Absolute standard hydrogen electrode potential and redox potentials of atoms and molecules: machine learning aided first principles calculations

Ryosuke Jinnouchi,Ferenc Karsai,Georg Kresse

2024-09-17

Abstract:Constructing a self-consistent first-principles framework that accurately predicts the properties of electron transfer reactions through finite-temperature molecular dynamics simulations is a dream of theoretical electrochemists and physical chemists. Yet, predicting even the absolute standard hydrogen electrode potential, the most fundamental reference for electrode potentials, proves to be extremely challenging. Here, we show that a hybrid functional incorporating 25 % exact exchange enables quantitative predictions when statistically accurate phase-space sampling is achieved via thermodynamic integrations and thermodynamic perturbation theory calculations, utilizing machine-learned force fields and $\Delta$-machine learning models. The application to seven redox couples, including molecules and transition metal ions, demonstrates that the hybrid functional can predict redox potentials across a wide range of potentials with an average error of 80 mV.

Chemical Physics

What problem does this paper attempt to address?

The problem this paper attempts to address is the accurate prediction of the Absolute Standard Hydrogen Electrode Potential (ASHEP) and the potentials of various redox pairs. Specifically, the main challenges faced by the researchers include: 1. **Uncertainty in experimental measurements**: Although ASHEP is a fundamental reference point, results obtained from different experimental methods vary, ranging from −4.80 to −4.28 volts. 2. **Difficulty in theoretical calculations**: First-principles calculations of ASHEP are very challenging, requiring substantial computational resources. Traditional methods such as Quantum Mechanics/Molecular Mechanics (QM/MM) methods and static first-principles calculations have certain limitations and errors. To overcome these challenges, the researchers developed a method combining Machine Learning (ML) and Thermodynamic Integration (TI) through the following steps: - Using hybrid functionals (including 25% exact exchange) and obtaining statistically accurate phase space sampling through thermodynamic integration and thermodynamic perturbation theory calculations. - Utilizing Machine Learning Force Fields (MLFF) and Δ-machine learning models to accelerate molecular dynamics simulations, thereby improving computational efficiency. - Applying this method to calculate seven different redox pairs, including molecules and transition metal ions, demonstrating its ability to predict redox potentials over a wide potential range with an average error of only 80 millivolts. In summary, this paper aims to establish a framework that can accurately predict ASHEP and the potentials of various redox pairs by combining machine learning techniques with first-principles calculations.

Absolute standard hydrogen electrode potential and redox potentials of atoms and molecules: machine learning aided first principles calculations

Machine learning-aided first-principles calculations of redox potentials

Redox potentials and acidity constants from density functional theory based molecular dynamics.

Predicting redox potentials by graph‐based machine learning methods

Machine Learning Potential for Electrochemical Interfaces with Hybrid Representation of Dielectric Response

Predicting Redox Potentials by Graph-Based Machine Learning Methods

ChemInform Abstract: Redox Potentials and Acidity Constants from Density Functional Theory Based Molecular Dynamics

Deepks: A Comprehensive Data-Driven Approach Toward Chemically Accurate Density Functional Theory

Grand-Canonical First Principles-Based Calculations of Electrochemical Reactions

Machine learning interatomic potentials for reactive hydrogen dynamics at metal surfaces based on iterative refinement of reaction probabilities

Machine Learning Prediction of Hydrogen Atom Transfer Reactivity in Photoredox-Mediated C–H Functionalization

Molecular structure–redox potential relationship for organic electrode materials: density functional theory–Machine learning approach

Ab Initio Calculations of the Redox Potentials of Additives for Lithium-Ion Batteries and Their Prediction through Machine Learning

Towards Predicting Metal Extraction from Oxides with Machine-Learning Augmented First-Principles Calculations

A constant potential reactor framework for electrochemical reaction simulations

Teaching oxidation states to neural networks

Using machine learning to go beyond potential energy surface benchmarking for chemical reactivity

Prediction of organic homolytic bond dissociation enthalpies at near chemical accuracy with sub-second computational cost

Development of Accurate DFT Methods for Computing Redox Potentials of Transition Metal Complexes: Results for Model Complexes and Application to Cytochrome P450

Augmenting zero-Kelvin quantum mechanics with machine learning for the prediction of chemical reactions at high temperatures