An experimental study of the specific heat at the glass transition during cooling

Abstract

In the preceding chapters we have demonstrated the continuous cooling method as a tool for the calorimetric study at the glass transition, bringing out several novel aspects observed in the specific heat measured at the glass transition during cooling. In this concluding chapter, we summarize the important observations, point out unfinished aspects of the investigation and indicate the directions for future work. Summary of the results

We have obtained the specific heat of several organic materials as the supercooled liquid is cooled into the glassy state. In the glass transition region, the specific heat measured during cooling differs from that obtained during the heating of a quenched glass as observed earlier using DSC and adiabatic calorimetry. The specific heat obtained during continuous cooling does not show the peak in the glass transition interval which is obtained in the heating experiments.

From the analysis of the specific heat data from our experiments using a phenomenological model, we have obtained the structural relaxation parameters. This analysis has led to the following important observations:

The equilibrium configurational specific heat seems to follow a relation of the type a+b/Ta + b/Ta+b/T for temperatures down to and slightly below TgT_gTg?. This rise indicates that the drop in the specific heat observed at the glass transition is purely relaxational in origin. We observe that near or below the glass transition, the Vogel-Fulcher (VTF) law may not be strictly valid. The parameter T0T_0T0? in the VTF equation becomes smaller as we go down to lower temperatures. This behavior of T0T_0T0? seems to suggest a tendency of the relaxation times to move towards Arrhenius behavior as has also been observed in some earlier studies on other glassy systems. Our analysis brings out the role of the delayed enthalpy relaxation ("the trapped heat release") in determining the observed specific heat.

In our specific heat measurements on partially crystallized glasses, we have observed the effect of partial crystallization on the structural relaxation in the residual vitreous phase. We find that in the glass transition region, the specific heat of a partially crystallized glass is not the same as the specific heat of a simple mixture (taken in the appropriate ratio) of glass and crystal. The annealing process leading to crystallization changes the relaxational parameters of the uncrystallized phase as seen in our measurements.

A significant observation in this study is the broadening of the glass transition interval in the later stages of crystallization. We suggest that this may be related to the 'size dependence' of the relaxational behavior. A wider spectrum of relaxation times observed in the later stages of crystallization may be related to the decrease in the size of the residual uncrystallized regions and the resulting heterogeneity. We have also observed a correlation between the width of the glass transition interval and the parameter ?\beta? in the KWW relaxation function. In the partially crystallized glasses, we find that the width of the glass transition interval (?\Delta?) varies linearly with (1??)(1 - \beta)(1??).

The specific heat during cooling shows an interesting dependence on the cooling rate during the measurement. For slower cooling rates, the specific heat is lower in the whole temperature range covering the supercooled liquid, the glass transition interval and the sub-TgT_gTg? regions. This may be a consequence of the dependence of the vibrational specific heat on the configurational state of the system.

In the thermal cycling experiments, we observe an anomalous dip in the specific heat below the glass transition. This dip as seen in our experiments implies an increase in the cooling rate. The anomaly is sensitive to the thermal history and the experimental kinetics. We have done systematic studies of the dependence of the anomaly on annealing above the glass transition. We have suggested the possible origin of this anomaly in the decoupling of a part of the configurational component (of the enthalpy change) from the rest, the relaxation times associated with this component being different (both in magnitude and the temperature dependence) from the average relaxation times in the system. Such a decoupling can act as an effective 'heat sink' internal to the sample. This would lead to an increase in the observed anomaly.