Due to the existence of surface stress, the interior of a free standing nanowire is in non-zero stress state especially near free surface (see attached). This stress state corresponds to strain field in nanowires. A reference configuration would then be necessary to define the strain. Any good suggestions on the selection of the reference configuration?

Copper nanowires are relaxed with axial periodic and lateral free traction B.C.. However, the results give non-zero radial stress Sigma_rr at free surface, which does not valish even after average over the nanowire volume. The radial distribution from nanowire center to free surface of Sigma_rr is as given below. The average Sigma_rr over entire nanowire gives compressive value correlated linearly with inverse nanowire diameter, which makes it not like numerical error. Anybody share some information why is that? The atomic stress is calculated using Virial stress.The nanowire has circular cross-section.