New Scientist

[New Scientist, vol. 160 no. 2156, p.30(1998)]
(This article was taken from New Scientist magazine Copyright New Scientist, RBI Ltd 1998)

There's the rub

Friction is everywhere-sometimes welcome but often not. So isn't it time we got to grips with it? Charles Seife and Ben Crystall investigate

WITHOUT friction, life as we know it would be impossible. Raising a glass of wine or a fork of spaghetti to your lips would be a distant dream. And you could forget about driving your car down the street-even walking across the room would be tricky. Without the inherent stickiness of surfaces when they rub against each other, we'd get absolutely nowhere.

As vital as friction is, nobody really knows how it works. Sure, we may have an intuitive feel for it. Everyone knows that an engine grinds to a halt without oil, and that a blob of butter helps remove a ring from a finger. But when it comes to understanding friction at the tiniest scale-where atom meets atom-we know surprisingly little.

That, however, is beginning to change. With the help of incredibly sensitive equipment, researchers are working out just what makes things cling together or slide effortlessly apart. And to explain their often unexpected results, they've had to come up with weird theories. Some suggest that things stick together because of imaginary particles (see "Virtually friction", p 33). Others blame it all on waves or claim that electrons behave like minute anchors.

Unravelling these mysteries could allow us to control friction, not simply by adding lubricants, but by tweaking the properties of a surface. It might even be possible to harness friction in strange new ways, such as building minuscule "conveyor belts" that could carry a single molecule, or triggering specific chemical reactions by rubbing two surfaces together.

For hundreds of years, tribologists, as those who study sliding surfaces are called, suspected that friction was caused by the roughness of surfaces. Just as it's difficult to slide two sheets of sand-paper over each other, perhaps the atomic-scale pits and knurls on two surfaces catch together, preventing them gliding smoothly past each other. But in the 1950s, when tribologists took a closer look, the idea fell through. It turns out that you can sometimes reduce the friction between two surfaces by making one of them even rougher.

If roughness isn't always responsible, what about chemical bonding? When two perfectly smooth crystal surfaces are pressed together, they may "cold weld" as their atoms bond. To slide the surfaces apart, the bonds must be torn asunder. Could this be the cause of friction?

The only way to find out is to look at what individual atoms do as they move past each other. So in the 1970s, Jacob Israelachvili, a tribologist now at the University of California at Santa Barbara, designed a device to study friction on the tiniest scale-the surface-force apparatus. And what Israelachvili learnt caused consternation. It turns out that even when surfaces don't bond together, there's still plenty of friction between them.

If it's not roughness or molecular bonding, what does cause friction? In the 1980s, Gary McClelland at IBM's Almaden Research Center in California and Jeffrey Sokoloff of Northeastern University in Boston independently suggested an answer: sound waves.

This may seem bizarre, but it's not the waves themselves that cause friction. Rather, it is the energy lost creating the vibrations. When something rubs against a crystal of mica for instance, the atoms of the crystal are "plucked" like the strings of a guitar (see Diagram). "As one layer moves across another, you make the atoms move," says Mark Robbins, a physicist at Johns Hopkins University in Baltimore. The atoms start bobbing back and forth as vibrations, or phonons, travel along the surface of the crystal. The energy of motion of one layer is turned into waves-and the more energy is lost, the harder it is to move two surfaces over each other.

The phonon theory has exciting possibilities. Just as the strings of a guitar vibrate only at certain frequencies, so you can play only certain "notes" on a crystal. What's more, just as acoustic engineers can design auditoriums that enhance sounds or stifle them, a physicist can theoretically design a pair of crystals so that one can't generate phonons in the other. "You can have an almost loss-free system," says Jacqueline Krim, a physicist at North Carolina State University in Raleigh.

Both Krim and Robbins have already shown that cooling a material or increasing the density of its surface-squeezing more atoms into the same space-makes the surface stiffer and reduces the number of vibrational frequencies that are able to pass through it. "You end up getting fewer phonons," says Robbins. By increasing the density of a layer of krypton atoms on a metal surface, for instance, they decreased the friction between the krypton and the metal by a factor of seven.

Each new experiment and theoretical calculation seemed to confirm that phonons are responsible for friction when atoms rub against atoms. But six years ago, Mats Persson, a physicist at the Chalmers University of Technology at Gothenburg, Sweden, suggested that electrons might also be involved.

The electron clouds that surround the nucleus of each atom fluctuate randomly. For example, a dense-more negative-part of the cloud will repel the electron cloud on another atom, inducing a positive charge. These will then attract each other. Such forces are known as van der Waals forces.

Persson suggested that over a very short range, the free electrons in the surface of a metal might be attracted by van der Waals forces from an adjacent surface. Move this surface, and the electrons in the metal will be dragged along with it. This creates a weak current in the surface-and, as with most electrical currents, there will be resistance as the moving electrons are impeded by the atoms or tiny imperfections they meet. They behave like tiny anchors, causing drag (see Diagram).

Testing times

It was an exciting idea, but was it right? Krim had just the tool to test this hypothesis: a quartz microbalance. This sensitive apparatus can detect minute frictional forces between a quartz electrode and a layer of material just atoms thick.

Like the quartz crystal in a digital watch, the quartz electrode vibrates back and forth under the influence of an electric field. By measuring its resonance frequency and seeing how it changes when various materials are deposited on the quartz, Krim can work out how much friction there is between the layers. In an experiment last year, Krim covered the quartz with lead and deposited a layer of nitrogen atoms on top (see Diagram). The idea was to cool the lead to below 72 K, the temperature at which it starts to become superconducting. Cooling the quartz crystal shouldn't change its structure much, but as the lead's resistance falls, any electronic component to friction should disappear too.

But when Krim cooled the lead to less than 72 K, the friction felt by the nitrogen layer suddenly dropped by half. Either the way the phonons travel through the lead changes, or the electronic component of friction is greater than expected-or something even stranger is happening. "Nobody knows for sure," says Israelachvili.

Despite the puzzle, Krim is convinced that electrons do contribute to friction. For her, the only question is, by how much? Scott Perry, a chemist at the University of Houston, thinks that in some materials, the electronic component is substantial. By covering vanadium carbide with oxygen atoms, he recently found he could halve the friction it experienced. He believes the effect is due solely to electrons. "The oxygen bonds to the vanadium and ties up the electrons involved in conduction," he says. With fewer free electrons to anchor the surfaces together, there is less friction between them.

Krim's and Perry's experiments have not only shed new light on friction, but they also raise the prospect of changing the frictional properties of a conducting surface with the flick of a switch. For example, changing the electrical resistance of a surface could reduce or increase friction.

What's more, if you turn the whole concept on its head, you could have a way to make nanoscale conveyor belts. "If there's a frictional drag force in one material due to electrons, then the reverse is true-if electrons move, then they'll exert a force on a material," says Krim.

Pass a current one way, and the electrons will pull atoms on the surface with them. Reverse the direction of the current, and the atoms should move in the opposite direction. "Since electronic interactions are so very short range, my feeling is that it will only work for atoms, small molecules, lightweight particles or little three-dimensional structures," Krim says. She is investigating the effect at present, and hopes to have a working device within a couple of years. "There's absolutely nothing fanciful about it," she says.

Learning more about the flow of energy during friction might even offer chemists new ways to trigger reactions at a surface. When two materials rub together, small regions at their surfaces become incredibly hot. The phonons generated by friction interfere with each other or are scattered, making atoms vibrate randomly-which is what heat is. "Normally to catalyse the reaction of ethylene on a platinum surface, you must heat the platinum to 500 C," says Krim. But similar reactions can occur between two surfaces when they are rubbed together, even though the temperature of the bulk of the material is much lower.

No one really knows how hot the interface gets. But the hope is friction could be used not just as a way of heating surfaces, but to trigger specific reactions. Most reactions begin when one molecule is split in two, allowing the bits to recombine with other atoms or molecules. To break a specific chemical bond requires exactly the right amount of energy, but simply heating a catalyst makes its atoms vibrate at all sorts of different energies, so it's difficult to determine which bond is broken.

However, if it were possible to make a surface vibrate at just the right frequency-by rubbing together crystals through which only certain phonons could pass-chemists might be able use friction to break specific bonds and promote the reactions they choose. "It could allow us an enormous control over surface chemistry," says Krim.

It may be a while before such applications find commercial uses. But there's one breed of machine for which our new knowledge of friction could be vital-the tiny mechanical devices called microelectromechanical systems (MEMS). With the smallest devices just a few hundred atoms across, even the tiniest amount of friction is a killer, says Roya Maboudian, a chemical engineer at the University of California at Berkeley, who is building tiny gears, motors and engines from silicon (see "Invasion of the micromachines", New Scientist, 29 June 1996, p 28).

One way to defeat friction is to cover these minute devices with specially designed coatings. Maboudian has already found that a single layer of long-chain hydrocarbons bound to the surface of a miniature gear can reduce friction a thousandfold. Since the molecules form an ordered array, it becomes hard for phonons to move through the layer. And, according to calculations by Perry, as you lengthen the carbon chains in the layer, the molecules will pack closer, support fewer phonons and reduce friction even further.

On the atomic scale, then, controlling friction by changing the stiffness of a surface or its electronic properties seems to be relatively easy. Researchers have even created surfaces that for all practical purposes are friction free. But extending this understanding to the macroscopic scale is another thing.

The problem is that atomic-scale tests reveal only part of the picture. "In the real world, two objects touch each other at lots of tiny contacts-billions of them," says Krim. Push a mug along a table and the contact points between the two change shape. Stop pushing it and they change again. "If you put a book on a table, do you really know where the atoms are touching?" she asks. "It's hard to get a probe in there."

To study friction at every level, what tribologists really need is a way to make the link between the nanoscale world and real machinery. And that could happen shortly. The micromachines that are benefiting from the latest research could lead to a new generation of devices capable of measuring friction at a variety of scales. Maboudian has already succeeded in making one such tool-a tiny comb-like structure that slides across a surface when a small voltage is applied across it. These devices could provide some useful answers, says Krim. "And when we understand a bit more, almost anything could be possible."

Virtually friction

One of the many strange consequences of quantum mechanics is that every place in the Universe is seething with particles that wink in and out ofexistence even in the deepest vacuum. These virtual photons can collide with objects, passing on their momentum to them. But since objects are struck from all sides, these forces cancel out, so virtual particles have no effect on their surroundings.

But between two parallel plates in a vacuum, it's a different situation, says John Pendry, a physicist atImperial College, London. Pairs of virtual photons bounce between the two plates ("Running on empty", New Scientist, 25 April, p 36). If the plates aren't moving, the pairs will have equal and opposite momentum-they bounce in a symmetric way.

But if the surfaces are sliding over each other, the situation will change. Some photons will be red-shifted and others will be blue-shifted, so the forces they produce will not cancel out. The resulting forces tend to act against the direction of relative motion, says Pendry, slowing the plates' slide. And in many cases, he believes, "quantum friction" may be as important as electronic or phonon-based friction.

Further Reading