Glass transformation in vitreous As2Se3 studied by conventional and temperature-modulated differential scanning calorimetry

We have studied the glass transition behavior of vitreous As2Se3 by carrying out temperature-modulated differential scanning calorimetry (TMDSC) and conventional differential scanning calorimetry (DSC) experiments to measure the glass transition temperature Tg. In TMDSC experiments we have examined the reversing heat flow (RHF), that is the complex heat capacity CP in the glass transition region as the glass is cooled from a temperature above the glass transition temperature (from a liquidlike state) and also as the glass is heated starting from room temperature (from a solidlike state). The RHF, or CP versus T, in TMDSC changes sigmoidally through the glass transition region without evincing an enthalpic peak which is one of its distinct advantages for studying the glass transformations. The Tg measurements by TMDSC were unaffected by the amplitude of the temperature modulation. We have determined apparent activation energies by using Tg-shift methods based on the Tg-shift with the frequency (ω) of temperature modulation in the TMDSC mode and Tg-shift with heating and cooling rates, r and q, respectively, in the DSC mode. It is shown that the apparent activation energies ∆h* obtained from ln ω versus 1/Tg and ln q versus 1/Tg plots are not the same, but nonetheless, they are approximately the same as the apparent activation energy ∆hn of the viscosity over the same temperature range where the empirical Vogel expression of Henderson and Ast, η = 12.9 exp[2940/(T - 335)], was used for the viscosity. The latter observation is in agreement with the assertion that the structural relaxation time Ʈ is proportional to the viscosity h. The apparent activation energy ∆hr obtained from the ln r versus 1/Tg plot during heating DSC scans is lower than ∆h* observed during cooling scans. The results are discussed in terms of a phenomenological Narayanaswamy type relaxation time. It was observed that Tg obtained from TMDSC cooling experiments did not depend on the underlying cooling rate for q ≤ 1 °C min-1; and for temperature amplitudes 0.5–5 °C. The transition due to the temperature modulation was well separated from the transition due to the underlying cooling rate. Further, the apparent activation energies obtained from ln ω versus 1/Tg during cooling and heating scans for q and r ≤ 1 °C min−1 are approximately the same as expected from Hutchison's calculations using a single relaxation time model of TMDSC experiments.