Stiction is a problem in nano- and micro-electromechanical systems (NEMS and MEMS) whereby the tiny components stick together, often greatly reducing the reliability and long-term durability of these devices. It occurs when capillary, van der Waals and electrostatic forces between surfaces overpower the built-in restoring forces of the overall structure. In smaller systems the effect is more pronounced because of a larger surface-area-to-volume ratio.

Practical methods to eliminate stiction-related failures involve designing devices with high mechanical restoring forces, or by using “passivants” - special treatments for reducing surface energy. Some researchers have employed time consuming and costly molecular-dynamics simulations to see how roughness affects stiction, but so far these have not given any useful insights. Liu and co-workers, however, have performed experiments that demonstrate a correlation between surface roughness and stiction - a result that they hope could be used to minimize stiction in MEMS, and possibly NEMS, components.

The US team started with a series of silicon wafers, each with a different average roughness - that is, with different sized lumps or “asperities” on the surface. They then brought the cantilevers of an atomic force microscope with various tip radii into contact with a single asperity on the surface and measured the size of the adhesive force.

The researchers found that the adhesive force falls quickly as the average roughness is increased, but reaches a minimum beyond which it steadily rises again - in other words, there exists an optimal roughness. This value increases with the radius of the cantilever tip.

According to Liu and co-workers, this is because a tip resting on a completely smooth surface is strongly attracted by the majority of the surface’s adhesive forces. A small asperity on a slightly roughened surface acts to distance the tip from the surface, so these forces are less pronounced. But too large an asperity on a very rough surface will have its own strong adhesive forces which cancel the distancing effect.

“Our work suggests a promising way to minimize adhesion between two surfaces by tuning asperity height to feature-size in MEMS devices,” Liu told “We didn't quantify by how much stiction could be reduced, but our model can provide a useful predictor of the behaviour of adhesive contacts down to the nano scale.”