Lars Banko

The discovery of high‐performance electrocatalysts is crucial for advancing sustainable energy technologies. Compositionally complex solid solutions comprising multiple metals offer promising catalytic properties, yet their exploration is challenging due to the combinatorial explosion of possible compositions. In this work, human‐decision driven combinatorial sputtering of thin‐film materials libraries and high‐throughput characterization is combined with Bayesian optimization to efficiently explore the quaternary composition space Ni‐Pd‐Pt‐Ru for the oxygen evolution reaction (OER) in alkaline media. Using this method, the global activity optimum of pure Ru is identified after covering less than 20% of the complete composition space with six materials libraries. Six additional libraries are fabricated to validate the activity trend. The resulting dataset is used to formulate general guidelines for the efficient composition of space exploration with combinatorial synthesis paired with Bayesian optimization.

Information technology and data science development stimulate transformation in many fields of scientific knowledge. In recent years, a large number of specialised systems for information and knowledge management have been created in materials science. However, the development and deployment of open adaptive systems for research support in materials science based on the acquisition, storage, and processing of different types of information remains unsolved. We propose MatInf—an extensible, open-source solution for research digitalisation in materials science based on an adaptive, flexible information management system for heterogeneous data sources. MatInf can be easily adapted to any materials science laboratory and is especially useful for collaborative projects between several labs. As an example, we demonstrate its application in high-throughput experimentation.

The discovery of high‐performance electrocatalysts for water electrolysis is highly important. We use a strategy for catalyst discovery based on high‐throughput screening of a broad range of materials compositions on a thin‐film materials library to identify noble‐metal‐lean multi‐metal compositions with high activity towards the hydrogen evolution reaction in alkaline electrolyte. We demonstrate this strategy using a quaternary materials library containing Mo, Ag, Ti, and Ru fabricated by combinatorial magnetron sputtering on a 10 cm diameter wafer providing 342 measurement areas. Surprisingly, binary Mo−Ru‐containing catalyst compositions with comparatively low Ru content exhibited the highest activity. Using the polymer/metal precursor‐based spray technique, Ni foam electrodes were modified with the Mo−Ru hit compositions and evaluated in a model electrolyzer in membrane electrode assembly (MEA) configuration. The electrodes showed a very low overpotential of only 132 mV at a comparatively high current density of −2 A cm⁻² and a 24 h electrolysis stability at −1 A cm⁻² with no observable degradation after the initial electrode conditioning.

Developing a hydrogen evolution reaction (HER) electrocatalyst that can steadily drive a large current density is of great significance for industrial electrochemical water‐splitting technology. Herein, the compositionally complex noble‐metal‐lean system Co–Cu–Mo–Pd–Re is investigated for alkaline HER by a scanning droplet cell‐assisted high‐throughput screening of co‐sputtered thin‐film material libraries. The identified hit compositions are used to prepare a catalyst‐coated Ni foam‐based layer that is hot‐pressed with a specifically designed PEEK‐covered carbon‐based gas diffusion layer to obtain a novel‐design gas diffusion electrode (GDE). H₂ bubbles that collect on the electrode surface are effectively reduced by controlling the hydrophobicity of the catalyst layer of the GDE, hence enabling a high current (−1 A cm⁻²) operation for HER in alkaline water electrolysis.

Information technology and data science development stimulate transformation in many fields of scientific knowledge. In recent years, a large number of specialized systems for information and knowledge management have been created in materials science. However, the development and deployment of open adaptive systems for research support in materials science based on the acquisition, storage, and processing of different types of information remains unsolved. We propose MatInf - an extensible, open-source solution for research digitalisation in materials science based on an adaptive, flexible information management system for heterogeneous data sources. MatInf can be easily adapted to any materials science laboratory and is especially useful for collaborative projects between several labs. As an example, we demonstrate its application in high-throughput experimentation.

High-throughput experimentation enables efficient search space exploration for the discovery and optimization of new materials. However, large search spaces, e.g. of compositionally complex materials, require decreasing characterization times significantly. Here, an autonomous measurement algorithm was developed, which leverages active learning with a Gaussian process model capable of iteratively scanning a materials library based on the highest uncertainty. The algorithm is applied to a four-point probe electrical resistance measurement device, frequently used to obtain indications for regions of interest in materials libraries. Ten libraries with different complexities of composition and property trends are analyzed to validate the model. By stopping the process before the entire library is characterized and predicting the remaining areas, the measurement efficiency can be improved drastically. As robustness is essential for autonomous measurements, intrinsic outlier handling is built into the model and a dynamic stopping criterion based on the mean predicted covariance is proposed. A measurement time reduction of about 70–90% was observed while still ensuring an accuracy above 90%.

Compositionally complex materials such as high‐entropy alloys and oxides have the potential to be efficient platforms for catalyst discovery because of the vast chemical space spanned by these novel materials. Identifying the composition of the most active catalyst materials, however, requires unraveling the descriptor‐activity relationship, as experimentally screening the multitude of possible element ratios quickly becomes a daunting task. In this work, we show that inferred adsorption energy distributions of *OH and *O on complex solid solution surfaces within the space spanned by the system Ag−Pd−Pt−Ru are coupled to the experimentally observed electrocatalytic performance for the oxygen reduction reaction. In total, the catalytic activity of 1582 alloy compositions is predicted with a cross‐validated mean absolute error of 0.042 mA/cm² by applying a theory‐derived model with only two adjustable parameters. Trends in the discrepancies between predicted electrochemical performance values of the model and the measured values on thin film surfaces subsequently provide insight into the alloys’ surface compositions during reaction conditions. Bridging this gap between computationally modeled and experimentally observed catalytic activities, not only reveals insight into the underlying theory of catalysis but also takes a step closer to realizing exploration and exploitation of high‐entropy materials.

Compositionally complex materials such as high‐entropy alloys and oxides have the potential to be efficient platforms for catalyst discovery because of the vast chemical space spanned by these novel materials. Identifying the composition of the most active catalyst materials, however, requires unraveling the descriptor‐activity relationship, as experimentally screening the multitude of possible element ratios quickly becomes a daunting task. In this work, we show that inferred adsorption energy distributions of *OH and *O on complex solid solution surfaces within the space spanned by the system Ag−Pd−Pt−Ru are coupled to the experimentally observed electrocatalytic performance for the oxygen reduction reaction. In total, the catalytic activity of 1582 alloy compositions is predicted with a cross‐validated mean absolute error of 0.042 mA/cm² by applying a theory‐derived model with only two adjustable parameters. Trends in the discrepancies between predicted electrochemical performance values of the model and the measured values on thin film surfaces subsequently provide insight into the alloys’ surface compositions during reaction conditions. Bridging this gap between computationally modeled and experimentally observed catalytic activities, not only reveals insight into the underlying theory of catalysis but also takes a step closer to realizing exploration and exploitation of high‐entropy materials.

Compositionally complex perovskites provide the opportunity to develop stable and active catalysts for electrochemical applications. The challenge lies in the identification of single‐phase perovskites with optimized composition for high electrical conductivity. Leveraging a recently discovered effect of self‐organized thin film growth during reactive sputtering, La–Co–Mn–O and La–Co–Mn–Fe–O perovskite (ABO₃) thin‐film materials libraries are synthesized. These show phase‐pure La perovskites over a wide range of chemical composition variation for the B‐site elements for deposition temperatures ≥300 °C. It is demonstrated that this approach enables the discovery and tailoring of chemical compositions for desired optical bandgap and electrical conductivity and thereby opens the path for the targeted development of, for example, new high‐performance electrocatalysts.

Compositionally complex materials such as high-entropy alloys and oxides have the potential to be efficient platforms for catalyst discovery because of the vast chemical space spanned by these novel materials. Identifying the composition of the most active catalyst materials, however, requires unraveling the descriptor-activity relationship, as experimental screening the multitude of possible element ratios quickly becomes a daunting task. In this work, we show that inferred adsorption energy distributions of *OH and *O on complex solid solution surfaces within the space spanned by the system Ag-Pd-Pt-Ru are coupled to the experimentally observed electrocatalytic performance for the oxygen reduction reaction. In total, the catalytic activity of 1582 alloy compositions is predicted with a cross-validated mean absolute error of 0.042 mA/cm2 by applying a theory-derived model with only two adjustable parameters. Trends in the discrepancies between predicted electrochemical performance values of the model and the measured values on thin film surfaces subsequently provide insight into the alloys’ surface compositions during reaction conditions. Bridging this gap between computationally modeled and experimentally observed catalytic activities, not only reveals insight into the underlying theory of catalysis but also takes a step closer to realizing exploration and exploitation of high-entropy materials.