The most recent Proceedings of the National Academy of Sciences (Vol. 104, No. 4, pp 1130-1133) has a free access article which is visually striking, although I was really disappointed in the analysis. PDMS was exposed to a beam of Ga+ ions, which then forms a stiff upper layer (SiOx - a depletion of carbon is noted) atop the softer underlayer, with the upper layer forming a buckled surface at high enough Ga+ fluences.
Why? The authors are silent on the subject.
Having a stiff layer of anything over a softer layer doesn't mandate buckling unless there is a difference in stress/strain that is allowed to relax. I suspect that that is the ultimate cause here - residual stress from the manufacturing of the original PDMS. This would be easy to verify qualitatively.