By most recent estimates, improved attention to friction and wear would save developed countries up to 1.6% of their gross national product—over $100 billion annually in the United States alone. It is thus not surprising that tribomaterials, materials designed for use in moving contact (sliding, rolling, abrasive, etc.) have for decades attracted the interests of materials scientists and mechanical and chemical engineers. However the field of tribology is hardly a recent one. Such tribological advances as Leonardo da Vinci's design of intricate gears and bearings (some of which were not built until the Industrial Revolution provided sufficiently strong materials) and the landmark 18th century development of a timepiece allowing accurate longitudinal positioning of ships at sea (accomplished via a self-lubricating wooden gear) could easily be termed “modern,” given the overall longevity of the field.

As important as tribomaterials are to technology, their discovery has usually been serendipitous. Materials scientists have frequently been able to provide explanations for why tribomaterials perform as well as they do. They have also been able to substantially improve the performance of tribomaterials through the development of new alloys, composites, and/or improved surface-engineering methods. They have however been far less successful at a priori design of tribomaterials with improved performance, largely because friction and wear processes have not been understood at the fundamental level.

The late 1980s marked the advent of renewed interest in fundamental areas of tribology, sparked by a number of new experimental and theoretical techniques that made it possible to study the force of friction in geometries that are well-defined at the nanometer scale.

In order to gain a fundamental understanding of friction, one must understand how the energy associated with work performed to overcome the frictional force is converted into heat at the molecular level. One of the simplest possible geometries in which friction can occur, and thus be studied, is that of a fluid or crystalline layer adsorbed on the surface of an ideal, atomically flat crystal. This system is ideal for studies of “interfacial” friction—that is, friction attributable to atoms and molecules immediately adjacent to the plane along which sliding occurs. Moreover it is directly accessible to experiments with a quartz crystal microbalance (QCM) and to theoretical studies through analytic calculations or molecular-dynamics (MD) simulations. The geometry is vastly simpler than that of contact between macroscopic objects in which the friction necessarily reflects the collective behavior of a multitude of buried contacts. (See the article by T. Baumberger and C. Caroli in this issue.) It is also far simpler than a case in which shear occurs within the bulk of a material rather than being confined to an interface. (See the article by D.A. Rigney and J.E. Hammerberg in this issue.) Nonetheless even in this simple geometry, friction can be far from negligible. A remarkable range of shear stresses, from 10−2 to 1010 N/m2, have been measured by a variety of techniques for wear-free geometries involving contact between crystalline interfaces.

Two of the most widely employed approaches for enhancing the lifetime and tribological performance of mechanical devices involve the introduction of lubricants and solid/hard coatings to the interface between two contacting surfaces. These approaches have been employed for centuries, and their design has been widely accomplished through empirical methods. In the last 20 years however, the general trend of device miniaturization and the development of advanced technology applications have placed more stringent demands on the performance of many tribological interfaces. For example, many of the hard-disk storage products now shipped in computers rely upon a single layer of lubricant molecules adsorbed on a submicron-thick layer of amorphous carbon to control the static friction and wear between the read/write head and the disk surface. In other advanced technologies, hard-coating materials are introduced into tribological interfaces that will operate in remote or harsh environments to improve the durability and lifetime of the devices. In the previous examples, the introduction of a thin coating or a change in surface composition dramatically influences the tribological properties and indicates that many tribological interfaces can be understood and described on a molecular scale. This article will review recent progress made in understanding the molecular events associated with tribological phenomena from both experimental and theoretical approaches.

In general, advances in experimental studies have been realized through developing a greater sensitivity to the relatively few number of species found within a tribological contact.

Scientists and engineers still debate the relative contributions of roughness and interlocking of asperities, adhesion and interatomic forces, and elastic and plastic deformation in friction phenomena. However a central point of agreement does exist: Friction involves energy dissipation, or the conversion of mechanical work into heat. Therefore any mechanism that allows for this energy dissipation during sliding interactions of materials deserves consideration, whether it occurs strictly at the nominal sliding interface or within material at some distance from that interface.

If the interactions are limited to atomic distances, it is entirely appropriate to focus attention on atoms at the two surfaces (see the articles by M.O. Robbins and J. Krim and J.A. Harrison and S.S. Perry in this issue). Friction is then a surface phenomenon. However if the sliding conditions lead to extensive plastic deformation of either material—over distances readily detectable by eye or by optical microscopy—then the associated energy dissipation is likely to account for most of the frictional losses. In such cases, friction involves more than processes occurring at the surface.

Materials designed for rolling or sliding contact, like corrosion-resistant materials, can provide great cost savings to industry. So why can't such “tribomaterials” be designed based on materials properties that control friction and wear? In recent years, it has become clear that the properties we need to understand are not only those of the starting materials—whether bulk solids or engineered surfaces—but also those of the materials generated within the sliding (or rolling) contact, the so-called “third-body” materials. This article reviews third-body processes and their role in controlling friction and wear of practical surface treatments.

It may seem odd that properties like slipperiness and wear resistance can be easily described but are not scientifically understood. We can feel that polytetrafluoroethylene is slippery but know that it is not very wear-resistant because we can scratch it with a fork. We can sense low friction when we rub talcum powder between our fingers but know it cannot lubricate indefinitely because it gets ejected as we rub. Is it contradictory to ask for materials that are both slippery and wear-resistant? Slipperiness is associated with low shear strength whereas wear resistance is modeled on high hardness—that is, high shear strength. Before we can answer this seeming contradiction, it is useful to review some aspects of friction and wear.

More than 50 years ago, Bowden and Tabor explained how a low shearstrength film can reduce the friction coefficient between two higher shearstrength materials in sliding contact.

In this article, we review the present status of experimental and theoretical work on friction at the interface between extended macroscopic bodies, rough on the micrometer scale. We show that systematic detailed studies of low-velocity friction using dynamical systems analysis, together with their shear response in the static state, provide a tool for investigating the physical processes taking place on the mesoscopic scale of real contacts between rough surfaces. This approach should shed light on the enduring question of the relationship between macroscopic friction and microscopic dissipative mechanisms. This still open issue has come back to the fore during the last decades, following considerable progress due to the development of “molecular tribometers.

Bowden and Tabor pointed out that, because nominally flat surfaces are in general rough on small scales, the real area of contact Ar (Figure 1) is only a small fraction ϕ of apparent contact area A0. On the other hand, they postulated the existence of a stressσs characteristic of the shear strength of the interface between a given couple of solids. Hence the friction force:

In this framework, the Amontons-Coulomb (AC) law F = μFN amounts to stating that Ar is proportional to the normal load FN where μ is the coefficient of friction.

When considering soft metals, Bowden and Tabor noticed that ϕ ≪ 1 entails that the nominal local pressure p on the real contacts—of the order of FN/(ϕA0)–generally overcomes the yield strength Y so that the contacting asperities flow plastically until p = H≈ 3Y, the “hardness” of the (softer) material. So, Ar = FN/H.

A growing interest exists in developing technologies that use silicon and other electronic materials as mechanical materials. Using standard processes of the integrated-circuit industry, researchers have successfully fabricated miniature mechanical components (micromachines) such as membranes, gears, motors, pumps, and valves. The integration of miniaturized mechanical components with microelectronic components has spawned a new technology known as microelectromechanical systems (MEMS). It promises to extend the benefits of microelectronic fabrication to sensing and actuating functions. Early applications of this technology include the digital mirror display, which has of the order of 106 aluminum thin-film micromirrors fabricated on top of a complementary-metal-oxide-semiconductor static random-access-memory integrated circuit. Other applications include integrated accelerometers for tasks such as air-bag deployment.

A number of fabrication techniques have been developed for this technology and have been reviewed elsewhere. In this review, I focus on surface-micromachining technology and adhesion and friction problems in surface-micromachined polycrystalline silicon (polysilicon) structures, though many of the principles discussed will also apply both to single-crystalline silicon and nonsilicon-based structures. Surface micromachining, defined as the fabrication of micromechanical structures from deposited thin films, is one of the core technological processes underlying MEMS. Surface microstructures have lateral dimensions of 50-500 μm with thicknesses of 0.1–2.5 μm and are offset 0.1–2 μm from the substrate. The basic steps in a surface-micromachining process appear in Figure 1. First the substrate is typically coated with an isolation layer (Figure la) that protects it during subsequent etching steps. A sacrificial layer is then deposited on the substrate and patterned. For simplicity, Figure 1b shows that the opening of the sacrificial layer is terminated on the isolation layer. The microstructural thin film is then deposited and etched (Figure 1c). Finally selective etching of the sacrificial layer creates the freestanding micromechanical structures such as the cantilever beam shown in cross section in Figure 1d. The technique can be extended to make multiple-layer microstructures.