“…1 Subsequently, several groups have confirmed that both micron-sized samples of single-crystal and polycrystalline silicon undergo fatigue failure. [2][3][4] However, a recent experiment calls into question these studies by observing similar cracking behavior in both cyclic-and monotonic-loaded micromachined silicons. 5 Kahn et al 6 has compiled and summarized the main experimental observations of silicon fatigue failure as follows: ͑1͒ morphological changes in areas of high stress, including surface roughening and oxide thickening, ͑2͒ a decrease in high cycle fatigue lifetime when the peak stress is increased, ͑3͒ a decrease in low cycle fatigue lifetime when R, the ratio between minimum and maximum stress ͑negative values denote compression͒, is decreased, ͑4͒ a decrease in fatigue lifetime associated with a larger initial flaw, which is assumed to be caused by subcritical crack growth, ͑5͒ no dependence on vacuum or air in fatigue lifetime for low cycle failure, but a significant decrease in lifetime during high cycle fatigue if water is present, and ͑6͒ a fatigue lifetime that depends only on the number of cycles, not the elapsed time or the test frequency.…”