Comments on "Thermal decomposition of pyridoxine: an evolved gas analysis-ion attachment mass spectrometry study". About the application of model-fitting methods of kinetic analysis to single non-isothermal curvesIn a paper recently published in this journal, [1] the thermal decomposition of pyridoxine was studied by means of evolved gas analysis-Li + ion attachment mass spectrometry (EGA-Li + IAMS), which is a technique with great potential for the kinetic study of thermal decomposition reactions. In that paper, the decomposition products were identified and, in addition, a kinetic analysis of the process was carried out. For that purpose, these authors used the total ion monitoring (TIM) curve, which indicates the amount of products released at any given time and, consequently, can be related to the rate of reaction as a function of time or temperature. The extent of conversion at a certain time or temperature was then determined by integrating the area under the TIM curve from the beginning of the signal until the selected time/temperature. [1] With that information, the kinetic parameters driving the reaction, i.e. the activation energy and the pre-exponential factor, were calculated by fitting the experimental data extracted from the TIM curve to a first-order kinetic model. At a first glance, this was a standard model-fitting method of kinetic analysis but, as employed, it had two significant flaws that we aim to clarify here: performing the analysis with data from a single nonisothermal experimental run, and fitting the data to a first-order model, without testing other possible models.It is known that a reaction can be kinetically described by three parameters: the activation energy (E a ), the preexponential factor (A) and the kinetic model. These constitute the so-called kinetic triplet. The kinetic model is an algebraic function that reflects the relationship between the reaction rate and conversion and it can be related to a reaction mechanism. During the last decade a number of theoretical kinetic models have been proposed. [2][3][4] Model-fitting methods of kinetic analysis are widely used because of their simplicity and they basically consist of fitting the experimental data to several such kinetic models. The model providing the best fit is usually regarded as the correct one, and the activation energy is deduced from the slope of the fit. Unfortunately, it has been long established that the kinetic parameters cannot be reliably determined from a single non-isothermal curve because the experimental data provide a reasonable fit regardless of the kinetic model employed. [5][6][7][8] As an example, Fig. 1 shows a simulated a-T curve, constructed using the following kinetic parameters: E a = 100 kJ mol -1 , A = 10 10 s -1 and an Avrami-Erofeev A2 kinetic model. The simulation was performed using a Runge-Kutta 4 th order numerical integration method, programmed using the Mathcad engineering calculations software (Mathsoft; PTC, Needham, MA, USA). For the sake of comparison, the simulated curve has bee...