes of properties taking place at Ar 3 2 1.

LESSON X

CAST STEEL

The structural and other changes taking place at the thermal critical points of
steel account for the deep changes of properties resulting from the treatments to
which steel is subjected in the process of manufacture of steel objects. We are now
in a position to understand these changes, to anticipate them, and to arrive at the
rationale of the treatment of steel which for so many centuries remained purely empir-
ical.

It is logical that we should first consider the structure of steel before it has re-
ceived any treatment whatsoever, namely, the structure of the metal in its cast con-
dition. To this study the present lesson will be devoted.

The structure of cast steel is different from what, in these lessons, has been termed
the normal structure of the metal because, having been developed during very slow
and undisturbed cooling from the molten condition, crystalline growth has been
promoted, whereas in working and reheating such large growth is hindered or cor-
rected. It may well be expected then that the structure of steel castings will be
coarser, as is generally expressed, that is, made up of larger crystalline grains, than
the normal structure so far considered, and therefore that steel castings will suffer
from all the ills that pertain to a coarse structure, namely, weakness, lack of ductility,
or even brittleness, etc. It will be profitable at the outset to consider in a general
way the genesis of the structure of cast eutectoid, hypo-eutectoid, and hyper-eutectoid
steel.

Structure of Cast Eutectoid Steel. Let us first look into the genesis of the struc-
ture of eutectoid steel in the cast condition. Above its melting-point this steel, like all
steels, consists of a liquid solution of carbon or of the carbide Fe 3 C in iron (see Fig. 4,
Lesson VIII). Upon solidifying this liquid solution is converted into a solid solution,
that is, the carbon or carbide remains dissolved in the iron, known now as gamma
iron. This solid solution of iron or carbide of iron in gamma iron is called austenite.
Solidification means crystallization: crystals or crystallites of austenite form during
the solidification and, as is usual, the slower the solidification the larger will the crys-
tals be. Osmond has shown that these crystals belong to the cubic system and that
they are chiefly octahedra (although this is doubted by Le Chatelier). These octa-
hedra of austenite continue to grow on slow cooling below the solidification, this growth
being effected through several adjacent crystalline grains taking the same orientation
and therefore merging into a larger crystalline grain. It is evident, therefore, that in
the process of making steel castings the slow and undisturbed cooling prevailing both
during and after solidification promotes the formation of large grains of austenite,
these grains being made up of small octahedric crystals, and that the larger the cast-
ings the larger generally the crystalline grains, that is, the coarser the structure.
When slowly and undisturbedly cooled eutectoid steel then reaches its single critical

2 LESSON X CAST STEEL

point, As.2.1, it is composed of relatively large grains of austenite. In passing through
this point the austenite grains are converted bodily into as many pearlite grains, as
explained in Lesson VIII, a coarse austenitic structure acquired at a high temperature
giving rise to a coarse pearlitic structure at ordinary temperature. The polyhedric
structure, therefore (Fig. 1), observed after complete cooling indicates the original poly-
hedric structure of austenite formed above the critical point. The meshes of the net-
work are sections through pearlite grains, the net merely boundary lines between such
grains, originally boundary lines between austenite grains. This polyhedric structure
of slowly cooled steel proves the polyhedric structure of austenite at a high tempera-
ture. Because of a coarser grain and coarser pearlite cast eutectoid steel is weaker
and less ductile than eutectoid steel properly worked or annealed or both.

Structure of Cast Hypo-Eutectoid Steel. Let us now consider the genesis of the
structure of hypo-eutectoid steel, and let us select as an example steel containing

Pig. 1. Eutectoid steel. Cast. Magnified
500 diameters. (Boylston.)

Fig. 2. Hypo-eutectoid steel. Cast. Free f en-it o
rejected chiefly to the boundaries. Magnified
100 diameters. (H. C. Cridland in the author's
laboratory.)

0.60 per cent carbon and, therefore, composed after complete slow cooling of 72 per
cent of pearlite and 28 per cent of free ferrite. The formation of the structure of
this steel has been depicted in Figure 2, Lesson VIII, to which the reader is referred.
This steel on solidifying passes, like all steels, from the condition of a liquid solution
of iron and carbon to that of a solid solution of carbon (or more probably Fe 3 C) in
gamma iron, this solid solution, or austenite, being made up of crystalline grains. The
austenite grains formed during solidification continue to grow as the steel cools slowly
to its upper critical point Ar 3 . 2 , when, as explained in Lesson VIII, ferrite begins to be
liberated and continues to be liberated as the metal cools to its lower point Aiv This
setting free of ferrite is apparently brought about by each grain of austenite rejecting
the ferrite in excess of the eutectoid composition, so that by the time the point Ari
is reached each residual grain of austenite has the eutectoid composition and on cooling
through Ari is converted bodily into a grain of pearlite. Microscopical examination
reveals the fact that the pro-eutectoid ferrite is rejected (1) to the boundaries of

LESSON X CAST STEEL 3

the decreasing austenitic grains and (2) between the cleavage or crystallographic
pianos of these crystalline grains, so that three types of structures may be distin-
guished in cast hypo-eutectoid steel, (a) structures in which the free (pro-eutectoid)
ferrite has been rejected chiefly to the boundaries of the austenitic grains (Fig. 2),
clearly indicating that these grains were polyhedric, (6) structures in which the free
ferrite has been rejected chiefly between the cleavage planes of austenite (Fig. 3),
proving the crystalline character of that constituent and suggesting as later explained
that its crystallization is cubic, and (c) structures in which the free ferrite has been
rejected partly to the grain boundaries and partly between the cleavage planes.
Long exposure to high temperatures followed by slow cooling _apj3ears to favor
the massing of free ferrite between crystallographic planes, whereas short exposure

Fig. 3. Hypo-eutectoid steel. Cast. Free ferrite
rejected chiefly between cleavage planes. Mag-
nified 100 diameters. (W. J. Burger, Corres-
pondence Course student.)

and more rapid cooling promotes the expulsion of free ferrite to the grain boundaries,
resulting in sharply defined network structures.

The structure of cast hypo-eutectoid steel is coarse (1) because its slow and un-
disturbed cooling promotes the formation of large austenite grains and hence, later,
of large pearlite grains, (2) because its slow cooling between the upper and lower
critical points favors the rejection of a maximum amount of free ferrite which rejec-
tion makes for coarseness of structure, and (3) because its slow cooling from the upper
critical point to atmospheric temperature promotes the crystallization of free ferrite
into large grains, this influence, however, being material only where there is a large
amount of free ferrite, i.e. in very low carbon steel.

Because of its coarser structure cast hypo-eutectoid steel is less tenacious and less
ductile than forged or properly annealed steel of similar composition.

Structure of Cast Eutectoid vs. Structure of Cast Hypo-Eutectoid Steel.
Although the pearlite grains of eutectoid steel may be and often are larger than the
pearlite grains of hypo-eutectoid steel, the latter, especially when judged by its frac-

4 LESSON X CAST STEEL

ture, is the coarser of the two. This greater coarseness of hypo-eutectoid steel in
spite of smaller pearlite grains is due to the presence of free ferrite, relatively small
pearlite grains surrounded by coarse ferrite envelopes or holding coarse ferrite particles
imparting a coarse appearance to the fracture of steel. The dimension of the pearlite
grains, therefore, while not without influence, is not the criterion by which to judge
of the coarseness or fineness of the structure and fracture of hypo-eutectoid steel, the
amount of free ferrite present and its mode of distribution having to be taken into
consideration. In very low carbon steel, moreover, there is but little pearlite and
the small amount present occurs as small irregular particles (Fig. 4) exerting but little
influence upon the character of the fracture which now depends quite exclusively upon
the dimension of the ferrite grains. As ferrite grains, however, no matter how small
never impart as fine a structure or fracture to steel as pearlite grains, it follows that

Fig. 4. Hypo-eutectoid steel. Cast. Carbon 0.20%.
Magnified 28 diameters. (Boylston.)

low carbon (ferritic?) steels can never have as fine a structure or fracture as higher
carbon (pearlitic) steels.

Structure of Cast Hyper-Eutectoid Steel. The genesis of the structure of cast
hyper-eutectoid steel has been depicted diagrammatically in Figure 3, Lesson VIII.
Between its solidification point and its upper critical point (A cm ) this steel is com-
posed, like all steels, of crystalline austenite grains formed on solidifying and subse-
quent slow cooling. Upon reaching the point Ar cm the setting free of cementite
begins, ending only at the lower point Ar 3 . 2 1. This free cementite, like the free ferrite
of hypo-eutectoid steel, is rejected (1) to the boundaries of the diminishing austenite
grains and (2) between the cleavage planes of this crystalline austenite, giving rise
to the three types of structure described in the case of hypo-eutectoid steel, but in
which free ferrite is replaced by free cementite, namely, (a) structures in which the free
cementite is found chiefly at the grain boundaries, (6) structures in which the free
cementite is chiefly located between cleavage planes, and (c) structures in which the
free cementite is partly at the boundaries and partly between crystallographic

LESSON X CAST STEEL 5

planes (Fig. 5). Like the structure of cast hypo-eutectoid steel, the structure of
cast hyper-eutectoid steel bears witness (1) to the polyhedric form of the austenite
grains, (2) to the crystalline character of these grains, and (3) to their probable cubic
crystallization.

Long exposure to high temperatures followed by very slow cooling promotes in
hyper-eutectoid steel the rejection of cementite to the cleavage planes, while short
exposure and more rapid cooling favor the rejection of cementite to the boundaries.

The rejection of free cementite like the rejection of free ferrite makes for coarseness
of structure and fracture, from which it follows that cast and slowly cooled hyper-
eutectoid steel will be coarser than cast and slowly cooled eutectoid steel, and that
the more free cementite it contains, that is, the higher the carbon, the coarser it will
be. The structure and fracture of hyper-eutectoid steel, however, will generally be

Fig. 5. Hyper-eutectoid steel. Cast. Free cementite rejected
partly to the boundaries and partly between cleavage planes.
Magnified 114 diameters. (Boylston.)

decidedly finer than that of hypo-eutectoid steel because of the very small amount of
free cementite present in the former compared to the amount of free ferrite in the latter,
unless, indeed, the hypo-eutectoid steel be very near the eutectoid composition. This
is due to the fact, now well understood, that starting from the eutectoid composition
(carbon 0.85 per cent), as the carbon decreases the amount of free ferrite increases
rapidly, while as the carbon increases above the eutectoid ratio the amount of free
cementite increases slowly and remains small even with high carbon content. Steel
with 0.50 per cent carbon, for instance, contains 40 per cent of free ferrite, hence
its coarseness both of structure and fracture, while steel with say 1.25 per cent carbon
contains but 6.4 per cent of coarsening free cementite, hence the relative fineness of
both its structure and fracture.

Ingotism. Howe has suggested the term "ingotism" to designate the structure
of cast steel described in the foregoing pages and characterized (1) by large pearlite
grains and (2) by coarse ferrite or cementite membranes surrounding them and by
irregular masses of these constituents located in some of the cleavage planes of the
original austenite grains.

6 LESSON X CAST STEEL

Structure of Cast Steel vs. Structure of Meteorites. The structure of meteo-
rites thro^ 'additional light upon the mechanism of the formation of the structure
of cast stei -,*, outlined in the foregoing pages. The most characteristic feature of the
structure o> hypo- and hyper-eutectoid cast steel, namely, the coarse massing of free
ferrite or ceuentite at the boundaries of the austenite grains or between the cleavage
planes of the crystalline austenite is exhibited most strikingly in the structure of
some meteorites known as the Widmanstatten structure. This structure, however,
is on a much larger scale than the similar structure of steel castings; which should

Fig. 6. Steel. Carbon 0.55%. Widmanstatten structure. Magnified 6
diameters. (Belaiew.)

not be a source of surprise when it is considered that the conditions required for the
formation of such structure are greatly exaggerated and intensified during the cool-
ing of meteorites, namely, (1) a very slow solidification peroid, (2) very long exposure
to high temperatures, and (3) very slow cooling from these high temperatures.

Belaiew succeeded in a remarkable manner in reproducing the Widmanstatten
structure by subjecting carbon steels to a high temperature for a very long time and
cooling them extremely slowly, the fall of temperature from 1500 to 300 deg. C. last-
ing 60 hours, an evident attempt at reproducing the conditions which must prevail
during the solidification and further cooling of meteorites. The structures obtained
by Belaiew in the case of steel containing 0.55 per cent carbon and otherwise of com-

LESSON X CAST STEEL 7

mercial quality are shown in Figures 6, 8, 10, 12, and 13. They ar >ical struc-
tures of steel castings of the same grade but on a much larger scale should be
noted that the magnification of Figures 8, 10, 12, and 13 is only 30 eters while
Figure 6 is magnified but 6 diameters. Figure 6 is a beautiful illus tion of that

Fig. 7. Section parallel to the surface
of a cube. (Tschermak).

Fig. 8. Steel. Carbon 0.55%. Section parallel to the
surface of a cube. Magnified 30 diameters. (Belaiew.)

type of structure in which the free ferrite has been rejected both to the grain boun-
daries forming a sharply outlined network and between crystallographic planes.
In Figures 8, 10, 12, and 13 the free ferrite is seen massed between cleavage planes.

Octahedric Crystallization of Austenite. It will be noted that the ferrite bands shown in Figures
8, 10, 12, and 13 cut each other at right angles or, more frequently, form equilateral triangles. Ac-
cording to crystallographers these are indications that austenite crystallizes in regular octahedra.
That this inference is correct appears to be conclusively proven by the following remarks of Belaiew.

LESSON X CAST STEEL

" Let us consider an octahedron and let us assume that four systems of lamellae locate themselves
in this octahedron along its crystallographic planes, that is, parallel to the four pairs of its surfaces,
a fact that has been long known in the case of meteoric irons.

"If we now examine any section of the octahedron, we shall find that not only the angles formed
liy the projections of the lamellae vary in different sections, but that the number itself of different

Fig. 9. Section parallel to the surface
of an octahedron. (Tschermak.)

Fig. 10. Steel. Carbon 0.55%. Section parallel to tho sur-
face of an octahedron. Magnified 30 diameters. (Belaiew.)

sections varies likewise from two to four. For instance, when the section is parallel to the surface
of the cube, the number of different sections is minimum, that is, two, and in the entire section we
find only two systems of lamellae forming right angles. Figure 7 is a diagram of such section
and Figure 8 a corresponding section of the steel.

" A section parallel to one of the surfaces of the octahedron will yield equilateral t riangles formed
by three systems of lamellae forming 60 angles; the fourth system coincides with the section con-
sidered (see Figs. 9 and 10).

LESSON X CAST STEEL

" In a section parallel to the surface of the dodecahedron, two systems of lamella; are observed
forming an angle of 109 28' 16"; the other two systems coincide and divide this angle in half
(Pigs. 11 and 12). Finally any section will give four different systems cutting each other at differ-
ent angles (Fig. 13).

"All these cases, as we have just seen, can very well be illustrated by different samples of our

Fig. 11. Section parallel to the surface
of a dodecahedron. Magnified 30 di-
ameters. (Tschermak.)

Fig. 12. Steel. Carbon 0.55%. Section parallel to the sur-
face of a dodecahedron. Magnified 30 diameters. (Belaiew.)

alloy which, firstly, affords a rather weighty proof of the octahedric crystallization of steel and, sec-
ondly, brings out the remarkable analogy of this structure with that of meteorites and warrants us
to allude to the synthesis of the structure so called of Widmanstatten . . . this structure is the neces-
sary consequence of the uniform orientation of the elementary octahedra within a volume of
greater or less dimension; it is, therefore, in no way related with the carbon content and must be

LESSON X CAST STEEL

obtained in any alloy of iron and carbon whenever the conditions are favorable to the formation of
that structure.

"Moreover, in practice this structure is met (although certainly much less developed than in
our alloys) every time that the metal is subjected to an intense heating followed by slow cooling as
is the case with cast steel or, better still, with burnt or overheated steel.

"The very brittleness of these steels may be due to a certain extent to the uniform orientation
of the elementary octahedra which are the cause of that structure."

Let it be noted that in the case of meteorites the length of time during which the metal is main-
tained at a high temperature is so long that generally but one crystal is formed, that is, all the
elementary octahedra formed on solidification have assumed the same orientation. In steel the
conditions being less favorable to uniformity of orientation we have several grains.

/*/.,

Fig. 13. Steel. Carbon 0.55%. Four systems of lamellae.
Magnified 30 diameters. (Belaiew.)

Experiments

The student should procure samples of cast steel containing the following pro-
portions of carbon: from 0.20 to 0.50 per cent, from 0.70 to 0.90 per cent, and 1.25
or more per cent. These should be prepared for microscopical examination in the
usual way, and examined both with low and high powers. They should be photo-
graphed under a magnification not exceeding 100 diameters as the aim should be to
bring out the structural characteristics described in this lesson, for which purpose
a high magnification is not necessary or, indeed, desirable. In the case of the
hypo-eutectoid steel several specimens may be prepared and examined to illustrate
the different aspects which cast steel of that grade may assume as explained in
the lesson.

Examination

Describe briefly the genesis of the structure of (1) cast hypo-eutectoid steel, (2)
cast eutectoid steel, and (3) cast hyper-eutectoid steel.

LESSON XI
THE MECHANICAL TREATMENT OF STEEL

Forged steel objects are manufactured by subjecting the cast metal (1) to pres-
sure exerted by rolls, presses, or dies, or to blows from hammers and (2) by reheating
it to various temperatures for various lengths of time and cooling TE at various rates.
In other words we have to consider (1) the mechanical treatment of steel and (2) its
heat or thermal treatment. The machining of steel is not here mentioned since it
can evidently have no effect upon the structure and properties of the metal, unless it
be a very superficial one.

While the primary purpose of working steel is to shape it into useful appliances
and while the primary purpose of reheating it may be, and often is, to impart to the
metal such plasticity as will facilitate its being so shaped, both mechanical and heat
treatments deeply affect the structure of steel and therefore its physical properties.
To this very important subject the next four lessons will be devoted. We shall first
consider the influence of mechanical treatment and then that of heat treatment.

The effect of work upon the structure and properties of steel greatly depends
upon the temperature of the metal while it is being worked, the expressions "hot work-
ing" and "cold working" being common ones, the former meaning working the steel
while hot, and the second working it while cold, more specifically at atmospheric
temperature. In these lessons the expression hot working will be applied to working
the metal while above its critical range, and cold working to work performed below
that range. In justification of this course it will be shown that the effect of work
changes quite sharply as the critical temperature of steel is passed.

Hot Working. Hot working may be applied to steel (1) after it has solidified
but before it has cooled to a much lower temperature so that it still possesses the
necessary plasticity, or (2) the steel ingot may be allowed to cool to atmospheric tem-
perature or at least to a temperature so low that reheating is required as a preliminary
step to successful hot working. The following considerations will show that so far as
the influence of hot work is concerned it is quite immaterial whether the steel ingot
has or has not been completely cooled before being brought to the forging tempera-
ture. Let us assume that the steel ingot be allowed to cool to atmospheric tempera-
ture, that it is then reheated to a temperature well above its critical range and then
subjected to hot working.

An attempt has been made in Figure 1 to depict graphically the influence of hot
work on the structure of steel. The diagram will be readily understood. The critical
range both on heating and cooling is represented by a double line, but the reader
will, of course, bear in mind that the critical range on heating does not coincide with
the range on cooling and that each range may include one, two, or three critical points,
as fully explained in previous lessons. For the present purpose it is preferable to
represent both ranges irrespective of the number of critical points included by the
two parallel lines shown in the diagram. The solidification range is likewise indi-
cated by a double line. The widths of the shaded areas are intended to be propor-

LESSON XI THE MECHANICAL TREATMENT OF STEEL

+-
t
h
+

_J-

a
o

&' 1
Q -p

|!|

S>!1

~6 (J ^
l" r

| t b,

U S^

I* It

J|ft

^^ vStf cS rtivO-r-

xi s 5: 1 <8 ^^o

LESSON XI THE MECHANICAL TREATMENT OF STEEL 3

tional to the grain size resulting from the various treatments; when an area is reduced
to a mere line the corresponding grain is very small, i.e. the structure very fine. As
the steel ingot solidifies at A crystalline grains are formed which increase in size on
slow and undisturbed cooling to the critical range B. In passing through that range
the austenite grains are converted into pearlitic grains with or without rejection of
free ferrite or of free cementite according to the carbon content of the steel. On
cooling from B to C there is no further growth and the size of the final grain may be
represented by the width of the shaded area at C. The metal has now the usual
coarse structure of steel castings described in Lesson X. Upon reheating this coarsely
crystalline steel ingot from S to R and through its critical range, i is. converted from
the condition of an aggregate of ferrite and cementite into a nearly amorphous solid
solution, so that as the metal emerges from its range it has a very fine structure.
As steel just above its critical range, however, would not be plastic enough for hot
working and would not afford the necessary range of temperature through which to
cool while being worked and still remain, as it should, above the range, it is generally
necessary to heat the metal to a much higher temperature, say to some 1200 deg. C.
or even more (W in Fig. 1). As the metal is heated from its critical range to this
high temperature, it probably crystallizes so that when work begins at w it is in a
somewhat crystalline condition. This grain growth on heating from R to W is de-
picted in the diagram. The heavy pressure or blows which are now applied, how-
ever, soon break up this crystallization while preventing a new one from forming, so
long at least as the work continues sufficiently vigorous, for undisturbed cooling is a
condition necessary for the ready growth of crystals. As soon as work ceases, how-
ever, the metal being now left to cool undisturbedly, if its temperature is then above
its critical range, for instance at /, it will immediately begin to crystallize and its
crystallization will continue from / to r, that is, until its critical range is reached, when
it will stop and the austenite grains just formed will be converted into as many pearl-
ite grains with or without rejection of free ferrite or free cementite. The width of
the band at s is intended to represent the size of the final grain. It is smaller than the
grain of the steel ingot but it is still large.

If work ceases, then, while the metal is still above its critical range it is evident
that it will crystallize and that the resulting grains will be the larger, that is, its
structure will be the coarser, the higher the temperature at which the work was
stopped.

Finishing Temperatures. The temperatures at which work ceases are known as
the finishing temperatures and the above considerations will show the importance of
proper finishing temperatures if it be desired to impart to steel implements the finest
grain that it can acquire through working together with the desirable properties
inherent to it.

Should hot work be continued, for instance, until the temperature of the metal
is but slightly above its critical range, /' in Figure 1, the austenitic grains that will
form on undisturbed cooling to the range will be small, so that the final grain of the
metal at .s' will likewise be small, as indicated by the width of the shaded area. If
the finishing temperature is exactly at the critical range, /", the final grain at s" will
be very fine. If working be continued until the metal has cooled to a temperature
lower than its critical range, for instance /'" in the diagram, the structure will be
fine since crystallization was prevented while the metal was cooling above its critical
range, but it will be distorted because the effect of working below the range, that is,

4 LESSON XI THE MECHANICAL TREATMENT OF STEEL

of cold working, is to distort the structure as explained in the following pages. A
distorted structure in turn means decreased ductility and eventually brittleness. It
seems evident, therefore, that hot worked steel implements should be finished exactly
at their critical ranges or at temperatures but slightly superior or inferior to them.

Structure of Hot Worked Eutectoid Steel. If eutectoid steel be subjected to
hot work until it has nearly reached its critical range (now consisting of the single
point Ars.2.i), /" in Figure 1, it will have a very fine austenitic structure which in
cooling slowly and undisturbedly through the critical range at r" will be converted
into a fine pearlitic structure. The metal will then have as fine a structure as can
be imparted to it by work alone, followed by slow cooling through the range.

If eutectoid steel be worked until its temperature is still considerably above its
critical point (/, Fig. 1), and then allowed to cool undisturbedly, austenite grains begin

Fig. 2. Hot worked hypo-eutectoid steel.
Carbon 0.50%. Finishing temperature near
critical range. Magnified 100 diameters.
(Burger, Correspondence Course student.)

to form and increase in size as the metal cools to its critical point, i.e. from f to r
when these austenitic grains are converted into as many pearlitic grains. The final
grain size therefore will depend upon the temperature at which work ceased and will
be the greater the higher that temperature.

Structure of Hot Worked Hypo-Eutectoid Steel. If hypo-eutectoid steel be
worked until its temperature is but very slightly above its critical range (/", Fig. 1)
and then allowed to cool undisturbedly, small grains of austenite are formed which
on passing through the critical range at r" are converted into pearlite grains with
rejection of excess free ferrite as now well understood. Because of the small size of
the austenite grains and because of relatively quick cooling (in air) the rejected fer-
rite generally locates itself at the boundaries of the grains and network structures
are produced (see Fig. 2). In the case of low carbon steel, however, containing say
less than 0.30 per cent carbon the proportion of free ferrite is so large, i.e. the ferrite
net so thick, that the structure consists of particles of pearlite embedded in a matrix
of ferrite (see Fig. 3).

LESSON XI THE MECHANICAL TREATMENT OF STEEL 5

If tho working of hypo-eutectoid steel ceases at a temperature considerably above
its critical range, / in Figure 1, the austenite grains formed on cooling to the range
be relatively large and therefore also the pearlite grains of the completely cooled

Fig. 3. Hot worked hypo-eutectoid steel. Carbon
0.05%. Magnified 114 diameters. (Boylston.)

Fig. 4. Hot worked hypo-eutectoid steel. Carbon
0.50%. Finishing temperature considerably above
the critical range. Magnified 56 diameters.

metal. The free ferrite will still be found chiefly at the boundaries (Fig. 4), the
meshes of the network structure, however, being larger than in similar steel finished
at a lower temperature, as will be apparent from a comparison of Figures 2 and 4.

6 LESSON XI THE MECHANICAL TREATMENT OF STEEL

It will be seen 'that in forged hypo-eutectoid steel forming network structures on
slow cooling through the critical range a close relation must exist between the size of
the meshes and the finishing temperature.

Structure of Hot Worked Hyper-Eutectoid Steel. If hyper-eutectoid steel be
hot worked until its temperature is very near its critical range (/", Fig. 1) and
then allowed to cool undisturbedly, small austenite grains are formed which on
cooling through the range are converted into small pearlite grains with rejection of
free cementite, as shown in Figure 5. If the work be stopped at a temperature con-
siderably above the critical range (/, Fig. 1), the final pearlite grains will be larger
while the free cementite will be located chiefly at the grain boundaries, a network
structure being produced. It will be evident that a close relation exists between

Fig. 5. Hot worked hyper-eutectoid steel. Carbon
1.50%. Finishing temperature near critical range.
Magnified 100 diameters. (Reinhardt in the
author's laboratory.)

the size of the meshes of these network structures and the corresponding finishing
temperatures.

Sorbite. High magnification of the structure of the meshes of both hypo- and
hyper-eutectoid steel described in the foregoing pages often fails to reveal the char-
acteristic features of pearlite, namely, (1) sharply defined parallel plates alternately
of ferrite and cementite and (2) a constant or nearly constant carbon content. The
structure of these meshes instead remains indistinct and presents a granular rather
than a lamellar aspect (Fig. 6). It is also frequently noted in connection with these
network structures that the full amount of free ferrite or of free cementite has not
been rejected, the pearlitic grains having retained some of the constituent in excess of
the eutectoid ratio. To this imperfectly developed pearlite the name of sorbite has
been given (Osmond) and quite universally adopted in spite of recent and regrettable
efforts to eliminate it from metallographic nomenclature. It will be apparent that
the formation of sorbite results from a relatively quick cooling through the critical

LESSON XI THE MECHANICAL TREATMENT OF STEEL 7

range, time being denied for the crystallization of distinct lamellae of ferrite and
cementite, and, in the cases of hypo- and hyper-eutectoid steel, for the rejection of
the full amount of free ferrite or free cementite. The cooling in air of hot worked
pieces, especially when of small size, is often sufficiently rapid to cause the formation
of sorbite rather than of pearlite.

The production and nature of sorbite will be dealt with at greater length in an-
other lesson. It should be mentioned here, however, that while sorbite is less ductile
than pearlite it has a higher tenacity, higher elastic limit, and greater hardness (hence
greater wearing power). When these qualities are required, in hot forged objects,
they may consequently be obtained, although necessarily at the~ sacrifice of some

Fig. 6. Hypo-eutectoid steel. Carbon 0.50%. The ill-defined constituent is
sorbite. Magnified 1000 diameters. (Boynton.)

ductility and softness, by hastening the cooling through the critical range, after work
has ceased, when sorbitic rather than pearlitic steel will be produced.

Hot Working of Steel vs. Its Critical Range. In conducting the hot working of
steel so as to impart to the metal the finest grain that can result from finishing at
suitable temperatures, it is generally necessary only to consider its lower critical
point on cooling, namely, Ari or Ars.a-i. The following considerations will justify
this statement. In the case of hypo-eutectoid steel if the working be continued while
the metal cools from its upper point or points to its lower point, it will make for fine-
ness of structure by preventing a coarse massing of the free ferrite while the cold
working of that constituent, if taking place at all, must be very slight. The same
reasoning applies, with greater force, to the hot working of hyper-eutectoid steel from
its upper point (denoting the formation of free cementite) to its lower point Ar 3 . 2 .i.
Greater fineness of structure will result with very little, if any, distortion of the free
cementite.

It is apparent, therefore, that in order to secure the finest grain obtainable through

8 LESSON XI THE MECHANICAL TREATMENT OF STEEL

mechanical refining without appreciable structural distortion steel objects should be
finished slightly above their lower critical point, that is, in the vicinity of 700 deg. C.
for all grades of commercial carbon steel.

Cold Working. By the cold working of steel is meant in these pages the work-
ing of it while its temperature is below its critical range. It will now be shown that
the effect of cold working upon the properties of the metal is very different from that
of hot working. This should not be a cause for surprise if it be borne in mind that
steel above its critical range is in a condition totally different from its condition
below it. Above the critical range we have to deal with a solid solution of iron and
carbon, below it with an aggregate of ferrite and cementite. The solid solution exist-
ing above the range will crystallize if allowed to cool undisturbedly and it has been
shown in the foregoing pages that working in this range, i.e. hot working, is effective

Fig. 7. Cold worked hypo-eutectoid steel.
Carbon 0.30%. Magnified 100 diameters.
(Burger, Correspondence Course student.)

in preventing or at least retarding this crystallization, thus imparting a smaller grain
to the metal. The aggregate of ferrite and cementite existing below the range, on
the contrary, exhibits no tendency to crystallize during slow and undisturbed cooling,
because this aggregate was formed and fully developed while passing through the
range, the size of its elements, that is, its coarseness depending (1) upon the coarse-
ness of the solid solution immediately before its transformation and (2) upon the
time occupied in cooling through the range. Working this aggregate, therefore, as it
cools to atmospheric temperature, or working it while at atmospheric temperature,
i.e. cold working steel, does not prevent its crystallization. Its effect consists in dis-
torting the existing aggregate structure, chiefly through the stretching or elongation
of its crystalline elements (free ferrite, free cementite, pearlite) in the direction of the
forging, and such distortion in turn means decreased ductility and eventually extreme
brittleness. The effect of cold working upon the structure of steel is illustrated in
Figures 7 and 8 in the case of hypo-eutectoid steel. It is also depicted in the diagram
of Figure 1. While the structural distortion caused by cold working is very slight
near the critical range of the metal, it rapidly increases as the temperature decreases,

LESSON XI THE MECHANICAL TREATMENT OF STEEL

becoming very pronounced at atmospheric temperature. The manufacture of wire by
cold drawing affords a familiar instance of the effect of work performed at atmos-
pheric temperature both on the structure and properties of the metal. It is well
known that after the wire has been passed through several dies it becomes so brittle
that annealing is necessary in order to make further reduction in size possible, the
annealing operation removing the structural distortion and brittleness produced by
working at atmospheric temperature.

Mechanical Refining. It would seem as if with the use of pyrometers at least it
should be a relatively simple matter to finish steel objects very near the desirable

Fig. 8. Cold worked hypo-eutectoid steel. Carbon 0.30%. Magnified
150 diameters. (Buck, Correspondence Course student.)

temperature and thus secure for them the best structure that can be imparted by
work alone. Upon reflection, however, it will be manifest that the problem is on the
contrary an insoluble one, for the reason that unless the objects are of very small
cross-sections, it is quite impossible to finish them so that their temperature will be
uniform throughout, the central portions being necessarily hotter than the outside.
Should the forging be so conducted that the temperature of the outside be very near
the critical range, the center, being materially hotter, will coarsen on cooling, while if
the implements, on the contrary, are finished so that their center may have the fine
structure produced by ceasing the work at the proper temperature, their outside must
necessarily suffer from cold working. The limitations of work alone as a means of
imparting the best possible structure to steel are therefore quite evident.

While a uniformly fine grain cannot be imparted to steel objects of considerable
size through hot working alone, the value of hot work as a means of refining the
structure of steel remains very great as exemplified by the structure of properly hot
forged steel when compared with the structure of steel castings of similar composi-
tion. The finer grain imparted to steel by working it has been called by some writers

10 LESSON XI THE MECHANICAL TREATMENT OF STEEL

"mechanical" refining to distinguish it from the refining produced by heat, i.e. from
"thermal" refining. In practise hot work should be so conducted, that is, the finish-
ing temperatures so regulated, that the central portions of the finished implements
will not suffer unduly from the coarsening influence of too high a finishing tempera-
ture, while at the same time the outside will not suffer unduly from the effect of cold
working. The natural tendency of rolling and other forging mills is to finish work
at too high temperatures for the simple reason that the metal is then more plastic
and consequently requires less power for its working. In some manufactures, how-
ever, and, especially in the rolling of rails, the importance of proper finishing tempera-
tures has been given careful attention and the rolling operation so modified as to
deliver rails of much finer grain and therefore better physical quality, than formerly.
Besides its important grain refining influence hot work further improves the quality
of steel by closing and, if the carbon be low enough, welding, the blow-holes and
otherwise increasing its soundness and by removing cooling strains.

Experiments

The student should prepare for microscopical examination samples of hot forged
hypo-eutectoid, eutectoid, and hyper-eutectoid steel. Their structure should be com-
pared with the structure of similar steels in the cast condition (Lesson X) and also
with the structure of similar normalized steels (Lessons IV and V). If the forging of
these samples was finished at a fairly high temperature their structure should be
quite similar to the normal structure of like steels described in Lessons IV and V.
Finer structures point to lower finishing temperatures.

A sample of cold worked steel preferably the longitudinal section of an unannealed
cold drawn wire should likewise be prepared and its structural distortion noted.

If a cross-section of a rail or of some other rolled shape can be obtained, a piece
should be cut from the center and one near the edge (advisably also from the web
and extremity of flange in the case of a rail section). These pieces should be pre-
pared and examined and the coarser grain of the central portion noted.

All specimens should be photographed preferably under a magnification not ex-
ceeding 100 diameters.

Examination

Describe briefly the effect upon the structure of steel (1) of working above the
critical range (hot work) and (2) of working below that range (cold work).

LESSON XII

THE ANNEALING OF STEEL

Purpose of Annealing. The purpose of annealing steel may be (1) to increase
its softness and ductility that it may, for instance, be more easily- machined or (2) to
secure a desirable combination of high strength and elastic limit with fair ductility
that it may successfully stand the strains to which it is to be subjected. These changes
of physical properties result from corresponding changes in the structure of the metal
brought about by proper heat treatment. In annealing steel it is generally intended
to impart to it as fine a structure, that is as small a grain, as is consistent with the
nature of the treatment and the grade of the steel. Hot forged steel objects maybe
improved by annealing for certain purposes, because of their structure being often (1)
relatively coarse owing to high finishing temperature and (2) heterogeneous as ex-
plained in Lesson XI. The structure of cold worked steel, at least when severely cold
worked, is so defective that the metal must be annealed before it can be put to useful
purposes. Finally, steel castings have so coarse a structure as to be very deficient both
in strength and ductility and should always be refined by annealing.

Nature of the Annealing Operation. The annealing operation comprises three
distinct steps: (1) heating the steel, (2) keeping its temperature constant at the an-
nealing temperature, and (3) cooling it from the annealing to atmospheric tempera-
ture. These steps will be considered separately.

Heating for Annealing. The first step in annealing always consists in heating
the metal past its critical range because by so doing the preexisting structure, how-
ever coarse, is obliterated; the metal, for the time being, assuming a nearly amorphous
structure. This important structural change is due, as we now understand it, to the
passage of the steel from the state of an aggregate of ferrite and cementite to that of
a homogeneous solid solution, and it is not to be wondered at that so radical a struc-
tural change should destroy effectively any preexisting crystallization. The annealing
of steel castings, however, constitutes an apparent exception to the rule that heating
just through the range is sufficient to break up effectively the preexisting structure,
for their successful annealing often requires a materially higher temperature. Should
the temperature of the steel remain below its critical range, no structural change
would take place and the annealing would be ineffective (I, Fig. I). 1 Should, on the
contrary, the temperature of the steel be carried considerably above the range, its
structure, which was finest as it emerged from the range, begins to coarsen on further
heating and continues to grow as the metal cools slowly to the range, so that its final
structure would be at least relatively coarse (II, Fig. 1). Clearly, therefore, to an-

1 AYhen steel contains hardening carbon it may be softened and made more ductile by heating
it to temperatures lower than its critical range (as in the tempering of hardened steel) but such treat-
ment is not, or at least should not be, called annealing. Cooling strains may also be removed, at
least in part, by heating below the range.

LESSON XII THE ANNEALING OF STEEL

neal steel forgings they should be heated through their critical range and kept at a
temperature as close to the upper part of that range as possible (III, Fig. 1). The
annealing temperature will, of course, vary with the carbon content since the position
of the critical range, or rather its width, varies likewise. The following ranges of tem-
peratures are recommended by the Committee on Heat Treatment of the American
Society for Testing Materials. The report of the Committee states that for steels
containing more than 0.75 per cent manganese slightly lower temperatures suffice.

RANGE OF CARBON CONTENT
Less than 0.12 per cent
0.12 to 0.25 per cent
0.30 to 0.49 per cent
0.50 to 1.00 per cent

RANGE OF ANNEALING TEMPERATURE
875 to 925 deg. C. (1607-1697 deg. F.)
840 to 870 deg. C. (1544-1598 deg. F.)
815 to 840 deg. C. (1499-1.544 deg. F.)
790 to 815 deg. C. (1454-1499 deg. F.)

sr
r\

Cr/r/co/

Rnnnf*

\ V

^ v Vi

8 V v

. \% \n

O \^ \o

Q ^

\ V"

IK (-,

i O

\ \

Structure Coarse OAF Very f/ne structure

unchanged sfruc/vre /~/ne structure Strong, e/ast/c, and tough

O Hardest, strongest, one/ /east ducft/e

F Softest, w&okest, ond most ducf//e

Fig. 1. Diagram depicting the annealing of steel.

Time at Annealing Temperature. The steel object should be kept at the anneal-
ing temperature long enough to be heated right through to that temperature. The
Committee on Heat Treatment, referred to above, states that an exposure of one hour
should be long enough for pieces twelve inches thick. Thicker pieces, of course, need
a longer heating.

The usefulness of pyrometers in conducting annealing operations is obvious.
Their use is to be strongly recommended.

Cooling from Annealing Temperature. Having imparted a fine structure to the
steel the next step must be to retain it. The most effective way of accomplishing
this consists in cooling the steel very quickly, by quenching it in water for instance,
as time is then denied for the structure to coarsen at all while the metal cools to at-
mospheric temperature. Such rapid cooling, however, as is well known, hardens the
metal and deprives it of ductility (unless, indeed, it contains very little carbon) , and

LESSON XII THE ANNEALING OF STEEL 3

this would defeat the purpose of annealing which always demands the retention of
considerable ductility. It follows from these considerations that, in annealing, cool-
ing from the annealing temperature cannot be so rapid as to very materially harden
the steel. Its rate should, moreover, be regulated in accordance with the kind of
properties we most desire the steel object to possess. For instance, (1) if softness
and ductility are wanted (for ease in machining) , necessarily at a certain sacrifice of
strength and elasticity, the cooling should be very slow, to wit, with the furnace in
which the object was heated, (2) if greater hardness (for wearing power), strength,
and elasticity are desired, at the necessary sacrifice of some ductility, the cooling
should be more rapid as, for example, in air or, in the case of low carbon steel, in oil
or, with very low carbon steel, even in water (III, Fig. 1).

Rate of Cooling vs. Carbon Content. The lower the carbon content the more
rapid may be the cooling from the annealing temperature without affecting too deeply

Fig. 2. Steel. Carbon 0.50 per cent. Magni-
fied 100 diameters. Heated to 1000 deg. C.
and slowly cooled in furnace. (W. J. Burger,
Correspondence Course student.)

the ductility of the metal. For instance, (1) steel containing not over 0.15 per cent
carbon may be quenched in water, thereby increasing its strength and elastic limit
and still remain very ductile, (2) steel with less than 0.20 or 0.30 per cent carbon may
be quenched in oil with satisfactory results, (3) with a larger proportion of carbon
such rapid cooling is no longer possible, as it would destroy the ductility of the metal,
recourse having then to be had to cooling in air for the desired combination of strength
and ductility or to the double annealing treatment soon to be described.

Rate of Cooling vs. Size of Object. Since large objects necessarily cool more
slowly than smaller ones when subjected to the same cooling influences, it is evident
that the external conditions should also be regulated in accordance with the dimen-
sions of the objects treated. To secure maximum softness and ductility, for instance,
the cooling of small objects should be more effectively retarded than the cooling of
larger ones. Assume, for example, two objects made of the same steel, one large and
one small, and both cooled in air from the annealing temperature; the smaller object

4 - LESSON XII THE ANNEALING OF STEEL

will be harder and less ductile than the larger one, because of its quicker cooling. To
render it as soft and ductile as the larger object cooling in the furnace may be neces-
sary. Similarly, to give strength and high elastic limit the cooling of large objects
must be more vigorously hastened than that of smaller objects as, for instance, cooling
in oil against cooling in air for the smaller piece.

Furnace Cooling from Annealing Temperature. As an example of the effect of
furnace cooling upon the structure of steel, let us take a steel bar J/ in. square, con-
taining 0.50 per cent carbon, heated to 1000 deg. C. and slowly cooled with the fur-
nace. Its structure is shown in Figures 2 and 3. It will be seen to be composed of
the normal proportions of pearlite and free ferrite, namely, some 60 per cent of the
former, and it will also be noted that the pearlite is distinctly laminated (Fig. 3),

Fig. 3. Steel. Carbon 0.50 per cent. Magnified 670 diameters. Heated
to 1000 deg. C. and slowly cooled in furnace. (C. C. Buck, Correspondence
Course student.)

and that in places at least the ferrite forms characteristic polyhedric grains. This
structure is due to the slow cooling of the steel through its critical range, which per-
mits the rejection of the full amount of free ferrite and a distinct crystallization of
the constituents of the residual austenite into plates of ferrite and cementite. The
relative softness and great ductility of the steel in this condition is due (1) to the
presence of the full amount of soft ferrite in relatively large areas and (2) to the pres-
ence of distinctly laminated pearlite indicating the absence of hardening carbon as
explained later.

Air Cooling from Annealing Temperature. To illustrate the influence of air
cooling upon the structure of steel, let us take likewise a steel containing some 0.50
per cent carbon, heated to 1000 deg. C. and cooled in air. Its structure is shown in
Figure 4. It will be found quite unlike the structure of the same steel after
furnace cooling (Figs. 2 and 3). It contains a much smaller proportion of free ferrite,
apparently not over 20 per cent, in the form of a distinct net surrounding dark meshes

LESSON XII THE ANNEALING OF STEEL 5

which a high magnification fails to resolve into distinct parallel plates. '- Relatively
quick cooling through the critical range has prevented the separation of the normal
amount of free ferrite, from which it necessarily follows that the dark constituent
contains more ferrite than true pearlite; nor has it the structure of true pearlite, time
also having been denied on cooling through the range for the formation of distinct
plates of ferrite and cementite. Sorbite is the name of this constituent. The
structure of pearlite passing into sorbite is shown in Figure 5.

Properties of Sorbite. Sorbite has already been briefly described in Lesson XI,
where it was shown that it could be produced in steel forgings of small sections
through simple air cooling from a finishing temperature superior, to the critical range,

Fig. 4. Steel. Carbon 0.50 per cent. Magnified 100
diameters. Heated to 1000 deg. C. and cooled in
air. (W. J. Burger, Correspondence Course student.)

and in larger sections by hastening somewhat their cooling through that range. It
has also been stated that sorbite is harder, stronger, and less ductile than pearlite.
By so regulating the cooling from the annealing temperature, therefore, that sorbitic
steel is produced, hardness, strength, and elasticity will be promoted at the sacrifice
of some ductility (III, Fig. 1). It will be explained in another lesson that sorbite is
generally regarded as one of the transition stages assumed by the metal as it passes
from its austenitic condition, stable above the critical range, to its pearlitic condition,
stable below that range.

Influence of Maximum Temperature. The influence of the maximum tem-
perature to which steel is heated before being allowed to cool is well shown in
Figures 6 to 9 which should be compared with Figures 2 to 5. They refer to steel
containing 0.50 per cent carbon and heated to 800 deg. C., while the structures
illustrated in Figures 2 to 5 refer to the same steel but heated to 1000 deg. The
constitutents are the same, namely, ferrite and pearlite in the furnace cooled
samples, ferrite and sorbite in the air cooled samples, but the higher temperature
resulted in the formation of larger particles of pearlite or sorbite, evidently because
of the formation above the critical range of larger austenitic grains.

6 LESSON XII THE ANNEALING OF STEEL

Influence of Time at Maximum Temperature. Maintaining steel for a long
time at a high temperature causes the formation of large austenite grains, which in
passing through the range are converted into large pearlite or sorbite grains
with rejection of free ferrite in hypo-eutectoid steel and of free cementite in hyper-
eutectoid steel. It is also noted that the more prolonged the heating the smaller
the amount of the excess constituent (free ferrite or free cementite) separating on
rapid (air) cooling through the range. This is shown in Figure 10 in which is
depicted the structure of steel containing 0.50 per cent carbon heated to 1150 deg. C.
for two hours and air cooled. The very large sorbitic grains should be noted as
well as the very small proportion of free ferrite.

Fig. 5. -Steel. Carbon 1.00 per cent. Magnified 1500 diameters.
Pearlite (laminated) passing into sorbite. (Osmond.)

Oil and Water Quenching from Annealing Temperature. As already explained
only steel containing very little carbon may be quenched in oil or water for purposes
of annealing, unless, indeed, the double treatment soon to be described be employed
when higher carbon steels may be so quenched. The structure of steel containing
0.10 per cent carbon, heated to 950 deg. and quenched in water, is shown in Figure 11,
while in Figure 12 is seen the structure of steel containing 0.20 per cent carbon
quenched in oil from a temperature of 850 deg. Rapid cooling through the range did
not prevent the separation of the bulk of the large amount of excess ferrite present in
these steels, hence their softness and ductility even after quenching. They are, how-
ever, somewhat stronger and more elastic than similar steels more slowly cooled, (1)
because they contain a somewhat smaller proportion of soft free ferrite, (2) because
the free ferrite they contain has crystallized into smaller grains, and (3) because their
carburized constituent is sorbitic or even martensitic ' rather than pearlitic.

1 Martensite is the ordinary constituent of steel hardened by quenching. It is hard and de-
prived of ductility.

Fig. 6. Magnified 100 diameters. Heated to
800 deg. C. and slowly cooled in furnace.

Fig. 8. Magnified 100 diameters. Heated to
800 deg. C. arid cooled in air.

Fig. 7. Magnified 670 diameters. Heated to Fig. 9. Magnified 670 diameters. Heated to
800 deg. C. and slowly cooled in furnace. 800 deg. C. and cooled in air.

Figs. 6-9. Steel. Carbon 0.50 per cent. (C. C. Buck, Correspondence Course student.)

LESSON XII THE ANNEALING OF STEEL

Double Annealing Treatment. It has been stated that the most effective way
of retaining in the cold the very fine structure acquired by steel in passing through
its critical range consisted in cooling it very rapidly as soon as it emerged' from that
range, as, for instance, by quenching it in water. This treatment, however, unless the
metal contains very little carbon, hardens the steel and deprives it of ductility, where-
as' annealed steel should not be very hard and should possess much ductility. If
this fine grained but hard steel, however, be reheated to a temperature close to but
below its critical range, say to from 500 to 650 deg. C., it loses its hardness but re-
tains its fine structure and again becomes ductile (IV, Fig. 1).

The double treatment outlined above fulfils admirably the aims generally sought

Fig. 10. Steel. Carbon 0.50 per cent. Magnified 100 diameters.
Heated to 1150 deg. C. for two hours and cooled in air. (Boynton.)

in annealing, namely, the production of a very fine structure possessing high strength
and elastic limit with fair ductility, in other words toughness and high resistance to
wear and to shock. The change of structure taking place on heating hardened steel
close to the lower limit of its critical range will be considered at some length in
Lesson XIV. It will suffice to note here that the metal passes from a fine martensitic
or troostitic condition (the ordinary condition of well-hardened steel) to an equally
fine sorbitic condition, possessing in a high degree the physical properties desired.
The first heating is sometimes called "grain refining" treatment and the second
"toughening" treatment.

The quenching of a piece of steel from above its critical range, while simple enough
in the case of very mild steel, presents increasing difficulties as the carbon increases.
It should be conducted with care and intelligence and only by experts. Steel con-

LESSON XII THE ANNEALING OF STEEL

taining very little carbon, say not over 0.15 per cent, may be quenched in water,
others should be quenched in oil. The Committee on Heat Treatment of the Amer-
ican Society for Testing Materials recommends, in order to lessen the danger of

Fig. 11. Steel. Carbon 0.10 per cent. Magnified 100
diameters. Heated to 950 deg. C. and quenched in
water. (Boylston.)

Fig. 12. Steel. Carbon 0.20 per cent. Magnified
100 diameters. Heated to 850 deg. C. and quenched
in oil. (Boylston.)

cracking, that the object be removed from the oil or water bath before its tempera-
ture has fallen below 160 deg. C., or in any event below 100 deg., and that the second
treatment be applied within a few hours after the quenching, preferably without ever

10 LESSON XII THE ANNEALING OF STEEL

allowing the piece to cool below 100 cleg, and certainly not below 20 deg. The final
properties of the steel will depend upon the temperature of the second heating; the
higher that temperature the softer and more ductile will it be, but also the less strong
and elastic. For great strength, high elastic limit, and little ductility reheating to
500 deg. should be applied, while for great ductility, at the sacrifice of considerable
strength, the reheating should be carried to 700 or 725 deg. For intermediate tensile
strength, elastic limit, and ductility such as are desired in the majority of cases, the
temperature of the second treatment should be between 600 and 650 deg. C. While
from purely theoretical considerations it might be argued that the rate of cooling
from this second treatment is immaterial, there is little doubt but that the strength
of the steel increases somewhat and its ductility decreases with the rapidity of cool-
ing. This cooling may be performed in the furnace, in air, in oil, or in water.

Fig. 13. Steel. Carbon 0.50 per cent. Magnified
500 diameters. Heated to 850 deg. C., quenched in
water, reheated to 600 deg., and cooled in air. (W. H.
Knight in the author's laboratory.)

The double annealing treatment described in the foregoing paragraphs was first
suggested by Wallerant of the Creusot Steel Works, France. It was also described
by Andr6 Le Chatelier and adopted by the French navy. Its use is now general
when high physical requirements are to be met.

In Figure 13 is shown the structure of steel, containing some 0.50 per cent carbon,
after double annealing. The fineness of the structure should be noted as well as the
lack of laminations and the absence of free ferrite. This steel is composed wholly
of finely divided sorbite.

Annealing Eutectoid Steel. W T hile the mechanism of the structural changes
taking place on annealing steel has been made clear in the preceding pages, it may
not be without interest to consider further and in succession the annealing of eutee-
toid, hypo-eutectoid, and hyper-eutectoid steel, as these three types of steels have
different structures and their annealing involves different structural changes.

LESSON XII THE ANNEALING OF STEEL

Slowly cooled eutectoid steel is composed wholly of pearlite which, upon being
heated through the single critical point of the metal, namely Ac 3 . 2 .i, is converted into
a solid solution (austenite). The grains of this austenite are very fine as the steel
emerges from its range and they are kept from growing by preventing the steel from
reaching a higher temperature. On cooling through Ar 3 . 2 .i the metal again becomes
pearlitic if it be given time, as, for instance, in cooling in the furnace, while it becomes
sorbitic if cooled more quickly, as, for instance, in air in the case of small objects.
Should the steel be quenched in water or oil from the annealing temperature and
then reheated near but below the point Acs.2.i, the finely martenso-troostitic struc-
ture produced by quenching from above the range is converted into very fine sorbite.

Fig. 14. Steel. Eutectoid. Magnified 412 diameters. Heated to 800 deg.
C. and slowly cooled in furnace. (C. C. Buck, Correspondence Course
student.)

It has been explained in Lesson X that, for like treatments, the structure of eutec-
toid steel is finer than that of either hypo-eutectoid or hyper-eutectoid steel and this
holds true in the case of annealed samples, although the difference may not be notice-
able when comparing the structure of eutectoid steel with that of hyper-eutectoid
steel containing but a slight excess of free cementite.

In Figure 14 is shown the structure of eutectoid steel heated to 800 deg. C. and
slowly cooled in the furnace. It is made up of well-developed pearlite. The struc-
ture of the same steel, quenched in oil at 825 deg., reheated to 650 deg., and cooled
in air, is exhibited in Figure 15. The metal is now composed of fine grained sorbite.

Annealing Hypo-Eutectoid Steel. Slowly cooled hypo-eutectoid steel is an ag-
gregate of pearlite and free ferrite. On being heated through its critical range, as
soon as the point Aci is reached, the pearlite is bodily converted into austenite, while
the ferrite still remains free. On further heating, however, it begins to be absorbed
by austenite, its absorption being completed as the metal emerges from its Ac 3 point.

12 LESSON XII THE ANNEALING OF STEEL

Above Acs the steel is composed wholly of homogeneous austenite. On cooling
through the critical range, unless, indeed, the cooling be very rapid and sufficient
carbon be present, ferrite is again liberated in amount proportional to the slowness
of the cooling up to the maximum quantity consistent with the carbon content in
the steel. If the cooling be very slow then, for instance in the furnace, the totality
of the excess ferrite will be rejected and the residual austenite converted into well-
defined pearlite (Figs. 3 and 7), while if the cooling be more rapid, for instance in
air in the case of small objects or in oil with larger ones, a portion only of the excess
ferrite is liberated while the residual austenite is converted into sorbite (Figs. 4
and 9). The liberation of ferrite taking place during the slow cooling of hypo-eutectoid
steel coarsens its structure and is the chief reason why annealed hypo-eutectoid steel

Fig. 15. Steel. Eutectoid. Magnified 720 diameters.
Heated to 825 deg. C., quenched in oil, reheated to
650 dog., and cooled in air. (Boylston.)

cannot have as fine a structure as annealed eutectoid steel. Howe further contends
that as hypo-eutectoid steel is heated from Acj to Ac 3 a new crystalline growth takes
place which is the coarser the greater the distance between AI and A 3 , that is, the less
carbon in the steel, so that by the time the old structure has been obliterated, i.e. at
Ac 3 , a new grain has formed which is an additional reason why the structure of hypo-
eutectoid steel cannot be refined to the same extent as that of eutectoid steel.

The structure of hypo-eutectoid steel after double annealing has been shown in
Figure 13. The rapid cooling through the range prevented the liberation of ferrite,
while the second treatment produced sorbite, but this sorbite is not as fine grained
as that produced in eutectoid steel by similar treatment.

Annealing Hyper-Eutectoid Steel. Slowly cooled hyper-eutectoid steel is an
aggregate of pearlite and free cementite. On heating it through its critical range, i.e.
through its Aca.2.i. and Ac cm points, pearlite is converted into austenite at the lower
point and this austenite absorbs the free cementite as the metal is further heated

LESSON XII THE ANNEALING OF STEEL

from Ac 3 .2.i to Ac cm . At Ac cm the absorption is complete and the metal composed
entirely of austenite. On cooling through the range, if time be given, as for instance
in cooling in the furnace, the full proportion of free cementite is again liberated and
the residual austenite converted at Ar 3 . 2 .i into clearly laminated pearlite (Fig. 16).
If the cooling be more rapid, as for instance in cooling small pieces in air, a portion
only of the free cementite is set free, while the residual austenite is converted into
sorbite. This setting free of cementite, like the liberation of ferrite in hypo-eutec-
toid steel, coarsens the structure. The coarsening influence of free cementite, how-
ever, is far from being as marked as that of free ferrite, chiefly because free cementite
is generally present in much smaller proportions. Steel containing as much as 1.50

Fig. 16. Steel. Carbon 1.43 per cent. Magnified 500 diameters.
Heated above critical range and slowly cooled in furnace. (Boyn-
ton.)

per cent carbon, for instance, contains but 11.50 per cent of coarsening cementite,
while steel with 0.40 carbon contains 52 per cent of coarsening ferrite.

In Figure 17 is shown the structure of hyper-eutectoid steel subjected to the
double annealing treatment. This treatment prevented the separation of any free
cementite and resulted in the production of fine grained sorbite.

Annealing Steel Castings. It has been mentioned that the very coarse struc-
ture of steel castings, called "ingotism" by Howe, was not as readily refined as the
structure of steel forgings, its satisfactory annealing often necessitating heating to
temperatures considerably higher than the critical range. Notwithstanding their
greater resistance to the annealing treatment, successfully annealed castings may
possess physical properties fairly equal to those of forgings. In Figure 18 is shown
the structure of cast steel containing some 0.30 per cent of carbon and properly an-
nealed. The presence of a relatively small amount of free ferrite will be noted. When
highly magnified the carbon-holding constituent should have a sorbito-pearlitic ap-
pearance.

LESSON XII THE ANNEALING OF STEEL

Spheroidizing of Pearlite-Cementite. On slow cooling through the critical range,
austenite of eutectoid composition is converted into pearlite made up of distinct

Fig. 17. Steel. Carbon 1.25 percent. Magnified 670 diameters. Heated
to 800 deg. C. and quenched in oil, reheated to 600 deg. and air cooled.
(C. C. Buck, Correspondence Course student.)

<#*&::.'

Fig. 18. Steel. Cast. Carbon 0.30 per cent.
Magnified 100 diameters. Annealed. (W. J.
Burger, Correspondence Course student.)

parallel plates or lamellae alternately of ferrite and cementite. This condition of the
cementite of pearlite, however, is not final, it is not structurally stable, if it can be
thus expressed, for if the steel be kept for a sufficiently long time at a temperature

LESSON XII THE ANNEALING OF STEEL 15

but slightly below the range, preferably between 600 and 700 deg. C., the cementite
shows a marked tendency to form rounded particles or, as it has been well said, to
"spheroidize." This phenomenon is shown in Figure 19 in the case of 1.24 per cent
carbon steel. The cementite now occurs as small irregular grains embedded in
ferrite. This variety of pearlite, if we can still speak of it as pearlite, is sometimes
called "granular" pearlite. Some writers state that the essential factor in this
spheroidizing process is extremely slow cooling between 700 and 600 deg. Since
the point Ari generally occurs a little below 700 deg. it may well be asked whether
(1) the spheroidizing of pearlite is due to excessively slow cooling through that
point or whether (2) lamellar pearlite must first be formed by moderately slow cool-
ing through Ari, being afterwards converted into Spheroidized- pearlite below An.
The evidences at hand are not conclusive. Spheroidized pearlite is softer, less
tenacious, and more ductile than lamellar pearlite. The author has heard of this

Fig. 19. Steel. Carbon 1.24 percent. Mag-
nified 1000 diameters. Spheroidized cemen-
tite. (Osmond.)

spheroidizing treatment having been applied to high carbon steel in order to soften
it so as to facilitate its machining, it being afterwards reheated above its range and
made pearlitic, sorbitic, or martensitic according to requirements.

Varieties of Pearlite. From the foregoing it will be evident that several varieties
of pearlite are to be considered and that the physical properties of steel will depend
greatly upon the character of the pearlite it contains. Arnold considers four varieties
of pearlite which are well illustrated in Figure 20. His first phase, which he calls
"sorbitic" pearlite, is generally called sorbite by other writers. The character and
physical properties of sorbite have been described, as well as some of the conditions
necessary to its formation. His second and third phases, to which he gives the names
respectively of "normal" and "laminated" pearlite, are both true pearlite, the thicker
lamella of the latter being due to a slower cooling through the critical range. His
fourth phase is pearlite in the process of spheroidizing. l

Graphitizing of Cementite. It will be explained in another lesson that the car-
bide FeaC (cementite) is not the most stable form that can be assumed by carbon

LESSON XII THE ANNEALING OF STEEL

Showing the Properties of Pearlite and its Decomposition Product.
Fe 3 C represented Black.

Mechanical Properties of

Mass.

Microstructure.

Segregation Stages.

Maximum tensile stress
about 70 tons per square
inch. Elongation on 2
inches =about 10 per cent.

1ST PHASE.
"Sorbitic" ! pearlite with

emulsified Fe 3 C.
dark on etching.

Verv

Maximum tensile stress
about 55 tons per square
inch.. Elongation on 2
inches=about 15 per cent.

Maximum tensile stress
about 35 tons per square
inch. Elongation on 2
inches=about 5 per cent

2ND PHASE.

Normal pearlite with semi-
segregated FejC. Dark
on etching.

3RD PHASE.

Laminated pearlite with
completely segregated
Fe 3 C. Exhibiting a play
of gorgeous colours when
lightly etched.

Maximum tensile stress
about 30 tons per square
inch.

4TH PHASE.

Laminated pearlite passing
into massive Fe 3 C and
ferrite.

NOTE. It is important to remember that in a single section of steel two or even all
three phases of pearlite may be observed in juxtaposition gradually merging into each
other.

Fig. 20. (Arnold).

LESSON XII THE ANNEALING OF STEEL 17

when alloyed with iron. It will he shown that cementite tends to break up into iron
and graphite according to the reaction

Fe 3 C = 3Fe + C

I I I

cementite ferrite graphite

and that the graphite form is the final stable condition of carbon. This graphitizing
tendency of cementite remains latent unless the conditions be favorable to its activity.
These conditions are (1) long exposure to a temperature exceeding the critical range
and slow cooling, 1 (2) the presence of much carbon, and (3) the presence of silicon or
of some other elements exerting a similar influence. It will be~shown that this ten-
dency of cementite to be converted into graphite and iron is responsible for the pro-
duction of so-called malleable castings and, probably, also for the production of gray
cast iron. In the case of steel, because of the relatively small amount of carbon
present (not exceeding 1.75 or at the most 2 per cent), the graphitizing tendency is
slight. Long exposure of hyper-eutectoid steel, especially if it contains more than
one per cent carbon, however, to a temperature exceeding its critical range, is always
likely to produce a small amount at least of graphitic carbon greatly impairing
thereby, if not ruining, the metal. An instance of graphite formation in high carbon
steel is shown in Figures 21 and 22. There is little doubt but the free cementite pres-
ent in hyper-eutectoid steel, and formed as the steel cools from its Ar cm to its Ar 3 . 2 .i.
point, is more readily converted into graphite than the cementite included in the
pearlite. Once the graphitizing is started, however, it may be carried to completion
and include the whole of the cementite present. This indeed is what happens in
certain grades of malleable cast iron and of gray cast iron which contain practically
the totality of their carbon in the form of graphite. The presence of some free cemen-
tite appears to be necessary to start the graphitizing, which would explain why it
does not take place in hypo-eutectoid steel. The author at least never had a case
of graphite formation in such steels brought to his attention nor was he ever able to
produce graphite in hypo-eutectoid steel.

Burnt Steel. When high carbon steel is heated to a temperature approaching
its melting-point, it becomes extremely red short as well as cold short, its fracture
becomes very coarse and shiny and these defects cannot be cured short of remelting
the metal. The steel in such condition is said to be burnt. These results are ap-
parently brought about by the evolution of gases under the influence of a high tem-
perature, chiefly carbon monoxide, resulting from atmospheric oxygen finding its way
through the pores of the metal and combining with some of the carbon, although
according to Howe other occluded gases, such as hydrogen and nitrogen, may also
contribute. These gases force the crystalline grains apart, destroying their cohesion,
hence the brittleness of the metal. Oxidized membranes are also frequently found
surrounding some of the grains, their presence readily explaining the impossibility
of restoring burnt steel by forging since they would prevent the welding of adjacent
grains. Instances of the structure of burnt steel are shown in Figures 23 and 24. Howe
defines burning as being "a mechanical separation of the grains on extreme overheat-
ing." Some writers have argued, apparently on good ground, that burning will not
take place unless the steel has been heated to so high a temperature that it has actu-
ally begun to melt, the explanation being perfectly consistent with the well-known

1 One or two instances have been cited of graphite having been formed below the critical range.

LESSON XII THE ANNEALING OF STEEL

S^mmM^^ *4f wl

' : iit

p*se

Fig. 21. Stool. Carbon 1.25 per cent. Magnified 150 diameters. An-
nealed five hours at 830 dog. C. (C. C. Buck, Correspondence Course
student.)

f8&W*^lPijli

Fig. 22. Steel. Carbon 1.25 per cent Magnified 670 diameters. An-
nealed five hours at 830 dog. C. 'C. C. Buck, Correspondence Course
student.)

LESSON XII THE ANNEALING OF STEEL 19

fact that high carbon steel burns much more readily than low carbon steel. To make
the matter clear let us consider the diagram of Figure 25 in which the solidification
period of steel is shown as influenced by its carbon content. This diagram will be
discussed at greater length in another lesson. Let us for the present note (1) that
as the carbon increases from to 2.0 per cent the solidification of the steel is lowered
from A to B, that is from 1500 deg. C. to 1325 cleg., (2) that while carbonless iron
solidifies at a constant temperature, namely 1500 deg., as the carbon increases, the
range of temperature covered by the solidification period increases likewise, extend-
ing from B to C with 2 per cent carbon, that is from 1325 to 1130 deg. C. ABC then
represents the solidification zone of steels of increasing carboii content and the
heating of the metal to any point within this zone, when it is partly melted, will
cause it to burn. It follows from this that carbonless iron and very low carbon steel

Fig. 23. Burnt steel. Carbon 1.24 per cent. Fig. 24. Burnt steel. Magnified 30

Magnified 20 diameters. Quenched at a white diameters. (Stead.)

heat. Unetched. (Osmond.)

can be heated to a very high temperature without burning, while the danger of burn-
ing increases with the carbon. With 0.50 per cent carbon, for instance, the burning
zone extends from 1400 to 1450 deg., with 1.0 per cent it extends from 1310 to 1400
deg., with 1.50 per cent carbon from 1210 to 1360 deg. In short, as the carbon in-
creases the steel burns more readily (1) because its melting-point is lowered and (2)
because its solidification zone, which is also its burning zone, is widened. According
to the theory it should not be possible to burn carbonless iron, and indeed the author
docs not know that the claim has ever been made that carbonless iron could be burnt.
If ABC represents a burning zone into which steel cannot be brought without
having its useful qualities destroyed, we naturally ask why all steels are not so injured
seeing that they must pass at least once through this zone in cooling from the molten
condition. The reason why steel does not burn on solidifying and further cooling is
explained by Howe on the ground that while steel ingots or other castings solidify,
much hydrogen is given out which may mechanically restrain the oxygen from enter-
ing and also counteract it, preventing thereby the evolution of CO from within and
the formation of oxidized films, the chief causes of burning. It is also possible, Howe
says, that the greater kneading which an ingot undergoes cures burning, while the

LESSON XII THE ANNEALING OF STEEL

slight kneading possible in reworking a steel bar does not. This, however, would not
explain why steel castings, which undergo no work at all, are not burnt, unless it is
because they are generally protected from atmospheric oxidation by their molds
(Howe).

Burning should not be confounded with overheating. Overheated steel has a
very coarse structure and fracture, but it can be restored by heat treatment alone, or
at least by heating and forging, while burnt steel is incurable. Overheating results
from heating close to but below AC (Fig. 25), generally for a considerable length of
time, while, as explained, the temperature in burning is carried above AC.

/5OO

Carbon per cenf

Fig. 25. Diagram depicting the burning temperature range.

Important results recently obtained by Gutowsky would place the end of the
solidification of various carbon steels as indicated by the dotted line in Figure 25.
If these are the correct temperatures at which solidification is complete, it follows
that the burning zone is wider than was generally believed before the publication of
these results.

Crystalline Growth of Austenite Above the Critical Range. Above its critical
range steel is composed of polyhedric crystalline grains of austenite, which are made
up of small crystals (probably octahedra) similarly oriented in the same grain but
whose orientation changes from one grain to the next. Indeed it is this lack of uni-
formity of the orientation of the crystalline matter building up the grains that gives
existence to these grains, for if all the small crystals of which they are composed were
similarly oriented, clearly there would be but a single allotrimorphic crystal or grain.

LESSON XII THE ANNEALING OF STEEL 21

If steel be maintained for a long time above its critical range, the austenite grains of
which it is composed show a tendency to grow in size through adjoining grains assu-
ming like crystalline orientation and, therefore, merging into a single and correspond-
ingly larger grain. This growth increases with the temperature and with the duration
of the treatment. Given a sufficiently long time and sufficiently high temperature,
but one grain must be formed. This, as already seen in Lesson X, is actually what
takes place in meteorites during the cooling of which the prevailing conditions are
such as to produce this uniformity of orientation. It follows from the above con-
siderations that on annealing, if the metal be kept a long time above its critical range,
even but slightly above it, a coarser austenite will be formed which in turn implies,
after slow cooling, a coarser pearlitic or sorbitic structure. In hypo-eutectoid steel
the grains of austenite will expel some free ferrite, and in hyper-eutectoid steel some
free cementite, before being converted into pearlite, but the final pearlite grains will
nevertheless increase in size with the size of the original austenite grains. The
structure of steel containing 0.50 per cent carbon kept two hours at 1150 deg. C.
and cooled in air has been shown in Figure 10. The very large sorbitic grains
formed prove the existence above the range of equally large, or even larger, austenitic
grains. The cooling through the range was so rapid that but a small amount of
free ferrite was separated, the sorbite grains, therefore, representing nearly the exact
size of the austenite grains.

In Figure 26 an attempt has been made to depict this relation between the
austenitic structure above the range and the corresponding pearlitic or sorbitic struc-
ture below the range. The steel considered is supposed to be hypo-eutectoid and to
contain, after slow cooling, a large proportion of free ferrite. A is intended to rep-
resent a piece of this steel made up of nine relatively small austenitic grains formed
on short exposure above the range. On cooling through the range ferrite is liberated
and the residual austenite grains converted into as many pearlite grains as shown in
A'. The small squares of the matrix surrounding the pearlite grains represent as
many small ferrite grains. After a longer exposure, possibly at a higher temperature,
the steel will be made up of larger austenite grains, say of four grains as shown in B,
and these, on slow cooling through the range, will be converted after rejection of fer-
rite into four pearlite grains as indicated in B'. Theoretically, at least, we may assume
that the temperature above the range may be so high and the exposure at that tem-
perature so long that but a single austenite grain is formed, the entire mass having
assumed a uniform crystalline orientation as shown in C. On slow cooling a single
pearlite grain would then be formed surrounded by free ferrite in the form of small
grains as depicted in C'. An exceedingly slow cooling through and below the range
would have a tendency to cause the free ferrite to crystallize into larger grains and
eventually to form but a single grain as indicated in C". 1 Finally, as will be later
explained, on very long exposure to a high temperature the cementite should, theo-
retically at least, be converted into as many graphite particles as there are austenite
grains and, therefore, into a single graphite particle in case there is but a single grain
of austenite, the steel then consisting, after slow cooling, of a kernel of graphite sur-
rounded by uniformly oriented ferrite as shown in C'". As already explained,
however, this breaking up of cementite into graphite and ferrite does not take place
unless a considerable amount of carbon is present, namely over 1 per cent.

1 Such slow cooling, as previously explained, would also have a tendency to cause the spheroi-
dizing of cementite.

LESSON XII THE ANNEALING OF STEEL

LESSON XII THE ANNEALING OF STEEL 23

The conditions depicted in A', B' , C', C", and C'" are conditions of equilibrium,
according to the phase rule, since but two phases are present. It is now believed,
however, that A', B', C', and C" represent metastable equilibrium while D, only,
represents stable equilibrium. The phase rule will be considered in another lesson
when these remarks will be made clear.

Crystalline Growth of Ferrite Below the Critical Range. Stead in 1898 showed
that strained ferrite, when heated close to but below the critical range of the metal,
undergoes a marked crystalline growth caused by adjoining ferrite grains assuming
the same crystalline orientation and therefore merging into a larger grain. It seemed
evident from Stead's experiments that unless the ferrite be strained (through cold
working or otherwise) it will not grow on annealing, and also -Htafc the presence of a
relatively small amount of carbon (some 0.15 per cent or more) effectively prevents
the growth of the ferrite grains, apparently because of the pearlite particles standing

Fig. 27. Steel. Carbon 0.05 per cent. .Magnified 6 diameters. Subjected to
Brinell ball test under pressure of 6000 kilograms and heated to 650 deg. C. for
seven hours. Vertical section through bottom of spherical depression. (J. O.
Connolly in the author's laboratory.)

in their way and preventing their merging. The author recently conducted in his
laboratory some experiments, the results of which throw additional light upon this
crystalline growth. 1

The following extracts from a paper by the author to be presented at the sixth
Congress of the International Association for Testing Materials describe briefly some
of the most significant results obtained.

In Figure 27 is shown the slightly magnified structure of a steel containing 0.05
per cent carbon which had been subjected to the Brinell ball test 2 under a pressure of
6000 kilograms and then annealed at 650 dog. C. for seven hours. The section shown
is a vertical one passing through the bottom of the spherical depression made by the

1 These experiments were made by J. O. Connolly, at the time a research student at Harvard
University, now with the American Steel and Wire Co.

2 In using the Brinell ball test as a suitable means of straining ferrite in order to study its
growth on annealing the author followed Charpy.

24 LESSON XII THE ANNEALING OF STEEL

10mm. ball used. It will be obvious that the strain was most severe at A, that is
at the very bottom of the depression, and that it decreased gradually in intensity
from A to D. The following features of the structure should be noted as having
special significance: (1) At D where the metal was but slightly if at all strained no
crystalline growth occurred, (2) at C where the strain must have been more severe a
sudden growth of maximum intensity has taken place, (3) from C to B as the severity
of the strain increases the crystalline growth shows a gradual decrease, and (4) from
B to A, that is, with further increase of the intensity of the strain, no crystalline
growth has taken place. These observations point to the conclusion that ferrite
grains will not grow on annealing below the critical range unless they have been
subjected to a certain stress creating a certain strain, and that they will not grow if

Fig. 28. Steel. Carbon 0.05 per cent. Magnified 6
diameters. Subjected to Brinell ball test under
pressure of 3000 kilograms and heated to 650 deg. C.
for seven hours. Vertical section through bottom of
spherical depression. (J. O. Connolly in the author's
laboratory.)

that stress, and therefore the resulting strain, has been exceeded. In other words,
they point to the existence of a critical strain producing growth, strains of greater
or less magnitude being ineffective. The narrow region occupied by the critically
strained metal should also be noted as well as the very sharp line of demarcation be-
tween the critically strained and the under-strained metal. The separation of the
critically strained metal from the over-strained is not so sharp.

Similar experiments were repeated many times and like results always obtained.

A piece of the same steel was subjected to exactly the same treatment, except that
the ball pressure applied was 3000 kilograms or half the pressure applied to the
previous one. The crystalline growth resulting from the annealing of this sample is
shown in Figure 28. Because of the smaller stress applied, the critically strained por-
tion of the metal is nearer the depression. This would naturally be expected. Here,
as in the previous case, we have three distinct regions: (1) the metal surrounding the
depression and extending to a certain distance which was too severely strained to

LESSON XII THE ANNEALING OF STEEL 25

grow, (2) the critically strained metal in the form of a spherical shell, and (3) the
rest of the metal unstrained or too feebly strained for the growth to take place. Fig-
ure 29 is a section through a similar sample, the specimen having been ground level
with the bottom of the depression. The occurrence in this section of a ring showing
crystalline growth will be readily understood.

In Figure 30 is shown the structure of a bar of the same steel which after having
been completely bent was subjected to annealing (seven hours at 650 deg. C.). A
piece of the bent portion of this bar was then cut and a longitudinal section through
its center prepared for microscopical examination. It will be obvious that the upper
part of the bent portion of the bar was subjected to severe tension and the under
part to severe compression. Somewhere between the upper and lower parts a

Fig. 29. Steel. Carbon 0.05 per cent. Magnified
6 diameters. Subjected to Brinell ball test under
pressure of 3000 kilograms and heated to 650 deg.
C. for seven hours. Horizontal section through
bottom of spherical depression. (J. O. Connolly in
the author's laboratory.)

neutral plane existed which was subjected neither to tension nor to compression, and
in the vicinity of this plane the metal was but slightly strained. Moving in both
directions from this neutral plane, the metal becomes gradually more severely strained.
Figure 30 shows that (1) in the center of the bar no growth took place, the metal
being here under-strained, (2) as soon as the critically strained portion was reached
a very abrupt growth occurred of maximum intensity, and (3) this growth decreased
gradually as the metal became more severely strained, being very slight if existing
at all where the strain was maximum, that is, near the upper and under parts of the
bend. The widening of the central zone, free from growth as the distance from the
bend increases, is also consistent with the existence of a critical strain for it is evident
that the portion of unstrained or under-strained metal increases with that distance.
Incidentally this experiment shows that tension is apparently as effective as com-

26 LESSON XII THE ANNEALING OF STEEL

pression in producing crystalline growth, there being apparently no difference in the
size of the ferrite grains between the upper and under parts of the bar. The criti-
cally strained portions occupy also nearly the same position with regard to the out-
side surfaces, that is, they occur at the same depth. Their width also appears to be
the same.

Bars of the same steel were subjected to compression, shearing, and twisting
stresses and the results obtained were in every case consistent with the existence of
a critical strain.

With a view of securing some data in regard to the magnitude of the critical stress
needed to induce growth on annealing, a number of bars of the same steel were sub-
jected to tensile stresses of increasing intensity, annealed at 650 deg. seven hours, and
in every case a cross section of the strained and annealed bar was prepared for micro-

Fig. 30. Steel. Carbon 0.05 per cent. Magnified 3 diameters. Bar bent
double and heated to 650 deg. C. for seven hours. Longitudinal section through
center of bent portion. (J. O. Connolly in the author's laboratory.)

scopical examination. The elastic limit or rather yield point of the metal was in the
vicinity of 28,000 Ibs. per square inch, and its ultimate strength 45,600 Ibs. per square
inch.

In Figures 31, 32, 33, and 34 are shown the structures of the bars subjected re-
spectively to stresses of 38,000, 40,000, 42,000, and 44,000 Ibs. per square inch.

These tensile tests afford another conclusive evidence of the existence of a critical
strain and the fact that a stress of 40,000 Ibs. per square inch produces a marked
growth while stresses but slightly inferior or superior, namely, 38,000 and 42,000 Ibs.
per square inch do not induce growth is an indication of the narrowness of the range
of the critical stress.

Brittleness of Low Carbon Steel. The crystalline growths and other structural
changes described in the foregoing pages lead naturally to the consideration of the
brittleness occasionally exhibited by low carbon steel. Since steel containing very
little carbon is essentially made up of ferrite, its occasional brittleness must be due to
the occasional brittleness of ferrite, a constituent by nature soft and ductile. Stead
has indicated two kinds of brittleness from which ferrite may, and occasionally does.

LESSON XII THE ANNEALING OF STEEL

suffer, namely, (1) "inter-granular" brittleness and (2) "inter-crystalline" or "cleav-
age" brittleness.

By inter-granular brittleness is meant a lack of cohesion between the ferrite grains
leading to ready fracture under shock, the line of fracture following the boundary

Fig. 31. Tensile stress, 38,000 Ibs.'per sq. in.

Fig. 33. Tensile stress, 42,000 Ibs. per sq. in.

Fig. 32. Tensile stress, 40,000 Ibs. per sq. in.

Fig. 34. Tensile stress, 44,000 Ibs. per sq. in.

Steel. Carbon 0.05 per cent. Magnified 6 diameters. Strained by tension and heated to 650 deg.
C. for seven hours. Cross sections of strained bars. (J. O. Connolly in the author's laboratory.)

lines of the grains. Such brittleness is usually due to the presence of impurities form-
ing brittle and more or less continuous membranes surrounding the grains. The
presence of much phosphorus, however, appears to produce inter-granular brittleness
without producing surrounding membranes.

LESSON XII THE ANNEALING OF STEEL

Inter-crystalline or cleavage brittleness is caused by the ferrite grains assuming
nearly the same crystalline orientation so that the plane of fracture follows the cleav-
age planes and passes from grain to grain almost in a straight line. The diagram
shown in Figure 35 will make this clear. The cross lines represent the cleavage
planes in each grain. In B the metal is made up of large ferrite grains but the crys-
talline orientation of these grains is so heterogeneous that a line of fracture cannot
readily be developed and pass from grain to grain, the abrupt change of crystalline
orientation encountered at each boundary line acting as an effective obstruction to
its advance. In A, on the contrary, while the grains are smaller they have nearly the
same orientation, hence fracture may proceed from grain to grain with much greater
ease and in a nearly straight line. Fortunately, so uniform a crystalline orientation

Fig. 35. (Stead.)

is not often met with and cleavage brittleness is a rather rare occurrence. It can be
cured by reheating the steel to 900 deg. or higher.

Stead also noticed that low carbon steel plates rolled below the critical range and,
therefore, strained, and annealed likewise below the range, often exhibit a tendency
to break in three directions, namely, at 45 deg. to the direction of rolling and at right
angles with the surface of the plates, that is, in the directions of the three cleavage
planes of a cube having four faces at 45 deg. to the edges, and two faces parallel
to the surface of the plates. This he calls "rectangular" brittleness. "We are led
from this to conclude," Stead writes, "that, just as light impresses a latent image on
a bromide photographic plate which cannot be seen but is developed and made mani-
fest by the action of certain chemical agencies, so the rolling appears to impress a
latent disposition in the steel to crystallize in certain fixed positions, and annealing
develops it afterwards." The brittleness here referred to is undoubtedly caused by
the crystalline growth of strained ferrite when annealed below its critical range as
fully explained in this lesson, the formation of large ferrite grains naturally causing
brittleness. This kind of brittleness is sometimes called "Stead's brittleness." No
very satisfactory explanation has so far been offered to account for this greater brit-

LESSON XII THE ANNEALING OF STEEL 29

tleness in certain directions. It may be that the large crystalline grains of ferrite
produced have nearly the same orientation and that they are so oriented as to lead
to easy inter-crystalline rupture in the directions indicated.

Experiments

The student should procure samples of hypo- and hyper-eutectoid steels, cast,
hot worked, and cold worked. These should be microscopically examined and, if
possible, photographed. They should then be annealed, following the instructions
given in this lesson, and cooled at various rates, to wit, in furnace, in air, and, in the
case of low carbon steels, in oil or even in water if the steel does not contain over
0.15 per cent carbon. The structure of each sample should be compared to the struc-
ture of the same steel before annealing and the structural changes resulting from the
annealing operation carefully noted.

It is advisable to subject at least one sample of hypo-eutectoid steel and one sam-
ple of hyper-eutectoid steel to the double annealing treatment, noting the finely
sorbitic structure produced by this operation.

Samples of hyper-eutectoid steel, preferably containing not less than 1.25 per cent
carbon, should be subjected to (1) spheroidizing treatment and (2) graphitizing treat-
ment, and their structure carefully examined.

A sample of hyper-eutectoid steel should be heated into its burning zone and the
resulting structure examined.

A small sample of steel containing not more than 0.10 per cent carbon should be
filed very smooth, subjected to the ball test under a pressure of some 3000 kilograms,
and annealed to 650 or 700 deg. for at least five hours. A vertical section and an
horizontal section passing through the bottom of the spherical depression should be
prepared for microscopical examination and the location and magnitude of the
crystalline growth noted.

It is desirable that all samples should be photographed, using suitable magnifica-
tions.

Examination

I. Assuming a forged steel containing 0.50 per cent carbon, what annealing treat-
ment would you recommend in order to produce: (1) great softness, (2) hard-
ness and great strength but little ductility, and (3) a fair combination of
strength, elasticity, and ductility?

II. Assuming a steel containing 0.10 per cent carbon, what treatment would you
recommend to produce maximum strength?

III. Assuming a steel containing 1.25 per cent carbon, what treatment would you

recommend to produce a desirable combination of strength, elasticity, and
ductility so that the metal while tenacious will satisfactorily stand wear and
shocks?

IV. Explain how steel can be made sorbitic by annealing.

V. Explain the difference in physical properties between sorbitic and pearlitic
steel containing the same amount of carbon.

30 LESSON XII THE ANNEALING OF STEEL

VI. Explain why the structure and fracture of annealed hypo-eutectoid steel can-
not be made as fine as the structure and fracture of annealed eutectoid steel.

VII. Describe and explain the burning of steel.

VIII. Explain the relation existing between the dimension of austenite grains formed
above the critical range and the pearlite grains formed on passing through
the range.

LESSON XIII

THE HARDENING OF STEEL

References have already been made in these lessons to the4nvaluable property
possessed by iron, containing a sufficient amount of carbon, of becoming extremely
hard when suddenly cooled from a high temperature as, for instance, by quenching
in water from a bright red heat. This operation is known as the hardening of steel.
The close relation existing between the hardening of steel and its critical range, which
has also been alluded to, provides the key to the rationale of the hardening opera-
tion. This operation consists of two distinct steps (1) heating to the hardening
temperature and (2) cooling from that temperature.

Heating for Hardening. In order to harden steel it is necessary first to heat
it above its critical range, because it is in passing through that range that it acquires
hardening power. Any attempt at hardening it by cooling it suddenly from a temper-
ature inferior to its critical range would result in but a very slight, if any, increase of
hardness. It is evident, therefore, that to possess hardening power steel must be in
the condition of a solid solution since the aggregate of ferrite and cementite formed on
slow cooling through the critical range cannot be hardened by sudden cooling. The
metal should not be heated much above the top of its range, because in so doing we
coarsen its structure as explained in previous lessons, while we do not increase, mate-
rially at least, its hardening power, and our aim in hardening should be to secure
maximum hardness and finest possible structure. Quenching from a temperature
greatly exceeding the critical range, moreover, increases the danger of warping and
cracking the objects in the quenching bath. Nor should the steel be heated to a
temperature much above its critical range and then cooled to that range before quench-
ing, as sometimes recommended, because its structure is then likewise coarsened by
the heating and slow cooling preceding the quenching. Clearly the rationale of the
hardening operation consists in heating the metal just through its critical range,
thus conferring to it both full hardening power and finest possible structure, and
then in cooling it suddenly as soon as it emerges from its range, lest its structure be
coarsened by heating above the range or by prolonged exposure at the quenching
temperature. This judicious method of conducting the hardening operation is some-
times described as "hardening on a rising temperature." Let it be borne in mind
that, since the position and width of the critical range vary in different steels, the
most desirable quenching temperature will vary likewise. Low and medium high
carbon steels should be quenched at higher temperatures than high carbon steel,
for in order to acquire full hardening power they should be heated past their upper
critical points, namely, Ac 3 or Ac 3 .2, as the case may be.

Cooling for Hardening. To harden the steel the metal should be cooled very
quickly from the temperatures mentioned in the above paragraph to atmospheric
temperature, generally by immersing it in a medium capable of rapidly abstract-
ing heat from it. The increase of hardness will lie the greater the higher the carbon

2 LESSON XIII THE HARDENING OF STEEL

content, at least up to the eutectoid point, and the more rapid the cooling, the
latter, in turn, depending upon the size of the object hardened and the nature of the
quenching bath, i.e. its power of abstracting heat from the cooling mass. It was long
thought that this so to speak cooling power of the bath depended chiefly, if not
solely, upon its temperature at the time of immersion and upon its heat conductivity.
It was believed, for instance, that mercury was a more effective cooling medium than
water, because of its greater conductivity for heat, that cold water was more effec-
tive than tepid water, because of its lower temperature, etc. Recent investigations
appear to show, however, that the cooling power of a quenching bath is, within limits,
quite independent of its actual temperature and of its heat conductivity, and even of
its specific heat. Benedicks contends that it depends almost exclusively upon its
latent heat of volatilization. Its temperature, however, should be low enough to

^fc ^^^^ *^5it J^^Wf-*JJ jf fi^r i * *

Fig. 1. Steel. Carbon 0.45 per cent. Mag-
nified 1000 diameters. Heated to 825 deg. C.
and quenched at 720 deg. (Osmond.)

prevent the adherence of vapor bubbles to the metal. In accordance with these
views mercury, in Benedicks' opinion, is inferior to water while saline solutions are
not superior to it. Methyl alcohol, on the contrary, is a more effective cooling medium
for hardening than water. According to Le Chatelier, also, mercury is less effective
than water but in his opinion because of its lower specific heat. Le Chatelier be-
lieves that the specific heat of the liquid is the most important factor influencing its
value as a cooling medium, its conductivity being of secondary importance, the loss
of heat taking place more through circulation than through conductivity. On the
other hand Benedicks contends that the rate of flow has very little influence.

Structural Changes on Hardening. Bearing in mind the enormous differ-
ence between, the properties of hardened steel and those of the same metal un-
hardened, we should naturally expect to find the structure of hardened steel likewise
totally different from that of unhardened steel. And so indeed it is as shown in
Figure 1, in the case of hardened steel containing some 0.50 per cent carbon. To
account for this structure let us remember that, initially, this steel consisted of an
aggregate of ferrite and cementite which, upon being heated through its critical

LESSON XIII THE HARDENING OF STEEL 3

range, was converted into a solid solution (austenite) of carbon or carbide in gamma
iron. This was necessary to impart hardening power. Had the metal been allowed
to cool slowly through its critical range it would have been converted back into a
mixture of ferrite and cementite. On rapid cooling, however, this transformation
was prevented, at least in part, the time necessary for its completion having been
denied. A conclusive evidence that the transformation does not occur in its entirety is
afforded by the absence of a marked critical range on quick cooling. If the transforma-
tion of the solid solution could be effectively prevented austenite should be the constit-
uent of hardened steel. In the commercial hardening of steel, however, the cooling
is not sudden enough to prevent at least a partial transformation of austenite, not
into ferrite and cementite but into a more or less transitory form, marking the first
step of that transformation and called "martensite." Very frequently the rate of

Fig. 2. Steel. Carbon 1.57 per cent. Mag-
nified 1000 diameters. Heated to 1050 deg.
C. and quenched in ice-water. (Osmond.)

cooling is not sufficiently rapid to prevent the martensite from further partial '-ans-
formation into a second transition constituent known as "troostite." Martensite and
troostite, then, are the ordinary constituents of commercially hardened steel. It will
now be profitable to consider at some length the occurrence, nature, and properties
of the three constituents chiefly concerned in dealing with hardened steel, namely,
austenite, martensite, and troostite.

AUSTENITE

Nature of Austenite. 1 Austenite is universally considered as a solid solution
of carbon or, more probably, of the carbide Fe 3 C in gamma iron. 2 All steels above
their critical range are made up of this solid solution. It follows that the carbon
content of austenite varies, like that of steel, from a mere trace to some 1.75 or 2 per
cent. It is not, therefore, a constituent of constant composition.

1 This name was suggested by Osmond in honor of the late Sir William Roberts-Austen. Aus-
tenite has also been called mixed crystals and gamma iron and by some writers, wrongly, martensite.

2 Arnold believes that austenite is the carbide Fe 24 C (hardenite) holding in solution ferrite in
hypo-eutectoid steel and cementite in hyper-eutectoid steel.

4 LESSON XIII THE HARDENING OF STEEL

Occurrence of Austenite. While present in all steels above their critical range
austenite is very rarely found in ordinary steels cooled to atmospheric tempera-
ture. This is due to the rapidity with which austenite is transformed on cooling
through the critical range if not into an aggregate of ferrite and cementite, at least
into some transition stages. In the commercial hardening of ordinary carbon steel
the passage of the metal through its range is never sufficiently rapid to retain in the
cold a small amount even of undecomposed austenite. To prevent the transforma-
tion of a portion of the austenite the conditions generally affecting the hardening of
the metal must be, so to speak, greatly intensified: (1) the steel should be highly car-
burized, (2) quenching should be from a high temperature (1000 deg. C. or more),
and (3) a very effective quenching bath should be used such as ice-cold water. In

Fig. 3. Steel. Carbon 1 .57 per cent. Magnification not stated. Heated to
1050 deg. C. and quenched in ice-water. (Osmond.)

Figure 2 is shown, after Osmond, the structure of steel containing 1.57 per cent carbon
heated to 1050 deg. C. and quenched in ice-water. The magnification is 1000 diam-
eters. The dark-colored, zigzag constituent is martensite; the light matrix, or back-
ground, is austenite. The structure of the same steel, under lower magnification, is
seen in Figure 3. By the drastic quenching treatment just described it is possible,
in the case of high carbon steel, to retain more than one half of the steel in its aus-
tenitic condition.

The retention of austenite in the cold is greatly helped by the presence of some
elements such as manganese and nickel which lower the position of the transforma-
tion range, eventually depressing it below atmospheric temperature and, therefore,
causing the steel to remain austenitic even after slow cooling. This actually takes
place in the presence of some 12 per cent manganese or 25 per cent nickel. When
high carbon steel contains over one per cent of manganese it can be retained in its aus-

LESSON XIII THE HARDENING OF STEEL 5

tonitic condition upon very rapid cooling through its range. In Figure 4, for instance,
is seen, after Robin, the structure of a very small sample of steel (1 to 2 cubic centi-
meters) containing from 1.5 to 1.7 per cent carbon and one per cent manganese
quenched in ice-cold water from a temperature of 1400 deg. C. It consists entirely of

Fig. 4. Steel. Carbon 1.60 percent, manganese 1.00 per
cent. Magnified 300 diameters. Heated to 1400 deg C. and
quenched in ice-cold water (Robin.)

Fig. 5. Steel. Carbon 1.94 per cent, manganese 2.20 per cent.
Heated to 1100 deg. C. and quenched in ice-cold water.
(Maurer.)

austenite. Maurer, likewise, succeeded in retaining in its austenitic condition a steel
containing 2 per cent manganese and 2 per cent carbon by quenching it in ice-cold
water from a temperature of 1100 deg. (Fig. 5). As the manganese increases the
retention of austenite becomes easier, that is, the quenching need not be so drastic nor
the carbon content so high. Finally with 10 or more per cent of manganese and one

6 LESSON XIII THE HARDENING OF STEEL

or more per cent carbon the steel remains austenitic after slow cooling. The struc-
ture and properties of manganese steel will be considered in another lesson.

To sum up: (1) austenite is never produced in the commercial hardening of
ordinary carbon steel; (2) it may be retained in the cold, however, associated
with considerable martensite in quenching very high carbon steel, from a very high
temperature in ice-cold water, as, for instance, by quenching steel containing not less
than 1.50 per cent carbon from 1000 deg. C. or higher; (3) in the presence of 1 per
cent of manganese very small pieces of very high carbon steel may be retained wholly
in their austenitic condition by quenching them from a very high temperature, as,
for instance, by quenching in ice-cold water from 1400 deg. C. small pieces of steel
containing 1 per cent of manganese and .not less than 1.5 per cent carbon (Robin);
(4) with increasing proportions of manganese the transformation of austenite may
be prevented in steel containing less carbon and quenched from lower temperatures
(Maurer); (5) manganese steels containing, for instance, 10 or more per cent manganese
and one or more per cent carbon remain austenitic after slow cooling; (6) nickel
steel containing some 25 per cent of nickel likewise remains austenitic on slow cooling.

Benedicks contends that in the preservation of austenite in carbon steel by
rapid cooling an important part is played by the very great pressure to which the
metal is subjected, (1) because of the shrinkage of the exterior portion or outer shell
on the interior and (2) because of the dilatation accompanying the change from
gamma to beta iron. Were it not for this pressure Benedicks believes that the trans-
formation of austenite could not be prevented. As an evidence of this he shows that
austenitic steels produced by quenching are austenitic only in their interior, i.e. where
the pressure had been greatest, the outside layers in which the pressure was small or
nil being martensitic. He shows, further, that on removing by grinding the marten-
sitic shell the austenitic core, in turn, becomes martensitic owing to the removal of
the pressure exerted upon it by that shell. Again the quenching of steel cylinders
surrounded by cast iron shells resulted in the formation of austenite close to the
skin of the steel cylinders owing apparently to the very great pressure exerted upon
the steel by the contraction of the iron shells.

Etching of Austenite. The etching reagents usually applied to bring out the
structure of unhardened steel, namely, picric acid, nitric acid, tincture of iodine, etc.,
do not always yield satisfactory results in the case of hardened steel. Kourbatoff
discovered a complex reagent which often produces greater contrasts between the
various constituents. It is made up by mixing one part of amyl alcohol, one part of
ethyl alcohol, one part of methyl alcohol, and one part of a 4 per cent solution of
nitric acid in acetic anhydride and should be prepared just before use.

Heyn recommends for etching hardened steel a solution containing one part of
hydrochloric acid and 99 parts of absolute alcohol. More uniform results are ob-
tained if a weak current of electricity be passed through the solution, the samples to
be etched forming the positive pole while the negative electrode may consist con-
veniently of a piece of sheet lead. With the assistance of the electric current the use
of a very dilute aqueous solution is advisable, namely, one part of hydrochloric acid
in 500 parts of distilled water.

Osmond, likewise, used successfully a solution of 10 per cent of hydrochloric acid
in water by which the martensite is colored darker than austenite, the treatment re-
quiring several minutes. Osmond writes: "There is more regularity obtained by
having the specimen connected, by means of a platinum wire, with the positive pole
of a bi-chromate cell, a strip of platinum placed in the acid being connected with the

LESSON XIII THE HARDENING OF STEEL 7

negative pole. In this way the specimen becomes the anode, and the platinum the
cathode."

Benedicks recommends for the etching of martenso-austenitic steel a 5 per cent
alcoholic solution of metanitrobenzol-sulphonic acid which always darkens martensite
more than austenite. Immersions of some fifteen seconds are generally sufficient.

Structure of Austenite. When austenite and martensite occur in the same sam-
ple the latter is generally colored darker than the former (Figs. 2 and 3). Martensite,
moreover, is readily distinguishable because of its zigzag or needle shape. Some
writers claim that martensite is sometimes colored less than austenite. Indeed Maurer
contends that this is always so, arguing that if most photomicrographs indicate the
contrary it is because the martensite had. undergone a certain amount of tempering
resulting in the formation of some troostite as later explained. According to this
writer, in order to prevent any tempering of the martensite, and therefore the forma-
tion of dark-colored troostite, great care must be exercised in sawing, polishing, etc.
To this Benedicks replies that it cannot always be so for in quenching austenite in
liquid air martensite is formed which must be free from troostite and which, never-
theless, is darker than austenite. It may be asked, however, whether it is certain
that the martensite produced in this way is actually free from troostite. Evidences
of a more conclusive nature are needed to account satisfactorily for the shifting in
the relative coloration of austenite and martensite when occurring side by side. Pure
austenite is made up of polyhedric grains l (see Figs. 4 and 5) which, as explained in
previous lessons in connection with the structure of gamma iron, are undoubtedly
made up of true crystals, small octahedra according to Osmond. It should be noted
that when austenite occurs in the presence of much martensite (Fig. 2) its polyhedric
structure is not brought out. Twinnings are frequently observed in austenite (see
Lesson II, Fig. 11) although it has been claimed that they form only after straining,
especially if followed by annealing.

Baykoff succeeded in etching austenite above the critical range of the steel, that
is, in a range of temperature where it is stable. He accomplished this by heating
polished steel samples in a porcelain tube through which a current of hydrogen was
kept circulating and by passing through it gaseous hydrochloric acid when the de-
sired temperature had been obtained. The resulting structures were found to be
polyhedric even in the presence of very little carbon, thus confirming the previous
belief as to the crystalline character of austenitp.

Properties of Austenite. Since the carbon content of austenite varies from a
mere trace to nearly 2 per cent it may well be expected that its physical properties
will likewise vary, i.e. that it will increase in hardness and strength and decrease in
ductility as the carbon increases. Osmond has shown conclusively that austenite
was softer than martensite of identical carbon content. When it is remembered that
in order to produce austenite in ordinary carbon steel all the factors generally in-
creasing the hardness of the metal must be intensified, it is at first surprising that
the energetic quenching treatment required should yield a softer metal. The con-
clusion must be that gamma iron is softer than beta iron. This relative softness of
austenite is well shown by Osmond in Figure 6 which represents the structure of a
bar of steel containing 1.55 per cent carbon in the center and a gradually decreasing
amount towards the outside. This bar was heated to 1050 deg. C. and quenched in

1 Because of this structure Guillet and some other writers refer to steels composed of austenite
:i- polyhedric" steels. This does not seem advisable as it may lead to confusion, for other steels
also have polyhedric structures, to wit, very low carbon (ferritic) steels.

LESSON XIII THE HARDENING OF STEEL

mercury at a temperature of 9 cleg. C. After polishing hut before etching a needle
was repeatedly drawn across it from end to end with even pressure. The photograph
clearly shows that the needle scratched the steel (1) where it contains so little carbon
(0.40 to 0.60 per cent) that it was only partly martensitic and hence relatively soft,
(2) in those regions which because of very high carbon content (1.30 to 1.55 per cent)

Carbon
per cent

0.8C
0.90
1.00
1.10
1.20
1.30
1.40

1.50
1.55
1.50
1.40
1.30
1.20
1.10
1.00
0.90
0.80
0.70
0.60
0.50

0.40
0.36

0.35

. Hardenite and
Martens! te

Austenite and
Hardenite

Hardenite and
' Martensite

Fig. 6. Showing the relative softness
of austenite.

were partly austenitic, and (3) that it failed to scratch it in those regions which because
of a more moderate amount of carbon (0.70 to 1.20 per cent) were fully martensitic
and, therefore, very hard. It is also well known that high carbon austenitic manga-
nese steel, while extremely difficult to machine, can be readily scratched by a needle,
being mineralogically softer, therefore, than high carbon, martensitic steel. Rosen-
hain and Humphrey have shown that above the critical range austenite (gamma
iron) was much softer than beta iron. Since steel above its critical range is non-

LESSON XIII THE HARDENING OF STEEL 9

magnetic we should expect steels which remain austenitic in the cold to be non-mag-
netic. This we know to be the case, for manganese as well as nickel austenitic steels
are non-magnetic.

Some of the physical properties of austenite may be inferred from the known
properties of austenitic steels such as manganese and high nickel steels. These are
known to be very ductile (after suitable heat treatment), tenacious, of low elastic
limit, to possess very high resistance to wear although their mineralogical hardness is
not excessive and to be machined only with great difficulty. They have, like gamma
iron, a very high electrical resistance.

It has already been pointed out that the crystallization of austenite is probably

Fig. 7. Austenitic steel quenched in liquid air. Magnified 250
diameters. (Osmond.)

cubic, the octahedron being its prevailing crystalline form. Le Chatelicr, however,
believes that austenite crystallizes in the orthorhombic system with octahedral cleav-
age. On slow cooling through the critical range in the absence of considerable quan-
tities of retarding elements such as manganese and nickel, austenite rejects a sufficient
amount of ferrite in hypo-eutectoid, or of cementite in hyper-eutectoid, steel to assume
the eutectoid composition (0.85 per cent C. or thereabout) when it is converted bodily
into pearlite. This transformation is not sudden, however, several transition con-
stituents being formed, namely, martensite, troostite, and sorbite.

It will be seen in another lesson that on tempering austenite, that is, on reheating
it below the critical range of the metal it is likewise converted gradually and succes-
sively into martensite, troostite, and sorbite or according to some writers directly
into troostite and then into sorbite.

Quenching austenite in liquid air results in the formation of martensite with in-
creased volume causing swellings of the polished surface as shown in Figure 7.

10 LESSON XIII THE HARDENING OF STEEL

MARTENSITE

Nature of Martensite. 1 It is very generally believed that martensite corres-
ponds to an early stage in the transformation of austenite in passing through the
critical range. Opinions differ, however, as to its exact nature. Accepting the possi-
bility of iron existing under three allotropic forms, namely, as gamma, beta, and alpha
iron, and the carbon under two distinct conditions, namely, as the crystallized car-
bide FesC or cement carbon and of this carbide or possibly elementary carbon being
dissolved in iron, i.e. as hardening carbon, what are the probable conditions of these
two constituents in martensite? Osmond and many others believe that in martensite
iron is present chiefly in its beta condition, holding carbon in solution, hence the
great hardness of that constituent. Since martensite is magnetic, however, it must
also contain an appreciable quantity of magnetic alpha iron. This theory, which may
be called the allotropic theory, is the one most widely held. Le Chatelier, not believ-
ing in the existence of beta iron, considers martensite as essentially a solid solution of
carbon in alpha iron, owing its great hardness to its state of solid solution and its
magnetism to the presence of alpha iron. Edwards contends that austenite and
martensite are in reality the same constituent, namely, a solid solution of carbon in
gamma iron, differing only in structural aspect, the needles of martensite resulting
from the twinning of austenite caused by the severe pressure exerted upon it during
rapid cooling. He bases his view chiefly upon the absence of the point A 2 in medium
high and high carbon steels, from which he infers that beta iron does not form in
those steels, losing sight of the fact that the points As. 2 and Aa. 2 .i may very well,
and probably do, include the A 2 changes. Kroll also speaks of martensite as repre-
senting the "mutilated structure of austenite due to twinning." Arnold believes that
martensite is, like austenite, the carbide Fe 24 C holding in solution ferrite in hypo-
eutectoid steel and cementite in hyper-eutectoid steel.

Careful consideration of the evidences at hand leads to the adoption of the first
theory (Osmond's) as the one best supported. That martensite is to a great extent
a solid solution seems evident from the fact that it contains a great deal of hardening,
i.e. dissolved carbon, and it seems probable that beta iron is the solvent, for if gamma
iron were the solvent it would not explain the greater hardness of martensite com-
pared to that of austenite while we have good reason to doubt the power of alpha iron
to dissolve carbon seeing that below the critical range, i.e. when in its alpha form,
iron will not absorb carbon.

Occurrence of Martensite. Martensite is most readily obtained through the
quenching of small pieces of high carbon steel in cold water; in the case of largo
pieces, while the outside portion may be martensitic their center is likely to be partly
troostitic. In low carbon steel it is more difficult still to prevent the formation of
some troostite while in steel containing very little carbon free ferrite as well is likely
to be present. In very high carbon steel some free cementite is generally associated
with the martensite.

Etching of Martensite. Dilute alcoholic solutions of picric, nitric, or hydro-
chloric acid generally bring out satisfactorily the structure of martensite but the
Kourbatoff reagent, already described, sometimes yields better results. Martensite
generally darkens more quickly than austenite but always remains much lighter than
troostite.

1 This name was selected by Osmond in honor of A. Marteas a distinguished German metallur-
gist and testing engineer.

LESSON XIII THE HARDENING OF STEEL 11

Structure of Martensite. Martensitic structures are shown in Figures 1 and 8.
Osmond describes the structure of martensite as consisting of three systems of fibers,
respectively parallel to the three sides of a triangle and crossing each other frequently.
Osmond also states that when the metal contains less carbon the needles are longer
and more clearly differentiated, other things being equal. According to crystallog-
raphers these markings, in reality cleavages of octahedra, indicate crystallites of
the cubic system and, therefore, afford an additional evidence of the cubic crystalliza-
tion of austenite from which martensite is derived. Osmond and Cartaud refer to
them as probable pseudomorphs of twinnings due to tension, occurring in gamma
iron through partial formation of the bulky beta and alpha modifications.

Properties of Martensite. The carbon content of martensite varies from a mere
trace to as much as one per cent, and possibly more, in very suddenly cooled hyper-
eutectoid steels. In high carbon steels, however, it is difficult to prevent the setting

Fig. 8. Steel. Carbon 1.25 per cent. Mag-
nified 150 diameters. Heated to 1232 deg.
C. and quenched in oil. (C. C. Buck, Cor-
respondence Course student.)

free of much of the excess cementite even on very quick cooling. From this varia-
tion of its percentage of carbon it follows that the properties of martensite must also
vary. As the carbon increases its hardness and strength increase while its ductility
decreases, martensitic steels being generally hard and brittle and, therefore, unforge-
able in the cold.

It will be seen in another lesson that on heating martensite below the critical
range, i.e. on tempering it, it is converted first into troostite and then into sorbite.

TROOSTITE

Nature of Troostite. 1 While most writers believe that troostite represents a con-
dition of the steel resulting from the transformation of martensite and, therefore, a
further step in the transformation of austenite, much difference of opinion exists as
to its exact nature. The controversy has given rise to a very large and apparently
exaggerated amount of discussion. Here, as in the case of martensite, we must con-

1 The name troostite was selected by Osmond in honor of the French chemist Troost.

12 LESSON XIII THE HARDENING OF STEEL

sider the possibility of the iron existing in the gamma, beta, or alpha form or in two
or even all three of these conditions, while the carbon may exist as cement carbon or
as hardening carbon or partly as cement and partly as hardening carbon. Then
the association between iron and carbon may be of the nature of an aggregate or of a
solid solution or partly aggregate and partly solution, or, indeed, half way between
aggregate and solution, namely, resembling a colloidal solution, an emulsion, or an
^uncoagulated substance. Nearly every conceivable hypothesis has been suggested to
account for the nature of troostite. It has been described as a solid solution of
carbon or of carbide in gamma iron, in beta iron, and in alpha iron. It has also been
suggested that it might be pure beta iron.

In later years, thanks chiefly to the enlightening discussions of Benedicks sup-
ported by the weighty evidence of skilfully conducted experiments, metallographists
have come to regard troostite as an uncoagulated mixture of the constituents of mar-
tensite and sorbite, that is, of (1) carbide dissolved in beta iron, (2) crystallized Fe 3 C.
and (3) crystallized alpha iron clearly martensite passing to sorbite. Benedicks
compares it to a colloidal solution 1 while Arnold had previously described it as
"emulsified" pearlite. 2 The existence of considerable dissolved (hardening) carbon
in troostite is proven by analysis as well as the existence of considerable crystallized
Fe 3 C (cement carbon). Its relatively great hardness points strongly to the presence
of a considerable amount of beta iron while its magnetism demands the presence of
alpha iron. Benedicks' hypothesis is consistent with what we know of the formation
of troostite and of its properties.

In the report of the Committee on the Nomenclature of the Microscopical Con-
stituents of Iron and Steel of the International Association for Testing Materials,
troostite is defined as follows: "probably aggregate. In the transformation of aus-
tenite, the stage following martensite and preceding sorbite . . . An uncoagulated
conglomerate of the transition stages."

Occurrence of Troostite. In order to produce troostite on cooling steel from
above its critical range, it is necessary that the cooling through the range should
be so regulated as to allow it to form and at the same time prevent its further
transformation (into sorbite and pearlite). These conditions may prevail (1) in cool-
ing slowly to the middle of the range, thus permitting the formation of troostite (see
Fig. 13), and then quickly to atmospheric temperature, thus retaining troostite, and
(2) in cooling through the range at a rate uniform throughout but so regulated as to
cause the production and retention of troostite (Fig. 13) as, for instance, quenching
large pieces in water when the central portions at least will be troostitic, or quench-
ing smaller pieces in oil. It will be explained in the next lesson that troostite may
also be produced by tempering (i.e. reheating below the critical range) austenitic
and martensitic steels.

Troostite is readily produced by heating a bar of steel, containing 0.50 per cent
carbon or more, white hot at one end and quenching it in water, when at some
distance from the heated end the temperature must necessarily have been such as to
produce troostite. This can generally be detected by means of a file, the martensitic

1 A colloid may be regarded as a substance passing from the state, of solution to that of an
aggregate or vice versa; it is no longer a solution but not yet an aggregate. To express it more scien-
tifically, while not a true solution the particles of solvent and solute are ultra-microscopic. Accord-
ing to Le Chatelier so-called colloidal solutions are in no way solutions, but merely liquids holding
in suspension very finely divided particles; the expression, he says, should not' be used.

2 "Emulsified carbide present in an excessively fine state of division in tempered steels." (1895.)

LESSON XIII THE HARDENING OF STEEL

portion of the bar being too hard to be marked while the troostitic part, although
hard, can be scratched. The sorbitic and pearlitic portions are decidedly softer.

Properties of Troostite. It will be obvious from the foregoing description of the
nature and formation of troostite that its physical properties must be intermediate
between those of martensite and of sorbite. For like carbon content troostite is
softer and more ductile than martensite but harder and less ductile than sorbite. It
will be shown that at some 400 deg. C. it begins to be transformed into sorbite.

Etching of Troostite. Troostite is colored decidedly darker than any other con-
stituent by the ordinary etching reagents. While dilute alcoholic solutions of nitric,
picric, or hydrochloric acid yield satisfactory results, Kourbatoff s reagent is pre-
ferred by some.

Structure of Troostite. Troostite generally occurs as dark-colored, irregular
areas, representing sections through nodules generally accompanied by martensite or

Fig. 9. Steel. Quenched during critical range.
Magnified 200 diameters. Slightly etched.
Troostite and martensite. (Guillet.)

Fig. 10. Steel. Carbon 0.4o per cent. Magni-
fied 1000 diameters. Troostite and martensite.
(Osmond.)

sorbite or both or as membranes surrounding martensite grains (Figs. 9 to 12). In
hypo-eutectoid steel free ferrite, and in hyper-eutectoid steel free cementite, may also
be present and, indeed, even well-developed pearlite (Fig. 12). Osmond describes the
structure of troostite as "almost amorphous, slightly granular, and mammilated."

Sorbite. Sorbite is not, properly speaking, a constituent of hardened steel. It
seems appropriate, however, to again mention it here seeing that, it constitutes the
connecting link between annealed (pearlitic) steels and hardened (troosto-marten-
sitic) steels, and also because it results from the transformation of troostite, thus com-
pleting the various stages assumed by iron carbon alloys in passing from the condition
of austenite, stable above the range, to that of pearlite, stable below that range.
These stages are (1) austenite, (2) martensite, (3) troostite, (4) sorbite, and (5)
pearlite.

Sorbite is now generally regarded as an uncoagulated mixture of the constituents
of troostite and of pearlite; it apparently contains (1) some hardening carbon, i.e.
carbon or FesC dissolved in beta iron, hence the greater hardness and strength of
sorbite compared to the hardness and strength of pearlite, (2) a considerable quan-

LESSON XIII THE HARDENING OF STEEL

tity of alpha iron, hence its magnetism and relative softness, and (3) a considerable
quantity of crystallized FesC (cement carbon) as proven by analysis. While sorbite
probably contains the same constituents as troostite it holds considerably less unde-
composed solid solution and considerably more alpha iron, hence it is much softer

Fig. 11. Steel. Carbon 0.54 per cent. Magni-
fied 100 diameters. Troostite and martensite.
(Boynton.)

Fig. 12. Steel. Carbon 0.54 per cent. Magnified 1000 diameters. Martensite, troostite,

sorbite, and pearlite. (Boynton.)

and more ductile than troostite. In other words the transformation which eventually
must lead to the formation of pearlite is more advanced in sorbite than it is in troost-
ite. The nomenclature committee, already referred to, describes sorbite as follows:
"Aggregate ... In the transformation of austenite, the stage following troostite
. . . and preceding pearlite. Most writers believe it essentially an uncoagulated
conglomerate of irresoluble pearlite with ferrite in hypo- and cementite in hyper-
eutectoid steels respectively."

LESSON XIII THE HARDENING OF STEEL 15

The occurrence, etching, structure, and properties of sorbite have been described
in Lessons XI and XII when it was shown that it is formed (1) in small pieces of steel
cooling in the air from above their critical range, (2) in larger pieces quenched in oil
from above the range, or (3) in small pieces quenched in water from near the bottom
of the range. In other words to form sorbite we must so regulate the cooling through
the critical range that it is allowed to form but prevented from further transforma-
tion (into pearlite). It will be seen in the next lesson that sorbite is also formed on
tempering austenitic, martensitic, and troostitic steels.

By its physical properties sorbite occupies an intermediate position between
troostite and pearlite; as previously mentioned it is stronger, harder, and less ductile
than pearlite but softer and more ductile than troostite.

Troosto-Sorbite. Kourbatoff gives the name of troosto-sorbite to a constituent
associated with martensite and austenite in quenching, from a high temperature, steels
very high in carbon. It is not clear that this constituent is more than a mixture of
troostite and sorbite. We may talk of troosto-sorbite as we do of a greenish blue
tint to indicate shades intermediate between green and blue, and similarly the ex-
pressions martenso-austenite, troosto-martensite, and sorbitic-pearlite, or like expres-
sions, are useful and their meanings obvious. In the report of the Committee on the
Nomenclature of the Microscopical Constituents of Iron and Steel of the International
Association for Testing Materials, troosto-sorbite is thus denned: "Indefinite aggre-
gate, the troostite and the sorbite which lie near the boundary which separates these
two aggregates."

Hardenite. The name of hardenite is frequently given both to austenite and to
martensite of eutectoid composition, 1 i.e. to the original austenite of eutectoid steel
and to the residual austenite of hypo- and hyper-eutectoid steel after rejection of the
full amount of free ferrite or of free cementite. In other words the name is applied

(1) to the condition of austenite in slowly cooled steels immediately preceding its
conversion into martensite and (2) to the resulting martensite (necessarily of eutec-
toid composition if the cooling to the range has been sufficiently slow). It is unfor-
tunate that the same term is used to designate both austenite and martensite, two
apparently sharply different constituents, as it is likely to lead to confusion. Its use
should be restricted to the designation of austenite of eutectoid composition. Giving
it this meaning it will be apparent, as later explained, that hardenite possesses maxi-
mum hardening power and, therefore, that steel made up exclusively of hardenite, i.e.
eutectoid steel, possesses maximum hardening power.

Rate of Cooling through Critical Range vs. Structure of Steel. It has been made
clear in the foregoing pages (1) that in order to retain some austenite in the cold the
metal should be highly earburized and very quickly cooled from a high temperature,

(2) that pearlite is produced by very slow cooling through the critical range, and

(3) that in order to cause the formation of any of the three recognized transition con-
stituents, namely, martensite, troostite, and sorbite, the steel should be cooled through
its critical range in such a way as to allow the formation of the desired constituent
while preventing its further transformation as, for instance, (a) by cooling the metal
slowly to that portion of the range in which the constituent is formed and then quickly
to atmospheric temperature or (6) by cooling the metal through its range at a uni-

1 Originally the name hardenite was applied by Howe to austenite and martensite of any com-
position (18S8). Osmond used it to designate austenite saturated with carbon (1897). Both these
m>;aninu;s have been withdrawn by their proposers. Arnold calls hardenite the carbide Fe^C which
he believes exists above the critical range.

LESSON XIII THE HARDENING OF STEEL

form speed but so regulated that the transformation of austenite proceeds only to
the desired extent, to wit, cooling in water for martensite, in oil for sorbite.

An attempt has been made in Figure 13 to give a graphical illustration of the
cooling conditions needed for the production of the various constituents of steel.
Its interpretation will be obvious. The critical range, or rather the lower critical
point, Ari or Ar 3 . 2 .i is represented as covering a considerable range of temperature
so as to afford the necessary room for the diagrammatical representation of the for-
mation, within that range, of the transition constituents. The diagram indicates that
as the metal cools slowly through its range it does not pass abruptly from an austenitic
to a martensitic condition and then to troostite, etc., but that these transformations

X
P

/ / /

s T M A

Avsfenite

A M T

A = Austenite
M = Martensite
T = Troostite
S = Sorbite
P =Pearlite

Fig. 13. Diagram depicting the formation of austenite, martensite, troostite, sorbite, and pearlite

in steel cooling through its critical range.

are, on the contrary, gradual, the following types of structure being formed, theoreti-
cally at least : austenite, austenite plus martensite, martensite, martensite plus troost-
ite, troostite, troostite plus sorbite, sorbite, sorbite plus pearlite, and pearlite. The
transformations depicted refer to eutectoid steel or to the residual austenite (neces-
sarily of eutectoid composition) of hypo- and hyper-eutectoid steel formed on slow
cooling to Ari after rejection of free ferrite or free cementite. In the cases of these
steels, therefore, free ferrite or free cementite is present in the above structures,
unless, indeed, cooling between Ar 3 and Ari or between Ar cm and Ari has been so rapid
as to prevent their separation. While, theoretically, very quick cooling from a to at-
mospheric temperature should retain the steel in its austenitic condition, even under
the most favorable conditions, this can be done but partially, considerable martensite
being produced. Cooling slowly to m and then quickly should produce martensite

LESSON XIII THE HARDENING OF STEEL 17

while slow cooling to t or s followed by quick cooling should produce, respectively,
troostite and sorbite. Slow cooling to p, followed or not by quick cooling, results in
the formation of pearlite. Slow cooling to intermediate points between m and t or
t and s, etc., should, theoretically, cause the formation of martensite and troostite,
troostite and sorbite, etc.

These conditions may be realized in the same piece of steel by heating one end of
a steel bar, preferably of eutectoid or hyper-eutectoid composition, well above the
critical range and quenching the whole bar in ice-cold water. It is evident that, since
at the time of quenching the temperature of the bar decreased gradually from the
hot to the cold end, a portion of the bar must have been quenched while in the mar-
tensitic condition, another while in the troostitic condition, etc. The preparation
and microscopical examination of longitudinal sections through the center of the bar
should reveal the existence of the various constituents indicating as many stages in
the transformation of austenite.

On the left of the diagram five lines starting from the point R above the criti-
cal range represent coolings through the range at different speeds as, for instance,
(1) very quickly in ice-cold water, (2) quickly in water, (3) less quickly in oil, (4)
slowly in air, and (5) very slowly in furnace, resulting, respectively, in the forma-
tion of austenite (or rather austenite and martensite), martensite, troostite, sor-
bite, and pearlite or, more frequently, of mixtures of two or even three of these
constituents.

These conditions may be realized in the same piece of metal by heating a steel
bar of considerable cross section (not less than one inch in diameter) and preferably
of eutectoid or hyper-eutectoid composition to a temperature well above the critical
range and quenching it in water. The cooling should be rapid enough to cause the
formation of martensite near the outside of the piece while the cooling of the center
should be so slow (because of the size of the bar) as to permit the formation of pear-
lite. Between the martensitic shell and the pearlitic core the metal should be com-
posed of the other transition constituents, that is, starting from the center, of sorbite
and then of troostite. The microscopical examination of a cross section through the
bar should reveal this gradual change of structure from center to outside.

The formation of a transition constituent through the tempering of a constituent
representing a less advanced, and therefore less stable, stage of transformation has
already been alluded to and will be considered at greater length in the next lesson.

Are the Transition Stages Distinct Constituents? It would appear from our con-
sideration of the formation of the transition constituents that they must represent
as many stages in the progressive transformation of austenite and that sharp lines of
demarcation between them are not to be expected or, rather, that they must be linked
together by an unbroken chain of transition stages just as the blue and yellow colors
are connected through an unbroken series of bluish, greenish, and yellowish shades.
This logical inference, however, is not supported by microscopical evidences in the
cases of austenite, martensite, and troostite for these constituents are sharply differ-
entiated from each other whenever they occur together. Stages or structural condi-
tions representing the gradual transformation of austenite into martensite or of
martensite into troostite are not observed; it is as if these transformations had actu-
ally taken place by rather sudden steps.

The transformations of troostite into sorbite and of sorbite into pearlite, on the
contrary, appear to be very gradual and easy to follow. In other words while troost-
ite, sorbite, and pearlite are readily distinguishable, each having marked charac-

18 LESSON XIII THE HARDENING OF STEEL

teristics of its own, structural arrangements are frequently observed which undoubtedly
correspond to intermediate stages.

The five constituents of steel, resulting from various modes and rates of cooling',
might be compared to a spectrum of five elementary colors, i.e. violet, blue, yellow,
orange, and red, with gaps existing between the first three (our austenite, martensite,
and troostite) while the third and the fifth, yellow and red (troostite and pearlite)
are closely linked together by the fourth, orange (our sorbite), and an infinity of in-
termediate shades.

Metarals and Aggregates. Howe suggests dividing all microscopical constit-
uents of iron and steel into (1) metarals and (2) aggregates.

These he describes as follows: "Metarals, true phases like the minerals of nature.
These are like definite chemical compounds, or solid solutions, and hence consisting
of definite substances in varying proportions . . . Aggregates, like the petrographic
entities as distinguished from the true minerals. These mixtures may be in definite
proportions, i.e. eutectic or eutectoid mixtures (ledeburite, pearlite, steadite) or in
indefinite proportions (troostite, sorbite)." 1 Under these two headings the constit-
uents of iron-carbon alloys would be classified as follows:

Metarals: ferrite, cementite, austenite, graphite.

Aggregates: pearlite, sorbite, troostite, ledeburite, steadite.

Opinions differ as to the nature of martensite; if it is a solid solution it is a metaral,
if not it must be classified with the aggregates. Should we recognize the existence
of solid colloidal solutions, it is not clear whether these should be grouped with the
metarals or should form a distinct class between the metarals and the aggregates.

Hardening Eutectoid Steel. It will now be profitable to consider separately
the hardening of eutectoid, hyper-eutectoid, and hypo-eutectoid steel. In hardening
eutectoid steel the metal should be heated through its critical range, i.e. through its
single critical point Ac 3 . 2 .i. By so doing we confer upon it full hardening power and
finest possible structure. The steel should then be cooled from that temperature as
promptly as possible avoiding heating much above the range or long exposure at any
temperature above the range, as either procedure would tend to increase the grain
size of the metal. A temperature of some 775 to 825 deg. C. will generally be, there-
fore, the best temperature to which to heat and from which to cool eutectoid steel for
the purpose of hardening. By this treatment the steel passes from a finely austenitic
to a finely martensitic or troosto-martensitic condition (Fig. 8).

Hardening Hyper-Eutectoid Steel. Let us assume a steel containing 1.25 per
cent carbon and, therefore, composed approximately of 93 per cent of pearlite and
7 per cent of free cementite. Upon heating this steel through its lower point Ac 3 .2.i its
93 per cent of pearlite are converted into 93 per cent of austenite possessing hardening
power, but the metal still contains 7 per cent of free cementite deprived of hardening
power. If we heat it past its upper point Ac cm the free cementite is absorbed and the
whole mass becomes austenitic. A little reflection will show, however, that the steel
should be quenched as soon as it rises above the point Ac 3 .2.i, because we then pro-
duce, theoretically at least, 93 per cent of fine grained martensite while retaining, to
be sure, the original 7 per cent of cementite, but as this constituent is harder than

1 Howe further writes: "Many true minerals, such as mica, felspar, and hornblende, are divisible
into several different species. Such minerals are definite chemical compounds, in which one element
may replace another. Others, such as obsidian, are solid solutions in varying proportions and in these
also one element may replace another. Metarals like minerals differ from aggregates in being sever-
ally chemically homogeneous."

LESSON XIII THE HARDENING OF STEEL 19

martensite its presence adds to, rather than takes away from, the hardness of the
quenched metal. Should we, on the contrary, heat to above Ac cm before quenching
the whole mass would be converted into martensite but it would be less hard if any-
thing than the metal quenched at a lower temperature while its structure would be
coarser and the danger of cracking the objects in the quenching bath would be greater.
The best hardening temperature for hyper-eutectoid steels, therefore, is the same as
that for eutectoid steel, namely, some 775 to 825 deg. C.

The structure of a properly hardened hyper-eutectoid steel is shown in Figure 14.
Like hardened eutectoid steel it consists of very fine martensite.

Hardening Hypo-Eutectoid Steel. Let us take a steel containing some 0.50 per
cent carbon and exhibiting therefore the points AI and A 3 . 2 . ~S'urh steel is made up

Fig. 14. Steel. Carbon 1.10 per cent. Magnified 100 diameters. Quenched
in water from above its critical range. (Boylston.)

of 60 per cent of pearlite and 40 per cent free ferrite. Upon heating it through its
Aci point the pearlite is converted into austenite, so that at this temperature some
60 per cent of the mass of the metal is endowed with hardening power. Should we
quench this steel, therefore, as soon as it rises above its lower critical point but 60
per cent of its bulk would be hardened ; it would still retain 40 per cent of soft ferrite.
.If, on the contrary, the heating be carried to just above Ac 3 . 2 the free ferrite is absorbed
by the austenite and the whole mass becomes harrienable. Upon quenching the steel
from that temperature its entire bulk may be converted into martensite or troosto-
martensite, according to rate of cooling. While this martensite will not be quite as
hard as the martensite produced by quenching from just above Aci the metal as a
whole will be harder and of a more uniform and finer structure because of the absence
of free ferrite, or at least of any considerable amount of it. It follows from these con-
siderations that for the purpose of hardening, hypo-eutectoid steel should be quenched
from just above its upper critical point, namely, Ac 3 . 2 or Ac 3 (825 to 925 deg. C. ac-
cording to carbon content). Hypo-eutectoid steel containing very little carbon, say

20 LESSON XIII THE HARDENING OF STEEL

less than 0.25 per cent, cannot be very materially hardened by the ordinary quench-
ing methods because of the large amount of soft ferrite which it contains in excess of
the eutectoid ratio and which cannot be retained in solution, even on very quick
cooling (see Lesson XII, Fig. 11). The structure of hardened hypo-eutectoid steel
is shown in Figure 1.

Steel of Maximum Hardening Power. From the above considerations it will be
obvious that the hardening of steel consists in preventing the formation of relatively
soft pearlite and in causing, instead, the formation and retention of hard martensite
or troostite or, more often, of both. It follows from this that the steel possessing
maximum hardening power must be that steel which in slow cooling would contain
most pearlite, namely, eutectoid steel. It does not, of course, mean that quenched
eutectoid steel is harder than quenched hyper-eutectoid steel but merely that the in-
creased hardness produced by quenching is greatest in the case of eutectoid steel.
Quenched hyper-eutectoid steel is harder than quenched eutectoid steel because of
the presence in the former of some free cementite or of more highly carburized mar-
tensite but the difference of hardness between the two steels is greater before quenching.

Hardening Large Pieces. In hardening pieces of considerable cross section it is
evident that the central portions will not cool as quickly as the outside and will not,
therefore, be as hard. Indeed the center may cool so slowly that it will fail to harden
at all. The limitation of the hardening process, as applied to large pieces, will, there-
fore, be evident. It is seldom desirable, however, to harden large pieces to their
very core. When large steel objects are to be hardened, as for instance in the case of
armor plates, superficial hardness only is desired, or at least hardness penetrating to
but a relatively small depth, and this is readily secured through the case hardening
process.

Hardening and Tempering in One Operation. A peculiar but frequent way of
hardening tools consists in heating the tool to the proper temperature and then cool-
ing quickly to a black heat only that portion of it which should be hard, while keep-
ing the other portion out of the bath and, therefore, at a high temperature. Upon
withdrawing the tool from the bath the heat stored away in the hot portion diffuses
to the cooled portion which is in this way reheated, that is, tempered, as later ex-
plained, the amount of tempering being regulated at will by again quenching the
metal as soon as the desired tempering color has been obtained.

Experiments

The following experiments will prove instructive. A steel bar about one half inch
square or round and containing some 0.50 per cent carbon should be heated at one
end, conveniently in a forge, in such a way that the extreme end will be at a very
bright red or yellow heat while its color three inches from the end should not be more
than a very dull red. The whole bar should then be quenched in cold water. Three
pieces one inch long should be detached from the heated end of the bar and a
longitudinal section of each piece polished, etched, and microscopically examined.
Starting with the piece that was heated to the highest temperature the constituents
of hardened steel should be noted in the ordinary order, i.e. martensite, troostite,
sorbite, and in the unhardened portion, pearlite, or association of two or more of
these. Free ferrite will occur, of course, in the pearlitic portion while it may also be
observed, but in smaller quantity, in the sorbitic and even troostitic zones.

Small pieces of hypo-eutectoid, eutectoid, and hyper-eutectoid steels should be

LESSON XIII THE HARDENING OF STEEL 21

heated above their respective critical range and quenched in cold water and some in
oil. Cross sections of all pieces should be prepared for examination which should show
that the pieces quenched in water are martensitic or troosto-martensitic while those
quenched in oil are chiefly troostitic or troosto-sorbitic. In hypo-eutectoid steel free
ferrite and in hyper-eutectoid steel free cementite are likely to occur.

If a sample of steel can be obtained containing not less than 1.50 per cent carbon
and not less than 1 per cent manganese (preferably 2 per cent) a small piece of this
steel should be heated to 1100 deg. C., or higher, and quenched in ice-cold water.
This treatment should produce a martenso-austenitic structure.

Etching. The reagents already described for etching pearlitic and sorbitic steels,
namely, solutions of nitric and picric acid in alcohol, may be employed with generally
satisfactory results for the etching of hardened steels. As these etch much more
quickly, however, the immersions should be correspondingly shorter, especially when
the metal contains troostite. Picric acid being slower in its action than nitric acid is
preferred by some for etching hardened steel. The Kourbatoff reagent may be
tried.

Examination
I. Describe the treatment necessary to impart hardening power to steel.

II. Explain the influence of the rate of cooling through the critical range on the
structure of steel and describe briefly the constituents resulting from cooling
through that range at various speeds.

LESSON XIV

THE TEMPERING OF HARDENED STEEL

Steel that has been hardened by rapid cooling from above its critical range, as ex-
plained in the preceding lesson, is often harder than necessary and generally too brittle
for most purposes. In order to decrease its brittleness, that is, to toughen it without
very material diminution of hardness, the metal is generally " tempered," that is, re-
heated to a temperature considerably below its critical range. This operation is called
tempering because it somewhat mitigates or tempers the effects of the previous harden-
ing treatment.

Tempering Temperatures. The hardening of steel causes increased hardness,
brittleness, and elastic limit, all of which are somewhat lowered by the tempering
operation. The effect of tempering begins to be noticeable at about 100 deg. C.
and increases in intensity as the temperature rises, until finally at some 600 deg. the
metal assumes again the physical properties characteristic of the unhardened con-
dition. The temperature to which hardened steel should be heated for tempering
varies, therefore, with the use to which it is destined. If it is desired to retain the
greatest possible hardness, necessarily with its accompanying brittleness, the steel
should be reheated but slightly above 200 deg. C., as, for instance, in tempering razor
blades when extreme hardness is essential and brittleness relatively immaterial. If,
on the contrary, considerable toughness is indispensable, at the necessary sacrifice of
some hardness the steel should be tempered to some 300 deg. C. or even to a higher
temperature. The great majority of tools are tempered between 200 and 300 deg.

Tempering Colors. Hardened steel objects subjected to tempering being gen-
erally quite bright and their heating being generally conducted in an oxidizing atmos-
phere, very thin films of oxides form upon their surfaces. The colors of these films
vary with the temperature, that is, to each tempering temperature corresponds a cer-
tain color, and blacksmiths generally depend upon these colors for the tempering of
their tools, the use of pyrometers for this operation being far from general. Accord-
ing to Howe the tempering colors and corresponding temperatures are as follows:

Pale yellow 220 deg. C. or 428 deg. F.

Straw 230 " " " 446 ""

Golden yellow 243 """ 469 ""

Brown 255 " " " 491 ""

Brown dappled with purple 265 " " " 509 " "

Purple 277 """ 531 " "

Bright blue 288 " " " 550 ""

Pale blue 297 " " " 567 ""

Dark blue 316 " " " 600 " "

In order that the tempering colors may be plainly seen the steel objects should be
smooth and bright, preferably polished.

Time at Tempering Temperature. It is the common belief that once the desired
temperature is obtained, as indicated by the color, little is to be gained by maintaining

2 LESSON XIV THE TEMPERING OF HARDENED STEEL

the steel at that temperature any length of time on the ground that it will not result
in producing additional tempering. The tempering of steel has been compared to the
releasing of a spring permitting a certain structural rearrangement, that is, a certain
tempering of the metal at any temperature but not more. To produce additional
tempering the temperature must be increased, that is, the spring must be further
released. Recent investigations, however, have shown that the maintenance of hard-
ened steel at a certain tempering temperature often does produce additional temper-
ing effect. It was further ascertained that the color, instead of remaining unchanged
at any given temperature, advances in the tempering color scale as it would with in-
creasing temperature. In other words, the tempering colors, contrary to the view
generally held, are not an absolute criterion by which to judge of the temperature of
the steel, since they vary with the length of time during which the steel is kept at any
temperature. These experiments seem to show, however, that the amount of temper-
ing effected is closely related to the color, that is, that to each shade corresponds a
certain amount of tempering. It should, however, be borne in mind that these colors,
with the corresponding tempering they imply, may be obtained in two ways, (1) through
short exposure at a certain temperature and (2) through longer exposure at lower
temperatures. The same amount of tempering, for instance, would result (a) from
heating hardened steel to 288 deg., when its color is bright blue, followed by immediate
cooling and (6) from heating it to 255 deg., when its color is brown, and maintaining it
at that temperature until its color becomes bright blue. Unless baths kept at con-
stant temperatures are used, however, it is evident that in practise the steel should
be quenched as soon as the desired color is produced and while its temperature is
rising, because it is simpler and more convenient to produce that color on a rising
temperature than by maintaining the metal at a constant temperature.

Some writers doubt the existence of so close a relation between the color and the
resulting tempering. According to Barus and Strouhal to each tempering tempera-
ture corresponds a maximum tempering effect which is the more quickly reached the
higher the temperature. At 100 deg., for instance, the maximum effect was not ob-
tained after one hour although maintaining the steel at that temperature two more
hours had but little additional effect. At 200 deg. the maximum effect was obtained
in ten minutes, while at 300 deg. one minute was sufficient.

Rate of Cooling from Tempering Temperature. Once the desired amount of
tempering is effected, as indicated by the color or otherwise, the rate of cooling to at-
mospheric temperature appears to be quite immaterial. In practise the piece is gener-
ally quenched, merely for convenience. The theory is that while by keeping the
metal at a certain temperature its tempering may be carried farther, on cooling
tempering ceases, for the spring is now tightened, to use the simile already referred
to, so that the rate of cooling is without influence.

Hardening and Tempering Combined. A method of hardening and tempering
combined, frequently employed when but one end of a tool must be hardened as in
the case of chisels and drills, has been described in the preceding lesson. It consists in
heating the tool above its critical range, quenching that portion only which is to be
hard, removing it from the bath, and allowing the heat stored in the unquenched por-
tion to heat by conduction the quenched part until the desired temper color is ob-
tained, when it is again quenched lest the tempering be carried too far.

Explanation of the Tempering of Steel. The theories accounting for the temper-
ing of steel will be considered in the next lesson together with the hardening theories.
It will suffice for the present to point out that hardened steel is generally considered

LESSON XIV THE TEMPERING OF HARDENED STEEL 3

to be in an unstable condition and, therefore, eager to return to a more stable form and
actually undergoing this change whenever given an opportunity, that is on raising
its temperature. At atmospheric temperature the passage of an unstable martensitic
condition into a more stable troostitic or sorbitic form is prevented by the rigidity of
the metal. A slight heating, however, produces some plasticity and such transforma-
tion takes place to a small extent. On increasing the temperature the rigidity dimin-
ishes farther and the transformation advances. The tempering of hardened steel, in
other words, is due to its transformation from an unstable condition to one more
stable. It will be seen that at some 600 deg. C. the metal assumes a stable condition,
i.e. is fully tempered.

Tempering Austenitic Steels. It has been explained in the preceding lesson that
austenite, martensite, troostite, and even sorbite were the constituents formed in
hardened steel according to the rapidity with which the metal is cooled through
and below its critical range. It will now be instructive to consider separately the
tempering of steels having these different types of structure.

Austenitic carbon steel, as already stated, is not a commercial article, as it requires
for its production, at least in the absence of a large proportion of manganese, the
presence of much carbon, an excessively high quenching temperature, and a quench-
ing bath at a very low temperature. And even when these conditions prevail, only
one half or so of the bulk of the steel can be retained in an austenitic condition, the
other half being martensitic. The tempering of austenite should, nevertheless, be con-
sidered. Since in austenitic steel the condition of the metal stable only above the
critical range has been retained in the cold, it follows that cold austenitic steel must
be in a very unstable condition. At atmospheric temperature the rigidity of the
metal is so great that a return to a more stable form is not possible but, on heating it
very slightly, sufficient plasticity is produced to permit a partial transformation of
austenite. This partial transformation, theoretically at least, should result in the
formation of martensite, troostite, and sorbite in the order named as the tempering
temperature increases. This has been depicted in I, Figure 1. In this diagram it is
shown (1) that as soon as austenite is heated above atmospheric temperature it begins
to be converted into martensite, (2) that it is entirely converted into martensite at
200 deg. C., (3) that in heating above 200 deg. martensite begins to pass to troostite,
(4) that at 400 deg. the transformation of martensite into troostite is complete, (5) that
above 400 deg. troostite is gradually converted into sorbite, (6) that at 600 deg. C.
the transformation of troostite into sorbite is complete, and (7) that the sorbite con-
dition is the final condition acquired by hardened steel when reheated close to, but
below, its critical range. While it is certain that lamellar, i.e. true pearlite, cannot be
formed by reheating hardened steel, it is claimed by some that long heating of sorbite
near the critical range, that is, between 600 and 700 deg. C., will result in the forma-
tion of granular pearlite brought about as explained in another lesson by the spheroi-
dizing of the cementite. The tempering of austenite depicted in I, Figure 1, represents
the transformation which, according to theoretical consideration, should be expected
to take place. Most observers, however, report that on tempering austenite it is at once
converted into troostite as soon as the rigidity of the steel has been sufficiently relaxed,
the martensitic stage not being assumed. This is shown diagrammatically in II,
Figure 1. Here, again, it is indicated that at 400 deg. the steel is entirely troostitic
and at 600 deg. entirely sorbitic. Benedicks thinks that it is quite possible that aus-
tenite is first transformed into martensite but that the resulting martensite is so
readily and quickly converted into troostite that its short existence easily escapes

LESSON XIV THE TEMPERING OF HARDENED STEEL

* I

t
I

LESSON XIV THE TEMPERING OF HARDENED STEEL 5

observation. His belief is based upon the following considerations: (1) the austenite
retained in high carbon steel by very sudden cooling is subjected to great pressure
caused by the accompanying martensite having been formed with considerable dilata-
tion, (2) on reheating this martenso-austenitic steel the martensite is first converted
into troostite, because this transformation taking place with contraction is readily
induced, (3) once this transformation started, the pressure upon the austenite is re-
leased and this constituent, in turn, passes first to the martensitic stage (with increase
of volume) and then quickly to the troostitic stage (with decrease of volume) the
martensitic condition being of so short duration as to readily escape detection.

Tempering Martensitic Steel. It has been shown that martensite is generally
present in commercially hardened steel. Since this constituent represents a partial
transformation of austenite it follows that it must be more stable than austenite at
atmospheric temperature. It is sufficiently unstable, however, to be readily con-
verted first into troostite and then into sorbite on tempering as indicated in III,
Figure 1. At 400 deg. the transformation of martensite into troostite is complete,
while at 000 deg. troostite is replaced by sorbite. Bearing in mind the physical prop-
erties of martensite, troostite, and sorbite, it will be readily understood why, on temper-
ing martensitic steel, its hardness gradually decreases while it becomes less brittle and
indeed quite ductile if made sorbitic.

Tempering Troostitic Steel. Commercially hardened steel frequently contains
large proportions of troostite. The tempering of troostite is depicted in IV, Figure 1.
This constituent 1 icing decidedly less unstable than martensite requires greater plastic-
ity, i.e. a higher temperature, before being transformed into a still more stable condi-
tion. Experimental evidences seem to show that the tempering of troostite, i.e. its
transformation into sorbite, requires a temperature of at least 400 deg. and that at
600 deg. the transformation is complete. Since in practise the tempering of steel is
seldom carried above 300 deg. it would seem as if steels made up of troostite do not
need to be tempered, being sufficiently tough. There seems to be no reason to doubt
the accuracy of the above inference. Commercially hardened steels, however, are
generally either entirely martensitic or, more frequently, partly martensitic and partly
troostitic and are in need of tempering because of the large proportion of the exces-
sively hard and brittle martensite they usually contain.

Tempering Troosto-Martensitic Steel. In V, Figure 1, the tempering of hard-
ened steel containing both martensite and troostite has been depicted. It is assumed
that the martensite present begins to be converted into troostite as soon as the tem-
perature of the metal rises, while the transformation of the troostite into sorbite begins
only at 400 deg.

Tempering Troosto-Sorbitic Steel. From the diagram used to illustrate the
tempering of steel it will be apparent that sorbite is relatively so stable a constituent
that its transformation into pearlite cannot be effected below the critical range or, in
other words, that it cannot be tempered. Indeed sorbitic steels not being hardened
steels need not be considered in connection with the tempering operation. A graphi-
cal representation of the tempering of troosto-sorbitic steel, however, has been
included in Figure 1 . It shows that such steel remains unchanged until its tempera-
ture reaches 400 deg. when the troostite it contains begins to be transformed into
sorbite, the transformation being, as usual, complete at 600 deg.

Osmondite. It has been shown that on tempering hardened steel it is entirely
converted into troostite at about 400 deg. C. Below that temperature some marten-
site remains in the structure, while above it some sorbite is present. To the condition

LESSON XIV THE TEMPERING OF HARDENED STEEL

of steel when made up wholly of troostite Heyn gives the name of "osmondite." It
will he apparent that osmondite does not represent a new constituent but merely a
condition assumed by the steel at a certain temperature and the wisdom of giving a
specific name to that condition may well be questioned.

Troostite is more readily colored by the usual etching reagents than any other
constituent of steel, from which it follows that steel made up exclusively of troostite,
i.e. in the osmondite condition, must exhibit maximum coloration. Again Heyn has
shown that the solubility of steel in dilute sulphuric acid increases with the amount
of troostite present, being maximum in steel tempered to about 400 deg. C. From

1000

2.00

Hardened Sfeel

Fig. 2. Diagram depicting the constituents formed (I) on slow cooling, (II) on quick cooling, and

(III) on reheating hardened steel.

these observations Heyn described osmondite as a constituent of steel characterized
by maximum solubility in acids and by maximum coloration under the action of acid
metallographic reagents. In the report of the Committee on the Nomenclature of
the Constituents of Iron and Steel of the International Association for Testing Mate-
rials osmondite is described as follows: "Probably aggregate. That stage in the
transformation of austenite at which the solubility in dilute sulphuric acid reaches its
maximum rapidity. Arbitrarily taken as the boundary between troostite ami sor-
bite . . . The following hypotheses have been suggested, none of which has sub-
stantial experimental foundation: (1) A solid solution of carbon or an iron carbide in
alpha iron; (2) The colloidal system of Benedicks in its purity, troostite being this
system while forming at the expense of martensite, and sorbite being this system

LESSON XIV --THE TEMPER ING OF HARDENED STEEL 7

coagulating and passing into pearlite; (3) The stage of maximum purity of amorphous
alpha iron in the way to crystallizing into ferrite."

Structural Changes on Slow Cooling, Quick Cooling, and Reheating. It seems
helpful and instructive to depict graphically in a single diagram the structural changes
taking place in eutectoid steel (1) on slow cooling through its critical range, (2) on
quick cooling through that range, and (3) on reheating quickly cooled (hardened) steel
above the range. The changes indicated in I (Fig. 2) show, as already explained, that
steel, on cooling slowly through the critical range, is converted successively into mar-
tensite, troostite, sorbite, and pearlite. On heating the same steel from below to
above the range the same changes would take place but in the_reyerse order. In II
the steel has been cooled through the range at such speed that martensite was formed
but prevented from further transformation, hardened martensitic steel being pro-
duced. The reheating of this martensitic steel is depicted in III. Below the range
martensite is gradually converted first into troostite and then into sorbite. On enter-
ing the range the steel remains sorbitic but on further heating the sorbite is converted
back into troostite and then into martensite. Near the top of the range austenite
begins forming, the transformation being complete as the steel emerges from its range.
Similar structural transformations would take place in subjecting hypo-eutectoid
and hyper-eutectoid steels to like treatments, but free ferrite or free cementite would
generally be present.

Microstructure of Hardened and Tempered Steel. The structural changes
corresponding to the transformations taking place on tempering hardened steel de-
scribed in the foregoing pages are not always readily detected by microscopical ex-
amination. This is due to the fact that, structurally speaking, these changes are often
pseudomorphic changes, the crystalline forms of the original constituent or constit-
uents having been retained, although the nature of the crystals themselves has been
altered.

Referring to pseudomorphism Dana writes, "The crystalline forms under which a
species occur are sometimes those of another species." Bayley defines pseudomorphs
as bodies possessing forms borrowed from another substance, or as a body possessing
the form of one substance and the chemical and physical properties of another. In
the formation of a pseudomorph the material of the original substance is replaced
by the new substance but its external form remains unchanged.

According to Heyn the following appearances are observed in the case of eutectoid
steel :

(1) After hardening but before tempering: martensite with well-developed needles
remaining uncolored after an immersion of ten minutes in a solution of one per cent of
hydrochloric acid in alcohol.

(2) After tempering between 100 and 200 deg. : martensitic structure unchanged
but colored yellow or brown.

(3) After tempering to 275 deg. : the needle structure becomes coarser and recalls
mixtures of austenite and martensite, one of the constituents remaining uncolored,
the other assuming a dark coloration.

(4) After tempering to 405 deg. : the needles have disappeared and the sample
appears dark and mottled suggesting a mixture of two relatively dark constituents;
this is troostite.

(5) After tempering to 500 deg. : the light areas become more abundant.

(6) After tempering to GOO deg. : irregular meshes are observed rounded and light,
partially surrounded by a darker network which appear to stand in relief.

8 LESSON XIV THE TEMPERING OF HARDENED STEEL

Photomicrographs of the structure of hardened and tempered steels are repro-
duced in Figures 3 and 4. The structure of hardened steel reheated to 600 deg.
has been shown in Figure 13, Lesson XII.

Carbon Condition in Tempered Steel. It is generally admitted that the com-
bined carbon present in steel may exist under two different conditions, (1) as harden-
ing carbon, that is, as carbon or the carbide Fe 3 C dissolved in iron, and (2) as cement
carbon, that is, as the crystallized carbide FesC or cementite. It is further believed
(a) that pearlite is quite, if not altogether, free from hardening carbon, (6) that mar-
tensite contains, for a given steel, the maximum amount of hardening carbon that can
be produced and retained in that steel by the ordinary hardening operation, and

Fig. 3. Steel. Carbon 0.45 per cent. .Minified 1000 diameters.
Heated to 825 deg. C., quenched from 720 deg., and tempered between
blue and brown (275 deg.?). (Osmond.)

(c) that on tempering martensite the amount of hardening carbon decreases as the
tempering temperature increases, while the proportion of cement carbon increases
correspondingly. Heyn, however, found that on analyzing the residue remaining on
dissolving, in dilute sulphuric acid, hardened eutectoid steel tempered below 400 deg.
it contained no carbon in the form of the carbide Fe 3 C, that is no cement carbon.
For the carbon remaining in the residue and which, in his opinion, is different from
cement carbon, Heyn suggested the notation Cf. Not until a temperature of 400
had been reached on tempering was cement carbon detected in the residue. Heyn
infers from these observations (1) that troostite contains no cement carbon, its resid-
ual carbon being in the hypothetical Cf condition, (2) that osmondite contains the
maximum amount of Cf carbon, and (3) that sorbite must be formed, that is, the steel
must be tempered above 400 deg. in order to produce cement carbon. Heyn further
argues that the deep coloration produced on etching hardened and tempered steel is
caused by the separation of Cf carbon, being, therefore, maximum in osmondite, that
is, when the steel contains nothing but troostite.

LESSON XIV THE TEMPERING OF HARDENED STEEL 9

Osmond very appropriately remarks that it is not necessary in order to explain
Heyn's results to admit the existence of a new form of carbon, it being quite possible
that the Cf carbon of Heyn is cement carbon, that is, Fe 3 C so finely divided when
formed below 400 deg. that it is readily decomposed by the acid, while above 400 it
becomes coarser and, therefore, resists better the action of the acid.

Decrease of Hardness on Tempering. According to Boynton the decrease of
hardness taking place on tempering is gradual up to 350 deg., quite sudden between
350 and 550, and nil above 550 deg.

Heyn found the following values for the loss of hardness on tempering expressed
in per cent of the original increase produced by the hardening operation:

at 100 deg. 2.5 per cent at 400 deg. 70.0 per cent

" 200 " 14.0 " " " 500 " 87.5 " "

" 300 " 41.0 " " " 600 " 97.5 " "

Heyn also observed that the loss of hardness takes place most quickly at 300 deg.

Fig. 4. Steel. Carbon 0.50 per cent. Magnified
500 diameters. Heated to 850 deg. C., quenched in
water, reheated to 400 deg., and quenched in water.
(\V. H. Knight in the author's laboratory.)

Heat Liberated on Tempering. In hardening steel the cooling through and
below its critical range is so rapid that the transformation of austenite which would
have taken place had time been given can proceed but partially, namely, to the mar-
tensitic or troosto-martensitic stage. The heat which would have been generated had
the transformation been complete remains latent in hardened steel. On tempering,
however, the partially suppressed transformation is permitted to proceed farther,
this return to a more stable condition being accompanied by some evolution of heat.
According to Osmond this latent heat can be made apparent by dissolving hardened
steel in double chloride of ammonium and copper when it evolves more heat than
unhardened steel.

Heyn made some careful determinations of the heat generated on tempering
hardened steel. The greater acceleration in heating hardened steel was made apparent

10 LESSON XIV THE TEMPERING OF HARDENED STEEL

by the differential method, using as neutral bodies similar steels in their pearlitie, that
is unhardened, condition. It was observed that the heat generated on tempering is
maximum at 360 deg.

Experiments

A bar of steel 6 inches long and about \ /l 2. inch square or round, preferably of eutec-
toid composition, should be heated well above its critical range and quenched in cold
water. It should then be freed from scale and brightened by rubbing it with emery-
paper, and one end of it heated in the flame of a bunsen burner. At some distance
from the heated end the tempering colors will be observed in their usual order.

Pieces should now be detached from the treated bar corresponding to several
tempering colors, for instance, yellow, brown, and blue, as well as a piece near the
cold end and therefore not tempered. These should be polished, etched, and micro-
scopically examined. In polishing hardened steel great care must be taken to prevent
the heating of the samples, since this would necessarily cause a certain amount of
tempering and therefore of structural transformation. A liberal supply of water
should be used.

Two small pieces of the same steel should be hardened and reheated, one to
400 deg. C. and the other to 600 deg. The first sample should contain a great deal of
troostite, the last should be sorbitic.

All samples should be photographed.

Examination

Describe the tempering (1) of martensitic steel, (2) of troostitic steel, and (3) of
troosto-martensitic steel.

LESSON XV

THEORIES OF THE HARDENING OF STEEL

Many theories have been put forward to explain the hardening_of steel through
sudden cooling from a high temperature. They may be divided into two classes,
(I) the "retention" theories and (II) the stress theory. The retention theories in-
clude (A) the solution theories and (B) the carbon theories. Three solution theories
at least have been proposed, (a) the gamma iron theory, (6) the beta iron theory, and
(c) the alpha iron theory, while two carbon theories should be mentioned, (a) the hard-
ening carbon theory and (b) the subcarbide theory. This classification of the harden-
ing theories is given below in a tabular form, as well as the names of their proposers
and of some well-known scientists supporting them.

(a) Gamma Iron

Edwards

theory

(6) Beta iron or

Osmond, Roberts-Austen,

(A) Solution

allotropic

Howe, and a majority

theories

theory

of writers

(I) Retention

Le Chatelier, Guillet

theories

theory

(a) Hardening

(B) Carbon

carbon theory

theories

(6) Subcarbide

Arnold

theory

(II) Stress

theory

I Andre Le Chatelier
Charpy, Grenet

Retention Theories. The retention theories claim that in hardening steel a con-
dition or set of conditions existing normally above its critical range is retained un-
changed in the cold or but partially changed because of the rapid cooling through and
below the critical range. In other words, such very quick cooling through the range
denies the necessary time for the transformations to take place, at least fully, and a
condition is preserved in the cold which is stable only above or within the critical
range. According to these theories hardened steel, therefore, is in an unstable condi-
tion, hence the possibility of tempering it. That the transformations which would
have taken place on slow cooling through the range are suppressed, partly at least, in
hardening is made evident by the absence of critical points during very rapid cooling
or by the appearance of feeble points only at greatly lowered temperatures. Other
evidences of the partial suppression of the transformations are afforded (1) by the

2 LESSON XV THEORIES OF THE HARDENING OF STEEL

condition of the carbon in hardened steel which is different from the condition of
that element in unhardened steel, (2) by the evolution of heat taking place on tem-
pering hardened steel as explained in Lesson XIV, and (3) by the structure of
hardened steel.

To assume that hardened steel owes its hardness to the suppression, partial at
least, of the transformations taking place on slow cooling through the critical range
is, therefore, both natural and logical.

When we come to look into the nature of the constituents stable above or within
the range, which on being retained in the cold impart extreme hardness to steel, we
find very great differences of opinion among competent authorities.

Solution Theories. The majority of writers believe that hardened steel is in the
condition of a solid solution, quick cooling through the range having prevented the
formation of the ferrite-cementite aggregate. This is strongly supported by micro-
scopical and other evidences. They all agree that the solution consists of carbon dis-
solved in iron but different views are held in regard to the condition of the carbon and
of the iron. While it is now the general belief that the carbide FesC rather than ele-
mentary carbon is dissolved in iron, some think that in hardened steel the iron is
chiefly present as gamma iron, others that it is present chiefly as beta iron, while others
still believe that it exists mainly in the alpha condition. These three contentions will
be briefly considered.

Gamma Iron Theory. Edwards claims that on rapid cooling through the range
iron is retained in the gamma condition. In other words that martensite, the usual
constituent of hardened steel, is identical in composition if not in structure to austen-
ite, the solid solution of carbon in gamma iron stable above the range. His conten-
tion is based solely upon the absence of the point A 2 in medium high and in high
carbon steel from which he argues that beta iron does not form in these steels, dis-
missing the very reasonable possibility of beta iron forming at the points As.2 and
Aa.g.i. Edwards' theory does not explain, at least satisfactorily, the marked difference
of hardness between austenite and martensite nor their totally different structure.
Benedicks, moreover, has shown quite conclusively that austenite cannot exist in
the cold unless subjected to very great pressure, as explained in Lesson XIII.
Razor blades, for instance, although made extremely hard by quenching cannot
be austenitic.

Beta Iron or Allotropic Theory. This theory put forward with great vigor and
brilliancy by Osmond, championed first by the late Roberts-Austen and later by
Howe and many other eminent metallurgists, contends that in hardened steel iron is
present chiefly in the beta condition. It is often referred to as the allotropic theory
of the hardening of steel. It should be borne in mind, however, that Osmond's allo-
tropic theory of iron and his allotropic theory of the hardening of steel are two differ-
ent conceptions. One may believe in the former without being an adherent of the
latter. The allotropic theory of iron claims that iron exists in at least two and prob-
ably in three allotropic forms. No one doubts the allotropy of iron although some
writers, notably Le Chatelier, believe that only two allotropic forms, namely gamma
and alpha iron, have been shown to exist. The allotropic theory of the hardening of
steel claims (1) that in hardened steel carbon, or more probably the carbide FeaC, is
in solution chiefly in beta iron, hence its hardness, beta iron being very hard, (2) Iliat
hardened steel contains alpha iron, hence its magnetism, alpha iron being the only
allotropic form under which iron is magnetic.

LESSON XV THEORIES OF THE HARDENING OF STEEL 3

While the allotropists regard the retention of beta iron as the chief cause of the
hardening of steel, they do not ignore the very important part played by carbon.
They realize that the presence of carbon is essential to the retention of iron in its hard
allotropic form, for in the absence of carbon it is not possible to harden iron. It is
customary to compare this action of carbon in preventing the transformations to that
of a brake, the more carbon the more powerful the brake action, hence the harder the
steel because of the retention of a larger quantity of beta iron. They believe, however,
that the hardness of quenched steel increases with the proportion of the beta iron it
contains rather than with the proportion of carbon. In other words, that if it were
possible to retain the same amount of beta iron with less carbon or even in the com-
plete absence of carbon the metal would be equally hard. The ajlotropists do not
claim that steels hardened in the ordinary way, that is, martensitic steels, are abso-
lutely free from gamma iron. Evidences are lacking to settle this point. Again the
allotropists do not deny that the internal pressure created by the transformation of
austenite into the more bulky martensite may contribute to the hardness of the
metal.

Summing up, in the light of this theory, the hardening of steel by rapid cooling is
thus explained: (1) the bulk of the iron passes from the gamma to the beta condi-
tion, hence the great hardness produced, (2) some of the beta iron is further trans-
formed into alpha iron, hence the magnetism of hardened steel, (3) a large proportion
of the carbon or more probably of the carbide FesC remains dissolved in the beta
iron, the presence of this dissolved (hardening) carbon in hardened steel being proven
by chemical analysis, (4) the internal pressure created by the transformation of aus-
tenite into martensite, that is, of gamma into beta iron, may contribute to the final
hardness. Osmond's theory of the hardening is, therefore, based on the belief (1) in
the existence of beta iron and (2) in the hardness of beta iron. Although the very
existence of beta iron has been challenged by Le Chatelier and Guillet the author
believes that the evidences at hand point strongly to its reality, while direct evidences
of its hardness have been obtained. Rosenhain and Humfrey on straining polished
bands of pure iron heated in vacuum obtained three sharply distinct structures, each
one possessing different mechanical properties, especially as to hardness, and each
structure representing the condition of the iron, respectively, above A 3 , between A 3
and A 2 , and below A 2 , that is, the structure of gamma, beta, and alpha iron. The
iron when in the beta condition was found to be decidedly harder than either gamma
or alpha iron. These experiments point, therefore, both to the existence of beta iron
and to its hardness.

On tempering hardened steel Heyn found that 70 per cent of the increased hard-
ness produced by the hardening operation were lost in tempering below 400 deg.,
the remaining 30 per cent being possibly due, according to Osmond, to the internal
pressure already alluded to. He further observed that hardening (dissolved) carbon
did not begin to be converted into cement carbon (crystallized Fe 3 C) until a tempera-
ture of 400 deg. was reached. Steel then loses most of its hardness while its carbon
remains in the hardening condition. From this coexistence of softness and harden-
ing carbon it logically follows that steel does not owe its hardness to the presence of
hardening carbon, and that the presence of allotropic beta iron remains the only
possible explanation. This conclusion is further supported by the fact that on tem-
pering steel it is chiefly below 400 deg. that heat is liberated and this liberation must
necessarily be ascribed to the iron returning from the beta to the alpha condition.

4 LESSON XV THEORIES OF THE HARDENING OF STEEL

Alpha Iron Theory. Le Chatelier and Guillet believe that on quick cooling
through the critical range the allotropic transformation of iron from its gamma to
its alpha condition is not prevented but that the steel remains, nevertheless, in the
condition of a solid solution, hardened steel in their opinion being a solid solution of
carbon (or of the carbide Fe 3 C) in alpha iron owing its hardness to its state of solu-
tion and its magnetism to the presence of alpha iron. This view is based chiefly
upon these writers' belief that the point A2 is not an allotropic point and that there-
fore beta iron does not exist. The most serious objection to this theory lies in the
conclusive nature of the evidences pointing to the allotropic character of the point A 2 .

Carbon Theories. The carbon theories contend that the hardness of rapidly
cooled steel is due primarily to the retention in the cold of a very hard condition of
the carbon normally stable only above the range, the allotropic transformation of
iron playing no, or but an unimportant, part in the phenomenon. As supporting
their claims they point to the fact that carbonless iron cannot be hardened and that
the more carbon present the greater the increased hardness produced by quick cool-
ing, at least up to the eutectoid carbon ratio.

These theories differ in regard to the exact condition of the carbon thus retained
by quenching and imparting great hardness to the metal. The hardening carbon
theory and the subcarbide theory should be briefly described.

The Hardening Carbon Theory. It was held for many years by the majority of
writers that hardened steel owed its hardness to the presence of hardening carbon, a
form of carbon stable only above the range but which could be retained, in part at
least, by quick cooling. This belief rested on the apparent difference existing between
the condition of the carbon in hardened and in unhardened steels as proven by dis-
solving them in cold dilute acids when a large proportion of the carbon of hardened
steel escapes as hydrocarbons, whereas nearly the totality of the carbon of unhard-
ened steel remains as a residue which, upon being analyzed, is found to consist of the
carbide FesC. As to the exact nature of hardening carbon, vague, conflicting, and
often extraordinary statements appeared, it being claimed by some, for instance, that
hardening carbon was carbon in a diamond-like condition. It is at present believed
by most that hardening carbon is carbon (or more probably the carbide Fe 3 C) dis-
solved in iron, its escape as hydrocarbons upon being subjected to the action of dilute
acids being due to its extremely fine state of division. If this be the nature of harden-
ing carbon, the hardening carbon theory becomes, of course, a solution theory.

It is obvious that carbon as such, no matter how great its hardness, could not
impart extreme hardness to steel in which it may be associated with 199 times its
weight of soft ferrite, as, for instance, in steel containing 0.50 per cent carbon. The
contention that it is not carbon, as such, which is retained by quick cooling but a very
hard carbide constituting the whole or a large part of hardened steel is not, of course,
open to the same objections. This is the claim of the subcarbide theory.

The Subcarbide Theory. Arnold contends that eutectoid steel above its critical
range exists as the carbide Fe^C, a chemical compound containing about 0.89 per
cent carbon. This carbide which he calls "hardenite" is very hard and being re-
tained by quick cooling imparts hardness to quenched steel. In hypo-eutectoid steel
some ferrite and in hyper-eutectoid steel some cementite are dissolved in this sub-
carbide. It follows from this theory that austenite and martensite correspond to
different structural appearances of the same constituent, namely, the carbide Fe2.jC
when of eutectoid composition, the same carbide plus ferrite or cementite in hypo- or

LESSON XV THEORIES OF THE HARDENING OF STEEL 5

hyper-eutectoid steel. This theory is purely of a speculative character, the existence
of the carbide Fe24C not being supported by a single direct evidence. It is, moreover,
strongly opposed by the universally accepted theory of metallic alloys which holds
that eutectic (and eutectoid) alloys immediately before their formation are not defi-
nite chemical compounds but liquid or solid solutions. On forming, whether or
not it implies a change of state, the solution is transformed into an aggregate of the
solute and solvent, ferrite and cementite in the case of iron-carbon alloys. The
breaking up at a certain critical temperature of a definite chemical compound,
as demanded by Arnold's theory, into a eutectoid aggregate of the elements of that
compound is contrary to our firmly established knowledge of the mechanism of the
formation of such aggregates. It implies a return to Guthrie's original error.

The Stress Theory. In cooling steel quickly from above its critical range it is
subjected to two kinds of stresses, (1) stresses due to the shrinkage of its outer shell
on its interior and (2) stresses due to the transformation with increased volume of
gamma into beta and alpha iron. The existence of the strains resulting from these
stresses have been claimed to account satisfactorily for the hardening of steel by
sudden cooling. It was argued long ago, for instance, that the hardening of steel by
sudden cooling might be due to the metal being in a severely strained condition, be-
cause of the quicker cooling of the outer layers, these layers through their contrac-
tion exerting a severe pressure upon the central portion of the steel objects. The
advocates of this theory pointed to the increased hardness resulting from cold work-
ing steel as a proof that severe straining produces hardness. Some went as far as to
claim that cold worked steel and hardened steel are practically in the same physical
condition, the quenching of steel producing, so to speak, an internal cold working
(straining) of the metal. They seem to have overlooked the enormous difference be-
tween the relatively small increase of hardness produced by cold-working and the
hardness resulting from quenching. In the fact that both cold-working and harden-
ing increase the elastic limit and decrease the ductility they found additional support
for their view. According to Osmond the belief once held that cold working causes
an allotropic transformation is now abandoned. While it is not unreasonable to
assume that the strained condition of the metal adds to its hardness it is hardly think-
able that the sudden and very great increase of hardness produced by quick cooling
is due altogether to this straining. If it were so the outer layers of quenched steel
implements should not be hard and razor blades could not be hardened. Again, it is
impossible to reconcile this theory with the fact that it is necessary to quench steel
from above its critical range in order to harden it, for one cannot conceive why, if the
steel be quenched slightly below the range, the strain created in the quenching bath
would be so slight as to have no hardening effect, whereas quenching from a tempera-
ture but a few degrees higher would produce very severe straining. Finally, if the
contraction of the outer layers due to their rapid cooling can induce such hardening
strains in the case of steel, it is surprising that a similar phenomenon is not observed
in the case of other metals. The internal strains resulting from the transformation
of gamma into beta iron (austenite into martensite) with increased volume afford a
more acceptable explanation of the hardening of steel. It was offered long ago and
has recently been revived and presented in a more scientific way, notably by Andre
Le Chatelier, Charpy, and Grenet. It is argued that on quick cooling the allotropic
transformations take place, partially at least, at a temperature so low that the metal
lacks the necessary plasticity to yield to the severe stress excited by these transforma-

6 LESSON XV THEORIES OF THE HARDENING OF STEEL

tions, remaining, therefore, severely strained. It becomes internally "ecroui," as the
French express it. Here again, however, it would seem as if the outer layers should
not be strained and should, therefore, remain soft, which of course is contrary to facts.
Nor is the failure of carbonless iron to harden satisfactorily explained by this theory.

Grenet believes that in hardened steel the allotropic transformations are complete,
that is, that its iron exists only in the alpha condition but so severely strained (ecroui)
as to be very hard. He rests his belief chiefly on his assertion that on quick cooling
the dilatation indicative of the allotropic transformations are not suppressed and
that the metal does not remain non-magnetic. He overlooks the claims of the advo-
cates of the retention theories, so strongly supported, that the transformations are
not completely suppressed, hence the occurrence of a dilatation and of magnetism.
When they are completely prevented, as in austenitic steels, the metal neither expands
nor becomes magnetic on cooling.

The liberation of heat observed on tempering hardened stee.1 points to a return to
a more stable condition, supporting, therefore, the retention theories and opposing the
stress theory.

Tempering and the Retention Theories. The tempering of hardened steel, as
already explained, is readily accounted for by the retention theories on the ground
that the metal being in an unstable condition is ever eager to assume a more stable
form, implying a return, partial at least, of the iron to the alpha condition and of the
carbon to the cement condition. On heating the steel but slightly above atmospheric
temperature its rigidity is sufficiently diminished to permit a slight transformation
of this kind, the higher the temperature the more pronounced of course being its
tempering.

Tempering and the Stress Theory. The stress theory, likewise, satisfactorily
accounts for the tempering of hardened steel on the ground that upon slight reheat-
ing the internal strains are sufficiently released to produce an appreciable decrease of
the specific effects of hardening, namely, decrease of hardness, of strength, of elastic
limit and increased ductility.

Summary. From the above short description of the various theories advanced
to explain the hardening of steel the reader will probably gather the impression that
the retention theories, especially the beta iron theory, are the most acceptable ones.
It seems quite possible, however, even probable, that the various theories, while ap-
parently antagonistic, bring each their contribution to the elucidation of the problem.
Should we not believe with the allotropists that the hardness of steel is due chiefly to
the retention of a large quantity of a hard allotropic variety of iron, probably beta
iron, and that this iron contains in solution the hardening carbon of the carbonists,
the presence of which is absolutely essential to the existence of beta iron in the cold?
Should we not with the advocates of the stress theories believe in the hardening in-
fluence of the strains created on quick cooling (a) because of the shrinkage of the
outer layers of the metal and (6) because of the expansion accompanying the trans-
formation of gamma into beta iron? None of these theories alone gives a fully satis-
factory explanation: Beta iron cannot be retained in the absence of carbon and if it
could be it is not certain that it would be intensely hard; the presence of intensely
hard carbon or iron carbide as the chief cause of hardening is contrary to evidences;
the strained condition of hardened steel does not account satisfactorily for its hard-
ness; Le Chatelier's contention that quickly cooled steel is hard although its iron is
in the soft alpha condition because of its being in a state of solution is opposed by

LESSON XV THEORIES OF THE HARDENING OF STEEL 7

the evidences at hand (a) of the existence of beta iron and (6) of the hardness of beta
iron; Arnold's theory that hardened steel owes its hardness to the retention of a hard
subcarbide of iron lacks experimental support and is scientifically untenable.

Examination

Describe briefly the various theories of the hardening of steel indicating your
preference and your arguments supporting it.

LESSON XVI

THE CEMENTATION AND CASE HARDENING OF STEEL

The affinity of iron for carbon is so great that when heated -to -a sufficiently high
temperature in contact with some suitable carbonaceous matter it readily absorbs
carbon. If the heating be protracted (several days) and the amount of carbon ab-
sorbed considerable, the operation is known as "cementation" and the resulting metal
as "cemented," "converted," or "blister" steel, or in Sheffield, England, as "blister
bar," while if the treatment be of relatively short duration (a few hours) and the
absorption of carbon in consequence superficial, it is called "case hardening."

Cementation is generally applied to wrought-iron bars which are afterwards
melted (crucible process) and shaped into finished articles by casting or forging,
while case hardening is applied directly to finished objects generally of low carbon
steel. The purpose of cementation is to introduce carbon into wrought iron, thereby
converting it into steel, the subsequent treatments (melting, forging) producing a
uniform distribution of the carbon, whereas the purpose of case hardening is to
manufacture steel objects with hard skins or cases while retaining their soft and
tough centers or cores.

The quantity of carbon thus absorbed by iron at a high temperature but below its
melting-point depends chiefly upon (1) the composition of the iron or steel subjected
to carburizing, (2) the carburizing temperature, (3) the length of time at that tem-
perature, and (4) the nature of the carburizing material.

Composition of the Iron or Steel Subjected to Carburizing. It is probably true
that the smaller the proportion of carbon in the iron the more eagerly will it take up
carbon, from which it follows that as the carburizing proceeds, that is, as the metal
becomes more highly carburized, additional introduction of carbon requires progres-
sively longer time, the metal acting in this way not unlike a solution approaching
its saturation point.

In the cementation process bars of very pure wrought iron and in "case harden-
ing" steel objects containing at the most 0.20 per cent carbon are subjected to the
carburizing treatment. The steel should not contain over 0.25 per cent of manganese
lest the case be too brittle. The presence of certain elements appear to hinder the
carburizing operation while others facilitate it.

According to Guillet the absorption of carbon is favored by those special elements
which exist as double carbides such as manganese, tungsten, chromium, molybdenum,
and opposed by those which form solid solutions with iron such as nickel, silicon, and
aluminum.

Carburizing Temperature. While it has been claimed that iron below its critical
range will absorb some carbon this absorption, if taking place at all, is very slow,
from which it is logical to infer that alpha iron has very little, if any, dissolving power
for carbon. In order to produce quick and intense carburization the iron should be
in its beta or, more probably, in its gamma condition, and steel, therefore, in the condi-

2 LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL

tion of a solid solution. Cementing and case hardening operations must consequently
be conducted above the critical range of the iron or low carbon steel treated, that is,
at a temperature exceeding 800 deg. C. It is also certain that the higher the tem-
perature the quicker will carbon be absorbed and the deeper will it penetrate into
the steel, that is, the deeper the "case." At Sheffield, England, where the cementation
process is used more extensively than anywhere else the carburizing temperature is
in the vicinity of 950 to 1000 deg. Most case hardening treatments are probably con-
ducted in the vicinity of 900 to 950 deg. C.

Time at Carburizing Temperature. The amount of carbon absorbed, and there-
fore the thickness of the case as well, increases, of course, with the length of the
operation but, as already mentioned, carburization takes place more and more slowly
as the carbon content increases. The maximum amount of carbon which iron can
take up while in the solid state is probably not far from 2.50 per cent, this, however,

Fig. 1. Steel. Case hardened. Magnified 20 diameters.

requiring a protracted treatment at a very high temperature. While in the manu-
facture of blister steel considerably more than one per cent of carbon is frequently in-
troduced into the wrought-iron bars, in carburizing finished steel articles it is seldom
desired to produce a case containing more than one per cent of carbon near the outside,
a superficial, carburized layer of eutectoid composition (0.85 per cent C.) being gen-
erally considered to yield the best results. The length of time needed to produce the
desired degree of carburization and desired depth of case must necessarily depend upon
the nature of the metal, the kind of carburizing material used, and the temperature.

Distribution of the Carbon. It will be apparent from the nature of the opera-
tion that in this carburizing of solid iron carbon travels slowly from the outside
towards the center and that, therefore, the proportion of carbon absorbed must
decrease from outside to center, unless indeed the objects treated are very thin or
the treatment so long and conducted at so high a temperature as to cause even the
center to absorb the maximum amount of carbon. The decrease of carbon as one
approaches the core of the object is well illustrated in Figures 1 and 2. A band of
hyper-eutectoid steel characterized by the presence of free cementite is frequently

LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL 3

noted (Fig. 1) followed by a band of cutectoid composition characterized by the
absence of both free cementite and free ferrite and this in turn is followed by a
band showing abrupt and rapid decrease of carbon characterized by an increasing
amount of free ferrite.

In case hardening operations the penetration of the carbon may be very slight
indeed, not exceeding 0.5 mm., while it may measure as much as 5 mm. In the
majority of instances the penetration does not exceed 2 mm. This depth of pene-
tration or thickness of case must be regulated according to requirements. It will
depend upon temperature, time, composition of steel, and kind of carburizing material.
Lake mentions 0.87 mm. per hour as an average speed of penetration. As already
stated, it is not generally advisable to produce a case containing "more than some
0.90 per cent carbon.

Fig. 2. Steel. Case hardened. Magnified 100 diameters. (G. A. Rein-
hardt in the author's laboratory.)

The production of a deep case, while at the same time keeping the carbon content
of the outside of the case below one per cent, may be brought about by a rather long
treatment at a relatively low temperature, namely, some 850 deg. Some results ob-
tained by Guillet in regard to the influence of temperature and of time on the depth
of penetration are shown graphically in Figure 3 as plotted by Bauer. The carbu-
rizing material used was not stated. The full line represents relative penetrations at
1000 deg. after different lengths of time, namely, one, two, four, and six hours, while
the broken line represents the depths of penetration resulting from heating for eight
hours at different temperatures.

It will be obvious that the process of case hardening can be controlled by the
microscopical examination of test pieces much more readily and accurately than by
chemical analysis.

Carburizing Materials. A great variety of carbonaceous materials are used for
introducing carbon in iron and steel in the solid state. These substances may be
solid, liquid, or gaseous. Solid materials are used more extensively than liquid or

LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL

gaseous ones, the most important being charcoal (both wood and bone), charred
leather, crushed bone, horn, mixtures of barium carbonate (40 per cent) and charcoal
(60 per cent) or of salt (10 per cent) and charcoal (90 per cent), both recommended
by Guillet, and for quick but very superficial hardening, powdered potassium cyanide
and potassium ferro-cyanide or mixtures of potassium ferro-cyanide and potassium
bichromate. A molten bath of potassium cyanide heated to 850 deg. and in which

IIOO'C a

1000* 6

900 *

800 2

IO 15 2O 25 30 35

Penetration, in, m /m.

Fig. 3. Temperature and time-penetration curve. (From Brearley's
"The Heat Treatment of Tool Steel.")

v s a 10

Time, of Heating, (hows)

Fig. 4. Time-penetration curve. (From Brearley's "The
Heat Treatment of Tool Steel.")

the steel articles are immersed produces quickly superficial but hard and even cases.
The poisonous character of the escaping gases, however, is a serious objection to the
use of this method. The carburizing of iron may also be performed at the proper
temperature by means of gases such as illuminating or other coal or oil gases rich in
hydrocarbons. At the Krupp works in Germany gases are used for carburizing the
faces of armor plates.

The relative merits of wood charcoal, charred leather, and a mixture of barium
carbonate and of wood charcoal for carburizing are shown graphically in Figure 4, in

LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL 5

which arc plotted some results obtained by Shaw-Scott. While wood charcoal causes
a slow carburization it is the best material and the one invariably employed for the
production of very deep cases as, for instance, in making blister steel.

Many so-called secret mixtures are offered for sale as case hardening substances
for which extraordinary virtues are claimed, the usual statement being that by their
use steel of ordinary or inferior quality may be converted into high grade metal com-
parable to the best crucible tool steel. On investigation they are generally found to
be chiefly mixtures of carbonaceous and cyanogen compounds possessing the well-
known carburizing properties of those substances.

Mechanism of Cementation. It was held for many years that in the cementa-
tion of iron solid carbon passed bodily from the packing materiarinCo the metal, fol-
lowed hy a slow migration towards the center. Recent investigations, however, have
made it evident that the transfer of the carbon from the packing material to the
metal is accomplished chiefly, if not altogether, by means of some gases liberated or
formed during the annealing treatment. It has been shown quite conclusively, for
instance, that if a piece of steel surrounded by pure carbon be heated in vacuum, thus
precluding the formation of gases, it will not take up carbon, although one observer
has noted that if decided pressure be applied some carbon will pass into the iron even
in the absence of gases. Whether this be so or not it is apparently certain that the
carbon must first be volatilized before becoming very active as a carburizing agent in
the cementation and case hardening treatments.

Carbon monoxide (CO) and volatilized cyanogen (CN) compounds are the gases
which seem most effective. The carbon monoxide is derived from a partial combus-
tion of the carbon of the cementing material by atmospheric oxygen while the cyan-
ogen results from a combination of that carbon with atmospheric nitrogen or from
the decomposition of cyanide compounds such, for instance, as potassium cyanide
and ferro-cyanide. It may be assumed that the carbon monoxide once formed gives
up its carbon to the iron according to the reaction,

2CO + 3Fe = Fe 3 C + CO 2 ,

the resulting Fe 3 C or cementite being dissolved by the austenite very much as salt is
dissolved in water and the CO 2 being again reduced to CO on coming in contact with
fresh carbon (CO 2 + C = 2CO). The marked activity of cyanogen compounds com-
pared t'o the slower action of charcoal have led some to believe that cyanogen gases
are especially effective in carburizing iron. It should be noted, however, that while
cyanide compounds produce a much quicker carburization they soon lose their
carburizing power so that when deep cases are needed, as in the manufacture of
blister bars, charcoal, acting chiefly through the production of carbon monoxide, is
preferable.

Cooling from Carburizing Temperature. It is generally desired that articles sub-
jected to the case hardening treatment should have a very hard surface. To produce
this hardness the case hardened articles should be quenched from above their critical
range. The prolonged heating at a very high temperature to which these articles
have been exposed, however, has developed a coarseness of structure both in the
core and in the case which would be retained if they were, as they sometimes are,
quenched from the carburizing temperature or after cooling to a somewhat lower
temperature. It is obvious that in order to impart a fine structure both to the core
and to the case the articles should be cooled and then subjected to suitable heat
treatments.

6 LESSON XVI THE CEMENTATION AND CASE HARDENING OF STEEL

Heat Treatment of Case Hardened Articles. In order to refine the structure of
the core which has been coarsened by a long exposure to a high temperature the
metal should be reheated slightly above the critical range of that core and since its
carbon content seldom exceeds 0.15 per cent carbon a temperature of at least 900
deg. C. should be used. Guillet recommends 1000 to 1025 deg. The finer structure
thus imparted to the core will then be retained most effectively by quenching the
metal in water or oil. By such treatment, however, the case, although hardened, is
still relatively coarse since its quenching was effected at a temperature considerably
exceeding its critical range. In order to refine it while leaving the structure of the
core undisturbed the article should now be reheated slightly above the critical range
of the case, that is, to some 750 or 800 deg. C., and then quenched in oil or water. By
this double treatment we have hardened the case while conferring to it as well as to
the core a fine structure.

Tempering Case Hardened Steel. It has been seen that hardened high carbon
steel is generally subjected to a tempering process, i.e. reheated to some 200 or 300
deg. C. in order to decrease its brittleness while losing but little hardness.

There seems to be at first sight no apparent reason why case hardened articles
could not be likewise improved by suitable tempering following the hardening of
the case. On second thought, however, it will be realized that since the chief pur-
pose of tempering is to toughen the hardened steel and since case hardened articles
depend for their toughness on the toughness of their cores little is to be gained by
tempering them.

Experiments

A sample representing the cross section of a case hardened bar or other steel object
should be polished, etched, and microscopically examined. The structures of the case
and core described in the lesson should be noted.

If the bar has not been retreated after cooling from the carburizing heat it should
be subjected to the two heat treatments described, namely, (1) reheating to 950 deg.
C. followed by quenching in water or oil and (2) reheating to 800 deg. followed by
quenching in water. A transverse section should be prepared for microscopical
examination and the refining of both the. core and case noted.

Photomicrographs of the various structures should be taken, for which a magnifi-
cation of 50 to 100 diameters will be sufficient.

Examination

I. Describe the absorption of carbon by iron above its critical range.
II. Describe the best treatments to be applied to case hardened articles after car-
burizing and explain why they are needed.

LESSON XVII

SPECIAL STEELS
GENERAL CONSIDERATIONS

The steels so far considered in these lessons are the ordinary steels of commerce, at
present often called "carbon" steels to distinguish them from the "special" steels
of relatively recent origin but of rapidly growing importance. By special steels is
meant those steels which owe their properties in a marked degree to the presence of
one or more special elements whereas the properties of carbon steels depend chiefly,
if not exclusively, for like treatment, upon the proportion of carbon present. Special
steels containing but one special element are commonly called "ternary" steels,
being considered to be made up of three constituents, namely iron, carbon, and the
special element, while steels containing two special elements are called " quarternary "
steels because of the presence of four constituents: iron, carbon, and the two special
elements. These two classes of special steels will be considered separately.

Ternary Steels. We are indebted to Guillet for a brilliantly conceived and
vigorously developed theory of the ternary steels. Too rigorous an application of the
theory, however, should not be insisted upon for there are some facts not yet satis-
factorily explained by it. Its use, nevertheless, will be found an invaluable guide
in directing researches dealing with the manufacture and the application of these
steels.

Guillet's theory of the structure and properties of ternary steels may be briefly
formulated by a few propositions. It is also represented graphically in Figure 1.

(1) On the introduction of a special element in carbon steel the latter remains at
first pearlitic, but as the proportion of the special element increases, the carbon re-
maining constant, it becomes first martensitic and then austenitic (polyhedric), as
shown graphically in Figure 1, and sometimes cementitic (carbide steel) 1 as later
explained.

(2) By increasing the amount of carbon present in a special steel, the proportion
of the special element being kept constant, it is generally converted from a pearlitic
into a martensitic condition or, if already martensitic, into an austenitic condition.

(3) The greater the amount of carbon the smaller the proportion of the special
element needed to cause a structural transformation, as for instance pearlite into mar-
tensite or martensite into austenite. This is indicated in Figure 1.

(4) The greater the amount of the special element the smaller the proportion of
carbon needed to cause a structural transformation. This is also shown in Figure 1.

1 Guillet uses the term polyhedric to designate an austenitic structure and carbide steel (acier a
rnrhiire) to indicate the presence of ccmcntite (generally in special steels a double carbide of iron and
the special element). It seems to the author that the terms austenitic and cementitic are preferable
because they suggest unmistakably the nature of the constituents. Austenitic steels are not the only
ones exhibiting u polyhedric structure; ferritic (low carbon) steels for instance are also polyhedric.

2 LESSON XVII SPECIAL STEELS

(5) No very sharp lines of demarcation are observed between the different types
of structures mentioned in the preceding propositions, relatively wide ranges of com-
position existing, on the contrary, in which the steel may be partly pearlitic and partly
martensitic or partly martensitic and partly austenitic, etc. These transition ranges
are indicated by shaded areas in the diagram of Figure 1. Greater refinement in the
construction of this diagram would undoubtedly lead to the introduction of a troo-
stitic zone between the pearlite and martensite areas and possibly also of a sorbitic
zone between pearlite and troostite.

To sum up, constituents may be formed during the slow cooling of many special
steels which in carbon steels can only be produced by very rapid cooling through the

.6 .3

Percent carbon.

Fig. 1. Constitutional diagram of special steels.

critical range. Carbon steels, moreover, even after very rapid cooling cannot be re-
tained wholly in an austenitic condition while several special steels remain austenitic
after slow cooling. It is evident from the above and from the diagram that in order
to produce a certain structure, (1) the proportion of carbon may be kept constant
while the proportion of the special element is increased until the desired structure is
obtained, or (2) the proportion of the special element may be kept constant and the
proportion of carbon increased, or (3) both the proportion of carbon and of the special
element may be increased when the desired structure will be obtained more quickly.

The usefulness of Quillet's diagram is obvious. Should we desire, for instance, to
know the kind of structure, and therefore the physical properties, of a steel containing
0.60 per cent carbon and 8 per cent of the special element, the diagram shows that-
such composition falls within the martensitic range. Likewise a steel containing one
per cent carbon and 15 per cent of the special element would be austenitic according
to the diagram. Or one may wish to know what proportion of the special element

LESSON XVII SPECIAL STEELS 3

should be added to a carbon steel containing, say, 0.5 per cent carbon, to make it
martensitic; the diagram shows that 7 per cent will be needed. Again, having an
austenitic steel containing 10 per cent of the special element it may be desired to know
the minimum amount of carbon that may be present without causing the steel to
become martensitic; the diagram shows 0.80 per cent of carbon to be the smallest
proportion of carbon permissible.

The construction of such diagrams requires the preparation of a number of alloys
varying in their contents of carbon and of the special element, their microscopical
examination and the plotting of their structure.

It is quite essential to know the rate of cooling adopted in the construction of the
diagram, i.e. whether the samples were cooled in air or more slowly in the furnace,
for it is evident that their structure may be deeply affected by thus varying the speed
at which they cool. Some special steels, for instance, may be pearlitic when cooled
very slowly in the furnace, martensitic when cooled in air, and austenitic after water
quenching.

Influence of the Special Element upon the Location of the Critical Range. The
production of martensitic and austenitic structures on slow cooling is due to the fact
that the special element lowers the position of the critical point to a temperature so
low (1) as to permit only a partial transformation, namely of austenite into marten-
site, the steel being too rigid to allow a more complete transformation, or (2) as to
prevent even a slight transformation, the steel in that case remaining austenitic.
This influence of the special element in lowering the position of the critical range is
depicted in Figure 2 in which it is assumed that the proportion of carbon remains
constant. It has been further arbitrarily assumed in this diagram that the critical
point was progressively and uniformly lowered from 700 deg. C. to deg., as the pro-
portion of the special element increased from to 6 per cent. From many observa-
tions it appears (1) that as long as the critical point remains above 300 deg. C. the steel
becomes pearlitic on slow cooling, (2) that when the critical point is lowered below
300 deg. it becomes martensitic, the rigidity of the metal preventing further trans-
formation, and (3) that when the critical point is lowered to atmospheric temperature
or below it the metal remains untransformed, that is, austenitic. These inferences
are offered here because of their apparent usefulness and suggestiveness, but the au-
thor realizes that the lines indicating the relation between the position of the critical
points and the corresponding structures cannot be sharply drawn, for they are likely
to shift according to the nature of the special element, the rate of cooling, etc. Again,
troostitic and possibly also sorbitic steel are likely to form between pearlite and
martensite, that is, whenever the critical point is lowered, say below 400 or possibly
below 500 deg.

To make the meaning of the diagram of Figure 2 clear let us consider three steels:
I, II, and III, all containing one per cent of carbon, but respectively 1, 4.50, and 7
per cent of the special element. As steel I cools it undergoes its transformation at
about 600 deg. At that temperature the metal is so plastic that the transformation
of austenite into pearlite readily takes place; the steel becomes pearlitic. The critical
point of steel II is slightly below 200 deg. At this temperature the transformation
of austenite into martensite will take place, but the metal is now too rigid to permit
further transformations; the steel remains martensitic. Steels which remain marten-
sitic after slow (air) cooling are said to be "self-hardening." In the case of steel
III, since its critical point is lowered below atmospheric temperature it necessarily re-

4 LESSON XVII SPECIAL STEELS

mains austenitic. Since austenitic special steels have their points of transformation
situated below atmospheric temperature, it should be possible through cooling to a
sufficiently low temperature, as for instance by immersion in liquid air, to cause at
least their partial transformation, that is, they should become martensitic after such
treatment and this indeed is precisely what happens. The transformation of aus-

floo.

3 4 S

/ Special element.
Carbon I %

Fig. 2. Influence of special element on the position of the critical point.

tenite into martensite takes place with increased volume and the steel from non-
magnetic becomes magnetic.

In the presence of a special element lowering the critical points the influence of
carbon is cumulative, i.e. the greater the proportion of carbon the more marked the
action of the special element. An attempt has been made in Figure 3 to show this
graphically. The diagram indicates the position of the critical point corresponding
to any combination of carbon content between 0.25 and 1.50 and of the special ele-
ment between and 12 per cent. If the point is above 300 deg. we may assume that
the steel is pearlitic, if below 300 that it is martensitic, if at or below atmospheric

LESSON XVII SPECIAL STEELS

6 LESSON XVII SPECIAL STEELS

temperature that it is austenitic. In this illustration arbitrary values have been given
to the combined influence of various proportions of carbon and of the special element
upon the position of the critical points. The diagram shows, for instance, that while
with 0.50 per cent carbon 10 per cent of the special element are required to lower the
critical points to deg. C., if the steel contains 1.25 per cent carbon, 4 per cent of the
special element suffice. The production of pearlitic, martensitic, and austenitic struc-
tures according to the position of the critical point has also been indicated. A diagram
of this kind may be even more useful than Quillet's for, while giving the same kind of
information as his, it shows in addition (1) the relation between the composition of the
.special steel and the position of the critical point and (2) the influence of the position
of the critical point upon the structure. Its construction calls for many determina-
tions of the position of the critical point in steels of varying composition and for the
microscopical examination of the corresponding structures. It is of course quite
likely that as experimentally constructed it would consist of more or less smooth curves
rather than of straight lines.

It will soon be shown that the influence of special elements in lowering the critical
points varies greatly, from which it follows that some elements cause the production
of martensitic and austenitic steels much more readily than others. Indeed some
elements never cause a sufficient depression of the critical points to yield austenitic
or even martensitic steels. A few elements even actually raise the location of the
critical points in which case, of course, the steel always becomes pearlitic on slow
cooling, regardless of its composition.

From the foregoing considerations it appears that according to their structural
composition special steels may be divided into at least four classes, (1) pearlitic steels,
(2) martensitic steels, (3) austenitic (polyhedric) steels, and (4) cementitic (Quillet's
carbide) steels. These should now be farther considered.

Pearlitic Steels. The pearlitic steels, as we have seen, are those which generally
contain but a relatively small amount of the special element, although in case of steel
very low in carbon the proportion of the special element may be quite large. In these
steels the special element may (1) be dissolved in the ferrite forming with it a solid
solution, (2) be combined with carbon in cementite as a double carbide of iron and the
special element, or (3) be partly dissolved in ferrite and partly combined with carbon.

According to Quillet nickel and silicon, for instance, are entirely dissolved in
ferrite, while manganese, chromium, tungsten, vanadium, and molybdenum are partly
held in solution by ferrite and partly present in cementite as double carbides. Such
terms as nickel-ferrite, silico-ferrite, mangano-ferrite, etc., have been suggest eel to
designate ferrite holding in solution large proportions of nickel, silicon, manganese,
etc., respectively.

The structure of pearlitic special steels is generally quite similar to that of pearlitic
carbon steels, although the pearlite particles of special steels frequently are more
angular than those of carbon steel, often exhibiting many straight sides and sharp
corners, whereas the pearlite particles of carbon steel are more rounded. The lami-
nation of pearlite is also often more minute in special steels while for same carbon
content it often appears to occupy a larger bulk. From their similarity of struc-
ture, it might reasonably be inferred that pearlitic special steels should not differ
much in physical properties from ordinary carbon steels. As a matter of fact, however,
pearlitic special steels are often greatly superior to carbon steels generally because
they possess in a much greater degree that desirable combination of properties, strength,

LESSON XVII SPECIAL STEELS 7

or rather high elastic limit, and ductility. They are also frequently harder for like
ductility, and therefore better adapted to resist wear. Finally their ability to resist
shocks is often markedly superior to that of carbon steels. Their uses make for greater
efficiency and their greater strength permits an often welcome reduction in bulk and
weight of certain parts of machinery. This greater strength and stiffness of special
pearlitic steels may be due to the special element dissolving, in part at least, in the
ferrite thereby increasing its strength, elastic limit, and hardness, a stronger and harder
ferrite resulting in turn in a stronger and harder pearlite. The superior physical qual-
ities of these steels may also be due, at least partly, to a finer, closer ferrite-cementite
aggregate.

The critical points of special pearlitic steels generally occur atrtemperatures some-
what lower than those at which the critical points of carbon steel are located, this
being in accord with the usual influence of special elements upon these points as
already explained.

Martensitic Steels. For the same carbon content martensitic steel contains
necessarily more of the special element than pearlitic steels, while for a given propor-
tion of the special element they must contain more carbon; they generally contain
both more carbon and more of the special element than pearlitic steels. As already
stated the influence of some special elements in lowering the critical points is not suffi-
ciently pronounced to result in the formation of martensite on slow cooling. Indeed
some elements raise the position of the critical points in which case pearlitic steel
must necessarily always be formed on slow cooling.

The properties of martensitic special steels are not unlike the properties of marten-
sitic carbon steels, that is of hardened carbon steel. These steels are hard and brittle
and unforgeable in the cold. Their uses are very limited, chiefly because of their
brittleness and of the difficulty of machining them. While resembling hardened carbon
steels they are quite stable above atmospheric temperature, being little affected by
tempering, i.e. by reheating to 200 or 300 deg. C. This property suggests one impor-
tant application at least of martensitic special steels later to be considered, namely,
their use for the manufacture of cutting tools, their greater stability permitting the
tools to be heated to a higher temperature, i.e. the cutting being performed at greater
speed without breaking down through excessive tempering.

The martensite of special steels probably is, like the martensite of carbon steels,
chiefly a solid solution in beta iron of the carbide FesC or more often of a double car-
bide of iron and of the special element, the magnetism of the metal being due to the
presence of some alpha iron.

Austenitic (Polyhedric) Steels. For a given carbon content austenitic steels
necessarily contain more of the special element than martensitic steels, while for a
given proportion of the special element they are necessarily more highly carburized.
Austenitic steels generally contain both more carbon and more of the special element
than martensitic steels. Their properties are as might be expected similar to those of
austenitic carbon steels, that is, of high carbon steels cooled extremely quickly from
a very high temperature. Austenitic steels are moderately tenacious but very ductile;
they have a low elastic limit but possess a remarkable power of resisting wear by abra-
sion as well as rupture by shocks. Their mineralogical hardness, however, is gener-
ally inferior to that of martensitic steels.

Unlike quenched austenitic carbon steel these special steels are stable at all tem-
peratures below their point of solidification and are not therefore greatly affected by

8 LESSON XVII SPECIAL STEELS

heat treatment unless of a protracted nature. They should be, however, free from
separated carbides. If not they should be heated to a high temperature so as to cause
the solution of these carbides, and then quenched to prevent their separating again.

The austenite of special steels undoubtedly consists of a solid solution in gamma
iron of carbon and of the special element, probably of a double carbide. Because of
the absence of alpha iron austenitic steels are non-magnetic.

Austenitic special steels find useful application for parts of machinery and the like
subjected to very severe wear by abrasion and to shocks. Their low elastic limit and
the difficulty of machining them are the chief reasons preventing their wider use.

Cementitic (Carbide) Steels. Some special elements on being introduced in
increasing proportions fail to convert the metal into austenite, free particles of a double
carbide of iron and the special element being formed instead and embedded in a mar-
tensitic, troostitic, sorbitic, or pearlitic matrix. Guillet calls these "carbide" steels.
Such elements as chromium, tungsten, molybdenum, and vanadium when present in
sufficient quantity produce cementitic steels. The most valuable property of these
steels is their power, when the carbides are embedded or rather dissolved in a mar-
tensitic matrix, of retaining their hardness when heated to such temperature as
would readily cause the softening of hardened carbon steel, thus permitting their use
in the shape of tools at such speed as to cause their cutting edges to become visibly
hot. This phenomenon will be further explained when describing self-hardening and
high speed steels.

Treatments of Special Steels. Special steels are subjected to the same treat-
ments as carbon steels, i.e. to hot and cold working, annealing, hardening and temper-
ing, and to case hardening. Since the special elements, however, often have a marked
influence on the position of the critical points it is obvious that the temperatures in-
dicated as the most suitable ones for the annealing and hardening of carbon steels
may not be satisfactory in the case of special steels. The position of the critical
points should in every case be determined and the heat treatments conducted accord-
ingly. Greater care is also frequently needed in the forging of special steels, many
of them not being quite as malleable as carbon steels. Finally some of the special
elements promote the absorption of carbon by iron below its solidification-point while
others oppose it and these influences must be considered in case hardening special
steels. The treatment of special steels will be considered further in the next lesson in
connection with the description of some of the most important commercial types.

Treatment of Pearlitic Steels. Pearlitic special steels may, like carbon steels,
be subjected to annealing, hardening, tempering, and case hardening. Their critical
points, however, being generally lower, the proper temperatures for these operations
are likewise lower. They should be determined for each steel. According to Guillet,
however, the steel should be heated quite a little above its critical range because in
the case of the pearlite of special steels its transformation into a solid solution does not
take place as readily. The steel should then be cooled to near its critical range
before quenching. Guillet writes that pearlitic special steels may be divided into
(1) those that are not very sensitive to annealing, namely nickel and silicon steels,
and (2) those that are very sensitive to annealing, namely manganese, chrome, van-
adium, tungsten, and molybdenum steels. It should be noted that in the first group
the special elements are supposed to be entirely dissolved in the iron, while in the
second group they are partly dissolved and partly present as carbides. The case harden-
ing of pearlitic special steels may result in the production of martensitic or even austen-

LESSON XVII SPECIAL STEELS 9

itic cases without the necessity of rapid cooling from above the critical range. This
will be readily understood by referring to Figure 1 where it will be seen that by keep-
ing the proportion of the special element constant and increasing the carbon the
steel may be converted from a pearlitic to a martensitic and even to an austenitic
condition. The nearer the steel to the boundary between the pearlitic and marten-
sitic zones the more readily, of course, will it become martensitic on case hardening
because the smaller the amount of carbon needed to produce that transformation.
This possibility of producing steel objects with a pearlitic soft core and a hard marten-
sitic shell without quenching from a high temperature and therefore without exposing
the objects to the dangers of the quenching bath does not seem to have received the
attention it deserves for it suggests important practical applications. And likewise
the production of a soft pearlitic core surrounded by a hard martensitic steel, itself
surrounded by a tough austenitic steel. It points at least to the manufacture of pearl-
itic steel objects which can be readily machined, etc., and as a last treatment made
austenitic to a certain depth, being in this way greatly superior to the austenitic steels
at present used, which being cast in an austenitic condition can be machined only
with very great difficulty.

Treatment of Martensitic Steels. Unlike martensitic carbon steels martensitic
special steels being quite stable below the critical range of the metal are not readily
affected by tempering treatments. It should be borne in mind, however, that some
special steels which are martensitic after air cooling may become pearlitic, in part at
least, after very slow cooling in the furnace and austenitic or martenso-austenitic
after quenching in water. By selecting a special steel of suitable composition, for in-
stance, and allowing it to cool in a furnace it becomes pearlitic and can in consequence
be machined; after machining the finished object may be made martensitic by cool-
ing in air, doing away with the necessity of the quenching bath and its inherent evils.
It is evident that for this purpose the composition of the steel should be near the bound-
ary line between the pearlitic and martensitic zones. The author believes that the
practical possibilities of this procedure have been overlooked. When working near the
pearlito-martcnsitic boundary line the formation of troostite is of course always likely.

Treatment of Austenitic Steels. Austenitic special steels are stable theoretic-
ally at least at all temperatures and should not, therefore, be affected by heat treat-
ment of any kind. Some special steels, however, may require air cooling to be truly
austenitic, in which case very slow cooling in the furnace may result in the produc-
tion of some martensite or troostite and even of some pearlite, accompanied by
the reappearance of magnetism. It also frequently happens that during the rela-
tively slow cooling of austenitic steels some free cementite may be formed, consist-
ing generally of a double carbide of iron and the special element, this setting free of
cementite being generally accompanied by a decided decrease of strength and ductil-
ity. In order to cause the reabsorption of the separated carbide heating to a high
temperature (1000 cleg. C. or higher) is generally required followed by rapid cooling
in water or oil so as to prevent its separating again on cooling. This treatment is
sometimes called " water toughening."

Treatment of Cementitic Steels. Cementitic steels contain many particles of
cementite or double carbide embedded in a matrix which may be martensitic, troo-
stitic, sorbitic, or pearlitic according to the rate of cooling. It is often desirable to
cause the disappearance in part at least of these particles while producing a finely
martensitic structure, and for this purpose heating to a hisrh * ^vre (1000 deg.

10 LESSON XVII SPECIAL STEELS

or more) followed by relatively quick cooling is necessary. Cooling in air is often
sufficiently rapid to retain the carbide in solution, as for instance in the case of the
high speed steels soon to be described.

Quaternary Steels. Quaternary steels like ternary steels may be pearlitic, mar-
tensitic, austenitic, or cementitic as well as sorbitic and troostitic. If the two special
elements are present in small quantities the steels remain pearlitic. If they contain
one or two cementite forming elements such as chromium, tungsten, molybdenum, etc.,
they are likely to be cementitic, that is, to contain many particles of a double or triple
carbide. These should generally be made to dissolve in the matrix by heating to a
high temperature followed by rapid cooling when a finely martensitic structure, quite
free from cementite, may be produced as in the treatment of high speed steel. If
the quaternary steels contain considerable proportions of two special elements capable
of forming solid solutions with iron, as for instance nickel and manganese, they are
frequently martensitic or austenitic. A large proportion of an element which is partly
dissolved in ferrite and partly present in cementite as a double carbide, manganese,
for instance, may result in the occurrence of cementite particles embedded in an
austenitic matrix.

Examination

I. Describe and explain Guillet's diagram showing the structural composition of
special steels corresponding to varying proportions of carbon and of the special
element.

II. Explain the influence of some special elements on the position of the critical
points.

III. Explain the superiority of pearlitic special steels over pearlitic carbon steels.

IV. Explain any difference which may exist between the martensite of special

steels formed on slow cooling and the martensite of carbon steel produced by
quick cooling.

V. Explain the possible production in the case hardening of some pearlitic special
steels of martensitic or even austenitic cases without quenching.

LESSON XVIII

SPECIAL STEELS

CONSTITUTION, PROPERTIES, TREATMENT, AND. USJSS_ OF MOST
IMPORTANT TYPES

The present lesson is devoted to a brief consideration of the composition, struc-
ture, properties, treatments, and uses of those special steels which have been found to
be of commercial value, namely, nickel, manganese, tungsten, chromium, vanadium,

Aus1~en i fie

Martens it ic

Pearl iti c

O 0,4 0.8 1.2. /.6

%Carbon

Fig. 1. Nickel steel. Constitutional diagram. (Guillet.)

silicon, chrome-nickel, chrome-vanadium, chrome-tungsten, and chrome-molybdenum
steels.

Nickel Steel. Nickel apparently dissolves in iron in all proportions. The con-
stitutional diagram of nickel steel is illustrated in Figure 1 after Guillet. In view of
the explanation of such diagrams given in the preceding lesson it will be readily under-
stood. It shows that as the carbon increases from to 1.60 per cent and the nickel
from to 30 per cent the metal which at first remains pearlitic becomes martensitic

LESSON XVIII SPECIAL STEELS

and finally austenitic. The nickel steels of greatest commercial importance seldom
containing more than 5 per cent nickel and one per cent carbon are pearlitic. This
influence of nickel in preventing partly or wholly the transformation of austenite into
pearlite is due to its lowering the critical point of the steel as fully explained in the last
lesson and as illustrated graphically in Figure 2 in the case of iron-nickel alloys con-
taining small proportions of carbon. It should be remembered that the presence of
larger quantities of carbon would intensify the influence of nickel, i.e. would cause
the critical points to be further depressed. The diagram shows that as the nickel in-
creases from to some 25 per cent both transformations, AB on cooling and A'B' on
heating, are depressed, the former, however, much more quickly than the latter, re-
sulting in a rapidly increasing gap between the two transformations. In other words,
nickel up to 25 per cent greatly increases the hysteresis. Taking a steel, for instance,
with 10 per cent nickel cooling from a high temperature, it remains non-magnetic

A
700

600
500

.oo

300
200
100

._5

SO 30 liO SO 60 '0 80 90 IOQ

Fig. 2. Influence of nickel on the critical points of
iron. (Osmond.)

until a temperature of some 400 deg. C. is reached when it undergoes the magnetic
and other transformations. On reheating this magnetic steel, however, it does not
again lose its magnetism until a temperature of some 675 deg. is attained. Between
400 and 675 deg. this nickel steel will be magnetic, therefore, in case its last transfor-
mation resulted from cooling below 400 deg. and it will be non-magnetic if it resulted
from heating above 675 deg. When this hysteresis gap between the transformation is
considerable the alloys are said to be irreversible, meaning by that expression that the
reverse transformation cannot be produced at or near the same temperature. Nickel
steels containing between and 25 per cent are therefore often spoken of as irreversible
alloys. It should be noted, however, that when the nickel content does not exceed
some 3 per cent the alloys are really reversible, that is, the gap between the critical
transformations on heating and cooling is not excessive. According to Osmond, for
instance, with 3.82 per cent nickel the critical point on heating occurs at 710 deg.
and on cooling at 628 deg. A gap of 100 deg. might be arbitrarily selected as a line
of demarcation between reversible and irreversible alloys. Returning to Figure 2 it
will be seen that as the nickel content increases above 25 per cent the transformations

LESSON XVIII SPECIAL STEELS 3

become abruptly reversible (the gap between them not exceeding 50 deg.), that
their position is now gradually lifted, reaching a maximum for about 70 per cent
nickel, and that it is then again lowered. Iron-nickel alloys containing more than 25
per cent nickel are therefore reversible.

The diagram also shows that with some 25 per cent nickel the transformation is
lowered below atmospheric temperature which means that the metal on cooling from
above B' remains non-magnetic at atmospheric temperature and that its iron, there-
fore, is in the gamma condition and its structure austenitic.

An attempt has been made in Figure 3 to construct a diagram indicating the re-
lation existing between carbon content, nickel content, positiorrof- the critical points
on cooling, and corresponding types of structure as explained in Lesson XVII.

700

75 Id. 125 15. 175 2O. 225 25.

Nickel % O 23

Fig. 3. Influence of nickel and carbon on the position of the critical point An and corresponding

types of structure.

As already stated the pearlitic nickel steels are those most widely used. In the
majority of cases the nickel content does not exceed 3.50 per cent while the carbon
content is seldom over 0.50 per cent. These steels compared with carbon steels of
equal ductility have a considerably higher strength and especially higher elastic limit,
while compared with carbon steels of like elastic limit they have much greater duc-
tility. To explain this in another way, the introduction of some 3.50 per cent nickel
in a 0.50 per cent steel, for instance, raises its elastic limit very considerably while
decreasing its ductility but slightly. Pearlitic nickel steels are also somewhat harder
than carbon steels of like properties, hence better able to resist wear. When properly
heat treated their ability to resist shocks is likewise greater.

The structure of pearlitic nickel steel is shown in Figure 4. On comparing it with
that of carbon steel of like carbon content it will be noted that the pearlite particles
arc somewhat sharper and more angular and the fcrrite grains smaller. When exam-

4 LESSON XVIII SPECIAL STEELS

ined under high magnification the nickel pearlite is seldom found as distinctly lami-
nated as ordinary pearlite.

The hardening and annealing of nickel steels should be conducted at lower temper-
atures than the hardening and annealing of ordinary steels of similar carbon content
since their critical points occur at lower temperatures. From the evidences at hand
it would seem as if between and 5 per cent nickel, and in the case of low carbon
steels, each one per cent of nickel lowered the Ar! point some 20 deg. C. and the Aci
point some 10 deg. In the nickel pearlitic steels of commerce, therefore, the points
Ari and Aci should occur at or near the temperatures indicated in the following table
according to their percentage of nickel.

Fig. 4. Nickel steel. Carbon about 0.30 per cent.
Nickel about 3 per cent. Magnified 100 diameters.
(G. A. Reinhardt in the author's laboratory.)

% Ni

Aci

750

700

0.50

745

690

1.00

740

680

1.50

..... 735

670

2.00

730

660

2.50

725

650

3.00

720

640

3.50

715

630

4.00

710

620

4.50

705

610

5.00 .

. 700 .

. 600

While nickel retards the carburization of iron by case hardening, the cores of nickel
steel articles are not coarsened by the high temperature of the carburizing operation
to the same extent as carbon steel cores, so that one treatment is often sufficient,
namely reheating to and quenching from a temperature slightly superior to the crit-

LESSON XVIII SPECIAL STEELS 5

ical range of the case, that is, to some 700 to 750 deg. in the presence of some 3 or 3.5
per cent nickel. Higher nickel contents call for lower quenching temperatures.

The case hardening of nickel steels offers the possibility already alluded to of pro-
ducing a martensitic case without quenching. Nickel steel, for instance, containing
not over 0.25 per cent carbon and some 3.50 or more per cent nickel can readily be
made martensitic near the outside by case hardening followed by air cooling as shown
in Figures 5 and 6. The martensitic grains owe their polyhedral form to the original
austenitic grains from which they are derived. The thickness of the mastensitic case
is about 0.5 mm. The occurrence of troostite should be noted. Under lower mag-
nification (Fig. 6) a solid troostite band is seen to separate the_martensitic and the
sorbito-pearlitic portions. With a little more carbon and nickel martenso-austenitic
cases may be produced as shown in Figures 7 and 8.

Nickel steels that are martensitic as cast are not utilized because like all mar-
tensitic steels they are hard, brittle, and cannot be machined. Their case hardening
should result in the formation of ductile, austenitic cases. Nickel steels which are
martensitic after air cooling may be troostitic, sorbitic, or even pearlitic after very
slow cooling in the furnace, while they may become austenitic on water quenching.
The structure of martensitic nickel steel is shown in Figure 9.

Austenitic nickel steels are not widely used, the high carbon, high manganese
steels being preferred when an austenitic steel is desired, in part at least because of
their lower cost. Like all austenitic steels they are non-magnetic, ductile, very diffi-
cult to machine and have a low elastic limit. Their structure is polyhedric (see Fig. 10),
Some types of austenitic nickel steels have, however, found interesting applications
based chiefly on the marked influence of nickel on the dilatation of the metal. With
36 per cent nickel, for instance, the dilatation is nearly nil and the resulting alloy,
discovered by Guillaume, and called by him "invar" is used successfully for the con-
struction of clocks and other instruments of precision. With some 46 per cent of
nickel and 0.15 per cent carbon the coefficient of dilatation is nearly the same as that
of glass and alloys of that composition called "platinite" are used in place of plati-
num for the construction of incandescent electric lamps. Austenitic nickel steel, like
all austenitic special steels, may be made martensitic and thereby regain its magnetism
by immersion in liquid air. The increase of volume which accompanies this trans-
formation produces a swelling of the polished surface which because of the resulting
relief effect renders the structure of the metal apparent without etching, as shown in
Figure 11.

Manganese Steel. Manganese, when alloyed with iron and carbon in large pro-
portion is partly dissolved in the iron and partly present as a double carbide of iron
and manganese. From this behavior of manganese the structural types formed by in-
creasing both carbon and manganese may be anticipated. The steel should at first
remain pearlitic and then become in succession martensitic and austenitic. With
much manganese and carbon, however, the separation of carbide is to be expected.
The constitutional diagram of manganese steels is shown in Figure 12 after Guillet,
while a critical point structural diagram has been constructed tentatively in Figure 13.
By comparing the constitutional diagram of manganese steel with that of nickel steel
it will be noted that manganese is, roughly stated, twice as effective as nickel in pro-
ducing a certain type of structure, as for instance in converting pearlitic into marten-
sitic steel.

Manganese is present in appreciable quantities in all ordinary carbon steels but

LESSON XVIII SPECIAL STEELS

Fig. 5. Nickel steel. Nickel 3.44 per cent. Carbon 0.176 per cent. Case
hardened and air cooled. Magnified 100 diameters. (G. A. Reinhardt in
the author's laboratory.)

Fig. 6. Same steel as in Fig. 5. Same treatment. Magnified 50 diameters.
(G. A. Reinhardt in the author's laboratory.)

LESSON XVIII SPECIAL STEELS

Fig. 7. Nickel steel. Nickel 4.86 per cent. Carbon 0.115 per cent. Case
hardened and air cooled. Magnified 100 diameters. (G. A. Reinhardt in the

author's laboratory.)

Same steel as in Fig. 7. Same treatment. Magnified 300 diameters.
(G. A. Reinhardt in the author's laboratory.)

LESSON XVIII SPECIAL STEELS

unless the latter contain considerably more than one per cent of that element they are
not regarded as manganese steels. With carbon not exceeding 0.80 per cent and man-
ganese not exceeding some 3 per cent the steels remain pearlitic and, therefore, not
unlike the pearlitic nickel steels so widely used. Manganese pearlitic steels, however,

Fig. 9. Nickel steel. Cast. Nickel 10 per
cent. Carbon 0.80 per cent. Magnified 300
diameters. (Guillet.)

Fig. 10. Nickel steel. Nickel 25 per cent.
Carbon 0.80 per cent. Magnified 300
diameters. (Osmond.)

Fig. 11. Nickel steel. Nickel 15 per
cent. Carbon 0.80 per cent. Cooled in
liquid air (-180 deg. C.). Not etched.
Magnified 300 diameters. (Guillet.)

are practically ignored by steel manufacturers and users apparently (1) because of
the wide-spread belief that such steels are brittle, and (2) because of the difficulty of
manufacturing low carbon manganese steels. The belief in the brittleness of pearlitic
manganese steels is founded on Hadfield's statement that between 2 and 6 per cent of
manganese the steels are hopelessly brittle. On closer examination, however, it would

LESSON XVIII SPECIAL STEELS

0) /a

A u s /e n i f i c

Martens/ 1 ic

Pea rl iti c

0.4- O.d 1.2.

% Carbon

Fig. 12. Manganese steel. Constitutional diagram. (Guillet.)

700.

60Q

Fig. 13. Influence of manganese and
carbon on position of critical point Ari
and corresponding types of structure.

LESSON XVIII SPECIAL STEELS

seem as if this statement was true only in the case of rather high carbon steels cooled
relatively quickly. Evidences have since been offered, notably by Guillet, showing
that low carbon pearlitic manganese steel slowly cooled is not brittle. These steels have
been difficult to manufacture because of the necessity of using ferro-manganese from
the blast furnace and therefore high in carbon, thereby introducing much carbon in
the steel. At present, however, nearly carbonless ferro-manganese is produced in
electric furnaces and also by the thermit process. It is also claimed that low carbon
manganese steels can be successfully produced in the electric furnace under suitable
oxidizing conditions and at high temperature when carbon may be oxidized in prefer-
ence to manganese. If the physical properties of low carbon pearlitic manganese
steel are at all comparable to those of pearlitic nickel steel and if manganese steel
can be manufactured more cheaply than nickel steel of like properties the manufac-

Fig. 14. Manganese steel. Austenitic. Cast. Magnified 50 diameters.

ture and testing of low carbon manganese steel should receive more attention. The
possibility of producing by case hardening articles having pearlitic cores and austen-
itic cases also deserves some consideration. Martensitic manganese steels are not
utilized because of the hardness and brittleness which they share with all martensitic
steels. Those manganese steels whose composition is near the boundary line between
the pearlitic and martensitic regions while martensitic after air cooling may be troos-
titic or even pearlitic after very slow cooling while they may become austenitic on
quenching.

Austenitic manganese steel is of considerable industrial importance. It is often
called from the name of its inventor " Hadfield " steel. It generally contains from
10 to 15 per cent manganese and from one to 1.5 per cent carbon. In its cast condition
it is weak and has but little ductility probably because of the presence of a consider-
able quantity of free carbide. On being heated to a high temperature (1000 deg. C.
or more), however, and quenched in water or oil its tenacity is greatly raised and it

LESSON XVIII SPECIAL STEELS

becomes very ductile; the treatment being often called on that account " water tough-
ening." The marked change of properties resulting from it is probably due to the ab-

Fig. 15. Same as in Fig. 14. Magnified 300
diameters.

sorption of the carbide at a high temperature and its retention in solution by quick
cooling. The structure of manganese steel both in its cast and in its water quenched
condition is shown in Figures 14 to 16. In the cast sample the carbide is seen to occur

Fig. 16. Manganese steel. Austenitic.
Water quenched. Magnified 200 diam-
eters. (Guillet.)

as thick membranes surrounding the austenitic grains and here and there in chunks.
The treated sample is nearly free from carbide and possesses the polyhedric structure
characteristic of gamma iron and of austenite. The properties of austenitic man-

LESSON XVIII SPECIAL STEELS

ganese steel are those of austenite, namely low elastic limit but great hardness and
wearing power combined with much ductility.

Tungsten Steels. Tungsten appears to raise rather than lower the critical points
of iron while it forms with it a double carbide of iron and tungsten from which it may
be safely inferred that tungsten steels will at first remain pearlitic on slow cooling
and that as the percentage of tungsten increases it will become cementitic, that is
it will contain carbide particles. This is shown in Figure 17, which is a reproduction
of the constitutional diagram of tungsten steels according to Guillet. In the presence
of a considerable proportion of tungsten, however, the position of the Ari point

Cemenf/ti c

Pearl i tic

04- Q-Q

/o Carbon

I.G

Fig. 17. Tungsten steel. Constitutional diagram. (Guillet.)

seems to be greatly affected by the temperature from which the steel cools. Os-
mond, for instance, found in the case of a steel containing 0.42 per cent carbon and
6.25 per cent tungsten that heating to and cooling from some 900 deg. reveals the
existence of two critical points respectively at 690 and 650 deg. As the temperature
from which cooling begins increases, however, two interesting phenomena are ob-
served, (1) the upper point occurs at practically the same temperature but becomes
fainter and finally disappears, and (2) the lower point remains pronounced but its
position is gradually lowered. Cooling from 1015 deg., for instance, resulted in a faint
critical point at 670 deg. while the lower point was depressed to 625 deg. ; cooling from
1210 deg. caused the disappearance of the upper point while the lower remained very
pronounced but now occurred at 500 deg. Bohler, likewise experimenting with a steel
containing 0.85 per cent carbon and 7.78 per cent tungsten, reports the existence of a
point at 710 deg. and one at 550 deg., the upper one, however, occurring only when

LESSON XVIII SPECIAL STEELS

the metal has not been heated above 1 100 deg. while the second is only to be detected
when the temperature exceeds 1000 deg. In other words on cooling from above
1100 deg. the lower point only occurs, on cooling from below 1000 deg. only the upper
point is visible, while heating to and cooling from a temperature situated between
1000 and 1100 deg. causes the appearance of both points. It will be shown soon that
this indirect influence of tungsten upon the critical points affords an explanation of
the remarkable properties of self-hardening and high speed steels which are chiefly
tungsten steels.

On heating cementitic tungsten steels to a high temperature the particles of free
carbides are dissolved the more completely the higher the temperature. Air cooling
is often sufficient to retain the carbides in solution while the metal becomes finely
martensitic. Such steels are said to be "self-hardening." 1 The structure of two

Fig. 18. Tungsten steel. Tungsten 27.75
per cent. Carbon 0.276 per cent. Mag-
nified 200 diameters. (Guillet.)

Fig. 19. Tungsten steel. Tungsten 39.96
per cent. Carbon 0.867 per cent. Mag-
nified 200 diameters. (Guillet.)

cementitic tungsten steels is reproduced in Figures 18 and 19. The white particles
are the double carbide.

Tungsten steels are used (1) for springs generally after hardening followed by
tempering, (2) for magnets after hardening only, and (3) for tools as self-hardening
steels. In the latter case, however, considerable manganese is always present, the
resulting alloy being in reality a quaternary steel. High speed steels are quaternary
steels generally containing a large proportion of tungsten; they will soon be described.

Chrome Steels. Chromium forms a double carbide with iron and carbon, while
it has little if any direct influence on the position of the critical points. 2 The presence

1 The presence of manganese or of a little chromium is necessary, however, to impart self-hard-
ening properties to tungsten steel. The original "self" or "air hardening" steel, that is "Mushet"
steel, always contained considerably more than one per cent manganese and was high in carbon.

2 According to some recent observations of Nesselstrauss, chromium lowers the point Ar 3 of hypo-
eutectoid steels, eventually causing its disappearance while it raises a little the point Ari. The
proportion of chromium needed to cause the point Ar 3 to disappear is the smaller, the higher the
carbon content. With 0.20 per cent carbon, 5 per cent chromium are needed. With more carbon a
correspondingly smaller percentage of chromium suffices.

LESSON XVIII SPECIAL STEELS

of chromium, however, like that of tungsten causes the point on cooling to be markedly
bwered as the temperature from which the metal cools increases. Osmond, for in-
stance, found the following relations between the maximum temperature and the po-
sition of the critical point on cooling:

MAXIMUM
TEMPERATURE

835
1030

CRITICAL POINT
ON COOLING

713-716
682-692

MAXIMUM
TEMPERATURE

1220
1320

CRITICAL POINT
ON COOLINO

635-643
600-640

o.Q 1.2,

% Carbon .

Fig. 20. Chrome steel. Constitutional diagram. (Guillet.)

/.6

The constitutional diagram of chrome steels is shown in Figure 20 after Guillet.
Reheating cementitic chrome steel to a high temperature followed by quick cooling
(air or water) results generally in the disappearance of some of the particles of free
carbide. The case hardening of chrome steels yields very hard cases. As the core
coarsens, however, these steels should always receive the double treatment described
in Lesson XVI, that is, (1) reheating and quenching for refining the core, and (2) re-
heating and quenching for refining and hardening the case.

The chrome steels that are utilized seldom contain more than 3 per cent chromium
and are therefore pearlitic after slow cooling; they are used for the manufacture of
armor piercing projectiles, of steel balls, of files, and of some other tools, the presence
of chromium increasing the hardness and the hardening power of the metal.

Becker writes that the hardness imparted by chromium is not accompanied by as
much brittleness as that induced by carbon. According to the same author chromium

LESSOX XVIII SPECIAL STEELS

also has the effect of increasing the elastic limit of steel, especially when it is com-
bined with vanadium. The structure of pearlitic chromium steel resembles that of
pearlitic carbon steel.

Vanadium Steels. Vanadium forms double carbides with iron and has no marked
influence on the positions of the critical points. Unlike tungsten and chromium it
dues not seem to cause the lowering of the Ar points with increasing temperature.
The constitutional diagram of vanadium steels is shown in Figure 21 after Guillet.
Two types of structures are produced, pearlitic and cementitic. Guillet describes the
appearance of the particles of free carbide as being triangular.

According to Guillet heating cementitic vanadium steels to a high temperature

Oeme n // 1 1

o o.s i.o 1.5 2.0

%Carbon

Fig. 21. Vanadium steel. Constitutional diagram. (Guillet.)

fails to cause the absorption of the free carbide, vanadium steels differing in this re-
spect from other steels in which double carbides are formed.

The vanadium steels commercially utilized seldom contain more than 0.50 per
cent vanadium and are, therefore, pearlitic and free from carbides. Their proper-
ties recall those of pearlitic nickel steels, namely high combination of elastic limit and
ductility and high resilience. The very small amount of vanadium sufficient to pro-
duce these results should be noted.

Silicon Steels. Silicon, probably as an iron silicide, FeSi, forms a solid solution
with iron in all proportions and has no very marked influence upon the position of the
critical points from which we may infer that slowly cooled silicon steels will be neither
martensitic nor cementitic. It is a well-known fact, moreover, that silicon has a
marked tendency to cause the formation of graphitic carbon when present over a

LESSON XVIII SPECIAL STEELS

certain proportion especially in high carbon steel. The constitutional diagram of
silicon steels is shown in Figure 22 according to Guillet. It will be seen that the
structure is independent of the carbon content being entirely regulated by the pro-
portion of silicon. As long as the proportion of silicon does not exceed 5 per cent the
steel is pearlitic and the whole of the carbon remains in the combined condition.
Between 5 and 7 per cent of silicon some pearlite is still present and, hence, some com-
bined carbon, but graphitic carbon also occurs; between 7 and 20 per cent of silicon
the whole of the carbon is in the graphitic condition, the balance of the steel consist-
ing of a solid solution of the silicide FeSi in iron (silico-ferrite), and also, according to

^o.

o
o

10.

Graphite +Fe Si.

Grap kite + s>oluf/on

Pearl ite+graphii~e + so Sufi on Fe 3 i.

Pearlfte+solufion

1.0 A5

% Carhon

Fig. 22. Silicon steel. Constitutional diagram. (Guillet.)

Guillet, of some Fe 2 Si likewise dissolved in iron. With more than 20 per cent of sili-
con the steel is composed of graphite and of FeSi.

It should be noted that the only silicon steels utilized contain less than 5 per cent
of silicon and are, therefore, pearlitic and free from graphitic carbon unless indeed
subjected to prolonged annealing or very slow cooling when some graphitic carbon
will form, especially in high carbon steels. This formation of graphitic carbon takes
place the more readily the higher the temperature, the longer the time at a high tem-
perature, the more silicon and the more carbon in the steel.

Silicon steels are chiefly used in the construction of dynamos because of their
low magnetic hysteresis and high permeability and, with the addition of quite a little
manganese, for springs and for certain parts of automobiles.

Chrome-Nickel Steels. From our knowledge of the constitution of nickel steels
and of chromium steels it is possible to foretell the constitution of the quaternary
chrome-nickel steels. Nickel steels being pearlito-martenso-austenitic and chrome

LESSON XVIII SPECIAL STEELS 17

steels being chiefly pearlito-cementitic, we may expect that in chrome-nickel steels,
as the proportions of carbon, nickel, and chromium increase the steels, at first pear-
litic, will become in turn martensitic, austenitic, and cementitic. In the case of ce-
mentitic steels the free carbides will be embedded in a martensitic or austenitic matrix
according to their composition. It should also be expected that increasing the pro-
portion of nickel (and carbon) will cause the steel to pass from the pearlitic to the
martensitic condition, etc., while increasing the chromium (and carbon) will make it
cementitic. The presence of both nickel and chromium in the same steel produces a
metal possessing the valuable qualities of both nickel and chromium steels, namely,
high elastic limit combined with high ductility, greater hardness^ hardening power,
resilience, and better wearing qualities than carbon steels. Chrome-nickel steels are
especially valuable in the construction of parts to be hardened and tempered when
they yield a finely martensitic structure having greater shock resisting power than the
martensite of carbon steels. Practically the only chrome-nickel steels utilized are the
pearlitic ones containing therefore moderate amounts of carbon, nickel, and chro-
mium. They are used extensively in automobile construction and for the manufacture
of armor plates. In the latter case, of course, one face of the plates is subjected to
the case hardening treatment. The case hardening of nickel-chromium steel resembles
that of nickel steel. The metal should be reheated, after the case hardening operation,
slightly above the critical range of the case and quenched. Because of the presence
of nickel it is not so imperative to heat to and quench from a temperature superior to
the critical range of the core before hardening the case, although such procedure would
probably yield a tougher core.

Quaternary Vanadium Steels. The introduction of a small amount of vanadium
into the various special steels has been strongly urged and nickel-vanadium, chrome-
vanadium, chrome-nickel-vanadium and chrome-tungsten-vanadium steels have been
used. It is claimed that the presence of a small proportion of vanadium (less than
0.50 per cent) increases the soundness of castings and their freedom from occluded
gases and that it adds to the desirable physical qualities of forgings such as strength,
resilience, ductility, etc. Guillet writes that nickel-vanadium steel properly hardened
by quenching is relatively so tough that, unlike other hardened steels, it does not
require tempering. Since vanadium forms a double carbide with iron its presence in
steel is likely to make it cementitic. In nickel steel and in nickel-chromium steel the
matrix will be pearlitic, martensitic, or austenitic in accordance with the proportion
of nickel and carbon present; in chrome steels it will be pearlitic or martensitic.

Chrome-Tungsten or High Speed Steels. Since both chromium and tungsten
form carbides and since they do not lower the Ari points, at least directly, it may
be fairly anticipated that slowly cooled chromium-tungsten steels will be cementitic
with a pearlitic, sorbitic, or even troostitic matrix. Upon being heated to a high tem-
perature the carbide particles are dissolved and if the cooling that follows be suffi-
ciently rapid they are retained in solution, the metal acquiring a finely martensitic
structure. To cause a complete absorption of the free carbide, however, a very high
temperature is often required, in some cases approaching the melting point of the steel,
while air cooling is frequently sufficiently rapid to prevent the carbide from again
forming. After such treatment these steels although fully hardened are in a condition
relatively so stable that they may be heated to a visibly red heat, i.e. to some 600
deg. C., before their martensite undergoes any marked transformation. This invalu-
able property makes it possible, with tools made of such steels suitably treated, to cut

LESSON XVIII SPECIAL STEELS

steel and other hard metals at such speed that the cutting edge of the tool becomes
visibly red hot before breaking down. These steels are known in consequence as
high speed steels. Their discovery by Taylor and White, at the time in the employ
of the Bethlehem Steel Company, South Bethlehem, Pennsylvania, marks one of the
most distinct and revolutionary advances ever made in the metallurgy of iron and
steel. The composition of these steels varies greatly: they may contain from 0.25 to
one per cent carbon, generally not over 0.60 per cent; from 5 to 25 per cent of tungsten,
generally between 10 and 20 per cent; from 2 to 10 per cent of chromium, generally
between 2 and 8 per cent, and seldom over 0.40 per cent of manganese. In some types
tungsten is replaced in part or wholly by molybdenum; in others a small amount of
molybdenum is present in addition to the tungsten and chromium, while in others still
a small amount of vanadium (0.2 to 0.4 per cent) occurs. Properly treated high speed

Fig. 23. High speed steel. Typical
structure after correct heat treat-
ment. Magnified 1000 diameters.
(Edwards.)

Fig. 24. High speed steel. Typical
structure after annealing. Magnified
150 diameters. (Edwards.)

steels often exhibit a polyhedric structure (Fig. 23) quite, if not altogether, free from
carbide particles. The structure of a similar steel annealed is reproduced in Figure 24.

The inventors of high speed steels recommended the following treatments as yield-
ing the best results: (1) heating the tool slowly to about 815 deg. C., then quickly
until its extreme edge showed indications of melting, (2) cooling the tool quickly to
below 860 deg. and then quickly or slowly to atmospheric temperature, and (3) re-
heating the tool to about 640 deg. for five minutes (in a lead bath) followed by cooling
in air. The author believes, however, that tools of high speed steel are now generally
heated to near their melting-point followed by cooling freely in air or in an air blast,
a second treatment being rarely applied.

The remarkable properties of high speed steels briefly outlined in the foregoing
pages must be ascribed to the formation in those steels of a martensitic structure
more stable than that of quenched high carbon steel and possibly also possessing su-
perior cutting qualities. An explanation of the greater stability and better quality
of the martensite of high speed steels is suggested at least by the well-known indirect
influence of chromium and tungsten on the critical points. It has been explained that
while these elements have no marked direct influence upon the location of the critical
points their presence causes the critical point on cooling to be lowered as the tempera-

LESSON XVIII SPECIAL STEELS

ture from which cooling begins, increases. 1 This influence is shown graphically in
Figure 25 in which the line AB represents the maximum temperatures reached before
cooling and CD the position of the critical point corresponding to these temperatures.
As AB ascends CD descends. If it be considered that martensite forms while the steel
cools through its critical point it follows that as the steel is heated to higher tempera-
tures its martensitic structure forms at gradually decreasing temperatures. And may
we not conceive that the lower the temperature at which martensite forms the greater
its stability and the better its cutting properties? Might not its superior cutting qual-

/ooo.

00.

500.

-4-00.

Fig. 25. High speed steel. Relation between heating temperature, po-
sition of the ATI point, and the stability of resulting martensite.

ity be due to its having been formed while the metal was cooler, i.e. stiffer and there-
fore opposing more effectively its transformation? It should be more stable on the
ground that the greater the range of temperature between the temperature of a con-
stituent and the point at which its transformation was due the less stable its condi-
tion. In Figure 25, for instance, the decreasing distances MN ', M'N', and M"N"
may be considered as proportional to the instability at atmospheric temperature of
the martensitic structures resulting from cooling respectively from L, L', and L"
and formed in passing respectively through M , M', and M" . If this reasoning be cor-
rect it follows that the lower the temperature at which martensite is formed the greater

1 This was shown by Osmond several years before the introduction of high speed steel and it is
natural to infer that Taylor and White were guided by Osmond's discovery.

20 LESSON XVIII SPECIAL STEELS

must be its stability and, therefore, the higher the temperature to which it may be
heated before undergoing the tempering transformation which deprives it of its cutting
properties.

Edwards believes (1) that heating high speed steel to 1200 deg. C. results in the
formation of a carbide of tungsten which is dissolved by the iron, (2) that the critical
point then exhibited by the steel at about 380 deg. is not the point Ar t lowered by
the presence of tungsten and chromium but that it marks a change occurring in the
carbide of tungsten, (3) that heating high speed steel to 1320 deg. C. causes the for-
mation of a double carbide of tungsten and chromium held in solution by the iron,
even in slow cooling, thus explaining the absence of critical point below 900 deg. in
steel so treated, and (4) that the failing of a high speed tool is to be attributed to the
formation of a new brittle constituent which can be produced by reheating the steel to
700 deg.

Experiments

The student should procure some samples of special steels, preferably the follow-
ing: nickel steel containing between 3 and 3.50 per cent nickel and between 0.25 and
0.50 per cent carbon; cast manganese steel containing between 10 and 15 per cent
manganese and between one and 1.5 per cent of carbon; chrome-tungsten (high speed)
steel of good commercial quality.

Nickel Steel. In its cast or forged condition the nickel steel selected is pearlitic.
Its structure should be normalized by heating to 900 or 1000 deg. C. followed by slow
cooling and examined. The differences in appearance between the pearlite particles
and those of ordinary carbon steel subjected to like treatment (Lesson V) should be
noted. Samples of this steel may be hardened and case hardened to verify the ac-
curacy of the statements made (1) as to the lower temperature needed for hardening,
and (2) as to the possibility of producing martensitic cases without quenching.

Those students who have the necessary apparatus are advised to determine the
critical points of this steel comparing their results with the temperatures indicated in
the lesson.

Manganese Steel. The sample of manganese steel is austenitic. Its structure
in the cast condition should be examined and should be found to contain many car-
bide particles, possibly forming continuous membranes around the austenite grains.
A sample of this steel should be heated to 1000 or 1100 deg. and quenched in water.
Its structure should now be purely austenitic (polyhedric) quite if not altogether
free from carbide particles. The absence of any critical point in this steel should be
ascertained; cooling from a high to atmospheric temperature should yield a smooth
curve free from heat evolutions.

High Speed Steel. The structure of this steel should be examined both in the
forged condition and after reheating to some 1200 deg. followed by air cooling. The
influence of tungsten and chromium in causing the lowering of the critical points as
the temperature of cooling increases may be verified by those having the necessary
facilities.

Photomicrographs should be taken from all structures at suitable magnifications.

Examination

Describe briefly the constitution, properties, treatments, and uses of (1) nickel steel,
(2) manganese steel, (3) nickel-chromium steel, and (4) high speed steel.

LESSON XIX

CAST IRON

Cast iron differs from steel in being deprived of malleability. This lack of malle-
ability is due to the presence of a large quantity of carbon, generally between 2.50 and
4.00 per cent. The carbon present in cast iron may be (1) wholly in the graphitic
condition, (2) wholly in the combined condition, and (3) partly in the graphitic and
partly in the combined condition. These three types of cast iron should be separately
considered. The factors influencing the formation of graphite or of combined carbon
should, however, first be recalled.

Formation of Combined and Graphitic Carbon. It will be explained in Lesson
XXIII that when cast iron solidifies the carbon probably remains in the combined
condition and that the resulting carbide, Fe 3 C, is partly free and partly in solid solu-
tion in the iron. This FesC, however, is an unstable compound and when formed at
a high temperature it is readily decomposed into graphite and iron according to the
reaction

' Fe 3 C = 3 Fe + C

hence the formation of graphitic carbon in cast iron. Two factors are conspicuous
in promoting the formation of graphitic carbon, (1) a slow rate of cooling through and
below the solidification period and (2) the presence of silicon. The gray, i.e. graphi-
tic, cast irons are generally those which have been cast in sand and hence slowly
cooled and which contain a relatively large percentage of silicon. It follows that
under otherwise identical conditions and compositions a large casting will become
more graphitic on solidifying than a smaller one since it will cool more slowly; also
that of two castings of equal size, cooled under like conditions and of identical com-
position except as to their silicon contents, the one richer in silicon will contain more
graphitic carbon. The conditions most effective in preventing the formation of
graphitic carbon and in promoting, therefore, the retention of carbon in its combined
form are (1) quick rate of cooling through and below the solidification range and (2)
the presence of much sulphur or manganese.

Cast Iron Containing only Graphitic Carbon. Cast irons containing a consider-
able amount of graphitic carbon are known as gray cast irons because of the appear-
ance of their fracture which is grayish or blackish and coarsely crystalline. Cast
irons containing the whole of their carbon in the graphitic condition and therefore
free from combined carbon are extreme types seldom produced. Their structure,
however, should be considered.

Proceeding as we did in the case of steel we shall first assume cast iron to be a pure
alloy of iron and carbon, free, therefore, from its usual impurities. If the whole of
the carbon is in the graphitic condition it is evident that cast iron can only contain
the two constituents graphite and iron or ferrite. We may therefore anticipate its
structure. It will, however, be interesting to study the mode of occurrence of the

2 LESSON XIX CAST IRON

graphitic carbon. The structure of cast iron practically free from combined carbon
is illustrated both before and after etching in Figures 1 and 2. The metal will be seen
to consist of an iron or ferrite matrix in which are embedded many irregular and
generally elongated and curved plates of graphite. These graphite plates break up
so effectively the continuity of the metallic mass as to completely destroy the ductility
and malleability of a substance (ferrite) by nature very ductile and malleable. The
brittleness of highly graphitic cast iron is not due so much to the brittleness of the
graphite it contains nor even to its large proportion of graphite as to the thorough
manner in which the continuity of its otherwise ductile matrix is destroyed by the
shape and distribution of the graphite particles.

Fig. 1. Gray cast iron free from combined carbon. Magnified
100 diameters. Not etched. (F. C. Langenberg in the author's
laboratory.)

It will be seen in another lesson that when the graphite occurs in small rounded
particles as it does in malleable cast iron the ferrite matrix may retain considerable
ductility and malleability.

The ferrite matrix of this highly graphitic cast iron (Fig. 2) will be seen to be made
up of the polyhedric crystalline grains characteristic of carbonless iron, the ferrite of
cast iron being similar in this and other respects to the ferrite of wrought iron and of
hypo-eutectoid steel. In impure cast iron it undoubtedly holds in solution silicon
and possibly, to some extent, other impurities.

Highly graphitic cast iron is brittle and deprived of ductility and malleability
because of the presence of numerous plates of graphitic carbon; it is weak because of
the presence of graphite plates and because of the relative weakness of its matrix; it
is soft and therefore easily machined because of the softness both of its matrix and of
the graphite it contains; it expands on solidifying because of the formation, with

LESSON XIX CAST IRON 3

increase of hulk, during solidification of a large amount of graphitic carbon. It
should be noted that because of its low specific gravity graphite will occupy a rela-
tively large proportion of the bulk of the metal. Cast iron, for instance, containing
by weight 3 per cent of graphite contains by volume some 12 per cent of that element.
As might lie expected the rate of solidification and further cooling has some influ-
ence both upon the shape and size of the graphite particles as well as upon the size of
the ferrite grains, and therefore upon the physical properties of the metal, very slow
solidification promoting the formation of large graphite plates and of large ferrite
grains. Were it possible to cause the graphite in cast iron to occur in small rounded

Fig. 2. Same metal as in Fig. 2. Magnified 100 diameters
Etched. (F. C. Langenberg in the author's laboratory.)

particles instead of as sharp, curved plates, its ductility and strength would undoubt-
edly be greatly increased.

The diagram of Figure 3 shows graphically the structural composition of iron-
carbon alloys in which the whole of the carbon occurs as graphite. Only those alloys,
however, containing from 3 to 4.5 per cent carbon can be produced. Indeed in the
absence of silicon even these are quite unobtainable. With less than 3 per cent car-
bon it is well nigh impossible to prevent the retention of some combined carbon, while
with less than 2 or at least with less than 1.50 per cent carbon the whole of the carbon
is likely to be in the combined condition. The diagram, therefore, is only a theoretical
one. It has, nevertheless, its interest for it will be shown in another lesson to repre-
sent the stable and final equilibrium of the iron-carbon system. The percentage of
graphite by volume has also been indicated.

Cast Iron Containing only Combined Carbon. Cast iron containing only com-
bined carbon and free, therefore, from graphitic carbon is called "white" cast iron

4 LESSON XIX CAST IRON

from the aspect of its fracture which is white, brilliant, and highly metallic. The
absence of graphitic carbon is generally due (1) to the presence of much manganese
and sulphur and of little silicon, (2) to quick cooling through and below the solidifica-
tion period, or (3) to both low silicon, high manganese and sulphur, and quick solidi-
fication. Cast iron, for instance, may contain so much sulphur and manganese and
so little silicon as to be white even after slow solidification or it may solidify so quickly
as to be white even in the presence of much silicon and little manganese and sulphur.
A familiar instance of the marked influence of the rate of cooling is afforded by the
casting in the metal'molds of casting machines of cast iron which if cast in sand would

Cast /ron
free from

IOO

25.

Graph /Ye
weighi".

Ferr/fe.

0~ ~/~ ~~ ~J~ 4

Percent Carbon.

Fig. 3. Structural composition diagram of iron-carbon alloys free from combined carbon.

have been gray, whereas it is now white. Small castings since they cool more quickly
become white more readily than larger ones.

In the absence of graphitic carbon the structure of cast iron should resemble the
structure of a very high carbon steel, i. e. it should consist after slow cooling of pearlite
and of a large amount of free cementite. This is found to be the case as shown in
Figure 4 in which is illustrated the structure of white iron containing about 3 per cent
of combined carbon. Theoretically this alloy should contain nearly 63 per cent of
pearlite and 37 per cent of free cementite. The dark constituent in the photograph
is pearlite, the light one, visibly in relief, cementite. The structure of white cast
iron is also shown in Figure 5 under high magnification, the laminations of pearlite
being clearly seen. The structure of white cast iron is further illustrated in Figures
6 and 7 after Guillet. These photomicrographs are reproduced here because they

LESSON XIX CAST IRON 5

afford an interesting example of the action of sodium picrate (Lesson V, page 7)
in coloring free cementite while leaving pearlite uncolored.

The structural composition of white cast iron is to be calculated like the composi-
tion of any hyper-eutectoid steel of known carbon content as explained in Lesson V,
the following relation existing between the percentage of pearlite and that of carbon
in the iron, on the assumption that pearlite contains 0.834 per cent carbon:

P =

800 - 120 C

the balance of the metal being of course free cementite. While -structurally it re-
sembles high carbon steel, white cast iron is deprived of malleability being indeed very
brittle and very hard. This brittleness and hardness are due to the very large pro-
portion of free cementite present which itself is very hard and brittle.

Fig. 4. White cast iron. Magnified 56 diameters.

Fig. 5. White cast iron. Magnified
500 diameters. (Wtist.)

It will be evident that, starting from carbonless iron, as the carbon increases at
first low carbon steel is produced and then in succession medium high carbon steel,
high carbon steel, and finally white cast iron, each metal passing gradually into the
next without any sharp line of demarcation between them. It is logical to base the
distinction between high carbon steel and white cast iron upon the malleability of
the former and the non-malleability of the latter and this is altogether a question of-
carbon content. The dividing line may be drawn somewhat arbitrarily at 2 per cent
carbon. As a matter of fact steels are very seldom manufactured containing more
than 1.75 per cent carbon while white cast iron rarely contains less than 2.25 per cent
carbon. Between the steel series, therefore, and the white cast-iron series there is a
natural gap, the existence of which generally removes any doubt as to the nature of
the metal under examination.

Again if the process of manufacture be known there need be no doubt as to the
classification of any highly carburized iron alloy: if made in the blast furnace from

LESSON XIX CAST IRON

the reduction of iron ore, it is cast iron, while if it is the product of refining cast iron
(by the Bessemer or the open hearth processes), or of the remelting under oxidizing
conditions of iron or steel scrap with or without cast iron (open hearth process), or of
the carburizing of wrought iron (cementation process), or of the carburizing and melt-
ing of wrought iron (crucible process), it is steel.

The diagram of Figure 8 indicates graphically the structural composition both
proximate and ultimate of any iron carbon alloy containing from to 6.67 per cent
combined carbon, that is, from 100 per cent ferrite to 100 per cent cementite assuming
as it has been done before that pearlite contains 0.834 per cent carbon.

It will be noted that two sources of ferrite are indicated in the diagram, namely,
(1) pearlite (eutectoid) ferrite and (2) pro-eutectoid or free ferrite, the sum of both
being known as total ferrite, while four sources of cementite are to be considered,
namely, (1) pearlite (eutectoid) cementite, (2) pro-eutectoid cementite, (3) eutectic
cementite, and (4) pro-eutectic cementite, the sum of all four being known as total

Fig. 6. White cast iron. Magnified 200 diam-
eters. Etched with picric acid. (Guillet.)

Fig. 7. .Sumo metal as in Fig. 6. Magni-
fied 200 diameters. Etched with sodium
picrate. (Guillet.)

cementite and that of (2), (3), and (4) as free cementite. Let us recall the meaning
of these terms:

Pearlite or eutectoid ferrite is the ferrite included in pearlite.

Free or pro-eutectoid ferrite is the ferrite liberated as hypo-eutectoid steel cools
slowly from its upper critical point (Ar 3 or Ar 3 . 2 ) to its lower point (Art).

Pearlite or eutectoid cementite is the cementite included in pearlite.

Pro-eutectoid cementite is the cementite that is liberated as hyper-eutectoid metal
cools slowly from its upper critical point (Ar cm ) to its lower point (Ar 3 . 2 .i).

Eutectic cementite is the cementite included in the austenite-cementite eutectic
which forms at the end of the solidification of alloys containing more than some 1.70
per cent carbon as explained in Lesson XXIII.

Pro-eutectic cementite is the cementite which forms between the bcgmnin: ; ml
the end of the solidification of alloys containing more than 4.3 per cent carbon as
explained in Lesson XXIII. Howe calls it "primary" cementite.

The portions of the diagram corresponding respectively to the steel series and to
the white cast-iron series have also been indicated leaving two groups of alloys un-

LESSON XIX CAST IRON

Fig. S. Structural composition diagram of iron-carbon alloys free from graphitic carbon.

LESSON XIX CAST IRON

represented by industrial products, namely, those containing between 1.70 per cent
and 2.25 per cent carbon and those containing more than 5 per cent of carbon. 1 The
steel portion of this diagram has been used in Lesson V.

Cast Iron Containing both Combined and Graphitic Carbon. Cast-iron castings
nearly always contain both combined and graphitic carbon. In the majority of cases
they contain from 0.25 to 1.50 per cent of combined carbon, the balance of that ele-
ment being in the graphitic condition. The chief factors affecting this distribution
of carbon between the combined and the graphitic states have already been alluded
to; they are (1) the rate of cooling during and below solidification (hence the size of
the castings) and (2) the presence of silicon, manganese, and sulphur, the first element
promoting, and the last two opposing, the formation of graphitic carbon. If it be

Fig. 9. Gray cast iron. Hypo-cutectoid matrix
(0.25 per cent combined carbon). Magnified 100
diameters. (Boylston.)

considered (1) that graphitic carbon is soft, (2) that the presence of much graphite,
since it implies little combined carbon, means the occurrence of soft ferrite in the cast
iron, (3) that combined carbon (cementite) is very hard, and (4) as later explained,
that the presence of combined carbon, at least up to a certain proportion, greatly in-
creases the strength of cast iron, it will be evident that the physical properties of cast
iron, especially its strength and hardness, will depend chiefly upon the proportion of
combined and graphitic carbon it contains.

Let us first consider the structure of cast iron containing a small amount,tsay 0.25

1 It should be remembered that we are considering here pure alloys of iron and carbon. In the
presence of much manganese (or chromium) iron may contain as much as 6.50 per cent and even much
more carbon, while in their absence cast iron very seldom contains more than 4.5 per cent carbon.
Again some unimportant products are offered for sale under the name of semi-steel which may contain
between 1.75 and 2.50 per cent of total carbon entirely combined, forming, therefore, a sort of connects
ing link between the steel and the cast-iron series. They are frequently obtained by remelting cast
iron in cupola furnaces in the presence of considerable iron or steel scrap. Washed metal likewise
may contain between 1.75 and 2.50 per cent of total and entirely combined carbon but this is a
semi-finished product resulting from a partial refining only of cast iron.

LESSON XIX CAST IRON 9

per cent, of combined carbon and some 3 per cent of graphitic carbon (Fig. 9). The
combined carbon will yield 15 times its own weight of cementite (0.25 X 15 = 3.75
per cent) and the resulting cementite eight times its own weight of pearlite
(3.75 x 8 = 30 per cent), or more quickly,

0.25 X 120 = 30 per cent pearlite

exactly as in the case of steel.

The cast iron under consideration will therefore contain 30 per cent pearlite, 3
per cent of graphite, and the balance, 67 per cent, necessarily free ferrite. Cast iron
containing graphitic carbon may be considered as being made up_oLtwo distinct parts,
namely, (1) graphite and (2) a metallic matrix in which the graphite particles are
embedded. It will be obvious, moreover, that the metallic matrix of gray cast iron
is in reality a steel matrix since it necessarily consists, like steel, of an aggregate of

Fig. 10. Gray cast iron. Eutectoid matrix. Magni-
fied 500 diameters. (Boylston.)

ferrite and cementite partly associated to form pearlite. In cast iron free from com-
bined carbon the matrix is pure ferrite; with a little combined carbon it is of the nature
of a low carbon steel; as the combined carbon increases the metallic matrix is con-
verted into steel of increasing carbon content; with less than some 0.80 per cent car-
bon the matrix resembles an hypo-eutectoid steel; with some 0.80 per cent carbon
the matrix is pure pearlite, i.e. eutectoid steel; with more than 0.80 per cent carbon
some free cementite is formed, the matrix consisting of hyper-eutectoid steel. In
other words, as the proportion of combined carbon increases, structural changes take
place in the metallic matrix of cast iron identical to those observed and described in
the case of steel. The structure of cast iron containing increasing proportion of com-
bined carbon is illustrated in Figures 9 to 11.

The above considerations justify us in considering gray cast iron as steel, that is,
as an aggregate of ferrite and cementite, the continuity of which is destroyed by the
presence of numerous graphite plates, the enormous difference in properties between

LESSON XIX CAST IRON

cast iron and steel being due solely to the presence of this graphitic carbon or rather
to the breaking of the continuity of the mass which it implies. To illustrate further,
if it were possible to remove bodily from a chunk of cast iron every particle of graphite,
its strength and ductility would not probably be greatly increased because its con-
tinuity would still be effectively destroyed by the empty spaces once occupied by
graphite. If not too high in carbon, however, it would now be weldable and could be
forged into a small piece of steel, or it could be remelted and cast into a sound steel
casting.

Mottled Cast Iron. Cast irons are sometimes produced that are partly gray and
partly white, that is made up of particles containing graphitic carbon and of particles
free from graphite their structure is well shown in Figure 12. They are called
"mottled" because of the appearance of their fracture.

Fig. 11. Gray cast iron. Hyper-eutec-
toid matrix. Magnified 500 diameters.
(Wust.)

Fig. 12. Mottled cast iron. Magnified
500 diameters. (Wiist.)

Structural Composition of Cast Iron. The structural composition of any cast
iron of known percentages of graphitic and combined carbon can readily be calculated
by following the method employed in the case of steel and assuming pearlite to con-
tain 0.834 per cent carbon, that is, exactly one part by weight of cementite and seven
parts of ferrite. It should be noted, however, that in the presence of graphite it
requires a smaller proportion of carbon to convert the whole matrix into pearlite since
there is less iron to be so converted. To make the matter clear in order that cast
iron may be free from both excess ferrite and excess cementite it must contain ferrite
and cementite in the exact proportion of one part of cementite to seven of ferrite. If
the cast iron contains G per cent of graphite and C per cent of combined carbon
forming 15C per cent of cementite the balance of the metal, 100 - G - 15C, will be
ferrite. Consequently if

100 - G - 15C = 7 x 15C = 105C
or 100 - G - 120C =

that is, if the proportion of ferrite equals seven times that of cementite the matrix
will contain pearlite only; if 100 - G - 120C is greater than the matrix will be hypo-
eutectoid; if 100 G 120C is smaller than the matrix will be hyper-eutectoid.

LESSON XIX CAST IRON 11

In the presence of 3 per cent graphite, for instance, the following relation will indicate
the needed percentage of combined carbon to make the matrix entirely pearlitic:

100 - 3 - 120C =

which gives for C very nearly 0.80 per cent.

It will be sufficiently accurate to assume in every case that gray cast iron contain-
ing less than 0.80 per cent combined carbon has an hypo-eutectoid matrix while cast
iron containing more combined carbon has an hyper-eutectoid matrix.

In calculating the structural composition of cast iron two cases then should be
considered, (1) the cast iron contains less than 0.80 per cent combined carbon; it has
an hypo-eutectoid matrix and (2) it contains more than 0.80 per cent combined car-
bon; it has an hyper-eutectoid matrix. In the first instance we have the following
relations between the percentage of graphitic carbon, G, the percentage of combined
carbon, C, the percentage of ferrite, F, and the percentage of pearlite, P,

(1) F + P + G = 100

(2) P = 8 x 15C = 120C

If C = 0.50 per cent, for instance, and G = 3 per cent, the iron would contain 60
per cent pearlite, 37 per cent ferrite, and 3 per cent graphite.

If the iron has an hyper-eutectoid matrix, that is if it contains more than 0.80
per cent of combined carbon, we can write the following equations, Cm representing
the percentage of free cementite,

(1) P + Cm + G = 100

(2) P = fF = f (100 - G - 15C)

the second equation expressing the fact that the totality of the ferrite (100 G 15C)
is included in the pearlite and that the percentage of pearlite is equal to f- times that
of ferrite. If cast iron contains 2 per cent of graphite and 1.25 per cent of combined
carbon, for instance, the foregoing equations indicate the following structural com-
position: pearlite 90.60 per cent, free cementite 7.40 per cent, and graphite 2 per cent.

The structural, graphical diagram of cast iron containing both combined and
graphitic carbon has been constructed in Figure 13 in accordance with the scheme
followed in these lessons. It is assumed in this diagram that the total carbon re-
mains constant at 3.50 per cent and that the amount of combined carbon increases
from to 3.50 per cent, in this way including the two extreme cases corresponding
respectively to absence of combined carbon and of graphitic carbon. If this diagram
be compared with that of the structural composition of steel, Lesson V, p. 13, the
steel nature of the metallic matrix of cast iron will be apparent. It will be noted that
in the present diagram when the proportion of combined carbon exceeds 1.7 per cent
there are two sources of free cementite indicated, namely, pro-eutectoid cementite and
eutectic cementite; the origin of the latter will be made clear in Lesson XXIII. Both
of these cementites constitute the free cementite present in cast iron, containing more
than 1.7 per cent of combined carbon; while formed, as later explained, at different
periods of the cooling, they appear to coagulate together and cannot be distinguished
from each other under the microscope.

Physical Properties of Cast Iron vs. its Structural Composition. The physical
properties of cast iron must necessarily depend to a very great extent upon the prop-
erties of its steel matrix from which it follows that its hardness and strength will
increase with increasing combined carbon, the hardness indefinitely, the strength up

LESSON XIX CAST IRON

to the eutectoid carbon ratio. It is evident, therefore, that cast iron of maximum
strength (1) should have a steel matrix of maximum strength, i.e. should contain
in the vicinity of 0.80 per cent combined carbon and (2) should contain as little
graphitic carbon as possible since every graphite particle is a source of weakness;
in other words, the nearer cast iron approaches a steel of maximum strength the greater
will be its strength. After having secured the desired amount of combined carbon to
give strength it is evident that a reduction of the graphitic carbon must mean a cor-
responding reduction of the total carbon in cast iron. In ordinary cupola practise for
the production of cast-iron castings, however, which consists in remelting pig iron of

Cos f /ron
W/fh Hypo-eufecfoid
matrix

Cosf /ron

with Hyper-eufecfo/d
matrix

Graphite

/oo

' Pear/ite

Free

yo-G</1t

rerrtfe.

Combined C
Grvphite C

0.50

aoo

/.oo

z.oo

200
1.50

2.JO

/.oo.

-500

oso

3.50
O

Fig. 13. Structural composition diagram of iron-carbon alloys containing a constant propor-
tion of total carbon (3.50 per cent), but varying percentages of combined carbon (from ft to 3.50
per cent.)

suitable composition, the proportion of total carbon is difficult to control, being neces-
sarily between 3 and 4 per cent and we must depend to produce strength almost alto-
gether upon the retention in the combined condition of a suitable proportion of carbon.
The total carbon may be decreased, however, by the use of iron and steel scrap as part
of the burden of the cupola resulting in increased strength, for same percentage of
combined carbon, or by remelting in a so-called "air furnace," i.e. under oxidizing con-
ditions when part of the carbon is burnt out. These low total carbon, and therefore
tenacious, cast-iron castings are sometimes offered for sale under the name of semi-
steels, a practise somewhat misleading for they are not steel in any sense of the word
since they are not malleable, have very little ductility, and generally contain a con-
siderable amount of graphitic carbon.

If soft cast-iron castings are desired so that they may be easily machined they
should contain as little combined carbon as possible. In the presence of but little

LESSON XIX CAST IRON 13

combined carbon, however, the iron will not be very tenacious, strength and softness
being antagonistic. If the castings are to be hard they should contain much combined
carbon and, therefore, little graphite. In extreme cases they will be free from graphite,
when their hardness will be very great, but they will then also be very brittle. In
the majority of cases castings are wanted soft enough to be easily machined and at
the same time of fair strength. This combination of properties is evidently to be
obtained by producing a matrix corresponding to a medium high carbon steel, i.e. by
causing the cast iron to retain some 0.30 to 0.60 per cent of combined carbon.

The percentage of combined carbon in cast iron upon which its physical properties
primarily depend, can be ascertained more quickly and readily_by microscopical ex-
amination than by chemical analysis and quite as accurately.

Chilled Cast-iron Castings. It is sometimes desired to produce cast-iron cast-
ings very hard near their outside but soft and relatively tough near their center. This
may be done by so regulating the composition and solidification as to prevent the
formation of graphite in those portions that should be hard while allowing it to form
in the portions that should be soft. The means generally employed consist in using
iron plates for those parts of the molds corresponding to the parts of the castings that
are to be hard and sand for the other parts, the quicker solidification and further cool-
ing of the metal coming in contact with the iron plates causing the retention of much
combined carbon. The resulting castings are known as "chilled" castings. Impor-
tant instances of the application of this method are to be found in the manufacture
of chilled cast-iron wheels and of chilled rolls. It will be evident that the presence
of sulphur and manganese in the cast iron should promote the retention of combined
carbon on quick cooling while the presence of silicon and of large percentages of total
carbon should hinder it. The chemical composition of cast iron to be converted into
chilled castings should therefore be carefully regulated. Microscopical examination
should prove of much value in examining the depth and quality of "chills."

Examination

I. Describe the structure of cast iron containing both graphitic and combined car-
bon and compare it to that of steel.

II. Assuming pearlite to contain 0.83 per cent carbon what will be the structural
composition of cast iron containing 3.25 per cent graphite and 0.40 per cent
combined carbon and of cast iron containing 1.90 per cent graphite and 1.60
per cent combined carbon?

III. Describe with explanation what should be the structure and composition (as
far as carbon is concerned) of cast iron with the following properties: (1) Very
soft but weak, (2) very hard but brittle, (3) very strong but lacking in softness,
and (4) combining moderate strength with moderate softness.

LESSON XX

IMPURITIES IN CAST IRON

In the preceding lesson cast iron has been considered as a pttre- alloy of iron and
carbon, but like steel, commercial cast iron always contains varying proportions of
silicon, manganese, phosphorus, and sulphur, and we should now examine the influence
of these impurities on its structure and consequently on its properties.

Silicon in Cast Iron. Cast iron seldom contains less than 0.50 per cent silicon and
frequently as much as 3 or 3.50 per cent. As in the case of steel this silicon probably
combines with some of the iron to form the silicide of iron, FeSi, which is then dissolved
in the balance of the iron. The ferrite of cast iron, therefore, always holds a consider-
able amount of silicon or rather of the silicide FeSi in solution. It has been seen that
silicon produces exactly three times its own weight of FeSi; cast iron with 2 per cent
of silicon, for instance, will contain 6 per cent of FeSi dissolved in its ferrite. The
influence of silicon on the properties of cast iron is very important chiefly through its
promoting the formation of graphitic carbon and, therefore, increasing the softness
and, if carried too far, decreasing the strength of cast iron. It is why foundrymen
often speak of silicon as a "softener." That it increases fluidity while decreasing
shrinkage and chill is also well known.

The occurrence of varying amounts of silicon in cast iron cannot be detected under
the microscope, unless indirectly and roughly through the presence of more or less
graphitic carbon. It is probably true, however, that under otherwise similar con-
ditions, the more silicon present in ferrite the slower the etching of that con-
stituent. The ferrite of cast iron with hypo-eutectoid matrix, for instance, possibly
because of its greater silicon content, often remains brilliant after the usual deep
etching treatment which would color decidedly some of the grains of the ferrite of
wrought iron or of low carbon steel.

Sulphur in Cast Iron. Cast-iron castings of good quality should not contain
more than 0.1 per cent of sulphur while but a trace of that element may be present.
In the manufacture of chilled castings, however, as much as 0.2 per cent sulphur is
sometimes allowed. It has been explained in Lesson VI that because of the great
affinity of manganese for sulphur these two elements readily combine to form the
manganese sulphide MnS, each part by weight of sulphur giving rise to the forma-
tion approximately of 2 1/2 parts of MnS. Cast iron with 0.05 per cent sulphur, for
instance, would contain about 0.125 per cent of MnS. As in the case of steel this
sulphide occurs in the form of rounded particles of a dove gray or slate color embedded
in the metallic matrix (Fig. 1). Should there not be enough manganese present to
combine with all the sulphur, the remaining sulphur would unite with iron to form the
sulphide FeS which would occur as rounded yellow areas. It should be noted, how-
ever, that since it requires less than two parts by weight of manganese to combine
with one part of sulphur, when cast iron contains twice as much manganese as it does
sulphur no iron sulphide can be formed, theoretically at least. Since it is seldom that

2 LESSON XX IMPURITIES IN CAST IRON

cast iron does not contain a considerably greater proportion of manganese the occur-
rence of FeS in cast iron should be rare.

The influence of sulphur in opposing the formation of graphitic carbon has already
been mentioned; it may consequently be said to harden cast iron. It has also a well-
known influence in increasing the depth of "chill" in solidifying cast iron against a
metal wall, that is the thickness of metal free from graphitic carbon produced by the
cooling action of that wall. Its other influences are harmful as it increases shrinkage,
causes the molten metal to be sluggish, and induces unsoundness.

Manganese in Cast Iron. Special cast irons are made, known as spiegeleisen,
ferro-manganese, etc., containing very large proportions of manganese, but in ordi-
nary castings the amount of manganese seldom exceeds 2 per cent and may be as low
as 0.10 per cent. It has been shown in Lesson VI that when manganese is present in

Fig. 1. Partial!}' malleablized cast iron. Magnified 670 diameters. Sul-
phur about 0.2 per cent, manganese 0.50 per cent. (C. C. Buck, Cor-
respondence Course student.)

these relatively small proportions it first combines with sulphur to form the sulphide
MnS, and then with carbon to form the carbide Mn 3 C, this carbide uniting with the
carbide Fe 3 C to form cementite. The cementite of cast iron, therefore, like that of
steel nearly always contains some Mn 3 C. Since iron and manganese, however, have
nearly the same atomic weights it remains true that to obtain the percentage of cemen-
tite in any commercial iron-carbon alloy it suffices to multiply its percentage of com-
bined carbon by fifteen.

The influence of manganese in opposing the formation of graphitic carbon has
already been noted; like sulphur it is a hardener, its presence in large proportions in-
creasing the difficulty of machining castings. It promotes the absorption of carbon
by iron. Some believe that it increases shrinkage and that while it has no marked
influence on the depth of the chill it increases its hardness.

Phosphorus in Cast Iron. It has been explained in Lesson VI that when phos-
phorus occurs in very small quantities as it does in steel, the totality of it probably

LESSON XX IMPURITIES IN CAST IRON 3

forms the phosphide Fe 3 P, which is then dissolved by the iron. In cast iron, how-
ever, because of the frequent presence of a considerable proportion of phosphorus
and of a larger proportion of carbon this element assumes another condition. The
occurrence of phosphorus in cast iron was first studied and described by Stead. The
results of his important investigations are briefly summarized below:

(1) When phosphorus is alloyed with carbonless iron in amount varying from
traces to 1.70 per cent, it forms a phosphide corresponding to the formula Fe 3 P,
which is held in solid solution by the iron. All the metals used commercially, such as
wrought iron and steels containing very little carbon, may be included in this class.
They consist essentially of polyhedric grains of ferrite holding Fe 3 P in solution.

(2) When the metal contains from 1.70 to 10.2 per cent phospliorus it consists of
a saturated solution of Fe 3 P in iron (1.70 per cent P) and of a eutectic alloy containing
about 10.2 per cent P and made up of about 61 per cent Fe 3 P and 39 per cent of the

Fig. 2. Alloy of iron and phosphorus. Phos-
phorus 1.8 per cent. Magnified 350 diam-
eters. (Stead.)

Fig. 3. Alloy of iron and phosphorus. Phos-
phorus 8 per cent. Magnified 250 diameters.
(Stead.)

saturated solution of Fe 3 P in iron. To account readily for this structure and that of
the following groups, it is only necessary to consider these metals as alloys of two
constituents: one the phosphide Fe 3 P, and the other a saturated solution of Fe 3 P in
iron. It is well known that a certain class of binary alloys when solidifying give rise
to the formation of a eutectic alloy, that is, of a mechanical mixture made up in
definite proportions, of extremely small plates alternately of one and the other con-
stituents, the balance of the mass consisting of that constituent which is present in
excess over the amount required to form the eutectic alloy. It is precisely what
happens in the case of iron containing over 1.70 per cent phosphorus. The formation
of eutectic alloys will be further described in Lesson XXII. For the iron-phosphide
eutectic discovered by Stead the author in 1901 suggested the name of "Steadite."

Figures 2, 3, and 4 illustrate the structure of iron containing respectively 1.8, 8,
and 10.2 per cent phosphorus. The mottled constituent made up of two structural
elements in close juxtaposition corresponds in every case to the phosphide eutectic.
The background of Figure 2 and the clear regions of Figure 3 are composed of the
solid saturated solution, while Figure 4 is composed entirely of the eutectic alloy.

4 LESSON XX IMPURITIES IN CAST IRON

(3) When the iron contains from 10.2 per cent to 15.58 per cent phosphorus it is
composed of crystals of FesP surrounded by the eutectic mixture just described, as
illustrated in Figure 5, in which the white angular areas represent Fe 3 P and the back-
ground the eutectic alloy.

(4) Describing the influence of carbon on the structure of iron-phosphorus alloys,
Stead wrote:

"On melting saturated solid solutions of phosphide of iron in iron with carbon,
the latter causes a separation of the phosphide near to the point of solidification,
which appears in the solid metal as a eutectic in irregular-shaped areas, if the carbon
present is small, and in envelopes, increasing in thickness with the amount of carbon
present, but is incapable of throwing the whole of the phosphide out of solution even
when 3.5 per cent C is present. A residuum always remains in solid solution, This
residuum is smallest, however, when the carbon is at a maximum."

Fig. 4. Alloy of iron and phosphorus. Phos-
phorus 10.2 per cent. Magnified 350 diam-
eters. (Stead.)

P'ig. 5. Alloy of iron and phosphorus. Phos-
phorus 11.07 per cent. Magnified 60 diam-
eters. (Stead.)

These conclusions are illustrated in Figures 6, 7, and 8, showing the structure of
some samples of iron with 1.7 per cent phosphorus, and containing respectively 0.18,
0.71, and 1.4 per cent C.

It will be obvious from the above description of the behavior of phosphorus that
in cast iron, because of the presence of a large amount of carbon, nearly the whole of
the phosphorus is liberated from its solution with iron and caused to occur as the
phosphide eutectic or steadite, even if the metal contains less than the necessary
amount of phosphorus needed to saturate the iron, namely 1.70 per cent. Indeed so
marked is this action of carbon that Stead tells us that in steel containing but 0.1
per cent phosphorus a portion of it is liable to be thrown out of solution in the pres-
ence of 0.90 per cent carbon.

To sum up, the phosphorus in ordinary steels occurs chiefly and probably alto-
gether as the phosphide Fe s P dissolved in iron while in cast iron it occurs chiefly if not
entirely as a eutectic the constituents of which are (1) a solid solution of iron and
1.70 per cent of phosphorus and (2) the phosphide Fe 3 P. While Stead writes that the
whole of the phosphorus is not liberated from solution even in the presence of much

LESSON XX IMPURITIES IN CAST IRON

carbon, the amount retained in solution in the presence of some 3 per cent or more
carbon is apparently very small and it may probably be assumed for all practical pur-
poses that in cast iron the whole of the phosphorus is present as a eutectic, for Stead
says that the phosphide eutectic may be detected in cast iron containing as little as

Fig. 6. Alloy of iron, phosphorus, and carbon. Fig. 7. Alloy of iron, phosphorus, and carbon.
Phosphorus 1.74 per cent, carbon 0.18 per Phosphorus 1.70 per cent, carbon 0.71 per
cent. Magnified 60 diameters. (Stead.) cent. ^Magnified 250 diameters. (Stead.)

Fig. 8. Alloy of iron, phosphorus, and carbon.
Phosphorus 1.70 per cent, carbon 1.40 per
cent. Magnified 250 diameters. (Stead.)

0.03 per cent phosphorus. The structure of phosphoretic cast iron is illustrated in
Figures 9 and 10. The various constituents, graphite, pearlite, free ferrite, and
steadite are easily distinguishable. The first sample has a hypo-eutectoid matrix,
the .second a eutectoid matrix.

Stead recommends heat tinting as a suitable treatment for bringing out the
phosphide eutectic especially in white cast iron when there is danger of confounding

LESSON XX IMPURITIES IN CAST IRON

it with cementite. The heat tinting method of Stead will be found described in an
Appendix to these lessons.

Stead explains as follows the fact that a relatively high proportion of phosphorus
in cast iron does not produce extreme brittleness.

Fig. 9. Cast iron. Magnified 100 diameters. Graph-
ite, ferrite, pearlite, and steadite. (Boylston.)

Fig. 10. Cast iron. Magnified 100 diameters.
Graphite, pearlite, and steadite. (Boylston.)

"The reason why phosphoretic pig irons are not more brittle than they are is
because the eutectic separates into isolated segregations, and does not form con-
tinuous cells round the crystalline grains. When the phosphorus does not exceed
1.7 per cent the metal is comparatively strong, but an addition of 0.3 per cent reduces
the strength materially, the reason of which is that the eutectic brittle areas in metal

LESSON XX IMPURITIES IN CAST IRON 7

with 2 per cent phosphorus approach each other closely, leaving less of the strong
ground mass intervening."

Phosphorus has no marked influence upon the condition in which carbon occurs
in cast iron but it increases the fluidity of the metal probably because of the forma-
tion of a large quantity of fusible and fluid phosphide eutectic.

Structural Composition of Phosphoretic Cast Iron. Since the presence of
10.2 per cent phosphorus causes the production of 100 per cent steadite it follows
that the phosphorus in cast iron gives rise to the formation of approximately 10
times its own weight of steadite.

In calculating the structural composition of cast iron, therefore, the amount of
phosphorus present must be considered as it may very materially lower the percent-
age of combined carbon needed to convert its matrix into pearlite. Bearing in mind
that cast iron to be free from both free ferrite and free cementite must contain ferrite
and cementite in the proportion of seven to one the following relation will indicate
the needed amount of combined carbon:

Ferrite = 100 - G - lOPh - 15C = 7 x 15C = 105C

i
cementite

or 100 - G - lOPh - 120C =

in which G, C, and Ph represent respectively the percentage of graphite, combined
carbon, and phosphorus, 15C representing, of course, the proportion of cementite and
lOPh that of steadite.

If the first term of the above equation is greater than the metal will contain free
ferrite, i.e. its matrix will be hypo-eutectoid; if it is smaller than it will contain pure
cementite, i.e. its matrix will be hyper-eutectoid.

In the presence of 3 per cent of graphite and 1 per cent phosphorus, for instance, the
percentage of combined carbon needed to t produce a eutectoid matrix will readily be
obtained in solving the equation

100 - 3 - 10 - 120C =

which calls for 0.75 per cent combined carbon. Less combined carbon would produce
free ferrite while more would cause the formation of free cementite.

In calculating the structural composition of any phosphoretic cast iron of known
percentage of phosphorus, graphite, and combined carbon, it should first be ascertained
therefore whether its matrix will be hypo- or hyper-eutectoid. Let us assume, for
instance, a cast iron containing 1.50 per cent phosphorus, 3.25 per cent graphite, and
0.40 per cent combined carbon. Since 100 - 3.25 - 15 - 120 x 0.40 is greater than
the matrix of the iron will be hypo-eutectoid, that is the metal will contain free
ferrite. The following equations will then permit the ready calculation of its struc-
tural composition.

(1) P + F + S + G = 100

(2) P = 120C

(3) S = lOPh
which give

Pearlite (P) = 48.00 per cent

Free ferrite (F) = 33.75 " "

Steadite (S) = 15.00 " "

Graphite (G) 3.25 " "

100.00

8 LESSON XX IMPURITIES IN CAST IRON

Taking another example, a cast iron containing 2 per cent graphite, 1.50 per cent
combined carbon, and 0.75 per cent of phosphorus, since 100 - 2 - 7.50 - 120 x 1.50
is less than 0, the iron will contain free cementite. Its structural composition will
be obtained by solving the equations

(1) P + Cm + S + G = 100

(2) P = f- (100 - 15C - lOPh - G)

(3) S = lOPh
which give

Pearlite (P) = 77.78 per cent

Free cementite (Cm) = 12.72 " "
Steadite (S) = 7.50 " "

Graphite (G) =__ 2 ^0 " "

100.00

Chemical vs. Structural Composition. It will now be instructive to compare the
chemical composition of cast iron both ultimate and proximate with its structural
composition. To that effect let us assume a cast iron of the following ultimate chemi-
cal composition:

Graphitic carbon 3.00 per cent

Combined carbon 0.50 " "

Silicon 2.00 " "

Phosphorus 1.50 "

Manganese 0.40 " "

Sulphur 0.02 " "

Iron (by difference) 92.58 "
100.00

In view of the foregoing considerations the following proximate compounds will be
formed:

(1) 0.02 per cent S will produce about 0.05 per cent MnS.

(2) 0.05 per cent MnS contains about 0.03 per cent Mn.

(3) This leaves 0.40 0.03 = 0.37 per cent Mn in excess to combine with carbon.

(4) 0.37 per cent Mn will form 0.39 per cent MiisC.

(5) 0.39 per cent Mn contains about 0.02 per cent carbon.

(6) This leaves 0.50 - 0.02 = 0.48 per cent carbon to combine with iron.

(7) 0.48 per cent carbon will form 7.20 per cent Fe 3 C.

(8) 1.50 per cent phosphorus will form 9.63 per cent Fe 3 P.

(9) 2.00 per cent silicon will form 6 per cent FeSi.

The proximate chemical analysis of the cast iron considered will consequently be:

Graphitic carbon 3.00 per cent

Fe 3 C 7.20 " "

Mn 3 C 0.39 " "

Fe 3 P 9.63 " "

FeSi 6.00 " "

MnS 0.05 " "

Iron (by difference) 73.73 " "
100.00

LESSON XX IMPURITIES IN CAST IRON 9

Knowing the proximate chemical constituents the structural composition can be
readily calculated. The Fe 3 C and Mn 3 C form the cementite, hence the cast iron con-
tains 7.20 + 0.39 = 7.59 per cent cementite. The whole of this cementite, since the
iron evidently has an hypo-eutectoid matrix, will combine with ferrite in the propor-
tion of 7 to 1 to form pearlite, hence the cast iron will contain 7.59 x 8 = 60.72 per
cent pearlite; 1.50 per cent phosphorus means 15 per cent of steadite; the 6.00 per
cent of FeSi will be dissolved in the ferrite while the small quantity of MnS will occur
as independent minute particles. The structural composition will therefore be:

Pearlite (P) = 60.72 per cent

Free ferrite (F) (by difference) = 21.23 " - -1L-.
Steadite (S) = 15.00 " "

Graphite (G) = 3.00 " "

MnS = 0.05 " "

100.00

The quicker method for the calculation of the structural composition of cast iron
given in foregoing pages, and in which the presence of manganese, silicon, and sulphur
is ignored would have given:

Pearlite (P) 60.00
Free ferrite (F) 22.00
Steadite (S) 15.00
Graphite (G) 3.00
100.00

For practical purposes these values are identical to those obtained from a knowledge
of the complete analysis of the iron.

Experiments

Samples of gray cast iron having respectively hypo- and hyper-eutectoid matrix as
well as a sample of white cast iron should be prepared for microscopical examination.
They should be examined both before and after etching. In the unetched condition
the graphite plates will stand out sharply on a brilliant, white background. After
etching the following features should be noted, (1) the polyhedric character of the
ferrite grains in cast iron containing a small amount of combined carbon, (2) the in-
creasing proportions of pearlite in cast iron with hypo-eutectoid matrix as the per-
centage of combined carbon increases, (3) the laminations of the pearlite areas,
(4) the occurrence of free cementite in cast iron containing more than some 0.80 per
cent combined carbon and (5) the occurrence of phosphide eutectic (steadite) in all
cast irons containing an appreciable proportion of phosphorus.

The samples should be photographed under a magnification not exceeding 100
diameters.

Examination

I. Explain the occurrence of phosphorus in cast iron.

II. A cast iron contains 2.50 per cent graphite, 0.75 per cent combined carbon, and
1.25 per cent phosphorus. Till its matrix be hypo- or hyper-eutectoid?
Explain your answer.

10 LESSON XX IMPURITIES IN CAST IRON

III. What will be (1) the proximate chemical composition and (2) the structural
composition of a cast iron having the following chemical ultimate composition:

Graphitic carbon 2.50 per cent

Combined carbon 1.00 "

Silicon 1.50 " "

Phosphorus 1.25 " "

Manganese 0.50 " "

Sulphur 0.04 " "

Iron (by difference) 93.21 " "
100.00

LESSON XXI

MALLEABLE CAST IRON

Graphitizing of Cementite. The unstability of cementite has already been alluded
to. It has been mentioned that the prolonged annealing of high carbon steel above
its critical range was always likely to result in the formation of some graphitic carbon
through the dissociation of the unstable cementite according to the reaction:

Fe 3 C = 3Fe + C

The graphitic carbon formed in this way is often called, according to Ledebur,
"temper" carbon to distinguish it from the graphite formed during the solidification
of cast iron. This breaking up of cementite into ferrite and graphite takes place the
more readily (1) the more combined carbon in the metal, (2) the higher the tempera-
ture, (3) the longer the exposure to a high temperature, (4) the more silicon and the
less manganese and sulphur present. The influence of a large amount of combined
carbon in promoting the graphitizing of cementite is made evident by the facts (1)
that iron-carbon alloys containing less than some 0.50 per cent carbon cannot be made
graphitic under the most favorable annealing conditions even when containing much
silicon, (2) that in steel containing in the vicinity of one per cent carbon the graphi-
tizing proceeds very slowly and remains partial, and (3) that in alloys containing 2.50
per cent or more of combined carbon and a sufficient amount of silicon the conversion
of cementite into iron and graphite takes place readily and can be carried to comple-
tion, combined carbon disappearing altogether. The influence of temperature and
time upon the dissociation of cementite was to be expected. Evidences will be pre-
sented in Lesson XXIII to show that the higher the temperature at which cementite
forms the more readily is it converted into iron and graphitic carbon, during solidi-
fication and subsequent cooling. The influence of silicon in graphiti.zing cementite
could likewise have been anticipated because of its well-known power to cause the
formation of graphitic carbon during and below solidification. It will be argued in
Lesson XXIII that the formation of graphite during solidification always results
from the dissociation of cementite, that is, that cementite (FesC) always forms first
but being very unstable readily breaks up into iron and graphite, its dissociation being
promoted (1) by slow cooling and (2) by the presence of silicon.

The conversion of combined carbon into graphitic or rather temper carbon finds
an important industrial application in the manufacture of so-called malleable cast-
iron castings, also termed "malleable castings," "malleable cast iron," and even
"malleable" pronounced "mallable."

The metallography of these castings should now be considered.

Malleable Cast-Iron Castings. The adjective "malleable" always used in de-
scribing these castings is misleading for it suggests a malleability, and other proper-
ties akin to those of steels, which malleable castings are far from possessing. By
malleability the metallurgist always understands that property which makes it pos-

2 LESSON XXI MALLEABLE CAST IRON

sible to convert a metallic mass into commercial shapes by rolling, hammering, etc.,
and such degree of malleability is not present in malleable cast iron. It is why in the
author's opinion this product should continue to be classified as cast iron in spite of
the arguments recently presented to classify it as steel on the ground that it is more
malleable than ordinary cast iron. The manufacture of malleable cast-iron castings
consists in subjecting to a long annealing treatment cast-iron castings of suitable com-
position whereby some of the carbon may be expelled by oxidation while most if not
the whole of the remaining is converted into graphite (temper carbon). The various
factors influencing this operation will be briefly described.

Original Castings. The castings to be subjected to the malleablizing treatment
are often called "hard" castings because they are invariably made of white cast iron
and are therefore very hard and brittle. The reason why the original castings must
be made of white cast iron will be obvious if it be considered that the malleability to
be imparted to these castings is to result chiefly, and in some cases altogether, from a
conversion of combined carbon into temper carbon as explained above. Any graphite
particle existing in the casting will be unaffected by the annealing treatment and will
be a source of weakness in the finished casting. Clearly, therefore, the hard castings
should be free from graphitic carbon. As to the amount of carbon that should be
present, theoretical considerations lead us to conclude that the less carbon the more
ductile should be the casting after thorough malleablizing, for, after all, the particles of
temper carbon while much less injurious than the hard brittle cementite from which
they are derived are nevertheless a source of weakness chiefly because of their breaking
up the continuity of the metallic mass. Again a smaller proportion of carbon neces-
sarily means a shorter annealing operation. The desirability of low carbon content in
cast-iron castings to be malleablized is universally recognized and it is partly why a
large proportion of these castings are made in the "air" furnace, as by its use the per-
centage of carbon can be lowered. On the other hand, as already mentioned, if the
proportion of combined carbon be small the graphitizing takes place with greater diffi-
culty. These considerations point to the existence of a lower as well as of an upper
limit for the carbon content of hard castings. In the majority of cases the percentage
of carbon varies between 2.50 and 3.00 per cent.

The beneficial action of silicon has been referred to; it greatly promotes the graphi-
tizing of the cementite. It might seem then as if the more silicon in the hard casting
the better at least up to some 2 or 3 per cent. Castings of white cast iron cannot be
made, however, in the presence of much silicon since this element would cause the
formation of graphitic carbon on solidification, even during quick cooling, and the
casting would not be white. And since large castings will solidify more slowly than
smaller ones it is evident that in large castings especially but a relatively small amount
of silicon can be allowed. It is probably true that as much silicon should be intro-
duced in the cast iron as will permit the making of white castings and this proportion
of silicon will necessarily decrease as the size of the casting increases. In practise it
is found to vary between 0.30 per cent in large castings (one inch thick or more) and
1.25 per cent in very small castings for like solidification conditions and subsequent
cooling. In the majority of cases the silicon content varies between 0.50 and 1 per
cent.

In Moldenke's, opinion the manganese should not exceed 0.60 per cent as a larger
amount is liable to cause trouble on annealing and difficulty in machining. Phos-
phorus according to the same authority should not exceed 0.225 per cent while sulphur
should not exceed 0.07 and preferably not 0.05 per cent. These figures represent

LESSON XXI MALLEABLE CAST IRON 3

American practise only, in Europe cast-iron castings being malleablized containing
considerably more sulphur and phosphorus.

The structure of a hard casting of suitable composition for the malleablizing process
is shown in Figure 1. It will be seen to be characteristic of the structure of white cast
iron.

Annealing Operation. The hard castings are placed in an annealing box and
firmly packed in a suitable material. The boxes are then placed in the annealing
furnaces and with their contents exposed to a desirable temperature for a suitable
length of time. These various steps should be briefly considered.

Packing Materials. The packing material affords a support for the castings
while it may or may not play an important part in the process Tfself according to its
being (1) oxidizing or (2) non-oxidizing. In his pioneer work, in 1722, Reaumur
surrounded white cast iron with crushed iron oxide, the oxygen of which removed a

Fig. 1. Hard casting (white cast iron). Magnified
67 diameters. (Boylston.)

large proportion of the carbon during the subsequent long annealing treatment at a
high temperature. Given a sufficiently high temperature and sufficiently long expo-
sure nearly the whole of the carbon could in this way be eliminated from small castings,
a small amount of it only remaining as temper carbon. It will be evident that in the
operation conducted in this way the oxidizing action of the packing is of more im-
portance in producing malleability than the graphitizing of the cementite. The
elimination of the carbon takes place through the well-known reaction between the
iron oxide (from the packing) and the carbon in the steel,

Fe 3 C + 3O = 3Fe + 3CO

the cementite in the center migrating towards the outside and being in turn acted
upon by the oxygen of the packing material.

It will be evident that this elimination must necessarily proceed from the outside
to the center, so that if one should interrupt the operation after a relatively short
time he would find an increasing proportion of carbon from shell to center. The use

4 LESSON XXI MALLEABLE CAST IRON

of an oxidizing and therefore chemically active packing material was for a long time
considered essential in the malleablizing process. Later investigations, however,
showed that white cast iron could be made malleable solely through the conversion
of combined into graphitic carbon with very little if any removal of the carbon and
that an oxidizing packing was not therefore necessary. Inert packing such as sand
and clay may be used and satisfactory malleable castings produced. It is customary,
however, in American practise, to use oxidizing packings but to depend chiefly on the
graphitizing of cementite brought about by annealing for the desired malleability and
strength. Although realizing that an oxidizing substance as a packing material is
not indispensable it is still preferred because of the apparent greater strength its use
confers to the castings. The substances most used are powdered iron oxides in the
shape of iron ore, mill scale, "bull-dog," etc., and, according to Moldenke, the crushed
flakes detached from the annealing pots themselves.

Annealing for Malleablizing. The malleability imparted to white cast iron by
annealing results fro*n, as explained in the foregoing pages, (1) the hard brittle cemen-
tite being converted into minute rounded particles of soft graphitic (temper) carbon
and (2) the burning of some carbon by the oxygen of the packing materials. If we
depend chiefly upon the burning of the carbon for the desired malleability, so-called
"white heart" castings are produced, while if the graphitizing of cementite is mainly
sought, so-called "black heart" castings are obtained. The temperature and length
of the annealing operation vary according to the kind of castings wanted.

The annealing operation is always conducted above the critical point Ac3.2.i of the
white cast iron and of course below its solidification, that is when the metal exists as
an aggregate of cementite and solid solution (saturated austenite) as explained in
Lesson XXIII. It is not known whether the free or the dissolved cementite is graph-
itized first or whether the dissociation of both kinds proceeds simultaneously. Noting,
however, that the graphitizing takes place more readily in the presence of a consider-
able quantity of combined carbon and, therefore, of free cementite, whereas in the
presence of but little free cementite (cast iron with less than 2 per cent carbon) it is at
best very slow and remains partial, it seems logical to infer that it is the free cementite
which is first decomposed, indeed that its presence is necessary to cause the graphiti-
zing of the dissolved cementite.

Annealing for "White Heart " Castings. In the making of white heart castings a
very large proportion of the carbon is removed by oxidation through the use (1) of an
oxidizing packing, (2) of a high annealing temperature, and (3) of a long annealing.
The original malleable castings made by Reaumur were of this type and his method
is still the prevailing one in Europe. The castings are annealed for four or five days
at a temperature of 800 to 900 deg. C. By removing some castings from the annealing
box from time to time and examining their structure the progress of the operation
may be readily observed. The following transformations are noted as the operation
progresses: (1) a narrow white rim of decarburized metal caused by the burning of
the carbon from the outside of the casting and a dark center caused by the graph-
itizing of some of the cementite, (2) a broader white rim and a smaller and darker
core due to the burning of more carbon and to the formation in the center of the casting
of more graphitic carbon, the graphitizing of the cementite being now indeed possibly
complete, (3) the fracture of the metal is now white and steely to the very center
showing that most of the carbon has been removed by oxidation. These conclusions
are fully confirmed by the microscopical examination of the structure of castings
subjected to this annealing treatment for various lengths of time. As this oxidation

LESSON XXI MALLEABLE CAST IRON 5

of the carbon is necessarily a very slow process, white heart castings of small size only
are made, generally not over Yi inch thick. For malleablizing larger castings the
"black heart" process soon to be described is decidedly superior.

White heart castings have a rather coarse fracture and structure because of the
high and prolonged heating to which they were exposed. Since they resemble low
carbon steels in composition and structure it would seem as if their properties could
be very materially improved by suitable heat treatment as, for instance, by annealing
followed by cooling in oil or in air according to their carbon content. Moldenke
states that the European white heart malleable castings bend excellently and are very
good, though slightly weaker than black heart castings.

Annealing for "Black Heart" Castings. For the production of black heart cast-
ings, since the oxidation of carbon is of secondary importance, the annealing tem-
perature needs not be so high nor indeed is it necessary to use an oxidizing packing.
As already mentioned, oxidizing packings are generally used, however, because of the
apparent greater strength they impart to the castings. A certain amount of carbon,
therefore, is burnt at the surface of the casting causing a narrow white rim while the
core is very dark, hence the name of black heart given to these castings. This dark-
ness of the core is due as we now understand it to the transformation of cement
carbon into temper carbon.

The annealing temperature for the production of black heart castings is generally
between 750 and 800 deg. C. in the case of air furnace iron, that is, but slightly above
the critical range of the metal and the length of time at the full converting tempera-
ture between 2]/2 and 3 days. Moldenke states that cupola iron castings should be
annealed at temperatures some 65 to 120 deg. higher and writes:

"Just why there should be a difference in the temperature required for castings
of the same composition when made in the cupola or in the air furnace, is one of the
unsolved problems. It may be chemical in that the degree of oxidation has its effect
on the opening up of the structure under the influence of heat. It may, on the other
hand, be a matter of molecular physics, and depend on the constitution and structure
of the castings as made, either in the contact of fuel with iron, or not. Possibly it
may be a combination of both the chemical and the physical. Yet the problem still
remains to be solved."

In completely malleablized black heart castings practically the whole of the car-
bon should be present in the graphitic condition, the castings consisting then of a
narrow case of decarburized iron and of a core made up of ferrite and particles of
temper carbon, as shown in Figures 2 and 3. If the malleablizing is but partial
because of too low a temperature, too short a treatment, too little silicon, or for some
other reason, considerable dissolved carbon may remain which in cooling through the
critical range will give rise to the formation of pearlite. The structure of such par-
tially malleablized cast iron is illustrated in Figures 4 to 6. It will be seen to con-
sist of graphitic carbon, pearlite, and ferrite. The mechanism of the formation of
such structures is obvious. At the end of the annealing treatment the metal con-
sisted of a mass of austenite in which were embedded a number of particles of
temper carbon; on slow cooling through the range the hypo-eutectoid austenite
rejected some free ferrite until, its composition reaching the eutectoid carbon ratio,
it was converted intb pearlite. It will be noted that the pearlitic areas which should
indicate the location of the residual austenite are situated away from the temper
carbon particles, the latter being surrounded by ferrite.

The annealing of white cast iron may be so incomplete as to retain so much dis-

6 LESSON XXI MALLEABLE CAST IRON

solved carbon that in slow cooling free cementite as well as pearlite will be formed.
This is well shown in Figure 7. In this case the austenito which existed above the

Fig. 2. Black heart casting. Magnified 100 diameters. Not etched
(F. C. Langenbcrg in the author's laboratory.)

g%^prA,<,

*5k\* k /* k-v j
.^\' V - J
" '-W- -r v .

**/* -IvM^

Fig. 3. Same metal as in Fig. 2. Magnified 100 diameters. Etched
(F. C. Langenberg in the author's laboratory.)

range at the end of the annealing operation was hyper-eutectoid, that is, it contained
more than 0.85 per cent, or thereabout, of dissolved carbon and on cooling through

LESSON XXI MALLEABLE CAST IRON

the range, therefore, liberated some free cementite. It is evident that partially
malleablized cast iron, that is, cast iron still containing considerable combined carbon,
cannot be as malleable as malleablized cast iron free from combined carbon since its
metallic matrix is necessarily less malleable.

Fig. 4. White cast iron partially malleablized. Magnified 100 diam-
eters. Particles of temper carbon surrounded by white ferrite. The
dark background is pearlite. (F. C. Langenberg in the author's
laboratory.)

Fig. 5. White cast iron partially mal-
leablized. Magnified 50 diameters.
(Wiist.)

Fig. 6. Same metal as in Fig. 5. Magni-
fied 500 diameters. (Wiist.)

Gray Cast Iron vs. Malleable Cast Iron. From the foregoing description of the
nature and manufacture of malleable cast-iron castings it is evident that gray cast
iron and malleable cast iron may have exactly the same chemical composition, although,
of course, the fc.rmer will generally contain more silicon and total carbon. In spite
of identical or nearly identical composition, however, these two metals differ enor-
mously in physical properties, gray cast iron being weak and brittle, malleable cast iron

8 LESSON XXI MALLEABLE CAST IRON

much stronger, and endowed with remarkable shock-resisting qualities. To account
for this we must look into the microstructure of both metals when it will be observed
that in gray cast iron the graphite occurs in large, generally curved, flakes and plates
whereas in malleable cast iron it is present in small rounded particles, and it may well
be conceived that the mode of occurrence of graphite in gray iron breaks up the
continuity of the metallic matrix much more effectively, therefore, weakening it and

Fig. 7. White cast iron slightly malleablized. Mag-
nified 100 diameters. (H. F. Miller in the author's
laboratory.)

destroying its ductility. Were it possible during the solidification of cast iron to
cause the graphite to occur as it does in malleable cast iron, there is no reason to doubt
but that it would be as strong and as ductile.

Experiments

Samples of fully and of partially malleablized cast iron should be prepared for
microscopical examination as well as a sample of white cast iron suitable for the nialle-
ablizing process. They should be examined both before and after etching and their
structures compared with the illustrations of this lesson. The following features
should be noted: (1) the decarburized shells surrounding, the graphitic cores of the
malleablized castings, (2) the polyhedric structure of the ferrite of fully malleablized
cast iron, (3) the ferrite areas surrounding the temper carbon particles of partially
malleablized samples having an hypo-eutectoid matrix, and (4) the occurrence of free
cementite in very incompletely malleablized castings.

Examination

I. Explain the graphitizing of cementite.
II. Describe the structure of (1) fully malleablized black heart castings and (2) of

partially malleablized black heart castings.

III. Explain the apparent reason for the malleability of annealed white cast iron
and the lack of malleability of gray cast iron containing the same proportion
of graphitic carbon.

LESSON XXII

CONSTITUTION OF METALLIC ALLOYS

For many years vague and conflicting opinions were entertained in regard to the
nature of metallic alloys. It was not known whether these intimate associations of
two or more metals were merely mechanical mixtures or chemical compounds while
the existence of solid solutions was unsuspected. The application to the study of
metallic alloys of the determination of the solubility curves which had proved so
fruitful in investigating the mechanism of the solidification of ordinary (liquid) solu-
tions and of mixtures of melted salts, soon followed by the microscopical examination
of their structure have at last revealed their true constitution. We know now that
metallic alloys may be considered as solution of high freezing (solidification) point
and, therefore, solid at ordinary temperature, whereas the old conception of solution
applied only to substances liquid at that temperature. It is evident, however, that
the location of the freezing-point of a substance in the temperature scale can have
no bearing whatever upon its constitution, that is, upon the mode of occurrence
of its constituents and the nature of the bond uniting them.

In these lessons the constitution of metallic alloys will be considered only so far as
necessary to understand the equilibrium diagram described in Lesson XXIII in which
steel and cast iron are considered as alloys of iron and carbon, i.e. as solutions of these
elements, liquid at a very high temperature, frozen at ordinary temperature.

Since moreover steel and cast iron are considered as pure iron-carbon alloys it will
suffice for our purpose to deal only with alloys of two metals, that is, with binary
alloys.

The constitution of alloys is revealed chiefly (1) through the mechanism of their
solidification as disclosed by their "fusibility" curves, and (2) through the microscopi-
cal examination of their structure after solidification.

Solidification of Pure Metals. Let us first consider the solidification of a pure
metal by observing its rate of cooling from the molten to the solid condition. This
involves the use of a pyrometer, preferably a thermo-electric (Le Chatelier) instru-
ment, the hot junction of which, suitably protected, is embedded in the cooling metal.
By recording the successive intervals of time in seconds required for each successive
cooling through ranges of temperature say of 10 deg. C., and plotting time against
temperature a cooling curve of the type shown in Figure 1 is obtained. 1 The teach-
ings of such curves are obvious. Starting with the molten metal at A, its temperature
being T, its cooling from A to B, while its temperature is falling from T to Ts is uni-
formly retarded. This results in the smooth, nearly straight portion AB of the curve.
The cooling of a molten metal above its solidification point is in this respect similar
to the cooling of any substance free from thermal critical points; the cooling curves

1 Self-recording pyrometers may also be used.

2 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

obtained in such cases are always smooth curves indicating a uniform increase in
time as the temperature of the substance falls. The curve of Figure 1 indicates the
occurrence of a sharp critical point at the temperature Ts corresponding to the hori-
zontal portion BC of the curve. It is evident that on reaching Ts the temperature
of the metal suddenly ceased to fall and remained stationary during an interval of
time represented by tt'. The metal then resumed a normal rate of cooling which was
continued to atmospheric temperature as indicated by the smooth portion CD which,
were it not for the jog BC, would be a continuation cf AB. We naturally connect this
sadden appearance of a critical point in the cooling curve with the solidification of the

Fig. 1. Typical cooling curve of pure metal.

metal. It clearly indicates (1) that the metal begins to solidify at the temperature
Ts, (2) that while it is solidifying its temperature remains constant, (3) that the
solidification lasts t' - 1 seconds. It is evident that during its solidification the metal
was exposed to the same cooling influences as those prevailing above and below it, and
since its temperature nevertheless remains constant it must be that heat is here
liberated in amount exactly sufficient to make up for the heat lost by radiation and
conductivity. Any attempt at increasing the rate of cooling during solidification would
result in hastening solidification and not in lowering the temperature of the metal.
The heat evolved during the solidification of a substance is known as its "latent heat
of solidification." In Figure 1, AB then represents the cooling of the liquid metal,
BC its solidification, and CD the cooling of the solidified metal. On heating a pure
metal from below to above its melting-point, a similar curve is obtained indicating
(1) that the melting-point coincides with the solidification point, at least under normal

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

conditions, and (2) that the melting takes place at a constant temperature, being,
therefore, accompanied by an absorption of heat.

In Figure 2 the cooling curves of several pure metals are shown. They differ only
in regard to the location of the horizontal portions indicating the temperatures of
solidification.

The structure of pure metals has been described in Lesson I, when they were shown
to be made up of polyhedral crystalline grains, each grain consisting of true crystals
of uniform orientation. As an instance of the structure of pure metals, the struc-
ture of pure lead is shown in Figure 3.

2.OOO

7~ / nn e

Fig. 2. Cooling curves of various pure metals.

Solidification of Binary Alloys the Constituents of which Form Solid Solutions.
The cooling or solidification curves of alloys of two or more metals may be constructed
exactly like the cooling curve of a pure metal, namely, by observing the rate of cooling
as the temperature is lowered from above the melting-point to atmospheric temperature
and plotting the intervals of time against the corresponding temperature falls. A
number of binary alloys are then found to yield cooling curves of the type shown in
Figure 4. From A to B, that is, as the alloy cools from T to Tb, the curve is smooth
and, therefore, indicative of normal cooling. At B there is a sudden change of direction
and from B to C, that is, from the temperature Tb to the temperature Tc, the cooling of
the alloy is evidently abnormally slow. From C to D, that is, from the temperature
Tc to atmospheric temperature, the cooling is again normal. Since the portion BC
of the curve clearly indicates spontaneous evolutions of heat causing a marked retarda-
tion in the cooling of the alloy and lasting t'~t seconds, we naturally infer that it cor-
responds to its solidification. It follows from the appearance of the cooling curve

4 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

that alloys yielding such curves do not, like pure metals, solidify at a constant tempera-
ture but that their solidification, on the contrary, lasts t'-t seconds while their tempera-
ture falls from Tb to Tc. Summing up, AB indicates the cooling of the molten alloy,
B the beginning, and C the end of its solidification, tt' the time required for its solidifi-
cation, Tb Tc the fall of temperature during solidification, and CD the cooling of the
solidified alloy. Above the point B, therefore, the alloy is entirely liquid, below C it
is entirely solid, while between B and C it is partly liquid and partly solid. The point B
is accordingly called the liquidus point and C the solidus point.

Binary alloys whose cooling curves are of the type shown in Figure 4 are known
to be solid solutions. In these alloys the component metals which are completely
merged in the liquid condition remain likewise so completely merged after solidifica-
tion that their separate existence cannot be detected by microscopical examination or

Fig. 3. Pure lead. Magnified 20 diameters. (F. C.
Langenberg in the author's laboratory.)

other physical means. They formed, on solidifying, homogeneous crystals containing
both metals in indefinite proportions. These crystals are sometimes called "mixed
crystals" and substances yielding them " isomorphous " mixtures by which it is meant
that only isomorphous substances ' can yield mixed crystals or in other words can form
solid solutions. The expression "solid solution" is much preferable and is now quite
universally used.

The mechanism of the formation of solid solutions of two metals should be ex-
amined more closely. Let us assume that a certain proportion of the metal M of
relatively low melting-point is alloyed with, or dissolved in, the metal M' of higher
melting-point. The metal M may be considered as the solute and M' as the solvent.
It is believed that when solidification begins homogeneous crystals of M and M' are
formed but that they contain a smaller proportion of the fusible metal M than the

1 Isomorphous substances are those that are capable of crystallizing in the same crystallographic
forms.

LESSON XXII CONSTITUTION OF METALLIC ALLOYS 5

liquid bath, which is thereby enriched in M .' On further cooling these crystals grow
but the crystalline matter now deposited contains more of the metal M than the crystals
first formed, although still less than the molten bath which is further enriched in M
and so on, the crystals growing through successive additions of crystalline matter con-
taining increasing proportions of the dissolved and relatively fusible metal M, and
approaching, therefore, although not reaching, the composition of the molten metal
until finally the last drop solidifies. Meanwhile, as the temperature is lowered through
and below the solidification range, diffusion takes place within the crystals so that
finally they become chemically homogeneous provided time be given (through slow

Fig. 4. Typical cooling curve of binary alloy whose component metals form a

solid solution.

cooling) for complete diffusion. Like pure metals, alloys whose component metals
form solid solutions, are composed of polyhedral crystalline grains (see Fig. 5).

Fusibility Curves of Binary Alloys whose Component Metals are Completely
Soluble in each Other when Solid. So-called "fusibility curves" or "equilibrium
diagrams" are obtained from any series of alloys by constructing the solidification
curves, as explained above, of a number of alloys of that series and plotting on a single

1 It is because the bath becomes richer in the more fusible metal M that its melting-point is
lowered, resulting in its solidification covering a considerable and falling range of temperature. If
the crystals first formed had the same composition as the bath, solidification would take place at a
constant temperature. Witness the solidification of pure metals, of eutectic alloys, and of chemical
compounds: the solidifying metal having the same composition as the liquid from which it forms,
solidification takes place at a constant temperature.

6 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

diagram the evolutions of heat observed. This has been done in Figure 6 by uniting
the BB'B" . . . and the CC'C" . . . points denoting respectively the beginning and the
end of the solidification of a number of alloys A A' A" . . . Leaving out the independent
cooling curves used for the construction of the fusibility curve, the diagram shown in
Figure 7 is obtained, the co-ordinates being now temperature and composition. The
figure is typical of the fusibility curves of binary alloys whose component metals are
entirely soluble in each other when solid. They are composed of two branches, one
concave the other convex, uniting the melting-points of the constituent metals.
MLM' is known as the liquidus because any alloy of the series above that line is
entirely liquid, while MSM' is called the solidus because any alloy below it is entirely
solid. Within the area MLM'SM the alloys are partly liquid and partly solid. The
solidification of these alloys should be further described with the help of this diagram.
Let us assume an alloy the composition of which is represented by the point P in the

Fig. 5. Copper-zinc alloy. Copper 50 per
cent. Magnified 200 diameters. Homogene-
ous solid solution. (Guillet.)

diagram and containing, therefore, 75 per cent of the low melting-point metal M and
25 per cent of the less fusible metal M'. An alloy of that composition at the tempera-
ture T is entirely liquid since its condition is represented by a point situated above the
liquidus. As the alloy cools from P to I its temperature falling from T to t, it still
remains entirely liquid. At I, temperature t, solidification begins through the forma-
tion of crystals whose composition must be represented by the point s at the inter-
section of the solidus and of a horizontal line through I, for it is evident that the only
crystals that can be in equilibrium at the temperature t with the liquid of composition
I must have the composition s. To clarify, if the crystals in equilibrium with the
liquid I at the temperature t have not the composition s, then they must necessarily
contain either more of the metal M ' or less of that metal. In other words, their com-
position may be represented by the point x to the left of s or by the point y to its
right. It is evident that the composition of the crystals forming at the temperature t
from a liquid of composition / cannot have the composition x since the corresponding
point falls within the area MLM'SM and since any point in this area cannot represent
the composition of the crystals existing at the corresponding temperature but must

LESSON XXII CONSTITUTION OF METALLIC ALLOYS
fl _ fl' ft"

Fig. 6. Diagram showing how fusibility curves are constructed.

2.5 SO

Composition.

Fig. 7. Typical fusibility curve of binary alloys whose component metals form solid solution.

represent on the contrary the average composition of the partly solidified alloy, hence
the crystals forming at t in the case we are considering cannot have a composition
represented by a point to the left of s. Assuming the composition of these crystals to

8 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

be represented by the point y to the right of s, then it is evident that these crystals
must have formed at z, that is, at a temperature considerably higher than t, which
cannot be since our alloy does not begin to solidify before the temperature t is reached.
Hence the composition of the crystals forming at t, when the alloy P begins to solidify,
cannot be represented by a point to the right of s. Clearly the point s must indicate the
composition of those crystals. As the alloy cools from I to o, that is, from t to t', the
crystals which began to form at I continue to grow by gradual deposition of crystalline
matter progressively richer in M while the remaining liquid bath likewise becomes
richer in M and consequently more fusible, its varying composition being represented
by II'. For any point o within the area MLM'SM, that is, when the alloy is in the proc-
ess of solidification, the composition of the crystals in equilibrium at the correspond-
ing temperature t' with the solution I' is represented by the point s'. This must

Fig. 8. Iron-copper alloy. Copper 10 per cent.
Magnified 125 diameters. Heterogeneous crys-
talline grains. The dark parts are richer in
copper, the lighter parts richer in iron. (Stead.)

necessarily mean that as the metal cools from t to t' and while the crystals are grow-
ing, diffusion must necessarily take place in each crystal so that the concentric layers
of varying composition of which we may conceive that they are initially composed,
assume the same composition s', the crystals being now homogenous. At s", tem-
perature t", the solidification is complete. The last drop of liquid to solidify had the
composition I". By diffusion the crystals, which began to form at s and grew from s to
s", that is, as the temperature fell from t to t", have assumed a homogenous chemical
composition. While this diffusion, however, needed to produce homogenous crystals,
readily takes place during the crystallization of liquid solutions, the case is different
with solid solutions. Unless solidification and subsequent cooling have been sufficiently
slow, solid solutions may remain heterogenous, i.e. the different layers of crystalline
matter of which each crystal is composed may not be of identical composition, the
proportion of the most fusible metal increasing from center to outside (see Fig. 8).
As an instance of binary alloys forming solid solutions the fusibility curve of gold-
platinum alloys is reproduced in Figure 9.

Binary Alloys Forming Definite Compounds and Solid Solutions. When two

LESSON XXII CONSTITUTION OF METALLIC ALLOYS 9

metals unite to form solid solutions, it does not necessarily imply that they may not
chemically combine with each other as well. Important instances are known of metals
forming a definite chemical compound, which compound is then dissolved in the re-
maining excess metal. Two metals M and M', for instance, may unite chemically to
form the compound MxM'y and unless the alloy contains the two metals exactly in
the required atomic proportions, the compound may form a solid solution with the
excess of metal M or of metal M', as the case may be. In such cases the alloys should
be considered not as alloys of two metals but of one metal and of one chemical com-
pound of two metals, when the mechanism described to explain the solidification of
solid solutions will be found applicable. Indeed we may conceive the existence of
alloys of two metals in which two definite compounds are formed, mutually soluble
and, therefore, forming solid solutions. Such alloys should be considered not as alloys

leoo
noo
jeod

JSOO
IfOO

1300

200

/too

IOOO
SCO
SCO
TOO

Pe, (174-4-)

too %fr-
fy weight

Liquid,

Solid, solution,
fAu, + Pt)

Fig. 9. Fusibility curve alloys of gold and platinum. (Desch.)

IO 2O 3O +O SO 60 7O SO 9O fOO

Atom,%Pt.

of two pure metals but of two chemical compounds, the mechanism of their solidifica-
tion being then identical to that of alloys of two metals.

Binary Alloys whose Component Metals are Insoluble in Each Other in the Solid
State. If two metals, although soluble in the liquid state, are insoluble when solid, it
is evident that on solidification they must crystallize separately into distinct crystals
readily distinguishable under the miscroscope, in other words, that the solid alloy must
be an aggregate of the two metals. The study of the mechanism of the solidification
of such alloys and of their microstructure should receive some attention.

The typical cooling or solidification curve of an alloy of two metals insoluble in
each other when solid is shown in Figure 10. The curves are obtained as previously
explained by observing the time required by the alloy to cool through successive and
equal ranges of temperature and by plotting the times against the corresponding falls
of temperature, or more conveniently and accurately by the use of self-registering
pyrometers. The curve of Figure 10 will be seen to consist of four parts, namely, AL

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

indicating a normal rate of cooling, LL' a retarded cooling, L'S a stationary tempera-
ture, and SB a normal rate of cooling to atmospheric temperature. It is logical to
infer that AL represents the uniform cooling of the liquid alloy and SB the uneventful
cooling of the solid metal, L being the liquidus point, and S the solidus. The portion
of the curve LL'S represents the solidification of the alloy as the temperature falls
from t to t'. It will be observed (1) that the solidification begins at L, temperature t,
(2) that it proceeds from L to L' as the temperature is falling, and (3) that the end of
the solidification takes place at a constant temperature, namely, t', as indicated by the

T T

Tim e

Fig. 10. Typical cooling curve of binary alloy whose component metals are
insoluble in each other in the solid state.

horizontal portion L'S, the temperature of the alloy remaining constant during T'-T
seconds.

In Figure 11 the solidification curves of a number of alloys of the same series have
been constructed as explained above, the alloys arbitrarily selected containing respec-
tively 10, 20, 40, 60, and 80 per cent of metal M. It will be noted that, with the
exception of the alloy containing 40 per cent of M , each alloy exhibits two evolutions
of heat, namely an upper evolution, L, at varying temperatures and a lower, E, always
at the same temperature regardless of the composition of the alloy. By uniting the
upper points and the lower ones as indicated by dotted lines in Figure 11, the so-called
fusibility curve or equilibrium diagram of that series of alloys is obtained. In Figure 12
the fusibility curve only is represented, the independent cooling curves used for its

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

construction having been omitted. The co-ordinates are now composition and tem-
perature. The curve indicates clearly the beginning and the end of the solidification

IO/o

Fig. 11. Diagram showing how fusibility curves are constructed.

O
/OO

30 4O
7O 6O

Composition.

/oo
o

Fig. 12. Typical fusibility curve of binary alloys whose component metals are insoluble in each

other in the solid state.

of any alloy of the series. The solidification of an alloy whose composition corresponds
to the point R, for instance, and which, therefore, contains 20 per cent of metal M and
80 per cent of metal M' evidently begins at N and ends at P. The fusibility curve is

12 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

made up of three branches, namely, the two intersecting lines LE and L'E starting
respectively from the melting-points of the two constituent metals and a horizontal
line SS' passing by the point of intersection E of the first two. This is the typical
fusibility curve of binary alloys whose component metals are insoluble in each other
in the solid condition. The solidification of these alloys should now be examined more
closely. Several features are obvious. The different alloys begin to solidify at differ-
ent temperatures according to their composition. At first, as the percentage of M
increases from to 40 (a proportio'n arbitrarily selected), the solidification point is
lowered from L to E, while with further increase of M from 40 to 100 per cent, the
solidification point is raised from E to L'. The solidification of all alloys, however,
ends at the same temperature, namely, at the temperature S. Clearly, LEU is the
liquidus and SES' the solidus. The alloy containing 40 per cent of M is evidently the
most fusible alloy of the series since it remains liquid until the temperature S is reached
while other alloys begin to solidify at higher temperatures. This alloy of lowest
melting-point is known as the "eutectic" alloy, from the Greek meaning "well melt-
ing." It is evident that, like pure metals, eutectic alloys solidify at a constant tempera-
ture, namely the eutectic temperature. Many aqueous solutions also give rise to the
formation of solutions of lowest freezing-points called " cryohydrates " and which were
at first supposed to be true chemical compounds, that is, hydrates containing salt and
water molecules in atomic proportions. And, likewise, eutectic alloys were at first
supposed to be definite chemical compounds of the two metals. They are now known
to be aggregates of these metals generally very finely divided. This will be made clear
by following the solidifications of three alloys, R, R' , and R" (Fig. 12). The alloy R
contains, according to the diagram, 20 per cent of the metal M and 80 per cent of the
metal M' . Since it contains less of the metal M than the eutectic alloy, we may for
convenience refer to it as an hypo-eutectic alloy, although of course in regard to the
content of M' it would be hyper-eutectic. In cooling from R to N the alloy remains
liquid. At N solidification begins through the formation of pure crystals of M', that
is, of the metal which is present in excess above the eutectic ratio. The formation of
pure crystals of M' continues as the alloy cools from N to P and meanwhile the portion
of the alloy remaining liquid (we may call it the mother metal) becomes gradually
richer in M , i.e. it approaches gradually the composition of the eutectic alloy. Finally
at P, temperature S, the remaining liquid has exactly the eutectic composition and
now solidifies at a constant temperature, the solidification temperature of the eutectic
alloy. In other words, as the alloy cools from N to P with formation of pure crystals
of metal M' the composition of the portion of the alloy remaining liquid varies ac-
cording to the line NKE, reaching the composition E, that is, the eutectic composition,
always at the same temperature regardless of the initial composition of the alloy.
When the alloy has cooled to 0, for instance, a point between LE and SE, it is partly
liquid and partly solid, its temperature is T and the composition of the liquid portion
is represented by K, that is, it contains 30 per cent of the metal M. On further cooling
from to P the composition of the mother metal shifts from K to E.

In the case of the alloy R", containing a larger proportion of the metal M than the
eutectic alloy and which we may, therefore, consider to be hyper-eutectic, when it
reaches the point N' its solidification begins, pure crystals of the metal M being
formed while the molten bath becomes gradually richer in M' gradually approaching,
therefore, the composition of the eutectic alloy, until at the temperature S that com-
position is reached when the remaining liquid solidifies at a constant temperature.

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

Any point 0' situated between L'E and S'E indicates an alloy in part solid and in
part liquid; its temperature is T' and the composition of the liquid is represented by
K'. On cooling from 0' to P' additional crystals of pure M are formed, or those already
formed continue to grow while the composition of the molten metal shifts from K' to E.

Starting with the alloy R' of eutectic composition, it remains liquid until at E,
temperature S, it solidifies at a constant temperature, there being no excess metal to
be rejected.

Seeing that the branch LE corresponds to the formation of pure crystals of the
metal M' and the branch L'E to the formation of pure crystals of the metal M, the con-
clusion seems irresistible that their point of intersection E must correspond to the

M'-

M-

.01

M%
M'%

o
too

a Ac
2O

Composition.

Fig. 13. Diagram depicting the mechanism of the solidification of alloys whose component metals
are insoluble in each other in the solid state.

simultaneous formation of crystals of metal M and of metal M' and that the eutectic
alloy, therefore, must be a finely divided aggregate of M and M'. In other words, at
any point on the branch LE, the alloy is saturated with the metal M' so that the lower-
ing of its temperature must cause the separation of M' crystals with corresponding
lowering of the saturation point, that is, of the solidification point of the bath. In a
similar way, at any point on the branch L'E, the alloy is saturated with M and a fall
in its temperature must result in the formation of M crystals while the solidification
point of the portion remaining liquid is thereby lowered. At E the alloy is saturated
with both metals so that any attempt at lowering its temperature must result in the
simultaneous deposition of crystals of M and of M', and since the composition of the
bath remains the same, solidification now takes place at a constant temperature.
Hence the constitution of eutectic alloys and the reason for their constant freezing
temperature.

It has been attempted in Figure 13, to depict graphically the mechanism of the

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

solidification of hypo-eutectic, eutectic, and hyper-eutectic alloys. As diagrams of this

kind have already been used in these lessons the present one will be readily understood.

When the alloy R, for instance, reaches the point N, crystals of M' form, their

.8
is

io
O

o
O

M /
M'/

IOO

75.

SO.

a a'

Metaf
M'

^ ^, ft

f ^

^ -Eutecfic ^^
M + M' = ^^

fefa/
M

2.5.

;{.

M. =c

i .

=^,=^==^-^ ~- ,

O ' ^N

IOO

<3O QO

Chemical compos/ f /on.

-40

<30

/OO
o

Fig. 14. Diagram showing the structural composition of binary alloys whose component metals
are insoluble in each other in the solid state.

%M o
% M ' /oo

Chemical composition.

Fig. 15. Diagram showing the fusibility curve and the structural composition of binary alloys
whose component metals are insoluble in each other in the solid state.

formation as the metal cools from N to P being represented by the area NPP'. At P
the residua! liquid, now of eutectic composition, solidifies at a constant temperature.
At any point the composition of the portion still liquid is represented by A'. The
completely solidified metal, will be made up of ob per cent of metal M and be per cent
of eutectic alloy.

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

It will not be necessary to describe at greater length the graphical representation
of the solidification of the alloys R' and R".

The structural composition of alloys of two metals completelyinsoluble in each other
in the solid condition may conveniently be represented graphically as shown in Figure

E
3

326

228

Sb^

632

600

500
400
300
200

x"*^

Pb
\1

90 100

20 30 40 50 60 70
Percentages of antimony by weight

Fig. 16. Fusibility curve of alloys of lead and antimony. (Roland-Gosselin.)

Fig. 17. Typical structures of alloys whose component metals
are insoluble in each other in the solid state, x, excess metal
M' and eutectic; y, eutectic; z, excess metal M and eutectic.
(Gulliver.)

14. The interpretation of this diagram is obvious. An alloy containing 20 per cent
of the metal M, for instance, will be made up of ab per cent of M' and be per cent of eu-
tectic alloy. With 40 per cent of M the alloy is wholly of eutectic composition while with
80 per cent of M , for instance, it contains a'b' of M crystals and b'c' per cent of eutectic.

16 LESSON XXII CONSTITUTION OF METALLIC ALLOYS

The composition diagram may with advantage be combined with the equilibrium
diagram as it has been done in Figure 15. Taking an alloy whose composition is
represented by R, for instance (20 per cent M and 80 per cent M'), it is seen (1) that
it begins to solidify -at N, (2) that as it cools from NtoP pure crystals of M' are formed,
(3) that the percentage of M' thus liberated is represented by PK, (4) that at P the
remaining molten alloy now of eutectic composition solidifies, and (5) that KO repre-
sents the percentage of this eutectic alloy.

The graphical method used in these lessons for representing the structural com-
position of alloys was first suggested by the author in 1896. It has since been widely
used and Tammann employs it to represent the heat liberated by the solidification, of
the eutectic or the time taken for its solidification, it being evident that both heat
and time must necessarily be proportional to the amount of eutectic alloy formed. In

T. "A -'' ~

;-?> v=

Fig. 18. Eutectic alloy of bismuth and tin.
Magnified 200 diameters. (Desch.)

Figure 15, therefore, the vertical distances of the shaded area are proportional to the
times during which the temperature of the alloys remained constant while the residual
baths of eutectic composition were solidifying. The fusibility curve of lead-antimony
alloys is reproduced in Figure 16 as an instance of alloys whose component metals are
entirely insoluble in each other after solidification.

From the foregoing it appears that solidified alloys of two metals insoluble in each
other are aggregates of these two metals and that three types of structure are to be
expected (1) the structure of hypo-eutectic alloys composed of crystals or crystalline
grains or particles of one metal embedded in, or surrounded by, some eutectic alloy,
(2) the structure of hyper-eutectic alloys consisting of crystalline particles of the
other metal and of eutectic alloy, and (3) the structure of eutectic alloys consisting
of a finely divided aggregate of minute particles of both metals. These three types
are represented diagramatically in Figure 17. Eutectic alloys are often made up of
very thin alternate and parallel plates or lamellae of each of the two constituents,
but in some cases they consist of rounded or elongated particles of one of the con-
stituents embedded in a matrix of the other constituents. The structures of some

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

eutectic alloys are shown in Figures 18 to 20. It will be noted that the constituents
of eutectic alloys are not always pure metals.

Binary Alloys whose Component Metals are Partially Soluble in each Other when
Solid. Metals are seldom absolutely insoluble in each other when solid, each metal

Fig. 19. Eutectic (eutectoid) alloy of iron and FesC. Pearlite. Magni-
fied 1000 diameters. (Law.)

Fig. 20. Eutectic alloy of SnCu 4 and Cu 3 P. Magnified 1000
diameters. (Law.)

in the majority of cases being capable of retaining in solid solution a small percentage
of the other metal. The modifications which such partial solubility introduces in the
fusibility curve should be considered. Let us suppose two metals M and M' and let
us assume that the metal M' is capable of retaining in solution immediately after

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

solidification, i.e. at the eutectic temperature, 5 per cent of the metal M and that the
metal M can retain in solid solution 10 per cent of the metal M'. The fusibility curve
of these alloys constructed in the usual way will have the appearance indicated in

M % O

M'% (00 Q7 SO 60

Compos! ti'on .

Fig. 21. Typical fusibility curve of alloys whose component metals are partially soluble in each

other in the solid state.

loo

i So/ id Solution / \

. Solid solution

M -t- 3%> fVI /

YK. A^ + G /o M

^ M\

TfJK C

o yf HI

N. "

"i" A Eu , +ec

+ '' c IfflK i

f\1 JT< '

>oM fllTiv ^

A M +60>

' M 'i lltflfk "

'A A

"TV * ***

o' yf

TK >s>

^o ' y*f

' TV

1 yl^

Phv ^0

Lri

|j|K

ullll

1 i II 1 1 1 1 11 1 liTv

M /o O 3 2O *4O 6O <3O 94/OO

M'/o 100 91 dO 6O 40 20 O

i~f~ion

Fig. 22. Diagram showing the structural composition of binary alloys whose component
are partially soluble in each other in the solid state.

met;d

Figure 21. By comparing it with the fusibility curve of metals entirely insoluble
(Fig. 12) it will be noted that the only differences between them are (1) that the
eutectic line SES' instead of extending over the whole length of the diagram now stops
at the points S and S' corresponding respectively to 5 per cent of metal M and to 10

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

per cent of metal M', and (2) the introduction of the branches SA and S'B indicates the
changes of solubility of the metals as the alloys cool from the eutectic to atmospheric
temperature. These curves show that at atmospheric temperature M' retains in
solution 3 per cent of M and that M retains in solution 6 per cent of M' . LEU is
the liquidus, LSES'L' the solidus. Clearly, any alloy containing less than 5 per cent
of the metal M or less than 10 per cent of the metal M' solidifies as a solid solution,
the diagram showing the absence of eutectic alloys. In other words, alloys contain-
ing from to 5 per cent of M may be considered as alloys of M' and of a solid solution
of M' plus 5 per cent of M, LIP being the liquidus and LS the solidus of such alloys
and their solidification taking place as explained in the case of any two metals form-
ing solid solutions. And, likewise, alloys with less than 10 per cent of M' may be re-
garded as alloys of the metal M and of a solid solution of M plus 10 per cent of M',
their liquidus being represented by L'M and their solidus by L'S'. For all alloys
containing between 5 and 90 per cent of M the diagram shows that a eutectic alloy is
formed and that the mechanism of the solidification is apparently the same as the one

'Co

1100 i

so 30 fa so 60 jo go 90 100% Cu

IOOO-

900

coo
700
600
00

lost",

962

Solid- solution. A. -f-
Solid. soUiticn B

\ B

10 20 3O 40 SO SO 7O 80 9O IOO

Atom../aCu.

Fig. 23. Fusibility curve of alloys of silver and copper.
(Desch.)

described in the case of alloys of insoluble metals. It should be noted, however, that
the two components of these alloys are no longer the pure metals but two solid solu-
tions, namely, a solution of M' containing 5 per cent of M and a solution of M con-
taining 10 per cent of M', one of these solid solutions, therefore, crystallizing when the
temperature of the alloy reaches any point on the lines LEL' and the eutectic alloy
being made up of a fine conglomerate of these two solid solutions. To clarify let us
consider an alloy represented by the point R, Figure 21. At N its solidification begins,
crystals of a solid solution of M' containing 5 per cent of M being formed. At P the
residual molten mass, having reached the eutectic ratio, crystallizes at a constant
temperature as a fine aggregate of the two solid solutions. Since the mutual solubili-
ties of the metals M and M' decrease, however, as the alloy cools below the eutectic
temperature, it is evident that each crystal must undergo a gradual transformation.
These transformations are indicated by the branches SA and S'B. After the
solidification of the eutectic, all alloys are aggregates of the two solid solutions whose
compositions are represented by the points S and S', that is, in the case under con-
sideration, arbitrarily, M' plus 5 per cent of M and M plus 10 per cent of M'. At atmos-
pheric temperature all alloys are aggregates of two solid solutions whose compositions
correspond to the points A and B, that is in the case considered M' plus 3 per cent of M

LESSON XXII CONSTITUTION OF METALLIC ALLOYS

and M plus 6 per cent of M'. At any point below the eutectic line, the corresponding
alloy is an aggregate of two solid solutions whose compositions are indicated by the
corresponding points on the branches SA and S'B. At C, for instance, in case of alloy 7?,
the structure is composed of free crystals of solid solution D and of eutectic, the com-

Fig. 24. Silver-copper alloy. Copper 15 per cent.
Magnified 600 diameters. Dark constituent is
silver containing a little copper. (Osmond.)

Fig. 25. Silver-copper alloy eutectic.
Copper 28 per cent. Magnified 600
diameters. (Osmond.)

Fig. 26. Silver-copper alloy. Copper 65 per cent.
Magnified 600 diameters. Light constituent is
copper containing a little silver. (Osmond.)

ponents of which are solid solution D and solid solution F. In other words, as the alloy
cools from the eutectic line to atmospheric temperature, the composition of its two
constituents shifts respectively from S to A and from 8' to B.

The structural composition of alloys whose component metals are partially soluble
when solid may be represented in the usual way as shown in Figure 22. Its interpre-

LESSON XXII CONSTITUTION OF METALLIC ALLOYS 21

tation does not call for further elaboration. Between and 3 per cent of M and
between and 6 per cent of M', solid solutions are formed of corresponding composi-
tions. Between 3 and 40 per cent of M , the free solid solution formed is saturated with
M, and between 40 and 94 per cent of M it is saturated with M', while the eutectic
is made up of the two saturated solutions. The fusibility curve of silver-copper alloys
is shown in Figure 23 as an example of alloys whose components remain partly soluble
in each other after solidification and typical structures of these alloys are given in
Figures 24 to 26.

Examination

Describe briefly the mechanism of the solidification of binary alloys (1) completely
soluble, (2) completely insoluble, and (3) partly soluble in each other in the solid state.

LESSON XXIII

EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

Fusibility Curve of Iron-Carbon Alloys. Steel and cast iron are essentially alloys
of iron and carbon, and their fusibility curve or equilibrium diagram may be con-
structed as in the case of any binary alloy, namely, by determining the independent
cooling curves of a number of alloys of the series and plotting the evolutions of heat

/S"oo

looo
Percen f C O

PercenfFe,C O

l.O t.7 2.O v3.0 4.O 4.3 5O 6O 6.67

IS vSO -4S GO 7-5" 9O IOO

Fig. 1. Fu-sibility curve of iron-carbon alloys.

noted against the corresponding temperatures. The resulting curve is shown in the
diagram of Figure 1.

While it is not generally possible for molten iron to dissolve more than 5 per cent
of carbon, the diagram has been constructed so as to include a percentage of carbon up
to 6.67 per cent which corresponds to 100 per cent Fe 3 C. That portion of it, however,
corresponding to more than 5 per cent of carbon, has been drawn in dotted lines.
The complete equilibrium diagram should include the evolutions of heat occurring
after solidification, known as the critical points, which have been fully described in
these lessons, but they are purposely left out of the diagram of Figure 1, it being
desired first to confine our attention to the mechanism of the solidification of iron-

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

carbon alloys. In view of the explanation of the meaning of fusibility curves given
in Lesson XXII, it will be evident that iron and carbon alloys are partially soluble in
each other when solid, that LEU is their liquidus and LSS' their solidus and that the
point E indicates the formation of a eutectic alloy. As carbon increases from to
about 1.70 per cent, the alloys solidify as solid solutions of corresponding carbon
content, LA being the liquidus and LS the solidus. These solid solutions considered as
microscopical constituents are called austenite. The solidification of alloys containing
between 1.7 and 4.3 per cent carbon begins when their temperature reaches the line
LE, crystals of a solid solution containing 1.70 per cent carbon (sometimes called
saturated austenite), being then formed which keep on growing until the line SS',
temperature 1130 deg., is reached when, clearly, a eutectic is formed composed of

Hypo -evtecf/ c a/ toys /

~/yper- eufecTic

too

evrectoid Hyper- e u te c tte id a 1 toys

Pro-eutectic / \

Pro -eut~ect~i c

-Saturated /

\ Cemenf/fe.

Austenite A

-5 15

A\\ I

c .,, ,,. A\\vfec-t
Solid solution /

"\
<c \

O ,r>

(Austen ite) ASofurated At.

istenite V

i, SO

A + Cement/

t e ) \

54f

y K

M\\

/oC O '/.O 2.0 ^.0 4.O

SO >.O G.G7

%fc 3 C /5 30 45 60 7S 90 100

Fig. 2. Structural composition of iron-carbon alloys immediately after their solidification.

that solid solution and of another constituent. The nature of the other constituent
present in the eutectic alloy formed during the solidification of iron-carbon alloys has
been in dispute. It seems at first natural to infer that elemental carbon, i.e. graphite,
is that constituent, in which case the eutectic alloy would be made up of minute crys-
talline particles alternately of saturated austenite and of graphite. Many evidences,
however, point to carbon being dissolved in molten iron as the carbide Fe 3 C (cement ite)
and to its always solidifying as Fe 3 C, although, as later explained, it may break up
into iron and graphite (FesC = 3Fe + C) immediately after its solidification. If this
hypothesis be correct, it follows that the eutectic alloy must be a mechanical mixture
of minute particles of saturated austenite and of Fe 3 C. It would seem at first as if
the microscopical examination of alloys of suitable compositions should readily reveal
the nature of the eutectic alloy. It will soon be seen, however, that both cementite
and graphite are generally found in solidified eutectiferous alloys, the microscopical
test leaving us in doubt as to which of the two constituents formed first. In the

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 3

author's opinion it seems more probable that when an iron-carbon alloy containing
more than some 1.7 per cent carbon solidifies, a eutectic of saturated austenite and of
cementite is always produced, or in other words, that in the diagram of Figure 1 the
curve L'E indicates the crystallization of cementite and not of graphite. The opposite
view will be considered later. Let us now follow the solidification of four typical
alloys, namely, R, R', R", R'", containing respectively 1 per cent, 3 per cent, 4.3 per
cent, and 4.8 per cent of carbon, the first two being, therefore, hypo-eutectic alloys,
the third the eutectic alloy, and the fourth a hyper-eutectic alloy. As the alloy R
cools, it begins to solidify at M through the formation of crystals of a solid solution
the composition of which is represented by the point T on the solidus. On cooling from
M to N these crystals grow through successive additions of crystalline matter, the
composition of which varies from T to N while the composition of the molten bath
shifts from M to U, the last drop solidifying having the composition U. As soon as the
crystalline matter is deposited, however, at least if time be given, diffusion takes place
through each crystal so that finally they are chemically homogeneous and of com-
position .V, the completely solidified metal being composed of homogeneous crystalline
grains of austenite containing one per cent carbon.

At any temperature V between the solidus and the liquidus the crystals in equilib-
rium with the molten metal must have the composition Q. It may at least be con-
ceived that if the cooling through and below the solidification period be rapid, the
crystalline grains of austenite will not be chemically homogeneous, complete diffusion
having been prevented.

In the case of an alloy whose composition is represented by the point R' in the
diagram, it begins to solidify at M' (some 1230 degrees C.), when crystals of iron
containing 1.70 per cent of carbon (saturated austenite) begin to form, the composition
of the molten metal meanwhile shifting from M' to E. At 0', temperature 1130, the
residual molten mass has reached the eutectic composition and now solidifies at a
((jnstant temperature, namely, the eutectic temperature, the completely solid metal
being made up of crystalline grains of saturated austenite surrounded bya eutectic alloy
composed of minute crystals of saturated austenite and minute crystals of cementite.

The alloy R" has the eutectic composition (4.3 per cent carbon). It remains
liquid until its temperature falls to 1130 deg. C. when it solidifies at a constant
temperature after the fashion of eutectic alloys.

The alloy R'" contains more carbon than the eutectic alloy. On reaching its
solidification point M'" crystals of cementite begin to form, increasing in size as the
metal cools from M'" to 0'" while the composition of the bath shifts from M"' to E.
At 0'" the residual molten mass having now the composition of the eutectic alloy,
freezes at a constant temperature, the completely solidified alloy being made up of
crystals of cementite embedded in a ground mass consisting of the eutectic alloy.

Structural Composition of Iron-Carbon Alloys Immediately after Solidification.
The structural composition of iron-carbon alloys immediately after their solidification,
assuming that Fe 3 C and not graphite forms, may be represented in the usual way by
the diagram of Figure 2. The diagram clearly shows (1) that alloys containing less
than 1.70 per cent of carbon are made up wholly of solid solutions, (2) that alloys
containing between 1.70 and 4.3 per cent carbon are made up of decreasing propor-
tions of saturated austenite and increasing proportions of eutectic, (3) that alloys
containing exactly 4.3 per cent of carbon are composed entirely of eutectic, and (4) that
alloys containing more than 4.3 per cent carbon contain an increasing amount of free

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

cementite, and decreasing amount of eutectic the latter disappearing altogether when
the metal contains 6.67 per cent carbon.

A modified form of the structural composition diagram may profitably be combined

ISOO

/oo

-50

O
O

/=>
3

F"
6.67

F>erceni~ Carbon.

Fig. 3. Fusibility curve and structural composition diagram of iron-carbon alloys.

with the equilibrium diagram as shown in Figure 3. Its interpretation should be
obvious. The area ABCD represents the structural composition of all alloys contain-
ing less than 1.7 per cent of carbon, and shows that they are made up of 100 per cent

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

of austenite. By dividing this area by the line AC, however, it is further shown graphi-
cally that the composition of the austenite of these alloys varies, and that they may be
considered as being made up of ABC saturated austenite diluted by ADC pure iron,
clearly indicating that with carbon the alloy is entirely made up of iron and with
1.7 per cent carbon entirely of saturated austenite. The area BCH represents the
proportion of free (pro-eutectic) saturated austenite formed during the solidification
of any alloy containing between 1.70 and 4.3 per cent carbon. The area BEFH rep-
resents the proportion of eutectic present in any alloy containing more than 1.7 per
cent of carbon. By dividing this area in two portions, moreover, by means of the line
BF we show graphically the relative proportions of saturated austenite and of cemen-
tite in the eutectic. Finally, the area EOF indicates the percentage of pro-eutectic

\^L

\ X /

\ x ^5^ ^

\ N \ > -**' ^

14-00.

\ x ^

*V j'

^ ^ <<>

\ x c-,i ,</

\ ^^ v

\ <p ^ K /

-K

^\> $''

^ C* X ^3 {/

\ <>^ ^> M/

\ "^. '^ ,'

12.00

\ ^^^^, /

"^ ^\/^

,,00

Ausfeni te-graphifs eafectic, solidifies.

looc

PercentC 1.0 20 JO <4.O 5O 6O 667

Percent fe Co is do 45 6O 75 90 IOO

Fig. 4. Iron-graphite fusibility curve of iron-carbon alloys.

ccmentite in any alloy containing more than 4.3 per cent carbon. To clarify, let us
consider the alloy of composition R (3 per cent carbon). As it cools from M to N an
amount of saturated austenite crystallizes, represented in percentage by the line OP
of the structural composition diagram. At ./V an amount of eutectic alloy is formed,
represented by the distance KO made up of K L per cent of cementite and LO per cent
of saturated austenite.

The percentage of cementite and of saturated austenite in the eutectic may be
readily calculated by solving the equations

(1) A + Cm = 100

in which A represents the percentage of austenite, and Cm the percentage of cementite
in the eutectic alloy and which express the facts that the carbon present in the eutectic

6 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

(4.3 per cent) is divided between the austenite and the cementite, the former contain-
ing 1.7 per cent carbon, the latter 6.67 per cent. The resolution of these equations
shows that the eutectic alloy contains nearly 47.7 per cent of saturated austenite and
52.3 per cent of cementite. The structural composition of any iron-carbon alloy
immediately after its solidification may likewise be readily calculated. If it contains
less than 1.70 per cent carbon it is entirely made up of austenite. If more highly
carburized, two cases are to be considered: (1) the alloy contains between 1.7 and 4.3
per cent carbon when it is made up of saturated austenite (A) and of eutectic (E) and
(2) the alloy contains between 4.3 per cent and 6.67 per cent carbon when it is com-
posed of cementite (Cm) and of saturated austenite (A).

/SO a.

I4OO .

/300.

-k

I2OO .

I-

1100

IOOO

Percent C O
Percent Fe^C O

Fig. 5. Combined graphite-cementite fusibility curves of iron-carbon alloys.

In the first instance the two following equations

(1) A + E = 100

will give the values of A and E for any known carbon content (C) while in the latter
case the equations

(1) Cm + E = 100

will likewise give the values of Cm and E.

An alloy, for instance, containing 3 per cent of carbon will be found to contain 50
per cent of eutectic and 50 per cent of saturated austenite, while an alloy with 5 per
cent carbon is composed of 70.5 per cent of eutectic and 29.5 per cent free cementite.

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 7

Iron-Graphite Fusibility Curve. It has already been mentioned that some
writers claim that graphite instead of cementite may, and if time be given does, form
on solidification; in other words, that the eutectic alloy may be composed of saturated
austenite and graphite, and that free graphite may separate during the solidification
of alloys containing more than 4.3 per cent carbon. The diagram interpreting this
assumption which may be called the iron-graphite fusibility curve is shown in Figure 4.
It will be seen to be similar to the iron-cementite diagram (Fig. 1).

Combined Graphite-Cementite Diagram. Recognizing the possibility of the
formation of a graphite-austenite eutectic and of a cementite-austenite eutectic ac-
cording to the rate of cooling, some writers, notably Charpy and Grenet, Heyn, and
Benedicks, recommend the use of double solidification curves ifTtftc equilibrium dia-
gram of iron-carbon alloys. These double curves are shown in Figure 5, the dotted
lines referring to the austenite-graphite system. It will be noted that free graphite
and the graphite-austenite eutectic form respectively at temperatures slightly higher
than those at which free cementite and the cementite-austenite eutectic form, the
solidification of the latter constituents being regarded as due to surfusion or under-
cooling. It is accordingly believed that only the iron-graphite system is stable, the
iron-cementite system being "metastable." Our reasons for believing that graphite
and not cementite is the final condition to be assumed by carbon are based on repeated
and concordant observations that any condition promoting stable equilibrium results
in the transportation of cementite into graphite as, for instance, very slow cooling
during and below solidification or long exposure of cementite (as in the manufacture
of malleable cast-iron castings) to a high temperature, while on the contrary, treat-
ments opposing equilibrium, such as quick cooling, always result in the formation or
retention of cementite. Roozeboom, when he first took up the study of the iron-
carbon diagram, believed that cementite was the final stable condition of carbon, any
graphite having formed during solidification combining with iron at some 1000 degrees
C. to form cementite. The error of this view soon became apparent, however, to
Roozeboom himself.

Graphitizing of Cementite. Although recognizing the fact that graphite and not
cementite must be the final condition assumed by carbon, the author believes with
some other observers that graphite never forms directly as iron-carbon alloys solidify,
its occurrence always resulting from the breaking up of cementite according to the
reaction

Fe 3 C = 3Fe + C

from which it would follow that the iron-graphite fusibility curve need not be in-
cluded in the equilibrium diagram. Even those who believe in the possibility of the
direct formation of graphite do not deny that cementite is the constituent which
generally forms first on solidification; they state that the separation of graphite from
molten iron is possible only in the case of very slow cooling. As a matter of fact, how-
ever, they offer no conclusive evidences that such separation ever takes place. From
the formation of "kish," that is, of graphite floating on the surface of a ladleful of
molten cast iron, it does not follow that such graphite formation was not preceded by
the formation of cementite. If, on very slow cooling, graphite separated directly from
molten iron, the bulk of it at least should rise to the top of the molten bath and the
solidified mass should be found much richer in graphite near its surface than at some
distance from it. The author does not understand such heterogeneity in the dis-

8 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

tribution of graphitic carbon to be observed in the case of gray cast-iron castings. On
the contrary, on the assumption that graphite results from the breaking up of cemen-
tite soon after its solidification, it is readily understood why, in spite of their very
great difference in specific gravity, iron and graphite are found uniformly distributed
in the various parts of castings. The microscopical examination of the structure
of very slowly cooled castings does not reveal the existence of a graphite-austenite
eutectic.

That cementite is unstable, being readily converted into iron and graphitic carbon,
is also generally admitted. It is upon this instability of cementite that the important
industrial operation of converting white cast-iron castings into graphitic malleable
castings is based. And there is abundant evidence that the higher the temperature,
the more readily is cementite dissociated, from which it follows that the higher the
temperature at which cementite forms the more readily will it be converted into
graphitic carbon. Bearing this in mind, and with the assistance of the diagram of
Figure 3, let us look more closely into the graphitizing of cementite. The diagram
shows clearly that, during the solidification of alloys containing more than 1.70 per
cent of carbon, (1) some cementite forms as pro-eutectic cementite if the metal con-
tains more than 4.3 per cent carbon, (2) some cementite forms as eutectic cementite in
all alloys, (3) some cementite remains dissolved in the eutectic-austenite of all alloys,
and (4) some cementite remains dissolved in the free austenite of alloys containing less
than 4.3 per cent carbon. Considering first the free cementite, that is, the pro-eutectic
and the eutectic cementite, it is evident that the former is formed at a higher tempera-
ture, and that the more carbon in the alloy the higher the temperature at which it
begins to form. It seems safe to infer, therefore, that pro-eutectic cementite will break
up into graphite and ferrite more readily than eutectic cementite, this being consistent
with the well-known fact that hyper-eutectic alloys are generally rich in graphite even
after relatively quick cooling. The presence of pro-eutectic cementite may also pro-
mote the formation of graphitic carbon because once this graphitizing process is
started, it is likely to extend, if time be given, to the bulk of the cementite, first the eu-
tectic cementite and later the austenite-cementite undergoing the change. Alloys con-
taining less than 4.3 per cent carbon and consequently free from pro-eutectic cementite
should not become graphitic as readily because of the lower temperature at which
eutectic-cementite forms. If a large proportion of cementite be formed, however,
that is, if the alloys contain more than 3 or 3.5 per cent of carbon, a certain amount
of graphitizing is readily induced through slow cooling. With decreasing carbon
the breaking up of cementite becomes progressively more difficult until, in alloys con-
taining less than 1.7 per cent carbon (the steel series), and, therefore free from eutectic
cementite, graphitic carbon is very seldom formed.

It should be borne in mind that while those alloys which contain but a small pro-

)rtioi. oi^carbon cannot be made graphitic, when a large proportion of carbon is

'he graphitizing once started may be made to include the totality of the

'iiib^iue, thus explaining the freedom of steel from graphite and the freedom of

some cast irons from cementite.

The foregoing remarks apply to pure iron-carbon alloys, the influence of the
elements generally present in commercial products having been purposely ignored.
When dealing with commercial steel and cast iron, the well-known influence of silicon
in promoting the formation of graphitic carbon should be remembered as well as the
opposite influence of sulphur and manganese. Because of the presence of a notable

f , .

*1S|' > X

.-. w^^ffifciv''. ^ -

Fig. 7. Magnified 750 diameters.

Fig. 6. Magnified 50 diameters.

Fig. 9. Magnified 750 diameters.

Fig. 8. Magnified 50 diameters

Fig. 10. Magnified 50 diameters. Fig. 11. Magnified 750 diameters.

Figs, ft and 7. Iron-carbon alloy. Hypo-cutcctic. Structure immediately after solidification. Dark crystallites of
saturated austenitc in a matrix of austenite-cementitc eutectic. Figs. 8 and 9. Iron-carbon alloy. Austenite-
cementite outrctic. I r igs. 10 and 11. Iron-carbon alloy. Hyper-eutectic. Structure immediately after solidifi-
cation. Needles of cementite in a matrix of austenite-cemcntite eutectic. (Gcerens.)

10 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

proportion of silicon, commercial cast irons after slow cooling are necessarily more
graphitic than pure alloys of same carbon content.

Structure of Iron-Carbon Alloys Immediately after Solidification. If the alloy
contains less than some 1.70 per cent carbon it is made up after complete solidification
of crystalline grains of austenite. It has been explained, however, that in the absence
of manganese or other "retarding" elements it is not possible to prevent, even through
very rapid cooling, the transformation of some of the austenite at least into martensite.
The polyhedric structure of austenite has been illustrated in these lessons in the case
of special steels (manganese and nickel steels) and it is now well understood that the

ISOO

6.67

Fig. 12. Equilibrium diagram of iron-carbon alloys.

frequent network structures of slowly cooled steel are due to the existence of poly-
hedric austenitic structures above their critical range.

If the alloy contains from 1.70 to 4.3 per cent carbon it is made up, after solidifi-
cation, of crystals of saturated austenite and of eutectic alloy. This is well shown
after Gcerens in Figure 6, in which the dark "pine tree" crystals consist of saturated
austenite, while the ground mass is the cementite-austenite eutectic. In Figure 7 the
same structure is shown more highly magnified. If the alloy contains exactly 4.3 per
cent carbon, it consists wholly of eutectic as shown under different magnifications in
Figures 8 and 9. It has been seen that, theoretically, this eutectic contains 47.7 per
cent of saturated austenite (the dark constituent), and 52.3 per cent of cementite.
Alloys containing more than 4.3 per cent carbon consist after solidification of free
cementite in the form of needles embedded in a eutectic matrix as shown in Figures 10
and 11.

It should be noted that the dark constituent occurring in these structures and
described as saturated austenite may not be absolutely unaltered austenite because of

LES30X XXIII EQUILIBRIUM DIAGRAM OF IROX-CARBOX ALLOYS 11

S *

12 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

the difficulty of completely preventing the transformation of that constituent even
in the presence of a large amount of carbon and by very sudden cooling. If the aus-
tenite has undergone some transformation, however, so that it contains some mar-
tensite and even troostite those transformations must have taken place in situ and the
structures reproduced in Figures 6 to 11 must represent accurately the structural
aspect of the corresponding alloys after solidification.

Complete Equilibrium Diagram. In the foregoing pages only the solidification
curves of iron-carbon alloys have been considered and the probable mechanism of
their freezing explained. Their equilibrium diagram, however, must include all heat
evolutions observed on cooling from the liquid condition to atmospheric temperature ;
in other words, the thermal critical points fully described in previous lessons are part
of the complete equilibrium diagram as indicated in Figure 12.

Since the meaning of every curve of this diagram has been discussed, it only re-
mains to inquire into any possible structural or other changes taking place after-
solidification and before the alloys reach their respective thermal critical point or
points, that is, while they cool from the solidus LSS' to the eutectoid line CDF . The
changes which do or may take place as the alloys cool in this range are clearly stated
in Figure 13. In this diagram the most likely significance of every curve is indicated
as well as the nature of all structural transformations, and of all possible resulting
structures after complete cooling. The author believes that it embodies those infer-
ences best supported by analogy and by experimental evidences. Although neces-
sarily involving some repetition, a methodical examination of the various parts of
this complete diagram seems advisable, as it will permit a recapitulation of the various
matters previously discussed.

Let us consider (1) the solidification of iron-carbon alloys, (2) their cooling from
the solidus LSES' to the eutectoid temperature CDF, and (3) their cooling through
the eutectoid temperature and their final structures.

According to the mechanism of their solidification iron-carbon alloys are divided
into three classes, namely, (1) alloys containing less than 1.70 per cent of carbon, (2)
alloys containing between 1.70 and 4.3 per cent carbon, and (3) alloys containing
more than 4.3 per cent of carbon. Alloys containing less than 1.70 per cent of carbon
and including, therefore, all the steels of commerce solidify as solid solutions of the
carbide Fe 3 C (cementite) in gamma iron, these solutions being known as austenite.
LA is the liquidus and LS the solidus of these alloys. Alloys containing between 1.70
and 4.3 per cent of carbon solidify through the formation of crystals of saturated
austenite at gradually decreasing temperatures and through the final solidification,
at the eutectic temperature, of the residual molten metal necessarily of eutectic com-
position. Alloys containing more than 4.3 per cent of carbon solidify through the
formation of cementite crystals at gradually decreasing temperatures, and through to
final solidification, at the eutectic temperature, of the residual molten metal necessarily
of eutectic composition.

In cooling below their solidus, LS, alloys with less than 1.70 per cent carbon undergo
no change until they reach their thermal critical points Ar 3 , Ar 3 . 2 , Ar 3 . 2 .i or Ar cm as
the case may be, when, if they contain less than some 0.85 per cent of carbon (hypo-
eutectoid steels), some iron is set free and converted into beta iron, while if they
contain more carbon (hyper-eutectoid steels), cementite is liberated. In either case
when the eutectoid temperature is reached the residual austenite, now of eutectoid
composition (0.85 per cent carbon), is converted into pearlite.

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 13

iSOO. 2 ^ . . 6

Pro-eufecf/c
Cemen t/fe

Iron
d Hut in
saturate
Austenrfe

Struct u rot
c.orrt*po<s 1 f/ort
ec//a/e/y
offer
3o lie/if i cation

Co m po*s i fton
i mm edict te/y
above
Ar,

Structural
compos it ion

6 6.57
Carhon /o

Fig. 14. Equilibrium and structural composition diagram of iron-carbon alloys.

14 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

After solidification, alloys containing between 1.70 and 4.3 per cent carbon are
aggregates of saturated austenite (austcnite containing 1.70 per cent C or 25.5 per
cent cementite) and of cementite-austenite eutectic. On cooling below their solidus,
cementite (pro-eutectoid cementite) is liberated both from the free and from the
eutectic-austenite, while if the cooling be sufficiently slow both the eutectic and pro-
eutectoid cementite may be partly or wholly dissociated into graphite and iron (f errite) .
Indeed, the graphitizing may even include the eutectoid cementite, in which case the
alloy is made up solely of ferrite and graphite.

Alloys containing more than 4.3 per cent of carbon are, immediately after solidi-
fication, aggregates of cementite (pro-eutectic cementite) and of cementite-austenite
eutectic. On slow cooling below their solidus, cementite (pro-eutectoid cementite) is
liberated from the eutectic austenite while the pro-eutectic, eutectic, and pro-eutectoid

Fig. 15. The author's early equilibrium diagram (1896).

cementite may be partly or wholly dissociated into graphite and ferrite, in some ex-
treme cases the eutectoid cementite even being graphitized.

On cooling through the eutectoid temperature, any remaining austenite is bodily
converted into pearlite.

The above consideration clearly shows that in alloys containing more than some
1.70 per cent carbon four types of structure may be formed according to the rate of
cooling below the solidus:

(I) Cementite plus pearlite, the structure of white cast iron, readily produced by
quick cooling and representing a metastable equilibrium. (II) Ferrite and graphite,
an extreme case, possible only after very slow cooling and in the presence of much
silicon, little manganese, and sulphur, and representing stable equilibrium. (Ill) Cem-
entite plus pearlite and graphite, the structure of gray cast iron with hyper-eutectoid
matrix, resulting from slow cooling, promoted by the presence of silicon and opposed
by sulphur and manganese. (IV) Pearlite, ferrite, and graphite, the structure of gray
cast iron with hypo-eutectoid matrix, produced because of slower cooling or because
of the presence of more silicon or of less sulphur and manganese.

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 15

It will be explained in the next lesson that, according to the phase rule, only two
components may be present in a binary alloy in a state of equilibrium from which
it follows that gray cast irons, since they contain besides iron both cementite and
graphite, are out of equilibrium. One of the components must disappear if the alloy
is to assume equilibrium. Cementite is undoubtedly that component as proven by
the malleablizing of cast iron when cementite is readily dissociated into graphite and
ferrite on prolonged heating to a high temperature.

In Figure 14 the complete equilibrium diagram is shown combined with three
constitutional diagrams showing graphically the structural composition of iron-car-
bon alloys (1) immediately after their solidification, (2) immediately before the eutec-
toid temperature, and (3) below the eutectoid temperature, ^nlhe assumption that
no graphitic carbon is formed. The structural changes taking place while the alloys

1.600-
1,500-
1.400-
| 1.300-
C 1.200'

1,100-

^ IjOOO"

^ft

a. 300-
soar

700'

eoo-

A^-~.

2,732'

2^72' 'o

2,012' ^

1,832' J

-*-

1,292'
MB*

'"*

"^^^

i^id,

solutior

t, of

CcLrbon/

irt' /r

~~~

-*-,

---^

*** ^.^

"* 2^

r*"**

Jj t-O UZ,

< eu^4

*tio

^- E

X
* Sol

<t solatia

^ /

3-V

of Carbon

'in. Iron

--.-*-
P

-"" Y~

"*""*"*"

Solid,

eutectic

per

Fig. 16. Roberts-Austen's first equilibrium diagram (1897).

cool below their solidus down to the eutectoid temperature are, in this way, clearly
depicted. The following facts, for instance, are graphically shown, (1) the pro-eutectic
cementite formed during the solidification of hyper-eutectic alloys and the eutectic
cementite present in all alloys containing more than 1.70 per cent carbon remain un-
changed as the alloy cools to atmospheric temperature, (2) the free saturated austenite
of hypo-eutectic alloys, as well as the eutectic-austenite, are converted into eutectoid
austenite through the liberation of cementite (pro-eutectoid cementite), the area
EDH representing the cementite thus set free, and (3) in hypo-eutectoid alloys iron
is set free as shown by the area FJG. The lower diagram shows that, in cooling
through the eutectoid temperature, the remaining austenite, now necessarily of"
eutectoid composition, and sometimes called hardenite is converted into pearlite.
Taking, for instance, the metal whose composition and temperature are represented
by the point R, its transformations and final structure are clearly shown. At M it
begins to solidify through the formation of crystals of saturated austenite; from M to
N the austenite crystals continue to grow, the percentage of free austenite present in
the solidified metal being represented by the distance OP; at N the residual bath

16 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

solidifies as a eutectic alloy, the percentage of which is proportional to the distance
KO; this eutectic contains KL = PQ = TU per cent of cemeiitite and LO per cent
of saturated austenite; LP is the total amount of austenite in the alloy; after solidifi-
cation in cooling from N to K, pro-eutectoid cementite is liberated both from the
free and from the eutectic-austenite, QS representing the percentage of cementite
finally expelled; on reaching the point K, the remaining austenite, ST, is of eutectoid
composition, when it is sometimes called hardenite, and in cooling through K this
austenite is converted into pearlite, the metal being finally made up of TU per cent
of eutectic cementite, UV per cent of pro-eutectoid cementite, and VX per cent of
pearlite, the latter containing VW per cent of cementite, and WX per cent of ferrite,
or of TV per cent of free cementite and VX per cent of pearlite, or again of TW
per cent of total cementite and WX per cent of ferrite.

Historical. In view of the scientific and practical importance of the equilibrium

VSOO

2-S If 3-0 3-2 3-f 3-S

CARBON PER CCNT

5-0 SI 54 *6

Fig. 17. Roberts-Austen's second equilibrium diagram (1899).

diagram of iron-carbon alloys, a brief historical sketch of its evolution should be of
interest to the reader. The first diagram was published by the author in 1896. ' It is
reproduced in Figure 15. It will be noted that, although the diagram includes only
the thermal critical points, it is otherwise substantially accurate. In describing it
the author wrote in part :

"Figure 1 shows graphically the position of the critical points in cooling steels of
various carbons. The width of the black lines does not refer to the intensity of the
retardations, but only indicates the range of temperature which they cover. For
instance, it shows that the single retardation of high carbon steel begins at about
680 deg. C. and ends at about 640 deg. C. The maximum evolution of heat lies some-
where between these limits, but not necessarily in the middle.

" This graphical representation was obtained by plotting the results of the investi-
gations of Osmond, Howe, Roberts- Austen, Arnold, and the writer; and, with one or
two exceptions all their figures fall very nearly within the limits here indicated."

1 "The Microstructure of Steel and the Current Theories of Hardening," ALBERT SAUVEUK,
Transactions American Institute of Mining Engineers. 1896, p. 867.

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 17

It is from this modest beginning that the present diagram was evolved.

In 1897 Roberts-Austen published in his fourth report of the Alloys Research Com-
mittee of the Institution of Mechanical Engineers the diagram reproduced in Figure 16.

Two years later, in 1899, the diagram shown in Figure 17 was published by Roberts-
Austen and Stansfield in the fifth report of the Alloys Research Committee. Some of
the conspicuous features of this diagram should be noted. The solidification point of
pure iron was indicated to be 1600 degrees C. whereas we know now that it is nearly
1500 deg. No attempt had been made yet at ascertaining the end of the solidification,
that is, the solidus, of alloys forming solid solutions; the formation of a eutectic on
solidification was indicated as taking place in alloys containing more than one per cent
carbon; graphite was supposed to crystallize during the solidification of alloys con-

1600"
IBOO

1 S.

3 4 S

'j~t>on

Per

1400
I3OO

>aao
1100
1000

90O"
SOO"
7OO
&XP

r~r '

^^'

?^Ba

tchA

WZ&.

*>00f)

Lrsj

riff

Lit)

u2c^

Liqu.
rafif,

tie

UUH.

^B/

Warte

tsite

7%z^ft

*nsitt

"** ^

&''

Graft,

fuke

Ma)

ite+t

)

prt

t f

Krr

lie*

icLite.

Perl*

Fig. 18. Roozeboom's equilibrium diagram (1900).

taining more than 4.3 per cent carbon, there being in the diagram no indications of
possible formation of cementite; the eutectic alloy was assumed to be a graphite-iron
eutectic; critical points occurring below the eutectoid temperature were represented
in the diagram and marked "hydrogen points" (See Lesson VII, page 8, "Minor
critical points"); the Ar cm curve was arbitrarily extended to yield a V-shaped curve.
Roberts-Austen mentioned the formation of a solid solution, free in hypo-eutectic
steels, and as a constituent of the eutectic in alloys of eutectic composition, and he
ascribed the presence of free cementite in cast iron to the liberation of that constituent
from solid solution.

In 1900 Roozeboom took up the study of Roberts-Austen's diagram, and applying
to it the teachings of the phase rule published the diagram of Figure 18 as a probably
accurate representation of the solidification mechanism of iron-carbon alloys and of
the structural transformations taking place after solidification. In this diagram, the
line ba, that is, the solidus of alloys forming solid solutions, is for the first time indicated;

18 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS

1500

1400

1300

1200

ft
o

fliopo

a
t

K .

a 90

9 HOC

600

CARBON PER CENT

Fig. 19. Carpenter's and Keeling's equilibrium diagram (1904).

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 19

the word martensite is used instead of austenite to denote the solid solution of iron
and carbon; free graphite is assumed to form during the solidification of alloys con-
taining more than 4.3 per cent carbon and the constituents of the eutectic alloy are
supposed to be martensite (solid solution) and graphite. The Ar cm curve of hyper-
eutectoid steels is arbitrarily extended as an horizontal line starting from E (1.75 per
cent carbon and 1000 deg. C.) and extending to the end of the diagram. Roozeboom
argued that, while graphite formed on solidification in all alloys containing more than
2 per cent carbon, this graphite at 1000 deg. (line EF) recombined with iron to yield
cementite so that, finally, alloys in equilibrium would contain only ferrite and cementite,
thus conforming to the phase rule which forbids the presence of more than two com-
ponents in binary alloys. It has since been conclusively shown that Roozeboom was
in error, that while the ferrite-cementite system is in equilibrium according to the

TEMPER-
ATURE

1500

1400

1300

MOO

1100

1000

900

BOO

700

600

500

- ,

\ LI

3UIC

""^

tr\i

^D'

QUH
ENTI
\\ 1

& M
\ <

XED

:RSI

ALS

f> *'

'/*

CEM

V*CT

IXE

?//

CR\

'STA

i
i

t
f

* >

IXEI

) CR

YSTfi

JTE'

:TIC

/
/

& E

JTE

STIC

=MEr<

TITE

CRY:

STAL

PER

?ITE

EAR

LITE

& C

EME

NTI

PEAR

LITE

if C.

~/vr ,

CAR6

/ e J + s 6 CEMEK-

Ca TITE

Fe Fe,C

Fig. 20. Benedicks' double equilibrium diagram.

phase rule, it is in metastable equilibrium, the ferrite-graphite system being the only
stable one. The hypothetical horizontal line EF is now consequently omitted from
the equilibrium diagram, and the Ar cm curve made to join the eutectic line at its
origin (a).

In 1904 Carpenter and Keeling made a series of very careful experiments in order
to ascertain the evolutions of heat taking place in cooling pure iron-carbon alloys from
the liquid state to atmospheric temperature. By plotting their results, the equilibrium
diagram reproduced in Figure 19 was obtained. The solidification of pure iron is
shown to take place at 1500 deg. C. The curves are otherwise identical to those of
Roozeboom, the horizontal line EF having been introduced. The faint evolutions of
heat occurring in the vicinity of 600 deg. C. already discovered by Roberts-Austen
and ascribed by him to the presence of hydrogen, were also observed by Carpenter
and Keeling, as well as some faint evolutions in the vicinity of 775 deg., the meaning
of which remains uncertain.

20 LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBOX ALLOYS

When it became apparent that graphite and not cementite must be the final stable
form of carbon, several authorities argued that two equilibrium conditions could exist
according to the rate of cooling during solidification, one of them stable, the other
metastable, and that this should be indicated in the diagram. This view was presented
notably by Charpy and Grenet, by Benedicks and by Heyn. The double diagram
advocated by them is represented in Figure 20. The solidification of free cementite
and of the cementite-austenite eutectic being assumed to be due to the well-known
phenomenon of surfusion or undercooling, the corresponding curves are arbitrarily
outlined at temperatures slightly lower than those pertaining to the formation of free

500'

Cent,.

Fig. 21.

Carbon, per
Rosenhain's equilibrium diagram (1911).

graphite and of graphite-austenite eutectic. The author has already shown why, in
his opinion, the graphite curves should be left out. The view that cementite always
forms during the solidification of iron-carbon alloys but that being unstable it is
readily dissociated into ferrite and graphite, seems to be better supported by experi-
mental evidences and more consistent with practical facts.

Rosenhain has recently plotted the experimental results of Carpenter and Keel-
ing, of Gutowsky and of himself, obtaining the diagram reproduced in Figure 21. He
considers Gutowsky's results in regard to the form of the solidus curve of alloys form-
ing solid solutions as more accurate than those previously published, and he incorpo-
rates them in the diagram as shown in Figure 21, the solidus line being rounded instead
of straight as heretofore represented. In justification of his course, Rosenhain writes:

"We have now to consider the curved portion of the 'solidus,' the line AD. This

LESSON XXIII EQUILIBRIUM DIAGRAM OF IRON-CARBON ALLOYS 21

represents the temperatures at which the alloys have just completed their freezing
process, that is, have just become completely solid, or, conversely, it represents the
temperature of incipient fusion on heating. In the earlier investigations, and even in
those of Messrs. Carpenter and Keeling, these temperatures were obtained by esti-
mating the point on each of the cooling-curves where the heat-evolution due to solidi-
fication came to an end. Unfortunately, the end of a heat-evolution is never sharply
indicated on the curves, so that this estimation was admittedly vague. Quite recently
that determination has been repeated, and with considerable greater accuracy, be-
cause a very much more satisfactory method was available . . .

"The method of determining the 'solidus' was to take small pieces of steel, of
known composition, heat them, and suddenly cool them from successively higher
temperatures; afterward each specimen was examined by means of the microscope.

ig. 22. Cooling curves of carbon steels replotted from the data of Carpenter and Keeling. (Rosenhain.)

It is easy, as the photographs show, to determine what is the particular point at which
you have reached a temperature where there was a small quantity of liquid metal
present at the moment of quenching."

The data obtained by Carpenter and Keeling have been given in the form of a table
in Lesson VII, page 9, and some of these curves reproduced in Figure 8, page 17 of the
same lesson. Rosenhain has recently replotted some of the figures of Carpenter and
Keeling by his derived differential method (Lesson VII, page 16) and obtained the
sharp curves shown in Figure 22.

Examination

Describe briefly (1) the solidification, (2) the transformations after solidification
and according to rate of cooling, and (3) the final structures of iron-carbon alloys con-
taining respectively 1.25, 3.50, and 5.00 per cent carbon.

Calculate the structural composition of these alloys, assuming that graphitic
carbon does not form, (1) immediately after solidification, and (2) below the eutectoid
temperature.

LESSON XXIV

THE PHASE RULE

The Phase Rule to which references have been made in the preceding lessons
should now be considered as it has been found of much assistance in interpreting cor-
rectly the iron-carbon equilibrium diagram.

Enunciation of the Phase Rule. The phase rule was enunciated in 1878 by Wil-
lard Gibbs, at the time Professor of Physics in Yale University. It is one of the most
notable contributions ever made to physical chemistry.

The phase rule deals with the equilibrium of systems and is generally expressed

by the formula :

F = C + 2-P

showing the relation existing between the degrees of freedom (F) of a system, the num-
ber of components (C), and the number of phases (P); it tells us that the number of
degrees of freedom of any system is equal to the number of its components plus two,
minus the number of phases present. In order to understand the phase rule and its
application, it is necessary and sufficient to have an accurate understanding of the
meaning of the terms employed in its enunciation, namely, equilibrium, degrees of
freedom, components, and phases.

Equilibrium. A substance or system may be said to be in a state of equilibrium
when it is chemically and physically at rest, meaning by chemical rest that chemical
compounds are neither being tlissociated nor formed, and by physical rest, not the
absence of motion but the absence of molecular transformation, such as changes of
state or allotropic changes. It is necessary, however, to distinguish between homo-
geneous and heterogeneous substances. A substance is said to be homogeneous when it
is chemically and physically uniform throughout, i.e. when any two portions of it pos-
sess identical chemical and physical properties. Homogeneous substances are neces-
sarily gaseous mixtures, elements, chemical compounds, or liquid and solid solutions.
The equilibrium of a homogeneous system is sometimes called homogeneous equi-
librium. A heterogeneous substance is made up of two or more physically separate
parts, that is, of parts having different physical properties. Ice and water, many
rocks, and many alloys are instances of heterogeneous substances. If the com-
ponent parts of heterogeneous substances may be present in indefinite proportions,
the substances are mechanical mixtures; if they occur in definite proportions, the
substances are eutectic or eutectoid alloys. The equilibrium of heterogeneous systems
is sometimes called heterogeneous equilibrium.

Howe has recently suggested that the homogeneous constituents of alloys be called
"metarals" because of the great analogy between the constitution of metallic alloys
and of rocks, the minerals being the homogeneous components of the latter, while the
word aggregate is very frequently used to designate heterogeneous alloys. In metal-

2 LESSON XXIV THE PHASE RULE

lography, therefore, metarals and aggregates may conveniently replace the equivalent
terms, homogeneous and heterogeneous substances, of the physical chemist.

Only three independently variable factors can affect the equilibrium of a system,
namely, (1) the temperature, (2) the pressure, and (3) the concentration or composi-
tion. If arbitrary values may be given to one or more of these factors without destroy-
ing the chemical and physical rest of the system, its equilibrium is said to be stable.
On the contrary, if a change in value of any one of these three factors results in chemi-
cal or physical transformation, i.e. in atomic or molecular activity such as dissociation
or formation of chemical compounds, changes of state, or allotropic changes, the
equilibrium of the system was unstable. Water under atmospheric pressure is in
stable equilibrium, for we may change its temperature within wide limits without
causing it to undergo a change of state, while of course its chemical composition
remains likewise unaffected. All elements are generally in a state of stable equilib-
rium, as well as an infinite number of substances composed of two or more elements,
for they may be heated, for instance, through wide ranges of temperatures without
upsetting their physico-chemical equilibrium. Examples of unstable equilibrium,
however, are far from uncommon. During the solidification of substances, for in-
stance, stages must generally be passed through during which the equilibrium of the
substance is unstable, and it is often possible through very rapid cooling to retain
in the cold the unstable conditions, because of the rigidity of the substance now
opposing the changes needed for a return to stable equilibrium. It has been seen
in these lessons that hardened steel is, for the above reason, unstable, hence the
possibility of tempering it by slight reheating.

The kind of equilibrium known as metastable remains to be described. Liquids
may be cooled, when taking the necessary precautions, below their normal freezing-
point, without freezing, the phenomenon being known as superfusion, surfusion, or
undercooling, and the substance being said to be in metastable equilibrium.
Water, for instance, may be cooled below deg. C. and still remain liquid.
The introduction of a solid fragment of the substance, a piece of ice in the case
of water, results in the solidification of the liquid while its temperature rises to its
normal freezing-point. Otherwise, the substance may be kept liquid below its
solidification point for any length of time. If the temperature of the liquid con-
tinues to fall, however, a point is reached when its .equilibrium becomes unstable, i.e.
when further lowering of temperature causes the liquid to solidify. To state the case
broadly, the failure on the part of a system to undergo a certain chemical or physical
transformation when that transformation is due, although given the necessary time,
results in metastable equilibrium, while its failure to undergo a transformation because
of the necessary time being denied, as in quenching, results in unstable equilibrium.
Metastable equilibrium is stable, at least within narrow limits of temperature, while,
theoretically at least, slight heating of a substance in unstable equilibrium should
result in a partial return to a more stable condition, that is, in a partial occurrence of
the transformation that was suppressed by quick cooling.

Degrees of Freedom. By the degrees of freedom (sometimes called degrees of
liberty), of a system is meant the number of the three independently variable factors,
temperature, pressure, and concentration, which may arbitrarily be made to vary
without disturbing the system's physico-chemical rest. It has already been noted that
a system, in order to be in stable equilibrium, must have at least one degree of free-
dom. It will also be understood that no system can have more than two degrees of
freedom because in the case of arbitrary values being given to two of the factors, the

LESSON XXIV THE PHASE RULE 3

value of the third is necessarily fixed, this being due to the known rigid relations
existing between temperature, pressure, and concentration.

Systems whioh have no degree of freedom are said to be "un variant" or ''non-
variant." Their equilibrium is necessarily unstable. Systems having one degree of
freedom are called "univariant" or "mono variant," and those with two degrees of
freedom "bivariant" or "divariant."

Phases. By the phases of a system are meant the homogeneous, physically dis-
tinguishable, and mechanically separable constituents of that system. Water and ice,
for instance, are possible phases of the water-ice system; quartz, felspar, and mica are
phases of quartz, that is, of the silica-alumina-potash system. It will be apparent that
phases must necessarily be gaseous mixtures, elements, definite-chemical compounds,
or solutions. As previously mentioned, Howe, following the petrographical nomencla-
ture, and noting that the minerals are the phases of rocks, calls "metarals" the phases
of metals and alloys.

Components. The components of a system are described by Findlay as "those
constituents the concentration of which can undergo independent variation in the
different phases," by Bancroft as "substances of independently variable concentra-
tion," by Mellor as those "entities which are undecomposable under the conditions
of experiments," by Howe as "free elements and those compounds which in the nature
of the case are undecomposable under the conditions contemplated, and so play the
part of elements." The components of a system may be either chemical compounds
or elements, but there is at times some difficulty in grasping the distinction between
the components of a system and its ultimate chemical constituents. The criterion by
which to decide whether an entity is or is not a component, is the possibility of in-
dependent variation in the different phases. Take the system water, for instance:
evidently water and not hydrogen and oxygen is the component, because any
variation in the proportion of hydrogen would necessarily imply a corresponding
and well-defined variation in the proportion of oxygen and vice versa. Findlay
writes:

"In deciding the number of components in any given system, not only must the
constituents chosen be capable of independent variation, but a further restriction is
imposed, and we obtain the following rule: As the components of a system there are to be
chosen the smallest number of independently variable constituents by means of which
the composition of each phase participating in the state of equilibrium can be expressed
in the form of a chemical equation."

In the case of alloys, however, such difficulty does not arise, for the constituent
metals are always the components of the systems.

The Phase Rule Applied to Alloys. In dealing with alloys we may for all practi-
cal purposes ignore the influence of pressure, seeing that because of their feeble vola-
tility they are practically always subjected to atmospheric pressure. Omitting the
influence of pressure necessarily reduces by one the possible number of degrees of
freedom so that in the case of alloys the phase rule may be expressed by the formula :

F=C+1-P

signifying that the number of degrees of freedom is equal to the number of components
plus one, minus the number of phases. Since to be in stable equilibrium a system must
have at least one degree of freedom, it is obvious that an alloy made up of n metals
cannot have more than n phases. If it had n + 1 phases it would have no degree of
freedom, that is, its equilibrium would be unstable. With n- I phases it would have

4 LESSON XXIV THE PHASE RULE

two degrees of freedom. It could not have less than n - 1 phases, since it cannot have
more than two degrees of freedom.

The Phase Rule Applied to Pure Metals. Pure metals have only one component,
hut their possible phases are (1) liquid metal, (2) solid metal, (3) several allotropic con-
ditions of the solid metal. Let us consider Figure 1, which represents the solidification
of a pure metal as explained in Lesson XXII.

Above the temperature T the metal is entirely liquid; it has but one phase, and
consequently one degree of freedom (F =1 + 1-1 = 1). The system above T is
univariant; its temperature may be altered within wide limits without disturbing its

F- /+/-/=/
un/ vctr/anf

F=/+/-2=o
non- var/an t

F~- /+/-/=/

/ / me

Fig. 1. Equilibrium of pure metals according to the Phase Rule.

equilibrium: it remains liquid. At the temperature T two phases are present, solid
metal and liquid metal, the metal having, therefore, no degree of freedom (F = 1 + 1 -
2 = 0): it is non-variant. Liquid and solid metal can exist only at one temperature,
the critical temperature of solidification, any change of its temperature resulting in
the disappearance of one of the phases, that is, in a return to stable equilibrium.
Increasing the temperature must result in the disappearance of the solid phase, while
lowering the temperature must cause the disappearance of the liquid phase. Below
the temperature T the system contains only the solid phase, being, therefore, univari-
ant : its temperature may be varied arbitrarily.

The Phase Rule Applied to Binary Alloys. Binary alloys having for components
the two alloying metals, the formula of the phase rule becomes:

F = 2+ 1 -P
or F = 3 - P

LESSON XXIV THE PHASE RULE 5

Clearly binary alloys when in a condition of stable equilibrium cannot have more than
two phases. With one phase they will be bivariant, with two phases univariant, and
with three phases non-variant. Let us apply the rule to the fusibility curves of binary
alloys of metals partially soluble in each other when solid (Fig. 2). Above the liquidus
MEM' there is but one phase present, namely the liquid phase, the system being,
therefore, bivariant (F = 3 - 1 = 2), i.e. both temperature and concentration may be
varied arbitrarily without upsetting the equilibrium of the system, which means, in
the case under consideration, without causing its solidification. On reaching any
point L of the liquidus the alloy begins to solidify, and two phases are now present,
namely, solid solution and liquid alloy, the system becomes univariant (F = 3 - 2 = 1).
Having but one degree of freedom only the temperature or the TOircentration may be

M / O
M'/ too

Fig. 2. Equilibrium according to the Phase Rule of binary alloys whose component metals are
partially soluble in each other in the solid state.

arbitrarily varied. Should we, for instance, lower the temperature of alloy R from
T to T' (Fig. 2) the composition of the liquid phase necessarily shifts from L to L', and
that of the solid phase in equilibrium with it from s to s'. In the region MSES'M' of
the diagram bounded by the liquidus and solidus lines, therefore, the alloys are uni-
variant, any arbitrary change of temperature resulting in a well-defined change of
concentration and vice versa. At E, corresponding to eutectic composition and
eutectic temperature, three phases are present, namely two solid solutions and liquid
alloy, the system having no degree of freedom (F = 3 - 3 = 0). Neither the tempera-
ture nor the concentration may be altered without causing the disappearance of at
least one of the phases. Increasing the temperature must result in the disappear-
ance of both solid solutions, the system becoming bivariant, while lowering it must
be followed by the disappearance of the liquid phase. Again, shifting the concen-
tration to the left or right of E must yield the univariant system solid solution plus
liquid. Clearly two solid phases and a liquid phase can only exist at one critical
temperature and for one critical composition of the alloy; in the case of a eutectic

6 LESSON XXIV THE PHASE RULE

alloy these three phases can exist only at its freezing temperature. In the areas
AM SB and DM'S'C single homogeneous solid solutions only are present, that is, but
one phase exists, and the corresponding alloys have, therefore, two degrees of freedom.
Arbitrary changes both of temperature and composition within these areas do not
disturb the equilibrium of the system. Within the region BSS'C two phases occur,
solid solution M and solid solution M', the corresponding alloys having, therefore, but
one degree of freedom. Increasing the temperature from P to P', for instance, must
result in shifting the composition of the solid solutions respectively from R to R' and
from to 0'.

The Phase Rule Applied to Iron-Carbon Alloys. Since iron-carbon alloys belong
to the class of binary alloys the constituents of which are partially soluble in each
other in the cold, the application of the phase rule to their equilibrium diagram should
not present any difficulty, but we have now to consider allotropic changes as well as
changes of state. Their possible phases or metarals are: (1) liquid iron, (2) liquid
solution of carbon (or Fe 3 C) in iron, (3) solid solution (austenite) of carbon (or Fe 3 C)
in gamma iron, (4) solid gamma iron, (5) solid beta iron, (6) solid alpha iron (ferrite),
(7) solid solution (martensite) J of carbon (or FeaC) in beta iron, (8) solid cementite,
(9) graphite, and possibly others. The exact nature of troostite and sorbite being
still in doubt, they are not here classified as phases, seeing that they may be, and
probably are, aggregates of two or more phases, unless indeed they be, according to
Benedicks, emulsions or colloidal solutions. Scientists do not agree, however, as to
whether colloidal solutions are or are not phases, opinions differing in regard to their
homogeneity. Indeed some writers like Le Chatelier question the existence of colloidal
solutions which they consider as finely divided aggregates. Pearlite evidently is not
a phase, but an aggregate of the two phases, ferrite and cementite, in constant pro-
portion after the fashion of eutectic and eutectoid mixtures.

Let us now apply the teachings of the phase rule to the iron-carbon equilibrium
diagram (Fig. 3). The number and kinds of phases existing at different temperatures,,
and for different proportions of the components, iron and carbon, have been clearly
indicated and will be readily understood in view of the foregoing considerations.
Above the liquidus LEL' all alloys are composed of but one liquid phase, and have,
therefore, two degrees of freedom; between the liquidus and solidus, that is, in the region
LSE and L'S'E, two phases are present, liquid solution plus solid solution (austenite),
or liquid solution plus solid Fe 3 C, hence the corresponding alloys have here but one
degree of freedom; alloys of composition E and at the corresponding temperature are
evidently made up of three phases, namely liquid solution plus solid austenite plus
solid cementite, being, therefore, non-variant; in the region LADS all alloys being com-
posed of but one phase, namely, solid austenite, are bivariant; at D the alloy contains
three phases, ferrite, cementite, and austenite, and is, therefore, non-variant; in the
area DSS'F two phases are present, solid solution (austenite) plus cementite, and the
system has but one degree of freedom. If, as is often the case, cementite is in this
region decomposed into iron and graphite the alloys are for the time being non-variant,
becoming again univariant on the complete disappearance of cementite. In the region
ABU beta iron and austenite are present, the alloys having in consequence but one
degree of freedom; in the region BCDH alpha ferrite and austenite are present and
the alloys, therefore, are univariant. Finally below CDF three possible cases should
be considered: (I) the cementite formed during solidification and subsequent cooling

1 All investigators do not agree as to the homogeneity, that is the phase-like character of marten-
site, some still regarding it as an aggregate.

LESSON XXIV THE PHASE RULE

8 LESSON XXIV THE PHASE RULE

remains unchanged, in which case the alloy's are made up of the two phases ferrite and
cementite, being, therefore, univariant, their equilibrium, however, as previously ex-
plained, is supposed to be metastable; (II) the cementite has been completely con-
verted into ferrite and graphite, only those two phases being present, undoubtedly
representing the stable equilibrium of all iron-carbon alloys; (III) the dissociation of
cementite has been incomplete, both cementite and graphite being present, which
with ferrite give three phases, the corresponding alloys being non-variant and, there-
fore, their equilibrium unstable. Condition (I) generally prevails in all grades of steel,
and is readily produced in cast iron by rapid cooling especially in the absence of con-
siderable silicon, the resulting alloys being known as white cast iron. Condition (II)
never obtains in steel, but may be produced in highly carburized alloys by very slow
cooling through and below solidification, especially in the presence of much silicon
and in the absence of manganese and sulphur. Condition (III) is the condition of the
gray cast irons of commerce, their compositions and other influences prevailing during
their cooling being such as to cause the graphitizing of varying proportions of cementite.

Examination

Describe the application of the phase rule to iron-carbon alloys containing respec-
tively 0.60, 1.25, 3.00, and 5.00 per cent carbon as they cool from the molten condition
to atmospheric temperature.

APPENDIX I

MANIPULATIONS AND APPARATUS

In the foregoing pages the author has described at length those apparatus and
manipulations which in his laboratory he has found to yield the best results. In the
present appendix the apparatus and manipulations of some other workers are briefly
described.

POLISHING AND POLISHING MACHINES

Sorby in his pioneer work polished his samples on emery-papers of increasing
fineness followed by rubbing with tripoli, crocus, or Water-of-Ayre stone, and finally
with jeweler's rouge. Emery-papers are still used, but for quick polishing they
are often replaced by emery-powders spread wet on revolving wheels; the author
has retained the use of tripoli powder for the treatment preceding the final polish-
ing but others now prefer specially prepared flour-emery or diamantine; jeweler's
rouge is still widely used for the final treatment, although some prefer specially pre-
pared alumina, as first suggested by Le Chatelier.

In 1904 Osmond's polishing method consisted in roughing off with emery and
polishing with rouge. Emery-papers of increasing fineness were stretched over glass
plates. The papers used were prepared by mixing with water levigated "120
minutes" emery 1 and collecting the deposits formed at the end of increasing periods
of time in precipitating glasses. The classified powders, after drying, were mixed
with a mucilage of albumen (made up of 72 cubic centimeters of albumen and 28
cubic centimeters of water beaten to a froth and, after 12 hours, strained through a
fine-meshed sponge) and spread on paper of the best quality. Osmond also pre-
pared his own rouge by calcining copperas at as low a temperature as possible and
separating the finest product by levigation. The rouge was spread on a piece of
cloth stretched over the cast-iron table of a small horizontal polishing machine and
used wet.

In 1900 Le Chatelier's method of polishing specimens of iron and steel previously
rubbed upon emery-papers, including the finest grades, consisted in rubbing them
successively (1) on emery-paper prepared with albumen, according to Osmond, with
the deposit obtained in between a quarter of an hour and one hour in the ammoniacal
washing of flour-emery, (2) on a felt disk covered with some soap paste prepared
with the deposit of alumina or of emery, obtained in between one and three hours,
(3) on a flat disk made of wood, metal, or ebonite, covered with cloth, velvet, or
leather strongly glued upon it; upon this covering the soap preparation, obtained
with the deposit of alumina after twenty-four hours, was spread. The last two disks

1 By "120 minutes" emery is meant emery which has taken 120 minutes to settle in a vessel of
water of certain dimensions.

APPENDIX I MANIPULATIONS AND APPARATUS

were rotated by some mechanical devices producing great speed. All disks must be
frequently moistened with a brush or sponge.

According to Gcerens, Le Chatelier's method in 1908 consisted in the use of (1)
small sheets of French emery-paper, Hubert grades IG and 00 on ground glass plates,
(2) flannel stretched over glass covered with "one minute" emery previously passed
through a fine sieve (1200 meshes per sq. cm.), some soap solution being also poured
over the cloth, (3) a similar support covered with "120 minutes" emery previously
passed through a very fine sieve (2600 meshes per sq. cm.) and washed, and (4) a
vertically revolving brass disk covered with flannel and washed alumina. The fine
alumina mixed with water and soap solution may be sprayed on the disk by means
of the sprayer shown in Figure 1.

The preparation of fine alumina powder for the final polishing of iron and steel

Fig. 1. Sprayer for emulsified alu-
mina. (Gcerens.)

Fig. 2. Pipette for the levigation
of alumina. (Gcerens.)

samples was first described by Le Chatelier in 1900. The method used is that em-
ployed by Schloesing for the analysis of kaolins. The following description of Le
Chatelier's manipulation is from Gcerens (1908).

The purest precipitated alumina, from ammonia alum, is passed through a sieve
of 2600 meshes per sq. cm., and 100 grains of it in 300 c. c. of distilled water are trit-
urated in a mill for three hours. The whole is then poured into a liter flask, well
shaken, and about 200 c. c. pipetted off into a flask closed with a rubber stopper.
To this are added 1800 c. c. of distilled water and 2 c. c. of concentrated nitric acid
(1.4 sp. gr.), the mixture well shaken, and allowed to settle; the settling is complete
in a short time (about two hours). The clear supernatant liquid is siphoned off with
an S-shaped siphon; with careful manipulation this is possible to the extent of -fo
of the total amount. The liquid drawn off is replaced by distilled water, the mixture
well shaken several times, and allowed to settle again, after which the wash water is

APPENDIX I MANIPULATIONS AND APPARATUS 3

again drawn off as before. This is repeated three or four times more. At last the
supernatant liquid remains milky for a whole day, which is an indication of the per-
fect removal of acid. Finally, distilled water is added for the last time up to about
2 liters, the mixture thoroughly shaken, and the alumina separated from the liquid
in the apparatus shown in Figure 2. A pipette a of about 500 c. c. capacity is drawn
out below to an opening of about 3 mm. internal diameter. The alumina is prevented
from clinging by giving an inclination of at least Y% to the sloping sides of the tube.
The piece 6 is connected to the water pump. The end (of the pipette) is dipped
into the vessel containing the emulsified alumina, and the pipette sucked full, where-
upon the opening b is closed with a screw cap so far that one dmpjuns out about every
fifteen seconds. The material obtained during the first quarter of an hour is very
heterogeneous and still scratches the surface of the section markedly, so that it can-
not be used. After a quarter of an hour has expired the tap is closed and the alu-
mina allowed to settle complete!}'. After three hours the material is placed in the flask
A (Fig. 1) provided with a spraying arrangement. Soap solution 1 is added and the
mixture diluted with distilled water. The material thus prepared is ready for use,
and is suitable for steel and pig iron. The residue which settles in 3 to 24 hours is
treated similarly, and serves for polishing softer materials (iron, copper, etc.). The
portion which still remains in suspension after 24 hours is too fine and is poured away.

The same method has been applied to commercial flour-emery, oxide of chromium,
and oxide of iron, but the resulting products are far from being as satisfactory as the
alumina powders.

Revillon has recently described a rapid method of preparing alumina suitable for
polishing. A certain amount of alumina is suspended in a large volume of water,
well shaken, and allowed to stand for five minutes. The liquid is then siphoned off
and with the particles of alumina still held in suspension may be used for polishing.
To obtain finer powders 15 to 20 grams of finely ground alumina should be mixed
with one liter of distilled water, shaken, and allowed to settle five minutes. Most of
the liquid is then siphoned off, transferred into another vessel, and allowed to stand
fifteen minutes; the decantation is repeated, etc., a clearer liquid, that is one holding
finer particles in suspension, being obtained every fifteen minutes. The final liquid,
from which no powder is deposited, may be used for the finest polishing, the inter-
mediate products for rougher work.

Robin has described the preparation of alumina powder by a method based upon
the catalytic action of mercury in causing the oxidation of pure aluminum. Strips
of aluminum are immersed in mercury for a short time and then exposed to moist
air when the small amount of mercury they have absorbed causes the oxidation of
the metal, growth of A1 2 3 taking place, the increase of which is visible with the naked
eye. This alumina can be readily detached and as a fine powder may be used for
polishing. Robin claimed for his method the advantages of greater simplicity and
lower cost.

In 1900 Stead recommended for polishing iron and steel samples the use of emery-
papers, Hubert grades Nos. 0, 00, and 000, followed by rubbing with one grain of
diamantine powder 2 spread wet over a smooth black cloth and, for final treatment,
gold rouge used dry on chamois leather or, for finer structures, wet on parchment or

1 The .soap solution is prepared by dissolving pure (Venetian) soap in hot water and filtering
through a filter paper into a flask. After cooling the solution should he sirupy.

2 Diamantine powder consists of pure alumina and is used by jewelers for polishing steel.

4 APPENDIX I MANIPULATIONS AND APPARATUS

kid leather. He used a simple, hand polishing machine in which one block at a time
was made to rotate horizontally (Fig. 3).

A foot polishing machine also designed by Stead is shown in Figure 4 and a

cab

larger one to be run by power in Figure 5. In these machines brass disks carry-
ing conical wooden blocks are attached to vertical spindles and driven from below.
Emery-papers are fastened to some of these blocks by means of brass rings slipping

APPENDIX I MANIPULATIONS AND APPARATUS 5

over them while others are covered with cloth in a similar way. Clamps are provided
for holding the samples against the revolving disks. The central vessel contains the
water needed for wet polishing, a small tap projecting over each disk. The excess
water is caught by brass water guards and discharged into a trough below the level
of the disks. These machines are made by Carling and Sons of Middlesbrough, Eng-
land.

Martens, according to Gcerens (1908), uses vertically rotating disks upon which

Fig. 4. Foot power polishing machine.
(Stead.)

Fig. 5. Multiple polishing machine. (Stead.)

are pasted emery-papers, Hubert brand, grades 3, 2, 1G, 1M, IF, 0, and 00 and, for
final treatment, levigated jeweler's rouge on cloth. The disks make 400 revolutions
per minute. The average time needed to polish a specimen varies between \Yi and
2 hours.

Gulliver (1908) recommends for polishing the use of emery-papers grades No. 1,
0, and 00 on hard wood or plate glass and for final treatment the finest rouge or dia-
mantme powder on cloth stretched over hard wood.

The polishing machines shown in Figures 6 and 7 are made by P. F. Dujardin of
Dusseldorf. It will be noted that one side only of the disks is utilized. A machine
like the one of Figure 7 is also made for belt driving.

APPENDIX I MANIPULATIONS AND APPARATUS

Sexton describes the polishing machine Figure 8, made by Baird and Tatlock.
Its construction is obvious.

A simple polishing machine consisting of an horizontally revolving disk (Fig. 9)
was described in 1899 by Ewing and Rosenhain. A is the spindle of an electric motor
carrying a small driving disk B, fitted with a rubber ring to increase the driving fric-

Fig. G. Foot power polishing machine. (P. F. Dujardin and Co.)

tion. The polishing disk C has a vertical axis running in a bearing on the casting D.
The under side of the polishing disk bears upon the driving wheel B and takes motion
from it.

A. Kingsbury in 1910 described his polishing method. He prepares his support-
ing blocks by pouring paraffin on brass disks. After solidifying these paraffin
blocks which are about J/ inch thick and 8 inches in diameter have their upper face
dressed flat. They are made to rotate horizontally in a suitable machine and upon
them emery of increasing fineness and finally rouge are used in succession. The

APPENDIX I MANIPULATIONS AND APPARATUS 7

speed of the polishing machine is 200 revolutions per minute. The time needed to
polish a sample of ordinary steel is given as fifteen minutes.

C. Campbell in 1902 described the polishing operation as consisting in rubbing the
sample, previously filed smooth, successively on emery-cloth, grades and 00, and on
French emery-papers, grades 0, 00, 000, and 0000. The specimen is then polished on

Fig. 7. Polishing motor. (P. F. Dujardin and Co.)

broadcloth or chamois leather with well washed rouge and water. Some workers,
the writer says, use an intermediate stage with diamantine powder.

C. H. Risdale in 1899 described his polishing operation as consisting in (1) rough
filing, (2) fine filing, (3) rubbing with rough commercial emery-cloth stretched on a
board, (4) rubbing with fine emery-cloth stretched on a board, (5) rubbing on fine
specially prepared paper on disks of Stead's polishing machine, (6) rubbing on dia-
mantine on cloth stretched on disks of Stead's machine, (7) rubbing on rouge on

APPENDIX I MANIPULATIONS AND APPARATUS

washed leather similarly mounted or, for very fine work, on rouge on wetted parch-
ment.

Guillet in 1907 recommended for polishing two carborundum wheels and two
suitably selected emery-papers and, for final treatment, alumina on cloth stretched
over a revolving disk. He places smooth sheets of zinc between the wooden disks of
his polishing machines and the polishing cloths.

In 1901 Arnold described as follows a quick polishing and etching method: "Take

Fig. 8. Polishing machine. (Baird and Tatlock.)

Fig. 9. Polishing machine. (Ewing and Rosenhain.)

two pieces of hard wood, 12" x 9" x 1", planed dead smooth on one side; then by
means of liquid glue evenly attach to the smooth faces two sheets of the London
Emery Works Go's atlas cloth No. 0. Allow the glue to set under strong pressure.
Next, by means of a smooth piece of steel, rub off from one of the blocks as much as
possible of the detachable emery. This is No. 2 block, the other, necessarily, No. 1
block.

"The steel section, say J^ inch thick and 3/2 inch diameter, is rubbed for one minute

APPENDIX I MANIPULATIONS AND APPARATUS 9

on No. 1 block, the motion being straight and not circular; then, for the same time
and in the same manner rub on No. 2 block. Next place the bright but visibly
scratched sections in a glass etching dish 3" X 1" X H"> an d cover the steel with nitric
acid sp. gr. 1.20.

"Watch closely until in a few seconds the evolved gases adhering to the section
change from pale to deep brown and effervescence ensues. Then, under the tap,
quickly wash away the acid and for a minute immerse the piece in a second dish con-
taining rectified methylated spirits. Dry the section by pressing it several times on
a soft folded linen handkerchief, when it will be ready for examination. The struc-
ture will be clearly exhibited, the innumerable fine scratches^ visible before etching
having virtually vanished."

DEVELOPMENT OF THE STRUCTURES

The methods which, in common with many workers, the author has found most
satisfactory for revealing the structure of polished iron and steel specimens have been
described in these pages. They include etching with concentrated nitric acid, with
very dilute alcoholic solutions of nitric acid aqd of picric acid (Lesson III), with
sodium picrate and ammonium oxalate (Lesson V), and with the Kourbatoff reagent
(Lesson XIII). Other methods have been used that should be mentioned.

Polishing in Relief. So-called relief polishing has been used successfully by
Sorby, Martens, Behrens, and especially by Osmond. It consists in rubbing the
specimen on a soft, yielding support with some suitable polishing powder, the softest
constituents being, so to speak, dug out, leaving the harder ones standing in relief.
These differences of level make it possible to distinguish the constituents under the
microscope without further treatment. It is evident that only those samples which
are made up of constituents differing much in hardness can be so treated. The free
cementite of hyper-eutectoid steel or of white cast iron, for instance, can be made to
stand strongly in relief because it is so much harder than the accompanying pearlite
or other constituents.

Osmond polishes his samples on a damp piece of parchment stretched over a piece
of well-planed wood. It is sprinkled with rouge which is rubbed strongly on the
parchment. The block is then put under the tap and washed so that only those part-
icles of rouge that have found their way into the pores of the parchment are retained.
To distinguish between raised portions and cavities the luminous rays are strongly
diaphragmed and the objective placed a little below the focusing point, is slowly
raised. The reliefs, which at first appear brilliant and yellowish on a relatively
darker ground, gradually become dark on a bright ground; the cavities present in-
verse appearances so perfectly that two photographs of the same preparation, taken
one a little below and the other a little above the mean focusing point, are almost
positive and negative to one another.

Polish-Attack. For many years Osmond obtained his best preparations by a
combined polishing and etching method (polissage-attague) consisting in rubbing the
polished sample upon a piece of parchment covered with some aqueous extract of
liquorice root, with the addition of precipitated calcium sulphate. In 1899 Osmond
and Cartaud recommended replacing the extract of liquorice by a diluted solution of
nitrate of ammonium (2 parts by weight of the crystallized salt to 100 parts of
water). A piece of parchment spread tightly over a smooth board is soaked with the

10 APPENDIX I MANIPULATIONS AND APPARATUS

solution and the specimen rubbed upon it until sufficiently etched. It is not necessary
to add any sulphate of calcium.

Etching. Sorby etched his specimens with very dilute solutions of nitric acid in
water and this reagent was widely used for many years by other metallographists.
The water has now been replaced by absolute alcohol (Lesson III, page 7). Le Chate-
lier has mentioned the use of glycerine as a satisfactory non-oxidizing vehicle for
nitric as well as for picric and hydrochloric acid.

Osmond, the author believes, was the first to use tincture of iodine. This tinc-
ture is applied in the proportion of one drop per square centimeter of surface and
allowed to act until it is decolorized, the treatment being repeated after examina-
tion if needed. Le Chatelier recommends applying the tincture with the tip of the
finger and gently rubbing the specimen.

Stead uses a solution made up of 1.25 grains of iodine, 1.25 grains of iodide of
potassium, 1.25 grains of water and alcohol to make up 100 c. c. After the iodine has
lost its color the sample should be washed in water, then in alcohol, and finally dried
in a blast of hot air.

Martens and Heyn in 1904 recommended the use (1) of an etching solution con-
taining one part of hv-Jrcchloric acid (1.19 sp. gr.) and 100 parts of absolute alcohol,

Fig. 10. Arrangement for electrolytic etching.

and (2) of one part of hydrochloric acid in 500 parts of water with the assistance of
the electric current.

Heyn used a solution of double chloride of copper and ammonium containing 12
grains of the salt and 100 grains of distilled water.

To distinguish with certainty between iron phosphide and cementite, Matweieff
recommends neutral sodium picrate washed several times with distilled water to
eliminate the excess of picric acid or of sodium that might be present. The sample
is immersed in the boiling solution for 20 minutes, a treatment by which the iron
phosphide is strongly attacked while the cementite and pearlite remain unaffected.

For etching austenite and martensite Robin recommends the use of a saturated
solution of picric acid in alcohol, an immersion of thirty seconds to one minute, wash-
ing with water without touching the specimen and drying by air blast or simply in
air. Films of various tints are formed, ferrite remaining uncolored.

Le Chatelier has used bitartrate of potassium as an etching reagent. It leaves
cementite and pearlite uncolored, while it imparts a dirty coloration to ferrite.

The same author has described the use of a freshly prepared reagent made up of
equal parts of a solution containing 50 per cent of soda and of a solution containing
10 per cent of lead nitrate. Cementite is quickly colored by it while the phosphides
and especially the silicides are also attacked. The reagent is recommended for highly
carburized metals. Medium high carbon steels of great purity are not affected by
this solution, but when impure, the pearlite is energetically acted upon, probably
because of the presence of impurity in that constituent.

APPENDIX I MANIPULATIONS AND APPARATUS 11

Le Chatelier has also mentioned the use of a solution of 10 per cent gaseous hy-
drochloric acid in absolute alcohol to which is added 5 per cent of cupric chloride for
annealed steels and one per cent of the same salt for hardened steels. Ferrite and
cementite are not colored, martensite very little, austenite a little more, troostite
and sorbite decidedly.

Hilpert and Colver-Glauert have described the use of sulphurous acid for non-
pearlitic steels and for pig iron. A saturated solution of sulphur dioxide in water is
prepared and 3 or 4 per cent of tliat solution in water used. The time of etching
varies between seven seconds and one minute. Alcohol may be substituted for water
in which case the etching lasts several minutes. The treatment causes the deposi-
tion of layers of iron sulphide of different thickness and, therefore, of different colors,
on the various constituents.

Electrolytic Etching. Lc Chatelier was one of the first to advocate the use of
the electric current in order to obtain a more uniform action in etching iron and steel
samples. Sheet lead may be used for the positive electrode, and, as electrolyte, a 10
per cent solution of chloride or sulphate of ammonium gives good results. The cur-
rent needed varies between 0.001 and 0.01 amperes per square centimeter.

Electrolytic etching has been described by Cavalier (1909). A few cubic centi-
meters of the electrolyte are placed in a platinum dish C (Fig. 10) connected with
one pole of the battery P; the specimen E connected with the other pole is placed
in the solution, a piece of filter paper A being inserted between the dish and the
polished surface of the specimen. The current is regulated through the rheostat R.
Four or five volts are required with an intensity of 0.001 to 0.01 amperes per square
centimeter. The attack lasts from a few seconds to a few minutes.

Heat Tinting. Heat tinting as a means of imparting different appearances to
the various constituents of iron and steel was first used by Behrens and Martens and
later, with much success, by Stead. When a polished piece of iron or steel is heated
in an oxidizing atmosphere oxidized films are formed, the color of which varies with
the thickness, that is, with the temperature and duration of treatment. It is also
found that the various constituents are differently colored because oxidized at dif-
ferent speeds. According to Stead the metal should be first well rubbed with a piece
of linen or chamois leather and placed on an iron plate heated by a Bunsen burner.
It is best to heat gradually and examine periodically under the microscope and stop
when the structure appears to be most perfectly colored. After each heating the
section may be placed in a dish of mercury so as to cool it rapidly and check further
oxidation. The oxidized films assume in succession the following tints as they in-
crease in thickness: pale yellow, yellow, brown, purple, blue, and steel gray. The
method is especially useful for identifying phosphides, sulphides, and carbides in cast
iron and for detecting the more highly phosphorized portions of iron and steel. Free
cementite colors less readily than iron but more rapidly than phosphide of iron. Iron
containing phosphorous in solid solution colors more rapidly than pure iron or than
iron containing less phosphorus.

Hot Etching. Steel while at a high temperature (red heat) has been etched

(1) by Saniter in molten calcium chloride heated to the desired temperature, and

(2) by Baykoff in a current of gaseous hydrochloric acid.

Washing and Drying. After removing the specimens from the etching bath, the
author washes them in alcohol and dries them in an air blast. They are then rubbed
once or twice very gently on a block covered with a fine piece of chamois skin and

APPENDIX I MANIPULATIONS AND APPARATUS

carefully kept free from dust. Washing in water, in caustic potash, in lime water,
and in ether has also been recommended as well as the use of fine linen cloth and of
a hot blast for drying.

Preserving. The author preserves his etched specimens in dessicators and in
air-tight cabinets.

Several protective coatings have been described. Stead covers them with paraf-
fin wax dissolved in benzole, which is removed by wiping with a clean linen rag mois-
tened with benzole, when it is desired to examine the specimens. Le Chatelier applies
a coating of "zapon," a solution of gun cotton in amyl acetate sufficiently transparent
to allow examination with the highest powers.

By keeping the specimens in mercury their tarnishing should be effectively pre-
vented while they would be at all times accessible for immediate examination. Nor
should this scheme call for the use of a large amount of mercury nor for much space;
flat glass trays might be used containing just enough mercury to cover their smooth
bottom and the specimens placed in them polished face down. In this way a large

.Glass slide

Brass ri

Class

Fig. 11. Stead's mounting device.
(C. H. Desch's Metallography.)

Fig. 12. Gulliver's mounting device.

number of samples could be preserved in a small place and in a small quantity of
mercury. In a tray measuring 12 by 12 inches, for instance, nearly 200 samples of
ordinary size (^ to % inch in diameter) could be kept.

MOUNTING AND MOUNTING DEVICES

The author's special holders for placing the prepared samples on the stage of the
microscope have been described (see Apparatus for the Metallographic Laboratory,
page 7). Other methods have been used and are still employed by some workers,
namely (1) mountir. " in some plastic material, and (2) the use of leveling stages.

Plastic Mounting. Osmond mounts his specimens by embedding them in a
little soft wax placed upon a glass plate. The leveling is managed by means of two
pieces of glass tube of equal height, one on each side of the sample.

Stead places the specimens polished face down on a piece of plate glass (Fig. 11)
and surrounds them with brass cylinders accurately turned. A piece of plastic wax
is stuck upon the center of a glass microscope slide and is then pressed upon the sec-
tion till the glass slide comes in contact with the brass ring. The specimen adheres
to the wax and the mounting is complete.

Gulliver (1908) describes the device (Fig. 12) for mounting specimens. It consists

APPENDIX I MANIPULATIONS AND APPARATUS

of a circular ring faced on its upper surface A, and screwed internally at B to fit the
foot, of which the upper end C is also faced. The distance between the parallel faces
A and C can thus be adjusted. The specimen is placed at D and a glass slide E with
some soft modeling clay or wax is pressed upon it until the glass touches the ring
at AA.

Mechanical mounting devices working on the principle of the microtome have
also been used. They have been described by M. A. Richards: "Projecting from a
cylindrical metal base three inches in diameter, is a threaded upright three and one-

Fig. 13. Watson and Sons' mounting
device.

Fig. 14. Watson and Sons'
leveling stage.

Fig. 15. Huntington's leveling stage.

half inches in diameter. A cylindrical nut or collar three inches high and two and
one-half inches outside diameter screws on the threaded uprigL A small circle of
chamois skin is placed on the top of the thread upright to protect- the etched face of
the micro-section. To mount a section, place it face down on the chamois skin, press
upon the upper projecting portion a lump of beeswax, and upon this place the ground
glass (ground surface down). A few revolutions of the collar will cause the glass to
rest upon the upper edge of the collar, and the adhesion of the glass and beeswax to
the specimen may be made complete by slowly turning the collar down with one hand
while keeping the glass base in close contact with the collar-top with the other hand.
In this manner, no matter how irregular the section, the parallelism of the etched
surface and the glass base may be very quickly and accurately obtained."

APPENDIX I MANIPULATIONS AND APPARATUS

The mounting device (Fig. 13) is constructed by Watson and Sons. It consists of
two horizontal plates, the upper one being capable of vertical movement but always
remaining parallel to the lower one. The specimen is placed with its polished surface
on the lower plate, and the upper plate carrying a glass slip to which some suitable
clay or wax is attached is lowered into contact.

Leveling Stages. The leveling stage (Fig. 14) is constructed by Watson and

Fig. 1G. Le Chatelier's inverted metallurgical microscope.
Early form.

r 1 Ol

f l 1

1 .

F I 1

"~1 i

Fig. 17. Le Chatelier's inverted metallurgical microscope.

Sons, London. The specimen is held by two rotating jaws and can bo leveled by
means of the screws A and B BI.

Professor A. K. Huntington devised the leveling stage shown in Figure 15. It is
provided with a ball and socket joint for leveling, permitting the placing of the sam-
ple in any position.

Other forms of leveling stages are shown in some of the illustrations in the follow-
ing pages as part of some metallurgical microscopes.

APPENDIX I MANIPULATIONS AND APPARATUS

METALLURGICAL MICROSCOPES

The microscopes and accessories used by the author have been fully described.
In the following pages instruments used by some other workers or described by them,
as well as those manufactured by well-known makers, are mentioned.

Le Chatelier. In 1897 Le Chatelier devised an inverted microscope which later

Fig. 19. Le Chatelier's inverted metallurgical microscope.

Fig. 20. Device for placing
specimens on the stage of
the Le Chatelier micro-
scope in a fixed position.
(Le Grix.)

he greatly improved and which is now constructed with unimportant modifications
by several microscope makers. An early form of Le Chatelier's instrument is shown
in Figure 16 and its more recent construction in Figures 17 and 18. The objective
B (Fig. 17) is directed upwards while the eye-piece 0, placed horizontally, receives
the image by the reflection of a totally reflecting prism F placed below the objective.
The prism F may be rotated by means of the milled head P and the light reflected by
the objective turned at will into the tube G and the eye-piece O for visual examination

APPENDIX I MANIPULATIONS AND APPARATUS

or into another tube connected with a camera for photographing (Fig. 18). The
light is condensed by the lens A and, being deflected at right angles by the prism /,
passes through the objective B and reaches the object M placed on the stage E. In
case the light is placed at a higher level than the condensing lens A, it must be re-
ceived by a totally reflecting prism H which directs it into the condenser A. D is a
diaphragm placed at the principal focus of the complex optical system composed of
the objective R, the illuminating prism J, and the lens A, The opening as well as the
position of the diaphragm may be altered. Another diaphragm placed at 7 affords
a means of stopping the light which would fall upon parts of the preparation outside
of the portion examined and which would increase the blur resulting from the reflec-
tion of useless rays by the back lenses of the objective. In the early construction of
this instrument when the object was to be photographed the prism F was withdrawn
from the path of light and the image allowed to form on a photographic plate placed
below (Fig. 16).

A slightly different construction is shown in Figure 19. For photographic pur-
poses the image forms on a plate placed in a holder rigidly connected with the instru-

Fig. 21. Inverted metallurgical microscope constructed by E. Leitz.

ment, no eye-piece being used. As the distance between the photographic plate and
the objective is short, very small photomicrographs are obtained, which must gen-
erally be subsequently enlarged. Z is a plate carrying an eye-piece for use with the
long bellows camera (Fig. 18). The Le Chatelier microscopes are constructed by
Ph. Pellin of Paris.

In order to be able to examine identical portions of the same specimen at different
times with the Le Chatelier microscope, Le Grix (1907) suggested the arrangement
shown in Figure 20. A circular metallic disk with rectangular opening RR' and carry-
ing two pointed stops A and B is fitted to the stage. A file mark E is made in
the specimen M, which is then placed on the stage so that the stop A enters the
groove E while the specimen presses against the other stop B, in this way securing
a constant position for the object.

Ernst Leitz. A slightly modified form (Fig. 21) of the Le Chatelier inverted
microscope is made by Ernst Leitz of Wetzlar, Germany. The modifications were
suggested by Guertler. The stage and illuminating appliances are shown on a larger
scale in Figure 22.

The same maker also manufactures the microscope shown in Figures 23 and 24

APPENDIX I MANIPULATIONS AND APPARATUS

designed by W. Campbell. The stage can be removed and the upper part of the in-
strument attached to the base for the examination of large surfaces.

P. F. Dujardin. P. F. Dujarclin and Co. of Diisseldorf construct a Le Chatelier

Fig. 22. Inverted metallurgical microscope constructed by E. Leitz.

Fig. 23. Metallurgical microscope con-
structed by E. Leitz.

Fig. 24. Metallurgical microscope
constructed by E. Leitz.

inverted microscope as shown in Figure 25. They also make the microscope (Fig. 26)
in which the vertical illuminator carries its own source of light and condenser.

C. Reichert. The metallurgical microscope (Fig. 27) designed by Professor
Rejto is made by C. Reichert of Vienna. The position of the vertical illuminator
immediately below the eye-piece should be noted. The stage is provided with a level-

APPENDIX I MANIPULATIONS AND APPARATUS

ing mechanism. The same maker manufactures an inverted Le Chatelier microscope
as shown in Figure 28. According to Desch, in this microscope, two right-angled
prisms are cemented together to form the cube P (Fig. 29). The upper prism is sil-

vered over an elliptical area, as shown by the central dark line. A portion of the
light proceeding from the mirror M passes through the clear portion of the glass cube
P and falls upon the object S. The light reflected back by the object upon striking

APPENDIX I MANIPULATIONS AND APPARATUS

the silvered portion of the prism is deflected at right angles into the tube which
conducts it to the eye or to a photographic plate.

R . Fuess. A metallurgical microscope practically identical in construction to
the Le Chatelier inverted instrument is made by R. Fuess of Steglitz, near Berlin.

Robin. The microscope and photographic attachment shown in Figure 30 was

Fig. 26. Metallurgical microscope constructed by P. F. Dujardin

and Co.

ERT,WIEN.

Fig. 27. Metallurgical microscope con
structed by C. Reichert.

designed by Robin. Visual examination is possible only on the screen of the camera.
The stage consisting of a smooth disk is tilting and the specimen is fastened upon it
with wax. To secure an accurately horizontal position of the polished surface, a
plug with a perfectly flat surface is screwed into the microscope in place of the ob-
jective and the stage raised until the specimen coming in contact with the plug, the
latter through gentle pressure causes the polished surface to assume a horizontal
position. The plug is then removed and the objective inserted.

APPENDIX I MANIPULATIONS AND APPARATUS

Martens. The Martens metallurgical microscope (1899) made by Zeiss of Jena is
shown in Figure 31. It can be used horizontally only, the tube is very wide and

Fig. 28. Inverted metallurgical microscope constructed by C. Reichert.

Fig. 29. Illuminating prisms of
Reichert's inverted microscope.

Fig. 30. Metallurgical microscope designed by Robin.

the vertical, mechanical stage is provided with both coarse and fine adjustments Y
and Z and with leveling screws aa. The flexible connection / permits the focusing

APPENDIX I MANIPULATIONS AND APPARATUS

of the object from the camera screen. The instrument is designed especially for
photography.

A complete Zeiss equipment including a large' electric arc lamp is shown in
Figure 32. It will be noted that the mounting of the camera is entirely separate
from that of the other parts.

Martens also designed the ball-jointed microscope (Fig. 33) which he used prin-
cipally for observing the progress of etching.

Rosenhain. The microscope shown in Figure 34 was constructed by R. and J.
Beck for Rosenhain. The stage is mechanical and provided with coarse and fine

Fig. 31. Martens metallurgical microscope.

adjustments, and all controlling heads are placed beneath. Appliances are provided
for various kinds of illumination. The necessary alteration of focus for photograph-
ing can be done at the eye-piece by a suitable arrangement provided for that purpose.

Osmond. Osmond used a Nachet microscope of the ordinary type connected
with a vertical camera and a prism illuminator. He writes, however, that special
metallurgical microscopes "are certainly to be preferred." In Osmond's opinion the
vertical is very much superior to the horizontal camera for studying metals.

Nachet. Nachet of Paris constructs the metallurgical microscope (Fig. 35) .
The vertical illuminator carries a tube provided with an iris diaphragm. The stand
is to be used in the vertical position only. The stage has a coarse vertical adjust-
ment. A similar microscope is made with mechanical stage provided with both
coarse and fine adjustments.

The prism illuminator (Fig. 36) designed by Guillemin is made by Nachet. A

APPENDIX I MANIPULATIONS AND APPARATUS

lateral as well as a slight tilting motion may be imparted to the prism through the
milled heads B and C.

Fig. 32. Martcns-Zeiss metallurgical microscope and camera.

Fig. 33. Martens ball-jointed
microscope.

Nachet's illuminating objectives have been described and illustrated (Apparatus
for the Metallographic Laboratory, page 17).

Cornu-Charpy. The arrangement shown in Figure 37 was used by Charpy. The
vertical illuminator G consists of four thin glass plates placed at an angle of 45 deg.

APPENDIX I MANIPULATIONS AND APPARATUS

immediately below the eye-piece and it receives the light reflected by the totally
reflecting prism P. This prism is so mounted that it can rotate freely around the
axis of the microscope and also around the axis GP of the tube to which it is at-
tached, thus making it possible to receive upon it the light proceeding from a source
of light placed anywhere.

Fig. 34. Rosenhain metallurgical microscope.

Watson and Sons. The metallurgical microscope (Fig. 38) was constructed
in 1904 by Watson and Sons of London. The stage is provided with both coarse and
fine adjustments. The same makers following Martens construct the horizontal
metallurgical microscope (Fig. 39). The plain glass vertical illuminator (Fig. 40)
provided with iris diaphragm is also made by Watson and Sons.

APPENDIX I MANIPULATIONS AND APPARATUS

Fig. 35. Nachet metallurgical microscope.

Fig. 37. Cornu-Charpy metallurgical
microscope.

Fig. 36. Guillomin-Nachet
prism illuminator.

APPENDIX I MANIPULATIONS AND APPARATUS

Fig. 38. Metallurgical microscope constructed by
Watson and Sons.

Fig. 39. Horizontal metallurgical microscope constructed by
Watson and Sons.

APPENDIX I MANIPULATIONS AND APPARATUS

It. and ./. Beck. In 1904 R. and J. Beck of London constructed the prism ver-
tical illuminator shown in Figure 41. The device is fitted with an iris diaphragm

Fig. 40. Watson and Sons
vortical illuminator.

Fig. 41. Bock prism
illuminator.

Fig. 42. Beck surface microscope.

Fig. 43. Metallurgical microscope. (Queen and Co.)

beneath the prism for cutting off outside light, and a plate of stops so arranged that
the position of the beam of light impinging on the prism can be varied until parallel
light of the right angle is obtained. The same makers construct the instrument

APPENDIX I MANIPULATIONS AND APPARATUS

shown in Figure 42 for the examination of large metallic surfaces. The Rosenhain
microscope described in these pages is likewise made by R. and J. Beck.

Queen and Co. Queen and Co. of Philadelphia at one time (1898) placed on the
market the microscope and camera shown in Figure 43. The camera could be tilted
on one side for ocular examination. The same makers now construct the microscope
shown in Figure 44.

Arthur H. Thomas Co. Arthur H. Thomas Co. of Philadelphia are offering

J 1

Fig. 44. Metallurgical microscope. (Queen
and Co.)

for sale an illuminator designed by Wirt Tassin (Fig. 45). A condensing lens and
acetylene burner are attached to the vertical illuminator.

F. Koristka. The prism vertical illuminator (Fig. 46) was described by F.
Koristka of Milan in 1905. An iris diaphragm placed in front of the prism controls
the light which it receives. By pulling out the arm carrying the prism the latter may
be removed from the field.

Ph. Pellin. The Le Chatelier inverted microscope is constructed by Ph. Pellin
of Paris. The same makers also manufacture a portable microscopic outfit designed

APPENDIX I MANIPULATIONS AND APPARATUS

by Guillet (Trousse de Metallographie). It includes a small electric motor for polish-
irg, a vertical microscope so constructed that it can be fastened to any object it is

Fig. 45. Microscope and camera with Tassin
illuminator attached.

Fig. 40. Koristka prism illuminator.

desired to examine, files, emery-papers, etching reagents, etc. All parts are com-
pactly placed in a carrying case.

30 APPENDIX I MANIPULATIONS AND APPARATUS

Carl Zeiss. The instruments used by Martens and Heyn already described in
these pages are constructed by Carl Zeiss of Jena. The prism vertical illuminator
made by the same firm has been described and illustrated in the introductory chapter
on Apparatus.

Spencer Lens Co. The Spencer Lens Co. of Buffalo, N. Y., manufacture a
vertical metallurgical microscope with movable stage.

Bausch and Lomb Optical Co. The microscopes and accessories used and de-
signed by the author and fully described in these .pages are manufactured by the
Bausch and Lomb Optical Co. of Rochester, N. Y.

APPENDIX II

REPORT OF COMMITTEE 53 OF THE INTERNATIONAL ASSOCIATION FOR

TESTING MATERIALS

ON THE NOMENCLATURE OF THE MICROSCOPIC SUBSTANCES
AND STRUCTURES OF STEEL AND CAST IRON

Presented by the Chairman H. M. HOWE and the Secretary of the Committee ALBERT SAUVEUB
at the VI th Congress, New York, September, 1912

The Committee for studying this problem is constituted as follows:

Professor H. M. HOWE, Chairman, New York.

Professor ALBERT SAUVEUR, Secretary, Cambridge, Mass.

Members: F. OSMOND, Paris; Dr. H. C. H. CARPENTER, Manchester; Prof. W.
CAMPBELL, New York; Prof. C. BENEDICKS, Stockholm; Prof. F. Wiisx, Aachen;
Prof. A. STANSFIELD, Montreal; Dr. J. E. STEAD, Middlesbrough; Prof. L. GUILLET,
Paris; Prof. E. HEYN, Berlin-Lichterfelde; Dr. W. ROSENHAIN, Teddington.

I. GENERAL PLAN

We first enumerate the substances of such importance as to warrant it, indicating
roughly their constitution, and then define and describe certain of them.

The conditions which we meet a^e (1) that we need definitions on which all can
agree; and this implies that they must be free from all contentious matter and be
based on what all admit to be true; (2) that the reader must needs know the current
theories as to the constitution of these substances, and these theories are necessarily
contentious. We meet these conditions by the plan of giving (1) the Name which we
recommend for general use, followed immediately in parentheses by the other names
used widely enough to justify recording them; (2) the Definition proper, based on an
undisputed quality, e.g. that of austenite, which we base on its being an iron-carbon
solid solution, purposely omitting all reference to the precise nature of solvent and
solute; and (3) Constitution, etc., etc., in which we give the current theories as to the
nature of solvent and solute and appropriate descriptive matter.

The distinction between these three parts should be understood. (1) The Names
actually used are matters of record and indisputable. (2) The Definitions are matters
of convention or treaty, binding on the contracting parties, though subject to de-
nouncement, preferably based on some determinable property of the thing defined as
distinguished from any theory as to its nature, or if necessarily based on any theory
it should be a theory which is universally accepted. It is a matter purely of conven-
tion and general convenience what individual property of the thing defined shall
form the basis of the definition. The name and the definition should endure perma-
nently, except in the case of a definition based on an accepted theory, which must be

2 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

changed if the theory should later be disproved. (3) Theories and Descriptions are
not matters of agreement or convention but dependent on observation, and therefore
always subject to be changed by new discoveries. They are temporary in their nature
:is distinguished from the names and definitions which should be fixed, at least rela-
tively.

This case of austenite illustrates the advantage of non-indicative names. The
names which we propose to displace, "gamma iron" and "mixed crystals," imply
definite theories as to the nature of austenite, and hence might have to be abandoned
in case those theories were later disproved. The name "austenite" implies nothing,
like mineralogical names in general, and hence is stable in itself. Our infant branch
of science may well learn from its elder sister, which has tried and proved the advan-
tage of this non-indicative naming.

In those cases in which a name has been used in more than one sense we advise
the retention of one and the abandonment of the others, having obtained the consent
of the proposers of such names for their abandonment.

Many whose judgment we respect object to our including certain of the less used
names, e.g. from i to n in our list, holding them either to be confusing or to be needless.

It is true that several names (hardenite, martensite, sorbite, etc.), have been used
with various meanings, and hence confusingly, in spite of which most of them should
Lie retained, each with a single sharp-cut definition, because they are so useful.

As regards the alleged needlessness of certain names it is for each writer to decide
whether he does or does not need names with nice shades of meaning, such as osmon-
dite and troosto-sorbite. Those who look only at the general outlines and not at the
details have no right to forbid the workers in detail from having and using words
fitting their work; nor have those whose needs are satisfied by the three primary
colors a right to forbid painters, dyers, weavers, and others from naming the many
shades with which they are concerned. Like the lexicographer we must seive the
reader by explaining those words which he will meet, whether we individually use or
condemn them. We feel that we have exhausted our powers in cautioning writers
that certain words are rare and not likely to be understood by most readers, or are
improper for any reason, and in urging the complete abandonment of those with-
drawn by their proposers.

Needless words will die a natural death; needed ones we cannot kill. The good
we might do in hastening the death of the moribund by omitting them from this re-
port is less than the good we do by teaching their meaning to those who will meet
them in ante-mortem print. These readers have rights. We serve no class, but the
whole.

Illustrations. At the end of the several descriptions the reader is referred to
good illustrations in Osmond and Stead's "Microscopic Analysis of Metals," Griffin
& Co., London, 1904.

II. LIST OF MICROSCOPIC SUBSTANCES

The microscopic substances here described consist of

1. Meiarals, true phases, like the minerals of nature. These are either elements,
definite chemical compounds, or solid solutions and hence consisting of definite sub-
stances in varying proportions. These include austenite, ferrite, cementite, and
graphite.

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS 3

2. Aggregates, like the petrographic entities as distinguished from the true minerals.
These mixtures may he in definite proportions, i.e. eutectic, or eutectoid mixtures
(ledeburite, pearlite, steadite), or in indefinite proportions (troostite, sorbite). Those
aggregates which are important for any reason are here described.

(Many true minerals, such as mica, felspar, and hornblende, are divisible into
several different species so that these true mineral names may be either generic or
specific. These genera and species are definite chemical compounds in which one
element may replace another. Other minerals, such as obsidian, are solid solutions
in varying proportions, and in these also one element may replace another. Metarals
like minerals differ from aggregates in being severally chemically homogeneous.)

These two classes may be cross classified into:

(A) The iron-carbon series, which come into being in cooling and heating.

(B) The important impurities, manganese sulphide, ferrous sulphide, slag, etc.

The most prominent members of the iron-carbon series are :

I. molten iron, metaral, molten solution, but hardly a microscopic constituent;

II. the components which form in its solidification:

(a) austenite, solid solution of carbon or iron carbide in iron, metaral,
(6) cementite, definite metaral, Fe 3 C,
(c) graphite, definite metaral, C;

III. the transition substances which form through the transformation of austenite
during cooling:

(W) martensite, metaral of variable constitution; its nature is in dispute;
(c) troostite, indefinite aggregate, uncoagulated mixture,

(/) sorbite, indefinite aggregate, chiefly uncoagulated pearlite plus ferrite or ce-
mentite ;

IV. products 1 of the transformation of austenite:
(g) ferrite,

(/) pearlite.

This transformation may also yield cementite and graphite as end products in
addition to those under b and c.

In addition to the above, the names of which are universally recognized and in
general use, the following names have been used more or less:

(i) ledeburite (Wiist), definite aggregate, the austenite-cementite eutectic;

(j) ferronite (Benedicks), hypothetical definite metaral, /3 iron containing about
0.27 per cent of carbon;

(/,) steadite (Sauveur), definite aggregate, the iron-phosphorus eutectic (rare);

and three transition stages in the transformation of austenite, viz. :

(/) hardenite (Arnold), collective name for the austenite and martensite of eutec-
toid composition;

(m) osmondite (Heyn), boundary stage between troostite and sorbite;

(n) troosto-sorbite (Kourbatoff) , indefinite aggregate, the troostite and the sorbite
which lie near the boundary which separates these two aggregates (obsolescent).

1 In hypo-cut ectoid steels these habitually play the part of end products, though according to
the belief of most the true end of the transformation is not reached till the whole has changed into
a conglomerate of ferrite plus graphite.

4 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

III. DEFINITIONS AND DESCRIPTIONS

Carbon-iron Equilibrium Diagram, Figure 1. Under the several substances about
to be described an indication will be given of the parts of the carbon-iron equilibrium
diagram Figure I to which they severally correspond.

Austenite, Osmond (Fr. Austenite, Ger. Austenit, called also mixed crystals and
gamma iron. Up to the year 1900 often called martensite and wrongly sometimes
still so called). Metaral of variable composition.

Definition. The iron-carbon solid solution as it exists above the transformation

1500

KtOO-

1300-

1200-

1100-

21000-

900-

800-
M

700-
600-

500-

1.
Molten Iron

(Per Fondu)

Molben Iron
(Per Fondu)

flusbenite + Cementibe

a A.

oc-Ferrite

Pearl ite

8.B.

Cementibe

Pear I i be

1 2 3 4 5

Carbon per cent

Fig. 1. A,: The line PSK is often called "A,". A 3 : The line COS is often called "A 3 '
this name is sometimes applied to the line SE.

and

range or as preserved with but moderate transformation at lower temperatures, e.g.
by rapid cooling, or by the presence of retarding elements (Mn, Ni, etc.), as in 12 per
cent manganese steel and 25 per cent nickel steel.

Constitution and Composition. A solid solution of carbon or iron carbide (prob-
ably Fe 3 C) and gamma iron, normally stable only above the line PSK of the carbon-
iron diagram. It may have any carbon content up to saturation as shown by the line
SE, viz.: about 0.90 per cent at S (about 725 deg. C.) to 1.7 per cent at E (about
1130 deg.). The theory that the iron and the carbide or carbon, instead of being dis-
solved in each other, are dissolved in a third substance, the solution of eutectoid com-

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS 5

position (Fc 24 C, called hardenite) is not in accord with the generally accepted theory
of the constitution of solutions, and is not entertained widely or by any member of
this committee.

Crystallization. Isometric. The idiomorphic vug crystals are octahedra much
elongated by parallel growth. The etched sections show much twinning. (Osmond
and most authorities.) Le Chatelier believes it to be rhombohedral. Cleavage
octahedral.

Varieties and Genesis. (l) Primary austenite formed in the solidification of carbon
steel and hypo-eutectic cast iron; (2) eutectic austenite, interstratified with eutectic
cementite, making up the eutectic formed at the end of the solidification of steel con-
taining more than about 1.7 per cent of carbon, and of all cast iron.

Equilibrium. It is normal and in equilibrium in Region 4, and also associated
with beta iron in Region 6, with a iron in Region 7, and with cementite in Region 5.
It should normally transform into pearlite with either ferrite or cementite on cooling
past AI into Region 8.

Transformation. In cooling slowly through the transformation range, Ar 3 - Ari,
austenite shifts its carbon content spontaneously through generating pro-eutectoid
ferrite or cementite, to the eutectoid ratio, about 0.90 per cent, and then transforms
with increase of volume at Ari into pearlite. q.v., with which the ejected ferrite or
cementite remains mixed. Rapid cooling and the presence of carbon, manganese, and
nickel obstruct this transformation, (l) retarding it, and (2) lowering the temperature at
which it actually occurs, and in addition (3) manganese and nickel lower the temperature
at which in equilibrium it is due. Hence, by combining these four obstructing agents
in proper proportions the transformation may be arrested at any of the intermediate
stages, martensite, troostite, or sorbite, 1 q.v., and if arrested in an earlier stage it
can be brought to any later desired stage by a regulated reheating or "tempering."
For instance, though a very rapid cooling in the absence of the three obstructing ele-
ments checks the transformation but little and only temporarily, yet if aided by the
presence of a little carbon it arrests the transformation wholly in the martensite
stage; and in the presence of about 1.50 per cent of carbon such cooling retains about
half the austenite so little altered that it is "considerably" softer than the usually
darker needles of the surrounding martensite, with which it contrasts sharply. Again,
either (a) about 12 per cent of manganese plus 1 per cent of carbon, or (6) 25 per cent
of nickel, lower and obstruct the transformation to such a degree that austenite per-
sists in the cold apparently unaltered, even through a slow cooling. (Hadfield's man-
ganese steel and 25 per cent nickel steel, manganiferous and nickeliferous austenite
respectively.)

Occurrence. When alone (12 per cent manganese and 25 per cent nickel steel
and Maurer's 2 per cent carbon plus 2 per cent manganese austenite) polyhedra, often
coarse, much twinned at least in the presence of martensite, and readily developing
slip bands. In hardened high-carbon steel it forms a ground mass pierced by zigzag
needles and lances of martensite.

Etching. All the common reagents darken it much more than cementite, less

1 Though the transformation can be arrested in such a way as to leave the whole of the steel
in the condition of martensite, it is doubted by some whether it can be so arrested as to leave the
whole of it in any of the other transition stages. Troostite and sorbite caused by such arrest are
habitually mixed, troostite with martensite or sorbite or both, and sorbite with pearlite or troostite
or both.

6 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

than troostite or sorbite, and usually less, though sometimes more, than martcnsite,
which is recognized by its zigzag shape and needle structure. With ferrite and pearlite
it is never associated.

Physical Properties. Maurer's austenite of 2 per cent manganese plus 2 per
cent carbon is but little harder than soft iron, and 25 per cent nickel steel and Had-
field's manganese steel are but moderately hard. Yet as usually preserved in hardened
high carbon steel, the hardness of austenite does not fall very far short of that of the
accompanying martensite, probably because partly transformed in cooling. (Os-
mond's words are that it is "considerably" softer than that martensite.)

Specific Magnetism. Very slight unless perhaps in intense fields. In Hadfield's
manganese steel and 25 per cent nickel steel, very ductile.

Illustrations. "Microscopic Analysis of Metals," Figures 20, 50, and 51 on
pp. 39, 100, and 101.

Cementite (Sorby, "intensely hard compound"; Ger. Cementit, Fr. Cementite;
Arnold, crystallized normal carbide). Definite mctaral.

Definition. Tri-ferrous carbide, FeaC. The name is extended by some writers
so as to include tri-carbides in which part of the iron is replaced by manganese or
other elements. Such carbides may be called " manganiferous cementite," etc.

Occurrence. (a) Pearlitic as a component of pearlite, q.v.; (b) eutectic;
(c) primary or pro-eutectic; (d) pro-eutectoid; (e) that liberated by the splitting up of
the eutectic or of pearlite; and (/) uncoagulated in sorbite, troosite, and perhaps mar-
tensite. (c), (d), and (e) are grouped together as "free" or "massive."

Primary cementite is generated in cooling through Region 3; eutectic cementite
on cooling past the line EBD; pro-eutectoid cementite in cooling through Region 5;
pearlitic cementite on cooling past the line PSK, or AI. Though the several varieties
of cementite are generally held to be all metastable, tending to break up into graphite
plus either austenite above AI or ferrite below AI, yet they have a considerable and
often great degree of persistence. The graphitizing tendency is completely checked
in the cold but increases with the temperature and with the proportion of carbon
and of silicon present, and is opposed by the presence of manganese.

Crystallization. Orthorhombic, in plates.

Structure. (a) Pearlitic, in parallel unintersecting plates alternating with plates
of ferrite; (b) eutectic, plates forming a network filled with a fine conglomerate
of pearlite with or without pro-eutectoid cementite; (c) primary, in manganiferous
white cast iron, etc., in rhombohedral plates; (d) in hyper-eutectoid steel, pro-eutec-
toid cementite forms primarily a network enclosing meshes of pearlite through which
cementite plates or spines sometimes shoot if the network is coarse; (e) cementite
liberated from pearlite merges with any neighboring cementite; (/) the structure of
uncoagulated cementite cannot be made out. On long heating the pro-eutectoid
and pearlitic cementite spheroidize slowly, and neighboring particles merge; (o) in
white irons rich in phosphorus in flat plates embedded in iron-carbon-phosphorus
eutectic.

Etching, etc. After polishing stands in relief. Brilliant white after etching with
dilute hydrochloric or picric acid; darkened by boiling with solution of sodium picrate
in excess of sodium hydrate.

Physical Properties. Hardest component of steel. Hardness = 6 of Mohs scale.
Scratches glass and felspar but not quartz; very brittle. Specific magnetism about
two thirds that of pure iron.

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS 7

Illustrations. "Microscopic Analysis of Metals," Figures 42 and 43 on
pp. 84, 85.

Martensite (Fr. Martensite, Ger. Martensit). Metaral. Its nature is in dispute.

Definition. The early stage in the transformation of austenite characterized by
needle structure and great hardness, as in hardened high-carbon steel.

Constitution. I. (Osmond and others.) A solid solution like austenite, q.v., ex-
cept that the iron is partly beta, whence its hardness, and partly alpha, whence its
magnetism in mild fields. II. (Le Chatelier.) The same except that its iron is essen-
tially alpha, and the hardness due to the state of solid solution. III. (Arnold.) A spe-
cial structural condition of his "hardenite" (austenite); not \videlyjield. IV. A solid
solution in gamma iron. V. (Benedicks.) The same as I, except that the iron is wholly
beta and that beta iron consists of alpha iron containing a definite quantity of gamma
iron in solution.

Equilibrium. It is not in equilibrium in any part of the diagram, but represents
a metastable condition in which the metal is caught during rapid cooling, in transit
between the austenite condition stable above the line AI and the condition of ferrite
plus cementite into which the steel habitually passes on cooling slowly past the
line AI.

Occurrence. The chief constituent of hardened carbon tool steels, and of medium
nickel and manganese steels. In still less fully transformed steels (1.50 per cent
carbon steel rapidly quenched, etc.) it is associated with austenite; in more fully
transformed ones (lower carbon steels hardened, high carbon steels oil hardened, or
water hardened and slightly tempered, or hardened thick pieces even of high carbon
steel) it is associated with troostite, and with some pro-eutectoid ferrite or cementite,
q.v., in hypo- and hyper-eutectoid steels respectively. In tempering it first changes
into troostite; at 350 deg. -400 deg. it passes through the stage of osmondite; at
higher temperatures it changes into sorbite; and at 700 deg. into granular pearlite.
On lipating into the transformation range this changes into austenite, which on cool-
ing again yields lamellar pearlite.

Characteristic specimens are had by quenching bars 1 cm. square of eutectoid
steel, i.e. steel containing about 0.9 per cent of carbon, in cold water from 800 deg. C.
(1472 deg. F.).

Structure. When alone, habitually in flat plates made up of intersecting needles
parallel to the sides of a triangle. When mixed with austenite, zigzag needles, lances,
and shafts.

If produced by quenching after heating to 735 deg. C., it consists of minute crystal-
lites resembling the globulites of Vogelsang, which are rarely arranged in triangular
order. At times so fine as to suggest being amorphous.

Etching. With picric acid, iodine or very dilute nitric acid etches usually darker
than austenite, but sometimes lighter, always darker than ferrite and cementite, but
always lighter than troostite.

Illustrations. "Microscopic Analysis of Metals," Figure 19 on p. 38, Figure 52
on p. 102.

Ferrite (Fr. Ferrite, Ger. Ferrit). Definite metaral.

Definition. Free alpha iron.

Composition. Nearly pure iron. It may contain a little phosphorus and
silicon but its carbon content, if any, is always small, at the most not more than 0.05
per cent, and perhaps never as much as 0.02 per cent.

8 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

Occurrence. (a) Pearlitic as a component of pearlite, q.v.; (6) pro-eutectoid
ferrite generated in slow cooling through the transformation range; (c) that segre-
gated from pearlite, i.e. set free by the splitting up of pearlite, especially in low car-
bon steel; (d) uncoagulated as in sorbite, and probably troostite. (6) and (c) are
classed together as free or massive.

Thus ferrite is normal and stable in regions 7 and 8.

Crystallization. Isometric, in cubes or octahedra.

Structure. (a) Pearlitic ferrite, unintersecting parallel plates alternating with
plates of cementite; (6) pro-eutectoid ferrite in low carbon steel forms irregular poly-
gons, each with uniform internal orientation. In higher carbon steel after moderately
slow cooling, especially in the presence of manganese, it forms a network enclosing
meshes of pearlite. In slower cooling this network is replaced by irregular grains
separated by pearlite; (c) the ferrite set free by the splitting up of pearlite merges
with the pro-eutectoid ferrite, if any; (d) the structure of the ferrite in sorbite, etc.,
cannot be made out.

Etching. Dilute alcoholic nitric or picric acid on light etching leaves the
ferrite grains white with junctions which look dark. Deeper etching, by Heyn's
reagent or its equivalent, reveals the different orientation of the crystals or. grains,
(a) as square figures parallel to the direction of the etched surface, (6) as plates which
dip at varying angles and become dark or bright when the specimen is rotated under
oblique illumination. Still deeper etching reveals the component cubes (etching
figures, Atzfiguren), at least if the surface is nearly parallel to the cube faces.

Physical Properties. Soft; relatively weak (tenacity about 40,000 Ibs., per
sq. in.); very ductile; strongly f erro-magnetic ; coercitive force very small.

Grain Size. For important purposes (1) etch deeply enough, e.g. with copper-
ammonium' chloride, to reveal clearly the junctions of the grains; (2) count on a photo-
graph of small magnification the number of grains in a measured field so drawn as to
exclude fragments of grains; after (3) determining the true grain boundaries by ex-
amination under high powers (Heyn's method). Deep nitric acid etching is inaccurate,
because an apparent grain boundary may contain several grains.

Illustrations. "Microscopic Analysis of Metals," Figures 41, 56 on pp. 79, 116.

Osmondite (Fr. Osmondite, Ger. Osmondit).

Definition. That stage in the transformation of austenite at which the solubility
in dilute sulphuric acid reaches its maximum rapidity. Arbitrarily taken as the
boundary between troostite and sorbite.

Earlier Definition. Defined by the V th Congress as having the "maximum sol-
ubility in acids and by a maximum coloration under the action of acid metallographic
reagents." The present definition is confined to maximum rapidity of dissolving,
because we do not yet know that this in all cases co-exists with the maximum depth
of coloration, and in any case in which these two should not co-exist, the old defini-
tion does not decide which is true osmondite.

Constitution. The following hypotheses have been suggested, none of which
has firm experimental foundation: (1) A solid solution of carbon or an iron carbide
in alpha iron. (2) The colloidal system of Benedicks in its purity, troostite being this
system while forming at the expense of martensite, and sorbite, being this system
coagulating and passing into pearlite. (3) The stage of maximum purity of amor-
phous alpha iron on the way to crystallizing into ferrite.

Occurrence. Hardened carbon steel of about 1 per cent of carbon when reheated

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS 9

(tempered) to 350^00 deg. C. passes through the stage of troostite to that of
osmondite, and on higher heating to that of sorbite. What variation if any from this
temperature is needed to bring hardened steel of other carbon content to the osmond-
ite stage is not known. In that it represents a true boundary state between troostite
and sorbite it differs in meaning from troosto-sorbite, which embraces both the troost-
ite and the sorbite which lie near this boundary. Indeed osmondite has sometimes
been used in this looser sense. Writers are cautioned that, however useful these terms
may prove for making these nice discriminations, they are not likely to be familiar
to general readers.

Etching. According to Heyn it differs from troostite and sorbite in being that
stage in tempering which colors darkest on etching with alcoholtc^iydrochloric acid.

The present definition and description of osmondite should displace previous ones,
because they have the express approval of Professor Heyn, the proposer of the name,
and M. Osmond himself.

Ferronite (Fr. Ferronite, Ger. Ferronit) (Benedicks). Hypothetical definite metaral.

Definition. Solid solution of about 0.27 per cent of carbon in beta iron.

Occurrence (hypothetical) . In slowly cooled steels and cast iron containing
0.50 per cent of combined carbon or more, that which is generally believed to be fer-
rite, whether pearlitic or free, is supposed by Benedicks to be ferronite.

Hardenite (Fr. Hardenite, Ger. Hardenit).

Definition. Collective name for austenite and martensite of eutectoid composi-
tion. It includes such steel (1) when above the transformation range, and (2) when
hardened by rapid cooling.

Observations. On the generally accepted theory that austenite is a solid solution
of carbon or an iron carbide in iron, hardenite is the solution of the lowest transforma-
tion temperature, i.e. the eutectoid. The theory that instead it is a definite chemical
compound, Fe 2 ,iC, is considered under Austenite. Its proposer includes under
hardenite both eutectoid (0.90 per cent carbon) austenite when above the transforma-
tion range and the martensite into which that austenite shifts in rapid cooling (hard-
ening) .

Other Meanings. Originally (Howe, 1888) collective name for aastenite and
martensite of any composition in carbon steel. Osmond (1897), austenite saturated
with carbon. Both these meanings are withdrawn by their proposers.

Pearlite (Sorby's "pearly constituent." At first written "pearlyte" Fr. Perlite,
(!er. Perlit). Aggregate.

Definition. The iron-carbon eutectoid, consisting of alternate masses of ferrite
and cementite.

Constitution and Composition. A conglomerate of about 6 parts of ferrite to 1 of
cementite. When pure, contains about 0.90 per cent of carbon, 99.10 per cent of
iron.

Occurrence. Results from the completion of the transformation of austenite
brought spontaneously to the eutectoid carbon content, and hence occurs in all
carbon steels and cast iron containing combined carbon and cooled slowly through
the transformation range, or held at temperatures in or but slightly below that range,
long enough to enable the ferrite and cementite to coagulate into a mass microscopic-
ally resoluble. Hence it is the normal constituent in Region 8. Its ferrite is stable
but its cementite is metastable and tends to transform into ferrite and graphite.

Varieties and Structure. Because pearlite is formed by the coagulation of the

10 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

ferrite and cementite initially formed as the irresoluble emulsion, sorbite, (Arnold's
sorbitic pearlite) there are the indefinitely bounded stages of sorbitic pearlite (Arnold's
normal pearlite), i.e. barely resoluble pearlite, in the border-land between sorbite and
laminated pearlite; granular pearlite, in which the cementite forms fine globules in a
matrix of ferrite; and laminated or lamellar pearlite, consisting of fine, clearly defined,
non-intersecting, parallel lamellae alternately of ferrite and cementite. The name
granular pearlite was first used by Sauveur to represent what is now called sorbite.
This meaning has been withdrawn.

An objection to Arnold's name "normal pearlite" is that it is likely to mislead.
"Normal" here apparently refers to arising under normal conditions of cooling, but
(1) it rather suggests structure normal for pearlite, which surely is the lamination
characteristic of eutectics in general, and (2) the general reader has no clue as to what
conditions of cooling are here called normal. Many readers are not manufacturers,
and even in manufacture itself air cooling is normal for one branch and extremely
slow furnace cooling for another. Arnold calls troostite "troostitic pearlite" and
sorbite "sorbitic pearlite." This is contrary to general usage, which restricts pearlite
to microscopically resoluble masses.

Etching. After etching with dilute alcoholic nitric or picric acid it is darker than
ferrite or cementite but lighter than sorbite and troostite. A magnification of at
least 250 diameters is usually needed for resolving it into its lamellae, though the
pearlite of blister steel can often be resolved with a magnification of 25 diameters.
The more rapidly pearlite is formed, the higher the magnification needed for re-
solving it.

Illustrations. Lamellar pearlite. Osmond and Stead, "Microscopic Analysis,"
Figure 11 on p. 19, Granular pearlite, idem, Figure 18 on p. 36; Heyn and Bauer,
"Stahl und Eisen," 1906, Figure 14, opposite p. 785.

Graphite (Ger. Graphit, Fr. Graphite). Definite metaral.

Definition. The free elemental carbon which occurs in iron and steel.

Composition. Probably pure carbon, identical with native graphite.

Genesis. Derived in large part, and according to Gosrens wholly, from the de-
composition of solid cementite. Others hold that its formation as kish may be from
solution in the molten metal, and that part of the formation of temper graphite may
be from elemental carbon dissolved in austenite. It is the stable form of carbon in
all parts of the diagram.

Occurrence. (1) as kish, flakes which rise to the surface of molten cast iron and
usually escape thence;

(2) as thin plates, usually curved, e.g. in gray cast iron, representing carbon which
has separated during great mobility, i.e. near the melting range;

(3) as temper graphite (Ger. Temperkohle, Ledebur) pulverulent carbon which
separates from cementite and austenite, especially in the annealing process for mak-
ing malleablized castings.

Graphite and ferrite are sometimes associated in a way which suggests strongly
that they represent a graphite-austenite eutectic. But the existence of such a true
eutectic is doubted by most writers.

Properties. Hexagonal. H. 1-2. Gr. 2.255. Streak black and shining, luster
metallic; macroscopic color, iron black to dark steel gray, but always black when seen
in polished sections of iron or steel under the microscope; opaque; sectile; soils paper;
flexible; feel, greasy.

APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS 11

Troostite (Fr. Troostite, Ger. Trcostit). Probably agrregate. (Arnold, troostitic
pearlite.)

Definition. In the transformation of austenite, the stage following martensite
and preceding sorbite (and osmondite if this stage is recognized).

Constitution and Composition. An uncoagulatecl conglomerate of the transition
stages. The degree of completeness of the transformation represented by it is not
definitely known and probably varies widely. Osmond and most others believe that
the transformation, while generally far advanced, yet falls materially short of comple-
tion; but Benedicks and Arnold (9) believe that it is complete. The former belief
that it is a definite phase, e.g. a solid solution of carbon or an iron carbide in either /3
or 7 iron, is abandoned. Its carbon content like that of austeTiitc and martensite
varies widely.

Occurrence. It arises either on reheating hardened (e.g. martensitic steel) to
slightly below 400 deg., or on cooling through the transformation range at an inter-
mediate rate, e.g. in small pieces of steel when quenched in oil, or quenched in water
from the middle of the transformation range, or in the middle of larger pieces quenched
in water from above the transformation range. With slightly farther reheating it
changes into sorbite; with higher heating into sorbitic pearlite, then slowly into granular
pearlite, and probably indirectly into lamellar pearlite. It occurs in irregular, fine-
granular or almost amorphous areas, colored darker by the common etching reagents
than the martensite or sorbite accompanying it. A further common means of dis-
tinguishing it from sorbite is that it is habitually associated with martensite, whereas
sorbite is habitually associated with pearlite.

Areas near the boundary between troostite and sorbite are sometimes called
troosto-sorbite.

Properties. Hardness, intermediate between that of the martensitic and the
pearlitic state corresponding to the carbon content of the specimen. In general the
hardness increases, the elastic limit rises, and the ductility decreases, as the carbon
content increases. Its ductility is increased rapidly and its hardness and elastic limit
lowered rapidly by further tempering, which affects it much more markedly than
sorbite.

Sorbite (Fr. Sorbite, Ger. Sorbit). Aggregate. (Arnold, sorbitic pearlite.)

Definition. In the transformation of austenite, the stage following troostite
and osmondite if the stage is recognized, and preceding pearlite.

Constitution and Composition. Most writers believe that it is essentially an un-
coagulated conglomerate of irresoluble pearlite with ferrite in hypo- and cementite
in hyper-eutectoid steels respectively, but that it often contains some incompletely
transformed matter.

Occurrence. The transformation can be brought to the sorbitic stage (1) by re-
heating hardened steel to a little above 400 deg., but not to 700 deg. at which tem-
perature it coagulates into granular pearlite; (2) by quenching small pieces of steel
in oil or molten lead or even by air cooling them; (3) by quenching in water from just
above the bottom of the transformation range, Ari. Sorbite is ill-defined, almost amor-
phous, and is colored lighter than troostite but darker than pearlite by the usual
etching reagents. It differs further from troostite in being softer for given carbon
content, and usually in being associated with pearlite instead of martensite, and
from pearlite in being irresoluble into separate particles of ferrite and cementite.

As sorbite is essentially a mode of aggregation it cannot properly be represented

12 APPENDIX II NOMENCLATURE OF THE MICROSCOPIC CONSTITUENTS

on the equilibrium diagram. Its components at all times tend to coagulate into
pearlite, yet it remains in its uncoagulated state at all temperatures below 400 dog.

Properties. Though slightly less ductile than pearlitic steel for given carbon
content, its tenacity and elastic limit are so high that a higher combination of these
three properties can be had in sorbitic than in pearlitic steels by selecting a carbon
content slightly lower than would be used for a pearlitic steel. Hence the use of
sorbitic steels, e.g. first hardened and then annealed cautiously, for structural pur-
poses needing the best quality.

Manganese Sulphide (Fr. Sulphur de Manganese, Ger. Schwefelmangan), MnS
(Arnold and Waterhouse). Metaral.

Occurrence, etc. Sulphur combines with the manganese present in preference to
the iron, forming pale dove or slate gray masses, rounded in castings, elongated in
forgings.

Ferrous Sulphide (Fr. Sulphure de Fer, Ger. Schwefeleisen), FeS. Metaral.

Occurrence. The sulphur not taken up by the manganese forms ferrous sulphide,
FeS, which, probably associated in part with iron as an Fe-FeS eutectic, forms by
preference more or less continuous membranes surrounding the grains of pearlite.
Color, yellow or pale brown.

Sulphur Prints. When silk impregnated with mercuric chloride and hydrochloric
acid (Heyn's and Bauer's method) or bromide paper moistened with sulphuric acid
(Baumann's method) is pressed on polished steel, the position of the sulphur-bearing
areas, whether of FeS or MnS, records itself by the local blackening which the evolved
H 2 S causes. Phosphorus bearing areas also blacken Baumann's bromide paper.

MISCELLANEOUS

Eutectoid, Saturated, etc. The iron-carbon eutectoid is pearlite. Steel with more
carbon than pearlite is called hyper-eutectoid, that with less is called hypo-eutectoid.
Arnold's names "saturated," "unsaturated," and "supersaturated," for eutectoid,
hypo-eutectoid, and hyper-eutectoid steel respectively, have considerable industrial
use in English-speaking countries, but are avoided by most scientific writers on the
ground that they are misleading, because, e.g. there is only one specific temperature,
AI, at which eutectoid steel is actually saturated, and, if any other temperature is in
mind, that steel is not saturated. Above AI it is clearly undersaturated.

The objection to the names sorbite, troostite, martensite, and austenite, that
each of them covers steel of a wide range of carbon content, is to be dismissed because
a like objection applies with equal force to every generic name in existence.

The theoretical matter in this report is given solely for exposition and the com-
mittee disclaims the intent to impose any theory. This report is offered for adop-
tion subject to this disclaimer on the ground that the adoption of theories is beyond
the powers of a Congress.

INDEX

The Roman numerals refer to the numbers of the lessons, the letter A to the chapter
on " Apparatus for the Metallographic Laboratory."

A, AT, Ac, Ar 3 , Ac 3 , Ar 3 . 2 , Ac 3 . 2 , Ar 3 . 2 .i, Ac 3 . 2 .i, Ar cm , Ac cm . See critical

points, notation

Allotrimorphic crystals, definition of, I, 2
Allotropic theory of the hardening of steel, XV, 2
Allotropy, definition of, II, 4

of cementite, VIII, 9
iron, II, 4; VIII, 1, 14
sulphur, II, 6

Alloy steels. See special steels
Alloys, constitution of, XXII, 1 to 21

, fusibility curves of, XXII, 5 to 21
, microstructure of, XXII, 5 to 21

of iron and carbon, equilibrium diagram of, XXIII, 12 to 21
, fusibility curves of, XXIII, 1 to 21
, phase rule applied to, XXIV, 6 to 8
, structural composition immediately after soli-
dification of, XXIII, 3
, phase rule applied to, XXIV, 3 to 8
, solidification of, XXII, 3 to 21
, structural composition of, XXII, 15 to 20
whose component metals form solid solutions, solidification and

constitution of, XXII, 3 to 9
are insoluble in each other in the solid
state, solidification and constitution
of, XXII, 9 to 17
partially soluble in each other in
the solid state, solidification and
constitution of, XXII, 17 to 21
Alpha iron, VIII, 1, 14; IX, 8

, crystallization of, II, 7
, description of, II, 6

theory of the hardening of steel, XV, 4

Alumina powder for polishing, preparation of, Appendix I, 2
Ammonium oxalate etching, V, 7
Anhedrons. See allotrimorphic crystals
Annealing, air cooling in, XII, 4
, cooling in, XII, 2
, double treatment in, XII, 8

for malleablizing cast iron, XXI, 3 to 7
, furnace cooling in, XII, 4
, heating for, XII, 1

, influence of maximum temperature in, XII, 5
, influence of time at maximum temperature in, XII, 6
, nature of operation, XII, 1
1

INDEX

Annealing of steel, XII, 1 to 30

, oil and water quenching in, XII, 6
, purpose of, XII, 1

, rate of cooling vs. carbon content in, XII, 3
size of objects in, XII, 3
steel castings, XII, 13
temperatures for steel, XII, 2
Arnold on the hardening of steel, XV, 1, 4
Arnold's view of the nature of martensite, XIII, 10

troostite, XIII, 12
Austenite, crystallization of, X, 1 to 10

, definition, description, occurrence, and structure of, XIII, 3 to 9
, growth above the critical range of, XII, 20
, Osmond's test showing relative softness of, XIII, 7
, production by Maurer of, XIII, 5
Osmond of, XIII, 4
Robin of, XIII, 5
, relative softness of, XIII, 7
, saturated, XXIII, 2

Austenitic and pearlitic structures, relation between, XII, 21
special steels, XVII, 7
steel, tempering of, XIV, 3

Belaiew on the structure of steel and of meteorites, X, 6 to 10
Benedicks' equilibrium diagram of iron-carbon alloys, XXIII, 19

view of the nature of troostite, XIII, 12
Beta iron, VIII, 1, 11, 14; IX, 8
, crystallization of, II, 7
, description of, II, 6

theory of the hardening of steel, XV, 2
Binary alloys. See alloys.
Bivariant equilibrium, definition of, XXIV, 3
Black heart castings, XXI, 4

, annealing for, XXI, 5
Brass, twinnings in, II, 7
Brittleness, intercrystalline, XII, 27
, intergranular, XII, 27

of low carbon steel, XII, 26
Burnt steel, production and structure of, XII, 17 to 20

Cameras, A, 22 to 28
Carbide steel, XVII, 1

Carbon, condition of, in hardened and tempered steel, XIV, 8
, hardening and combined in steel, XIV, 8
in pearlite, V, 8
in steel, IV, 3
temper, XXI, 1

theory of the hardening of steel, XV, 1, 4
Carpenter and Heeling's cooling curves of steels, VII, 17

determinations of the critical points, VII, 8, 9
equilibrium diagram of iron-carbon alloys, XXIII, 18, 19
Case hardened articles, heat treatment of, XVI, 6

steel, tempering of, XVI, 6
hardening, composition of iron or steel subjected to, XVI, 1

INDEX

Case hardening, cooling after, XVI, 5

, distribution of carbon after, XVI, 2

, duration of, XVI, 2

, materials used for, XVI, 3

, mechansim of, XVI, 5

of steel, XVI, 1 to 6
, temperatures for, XVI, 1

Cast iron, calculation of structural composition of, XIX, 5, 10 to 13; XX, 7 to 9
, chilled castings of, XIX, 13

, constitution, properties, and structure of, XIX, 1 to 13; XX, 1 to 10
containing only combined carbon, XIX, 3
graphitic carbon, XIX, 1

, formation of combined and graphitic carbon in, XIX, 1
, impurities in, XX, 1 to 10

, influence and occurrence of manganese in, XX, 2

phosphorus in, XX, 2
silicon in, XX, 1
sulphur in, XX, 1
, malleable, XXI, 1 to 8

, structural composition vs. physical properties of, XIX, 11
steel, structure of, X, 1 to 10
Castings suitable for malleablizing, XXI, 2
Cement carbon, definition of, XIV, 8
Cementation. See case hardening

of iron and steel, XVI, 1 to 6
Cementite, allotropy of, VIII, 9

, definition and description of, IV, 5

, etching of, V, 7; XIX, 4

, formation of, X, 4

, free, definition of, V, 5

, graphitizing of, XII, 15; XXI, 1; XXIII, 7

in high carbon steel, V, 4
, primary. See cementite, pro-eutectic
, pro-eutectic, XXIII, 5
, spheroidizing of, XII, 14
Cementitic special steels, XVII, 1, 8
Charpy and Grenet on the equilibrium diagram of iron-carbon alloys, XXIII, 20

on the hardening of steel, XV, 1, 5
Chilled castings, XIX, 13
Chrome-nickel steel, XVIII, 16
steel, XVIII, 13 to 1,5

, uses and properties of, XVIII, 14
-tungsten steel. See high-speed steel.

Chromium, influence on critical points of iron of, XVIII, 13
Cleavage, definition of, I, 1

brittlcness. See intercrystalline brittleness, XII, 27
Cold working, crystalline growth after, I, 8

, influence on structure and properties of steel of, XI, 8
Colloidal solution, XIII, 12
Combined carbon in cast iron, XIX, 1, 3
Components, Bancroft's definition of, XXIV, 3
, definition of, XXIV, 3
, Findlay's definition of, XXIV, 3
, Howe's definition of, XXIV, 3
, Mellor's definition of, XXIV, 3
Condensers, A, 21

Cooling and heating curves of iron and steel, VII, 10 to 19
curves of pure metals, XXII, 1

INDEX

Copper, microstructure of, I, 1
Critical points and crystallization, IX, 2, 5
dilatation, IX, 1 to 3
electrical conductivity, IX, 1
magnetic properties, IX, 3
in high carbon (hypcr-cutcctoid) steel, VII, 7
iron, description of, II, 6
medium high carbon steel, VII, 6; VIII, 4
pure iron, VII, 5, 10; VIII, 1
very low carbon steel, VII, 6; VIII, 3
, Carpenter and Reeling's determination of, VII, 8
, causes of, VIII, 1 to 16
, definition of, VII, 1
, determination of, VII, 10

, graphical representation of the position and magnitude of, VII, 10
, heat absorbed or evolved at, VII, 8
, influence of chemical composition on position of, VII, 5

speed of heating and cooling on, VII, 4
, instruction for detection of, VII, 19
, merging of, VII, 6, 7, 8
, minor, VII, 8
, notation, VII, 2
, occurrence of, VII, 1 to 20
, relation between structure of steel and, VIII, 12
' , their effects, IX, 1 to 8

, use of neutral bodies in detecting, VII, 14
range. See critical points,
temperatures. See critical points.
Crystalline grains. See grains

growth in metals on annealing, I, 7

of strained ferrite, XII, 23 to 26
Crystallite of iron, II, 5
Crystallites, definition of, I, 2
Crystallization and critical points, IX, 2, 5
, cubic, of metals, I, 4
of austenite, X, 1 to 10

iron, II, 2
, process of, I, 1

Crystallography, systems of, I, 4
Crystals, allotrimorphic, definition of, I, 2
, cubic, of iron, II, 3, 4
, definition of, I, 1
, formation of, I, 1
, idiomorphic, definition of, I, 2
, mixed. See mixed crystals.
Cubic crystallization of iron, II, 2
metals, I, 4

Degrees of freedom, definition of, XXIV, 2
liberty. See degrees of freedom
Desch's types of cooling curves, VII, 18
Dilatation and critical points, IX, 1 to 3
Divariant equilibrium. See bivariant equilibrium
Double annealing treatment, XII, 8
Ductility of steel, structural composition vs., V, 17

INDEX

Edwards on high speed steel, XVIII, 20

the hardening of steel, XV, 1, 2

Edwards' view as to the nature of martensite, XIII, 10
Electric arc lamps, A, 19 to 21

furnaces, A, 35

Electrical conductivity and critical points, IX, 1
Electrolytic iron, microstructure of, II, 1
Electromagnetic stages, A, 11
Equilibrium, bivariant, definition of, XXIV, 3
, definition of, XXIV, 1
diagram. See fusibility curves

of iron-carbon alloys, XXIII, 12 to 21

, Benedicks' diagram, XXIII, 20
, Carpenter and Keeling's diagram,

XXIII, 19

, Roberts-Austen's diagrams, XXIII, 17
, Roozeboom's diagram, XXIII, 17
, Rosenhain's diagram, XXIII, 20
, the author's early diagram, XXIII, 16
, metastable, definition of, XXIV, 2
, stable, definition of, XXIV, 2
, univariant, definition of, XXIV, 3
, unstable, definition of, XXIV, 2
, unvariant, definition of, XXIV, 3
Etching, III, 6; Appendix I, 10

figures. See etching pits
of cementite, V, 7; XIX, 4
pits, formation of, I, 4

in iron, II, 3

with ammonium oxalate, V, 7
nitric acid, III, 7
picric acid, III, 6
sodium picrate, V, 7
Eutectic alloys, I, 5, 6

, constitution and occurrence of, XXII, 12 to 21
, definition of, XXII, 12
, iron-carbon, XXIII, 2
Eutectoid, definition of, IV, 3

steel, definition and structure of, V, 4
Ewing and Rosenhain, straining of iron by, II, 11

Ewing and Rosenhain's theory of crystalline growth of metals on annealing, I, 7
Eye-pieces, A, 3

Ferrite, crystalline growth of, XII, 23 to 26
, definition of, II, 4
, free, IV, 4

in cast iron, XIX, 1 to 10
low carbon steel, IV, 2
wrought iron, III, 1
grains, II, 1

, orientation of, II, 2
Ferro-ferrite, II, 4
Fibers in wrought iron, III, 2
Finishing temperatures, influence on the structure and properties of steel of, XI, 3

INDEX

Free cementite, definition of, V, 5

ferrite, IV, 4
Furnaces, A, 35
Fusibility curves of alloys, XXII, 5 to 21

iron-carbon alloys, XXIII, 1 to 21

Gamma iron, VIII, 1, 14; IX, 8

, crystallization of, II, 7
, description of, II, 6

theory of the hardening of steel, X7, 2
, twinning in, II, 7
Ghost lines in steel, VI, 10
Gold, microstructure of, I, 1
Grading of steel vs. its carbon content, IV, 1
Grain refining treatment, XII, 8
Grains, crystalline orientation of, I, 3
, ferrite, II, 1

, orientation of, II, 2
, growth of, on annealing, I, 7
of metals, definition and formation of, I, 3

, heterogeneousness of, I, 3
Graphitic carbon, factors influencing formation of, XIX, 1

in cast iron, XIX, 1, 2, 3
Graphitizing of cementite, XII, 15; XXI, 1; XXIII, 7

in malleablizing cast iron, XXI, 1
Gray cast iron, XIX, 8

vs. malleable cast iron, XXI, 7
Grenet on the hardening of steel, XV, 1, 5
Guillaume on nickel steel, XVIII, 5
Guillet on case hardening, XVI, 3, 4, 5
chrome steel, XVIII, 14
manganese steel, XVIII, 5
nickel steel, XVIII, 1
silicon steel, XVIII, 15
the hardening of steel, XV, 1, 4
tungsten steel, XVIII, 12
vanadium steel, XVIII, 15, 17
Guillet's theory of special steels, XVII, 1
Gutowsky on the equilibrium diagram of iron-carbon alloys, XXIII, 20

Hadfield steel, XVIII, 10
Hard castings, XXI, 2

Hardened and tempered steel, microstructure of, XIV, 7
Hardening and tempering in one operation, XIII, 20; XIV, 2
carbon, definition of, XIV, 8

theory of the hardening of steel, XV, 4
, cooling for, XIII, 1
, heating for, XIII, 1
of steel, XIII, 1 to 21

, theories of, XV, 1 to 7
, structural changes on, XIII, 2

theories of the hardening of steel, classification of, XV, 1
Hardenite, definition, occurrence, and properties of, XIII, 15
Heat tinting, Appendix I, 11

INDEX

Heat treatment of case hardened articles, XVI, 6
iron, influence of, II, 10
metals, influence of, I, 7

Heating and cooling curves of iron and steel, VII, 10 to 19
Heraeus electric furnace, A, 35

Heyn on decrease of hardness in tempering, XIV, 9
heat liberated on tempering steel, XIV, 9
osinondite, XIV, 6

the condition of carbon in hardened and tempered steel, XIV, 8
equilibrium diagram of iron-carbon alloys, XXIII, 20
structure of hardened and tempered steel, XIV, 7
High-speed steel, XVIII, 17 to 20

, composition of, XVIII, 18
, discovery by Taylor and White of, XVIII, 18
, microstructure of, XVIII, 18
, properties of, XVIII, 17
, theory of, XVIII, 18
, treatment of, XVIII, 17, 18

Hot working, influence on structure and properties of steel of, XI, 1 to 7
Howe on tempering colors, XIV, 1

the burning of steel, XII, 17
hardening of steel, XV, 1

Humfrey and llosenhain. See Rosenhain and Humfrey
Hypcr-eutectoid steel, definition and structure of, V, 4
Hypo-eutectoid steel, definition and structure of, V, 4

Idiomorphic crystals, definition of, I, 2
Illuminating objectives, A, 17
Illumination for microscopical work, A, 14 to 22
, oblique, A, 14 to 16
, vertical, A, 14 to 18
Illuminators, vertical, A, 14, 16 to 18
Impurities in cast iron, XX, 1 to 10

, influence on iron of, II, 10
in metals, influence of, I, 5
steel, VI, 1 to 12

, segregation of, VI, 10
Ingot iron, II, 1
Ingotism, X, n

Intcrcrystalline brittleness, XII, 27
Intergranular brittleness, XII, 27
Invar (nickel steel), XVIII, 5
Inverted microscope, A, 28
Iris diaphragms, A, 7
Iron, affinity for carbon of, XVI, 1
, allotropy of, II, 4; VIII, 1, 14-
, alpha, VIII, 1, 14; IX, 8

, description of, II, 6
, beta, VIII, 1, 11, 14; IX, 8

, description of, II, 6

-carbon alloys, equilibrium diagram of, XXIII, 12 to 21
, fusibility curves of, XXIII, 1 to 21
, phase rule applied to, XXIV, 6 to 8
, structural composition immediately after solidificat ion

of, XXIII, 3
eutectic, XXIII, 2

INDEX

Iron, cementation of, XVI, 1 to 6

-cementite fusibility curve, XXIII, 1

, cooling and heating curves of, VII, 10 to 14

, critical points of, VII, 5, 10; VIII, 1

crystallite, II, 5
, crystallization of, II, 2
, cubic crystals of, II, 3, 4
, electrolytic, microstructure of, II, 1
, etching in hydrogen, II, 10

pits in, II, 3, 4
, gamma, VIII, 1, 14; IX, 8

, description of, II, 6
-graphite fusibility curve, XXIII, 7
, influence of chromium on critical points of, XVIII, 13
heat treatment of, II, 10
impurities on, II, 10
mechanical treatment of, II, 11
nickel on dilatation of, XVIII, 5
tungsten on critical points of, XVIII, 12
, microstructure of, II, 1

oxide in steel, VI, 8
, slip bands in, II, 11
, straining of, II, 10, 11
sulphide in steel, VI, 3
Irreversible steels, XVIII, 2
Isomorphous mixtures, definition of, I, 5

Kourbatoff's etching to color cementite, V, 7
Kroll, etching of pure iron in hydrogen by, II, 9

Le Chatelier, Andre, on the hardening of steel, XV, 1, 5
Le Chatelier on the hardening of steel, XV, 1, 4

thermo-electric pyrometer for the determination of critical

points, VII, 10; A, 30

Le Chatelier's view of the nature of martensite, XIII, 10
Ledebur's temper carbon, XXI, 1
Lieberkiihn, A, 14

Lights for microscopical work, A, 14 to 22
Liquidus, definition of, XXII, 4

Magnetic properties and critical points, IX, 3

specimen holders, A, 9, 11
Malleable cast iron, XXI, 1 to 8

, annealing for the manufacture of, XXI, 3

, packing materials for the manufacture of, XXI, 3

vs. gray cast iron, XXI, 7
castings. See malleable cast iron
Manganese in cast iron, influence and occurrence of, XX, 2

steel, VI, 5
oxide in steel, VI, 8
steel, XVIII, 5 to 12

, austenitic, XVIII, 10
, martensitic, XVIII, 10

INDEX

Manganese steel, pearlitic, XVIII, 8

, properties of austenitic, XVIII, 11
, treatment of austenitic, XVIII, 11
, water-toughening of, XVIII, 1 1
sulphide in steel, VI, 3
Marble, twinnings in, II, 7
Martensite, Arnold's view as to the nature of, XIII, 10

, definition, description, occurrence, properties, etching, and

structure of, XIII, 10

, Edwards' view as to the nature of, XIII, 10
, Le Chateh'er's view as to the nature of, XIII, 10
, Osmond's view as to the nature of, XIII, 10
Martrnsitic special steels, XVII, 7, 9

steel, tempering of, XIV, 5
Matweieff's etching to color cementite, V, 7

method of etching slag in wrought iron, III, 3
Maurer, production of austentite by, XIII, 5
Mechanical refining, XI, 9
stages, A, 3, 12
treatment of iron, influence of, II, 11

steel, XI, 1 to 10

Metalloscope, universal, A, 10 to 13, 28
Metals, cooling curves of, XXII, 1

, crystalline growth on annealing, I, 7

, crystallization of, I, 1

, cubic crystallization of, I, 4

, definition and formation of grains of, I, 3

, influence of heat treatment, I, 7

mechanical treatment of, I, 8
, latent heat of solidification of, XXII, 2
, phase rule applied to, XXIV, 4
, solidification of, XXII, 1
Metallic alloys. See alloys

, constitution of, XXII, 1 to 21
Metarals, definition of, XIII, 18
Metastable equilibrium, definition of, XXIV, 2
Meteorites, microstructure of, X, 6 to 10
Microscopes and accessories, A, 1 to 30; Appendix I, 16 to 30

, inverted, A, 28

Microstructure of cast steel, X, 1 to 10
electrolytic iron, II, 1
hardened and tempered steel, XIV, 7
high carbon steel, V, 4,
sulphur steel, VI, 12
vs. low phosphorus steel, VI, 11
impure gold, I, 6
low carbon steel, IV, 2
medium high carbon steel, V, 1
meteorites, X, 6 to 10
oxidized Bessemer metal, VI, 12
pure copper, I, 1
gold, I, 1
iron, II, 1
metals, I, 1
platinum, I, 3
worked steel, XI, 1 to 10
wrought iron, III, 1, 2
Mixed crystals, definition of, I, 6

10 INDEX

Monovariant equilibrium. See univariant equilibrium

Mottled cast iron, XIX, 10

Mounting samples, Appendix I, 12 to 15

N
Nachet illuminating objectives, A, 17

prism vertical illuminator, A, 17
Nernst lamp, A, 20

Neutral bodies for the detection of critical points, VII, 14
Nickel, influence of, on critical points of iron, XVIII, 2

dilatation of iron, XVIII, 5
steel, XVIII, 1 to 5

, austenitic, XVIII, 5
, case hardening of, XVIII, 4
, critical points of commercial, pearlitic, XVIII, 2
, hardening and annealing of, XVIII, 4
, martensitic, XVIII, 5
, pearlitic, XVIII, 2
, properties of pearlitic, XVIII, 3
Nitric :icid etching, III, 7
Non-variant equilibrium. See unvariant equilibrium

Objectives, A, 3

Oblique illumination, A, 14 to 16

Orientation of crystalline grains, definition of, I, 3

ferrite grains, II, 2
Osmond on the hardening of steel, XV, 1, 2

, production of austenite by, XIII, 4
Osmond's view of the nature of martensite, XIII, 10
Osmondite, definition, description, and occurrence of, XIV, 5
Oxalate of ammonium etching, V, 7

Parabolic reflector, A, 14
Pearlite, carbon content of, V, 8

, definition and description of, IV, 3; VIII, 7
, formation of, X, 1 to 10
in high carbon steel, V, 4
low carbon steel, IV, 3
, varieties of, XII, 15
Pearlitic special steels, XVII, 6, 8
Phase rule applied to alloys, XXIV, 3 to 8

iron-carbon alloys, XXIV. 6 to 8
metals, XXIV, 4
, definition of,' XXIV, 3

, enunciation and explanation of, XXIV, 1 to 4
Phosphorus in cast iron, influence and occurrence of, XX, 2

steel, VI, 2

Photomicrographic cameras, A, 22 to 28
Photography. See photomicrography
Photomicrography, IV, 6
Picrate of sodium etching, V, 7
Picric acid etching, III, 6
Pits. See etching pits
Planes of cleavage. See cleavage
Platinite (nickel steel), XVIII, 5
Platinum, microstructure of, I, 3

INDEX

Point of rccalescence. See recalescence point.
Polishing, III, 4; Appendix I, 1 to 9

machines, A, 28; Appendix I, 4 to 9
Polyhedric special steels, XVII, 1, 7
Polymorphism. See allotropy
Preserving samples, Appendix I, 12
Prism vertical illuminator, A, 16, 17
Pseudomorphism, definition of, XIV, 7
Pure metals, microstructure of, I, 1
, crystallization of, I, 1
Pyrometer, Le Chatelier thermo-electric, for the determination of the

critical points, VII, 10; A, 30
Pyrometers, A, 30

, self-recording, VII, 18; A, 33, 34

Quenching in annealing, XII, 6
Quaternary steels, XVII, 10. See also special steels
vanadium steels, XVIII, 17

Recalescence point, description and occurrence of, VII, 1

Refining, mechanical, XI, 9

Retardations. See critical points

Reversible steels, XVIII, 3

Retention theories of the hardening of steel, XV, 1

Roberts-Austen on the hardening of steel, XV, 1, 2

Roberts-Austen's equilibrium diagrams of iron-carbon alloys, XXIII, 17

use of neutral bodies for detecting critical points, VII, 14; A, 35
Robin, production of austenite by, XIII, 5

Roozeboom's equilibrium diagram of iron-carbon alloys, XXIII, 17
Rosenhain and Ewing. See Ewing and Rosenhain

Humfrey, straining of iron by, II, 10
Rosenhain's equilibrium diagram of iron-carbon alloys, XXIII, 20

Saladin self-recording pyrometer, A, 33
Saladin's cooling and heating curves of steels, VII, 15, 16
Segregation of impurities in steel, VI, 10
Self-hardening steel, XVIII, 13

-recording pyrometers, VII, 18; A, 33, 34
Silicates in steel, VI, 8
Silicon in cast iron, influence and occurrence of, XX, 1

steel, VI, 1
steel, XVIII, 15
Slag in wrought iron, III, 2

, composition of, III, 3

Matweieff' s method of etching, III, 3
, microstructure of, III, 3
Slip bands, description and production of, II, 10

in iron, II, 11

Sodium picrate etching, V, 7
Solid solutions, XXII, 4 to 9

, definition of, I, 5
Solidus, definition of, XXII, 4
Solution theories of the hardening of steel, XV, 2

12 INDEX

Sorbite, definition, description, and formation of, XI, 6; XII, 5; XIII, 13
Sorby-Beck parabolic reflector, A, 10
Special steels, XVII, 1 to 10; XVIII, 1 to 20
, austenitic, XVII, 7, 9
, cementitic, XVII, 1, 8, 9
constitution, properties, treatment, and uses of most important types

XVIII, 1 to 20

, definition and general character of, XVII, 1 to 10
, influence of special elements on position of critical range in, XVII, 3
, martensitic, XVII, 7, 9
, pearlitic, XVII, 6,. 8
, polyhedric, XVII, 1, 7, 9
, treatment of, XVII, 8
Specimen holders, A, 7 to 9
Spheroidizing of cementite, XII, 14
Stable equilibrium, definition of, XXIV, 2
Stages, electromagnetic, A, 11

, mechanical, A, 3, 12
Stead on phosphorus in cast iron, XX, 3 to 7

the brittleness of low carbon steel, XII, 26

crystalline growth of very low carbon steel, XII, 23
Steadite, definition and description of, XX, 3
Stead's brittleness, XII, 28
Steel, annealing of, XII, 1 to 30

, temperatures of, XII, 2
, brittleness of low carbon, XII, 26

, calculation of structural composition of, V, 8 to 12; VI, 6, 7
, carbon in, IV, 3
, case hardening of, XVI, 1 to 6
castings, annealing of, XII, 13
, causes of critical points in, VIII, 1 to 16
, cementation of, XVI, 1 to 6
, chemical tests for the detection of sulphur in, VI, 4, 5

vs. structural composition of, VI, 6
, chrome, XVIII, 13 to 15
-nickel, XVIII, 16
, constitution, properties, treatment, and uses of most important

types of special, XVIII, 1 to 20
, cooling and heating curves of, VII, 10 to 19
, ductility vs. structural composition of, V, 17
, eutectoid, definition and structure of, V, 4
, effects of critical points in, IX, 1 to 8
, formation of graphite in high carbon, XII, 15
, ghost lines in, VI, 10
, hardening of, XIII, 1 to 21
, high carbon, cementite in, V, 4

, microstructure of, V, 4
, pearlite in, V, 4
-speed, XVIII, 17 to 20

, hyper-eutectoid, definition and structure of, V, 4
, hypo-eutectoid, definition and structure of, V, 4
, impurities in, VI, 1 to 12
, influence of cold working on the structure and properties of, XI, 8

finishing temperatures on the structure and properties of, XI, 3
hot working on the structure and properties of, XI, 1 to 7
, iron oxide in, VI, 8

sulphide in, VI, 3
, irreversible, XVIII, 2

INDEX 13

Steel, low carbon, etching of, IV, 6
, ferrite in, IV, 2

, microscopical examination of, IV, 6
, microstructure of, IV, 2
, pearlite in, IV, 3
, vs. wrought iron, IV, 1
, manganese, XVIII, 5 to 12
in, VI, 5
oxide in, VI, 8
sulphide in, VI, 3
, maximum strength of, V, 17
, mechanical treatment of, XI, 1 to 10
, medium high carbon, pearlite in, V, 1

, microstructure of, V, 1
, microstructure of high vs. low phosphorus, VI, 11

sulphur, VI, 12
, nickel, XVIII, 1 to 5
, normal structure of, IV, 1
, occurrence of critical points in, VII, 1 to 20

of maximum hardening power, XIII, 20
, phosphorus in, VI, 2

, physical properties of constituents of, V, 14
, production and structure of burnt, XII, 17 to 20
, relation between structure and critical points of, VIII, 12

above and below the critical range of,

XII, 20

, reversible, XVIII, 3
, segregation of impurities in, VI, 10
, self-hardening, XVIII, 13
, silicates in, VI, 8
, silicon, XVIII, 15

in, VI, 1

, special, XVII, 1 to 10; XVIII, 1 to 20
, structural changes on cooling in, VIII, 5 to 16
, structure of cast, X, 1 to 10

worked, XI, 1 to 10
, sulphur in, VI, 2

, tenacity vs. structural composition of, V, 15
, tempering of hardened, XIV, 1 to 10
, theories of hardening of, XV, 1 to 7
, tungsten, XVIII, 12, 13, 17
, vanadium, XVIII, 15, 17

vs. carbon content, grading of, IV, 1
Straining, crystalline growth after, I, 7, 8

of iron, II, 11

Stress theories of the hardening of steel, XV, 5
Structural composition of alloys, XXII, 15 to 20

cast iron, calculation of, XIX, 5, 10 to 13;

XX, 7 to 9

iron-carbon alloys immediately after solidifica-
tion, XXIII, 3

steel, calculation of, V, 8 to 12; VI, 6, 7
Subcarbide theory of the hardening of steel, XV, 4
Sulphur, allotropy of, II, (i

in cast iron, influence and occurrence of, XX, 1
steel, VI, 2

, chemical tests for the detection of, VI, 4, 5

INDEX

Taylor and White's discovery of high-speed steel, XVIII, 18

Temper carbon, XXI, 1

Temperatures for annealing steel, XII, 2

Tempering and the retention theories of the hardening of steel, XV, 6

stress theory of the hardening of steel, XV, 6
colors, XIV, 1

, decrease of hardness on, XIV, 9
, explanation of, XIV, 2
, heat liberated on, XIV, 9
, influence of rate of cooling in, XIV, 2

time in, XIV, 1
of austenitic steel, XIV, 3
case hardened steel, XVI, 6
hardened steel, XIV, 1 to 10
martensitic steel, XIV, 5
troostitic steel, XIV, 5
temperatures, XIV, 1

Tenacity of steel, structural composition vs., V, 15
Ternary steels, XVII, 1. See also special steels
Thermal critical points. See critical points

treatment. See heat treatment
Toughening treatment, XII, 8
Transformation points. See critical points
range. See critical points
Transition constituents. See also martensite, troostite, and sorbite

, definition and formation of, XIII, 15, 17
Troostite, Arnold's view as to the nature of, XIII, 12

, Benedicks' view as to the nature of, XIII, 12

, definition, description, occurrence, properties, etching, and

structure of, XIII, 11
Troostitic steel, tempering of, XIV, 5
Troosto-sorbite, XIII, 15
Tschernoff iron crystallite, II, 5
Tungsten, influence on the critical points of iron of, XVIII, 12

steel, XVIII, 12, 13
Twinnings, definition of, II, 7
in brass, II, 7

gamma iron, II, 7
marble, II, 7
produced by pressure, II, 7

Univariant equilibrium, definition of, XXIV, 3
Universal metalloscope, A, 10 to 13, 28
Unstable equilibrium, definition of, XXIV, 2
Unvariant equilibrium, definition of, XXIV, 3

Vanadium steel, XVIII, 15, 17
Vertical illumination, A, 14 to 16

illuminators, A, 14, 16 to 18

Water-toughening of manganese steel, XVIII, 11
Welsbach lamp, A, 19
Widmaustatten structure, X, 6

INDEX 15

White, AJaunsel. See Taylor and White
White cast iron, XIX, 3

heart castings, XXI, 4

, annealing for, XXI, 4
Wrought iron, composition of, III, 1
, definition of, III, 1
, fibers in, III, 2

, microscopical examination of structure of, III, 7
, microstructure of, III, 1, 2
, slag in, III, 2
vs. low carbon steel, IV, 1

Zeiss prism illuminator, A, 17

THIS BOOK IS DUE ON THE LAST DATE
STAMPED BELOW

AN INITIAL FINE OF 25 CENTS

WILL BE ASSESSED FOR FAILURE TO RETURN
THIS BOOK ON THE DATE DUE. THE PENALTY
WILL INCREASE TO SO CENTS ON THE FOURTH
DAY AND TO St.OO ON THE SEVENTH DAY
OVERDUE.

UJN tB 1944

JUN 16 1944

t fEB 2 1 946"

APR 23

111946

LD 21-10m-5,'43 (6061s)

THE UNIVERSITY OF CALIFORNIA LIBRARY

tfiil^^i