Session: Research Posters

Paper Number: 173001

Multiscale Characterization of Epsomite Powder-Bed Thermal-Convective Dehydration: Metamorphoses and Kinetics Modeling 

Thermochemical material (TCM) storage utilizes chemical bond energy for endothermic dehydration and exothermic hydration reactions, storing and releasing heat. Among TCM hydrated salt, Epsomite (MgSO₄-7H₂O) has a hydration chemical-bond energy around 0.7 eV, compared to 0.39 eV for the heat of water evaporation (associated with hydrogen bonding). This makes TCM a compact and high-energy-density solution for thermal energy storage. For example, in seasonal thermochemical energy storage systems, over the summer, dry hot air is blown into the reactor to dehydrate the Epsomite and charge the storage tank, while in the winter, humid air rehydrates the salt and releases the stored heat. In electric vehicle technology, Epsomite shows potential for mitigating EV battery thermal runaway due to its large energy storage density, non-toxicity, and availability.

Structural changes such as clotting (caking) and volume changes occur during both dehydration and hydration due to high vapor pressure and large reaction rate. The Epsomite clogs the pores in the particle bed, reducing the bed permeability and blocking the airflow and convection. Thus, understanding the operating limits, e.g., temperature, humidity, and air flow velocity is essential. Furthermore, existing reaction kinetics models that assume constant activation energy are not supported by experiments.

This research conducts multiscale investigation of Epsomite including examining the caking behavior of the packed bed, tracking morphological transitions across dehydration stages, and proposing empirical reaction kinetics for numerical simulations. At the atomic scale, first-principles calculations are performed to evaluate the crystal density, reaction enthalpy, and crystal structure (bond types and lengths). Experimentally, hot air is passed through Epsomite packed bed to measure and analyze its permeability, thermal behavior, and dehydration reaction conversion. At the continuum scale, CFD simulations use a newly developed dehydration-completion activation energy relation validated against the experimental results. Dehydrated samples are also analyzed using optical microscopy and X-ray microtomography (XRM).

With the large surface particle surface area and convective removal of water vapor molecules, significant heat can be absorbed rapidly. However, experimental results show that water vapor release can cause particle caking and significant loss of permeability. The dehydration kinetics reveals a completion-dependent activation energy ΔE_a(n), as n is the number of water molecules bound to one MgSO₄ unit. As n decreases from 7 to 0 during dehydration, activation energy remained almost unchanged in the early stage and then increases with completion. The XRM images show that dehydration of particles results in formation of pore within the particle and diffusion of vapor through the cracks (channels) formed on the shell. High heating rate results in dissolution of the outer portion of the shell by pore water, and this results in recrystallization and caking. A controlled heating scenario, starting with moderate temperature air permeation and increases later, enables complete dehydration without caking. With constant heating rate, an inlet temperature of over 110^oC and a velocity of 1.5 m/s, the particle bed exhibits caking (cementing) and significant loss of permeability.

Presenting Author: Fan Lu University of Michigan

Presenting Author Biography: Fan is a PhD candidate at University of Michigan focusing on renewable energy and heat transfer in multi-scale. He has publications on wet pipe insulation, oscillating heat pipe (a high-performance heat transfer device), and solutions for battery thermal runaway. Currently he is studying heat transfer in microscale and hopefully will reveal more physics about heat transfer and renewable energy applications.

Authors:

Fan Lu University of Michigan
Massoud Kaviany University of Michigan
Suhasini Ramshankar University of Michigan
Erik Yen General Motors
Peter Andruskiewicz General Motors