The influence of the temperature increase rate on the accuracy of diffusion parameters estimated under non-isothermal conditions

The accuracy of the diffusion parameters estimated under non-isothermal conditions was studied using computer-generated pseudo-experimental points. The data sets were obtained from Fick's 2nd law and an Arrhenius-type temperature dependence of the diffusivity. A normally distributed error was randomly attributed to the model data. The pseudo-experimental points thus obtained were then fitted to the model equation, with non-linear regression using the Simplex algorithm, yielding the estimates of the pre-exponential factor and activation energy. Several linear temperature profiles were studied, from 0.014 to 16°C min ¹. It was concluded that the accuracy of the parameters estimation depends on the temperature increase rate. The effects of the pre-exponential factor, the activation energy, the range of temperature and the magnitude of the experimental error on the value of the temperature increase rate that led to maximum accuracy of the parameters estimates were also studied.