Elsevier

Wear

Volume 250, Issues 1–12, October 2001, Pages 420-430
Wear

Sliding behavior of metallic glass: Part II. Computer simulations

https://doi.org/10.1016/S0043-1648(01)00607-XGet rights and content

Abstract

Molecular dynamics (MD) calculations were used to simulate the sliding of a two-component 2D amorphous system interacting via Lennard–Jones potentials. The friction coefficient showed a transient before reaching an average steady state value. The steady state friction coefficient was observed to decrease with an increasing sliding velocity. Mixing was observed at the sliding interface. The mixed layer grew at a rate that scaled with the square root of time. A density decrease was recorded in the region adjacent to the sliding interface. This spatially corresponded to the softer layer detected experimentally near the worn surface in a Zr41.2Ti13.8Cu12.5Ni10.0Be22.5 bulk metallic glass alloy after sliding. Subsurface displacement profiles produced in these simulations were similar to those observed in other material systems. The Navier–Stokes equation was used to analyze the material flow pattern, with results in agreement with data obtained from simulations. This suggests that the observed subsurface displacement profile may be a generic material flow pattern under combined compression and shear.

Introduction

In the preceding paper (Part I), the results of experimental studies on the tribological properties of Zr41.2Ti13.8Cu12.5Ni10.0Be22.5 bulk metallic glass (BMG) alloy were described. These included the morphology of the wear track, pin and debris, the effects of devitrification on friction and wear, re-amorphization after sliding of devitrified BMG, the generation of a mixed layer at the interface and the evolution of subsurface microstructure. However, because of the amorphous structure of the BMG system, certain experimental techniques proven effective for crystalline materials are not available for a BMG system. For example, for crystalline materials, useful information on microstructure evolution can be obtained by examining cross-sections of worn specimens by SEM or TEM. However, the structural inhomogeneities that give rise to the contrast in various electron microscopies do not exist in BMG alloys. An alternative approach involves computer simulation. Molecular dynamics (MD) simulation provides a powerful tool to complement experimental studies. In some cases, simulations can provide information that is difficult to obtain experimentally. The sliding behavior of BMG materials provides a good example for the complementary applications of experimental and simulation techniques.

There is growing interest in the applications of molecular dynamics simulation in tribology [1], [2]. Among the phenomena studied by MD have been static friction problems [3], indentation and cutting phenomena [4], [5], lubrication at asperities [6] and the response of lubricant molecules bonded to the surface [7]. The MD simulation was also used to study microstructure evolution during the highly nonequilibrium and irreversible processes taking place at a sliding interface [8], [9], [10], [11], [12]. In that study, the sliding of two blocks of single crystal copper was simulated. Both mechanical mixing and the development of nanocrystals adjacent to the interface were observed.

In this paper, the application of MD simulation to an amorphous solid in sliding contact is described. Connections between the experimental observations described in Part I and the simulation results will also be presented.

Section snippets

Potential and composition

Ideally, one uses potentials and compositions closest to those of the alloy system investigated experimentally. However, for the bulk metallic glass alloy system Zr41.2Ti13.8Cu12.5Ni10.0Be22.5, the precise forms of the interatomic potentials are unavailable for many of the atomic pairs involved. In addition, a large scale 3D simulation of a five component system would involve excessive computational time and resources. Therefore, a two-component 2D model amorphous system was used in the

Friction coefficient

The friction coefficient μ at any time t was calculated usingμ(t)=F(t)W(t)where F(t) is the friction force and W(t) the normal load, determined from the normal pressure in the Y-direction.

Fig. 3 is a typical plot of friction coefficient evolution observed in the simulation. The friction coefficient exhibits a transient overshoot before it settles down to an average steady state value. The maximum friction coefficient reached during overshoot is about 0.5 for all sliding speeds tested. However,

Conclusions

  • 1.

    The friction coefficient in the simulation has an initial transient before reaching an apparent steady state value.

  • 2.

    The steady state friction coefficient decreases as the sliding speed increases.

  • 3.

    Atomic scale mixing takes places at the sliding interface. The growth of the mixed layer scales with the square root of time for a rescaled sliding speed of Vs=1.0.

  • 4.

    The density near the sliding interface decreases after the start of sliding, probably due to the accumulation of shear-induced excess free

Acknowledgements

We are pleased to acknowledge research support from the National Science Foundation, Solid Mechanics and Materials Engineering, most recently under Grant No. CMS-9812854, and from Howrnet Corp. We also thank the Ohio Supercomputer Center and S. Glotzer at NIST for the use of computer facilities, and M.J. Mills, P.K. Gupta and A. Markworth at The Ohio State University and J. Hammerberg at Los Alamos National Laboratory for helpful discussions and suggestions.

References (33)

  • M.O. Robbins, M.H. Muser, Computer simulations of friction, lubrication and wear, in: B. Bhushan (Ed.), Handbook of...
  • G. He et al.

    Absorbed layers and the origin of static friction

    Science

    (1999)
  • M.R. Sorensen et al.

    Simulation of atomic-scale sliding friction

    Phys. Rev. B

    (1996)
  • J. Gao et al.

    Nano-elastohydrodynamics: structure, dynamics, and flow in nonuniform lubricated junctions

    Science

    (1995)
  • J.A. Harrison et al.

    Friction in the presence of molecular lubricants and solid/hard coatings

    MRS Bull.

    (1998)
  • B.L. Holian et al.

    The birth of dislocations in shock waves and high-speed friction

    J. Comp.-Aided Mat. Design

    (1998)
  • Cited by (91)

    • Mechanical and tribological performance of CoCrNiHf<inf>x</inf> eutectic medium-entropy alloys

      2021, Journal of Materials Science and Technology
      Citation Excerpt :

      Due to the chemically active elements contained such as Co, Hf, and so on, the roughened surface materials were oxidized quickly [58], which is consistent with the results in Table 2. Previous studies show that the oxygen-containing materials are usually brittle [59], thereby can be removed easily due to the tensile stress developed at the trailing of the sliding particles [60], resulting in a large wear volume. But, as Hf0.3 and Hf0.4 alloys have higher hardness, plastic deformation could be effectively restrained under a low normal load (Fig. 8) and the subsequent oxidation is relatively slight, as can be seen in Table 2.

    • Effect of W addition on the glass forming ability and mechanical properties of Fe-based metallic glass

      2018, Journal of Alloys and Compounds
      Citation Excerpt :

      Bulk metallic glasses (BMGs) have been studied extensively for their unique properties, such as high strength and hardness, as well as excellent corrosion and wear resistance [1–6].

    View all citing articles on Scopus
    1

    Address: Merck Research Laboratories, WP78-304, Sumneytown Pike, West Point, PA 19486, USA.

    View full text