“…[9] These approaches are based on some phenomenological assumptions: the linearization of the relaxation process using the reduced time concept, [7] an asymmetric distribution of relaxation times [discrete for KAHR, continuous (through the stretched exponential function) for TNM] to account for the non-exponential character of relaxation, the introduction of the fictive temperature [10] as structural parameter to account for non-linearity, and the generalization of the timetemperature superposition principle. Whilst each of the assumptions of the TNM model can be criticized, [11,12] this approach is generally able to reproduce the differential scanning calorimetry (DSC) scans recorded on heating the sample after annealing in the glass. On the other hand, clear discrepancies have been observed, notably for polymers, when several thermograms were considered for simultaneous fitting.…”