Comment on “Nonlinear current-voltage curves of gold quantum point contacts” [Appl. Phys. Lett. 87, 103104 (2005)]

In a recent Letter [Appl. Phys. Lett. 87, 103104 (2005)], Yoshida et al. report that nonlinearities in current-voltage curves of gold quantum point contacts occur as a result of a shortening of the distance between electrodes at finite bias, presumably due to thermal expansion. For short wires, the electrode displacement induces a thickening of the wire, as well as nonlinearities of the IV curve, while the radius of long wires is left unchanged, thus resulting in a linear IV curve. We argue here that electron shell effects, which favor wires with certain "magic radii," prevent the thickening of long wires under compression, but have little effect on wires below a critical length.