While “super-sizing” seems to be the driving force of our food industry, the direction of materials research has been quite the opposite:​ the dimensions of many technological devices are becoming ever smaller each year. This continuous reduction in size drives a great demand for understanding the mechanical behaviour, and in particular, the strength of materials at the sub-micron scale. In bulk form, the yield stress and strength of the material remain nearly constant regardless of the sample size because the sample dimensions are large compared to the length scale characterizing the material’s microstructure. However, when the geometries of critical dimensions on a device approach the size of material’s microstructure, the size effects prevail, and the bulk properties can no longer be used to predict mechanical behaviour.

Pure metals, as well as some alloys, have been found to exhibit strong size effects at the sub-micron scale:​ smaller samples consistently yield at higher stress levels. In many earlier experimental studies, the size effects in indentation, torsion and bending were understood in terms of plastic strain gradients that create geometrically-necessary dislocations leading to hardening. Even without the presence of strong strain gradients, the strengths of thin films on substrates are typically found to scale inversely with film thickness. The higher strengths observed in thin metal films relative to their bulk counterparts are usually attributed to the confinement of dislocations within the film by the substrate and in some cases, the passivation, as well. While these studies of thin films constitute important size effects on plasticity, they all arise from the constraining effects of surrounding layers. Another set of size effects has been observed for crystals that are initially dislocation-free. Classic experiments in the 1950’s showed that initially pristine metal whiskers yielded at nearly theoretical strengths. In addition, in the earliest stages of nanoindentation, the crystal volume being probed is extremely small and can be dislocation-free. Therefore, in the initial stages of deformation very large indentation stresses are needed to nucleate new dislocations, which leads to a size effect for plastic flow. Finally, several molecular dynamics simulations also agree with the tenet that smaller is stronger. However, in spite of much progress on size effects for plasticity there is still no unified theory for plastic deformation at the sub-micron scale.

In the work presented here, the attention is focused on size effects that arise in unconstrained geometries, in the absence of strong strain gradients, and with non-zero initial dislocation densities characteristic of well-annealed crystals. In this work, gold nanopillars ranging in diameter between 200 nm and several microns were fabricated using Focused Ion beam (FIB) machining with Ga+ ions and microlithography followed by electroplating. These small pillars were found to plastically deform in uniaxial compression at stresses as high as 800 MPa, a value ~50 times higher than for bulk gold. We believe that these high strengths are controlled by the process of hardening by dislocation starvation, unique to very small crystals. In this mechanism, the mobile dislocations have a higher probability of annihilating at a nearby free surface than being pinned by other dislocations. When the starvation conditions are met, plasticity is accommodated by the nucleation and motion of new dislocations rather than by motion and interaction of existing dislocations, as is the case for bulk crystals