Platinum and the Refractory Oxides I - COMPATIBILITY AND DECOMPOSITION PROCESSES AT HIGH TEMPERATURES By A. S. Darling, G. L. Selman and R. Rushforth Research Laboratories, Johnson Matthey & Co Limited

countered and although such contamination Under normal atmospheric conditions problems have been studied by several platinum is exceedingly inert with investigators (3, 4) few general conclusions respect to the more refractory oxides. have been arrived at. Most of the reactions Severe reactions can occur, however, observed were caused because either the when the oxidising potential of the platinum or the refractory oxide had been surrounding atmosphere is reduced accidentally contaminated by materials which below a critical level. Alumina, were not intentional constituents of the zirconia and thoria dissociate, oxygen experimental environment, and the import- is evolved and platinum extracts the ance of vapour phase reactions which transmit meial from the refractory to form diluta volatile suboxides and sulphides over dis- alloys and low melting point phases. tances of several inches is now well Magnesia is very resistant to this type appreciated (5). of decomposition, and appears to This article describes some of the more be an excellent refractory for use in interesting results obtained during a series of contact with platinum, This article systematic experiments in which every effort describes in general terms the reactions was made to avoid the complications of platinum with alumina, zirconia, of spurious contamination. In view of the magnesia and thoria at temperatures up successful attainment of this objective, a brief to 17OO"C. Subsequent articles will description of the apparatus employed may discuss the efects of geometry and be of value. environment on such reactions, and The basic test cell employed ensured that &ally their practical implications so the platinum metal or alloy and the refractory far as the user of platinum is concerned. oxide immediately surrounding it were the only two components at the experimental temperature. Platinum and its alloys can usually be The body of the cell shown in Fig. I is safely heated for long periods in contact with made of Pyrex glass and remains fairly cool the more refractory oxides without serious throughout the experiment. The wire risk of contamination. The embrittlement specimens are conveniently tested in the form which sometimes occurs when platinum wires of thermocouples which are directly heated by are heated in contact with alumina insulators mains frequency alternating current, the under reducing conditions at temperatures thermocouple hot junction being immersed above 1200°C is caused by siliceous impurities below the surface of a fairly loose bed of the in the refractory (I, 2) and platinum has been powdered refractory under test. By using a successfully melted in pure alumina crucibles polarised relay the heating current can, in air, and in vacuum, for many years. when required, be momentarily interrupted Inexplicable failures of high temperature every half cycle so that the hot junction apparatus are, however, occasionally en- temperature can be monitored with a potentio-

Platinum Metals Rev., 1970, 14, (2), 54-60 54 meter. The circuits employcd for self-heated thermocouples of this type have been TO ICE described by Welch (6, 8) and others (7). FLASK The impurities present in the oxide refractories used in this series of tests are given in the table. The platinum and platinum alloy wires conformed to the normal standards of thermocouple purity. Effect of Atmosphere Rhodium-platinum alloys were the first to be systematically tested in this apparatus and the protective atmosphere employed was argon containing approximately 50 p.p.m. of impurities. Under such conditions the standard platinum: 13 per cent rhodium- platinum thermocouple survived 1000 hour test runs at 16oo0C with little change in ISOLATION VALVE output and no significant changes in the composition of either limb were detected after such periods of test. Some aluminium contamination was detected after prolonged testing at 17oo0C,but this was not severe and had little effect on the output of the thermo- Fig. 1 Experimental test cell arrangement for studies of the compatibility of platinum couple. with refractory oxides When molybdenum-platinum thermo- couples were tested in this apparatus, Rapid failure was soon encountered when however, fairly severe surface erosion was tests were made in this purified argon encountered after the wires had been held at atmosphere. The platinum wires tended to 1600°C for a few hours only. As oxygen was fuse after two or three hours at 16oo0C, suspected as being the main cause of this because of severe reactions between platinum effect, some of the test cells were fitted with and the refractory oxide. Large volumes of zirconium filaments, which by electrical oxygen were evolved by the refractory and heating functioned as getters about 2 in. adsorbed by the filament, while the plztinum above the surface of the refractory bed. wires dissolved substantial quantities of

Typical Analyses of Refractory Oxides Used for the Compatibility Tests

Refractory I Impurity Content in p.p.m. Oxide Si Fe A1 Mg R Ag P Hf

A1,0, 55 - 3 2 I n.d. n.d.

ZrO 10 3 2 I n.d. n.d. n.d. ~200 MgO 2 I0 I0 - 2 I n.d. n.d.

Tho, 8 10 n.d. 2 n.d. 10 n.d. n.d.

n.d. - Not detected

Platinum Metals Rev., 1970, 14, (2) 55 Fig. 2 Microstructure of a 1 yo molybdenum- Fig. 3 Microstructure of a pure platinum platinum wire which fused after 2 hours at wire which failed after 15 hours in alumina lGOO°C in aluminu under a gettered argon at 1400'C under gettered argon. x 150 atmosphere. x 250 aluminium thus forming alloys of low melting the wire temperature, appreciably below that point. of the hot junction, was about I 100°C. These reactions were not confined to molybdenum-platinum alloys and systematic Experiments in Vacuum testing showed that they occurred whenever Since these reactions appeared to be platinum or platinum alloys were heated in associated with the removal of oxygen from contact with powdered alumina under a the test cell atmosphere, it seemed reasonable highly purified argon atmosphere. Figure 2 to assume that they would occur in vacuum illustrates the microstructure of a I per cent as well as in pure argon. As platinum can be molybdenum-platinum wire which fused after safely melted in vacuum in alumina crucibles two hours at 1600OC under such conditions. without serious contaminatio7 this conclusion Depending on the area studied the con- seemed a little anomalous but it was decided taminated wire below the refractory contained to investigate any solid state reaction which 0.5 to 1.5 per cent by weight of aluminium. might cccur. The preliminary vacuum tests Failure was less rapid at lower tempera- were made in the standard test cells, which tures, and the combination of increased time were continuously evacuated by an oil and lower temperatures allowed thermocouple diffusion pump to pressures of the order of wires to take up rather more aluminium. IO-~ Torr as measured by a cold cathode Figure 3 illustrates the microstructure of the ionisation gauge. fused zone of a 13 per cent rhodium-platinum Reactions did occur in vacuum although thermocouple which had failed in gettered they were rather less rapid than those ob- argon after 15 hours at 1400°C. The portion served in argon and somewhat different in of wire in this photograph contained 2 to 3 character. Whereas in argon the attack had per cent by weight of aluminium. At 1200°C been fairly uniform along thc submerged reactions still occurred although contamin- length of the rest wire, the severe reactions in ation with aluminium was most severe in vacuum conditions were confined to those portions close to the surface of the alumina portions of the wire immediately below the bed. Figure 4 shows the extent of the refractory surface. This effect is illustrated in contamination observed in this region where Fig. 5 which shows the reacted region of a

Platinum Metals Rev., 1970, 14, (2) 56 molybdenum-platinum thermocouple wire will be of the order of IO-~~atmospheres. It after contact with alumina powder for 100 was difficult to believe that such low oxygen hours at 1600OC in vacuum. Depending upon pressures had been achieved in simple the area analysed, this reacted zone in the experimental apparatus of the above type. I per cent molybdenum-platinum limb con- A more realistic approach involves a pro- tained 0.4 to 1.4 per cent by weight of cess of dissociation which yields both aluminium. aluminium and oxygen in gaseous form. The For some time the reasons for this region of saturated vapour pressure of aluminium at local attack at the refractory surface were not 1600OC is approximately I Torr. Lower fully appreciated and it was felt initially that pressures of aluminium would therefore at the oxidising potential in vacuum might this temperature behave as superheated possibly, although this was thought unlikely, vapours, and on this basis, oxygen and alumin- be rather higher than that obtained in argon. ium vapour pressures of the order of 10-l" Further vacuum tests were made therefore atmospheres would in combination satisfy with the zirconium getter filaments in the conditions necessary for dissociation. If operation immediately above the refractory the aluminium vapour pressure is maintained bed, but the results obtained were invariably continuously at this low level wc must the same as those encountered in the un- conclude that the activity of aluminium gettered vacuum. dissolved in platinum is very low indeed. The reduction of alumina by hydrogen in the Mechanisms of Dissociation presence of platinum has been reported (9). Two extreme conditions can be dis- Considerations of this sort suggested that tinguished when the dissociation of refractory experiments should be made on a refractory oxides is considered. It can, for example, be oxide for the metal of which platinum had a postulated that the dissociation of alumina higher affinity than it had for aluminium. results in the formation of aluminium metal and oxygen. In this instance the metal will Platinum-Zirconia Reactions have unit activity and wc must envisage An important feature of the Engel-Brewer partial pressures of oxygen which at 1600°C theory (10) is the prediction of strong

Fig. 4 Comamination in a 13% rhodium- Fig. 5 Reacted zone of a 1 yo molybdenum- platinum thermocouple wire after heating in platinum thermocouple wire after immersion ahminu under gettered argon for 126 hours in alumina powder for 100 hours at 1600°C at 1200°C. ~150 in vacuum. x 250

Platinum Metals Rev., 1970, 14, (2) 57 Fig. 6 Reactions between a platinum wire Fig. 7 Local surface reaction between R and zirconia powder after heating for platinum thermocouple limb and a large 300 hours at ll00"C in gettered argon. zirconia particle after 300 hours at 1100°C X 180 in gettered argon. x 180 reactions between elements from opposite some distance away from the wire surface as ends of the second and third long periods of shown in Fig. 6. Platinum vapour was the transition series. These transition element obviously moving outwards and reacting with reactions should, it was felt, provide powerful the zirconia particles. In regions of severe driving forces for platinum/refractory oxide attack fairly thick layers of the compound reactions, and it was decided to experiment Pt,Zr were deposited all over thc zirconia with zirconia which, when considered in particle surfaces. isolation, is a more stable oxide than alumina. Zirconia-platinum reactions were, however, Zirconium metal has moreover a vapour not entirely confined to those regions away pressure which is negligible compared to that from the wire surface. Where fairly large of aluminium, and it was hoped that differ- particles of zirconia were definitely in physical ences in the ways in which zirconia and contact with the wire surface severe local alumina reacted with platinum might lead to attack occurred. Figure 7 illustrates such a a better undcrstanding of the processes region where the reaction appears to have involved. progressed by solid state diffusion processes. The experiments did in fact confirm a very high affiniry of platinum for zirconium. When Platinum-Magnesia Reactions platinum and zirconium were heated together Although magnesia is a very stable re- in the argon arc furnace in proportions fractory the vapour pressure of magnesium suitable for the formation of the compound is very much higher than that of aluminium. Pt,Zr the reactions observed were comparable Surprisingly enough very few signs of in intensity to Thermit reductions although reactions between platinum and magnesia not quite so violcnt. have been detected even under conditions of When platinum thermocouples were heated very low oxidising potential where disastrous in contact with zirconia powder under failure would have occurred in alumina after gettered argon severe reactions occurred at only a few hours at temperature. temperatures as low as IIOO"~. As in the After IOO hours of test at 1650°C in alumina experiments the vacuum reactions magnesia, under an atmosphere of gettered although severe were rather slower than those argon, the calibration of a standard platinum in argon. Whereas in alumina most of the 13 per cent rhodium-platinum thermocouple reactions occurred on or within the platinum changed by less than I "C when retested at the wire the zirconia reactions tended to occur gold point.

Platinum Metals Rev., 1970, 14, (2) 58 A similar thermocouple was run for 450 hours at 1650°C in magnesia under purified argon. Recalibration at the palladium point indicated a decrease in output equivalent to 4"C, which at this temperature cannot be considered a scrious deterioration. Under similar conditions of test thermocouples immersed in alumina remain intact for a few hours only. It can be concluded, therefore, that the affinity of platinum for magnesium is ob- viously much lower than its affinity for aluminium. Although aluminium is rapidly taken up from vapours having pressures of Fig. 8 Reaction between a platinum wire the order of 10-l~Torr, few signs of mag- and thoria powder after heating for 84 hours nesium contamination have been detected at 1700°C in vacuum. X 150 even when magnesium vapours of the order of IO-~Torr must have been present in the vicinity of the platinum wire surface. In vacuum and gettered argon, however, severe reactions were encountered. The Platinum-Thoria Reactions character and intensity of the reactions were The extremely high free energy of forma- similar to those which occurred with alumina. tion of thorium oxide and the relatively low Figure 8 illustrates the microstructure of a vapour pressure of thorium metal made this platinum wire after exposure to pure thoria compound a logical refractory for detailed under a continuously pumped vacuum for 84 examination. The preliminary test results, hours at a hot junction temperature of 17ooOC. however, provided a good illustration of rhe The section shown was taken from a portion confusion which can be introduced into of wire close to the surface of the refractory studies of this sort by accidental and un- bed where the most severe reaction had taken expected contamination. Platinum-rhodium place. The thorium concentration in this thermocouples heated in the first test batches area was approximately 3 per cent by weight. of this oxide failed catastrophically in a These reactions occurred at the wire surface commercially pure argon atmosphere after and not on the surface of the oxide particles only a few hours. Careful metallographic as in zirconia. While the reactions between examination and analysis of the fused zone on thoria and platinum are more complex than the platinum wires revealed eutectiferous those encountered with other oxides, there is networks within the microstructure similar no doubt that platinum reacts with pure in appearance to those which occurred when thoria in atmospheres having far higher platinum reacted with pure thoria. In addi- oxygen concentrations than those which would tion to thorium, high concentrations of prevent any reaction with alumina and zir- phosphorus were detected in these networks conia. Experiments with thoria particles and it was subsequently found that inter- covering a wide size range have shown that actions of this type could be promoted by the alloying effects are due to vapour phase phosphorus concentrations in the refractory as reactions although the vapour species involved low as 75 p.p.m. At lower concentrations of have not yet been identified. phosphorus platinum-thoria reactions did not The affinity of platinum for thorium occur in commercially pure argon atmos- appears, therefore, to be higher than its pheres. affinity for aluminium and zirconium, a

Platinum Metals Rev., 1970, 14, (2) 59 finding very much in line with the Brewer Acknowledgements predictions. Acknowledgements are due to the Atomic Energy Authority under whose aegis most of this experimental work was undertaken. We are Conclusions particularly indebted to Mr J. P. Evans and Dr L. E. Russell of the Ceramic Division at Harwell The work described above has shown that for stimulating discussions. at high temperatures platinum will react strongly with refractories such as alumina, References zirconia and thoria when oxygen is effectively I T. Land, 3. Iron Steel Inst., 1947, 155, (2), removed from the surrounding atmosphere. 214 2 L. Reeve and A. Howard, J. Iron Steel Inst., Magnesia is the only refractory so far 1947, 155, (2), 216 examined which resists this type of decom- 3 H. J. Svec, J. Sci. Instrum., 1952, 29, 100 position, which can occur at temperatures as 4 M. Chaussain, Proc. Inst. Br. Foundrymen, 19~1,443A60477 low as 1200°C. 5 J. C. Chaston, R. A. Edwards and F. M. The intensity and rate of these decom- Lever,J. Iron Steel Inst., 1947, 155, (2), 229 6 J. H. Welch,J. Sci. Instrum., 1954, 458 position processes depends upon the geometry 31, 7 E. Aruja, J. H. Welch and W. Gutt, J. Sci. and environment of the reacting system. Instrum., 1959, 36, 16 These subjects and their practical impli- 8 J. H. Welch, British Patent No. 961,019 9 W. Bronger and W. Klemm, Z. Anorg. cations will be discussed in subsequent Allgem. Chem., 1962,319,58 articles. I0 L. Brewer, Acta Metall., 1967, 15, 553

Strength and Ductility of Iridium CAUSES OF FABRICATION DIFFICULTIES Iridium is harder and more difficult to work isms, and although microscopic deformation than any of the other face centred cubic twinning was observed, the stress-strain metals and, although it has a high shear curves showed that this had no significant modulus and low Poisson’s ratio, these fac- effect upon the macroscopic behaviour. tors alone do not account for the troublesome Like all prcvious workers, therefore, the fabrication characteristics of this refractory authors of the present paper conclude that noble metal. the difficulty in fabricating iridium is asso- The anomalous deformation behaviour of ciated with the presence of impurities. The iridium has stimulated a good deal of investi- segregation to grain boundaries of inter- gation, and a recent paper (I) presents the stitial impurities is suggested as a possible results of tensile tests carried out in the reason for the poor ductility of polycrystalline Cavendish Laboratory at Cambridge on single iridium, although the capricious behaviour of crystal test pieces which were pulled at tern- single crystals still remains unexplained. peratures ranging from 20 to 2040’C. Although the practical fabricator of iridium Test pieces were prepared by electro-spark will find little in the paper to assist his machining from zone-refined rods which had activities, possible lines for future investiga- been remelted up to 25 times, in vacuum, to tion can be discerned from the experimental expel all the dissolved and entrapped gases. results. It seems obvious, however, that a The critical resolved shear stress of these significant “breakthrough” has yet to be carefully refined crystals was approximately achieved, seven times higher than that of iridium of A. S. D. comparable purity, reported upon by Haasen (2)in 1965. The rate of work hardening was, References however, much less. The Cambridge investi- I C. A. Brookes, J. H. Greenwood and J. L. gators find it difficult to account for these Routbort, J. Inst. Metals, 1970, 98, 27-31 2 P. Haasen, H. Hieber and B. L. Mordike, differences in terms of deformation mechan- Z. Metallkunde, 1965, 56, 832

Platinum Metals Rev., 1970, 14, (2) 60