Mass diffusivity and thermal conductivity estimation of chloride-based salt hydrates for thermo-chemical heat storage: A molecular dynamics study using the reactive force field
Summary¶
Thermochemical heat storage in chloride salt hydrates couples dehydration and hydration kinetics to heat and mass transport inside grains and at surfaces. Pathak et al. develop transferable ReaxFF models for MgCl₂ and CaCl₂ hydrates spanning multiple hydration states so that reactive molecular dynamics can estimate water diffusivity and thermal conductivity during cycling. The study targets MgCl₂·nH₂O with n from 1 to 6 together with CaCl₂ hydrate models described in the paper, using the same methodological backbone to compare the two salts on equal footing—motivated by physical mixtures proposed for improved usability in thermal storage systems. The introduction argues that reactor design requires consistent transport coefficients across hydration states encountered during seasonal charging and discharging.
Methods¶
1 — MD application (ReaxFF-MD + SS-NEMD). Int. J. Heat Mass Transfer 149, 119090 describes ReaxFF-MD (reactive force field molecular dynamics) in LAMMPS-class workflows: H₂O diffusivity is computed through MgCl₂·4H₂O and MgCl₂·6H₂O and compared to prior n = 0–2 work; spatial diffusivity is analyzed from bulk toward the surface to probe dehydration-relevant surface vs bulk effects. Steady-state non-equilibrium MD (SS-NEMD), using the same reactive model, gives thermal conductivity of MgCl₂·nH₂O (n = 0, 1, 2, 4, 6) and CaCl₂·2H₂O; cross-section and length along the heat-flow direction are noted in the article as size-sensitive for κ. Engine: LAMMPS with ReaxFF (as stated in the work). System, PBC, unit cells: hydrate supercells per Methods/SI. Ensemble, timestep, equilibration, production, SS-NEMD thermal gradient, thermostats: full tables in the PDF+SI; N/A to re-list every value here. Barostat for any NPT segments: as in the article. Shear, shock, static external E-field, umbrella, metadynamics: N/A for the reported transport setup unless the SI states otherwise.
2 — Force-field training (ReaxFF, chloride hydrates). The paper proposes new transferable ReaxFFs extending MgCl₂·nH₂O to n = 4, 6 and a new CaCl₂·nH₂O model (n = 0, 2 in the text abstract), with DFT/QM reference data, training structures, and ParReaxFF/least-squares-style optimization as in IJHMT+SI. Reference expt/DFT comparisons for D and κ are part of the validation.
3 — Static QM / DFT-only as the sole result of the study. N/A; transport comes from RMD/SS-NEMD.
4 — Review or experiment-only. N/A.
Findings¶
Mechanisms and transport trends¶
Water diffusivities in MgCl₂·nH₂O span ~10⁻¹¹–10⁻⁹ m²/s, comparable to experiment. Surface effects are minor for MgCl₂·6H₂O dehydration in their treatment but matter for other hydrates. Thermal conductivities increase with hydration (~0.3–0.9 W/(m·K) across the MgCl₂ series); MgCl₂·6H₂O shows strong thermal anisotropy.
Mixed salts and durability¶
MgCl₂/CaCl₂ blends need not degrade bulk κ for the composite but can alter hydrolysis chemistry relevant to durability—separating thermal vs chemical design concerns.
Future workflow (as demonstrated)¶
Parameterize hydrates → validate D → add NEMD without changing the reactive core, easing mixed-salt extensions.
Limitations¶
Reactive FF transferability should be reassessed for interfaces, dopants, and extreme temperature cycles outside training.
Relevance to group¶
Extends ReaxFF into chloride hydrate thermochemical storage with van Duin-group parameterization involvement.