Session: 08-09-01: Thermal Energy Storage

Paper Number: 99532

99532 - Statistical Analysis of Liquid-to-Solid Nucleation of Magnesium Chloride Hexahydrate for Prediction of Supercooling in Thermal Energy Storage Applications 

Phase change materials (PCMs) have a variety of industrial applications such as thermal energy storage, thermal management, and dry cooling of power plants. While most studies have focused on improving thermophysical properties of PCMs, e.g., reducing a supercooling temperature or enhancing thermal conductivities with additives, one of the major bottlenecks for the implementation of PCM in the industry has been the stochastic supercooling behavior of PCMs. Previously, we proposed a framework, by combining stochastic analysis of nucleation and heat transfer within a PCM, to predict the average supercooling temperature associated with the volume and cooling rate of the PCM. Specifically, we demonstrated the validity of the framework with solid-solid phase change of neopentyl glycol (NPG). In this presentation, we report the extension of our framework to more commonly used PCMs such as salt hydrates; here we tested magnesium chloride hexahydrate (MgCl₂6H₂O). We first characterized the stochastic nucleation behavior of liquid-to-solid phase change of MgCl₂6H₂O at three different conditions using differential scanning calorimetry (DSC). Two different mass of 1.4 and 19.8 mg samples were cycled from 80 to 130 ℃ for 150 times with a cooling rate of 10 ℃/min. The 19.8 mg sample went through another 150 cycles at a 1 ℃/min cooling rate to investigate the effect of cooling rate on supercooling behavior. The supercooling temperature (ΔT) was determined as the onset of nucleation temperature, i.e., the first deviation point from the linear baseline of heat flow during a cooling run. DSC results showed that the degree of supercooling increases with a higher cooling rate and smaller mass, which is consistent with our understanding. We characterized a survival function (χ) of supercooling with the DSC data by counting the number of non-nucleated samples at a given temperature; then, the survival function was converted to a nucleation rate (J) as a function of supercooling temperature. We fitted the power-law approximation of nucleation rate (J(ΔT)=γΔTⁿ) with the experimental data, which results in the fitting parameters γ and n of 1.50×10^-9 [1/(Kⁿm³s)] and 12.01, respectively. The fitting line and experimental data showed a good agreement. Given the γ and n values from the fitting, our framework can predict the average and standard deviation of the supercooling temperature of a PCM of any size at any cooling rate. To validate the prediction capability at system-scale applications, we ran cooling experiments using a convection oven for MgCl₂6H₂O with volumes ranging across three orders of magnitude and cooling rates from 0.1 to 10 ℃/min, all of which showed reasonable agreement with our supercooling prediction. Further, in regards to the standardization of our framework for general PCMs, we propose a standardized protocol to obtain the fitting parameters using DSC and apply them to predict the supercooling of an application of interest. This work provides guidelines for the optimized design of PCM-based applications and has important implications for understanding the stochastic phase-change behavior of PCMs.

Presenting Author: Youngsup Song Lawrence Berkeley National Laboratory

Presenting Author Biography: Youngsup received his B.S. and M.S. in Mechanical Engineering from Yonsei University, where he developed nanomaterial-integrated MEMS devices. During his PhD in Mechanical Engineering at Massachusetts Institute of Technology (MIT), he focused on mechanistic understanding and enhancing pool boiling heat transfer via surface property and structure design. Prior to his PhD study, he also worked at Korea Institute of Materials Science (KIMS) in the Electrochemistry Department.

Authors:

Youngsup Song Lawrence Berkeley National Laboratory
Drew Lilley Lawrence Berkeley National Laboratory
Sumanjeet Kaur Lawrence Berkeley National Laboratory
Ravi Prasher Lawrence Berkeley National Laboratory