# Multistage Thermal Decomposition Kinetics of Glycidyl Azide Polymer-Based Thermoplastic Elastomers: A Constrained Deconvolution Approach

## TL;DR

This paper introduces a new method to analyze the complex thermal decomposition of a polymer used in rocket propellants, enabling accurate modeling of its behavior under heat.

## Contribution

A constrained deconvolution approach is developed to resolve overlapping decomposition reactions in energetic thermoplastic elastomers.

## Key findings

- The decomposition process was split into five distinct stages with high accuracy (R2 > 0.998).
- Activation energies for each stage were determined without assuming reaction mechanisms.
- The Šesták–Berggren model fit three major stages well, offering mechanistic insight.

## Abstract

Glycidyl azide polymer (GAP)-based polyurethane, a kind of energetic thermoplastic elastomer (ETPE), is a promising binder for advanced solid propellants, but its thermal decomposition involves overlapping competitive reactions that conventional single-step kinetic models cannot characterize accurately, limiting its engineering applications. To address this limitation, a constrained asymmetric Gaussian deconvolution strategy with fixed peak area ratios and shape constraints was developed in this work. This strategy was applied to resolve overlapping reaction rate curves converted from derivative thermogravimetric data of GAP-based ETPEs with 50 wt% GAP content at four heating rates of 5, 10, 15 and 20 K·min−1. The complex decomposition process was successfully split into five stages, assigned to azide cleavage, polyether backbone scission, carbamate cleavage, hydrocarbon product degradation and residue decomposition, with a goodness of fit of R2 > 0.998. Apparent activation energies of the five stages were determined through cross-validation by the Friedman and Flynn–Wall–Ozawa methods without prior assumption of reaction mechanisms, following the order of residue decomposition (181.4 ± 1.0 kJ·mol−1) > hydrocarbon product degradation (159.9 ± 1.0 kJ·mol−1) ≈ azide cleavage (156.5 ± 0.6 kJ·mol−1) > backbone scission (135.1 ± 0.7 kJ·mol−1) > carbamate cleavage (111.9 ± 1.1 kJ·mol−1). Pre-exponential factors with lnA0 values ranging from 22.2 to 34.0 were derived via the kinetic compensation effect. Finally, generalized master plots were employed to compare with classic solid-state reaction models for mechanistic insight, and the Šesták–Berggren model fit three major stages excellently (R2 > 0.996) by accounting for synergistic nucleation-growth and phase boundary mechanisms, enabling high-precision kinetic equations. It should be noted that the constrained deconvolution method proposed in this work has general applicability for kinetic analysis of GAP-based ETPEs with different formulations and other complex energetic polymer systems, while the obtained kinetic parameters are composition-specific and only applicable to the corresponding ETPE formulation studied herein.