Simulated Tribochemistry: An Atomic-Scale View of the Wear of Diamond

Molecular dynamics simulations are used to explore the atomic-scale chemistry and associated wear that occurs when diamond surfaces are placed in sliding contact. The simulations predict complex radical chemistry initiated by the shearing of hydrogen atoms from chemisorbed molecules. Observed chemical mechanisms include hydrogen abstraction from the surfaces, radical recombination, transient surface adhesion, and formation of debris at the interface. These simulations provide the first glimpse into the rich, nonequilibrium tribochemistry that occurs at diamond and related covalently-bonded interfaces. Placing two solid bodies in sliding contact usually results in wear of one or both of the bodies, often with the formation of The nascent particles that form the debris can further interact, either physically or chemically, with the solid bodies or with other nascent particles. This interaction of chemistry and friction is known as trib~chemistry,~,~ the results of which are well-known in everyday life-combustion engines break down, cutting tools become dull, and bearings fail. Despite the obvious importance of these consequences, and the long history of the field of tribology, much of the atomic-scale dynamics responsible for friction-induced wear remain elusive. There have been numerous experimental attempts to characterize tribochemical reactions and resulting wear. For example, Heinicke and co-workers4 have examined the tribochemical reactions which take place in ball mills; Singer5 has developed a thermochemical analysis to describe the reactions which take place during the wear of MoS2 films, TiN and Tic coatings, and steel implanted with Ti+; and Fishefi has examined the kinetics of tribochemical reactions. However, charcterization of these reactions has proven difficult because traditional analysis methods such as optical microscopy, electron microscopy, cathodoluminescence, and infrared ~pectroscopy’-~ are restricted to examination of the contact region subsequent to sliding. This limitation has led to a historic lack of information regarding the mechanisms of tribochemical reactions. Even new atomic-scale proximal probe techniques, such as atomic-force1° and friction-force microscopies,ll have yet to characterize specific, atomic-scale, tribochemical reaction mechanisms. * To whom correspondence should be addressed. + U S . Naval Academy. * Naval Research Laboratory. @ Abstract published in Advance ACS Abstracts, October 1, 1994. (1) Bowden, F. P.; Tabor, D. Friction: An Introduction to Tribology; Anchor Books, Anchor PressDoubleday: Garden City, NY, 1973; pp 9-24. (2) Rabinowitz, E. Friction and Wear of Materials; John Wiley & Sons: New York, 1965; pp 52-167. (3) Granick, S. Mater. Res. Bull. 1991, 16, 33-35. (4) Heinicke, G. Tribochemistry; Carl Hanser Verlag: Munich, 1984. (5) Singer, I. L. Surf. Coatings Technol. 1991, 49, 474-481. (6) Fisher, T. E. Fundamentals of Friction: Macroscopic and Microscopic Processes; Singer, I. L., Pollack, H. M., Eds.; H. M. Kluwer Academic Publishers: Dordrect, The Netherlands, 1992; pp 299-310. (7) Tabor, D.; Field, J. E. The Properties of Natural and Synthetic Diamond; Field, J. E., Ed.; Academic Press: London, 1992; pp 547-571 and references therein. (8) Feng, Z.; Field, J. E. Su$. Coatings Technol. 1991, 47, 631-645. (9) Wilks, J.; Wilks, E. M. The Properties of Diamond; Field, J. E., Ed.; Academic Press: London, 1979; pp 351-381. Macroscopic experiments have shown that, for diamond sliding on diamond in ultrahigh vacuum, friction is initially low. As sliding progresses, however, friction increases dramati~ally.~~~ It has been asserted that the friction is low initially due to saturation of surface radical sites by hydrogen and perhaps oxygen. Sliding is thought to wear away particles from the surfaces, creating radical sites. Once these radical sites have been formed, the surfaces adhere through the formation of chemical bonds. The formation of wear debris is then thought to result from the shearing of these bonds during continued sliding. This debris appears opaque and powdery and is composed of mostly amorphous or unsaturated carbon with little hydrocarbon and g r a ~ h i t e . ~ , ~ Wear rates associated with this process have been observed to depend on the direction of a b r a s i ~ n . ~ , ~ Ideally, one would like to be able to monitor reactants and the formation of the various intermediates as sliding progresses. Theoretical techniques such as molecular dynamics, where the precise positions of all atoms are known as a function of time, are uniquely suited to this task.12,13 Indeed, in this work, we report specific examples of tribochemical reaction sequences that can occur when two diamond surfaces are placed in sliding contact. These simulations provide the first glimpse into the rich, nonequilibrium tribochemistry that is possible in this and related covalently-bonded systems. The molecular dynamics friction experiment was carried out in the following way. Two diamond lattices consisting of 10 carbon atom layers, each layer containing 16 atoms, were placed in contact. The contacting surfaces of both lattices were the (1 11) surfaces of the diamond lattices. Two series of simulations were carried out. In the first series, the (1 11) contacting surfaces of both lattices were terminated with hydrogen only. In the second series, two hydrogen atoms from the upper surface were removed and replaced with ethyl (-CHzCH3 or R) groups (10) Binnig, G.; Quate, C. F.; Gerber, Ch. Phys. Rev. Lett. 1986, 56, 930-933. Bumham, N. A,; Colton, R. J . Scanning Tunneling Microscopy and Spectroscopy; Bonnell, D. A., Ed.; VCH: New York, 1993; pp 191249 and references therein. (1 1) Mate, C. M.; McClelland, G. M.; Erlandsson, R.; Chiang, S . Phys. Rev. Lett. 1987, 59, 1942-1945. Meyer, E.; et al. Thin Solid Films 1992, 220, 132-137. German, G. J.; Cohen, S . R.; Neubauer, G.; McClelland, G. M.; Seki, H.; Coulman, D. J . Appl. Phys. 1993, 73, 163-167. (12) For reviews of simulations related to friction, see: Mater. Res. SOC. Bull. 1993, 18. See also: Glosli, J. N.; McClelland, G. M. Phys. Rev. Lett. 1993, 70, 1960-1963. (13) Harrison, J. A.; White, C. T.; Colton, R. J.; Brenner, D. W. J . Phys. Chem. 1993, 97, 6573-6576. This article not subject to U.S. Copyright. Published 1994 by the American Chemical Society 10400 J. Am. Chem. Soc., Vol. 116, No. 23, 1994 Harrison and Brenner