Alloy Preparation

Platinum and rhodium sponges, and gold grain, all of the highest available purity, were induction melted, in air, in alumina crucibles before being chill cast into thick copper moulds. The ingots were homogenised at 1250°C for periods up to 64 hours, slowly cooled to 800°C and held at this temperature for 16 hours before quenching into water. After this treatment all the alloys studied could be cold worked into sheet or wire, although it was noted that the ease of working declined rapidly when the gold content exceeded 5 per cent.

The total base metal impurity content of the alloys was of the order of 0.005 per cent. Cold rolled sheet examined after the final stage of fabrication contained typically the following impurities:

Metallographic Studies

Solidus determinations were undertaken by the “bracketing” method, a 20 per cent rhodium-platinum: 40 per cent rhodium-platinum thermocouple being employed because of the high temperatures involved.

Samples of homogenised and rolled sheet were rapidly heated, in intimate contact with this thermocouple, to a temperature 20°C below the desired level, held constant for one hour and then taken up to the quenching temperature at a rate of 20°C per hour. Quenching was accomplished in water from a vertical furnace, and the quenching temperature was increased by increments of 20°C until signs of fusion became apparent in polished and etched microsections.

Attempts were also made to determine by metallographic methods the boundary between the homogeneous and two-phase regions. This was found to be rather an approximate technique in view of the steeply sloping nature of the boundary curve. Alloys containing more than 7 per cent of gold were, moreover, very highly stressed by rapid quenching. Grain boundary precipitation was not completely suppressed and this led to what might almost be described as stress corrosion. In some instances etching removed complete grains from the surface of polished microsections, and this behaviour complicated the search for small quantities of a second phase at the grain boundaries.

Figs. 2, 3 and 4 show typical microstructures of the alloys containing 3, 5 and 13 per cent by weight of gold after quenching from 900°C. At this temperature the phase boundary evidently occurs between 3 and 5 per cent of gold. Duplex structures are shown in Figs. 3 and 4. Globular precipitates of the gold-rich phase can be distinguished round the grain boundaries of these microsections, while the constituent in the body of the grains is too fine for complete resolution. The gold denuded zones round the grain boundaries and the tendency towards sub-grain formation are characteristic features of the structure of alloys heat treated at this temperature level.

The homogeneous structures shown in Figs. 5 and 6 were obtained by quenching from 1200°C. The well-defined cavity shown in Fig. 6 occurred when a complete grain fell out of the surface during the process of etching.

X-ray Diffraction Work

Filings taken from homogenised alloys were sealed in silica quills and heat treated at the same time as the metallographic specimens. After rapid quenching, these filings were used for X-ray diffraction studies with Cu Kα radiation and a Phillips powder camera. The patterns obtained were those of either a homogeneous face centred cubic solid solution or of a mixture of two such phases. The lattice parameters of the phases present in alloys quenched from 850°C are plotted in Fig. 7. These results confirmed the sudden emergence of the gold-rich phase when the gold content exceeded 3 per cent by weight. They agreed closely with values obtained by Raub and Falkenburg (4) at 800°C, but they did not permit an accurate determination of the whole phase boundary.

Resistometric Studies

The boundary between the single- and two-phase regions was finally determined by electrical resistance methods. The differential resistance bridge employed was first described by Stockdale (6), with modifications by Haughton (7). Two test specimens of very similar temperature/resistance characteristics were connected in opposition, the platinum leads having resistances which were low compared to those of the test specimens.

In general, the “standards” used for this work were the 15 per cent gold-platinum alloy and the 3 per cent gold, 12 per cent rhodium-platinum alloy. The test specimen and an appropriate “standard” were placed in a horizontal tube furnace, in a region of uniform temperature which was adjusted to about 700°C. The bridge was then balanced as closely as possible, and the furnace temperature changed by increments of 30°C at half-hourly intervals. The out of balance e.m.f. corresponding to each temperature level was carefully measured with a vernier potentiometer. A typical example of the graphs obtained by plotting these millivolt readings as a function of temperature is provided in Fig. 8. The lower inflection on this curve is caused by the 3 per cent gold, 12 per cent rhodium-platinum alloy “standard” entering the single-phase region. The upper inflection is attributable to the test specimen undergoing the same transformation.

At least three sets of determinations were made on each alloy, and the results obtained defined the position of the phase boundaries involved within limits of ±20°C. Support for the accuracy of these resistometric determinations was provided by the close agreement between the value of 1028°C determined for the 15 per cent gold-platinum alloy and the established value of 1030°C (8). Similarly, the value of 790°C determined by the resistance bridge on the 3 per cent gold, 12 per cent rhodium-platinum alloy is close to the 800°C obtainable from the X-ray work of Raub and Falkenburg (4).

Section through the Diagram at 85 per cent of Platinum

The results obtained by application of the previously described techniques made it possible to construct the vertical section shown in Fig. 9. It is obvious from this diagram that alloys used in contact with molten borosilicate glass will all be working in the single-phase condition. Differences in contact angle behaviour cannot therefore be attributed to discontinuous microstructural changes and must therefore be caused by fundamental modifications in the nature of the solid solution induced by increasing quantities of gold.

High Temperature Mechanical Properties

Stress-rupture data for the alloys containing 3 and 5 per cent of gold are presented in Fig. 10. These tests were made on specimens machined out of rolled sheet 0.060 inch thick and 0.25 inch wide. The gauge length was 1.5 inch. When stressed in tension at 700 p.s.i. at 1400°C the 5 per cent gold, 10 per cent rhodium, 85 per cent platinum alloy resists failure for approximately 150 hours. This compares with 60 hours for the binary 10 per cent rhodium-platinum alloy and 0.5 hour for pure platinum tested under the same conditions. The 3 per cent gold, 10 per cent rhodium, 87 per cent platinum alloy has a slightly shorter test life than the 10 per cent rhodium-platinum alloy although the difference is not pronounced. As shown by the elongation curves on Fig. 11, the gold-bearing alloy creeps less rapidly than the binary rhodium-platinum alloy, and although it has a lower elongation at fracture, this is unlikely to have an adverse effect upon its performance in service.

All the gold-bearing alloys were found to fail at the grain boundaries after high temperature stressing. This behaviour occurred at all stress levels, although rhodium-platinum is perfectly ductile at stresses above 400 p.s.i. at 1400°C. At lower temperatures, however, the 10 per cent rhodium-platinum alloy frequently exhibits intercrystalline failure and this has not imposed any restrictions upon its industrial employment.

Components in the glass industry are invariably taken out of service when distortions prevent efficient operation of the process. As indicated by the table, the fracture elongations of the gold-bearing alloys at 1400°C are high enough to ensure appreciable ductility under industrial conditions.