Long-term thermal energy storage balances the seasonal variations in renewable energy supply and demand, but applied storage concepts require improved performance in efficiency, reliability and capacity. In principle, supercooling and cold-crystallization offer a way to store heat for an extensive amount of time. In this approach, crystallization behaviour of the material governs the storage performance, as it directly relates to optimal efficiency, length of the storage period and heat release properties. This work explains the unique cold-crystallization behaviour of erythritol in cross-linked sodium polyacrylate. To this end, isothermal cold-crystallization was measured experimentally and analysed with the Avrami equation. Although the cold-crystallization rate constant follows the Arrhenius equation, it drastically decreases near the glass transition region and diverges from the equation. Thermal history also influences the cold-crystallization behaviour. Increases in cooling end-temperature reduce the subsequent crystallization time and promote metastable polymorph formation. These findings stem from the peculiar energy landscape of erythritol in cross-linked sodium polyacrylate. The landscape is classified as kinetically strong and thermodynamically fragile, which facilitates long-term thermal energy storage. Consistent supercooling and cold-crystallization behaviour of the material enables predicting the time-dependent crystallization rate at different temperatures. This confirms applicability of the two-stage Arrhenius-VFT model for temperature dependence and supports storage design in real-life applications. • Cold-crystallizing PCM for long-term TES with 230–350 MJ/m 3 storage capacity. • Spherical cold-crystallization enables modelling and predicting the kinetics. • Cold-crystallization follows Avrami equation and combined Arrhenius-VFT dependency. • Crystallization rate decreases drastically when approaching glass transition region. • Reducing cooling temperature increases following isothermal crystallization rate.
Read full abstract