The endurance of MoS2 sputter-coated steel balls was measured in conti
nuous and stop-start sliding tests in a four-ball wear tester. Test va
riables were counterface materials, uncoated steel and two ceramics (C
o-bonded tungsten carbide and sapphire), and gaseous atmospheres, dry
Ar and dry air. In continuous tests, coating endurance increased from
14 to 80 min; combinations were ranked as follows: steel in air=cerami
cs in air < steel in Ar < ceramics in Ar. Coating endurance in stop-st
art tests was the same as in continuous test for all combinations exce
pt steel in Ar, which failed sooner in stop-start tests. MoS2 wear tra
cks were analyzed at intervals from 1 min to failure by optical micros
copy, energy-dispersive X-ray spectroscopy (EDX) and Auger electron sp
ectroscopy (AES). Tracks run in Ar were smoother and showed ductile fl
ow; tracks in air were rougher and developed ragged scratches. Blister
s formed and contributed to failure in both atmospheres: in Ar they le
d to a slower ductile failure whereas in air they led to more rapid br
ittle failure. Abrasion by the worn steel ball accelerated coating wea
r in Ar. EDX of wear tracks showed that 2/3 of the coating was removed
within the first 10 min in both atmospheres; AES of the same tracks f
ound higher levels of oxygen in MoS2 run in air. Mechanisms of wear, t
ransfer and lubrication are discussed, and a quantitative model for bl
ister formation in ductile MoS2 coatings is presented. It is recommend
ed that hard ceramics replace abrasion-prone steel to enhance the wear
life of MoS2-coated steel in oxygen-free environments.