Improved prediction of solvation free energies by machine-learning polarizable continuum solvation model

Alibakhshi, Amin; Hartke, Bernd

doi:10.1038/s41467-021-23724-6

Cited by 72 publications

(50 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Widening the redox potential ranges considered can account for applicability of additional radicals in organic RFBs, which can also allow for higher overall cell voltage. Optimizing solubility with predictive models 72,73 and including charged moieties in both the training data and action space will be particularly important in achieving a high charge density. The stability metric employed also may have limitations that will need to be refined following experimental investigation.…”

Section: Discussionmentioning

confidence: 99%

Multi-objective goal-directed optimization of de novo stable organic radicals for aqueous redox flow batteries

Law

Tripp

et al. 2022

Nat Mach Intell

View full text Add to dashboard Cite

Advances in the field of goal-directed molecular optimization offer the promise of finding feasible candidates for even the most challenging molecular design applications. One example of a fundamental design challenge is the search for novel stable radical scaffolds for an aqueous redox flow battery that simultaneously satisfy redox requirements at the anode and cathode, as relatively few stable organic radicals are known to exist. To meet this challenge, we develop a new open-source molecular optimization framework based on AlphaZero coupled with a fast, machine-learning-derived surrogate objective trained with nearly 100,000 quantum chemistry simulations. The objective function comprises two graph neural networks: one that predicts adiabatic oxidation and reduction potentials and a second that predicts electron density and local three-dimensional environment, previously shown to be correlated with radical persistence and stability. With no hard-coded knowledge of organic chemistry, the reinforcement learning agent finds molecule candidates that satisfy a precise combination of redox, stability and synthesizability requirements defined at the quantum chemistry level, many of which have reasonable predicted retrosynthetic pathways. The optimized molecules show that alternative stable radical scaffolds may offer a unique profile of stability and redox potentials to enable low-cost symmetric aqueous redox flow batteries.

show abstract

Section: Discussionmentioning

confidence: 99%

Multi-objective goal-directed optimization of de novo stable organic radicals for aqueous redox flow batteries

Law

Tripp

et al. 2022

Nat Mach Intell

View full text Add to dashboard Cite

show abstract

“…To further improve the inherent accuracy of the implicit solvation models, use of machine-learning oriented approaches can be an attractive subject for future development as has been recently reported. 107 ■ ASSOCIATED CONTENT * sı Supporting Information…”

Section: Discussionmentioning

confidence: 99%

Multireference Perturbation Theory Combined with PCM and RISM Solvation Models: A Benchmark Study for Chemical Energetics

et al. 2021

View full text Add to dashboard Cite

The polarizable continuum model (PCM) has been one of the most widely used approaches to take into account the solvation effect in quantum chemical calculations. In this paper, we performed a series of benchmark calculations to assess the accuracy of the PCM scheme combined with the second-order complete-active-space perturbation theory (CASPT2) for molecular systems in polar solvents. For solute molecules with extensive conjugated π orbitals, exemplified by elongated conjugated arylcarbenes, we have incorporated the ab initio density matrix renormalization group algorithm into the PCM-CASPT2 method. In the previous work, we presented a combination of the DMRG-CASPT2 method with the reference interaction site model (RISM) theory for describing the solvation effect using the radial distribution function and compared its performance to the widely used densityfunctional approaches (PCM-TD-DFT). The work here allows us to further show a more thorough assessment of the RISM model compared to the PCM with an equal level of the wave function treatment, the (DMRG-)CASPT2 theory, toward a high-accuracy electronic structure calculations for solvated chemical systems. With the exception that the PCM models are not capable of properly describing the hydrogen bondings, accuracy of the PCM-CASPT2 model is in most cases quite comparable to the RISM counterpart.

show abstract

“…Deviation from first principles turns hard theory into soft theory, such as continuum solvation models. 15–17…”

Section: Soft Theories Versus Hard Theories7mentioning

confidence: 99%

Understanding chemistry: from “heuristic (soft) explanations and reasoning by analogy” to “quantum chemistry”

Seeman

Tantillo

2022

Chem. Sci.

View full text Add to dashboard Cite

show abstract

Improved prediction of solvation free energies by machine-learning polarizable continuum solvation model

Cited by 72 publications

References 30 publications

Multi-objective goal-directed optimization of de novo stable organic radicals for aqueous redox flow batteries

Multi-objective goal-directed optimization of de novo stable organic radicals for aqueous redox flow batteries

Multireference Perturbation Theory Combined with PCM and RISM Solvation Models: A Benchmark Study for Chemical Energetics

Understanding chemistry: from “heuristic (soft) explanations and reasoning by analogy” to “quantum chemistry”

Contact Info

Product

Resources

About