Periodic Reporting for period 1 - DEMI (Directed Evolution of Metastable Electrocatalyst Interfaces for Energy Conversion)
Berichtszeitraum: 2024-01-01 bis 2025-06-30
Our goal is to transform electrocatalysis research from the conventional “as-synthesized” view of catalysts to a data-driven understanding of the metastable active interface—the dynamic state of electrocatalysts under reaction conditions. Instead of being limited to elemental or binary alloys, we aim to harness high entropy materials (HEMs) as a discovery platform. Their vast multidimensional composition space holds the potential for identifying electrocatalysts that are both stable and highly active.
To achieve this, we integrate the complementary expertise of the PIs in theoretical modeling, high-throughput synthesis and characterization, nanoparticle synthesis, operando electrochemistry, and machine learning. This synergy enables key conceptual advances:
Evolutionary screening of micro-libraries to efficiently map stability across the HEM space.
Accelerated atomic-scale characterization of HEM surfaces via combinatorial synthesis combined with atom probe tomography.
High-throughput operando experiments using thin-film material libraries.
Inverse activity–structure relationships and novel theoretical descriptors for metastability.
Active learning approaches using materials informatics and a semantic data lake.
Through this integrated strategy, we will establish descriptors for metastability for electrocatalysts, focusing on the most critical energy conversion reactions: oxygen reduction, oxygen evolution, and CO2 reduction. Instead of passively accepting catalyst degradation, we will steer the evolution of HEM interfaces within their vast composition space toward long-lived, active states.
To achieve this, we integrate the complementary expertise of the PIs in theoretical modeling, high-throughput synthesis and characterization, nanoparticle synthesis, operando electrochemistry, and machine learning. This synergy enables key conceptual advances:
Evolutionary screening of micro-libraries to efficiently map stability across the HEM space.
Accelerated atomic-scale characterization of HEM surfaces via combinatorial synthesis combined with atom probe tomography.
High-throughput operando experiments using thin-film material libraries.
Inverse activity–structure relationships and novel theoretical descriptors for metastability.
Active learning approaches using materials informatics and a semantic data lake.
Through this integrated strategy, we will establish descriptors for metastability for electrocatalysts, focusing on the most critical energy conversion reactions: oxygen reduction, oxygen evolution, and CO2 reduction. Instead of passively accepting catalyst degradation, we will steer the evolution of HEM interfaces within their vast composition space toward long-lived, active states.
At UCPH, we developed a theoretical framework to estimate the dissolution of nanoparticles, enabling us to create maps of stability as a function of composition. This allows for comparison between theory and experiments. The framework partially includes surface segregation; however, we have studied binary surface alloys to better understand surface stability against segregation and to model segregation within the surface. We are making progress in developing descriptors for stability against dissolution and segregation at the surface, although we currently lack descriptors for stability against segregation within the surface.
At RUB, we successfully worked on the combinatorial synthesis, high-throughput characterization, and activity determination of thin-film high-entropy material libraries (alloys and oxides). Electrochemical activity screening was expanded to include automated scanning electrochemical cell microscopy (SECCM), alongside automated scanning droplet cell (SDC), in collaboration with Prof. Dr. Wolfgang Schuhmann. We also developed a workflow that combines high-throughput characterization of thin-film libraries—covering composition, structure, and properties—with an active learning algorithm to efficiently explore entire quaternary composition spaces. This approach was applied to multiple quaternary systems to screen their activity for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER), producing a large dataset for benchmarking simulations and nanoparticle comparisons.
We further introduced the simulation technique SIMTRA, designed to reliably predict the compositional outcomes of combinatorial sputtering processes. These simulations facilitate rational mask design for microlibraries and offer predictive insights into compositional gradients and their coverage of the composition space, also applicable to conventional material libraries. The sputter deposition simulation enhances efficiency by reducing the number of measurements needed post-fabrication, thus lowering experimental effort.
Prior to DEMI, our focus was on composition-activity relationships using volume composition data obtained from high-throughput Energy-dispersive X-ray spectroscopy (EDX). Recognizing the surface relevance of electrocatalytic activity, we developed a novel X-ray photoelectron spectroscopy (XPS) workflow that predicts surface composition from EDX mappings combined with limited XPS measurements, leveraging a multi-output machine learning algorithm. This method accelerates the surface analysis of thin-film libraries and enables characterization across entire composition spreads rather than single points.
To advance our active learning strategies, we began developing a new autonomous robot-assisted scanning electrochemical cell microscopy device, which will significantly speed up electrochemical characterization of our conventional materials libraries.
We successfully conducted electrochemical experiments using the combinatorial processing platform (CPP). CPPs were also prepared for Erlangen and tested for online dissolution analysis. Initial experiments focused on evolutionary screening.
At UBERN, we established and operated two synthesis platforms: one for ORR and one for OER catalysts. The first is based on electrodeposition and has been used for medium- and high-entropy alloy exploration, combined with theoretical studies at UCPH. One publication has already appeared, and the data from the second platform are currently being compared to libraries from RUB. At FZJ, electrodeposition is being automated to create nanoparticle libraries. Dissolution measurements are ongoing to generate Pareto plots for comparison with UCPH theory. The second platform has achieved a proof of concept for the statistical analysis of benchmark catalysts (Ir, Ru, IrRu) and MnO2-based OER catalysts. These measurements are under analysis, with two publications planned. Additionally, a first set of HEA alloys, prepared by incipient wetness impregnation, has been synthesized, analyzed, and sent to FZJ, where stability measurements have been successfully performed. These results are currently being compared with operando XRD and SAXS degradation data.
At FZJ, we developed a miniaturized electrochemical flow cell (µSFC) for in situ measurements of atomic probe tomography (APT) tips coated with thin-film libraries by RUB. The setup is in the proof-of-concept stage, with initial results related to dissolution and activity measurements on Cantor alloys.
Some of these project results have been published in peer-reviewed journals and presented at conferences through talks and posters. Further publications are in preparation.
Additionally, we implemented a research data management system (RDMS) and a website for the DEMI project.
At RUB, we successfully worked on the combinatorial synthesis, high-throughput characterization, and activity determination of thin-film high-entropy material libraries (alloys and oxides). Electrochemical activity screening was expanded to include automated scanning electrochemical cell microscopy (SECCM), alongside automated scanning droplet cell (SDC), in collaboration with Prof. Dr. Wolfgang Schuhmann. We also developed a workflow that combines high-throughput characterization of thin-film libraries—covering composition, structure, and properties—with an active learning algorithm to efficiently explore entire quaternary composition spaces. This approach was applied to multiple quaternary systems to screen their activity for the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER), producing a large dataset for benchmarking simulations and nanoparticle comparisons.
We further introduced the simulation technique SIMTRA, designed to reliably predict the compositional outcomes of combinatorial sputtering processes. These simulations facilitate rational mask design for microlibraries and offer predictive insights into compositional gradients and their coverage of the composition space, also applicable to conventional material libraries. The sputter deposition simulation enhances efficiency by reducing the number of measurements needed post-fabrication, thus lowering experimental effort.
Prior to DEMI, our focus was on composition-activity relationships using volume composition data obtained from high-throughput Energy-dispersive X-ray spectroscopy (EDX). Recognizing the surface relevance of electrocatalytic activity, we developed a novel X-ray photoelectron spectroscopy (XPS) workflow that predicts surface composition from EDX mappings combined with limited XPS measurements, leveraging a multi-output machine learning algorithm. This method accelerates the surface analysis of thin-film libraries and enables characterization across entire composition spreads rather than single points.
To advance our active learning strategies, we began developing a new autonomous robot-assisted scanning electrochemical cell microscopy device, which will significantly speed up electrochemical characterization of our conventional materials libraries.
We successfully conducted electrochemical experiments using the combinatorial processing platform (CPP). CPPs were also prepared for Erlangen and tested for online dissolution analysis. Initial experiments focused on evolutionary screening.
At UBERN, we established and operated two synthesis platforms: one for ORR and one for OER catalysts. The first is based on electrodeposition and has been used for medium- and high-entropy alloy exploration, combined with theoretical studies at UCPH. One publication has already appeared, and the data from the second platform are currently being compared to libraries from RUB. At FZJ, electrodeposition is being automated to create nanoparticle libraries. Dissolution measurements are ongoing to generate Pareto plots for comparison with UCPH theory. The second platform has achieved a proof of concept for the statistical analysis of benchmark catalysts (Ir, Ru, IrRu) and MnO2-based OER catalysts. These measurements are under analysis, with two publications planned. Additionally, a first set of HEA alloys, prepared by incipient wetness impregnation, has been synthesized, analyzed, and sent to FZJ, where stability measurements have been successfully performed. These results are currently being compared with operando XRD and SAXS degradation data.
At FZJ, we developed a miniaturized electrochemical flow cell (µSFC) for in situ measurements of atomic probe tomography (APT) tips coated with thin-film libraries by RUB. The setup is in the proof-of-concept stage, with initial results related to dissolution and activity measurements on Cantor alloys.
Some of these project results have been published in peer-reviewed journals and presented at conferences through talks and posters. Further publications are in preparation.
Additionally, we implemented a research data management system (RDMS) and a website for the DEMI project.
We have identified a resonance phenomenon between hydrogen bonded OH and water on alloy surfaces. It turns out that water, which alone on the surface is physisorbed, forms a fraction of a covalent bond to the surface when the water molecule is donor of a hydrogen bond to OH. This is a big surprise for us and we found it partly by serendipity.
The electrodeposition approach is beyond state of the art and its automation bears tremendous potential for medium or high throughput screening of HEA nanoparticles.
With the CPP approach we identified the formation of a few-nanometer thick oxygen-influenced layer on top of a high entropy alloy film which seems to protect the underlying film and is still active.
The electrodeposition approach is beyond state of the art and its automation bears tremendous potential for medium or high throughput screening of HEA nanoparticles.
With the CPP approach we identified the formation of a few-nanometer thick oxygen-influenced layer on top of a high entropy alloy film which seems to protect the underlying film and is still active.