VOL.'3 DECEMBER 1938 Philips Technica.1 ·Rev.iew, DEALING WITH TECHl\'lCAL PROBLEMS RELATING. TO THE PRODUCTS,. PROCESSES ~D INVESTIGATIONS OF N.V...PHILIPS' GLOEILAMPENFABRIEKEN EDITED BY THE RESEARCH LABORATORY OF N.V. PHILIPS' GLOEll.AMPENFABRIEKEN, EINDHOVEN, HOLLAND

ZIRCONIUM' AND ITS COMPOUNDSWITH A HIGH ·MELTING POINT

by J. D. F~ST.

In this article the preparation of zirconium and its mechanical, chemical and electrical properties are dealt with. The most important applications ofzirconium and its compounds with high. melting points are then discussed. Zirconium is often used in discharge tubes where its high melting point, small value of secondary emission and, when use'das f1 getter, its property pf being able to take up large quantities of oxygen, nitrogen and' hydrogen are of significance. Zirconium oxide and zirconium' carbide are 'important as refractory materials.

Introduction . ; Even since the beginning of the development of tungsten as filament in incandescent lamps, and the the manufacture of electric incandescent lamps, use of this element in the place of all the other much research has been carried out on metals with substances which were used for that purpose. high melting points in the laboratories of such fac- Several others of the metals mentioned, such' as ,tories. The metals in the following groups of the Zirconium, tantalum and molybdenum, have also periodic system drew particular attention: found uses, although less important ones. 1) The platinum group: platinum, iridium and The aim of this article is to make several state- osmium, the last of which was used for some ments about the properties and. ~.applications. of time as filament in the sources of electric light. the element. zirconium an? severa_l~ts comi)Qun~s. 2) The fourth main group: titanium, zirconium, We shall discuss, successively the metal, the oxide

hafnium and thorium. (Zr02) and the carbide (ZrC). 3) The fifth main: group: vanadium, niobium and The metal in the pure state is used chiefly in tantalum, the last of, which was used for some transmitting valves and other discharge tubes, time in so-called tantalum lamps. where use is made of its great chemical affinity 4), The sixth main group: chromium, molybdenum for various other elements such as oxygen, nitrogen, and tungsten. carbon and hydrogen. The oxide is chiefly used as a At the beginning óf the development of the elec- refractory and as an opacifier for glazes and tric incandescent lamp industry (around 1880) prac- enamels. The carbide is stilllittle used, but it shares tically nothing was known about the mechanical and the interest which has existed during the last ten or physical properties of the last th~ee groups men- fifteen years for the metallic car"bides with a very tioned, and very little about their chemical prop- high melting point and very great hardness. erties. The interest of this industry, however, contributed very much to the research which has Brittle and ductile zirconium led to the fairly extensive knowledge of these Twenty-four years ago the first scientific pub- metals which we now possess. This interest was at lication from the Philips Laboratory appeared 1). first chiefly directed toward the melting point It contained statements about the preparation of and the speed of evaporation, but when the electric metallic thorium, uranium, zirconium and titanium. lamp industry also began to manufacture X-ray Zirconium was prepared by reduction of zirconium tubes, transmitting and receiving valves for radio tetrachloride (ZrCla) with sodium. The metal was purposes, rectifier valves etc.v-It began to include 1) D. Lely and L. Hamburger, Herstellung der Elemente all the properties of the metals under consideration. Thorium, Uran, Zirkon und Titan, Z. anorg. allg, Chem. 87, The technical result was the present general use of . 209, 1914. -,.~,. fi

PHILIPS TECHNICAL REVIEW Vol. 3, No. 12 s, ( i obtained in the form of a powder and the size detrimental influence on the mechanical properties j , of the grain depended upon the reduction tem- of metals, so that very small amounts are already perature, the nature of the chloride, etc. Later enough to make the metal brittle. Titanium which I the ~ethod was improved upon by previously can also contain considerable quantities of oxygen subliming the chloride in hydrogen and thus and nitrogen in solid solution, exhibits the same obtaining it in large compact pieces. If this sub- phenomena as zirconium. Thorium, however, which limed chloride is used, and if it is reduced in belongs to the same group of the periodic system as large porti~ns at a fairly high temperature, the titanium and zirconium, is found to have no ap- nietal is obtained in, the form of chunks with di- preciable dissolving power for oxygen and nitrogen. mensions of several, centimetres, which are, how- The result is that this last metal does not exhibit ever, porous. This method of preparation pro- the phenomenon mentioned': thorium rods can also duces metal of great purity and is still one of the be obtain~d in the ductile state by compression best methods of preparing zirconium in the form and sintering of the powder. of a powder or of porous chunks. The best We must assume that grains of metal obtained samples we1;~ fo~nd upon analysis to have a by the reduction of the chlorides of the three metals purity of 100 per cent. The separate grains showed in question are covered with a film of oxideand (or) a certain ,ductility (mechanical deformability). It nitride. In the sintering of rods pressed from the was, however, very remarkable that when a rod was powders, the oxygen and nitrogen dissolve in the pressed from the pure metal powder, according to metal in the, case of zirconium and titanium, and the method used in working tungsten, and when the metal thus becomes brittle, while in the case this rod was sintered at a high temperature in a of thorium,' the oxide and nitride remain present as very high vacuum, '1;he, rod showed practically such in the rod. It is true that in the sintering of .~.. no sign of ductility. Working to sheet or wire by pressed thor~um rods a change in the structure also rolling, hammering and drawing was' therefore app,ears, but this change in the structure has a .impossible. Since however chemical analysis indi- favourable influence on the mechanical properties . cated a very highdegreeof purity, as was mentioned It consists namely in the fact that the oxide or above, it was for many years generally assumed nitride is collected into separate grains, which are, that zirconium is a very brittle metal and must be then ,present after sintering surrounded by a considered to belong to the so-called half-metals. coherent basic mass of ductile metal. This makes However, thirteen years ago an entirely new it understandable that much greater quantities of method of preparing zirconium in the form of oxygen and nitrogen are required to render thorium rods was discovered in this lahoratory 2), namely brittle than in the case of zirconium and titanium. by thermal decomposition of zirconium tetraiodide The grains of zirconium and titanium obtained on a glowing filainent (see below) and the rods as by reduction usually contain small traces of oxygen ! and nitrogen already dissolved, so that each in- I prepared were fouhd to possess a high degree of I ductility. The d~f~r~~bility is' so great that it dividual grain has a ductility considerably lower is now even possible', to cold-work rods of 7 mm than that of the absolutely pure metals, but con- \ i thickness to very thin .wire and sheet. The crystal siderahly higher than that of the pressed and sin- i structure of duétil~ zirconium was found to be the tered rods. ! same as that of brittle Zirconium, so that it was j' 'I not a question of, two allotropie modifications. Preparation of ductile zirconium in rod form It has therefore' 1;0 be assumed that the brittle Ductile zirconium is prepared by thermal deeoni- zirconium contained one or more impurities of position of zirconium tetraiodide. At a temperature of which chemical analysis gave no indication. 431°C the vapour pressure of this compound reaches Only in recents years has it become possible to a value o~ one atmosphere. If the gaseous iodide arrive at a, reasorîably satisfactory explanation comes into contact with a surface whose temperature of this phenomenon. .I't has been found that zir- is higher than 1l00°C, it decomposes partially into conium is able to ,tàke' up quantities of oxygen zirconium and iodine. When the process is carried and nitrogen in solid solution. When these elements out in practice the decomposition takes 'place are present in the dissólved state they have a very on a thin wire which is heated by the passage' of current to a, suitable temperature, for instance I ,2) J. H. de Boer und J. D. Fast, Über die Darstellung der 1300°C. Tlie reaction takes place in an apparatus reinen Metalle der Titangruppe durch thermische Zerset- I" zung ihrer Jodidè, I. Zirkonium, Z. anorg. allg, Chem. 153, made of pyrex' glass into which thick tungsten I' 1, 1926. terminals are fused. The required amount of crude I: - '_ '.,~ ',."

. DECEMBER 1938 ZIRÇONIUM . 347

zirconium is introduced into this apparatus in the For practical uses zirconium is made in rods form of porous chunks (obtained by the reduction about 7 mm thick which weigh about 200 gr. of zirconium tetrachloride or of sodium zirconium The oven .temperature is kept at 250° to 350°C fluoride with sodium) together with a relatively, during the preparation. A tungsten wire 40 microns small amount of iodine. The apparatus is then thick is often used as core wire. The tungsten evacuated to a pressure of about lp-6mm. During content of the zirconium rods then amounts' to '. .evacuation the crude' zirco~ium is heated to 0.01 per cent. For applications.In which this tung-, drive off adsorbed gases, while the iodine is cooled. sten content would be detrimental a core wire of After. outgassing. the apparatus is sealed off and zirconium itself is used. heated in an electric oven. At a temperature of about 100°C (depending on the nature of the 200 -2SS72 crude zirconium). the iodine combines with part ~ - .S: --...... of the zirconium to give ZrI4, and the substance .... evaporates and is decomposed on the wire into ~ r, zirconium and iodine when the temperature of 150 s the wire is sufficiently high. In this way crystals \ of zirconium begin to grow on the wire. f The vapour pressure of the iodide at 100°C is so small that it may be neglected, so that upon heating the core wire to 1DO I 1300°C, for instance, no appreciable growth takes place. At an oven temperature of 150°C the vapour pressure is already so high (about 0.001 mm Hg) that zirconium slowly begins to be deposited on the glowing wire, With increasing temperature of the oven the speed of growth increases and 50 I1 reaches a high value at 250°C. The vapour pressure of the tetraiodide then amounts to about 0.5 mm Hg. From about

250°C, however, the ZrI4 pressure rises only slightly, because of the fact that above this temperature the tetraiodide begins o -r-femperature in oe 0 to react with the excess crude zirconium giving the tri-iodide / " 100° 200 (ZrIs), which has no appreciable vapour pressure in this temperature range. At still higher temperatures the di-iodide . Fig. L Speed of formation of zirconium obtained by decompo- sition of zirconium tetraiodide, as a function of the oven (ZI2) is also formed, a compound which also has no appre- temperature. The weight obtained at the corresponding oven ciable vapour pressure at the temperatures attainable in pyrex temperature and at a wire temperature of 1300°C after a glass. Above 310°C the tri-iodide decomposes according to time of growth of 30 hours is plotted as ordinate. the equation . ~ 2 z-r, «- ZrI2 + ZrI4' As stated above only a relatively small amount of iodine is used in. the preparation of ductile while above 430°C the di-iodide decomposes according to the. equation: zirconium. It serves only as means of transporring ~ , the metal from the crude, zirconium to the growing 2 ZrI2 ~ Zr + ZrI4 rod. It is continually freed at the growing rod, At higher temperatures therefore the p~essure of ZrI is 1 combines again with part of the excess crude metal' determined by these equilibria. - 4 The result of these phenomena is that the speed of growth and "loaded with zirconium" it enters the gas does not increase appreciably above 250°C. and that above atmosphere again. Care must. be taken that the 350°C it finally begins to decrease again. crude zirconium does not contain any great amou,nt Fig. 1 gives a graphic representation of the speed of substances like iron, aluminium and silicon, of formation of ductile zirconium as a function since these substances .also form relatively volatile of the oven .temperature. The temperature is plot- iodides and their presence .might lead to contam- ted as abscissa, and as ordinate the weight of the ination of the rods. The oxygen and nitrogen zh:conium rods which were obtained after ~ time content of the crude -zirconium on the other hand of growth of 30 hours. In each separate experiment is not harmful, since these substances remain the, oven temperature was kept constant during behind in the crude metal. this time of growth. All the apparatus which were The rods which are prepared according to' the used in these experiments were filled with 250 g. of process described have a completely compact struc- crude zirconium from the saine prepa;ration (ob- ture. The specific weight (6.54) determined, with ' tained by. reduction of the chloride) and with 12 g. the aid of a hydrostatic balance, agrees accurately , of iodine .. The wire temperature was the' same, with. the specific weight deduced theoretically fróm~ 1300°C, in all the experiments.. .. röntgenographic data. The surface of the rods is 348 PHILIPS TECHNICAL REVIEW VoL 3, No. 12 composed of well formed crystal planes. Fig. 2 of this phenomenon is that even with careful shows a photograph of zirconium rods 7 mm thick. lubrication particles are torn loose from the sur- face due to the adhesion between the metal and the surface of the die. These particles remain stuck to the die and are the cause not only of the seizing Lut also of the continuous r'ise in the required drawing force to a value which exceeds the tensile strength 3). If the wire is hot drawn (temperature about SOO°C)an oxide film is formed on the wire which is indeed very hard, but which makes draw- ing possible by preventing seizing. Since we are concerned with a surface phenomenon we can sur- round the wire with a metal covering instead of the oxide covering. The entirely cold hammering and drawing of rods about 7 mm thick to wire Fig. 2. Zirconium rods 7 mm thick. of 30 microns for instance is possible on a techni- cal scale only when a covering of iron, nickel or Mechanical working of zirconium copper is given to the rods. This covering is then We have already mentioned that the zirconium removed chemically (by solution in acid) after rods prepared by decomposition of the iodide are the working. Zirconium wire prepared according so ductile that they may be deformed into thin to this method is entirely free of oxygen and ni- wire and sheet without it being necessary to heat trogen, while hot drawn zirconium wire always

Fig. 3. Zirconium sheet 20 microns thick, obtained by cold rolling of the material from the thickness of 0.6 mm. them. In the preparation of zirconium wire the contains very small quantrties of these elements. rods are first swaged with swagging machines to The rolling of the rods to sheet can also be done wire of about 0.5 mm thickness. This treatment cold. Usually, however, the rods are hot rolled to can be carried out entirely without heating. a thickness of 1 to 0.5 mm. The surface is then From the diameter of 0.5 mm the wire is drawn cleaned and the rest of the rolling is done cold. through diamond dies to smaller diameters. If It was found possible in this way to make strips an attempt is made to carry out this process 20 microns thick, 30 cm wide and I to 2 ID long. also entirely cold, great difficulties arise. A grat- 3) It might be expected that the cause of the difficulties in ing noise is heard during the drawing, while the drawing lays in a mechanical hardening occurring during wire repeatedly breaks. If the wire is examined the process. This was, however, found not to he the case, since the same difficulties are also encountered when a after drawing under a microscope, grooves are seen zirconium wire is drawn immediately after being annealed in the surface along the whole length, The cause in a high vacuum. DECEMBER 1938 ZIRCONIUM 349

If the rolling is done on polished rollers, fine spe- Chemical properties of zirconium cular sheet is obtained. Fig. 3 shows a photograph The most characteristic chemical property of of zirconium sheet 20 microns thick which was rolled metallic zirconium is its great affinity for various from the thickness of 0.6 mm. other elements. Finely-divided zirconium powder It has been found possible to make cold rolled shell ignites when heated above about 200°C in the bodies of zirconium by the deep stamping process. air, and burns with a blinding white light to No partienlar difficulties arise during drawing give zirconium oxide (Zr02). When zirconium is when the zirconium sheet which serves as raw heated in nitrogen a very stable nitride is formed material has been rolled in the correct way. Fig. 4 having the formula ZrN. With hydrogen the com-

pound ZrH2 is formed, which however decomposes again into its elements at relatively low tempera- tures (above 600°C). Table I gives the heats of formation of the three compounds mentioned.

Table I

Heat of formation Compound k. cals. per mol.

258 82 Fig. 4. Shell bodies made of zirconium sheet by application of a deep drawing process. 40.5

shows a photograph of three different types of If the metal is not allowed to react with an excess drawn products made of zirconium. Fig. 5 shows of gas, but with quantities in a closed space, it the various stages in making such an article. The is found that large quantities of gas can be taken processes were carried out cold. Just as in the draw- up without a new phase being formed. In the ing of wire it was found that barc metal is very reaction with nitrogen the cubic phase of the ni- difficult to draw, since seizing takes place not tride is formed in addition to the hexagonal phase only with cold-rolled sheet but also with sheet of the metal only after the nitrogen content has annealed in a high vacuum. If, however, the mate- exceeded about 20 atom per cent. The first 20 per rial is heated previous to the stamping for a very cent are thus taken up in solid solution. For oxygen short time to 650°C in air, a film of oxide is formed the limit of solid solubility even lies at about 40 which facilitates drawing. The best results are atom per cent, while for hydrogen it is already obtained upon heating the metal for only a few reached at 5 atom per cent. seconds in the colourless flame of a Bunsen burner, In the investigation of the systems zirconium- and then quenching it in water. The metal is then oxygen, zirconium-nitrogen and zirconium-hy- in the form of very fine crystals, since the cold- drogen it is best to begin with ductile zirconium worked sheet recrystallizes upon heating, while rods or bare wires which are made by cold working due to the extremely short time of heating there of the rods. In the case of the systems with oxygen is no appreciable growing together of the very or nitrogen the wires must be heated for some fine crystals formed. This fine grained structure time at a high temperature after taking up the IS an advantage for the stamping process. -'different amounts of gas in order to reach complete homogeneity. The simplest method of heating is by the passage of a current, When the content of gas is not too high the solutions formed cannot be distinguished in appearance from the pure metal. When unworked rods are used the solutions still exhibit the same crystal planes and the same high metallic lustre. Fig. 6 shows two pieces of the same zirconium rod (thickness about 3 mm). The shorter piece has been kept in the original state. and is therefore quite pure and very ductile. The Fig. 5. Different stages in the drawing of shell bodies. The manufacture of the cap (right) from the zirconium sheet longer contains 13 atom per cent of nitrogen and (left) requires three processes which can be carried out cold. is very brittle. Superficially, however, the two 350 PHILIPS TECHNICAL REVIEW Vol. 3, No. 12 pieces cannot be distinguished from each other. Table ill If wires containing several atom per cent of Composition in Temperature coefficient oxygen, nitrogen or hydrogen are heated in a atom per cent of the electrical resistance high vacuum, it is found that the hydrogen can be driven out quite easily. (The removal of the last 100% Zr 0.00445 traces, however, takes a long time even at a high 98% Zr + 2% 0 0.00264 temperature). Oxygen and nitrogen, however, 98% Zr + 2% N 0.00261 cannot be driven out. Upon increase of temperature the gas pressure remains zero, and at t.emperaturee For oxygen and nitrogen therefore the same near the melting point metal atoms begin to evap- lowering of the temperature coefficient was found. orate, while the oxygen or nitrogen remains in the wire. These gases are therefore also bound by Applications of metallic zirconium strong chemical forces when they are present in the metal in the dissolved state. Mixed with one or more oxidation agents finely- The specific weight of zirconium is increased by divided zirconium is used as a flash light powder. the introduction of oxygen or nitrogen atoms into Several other applications of zirconium are also the metal lattice. From this it may be concluded based upon the property of this element of reacting that- these atoms lie in the open spaces of the metal easily and rapidly with varrous gases, with the

Fig. 6. Two pieces of the same zirconium rod. The shorter piece is chemically pure and very ductile. The longer piece was heated in an atmosphere of nitrogen, it contains 13 atom per cent of nitrogen and is very brittle. In appearance the two pieces cannot be distin- guished from each other.

lattice. They are probably quite irregularly distrib- result that the gas pressure falls to zero if there uted throughout the lattice. In table I I are the is an excess of metal. For this reason it is used as a results of several determinations of volume and getter in discharge tubes, especially in transmitting specific weight (ofrods 3 mm thick) before and after valves. The zirconium must be placed in the tube the taking up of known amounts of nitrogen or at such a spot that it assumes the required temper- oxygen. ature at the moment when its function of com- Table II bining with the gas must be exercised, or it must be at such an electric potential that the gas ions Cornposition in I yolume Specific formed strike the zirconium and are then held atom per cent _j__ Increase weight bound. The favourable action of the zirconium is manifested in two ways. In the first place the 100% Zr 6.540 times of evacuation during manufacture are con- 98 % Zr + 2 %0 0.15% 6.555 98 % Zr + 2 %N 0.23% 6.546 siderably shortened because the zirconium also 90 % Zr + 10 %0 1.05% 6.600 acts as a pump. In the second place the small 87 % Zr + 13 %N 1.72% 6.577 amounts of gas which may be freed when the dis- 78.5% Zr 21.5% 0 1.51% 6.755 + charge tube is in use are immediately bound. The t.emperature which the zirconium must assume The temperature coefficient of the electrical depends upon the nature of the gas to he absorbed resistance decreases upon the solution of oxygen and in addition on the nature of the zirconium or nitrogen in the metal. The values given in itself. If ductile (and therefore compact) zirconium table I I I were measured. is used, and if the freeing of gases like oxygen, DECEMBER 1938 ZIRCONtUM: 351

nitrogen, carbon monoxide and carbon dioxide the compact metal is remarkably inert at ordinary must be taken into. account, it is desirable to have . temperatures. This inertness must be ascribed to a the temperature as high as possible (1400°C for film of oxide which is formed on the surface and example). Hydrogen, however, is given off again cannot dissolve in the metal at ordinary tempera- for the most. part at this high température (an tures. This oxide film, which is quite invisible, equilibrium pressure is established). The optimum protects the metal to such an extent, that zir- temperature .for taking up hydrogen. is 300 to conium keeps its lustre under all circumstances· 400°C, in which temperature range a solid solution and is not attacked by acids, bases or other corrosive . of hydrogen in zirconium has no appreciable gas agents (except hydrogen fluoride). Thanks to this pressure, while the absorption velocity is already property zirconium i~ a very suitable metal for very high. Therefore if provision must, be made use in chemical processes where other metals not only against the freeing of oxygen, etc., bu"t would quickly corrode. For example in the etching also against the freeing of hydrogen and water or cleaning of metal articles which must be immersed vapour; it is desirable to have present in the tube, successively in alkali and acid. baths, zirconium' in addit~on to the zirconium heated to a high temper- wire. is used to hang up the articles as it does not ature, a portion of zirconium with a temperature . corrode and does not contaminate the baths, while of 300 to .400°C. It is often possible to combine in· electrolysis it mayalso serve as supply line these two requirements by introducing the, zir- for the. current. . conium at such a spot that different parts of the metal are raised to different temperatures so that The refractory compounds of zirconium the temperature range extends from 300° to for The oxide, the nitride and the carbide ofzirconium instance 1400"C., melt at considerably higher temperatures than the In many cases metallic zirconium can also be element .itself, as. may be seen from table IV. used in the form of a fine powder which may for instance be applied to the anode of a transmitting Table IV valve. The finely-divided zirconium reacts with the different gases at a considerably lower tempe- Melting point, °C rature than compact zirconium. . ____ s_uh_st_u_n_ce :I_ . The fact that zirconium may contain large Zirconium, Zr 1860°

quantities of oxygen and nitrogen in solid solution Zirconium oxide, Zr02 2700° is a great advantage when the metal is used as a . Zirconium nitride, ZrN 2980° Zirconium carbide, ZrC. getter. At sufficiently high temperature there is 3530° rio layer of oxide or nitride formed which hinders further taking up of gas, but the oxygen or nitrogen The oxide dissolves poorly in glass and is there- diffuses to the inside and the metal surface always fore used to render glaze and' enamel opaque. remains clean. It is also' used in the manufacture of very reo' A quite different property of zirconium which i~ fractory articles in the form of crucibles, tubes, important for its use in transmitting valves is the boats, fire bricks, insulators, etc. These articles small value of its secondary emission. By making can be used at temperatures considerably higher the control grid of a transmitting valve of zirco- than 2000°C. In addition to its high melting point nium, or covering it with zirconium, the secondary the small coefficient of expansion ~nd the chemical emission of this grid can be diminished to a negligi- stability are important in this connection. If the

ble value. The zirconium used for this purpose also products are made of pure Zr02, they show a tendency' retains its property of combining with gases. By to develop cracks upon heating due to the transi- covering a grid of molybdenum or tungsten with tion at 1000°C from a monoclinic to a tetragonal zirconium oxide, the 'secondary emission of the grid modification, which is accompanied by a contrac- mayalso be sufficiently reduced. For further par- tion of 7 per cent in volume. If the zirconium oxide ticulars about this action of zirconium the reader is is mixed with a few per cent of magnesium oxide, referred to an article by H. G. Boumeester in a homogeneous cubic solid solution is formed this periodical (Philips techno Rev: 2, 115. 1937). upon heating above 1700°C, which is subject to no Although zirconium, as may be seen from the .changes in modification, and which therefore can be above, combines readily with all kinds of elements used for the manufacture of very highly resist- or reacts with all kinds of compounds at higher ant refractory articles. Instead of magnesium temperature (depending on the form of the metal), oxide other substances mayalso be used, calcium 352 - PHILIPS TECHNICAL REVIEW Vol. 3, No. 12 oxide or thorium oxide for example. Beryllium addition to a high melting point it has metallic . oxide with zirconium oxide forms an eutectic sys- . electrical conductivity so that it could be used fór tem without the formation of solid solutions. filaments of electric lamps. Upon the invention With the help of this compound it is therefore of the tungsten lamp, however, the compound impossible to suppress the reversible change in passed out of use. It was discovered later that a modification of Zr02' since we have to do with mixture of 4 parts of tantalum carbide and I' part mechanical mixtures in the whole system. How- of zirconium carbide forms a material "":"iththe ever, when BeO is used in combination with highest melting point of any known' substance. MgO (for instance 2 per cent MgO + 0.5 per cent The two carbides form an uninterrupted ~eries BeO), it seems to give a more satisfactory ~esult of solid solutions with a maximum melting point than MgO alone. a.t the composition mentioned. The melting point Because of the increasingly high requirements of this. mixture is almost 4000°C (At the saine being made of many ceramic products with resp- high temperature a- mixture of 4 parts tantalum ect to their resistance to heat, it is probable that carbide and 1 part hafnium carbide melts). Zirco- many _ possibilities for the application of Zr02 nium carbide is also one of the metallic carbides will still be found. with a very great hardness, for which compounds Zirconium carbide, ZrC, is one of the few sub- there has been great interest in recent years in stances which have a higher melting point than connection with the important applications which tungsten. In the history of the electric light bulb the carbides of tungsten, tantalum and titanium industry this compound has played a part. In have-found in the working of metals.

QUANTITATIVE CONSIDERATIONS OF ELECTRIC WELDING

by J. ter BERG.

In ordinary welding with a rod provided with an oxidizing coating, it could be shown that 82 per cent of the fused metal core is used in the weld, 12.5 per cent is lost by spattering and about 4.5 per cent is lost due to oxidation by the coating, while evaporation plays only a minor part. If instead of the ordinary atmosphere, nitrogen or compressed air is used as gas environment, little change is observed: more or less iron is then also oxidized by the flux. When a mixture of hydrogen and nitrogen is used, this reaction does not take . place at all. It has been shown clearly by these experiments, that in the attempt to in- crease efficiency care must he taken not only to have a low loss from spattering, but also that the coating of the rod should possesslittle or no oxidizing power, or that this coating should give off gases during welding which form a reducing atmosphere around the are.

Introduction efficiency of a welding rod we mean the ratio be- It is a well known fact that in electric welding tween the weight of the metal found in the weld ;;t certain percentage of the metal used is lost in and the weight" of the core metal consumed in some way or other. What may be the causes of this making the weld. loss of material, and to what extent each cause The rods specially made for our experiments con- Il].ay contribute to the total loss, has not, a's far tained a malleable iron core with a diameter of 5 mm, as is known to the author, been dealt with in the while the coating consisted of ferrous and ferric literature in the case of coated welding rods. The oxide, quartz and aluminium silicate with potas- experiments and calculations described below sium water glass as a binder. The average ef- were carried out for the purpose of procuring data ficiency of several tests carried out under ordinary on this problem. conditions was low and amounts for these rods The term efficiency is used here, and b-r the about 82 per cent.