A critical reflection upon the application of the Arrhenius model to non-isothermal thermogravimetric curves