Hyperstabilization of martensites

The effect of hyperstabilization of martensite implies that the reverse martensitic transformation proceeds in two well separated stages. Namely, a small fraction of martensite (of the order of 10%) retransforms upon heating into the parent phase over a temperature range slightly higher than the nominal reverse transformation temperature, whereas the rest of the martensite retransforms through a renucleation of fine lamellae of the parent phase. The renucleation stage of the transformation is well defined and requires strong overheating of the order of 300 K with respect to the nominal transformation. In this letter, the results are discussed of a study of the hyperstabilization effect in different martensitic structures: faulted β'1 martensite in Cu–Al–Be system and twinned martensite in ferromagnetic Ni–Fe–Ga crystals by means of differential scanning calorimetry, transmission electron microscopy and internal friction. The conclusion has been drawn that hyperstabilization implies a severe blocking of the motion of interphase boundaries during the reverse transformation, which can be produced either due to a high concentration of highly mobile quenched-in defects ("sweeping" of defects during the reverse transformation) or due to a creation of obstacles by preliminary plastic deformation. The former mechanism requires very intense diffusion of quenched-in defects assisted by dislocations/interfaces, which has been confirmed by internal friction studies. It has been shown that the renucleation stage, which occurs at around 600 K for different alloys, is preceded by a relaxation internal friction peak. A possible role of this relaxation in renucleation of the parent phase is discussed.status: publishe