Machine learning for the modeling of interfaces in energy storage and conversion materials

Artrith, Nongnuch

doi:10.1088/2515-7655/ab2060

Cited by 66 publications

(43 citation statements)

References 110 publications

(130 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The examples cited and many similar ones reported in the literature can take years to develop as broad regions of phase space have to be sampled by an appropriate balance of training points to provide reliable predictions across the board. As a consequence, relatively few studies exist in which complex solid-liquid systems have been described with MLPs (35)(36)(37)(38)(39). These limitations have hampered progress in understanding solid-liquid interfaces, where accurate MLPs are urgently needed and offer many opportunities for deepening our understanding of processes like wetting, ice formation, or liquid flow and friction under confinement.…”

mentioning

confidence: 99%

Machine learning potentials for complex aqueous systems made simple

Schran

Thiemann

Rowe

et al. 2021

Proc. Natl. Acad. Sci. U.S.A.

138

117

View full text Add to dashboard Cite

Simulation techniques based on accurate and efficient representations of potential energy surfaces are urgently needed for the understanding of complex systems such as solid–liquid interfaces. Here we present a machine learning framework that enables the efficient development and validation of models for complex aqueous systems. Instead of trying to deliver a globally optimal machine learning potential, we propose to develop models applicable to specific thermodynamic state points in a simple and user-friendly process. After an initial ab initio simulation, a machine learning potential is constructed with minimum human effort through a data-driven active learning protocol. Such models can afterward be applied in exhaustive simulations to provide reliable answers for the scientific question at hand or to systematically explore the thermal performance of ab initio methods. We showcase this methodology on a diverse set of aqueous systems comprising bulk water with different ions in solution, water on a titanium dioxide surface, and water confined in nanotubes and between molybdenum disulfide sheets. Highlighting the accuracy of our approach with respect to the underlying ab initio reference, the resulting models are evaluated in detail with an automated validation protocol that includes structural and dynamical properties and the precision of the force prediction of the models. Finally, we demonstrate the capabilities of our approach for the description of water on the rutile titanium dioxide (110) surface to analyze the structure and mobility of water on this surface. Such machine learning models provide a straightforward and uncomplicated but accurate extension of simulation time and length scales for complex systems.

show abstract

mentioning

confidence: 99%

Machine learning potentials for complex aqueous systems made simple

Schran

Thiemann

Rowe

et al. 2021

Proc. Natl. Acad. Sci. U.S.A.

138

117

View full text Add to dashboard Cite

show abstract

“…Nevertheless, these costs will most likely plummet in the future with the advent of accurate reactive machine learning potentials. [66][67][68]…”

Section: Generality Of the Reaction Mechanismsmentioning

confidence: 99%

The Role of Water Loading and Germanium Content in Germanosilicate Hydrolysis

Jin¹,

Veselý²,

Heard³

et al. 2021

Preprint

View full text Add to dashboard Cite

New zeolitic frameworks can be prepared through the Assembly-Disassembly-Organisation-Reassembly (ADOR) process by exploiting the lability of Ge-O bonds in germanosilicate zeolites to control their hydrolysis. In the disassembly step, two key factors are water and germanium content, but their exact roles remain unknown. Nevertheless, we combined experimental water-vapor adsorption with first principles simulations to identify the mechanism of germanosilicate zeolite disassembly. The results showed that water vapor adsorption on UTL germanosilicate proceeds in reversible (at low partial pressures) and irreversible (at higher partial pressures) modes. Based on our ab initio molecular dynamics simulations, we related these two modes to weak physisorption at low water loading and to reactive transformations at high water loading, via collective mechanisms requiring high local water concentrations. This bimodal behavior also depends on the germanium content as high Ge-content further decreases UTL hydrolytic stability by opening up yet another low-energy disassembly pathway at high water loading. Overall, we discovered, verified and explained the mechanisms of UTL disassembly and its factors. These findings will likely be generalized to other ADORable germanosilicate zeolites and help to find the optimal protocol for the synthesis of new zeolites.

show abstract

“…In correspondence to drug discovery applications, the application of ML models to materials discovery has also seen a steep rise of research activity during the last decade, owing to the availability of methods, public implementations, and increased computer power [ 190 ]. In this section, we review some of the recent successful applications in the area with a focus on inorganic solid materials.…”

Section: Applications To Industrymentioning

confidence: 99%

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Morawietz

Artrith

2020

J Comput Aided Mol Des

Self Cite

View full text Add to dashboard Cite

Atomistic simulations have become an invaluable tool for industrial applications ranging from the optimization of protein-ligand interactions for drug discovery to the design of new materials for energy applications. Here we review recent advances in the use of machine learning (ML) methods for accelerated simulations based on a quantum mechanical (QM) description of the system. We show how recent progress in ML methods has dramatically extended the applicability range of conventional QM-based simulations, allowing to calculate industrially relevant properties with enhanced accuracy, at reduced computational cost, and for length and time scales that would have otherwise not been accessible. We illustrate the benefits of ML-accelerated atomistic simulations for industrial R&D processes by showcasing relevant applications from two very different areas, drug discovery (pharmaceuticals) and energy materials. Writing from the perspective of both a molecular and a materials modeling scientist, this review aims to provide a unified picture of the impact of ML-accelerated atomistic simulations on the pharmaceutical, chemical, and materials industries and gives an outlook on the exciting opportunities that could emerge in the future.

show abstract

Machine learning for the modeling of interfaces in energy storage and conversion materials

Cited by 66 publications

References 110 publications

Machine learning potentials for complex aqueous systems made simple

Machine learning potentials for complex aqueous systems made simple

The Role of Water Loading and Germanium Content in Germanosilicate Hydrolysis

Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications

Contact Info

Product

Resources

About