High-entropy materials, first demonstrated in metallic alloys and later extended to oxides and other systems, unlock a vast compositional space with properties suited for catalysis, energy, and structural materials. However, the high compositional complexity makes systematic exploration challenging, and only a small portion of the design space has been studied. To address this, we introduce an active learning strategy that integrates predictive modeling, uncertainty estimation, and iterative sampling to efficiently navigate embedded compositional material spaces. This approach continuously learns from previous evaluations, focusing subsequent searches on the most promising regions while reducing both time and data requirements. We demonstrate this methodology in the search for high-entropy oxygen carriers for chemical looping, where it rapidly accelerates discovery and identifies promising candidates more effectively than conventional trial-and-error or grid-search approaches. Importantly, this strategy is general and well-suited to exploring the vast space of multicomponent materials.

The discovery and design of new materials are paramount in the development of green technologies. High entropy oxides represent one such group that has only been tentatively explored, mainly due to the inherent problem of navigating vast compositional spaces. Thanks to the emergence of machine learning, however, suitable tools are now readily available. Here, the task of finding oxygen carriers for chemical looping processes has been tackled by leveraging active learning-based strategies combined with first-principles calculations. High efficiency and efficacy have, moreover, been achieved by exploiting the power of recently developed machine learning interatomic potentials. Firstly, the proposed approaches were validated based on an established computational framework for identifying high entropy perovskites that can be used in chemical looping air separation and dry reforming. Chief among the insights thus gained was the identification of the best performing strategies, in the form of greedy or Thompson-based sampling based on uncertainty estimates obtained from Gaussian processes. Building on this newfound knowledge, the concept was applied to a more complex problem, namely the discovery of high entropy oxygen carriers for chemical looping oxygen uncoupling. This resulted in both qualitative as well as quantitative outcomes, including lists of specific materials with high oxygen transfer capacities and configurational entropies. Specifically, the best candidates were based on the known oxygen carrier CaMnO3 but also contained a variety of additional species, of which some, e.g., Ti; Co; Cu; and Ti, were expected while others were not, e.g., Y and Sm. The results suggest that adopting active learning approaches is critical in materials discovery, given that these methods are already shifting research practice and soon will be the norm.

Reactions between nitrogen-containing compounds and free chlorine in swimming pools produce several undesirable by-products, with trichloramine (TCA) being especially problematic due to its odor and hazardous properties, causing respiratory symptoms and skin and eye irritation for swimmers and personnel. To address this issue, this study investigates electrochemical oxidation as a potential technique for reducing TCA concentrations in various pool water environments. Specifically, the results show that electrolysis, using a mixed metal oxide (MMO) electrode as the anode, affects the amount of free chlorine and significantly reduces TCA in synthetic pool water containing urea and sodium hypochlorite. This approach is further validated in large-scale systems with real swimming pool water. In two long-term experiments, TCA released from the pool water falls from a baseline of 1.42 mg/m3 to 0.94 mg/m3 in a 680 m3 pool, a reduction of 34% (p<0.001), and from 0.77 mg/m3 to 0.35 mg/m3 in an 85 m3 pool, a reduction of 55% (p<0.001). Water from a third pool was used to map the potential dependence. Importantly, the anode potential is critical for TCA reduction, with 1.8 V vs. RHE identified as optimal and a working window of 1.7–1.9 V. The system operates at approximately 1 A/m2 and 2 W/m2. Overall, this electrolysis process reduces the trichloramine burden of swimming pool water, which is expected to lower airborne trichloramine in swimming halls and thereby improve conditions for users and personnel.

Zeolitic catalysts used in the industrial manufacture of commodity chemicals excel thanks to their thermal stability, chemical tunability, and microporosity, aiding catalytic functions and shape- and size-selectivity that can be utilised also in reactions such as the conversion of biomass-derived furans into green aromatics, i.e., benzene, toluene, and xylenes (BTX). However, excessive aromatization into polyaromatics and other carbonaceous species results in blocking of both active sites and micropores, ultimately deactivating the catalyst. As a solution, here, three classes of microporous materials comprising complementary meso- and/or macroporous domains have been synthesised and explored for the conversion of 2,5-dimethylfuran to BTX. First, meso- and macropores have been introduced in microporous Ga-MFI zeotype catalysts, with varying gallium content, via post-synthesis template-assisted base leaching. This increases the catalytic activity per acid site thanks to enhanced mass transfer properties, although at the expense of the acid site density, ultimately leading to similar overall production of aromatics. Second, gallium doped mesoporous silica, Ga-MCM-41, is shown to decrease the selectivity towards monoaromatics while increasing the formation of coke, due to the mere presence of mesopores. Through a one-pot, two-temperature step hydrothermal synthesis, Ga-MFI/MCM-41 composites were formed by partial transformation of the Ga-MCM-41 mesopore walls into microporous MFI-framework domains. Thereby, the selectivity towards benzene and isomerisation products increases while that for coke decreases. Third, Ga-containing self-pillared pentasil units (SPP) have been synthesised with varying gallium content. These materials possess both microporous MFI-domains and a broad size range of meso- and macropores in between said domains. However, due to its low acidity, almost no conversion of 2,5-dimethylfuran into benzene is observed. By comparing these three classes of materials, the critical role of the micropores is highlighted, as well as a potentially beneficial role of meso- and macropores. Moreover, the acid site density, generated by gallium, plays an important promoting role to convert 2,5-dimethylfuran into aromatics. Only when the majority of the material consists of microporous MFI, the selectivity is towards monocyclic aromatics such as benzene, instead of polycyclic aromatics and other coke species.

Understanding the role of oxygen vacancies in perovskite oxides is essential for tailoring their functional properties. In this study, we employ density functional theory (DFT) calculations to investigate the structural, electronic, and defect-related properties of CaMnO3-delta across a wide range of oxygen vacancy concentrations. Although prior studies have examined oxygen vacancies in CaMnO3-delta, here we present a systematic investigation across the full range from delta = 0 to delta = 0.5 in increments of Delta delta = 0.0625, thereby providing detailed new insights into its defect chemistry. Using cluster expansion techniques and high-accuracy relaxations, we identify ground-state configurations for each oxygen vacancy concentration and analyze their lattice distortions, bond length variations, and charge redistribution. Our results reveal a nonlinear progression of lattice parameters and bond environments with increasing oxygen vacancy concentration, accompanied by significant changes in the electronic band structure. Bader charge analysis indicates a progressive reduction in Mn oxidation state and charge compensation among all atomic species. As the estimated cost of creating oxygen vacancies increases at high concentration, from 2.03 to 2.26 eV, there exists a thermodynamic limit to vacancy incorporation at high oxygen vacancy concentrations. These findings provide a comprehensive understanding of how oxygen vacancies influence the stability and electronic behavior of CaMnO3-delta, offering insights relevant to its use in different types of applications.

The discovery and design of new materials are paramount for advancing green technologies. High-entropy oxides represent one such group that has been only tentatively explored, mainly due to the inherent problem of navigating vast compositional spaces. Here, oxygen carriers for chemical looping processes have been identified using active learning-based strategies and first-principles-informed calculations. The proposed approaches were validated using an established computational framework for identifying high-entropy perovskites suitable for chemical looping air separation and dry reforming. The central insight gained was the identification of effective strategies, including greedy and Thompson-based sampling, informed by uncertainty estimates from Gaussian processes. Building on this knowledge, the concept was applied to the challenge of discovering high-entropy oxygen carriers for chemical-looping oxygen uncoupling. This resulted in both qualitative and quantitative outcomes, including lists of materials with high oxygen transfer capacities and configurational entropies. The top candidates were based on the known oxygen carrier CaMnO3 and included expected elements such as titanium, cobalt, and copper, as well as unexpected ones such as yttrium and samarium. The results suggest that adopting active learning approaches is critical for materials discovery, as these methods are already reshaping research practice and will soon become the norm.

Ethylene epoxidation remains one of the most industrially significant catalytic processes, as it serves as the primary route to ethylene oxide—an essential precursor chemical in the production of ethylene glycol, surfactants, and other chemicals. Despite nearly a century of commercial operation, the detailed reaction mechanism of silver-based catalysts remains the focus of intensive research. This chapter offers a detailed overview of the mechanistic pathways involved in ethylene epoxidation, particularly the competition between selective epoxide formation and total oxidation routes. In particular, it explores the role of different silver surface facets, the dynamic behavior of chemisorbed oxygen species, and the influence of surface modifiers such as chlorine and alkali metals. Special attention is given to the nature of electrophilic and nucleophilic oxygen species and their role in determining selectivity. Results from advanced experimental techniques, including in situ and operando spectroscopy, alongside density functional theory (DFT) and microkinetic modeling, are discussed and provide a multi-scale perspective on the catalyst behavior. Finally, we outline ongoing challenges and future directions, including integrating machine learning and data-driven approaches to unravel complex reaction networks.

Chemical-looping is a technology that increases combustion efficiency and enables carbon capture with a low energy penalty. A central component in chemical-looping is the oxygen carrier particles, which transfer oxygen from the air reactor to the fuel reactor. It is, therefore, important to look for new materials that might perform better than those used today. How suitable a material is as an oxygen carrier is largely determined by its thermodynamic properties, as this governs the maximum extent of fuel conversion in the fuel reactor. Gathering data to estimate thermodynamic properties is sometimes difficult or even impossible using experimental methods. CaMnO3-delta is an example of a complex material that has shown promise as an oxygen carrier for chemical-looping combustion and chemical-looping with oxygen uncoupling. Still, the fundamentals of this system are not fully understood, and available thermodynamic data show significant variations. Computational modeling is a method for generating thermodynamic data of equivalent accuracy to experimental investigations for complex systems. In this study, first-principles calculations by using density functional theory (DFT) and three different functionals (GGA + U, SCAN, r2SCAN) have been used to investigate the perovskite CaMnO3-delta. Our unique approach allows for the detailed examination of individual oxygen vacancies and their impact on thermodynamic properties, which is critical for understanding oxygen transport in chemical-looping. This is important since vacancies largely govern the oxygen transport needed in chemical-looping. The generated data is used in multiphase thermodynamic calculations to achieve greater detail in the predicted Ca-Mn-O phase-diagrams. The calculated results using SCAN show good agreement with data found in commercial databases and experimental data when comparing heat capacity and entropy, with differences of -1.17 and -11.17 J/(molK), respectively, for delta = 0 and 298.15 K. However, a larger discrepancy is shown for the estimated formation enthalpy at the same conditions, +26 kJ/mol compared to data from a commercial database. The results using r2SCAN are within 0.5% of the results using SCAN. The functional GGA + U estimates larger differences than SCAN and r2SCAN, +0.77, -5.43 J/(molK), and +48 kJ/mol for heat capacity, entropy and formation enthalpy, respectively. This indicates that CaMnO3 is less stable than previously proposed in the literature.

The large-scale ammonia synthesis using the Haber-Bosch process is crucial in modern society and the reaction is known to be facile over Ru-based catalysts. Herein, first-principles kinetic Monte Carlo (kMC) simulations are utilized to explore the reaction kinetics on Ru nanoparticles (NPs), extending the current knowledge that is mainly based on calculations of single crystal surfaces. It is only by accounting for the effects of kinetic couplings between different sites and inherent strain in the NPs that experimental turnover frequencies (TOFs) can be reproduced. The enhanced activity of inherently strained NPs is attributed to the co-existence of sites with both tensile and compressive strain, which simultaneously promotes N2 dissociation and NHx (x = 0, 1 and 2) hydrogenation. We propose that kinetic couplings on Ru NPs with tailored strain-patterns offer a strategy to break the limitations of linear scaling relations in the design of ammonia synthesis catalysts.

The dynamic character of heterogeneous catalyst particles makes direct comparisons between first-principles kinetics and experimental results obtained for technical catalysts challenging. First-principles kinetics is commonly based on a single model structure and constant reaction conditions, whereas experiments are performed over a particle distribution with shapes that respond to the reaction conditions. Here, we develop a framework for particle-shape adaptive kinetic Monte Carlo simulations in a reactor model (PAKS-R), which integrates first-principles-based kinetic Monte Carlo (kMC) simulations with a reactor model. The framework bridges the gap between the experimental situation by allowing for (i) particle size distributions, (ii) reaction conditions that change along the reactor, and (iii) dynamic shape changes of the NPs as a response to the coverages. The method is applied to ammonia synthesis over Ru NPs, reproducing the previous experimental reaction kinetics. The results show that the activity depends sensitively on the particle size and reaction conditions. The effect of dynamical shape changes is, on average, limited but strongly particle dependent. The PAKS-R approach is robust and general and can be used to explore the reaction kinetics of complex, technical catalysts for a range of reactions.

To address the need for a sustainable chemical industry, commodity chemicals including the aromatics benzene, toluene, and xylenes (BTX) must be produced from renewable feedstocks such as biomass-derived furans. Here, the conversion of 2,5-dimethylfuran (2,5-dmf) into aromatics was studied by step-response experiments in a chemical flow reactor, catalyzed by a series of phase-pure MFI framework Ga-zeotype catalysts with a gallium content ranging from 0.5 to 11 wt %. The lifetime of the catalyst and its aromatic production increase with increasing gallium content, demonstrating a near-40-fold increase in benzene production when increasing the gallium content from 0.5 to 8.6 wt %, while a further increase to 11 wt % leads to a decrease in benzene production due to rapid deactivation of the catalyst by coke. Acid site analysis reveals that aromatization occurs on strong Br & oslash;nsted acid sites, promoted by strong Lewis acid sites, while isomerization occurs on weak Br & oslash;nsted acid sites. At high gallium content (>= 6.0 wt %), gallium-based nanoparticles are formed, whose presence results in faster catalyst deactivation. The catalysts were active for five consecutive cycles and were readily regenerated, recovering the majority of their initial acid sites.

Methanol is a liquid energy carrier that has the potential to reduce the use of fossil fuels. Industrial production of methanol is currently a multistep high-temperature/high-pressure synthesis route. Direct conversion of methane to methanol under low-temperature and low-pressure conditions is an interesting but challenging alternative, which presently lacks suitable catalysts. Here, the complete reaction cycle for direct methane-to-methanol conversion over transition-metal dimers in the chabazite zeolite is studied by using density functional theory calculations and microkinetic modeling. In particular, a reaction mechanism previously identified for the Cu(2 )dimer is explored under dry and wet conditions for dimers composed of Ag, Au, Pd, Ni, Co, Fe, and Zn and the bimetallic dimers AuCu, PdCu, and AuPd. The density-functional-theory-based microkinetic modeling shows that Cu-2, AuPd, and PdCu dimers have reasonable turnover frequencies under technologically relevant conditions. The adsorption energy of atomic oxygen is identified as a descriptor for the reaction landscape as it correlates with the adsorption and transition-state energies of the other reaction intermediates. Using the established scaling relations, a volcano plot of the rate is generated with its apex close to the Cu-2, AuPd, and PdCu dimers.

As humankind searches for sustainable energy solutions, the demand for electrochemistry has increased. Thus, new and more advanced electrode materials are required. However, finding electrodes that meet the necessary performance is a challenge. Machine learning models can predict key properties such as catalytic activity and stability with surprisingly good accuracy, thus accelerating the process of evaluating materials. However, in most cases, the same models cannot explain how to generate new material compositions. Here, deep generative models can become very valuable. Although issues related to data availability and understanding how these models work still exist, combining deep generative models with computer simulations and laboratory experiments holds great potential for developing the next generation of electrodes. This short review will show recent progress in using deep generative models in related material fields and stress how these models can accelerate the discovery of electrode materials.

Chemical looping combustion (CLC) is an innovative technology suitable for converting waste-derived fuels into heat and power. The process inherently produces pure CO2, which is highly favorable for carbon capture and storage and could be instrumental for achieving negative emissions. CLC operates by utilizing solid oxygen carriers (OCs) to transfer heat and oxygen between two reactors. The OC play a crucial role in achieving an efficient combustion. Manganese-based OCs are particularly interesting, due to their ability to release gaseous oxygen. However, ash components from solid fuels could alter their oxygen transfer capacity, and cause problems related to corrosion and agglomeration. The objective of this work is to obtain in-depth insights about Mn-based OCs for CLC of waste-derived fuels. This is achieved by investigating phase transitions during CLC of solid fuels when utilizing two manganese-based OCs: manganese oxide and a representative manganese ore. For this purpose, thermodynamic modeling is employed, and a specific focus is given to K, Na, Cu, Zn, and Pb, due to their important role in corrosion and/or agglomeration. Thermodynamic databases are expanded by calculating properties from first-principles. It is shown that Mn-based OCs are suitable for effectively converting waste-derived fuels while limiting corrosion. Furthermore, the iron in manganese ores is found to have positive implications for oxygen-transfer reactions. In terms of alkali release to the gas phase, manganese ore seems to be a more promising material compared to manganese oxide. The pathways for the heavy metals Zn, Cu, and Pb were, meanwhile, independent of the OC type.

2,4-dimethylfuran has a rare disubstitution pattern in the five-membered heterocyclic furan ring that is highly interesting chemically but challenging to access synthetically. We present a heterogeneously catalysed route to synthesise 2,4-dimethylfuran from commonly available 2,5-dimethylfuran using a zeolite packed-bed flow reactor. As supported by DFT calculations, the reaction occurs inside the zeolite channels, where the acid sites catalyse proton transfer followed by migration of a methyl group. The zeotype Ga-silicate (MFI type) appears superior to an aluminium-containing ZSM-5 by demonstrating higher selectivities and slower catalyst deactivation. This work provides new opportunities for the continuous valorisation of bio-feedstock molecules in the perspective of the emerging biorefinery era. MFI framework zeolites and zeotypes enable the isomerisation of 2,5-dimethylfuran to 2,4-dimethylfuran through a heterogeneously catalysed route. Incorporating gallium instead of aluminium increases the selectivity and durability of the catalyst. This reaction is promoted by suitable acidity of the catalyst and steric confinement in the micropores. image

In this study, an efficient first-principles approach for calculating the thermodynamic properties of mixed metal oxides at high temperatures is demonstrated. More precisely, this procedure combines density functional theory and harmonic phonon calculations with tabulated thermochemical data to predict the heat capacity, formation energy, and entropy of important metal oxides. Alloy cluster expansions are, moreover, employed to represent phases that display chemical ordering as well as to calculate the configurational contribution to the specific heat capacity. The methodology can, therefore, be applied to compounds with vacancies and variable site occupancies. Results are, moreover, presented for a number of systems of high practical relevance: FeKTiO, KMnO, and CaMnO. For the reference materials, the agreement with experimental measurements is exceptional in the case of ilmenite (FeTiO3) and good for CaMnO3. When the generated data is used in multi-phase thermodynamic calculations to represent materials for which experimental data is not available, the predicted phase-diagrams for the KMnO and KTiO systems change dramatically. The demonstrated methodology is highly useful for obtaining approximate values on key thermodynamic properties in cases where experimental data is hard to obtain, inaccurate or missing.

Ethylene epoxidation is industrially and commercially one of the most important selective oxidations. Silver catalysts have been state-of-the-art for decades, their efficiency steadily improving with empirical discoveries of dopants and co-catalysts. Herein, we perform a computational screening of the metals in the periodic table, identify prospective superior catalysts and experimentally demonstrate that Ag/CuPb, Ag/CuCd and Ag/CuTl outperform the pure-Ag catalysts, while they still confer an easily scalable synthesis protocol. Furthermore, we show that to harness the potential of computationally-led discovery of catalysts fully, it is essential to include the relevant in situ conditions e.g., surface oxidation, parasitic side reactions and ethylene epoxide decomposition, as neglecting such effects leads to erroneous predictions. We combine ab initio calculations, scaling relations, and rigorous reactor microkinetic modelling, which goes beyond conventional simplified steady-state or rate-determining modelling on immutable catalyst surfaces. The modelling insights have enabled us to both synthesise novel catalysts and theoretically understand experimental findings, thus, bridging the gap between first-principles simulations and industrial applications. We show that the computational catalyst design can be easily extended to include larger reaction networks and other effects, such as surface oxidations. The feasibility was confirmed by experimental agreement.

Chemical looping is an innovative technique that relies, to a large extent, on the possibility of finding new oxygen carriers. Until now, these materials have primarily been identified via experimental techniques and therefrom derived insights. However, this is both costly and time-consuming. To speed-up this process, we have applied a computational screening approach based on energetic data retrieved from the Open Quantum Materials Database. In particular, we have considered combinations of all mono-, bi-, and trimetallic alloys and mixed oxides with up to three distinctive phases. Here, we specifically focus on a technique referred to as chemical looping oxygen uncoupling, which is especially suitable for solid fuels, e.g., combustion of biomass for negative CO2 emissions. The formation energies obtained for the materials of interest were used to identify phase transitions that are likely to occur under conditions relevant for chemical looping oxygen uncoupling. Given these criteria, the initial list of 300000 materials is reduced by a factor of 20, and after filtering out rare, radioactive, toxic, or harmful elements only 1000 remain. When considering the abundance of elements in the ranking criteria, most of the highest ranking phases include Cu, Mn, and Fe. This adds credibility to the procedure, as many viable oxygen carriers for chemical looping oxygen uncoupling that have been studied experimentally contain these elements. While Cr-based materials have not been widely explored for this application, our study suggests that this might be worthwhile since these occur more frequently than Fe. Other elements that would be interesting as additional components include Ba, K, Na, Al, and Si.

Catalyst activity can depend distinctly on nanoparticle size and shape. Therefore, understanding the structure sensitivity of catalytic reactions is of fundamental and technical importance. Experiments with single-particle resolution, where ensemble-averaging is eliminated, are required to study it. Here, we implement the selective trapping of individual spherical, cubic, and octahedral colloidal Au nanocrystals in 100 parallel nanofluidic channels to determine their activity for fluorescein reduction by sodium borohydride using fluorescence microscopy. As the main result, we identify distinct structure sensitivity of the rate-limiting borohydride oxidation step originating from different edge site abundance on the three particle types, as confirmed by first-principles calculations. This advertises nanofluidic reactors for the study of structure-function correlations in catalysis and identifies nanoparticle shape as a key factor in borohydride-mediated catalytic reactions.