I'll I i 
 
 
 ! 
 
 iii ilii i 
 
 
 i 
 
 II 
 
 i 
 
M * 
 
 LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF CALIFORNIA. 
 
 

 I 
 
 
 

Electric Furnaces and their Industrial 
 Applications 
 
ELECTRIC FURNACES AND 
 
 THEIR INDUSTRIAL 
 
 APPLICATIONS 
 
 J. WRIGHT 
 it 
 
 WITH 57 ILLUSTRATIONS 
 
 or THE ^ 
 
 UNIVERSITY 
 
 NEW YORK 
 THE NORMAN W. HENLEY PUBLISHING CO 
 
 132 NASSAU STREET 
 
wg- 
 
INTRODUCTION 
 
 THE development of the electric furnace, and the various 
 industries with which it is associated, as a necessary auxiliary 
 to the processes involved, is making vast strides, and is 
 regarded with ever-increasing interest by the electro- 
 chemist and metallurgist, who see in it possibilities, far 
 beyond those offered, until some few years back, by the blast 
 furnace on a commercial scale, and the oxy-hydrogen flame 
 in the laboratory. 
 
 The limitations of temperature imposed by the methods 
 available prior to the introduction of electricity as a heating 
 agent, were such as to render commercially impossible 
 many of the processes now % carried out by its aid. On the 
 other hand, the introduction of the electric furnace with its 
 vast possibilities in the field of exceedingly high tempera- 
 tures, gave rise at first to misuse of the power available ; 
 i.e. the temperatures required for the various reactions were 
 over, rather than under estimated, and, as a result, the sub- 
 stances which should have 'been produced by the furnace 
 were again split up into constituent elements, or built up 
 into compounds other than those which it was originally 
 intended to produce. Many of the early failures were due 
 to this cause, especially in furnaces of the "arc" type y 
 but experience has taught the usual lesson, and it is now 
 possible to regulate the temperature of an electric furnace 
 within far narrower limits than is possible with furnaces of 
 the ordinary type, consuming coke and coal fuel. 
 
 There is hardly an electro-metallurgical process to which 
 the electric furnace has not been applied, either experi- 
 
 
INTRODUCTION 
 
 mentally, or on a commercial scale, and, though many of the 
 experimental attempts have, thus far, proved abortive, the 
 success of the remainder justifies the hope that perseverance, 
 aided by the incentive to progress in the shape of much 
 valuable material, such as low-grade ores, so-called " waste 
 products," etc., at present unworkable, may lead to increased 
 research in this promising field, in which it is fairly safe to 
 predict that the indefatigable worker will not go unrewarded. 
 The introduction of acetylene gas for illuminating 
 purposes, has of course, been responsible for the great 
 advance in the carbide industry, which at present con- 
 stitutes the largest branch of electric furnace work, 
 although at one time the boom, and consequent over 
 production, threatened ruin to many works. So great, in 
 fact, was the reaction consequent upon an overstocked 
 market, that many factories possessing furnaces and plant 
 for the manufacture of calcium carbide, turned their atten- 
 tion to other substances, which, with a little adaptation, 
 their furnaces could be made to produce. Alloys of the 
 ferro-chrome type are amongst the substituted products, 
 being largely used in hardening armour plates, etc. 
 
 Electric smelting furnaces for the reduction of iron and 
 other ores to the metallic state, have also provided a subject 
 for extensive research, with no small measure of success, it 
 having been found possible to apply the electric smelting 
 process in many cases where the blast furnace, either from 
 scarcity of fuel or similar causes, was out of the question. 
 
 It is naturally impossible at the present stage of our 
 knowledge concerning the generation of electrical energy 
 from coal and other fuels, for the electric smelting furnace 
 to compete with smelting furnaces using fuel direct, but 
 there are many instances in which the electrical process 
 can be introduced with advantage, and a fair promise of 
 profit to the metallurgist. 
 
 Metalliferous ores frequently abound in localities where 
 fuel of any description is scarce, but water power plentiful ; 
 
INTRODUCTION 
 
 in such cases, a hydro-electric generating plant, though 
 expensive to instal, frequently provides a ready way out of 
 the difficulty. Water power has, in fact, been the salvation 
 of the electro-chemical and electro-metallurgical industries 
 which for the greater part involve a vast expenditure of 
 power within a comparatively small compass. Hence we 
 find almost all the large works concerned in the foregoing, 
 allied industries, confined to America and the Continent, 
 and frequently to localities where fuel is comparatively 
 scarce but water power plentiful. 
 
 Niagara, with its 50,000 H.P. derived from the celebrated 
 Falls, is a case in point, and it is worthy of note that by far 
 the greater proportion of this huge power is utilized in the 
 two industries enumerated above. 
 
 It is unnecessary to enlarge further, at this point, upon the 
 many applications of the electric furnace in modern in- 
 dustry, in that the various processes will be alluded to iifc 
 connexion with the several furnaces to be described later, 
 
 There is also a type of furnace which the author proposes 
 to include in this work, with the object of rendering it 
 complete, and that is the electrolytic type. Although not, 
 strictly speaking, furnaces, in the ordinary acceptation of 
 the term, there are several forms of apparatus which depend 
 for their action on a combination of thermal, electrolytic, and 
 chemical effects. Since the office of the furnace proper, 
 viz., that of heating the raw materials, is an essential 
 auxiliary in such cases, it has been resolved to include them 
 in the pages which follow, the necessary line of demarcation 
 between a thermal and an electrolytic process pure and 
 simple being determined by the presence of an aqueous 
 solution. Only such methods and apparatus will be dealt 
 with as involve the presence of a fused electrolyte. 
 
 Several of the types of electric furnace construction, de- 
 scribed in detail in the pages which follow, have never pro- 
 gressed beyond the experimental stage. Their descriptions 
 are included with a view to showing the extraordinary 
 
INTRODUCTION 
 
 amount of ingenuity which has been expended upon electric 
 furnace design, with a view to rendering them efficient, and, 
 as far as possible, automatic in action. It is this very 
 ingenuity, entailing a certain elaboration of detail which 
 has militated against the commercial success of several 
 very promising furnace inventions. To put the matter in 
 a nutshell, the aspiring designers of some electric furnaces 
 have attempted to apply principles of construction and 
 operation comparable with the delicate mechanism and 
 controlling principles of the self -regulating arc lamp. 
 
 It is obviously a mistaken feature in electric furnace 
 design, this elaboration of detail. In an apparatus whose 
 parts may be, and frequently are, subjected to extremely high 
 temperatures, it is necessary that every part shall possess 
 stability, in order to resist the destructive tendency of such 
 great heat, a stability which the nature of some of these 
 inventions, renders impossible of attainment. 
 
 The illustrations in this book are essentially in the nature 
 of sectional diagrams, representing " principles of construc- 
 tion " rather than views of the objects as they actually 
 appear to the observer. Photographs of the majority of 
 electric furnaces would convey little or no information to 
 the reader ; a mass of brickwork, with perhaps some iron 
 plates ; a series of heavy cables leading in ; and one or two 
 flues for carrying off the gaseous products : that is all. 
 
 J. W. 
 
 viii 
 
SUMMARY OF CONTENTS 
 
 SECTION I PAGES 
 
 HISTORICAL AND GENERAL 1-20 
 
 SECTION II 
 ARC FURNACES ....... . . 21-26 
 
 SECTION III 
 RESISTANCE FURNACES AND TYPICAL PROCESSES . . 27-57 
 
 SECTION IV 
 
 CALCIUM CARBIDE MANUFACTURE . . . 58-106 
 
 SECTION V 
 
 IRON AND STEEL PRODUCTION IN THE ELECTRIC 
 
 FURNACE 107-164 
 
 SECTION VI 
 
 PHOSPHORUS MANUFACTURE IN THE ELECTRIC 
 
 FURNACE . 165-169 
 
 SECTION VII 
 GLASS MANUFACTURE IN THE ELECTRIC FURNACE . 170-175 
 
SUMMARY OF CONTENTS 
 
 SECTION VIII PAGES 
 
 ELECTBOLYTIC FURNACES AND PROCESSES . . . 176-213 
 
 SECTION IX 
 
 MISCELLANEOUS ELECTRIC FURNACE PROCESSES . 214-220 
 
 SECTION X 
 LABORATORY FURNACES AND EXPERIMENTAL RESEARCH 221-234 
 
 SECTION XI 
 TUBE FURNACES * . . . . . . 235-245 
 
 SECTION XII 
 TERMINAL CONNEXIONS AND ELECTRODES . . . 246-256 
 
 SECTION XIII 
 EFFICIENCY AND THEORETICAL CONSIDERATIONS . 257-264 
 
 SECTION XIV 
 MEASUREMENT OF FURNACE TEMPERATURES 265-283 
 
f\8 R A 
 OF THE 
 
 { UNIVERSITY } 
 
 SECTION I 
 
 HISTORICAL AND GENERAL 
 
 Definition. Strictly speaking, an electric furnace is an 
 apparatus for bringing about a physical or chemical change 
 in materials by the aid of heat obtained from the transforma- 
 tion of electrical energy. There is, however, another 
 class of apparatus, which the writer has seen fit to include, 
 under the title " Electrolytic Furnaces," in which the action 
 is in part electro-thermal, as in the electric furnace pure and 
 simple, and, for the rest, electrolytic. This class of furnace 
 is mainly employed for the electrolysis of fused salts, as, 
 for instance, in the manufacture of aluminium, where heat 
 is a necessary auxiliary if the requisite fusion of the electro- 
 lyte is to be maintained. 
 
 Historical. As early as 1853 a form of arc furnace con- 
 struction, devised by Pichon, was described in the Practical 
 Mechanics' Journal. It consisted of a series of arcs, set up 
 between electrodes of large size, and through the several 
 independent heat zones of which, metallic ores, mixed with 
 carbon, were passed, with the object of reduction. There 
 is no record of such a furnace having ever existed, except on 
 paper. 
 
 Sir William Siemens may justly be credited with having 
 been the first to suggest the employment of electric furnaces 
 on a commercial scale. For some time (1878-1879) he 
 conducted experiments, with a view to determining the 
 possibilities of the arc furnace as an auxiliary to certain 
 industrial processes, and in June, 1880, embodied the results 
 of his researches in a paper which he read before the then 
 
ELECTRIC FURNACES AND 
 
 Society of Telegraph Engineers. In the course of his experi- 
 ments with the furnace described below, he succeeded in 
 obtaining an efficiency of 33 per cent., and summed up his 
 conclusions as follows : 
 
 1. The degree of temperature attainable in the electric 
 furnace is theoretically unlimited. 
 
 2. Fusion may be effected in a perfectly neutral atmo- 
 sphere. 
 
 3. Furnace operations can be carried on in a laboratory, 
 
 FIG. l. 
 
 without much preparation, and under the eye of the 
 operator. 
 
 4. The limit of heat practically obtainable, with the use of 
 ordinary refractory materials is very high, because, in the 
 electric furnace, the fusing material is at a higher tempera- 
 ture than the crucible, whereas, in ordinary fusion, the tem- 
 perature of the crucible exceeds that of the material fused 
 within it. 
 
 The historical furnace used in connexion with these early 
 experiments is depicted in Fig. 1, and consisted of a 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 refractory crucible A, of plumbago, magnesia, lime, or other 
 suitable material, which may be varied according to the nature 
 of the substance to be treated within it. It is supported at 
 the centre of a cylindrical jacket B, and is packed around 
 with broken charcoal, which, being a poor conductor of heat, 
 isolates it from the surrounding atmosphere, and conserves 
 the heat developed within the crucible to such an extent 
 that there is very little loss due to radiation or diffusion. 
 The negative electrode consists of a massive carbon rod C, 
 passing axially through the centre of the crucible lid, and 
 free to move vertically therein, the clearance opening being, 
 for obvious reasons, very small. 
 
 The cathode, C, is suspended from the lower extremity of 
 a copper strap D, which conducts the current from it, being 
 attached at its upper end to the curved extremity of a hori- 
 zontal beam E. The positive electrode F, which may 
 be of iron, platinum or carbon, consists of a cylindrical rod 
 of one or other of these materials, passing up through the 
 centre of the crucible base. The other side of the beam, E, 
 carries, suspended from its extremity by a hinged joint, 
 a hollow soft iron cylinder G, forming the core of the solenoid 
 S. The core G works in a dash-pot P, the tendency of 
 the solenoid S, when active, being to raise it against the 
 counteracting force of the adjustable counterweight W, 
 thus lowering the cathode G into the crucible. The solenoid 
 winding is connected in shunt to the two electrodes. 
 
 The furnace^ was originally designed by Siemens for the 
 fusion of refractory metals, and their ores ; consequently, 
 once the action is started, electrical connexion is established 
 between the lower electrode F and the semi-metallic mass 
 in the crucible, and the arc continues to play between the 
 surface of the mass and the movable carbon rod C. As 
 the current through the furnace increases, that through the 
 shunt winding of the solenoid diminishes, and the weight 
 W, coming into play, causes its end of the beam to descend, 
 thus raising the cathode C, and restoring equilibrium. 
 
 3 
 
ELECTRIC FURNACES AND 
 
 To Moissan we are, of course, indebted for much valuable 
 information to as the possibilities opened up by the intro- 
 duction of the electric furnace, and the results of his researches 
 into the subject, duly assembled in convenient form in his 
 book Le Four Electrique, are well worthy of careful study 
 by those interested in the subject of high temperatures. 
 
 Moissan's most valuable researches into the chemistry 
 of high temperatures, rendered possible by the introduction 
 of the electric furnace, were carried out during 1893 and 1894. 
 Many conclusions may be drawn from an exhaustive study 
 of Moissan's work, and those having, or promising to have 
 an industrial importance are well summarized by Blount, in 
 his book Practical Electro-Chemistry , as follows 
 
 " The stable form into which carbon, whether amorphous, 
 or crystallized as diamond, tends to pass, is graphite. Under 
 ordinary conditions, carbon does not melt, but passes 
 directly into the gaseous state ; if subjected to high pressure, 
 as it may be by suddenly cooling a liquid, e.g. iron, in which 
 it is dissolved, it may be liquefied, and then may crystallize 
 as diamond. 
 
 " Lime, magnesia, molybdenum, tungsten, vanadium, and 
 zirconium may be fused. Silica, zirconia, lime, aluminium, 
 copper, gold, platinum, iron, uranium, silicon, boron and 
 carbon may be volatilized. The oxides among these 
 substances may be deposited in a crystalline form. Oxides 
 usually regarded as irreducible, e.g., alumina, silica, baryta, 
 strontia and lime, uranium oxide, vanadium oxide, and 
 zirconia may be reduced by carbon in the electric furnace. 
 Many metals which are reduced with difficulty in ordinary 
 furnaces, such as manganese, chromium, tungsten, and 
 molybdenum, may be prepared in quantity. Moreover, 
 in the electric furnace, these metals may be obtained of 
 approximate purity, in spite of their great tendency to 
 unite with the oxygen and nitrogen of the air. It often 
 happens that, when a metallic oxide is reduced with excess 
 of carbon in the electric furnace, a carbide of the metal is 
 
 4 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 first formed. From this the pure metal can usually be 
 prepared by fusing the carbide with the oxide of the metal. 
 The carbon is oxidized and an equivalent of the metal 
 reduced. The behaviour of such metals in dissolving 
 carbon at high temperatures, in rejecting it on cooling, and 
 in losing it when subjected to selective oxidation, in general 
 resembles that of iron, which is well known. One class of 
 bodies is particularly stable at the high temperatures 
 attainable by the electric furnace to wit, that comprising 
 the carbides, borides, and silicides. These substances are 
 usually of simple composition : SiC (silicon carbide), CaC 2 
 (calcium carbide), Mn 3 C (manganese carbide), Fe 2 Si (iron 
 silicide), FeB (iron boride), CB 6 (carbon boride), will serve 
 as examples. Some members of the group are extremely 
 hard. Thus carbon silicide (or silicon carbide) is harder than 
 emery, while boron carbide and titanium carbide may 
 actually serve to cut a diamond not merely to polish it, as 
 does silicon carbide, but to produce definite facets. Others 
 of the carbides have another claim to interest from an in- 
 dustrial as well as from a scientific standpoint. Everyone 
 knows nowadays that calcium carbide is decomposed by 
 water and yields acetylene ; but it is not always realized that 
 the property of thus giving rise to a hydrocarbon is general 
 for a large number of similar bodies, e.g. the carbides of 
 lithium, aluminium, thorium, and cerium. Lithium carbide 
 (Li 2 C 2 ) yields acetylene ; aluminium carbide (A1 4 C 3 ) gives 
 methane ; cerium carbide (CeC 2 ), a mixture of the gases 
 acetylene, ethylene, and methane, and a notable proportion 
 of liquid hydrocarbons. This brief catalogue of facts will 
 show how large a field for industrial research exists, and how 
 well mapped are the paths by which it may be entered." 
 
 Passing record of the earlier researches into the possibilities 
 of the electric furnace as an industrial auxiliary may be 
 gleaned from the Report of the Franklin Institute, July, 
 1898, on the researches of M. Henri Moissan. 
 
 The electric furnace utilized by Moissan for conducting 
 
 5 
 
ELECTRIC FURNACES AND 
 
 these researches was devised by him, and is of very simple 
 construction. It belongs to that class known as " Indirect 
 arc furnaces," so called from the fact that the arc itself is 
 not brought into actual contact with the material under 
 treatment, which receives its heat, instead, by reflection from 
 the furnace walls or roof. It is represented in Fig. 2, 
 and consists of two blocks of chalk, A, A, so hollowed out 
 that, when placed together, they form a cavity for the 
 reception of the carbon crucible C, which contains the 
 material to be treated, and constitutes, in point of fact, the 
 hearth of the furnace. Two carbon electrodes E E pro- 
 ject through the sides of the blocks, and meet, with the 
 exception of an arcing space, at a point just above the mouth 
 
 FIG. 2. 
 
 FIG. 3. 
 
 of the crucible C. Electrical connexion with the source of 
 current is secured through the metal clamps M M. Metal 
 bands B serve to hold the chalk blocks together whilst the 
 furnace is active. 
 
 As a laboratory type, this simple furnace has many 
 advantages ; it is built up of refractory material, and com- 
 prises comparatively few parts, which are easily taken to 
 pieces and reassembled. Furthermore, the centre of activity, 
 surrounded as it is, on all sides, by a considerable thickness 
 of refractory and non-conducting material, can be brought 
 to an extremely high temperature, with very little accom- 
 panying loss of energy. 
 
 6 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 One of the earliest resistance furnaces or muffles, which 
 depended for its action on the heat generated in a conductor 
 of reduced cross-section embedded in the substance of its 
 walls, was that patented by Faure in 1883. It was intended 
 for the manufacture of sodium, and a sectional view is repre- 
 sented in Fig. 3, from which it will be seen that the 
 conductors c c were embedded in the hearth. One or two 
 later modifications of this early type are described in the 
 section dealing with laboratory furnaces. 
 
 The commercial resistance furnace is the outcome of the 
 inventive genius of Messrs. Eugene H. and Alfred H. 
 Cowles, who, after numerous experiments, selected coarsely 
 powdered carbon as a suitable material for the resistance 
 core, whilst, at the same time, a necessary ingredient, for the 
 reduction of oxides. 
 
 From an historical point of view, the following extract 
 from a paper read before the American Association for 
 the Advancement of Science, in 1885, by Professor Chas. 
 F. Mabery, may prove of interest : 
 
 " A short time since, Eugene H. Cowles, and Alfred H. 
 Cowles, of Cleveland, conceived the idea of obtaining a con- 
 tinuous high temperature, on an extended scale, by intro- 
 ducing into the path of an electric current, some material 
 that would afford the requisite resistance, thereby pro- 
 ducing a corresponding increase in the temperature. After 
 numerous experiments, that need not be described in detail, 
 coarsely pulverised carbon was selected as the best means 
 for maintaining a variable resistance, and at the same time 
 the most available substance for the reduction of oxides. 
 When this material, mixed with the oxide to be reduced, 
 was made a part of the electric circuit in a fire-clay retort, 
 and submitted to the action of a current from a powerful 
 dynamo machine, not only was the reduction accomplished, 
 but the temperature increased to such an extent that the 
 whole interior of the retort fused completely. In other 
 
 7 
 
ELECTRIC FURNACES AND 
 
 experiments, lumps of lime, sand, and corundum were fused, 
 with indications of a reduction of the corresponding metal ; 
 on cooling, the lime formed large, well-defined crystals, 
 the corundum, beautiful red, green, and blue, hexagonal 
 crystals. 
 
 " Experiments already made show that aluminium, silicon, 
 boron, manganese, magnesium, sodium, and potassium can 
 be reduced from their oxides with ease. In fact there is no 
 oxide that can withstand temperatures attainable in this 
 electrical furnace. Charcoal in considerable quantities is 
 changed to graphite ; whether this indicates fusion, or 
 solution of carbon in the reduced metal, has not been fully 
 determined. 
 
 " As to what can be accomplished by converting enormous 
 electrical energy into heat within its limited space, it can only 
 be said that it opens the way into an extensive field for 
 pure and applied chemistry. It is not difficult to conceive 
 of temperatures limited only by the capability of carbon to 
 resist fusion." 
 
 The Cowles furnace made its first appearance in public 
 in 1885, its initial application to the needs of industry being 
 in the reduction of oxides (vide the zinc furnace, in which a 
 graphite crucible forms one electrode). 
 
 In 1887, the Cowles Brothers took out a patent on a 
 furnace with an arrangement for continuous feeding of 
 the charge. 
 
 Temperatures Attainable in the Electric Furnace. The 
 maximum temperature attainable by the combustion of 
 fuel, either in solid, liquid, or gaseous form, and under the 
 most favourable conditions for the conservation of the heat 
 developed, is in the neighbourhood of 2,000 C.=3,632F., 
 although Heraeus, in a paper before the German Bunsen 
 Society in 1902, claimed that he had succeeded in con- 
 structing a non-electric furnace, in which temperatures up to 
 2,200 C.=3,992F. could be produced. He employs an 
 
 8 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 iridium tube, suitably mounted in a furnace, and heated by 
 means of an oxy-hydrogen flame. The temperature was 
 measured by the aid of a thermo-couple, consisting of abso- 
 lutely pure iridium, and an alloy of 90 per cent, iridium with 
 10 per cent, ruthenium. Up to 1,650C.=3,002F. a direct 
 comparison was made between this thermo-couple, and a 
 standard calibrated by the Reichsanstalt. Above this 
 temperature the several values were arrived at by calcula- 
 tion. 
 
 The temperature of the electric arc itself has never been de- 
 termined, the only available data on the subject of such high 
 temperatures being the results of temperature or calori- 
 metric measurements, made on the active extremities of 
 the carbons. Thus, in 1893, Violle tested the temperature 
 of the positive carbon crater by photometric methods, and 
 found it to be 3,500C.=6,332F., and independent of the 
 magnitude of the current producing it, between 10 and 400 
 amperes. This estimate is only subject to error through a 
 corresponding miscalculation of the specific heat of carbon, 
 and was subsequently modified by the investigator to 
 3,600C.:=6,512 F., as the result of assigning a slightly 
 different value to this specific heat. 
 
 It is probable that the temperature of the arc itself is 
 slightly higher than the above figure, which may nevertheless 
 be taken as the approximate limiting temperature of 
 furnaces operating on the arc principle, at atmospheric 
 pressure, and with carbon electrodes. Assuming this to be 
 the temperature at which carbon vaporizes, it is obvious that 
 a limit is similarly set upon the temperature obtainable in 
 furnaces of the " resistance " type, in which a carbon core 
 is employed. 
 
 Basing his deductions upon an interesting and instructive 
 experiment performed by Moissan, Townsend (Electrical 
 World, April 6, 1901), suggests defining the limiting tempera- 
 ture of the arc between carbon electrodes, as that tempera- 
 ture at which the complex carbon molecule breaks down ; 
 
 9 
 
ELECTRIC FURNACES AND 
 
 and that of the resistance furnace, with carbon core, as th6 
 true point of vaporization of carbon, the former tempera- 
 ture being distinctly lower than the latter. 
 
 The experiment referred to as having been performed by 
 Moissan is mentioned in Comptes Eendus, vol. cxix, p. 776, 
 and consisted in exposing to the direct heat of a 2000 ampere, 
 80 volt arc, a carbon tube having an internal diameter or bore, 
 of one centimetre. The experiment served to demonstrate 
 the volatilization and condensation of carbon, the interior 
 of the tube becoming filled, under the intense heating effect 
 of the arc, with carbon vapours, which subsequently con- 
 densed upon its walls in the form of graphite. Crystallized 
 silicon, placed at the lower extremity of the tube, fused and 
 volatilized, with the result that its ascending vapour, meet- 
 ing the descending carbon vapour, combined with it to form 
 transparent needle-like crystals of silicon carbide (carbor- 
 undum). 
 
 Townsend argues that, since silicon carbide was formed in 
 this manner the temperature of the vapours was below that at 
 which this compound is decomposed, e.g. below that of 
 the Acheson graphite furnace ; hence his deductions as to 
 the relative limiting temperatures of arc and resistance 
 furnaces as set forth above. 
 
 The heat intensity, or temperature attainable, in an electric 
 furnace depends, among other things, on the heat-conserving 
 qualities of the materials of which the furnace is constructed. 
 In furnaces of the Moissan type, but lined with blocks of 
 pure carbon, and reinforced, on the outside, with a re- 
 fractory non-conductor of heat, such as chalk or magnesia, 
 it is possible, therefore, to obtain a temperature of approxi- 
 mately 4,000C.=7,232P. whilst temperatures ranging 
 from 2,000-3,500C.=:3,632 -6,332 F.,are easily reached 
 and maintained in the commercial electric furnace. 
 
 Classification. There are two leading types of electric 
 furnace, distinguished from one another by the method 
 in which the heat energy is produced. They are known 
 
 10 
 
^ \ * B 
 X- V OF THE 
 
 { UNIVERSITY ) 
 
 \^ <?4 iroftH^^^ 
 
 THEIR INDUSTRIAL APPLICATIONS 
 
 respectively as " Arc " and " Resistance " furnaces. The 
 former performs its work by virtue of the intense heat 
 generated by the arc set up between two or more electrodes, 
 one of which is frequently constituted by the lining, or 
 base, of the furnace chamber itself. The latter derives 
 its heating power from the passage of a heavy current 
 through an attenuated conductor of high resistance, such 
 as a carbon pencil. 
 
 These two main classes may be subjected to further 
 subdivision, depending upon the principle of their con- 
 struction and working. The arc furnaces, for example, 
 may be divided into those 
 in which the raw material 
 to be heated is subjected 
 to the direct heat of the arc, 
 and is brought into actual 
 contact with it, and those 
 in which the heating is 
 effected indirectly, either by 
 conduction, or reflection from 
 the roof or walls. In this 
 latter type, to which belongs 
 the Moissan furnace already 
 described, the charge does 
 
 not come into contact with the electrodes at all, but is 
 situated below the plane of the arc, and sometimes partially 
 protected from it, thus reducing the chances of contamina- 
 tion by particles which might become detached therefrom, a 
 very necessary precaution in some processes, such as glass 
 manufacture, for example. 
 
 Another way of classifying furnaces, adopted by some 
 writers, is according to their construction, and the arrange- 
 ment of the electrodes. Thus arc furnaces may be sub- 
 divided under separate headings, according to whether 
 the electrodes are horizontal, vertical, or parallel. A 
 typical arc furnace, with horizontal electrodes, is shown 
 
 II 
 
 FIG. 4. 
 
ELECTRIC FURNACES AND 
 
 in diagram in Fig. 4, E E being the electrodes, and H 
 the refractory hearth, or crucible, provided with a tap- 
 hole, as shown, at its lowest point, for drawing off the molten 
 products of the reaction. 
 
 Resistance furnaces, sometimes called " Incandescent 
 Furnaces," may be similarly subdivided, according to 
 whether the heating is directly or indirectly effected ; a 
 further distinction also arises out of the nature of the 
 resistance itself ; in some cases, this is independent, and 
 takes the form of a carbon pencil, or even a glower of the 
 Nernst type, whilst in others it is constituted by a 
 poorly conducting column of the unconverted charge it- 
 self. 
 
 Electrolytic furnaces are essentially of the resistance 
 type, the resistance in this case being the electrolyte, 
 usually a fused salt, under treatment. 
 
 Another modern development of the resistance furnace 
 principle is that class of apparatus known as " Induction 
 Furnaces," of which the Kjellin furnace, for the manu- 
 facture of steel, is a typical example, and will be described 
 later on. These induction furnaces depend for their 
 action upon the heating effect of a current induced in them 
 by a neighbouring conductor through which an alternating 
 current is flowing. They are, in effect, transformer fur- 
 naces, the furnace itself, or rather its contents, forming 
 a closed-circuit secondary winding, in which is induced 
 a heavy current, by the comparatively small current flowing 
 through a neighbouring primary winding of many turns. 
 The principal advantage of this class lies in the complete 
 absence of troublesome electrodes ; no minor consideration, 
 as will be seen when dealing with the various electric furnace 
 processes. 
 
 There is yet another type, which is, however, confined, 
 for the greatei part, to small units and laboratory apparatus, 
 and is known as the " Tube Furnace." A cylindrical or 
 tubular construction offers several advantages, more 
 
 12 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 especially for laboratory work and experimental research. 
 Tube furnaces operate either on the resistance principle, 
 or a combination of both arc and resistance heating, and 
 are dealt with in a special section. 
 
 Two qualifying terms often used in connexion with 
 electric furnace work are " Intermittent " and " Con- 
 tinuous." An intermittent furnace is one in which only 
 one charge, equal to the capacity of the furnace, can be 
 dealt with at a time, without interrupting the operation 
 for the purpose of removing the product. A continuous 
 furnace, on the other hand, is so constructed with regard 
 to its tapping and feed arrangements, that a process can 
 be carried on continuously without shutting down, the 
 material under treatment either passing through at a 
 regular rate, or the furnace being partially relieved of its 
 contents at regular intervals without necessitating a stop- 
 page of the current. Continuous furnaces are sometimes 
 referred to as " Tapping " furnaces. 
 
 As regards the relative advantages and disadvantages 
 of the arc and resistance principles of electric furnace 
 working, it may be stated in general that the intense and 
 more or less concentrated heat developed in the arc is 
 unsuitable for many commercial processes, which call for 
 a less violent and more readily controllable source of heat, 
 such as is provided in the resistance principle of con- 
 struction. 
 
 Here, again, we are confronted with further difficulties, 
 for, if the resistance core be composed of the material 
 itself under treatment, such, for example, as a mixture of 
 ore and slag, in an electric smelting furnace, the current, 
 and therefore the heating effect, will be subject to extreme 
 variations, depending upon the changes brought about 
 in the conductivity of the core as the operation proceeds, 
 and the liberated metal asserts its presence. To mitigate 
 this drawback to a certain extent, the molten product 
 must be continually tapped off, such that a constant cross- 
 
 13 
 
ELECTRIC FURNACES AND 
 
 sectional area of resistance core may be, as far as possible, 
 maintained. 
 
 Better conditions for the control and regulation of the 
 temperature obtain in resistance furnaces with independent 
 cores, i.e. furnaces in which the resistance column does 
 not form part and parcel of the charge under treatment, 
 but is independent of it, and maintains its shape and cross- 
 section, within reasonable limits, throughout the process. 
 The main drawback to their use is, of course, the extra 
 expense involved by the introduction of such a core, and 
 its subsequent periodical renewal. 
 
 Rasch (Zeit. /. Elektrochemie, February 19, 1903) deprecates 
 the use of the arc furnace, with carbon electrodes, for the 
 majority of processes. He cites the following requirements 
 as essential to the success of pyrochemical reactions in 
 general 
 
 " The electrodes must be capable of withstanding a high 
 energy density, and a high temperature ; they must not 
 be oxidized, and must not exercise a reducing action on 
 the reaction to be obtained." 
 
 He suggests the substitution of conductors of the second 
 class for carbon ; e.g., rods, or tubes, of magnesia, aluminium 
 oxide, etc., which require to be pre-heated, after the manner 
 of a Nernst lamp glower, before they become conductors. 
 This principle has already been applied to laboratory fur- 
 naces, and small units, but its adoption on a more extensive 
 scale is at present limited by the cost of the material, and 
 the difficulties incidental to the construction of a tube 
 furnace of any magnitude. 
 
 General Remarks. Employed for heating purposes, 
 electricity is only capable of evolving that quantity of 
 thermal energy represented by Joule's equivalent of the 
 energy originally consumed in generating the current. 
 In the case of a steam-driven plant, the 'heat energy avail- 
 able at the furnace terminals is only a small fraction (from 
 yo to sV) of the calorific value of the coal used in the boiler 
 
 14 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 furnaces. Of this latter about one-fifth is lost in the flues 
 during combustion : the waste gases account for a con- 
 siderable proportion, of far greater value, from a heating 
 point of view, than the electricity generated ; whilst some 
 75 per cent, of fuel energy is present in the exhaust steam. 
 It will readily be seen, therefore, that electric heat, 
 derived from steam-driven plant, is only commercially 
 applicable in special cases, involving especially high tem- 
 peratures, which are unattainable by ordinary combustion 
 methods, concentrated heating effect, or where the value 
 of the substances under treatment, and the necessity for 
 maintaining them in a pure state, outweigh the cost of the 
 heating itself, and render it a secondary consideration. 
 
 With natural water power, however, the matter assumes 
 a different aspect. Here we have a frequently unlimited 
 supply of energy, which is only costing the interest on 
 capital outlay for plant and maintenance, so that the com- 
 paratively wasteful transformation of water power into 
 thermal energy becomes a minor consideration. 
 
 In nearly all electro-chemical and electro-metallurgical 
 processes, the cost of the electrical energy constitutes a 
 considerable percentage of the total cost of operation ; 
 usually 25 per cent, or more, sometimes as much as 90 per 
 cent. Cheap electrical energy is therefore an essential 
 requirement for the economical working of an electric 
 furnace process. 
 
 The principal advantages of the electric furnace, from 
 an industrial standpoint, may be summarized as follows : 
 
 1. The time required for a reaction is reduced, and the 
 yield increased for a given time. 
 
 2. Reactions take place more completely. 
 
 3. The heating effect is concentrated at the point where 
 it is most required. 
 
 4. It is capable of bringing about high temperature 
 reactions impossible at ordinary temperatures. 
 
 5. Operations may be readily conducted in the presence 
 
ELECTRIC FURNACES AND 
 
 of various gases, bringing their chemical action into play 
 on other substances at high temperatures, or causing one 
 or more gases to react together. 
 
 In laying out an electric furnace plant, it is essential 
 that the furnaces be as near to the source of current as is 
 compatible with the safety of the generators, and accessi- 
 bility of the former. No undue loss of energy is then 
 likely to take place in the main connecting cables, whilst 
 the quantity of copper required for the latter is reduced 
 to a minimum. Some appreciable separation between 
 the two is, of course, essential, owing to the intense heat 
 which prevails in the immediate neighbourhood of the 
 furnace, and which would prove prejudicial to the dynamos. 
 M. Keller advocates a separation of two metres, with a 
 dividing wall between generator and furnace. The cables 
 would then be from 18 to 20 metres long, and the power 
 factor as high as 0'9. 
 
 For electric furnace work, in the absence of electrolytic 
 action, an alternating current is more easily regulable, 
 and yields better heating effect than a direct current. 
 
 According to Mr. J. B. C. Kershaw (Electrical Review, 
 July 7, 1899), the first instance of three-phase currents 
 being commercially employed in electric furnace work was 
 in the manufacture of calcium carbide, at Langenthal, in 
 Switzerland. The furnaces are of the movable hearth 
 type, and three are worked simultaneously, each taking 
 current of one phase, viz. 1,000 to 1,500 amperes at 75 
 volts. 
 
 The repulsion exerted by a strong magnetic field upon 
 the electric arc has not been without its attractions for 
 the furnace inventor, as instance a patent granted in 1897 
 to W. H. Monk. The principle of the invention is equally 
 applicable to either arc or resistance furnaces, and consists 
 in the provision of a rotating magnetic field in the neigh- 
 bourhood of the arc or heating current, whereby the latter 
 is also caused to rotate or circulate through the mass of 
 
 16 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the material to be heated, and thereby extend its sphere 
 of influence. 
 
 In one form of arc furnace on this principle, detailed 
 in the published specification, arcs were set up between 
 concentric tubular carbons, which also served as feed 
 passages, and a solid electrode, mounted on an iron plate. 
 Surrounding this was an iron ring, suitably protected from 
 the intense heat, and wound continuously, with four branch 
 connexions, whereby it could be supplied with two alternat- 
 ing currents in quadrature, to produce the rotating field. 
 
 A rotating field is similarly produced around the central 
 resistance core or cores in a furnace working on the latter 
 principle: 
 
 The project is an ingenious one, and worthy of mention, 
 but it is questionable whether the extra expense involved 
 in the furnace construction, to say nothing of the power 
 consumed in producing the required magnetic field, is 
 overset by the attendant advantages. At all events, there 
 is no record of the principle having been adopted in practice. \ 
 
 Another inventor, I. L. Roberts, adopts the travelling 
 belt, or conveyor principle, with a view to securing con- 
 tinuity of furnace action. His invention relates to a 
 delivery hopper, and flue, or casing, covering a slowly 
 travelling belt, woven from asbestos coated wire, on which 
 the raw material to be treated is deposited from the hopper, 
 and carried into the heat zone created by the arc set up 
 between two carbon electrodes, held in adjustable clamps 
 and projecting through the walls of the flue. 
 
 In 1898, Messrs. Siemens Bros. & Co. took out a patent 
 solely on a form of furnace construction which prevented 
 the ingress or egress of atmospheric air either to or from 
 the interior of the furnace. To this end the raw material 
 under treatment was built up in the form of a thick layer, 
 capable of preventing the permeation of air, whilst the 
 upper carbon electrode was given a tubular form and 
 utilized for the secondary purpose of carrying off the gases 
 
 17 c 
 
ELECTRIC FURNACES AND 
 
 produced in the reaction, without permitting combustion. 
 This tubular construction for the upper electrode has since 
 been applied in many different designs of furnace, and is 
 used, not only as a flue for the gases, but also as a feed 
 channel for the raw materials to be treated. 
 
 Statistics. In the words of an old saw, " The proof of 
 the pudding is in the eating," and, despite the somewhat 
 wasteful transformation of energy entailed by the industrial 
 application of the electric furnace, a few statistics from 
 various sources will serve to show that in many cases, and 
 for various processes, the advantages of the electric furnace 
 have so far outweighed this drawback as to warrant their 
 installation on a scale of considerable magnitude. 
 
 According to Prof. Borchers, the aggregate power utilized 
 in electro-chemical and electro-metallurgical industries 
 employing electric furnaces in 1900 was 
 
 For Calcium carbide manufacture . . 180,000 h.p. 
 Aluminium . * 27,000 
 
 Carborundum . 2,600 ., 
 
 Total 209,600 h.p. 
 
 The total output of electrical power at Niagara in 1901 
 was 50,000 H.P., of which 23,000 H.P. was consumed in 
 electro-chemical and electro-metallurgical processes. Among 
 the principal consumers for electric furnace operations are 
 the following, the statistics being taken from Cassier's 
 Magazine for May, 1901 
 
 The Acheson International Graphite Company 1,000 
 H.P., utilized in the conversion of anthracite into graphite, 
 at a pressure of 80 volts. 
 
 The Pittsburg Reduction Company 5,000 H.P., utilized 
 in the extraction of aluminium from bauxite by the Hall 
 process. Direct current at 160 volts. 
 
 The Carborundum Company 2,000 H.P., utilized in 
 the manufacture of silicon carbide. Alternating current at 
 J 10 volts. 
 
 18 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The Niagara Electro-Chemical Company 500 H.P., 
 utilized in the production of sodium, and sodium peroxide 
 from caustic soda by the Castner process ; 165 volts, con- 
 tinuous current. 
 
 The Union Carbide Company 10,000 H.P., utilized in 
 the manufacture of calcium carbide. Each furnace 200 H.P. 
 Alternating current at 110 volts and 25 alternations per 
 secor'I. 
 
 French statistics (1904) show that out of a total of 
 238,703 H.P., available from the falls of the Alps, 22,536 H.P. 
 is employed in the manufacture of aluminium ; 20,485 H.P. 
 in other metallurgical industries ; and 104,466 H.P. in 
 calcium carbide manufacture. 
 
 Scientific Deductions. The principle of the electrical 
 furnace, and the reactions brought about by the extremely 
 high temperatures produced within it, have been advanced 
 in explanation of many geological facts, hitherto unex- 
 plained, e.g., the formation of natural gases, petroleum, 
 bitumen, graphite, corundum, etc., all of which occur in 
 nature, and can be artificially produced, under similar 
 conditions, by the direct, or indirect aid of the electric 
 furnace. 
 
 In a paper before the French Academie des Sciences 
 (Comptes Rendus, 134, pp. 1185-1188), MM. Sabatier and 
 Sendereus, writing on the synthetic production of petroleum, 
 incidentally call attention to the existence of the carbides 
 of the alkaline earth metals in nature. In addition to 
 opening up possibilities of artificial petroleum becoming 
 an electric furnace product, this fact also lends colour to the 
 oft-suggested theory that processes analogous to those of 
 the high temperature reactions brought about in the electric 
 furnace also take place as the result of natural phenomena. 
 
 Moissan's original theory of the natural formation of 
 petroleum, which is supported by the results obtained by 
 the above investigators, is that it results from the action 
 of water upon natural carbides, under certain conditions 
 
 19 
 
ELECTRIC FURNACES AND 
 
 involving the presence of metals, and a temperature not 
 exceeding 149 C=300F. 
 
 The above authorities summarize their results as follows. 
 A liquid, resembling Caucasian petroleum, in all its physical 
 properties is obtained by passing a mixture of acetylene 
 gas and hydrogen, in certain proportions, over finely divided 
 nickel, or a metal belonging to the same chemical group, 
 maintained at a temperature of from 200-300C = 392- 
 572F. 
 
 Variations in the physical conditions of the experiment 
 resulted in the formation of an oil resembling Galician 
 petroleum, and, like it, containing aromatic hydrocarbons, 
 whilst lowering the temperature to 180C=356F. furnished 
 a liquid resembling American petroleum. 
 
 The passage of acetylene gas per se over the finely divided 
 metals resulted in the formation, by exothermic reactions, 
 of hydrocarbons of the unsaturated series, which, when 
 mixed with hydrogen, and again passed over the heated 
 metal, led to the formation of a similar product to the first, 
 resembling Caucasian oil. 
 
 20 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION II 
 
 ARC FURNACES. 
 
 An indirect arc furnace, for the treatment of metallic 
 oxides and metalloids, such as silica, ferrous chroma te, 
 bauxite, etc., and converting them into a vitreous and 
 homogeneous product of commercial value, is repre- 
 sented in Fig. 5. The process in its entirety consists in 
 subjecting the raw materials to the radiated or luminous 
 heat of an arc, which fuses them, and then subsequently 
 cooling the resultant fused mass. The furnace in which 
 the fusion is effected is of simple construction, its design 
 being such that the charge never comes in actual contact 
 with the arc, but is heated solely by radiation and reflection. 
 The materials to be treated are fed in through a central 
 chimney or flue F, fitted with a gas-tight valve V, and a 
 connexion, C, for producing a vacuum in the furnace chamber. 
 On reaching the bottom of F, the charge divides between 
 the two inclined chutes BB, which form a bridge or dome 
 immediately over the arc A. In passing down these, it 
 receives a preliminary heating, the fusion being completed 
 when it reaches the hearth, H, of the chamber, by the direct 
 radiant heat from A. 
 
 An indirect arc furnace devised by G. de Chalmont 
 comprises a refractory crucible, supported on one carbon 
 electrode, which passes vertically up through the base of 
 the furnace chamber, whilst the remaining electrode is 
 horizontally disposed, in the shape of four radial sections 
 which bear on the outer surface of the wall of the crucible. 
 The charge is thus screened from the direct heat of the arcs, 
 which play over the exterior of the crucible. 
 
 21 
 
ELECTRIC FURNACES AND 
 
 The de Chalmont furnace, invented by the late Dr. de 
 Chalmont, and employed by him in his researches on metallic 
 silicides, presents several novel and interesting features in 
 furnace construction. It is represented in diagrammatic 
 
 FIG. 5. 
 
 FIG. 6. 
 
 section by Fig. 6, and consists of a cast-iron trough, T, 
 made in two sections, upper and lower respectively, which 
 are connected by a flanged joint at F. The whole is mounted 
 on wheels, W, for portability. A carbon lining, C, constitutes 
 one electrode and also the hearth of the furnace proper, 
 which is provided with the usual tapping hole t, and flue /. 
 J is a layer of refractory insulating material, supporting the 
 lid L, which, as represented in the figure, is given a sec- 
 tional construction, each separate portion being furnished 
 with a distinct conduit for water-cooling purposes. The 
 upper adjustable electrode A, also of carbon, enters through 
 a stuffing box B, in the centre of the cover, which effectually 
 prevents ingress of air, whilst at the same time permitting 
 the necessary adjustments of the movable electrode. Sh 
 is a shunt interposed between the upper electrode and the 
 cover itself, its object being to prevent accidental arcing 
 
 22 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 between them, by maintaining them both at the same 
 potential, thus protecting the lid from injury. 
 
 The Denbergh arc furnace, invented by Dr. F. P. van 
 Denbergh, is specially adapted to a variety of operations, 
 chief among which may be cited the manufacture of sul- 
 phuric and phosphoric acids, and alkaline silicates, such as 
 " water-glass." The furnace is represented in sectional 
 elevation in Fig. 7. It comprises a fire-brick structure F, 
 contracted below, as shown, to form a crucible for the 
 reception of the molten pro- 
 ducts. The upper chamber 
 where the actual fusion takes 
 place is lined interiorly with 
 a refractory layer R, which 
 remains unaffected by the 
 gases and vapours produced. 
 E E are carbon electrodes, 
 each mounted in a species of 
 universal ball joint, built 
 into the furnace wall, which 
 permits a longitudinal or feed 
 adjustment, and, at the same 
 time, a swivelling motion of 
 the electrodes, whereby the 
 
 position of the arc may be regulated at will. The base B is 
 removably rabbited into the furnace walls, and supported 
 in position by the loose bricks 6, b. A flue, the opening to 
 which is shown at o, leads up from the base of the furnace 
 within the walls to a point which determines the level of 
 the molten mass within ; its position, within the heated 
 walls, effectually prevents " freezing " and consequent 
 stoppage of the outlet. A is an outlet at the top of the 
 furnace, for gaseous products. The raw material is fed 
 in at a regular rate, from the hopper H, by a reciprocating y 
 piston-feed mechanism P, driven by the belt D, and falls 
 vertically through the central heat zone of the arc. 
 
 23 
 
ELECTRIC FURNACES AND 
 
 In this furnace, the inventor manufactures phosphoric 
 acid from apatite, etc. The mineral is first crushed, and a 
 quantity of sand added to it to serve as a flux. Under the 
 heat of the arc, phosphorus is liberated from the mixture. 
 The atmosphere of the furnace is rich in oxygen, with 
 which the phosphorus immediately enters into combination, 
 forming oxides. These latter are subsequently hydrated 
 by steam, which is introduced into the furnace chamber for 
 the purpose. 
 
 A somewhat ingenious continuous arc furnace is the 
 invention of Mr. F. J. Patten. The principal mechanical 
 and electrical features of its construction are represented 
 
 FIG. 8. 
 
 in Fig. 8. It consists, in the main, of a vertical centre 
 spindle S, suitably supported for free rotation, in bearings, 
 and carrying a flat annular table T, which, protected by a 
 refractory bed B, constitutes the moving hearth or floor 
 of the furnace, upon which is spread the raw material to be 
 subjected to the action of the arc. The revolving table is 
 additionally supported on ball bearings b b. The two carbon 
 electrodes E E are carried by vertical supports, capable 
 of lateral adjustment by means of a screw carriage, mounted 
 on a radial arm, which also permits of their rotation inde- 
 pendently of the hearth. A greater number of electrode 
 carriers may be added, if it be wished to increase the 
 capacity of the furnace. 
 
 An additional feature in the construction consists of 
 
 24 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 electro-magnets m m attached to the electrode terminals, 
 which, when active, deflect the arcs down on to the charge 
 after the manner of a blow-pipe. 
 
 A patent, dated July 17, 1897, granted to C. Bertolus, 
 covers the application of polyphase currents to electric 
 furnace working on the arc principle. 
 
 The advantages of three-phase currents for electric 
 furnaces of the arc type are 
 
 1. More uniform distribution of heat throughout the 
 mass of material under treatment. 
 
 2. Accidental extinction of the arc a very remote con- 
 tingency. 
 
 3. Polyphase generators are of simpler construction, and 
 less liable to breakdown, consequent upon the sudden 
 variations in load, incidental to arc furnace working. 
 
 Koller's arc furnace construction is very simple, and was 
 devised with a view to securing better distribution of the 
 heating effect throughout the charge, whilst at the same 
 time rendering the furnace suitable for use on circuits of 
 higher voltage than usual. It consists of a longitudinal 
 chamber, with massive carbon blocks, projecting through 
 the end walls. A series of carbon blocks, supported in line 
 with these terminal electrodes, are arranged along the 
 furnace chamber at regular intervals, their number varying 
 according to the voltage. The arc is thus split up into a 
 series, and a number of heated regions are secured in the 
 centre of the mass of raw material, which is packed around 
 the blocks. 
 
 A rather ingenious application of the arc furnace principle 
 to the carbonizing of electric lamp filaments is the outcome 
 of a patent granted to W. L. Voelker, the inventor of the 
 Voelker Carbide Lamp. The device consists of a fire- 
 proof crucible, enclosing the arcing space between two 
 vertical carbon electrodes, which constitute the terminals 
 of the furnace. The untreated filament is reeled off from 
 a drum, through orifices in the walls of the crucible, on to 
 
 25 
 
ELECTRIC FURNACES AND 
 
 another receiving drum, its line of passage through the 
 crucible being so disposed as not to pass through the axis 
 of the arc itself, but to one side of it, the arc being deflected 
 on to it. The crucible is kept charged, either with carbon 
 vapour, or a hydrocarbon reducing gas, during the treatment. 
 A patent for a similar method of procedure in the manu- 
 facture of filaments for incandescent lamps, has been granted 
 to R. A. Nielsen, and is applicable to oxides of the rare 
 earths, such as are utilized in Nernst lamp glowers. These 
 oxides are subjected to the heat of the arc, whilst shielded 
 from its electrolytic action by a tube of refractory material, 
 until they melt or soften, whereupon they are drawn out 
 into threads of the required diameter. 
 
 Action of the Arc Furnace. A paper by MM. Gin and 
 Leleux, before the Paris Academy of Science in 1898, dis- 
 cussed the action of the electric arc furnace, from a theo- 
 retical standpoint, as compared with the results obtained 
 in actual practice. We may regard the fall of potential 
 between the electrodes of an arc furnace as taking place 
 through the resistance column of gas which lies between them, 
 and is the product of the volatilization of either the elec- 
 trodes themselves, or a portion of the raw material under 
 treatment. Regarding this column of gas as a theoretically 
 perfect cylinder, having a sectional area equal to that of 
 the electrodes, and surrounded by a perfect heat-conserving 
 screen or jacket, we can prove that the temperature would 
 increase in direct proportion to the square of the current 
 density, and the ratio of the resistivity to the specific heat 
 per unit volume, of the atmosphere of the arc. 
 
 In actual practice, however, this result is never obtained ; 
 the arc forms a species of pocket or envelope, with a small 
 orifice in its upper portion, through which the various gases, 
 resulting from the reaction, make their escape. The size 
 of this envelope, or cavity, increases, until a point is reached, 
 at which the resultant heat of the arc is equivalent to the 
 heat dissipated in the surrounding raw material. 
 
 26 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION III 
 
 RESISTANCE FURNACES AND TYPICAL PROCESSES 
 
 The heat developed in a conductor, such as the core of a 
 resistance furnace, by the passage of an electric current 
 through it, depends upon 
 
 1. The ohmic resistance of the conductor. 
 
 2. The magnitude of the current. 
 
 3. The duration of its flow. 
 
 All three quantities are open to regulation in the majority 
 of resistance furnaces, hence the facility with which this 
 type lends itself to exact temperature regulation and control. 
 
 The temperature of a heating resistance core depends 
 upon the rate of development of heat within it, and upon 
 the rate at which it is capable of dissipating that heat by 
 radiation and conduction. The ultimate fixed temperature 
 of a resistance furnace core is attained when the above two 
 quantities are exactly equal, i.e. when the rate at which 
 heat is dissipated or given off by the conductor is equal to 
 the rate at which it is developed. 
 
 The limit of temperature attainable in such a conducting 
 core depends, of course, upon the fusing point of the material 
 of which it is composed. 
 
 An excellent example of the various possibilities of a 
 suitably designed resistance furnace, in which the tempera- 
 ture regulation is well under control, and can be altered 
 with a fair degree of accuracy over an extended range, is 
 to be seen in the various Acheson furnace processes, in which 
 
 27 
 
ELECTRIC FURNACES AND 
 
 silicon and carbon constitute the raw materials under 
 treatment. 
 
 By certain variations in the conditions and temperatures 
 at which the reactions are brought about, it is possible to 
 produce carborundum, " white stuff," silicon, siloxicon, or 
 graphite, at will. 
 
 With resistance furnaces generally, and those of the 
 Acheson type, with only a partially independent core, in 
 particular, the question of efficient temperature regulation 
 is an important one, in that it cannot be effected by any 
 change in the internal conditions of the furnace charge, or 
 core ; the latter, once the operation is started, is unalter- 
 able, and it is only by external means that the current, 
 and consequently the temperature, can be governed and 
 regulated. 
 
 A feature peculiar to this type of furnace is a high internal 
 resistance, which gradually falls during working, until, 
 at the end of an operation, it is fairly low, a condition of 
 things which can only be compensated by corresponding 
 regulation of the voltage at the furnace terminals. 
 
 The actual resistance which a furnace is desired to have, 
 can only be procured, or reproduced, by experiment, a 
 process in which the choice of many different grades of core 
 carbon, together with a variation in the size and shape of the 
 constituent particles, provides a wide field for initial 
 adjustment. 
 
 Herr Otto Vogel, a German authority, has made the 
 following suggestions with regard to resistance furnaces. 
 The heating resistance or core should be given the largest 
 possible surface area, in order to facilitate the transference 
 of its heat to the materials of the furnace charge. 
 
 Assuming, for the sake of example, three distinct cores, 
 each of 1,200 square millimetres cross-sectional area, but 
 having different formations, viz., circular, square, and rect- 
 angular ; their efficiencies, from the point of view of heat 
 transference, will vary considerably. 
 
 28 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Taking the diameter of the first (circular) core as 39 m.m., 
 its effective surface area, per metre of length, will be 132,000 
 square m.m. The second (square), with 35 m.m. side, will 
 expose an area of 140,000 square m.m., for the same length, 
 whilst the third (rectangular), 80 m.m. by 15 m.m., will, 
 under similar conditions, have a surface area of 190,000 
 square m.m. Their efficiencies will therefore stand in the 
 ratio 6:7:9. 
 
 Again, by constructing the circular electrode in the form 
 of a hollow cylinder, 30 m.m., internal, and 46 m.m. external 
 diameter, its effective surface area will, for the same cross- 
 section as formerly, be increased to 146,000 square m.m. 
 for one metre of length, whilst its mechanical strength will 
 be superior to that of the original solid cylinder. 
 
 The same authority quotes the results of tests made on 
 independent carbon resistance cores, with a view to ascer- 
 taining their rate of oxidation, or combustion, in air. 
 
 To this end, a current of 500 amperes was passed through 
 a rectangular carbon core, 400 by 80 by 15 m.m., which, 
 with a gradually increasing current density, consequent 
 on the reduction of its cross-sectional area by combustion, 
 was raised, during a period of five hours, to a white heat. 
 
 Current was then switched off and the block examined, 
 when it was found that the original dimensions of 80, and 
 15 m.m., had decreased to 60, and 7 m.m. respectively, 
 whilst the under side of the block had been reduced to a 
 knife edge ; the original cross-sectional dimension of 1,200 
 square m.m. had been reduced to 210 square m.m., with a 
 corresponding increase in current density, from 0*41 to 2*38 
 amperes per square m.m. 
 
 A further test was then made with a view to ascertaining 
 the temperature limitations imposed by this method of 
 resistance heating with independent carbon core. A second 
 carbon block, having the same dimensions as the first, was 
 totally embedded in powdered lime, with the exception of its 
 flat upper surface, which was left exposed to the atmosphere. 
 
 29 
 
ELECTRIC FURNACES AND 
 
 A current of 300 amperes was first applied for a period of 
 15 minutes, thus bringing the temperature of the carbon 
 up to a p.oint suitable for the commencement of the tests, 
 after which the current was increased by regular increments, 
 at intervals of five minutes, the temperature resulting from 
 each augmentation of current being measured by the col- 
 lapse of Seger cones placed upon its exposed surface. 
 
 The following results were obtained : 
 
 CURRENT DENSITY IN 
 
 SEGER CONE 
 
 RESULTANT 
 
 TEMPERATURE. 
 
 AMPERES PER SQ,. M.M. 
 
 No. 
 
 CENTIGRADE. 
 
 FAHRENHEIT. 
 
 0-25 
 
 0-22 
 
 590 
 
 1,094 
 
 0-33 
 
 0-18 
 
 710 
 
 1,310 
 
 0-45 
 
 0-8 
 
 990 
 
 1,814 
 
 0-54 
 
 0-2 
 
 1,110 
 
 2,030 
 
 0-66 
 
 4 
 
 1,210 
 
 2,210 
 
 1-2 
 
 24 
 
 1,600 
 
 2,912 
 
 1-6 
 
 36 
 
 1,850 
 
 3,362 
 
 2-0 
 
 
 
 2,300 
 
 4,172 
 
 The block was then replaced by one of circular section, 
 having a diameter of 30 m.m. With this, the current 
 density was increased to as much as 4 to 6 amperes per 
 square m.m., without disturbing its solidity. 
 
 At 8 amperes per square m.m., however, the carbon 
 
 FIG. 9. 
 
 commenced to volatilize, though no intermediate fluid 
 state was noticed. Its original diameter was reduced, by 
 this treatment, to 8 m.m. 
 
 Furnaces. An early form of electric furnace devised by 
 Borchers may be taken as typical of the resistance principle 
 of construction. It consists (Fig. 9), of a cavity A in the 
 centre of a block of refractory material B, through the 
 sides of which pass two large carbon electrodes E E. 
 
 30 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 These latter are fitted with metal terminal clamps M for 
 attachment to the source of current, and extend to a point 
 just within the cavity A, where their extremities are bridged 
 by a thin carbon pencil C, which becomes heated by the 
 passage of the current, and imparts its heat to the raw 
 material, which is either placed within, or caused to pass 
 through the cavity. 
 
 A modification of what is more familiarly known as the 
 " Cowles Furnace " is represented in Fig. 10. It consists 
 of a rectangular fire-brick structure F, with hearths sloping 
 to a central tapping hole T. L is a lid, fitted with feed 
 hoppers H H. The current enters by way of the end 
 
 FIG. 10. 
 
 electrodes E E, and passes through the core C, which 
 consists of a mixture of the ore, or other material to be 
 treated, with carbon. The latter, in this instance, fulfils 
 the double office of conductor and reducing agent. The 
 core is surrounded by a layer of granular charcoal G, which 
 conserves the heat generated, and is, at the same time, 
 highly refractory. The hopper openings in the lid L en- 
 able a constant and evenly distributed charge to be main- 
 tained within the furnace. 
 
 L. S Dumoulin's improved construction for furnaces of 
 the independent resistance type relates to the resistances 
 
 31 
 
ELECTRIC FURNACES AND 
 
 themselves, and consists in a means for protecting them, 
 and thereby prolonging their useful life. 
 
 To this end, the heating resistances, which are of the 
 usual carbon rod type, with a reduced central cross-section, 
 where they come within the furnace, and enlarged extremi- 
 ties by which they are supported within the furnace walls, 
 and to which the terminal connexions with the external 
 circuit are effected, are completely encased in a fire-clay 
 jacket, which is applied in a plastic state, and moulded 
 around the resistance rod so as to act as a complete pro- 
 tection from the disintegrating effects of the charge under 
 treatment. 
 
 If necessary, additional protection is secured by thread- 
 ing the rods, thus previously coated, through additional 
 jackets, or tubes of fire-clay, which extend through the 
 furnace proper, and are, like the rods themselves, supported 
 at their extremities by the walls of the structure. 
 
 A resistance furnace, patented by W. T. Gibbs, Canada, 
 is based on the principle of indirect heating, where the 
 charge does not come into actual contact with the heated 
 resistance, but receives its heat by reflection from the domed 
 roof. The furnace, which is too simple to need illustration, 
 consists of a built-up rectangular structure of carbon blocks, 
 or fire-brick, and having, as already stated, a domed roof 
 for reflecting the heat downwards on to the charge. Suit- 
 able inlet and outlet orifices are provided, both for solids 
 and gases, while the heating resistances consist of carbon 
 rods, placed transversely across the furnace, about two- 
 thirds up, and clamped between massive carbon terminal 
 blocks, which are held, in turn, by cast iron sockets, 
 let into the walls, and connected to the source of cur- 
 rent. 
 
 The heating capacity of the Gibbs furnace is governed 
 by its size, and the number of carbon rod heaters employed, 
 whilst its adaptability to currents of varying character 
 lies in the external means for changing the connexions of the 
 
 32 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 rods, which may be either in series, parallel, or a combina- 
 tion of the two. 
 
 The resistance furnace invented by M. Girod, is capable 
 of a wide range of application, whilst, at the same time, of 
 comparatively simple construction. It consists of a cruci- 
 ble of graphite, or refractory earths, encased on the outside 
 by a graphite sheathing, which serves the purpose of a high 
 resistance path between the electrodes, and in which heat 
 is developed by the passage of a suitable current. The 
 structure is mounted on a horizontal shaft, so as to admit 
 of tipping, or canting, during operation, after the manner 
 of a Bessemer converter. The voltage necessary ranges 
 from 20 to 80, and the temperature can be regulated be- 
 tween 500 and 3, 500 C. = 932 and 6,332 F., or even 
 higher, if necessary. A simple furnace, such as the above 
 appears to be, should be readily adapted to foundry opera- 
 tions, where refractory metals have to be dealt with. 
 
 In the light of Dr. Nernst's discoveries regarding the 
 conducting power of certain solids, when raised to a high 
 temperature, the question of a satisfactory material for 
 the construction of the walls of resistance furnaces is an 
 important one, in that, if the material chosen be in the 
 nature of a "solid electrolyte," considerable loss of energy 
 is likely to arise from the passage of useless currents through 
 the walls of the structure, when raised to the temperatures 
 necessary to bring about a re-action. 
 
 Frohlich seems to have recognized the need for investi- 
 gation in this direction, and, in an article in Zeitschrift fur 
 Elektrochemie, August 6, 1903, he describes his investigations 
 in search of a suitable material. 
 
 The necessary properties for such are a high melting point, 
 chemical indifference , high electrical resistance, and plasti- 
 city, enabling the substance to be moulded into various 
 shapes and sizes. 
 
 He (Frohlich) claims to have discovered a number of 
 chemical compounds which fulfil these conditions, their 
 
 33 D 
 
ELECTRIC FURNACES AND 
 
 electrical resistance being much higher than that of carbon, 
 whilst they are decomposed but little, if at all, by uni- 
 directional currents. 
 
 His sugggestion, which he has since carried into practice, 
 is to construct conducting plates of these materials, which, 
 furnished with suitable electrodes, would constitute the 
 heating units of a resistance furnace of any required shape 
 or form. The composition of the material is not stated. 
 
 In Dr. 0. Frohlich's lately patented resistance furnace, 
 constructed on this principle, the walls of the structure 
 itself constitute the heating resistances, rendering the em- 
 ployment of an independent core superfluous. The material 
 used in the construction of the resistance walls, the com- 
 position of which is not divulged, is said to have sixteen 
 times the resistivity of carbon when hot, and twenty-five 
 times its resistivity when cold. A suitable material has 
 also been discovered for insulating the semi-conducting 
 walls, the principal merit of which is that it fails to become 
 a conductor, even when raised to a white heat. 
 
 The furnace is surrounded by a refractory heat-conserving 
 jacket, an air-space being left between the two. 
 
 Experimental trials have demonstrated the possibility 
 of attaining a temperature of at least 1,600C. r=2,912F., 
 by means of this furnace. The resistance material itself 
 does not fuse until a temperature well above 2,000C.= 
 3,632F., has been reached. 
 
 The application of polyphase currents to electric re- 
 sistance furnaces is the subject of a patent, of which the 
 details were published in 1897 by Messrs. H. Maxim and 
 W. H. Graham. The projected furnace is of fire-brick 
 with a series of parallel horizontal electrodes projecting 
 through one of its walls. These are connected individually, 
 through metres and switches, with the several circuits of a 
 polyphase generator. Opposite their inner extremities, 
 on the other side of the furnace, is arranged a common 
 electrode, to act as a return conductor. 
 
 34 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 For the manufacture of calcium carbide in this furnace, 
 small resistance rods or pencils of carbon are so disposed 
 as to bridge the space between each individual electrode, 
 and the common return, whilst the raw mixture of coke and 
 lime is packed around them. These rods become heated 
 with the passage of the current, and convert the mixture 
 into coherent masses of calcium carbide, which are subse- 
 quently removed with the aid of tongs. 
 
 The electrodes are carried on threaded rods, passing 
 through suitable handles, which permit the necessary with- 
 drawal of the electrodes pending the insertion of new carbon 
 pencils, and, when in place, an even pressure is maintained 
 by springs. 
 
 If so desired, the return, or common electrode, may be 
 subdivided, and each pah- enclosed within a separate com- 
 partment of the furnace. 
 
 A novel type of resistance, or, as it is sometimes termed, 
 " Induction " furnace, patented independently, with some 
 slight variation of detail, by Colby, Ferranti, and Kjellin, 
 consists of an annular, or helical channel, in a refractory base, 
 filled with a conducting, or semi-conducting medium which 
 constitutes the furnace charge, and has a heavy current 
 induced in it by a surrounding coil of many turns, carrying 
 an alternating current. 
 
 The device, in point of fact, acts as the closed-circuit 
 secondary of a step-down transformer, and is said to be 
 admirably adapted for the fusing of such metals as pla- 
 tinum, which, if exposed to the atmosphere during the 
 process, as in the ordinary type of furnace, occlude oxygen, 
 and other gases in their mass, which lead subsequently to 
 blow holes, and other imperfections in the casting. The 
 Kjellin patent has since been considerably developed, and 
 is now applied to the smelting of iron ore and the manu- 
 facture of a special quality of crucible steel, at Gysinge, 
 Sweden. 
 
 The process is worked on a commercial scale, and further 
 
 35 
 
ELECTRIC FURNACES AND 
 
 particulars will be found in the section devoted to steel 
 production. 
 
 The Acheson Furnace Group. One of the most indefati- 
 gable workers in the field represented by the industrial 
 application of the resistance furnace is E. G. Acheson, of 
 Niagara. The many processes with which his name is 
 associated as inventor and originator are all capable of 
 being conducted, with some slight modifications, in the 
 same type of furnace, which has, in fact, come to be gener- 
 ally known as the " Acheson Furnace." 
 
 The form of construction is unique in being of a tem- 
 porary nature. The whole structure is built up afresh, 
 of loose fire-bricks, for each operation, and pulled down 
 again at the conclusion of a run. Such a furnace con- 
 struction offers several advantages ; it is not costly to 
 construct, is devoid of all considerations implied by the 
 terms, wear and maintenance, and is, moreover, simplicity 
 itself, a very important desideratum in electric furnace 
 construction. 
 
 Carborundum. The manufacture of carborundum, or 
 carbide of silicon, is one of the simplest industrial appli- 
 cations of the resistance principle. The furnace itself, 
 is not a permanent structure, but built up loosely of fire- 
 brick for each run, only to be pulled down again at the end 
 of the operation. The process is purely synthetic, carbon, 
 in the form of coke, being caused to combine, under stress 
 of electrical heat, with the silicon contained in sand as an 
 oxide. The oxygen of the silica combines with the excess 
 of carbon present, forming carbon monoxide gas, which is 
 given off, and burns at the various openings in the furnace 
 wall, with a blue flame. The silicon thus set free combines 
 with more carbon to form the carbide. 
 
 The furnace, Fig. 11, is built in the form of a species of 
 trough T, with massive electrode clamps, E E, projecting 
 through the end walls, and carrying each a bundle of carbon 
 rods C, which establish electrical connexion with the 
 
 36 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 core proper R running through the centre of the mass. 
 The core may consist of granular carbon, coke, or the 
 carbon rods themselves. The furnace is well stacked with 
 the raw mixture M of powdered coke and sand, to which 
 are also added sawdust and a small quantity of common 
 salt, the actual proportions of each ingredient contained in 
 a ten ton charge, which is the usual quantity dealt with 
 in one operation at Niagara, being coke, 34' 2 per cent. ; 
 sand, 54*2 per cent. ; sawdust, 9' 9 per cent. ; and common 
 salt, 1*7 per cent. 
 
 The object of the sawdust is to secure porosity of the 
 furnace charge, and that of the salt to act as a flux. 
 
 FIG. 11. 
 
 The subsequent conversion, which occupies about thirty- 
 six hours, results in some two tons of carborundum, as against 
 a theoretical yield of four tons, from which it will be seen 
 that the process is anything but efficient. 
 
 The carborundum furnaces in use at Niagara are 15 by 
 7 by 7 ft. The electrodes consist of bundles of sixty 
 3-in. carbon rods, each 2 ft. long. These are mounted in 
 heavy bronze sockets, supported by the brickwork. The 
 coke resistance core is 9 ft. long by 2 ft. in diameter, and the 
 grains of coke utilized in its formation vary from J to f in. 
 in diameter. An alternating current is employed. 
 
 The initial voltage at the furnace terminals is 200, which, 
 however, falls to 80 or thereabouts as the operation pro- 
 ceeds and the charge becomes heated. Regulation is 
 
 37 
 
ELECTRIC FURNACES AND 
 
 effected by varying the voltage so as to keep the furnaces 
 working at their full capacity of 1,000 h.p. apiece. 
 
 The charge, when broken up at the end of the run, is 
 found to consist of several layers, the centre one of which 
 is graphite, into which the original coke, or carbon of the core, 
 has been partially converted. Immediately surrounding 
 this is a layer, about 16 in. in thickness, of crystallized 
 carborundum, which constitutes the useful portion of the 
 charge, and consists approximately of 70 per cent, silicon, 
 and 30 per cent, carbon. Next to this will be found a 
 greenish layer of amorphous silicon carbide, mixed with 
 unconverted portions of the charge ; this layer is known 
 as " white stuff," and it is presumable that its temperatures 
 of formation and decomposition are fairly near together. 
 The extreme outer layer consists mainly of unconverted 
 material, together with a somewhat higher percentage of 
 common salt, than before, that substance having been 
 volatilized, and driven out in the process of conversion. 
 These latter layers are worked up again into the next charge. 
 
 In connexion with the low efficiency of the process, it is 
 stated that the electrical energy necessary for producing 
 carborundum has been reduced from 15' 5 to 8' 6 kilowatt 
 hours per kilogramme. 
 
 The temperature at which the operation is conducted 
 needs very careful regulation, in that, if allowed to reach 
 too high a value, the carborundum is again decomposed 
 and the silicon volatilized. 
 
 Pure carborundum is a colourless, crystalline compound, 
 having the formula, SiC ; the commercial article is dark 
 brown or black; its specific gravity is 3' 12, and it con- 
 sists of 70 per cent, silicon and 30 per cent, carbon. 
 
 Its formation in the electric furnace by the simultaneous 
 reduction of silica, and synthesis of the resulting elements, 
 is represented by the equation : 
 
 The furnace charge, which is packed around the core of 
 
 38 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 granulated coke, and fills up the entire furnace cavity, 
 weighs about ten tons, as already intimated. 
 
 Carborundum was first discovered by Acheson in 1893. 
 In that year, only 6$ tons were manufactured, but, in 
 1901, this figure had been increased to 1,690 tons, whilst 
 an extension of the plant (1902) from 2,000 to 3,000 h.p. 
 raises the possible output to 2,690 tons. It is largely 
 employed as an abrasive, a refractory lining for steel, 
 and cement furnaces, and a deoxidiser in steel manufac- 
 ture. 
 
 In connexion with its use for lining furnaces, Dr. W. 
 Engels has lately (1900) patented a method of applying 
 carborundum to the interior walls of furnaces in the form 
 of a paint. To this end, powdered carborundum is mixed 
 with one- third its weight of water glass, and water added 
 to the mixture until it assumes the consistency of cream, 
 in which form it is applied by means of a stiff brush to 
 the surfaces it is required to protect. If the furnace is 
 to be employed in the treatment of basic substances, e.g. 
 basic slag, fire-clay is substituted for the water-glass. It 
 is then made up of 85 per cent, carborundum, and 15 per 
 cent, clay, and applied in the same manner, viz., with a 
 brush, the surfaces to be protected having been pre- 
 viously dried thoroughly by gentle heating, the presence 
 of moisture being fatal to adhesion of the coat. 1,200 
 grams of carborundum are said to be required per square 
 metre of surface to be covered. 
 
 The manufacture of what is known as " white stuff," 
 a material largely consisting of silicon and carbon, which 
 is formed, to a limited extent, in the ordinary carborundum 
 furnace, as a thin layer between the carborundum and the 
 unconverted charge, has been made the subject of a patent 
 by Acheson. The respective temperatures of formation and 
 decomposition of this substance are comparatively near 
 to one another, rendering it necessary to devise a process 
 of manufacture in which these upper and lower tempera- 
 
 39 
 
ELECTRIC FURNACES AND 
 
 ture limits should be the limiting temperatures of the re- 
 action. Acheson achieves this result in a resistance furnace 
 with subdivided core, consisting of a number of equally 
 spaced carbon rods, distributed throughout the mass of 
 the furnace charge. The success and efficiency of the 
 process depend, of course, upon the spacing of these several 
 cores, which are so disposed that the temperature attained 
 in their immediate vicinity is not in excess of the maximum 
 whilst that at intermediate points does not fall below the 
 minimum temperature necessary for the formation of the 
 compound. The result of such a disposition is the con- 
 version of the entire charge into " white stuff." 
 
 Fitzgerald prepares refractory carborundum articles, 
 such as crucibles, furnace bricks, etc., by moulding crystal- 
 lized carborundum into the desired form, and subsequently 
 subjecting the articles to such a temperature, in the electric 
 furnace, that the mass re-crystallizes. 
 
 F. J. Tone, of Niagara Falls, has adopted " white stuff," 
 or amorphous carborundum, to a similar purpose, the pro- 
 duct, owing to its greater porosity, being especially suitable 
 for furnace construction, where it will be subjected to sudden 
 and excessive temperature changes. A temporary binder 
 of glue, or water-glass, is added to the mass previous to 
 moulding, and the articles may be conveniently heated 
 by embedding them in the usual charge of a carborundum 
 furnace. 
 
 Siloxicon is the name given to a new class of compounds, 
 the commercial value and mode of preparation of which 
 have been recognized and patented by Acheson, and 
 assigned to the Acheson Company of Niagara Falls. 
 
 Generally speaking, it consists of carbon, silicon, and 
 oxygen in chemical combination, and is described as 
 amorphous, grey-green, when cold, and light yellow when 
 heated to 149C.=300F. or above; density 2' 75 ; very 
 refractory to heat ; insoluble in molten iron ; neutral 
 towards acid and basic slags ; indifferent to all acids save 
 
 40 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 hydrofluoric ; unattacked by hot alkaline solutions, and 
 self-binding to such a degree that the use of a separate 
 binding agent is not essential in forming it into crucibles, 
 furnace-linings, fire-bricks, and such other articles as may 
 be manufactured from it. The articles may be formed by 
 merely moistening the pulverulent material with water, 
 moulding, and firing, or, if desired, a carbonaceous or other 
 binding material may be used. 
 
 It is a by-product of the process of carborundum manu- 
 facture, and is prepared in an Acheson resistance furnace 
 with multiple core. The manufacture of siloxicon bears 
 a striking resemblance to that of carborundum. The 
 process is patented by E. G. Acheson, and is being worked 
 by the Acheson International Graphite Company. The 
 raw materials are ground coke and sand, mixed in the 
 proportion of one part carbon to two parts silicon, with a 
 certain quantity of sawdust to impart porosity to the mass. 
 In adding the latter, due regard is paid to its carbon con- 
 tent ; the correct apportioning of the carbon percentage 
 in the furnace charge is of great importance, as, if it be 
 present in excess, the silica will be completely reduced, 
 and no formation of siloxicon will take place. Siloxicon 
 has a chemical composition approximately represented 
 by the formula Si 2 C 2 0. 
 
 As in the case of carborundum manufacture, the furnace 
 is loosely built of fire-brick, its approximate dimensions 
 being : length, 30 ft. ; width, 8 ft. ; and depth, 6 ft. To 
 operate a siloxicon furnace of this capacity 1,000 e.h.p. 
 are required. 
 
 The original siloxicon furnace had a double flat heating 
 core, but the writer understands that later furnaces have 
 been constructed with three cores. 
 
 The temperature at which the reaction takes place is 
 lower than that required for the production of carborundum, 
 the formation of siloxicon occurring at a temperature 
 between 2,480 and 2,757C.= 4,500 and 5,000F., whereas 
 
 41 
 
ELECTRIC FURNACES AND 
 
 carborundum manufacture entails a temperature of approx- 
 imately 3,876C. = 7,000F. 
 
 Freshly made siloxicon is a greyish green, loosely co- 
 herent mass, and, on leaving the furnace, it is subjected 
 to a milling process which reduces it to such a grade that 
 it will pass through a No. 40 sieve. 
 
 It is a self-binding material, and only needs forming 
 into a paste with water, when it can be moulded into any 
 desired form, which is rendered permanent by subsequent 
 firing. For some purposes, an admixture of 25 per cent. 
 clay may be introduced into the mass, prior to moulding, 
 or, if preferred, additional binding materials, such as liquid 
 tar, asphaltum, pitch, molasses or glue ; in fact, any similar 
 carbonaceous binding material may be employed, in which 
 case the siloxicon powder and agglutinant are thoroughly 
 mixed whilst hot, after which the mixture is moulded into 
 the desired form and burned, or, as an alternative, may be 
 used without burning, the subsequent heating to which it 
 is subjected in use being sufficient for the purpose. 
 
 It has a promising future in the manufacture of all 
 kinds of refractory articles, such as furnace linings, 
 crucibles, fire-bricks, etc. 
 
 For lining furnaces, the paste or mixture, prepared as 
 above, is applied to the surface to be protected, and well 
 rammed into place, where subsequent burning converts it 
 into a homogeneous refractory layer. 
 
 Siloxicon, when heated to a temperature of 1,466C. = 
 2,674F., or above, in an atmosphere containing free 
 oxygen in excess, decomposes, presumably, according to 
 the equation : 
 
 + 70=2Si0 2 
 
 If the siloxicon thus heated be agglomerated or moulded 
 into a coherent mass, the reaction is confined to the surface 
 of the latter, and produces a vitreous glaze, tinged light 
 green by the presence of iron as an impurity. In the absence 
 
 42 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 of free oxygen, or in a reducing atmosphere, no decomposi- 
 tion occurs at the above temperature, and the siloxicon 
 may be heated to the formation temperature of carborun- 
 dum, approximately 3,867C.=7,000F., without change. 
 At this point, however, it dissociates ; the equation repre- 
 senting its decomposition is supposed by Acheson to be : 
 
 Si 2 C 2 0=SiC+Si + CO. 
 
 Solid carborundum is left behind, whilst siloxicon vapour 
 and carbon monoxide gas are driven off. Carborundum, 
 subjected to the same treatment, behaves similarly. 
 
 Artificial Graphite. The electric furnace devised by 
 E. G. Acheson for the manufacture of graphite is, like the 
 other designs which owe their origin to his inventive genius, 
 of the resistance type. It takes the form of a trough-like 
 structure, in which is placed the rough charge of anthracite 
 coal, or coke, which it is required to convert into graphite. 
 The core is independent of the charge itself, and consists 
 of a series of carbon rods, extending through the mass to 
 be heated, and forming a bridge for the current between 
 two end electrodes. The raw charge itself is, at first, a 
 non-conductor, and the current, when switched on, passes 
 entirely through the carbon rods, raising them to a high 
 temperature. The heat thus evolved converts the layer of 
 coal or coke, immediately surrounding the core, into graphite, 
 which, in turn, becomes a conductor, and performs a similar 
 office for the next layer of raw material, and so on until 
 the whole charge has been converted. 
 
 A study of the various allotropic modifications of carbon, 
 and the changes they severally undergo when subjected 
 to high temperatures, provides a wide and interesting field 
 for the investigator, and many scientists, past and pre- 
 sent, recognizing the possibilities opened up by a more 
 advanced knowledge of the subject, have, from time to 
 time, devoted considerable attention to it. The principal 
 results of their experiments, in so far as they bear upon 
 
 43 
 
ELECTRIC FURNACES AND 
 
 the artificial production of graphite, have been assembled 
 in the form of a very interesting article by C. P. Townsend, 
 which appeared in the Electrical World, April 6, 1901, 
 and to which the reader is directed for the original refer- 
 ences. The writer has taken the liberty of extracting 
 therefrom the following interesting and salient points in 
 the history of the manufacture of artificial graphite. 
 
 Scheele, as far back as 1778, discovered the solvent 
 power of molten iron for carbon, and also the additional 
 fact that a portion of the carbon so dissolved separates out 
 on cooling in the form of graphite, a discovery which has 
 since exercised an important influence upon the iron 
 industries. 
 
 This knowledge was further supplemented in 1896 by 
 Moissan, who showed that the purity of the graphite thus 
 formed by separation on the cooling of the iron solvent 
 is proportional to the temperature to which the solvent 
 metal has been raised, and to the pressure exerted by the 
 iron in cooling. Moreover, it has been demonstrated that 
 graphite, from whatever source, invariably contains hydro- 
 gen as a constituent, though not, apparently, in chemical 
 combination with it. 
 
 In 1849 Despretz commenced a series of careful inves- 
 tigations into the changes undergone by carbon at high 
 temperatures, his apparatus consisting of a closed cast-iron 
 furnace, through the walls of which, and rendered air- 
 tight by leather stuffing boxes, projected two carbon 
 electrodes. 
 
 Inlet and outlet pipes for the introduction and with- 
 drawal of gases were provided, and the effects of variations 
 in the pressure duly noted. 
 
 This experimental furnace was operated both on the " arc " 
 and " resistance " principles, by the current from 600 
 Bunsen cells, in series of 25, 50, or 100, as occasion deter- 
 mined. 
 
 Various forms of carbon were experimented upon, and 
 
 44 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Despretz thus summarized the results of his investiga- 
 tions 
 
 (1) Carbon, in a vacuum, volatilizes at temperatures 
 produced by 500 to 600 Bunsen cells. Under pressure 
 of nitrogen at 2| atmospheres, volatilization occurs more 
 slowly. 
 
 (2) Carbon, at these temperatures, may be bent, welded, 
 or fused. 
 
 (3) Carbon from all sources becomes progressively softer, 
 finally turning to graphite. 
 
 (4) Graphite volatilizes slowly. 
 
 (5) Diamond is converted into graphite. 
 
 Strangely enough, this early investigator appears to have 
 actually applied the modern principles of artificial graphite 
 manufacture, and yet failed to produce it, for he asserts, 
 in the course of his report : " I have enveloped, I have 
 impregnated, acicular rods of carbon, with more fusible 
 materials, silica, alumina, magnesia, to determine whether 
 the presence of a more fusible body would render the fusion 
 of the carbon more easy. The silica, magnesia, and alumina, 
 escaped as vapours and the carbon remained unmodified." 
 
 In the case of anthracite coal, however, Despretz appears 
 to have met with a measure of success, for he speaks of its 
 conversion into " well characterized graphite," the con- 
 version having been effected in a crucible by the heat of 
 an arc struck between an independent electrode and the 
 anthracite under treatment. 
 
 In 1870 Berthelot studied the direct combination and 
 conversion of carbon ; he was unable to effect the conver- 
 sion of amorphous carbon into graphite by any other heat 
 save that of the arc, nor could he convert graphite into 
 amorphous carbon. Berthelot was responsible for the 
 discovery that graphite could be obtained as the result of 
 decomposing a carbide. He heated boron carbide in an 
 atmosphere of dry chlorine gas, and found that at a tempera- 
 ture below the softening point of glass, carbon separates 
 
 45 
 
ELECTRIC FURNACES AND 
 
 out as amorphous graphite, and, at the fusing point of porce- 
 lain as hexagonal crystals. 
 
 In 1893 Girard and Street introduced their system of 
 superficially graphitizing electrodes and other carbon 
 articles, by causing an arc or arcs to play over the entire 
 surface of the object, the process being carried out either 
 in an atmosphere of carbon monoxide, nitrogen, or similarly 
 neutral gas, or in a vacuum. 
 
 In 1896 Moissan pointed out that diamond, when sub- 
 jected to the heat of the arc, was converted into graphite, 
 a conclusion previously arrived at by Despretz. Moissan 
 also investigated the solution of carbon by many metals, 
 and its subsequent separation as graphite, on cooling. 
 Among the metals experimented with, were aluminium, silver, 
 manganese, nickel, chromium, tungsten, molybdenum, 
 uranium, zirconium, vanadium, titanium and silicon. 
 
 At normal pressures, carbon has no intermediate liquid 
 state. 
 
 The results of Moissan's varied and extensive researches 
 on the subject of graphite and its formation have been 
 thus summarized by him (Comptes Rendus, vol. 119, p. 980) 
 
 1. Whatever the variety of carbon, elevation of tempera- 
 ture always suffices to convert it into graphite. This 
 graphite may be amorphous, or crystallized, has a specific 
 gravity of 2' 10 to 2' 25, and a temperature of combustion 
 in oxygen of 660C.=1,220F. 
 
 2. There are several varieties of graphite, just as there are 
 several varieties of amorphous carbon and of the diamond. 
 
 3. The stability of graphite (its resistance to oxidation) 
 increases in proportion to the temperature to which it has 
 been subjected. 
 
 4. This fact is clearly shown by the resistance opposed 
 by different graphites to transformation to graphitic oxide. 
 The difficulty of this oxidation increases in proportion to 
 the fusing point of the metal, from solution in which the 
 graphite has been derived. Similarly, a readily oxidizable 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 graphite, like that of Ceylon, becomes much more resistant 
 after heating. 
 
 Acheson, who is associated with the most modern methods 
 of treating carbon and certain of its compounds in the electric 
 furnace, patented, as early as 1895, a process of coke puri- 
 fication, which consists in subjecting it to the direct heat 
 of the current, thus volatilizing the impurities and leaving 
 the carbon in a practically pure state, and, in all probability 
 partially converted into graphite. In the same year, he 
 patented a method for the production of graphite by the 
 decomposition of silicon carbide (carborundum), that com- 
 pound being heated to such an extent as to volatilize the 
 silicon. Moissan has similarly produced graphite by the 
 high temperature decomposition of calcium carbide, em- 
 ploying a 1,200 ampere, 60 volt, arc furnace. 
 
 In 1899 Acheson, who may be regarded as an authority 
 with considerable practical experience of his subject, 
 published in the Journal of the Franklin Institute his 
 observations on the yield of graphite, and the conclusions 
 to be drawn therefrom. They run as follows 
 
 1. Comparatively pure petroleum coke produces prac- 
 tically no graphite. 
 
 2. Impure bituminous coal coke produces large quantities 
 of graphite. 
 
 3. The larger the percentage of impurities in the bitumin- 
 ous coal core, the greater the yield of graphite. 
 
 4. Only part of the carbon core of carborundum furnaces 
 is converted into graphite (that part which carries an 
 admixture of slate, or a high ash content). This conversion 
 is not increased, even by repeated use of the same grains in 
 successive furnaces. 
 
 The conclusions drawn by Acheson from these observa- 
 tions are 
 
 1. Graphite is the form carbon assumes when freed from 
 chemical associations, under conditions of low pressure 
 and protection from chemical influence. 
 
 47 
 
ELECTRIC FURNACES AND 
 
 2. Diamond is the form carbon assumes when freed from 
 chemical associations, under the conditions of high pressure 
 and protection from chemical influence. 
 
 And, by " inference " 
 
 3. Amorphous carbon is the form carbon assumes when 
 freed from chemical associations, under conditions of high 
 or low pressure and exposure to chemical influence. 
 
 It will have been noted, from the foregoing context, that 
 these various investigators are far from being of one accord 
 as regards the theory of the formation of graphite. Com- 
 menting upon this, and basing his deductions upon the fact 
 that Moissan, with an arc furnace, succeeded in converting 
 all forms of carbon into graphite, whereas Acheson, working 
 with his well known resistance type of furnace, failed to 
 effect such conversion in the case of pure carbon, Townsend 
 suggests that the preliminary ionization of the carbon is 
 essential to the formation of graphite ; he thus advances 
 his theory 
 
 " The essential condition for the formation of graphite 
 may be considered to be the presence of carbon in the 
 ionized state, and its separation therefrom. This, so far 
 as we know, may be accomplished by any one of four 
 methods 
 
 "1. By solution in metals or carbides. 
 
 " 2. By the direct ionizing action of the electric discharge, 
 and notably the arc, upon carbon. 
 
 "3. By the action of the electric discharge upon gaseous 
 carbon compounds. 
 
 "4. By the dissociation, by heat, of certain carbon 
 compounds, notably the carbides." 
 
 Borchers and Mogenburg, investigating the conversion 
 of amorphous carbon into graphite by electric furnace 
 methods, have also found the process, with pure carbon, 
 very difficult of accomplishment. The conversion is, 
 however, facilitated by the presence of metals or their 
 compounds, notably aluminium. These combine with the 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 carbon to form carbides, which are subsequently decomposed, 
 and the carbon released from combination in the form of 
 graphite. The quantities of metal thus required are very 
 small, and aluminium has been found most suitable and 
 effective for the purpose. 
 
 In the course of his researches into the manufacture of 
 artificial graphite, Acheson has demonstrated that the 
 carbide forming impurities need not be present in propor- 
 tion sufficient to react at once with the whole of the carbon 
 to be converted, but that the graphite formation may be of 
 a progressive character, small quantities of carbide being 
 formed, decomposed, and the volatilized impurity again 
 combine with an adjacent portion of carbon, to repeat the 
 cycle of transformation. 
 
 He also found that the introduction of additional carbide 
 forming impurities, in the case of non-coking coal, and 
 certain kinds of charcoal, was unnecessary, in that they 
 contained sufficient mineral impurities in themselves to 
 effect their conversion into graphite, when raised to the 
 required temperature. 
 
 It is not essential that the preliminary mixing in of the 
 impurities with the mass of carbon be thorough, as the 
 metallic vapours formed during the process thoroughly 
 permeate the entire mass. 
 
 According to Blount (Paper read before Manchester 
 Section I.E.E., March 4, 1902) there are three methods 
 available for the artificial manufacture of graphite in the 
 electric furnace. 
 
 1. By direct conversion at a very high temperature. 
 
 2. By crystallization, on cooling, from supersaturated 
 solution of carbon in some suitable metal (e.g., aluminium, 
 manganese, nickel, chromium, tungsten, molybdenum, 
 uranium, zirconium, vanadium and titanium). 
 
 3. By crystallization of dissolved carbon from its metallic 
 solution, by the addition of an element, such as silicon or 
 boron, capable of displacing it. 
 
 49 
 
 J^\ \ & R A ft rSfe 
 u or THIE 
 
 ( UNIVERSITY ) 
 
ELECTRIC FURNACES AND 
 
 Blount, in speaking of Acheson's method of manufacturing 
 artificial graphite, criticizes the wording of his patent claims, 
 which nominally cover the production of graphite by the 
 decomposition of a carbide, and points out that, if put to 
 the test in a patent litigation case, they would probably 
 prove untenable, in that they refer to processes discovered 
 and made public many years ago. In this connexion it 
 may be mentioned that Acheson's own theory to account 
 for the conversion of amorphous carbon into graphite at 
 high temperatures is that it follows from the decomposition 
 of a carbide, which is formed in the first instance, the carbon 
 of the latter being deposited in the form of graphite. 
 
 A good mixture for the production of artificial graphite 
 is said to consist of 97 parts finely divided amorphous carbon 
 mixed with three parts iron oxide. It is suitable for 
 moulding into any desired form previous to graphitizing. 
 
 The output of artificial graphite in America was 81 tons 
 in 1897 ; in 1901, it had increased to 1,200 tons. 
 
 The total yearly production of the International Acheson 
 Graphite Co. alone, in 1902, was said to be in the neigh- 
 bourhood of 500 tons. 
 
 Graphitizing Electrodes. The Acheson resistance furnace 
 for graphitizing electrodes, as employed at the works of the 
 International Acheson Graphite Company, Niagara Falls, 
 is of the usual type, as already described in connexion with 
 the manufacture of carborundum. The process of graphitizing 
 is similar to that already dealt with, and consists in causing 
 the pure carbon, under the influence of electric heat, to 
 combine with certain carbide forming impurities which are 
 present in the mass prior to the process. 
 
 The electrodes are first made up in the usual way, from 
 petroleum coke and pitch, like ordinary arc light carbons, 
 the only addition being a certain amount of carbide- forming 
 material such as silica, or iron oxide. They are baked in 
 the usual way, and then subjected to the graphitizing 
 furnace, in which process they are raised to a temperature 
 
 50 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 well above that of the volatilization of iron, aluminium, 
 or silicon. 
 
 The necessary temperature for the reaction is higher than 
 that required for the decomposition of the carbides formed, 
 and is produced in the furnace construction represented 
 in Fig. 12, where F is the outer rectangular structure of 
 fire-brick, through the end walls of which are introduced 
 the electrodes e e, which conduct the current to the charge C ; 
 this latter may either consist of cylindrical rods, as repre- 
 sented in transverse section in the figure, or rectangular 
 plates. In the former case the rods are, as shown, packed 
 transversely to the direction of current flow, the necessary 
 
 FIG. 12. 
 
 high resistance path from one to the other being estab- 
 lished along the lines of contact between the cylindrical 
 surfaces. 
 
 It is chiefly along these lines that the heat is generated, 
 and this mode of packing the articles to be graphitized 
 gives rise to the evolution of heat exterior to the articles, 
 rather than in the mass itself, as would be the case were 
 they arranged with their axes parallel to the line of direction 
 of the current. This constitutes the principal feature of 
 the invention, and effects considerable economy in the 
 current required for the operation of the furnace. 
 
 If the objects to be graphitized are rectangular, they are 
 arranged in a series of regular piles along the path of the 
 current, each pile being separated from its neighbour by 
 a filling of ground or pulverized coke. 
 
 The latter is also employed as a filling, in both cases, and 
 
 51 
 
ELECTRIC FURNACES AND 
 
 is introduced between the ends of the charge and the elec- 
 trodes. The base of the furnace is lined with a refractory 
 layer of carborundum R, or similar conducting material, 
 and the whole is covered in at the top with a layer of ground 
 coke and sand S. 
 
 One of the latest resistance furnace methods patented 
 by Acheson relates to the manufacture or baking of 
 moulded carbon articles, such as electrodes, arc-light 
 carbons, etc. To this end they are embedded in a heating 
 resistance composed of granular carbon, or a mixture of 
 carbon with silicon, etc. The articles to be baked are 
 placed with their longest dimensions at right angles to the 
 path of the current, as in the foregoing graphitizing furnace, 
 and the principle of the method consists in confining the 
 evolution of heat to the resistance core of packing, rather 
 than to the articles themselves. A very uniform and easily 
 regulable heating effect is said to be obtained in this manner. 
 
 Coke Purification. It is necessary that the carbon used 
 in electro-metallurgical operations be pure, and, more 
 especially, free from silica and the compounds of silicon, 
 which would, if present, combine with the metal under 
 treatment, and result in the re-introduction of like impuri- 
 ties into the finished article. 
 
 Mr. C. M. Hall, the patentee of the aluminium process 
 which bears his name, has devised and patented a form of 
 resistance furnace for ridding the coke of these undesirable 
 impurities before it is employed in connexion with the 
 extraction of metals. 
 
 The process consists in powdering the coke, and mixing 
 it intimately, whilst thus powdered, with a metallic fluoride, 
 such as sodium fluoride, cryolite, or fluor-spar. This 
 mixture, when subjected to heat in a suitable furnace, 
 becomes pure carbon, owing to the reaction set up between 
 the fluoride and the compounds of silicon, which combine 
 to form silicon fluoride, the latter being driven off in the 
 form of gas. 
 
 52 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 By adding a suitable proportion of pitch to the mixture, 
 as a binding material, it can be moulded into electrodes of 
 any desired form, and the purification and baking process 
 carried out in one operation of the furnace. 
 
 The latter is built of brick, and filled with moulded carbon 
 blocks, packed symmetrically, and insulated from one 
 another by suitable packing. Through the centre of the 
 mass, and in close proximity to the blocks, runs a resistance 
 core with end electrodes. An alternating current is supplied 
 to this core by means of the electrodes, and the heat ad- 
 justed to any desired temperature, by varying the current. 
 
 The actual temperature is necessarily slightly higher than 
 that required for baking alone, in order to produce the 
 necessary thermo-chemical action for the purification of 
 the carbon. 
 
 Peat Coal. Of late years the problem of adapting peat 
 to the same fuel uses as coal has engaged considerable 
 attention, and several processes for manufacturing what is 
 known as " Peat Coal " have been evolved. In this con- 
 nexion the possibilities of electric heating have not been 
 overlooked. 
 
 The Jebsen process, invented by P. Jebsen, Norway, 
 and commercially exploited by him at Stangfjorden, has 
 been in successful practice for over three years. The peat 
 is first kneaded and compressed into rectangular blocks by 
 special machinery, after which it is subjected to a drying 
 process, by exposure to blasts of the hot gases produced 
 in the carbonizing retorts. These gases, which have an 
 initial temperature of from 90 to 100C. = 194 to 212F., 
 reduce the quantity of moisture present in the peat from 
 80 per cent, to 25 per cent, or thereabouts. 
 
 It is then ready to undergo the electrical carbonization 
 process, which is carried out in what is virtually a resistance 
 furnace with independent cores. Each furnace or retort, 
 consists of an iron cylinder, about six feet long, by three 
 feet in diameter, with removable end covers. These 
 
 53 
 
ELECTRIC FURNACES AND 
 
 retorts are mounted in a vertical position and lined interiorly 
 with asbestos or fire-brick. 
 
 The heating resistances are in spiral form, supported 
 from the walls of the retort, and so disposed that the peat 
 under treatment, when packed into the cavity, comes into 
 intimate contact with them. The disposition is such that 
 practically all the heat evolved is utilized and absorbed by 
 the peat, the loss, by radiation, being, to all practical 
 intents and purposes, nil. Experience has determined the 
 time and current necessary to effect the carbonization process, 
 after which the current is switched off, and the retort, with 
 its charge, allowed to cool. The lower cover is then removed, 
 and the carbonized peat allowed to fall into a truck placed 
 to receive it. 
 
 The temperature required to thoroughly effect the car- 
 bonization is from 400 to 500C.=752 to 932F. By 
 lengthening the retorts, the process can be made continuous, 
 the dried peat being continuously fed in above and the car- 
 bonized product removed, periodically, from below. 
 
 The Jebsen process of peat coal manufacture results in 
 several marketable by-products, among which may be men- 
 tioned peat oil, paraffin, gas oil, coke, and gas, the latter 
 being utilized, as already stated, in the preliminary drying 
 of the peat. The peat coal produced by this process has a 
 heating value of 7,500 calories, burns with intense heat, and 
 produces very little ash, or soot. Samples, subjected to 
 analysis at the Royal Norwegian High School, in Christiania, 
 showed the following composition 
 
 Carbon 
 
 Hydrogen 
 
 Oxygen 
 
 Nitrogen 
 
 Sulphur 
 
 Ashes 
 
 Humidity 
 
 76-91 per cent. 
 4-64 
 8-15 
 1-78 
 70 
 3-00 
 4-82 
 
 100-00 
 
 54 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Peat carbon, produced by the Jebsen process, is said to 
 be especially suited for calcium carbide manufacture. The 
 product of the process is, roughly, 33 per cent, charcoal, 
 4 per cent, tar, 40 per cent, liquor, and 23 per cent, gas, 
 the figures being based on the treatment of initially dried 
 peat. The carbon is very hard, and of a deep black colour. 
 The current necessary for the process is derived from five 
 80 k.w. generators driven by turbines of 128 h.p. 
 
 A process analogous to that of Jebsen, but not carried 
 to such a high degree of carbonization, has been invented 
 by Mr. Bessey, and was publicly demonstrated before a 
 committee of press representatives and technical experts, 
 at the works of Messrs. Johnson and Phillips, Charlton, 
 Kent, in 1903. 
 
 The peat is first placed in a centrifugal drying apparatus, 
 which drives out a considerable portion of the 80 per cent, 
 moisture which it contains. Electrodes are then introduced, 
 and the drying pushed to a higher degree by the resistance 
 process of electrical heating, the necessary initial conduc- 
 tivity of the mass being secured by the addition of certain 
 chemicals. 
 
 A company, known as the Electro-Peat-Coal Syndicate, 
 has since been formed for the purpose of exploiting this 
 process on a commercial scale, and it is estimated that peat 
 coal can be produced at a cost of 5s. per ton inclusive. 
 
 Carbon Bisulphide. The resistance furnace utilized in 
 the manufacture of carbon bisulphide at Penn Yan, N.Y., 
 U.S.A., is illustrated in Fig. 13. It is the invention of 
 Mr. E. R. Taylor, and is the outcome of considerable fore- 
 thought and experiment. In the original furnaces, inde- 
 pendent carbon electrodes were utilized as terminals, being 
 protected from a too rapid combustion by broken carbon, 
 which was fed into the furnace through orifices immediately 
 surrounding the electrodes. In the latest type of furnace 
 the carbon block terminals have been entirely dispensed 
 with, electrical connexion being made directly with the 
 
 55 
 
ELECTRIC FURNACES AND 
 
 carbon feed hoppers, through which the broken carbon is 
 introduced into the furnace. 
 
 In brief, the action is as follows : the upper, cylindrical 
 portion is filled with closely packed carbon C, through 
 
 which the sulphur vapours, 
 produced by the action of 
 electric heat upon the 
 fused mixture of carbon 
 and sulphur, F at the 
 hearth of the furnace, rise, 
 and in passing are con- 
 verted into carbon bisul- 
 phide, which is suit- 
 ably collected and con- 
 densed. 
 
 Terminal con n e x i o n 
 with the furnace is se- 
 cured through the carbon 
 hoppers M M, which are 
 four in number, situated 
 at equal distances around 
 the circumference of the 
 hearth, and through which 
 a constant supply of 
 broken carbon is fed. 
 
 The raw sulphur is made to perform a primary duty before 
 finally entering into combination with the carbon. It is 
 fed in cold, through hoppers H H, which convey it into 
 annular chambers, entirely surrounding the furnace body 
 and hearth : in these, the sulphur, whilst still in a cold state, 
 acts as a heat-conserving jacket, and only becomes molten 
 when it approaches the bottom of the annular chambers, 
 which communicate with the furnace hearth, and through 
 which the sulphur feed is effected. It thus reaches the 
 centre of activity in a heated condition, and no undue 
 lowering of the general temperature results. 
 
 56 
 
 Fid. 13. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 By varying the electrical connexions to the four terminal 
 hoppers, a very thorough and complete fusion of the carbon 
 and sulphur is effected in the region of the hearth. 
 
 The dimensions of these furnaces, as used at Penn Yan, 
 are : height, 41 feet ; diameter, 16 feet ; and they require 
 a current of 4,000 amperes, at from 40 to 60 volts, to operate 
 them. The regulation is, to a certain extent, automatic ; 
 as the temperature, and consequently the degree of fusion, 
 increases, more molten sulphur flows in to the bottom of the 
 hearth from the sulphur jackets, and the level of the molten 
 mass rises until it encircles the electrode hoppers, and gives 
 rise to an increase in the electrical resistance of the active 
 column, with a consequent decrease in the current until the 
 working conditions again become normal. 
 
 A later modification consists in locating the active heat 
 zone nearer to the furnace base, a procedure which renders 
 it possible to tap off accumulations of slag resulting from 
 impurities in the charge. 
 
 Early experiments in 1903 showed an output of 5,000 
 kgs. of carbon bisulphide per furnace in 24 hours, but this 
 figure has since been considerably increased. The output 
 from the Penn Yan furnaces, to June 1902, was 1,500 tons. 
 
 The carbon bisulphide furnaces at Penn Yan are the 
 largest electric furnaces at present in use. 
 
 57 
 
ELECTRIC FURNACES AND 
 
 SECTION IV 
 
 CALCIUM CARBIDE MANUFACTURE 
 
 General. Calcium carbide, expressed in chemical formula, 
 CaC 2 , is produced by heating together a mixture of 65 per 
 cent, lime, with 35 per cent, carbon, or coke. 
 
 A very high temperature, approximately 3,312C. = 
 6,000F., is required to effect the combination, the oxygen 
 of the lime being driven off, and uniting with a certain 
 percentage of carbon, to form monoxide (CO) and dioxide 
 (C0 2 ) of carbon. Some heat is necessarily evolved as a 
 result of the chemical combination pure and simple, but 
 it is insufficient to render the mixture self-heating ; con- 
 sequently artificial means, in the shape of the electric 
 furnace, have to be resorted to, in order that the mass may 
 be raised to the requisite temperature. 
 
 As early as 1862, Wohler prepared calcium carbide, by 
 heating an alloy of zinc and calcium in the presence of an 
 excess of carbon, but failed to isolate the compound. In 
 1893, Travers also succeeded in manufacturing calcium 
 carbide, though in admixture with other substances, by 
 heating a mixture of calcium chloride, carbon, and sodium. 
 jr; Moissan, however, must be credited with having been the 
 first to manufacture and isolate calcium carbide with the aid 
 of the electric furnace. He published his discovery on 
 December 12, 1892, in the form of a paper communicated 
 to the Academie des Sciences, which ran thus 
 
 " If the temperature (in the electric furnace) reaches 
 3,000C. 5,432F., the lime forming the furnace melts 
 
 58 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 and runs like water. At this temperature carbon quickly 
 reduces calcium oxide, and the metal is separated in quantity ; 
 it unites easily with the carbon of the electrodes to form a 
 carbide of calcium, liquid at a red heat, and easily collected." 
 
 At the end of 1892 Willson announced his discovery of 
 calcium carbide, which had been made independently, and 
 apparently without knowledge of Moissan's researches on 
 the subject. Thos. L. Willson, the original discoverer of a 
 commercial process of calcium carbide manufacture, was, in 
 1885, an employe of the Brush Company at Cleveland, 
 where the Cowles Bros, were then operating their electric 
 furnace. He subsequently experimented with the electric 
 furnace at Spray, N.C., U.S.A., where the first successful 
 production of calcium carbide was effected in 1888. 
 
 His attempts to patent its methods of production in the 
 electric furnace proved futile, and the only patents now 
 standing, which bear on carbide manufacture, either relate 
 to certain details in the design of the furnaces themselves, 
 or to such subsidiary processes as tend to promote the 
 efficiency of manufacture, e.g. the preheating of the raw 
 materials by suitable combustion of the reaction gases. 
 
 With reference to the early history of calcium carbide, 
 Vivian B. Lewes, Acetylene, tells us that in 1886-1887 the 
 lads employed at the Cowles' aluminium works were actually 
 playing with the substance, and experimenting with the 
 acetylene gas evolved on bringing it into contact with water. 
 Yet it was not until 1892, five years later, when its commercial 
 value and mode of preparation had been independently 
 studied by Willson and Moissan, that it became an industrial 
 product. 
 
 In France, M. Louis Michel Bullier claims to be the 
 orignal discoverer of crystalline calcium carbide, and his 
 claims have been upheld by the French Courts. The 
 Bullier patents belong to the Societe des Carbures Metal- 
 liques, a Parisian company founded by Bullier. In 1903, 
 this concern, which controls the output of 17 carbide fac- 
 
 59 
 
ELECTRIC FURNACES AND 
 
 tories in the Alpine districts, compelled its dealers to sell 
 at a uniform price of 14 per metric ton. The annual out- 
 put is estimated at 18,000 tons. 
 
 Commercial calcium carbide, in which iron is usually 
 present as an impurity, is a grey mass, with a metallic 
 appearance and a crystalline fracture. Its specific gravity 
 is 2*22, and its crystals, examined under the microscope, 
 are seen to be transparent, and of a reddish tint, due to the 
 impurities present. 
 
 When brought into contact with water, calcium carbide 
 decomposes rapidly into lime, and acetylene gas, which 
 latter is given off. The equation 
 
 CaC 2 +2H 2 = Ca (OH) 2 +C 2 H 2 , 
 
 represents the reaction which takes place. Acetylene has 
 also been produced directly in the arc furnace by Berthelot, 
 who succeeded in bringing about the direct union of carbon 
 and hydrogen in the requisite proportions. 
 
 M. Gin has made an extensive study of the reactions which 
 take place in a carbide furnace during the formation of 
 calcium carbide. Regarding the furnace contents as a 
 series of layers, in which the temperature rapidly decreases 
 from the interior to the exterior, he finds that, in these 
 layers, different chemical equilibria are formed, the reaction 
 being more endothermic the higher the temperature. The 
 gases evolved in different parts of the furnace contain both 
 free oxygen and free calcium. The former is liberated 
 at the points of maximum temperature, or, in other words, 
 its quantity increases with the current density. Calcium 
 vapours, on the other hand, originate at points of lower 
 temperature. The presence of free oxygen sets up combus- 
 tion of the electrodes, immediately below the surface of 
 the charge, whilst the calcium is deposited as the fine dust 
 so familiar to carbide workers. 
 
 Gin explains the presence of these two elements in a free 
 state, by referring to the researches of Berthelot, who proved 
 
 60 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 that carbon monoxide is dissociated at high temperatures ; 
 in the majority of commercial formulae for carbide furnace 
 charges there is an excess of lime. The formation of 
 calcium carbide consists in simple substitution of carbon for 
 the oxygen of the latter. The carbon monoxide formed 
 as a secondary product of the reaction is, in the hottest 
 parts of the furnace, entirely dissociated, whilst calcium 
 vapours are evolved at the boundary surfaces of calcium 
 oxide, and calcium carbide. 
 
 In a resistance furnace, calcium carbide commences to 
 form at a current density of 1*2 amperes per square milli- 
 metre, through the core, whilst, if the density be increased 
 to 2 amperes per square m.m., the product assumes a fluid 
 condition. According to Vogel, at the latter density in 
 the core of a resistance furnace, fluid calcium carbide 
 commences to form rapidly in the immediate neighbourhood 
 of the core. This rapid formation, however, only extends 
 over a region of some three or four centimetres from the 
 core, the remainder of the charge, though incandescent, 
 remaining in a semi-solid condition, in which state it forms 
 a gradually increasing mass, disposed in shunt to the resist- 
 ance core proper, conducting a considerable portion of the 
 main current through itself, and thereby decreasing the 
 heating efficiency of the core. 
 
 The result is a mixture of true carbide, with a partially 
 converted remainder, which offers difficulties in the way of 
 economical separation. It is therefore necessary, in the 
 operation of carbide furnaces of this description with a single 
 core, that the furnace charge be periodically broken 
 up, or agitated, and the crust formation thereby pre- 
 vented. 
 
 To obviate this defect, it has been suggested, and, in 
 fact, a German patent (No. 120,831, February 8, 1900) has 
 been taken out on a resistance carbide furnace, of which 
 the core, in multiple, is disposed in the form of a gridiron, 
 which effectually prevents the formation of a superimposed 
 
 61 
 
ELECTRIC FURNACES AND 
 
 and uneconomical crust, whilst the carbide formed falls 
 away, by virtue of its own weight, to the hearth of the 
 furnace, the raw materials descending from above to take 
 its place. 
 
 Of such a gridiron formation of five electrodes, 100 m.m. 
 deep, 15 m.m. thick, and 1,200 m.m. long, the total weight 
 is 13' 5 kgs. ; the cross-sectional area, 7,500 square m.m. ; 
 and the electrical resistance, 0*003 ohms. The current 
 required, at a density of 2 amperes per square m.m., is 
 15,000 amperes, at a pressure of 45 volts, or 1,100 (German) 
 H.P., of which 920 are utilized in the furnace. This disposi- 
 tion is secured by arranging three furnaces in series, the 
 1,200 m.m. of resistance core length, being equally divided 
 between them. In the three furnaces, with gridiron cores, 
 proportioned as above, there would be a total effective 
 heating surface of 13,800 square c.m., as compared with 
 675 square c.m. in a three-electrode arc furnace with square 
 carbons, each of 150 m.m. side. 
 
 Assuming a yield of 0'2 kg. of carbide per H.P. hour, the 
 expenditure of 900 H.P. would produce 180 kgs. per hour 
 in each furnace, or a total of 540 kgs. 
 
 MM. Vulitch and d'Orlowsky, of Paris, suggest the pre- 
 paration of calcium carbide by first fusing the lime, per se, 
 in an arc furnace, and then pouring it into an excess of some 
 heavy hydrocarbon, such as the residue from petroleum 
 distillation, in the body of which the actual carbide forma- 
 tion takes place. The product is loose and non-coherent, 
 and saturated throughout with the hydrocarbon, which 
 renders it practically non-hygroscopic and adapted for the 
 intermittent generation of acetylene, evolution of the latter 
 only taking place during actual immersion in water. 
 
 In the manufacture of calcium carbide, as in other electric 
 furnace processes, the tendency has been to revert to the 
 more moderate and easily regulable " resistance " principle 
 of working. It is, however, questionable whether calcium 
 carbide has ever been manufactured, commercially, without 
 
 62 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the development of an arc or arcs, within the mass of ! 
 material bridging the electrodes. 
 
 As M. Pradon points out, the presence of vapours of 
 calcium and carbon in the neighbourhood of the electrodes 
 reduces the E.M.P. necessary for the striking of an arc to 
 from 8-10 volts, a pressure at which, owing to the imprac- 
 ticable increase in cross-sectional area of the furnace elec- 
 trodes and connexions, it would be impossible to operate. 
 
 It is, of course, possible to smother the arc by judicious 
 feeding of the raw materials, but the constantly changing 
 nature and irregular movement of the charge permit it to 
 strike again. 
 
 Bearing these facts in mind, it would seem probable that 
 the whole commercial process of carbide manufacture 
 depends upon a combination of the arc and resistance 
 principles of furnace working, which operate irregularly, 
 and irrespective of the design of the furnace itself, in so far 
 as the latter may be intended to confine the reaction to one 
 or the other of the two principles of operation. 
 
 The calcium carbide of commerce cannot be purified by 
 any of the usual methods adopted with fusible and soluble 
 products, in that it is, to all practical intents and purposes, 
 infusible, and insoluble in any known solvent. It remains, 
 therefore, to produce the carbide in as pure a state as possible, 
 such that subsequent purification is rendered unnecessary. 
 
 There are, at present, two known methods of achieving 
 this result, both of which consist in the addition of foreign 
 materials to the charge of carbon and lime undergoing 
 treatment in the furnace, which, either by combination 
 with the impurities, or reacting upon them during the pro- 
 cess, render them innocuous or convert them into such a 
 form as permits of their ready removal from the finished 
 product. 
 
 Hewes' method consists in employing, as a furnace charge, 
 a mixture of 26 parts carbon, 64 lime, 8 calcium carbonate, 
 and 2 peroxide of manganese. The function of the latter 
 
 63 
 
ELECTRIC FURNACES AND 
 
 is to form carbide of manganese (MnC 3 ), which reduces the 
 temperature of the reaction, and tends to increased fluidity 
 of the product. Methane (CH 4 ) is also produced, and serves 
 as a diluent, lessening the tendency to a sooty deposit from 
 the acetylene subsequently produced from the carbide. 
 
 The object of adding calcium carbonate to the mixture 
 is to produce a generous yield of gas, which, bubbling up 
 through the carbide, carries with it such impurities as the 
 sulphides and phosphides of calcium to the surface, leaving 
 them there in the form of a crust, which is readily separated 
 from the charge when the furnace is opened. 
 
 Dr. W. Rathenau's process is specially adapted to the 
 getting rid of silicious impurities in either the lime or carbon. 
 He adds to the charge a metal, preferably iron, or its oxide, 
 in sufficient quantity to take up and combine with all the 
 silicon present as impurities. The result is the formation 
 of iron siticide, which collects at the bottom of the furnace 
 below the carbide, and may be drawn off, or subsequently 
 separated from the charge, as circumstances may deter- 
 mine. 
 
 Silicide of iron, or ferro-silicon, thus produced, constitutes 
 a useful by-product and may be marketed for various 
 purposes. 
 
 Temperature of Formation. There has been considerable 
 discussion as to the temperature of formation of calcium 
 carbide. Several experimenters have tackled the subject 
 with a view to ascertaining the exact degree of heat neces- 
 sary ; among others, Rothmund and Borchers (Zeitschrift 
 fur Elektrochemie, May 29, 1902). The former proved, by 
 experiment, that the lowest temperature at which calcium 
 carbide can be formed is 1,620C. = 2,948F. Below this 
 temperature, again, at 1,560C.= 2,840F. ; and in the 
 presence of carbon monoxide gas, the carbide already formed 
 is disintegrated, or split up, into its constituent elements. 
 
 Borchers, who conducted his experiments with the aid 
 of blow-pipe heating, found that a temperature of approxi- 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 mately2,000C. = 3,632F.,was necessary for the formation 
 of crystalline calcium carbide, from which it will be seen 
 that although, theoretically, only 1,620C. is necessary, a 
 temperature of at least 2,000C. is required in practice. 
 
 Quality. The commercial quality of calcium carbide is 
 determined by its capacity for evolving acetylene gas. 
 One kilogramme of pure calcium carbide evolves 348' 4 litres 
 of acetylene, the latter being determined at a pressure of 
 760 m.m., and a temperature of 0C. = 32F. This is 
 equivalent to 5' 587 cubic feet of gas per pound of carbide. 
 The calcium carbide of commerce seldom yields more than 
 300 litres per kg., the average figure being between 280 and 
 290, or 86 per cent. 
 
 Current Used in Calcium Carbide Manufacture. Either 
 continuous or alternating current may be used in the pro- 
 duction of calcium carbide, and an alternating plant is, 
 in many cases, preferable, in that it lends itself more readily 
 to long distance transmission, and is, withal, better capable 
 of withstanding the strains incidental to large variations 
 in the load, such as are invariably encountered in electric 
 furnace working. 
 
 An interesting article by M. J. Pradon, in the Zeitschrift 
 fur Calciumcarbid Fabrication, 1901, entitled " The In- 
 fluence of the Nature of the Electric Current on the Manu- 
 facture of Calcium Carbide," throws some interesting 
 light on many debatable points connected with this exten- 
 sive application of electric furnace methods. 
 
 Discussing the probability of electrolytic action taking 
 place in a carbide furnace when continuous currents are 
 employed, M. Pradon discounts the greatly magnified 
 drawbacks incidental to continuous current working. He 
 states, however, that, under exactly similar conditions, and 
 given identical charges of raw material, the carbide pro- 
 duced in a continuous current furnace is of a lower gas- 
 producing quality than that resulting from the action of an 
 alternating current ; the former product is of a compact 
 
 65 F 
 
ELECTRIC FURNACES AND 
 
 nature, which offers considerable resistance to fracture, 
 whilst the latter is crystalline, and breaks readily. 
 
 Another interesting fact which has been noted in connexion 
 with the operation of both types of furnace is, that in the 
 case of direct current it is necessary that the preliminary 
 mixing of the charge be thorough, in order to ensure homo- 
 geneity, whereas, with an alternating current furnace, a 
 very casual mingling of the constituents suffices. 
 
 In connexion with the use of alternating currents for 
 carbide furnace work, M. Pradon lays considerable stress 
 on the lag due to self-induction, and, to obviate losses from 
 this cause, he has designed a furnace operating on the " con- 
 centric " principle. 
 
 It consists of a movable hearth, in the form of a truck 
 mounted on wheels, which receives and retains the carbide 
 formed, and can be removed and replaced when full. When 
 placed in position in the furnace, this truck is located below 
 the centre of a metal chimney or flue, which serves to conduct 
 away the gases of combustion. Down the centre of this 
 chimney passes the vertical electrode. The side walls of 
 the furnace are of metal, and one connexion from the alter- 
 nating current generator is made to the upper portion of 
 these walls, whilst the other is made to the lower extremity 
 of the chimney. 
 
 The path taken by the current is thus downward through 
 the metallic walls of the furnace, into the truck or hearth, 
 and up through the^ unconverted mixture therein to the 
 vertical electrode, through which it passes to the upper 
 extremity of the chimney, and down again, being thus, 
 through every portion of its circuit, subjected to the neutral- 
 izing effect of a return current. 
 
 An alternative method for the elimination of self-induction 
 losses in an alternating current furnace is to do away with 
 the arc and effect the conversion on the " resistance " 
 principle. According to M. Pradon, this plan is impractic- 
 able, as, owing to the presence of vapours of calcium and 
 
 66 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 carbon, an arc is struck at a very low voltage, even 8 to 10 
 volts being sufficient. To operate below this voltage would 
 entail an impracticable increase in the cross-sectional area 
 of the furnace electrodes and connexions thereto. 
 
 Raw Materials. As regards the raw materials used in 
 calcium carbide manufacture, certain advantages, such as 
 freedom from ash, and consequent contamination of the 
 carbide with silicon, magnesium, iron, and similar impurities, 
 are claimed for the carbon resulting from woody tissue, 
 which is utilized, notably in the three-phase Memmo 
 furnaces, to be described later. 
 
 In this connexion, a brief allusion to Prof. V. L. Emerson's 
 invention for the conversion of saw-dust and saw-mill 
 refuse into commercial products including a special form 
 of carbon for carbide manufacture, may prove of interest. 
 
 The saw-dust, having first been subjected to a drying 
 process, which is effected by the combustion gases derived 
 from its subsequent treatment, is next charged into a retort 
 consisting of a vertical iron cylinder, some 15 to 20 ft. 
 high, by 3 ft. in diameter, surrounded by brickwork. 
 
 Within the cylinder are a series of perforated hoods, 
 mounted on a central, hollow, revolving shaft, which is also 
 provided with perforations between the hoods. The 
 gaseous products from the distillation process which is 
 carried on in this apparatus make their way out at the lower 
 end of the cylinder, and, having passed through a purifying 
 process, are partly used in carbonizing the charge by being 
 forced through it as it passes down the retort, and, for the 
 rest, serve to dry the sawdust, as already stated. The 
 liquid products of the distillation are also collected and 
 separated, being subsequently drawn off through two out- 
 lets ; of these, one is a wood creosote product which can be 
 utilized for a variety of purposes, and the other, crude 
 pyroligneous acid, from which wood alcohol, acetic acid, 
 and various other products can be prepared. The carbon 
 remainder passes out through a separate opening, and is 
 
ELECTRIC FURNACES AND 
 
 said to be a very pure form of carbon, especially adapted 
 for carbide manufacture. 
 
 Experience has demonstrated that a carbide furnace 
 works at its best when fed with a raw mixture of carbon 
 and lime which has been reduced to granular form rather 
 than powdered. Extreme comminution of the charge is 
 inadvisable in that, setting aside the principal question of 
 the expense involved in thus reducing the ingredients to a 
 powder, there remains the additional disadvantage that, 
 when thus finely divided, they agglomerate round the 
 electrodes, forming an impermeable coating, which hinders 
 the free escape of the gases formed in the reaction. 
 
 The action which takes place in a carbide furnace in the 
 immediate neighbourhood of the electrodes, when fed with 
 a charge which has been too finely divided, is thus described 
 by Berger Carlson, in an article which appeared in Zeit- 
 schrift fur Elektrochemie, about the end of 1899. 
 
 Carbon monoxide gas is formed as a natural result of the 
 reaction, and by reason of the impermeability of the finely 
 powdered charge, forms a cavity round the electrodes, the 
 inner walls of which, from exposure to the radiant heat of 
 the arc, become glazed by the fusion of the lime and its 
 subsequent reaction upon the particles of carbon in its 
 immediate neighbourhood. The internally glazed cavity 
 thus formed still further encloses the reaction gases, with 
 the result that the pressure gradually rises, until the walls 
 of the cavit}^ give way, and allow a portion of the carbon 
 monoxide to escape, carrying with it a portion of the charge 
 in the form of dust. Channels are thus formed within the 
 general mass of the charge and, by virtue of the highly 
 heated carbon monoxide from the region of the arc passing 
 through them, also become glazed, and serve to conduct 
 away from the furnace a considerable quantity of heat which 
 would otherwise serve a useful purpose. 
 
 This constant honeycombing of the mass of raw material 
 and subsequent caving in of the recesses formed gives rise 
 
 68 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 to serious fluctuations in the load, by setting up constant 
 changes, having a considerable range, in the resistance of 
 the fused mass which bridges the electrodes. 
 
 In order to eliminate these various drawbacks, the same 
 authority advocates the use of a mixed charge, the particles 
 of which have been reduced to the size of hazel nuts, and 
 quotes statistics which go to prove that, whereas, with a 
 finely powdered charge, 3,000 kgs. are necessary for the 
 production of 1,000 kgs. of carbide, only 1,700 kgs. are 
 required for the production of a similar quantity when the 
 charge is of a coarse or granular nature. 
 
 It may be mentioned that the quantity of raw material 
 theoretically required for the manufacture of 1,000 kgs. of 
 calcium carbide, is 1,440 kgs. 
 
 Furnaces. The improvement of electric furnaces for 
 the production of calcium carbide, judging from the number 
 of applications for patents, seems to have been a favourite 
 theme for inventors. Thus, according to Kershaw (Electrical 
 Review, July 7, 1899), the number of applications for patents 
 on inventions for the production and utilization of calcium 
 carbide numbered, in 1896, 159 ; in 1897, 172 ; and in 1898, 
 124. 
 
 Many of the furnace constructions dealt with are most 
 ingenious, the principal aim on the part of the designers 
 being, apparently, continuity of furnace action, whereby 
 the plant may be worked continuously, day and night, 
 with a consequent gain in efficiency ; and auxiliary heating 
 devices, whereby the raw material is heated to a certain 
 extent, prior to its introduction into the furnace, thereby 
 involving a lesser expenditure of electrical energy than 
 would otherwise be necessary for a given output of carbide. 
 
 Blount, in his paper before the Manchester Section of 
 the Institution of Electrical Engineers, suggests a suitable 
 design for a carbide furnace, based on the following re- 
 quirements 
 
 1. The furnace should make carbide sufficiently fluid to 
 
 69 
 
ELECTRIC FURNACES AND 
 
 be tapped and run away from the sphere of action ; i.e. it 
 must be continuous in operation. 
 
 2. It should not have an exposed arc, and the heat 
 electrically generated should be conserved as much as 
 possible. 
 
 3. It should be so constructed that the containing vessel 
 is hardly affected by the high temperature necessary to 
 form carbide. 
 
 Basing his design upon the three conditions set forth 
 
 above, Blount suggests the 
 furnace construction illus- 
 trated in Fig. 14. The fol- 
 lowing is his own description 
 of the apparatus, which would 
 certainly appear to possess 
 several advantages over exist- 
 ing types. 
 
 " The furnace consists of a 
 fire-brick casing A, with a 
 magnesia lining B . The shape 
 is conical, and at the bottom 
 the furnace is contracted to 
 form a hearth for the fused 
 carbide. The tapping hole is 
 at the bottom of this con- 
 tracted part. The lower elec- 
 trode is a carbon plate C, and the upper electrode, a massive 
 carbon rod D, of circular section. 
 
 " The raw material is fed into the annular space between 
 the upper electrode and the magnesia lining in sufficient 
 quantity to enclose and smother the zone of highest tem- 
 perature. This feeding may be done mechanically, and 
 various devices suggest themselves, but it is probable that 
 hand-feeding will be at least as effective, because the 
 amount of labour required is not great, and the variation 
 of conditions is so large, that the greater elasticity of hand- 
 
 70 
 
 FIG. 14. 
 
THEIR INDUSTRIAL APPLICATIONS \ 
 
 feeding may be a positive advantage. For the safety of 
 the workmen it may be necessary to draw off the carbon 
 monoxide by a fan through a side flue, leaving the top of 
 the furnace open and uninterrupted for manipulating the 
 charge. In operating the furnace, a pool of carbide will 
 be formed and maintained of such depth as to cover the 
 hearth and to allow of tapping the pure and fairly fluid 
 product, but the depth of this pool will be kept small, and 
 special attention will be directed to maintaining its fluidity 
 by the means indicated below. It will be observed that 
 the cross-section of that part of the furnace immediately 
 surrounding the lower end of the upper electrode is con- 
 siderably greater than the cross-section of the hearth on 
 which the carbide collects. Thus the smaller section of 
 the column of carbide compensates for its conductivity 
 being greater than that of the raw materials ; the desired 
 temperature can be maintained, and the carbide will re- 
 main sufficiently fluid to be tapped." 
 
 Three-phase furnaces, as employed at St. Marcello 
 d'Aosta, enable a more evenly distributed heating effect 
 to be obtained, whilst, in practice, it has been found that 
 the highest output of carbide is obtained from the inter- 
 mittent type of furnace. 
 
 Kicardo Memmo, commenting on calcium carbide manu- 
 facture in Italy, sets down the following conditions as 
 essential to the commercial success of the industry 
 
 1. The price of the plant, per horse power installed, 
 including machinery, furnaces, and all the necessary supplies, 
 must not be higher than 500 francs =20 approximately. 
 
 2. A certain ratio between the number of tons produced, 
 and the number of effective H.P. produced, must exist. The 
 price of calcium carbide, including general expenses, must 
 not be more than 120 francs =4 15s. per ton. 
 
 3. The capital expended must be liquidated in ten years. 
 This corresponds to a sinking fund of 40 francs =1 12s. 
 per ton produced. 
 
 71 
 
ELECTRIC FURNACES AND 
 
 4. The selling price of the carbide must be 250 francs 
 
 =10 per ton, so that, including freight, loss, profits, etc., 
 
 the price to consumer must not exceed 300 francs =12 
 
 per ton. On the capital invested, at least 6 per cent. 
 
 must be paid. 
 
 The Willson carbide furnace, as originally constructed, 
 is represented in Fig. 15, and consists of an outer struc- 
 ture F of fire-brick, with a lining of carbon C, which 
 
 constitutes the crucible, or 
 hearth of the furnace, and 
 was covered in by a lid L, 
 also of carbon, through the 
 centre of which passed the 
 upper carbon electrode E, 
 adjustment being effected by 
 means of a hand-wheel W 
 and screwed rod R, attached 
 to the clamp D supporting 
 the carbon. T is a tapping 
 hole, for drawing off the car- 
 bide formed, in a molten 
 state. Incidentally, it may 
 be recorded that this furnace 
 was not originally designed 
 FIG. 15. for carbide manufacture, but 
 
 was intended for the reduc- 
 tion of aluminium, and was found unsuitable for the pur- 
 pose. 
 
 The charge was introduced in small quantities at a time, 
 the electrode E being gradually raised, meanwhile, until 
 a layer of molten carbide was produced at the bottom of 
 the carbon crucible C, which then took the place of the 
 second electrode. 
 
 This was followed by an improved form of the Willson 
 furnace, in which a removable truck, mounted on wheels, 
 took the place of the fixed carbon-lined crucible, and was 
 
 72 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 provided with a massive cast-iron base, which serves as 
 the second electrode. This modified furnace was not 
 hermetically sealed, the whole of one side being merely 
 closed by iron doors whilst the furnace was in action. The 
 free circulation of air possible with this design led to con- 
 siderable loss and trouble, arising out of the waste of 
 material and rapid consumption of the upper electrode ; 
 it was, therefore, still further improved, prior to its adoption 
 for carbide manufacture at Foyers. To this end two sides 
 of the truck hearth were made removable, and consisted 
 of massive cast-iron bars, each provided with a central 
 opening for the escape of the carbon monoxide gas. This 
 horizontal diffusion of the gases produced in the reaction 
 was found to be an improvement upon the more usual 
 vertical arrangement. 
 
 The writer understands that even this third stage in 
 the development of the original Willson furnace has not 
 proved altogether satisfactory to the British Aluminium 
 Company who control the manufacture of calcium carbide 
 at Foyers, and that a new continuous type of furnace, 
 yielding 2,500 kgs. of 80 per cent, carbide per 1 E.H.P. 
 year, has since been installed. No particulars concerning 
 this later furnace are, however, available. 
 
 The Union Carbide Company, of Niagara, U.S.A., owns 
 and works the Willson furnace patents in America. It 
 is the only concern manufacturing calcium carbide in the 
 United States, and its capacity is in process of extension 
 to 100 tons per 24 hours. The cost of manufacture is 
 stated to be in the neighbourhood of 5 per ton, and the 
 product is graded into three sizes, varying from 3J in. 
 to 2 in., 2 in. to J in., and \ in. to ^ in. respectively. 
 Rotary continuous furnaces are employed, and the dust is 
 worked up afresh with each charge. 
 
 The Gin and Leleux carbide furnaces, as used at Meran, 
 in the Austrian Tyrol, are semi-intermittent in action, 
 but disposed in pairs, and furnished with an ingenious 
 
 73 
 
ELECTRIC FURNACES AND 
 
 overhead arrangement, whereby a single upper electrode 
 is made to serve for either furnace in turn. By thus 
 working one furnace whilst the other is cooling and being 
 charged afresh with raw material, a practically continuous 
 operation is secured. 
 
 In general principle, the furnace resembles the modified 
 Willson type, with removable truck hearth. There is, 
 however, an additional feature in connexion with these 
 furnaces which renders their mode of working distinct 
 from that of the Willson type. It consists in a revolving 
 fan. which extracts the gases produced in the reaction 
 from the furnace chamber by suction, and passes them on 
 to be utilized in a preliminary heating of the next charge ; 
 a considerable saving in the upper carbon electrodes is 
 thereby effected as the gases are withdrawn, or, at all 
 events, diluted with the ingoing air, to such an extent that 
 combustion is prevented. 
 
 At Meran these furnaces have each a capacity of 260 
 k.w., in three phases of 2,500 amperes at 33 volts. Each 
 furnace operation lasts from 10 to 12 days without inter- 
 mission, the only stoppages necessary being for the replace- 
 ment of the upper electrode. The carbide is tapped from 
 the furnaces in a molten condition, through two orifices, 
 or nozzles, so designed as to prevent splashing. When 
 these orifices become finally choked, as they do after several 
 days, the remainder of the operation is performed, and 
 the furnace filled, by raising the upper electrode until the 
 chamber is charged to its utmost capacity. The furnace 
 is then cooled and opened, and the block of carbide formed 
 removed. 
 
 Five kgs. of carbide per k.w. day of 24 hours is the 
 output claimed for these furnaces. 
 
 In 1900, according to M. Keller, this figure had been 
 increased to 6*2 kgs. of carbide per k.w. day, representing 
 a thermal efficiency of about 75 per cent. 
 
 The Gin and Leleux furnaces as employed at Milan, 
 
 74 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Italy, and taking their energy from the Paderno three- 
 phase sub-station, are each of 225 k.w. capacity, and work 
 at a terminal E.M.F. of 28-33 volts. They are run in sets 
 of three, star connected, one in each leg, or phase, of the 
 supply system. The hearth of each furnace, which is in 
 two layers, of high and low resistance respectively, con- 
 stitutes one electrode, the other being arranged axially, 
 and capable of vertical feed adjustment. 
 
 The high resistance layer of the hearth, heated by the 
 passage of the current, serves to maintain the carbide in a 
 state of fusion. 
 
 Graphite, or a mixture of graphite and anthracite, is the 
 substance from which the electrodes are manufactured, and 
 they are subjected to a baking heat of 1,200C. = 2,192F. 
 for ten hours before use. They last about 150 hours. 
 
 Both the lime and coke used at Milan contain at least 
 85 per cent, lime and carbon respectively, and are mixed, 
 prior to being fed into the furnaces, in the proportion of 
 64-68 parts coke to 100 lime. The resultant carbide is 
 run off every two hours in quantities varying from 115- 
 175 kgs. for each furnace. Each run lasts 80 hours, the 
 final yield being from 1,100-1,350 kgs. of carbide. 
 
 The mean yield at this factory in 1900 was estimated at 
 5' 6 kgs. per k.w. day. 
 
 The furnace construction devised by William P. Parks, 
 of Chicago, was based upon the practicability of obtaining 
 the carbide in a molten state, and drawing it off at the 
 base of the furnace as formed, thereby rendering the pro- 
 cess continuous. 
 
 A section of the furnace, as originally patented, is shown 
 in Fig. 16. It comprises a cylindrical structure F, con- 
 tracted half-way down as shown to form a hearth, which, 
 in the shape of a circular block of carbon C, serves, at the 
 same time, as the lower fixed electrode. 
 
 In the upper face of this block, which is flush with the 
 contracted portion of F, an annular channel c is cut, and 
 
 75 
 
ELECTRIC FURNACES AND 
 
 led into a tapping opening t, its object being to collect the 
 molten carbide formed, and convey it out of the furnace. 
 
 The upper movable electrode 
 is in the form of a hollow 
 carbon cylinder B, its lower 
 extremity, where it penetra/tes 
 into the furnace, being sub- 
 divided by a series of radial 
 slots as shown ats. The arc is 
 thus split up into a number 
 corresponding with the number 
 of slots, and the heating effect 
 thereby distributed throughout 
 the charge instead of being con- 
 fined to a single point, as 
 would be the case were the 
 extremity of the electrode left 
 solid. T is a feeding tube, into 
 which the raw material is fed 
 by a screw conveyor S from a 
 feed hopper H, and passes 
 down the centre of the elec- 
 trode B into the zone of 
 
 greatest activity. In passing down the tubular electrode it 
 receives a preliminary heating from the hydrocarbon burners 
 n n, which are introduced through the side walls ; a con- 
 siderable economy in the current, which would otherwise 
 be required to bring the charge up from the normal tem- 
 perature at which it enters the hopper, is thereby effected. 
 The resistance furnace for the manufacture of calcium 
 carbide, invented by Mr. Hudson Maxim, is of an inter- 
 mittent type, and produces the carbide in " block " or 
 " ingot " form instead of the semi-molten state aimed at 
 by many inventors. 
 
 It consists (Fig. 17) of a rectangular, refractory chamber, 
 R, provided with means for the introduction and with- 
 
 FIG. 16. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 drawal of the truck T, which constitutes the hearth of the 
 furnace, and, running on rails, is easily removed, when full 
 of carbide, and can be replaced by a similar empty truck, 
 without any serious delay in the process of manufacture. 
 It is provided with a refractory lining, B, and, through per- 
 forations in one of its side walls pass a series of horizontal 
 pairs of electrodes, E, E2, E3, whilst the opposite side is 
 slotted to permit the passage of the plates of carbide, p p, 
 which serve as temporary short-circuiting pieces, to start 
 the action of the furnace. The electrodes, E, E2, etc., 
 take the form of 
 hollow carbon 
 cylinders, each 
 mo unted in a 
 metal sleeve, or 
 combina t i o n of 
 sleeves, s, the in- 
 tervening sp a c e 
 between the metal 
 of the sleeve and 
 the carbon elec- 
 trode, being filled 
 in with ground car- 
 bon to ensure good 
 
 electrical connexion. Through the tubular centre of the 
 electrodes a supply of pulverized carbon is fed towards the 
 interior of the furnace, and serves to protect their active 
 extremities. The raw material is fed into the furnace 
 through the central opening C, its rate of admission being 
 regulated by the sliding shutters t t, whilst a preliminary 
 heating of the charge is secured from an outer flue, kept hot 
 by an oxy-hydrogen burner b. The action of the furnace 
 is started by short-circuiting the various horizontal pairs 
 of electrodes by means of the plates p p ; once started, the 
 action is maintained through the carbide of the charge 
 itself, and, by gradually withdrawing the electrodes E, E2, 
 
 77 
 
 FIG. 17. 
 
ELECTRIC FURNACES AND 
 
 etc., whilst at the same time feeding in the raw material, a 
 block of carbide is formed which extends practically the 
 whole width of the truck. When this has been accom- 
 plished, the electrodes are entirely withdrawn, the truck 
 and its contents removed, and another empty one inserted 
 in its place, the same process being then repeated. 
 
 A carbide furnace, 
 designed and patented 
 by Messrs. J. Zim- 
 merman and I. S. 
 Prenner, poss esses a 
 novel distinction in the 
 matter of feeding ar- 
 rangements for the raw 
 material, which, instead 
 of being introduced into 
 the heat zone of the arc 
 from above, as usual, is 
 carried up by a screw 
 conveyor, from below 
 and extruded just below 
 the active extremities of 
 the electrodes. The feed 
 principle may be likened to that of a fountain. 
 
 The furnace is represented in Fig. 18 ; the mixture of 
 carbon and lime is fed into the lateral hopper H, and passes 
 thence, down the inclined conduit C, to the bottom of the 
 screw conveyor A, the discharge orifice O of which is of 
 smaller diameter than the conveyor itself, and, as shown, 
 is situated at the top, just below the arcing space between 
 the electrodes E E. In passing through the conveyor, 
 which is driven from an external source by the belt or chain 
 B, the charge undergoes a certain amount of compression, 
 exuding finally into the intensely heated area of the arc, 
 where it becomes converted into carbide, and passes, by 
 gravitation, down the annular space between the conveyor 
 
 78 
 
 FIG. 18. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 casing, and the inner wall of the furnace, into the receiving 
 truck T, which can be removed or replaced when full, thus 
 rendering the process continuous. 
 
 A patent, granted some time back to Dr. Hugo Koller, 
 includes an arc carbide furnace of a special type, adapted for 
 use with currents of higher voltage than it is customary to 
 employ for the purpose of carbide production. The furnace 
 itself takes the form of a long narrow conduit of refractory 
 material, the length of which is governed, to a certain extent, 
 by the voltage to be employed in the reaction. Massive 
 carbon electrodes are introduced through the end walls, 
 whilst a number of intermediate carbon blocks, spaced at 
 regular intervals apart along the length of the furnace, 
 and depending in number also upon the voltage, serve to 
 form a chain of inter-connecting bi-polar electrodes, the 
 space between each pair constituting a receptacle for the 
 raw material and carbide formed therefrom. 
 
 The intermediate blocks are raised on suitable supports 
 out of contact with the floor or hearth of the furnace, and 
 the latter is lined, along the intervening spaces, with shallow 
 carbon crucibles, or troughing, which assist in conducting 
 the current, and at the same time serve as receptacles for 
 the carbide. Only a small portion of the current is con- 
 ducted along these carbon troughs, which are not in actual 
 contact with the electrode blocks ; the main reactions 
 occur through the heat developed in the several arcs formed 
 between each neighbouring pair of blocks. 
 
 This form of furnace construction is, from the very fact 
 of its adaptability to circuits of comparatively high voltage, 
 especially applicable to large units for the commercial 
 manufacture of calcium carbide. 
 
 A carbide arc furnace, patented by J. H. Morley, has 
 the usual interchangeable hearths and automatic feed 
 arrangements, the distinctive novelty in its construction 
 being the upper carbon electrode, which takes the form of 
 a hollow rectangular conduit, descending vertically into 
 
 79 
 
ELECTRIC FURNACES AND 
 
 the furnace. The central aperture, or bore, of the carbon 
 is open to the atmosphere, and furnishes a species of flue, 
 or chimney, through which a downward current of air is 
 induced. This meets the active charge, at the very point 
 of greatest activity, and is instrumental in preventing over- 
 heating and explosions of gas produced in the reaction. 
 
 The process is rendered more thorough, and a more or 
 less continuous furnace action secured, by a reciprocating 
 motion communicated to the central carbon electrode. 
 In addition to the up and down motion thus imparted, a 
 certain turning movement is also given to the conduit, the 
 motion of which thus resembles the turn of a screw, and 
 tends to loosen the charge of carbon and lime immediately 
 surrounding it, and bring it within the heat zone created 
 by the arc. 
 
 We now come to a series of rotary furnaces, in which the 
 hearth, or container, for the raw materials under treatment 
 is constructed in circular form, and mounted in such a 
 manner as to permit of its being revolved like a wheel. 
 Continuity of action and a more homogeneous product are 
 claimed for this class of furnace. 
 
 The Nikolai principle of carbide furnace construction 
 provides for a continuous process, the furnace hearth being 
 revoluble, and the raw material fed in at one point, whilst 
 the carbide is withdrawn at another, the arc, or zone, of 
 activity intervening. 
 
 The furnace itself consists of a circular, refractory trough, 
 which constitutes in effect the tyre of a wheel, the latter 
 being driven at a suitable speed, by gearing. This wheel 
 and the attached revoluble hearth constitute one electrode 
 of the furnace, the other being fixed at a given point above 
 the trough, with sufficient clearance for the charge to pass 
 beneath it, through the heated belt formed by the arc. 
 
 The mixture of coke and lime is fed into the trough at a 
 regular rate from a point behind the fixed electrode, which 
 latter is protected by a heat-conserving hood, and, having 
 
 80 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 been carried under it by the revolution of the wheel, is 
 converted into calcium carbide by the arc. The carbide 
 thus produced then passes on, and is removed at a point 
 in advance of the fixed electrode by a scraper, leaving the 
 trough free for the reception of a fresh charge when it again 
 reaches the feeding point. 
 
 The Horry carbide furnace was designed by Mr. W. S. 
 Horry with a view to overcoming many of the obvious 
 defects of the earlier types of carbide furnace. Among 
 the principal drawbacks which the inventor set himself to 
 overcome are the intermittent nature of the furnace action, 
 which, in many cases, necessitates shutting down the plant 
 as soon as the furnace is full, in order to allow the contents 
 to cool sufficiently for removal. 
 
 Another defect arising out of the mode of adjustment 
 and general disposition of the electrodes is a constantly 
 changing internal resistance, and consequently varying 
 current, under which circumstances the mass of carbide 
 already formed is again in part subjected to the further 
 action of intense heat, tending to decompose it, whilst 
 other portions of the charge, out of the immediate zone of 
 heat, remain unconverted, thus leading to a somewhat 
 mixed, and by no means homogeneous product. 
 
 The constructional principle of the Horry furnace is re- 
 presented in diagram by Fig. 19, where H is an open 
 fire-brick hopper, containing the mixture of lime and coke, 
 and supporting at its lower or discharge extremity the 
 two carbon electrodes E E, which are bevelled at their 
 active edges, in order to present a vertical surface to 
 the arc. The length of the latter is thus maintained 
 constant, an important feature, which greatly facilitates 
 regulation. 
 
 The hopper discharges its contents between the flanges 
 of a rotatable metal drum D, which receives the carbide 
 formed in the arc, and retains it by virtue of removable 
 cover plates or battens B, which can be slid into place, or 
 
 81 G 
 
ELECTRIC FURNACES AND 
 
 removed, at will, As the formation of carbide goes on, 
 the drum is slowly rotated by means of the hand-wheel 
 and worm W, thus conveying the carbide, in a plastic con- 
 dition, away from the region of the arc. 
 
 When it reaches a point diametrically opposite to the 
 arc, the cover plates are removed, and the finished product 
 scraped out, leaving that portion of the drum free for the 
 
 reception of a fresh charge when 
 it again comes under the mouth 
 of the hopper. The cooling of 
 the product is thus effected grad- 
 ually as it gets farther and 
 farther away from the arc, the 
 heat not being entirely lost, but 
 in part communicated to the un- 
 converted portion behind. 
 
 The action need never be 
 stopped entirely, and the speed 
 of rotation of the wheel is regu- 
 lated in conformity with any 
 variations in the current con- 
 sumed by the furnace, which in- 
 dicates sufficiently how the con- 
 version is proceeding at the par- 
 ticular point of the drum which 
 is then under the hopper. The 
 carbide is removed in a plastic 
 state. 
 
 At Sault Ste. Marie a Horry carbide plant of 20,000 H.P. 
 has been installed. 
 
 The Bradley furnace, patented by C. S. Bradley in 1897 j 
 bears a striking resemblance to the Horry carbide furnace, 
 in that it consists of a slowly revolving wheel, which, by 
 bringing fresh materials into the region of the arc and 
 removing the resulting carbide therefrom, renders the pro- 
 cess continuous. 
 
 82 
 
 FIG. 19. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The arrangement consists of a wheel, with semi-circular 
 rim, forming one-half of a circular channel for the reception 
 of the raw materials and the carbide formed therefrom, 
 the other half being constituted by semi-circular sections, 
 or cover plates, which are latched, or otherwise fastened 
 on the descending side of the wheel as it leaves the region 
 of the arc, and removed again, on the ascending side, for the 
 withdrawal of the product and unconverted portion of the 
 charge. The wheel is revolved very slowly through the 
 medium of worm gearing, driven by an electro-motor ; a 
 structure, 15 ft. in diameter, may, according to the specifi- 
 cation, be revolved once in five days. 
 
 The arc is either created between a single carbon electrode, 
 4 in. in diameter, adjustably mounted on a stand, and 
 arranged to strike an arc with a temporary carbon core 
 embedded in the raw material, or may be arranged between 
 two oblique electrodes, operating at the mouth of the closed 
 portion of the rim. In the former case, electrical connexion 
 with the temporary core is secured through the medium 
 of copper bolts, a commutator, and fixed brushes. The 
 arc, once struck, is, of course, maintained between the 
 fixed electrode and a portion of the core, rendered conduct- 
 ing by the heat of the reaction. At the discharge point, 
 the cover plates, as before stated, are removed, and the 
 unconverted portions of the charge fall into a conveyor, 
 which returns them to the feed side of the wheel to again 
 undergo the heating process. 
 
 By treating a mixture of powdered sand, and the oxides 
 of calcium, strontium and barium respectively, silicides of 
 these metals have been produced in the Bradley furnace, 
 and it is suggested that they may prove of ultimate value 
 in the various dye industries from the fact of their being 
 powerful reducing agents. 
 
 Mr. C. S. Bradley has recently modified his rotary furnace 
 construction, the modification being embodied in a patent 
 taken out during 1903, which provides for lining the sides 
 
 83 
 
ELECTRIC FURNACES AND 
 
 and " belly " of the drum hearth with carbon or graphite 
 plates, thus enabling the area of thermal activity to be 
 extended throughout the entire mass of material under 
 treatment, without fear of consequent injury to the struc- 
 ture itself, a contingency previously avoided by a layer of 
 unconverted material. 
 
 A rotary carbide furnace construction has also been 
 patented by Hudson Maxim. It consists of a drum of 
 somewhat complicated construction, in which, according 
 to the claims, the fused carbide is held against the inner 
 surface, whilst any excess passes out through the open 
 end of the drum. 
 
 A somewhat ingenious, but seemingly commercially 
 impracticable form of construction for continuous carbide 
 furnaces was that embodied in a patent, the specification 
 of which, by Messrs. J. W. Kenevel, C. A. Spofford, and J. H. 
 Mead, is dated August 24, 1897. 
 
 In general principle the furnace may be likened to a 
 domestic mangle, or rolling mill, and consists of two cylin- 
 drical electrodes, which may be of carbon, between which 
 the raw mixture of coke and lime is fed from a suitable 
 hopper, the rate of feed being controlled by a slide. 
 
 Kerosene, fed on to the moving surfaces of the rollers by 
 suitably disposed tubes, acts as a lubricant, and prevents 
 the carbide from adhering thereto. The spindles of the 
 rollers are insulated from them by glass sleeves, and are 
 carried in sliding bearings, pressed together by springs, 
 after the manner of a mangle, and driven by gear wheels 
 carried on link pieces. 
 
 In passing between the electrode rollers, the mixture is 
 converted into carbide by the current, which is conveyed 
 to them through the medium of brushes and collector rings, 
 a somewhat difficult problem to tackle effectually, when 
 dealing with such large currents as are called for in electric 
 furnace methods. The carbide, when formed, drops below 
 on to baffles, placed to receive it. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Messrs. Lamoth and McRae, of Ottawa, are the inventors 
 of a resistance furnace for the production of calcium car- 
 bide, the two electrodes being bridged by carbon pencils. 
 
 M. Raoul Pictet has devoted considerable time and 
 attention to the study of carbide manufacture, and has 
 designed more than one furnace for its production. His 
 first practical experiment was conducted with a species of 
 compound furnace, whereby the necessary heat was obtained 
 in a series of three stages. 
 
 At the top was an ordinary blast furnace, consuming 
 coke as fuel, and into which the cold charge of coke and 
 lime was primarily fed. In this section, it attained a temper- 
 ature of 1,980C. 3,600F. It then passed on by gravitation 
 to a second section, where, under the influence o an oxy- 
 hydrogen blow-pipe flame, its temperature was still further 
 increased to 2,368C. = 4,300F. Finally, it passed into 
 the region of the electric arc, where a temperature of from 
 2,757 to 3,312C. = 5,000 to 6,000F. was reached, and 
 the combination of carbon and calcium effected. 
 
 In a subsequent design, the blast furnace and somewhat 
 costly blowpipe sections were dispensed with, the gradual 
 preliminary heating of the charge being effected instead 
 by the waste gases from the furnace, assisted by the natural 
 heat arising out of the chemical combination. 
 
 Two carbide furnaces of this description, constructed 
 under M. Pictet's patents, were built some few years back 
 at Ingleton, for the then Ingleton Carbide Company. 
 Each furnace required 2,000 amperes to work it, and con- 
 sisted of a hearth some 2 ft. square, by 3 ft. high, built of 
 bauxite bricks, cased on the outside with a layer of fire- 
 brick, and mounted on foundations of ordinary brick, 
 sufficient space being left below for the introduction and 
 withdrawal of a truck for the reception of the carbide 
 formed therein. 
 
 The carbon electrodes were square in section, with 6-in. 
 sides, the negative being mounted at an angle of 30 
 
 85 
 
ELECTRIC FURNACES AND 
 
 with the horizontal, whilst the positive was horizontal, and 
 adjustable for feed purposes, the latter being effected by a 
 hand wheel and screw motion. 
 
 Each electrode entered the furnace proper through a 
 sheet iron, water- jacketed gland, whilst the terminal con- 
 nexion itself, described separately in another portion of 
 the book, was also water- jacketed. 
 
 Water circulation was secured by means of high level 
 tanks, one to each electrode circuit. The electrical insula- 
 tion of the circulatory system was further secured by mount- 
 ing the water tanks on wooden bases, and by the introduc- 
 tion of 6-in. lengths of rubber tubing, in all pipe connexions, 
 both supply and waste. 
 
 An opening was provided in the bottom of the furnace 
 proper for the discharge of carbide into the truck placed 
 to receive it, whilst a flue, also in connexion with the dis- 
 charge orifice, conducted such heated gaseous products as 
 escaped with the carbide through a space between the 
 furnace linings to the top, whence an iron tube, enclosed 
 in a fire-brick flue, and inclined at an angle of 25, led up 
 from the furnace to a suitable point above, where the iron 
 tube penetrated the fire-brick wall of the flue, the latter 
 being bent at right angles, for the purpose, and continued 
 upwards to form a feed hopper or chute, down which the 
 mixture of coke and lime was fed in the form of compressed 
 briquettes. 
 
 The hot gases of combustion passed up between the 
 inner flue wall and the iron tube to the outer end, which 
 effected a junction with the chimney shaft. In their pas- 
 sage they served to impart a preliminary degree of heat to 
 the descending charge in the iron tube. 
 
 The furnace action was started, and a primary draught 
 created by an ordinary fire, which was lit within a space 
 provided for the purpose between the two furnaces. 
 
 The original Pictet furnaces at Ingleton have since been 
 replaced by Parker furnaces, described on page 94. 
 
 86 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Pictet estimates that by the economies consequent on 
 efficient preheating , of the raw materials in calcium carbide 
 manufacture, the cost can be reduced to 3 85. per ton. 
 
 Dr. Borchers' improved method of carbide production 
 involves several distinctly novel points, chief among which 
 may be mentioned the utilization of the excess heat evolved 
 for producing steam in an encircling boiler, or large water 
 jacket, which surrounds the furnace ; any excess of heat 
 over and above that required in the process of carbide 
 formation is thus utilized in the generation of steam, which 
 may serve in turn for driving an engine, thus effecting a 
 considerable saving. 
 
 This suggestion, of totally enclosing the furnace and its 
 contents in a water jacket, possesses also the attendant 
 advantage that any unconverted charge remaining at the 
 end of the operation can be withdrawn cool, and its ad- 
 mixture of carbon thereby preserved from the combustion 
 which usually ensues when the unconverted remainder of 
 an ordinary carbide furnace charge is exposed to the air. 
 
 The drawbacks incidental to existing types of carbide 
 furnace, which Dr. Borchers' improved construction is 
 intended to overcome, are 
 
 (1) Partial combustion of unconverted carbon when the 
 charge is withdrawn, and consequent contamination of 
 the remainder by ash, together with the necessary addition 
 of fresh carbon. 
 
 (2) Particles of dust are carried over by the resultant 
 gases and inconvenience the attendants, besides adding, 
 by their combustible nature, to the danger of fire. 
 
 (3) The amount of power consumed in carbide manu- 
 facture is abnormal. 
 
 In addition to the above novel feature of water-jacket- 
 ing the entire furnace, which is accordingly constructed 
 with very thin walls, Dr. Borchers proposes to work his 
 carbide furnaces on the following plan 
 
 In the resistance type, a heating core of carbon is packed 
 
 87 
 
ELECTRIC FURNACES AND 
 
 between two carbon electrodes, situated at opposite ends 
 of the furnace, and surrounded by a filling of lime. Cur- 
 rent is then switched on and heats the carbon column, 
 with the result that the layer of lime in immediate contact 
 with it is fused, and enters into combination with the carbon 
 to form carbide, which flows away to the bottom of the 
 furnace. A fresh layer of lime takes its place, and the 
 reaction is repeated, or rather continued, until the whole 
 of the carbon forming the resistance is used up, when the 
 current is switched off, and the resulting carbide, after 
 having been allowed to cool, removed. 
 
 In Dr. Borchers' arc furnace the electrodes are disposed 
 vertically, the upper descending axially through a hopper, 
 which also serves to feed the charge into the furnace, and 
 the lower one passing up to meet it, through a hinged trap- 
 door in the hearth or base of the furnace, which can be 
 opened at the end of the operation, for withdrawal of the 
 carbide formed. 
 
 To commence with, a mixed charge of coke and lime is 
 well rammed into the furnace around the electrodes and 
 the hopper, whilst the latter is also filled with a coarsely 
 ground charge of raw material, to such a depth as to pre- 
 vent the escape of the gases evolved in the reaction. In 
 the upper portion of the furnace, remote from the actual 
 heat zone, is located an inclined metal plate, perforated to 
 allow the passage of the furnace gases, and packed, above, 
 with a layer of wood shavings, or other suitable material, 
 to act as a filter on the gases passing through, and free 
 them from the dust and solid particles which they would 
 otherwise carry with them out of the furnace. 
 
 The dimensions of the furnace are such that when the 
 operation is completed, a block of carbide is formed in the 
 centre, surrounded on all sides by a layer of unconverted 
 charge, which remains non-conducting by virtue of the water- 
 cooled walls. This block is left to cool for a few hours before 
 withdrawal, the excess heat stored up in it being, as before 
 
 88 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 mentioned, absorbed by the surrounding water and thereby 
 utilized in the generation of steam for power purposes. 
 
 A patent taken out by Ricardo Memmo in 1897 relates 
 more especially to two improvements, continuity of furnace 
 action, and pre-heating of the charge, in the efficiency of the 
 process of carbide manufacture, and covers a form of construc- 
 tion in which the furnace proper is tubular, with a movable 
 graphite hearth. This latter is supported on an iron 
 foundation, which can be raised or lowered as required by 
 either a screw, or rack and pinion motion. To commence 
 with, the hearth is adjusted at a point, just below the upper 
 electrode, the raw mixture to be converted being fed in 
 by a screw conveyor, from a hopper, provided with a 
 refractory lining. 
 
 In order to effect the pre-heating of the charge, water is 
 included in the mixture supplied to the hopper, and forms 
 water gas, which is burnt, to heat the air supplied to a 
 circulatory system, which includes both the jacketed walls 
 of the furnace, and the hopper itself. 
 
 La Societa Italiana del Forni Electricci, Italy, about 
 this time, took out a patent for a carbide furnace, in which 
 the gases of combustion were similarly employed, after 
 purification and storage, to pre-heat the raw material, 
 which was delivered through an elliptical tube by a screw 
 feed motion. The electrodes consisted of carbon blocks, 
 one of which was adjustable by means of a lever to which 
 it was attached by a metal clamp, or plate. The furnace 
 was lined with carbon, and the combustion gases passed out 
 through orifices in the walls of the chamber to a suitable 
 collecting point, where they were duly collected, purified, 
 and stored, as described above, to be ultimately consumed 
 in pre-heating the charge before its entry into the furnace 
 proper. 
 
 The three-phase carbide furnaces designed and patented 
 by Ricardo Memmo have been in use at St. Marcello d'Aosta, 
 Italy, since 1897. Charcoal is employed in place of the 
 
ELECTRIC FURNACES AND 
 
 more usual coke to provide the carbon necessary to the 
 combination. Wood, from the abundant supply pro- 
 vided by the vast forests of the Alps, is converted into 
 charcoal in gas retorts, the resultant gas being again utilized 
 in heating other furnaces for baking both the lime and 
 mixture of the latter with charcoal. Charcoal and lime, 
 having been individually subjected to the baking process, 
 are next pulverized, and subsequently mixed in the correct 
 proportion for the production of calcium carbide. Small 
 quantities -of water and tar are added, to secure agglomera- 
 tion, and the mass is 
 formed, by a machine, 
 into bricks, which are 
 subsequently baked in 
 a furnace heated by the 
 gaseous products of the 
 electric furnaces. These 
 bricks constitute the 
 
 FIG. 20. 
 
 raw material with which 
 the furnaces are fed, 
 and it is worthy of 
 note that, treated in 
 this manner, little or no 
 dust, which gives rise to 
 considerable trouble in 
 carbide furnaces fed in 
 the ordinary manner with a pulverized charge, is formed. 
 
 The types of carbide furnace in use at St. Marcello, are both 
 intermittent and continuous in action. The former, Fig. 20, 
 is extremely simple in construction, and consists of a cubical 
 chamber A, built of refractory brick, cased externally with 
 ordinary brick B. The floor or hearth H is compressed 
 magnesium oxide, or lime, and an iron door, lined internally 
 with refractory brick, serves for the removal of the carbide 
 formed. 
 
 Above the door is a feed orifice, through which pro- 
 
 90 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 jects a cast-iron chute, down which the bricks are fed into the 
 furnace, whilst in the opposite wall is a flue F for convey- 
 ing away the gases formed in the conversion, to be ulti- 
 mately used in baking the bricks, as already described. The 
 three carbon electrodes C C C enter through the arched 
 roof R, which is also of refractory brick ; they are each 
 five inches in diameter, and are supported by holders, on 
 the extremities of large iron rods, threaded for the greater 
 portion of their length, and passing through bronze nuts N, 
 supported on cast-iron plates P, which cover the openings in 
 the furnace, any remaining space being cemented up with a 
 mixture of graphite and glue. Hand wheels W are attached 
 to these nuts, and serve to adjust the relative positions of 
 the three carbons inside the furnace. 
 
 The capacity of the latter is seventy cubic feet. 
 
 To commence with, the carbons are lowered until they 
 make contact with a short-circuiting plate of graphite, and 
 the interior of the furnace is filled with bricks, duly prepared 
 as described above ; a lining, about half an inch thick, of 
 unconverted mixture is usually left adhering to the walls 
 from the previous operation ; it protects the structure from 
 excessive heat, and, incidentally, prevents the newly 
 formed carbide from adhering to the walls. The furnace 
 door is closed, and all spaces luted with clay or lime. 
 
 The generating plant is then started up, and the load 
 soon attains its normal value, viz., 1,200 amperes on each 
 circuit, at 145 volts. After about half an hour, the furnace 
 is fed through the main door, and subsequent regulation 
 is effected according to the readings on the ammeters, 
 which are included, one in each circuit. 
 
 The furnaces are charged once every quarter of an hour, to 
 make up for the decrease in volume of the charge arising 
 out of the conversion. A complete operation lasts about 
 five hours, after which the current is switched on to the 
 next furnace. Another interval of five hours is then necessary 
 in order to permit the contents of the furnace to solidify, 
 
ELECTRIC FURNACES AND 
 
 when they are removed, red hot, by means of iron bars, 
 
 Each block of carbide thus formed weighs from 150 to 
 
 250 kilogrammes, and is left for some considerable time to 
 
 cool before it is broken up. 
 
 The output of carbide manufactured at St. Marcello 
 
 works out at 4' 37 kgs. per k.w. day of twenty-four hours, 
 
 but, according to UElectri- 
 cien, July 8, 1899, this fig- 
 ure has been increased to 
 4*88 kgs. per k.w. day. 
 
 The consumption of elec- 
 trodes during one opera- 
 tion of the furnace is 4*5 
 to 5 kgs., or 24s. per ton of 
 carbide produced. The 
 figure is a high one, and is 
 attributed mainly to the 
 frequent stoppages for dis- 
 charging and recharging, 
 during which a wasteful 
 combustion of the electrode 
 carbon goes on. 
 
 In the continuous carbide 
 furnace operated at St. 
 Marcello, the three elec- 
 trodes E E (Fig. 21) are 
 grouped " star " fashion, 
 and the arcs are struck be- 
 tween them and a common 
 
 carbon electrode C, which constitutes the hearth of the 
 
 furnace. 
 
 The latter is a cylindrical structure R of refractory brick, 
 
 the three carbon electrodes entering at an angle through the 
 
 upper walls, and being supported and fed downwards in a 
 
 similar manner to the foregoing. 
 
 The feed hopper H is situated over the centre of the 
 
 92 
 
 FIG. 21. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 furnace, above the carbons, whilst the hearth, which con- 
 sists of a cast-iron plate F covered with several layers of 
 graphite C, is supported on a central screwed stem S, and 
 can be raised or lowered throughout the entire height of the 
 furnace by means of a hand wheel and gearing W. To 
 commence with, it is raised until the graphite plate makes 
 contact with the three carbon electrodes, and is gradually 
 lowered by means of its hand wheel, as the operation pro- 
 ceeds, and fresh material is fed in. 
 
 As the process of carbide formation goes on, the hearth, 
 as already stated, is gradually lowered, making room for more 
 carbide, and a further supply of raw material. At a certain 
 point, the remaining, partially converted and unconverted 
 charge is gripped and held by three clamping jaws, operated 
 by screws, whilst the carbide already formed is removed at 
 the base of the furnace, and the hearth re-adjusted for the 
 reception of a fresh charge. 
 
 Six or seven hours' continuous working suffice to bring 
 the hearth to its lowest point, with a column of finished 
 carbide above it, surmounted by a layer around the arcs, in 
 which the process of conversion is still going on. A door at 
 the base of the structure permits of the removal of the car- 
 bide first formed, and the process is thereby rendered 
 continuous. 
 
 Regulation is effected by moving the hearth only, the 
 adjustment of the carbons themselves being only of a nature 
 to compensate for their consumption. 
 
 The Parker carbide furnace, designed and patented by 
 Mr. A. Parker, of Chorley, has, for its main object, the 
 securing of a more wide-spread action throughout the mass 
 of unconverted material, with a consequent lessening of the 
 crust or unconverted residue, which is always present in car- 
 bide furnaces, and has to be either mingled with the carbide 
 formed, thereby decreasing its efficiency as a gas producer, 
 or again subjected to the action of the furnace in conjunc- 
 tion with the next charge to be introduced. 
 
 93 
 
ELECTRIC FURNACES AND 
 
 The Parker furnace, the principle of which is represented 
 in sectional plan by Fig. 22, consists of a slowly rotating 
 cylindrical crucible A, lined with carbon C, which con- 
 stitutes the negative electrode, and a rectangular vertical 
 block, B, which forms the positive electrode, and is placed 
 axially within the crucible, A. The slow rotation of the 
 latter, coupled with a simultaneous upward motion, com- 
 municated to the electrode B, 
 serves to build up a solid and 
 homogeneous block of carbide, 
 surrounded by a very thin crust 
 of unconverted material. The 
 corners of the electrode B, com- 
 ing, as they do, into close prox- 
 imity with the inner wall of A, 
 effectually agitate the contents 
 FIG. 22. during the process of conversion, 
 
 whilst the arc is constantly moved 
 
 from one point to another, thereby tending to homogeneity 
 of action. The feed hopper inlets i i are arranged, one on 
 either side of the positive electrode B at the upper portion 
 of the furnace, and a regular feed is automatically secured. 
 The process of removing a full crucible, substituting an 
 empty one, and readjusting the electrode B preparatory to 
 again switching on the current, is said to occupy only three 
 minutes. 
 
 In a modified form of this furnace, a flattened or ovoid 
 electrode is substituted for the rectangular formation shown 
 in the figure. 
 
 The Parker carbide furnace is now in use by the Acetylene 
 Gas and Electric Smelting Company, Ltd., who have pur- 
 chased the patent rights, having superseded the original 
 Pictet furnaces which were installed by the Ingleton Water 
 Power Company. 
 
 The type of furnace employed by the United Alkali Com- 
 pany, at Widnes, resembles that devised by Parker. It 
 
 94 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 operates on the resistance principle, the upper electrode being 
 continuously embedded in the charge, whilst the remaining 
 electrode is integral with the crucible containing it. 
 
 In operation, the latter is slowly rotated, and the upper 
 electrode gradually raised, raw material being fed in mean- 
 while, until a block of carbide, practically equal to the capa- 
 city of the crucible, has been built up. By this continuous 
 relative movement of 
 the charge and elec- 
 trodes, equivalent, in 
 point of fact, to a slow 
 stirring motion, very 
 complete action is se- 
 cured, and but little of 
 the contents of the 
 crucible remain uncon- 
 verted. 
 
 In 1901, Dr. Oscar 
 Frohlich, of Steglitz, 
 Germany, patented and 
 assigned to the Siemens 
 and Halske Company, 
 
 of Berlin, a somewhat 
 ingenious carbide fur- 
 nace, the principal nov- FIG. 23. 
 elty in which consists 
 
 in the collection and utilization of the otherwise waste gases, 
 mainly carbon monoxide, resulting from the reaction itself. 
 The furnace in which Dr. Frohlich effects this, is repre- 
 sented in Fig. 23, and consists of a cylindrical iron 
 crucible A, lined interiorly with fire-clay F, and a sloping 
 funnel-shaped carbon hearth C, which forms one electrode 
 of the device, and centres in a discharge orifice or outlet 
 for the carbide. The other electrode T, also of carbon, 
 takes the form of a vertical tube, or flue, the upper end of 
 which is capped and fitted with the tubular conduits c c, 
 
 95 
 
ELECTRIC FURNACES AND 
 
 which curve over as shown, and descend to an annular flue 
 /, disposed around the lower portion of the crucible. 
 
 The carbon monoxide and other reaction gases pass up 
 through the central tubular carbon electrode T, and by 
 way of the conduits c c to the flue /, where they are burnt 
 in admixture with air, which enters through holes in the 
 flue walls. The products of combustion pass off by way of 
 the pipe P. The heat of combustion of the reaction gases, 
 which, collected in this manner, are. almost entirely free from 
 dust, is thus utilized in auxiliary heating of the crucible 
 contents. 
 
 The carbide formed falls through the orifice O on to the 
 outer surface of the conical table D, which is lipped to 
 retain it. This receiving cone is carried by a vertical stem, 
 which can be raised or lowered in its standard S by means 
 of the lever H ; its position, relatively to the crucible, can 
 thus be adapted to either intermittent or continuous action 
 of the furnace, as may be desired. 
 
 Mr. Alfred H. Cowles has recently patented some details 
 in the process of calcium carbide manufacture, which, among 
 other things, include the utilization of the carbon monoxide 
 gas given off for the purpose of pre-heating the charge. In 
 Dr. Frohlich's invention, already described, the gas is 
 used rather as an auxiliary to the electric heating in the 
 furnace itself. 
 
 A form of carbide furnace construction, based upon 
 Cowles' latest patent, is depicted in Pig. 24. It consists 
 of an inverted cone-shaped crucible A, of iron, lined 
 interiorly with refractory non-conducting material R. 
 The hearth of the furnace consists of a solid carbon block 
 B, also protected by an iron casing, and electrically separated 
 from the main body of the furnace by a layer of insulating 
 material L. This carbon block also constitutes one fixed 
 electrode, the others E E consisting of carbon rods, dis- 
 posed radially, like the spokes of a wheel, through the upper 
 part of the furnace walls. 
 
 96 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 A cursory study of this form of construction will show 
 that the heating, being effected on the resistance principle, 
 is most concentrated at, 
 or near, the hearth itself, 
 where the various current 
 paths converge. The ob- 
 ject of this concentration 
 is to secure a molten pro- 
 duct which can be run off 
 at the tap hole T pro- 
 vided in the hearth. 
 
 Mounted axially in the 
 centre of the lid D, which 
 effects a gas-tight junction 
 with the furnace proper, 
 is a pre -heating chamber 
 or tower H, having a feed 
 aperture at the top. It 
 
 consists, like the furnace, FlG 2 4. 
 
 of an iron shell, lined with 
 
 refractory material, and through its walls, a short distance up 
 from the furnace lid, pass pipes, P P, which communicate 
 with the main furnace chamber as shown, and serve to convey 
 the carbon monoxide and other gases to burners in the 
 pre-heating tower, where they are burnt in conjunction with 
 a blast of air introduced through the smaller pipes p p 
 after the manner of a blow-pipe. 
 
 The descending charge thus receives a preliminary heating 
 before entering the furnace, an initial elevation of its 
 general temperature, which should effect considerable 
 economy in the current necessary to complete the process 
 of carbide formation. 
 
 A form of arc furnace, invented by R. C. Contardo, which 
 is applicable, with some slight modifications, to either 
 calcium carbide manufacture, or the smelting of metallic 
 ores, consists of a refractory chamber, with curved, or 
 
 97 H 
 
ELECTRIC FURNACES AND 
 
 contracted hearth, immediately above which are disposed the 
 two electrodes. Above these again, and meeting at an 
 angle over the arc, are two inclined plates which may be 
 of graphite, or other equally suitable material. These form, 
 together with the downwardly sloping walls of the furnace 
 itself, a couple of inclined passages or chutes, down which 
 the raw material is fed from a central orifice above. 
 
 By virtue of their position, immediately above the arc, 
 the sloping plates become intensely hot, and communicate 
 their heat to the descending charge, which is thus pre-heated 
 before entering the active centre of the furnace, immediately 
 
 below the electrodes, where 
 the operation is concluded. 
 
 In general construction, the 
 above furnace is very similar to 
 that illustrated in Fig. 32. 
 
 The King carbide furnace 
 turns out its product in "block " 
 form. The hearth consists of 
 a truck T (Fig. 25), at the 
 bottom of which is a slab of car- 
 bon C, forming one electrode 
 of the furnace, the other, E, be- 
 ing vertically adjustable. The 
 truck is mounted on rails, and 
 provided with a mechanical 
 attachment, not shown in the 
 figure, whereby a small recip- 
 rocating movement is communicated to it, during the 
 action of the furnace, tending to shake down and level the 
 charge. It is enclosed in a brickwork chamber B, down 
 inclined channels a a, in the walls of which the raw 
 material is fed. As the operation proceeds, the upper 
 electrode E is gradually raised, whilst fresh material is 
 fed in, until the truck is full of carbide together with partially 
 converted raw material. It is then allowed to cool, run out 
 
 98 
 
 FIG. 25. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 of the chamber B, and emptied, the carbide being separated 
 from the unconverted material, and the latter worked up 
 into the next charge. 
 
 Air ducts, not shown in the figure, are provided within the 
 walls of the structure, which, by circulation of the air 
 within them, serve to keep the inactive portions of the 
 furnace fairly cool, and confine the centre of thermal 
 activity to the truck hearth. 
 
 The furnace employed at the works of the Deutsche 
 Gold und Silberscheide Anstalt consists of a removable 
 hearth, or coke-lined truck, which also constitutes one 
 electrode, and a fixed upper structure of refractory brick- 
 work, which surrounds the truck when in its working 
 position, and assists in conserving the furnace heat. 
 
 The upper carbon electrode is adjustable, and passes 
 through an aperture in the roof, as do also the feed inlet, 
 and a pipe for conveying away the carbon monoxide gas 
 formed in the reaction. 
 
 To commence with, the truck hearth, previously filled 
 with a granular mixture of coke and lime, is run, on rails, 
 to a point immediately beneath the hood, into which it is 
 then raised, vertically, by chains and pulleys, until its lower 
 projecting flange effects a gas-tight junction with the 
 lower rim of the fixed structure above. 
 
 The upper adjustable electrode is then lowered into 
 position through the opening provided for it in the hood, 
 and the current switched on. A current of from 2,000 to 
 3,000 amperes at 60 to 65 volts is required to operate the 
 furnace, and an output of five kilogrammes of carbide per 
 kilowatt day is said to be secured, the resultant product 
 being 85 per cent, pure carbide. The upper carbon electrode 
 is consumed at the rate of about 50 kgs. per ton of carbide 
 produced. 
 
 In carbide furnaces designed for continuous action, some 
 difficulty is experienced in securing the requisite regularity 
 of action, owing to the time expended in raising each batch 
 
 99 
 
ELECTRIC FURNACES AND 
 
 of raw material which arrives at the heat zone to the 
 temperature necessary for effecting combination. 
 
 The action is thus jerky and intermittent, rather than 
 continuous. 
 
 A German patent, the principle of which is illustrated 
 in Fig. 26, has in view the elimination of this defect. It 
 relates to a continuous carbide furnace, in which one 
 electrode, A, is vertical, whilst the other, B, is inclined to 
 it, meeting it at the lower extremity, with the exception 
 of the necessary arcing space. The furnace hearth F 
 
 is also inclined at an angle, which 
 coincides with that of the elec- 
 trode B, and its walls, moreover, 
 converge, to form a species of 
 funnel, or tapering conduit, with 
 its smallest extremity situate at 
 the meeting point of the elec- 
 trodes, where also the discharge 
 orifice for the finished carbide is 
 located. 
 
 The raw charge, fed in at the 
 _^ widest dimension of the furnace, 
 thus forms a bridge between the 
 electrodes for the greater part of 
 their length, the resistance of 
 which varies inversely with the 
 cross-section. The heating effect 
 
 is thus gradually increased, as the mixture of coke and lime 
 travels towards the arcing zone, and is finally concen- 
 trated in the region of the arc, where the operation is com- 
 pleted and the carbide withdrawn. 
 
 Energy Required in Calcium Carbide Manufacture. 
 Blount's theoretical estimate of the energy necessary to 
 produce calcium carbide is thus arrived at 
 
 " Moissan has shown that the heat of formation of calcium 
 oxide is 145 Calories, and that the reaction, 
 
 100 
 
 FIG. 26. 
 
THEIR INDUSTRIAL APPLICATIONS 
 CaO+ 3C=CaC 2 + CO, 
 
 takes place at 3,300C= 5,972F. The specific heat 
 of CaO may be taken as approximately 0-12 ; that of 
 carbon as 0-47. The energy necessary to raise 56 grammes 
 of CaO and 36 grammes of C to this temperature 
 is 79*5 Calories. The formation of calcium carbide from 
 Ca and C is esteemed an endothermic reaction, requiring 
 48 Calories. The total energy needed is, therefore, 
 
 79-5 + 145 + 48 Calories = 272-5 Calories. 
 
 From this must be deducted the energy evolved by 
 the oxidation of carbon to CO, i.e., 29 Calories. There- 
 fore, the energy to be supplied to form 64 grammes of 
 CaC 2 is 243-5 Calories. 
 
 " In this calculation, the energy evolved or absorbed by 
 the formation of CaC 2 from Ca and C 2 is a doubtful quantity. 
 No reliable data have been arrived at. The estimate given 
 is likely to err on the right side, the more so as no credit 
 has been taken for possible regeneration by utilizing the 
 sensible heat of one charge for warming up the next. 
 
 " Thus it may be taken, for practical purposes, that the 
 formation of one ton of calcium carbide requires 5,889 H. P. 
 hours, or conversely, for each H.P. day of 24 hours, 4-1 kgs. 
 of calcium carbide may be formed." 
 
 The Progressive Age, in 1896, deputed Messrs. Houston, 
 Kennelley, and Kinnicutt to carry out some experiments 
 in carbide manufacture on a commercial scale at Spray, 
 N. Carolina, with a view to determining the actual cost of 
 production. 
 
 The plant was of 300 H.P. and the current was delivered 
 at the furnace terminals at an E.M.F. of 100 volts. There 
 were two furnaces, each with a hearth area 3 ft. by 2 ft. 
 6 in., the hearth, and fixed electrode, consisting of a carbon 
 plate, 8 in. thick, supported on an iron base plate. The 
 
 101 
 
ELECTRIC FURNACES AND 
 
 upper electrode was vertically adjustable, and . consisted 
 of a composite carbon block, 12 in. by 8 by 3 : it was 
 consumed at the rate of ^ in. per hour. The charge con- 
 sisted of the usual coke and lime, containing 37 per cent, 
 carbon, and 52 per cent, lime, respectively. The furnaces 
 were started by striking an arc between the upper electrode, 
 and a small quantity of the raw charge placed upon the 
 hearth. The upper electrode was gradually raised, and 
 fresh material added, until a pyramidal mass of carbide 
 was built up, equal, or nearly so, to the cubic capacity of 
 the furnace. 
 
 Two runs were made, the weight of raw material treated 
 in each case being 900 kgs., whilst an output of about 
 90 kgs. of calcium carbide resulted from each. The con- 
 sumption of energy was, in the first instance, 193-1 H.P. 
 for three hours, or 579-3 h.p. hours = 432 k.w. hours; and, 
 in the second, 195-3 H.P. for 2 hours 40 minutes, or 520-8 
 h.p. hours =388-5 k.w. hours. 
 
 The respective yields were thus 3-75 kgs. of 80 to 85 per 
 cent, carbide per h.p. per 24 hours, and 4-15 kgs. for the 
 same expenditure of power respectively. 
 
 Many and varied are the figures relating to the energy 
 required for the production of calcium carbide, not only on 
 a commercial scale, but also from a theoretical standpoint 
 based on calculations involving thermal data regarding 
 the component materials themselves. In this connexion, 
 Mr. J. B. C. Kershaw, F.I.C., contributed a very interesting 
 series of articles which first appeared in the Electrician, 
 November 23, 1900. 
 
 Dealing, primarily, with the theoretical side of the ques- 
 tion, the following table, excerpted from the Electrician 
 of that date, gives his figures, in tabular form, as gleaned 
 from various authorities on the subject of calcium carbide 
 manufacture, the actual references being quoted in the 
 original article 
 
 102 
 
Of 
 
 THEIR INDUSTRIAL APPLICATIONS 
 
 AUTHORITY. 
 Sieber 
 Addicks 
 Bredel 
 Wolff 
 Pictet . 
 Haber . 
 Gin . 
 Lewes . 
 Allen 
 
 KILOWATT HOURS. 
 1,523 
 1,941 
 1,960 
 2,043 
 2,640 
 2,670 
 3,069 
 3,290 
 3,293 
 
 In the above table, the figures represent the kilowatt 
 hours theoretically necessary to produce one metric ton 
 (2,204 Ibs.) of calcium carbide containing 80 per cent, 
 carbide. 
 
 As Mr. Kershaw observes, there is a wide disparity be- 
 tween the figures quoted above, which is in the main due 
 to the fact that no allowance has been made for waste of 
 material in the furnace, nor for the heat required to raise 
 the temperature of the charge, in the first instance, to that 
 at which the actual carbide formation takes place. 
 
 In the course of his article, the Author discusses the 
 various possible sources of error, and, in summing up, 
 decides in favour of the calculation made by M. Gin, and 
 published in UEclairage Electrique, May 6, 1899. 
 
 The following is the reasoning referred to 
 
 " The temperature of the reaction is taken as 3,300C. = 
 5,972F., and the following formulae are used for calculating 
 the specific heats of the carbon and lime at that tempera- 
 ture 
 
 Carbon (atomic heat) ... . 4-26 +0-00072t. 
 
 Lime (gramme molecule) . . . 11-4+0-OOlt. 
 
 Using these data, the following figures are obtained for 
 the production of the gramme molecule of calcium carbide 
 
 Heat necessary to raise 56 grammes of 
 
 limeto3,300C 
 
 Heat necessary to raise 36 grammes of 
 
 carbon to 3,300C. 
 
 Heat necessary to split up 56 grammes 
 lime into Ca and O 
 
 43,060 Calories. 
 53,940 
 145,000 
 
 Total 
 
 . 242,000 
 
 103 
 
ELECTRIC FURNACES AND 
 
 Less 
 
 Heat produced by formation of 28 
 
 grammes of CO . . . . 26,100 Calories 
 
 Heat produced by formation of 64 
 
 grammes of CaC 2 . . . 3,900 
 
 Net Heat required . . 212,000 
 
 212,000 Calories = 245*5 watt hours electrical energy 
 = 3,837 k.w. hours per ton of carbide. 
 
 Mr. Kershaw next proceeds to deal with the various 
 experimental and estimated yields of carbide, the results 
 of which are embodied in the following table 
 
 AUTHORITY. 
 
 Lewes . * 
 
 Korda 
 
 Houston and Kennelly 
 
 Bullier 
 
 Liebetanz 
 
 Wolff 
 
 KILOWATT HOURS. 
 
 . 4,105 
 
 . 4,380 
 
 . 4,393 
 
 . 4,470 
 
 . 4,800 
 
 . 5,111 
 
 . 6,514 
 
 The figures in the above table represent the kilowatt 
 hours necessary to produce one metric ton (2,204 Ibs.) of 
 calcium, containing 80 per cent, carbide, as determined 
 by actual experiment, and estimates based on the results 
 of the same. 
 
 Finally, in Table No. 3, below, Mr. Kershaw gives us the 
 results obtained in actual practice 
 
 AUTHORITY. 
 
 * Lewes . 
 *Keller 
 *Carlson 
 Patten 
 Borchers 
 Keller . 
 Lewes 
 Haber 
 Pierard 
 Nikolai 
 Carlson 
 Haber 
 
 KILOWATT HOURS. 
 
 . 3,576 
 
 . 3,788 
 
 . 3,871 
 
 . 4,104 
 
 . 4,105 
 
 . 4,157 
 
 . 4,291 
 
 . 4,351 
 
 . 4,470 
 
 , 4,920 
 
 . 5,066 
 
 . 5,616 
 
 . 5,960 
 
 104 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The above figures represent the kilowatt hours necessary 
 to produce one metric ton (2,204 Ibs.) of calcium carbide, 
 containing 80 per cent, carbide, in actual commercial 
 practice. 
 
 Summing up these various data, Mr. Kershaw finally 
 decides in favour of those marked with an asterisk as being 
 the most reliable. They represent results obtained with 
 the Deutsche Gold und Silberscheide Anstalt form of fur- 
 nace ; the Willson furnace, as used at Foyers ; and the 
 Gin and Leleux production, in use at Meran, in Austria. 
 These three types are described elsewhere, and may be 
 taken as representing three of the most efficient carbide 
 furnaces then in practical use. 
 
 Cost of Calcium Carbide Manufacture. The cost of 
 calcium carbide manufacture must necessarily be largely 
 governed by local conditions, source of power, cost of raw 
 materials, etc., so that it is impossible to formulate any 
 hard and fast rule on the subject. 
 
 Some interesting figures, from various sources, are 
 assembled in the form of a statistical article on the progress 
 of the industry by Kershaw, in the Electrical Review, Sep- 
 tember 22, 1899. 
 
 At Lourdes, in the Hte. Pyrenees, the cost is estimated 
 at about 12 per ton. In Switzerland, at that time, car- 
 bide was selling at 14 per ton, and, in Paris, at 19. 
 
 At Meran, in the Austrian Tyrol, it was estimated to cost 
 7 5s. per ton, f.o.r., and that after allowing for interest 
 on capital and depreciation of plant. 
 
 At Niagara, in 1899, it was estimated that carbide could 
 be produced at from 7 5s. to 8 per ton. 
 
 Liebetanz, in his paper before the Buda Pesth Congress, 
 estimates the cost of production with steam power at 
 11 145. per ton, and with water power at 8 18s. 
 
 According to Carl Hering (ISEclairage Electrique) it 
 requires, theoretically, 1,900 Ibs. of lime, and 1,230 Ibs. 
 of carbon to produce one ton of calcium carbide ; in actual 
 
 105 
 
ELECTRIC FURNACES AND 
 
 practice, 2,050 Ibs. of lime, and 1,420 Ibs. of carbon are 
 necessary. These figures apply particularly to the carbide 
 factory at Meran, in Austria, where the cost of lime, per 
 ton, was then about 16$., and of carbon, 32s. 
 
 One electrode suffices for ten tons of carbide, and costs 
 6 12s. or approximately 13s. per ton. 
 
 The electrical energy required per ton of carbide produced 
 was 6,400 h.p. hours, which, at 2 per E.H.P. year, works 
 out at a little above 1 16s. per ton. 
 
 Accessory machinery, losses, etc., account for about 
 200 H.P., or 4s. per ton, the output being 6-5 tons per diem. 
 
 Labour, at 3s. to 3s. 4d. per day, amounts to 15s. per ton. 
 Amortisation 1 per ton, and general expenses 1 per ton ; 
 maintainance of plant, 6s. per ton. Total cost, at Meran, 
 7 5s. per ton. 
 
 106 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION V 
 
 IRON AND STEEL PRODUCTION IN THE ELECTRIC FURNACE 
 
 Introductory. M. Keller, of the electro-metallurgical 
 firm, Keller, Leleux and Co., France, in a paper before the 
 Iron and Steel Institute, 1903, states the conditions under 
 which electrical reduction processes may be economically 
 adopted ; they are as follows 
 
 " Although the reduction of iron ores by electricity has 
 formed the subject of a practical study by the Author, the 
 results of which will be described further on, it may be 
 stated at once that the employment of electricity as a 
 reducing agent is only practicable from an economic point 
 of view : first, when it is a question of manufacturing special 
 qualities of iron, from pure ore, delivered at the works 
 on favourable terms. Secondly, when it is desired to foster 
 an iron and steel industry in a country hitherto undeveloped 
 in this respect, into which all the coal must be imported, 
 where iron ore of good quality abounds, and where natural 
 sources of power are available in the immediate neighbour- 
 hood of the ore deposits. 
 
 " The Author has determined, experimentally, that one 
 k.w. year, utilized in an electric reducing furnace, is capable 
 of yielding about four tons of steel-making pig-iron. In 
 a general way, therefore, the reduction of iron ore by 
 electricity in a country possessing metallurgical works 
 and equally good conditions of transport throughout, is 
 only theoretically practicable if the ore can be obtained 
 on equal terms, and if the cost of the kilowatt year does not 
 
 107 
 
ELECTRIC FURNACES AND 
 
 exceed 1 55. 6d. It must be borne in mind, however, 
 that the scale of production of an electric furnace is less 
 than that of blast furnaces, and that, on this account, 
 establishments equipped for electric working would be 
 of less capacity ; the working expenses being also pro- 
 portionately higher, it does not appear as if the reduction 
 of iron ores, so far as the smelting of ordinary pig is con- 
 cerned, will ever enter the sphere of practice in any European 
 hydro-electric works." 
 
 The advantages of the electric reduction process are thus 
 detailed by the same authority 
 
 " A high degree of purity can be reached by the electric 
 method of production, owing to the method of generating 
 the heat, which enables the action of the fuel to be limited 
 strictly to that of a reducing agent. By this means the 
 action of the sulphurous gases is, to a great extent, avoided ; 
 and, since the quantity of coke required per ton of iron is 
 so small, the very best qualities can be utilized for the 
 purpose. It is therefore possible, with good ore, to obtain 
 a pig-iron which will compare in purity with Swedish iron. 
 A further advantage consists in the higher temperature of 
 working, which the use of electricity permits, without, 
 however, causing an excess of carbon in the material. It 
 is also possible, in consequence of the hot working, to form 
 slags of ultra-basic character." 
 
 In steel production, on the other hand, the cost of the 
 electrical energy is a secondary consideration. In melting 
 and fining one ton of steel produced from scrap iron and 
 steel, M. Keller estimates that about 1-10 k.w. year is neces- 
 sary, and, even using steam as the primary source of power, 
 the cost of energy will only amount to about 1 12<s. per 
 ton of steel. 
 
 According to Heroult, 2,500 tons of steel have been 
 already produced in the electric furnace, whilst plant, 
 with an aggregate capacity of 400 tons of steel per day, 
 is in operation.. 
 
 1 08 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 In both the Heroult and Keller electric furnace steel 
 processes, the heating arcs are struck between the metallic 
 charge and the electrodes. The Heroult furnace is in opera- 
 tion at La Praz, and the Keller at Li vet, on the Bomanche. 
 As will be seen from the above, the cost of smelting iron 
 ores by electrical means is somewhat large, and, in the 
 present stage of its development, can only hope to com- 
 pete with ordinary blast furnace processes, in special cases, 
 as for example, where power can be had at a comparatively 
 low figure, whilst pure ore is plentiful. 
 
 Dr. H. Goldschmidt, who has acted in the capacity of 
 inspector, appointed by the German Patent Office to 
 report on the various electric smelting processes, has fur- 
 nished some interesting information and data on the sub- 
 ject (Zeitschrift fur Elektrochemie, 1903). 
 
 With reference to the earliest industrial application of 
 the electric furnace to smelting processes, it may be men- 
 tioned that to Heroult belongs the honour of having made 
 the pioneer move, for, according to Goldschmidt, he de- 
 livered his first car-load of steel, consisting of bars weigh- 
 ing 400 kgs. each, and amounting, in all, to 8,890 kgs., to 
 Messrs. Schneider & Co., Creusot, on December 28, 1900. 
 This consignment was the product of his electric furnace 
 at Froges. 
 
 The electric furnace has hot been commercially applied 
 in America to iron and steel manufacture, but Europe has 
 been the scene of many experimental and commercial 
 trials, with some of which very successful results have 
 been obtained. This latter fact, combined with the exist- 
 ence of rich iron deposits in Canada, a country in which 
 water power is abundant, has prompted the Canadian 
 Government to appoint a Commission to visit Europe, 
 and study the possibilities of electric smelting, ore reduction, 
 and steel manufacture, with a view to its industrial develop- 
 ment in Canada. 
 
 Sweden, France, and Italy are the principal countries 
 
 109 
 
ELECTRIC FURNACES AND 
 
 to be visited by this Commission, which, at the time of 
 writing, is actually engaged upon its work of investigation. 
 Some valuable information should be the outcome of 
 this enterprise, and the advent of the official report will 
 doubtless be awaited with interest by metallurgists in 
 general. 
 
 The resistance furnace, with carbon core, has, until 
 comparatively recently, been commercially inapplicable 
 to the smelting or reduction of non- volatile metals, for the 
 simple reason that the mere presence of carbon in the 
 core has led to the formation of a carbide by reaction with 
 the metal liberated from the ore. 
 
 Acheson has, however, surmounted this obstacle by 
 protecting the carbon core from contact with the charge 
 of metallic ore by an impervious sheath, e.g., carborundum. 
 The arrangement is very simple ; around the carbon or 
 granular coke core of an ordinary resistance furnace of 
 the well-known Acheson type is packed either a quantity 
 of carborundum crystals, already formed, or sufficient 
 silica and carbon, mixed in the correct proportions, to form 
 carborundum. Outside this, again, is placed the mass 
 of ore to be reduced. In action, the carborundum forms 
 a coherent impermeable sheath, which is self-supporting, 
 and maintains its position around the core, even after the 
 encircling charge has fallen away from it by virtue of the 
 reduction taking place. With the aid of this protecting 
 sheath, Acheson has succeeded in effecting the direct 
 reduction of silicon and aluminium in the electric furnace. 
 
 Processes. In connexion with the various thermo- 
 electric processes, experimental and otherwise, for the 
 manufacture of steel, the following table, containing results 
 and data furnished by Goldschmidt, of Essen, at the Fifth 
 International Congress of Applied Chemistry, may prove 
 of interest. 
 
 no 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Sa 
 
 O H ^ .J 
 
 isfi 
 
 o o 
 
 (M 
 
 CO 00 
 
 *"* O 
 
 IS S.S. 
 
 *l 
 
 -M I 
 
 a ^ 
 
 CO 
 
 O 
 
 03 tS 
 
 s 
 
 . I 
 
 1 1 1 1 I I I 
 
 d 
 
 ^3^0^73 73 73^73 73 
 QigoOo-^CDgfligogftgogo 
 
 a^aJTjocp-So^c-^o-JSo-So 
 coccC-go3<i;o3oceoo30o3Ocoaa3 
 fi^tfSC'aSC^C^fl^a^fl^rt 
 
 STBS'SiiTBSUi^iTB'SS 
 
 fe fe fe ^ 
 
 -! 
 
 fe fe 
 
 
 
 i 
 
 III 
 
ELECTRIC FURNACES AND 
 
 According to Ruthenberg, the production of one ton of 
 metallic iron requires, by the 
 
 Stassano process .... 3,000 h.p. hours. 
 
 De Laval .... 3,500 
 
 Rossi .... 4,800 
 
 Ruthenberg .... 500 k.w. 
 
 Comparing the economy of the blast furnace process 
 with that of the electric furnace, Ruthenberg points out 
 that a ton of coke is required for the production of one ton 
 of pig-iron, of which only a little more than 150 kgs. are 
 actually required for the reduction of the ore, the remainder 
 being necessary to melt the reduced iron, and, incidentally, 
 to form cinder, whilst no inconsiderable portion is lost or 
 wasted in the form of combustible gas. 
 
 One of the earliest attempts to apply the electric furnace 
 to smelting operations, was that made in 1892 by De Laval, 
 of Stockholm, his furnace being of the resistance type, but 
 differing from most others of a like kind, in that the core 
 consisted of fused magnetite. The mass of this latter, 
 which constituted a molten bath, was disposed between, but 
 separated from, two pockets of molten iron, which served 
 the double purpose of terminal electrodes and receptacles 
 for the reduced metal which found its way into them 
 during the process, and was constantly tapped from them 
 in order to ensure a constant level. 
 
 An alternating current was used with this device, the 
 ores to be smelted, mixed with the necessary proportion of 
 carbon, being fed into the bath of molten magnetite, where 
 they underwent reduction by the heat generated electri- 
 cally in the molten mass. 
 
 The Gin furnace for steel manufacture is based on the 
 same resistance principle, and is probably one of the simplest 
 in point of application. 
 
 It comprises a long trough, shown at T, in transverse 
 section, Fig. 27, lined with refractory material R. This 
 
 112 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 trough is filled with fused cast iron F, which, in addition 
 to being the material under treatment, also constitutes 
 the heating resistance column. The terminal connexions 
 
 FIG. 27. 
 
 through which current is conveyed to the molten column 
 are large blocks of steel, kept cool by internal water cir- 
 culation, and also by virtue of their greater cross-section. 
 
 The inconvenience attendant on so unwieldy a construc- 
 tion as would be entailed by arranging the resistance channel 
 or furnace hearth, in one straight length, is eliminated by 
 doubling it back upon itself several times in zig-zag fashion, 
 as shown at E, Fig. 27, where t is the trough, and e e the 
 terminal electrode blocks. Thus constituted, the furnace 
 may be likened to a huge incandescent electric lamp fila- 
 ment ; it is mounted, for portability, on a trolley, which 
 can be run in and out of an arched furnace, for the recep- 
 tion of the charge of molten pig-iron, which has been 
 primarily fused in the latter. 
 
 There are two distinct methods of manufacturing steel 
 by the Gin process known respectively as the " dilution " 
 method, and the " ore process." They are employed 
 either independently, or in combination. The former is 
 very simple, and consists in adding a calculated percentage 
 of scrap iron to the molten mass, in which it dissolves, 
 followed by the speedy diffusion of the carbon throughout 
 the mass, and its consequent conversion into steel. 
 
 The " ore process " entails the addition of iron oxide to 
 the mass, the oxygen of which assists in the elimination of 
 undesirable impurities. 
 
 M. Keller's electric blast furnace for the reduction of 
 
 113 i 
 
ELECTRIC FURNACES AND 
 
 iron ores resembles an ordinary blast furnace in general 
 appearance, its hearth being surmounted by a shaft of 
 refractory brick for the reception of the ores, flux, and 
 reducing materials, which are fed in at the top through a 
 gas-tight valve. 
 
 The electrodes are vertical, and laterally disposed, each 
 one being capable of independent adjustment, which enables 
 their several heating effects to be equalized by means of an 
 ammeter included in each circuit. 
 
 There are at least four of these electrodes to a furnace, 
 and any one of them can be removed and replaced without 
 stopping the run. At the commencement of a run the 
 current is switched on, and the furnace chamber filled with 
 raw materials. Fusion commences at the bottom, in the 
 neighbourhood of the hearth, but soon extends upwards 
 throughout the entire mass, the resultant gases being led 
 off and burnt, to dry and preheat the ores. 
 
 The steel fining furnace is placed at a lower level, its 
 inlet being just below the tap-hole of the blast furnace, 
 so that the reduced iron can flow into it. Two vertical 
 electrodes are used, which just establish contact with the 
 surface of the slag, but do not penetrate to any appre- 
 ciable depth. 
 
 The E.M.F. at the terminals of the reducing furnace is 
 25-30 volts, and of the steel furnace, 50-75 volts. 
 
 M. Keller estimates that an electro-metallurgical works, 
 with 10,000 H.P. available, could produce, by his process, 
 60 tons of steel per diem, of which 50 tons would accrue 
 from 55 per cent, ore, and the remaining ten tons from the 
 fusion of scrap. 
 
 The cost of the process, assuming a k.w. year of 8,400 
 hours to run into 2, would be from 3 125. to 4 per ton 
 of steel, exclusive of royalties, payable on the patent 
 rights ; the electrical energy would be responsible for 
 only 135. 3d. of the above total cost. 
 
 The electrical manufacture of steel by the Keller process, 
 
 114 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 as carried on at Kerrouse, France, has proved highly suc- 
 cessful. All varieties of iron ore have been electrically 
 fused and treated. Practical trials have demonstrated 
 that for the manufacture of one ton of steel from ordinary 
 ores, electrical energy to the extent of 2,800 k.w. hours, or 
 3,800 e.h.p. hours is required. 
 
 The total cost of manufacture, including electrodes, coke, 
 ore, refining materials, etc., and general expenses, was 
 found to be from 3 15s. to 4. 
 
 The ultimate disposition at Kerrouse will consist of a 
 battery of electric furnaces, served by one portable refining 
 furnace, which, in the form of a large ladle, is placed before 
 each furnace in turn, and the electrodes lowered into its 
 contents, tapped from the reducing furnaces. 
 
 The Simon electric furnace, for smelting iron and other 
 ores, consists of a movable truck hearth, which carries the 
 negative electrode, consisting of horizontal carbon blocks 
 embedded in a layer of powdered ferro- manganese. Small 
 priming blocks of carbon rest on the upper surfaces of 
 these, and serve as a temporary bridge between the cathode 
 and the vertical anodes, which are similar carbon blocks 
 embedded in a compressed mass of slag, tar, and resin. 
 The upper electrodes slide vertically in a frame, and the 
 active hearth of the furnace is enclosed. 
 
 The electrical process of iron smelting and steel produc- 
 tion, originally devised by M. Henri Harmet, consisted of 
 three distinct operations, viz., the reduction of the ore ; 
 the fusion of the ore ; and, finally, the refining of the 
 resultant metal. In a later patent, the order of the first 
 two processes has been reversed, the ores being first fused 
 and then reduced. 
 
 Electrical heat is largely employed in the process, but is 
 supplemented by the heat of combustion of the gases pro- 
 duced in the intermediate, or reduction chamber, which are 
 mixed with air, and forced, under pressure, into the lower 
 portion of the first, or fusion chamber. 
 
ELECTRIC FURNACES AND 
 
 The three furnaces are arranged with their hearths 
 sloping in cascade, so that the fused matter may run from 
 the fusion chamber into the reduction furnace, and the 
 resultant metal from the latter into the refining oven. 
 
 In the fusion chamber, the electrodes are placed at the 
 bottom, or hearth, whilst in the intermediate, or reduction 
 furnace, they are vertically arranged, with their lower 
 extremities touching, but not penetrating, the upper layer 
 of slag. 
 
 The coke is fed into the reducing furnace down a vertical 
 tower, or cylindrical structure, resembling an ordinary 
 blast furnace, emerging finally, upon the sloping hearth, 
 at a point just in advance of the tapping holes leading to 
 the refining oven ; it thus interposes an incandescent 
 screen, through which the molten metal and slag have per- 
 force to filter, before passing into the latter. 
 
 The initially smelted metal forms a thin layer on the 
 hearth of the reducing chamber, upon which floats the slag, 
 and also a coating of coke, detached from the main column. 
 The molten ores from the fusion chamber, flow on to this 
 mixture of coke and slag, and in the intense heat generated 
 by the electric current are reduced, as in an ordinary 
 blast furnace, passing through the descending screen 
 of heated coke, before finally reaching the refining oven, 
 which operates on a similar principle to that of the 
 Martin type. 
 
 A later patent taken out by M. Harmet covers an electric 
 blast furnace with a lower " fusion zone," and an upper 
 " reduction zone." The electrodes pass through the walls, 
 and the excess heat, generated in process of fusion, is 
 utilized for purposes of reduction. To this end, the gases 
 from the top of the furnace are collected, and led off by 
 pipes, to a fan or blower, which re-introduces them, under 
 pressure, into the furnace, in the neighbourhood of the 
 fusion centre. They again pass upward, carrying with 
 them the excess heat energy into the reduction zone. 
 
 116 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The gases, on coming into contact with the incandescent 
 coke, are converted into carbon monoxide, which effects 
 the necessary reduction of the descending charge as it 
 diffuses upwards. It will be seen from this that the opera- 
 tion is continuous. 
 
 A patent granted to Prof. W. S. Franklin, of Lehigh 
 University, towards the latter 
 end of 1903, covers a somewhat 
 ingenious form of electric fur- 
 nace construction, primarily 
 intended for the reduction of 
 iron ore, but also applicable to 
 the manufacture of glass. The 
 furnace, illustrated in Fig. 28, 
 consists of a pear-shaped de- 
 pression, or hearth, in a refrac- 
 tory base B, extending up- 
 wards into a funnel-shaped 
 opening or feed hopper. 
 
 The novelty of the design 
 lies in the construction of the 
 upper electrode E, which may T\ 
 be of carbon, or metal, ac- 
 cording to circumstances. Its 
 outer surface is furnished with 
 downwardly inclined teeth, t t 
 
 like those of a saw, whilst it is clamped, at its upper 
 extremity, to the moving member of an eccentric or re- 
 ciprocating motion R. 
 
 The raw charge is fed into the funnel-shaped mouth H, 
 around this central electrode, which latter, by its vertical 
 movement, carries it down into the furnace proper, the 
 return stroke, owing to the slope of the teeth t t, moving 
 the charge in the hopper but little, if at all. In this manner, 
 a constant feed is secured. 
 
 When applied to the reduction of iron ores, a layer of 
 
 117 
 
 FIG. 28. 
 
ELECTRIC FURNACES AND 
 
 carbon C is placed at the lower portion of the hearth, 
 electrical connexion with it being secured by an iron rod, 
 passing through the refractory wall. Over this is packed 
 a filling of slag, and the current turned on. The resultant 
 heat is regulated by the height of the upper electrode ; 
 if the latter dips into the molten slag, the furnace operates 
 on the resistance principle, and the heating effect is moderate ; 
 if, however, it be raised out of contact with the surface 
 of the slag, an arc is struck between the two, and the tem- 
 perature augmented. 
 
 As the ore descends on to the surface of the molten slag, 
 it melts, and the metal filters through to the lower portion 
 of the hearth, whence it is drawn off by way of the tap- 
 hole T. A similar opening S, at a higher elevation, pro- 
 vides for the removal of the superfluous slag. 
 
 When utilized for the electrical manufacture of glass, 
 an initial charge of that substance is given to the furnace, 
 in order to start the action, after which the raw mixture 
 is fed in as before, the resultant glass being drawn off in 
 a molten condition at the lower tap-hole T. 
 
 A simple and comprehensive patent, taken out in 1899 
 by Dr. Borchers, relates to resistance furnace methods 
 of smelting and ore reduction, in which carbon is employed 
 as the reducing agent. The process consists in packing 
 the whole of the carbon employed in the reaction between 
 two electrodes, to act as a heating resistance core, and 
 surrounding the mass with an encircling filling of the ore 
 or substance to be acted upon. No preliminary mixing 
 of the two ingredients is necessary in this method, which 
 is closely allied to that of Dr. Borchers, described under 
 the section dealing with carbide furnaces. 
 
 In 1901 a French patent was taken out by the Societe 
 Electrometallurgique Fran9aise on an arc furnace for the 
 manufacture of all varieties of iron and steel in one opera- 
 tion, by the direct treatment of specific mixtures of iron 
 ore, carbon, and such other ingredients as are determined 
 by the nature of the product required. 
 
 118 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The furnace construction is very simple, consisting of 
 a refractory crucible, provided with a domed cover of 
 loose fire-bricks, which can be removed for charging pur- 
 poses, and through which pass two similar vertical elec- 
 trodes. These latter do not extend to the bottom of the 
 crucible, but only to a level with the upper tapping-hole, 
 provided for the removal of the slag formed in the smelting 
 process. 
 
 The molten metal collects at the bottom of the crucible 
 below the level of, and consequently out of contact with, 
 the electrodes, and is drawn off at a separate tapping-hole, 
 situated at the lowest point of the hearth. 
 
 The principle of the arc smelting furnace, devised by 
 Dr. F. C. Weber, consists in allowing the raw material, 
 or ore, to fall vertically through the intense heat zone pro- 
 duced by a number of arcs, into a suitable receptacle placed 
 below. 
 
 The furnace comprises a tower-like structure, with the 
 usual feed hopper at its upper extremity, around the point 
 of entry of which, into the furnace, is arranged an outer per- 
 forated jacket, to permit the escape of the combustion gases. 
 
 Below, the structure broadens out into a species of tunnel 
 for the reception of a removable truck to contain the molten 
 metal smelted by the arcs. The electrodes for these latter 
 pass through the side walls of the furnace at varying heights, 
 just above the receiving chamber, being arranged in equal 
 horizontal rows of opposing carbons. Several of these 
 rows are arranged, one above the other, constituting, when 
 active, a perfect sheet, or bank of arcs, through which the 
 raw material has to pass in falling to the base of the furnace. 
 
 Each arc is given independent control, so that its in- 
 tensity and consequent heating effect may be regulated 
 at the will of the operator, the usual procedure being to 
 graduate them in such manner that the most intense heat 
 is at the top, where the charge first comes into^contact 
 with them. 
 
 119 
 
ELECTRIC FURNACES AND 
 
 To prevent current leakage between the electrodes other 
 than through the legitimate path of the arc, they are passed 
 through the furnace walls in such directions as to take the 
 longest possible path, viz. at an angle to both horizontal 
 and vertical planes. 
 
 A convenient arc furnace for the reduction of metallic 
 ores, manufacture of steel, etc., has been devised by M. 
 Paul Heroult. It involves the production of two arcs 
 in series, or may, if the electrodes are submerged in the 
 raw material to be treated, be regarded as a resistance 
 furnace, in which the conducting path is constituted by 
 an upper layer of the unconverted charge. 
 
 In general principle it is extremely simple, comprising 
 an open crucible with a refractory lining, provided with 
 the necessary tapping holes, and two similar vertical elec- 
 trodes, between the respective extremities of which, and 
 the surface of the conducting charge, the two arcs are struck. 
 
 The electrodes are mounted in massive holders attached 
 to screwed stems permitting vertical adjustment, means 
 for which are provided in the shape of bevel gear, and 
 hand- wheels, mounted on swinging brackets, attached to 
 suitable supports on either side of the crucible. Two 
 separate voltmeters, connected between the two electrodes 
 and the molten mass under treatment, respectively, serve 
 as gauges of the condition of the two arcs, whereby an 
 equal adjustment can be secured. 
 
 One form of the Heroult apparatus for the direct pro- 
 duction of steel resembles, in general construction, its 
 prototype, the Bessemer converter. 
 
 It consists of a steel structure, shaped somewhat like 
 a converter, and furnished with the usual pouring lip and 
 chimney, or gas outlet. It is lined throughout with re- 
 fractory material, such as fire-clay, and, rounded at its 
 base, to permit the necessary tilting, is mounted for con- 
 venience on toothed rockers, gearing with fixed rails, 
 similarly cogged. 
 
 120 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Rigidly attached to the back of the crucible, and thus 
 moving with it, are vertical supports, carrying horizontal 
 arms for the reception of the carbon electrodes, which are 
 two in number, placed side by side, the arcs being struck 
 in series, thus obviating the necessity for any electrical 
 connexion to the crucible itself. 
 
 A rack and pinion motion, worked by a hand-wheel, in 
 combination with a clamping device, serves to adjust the 
 arcs, and fix them in position. The mode of terminal 
 attachment to the carbon electrodes, which are square in 
 section, is somewhat novel, and deserves special mention. 
 
 The electrodes themselves abut, through the medium 
 of one face only, against the plane surface of the clamp 
 attached to the horizontal supporting arm. Around them 
 pass, loosely, two flexible metal bands, also attached at 
 their free extremities to the supporting arms. The con- 
 tact between electrode and support is made mechanically 
 and electrically secure by driving in copper wedges between 
 these bands and the electrode surfaces, a proceeding which 
 not only renders the whole construction rigid, but provides 
 a path of very low electrical resistance between metal and 
 carbon. 
 
 In larger models of the above converter the offices of 
 oscillating the crucible and adjusting the carbon electrodes 
 may be performed by hydraulic power instead of manually, 
 whilst the pouring lip may be extended into a ladle, per- 
 mitting the resultant metal to be cast direct from the 
 furnace. 
 
 According to Goldschmidt, Heroult produces in his 
 furnace a tool steel of the finest quality from a raw mixture 
 of cast iron and steel scrap. 
 
 Each charge weighs approximately three tons, the elec- 
 trodes being placed above, and an alternating current of 
 4,000 amperes at 60 volts, employed in producing the arcs. 
 
 The percentage of impurities in the product is estimated 
 to be 
 
 121 
 
ELECTRIC FURNACES AND 
 
 Sulphur 
 
 Silicon 
 
 Manganese 
 
 Phosphorus 
 
 Carbon 
 
 0-007 per cent. 
 0-003 
 0-150 
 0-003 
 0-60 to 1-80 
 
 At the time of writing a plant is in contemplation, having 
 a capacity of 600 H.P., and capable of dealing with 15 tons 
 of metal every twenty-four hours, in three operations of 
 the furnace. 
 
 Fifteen hundred tons of steel have already been made 
 by the Heroult method in France and Sweden, the process 
 being exploited, in the latter country, by the Aktiebolaget 
 Heroults Electriska Stal. 
 
 Heroult's procedure in making steel by the electric 
 furnace method consists in first melting a charge of scrap 
 iron, either per se, or mixed with pig-iron. The molten 
 mass is then covered with an artificial slag, and subjected 
 to the action of the arc, the sole object of which is to raise 
 it to the necessary temperature. The first slag is then 
 poured off, and new fluxes introduced, whereby all im- 
 purities are removed, and, on adding carbon and other 
 ingredients incidental to the quality of steel required, 
 perfect deoxidation is obtained. 
 
 The works of the Societe Electrometallurgique Franaise, 
 at La Praz, Savoy, who manufacture aluminium, and also 
 ferro-chrome and steel, by the Heroult process, utilize a 
 maximum of 14,000 H.P. The machines recently installed 
 for electrical steel manufacture are 500 H.P. single-phase 
 alternators, with an output of 2,750 amperes at 110 volts, 
 and 33 periods per second. 
 
 A later design of electric furnace, recently patented by 
 Heroult, for the production of ferro-silicon, ferro-manganese, 
 cast iron, etc., presents several novel points. It is, how- 
 ever, not an independent piece of apparatus, but is intended 
 to be used in conjunction with a preparatory furnace of 
 
 122 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the ordinary type, in which the ores to be reduced are 
 primarily heated and softened before being subjected to 
 the action of the electric furnace described below. 
 
 A sectional elevation is represented in Fig. 29. It 
 works on the resistance principle, the heating resistance 
 taking the form of a column of coke C, which is fed into 
 the furnace conjointly with 
 the ores, and at the same 
 time serves as a reducing 
 agent. A B are the two 
 inlet or feed passages ; the 
 former sloped as shown to 
 facilitate the entry of the 
 already softened ore, r, and 
 the latter, a vertical funnel, 
 or chimney, for the coke 
 feed. 
 
 The hearth, where the 
 principal fusion takes place, 
 is provided by the carbon 
 crucible D, which also con- 
 stitutes the negative elec- 
 trode, being provided with 
 suitable means for connexion FIG. 29. 
 
 to the external circuit. 
 
 The positive electrode E also consists of a carbon block, 
 situated in the upper portion of the structure, just above 
 the mouth of the ore inlet A. In the upper wall of this 
 latter is arranged a flue F, through which the hot gases 
 evolved during the process make their escape, and pass 
 over into the preparatory furnace, where they are utilized 
 in maintaining the necessary heat. H is an intermediate, or, 
 as the inventor terms it, " false " electrode, also of carbon. 
 
 The course of the current is from the upper electrode E, 
 across the ore feed inlet to H, and thence through the 
 column of mixed coke and ore to D. 
 
 123 
 
ELECTRIC FURNACES AND 
 
 A primary heat zone is thus created at the point of entry 
 of the ore, where it is most needed, and the process is con- 
 tinued through the mixed column as it passes down to the 
 crucible D, where the molten metal is collected, and drawn 
 off at the tap-hole t ; s is a similar tap-hole, at a higher 
 elevation, for the removal of the slag. 
 
 Marcus Ruthenberg has confined his attention mainly 
 to the design and perfecting of furnaces and processes for 
 the treatment of such ores as, by virtue of their finely 
 divided state, offer special difficulties in the way of treat- 
 ment and reduction by ordinary smelting furnace methods. 
 Such ores are, not infrequently, exceedingly rich in metal ; 
 any process, therefore, tending towards their successful 
 commercial reduction, should be welcomed by metallurgists 
 and the mining world generally. 
 
 An ingenious furnace for agglomerating magnetic sand, 
 
 an ore extremely rich in 
 pure iron, is represented in 
 Fig. 30. It takes the form 
 of twin cast-iron hoppers, 
 H H, hinged together at 
 the top, as shown, by an 
 insulating joint, and pro- 
 vided with a ring R for 
 suspension from a suit- 
 able support. These hop- 
 pers converge at the bot- 
 tom to opposing discharge 
 orifices, d d, which are 
 water- jacketed as shown, 
 for the purpose of keep- 
 ing them cool, and thereby 
 preserving them from de- 
 struction by the great 
 heat evolved. 
 
 FIG. 30. Terminal connexion with 
 
 124 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the hoppers is made at points near these discharge open- 
 ings, and a screwed rod, and hand wheel adjustment W, 
 about half-way up, serves to regulate the distance between 
 them at the will of the operator. 
 
 M M are electro-magnet windings encircling the upper 
 portion of the hoppers, and energized by a current when 
 the furnace is working. Both hoppers are thus rendered 
 magnetically active, and communicate their magnetism 
 to the particles of magnetic ore, or sand, which they con- 
 tain. Under the magnetic influence, these particles adhere 
 together, and, as they emerge from the discharge orifices 
 below, agglomerate into masses, which are then subjected 
 to the heating effect of the main current passing through 
 them from hopper to hopper, the latter acting as terminal 
 electrodes. 
 
 The agglomeration is thus rendered permanent, and the 
 ore falls away, in plastic masses, into the crucible C placed 
 below to receive it. In this form it is more readily treated 
 by ordinary smelting processes. 
 
 Another patent, granted to the same inventor, for 
 agglomerating magnetite, is equally ingenious in its design 
 and construction. It consists of a cast-iron hopper situated 
 above, and discharging immediately on to the periphery 
 of a cast-iron cylinder below. The cylinder and hopper 
 form the electrodes of the main furnace circuit, and the 
 former, driven by a motor, through a worm gear, auto- 
 matically draws the issuing magnetite away from the dis- 
 charge orifice of the hopper, whilst at the same time it is 
 agglomerated, as in the previous process, by the heating 
 effect of the current. The discharge orifice of the hopper 
 is, of course, protected, as in the previous furnace, by a 
 cooling water-jacket. 
 
 If so desired carbon and the necessary flux ingredients 
 may be introduced to form a mixture with the crude mag- 
 netite before subjecting it to the action of the furnace, and 
 the ore thus reduced to metal in a single operation. 
 
 125 
 
ELECTRIC FURNACES AND 
 
 The Ruthenberg process, which was demonstrated before 
 a committee of experts in January, 1903, at the works of 
 the Cowles Electric Smelting Co., West Lockport, N.Y., 
 consists, briefly, in first purifying the raw ore to the highest 
 possible degree, incorporating it with powdered reducing 
 material in the shape of charcoal, or coke dust, and 
 then subjecting it to the action of one of the furnaces de- 
 scribed above, from which it emerges as a reduced, fritted 
 mass, suitable for direct use in steel manufacture without 
 any further treatment. 
 
 Some 33 per cent, of fuel, and all the limestone incidental 
 to the ordinary blast furnace process is thus saved, whilst 
 the resultant steel is stated to be of a better and purer 
 quality than that produced by ordinary combustion methods 
 pure and simple. 
 
 The process entails the combustion of 500 kgs. of soft 
 coal per ton of product, of which 150 kgs. are necessary to 
 effect the reduction. The expenditure of energy necessary 
 in other processes to maintain the fusion of the entire mass 
 is, in the Ruthenberg process, dispensed with, only a small 
 portion of the ore, that bridging the actual terminals, and 
 constituting the resistance core, being treated at a time. 
 
 The terminals of the electro-magnetic arrangement are 
 identical with those of the furnace proper, there being no 
 electrodes, in the ordinary acceptation of the term, though 
 the water-cooled discharge openings of the hoppers really 
 act in this capacity. 
 
 Prior to its introduction into the furnace hoppers the 
 ore is purified as far as possible, all earthy and cinder- 
 forming matter being eliminated. It is then crushed, the 
 degree of granulation, or comminution, being largely deter- 
 mined by the quality of the ore ; low grade ores require 
 to be finer than those rich in metal. The crushed ore is 
 then mixed with the necessary ingredients, and in the 
 proportion indicated by the desired analysis of the product, 
 and introduced into the hoppers. The high temperature 
 
 126 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 attained at the discharge orifices fuses it into a homo- 
 geneous mass, denuded of a portion of its oxygen. The 
 remaining complement of oxygen is eliminated in course 
 of cementation, which is secured by keeping the mass 
 heated after its discharge from the electric furnace. The 
 final treatment or fusion is effected in an ordinary open 
 hearth furnace. 
 
 Ruthenberg states that the energy required in smelting 
 one ton of iron ore by his process is 500 k.w. hours. He 
 further claims that his process of electrical steel manu- 
 facture reduces the cost 20 per cent., and that the product 
 is superior to that manufactured by the Bessemer process. 
 
 He has, in addition to the apparatus already described, 
 also patented another resistance furnace process for the 
 smelting and reduction of iron ores ; the process consists 
 essentially in thoroughly incorporating a carbonaceous 
 reducing material with the broken fragments of ore, coking 
 it in place, and reducing the resultant mass to spongy iron, 
 which is finally fused. 
 
 The resistance furnace in which these several phases of 
 the complete process are effected has horizontal electrodes, 
 which pass through the walls in the neighbourhood of the 
 hearth, the heating resistance circuit being completed by 
 a mass of the molten metal. Immediately above this is 
 the fusion zone, in which is effected the fusion of the spongy 
 metal, which has already been reduced, at a still higher 
 level, in the reduction zone. The coking zone is above 
 this latter again, and the process is effected by the heat 
 from the waste gases, which play over the mass immediately 
 after it has been charged into the furnace. 
 
 It is rumoured that a number of Canadian capitalists 
 have purchased the patent rights of the Ruthenberg pro- 
 cess of electric smelting and steel production, which lends 
 itself readily to the treatment of iron, sand, and magnetic 
 iron ores. They will construct a plant on the Welland 
 River, at Chippewa, utilizing Niagara power from the 
 
 127 
 
ELECTRIC FURNACES AND 
 
 Canadian side. Their object is to work the extensive 
 deposits of magnetic ore in Ontario and Quebec, in the 
 vicinity of which fuel is scarce, but water-power plentiful, 
 a double recommendation for the installation of an electric 
 furnace plant. The outcome of this industrial venture 
 will be awaited with interest. 
 
 The development of the process by its inventor has been 
 of an encouraging nature. In a paper before the American 
 Electro-Chemical Society in 1902, he quoted the work per- 
 formed upon one ton of ore by the electric furnace as in 
 the neighbourhood of 500 k.w. hours. In a subsequent 
 paper (1903) the inventor was able to announce that he 
 had brought the consumption of electrical energy, for a 
 similar quantity of ore, down to one-half the original figure, 
 or 250 k.w. hours. 
 
 The Conley furnace, so named after its inventor, Michael 
 R. Conley, and exploited by the Conley Electric Furnace 
 Company, is of a simple design, which lends itself to a 
 variety of operations. 
 
 The furnace, Fig. 31, is cylindrical, being constructed, 
 as usual, of fire-brick F. The upper portion is funnel- 
 shaped, as shown, and constitutes a feed hopper, the walls 
 of which converge to a contracted portion or neck, n, 
 about two-thirds down, where the primary heat zone is 
 situated. Below this contraction the furnace expands 
 into a hearth H provided with a tapping hole t, where 
 the molten furnace products collect, and are drawn off. 
 Midway between the contraction and the floor of the hearth 
 is placed a second heat zone, n2, which either per se, or in 
 combination with the primary zone, n, permits a variety 
 of combinations in the way of electrical connexion of the 
 mains carrying the current. 
 
 The furnace is of the resistance type, the incandescent 
 portion consisting partly of the electrodes e e and, within 
 the furnace, of the charge itself. The electrodes e e are 
 segmental in form, and consist of a graphite clay mixture ; 
 
 128 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 they are mounted between segmental pillars of the fire* 
 brick structure, in such manner as to permit of consider- 
 able adjustment to compensate for 
 combustion and wear. They meet 
 at the point of entry into the fur- 
 nace chamber, and form a contin- 
 uous band, which, in conjunction 
 with the conducting charge, be- 
 comes incandescent when the re- 
 quired current is passed through. 
 Each segmental electrode is pro- 
 vided with a terminal, and it will 
 be seen that, by varying the elec- 
 trical connexions to the furnace, 
 as stated above, the direction of 
 maximum heat can be varied at 
 will, in either a horizontal or ver- 
 tical direction. 
 
 As already mentioned, the furnace 
 is designed for a number of uses, 
 chief among which are the reduc- 
 tion of iron ores and the manu- 
 facture of steel. 
 
 In Conley's latest form of resist- 
 ance furnace for the smelting of iron 
 ores, and the manufacture of steel, 
 
 with facilities for direct casting from the furnace itself, certain 
 modifications have been made in the original design, de- 
 scribed above. It consists, as before, of a fire-brick struc- 
 ture, the upper portion of which is funnel-shaped, and has 
 a central rib, or partition, also of fire-brick. The walls 
 of the funnel converge to a neck, where the principal heat 
 zone is located. Three resistance plates, consisting of 
 one part graphite to three of fire-clay, are introduced at 
 this point, and provided with terminals for connexion to 
 the external circuit. The section of the central portion 
 
 129 K 
 
 FIG. 31. 
 
ELECTRIC FURNACES AND 
 
 of these plates is reduced, in order to increase the resist- 
 ance, and, consequently, the heating effect. Below this 
 heat zone the furnace expands, as before, into a hearth of 
 larger diameter, where the molten metal collects and is 
 maintained in a fused state by a carbon resistance belt, 
 also provided with external terminals, as in the previous 
 design. 
 
 An especial feature of the furnace consists in a magnesite 
 lining, which is applied to all exposed carbon surfaces of 
 the resistances, and prevents direct contact between them 
 and the charge of ore to be reduced ; the carbon required 
 for the reaction can thus be exactly computed, and the 
 necessary percentage added to the ore without fear of 
 subsequent contamination by carbon particles which might 
 otherwise become detached from the electrodes themselves. 
 
 Conley's estimate of the cost of steel production by his 
 process runs as follows : Working on ores, and producing 
 at the rate of 100 tons per day : 
 
 
 Electrical energy . . . . . . .50 
 
 Thirty tons of coke at 8s. . . . . . 12 
 
 Two hundred tons of ore (65 per cent, metal) at 14s. . 140 
 Repairs and maintenance . . . ' 10 
 
 Labour ,~ .-. . . . . v . 25 
 
 Total . . . ; ... 237 
 or an average cost of 2 7s. 5d. per ton. 
 
 Working on pig iron and scrap, and producing at the 
 
 rate of 24 tons per day 
 
 s. d. 
 Electrical energy . . . . 12 10 
 
 Twelve tons of wrought iron at 112s. 
 
 ,, ,, cast iron scrap at 64s. 
 
 Repairs and maintenance 
 Labour . '' . 
 
 67 4 
 38 8 
 500 
 13 
 
 Total . ' . ; . 136 2 
 
 or an average cost of 5 13s. Qd. per ton. 
 
 Contardo's furnace for the reduction of metals such as 
 
 130 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 iron from their ores and the manufacture of steel, consists 
 (Fig. 32) of three distinct parts, or chambers : (A) a 
 funnel, or feed hopper, with suitable regulating device, and 
 gas outlet, for the introduction and control of the charge 
 of raw material or ore to be reduced. (B) an intermediate 
 portion, or hearth, which is the seat of the necessary heat, 
 provided by an arc or arcs, 
 struck between diametrically 
 opposed electrodes E, and 
 shielded from direct contact 
 with the charge by an inverted 
 V-shaped roof R of refractory 
 material, thus protecting the 
 charge from the introduction of 
 impurities, such as might be- 
 come detached from the elec- 
 trodes themselves ; and (C), a 
 receiving chamber, or crucible, 
 for the reception of the molten 
 metal, provided with the usual 
 tapping hole at its lowest point, 
 and within the walls of which, 
 communicating with suitable 
 inlets, are passages, or flues F 
 for the introduction of reducing 
 and carburetting gases, respec- 
 tively, into the presence of the 
 heated mass. The nature of 
 the gases introduced through 
 
 these flues is determined by the required composition of 
 the finished product, and may, if the process be one of 
 smelting pure and simple, be either omitted altogether, 
 or regulated as desired. 
 
 The Stassano process for the electrical production of 
 iron and steel, consists, briefly, in first subjecting the ore 
 to a process from which it emerges in a state of fine sub- 
 
 FIG. 32. 
 
ELECTRIC FURNACES AND 
 
 division. Lime and coke, also ground to a fine powder, 
 are then mixed with it in the necessary proportions, and 
 the mixture, aided by a suitable binding material, is moulded 
 into briquettes, in which form it is subjected to heat in 
 specially designed arc furnaces. 
 
 The process is continuous, the resultant metal and slag 
 being tapped off at regular intervals. An experimental 
 plant was installed at Cerchi, in Italy, the furnace being 
 three metres high, and calling for an expenditure of 1,800 
 amperes at 50 volts for an output of 30 kgs. of metal per hour. 
 
 In the original experiments with the Stassano process 
 of iron and steel production, 1 kg. of manganiferous steel 
 involved an expenditure of 4'08 E.H.P. hours ; this figure, 
 has, however, since been reduced to 2*7 E.H.P. hours for 
 an equivalent output. 
 
 The process consists, in the main, in substituting an 
 arc furnace for the blast furnace at present employed, and 
 feeding the material to be treated, into it, in the form of 
 briquettes, consisting of iron ore and carbon, lime, and 
 tar, as a binding material. The process is synonymous 
 with that which takes place in a blast furnace, the ores 
 being reduced by the carbon and the siliceous remainder 
 removed in the form of slag. 
 
 The ores most usually treated are those in which the 
 metal exists as a carbonate or oxide ; if the former, the ore 
 requires to be initially roasted. Previous to subjecting 
 them to the action of the arc, all ores are powdered and 
 mixed with the proportions of carbon, lime, and silica, 
 which a prior analysis of the ore may show to be necessary. 
 From 5 to 10 per cent, of pitch is then added as a binding 
 material, and the mass moulded into briquettes by hydraulic 
 pressure. 
 
 About half an hour is the time required in producing suffi- 
 cient metal to form a small ingot, the power consumption 
 being in the neighbourhood of 3,000 H.P. for each ton of 
 metal produced. 
 
 132 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The Stassano electric smelting works are in operation at 
 Darfo, near the Lago d'Isco. According to Goldschmidt, 
 Stassano has succeeded in producing soft iron, with a carbon 
 content less than 0'2 per cent., directly from iron ore on a 
 commercial scale in the electric furnace. It is necessary to 
 add that very pure ores are available, as the following table 
 of chemical compositions will show 
 
 CHEMICAL COMPOSITION OF ITALIAN ORES. 
 
 SUBSTANCE. MAGNETITE. KED HAEMATITE. LIMONITE. LIMONITE. 
 Fc 3 O 4 . 78-400% 
 
 88-850% 80-930% 73-840% 
 
 0-470% 0-567% 0-567% 
 
 Fe O 
 
 2^3 
 
 MnO 
 
 8i0 2 
 
 Al a O : 
 
 CaO 
 
 MgO 
 
 S . 
 
 P . 
 
 8-650% 2-960% 1-970% 1-970% 
 
 7-330% 3-420% 2-152% 2-152% 
 
 0-870% 0-590% 0-590% 
 
 1-030% 
 
 0-078% 0-070% 0-070% 
 
 0-093% 0-124% 0-124% 
 
 A consideration of the energy required, and cost, of the 
 electrical process devised by Stassano, yields the following 
 figures. Working at an efficiency of 80 per cent., 1 H.P. hour 
 is equivalent to 508' 2 Calories, or the heat derived from 
 1 kilogramme of coal is equal to 3 H.P. hours. Cost of insta- 
 lation, per E.H.P., 12 ; annual cost (10 per cent, for interest 
 and depreciation, 1 4s., maintenance, 8s.) 1 12s. The cost 
 of 1 H.P. hour, on the basis of 7,000 working hours per 
 annum, is therefore 0'055d. 
 
 In the Stassano furnace, the charge is heated solely by 
 radiation, the heat of the arc, which is maintained by 2,000 
 amperes at 170 volts, alternating current, being reflected on 
 to the charge by a domed roof. 
 
 The electrodes pass through water-cooled sleeves, and are 
 supported by conducting rods, sliding in gas-tight glands, 
 and adjustable therein, by means of small hydraulic cylin- 
 ders, which afford a simple and convenient means of feeding 
 or withdrawing the carbons. Once the action is started, the 
 arc is drawn out to as great a length as one metre, the oper- 
 
 133 
 
ELECTRIC FURNACES AND 
 
 ation of the furnace being accompanied by considerable 
 noise. 
 
 A great measure of the success attained by Captain Stas- 
 sano in his electric smelting process is attributed by Gold- 
 schmidt to the care bestowed upon the composition of the 
 furnace charge, with a view to securing complete reduction 
 and a maximum yield. The following table represents the 
 percentage composition of a typical Stassano furnace charge, 
 which, thoroughly incorporated, and mixed with tar as a 
 binding material, is moulded into briquettes prior to its in- 
 troduction into the furnace. 
 
 ORES. FLUXES. COAL. TAR. 
 
 Fe 2 O 3 93-02% CaO . 51-21% C . 90-42% C. . 59-2% 
 
 MnO . 0-610% MgO. 3-11% Ash . 3-88% Hydro- 
 carbons 40-5% 
 
 Si0 2 . 3-79% A1 2 3 \ - 0/ H 2 . 5-70% Ash . 0-27% 
 
 S . . 0-058% Fe 2 O 3 J /0 
 
 P . . 0-056% SiO 2 . 0-9% 
 
 0-5% C0 2 ._43.43% 
 
 H 2 . 1-72% 
 
 A charge consists of ores, 100 kgs. ; coal, 23 kgs. ; and 
 fluxes, 12' 5 kgs., the various constituents being apportioned 
 as follows 
 
 ORES. COAL. FLUXES. 
 
 Fe 2 3 
 
 . 93-02 kgs. 
 
 C . 
 
 20-9 kgs. 
 
 CaO . 
 
 6-401 kgs. 
 
 MnO. 
 
 . 0-619 ~, 
 
 Ash . 
 
 0-892 
 
 MgO . 
 
 0-389 
 
 Si0 2 . 
 
 s 
 
 . 3-79 
 . 0-058 
 
 H 2 0. 
 
 1-21 
 
 A1 2 3 ) 
 Fe 2 0j 
 
 0-062 
 
 P . 
 
 . 0-056 
 
 
 
 Si0 2 . 
 
 1-125 
 
 CaO \ 
 MgO J 
 
 0-5 
 
 
 
 C0 2 . 
 
 5-429 
 
 H 2 
 
 . 1-72 
 
 
 
 
 
 
 
 The resultant metal has the following composition- 
 
 Iron 
 
 Manganese 
 Silicon 
 Sulphur . 
 Phosphorus 
 Carbon 
 
 99-764 percent. 
 0-092 
 Trace. 
 
 0-059 
 0-009 
 0-090 
 
 134 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The following thermal data regarding the heat required 
 for bringing about the various stages of the process are 
 quoted 
 
 For the reduction of 1 gramme-molecule of 
 Fe 2 3 . . . . 
 
 For the conversion of 1 kg. of water at 100C. 
 = 212F. into steam, at the same tem- 
 perature ...... 
 
 For raising the temperature of 1 kg. of steam 
 at 100 C., by 1C 
 
 For calcining 1 kg. of flux 
 
 For raising the temperature of 1 gramme- 
 molecule of CO 2 , 1C. . . 
 
 For raising the temperature of 1 gramme- 
 molecule of CO, 1C. 
 
 For fusing 1. kg. of iron .... 
 
 For fusing 1 kg. of slag .... 
 
 The combustion of 1 kg. of C to CO, generates 
 
 Required, for the decomposition of ferric 
 92-02x192 
 
 oxide 
 
 0-16 
 
 For the evaporation of the water contained in 
 the ore and coal (1-72 + 1-21) 637 . 
 
 Required, to superheat the steam to 500C. 
 = 932F., 2-93 x 0-048 x 400 
 
 Required, for calcining the flux 12-5 x425 . 
 
 Required, for raising the temperature of the 
 C0 2 toSOO'C. 51429 x 0-016x500 
 
 0-044 
 the CO 
 
 generated, 
 
 Required, for heating 
 
 On.Q 
 
 0^12 x 0-0068x500= . 
 
 Required, for fusing the iron produced, 
 
 65 x 350 . . . . . 
 Required, for fusing the slag, 13-89 x600 . 
 
 Total . 
 
 The oxidation of C, to CO, produces 20-9 x 
 2,175= 
 
 Balance 
 
 CALORIES. 
 192-000 
 
 637-000 
 
 0-480 
 425-000 
 
 0-016 
 
 0-0068 
 350-000 
 600-000 
 2,175-000 
 
 111,552-000 
 
 1,866-41 
 
 562-56 
 5,312-5 
 
 987-09 
 5,921-667 
 
 22,775-2 
 8,334-0 
 
 157,311-427 
 
 45,457-500 
 111,853-927 
 
 Assuming, therefore, an efficiency of 80 per cent., this 
 total of 
 
 135 
 
ELECTRIC FURNACES AND 
 
 111,853-927 Calories, = = 219'08 H.P. hours, 
 
 635-3x0-80 
 
 or, the yield being 65'114 kilogrammes of iron, the energy 
 required for the production of 1 kg. is 
 
 219-080 = 3 . 364 p h 
 
 65-114 
 
 and the cost, based on the foregoing estimate of 0'055d. per 
 H.P. hour, works out at 155. 5d. per ton. 
 
 A smaller model of the Stassano furnace, which is used 
 for experimental purposes at Darfo, requires 1,000 amperes 
 at 80 volts. 
 
 Turin is at present the scene of Stassano' s electric smelt- 
 ing operations, where he is conducting experiments with his 
 process, on behalf of the Italian Government. 
 
 Assuming a plant of 5,000 H.P., producing steel at the rate 
 of 30 tons per day, with a thermal efficiency of about 66 per 
 cent., Stassano estimates the cost of manufacturing 1,000 
 
 kgs. to be as follows 
 
 s. d. 
 
 1,600 kgs. of ore at 12s. per ton . . . 19 2 
 
 For pulverization of same at 2s. 5d. per ton. 3 10 
 
 200 kgs. of flux at 4s. per ton . . 10 
 
 250 kgs. of coke at 36s. per ton 
 
 Crushing same at Is. 7d. per ton 
 
 190 kgs. of other materials (tar, etc.) at 56s 
 
 per ton 
 
 Mixing at 2s. 5d. per ton 
 Wear of electrodes, 12 kgs. 
 Labour 
 Maintenance of furnace 
 
 090 
 005 
 
 10 8 
 055 
 2 lO 
 4 9 
 097 
 025 
 
 Accessories 
 
 Electrical energy, 4,000 H.P. hours at 0-0o5d. . 18 3 
 
 Other costs . 0. . . . . 2 4 
 
 Total 498 
 
 Less the value of inflammable gases, 900 cubic 
 
 metres at 0-1 92d. . . . . 14 5 
 
 Nett Cost . . . . 3 15 3 
 
 Dr. Goldschmidt, in a paper before the Electro-Chemical 
 
 136 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Society of Cologne (1903), gave some interesting particulars 
 regarding the Stassano process. The following are a few of 
 the salient points mentioned. 
 
 The current used in the experiments which have been 
 carried out at Darfo, in Italy, is alternating, and is derived 
 from two large dynamos, each of 500 H.P., and a smaller one 
 of 100 H.P. The process, like other reduction methods, con- 
 sists essentially in reducing the iron ores with charcoal or 
 coke, by virtue of the heat derived from an electric arc, work- 
 ing above the charge, suitable fluxes being introduced, to 
 eliminate undesirable impurities. 
 
 The latest form of Stassano furnace consists of a closed 
 cylindrical structure, with domed roof, constructed of 
 masonry, and having a cavity one metre in diameter and 
 one metre high. A sloping feed channel serves for the intro- 
 duction of the ore, whilst two tapping holes, at different 
 elevations, provide for the withdrawal of the molten metal 
 and slag respectively. The entire furnace can be revolved 
 around a vertical axis. 
 
 The electrodes are introduced through the walls at oppo- 
 site extremities of a diameter, and at or about the centre of 
 the furnace cavity; at the commencement of the process 
 their inner extremities are close together, but, as the opera- 
 tion proceeds and the temperature increases, they are with- 
 drawn until the arc extends across the entire width of the 
 furnace cavity. 
 
 With the arc at its maximum, the current is so regulated 
 that in 
 
 20 minutes, the E.M.F. is 80 volts, and the current 800 amperes. 
 40 100 1,000 
 
 70 70 600 
 
 100 50 500 
 
 The entire reduction process occupies two hours, from 
 first switching on, until the reduced metal is ready for cast- 
 ing, during the first half of which period the charge is being 
 introduced. 
 
 137 
 
ELECTRIC FURNACES AND 
 
 The following is the composition of a typical charge, which 
 was reduced by a current of 1,000 amperes at 80 volts- 
 Iron ore . . 1,000 parts. 
 
 Limestone . . . . . . 125 
 
 Carbon . . . . . . . 160 
 
 Miscellaneous (hydrocarbons, etc.) . . 120 
 
 The total weight of the above charge was 70' 25 kgs., and 
 the following table represents the original weight of redu- 
 cible substances, together with the weights actually obtained 
 in the process 
 
 SUBSTANCE. ORIGINAL WEIGHT (GRAMMES). PRODUCT (GRAMMES). 
 Iron 32,557 30,727 
 
 Manganese 239-7 28-3 
 
 Silicon 910 Traces 
 
 Sulphur 29 15 
 
 Phosphorus 28 2-7 
 
 The actual yield was thus 30'8 kgs. of wrought iron. 
 
 Below will be found a tabular analysis of the impurities 
 in four different samples of steel manufactured by the Stas- 
 sano process. 
 
 I. II. III. IV. (CHROME 
 
 STEEL). 
 
 Carbon . . 0-04% 0-09% 0-17% 1-51% 
 
 Manganese . 0-05% 0-18% 0-17% 0-26% 
 
 Silicon . Trace Trace 
 
 Sulphur . 0-05% 
 
 Phosphorus 0-029% 
 
 Chromium . 1-22% 
 
 An iron and steel plant, at Gysinge, Sweden, employing 
 electric furnaces designed by M. F. A. Kjellin, is now produc- 
 ing steel of excellent quality, which is unusually dense and 
 homogeneous. It is very tough, and, when annealed, can 
 be readily worked in a cold state, without any tendency to 
 distortion consequent on hardening, as is frequently the case 
 with ordinary grades of steel. 
 
 The inventor attributes the improved quality of the steel, 
 manufactured by his process, to the almost complete absence 
 of gaseous matter, especially hydrogen. 
 
 138 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The capacity of the latest Kjellin furnace, installed at 
 Gysinge, is 1,880 kgs., and, charged with raw materials in a 
 cold state, it is capable of producing 1,500 tons of steel per 
 annum. It is supplied with current from a single-phase 
 alternating current generator, driven by a vertical turbine, 
 capable of developing 300 H.P. 
 
 The principal considerations which led to the design of 
 the Kjellin furnace are as follows. M. Kjellin early recog- 
 nized the fact that in operating steel furnaces on the ordinary 
 resistance principle, with independent electrodes, the molten 
 metal absorbs impurities from the latter, which, especially 
 if of carbon, are also consumed, and thus add to the expense 
 of the operation, whilst the carbon monoxide gas formed as 
 a result of the combustion, hinders the elimination of this 
 gas from the mass of the steel itself. On the other hand, if 
 carbon be dispensed with, and the current led direct to the 
 mass of metal under treatment, the enormous currents re- 
 quired necessitate the use of copper leading-in cables at 
 least equal, in cross-sectional area, to the resistance column 
 of metal. 
 
 Water-cooled iron terminals have been suggested, but 
 fresh difficulties arise out of the magnetization effects ; more- 
 over, the use of alternating currents, one of the desirable 
 incidentals to such a large power conversion, leads to fresh 
 complications, owing to skin effect, hysteresis, and conse- 
 quent loss of power. 
 
 With a view to surmounting these various difficulties, 
 Kjellin, in 1899, suggested to M. Benedicks, general manager 
 of the works at Gysinge, the form of construction now known 
 as the Kjellin furnace. It consists of an annular groove, or 
 hearth, built of refractory bricks, and closed in at the top 
 by refractory covers, which are removable for charging 
 purposes. 
 
 Centrally, within this ring is placed one limb of a rectan- 
 gular iron core, built up of thin sheet-iron laminations, upon 
 which is wound a cojpper wire coil, which constitutes the 
 
 139 
 
 OF 
 
 UNIVERSITY 
 
 f UN 
 ^S 
 
 
ELECTRIC FURNACES AND 
 
 primary winding of a closed-circuit, step-down transformer, 
 the secondary, in the shape of the contents of the annular 
 hearth, having only one turn. Current is supplied to this 
 primary winding at a pressure of 3,000 volts, and the ratio 
 of transformation is such that the current through the 
 furnace charge is approximately equal to the primary cur- 
 rent multiplied by the number of turns in the primary 
 winding, or about 30,000 amperes. 
 
 The first Kjellin furnace was installed at Gy singe in Feb- 
 ruary, 1900, and the first ingot of steel successfully cast on 
 the 18th of the following month, the expenditure of energy 
 being 78 kilowatts for 24 hours, and the yield 270 kgs. An 
 improved efficiency was secured in November, 1900, when a 
 second furnace was completed, capable of producing from 
 600 to 700 kgs. of steel in 24 hours, with an energy expendi- 
 ture of 58 k.w. With this construction, however, the avail- 
 able cooling surface, and consequent radiation losses were 
 out of all proportion to the furnace capacity, and, in 1902, 
 a neighbouring sulphite pulp mill having been destroyed by 
 fire a year previously, its available water power was adapted 
 to steel manufacture on the Kjellin principle, with the result 
 that a third furnace was constructed, and has been work- 
 ing satisfactorily since May, 1902. 
 
 It has a capacity of 1,800 kgs. of charge, 1,000 kgs. of steel 
 being tapped off at a time, leaving the remaining 800 kgs. to 
 complete the secondary circuit and avoid shutting down the 
 power plant. The output is 4,100 kgs. of steel in 24 hours, 
 with an energy expenditure of 165 k.w., or 225 E.H.P., the 
 charge being introduced into the furnace in a cold state. 
 
 The raw materials employed are Dannemore pig iron and 
 wrought iron, which are fed into the furnace along with scrap 
 steel, and in the proportions indicated by experience, as 
 yielding the best percentage of carbon. 
 
 M. Kjellin is reported to have spoken as follows regarding 
 his invention 
 
 " By the electric furnace described, the steel has no op- 
 
 140 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 portunity of taking up such gases, or other impurities, and 
 the quality is also better than that of crucible steel with the 
 same analysis. To make special steels with nickel, tungsten 
 or chromium, offers no difficulties, and the alloys are quite 
 homogeneous. The cost of production depends, principally 
 on the efficiency of the furnace and the price of power. 
 
 " At the furnace now in use at Gysinge, the losses have 
 been experimentally proved to be 87*5 k.w., so that the 
 effective power absorbed by the steel is 165 87*5 = 77*5 k.w., 
 and, as these produced 4,100 kgs. of steel in 24 hours, one 
 effective kilowatt produces about 53 kgs. of steel ingots in 
 the same time. Every kilowatt more in the furnace, when 
 the size is not altered, increases, then, the output with 53 kg. 
 steel ingots, and we calculate, when, within a few months, 
 we get a stronger water wheel, to produce about 6,000 kgs. 
 of steel ingots with 200 k.w. in 24 hours. 
 
 " As the absolute cost of labour and repair will be the 
 same, these costs for one ton of steel ingots will be about two- 
 thirds of the cost now, and the price of power, per ton, will 
 also be sensibly diminished. 
 
 " Thus, a furnace of 736 k.w., or 1,000 E.H.P. will produce 
 30,000 kgs. of steel ingots in 24 hours when charged with cold 
 materials. With hot materials the output is much greater. 
 For instance, if 250 kgs. of molten pig iron are charged for 
 each ton of steel ingots produced, the output is increased 
 from 30,000 to 36,000 kgs. in 24 hours with the same ex- 
 penditure of energy. 
 
 " The cost of labour and repair for such a furnace will, 
 in my opinion, be less that those of an open-hearth furnace 
 of the same size, so that, where power is cheap, there is a 
 possibility of producing a steel at a smelting cost not ex- 
 ceeding that of the open-hearth furnace." 
 
 The following is a detailed estimate of the cost of the 
 Kjellin process, based on data obtained in actual practice at 
 Gysinge 
 
 4,100 kgs. of steel are produced in 24 hours, with an 
 
 141 
 
ELECTRIC FURNACES AND 
 
 energy expenditure of 225 x 24 H.P. hours ; 1,000 kgs. re- 
 quire, therefore, 1,320 H.P. hours. 
 
 *. d. 
 Cost of electrical energy, assuming 1 H.P. hour 
 
 to cost 0-18d 100 
 
 Charge (pure charcoal, cast iron and wrought iron) 6 10 
 Repairs and lining . . . . . 9 3| 
 
 Depreciation and interest . . . .023 
 
 Labour 10 
 
 Total cost, excepting royalties, business charges, 
 
 etc . 8 11 6J 
 
 In the Kjellin process the furnace gases do not come into 
 contact with the steel. The cost of the furnace is given as 
 15,000 kronen (832). 
 
 The following table gives the analyses of three samples 
 of steel manufactured by the Kjellin process at Gy singe 
 
 I. ii. in. 
 
 Carbon . . 1-45% 1-20% 0-95% 
 
 Silicon . . 0-47% 0-74% 0-35% 
 
 Manganese . 0-49% 0-46% 0-33% 
 
 Phosphorus . 0-011% 0-013% 0-014% 
 
 Sulphur . 0-010% 0-010% 0-015% 
 
 Other Metals. One of the difficulties incidental to re- 
 duction processes, as carried out in the electric furnace, is 
 that the molten metal formed, readily absorbs the carbon 
 of which the electrodes are composed, and thereby becomes 
 contaminated. 
 
 Several inventors, realizing the importance of this draw- 
 back to electric smelting, have set themselves the task of 
 providing a remedy. E. G. Acheson has patented a method 
 of protecting the carbon core in resistance furnaces by coat- 
 ing it with a refractory sheath of a carbide, which, whilst 
 preventing actual contact between core and charge, never- 
 theless conducts the heat generated by the former to the 
 mass of the latter. 
 
 M. Simon, whilst investigating the possibilities of pre- 
 paring metallic manganese in the electric furnace, also 
 experienced this difficulty, but succeeded in overcoming it by 
 
 142 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 enveloping the anode in a sheath of manganese slag, in 
 which the metal existed as the lower oxide, and was there- 
 fore inactive in the presence of carbon. 
 
 George Egly protects the lower electrode or cathode, in 
 his furnace for the reduction of protoxide of nickel, by 
 covering it with a layer of magnesia, about 2 c.m. in thick- 
 ness. 
 
 In 1895, Moissan succeeded in producing silicon in the 
 electric arc furnace by reducing silica, in the form of quartz, 
 with carbon, in a coke crucible. Acheson has since, 1902, 
 applied the resistance furnace principle to this process, 
 utilizing graphite as a reducing agent. To this end, he 
 mixes silica and graphite in the proportions necessary to the 
 reaction, and forms the resultant mixture into a column 
 or pyramid, which he utilizes as a conducting core in a 
 resistance furnace. This admixture of graphite with the 
 silica, imparts sufficient conductivity to the mass ; thus, 
 it was found by experiment that a core two by two, by 
 three inches, had an initial resistance of 33 ohms, which 
 fell to 5 ohms as the reduction proceeded. 
 
 The success of the above and similar experiments has 
 led Acheson to take out a patent covering the applications 
 of graphite as a reducing agent for use in electric furnaces 
 of the resistance type. The essential properties which 
 fit it for use as such are 
 
 (1) Its high purity, and consequent inability to con- 
 taminate any ores heated in its presence. 
 
 (2) High electrical and thermal conductivity, whereby 
 the necessary current is transmitted and the resultant heat 
 evenly distributed. 
 
 (3) Extreme divisibility, permitting an advanced degree 
 of comminution, thereby lending itself to the production of 
 an intimate mixture with the ore to be reduced. 
 
 The Tone furnace, invented by Mr. P. J. Tone, engineer 
 to the Carborundum Company, Niagara Falls, has been 
 devised with a view to overcoming one of the drawbacks 
 
 143 
 
ELECTRIC FURNACES AND 
 
 incidental to electric smelting, where carbon is used as a 
 reducing agent, viz., the tendency on the part of the liber- 
 ated metal to recombine with the excess of carbon present 
 in the furnace charge, to form a carbide. As already stated, 
 this has been a common experience with metallurgists in 
 treating the ore of a metal whose temperatures of reduction 
 and volatilization are approximately the same. 
 
 In the Tone furnace, illustrated in Fig. 33, this ten- 
 dency is overcome by a more general distribution of the 
 
 heat generated in the furnace, 
 combined with a judicious dis- 
 tribution of the charge about 
 the central source of that heat. 
 The charge C is thoroughly 
 mixed, and so arranged about 
 the resistance core R as to 
 permit globules of the reduced 
 metal to form, and descend by 
 gravitation to the lower portion 
 of the structure, whence they 
 drip, finally, into the recept- 
 acles A A. E E are the ter- 
 minal electrodes, arranged at 
 the upper and lower extremities 
 of the core B, which consists 
 of a pyramidal pile of carbon 
 blocks, with intervening spaces. 
 
 The temperature of the furnace needs to be very carefully 
 regulated, such that the reduced metal is not volatilized, 
 whilst the heating effect is evenly distributed and general 
 throughout the mass. 
 
 Copper. The Heroult process has been applied to the 
 extraction of metallic copper from its ores. In 1903, ex- 
 periments conducted at La Praz, France, under the personal 
 supervision of the inventor, yielded satisfactory results. 
 In connexion with these experiments, the following 
 
 144 
 
 FIG. 33. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 details have been made public. Two carbon electrodes, of 
 square cross-section, were used, each side being 25 c.m. 
 The mineral, which contained 7 per cent, metallic copper, 
 was placed in a rectangular crucible, the electrodes de- 
 scending vertically into it. Eighteen tons of ore were 
 treated in 24 hours, the resultant matte containing from 
 43 to 45 per cent, copper, whilst the slag only retained from 
 O'l to 0*2 per cent, of the metal. The current required to 
 effect this result was from 3,500 to 4,000 amperes at 110 
 volts. 
 
 M.M. Keller, Leleux, & Co. have also conducted experi- 
 ments on the electrical extraction of metallic copper from 
 its ores, utilizing a Keller furnace with two centres of ac- 
 tivity, each provided with an independent pair of electrodes. 
 The actual reduction is effected in the first heat zone, the 
 second serving to maintain the high temperature of the 
 mass, pending the thorough separation of matte and slag. 
 
 Twenty-five tons of ore were treated in 24 hours, the 
 power consumed being 500 k.w. The consumption of 
 electrodes amounted to from 5 to 7 kgs. per ton of ore. 
 
 The composition of the resultant matte was as follows 
 
 per cent. 
 
 Silica 
 
 Aluminium 
 
 Iron 
 
 Manganese 
 
 Sulphur 
 
 Phosphorus 
 
 Copper 
 
 And of the slag 
 
 Silica 
 
 Alumina 
 
 Lime 
 
 Magnesia 
 
 Iron 
 
 Manganese 
 
 Sulphur 
 
 Phosphorus 
 
 Copper 
 
 0-8 
 
 0-5 
 24-3 
 
 1-4 
 22-96 
 
 0-005 
 47-9 
 
 27-2 
 
 5-2 
 
 9-9 
 
 0-39 
 32-5 
 
 8-23 
 
 5-7 
 
 0-062 
 
 0-1 
 
 per cent. 
 
 The above trials were carried out in the presence of M. Ch. 
 
 145 L 
 
ELECTRIC FURNACES AND 
 
 Vattier, who, acting under a commission from the Chilian 
 Government, has lately visited France for the purpose of 
 carrying out experiments in the electrical smelting of copper 
 ores. 
 
 M. Vattier's report contains some interesting comparative 
 figures relating to costs, etc. In the coke furnaces at present 
 in vogue in Chili, 3,200 kgs. of coke, costing 13, are required 
 for each ton of copper produced. On the other hand, a 
 consumption of 1*25 k.w. year, is estimated as sufficient 
 to smelt 16 tons of ore, corresponding to one ton of copper, 
 in the electric furnace. Assuming the cost of 1 k.w. year to 
 be 24s., a figure based on the abundance of water power in 
 Chili, the cost of energy per ton of copper produced would 
 amount to 30s., and, allowing 36s. for electrodes, etc., the 
 total cost of production per ton of metal works out at 
 3 6s. 
 
 Arsenic. The importance of arsenic as a marketable 
 commodity may be gauged from the facts, published by the 
 Mineral Industry, that the total output for 1900 was 7,300 
 metric tons, produced by Great Britain, Germany, Italy, 
 Spain, and Canada, the two first named being the principal 
 manufacturers. The demand is said to be quite equal 
 to the supply, so that there is little doubt as to the need for 
 this useful metal. 
 
 Unfortunately, the ordinary metallurgical processes for 
 its extraction from the various arsenic - bearing ores possess 
 many attendant disadvantages, not the least of which 
 arise out of the poisonous nature of the fumes given off 
 during the process, and their deleterious effect on the 
 surrounding animal and vegetable kingdom. 
 
 This is all the more unfortunate in that many of the 
 well-known arsenic bearing ores are also rich in gold and 
 other noble metals. An improved process, therefore, 
 calculated to increase the output and efficiency of ex- 
 traction, would doubtless be welcomed by metallurgists 
 in general. 
 
 146 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Such a process, applicable to at least one rich grade of 
 arsenic ore, has been invented by Mr. G. M. Westman, of 
 New York. It comprises, in effect, a resistance furnace 
 process, and the ores capable of treatment by it are known 
 as " mispickel," or arseno-pyrite ores, which also contain 
 ores of gold, silver, and other metals of value. They are 
 very rich in arsenic, averaging 46 per cent, by weight of that 
 metal, which exists in the form of sulph-arsenide of iron 
 (FeS 2 +FeAs 2 ). 
 
 In brief, the process of reduction consists in heating the 
 ore electrically, in a closed furnace, from which atmospheric 
 oxygen is hermetically excluded. The iron combines with 
 the sulphur, forming ferric and ferrous sulphides, which 
 are thrown down in the form of a fused matte, and include 
 the precious metals ; the arsenic itself is liberated as a heavy 
 metallic vapour, and collected in suitable condensers as a 
 fine, metallic powder. 
 
 The process is carried on in a circulating atmosphere of 
 nitrogen gas, formed, in the first place, from the atmosphere 
 by extracting the oxygen, by combination with a small 
 quantity of arsenic vapour, to form the oxide. Once pro- 
 duced in this manner, the same volume of nitrogen is, by 
 circulation, used repeatedly in the furnace. 
 
 The general construction of the latter is delineated in 
 Fig. 34, where F is the furnace proper, with a refractory 
 hearth or lining R of fire-brick, in which are embedded 
 the two cast-iron electrodes E E. 
 
 A central hopper H serves for the introduction of the 
 ore, but, at the same time, prevents ingress of air. M is 
 the mass of fused ore, which constitutes the resistance path 
 between the electrodes ; C C C C are condensers, with 
 bottom traps t t, in which the powdered arsenic collects 
 and is removed. The tube T completes the circulatory 
 system for the nitrogen gas, which is kept in motion by a 
 mechanically driven blower B. This latter is reversible, 
 such that when one set of condensers becomes heated, the 
 
 147 
 
ELECTRIC FURNACES AND 
 
 current of gas may be reversed, and the heat energy, thus 
 stored up, utilized for maintaining the general temperature 
 of the furnace, which would otherwise tend to fall. 
 
 FIG. 34. 
 
 An exhaustive test was conducted on an experimental 
 furnace of this type by Mr. Carl Hering, the results of which 
 were detailed in the Electrical World, April 27, 1901. The 
 following essential facts are extracted therefrom. 
 
 The process, owing to the small size of furnace employed, 
 was naturally wasteful of energy, but, by measurement 
 of some losses, and computation of others, a fair estimate 
 of the probable cost of reduction in a larger furnace was 
 arrived at. 
 
 An alternating current of 8,000 to 10,000 amperes at a 
 periodicity of 120, was used, the electrodes, which, as 
 already stated, were of cast iron, being about six square 
 inches in cross-sectional area, and rather long. 
 
 The losses, which totalled more than half the energy 
 supplied to the furnace, were due to several contributory 
 causes, viz., skin effect, hysteresis, resistance, Foucault, or 
 " eddy " currents, and conduction. The losses in leads alone, 
 from the secondary of the transformer to the furnace ter- 
 minals, amounted to 44 kilowatts, whilst 9 to 10 kilowatts 
 were lost by radiation and conduction. 
 
 In one trial 90 kgs. of ore were supplied to the furnace in 
 the hour, the power consumed in the conversion being about 
 
 148 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 105 k.w. ; the current varied from 4,000 to 8,400 amperes, 
 and the voltage from 22 to 12, representing a rate of approx- 
 imately 1,140 k.w. hours per metric ton of ore. 
 
 According to Mr. Hering, it is quite safe to assume that 
 in a large, well-designed furnace, the engine power required 
 in the reduction process would not be more than 1,000 k.w. 
 hours per metric ton of ore. 
 
 The theoretical heat energy required has been variously 
 estimated at from 200 to 400 k.w. hours per ton, which 
 points a margin for improvement. 
 
 The patent rights in the Westman process are the property 
 of the Arsenical Ore Reduction Company of Newark, New 
 Jersey, who are applying it to the large deposits of ore in 
 what is known as the Big Dan claim in Ontario, recently 
 acquired by them. 
 
 The following extract from Mr. Hering' s original article 
 contains some general information acquired during the 
 tests, which should prove useful in the design of electric 
 furnaces generally. 
 
 " Most of the losses point to the importance of using 
 continuous currents, if possible, as it is very difficult to 
 lead such very large alternating currents through conductors. 
 When, however, alternating currents must be used to 
 avoid electrolysis, as is often the case, then many difficulties 
 are reduced by proportioning the furnace so that the current 
 is as small as possible and the voltage as great as possible ; 
 i.e., the resistance column of the material should be made 
 as long, and as small in cross-section, as the circumstances 
 permit. The loop formed by the alternating current circuit 
 should have as small an area as possible, so as to diminish 
 the induction. The leads should, of course, be made as 
 short as possible, every inch saved in length being important. 
 The transformer should be placed within as few inches of the 
 furnace, as possible. Those parts of the conductors which 
 are unavoidable, as, for instance, where they pass through 
 the walls of the furnace, should be made of insulated lamina- 
 
 149 
 
ELECTRIC FURNACES AND 
 
 tions alternately of opposite polarity, otherwise the skin 
 and inductive effects will become important. For these 
 parts, the best electrically conducting material should be 
 used, so as to reduce the cross-section as much as possible, 
 as such large pieces of metal conduct considerable heat 
 away from the interior of the furnace. The frequency 
 should also be made as low as possible when such very large 
 conductors have to be used. 
 
 " Metallic casings of the furnace should be avoided, 
 particularly if of iron. If bands are necessary, they should 
 be made of a non-conducting material, or have insulated 
 joints in them to prevent closed circuits in the neighbour- 
 hood of the large alternating field necessarily produced 
 in the part of the circuit in the furnace itself. These are 
 some of the considerations underlying the design of large 
 alternating current resistance furnaces. 
 
 " The heat led off through the walls of the furnace, which 
 was about one metre cube, amounted to about 5 k.w., 
 the inner temperature being from 1,000 to 1,200C.= 1,832 
 to 2,192F., that necessary for the vaporization of arsenic 
 being only 450C. = 842F. The channel for the liquid 
 matte was about 50 c.m. long, and 10 c.m. wide. This 
 loss, which is not large for a furnace of about 150 H.P., could 
 be diminished only by having thicker walls, but it becomes 
 relatively less the larger the furnace. 
 
 " Theoretically, the following relations can be shown 
 to exist for different sizes of furnace, the assumption for 
 enabling a general law to be stated being, that the furnace 
 is a hollow sphere with relatively thick walls, and that the 
 temperature of the outside surface remains the same. The 
 capacity, or interior volume increases as the cube of the 
 inner diameter. The thickness of the walls will increase 
 with the square root of this diameter, and the radiation 
 losses, per unit volume of the capacity, or interior space, 
 will diminish in inverse proportion to the square of the 
 diameter. That is, for twice the inside diameter, the capa- 
 
 150 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 city becomes eight times as great, the walls 1*4 times as 
 thick, and the radiation loss per unit volume of the inside 
 becomes only one quarter as great, showing the great ad- 
 vantage of larger furnaces in diminishing this loss. For 
 rectangular furnaces, especially if quite long, these rela- 
 tions will not be quite as favourable, and these general 
 laws then apply only approximately. 
 
 " Heating the material quickly, i.e., making each particle 
 pass through the process as quickly as is consistent with the 
 proper operation, will, of course, reduce the size of the 
 furnace for the same output, but, assuming that a certain 
 definite number of heat units are required per ton of material, 
 as is generally the case, the larger furnace would require 
 the smaller current. Both should therefore be considered 
 in designing a furnace. 
 
 " In all electric furnaces, the fire-brick lining, when very 
 hot, becomes a conductor, a quality which is made use 
 of in the Nernst lamp. In resistance furnaces, this should 
 be taken into consideration, as it lowers the resistance 
 between the electrodes, and necessitates a greater current, 
 at a lower voltage, to produce the same heating effect. 
 This means that greater flexibility in the regulating apparatus 
 is required. The heat generated in this lining is not lost, 
 except in so far as it diminishes the thickness of the heat 
 insulating walls of the furnace. The conducting of the 
 lining can be prevented only by keeping it cool, which is 
 wasteful." 
 
 Zinc. The electrical distillation of metallic zinc from 
 its ores was a problem attacked by Messrs. Cowles Bros, 
 in the early eighties, the original type of resistance furnace, 
 invented by them for this purpose, being depicted in Fig. 35, 
 where F is the furnace proper, constructed of fire-clay, 
 in cylindrical form, and embedded for heat conserving pur- 
 poses in a refractory non-conducting bed B of considerable 
 thickness. The back of the furnace consisted of a carbon 
 plate or disc C, which also constituted a terminal elec- 
 
ELECTRIC FURNACES AND 
 
 trode, the other taking the form of a graphite crucible G, 
 which served the treble purpose of electrode, removable 
 
 cover, or plug, 
 and condensing 
 chamber for the 
 metallic zinc, 
 which distilled 
 over into it 
 through the 
 orifice o. It was 
 closed by a lid 
 L, and fitted with 
 FIG. 35. an outlet pipe 
 
 P for the escape 
 
 of the gases generated. The resistance heating column 
 was constituted by the mass of ore R itself. 
 
 It is probable that the practical difficulties mentioned in 
 connexion with resistance furnaces in the introductory 
 chapter of this book militated against the adoption of this 
 furnace on anything like a commercial scale. 
 
 De Laval's furnace for the distillation and reduction of 
 the ores of volatile metals, such as zinc, is adapted for use 
 on either the arc or resistance principles, as circumstances 
 may determine. 
 
 It is represented, in sectional elevation by Fig. 36, and 
 is so constructed that the charge does not come into direct 
 contact with the arc or heated resistance, but passes gradu- 
 ally beneath it. under the action of a reciprocating feed 
 arrangement, and receives, by radiation, a gradually aug- 
 mented supply of heat. 
 
 The apparatus consists of a refractory structure F hav- 
 ing the form shown in the figure, a feed hopper H com- 
 bined with a reciprocating piston P, and an outlet 0. 
 A is the arc, or heated resistance as the case may be, the 
 gradations of heat from which are secured by the contour 
 of the charge ; this latter, under the influence of the reci- 
 
 152 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 procating feed, assumes the slope shown ; it thus receives a 
 preliminary moderate heating at the apex A of the mass, 
 which gradually 
 increases in in- 
 ten sity as it 
 moves forward 
 under the source 
 of heat, the pro- 
 cess being finally 
 concluded im- 
 mediately below p IG 
 the latter. 
 
 A steady reaction is thus secured without a too violent 
 liberation of gas. 
 
 The process, as applied to the distillation of metallic 
 zinc from zinc lead ores, consists in introducing the pul- 
 verized ore, in stack form, as illustrated above. The radiant 
 heat of the arc, struck between two lateral electrodes, 
 causes the non-volatile constituents of the charge to melt 
 and flow down into the well of the hearth, whence they 
 are tapped off in the usual manner ; fresh surfaces of the 
 sloping stack are thus continually exposed to the arc as the 
 fusion proceeds. The volatilized zinc, together with cer- 
 tain gases, passes off through the outlet shown, and is 
 suitably collected and condensed. 
 
 To eliminate the sulphur in an untreated zinc sulphide 
 the inventor suggests mixing iron ore with the charge, prior 
 to its introduction into the furnace. 
 
 Messrs. 0. W. Brown and W. F. Oesterle have taken out 
 a patent for the electrical smelting of blended zinc sulphides, 
 one of the claims being that valuable by-products (calcium 
 carbide and carbon bisulphide) are formed in the process. 
 Coke, or carbon, is mixed with the ore, and the mixture 
 heated in an electric furnace to a temperature " sufficiently 
 high to produce the desired products." Metallic zinc 
 separates out, is volatilized and collected by condensation 
 
 153 
 
ELECTRIC FURNACES AND 
 
 in the usual manner, whilst the carbon, calcium, and 
 sulphur, also present in the ore, combine to form the above- 
 named by-products. 
 
 These details furnished by the patent specifications are 
 somewhat meagre, and there is no mention of its practical 
 application on an industrial scale. 
 
 Salgues describes (Paper before the Societe des Ingenieurs 
 Civils de France, 1903) a process of treating zinc ores in 
 the electric furnace. 
 
 The ore, in the form of the oxides and sulphides, is mixed 
 with a suitable flux, and subjected to the action of the 
 furnace in which it is fused, and a reaction started, either 
 in itself, or with suitable reducing agents, carbon, iron, etc. 
 The resulting slag contains practically no zinc, whilst the 
 latter is either tapped off, or distils over according to the 
 conditions of operation. 
 
 Treating ores containing 40 per cent, metallic zinc, in 
 furnaces of 100 k.w. capacity, nearly 5 kgs. of metal per 
 k.w. day have been obtained. 
 
 The operation is conducted in a closed furnace, and has 
 been on practical trial at a carbide factory at Campagna, in 
 the French Pyrenees. 
 
 Dorsemagen has devised a process for the electric furnace 
 reduction of siliceous zinc ore, the products being metallic 
 zinc and carborundum (silicon carbide). A mixture of the 
 ore with carbon is heated in an ordinary electric furnace, 
 with carbon electrodes, the result being that silicon carbide 
 is formed, whilst the metallic zinc distils over. 
 
 An analogous process is due to Borchers and Dorsemagen, 
 and is employed in the treatment of compound ores con- 
 taining both iron and zinc, the products being ferro-silicon 
 and zinc. The process is practically identical with the 
 foregoing. 
 
 An Italian process of zinc smelting in the electric furnace, 
 invented by Casaretti and Bertani, is worked by an electro- 
 metallurgical company. One kilogramme of zinc is pro- 
 
 154 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 duced per 2 H.P. hours, and the consumption of coal in 
 the furnace is 15 kgs. per 100 kgs. of ore. 
 
 Chromium. Herr Aschermann has succeeded in pre- 
 paring chromium in the electric furnace. He uses a gas- 
 tight steel structure, containing a graphite crucible, which 
 forms the hearth of the furnace, and into which passes from 
 the outside, a movable electrode. The raw material con- 
 sists of a mixture of oxide of chromium and sulphide of 
 antimony, in the proportion of ten parts to twenty- three. 
 This is placed in the crucible, and the furnace closed, where- 
 upon a current of about twenty amperes is passed through, 
 and suffices to fuse the charge. 
 
 Alloys of chromium and antimony result from the fusion, 
 whilst an amorphous mass of the suphides and oxides of 
 antimony remains on the upper walls of the crucible. The 
 antimony may be driven off in its entirety by reheating, 
 whilst the carbon, an appreciable quantity of which is 
 absorbed and dissolved by the chromium, separates out as 
 graphite on cooling. 
 
 The process was exploited at Cassel, Germany, in 1897-98. 
 
 Molybdenum. Can be prepared in a similar manner from 
 its dioxide (Mo0 2 ) by heating the latter in conjunction 
 with a small percentage of carbon. 
 
 Tungsten. Is prepared by reducing tungstic acid (H 2 WO 4 ), 
 with a small percentage of carbon, in the electric furnace. If 
 the carbon be in excess or the mass too thoroughly fused, a 
 cast metal or carbide is the result. Care must therefore be 
 exercised, in preparing pure metallic tungsten, to regulate 
 the temperature, and allot the proportion of carbon required 
 for the reduction most carefully. 
 
 Nickel. An electric furnace process for the direct pro- 
 duction of nickel from its ores has been exploited at Sault 
 Ste. Marie, Ontario, U.S.A. 
 
 A revolving electric furnace, designed by F. H. Clergue, 
 was employed, in which the ores were smelted by the heat 
 of the arc. A ferro-nickel alloy was said to be obtained in 
 
 155 
 
ELECTRIC FURNACES AND 
 
 one operation, a sample of which, submitted to Messrs. 
 Krupp, of Essen, gave great satisfaction, and is rumoured 
 to have led to the placing of a large order for the alloy. 
 
 Sodium. A. H. Cowles has taken out a patent on an 
 electric furnace process for smelting or reducing sodium 
 aluminate. To this end, the latter is broken up and intim- 
 ately mixed with granular carbon, and the mixture subjected 
 to heat in a closed electric furnace. Metallic sodium is 
 liberated in the form of vapour, and passes over to be 
 collected and condensed in a suitable receiver, whilst 
 the aluminium of the compound combines with the carbon 
 to form aluminium carbide, which can be tapped off from 
 the furnace hearth. 
 
 Alloys of Iron and Steel. In his presidential address to the 
 Society of Chemical Industry, in July, 1901, Mr. J. W. Swan 
 touched upon the growing importance of this industry, not 
 only from the steel manufacturer's point of view, but also 
 as an alternative industry to which many idle carbide 
 plants might, for the nonce, be applied. 
 
 Chromium and chrome iron are being made at Essen by the 
 Goldschmidt process, by the Willson Company in America, 
 and by several of the French metallurgical companies. 
 Ferro-silicon is being produced at Meran, in the Austrian 
 Tyrol, and also at a few of the carbide works in France. 
 
 According to Mr. Swan, the raw ingredients consist of 
 scrap iron, quartz and coke. The daily yield of each 
 furnace is 1,200 kgs., and the product contains 77-5 per cent, 
 iron, and 21-5 per cent, silicon, and costs, at Meran, 8 per 
 ton. The yield, in the electric furnace is one ton per 5,000 
 k.w. hours. 
 
 Ferro-titanium is produced by heating together scrap iron, 
 titaniferous ore, and aluminium, in an electric furnace. 
 Over 30,000 H.P. were available in 1901 for the manufacture 
 of these alloys in the electric furnace. 
 
 In this connexion, it may be mentioned that the Willson 
 Aluminium Company recently installed a 3,000 H.P. plant 
 
 156 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 for the manufacture of ferro-chrome, at Great Kanawha 
 Falls, thirty-six miles above Charleston, W.Va., U.S.A. 
 The chrome ore used in the manufacture is imported from 
 Australia and Asia Minor, and the resultant alloy contains 
 as much as 70 per cent, chromium. 
 
 In 1901, the entire output of the plant was being taken by 
 the Carnegie and Bethlehem Steel Companies, who employed 
 it for hardening armour plates for the U.S. Navy. 
 
 Gin (U Industrie Electrique, April 25, 1901), gives some 
 interesting statistics relating to the thermo-electric manu- 
 facture of ferro-silicon at Meran. The materials and pro- 
 portions used were 
 
 Iron scale (forge) ..... 1,000 parts. 
 
 Quartz 410 
 
 Coke 398 
 
 The respective percentages of useful material were : iron, 
 71-9 per cent. ; silica, 91-3 per cent. ; and carbon, 63-9 per 
 cent, of the raw material. The normal consumption of 
 energy was at the rate of seventy watts per square centimetre 
 of section. The material was cast in lots of 776 kgs., an 
 interval of 15 hours separating two consecutive casts. The 
 efficiency of the process was 80 per cent., and the yield, 
 200 grammes per k.w. hour. The fusion temperature 
 varies from 1,130 to 1,400 C.= 2,066 to 2,552F. 
 
 The resultant product had the following percentage com- 
 position 
 
 Silicon . . . . . .21-45 per cent. 
 
 Iron 77-50 
 
 The working costs at that period were in the neighbourhood 
 of 200 francs per metric ton, or, in English currency, 7 18s, 
 
 In the same article, M. Gin describes a method of manu- 
 facturing ferro-silicon from steel furnace slags. The latter, 
 which, in this particular instance, was obtained from a 
 Martin furnace, had the following composition 
 
 157 
 
ELECTRIC FURNACES AND 
 
 Silica 
 Alumina 
 Ferrous oxide 
 Manganese monoxide 
 Lime, magnesia, etc. 
 
 50-42 per cent. 
 
 2-26 
 34-1 
 9-42 
 3-30 
 
 The above, mixed with coke, in the proportion of 1,680 kgs. 
 of slag to 600 kgs. of coke, containing 80 per cent, carbon, 
 yielded a product which analyzed as follows 
 
 Silicon 
 Iron . 
 Manganese 
 Carbon 
 Miscellaneous 
 
 29-64 per cent. 
 
 53-7 
 
 13-18 
 
 vO 
 
 2-96 
 
 The yield consisted of 4,090 kgs. in 110 hours, the current 
 required being 6,950 amperes at 29-1 volts, or an energy con- 
 sumption of 5,380 k.w. hours per metric ton. 
 
 Ferro-silicon is a far more active agent than carbon in 
 the process of steel manufacture. The combustion of 1 per 
 cent, evolves heat to the extent of 7,830 calories, as compared 
 with 2,473 calories for the combustion of a similar quantity 
 of carbon. 
 
 According to Keller (Paper before Iron and Steel Institute, 
 1903) the degree of purity attainable with ferro-silicon 
 manufactured in the electric furnace is governed by the 
 percentage of silicon present ; in other words, the larger the 
 silicon content, the greater the purity of the alloy. 
 
 He explains this by the fact that it is the iron itself which 
 is responsible for the introduction of impurities, and that, 
 owing to the higher temperatures necessary for the produc- 
 tion of alloys with a large silicon content, the impurities are 
 either volatilized, or eliminated by secondary reactions. 
 
 He has obtained the following analysis for a 50 per cent, 
 ferro-silicon 
 
 Phosphorus . . . . . . 0-02 per cent. 
 
 Sulphur . . . . V . Trace 
 
 Carbon . . . . 
 
 Actual working results have demonstrated that the energy 
 
 158 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 required for the production of one ton of 30 per cent, ferro- 
 silicon in the electric furnace is 3,500 k.w. hours. Messrs. 
 Keller, Leleux, &Co.,in their works at Livet, Isere, France, 
 have an available total of 15,000 H.P. It has been found 
 practicable, with an expenditure of 4,000 H.P., to turn out 
 twenty tons of 30 per cent, ferro-silicon per day, whilst alloys 
 with as high a silicon content as 70 to 80 per cent, have been 
 manufactured. The furnaces employed at the Livet works 
 are of the resistance type, and are of 650 H.P. capacity. 
 
 Ferro-Chrome.The Willson Company, in 1898, were 
 stated to be manufacturing ferro-chrome at the rate of 60 
 tons per month. 
 
 Aluminium Alloys. The Cowles furnace, for the manu- 
 facture of aluminium alloys, consisted of a fire-brick struc- 
 ture, with massive fire-clay lid, in which latter, orifices were 
 left for the escape of the gases. The electrodes entered 
 through two opposite walls at an angle, and consisted of 
 bundles of carbon rods mounted in massive metal caps, the 
 material of which was a constituent of the alloy under 
 preparation, in order to guard against contamination of the 
 product in the event of fusion, which sometimes occurred 
 towards the end of a run. The electrodes entered through 
 tubular conduits, and were provided with screwed rod 
 adjustment. 
 
 The furnace was lined interiorly with broken charcoal, 
 which latter was protected from any tendency to agglom- 
 erate under the intense heat of the furnace, by a previous 
 immersion in lime paste, which coated each particle and 
 separated it from its neighbour. 
 
 In this connexion it is interesting to note that here were 
 the elements necessary for the production of calcium 
 carbide, then a comparatively unknown compound ; and, 
 indeed, it is probable that calcium carbide, in minute 
 quantities, was frequently formed, incidentally, during the 
 action of this early Cowles furnace. 
 
 Alumina, in the form of corundum (crystallized) was 
 
 159 
 
ELECTRIC FURNACES AND 
 
 employed in this process, the initial charge consisting of 
 15 kgs. corundum, 30 kgs. of granulated copper, and suffi- 
 cient carbon to impart the necessary conductivity to the 
 mixture. The current required for the operation was 3,000 
 amperes at 50 volts, which was maintained throughout the 
 run, the electrodes being adjusted according to the variation 
 in electrical resistance of the mass under treatment. Each 
 run lasted from two to three hours, the reaction which 
 took place being represented by the equation 
 
 3 C = 3CO+A1. 
 
 The carbon monoxide gas formed was burnt under a 
 chimney, and the resultant products of combustion carried 
 through a special flue, in which they deposited any metallic 
 particles carried over by them from the furnace. 
 
 The product contained from 15 to 20 per cent, of alumin- 
 ium, and the output of the later trials at Milton were 33 
 grammes per k.w. hour, as against a theoretical yield of 
 152 grammes, or an efficiency of only 22 per cent. 
 
 This aluminium bronze, or aluminium gold, as it is some- 
 times termed, consists of one part aluminium and nine parts 
 copper. 
 
 The introduction of the Heroult and Hall processes for the 
 manufacture of pure aluminium have entirely superseded the 
 Cowles process for the production of aluminium alloys, the 
 pure metal being first obtained, and then alloyed in the 
 desired proportion, thus yielding a much purer and more 
 homogeneous product. 
 
 A later process, patented in 1901 by A. H. Cowles, for 
 the manufacture of aluminium bronze in the electric 
 furnace, though far from efficient, presents several novel 
 points. The furnace consists of a porous carbon crucible, 
 hermetically sealed by a close-fitting lid, through which 
 passes a carbon rod. This latter, and the crucible itself, 
 constitute the two electrodes. Volatile products, e.g. sodium, 
 resulting from the reduction of sodium aluminate in the 
 
 1 60 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 crucible by carbon, pass through the porous walls of the 
 latter, under pressure of the gases produced in the reaction 
 and are condensed on the inner walls of a cooling tank, or 
 water jacket, which surrounds the crucible, and with it 
 forms an annular space, in which the volatile product is 
 collected. The aluminium carbide remains behind in the 
 crucible, or, if copper, or other metal be employed, an alu- 
 minium alloy is the result. 
 
 In order to prevent clogging of the pores in the walls of 
 the crucible a very high temperature is necessary, a fact 
 which would presumably detract from the efficiency of the 
 process. 
 
 The Comminution, or Pulverization of Metals, with or 
 without the subsequent production of their oxides by com- 
 bustion in oxygen, is a branch of metallurgy to which the 
 electric furnace, duly modified, is admirably adapted. 
 
 The general method of procedure consists in heating the 
 metal, either by means of an arc or resistance furnace, to 
 the temperature of volatilization ; the resulting vapour 
 burns in air or oxygen, and the oxide of the metal is pro- 
 duced. In the event of comminution of the metal, pure and 
 simple, being required, the operation is carried out in a 
 neutral atmosphere. 
 
 One of the earliest applications of the electric furnace to 
 these requirements was made in 1896 by the Societe Civile 
 d'Etudes du Syndicat de 1'Acier Gerard, of Paris, who took 
 out a patent on a process for the production of steel from 
 pig iron. In the main, the process consisted in pulverizing 
 the metal, and afterwards subjecting it to an air blast, which 
 resulted in the oxidation of silicon, sulphur, phosphorus, 
 and carbon, and in due course the conversion of the iron 
 into steel. 
 
 To effect the necessary pulverization, an arc was suggested, 
 the primarily fused metal being poured, in a falling stream, 
 between the two carbons and consequently through the heat 
 zone of the arc, after which it passed through an air blast, 
 
 161 M 
 
ELECTRIC FURNACES AND 
 
 and was subsequently collected in the hearth of the fur- 
 nace. 
 
 In a modified method, the metal was heated in a resistance 
 furnace, having water-cooled terminals, and oxidation of 
 the vapour effected by a blast of air projected against the 
 surface of the molten mass at the point of maximum 
 temperature. 
 
 The process, patented in 1902 by C. S. Lomax, for the 
 manufacture of the oxides of lead and tin, is almost identical 
 with the above, the general principle of furnace and operation 
 being illustrated in Fig. 37, where H is a shallow 
 hearth with enlarged or expanded end cavities, the metal 
 in which remains cooler by virtue of their greater cross- 
 section and consequent carry- 
 ing capacity, and in which ter- 
 minal electrodes are inserted 
 for connexion to the source of 
 current. P is a supply pipe, 
 for the delivery of air under 
 pressure through a series of 
 branch nozzles n at or near 
 the surface of the molten metal 
 in H. The oxides, thus formed, 
 
 pass over, under direction of the controlling air blast and 
 cover C, into a receiving] chamber R, where they are 
 collected. 
 
 It is claimed for this furnace that pure dioxide of tin may 
 be produced by the action of an air blast, the temperature 
 of which has been primarily raised to 400C. 752F., the 
 molten mass being at 1,200C.=2,192F., whilst, under 
 similar conditions, but with a cold air-blast, putty powder, 
 a mixture of the dioxide and monoxide of tin, is formed. 
 In the manufacture of lead oxides a cold air blast is employed. 
 M. Paul Bary has devoted considerable attention to this 
 subject, and his more recent apparatus is based, in principle, 
 upon the volatilization of a stream of molten metal carrying 
 
 162 
 
 FIG. 37. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 a current by the intermittent arc which follows a rapid and 
 continuous series of interruptions. 
 
 It is a well known fact that if a current of certain density 
 be passed through a metallic film, the continuity of the latter 
 is momentarily destroyed, and immediately restored again 
 at a rate amounting to some hundreds of interruptions per 
 second ; it is upon this phenomenon that M. Bary has based 
 the construction of his latest apparatus, which presents several 
 points of interest. 
 
 A sectional elevation of the furnace is represented in 
 Fig. 38 ; it comprises 
 an upper chamber A 
 divided from a lower 
 receiving chamber B 
 by a horizontal par- 
 tition p. The con- 
 struction is circular, 
 and, through equi- 
 distant points around 
 the circle formed by 
 the partition p, pass, 
 vertically, a series 
 of small tubes n n, 
 extending upwards 
 
 above the surface of the molten metal in A, but provided 
 with side perforations at a a to permit the passage of the 
 metal into the lower chamber B. At their lower extremities 
 s s, which extend into B, the tubes n converge into fine 
 nozzles, which serve to concentrate the falling streams upon 
 a series of metal buttons m m placed immediately 
 beneath them. An insulating partition D is introduced 
 just above the line of nozzles, which electrically isolates the 
 upper portion A from the lower B, the only connexion 
 between the two, being through the tubes n, nozzles s, 
 buttons m, and the passing streams of molten metal, which, 
 as already indicated, automatically break up and reform 
 
 163 
 
 FIG. 38. 
 
ELECTRIC FURNACES AND 
 
 themselves several hundred times per second under the action 
 of the current which they carry. R R are resistances located 
 between the upper and lower chambers, and included in the 
 main circuit. These, heated by the current, serve to main- 
 tain the contents of both chambers in a molten state. The 
 metallic buttons m, and the conductors leading to them, 
 are embedded in a refractory base E, which slopes towards 
 the outlet 0. The oxidized products pass over through 
 this outlet into a suitable receiving chamber, any excess of 
 unconverted metal being caught, in passage, by a trap T. 
 
 F is an inlet for the introduction of the necessary oxidiz- 
 ing gas, and serves also as an electrode connexion to the 
 external circuit, the remaining connexion being made to the 
 upper chamber A. 
 
 The Bary process for the manufacture of stannic acid 
 comprises a resistance furnace in which the tin is vaporized, 
 and the resultant vapour, passing through an orifice in the 
 upper part of the structure, is ignited by a blast of air, 
 forming, by combustion, dioxide of tin. The patent is con- 
 trolled and applied by the Tin Electro-Smelting Co., Ltd., of 
 Paris. 
 
 The electric furnace process for the manufacture of white 
 lead, patented by Messrs. Bailey, Cox, & Hey, consists in 
 striking an arc between a tubular electrode, and the surface 
 of the molten lead to be converted. The tubular electrode 
 serves also as a conduit, through which is introduced a 
 mixture of steam, carbon dioxide gas, and acetic acid 
 vapour, the reaction between which, and the vaporized 
 lead at the central heat zone, results in the formation of 
 commerical white lead. 
 
 By a new method, pigments of metallic oxides are pro- 
 duced by burning, in special flues, the waste vapours from 
 electric reduction furnaces. The varying mixtures, from 
 different ores, give a great variety of colours, waste is avoided, 
 and the products are in extremely fine powder without 
 grinding. 
 
 164 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION VI 
 
 PHOSPHORUS MANUFACTURE IN THE ELECTRIC FURNACE 
 
 Speaking generally, the main difficulty experienced in the 
 manufacture of phosphorus, per se, in the electric furnace, 
 is the necessarily high temperature at which the vapour 
 leaves the furnace chamber. This vapour was originally 
 condensed immediately after leaving the furnace by means 
 of a water circulatory system, but improvements have since 
 been effected by the Compagnie Electrique du Phosphor, 
 Billaudot & Cie. 
 
 In the method adopted by this firm, the vapour is led 
 through a system of cooling towers, whereby the condensation 
 is carried on, by air cooling, at a slower rate, and it is said 
 to yield not only a better quality of phosphorus, but also a 
 larger quantity, and at a reduced cost. 
 
 In 1899, it was stated (Tatlock) that one-half the world's 
 production of phosphorus, approximately 1,000 tons, was 
 manufactured by electric furnace methods. 
 
 A process for the manufacture of phosphorus in the electric 
 furnace was invented independently by Dr. Headman and 
 Mr. Thos. Parker, in 1888. 
 
 The two inventors subsequently combined interests, and 
 the Parker furnace, illustrated in Fig. 39, was the outcome. 
 
 It consists of a fire-brick structure F, with a domed roof, 
 through an aperture in the centre of which the raw material 
 is fed from the hopper H by the mechanically driven screw 
 conveyor S. The disposition of all parts is such that air is 
 
 165 
 
ELECTRIC FURNACES AND 
 
 hermetically excluded from the interior of the furnace. The 
 lower portion of the structure is contracted to form a hearth 
 A, as shown, and, through the side walls at this point, are 
 
 introduced the terminal 
 carbon electrodes E E 
 in parallel sets, which serve 
 to convey the current to 
 the charge, the furnace 
 operating on the resist- 
 ance principle. This linear 
 distribution of electrodes 
 results in distributing the 
 heat area, and rendering 
 the effect more uniform. 
 
 The two rows of smaller 
 electrodes e e, placed be- 
 * low the main electrodes, 
 suffice to start the heat- 
 ing of the charge, and 
 bring it up to a certain 
 point, after which it is 
 
 maintained by the current passing between E E. D is 
 a flue for conducting away the gases and vapours formed 
 during the process. 
 
 The Readman and Parker process was perfected commer- 
 cially in 1888. It was exploited on a large scale at the works 
 of the Electric Construction Company, Wolverhampton, 
 and has since been adopted by Messrs. Albright and Wilson, 
 at Oldbury. The mineral phosphate is first pulverized, then 
 mixed with carbon and sand, and finally heated in the 
 closed electric furnace described above, whereupon the 
 phosphorus distils over and is collected under water. 
 
 The capacity of the furnace is 80 kgs. of phosphorus per 
 furnace per day. The process is fairly economical, in that 
 from 80 to 90 per cent, of the phosphorus contained in the 
 charge is recovered. In 1898, the Oldbury works utilized 
 
 166 
 
 FIG. 39. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 some 700 H.P., and the Niagara works 300 H.P., in the 
 manufacture of this element, but the plant and output have 
 since been considerably extended. Phosphorus, to the 
 amount of fifteen tons per month, is said to be produced at 
 Niagara Falls by the Oldbury Chemical Company. 
 
 Similar plants are in existence at Geneva (Switzerland), 
 and Greisheim (Germany), in which latter country the 
 refusal of patent rights has left the electric furnace method 
 open to all comers. 
 
 The resistance furnace, invented by H. A. Irvine for the 
 electrical manufacture of phosphorus, is closely allied to the 
 original De Laval furnace, described in the section on 
 smelting, in that the resistance column is in a state of fusion. 
 The furnace consists of a refractory chamber, with a domed 
 roof, in the centre of which is fitted a feed hopper, and 
 through which, on either side of the hopper, descend ver- 
 tical carbon electrodes, their lower extremities reaching to 
 within a short distance of the floor or hearth of the furnace, 
 which latter, together with the side walls, is composed of 
 packed carbon. 
 
 Tap holes are provided, whereby the level of the molten 
 mass on the furnace hearth, and consequently the cross- 
 section of the conducting column between the electrodes, 
 is maintained constant. Two lower horizontal carbon 
 electrodes, projecting through the furnace walls, on a level 
 with the hearth, also provide a ready means of changing 
 the direction of the current path, if such be necessary. 
 
 The action is first started through a layer of granular 
 coke, so placed as to bridge the lower extremities of the 
 electrodes, and is subsequently maintained through a mass 
 of conducting slag, formed from the fusion of the raw 
 materials. The charge consists of a mixture of phosphate 
 rock, carbon, and the necessary proportion of a suitable 
 flux, and, in its molten state, covers the furnace hearth to 
 a constant level, determined by the aforementioned tap 
 holes. The unfused charge rests on the top of this molten 
 
 167 
 
ELECTRIC FURNACES AND 
 
 mass, and, as it melts, flows down to take the place of that 
 already fused and drawn off. 
 
 The principal difficulty experienced in the application 
 of furnaces of this type to industrial operations on a com- 
 mercial scale is that of finding a sufficiently refractory and 
 non-conducting lining, which shall, at the same time, be 
 proof against oxidation by the furnace contents. 
 
 The Machalske electric furnace process for the manu- 
 facture of phosphorus, invented by Dr. F. J. Machalske, of 
 Long Island City, New York, and exploited by the Anglo- 
 American Chemical Company, is illustrated in Fig. 40. 
 The furnace consists of a central chamber, or crucible C, 
 36 by 12 by 18 in., built up of carbon blocks, and lined 
 interiorly with calcined magnesia and a special mixture, 
 whilst outside it is jacketed with fire-clay, red brick, 
 and a mixture of borax and asbestos flour. The upper 
 electrode E is mounted in a special holder or clamp A, 
 and provided with a hand wheel and worm W for feed 
 adjustment. It is 4 in. in diameter, and 8 ft. long. The 
 lower electrode E2 passes vertically up through the floor 
 of the furnace, and is also provided with a screw and hand 
 wheel adjustment, W2, as shown. 
 
 The raw material, or phosphate, is placed in the hopper 
 H, and fed into the furnace chamber by the screw conveyor 
 S. Once the action is started the arc can be drawn out to 
 as great a length as 15 in. 
 
 An alternating current is employed, with a voltage 
 ranging from 30 to 120. Each furnace takes from 1,000 
 to 4,000 amperes, and a temperature of 3,867C.= 7,000F. 
 is available within five minutes of switching on. 
 
 A molten slag, consisting mainly of calcium silicate, is 
 produced, and run off at the tap hole T, whilst the phos- 
 phorus is driven off as vapour, and passes, by way of the 
 outlet O, into suitable condensers, where it is solidified in 
 the form of dark yellow shavings. These are subsequently 
 treated with sodium hypobromide, whereby the red phos- 
 
 168 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 phorus is converted into yellow, and impurities elimin- 
 ated. 
 
 The chemical equation representing the reaction which 
 takes place, is as follows : 
 
 Ca 3 (P0 4 ) 2 + 3Si0 2 + 50 =3CaSi0 3 + 5CO + 2P. 
 
 The calcium monosilicate, as already stated, accumulates 
 on the furnace hearth, 
 whence it is run off in 
 a syrupy condition. 
 
 In the Machalske 
 furnace, as described 
 above, 68 kgs. of the 
 raw phosphate can be 
 treated in a quarter of 
 an hour, yielding yellow 
 phosphorus at 3%d. per 
 pound, reckoned on a 
 basis of 2d. per B.O.T. 
 Unit. 
 
 Two of these furnaces, 
 each consuming 2,000 
 amperes, are in use by 
 the Anglo - American 
 Chemical Company in 
 the manufacture of yel- 
 low and red phosphorus. 
 
 Dr. Machalske states, that in the course of his experiments 
 with the above process of phosphorus manufacture, he 
 discovered that, by means of an electric furnace, connected 
 to special condensing apparatus, chlorides of carbon, more 
 especially carbon tetrachloride, , could be produced by 
 treating a mixture of common salt, carbon, and sand. 
 
 169 
 
ELECTRIC FURNACES AND 
 
 SECTION VII 
 
 GLASS MANUFACTURE IN THE ELECTRIC FURNACE 
 
 One of the more recent applications of the electric furnace 
 is in the manufacture of glass, a process which entails con- 
 siderable expenditure of heat, and a comparatively clean 
 source for the latter, such that no impurities, in the shape 
 of combustion products, etc., shall enter into the fused 
 mass, and destroy the purity and transparency of the 
 finished article. 
 
 Nernst's discoveries in connexion with such earths as 
 become conductors when heated to a certain degree, have 
 an important bearing on the development of this industry, 
 in that glass itself may be numbered among those sub- 
 stances ; molten glass is, in point of fact, an electrolyte, and 
 thus lends itself readily to electric furnace methods of 
 manufacture. 
 
 One of the earliest electric furnaces for glass production 
 was patented in Germany, in 1882, by Messrs. S. Reich & 
 Co. It was of the resistance type, and consisted essentially 
 of a carbon crucible, open at the base, and lined internally 
 with a net or bag of platinum wire. 
 
 The raw material was fed into this, and, having been 
 fused, by the heat developed in the carbon walls, dripped 
 through into refining vessels placed underneath. 
 
 The arc furnace, invented by Albert A. Shade, is specially 
 adapted to the fusion of silica or its compounds, as in glass 
 manufacture, but is also applicable to other electric furnace 
 processes. The hearth is a long inclined trough, of circular 
 
 170 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 section, lined with refractory material, and provided, at 
 its upper extremity, with a screw conveyor or other device, 
 for securing a continuous feed, and at its lower end with a 
 discharge orifice and pouring lip. 
 
 The diameter of the lower half of the trough-shaped 
 hearth is less than that of the upper, being proportional 
 to the desired rate of passage of the material under treatment, 
 and the arc electrodes, arranged in diametrically opposed 
 pairs, enter through the side walls at a point just above 
 the reduced diameter. They are disposed in the lower 
 half of the inclined hearth or chute, the last pair of elec- 
 trodes being very near the discharge opening. In line with 
 the electrodes, but at a higher elevation, are a series of 
 pipes, opening into the furnace chamber, and suitably 
 protected from the heat. These convey exhaust furnace 
 gases into the interior, where they are burnt, and serve to 
 dry and pre-heat the charge during its descent, thus pre- 
 paring it for the more advanced heating effect of the arcs. 
 
 Vertically disposed within, and entirely surrounded and 
 protected by the masonry of the furnace structure, are a 
 series of electro-magnets, one to each pair of electrodes, 
 with its active poles immediately below the arcing space. 
 Sandwiched between each neighbouring pair of magnets 
 is an iron shield plate, also embedded in the masonry, and 
 serving to confine the effect of each particular magnet to 
 its own arc. The general tendency of the magnets, when 
 active, is to deflect the arcs downward on to the gradually 
 moving charge, thereby securing an enhanced heating 
 effect. 
 
 A flue leads from the top of the furnace chamber to a 
 jacket surrounding the conveyor cylinder, which is thus, 
 with its contents, also subjected to a preliminary drying 
 process, the jacketing being even extended to the hopper 
 itself. 
 
 Although of somewhat complex construction, this furnace 
 would certainly appear to be designed with a view to taking 
 
 171 
 
ELECTRIC FURNACES AND 
 
 the fullest advantage of the otherwise waste heat of the 
 furnace gases, and its general disposition, in addition to 
 being well thought out, indicates that the inventor is fully 
 alive to the economies attendant on an efficient system of 
 pre-heating the raw materials, and reducing the duties of 
 the electrical portion of the apparatus to a minimum. 
 
 The Henrivaux furnace for the electrical manufacture 
 of glass comprises three steps or terraces, composed of a 
 refractory material. Immediately over the surface of 
 each step is disposed a powerful arc, and the raw material, 
 fed from a hopper into the heat zone of the topmost arc, 
 is fused, and flows down, passing in turn through each of 
 the two remaining arcs, and finally reaching a receiver, 
 or trough, into which it drips. 
 
 The process is far from efficient in that a considerable 
 quantity of the heat generated is dissipated without doing 
 any useful work. 
 
 Modern gas-fired glass furnaces call for a consumption 
 of 1-5 to 2-8 kgs. of coal for every kilogramme of glass 
 produced, whereas the current consumption, in the Henri- 
 vaux furnace, for a similar result, is equivalent to 9-3 kgs. 
 of coal. A considerable improvement in the efficiency 
 of the process is, however, predicted. 
 
 In F. A. Becker's electric furnace method of glass manu- 
 facture, the raw materials are passed between three pairs of 
 carbon electrodes. The first two pairs serve to fuse the 
 materials, with an expenditure of 100 amperes at 40 volts. 
 The third pair of carbons is located below the other two 
 and maintains the fusion, with an energy expenditure of 
 50 amperes at 40 volts. Alternating currents are used, and 
 the process is continuous. 
 
 Among the most successful of electric glass furnace 
 methods, may be mentioned the Voelker process, which 
 has been exploited on a commercial scale at Plettenberg, 
 Westphalia. 
 
 The furnace invented by August Voelker is represented 
 
 172 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 in Fig. 41, and is a combination of the arc and resistance 
 types, the upper portion A being utilized as an arc furnace 
 for melting the raw 
 materials ; the inter- 
 mediate B as a resist- 
 ance furnace for a species 
 of refining process which 
 the molten mass subse- 
 quently undergoes be- 
 fore it finally overflows 
 into the lower receptacle 
 or trough C. 
 
 The arc furnace A is 
 constructed, as usual, of 
 refractory material, and 
 has a dome d, which 
 reflects the heat of the 
 arc set up between the 
 two electrodes e e on 
 
 to the raw material, 
 
 which is fed by screw FlG - 41 - 
 
 conveyors s s from the 
 
 hoppers h k, and delivered, as shown, just beneath the 
 
 electrodes. 
 
 The mass, having been fused by the reflected heat of the 
 arc, flows down the central opening or chimney c into the 
 intermediate resistance furnace B, which is subdivided 
 by the vertical perforated partitions p p into one central, 
 and two outer, chambers. In these latter are placed the 
 two electrodes / / of the resistance furnace, the circuit 
 between them being completed by the molten glass. The 
 object of the partitions is to prevent contamination of the 
 fluid mass by particles which might become detached 
 from / /. 
 
 In this intermediate chamber the glass becomes more 
 fluid, and the bubbles of air and gas carried over by it from 
 
 173 
 
ELECTRIC FURNACES AND 
 
 the first fusion in A are driven off. It then overflows 
 into the cavity C, whence it is drawn off as pure glass, and 
 is ready for use as such. 
 
 In a modified form of the Voelker glass furnace construc- 
 tion the arcs for the preliminary fusion of the raw materials 
 are disposed in radial channels, converging to the inter- 
 mediate refining chamber, their heated gases being burnt 
 below the final reservoir, where they assist in maintaining 
 the fluidity of the molten product. 
 
 The Bronn process for the manufacture of glass in the 
 electric furnace has been devised with a view to overcoming 
 several of the drawbacks incidental to a continuous process, 
 chief among which may be mentioned the splitting up of 
 the charge, whilst passing through the furnace, into unequal 
 masses, and a consequent lack of homogeneity in the manu- 
 factured product ; contamination of the molten glass by 
 particles detached from the electrodes, etc. 
 
 In the Bronn furnace, invented by J. Bronn, of Cologne, 
 these drawbacks are avoided. Its general principles of 
 construction and operation are represented in Fig. 42, 
 where H is a hopper, in which the raw material in the form 
 of a powder is mixed with a suitable binding substance, 
 such as water-glass, hydraulic lime, or plaster, which will 
 not affect the transparency or purity of the resultant 
 glass. 
 
 Having undergone the process of mixing, it is fed down 
 the chute C to the rollers R R, between which it passes, 
 and is thereby transformed into a continuous and homo- 
 geneous sheet, or rod, as the case may be, the particles 
 being held together by the water-glass or lime before men- 
 tioned. 
 
 It next passes over a heated roll r, which drives off all 
 moisture, and finally emerges on to the upper extremity of 
 an inclined plane p, forming the hearth or floor of the 
 furnace. Down this it travels at a regular rate, dependent 
 upon the speed of fusion, and consequent glass formation, 
 
 174 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 passing, for that purpose, under the arcs playing between 
 the electrodes E' E, or, if in rod form, travelling down the 
 hearth in like manner between opposite pairs of arc elec- 
 trodes. 
 
 A comparatively recent patent granted to J. C. T. Kess- 
 
 FIG. 42. 
 
 meier relates to the application of an electric furnace method 
 as an auxiliary in the casting of glass articles. The extreme 
 fluidity required for the successful casting of molten glass 
 is brought about by arranging a pair of carbon electrodes, 
 through which a current is caused to pass, in the walls 
 of the outlet channel from the ordinary melting crucible 
 to the moulds. 
 
 Having first been reduced to a molten condition in the 
 crucible by external heat in the ordinary manner, the glass 
 is then run off through this channel, and, in passing between 
 the two electrodes, completes the circuit and, by virtue 
 of its conducting properties, carries sufficient current to 
 raise it to a temperature of 1,924C.=3,500F., or even 
 higher, at which stage it becomes extremely fluid. 
 
 No mention is made of electrolytic effects produced upon 
 the glass by the current in passing, though it seems prob- 
 able that such would occur with a uni-directional current. 
 
 175 
 
ELECTRIC FURNACES AND 
 
 SECTION VIII 
 
 ELECTROLYTIC FURNACES AND PROCESSES 
 
 Aluminium. Monkton, as early as 1862, was granted 
 an English patent covering the reduction of alumina by 
 carbon with the aid of electric heat. 
 
 Moissan has, however, shown, in the course of his re- 
 searches, that such a process is commercially impracticable, 
 in that alumina, even when in a liquid or molten state, is 
 not reducible by carbon, it being necessary to vaporize 
 both substances, and, moreover, heat their mixed vapours 
 to a very high degree before reduction is effected ; even 
 then the product is partly aluminium carbide. 
 
 Aluminium was first prepared by Wohler in 1828, by 
 heating aluminium chloride in the presence of sodium. The 
 only commercial process for its extraction, prior to the 
 introduction of the thermo-electrolytic method, was that 
 of heating the double chloride of aluminium and sodium 
 (2NaCl, A1 2 C1 6 ), or the native double fluoride (cryolite), 
 with sodium. 
 
 One of the earliest instances of the manufacture of alumi- 
 nium by electrolytic methods was that of Lacassagne and 
 Thier's primary battery, in which the metal was produced 
 as a by-product of the battery reaction. 
 
 The latter consisted of an outer crucible, heated by an 
 ordinary furnace and containing chloride of sodium, in 
 which was immersed an iron electrode. In this was placed 
 a porous pot containing a carbon electrode immersed in 
 chloride of aluminium. The furnace served to bring both 
 salts to a condition of fusion, when a current was available 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 at the electrode terminals, and, as before stated, aluminium 
 was produced as a by-product. 
 
 The price of aluminium in 1855 was approximately 
 56 per kg., as against its present market price of 2s. Id. 
 per kg., a reduction which has been mainly brought about 
 by the inception of the electrolytic processes for its ex- 
 traction from alumina. 
 
 Aluminium is one of the most abundant metals in nature, 
 constituting from ten to twenty per cent, of all varieties 
 of clay, in the form of silicates. No method has, up to the 
 present, been forthcoming for the economical reduction of 
 aluminium silicates, and the metallurgist has consequently 
 to fall back upon bauxite, a hydrated alumina, which is 
 found in France, Ireland, and the United States ; and 
 cryolite, the native double fluoride of aluminium and sodium, 
 which has the formula, 6NaP.Al 2 F 6 . The principal 
 source is Arksut Fiord, West Greenland. These two sub- 
 stances are at present the most suitable raw materials for 
 the production or separation of metallic aluminium. 
 
 Modern progress in aluminium manufacture has indicated 
 the desirability of dispensing with a carbon cathode, and 
 employing the molten metal itself in its stead. The yield 
 is about 1 kilogramme per 14 k.w. hours, the power being 
 only a small item in the total cost of manufacture. 
 
 The principal expense is entailed in the preparation of 
 the pure alumina from crude bauxite, a process which 
 accounts for fully 40 per cent, of the total cost of manu- 
 facture. Crude bauxite contains iron, silicon, titanium, 
 and several other undesirable impurities. 
 
 The heat of formation of A1 2 3 is 392,600 Calories; 
 therefore, according to Thomson's rule, a minimum E.M.F. 
 of 2-8 volts is required at the terminals of the aluminium 
 furnace in order to bring about the separation of the metal 
 from its compound, alumina. This voltage, multiplied 
 by the number of coulombs passed through the furnace, 
 represents the electrical energy which is, during the reaction, 
 
 177 N 
 
ELECTRIC FURNACES AND 
 
 converted into chemical energy, and stored, as such, in the 
 metallic aluminium. It may, when desired, be converted 
 into heat energy by causing the aluminium to again enter 
 into combination with oxygen, as, for example, in the 
 " Thermite " process. 
 
 The electrical reduction of aluminium from alumina was 
 first effected at Cleveland, Ohio, U.S.A., the site of the 
 Cowles Electric Smelting and Aluminium Co. The Cowles 
 syndicate was organized in 1886, at Stoke-on-Trent, where 
 electric smelting works were erected. It was re-organized 
 in 1894 as the British Aluminium Company, with works at 
 the Falls of Foyers, in Scotland. 
 
 Electrolytic processes for the manufacture of aluminium 
 were patented during the period 1886-1887, by Hall in 
 America, and Heroult in England and France. It was 
 worked on a commercial scale at Pittsburg by Hall as 
 early as 1888. 
 
 According to Roberts-Austen, the cost of purified alumina, 
 as manufactured from bauxite, constitutes 45 per cent, 
 of the total cost of aluminium manufacture. Crude 
 bauxite often contains, in addition to alumina, oxides of 
 iron, silicon, and titanium. 
 
 One of the disadvantages of the Bayer process (described 
 further on) for extracting alumina from bauxite, lies in 
 the presence of sodium carbonate, one of the auxiliary 
 compounds necessary to bring about the initial fusion. 
 Once mixed with this compound, it is almost impossible 
 to entirely eradicate it from the otherwise purified alumina. 
 
 Hall's process for the extraction of pure alumina from 
 bauxite, patented in 1901, has led to the production of a 
 grade of alumina, very different, both in appearance and 
 characteristics, to that resulting from Bayer's method. 
 
 In the specifications relating to Hall's patents for the 
 purification of bauxite, two processes are dealt with. The 
 bauxite to be treated is represented as having the following 
 composition 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Alumina 
 Ferric oxide 
 Silica 
 
 Titanic oxide 
 Water 
 
 60 per cent. 
 
 18 
 2-3 
 3-4 
 
 17 
 
 The first method consists, briefly, in calcining the bauxite, 
 after which, from eight to ten per cent, of carbon is added, 
 and the mixture fused in a carbon-lined electric furnace, 
 with carbon electrodes. The fusion is maintained for about 
 an hour, the current being either alternating or continuous. 
 The outcome of thus heating the molten mass in the pre- 
 sence of carbon, is that the impurities are reduced, and sink 
 to the hearth of the furnace in the form of an alloy. 
 
 If the percentage of iron present as an impurity be low, 
 or the other impurities, silica and titanic oxide, corre- 
 spondingly high, it is an advantage to add more iron to the 
 fused mass, either as oxide, or in the metallic state, to aid 
 the separation. 
 
 The second method is practically identical, in principle, 
 with the first, the only difference being the substitution 
 of metallic aluminium for carbon as the reducing agent. 
 As before, a preliminary analysis of the bauxite is made, 
 and sufficient aluminium added to combine with the whole 
 of the oxygen known to be present as a constituent of the 
 various impurities. The mass is then subjected to fusion 
 in the electric furnace for about one and a half hours, during 
 which the latter separate out, as in the previous case, but 
 without evolution of gas. The same procedure, in the 
 event of a low iron content, should be adopted here, as 
 in the first case, and the iron may conveniently be associated 
 with the added aluminium in the form of a powdered alloy 
 containing about 50 per cent, of each metal. 
 
 The theoretical output of aluminium at maximum 
 efficiency is 2-85 kgs. per 1 k.w. per 24 hours. In actual 
 practice, however, the efficiency of the process is very low, 
 and '6 kg. of aluminium per k.w. day is probably nearer 
 the mark. 
 
 179 
 
ELECTRIC FURNACES AND 
 
 About 2 kgs. of anhydrous alumina are required for 
 the manufacture of 1 kg. of metallic aluminium, from which 
 it will be seen that the cost of the raw material is no in- 
 considerable item in the total cost of manufacture. The 
 consumption of carbon electrodes is also a source of expense, 
 the theoretical figure, 66 per cent, of the weight of alu- 
 minium manufactured, usually increasing, in practice, to 
 100 per cent., or weight for weight with the metallic product. 
 
 The following are two estimates of the cost of aluminium 
 manufacture, the latter being based on the Hall process 
 
 PER KILOGRAMME 
 OF ALUMINIUM. 
 
 Power . . , . , * . . ' 4-8cZ. 
 
 Alumina . . . ..... . 8-8d. 
 
 Carbon electrodes . . . . T 4-4d. 
 
 Labour, superintendence, interest on, and repairs 
 
 to furnaces ...... 4-4d. 
 
 Total . . .. . . 22-4d. 
 
 Power . . . .... . 4-6d. 
 
 Alumina . . , . . . . 13-9rf. 
 
 Carbon electrodes . . . . . . 3-5d. 
 
 Miscellaneous 1 3d. 
 
 Total 23-3d. 
 
 There is one important point in connexion with the 
 electrolytic manufacture of aluminium, and that is the 
 necessity for maintaining that difference between the 
 specific gravities of the liberated metal, and the fused 
 electrolyte, which will ensure their occupying correct rela- 
 tive positions in the furnace, or, in other words, to keep 
 the aluminium at the bottom. 
 
 The following table, published by Richards, shows the 
 specific gravities of the various substances employed in 
 aluminium manufacture, in both a fused and solid state, 
 and it will be seen that the margin of weight, causing the 
 liberated aluminium to remain at the bottom, is not great. 
 
 180 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SPECIFIC GRAVITIES. 
 FUSED. SOLID. 
 
 Commercial aluminium . . . 2-54 . . 2-66 
 
 Commercial Greenland cryolite . . 2-08 . . 2-92 
 
 Cryolite saturated with alumina . . 2-35 . . 2-90 
 
 Cryolite mixed with aluminium fluoride 
 in the proportion indicated by the 
 
 formula Al 2 F 6 -2NaF . . . 1-&7 . . 2-96 
 The above saturated with alumina 2-14 2-98 
 
 In the Zeitschrift fur Elektrochemie, January 2 and 9, 
 1902, were published some details of laboratory experiments, 
 carried out by Messrs. Haber & Geipert in regard to the 
 electrolytic production of aluminium. 
 
 The object of the experiments was to determine the actual 
 E.M.F., current density, etc., necessary to bring about the 
 reaction. They were carried out in the laboratory of the 
 Technical College at Karlsruhe, the apparatus employed 
 being a small carbon crucible, with a central carbon anode 
 arranged axially therein. In this furnace the authors suc- 
 ceeded in obtaining metallic aluminium from an electrolyte, 
 consisting of a mixture of two-thirds aluminium and sodium 
 fluorides, and one-third alumina, the current employed 
 being from 300 to 400 amperes at 7 to 10 volts, representing 
 a current density of 3 amperes per square centimetre. 
 
 The central anode was gradually raised as the operation 
 proceeded, whilst the high percentage of aluminium fluoride 
 present rendered the bath very fluid. 
 
 The resultant metal contained only 0-05 per cent, carbon 
 and 0-034 per cent, silicon, and had a tensile strength of 
 from 13-7 to 17-3 kgs. per square millimetre. 
 
 In 1903, there were, according to Kershaw (Electrical 
 Review, March 20, 1903), nine factories engaged in the pro- 
 duction of aluminium, either by the Hall or Heroult process ; 
 they are located as follows : America, 3 ; United Kingdom, 
 1 ; France, 2 ; Germany, 1 ; Switzerland, 1 ; Austria, 1. 
 
 The aggregate power available at these several factories 
 is between 36,000 and 40,000 H.P., all derived from water. 
 
 181 
 
ELECTRIC FURNACES AND 
 
 A reliable estimate places the output for 1901 and 1902 
 of the : 
 
 1901 1902 
 
 Six European works, at . . 4,000 tons . . 3,800 tons 
 Three American works, at . 3,240 . . 4,200 
 
 Totals . . 7,240 . . 8,000 
 
 In November, 1902, the prices of the Pittsburg Reduction 
 Company were as follows 
 
 No. 1 Metal, guaranteed over 99 per cent. 16^d. to 18^d. per Ib. 
 
 2 90 15id. lid. 
 Nickel-aluminium alloy (less than 10 per 
 
 cent, nickel) .... 16|d. 19|d. 
 
 Powdered aluminium . . , 45(7. ,, 50d. ,, ,, 
 
 Aluminium castings . . . 22|d. ,, ,, 
 
 All the above prices were subject to discounts ranging 
 from 10 to 15 per cent., whilst European competition is 
 stifled by an import duty of 4d. per Ib. on ingot metal, and 
 . per Ib. on sheet. 
 
 American prices for aluminium rod and wire, ranged from 
 . to 26d. per Ib., according to gauge, a rebate of l\d. to 2d. 
 being allowed off list price, according to the magnitude of 
 the order. 
 
 The (1903) prices of the British Aluminium Company were 
 as follows 
 
 Ingot metal (guaranteed over 
 
 99 per cent, aluminium) . Ifi^d. per Ib. less 7^ per cent. 
 Ingot metal (98 to 99 per cent.)15rf. 7| 
 (No. 4 alloy) 16d. 2| 
 (No. 6 alloy) . 13fd. 
 Wolframinium alloy . . 17M 
 
 Aluminium wire, Nos. 1-14 
 S.W.G. . ... , 
 
 The American output of aluminium is controlled by the 
 Pittsburg Reduction Company, with two factories at Niagara 
 Falls, and one at Shawinigan Falls. 
 
 The European manufacturers are the Aluminium Industrie 
 Aktien-Gesellschaft, with works at Neuhausen, Rheinfelden, 
 and Lend Gastien ; the Societe Electrometallurgique Fran- 
 
 182 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 9aise, with works at Froges and La Praz ; and the Compagnie 
 des Produits Chimiques d'Alais, with works at St. Michel, 
 France. The only British concern engaged in the industry 
 is the British Aluminium Company, with works at Foyers, 
 N.B. 
 
 The United States, represented by the Pittsburg Reduction 
 Company, who control the whole of the patents therein, is, 
 of course, the principal producer. It has in operation at 
 Niagara Falls, plant utilizing some 11,000 H.P., and at 
 Shawinigan Falls, Quebec, operating under the title of the 
 Royal Aluminium Company, 5,000 H.P., the two works being 
 capable of manufacturing 4,500 tons of aluminium per 
 annum. In addition, the former Company has lately pur- 
 chased a site at Massena, New York, where it is installing 
 a plant of 12,000 H.P. 
 
 The British Aluminium Company employs the Heroult 
 process, and has an available plant of 14,000 H.P. The 
 Heroult process is also exploited by the Societe electro- 
 metallurgique Fran9aise, at La Praz. The Compagnie des 
 Produits Chimiques d'Alais, at St. Michel, employ the Hall 
 and Minet process, whilst the Aluminium Industrie Aktien- 
 Gesellschaft, with an aggregate of 14,000 H.P., also utilises 
 the Heroult process. 
 
 The following table represents the production of alumin- 
 ium in the United States, from 1883 to 1902, inclusive. 
 
 YEAR. 
 
 
 
 POUNDS. YEAR. 
 
 
 POUNDS. 
 
 1883 
 
 
 
 83 1893 
 
 
 333,629 
 
 1884 
 
 
 
 ' 150 1894 
 
 
 550,000 
 
 1885 
 
 
 
 -283 1895 
 
 
 920,000 
 
 1886 
 
 
 
 3,000 1896 
 
 
 1,300,000 
 
 1887 
 
 
 
 18,000 1897 
 
 
 4,000,000 
 
 1888 
 
 
 
 19,000 1898 
 
 
 5,200,000 
 
 1889 
 
 
 
 47,468 1899 
 
 
 6,500,000 
 
 1890 
 
 
 
 61,281 1900 
 
 
 7,150,000 
 
 1891 
 
 
 
 150,000 1901 
 
 
 7,150,000 
 
 1892 
 
 
 
 250,885 1902 
 
 
 7,300,000 
 
 The Heroult Process. In the Heroult process of aluminium 
 manufacture, the lining of the furnace hearth is protected 
 
 183 
 
ELECTRIC FURNACES AND 
 
 from electro-chemical action by a layer of cooled, solidified 
 electrolyte. 
 
 The voltage theoretically required to bring about the 
 electrolysis of the alumina in the Heroult process, is 2-2. 
 Very good quality carbon is required for the anodes, which 
 are, of course, consumed in the process, and, if at all impure, 
 would tend to contaminate the resultant metal. 
 
 The bauxite used in the manufacture of aluminium by the 
 Heroult process at Foyers, is obtained from Larne, in Ireland, 
 and, according to Blount, has the following percentage com- 
 position 
 
 56 per rent. 
 
 Alumina 
 Ferric oxide 
 Silica 
 
 Titanic acid 
 Water 
 
 3 
 
 12 
 3 
 
 26 
 
 100 
 
 The method of preparing alumina from this, being an 
 essential part of the Heroult process of aluminium manu- 
 facture, a brief description, extracted from Practical 
 Electro-Chemistry, by Bertram Blount, is included. 
 
 " The material is crushed so as to pass a quarter-inch 
 mesh sieve, and is gently roasted in a revolving calcining 
 furnace, the temperature being regulated so as to destroy 
 any organic matter, and ensure that all iron shall be present 
 as Fe 2 3 , and nevertheless not to render the alumina in- 
 soluble. The roasted material is powdered so as to pass a 
 sieve having thirty meshes per linear inch, and is digested 
 with a solution of caustic soda of specific gravity 1-45 at a 
 pressure of 70 to 100 Ibs. per square inch. After digestion 
 for two or three hours, the solution is diluted to a specific 
 gravity of 1-23, and is passed through filter presses, and 
 afterwards through cellulose filters, consisting of sieves 
 carrying a layer of cellulose pulp, the whole contrivance 
 somewhat resembling the laboratory apparatus known as a 
 Gooch crucible. 3y this double filtration a satisfactory, 
 
 184 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 clear liquor is obtained. In former processes for the manu- 
 facture of alumina, the alkaline aluminate was decomposed 
 with C0 2 , and the alumina was thus precipitated. The dis- 
 advantage of this process, apart from the cost of the C0 2 , 
 is, that any silica present in solution is also thrown 
 down and contaminates the alumina, and, moreover, the 
 alkali is converted into carbonate, and has to be recausticised 
 before it can be used again for extraction. By Bayer's pro- 
 cess, which is that now in use, the caustic solution of alumina 
 is treated with a small portion of alumina, precipitated in a 
 previous operation ; it is thereby caused to deposit about 
 70 per cent, of its dissolved alumina, if the solution is well 
 agitated and the precipitation allowed to continue for about 
 36 hours. The clear liquor is drawn off, and the alumina 
 washed in a filter press, and dried to some extent by a blast 
 of air, being then roasted at about 1,100C. = 2,012F., in 
 order to render it both anhydrous and non-hygroscopic. 
 The latter quality is necessary, as otherwise the alumina 
 would absorb water during storage, and would not be fit to 
 feed into the electrolytic cell. 
 
 " The caustic soda solution, diluted by retaining a 
 portion of its alumina, is concentrated in a triple effect 
 vacuum evaporator, to its original specific gravity of 1-45, 
 and is then ready for the extraction of another portion of 
 bauxite. It will be seen that the caustic soda serves merely 
 to pick out the alumina from its accompanying impurities, 
 and to deposit it, as it were at the word of command, in a 
 pure state." 
 
 The Heroult furnace in use at Foyers consists of a square 
 iron casing A, Fig. 43, lined with carbon C, and containing 
 the cathode F, which, in the form of a cast-iron plate, rests 
 on the bottom -of the furnace, and is connected externally 
 with the source of current. The anode consists of a bundle, 
 or fagot of carbon rods B suspended over the furnace 
 cavity and adjustable as to its height and consequent depth 
 pf immersion therein. 
 
 185 
 
ELECTRIC FURNACES AND 
 
 The furnace is first filled with cryolite, which is speedily 
 fused when the current is switched on, after which powdered 
 
 alumina is fed in at a 
 regular rate. The differ- 
 ence of potential between 
 the furnace terminals is 
 from 3 to 5 volts, the 
 metallic aluminium col- 
 lecting at the negative 
 electrode, whilst oxygen 
 _ gas is set free at the posi- 
 tive, or carbon electrode, 
 with which it enters into 
 combination, forming the 
 gases monoxide and di- 
 oxide of carbon. The 
 Current density is 1'9 
 amperes per square centi- 
 metre of cathode surface, amounting to about 8,000 
 amperes per furnace, and the temperature at which the 
 operation is carried out varies from 750 to 850C.= 1,382 
 to 1,562F. 
 
 The electrolyte employed in the Heroult process consists of 
 a mixture of 
 
 FIG. 43. 
 
 Fluoride of calcium 
 
 Double fluoride of aluminium and 
 
 sodium (cryolite) 
 Fluoride of aluminium . 
 
 234 parts by weight. 
 
 421 
 
 845 
 
 From 3 to 4 per cent, of a suitable chloride, that of calcium, 
 for example, is added, together with sufficient alumina to 
 form a very stiff mixture. 
 
 The Hall Process. The distinguishing features between the 
 Heroult and Hall processes may be regarded as almost 
 entirely of a mechanical nature, including the method of 
 operation, and such minor details incidental thereto as are 
 essential to the efficiency of the process. 
 
 186 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Mr. Chas. M. Hall, the originator of the aluminium pro- 
 cess now exploited by the Pittsburg Reduction Company, 
 first experimented with his process at the works of the 
 Cowles Electric Smelting and Aluminium Company, Lock- 
 port, New York, U.S.A. 
 
 The furnace hearths are of cast iron, lined interiorly to 
 some considerable thickness with carbon ; they are 6 ft. 
 long by 3 wide by 10 in. deep, and constitute the cathode. 
 The anodes consist of a number of carbon rods, mounted in a 
 special holder, and immersed, at their lower extremities, 
 in the molten electrolyte. Each individual unit of the 
 anode carries a current of 250 amperes, and the total current 
 required for the operation is in the neighbourhood of 10,000 
 amperes. 
 
 The E.M.F. is 5 volts, the power consumed being approxi- 
 mately 65 H.P. The furnaces are worked in series. 
 
 The process consists in the electrolysis of purified alumina, 
 dissolved in a molten bath of the double fluoride of alumin- 
 ium and sodium (A1 2 F 6 .2NaF). The inventor has pub- 
 lished the following data relating to the cost of the materials 
 for the process of aluminium manufacture patented by him 
 
 Alumina . * . 2|d. to 3d. per pound. 
 
 Cryolite . . ... 3d. 
 
 Hydrofluoric acid . . . 2d. to 2%d. ,, 
 
 Carbon for lining furnace hearths l|d. 2d. ,, ,, 
 
 The electrolyte used in the Hall process is prepared in a 
 lead-lined vat, by treating a mixture of alumina, cryolite, 
 and fluorspar with hydrofluoric acid ; the mass is then dried 
 and placed in the furnaces, where it is subjected for some 
 time to the heating action of the current, in order to reduce 
 it to a state of complete fluidity before the alumina is fed in. 
 
 The proportions most commonly used are 
 
 Aluminium fluoride . . .,677 parts by weight. 
 Sodium ,, ... 251 ,, 
 
 Calcium . . .234 
 
 When fusion has been completely effected, pure alumina 
 
ELECTRIC FURNACES AND 
 
 is fed in, in the proportion of 20 per cent, by weight of the 
 solvent, and the like proportion maintained as the operation 
 proceeds. The temperature at which the Hall furnace 
 operates is below 982C.=1,800F. 
 
 The Pittsburg furnaces are tapped once a day, the removal 
 of the molten metal being accompanied by a corresponding 
 increase in the electrical resistance of the furnace charge. 
 Visual indication of the increased resistance is obtained by 
 means of an ordinary incandescent lamp, which is connected 
 in shunt between the electrodes, and, by its increased 
 brilliancy, indicates that the resistance of the fused elec- 
 trolyte is rising and that fresh alumina must be fed in. 
 
 The following rules, to be observed in the thermo-elec- 
 trolytic manufacture of aluminium, have been formulated 
 and published by Mr. Hunt, President of the Pittsburg 
 Reduction Company. 
 
 (1) The solvent, with its dissolved alumina, must be 
 fusible at a moderate temperature. 
 
 (2) The solvent must dissolve alumina freely, e.g., must 
 take up at least 20 per cent, at the working temperature. 
 
 (3) The critical voltage for the solvent must be higher 
 than that for the alumina. 
 
 (4) The specific gravity of the solvent at its working 
 temperature must be lower than that of the aluminium 
 at the same temperature, so that the metal may collect at 
 the bottom of the cell. 
 
 Extended litigation has resulted in the Cowles and Bradley 
 patent claims in the manufacturing process, being upheld 
 in the United States Courts, with the result that the Pitts- 
 burg Reduction Company is now reported to be under a 
 working agreement with the owners of ^ these patents as to 
 royalties, etc. 
 
 Theoretical Considerations. M. Gustav Gin contributed a 
 theoretical paper on the subject of aluminium manufacture 
 at the Fifth International Congress of Applied Chemistry, 
 the salient features of which mav be summarised as follows 
 
 188 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 According to Minet, the physical conditions of the fused 
 electrolyte employed should be 
 
 (1) Low fusion point. 
 
 (2) Sufficient fluidity during the process to permit the 
 molten aluminium to separate out. 
 
 (3) Fusion density to be lower than that of aluminium. 
 
 (4) Low vapour tension, and low ohmic resistance. 
 From a chemical point of view, and for a minimum ex- 
 
 penditure of energy, the aluminium compound selected for 
 electrolysis should have a low heat formation value. The 
 E.M.F. required for the electrolysis of fused aluminium 
 compounds may be divided under two headings, viz. 
 
 (1) That necessary to overcome the ohmic resistance 
 of the fused electrolyte, and the polarisation of the elect- 
 rodes, and secure the required current density, and 
 
 (2) That required to effect the electrolysis and bring 
 about the discharge of the ions. 
 
 The first is easily obtainable, whilst the second is a func- 
 tion of the heat of formation, and can also be arrived at 
 by calculation. Minet has proved experimentally that, 
 at the temperature of the molten mass in the aluminium 
 extraction process, aluminium fluoride is completely dis- 
 sociated into its constituent elements. The following 
 thermal equation therefore applies 
 
 Q=jF+J A1 * 
 
 =50,750 + 40,100 C. gr. d. 
 =90,850 C. gr. d. 
 
 from which, since 23,0670. gr. d. are equivalent to one volt, 
 we get E=3-93 volts. 
 
 Assuming the fluorine thus liberated enters into com- 
 bination with carbon at the anode to form carbon-tetra- 
 fluoride (CFJ, the equation then becomes 
 
 = (50,750 + 40,100) -33,400 
 = 57,450 C. gr. d. 
 189 
 
ELECTRIC FURNACES AND 
 
 from which, E 2-49 volts, a value more nearly in agree- 
 ment with that obtained experimentally by Minet (2-5 
 volts). 
 
 Passing on to a consideration of the electrolysis of alumina, 
 we have 
 
 Q = Jo* + JAI* -J(CO*) 
 =: (34,950 + 40, 100) -24,400 
 = 50,650 C. gr. d. 
 and E=2-19 volts. 
 
 The above calculations have been experimentally con- 
 firmed by Gin, who obtained the figure, 2-30 volts, as the 
 mean of four observations in Le Blanc's method of deter- 
 mining potential differences. 
 
 The thermal equation for aluminium sulphide is : 
 
 Q=Js* + J** -J (C* S*) 
 =(-6,300 + 40,100) -4,350 
 =29,450 C. gr. d. 
 or E = l-27 volt. 
 
 From these data we arrive at the energy required to pro- 
 duce one kilogramme of aluminium from any of the three 
 compounds mentioned. They are as follows 
 
 Al 2 F 6 =Kr + 8-5 K.W. Hours 
 Al 2 3 =Ko + 7-5 
 
 Al 2 S 3 =Ks+4-4 
 
 where KF, Ko, and Ks are quantities varying between 16 
 and 19 k.w. 
 
 The difference in the energy required by any of these 
 three compounds does not therefore exceed 5 k.w. hours 
 per kg. of aluminium produced, and M. Gin therefore con- 
 cludes that any future improvement in the economy of 
 aluminium manufacture will depend upon the cost of the 
 raw materials, and not upon a corresponding improvement 
 in the efficiency of the process itself. 
 
 Chlorine Smelting with Electrolysis. The Swinburne- 
 
 190 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Ashcroft or " Phoenix " process of so-called chlorine smelt- 
 ing, which is the joint invention of Mr. Jas. Swinburne, 
 M.I.G.E., and Mr. E. A. Ashcroft, is a cyclic process for 
 obtaining various metals from their ores, which can be con- 
 ducted with little or no loss, and has, therefore, a promising 
 future in competition with the present wasteful methods 
 of smelting by reduction processes. 
 
 It is appropriately called " chlorine smelting," because 
 chlorine is the vehicle employed to displace the sulphur 
 from the sulphide ores under treatment, only to be re- 
 covered by a subsequent stage in the process, and used over 
 again. 
 
 Various papers and articles dealing with the process have 
 been published from time to time, including a recent one 
 by Mr. Swinburne himself, read before the Faraday Society 
 on June 30, 1903, which deals mainly with the financial 
 aspect of the process. 
 
 The principal features are as follows : The process con- 
 sists in three main stages, which are thus rendered by Mr. 
 Swinburne in the form of chemical equations. Employing 
 the usual chemical symbols for zinc, chlorine, and sulphur, 
 and substituting M and N for other metals present in mixed 
 sulphide ores, Stage I. is represented by the equation 
 
 ZnS +MS +NS + 3C1 2 = ZnCl 2 +MC1 2 +NC1 2 + 3S. 
 Stage II., after the removal of the sulphur, by 
 ZnCl 2 + MC1 2 + NC1 2 + 2Zn = 3ZnCl 2 + M + N 
 
 and, finally, Stage III., consisting in electrolysis of the zinc 
 chloride 
 
 3ZnCl a =3Zn + 3Cl 2 
 
 To put it briefly into words, the process consists in first 
 treating the sulphide ores with chlorine, which displaces 
 the sulphur, and itself enters into combination with the 
 metals thus set free to form chlorides, which latter are 
 then subjected to electrolysis, whereby the metal is sepa- 
 
 191 
 
ELECTRIC FURNACES AND 
 
 rated out, and the chlorine recovered, to be used over 
 again. 
 
 Take galena as an example. It may or may not contain 
 silver, and even gold. It is first treated with chlorine in 
 the presence of heat, which leads to the formation of lead 
 chloride and sulphur, the latter being thrown down or 
 condensed in the form of brimstone, whilst the former, if 
 it be known to contain admixtures of the precious metals, 
 is again treated with metallic lead, which replaces them. 
 The precious metals, if present, form an alloy with the 
 excess of metallic lead present, producing a metallic mix- 
 ture of any desired richness, limited only by its melting point. 
 
 With the exception of a slight loss, which arises in prac- 
 tice, out of the leakage of air into the vats, forming sulphur 
 dioxide, the sulphur is also saved. 
 
 The process is precisely similar with other ores, or mix- 
 tures of ores, and is especially applicable to such refractory 
 samples as have hitherto presented untold difficulties to 
 the metallurgist. 
 
 Though consisting partly in a thermo-chemical, and 
 partly in an electrolytic process, or combination of processes, 
 and therefore not coming, in its entirety, strictly under the 
 category of electrolytic furnace methods, it will be neces- 
 sary to treat of the process in full, in order to render it 
 clear to the reader. 
 
 The crude ores are first crushed, and then fed into a re- 
 ceptacle, called for convenience a transformer, and resem- 
 bling in general appearance a small blast furnace. It is 
 constructed of iron, and lined with fire-clay. Like a blast 
 furnace, it has a small cone at the top, whilst an opening 
 at the base, fitted with a carbon tube, serves for the intro- 
 duction of the chlorine. Outside the transformer, con- 
 nexion is made between this tube and an iron pipe, which 
 conveys the chlorine to it. The action of the apparatus 
 is continuous, and we will assume that it is partly filled 
 with fused chlorides, and gangue floating on top. 
 
 192 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Fresh ore is fed in at the top, and, simultaneously, chlorine 
 is pumped in at the base by way of the iron pipe and car- 
 bon tube. It bubbles up through the fused mass, and in 
 its passage combines with the metals present to form chlo- 
 rides, liberating the sulphur, which passes off as a vapour, 
 and is suitably condensed. 
 
 This chemical action produces a great deal of heat, 
 which serves to maintain the contents of the apparatus at 
 the necessary fusion temperature, and thus render it self- 
 heating. The temperature regulation is an important 
 point ; if it be too low, chloride of sulphur is formed, while, 
 on the other hand, if it be too high, some of the chlorides 
 distil over, and are condensed with the sulphur. It 
 is regulated within the limits of practice, which allows 
 a fairly wide margin, by controlling the rate of admission 
 of the chlorine. The size of the transformer is also a con- 
 trolling feature, and must be such that, when running at 
 full load, sufficient heat is driven off by conduction, and 
 radiation from the outer surface, to keep the contents at 
 or about the required temperature. 
 
 In treating Broken Hill slimes, which have already been 
 tried, for purposes of experiment, by Messrs. Swinburne 
 and Ashcroft, a transformer, having a capacity of 10 tons, 
 was successfully employed, but Mr. Swinburne ventures 
 the opinion that a larger size, with thinner lining, would 
 probably yield improved results. 
 
 Visual indication as to the state of the transformer con- 
 tents during the progress of the operation, and whether a 
 further supply of ore is required, may be secured by making 
 a hole at the top of the transformer. The sulphur vapour 
 and gaseous products in passing over through the legitimate 
 outlet to the condensers and chimney create a slight indraft 
 of air at this opening, with the result that sulphur dioxide 
 is formed, and burns with a blue name. This latter indi- 
 cates that all is well, and that sufficient ore is present. If 
 there be an insufficiency of ore in the transformer, fumes 
 
 193 o 
 
ELECTRIC FURNACES AND 
 
 of ferric chloride are formed at the top of the chamber, and 
 pass over with the sulphur vapour ; it is necessary to avoid 
 this contingency, and the ore feed must consequently be 
 watched until the process is completed. 
 
 When the transformer is full, the supply of ore is stopped, 
 but the process is continued until brown fumes begin to 
 make their appearance. The charge is then tapped off, 
 and permitted to cool. 
 
 The intermediate or chemical stage of the process varies 
 with the ores under treatment, depending naturally upon 
 the metals present. It is best described in Mr. Swinburne's 
 own words, the Broken Hill slimes being again taken as an 
 example. 
 
 " We now come to the intermediate or chemical stage 
 of the process. This varies with the ore used. We may 
 take the Broken Hill slimes, and imagine there is copper 
 too. This would be about as troublesome an ore as we 
 could have. We have used the Broken Hill ore, and mixed 
 it with a Tasmanian copper ore, but we have chiefly worked 
 on it alone. The fused mass from the transformer con- 
 sists of chlorides of lead, zinc, iron, manganese, copper, 
 silver and gangue. It is run into water, and through a 
 filter press when cool enough. This takes out the gangue 
 and lead chloride, carrying most of the silver. The gangue 
 is easily separated from the lead and silver chlorides, and 
 these chlorides are then dried and fused in contact with 
 lead, which extracts the silver and any gold ; and then 
 with zinc, which gives lead, practically pure, and anhy- 
 drous neutral zinc chloride, which is ready for the electro- 
 lysis vats. 
 
 " The filtrate from the press contains a little lead and 
 silver, in solution, and copper, iron, manganese, and zinc. 
 The lead and silver are taken out with spongy copper. 
 The copper is taken out as sponge or ' cement ' by zinc, 
 and we have left iron, manganese and zinc chlorides. 
 
 " The iron is chlorinated up to the ferric state, and zinc 
 
 194 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 oxide is added to cause precipitation. This throws down 
 hydrated ferric oxide. This is the base of iron paint, and 
 is marketable, its value depending on the colour obtained. 
 The solution is further chlorinised in the presence of 
 more zinc oxide, and the manganese goes down as per- 
 oxide. 
 
 " We have now substituted zinc for all the other metals 
 in their chlorides, and have nothing left but zinc chloride. 
 This is evaporated down carefully, and fused. This de- 
 composes some of the chloride, and makes an oxy chloride. 
 Steinhart evaporates in vacuo, to produce neutral anhydrous 
 zinc chloride. We have not done this ; we find that with 
 cautious boiling down there is not much oxygen in the 
 final result. 
 
 " This is got rid of in open preliminary vats, which use 
 inexpensive anodes, which are gradually used up. The 
 consumption is less than if all the oxygen went off as mon- 
 oxide. The anhydrous neutral chloride from these vats is 
 then added to that from the lead chloride substitution, 
 and is taken to the final electrolysis vats." 
 
 The final stage, which constitutes the part actually played 
 by the electrolytic furnace, has for its object, as before 
 stated, the final separation of the remaining metal (zinc), 
 and the recovery of the chlorine. 
 
 The furnace itself is a very simple affair, consisting of an 
 iron chamber, lined with fire-brick ; in course of time the 
 fused chloride permeates the pores of this lining, and solidi- 
 fies therein, so that, in reality, the lining itself ultimately 
 becomes one of chloride. 
 
 The cathode is of fused zinc, and the anode of carbon, 
 which latter is not attacked by chlorine. Some care is 
 necessary in the selection of a suitable grade of carbon, 
 but, once found, the anodes stand extremely well, and are 
 permanent, the temperature not being sufficiently high to 
 cause combustion. 
 
 In practice, a slight suction is kept on the furnace during 
 
 195 
 
ELECTRIC FURNACES AND 
 
 the electrolysis, so that, if there be any leakage, it is in the 
 nature of ingress of air rather than egress of chlorine. 
 
 In the experiments which have been conducted up to 
 the present, a comparatively small furnace, consuming 
 3,000 amperes, has been used ; but a 10,000 ampere furnace 
 is being tried, this being the maximum current density 
 which can conveniently be dealt with. An electromotive 
 force of approximately 4 volts per furnace is necessary to 
 maintain the action. 
 
 The furnace used in the experimental trials of the process, 
 carried out at the old Cowles Aluminium works at Milton, 
 
 Staffordshire, is illus- 
 trated in section in 
 Fig. 44. It consists 
 of an outer steel shell 
 S, 6 ft. in diameter, 
 lined for a depth of 
 18 in. with fire-brick, 
 the inner layers be- 
 ing set in a special 
 cement. The brick- 
 work B forms a 
 cylindrical chamber 
 
 or hearth H, at the bottom of which is placed the 
 cathode Z, in the shape of a mass of molten zinc, weighing 
 approximately one ton. Electrical connexion with this 
 latter is secured by means of a massive steel block A let 
 into a recess in the brickwork, and connected with the 
 external circuit by fire screwed copper tubes t t, which are 
 also made part of a system through which air is caused to 
 circulate for cooling purposes. F is the anode, and con- 
 sists of an iron cover plate grouted in place with a special 
 heat-resisting cement, and serving as a terminal support 
 for 120 2 J-inch hard carbon rods c c, which are submerged 
 in the molten electrolyte to a depth of about six inches. 
 The current required to operate a furnace of this descrip- 
 
 196 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 tion is from 3,000 to 4,000 amperes at 4 to 4-5 volts, the 
 working temperature being 475C= 887F., and the out- 
 put one ton of metallic zinc per week. 
 
 The cost of working the process is, roughly, 30<s. per ton 
 of ore, plus the cost of the electrical energy. 
 
 Sodium and Caustic Soda. Dr. Rogers, as early as 1889, 
 studied the possibilities of preparing lead-sodium alloys by 
 electrical means, and actually succeeded in manufacturing 
 an alloy containing 17 per cent, of sodium, by the electro- 
 lysis of fused sodium chloride in a 30 Ib. crucible. A lead 
 cathode was employed, and the process called for an ex- 
 penditure of 72 amperes at 21 volts for a period of 2 
 hours. 
 
 To this investigator belongs the credit of having first 
 suggested a basic lining for the furnace used in the electro- 
 lysis of fused sodium chloride. 
 
 Dr. Borchers's process of sodium extraction by deposi- 
 tion of the metal upon a flowing stream of lead involved a 
 current density of 3-2 amperes per square inch of cathode, 
 the E.M.F. being 6 to 8 volts. 
 
 C. T. J. Vautin's process was originally based upon an 
 attempt to obtain caustic solutions and chlorine, but, owing 
 to certain natural and unforeseen tendencies on the part 
 of the various constituents of the electrodes and electro- 
 lyte, the system, when tried at Keighley, Yorkshire, in 
 1893 to 1894, proved commercially impracticable. 
 
 Vautin, however, suggested the maintenance of fusion 
 of the electrolyte by electric heat, and both this suggestion, 
 and that of Dr. Rogers concerning basic linings for the 
 furnaces, constitute two at least of the essential features 
 of Acker's successful solution of the problem. 
 
 The Vautin process consisted in the electrolysis of fused 
 sodium chloride, employing a lead cathode, which was 
 intended to alloy with the sodium at the moment of libera- 
 tion, and thus protect it from further action, electrolytic 
 or otherwise. 
 
 197 
 
ELECTRIC FURNACES AND 
 
 The practical difficulties which led to the abandonment 
 of the process were the destruction of the containing vessel 
 by the combined action of the contents at high tempera- 
 tures, and the failure of the cathode surface to remain 
 active once the action had been started, owing to the forma- 
 tion of a crust of the lead-sodium alloy, which impeded 
 further combination between the two metals ; the sodium 
 subsequently set free either burning away on rising to the 
 surface of the fused electrolyte, or re-combining with the 
 nascent chlorine to form sub-chlorides. 
 
 Leon Hulin, in 1890, commenced experimenting at Modane, 
 France, independently of Vautin, and without knowledge 
 of what had already been accomplished in the igneous 
 electrolysis of sodium chloride. In order to get over the 
 crust formation difficulty, which led to the failure of Vautin's 
 process, a compound electrolyte was employed, contain- 
 ing a definite proportion of lead chloride. Electrolysis 
 resulted in the simultaneous liberation of lead and sodium 
 at the cathode, which immediately combined to form the 
 alloy, and, in this condition, were immune from the action 
 of the chlorine set free at the anode. 
 
 The electrolyte consisted of the mixed chlorides of lead 
 and sodium, whilst the anodes were also compound, con- 
 sisting of carbon and lead, the latter, or metallic portion, 
 being connected in circuit with a controlling resistance or 
 rheostat, by means of which the proportion of the main 
 current passing through it could be controlled, and the 
 amount of lead dissolved and deposited similarly governed. 
 
 The Hulin process has been worked on a limited indus- 
 trial scale at the paper works of MM. Matussiere et Forest, 
 Modane. Four furnaces were employed, the terminal 
 E.M.F. of each being 7 volts, and the current density 
 between 1-5 and 2-5 amperes per square centimetre of 
 cathode surface. The best results were obtained with a 
 current density of 1-9 amperes per square centimetre. A 
 composite lead-carbon anode was used, the metallic portion 
 
 198 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 only conveying a fraction (12 per cent.) of the total cur- 
 rent, and by its combination with the chlorine liberated 
 serving to maintain the requisite proportion of lead chlo- 
 ride in the electrolyte. The yield per k.w. hour is said to 
 have been 108 grammes chlorine, and 72 grammes sodium, 
 or 2-48 kgs. chlorine, and 1-66 kgs. sodium per k.w. day 
 of 23 hours. 
 
 According to Kershaw, the theoretical voltage required 
 in the process 
 is 4-2, and the 
 yield, per k.w. 
 hour, 356 
 grammes of 
 caustic soda and 
 314 grammes 
 of chlorine. 
 Against these 
 may be set the 
 actual figures 
 obtained in 1 
 practice, viz., 7, 
 150, and 129, 
 
 respectively, showing a current efficiency of 69-3, and an 
 energy efficiency of 41-5 per cent. 
 
 The Acker electrolytic furnace, for the manufacture of 
 caustic soda from fused electrolytes, as exploited by the 
 Acker Process Company at Niagara Falls, is shown in 
 diagrammatic section in Fig. 45. It consists, essentially, 
 of a rectangular trough-like structure or hearth F, having 
 a steel base s, refractory side walls m of magnesia, and a 
 removable horizontal false bottom, or dividing partition p, 
 also of steel. A flue / provides for the escape of the chlo- 
 rine gas evolved during the process, which may be used 
 for the manufacture of bleaching powder, whilst one end 
 of the structure slopes to a well W fitted with an injector A, 
 for the introduction of steam under pressure. The catho- 
 
 199 
 
ELECTRIC FURNACES AND 
 
 die or negative electrode connexion with the source of 
 current is made through a metal tube T communicating 
 with the well W. The anodes C C are cylindrical graphite 
 rods, projecting vertically through the cover to within a 
 short distance of the steel partition p. N is a mass of the 
 raw material, sodium chloride or common salt, which is 
 packed on top of the furnace, and performs a triple duty, 
 viz., that of conserving the heat generated, hermetically 
 sealing all except the legitimate gas outlets, and providing 
 a source of supply as the charge becomes exhausted. P is 
 a mass of molten lead, extending upwards to a level slightly 
 above that of the false bottom, or partition p, and in elec- 
 trical connexion with the cathode. 
 
 The process of operation is as follows : Steam being 
 admitted to the injector through the inlet pipe i, carries 
 up with it, from the base of the injector well, a mixture of 
 lead, caustic soda, and hydrogen gas, the resultant products 
 of[the electrolytic process going on in F. On reaching the 
 upper chamber above the well, W, these three constituents 
 part company, the alloy flowing back, over an inclined par- 
 tition, into the main furnace F, whilst the caustic liquid 
 passes over into the receiver R. Hydrogen gas is available 
 at the outlet o, and may be utilized for a variety of purposes, 
 chief among which may be mentioned its application to 
 an oxyhydrogen burner or burners for the preliminary 
 fusion of the charge. 
 
 The following facts, in connexion with the Acker process, 
 as carried on at Niagara Falls, were published by the in- 
 ventor in the course of a paper read by him before the 
 American Electro-Chemical Society at Philadelphia, in 1902. 
 Each furnace has four anodes, and consumes a total current 
 of 8,000 amperes or 2,000 amperes to each anode, represent- 
 ing a current density of 3 amperes per square centimetre. 
 
 The efficiency of the process is 100 per cent., and the 
 works have been in continuous operation since 1900, with 
 but few temporary accidental interruptions. 
 
 20Q 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The only trouble experienced is said to have been in the 
 electrode connexions at the carbon-copper junction, the 
 metal being gradually corroded or eaten away. An im- 
 proved method of construction has, however, minimised 
 this drawback. 
 
 The anodes consist of graphitised carbon blocks, and are 
 manufactured by the International Acheson Graphite Com- 
 pany. There are four to each furnace, each block having 
 a section approximately 7J by 14 inches, and carrying 
 2,000 amperes. They are unattacked by the electrolyte, 
 the only weak point being the junction between the metal 
 terminal clamps and the carbon. This is protected by a 
 sleeve of basic cement. 
 
 The furnaces are lined with magnesite bricks, placed 
 loosely in position, without cement, but held together firmly 
 by the solidified electrolyte, which permeates them when in 
 a fused condition, and, on solidifying, locks them rigidly 
 together. 
 
 In 1902 Acker furnaces to the number of forty-five were 
 in operation, consuming 3,250 E.H.P., or 8,000 amperes 
 at 300 volts. Each individual furnace requires a current 
 of 8,000 amperes at 7 volts terminal E.M.F., or 75 E.H.P., and 
 produces, on an average, 290 kgs. of anhydrous caustic daily. 
 
 The successful conduct of the process consists in con- 
 tinuously removing the sodium alloy as quickly as it forms, 
 the reason being that it is unstable in the presence of the 
 fused salt. 
 
 The steam from the injector performs a double office, 
 in that it sets up a rapid circulation, and also oxidises the 
 sodium of the alloy formed, producing fused, anhydrous, 
 sodium hydrate (NaOH). 
 
 The chlorine gas evolved during the process is drawn off 
 by means of a fan, and utilized in the manufacture of bleach- 
 ing powder. The furnaces are of cast iron, lined to a level 
 above that of the fused lead ; they are arranged in pairs, 
 one on either side of a common flue. 
 
 201 
 
ELECTRIC FURNACES AND 
 
 The Elliott-Cresson Medal of the Franklin Institute has 
 been awarded to Mr. Chas. E. Acker for his electrolytic 
 method of manufacturing caustic soda and chlorine as 
 described above. 
 
 The advantages of the Acker process, as compared with 
 other purely electrolytic methods, are thus epitomised by 
 the Committee who recommended the award 
 
 (1) Its directness. 
 
 (2) The very heavy current sent through each pot. 
 
 (3) The absence of the evaporation of caustic and boiling 
 down to dryness. 
 
 (4) The absence of water solutions, requiring special 
 pumps, and a complicated circulating system. 
 
 (5) The absence of mercury. 
 
 The disadvantages, on the other hand, are 
 
 (1) A large power consumption, 6-75 volts, in place of 4-5. 
 
 (2) The rapid destruction of the apparatus. 
 
 (3) The rapid destruction of the anodes. 
 
 (4) The more arduous work of the pot-men. 
 
 Calculations on the part of the committee show the fol- 
 lowing distribution of the total energy required, throughout 
 the various stages of the process 
 
 54 per cent, is usefully applied to chemical separation of 
 the sodium from the chlorine. 
 
 9 per cent, in fusing the salt. 
 
 Lost by radiation, 37 per cent. 
 
 In the later types of furnace for the application of the 
 Acker process of caustic soda manufacture, a more rapid 
 circulation of the fused lead is secured by means of a bladed 
 propeller, which revolves at the bottom of the injector well. 
 The effect of this additional feature is said to be a more con- 
 centrated product, in that less steam is required than was 
 the case in the original apparatus, where it was depended 
 upon to produce the requisite mechanical agitation of the 
 lead. 
 [ Castner's sodium process consists in the electrolysis of 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 fused sodium hydrate. An iron pan, or crucible, contains 
 the electrolyte, which is initially fused by a furnace below 
 it. The cathode, of nickel, passes vertically up through an 
 opening in the base, and is surrounded, at its upper extre- 
 mity, by an annular anode, also of nickel. The fused con- 
 dition of the electrolyte is partially maintained by an 
 excess current, and, for the remainder, by the auxiliary com- 
 bustion furnace. Sodium is set free, in liquid form, and 
 rises to the surface of the molten mass, whence it is periodi- 
 cally removed. 
 
 Temperature regulation needs to be very exact, as, if the 
 heating effect be only slightly in excess of that required, 
 recombination of the liberated sodium takes place, and no 
 metal is set free. 
 
 The current efficiency of the process is from 70 to 90 per 
 cent., and it is in commercial operation at Runcorn, in 
 this country, Niagara, U.S.A., and in Germany and France. 
 
 The Niagara plant consists of some 120 furnaces, each 
 of which requires a current of 1,200 amperes, at an E.M.F. 
 of 5 volts to operate it. 
 
 The total output of sodium in 1897 was 260 tons. 
 
 Commenting, in Zeitschrift fur Elektrochemie, about the 
 beginning of 1903, on the Castner process for the extraction 
 of sodium, Professor Le Blanc and Herr Erode remark that 
 neither the inventor nor those associated with the conduct 
 of the process on an industrial scale understand exactly 
 what goes on in the furnace during the reaction. They 
 therefore conducted experiments with a view to ascertain- 
 ing this, and summarise their results as follows 
 
 (1) Molten sodium hydrate, containing water, shows two 
 decomposition values : 1-3, and 2-2 volts. When water is 
 absent, the lower of these two values disappears. 
 
 (2) At the lower E.M.F. , hydrogen and oxygen are sepa- 
 rated ; at the higher E.M.F., sodium and oxygen. The 
 yield of oxygen is never quantitative, but it increases with 
 the current. The yield of hydrogen is quantitative under 
 
 203 
 
ELECTRIC FURNACES AND 
 
 2-2 volts, so long as no free sodium is separated at the 
 cathode. When no water is present in the hydrate as an 
 impurity, only sodium is obtained, with an E.M.F. above 
 2-2 volts. This fact proves that pure molten sodium 
 hydrate contains only the ions Na and OH, and that 
 neither O nor H ions are present in the electrolyte. 
 
 (3) Molten caustic soda quickly arrives at a state of 
 equilibrium as regards moisture exchanges with the atmo- 
 sphere. Normally it contains a considerable amount of 
 water ; if it contains no water, it is strongly hygroscopic. 
 
 (4) Molten caustic soda, containing free sodium as an 
 impurity, yields hydrogen at the anode, in addition to 
 oxygen, on electrolysis with high current densities. This 
 evolution of hydrogen can only result from the separation 
 and discharge of OH ions. 
 
 The same authorities are responsible for the statement 
 that potassium cannot be similarly separated by electro- 
 lysis of its fused hydrate, as claimed in the Castner patent, 
 but that oxidation has to be guarded against by depositing 
 a layer of petroleum on the surface of the fused mass, when 
 globules of metallic potassium are set free. 
 
 G. P. Scholl has improved upon the original method by 
 effecting an economy in the current required, and, at the 
 same time, eliminating the evolution of hydrogen, which 
 takes place at the cathode in Castner's process. He adds to 
 the fused electrolyte some 50 per cent, of sodium sulphide, 
 and reduces the voltage of the operation to that required 
 for the decomposition of the sulphide only ; theoretically, 
 1-8 volt. Sodium is, as before, set free at the cathode, 
 whilst the liberated sulphur, reacting with the fused caustic, 
 again forms sulphide of sodium, which is again electrolysed, 
 and so the cycle goes on, fresh caustic being added from 
 time to time in quantities proportional to the sodium set 
 free. The process is said to effect a considerable economy 
 in current, as compared with the original Castner method. 
 
 The Fischer process for the electrolytic manufacture of 
 
 204 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 sodium is represented diagrammatically in Fig. 46, where 
 C, is a shallow crucible, or furnace hearth, divided into two 
 portions by a central partition p, which does not extend 
 quite to the bottom. The electrodes E E 2 are horizontally 
 disposed, being introduced through the walls of the crucible. 
 The anode E is a carbon rod, whilst the cathode E 2 is a 
 metal tube, placed with its axis on a level with the surface 
 of the fused electrolyte. The side of the crucible contain- 
 ing the negative tubular electrode E 2 is kept cool by a 
 water jacket, not shown in the figure. The raw charge 
 consists of a mix- 
 ture of 79 parts by 
 weight of potassium 
 chloride, with 59 of 
 sodium chloride, or 
 common salt. The ^ 
 
 resultant sodium Fi G . 46. 
 
 contains an admix- 
 ture of about 1 per cent, of potassium, and is drawn off 
 through the tubular electrode E 2 . 
 
 The Darling electrolytic furnace, which is employed in 
 the manufacture of nitric acid and metallic sodium from 
 nitrate of soda, is of comparatively simple construction. 
 The anode consists of an outer iron vessel, within which 
 are placed two concentric cylinders of perforated iron. A 
 shunt connexion between these latter and the outer iron 
 anode protects them from injury during the action of the 
 furnace, whilst the space between the two perforated walls 
 is filled in with powdered magnesia. 
 
 The fused salt, sodium nitrate, is contained within the 
 inner of the two cylinders, the cathode, a carbon rod, being 
 immersed in it. Nitric acid gas is set free in the outer 
 space between the walls of the anode and the concentric 
 cylinders, and is conveyed, by an outlet pipe, to condensing 
 apparatus, where it is converted into the liquid acid. The 
 sodium, on the other hand, rises to the surface of the fused 
 
 205 
 
ELECTRIC FURNACES AND 
 
 salt contained within the inner perforated cylinder, whence 
 it is removed at intervals by means of a ladle, and placed 
 in tin vessels containing a small quantity of paraffin, which 
 effectually protects it from the action of the atmosphere. 
 The fall of potential between the terminals of each furnace, 
 whilst the operation is in progress, is 15 volts. A current 
 of 400 amperes per furnace is required. 
 
 The Darling process is worked by Harrison Bros. & Co., 
 of Philadelphia. 
 
 Calcium and Strontium. Calcium, as is well known to 
 the chemist and metallurgist, is of great value in various 
 industries as a reducing agent, the only drawback to its 
 widespread use being its present comparatively high market 
 value, which renders it unsuitable for all but experimental 
 purposes, where the question of cost is, within certain 
 limits, no object. 
 
 The high price charged for the metal calcium is due to 
 the expensive method of production, which consists in 
 subjecting a dilute solution of calcium chloride to electro- 
 lysis in the presence of a mercury cathode ; an amalgam 
 of calcium with mercury is thus obtained, from which the 
 mercury is subsequently driven off by the aid of heat. 
 
 Apart from the cost of the process, pure calcium is never 
 produced by this method, there being always a small per- 
 centage of metallic mercury present in the resultant product. 
 
 Messrs. Borchers and Stockem, recognising the need for 
 a cheaper and more efficient method of production, turned 
 their attention to the possibilities offered by an electric 
 furnace method, and have recently succeeded in designing 
 an electrolytic furnace in which pure metallic calcium is 
 readily manufactured from its fused chloride. 
 
 The fundamental principle involved consists in the 
 employment of a small cathode, and correspondingly 
 large anode, between which the fused mass is brought to 
 a red heat, when -the metallic calcium makes its appearance 
 around the former as a spongy mass. On removing this 
 
 206 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 latter from the furnace, and immersing it in petroleum, a 
 porous residue is obtained, some 50 to 60 per cent, of which 
 consists of pure calcium. The mass is compressed or 
 squeezed whilst still warm, in order to get rid of the 
 chloride with which it is saturated, and a product containing 
 fully 90 per cent, of pure metal is the result. The latter 
 is then fused in an hermetically sealed chamber, and there- 
 by converted into a firm silvery mass of metallic calcium. 
 Calcium is claimed to have been produced by this method at 
 a cost of approximately 3<9. per kilogramme, about 1/5, 000th 
 of the cost of the original process. 
 
 To pass on, however, to a consideration of the furnace 
 employed ; let us turn our 
 attention to Fig. 47, which 
 represents a section of the 
 latter. It consists of an 
 outer carbon cylinder C 
 built up of several longi- 
 tudinal sections, each 
 keyed into its neighbour, 
 and the whole held to- 
 gether by a metal ring or 
 band B, which, at the 
 same time, constitutes a FlG - 47 - 
 
 terminal connexion, the 
 carbon cylinder being the anode. 
 
 The upper portion of the cylinder is open, whilst its lower 
 end is closed by a fire-clay cylinder c surrounding and 
 insulating a metal chamber A, which performs the double 
 office of a circulating water tank for cooling purposes, and 
 a support for the cathode Fe, which consists of an iron rod 
 placed axially with regard to the cylindrical anode, and 
 screwed at its base into the upper portion of the cooling 
 chamber A. Inlet and outlet pipes i and o respectively, 
 serve to convey the water to and from the chamber, whilst 
 it is protected from the heat of the furnace by a layer of 
 
 207 
 
ELECTRIC FURNACES AND 
 
 fluor-spar s, which, owing to its high melting point, remains 
 solid during the operation of the furnace. 
 
 The mass of calcium chloride M is placed above this, the 
 action of the furnace being started by several thin carbon 
 rods, placed radially, like the spokes of a wheel, between 
 the iron cathode Fe and the inner surface of the carbon 
 cylinder. The heat set up by the flow of current through 
 these rods serves to start the fusion of the upper layer of 
 chloride, and they are subsequently removed, leaving the 
 process of electrolysis to continue through the initially 
 fused mass. 
 
 The position taken up by the resultant spongy calcium 
 is indicated in the figure by Ca. 
 
 Strontium. The apparatus for the manufacture of 
 strontium, is, with some slight modifications, incidental to 
 the properties of the separated metal itself, similar to that 
 detailed above for the production of calcium. Unlike the 
 latter, strontium separates out from its compounds in the 
 form of small spherical masses, which tend to rise to the 
 surface of the molten salt, and there again enter into 
 chemical combination with the chlorine from which they 
 have just been liberated. 
 
 To obviate this drawback, the iron cathode is made 
 shorter, such that its upper extremity only reaches to a 
 point just above the lower edge of the cylindrical carbon 
 anode. The latter rests upon a cylindrical fireproof struc- 
 ture of insulating material, such as fire-clay, which, in turn, 
 is supported in a cup-shaped depression in the upper portion 
 of the cooling chamber. The latter is, in the strontium 
 furnace, given a larger diameter than the carbon anode, 
 and it is in the central depression or basin, around which 
 the anode rests, that the metallic strontium collects during 
 the action of the furnace, and is solidified by the cooling 
 effect of the circulating water. 
 
 The temperature of the operation in Borchers' and 
 Stockem's electrolytic manufacture of metallic calcium, is 
 
 208 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 a red heat, above the melting point of calcium chloride, 
 but below that of metallic calcium. 
 
 Dr. K. Arndt, commenting on the former method for the 
 manufacture of metallic calcium, in Zeitschrift fur Elektro- 
 chemie, November 13, 1902, states that he has obtained 
 a similar result with simpler apparatus. 
 
 His furnace consisted of a coated iron crucible, a carbon 
 anode, and, as cathode, a thick iron wire. With this com- 
 paratively simple outfit, he succeeded in obtaining metallic 
 calcium in large well-fused pieces, capable of being easily 
 worked with a file or hammer. An analysis of samples 
 showed 99 per cent, calcium, and 1 per cent, silicon, but no 
 iron or aluminium. 
 
 Arndt's furnace for the separation of metallic calcium by 
 igneous electrolysis, consists of an iron crucible, lined in- 
 teriorly with a paste made up of kaolin and water. The 
 base is additionally protected by a layer of calcium fluoride, 
 the whole being dried for a period of twenty-four hours in 
 a warm place. The anode is a carbon rod, and the cathode, 
 an iron wire, sheathed to within 4 c.m. of its lower extremity 
 in a porcelain tube. 
 
 The furnace is cooled by conduction and radiation, being 
 mounted, for that purpose, on a massive cast-iron base. It 
 is filled with anhydrous calcium chloride, and the action 
 started by striking an arc between the anode and an inde- 
 pendent carbon rod held in the hand, until the temperature 
 has risen sufficiently to permit the current to flow through 
 the alternative path provided by the fused electrolyte. 
 The current is then maintained constant at an E.M.F. of 
 from 20 to 25 volts, the reaction taking place quite quietly. 
 The product is obtained in large well fused masses, a fresh 
 surface of which is nearly white, but quickly assumes a 
 yellowish tinge. The metal is readily workable, and is of 
 99 per cent, purity. 
 
 Manganese, and its alloy with iron, commonly known as 
 ferro-manganese, has lately proved a subject for extensive 
 
 209 p 
 
ELECTRIC FURNACES AND 
 
 electric furnace operations, both experimental and industrial. 
 
 The Simon electric furnace process of ferro-manganese 
 manufacture, which has been successfully applied in France, 
 is electrolytic. It consists in the electrolysis of a mass of 
 fused calcium fluoride, in which oxide of manganese has 
 been previously dissolved. In this respect it bears a re- 
 semblance to the aluminium extraction processes of Heroult 
 and Hal). 
 
 Careful temperature regulation is here, as in many other 
 furnace processes, an essential feature of the operation, 
 metallic manganese being volatile at a temperature but 
 slightly in excess of its melting point. 
 
 A raw charge having the following composition is fed 
 into the furnace 
 
 Oxide of Manganese 
 Silica . 
 
 Lime .... 
 Magnetic oxide of Iron . 
 
 6 parts. 
 6 
 3 ,. 
 
 0-6 ,. 
 
 Under the thermo-electrolytic action of the furnace, the 
 oxide of manganese is reduced in part by electrolysis, and 
 also by chemical combination of its oxygen with the carbon 
 of the electrodes. The results are found to be most favour- 
 able when carbon is mixed with the oxide before its intro- 
 duction into the furnace. The product has the following 
 composition 
 
 Manganese . 
 
 Iron 
 
 Silicon 
 
 Carbon 
 
 Phosphorus 
 
 It may be added that, should silicon and phosphorus 
 be present in the bath as impurities, they combine with 
 the fluorine as the result of secondary reactions and pass 
 off as volatile by-products. 
 
 The above process, as carried out in practice, involves a 
 power expenditure of 3,475 k.w. hours per ton of ferro- 
 manganese, each furnace having a capacity of at least 
 
 210 
 
 84-00 per cent. 
 8-30 
 0-20 
 7-10 ., 
 0-10 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 150 k.w., and a terminal E.M.F. of 30 volts. Of this, 
 only 7-5 volts is actually required to effect the electrolysis, 
 the balance being neccessry to ensure the passage of the 
 current required to maintain the general high temperature 
 of the mass. 
 
 In 1902 it was decided to erect a plant for the manu- 
 facture of ferro-manganese by this method, at Orlu, in the 
 Pyrenees. The cost of production, with cheap water 
 power available, was estimated at 8 85. Qd. per ton. Further 
 particulars are wanting. 
 
 Metallic Lead has been prepared, though not on a com- 
 mercial scale, by electrolysis of its fused salts. The appa- 
 ratus and method are due to Borchers ; the furnace is of 
 cast iron, and is in two parts, separated from one another 
 by a water-cooled insulating joint, which, by virtue of its 
 lower general temperature, is surrounded and protected 
 from the action of the fused salt and metal by a coating 
 of solidified salt. 
 
 One side of the furnace, comprising the anode, is placed 
 at an angle, its inner surface being interrupted by a series 
 of deep horizontal grooves or corrugations, serving to 
 retain a portion of the molten lead, which is constantly 
 flowing over them from a reservoir above, and is simul- 
 taneously run off from the hearth of the furnace by way 
 of a siphon. 
 
 The furnace, or electrolytic vat, is itself placed in a flue 
 of an auxiliary furnace, which maintains its contents in a 
 state of fusion. The electrolyte consists of a mixture of 
 potassium and sodium chlorides, and lead oxy chloride ; 
 the resultant lead collects in the cathode side of the hearth, 
 whence it is continuously removed by a second siphon. 
 
 Borchers employed a current density of one ampere per 
 square centimetre, at a pressure of 0-5 volt, the yield being 
 5 kilogrammes of metallic lead her 1 e.h.p. hour. 
 
 Magnesium is produced by the electrolysis of fused 
 carnallite, a double chloride of magnesium and potassium, 
 
 211 
 
ELECTRIC FURNACES AND 
 
 which gives up its water of crystallisation, and assumes a 
 clear fluid condition at a temperature below 700C = 
 1,292F. The electrolysis is carried out in a closed cham- 
 ber, an inert gas being passed through during the process. 
 Chlorine gas is evolved, whilst metallic magnesium is set 
 free, and floats on the surface of the electrolyte. It is 
 essential that these two products of the reaction be kept 
 separate from one another, and that the magnesium be 
 not allowed to come into contact with oxygen, hence the 
 necessity for passing an inert gas through the furnace 
 chamber. A current density of -15 ampere per square 
 centimetre of cathode cross-section is required to bring 
 about the electrolysis. 
 
 Zinc Borchers has experimented on the extraction of 
 metallic zinc by electrolysis of the fused chloride, but the 
 process, owing to certain difficulties in the way of its 
 practical application, has never been commercially devel- 
 oped. The electrolytic furnace consisted of a leaden vessel, 
 with close-fitting lid, hermetically sealed in position by 
 solidified zinc chloride. A zinc lining, bent to fit the interior 
 of the crucible, formed the cathode, whilst the anode con- 
 sisted of a vertical carbon rod 
 
 Means were provided for replenishing the charge of zinc 
 chloride and conducting away the chlorine gas evolved. 
 Extraneous heat was requisitioned to start the fusion, 
 which was subsequently maintained by the passage of the 
 current, together with the heat of the reaction. 
 
 M. Gin has succeeded in preparing Vanadium and its 
 compounds by an electrolytic furnace method which he 
 described in a paper before the 1903 International Congress 
 of Chemistry at Berlin. The principle of the process 
 depends upon the great conductivity of vanadium trioxide, 
 and on the facility with which the tri-fluoride is formed, 
 when the trioxide is reacted upon by fluorine in the pre- 
 sence of carbon. 
 
 The most important preliminary to the actual process 
 
 212 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 is the preparation of the anodes. Vanadium trioxide, 
 prepared by calcining vanadic acid in the presence of carbon, 
 is mixed with a suitable proportion of retort carbon and 
 powdered resin, and the mass worked up, under heat, to 
 a plastic paste. The latter is then subjected to hydraulic 
 pressure, under which it is extruded in either cylindrical 
 or prismatic form, through dies. 
 
 The shaped rods, thus formed, are then subjected to a 
 high temperature in an hermetically sealed oven. The 
 resultant electrodes are said to be capable of withstanding 
 a current density equal to 0'7 of that of carbon electrodes 
 of the same section. A bundle of these rods constitutes 
 the anode, whilst the cathode is a block of iron. 
 
 The electrolyte consists of fused vanadium fluoride, 
 and, under the action of the furnace, vanadium is set free 
 at the cathode and fluorine at the anode. The latter again 
 enters into combination to form a fresh quantity of vanadium 
 fluoride, and the electrolysis is repeated. The most suitable 
 current density is said to be 2 amperes per square 
 centimetre of anode section, and 6 amperes per square 
 centimetre of cathode section, the terminal E.M.F. being 
 11 to 12 volts. 
 
 For alloys containing more than 25 per cent, vanadium, 
 the cross section of the anode should be appreciably less 
 than the total active surface of the "anodes. 
 
 Iron-vanadium alloys, and an increased fluidity, may be 
 obtained by adding fluoride of iron to the fused mass, but, 
 if no iron be added, almost pure vanadium is available, 
 which, however, has to be removed from the furnace in a 
 solid state, owing to the difficulty of tapping its semi-liquid 
 mass. 
 
 The equations representing the two stages of the reaction 
 
 are 6F + V 2 O 3 + 3C=2VF 3 + 300 
 
 and 
 
 213 
 
ELECTRIC FURNACES AND 
 
 SECTION IX 
 
 MISCELLANEOUS ELECTRIC FURNACE PROCESSES 
 
 " Alundum" as distinguished from corundum is an 
 electric furnace product, the process, invented by Jacobs, 
 being exploited by the Norton Emery Wheel Company, 
 of Niagara Falls, who utilize some 500 H.P. in its manu- 
 facture, the daily yield being from 4 to 5 tons. 
 
 The process consists in thoroughly calcining bauxite in 
 ordinary furnaces, after which it is fused in an electric furnace 
 of the arc type, and crystallizes, on cooling, as alundum. 
 
 Baryta is another electric furnace product. It is manu- 
 factured at Niagara, from barytes, by the United Barium 
 Company ; a suitable reducing agent, e.g., carbon, is added 
 to the barytes, and the mixture subjected to electric heat 
 in 500 k.w. furnaces, from which the product, a mixture of 
 oxide and sulphide of barium, is tapped off. 
 
 The equations representing the reactions which take 
 place are as follows 
 
 4BaSO 4 + 4C=BaS + 3BaSO 4 + 4CO. 
 
 The barium sulphide then reacts with the sulphate, 
 forming anhydrous baryta, thus 
 
 BaS + 3BaS0 4 =4BaO + 4S0 2 . 
 
 The mixed product is separated by aqueous solution, the 
 baryta crystallizing out as barium hydrate Ba(OH) 2 +8 H 2 0, 
 whilst the sulphide is a by-product. 
 
 The charge consists of a mixture of ground barytes and 
 coke, mixed in the proportions of four molecular equiva- 
 lents of each. When heated in the electric furnace, a 
 
 214 
 
UNIVERSITY I 
 
 
 THEIR INDUSTRIAL APPLICATIONS 
 
 reaction occurs between the carbon and barium sulphate, 
 according to the above equation, 25 per cent, of the latter 
 being reduced to barium sulphide. There is thus created 
 a mixture of approximately three molecular equivalents 
 of barium sulphate with one of the sulphide, and a secondary 
 reaction occurs, in which barium hydrate is formed, only 
 2 to 3 per cent, of unconverted sulphate remaining, and 
 a small quantity of the sulphide. 
 
 The maximum output is 8 tons of product per 24 
 hours, which is tapped off periodically, and dissolved 
 in hot water. The insoluble impurities are removed by 
 filtration, and the barium hydrate crystallized by cooling, 
 whilst the sulph-hydrate remains as a by-product. 
 
 The crystallized barium hydrate is washed with a cold 
 water spray, and subsequently dried and fused, in which 
 latter condition it is poured into iron drums, and solidifies, 
 on cooling, into masses of approximately 250 kgs. apiece, 
 containing less than 1 per cent, of impurities. 
 
 The principal applications of barium hydrate are in the 
 sugar industry, water purification, tanning, and the manu- 
 facture of pigments. 
 
 The plant of the United Barium Company at Niagara 
 has a capacity of 60 tons of the hydrate. 
 
 Barium Cyanide is another commercial electric furnace 
 product. It is manufactured by the Cyanide Company at 
 Niagara Falls. As a raw material, barium carbide, mixed 
 with coke, is used, and subjected to heat in the electric 
 furnace, whilst, at the same time, a current of producer 
 gas is passed through the mixture, with the result that 
 the necessary combination with nitrogen is secured. 
 
 Calcium Cyanide. A patent has been taken out by a 
 German Company, on an electric furnace process for the 
 production of * calcium cyanide. A mixture of lime and 
 coke is raised to a temperature of 2,000C.= 3,632F., 
 whilst nitrogen gas, or air, is passed over the mixture. The 
 necessary proportion of nitrogen enters into chemical com- 
 
 215 
 
ELECTRIC FURNACES AND 
 
 bination with these ingredients to form the cyanide, which 
 is stated to be of value as a fertiliser. 
 
 An electric furnace process, for the casting of articles from 
 refractory materials, e.g., fire-bricks, tiling, conduits, etc., 
 has been devised by Mr. Chas. B. Jacobs. It consists in 
 subjecting acid silicates, such, for example, as the compound 
 represented by the formula A1 2 3 . 2Si0 2 , to continued fusion 
 in any ordinary electric furnace. As a result, a certain pro- 
 portion of the silica is volatilized, and the residue, a tough 
 mass, of waxy appearance, having distinct physical properties, 
 may be run into moulds of any desired form. 
 
 As an example of the energy necessary to effect the fusion, 
 it is stated that a current of 1,500 amperes at 100 volts 
 will suffice to melt 90 kgs. of the charge in 20 minutes, the 
 continuation of the fusion for another 40 minutes being 
 sufficient to volatilise 50 per cent, of the silica. 
 
 Silicides. The silicides of the alkaline earth metals, 
 calcium silicide (CaSi 2 ), barium silicide (BaSi 2 ), and 
 strontium silicide (SrSi 2 ), were discovered by the Ampere 
 Electro-Chemical Company, in July 1899. They are the 
 silicon analogues of the alkaline earth carbides, and are 
 manufactured in the electric furnace under similar conditions, 
 though at a rather higher temperature than is called for hi 
 the formation of the carbides. 
 
 The raw materials consist of carbonates, oxides, sul- 
 phates, or phosphates of the alkaline earth metals, mixed 
 with silica and carbon, in the requisite proportions to bring 
 about the reduction ; or, as an alternative mixture, silicates 
 of calcium, barium, and strontium, mixed with sufficient 
 carbon to combine with the oxygen present in the compounds. 
 
 The resultant silicides, according to Jacobs, are white, 
 or bluish white, of metallic appearance, and with a distinct 
 crystalline structure. They oxidise slowly in air, at normal 
 temperatures, and more rapidly when heated, yielding silicon 
 dioxide, and the oxide of the particular alkaline earth metal 
 content. 
 
 216 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Like the carbides, they decompose when brought into 
 contact with water, evolving hydrogen gas, according to the 
 equation 
 
 BaSi 2 + 6H 2 0=Ba(OH) 2 -2Si0 2 +10H. 
 
 The following are the quantities of hydrogen gas, which 
 are obtainable from one pound of chemically pure silicides, 
 on treatment with water 
 
 One pound calcium silicide . . . 18-73 cubic feet 
 
 ,, ,, strontium silicide . . 12-36 ,, ,, 
 
 ,, ,, barium silicide . . . 9-15 ,, ,, 
 at 0C. = 32F., and 760mm. pressure. 
 
 As a hydrogen producer, barium silicide is stated to be 
 the cheapest and most convenient source, it being only 
 necessary to introduce the material into an ordinary acety- 
 lene gas generator, and decompose with water in the usual 
 way. 
 
 All the silicides are strong reducing agents, and that of 
 barium has been largely used for the reduction of indigo by 
 making the indigo blue into a thin paste with water, and 
 introducing the barium silicide, which has been previously 
 ground fine ; the solution of indigo white, thus produced, 
 may be applied directly to the fibre of the substance to be 
 dyed. 
 
 If kaolin, or china clay, be subjected to heat in the electric 
 furnace, and, simultaneously, to the action of a current of 
 hydrogen gas, silicon hydride is formed, according to the 
 equation 
 
 Al 2 Si 2 O 7 + 8H = SiH 4 + 2H 2 + Al 2 Si0 6 . 
 
 If continued still further, the reaction results in the re- 
 duction of the aluminium silicate to aluminium oxide, either 
 of which are of value as abrasives. The silicon hydride 
 ignites spontaneously on coming into contact with air, 
 forming silicon dioxide and water vapour. The former, 
 being in a state of extreme comminution, is valuable as a 
 polishing powder for fine metal work. 
 
 Combination Processes. Several inventors in the electric 
 
 217 
 
ELECTRIC FURNACES AND 
 
 furnace field have essayed to combine, efficiently and practi- 
 cally, two processes in one and the same operation, using 
 the same furnace structure, and collecting the products 
 independently. Such combination methods can hardly be 
 regarded as successful in a commercial sense, in that the 
 conditions essential to one process are seldom, if at all, 
 allied, or in any way similar to those governing the successful 
 conduct of another. 
 
 The natural consequences of an attempt to combine two 
 such processes in one and the same operation are therefore 
 impracticable complication of the furnace, and commercial 
 inefficiency of one or both processes. 
 
 An instance of such combination methods is that suggested 
 by Mr. A. Dorsemagen, of Wesel, Germany, for combining 
 carborundum manufacture and the reduction of zinc ores. 
 In order to effect this, silicate of zinc is substituted for the 
 sand of the usual carborundum furnace charge, with the 
 result that carborundum is formed in the usual way, or said 
 to be formed, whilst the metallic zinc, liberated from com- 
 bination with the silica, distils over into a suitable receiver. 
 
 In order to bring about this double reaction, a special 
 form of furnace construction, differing materially from the 
 ordinary open carborundum furnace, is an obvious necessity ; 
 such a furnace must be closed in on all sides, and the 
 material of its walls must be impervious to zinc vapours. 
 It seems hardly likely that the manufacture of carborundum, 
 comparatively inefficient as is the simple procedure at 
 present adopted, would prove commercially feasible 
 if worked in conjunction with the electrical reduction of 
 zinc ore, which latter is, to electro-metallurgists, still in 
 the nature of a partially solved problem. 
 
 Other combined electric furnace processes worthy of 
 mention, are that of Rathenau, for the manufacture of iron 
 silicides by the addition of iron to the charge of a calcium 
 carbide furnace ; this method, which has for its object, 
 the purification of the resultant carbide, has already been 
 
 218 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 described, and cannot be regarded in the light of a commercial 
 failure, in that it effects the desired object, whilst the by- 
 product is useful, and does not involve unnecessary com- 
 plication of the furnaces. 
 
 The same applies to Heibling's simultaneous manufacture 
 of chrome-iron and calcium carbide, which is also a purifi- 
 cation method pure and simple ; the latter, which forms as 
 a slag, depriving the chrome-iron of carbon as an undesirable 
 impurity. 
 
 Desjardins has also proposed the joint production of water 
 glass and phosphorus by treating, in the electric furnace, 
 a mixed charge of sodium phosphate, silica, and carbon. 
 
 Joudrain, Billaudot, Jacobsen, Hilbert, Bradley, Read, 
 and Jacobs, have patented joint processes for the simul- 
 taneous manufacture of calcium carbide and phosphorus. 
 The process, in general, consists in treating calcium phos- 
 phate in the furnace with excess of carbon, the latter reacting 
 with the calcium to form carbide, and the liberated phos- 
 phorus distilling over. 
 
 The disadvantages incidental to such a combination are, 
 that in addition to complicating the carbide furnace construc- 
 tion, calcium phosphide is formed conjointly with the carbide, 
 and any acetylene gas subsequently generated from the 
 latter is contaminated with phosphoretted hydrogen. 
 
 The last three inventors enumerated above recognize 
 this fact, and claim, in their patent specification, the manu- 
 facture of calcium " carbophosphide," a mixture of carbide 
 and phosphide of calcium in such proportions that, when 
 reacted upon by water, a spontaneously inflammable mix- 
 ture of gases is produced. This compound is especially 
 suitable for the " marine light," a modern auxiliary in 
 naval and military operations, and is therefore in demand. 
 
 M. Gustav Gin, whose name is associated with modern 
 research into the possibilities of electric smelting, has de- 
 vised a joint process for the simultaneous production of iron 
 alloys, and alkaline oxides ; e.g., ferro-silicon and baryta. 
 
 219 
 
ELECTRIC FURNACES AND 
 
 The process is carried out in two stages, the first of which 
 does not necessarily entail the use of an electric furnace 
 in that a very high temperature is not required. A mixture 
 of broken quartz, barium sulphate, and charcoal, is sub- 
 jected to moderate heat, with the resultant formation of 
 barium silicate. The latter is then mixed with metallic 
 iron or its oxide, and carbon, and again subjected, this 
 time to a high temperature, in the electric furnace. Ferro- 
 silicon and barium oxide are formed, the latter being col- 
 lected as a sublimate. 
 
 An English patent, dated May 9, 1901, was granted to 
 A. H. Cowles and the British Aluminium Company on a 
 double electric furnace process, in which an alloy, or carbide 
 of aluminium, and metallic sodium, are produced simul- 
 taneously in one operation of the furnace. 
 
 The latter has one wall which is pervious to gases, and 
 forms a partition between the furnace proper and a con- 
 densing chamber, where the metallic sodium is collected. 
 In the furnace is placed a mixture of sodium aluminate 
 and carbon ; under the influence of electrical heat the 
 sodium is set free, and passes, as a vapour, through the 
 porous wall into the condensing chamber, where it is duly 
 condensed and collected as metallic sodium The aluminium 
 combines with the carbon to form aluminium carbide, or, 
 if a non- volatile metal be present, unites with it to form 
 an alloy. An alloy of sodium may be similarly produced 
 by including a volatile metal among the constituents of 
 the furnace charge. 
 
 220 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION X 
 
 LABORATORY FURNACES AND EXPERIMENTAL RESEARCH 
 
 Some interesting and instructive laboratory furnaces for 
 experimental use have been designed for, and are in use 
 at, Owens College, Manchester. 
 
 They were aptly described and illustrated in a paper 
 read by Prof. R. S. Hutton before a meeting of the Institu- 
 tion of Electrical Engineers, November 25, 1902. 
 
 A 40-kilowatt Moissan arc 
 furnace, which is typical of arc 
 furnaces with indirect or re- 
 flected heat, is shown in sec- 
 tion in Pig. 48. It consists of 
 two blocks, A and B, of Monk's 
 Park bath stone, one, A, form- 
 ing the cover, or domed lid, 
 and the other, B, the body of 
 the furnace. 
 
 These two blocks are grooved 
 semi-circularly where they fit 
 together, in order to form a 
 circular channel for the recep- 
 tion of the carbon electrodes E E, which pass hori- 
 zontally through the walls, and meet at the centre, with the 
 exception of an arcing space, which is situated just above 
 the centre of the recess or hearth in the lower block B. 
 This latter is from 2J to 4 in. in diameter, and in it is 
 placed a carbon crucible C, resting on a layer of powdered 
 magnesia, which isolates it from contact with block B. 
 A clearance space is also left around the crucible as shown ; 
 
 221 
 
 FIG. 48. 
 
ELECTRIC FURNACES AND 
 
 this facilitates the heating, and also prevents combination 
 between the carbon of the crucible and the limestone, 
 which might otherwise take place under the influence of 
 the intense heat generated. F F are iron bands, which 
 serve to hold the blocks together, and prevent them from 
 splitting, whilst, at the same time, they afford a ready 
 
 means of lifting either the 
 entire furnace, or its cover, 
 by means of attached eye- 
 bolts. The remaining di- 
 mensions of the furnace are 
 indicated in the sketch. 
 
 Another form of experi- 
 mental arc furnace in use 
 at Owens College comprises 
 a vertical construction, 
 which, with some slight 
 and easily effected modifi- 
 cations, can be adapted to 
 a variety of purposes. It 
 is represented in Fig. 49, 
 and consists of a rect- 
 angular cast-iron base, B, 
 some 19 in. long, by 14 in. 
 wide, strengthened by cross 
 ribs cast on to it belqw, 
 FIG. 49. an d provided at one corner 
 
 with a levelling screw, and 
 
 at the other with a |-in. clamping bolt, to serve as a ter- 
 minal connexion. 
 
 Rising vertically from the centre of one of the lesser sides 
 of this base, is a hollow cast-iron standard, S, 2J in. in 
 diameter, flanged at its lower extremity, and bolted to the 
 base B, a sheet of vulcanized fibre F being interposed for 
 the purpose of insulation. A steel plunger P, IJin. in 
 diameter, fits inside this standard, and is capable of vertical 
 
 222 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 adjustment therein, being firmly secured in any desired 
 position by the hand wheels and set screws W W Clamped 
 at right angles to the upper extremity of this plunger is a 
 cross bar or bracket A, which carries, at its outer end, the 
 screw feed adjustment H, and clamp K, of the upper 
 carbon D. Means for two terminal connexions to the 
 pillar and carbon holder are provided, that shown at a 
 being employed in cases where the current exceeds 600 
 amperes. 
 
 The carbon holder K is fitted with a set of clamping 
 collars of varying diameter, which will accommodate car- 
 bons up to 3 in. ; it is carried by a phosphor bronze stem 
 T, 1 J in. in diameter, and square threaded (four to the inch) 
 for feed adjustment. About 12 in. vertical adjustment 
 is obtainable through this construction, and is effected 
 by means of the hand wheel J. The electric circuit to the 
 upper carbon is not completed through the screw threads 
 and bushing, but is provided by four \ in. flexible cables 
 c c clamped to a lug L cast on the bracket, and to the 
 upper extremity of the phosphor-bronze stem. 
 
 The hearth or crucible may be varied to suit the opera- 
 tion it is required to perform, and is supported on, or clamped 
 to the cast-iron base B, with its centre immediately in line 
 with the axis of the upper carbon electrode. 
 
 A 40-kilowatt experimental carbide furnace, with 
 parallel electrodes is also described by Prof. Hutton. It 
 is a very simple affair, as the diagram, Fig. 50, will indi- 
 cate. E E are the two parallel electrodes clamped 
 side by side, but insulated from one another in a cross-bar 
 or yoke B. Electrical connexion with the carbons is 
 secured by flat copper straps c c, which also serve as 
 supports for the carbons and yoke piece, and are slung, by 
 a second upper yoke, to a crane C. They slide in recesses 
 provided for them in the bracket D, and are thus pre- 
 vented from swaying. The furnace hearth is either built 
 up of fire-brick, or takes the form of a cast-iron pot, the 
 
 223 
 
ELECTRIC FURNACES AND 
 
 FIG. 50. 
 
 material under treatment 
 providing a sufficiently re- 
 fractory lining. 
 
 A novel feature of the 
 laboratory equipment at 
 Owens College, consists in a 
 high pressure electric furnace, 
 which can be adapted to 
 any of the many and varied 
 electric furnace processes, and 
 is designed to operate under 
 a pressure of 200 atmospheres. 
 It consists of a massive steel 
 cylinder which can be used 
 either in a horizontal or vertical position. 
 
 Screw gearing, actuated through high pressure glands, 
 serves to feed the electrodes together inside, whilst two 
 small windows at opposite ends of a diameter are constructed 
 to withstand the maximum pressure, and permit inspection 
 of the interior during the progress of an experiment, or 
 may be replaced, if so desired, by gas inlet and outlet pipes, 
 where it is desired to conduct an operation in an atmosphere 
 of any given gas. The whole cylinder is water- jacketed, 
 the actual hearth of the furnace consisting of a refrac- 
 tory lining contained within, and supported by a cast- 
 iron enclosure, between which and the outer steel 
 cylinder the water circulation takes place. 
 
 Suitable valves and pressure gauge connexions are also 
 provided. 
 
 The furnace was constructed by Messrs. Lennox, Reynolds, 
 and Fyfe, to designs provided by the College authorities. 
 
 The Physical Chemistry department of the McGill 
 University, Montreal, is equipped with two electric 
 furnaces. They have a current capacity of 100 amperes 
 at 110 volts, and are principally employed in the analysis 
 of refractory compounds. 
 
 224 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 A convenient resistance, crucible furnace, for the deter- 
 mination of the cooling curves of steel, and of melting points, 
 has been devised by H. M. Howe, Ph.D., for the School 
 of Mines, Columbia University. 
 
 It is represented in sectional elevation by Fig. 51, and 
 consists of a cylindrical structure M of magnesia, made 
 in two semi-cylindric halves in order to permit ready access 
 to the interior, in which is placed the magnesia crucible C. 
 Surrounding this crucible, and embedded in a groove in the 
 magnesia blocks M, specially prepared for its reception, 
 is a spiral platinum coil P, of 
 U-shaped section, the edges of 
 the U fitting closely against the 
 outer wall of the crucible. The 
 extremities of this heating spiral 
 are led out of the furnace at 
 a and b for connexion to 
 suitable terminals, whilst an 
 inlet pipe p provides for the 
 introduction of any desired gas 
 into the interior of C. A 
 magnesia lid L is fitted over 
 the top, and luted on with fire- 
 clay. S is a removable stop- 
 per, making a fairly close joint 
 
 with the lid, and serving for the introduction, through a 
 central vertical bore, of the insulated and protected twin 
 leads I of the thermo-couple T, which, situated in the 
 centre of the crucible, and, consequently, of the heat zone, 
 serves to indicate the temperature. 
 
 Small leakages through the various points of entry of 
 the platinum wires and thermo-couple connexions are 
 counterbalanced, in the main, by packing, and also by 
 maintaining a suitable pressure of gas at the inlet pipe p. 
 
 A special taper mandrel, grooved to correspond with the 
 interior groove in the magnesia blocks M, is utilized for 
 
 225 Q 
 
ELECTRIC FURNACES AND 
 
 the introduction of the spiral, the two halves of M being 
 brought together so as to encircle it, and the mandrel itself 
 subsequently withdrawn by a left-hand screw motion, 
 leaving the wire in its place in the groove. 
 
 A simple and convenient form of electric oven, or furnace 
 for dentists and laboratory use, is the invention of Mr. 
 August Eimer. It takes the form of a rectangular base 
 plate and semi-cylindrically domed cover, both of re- 
 fractory material. One end is permanently closed by a 
 semi-circular plate, which forms the back of the furnace, 
 whilst the front, which is removable, is vertically mounted 
 at the end of a sliding plate, which covers the furnace hearth, 
 and serves as a container or tray, for the reception of the 
 articles to be heated ; it is removed, bodily, with the front 
 door of the furnace, in the centre of which is an inspection 
 aperture. 
 
 The heating elements consist of wires threaded back- 
 wards and forwards, at regular intervals, through longitu- 
 dinal holes in the thickness of the domed cover. A sliding 
 base plate, fitting inside, and constituting the hearth of the 
 oven, also carries similar wires embedded in it. 
 
 The Hammond dental muffle, invented by Dr. J. F. 
 Hammond, is not based on any new fundamental principle, 
 its novelty lying more in the elaboration of detail, and pro- 
 vision for interchangeability of parts, and consequent free- 
 dom from serious breakdown. 
 
 It is heated on the resistance principle, the heater being 
 a platinum wire coil, embedded in, and protected by, a fire- 
 clay foundation. The heat developed is retained by a 
 heat-conserving jacket, consisting of alternate layers of 
 sand, fire-clay, and metal, arranged concentrically round 
 the central hearth. A mica door is provided for convenience 
 in the introduction of the objects to be heated, and also 
 provides a ready means of watching the process, whilst the 
 whole is mounted on a convenient base with the necessary 
 controlling resistance and switch. 
 
 226 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The principle novelty consists in the arrangement of the 
 resistance coil itself, which is provided, at regular and fre- 
 quent intervals with loops, projecting outwards through 
 the walls of the furnace. 
 
 These loops permit the necessary expansion and contraction 
 of the wire, due to heating and cooling, whilst any incidental 
 interruption in the circuit is readily located and repaired. 
 
 A resistance muffle, invented by J. Weiss, presents some 
 points of similarity to the laboratory crucible of H. M. Howe, 
 described in another paragraph. It has, for its object, 
 the rendering of all parts, such as are subject to wear and 
 tear, readily accessible for repair and replacement. To this 
 end, the muffle itself, which is tubular in form, with the 
 platinum resistance wire wound in a groove around its 
 outer periphery, is encased, bodily, in a foundation block, 
 which is divided, vertically, into two halves, along a plane 
 corresponding with the axis of the muffle. The latter is 
 flanged at its two extremities, the flanges fitting into cor- 
 responding recesses in the supporting block, whilst a recess 
 in each block, somewhat larger than the outer diameter 
 of the muffle, provides an air-space, and prevents undue 
 heating of the casing, such as would otherwise result from 
 direct contact with the platinum resistance. 
 
 The extremities of the latter are led out to terminals 
 on the upper surface of one of the two casing blocks, and 
 the whole, when mounted up and ready for operation, is 
 held together by a suitable frame or strap, furnished with a 
 set screw. 
 
 In the form of dental muffle, or laboratory resistance 
 furnace, devised by R. Winter, a considerable reduction 
 in the size of the heating units, and correspondingly lessened 
 liability to fracture under extreme variation of tempera- 
 ture, are secured by embedding the resistance wires in 
 porcelain tubes, and supporting the latter in horizontal 
 grooves in opposite interior walls of the furnace, even 
 spacing being secured by lateral projections. 
 
 227 
 
ELECTRIC FURNACES AND 
 
 The electric furnace in miniature has been adapted to 
 the requirements of microscopical research. The apparatus 
 was described by Prof. C. Doelter to the Vienna Academy 
 of Sciences. An electric oven, 2 in. high, is mounted 
 on the object stand of the microscope, and yields tempera- 
 tures up to 1,200C.=2,192F. In use, the lens is separated 
 from the heated object by about 1 in. Even at the 
 highest temperature of the subject under examination, 
 however, both microscope and objective are kept quite 
 cool by a special arrangement of asbestos plates, and a 
 spiral tube carrying ice-cold water. 
 
 A. Kalahne has experimented with various stoves of 
 the embedded resistance type, with a view to determining 
 their limitations, and energy consumption. The results 
 of his researches were published in Ann. der Physik, No. 6, 
 1903, in the form of graphic curves, of which the verti- 
 cal ordinates represented watts consumed, and the hori- 
 zontal ordinates, temperatures attained, in degrees Centi- 
 grade. 
 
 Six different furnaces were tested, of which two were 
 provided with double coils, whilst the remaining four had 
 only one coil apiece. As regards the foundations, or bodies 
 of the furnaces, which were in tube form, five were of porce- 
 lain, 40 to 70 c.m. long, three being glazed interiorly, but 
 not exteriorly. A sixth tube, 20*2 c.m. in length, was of 
 Marquardt composition. 
 
 Kalahne found asbestos a preferable material to fire- 
 clay for covering in the resistance wires, which were of 
 nickel. If fire-clay be employed, its different coefficient 
 of expansion leads to cutting or tearing of the resistance 
 wire at high temperatures. Asbestos possesses the ad- 
 ditional recommendation that it can be easily renewed. 
 
 Application of the Electric Furnace to Scientific Research. 
 By no means the least important of the many uses to which 
 the electric furnace has been applied is that of an auxiliary 
 to scientific research, especially among such refractory 
 
 228 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 substances as the diamond, and metals with exceedingly 
 high melting points. 
 
 As an example of what can be done with the aid of electric 
 heat, tantalum, which has hitherto been known only as 
 a somewhat impure powder, having a density of 10' 50, 
 has now been produced from tantalic acid in the electric 
 furnace by Henri Moissan. The acid was reduced to metallic 
 tantalum by heating electrically in the presence of sugar 
 carbon as a reducing agent, and the product, which shows 
 a density of 12' 7 9, has a brilliant metallic lustre. It 
 scratches quartz, is infusible in the oxy-hydrogen blow- 
 pipe flame, has a crystalline fracture, and is classed with 
 the metalloids. 
 
 At the Fifth International Congress of Applied Chemistry, 
 held in Berlin, W. Hempel described some experiments 
 on melting-point determinations at high temperatures, 
 which had been carried out with the aid of an electric 
 furnace built up of carbon rods surrounded by Kieselguhr. 
 With this apparatus the melting-point of magnesia was 
 found to be in the neighbourhood of 2,250C.= 4,082F., 
 and of lime, 1,900C. 3,452F. 
 
 With the Heraeus platinum-foil tube furnace the melting- 
 point of 99 per cent, manganese was investigated and 
 found to be 1,245C.= 2,273F. 
 
 The Heraeus platinum-foil tube furnace, and the new 
 quartz tubes, supplied by Heraeus, of Hanau, and by Siebert 
 and Kuhn, of Cassel, have provided means for the conduct 
 of numerous experiments on the volatilization of metals, 
 hitherto impossible with glass tubes. 
 
 The advantages of quartz for this purpose are 
 
 1. That it will stand a very high temperature without 
 softening, and vessels made of it can consequently be 
 exhausted of air, even when simultaneously subject to 
 extremely high temperatures. 
 
 2. One portion of a quartz vessel may be cooled whilst the 
 other is glowing, thus rendering it possible to evaporate 
 
 229 
 
ELECTRIC FURNACES AND 
 
 metals in the hotter region of the tube, and condense them 
 in the cooler portion, without risk of fracture, whilst gas- 
 tight connexions with the cooler portions can be made 
 with the aid of a mixture of two parts wax and one part 
 wool grease. 
 
 Metals, per se, do not attack quartz, but their oxides 
 do, so that a certain amount of care is essential, especially 
 in dealing with lead. 
 
 The quartz vessels are L-shaped, and the metal is, as 
 a rule, placed in a bulb, blown to the top of the L, and 
 bent back. 
 
 With this apparatus Drs. F. Krafft, Kuch, and Haa-gn, 
 of Hanau, have conducted a series of experiments to deter- 
 mine the boiling, volatilization, and distillation tempera- 
 tures of various metals. The full text of their researches 
 is published in the Berichte der Deutschen Chemischen Gesell- 
 schaft, vol. XXXVI. The vacuum employed is described 
 as one suitable for cathode glows, whilst the temperatures 
 were determined by the aid of platinum platinum-rhodium 
 thermo-couples. 
 
 The following are a few of the more interesting results 
 obtained by these investigators : Zinc began to distil dis- 
 tinctly at 430C.= 806F. without melting, and distillation 
 afterwards continued at a temperature of 300C. = 572F. 
 The boiling zinc displayed the Leiden-frost phenomenon, 
 which was also observed in the case of cadmium. The 
 latter began to evaporate at 320C.= 608F., and dis- 
 tilled visibly at 448C. = 838F. Selenium was easily 
 volatilized at 380C.= 716F. TeUurium distilled and 
 boiled briskly at 540C. = 1,004F. ; lead volatilized 
 sufficiently at 801C.=: 1,472F., to form a mirror which 
 could be melted, and at 1,160C.= 2,120F. it boiled 
 regularly and distilled. Antimony was volatilized at 
 670C.= 1,238F., and distilled at 780C.= 1,436F. ; 
 bismuth volatilized at 540C.=: 1,004F., and distilled 
 at 1,140C.= 2,084F. ; tin could not be volatilized at 
 
 230 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 1,100 C.= 2,012F., however. Silver began to volatilize 
 at 1,200C.:=2,192 F. ; at 1,340C.= 2,444F. it distilled 
 at the rate of 0'09 grammes in 12 minutes. Copper began 
 to distil at 1,315C.= 2,399F. Gold was the most refrac- 
 tory metal experimented with ; it formed a mobile liquid 
 at 1,180C.= 2,156F. ; at 1,300C.= 2,372F. some silver 
 distilled over from the impure gold, whilst at 1,375C. = 
 2,507F. sufficient gold had distilled to produce a gold 
 mirror. 
 
 The high temperature researches carried out by Moissan 
 with the aid of the electric furnace are detailed in his book 
 Le Four Electrique, to which the reader is referred for 
 particulars of his experiments. He is essentially a scien- 
 tific enthusiast, and pursued his investigations with a view 
 of adding to the scientific knowledge of the day. His 
 work, which is mainly of a laboratory and experimental 
 character, has nevertheless been instrumental in suggesting 
 several of the industrial electric furnace processes which 
 are in operation to-day. 
 
 With the aid of furnaces devised by himself, Moissan 
 has succeeded in fusing, distilling, and crystallizing many 
 of the most refractory substances, among which may be 
 mentioned magnesia, lime, platinum, and carbon. 
 
 In connexion with the last-named element he has con- 
 ducted extensive investigations into the properties of its 
 allotropic modifications, and last, but not least, has actually 
 succeeded in manufacturing diamond. 
 
 His researches in the preparation of carbides, borides, 
 and silicides, resulted in quite a quantity of hitherto un- 
 known compounds, among which may be mentioned the 
 now familiar calcium carbide and silicon carbide (car- 
 borundum) of commerce. 
 
 Moissan originally described the carbides of calcium, 
 strontium, and barium as opaque, except when obtained 
 in very thin sheets. 
 
 He has since modified this description in accordance 
 
 231 
 
ELECTRIC FURNACES AND 
 
 with the results of later experiments (Comptes Eendus, 
 December 5, 1898). He ascribes the opacity of calcium 
 carbide to impurities, principally iron, which are present 
 in the mass. When obtained in an absolutely pure 
 state it is as transparent as lithium carbide or sodium 
 chloride. 
 
 That this is so may be demonstrated by a simple ex- 
 periment. A mixture of metallic calcium and amorphous 
 carbon, obtained by the rapid combustion of acetylene, 
 is heated to a dull red, yielding a white product, pure 
 calcium carbide, which, under the microscope, proves to 
 be a collection of transparent crystals. 
 
 In this connexion Moissan further points out that, under 
 these conditions, the heat of combination between the 
 carbon and the calcium is sufficient to fuse the carbide, 
 a result never yet achieved in commercial practice. 
 
 The reddish-brown tint of commercial carbide may be 
 reproduced in this pure white mass by heating it with a 
 small quantity of iron. 
 
 The carbides of potassium, K 2 C 2 , and lithium, Li 2 C 2 
 may also be obtained in transparent crystalline plates, 
 whilst aluminium carbide A1 4 C 3 sometimes crystallizes 
 in transparent plates, having a yellowish tinge. 
 
 MM. Moissan and Kouznetzow have succeeded in pre- 
 paring, by direct reduction in the electric furnace, a double 
 carbide of chromium and tungsten, having the formula, 
 W 2 C.3O 3 C 2 , which is so hard as to scratch glass, topaz 
 and ruby, though not diamond. It is unattacked by 
 acids, either individually or mixed, but is dissolved by 
 fused chlorate and alkalies. 
 
 Messrs. Tucker and Moody have succeeded in producing 
 the borides of zirconium, chromium, tungsten, and 
 molybdenum, by heating these metals in conjunction with 
 borax, in the electric furnace. The current required was 
 from 200 to 275 amperes. Attempts were made by the 
 same experimenters to prepare borides of copper and 
 
 232 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 bismuth in a similar manner, but were not attended with 
 success. 
 
 Bolton has succeeded in bringing about a direct union 
 between carbon and chlorine by striking an arc between 
 two carbon electrodes in an atmosphere of the latter gas. 
 
 Hutton has conducted a series of experiments on the 
 fusion of quartz, and the results of his efforts were em- 
 bodied in the form of an article in the Electro-Chemist and 
 Metallurgist, May, 1902. 
 
 He first employed an open arc to effect the fusion, and 
 found that reduction took place in the immediate neigh- 
 bourhood of the arc, giving rise to a black stain on the 
 surface of the quartz, which, however, disappeared on 
 holding the mass away from the centre of the flame for 
 a short time. 
 
 He subsequently continued his experiments with a closed 
 furnace, using one of the Moissan type, modified by cutting 
 openings in its side walls to permit the passage of a carbon 
 vessel charged with quartz, in a line at right angles to the 
 arc. A current of 300 amperes at 50 volts was employed, 
 and thick- walled quartz tubes, having an internal diameter 
 or bore of about J in., were easily cast in a rough carbon 
 mould, a carbon core at the centre securing the correct 
 position of the bore. 
 
 The charge for the mould consisted of broken quartz, 
 not too finely powdered, and the resultant tubes were 
 easily withdrawn from the carbon mould and freed from 
 the cores. They were not quite free from air-bubbles, 
 but were improved in appearance by subsequent re-heating 
 under the arc, the tube being meanwhile rotated. 
 
 The new quartz tubes for laboratory research at high 
 temperatures have considerably extended the field of 
 investigation. Metals may be boiled and evaporated in 
 them at temperatures up to 1,200C.= 2,192F., and, 
 with care, up to 1,400C. = 2,552F., even in a high vacuum, 
 and the tubes containing metal at the former temperature 
 
 233 
 
ELECTRIC FURNACES AND 
 
 may be safely removed from the furnace, cooled, and 
 replaced, without fear of breakage. Using an electric 
 furnace, it is possible to regulate the temperature, within 
 two or three degrees, between 18 and 1,400C. = 64 
 and 2,552F., whilst the air-pump connexions to the tubes 
 may be hermetically sealed with wax without risk of melt- 
 ing, although within a few inches of the intense heat. 
 
 234 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION XI 
 
 TUBE FURNACES 
 
 Mr. H. N. Potter, of New York, has devoted considerable 
 attention to the perfecting of a series of tube furnaces, 
 in which the hearth takes the form of a refractory con- 
 ducting tube, through the walls of which the current is 
 passed in a longitudinal direction, heating them in its 
 passage. 
 
 There are several difficulties connected with the efficient 
 operation of this type of furnace, among which may be 
 mentioned 
 
 1. The attainment of an even heating effect in all parts 
 of the tube. 
 
 2. Interchangeability of parts, whereby those most 
 subject to wear are rendered easily replaceable. 
 
 3. Satisfactory terminal connexions, which shall allow 
 for the necessary expansion and contraction of the tube 
 itself. 
 
 These are the main difficulties which Mr. Potter has 
 realized, and set himself to overcome, and we will now 
 proceed to consider the various forms of tubular furnace 
 construction advocated by him, and the manner in which 
 they severally contribute towards the desired result. 
 
 The discovery of the Nernst principle has, as already 
 stated, opened up another promising field for electric 
 furnace construction, to which the tubular type readily 
 lends itself ; the several improvements with which we 
 shall now proceed to deal are equally applicable to furnaces 
 on the Nernst principle, and to the older forms of carbon 
 tube. 
 
 Equalization of Heating Effect. Mr. Potter suggests three 
 
 235 
 
f ELECTRIC ^FURNACES AND 
 
 distinct forms of construction for the tube, with a view to 
 securing a uniform distribution of heat to all parts of the 
 latter. His first construction consists of a simple tube, 
 the terminal connexions to either extremity of which are 
 subdivided. Thus, a series of equidistant points to the 
 number of six, or upwards, are marked off around the 
 annular surface at each extremity of the tube ; separate 
 and distinct terminal connexions are made to each of these 
 points, and the current may be derived from a transformer 
 with subdivided secondary winding, the number of sub- 
 divisions being equal to the number of pairs of terminal 
 points to the furnace tube. 
 
 The second type of tube to secure equal heating takes 
 the form of a partially divided cylinder ; longitudinal 
 divisions or grooves, spaced at equal distances apart, 
 around the external periphery of the tube, serve to con- 
 centrate the flow of current along definite lines, repre- 
 sented by the intervening sections of tube. 
 
 The third form of construction consists in a total longi- 
 tudinal subdivision of the tubular walls into a number 
 of similar parts, the intervening spaces being filled in with 
 refractory insulating plates after the manner of a dynamo 
 commutator. 
 
 A further improved construction of tube furnace, designed 
 by Mr. Potter with a view to the additional strengthening 
 of parts, prevention of distortion of the tube under the 
 effects of heat, and conservation of the heat produced, 
 is shown in Fig. 52. It comprises a carbon tube C, 
 which may or may not be provided with an interior re- 
 movable lining T, also tubular ; carbon terminal discs 
 D D, and a series of annular ribs r r of the same material, 
 which are slipped over the tube, and serve as an additional 
 support thereto. The spaces between the ribs r r are 
 filled in with magnesia m, covered by a layer of asbestos, 
 and the whole is enclosed in a glazed earthenware jacket J. 
 As may be readily imagined, this provides a very rigid 
 
 236 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 FIG. 52. 
 
 construction, whilst the 
 loss of heat, resulting 
 from radiation, is com- 
 paratively small. 
 
 Terminal Connexion 
 and Mounting of Tubu- 
 lar Furnaces. Mr. Pot- 
 ter has also tackled this 
 problem, and the resul- 
 
 tant design is that of a form of terminal connexion and mount- 
 ing for tubular furnaces, which enables the latter to be placed 
 in any position, and allows for the necessary expansion 
 without resultant disturbance of the electrical connexions. 
 It is represented in Fig. 53, where C is the carbon 
 tube, as before ; D D terminal discs, or rings of the same 
 material, the outer surfaces of which are double-tapered 
 as shown, to form a seating for the metal clamping rings 
 r r, held together, and in good electrical contact with the 
 carbon, by the bolts & & ; A A are sheet metal diaphragms, 
 through which the bolts 6 also pass, thus securing them 
 to the terminal rings of the tube. The outer edges of the 
 diaphragms are clamped between metal end plates p p, 
 
 and the flanged 
 extremities of 
 an outer casing 
 B, suitable in- 
 sulation being 
 inserted at the 
 joint. The 
 flexibility of the 
 metallic dia- 
 phragms A per- 
 mits^the neces- 
 sary expansion 
 of the tube 
 under heat, 
 237 
 
 - 53 - 
 
ELECTRIC FURNACES AND 
 
 whilst the actual electrical connexions from the source of 
 energy are made to the end plates p, and are thus ren- 
 dered independent of any movement in the tube itself. 
 
 Another departure in tube furnace construction consists 
 in the provision of an additional source of heat within the 
 tube itself ; this takes the form of an arc between two 
 carbon electrodes, the position of which is adjustable 
 longitudinally along the axis of the tube, so as to bring 
 the arc to any desired point. 
 
 A revolving tube furnace, boasting the above com- 
 bination of arc and resistance principles, is represented 
 in Fig. 54. The tube proper consists of separate graphite 
 sections g g, pieced together to form a complete cylinder, 
 and surrounded by an insulating jacket j, which latter 
 is, in turn, enclosed in a sheet metal casing c. Around 
 the periphery of the casing are arranged two annular metal 
 
 FIG. 54. 
 
 bearing pieces b b, which serve as journals to the tube, 
 being mounted on rollers r r. The necessary motive power 
 to rotate the furnace is supplied through a pulley P, also 
 mounted on the outer casing. E E are fixed end pieces, 
 or covers, lined with graphite blocks a a which make 
 rubbing contact with the tube sections g, outer insulation 
 i, and metallic casing ra m. These latter form the ter- 
 minal attachments for the resistance portion of the furnace. 
 The arc electrodes A, A2 are introduced through in- 
 
 238 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 sulating bushes in the centre of the end covers, E E and, 
 as stated above, are adjustable as regards their axial position 
 in the furnace. I and are inlet and outlet orifices re- 
 spectively for the introduction of the raw material, and 
 withdrawal of the product. They penetrate the stationary 
 end covers, as shown, and, the tube being preferably 
 mounted at an angle, the process is rendered continuous, 
 the charge passing through by virtue of rotation and 
 gravitation from one end of the tube to the other. 
 
 Mr. Potter's design for a tube furnace on the Nernst 
 principle, especially adapted to the baking of " glowers " 
 for Nernst lamps, comprises a simple tube, formed from 
 a mixture of the electrolytic oxides of magnesia, or zirconia, 
 and yttria, fitted with terminal connexions, and enclosed 
 in, and strengthened by an encircling jacket, consisting 
 of one only of the oxides named above. 
 
 Over this again is wound the heater coil, a layer of mica 
 being interposed between them. The actual heat is 
 developed, once the action is started by the coil, in the 
 inner tube of mixed oxides, the outer jacket remaining, 
 to all intents and purposes, a non-conductor, whilst, at 
 the same time, its chemical similarity to the furnace tube 
 proper prevents any detrimental action from taking place 
 between them. 
 
 This notion of a single oxide for a protective and sup- 
 porting jacket is one of the essentials of Potter's patent, 
 any other refractory substance being unsuitable on account 
 either of its electrolytic properties when heated, or of its 
 tendency to unite chemically with the substance of the 
 furnace tube proper. 
 
 A resistance tube furnace on the Nernst principle, devised 
 by Drs. Nernst and Glaser, is illustrated in Fig. 55. The 
 resistance heater is in the form of a hollow cylinder, or 
 tube R, composed of the usual electrolytic oxides, pre- 
 ferably a mixture of magnesia, with traces of calcium 
 carbonate, silica or kaolin, and alumina. It is, like one 
 
 239 
 
ELECTRIC FURNACES AND 
 
 of the tube furnaces already described, enclosed in a second 
 cylinder C, of loose oxides, held in place by an outer casing 
 J, which serves to retain the heat. The electrodes are 
 an especial feature of the construction, and consist of 
 oxide of iron F packed closely round the two ends of the 
 tube in annular form, which latter is preserved, and the 
 oxide held in position by iron hoops /, secured and drawn 
 
 together by bolts and nuts. 
 A refractory lid L is pro- 
 vided, and the furnace stands 
 on a base B. The raw 
 material to be treated therein 
 may either be placed directly 
 in the tube, in which case it 
 is lined interiorly with mag- 
 nesia, or in a separate cruc- 
 ible c, supported by a re- 
 movable portion of the base. 
 The action of the furnace is 
 started by a preliminary heat- 
 ing of the tube, either with 
 
 a gas flame, as in torch-actuated Nernst lamps, or by a 
 carbon resistance rod introduced axially through the tube 
 and heated by the passage of the current. 
 
 The Eddy tube furnace, invented and patented by A. 
 H. Eddy, was specially designed for the fusion of enamels 
 upon earthen and metallic ware, a process calling for ex- 
 tremely fine temperature regulation, and freedom from 
 ash and combustion vapours. It consists of a series of 
 tubes, composed of such electrolytic oxides as conduct 
 when heated. Axially within these tubes, but not in 
 electrical contact with their inner walls, are arranged 
 carbon rods, which, heated by the current, serve for the 
 preliminary heating of the tubes themselves. 
 
 The combinations of tubes and rods are mounted in 
 batches, of varying number, between water-cooled ter- 
 
 240 
 
 FIG. 55. 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 minal blocks. The furnace itself is constituted by arranging 
 a number of these combinations transversely across a long, 
 narrow hearth, over which, and beneath the tubes, are 
 packed the articles to be heated. 
 
 Varying gradations of temperature at different points 
 along the length of the furnace are secured by connecting 
 the several sections in series, their respective heating 
 capacities being regulated by the number of tubes com- 
 prising them. 
 
 The initial heating of the tubes is effected by the passage of 
 a current through the carbon rods, thus heating the tubes, 
 which, in turn, become conductors, and shunt a portion 
 of the current. The rods may subsequently be entirely 
 withdrawn, when the maximum heating effect is obtained 
 within and around the substance of the tubes themselves. 
 
 What is known as an electric combustion furnace, being 
 a combination of an ordinary combustion tube, with 
 electrical means for starting and maintaining high degrees 
 of temperature therein, has been designed by Mr. W. M. 
 Carr. It consists of a combustion tube proper, with copper 
 gauze at 'one extremity and granular oxide of copper 
 between perforated platinum discs at the other, surrounded 
 by an electrical .portion, which serves to produce and 
 maintain the necessary heat. 
 
 It comprises a porcelain cylinder surrounding the com- 
 bustion tube, and having three heating coils wound upon 
 it, one at either extremity, and the other at the centre. 
 A heat conserving jacket of magnesia encloses the whole 
 apparatus. The mode of operation is as follows : The 
 material to be treated is placed in a platinum boat at the 
 centre of the inner combustion tube ; by means of inde- 
 pendent switching arrangements, a current is first passed 
 through the two end coils in series, and raises the tube 
 and its contents to the required temperature, the latter 
 being subsequently maintained, so long as desired, by all 
 three coils connected in series. 
 
 241 R 
 
ELECTRIC FURNACES AND 
 
 The usual means are provided for the introduction of 
 oxygen gas to, and the withdrawal of the products of com- 
 bustion from, the tube. 
 
 A convenient form of laboratory resistance furnace, 
 exhibited by Mr. Otto T. Louis, at the Conversazione of 
 the American Institute of Electrical Engineers, in 1901, 
 has a tubular construction, and consists of several terra 
 cotta cylinders, each about two inches in diameter, and 
 half an inch in thickness. Around these are wound the 
 heating resistance spirals, consisting of platinum wire, 
 about No. 23 S.W.G. Switching facilities are provided, 
 whereby the wire, which is arranged in four distinct circuits, 
 can be connected in series to start the furnace, and then, 
 as the resistance rises with increasing temperature, in 
 parallel to maintain the heat. 
 
 A so-called electric oven, or small resistance furnace 
 of the tube type, is employed in the German Reichsanstalt 
 for the calibration of thermo-couples used in high tem- 
 perature thermometry. 
 
 It consists of four concentric porcelain tubes, the outer 
 being glazed, and jacketed with asbestos. The inner 
 tubes are composed of a very refractory material, devised 
 by Hecht. The second tube serves as a mandrel or base 
 for the heating resistance, which, in the form of wire, 
 is wound upon it. For procuring temperatures up to 
 1,400C.= 2,552F., a pure nickel wire, having a diameter 
 of 2 m.m., is used, and is protected by a layer of fire-clay 
 paste laid on so as to embed the turns. For higher tem- 
 peratures a platino-iridium wire is used, which increases 
 the range to 1,600C.= 2,912F. Above this the wire 
 fuses into the porcelain. 
 
 Temperature regulation is effected with a great degree 
 of accuracy by means of an adjustable rheostat, the sliding 
 contact of which moves over and makes contact with a 
 constantan band. 
 
 The thermo-couples are calibrated in sets by a com- 
 
 242 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 pensation method, being mounted, for the purpose, in small 
 circularly gapped sockets of platino-iridium. The errors 
 do not exceed O'l per cent. 
 
 The power consumed in the furnace is low, being sixty 
 watts for a temperature of 200C. = 392F., and rising 
 to 1,540 watts for 1,300 C.= 2,372F. 
 
 Observations on the melting-point of the element man- 
 ganese were carried out at the beginning of 1902 by Heraeus, 
 of Hanau, in a specially constructed furnace of the tube 
 pattern, devised by this firm. The furnace consisted of 
 a porcelain tube, 16 m.m. in diameter and 30 c.m. long. 
 Around a portion of this tube is wound, spirally, the heating 
 resistance, in the shape of a strip of platinum foil. 
 
 Efficient and gradual regulation of the temperature is 
 secured by means of an adjustable resistance connected 
 in series with the foil. With this arrangement a tempera- 
 ture of 1,400C.= 2,552F. could be attained within three 
 minutes of switching on the current, whilst the maximum 
 temperature available was 1,600C.= 2,912F. 
 
 Addressing a meeting of the German Electro-Chemical 
 Society in the same year, Dr. Haagn, chemist to the firm 
 of Heraeus, expatiated upon the advantages of this new 
 form of laboratory tube furnace construction. 
 
 The heating resistances in previous furnaces of a similar 
 type have been wire, and Dr. Haagn estimates the weight 
 of platinum required when foil is used to be only T ^ to 
 -^Q that necessary in a wire resistance. 
 
 Referring to the tubes, the same authority stated that 
 glass or porcelain was preferable to magnesia, as a safeguard 
 against electrolytic dissociation and consequent injury 
 to the platinum heater in the neighbourhood of the cathode. 
 Above a temperature of 1,500C. = 2,732F., the same 
 effect is observed even with tubes constructed of the best 
 Berlin porcelain, but, with specially manufactured tubes, 
 a temperature as high as 1,700C. 3,092F. has been 
 maintained for a short period without injury. 
 
 243 
 
ELECTRIC FURNACES AND 
 
 Dr. Haagn recommends this type of furnace for use in 
 chemical and physical laboratories, and points out its appli- 
 cability to such operations as ash determinations, organic 
 combustion analysis, the direct determination of carbon 
 in steel by the Lorenz method, etc. 
 
 A subdivision of the special heating resistance, with 
 suitable provision for separate terminal connexions to each 
 section, largely increases the range of utility of this furnace, 
 in that it enables the heating of various sections of the 
 tubular hearth to be carried out independently or together, 
 at the will of the operator. 
 
 The platinum foil employed in the Heraeus tube furnace 
 is 0*007 m.m. thick. The furnace is easily regulated by 
 means of a series resistance. These furnaces can be made 
 for any voltage up to 220, but, for very high temperatures, 
 low voltages are said to answer best. 
 
 A modification of the Heraeus furnace, for heating porce- 
 lain tubes, etc., was described in the Journal de Chimie 
 Physique (1903, I. 177) by M. A. Guntz. The foundation 
 consists of a tube of refractory earths, several millimetres 
 thick ; 60 m.m. longer, and with an internal diameter or bore, 
 from four to five m.m. larger than the external diameter 
 of the tube or article to be heated. This tube is given several 
 coats of a paste made up of magnesia and alumina, mixed 
 either in the proportion of equal molecules of the two 
 oxides, or eight molecules, of magnesia to one of alumina, 
 to which some gelatinous alumina, or a weak solution of 
 aluminium acetate, is added as a binding material. 
 
 An alternative paste is made up of one part strongly 
 ignited alumina ; one part of alumina obtained by incom- 
 plete calcination of ammonium alum, and one part calcium 
 aluminate, prepared by heating to redness, equal weights 
 of alumina and pure chalk. No appreciable trace of silica 
 should be present in the ingredients. 
 
 The layers of paste are dried naturally in air, as the 
 result of which, they adhere well. The dimensions of the 
 
 244 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 platinum wire required to produce a pre-determined heating 
 effect are then calculated from the equation 
 
 Where E represents the maximum available voltage, 
 C current, 
 
 S section of the wire, 
 
 ,, 1 ,, ,, length of the wire, 
 
 and R ,, resistivity of one metre of plati- 
 
 num wire, having a cross-section of one square millimetre, 
 a quantity which requires to be ascertained for the particular 
 sample available. 
 
 Having thus found the length and diameter, a simple 
 calculation will serve to determine the number of turns 
 to be coiled on the tube ; suffice it to say that, unless a 
 specially rapid heating effect be desired, neighbouring 
 turns should not be less than 5 m.m. apart. Having deter- 
 mined the number and pitch of the spirals, a suitable 
 thread or groove, about '5 m.m. deep, is chased in the pre- 
 pared coating of the tube. The platinum wire is laid in 
 place in the groove thus prepared for its reception, the two 
 end turns being double, and tied, or otherwise secured in 
 place. 
 
 A number of coats of the original paste are then applied 
 over all, to a thickness of four or five millimetres, the whole 
 being then encased in a sheet of asbestos. A weak current 
 is first passed through the spiral, warming it sufficiently 
 to drive the moisture out of the paste, which should remain 
 firm. In the event of cracking, the openings mus.t be 
 filled in with paste. A final jacket of asbestos sheet, on 
 which are mounted suitable terminals, completes the con- 
 struction. 
 
 245 
 
ELECTRIC FURNACES AND 
 
 SECTION XII 
 
 TERMINAL CONNEXIONS AND ELECTRODES 
 
 The terminal connexions to the electrodes of electric 
 furnaces have long been a source of trouble to the operator, 
 and an object for the application of the inventor's genius 
 and forethought. There are several important points to 
 be considered in the design of an efficient terminal connexion 
 to carry the large currents met with in electric furnace 
 work. 
 
 (1) They must possess a certain amount of flexibility 
 in order to allow of the necessary adjustments. 
 
 (2) They must establish good electrical connexion with 
 the electrode on the one hand, and with the conducting 
 cables on the other. 
 
 (3) They must not become unduly heated when the 
 furnace is at work. 
 
 (4) They must not be of such a nature as to radiate a 
 considerable percentage of heat and thereby lead to a 
 serious reduction in the efficiency of the furnace. 
 
 Many and varied are the designs which have been evolved 
 from time to time, with a view to meeting all these require- 
 ments in one and the same device, and, although some con- 
 siderable improvement in the means for transmitting the 
 current to the furnace electrodes has undoubtedly been 
 effected during the last few years, there is still a good deal 
 of room for more, and those interested in furnace industries 
 might do worse than turn their attention at spare moments 
 to the improvement of this important accessory, which, if 
 badly designed, is quite capable of wasting several horse- 
 power in a plant of moderate capacity. 
 
 246 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 One of the earlier Cowles' furnaces was provided with a 
 somewhat novel terminal arrangement, which has since 
 fallen into disuse. It consisted of a species of metallic 
 gland or stuffing box, filled with copper shot, through the 
 centre of which passed the electrodes. These latter set 
 up a good rubbing contact with the shot, and, through them, 
 with the metallic sleeves ; a contact which was, in fact, 
 improved by use, in that the motion of the electrodes, in 
 feeding them forward into the furnace, revolved the shot, and 
 brought fresh clean surfaces into contact with one another. 
 In furnaces having a suspended electrode it is frequently 
 arranged that the support for this latter shall also serve 
 to convey the current to it. This construction answers 
 very well for furnaces of moderate size, but is not suitable 
 for larger types, in that the two governing features, con- 
 ductivity and strength, do not increase in like proportion. 
 A form of terminal connexion, devised with a view to 
 overcoming this difficulty, has been patented by Fausto 
 Morani, his apparatus consisting of a supporting frame for 
 the electrode, which is separate and distinct from the electrical 
 connexion thereto. Both attachments are furnished with 
 means for the circulation of water for cooling purposes. 
 
 E. G. Acheson, one of the recognized leaders in the field 
 of electric furnace work, has also attacked the problem, 
 his latest design of terminal for resistance furnaces being 
 capable of dealing efficiently with so large a current as 
 37,500 amperes, and this with a total cross-section of only 
 1,152 square inches, representing an approximate current 
 density of 32*5 amperes to the square inch. 
 
 A transverse section of the terminal is shown in Fig. 56. 
 It comprises 72 square blocks C C of graphitized carbon, 
 each having a cross-sectional area of 16 square inches. 
 These are placed in block form, as close together as possible, 
 and end on to a copper plate A, to which they are clamped 
 by studs S, and nuts N. One of the attachments is shown 
 in section in the figure, and it will be seen that the portion 
 
 247 
 
ELECTRIC FURNACES AND 
 
 FIG. 56. 
 
 of the stud s, which screws into the carbon block C, is 
 
 larger in diameter than 
 the portion projecting 
 through the copper 
 plate, thus furnishing a 
 shoulder which butts 
 up against a raised por- 
 tion of the latter, and 
 takes all strain, arising 
 out of the clamping, off 
 the carbon itself. 
 
 The copper plate A 
 constitutes one wall 
 of a flat water tank, 
 
 he remaining wall D being of sheet metal. The sides 
 and bottom of the tank are formed of flanged pieces 
 E, into circular openings p p, in which, a series of cable 
 ends, forming the main electrical connexions to the elec- 
 trode, are sweated. A tank T, in which a constant level 
 of water is maintained, communicates with the flat tank 
 or base of the electrode by a pipe t, and a constant circula- 
 tion of cooling water is thus secured. 
 
 The following are among the difficulties experienced 
 in designing an efficient and convenient form of terminal 
 connexion for electric furnaces employing carbon electrodes. 
 In the first place, metal has a higher expansion co-efficient 
 than carbon ; consequently the heat conducted from the 
 furnace itself, together with the hot gases produced, and 
 which tend to make their escape, among other places, at the 
 point of entry of the electrodes through the furnace walls, 
 both tend to expand the metal portions to a greater extent 
 than the corresponding parts of the carbon electrodes. 
 A small air-gap is consequently created between the two, 
 and the electrical resistance at the surfaces of contact 
 rises very rapidly. This, in itself, sets up a further heating 
 effect, which may, in extreme cases, be heightened by arcing, 
 
 248 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 set up between two widely separated points, which ulti- 
 mately tends to destroy the connexion entirely from an 
 electrical point of view. 
 
 Chavarria Contardo's patent form of terminal connexion 
 for electric furnaces is both ingenious and convenient, in 
 that it permits the necessary sliding movement of the car- 
 bons, to compensate for combustion and wear, without dis- 
 turbing the electrical connexions thereto. 
 
 Cylindrical carbons are employed, and they slide in 
 water- jacketed sleeves, let into the wall of the furnace 
 structure. Electrical connexion with the sleeve is rendered 
 secure, and a low resistance path provided for the current, 
 by prolongations of the metal sleeves, consisting of a series 
 of bronze segments, which encircle the carbon electrode, 
 and are maintained in intimate contact with it by means of 
 set screws. A thin packing of sheet copper is usually inter- 
 posed between the bronze segments and the surface of the 
 carbon, for the better protection of the former. 
 
 Some saving in space, and length of active circuit, may 
 be effected by employing hollow segments, which are them- 
 selves water-cooled, thus avoiding the necessity for the 
 complete metal sleeve alluded to above. 
 
 The connecting cable is either in multiple, or, if in single 
 form, is split up into a series of strands corresponding with 
 the number of bronze segments, one such strand or cable 
 being connected to each segment by a suitable mechanical 
 joint, thus ensuring an even distribution of the current 
 through the mass of the electrode. 
 
 This construction not only permits, as already stated, 
 a sliding feed adjustment of the electrodes, but also a con- 
 stant feed, even when a length of carbon becomes too short 
 to be of any further use by itself. The procedure then 
 consists in making a joint outside the furnace, between it 
 and the new electrode and rendering the same secure by a 
 species of graphitic cement. 
 
 The Cowles' form of terminal connexion for open furnaces 
 
 249 
 
ELECTRIC FURNACES AND 
 
 consists in a metal socket for the reception of the carbon 
 electrode. A water-cooling jacket is arranged in close 
 proximity to the metal socket, and an inlet pipe is fitted 
 in such a position that the ingoing water strikes against 
 that wall of the jacket which adjoins the electrode socket. 
 
 An outlet is also provided so that a constant circulation 
 may be maintained. 
 
 A form of water-cooled terminal connexion, designed by 
 O. Imray, is adapted to electrodes of the suspended type. 
 The carbon block forming the electrode is suspended from a 
 species of metal cross-head, above, by yoke pieces, shaped 
 to suit the block, and secured to the supporting cross-head 
 by bolts and nuts. The length of the yoke pieces is such 
 that the cross-head and its attached parts are sufficiently 
 removed from the heat-zone which surrounds the upper 
 part of the carbon electrode. 
 
 Electrical connexion with the carbon block is secured by 
 a metal capping, which either butts up flat against its upper 
 extremity, or is shaped so as to encircle the block for a 
 short distance down. This capping is connected with the 
 cross-head by a central metal tube, and is water-jacketed 
 as regards its surface of contact with the carbon. The 
 upper end of the tube is jointed to the main conductor, 
 whilst axially through its centre runs a second smaller tube 
 or inlet for the water supply, which it delivers, as already 
 stated, just above the metallic wall which is in contact with 
 the carbon. Openings or outlets for the water are pro- 
 vided in the outer tube, and provision is made for a com- 
 plete circulatory system, whereby the yoke pieces themselves 
 are also kept cool. 
 
 In conjunction with this patent is described a special 
 form of outlet for molten furnace products. Instead of 
 the usual tapping hole, with its attendant drawbacks, 
 slits are made in the wall of the furnace, and lined, for pro- 
 tection, with steel, or equally suitable metal, which is, in 
 turn, water- jacketed. 
 
 250 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 The form of water-cooled terminal construction originally 
 adopted for the Pictet carbide furnaces at Ingleton. is shown 
 in Fig. 57. The main cable M, consisting of stranded 
 No. 19 B.W.G. copper wire, made up to a cross-section 
 equal to three square inches of solid copper, is compressed 
 into rectangular form, with dimensions, 2J by 2 in., and 
 sweated into the metal socket S, which is water- jacketed 
 as shown. The square extremity of the carbon electrode 
 C, 6 by 6 in., is socketed into the other extremity of the 
 terminal, which is mounted on a travelling carriage A, 
 moved along the 
 slide B, as occasion 
 requires, by means 
 of a screw and hand- 
 wheel W. 2,000 
 amperes were effect- 
 ually dealt with by 
 this form of terminal 
 construction. 
 
 Electrodes. As may be readily imagined, the consump- 
 tion of electrodes in the electric furnace, especially where 
 such furnaces are worked on an extensive scale, as in carbide 
 manufacture, is by no means inconsiderable. 
 
 Several firms have, in consequence, taken up the manu- 
 facture of electrodes for use in their own furnaces. Among 
 these may be mentioned the Societe des Carbures Metal- 
 liques, which owns carbide factories situated at Notre Dame 
 de Briancon, Le Castelet, Bellegarde, and Berga. The 
 electrode factory is situated at the first named, and the 
 following essential features of the manufacturing process 
 are extracted from an article on the subject by M. G. Strauss, 
 which appeared in the Revue General de Chimie Pure et 
 Applique in 1901. 
 
 The raw materials are retort carbon, petroleum coke, 
 and coal tar. The first named is obtained in varying 
 qualities, and calls for thorough mixing to ensure homogeneity 
 
 251 
 
ELECTRIC FURNACES AND 
 
 in the finished product. The second, petroleum coke, is a 
 porous residue, produced in the distillation of crude petro- 
 leum. It is employed specially in the manufacture of very 
 hard and dense electrodes such as are used in electrolytic 
 furnaces with fused electrolytes, and is very pure. The 
 coal tar, used as a binding material, is either produced 
 in process of refining ammonia, from which it distils over 
 at a temperature of from 240 to 330F. = 115 to 166C., 
 or it may take the form of a mixture, artificially prepared, 
 by dissolving pitch at steam heat in a heavy tar oil. 
 
 The process itself is briefly as follows. The carbon or 
 coke, as the case may be, is first crushed in special machinery, 
 and subsequently sieved into powders of varying degrees 
 of comminution, their classification and ultimate use 
 depending upon the size of electrode to be made from 
 them. A quantity of powder of the required grade is 
 then selected and mixed with the requisite quantity of 
 coal tar, in steam-jacketed vessels, after which the mixture 
 is ground under heavy edge runners. From the grinding 
 process it emerges as a thin, flat, plastic cake, ready for 
 compressing and moulding into shape. To this end, it is 
 formed, by hydraulic pressure, into " cartridges," or ingots, 
 which may weigh anything from 180 kgs. to a ton, according 
 to the size of electrode required. The latter are finally 
 shaped by extrusion from a die under enormous hydraulic 
 pressure, amounting to as much as 2,000 tons for the larger 
 sizes. 
 
 All that remains before the electrodes are finally ready 
 for use in the furnaces, is to stove them, a process carried 
 out in a Siemens regenerative furnace, worked continuously, 
 the electrodes being packed for the purpose in muffles, and 
 the interstices filled in with carbon dust. This final stoving 
 occupies a period of several days. 
 
 To ensure a satisfactory product, periodical tests for 
 electrical conductivity are made, by fastening a sample of the 
 electrode in a vertical position in a metal mould. Molten 
 
 252 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 lead is then poured in at either end to establish electrical 
 contact with the carbon surfaces, and holes, drilled in the 
 two lead castings, at a certain pre-determined distance 
 apart, serve as mercury cups for terminal connexion to the 
 testing apparatus, which takes the form of a Thomson low 
 resistance bridge. 
 
 The conductivity is an extremely variable quantity, 
 depending in turn upon the nature of the raw materials 
 used, and upon each and every stage of the manufacturing 
 process. 
 
 Current density and temperature have a most important 
 bearing upon the life of electrodes, more particularly anodes, 
 in an electrolytic bath. Graphitized carbon is less susceptible 
 to disintegration than ordinary carbon. The Castner 
 alkali process was originally worked with carbon electrodes 
 graphitized by the inventor's own process, which consisted 
 in packing ordinary carbons in carbon dust, and subjecting 
 them to a white heat in an electric furnace. 
 
 Graphitized carbon electrodes, as manufactured at 
 Niagara Falls, under the Acheson patents, are now largely 
 used in electrolytic works. 
 
 The electrodes used in the manufacture of aluminium 
 must be free from silica and silicates, which, if present as 
 impurities, lead to the formation of silicon fluoride, by re- 
 action with the fluorine, thus robbing the bath of the latter 
 element. Silica also is liable to reduction by the lower 
 strata of aluminium in the bath, a reaction which would 
 destroy the purity of the product. Hydrocarbons may be 
 present in the carbon electrodes used for aluminium manu- 
 facture, and have no deleterious effect ; the carbon should, 
 however, be free from ash. 
 
 In carbide manufacture, on the other hand, the presence 
 of hydrocarbons and silica in the carbons of the electrodes 
 is immaterial, but phosphate impurities are debarred, as 
 they lead to the formation of calcium phosphide, with the 
 result that the acetylene gas generated from the resultant 
 
 253 
 
ELECTRIC FURNACES AND 
 
 carbide is rendered impure by admixture with phosphoretted 
 hydrogen. 
 
 The choice of suitable carbon electrodes for the various 
 electric furnace operations detailed in this book, though 
 seldom thus regarded, is, in reality, one of the most important 
 features in electric furnace design, contributing, as it does, 
 in no small measure, to the efficiency of the process, purity 
 of the product, and last, but by no means least, as a refer- 
 ence to some of the cost statistics already given, will show, 
 the cost of the process. 
 
 Carbon exists in various polymeric forms, amorphous 
 carbon, sp. gr. 1*7; graphite, sp. gr. 2'25 ; diamond, sp. gr. 
 3'5 ; and various other equally well-known varieties. 
 
 The following is an account of the method of electrode 
 manufacture as adopted by the British Aluminium Company, 
 and furnished by Mr. John Sutherland to the Electro-Chemist 
 and Metallurgist, for April, 1901. 
 
 " Carbon electrodes for the reduction of aluminium may 
 be made from any form of pure carbon. The materials 
 generally used by the various carbon works are retort carbon, 
 petroleum coke, soot, anthracite, and Ceylon graphite. 
 The crude material is first dried or roasted, to eliminate 
 water, or any organic impurities, and then ground in a roller 
 or ball mill. The ground carbon is then passed to the 
 mixer, where it is thoroughly mixed with anhydrous taiy 
 in sufficient quantity to make it bind together properly. 
 The mixers in use are various, but those made by Werner 
 and Pfleiderer seem to answer best. The mixture of tar 
 and carbon having been thoroughly incorporated, so that 
 each grain of carbon has its own coating of tar, is ready 
 for formation into blocks, which is effected in an hydraulic 
 press, and the carbons may be either squirted, or pressed 
 direct. 
 
 " The next step in the process is the baking of the car- 
 bons, which is done in a furnace, heated, preferably, by 
 gaseous fuel. The style of furnace is not important, pro- 
 
 254 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 vided the goods are not heated or cooled too quickly, and a 
 high temperature is obtained. During the baking, the 
 carbons are packed in carbon powder to protect them from 
 oxidization. 
 
 " Good carbons should emit a distinct metallic sound 
 when struck, and should not be porous. Unfortunately 
 there is no distinguishing test for electrodes ; the only 
 means to determine between good and bad ones is to put 
 them through the ordinary process in the electric furnaces, 
 and observe their behaviour therein." 
 
 Mr. F. A. Fitzgerald, of the International Acheson 
 Graphite Company, has pointed out (Paper before Niagara 
 Falls Convention of American Electro- Chemical Society) 
 that in both electro-chemical and electro-metallurgical pro- 
 cesses, involving the use of carbon electrodes, the deter- 
 mination of the density of the latter is an important point, 
 in that the general efficiency of the electrode depends upon 
 the nature of the carbon, and consequently on its density. 
 It is desirable to know both the real and apparent densities, 
 for, having given these two quantities, the porosity of the 
 electrode can be arrived at by calculation. In this con- 
 nexion, it may be added that the real density is that of the 
 carbon composing the electrode, whilst the apparent density 
 is the ratio of the weight of the electrode to its volume taken 
 as a whole. The porosity is the ratio of the difference of 
 the volume of the electrode taken as a whole, and the volume 
 of the carbon of which it is composed, to the volume of the 
 electrode taken as a whole. 
 
 Owing to the retention of air by amorphous carbon and 
 graphite, the determination of the real density is not easily 
 accomplished. Mr. Fitzgerald employs a volumeter, fitted 
 with a rubber stopper and air-pump connexion. The volu- 
 meter, partly filled with kerosene, is placed in a water bath 
 and a reading taken. A specimen of the electrode to be 
 tested is weighed and placed in the volumeter, which is 
 then exhausted of air by means of the pump. When the 
 
 255 
 
ELECTRIC FURNACES AND 
 
 maximum attainable vacuum has been reached, and main- 
 tained for a period of ten minutes, the pump is disconnected 
 and a reading again taken, after which the operation is 
 repeated until two equal consecutive readings are obtained. 
 Then the difference between this last and the original 
 reading of the volumeter represents the actual volume of 
 the sample, and the real density is calculated from it. 
 
 The apparent density of the same sample is now deter- 
 mined by coating it lightly with shellac, and immersing it 
 in water in the volumeter, which then indicates the apparent 
 volume by displacement ; the apparent density is arrived 
 at, as before, by calculation. 
 
 Mr. Fitzgerald also points out that, in view of the fact 
 that, for electro-metallurgical purposes, electrodes of pure 
 graphite are preferable, it is desirable that some test be 
 instituted to determine whether such electrodes are entirely 
 composed of graphite, or whether they include amorphous 
 carbon in their composition, in which latter case they are 
 especially subject to disintegration. 
 
 Mr. C. M. Hall, the inventor of the aluminium process 
 which bears his name, has devised a method of baking carbon 
 electrodes, especially those intended for aluminium manu- 
 facture, in an electric furnace of the resistance type. To this 
 end, the carbons are disposed either in horizontal or vertical 
 rows, around a carbon core, from which they are insulated, 
 and from one another by a refractory and non-conducting 
 powder, such as bauxite, purified alumina, magnesia, etc. 
 The core is preferably made up of granular charcoal or 
 coke, but a fused electrolyte is also suggested. 
 
 By passing a suitable current through the core, the car- 
 bons are raised to a temperature of 1,647 to 2,202C. = 3,000 
 to 4,000F., very little current actually passing through 
 the carbons themselves. The following data are suggested. 
 A core, 4 or 5 ft. in length, and 12 by 30 in. in cross-sectional 
 area, requires to have developed therein 400 to 500 H,P., 
 by an alternating current at 35-40 volts, 
 
 256 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION XIII 
 
 EFFICIENCY AND THEORETICAL CONSIDERATIONS 
 The efficiency of a simple electric furnace is the ratio of 
 
 the heat energy of the electric current usefully applied to 
 
 bring about the operation of the furnace, to the total heat 
 
 energy generated. 
 
 The heat energy usefully applied in an electric furnace 
 
 operation is absorbed by one or both of two requirements, 
 
 viz. : 
 
 (1) The heat energy necessary to raise the temperature 
 of the furnace charge to that point at which the desired 
 reaction occurs. 
 
 (2) The heat energy taken up by the reaction itself. 
 Both quantities can be determined by calculation ; the 
 
 former, from the mass, specific heats, latent heat of fusion, 
 and temperature, whilst the latter can be obtained from 
 known thermo-chemical data. 
 
 The efficiency of an electric furnace is governed by many 
 factors, for the greater part incidental to the particular 
 process and type of furnace under consideration. Such 
 general determinants as size, temperature of reaction, 
 radiation, disposition of terminals, etc., may, however, 
 be cited as controlling factors, especially the first named, 
 size, or dimensions of the furnace. 
 
 A little consideration will serve to show the importance 
 of this point. The capacity or cubic contents of a furnace 
 increase approximately as the cube of the linear dimensions, 
 whilst the surface available for radiation, one of the principal 
 channels through which heat energy is lost or wasted, 
 
 257 s 
 
ELECTRIC FURNACES AND 
 
 only increases as the square of the linear dimensions ; 
 in other words, the possibility of radiation losses does not 
 increase in direct proportion to the increase in capacity 
 of the furnace, so that a large furnace is far more efficient 
 than a small one. The general tendency therefore should 
 be towards an increase in the dimensions and capacity of 
 electric furnaces, only limited by mechanical or other 
 considerations. 
 
 The subject of electric furnace efficiency was most ably 
 dealt with by Prof. J. W. Richards, Ph.D., in a Paper read 
 before the American Electro-Chemical Society in 1902. 
 For purposes of discussion, Prof. Richards classifies electric 
 furnaces generally under two main heads, viz. 
 
 (1) Those in which the charge is simply heated without 
 any chemical change taking place ; and 
 
 (2) Those in which, beside the heating of the charge, 
 there is also a chemical change. 
 
 A further subdivision is based upon the solid or fused 
 condition of the charge. We thus have four sub-headings, 
 viz. 
 
 (1) Heating without fusion and without chemical change, 
 
 (2) with 
 
 (3) without with 
 
 (4) with 
 
 which include nearly all the furnaces and furnace operations 
 described in this book, always excepting, of course, the 
 electrolytic section, in which the efficiency of the electrolysis 
 itself is a controlling feature, and must be duly considered 
 in arriving at the total efficiency of the furnace or process. 
 As an example of No. 1 of the above sub-headings, we 
 have the Acheson graphite furnace, in which anthracite 
 coal is converted into graphite. Chemical changes do, of 
 course, accompany the process, in the shape of progressive 
 formation and decomposition of carbides, but, from the 
 point of view of efficiency, they are negligible, in that the 
 
 258 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 heat energy absorbed exactly counterbalances that evolved. 
 The actual conversion of amorphous carbon into graphite 
 is accompanied by the evolution of heat, and therefore once 
 the temperature of reaction is reached, actually assists the 
 current in bringing about the desired result. 
 
 Roughly speaking, 5,450 kgs. of anthracite are converted 
 into 4,550 kgs. of graphite by an expenditure of 1,000 H.P. 
 for twenty hours. The heat actually evolved by the re- 
 action itself is some 10 per cent, of that supplied by the cur- 
 rent ; the total heat energy is therefore 110 per cent, of the 
 heat energy available from the current, of which 82-5 per 
 cent, is utilized, showing an efficiency of 82-5-^110=75 
 per cent. 
 
 Passing on to a consideration of the allied Acheson pro- 
 cess of graphitizing electrodes, described in the resistance 
 furnace section, 3,175 kgs. of electrodes, embedded in the 
 same weight of granular carbon, call for an expenditure of 
 1,000 H.P., the efficiency of the process being 38 per cent. 
 
 Class No. 2, above, is exemplified in the Jacobs process 
 of fusing calcined bauxite. The process is one of simple 
 fusion, carried out in a cylindrical crucible or hearth, by the 
 heat of the arc, there being no accompanying chemical 
 action. The efficiency, with a 1,360 kg. charge, is 74 per 
 cent. 
 
 Carborundum manufacture is cited as an example in 
 Class 3, with a calculated efficiency of 76-5 per cent. 
 
 Under Class 4, the manufacture of calcium carbide is the 
 best known example, with a net calculated efficiency of 
 63 per cent. 
 
 The principal difficulty attendant on the calculation of 
 the efficiency of many electric furnace processes is the pre- 
 sent scarcity of data concerning the specific heat of the 
 various substances treated, at high temperatures. An ex- 
 tensive scheme of research in this direction would doubtless 
 result in an increase of reliable data ; it is, however, a 
 matter of time, and, until this information has been gathered 
 
 259 
 
ELECTRIC FURNACES AND 
 
 together, the calculation of electric furnace efficiencies 
 must necessarily be approximate, rather than exact. 
 
 Summarizing the facts at his command, Prof. Richards 
 sets down a general figure of 60-75 per cent, as the average 
 commercial efficiency of furnaces varying in power from 
 200-1,000 H.P. 
 
 Taking the case of the Acker process, the 8,000 amperes 
 consumed in each furnace should yield, theoretically, 287 
 kgs. of caustic per twenty-four hours. In actual practice, 
 with a yield of 264 kgs., the current efficiency is 93 per 
 cent. The voltage theoretically required for the electro- 
 lysis of fused sodium chloride is 4-2, therefore the total or 
 energy efficiency of the process is in the neighbourhood of 
 55 per cent. This low efficiency is in part due to the heat 
 energy consumed in fusing the cold salt, and bringing it up 
 to the reaction temperature. 
 
 The general principle and disposition of a commercial 
 electric furnace would not, at first sight, appear to lend 
 itself very readily to theoretical consideration ; neverthe- 
 less, Mr. F. A. Fitzgerald, in a Paper entitled " Note on 
 Some Theoretical Considerations in the Construction of 
 Resistance Furnaces," which he read before a Niagara 
 Falls meeting of the American-Electro-Chemical Society in 
 1903, treated his hearers to a mathematical disquisition on 
 the theory of resistance furnaces, taking as his subject the 
 well known Acheson type, used in the manufacture of 
 carborundum, graphite, etc. 
 
 In all these furnaces, the electrical resistance is a variable 
 quantity, usually diminishing towards the end of a run, 
 and it is obviously impossible to effect regulation in the 
 interior of the furnace, except, perhaps, in the case of that 
 particular example used in the graphitizing of electrodes, 
 where some little adjustment is possible by varying the 
 spaces between neighbouring piles of electrodes under 
 treatment. 
 
 Regarding the distribution of heat in these furnaces 
 
 260 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Fitzgerald reasons as follows : If a cylinder (core) with 
 radius, p l and length L, is kept at a constant temperature, 
 , the quantity of heat, Q, passing through the surrounding 
 material, the inner surface of which is at the temperature 
 2 per second, is Q=2 < 7rp 1 LK (#! 2 )> where K is a 
 
 constant. Thus fl, 2 o~~ 
 
 If the material outside the core is also a cylinder, with a 
 radius p 2 , a heat conductivity of K 15 and a temperature at 
 
 its outer surface of 6 Z ,Q = 
 
 while between the outside of the cylinder of material under 
 treatment and the walls of the furnace, the passage of heat 
 
 per second is similarly q _ ^ 3 ~ *' 
 
 Pz 
 
 or, if p 2 p 3 =t, and is very small 
 
 84 
 Now, since the efficiency of the furnace is 9~? K ' and 
 
 Mo 
 
 03-04 must be kept small to make q small. Graphite 
 and electrode furnaces, however, cannot be made indefi- 
 nitely large, because the last equation shows q to vary 
 directly as the product of L, and p 2 . 
 
 In the case of the carborundum furnace, the chief 
 equation may be written 
 
 To make the largest amount of carborundum, K t must 
 be as great as possible, this being effected by having the 
 mixture surrounding the core, of high density, for then the 
 crystal mass formed will be also dense. 
 
 261 
 
ELECTRIC FURNACES AND 
 
 ~ # 3 should also be large ; but the latter is fixed at a 
 temperature just below that of the formation of carborun- 
 dum, while # 2 can only be raised to a point short of its 
 decomposition. As regards the outside of the carborundum 
 cylinder, nothing can be done to diminish k, but 3 4 
 may be kept small by having a good thickness of raw material 
 always present. Since the value of p 2 increases with the 
 length of the run, a point will eventually be reached where 
 q=Q in the efficiency equation, so that a time comes when 
 no more carborundum is made. To work efficiently, the 
 furnace must be stopped long before this happens. 
 
 According to MM. Gin and Leleux (Comptes Eendus, 126, 
 pp. 36), the temperature of the arc, when employed for 
 heating purposes, may be computed from the equation 
 
 where p is the resistance of the envelope of gas, 
 
 c is the specific heat of gas per unit of volume, 
 $ is the sectional area of the electrodes, 
 t is the temperature of the arc. 
 
 Expressed in words, this formula tells us that the tem- 
 perature (t) increases directly as the square of the current 
 density, and directly as the ratio of the total resistance of 
 the gaseous envelope to the specific heat per unit volume 
 of the same. 
 
 M. Gin has since supplemented this formula by others 
 (Elelctrochemische Zeitschrift, May, 1902) representing tem- 
 perature changes and efficiency of the arc furnace, assuming, 
 contrary to the above reasoning, that the medium sur- 
 rounding the electrodes, in the neighbourhood of the arc, 
 is a conductor. They are as follows 
 
 Let I be the length of the separating medium 
 s diameter 
 
 p ,, resistance ,, ,, 
 
 c ,, specific heat ,, ,, based 
 
 on unit volume. 
 
 262 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 Then the amount of energy converted into heat in unit 
 time is C 2 R, and the corresponding heat evolved is 
 
 I /I \ a 
 A\s ) 
 
 If the arc be surrounded by a heat insulating substance, 
 
 the equation becomes - f ] pis cist, from which it will 
 
 A\sJ 
 
 be seen that the temperature of the arc in the mass, will in- 
 crease in proportion to the square of the current density, a 
 condition which holds good, whether the intervening medium 
 between the electrodes be in a gaseous or fluid state. 
 
 M. Gin's formulae for calculating the temperature of a 
 resistance furnace are somewhat complicated, and are, 
 moreover, based on certain assumptions, which render 
 results more or less approximate. 
 
 In applying the following equations, the quantities must 
 be expressed in gramme-calories. 
 
 Let Cs be the mean specific heat of the core, or other 
 
 substance, in a solid state. 
 Cf be its latent heat of fusion. 
 Cl be its mean specific heat in a fluid state. 
 Or be the heat absorbed as a result of the chemical 
 
 changes. 
 P be the weight of the substance passing through the 
 
 furnace in unit time, when in full operation. 
 Tf be its temperature of fusion. 
 
 Let Tr be the temperature at which the reaction occurs, 
 and 2 be the total superficial area of the exterior of the 
 
 furnace from which radiation occurs. 
 Then 
 
 EJ = M6 [P(CsTf + Cf + Cl (Tr-Tf) + Cr)+K2STr]. 
 
 From this, Tr can be approximately calculated. Since 
 
 K Z S is a constant for one type of furnace, it may be written 
 
 K ; then, omitting CsTf, the equation simplifies down to 
 
 EJ = M6 [P(Cf+Cr + Cl (Tr-Tf) +KTr]. 
 
 263 
 
ELECTRIC FURNACES AND 
 
 whilst the effective work of the furnace is represented by 
 
 the formula 
 
 P[ Cf + Cr + Cl (Tr-Tf)] 
 
 P [Cf +O+C1 (Tr-Tf )]+KTr, 
 
 from which it follows that the lower the temperature at 
 which the desired reaction takes place, the lower will be the 
 thermal efficiency of the furnace, whilst the latter will in- 
 crease in proportion to the heat absorbed by the various 
 changes, both chemical and physical, taking place in the 
 substance of the furnace charge as a result of the reaction. 
 
 264 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 SECTION XIV 
 
 MEASUREMENT or FURNACE TEMPERATURES 
 
 It is necessary for the sake of uniformity, as also for 
 purposes of research, and the general advancement of 
 scientific knowledge concerning the reactions which take 
 place in the electric furnace at high temperatures, that 
 some convenient and reliable means be provided for indica- 
 ting, and, if necessary, recording, the temperature at which 
 any particular operation is being carried out. The provi- 
 sion of such means for high temperature measurement is 
 by no means so simple as would appear at first sight. 
 
 Owing to the extremely high degree of heat which a 
 furnace of the electrical variety is capable of evolving, it 
 is impossible to introduce anything in the nature of a ther- 
 mometer structure into the furnace itself, without risk of 
 its speedy destruction, either by fusion or combustion. 
 
 The ordinary types of thermometer which depend, for 
 their action, on the expansion, under heat, of a column of 
 liquid are, of course, out of the question, and it becomes 
 necessary to fall back upon some method, which either per- 
 mits the use of a very refractory material in the thermo- 
 meter bulb proper, or, for still higher temperatures, involves 
 a comparison by simple observation, between the tempera- 
 ture of the furnace and that of a known standard, which 
 can be safely manipulated from a point external to the 
 furnace itself. 
 
 So far as our present knowledge carries us, the only two 
 reliable and exact methods of direct furnace temperature 
 measurement depend upon a portion of the apparatus being 
 
 265 
 
ELECTRIC FURNACES AND 
 
 subjected to the direct heat of the furnace, and their range 
 is, in consequence, somewhat limited. Some very accurate 
 and valuable records of temperature have nevertheless 
 been obtained by their aid. They are both of an electrical 
 character, and depend for their action, the one upon the 
 rise in electrical resistance of a short length of platinum 
 wire, when subjected to heat, and the other, upon the 
 electro-motive force set up at the hot junction of a refrac- 
 tory thermo-couple. 
 
 Much valuable work in the perfecting of these two thermo- 
 electric methods of high temperature measurement, and 
 the means for accurately indicating and recording the 
 temperatures registered by them, has been done by Profs. 
 Callendar and Griffiths, their various designs being worked 
 out and manufactured on a commercial scale by the Cam- 
 bridge Scientific Instrument Company. No work on the 
 industrial application of high temperatures would be com- 
 plete without a description of these various forms of appa- 
 ratus, and we will therefore proceed to briefly discuss 
 them in their bearing on electric furnace work. 
 
 Of the two methods of electric thermometry, just con- 
 sidered, the resistance method yields the more accurate 
 results up to about 600C.= 1,112F. ; whilst, on the other 
 hand, the thermo-couple method has a greater range, and 
 permits the measurement of high temperatures in very 
 small enclosures. Its comparatively negligible time lag 
 also renders it especially suitable in cases where a con- 
 tinuous record of rapidly varying temperatures is desired. 
 
 Reverting for the moment to a consideration of the 
 platinum resistance thermometer, Professor H. L. Callendar, 
 in an article in the Philosophical Magazine for December, 
 1899, adduces several reasons for its adoption as a practical 
 standard for high temperature measurements, chief among 
 which may be cited the facility for establishing such a 
 standard in any part of the world, it only being necessary 
 to send a few grammes of the standard wire in an ordinary 
 
 266 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 letter to any desired spot in order to reproduce the scale 
 with great accuracy. 
 
 A paper by Mr. H. M. Tory, on his investigations into 
 the probable order of accuracy obtainable in comparing 
 high temperatures by means of commercial samples of 
 platinum wire, was read by Professor Callendar before the 
 Physical Society on June 22, 1900. Five different samples 
 in all were subjected to comparison, each wire being directly 
 compared with a pure platinum standard by winding the 
 two side by side on the same thermometric tube. It was 
 found that between 400 and 1,000C.=752 and 1,832F., 
 the fundamental coefficients of the wires varied within 40 
 per cent, of the maximum value, but that the temperatures 
 registered by them, when calculated on the platinum scale 
 by means of the ordinary formula, did not differ by more 
 than 9 at 1,000C. 
 
 Curves were plotted from the results obtained, having 
 the platinum temperatures of the standard wire as abscissae 
 and the difference between the temperatures indicated by 
 the two wires under comparison, as ordinates. Within the 
 limits of observation these were all straight lines, thus 
 indicating that it is only necessary to determine two con- 
 stants in order to compare a commercial platinum resist- 
 ance thermometer with the standard, and therefore with 
 the scale of the gas thermometer usually accepted as a 
 primary standard. These two constants are obtainable 
 from observations at the boiling point of sulphur and the 
 freezing point of silver, permitting the construction of a 
 practical thermometric scale, which, between and 1,000 
 does not vary from the gas scale by more than two or three 
 degrees. 
 
 In No. 435, Proceedings of the Royal Society, C. Chree 
 treats of some important investigations carried out at Kew 
 Observatory with a view to determining the accuracy of 
 platinum resistance thermometry. Six of these thermo- 
 meters were tested by means of a Callendar-Griffiths resist- 
 
 267 
 
ELECTRIC FURNACES AND 
 
 ance bridge, and no less than thirteen possible sources of 
 error were discovered. Among the principal ones are 
 thermo-electric currents, set up at the various junctions ; 
 heating of the resistance wire by the battery current em- 
 ployed in taking the temperature resistance measurements ; 
 errors in the temperature coefficient for the particular 
 sample of platinum wire used ; insufficient immersion in 
 the substance or space, whose temperature it is required 
 to measure, and excessive time lag, or slowness in acquiring 
 the temperature of the surrounding medium. 
 
 The effects of insufficient immersion, resulting in heat 
 being conducted away through the leads and external con- 
 nexions is, for commercial work, negligible, though it may, 
 if the total immersion be less than 10 c.m., amount to as 
 much as 0'01C., or even more. 
 
 The time lag calls for more consideration, especially if 
 the thermometer tube be only inserted in the furnace at 
 intervals, and thus not left continually immersed. 
 
 The six thermometers tested were found to take from 
 four to five times as long in acquiring the temperature of 
 the surrounding furnace atmosphere, as a mercurial thermo- 
 meter. 
 
 So important a value has the platinum resistance method 
 of high temperature measurement acquired of late years, 
 that in 1887 Professor Callendar, by dint of elaborate and 
 careful investigations which he then made into the pos- 
 sibilities of this method of thermometry, introduced the 
 scheme of " Platinum Temperatures," or the Platinum 
 Scale, which is a standard for the platinum resistance 
 method of thermometry, just as the Fahrenheit, Centigrade, 
 and Reaumur scales are standards for the previously exist- 
 ing mercurial instruments. These platinum temperatures 
 are obtained from the formula 
 
 where pt is the platinum temperature, corresponding with 
 
 268 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the ohmic resistance R of a given wire, the resistance of 
 which at 100C. and 0C.=212F. and 32F. are Rl and Ro 
 respectively. 
 
 The researches of Professors Callendar, Dewar, Fleming, 
 Griffiths and others show that this law holds good for all 
 temperatures between -200C. and 1,300C. = -328F. 
 and 2,372F., irrespective of the extremes and rapid varia- 
 tions of temperature to which the wire may be subjected, 
 provided it be carefully annealed ; and that the platinum 
 wire invariably offers the same resistance at the same 
 temperature. 
 
 So far, however, as the actual variation of resistance with 
 temperature goes, the law appears to be of a complicated 
 nature. 
 
 Among others, J. D. Hamilton has investigated this 
 subject, and in the Philosophical Magazine for December, 
 1897, he suggests the formula 
 
 (R+) 2 = p (t+b), 
 
 where a, p, and b are constants, and R and t the resistance 
 and temperature respectively. This most nearly repre- 
 sents the results hitherto obtained. The constants have, 
 however, to be individually obtained for each wire by a 
 series of careful observations. 
 
 Messrs. Crompton & Co., of Chelmsford, in this country, 
 and Messrs. Hartmann & Braun, of Frankfort, both manu- 
 facture a direct reading electrical thermometer based on 
 the platinum resistance principle. The indicator itself, 
 resembles a direct-reading ohm-meter, and consists of two 
 intersecting coils, mounted, and capable of rotation in a 
 non-homogeneous magnetic field. A standard resistance 
 of known value is included in the circuit of one coil, whilst 
 the other is connected with the coil of the resistance thermo- 
 meter. The latter, for high temperatures, is constructed, 
 as usual, of platinum, whilst, for lower degrees of heat, a 
 " nickelin " wire is employed. The scale is graduated 
 
 269 
 
ELECTRIC FURNACES AND 
 
 directly in degrees, and has a range extending up to 1,200C., 
 the highest temperature that can be measured by its aid. 
 
 For regular and constant use an open scale in the neigh- 
 bourhood of the most usual temperatures under measure- 
 ment can be obtained by a suitable adjustment of the shape 
 of the pole pieces. 
 
 The current required to actuate the indicator is only 
 0*03 ampere at 5 volts. 
 
 Messrs. Siemens Bros, also manufacture a conven- 
 ient form of platinum resistance pyrometer, in which the 
 resistance wire is wound upon a refractory cylinder, and 
 protected, for the greater part of its length, by an iron 
 tube ; being additionally safeguarded at its active extre- 
 mity, which is exposed to the direct heat of the furnace, by 
 a platinum shield. An increase in the resistance of the 
 platinum spiral, from 10 ohms to 44'9 ohms, corresponds 
 with a rise in temperature from 14C. to 1,205C.=58F. 
 to 2,204F. Two methods of reading the indications are 
 provided by the makers, either one of which can be adopted. 
 One consists of a differential galvanometer and a set of 
 resistance coils, which, used in conjunction with a battery, 
 give the required resistance in ohms, the corresponding 
 temperature being read off a scale supplied with the instru- 
 ment. The other apparatus consists of a Wheats tone 
 bridge combined with a D'Arsonval galvanometer, the 
 variable arm of the former being arranged in the form of a 
 circle, traversed by a sliding contact arm, carrying an index 
 which moves over a circular scale calibrated to read directly 
 in degrees Fahrenheit. 
 
 The use of the thermo-couple for high temperature ther- 
 mometry was proposed as far back as 1826 by the late M. 
 Becquerel. He adopted it for the measurement of the 
 underground temperature in the Natural History Museum 
 of Paris ; an iron-copper couple was employed, one junction 
 of which was situated in the underground area, whose 
 temperature was desired, whilst the other was immersed 
 
 270 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 in a bath, the temperature of which was capable of regula- 
 tion either by artificial heating or cooling, and could be 
 measured by means of an ordinary mercurial thermometer. 
 A galvanometer was included in the thermo-couple circuit, 
 and the temperature of the bath regulated until its deflection 
 became zero, a sure indication that both junctions were 
 at the same temperature. The temperature of the bath, 
 and consequently of the underground space, was then 
 measured by means of an ordinary mercurial thermometer. 
 
 M. Henri Becquerel, the grandson of the above, has 
 suggested a more direct method of using the thermo-couple 
 for this purpose, which he designates as the " Sliding Scale " 
 method. A high resistance D'Arsonval galvanometer is 
 included in the circuit from the thermo-couple, and has its 
 scale graduated to read directly in degrees of temperature. 
 In setting up the instrument to indicate the temperature 
 of a given enclosure, such as a furnace, the scale is so placed 
 that the zero or undeflected position of the galvanometer 
 mirror indicates upon it the temperature of the room in 
 which it is placed, and which has been previously deter- 
 mined by means of an ordinary mercurial thermometer. 
 Any subsequent deflection then gives the required tempera- 
 ture of the other junction directly, and without the neces- 
 sity for calculation, or for keeping the cold junction at zero 
 by immersion in an ice bath. 
 
 In Comptes Bendus, April 28, 1902, M. D. Berthelot gives 
 some very interesting and practical information respecting 
 the calibration of thermo-couples for use in high temperature 
 thermometry. According to him, the cheapest and most 
 satisfactory couple is formed of platinum, in conjunction 
 with a 10 per cent, alloy of platinum and iridium. To ensure 
 regularity in the curve of comparisons, the calibration should 
 be carried out in an atmosphere of air, nitrogen, or carbon 
 dioxide gas. 
 
 Between the temperatures of 400 and 1,100C.= 752 
 and 2,012F., the curve denoting the relationship between 
 
 271 
 
ELECTRIC FURNACES AND 
 
 Log. E.M.F. and Log. temperature is to all practical intents 
 and purposes a straight line. 
 
 M. Berthelot employed, for his limiting temperatures, 
 the melting point of zinc (419C.= 786F.) and the melting 
 point of gold (1,064C.= 1,947F.), the latter being deter- 
 mined automatically by the insertion of a piece of that metal 
 between the elements at the hot junction, and reading the 
 E.M.F. at the instant before the circuit is interrupted by 
 the fusion of the gold. 
 
 The Reichsanstalt apparatus for calibrating thermo- 
 couples for high temperature measurement is described 
 in the portion of the book dealing with laboratory furnaces. 
 
 Mr. H. J. Robinson, in letters to the Electrical Review, 
 March 6 and April 17, 1903, calls attention to some of the 
 drawbacks incidental to the use of thermo-electric pyro- 
 meters as at present constructed, and points out how they 
 may be in part eliminated. His experience appears to have 
 been gained in the use of platinum, platinum-rhodium, or 
 iridium thermo-couples, made up of wires "018 in. in dia- 
 meter, threaded through porcelain insulators, and protected 
 by a steel tube, merging, at its outer end, into a water-cooled 
 terminal box. In this type of pyrometer, the primary 
 trouble was experienced at the hot junction, which is effected 
 by fusing the ends of the two wires together. 
 
 Mr. Robinson found it more conducive to a constant 
 reading, when the ends forming the hot junction are tightly 
 twisted together for about a quarter of an inch, and fused 
 or not fused, in addition, as the case may be. Another 
 source of trouble was the rapid deterioration of the wires, 
 owing to the hot furnace gases entering the tube ; this was 
 circumvented by an extra sheathing of iron steam pipe, 
 slipped over the steel protecting tube, and its open extre- 
 mity welded up. This latter procedure also reduces the 
 trouble consequent on the liability of the platinum to 
 become brittle or " short " with continued usage. In this 
 connexion, he has found heating the wire white hot in air 
 
 272 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 a number of times, by passing a suitable current 
 through it, to be efficacious. A 120- volt supply was utilized, 
 the current being switched on and off about twenty times, 
 which removed all tendency to brittleness. The wire should 
 be well cleaned before replacing in its mounting. 
 
 Consequent on Mr. Robinson's letters came a communica- 
 tion from Mr. S. Weiss containing a suggestion that the steel 
 protecting tube, and water-cooled terminals be dispensed 
 with altogether, and the pyrometer mounting made entirely 
 of fire-clay, some 3 or 4 ft. in length, a form of construction 
 which, in the writer's opinion, is impracticable on the score 
 of weakness and liability to fracture. Mr. Weiss further 
 states that such a fire-clay tube may be rendered impervious 
 to furnace gases by washing it over with a mixture of pure 
 kaolin and water, mixed to the consistency of cream. 
 
 The possibilities of thermo-electric pyrometry are limited 
 by the fusing temperature of the metals employed in the 
 construction of the thermo-couples, and by the fact, recently 
 pointed out by Nernst, that all substances, commonly em- 
 ployed as refractory insulators, become good electrical con- 
 ductors at high temperatures. 
 
 As we travel higher in the temperature scale, we have to 
 fall back, as already stated, upon other methods, all of 
 which are founded on more or less direct comparison with 
 a negotiable standard situated outside the furnace. 
 
 One of the simplest and most ingenious of these methods, 
 which is also, unfortunately, limited as to its range of 
 applicability, consists in the comparison, by observation, 
 of the degree of incandescence of an electric lamp filament 
 with that of the heated interior wall of the furnace, an 
 inspection opening being provided for the purpose in the 
 wall of the latter. 
 
 The apparatus in general resembles a telescope, mounted 
 on a light tripod, and placed with its axis in line with the 
 opening in the furnace wall. A small, low-voltage, incan- 
 descent electric lamp, fed with current by two or more dry 
 
 273 T 
 
ELECTRIC FURNACES AND 
 
 cells, or small accumulators, is fitted within the tube of the 
 telescope, the current through it, and, consequently, its 
 degree of incandescence, being regulated by means of a 
 rheostat and switch. 
 
 The method of taking a temperature observation consists 
 in viewing the incandescent filament of the lamp against 
 the glowing background of the interior furnace wall ; if the 
 filament appear black, or dark, then its condition of incan- 
 descence, and consequently its temperature, is lower than 
 that of the furnace ; whereas if it appear as a white or light 
 line, its temperature is higher than that of the furnace 
 interior. By means of the regulating resistance and switch, 
 the current through the lamp is regulated until the filament, 
 as viewed against the incandescent furnace wall, apparently 
 disappears. This signifies that the two are in an equal state 
 of incandescence, and, consequently, that their temperatures 
 are equal. The actual temperature of the filament is then 
 read off from a scale, previously prepared by experiment, 
 which gives the relation between the current passing, as 
 registered on an ammeter included in the circuit, and the 
 temperature. 
 
 The range of this ingenious device is limited to 1,980C. 
 = 3,600F. 
 
 The optical pyrometer has lately been elaborated and 
 constructed on a more practical basis by Dr. H. Wanner. 
 It is being manufactured and marketed by Messrs. Emier 
 & Amend, of New York, and is applicable to temperatures 
 of 4,000C. = 7,232F., or above. 
 
 The following is Dr. Wanner's description of the appa- 
 ratus and principle involved 
 
 " The law, as established experimentally, permits at 
 least for a certain group of glowing bodies (the so-called 
 theoretically ' black bodies '), measurement up to the 
 highest temperatures. When a compact body is heated, 
 the rays emanating from the same may be observed by 
 the human eye, and the body's colour will change, with 
 
 274 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 rising temperature, from dark red to light red, to yellow, 
 and to bright white. This means that at first mainly red 
 rays are observed, to which, at the higher temperatures, 
 the other spectrum colours, orange, yellow, etc., are added, 
 until the rays appear white. 
 
 " In analyzing the rays by a prism it is found that with 
 the rise of temperature, some single colour, for instance, 
 red, undergoes a rise in intensity, which can be measured 
 progressively with a specially constructed photometer. 
 If we know the law for the mutual relation between the 
 determining factors (the temperature and the light intensity 
 of the single colour, and its wave length) we are enabled to 
 measure high temperatures by the photometric measure- 
 ment of the light intensity of a certain colour. 
 
 " The apparatus to be applied is therefore a photometer, 
 containing at the same time a prism, to separate a single 
 colour. A spectrum is produced in the ordinary manner 
 with the aid of a slit, lenses, and a prism, from which, by 
 means of a diaphragm, the light of a certain wave length 
 is separated, and the measurement of the light intensity 
 is made by polarization. To that part of the apparatus 
 which faces the radiation to be examined, a small incan- 
 descent lamp is attached, the light of which is used for 
 comparing the intensity of the light to be measured. 
 
 " On looking through the apparatus one observes the 
 circular field of vision, divided into two halves (like in a 
 one-half shade sugar- testing polariscope), one of which is illu- 
 minated by the small incandescent lamp, and the other by 
 the light of the glowing body being examined. Both halves 
 of the field show red colour. On turning the eye-piece 
 containing the Nicol prism both halves of the field of vision 
 can be easily brought to equal intensity, and on a circular 
 scale the number of degrees are read. The actual tempera- 
 ture is found from a table which accompanies each instru- 
 ment. The temperatures given on the table have been 
 calculated by means of the law mentioned before. 
 
 275 
 
ELECTRIC FURNACES AND 
 
 " The entire procedure is so simple that it can be readily 
 learned by any foreman or intelligent workman within a 
 short time. The whole apparatus is 30 c.m. long, built 
 like a telescope, and can be easily handled without a sup- 
 port. It does not matter how great the distance is from 
 which the measurements are taken, if only the field of 
 vision is properly illuminated by the light emanating from 
 the body to be examined. 
 
 " The exactness of this new method depends solely on 
 the accuracy of the observer, and on the degree to which 
 the body under test approximates what is called in the 
 theory of radiation, a ' black body.' Errors due to lack of 
 experience of the observer are practically eliminated, for 
 various parties, who have been asked to take measurements, 
 and who used this apparatus for the first time, have obtained 
 the same correct results. The nearest approximation to 
 the theoretical ' black body ' are the insides of the closed 
 furnaces, as muffles, etc., the glow of which is observed 
 through a small opening, which, however, must not be 
 covered by glass or mica during the measurement. The 
 measured temperature is equal to the real one within a few 
 degrees." 
 
 Another method of high temperature determination, 
 which verges more nearly on direct measurement than the 
 foregoing, is based, in principle, on Stefan's law, which runs 
 to the effect that the radiation from an absolutely black 
 body is proportional to the fourth power of its absolute 
 temperature. By an " absolutely black body " is meant, 
 in this case, one which receives and retains all the heat 
 imparted to it, and only gives it out again by radiation, 
 and not by reflection from its surface. 
 
 Kirchoff has shown that the interior walls of an enclosure, 
 such as a furnace, which are at a high uniform temperature, 
 behave as an absolutely black body, according to the above 
 definition, and its existence as such is not materially altered 
 by the presence of the small orifice necessary for taking 
 
 276 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 the temperature observation. Given, then, an absolutely 
 black body, in this case a portion of the incandescent wall 
 of the furnace, it is possible to concentrate, and measure, a 
 certain small definite proportion of the radiated heat, and 
 by this means determine the temperature. 
 
 The apparatus for achieving this in practice also takes 
 the form of a species of telescope, with an objective of fluor- 
 spar. The latter substance is chosen for this purpose by 
 the inventor, M. Fery, because its absorption of radiant 
 heat is very small, and the consequent error from loss, due 
 to the passage of the rays through it, is correspondingly 
 small. In actual practice, the sensitiveness of the device 
 is decreased by about 10 per cent. 
 
 The radiant heat thus concentrated by the fluor-spar 
 objective is received upon the junction of a miniature 
 thermo-couple placed at the focus ; the resulting E.M.F. is 
 measured, and, by simple computation, gives the required 
 temperature of the furnace. The lens is disposed for 
 parallel rays, and a diaphragm cuts off all but a certain pre- 
 determined cone, which is allowed to reach the thermo- 
 couple, thus rendering the device independent of its dis- 
 tance from the furnace. 
 
 M. Fery's apparatus has been standardized by comparison 
 with a Le Chatelier pyrometer, and exhibited an error well 
 below 1 per cent., between temperatures of 914C. and 
 1,450C.=1,677F. and 2,642F. respectively. Owing to 
 the minuteness of the mass to be heated (less than 1-1 00th 
 of a milligramme) there is practically no time lag, the appa- 
 ratus following quick variations in the temperature under 
 measurement with surprising accuracy. Neither has any diffi- 
 culty been experienced with the zero, which is very constant. 
 
 The temperature of the positive arc carbon, as measured 
 by this device, was found to be 3,490C.= 6,314F. 
 
 M. C. Fery has also experimented with a somewhat 
 similar method of high temperature measurement by radia- 
 tion, which is based on Wien's complex law. 
 
 277 
 
ELECTRIC FURNACES AND 
 
 Am T=A, where X is the wave length, T the absolute 
 temperature, and m refers to the maximum energy radiated 
 at a particular wave length. 
 
 This gives the radiation in terms of any chosen wave 
 length, instead of the total radiation as in the previous 
 method founded on Stefan's law. 
 
 M. Fery has compensated the complexity of Wien's 
 formula by a method of considerable ingenuity. He 
 reduces the standard radiation to equality with that under 
 measurement by interposing an acute-angled prism of 
 absorbent glass, in the path of the rays. This prism is so 
 displaced as to interpose varying thicknesses. The absorb- 
 ing power of the prism also follows a formula of the Wien 
 type, and is such that the displacement is inversely pro- 
 portional to the temperature under measurement. The 
 latter is thus read off from the degree of displacement of 
 the prism, when the two radiations have been equalized. 
 
 By this method, temperatures of 3,867C. and 3,897C.= 
 6,992F. and 7,046F., in red and green light respectively, 
 have been obtained for the positive carbon of the arc. 
 There is thus a discrepancy between these values, and the 
 3,490C.= 6,314F. obtained by the previous radiation 
 method, and it is accounted for by the fact that carbon 
 does not behave as an absolutely black body at the tem- 
 perature of the arc. 
 
 As far back as 1858 Balfour Stewart demonstrated that 
 the emissivity and the absorptive power of a body at a given 
 temperature, are equal for any radiation. This is known 
 in England as Stewart's, and on the Continent as Kir- 
 choff's, law, having been independently proved by the 
 latter. 
 
 By " absorptive power " is meant that portion of the total 
 radiation received by a surface, which is absorbed, whilst 
 the " emissivity " of a body at a given temperature, for 
 any given radiation, is the ratio of the quantity of that 
 radiation which it emits, to the quantity of the same radia- 
 
 278 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 tion emitted by an ideal black body at the same tempera- 
 ture and under the same conditions. 
 
 In 1879 Stefan evolved the law which is known by his 
 name, and is to the effect that the total radiation from any 
 ideal black body varies as the fourth power of the absolute 
 temperature. In 1884, a theoretical proof of this law was 
 furnished by Boltzmann, who based his reasoning on one of 
 Maxwell's fundamental laws regarding the electro-magnetic 
 theory of light, viz., that a beam of light exerts, in the direc- 
 direction of propagation, a pressure, per unit of area, equal 
 to the energy in unit volume of the radiations. Following 
 closely on the deductions of Boltzmann, came two laws, 
 formulated by Wien, and expressed by the equations : XT = a 
 constant, or \mT=a> constant, and EwT~ 5 = a constant. 
 The former, which is known as the " Law of Displacement," 
 has it that any monochromatic radiation is moved towards 
 the shorter wave length by a rise of temperature. 
 
 It has remained for MM. Lummer and Pringsheim to 
 supply the connecting link between theory and practice, 
 which they have recently done, by proving experimentally 
 the truth of these various relations between radiation and 
 absolute temperature. These investigators, taking the 
 mean of several readings, have proved the constant, for an 
 ideal black body to be \mT= 2,940. 
 
 These two investigators (Berichte der Deutschen Physi- 
 Icolischen Gesettschaft, I., 1903) state the three laws relating 
 to black radiation to be as follows 
 
 (l)/ 00 EXdX=:0-T 4 (Stefan-Boltzmann Law). 
 
 (2) AmT=A (contained in Wien's Displacement 
 
 Law). 
 
 (3) EmT 5 =B. 
 
 where EXdX is that portion of the black radiation contained 
 between X and X + dX at the temperature T of the gas ther- 
 mometer scale ; Xm is the wave length for which, at this 
 
 279 
 
ELECTRIC FURNACES AND 
 
 temperature, the emissive power EX, in the normal spectrum, 
 reaches its maximum Em ; m is the maximum energy 
 radiated at a particular wave length, whilst <r, A, and B are 
 constants, determinable with sufficient accuracy. 
 
 By means of a specially constructed carbon tube resist- 
 ance furnace, the interior of which represented the theo- 
 retical black body, and the radiation from which was mea- 
 sured simultaneously, by several different methods, including 
 a surface bolometer, spectrum bolometer, and spectrum 
 photometer, the above experimenters have succeeded in 
 verifying the foregoing laws, and rendering them applicable 
 to absolute temperature measurement up to 2,300C.= 
 4,172F., thus extending the possible range of correct 
 thermometry by about 1,000C.= 1,832F. 
 
 In order to apply the laws of radiation to the exact 
 measurement of high temperatures, all that is necessary is 
 to expose a bolometer or radiation meter to an orifice in the 
 heated furnace cavity, which is equivalent to an " absolutely 
 black body," and, by this means, measure the radiant energy 
 given off by it. 
 
 K. Ferrini has suggested the application of calorimetric 
 methods to the measurement of high temperatures, the 
 latter being calculated from results obtained with a calori- 
 meter. The following table gives the number of kilogramme 
 calories necessary to raise the temperature of a kilogramme 
 of platinum, or nickel, from zero, or freezing-point, to a given 
 temperature 
 
 280 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 FINAL TEMPERATURE. CALORIES ABSORBED BY 1 KILOGRAMME. 
 
 DEGREES C. PLATINUM. NICKEL. 
 
 100 3-23 12 
 
 200 6-58 24 
 
 300 9-95 37 
 
 400 13-64 50 
 
 500 17-35 63-5 
 
 600 21-18 ...... 75 
 
 700 25-13 90 
 
 800 29-20 103 
 
 900 ...... 33-39 117-5 
 
 ,000 37-70 134 
 
 ,100 42-13 
 
 ,200 46-65 
 
 ,300 51-35 
 
 ,400 56-14 
 
 ,500 61-05 . 
 
 ,600 66-08 
 
 ,700 71-23 
 
 1,800 76-50 
 
 Then the following formulae apply for the computation 
 of the temperature. Let ra be the weight of the metal, 
 M that of the water, ^ the initial, and t 2 the final tempera- 
 ture of the water, c the specific heat of the metal at t 2 , and 
 Q, the heat required to cool the body from the unknown 
 temperature, T down to zero, and we have 
 ra (Q-c^-y, 
 
 whence Q = ct 2 + (t 2 tj 
 
 Having thus determined Q, the value of T may be ascer- 
 tained from the table within 100 degrees. If greater accu- 
 racy be required, the following equations are applicable 
 
 For platinum 
 
 T + 52-84 T- Q -|- 6 =0. 
 
 For nickel 
 
 1+Q 
 
 T = 
 
 13 
 
 The results obtained are said to be correct between 700 
 and 800C.= 1,292, and 1,472F., whilst the errors intro- 
 duced when the temperatures are exceeded, are negligible. 
 
 281 
 
ELECTRIC FURNACES AND 
 
 For experimental, or laboratory work, the method is 
 worthy of consideration, but would appear altogether too 
 complicated for adoption on a commercial scale. 
 
 There is another way of estimating high temperatures, 
 which can, however, hardly be regarded in the light of 
 temperature measurement, in that it is of assistance only 
 in giving visible indication of the fact that a certain pre- 
 determined temperature has been reached. 
 
 It was devised by Prof. Seger, as a rough, practical 
 test for the pottery manufacturer, to determine the tem- 
 perature of his baking ovens, and consists of the well-known 
 " Seger Cones," or " Pyramids." As the result of a series 
 of exhaustive fusion experiments, Prof. Seger suc- 
 ceeded in producing a series of systematic mixtures of pure 
 silicates, aluminates, etc., each having a different melting 
 point, and each mixture differing, in its melting point 
 some 20 or 30C., from the one next above or below it in 
 the scale. A series of fifty-eight different mixtures or 
 compositions were thus worked out, with fusing points, 
 which varied from dull red, 590C. = 1,094F., to the 
 melting point of platinum, 1,850C.= 3,362F., in steps 
 of 20 or 30 C.=68 or 86 F. These mixtures, moulded 
 into conical or pyramidal form, are known as " Seger Cones," 
 and the method of using them to ascertain when a certain 
 pre-determined temperature has been reached, is to place 
 one or more of the cones, which, on the Seger scale of tem- 
 peratures, represent most nearly the desired temperature 
 at different points of the furnace cavity, suitably supported 
 and protected by refractory slabs. 
 
 When the fusing temperature of the composition of which 
 the cone is constructed has been reached, it softens, and 
 collapses, its upper portion, or apex, curving over. The 
 degree of curvature, if the temperature be not unduly ex- 
 ceeded, is, to the experienced eye, an additional gauge of 
 the exact degree which has been reached. Observation 
 holes in the walls of the furnace, and in line with the cones, 
 
 282 
 
THEIR INDUSTRIAL APPLICATIONS 
 
 are, of course, essential, and should be protected by mica 
 to prevent indraughts of air from vitiating the results. 
 
 The following table represents the chemical composition 
 of Seger cones over a limited portion of the scale, between 
 1,330C.=:2,426 F. and 1,530C. = 2,786F., inclusive 
 
 10 ' ' ' " 1<0 Al2 3 10Si 2- 
 
 ? CaOJ l ' 2 A1 *3 12 Si 
 12 ' ' ' (o'.7CaO| 1<4 A1 *3 14 
 
 13 cao 1 ' G M ** 16 Si 2 1 ' 390 
 
 14 ' ' ' 1<8 A1 2 18 Si 2 ' 1 ' 410 
 
 15 ' ' {o.7 Cao} 2-1 A1 23 21 Si 2 M30 
 
 16 ' ' ' {o^7 CaOJ 2 ' 4 A1 ^3 24 Si 2 1 ' 450 
 
 17 ' ' ' j 2 ' 7 ^3 27 Si 2 
 
 iQ vajv 2 v^| , . , o , s - o 14.00 
 
 i a . . . S f\n r^<-iO f -^ J 2 3 oiv^2 A,*{*VJ 
 
 10 (n^n'nl 3 ' 5 A1 2 3 3 5 Si0 2 . . 1,510 
 
 20 . 
 
 For full particulars, the reader is referred to the original 
 pamphlet, describing the cones, which was published by 
 the Thonindustrie Zeitung (Berlin). 
 
 Although essentially a rough and ready method of 
 recording high temperatures, the system has one great 
 advantage ; any workman, of average intelligence, can be 
 entrusted with the carrying out of temperature investiga- 
 tions by its aid. 
 
 283 
 
INDEX 
 
 Acheson Group of Electric Fur- 
 naces, 36 
 
 Carborundum Furnace, 36 
 
 Graphite Furnace, 43 
 
 Furnace Terminal, 247 
 Acker Sodium Process, 199 
 Advantages, Relative, of Arc and 
 
 Resistance Principles, 13 
 
 Principal, of Electric Furnace, 15 
 Alloys of Iron and Steel, 156 
 
 Aluminium, 159 
 Aluminium, Manufacture of, 176 
 
 Haber & Geipert's Experi- 
 
 ments, 181 
 - Heroult Process, 183 
 
 Heroult Furnace, 185 
 
 Hall Process, 186 
 
 Price of, 182 
 
 Theoretical Considerations in 
 
 Manufacture of, 188 
 Alundum, 214 
 Application of Electric Furnace to 
 
 Scientific Research, 228 
 Arc Furnaces, 21 
 
 Theoretical Action of, 26 
 
 Arndt's Calcium Process, 209 
 Arsenic Smelting, 146 
 
 Furnace, 147 
 Artificial Graphite, 43 
 
 History of, 44 
 Aschermann Chromium Process, 
 
 155 
 
 Barium Cyanide, 215 
 Baryta, 214 
 
 Becker Glass Furnace, 172 
 Blount's Carbide Furnace, 70 
 Borchers' Furnace, 87 
 
 Resistance Furnace, 30 
 
 and Stockem's Calcium Pro- 
 
 cess, 206 
 
 and Stockem's Strontium Pro- 
 
 cess, 208 
 
 Bradley Furnace, 82 
 Bronn Glass Furnace, 174 
 
 Calcium, Borchers' Process, 206 
 
 Arndt's, 209 
 
 Manufacture of, 206 
 
 Calcium Carbide Manufacture, 58 
 - Carbide, Purification of, 63 
 
 Temperature of Formation 
 
 of, 64 
 
 Quality of, 65 
 
 Current used in Manufacture 
 
 of, 65 
 
 Raw Materials in Manufacture 
 
 of, 67 
 
 Furnace, Blount's, 70 
 
 Willson's, 72 
 
 Gin & Leleux, 73 
 
 Parks', 75 
 
 Maxim, 76 
 
 Zimmerman, 78 
 
 Roller, 79 
 
 Morley, 79 
 
 Nikolai, 80 
 
 Horry, 81 
 
 Bradley, 82 
 
 Kenevel, 84 
 
 Pictet, 85 
 
 Borchers, 87 
 
 Memmo, 89 
 
 Parker, 93 
 
 Frohlich, 95 
 
 Cowles, 96 
 
 Contardo, 97 
 
 King, 98 
 
 Deutsche Gold und- Silber 
 
 Scheide Anstalt, 99 
 Calcium Cyanide, 215 
 Carbon Bisulphide Furnace, 55 
 Carborundum, 36 
 
 Furnace, Acheson, 36 
 
 for lining Furnaces, 39 
 Chromium, Preparation of, 155 
 Classification of Electric Furnaces, 
 
 10 
 
 Coal, Peat, Jebsen Process, 53 
 : Bessey Process, 55 
 Coke Purification, 52 
 Combustion Furnace, 241 
 Combination Furnace Processes, 217 
 Comminution of Metals, 161 
 Conley Smelting Furnace, 128 
 Contardo Smelting Furnace, 130 
 
 Carbide Furnace, 97 
 Copper Smelting, 144 
 
 285 
 
INDEX 
 
 Cores, Tests of Resistance Furnace, 
 
 29 
 Cowles, early experiments, 7 
 
 Furnace, 31 
 
 Carbide Furnace, 96 
 
 Zinc Furnace, 151 
 
 Sodium Process, 156 
 Crucible Furnace, Howe, 225 
 
 Darling Sodium Process, 205 
 De Chalmont Furnace, 21 
 Deductions, Scientific, 19 
 Definition of Electric Furnace, 1 
 De Laval Smelting Furnace, 112 
 
 Zinc Furnace, 152 
 Denbergh Arc Furnace, 23 
 Dental Muffle, Hammond, 226 
 Deutsche Gold und Silber Scheide 
 
 Anstalt Carbide Furnace, 99 
 Distillation of Zinc, 151 
 Dorsemagen Zinc Process, 154 
 Dumoulin Resistance Furnace, 31 
 
 Early History of Electric Fur- 
 nace, 1 
 
 Eddy Tube Furnace, 240 
 Efficiency, Furnace, 257 
 Eimer Electric Oven, 226 
 Electric Furnace, Definition of, 1 
 Electrodes, 251 
 
 Graphitizing, 50 
 
 Hall method of Baking, 256 
 Electrolytic Furnace Processes, 176 
 
 Faure Furnace, 7 
 
 Ferro -Manganese, Simon Process, 
 
 210 
 
 Fischer Sodium Process, 204 
 Franklin Smelting Furnace, 117 
 Frohlich Carbide Furnace, 95 
 
 Resistance Furnace, 34 
 Furnace, Electric, Definition of, 1 
 Siemens, 2 
 
 Moissan, 5 
 
 Faure, 7 
 
 Temperature attainable in, 8 
 
 Classification of, 10 
 
 Principal Advantages of, 15 
 
 Arc, 21 
 
 Three Phase, Advantages 
 
 of, 25 
 
 Carbonizing Lamp Fila- 
 ments in, 25 
 
 Resistance, 27 
 
 Application of Polyphase 
 
 Currents to, 34 
 
 Furnace, " Induction " 35 
 
 Acheson Group, 36 
 
 for Graphitizing Electrodes, 
 
 51 
 Baking Carbon Articles in, 
 
 52 
 
 Coke Purification in, 52 
 
 Tube, 235 
 
 - Terminals, 246 
 
 - Electrodes, 251 
 
 - Efficiency, 257 
 
 Gibbs Resistance Furnace, 32 
 Gin Steel Furnace, 112 
 Process, 113 
 
 - & Leleux Carbide Furnace, 73 
 Girod Resistance Furnace, 33 
 Glass, Manufacture of, 170 
 
 - Furnace, Shade, 170 
 
 Henri vaux, 172 
 
 Becker, 172 
 
 Voelker, 172 
 
 Bronn, 174 
 
 Graphite, Artificial, 43 
 
 Furnace, Acheson, 43 
 
 Artificial, History of, 44 
 Conditions for Formation 
 
 of, 48 
 
 Graphitizing Electrodes, 50 
 Guntz Tube Furnace, 244 
 
 Hall Aluminium Process, 186 
 Hammond Dental Muffle, 226 
 Harmet Smelting Furnace, 115 
 Henri vaux Glass Furnace, 172 
 Heraeus Tube Furnace, 243 
 Heroult Smelting Furnace, 120 
 
 - Steel Process, 122 
 
 Aluminium Process, 183 
 - Furnace, 185 
 
 History, Early, of Electric Fur- 
 nace, 1 
 
 - of Artificial Graphite, 44 
 Horry Carbide Furnace, 81 
 Howe Crucible Furnace, 225 
 Hulin Sodium Process, 198 
 
 Induction Furnace Principle, 35 
 Iron and Steel Production in Elec- 
 tric Furnace, 107 
 
 Processes, 110 
 
 Alloys, 156 
 
 Irvine Phosphorus Furnace, 167 
 
 Keller Smelting Furnace, 113 
 
 286 
 
INDEX 
 
 Kenevel Carbide Furnace, 84 
 King Carbide Furnace, 98 
 Kjellin Steel Process, 138 
 
 Furnace, 139 
 Roller Arc Furnace, 25 
 
 Carbide Furnace, 79 
 
 Laboratory Furnaces, 221 
 Lead, Production of in Electric 
 Furnace, 211 
 
 Magnesium, Production of, 211 
 Machalske Phosphorus Process, 168 
 Magnetic Field, Monk's Patent, 16 
 Manganese, Manufacture of, 209 
 Maxim Carbide Furnace, 76 
 Measurement of Furnace Tempera- 
 ture, 265 
 
 Memmo Carbide Furnace, 89 
 Metals, Pulverization of, 161 
 Lomax Method, 162 
 
 Bary Method, 162 
 
 Miscellaneous Electric Furnace Pro- 
 cesses, 214 
 Moissan's Researches, 4 
 
 Furnace, 5 
 
 Molybdenum, Preparation of, 155 
 Morley Carbide Furnace, 79 
 Muffle, Hammond Dental, 226 
 
 Weiss Resistance, 227 
 - Winter Resistance, 227 
 
 Nickel, Production of, 155 
 Nikolai Carbide Furnace, 80 
 
 Optical Pyrometer, 274 
 Oven, Eimer Electric, 226 
 
 Parks Carbide Furnace, 75 
 Parker Carbide Furnace, 93 
 Patten Arc Furnace, 24 
 Peat Coal, Jebsen Process, 53 
 
 Bessey Process, 55 
 
 Phoenix Process, 190 
 Phosphorus, Manufacture of, 165 
 
 Furnace, Readman-Parker, 165 
 
 - Irvine, 167 
 
 Machalske, 168 
 
 Pictet Carbide Furnace, 85 
 Platinum Resistance Thermometry, 
 
 268 
 Polyphase Currents, Application of 
 
 to Electric Furnaces, 34 
 Potter Tube Furnaces, 236 
 Pradon Carbide Furnace, 66 
 Pulverization of Metals, 161 
 
 Purification of Calcium Carbide, 63 
 
 Coke, 52 
 
 Readman-Parker Phosphorus Fur- 
 nace, 165 
 
 Relative Advantages of Arc and 
 Resistance Principles, 13 
 
 Researches, Sir Wm. Siemens', 2 
 
 Moissan's, 4 
 
 Application of Electric Furnace 
 
 to Scientific, 228 
 Resistance Furnaces, 27 
 
 Cores, Tests of, 29 
 
 Walls of, 33 
 
 Ruthenberg Smelting Furnace, 124 
 
 Process, The, 126 
 
 Salgue's Zinc Process, 154 
 Scholl Sodium Process, 204 
 Scientific Deductions, 19 
 Seger Cones, 282 
 Shade Glass Furnace, 170 
 Siemens' Researches, 2 
 
 - Furnace, 2 
 Silicides, 216 
 Siloxicon, 40 
 
 Furnace, 41 
 Simon Smelting Furnace, 115 
 
 - Ferro-Manganese Process, 210 
 Smelting with Electrolysis, Chlor- 
 ine, 190 
 
 Furnace, De Laval, 112 
 Gin, 112 
 
 Processes, 113 
 
 - Keller, 113 
 
 - Simon, 115 
 
 Harmet, 115 
 
 - Franklin, 117 
 Weber, 119 
 
 Heroult, 120 
 
 Ruthenberg, 124 
 
 Conley, 128 
 
 Contardo, 130 
 
 Stassano, 133 
 
 Kjellin, 139 
 
 Tone, 143 
 
 Smelting, Copper, 144 
 
 Arsenic, 146 
 
 Zinc, 151 
 
 Sodium and Caustic Soda, 197 
 
 Process, Cowles, 156 
 
 Vautin, 197 
 
 Hulin, 198 
 
 Acker, 199 
 
 Castner, 202 
 
 287 
 
INDEX 
 
 Sodium Process, Scholl, 204 
 
 Fischer, 204 
 
 Darling, 205 
 
 Stassano Process, The, 131 
 
 Furnace, 133 
 Statistics, General, 18 
 
 Steel Production in Electric Fur- 
 nace, 107 
 Strontium, Manufacture of, 206 
 
 Borchers Process, 208 
 
 Swinburne- Ash croft Process, 190 
 
 Temperature Attainable in Electric 
 Furnace, 8 
 
 of Formation of Calcium Car- 
 
 bide, 64 
 
 Measurement of Furnace, 265 
 By Radiation Methods, 276 
 
 by Calorimetric Methods, 
 
 280 
 
 by Seger Cones, 282 
 
 Terminal Connexions and Elec- 
 trodes, 246 
 
 Acheson, 247 
 
 Thermo-Electric Pyrometry, 270 
 Theoretical Action of Arc Fur- 
 nace, 26 
 
 Considerations, 257 
 
 Theory of Aluminium Manufac- 
 ture, 188 
 
 Three-Phase Currents, Advantages 
 of, 25 
 
 Furnaces, Memmo, 89 
 
 Tone Smelting Furnace, 143 
 Tube Furnaces, 235 
 
 Potter, 236 
 
 Terminal Connexion and 
 
 Mounting of, 237 
 
 Revolving, 238 
 
 on Nernst Principle, 239 
 
 Eddy, 240 
 
 Heraeus, 243 
 
 Guntz, 244 
 
 Tungsten, Preparation of, 155 
 
 Vanadium, Preparation of, 212 
 Vautin Sodium Process, 197 
 Voelker, Arc Furnace, 25 
 Glass Furnace, 172 
 
 Walls of Resistance Furnaces, 33 
 Weber Smelting Furnace, 119 
 Weiss Resistance Muffle, 227 
 Westman Arsenic Furnace, 147 
 White Stuff, 39 
 Willson Carbide Furnace, 72 
 Winter Resistance Muffle, 227 
 
 Zimmerman Carbide Furnace, 78 
 Zinc, Distillation of, 151 
 
 Electrolytic Extraction of, 212 
 
 Furnace, Cowles, 151 
 De Laval, 152 
 
 Process, Salgues, 154 
 
 Dorsemagen, 154 
 
 But!:r & Tanner, The Selwood Printing Works, r>ome, and London. 
 
 288 
 
MINERAL TEEHNOLQGY LIBRARY 
 UNIVERSITY OF CALIFORNIA LIBRARY 
 BERKELEY 
 
 Return to desk from which borrowed. 
 This book is DUE on the last date stamped below. 
 
 JUL 1 4 1990 
 1954 
 
 LD 21-100m-9,'48(B399sl6)476