UC-NRLF B 3 110 X X RAYS TABLES OF PHYSICAL AND CHEMICAL CONSTANTS AND SOME MATHEMATICAL FUNCTIONS G. W. C. KAYE, M.A. (Cantab.), D.Sc. (Lond.), A.R.C.Sc. (Lond.) AND T. H. LABY, M.A. (Cantab.) Royal 8vo, Qs. net LONGMANS, GREEN AND CO. LONDON, NEW YORK, BOMBAY, CALCUTTA, AND MADRAS PIG. 1. Photograph obtained by C. T. R. Wilson of the path of a beam of X rays in supersaturated air. The beam of rays, some 2 mm. in diameter, was sent through the moist air (from left to right in the figure) immediately after the expansion which produced the supersaturation (see p. 155). The axis of the camera was horizontal, and the magnification of the photograph is about 2 times. FiQ. 2. Photograph obtained by C. T. R. Wilson showing the passage of X rays (from left to right) through a thin copper plate (see p. 158). X RAYS BY G. W. C. KAYE, M.A., D.Sc., CAPT. R.E. (T.) HEAD OF THE RADIUM AND X-RAY DEPARTMENT AT THE NATIONAL PHYSICAL LABORATORY SECOND EDITION WITH ILLUSTRATIONS LONGMANS, GREEN, AND CO. 39 PATERNOSTER ROW, LONDON FOURTH AVENUE & 3 oTH STREET, NEW YORK BOMBAY, CALCUTTA, AND MADRAS 1917 All rights reserved PREFACE TO FIRST EDITION. THIS little book does not profess to be a treatise or hand- book on X rays. It aims merely at giving an account of such of the present-day methods and apparatus as appear valuable or novel, and which, in many cases, can only be found scattered throughout many journals ; it treats critically, and here and there somewhat comprehensively, some of the features which have laid claim to the interests of the writer from time to time ; it is concerned to some extent with the development of theory as well as of experi- ment ; and it attempts to convey a notion, however dis- connected and ill-proportioned, of the historical trend of events from Prof. Rontgen's world-famous discovery in 1895 down to the end of the year 1913. The author trusts that the form of the book will be accept- able, not only to the student of physics, but to the man of general scientific interests, and particularly to the members of the medical profession, most of whom are keenly alive to the possibilities of the rays which Rontgen placed at their service. He is aware from experience as teacher and examiner of medical students, at Cambridge and London, of their need of a book on the subject which is neither recondite nor mathematical. To two of his colleagues at the National Physical Labora- tory, the writer wishes especially to record his grateful thanks. Mr. E. A. Owen has revised both manuscript and proof, and has co-operated extensively in the treatment of Chapter XIII., on the Diffraction of X Rays by Crystals. Mr. W. F. Higgins has given freely and generously of his time and energies, and rendered invaluable aid in all the different stages of the work. He has executed with great care a large proportion of the diagrams, and is responsible for the preparation of the index and some of the more lengthy tables. 366710 viii PREFACE Finally, the author would wish to thank his wife and Mr. J. R. Willis for general criticism, and Mr. A. A. Robb of St. John's College, Cambridge, for permission to include his verses on Maxwell's famous equations and the birth of an X ray. Mr. Robb's skill in parody is not so well known outside Cambridge as his mathematical researches ; and the author ventures to hope that the ." Revolution of the Corpuscle," which first saw light in the Post-Prandial Proceedings of the Cavendish Laboratory, will serve to " temper the wind " of those critics who can see only the numerous shortcomings in the book. The writer will be content if his work can be regarded as one of the many tokens of esteem with which old students of the Cavendish School of Research have delighted to honour their distinguished professor, Sir J. J. Thomson. February 1914. PREFACE TO SECOND EDITION. IN this edition the writer, so far as his military duties would permit, has subjected the book to a thorough revision and incorporated original work of note published up to the middle of 1916. He would again record the generous and experienced help of Mr. W. F. Higgins, who has added to his many kindnesses by writing an additional chapter on X-Ray Equipment and Technique. The editor of the Journal of the Rontgen Society has kindly permitted the reproduction of a number of photographs. The author is happy to think that his little book has been of service to the many X-ray workers now labouring in the common cause of humanity in our hospitals both at home and overseas. Sept. 1916. CONTENTS. PAGE PREFACE - .... - - vii TABLE OF NATUBE AND PROPERTIES OF VARIOUS KINDS OF " RAYS " - xv " THE REVOLUTION OF THE CORPUSCLE " - - - xvi INTRODUCTION xix TABLE OF ABBREVIATIONS OF REFERENCES TO JOURNALS - xxii CHAPTER I. THE PHENOMENA OF A DISCHARGE TUBE. THE POSITIVE COLUMN - - .... 1 THE CATHODE DARK-SPACE - - 2 THE FLUORESCENCE OF THE TUBE - 3 CHAPTER II. CATHODE RAYS. HISTORICAL - - 4 THE NATURE OF THE CATHODE RAYS - 5 J. J. THOMSON'S EXPERIMENTS - 7 CORPUSCLES AND ELECTRONS 8 WEHNELT CATHODES - - ... 8 PROPERTIES OF CATHODE RAYS - 10 ELECTRON THEORY OF MATTER - 18 POSITIVE RAYS. NATURE OF THE POSITIVE RAYS - - 19 J. J. THOMSON'S EXPERIMENTS 20 THE " IONICS " OF AN X-RAY TUBE - 23 x CONTENTS CHAPTER III. X RAYS. PAGE THE DISCOVERY OF X RAYS . . 24 METHOD OF PRODUCTION OF X RAYS 25 CHAPTER IV. AN X-RAY BULB. EARLY X-RAY TUBES - 29 THE ANODE - - 33 THE CATHODE - - 34 THE ANTICATHODE - - 35 SUITABLE METALS FOR THE ANTICATHODE - - 38 METHODS OF COOLING THE ANTICATHODE - - 41 THE COOLIDGE TUBE - - - - 43 THE SNOOK HYDROGEN TUBE - -46 CHAPTER V. HIGH-POTENTIAL GENERATORS. INFLUENCE MACHINES - -51 INDUCTION COILS - - 53 CORE - 54 PRIMARY WINDING - - - - 55 CONDENSER - 55 SECONDARY WINDING - - 57 FEATURES OF COIL DESIGN - 59 STEP-UP TRANSFORMERS - . 63 BREAKS AND INTERRUPTERS . 66 THE HAMMER BREAK 66 THE WEHNELT INTERRUPTER 66 MERCURY BREAKS - - 68 RECTIFIERS AND VALVE -TUBES 70 CONTENTS xi CHAPTER VI. THE "HARDNESS" OF AN X-RAY BULB. PAGE THE FACTORS CONTROLLING THE HARDNESS OF AN X-RAY BULB 72 THE PROGRESSIVE HARDENING OF AN X-RAY BULB WITH USE 75 DEVICES FOR SOFTENING AN X-RAY BULB - 78 CHAPTER VII. THE BLACKENING OF AN X-RAY BULB. CATHODIC DISINTEGRATION OR " SPUTTERING " - 80 THERMAL VOLATILISATION OF THE ANTICATHODE 85 VIOLET COLORATION OF THE BULB 87 CHAPTER VIII. THE MEASUREMENT OF X RAYS. THE INTERNATIONAL AND BRITISH RADIUM STANDARDS - 88 METHODS OF MEASURING INTENSITY - 90 CURRENT THROUGH THE TUBE - 90 THERMAL METHODS OF MEASURING INTENSITY - - 91 IONISATION METHODS - 91 PHOTOGRAPHIC METHODS - 95 FLUORESCENCE METHODS - . - 97 METHODS USED IN MEDICINE : THE PASTILLE, ETC. 97 METHODS OF MEASURING HARDNESS - - 99 POTENTIAL APPLIED TO X-RAY TUBE - 99 WAVE -LENGTH - - - 102 ABSORPTION METHODS - - 103 BENOIST'S PENETROMETER, ETC. - - 107 ENERGY MEASUREMENTS - 109 CHAPTER IX. SCATTERED, CHARACTERISTIC AND SECONDARY CORPUSCULAR RAYS. SCATTERED X RAYS - - 112 DISTRIBUTION - 114 POLARISATION 114 xii CONTENTS PAGE CHARACTERISTIC OR "MONOCHROMATIC" X RAYS - 116 K AND L SERIES OF RADIATIONS - 117 CHARACTERISTIC RADIATIONS FROM RADIOACTIVE ELEMENTS 122 CHARACTERISTIC LIGHT RAYS - 123 DIRECT GENERATION OF CHARACTERISTIC RAYS - - 125 CONNECTION WITH CATHODE RAY VELOCITY - 130 ABSORPTION OF CHARACTERISTIC RAYS - 135 SECONDARY CORPUSCULAR RAYS - - 145 DISTRIBUTION - 145 VELOCITY - 146 ABSORPTION - 147 CHAPTER X. FURTHER PROPERTIES OF THE X RAYS. lONISATION PRODUCED BY X RAYS - 151 C. T. R. WILSON'S FOG CONDENSATION EXPERIMENTS - 155 VELOCITY OF X RAYS - 158 CHAPTER XI. PRACTICAL APPLICATIONS OF X RAYS. RADIOGRAPHY - 160 INSTANTANEOUS RADIOGRAPHY - - 165 PHYSIOLOGICAL AND CURATIVE ACTIONS OF X RAYS - - 167 X-RAY " BURNS " -167 PROTECTIVE DEVICES - 167 GLASSES ESPECIALLY TRANSPARENT TO X RAYS - 171 THERAPEUTIC USE OF HOMOGENEOUS X RAYS - 172 PRESENT-DAY RADIOLOGY AND RADIOTHERAPY 173 CHAPTER XII. X-RAY EQUIPMENT AND TECHNIQUE. HIGH-TENSION GENERATORS - 177 CONTROL SWITCHBOARDS - - - 179 CONTENTS xiii PAGE MEASURING INSTRUMENTS - 182 X-RAY TUBES - 184 HIGH-TENSION CIRCUITS - - 184 TUBE-STANDS AND COUCHES - 186 PHOTOGRAPHIC APPARATUS - 191 LOCALISATION - - - 191 CHAPTER XIII. DIFFRACTION OF X RAYS BY CRYSTALS. LAUE'S THEORY OF DIFFRACTION BY CRYSTALS - - 195 THE EXPERIMENTS OF FRIEDRICH AND KNIPPING - 197 BRAGG'S THEORY - - 205 THE X-RAY SPECTROMETER - ^211 CRYSTAL -STRUCTURE - 218 WAVE-LENGTHS OF X RAYS AND ATOMIC DISTANCES IN CRYSTALS - 223 MOSELEY'S EXPERIMENTS - - 224 WAVE-LENGTHS OF y RAYS - - 231 CHAPTER XIV. THE NATURE OF THE X RAYS. IDENTITY OF X RAYS AND ULTRA-VIOLET LIGHT RAYS 237 STOKES' SPHERICAL PULSE THEORY - - 239 THE LOCALISED PULSE THEORY OF J. J. THOMSON - - 244 PLANCK'S QUANTUM THEORY - 245 THE OUTSTANDING PROBLEM OF RADIATION - - 248 APPENDIX I. AN INTERVIEW WITH PROF. RONTGEN by Sir James Mackenzie Davidson - 249 APPENDIX II. THE PRODUCTION OF HIGH VACUA - - * 251 xiv CONTENTS APPENDIX III. PAGE ELECTRICAL INSULATORS - ... 255 APPENDIX IV. TABLE OF ATOMIC NUMBERS AND ATOMIC WEIGHTS - - 258 TABLE OF DENSITIES - 260 TABLE OF VALUES of e~ Kd - - 261 TABLE OF CATHODE-RAY VELOCITY AND POTENTIAL - - 262 TABLE OF CHARACTERISTIC y RAYS - - 263 APPENDIX V. RECOMMENDATIONS FOR THE PROTECTION OF X-RAY OPERATORS 264 INDEX 266 NATURE AND PROPERTIES OF VARIOUS KINDS OF "RAYS." f = charge carried by hydrogen ion in electrolysis, e =unit of electricity-^ =4-7x 10~ 10 electrostatic units (E.S.U.), [ =l-57x 10~ 20 electromagnetic units (B.M.U.). NATURE. RAYS. RATIO OF CHARGE TO MASS, (efm), ETC. VELOCITY. cms. /sec. RANGE, ETC. E.M.u.grm.-i E.s.tr.grm.- ] ELECTRI- I CALLY ( NEUTRAL. ^ Hertzian waves. Infra-red rays. Visible light rays. Ultra- violet light. Entladung- strahlen. 106 to 0-4 cms. A 0-031 to 7-7x10-5 f 7-7x10-5 to WOVP 1 3-6 x 10-5 IWave /3-6xlO-Ho l^g "I 6 x 10-6 About 10-6 ? ) 3 x low. From a few mm. of air to infinity. Xrays Wave length : 1-2 x 10-7 to 1-7 x 10-9 cm. 3 x iQio. From a few cms. to over 100 metres in air at N.T.P. y rays of Ra, U, Ac, Th, etc. Wave length ; 1-4 x 10-& to 7-0 x 10-!0 cm. 3 x IQio. Reduced to 1 per cent, by J mile of air at N.T.P. Ordinary atoms and molecules. H 2 atom w= 1-64 x lO- 24 grm. Diam. =2-2 x 10~ 8 cm. H 2 : 18-4 x 104 \ at O 2 : 4-6xl04/0C. Mean free path of Ho at C., 1-8 x 10-5cm. CARRIERS OF NEGATIVE / ELECTRI- \ CITY. Electrons. Corpuscles. Cathode rays. Lenard rays. Negative ion at low pressures. (For small 1-77 x 107 rm = 1/183( LDiam. =4 velocities.) 5-31 x 1017 )H 2 atom. 1 < 10-is cm. J Photoelectrons 107 to 108. Very small. Wehnelt cathode rays 108 to 109. Very small. Cathode rays 109 to lOio. Range in air a few mms. /3 rays of Ra, U, Th, Ac, K, etc. /3 rays of Ra lOio to 2-99 x IQio. Stopped by about 1 cm. of lead. S rays (slow /3 rays). As low as 3-2 x 108. Too slow to ionise. Negative ion. May have several charges (though usually one) and up to 30 molecules. In air ; 1-8 for unit electric field. Negatively charged atoms and molecules in discharge tubes. 10^ (for H 2 ). 3 x 101^ (for H 2 ). Up to 108. . CARRIERS OF POSITIVE / ELECTRI- \ CITY. a rays of Ra, U, Th, Ac, etc. (helium atoms with charge 2e). }4-8 x 10- [w =6-56 x 10-^] 1-4 x 1Q14. Initial value 1-6 x 109 to 2-2 x 109 (depend- ing on source). 3 to 8 cms. in air at N.T.P. Recoil atoms. 47 (RaB). 1-4 x lOi 2 . 5xl04(RaC). 1/10 mm. in air at N.T.P. Positively charged atoms and molecules ( Kanalstrahlen) in discharge tubes. 104 (for H 2 ). 3 x 1014 (for H 2 ). Up to 108. Positive ion. May have several charges (though usually one) and up to 30 molecules. In air ; 1'5 for unit electric field. THE REVOLUTION OF THE CORPUSCLE.' A CORPUSCLE once did oscillate so quickly to and fro, He always raised disturbances wherever he did go. He struggled hard for freedom against a powerful foe An atom who would not let him go. The aether trembled at his agitations In a manner so familiar that I only need to say, In accordance with Clerk Maxwell's six equations It tickled people's optics far away. You can feel the way it's done, You may trace them as they run dy by dy less d/3 by dz is equal K . dX/dt While the curl of (X, Y, Z) is the minus d/dt of the vector (a, b, c). Some professional agitators only holler till they're hoarse, But this plucky little corpuscle pursued another course, And finally resorted to electromotive force, Resorted to electromotive force. The medium quaked in dread anticipation, It feared that its equations might be somewhat too abstruse, And not admit of finite integration In case the little corpuscle got loose. For there was a lot of gas Through which he had to pass, And in case he was too rash, There was sure to be a smash, Resulting in a flash. Then dy by dy less d/3 by dz would equal K . dX/dt While the curl of (X, Y, Z) would be minus d/dt of the vector (a, b, c). The corpuscle radiated until he had conceived A plan by which his freedom might be easily achieved, 1 Air : " The Interfering Parrot " (Geisha). REVOLUTION OF THE CORPUSCLE xvii I'll not go into details, for I might not be believed, Indeed I'm sure I should not be believed. However, there was one decisive action, The atom and the corpuscle each made a single charge, But the atom could not hold him in subjection Though something like a thousand times as large. The corpuscle won the day And in freedom went away And became a cathode ray. But his life was rather gay, And he went at such a rate, That he ran against a plate ; When the aether saw his fate Its pulse did palpitate. And dy by dy less d/3 by dz was equal K . dX/dt. While the curl of (X, Y, Z) was the minus d/dt of the vector (a, b, c). A. A. R. INTRODUCTION. IN the early nineties, it was not infrequently maintained that the science of physics had put its house in complete order, and that any future advances could only be along the lines of precision measurement. Such pessimism has been utterly confounded by a sequence of discoveries since 1895, unparalleled in their fundamental nature and promise. Even many not specially concerned have had their atten- tion directed to the recent attempts at solving the riddle which has excited interest and taxed ingenuity since the beginning of civilisation the problem of the ultimate struc- ture of matter. The chemist and physicist have long built upon a theory of atoms and molecules ; though information as to the existence and behaviour of individual atoms was only based on speculation, however justifiable. But within the last decade we have not only isolated the atom, but we have learnt a great deal about its internal structure. Radioactivity has, for example, introduced us to an electrically charged atom of helium (the a ray) with characteristics such that it can, in spite of its extreme small- ness, 1 make individual appeal to our senses. The speed of a rays is so abnormally high, 2 that if, for instance, they are allowed to strike a fluorescent screen, as in the Spin- thariscope of Sir Wm. Crookes, each atom possesses enough energy to record its arrival by a visible flash of light. This provided what was probably the first instance of the regis- tering of a single individual atom. Rutherford and Geiger similarly turned to account the electric charge, and have 1 Mass about 7 x 10~ 24 gramme ; diameter about 2 x 10~ 8 cm. 2 About 12,000 miles or 2 x 10 !l cm. per sec. xx INTRODUCTION actually recorded the arrival of single atoms by means of a delicate electrometer. More recently, C. T. R. Wilson, by the aid of his most beautiful and ingenious experiments on fog condensation, has succeeded in rendering visible and photographing the paths, not only of single charged atoms, but of electrons and X rays as well. The emission of such charged atoms from the radioactive elements proceeds entirely regardless of let or hindrance on our part : we have, however, the ability and means to create similar abnormal carriers for ourselves they are, for instance, to be found in abundance in a rarefied gas through which an electric discharge is passed. Under these conditions, the gas molecules, which are ordinarily electri- cally inert, assume, in many cases, electrical charges and. in addition, may have their usual velocities l increased a thousand-fold, so that they acquire properties which single them out from their fellows. These facts are sufficiently attractive in themselves, but even those few among us who, severely practical, still look askance at ' ionics/ cannot but agree that the close study which has been given to the phenomena of a discharge tube has already been more than repaid by the further discovery of electrons by J. J. Thomson and of the X rays by Rontgen. The amazing properties of the X rays excited universal astonishment at the time of their discovery. An X-ray outfit is now indispensable to the surgeon and physician j and the debt which the world of humanity owes the X rays, to take the present war alone, is a heavy and increasing one. Through the efforts of a devoted band of workers, with an outlook on life not immediately utilitarian, the Rontgen rays have thrown a searchlight on many phases of atomic physics not susceptible to other methods of attack. And, last of all, and quite recently, X rays have come to the aid of the crystallographer and triumphantly dis- played in the hands of Laue, Friedrich and Knipping, Bragg and others, the regular grouping of the atoms in a crystal. 1 About 20 miles per minute in the case of air molecules. INTRODUCTION xxi The experiments, which have opened up an immense field of enquiry, have at the same time given the long-deferred answer to the problem of the nature of the unknown or ; ' X " rays. In this connection the work of Moseley stands out pre-eminently. The early death of this brilliant young physicist, while serving as an officer in the Gallipoli campaign, is not the least of the tragedies of the present war. The Geissler discharge tube 1 the former beautiful play- thing of the scientist has proved the pioneer of some of the most wonderful discoveries and speculations that physical science of this or any generation has known. Truly, as Maxwell predicted in the early seventies, the vacuum tube has shed light upon the whole domain of electrical science, and even upon the constitution of matter itself. Our present intent, that of the study of the X rays, is approached most naturally by way of a scrutiny of the general phenomena of a discharge tube and of electrons in particular. To these branches of the subject we accord- ingly first direct ourselves. 1 Known at different times and in different countries as the Pliicker, Hittorf, or Crookes tube. ABBREVIATIONS OF REFERENCES TO JOURNALS. A.d.P. Annalen dor Physik. A.Rt.R. Archives of the Rontgen Ray. C.R. Comptes Rendus. D.P.G.V. Verhandlungen der Deutschen Physikalischen Gesell- schaft. J.d.P. Journal de Physique. J.Rt.S. Journal of the Rontgen Society. N. Nature. P.C.P.S. Proceedings of the Cambridge Philosophical Society. P.M. Philosophical Magazine. P.P.S. Proceedings of the Physical Society. P.R. Physical Review. P.R.S. Proceedings of the Royal Society (Series A). P.R.S.E. Proceedings of the Royal Society of Edinburgh. P.T. Philosophical Transactions (Series A.). P.Z. Physikalische Zeitschrift. CHAPTER I. THE PHENOMENA OF A DISCHARGE TUBE. WHEN a current of electricity from an induction coil or influence machine is sent between two metal electrodes fused into the ends of a glass tube (say 12 inches long) from which the air is gradually withdrawn by a pump, the tube presents a continuous succession of striking appear- ances. At high pressures, air is a very bad conductor of electricity ; and a large force is necessary to produce a visible discharge while the pressure remains in the region of atmospheric. But a reduction of pressure facilitates the passage of the spark, which after a time loses its noisy character and is replaced by a collection of sinuous and irregular pink streamers which later broaden and fill almost the whole of the tube with a pink diffuse glow known as the positive column. Simultaneously the alternative spark-gap of the coil diminishes to a small fraction of an inch evidence that the rarefied air is now conducting well. Meanwhile the cathode the electrode by which the cur- rent leaves the tube l assumes at its tip a luminous tuft the negative glow violet in colour, which later grows until it completely envelops the cathode. Between these two luminous glows comes a darker ill-defined region called the Faraday dark-space. These general appearances correspond to a pressure of some 8 to 10 millimetres of mercury. As the pressure is further reduced, the alternative spark- gap begins to lengthen, the anode becomes tipped with a 1 The electrode by which the current enters the tube is the anode. A 2 X RAYS vivid speck of glow, and the positive column proceeds, if the current density is suitable, to break up into thin fluctu- ating pink discs or striae, which subsequently thicken and diminish in number, intensity and extent. The Faraday dark-space enlarges, and in the meantime (at about 1 mm. pressure), the violet negative glow increases in brightness and volume, 1 and the glass walls of the tube are seen to fluoresce with an olive-green light which, as J. J. Thomson (P.C.P.S. 1910) has shown, is due to the action of extremely active ultra-violet light from the negative glow. 2 Dark Sp&ce. FIG. 3. Discharge tube at low pressure, showing cathode dark-space and positive striations. As the exhaustion proceeds, this fluorescence disappears, * the negative glow detaches itself like a shell from the cathode, while a new violet film forms and spreads over the surface of the cathode. Thus the negative glow now consists of two parts : they are separated from each other by a narrow dark region called the Crookes or cathode dark-space, which has a sharply defined outline running parallel to that of the cathode. (See Fig. 3.) With a reduction of pressure, the dark space increases in width, and pushes the outer negative glow before it. The dark space is often used as a rough indication of the pres- sure, though its width depends also on both the current 1 The length of the glow on the cathode depends also on the current : the two are indeed roughly proportional (Hehl, 1903). 2 This is especially marked in hydrogen. The fluorescence is bluish- green with lead glass, THE PHENOMENA OF A DISCHARGE TUBE 3 density and the metal of the cathode, and is not really a reliable guide to the degree of exhaustion. 1 With higher rarefactions (say -fa mm.) both positive and negative glows become less bright and definite in outline, and finally lose almost all traces of luminosity. Meanwhile the cathode dark space has grown at the expense of all else, until finally it becomes so large that its boundaries touch the glass walls of the tube. It is at this stage that the tube begins to shine, first in the region of the cathode, and then (as the dark space extends) over its whole surface, with the brilliant apple-green fluorescence 2 well known to those who are accustomed to X-ray tubes. All this time the length of the alternative spark-gap has been steadily increasing. If the exhaustion is pressed still further, the fluorescence diminishes, and the resistance of the tube increases, until finally it becomes impossible for the discharge to pass at all. The pressure at which this comes about depends partly on the induction-coil and partly on the size of the discharge tube. A large tube always runs more easily than a small one at the same pressure. 8 The above are the more salient features to be observed in a discharge tube. At atmospheric pressure the discharge consists mainly of positive column ; at low pressures it is the phenomena of the cathode dark-space which are most conspicuous and have attracted widespread attention. 1 See Aston, P.R.S. 1912. 2 This is quite different in appearance from, and much more brilliant than the olive-green fluorescence at higher pressures (see above). The yellow-green colour holds only for soda-glass tubes. Lead and lithium glasses yield a bluish fluorescence. In point of fact, Prof. H. Jackson (N. July 1915) has shown that a truly vitreous glass exhibits little if any fluorescence. The green fluorescence of X-ray tubes is associated with the presence of a small amount of manganese ; and manganese dioxide is, in fact, deliberately added so as to yield the colour of fluorescence by which the X-ray worker has grown accustomed to judge his tubes. CHAPTER II. CATHODE RAYS. Historical. The study of the green fluorescence on the glass was com- menced by PI ticker as long ago as 1859, and carried on vigorously by Hittorf (1869) and Goldstein (1876) in Ger- many, by Crookes (1879) in England, and by Puluj in Austria. It was soon ascertained that the fluorescence was produced by something coming from the region of the cathode ; for a suitably interposed obstacle cast a sharp shadow on the walls of the tube. This " something " was given the name cathode rays (Kaihodenstrdhlen) by Gold- stein, who believed that the fluorescence was produced by waves in the ether for whose propagation the gas in the tube was not necessary. The expression has been retained, and rightly, for although the word " ray " has come to be associated with a wave- motion in the ether, the connection is quite fortuitous ; indeed Newton used the term in his corpuscular theory of light. The general properties of the cathode rays soon revealed themselves. Pliicker had found that their path was bent by a magnet. Goldstein, by suitably contrived shadow experiments, confirmed Hittorf's observation that the rays travel in straight lines, and showed further that they start at right angles to the surface of the cathode. Crookes, by the use of cathodes shaped like a concave mirror, demonstrated that the rays concentrate near the centre of curvature, and there display in a marked degree heating properties and an ability to excite phosphorescence in CATHODE RAYS 5 many substances. They tend also to push away any object against which they strike, and cause, for example, paddle- wheels of mica to rotate, although we now know that such experiments depend for their success, not on the momentum of the rays, but on heating effects such as prevail in a radio- meter. 1 Most important of all, cathode rays, when they strike matter, generate X rays, as Rontgen showed in 1895. The Nature of the Cathode Eays. For nearly thirty years the English and German schools of physics disagreed as to the nature of the cfathode rays. German physicists, follow- ing Goldstein's lead, held that the rays were similar to light a wave motion in the ether with which matter had nothing to do. This view received much ' Anode , P ,1 . FIG. 4. A Lenard tube for showing the pass- Support irom tne experi- age of cathode rays through a thin aluminium ments of Hertz (1892), and his friend and colleague Lenard (A.d.P. 1894), who showed that solid bodies were not absolutely opaque to the rays which could pass, for example, through gold and aluminium foil. Lenard's historic vacuum-tube was provided with a small " window " of aluminium foil, 0*000265 cm. thick. (See Fig. 4.) When the cathode rays were shot against this window, he found that they passed through without puncturing it, and were able to excite phosphorescence, etc., a few millimetres away in air. 2 There was no diffi- culty in accounting for the transparency of thin solids on the ether-wave theory, though it was apparent that the relation to ordinary optical transparency was slight ; for instance, gold leaf is more transparent than clear mica to 1 A paddle-wheel made of a good thermal conductor such as aluminium does not show the effect. In a radiometer the vanes are propelled by the recoil of the gas molecules from the warmer face of each vane. 2 The Lenard rays travel farther in attenuated air. We know now that part, at any rate, of the phosphorescence which these " Lenard rays " produced was due to X rays generated by the aluminium. X rays were not discovered until the following year. 6 X RAYS cathode rays. On the other hand, the normal emission of the rays from the cathode could scarcely have been antici- pated on any wave theory : one would have inferred a general emission in all directions. But the magnetic deflectibility of the cathode rays, and its unsatisfactory explanation on a wave theory, were re- garded as crucial by English physicists, who were unanimous in the view that the rays consisted of particles of matter charged with negative electricity and projected with im- mense speeds from the cathode. Varley in 1870 seems to have been the first to suggest, though somewhat vaguely, the presence of such particles in the electric discharge ; and Sir Wm. Crookes, in a series of papers in the Philosophical Transactions (1879-1885), definitely adopted this standpoint to explain and coordinate the many striking results of his experiments. In a flight of intuition, Crookes suggested that the matter constituting the cathode rays was neither gas nor liquid nor solid as ordinarily known, but in a fourth state tran- scending the gaseous condition an extremely happy sur- mise, as events proved. It is interesting to recall that Faraday as long ago as 1816 speculated on the possibility of the existence of such a fourth state of matter. Crookes supposed the cathode particles to have the dimensions of molecules, but this view became hard to re- concile with their penetrating power for metals found (as mentioned above) by Hertz and Lenard some years later. To meet the difficulty, the suggestion was put forward that the metal, under the bombardment of the cathode rays, acted as a pseudo-cathode, and itself emitted cathode rays. It was further pointed out that instances of the penetration of metals by molecules were not unknown ; to wit, the passage of hydrogen through hot palladium and platinum, the squeezing of water under hydraulic pressure through gold and other metals, and the gradual penetration of lead by gold when discs of the two are in close contact. 1 How- ever, these explanations were not very satisfying, nor did 1 The experiments of Roberts-Austen on the interdiffusion of metals were being conducted about this time (see P.T. 1896). CATHODE RAYS 7 they prove to be necessary, for about this time new evidence began to accumulate in favour of the charged particle theory. Sir J. J. Thomson came to the conclusion that the velocity of the cathode rays was appreciably less than that of light ; and the French physicist Perrin in 1895, by catching the rays in a Faraday cylinder, demonstrated that they carried a charge of negative electricity with them. But though Perrin's experiment carried conviction to the majority, it- was not regarded as conclusive by the extreme ether-wave supporters, who looked on the electrification as merely To Electroscope FIG. 5. J. J. Thomson's modification of Perrin's apparatus for proving that cathode rays are negatively charged. revealing the presence of electrified particles which were an accidental accompaniment of, and were not essentially connected with, the cathode waves. But J. J. Thomson in 1897, by suitably modifying Perrin's apparatus, showed that when a magnet was used to deflect the cathode rays the negative electrification followed exactly the same course as the rays which produced the fluorescence on the glass (Fig. 5). He further succeeded in deflecting the cathode rays by an electric field : Hertz had tried this experiment fourteen years previously, and had failed because the pressure of his gas was too high, the gas con- ducting sufficiently well to mask the effect of the electric field. After these results it could hardly be doubted that the cathode rays were negatively charged particles, and the objections to this view, on the score of their penetrating 8 X RAYS ability for solids, were finally set at rest by J. J. Thomson in 1897-1898. In a series of experiments, which for ingenuity and insight have rarely been equalled, Professor Thomson, by deflecting the cathode rays in magnetic and electric fields of known strength, was able to infer the size of the charged particles ; and later, to deduce their mass, velocity and electric charge. (See p. 14.) He showed that the cathode rays were neither atoms nor molecules, but something far smaller : the mass of each of the particles proved to be about T ^Voth part of the smallest mass hitherto recognised by chemists that of the atom of hydrogen. Their nature depends neither on the nature of the cathode nor on that of the residual gas in the discharge tube. The charge is invariable, and agrees with that carried by the hydrogen ion in liquid electrolysis : the velocity depends upon the electric force which is applied to the tube ; the speeds are found to be an appreciable fraction of the speed of light, 1 in some cases as much as one-third. Thus the cathode rays proved to be neither ethereal waves nor ordinary material particles, but bodies of sub-atomic size moving with prodigious velocities, a state of things so nearly realising Newton's long-abandoned conception in his corpuscular theory of light that J. J. Thomson called the small particles which constitute the cathode rays, corpuscles. Johnstone Stoney had previously suggested the name electron for the electrolytic unit, or atom of electricity, and the suitability of the expression for the cathode rays was at once recognised : both terms have since come into common use. It is perhaps difficult to realise the disproportion 2 between the size of an atom and the size of an electron : the two have been aptly compared to a fly in a cathedral, or a speck of dust in a room ! Wehnelt Cathode. If the cathode is constructed of a strip of platinum which can be raised to a bright red heat, 3 and on which is mounted 1 The velocity of light is 186,300 miles or 3 X 10 10 cms. per sec. 2 1 : 100,000 in linear dimensions ; 1 : 10 15 by volume. 3 By means of an independent current. CATHODE RAYS 9 a speck of lime 1 (see Fig. 6), all the ordinary phenomena of a discharge tube can be reproduced by means of quite small potentials 100 volts or less between the cathode and anode. The hot lime emits torrents of corpuscles and the pencil of cathode rays, owing to their low speed (see p. 12), produces vivid luminosity in the rarefied gas. The velocity of the rays is proportional to the square root of the potential applied : the velocities are of the order of T V^h of those in Platinum Srrtp Electrified PlaJe FlQ. 6. A discharge tube with a Wehnelt cathode, displaying the repulsion of the cathode rays by a negatively charged plate. an ordinary discharge tube ; e.g. with 50 volts, the velocity = 4- 2 x 10 8 cms. per sec. The magnetic and electrostatic deflection of these rays can be strikingly demonstrated owing to their relatively small velocity. It is possible with the hot-lime cathodes to send very large currents through the discharge tube currents of T ^ ampere and more are readily attained. With high cathode temperatures and voltages of 200 or 300 on the tube it requires precautions to prevent the discharge growing into an arc 2 with the consequent destruction of the cathode. 1 Or any of the alkaline earths (a speck of sealing-wax, ignited in situ, is very convenient). See Wehnelt (P.M. July 1905). 2 A water resistance in the potential circuit serves to prevent this. 10 X RAYS Transmission and Absorption of Cathode Bays. Lenard (Wied. Ann. 1895) showed that for fast- moving cathode rays, the extent of the absorption in different sub- stances is roughly proportional to the density. The pene- trating power of a cathode ray varies very greatly with its speed. The highest speed rays, which move at the rate of about 10 10 cms. per sec., can only penetrate 2 or 3 mms. of air at ordinary temperature and pressure. The fastest /3 rays from radium are cathode rays which have no more than about three times this speed, and yet their range in air is nearly 100 times as great. The range is, of course, increased by lowering the pressure of the gas, as the mole- cules are not so closely packed, and the cathode rays suffer fewer encounters with the atoms. During its journey, the cathode ray loses velocity both by ionising and by deflection. So long as its speed remains high, it pursues a fairly even course ; as it slows down, it becomes more and more liable to deflection by the encoun- tered atoms, until finally it loses so much energy that it becomes undistinguishable as a cathode ray. Thus a fine pencil of cathode rays gradually becomes fuzzy and scattered. Some of the cathode rays are actually swung completely round by the surface atoms, and so may be " reflected " with velocities up to the original velocity. The more obliquely incident the rays, the greater the number " re- flected." In regard to the transmitted rays, Whiddington (P.E.S. 1912) has recently shown experimentally the truth of a relation deduced theoretically by Sir J. J. Thomson some years ago. He finds that the maximum velocity ( V d ) with which a cathode ray may leave a material of thickness dis given by F. 1 - F/ = *,' where F is the initial velocity of the cathode ray and a is a constant (2 x 10 40 for air ; 732 x 10 40 for Al ; 2540 x 10 40 for gold ; all in cm. -sec. units). This fourth-power scatter- ing law holds also for Ra /3 rays, except the very fastest. Whiddington was unable to trace any simple connection between the value of a and the atomic weight or density of the material. CATHOBE RAYS 11 Cathode rays lose their speed very quickly in passing through solids, and thin metal leaf has to be used in experi- mental work. 1 The maximum thickness of aluminium or glass which transmits high-speed cathode rays to any appreciable extent is about 0-0015 cm. (Cf. p. 49.) Heating Effects produced by Cathode Bays. The bulk of the energy of the cathode rays is dissipated as heat when the rays strike an obstacle. A simple cal- culation shows that if in a tube of moderate vacuum the current carried by the cathode rays is a milliampere (10~ 3 amp.), the energy given up by the rays per minute is of the order of 100 calories. 2 Now a milliampere is only a moderate current ; as will be seen later, currents up to 50 or 60 milliamperes and even more (with momentary discharges) obtain in practice. No target can withstand such currents for any length of time if the rays are concentrated by using a concave cathode. Platinum may be fused, diamonds converted into coke ; even tantalum and tung- sten 3 with melting points in the neighbourhood of 3000 C. can be rendered molten. Owing also to the low pressure, most metals can be vaporised with ease. The heating effects reach a maximum at a certain pres- sure, which is not very low, and are not so marked in very high vacua. lonisation produced by Cathode Rays. A cathode ray has the property of ionising a gas, i.e. of rendering it electrically conducting. The ionising power is 1 Al, Cu, Ag, Sri, Pt, and Au can be got in the form of thin leaf. (See Kaye and Laby's Physical Constants, p. 35.) 2 The energy E = \i . . v 2 , where i is the current, v is the velocity of the rays, m is the mass of and e the charge on each ray. If E is expressed in calories per min., i in milliamperes, and v in cms. per sec., then E = 4. lQ~ 18 iv 2 . This assumes that all the cathode ray energy is turned into heat. 3 Von Wartenberg (1907) determined the melting point of tungsten by means of a cathode-ray vacuum furnace. He used a concave Wehnelt cathode (p. 8) for the purpose. See also Tiede (1913) for an account of a cathode-ray furnace with a water-cooled cathode. 12 X RAYS especially conspicuous with the slower rays the ionisation per centimetre of path was in fact found by Glasson (P.M. 1911) to vary approximately as l/( velocity) 2 . The faster rays have the greater energy, it is true, but they do not begin to ionise to any great extent until their velocity has dropped. 1 A cathode ray (like the a ray) ionises most towards-the end of its path, until finally it loses so much energy that it can no longer ionise, and ceases to be dis- tinguishable as a cathode ray. The strong ionisation is responsible for the luminosity which cathode rays produce in the residual gas of a dis- charge tube a luminosity which at higher pressures is displayed as the outer negative glow bounding the cathode dark-space (see p. 2), and at lower pressures lights up the path of the rays themselves (see p. 34). The luminosity of the track reaches a maximum at a certain pressure ; as the pressure is further reduced, the rays gradually become faster and less luminous, but simultaneously their power of exciting fluorescence in the glass walls of the tube increases, until finally the rays become quite invisible and are mani- fested solely by the fluorescence on the glass. Fluorescence produced by Cathode Rays. An ordinary X-ray tube affords abundant evidence of the fluorescing properties of glass subjected to cathode rays. Crookes found in 1879 that glass which had suffered pro- longed bombardment by the rays fatigued and lost a good deal of its fluorescing ability. Most of the fatigue is only temporary, but a portion is very permanent. Crookes found, for instance, that complete recovery was not brought about even by fusion of the glass ; and Campbell- Swinton (P.R.S. 1908) refers to a tube in which the fatigue persisted for more than ten years. Swinton showed that the fatigue is purely a surface effect, and is removed by grinding away the surface of the glass. He found that the thickness which had to be removed for this purpose was always about 015 mm. 1 The total ionisation of the faster ray is, of course, greater than that of the slower. CATHODE RAYS 13 There are many substances which afford striking and beautiful examples of the fluorescing ability of cathode rays. Among those which are useful in practice are barium platino-cyanide (a material of which fluorescent screens are usually made), the mineral willemite (a silicate of zinc), zinc blende (sulphide of zinc), and kunzite (a lithium felspar)'. The fluorescing power of a cathode .ray increases with its velocity, and does not seem to be possessed by the very slowest rays. Rays as slow as those from a Wehnelt cathode are, however, capable of causing fluorescence. Magnetic Deflection of Cathode Rays. When a charged particle (mass m, charge e) is projected along a line of magnetic force, it continues to move along it ; but if it is projected (with velocity v) at an angle to the magnetic field (of strength H), the deflecting force acting upon it is Hev sin 0. Hence, if p is the radius of curvature of the resulting path, Hev sin = mv 2 /p mv p = H^hTe> which represents a helix. So na. 7. that, in general, the effect of a magnetic field on the path of a moving corpuscle is to twist it into a helix wound on a cylinder with the lines of force as axis. With the slow cathode rays given out by a Wehnelt cathode (p. 8), the helix can be beautifully demonstrated : with strong fields, the helix becomes so long and attenuated that the rays appear to follow the lines of force. In the particular case when the cathode ray is projected at right angles to the magnetic lines of force, sin 0=1, and ^o = mv/He. Thus in a uniform magnetic field, the path of the particle is a circle in a plane at right angles to the magnetic force, i.e. the particle is bent away in a direction at right angles 14 X RAYS both to the field and to its former direction. 1 The extent to which cathode rays are bent by a magnetic field thus depends on the strength of the field, on the speed of the particles, and on the quantity denoted by e/m. As to the speed, that is related directly to the potential which is applied to the tube ; the velocity is, indeed, roughly pro- portional to the square root of the potential or alternative spark-length (see p. 100). Electric Deflection of Cathode Bays. If an electric field X is acting at right angles to the direction of projection of the cathode ray, the force on it is Xe in the direction of the field. Thus, as above, the radius of curva- ture is mv z The corresponding expression for the magnetic deflection was mv C&rhode PI a,res for Electric Field L Anode FIG. 8. J. J. Thomson's apparatus for measuring e/m of cathode rays. An electric and a magnetic field are contrived so that their effects on a beam of cathode rays balance each other exactly, in which case the fluorescent spot produced by the cathode rays remains undeviated. If we contrive things so that the magnetic and electric deflections are equal and opposite, we can, at once, from a knowledge of the fields, derive both the velocity and e/m 1 The following mnemonical rule is convenient for remembering in which direction a cathode ray is deflected : If the magnetic field H (i.e. the direction in which a N-seeking pole would move) is upwards towards the Head, and the cathode Ray is moving horizontally towards the Right hand, the mechanical Force on the ray is horizontally towards the Front, CATHODE RAYS 15 for the cathode ray ; for v = X/H, and e/m = X/(pH 2 ). It was in this way that J. J. Thomson first arrived at the nature of the cathode rays (p. 8). (See Fig. 8.) Braun Tube. A practical application, due to Braun, of the bending of cathode rays under magnetic or electric force has come into use in electrical engineering for the purpose of studying the wave-form of rapidly changing alternating currents. In the Braun tube, a narrow pencil of cathode rays is received on a fluorescent screen, and is subjected en route to both a magnetic and an electric field. The two fields are at right angles, and are both actuated by the alternating current. The cathode rays, having practically no inertia, are able to follow the most rapid vagaries of the fields, and so trace out on the screen a pattern, from which the wave-form can be deduced. Magnetic Spectrum of Cathode Rays. Each interruption of the primary current in an induction coil produces a small train of strongly-damped oscillations in the discharge (p, 62). Thus the potential on the tube is intermittent, and the result is a stream of cathode rays with a variety of speeds, each peaklet on the oscillatory potential curve producing a group of uniformly fast cathode rays, of which the speed diminishes with successive oscilla- tions. Accordingly, if the cathode rays are subjected to a mag- netic field, the different groups are differently deviated the greater the speed, the less the magnetic deviation. Thus a slit of cathode rays, when allowed to fall on a fluorescent plate, produces in a magnetic field a number of bright lines or bands which go to make up a " magnetic spectrum " an appearance first noticed by Birkeland (C.R.) in 1896. The brightness of each band is a measure of the number of rays moving with the same speed : the displacement of the band from the undeflected position is inversely proportional to the velocity, and directly proportional to e/m and the strength of the field. 16 X RAYS If the gas pressure in the discharge tube is not very low, the bands may be numerous (30 or more), but if the pressure is lowered, the oscillations are more strongly damped, and so the lines become fewer, group themselves more closely, are less deviated as a whole (the cathode rays being faster), while the least deviated line becomes the brightest ; thus Zero Position. Pressure 0-008 mm. Pressure 0-004 mm. FiO. 9. Examples of magnetic spectra of cathode rays. most of the cathode rays now possess the maximum velocity. Fig. 9 shows two magnetic spectra obtained by Birkeland at different pressures. Beatty (P.R.S. 1913) found that with a potential of about 60,000 volts on the tube, the main stream of cathode rays had a velocity corresponding to about two-thirds of the potential as given by a spark-gap. A spectrum of rays can also be produced by an electric field, in which case the deviations are inversely proportional to the square of the velocity. A tube driven by a Wimshurst machine or a battery of cells does not yield a magnetic spectrum, but only a single bright line, which is evidence of the fact that all the cathode particles have the same velocity. Constants of Cathode Rays. Measurements of the speed of cathode rays have been made by various experimenters ; the rays in an ordinary CATHODE KAYS 17 discharge tube have velocities ranging from about 10 9 to rather more than 10 10 cms. per sec., i.e. from one- thirtieth to one- third of the speed of light. The latest determinations of e/m (see p. 14) for the slowest rays give a value of I'77xl0 7 expressed in electro- magnetic units. Measurements of the ionic charge make e=l-57xlO~ 20 E.M.U., so that, for small velocities, m is 8'8xlO~ 28 gramme. Theory indicates that electrons owe all their mass to their velocity, and that for cathode rays moving, for example, with one-third the speed of light m would have a value about 6 % greater than the above. Now, according to the best authorities, the mass of the hydrogen atom is 1*6 x 10 ~ 24 gramme, so that the number of electrons equal in mass to the hydrogen atom is about 1830. Further constants will be found on p. xv. The Ubiquity of Electrons. Since their discovery in so artificial a source as a vacuum tube, electrons have been found literally to pervade the universe. Relatively low-speed electrons are emitted in many chemical reactions and by metals when exposed to light, especially ultra-violet. High-speed electrons with velocities almost up to that of light constitute the /5 rays of radium and the radioactive substances : the alkaline metals (at any rate K and probably Rb) also emit corpuscles. They are ejected in abundance from hot bodies, markedly so from the alkaline earths (lime, baryta, etc.) with velocities and in amounts depending on the temperature. Without doubt they play a part in cosmical physics : the most recent explanations of the aurora or northern lights regard them as due to enormously fast electrons ejected by the sun, which are collected and guided in long spirals (see p. 13) to the polar latitudes by the earth's magnetic lines of force. 1 They there ionise and cause luminosity in the 1 Aurora occur most abundantly not at the poles, but at about latitude 68. This requires for the electrons a velocity closely approaching that of light a velocity even greater than that possessed by the fastest known J3 particles from radium. B 18 X KAYS upper attenuated regions of the earth's atmosphere, just as they do in a vacuum tube. But whatever their origin, electrons have always been found to maintain their invariable* and indivisible char- acter : they carry the unit of electricity, and can indeed be regarded as the ultimate fundamental carrier of negative electrification. The Electron Theory of Matter. Rutherford's theory of the constitution of matter, which is a development of that of J. J. Thomson, regards an atom as built up of a minute nucleus of positive electricity 1 (about 10" 12 cm. in diameter) surrounded by an inner cluster of negatively charged electrons which rotate round the nucleus, and an outer group of electrons which also rotate and are less rigidly attached. The total negative charge of the electrons is equal to the positive charge of the nucleus. The outer electrons, by their number and arrangement, are responsible for the chemical and physical properties of the atom : the inner electrons have influence only on the pheno- mena of radioactivity. This explains why physical and chemical behaviour do not go hand in hand with X and y-ray phenomena. Present theory indicates that the number of corpuscles in an atom is equal to the " atomic number " of the element, i.e. the number of the element in the periodic table (see p. 227). As positive electricity has never been found to be associated with bodies less than atoms, it would appear that the atom owes most of its mass to its positive nucleus, which is capable of deflecting both a and ft particles out of their paths. The electron theory of matter has been elaborated by Lorentz and others, and extended to many departments of physics. For instance, the phenomena of magneto- optics show that electrons are intimately concerned with the spectrum lines. An electronic theory of magnetism has been developed by Larmor (1897) and Langevin (1905). 1 Rutherford (P.M. 1911). See also Bohr (P.M. July, Sept. and Nov. 1913, Sept. 1915), and J. J. Thomson, Engineering, March 21, 1913. POSITIVE RAYS 19 In the case of solids, there are supposed to exist un- attached and wandering electrons interspersed between the molecules. These can be ejected by ultra-violet light or heat : they are the important agents in thermo-electricity, and in the conduction of electricity and heat ; a good con- ductor, for example, is one which contains many of these free electrons. The electron theory has, among other things, led to important deductions concerning the specific heats of metals at low temperatures a subject to which Nernst and others have lately given attention. For an account of the electron theory see Lorentz, Theory of Electrons, 1909, and N. R. Campbell, Modern Electrical Theory, 2nd ed., 1913. POSITIVE RAYS. Though the ordinary cathode rays are the most con- spicuous of the rays existing in a discharge tube, there are others also present. As long ago as 1886, Goldstein, by perfor- ating the cathode of a discharge tube, observed that a stream of rays travelled through the tube in the reverse direction to the cathode rays. To these rays he gave the general name of Canal- rays (Kanalstrahlen). Wien showed that the Canal rays were deflected in the opposite direction to cathode rays when subjected to a magnetic force, and he came to the conclusion that they must consist of positively charged particles. The deflection detected in this case, however, was small compared with that obtained with cathode rays, 1 but the deflection in the electric field was of the same order of magnitude for both positive and negative particles. This pointed to the fact that the positive particle had greater mass than the cathode or negative particle ; Wien, in fact, found that the mass of the positive particle was of the same order of magnitude as that of a hydrogen molecule. The existence of the two kinds of rays in a discharge tube can very easily be shown by the different colours they produce in lithium chloride. This substance fluoresces blue 1 Positive rays require, roughly speaking, magnetic fields of 1000 gauss or more, i.e. at least forty times as strong as are needed to deflect cathode rays. 20 , X RAYS under the' action of cathode rays, and red under the action of positive rays, so that a small glass bead coated with lithium chloride and placed in a suitable position between the electrodes in the tube will appear red on the side towards the anode, and blue on the side facing the cathode. This bead can be utilised to explore the region between the electrodes in order to determine in what part of the tube the positive rays have their origin. Starting with the bead at the anode, we should find that it does not appear red on the side facing the anode until we arrive at the boundary between the negative glow and the Crookes dark-space (p. 2) ; it continues to fluoresce red throughout the Crookes dark-space. The amount of fluorescence in various parts of the dark-space, however, shows that most of the rays start from the boundary of the dark-space. The positive rays have strong ionising, fluorescing, and photographic actions. They cause soda glass to fluoresce a dull green ; willemite, a bright green : in both cases, the effects are much inferior to those produced by cathode rays. The positive rays show strong pulverising properties, and roughen or disintegrate any surface against which they strike. During the last few years the whole question of the electric discharge from the point of view of positive electricity has been taken up by Sir J. J. Thomson. He has shown that there exist in the tube high-speed atoms and molecules of the gases present, some positively charged, some nega- tively, and some uncharged. In no case has a positive ray been detected whose mass is smaller than that of the hydrogen atom ; a positive electron, if such exists, has hitherto eluded search. The velocities of these positive particles are in the neighbourhood of 10 8 cms. per second (the fastest have a speed of about 2 x 10 8 cms. per second) ; this is of the order of 1000 times the ordinary velocity of mole- cules as calculated from the kinetic theory of gases (p. 77). It is found that there are many more of these high-speed charged particles (or ions) moving towards the cathode than from it. 1 If a hole is made in the cathode, the positive 1 These latter "retrograde rays" travel with and among the cathode rays, but can be detected when the cathode rays are removed by a magnet. POSITIVE RAYS 21 rays stream through it and form the Canal rays of Goldstein ; and in this region, where they are separated from the cathode rays, they can be independently investigated. Professor Thomson received a very fine pencil of these rays on a photographic plate, and en route subjected them simultaneously to the action of magnetic and electric fields, the magnetic field deflecting the rays at right angles to PI A res for ElecrricReld Anode FIG. 10. J. J. Thomson's apparatus for measuring e/m of positive rays ( Kanalstrahlen) . the deflection caused by the electric field (see Fig. 10). 1 It is found that only a small portion of the beam is deflected, the main part goes on unaffected by the deflecting forces. If x and y are the deflections of a particle due to the action of the electric and magnetic fields respectively, and e, ra, and v the charge, mass, and velocity of the particle, then we have x = A mv e mv 2' where A and B are constants depending upon the strengths 1 In the corresponding cathode-ray experiment (p. 14) the magnetic force and electric force were parallel. 22 X RAYS of the electric and magnetic forces, and the distances the rays have to travel from the time they enter the field of force until they reach the photographic plate. The equation to the trace on the photographic plate becomes 2 jg2 e x m This is the equation of a parabola, so that we have on the plate a series of parabolas representing the loci of particles which have a constant value of e/m. (See Fig. 11.) Different particles register different loci on the plate, and for each locus m has a definite value, so that the method affords a means of determining accurately the atomic weights of substances which are present in the tube. The system of curves obtained on the screen depends, of course, upon the nature of the gas in the tube. Most positive rays carry only one ionic charge (equal to that carried by the cathode rays), though with some elements 2, 3, and up to 8 charges have been found. By this method, Professor Thomson has been able to de- termine the atomic weights of numerous elements, 1 and has, indeed, discovered substances as yet unknown to the chemist. When, for example, nitrogen is put into the tube there are seen on the photographic plate parabolas corresponding to N + (an atom of nitrogen with one charge), N ++ , N 2 + , N 3 + . A very interesting case is afforded by marsh gas (CH 4 ), which yields lines corresponding to the com- binations CH, CH 2 , CH 3 , and CH 4 . The great advantage FIG. 11. Photograph obtained by J. J. Thomson of parabolic loci of positive particles subjected to magnetic and electric fields at right angles to each other. The differ- ent traces represent different sub- stances. In the figure, the magnetic deflections are vertical. The mag- netic field was reversed half-way through the exposure, so that both halves of each parabola are re- corded. 1 e/m for the hydrogen molecule carrying a single charge is 1C 4 electro- magnetic units. POSITIVE RAYS 23 of the method lies in the fact that only a very small quantity of the gas investigated is required in the tube ; and, in addition, a combination of atoms in an unstable state has only to exist for a minute fraction of a second in the tube in order to allow it to be recorded on the photographic screen. For a fuller account of this ingenious method, see J. J. Thomson, Positive Rays (Longmans, 1913). The positive rays are important from the point of view of the X-ray worker, in that by their bombardment of the cathode, they liberate the cathode rays. Furthermore, the positive rays are responsible for the positive electrification which the inner surface of the glass walls of an X-ray bulb always assumes (p. 32). and which serves to heighten the fluorescence of the glass by attracting the secondary cathode rays from the antic athode. The positive rays doubtless also play a prominent part in the action of valve- tubes (p. 70). It may be noted that in an X-ray bulb many of the positive rays strike and disintegrate the glass walls round the cathode. The "Ionics" of an X-ray Tube. Before leaving the subject of positive and cathode rays it is convenient to anticipate matters slightly and set forth simply the " ionics " of an ordinary X-ray bulb. The process of evacuation of an X-ray tube is deliberately never completed, but an appreciable amount of gas is always left, the pressure ranging from say 0-Qill to 0*010 mm. of mercury. The number of ions (see p. 92) present in this residual gas is normally very small. But when a high potential is applied to the terminals of the tube, the effect is to accelerate the motion of these ions and cause them to produce many more ions by collision with the gas molecules. The electric field also serves to direct the positive ions towards the cathode which they bombard and cause to emit cathode rays. The electric field further impels these cathode rays towards the anticathode upon colliding with which their velocity is suddenly changed and X rays are generated. CHAPTER III. X RAYS. The Discovery of X Kays. 1 We have dealt above in some detail with many of the features which cathode rays possess ; we have, however, made no more than mention of their most striking property of all that of generating X rays. In the autumn of 1895 Professor Wilhelm Konrad Rontgen of Wiirzburg, Bavaria, discovered, it may be said almost accidentally, the rays which now bear his name. During the course of a search for invisible light rays, he turned on a low-pressure discharge tube, which for the purpose was completely enclosed in stout black paper, and to his surprise noticed that a fluores- cent screen lying on a table some 3 metres or so distant shone out brightly. The light-tight cover precluded any possibility of the effect being due to ordinary ultra-violet light ; there was evidently some curious radiation coming from the tube. If obstacles were interposed, Rontgen found that they cast shadows on the screen ; and in this way he traced back the unknown or " X " rays to their source, which proved to be the region of impact of the cathode rays on the glass walls of the tube. Further investigation revealed the fundamental fact that Rontgen or X rays are produced whenever and wherever cathode rays encounter matter. It was imagined by many that X rays were present in the original cathode ray beam, and were obtained by mere subtraction. But this was dis- proved by the discovery that when the cathode rays were magnetically deflected, the source of the X rays also moved. The experiment also put out of court the notion that X rays were due to the impact of particles of metal from the cathode. 1 See also p. 249. DISCOVERY OF X RAYS 25 But the fascinating feature of the new rays was their extraordinary ability to penetrate many substances quite opaque to light. The degree of penetration was found to depend on the density ; for example, bone is more absorbent than flesh, and if the hand is placed in the path of the rays, the bones stand out dark against the flesh in the shadow cast on a fluorescent screen. Rontgen at once appreciated the immense significance of his discovery to the surgical profession, and communicated his results to the Physico- Medical Society of Wiirzburg in November 1895. 1 It was soon ascertained that X rays affected a photo- graphic plate, 2 could not apparently be refracted or reflected (see p. 168), and, unlike cathode rays, were not bent by a magnetic or electric field, 3 a result which shows that the X rays do not carry a free electric charge/ In 1896, J. J. Thomson, Hurmuzescu, Benoist, Dufour, and others found that Rontgen rays shared with cathode rays (and ultra- violet light) the property of ionising or imparting temporary electrical conductivity to a gas, which ordinarily is a nearly perfect insulator. Before considering in any detail the advances that have been made in the various branches of the subject, it will probably be useful first to recount briefly the essential particulars of the working of a simple X-ray equipment. A BRIEF ACCOUNT OP THE PRODUCTION OF X RAYS. An X-ray Bulb. When a current of electricity from a Ruhmkorff induction coil is sent through an X-ray tube, a pencil of cathode rays from the concave cathode is focussed on the target or anti- 1 See also L 'Eclair Elect. 6. 241. 1896. For an account of Rontgen's later work, see Berl. Ber. 1897, and Ann. Phy. Chem. 1898. Rontgen's three memoirs are translated in the Electrician (Jan. 24, 1896 and April 24, 1897) and A.Rt.R. (Feb. 1899). 2 The inexplicable fogging of unopened packets of photographic plates in the neighbourhood of a Crookes tube was engaging the attention of more than one English physicist at the time of the discovery of the X rays. 3 Walter (A.d.P. 1904) used magnetic fields up to 19,000 gauss. Paschen (P.Z. 1904) similarly exposed Ra y rays to fields of 30,000 gauss. 26 t X RAYS cathode, the surface of which is inclined at 45 to the rays, and is usually made of a metal of high atomic weight, such as platinum (Fig. 12). An additional anode is usually pro- vided, but is not indispensable. The anode and cathode are generally of aluminium. From the point of impact of the cathode rays on the anticathode, X rays are given out in all directions. The anticathode tends with continued use to become very hot, and is often either made massive Anode V Anticathode / Cathode FIG. 12. A simple type of focus bulb showing the various electrodes. or cooled in some fashion. The pressure of the gas in an X-ray tube becomes lower with use, and a device for " soften- ing " the tube (i.e. raising the pressure) is therefore usually provided. The higher the pressure, the less the potential required to drive the tube and the less penetrating the X rays ; both the X rays and the tube are often termed " soft " if the pressure is high. The lower the pressure, the " harder " are the rays. In the X rays from any particular tube there are many qualities present ; this is shown by the fact that rays which have traversed one thickness of material are more penetrating to a second. X rays are invisible and do not make glass fluoresce ; the pale green hemisphere of fluorescence on the bulb is due to " reflected " cathode rays from the anticathode striking the glass walls. That this is so is shown by the distorting action of a magnet on the boundary of the fluorescence. PRODUCTION OF X RAYS 27 An Induction Coil. An induction coil is merely a device for transforming a lostpotential current, such as is yielded by a battery of a few cells, into a high-potential current of the kind suitable for driving an X-ray bulb. An induction coil consists essentially of a cylindrical iron core round which is wound a coil of stout insulated wire ; this coil, which is known as the primary, consists of relatively few turns. Outside this 'is the secondary coil consisting of many thousands of turns of finer wire carefully insulated. Fig. 13 shows diagram- Sp&rk Condenser FIG. 13. Diagrammatic representation of an induction coil. matically the various parts of a small coil. A hammer- break interrupter is shown in the primary circuit, and a condenser, usually mounted in the base of the coil, offers an alternative path to the break. The primary circuit is joined to a suitable battery ; and the object of the interrupter is to make and break the current in rapid succession. The consequence of this is at every " make " to induce in the secondary coil a momentary current, and at every " break " an equal momentary current in the opposite direction. But in X-ray work it is important that the current through the X-ray tube should be v all in one direction, and herein lies the chief function of the condenser. When the circuit is made, the condenser takes and stores the first rush of current, which therefore grows relatively slowly and magnetises the 28 X RAJS core ; at break, however, the condenser discharges its elec- tricity through the primary circuit with great rapidity and demagnetises the core. The induced potentials in the secondary are accordingly much feebler at make than at break : the currents resulting from the former are known as " reverse " or " inverse " currents, and in a good coil are nearly suppressed; Thus the sparks which pass between the terminals of the secondary circuit are due chiefly to the break and only pass one way. The power of a coil is often designated by the length of its longest spark, e.g. a 6-inch coil. The iron core serves to increase the number of lines of force through the coils. The condenser is made of alternate layers of tinfoil and paraffined paper. The hammer-break consists essentially of a steel strip on which is mounted a piece of soft iron ; this is attracted by the core when the current passes, and so breaks circuit between two platinum studs in the primary circuit. The spring is thus caused to vibrate backwards and forwards like the hammer of an electric bell, and so alternately makes and breaks the primary current. 1 An extended account of the induction coil is given on p. 53 et seq. 1 In a modern X-ray coil the hammer-break is rarely fitted, but terminals are provided for connecting up to an independent mercury or other type of break. CHAPTER IV. -I AN X-RAY BULB. Early X-ray Tubes. The vacuum tube with which Rontgen made his famous discovery in 1895 was pear-shaped, with a flat disc for cathode mounted in the body of the bulb at its narrow Anode GxThode I Su pporf FIG. 14. Type of tube with which Rontgen discovered X rays. The cathode V rays impinged on the broad end of the tube. end ; the anode was in a small side tube (Fig. 14). 1 The cathode rays impinged on the large end of the bulb, pro- ducing vivid fluorescence. This pattern of tube was widely copied, but it was soon found that it did not survive many of the prolonged exposures which were necessary to secure radiographs of any value. Moreover, owing to the large area of emission of the rays, the photographs were always blurred and somewhat indistinct. Experimenters set about 1 In another early form of X-ray bulb used by Rontgen, the anode consisted of a large ring in the body of the bulb. 30 X RAYS to find ways and means of prolonging the life of the tube, ot shortening the exposure, and of improving the definition. Under the impression, then prevailing, that active fluores- cence was essential for the genesis of the X rays, 1 various workers, about 1897, constructed tubes of fluorescent glass (e.g. uranium and didymium glasses) with the idea of en- hancing the output of the tube ; it was, however, found later that the fluorescence was quite immaterial. Campbell-Swinton in 1896 modified Bontgen's design of tube by inserting a sheet of platinum obliquely in the path of the cathode rays. The improvement was considerable, though the radiographs were still lacking in sharpness, and the exposures unduly protracted. Pr.T&rgef FIG. 15. Tube used by Crookes to display the heating effects of focussed cathode rays. The same year Professor H. Jackson of King's College/ London, turned to account a former discovery of Sir Wnn Crookes, and replaced the flat cathode by a concave oni. Crookes had shown in 1874 that a hollowed-out cathode brought the cathode rays to a focus, and five years later actually constructed a tube with a plate of platinum at the focus to display the heating effects of the rays (Fig. 15). The tube must have given out X rays in abundance, but they remained unnoticed. Professor Jackson mounted the platinum target at 45 to the rays (Fig. 16) ; in essential respects his tube agreed with that of Crookes. The new 1 It may be recalled that the late Henri Becquerel, at the suggestion of M. Poincard, was led to investigate whether X rays were an invariable accompaniment of phosphorescence in general. Among the substances he tried were uranium salts : the result was the discovery of radioactivity in 1896, two months after the discovery of X rays. AN X-RAY BULB 31 focus tube was a vast improvement on its predecessors ; the exposures were shortened enormously, and, owing to the small area of emission, the resulting photographs showed wonderful sharpness and detailJC It is remarkable how slight the subsequent changes have been ; many thousands of X-ray tubes have been made, but with one notable exception (p. 43) the design agrees essentially with that of fifteen years ago. It may be said that the X-ray bulb has scarcely kept pace with the very extensive improvements that have been made in the rest of the X-ray equipment. There is no gainsaying the fact Anode zxnd AnricaJtiode FIG. 16. Jackson's first focus tube, employing focussed cathode rays. that even now X-ray tubes are prone to be fickle, and it is scarcely possible to guarantee their behaviour. A bulb will be perfectly satisfactory one day, and yet refuse to work reliably the next ; x and of two bulbs apparently precisely similar, one may work well for months, the other may break down within a few days. Many X-ray workers take the precaution of resting a favourite bulb occasionally ; a bulb is often improved by being allowed to lie idle for a few weeks. In general, a large bulb is better than a small for passing a heavy current (see p. 72). Some makers have accordingly constructed monster bulbs, which have, however, little to commend them. Indeed, if certain difficulties could be sur- mounted, quite small bulbs would possess many advantages. In all the earliest tubes, the cathode was mounted in the body of the bulb, but by the end of 1896 it was withdrawn just within the neck of a side tube a design typical of all later makes, and one which conduces to greater steadiness 1 Pue probably to variations in the gas-pressure. 32 X RAYS and hardness l (see p. 73). In Jackson's bulb and its pre- decessors the target served also as anode ; we find an auxiliary anode introduced in a tube by Gundelach in 1896. Both forms persisted for some years, but nowadays the auxiliary anode usually finds a place. About 1902, makers vainly sought to improve the efficiency of their tubes by coating some or all of the electrodes with radioactive material. Since then they have devoted most of their attention to the design and .material of the anticathode. THE ELECTRODES OF AN X-RAY BULB. The electrodes are fixed to stout wires or rods which, for ease of manufacture and repair, are invariably mounted in side tubes projecting from the main bulb. It was found as early as 1896, that the discharge is materially steadied by completely sheathing the supporting wires with glass tubes. This largely serves to check the tendency for the discharge to pass along the walls of the tube, more especially at low pressures, and checks the "sputtering" which is pronounced with wires and points. The glass walls become highly charged, negatively in the region of the cathode (p. 74) and positively in the main body of the tube (p. 23) : these charges, particularly with a blackened tube, may cause the focal-point to wander and lead to. sparking along the glass. The passage of the discharge along the glass walls is not moreover confined to the inside of the tube. Alippi found in 1906 that if a large jet of steam were allowed to play on the bulb, the X-ray output and general fluorescence greatly increased. The effect is probably due to the removal of dust and the surface alkali in the glass with a consequent diminution of conductivity. Local surface electrification is probably responsible for the green wisp-like discharges which can often be seen on the inner surface of a bulb. This is distinct from the flickering due to insufficient voltage on the tube. 1 J. J. Thomson (P.M. 1912) finds that this position of the cathode is also the most favourable for the production of the positive " canal " rays. AN X-EAY BULB 33 The part that surface electrification plays, and the control it possesses over the hardness and steadiness of a discharge is not, I think, generally appreciated. THE ANODE. As remarked above, the modern X-ray bulb is always provided with an additional anode of aluminium which is joined externally to the anticathode (Fig. 12). The precise benefit of the anode is a little doubtful, though in some cases the result of disconnecting it from the anticathode is to soften the tube. C. E. S. Phillips (A.Rt.R. 1902) con- cluded that the auxiliary anode was helpful, probably by electrostatic action, in steadying the discharge. He remarks that the most advantageous position for the anode is behind the anticathode. The auxiliary anode is also probably bene- ficial during the passage of the inverse current which exists with all coil discharges : in these circumstances, the alu- minium anode, rather than the platinum anticathode, tends to act as a temporary cathode, and as aluminium exhibits much less cathodic sputtering (p. 82) than platinum, the walls of the tube are not blackened to the same extent. It id usually stated that the discharge is independent ofV the position of the anode. This is only true if the anode is outside the cathode dark-space : if the anode is within the dark-space, the discharge only passes with difficulty. \ Now, in an X-ray tube, working under ordinary conditions, the cathode dark -space is big enough to enclose the anti- cathode within its boundaries, and the presence of the anode, which is invariably inserted within a confined side tube, is therefore advantageous. There is this, too, to be remembered : the easiest direction for a discharge to cross an unsymmetrical tube is that which makes the less restricted electrode the cathode in other words, that direction which offers the cathode dark-space least obstruction ; the tube runs harder if the dark-space touches the walls. If the design of an X-ray bulb is borne in mind, it will be realised that this property (which is made use of in the various valve-tubes, p. 70) would, in the c 34 X RAYS absence of the confined anode, result in facilitating rather than in retarding the passage of the inverse current. Most workers have experienced this tendency of X-ray bulbs to act as rectifiers, and their refusal, on occasion, to let through the '' break " current at all. THE CATHODE. For the reasons given elsewhere (p. 82), the cathode is made of aluminium, and is mounted just within the neck of a side tube to the bulb. In a focus tube, the cathode is concave. Now, while the normal ejection of cathode rays holds for plane surfaces, it is not the case for concave cathodes except when the pressure is not very low. As the exhaus- tion proceeds, the focus of the rays recedes farther and farther from the cathode, and may reach a distance of something like four or five times the radius of curvature of the cathode : ordinarily the distance between the cathode and anticathode is some two or three times the radius of curvature. The size of the focal spot may vary somewhat capriciously in practice owing to variations in the gas- pressure. The harder the tube the smaller the spot (p. 32). The correct disposition of anticathode and cathode is a matter of some nicety for the maker, who has to be guided mainly by his experience and the hardness at which the tube is to be run. The anticathode is usually mounted a shade out of focus to avoid its premature destruction by fusion, though for radiographic purposes this entails some loss in definition. Some of the earlier X-ray tubes were provided with devices for moving the anticathode to suit the conditions of use. Campbell-Swinton (P.R.S. 1897) found that, at moderate pressures, the cathode rays do not form a solid cone of rays, ' but are condensed into a hollow conical shell. At low pressures, however, the rays are chiefly concentrated along the axis of the cathode. Owing to the ionising effect on the residual gas, this bundle of rays is displayed as a luminous pencil which stretches from cathode to anticathode. The origin of the pencil of rays, which usually is readily dis- AN X-RAY BULB 35 cernible in a soft X-ray bulb, is due to the repulsive effect of the electricity on the walls of the tube adjacent to the cathode. The same effect obtains also with plane cathodes (see p. 74). With a cathode made of a metal tube, a con- centrated pencil of rays emerges from each end along the axis ; such cathodes are sometimes convenient in experi- mental bulbs. It would appear from the work of some experimenters that to keep the cathode cool is of service in diminishing the tendency of the bulb to harden with use. In the Gaiffe- Barret tube, for instance, the cathode is cooled by directing an air blast on its back surface. One may mention here that cathodes made of the electro- positive metals conduce to smooth running of the discharge ; for example, an aluminium cathode faced with calcium metal permits a tube to be run with safety much harder than one with the plain aluminium cathode. This is prob- ably due to the comparative ease with which such metals emit electrons. See also the Coolidge tube (p. 43). THE ANTIC ATHODE. The desiderata in an anticathode intended for modern radiography are : (1) A high atomic weight to secure a large quantity of rays. (2) A high melting point to permit sharp focussing of the cathode rays without fusing the target. (3) A high thermal conductivity to diminish local heat- ing. (4) A low vapour pressure at high temperatures to avoid thermal "sputtering" on the walls (see p. 85). The Atomic Weight of the Anticathode. It was known almost from the first that the heavier metals, or rather those of high atomic weight, make the most efficient anticathodes. Rontgen himself found in 1896 that the rays from platinum are more intense than those from aluminium. Campbell-Swinton, Kaufmann, Roiti, Sir Oliver Lodge, S. P. Thompson, and Langer, all about 1897, did work connecting atomic weight and intensity of radiation. 36 X RAYS These earlier workers used photographic or fluorescence methods of measuring intensities, and consequently most of their observations are of qualitative rather than quanti- tative interest. In some experiments made by Kaye in 1908 l the metals used as anticathodes, some twenty in number, were mounted on a trolley inside the discharge tube (see Fig. 17). By To E&rfh lonis&hon Chamber To Pump FIG. 17. Apparatus for generating and measuring the X rays from different anticathode metals. means of a magnetic control, the trolley could be moved and any metal desired brought under the beam of cathode rays. The tube was provided with an aluminium window 0*0065 cm. thick, and the emergent rays, which thus suffered but slight absorption, were measured by an ionisation method. The discharge was maintained by an induction coil. The experiments showed that there are, in general, at least two classes of X rays given out by an anticathode 2 1 Phil. Trans. Roy. Soc. A, 209, p. 123. 2 The X rays from an anticathode will ordinarily be supplemented by at least two types of soft X rays one produced by the X rays in passing through the glass walls ; the other from the impact of " reflected " cathode rays against the glass and the residual gas molecules. AN X-RAY BULB 37 heterogeneous " primary " or " independent " X rays, and homogeneous X rays characteristic of the metal. The quality and amount of the latter rays are controlled by the nature of the anticathode and the potential on the tube ; if the tube is soft, with many metals the X rays are 125 100 75 cc *o 50 c CD ^25 OBi- OPb Pr 50 100 150 Atomic Weight of 200 250 FIG. 18. Graph connecting atomic weight of anticathode with intensity of " independent " X rays. (Pt =100.) almost wholly characteristic (see p. 125). In some cases, such rays are too soft to penetrate the glass walls of an ordinary tube. However, the aluminium window enabled their pre- sence to be readily detected. Their intensities did not follow any simple atomic weight order for example, the metals of the chromium-zinc group 1 emit radiations very rich in soft and ionising rays. 1 Cr, Mn, Fe, Ni, Co, Cu, Zn. 38 X RAYS To remove the characteristic rays, an aluminium screen 2 or more mms. thick was used, and it was then found that the intensity of the remaining harder " independent " rays increased with the atomic weight of the anticathode ; the two are indeed roughly proportional. 1 Fig. 18 shows the relation for a potential of about 25,000 volts on the tube. Very little change was produced in the relative intensities by increasing the thickness of the aluminium screen the rays from all the metals were, under these conditions, very fairly homogeneous and of the same quality. Thick screens of other metals yielded much the same sort of curve, modified a little here and there. When the potential on the tube was raised the heavy-atomed anticathodes became slightly more efficient ; with a diminished potential the lighter elements somewhat increased their relative intensity values. Suitable Anticathode Metals. The list in Table I. gives the atomic weights, the radiation values, the melting points, and thermal conductivities (where known) of those elements which by reason of their refractori- ness may be regarded as suitable for the anticathode of a focus bulb. The radiation values are for hard rays and are taken from Kaye's experiments (p. 36) ; in some cases the numbers have been obtained by interpolation. The thermal conductivities quoted are at room temperatures ; most metals diminish in conductivity as the temperature rises. The remaining constants are from Kaye and Laby's Physical Constants. The properties of some of the metals are not convenient, and to others the scarcity and price are at present an insurmountable objection. Among the metals which have been commonly employed as anticathodes in radiography are osmium, iridium, tung- sten, tantalum, and. of course, platinum. Platinum, which is most commonly used, has a melting-point none too high for the purpose, sputters badly, and its price, steadily becoming exorbitant, is being instrumental in directing attention to the properties of tantalum and tungsten, metals 1 Gray (P.R.S. 1911) obtained the same result for the 7 rays produced by the impact of radium /3 rays on different elements. AN X-RAY BULB 39 whose chemistry has become familiar through their extensive employment in electric lamps. Neither metal sputters so badly as platinum, 1 both have a very much higher melting point, and but a slightly inferior radiation value, while tungsten has a superior thermal conductivity, thus permitting sharper focussing of the cathode rays. It is only within the last few years that it has been possible to obtain forged pieces of pure, dense and malleable tungsten suitable for the purpose. Tungsten anticathodes have now firmly estab- lished themselves in popular favour. TABLE 1. Metal. Atomic Weight. Density. Intensity of Radiation. Melting Point. Thermal Conductivity. Uranium - (0-16) 238-5 grms./c.c. c. 18-7 (Pt = 100) c. 125 c. C.g.S. Thorium - 232-0 11-3 c. 120 Cold 197-2 19-3 101 1064 0-70 Platinum - 195-2 21-5 100 1750 0-17 Iridium 193-1 22-4 98 2290 0-17 Osmium - 190-9 22-5 97 2700 0-17 Tungsten - Tantalum - 184-0 181-0 19-3 16-6 91 90 3400 2900 0-35 0-12 Palladium 106-7 11-4 55 1550 0-17 Rhodium - 102-9 12-4 54 c. 1900 Ruthenium 101-7 12 3 53 1950 ? Molybdenum Niobium - 96-0 93-5 8-6 12-7 50 49 2500 2200 ? Zirconium 90-6 4-1 47 c. 1300 Yttrium 89-0 3-8? 46 Copper Cobalt 63-6 59-0 8-9 8-6 33 30 1084 1480 0-92 Nickel 58-7 8-9 30 1450 0-14 Iron - 55-9 7-9 27 1530 0-15 Manganese Chromium 54-9 52-0 7-4 6-5 26 25 1280 1520 Vanadium 51-1 * 5-5 24 1720 Titanium - 48-1 3-5 22 1800 Iridium is even more expensive than platinum, but appears to behave satisfactorily if there is no oxygen in the X-ray 1 It is important in the case of tungsten to get rid of water vapour or oxygen in the tube, if excessive sputtering is to be avoided. 40 X RAYS tube. Osmium, which was introduced in the very early days of X rays by Sir James Mackenzie Davidson, while excellent as an anticathode, is very scarce and expensive. Rhodium would seem to have much to recommend it as a material for anticathodes ; it has a high atomic weight and low volatility. Bragg has moreover shown that the rhodium radiation from a soft tube is remarkably homogeneous (see also pp. 125 and 216). Platinised Nickel Anticathodes. It should be remarked that almost all the cheaper X-ray tubes are fitted with nickel anticathodes faced with very FIG. 19. Photomicrograph of a platinised nickel anticathode fused by a discharge. thin platinum sheet (about T ^ mm. thick). The high price of platinum was responsible for the introduction (in 1897) of these composite anticathodes of which nowadays large numbers are turned out. There is no objection to the plan if the tube is intended only for moderate output ; but care should be taken that the platinum facing is not fused, as nickel is a greatly inferior radiator. Fig. 19 is a photo- micrograph (due to Mr. J. H. Gardiner) of a fused platinised nickel target. Design of the Anticathode. Some makers envelop the anticathode in a glass sleeve, others fit it within a porcelain ring. Both devices prevent AN X-RAY BULB 41 the incidence of cathode rays on the sides of the anticathode and the consequent generation of X rays which would interfere materially with sharp definition in photographic work. In some cases, the anticathode is made trough- shaped or is surrounded by a hollow aluminium cylinder to do away with the X rays produced by the reflected cathode rays striking the glass walls : the definition is described as being improved. In certain makes of bulb the anticathode is provided with a " focussing ring " consisting of a stout ring of copper mounted in front of the anticathode (Fig. 25) ^^BHBBBBIHHHBiHHK^^. QBHB ; FIG. 20. An X-ray bulb, showing water-cooled anticathode and automatic softening device. with which it is in metallic contact. The object is to prevent the wandering of the focal spot (p. 32) with the consequent blurring of definition. Cooling the Anticathode. Nowadays, the very pronounced heating of the anti- cathode is overcome .in many tubes by cooling the back surface by water (Fig. 20) or a stream of air. In some makes of tube, no attempt at cooling is made, the anti- cathode being designed for continuous use at a red heat. With other designs of target, the temperature is kept down by increasing the massiveness of the anticathode (Fig. 21) ; 42 X RAYS this is done by backing up the platinum or tungsten plate with copper, nickel, or iron. In some cases, the support extends to the outside of the tube, and is there provided with fin radiators (Fig. 22). FIG. 21. A bulb, showing massive anticathode and osmosis softening device. Andrews is responsible for a method of cooling the anti- cathode by means of tongs. A thin copper tube is closed at one end to which is fastened the platinum target : the other end of the tube is fused to the glass through the FIG. 22. A Cossor bulb with automatic softening device and fin radiator for cooling anticathode. intermediary of a platinum belt, and thus the inside of the copper tube is open to the air. Into the aperture can be introduced a pair of metal tongs, by means of which the massiveness of the anticathode can be greatly increased. AN X-RAY BULB 43 The point of impact of the cathode rays is generally not more than 1 or 2 mms. across, and with a heavy discharge the heating is so intensely restricted and rapid that the anti- cathode may be melted locally without damage to the rest of the plate. Fig 23 is a photomicrograph (kindly lent me by Mr. J. H. Gardiner) of the focus spot of a tantalum anti- cathode subjected to a momentary heavy discharge. The metal was liquefied, and the pool of molten metal was blown away from the cathode into a mound, where it solidified on the cessation of the current. FIG. 23. Photomicrograph of fused focus-spot in a tantalum anticathode. Gardiner (J.Rt.S. 1909), by the use of a small magnet to deflect the cathode rays to a new portion of the anticathode. showed that, without in any way impairing photographic definition, it is possible to prolong very greatly the life of an anticathode. ^Coolidge X-ray Tube. In 1913 Dr. W. Coolidge * (P.R. Dec. 1913) designed a new X-ray bulb which marked an important step in the pro- gress of the subject. The chief novelty is the fact that the gas pressure is so low that the residual gas plays little or no part in the ' ionics ' of the bulb. The source of the electrons 1 Of the General Electric Co/s Research Laboratory, Schenectady, New York. 44 X RAYS is an incandescent cathode. This consists of a small flat spiral of tungsten wire, surrounding which is a molybdenum tube, the two being electrically connected. The tungsten spiral is heated by a subsidiary electric current (as with a Wehnelt cathode, p. 8), and so becomes a source of electrons (or cathode rays) to an extent which increases rapidly with the temperature. The molybdenum tube serves to focus the stream of electrons (see p. 75) on the anticathode, which is of tungsten and unusually heavy. There is no . additional anode. The vacuum within the tube is extremely high about 1000 times that of an ordinary X-ray tube with the result that unless the cathode is heated, it is impossible to send Hewing Leads S^ ~^ > / ^ Heavy ' Tungsten Spiral C&Khode FIG. 24. Coolidge X-ray tube. a discharge through the tube. Furthermore, the greatest care is taken in freeing the electrodes and the glass walls from gas, before sealing off, after exhaustion. By reason of the care taken in exhausting the tube, there is no appreciable change in the vacuum, and, therefore, in the intensity of X rays, even after a run of many hours. The focal spot does not wander or vary in size/ The intensity of the X rays is precisely and readily con- trolled by adjusting the temperature of the cathode. A sensitive control over the cathode temperature is essential. At high temperatures (2300 C. or so) an enormous output of X rays is possible. For example, such a tube has been run, on a 7 cm. alternative spark-gap, for hours at a time, with a steady current of 25 milliamperes passing through the tube continuously. With heavy continuous discharges AN X-RAY BULB 45 the walls of the tube may well be kept cool by a blast of air. The penetrating p-wer of the rays depends, as in an ordinary tube, solely on the potential difference between the electrodes. The X rays are heterogeneous (see p. 128). Either direct or alternating potential can be used to excite the tube, for the hot cathode allows the current to pass only in one direction, except indeed when the anticathode also becomes red or white hot, as is usually the case after heavy or long discharges. As a result, with a coil discharge, the "inverse" current is entirely abolished under normal conditions. Owing to the low pressure, positive rays do not play an appreciable role, and there is in consequence no evidence of cathodic sputtering. There are slight traces of blackening due to vaporisation of the tungsten. The starting and running voltages are the same, and the tube* is remarkable in showing no fluorescence of the glass as in ,iie ordinary X-ray tube, so that its appearance affords little notion of the output or indeed of any activity at all. The walls of the tube become negatively charged, and so differ from those of an ordinary gas-filled bulb (see p. 32). It is found that in the type of hot-cathode tube now on the market there is a very considerable emission of X rays from all over the anticathode. A test with a pin-hole camera showed that the integrated value of these stray rays may amount to i of the intensity from the focal spot. This disturbing factor, which is doubtless produced by secondary cathode rays, may be reduced by the use of a glass mantle and a suitably perforated cap fitting on the front of the target. Coolidge (American Journal pf Rontgenology, Dec. 1915) has also experimented with anticathodes cooled by a current of water. Such tubes permit enormous X-ray outputs. One tube ran continuously for 68 hours at 100 milliamperes and 70,000 volts. Others have been run continuously at 200 milliamperes, the power input being 14 kilowatts. It is anticipated that this figure will be shortly increased to 50 kilowatts, 46 X RAYS Rutherford, Barnes and Richardson (P.M. Sept. 1915) found that with a cold cathode the perfection of the vacuum in a Coolidge tube under experiment was such that it with- stood 175,000 volts without breaking down. When the cathode was heated, they could not detect (by an ionisation method) any radiation with a voltage less than 10,000 on the tube. As the voltage was increased the penetrating- power of the radiation increased rapidly and regularly. At constant voltage the composition of the X-ray beam and hence the shape of the absorption curves were fovind to be independent of either the current sent through the X-ray tube or of the temperature of the cathode. The tube was excited in turn by a Wimshurst machine and an induction coil : the absorption curves for the same voltage proved to be practically identical. The hardest X rays obtained by Rutherford with a Coo- lidge tube had a penetrating power about \ of that of the hardest y rays from RaC. Even with the highest voltages employed (175,000 volts) it was found that the intensity observable through 3 mm. of lead was less than 1 > 1 of the initial value ; and this thickness of lead may be regarded as affording adequate protection for the worker against any of the X rays from a Coolidge tube. For further data concerning the Coolidge tube, see pp. 128 and 176. Snook Hydrogen Tube. In the tube recently introduced by H. C. Snook, the residual gas is pure hydrogen. One of the features of the tube is the method of regulating the gas-pressure, in which an ingenious application of the osmosis method (p. 78) is employed. This method of regulating the vacuum depends on the fact that hydrogen will pass through red-hot platinum or palladium from a region of higher to one of lower hydrogen pressure. This diffusion takes place quite irrespective of the presence of other gases on either side of the metal boundary. In the Snook tube two osmosis tubes are sealed into the bulb. Either may be heated (to a bright red heat) by causing a discharge to pass from the tip of the tube to AN X-RAY BULB 47 an adjacent metal electrode, the control obtained being rapid and convenient. The one regulating tube, which is usually of palladium, is exposed to the atmosphere. Now the amount of hydrogen in the air is negligible, and consequently when this tube is heated, hydrogen passes from the interior of the bulb and the vacuum is raised. The other regulating tube (usually of platinum) is sur- rounded by a small auxiliary bulb containing pure hydrogen at about atmospheric pressure, and the result of heating this tube is that hydrogen diffuses into the X-ray bulb and the vacuum is lowered. When necessary the auxiliary bulb can be replenished with hydrogen. Hydrogen Reservoir Osmosis Reguledor for lowering Vacuum Osmosis Regulator for raising Vacuum He&vy Copper^ Anhc&mode farced wihh Tungsten JfiG. 25. Snook hydrogen tube. Fig. 25 shows the Snook bulb and the two osmosis tubes. The anticathode is of tungsten. There is very little " sput- tering " of tungsten (see p. 80) in pure hydrogen, and, in passing, we may note that a tube filled with hydrogen runs harder for the same spark-gap than one containing air. Metal X-ray Bulbs. Sir Oliver Lodge designed an air minium X-ray bulb in 1897, but the idea was not pursu?;;. ehnder (Elect. Zeit. Feb. 1915) and Siegbahn (D.P.G. V. Dec. 1915) have de- scribed all-metal X-ray tubes with porcelain-insulated elec- trodes. Siegbahn's tube was provided with a silver window which acted also as anticathode (see p. 142). 48 X RAYS Obliquity of the Anticathode. The design of tube introduced by Prof. Jackson, in which the cathode rays are focussed on an anticathode inclined at 45 to the beam of cathode rays, has become the universal pattern. It has two disadvantages : (1) The obliquity of the anticathode to the cathode rays increases the area of emission of the X rays. This is to the detriment of definition in photographic work, though it must be conceded that in any case a point source of X rays is not feasible in practice. (2) It is impossible, with a coil discharge, to suppress entirely the reverse current at " make," during which time the cathode rays proceeding from the anti- cathode impinge on the glass walls, with the con- sequent risk of piercing the tube. Both objections could be met by mounting the anticathode parallel to the cathode and using normally incident cathode rays. The writer showed (P.R.S. 1909), in some preliminary experiments, that the output of a tube was almost inde- pendent of the obliquity of the anticathode. The fluores- cence of the bulb, which is due to the " reflected " cathode rays from the anticathode, increased very markedly as the angle of incidence (to the normal) of the cathode rays in- creased, but the X rays did not show any corresponding variation either in quality or quantity. Depth of Origin of X Rays in an Anticathode. Various observers have found that the mean depth at which Rontgen rays originate in an anticathode is directly proportional to the potential employed. Ham (P.R. 1910) found that with a potential of 21,500 volts the mean depth was 5-9 x 10" 5 cm. in the case of a lead anticathode. Davey (P.R. 1914) using a platinum target found the mean depth to be 0-00004B cm., where B is the Benoist hardness number (see p. 107). The writer showed (P.C.P.S. 1909) that with spark-gaps of from 1 mm. to 1 cm. a thickness of from 1 x 10~ 5 to AN X-RAY BULB 49 4 x 10 " 5 cm. of gold, copper, or aluminium, was more than sufficient to generate X rays. These distances may be compared with the minimum thicknesses which have been found essential for complete " reflection " of cathode rays of various velocities. These are as follows : TABLE II. Potential. Thickness of Metal. Authority. 11,000 volts 5-3 x 10~ 5 cm. Al Warburg 1905 16,500 19-0 Al ,, 21,800 24-4 Al M 27,800 <6'6 Cu 90,000 0-25 Pb Ham 1910 Distribution of the X Rays. The distribution of the X rays from a bulb of the ordinary^ type is not quite uniform. Ham (P.R. 1908), Bordier (1908), and Gardiner (J.Rt.S. 1910) agree that in a plane determined 100 FIG. liG. Graph showing distribution of X rays, the cathode rays being incident normally on anticathode. by the beam of cathode rays and the normal to the anti- cathode, the intensity reaches a maximum in a direction at about 60 from the normal. (Cf. p. 116.) A distribution curve obtained by the writer (P.R.S. 1909) for normally incident cathode rays is given in Fig. 26, in which the length of the radius vector in any direction is proportional to the 50 X RAYS intensity. It should be noted that as X rays are given off in all directions, they will be found proceeding from the back and sides of the anticathode as well as from the front. Thin Anticathodes. Some information on this point is afforded by the writer's experiments (P.C.P.S. 1909) on the emission of X rays in both backward and forward directions from anticathodes consisting of aluminium, copper, gold, or platinum leaf. The apparatus is shown in Fig. 27. The results indicate that window FIG. 27. Apparatus for measuring X rays emitted from each side of a very thin anticathode. the forward or " emergence " X rays exceed the backward or " incidence " rays both in intensity and hardness. In other words, the X rays tend to proceed in the same direction as the cathode rays which produce them. This is most pro- nounced in the case of aluminium, where with leaf about 0-00001 cm. thick, and a spark-gap of 1 to 2 cms., the emer- gence rays were two or three times as intense as the incidence. Stark (P.Z. 1909), using a photographic method, has obtained similar results for a carbon anticathode. It would be of interest to test the homogeneity of the X rays from thin anticathodes. In many cases the proportion of characteristic radiation might be expected to be unusually large. CHAPTER V. HIGH-POTENTIAL GENERATORS. THE various means of exciting X-ray bulbs may be con- veniently grouped into : (1) Influence machines. (2) Induction coils. (3) Step-up transformers. INFLUENCE MACHINES. Influence machines, which are nowadays almost always of the Wimshurst type, have been largely used in France, Germany, and the States for the production of X rays, but, probably owing to climatic reasons, have found little favour in this country. Very few influence machines, sold as such, are really suitable for the purpose ; nearly all of them need redesigning both from a mechanical and an electrical point of view. If glass is chosen for the material of the revolving plates, it should be free from excess of alkali, which in damp weather makes the surface conduct- ing : ordinary window glass is quite unsuitable. Alkali- free glass is now procurable ; it is, for example, used in the Moscicki condenser. Such glass should not be coated with shellac varnish according to the usual custom ; shellac is slightly hygroscopic, and, although it is a better insulator than bad glass, it is not so good as the best glass. Care should be taken to avoid undue fingering of the plates. Ebonite plates have advantages over glass (see p. 255), certainly on the score of safety for high-speed machines. 52 X RAYS With continued exposure, however, to the stray brush- discharges, the ebonite tends to deteriorate, probably owing to the ozone, which is always generated in abundance, and which many workers find objectionable. For leads, massive or india-rubber sheathed wires free from points and sharp bends, and as short as possible, should be u&ed, otherwise the leakage by brush-discharge, always considerable, will prove excessive. When an X-ray bulb is run by a machine, either two short spark-gaps or two Ley den jars should be put in series with the bulb, one on each side of it : this will prevent undue frittering away of the electricity. With a multiple-plate machine in good working order, a, beautifully steady X-ray discharge can be obtained. The current is, moreover, unidirectional, and is found to be not so destructive to the anticathode as pulsating or alternating current. The voltage from a Wimshurst machine is proportional to the speed of the plates : there is no theoretical limit to the potential obtainable, except such as is imposed by leakage or disruptive discharge. A Wimshurst machine is peculiar in that the current obtained is almost entirely independent of the voltage. The current output can be raised by increasing the number of plates. The voltage is readily controlled by altering the tilt of the rod supporting the brushes : a needle-point spark-gap is useful in regulat- ing minor variations of the potential. But, as has already been remarked, the idiosyncrasies and unreliability of influence machines have caused most workers to fight shy of them, at any rate for X-ray work. For instance, some machines refuse to work at all inside the glass cases provided for them ; yet, in their absence, the machines attract all the dust within reach and require con- tinual cleaning. It is a habit with nearly all machines to reverse their electrification if stopped and restarted : in at least one type, a device is provided to counteract this. It may be noted that although a Wimshurst machine generates homogeneous cathode rays, the resulting X rays are not homogeneous. INDUCTION COILS 53 As an example of the successful large design of machine, one may mention that of Hulst in America. The plates, fifty in number and small in diameter, are constructed of compressed mica, and are motor-driven at a very high speed about a vertical axis. Such a machine will send a current of some 15 to 20 milliamperes l through an X-ray tube, and yield rays of an intensity such as would require double the current from a coil. The machine is, however, excessively noisy, and there is, of course, the danger atten- dant on the high speed of the whirling plates. Villard and Abraham (C.E. 1911) describe a somewhat smaller 20-plate Wimshurst machine, whose construction allowed speeds of from 1200 to 1400 revolutions per minute. The plates were of ebonite 70 cms. across. The maximum current obtained was 3 milliamperes, the highest voltage about 320,000 volts, and the longest spark-gap 55 cms. Some workers have been successful with Wimshursts, which work in air-tight cases into which air or carbonic acid is pumped under pressure. The idea is to kill the losses due to brush discharge ; but the working difficulties are so great that the latest designs of Wimshursts have reverted to the simple unenclosed pattern. INDUCTION COILS. It is only within the last few years that makers of in- duction coils have stirred themselves to meet the special requirements of the X-ray worker. The improvements in design and performance are doubtless not wholly uncon- nected with the competition offered by the various step-up transformers. The present-day coil offers improvements even on its predecessors of only five years ago ; standardis- ing of proportions proceeds, and any differences of design among the different coil makers depend more on individual predilections than on theoretical grounds. It is not generally realised that the same coil cannot be equally efficient for all purposes ; it cannot, for example, 1 A milliampere = ^^ ampere. 54 X RAYS prove equally satisfactory for hard and soft bulbs, or for all speeds of interrupters. While all the ambitious efforts of the early coil maker were directed towards phenomenally long sparks, nowadays, for X-ray work, he is content with a 10 to 12 inch spark, provided it is a " fat " one and as unidirectional as possible. A fat spark means heavy current and intense X rays, and that satisfies the radiographer, who requires short exposures for much of his work, and finds that very long sparks mean rays too penetrating for his purpose. 1 Some of the later coils will pass through an X-ray tube sustained secondary currents up to 60 milliamperes with relatively small primary currents and but little inverse current. It will not be un- profitable to consider in some detail the various parts of a modern coil, a brief account of which was given on p. 27. Core. The aim of the coil-maker is to magnetise the core slowly (at make) and demagnetise it rapidly (at break). The spark-length depends on how quickly the core can be demagnetised. On the other hand, the output or power of the coil depends largely on the degree of magnetisation. With modern high-frequency interrupters the core is never either fully magnetised or demagnetised. The ideal size of core depends on the size of the primary and the current in it, on the frequency and character of the break, and on the output required : the heavy dis- charge coil of to-day has a conspicuously large and stout iron core whose length is some five or six times the diameter. The chief objects kept in mind in core design are (1) to diminish the inverse current, and (2) to reduce the losses due to eddy-currents and hysteresis in the iron. The inverse current is lessened by packing as much iron as possible into the space available for the core. The hysteresis loss is diminished by using iron as soft as can be obtained. The eddy-currents are reduced by using, instead of a solid iron core, closely packed wires or plates varnished to 1 .4 propos of long-spark coils, Carpentier showed in 1910 at Paris a monster coil capable of a 60-inch spark. INDUCTION COILS 55 diminish the electrical contact between them. Laminated plates have a better " space factor " than wire in a cylin- drical core in other words, there is less space unoccupied by iron and accordingly plates are used for nearly all large coils. Iron with very high resistivity is now avail- able, and so fairly thick plates can be employed. Primary. The primary is usually wound in three layers, either as a simple winding, or in some form of adjustable winding to secure adaptability to prevailing conditions. There are in common use three different methods of winding primary coils which permit adjustment. In one, the connections are arranged so that each of the three layers can be put in series or parallel with its fellows ; in a second, a number of " tapping-off " wires permit connection to different parts of the primary circuit ; in a third, the primary is wound with several wires " abreast," so that these multiple windings can be put either in parallel or series at will. A heavy-discharge coil has a primary stout enough to permit direct coupling to the electric light supply of 100 or 200 volts. Great care has to be paid to the insulation of the primary, owing to the induced E.M.F. from the secondary, of which all observers are well aware by reason of the shock which can be obtained from the primary of even a small coil in action. Nowadays, if a fault develops in a coil, it is usually in the primary rather than in the secondary ; the defect is probably due to nitric acid formed by brush-discharges induced by the secondary. Condenser. It was Fizeau, nearly a century ago, who, by the addition of a condenser, revolutionised the induction coil and ob- tained sparks of lengths hitherto unheard of. But Lord Rayleigh demonstrated some years ago that if the primary current is interrupted with sufficient rapidity e.g. by severing a wire with a rifle bullet it is possible to dis- pense altogether with the condenser without impairing the length of the spark from the coil. Owing to the increasing 56 X BAYS use of Wehnelt and high-frequency mercury breaks, the condenser, once paramount in importance, has become in such cases Unessential. With the older patterns of breaks the condenser is, of course, still important. Its functions are three in number : it performs each of them with incom- plete success. 1 (1) To increase the suddenness of the " break " and the slowness of the " make," and so to reduce the inverse current. (2) To suppress undue sparking and arcing at the inter- rupter. (3) To retard the formation of induced currents in the primary. It is important that the capacity of the condenser should be as nearly as possible adapted for the particular value of the inductance of the primary as well as for the magni- tude and frequency of the primary current. If the capacity is too large or too small, the secondary wave of potential will be neither so large nor so sudden. 2 The capacity required depends also very considerably on the type of break for instance, less capacity is required with a gas break than with an oil break and accordingly an adjust- able condenser should be used in the primary if a coil is required for a variety of purposes. But for coils restricted to X-ray work alone the invariable condenser is being increasingly fitted, on account of its simplicity. Condensers have improved out of all recognition during the last few years. With condensers of tin-foil and waxed- paper, this is chiefly due to a better knowledge of the hygroscopic properties of paraffin wax and of the impor- tance of manipulating it by machinery rather than by hand. Primary Tube. Between the primary and secondary coils comes the primary tube ; this is made of ebonite, micanite, or, less 1 See W. H. Wilson, P.R.S. March 1912 2 See Jones and Roberts (P.M. Nov. 1911). In one instance, by reducing the capacity to one-fourth its value, the maximum potential was increased two and a half times. INDUCTION COILS 57 commonly, porcelain. Ebonite has the advantage of being readily machined and worked, but micanite, on account of its greater electric strength, is generally used in large coils, though it is inconvenient mechanically. Secondary. It is in the methods of winding the secondary that the greatest improvements have been effected in the modern coil. Simple winding is never used, partly because of the dangerous strain on the insulation owing to contiguous CenfraJ Ebonite Primary Tube I FIG. 28. Diagrammatic representation of a bisectional winding of the secondary of an induction coil. layers being at very different potentials, and partly because one end of the wire finishes up a*t the innermost layer. An obvious way to avoid this, is to divide the secondary into two sections, wind each of them simply, mount them side by side, and connect the two innermost ends of the wires together at the adjacent faces (Fig. 28). This plan has several advantages. The electric strain on the primary tube is slight ; the tube may accordingly be very thin, so that the primary and secondary windings are close together, with a consequent gain in the efficiency and a diminution in the size and weight of the coil. The method is accordingly of special value for smaller and portable coils. Owing to the electric stress between the outermost points of the 58 X RAYS adjacent faces of the two sections, the intermediate ebonite plate has to be made thick and protruding from the body of the coil (Fig. 29). FIG. 29. A Cox coil wound on the bisectional principle. For large coils (such as is shown in Fig. 30), some form of sectional winding is used, in which a large number of FIG. 30. A Butt coil wound on a multisectional principle. circular flat sections, a few wires thick, are threaded side by side on the primary tube and separated by partitions INDUCTION COILS 59 of waxed or varnished paper. In some cases, these sections are connected up in series by joining the innermost wire of the first section to the innermost of the second, the outermost wire of the second section to the outermost of the third, and so on (Fig. 31), as in the bisectional method ; in others, by joining the innermost wire of one section to the outermost of the next, and so on. Much ingenuity has been exercised in devising methods of winding. 1 It may be noted that the method of sectional winding requires a thick primary tube. I,.... I....J...., I v.1. I -.1 iliiiliil FIG. 31. Diagrammatic representation of a method of multisectional winding of the secondary of an induction coil. Whatever the method of winding, the secondary coil, when complete, is immersed in hot paraffin wax in vacuo, It is highly important to exclude air bubbles from the wax, and the method of vacuum-exhaustion is absolutely essential, if a break-down in the secondary is to be avoided. Some Points in Coil Design. The chief objection to induction coils for X-ray work is the inverse current which all coils generate, chiefly at " make," but also to some extent at " break." The inverse current may be lessened (1) by making the number of turns in the primary as large as possible, 1 See a paper by R. S. Wright (J.Rt.S. 1913) to which the writer is much indebted. 60 X RAYS (2) by reducing the magnetic leakage between the primary and secondary : this means paying attention to the core. The inverse current is augmented by irregular interruption, and care should therefore be taken to keep the break in good order. The inverse current also tends to increase if the X-ray bulb is softened. Sparking at the interrupter, with its attendant waste of energy, may be reduced (a) by increasing the self-induction of the primary, (b) by lowering the frequency of the interruptions. (1) and (a) are consonant, but they both imply a large secondary if the coil is to give long sparks. This is objec- tionable from the coil maker's point of view who, to obtain a heavy discharge, is very desirous of keeping down both the resistance and the number of turns in the secondary. It is, however, possible to obtain long sparks with a secondary of reasonable size, by increasing the rate of interruption. (6), however, requires a low-frequency break ; and, more- over, eddy-current losses become considerable with very high frequencies. If a heavy output is required from a coil, and the voltage available for the primary is only low, the self-induction of the primary should be kept down. This is inconsistent with but more important than (a). In such cases the output can often be materially improved by taking care that the leads from the battery to the coil are kept as short and straight as possible, the object being to diminish the self- induction in the circuit. The efficiency of even the best induction coils, considered as transformers, is not high in the region of 50 to 70 per cent. It could, of course, be increased by using a com- pletely closed (ring) core instead of a straight one, and so diminishing the magnetic leakage. But the difficulty hitherto has been that, with a closed core, demagnetisation does not occur with the intermittent current which obtains in a coil discharge. The objection does not apply to true alternating current, in which there is a complete reversal, INDUCTION COILS 61 and for which, of course, closed-core transformers are always used. Enough has been said to indicate some of the problems which confront the coil designer. It is in reconciling necessarily antagonistic factors to suit the main purpose of the coil that his skill finds chief scope. The Wave-form of the Primary and Secondary Currents. The oscillograph l has been employed by a number of workers to investigate the shape of the waves of current FlQ. 32. Oscillograph record of a make and break of the primary current of an induction coil. and potential generated by a coil at each make and break of the interrupter. Fig. 32 shows a typical record (due to Salomonson, J.Rt.S. 1911) of a single make and break of the primary current in the case of a 13-inch coil giving a 10-inch spark : a mercury-oil break was used. As soon as the circuit is completed, the current starts from zero and rapidly grows in strength until the moment at which the circuit is broken. The current then falls to zero in about 1/1000 sec. In some cases, the curvature of the rising part of the curve is more marked than in Fig. 32. A close 1 An oscillograph is essentially a low-resistance, moving-coil galvano- meter of few turns and with a very short time of swing. 62 X RAYS inspection will show that superposed on the main current are extremely rapid oscillations : these are produced by the condenser. W. H. Wilson (P.E.S. 1912) noted that much longer sparks could be obtained from a coil when these high-frequency oscillations were pronounced in the primary _____ current. Fig. 33 illustrates them very well. The frequency of these rapid oscillations may reach many thousands a second. In regard to the secondary circuit, Duddell (J.Et.S. 1908) found that the discharge con- sisted of isolated groups of strongly-damped impulses very abrupt and short-lived. The in- terval between successive groups of waves was relatively long compared with the actual dura- tion of each group, which latter was of the order of 1/1000 sec. Fig. 34 'gives a general notion of the state of things that obtains with a medium vacuum in the X-ray bulb. 1 The upper graph shows the current, the lower the potential. In the latter curve, the upper peak is the potential tending to send the current in the right direction through the tube : the smaller and broader inverted peak is due to the inverse potential, which in this case is conspicuous. The maximum direct potential is about 60,000 volts, the maximum inverse potential about 33,000 volts. The current curve is very similar to the potential curve : a small inverse current is detectable. In Fig. 35 a rectifying spark-gap is inserted in the circuit : its ability to suppress the obnoxious inverse pulses is well 1 A graph showing greater detail is a good deal more complicated. Pip. S3. Oscillograph record of a primary current showing super- posed high-frequency oscillations. HIGH-TENSION TRANSFORMERS 63 displayed. The maximum direct potential now supplied to the bulb is 39,000 volts. Thus some 21,000 volts have been Current. Potential. FIG. 34. Oscillograph record of groups of impulses in the secondary circuit of an induction coil. used up in the spark-gap ; and the illustration serves to point out the loss of energy that occurs in a spark-gap, and Current. J__LJ_J_J Potential. PIG. 35. Conditions as in Fig. 33, but with rectifying spark-gap inserted. the desirability of avoiding its use by not generating the inverse -current at all, if that were possible. HIGH-TENSION STEP-UP TRANSFORMERS. About 1908 the first high-tension transformer for X-ray work was introduced by Snook (Fig. 36), and since then transformers have been largely used in X-ray work, more 64 X RAYS especially in instantaneous radiography. The machine is essentially nothing more than an oil-immersed step-up trans- former, which is supplied with alternating current from an alternator. A rotating pole-changing switch rectifies the high potential alternating current from the secondary of the FIG. 36. Present design of Snook high-tension transformer. transformer. To secure the perfect synchronism which is essential for rectification, the commutator is mounted on the same shaft as the alternator. The resulting current is, of course, not uniform, but pulsating as in B (Fig. 37) ; its amount can be varied at will from J to 100 milliamperes. The efficiency of the transformer, which is of the ring type, is considerably greater than that of an induction coil. The HIGH-TENSION TRANSFORMERS 65 chief objections to such transformers are the high cost and large size, the excessive noise, and the attention which moving machinery requires. On the other hand, they are capable of enormous output and easy control, there is little AAA/WW ' FIG. 37 A. Alternating current of sine form. B. Pulsating current pro- duced by rectification of A. or no inverse current, and no interrupter is needed. Recently, very considerable improvements in design and performance have been effected. It has been suggested that the sinusoidal current curve of the high-tension transformer is not quite as efficient, from an X-ray stand-point, as the long steep peaks of an induction coil (see p. 63), and that they are relatively more destruc- tive to X-ray tubes ; and doubtless there is something to be said for this point of view. In one direction it would appear that simplification is possible in the use of step-up transformers for X-ray work. Instead of sending into the primary of the transformer a sinusoidal current, use an alternator specially designed to give a very unsymmetrical wave form consisting of an abrupt high peak on one side and an almost suppressed loop 1 on the other. The necessity for the commutator thus disappears. Boas described such an arrangement in 1911, and found it to work well in practice. Cabot has recently designed a high-potential rotary converter, in which by the commutation of a symmetrical 9-phase system, the voltage fluctuates.no more than 1 to 2 per cent. The maximum voltage attainable with the machine is 100,000 volts, and the output up to 15 kilowatts. 1 Merely sufficient, in fact, to demagnetise the core after each reversal (see p. 60). E 66 X RAYS BREAKS AND INTERRUPTERS. The Hammer Break. The hammer break (see p. 28), the accompaniment of most of the earlier coils, has been greatly improved recently. Attention has been paid to its period and its mechanical stoutness. Some of the later types compare favourably in steadiness with any kind of interrupter, when only a light output is required, as with a soft X-ray tube. 1 But, on a heavy load, the hammer break is useless : it cannot carry the current without excessive sparking and disintegration of the platinum. This does not contribute to steadiness and economy of working. The frequency of a hammer break never reaches more than about 200 (per sec.), and is usually much less : with a large coil it may be as low as 25 to 30. Accordingly a variety of other breaks have been introduced from time to time. These include the electrolytic interrupters, and the various kinds of motor-driven breaks which employ mercury. The Wehnelt Electrolytic Interrupter. Wehnelt in 1899, turning to account an earlier observa- tion of Violle in 1892, devised his interrupter, which now enjoys extensive popularity. It consists of two electrodes immersed in dilute sulphuric acid. 2 The ca- thode is a large lead plate, the anode consists of one or more platinum points. The amount of the anode exposed to the liquid can be adjusted by means of a porcelain sleeve round each of the platinum points (Fig. 38). For efficient interruption, the PlG. 38. Wehnelt interrupter with onrrAnf rnnat IIA V^f -inr^^n oAr+nin single platinum anode (Siemens). 1 The parallelism and flatness of the contact-pieces should be seen to : a thin piece of flat wood faced on both sides with a fine grade of emery paper is useful for passing between the platinum studs. 2 A density of T2 is suitable. Some workers add a little CuSO 4 . INTERRUPTERS 67 limits ; if it is too small (below about 10 amperes) mere ordinary electrolysis occurs, if too great (say 40 amperes or more) the polarisation increases to such an extent that the current almost ceases and the anode becomes white- hot, and hisses and disintegrates in the liquid. With a suitable current the anode is normally surrounded by a violet light, and the interruptions are of an explosive and almost deafening character a very unpleasant feature of the break. 1 Electrolytic breaks will not work with voltages exceeding 80 to 120 volts ; they are capable of a larger output than any type of break, but the reverse current is considerable and the X-ray tubes suffer in consequence. Opinions are still very much divided as to the mode of action of the break : the usual explanation is that the inter- ruptions are brought about by the periodic sealing and unsealing of the anode by liberated bubbles of gas ; but this does not meet all the circumstances. There are many factors to take into account the size of the anode point, the current, the concentration and temperature of the acid, the inductance and capacity in the circuit : all these affect the interruptions. Compton (P.E. 1910) showed that just as with the ordinary hammer break, the " break " is more sudden than the " make." The Wehnelt break usually requires a little humouring, and works rather better when the acid is warm, a state of things which soon results in practice ; indeed, for prolonged use, it is necessary for regular interruption to cool the acid, e.g. by means of a water-cooled worm of lead tubing. The interruptions are extremely rapid as high as 1500 to 2000 per sec. when a very small anode point is used : even with very large currents the frequency may reach 200. The frequency is increased (1) by diminishing the size of the anode point, (2) by raising the temperature of the acid, (3) by diminishing the self-induction in circuit. Some self- induction is, however, essential or there will be no inter- ruptions. A condenser across an electrolytic break is not beneficial, and is, in fact, generally detrimental to the working of the break, which itself functions as a condenser. 1 Many makers now fit silencers to the break. 68 X RAYS FIG. 39. Wehnelt interrupter with perforated tube round lead anode (Schall). The energy required is diminished by raising the tempera- ture and (slightly) by using stronger acid. It is found that to get the same spark-length, a more powerful coil is required with the Wehnelt than with any other break. An electrolytic break does not, in fact, conduce to the highest efficiency in the working of a coil. In another form of electrolytic break, both electrodes are of lead (Fig. 39), but one is surrounded by a porcelain cylinder pierced with a number of small holes, at which the bubbles of gas are formed. This type permits no control over the current, 1 but the reverse current is said to be smaller. This latter break is also suitable for alternating current, in which case it may be noted that the frequency is always equal to the frequency of the supply current, and is not affected by any of the controllable features of the break. Mercury Breaks. There are many ingenious forms of these breaks on the market, some of which are extensively used. They are invariably motor- driven. The early forms depended on the rapid dipping of a plunger into a trough of mercury ; in some of the later types a jet of mercury is pumped against a series of rapidly revolving metal vanes. To these and other types of breaks, the various makers' catalogues do full justice. Two varieties of mercury break are illustrated in Figs. 40 and 41. 1 In the Caldwell-Swinton pattern, the cylinder is pierced with only one hole, the size of which can be regulated and the current thus varied. FlQ. 40. Sanax mercury-paraflin break. INTERRUPTERS 69 In the earlier forms the revolving system was immersed in paraffin oil or methylated spirit. With either liquid, but especially with the oil, the mercury emulsifies in most breaks, and the cleaning required is frequent and wasteful. Coal gas or hydrogen at 1 or 2 atmospheres is generally used nowadays in mercury breaks : the break needs less cleaning, and is usually more reliable and economical than with a liquid dielectric. Salomonson has shown (J.Rt.S. 1911), by means of the oscillograph, that stronger and more abrupt quenching of the spark is obtained with a gaseous dielectric than with a liquid in which a conducting charred FIG. 41. Mercury-gas break. (See also Fig. 88.) track persists after each spark. Less condenser capacity is required for a gas break than for an oil or spirit break. With most coils, these motor-driven breaks permit a heavier discharge current at the higher speeds. The mercury break is designed so that the current is " off " rather longer than " on " ; in this respect it is superior to the Wehnelt, in which the " on " period is equal to the " off," to the detriment of the demagnetisation of the core. When a high-speed break is employed it is beneficial for the bulb to receive periodic " rests." This can be brought about by inserting an additional break, but of low speed, in series with the high-speed break in the primary. Doubtless most workers would prefer a mercury break to any other kind for general use ; though for heavy instan- taneous work an electrolytic break is probably unequalled. 70 X HAYS A mercury break permits greater control, however, and the good types are not subject to current and voltage limits of working, such as obtain with an electrolytic break. RECTIFIERS AND VALVE-TUBES. The chief defect of the induction coil from the point of view of the X-ray worker is that it does not give unidirec- tional currents : the reverse current at " make " has a disastrous effect on the X-ray tube, and requires to be suppressed. For this purpose we may introduce into the circuit the simple point and plane spark-gap, which depends on the fact that the spark passes more readily when the point is positively charged than when it is negatively charged. The device is an old one, and is not always particularly efficient, more especially if the current is considerable. The greater the current which passes, the longer is the spark-gap required for rectification. For a current of about a milliampere, a spark- length of 1 cm. or more is suitable. Duddell (J.Kt.S. 1908) showed FlQ. 42. The Duddell spark-gap. that with a point anode and a given spark-length, a cup-shaped cathode will rectify a larger current than a plane, and a plane a larger current than a sphere. Duddell has accordingly designed a recti- fying spark-gap, in which the point electrode is surrounded by a hollow sphere, through which the point enters by means of a glass tube in a cork (Fig. 42). Correctly disposed, one rectifier in series with the X-ray tube and a second (reversed) in parallel with the tube, the arrangement is described as extremely efficient. For most purposes, especially when electrolytic breaks are used, the various valve- tubes are more efficient than spark- gaps. These consist of a large aluminium cathode, often spiral in form, mounted in an exhausted bulb : the anode is small, and is contained in a restricted side tube (see p. 33). The design is due to Villard : in Sir Oliver Lodge's modi- RECTIFIERS 71 fication (Fig. 13), the anode (of iron wire) is surrounded with a copper sheath, partly to prevent sputtering on the glass walls, and partly to increase the resistance of the tube Celhode Anode on Kod Copper She&fh Aluminium Wi're FIG. 43. Section of a Lodge valve-tube for the reverse current (see p. 73). Owing to the use of a phosphorus method of completing the exhaustion, the Lodge valve is red in colour (see p. 254). The Lodge tube is said not to harden with use, but other types of valves should be fitted with some softening device, as they harden con- siderably with use and do not rectify well if the pressure becomes very low. A valve- tube is only efficient over a limited range of pressures. The Wehnelt valve-tube employs a hot-lime cathode (p. 8). Such a tube, in series Fl Ja- Amu itipievaive-tube^ with an X-ray bulb, will transmit only . one phase of the discharge from the coil. Miller's neat mica-disc valve should also be noticed. This consists of a perforated disc of mica carried by an insulated rod on the shaft of the motor driving the interrupter and rotating with it. The disc is located in the high-tension circuit and is so arranged to permit the " break " current (but not the " make ") to pass through the apertures. For heavy work, multiple valve-tubes (Fig. 44) are advis- able, in series and parallel with the X-ray bulb (see p. 185). CHAPTER VI. THE HARDNESS OF AN X RAY BULB. Factors controlling the Hardness of an X-ray Bulb. The hardness of the X rays produced by a bulb is solely dependent on the potential difference between the electrodes of the bulb. There are a number of ways of controlling this potential difference : (1) The most generally recognised method is by varying 1 the degree of exhaustion of the bulb. The lower the pressure, the higher the voltage required and the harder the X rays. The range of effective pressures for producing X rays is very wide. It is, however, possible to make use of other methods which do not involve any change in the gas pressure. (2) By inserting a spark-gap or valve-tube (p. 70) in series with the bulb, the tube is hardened. With very soft bulbs, Winkelmann (A.d.P. 1900) states that the spark-gap should be placed between the cathode and the coil. At lower pressures, the position of the spark-gap is immaterial. (3) By employing Tesla or other currents of extremely high potential, the tube runs harder. Tesla currents are obtained by transforming up the secondary current from a coil by means of a special transformer immersed in oil. (4) By bringing the electrodes nearer together, the tube may be hardened (see p. 33). (5) By altering the nature of the gas in the tube. For the same pressure, a tube runs harder in hydrogen and still harder in carbon dioxide than in air. In other words, in order to generate X rays of equal hardness, a tube filled with air must run at a lower pressure than one containing hydrogen or carbon dioxide. THE HARDNESS OF AN X-RAY BULB 73 (6) By increasing the current density through the tube. This can be done : (a) By increasing the current in the primary of the coil. (b) By diminishing the size of the cathode. A tube with a fine wire cathode runs harder than one with a cathode of moderate size. (c) By diminishing the size of the tube. Winkel- mann in 1900 experi- mented with various sizes of tubes, and found that with a tube 5 mms. in diameter, he could get X rays at as high a pressure as 10 mms. of mercury with air as the residual gas. In the case of hydrogen and a tube 10 mms. in diameter, he obtained X rays at the remarkably high pressure of 30 mms. of mercury. If the tube is made too narrow, the hardening effect is spoilt. FIG. 45 . The discharge is hardened by withdraw- ing the cathode from B to A. FIG. 46. A Cossor bulb of lithium-glass with recessed cathode. (See Fig. 45.) (d) By diminishing the clearance between the cathode and the surrounding tube. It was pointed out on p. 31 that if the space round the cathode is restricted, the discharge passes with difficulty, so that if the cathode is withdrawn from the bulb into a side tube, the discharge hardens accord- ingly (Figs. 45 and 46). Precisely the same effect is obtained with a plane as with a concave cathode, and, indeed, with 74 X RAYS a tube in which the cathode is so inclosed the curvature of the cathode need only be very slight. A tube with a mov- able cathode employing this principle was described by Campbell-Swinton in 1897 (Electrician) ; the tube is in the Rontgen Society's collection of X-ray tubes in the South Kensington Museum. Swinton also employed an alterna- tive device consisting of a glass sleeve, a part of which was narrowed to slide along the glass rod which supported the cathode (Fig. -17). The remaining portion was widened so as to form a sheath round the cathode and project a varying dis- tance beyond it. Wehnelt (D.P.G.V. 1903) found that the arrangement allowed the alternative gap to be varied as much as eight times. Whid- FlQ. 47. Adjustable glass sleeve over the cathode dhlgton (P.C.P.S. 1913) for varying the hardness of the discharge. . ., . observed that, within limits, the distance the sheath projected beyond the cathode was proportional to the potential required to run the bulb. The hardening effect, as Goldstein remarked (D.P.G.V. 1901), is due to the glass round the cathode becoming nega- tively charged owing to leakage from the cathode. The cathode rays accordingly retreat to the centre of the cathode, where they form a concentrated pencil. In this way, the current-density and effective resistance of the tube are increased, and the more markedly if the adjacent glass is coated with sputtered metal. This charging up of the glass is responsible for a well- known effect produced by touching the tube near the cathode while the discharge is passing. The glass under the finger becomes vividly fluorescent, and a bundle of cathode rays is deviated towards the hand. Maltezos (C.E. 1897) showed that if the finger is replaced by the knob of a Ley den jar, the jar becomes positively charged, a clear indication of the negative electrification within that part of the tube (see THE HARDENING OF AN X-BAY BULB 75 p. 32). It is possible to vary the hardness of a tube by putting patches of tin-foil on the outside in suitable places. In the case of the hardening sleeve referred to above, Whiddington has shown that the tendency of the sleeve is to slide back into the side tube owing to electrostatic repul- sion by the cathode ; and, further, that if part of the sleeve is cut away, the cathode rays are bent away from the portion which remains. It can readily be demonstrated that a metal tube, if slid over the cathode inside the glass, will harden the discharge just like a glass tube. In fact, the cathode may be removed altogether and the cylinder alone used in its place ; a sharply defined pencil of rays will still proceed out along the axis of the cylinder (see p. 35). THE PROGRESSIVE HARDENING OF AN X-RAY BULB WITH USE. With a new discharge tube, the first effect of running the discharge is to cause an outburst -of gas. The effect, which may persist for some time, is due largely to gas ejected from the cathode. Aluminium almost always contains large quan- tities of gas, chiefly carbon compounds. Such gas is more readily reabsorbed than air let into the tube. In an X-ray tube, the anticathode also gives out considerable amounts of gas : indeed, the method* of bombardment by cathode rays is a most effective one for liberating the gas held by a metal. But, after some time, the gas-pressure becomes pro- gressively lower with continued running of the discharge. The cause of this has been a problem ever since the days of Pliicker (1858), and one to which a good deal of enquiry has been directed. The effect is undoubtedly not a simple one, and there appear to be several contributory causes. 1 Formerly, the responsibility for the absorption was thrown largely on the metal electrodes, more particularly on the anode ; and doubtless some such occlusion does take place, if only to a slight extent. 1 See Erode tsky and Hodgson (P.M. May 1916). 76 X RAYS But Hill (P.P.S. 1912) has recently shown that a marked absorption of gas occurs even with electrodeless discharges, and it would seem that it is to the glass walls of the tube we must look for the explanation. Campbell-Swinton (P.R.S. 1907 and 1908) concluded from his experiments that the ga& is actually driven into the glass by the dis- charge. He found that when the glass was subsequently fused, such gas (which proved to be chiefly hydrogen) segre- gated into small bubbles x whose depth below the surface did not exceed about 0'015 mm. This thickness of glass is, as Swinton points out, the greatest that will transmit cathode rays to any appreciable extent. A propos of this, it may be remarked that the effect appeared to be inti- mately associated with the fluorescence-fatigue which glass displays when subjected to prolonged bombardment by cathode rays (see p. 12). If the gas- permeated region of the glass is removed by grinding, the glass recovers its usual fluorescing ability. Hill (loc. cit.) found a similar absorption-fatigue ; and it would be interesting to test whether such removal of the fatigued surface promoted vigorous gas-absorption on further running of the discharge. Hill agrees with Willows (P.M. . 1901) in attributing the hardening of discharge tubes to chemical action between the gas and the glass. His experiments show that Jena glass gives the least absorption, lead glass coming next, while soda glass gives most of all. The greater stability of Jena glass is well known from its behaviour in other directions. Possibly fused silica 3 or alkali-free glass would prove to be superior even to Jena glass. It would be interesting to subject an ordinary soda glass bulb to steam or boiling-water treatment before exhaustion, to see if the removal of the alkali affected the rate of hardening. Ramsay and Collie (.AT". 1912) discovered helium (and a trace of neon) along with hydrogen in the deeply stained glass of an old X-ray tube. 2 This is suggestive, for hydrogen 1 The formation of bubbles in sucti circumstances was also noticed by Gouy (C.R. 1896) and Villard. 2 Sir J. J. Thomson (P.R.S. 1913) finds, however, that nearly all sub- stances when bombarded by cathode rays emit hydrogen and helium. 8 See Willows and George, P.P.S. 1916. THE HARDENING OF AN X-RAY BULB 77 and helium molecules have the highest speeds of all mole- cules. Under the electric discharge, these speeds may be increased a thousandfold, e.g. the average velocity of positive rays of hydrogen is 2 x 10 8 cms. /sec. (see p. 20). Gold- smith (P.R. July 1913) found that such high-speed mole- cules of hydrogen and helium can penetrate, for example, mica sheet from 0*001 to 0*006 mm. thick, though the slower air, argon, or C0 2 molecules cannot. But molecules which ^ould penetrate so great a distance as 0*015 mm. of glass would have to be considerably faster. How fast, we may infer from the fact that a particles (helium atoms) from RaC have a range of 0*04 mm. in glass. Such particles have an initial speed of about 2 x 10 9 cms. /sec., i.e. ten times the above velocity. It has, of course, never been shown that sufficiently high instantaneous velocities are not possessed by individual hydrogen molecules in a discharge tube one can only measure average velocities. But, in any case, it is obvious that any explanation such as this could only be a partial one ; it does not, for instance, explain the marked difference in the behaviour of different kinds of glass. The absorption may be due in part to chemical activity excited in the gas by the discharge, such as has recently been found by Strutt to be the case with nitrogen. It may be, too, that the action is stimulated by a species of electro- lysis of the glass produced by the high-tension discharge playing over its surface. It is well known that glass may be readily electrolysed by quite moderate potentials, if the temperature of the glass is raised, and it is a matter of experience that the discharge seems to have an ageing effect on the glass, to the detriment of subsequent working in the blowpipe. Such electrolysis might have a marked effect on the gas film which glass and other solids can condense on their surfaces. Possibly in such circumstances the gas film is capable of taking up abnormal amounts of the residual gas in the bulb. The hardening of an X-ray tube is well known to be pronounced with tubes whose walls have become blackened by metal sputtered from the electrodes (see p. 80). The finely divided metal behaves like spongy platinum in its 78 X RAYS absorptive properties for gases. 1 In most cases this is probably the right explanation of the hardening. To soften an X-ray Tube. It was early discovered that the resistance of a tube could be lowered by warming the bulb with a spirit lamp or gas burner, but the resulting benefit was only temporary, and various " softening " methods have been devised from time to time. Many of these methods involve the heating of some substance which has been inserted in the tube, e.g. sealing-wax, carbon, and red phosphorus have each been employed by various experimenters in the past : Sir William Crookes used caustic potash for this purpose as long ago as 1879. In many X-ray bulbs, this occlusion method is arranged to work automatically. A small alternative discharge tube communicates with the IX the resistance increases beyond a certain degree, the discharge chooses the FIG. 48. Osmosis tube for admitting hydrogen ol+ornafi* T^ofli anrl in into an x-ray buib. alternative patn, ana in so doing heats up some absorbent material such as asbestos, sheets of mica, or glass-wool enclosed in the small tube (see Fig. 22). The consequent liberation of gas (largely C0 2 and water vapour) lowers the resistance of the bulb, and the discharge resumes its proper path. But, in time, such substances " fatigue," having yielded all their available gas ; and the only course is to open up the tube (see p. 184). The plan often employed nowadays for softening bulbs is the " Osmosis " method, originated by Prof. Villard of Paris in 1898, and discovered independently by Profs. Winkelmann and Straubel of Jena in 1899. A small platinum or palladium tube closed at one end is sealed into the bulb, the unclosed end being open to the bulb (Figs. 48 and 25). 1 Soddy and Mackenzie (P.R.S. 1907) showed that helium was absorbed by aluminium scattered from the cathode of a discharge tube. In such case the gas may be mechanically trapped by a compact film of metal. TO SOFTEN AN X-RAY BULB 79 Wlolm di^rubber Bulb By applying a flame to the tube a small quantity of hydrogen diffuses through the hot metal, and the pressure of the bulb can be restored to the right amount. Palladium shows the effect so very markedly that care should be taken in the heating ; otherwise the result will be a bulb too soft for use. Indeed, this method should never be employed except when the discharge is running. The Bauer valve (J.Et.S. Jan. 1907) is a more recent contrivance for letting minute quantities of air into Rontgen bulbs. The valve (see Fig. 49) consists of a small unglazed porcelain disc, through the pores of which air can pass. Ordinarily the disc is sealed by mercury, but by means of a pneumatic piston the disc can be laid bare for a moment by pushing the mercury away (page 84). Porous Plug -> Filter Porous Plug \ Interior of Bulb FIG. 49. The Bauer valve for admitting air into an X-ray bulb. The filter is of gold leaf to absorb mercury vapour. To harden an X-ray Bulb. If by any mischance a bulb be- comes too soft for use, the only thing possible, apart from drastic re-exhausting, is to try and harden it by prolonged running with as large a coil as can be got. Often it is beneficial to send this hardening dis- charge in the reverse direction, i.e. from cathode to anode, temporarily disconnecting the anticathode for the purpose. Care should be taken in carrying out this operation, and the discharge should only be passed intermittently, to avoid puncturing the tube by local over-heating of the glass walls. An ingenious method of hardening a hydrogen tube is described on p. 47. CHAPTER VII. THE BLACKENING OF AN X-RAY BULB. WITH continued use, an X-ray bulb becomes blackened on its inner surface. The blackening is mischievous from several points of view. Firstly, the deposit not only tends greatly to increase the resistance of the tube to the discharge, but accelerates the absorption of the residual gas ; secondly, the discharge is wont to spark irregularly along the walls of the tube instead of through the gas ; and thirdly, the film of metal arrests the softest X rays. Two main causes are answerable for the blackening : (1) The disintegration or " sputtering " of the anticathode while acting as cathode during the inverse current ; and also of the cathode during the direct phase. (2) The volatilisation of the anticathode due to its high temperature under reduced pressure. CATHODIC SPUTTERING. Workers with discharge tubes have long been aware that when a high-potential current is passed through a vacuum tube provided with platinum electrodes, the glass adjacent to the cathode generally becomes coated with a mirror of platinum (Fig. 50). The anode, on the contrary, shows little or no such effect. This property of cathodic sputtering is common in greater or less degree to all metals. The effect was noticed in the very early days of THE BLACKENING OF AN X-RAY BULB 81 vacuum tubes : both Geissler and Pliicker (1858) remarked on it. Thus, quite apart from the cathode rays and positive rays, there is a cathodic emission which consists of particles of disintegrated metal from the cathode. These particles appear to be projected normally (at any rate, very approxi- mately) from the surface of the cathode, and to travel in straight lines. The streams of metal are negatively charged, CaJhode Anode FTO. 50. Illustrating cathodic sputtering. (From the Chemical World.) and it is found that they deposit more readily on surfaces which are positive with respect to the cathode. The positive electrification which the inner surface of an X-ray bulb usually possesses, is thus favourable to cathodic deposition. It does not appear that, in ordinary circumstances, the disintegration of the cathode plays any appreciable part in the passage of the current. Unlike the cathode rays, the sputtered particles require strong magnetic fields (2000 gauss and upwards) before any deviation of their path can be detected. The inference would be, either that the particles are very fast moving or that they are relatively large aggre- gates of molecules ; the latter view is supported by other evidence. The lower the pressure in the tube and the higher the potential applied, the farther are the particles hurled. There is no deposition within the cathode dark- space. The sputtered metal does not appear to excite fluorescence when it strikes the' glass walls of the tube. Cathodic disintegration is not a simple phenomenon, and the exact mechanism of the production of the sputtered particles is doubtful. It appears, however, to be connected with the bombardment of the cathode by the positive rays, the pulverising properties of which we have already noticed (p. 20). 82 X RAYS Experiment shows that the amount of metal shot from a cathode depends on (1) The nature of the metal of the cathode. (2) The temperature of the cathode. (3) The nature of the gas in the tube. (4) The current through the tube. (5) The fall of potential at the cathode. (1) The Metal of the Cathode. Sir William Crookes (P.R.S. 1891) was the first to investi- gate systematically the relative sputtering of a number of metals under like conditions of discharge. The residual gas was air ; the pressure, that corresponding to a dark-space 6 mms. thick (say *05 mm. Hg). A coil discharge was used, and in these circumstances the relative losses of weight at ordinary temperatures resulted as follows : TABLE III. CATHODIC SPUTTERING. (Palladium =100.) Palladium - Gold - 100 92 Copper Cadmium - 37 31 Silver - 76 Nickel 10 Lead 69 Iridium 10 Tin 52 Iron - 5 Brass - 47 Aluminium Platinum ... 40 Magnesium The order of these metals must not be regarded as in- violable. It is affected to some extent by a change in the pressure of the gas (which may, for instance, put platinum above gold), the nature of the gas, or the temperature of the cathode. Nevertheless, the sequence is of value to users of discharge tubes in general and of X-ray tubes in par- ticular. The reason for the invariable choice of aluminium l for the cathode is as readily apparent as the need for sup- pressing the inverse current through a tube and so prevent- ing the platinum anticathode from officiating as cathode. It 1 Geissler policed that aluminium did not sputter appreciably. THE BLACKENING OF AN X-RAY BULB 83 is not right, however, to assume that it is impossible to make aluminium sputter appreciably, 1 as will be evident from a scrutiny of the cathode of an old X-ray bulb : a brown deposit may usually be found on the central area of the cathode as well as on the glass in the vicinity. 2 Tantalum has also proved to be an excellent material for cathodes from the point of view of sputtering. I believe tungsten displays equally good properties. (2) The Temperature of the Cathode. Crookes showed that if the temperature of the cathode is raised appreciably, fcrr instance by the passage of the dis- charge, the sputtering of many metals is markedly increased. The electrodes tend to get very hot if the tube is at all soft, as more current is then passed by the gas. The rise of temperature of the cathode is roughly proportional to the current. This effect is distinct from that dealt with on p. 85. (3) The Nature of the Gas. The nature of the residual gas has a very marked effect both on the degree of sputtering that a metal exhibits and on the appearance of the deposit. Hydrogen, nitrogen, and carbon dioxide do not in most cases favour the effect, while oxygen, the halogens, and the monatomic gases, mercury vapour, He, A, Ne, Kr, and Xe bring about pronounced disintegration of most metals. Helium shows the effect least of all these gases, but argon is particularly potent, and metals so varied as Al, Ag, Cd, Pt, and Au are all excited to a maximum activity in this gas. Aluminium shows only feeble sputtering in hydrogen or nitrogen, and but little more in oxygen. Iron sputters a little in hydrogen ; silver and lead sputter markedly in this gas. Systematic work is needed to find the most suitable gas for an X-ray tube. Unless precautions to the contrary have been taken, the gas will probably consist largely of hydrogen and carbon dioxide liberated from the electrodes. Pt and especially Al (and Mg) emit large quantities of gas when used as cathodes. The point is also of importance in 1 See Campbell, P.M. Sept. 1914. 2 See Kaye, P.P.S. Ap. 1913. 84 X RAYS connection with the various methods of controlling the hardness of bulbs (p. 78). The automatic devices introduce chiefly carbon dioxide, and, in some cases, a little water vapour ; the osmosis valves, hydrogen ; the Bauer valve, air. So far as sputtering goes, hydrogen and carbon dioxide would appear to have advantages, though there is some diversity of opinion on the point. On the other hand, it may be remarked that a tube rendered unsteady by the hardening effect of hydrogen may often be caused to run smoothly by letting in a little air. (4) The Current through the Tube. The disintegration of a cathode increases with the current through the tube, apparently either as the first power or the square of the current. (5) The Fall of Potential at the Cathode. The volatilisation of the cathode is augmented by in- creasing the potential on the tube, and such control is 15 mm. Hg. 0'5 Pressure. I'O FIG. 51,-Relation between cathodic sputtering and pressure. (From the Chemical World.) readily obtained by lowering the pressure of the gas. Sput- tering is much more pronounced at low pressures than at THE BLACKENING OF AN X-RAY BULB 85 high, though at the very low pressures of an X-ray tube the disintegration is not quite so marked as at rather higher pressures, when the tube runs more easily. Fig. 51 displays the relation between the pressure and cathodic disintegration of a number of metals. It is due to Granquist (1898). The potential that is applied to an X-ray tube is not distributed evenly between the electrodes. The greater part is used up close to the cathode ; there is a gentle potential gradient in the space between the electrodes, and the remaining fall of potential occurs close to the anode. The amount of sputtering depends on the cathode-fall of potential, and this increases as the pressure of the gas is lowered. It appears to be essential that the potential fall at the cathode shall exceed a certain minimum value before the metal becomes ionised and disintegrated to any appreciable extent. Holborn and Austin (1904) found that this critical potential was about 500 volts- for a number of metals. VOLATILISATION OF THE ANTIC ATHODE. The high temperatures which anticathodes may attain in a focus tube are familiar enough, but the extent of the sublimation which most metals exhibit at temperatures well below their melting points may not have been brought home to many observers. A homely example of sublimation at low pressure is provided by the blackening which is a not uncommon feature of carbon and tungsten glow lamps. The subject has received attention at the hands of a number of workers, 1 and it appears that the degree of volatilisation is affected by : (1) The nature of the metal. (2) The temperature of the metal. (3) The nature of the surrounding gas. (4) The pressure of the gas. The disintegration of metals increases rapidly as the tem- perature rises. Of the platinum metals, platinum, rhodium, and iridium all disintegrate less as the pressure is reduced, 1 See Kaye, Chemical World, June 1913. 86 X RAYS and there is evidence to show J that in these cases the volatilisation is not a simple process, but is brought about by the formation of endothermic oxides more volatile than the metals themselves. It would seem that in order to reduce the sublimation of these metals to a minimum, the important thing is to ensure the absence of oxygen in the surrounding gas 2 a wise precaution, indeed, with most metals, as almost all observers agree. Hydrogen, nitrogen and argon do not in general favour disintegration. With palladium and most other metals, a reduction of pressure is favourable to volatilisation as would be antici- pated in cases of true sublimation. Table IV. 3 gives, for a number of metals, data con- cerning the effect of pressure on the boiling point, as well as the temperatures at which appreciable vaporisation has been detected (mostly at low pressures). The correspond- ing melting points are added for the sake of comparison. TABLE IV. Metal. Boiling Point. Volatilisation detectable at Melting Point at 1 Atmos. At 1 Atmos. In Vacuo. Cadmium 778 C. 450 C. 160C. 321 C. Zinc 918 550 180 419 Lead 1525 1150 360 327 Silver - 1955 1400 ? 680 961 Copper - Tin 2310 2270 1600 ? 1700 ? 400 360 1084 232 Gold 2530? 1800? 1370 1064 Iron 2450 950 1530 Platinum 2500? 1200 1750 Osmium 2300 2200 Iridium - 2600 ? 1400 2290 Tungsten 3700 ? 1800 3400 The table gives a notion of the extent to which volatili- sation occurs with metals, while still at temperatures well 1 See Roberts, P.M. 1913. 2 This is especially important in the case of indium. 3 See Kaye and Ewen, P.R.S. 1913. THE BLACKENING OF AN X-RAY BULB 87 below their melting points. There is scope for a good deal of systematic work on the volatility of platinum, tungsten, iridium, etc., when heated at low pressures in different gases. The results would be of great practical value to the user of X-ray bulbs. It is known that tungsten, for example, when heated, readily disintegrates and becomes brittle in the presence of oxygen or moisture. Irving Langmuir 1 has recently traced this to the formation of oxides. Coloration of the Glass of an X-ray Bulb with Use. The cathode rays " reflected " from the anticathode are responsible either directly or indirectly for the violet colour which the glass assumes in well used X-ray tubes. This coloration is most pronounced on the front side of the anticathode, and can be prevented by screening the glass with metal foil. Radium rays affect glass and quartz in the same way, though to a greater depth ; and cathode rays produce a similar colour in crystals of rocksalt or fluorspar. Possibly, therefore, the action is of the same nature in all these cases ; and may be the phenomenon is related to the violet permanganate coloration produced by ultra-violet light and sunlight in window glass. The violet colour is in all cases destroyed by heating. X-ray bulbs of lead glass become brown in colour rather than violet. Elster and Geitel (1898) have suggested that the various colorations are due to ultra- microscopic particles of reduced metal in the salt. l Proc. Amer. Inst. Elect. Eng. Oct. 1913. CHAPTER VIII. THE MEASUREMENT OF X RAYS. The International Radium Standard. The general desire to have a standard by which the output of an X-ray tube could be measured in a manner free from the defects of the usual methods, led the Rontgen Society in 1909 to appoint a Committee (with Dr. W. Deane Butcher as secretary) to consider the question. This Committee decided to initiate standards of radioactivity. These de- pended on the -y-ray activity of radium bromide and were prepared by Mr. C. E. S. Phillips. Largely owing to the efforts of Prof. Rutherford, the question was taken up by the Congress of Radiology at Brussels in September 1910. An International Committee was formed with Prof. Ruther- ford as President ; in March 1912 the Committee met at Paris and adopted as an International Radium Standard a specimen consisting of 2T99 milligrammes of pure radium chloride which had been prepared by Mme. Curie. The radium is contained in a thin- walled glass tube, and use is made of the 'y-ray ionisation. The International Standard is preserved at the Bureau International at Sevres near Paris. Secondary standards are obtainable by the various nations who require them. The British Radium Standard. The British Radium Standard, consisting of 2T10 milli- grammes of pure radium chloride, has been certified in terms of the International Standard, and is now deposited at the National Physical Laboratory at Teddington. The radium salt is contained in a small glass tube, through which a platinum wire is inserted to dissipate accumulated THE MEASUREMENT OF X RAYS 89 electric charges (Fig. 52). The standard serves as a means of standardising radioactive preparations as well as the energy output of X-ray bulbs. In this connection it may be noted that Winawer and St. Sachs (P.Z. July 1915) have suggested that a beam of X rays should be regarded as having unit energy when, by its complete absorption in air, it produces the same number of ions as the 7 rays from 1 gramme of radium (B-f-C) would produce under similar conditions. O I 2 3 cms. FIG. 52. The British Radium Standard at the National Physical Laboratory. Standardisation of X-ray Bulbs. 1 The difficulty of standardising the output of X-ray bulbs by means of such an ionisation standard is chiefly one of specifying and reproducing the working conditions of the bulbs. Possibly the various makers could be induced to work to standard dimensions, but few would assert that the design of an X-ray bulb has reached or even approached finality. Moreover, even if agreement in design were secured, the performance of a bulb is peculiarly susceptible to slight variations in the prevailing conditions (see p. 72), over some of which control is scarcely possible. The whole sub- ject is receiving attention at the present time. The output from an X-ray bulb must be specified with respect to (1) intensity, i.e. quantity per unit area, and (2) hardness or quality. The X rays from a bulb consist of two main classes : (1) the heterogeneous "general" or "independent" radiation which depends in quality solely on the speed of the parent cathode rays ; (2) the homogeneous " characteristic " or " monochro- matic " radiations which are characteristic of the metal of the anticathode (p. 116). 1 For a full account of the various methods of measuring X rays (more especially for medical purposes), see Christen, Messung und Doaierung der Rontgenstrahlcn. 90 X RAYS The proportions of these two classes depend on the con- ditions of discharge, and on the metal of the anticathode. The general radiation is always present, and has a range of hardnesses which depends on the range of speeds of the cathode rays. The characteristic radiations only appear when the cathode rays are sufficiently fast ; their hardness depends only on the material of the anticathode. METHODS OF MEASURING INTENSITY. The intensity of the X rays at a particular point is defined as the energy falling on one square centimetre of a receiving surface passing through the point and placed at right angles to the rays. Rontgen showed, and the fact has been amply confirmed by later workers, that the in- tensity of a beam of X rays from a focus -bulb falls off as the inverse square of the distance from the anticathode. General Remarks on Intensity Measurements. It may be noted that almost all the methods of intensity- measurement, as ordinarily practised, are unduly favourable to the soft rays when regarded from an energy standpoint. The ideal method of test would afford an exact comparison of the energy of a hard X ray with that of a soft ray ; but what almost always happens is that the hard rays are not wholly arrested by the testing instrument, and hence show up relatively badly. In order to make a fair com- parison between two bulbs, all the rays given out by both should be taken 1 into account. The hard rays as well as the soft ones should be completely absorbed, in which case the measurements would give a fair estimate of the relative amounts of energy emitted from the bulbs. (1) Current through the X-ray Tube. A measure of the intensity of the X rays from a bulb may be obtained by measuring (with a milliammeter) the current passing through the tube, provided the potential difference is kept constant (see p. 182). Kroncke (A.d.P. March 1914), Davey (P.R. Sept. 1914) and Rutherford, THE MEASUREMENT OF INTENSITY 91 Barnes and Richardson (P.M. Sept. 1915) all agree that for a given voltage the intensity of the radiation is directly proportional to the current through the tube. Kroncke found that, except at low voltages, the following relation holds with sufficient accuracy for all practical purposes : where 7 is the intensity of the X rays, i the current, V the applied voltage, F the breaking-down voltage, and k a constant for the tube. This agrees with J. J. Thomson's formula (p. 133) that the energy of an X ray is proportional to the 4th power of the speed of the generating cathode ray. It is important in a set of comparative observations to keep the current in the primary circuit of the coil constant, for an increase in the current through the primary not only augments the intensity of the rays, but hardens the tube. It is difficult to estimate how any particular current- measurer will average up the peculiar pulsating current of a coil discharge (see p. 63). Salomonson (J.Et.S. 1912) has recently shown that both the form and frequency of the interruptions must be controlled in exact measurements. In usual practice, the methods for measuring intensity depend on one or other of the properties of the rays : heating, ionising, fluorescing, photographic, or chemical. (2) Thermal Methods of Measuring Intensity. The heat produced when X rays are completely absorbed by a metal was first measured by Dorn in 1897. Angerer (A.d.P. 1907) and Bumstead (P.M. 1908) have shown that the same amount of heat is generated by a stream of X rays, no. matter what the absorbing metal. The heating effects are minute, and can only be detected by instruments as sensitive as the radiomicrometer, bolometer, or radiometer. It will be seen that the method is only fitted for the research laboratory, and does not enter into ordinary practice. (3) lonisation Methods of Measuring Intensity. The exact mechanism of ionisation is even now not fully comprehended, but the outcome is the formation of positively 92 X RAYS and negatively electrified particles ions the presence of which imparts to the gas a conductivity that persists for some little time. The extent of the ionisation depends on the number of ions produced, and this is reflected in the degree of excellence with which the gas conducts. The generally accepted view of the formation of ions is that a negative nucleus (the electron) is broken off from the atom, leaving a positive nucleus ; each of these charged nuclei gathers round itself a cluster of gas molecules sometimes in considerable numbers and the resulting molecular aggre- gates constitute the gaseous ions, both positive and negative. At low pressures, the negative ion exists as the electron unencumbered by any attached molecules. An ionisation method of evaluating X rays thus resolves itself into the measurement of an electric current an opera- tion which can be carried out with such delicacy and con- venience that practically all recent workers have utilised this property of the rays. The ionised gas is subjected to an electric field which drives the two classes of ions positive and negative in opposite directions with velocities which depend on the strength of the field. The magnitude of the current generated by the motion of these charged particles depends to some extent on the potential difference of the surfaces between which the field is applied ; with small potentials, the two are roughly proportional, just as in cases of metallic conduction ; but with higher potentials the current responds less and less to the potential, and finally reaches a constant value called the saturation current (see Fig. 53). This is the current which should always be measured in practice, and care should accordingly be exer- cised that the potential difference applied to the surfaces is sufficient to give the saturation current. The electric field necessary increases with the degree of ionisation, but for most cases likely to arise in X-ray work, 100 volts per cm. is adequate. The shape of the first part of the current-potential curve is explained by the liability of a charged particle to encounter and coalesce with another of opposite sign before reaching one of the bounding surfaces. But this tendency, which THE MEASUREMENT OF INTENSITY 93 militates, of course, against the growth of the current, will be lessened if the speed of the particles is increased by putting up the voltage between the surfaces. For the higher the speed, the shorter the time of passage, and the less likely are the chances of recombination. Finally, with the saturation voltage, all the ions reach the boundaries, and the number arriving exactly equals the number produced in the same time by the X rays passing through the gas. This Applied PotennaJ FIG. 53. Diagrammatic representation of the relation between current and potential for an ionised gas. is not the case with the lower voltages, and thus only from a knowledge of the saturation current can we infer the true degree of ionisation that the rays have produced. With still higher potentials, the current rapidly increases until the sparking point is reached. On this steep part of the curve, both positive and negative ions acquire sufficient speed to produce fresh ions by colliding with the atoms of the gas. Thus, by working with potentials just insufficient to cause the passage of a spark, the original ionisation may be greatly increased a hundredfold or so. The plan has been adopted for the measurement of very feeble ionisations. Before adopting one or other of the various forms of ionisation chamber for any particular purpose, it is necessary to decide what we want. If it is the total energy of the rays that is desired, then we must arrange for the rays to 94 X RAYS * To Electrometer be completely absorbed in the gas of the chamber, if necescary by contriving a suitably long path, or by increasing the pressure of the gas, or, again, by choosing a sufficiently dense gas 1 : the total ionisation, we have reason to believe, is a satisfactory measure of the total energy in the rays provided certain conditions are satisfied (see p. 154). If, on the other hand, we wish merely to ascertain the ionising power of a beam of rays at some particular point, then almost any form of ionisation chamber will suffice. One convenient design is shown in Fig. 54. A circular thin aluminium sheet is mounted midway between two similar sheets which are raised to a potential of a few hundred volts by a battery of cells. The central sheet is carefully insulated and joined to an electrometer. It is easy to calculate the electric field with this shape of vessel, a state- ment that does not apply to the very common design made up of a cylinder provided with an in- sulated wire electrode along the axis. 2 In this latter form, the field, which is very strong near the wire, falls off a great deal towards the surface of the cylinder ; the applied potential must be very considerable to ensure a saturating field throughout the chamber. Ionisation currents produced by X rays are usually of the order of 10 ~ 10 to 10 ~ 15 ampere ; the exact amount varies a great deal according to the circumstances. For the larger currents, it is sometimes possible to use a sensitive galvano- meter 3 ; but in general it is much more convenient to 1 E.g. sulphur dioxide or methyl iodide are very useful for the purpose. 2 See, for instance, the comparison ionisation chamber in Fig. .73. 3 The most sensitive galvanometers yet introduced are the P&schen and the lEinthoven. The former, with a low resistance and a short period, will readily indicate 10~ 10 ampere. See Camb. Sci. List. Co.'s list. XRays Fio. 54. An ionisation chamber, showing earthed guard-tube in the insulation. THE MEASUREMENT OF INTENSITY 95 deduce the current from the change of potential as measured by means of a Dolezalek quadrant electrometer or some form of gold-leaf electroscope. With an electrometer and a suitable condenser, currents from 10 ~ 8 to 10 ~ 14 ampere can be measured. For smaller currents down to 10 ~ 17 ampere an electroscope is better. Of the electroscopes, the C. T.. R. Wilson tilted variety l is convenient and sensitive, and possesses a small capacity. Some observers use electroscopes provided with aluminium windows, the X rays being sent directly into the electro- scope instead of into a separate chamber. The leaf in this case is charged to a high potential, and its rate of leak to the outer case is measured. There are on the market several " direct reading " X-ray quantimeters of this type, which are convenient for comparative measurements but are not capable of accurate absolute work. All the various instruments require to be calibrated, and their capacity (as well as that of the ionisation vessel) determined, before the currents can be deduced from the potential measurements. The French workers largely employ the late Prof. Curie's pidzo-electrique, in which the electricity generated by gradu- ally relieving the tension on a stretched quartz lamina is balanced against the ionisation current to be measured. The method requires considerable manipulative skill. 2 (4) Photographic and Fluorescence Methods of Measuring Intensity. Practically all the earlier workers used photographic or fluorescence methods of measuring the intensity of their X rays, but nowadays these methods, at any rate for most purposes, have been displaced by ionisation methods. An ordinary photographic plate is incapable of arresting and recording the hardest kinds of X rays, and therefore, from an energy standpoint, the softer rays are given undue weight when a heterogeneous beam is used. We need, therefore, to exercise care in drawing conclusions from the density of the photographic image as to the intensity of the rays. Moreover, Barkla and Marty n (P.M. 1913) have shown that 1 See, for example, the Camb. Sci. Inst. Co.'s list of electrometers, 2 3ee Rutherford's Radioactive Substances, 1913, 96 X RAYS if the X rays are just sufficiently hard to excite the radiations characteristic of silver or of bromine (the heaviest constituents of a photographic film), they are selectively absorbed and the photographic effect is greatly enhanced. X rays a little softer than this do not excite the charac- teristic rays, and are, therefore, recorded disadvantageously. Thus the photographic action may not be proportional to the absorption of the X rays by the sensitive film. 1 As far as practical difficulties are concerned, it should be remarked that the emulsion on an ordinary plate may vary in thickness by as much as 10 per cent., through want of flatness of the glass backing. This can be reduced to the order of 5 per cent, by the use of patent plate glass and the exercise of special care in the coating. The slower fine grained plates are to be preferred for more precise work, and, of course, one should adhere to some standard de- veloper and method of development. To the worker with limited resources the photographic method of measuring intensity offers advantages because of its simplicity. Some form of opacity-meter for obtaining a measure of the density of the image is the chief requirement. The opacity meter measures the extent to which a standard beam of light is cut down by the photographic film whose density is required. If 7 is the intensity of the testing light which is incident on the developed film, and I t that of the transmitted light, then, if \x is the fraction of the energy which is absorbed by a very small thickness, #, of the film > 7 _/ ,-Ad L t~ L tf> ; where d is the thickness of the film 2 (see p. 103). The film is assumed equally dense throughout its thickness. For films of uniform thickness, d is constant, so that X is proportional to log (/ // ( ). X is called the absorp- 1 See Voltz (P.Z. Aug. 1915), who investigated the point and considers the photographic measurement of X rays should be discarded. Kroncke (A.d.P. March 1914) came to a similar conclusion. He obtained 50 per cent, discrepancy between ionisation and photographic methods. 2 More precisely, this assumes monochromatic light. X is different for different wave-lengths. THE MEASUREMENT OF INTENSITY 97 tion coefficient ; (/ /^) * s known as the opacity, 1 and equals the number of times the incident light is cut down. Log (I /I t ) is termed the opacity-logarithm. Now, by definition, X is proportional to the density of the image, i.e. to the amount of silver per unit area of film. Thus the ratio of two opacity-logarithms gives the ratio of the film densities, and therefore the ratio of the photographic energies in the two cases. The opacity meter is graduated to read directly in opacity-logarithms. In fluorescence methods the luminosity is matched against some standard fluorescence excited by a steady source of radiation such as radium. The drawback to such methods is that the fluorescing salt becomes " tired " under the action of the rays. The sensitivity of a screen may also vary con- siderably from point to point, so that it is difficult to make a fair comparison. Barium platinocyanide is the material commonly used to sensitise a fluorescent screen. This salt, which has the formula BaPt(CN) 4 ,4H 2 exists in three different forms, of which the green crystalline variety is by far the most efficient for fluorescent purposes (Levy, J.Rt.S. 1916). (5) Methods of Measuring Intensity used in Medicine. In the therapeutic' use of X rays, various chemical re- actions brought about by the rays have been suggested and employed from time to time as aids to " dosage " ; for example, the discolouring of various alkaline salts (Holz- knecht, 1902) ; the liberation of iodine from a 2 per cent, solution of iodoform in chloroform 2 (Freund, 1904 ; Bordier and Galimard, 1906) ; the darkening of a photographic paper (Kienbock), see p. 96 ; the precipitation of calomel from a mixture of mercuric chloride and ammonium oxalate solutions 2 (Schwarz, 1907) ; and the change of colour of pastilles of compressed barium platinocyanide (Sabouraud- Noire and Bordier). X rays resemble light in their pro- perty of lowering the electrical resistance of selenium ; this property, if the pronounced fatiguing of the selenium could 1 The transparency is the reciprocal of the opacity. 2 X rays share this property with Ra rays and ultra-violet light, 98 X RAYS be overcome, would doubtless furnish the basis of a very convenient method of measurement. 1 It must be admitted that most of these methods, if not all, provide nothing more than the roughest notion of the intensity of a beam of ordinary heterogeneous X rays. Of all the various intensity-measurers, the pastille finds most favour with medical men. The barium-platinocyaiiide discs are some 5 mms. in diameter, and their colour, initially a bright green, changes, when exposed to the rays, to a pale yellow, and finally to a deep orange. The pastille is placed at a specified distance from the anticathode of the bulb, and the colour is matched against one of a number of standard tints. The method is extremely easy in prac- tice, and is fairly reliable as a guide for short exposures, but it is not very trustworthy for times exceeding ten minutes or so. The pastille method is defective in that it attempts to measure rays of all qualities by a surface colora- tion. Other platinocyanides show similar colour changes when exposed to X rays. Levy has shown that the change of colour is due to a change from the crystalline to the amorphous condition. If the pastille is put aside, the reverse change slowly takes place, especially in the presence of light, so that the pastille should not be exposed to full daylight during the X-ray treatment. Ultra-violet light and radium rays cause similar browning in such pastilles. The following table gives an idea of the relation between the different dosemeter scales : 5H units 2 (Holzknecht ; alkaline salt) = Tint B (Sabouraud-Noire ; pastille) = Tint 1 (Bordier ; varnished pastille) = 3 to 41 (Bordier and Galimard ; iodine solution) = 10X units (Kienbock ; photographic plate) = 3-5Kaloms (Schwarz ; mercury solution) . = Villard dose. 1 Fiirstenau (P.Z. Aug. 1915) claims to have got over this difficulty. 2 Unit 1H = one-third of the radiation necessary to set up the first signs of reaction in the healthy skin of the face. THE MEASUREMENT OF QUALITY 99 METHODS OF MEASURING QUALITY OR HARDNESS. The range of qualities of X rays is very wide, as would be inferred from the fact that, while some rays are unable to penetrate more than a centimetre or two of air at atmospheric pressure, others have been detected at dis- tances of 100 metres or more. The hardness of a bulb is solely dependent on the potential difference between the electrodes : an account of the various methods of controlling this potential difference is given on p. 72. There is some reason for believing that it is the maximum potential difference which is the important factor, but further experiments on this point are needed. The potential difference between the electrodes controls- the speed of the cathode rays. In the case of general X rays, the quality and energy depend on the fourth power of the speed of the exciting cathode rays. In the case of the characteristic radiations, the quality can be defined rigorously in terms of the atomic weight of the anti cathode. It is found that a certain minimum voltage on the tube is required to excite a particular radiation. There is thus a critical cathode-ray velocity for each char- acteristic X ray : slower cathode rays can only excite inde- pendent " rays " ; faster cathode rays are, within limits, increasingly effective generators of the characteristic rays, but with very high-speed rays the " independent " radiation is once again generated. The subject is dealt with later (p. 132), but it may here be mentioned that the critical cathode-ray speed is proportional to the atomic weight of the anticathode. If E is the potential difference to which a cathode ray owes its velocity (v), then the two are connected by the energy equation \m. v* = E .e, where e and ra are respectively the charge and the mass of the cathode ray. 100 X BAYS Taking e/m= 1-77 xlO 7 , E in volts and vm cms. /sec., ^ = 2-82^. 10~ 16 or v=5-95v/. 10 7 . A series of values of cathode-ray velocities and potentials up to 200,000 volts is tabulated on p. 262. The potential difference on a tube may be measured by ,a high-potential electrostatic voltmeter, of which there are now one or two excellent examples on the market. Or, . failing this, the length of the alternative spark-gap may be noted. It does not, however, follow that two bulbs having the same equivalent spark-gap will give out beams of the same composition. Usually they will not, owing to differ- ences in the gas-pressure and the disposition of the cathode, which two factors may or may not counteract each other. Table V. gives the approximate sparking voltages in air at atmospheric pressure and room temperature. Too much reliance must not be placed on the figures, as the results of different experimenters do not agree well, probably owing to the difficulty of measuring the potential. The values for the needle-point electrodes are for alternating current of sine form (see " Standardisation Rules," Amer. Inst. Elec. Eng. 1915). For alternating currents, the striking distance is most probably governed by the maximum voltage, which is accordingly given in the table rather than the effective (root-mean-square) value [*-(max.)/l-42]. For ball electrodes, the most recent and reliable measure- ments of the maximum spark-potentials for alternating current come out about 5 per cent, smaller l than those for direct current, the values for which are given in the remain- ing columns of the table. 2 These latter results refer to smooth polished metal balls of the same size. An inspection of the table shows that, in general, the spark passes more readily, the smaller the ball ; and that short spark-gaps require proportionately more potential than long. The measurements are taken in the absence 1 See Kowalski and Rappel (P.M. 1909), who employed balls up to 30 cms. diameter. 2 Based largely on the results of Algermissen (A.d.P. 1906) and Topler (A.d.P. 1907). THE MEASUREMENT L OF QUALMS 101 of any visible brush-discharge, a condition essential for definite sparking. It is better in practice to use mode- rately large balls than small, as with the latter, brush- discharge tends to occur, more especially at the negative pole : such glow is, of course, a prominent feature with needle- point electrodes. The needle-point spark-gap often supplied with induction coils, while it enhances the apparent capa- bilities of the coil, is not suitable for measuring purposes. If unequal-sized balls are used, the smaller electrode controls the spark-gap for moderate lengths of spark : the larger ball should be made the negative electrode (see p. 182). TABLE V. SPARKING POTENTIALS. Spark-gap. Diameter of Balls. Needle-pts. 0'5 cm. 1 cm. 2 cm. 5 cm. cm. inch. A.C. volts. B.C. volts. D.C. VOlts. D.C. VOltS. D.C. volts. (Max.) 0-1 04 1,000 5,000 5,000 5,000 5,000 0-2 08 2,000 8,000 8,000 8,000 9,000 0-3 12 4,000 11,000 11,000 11,000 12,000 0-4 16 5,000 14,000 14,000 14,000 15,000 0-5 20 6,000 16,000 17,000 17,000 18,000 0-6 24 7,000 17,000 20,000 20,000 21,000 0-7 28 8,000 18,000 22,000 23,000 24,000 0-8 31 10,000 19,000 24,000 26,000 27,000 0-9 35 11,000 20,000 26,000 29,000 30,000 1 39 12,000 21,000 27,000 31,000 33,000 2 79 24,000 24,000 36,000 48,000 57,000 3 1-18 34,000 26,000 42,000 58,000 77,000 4 1-58 42,000 27,000 45,000 65,000 93,000 5 1-97 49,000 Brush 47,000 71,000 105,000 6 2-36 55,000 discharge Brush 77,000 116,000 7 2-76 61,000 usually discharge 82,000 125,000 8 3-15 66,000 occurs. usually 87,000 133,000 9 3-54 71,000 occurs. 91,000 140,000 10 3-94 76,000 95,000 145,000 15 5-91 102,000 Brush 170,000 20 7-9 122,000 discharge 190,000 30 11-8 170,000 usually 40 15-8 220,000 occurs. - i 102 'X RAYS Trowbridge (P.M. 1898) found a spark-length of 200 cms. with a potential of 3,000,000 volts. With very long sparks, the shape of the electrodes (if of moderate size) is im- material. For instance, with a potential of 240,000 volts, Jona obtained the same sparking distance (47 cms.), whether a point and plate or two balls (2 cms. diam.) were used. It is probable that the minimum voltage required to produce a given length of spark is less for a Wimshurst machine than for an induction coil on account of the rapid variations in the potential of the latter. In the case of an X-ray tube, as the break-down voltage is higher than the running voltage, it is doubtful what precisely either voltmeter or spark-gap affords. With a pulsating current, we need to know the shape and abruptness of the potential curve, as well as the proportion of time between the impulses, before we can estimate the effective potential (see p. 63). It is probable, however, that, at any rate in the case of a hard tube, either instrument indicates a value not very far from the maximum potential (see p. 100), and that the bulk of the X rays are generated by cathode rays with a velocity which they owe to this maximum potential rather than to a mean potential (see p. 16). (1) Wave-length. We have good reason now for believing that X rays and light are identical, and that the hardness or penetrating power of an X ray is precisely defined by its wave-length : the shorter the wave-length, the harder the ray. The sub- ject is dealt with elsewhere (p. 194), but it has been shown by many observers that X rays may be diffracted by the invisible parallel planes of atoms in the interior of a crystal. From a knowledge of the distances separating the atoms, we can arrive at the wave-lengths of the X rays. W. L. Bragg (P.R.S. 1913) has calculated the atomic distances in the case of rock-salt (p. 223), and the wave-lengths of the X rays so far examined are found to lie between 10 " 7 cm. and 10 ~ 9 cm. A complete table will be found on p. 226. THE MEASUREMENT OF QUALITY 103 (2) Absorption- Coefficients. The customary way of specifying the character of X rays is to measure their absorption in a sheet of aluminium of definite thickness. Aluminium is not an ideal standard of reference, but it is chosen because it is readily procurable in convenient form, and, so far as we know, does not, in the majority of cases, complicate matters unduly by super- posing a characteristic radiation. Now it is found that if all the rays both entering and leaving a plate of material are homogeneous (that is, wholly of the same quality), then the rays are absorbed exponentially by the plate, i.e. if 1, 2, 3, ... similar sheets are successively introduced, each additional sheet absorbs the same fraction of what it receives. In other words, if there is no " scatter- ing " or transformation of the X rays, and if \x is the fraction of the intensity which is absorbed when the rays pass nor- mally through a very thin screen of thickness x (cm.), then for a plate of thickness d (cms.), 11 p~ Kd L J.Q . e , in which 7 is the intensity of the beam when it enters, and / that of the beam when it leaves the screen. e( = 2-72) is the 'base of the hyperbolic system of logarithms. X is termed the linear absorption coefficient. 1 2-3 It follows that X = =- (log 7 - log /) ; the logarithms are d to base 10. If in a set of observations with homogeneous rays, log I is plotted as ordinate against d, the graph is a straight line and X is 2-3 times the slope of the line. With ordinary heterogeneous rays, X is greater for thin screens than for thick, and so we can only deal with an average X, which, however, varies more and more slowly as the screen becomes thicker. The logarithmic curve of absorption for heterogeneous rays, such as are given out by an ordinary X-ray bulb, is not a straight line, but a curve which is steeper for thin screens 1 The precise physical interpretation of an exponential law of absorption is not so simple as its compact and convenient mathematical expression would lead one to suppose. 104 X RAYS than for thick. The general shape is rather steeper than the heavy curve in Fig. 64. For a method of finding analyti- cally the absorption coefficients of the constituents of a complex beam of rays, see J. J. Thomson, P.M. Dec. 1915. In the case of the characteristic radiations, an element exhibits a maximum transparency for each of its own char- acteristic radiations. For slightly harder rays than these, the absorption rapidly increases, the rays characteristic of the screen are produced and superposed on the transmitted rays to an extent which diminishes as the incident rays are increasingly hardened. For incident rays softer than the critical type, no characteristic rays are produced. Thus, as the incident rays are gradually hardened, the transmitted rays reach a maximum intensity when the incident rays are of the same quality as each of the characteristic rays in turn. A large value of X corresponds to easily absorbed rays, and a small one to very penetrating rays. X also varies with the nature of the absorbing screen, so that it is neces- sary to specify the material used. For medical purposes, it has recently been suggested that water should be chosen as the standard absorbing medium, since the absorptive power of water agrees closely with that of animal tissue. Some workers prefer to think in terms of the thickness, Z>, which reduces the intensity to half value. D is con- nected with X by the expression Z) = 0-69/X. A notion of the order of values of X may be got from the fact that for an X-ray beam of average hardness X A1 lies between 4 and 8 cm." 1 ; for hard rays between 2 and 4 cm.' 1 . X for fatty tissue varies from about 0-4 for hard rays to 0-7 for medium rays. 1 cm. of flesh absorbs from 30 to 90 per cent, of X rays. A table connecting I /I and Xd is given on p. 261. A more fundamentally important constant is obtained by dividing the absorption coefficient (X) by the density (p) of the absorbing screen. 1 This quantity, X/p usually called the mass- absorption coefficient gives a measure of the absorp- tion per unit mass of the screen for a normally incident pencil of rays of unit cross section. Since it is mass alone that affects absorption, at any rate as determined by the usual 1 For the standard material aluminium, p2'7. THE MEASUREMENT OF QUALITY 105 methods of measurement, it is more profitable to use mass- coefficients than linear-coefficients. If, as was at one time supposed, the absorbing powers of different materials were truly proportional to their densities, then for the same rays A/^o would be a constant, no matter what the substance used as screen. In point of fact, dense substances are a good deal more absorptive, mass for mass, than light, and \/p increases rapidly with the atomic weight of the screen. 1 The increase is more noticeable with hard rays than with soft (see also p. 138). Benoist (J.d.P. 1901) was the first to examine systemati- cally the absorption of a beam of ordinary heterogeneous X rays in various absorbing elements. For our purposes it is convenient to translate his results into quantities pro- portional to absorption coefficients ; and, when this is done, Fig. 55 is the result. It will be noticed that \jp increases steadily with atomic weight both for hard and soft X rays. For example, with hard rays, lead is twenty-five, silver eighteen, and copper eight times as absorbent as an equal mass of aluminium. There is a region of abnormal absorption round and about silver : the two beams of X rays, one hard, the other soft, are, moreover, absorbed almost equally by the silver group of elements. The explanation of this, as will be seen later (p. 142), is bound up with the amount of secondary radiation that silver emits under ordinary conditions. Absorption Coefficients of Heterogeneous Eays. But it may be urged that although characteristic rays have a perfectly definite X, the X rays from a coil-driven bulb are so very far from being homogeneous that the absorption coefficients as above defined are not particularly useful in X-ray practice. However, considerable guidance can be obtained from a knowledge of even an average value of X, calculated though it may be, on loose assumptions. 1 A similar relation holds for the soft y rays from radium. For hard y rays, a density law holds, and \/p is constant, except for the heaviest metals, which are a little more absorbent. In other words, these very penetrating rays almost entirely ignore atomic structure. For hard 7 rays, ^/p = 0-04 for all absorbing substances with an atomic weight less than 100. 106 X RAYS And, moreover, while it is true that the X rays from a bulb are in general very heterogeneous, they become less so as the spark-gap is increased. 1 The rays from a bulb with a spark-gap of some centimetres, which are transmitted by 06 50 100 ISO 200 250 Atomic Weight of Absorbing Element Fio. 55. Graph (derived from Benoist's transparency curve) connecting absorption with atomic weight and displaying a region of selective absorption. an aluminium screen 5 mms. or more thick, are very fairly though not strictly homogeneous. Absorption Coefficients in Air. On a density basis, the absorption by 1 cm. of air under 1 The explanation of this is probably bound up with the simpler char- acter of the magnetic spectrum of the cathode rays at low pressures (see p. 16). THE MEASUREMENT OF QUALITY 107 cm. of ordinary conditions is equivalent to that by water. Eve and Day (P.M. 1912) have measured at various ranges (up to 100 metres) the absorption coefficients in air of the rays from X-ray bulbs of different degrees of hard- ness. Table VI. contains some of their results. TABLE VI. ABSORPTION COEFFICIENTS IN Am. Distances from X-ray Bulb. Alternative Spark-gap. 4 to 10 metres. 20 to 40 metres. 40 to 60 metres. A A/p A A/p A A/p 1-5 to 5 cm. ro-oois p-4 (soft bulb) - to no l^o-ooio [0-8 11 cm. (med- ium bulb) - 0-00040 0-32 0-00040 0-32 0-00029 0-23 30 cm. (hard bulb) 0-00029 0-23 0-00027 0-21 0-00014(?) o.ii(?) Eve and Day note that X = 0-0004 is a good value for radiographic work ; but rays whose X is 0-0003 are too penetrating for such a purpose. The above values may be compared with those of Chadwick (P.P.S. 1912) for the Ra 7 rays when absorbed by air. His values of X for air are 0-000062 and 0-000059 cm.' 1 in the case of rays which have previously traversed 3 mms. and 10 mms. of lead respectively. The absorption coefficients for the various characteristic radia- tions are given on pp. 119and 138. (3) The Benoist Penetrometer. Among medical men Benoist's _ . ., radiochromometer orpenetrometer FIG. 56. Benoist s penetrometer. enjoys extensive use as a measurer of hardness. It consists of a thin silver disc 0-11 mm. thick, surrounded by twelve numbered aluminium sectors from 1 to 108 X RAYS 12 mms. thick (Fig. 56). The X rays are sent through the instrument, and the observations consist merely in matching on a fluorescent screen or photographic plate the image cast by the silver disc against the images of the aluminium plates : the thickness of the matching sector increases with the hardness of the rays. 1 " Benoist No. 4 to 5 " is a good average hardness for curative work, while for general radiography No. 6 on Benoist's scale is useful. A notion of the discharge potential across a tube may be got from the very rough relation that the voltage is from 8,000 to 10,000 times the Benoist reading of the X rays. Table VII. gives very approximately the relation between Benoist numbers and absorption coefficients in aluminium. TABLE VII. Benoist No. Mean A/p A j. 2 20 3 8 4 4 6 2 8 1-5 10 1 12 0-5 (4) Other Penetrometers. Walter Penetrometer. This consists of a number of holes in a lead disc, which are covered by a sequence of platinum discs of gradually increasing thickness. The Wehnelt Penetrometer agrees in principle with that of Benoist. An aluminium wedge is used instead of a series of stepped sectors. Rontgen in his third memoir described an instrument essentially similar to Benoist's, with the exception that the comparison metal was platinum instead of silver. 1 Benoist (C.R. 1902) based the theory of his instrument on the curves displayed in Fig. 55, which go to show that while the transparency of aluminium alters a good deal with the quality of the X rays, silver is almost equally transparent to both hard and soft rays. As Table XL shows, the assumption of a constant transparency for silver is by no means correct. THE ENERGY OF X RAYS 109 Christen' s Half-value Penetrometer. Christen adopts as a definition of quality the thickness of a layer of water (or, in actual practice, bakelite), which will reduce the intensity of a beam of rays to half its original value. The rays are sent through a stepped wedge of bakelite, alongside which is a perforated metal plate. This provides a standard of reference on a fluorescent screen, the two images being side by side. The holes in the plate are so designed that the area of the metal removed equals that which remains, so that the plate by this means reduces the intensity of a beam of rays to half -value. The holes are small enough to produce uniform illumination on a screen placed a short distance behind the plate. Bauer Qualimeter. This is an electrostatic voltmeter, which serves to measure the potential difference between the electrodes of a tube. (See p. 183.) The Klingelfuss Qualimeter consists of an auxiliary search coil and electrostatic voltmeter. The instrument works similarly to the Bauer. The hardness-numbers of the various penetrometers are all much the same as Benoist's, except those of Wehnelt, which are 50 per cent, bigger for the same quality of rays. THE " ENERGETICS " OF AN X-RAY BULB. When a stream of cathode rays strikes an anticathode, the different rays suffer a variety of fates. By far the greater number merely fritter away their energy until it becomes too small to render them distinguishable : the heat generated at the anticathode is ample proof that the energy of the cathode rays is mostly dissipated into heat. A large proportion of the cathode rays, in some cases as much as 75 per cent., are " reflected " by the anticathode in all directions against the glass walls of the tube with velocities which may be anything up to the original speed of the cathode rays. There is good evidence for believing that a cathode ray can pass through many atoms without being in any way deflected or transformed. The fate of the cathode ray is to some extent dependent on the material 110 X RAYS of the anticathode. The heavier the atom the more capable it is of swinging round a cathode ray which endeavours to pass it. Only a small proportion of the cathode rays are effective in producing X rays. The chances that a cathode ray will ultimately come into suitable conflict with some atom and so generate an X ray are slight. In fact, the efficiency of the present methods of generation of X rays is very low : the X ray is merely a small bye-product in the energy transformations of a Rontgen tube. Wien, Angerer, and Carter (A.d.P. 1905 and 1906) have worked independently at the subject. They agree that the ratio of the energy of the X rays to that of the exciting (heterogeneous) cathode rays is of the order of -nnnr ; Carter found that the efficiency increases with the hardness (is, in fact, proportional to the voltage on the tube), is independent of the current, but increases with the atomic weight of the anticathode. This value of the efficiency is not inconsistent with the estimate of Eve and Day (P.M. 1912), who remark that, of the energy supplied to an ordinary X-ray bulb, not more than about 1 in 20,000 is contained in the X rays as measured by their ionising ability. The efficiency of a soft bulb is probably even less than this, for more energy is converted into heat with low-speed cathode rays. The energy of general X radiation is pro- portional to the 4th power of the velocity of the exciting cathode rays (see p. 133). Winawer (P.Z. Nov. 1915) claims from his experiments that to work an X-ray bulb to greatest advantage, the exciiing current should for a given load : (1) be made up of short duration impulses as widely separated as possible, (2) produce the greatest heating of the anticathode. Beatty's Experiments. Beatty (P.R.S. Nov. 1913) carefully evaluated the energy of X rays in terms of the velocity of the parent cathode rays and of the atomic weight of the anticathode. Homogeneous rays of known velocity were sifted from a cathode stream THE ENERGY OF X RAYS 111 by a magnetic-spectrum method, and fell upon one of a number of anticathodes affixed to a sliding tray. The X rays so produced passed through a sheet of Al foil 0-0002 cm', thick, and were completely absorbed in an ionisation chamber consisting of a cylinder over a metre long filled with the vapour of methyl iodide. The resulting total ionisation was taken as measuring the energy of the X rays. The ionisation current was balanced against a fraction taken from the primary cathode-ray current by a variable shunt, so that a null deflection was obtained in the electroscope (see p. 134). Thus reliable readings could be obtained even when the cathode-ray current was very irregular. The ionisation which the cathode rays would have pro- duced in methyl iodide vapour was deduced from the work of Glasson (p. 12) and Whiddington (p. 10) on the passage of cathode rays in air. The result finally established by Beatty was where 1^ = energy of the X rays, E c =-- energy of the parent cathode rays, A= atomic weight of anticathode, 3 = velocity of the cathode rays expressed as a fraction of the velocity of light (i.e. 3xl0 10 cms./sec.). Thus, for an anticathode of platinum (with an atomic weight of 195) and a bulb of medium hardness, with cathode rays of speed, say 7-5 x 10 9 cms. /sec., TTr V manages to escape from the surface of the radiator. Thus the arrangement is very inefficient as a source of characteristic rays. The writer showed, however, in 1908 (P.T.) (see p. 36) that a large proportion of the radiation from the anticathode of an X-ray bulb may consist of the characteristic radiation of the metal of the anticathode, more especially if the bulb is soft. By the employment of screens of the same metal as the anticathode, the other radia- tions present are either absorbed or transformed into the char- acteristic radiation, the result being an intense and almost pure beam of characteristic rays . The potential on the tube should not be too high, otherwise the proportion of heterogeneous primary rays in the emitted beam will increase in amount. Figs. 62 to 65, taken from the above paper, give the log- absorption curves for three such different metals as Al, Cu, and Pt. The homogeneity of much of the radiation, when screen and anticathode are alike, will be apparent. With all three metals, there is a superposed softer homogeneous radiation, which is removed by quite thin screens. In the case of copper, the K radiation shows up prominently. Figs. 64 and 65 show the way in which the radiation from a platinum anticathode is absorbed by aluminium and platinum screens respectively. In the former case, the (thick) absorption curve betrays no apparent homogeneity in the rays. It is, however, possible to analyse the curve into three homogeneous components, having \/p=5'6, 23'7, and 70 respectively. These are represented both in amount and hardness by the three thin lines. The hardest is probably independent radiation, the second proves to be the character- istic L radiation of platinum. With the platinum screen, the independent radiation has disappeared, and the absorp- tion curve shows that the X rays transmitted by a screen 0*0005 cm. thick are almost entirely homogeneous L rays. 1 See, for example, Fig. 61. 126 X KAYS Al Anhc^thode , Al Screen 7mm Sp^rk, 23000 volts. O-Oi cm o 02 Thickness 005 of FIG. 02. The heavy curve is the log-absorption curve, for an Al screen, of the X rays from an Al anticathode. The curve can be resolved into two homogeneous components (indicated in amount and absorbability by the two thin straight lines). \ Cu AnNckrhode,Cu Screen >s' 5 i \ 7mm Sp^rk ,23000 volte. \ X _c O IO j \ x V \ ^Vv Yp-487 (K RAdi&rion) ' 00 ft 1 Q'5 \ \ N \-9IO Vp-102 ^ V \ X 50% 30% 20% 10% 5% 2% o-oo^cm o-oc4 o-ooe Thickness of Screen O-OOB FIG. 63. The heavy curve is the log-absorption curve, for a Cu screen, of the X rays from a Cu anticathode The homogeneity of much of the radia- tion is apparent. The curve can be resolved into two homogeneous com- ponents (indicated in amount and absorbability by the two thin straight lines), one of which is the K radiation of Cu. CHARACTERISTIC X RAYS 127 2C Pr Anhc\rhode , Al Screen 7mm.Spd.rk, 23OOOvolr5. (LR6.dio.rion) 100% 80%, 50% 3O% 20% A 10% < O-O4 O-O6 Thickness of Screen FIG. 64. The heavy curve is the log-absorption curve, for an Al screen of the X rays from a Pt anticathode. The curve can be resolved into three homogeneous components (indicated in amount and absorbability by the three thin straight lines), one of which is the L radiation of Pt. 4 On o 0-5 Pr Ann cathode , Pr Screen 7rnm Sp^rk , 230OO volts . X=6500 ^=300 X-2300 OOOO4CIT1 O OOO8 OOOI Thi duress -of Screen O-OOIG 100% 80% J 50% 30% 20% 5 7. 0-0020 FIG. 65. Conditions as in Fig. 64, except that the Pt X rays are absorbed by a Pt screen instead of Al. The homogeneous L radiation now predominates in the absorption curve, which is resolvable into the L radiation and a second softer component. 128 X RAYS The proportion of the L radiation from a platinum anti- cathode varies with the voltage on the tube. At about 11,000 volts the L rays constitute about 35 per cent, of the whole radiation transmitted through an aluminium window 0-0065 cm. thick ; at 32,000 volts, 63 per cent. ; and at 50,000 volts, 40 per cent. If the rays have to traverse the glass walls of an ordinary X-ray tube, the proportion of the L radiation is small, except at low voltages. With the higher voltages the proportion of hard radiation is con- siderable. This " end radiation " increases both in hardness and amount with the voltage. In the case of copper, nickel and iron radiations trans- mitted through an aluminium window as above, the pro- portion of K radiation amounts to between 80 and 90 per cent, at voltages between 20,000 and 30,000. In the case of all three metals the K radiation would be almost wholly absorbed by the glass walls of an X-ray bulb, which are usually between J mm. and 1 mm. thick. The relation between maximum hardness and voltage is of importance in connection with Planck's quantum theory (p. 245), from which we should anticipate that the maximum X-ray frequency would be reached with a voltage which diminishes with the atomic weight. On the experimental evidence available at present, we cannot say with certainty that there is a limit to the hardness of the X rays which may be generated by any particular anticathode. We know, however, that radiations much harder than the K radiations can be obtained from various metals. For example, the writer has obtained from aluminium, rays which have X//OAI = * 5 from iron, X//OAI = 13-5 ; from nickel, A//> A1 = 10-8 ; from copper, Coolidge Tube. Fig. 66, due to Rutherford, Barnes and Richardson (P.M. Sept. 1915), shows a set of log-absorption curves dealing with the absorption in aluminium of the X rays from a Coolidge tube (see p. 44) run at a variety of voltages. Just as with an ordinary tube, the radiation in all cases diminishes more rapidly with thin screens than with thick, CHARACTERISTIC X RAYS 129 owing to the absorption of the softer components. It is not until the radiation is reduced to about -5-577 of its initial value that the absorption curve becomes very nearly a straight line, at any rate with the higher voltages. This shows that the " end-radiation " is approximately homo- geneous. At the lowest voltages the main radiation proved to be the L characteristic radiation of tungsten, that being the material of the anticathode. At high voltages the radiation is mainly of the independent type. 10.000 |0 IS 20 25 30 x Thickness of Aft/minium fnmms. 35 40 FIG. 66. Log-absorption curves, for Al screens, of the X rays from a Coolidge tube with tungsten anticathode. The hardest X ray obtained by Rutherford (1916) had an absorption coefficient A A1 = 0-23 cm." 1 and X Pb = 9 cm." 1 . The following wave-lengths (see p. 195) were measured : Volts. 13,000 50,000 100,000 140,000 Shortest Wave-length Detectable. 1-02 xlO~ 8 cm. 0-299 0-188 0-171 130 X RAYS Quality of Characteristic Rays in terms of Parent Cathode Ray Velocity. The writer in 1909 (see J.Rt.S. 1913) attempted to associate the hardness of the characteristic radiation emitted by an anticathode, with the speed of the cathode ray required to excite the radiation. The underlying notion was that unless the cathode rays possessed a velocity greater than a certain critical value, no charac- teristic rays would be generated. If this were so, the X rays could, so to speak, be labelled in terms of the speed of the exciting cathode rays. Obviously a first sim- plification was to work with cathode rays of uni- form speed. This can be done by the use of either (1) an influence machine or (2) a magnetic-spectrum method applied to a coil discharge. In the latter plan, the cathode ray energy is, at suitable pressures, largely con- centrated in the fastest cathode rays (see p. 16) ; and the method had other obvious advantages which led to its adoption. The apparatus is indicated in Fig. 67. The cathode rays from (7. were spread by a magnetic field into a mag- netic spectrum, plainly visible along the plate anode, AS, which was coated with willemite. By varying the strength of the field, any part of the spectrum could be brought over the slit, 8. The pencil of cathode rays which passed through 8 impinged on the anticathode, T, below, and a bundle of X rays passed out through the thin aluminium window, W, and was measured by an ionisation method. Some half-dozen anticathodes were mounted on a trolley as o. 67. Apparatus for showing production of X rays with cathode rays of varying speed. CHARACTERISTIC X RAYS 131 described on p. 36. The additional cathode K was pro- vided to bombard the anticathodes so as to liberate the occluded gas, which otherwise, by its continued emission, softens the tube during the actual measurements. The experiments, which were arrested soon after their com- mencement, served, however, to show the extreme inefficiency of the slowest cathode rays as producers of X rays. As the different parts of the cathode spectrum were passed over the slit, and faster and faster cathode rays were brought into action, the rapid gain in the intensity of the X rays was very noticeable. The increase in intensity came in quite suddenly for some one speed of the cathode rays which did not appear to be the same for the different anticathodes employed. Whiddington's Experiments. In 1910 Whiddington carried out a research on somewhat similar lines, and obtained quantitative measurements of Brs.ss Cylinder Solenoid for Magnetic Field / 'Ammeter _^ GaWnomeferf/J ToEwrti ToEleclroscope' FIG. 68. Whiddington's apparatus for connecting the speed of cathode rays with the quality of various characteristic radiations. great importance (P.R.8. 1911) for the K radiations of a number of elements. His final apparatus is shown in Fig. 68. The cathode-ray spectrum was produced by a solenoid 132 X RAYS which yielded a uniform and calculable magnetic field. The anticathode was of silver, and the generated X rays struck a secondary radiator. The speed of the cathode rays was increased (by the hardening device described on p. 74) until the secondary radiator emitted its characteristic radia- tion, which of course was duly indicated in the ionisation chamber. Below this critical value of the velocity, there was little effect in the chamber ; above it, the ionisation current grew verj^ rapidly. Thus the cathode ray in the X-ray tube must possess a minimum velocity if it is to excite an X ray of given quality. Different radiators were tried, and the critical velocity was found to be roughly pro- portional to the atomic weight of the radiator : in point of fact, the speed in cms. per sec. was 100 million (10 8 ) times the atomic weight. Beatty (see next page) has since shown that the same result is true if the metal, instead of being used as a secondary radiator, is employed as an anti- cathode, as in Kaye's arrangement. Thus, to recapitulate, if V K is the critical velocity of the cathode rays in cms. per sec., and A is the atomic weight of the anticathode, then in the case of the K series of radia- tions, the empirical relation V K =A. 10 8 is approximately satisfied for a range of elements from Al to Se. By combining this expression with Chapman's formula (p. 120), it follows that for the L series In Table IX., Whiddington's experimental values for the K radiations are given in heavy type in columns 3 and 5. The values for the other K radiations and the whole of the L radiations are calculated by the formulae above. It must be understood that many of these radiations have not yet been discovered (see p. 119). CHARACTERISTIC X RAYS 133 TABLE IX. MINIMUM SPEED OF CATHODE RAYS REQUIRED TO EXCITE CHARACTERISTIC RADIATIONS. Radiator. Atomic Weight (0-16). Critical Velocity of Cathode Rays to excite Requisite Potential to impart Critical Speed to Cathode Rays. 1 K radiation. L radiation; K radn. L radn. cm./sec. cm./sec. volts. volts. Hydrogen - 1-01 1-tf xlO 8 3 Carbon 12-0 1-2 xlO 9 410 Aluminium 27-1 206 1200 Chromium - 52-0 509 2-0 xlO 8 7320 11 Iron - 55-8 583 39 9600 43 Nickel 58-7 617 5-4 10,750 80 Copper 63-6 626 7-8 11,080 170 Zinc - 65-4 632 8-7 11,280 210 Selenium .*." 79-2 738 1-56 xlO 9 15,400 690 Rhodium - 102-9 03 x 10 10 2-7 29,900 2,100 Silver 107-9 08 3-0 33,000 2,500 Tin - 119-0 19 3-6 40,000 3,600 Tungsten - 1840 84 6-8 95,000 13,000 Platinum "-, 195-0 95 7-4 108,000 15,000 Lead - 207-1 2-07 8-0 120,000 18,000 Uranium - 238-5 2-38 9-5 160,000 26,000 Energy of an X Eay. By slightly modifying the arrangement, and putting the ionisation chamber in place of the secondary radiator, 2 Whiddington was able to correlate the energy of the X rays with the velocity of the parent cathode rays, and so to establish the truth of a relation deduced theoretically by Sir J. J. Thomson in 1907, that the energy of an X ray is proportional to the fourth power of the velocity of the exciting cathode ray. Beatty (P.R.S f 1913) has recently proved that this relation is only true for " independent " X rays : if characteristic rays are generated, the expression no longer holds (see p. 111). Beatty's Experiments. Beatty (P.R.S. 1912) has shown that the bulk of the characteristic rays generated in Kaye's experiments (p. 125). 1 See p. 100 for relation between cathode-ray speed and potential. 2 See Fig. 68. 134 X RAYS is due to a direct transformation of the cathode radiation into characteristic radiation ; and that only a small re- mainder owes its origin, 00 *"" l ' ' as one would perhaps infer, to the indirect action of primary X rays in emerging from beneath the surface of the anti- cathode. Beatty obtained cathode rays of uniform speed by means of the mag- netic-spectrum method, and was able to show that the direct and in- direct effects occur simul- taneously as soon as the speed of the cathode rays exceeds the critical value (see p. 132). Fig. 69 shows for the case of a copper antic athode, the relative amounts of characteristic copper radiation generated directly and indirectly by cathode rays of different velocities. Both effects disappear if the speed falls below 6*25 xlO 9 cm. /sec., a value which agrees very closely with Whiddington's critical speed for copper. To overcome the difficulties of measuring the X rays due to the vagaries of a coil discharge, an ingenious null method was devised which consisted in balancing the current in the ionisation chamber against* part of that carried by the cathode-ray discharge. Fig. 70 shows the connections. The interior of the anticathode tube A was lined with aluminium and joined to the anticathode. The greater part of the cathode ray current passed to earth through the variable resistance P. A smaller fraction passed through the high resistance Q to the ionisation chamber. The Speed of CaJ-hode m cms. per sec. FIG. 69. Showing relative amounts of charac- teristic X rays generated directly and indirectly from a copper anticathode by cathode rays of a variety of s; CHARACTERISTIC X RAYS 136 resistance P was altered until the latter current just neutralised the leak in the ionisation chamber. Of the current leaving A, P/(P+Q) goes to the chamber. Since Cathode R*ys of known Speed AnhcaJ-hode Aluminium Window ,leAd Screen Chamber To Elecfroscope FIG. 70. Beatty's apparatus for measuring the characteristic X rays generated directly and indirectly by cathode rays of various speeds (see Fig. 69). Q was very large of the order of 1C 12 ohms we may write this, P/Q. Thus the relative intensity of the X rays is evaluated by determining P in each case. ABSORPTION OF CHARACTERISTIC RADIATIONS. Barkla and Sadler's Relation for Normal Absorption. As Table XII. shows, a constant absorption-ratio exists for each absorber no matter what the hardness of the X ray outside a certain limited range. Some of the departures from proportionality with the harder types of rays, in what should be regions of normal absorption, are due to the fact that the coefficients for many of the elements given in Table XI. have not been corrected for scattering. If we make the proper scattering correction (p. 113), it is found that the ratio of the absorptions of a radiation in any two particular elements is approximately constant, and does not depend on the quality of the radiation, provided only that such radiation does not excite the characteristic radiations 136 X RAYS of either element. This important relation was first pointed out by Barkla and Sadler. To take an example, Barkla and Collier have shown that in the case of carbon, the absorption values relative to aluminium which rise in Table XII. from 0-11 for soft rays to 0-41 for hard rays become, when corrected for scattering, a steady value of 0-11 for all types of X ray. Abnormal Absorption of Certain Qualities of Radiation by a Particular Element. It is found that an element exhibits a maximum trans- parency for X rays of a quality identical with that of either / c -. c OaOt--O-<*iCOiOiOt>iO(M 'oO-rfiM CO (N I I> t> O5 l> CO O 00 C1 CO l> H^oOi-HrticOl I I ! i- i I I I I I I I I I I ^co^cl^ I I I I 5 I I M I I I M I I CO (N co t> COOCOi-Hi-HTHdiO i | 1O vj u.j w v* vj-/ w I ^M CO rt< TH ^* <* Th rh 1C GO CO l> CO 4< | GOGOCJSpO' it^-O5> (^5Cp iM ^^^Ju^^j^^^^^^^^i, I IjuA^l^LA^l l-H l-H i ( i t Ca5COCOCOO5Coco I 11 o M I I I I I I i I OiOSCOGOGOGOGOCO 1 I l| II II M M M I CO OS GO t> CO GO CO GO O5 O CO Tt< | I I (M (N (M I O O O O i-H^-l^(N.in V&rious Elements o 7 r A ^ !K Rd.non L R&di&fion ? 40 8O \2O 160 200 24 Aromic Weighf- of Absorbing Flemenr FIG. 72. Graph showing relation between the absorption of Ni (K ) radiation by various elements and the atomic weight of the absorbing element. The absorption passes through a minimum for a screen of Ni and also for one of atomic weight of about 164 (whose L radiation is identical with the Ni (K ) rays). Compare Fig. 56. absorption falls once more. In other words, the trans- parency reaches a maximum with a nickel screen and also with one whose L characteristic radiation is identical with the nickel K radiation. Thus the regular curve of increas- ing \jp with atomic weight of the absorber, is modified by the addition of sudden drops at as many regions as there are elements having one or other of their characteristic radiations identical with the X rays which are being absorbed. Similar curves are obtained for any other characteristic radiation : if the radiation is harder, all the maxima and 142 X RAYS mininia are displaced to the right, and, if softer, to the left. It will be remarked that Fig. 72 is the analogue of Benoist's curve (p. 106) for heterogeneous X rays. Evidently, Be- noist's rays were rich in components approximating to the characteristic radiation of silver, and this explains the absorp- tion " loop " which is prominent in his curve for soft X rays and less pronounced in that for the harder rays. Absorption of Characteristic Radiations in Gases. E. A. Owen (P.R.S. 1912) measured the absorption of a number of characteristic radiations in light gases. To get To E lee fro -scope Com p pri I Onis _Q. * CO COlOCOrHr I o J ^SSSSSIIS^ o O5 ft i (O * t^ CO r ! c d -* CO id Oi CO Id "a CO CO ""O CO CO CO 3 O TH ^H * GO O5 00 OO CO t^ GO GO GO GO GO 1 OS 00 CO Ss 6 6 | 6 1 I If * O (N CO CO (N CO 1 bp * i o 00 Tt< p-H w C.o Id CO 1 C o t-- GO r 1 * g-ti l> CO TH O5 00 00 s "3 " O o S ,"S ^i 3 1 || < O 5 , r as is independent of the .temperature. The guard- ring device will be noticed. The apparatus is here shown as arranged for liquid-air ' ture. Ionisation and Temperature. In 1909Crowther(P^..) showed that the ionisation tempera- produced by X rays was IONISATION BY X RAYS 155 independent of the temperature, provided the density was kept constant. In these experiments, Crowther used a range of temperature from about 180 to +184 C., and took especial care that the X rays did not strike the testing electrodes. His apparatus as arranged for liquid air tem- peratures is shown in Fig. 74. C. T. R. Wilson's Condensation Experiments. C. T. R. Wilson (P.R.S. 1912), in a series of remarkable experiments, has recently succeeded in rendering visible and photographing the tracks of the charged ions which are produced when a beam of X rays (or radium rays) passes through a gas. The method is based on the fact that supersaturated water vapour deposits on ions just as it does on dust particles and forms tiny drops. Thus the trail of a beam of X rays, itself invisible, becomes marked by a crowded line of cloud. Wilson has been able to take instantaneous photographs of these condensation nuclei in the positions which they occupied immediately after their liberation by the X rays. Fig. 75 shows the apparatus. The air within the shallow condensation chamber was kept completely saturated with moisture by means of water in the bottom of the vessel. Supersaturation was produced by suddenly increasing the volume of the chamber by exposing the under side of the movable bottom to a vacuum chamber. This was effected by a sharp pull (to the left) on the cord sliown (in Fig. 75), which opened the valve below the condensation apparatus. After the release of the valve, the cord pulled up with a jerk, the heavy weight attached to it was thus suddenly arrested, and the fine thread below it carrying a steel ball, snapped and the steel ball fell. In its descent, the ball passed in succession through two spark-gaps. The first passage caused a Leyden jar flash through the X-ray bulb ; the second similarly excited the illuminating spark. The arrangements were such that a horizontal beam of X rays crossed the centre of the chamber ; the illuminating spark flashed a pencil of light at right angles to the beam of X rays, and horizontally, or nearly so ; and the camera 156 X RAYS was usually mounted horizontally on the opposite side of the chamber to the illuminating spark. An electric field was maintained between the upper and lower faces of the expansion chamber. I llumi nA.rin ti Sp&rk c Expansion Chamber -Thread Sreel BaJI X Roy Spo>.rk Illuminating Fia. 75. Diagrammatic representation of C. T. R. Wilson's apparatus for photographing the track of a beam of X rays in moist air. The order of events in an experiment was, therefore, (1) expansion producing supersaturation, (2) X-ray discharge producing ionisation in the cloud chamber, (3) condensation of water on the ions, (4) passage of the spark for photo- graphing the cloud tracks. Wilson's instantaneous photographs (see Figs. 1 and 76) IONISATION BY X RAYS 15 1 ; show the tracks of corpuscles starting within the beam of X rays and extending for some distance beyond it. There is no indication of any activity on the part of the X rays other than the production of corpuscles : and the track of the X ray is not distinguishable otherwise than as being the region in which corpuscles have their origin. The cloud FIG. 76. Photograph obtained by C. T. R. Wilson of the path of a beam of X rays in air supersaturated with moisture (see p. 155). The beam of rays, about 2 mm. in diameter, traversed the air (from left to right of the picture) immediately after the expansion which produced the supersaturation. The axis of the camera was horizontal, and the magnification of the photograph is 6 diameters. trails show that the corpuscles start in all directions from within the path of the primary beam : they do not appear to exhibit preference for any particular direction. The result is striking confirmation of the view r which Prof. Bragg has advocated for some years that the X ray is completely inoffensive and innocuous during its life, and that only on its disappearance does the effective agent the corpuscle come to life. lonisation by X rays appears, therefore, to be entirely a secondary process. Fig. 2 shows a pencil of X rays passing obliquely through a copper plate. The transmitted beam, though much less 158 X RAYS dense than the initial beam, can be plainly seen. From the copper issue corpuscular rays in all directions ; these, which are responsible for the " halo " round the sheet, prove to be mostly of relatively long range. The char- acteristic copper radiation also excites corpuscular rays in the air, the majority of which have only a range of about 1 mm. at atmospheric pressure, and are to be found scattered throughout the vessel. This may be compared with the 1 cm. to 3 cm. tracks of corpuscles from the primary X-ray beam. If silver is used instead of copper, the secondary corpuscles have a much longer path. The clear space shown on both sides of the copper sheet in Fig. 2 is due merely to the heat of absorption of the X rays and the consequent formation of a region of air which is not saturated. Wilson attempted to display the crystal-reflection of X rays (see p, 211) by means of the above apparatus, but, for some reason, the reflected beam was ineffective in producing ions, and the plan did not succeed. VELOCITY OF X RAYS. In 1906, Marx in Germany published the results of an ingenious and elaborate investigation on the speed of the Rontgen rays. He excited an X-ray bulb by means of electric waves from an electrical-wire system ; these waves also charged to a varying potential an insulated plate on which the X rays fell. The secondary corpuscles emitted from this plate were collected by a Faraday cylinder con- nected to an electrometer : the amount was obviously con- trolled by the phase-relation between the potential of the plate and that of the cathode of the Rontgen-ray bulb. If the various distances and the connecting wire lengths were adjusted so that the charge received by the Faraday cylinder was (say) a maximum, then it was found that if the distance of the X-ray bulb from the insulated plate was increased by a certain amount, the wire along which the waves travelled to the plate had to be lengthened by the same amount to restore the maximum. Thus, according to Marx, the Rontgen rays travel with the same velocity as VELOCITY OF X RAYS 159 electric waves along wires, and, therefore, with the velocity of light, at any rate to within 5 per cent. Marx's experimental arrangements were subjected to severe criticism by Franck and Pohl, who, having repeated the experiments, doubted the validity of the method. In reply, Marx (A.d.P. 1910 et seq.) has since carried out a new series of experiments which, he claims, support his original result, but which nevertheless do not appear to have satisfied his critics (A.d.P. 1911). All this work was carried out before the nature of the X rays was known ; and there is now no reason for believing that X rays travel with a velocity other than that of light. CHAPTER XI. PRACTICAL APPLICATIONS OF X RAYS. RADIOGRAPHY. AN extended treatment of this most important branch of the subject can be found in medical works ; Chapter XII. deals with a number of important points in connection with equipment and technique. A radiograph is, of course, nothing but a shadow picture, and naturally care must be taken to place the subject symmetrically with regard to the bulb, so as to avoid unnecessary distortion of the image. For perfectly sharp images, the X rays should obviously proceed from a single point on the anticathode, but this, as has been remarked, is impracticable, and so it is usually beneficial to stop down the rays as much as is feasible. For this purpose, lead tube diaphragms are often employed, and can, in some medical cases, be made to serve a double purpose for example, the kidneys, which are in continual periodic motion, can be arrested temporarily, for radiographic purposes, by pressing down such a tube tightly into the abdomen. The greater the distance of the bulb from the fluorescent screen or photographic plate, the more correct the picture ; in practice the distance is usually from 12 to 24 inches. The spark-gap should not exceed 10 to 12 inches. With longer spark-gaps, rays too hard for radiographic purposes result. 1 (See p. 178.) 1 In the same way, radiographs obtained by radium y rays show only slight contrast between substances of different density. RADIOGRAPHY 161 FIG 77. Early radiograph by Campbell JO IQ . 78.- Instantaneous radiograph of thorax, winton, Jan. 18, 1896. Exposure 4 mms. Exposure about 1/100 sec. (Knox, 1915.) FlQ. 79. Radiograph of the hip- joint L Siemens Brot. 162 X RAYS Photographic exposures naturally vary enormously with the coil, break, and tube used. With a hammer break, a 10-inch spark, and a tube in average condition, some 5 to Fia. 80. Microradiograph (magnified 17 diameters) of legs of a tiny lizard (Seps Tridactylus). By Pierre Goby. 10 seconds is suitable for the hand ; 20 to 30 seconds for the ankle ; and a minute or more for the thicker parts of the body. The latitude in X-ray photographic exposure is large, though it is important to avoid under-exposure. The photographic plates are placed with the film towards the bulb, and most photographers agree that slow develop- RADIOGRAPHY 163 ment is useful for work such as this, where full detail is required. Examples of modern radiography are shown in Figs. 78 and 79 and 83 to 87. As examples of uses other than medical for the X rays, one may notice their former employment in the Ceylon pearl-fishing industry to locate pearls in oysters without opening the shells. The degree of transparency to X rays serves as a means of differentiation between paste and real diamonds : the heavy lead glass is much more opaque than the natural gem. The X rays have also been used to search for contraband of war, and to detect blow-holes and flaws in steel plates and copper ingots. By the use of extremely soft X rays, radiographs have recently been obtained of, for example, the soft tissues of the body (showing the veins and nerves), the wings of insects, the venation of leaves, and the structure of flowers. Radiomicrography of tiny objects forms one of the latest achievements of X-ray manipulation, an example of which by M. Goby (A.Rt.R. Dec. 1913) is shown in Fig. 80. Bismuth Radiography. The alimentary system may be radiographed by rendering the required part temporarily opaque through the adminis- tration of bismuth salts or emulsions with the food. This produces contrast in the photograph. Fig. 81 shows a good illustration of the method. Thorium oxide and barium sulphate are also used. A word of caution should be added, for the pronounced and very soft secondary rays that bismuth and other heavy metals emit, may actually be injurious. Stereoscopic Radiography. In this work, two distinct pictures are taken in turn by moving the X-ray tube, between the exposures, 2 or 3 inches parallel to the surface of the plate, the distance between tube and subject being about 20 inches. The resulting photographs are examined in a stereoscope. The method affords a means of ascertaining the depth of a foreign substance in the body, and is often of great assistance in 164 X RAYS FIG. 81. Bismuth radiograph of the intestines. The black circular spot near the centre of the picture is produced by a metal disc which is placed on the umbilicus as a " landmark." diagnosis. There are other types of localisers, some of which display much ingenuity of design ; they can be found fully described in the makers' catalogues, (see p. 191). RADIOGRAPHY 165 Instantaneous Radiography. It is a far cry from the prolonged exposures in the early days of X rays to the instantaneous work that is possible with modern apparatus. Nowadays, snapshots can be taken through any part of the body, and almost any of the moving organs can be radiographed. The worker who requires exposures short and frequent enough for, say, cinemato- graph films now experiences no difficulties out of the ordinary. If a single rapid photograph is all that is required, it is possible to secure it by comparatively simple means, and to send through an X-ray tube momentary currents of a magnitude undreamt of a few years ago. One method for obtaining practically instantaneous radiographs is to join the primary of a modern heavy current induction coil, or other high-tension transformer, straight to the direct-current town-lighting mains using special "explosive" fuses. When the current is switched on, the fuses are immediately blown, and the consequent interruption of the current produces a powerful discharge through the secondary winding and the Rontgen tube in circuit with it. For such rapid exposures a simple X-ray tube without cooling and regulating devices suffices. Dessauer in 1909, by using a type of explosive fuse for the break, was able to take single flash radiographs with exposures of the order of T -jhr sec. The momentary current through the tube was some 200 milliamperes or more, and the alternative discharge in air consisted of a broad band of flame 40 to 50 cms. long. Sir James Mackenzie Davidson has recently succeeded in radiographing a bullet leaving the muzzle of a revolver. The bullet in its flight over the surface of a photographic plate broke the primary circuit of a coil somewhat after the fashion employed by Mr. Boys some twenty years ago in his flying-bullet photography. The resulting flash through a suitably disposed X-ray tube in the secondary circuit gave a shadow photograph of the bullet. Perhaps even more remarkable are Dr. Worrall's recent experiments with a monster coil having a core weighing some 3 hundredweights. With a primary current of from 40 to 80 amperes at 240 volts, and the use of an explosion 166 X RAYS break, flash currents of the order of 1*4 amperes lasting for an interval of from 2~oir to Tinnr second were sent through an X-ray tube. The intensity of the discharge was such as to be capable of chiselling out a piece of metal from the anticathode and leaving a pit behind. Dr. Worrall has obtained very beautiful instantaneous radiographs by means of his apparatus. The possibilities of the extension of such experiments as these are far from being exhausted. A transformer which weighs about half a ton was referred to by Mr. Duddell in his Presidential Address to the Institution of Electrical Engineers (1912). Given a closer co-operation between the medical profession and the electrical engineer, mammoth apparatus and extraordinary results may be looked for in the future. Intensifying Screens.^ But by the aid of intensifying screens (a device which dates back to 1897), instantaneous radiography is possible with a much less formidable equipment. The recent im- provements in such screens have removed the defects of grain, etc., which formerly militated against their extensive employment. The Sunic screen, for example, is coated with a tungstate of calcium, which, fluorescing as it does with a very actinic bluish light, is capable of reducing an exposure twentylold. The screen is placed in close contact with the film of the plate and the X rays are sent through the screen before reaching the plate. Owing to the after-luminescence, which persists for some minutes, the screen should either be removed immediately after the exposure or not be disturbed for some little time. X-ray Photographic Plates. The large demand for photographic plates in radiography has brought about the introduction, by several firms, of plates specially coated for X-ray work. Dr. Kenneth Mees is responsible for a photographic plate 1 which presents some novel features. The plate is coated with an unusually thick 1 The Wratten and Wainwright X-ray plate. THERAPEUTIC USE OF X BAYS 167 emulsion containing a heavy metal along with the silver. The emulsion is thus rendered dense enough to arrest and record most of the incident rays, and the confusing secondary radiation from the glass backing is avoided. The result is a gain in definition and detail without any sacrifice in speed and contrast. Plastic Prints. On account of the pictorial beauty of the results, this method of printing deserves mention, From the original negative, a positive is printed on a lantern plate. The positive and negative, which should be equally dense, are mounted in accurate register, glass sides together. A print is then taken by means of light incident at an angle of about 45, and a picture thus obtained which shows pseudo- relief. Fig. 82 shows an example of plastic printing. PHYSIOLOGICAL APPLICATIONS OF X RAYS. X-ray " Burns." The dangers of indiscriminate exposure to X rays are now common knowledge, but some of the pioneers in X-ray work bought their experience at the price of their lives. Undue exposure results in severe dermatitis or skin disease, followed in chronic cases by large and cancerous ulceration, scaling and shedding of the nails. Unfortunately, the ex- tremely painful progress of the disease does not appear to be arrested by avoiding further exposure to the rays. Nor is there any known means of hastening recovery, though, according to Sir James Mackenzie Davidson, some relief and improvement has been obtained in superficial cases by the application of radium to the affected part in " doses " of some minutes at a time. Protective Devices. It is now known that X-ray " burns " are mainly due to the absorption by the skin of the very soft rays ; such rays are easily arrested by screening. The various pro- tective devices (gloves, spectacles, aprons, etc.), now always 168 X RAYS employed for the safety of workers, rely on the absorptive properties of lead or lead salts in some form or other. C. Thurstan Holland. FIG. 82. Plastic print of hand, showing fracture of heads of fourth and fifth metacarpal bones. Impregnated rubber is often used, and Droit (C.E. 1912) has recently succeeded in heavily loading silk tissue with phospho-stannate of lead (up to about 68 per cent.), and THERAPEUTIC USE OF X RAYS 169 so producing a material which, while extremely light and supple, affords adequate protection against the rays. The Rontgen bulb is fitted with a lead glass sheath, or, in some instances, the bulb itself is made of lead glass provided with a window of soda or lithia glass to allow the rays to get out. Fluorescent screens used for examina- tion work should be faced with lead glass (not less than 5 to 10 mm. thick) on the side remote from the bulb. In all ordinary circumstances the following thicknesses of protective screen may be considered adequate for the opera tor (see p. 46) : Lead 2 mm. Lead-impregnated rubber - 8 mm. Lead-glass 10 to 20 mm. Physiological and Curative Action of X Rays. 1 It might be anticipated that an agency possessing such vital characteristics would, under control, find a wide field of application in the treatment of disease. This has proved to be the case, and, as their technique is being improved, the X rays are finding a sphere of activity quite distinct from that of radiography. The method of " dosage " is usually that of the pastille (see p. 98) assisted by a tacheo- meter (or speed-counter) on the mercury break. In many skin diseases, the action of the rays has been turned to account and has proved to be of notable service. The effects are not, however, confined to the skin ; some of the internal organs, notably the spleen, are found to be even more susceptible. Happily the nervous system gene- rally is not at all sensitive to the rays. One of the most striking physiological effects of Rontgen rays is their action on the growing cells of the young ; the growth of young animals is greatly stunted by the rays ; the adult animal shows a greater capacity for resistance. Sweat glands and hair follicles are attacked and ultimately destroyed a pro- perty which provides a signal cure for ringworm, and, with prolonged exposure, is capable of producing total baldness. 1 The writer is indebted to a lecture by Sir James Mackenzie Davidson, at the Royal Institution in 1912, for much of this section. 170 X RAYS The white corpuscles of the blood are affected by X rays, but the red corpuscles are very resistant. The treatment has been largely and successfully employed for rodent ulcers, but experience has shown that it does not provide a cure for malignant tumours and large cancerous growths, though it may arrest their rapidity of growth : this is equally true of radium treatment, though in this case the outlook seems more hopeful. Apparent success has resulted from the em- ployment of X rays in cases of tuberculosis of bones and joints. Curiously enough, Rontgen rays seem to have little or no action on bacteria, and, in this respect, stand out in marked contrast to ultra-violet light, which is most destruc- tive to all forms of bacteria. X rays (and 7 rays) induce a sensation of- luminosity in the retina, so that the shape of interposed obstacles can be made out by a blind-folded normal eye or even by a cataract-affected eye. A totally colour-bHnd eye may be abnormally sensitive to X rays. Suitable Rays for Therapeutics. In the therapeutic use of X rays, the one essential is that the rays shall be sufficiently hard to reach and be absorbed by the diseased tissue. In treatment of the skin the very softest rays are the useful ones, but to do any good to more deep-seated parts harder rays are required. In this case the less penetrating rays should be removed to avoid their prejudicial action on the skin. An aluminium screen J mm. thick is generally sufficiently thick for the purpose. One difficulty in treating deep-seated tissue is that the harder rays are mainly scattered instead of being absorbed by the tissue. This small energy absorption means that the curative effects must be feeble. They could, of course, be enhanced artificially in some cases by bismuth treatment or the like. It may be added that the therapeutic effect of X rays often evidences itself pronouncedly in the proximity of bones ; this is probably due in part to the characteristic radiation emitted by the calcium of the bone. Similarly, zinc and other metallic ointments might be employed to augment the effect in superficial treatment. THERAPEUTIC USE OF X RAYS 171 The practice, which is often followed, of employing the same X-ray bulb for both curative and radiographic work is, of course, wasteful. Rays which are useful in thera- peutics are obviously unsuitable for radiography, as in the latter case the essential thing is that the rays should not be absorbed, but should reach the photographic plate. 1 As was remarked on p. 37, the iron, nickel, copper group of metals, when used as anticathodes, emit radiations very rich in soft rays, such as are suitable for curative work. There is no necessity for using a point source of X rays in therapeutic work, and, in fact, the anticathode can advantageously be put out of focus, or, if necessary, a plane cathode used. Glasses specially Transparent to Soft X Rays. A bulb intended for skin treatment should be either made of a glass specially transparent to X rays or provided with a window of such glass. Schott in 1899 was the first to make up a glass of this kind a silico-borate of soda and alumina as the result of experiments on the trans- parency of various oxides and carbonates to X rays. His list reads in order of diminishing transparency Li, B, Na, Mg, Al, Si, K, Cu, Mn, As, Ba, and Pb a sequence which is that of atomic weight. Schott's glass was never put on the market, as at that time the radiographic properties of the X rays were the only ones considered, and in this respect the glass possesses no appreciable advantage over soda glass. C. E. S. Phillips' conducting glass (P.R.S.E. 1906), which is a mixture of silicate of soda and borax with a little lead glass, is also very transparent to X rays. Its coefficient of expansion is unusually high, but by the use of intermediate glasses, windows of it could probably be fused into X-ray tubes. Lindemann (1911) has recently constructed focus bulbs provided with windows of a glass of lithium borate, which, of all the glasses ever made, is probably the most transparent to soft X rays. This glass 1 The same thing occurs in the use of radium. The highly penetrating y rays have little medical value ; it is the softer 7 and the a and /3 rays which are arrested by the body. 172 X RAYS is not very permanent, however, but Messrs. Cossor have recently brought out an improved lithium glass which can be worked and permits joints with platinum, so that X-ray bulbs can be constructed entirely of it (Fig. 46). Therapeutic Use of Characteristic Radiations. It has been suggested that the various characteristic radiations (p. 116) would find application and lead to greater precision and efficiency in curative X-ray work. These radiations are each of uniform quality, and it is, therefore, only a question of choosing a suitably hard radiation for the purpose in hand. But characteristic X rays, as ordi- narily generated, are so feeble that hours of exposure are required in place of the minutes necessary with primary X rays from a bulb. The writer showed, however, in 1908 (p. 125), that, with a soft bulb, a considerable proportion of the rays from an anticathode may consist of its characteristic radiation. By this means, an intense beam could be obtained from a tube provided with a suitable metal for anticathode and a window of thin glass or aluminium. It is further advantageous to use a thin filtering screen of the same metal as the anticathode. Better still, perhaps, in some respects, would be to make the window itself of. the metal whose radiation is desired and to use the window also as anticathode. Such a tube with a window soldered to the glass was, in fact, used by Owen (see p. 142). Therapeutic Use of Cathode Eays. If. as Prof. Bragg has long maintained, and, as is now generally believed, the X ray is in itself ineffective and owes all its activity physical and chemical to the electrons which it produces when arrested, then the only purpose the X ray serves in therapeutics is to plant the action deeper in the body. To produce therapeutic action at any particular point, there must first of all be transformation of the X rays into corpuscular rays, and then absorption of these corpuscular rays. If cathode rays themselves were simply discharged at the skin by means, say of a Lenard tube (p. 5), they could not penetrate more than about THERAPEUTIC USE OE X RAYS 173 aV mm., i.e. about the thickness of a cigarette paper. Possibly such a treatment might be valuable for some surface ailments, more especially as the radiation would certainly be accompanied by an abundance of very soft X rays from the aluminium window. Present-day Eadiology and Radiotherapy. The great improvements of recent years in X-ray tech- nique have given the diagnostic methods of physicians and surgeons a facility and exactitude never dreamt of at one time. For instance, in the surgery of the bone, we can not only see fractures, but we can examine the intimate lamellar structure of the bone ; we have, moreover, learned that tumours and cysts in bones are not nearly so rare as was once supposed. We can detect and determine the position of tumours in the cranial cavity, even in the deepest parts. The diagnosis of diseases of all parts of the alimentary canal is now routine stricture of the oesophagus, stomachic dis- orders and movements, changes in the duodenum, diseases of the appendix, growths in the colon, etc., can all be demon- strated by " opacity " radiology. Diseases of the liver, gall-bladder and pelvic organs ; aneurisms, mediastinal tumours and malignant growths, tubercle in the lungs, can be diagnosed with certainty and without pain or danger. In regard to radiotherapy, the X rays (and radium rays) possess extremely valuable properties in the treatment of malignant disease. The living cell has the power of resisting or responding to X rays, and with suitable dosage the malignant cells disappear, leaving the healthy tissue quite unaffected. In regard to dosage, it should be remembered that the problem is biological and it will rarely, if ever, happen that a treatment can be specified regardless of the length and intensity of each of the doses into which the treatment may be divided. Some surgeons have reported that X rays have seemed to give relief from pain. It has also been noticed that X rays have induced broken bones to knit faster and more surely (see p. 170). 174 X RAYS FIG. 83a. Bullet found (accidentally) under the {jrcat toe. There was no visible wound ; the patient thought that he had struck his foot when crossing a field. BHttf FIG. 83&. Fracture of shaft of humerus by bullet. Bullet lying between fragments. FIG. 83c. Fragmentation of ulna by shrapnel. FiGg. 83o, b, c. are war radiographs by Sir Alfred Pearce Gould (3rd London Military Hospital). FIG. 84. Shrapnel bullets and fragments of shell in brain. (N. S. Finzi. FIG. 85. Ballet at hilum of right lung. The bullet had been there a month when this radiograph was taken. There was neither inflammation nor symptoms. (N. S. Finzi.) 176 X RAYS FIG. 86. Cherry twig. The blossoms have lost their petals and the cherry pits have started to form. (Davey ; Coolidge tube, see p. 44.) I . .. Lii J FlO. 87. Egg, nearly hatched. (Davey; Coolidge tube, see p. 44.) CHAPTER XII. X-RAY EQUIPMENT AND TECHNIQUE. THE various types of apparatus employed in the production of X rays have been described in previous chapters, but a brief account of the installation and procedure for taking X-ray photographs may be of interest, more especially in view of the importance the subject has assumed in the present war. For a fuller treatment and for guidance in the interpretation of radiographs, the various medical treatises should be consulted, e.g. Knox's Text-book of Radiology. The necessary installation may be conveniently described under the following headings : (a) High-tension generator. (b) Control switchboard. (c) Measuring apparatus. (d) X-ray tubes. (e) High-tension circuits. (/) Tube-stands and couches. (g) Photographic apparatus. (a) High-tension Generator. The high-tension generator consists either of an induction coil and its accessories, or of one of the more modern high- tension interrupterless transformers dealt with on p. 63 et seq. The latter type of apparatus is of particular value for instantaneous radiography, though it is possible to turn out the highest class of work of all kinds with an induc- tion coil. In the present war, induction coils are being 178 X RAYS exclusively employed in the base hospitals and in field and motor travelling outfits. For most purposes a 12-inch coil will suffice, but a 16 or 20-inch coil is better if ample power is desired. Moreover, although a potential corresponding to a 20-inch spark-gap is never required in X-ray practice, such coils give a much greater current-output (and less inverse current) at the lower voltages actually employed, and thereby enable reasonably short exposures to be given. Portable sets are often provided with smaller coils, with a view to reducing weight a-nd expense, but their scope is necessarily limited. The coil chosen should give a "fat" flaming discharge well up to its maximum sparking distance, and not a thin crackling spark such as characterises a small coil. As is remarked on p. 54, the attention of coil-makers has been directed towards designs which yield a large current-output ; and very efficient apparatus is now obtainable. The condenser is joined in parallel with (or " across ") the interrupter. In the case of an electrolytic break, no condenser is required. As mentioned on p. 55, modern coils are frequently provided with a subdivided primary winding. Suitable connections are then made according to the nature of the break employed. Interrupters. The several varieties of interrupters are dealt with on pp. 66 to 70. The type most generally em- ployed is the mercury-jet break, working in an atmosphere of coal-gas. The provision of a gas supply is usually of rio great difficulty, as rubber gas bags may be used where no supply is laid on. In all but experimental installations it is much better for the gas supply to be brought close to the interrupter by permanently fixed gas-piping, and the final connection made by a short length of flexible metallic or stout rubber tubing. Rigid connection is undesirable, as many interrupters vibrate considerably, a tendency which may be checked by fixing the apparatus to a heavy base standing on a thick felt pad. Before starting the coil, gas is allowed to flow through the interrupter for a few minutes to drive out the air. The exit tap is then closed, the inlet being left open in connec- X-RAY EQUIPMENT AND TECHNIQUE 179 tion with the supply, as most mercury-gas breaks work better under a small excess pressure. Should the air not be entirely replaced by gas, the switching on of the cur- rent and the consequent sparking in the interrupter usually fires the mixture, and a small explosion results, to prevent any damage from which a simple safety valve is generally fitted to the apparatus. In mercury-breaks the size of the contact plates against which the jet strikes is of importance, and these should be ad- justed to suit the coil with which the break is to be used. To work an induction coil to the best advantage a regulating resistance is desirable in order to adjust the speed of the driving motor, and so control the rate of interruption. For instantaneous radiography a Wehnelt electrolytic break is frequently employed. A very convenient arrange- ment in a permanent installation is to provide a triple Wehnelt interrupter for instantaneous radiography as an alternative to a mercury-break for screening and general work. A throw-over switch is then mounted on the control switchboard, to permit a quick change from one break to the other. (b) Control Switchboard. The ne-xt important feature to be dealt with is the control switchboard, and it is largely on the design of this that convenience in the working of an installation depends. The first essentials are simplicity and accessibility, as adjustments often have to be carried out in a darkened FIG. SS. Mercury -gas interrupter. (See al Fig. 41.) 180 X RAYS room. These features may most readily be secured by arranging the essential switches, resistances and measuring instruments on a marble or slate panel carried at a convenient height on a light iron framework mounted on castors (see Fig. 36). This movable switchboard is connected to a fixed terminal board, installed near the coil, by cables which for protection may conveniently be grouped together in a length of flexible metallic tubing. The switch-table is brought close to the X-ray couch or stand, and the cur- rent can then be regulated without the operator having to move about in a darkened room among high-tension leads. The essential controlling apparatus mounted on the switch-table consists of a main switch, an ammeter of suit- able range, and an adjustable resistance for regulating the current through the primary of the coil. If it is desired to get a measure of the energy-input, a voltmeter will have to be added, in which case it is necessary to employ the shunt method of regulating the coil, as, otherwise, the voltmeter will merely indicate the voltage of the supply- circuit. In the shunt method of regulation (see Fig. 89) the supply circuit is connected to each end of a resistance which plays the part of a potential-divider. The coil is connected to any point on this resistance by means of a moving contact, so that any desired voltage between zero and that of the main supply may be applied to the coil. To avoid the use of an inconveniently large number of studs, finer adjust- ment of the regulation may be secured by the inclusion in the shunt circuit of a small variable resistance in series with the coil. In designing the main resistance it should be remembered that the coils at one end will at times have to carry heavy currents ; furthermore, the individual coils of the resistance should be graded so as to ensure that the voltage applied to the coil may be increased by approxi- mately equal steps as the contact is moved from one stud to the next. The alternative method to the shunt connection just described is to connect the coil directly to the mains with a suitable resistance in series. X-RAY EQUIPMENT AND TECHNIQUE 181 For carrying out screen-examinations a well-darkened room is necessary. The artificial light should be under control from the switch-table. Where much work of this nature has to be undertaken it is a great convenience to install a "foot-switch." By such means it is possible, by one movement of the foot, to switch off the room lights and simultaneously turn on the coil ; a second movement reverses the operation, cutting off the coil and again turning on the lights. I M&in Supply T /Two Pole Swikh Pilof- Series Resistance Condenser FIG. 89. Shunt method of regulating an induction coil. A pilot light is frequently fitted to the switch-table, this lamp -being connected across the outgoing terminals of the main switch (see Fig. 89) so that whenever this switch is in the " on " position, the lamp is illuminated, and thereby indicates that the high-tension circuit is "live." To avoid interference with screen examinations, such pilot lamps are usually of small candle-power, and are of ruby or dark- blue glass. A pilot light is of particular value when working with a Coolidge tube, as this tube does not fluoresce. The timing of an exposure is usually carried out by means of a stop-watch, a method which is quite the most satisfactory for ordinary work. For short exposures of only a few seconds or a fraction of a second, an automatic timing switch is a great convenience. 182 X RAYS (c) Measuring Instruments. Reference has been made in the preceding section to the provision of an ammeter and voltmeter in the primary circuit of the coil. These enable the energy put into the system to be determined. Of greater importance, how- ever, to the radiographer is the measurement of the energy in the secondary circuit, since it is the current through the tube which controls the amount or intensity of the X rays generated, and the potential difference across the terminals which regulates the hardness of the rays. The current through the tube (see p. 90) is almost in- variably measured by means of a milliammeter. These instruments are usually of the moving-coil type. If any inverse current is present, the milliammeter records the difference between the direct and inverse currents. A reliable milliammeter is undoubtedly a very convenient instrument for measuring exposures in radiography. Ex- perience will show an operator the number of " milliampere- seconds " necessary to obtain a satisfactory radiograph of any particular subject, and this will afford useful data for future work with the same equipment. The values cannot, however, be applied, except as a very rough guide, to other installations. As already remarked on p. 100, the ordinary method of specifying hardness by means of the equivalent spark-gap is not completely satisfactory, but its convenience, together with the wide latitude permissible in almost all X-ray work, leads to the well-nigh universal adoption of this method in radiography and radiotherapy. The measurement is made by means of a spark-gap connected in parallel with the X-ray tube. The terminals of the gap are usually a point and a plane (the latter con- sisting of a metal disc some 3 or 4 inches in diameter), or alternatively two points. The terminals are mounted on insulating pillars, which may be moved backwards and forwards in a slide, an attached pointer and scale enabling the length of the gap to be read directly. In taking a reading, the gap is reduced until a spark just passes while the tube is running. The length of gap is X-RAY EQUIPMENT AND TECHNIQUE 183 then noted, and reference to a table such as that on p. 101 enables the voltage on the tube to be determined. Greater accuracy can be obtained by allowing the spark to take place between balls of definite size. For the same voltage the spark gap between balls is much smaller than between points (as Table V. reveals), and this is one reason why FIG. 90. Qualimeter of Bauer type. (See p. 100.) the point and point, or point and plane are more generally used, the longer distance being easier to measure. Hardness comparisons should be made for specified currents through the tube, and it must be again pointed out that the equivalent spark-gap may not be a true measure of the hardness of the X rays, as, for a given voltage, the composition of the rays emitted may vary very consider- ably from tube to tube. If, however, this is borne in mind, a radiographer will find the measurement of spark-length 184 X RAYS of great practical value as soon as he becomes accustomed to the particular apparatus under his control. The alternative to measuring the voltage on the X-ray tube is to determine directly the hardness of the rays emitted. The several types of penetrometer used for this purpose are dealt with on pp. 107 to 109, and require no further description here. A penetrometer of the Bauer type is illustrated in Fig. 90. (d) X-ray Tubes. Reference to makers' catalogues will show numerous varieties of X-ray tubes. The main features of the various types have been considered in detail in Chapters IV. and VI. For an ordinary installation a battery of six or more heavy tubes will be found useful. By having a number of tubes at one's disposal, a suitably conditioned bulb may be chosen for the required purpose. Each tube possesses its own characteristics, and it is only by actual experience with the "individual tubes that the most satisfactory choice of tube may be made for a particular purpose. X-ray bulbs which work erratically may frequently be restored to good condition by being rested for a month or two. In this connection reference may be made to Chapters VI. and VII. All tubes are now fitted with vacuum regulators ; the type chosen is mainly a matter of personal predilection. In this country the "occlusion method" (p. 78) is very largely employed. Continued use of this regulator should be avoided as far as possible, as this tends to make the tube unsteady in its running. It is found that most tubes will run for long periods for some definite values of current and hardness without much regulation. It is for this reason that the method of keeping a number of bulbs in current use is advocated. (e) High-tension Circuits. The connecting up of the X-ray tube to the secondary circuit of the induction coil or transformer is a point worthy of some consideration. It is obvious that, in order to carry X-RAY EQUIPMENT AND TECHNIQUE 185 out an X-ray examination of any part of the human body the X-ray tube must be capable of being placed in a variety of positions. In all these positions it must be possible for the high-tension leads to be brought to the tube with- out coming within sparking distance either of each other or of any metal parts of the tube-stand, etc. The safety of both patient and operator must also be considered by keeping the wires taut and as remote as possible. The plan now very largely adopted is to install an over- head circuit, consisting of a pair of stout conductors tightly stretched from one side of the X-ray room to the other. These wires must be kept sufficiently far apart to prevent any possibility of sparking across. To reduce the brush dis- charge which always tends to take place, sharp points and edges to all parts of the high-tension conductors should be avoided. The wires are insulated at each end by ebonite or fibre rods from 10 to 12 inches long. Wire strainers may be fitted to enable the conductors to be kept taut. The overhead circuit is connected at one end of the room to the high-tension generator, the milliammeter and valve- tube being conveniently inserted in these connections. The supply to the X-ray tube is then brought down from the overhead circuit at any desired point by spring connec- tions hooked on to the wires. A convenient spring connector is of the self-winding type, resembling closely a small self- winding steel measuring tape. By such means slack con- nections are avoided, and the danger of accidental sparking is greatly reduced. The several kinds of valve-tubes and rectifiers are dealt with on pp. 70 and 71. Care must be taken that these are connected in the circuit in the correct polarity. A useful accessory to show whether inverse Current is present Or not FIG. 91. Oscilloscope for detecting inverse is a small vacuum tube known as an oscilloscope (Fig. 91). 'When this is connected in the high-tension circuit one of the electrodes (the cathode) is covered with a blue glow. If, however, inverse current 186 X RAYS is present, the other electrode will also show this appear- ance to a certain extent. The proportion of inverse current may be roughly judged by the length of electrode over which this glow extends (see p. 2). (/) Tube-stands and Couches. The examination of a patient by X rays may be carried out either by the use of a fluorescent screen or by the taking of photographs. In most cases it is usual to combine these methods, a preliminary screen examination being made to enable the photographic plate to be placed in the most advantageous position. In order that this examination may be carried out rapidly and conveniently, it is necessary that the X-ray tube should be supported in such a way that it may be readily adjustable while the patient occupies some convenient stationary position. The various forms of apparatus by which this is accomplished may be divided into two classes upright screening stands and horizontal couches. Certain of these stands are convertible from one type to the other, while other more complicated forms constitute combined X-ray examination couches and surgical operation tables. We can only deal here with the main features of each type. Upright Screening Stands. In this form of apparatus the patient under examination stands or sits in contact with a vertical partition, behind which the X-ray tube is supported. The tube is carried in a protective box on a framework capable of motion in both transverse and vertical directions. Ease in working is obtained by the provision of a counter weight, while either movement may be arrested at will by means of clamping devices. The partition is of thin wood or of tightly stretched canvas, so as to be practically transparent to the X rays. The fluorescent screen is supported in front of the patient. It may either be carried by an arm rigidly fixed to the tube-holder and capable of moving with it in both directions, or it may be carried on a separate support, the two motions being again provided for. The former plan, of tube and screen moving simultaneously, is more satisfactory. The X-KAY EQUIPMENT AKD TECHNIQUE 187 fluorescent screen must, in addition, be capable of move- ment to and from the partition, and a clamp should be provided to maintain it at the right place. This adjust- ment further enables a slight pressure to be applied to the patient while an examination is being made, and so helps FIG. 92. Upright scruening stand (Watson). to prevent movement of the part which is being photo- graphed. If, now, a photograph is desired (the adjustment of position having been made), the screen is removed from the support- ing frame and replaced by a holder containing the photo- graphic plate. A second method which may often be conveniently employed is to slip the plate contained in a 188 X RAYS black-paper envelope between the object to be photographed and the fluorescent screen without the latter being removed. The envelope is then held in position while the exposure is made. Fig. 92 shows a modern type of upright screening stand. Fio. 93. Mackenzie Davidson X-ray couch, showing tube-box in position. Protective screen removed. Horizontal X-ray Couch. In a horizontal couch the patient is examined in a recumbent position on a horizontal table. The apparatus is of more general application than are the upright screening stands just described, as the physical condition of the patient is frequently the deciding factor. The X-ray tube must be capable of motion in either direction X-RAY EQUIPMENT AND TECHNIQUE 189 in a horizontal plane, but its position may be either above or below the patient. For screen examinations the " below position " is naturally more convenient, while for photo- graphic work the over-couch position is usually preferred, and in some cases is the only practicable course. Some of the methods for the localisation of foreign bodies are more easily carried out in the " over " position, while the com- pression necessary to prevent movement of the subject in photographing, for example, the kidney region is more readily possible in this position. Fig. 93 illustrates a modern X-ray couch. Where facilities are limited, an X-ray couch may' be replaced by a plain table (or a stretcher on trestles) and a tube-stand. X-ray couches should be constructed, as far as possible, of wood. With the all-metal type there is liability of electric shock both to patient and operator through faulty insula- tion or inattention to the position of high-tension wires. Such a shock may not be actually dangerous, but is always unpleasant, and the use of wood will greatly reduce the risk of its occurrence. The employment of metal in the tube- support should also be as restricted as possible, and in any case it must not be in contact with or close to the X-ray bulb, otherwise sparking and perforation of the bulb may occur, particularly with a hard tube. Protection of Operator and Patient. The X-ray tube itself must in all cases be surrounded by some form of protective cover or box, which will prevent, as far as possible, the passage of X rays in directions other than that in which they are required. In this connection it may be remarked that all forms of X-ray stand or couch should be provided with an adjustable diaphragm, by means of which the cross- section of the beam of rays may be controlled. This may be either a circular iris-diaphragm or rectangular with a separate adjustment in either direction. In both types the movable leaves should be practically opaque to the rays in order to afford protection for the operator. An adjustable diaphragm is of great assistance in searching for foreign bodies as it enables one to improve the definition by stopping down the rays. 190 X RAYS For couches in which the tube is carried under the table, it is possible to fit a box which will completely enclose the tube. This box is covered with lead-impregnated rubber, or in some instances with sheet lead. In the latter case it is important that ample space should be provided to avoid the danger of puncturing the X-ray bulb by sparks passing from the tube to the sides of the box. Further, the bare metal should be covered by thick paper, felt or leather, etc., to absorb the very soft but dangerous secondary rays produced. In the case of field sets, or where considerations of size and weight are important, the use of a box of this nature is not possible. The protection in such cases consists of a shield either of lead- glass or built up of protective rubber. Such shields are fairly satisfactory, but cannot be considered completely so. It is important that the quality of the material used should be good, and tests of its absorption- coefficient for rays of various hardness should be made. The fluorescent screen should be covered with lead-glass of proved quality, as otherwise the radiation falling on the operator's face may have dangerous results if much screen- ing is carried out. In some installa- FIG. 94. -Lead -glass shield tions, arrangements are made for mounted on castors (Watson). . . , , viewing the screen from the side by reflection in a mirror ; the face is then entirely protected from direct radiation. The latest types of horizontal couch are provided with a shield, which supplements and moves with the tube-box. The shield increases the protection afforded to the operator, who has perforce to stand within a short distance of the tube. A convenient type of lead-glass screen which can be wheeled about the room is shown in Fig. 94. X-RAY EQUIPMENT AND TECHNIQUE 191 (g) Photographic Apparatus. The purely photographic apparatus required for the production of radiographs comprises very little beyond the ordinary equipment of a dark room. Provision must be made for carrying on operations with photographic plates up to 12 by 15 inches, and adequate drying arrange- ments should be fitted, as finished prints may be required within an hour or two of taking the negative. The plate is generally exposed in two black-paper envelopes, which protect it from the action of light, while if an intensi- fying screen (p. 166) is used, a special holder to take both screen and plate is necessary. The question of the suitable placing of the photographic plate is too large to be considered here. The reader is referred to medical books on radiography. The question of the time of exposure of the plate can only be settled by the operator himself after the various factors concerning the individual characteristics of the installation and the nature of the subject to be photographed have been taken into account. A warning may be given against storing unexposed X-ray plates in the neighbourhood of an X-ray installation, as continued use of the apparatus may load to slight fogging of the plates, quite sufficient to mask some of the finer details which are looked for in a good negative. If exi- gencies of space render it necessary to store the plates in or near the X-ray room, they should be protected by a lead-lined box. Localisation. The difficulty of the interpretation of radiographs arises from the fact that such photographs merely record the positions of shadows which do not necessarily correspond in shape and size with the image that would be produced in an ordinary camera. This will be clearer from a consideration of Fig. 95, in which suppose GHK represents a body resting on a fluores- cent screen or photographic plate (EF), and A represents the source of X rays. On examination of the screen or 192 X RAYS developed plate, the images of foreign bodies situated at B and C will be indicated at L and M respectively. It will be seen at a glance that the distance from LtoM is not a measure of the distance between the bodies at B and C, while, if attention is turned to a third body at D, it will be seen that its shadow coincides with that of B, and there is nothing to distinguish the one from the other. It is apparent that the position of the shadow merely in- dicates the line on which the foreign body is situated and not its actual position. Suppose now the X-ray tube is moved to a point A f , all else remaining unchanged, the shadows of B and C will move to L' and M 1 ', while a new one corresponding to the body D will be seen at N r ; it no longer coin- cides with the shadow of B. Thus the new position of the tube marks out a new set of lines on which the bodies are situated. This suffices to give the exact position of each of the bodies, namely the points of intersection of each pair of corresponding lines. The foregoing em- bodies the fundamental principle of practically all the existing methods of locali- sation, of which that due to Sir J. Mackenzie Davidson was one of the first introduced. In practice, the points A and A f are maintained at the same vertical distance from the plate (EF) ; the triangles ABA' and LBL' are then similar, and from simple geo- metrical considerations the height of the foreign body at B above the plate can be readily calculated if the distances AA, LL', and the height of A or A f above the plate are known. LL' is the distance through which the image of the foreign body moves when the tube is moved through L 1 N'M'L M FIG. 95. Localisation by X rays. X-RAY EQUIPMENT AND TECHNIQUE 193 the distance A A'. The various methods which have been suggested for carrying out the process of localisation are merely devices for solving geometrically or analytically the above simple exercise in trigonometry. In some cases the several distances are actually measured and the calcula- tion performed by the aid of tables or some mechanical device. For these methods a couch must be employed such that the tube may be readily adjusted to any desired position, and can then be moved through a measured distance without upsetting the adjustments. In other methods an auxiliary object of known dimensions is photographed simultaneously with the foreign body whose position is required, and the shift of the tube and its distance from the plate are determined from the two images of the auxiliary body. Full details of most of the present-day methods of localisation are given in the various issues of the Journal of the Rontgen Society for 1915 and 1916. Stereoscopic methods of localisation are referred to on p. 163. CHAPTER XIII. DIFFRACTION OF X RAYS BY CRYSTALS. Early Attempts to diffract X Rays. From time to time, a good deal of ingenuity has been exercised by various experimenters in testing whether there are, on the boundaries of the shadows cast by small obstacles, variations in the intensity of the X rays corresponding to optical diffraction fringes. Rontgen (1898) could not satisfy himself on the point. Haga and Wind (Wied. Ann. 1899- 1901) experimented with a V-shaped slit, a few thousandths of a millimetre broad at its widest point, and. obtained, in their photographs of the slit, broadenings of the narrow part of the image : if the effect were due to diffraction, the same amount of broadening with light would be associated with a wave-length of about 1/3 X 10 ~ 8 cm. It must be confessed that the result is in accordance with those recently obtained by crystal-reflection methods (p. 226), but Walter and Pohl (A.d.P. 1908), who repeated Haga and Wind's experiments, found that the width of the image of the slit was largely affected by secondary effects in the photographic plate depending on the amount of energy sent through the slit, with the result that different times of exposure gave rise to images of different widths. They concluded that the diffraction effect was not proven, and that their own experiments went to show that the wave-length of an X ray does not exceed 10 ~ 9 cm. Attempts to refract X Bays. Many attempts have also been made to refract X rays. Rontgen, for example, tried prisms of a variety of materials DIFFRACTION OF X RAYS 195 such as ebonite, aluminium, and water. He also attempted to concentrate the rays by lenses of glass and ebonite. Chapman (P.C.P.S. 1912) experimented with a prism of ethyl bromide vapour a substance which is strongly ionised by X rays. Two distinct experiments were conducted, in which the conditions might have been expected to favour a positive result. In one, the X rays were such as to stimu- late markedly the radiation characteristic of bromine (p. 138) ; in the other, the rays were of a type that was selec- tively absorbed by the vapour. Injneither case, however, could any trace of refractionJbe-discpvereJ. ~" Reflection Experiments. Many fruitless efforts have also been made to reflect X rays. We now know that the obstacle in the way of success to such experiments was the extreme shortness of the wave-length of the X rays. The specular reflection of ordinary light waves is rendered possible by the fact that the irregularities remaining in a polished surface are small compared with the wave-length of light. But irre- gularities negligible for light waves become all important with X rays, and a reflecting surface, such as mercury or plate glass, deals with X rays in much the same way as a surface covered with innumerable facets scatters light rays in all directions with no trace of regular reflection as a whole. Diffraction of X Rays. It was Prof. M. Laue of Munich who, believing that X rays were short light rays with wave-lengths of an atomic order of magnitude, 1 conceived in 1912 the notion that the regular grouping of the atoms in a crystal, which modern crystallography affirms, should be capable of producing inter- ference effects with the X rays, in a way analogous to that in which diffraction gratings deal with light waves. Laue's theory was at once put to the test and triumphantly justified by Friedrich and Knipping (A.d.P. 1913). It will be inter- 1 Planck's theory of radiation had led Wien in 1907 and Stark in 1908 to values of 0-7 x 10~ 8 and 0-6 x 10~ 8 cm. respectively -for the wave-length of an X ray. 196 X RAYS esting to trace in some detail the historical development of these crystal experiments, which have also been pursued from a different standpoint in this country (see later). Laue's Theory. Crystallographers have gradually developed the theory introduced by Bravais in 1850, which contemplates the atoms of a crystal as residing at the angular points of a "space-lattice." In a crystal, like atoms are regarded as forming a perfectly regular system of points in space, each and every kind of atom present in the crystal conforming to its own independent system. These different point-systems, of course, interpenetrate, the result being a parallel net-like arrangement of points, to which the term ' ' space-lattice ' ' is applied. Thus the crystal naturally divides itself up into a large number of precisely identical elements, in all of which the same relative posi- tions of the atoms are main- tained. This elementary vol- ume is, in a sense, the brick from which the crystal pattern is built up everywhere after the same plan. FIG. 96. Representation of the diffrac- tion of X rays by the atoms at the corners of an elementary cube of a cubic crystal. The several atoms thus repeat themselves at definite intervals, and Laue's notion was that the resulting regular . avenues of atoms should be capable of acting as a three- dimensional diffraction grating for rays of suitably short wave-lengths. Laue first considered the case of a simple cubic crystal, and assumed that the atoms were arranged at the corners of little elementary cubes this being the simplest cubic point-system possible. As the incident X rays pass through the crystal, they influence the atoms en route, and a secondary wavelet spreads from each atom as a wave passes over it. Let us take for convenience axes of reference parallel to DIFFRACTION OF X RAYS 197 the sides of a cube and an origin at the centre of one of the atoms, 0. (Fig. 96 shows the atoms in the xz and yz planes of the "lattice.) For simplicity, consider a beam of X rays to enter the cube in the direction of the z axis. Let us ascertain the conditions which will ensure that the wave- lets from all the various atoms in the lattice shall co-operate or "be in phase " in some particular direction OP, whose direction cosines are a, /3, and y. 1 It is sufficient for the purpose to take the cases of the nearest atoms to on the axis, viz. A, B, and (7, and express the conditions that the wavelets from these atoms shall be in phase with that from 0. These conditions are aa -- /?! . I, (1) where a is the distance between neighbouring atoms (i.e. one side of the cube), I is the wave-length of the X rays, and h v h 2 , and h 3 are integers representing the number of complete wave-lengths that the waves from A, B, and C respectively are ahead of the wave from 0. From (1) we obtain h l ho h% a and therefore a, /3, and (1 y) ought to be in a simple numerical ratio. From a consideration of the other cubes grouped round the z axis, it is apparent that there is a number of other points of maximum intensity situated precisely like P with reference to the z axis, so that if a photographic plate is placed to receive the transmitted X rays, there should appear, where the waves co-operate, a group of spots of fourfold symmetry. The Experiments of Friedrich and Knipping. Laue's theory was put to the test of experiment at Laue's request by Friedrich and Knipping (A.d.P. 1913). All that 1 That is, a, (3, and y are the cosines of the angles which OP makes with the axes of x, y, and z respectively. 198 X RAYS was required was to arrange that a parallel beam of X rays should, after traversing a crystal, be received on a photo- graphic plate, so that any directions showing " interference maxima " would be registered as spots. The apparatus used is shown in Fig. 97. The X rays emitted from the bulb were cut down by lead stops, so that a narrow pencil of rays fell on the crystal, behind which a photographic plate was placed a few cms. distant. The first crystal that was tried gave the result anticipated from the theory. The photographic plate showed an intense undeflected spot round which was grouped a FIG. 97. Friedrich and Knipping's apparatus for showing diffraction of X rays by transmitting them through a crystal. number of diffracted spots, some of which were deviated by as much as 40 from the original direction of the rays (see Figs. 98 and 100). If the crystal were moved parallel to itself, the grouping of the spots was unaffected. By altering the distance of the photographic plate from the crystal, the spots, while showing but little alteration in size, in- creased or diminished their displacement from the centre. Further, if the crystal was rotated so as to make a different angle with the primary beam, the pattern on the plate was affected : by careful adjustment, it was possible to obtain positions in which the spots grouped themselves quite symmetric alty round the centre spot. The results were generalised for a number of different DIFFRACTION OF X -RAYS 199 FlQ. 98. Pattern of Laue spots obtained by Friedrich and Knipping when X rays are diffracted by a zinc-blende crystal. The incident rays are parallel to a cubic axis of the crystal. OH FIG t>9 W. L. Bragg's construction to explain the position of the Laue spots shown in Fig. 98 (see p. 207). (From the Proceedings of the Cambridge philosophical Society.) 200 X RAYS crystals. It was found that exposures of some hours were necessary to obtain good results, since by far the greater proportion of the rays was unaffected and undeviated by the crystal. Shorter exposures, however, sufficed to reveal the more intense spots. Laue's Eesults for Zinc-blende. Figs. 98 and 100 are reproductions of the results obtained in the case of zinc-blende when the rays travel along two different axes of symmetry in the crystal. Knowing the coordinates of any spot on the photographic plate relative to rectangular axes having their origin at the point where the primary beam strikes the crystal, we can get at once the direction cosines, a, /3, and (1 7) of the ray which gives rise to that particular spot, and hence we can deduce the values of the parameters h l9 h 2 , and h s . Now, as re- marked above, since h lt h 2 , and h s are whole numbers, these values of a, /3, and (1 7) should be in a simple numeri- cal ratio. This was actually found to be the case in all the photographs. In no instance was it necessary to assume a number for h lt h 2 , or h 3 greater than 10 to give the values of a, ft, and (1 7) a whole number ratio. This in itself is strong confirmation of the theory that the spots are due to interference. Each spot has its own values of h lt h 2 , and h 3 . These have to conform to equations (1). The associated values of a, /3, and 7 have further to obey the relation and so it follows that there is only one value which I/a can have to satisfy all the equations for each spot. Thus every spot gives a different wave-length, since the values of h l3 h 2 , and h 3 are different for the different spots. It is here that an important distinction arises between a crystal grating and a line-grating. In a line-grating an interference maxi- mum is always possible, no matter what the \yave-length ; that is to say, the grating yields a continuous spectrum with incident white light. But in the case 'of a three- dimensional grating, certain wave-lengths only are eligible DIFFRACTION OF X RAYS 201 \ FIG. 100. Pattern of Laue spots obtained by Friedricb. and Knipping, when X rays are diffracted by a zinc-blende crystal. The incident rays are parallel to a trigonalaxis of the cubic crystal (l.e .diagonal-wise through centre of cube). FlQ. 101. W. L. Eragg's method of stereographic projection, applied to the case in Fig. 100 (see p. 208). 202 X RAYS to form interference maxima, so that a continuous spectrum is impossible. A similar effect may be imitated by mounting half -silvered parallel plates in front of an ordinary line- grating. If white light is now thrown on the grating, the former continuous spectrum will be replaced by a line spectrum representing a series of definite wave-lengths. The Laue photographs seem to show that while, in general, the larger the values of the integers h l9 h 2) and h 3 , the fainter are the* spots to which they correspond, yet, at the same time, the smallest integers do not represent the most intense spots as one would be led to infer by analogy with a diffrac- tion grating, for which the low-order spectra are generally the brightest. Not only that, but certain spots associated with simple values of h lt h 2 , and h s are absent altogether. But if the pattern were the most general possible, then all values of the integers, at any rate up to a certain limit, should be represented on the plate. A satisfactory theory must account for these anomalies, and Laue sought to explain them by assuming that the primary beam was made up of a limited number of inde- pendent homogeneous constituents, the absence on the plate of a spot with simple parameters being ascribed to the absence of the particular wave-length, which alone is capable of forming the spot in question. It was pointed out above that any fixed values of h lt h 2 , and h 3 gave a definite value for I /a, but it is evident that if we took the same multiples of all these values, say, nh^ nk 2 , and nh 3 , the equations (1) on p. 197 would still be satisfied, but now by a wave-length l/n instead of I. By adjusting the values of h l9 h 2 , and h 3 in this way, Laue was able to account for all the spots in the photographs by assuming the existence of only five different wave-lengths in the incident beam. The explanation is not, however, entirely satisfactory, because these five wave-lengths should give many other spots which do not appear in the photographs. The Laue Spots for a Zinc-blende Crystal. A zinc-blende crystal belongs to the cubic system, and crystallography distinguishes between three elementary DIFFRACTION OF X RAYS 203 point systems of cubic symmetry, namely those con- taining : (1) points at each corner of the elementary cube, (2) points at each corner and one at the centre of the cube, and (3) points at the corners and at the centres of the cube faces. Laue assumed that zinc-blende belongs to the first system, but in point of fact it almost certainly belongs to the third, as Pope and Barlow have shown from other considerations . W . L. Bragg (P.C.P.S. 1912) was led to examine the Laue spots of zinc-blende from this point of view. Adopting this view of the structure, Bragg supposed, as before, that axes are taken with origin at an atom (Fig. 102 shows the atoms in the xz and yz planes), and that when the various atoms are stimulated KTT fVo Y r-QTre /inr>irloTif alrmo- FlG.102. Representation of the diffrac- by the A rays (incident along tion of x rays by the atoms at the corners the z axis), emits a wavelet Ta cubfc TryS. f an elementary cube which in the direction OP ish 1 wave-lengths behind that from atom A on the x axis, and so on. The equations (1) on p. 197 ensure that all the corner atoms (including that at the origin) shall emit wavelets which are in phase along OP. It is necessary to obtain the corresponding conditions for the centre-face atoms (such as D and E), so that their wavelets also shall be in phase with those from the corner atoms. The difference in phase between the wavelets from D and will be ( -j wave-lengths, since D is situated in the middle of the face of the cube. This must be a whole number of wave-lengths to give an interference maximum along OP ; and it follows that Ti^ and h 3 must either be both 204 X RAYS odd or both even. The same must also hold for h 2 and h 3 . This at once explains why the complete series of values of h l9 h 2 , and h 3 for the Laue spots is not represented on the photograph. Consider first of all the set of spots in the appropriate Laue photograph of zinc -blende (Fig. 98), which have & 3 = unity. The corresponding wave-lengths prove to have every possible value greater than a limiting wave-length of I = 0'034a, where a is the length of the side of an elementary cube. The sets of values corresponding to wave-lengths approaching I O 06a are responsible for the two very intense spots in the inner square of the pattern ; all other wave-lengths smaller or greater than 0'06a give fainter spots until, for the limiting wave-length 0'034a. they are barely visible. Bragg accordingly concluded that the X rays utilised in this particular Laue pattern formed a con- tinuous spectrum, with a maximum intensity in the region of Z=O06a. Exactly similar results are obtained for the sets of numbers having Ji 3 = 2. There are two very intense spots which form the outer square, and, in addition, a few others con- siderably fainter. Similarly for h 3 = 3, in which series there are still fewer spots. In Table XVIII. is displayed a typical set of values of I/a for the different spots corresponding to h 3 = 1. The table is very interesting because of its completeness ; within a certain range of wave-lengths, every spot ' anti- cipated from theory is registered on the photographic plate. Thus Bragg's results afford strong support to the atomic grouping which Pope and Barlow claim for the zinc-blende space-lattice. In later work, Bragg has shown that the zinc-blende diffraction pattern is due almost entirely to the heavier zinc atoms. The sulphur atoms are situated on a similar parallel lattice, which may be reached by stepping along one quarter of the diagonal of the elementary cube of the zinc-lattice. 1 1 See Braggs' X Rays and Crystal Structure for a complete account. DIFFRACTION OF X RAYS 205 TABLE XVIII. Zinc-blende crystal ; incident X rays parallel to a cubic axis. Values of wave-length for ^ 3 =1. Value of h z . Values of I/ a, for h 3 = 1. !=! ^-3 fcjWfl \ = 1 7^ = 9 1 (off the 0-178 (m) 0-073 (v) 0-039 (v) 0-024 photograph) (invisible) 3 0-178 (m) 0-104 (v) 0-057 (v) 0-034. (/) 0-022 (invisible) 5 0-073 (v) 0-057 (v) 0-039 (m) 0-027 (invisible) 7 0-039 (/) 0-034 (/) 0-027 (invisible) 9 0-024 0-022 ' (invisible) (invisible) [The letters v, m, and / indicate the intensity of the spots " v " signify ing very intense, " m " moderately intense, and "/ " faint.] W. L. Bragg's Theory of the Laue Spots. Bragg was led to bring forward an alternative explanation of the Laue interference phenomena from the point of view of the parallel and equidistant planes of atoms which can be pictured in a crystal. Many systems of planes can, of course, be chosen, but we can confine the choice to the relatively few systems in which the planes are rich in atoms. Contrary to the view of Laue, Bragg (as mentioned above) supposed that the incident beam of X rays contained (like white light) every possible wave-length over a wide range, and thus formed a continuous spectrum of rays. Imagine then that such a beam falls on a crystal, and let us assume that when it strikes a system of parallel planes of atoms a small amount of energy is reflected by each plane. The wave front of the reflected beam from a particular plane is formed by the wavelets sent out by the individual atoms in the plane. If the distance between successive planes is d, and the glancing angle of the rays is 6, the train of waves reflected from the different planes in the system 206 X RAYS will follow each other at intervals of 2d sin ; and if the wave-length is such that this distance is equal to a whole number of wave-lengths, the waves will reinforce each other, and we shall get an interference maximum in that direction. Hence in this case, when the incident beam contains every possible wave-length, a particular system of planes in the crystal picks out, so to speak, the right wave-lengths, and the result of the simultaneous working of all the various systems of planes is to resolve the beam into its constituents. If the angle of incidence is altered, then different wave- lengths will in general be selected to form the interference maxima. On this view, the different intensities which the various spots exhibit might be due either to an unequal distribution of the energy in the spectrum of the incident X rays or to a difference in the closeness of packing of the atoms in the various reflecting planes. Bragg's method of regarding the interference is, of course, analytically equivalent to that of Laue. The reflection- method has the great advantage of being more readily pictured, especially in considering what happens when a crystal is rotated, in which event the pattern of spots is distorted exactly as it would be if the spots were reflections in plane mirrors. By changing the angle of incidence, we alter the phase difference (2d sin 0) between waves from successive planes ; and so a spot produced initially by a certain wave-length continues to represent without break a sequence of the wave-lengths present in the incident beam. If, as Laue imagined, certain wave-lengths only were pre- sent in the incident X rays, then as the crystal was slowly tilted spots would suddenly appear and disappear on the plate ; but, on the contrary, when the experiment is tried, the same spots can be traced continuously across the plate. It is also interesting to notice that some spots are very much changed in intensity as the crystal is tilted. One spot, for instance, which is barely visible in the symmetrical pattern, becomes, in another position of the crystal, the most intense of all, because its new wave-length now coincides with the maximum in the spectrum of the X rays. DIFFRACTION OF X RAYS 207 The elliptical shape of the Laue spots is a direct con- sequence of the fact that the incident pencil of X rays is not strictly parallel but slightly conical, and so the reflected pencil, which is obliquely received on the photographic plate, shows an elliptical section. Elliptical Loci of Laue Spots. With prolonged exposures, many more spots appear on the photographic plate than can be detected in Fig. 98. As Fig. 99 shows, the various spots group themselves natu- rally on ellipses of various sizes, all of which pass through the X Rays FIG. 103. Construction demonstrating elliptical loci of Laue spots. central spot. These ellipses, which are nearly circular, are sections of circular cones having the incident beam as a generator. The elliptical locus is a consequence of the fact that the various systems of parallel planes which can be selected in a crystal may have all manner of orientations : the atoms are grouped on parallel straight lines as well as on parallel planes, and each of these rows has a set of planes parallel to it. For example, suppose as before that a beam 208 X RAYS of X rays travels along the z axis, and consider a closely packed plane of atoms passing through the line OS (Fig. 103), S being a point in the xz plane. If this plane of atoms contains the y axis, then the reflected beam will pass along OP. But there is a family or " zone " of planes of atoms which can also be selected as passing through 08, and as we pass in rotation from one of these planes to another, the reflected beam OP will sweep out a circular cone with 08 as ' zone-axis." The trace on the photographic plate (which is at right angles to the z axis) will accordingly be an ellipse which passes through Oz and touches the yz plane. Similarly, if the plane is rotated about a zone-axis which is in the yz plane and passes through the origin, the ellipse passes through the z axis and touches the xz plane. Now suppose that there is a plane of atoms in the crystal which contains both these zone-axes, then the reflected beam from this plane will give a spot at the intersection of the two ellipses obtained as above. We can in this way, by drawing ellipses corresponding to rotations about various axes through the origin, locate almost all the Laue spots. This is done in Fig. 99, which graphically displays a key to the spots for zinc-blende when the incident rays traverse a cubic axis of the crystal. The ellipses are marked in each case with the co-ordinates of the atom nearest the origin through which the axis of rotation passes. The scales of co-ordinates are measured in terms of a unit equal to half the distance between con- secutive points along the axes. This unit is chosen because the only system competent to account for all the Laue spots in the case of zinc-blende, is that in which there are points both at the corners and face centres of the elementary cube (see p. 203). Stereographic Projection of Laue Spots. In representing a Laue pattern diagrammatically, it is tedious and inconvenient to draw the various elliptical loci. A much easier method is, however, possible without unduly distorting the pattern. DIFFRACTION OF X RAYS 209 Suppose the X-ray beam AO (Fig. 104) traverses the crystal at 0, the undeviated beam striking the photographic plate ZD at Z. Let OS be a " zone-axis "; the rays reflected in the family of planes which pass through this zone-axis lie on a circular cone, of which 08 is the axis and OZ and OP are two generators. This cone cuts the sphere of which OZ is a radius, in a circle of which ZB is a diameter. The projection of this circle on the plane ZD from the " pole " A is also a circle, of which ZP' is a diameter and 8 is the centre. FIQ. 104. Geometrical construction to explain stereographic projection of Laue spots. Thus, if we consider the Laue pattern which is formed on the surface of the sphere ZBA, and then project this pattern on the plane ZD from the pole A, we shall have a new projection in which the ellipse with ZP as major axis is replaced by the circle on ZP' . The distortion of the pattern of spots by the transformation is very small except in the regions remote from the centre ; and we now have the convenience of drawing circles instead of ellipses. It is easy to calculate the positions of the centres of the circles, such as S, from the dimensions of the pattern when the crystal is symmetrically placed. 210 X RAYS An application of this method of projection to the case of zinc -blende is shown in Fig. 101. It will be observed how closely the diagram follows the corresponding photo- graph obtained by Friedrich and Knipping (Fig. 100). Display of Laue Spots by Fluorescent Screen. Terada (Proc. Math.-Phys. Roc. Tokyo, Ap. 1913) found that by the use of a sufficiently transparent crystal and a not too narrow beam of X rays, 1 he could detect the Laue spots visually by means of a fluorescent screen. On rotating the crystal, the elliptical loci of spots referred to above are strikingly displayed. The fluorescent method is likely to be especially .useful for a rapid initial examination of a crystal. It is also of value for watching the progressive behaviour of a crystal which is being subjected to physical or chemical treatment. For instance, it was found that on heating a crystal of borax, the spots remained visible until the crystal was almost entirely melted. Interference by Metallic Crystals. Little work has been done so far on metallic crystals. Keene (P.M. Oct. 1913) found that if a beam of X rays was passed through freshly rolled metal sheets, a sym- metrical pattern was formed on a photographic plate placed behind the sheet. The axis of symmetry of the pattern was parallel to the direction in which the sheet had been rolled. A rotation of the sheet in its own plane produced a corresponding rotation of the spots in the pattern. If the sheet were heated and allowed to cool, the pattern was replaced by a number of radial streaks arranged in a circular band around the undeflected spot. A very old specimen of metal gave the same result. These radial streaks are undoubtedly due to reflection from small crystals Formed in the one case by age and in the other by annealing. The subject was also investigated by Owen and Blake (N. Feb. 19, 1914), who adopted a reflection method. A piece of copper was cut in two and one of the pieces was 1 5 to 10 mm. diameter. DIFFRACTION OF X RAYS 211 annealed, while the other was untreated. A beam of X rays was allowed to fall in turn on each of the samples, with the result that a large number of spots were obtained on a photographic plate in the case of the annealed specimen, while no effect was produced with the other specimen. The difference is, in the one case, to be accounted for by the presence of innumerable small crystals variously oriented, each of which reflects its quotum of the original beam, while in the unannealed specimen there is an absence of crystalline structure and regular atomic grouping. W. L. Bragg (P.M. Sept. 1914) examined native copper crystals with the X-ray spectrometer (see below) and found that the atoms are arranged in simple face-centred cubic- lattices. Vegard (P.M. Jan. 1916) obtained the same result for crystals of native silver. THE X-RAY SPECTROMETER. In the preceding portion of this chapter we have followed for convenience the development of the subject of X-ray diffraction from the historical point of view. It will have been remarked how completely and satisfactorily the various Laue phenomena are interpreted on W. L. Bragg's view of reflection from planes of atoms. Bragg was led, at the suggestion of C. T. R. Wilson, to ascertain whether X rays were regularly reflected from cleavage planes in crystals : such planes are known to be very rich in atoms. Mica at once suggested itself, and the experiment, when tried, proved immediately successful, only a few minutes' exposure being required to give a visible impression on a photographic plate. We have spoken of " reflection," but it is apparent that the crystal in such experiments is playing the part of a diffraction grating. Prof. W. H. Bragg and his son, W. L, Bragg, attacked the subject from this point of view, and devised the X-ray spectrometer in which the crystal is used as a reflection grating. Their work, which has received inter- national recognition, is described in their book on X Rays and Crystal Structure (Bell), to which the reader is referred. 212 X RAYS Experiments on similar lines were conducted at the same time by Moseley and Darwin at Manchester, and later by Moseley, who also obtained results of the first importance. FIG. 105. Photograph of Bragg's X-ray spectrometer. B is a lead box con- taining an X-ray bulb. /Si and S 2 are adjustable slits which direct a beam of X rays on to the face of the crystal C. The reflected beam passes through the slit S% into the ionisation chamber /, where it is recorded by the tilted electroscope in the metal box E. K is an earthing key ; M , a mirror for illuminating the electroscope. C and / can each be rotated about the axis of the spectrometer. X-ray Spectrometer. Bragg's X-ray spectrometer is illustrated and described in Fig. 105. As will be seen, the apparatus is similar to an optical spectrometer in arrangement, an ionisation chamber taking the place of a telescope. The strength of the ionisa- tion current measures the intensity of the reflected beam. Moseley. de Broglie, and others have used a photographic instead of an ionisation method of registering the rays. ' DIFFRACTION OF X RAYS 213 The crystal is mounted with wax on an adjustable mount- ing fixed on the table of the spectrometer. The X-ray bulb should be very " soft," arid a pencil of X rays is employed which leaves the anticathode at a grazing angle, a plan which diminishes the evil effect of the wandering of the cathode spot, and with it the variation in the angle of incidence of the X-ray beam. 1 The ionisation chamber is filled with a heavy gas or vapour (S0 2 or methyl bromide) so as to yield a large ionisation. A variation of the photographic method was used by de Broglie (C.R. 1913 et seq.) in which the crystal is caused to revolve slowly and uniformly by clockwork. In a second method, camera and crystal move together, the former at twice the rate of the latter. De Broglie and Lindemann (see Gorton, P.R. 1916) have employed a concave reflection grating consisting of a sheet of mica bent to a cylindrical shape. X-ray Spectra. From what has been said already of crystal structure, we can picture a series of planes of atoms parallel to each natural face of a crystal. When X rays fall on this face, they appear to be reflected from the face itself, although in point of fact it is within a thin layer inside the crystal that reflection is occurring, at depths usually not exceeding one millimetre. If a train of X rays all of the same wave-length (I), falls on the crystal, it is only when the glancing angle (9) has certain values that " reflection " takes place. These values are given by l=2d sin O lt 2l = 2d sin 2 , and so on, where d is the distance apart of the atomic planes. The reflection at 9^ gives the first order spectrum, that at $ 2 the second order, and so on. At other angles there is, in general, no reflected beam. The above equations give us a relation 1 A Coolidge tube in which the cathode stream is normal to the anti- oathode offers many advantages for X-ray spectrometry. 214 X RAYS between I and d ; and so, by employing the same crystal face, the wave-lengths of different monochromatic X rays can be compared ; or by using the same wave-length, the spacing d can be compared for different crystals or different faces of the same crystal. We thus have the means not only of analysing a beam of X rays, but of investigating the structure of crystals. The X-ray spectrometer (which is the exact analogue of the optical spectrometer) has given us a measure of both the atomic spacings of crystals and the absolute wave-lengths of monochromatic X rays. In the case of general (or white) X rays, however the crystal is oriented there is always some value of which fits in with the possible values of d and I. Every set of planes reflects somewhat, but the amount of reflection diminishes with the complexity of the plane. Thus at every angle of reflection, a certain amount of reflected radiation can be detected, the quantity increasing very greatly as grazing incidence is approached. But with monochromatic rays the effect is more restricted, and it is at only a few special angles (given by the equations above) that reflection can be detected. Thus with a mixture of monochromatic and general radiation (such as is ordinarily emitted by an X-ray bulb), if the strength of the reflected beam is plotted against the glancing angle, the X-ray spec- trum consists of a background or smooth curve of white radiation, superposed on which are " peaks " corresponding to the monochromatic spectrum " lines." The position and form of these peaks depend only on the metal of the anti- cathode of which they are characteristic. It is not yet determined whether the independent or white radiation referred to above consists of a mixture of different characteristic rays or whether it represents a perfectly continuous spectrum of rays. It would be anticipated that the spectrum lines would be present in the " secondary " X rays produced by the original " reflection " method of Barkla. On trying the experiment de Broglie (C.R. May 1914) had no difficulty in detecting the lines. DIFFRACTION OF X RAYS 215 Fig. 106 is the curve obtained by the Braggs for the rays from a platinum anticathode. The reflector in this instance was a rock-salt crystal, though the general form and relative proportions were found to be the same for all the crystals examined. The curve shows three prominent peaks (marked A, B, and C in the figure) thrice repeated. The rays corre- sponding to each of these peaks are found to be homogeneous when tested by the usual absorption method. Correspond- E OJ CD CD DC 60 E I 25 1 6f "0rder AnNc^fhode, Rock-SaJl- Reflecl-or 5 10 15 20 25 30 35 40 Glancing Angles of Incidence of X Rays FIG. 106. Showing intensity-distribution of spectrum of X rays from platinum. There are three main spectrum lines, and a large proportion of " white " or general X rays. ing peaks in the different series prove to be closely related : not only are the absorption coefficients of the rays producing, for instance, B 19 B 2 , and B 3 identical, but the sines of their reflection angles are in simple ratio. For example, the several angles of reflection of the B peaks are ll-55, 23-65, and 36-65. The sines of these angles are 0-200, 0-401, and 0'597, which are very nearly in the ratio of 1, 2, and 3. 216 X KAYS There can be little doubt as to the interpretation of these results. The peaks A, B, and C represent three different sets of homogeneous rays which appear as first, second, and third order spectra. The three groups of rays are not manu- factured in the crystal, for their properties prove to be the same, no matter what crystal is used. The incident X rays consist, in fact, of " independent " rays of all wave-lengths with an admixture of homogeneous radiations characteristic of the platinum anticathode (compare Kaye and Beatty's results on pp. 36 and 133). It is clear, moreover, that the characteristic rays consist not of a single homogeneous con- stituent, but of several groups of component rays of different wave-lengths. Moseley and Darwin (loc. cit.), using rather more refined apparatus, similarly detected five homogeneous constituents in the platinum radiation : peaks B and C are, in fact, close doublets. The proportions of these constituents appeared to depend on the state of the X-ray bulb. The allied metals, osmium and iridium, yield X-ray spectra similar to that of platinum. Each contains three main groups of homogeneous rays, together with a good proportion of general radiation. The spectra of palladium and rhodium are very similar to each other ; each is very nearly homogeneous, at any rate with a soft bulb (see p. 125), and contains little general radiation. On this account, both radiations have been employed a great deal by the Braggs in their later crystal experiments. Fig. 107 shows the rhodium spectrum, the -principal line of which is in point of fact a very close doublet. It needs to be pointed out that Figs. 106 and 107 are examples of X-ray spectra of which the general form depends on the circumstances. While it is true that the spectral lines themselves are invariable in position, their relative in- tensity and that of the general radiation are modified by such factors as the hardness of the X-ray bulb, the presence of any filtering screens, the type of discharge, and, of course, on the resolving power of the spectrometer. The chemical nature of the crystal also exerts a marked effect on the distribution of the energy. Bragg has shown that this is DIFFRACTION OF X KAYS 217 due to the selective absorption by the atoms of the crystals of the various components of the X rays. E <$ <1> CO 0) DC o -C "fab c > ? Ill for the three principal directions. We can thus deduce both the form of the elementary parallelepiped and the value of * | 3 or , where V is the volume of the parallelo- I I piped. Now, if p is the density of the crystal, the mass associated with each parallelepiped, and so presumably with each diffracting centre, is Vp. If M is the molecular weight of the substance, the number of molecules associated with each centre is , which we may write Z 3 ( ). If, in a M \l M / series of comparative experiments, I is kept constant, then the expression within the brackets is proportional to the number of molecules per centre, and can, moreover, be evaluated by experiment. Bragg proceeded to do this for a number of different crystals, each of which contained one heavy atom, viz. zinc-blende (ZnS), fluorspar (CaF 2 ), calcite (CaC0 3 ), iron pyrites (FeS 2 ), and rock-salt (Nad). Using in every case the homogeneous rays of the B peak of Pt, he found that the value of the quantity was, within a few per cent., the same for all substances. His results are put out in Table XIX. Thus the number of molecules associated with each diffract- ing centre is the same, and if we take into consideration the very different constitution of these crystals, this fact seems to point to the association of one molecule, and one alone, with each diffracting centre. By combining this result with the deductions on p. 220, it would seem that, 222 X KAYS since there is only one heavy atom in each molecule, the pattern obtained with the various crystals is due to a space- lattice formed by the association of only one heavy atom with each centre. It will be noticed that potassium chloride gives a value for Vp/l 3 M equal to half that for the other crystals, the explanation being that the two atoms, being of nearly the same weight, are equally effective as diffracting centres, and that a parallelepiped with half the side is now the crystal unit. The above argument is obviously not a complete proof of this important point, but the probability of the truth TABLE XIX. 1 Crystal. Lattice. Density P Mol. Wt. M Face. d I V P vp VM Sylvine, KC1 Simple cubic 1-97 74-5 (100) (111) 10-2 18-0 2-86 1-62 23-4 22-2 0-605 Rock-salt, NaCl Face -centred cubic 2-15 58-5 (100) (110) (111) 11-4 16-0 9-8 2,53 1-82 2-95 32-5 33-9 33-5 1-22 Zinc -blende, ZnS Face -cent red cubic 4-06 97-0 (110) 16-5 1-76 30-8 1-28 Fluorspar, CaF 2 Face -centred cubic 3-18 78-0 (100) (111) 11-7 10-3 2-46 2-79 29-8 28-3 1-18 Calcite, CaCO 3 Rhombo- hedral 2-71 100-0 (100) (111) 10-5 11-2 2-74 2-60 44-8 1-22 Iron pyrites, FeS 2 Face -centred cubic 5-03 120-0 (100) 12-1 2-39 27-3 1-15 1 In this table = glancing angle of Pt B peak, first order. 1 = wave-length of Pt B peak. d = distance between planes parallel to the face investigated. V = volume of elementary parallelepiped, calculated from this value of d and a knowledge of the nature of the lattice. DIFFRACTION OF X RAYS 223 of the assumption that each centre represents a single atom has been strengthened by each and every one of the many varieties of crystals subsequently examined. In later work, Bragg has been able to allocate the positions of both the light and heavy atoms in many types of crystals, and has studied the motions of the atoms with heat. (See Braggs' X Rays and Crystal Structure.) Dimensions of Space-lattice and Wave-length of X Eays. If the arrangement assigned to the alkaline salts is correct, we are now in a position to calculate the wave-length of the B peak, Pt radiation, for Take the case of rock-salt (Nad), Molecular weight, M =5S'5xl'64xlO~ 24 grammes. Density, - p = 2-15 gm./c.c. V/l 3 = 33:3 (experimentally determined). .'. Z 3 (33'3x2-15) = 58-5 x 1-64 xl(T 24 ; whence Z 3 = 1'34 x KT 24 and 1 = MOxlO- 8 cm., which gives us the wave-length of the homogeneous radiation of the B peak of platinum. From the values of djl given in Table XIX., we can calcu- late the lattice-constants for the several crystals. Platinum L Radiation. The mass-absorption coefficient in Al of the rays con- stituting the B peak of platinum was measured by Bragg and found to be 23*7. From Fig. 60 this value corresponds either to a K characteristic radiation from an element of atomic weight 72*5 or an L characteristic radiation from one of atomic weight 198. The atomic weight of platinum is 195 : the agreement is too close to be fortuitous, and there can be little doubt that the B peak is due to the L radiation. 224 X RAYS We have the means of deriving further relations. From Whiddington's rule for K radiations (p. 132) we can cal- culate that the cathode-ray energy necessary to excite the K radiation from an atom of weight 72-5 is about 2 xlO~ 8 ergs. This energy should be equal to the energy of the X ray excited, which, if Planck's radiation formula holds in this connection, is given by Jiv 1 or hV/L h is Planck's constant (6'55xlO~ 27 erg sec.), v is the frequency of the radiation, and V is the velocity of light. Now we have just shown that Pt L radiation has a wave-length of 1*10 x 10~ 8 cm., and therefore hV^ _ 6-55 x IP" 27 X 3 x 10 10 T~ I'lOxlO' 8 = 1'78 xlO~ 8 ergs which is in fair agreement with the calculated value. Moseley's Experiments. Moseley (P.M. Dec. 1913, April 1914) examined photo- graphically the X-ray spectra of a large number of elements and obtained remarkable and important results. The ele- ments were mounted as anticathodes (as in the apparatus on p. 36) in an X-ray tube, the X rays being analysed by means of a crystal of potassium ferrocyanide. The discharge tube was provided with an aluminium window (0-0022 cm. thick), which in those cases where the radiation was very soft was replaced by one of goldbeater's skin. In some instances, the whole spectrometer had to be enclosed in an evacuated box, since the rays were too soft to penetrate more than 1 or 2 cms. of air. In the majority of cases the pure elements were used as anticathodes, but if the elements were rare or volatile, oxides, salts, or alloys were employed. There was usually no difficulty in such cases in sorting out the spectrum lines. Fig. 109 illustrates the X-ray spectra obtained by Moseley for some of the lighter elements which emit strong K char- acteristic radiations. The spectra are placed in approxi- mate register in the photograph. It will be noticed that 1 hv is the energy of a quantum, according to Planck's theory. DIFFRACTION OF X RAYS 225 the wave-length increases as the atomic weight diminishes, and that the spectrum consists in each case of two lines, of which the longer wave-length (the a line) is the more intense. The L characteristic radiations were found to be made up of at least five lines, a, /8, j. 3, e, reckoned in order of decreasing wave-length and decreasing intensity. Increasing Wave Length FIG. 109. Moseley's photographs of the X-ray or high-frequency spectra of a number of metallic anticathodes. The spectra, which are in the third order, are placed approximately in register in the figure. The wave-lengths are given on p. 226. For each metal, the more intense line, with the longer wave-length, is the K characteristic radiation. The brass shows the Zu and Cu lines ; the cobalt contained both nickel and iron as impurities. Later work has shown that the principal line of the two- line K spectrum is really a close doublet ; and with both the K and L series there are a good many other fainter lines which are now being examined. r 226 X RAYS TABLE XX. WAVE-LENGTHS OF LINES IN X-BAY SPECTRA. The values below are clue chiefly to Moseley. The most intense line is called a, the next /3. There are other lines present in both the K and L series. For later measurements, see Siegbahn, Friman, and Stenstrom (p. 229). There is a slight want of agreement in some cases between the results of different experimenters. Element. K Series. Element. L Series. a a j3 x 10-8 cm. x 10-8 cm. x 10-8 cm. x lO- 8 cm. Al 8-364 7-912 Zr 6-091 Si 7-142 6-729 Nb 5-749 5-507 Cl 4-750 Mo 5-423 5-187 K 3-759 3-463 Ru 4-861 4-660 Ca 3-368 3-094 Rh 4-622 . Ti 2-758 2-524 Pd 4-385 4-1C8 V 2-519 2-297 Ag 4-170 Cr 2-301 2-093 Sn 3-619 Mn 2-111 1-818 Sb 3-458 3-245 Fe 946 1-765 La 2676 2-471 Co 798 1-629 Ce 2-567 2-360 Ni 662 1-506 Pr 2-471 2-265 Cu 549 1-402 Nd 2-382 2-175 Zn 445 1-306 Sa 2-208 2-008 Y 0-838 Eu 2-130 1-925 Zr 0-794 Gd 2-057 1-853 Nb 0-750 Ds 914 1-711 Mo 0721 Er 790 1-591 Ru 0-638 Ta 525 1-330 Pd 0-584 W 486 Ag 0-560 Os 397 1-201 Sn 0-50 0-43 Ir 354 1-155 Sb 0-48 0-41 Pt 1-316 1-121 W 0-203 0-177 Au 1-287 1-092 Bragg has obtained the following results : Element. *i 2 /3 y Pt xlO-8cm. 1-316 x 10-8cm. x 10 -8 cm. 1-12 x 10 -8 cm. 0-96 Ag Rh 0-562 0-619 0-557 0-614 0-495 0-545 0-488 0-534 Pd 0-589 0-583 0-516 0-503 DIFFRACTION OF X RAYS 227 The wave-lengths of the more intense lines of both the K and L series are given in Table XX. It should be under- stood the values rest on W. L. Bragg's estimate of the atomic distances in rock-salt (see p. 223). But the outstanding feature of Moseley's work is the relationship which he established between the X-ray spec- trum and the order of an element in the periodic table. In WAVE LENGTH X IO 8 Cms 49|o 48 Co 45Rh. 44 Ru 43 42 Mo 41 Nb. 40Zr 39Y 38 Sr _ 37 Rb- m 36 Kr- is 35 Br S 34 Se S 33 As D 32 Ge z se o as: 5 27 Co. 26 Fe 2 25 Mn b 24Cr < 23V 22 T, 21 Sc 20 Co I9K IgA . ^ s /\ XlX^ -/">^ 23 1 //* s^S xt/i ! xfX S/^ /y s^s ^s ^s J^f ^ \ s \ ^ (X I/* l>^^ ^X* j/'f l/^ 1 ./^ J/^iX^ 1 ^ff Y\' \/ s^ /""/i t/^X^ i^r/n Xx^ 17 Cl xK^ i i 15P Xx^ i i 51 L^ \ \ 14^ \ ' 1 1 6 tt 10 12 14 16 18 20 22 2^ SQUARE ROOT OF' FREQUENCY X 10 Pio. 110. Moaeley's relation between atomic numbers and frequency of K radia- tions (Braggs' A' Rays and Crystal Structure). brief, he plotted the atomic number (i.e. the number which represents the order of the element in the periodic table) against the square root of the frequency of the X ray, and found that the points for the different elements lay ex- tremely well on a smooth line which was almost straight. In other words, the wave-length of the X ray is inversely proportional to (2V a) 2 where N is the atomic number and a a constant. Fig. 110 shows the result for the K radiations, and Fig. Ill for the Eradiations. The several curves refer to corresponding lines of the various elements. 228 X RAYS It will be observed that no element is unprovided for in the scheme, and the harmony of the relationship is such as to justify the assertion that the spaces which have been necessarily left are awaiting elements as yet undiscovered. Perfect regularity over the region extending from hydrogen WAVE LENGTH X 10 Cms 865 4. Z IS 19 -8 7 i 1 1 1 .11 1 III 79 Au 78 Pt. 77 Ir 760s 75 74 W 73 To. 72 7iU 7OVb. 69 T m 68Et 67 Ho 660s 65 Tb 64Gd 63 Eo. 62 Sa 61 60 Nd. 59 Pr 58 Ce. 57 La 56 Ba 55 Cs 54 Xe 531 52 Te 51 Sb 50 Sn. 49 In 48 Cd 47 AQ 46 PC 45 Rh 44 Ru 43 42 Mo 41 Nb 40Zr 39 Y 385r - 1 j ' u I 2 Z / jf i ! / IT 1 / // / / i/ 1 i/i/ / // / 2Z / / / / // / J-f 7 / / 7/ / / j M/ / jrr 7 /J /I ZZ / /// t/ i/y / / "r/ / *$/ //r/ / / / / / / / / / / / / Jj / // / // / !/' // YY f f 6 8 10 . 12 14 16 18, 20 22 2' SQUARE ROOT OF FREQUENCY XIO~ 8 FIG. 111. Moseley's relation between atomic number and frequency of L radia- tions (Bfaggs' X Rays and Crystal Structure). to gold can only be retained by allowing spaces for three new elements, one between molybdenum and ruthenium, another between lutecium and tantalum, and a third be- tween tungsten and osmium. These elements should not be difficult to find. The order of atomic numbers is that of atomic weights, except in the cases of argon, cobalt and tellurium. If in Figs. 110 and 111 atomic weights are employed instead of atomic numbers, the relationship is seen to be not nearly as DIFFRACTION OF X RAYS 229 perfect. A table of atomic numbers is given on p. 258 ; it will be noticed that the atomic number is approximately half the atomic weight. It is apparent from Moseley's experiments that the atomic number is something more than a mere integer : it evidently represents some fundamental attribute of the atom. Now several quite different methods of experimental attack all agree in indicating that the atomic number agrees closely with the number of positive charges carried by the nucleus of the atom, or, alternatively, the number of electrons in the atom (see p. 18). We may well suppose that the wave- length of a characteristic radiation depends directly on the magnitude of the nuclear charge. In passing we may note the development of a remarkable view that the same atomic number may be borne by each of several substances which may have different atomic weights (and in the case of the radioactive substances, different stabilities), but which may be quite inseparable by ordinary chemical or physical tests. Soddy calls the members of such a group of elements bearing a single atomic number and occupying therefore a single place in the periodic table " isotopes." That which is common to them all is the positive nuclear charge adopting Rutherford's theory of atomic structure. From what has already been said, it would be expected that isotopes would yield the same X-ray spectrum. To take an example, RaB, ThB, RaD,and lead are all isotopes (in spite of the fact that their atomic weights range from 214 to 207), and, when put to the test by Rutherford and Andrade (P.M. 1914), the spectrum of the soft 7 rays from RaB proved to be identical with the L characteristic radiation of lead. Bohr (P.M. 1913 et seq.) has developed Rutherford's theory of the constitution of the atom with remarkable success ; and linked it up quantitatively with the quantum theory (p. 245) and Moseley's experimental results. Moseley's work has been amplified by Siegbahn with Friman (P.M. Ap. Jy. Nov. 1916) and with Stenstrom (P.Z. 1916). They have extended the K series down to Na, and the L series down to Zn (l a = 1'2 x 10~ 7 cm.). 230 X RAYS Relation between Wave-lengths and Absorption Coefficients. Table XXI. gives a series of comparative values of wave- lengths and absorption coefficients in aluminium, derived from the results of Rutherford, Bragg, Moseley, and Barkla. A scrutiny of these results shows that if A is the absorption coefficient and I is the wave-length, then, except at high frequencies, \ __ jjn where k is a const., and n lies between f and 3. But from the above, we have approximately where N is the atomic number, and therefore, taking we have which, if we replace atomic number by atomic weight, is Owen's fifth power law, referred to on p. 119. Rutherford (1916) has, however, shown that absorption results are not a reliable guide for deducing wave-lengths, at any rate for high frequencies, on account of the pre- dominance of scattering over true absorption, see Hull and Rice, P.E. 1916. TABLE XXI. WAVE-LENGTHS AND ABSORPTION COEFFICIENTS. Wave-length, I. A/P A] . Wave-length, I. *V Ix 10 -'> cm. 0-04 12xlO- 9 cm. 22 2 0-21 13 28 3 0-57 14 35 4 1-20 15 43. 5 2-10 16 51 6 3-3 17 61 7 4-8 18 72 8 6-6 19 83 9 8-9 20 95 10 12-2 21 108 11 16-5 22 120 DIFFRACTION OF X RAYS The Two Methods of Analysis. To recapitulate, there are then two distinct methods of crystal- analysis depending on X rays. (1) The Laue transmission method, which uses the inde- pendent, heterogeneous (or " white ") X rays that commonly constitute the greater part of the output from an ordinary bulb. The crystal plays a part somewhat like that of a " crossed " transmission grating, and the structure of the crystal controls the pattern of the diffracted spots. (2) The Bragg spectrometer method, which employs the homogeneous X radiations and uses the crystal as a re- flection grating. The structure of the crystal evinces itself in the distribution and intensity of the spectrum lines of the various orders. The Bragg method gives the data by which the dimensions of the lattices of crystals can be com- pared, and the X-ray spectrometer has already proved itself a powerful instrument for examining crystal-structure. The Laue method, on the other hand, can only supply information concerning the nature of the lattices, and that in a limited degree. Wave-lengths of y Rays. Rutherford and Andrade (P.M. May 1914 and August 1914) carried out an investigation on the diffraction of y rays. They used as the source of their y rays a thin- walled a-ray tube containing 100 millicuries of radium emanation. 1 The y rays are given off by the products of the emanation, viz. RaB and RaC. A pencil of rays was allowed to fall on a crystal of rock-salt, just as in the X-ray spectrometer, the reflected beam being examined photographically. About 30 lines were investigated, extending over a range of glancing angles of from 44' to 14. The values of the wave-lengths of the various y rays were found to lie between 1-365 xlO~ 8 cm. and 7-1 x 10~ 10 cm., the smaller values being much less than any X-ray wave- lengths hitherto recorded. The wave-lengths of the most 1 A millicurie is the amount of emanation in equilibrium with 1 mgm. of radium. 232 X RAYS penetrating rays of RaC are undoubtedly much smaller still; and will probably be very difficult to measure. Rutherford and Andrade also employed an ingenious trans- mission method of measuring the small angle of reflection (about 1J) of the y ray. A point source of radium emanation was placed centrally in front of a crystal of rock-salt, a photographic plate being set up behind the crystal (Fig. 112). The y rays find their own reflecting planes at A, A, in the crystal and meet the Photographic Plate C Ra Em 3 3 C FIG. 112. Reflection of 7 rays. photographic plate at B, B. The reflecting planes also cast absorption shadows at C, C, since the transmitted beam is weakened by the loss of the rays that are reflected. A similar reflection occurs at two planes parallel to the paper, and as a result the photographic plate shows a square pattern made up of both dark and light lines as in Fig. 113, from which the angle of reflection of the 7 rays can be readily and accurately obtained. Wide Range of Electromagnetic Waves. With the addition of X rays to the list of electromagnetic waves already known, the table of wave-lengths is extended DIFFRACTION OF X RAYS 233 greatly in one direction. At the other end of the scale are the waves which were originally discovered by Hertz and are now used in wireless telegraphy. The longest wave- length generated up to the present is about 15,000 metres, or a little over nine miles : the shortest is a few milli- metres. The waves ordinarily used in " wireless " are a few thousand metres long, e.g. the wave-length of the wire- less time signals from the Eiffel Tower is 2000 metres ; of the Navy signals, from 600 to 1800 metres ; of trans- Atlantic signals, 7000 metres or more. FIG. 113. Reflection of y rays. Next to Hertzian waves, in order of magnitude, come the infra-red or heat rays, the greatest wave-length yet observed being J mm. We pass from these right through the visible spectrum to the ultra-violet rays, which have been explored by Schumann and Lyman as far as wave- length 6xlO~ 6 cm. ; these are examined photographically. An extreme form of ultra-violet rays is probably represented by the " Entladungstrahlen " which are emitted from electric sparks, or the negative glow in a discharge tube (see Laird, P.R. 1911). Next come X rays with wave-lengths of the order of 10 ~ 8 cm., and beyond them the most penetrating of all the y rays whose wave-lengths have only recently been measured. 234 X RAYS The various wave-lengths, as at present known, are sum- marised in Table XXII. ; they cover, as will be seen, the amazing range of about one thousand million million fold. Virtually the only gap in the sequence is that between ultra-violet light and X rays ; and doubtless some of the very soft X rays obtained by various experimenters (see p. 123) find a place in this gap. TABLE XXII. WAVE-LENGTHS OF ELECTROMAGNETIC RADIATIONS. Kind of Wave. Hertzian waves Infra-red rays Visible light rays Ultra-violet rays X rays - 7 rays - Wave-length in cms. 10 6 to 0-4 0-031 to 7-7 x lO- 5 7-7 x 10~ 5 to 3-6 x 10- 5 3-6 x 10- 5 to 6 x 10~ 6 1-2 x 10~ 7 to 1-7 x 10- 9 1-4 x 10~ 8 to 7-0 x 10- 10 CHAPTER XIV. THE NATURE OF THE X RAYS. THE discovery of the X rays by Rontgen, and their im- mediate application in surgery, excited the popular interest to an astonishing degree. Geissler tubes, no matter what their suitability, were in immediate demand by a strangely interested public. The scientific journals of 1896 bear wit- ness to the many workers, who turned, if only for a time, from their usual pursuits, eager to test the extraordinary properties of the new rays. Naturally enough, among such an army of enthusiasts, speculation as to the nature of the rays was not marked by any great restraint ; a few of the responsible suggestions may be briefly recalled. Rontgen, Boltzmann, and others regarded the rays as longitudinal ether- vibrations of short period and great wave- length : Jaumajin added to this a transverse component : Goldhammer, Sagnac, and many others believed that the new rays were extremely short transverse ether-waves akin to ultra-violet light ; on the other hand, Re took the view that the wave-length, far from being short, was infinitely long : Sutherland considered X rays to be due to internal vibrations of the electrons within the atom : other workers held that the rays were a manifestation of the breaking up of molecules into atoms at the target : Michelson suggested that Rontgen rays were ether vortices : Stokes put forward a theory of irregular pulses in the ether ; and finally, many physicists, including at one time, Rontgen himself, and more recently Prof. Bragg, inclined to the view that the rays were flights of material particles which resembled 236 X RAYS strongly, and were possibly an extreme though electrically- neutral form of, the parent cathode rays. It is only within the last few years that controversy has been stilled by the discovery that X rays can be re- flected and diffracted by crystals. There can scarcely be any doubt now that X rays are identical with ultra-violet light of extremely short wave-lengths ; wave-lengths, in fact, of the order of the diameter of the atom. Yet it is not quite all plain sailing, for while it seems certain from the extreme precision observed in the reflection experiments that X rays are regular light waves and occur in trains of great length, yet the difficulty is that in many of their properties the rays behave strangely like streams of discrete entities, the effects of which are localised in space in much the same way as are the effects of rifle bullets. The difficulty is not, however, unique ; it is now known to be common to all forms of radiation. The Newtonian laws implying perfect continuity and infinite divisibility of time and space have, until recently, found complete cred- ence ; but in the very nature of things they do not seem to be reconcilable with modern experiment, which suggests that energy radiation is essentially discontinuous and must take place by finite " jumps." As to the mechanism by which this is accomplished, it is at present obscure and still a matter for speculation. 1 To meet the difficulty, J. J. Thomson, in his nucleated pulse theory, has suggested that all the various light radia- tions consist of concentrated and localised electromagnetic impulses which travel with the speed of light in some one direction through the ether (see p. 244). Planck's quantum theory, as developed by Einstein and Stark (P.Zi. 1909 and 1910) similarly argues that X radiation (in common with all radiation) is made up of definite and indivisible increments which can travel without loss or alteration of form, the energy of these " bundles " being proportional to the fre- quency of the radiation (p. 245). The same difficulty was felt by Bragg when he put forward his corpuscular or neutral- 1 For an excellent treatment of this subject see N. R. Campbell, Modern Electrical Theory, 2nd ed. 1913. Also Rep. Brit. Assoc. Sect. A, 1913. THE NATURE OF THE X RAYS 237 pair theory of the X ray. This theory, which regarded an X ray as a neutral corpuscle, was conspicuously successful in predicting and interpreting the energy transfers met with in the inter-relations of the cathode rays and X rays. On the other hand, almost all the well established results of the undulatory theory of light seem to be irreconcilable with entity views such as these. Nucleated light does not appear to conform to the marvellous explanation of inter- ference and diffraction, which Young and Fresnel founded on a theory of spreading waves, nor does it obviously lead to the general laws of reflection and refraction which are apparently obeyed by all waves, from the shortest X rays to the longest Hertzian. Points of Kesemblance between Light Eays and X Rays. The essential identity of X rays and light rays cannot be denied, in view of the work on crystal-reflection, but, notwithstanding, it will be useful and not without interest to summarise the points of resemblance which previous experiment had revealed between X rays, 7 rays, and light rays. In many cases it had already been found that the effects differed only in degree. For example, the ionising effect of ultra-violet light on gases (first established by Lenard in 1900, and more recently and completely by Hughes, P.C.P.S. 1911) is relatively feeble when contrasted with the more vigorous activity of X rays. Again, all three agencies cause the ejection of corpuscles from metals, and experiment has shown that : (1) The intensity of the incident rays does not affect the speed, but merely the number of the ejected corpuscles (p. 146). (2) The speed is controlled by the quality of the incident rays, 1 but not at all by the metal (p. 147). [With 1 In the case of ultra-violet light, Hughes (P.T. 1912) finds experimentally that the energy rather than the speed of the fastest electrons is proportional to the frequency of the light. This confirms a deduction from Planck's quantum theory (p. 245), which regards the photoelectric effect as due to a quantum handing over its energy to an electron. Hughes found the velocity of emission to vary from metal to metal. 238 X RAYS ultra-violet light the range of speeds is from about 10 7 to 10 8 cms. per sec. ; with X rays, 10 9 to 10 10 ; with 7 rays, 10 10 to 2- 99 x 10 10 .] (3) The secondary corpuscles tend to continue in the line of flight of the original rays (p. 145). (4) There is a selective emission of corpuscles for certain wave-lengths of the rays (pp. 123 and 148). These results are common to all three rays. There are further points of resemblance. As was noticed on p. 112, if an element is exposed to X rays, then, in general, two different classes of X rays are given out by the sub- stance. Of these, one is identical in nature with the incident rays, and is nothing more than so much scattered radiation ; the other is a radiation characteristic of the element, and does not depend at all on the nature of the exciting rays, provided only that they are harder than the characteristic radiation. This latter feature at once recalls Stokes' law of fluorescence. Apart from some exceptions, Stokes' law that the frequency of the exciting light is always greater than that of the fluorescent light holds generally for light rays. The analogy with X rays is complete. An even more striking similarity is presented if the dis- tribution of the two secondary X radiations is compared with the distribution of light in kindred circumstances. When light is allowed to fall on minute particles in sus- pension, as in a fog, it is found that the scattered light is not uniformly distributed, but varies in amount in different directions. The scattered light emitted parallel to the original beam is double that at right angles ; the intermediate intensities are proportional to (l-fcos 2 $), where 6 is the angle measured from the original beam. But if the particles emit fluorescent light as well as scattered, the fluorescent light is equally intense in all directions. In just the same way, it is found (p. 114) that the intensity of the scattered X rays obeys, at any rate approximately, a (1 +cos 2 0) law over a considerable angular range ; and that the " fore and aft " intensity is very roughly twice that at THE NATURE OF THE X RAYS 239 right angles. And, to complete the parallel, the character- istic X radiations show a uniform distribution just as fluorescent light does. Other points of resemblance between X rays and light rays have been noticed from time to time in the preceding pages. One point of difference is provided by the pheno- mena of absorption. In the case of light, it is known that many of the dark lines in the absorption spectrum of a body are in the same position as the bright lines in its emission spectrum : in other words, a body, under suitable con- ditions, is capable of absorbing strongly its characteristic light radiations. But, with X rays, on the contrary, we find that an element is especially transparent to its char- acteristic X radiations (see p. 136), and it is only for rather harder rays than these that the absorption becomes abnor- mally large. We may now consider the case for the restricted entity hypothesis. It will be convenient first of all to recall the main features of Stokes' famous theory of the X rays. The Ether-Pulse Theory of Stokes. Sir George Stokes promulgated the pulse theory of the X rays in the Wilde Lecture before the Manchester Literary and Philosophical Society, on July 2, 1897. He considered that " when the charged molecules x from the cathode strike the target, it is exceedingly probable that by virtue of their charge they produce some sort of disturbance in the ether. This non-periodic disturbance or ' pulse ' would spread in all directions, so that, on this view, the Rontgen emanation consists of a vast and irregular succession of isolated and independent pulses starting from the points and at the times at which the individual charged molecules impinge on the target. We know of no reason beforehand forbidding us to attribute an excessive thinness to the pulses " ; and to the narrowness of these pulses Stokes attributed some of the differences between ordinary light and X rays, which, 1 This was in the days when the cathode rays were thought to be mole- cules. 240 X RAYS apart from this, resemble each other closely 1 : both consist of electric and magnetic forces at right angles to each other and to the direction of propagation, but in the X rays there is not that regular periodic character occurring in trains of waves of uniform wave-length. Thus a Rontgen pulse on Stokes' theory is somewhat analogous to the crack of a whip when it is suddenly stopped, or the flash of flame when a projectile strikes a target. Briefly, the theory claims that the energy of an X ray is contained within a thin spherical shell which travels outwards with the speed of light in all directions, from the place where the speed of a cathode ray is suddenly changed. The faster the cathode ray and the more abrupt the * change in its speed, the thinner and more energetic the pulse. By the laws of electrodynamics, such pulses of intense electric and magnetic forces are inevitable when rapidly moving elec- trified particles are suddenly stopped or started. The Polarisation of X Eays. The polarisation of X rays (p. 114) follows as an immediate deduction from the pulse theory. The theory contemplates secondary radiation as owing its origin to the disturbances produced in the corpuscles when the primary X rays pass over them. While the X rays are thus accelerating the corpuscles, each gives out a pulse of electric and magnetic 1 It should here be mentioned that, as a result of the work of Rayleigh, Schuster, and others, our notions of the nature of white light have been modified in recent years, and it is now generally accepted that white light (like " independent " X rays) consists of irregular pulses which are capable of being transformed into trains of sine-waves by the various diffracting or refracting instruments. FIG. 114. Representation of spreading pulse, showing kink in line of force OPQR attached to the charged particle 0, the velocity of which has been suddenly altered. THE NATURE OF THE X RAYS 241 force the secondary Rontgen pulse. A single primary pulse may produce a great number of secondary pulses with properties which depend on the grouping, etc., of the cor- puscles. Thus, on this point of view, there is, to use Sir J. J. Thomson's apt comparison, much the same difference between the primary and secondary rays as there is between the sharp crack of lightning and the reverberations of thunder. The argument in the polarisation experiments is that since in an X-ray tube the cathode rays are all travelling in the same direction, then in the resulting pulses the electric forces (which are at right angles to the direction of motion of the cathode rays) will lie in planes passing through that direction, and not at right angles to it (see Fig. 58). In other words, the particular pencil of X rays which is employed will be concentrated in the plane which contains both X rays and cathode rays. Hence the motion of the excited corpuscles in the radiator will also be mainly in this plane, and so the intensity of the secondary radiation will be a minimum in this plane, and a maximum at right angles to it, a result which agrees with that actually found. The fact that the X rays are only partially polarised, may be ascribed to the fact that the cathode rays in the X-ray tube are not stopped in a single collision, but describe many directions before finally coming to rest. Modification of Spreading-pulse Theory necessary. , But experiment has clearly established that the theory of the spreading pulse needs extensive modification. It will be profitable to review the trend of the results (to some of which we have already referred), that have led to the theory in its modified form. Categorically these are : (1) When X rays encounter a gas, only an exceedingly small fraction less than one in a billion of its molecules become ionised * (p. 151). 1 The same difficulty occurs in understanding /why, when ultra-violet light falls on metals which show photoelectric properties, such a very small proportion of the particles are liberated. 242 X RAYS (2) The extent of this ionisation is unaffected by tempera- ture (p. 154). (3) When X rays encounter a metal, the corpuscles ejected have a velocity which (a) does not depend on the intensity of the X rays, and so is independent of the distance of the metal from the X-ray bulb (p. 146), (b) increases continuously with the hardness of the X rays (p. 146), (c) does not depend on the nature of the metal (p. 147), (d) is equal, or nearly so, to the velocity of the cathode rays in the X-ray bulb (p. 148). (4) These ejected corpuscles are not evenly distributed, but tend to pursue in the main the original direction of the X rays. The effect is most pronounced with metals of small atomic weight and hard X rays (p. 145). In considering the first result, we may recall that according to the ether-pulse theory in its original form, all the molecules of a gas are equally exposed to the X rays, and we are led to infer that those few which become ionised must have been in a state very far removed from the average. Their ab- normal condition cannot be attributed to an exceptionally high kinetic energy, for the kinetic theory of gases would then require that the ionisation should vary rapidly with the temperature and we are immediately confronted with result (2). We might, however, claim that the ionisation is controlled by the internal conditions of the different atoms, rather than by their kinetic energy. The phenomena of radioactivity lead us to believe that atoms possess large stores of internal energy which are not readily unlocked by outside agencies ; and if it should be the case that the possession of an excep- tional amount of internal energy means weakened stability, then it might easily happen that only abnormal atoms would be ionised by X rays. Or, again, it might be that an atom is capable of collecting energy from many X rays until it has enough for one electron. On either view, the THE NATURE OF THE X BAYS 243 ejection of corpuscles from a metal subjected to X rays is interpreted as the outward sign of a sort of radioactive explosion of some of the atoms rendered temporarily un- stable. The X ray thus acts merely as a trigger to start the explosion ; the corpuscles come from the atom, and owe their energy to it alone. That their velocity should be independent of the intensity of the X rays follows at once, and is in accord with result (3a). But we have now to explain why the speed of the corpuscles is not independent of the quality of the X rays (result (4)). Why should the velocity be greater when the X rays are hardened, if their only effect is that of a trigger action ? and further, why should the path of a corpuscle be in- fluenced by the direction of the X ray, if the latter merely precipitates the disintegration of the atom ? On the explosion theory, neither result could be anticipated ; nor should we be unreasonable in expecting that the disintegra- tion of different metals would lead to very different velocities of the ejected corpuscles. The reverse is the case. We are, in fact, left with no alternative but to suppose that the energy of the corpuscle is derived from that of the X ray. Now, result (3a) remarks that the energy of the corpuscle is independent of the distance of the X-ray tube. But, for reasons similar to the above, the X ray must derive its energy directly from the parent cathode ray, and, according to the pulse theory, it distributes this energy over an ever- enlarging surface. The argument is fatal to the spreading- pulse theory. The energy of the X ray must, it is evident, be confined within very narrow bounds which do not widen as the X ray travels. 1 Combining this result with (3d) above, we are led to conclude that the X ray is a minute entity whose energy is not frittered away along its track, but is handed over completely to one corpuscle and no more on suitable encounter with an atom. This is a statement of the case for the entity hypothesis, and the difficulty remains of reconciling it with the ordinary electromagnetic theory of Maxwell. Of the attempts which 1 Sommerfeld (1911) maintains that the pulse theory is competent to explain part, at any rate, of this localisation of energy. 244 X RAYS have been made, we may refer briefly to Sir J. J. Thomson's nucleated pulse theory and Planck's quantum theory. The Nucleated or Localised Pulse Theory of J. J. Thomson. Sir J. J. Thomson's theory of the X ray assumes a fibrous structure in the ether, and pictures a corpuscle as the seat of a tube of force which stretches out into space. When a cathode ray has its velocity altered, the radiated energy runs along this tube, as a kink runs along a stretched wire. The energy is confined to the region of the kink, and it is not given up until it strikes a corpuscle, to which it can then transfer its energy without waste. The nucleated pulse is equivalent to Stokes' pulse, with the exception that instead of spreading out uniformly in all directions, it is confined to one direction only. Professor Thomson further believes that the energy of light is distributed in analogous fashion ; that individual light waves are not continuous, but correspond to a collection of wires along which the various disturbances travel ; and that if a wave-front could be made visible we should get, not con- tinuous illumination, but a series of bright specks on a dark ground. The energy is not, therefore, uniformly dis- tributed throughout the whole volume of the waves, but is concentrated in " bundles." The rays diminish in intensity with increasing distance owing to the greater separation of the " batches," and not to the enfeeble ment of individual units. The distribution of energy thus resembles that contemplated by the New- tonian emission theory of light, according to which the energy was located on moving particles sparsely dissemi- nated throughout space. In the case of X rays the phenomena are sharply defined, but with light rays they are much more involved. The dis- continuous wave-front theory, in fact, regards X and -y rays as light in its ultimate simplicity. This agrees with experi- ment : the general laws covering the behaviour of X rays are obeyed with fewer exceptions than is the case with light. THE NATURE OF THE X RAYS 245 Planck's Quantum Hypothesis. The quantum theory, which has been applied to many branches of physics, originated in an attempt to account for the spectrum of black-body radiation. The development of the theory and the evidence for its physical basis are excellently set out by Jeans in a Report on Radiation and the Quantum Theory (Physical Society, London), to which the writer is much indebted. Briefly the theory assumes that all matter contains large numbers of vibrators which can emit ether waves of different frequencies, but only spasmodically and in such a way that the quantity of energy emitted is an exact multiple of a certain unit or " quantum " (e). The amount of this quantum of energy increases with the frequency of vibration (v), or the energy, e = hi>, where h ( = 6-5xlO~ 27 erg. sec.) is a constant of nature, now known as Planck's constant. On Planck's theory the energy of an X ray is proportional to its frequency ; and the generation of secondary corpus- cular rays and photoelectrons is regarded as due to a quantum handing over its energy to an electron. In the same way the maximum frequency v of the independent radiation generated in an X-ray bulb can be calculated from the relation Ve = hv, where V is the voltage and e is the electronic charge. (See p. 128.) This important result has been verified experimentally by Hull and Rice (P.R. 1916, for voltages as high as 150,000. The relation does not appear to hold for characteristic radiations, the various lines of which appear simultaneously at a critical voltage. (See Webster, P.R. 1916.) Thus the quantum theory contemplates a certain dis- creteness or atomicity of an entity which is measured by h or some function of h. It may be noted that the physical dimensions of h are those of angular momentum, an identity which is of great importance in Rutherford's theory of the nucleated atom as developed by Bohr. It is probable that the atomicity of h is associated with the atomicity of the electronic charge, e, and in such event, a physical explanation of the quantum theory may be based 246 X RAYS on the atomicity and possible discrete existence of Sir J. J. Thomson's tubes of force or corpuscles of radiation referred to above. Arising out of this is the degree of substantiality of the ether, whether, as some of the modern school of British physicists contend, the ether is the primary real substance of the universe, or whether as the extreme rela- tivity school hold, the ether has no reality at all. Whatever role is assigned, however, to the ether, the main difficulty about the quantum theory is, as already men- tioned, that of reconciling it with the well-established results of the undulatory theory of light. " Fluctuation " Experiments with 7 Bays. Experimenters have naturally sought to establish by direct means the presumed discrete nature of light rays and X rays. As is well known, the spinthariscope of Sir Wm. Crookes exhibits for the a rays of radium fluctuations both in time and space. Similarly, the effects predicable for j3 rays have been observed ; and since 1910 a number of workers, among them von Schweidler (1910), E. Meyer (1910), Laby and Bur- bidge (1912), and Burbidge (1913), have endeavoured to detect corresponding fluctuations in the ionisation produced by 7 rays in a gas. For this kind of work, a steady source of rays is absolutely essential, and so 7 rays have been worked at rather than X rays. Laby and Burbidge (P.R.S. 1912) used two ionisation chambers, identical in all respects, and disposed them sym- metrically about and equidistant from the radium emitting the 7 rays. If the 7 radiation has a spherical wave surface, then the ionisations in the two chambers will have a con- stant ratio. If, on the other hand, the 7 rays are circum- scribed entities, emitted in random directions, as a rays are, then the number entering each chamber in a given time will fluctuate. There is one outstanding difficulty : if Prof. Bragg's view as to the indirect process of ionisation by 7 rays is correct (p. 152), the fluctuations might be produced by a variation in the number of /3 rays generated by each 7 ray. The fluctuations which Laby and Burbidge actually observed THE NATURE OF THE X BAYS 247 in their experiments cannot, therefore, be interpreted with certainty. More recently E. Meyer (A.d.P. March 1912), using some- what similar apparatus, finally concluded that a single y ray can produce ionisation in more than one direction and on more than one occasion : the numbers of y rays emitted by the same source. in two different directions do not appear to be independent. Meyer's experimental arrangements have been criticised by Burbidge (PM.S. 1913). Meyer's results are, however, in accord with those of Rutherford (P.M. Oct. 1912), who has recently found reasons for supposing that a swift /3 ray may give rise to several y rays in escaping from an atom, and still retain part of its original energy. We may here refer to the work of J. J. Thomson (P.P.S. Dec. 1914) and Chadwick (P.M. 1912), who have shown that positive and a rays are able to excite X and y rays when they fall on matter. This would suggest that it is kinetic energy rather than velocity which is the determining factor. " Fluctuation " Experiments with Light Rays. N. R. Campbell (P.C.P.S. 1909, 1910) attacked the problem of light emission by the " fluctuation " method, with the object of discriminating between the ordinary and entity light hypotheses. Unfortunately the difficulty of finding a source of light which is very intense and also extremely constant proved insurmountable. Taylor (P.C.P.S. 1909) approached the problem from a different standpoint. All ordinary optical phenomena are average effects, and are therefore incapable of differentiating between the usual electromagnetic theory of light and a restricted entity type. If, however, the intensity of light in a diffraction pattern were so greatly reduced that only a few of the indivisible bundles of energy could occur at once on a zone, the ordinary phenomena of diffraction would be modified or disappear altogether. Taylor's method of attack was to photograph the shadow of a needle under various illuminations, and with exposures chosen such that the total energy supplied was constant. Exposures ranging from a few seconds to three months were employed, but no 248 X RAYS variation in the sharpness of the diffraction pattern could be detected in the different photographs. Thus the more direct attempts to confirm the " discon- tinuous " nature of light and of X rays have not met with success. The Outstanding Problem of Radiation. It will be apparent that the problem of the nature of the X ray cannot yet be dismissed. We have succeeded in estab- lishing the essential identity of X rays and light rays, and the interest has, accordingly, shifted its ground. The diffi- culties, conspicuous with the X rays, have merged into those which all classes of electromagnetic waves are found to present in greater or less degree, and the full secret of the nature of X rays will doubtless be revealed when we find the key to the overshadowing problem of the mechanism of radiation in general. We have seen that the problem of the transference of energy by ether-waves involves us in the conception of a " quantum " of energy-radiation travelling undissipated through space. The reconciliation of the idea with the older and well-founded conception of spreading waves remains. The experimental evidence seems to indi- cate that both theories are true simultaneously : that radiant energy is both concentrated and indivisible, and at the same time spreads and is divisible. The keynote of the old mechanics is, in fact, continuity ; of the new mechanics, discontinuity. Any hope of a compromise between the two theories appears to involve concessions fatal to either. We are left confronted with the riddle of modern physics. APPENDIX I. IN connection with Rontgen's discovery, Sir James Mac- kenzie Davidson has been kind enough to write down for me his recollections of an interview with Prof. Rontgen not very long after the discovery of the X rays. " While travelling on the Continent in 1896 I made a pilgrimage to Wiirzburg, and called at Professor Rontgen's house in the evening, and was kindly granted an appoint- ment for the following morning. I presented myself about 11 a.m., and was shown into a laboratory which contained a coil and a small cylindrical-shaped X-ray tube. Pro- fessor Rontgen, a tall man with dark bushy hair, a long beard, and very kindly and expressive eyes, received me cordially. He could not speak much English ; I was still worse at German. However, by means of English and some Latin we made ourselves intelligible to one another. He excited the tube and showed me various shadows on a fluorescent screen. On each of the terminals of his coil he had a small aluminium ball, 1 cm. in diameter, which he told me he used as an alternative spark-gap to test the hardness of the tube. He incidentally remarked that he found a tube had its maximum photographic effect when it was working just at 2*5 cms. alternative spark a fact which I have always found to be correct. I asked some blunt questions as follows : Q. " 'What were you doing with the Hittorf tube when you made the discovery of the X rays ? ' A. " ' I was looking for invisible rays/ Q. "'What made you use a barium platino-cyanide screen ? ' 250 X RAYS A. " ' In Germany we use it to reveal the invisible rays of the spectrum, and I thought it a suitable substance to use to detect any invisible rays a tube might give off.' " He then detailed how he made the discovery. He said he had covered up the Hittorf tube with black paper so as to exclude all light, and had the screen (which was simply a piece of cardboard with some crystals of barium platino- cyanide deposited on it) lying on a table 3 or 4 metres from where the covered tube was situated, ready to be used. He excited the tube to ascertain if all light was excluded. This was so, but to his intense surprise he found the distant screen shining brightly ! " I asked him, ' What did you think ? ' He said very simply, ' I did not think, I investigated.' " Incidentally, he told me how he had taken a photograph through a pine door which separated two of his labora- tories. On developing the negative, he found a white band across it, which, he ascertained, corresponded to the beading on one of the door panels. He stripped the beading off, and found the band of shadow was due not to the increased thickness of wood but to the ' plumbum ' (white lead really) the doormaker had employed in attaching the strip of wood. " He seemed amused at my remonstrating with him about keeping the ' screen ' lying about in his laboratory. I told him it was a ' historical screen/ and should be pre- served in a glass case ! I hope he has carried out this suggestion. For the sudden shining of that * screen ' un- doubtedly led to one of the greatest discoveries in modern times." J. M. D. APPENDIX II. THE PRODUCTION OF HIGH VACUA. ..'. )i"3 A BRIEF notice may be taken of the present methods of exhausting vacuum tubes. The very highest vacua can be got by making use of the extraordinary absorptive powers for gases of charcoal (e.g. cocoanut char* coal) when immersed in liquid air-^a remarkably quick and effective method we owe to Prof. Dewar. Oxygen, nitro- gen, water vapour, etc., are absorbed to large extents, hydrogen rather less so, helium and neon least of all. It is essential that the charcoal should be previously heated in situ 1 and the emitted gas pumped off before applying the liquid air. Angerer (A.d.P. 1911) records a pressure of 8xlO~ 7 mm. Hg by the use of charcoal and liquid air. With one exception, all the various mechanical pumps for the production of vacua employ the plan of repeatedly abstracting and isolating, by means of a solid or liquid " piston," a certain from the vessel to be exhausted, and The exception is a strikingly recently introduced by Gaede. 2 FIG. 1 15. Tube contain- ing charcoal immersed in liquid air. fraction of gas delivering it elsewhere, ingenious " molecular pump " 1 As a practical precaution, a plug of glass-wool should be inserted above the charcoal, to stop the small carbon particles, which are expelled when the charcoal is heated, from passing over into the apparatus. 1 See Engineering, Sept. 20, 1913. 252 X RAYS It depends for its success on the viscous dragging of gas by the surface of a cylinder rotating within a second cylinder at a speed comparable with the velocities of the molecules of the gas, which are accordingly impoverished in one direction and accumulated in the other. The pump is extremely rapid in action, but requires the initial pressure to be reduced to a few mms. of mercury by an auxiliary pump. There is no piston, but always free communication, through the molecular pump, between the vessel to be ex- hausted and the auxiliary pump. With a speed of rotation of 12,000 revs, per min., and an initial vacuum of ^V mm., Gaede records the remarkably low pressure of 2 x 10 ~ 7 mm. It is very interesting to note the susceptibility of the pump to the molecular velocity of the gas present. For the same velocity of the cylinder, a lower pressure is attainable with air (molecular velocity 500 metres/sec.) than with hydrogen (molecular velocity 1800 metres/sec.), as may be shown by scavenging the vacuum with one or the other gas. The pump shares with the cooled-charcoal method the advantage of not requiring any drying agent vapours are sucked away as readily as gases. For such a rapid type of pump the connecting tube must not be restricted in bore ; the remark, indeed, applies to all pumps. Still more recently (1914), Gaede has put on the market a new hand-driven piston pump, which can produce very high vacua with great rapidity, and is capable of dealing with water- vapour. Next come the various types of mercury pumps the Topler and Sprengel in a variety of modifications, some of them designed .to be automatic in action. The rotary mercury pumps, such as the Gaede, have come into extensive use, and possess the great advantage that they can be motor- driven a feature commending itself to all who have worked with the hand-manipulated Topler. In regard to mercury (and oil) pumps it is well to remember that they will not pump vapours, 1 and that the vapour pressure of mercury 1 To obtain high vacua it is, therefore, necessary to remove water vapour by means of a drying agent such as phosphorus pentoxide. Other vapours can be frozen out by means of liquid air. APPENDIX II. 253 at ordinary temperatures is about TTHF^ mm. mercury a fact which does not always tally with the claims sometimes advanced on behalf of this or that pump. 1 Mercury pumps ought not to be set the task of exhausting from a high initial pressure they work best as finis hing-off pumps. For the earlier stages of exhausting there is available a variety of oil-pumps which can be motor-driven, arid some of which can deal with large quantities of gas. Ordinary heavy engine-oil works well in these pumps and has a low vapour pressure. A drying chamber should be used in con- junction with an oil-pump, or the oil may emulsify and the efficiency of the pump will suffer. Gas held by Walls of Tube. A great deal of gas mostly hydrogen and moisture is held by the electrodes and the walls of a vacuum tube. To liberate the gas, the discharge should be run for some time to suit the conditions under which the tube is intended to be used. 2 This, of course, ought to be done by the maker before the tube is sealed off from the pump. The walls of the tube hold this surface gas tenaciously it appears to be largely moisture which is held bound as a condensed surface layer. To get rid of it, the tube has to be heated to between 300 and 400 C., at which stage there is a great evolution of gas. If this is pumped off while the tube is hot, the vacuum will be found to improve greatly when the tube is cooled, and will not deteriorate with time so much as it otherwise would. " Finishing-off " Processes. There are one or two " finishing-off " processes to follow a pump, which are well known to research workers. Cocoa- nut charcoal, when used as anode, or the liquid alloy of potassium and sodium, when used as cathode, absorb 1 It is, however, possible for a pump to exhaust somewhat lower than the vapour pressure of the liquid used. A really good water injector (filter) pump will exhaust to about 7 mms. of mercury, whereas the vapour pressure of water at atmospheric temperatures is some 12 to 15 mms. 2 At higher pressures, more current can be passed through the tube, and the electrodes can be made hotter than at very low pressures. 254 X RAYS ordinary gases, if the discharge is not too heavy. Yellow phosphorus is converted to red by bombardment with cathode rays ; the change is accompanied by a diminution in pressure, due partly to the lower vapour pressure of the red allotrope. and partly to the fact that under the discharge the red phosphorus combines with any oxygen, nitrogen or hydrogen present, forming compounds with very small vapour pres- sures. This latter method is used in exhausting the Lodge vacuum valve (p. 71) : the presence of the phosphorus compounds is further useful in regulating the vacuum during the subsequent use of the valve. Merton (Chem. Soc. Journ. 1914) finds that " precipitated copper " heated in situ to about 250 C. absorbs gases and vapours with great readiness on cooling. In the earlier stages of exhausting a bulb, much time can generally be saved by omitting the usual constriction (for sealing-off purposes) in the connecting tube to the pump. When the exhaustion has reached a satisfactory stage, care- fully dried air is admitted, the constriction put in and the re-exhaustion proceeded with. Table XXIII. gives a notion of the capabilities of various pumps. TABLE XXIII. Pump. Gaede molecular - Gaede rotary (mercury) Improved Topler (mercury) Gaede piston Geryk(oil) - Sprengel (mercury) Injector (water) - Attainable Vacuum. mm. Hg. 0-000,000,2 0-000,01 0-000,01 0-000,05 0-000,2 0-001 7 APPENDIX III. ELECTRICAL INSULATORS. OF the available insulators, ebonite, sulphur, amber, sealing wax and fused silica are at present the only ones at all suitable for electroscopic work. With all of these, care should be taken to avoid fingering grease is fatal to insula- tion. In testing insulation, it should be remembered that a delicate electroscope may indicate signs of surface electri- fication for some hours after new insulation has been put in. Such electrification may be dissipated by means of a spirit lamp, or, better, by placing some uranium oxide near the insulator. To reduce the absorption of the electric charge which occurs to a greater or less extent with all insulators, 1 the size of insulating supports should be kept as small as possible in electrometer work. Ebonite that is really good is difficult to obtain nowadays ; it seems to be regarded by most. rubber manufacturers as a convenient means of using up rubber refuse unfit for any- thing else. Some of its defects are occasionally due to the materials used in polishing. Modern ebonite ages with some rapidity in sunlight, and on damp days may, owing to the film of sulphuric acid which forms on its surface, almost play the role of a conductor. In a room which gets much sun- shine most modern ebonite usually turns a dirty yellow colour in a few weeks, though some of the ebonite made ten or twenty years ago will exhibit no signs of deterioration. Notwithstanding its defects, ebonite which has had its 1 Paraffin wax, which is an excellent insulator, shows this soaking effect to a marked and objectionable degree. 256 X RAYS surface recently renewed is an excellent insulator. Ebonite offers the great advantage of being easily workable. Sulphur is convenient, in that it can be cast to shape. In this operation the vessels (glass or porcelain) and sulphur used should be clean, and the temperature should be raised but slightly above the liquefying point of the sulphur. In this limpid condition it can, for example, be poured or sucked into clean warmed glass l tubing, if sulphur plugs are required. The tubing can be readily slipped off later by slightly warming the outside. For some hours after solidification, sulphur can be turned to size or pared to shape with great ease. The insulating properties improve for some time after setting. There is no better insulator than sulphur, but, after a few months, especially in a room which gets much ^sunshine, its insulating qualities generally fall off to a considerable extent. Amber is an excellent insulator, and is almost always, reliable. It can, of course, be obtained in the form of pipe stems, which can be mounted in position with sealing wax. The Amberite and Ambroid companies supply amber pressed to convenient shapes and sizes. Amber has the disadvan- tages of being somewhat brittle and rather expensive. Sealing Wax is particularly useful in that it combines the qualities of an insulator and an air-tight cement with a very low vapour pressure. The insulating properties depend very much on the quality of the wax. One of the most reliable is " Bank of England." The usual care must be taken to avoid indiscriminate fingering. The insulating ability of the wax will be impaired, if in its manipulation it is allowed to catch fire and carbonise, or if a luminous flame is used. As shellac is hygroscopic, sealing wax as an insulator is somewhat susceptible to damp weather. Fused Silica yields place to none in its insulating qualities. Its specific resistance has been determined at the National Physical Laboratory to be greater than 2 x 10 14 ohm cms. at 16 C. Fused silica is practically independent of atmos- pheric humidity, and in the form of rod or tubing is par- ticularly convenient as an insulating material. It is the only 1 Not metal, unless lined, say, with paper. APPENDIX III. 257 high -class insulator which is unimpaired by moderate heat ; it is, however, spoilt if subjected to very high temperatures. Fused silica is now very cheap, but unfortunately the modern furnace methods of production cannot be relied upon to yield a product which possesses the insulating properties of the more expensive silica made by the oxyhydrogen flame. This remark applies alike to the clear transparent variety and the air-streaked satin-like kind. The furnace silica seems to be contaminated in some way, possibly by carbon from the electric furnace. Silica intended for insulation purposes should, of course, be alkali-free. APPENDIX IV. TABLE XXIV. THE ELEMENTS IN THE ORDER OF ATOMIC NUMBERS. International atomic weights for 1917 ; O = 16. The international atomic weights are fixed yearly by an international committee of chemists, consisting at present of Profs. F. W. Clarke (U.S.A.), T. E. Thorpe (Gt. Britain), and G. Urbain (France). The list below comprises 83 elements. For atomic numbers see p. 227. Atomic Num- ber. Atomic Weight. O = 16. Element. Sym- bol. First isolated by 1 1-008 Hydrogen H Cavendish 1766 2 4-00 Helium - He Ramsay and Cleve 1 1895 3 6-94 Lithium - Li Arfvedson 1817 4 9-1 Beryllium (Glucinum) - Be Wohler and Bussy 1828 5 11-0 Boron B Gay-Lussac & Thenard 1808 6 12-005 Carbon - C Prehistoric 7 14-01 Nitrogen N Rutherford 1772 8 16-00 Oxygen - Priestley and Scheele 1774 9 19-0 Fluorine F Moissan 1886 10 20-2 Neon Ne Ramsay and Travers 1898 11 23-00 Sodium - Na Davy 1807 12 24-32 Magnesium Mg Liebig and Bussy 1830 13 27-1 Aluminium Al Wohler 1827 14 28-3 Silicon - Si Berzelius 1823 15 31-04 Phosphorus P Brand 1674 16 32-06 Sulphur - S Prehistoric 17 35-46 Chlorine Cl Scheele 1774 18 39-88 Argon A Rayleigh and Ramsay 1894 19 39-10 Potassium K Davy 1807 20 40-07 Calcium - Ca Davy 1808 21 44-1 Scandium Sc Nilson and Cleve 1879 22 48-1 Titanium Ti Gregor . 1789 23 51-0 Vanadium V Berzelius 1831 24 52-0 Chromium Cr Vauquelin 1797 25 54-93 Manganese Mn Gahn 1774 26 55-84 Iron Fe Prehistoric 27 58-97 Cobalt - Co Brand 1735 28 58-68 Nickel - Ni Cronstedt 1751 29 63-57 Copper - Cu Prehistoric 30 65-37 Zinc Zn Mtd. by B. Valentine 15 centy. 31 69-9 Gallium - Ga L. de Boisbaudran 1875 32 72-5 Germanium - Ge Winkler 1886 33 74-96 Arsenic - As Albertus Magnus 13 centy. 34 79-2 Selenium Se Berzelius 1817 35 79-92 Bromine Br Balard 1826 36 37 82-92 85-45 Krypton Rubidium Kr Rb Ramsay and Travers 1898 Bunsen and Kirchhoff 1861 1 Jansseii and Lockyer (in sun), 1808. APPENDIX IV. 259 THE ELEMENTS IN THE ORDER OF ATOMIC NUMBERS Continued. Atomic Num- ber. Atomic Weight. O = 16. Element. Sym- bol. First isolated by 38 87-63 Strontium Sr Davy 1808 39 88-7 Yttrium Y Wohler 1828 40 90-6 Zirconium Zr Berzelius 1825 41 93-1 Niobium (Columbium) Nb Hatchett 1801 42 96-0 Molybdenum - Mo Hjelm 1790 44 101-7 Ruthenium Ru Claus 1845 45 102-9 Rhodium Rh Wollaston 1803 46 106-7 Palladium Pd Wollaston 1803 47 107-88 Silver - Ag Prehistoric 48 112-40 Cadmium Cd Stromeyer 1817 49 114-8 Indium - In Reich and Richter 1863 50 118-7 Tin Sn Prehistoric 51 120-2 Antimony Sb Basil Valentine 15 centy. 52 127-5 Tellurium Te v. Reichenstein 1782 53 126-92 Iodine - I Courtois 1811 54 130-2 Xenon - Xe Ramsay and Travers 1898 55 132-81 Caesium - Cs Bunsen and Kirchhoff 1861 56 137-37 Barium - Ba Davy 1808 57 139-0 Lanthanum La Mosander 1839 58 140-25 Cerium - Ce Mosander 1839 59 140-9 Praseodymium Pr Auer von Welsbach 1885 60 144-3 Neodymium - Nd Auer von Welsbach 1885 62 150-4 Samarium Sa L. de Boisbaudran 1879 63 152-0 Europium Eu Demarcay 1901 64 157-3 Gadolinium Gd Marignac 1886 65 159-2 Terbium Tb Mosander 1843 66 162-5 Dysprosium - Dy Urbain & Demenitroux 1907 67 163-5 Holmium Ho L. de Boisbaudran 1886 68 167-7 Erbium - Er Mosander 1843 69 168-5 Thulium Tm Cleve 1879 70 173-5 Ytterbium (Neo, Yb) - Yb Marignac 1878 71 175-0 Lutecium Lu Urbain 1908 73 181-5 Tantalum Ta Eckeberg 1802 74 184-0 Tungsten W Bros. d'Elhujar 1783 76 190-9 Osmium - Os Smithson Tennant 1804 77 193-1 Iridium - Ir Smithson Tennant 1804 78 195-2 Platinum Pt 16 centy. 79 197-2 Gold Au Prehistoric 80 . 200-6 Mercury - Hg Mtd. by Theophrastus 300 B.C. 81 204-0 Thallium Tl Crookes 1861 82 207-20 Lead Pb Mentd. by Pliny Prehistoric 83 208-0 Bismuth Bi Mtd. by B.Valentine 15 centy. 86 222 Radium Eman- ation (Niton) Ntr M. and Mme. Curie 1900 88 226-0 Radium - Ra Curies and Bemont 1898 90 232-4 Thorium Th Berzelius 1828 92 238-2 Uranium U Peligot 1841 260 X RAYS ATOMIC NUMBERS AND WEIGHTS OF THE EADIOACTIVE ELEMENTS. Element. At. No. At. Wt. Element. At.Wt. Uranium 1 92 238-2 Thorium - 90 232-4 2 92 234-2 Meso -thorium 228-4 ,. X, Y - 90,91 230-2 Radio -thorium 228-4 Ionium 90 230-2 Thorium X - 224-4 Radium 88 226-0 Emanation 220-4 Ra Emanation 86 222 A - 216-4 Radium A 84 218 B, C 19 C 2 82 212-4 B, C, - 82, 83 214 D* - 208-4 C 2 , D, E F (Polonium) l 82, 83 84 210 210 1 Probably converted into Pb. 2 Bi. The atomic weight of actinium is probably about 230. TABLE XXV. ATOMIC WEIGHTS AND DENSITIES OF THE ELEMENTS. Element. Atomic Weight. Density. Element. Atomic Weight. Density. Element. Atomic Weight. Density. Al 27 2-70 He 4 0-178 % Rb 85 1-532 Sb 120 6-62 H 1 0-08987J Ru 102 12-3 A 40 1-78 J In 115 7-12 Sa 150 7-8 As 75 5-73 I 127 4-95 Sc 44 Ba 137 3-75 Ir 193 22-41 Se 79 4-5 Be 9 1-93 Fe 56 7-86 Si 28 2-3 Bi 208 9-80 Kr 83 3-708 % Ag 108 10-5 B 11 2-5 ? La 139 6-12 Na 23 0-971 Br 80 3-10 Pb 207 11-37 Sr 88 2-54 Cd 112 8-64 Li 7 0-534 S 32 2-07 Cs 133 1-87 Lu 175 Ta 181 16-6 Ca 40 1-55 Mg 24 1-74 Te 127 6-25 C Mn 55 7-39 Tb 159 Diamond 12 12 3-52 Hg 201 13-56 Tl 204 11-9 Graphite 2-3 Mo 96 10-0 Th 232 11-3 Gas carbon 1-9 Nd 144 6-96 Tm 168 Ce 140 6-92 Ne 20 0-9002 % Sn 119 7-29 Cl 35 3-23 | Ni 59 8-9 Ti 48 4-50 'Cr 52 6-50 Nb 93 12-75 W 184 18-8 Co 59 8-6 N 14 1-2507 J U 238 18-7 Cu 64 8-93 Os 191 22-5 V 51 5-5 Dy 162 16 1-429 % Xe 130 5-851 Er 168 4-77 ? Pd 107 11-4 Yb 174 Eu 152 P 31 2-2 Y 89 3-8? F 19 1-69 { Pt 195 21-5 Zn 65 7-1 Gd 157 K 39 0-862 Zr 91 4-15 Ga 70 5-95 Pr 141 6-48 Ge 72 5-47 Ra 226 VJ TrO ? Paper 1-0 Au 197 19-32 Rh 103 12-44 J Grms. per litre at C. and 760 mm. APPENDIX IV. 261 TABLE XXVI. /=7 e-K TABLE CONNECTING J/I AND Xd. E.g. if Ad=-693, Ijl^'5. e=2-71828. (See p. 104.) (From Kaye & Laby's Physical Constants.) For values of x ,- ) * c. 1 Fatigue of fluorescence in glass - , " .. 12 ,, fluorescent screens - . 97 ,, in production of secondary rays 150 ,, of selenium - 97 Fi/eau, condenser of induction coil 55 Fluctuation experiments with y rays &-.. 247 light rays -iU : r!">,, 247 Fluorescence by cathode rays - 12 ,, in glass, fatigue of - :'*- 12 of glass at moderately high pressures 3 ,, ,, lithium chloride by cathode rays and positive rays 19 ,, methods of measuring intensity of X rays - 97 ,, of X-ray tubes, colour of - 3 Fluorescent screen - :.'>.' 97 ,, ,, display of Laue diffraction spots 1: : .' 210 Focus tube. Crookes' - , 30 ., Jackson's > - *< 30 ,, modern types of 41 Focussing distance of cathode rays - - - :<*>< 34 ring in X-ray tube 41 Fog experiments of C. T. R. Wilson 155 Foot-switch 181 Franck and Pohl, velocity of X rays , -.' 159 Frequency of interrupters - T - 66, 67 Freund, scale of X-ray dosage 97 Friedrich and Knipping, X rays and crystals - 195, 197 Friman and Siegbahn, X-ray wave-lengths 226, 229 Furnace, cathode-ray *.,- 11 Fiirstenau, fatigue of selenium 98 Fused silica, insulating properties of * 256 Gaede, mercury pump - - 252 ., molecular ,, - - 251 ,, piston '-],; *. ; 252 rotary ... . 252 Gaiffe-Barret X-ray tube - 35 Galimard and Bordier, scale of X-ray dosage v j. 97 Galvanometers ''-'-,' 94 y rays, absorption of - 105, 107, 263 characteristic radiations - 122 fluctuation experiments - -./ . 247 from a rays r-'3V,," 247 reflection of - 231 spectrum of soft rays - , - 229 Gardiner, distribution of X rays from bulb 49 magnetic displacement of anticathode focus-spot 43 ,, photomicrograph of anticathode - - 40, 43 Gases, absorption of characteristic rays in - 142 corpuscular ,, ,, - - 147 274 X RAYS PAGE Gases, absorption of y rays in 107 ,, held by glass surfaces - - 253 Geissler, discharge through gases - - xxi cathodic sputtering 81 Geitel and Elster, coloration of X-ray bulb 87 George and Willows, use of silica bulbs - 76 Glass, conducting - 171 ,, for influence machines 51 ,, insulating properties of 51 lithium 172 transparency of, for X rays - 171 Glasson, characteristic X rays from salts - 124 ionisation by cathode rays - 12, 111 Goby, radiomi orography - 163 Goldhammer, nature of X rays - - 235 Goldsmith, penetration of mica by positive rays - 77 Goldstein, canal-rays or positive rays 19 cathode-rays - 4 ,, electrification of glass round cathode - 74 Gorton, reflection of X rays by mica '" * ' 213 Gouy, gas bubbles in glass walls of X-ray tube - 76 Gowdy, fatigue in production of secondary rays 150 Granquist, cathodic sputtering 85 Gray, characteristic y rays - 123 Gundelach, introduction of auxiliary anode 32 Guest and Chapman, characteristic X rays from salts 124 Haga, polarisation of X rays u 115 Haga and Wind, diffraction of X rays 194 Half -value thicknesses for absorption - 104, 109, 119 Ham, depth of origin of X rays in anticathode - 48 ,, distribution of X rays from bulb 49 ,, polarisation of X rays 116 ,, thickness for reflection of cathode rays 49 Hammer break - 28, 66 Hardness of an X-ray bulb - 72 ,, ., ,, ,, methods of measuring 98 varying - 78 Heat generated by absorption of X rays 91 cathode rays 11 Hehl, cathode-glow 2 Helical path of cathode rays in magnetic field - 13 Helium, presence in old X-ray tubes 76 Hertz, deflection of cathode rays - 7 nature ,, ,, ,, 5 Hertzian waves, wave-length of " *':. 233 Herweg, polarisation of X rays 115 High-frequency oscillations in primary circuit of coil 62 spectra of metals - 225 High-tension circuits - -* 184 transformers - - 63, 177 Hill, hardening of discharge tube - 76 Hittorf, cathode rays - 4 tube ..... . . xxi INDEX 275 PAGE Hodgson and Brodetsky, hardening of X-ray tube- - - V 75 Holborn and Austin, cathodic sputtering - 85 Holzknecht, scale of X-ray dosage - 97 Homogeneous X rays - 116 Horizontal X-ray couch - - 186 Hot-lime cathode - 9 Hughes, energy of phot electrons - - 237 ,, ionisation by ultra-violet light - - 237 Hull and Rice, Planck's X-ray relation - - 245 wave lengths and absorption coefficients - - 230 Hulst, influence machines - 53 Hurmuzescu, discovery of ionisation by X rays - - 25 Hydrogen X-ray tube - 46 "I "radiation - 122 Induction coil, condenser of 55 core of 54 design of - 59, 177 detailed account of - 53 efficiency of - 60 elementary account of 27 primary tube of - > * 56 primary winding of - 55 secondary winding of - 57 Influence machines - 51 Innes, velocity of corpuscular rays - - 146 Instantaneous radiography - 165 Insulators, electrical - - 255 Intensifying screens - 166 Intensity of X rays, current method of measuring - 90 definition of - 90 fluorescence method of measuring 97 ionisation ,, ,, ,, 91 methods of measuring used in medicine - 97 photographic methods of measuring 95 thermal ,, ,, ,, 91 Interference of y rays - 231 Xrays - 195 International Radium Standard - 88 Interrupter, sparking at - 60 types used on induction coils - 66 use of 178 Wehnelt, electrolytic, design of 66 Inverse currents - 28 ,, ,, as affected by coil design - 59 square law for X-ray intensity 90 Ionics of an X-ray tube - 23 Ionisation chambers 94 by collision 93 in heavy gases - - 152 independent of quality of X rays - 152 methods of measuring X-ray intensity - 91 and pressure - 151 produced by cathode rays - 11 276 X RAYS PARE lonisation and temperature - - - - 154 by ultra-violet light 237 by X rays 151 .. ,, indirect action of 152 in mixed gases 151 Ions - xv , 20, 92 Iridium, use as anticathode 39 "Isotopes"- +K . ... 229 Jackson, fluorescence of X-ray tubes - ..*' ... 3 ,, focus bulb - .,-. 30 Jaumann, nature of X rays '' * 235 Jeans, Planck's quantum hypothesis - 245 Jona, sparking voltages - - 102 Jones and Roberts, function of condenser of induction coil ;' 56 K characteristic radiation - 117 Kanalstrahlen - 19 Kathodenstrahlen - 4 Kaufmann, fifth-power law of characteristic radiation . * ; ] 20 ,, various anticathodes - 35 Keene, diffraction of X rays by metallic crystals , 210 KienbSck, scale of X-ray dosage - 97 Kleeman, distribution of photoelectrons - 146 Klingelfuss, qualimeter - 109 Knipping and Friedrich, crystals and X rays 195, 197 Knox, radiology - 177 Kowalski and Rappel, sparking potentials - r .-- 100 Kroncke, intensity of X rays 90 selective absorption in photographic plates - 96 Kunzite, fluorescence of 13 L characteristic radiation - 117 Laby and Burbidge, y-ray fluctuation experiments - 246 Laird, entladungstrahlen - . - 233 Langer, various anticathodes 35 Langevin, electron theory of magnetism 18 Langmuir, disintegration of tungsten 87 Larmor, electron theory of magnetism 18 Lattice, space-, definition of 196 Laub, characteristic " I " radiation - ..." , 122 distribution of corpuscular rays 145 Laue, diffraction spots, displayed by fluorescent screen -\ , - 210 elliptical loci of - :5 f 207 shape of - - 207 ,, ,, ,, produced by zinc-blende - . * 200 ,, interference of X rays 195 , theory of crystal diffraction of X rays 196 Lead, use in protective devices - 168 Lenard, ionisation by ultra-violet light - :" 237 ,, law of absorption of cathode rays 10 rays 5 ,, therapeutic use of 172 Levy, barium platinocyanide pastilles - : 98 INDEX 277 PAGE Levy, barium platinocyanide screen - " '"" 97 Light rays and X rays, resemblance between - - 237 Lime cathode 8 Lindemann, lithium glass - 171 Lindemann and de Broglie, reflection of X rays by mica - >' 213 Lithium chloride, fluorescence of, by cathode and positive rays -^ 19 ,, glass 171 Localised pulse theory - 244 Localisation- 164, 191 Lodge, Sir Oliver, metal X-ray bulb 47 ,, ,, valve tube ' 71 various anticathodes : - -' 35 Lorentz, electron theory of matter - 18 Luminosity produced by cathode rays 12 Lyman, ultra-violet rays - - " 233 Machines, influence - 5i Wimshurst 52 Mackenzie and Soddy, absorption of gas by sputtered metal - 78 Magnetic deflection of cathode rays 13 displacement of focus-spot 43 field, effect on X rays - * 25 ,, spectrum of cathode rays - 15 Magnetism, electron theory of ' 18 Maltezos, electrification of glass near cathode - 74 Martyn and Barkla, selective action of X rays on photographic plate - 95 Marx, velocity of X rays - 158 Mass -absorption coefficients, definition of - 104 ,, ,, see absorption coefficients. Mass-iscattering coefficient, see scattering coefficient. Matter, electron theory of - - 18 Mechanism of production of X rays - 23 Medical applications of X rays - - 169 Mees, Kenneth, photographic plates for X-ray work - 166 Melting points of metals - - 39, 86 Mercury-breaks - 68, 178 Merton, precipitated copper method of producing vacua - - 254 Metallic crystals, diffraction of X rays by - -- 210 Metals used as anticathodes - 38 Metal X-ray tubes - 47 Meyer, y-ray fluctuation experiments 247 Michelson, vortex theory of X rays - - ' 235 Micro -radiography - -* 163 Miller, polarisation of X rays - 116 Miller, Leslie, mica-disc valve 71 Milliammeters 182 Mixed gases, ionisation in - - 154 Monochromatic X rays 116 Moore, corpuscular rays ~ - 145 More, fatigue in production of secondary X rays X* - 150 Moscicki condenser - *- J --v- '' - 51 Moseley, reflection of X rays - 212 X-ray spectra - - .',-*?& * 224 278 X RAYS PAGE Moseley and Bragg, characteristic rays - - - 122 Moseley and Darwin, characteristic rays from platinum - 122, 216 reflection of X rays - 211 Muller, X-ray bulbs 43 Multisectional winding of secondary of induction coil 59 National Physical Laboratory, Radium Standard - 88 Negative glow 1 ,, ion xv Nernst, specific heats at low temperatures - ; -_- . 19 Neutral-pair theory of the X ray - 236 Newton, corpuscular theory of light - 244 Nickel, platinised, use as anticathode 40 Nicol, mass-absorption coefficients - 119 Noire and Sabouraud, scale of X-ray dosage 97 Northern lights 17 Nucleated pulse theory of X rays - - - 244 Obliquity of anticathode - 48 Occlusion of gas by electrodes - 75 ., methods of softening X-ray bulb 78 Oil-pumps - <.,. - 253 Opacity, definition of 97 Opacity-logarithm - 97 Opacity meter 96 Oscillograph records of primary and secondary currents - 61 Oscilloscope - 185 Osmium, use as anticathode 40 Osmosis method of regulating X-ray bulbs - 46, 78 Owen, absorption of characteristic radiations in gases - 142 5th-power law of absorption - 119, 143, 230 ionisation (X-ray) and pressure 151 metallic window used as anticathode 142 relative X-ray ionisation in gases - 152 scattering of X rays - 114 total X-ray ionisation in gases - 154 Owen and Blake, diffraction of X rays by metallic crystals 210 Paschen, attempt to deflect y rays magnetically - 25 galvanometer 94 Pastille method of dosage - 98 Penetrometers 107 Perrin, negative charge on cathode ray - 7 Phillips, C. E. S., conducting glass - 171 function of anode 33 ,, radium standard - 88 Philpot and Barkla, ionisation independent of quality of X rays - 152 ,, ,, total ionisation in gases 154 ,, ,, ,, X-ray and corpuscular ionisation in gases - 152 Philpot, Barkla and Simms, generation of corpuscular radiation - 150 Photoelectric effect, selective 124 Photoelectrons, distribution of 146 energy of - -,,---. 237 speed of - - " 238 INDEX 279 , ' PAGE Photographic film, absorption by - - ''-'> 96 methods of measuring intensity of X rays - 95 plate, action of X rays on - 25 ,, selective action of X rays on - 95 plates for radiography 166 Photographs of X-ray tracks 155 Physiological action of characteristic rays - 172 Xrays 169 Piezo-electrique - 95 Planck, quantum theory of X rays - - 236, 237, 245 theory of radiation - 195 universal constant - 224 X-ray relation - - 245 Plastic printing - 167 Platinised nickel anticathode - 40 Platinocyanide of barium, fluorescence of - - 13 pastilles of --> ' ' 98 ,, screens of - 97 Platinum radiation - - -215,223 wave-length of - - -' 223 spectrum of X rays from - 215 use as anticathode 38 Pliicker, cathode rays 4 ,, cathodic sputtering 81 ,, hardening of discharge tube 75 tube xxi Pohl and Franck, velocity of X rays 159 *, Pringsheim, characteristic light rays - 123 Walter, diffraction of X rays - 194 Poincare, relation btween phosphorescence and X rays 30 Point and plane spark-gap rectifier - 70 Polarisation of X rays 1J4 theory of - - 240 Pope and Barlow, crystal structure of zinc blende - 203 Positive column - 1 ,, electron, absence of 20 ion - xv rays 19 e/moi- - 21 striae - 2 Potassium-sodium alloy, absorption of gases by - - 253 Pressure of gas in X-ray tube 23 ,, and ionisati on (X-ray) in gases - 151 Primary current wave-form - 6 1 tube of induction coil 56 winding ,, 55 Pringsheim and Pohl, characteristic light rays - J-- 123 Production of high vacua - "Y 251 ,, Xrays 25 Progressive hardening of X-ray bulb - 75 Protection of X-ray operators, recommendations for - 264 Protective devices against X-ray burns 167, 189 Pulse-theory of X rays, Stokes - 239 ,, ,, ,, modification necessary 241 280 X RAYS PAGE Pulse-theory of X rays, nucleated - - 244 Puluj, cathode rays 4 Quality of X rays, methods of measuring - 98 and potential on tube - 99 Quantum theory of X rays - - 236, 237, 245 Quartz, (fused) insulating properties of '"- - 256 Radiation, discontinuity of - 236 : Planck's theory of 195 Radioactive elements, table of atomic weights, etc. 260 Radiochromometer - 107 Radiographs - 161, 174-176 Radiography 160 ,, bismuth 163 instantaneous - 165 micro- - 163 stereoscopic - 163 Radiometer, Crookes' 5 Radio-micrography - 163 Radiotherapy, recent advances - ' * 173 Radium y rays, therapeutic uses -160, 171 ,, Standard, International and British 88 ,, treatment of X-ray burns 167 Ramsay and Collie, gas in glass of X-ray bulb ' 76 Rappel and Kowalski, sparking potentials - 100 Rayleigh, condenser of induction coil 55 nature of white light .i ' - 240 Re, nature of X rays " - . 235 Recoil atoms xv Rectifiers 70 Rectifying properties of X-ray bulb 34 ,, spark-gap, effect on maximum potential 62 References to Journals - xxii '' Reflection " of cathode rays 49 Xrays '... 211 ,, ,, early attempts - 195 y rays 231 Refraction of X rays 194 Regulation of X-ray tubes - - 78, 184 " Relief " photographs of X rays - ; 167 Retrograde rays - 20 Reverse currents (see also inverse currents) 28 " Revolution of the Corpuscle " xvi Rhodium, use as anticathode 40 X-ray spectrum - - 217 Rice and Hull, wave-lengths and absorption coefficients 230 ,, Planck's X-ray relation - ' - ? 245 Richardson, characteristic rays excited by /3 rays 123 Richardson and Rutherford, characteristic y rayg - 122, 263 Richardson, Barnes and Rutherford, absorption curves for Coolidge tube 128 ., break-down voltage of Coolidge tube - - '< -: 46 INDEX 281 PAGE Richardson, Barnes and Rutherford, intensity of X rays - 90 Rieman, fatigue in production of secondary X rays 150 Roberts, volatilisation of metals - 86 Roberts and Jones, function of condenser of induction coil 56 Roberts-Austen, interdiffusion of metals - 6 Rock-salt, atomic distances in /- 223 ,, crystal structure of 219 Rciti, various anticathodes - 35 Rontgen, diffraction of X rays 194 ,, discovery ,, ,, - 24, 249 nature ,, - 235 ,, penetrometer - 108 refraction of X rays 194 ,, various anticathodes - 35 Rontgen Society, radium standards 88 Russell and Chad wick, characteristic y rays 122 Rutherford, y rays from (3 rays 122, 247 International Radium Standard 88 theory of the atom - - 229 very hard X rays 46 ,, wave-lengths and absorption coefficients - 230 Rutherford and Andrade, reflection of y rays 229, 231 Rutherford and Barnes, energy of X-rays - "^ 111 Rutherford, Barnes and Richardson, absorption curves for Coolidge tube 128 ,, ,, break-down voltage of Coolidge tube 46 ,, ,, ,, intensity of X rays - 90 Rutherford and Richardson, characteristic y rays - 122, 263 Sabouraud and Noire, scale of X-ray dosage - 97 Sachs, St., and Winawer, standardisation ot X rays - 89 Sadler, absorption of corpuscular rays in gases - - 149 ,, mass-absorption coefficients - 119 Sadler and Barkla, absorption relation for characteristic radiations 136 ,, ,, ,, discovery of characteristic radiations - ' * 116 Sagiiac and Curie, corpuscular rays 145 Sagnac, nature of X rays - 235 Salomonson, current through X-ray bulb - 91 influence of dielectric in mercury breaks 69 ,, wave-form of current in primary of induction coil - 61 Sanax mercury-break 69 Saturation current - 92 Scattered X rays -112, 220 distribution of - -114,238 ,, polarisation of - 114 ,, scattering-coefficient - 113 ,, ,, and light, resemblance between - * 238 Schott, glass specially transparent to X rays / 171 Schumann, ultra-violet rays 233 Schuster, nature of white light - 240 Schwartz, scale of X-ray dosage - 97 Schweidler, y-ray fluctuation experiments - - 246 Screening stands - - - . 186 282 X RAYS PAGE Screens, barium platinocyanide - - - 97 intensifying - 166 protective - 190 Sealing-wax, insulating properties of 256 Secondary current wave-form - 61 winding of induction coil - 57 Xrays - - 112 Seitz, very soft X rays ] 23 Selective absorption 105 ,, of characteristic radiations . 136 ,, action of X rays on photographic plates 96 photoelectric effect - 124 Selenium, effect of X rays on 97 Separation distances of atomic planes in crystals - - 221 Shearer, X-ray ionisation in gases - 153 Shearer and Barkla, nature of corpuscular radiation 150 Shielding of anticathode - 41 " Shunt " regulation of induction coil 180 Siegbahn, metal X-ray bulb 47 ,, and Friman, X-ray wave-lengths 226, 229 Silica, fused, insulating properties of - 256 Silk, protective, for X-ray work - 168 Simms, Barkla and Philpot, generation of corpuscular radiation - 150 Size of X-ray tubes, effect of 31 Snook,, high-tension transformer - 63 ,, hydrogen tube 46 Soddy, "isotopes " - - 229 Soddy and Mackenzie, absorption of gases by sputtered metal - 78 Sodium-potassium alloy, absorption of gases by - - 253 Soft X rays - - 123 Softening of an X-ray bulb - 78 Sommerfeld, spreading-pulse theory of X rays . ..- - 243 Space-lattice, definition of - - 196 ,, dimensions of - 223 ,, of alkali-halogen salts - 218 Spark-gap 100, 182 effect on maximum-potential:: r" 62 rectifying 70 Sparking at interrupter - '.-.. 160 ,, voltages, table of - 101 Spectrometer for X rays - 212 Spectrum of cathode rays - . * 15 X rays - 224 Speed of X rays - 158 Sprengel pump - 252 Sputtered metal, absorption of gas by 77 Sputtering of cathode - 80 thermal - 85 Standard of radioactivity - 88 Standardisation of X rays - 89 Stark, distribution of X rays from thin anticathodes 50 ,, wave-lengths of X rays - 195 ,, quantum theory of X rays - - 236 Step-up transformers - - 63 INDEX 283 PAGE Stereographic projection of Laue diffraction spots - - 208 Stereoscopic radiography - 163 Stokes, ether-pulse theory of X rays - 235, 239 ,, law of fluorescence - - 238 Stoney, Johnstone, use of term " electron " 8 Straubel, osmosis valve - 78 Striae in discharge tube 2 Strutt, active nitrogen 77 Stuhlmann, distribution of photo -electrons 146 Sublimation of metals 85 Sulphur, insulating properties of - - 256 Sunic screen 166 Sutherland, nature of X rays - 235 Swinton, Campbell-, adjustable cathode 74 early radiograph 161 fluorescence in glass, fatigue of 12 hardening of discharge tube - 76 historical X-ray bulb - 30 shape of beam of cathode rays 34 various anticathodes - 35 Tantalum, use as anticathode 38 Taylor, fluctuation experiments with light - - 247 Temperature and X-ray ionisation. - 154 Terada, fluorescent screen observations of Laue spots - 210 Therapeutics, action of X rays in - 169 suitable rays for 170 use of corpuscular rays in - 172 ,, characteristic - 172 Thermal conductivities of metals - 39 ,, disintegration of anticathode 85 methods of measuring intensities of X rays 91 Thompson, S. P., various anticathodes - 35 Thomson, J. J., analysis of absorption curves 104 discovery of ionisation by X rays - 25 electric deflection of cathode rays - 15 e/m of cathode rays 14 e/m of positive rays 21 energy of the X ray - 133 , excitation of X and y rays by a rays - 247 fluorescence of walls of discharge tube - 2 4th-power absorption law for cathode rays 10 gas expelled by cathode-ray bombardment 76 nature of cathode rays - 7 negative charge on cathode rays - 7 nucleated or localised pulse theory of X rays 236, 244 positive rays 20 scattering law for X rays - 114 theory of the atom 18 very soft X rays - 123 Tiede, cathode-ray furnace - 11 Topler, pump , > 252 ,i spar king- voltages - 100 Total ionisation in gases - 'V. -93,154 284 X RAYS PAGE Transformers, high-tension - 63 Transmission of cathode rays 10 Transparency, definition of - 97 of substances to X rays - 104 Transparent glass for X-ray tubes - 171 Trinkle and Wehnelt, very soft X rays - 123 Trowbridge, long-spark voltages - 102 Tungsten, use as anticathode - 38, 44 cathode 44 Ultra-violet light, ionisation by - 237 Upright screening -stand - 186 Vacua, production of 251 Valve, Bauer 79 ,, osmosis 78 tubes 70 Varley, nature of cathode rays - 6 Vegard, crystal structure of silver - 211 polarisation of X rays 115 Velocity of cathode rays 16, 99 ,, ,, and potential, table of - - 262 ,, corpuscular rays 146 Wehnelt cathode rays - 9 X rays - 158 Villard, dosage scale 98 gas bubbles in wall of X-ray tube - 76 osmosis valve 78 valve tube - 70 Villard and Abraham, influence machines - 53 Violet coloration of X-ray bulb - 87 Violle, interrupter - 66 Volatilisation of anticathode 85 metals - 86 Voltz, selective absorption in photographic plates - 96 Walter, attempt to deflect X rays magnetically - 25 penetrometer 108 Walter and Pohl, diffraction of X rays 194 War radiographs 174, 175 Warburg, thickness required for reflection of cathode rays 49 Wartenburg, cathode-ray furnace - 11 Water-cooled anticathodes - 41 Wave-form of primary and secondary currents in induction coil - 61 Wave-lengths of y rays - - 231 various electromagnetic waves - - 233 X rays - 102, 223 ,, and absorption coefficients 230 ,, relation to atomic number 227 table of ;^ 226 Webster, Planck's X-ray relation - 245 Wehnelt, adjustable cathode 74 interrupter - 66, 179 lime cathode ..... 8 INDEX 285 PAGE Wehnelt, penetrometer - 108 valve tube 71 Wehnelt and Trinkle, very soft X rays 123 Whiddington, absorption of corpuscular rays 148 adjustable cathode - 74 characteristic X rays and cathode-ray velocity - 131 energy of X rays 133 ,, formula connecting K and L radiations - 120 fourth-power law of absorption of cathode rays - 10, 111 Wien, energy of X rays - 110 deflection of positive rays - 19 wave-length of X rays 195 Willemite, fluorescence of - 13, 20 Willows, hardening of discharge tube 76 Willows and George, use of silica bulbs 76 Wilson, C. T. R., condensation experiments 155 ,, ,, ionisation by X rays 152 reflection of X rays - 211 ,, tilted electroscope 95 Wilson, W. H., function of condenser of induction coil - 56 ,, oscillations in primary circuit of coil 62 Wimshurst machine - . 52 Winawer, efficiency of production of X rays 110 Winawer and St. Sachs, standardisation of X rays 89 Wind and Haga, diffraction of X rays 194 Winding of primary of induction coil 55 ,, secondary ,, ,, - 57 Window, metallic, use as anticathode 142 Winkelmann, effect of size on hardness of tube - 73 hardness of X-ray tube 72 osmosis valve 78 " Wireless " waves, wave-length of - 233 Worrall, instantaneous radiography 165 Wratten and Wainwright, X-ray plate 166 Wright, design of induction coil 59 Zehnder, metal X-ray bulb - 47 Zinc-blende, crystal structure of - 200 fluorescence of- - - - -13 PRINTED IN GREAT BRITAIN BY ROBERT MACLT5KOSE AND CO. LTD. AT THE UNIVERSITY PRESS, GLASGOW. THIS BOOK IS DUE ON THE LAST DATE STAMPED BELOW AN INITIAL FINE OF 25 CENTS WILL BE ASSESSED FOR FAILURE TO RETURN THIS BOOK ON THE DATE DUE. THE PENALTY WILL INCREASE TO 5O CENTS ON THE FOURTH DAY AND TO Sl.OO ON THE SEVENTH DAY OVERDUE. OCT19 OCT 9 1933 DEC 11 ; LD 21- VC I I 146 UNIVERSITY OF CALIFORNIA LIBRARY