Enthalpy relaxation studies of the glass transition in a metallic glass

The first ever enthalpy relaxation study has been conducted of the glass transition of the bulk metallic glass forming composition Zr65Al10Ni10Cu15 using differential scanning calorimetry (DSC). Both heating and cooling and isothermal annealing studies have been conducted. The glass transition temperature measured at 20 °C/min heating following a 20 °C/min cooling of 384 °C ± 2 °C agrees well with values previously published for this glass. Using the dependence of Tg on heating and cooling rate, the onset Tg activation energy was determined to be 120 kcal/mole ± 10 kcal/mole with the peak Tg activation energy, determined by the temperature of the peak in the Tg overshoot, was observed to be 98 kcal/mole ± 10 kcal/mole. Such increases in the activation energy, while not common for such transition range measurements, is consistent with this glass exhibiting non-Arrhenius enthalpy and possibly viscous relaxation rates. The heating and cooling and annealing DSC data were converted to derivative fictive temperature data and then fit to the Tool–Narayanaswamy–Moynihan–Hodge model utilizing a stretched exponential function to describe the non-exponentiality and the Tool–Narayanaswamy equation to describe the non-linearity of the relaxation. Best fit parameters for x and β were found to be 0.53 ± 0.05 and 0.79 ± 0.05, respectively. The individual fits to the data are not excellent, but the overall fit to all of the data is surprisingly good given the range of the thermal treatments given to this glass. The model fits worse, however, for the isothermal annealing data. The glass appears to anneal faster than the model predicts for relatively short anneals relatively close to the Tg, but appears to anneal much slower than the model predicts for relatively long anneals relatively far from Tg. These behaviors are consistent with this metallic glass exhibiting a non-Arrhenius activation energy that increases sharply with decreasing temperatures.