Correlating modeling and experimental atomic friction results

Dr. Neil Canter, Contributing Editor | TLT Tech Beat August 2015

Parallel replica dynamics slows down the modeling process and moving the sample instead of the AFM tip speeding up the experimental process to improve the correlation of the two methodologies.

KEY CONCEPTS
• Researchers studying atomic friction are having difficulty correlating modeling work with experimental results.
• Better correlation has been achieved by slowing down the modeling process and speeding up the movement of an AFM tip over an atomic surface.
• Discrepancies between modeling and experimental results still remain and might be due to the mass of an AFM tip and thermal effects.

RESEARCHERS ARE STEADILY GAINING A BETTER UNDERSTANDING OF how friction takes place at the atomic scale. At this point it is clear that the interaction of atoms to generate friction is different from the macroscale where two surfaces, such as those in machinery, interact with each other.

In a previous TLT article, researchers found that applying an electric field and adjusting the relative humidity enables the nanoscale friction of ionic solids to be adjusted (1). The water solubility of a particular ionic salt is an important factor because those salts with high solubility will display a friction reduction at a lower relative humidity. The researchers believe the friction reduction is due to the formation of an electric double layer of cations and anions at the junction between the top of an atomic force microscope (AFM) and the surface of the ionic salt. 

STLE board member Ashlie Martini, associate professor in the School of Engineering at the University of California-Merced in Merced, Calif., discusses the concept of atomic stick-slip sliding, which is related to understanding the mechanism for atomic friction. She says, “Stick-slip friction at the macroscale is the cause of the sound of a squeaky door as it is moving. At the atomic level, this phenomenon is seen with an AFM tip as it moves across a surface of atomic crystal lattices.”

Gaining a better understanding of friction at the atomic scale is extremely important. New technology under development to operate at the nanoscale will have to take into account atomic scale friction to operate effectively. This will impact many of the electronic devices that we depend upon. 

One problem that has faced researchers studying atomic friction is how to correlate the modeling work done to explain the phenomenon of atomic stick-slip friction with experimental results conduced with an AFM.

Martini uses molecular dynamics (MD) simulations, which can correlate atomic positions in stick-slip friction. She says, “MD simulations have been around for over 20 years and work effectively for a tiny volume of atoms. But scaling MD simulations up to capture the size of the AFM tip and the surface that it is scanning is a major challenge.”

In addition, the time scale gap between MD simulations and experimental results is large. Martini explains, “MD simulations calculate forces between atoms on a femtosecond scale. In contrast, an AFM tip moves across a surface of atoms in milliseconds. For a computer to reproduce the data generated by the AFM tip in such small steps, it would take about 30 years.”

The value of having modeling be consistent with experimental results cannot be minimized. Martini says, “Modeling helps to explain what is going on in experiments. However, a model for atomic scale friction that is not validated through experimentation is useless.”

Martini, in collaboration with STLE-member Robert Carpick, John Henry Towne professor and chair of mechanical engineering and applied mechanics at the University of Pennsylvania in Philadelphia, has now narrowed the gap between the modeling and experimental work done on atomic stick-slip friction.

PARALLEL REPLICA DYNAMICS
Martini and Carpick have taken steps to slow down the modeling process and speed up the movement of the AFM tip in order to improve the correlation of their two methodologies. The approach taken by Martini is to use a modeling technique known as parallel replica dynamics (PRD). 

Martini says, “PRD enables us to slow down the scan speed in our models by running parallel simulations in time using multiple computers.”

From the experimental standpoint, Carpick and his research group realized that there is a limitation in speeding up the movement of the AFM tip over an atomic surface. Their answer to speeding up the empirical process is to prepare a very compact shear piezo plate that moves the sample instead of the AFM tip (see Figure 1).


Figure 1. Adjusting atomic stick-slip friction experimentation so that the sample is moved instead of the AFM tip leads to better correlation with modeling studies. (Figure courtesy of the University of Pennsylvania.)

Experiments used in correlating the data were run on a gold surface oriented in a 111 plane and carried out in an ultrahigh vacuum at room temperature at a pressure of approximately 6 x 10^-10 torr. Evaporation of bulk gold onto muscovite mica was done to prepare the gold samples. The simulations were done to model the same system.

The researchers were able to model and run experiments at a scan speed that is in the micrometers per second range. Martini says, “This is the very first time that we are able to predict and measure atomic scale friction at the same speed.”

But discrepancies still remain with the atomic friction being larger in experiments than in simulations. Martini says, “We have identified two parameters that may have an impact on why our results do not completely line up with each other. The first is the mass of the AFM tip, which is involved in the stick-slip inertial process and not accounted for by our modeling. Thermal effects and other factors attributed to noise from the AFM instrument also may cause the modeling and experimental results not to completely match up.”

Still, this work represents significant progress in having model simulations validated by experiments and then used to explain what is occurring in those experiments. Future work will involve a shift to focus on the effect of temperature on atomic friction. Martini says, “The reasons why friction changes as a function of temperature are complex and not well explained. We also are going to continue work to match model simulations with experimentation.”

Additional information can be found in a recent article (2) or by contacting Martini at amartini@ucmerced.edu. 

REFERENCES
1. Canter, N. (2015), “Controlling friction at the nanoscale,” TLT, 71 (4), pp. 10-11.
2. Liu, X., Ye, Z., Dong, Y., Egberts, P., Carpick, R. and Martini, A. (2015), “Dynamics of atomic stick-slip friction examined with atomic force microscopy and atomistic simulations at overlapping speeds,” Physical Review Letters, 114, 146102.


Neil Canter heads his own consulting company, Chemical Solutions, in Willow Grove, Pa. Ideas for Tech Beat can be submitted to him at neilcanter@comcast.net.