Superlubricity: Seen at the macroscale for the first time

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

Computational simulations were conducted to explore the underlying mechanism of macroscale superlubricity.

KEY CONCEPTS
• The objective for improving machinery efficiency is to achieve superlubricity at coefficient of friction values below 0.005.
• Superlubricity is seen for a short time frame in computational simulations of a graphene-coated surface sliding against a DLC-coated surface due to the formation of nanoscrolls after the graphene wraps around the DLC surface.
• Introduction of nanodiamond particles to the simulations stabilizes the nanoscrolls and maintains superlubricity over a longer time frame.
ONE OF THE CHALLENGES FACED BY RESEARCHERS is to extrapolate the findings observed on minimizing friction at the atomic level to the macroscale. But there also have been problems with correlating results on atomic friction between modeling studies and experimental results.

In a previous TLT article, work conducted to bridge the gap between modeling and experimentation on atomic friction was described (1). A collaboration between two researchers has led to better correlation by taking steps to slow down the modeling process using a technique known as parallel replica dynamics and speeding up the movement of an atomic force microscope tip. The net result is that modeling and experimental data on atomic friction were taken at the same speed. 

The desired target for improving machinery efficiency is to attain superlubricity. Dr. Subramanian Sankaranarayanan, computational nanoscientist at the Argonne Leadership Computing Facility, a division of Argonne National Laboratory in Argonne, Ill., says, “Superlubricity is a friction state where the coefficient of friction drops to nearly a zero value. Technically, superlubricity occurs at coefficient of friction values below 0.005. Past research has found coefficient of friction levels as low as 0.001, primarily at the nanoscale.”

Previous efforts to develop superlubricity at the macroscale have centered on the use of highly oriented pyrolytic graphite surfaces, multiwalled carbon nanotubes and graphene. While superlubricity has been seen at the nanoscale, in most cases the phenomenon is due to the incommensurability of lattice planes sliding against each other leading to structural lubricity. This example can only be seen at the nanoscale. Defects and deformations seen in these materials at the macroscale have prevented superlubricity from being detected at the macroscale.

Initial experimentation at Argonne started with sliding a nickel surface against diamond-like carbon (DLC). Sankaranarayanan says, “While evidence of superlubricity was observed by our experimental collaborators, it was felt that the best course of action is to work with sp²/sp³-bonded carbons because they might provide incommensurate, amorphous surfaces where superlubricity has been found in the past.”

This led the researchers to slide a graphene-coated surface against a DLC-coated surface. Sankaranarayanan says, “Superlubricity was detected in some cases, but the friction levels fluctuated and were much greater when moisture was present.”

After the initial experimentation, Sankaranarayanan and his colleague, Argonne researcher Dr. Sanket Deshmukh, were asked to conduct computational simulations to better explain the experimental results and explore the underlying mechanism of macroscale superlubricity. 

NANOSCROLLS
In the initial set of simulations, the researchers found that the graphene wraps around the DLC surface to form nanoscrolls. Deshmukh says, “As soon as the graphene particles formed scrolls, the coefficient of friction dropped into the superlubricity range. The problem is that the graphene flakes collapsed due to the instability of the scrolls leading the coefficient of friction to go back up.”

To stabilize the graphene nanoscrolls, the researchers decided to introduce nanodiamond particles into the simulations. Deshmukh says, “Nanodiamond particles consist of undercoordinated carbon atoms that react with the graphene surface to facilitate the formation of stable nanoscrolls. Graphene forms defective sites under the stress of the process that easily react with the nanodiamonds stabilizing the nanoscrolls.

The process of nanoscroll formation is shown in Figure 1. Nanoscroll diameter is dependent upon the size of the particles with respect to graphene flakes. Deshmukh says, “We used the simulations to find the optimum ratio of graphene flakes to nanodiamond particles, which we consider the sweet spot.”


Figure 1. Superlubricity can be demonstrated when a graphene-coated surface containing nanodiamond particles slides against a DLC-coated surface forming nanoscrolls. (Figure courtesy of Argonne National Laboratory.)

Superlubricity originates from two factors according to Sankaranarayanan. He says, “Graphene flakes slide over an incommensurate DLC surface, which is energetically favored to reduce the coefficient of friction. The choice of materials is a consideration because the nanoscrolls reduce the contact area with the DLC surface in contrast to graphene flakes. This means that fewer carbon atoms on the graphene will interact with DLC. The only force of attraction in the all-carbon system is van der Waals forces, which are very weak and only operate over a short range between 3-6 angstroms.”

Experiments run to evaluate the interaction of graphene-containing nanodiamond particles with DLC confirmed the findings from the simulation studies. Superlubricity was observed over a range of loads, sliding rates at which the surfaces moved against each other and temperatures. 

But once the relative humidity increased to 30%, the coefficient of friction increased out of the superlubricity regime. Deshmukh says, “Our simulations found that water occupied the defects in the graphene leading to a more ordered material that suppresses the formation of scrolls.”

The researchers then scaled up the simulations so that they can occur on the macroscale. Deshmukh says, “We did simulations with 100 graphene flakes and 100 DLC nanoparticles and found that scroll formation takes place leading to superlubricity. Of interest is that not all of the nanoparticles need to form scrolls. In our simulations, we need to achieve a certain threshold of around 60% scroll formation to develop superlubricity at the mesoscale.”

This simulation marks the first time that superlubricity has been seen at the macroscale.

Current work is being done to determine how nanoparticle type, size, geometry and shape affect superlubricity. Sankaranarayanan says, “We will also be trying to understand computationally how superlubricity can occur in the presence of moisture. Water functionalizes surfaces by making them more hydrophilic. Use of very hydrophobic materials such as polytetrafluoroethylene may be needed to develop the conditions needed for superlubricity.”

Additional information can be found in a recent article (2) or by contacting Sankaranarayanan at ssankaranarayanan@anl.gov or Deshmukh at sanketkumar007@gmail.com. 

REFERENCES
1. Canter, N. (2015), “Correlating modeling and experimental atomic friction results,” TLT, 71 (8), pp. 10-11.
2. Berman, D., Deshmukh, S., Sankaranarayanan, S., Erdemir, A. and Sumant, A. (2015), “Macroscale superlubricity enabled by graphene nanoscroll formation,” Science, 348 (6239), pp. 1118-1122.


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.