A comparison of isoconversional and model-fitting approaches to kinetic parameter estimation and application predictions

Abstract

A variety of isoconversional and model fitting approaches, all of which use multiple heating schedules, are used to analyze selected data from the ICTAC kinetics and lifetime projects as well as additional simulated data sets created for this work. The objective is to compare the accuracy and suitability of various approaches for various types of chemical reactions. The various simulated data sets show that model fitting and isoconversional methods have comparable reliability for extrapolation outside the range of calibration. First, there is as much variability in prediction for various isoconversional methods as there is between isoconversional methods as a group and different plausible explicit models. Of the three isoconversional models investigated, the Friedman method is usually the most accurate. This is particularly true for energetic materials that have a drop in apparent activation energy in the latter stages of reaction, which leads to a delayed onset of rapid autocatalysis at lower temperatures. It is difficult to determine a priori whether isoconversional or model fitting approaches will give more accurate predictions. The greatest reliability is attained by using both the isoconversional and model fitting approaches on a combination of isothermal and constant heating rate data.