Predicting Ion Diffusion from the Shape of Potential Energy Landscapes.

We present an efficient method to compute diffusion coefficients of multiparticle systems with strong interactions directly from the geometry and topology of the potential energy field of the migrating particles. The approach is tested on Li-ion diffusion in crystalline inorganic solids, predicting Li-ion diffusion coefficients within 1 order of magnitude of molecular dynamics simulations at the same level of theory while being several orders of magnitude faster. The speed and transferability of our workflow make it well-suited for extensive and efficient screening studies of crystalline solid-state ion conductor candidates and promise to serve as a platform for diffusion prediction even up to the density functional level of theory.

14 examples of how LLMs can transform materials science and chemistry: a reflection on a large language model hackathon.

Large-language models (LLMs) such as GPT-4 caught the interest of many scientists. Recent studies suggested that these models could be useful in chemistry and materials science. To explore these possibilities, we organized a hackathon. This article chronicles the projects built as part of this hackathon. Participants employed LLMs for various applications, including predicting properties of molecules and materials, designing novel interfaces for tools, extracting knowledge from unstructured data, and developing new educational applications. The diverse topics and the fact that working prototypes could be generated in less than two days highlight that LLMs will profoundly impact the future of our fields. The rich collection of ideas and projects also indicates that the applications of LLMs are not limited to materials science and chemistry but offer potential benefits to a wide range of scientific disciplines.

A Robust Framework for Generating Adsorption Isotherms to Screen Materials for Carbon Capture.

To rank the performance of materials for a given carbon capture process, we rely on pure component isotherms from which we predict the mixture isotherms. For screening a large number of materials, we also increasingly rely on isotherms predicted from molecular simulations. In particular, for such screening studies, it is important that the procedures to generate the data are accurate, reliable, and robust. In this work, we develop an efficient and automated workflow for a meticulous sampling of pure component isotherms. The workflow was tested on a set of metal-organic frameworks (MOFs) and proved to be reliable given different guest molecules. We show that the coupling of our workflow with the Clausius-Clapeyron relation saves CPU time, yet enables us to accurately predict pure component isotherms at the temperatures of interest, starting from a reference isotherm at a given temperature. We also show that one can accurately predict the CO2 and N2 mixture isotherms using ideal adsorbed solution theory (IAST). In particular, we show that IAST is a more reliable numerical tool to predict binary adsorption uptakes for a range of pressures, temperatures, and compositions, as it does not rely on the fitting of experimental data, which typically needs to be done with analytical models such as dual-site Langmuir (DSL). This makes IAST a more suitable and general technique to bridge the gap between adsorption (raw) data and process modeling. To demonstrate this point, we show that the ranking of materials, for a standard three-step temperature swing adsorption (TSA) process, can be significantly different depending on the thermodynamic method used to predict binary adsorption data. We show that, for the design of processes that capture CO2 from low concentration (0.4%) streams, the commonly used methodology to predict mixture isotherms incorrectly assigns up to 33% of the materials as top-performing.

An Ecosystem for Digital Reticular Chemistry.

The vastness of the materials design space makes it impractical to explore using traditional brute-force methods, particularly in reticular chemistry. However, machine learning has shown promise in expediting and guiding materials design. Despite numerous successful applications of machine learning to reticular materials, progress in the field has stagnated, possibly because digital chemistry is more an art than a science and its limited accessibility to inexperienced researchers. To address this issue, we present mofdscribe, a software ecosystem tailored to novice and seasoned digital chemists that streamlines the ideation, modeling, and publication process. Though optimized for reticular chemistry, our tools are versatile and can be used in nonreticular materials research. We believe that mofdscribe will enable a more reliable, efficient, and comparable field of digital chemistry.

Tandem Hydroformylation-Aldol Condensation Reaction Enabled by Zn-MOF-74.

The tandem hydroformylation-aldol condensation (tandem HF-AC) reaction offers an efficient synthetic route to the synthesis of industrially relevant products. The addition of Zn-MOF-74 to the cobalt-catalyzed hydroformylation of 1-hexene enables tandem HF-AC under milder pressure and temperature conditions than the aldox process, where zinc salts are added to cobalt-catalyzed hydroformylation reactions to promote aldol condensation. The yield of the aldol condensation products increases by up to 17 times compared to that of the homogeneous reaction without MOF and up to 5 times compared to the aldox catalytic system. Both Co2 (CO)8 and Zn-MOF-74 are required to significantly enhance the activity of the catalytic system. Density functional theory simulations and Fourier-transform infrared experiments show that heptanal, the product of hydroformylation, adsorbs on the open metal site (OMS) of Zn-MOF-74, thereby increasing the electrophilic character of the carbonyl carbon atom and facilitating the condensation.

Highly selective CO2 photoreduction to CO on MOF-derived TiO2.

Metal-Organic Framework (MOF)-derived TiO2, synthesised through the calcination of MIL-125-NH2, is investigated for its potential as a CO2 photoreduction catalyst. The effect of the reaction parameters: irradiance, temperature and partial pressure of water was investigated. Using a two-level design of experiments, we were able to evaluate the influence of each parameter and their potential interactions on the reaction products, specifically the production of CO and CH4. It was found that, for the explored range, the only statistically significant parameter is temperature, with an increase in temperature being correlated to enhanced production of both CO and CH4. Over the range of experimental settings explored, the MOF-derived TiO2 displays high selectivity towards CO (98%), with only a small amount of CH4 (2%) being produced. This is notable when compared to other state-of-the-art TiO2 based CO2 photoreduction catalysts, which often showcase lower selectivity. The MOF-derived TiO2 was found to have a peak production rate of 8.9 × 10-4 µmol cm-2 h-1 (2.6 µmol g-1 h-1) and 2.6 × 10-5 µmol cm-2 h-1 (0.10 µmol g-1 h-1) for CO and CH4, respectively. A comparison is made to commercial TiO2, P25 (Degussa), which was shown to have a similar activity towards CO production, 3.4 × 10-3 µmol cm-2 h-1 (5.9 µmol g-1 h-1), but a lower selectivity preference for CO (3 : 1 CH4 : CO) than the MOF-derived TiO2 material developed here. This paper showcases the potential for MIL-125-NH2 derived TiO2 to be further developed as a highly selective CO2 photoreduction catalyst for CO production.

Metal-organic framework-based oxygen carriers with antioxidant activity resulting from the incorporation of gold nanozymes.

Blood transfusions are a life-saving procedure since they can preserve the body's oxygen levels in patients suffering from acute trauma, undergoing surgery, receiving chemotherapy or affected by severe blood disorders. Due to the central role of hemoglobin (Hb) in oxygen transport, so-called Hb-based oxygen carriers (HBOCs) are currently being developed for situations where donor blood is not available. In this context, an important challenge that needs to be addressed is the oxidation of Hb into methemoglobin (metHb), which is unable to bind and release oxygen. While several research groups have considered the incorporation of antioxidant enzymes to create HBOCs with minimal metHb conversion, the use of biological enzymes has important limitations related to their high cost, potential immunogenicity or low stability in vivo. Thus, nanomaterials with enzyme-like properties (i.e., nanozymes (NZs)) have emerged as a promising alternative. Amongst the different NZs, gold (Au)-based metallic nanoparticles are widely used for biomedical applications due to their biocompatibility and multi-enzyme mimicking abilities. Thus, in this work, we incorporate Au-based NZs into a type of HBOC previously reported by our group (i.e., Hb-loaded metal-organic framework (MOF)-based nanocarriers (NCs)) and investigate their antioxidant properties. Specifically, we prepare MOF-NCs loaded with Au-based NZs and demonstrate their ability to catalytically deplete over multiple rounds of two prominent reactive oxygen species (ROS) that exacerbate Hb's autoxidation (i.e., hydrogen peroxide and the superoxide radical). Importantly, following loading with Hb, we show how these ROS-scavenging properties translate into a decrease in metHb content. All in all, these results highlight the potential of NZs to create novel HBOCs with antioxidant protection which may find applications as a blood substitute in the future.

Machine learning for industrial processes: Forecasting amine emissions from a carbon capture plant.

One of the main environmental impacts of amine-based carbon capture processes is the emission of the solvent into the atmosphere. To understand how these emissions are affected by the intermittent operation of a power plant, we performed stress tests on a plant operating with a mixture of two amines, 2-amino-2-methyl-1-propanol and piperazine (CESAR1). To forecast the emissions and model the impact of interventions, we developed a machine learning model. Our model showed that some interventions have opposite effects on the emissions of the components of the solvent. Thus, mitigation strategies required for capture plants operating on a single component solvent (e.g., monoethanolamine) need to be reconsidered if operated using a mixture of amines. Amine emissions from a solvent-based carbon capture plant are an example of a process that is too complex to be described by conventional process models. We, therefore, expect that our approach can be more generally applied.

A data-science approach to predict the heat capacity of nanoporous materials.

The heat capacity of a material is a fundamental property of great practical importance. For example, in a carbon capture process, the heat required to regenerate a solid sorbent is directly related to the heat capacity of the material. However, for most materials suitable for carbon capture applications, the heat capacity is not known, and thus the standard procedure is to assume the same value for all materials. In this work, we developed a machine learning approach, trained on density functional theory simulations, to accurately predict the heat capacity of these materials, that is, zeolites, metal-organic frameworks and covalent-organic frameworks. The accuracy of our prediction is confirmed with experimental data. Finally, for a temperature swing adsorption process that captures carbon from the flue gas of a coal-fired power plant, we show that for some materials, the heat requirement is reduced by as much as a factor of two using the correct heat capacity.

SELFIES and the future of molecular string representations.

Artificial intelligence (AI) and machine learning (ML) are expanding in popularity for broad applications to challenging tasks in chemistry and materials science. Examples include the prediction of properties, the discovery of new reaction pathways, or the design of new molecules. The machine needs to read and write fluently in a chemical language for each of these tasks. Strings are a common tool to represent molecular graphs, and the most popular molecular string representation, Smiles, has powered cheminformatics since the late 1980s. However, in the context of AI and ML in chemistry, Smiles has several shortcomings-most pertinently, most combinations of symbols lead to invalid results with no valid chemical interpretation. To overcome this issue, a new language for molecules was introduced in 2020 that guarantees 100% robustness: SELF-referencing embedded string (Selfies). Selfies has since simplified and enabled numerous new applications in chemistry. In this perspective, we look to the future and discuss molecular string representations, along with their respective opportunities and challenges. We propose 16 concrete future projects for robust molecular representations. These involve the extension toward new chemical domains, exciting questions at the interface of AI and robust languages, and interpretability for both humans and machines. We hope that these proposals will inspire several follow-up works exploiting the full potential of molecular string representations for the future of AI in chemistry and materials science.

Characterization of Chemisorbed Species and Active Adsorption Sites in Mg-Al Mixed Metal Oxides for High-Temperature CO2 Capture.

Mg-Al mixed metal oxides (MMOs), derived from the decomposition of layered double hydroxides (LDHs), have been purposed as adsorbents for CO2 capture of industrial plant emissions. To aid in the design and optimization of these materials for CO2 capture at 200 °C, we have used a combination of solid-state nuclear magnetic resonance (ssNMR) and density functional theory (DFT) to characterize the CO2 gas sorption products and determine the various sorption sites in Mg-Al MMOs. A comparison of the DFT cluster calculations with the observed 13C chemical shifts of the chemisorbed products indicates that mono- and bidentate carbonates are formed at the Mg-O sites with adjacent Al substitution of an Mg atom, while the bicarbonates are formed at Mg-OH sites without adjacent Al substitution. Quantitative 13C NMR shows an increase in the relative amount of strongly basic sites, where the monodentate carbonate product is formed, with increasing Al/Mg molar ratios in the MMOs. This detailed understanding of the various basic Mg-O sites presented in MMOs and the formation of the carbonate, bidentate carbonate, and bicarbonate chemisorbed species yields new insights into the mechanism of CO2 adsorption at 200 °C, which can further aid in the design and capture capacity optimization of the materials.

Effects of Degrees of Freedom on Calculating Diffusion Properties in Nanoporous Materials.

If one carries out a molecular simulation of N particles using periodic boundary conditions, linear momentum is conserved, and hence, the number of degrees of freedom is set to 3N - 3. In most programs, this number of degrees of freedom is the default setting. However, if one carries out a molecular simulation in an external field, one needs to ensure that degrees of freedom are changed from this default setting to 3N, as in an external field the velocity of the center of mass can change. Using the correct degrees of freedom is important in calculating the temperature and in some algorithms to simulate at constant temperature. For sufficiently large systems, the difference between 3N and 3N - 3 is negligible. However, there are systems in which the comparison with experimental data requires molecular dynamics simulations of a small number of particles. In this work, we illustrate the effect of an incorrect setting of degrees of freedom in molecular dynamic simulations studying the diffusion properties of guest molecules in nanoporous materials. We show that previously published results have reported a surprising diffusion dependence on the loading, which could be traced back to an incorrect setting of the degrees of freedom. As the correct settings are convoluted and counterintuitive in some of the most commonly used molecular dynamics programs, we carried out a systematic study on the consequences of the various commonly used (incorrect) settings. Our conclusion is that for systems smaller than 50 particles the results are most likely unreliable as these are either performed at an incorrect temperature or the temperature is incorrectly used in some of the results. Furthermore, a novel and efficient method to calculate diffusion coefficients of guest molecules into nanoporous materials at zero-loading conditions is introduced.

Making the collective knowledge of chemistry open and machine actionable.

Large amounts of data are generated in chemistry labs-nearly all instruments record data in a digital form, yet a considerable proportion is also captured non-digitally and reported in ways non-accessible to both humans and their computational agents. Chemical research is still largely centred around paper-based lab notebooks, and the publication of data is often more an afterthought than an integral part of the process. Here we argue that a modular open-science platform for chemistry would be beneficial not only for data-mining studies but also, well beyond that, for the entire chemistry community. Much progress has been made over the past few years in developing technologies such as electronic lab notebooks that aim to address data-management concerns. This will help make chemical data reusable, however it is only one step. We highlight the importance of centring open-science initiatives around open, machine-actionable data and emphasize that most of the required technologies already exist-we only need to connect, polish and embrace them.

Using genetic algorithms to systematically improve the synthesis conditions of Al-PMOF.

The synthesis of metal-organic frameworks (MOFs) is often complex and the desired structure is not always obtained. In this work, we report a methodology that uses a joint machine learning and experimental approach to optimize the synthesis conditions of Al-PMOF (Al2(OH)2TCPP) [H2TCPP = meso-tetra(4-carboxyphenyl)porphine], a promising material for carbon capture applications. Al-PMOF was previously synthesized using a hydrothermal reaction, which gave a low throughput yield due to its relatively long reaction time (16 hours). Here, we use a genetic algorithm to carry out a systematic search for the optimal synthesis conditions and a microwave-based high-throughput robotic platform for the syntheses. We show that, in just two generations, we could obtain excellent crystallinity and yield close to 80% in a much shorter reaction time (50 minutes). Moreover, by analyzing the failed and partially successful experiments, we could identify the most important experimental variables that determine the crystallinity and yield.

Diversifying Databases of Metal Organic Frameworks for High-Throughput Computational Screening.

By combining metal nodes and organic linkers, an infinite number of metal organic frameworks (MOFs) can be designed in silico. Therefore, when making new databases of such hypothetical MOFs, we need to ensure that they not only contribute toward the growth of the count of structures but also add different chemistries to the existing databases. In this study, we designed a database of ∼20,000 hypothetical MOFs, which are diverse in terms of their chemical design space─metal nodes, organic linkers, functional groups, and pore geometries. Using machine learning techniques, we visualized and quantified the diversity of these structures. We find that on adding the structures of our database, the overall diversity metrics of hypothetical databases improve, especially in terms of the chemistry of metal nodes. We then assessed the usefulness of diverse structures by evaluating their performance, using grand-canonical Monte Carlo simulations, in two important environmental applications─post-combustion carbon capture and hydrogen storage. We find that many of these structures perform better than widely used benchmark materials such as Zeolite-13X (for post-combustion carbon capture) and MOF-5 (for hydrogen storage). All the structures developed in this study, and their properties, are provided on the Materials Cloud to encourage further use of these materials for other applications.

Toward Optimal Photocatalytic Hydrogen Generation from Water Using Pyrene-Based Metal-Organic Frameworks.

Metal-organic frameworks (MOFs) are promising materials for the photocatalytic H2 evolution reaction (HER) from water. To find the optimal MOF for a photocatalytic HER, one has to consider many different factors. For example, studies have emphasized the importance of light absorption capability, optical band gap, and band alignment. However, most of these studies have been carried out on very different materials. In this work, we present a combined experimental and computation study of the photocatalytic HER performance of a set of isostructural pyrene-based MOFs (M-TBAPy, where M = Sc, Al, Ti, and In). We systematically studied the effects of changing the metal in the node on the different factors that contribute to the HER rate (e.g., optical properties, the band structure, and water adsorption). In addition, for Sc-TBAPy, we also studied the effect of changes in the crystal morphology on the photocatalytic HER rate. We used this understanding to improve the photocatalytic HER efficiency of Sc-TBAPy, to exceed the one reported for Ti-TBAPy, in the presence of a co-catalyst.

Using collective knowledge to assign oxidation states of metal cations in metal-organic frameworks.

Knowledge of the oxidation state of metal centres in compounds and materials helps in the understanding of their chemical bonding and properties. Chemists have developed theories to predict oxidation states based on electron-counting rules, but these can fail to describe oxidation states in extended crystalline systems such as metal-organic frameworks. Here we propose the use of a machine-learning model, trained on assignments by chemists encoded in the chemical names in the Cambridge Structural Database, to automatically assign oxidation states to the metal ions in metal-organic frameworks. In our approach, only the immediate local environment around a metal centre is considered. We show that the strategy is robust to experimental uncertainties such as incorrect protonation, unbound solvents or changes in bond length. This method gives good accuracy and we show that it can be used to detect incorrect assignments in the Cambridge Structural Database, illustrating how collective knowledge can be captured by machine learning and converted into a useful tool.

Bias free multiobjective active learning for materials design and discovery.

The design rules for materials are clear for applications with a single objective. For most applications, however, there are often multiple, sometimes competing objectives where there is no single best material and the design rules change to finding the set of Pareto optimal materials. In this work, we leverage an active learning algorithm that directly uses the Pareto dominance relation to compute the set of Pareto optimal materials with desirable accuracy. We apply our algorithm to de novo polymer design with a prohibitively large search space. Using molecular simulations, we compute key descriptors for dispersant applications and drastically reduce the number of materials that need to be evaluated to reconstruct the Pareto front with a desired confidence. This work showcases how simulation and machine learning techniques can be coupled to discover materials within a design space that would be intractable using conventional screening approaches.

Enhanced Visible-Light-Driven Hydrogen Production through MOF/MOF Heterojunctions.

A strategy for enhancing the photocatalytic performance of MOF-based systems (MOF: metal-organic framework) is developed through the construction of MOF/MOF heterojunctions. The combination of MIL-167 with MIL-125-NH2 leads to the formation of MIL-167/MIL-125-NH2 heterojunctions with improved optoelectronic properties and efficient charge separation. MIL-167/MIL-125-NH2 outperforms its single components MIL-167 and MIL-125-NH2, in terms of photocatalytic H2 production (455 versus 0.8 and 51.2 µmol h-1 g-1, respectively), under visible-light irradiation, without the use of any cocatalysts. This is attributed to the appropriate band alignment of these MOFs, the enhanced visible-light absorption, and long charge separation within MIL-167/MIL-125-NH2. Our findings contribute to the discovery of novel MOF-based photocatalytic systems that can harvest solar energy and exhibit high catalytic activities in the absence of cocatalysts.