Deep Potential Molecular Dynamics Simulation of Local Structure and Properties of LiCl-KCl-CsCl-LaCl3 Molten Salt

Journal of Nuclear Materials, Год журнала: 2025, Номер unknown, С. 155749 - 155749

Опубликована: Март 1, 2025

Язык: Английский

---

Heterogeneous Structure, Mechanisms of Counterion Exchange, and the Spacer Salt Effect in Complex Molten Salt Mixtures Including LaCl₃

The Journal of Physical Chemistry B, Год журнала: 2024, Номер 128(16), С. 3972 - 3980

Опубликована: Апрель 16, 2024

Complex molten chloride salt mixtures of uranium, magnesium, and sodium are top candidates for promising nuclear energy technologies to produce electricity based on reactors. From a local structural perspective, LaCl3 is similar UCl3 hence good proxy study these complex mixtures. As fission products, lanthanide salts their also very important in own right. This article describes from an experimental theory perspective how different the roles MgCl2 NaCl with LaCl3. We find that, whereas becomes integral part multivalent ionic networks, separates them. In recent (J. Am. Chem. Soc. 2022, 144, 21751–21762) we have called disruptive behavior "the spacer effect". Because heterogeneous nature mixtures, there multiple motifs melt, each its particular free energetics. Our work identifies quantifies these; it elucidates mechanisms through which Cl– ions exchange between Mg2+-rich La3+-rich environments.

Язык: Английский

---

Enhanced exploration of LiF–NaF thermal conductivity through transferable equivariant graph neural networks

Journal of Applied Physics, Год журнала: 2025, Номер 137(6)

Опубликована: Фев. 10, 2025

Although molten salt reactors and thermal storage systems are attracting increasing interest, our understanding of the physicochemical properties salts is still incomplete. This largely due to difficulty conducting experiments under extreme temperatures with strict control impurities corrosion. Ab initio calculations, machine-learned force fields, classical molecular dynamics have helped alleviate some these issues. However, discrepancies between experimental theoretical computations conductivity fluoride become concern. In this paper, we present a modernized method for training transferable equivariant graph neural network fields model simple system, LiF–NaF, using minimal ab calculations. Using field, as well various other functions LiF–NaF were computed at chemical ratios in order gain new insights into limitations behaviors relation their conductivity. Results show function temperature but good agreement ratio. Secondary results compelling field first-principles ability interpolate extrapolate ratios.

Язык: Английский

---

Deep Learning Potential Assisted Prediction of Local Structure and Thermophysical Properties of the SrCl₂–KCl–MgCl₂ Melt

Journal of Chemical Theory and Computation, Год журнала: 2024, Номер 20(17), С. 7611 - 7623

Опубликована: Авг. 28, 2024

The local structure and thermophysical properties of SrCl2–KCl–MgCl2 melt were revealed by deep potential molecular dynamicsdriven machine learning to facilitate the development molten salt electrolytic Mg–Sr alloys. short- intermediate-range order melts was explored through radial distribution functions factors, respectively, their component temperature dependence discussed comprehensively. In MgCl2-rich system, is more pronounced, its evolution with exhibits a non-Debye–Waller behavior. Mg–Cl dominated 4,5 coordination Sr–Cl 6,7 coordination, geometries exhibit distorted octahedra pentagonal bipyramids, respectively. A database melts, including density, self-diffusion coefficient, viscosity, ionic conductivity, thus developed, covering range from 873 1173 K.

Язык: Английский

---

Insights into CaCl2-NaCl-KCl molten salt: A machine learning approach to unraveling structure and properties

Journal of Energy Storage, Год журнала: 2024, Номер 102, С. 114156 - 114156

Опубликована: Окт. 16, 2024

Язык: Английский

---

Deep potential molecular dynamic and electrochemical experiments to reveal the structure and behavior of Mn(II) in magnesium electrolysis

Brazilian Journal of Chemical Engineering, Год журнала: 2024, Номер unknown

Опубликована: Май 13, 2024

Язык: Английский

---

Unveiling ionic redox potentials: Advancing prediction through large-scale MLMD and FEP integration

Chemical Engineering Science, Год журнала: 2024, Номер 299, С. 120421 - 120421

Опубликована: Июль 7, 2024

Язык: Английский

---

Formation and Dynamics of Imidazole Supramolecular Chains Investigated by Deep Potential Molecular Dynamics Simulation

Langmuir, Год журнала: 2024, Номер 40(45), С. 23864 - 23871

Опубликована: Ноя. 1, 2024

Imidazole-based materials have attracted considerable attention due to their promising potential for facilitating anhydrous proton transport at high temperatures. Herein, a machine learning-based deep (DP) model bulk imidazole with first-principles accuracy is developed. The trained exhibits remarkable in predicting energies and forces, minor errors of 4.71 × 10–4 eV/atom 3.23 10–2 eV/Å, respectively. Utilizing DP molecular dynamics simulations, we systematically investigated the temperature-dependent formation supramolecular chains through partial radial distribution function, quantification hydrogen bond numbers, incoherent intermediate scattering diffusion coefficient. findings reveal influence temperature on path following either "Grotthuss" "vehicle" mechanism.

Язык: Английский

---

Understanding the local structure and thermophysical behavior of Mg-La liquid alloys via machine learning potential

International Journal of Minerals Metallurgy and Materials, Год журнала: 2024, Номер 32(2), С. 439 - 449

Опубликована: Дек. 26, 2024

Язык: Английский

---