A transmission electron microscopy study of wear tracks on nickel, nickel-30 weight percent chromium, and nickel-30 weight percent molybdenum wear test specimens
The wear of commercially pure nickel, nickel-30 weight percent chromium, and nickel-29 weight percent (20 atom percent) molybdenum was studied using transmission electron microscopy. The specimens studied were the subjects of previous investigations of surface effects of wear. The nickel and nickel-chromium test rings were wear tested in annealed condition; the nickel-molybdenum test rings were tested either in the disordered or in the ordered condition. The disordered alloy has molybdenum atoms at random sites in solid solution in the nickel. This is achieved by heating above the critical temperature (868°C) and cooling rapidly to room temperature. The ordered alloy has the molybdenum atoms located on specified sites relative to the nickel atoms. The ordered structure is obtained either by reheating the disordered structure for sufficient time at a temperature below.the critical temperature, or by slowly cooling from above the critical temperature. The average diamond pyramid hardness for the test rings were as follows: r nickel--77; nickel-chromium--156; disordered nickel-molybdenum--174; ordered nickel--molybdenum--362.
The wear tests were run in an apparatus which simulates the wear conditions in mechanical face seals. The test specimens were in the form of rings about two and one-half inches in diameter. The test rings were lapped to obtain a centerline average asperity height of less than 5µm, and flatness within two helium light bands (58.9 millionths centimeters). Test durations ranged from ten seconds to one hundred thirty-two seconds. The test between nickel and nickel-chromium was run at 9.0 meters per second sliding velocity, 1.24 x 10-3 kilograms per square millimeter face loading pressure and nitrogen as the sealant gas. The tests involving the nickel-molybdenum specimens were run at 9.75 meters per second sliding velocity, 1.55 x 10-3 kilograms per square millimeter face loading pressure and argon as the sealing gas.
Sections were cut from the test rings and sliced parallel to the wear surface in layers about O.025 centimeters thick. The slices were prepared for transmission electron microscopy by electrolytic thinning. The wear surface specimens were thinned from the unworn side only. This resulted in samples whose electron-transparent area was actually part of the wear track. Other samples for subsurface study were electrothinned from both sides of the sample.
The results of this study are:
(1) Both nickel-chromium and nickel-molybdenum alloy material were transferred during the wear test to the nickel test rings mated against them. The transferred alloy in the three alloy-nickel tests exhibited very fine recrystallized grains of the order of 0.1 µm. The test rings of the nickel-chromium alloy and the nickel-molybdenum alloy tested in the disordered condition both showed recrystallized grains. The ordered nickel-molybdenum test ring deformed by twinning and did not recrystallize.
(2) In the test between two disordered nickel-molybdenum test rings, recrystallization occurred in both rings. The recrystallized grains in the stator wear track were less than O.l um. In the rotor wear track, however, the dislocation density was high and individual grains could not be distinguished readily. The diffraction pattern proved the wear track to be recrystallized. The differences in the two wear track structures is believed to be related to the relative masses of the test ring sections adjacent to the wear interface. This point is detailed below.
(3) In the test involving ordered nickel-molybdenum and disordered nickel-molybdenum alloy test rings, recrystallization was observed in the disordered ring, but not in the ordered ring. Bands were formed in the ordered test ring as a result of wear reactions. Although these bands closely resembled twins, a cursory attempt to establish their nature was not successful.
(4) The stator of the ordered alloy versus ordered alloy test contained recrystallized grains which gave an ordered diffraction pattern. The rotor, however, was deformed by twinning and did not recrystallize. Here also, the difference in the structures of the two test rings is probably related to the mass effect on interface temperatures (see 5).
(5) The width of the stator at the wear interface is much narrower than that of the rotor. This would allow the heat generated at the rotor wear surface during the wear test to be conducted away from the wear interface much more quickly than the heat generated in the stator. The thermal gradient in the stator would be shallower than that in the rotor, and recrystallization of the deformed material in the stator would be more likely than in the rotor. This effect was observed, as discussed in points 2 and 4 above.
(6) The bulk stator temperatures recorded by Merrick (3) in no case exceeded 240°C. These temperatures were recorded at locations away from the interface and, of course, did not record the very local, brief temperature increases at the wear interface. The microstructural evidence in the present study indicates that interface temperatures ranged from above the recrystallization temperature of nickel (probably about 350°C) to above the order-disorder transition temperature of Ni4Mo (868°C).
Thesis75.B328.pdf
11.69 MB
Adobe PDF
95e7a8c4a5ca328074460d8d43e380c0