Search

Abstract Understanding organic reaction mechanisms is crucial for interpreting the formation of products at the atomic and electronic level, but still remains as a domain of knowledgeable experts. The lack of a large-scale dataset with chemically reasonable mechanistic sequences also hinders the development of reliable machine learning models to predict organic reactions based on mechanisms as human chemists do. Here, we present a high-quality and the first large-scale reaction dataset, denoted as mech-USPTO-31K, with chemically reasonable arrow-pushing diagrams validated by synthetic chemists, encompassing a wide spectrum of polar organic reaction mechanisms. We envision this dataset curated by applying a simple and flexible method that automatically generates reaction mechanisms using autonomously extracted reaction templates and expert-coded mechanistic templates to become an invaluable tool to develop future reaction outcome prediction models and discover new reactions.

Abstract Synthetic accessibility prediction is a task to estimate how easily a given molecule might be synthesizable in the laboratory, playing a crucial role in computer-aided molecular design. Although synthesis planning programs can determine synthesis routes, their slow processing times make them impractical for large-scale molecule screening. On the other hand, existing rapid synthesis accessibility estimation methods offer speed but typically lack integration with actual synthesis routes and building block information. In this work, we introduce BR-SAScore, an enhanced version of SAScore that integrates the available building block information (B) and reaction knowledge (R) from synthesis planning programs into the scoring process. In particular, we differentiate fragments inherent in building blocks and fragments to be derived from synthesis (reactions) when scoring synthetic accessibility. Compared to existing methods, our experimental findings demonstrate that BR-SAScore offers more accurate and precise identification of a molecule's synthetic accessibility by the synthesis planning program with a fast calculation time. Moreover, we illustrate how BR-SAScore provides chemically interpretable results, aligning with the capability of the synthesis planning program embedded with the same reaction knowledge and available building blocks. Scientific contribution We introduce BR-SAScore, an extension of SAScore, to estimate the synthetic accessibility of molecules by leveraging known building-block and reactivity information. In our experiments, BR-SAScore shows superior prediction performance on predicting molecule synthetic accessibility compared to previous methods, including SAScore and deep-learning models, while requiring significantly less computation time. In addition, we show that BR-SAScore is able to precisely identify the chemical fragment contributing to the synthetic infeasibility, holding great potential for future molecule synthesizability optimization.

Abstract Single‐atom catalysts (SACs), featuring highly uniform active sites, tunable coordination environments, and synergistic effects with support, have emerged as one of the most efficient catalysts for various reactions, particularly for electrochemical CO2 reduction (ECR). However, the scalability of SACs is restricted due to the limited choice of available support and problems that emerge when preparing SACs by thermal deposition. Here, an in situ reconstruction method for preparing SACs is developed with a variety of atomic sites, including nickel, cadmium, cobalt, and magnesium. Driven by electricity, different oxygen‐containing metal precursors, such as MOF‐74 and metal oxides, are directly atomized onto nitrogen‐doped carbon (NC) supports, yielding SACs with variable metal active sites and coordination structures. The electrochemical force facilitates the in situ generation of bonds between the metal and the supports without the need for additional complex steps. A series of MNxOy (M denotes metal) SACs on NC have been synthesized and utilized for ECR. Among these, NiNxOy SACs using Ni‐MOF‐74 as a metal precursor exhibit excellent ECR performance. This universal and general SAC synthesis strategy at room temperature is simpler than most reported synthesis methods to date, providing practical guidance for the design of the next generation of high‐performance SACs.

Abstract Atom-to-atom mapping (AAM) is a task of identifying the position of each atom in the molecules before and after a chemical reaction, which is important for understanding the reaction mechanism. As more machine learning (ML) models were developed for retrosynthesis and reaction outcome prediction recently, the quality of these models is highly dependent on the quality of the AAM in reaction datasets. Although there are algorithms using graph theory or unsupervised learning to label the AAM for reaction datasets, existing methods map the atoms based on substructure alignments instead of chemistry knowledge. Here, we present LocalMapper, an ML model that learns correct AAM from chemist-labeled reactions via human-in-the-loop machine learning. We show that LocalMapper can predict the AAM for 50 K reactions with 98.5% calibrated accuracy by learning from only 2% of the human-labeled reactions from the entire dataset. More importantly, the confident predictions given by LocalMapper, which cover 97% of 50 K reactions, show 100% accuracy for 3,000 randomly sampled reactions. In an out-of-distribution experiment, LocalMapper shows favorable performance over other existing methods. We expect LocalMapper can be used to generate more precise reaction AAM and improve the quality of future ML-based reaction prediction models.

ABSTRACT: The effect of cinnamon powder on the quality and mitigation of off-flavor in fried chicken drumsticks made from long-term thawed Korean native chicken (Woorimatdag No. 1, WRMD1) was investigated. The WRMD1 drumsticks were categorized into 5 groups: conventional thawing (16 h, CT), long-term thawing (48 h, LT), cinnamon powder added into ‘LT’ as marinade (0.03%, CM) or incorporated into the batter (1.35%, CB), and long-term thawing with cinnamon powder incorporated both in the marinade and batter (0.03% + 1.35%, CMB). The crude fat content was significantly higher in the CT and CMB than that of the CB. The CM, CB, and CMB showed significantly lower levels of 2-thiobarbituric acid reactive substance compared with the CT and LT. The predominant fatty acids in all treatments were C18:1n9, C18:2n6, and C16:0. The LT displayed lower total unsaturated fatty acid content than the CT (P < 0.05). The CM effectively decreased lipid oxidative volatiles, such as 1-octanol, 1-octen-3-ol, and 2-octen-1-ol, (E), in the LT (P < 0.05). Both the CM and CB showed an inclination to increase specific pyrazines associated with pleasant notes compared with the LT, and showed higher levels of pyrazines, such as pyrazine, 2-ethyl-6-methyl-, and pyrazine, 3-ethyl-2,5-dimethyl-, than those of the CMB (P < 0.05). The CM contained higher levels of 2,3-butanedione when compared with the other groups (P < 0.05). Multivariate analysis demonstrated that cinnamon had an effect in discriminating the treatment groups with cinnamon addition from both the CT and LT, whereas the CM, CB, and CMB formed distinct clusters. The CM and CMB received significantly higher aroma scores from panelists in comparison to the other groups. These findings suggest that the CM (0.03% cinnamon powder) can be used to enhance the aroma in fried WRMD1 drumsticks by reducing or masking the off-flavor volatiles associated with long-term thawing.

Abstract High-throughput virtual screening for crystals aims to discover new materials by evaluating the property of every virtual candidate in the database exhaustively. During this process, the major computational bottleneck is the costly structural relaxation of each hypothetical material on the large-scale dataset using density functional theory (DFT) calculations. Here, we present a generative domain translation framework that maps the unrelaxed structural domains to the relaxed domains, enabling data-driven structural translations. The model predicts the materials formation energy with a small mean absolute error without DFT relaxations, and furthermore can produce the atomic coordinates consistent with the DFT relaxed structures. The utility of the proposed concept is not restricted to the structural domains, and we expect that it can be extended to translate the domain of easy-to-compute properties into the domain of more difficult properties.

ABSTRACT: This study was aimed to identify and compare the taste-related compounds (nucleotide-related compounds, free amino acid contents, and fatty acid composition) and aroma (volatile organic compounds [VOC]) compounds in the chicken breast meat from 3 kinds of Korean native chicken (KNC), namely Hanhyup 3 (HH3), Woorimatdag 1 (WRMD1) and Woorimatdag 2 (WRMD2). Among the 3 breeds, WRMD1 had significantly higher IMP and AMP contents than HH3. WRMD2 exhibited higher levels of umami and sweet-taste amino acids and oleic acid composition compared to HH3 (P < 0.05). HH3 showed a higher composition of unsaturated fatty acids than WRMD2 (P < 0.05). On their discrimination by flavor composition, some compounds including aspartic acid were analyzed as important compounds. Regarding aroma compounds, unique aroma compounds were detected for each breed and some compounds such as isopropyl myristate, p-cresol, (S)-(+)-3-Methyl-1-pentanol, and cyclic octa-atomic sulfur were expected to be utilized as key compounds in discrimination of the 3 breeds. From the result of this study, the differences on the flavor compounds of three breeds were elucidated and key compounds for their discrimination were presented.

The purpose of this study was to evaluate the effect of temperature and time of sous-vide cooking method on the characteristics of Thoroughbred horse loin. Sliced portions (200 ± 50 g) were cooked by boiling (control) and sous-vide (65 and 70 °C for 12, 18, and 24 h). The samples were analyzed for proximate composition, pH, color, texture, microstructure, sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), microbiology, volatile organic compounds (VOCs), nucleotide content, and fatty acids composition. The color analysis showed decreased redness at elevated temperatures. Improved tenderness, demonstrated by reduced shear force values (36.36 N at 65 °C for 24 h and 35.70 N at 70 °C for 24 h). The micrographs indicated dense fiber arrangements at 70 °C. The SDS-PAGE revealed muscle protein degradation with extended sous-vide cooking. The VOC analysis identified specific compounds, potentially distinctive markers for sous-vide cooking of horse meat including 1-octen-3-ol, decanal, n-caproic acid vinyl ester, cyclotetrasiloxane, octamethyl, and 3,3-dimethyl-1,2-epoxybutane. This study highlights the cooking time’s primary role in sous vide-cooked horse meat tenderness and proposes specific VOCs as potential markers. Further research should explore the exclusivity of these VOCs to sous-vide cooking.

Abstract Perovskite is an important material type in geophysics and for technologically important applications. However, the number of synthetic perovskites remains relatively small. To accelerate the high-throughput discovery of perovskites, we propose a graph neural network model to assess their synthesizability. Our trained model shows a promising 0.957 out-of-sample true positive rate, significantly improving over empirical rule-based methods. Further validation is established by demonstrating that a significant portion of the virtual crystals that are predicted to be synthesizable have already been indeed synthesized in literature, and those with the lowest synthesizability scores have not been reported. While previous empirical strategies are mainly applicable to metal oxides, our model is general and capable of predicting the synthesizability across all classes of perovskites, including chalcogenide, halide, and hydride perovskites, as well as anti-perovskites. We apply the method to identify synthesizable perovskite candidates for two potential applications, the Li-rich ion conductors and metal halide optical materials that can be tested experimentally.

The discovery of heterogeneous catalysts for large molecule conversion has been lagging due to the combinatorial inventory of intermediates. Here, the author presents an automated framework to explore the chemical space of reaction intermediates.

Abstract Electrochemical CO2 transformation to high‐value ethylene (C2H4) at high currents and efficiencies is desired and yet remains a grand challenge. We show for the first time that coupling single Sb atoms and oxygen vacancies of CuO enable synergistic electrocatalytic reduction of CO2 to C2H4 at low overpotentials. Highly dispersed Sb atoms occupying metal substitutional sites of CuO are synthesized under mild conditions. The overall CO2 reduction faradaic efficiency (FE) reaches 89.3 ± 1.1% with an FE toward C2H4 exceeding 58.4% at a high‐current density of 500 mA/cm2. Addition of the p‐block metal is found to induce transformation of CuO from flakes to nanoribbons rich in nanoholes and oxygen vacancies, greatly enhancing CO2 adsorption and activation while suppressing hydrogen evolution. Further density functional theory calculations with in situ X‐ray diffraction reveal that combining Sb sites and oxygen vacancies prominently lessen the dimerization energy of adsorbed CO intermediate, thus boosting the conversion of CO2 to produce C2H4. This study provides a new perspective for promoting selective C–C coupling for electrochemical CO2 reduction.

Practical electrochemical N2 reduction reaction is challenged by competing side reactions. Here a combination of DFT and mikrokinetic modelling reveals the potential-dependent competition between electrochemical ammonia production and hydrogen evolution on a single-site iron catalyst embedded in N-doped graphene.

Chemical interaction between metal and oxide supports is an important molecular-level factor that influences the catalytic selectivity of a desirable reaction. Here, using Pt nanowires/TiO2 catalytic nanodiodes, the authors investigate an enhancement of both selectivity and hot electron generation on metal-oxide interfacial sites.

Reliable unzipping of the basal plane of mono-elemental graphene and phosphorene has only been reported thus far. Here, the authors demonstrate the unzipping of bi-elemental 2D MX2 transition metal chalcogenides as a general route to synthesize 1D nanoribbon structures.

Demand for ammonia continues to increase to sustain the growing global population. The direct electrochemical N2 reduction reaction (NRR) powered by renewable electricity offers a promising carbon-neutral and sustainable strategy for manufacturing NH3, yet achieving this remains a grand challenge. Here, we report a synergistic strategy to promote ambient NRR for ammonia production by tuning the Te vacancies (VTe) and surface hydrophobicity of two-dimensional TaTe2 nanosheets. Remarkable NH3 faradic efficiency of up to 32.2% is attained at a mild overpotential, which is largely maintained even after 100 h of consecutive electrolysis. Isotopic labeling validates that the N atoms of formed NH4+ originate from N2. In situ X-ray diffraction indicates preservation of the crystalline structure of TaTe2 during NRR. Further density functional theory calculations reveal that the potential-determining step (PDS) is ∗NH2 + (H+ + e–) → NH3 on VTe-TaTe2 compared with that of ∗ + N2 + (H+ + e–) → ∗N–NH on TaTe2. We identify that the edge plane of TaTe2 and VTe serve as the main active sites for NRR. The free energy change at PDS on VTe-TaTe2 is comparable with the values at the top of the NRR volcano plots on various transition metal surfaces. Public summary: • 2D TaTe2 is produced in large quantities • Jointly tuning the Te vacancies (VTe) and surface hydrophobicity of 2D TaTe2 enables efficient and stable electrocatalytic NRR with remarkable NH3 faradic efficiency • The edge plane of TaTe2 and VTe serve as the main active sites for NRR • The free energy change at the potential-determining step on VTe-TaTe2 is comparable with the values at the top of the NRR volcano plots on various transition metal surfaces

The real-time quantitative detection of hot electrons provides critical clues to understand the origin of the enhanced catalytic performance of bimetallic nanoparticles (NPs). Here, the authors investigate hot electrons generated on bimetallic PtCo NPs during H2 oxidation by measuring the chemicurrent on a catalytic nanodiode.

While lithium-oxygen batteries offer a green method to power vehicles, the sluggish decomposition of lithium peroxide limits device performance. Here, the authors direct lithium peroxide formation into amorphous nanostructures to enable its facile decomposition and improve charging efficiency.

Production of affordable, clean hydrogen relies on efficient oxygen evolution, but improving catalytic performance for the reaction in acidic media is challenging. Here the authors show how tuning the nanoporous morphology of iridium/iridium oxide leads to an improvement in activity/stability, compared with conventional iridium-based oxides.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.