Applications of AI for Catalysis and Energy Materials Discovery

Talk Abstracts:

Designing materials for thermoelectric applications: density functional theory and machine learning

Designing high-performance thermoelectric materials requires optimising a delicate balance of electronic transport, lattice thermal conductivity, and chemical stability. This talk will introduce a combined density functional theory (DFT) and machine-learning (ML) framework for accelerating this search. I will first outline how advanced first-principles workflows, including anharmonic force-constant modelling and Boltzmann transport calculations, provide detailed insight into carrier transport and heat flow in chalcopyrite compounds and related materials. I will then show how ML-based models such as hiPhive and CraTENet can dramatically reduce computational cost, enabling rapid prediction of electronic and thermal transport properties across very large chemical spaces. Finally, I will discuss how composition-driven screening coupled with generative AI crystal structure prediction techniques might open new routes for discovering Earth-abundant, high-zT thermoelectrics.

Optical Spectroscopic Analysis of Metal Oxide: Electrolyte Interfaces for Photo- and Electro-Catalysis

I will start by discussing the potential importance of both photoelectrocatalytic and electrocatalytic strategies in our transition to more sustainable energy systems. I will consider both the importance of water splitting to yield molecular hydrogen, and also the wider challenges of CO2 and nitrogen reduction. A key challenge is eluciation of the materials design requirements for efficient catalytic function. I will go on to discuss the use of optical spectroscopies to provide insight into (photo)electrocatalytic function, employing water oxidation as our most widely studied exemplar. In particular I will discuss how operando optical spectroelectrochemistry can used to determine redox state population densities and kinetics in metal oxides electrodes and photoelectrodes during water oxidation. This spectroelectrochemical approach is based on the idea that the redox states of most transition metal oxides are coloured, allowing the specific concentrations of each state to be tracked by their optical absorption/reflection as a function of material, applied bias, time, electrolyte etc. Such optical data can be complimentary to more widely employed electrochemical analyses such as J/V plots and Tafel analyses. In my talk I will discuss examples of the insights gained from operando spectroelectrochemistry into the design and function of a selection of materials, including Iridium Oxide and Nickel/Iron Oxyhydroxide electrocatalysts, as well as Hematite, Bismuth vanadate and Strontium Titanate photoelectrodes/photocatalysts.

Interfacial Design in Next-Generation Solid Electrolytes

Interfaces are central to the performance, functionality and longevity of energy materials and devices, whether they be batteries, fuel cells or a plethora of other crucial technologies. In solid-state batteries, interfaces govern how efficiently ions move between the solid electrolyte and the electrodes. Poorly engineered interfaces can lead to high resistance, reduced performance and even instability, while well-designed interfaces enhance conductivity, battery life and safety. In the first part of this talk, I will discuss how machine learning interatomic potentials have enabled us to explore interfaces and ion transport at scales far beyond the capabilities of ab initio molecular dynamics. This includes the investigation of solid electrolyte interphase formation between argyrodite Li6PS5Cl and Li metal and universal cation transport in LaCl and LaBr-based solid electrolytes. In the second part of this talk, I will present our recent atomistic simulations of grain boundaries in topical solid electrolytes, including garnet Li7La3Zr2O12 and argyrodite Li6PS5Cl. Specifically, I will discuss how the different electronic structures of Li7La3Zr2O12 and Li6PS5Cl grain boundaries lead to fundamentally different mechanisms of lithium ion incorporation and reduction and lithium atom nucleation in these two materials.

Data-driven Materials Science for Energy-Sustainable Applications

This lecture presents the development and application of data-extraction tools and
machine-learning (ML) models that are enabling data-driven materials science for
energy-sustainable applications. The talk describes how to source high-quality data to
build custom databases and apply them to (a) train ML models; (b) realise data-driven
materials discovery; (c) conduct data-driven materials science for the energy sector; (f)
create energy-efficient domain-specific language models for materials science to feed
agentic AI and help democratise AI for the materials research community. The case
studies presented track a theme of energy-sustainable materials

Engineering Local Chemical Environments in Electrolytes for Efficient Batteries

“Electrolytes play a crucial role in energy storage devices, impacting their environmental footprint, safety, cost, and performance. This talk will cover two areas of electrolyte research for improving batteries.

First, we will explore water-based batteries. Aqueous electrolytes, being non-flammable and less toxic, offer safer battery operation. However, their limited electrochemical stability window reduces energy density. To address this, highly concentrated “”water-in-salt”” (WIS) electrolytes have been developed, significantly expanding the stability window and enhancing the performance of Li-ion and Zn metal batteries for grid energy storage. Despite their advantages, WIS electrolytes have high viscosity and require large amounts of potentially toxic salts, which limits their usability. We will discuss how cation solvation, electrolyte structure, and hydrogen bonding influence the electrochemical properties and performance, particularly in Zn plating/stripping and electrolyte decomposition. Additionally, we will explore strategies for engineering relatively dilute electrolytes for efficient aqueous batteries.

Second, the talk will present a novel method for stabilizing interfaces in Li metal batteries (LiMBs). Conventional electrolytes result in low cycle life and safety issues due to “”dead”” lithium and dendrite formation. Prior research suggests that fluorine-rich interfacial layer chemistry is important for the stabilization of Li-metal anodes, which can be achieved when electrolytes with a high fraction of fluorinated solvents and/or salts are used. We propose an alternative approach using electrostatic attraction between positively charged readily reducible fluorinated cations and the negatively charged anode. This method enables the formation of a robust fluorine-rich interfacial layer with minimal additive (as low as 1 mM) facilitating dense, conformal Li deposition. This strategy can offers a cost-effective, environmentally friendly solution for enhancing high-energy batteries.”