CESAR - Computational Nanosciences - FRICTION AT THE NANOSCALE

Micro [1] and nano [2]-devices will have enormous impact on next generation civilian and military applications. This new ultra small technology will significantly improve the performance of already existing robots, computers, communication, and other electro/opto/mechanical devices. While initial efforts have been principally devoted to the fabrication and electrical performance of MEMS, recent studies have discovered a profound deleterious influence of friction and wear on the efficiency, power output, and steady state speed of micro-dynamics devices. Friction imposes serious constraints and limitations on the performance and lifetime of micro-machines [1-3] and, undoubtedly, will impose even more severe constraints on the emerging technology of nano-machines. Standard lubrication techniques used for large objects are expected to fail in the nano-world [4]. To make the needed future advances in micro- and nano-technology, a fundamental understanding of the operational, friction, and wear characteristics is paramount. Reducing wear and friction has a profound economic impact. 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 [3,5].

We propose combined experimental, theoretical and computational research to design efficient strategies to control and target frictional properties of sliding nano-objects. To succeed in this endeavor we have to significantly advance understanding of basic mechanisms of friction and frictional properties of sliding surfaces. Our experimental measurements will be performed using a high precision Surface Force Apparatus device that is capable to manipulate and control friction at the nanoscale level. Our theoretical/numerical research will combine molecular dynamics simulations (to guide and support experimental research) with modeling frictional behavior by reduced description to better understand the essence of frictional dynamics and suggest more efficient means to control friction. The expected outcome of this research will leapfrog our ability to define, predict, and control frictional properties of sliding nano-objects. The fundamental advance is to explain and control macroscopically dissipated energy by deliberate, externally driven, microscopic nonlinear manipulation. The practical significance of our effort will be in preventing nano-devices from breaking, cracking, faulting and having inaccurate performances. Successful completion of this research will help to design devices of higher reliability and, subsequently, lead to huge economical savings.

Micro-machine lubricant selection is complicated by considerations that hinder more conventional applications. In addition, due to the built-in-place nature of micro-mechanics, lubrication by the conventional means is prohibitive. The most significant differences between traditional approaches to lubrication, and the new approaches proposed here for control of friction, concern flexibility and accessibility. In traditional methods, the frictional properties of the lubricant depend on the parameters of the sliding objects (such as the load) and the external forcing. In contrast, in this newly proposed control of friction, frictional properties of the lubricant can be changed continuously by the frequency and the amplitude of the out-of-plane vibrations of the sliding surfaces. Even more importantly, responses of the sliding system to external vibrations might indicate faults of the micro-electro-mechanical or nano-device.

Recently, several groups [6-9] suggested that friction could be significantly reduced by applying small but intentionally focused control perturbations to the sliding system. Israelachvili [6] (experimental) and Landman [7] (full-scale molecular dynamics computer simulation), in considering thin-film boundary lubricated junctions, have coupled small-amplitude (of order of 1

) directional mechanical oscillations of the confining boundaries to the molecular degree of freedom of the sheared interfacial lubricating fluid. Load- and frequency- dependent transitions between a number of "dynamical friction" states were identified (Heuberger et. al. [6]). Grand-canonical molecular dynamics simulations (Gao et. al. [7]) confirmed the approach. Braiman et al.[8] and Krim et. al. [9] demonstrated numerically and experimentally that friction drops significantly when quenched disorder is introduced to the substrate. It has also been suggested [8,10] that frictional properties of sliding objects may significantly depend on the degree of phase synchronization of the nano-object. These results point to a completely new direction for realizing ultra-low friction in nano-mechanical devices.

It is not clear to what extent these preliminary studies describe a realistic frictional contact but the potential benefits following from this description are exciting. Moreover, the nanoscale situations of concern here present the most advantageous situation in which this should be possible, because of the extreme control over physical makeup of the sliding situation -- a degree of control that would be unthinkable in any situation involving macroscopic sliding. The potential high return of this proposed work is also relevant to MEMS applications.

We propose comprehensive research aimed at understanding the mechanisms of friction at the nanoscale. Based on this understanding, we intend to develop techniques to control friction. Our aim is to design surfaces with desired frictional properties. To succeed, the following challenges need to be overcome: (i) experimental measurements of frictional properties at the nanoscale; (ii) design of suitable control techniques to manipulate friction; (iii) rationally-designed surfaces with desired frictional properties; (iv) numerical simulations of sliding at the nanoscale; and (v) and optimization algorithms to track friction to desired behaviors.

One of the most intriguing possibilities concerns the hypothesis that motion in the lateral (shear) direction and in the normal (perpendicular) direction can be coupled in order to create a desired magnitude of frictional loss. The situation regarding the practical possibility of this is described in the following paragraph. The key experimental point is that our experiment in the Illinois lab is capable of testing this possibility. Lateral shear can be produced at shear rates up to 106 sec-1 [2,11]. Independently, vibrations of the film thickness can be induced over the same wide range [12] (see caption of Figure 1). This will allow direct and immediate test of the theoretical possibility that positive control on the magnitude of friction is possible using such superimposed oscillations. In the short term, the film thickness can be modulated in this fashion. In the longer term, it will be possible to incorporate feedback mechanisms, in which the amplitude and frequency of thickness modulation would be adjusted in response to transient friction response.

We have anticipated the possibility that the hypothesis of limited degrees of freedom will break down. To this end, the program includes an intensive molecular dynamics (MD) component, in order to distinguish theoretically between lubricants whose molecular makeup is different. To be explicit, the motivation here is that the additional relevant degrees of freedom, those that are not contained in present published strategies aimed to controlling friction, will involve the need to understand degrees of freedom of the confined fluid itself. Indeed, Landman [7] already presented MD evidence that a strategic approach to lower friction would require choosing impulse frequencies designed to thwart the ordering of the lubricant fluid molecules that would have otherwise have occurred. But these calculations involved an idealized system without direct application to an experimental situation. Extending these arguments, we can anticipate selecting conditions in which it might be possible to deliberately augment friction as desired. To accomplish these goals, it is essential to understand the internal dynamics [8,10], i.e. the internal rates of structural rearrangement, of confined molecules under shear. Comparison between experiment and simulation is particularly well suited to this problem because, since a nano-scale contact contains few (relatively few) molecules, simulations of relatively long duration become possible.

On the experimental side, two categories of lubricating nano-scale films are of particular relevance because they differ strongly in chemical makeup (see below). Existing theoretical models in this area have not yet incorporated chemical makeup as components of the model. One of the unique and most ambitious aspects of this proposal is to test our new theoretical ideas by intentionally varying the chemical makeup of the contact zone (an ambitious goal never before attempted on the theoretical side).

Friction measurements. To be of any relevance, the laboratory friction measurements should be done at the nanoscale level. When one considers friction in depth at the atomic and molecular levels, it is clear that custom-built devices at the University of Illinois (see figure 1) present unprecedented concomitant capacity to control and to characterize the surface chemistry, surface topography, and sliding dynamics of the sliding surfaces. Measuring frictional forces with such precision on custom-made surfaces will enable manipulation and control of friction on a nanoscale level. The instrument, using a homebuilt surface forces apparatus of unique design, allows one to form fluid films whose thickness ranges from sub-nanometer to several nanometers, or even as large as several micrometers -- the unique aspects of friction on the nano-scale can be clearly distinguished from bulk friction response.

We will contrast (a) a simple hydrocarbon liquid, squalane, which is a representative nonpolar liquid of appealing experimental simplicity because its film thickness is so readily controlled to be 1.6 nm; and (b) a simple aqueous situation, water containing electrolytes at relatively high ionic strength, ca. 0.01-1 M, which is readily controlled in the range 0.5-5 nm depending on the electrolyte. The hydrocarbon situation is of obvious energy significance and relevance owing to its connection to engine oils. The aqueous situation is of obvious energy significance and relevance owing to its connection to groundwater transport, to sliding in biomedical devices, to potential environmentally friendly technologies of the future, and to DNA-on-a-chip problems. The key scientific point is that the physics of these lubricating films is completely different -- the first being dominated by Lennard-Jones - van der Waals interactions, the second being dominated by electrostatic and hydrogen-bonding interactions. This comparison between two carefully chosen extreme situations will provide a stringent test of the theoretical predictions. It may also provide a benchmark for materials selection on the basis of which future choices of lubricant fluid in nanoscale machinery may be made.

Electromechanical Control of Friction. Friction, if manipulated by applying small adjustments (perturbations) to accessible elements and parameters of the sliding system, requires an a-priori knowledge of the strength and timing of the perturbations, similar in concept to chaos-control problems in which the small periodic perturbation technique [13] has shown great promise. We will develop rationally the following experimental agenda. First, displacements in the lateral (sliding) direction will be controlled by externally-applied perturbations in the normal direction, with motivation, provided earlier in this proposal, to reduce the magnitude of sliding friction at will. Secondly, friction in the sliding direction will be controlled by externally applied perturbations of the dimensions of sliding molecules, such as azo-benzene, whose geometry can be easily changed in response to externally applied optical excitation, thus opening the avenue to optically based control of friction. Third, friction in the sliding direction will be controlled by electric stimulus to the boundary layers in which sliding energy is dissipated -- an aspect of especial significance as concerns friction in aqueous systems where the electric double layer plays a paramount role. The experiments will be supported and guided by molecular dynamics simulations. Our theoretical/numerical research will be directed towards identification of parameter range to implement robust electromechanical control of friction.

Control of Chaos in Friction. In many situations, the behavior of the lubricant and the sliding surfaces is chaotic. Thus control of chaos and targeting towards desired behavior is essential. Techniques for controlling chaos in other systems and in other areas of scientific endeavor have been under systematic investigation since 1990 but never before have been applied to controlling friction at the experimental level. The main challenge is to develop efficient and robust methods to selectively drive low and high dimensional chaotic dynamics by using very tiny external perturbations. In particular, control of chaos in spatially extended system has not reached yet a high applicability level, and research in this direction is of very significant importance. We will study control of chaos in friction using both non-feedback [13] and feedback [14] methods.

Mechanisms for Friction and Lubrication at the Micro- and Nano-Scale. The mathematical foundations that underlie the proposed description of basic friction mechanisms are based on minimalist models with limited degrees of freedom. It is not clear to what extent they describe a realistic frictional contact but the potential benefits following from this description are exciting. Moreover, the nanoscale situations of concern here present the most advantageous situation in which this should be possible, because of the extreme control over physical makeup of the sliding situation -- a degree of control that would be unthinkable in any situation involving macroscopic sliding. We will study the dynamics of minimalist models of friction to describe surface force apparatus and quartz microbalance experiments. There exists strong evidence [8,10] that the dynamical properties of a sliding object (such as, for example, phase synchronization of the elements in the sliding object) can significantly affect friction. We will apply modern tools of nonlinear dynamics to study phase synchronization and its connection to frictional properties.

Proposed major milestones for the basic research tasks described in this section are given below. All work will be performed on a best-effort basis.

FY 01: Experimental apparatus will be built and tested (06/01). Experimental methods to measure shear-induced normal displacements will be implemented. The linear-responses of sliding systems during the process of nonlinear excitation will be demonstrated using novel breakthroughs, both experimental and theoretical, that are elaborated in this proposal. Methods to produce optical and electrical stimuli will be elaborated (09/01).

FY 02: Control of chaos in friction will be demonstrated (05/02). Chaos generated by conventional mechanically generated perturbations will be refined and controlled by optically and electrically controlled excitations on these same systems. (09/02).

FY 03: Nonlinear mechanisms for friction at the nanoscale will be formulated. The limits of the classical emphasis on linear responses will be rationally elucidated. Mechanisms by which to control the magnitude of friction by rational implications of newly understood nonlinearities will be demonstrated. Nonlinearities will concern not only mechanics, but also optical and electrical stimulation, both being of especial significance in MEMS devices. These newly developed approaches to control friction at the nano-scale will be contrasted explicitly with classical mechanisms of friction at the bulk where nano-scale control of this deliberate kind is not thinkable (06/03).

We will deliver to DOE/BES one copy of each publication submitted to the open literature, and one copy of each invention disclosure submitted for patent. Major technical highlights will also be provided on a timely basis.

H. Fujita, "Micro-actuators and Micro-machines", Proceedings of the IEEE 86, 1721-1732, (1998). A. M. de Aragon, "Space Applications of Micro/Nano Technologies", Journal of Micromechanics and Microngineering 8, 54-56, (1998). R. Maboudian, "Adhesion and Friction Issues Associated with Reliable Operation of MEMS", MRS Bulletin 23 (6), 47-51, (1998).
B. Bhushan, J. Israelachvili, and U. Landman, "Nanotribology: Friction, Wear, and Lubrication at the Atomic Scale", Nature 374, 607-616 (1995). See also: S. Granick, "Soft Matter in a Tight Spot," Physics Today 52, 26-31, (1999).
J. Krim, "Fundamentals of Friction", MRS Bulletin 23 (6) 20-21, (1998).
Yuan-Zhong Hu and S. Granick, "Microscopic Study of Thin Film Lubrication and its Contribution to Macroscopic Tribology", Tribology Letters 5, 81-88, (1998).
H. P. Jost, "Tribology - Origin and Future", Wear 136, 1-17, (1990).
M. Heuberger, C. Drummond, and J. Israelachvili, "Coupling of Normal and Transverse Motion during Frictional Sliding", J. Phys. Chem. B 102, 5038-5041, (1998).
J. Gao, W. D. Luedtke, and U. Landman, "Friction Control in Thin-Film Lubrication", J. Phys. Chem. B 102, 5033-5037, (1998).
Y. Braiman, F. Family, H. G. E. Hentschel, C. Mak, and J. Krim, "Tuning Friction with Noise and Disorder", Phys. Rev. E 59, R4737-R4740, (1999).
C. Mak and J. Krim, "Quartz Crystal Microbalance Studies of Disorder-Induced Lubrication", Faraday Discuss 107, 389-397, (1997).
H. G. E. Hentschel, F. Family, and Y. Braiman, "Friction Selection in Nonlinear Particle Arrays", Phys. Rev. Lett. 83, 104-107, (1999). F. Family, H. G. E. Hentschel, and Y. Braiman, "Friction at the Nanoscale", J. Phys. Chem. B 104, 3984-3987, (2000).
S. Granick, "Motions and Relaxations of Confined Liquids," Science 253, 1374-1379, (1991). A. Levent Demirel and S. Granick, "Friction Fluctuations and Friction Memory," Phys. Rev. Lett. 77, 4330-4333, (1996).
A. Dhinojwala and S. Granick, "Surface Forces in the Tapping Mode," Macromolecules 30, 1079-1085, (1997). 1
Y. Braiman and I. Goldhirsch, "Taming Chaotic Dynamics by Weak Periodic Perturbations", Phys. Rev. Lett. 66, 2545-2548, (1991).
E. Ott, C. Grebogi, and J. A. Yorke, "Controlling Chaos", Phys. Rev. Lett. 64, 1196-1199, (1990). K. Pyragas, "Continuous Control of Chaos by Self-Controlling Feedback", Phys. Lett. A 170, 421-428, (1992).