Platinum Metals Rev., 1976, 20, (3), 79
The Affinity of the Platinum Metals for Refractory Oxides
Reactions Under Reducing Conditions
The presence of a platinum metal enhances the ability of hydrogen, carbon, carbon monoxide and organic vapours to reduce the more stable refractory oxides. Such reactions, which may start at temperatures as low as 600°C, occur because of the high affinity of the platinum metals for the metal of the refractory oxide and result in the formation of intermetallic compounds or stable solid solutions. This account, based on experimental studies which yielded useful information on the thermodynamic behaviour of platinum metal alloys, presents results which should assist in preventing damage to industrial platinum apparatus. It also suggests that metals could be brazed to refractory oxides using a platinum containing brazing alloy under suitable reducing conditions.
Practical experience has shown that even platinum and palladium are not completely resistant to oxidation, and that they can, under certain conditions, also react with base metal oxides at high temperatures (1–5). When heated in air some of the platinum metals, initially develop an oxide skin which disappears by dissociation or evaporation above a critical temperature. Others, such as platinum itself, retain a clean surface, but suffer a more or less rapid decrease in weight because of volatile oxide formation.
The more stable and refractory base metal oxides can also react with platinum in an argon atmosphere, in vacuum, or even in air, the reactions being particularly severe with zirconia, thoria and alumina. Under such conditions magnesia can be considered the only oxide which is truly inert with respect to platinum (3, 4). These reactions are greatly accelerated by the presence of strong reducing agents, Bronger and his collaborators (6, 7, 8) having shown that many inter-metallic compounds can be synthesised by reacting the platinum metals with oxides in a hydrogen atmosphere. Similar techniques were also used for the production of palladium-tungsten alloys (9).
Within an industrial environment strongly reducing conditions are sometimes difficult to avoid. The platinum alloy bushings used for glass fibre production provide a good example of the hazards likely to be encountered, as they can be rapidly destroyed by silicon if the glass feed is contaminated by organic residues such as oil or wood. Platinum thermocouples can also react with their refractory insulators if the assembly is accidently brought into direct contact with luminous or other reducing flames.
Following earlier studies of the reactions between platinum and the transition metal carbides (10), the programme described below was undertaken to identify those industrial situations where reducing conditions caused by the presence of carbon monoxide, carbon or the hydrocarbons, could lead to the rapid failure of platinum plant or equipment.
|Si||1450 RuSi||1400 Rh2Si||1100 Pd2Si||−||1500 IrSi||+1100|
|Ti||1750 *||1600 Rh3Ti||1200 Pd3Ti||−||1600 Ir3Ti||1200 Pt3Ti|
|Al||−||1450 RhAl||Pd2Al||−||1600 IrAl||1500 Pt2Al3|
|1300 Pd3Al2||1250 Pt3Al3|
|Cr||1000 Pt3Cr (ordered)|
Reactions in Hydrogen
The principle of these reactions is best shown by the reaction of platinum and alumina in hydrogen (6):
For Equation 1, δH amounts to an oxygen vapour pressure at 1200°C of , for Eqn. 2 of torr making it possible to reduce Al2O3 with H2 in the presence of platinum only. This example shows the decisive importance of the heat of decomposition of the oxides and of the heat of formation of the various intermetallic compounds for these reactions. From their results, Bronger and Klemm (6) concluded that the heats of formation are highest for the platinum compounds and decrease on moving from platinum to palladium and from platinum via iridium to osmium. This makes ruthenium the metal least likely to react with base metal oxides in hydrogen.
A summary of the results of Bronger ’ s hydrogen reduction experiments is given in the Table. In general an increase in the reaction temperature decreases the concentration of the noble metal in the inter-metallic compound, as can be seen for platinum-aluminium alloys. With alumina, silica and titania the AB-phases RuSi, RhAl, PdAl, IrSi and IrAl were obtained, whereas in almost all other reactions only those phases containing more than 50 atomic per cent platinum metals were synthesised. The explanation may be that Al2O3, TiO2 and SiO2 have low heats of formation (−135.5, −112.5 and −105.0 kcals per mol of O2).
From these results one must conclude that refractories based on lime, magnesia or the rare earths should be preferred to aluminous materials for use in contact with industrial platinum equipment where reducing conditions are likely to be encountered. The behaviour of thoria and zirconia under such conditions has, unfortunately, not yet been investigated. Since it is known that the affinity of platinum for thorium and zirconium is stronger than its desire for aluminium, (3, 11) it seems fair to assume that zirconia and thoria will also react readily with platinum in hydrogen.
Reactions in Carbon Monoxide
Carbon monoxide is often present in combustion gases and may therefore be of some danger to the platinum metals at the higher temperatures. Unlike hydrogen, carbon monoxide is insoluble in the platinum metals and, since no dilute solid solutions are involved, the reaction conditions are therefore likely to be somewhat simplified. A detailed study was undertaken at this Research Institute to establish the way in which the platinum metals reacted with base metal oxides, in a carbon monoxide atmosphere, at various temperatures, under practical conditions (12, 13).
At the temperature levels investigated, carbon monoxide was found to reduce the stable base metal oxides in contact with platinum in much the same way as hydrogen (12). Silica and zirconia reacted with palladium even below 1000°C, and above 800°C small molten droplets provided evidence of liquid phase formation, as shown in Figure 1. Figure 2 shows how palladium vapour, formed above 1200°C, has reacted with the aluminous crucible material. Unusual whiskers were sometimes formed as a result of these reactions, those shown in Figure 3 being produced when palladium dissociated magnesia in the presence of carbon monoxide. Adorning the tip of each of these whiskers is a small ball having a duplex microstructure.
As in hydrogen, lime and magnesia reacted with the platinum metals under carbon monoxide less readily than silica and zirconia. The rare earth, Dy2O3 displayed remarkable resistance to decomposition, little evidence of reduction being seen even after 120 hours at 1400°C. When mixed with vanadium pentoxide, the platinum metals reacted quantitatively to produce f.c.c. solid solutions containing up to 40 atomic per cent vanadium.
Similarly quantitative reductions resulted from Pd–Cr2O3 and Pd–TiO2 mixtures containing an excess of palladium. In other instances, for example, when palladium reacted with zirconium oxide, X-ray analysis indicated the presence of phases not delineated on the binary equilibrium diagram.
The diffraction patterns obtained contained many lines which were difficult to index. This complexity might be attributed to the presence of ternary carbides, or oxygen or carbon stabilised pseudo-binary compounds. The more or less continuous transition phases, such as those observed by Bronger et al . (6), which bridge the gap between one structure and another in the series Pt–Pt7Mg–Pt3Mg might also be involved.
In carbon monoxide, as in hydrogen, platinum reacted strongly with the base metal oxides, being, in this respect at least as active as palladium. The hexagonal metals such as ruthenium were far more stable.
The reaction sintered pellets which resulted from these experiments sometimes displayed a fair amount of ductility even at high oxide concentrations, and it seems possible that technically viable methods of cermet production could be based on such reductions.
Reactions in the Presence of Carbon
These reactions, and especially their thermodynamics, have also been studied at this Research Institute, and some of the results obtained will be mentioned below.
When carbon is mixed with, for example, palladium and silica, the following reactions must be considered:
For δG5 we find
In addition the formation of carbides and solid solutions may add extra terms to Equation 6.
As has been mentioned earlier one of the decisive factors of the reaction is the free enthalpy of formation of the metallic phase (δG4). If this is high the CO pressure above the cermet pellet will be increased, or at a constant CO pressure the reaction will occur at lower temperatures. We therefore studied the reactions by CO equilibrium pressure measurements, weight loss determinations and X-ray methods.
Chromic Oxide and Carbon
Although chromic oxide should theoretically be reduced by carbon at 600°C at pressures of 10−5 torr, practical tests showed no significant reaction until 800°C was reached. The presence of palladium strongly enhanced this reaction, a palladium-chromium solid solution being formed. At 900°C the carbide Cr7C3 was also detected.
Figure 4 shows how the free energy of formation of the solid solution (corresponding to δG4 in Equation 6, the value of δG3 for Cr2O3+3C→2Cr+3CO being taken from the literature) varies with the chromium concentration of the alloy which is formed. Platinum reacted with chromium less readily than palladium. Compared with palladium, rhodium had little influence on the reduction of chromic oxide with carbon. Ruthenium, however was more active. The reaction began at 700°C, and after 200 hours at 900°C the ruthenium solid solution which was formed contained 18 atomic per cent chromium.
Silica and Carbon in Vacuum
Both palladium and platinum have a similar influence on this reaction which begins at about 700°C. The palladium and platinum phases formed are, respectively, Pd5Si, Pd4Si, Pd3Si; and Pt3Si, Pt7Si3, Pt2Si.
The formation of low melting point eutectics may complicate the reduction processes.
As shown in Figure 5, the partial molar free energy of formation of the PdxSiy phases is not greatly influenced by the amount of silica reduced to silicon. The value of 30 kcals at 1000K might be attributable to the formation of Pd3Si. With ruthenium a weak reaction starts at 900°C and at 1000°C small amounts of Ru2Si can be detected by X-ray diffraction methods. Rhodium has no influence upon the reduction of silica with carbon up to 1000°C.
Zirconia and Carbon in Vacuum
Zirconia is readily reduced by carbon in the presence of platinum and palladium. As can be seen from Figure 5, the driving force for the reaction may be the high free energy of formation of Pd3Zr. Over the temperature range investigated, the reaction ceased when the zirconium concentration in the reacted alloy exceeded 25 atomic per cent. Figure 6 shows how the surface reactions between the larger particles of zirconia, and the surrounding carbon and palladium powders, resulted in the formation of a reaction zone of Pd3Zr which was deficient in carbon.
Magnesia and Carbon in Vacuum
Although the reactions with magnesia have not yet been fully clarified, it seems evident that the high weight losses which occur at 900°C can not be entirely attributed to the reduction of magnesia with carbon. Our experimental work shows that this occurs only at temperatures of 1100°C and above. Similarly, a mixture of palladium and magnesia shows little weight loss when heated, in the absence of carbon, up to 1100°C. The phases formed when magnesia is heated with both carbon and a platinum metal at 900°C and above are Pd3Mg, Pt7Mg and Pt3Mg.
Above 1000°C, the weight losses which occur are unexpectedly high, and only traces of these intermetallic compounds can be detected. Similar effects are observed when magnesia is heated with carbon and ruthenium or rhodium. These effects might involve the formation of some volatile ternary phase, or the catalytic decomposition of the oxide which allows the magnesium to escape as vapour.
Reactions of Base Metal Oxides with Organic Vapours
Preliminary tests indicate that palladium-silica mixtures react readily at 900°C with propane and methyl alcohol. With platinum-silica mixtures the reaction proceeds rather slowly. A very low reaction rate is observed when mixtures of platinum or palladium are heated with zirconia or alumina in the above mentioned atmosphere.
The abilities of hydrogen, carbon, carbon monoxide and organic vapours to reduce the more stable refractory oxides are strongly enhanced by the presence of the platinum metals. Palladium and platinum tend, in general, to be more effective in promoting such reactions than ruthenium or rhodium.
Most of the reactions start in the temperature range 700 to 800°C, although reduction temperatures as low as 600°C have been observed. The driving force for these reactions is the high affinity of the platinum metals for the metal of the refractory oxide. This affinity results in the formation of intermetallic compounds or stable solid solutions, whose partial molar free energies of formation might be as high as −100 kcal per mol at 1000°C (the Pd/Pt-Mg systems).
The results of this experimental study are of value not only in helping to prevent damage to industrial platinum apparatus. They also suggest that metals could be brazed to refractory oxides with the aid of a platinum-containing brazing alloy, under suitably reducing conditions. Useful information on the thermodynamic behaviour of platinum metal alloys has also been obtained.
- 1A. R. Powell, Platinum Metals Rev., 1958, 2, ( 3 ), 95
- 2C. B. Alcock, Platinum Metals Rev., 1961, 5, ( 4 ), 134
- 3A. S. Darling,, G. L. Selman and R. Rushforth, Platinum Metals Rev., 1970, 14, ( 2 ), 54 and Platinum Metals Rev., 1971, 15, (1), 13
- 4W. Gorski, Z. Anorg. Allgem. Chem., 1968, 358, ( 3 / 4 ), 163
- 5E. Raub and W. Plate, Z. Metallkunde, 1957, 48, ( 10 ), 529
- 6W. Bronger and W. Klemm, Anorg. Allgem. Chem., 1962, 319, ( 1 / 2 ), 58
- 7W. Bronger, J. Less-Common Metals, 1967, 12, ( 1 ), 63
- 8H. Schulz,, K. Ritapal,, W. Bronger and W. Klemm, Z. Anorg. Allgem. Chem., 1968, 357, ( 4 / 6 ), 299
- 9H. R. Khan and Ch. J. Raub, J. Less-Common Metals, 1971, 25, ( 4 ), 441
- 10E. Raub and G. Falkenburg, Metall, 1973, 27, 669 and Z. Metallkunde, 1964, 55, ( 4 ), 190
- 11L. Brewer, Acta Metall., 1967, 15, ( 3 ), 553
- 12E.R Röschel and Ch. J. Raub to be published
- 13D. Ott and Ch. J. Raub to be published