Critical assessment: Martensite-start temperature for the transformation

It is increasingly important in the context of high-manganese steels of the kind that lead to twinning-induced plasticity to be able to estimate the temperature at which ε -martensite forms when austenite is cooled. We find that the thermodynamic method used in similar calculations for α ′ martensite cannot in many cases be implemented because of apparently imprecise thermodynamic data, a conclusion partly validated using limited first-principles calculations. Alternative, empirical methods are also evaluated. The austenite grain size dependence of the martensite-start temperature has also been rationalised in terms of existing theory for α ′ martensite. Experiments have also been conducted to show that the problem in dealing with the ε -martensite does not lie in the precision with which the transformation can be measured using dilatometry.