Machine-learned interatomic potentials (MLIPs) are revolutionizing computational materials science and chemistry by offering an efficient alternative to ab initio molecular dynamics (MD) simulations. However, fitting high-quality MLIPs remains a challenging, time-consuming, and computationally intensive task where numerous trade-offs have to be considered, e.g., How much and what kind of atomic configurations should be included in the training set? Which level of ab initio convergence should be used to generate the training set? Which loss function should be used for fitting the MLIP? Which machine learning architecture should be used to train the MLIP? The answers to these questions significantly impact both the computational cost of MLIP training and the accuracy and computational cost of subsequent MLIP MD simulations. In this study, we use a configurationally diverse beryllium dataset and quadratic spectral neighbor analysis potential. We demonstrate that joint optimization of energy versus force weights, training set selection strategies, and convergence settings of the ab initio reference simulations, as well as model complexity can lead to a significant reduction in the overall computational cost associated with training and evaluating MLIPs. This opens the door to computationally efficient generation of high-quality MLIPs for a range of applications which demand different accuracy versus training and evaluation cost trade-offs.

The selective conversion of CH4 and CO2 to liquid fuels is constrained in thermal catalysis by linear scaling relationships, overoxidation of oxygenate intermediates, and irreversible surface poisoning. Nonthermal plasma-enabled catalysis circumvents some of these limitations by introducing non-equilibrium excitation (e.g., vibrational, electronic, and dissociative) of molecular reactants. However, performance remains hindered by catalyst deactivation due to carbon and oxygen accumulation on active sites, driven by distinct mechanisms tied to metal-adsorbate interactions. Here, we show how continuous active-site regeneration during plasma-catalysis can be achieved by coupling plasma-driven dissociative chemisorption with targeted co-reactant activation. This strategy enables selective removal of surface poisons and is applicable across distinct catalyst regimes. On weakly binding catalysts like Cu, plasma-generated species remove adsorbed oxygen atoms, restore CH3O* intermediates and triple methanol selectivity. On strongly binding catalysts like Ni, accumulated CHx* fragments are selectively removed, doubling C2 hydrocarbon yields by reactivating C-C coupling. Unlike thermal systems, plasma excitation decouples regeneration from equilibrium constraints that correlate adsorption and desorption kinetics to a single gas temperature, enabling continuous, in situ removal of poisons without cycling and disrupting steady-state conversion. Using a combination of surface characterization methods and density functional theory calculations, we identify the mechanism by which plasma-co-reactant synergy tunes surface intermediates and redefines the design space of heterogeneous catalysis under mild conditions (Tsur = 473 K, Tvib = 4200 K, p = 1 atm).

To surmount the inherent limitations and fully harness the remarkable ultra-high specific capacity (2,596 mAh g −1 ) of phosphorus (P) anode for sodium-ion batteries (SIBs), we unveil an alternative fast and reversible electrochemical pathway based on Na 2 P 16 ↔Na 3 P, which transcends the barriers posed by sluggish reaction kinetics in solid-state red P. It entails the immobilization of dissolved sodium polyphosphide (Na 2 P 16 ) onto carbon cloth (CC) matrices via robust C─O─P bonding (Na 2 P 16 @CC), and the intrinsic superior malleability of Na 2 P 16 effectively mitigates the issue of electrode pulverization caused by volumetric changes of red P during (de)sodiation. Additionally, the profound chemical adsorption of surface oxygen-doped CC toward phosphorus species and the utilization of weakly solvating cyclic carbonate solvents synergistically inhibit the vexing dissolution of high-order polyphosphides in the electrolyte. By capitalizing on the advances of the novel reaction mechanism, the Na 2 P 16 @CC composite anode material achieves improved sodium storage performance with a high initial reversible capacity of 1.75 mAh cm −2 at 0.1 mA cm −2 and a capacity retention of 81% over 600 cycles. This work opens an avenue toward the rational design of P-based anodes for high-energy SIBs.

Layered Na x MO2 sodium oxide positive electrode materials have experienced renewed interest owing to the current commercial attention on sodium-ion batteries. Although there are many attractive qualities of these materials, they suffer from serious shortcomings owing to Na+ ordering and transition-metal layer gliding that cause a plethora of voltage plateaus during cycling. The P2-layered Na2+x Ni2-x/2TeO6 (0 ≤ x ≤ 0.5) system provides a framework for investigating the effect of dual Na+ substitution into the sodium layer and the transition-metal layer of the structure and its effects on the electrochemical properties of the materials. A careful investigation into the synthesis and properties of these materials reveals that the sodium content used during material preparation has a drastic effect on the composition and electrochemical profile of these materials. The sodium substitution disrupts ordering within the transition-metal layer, thereby disrupting Na+ ordering in the adjacent sodium layers. Beyond a critical sodium concentration, the layer stacking shifts, and all voltage plateaus of the P2-Na2Ni2TeO6 material are no longer observed at 4.4 V versus Na+/Na. These results also question the common belief that additional sodium precursor is required when preparing layered sodium oxide cathodes, providing new guidelines for material synthesis and characterization.

A new triaryl arsine (Ar3As)-based metal-organic framework (MOF) named AsCM-102 has been prepared by the reaction of As(C6H4-4-CO2H)3 with Co(BF4)2 and 4,4'-bipyridine. AsCM-102 contains pairs of staggered As donors that function as trans-chelators for the facile incorporation of organometallic RhI species via a single crystal-to-single crystal transformation. Coordination of RhI is achieved by soaking crystals in a solution of [Rh(CO)2Cl]2 at 70 °C. The originally closed and offset As2 pockets expand to facilitate the trans-As2 chelation of RhI. The resulting metalated MOF displays trans-[(Ar3As)2Rh(CO)Cln](1-n)+ complexes inside uniquely confined micropore reaction environments. Installation of the As-Rh-As moieties significantly enhances the internal porosity of the MOF. Crystalline RhI-AsCM-102 is an air-stable and recyclable hydroformylation catalyst, which is more active than its phosphine-based analogue. It is also selective toward the formation of iso-aldehydes over n-aldehydes with various C6-C8 olefin feedstocks. By leveraging the absolute atomic coordinates of RhI-AsCM-102 obtained from single-crystal X-ray diffraction analysis, density functional theory (DFT) explains the experimentally observed iso-favored hydroformylation regioselectivity due to pore confinement. RhI-AsCM-102 is resistant toward leaching of As into solution under forcing reaction conditions (40 atm of CO/H2, 70 °C). This work demonstrates the premise that incorporation of organo(arsines) into MOF scaffolds is a safer and more convenient strategy for their deployment in catalysis, by alleviating M-As bond lability and As toxicity issues, which prevents their widespread use in homogeneous catalysis.

Suppressing the interface deterioration and sodium dendrites growth is crucial for achieving long‐life polyethylene oxide (PEO)‐based all‐solid‐state sodium metal batteries. Herein, we systematically screen Sb 2 S 3 for use as a PEO‐based solid‐state electrolyte (PSE) additive through theoretical calculations, and in situ construct a highly stable solid electrolyte interphase (SEI) enriched with Na 2 S and Na 3 Sb. This SEI, characterized by its low reduction reaction activity, high ionic conductivity, and strong Na affinity, significantly inhibits interfacial side reactions, accelerates ion transport, and facilitates smooth Na + deposition. Moreover, the incorporation of Sb 2 S 3 effectively enhances the mechanical robustness, ionic transference number, and ionic conductivity of the composite solid‐state electrolyte film (Sb 2 S 3 @PSE), thereby mitigating the sodium dendrites formation. Consequently, remarkable electrochemical performances for the Sb 2 S 3 @PSE symmetric battery (achieving 5200 h at 0.1 mA cm −2 , 520 times longer than that of pristine PSE), and the Na 3 V 2 (PO 4 ) 3 |Sb 2 S 3 @PSE|Na full battery with high‐capacity retention of 91% after 1000 cycles, are demonstrated. This work, which emphasizes the in situ construction of a stable SEI, provides significant guidance to suppress interface degradation for long‐life solid‐state metal batteries.

Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. In this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid-liquid extraction mechanism of uranium using N,N-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO2(NO3)2(DEHiBA)2, agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.

Substituent group modifications can influence the conductivity of organic materials. However, achieving several orders of magnitude increases in conductivity through minor substituent changes remains a challenge. Here, we report the observation of such large changes in cocrystals of cyclo[8]pyrroles and polyiodides. Two cocrystals were prepared, one from all-ethyl-substituted cyclo[8]pyrrole (1•+), which forms a 2D stacked structure [(1•+)2⊃(I7)-•(I24)-], and the other from a methyl-ethyl-substituted analogue (2•+), which yields a 3D layered structure [2•+•(I16)-]. The methyl-ethyl-substituted cocrystal exhibited an approximately 1000-fold higher conductivity (6.1 × 10-1 S/cm) than its all-ethyl counterpart (4.2 × 10-4 S/cm). Cocrystal [2•+•(I16)-] demonstrated good stability, retaining the bulk of its conductivity after being exposed to air for four months or upon heating to 100 °C. The present findings highlight how substituent effects, a molecular feature readily amenable to modification, can have a profound effect on cocrystal conductivity. This work thus sets the stage for further optimization of high-performance organic conductors.

The precise control of mechanochemical activation within deep tissues using non-invasive ultrasound holds profound implications for advancing our understanding of fundamental biomedical sciences and revolutionizing disease treatments1-4. However, a theory-guided mechanoresponsive materials system with well-defined ultrasound activation has yet to be explored5,6. Here we present the concept of using porous hydrogen-bonded organic frameworks (HOFs) as toolkits for focused ultrasound (FUS) programmably triggered drug activation to control specific cellular events in the deep brain, through on-demand scission of the supramolecular interactions. A theoretical model is developed to potentially visualize the mechanochemical scission and ultrasound mechanics, providing valuable guidelines for the rational design of mechanoresponsive materials to achieve programmable control. To demonstrate the practicality of this approach, we encapsulate the designer drug clozapine N-oxide (CNO) into the optimal HOF nanocrystals for FUS-gated release to activate engineered G-protein-coupled receptors in the ventral tegmental area (VTA) of mice and rats and hence achieve targeted neural circuit modulation even at depth 9 mm with a latency of seconds. This work demonstrates the capability of ultrasound to precisely control molecular interactions and develops ultrasound-programmable HOFs to non-invasively and spatiotemporally control cellular events, thereby facilitating the establishment of precise molecular therapeutic possibilities.

Anode-free all solid-state batteries (AF-ASSBs) employ "empty" current collector with three active interfaces that determine electrochemical stability; lithium metal - Solid electrolyte (SE) interphase (SEI-1), lithium - current collector interface, and collector - SE interphase (SEI-2). Argyrodite Li6PS5Cl (LPSCl) solid electrolyte (SE) displays SEI-2 containing copper sulfides, formed even at open circuit. Bilayer of 140 nm magnesium/30 nm tungsten (Mg/W-Cu) controls the three interfaces and allows for state-of-the-art electrochemical performance in half-cells and fullcells. AF-ASSB with NMC811 cathode achieves 150 cycles with Coulombic efficiency (CE) above 99.8%. With high mass-loading cathode (8.6 mAh cm-2), AF-ASSB retains 86.5% capacity after 45 cycles at 0.2C. During electrodeposition of Li, gradient Li-Mg solid solution is formed, which reverses upon electrodissolution. This promotes conformal wetting/dewetting by Li and stabilizes SEI-1 by lowering thermodynamic driving force for SE reduction. Inert refractory W underlayer is required to prevent ongoing formation of SEI-2 that also drives electrochemical degradation. Inert Mo and Nb layers likewise protect Cu from corroding, while Li-alloying layers (Mg, Sn) are less effective due to ongoing volume changes and associated pulverization. Mechanistic explanation for observed Li segregation within alloying LixMg layer is provided through mesoscale modelling, considering opposing roles of diffusivity differences and interfacial stresses.

Mounting concerns regarding per-/poly-fluoroalkyl substances (PFAS) on human health are focusing attention on trace-level PFAS detection in aqueous environments. Here, we report a readily prepared small molecule, 2,6-bis(3,5-diethyl- 1H -pyrrol-2-yl)pyridine (receptor 1 ), that displays high binding affinities (log K a  = 4.9–6.2) and produces a strong “turn-on” emission response when exposed to representative PFAS in hexanes. The hydrophobic nature of 1 , and its strong affinity for various PFAS, allowed hexanes solutions of 1 to be used as “turn-on” emission sensors for dilute aqueous solutions of long-chain (≥C 8 ) PFAS under acidic conditions (pH 2) by liquid-phase extraction (LPE). In the case of perfluorooctanoic acid (PFOA), the response was rapid (under 10 min) and sensitive. Limits of detection (LOD) as low as 250 ppt were readily achievable by direct naked-eye observation. LOD as low as 40 and 100 ppt, respectively, could be reached for deionized and tap water solutions of PFOA using a smartphone color-scanning application. Little change in the sensitivity was seen in the presence of a range of inorganic and organic species that could act as potential interferants. Support for the present findings came from UV–vis absorbance, fluorescence, 1 H/ 19 F NMR spectroscopic analyses, density functional theory calculations, and single-crystal X-ray diffraction analyses.

We employed accumulative roll bonding to fabricate self-standing metallurgical composite of in situ formed alkaline potassium-bismuth-telluride intermetallic K2(Bi2/6Te3/6Vac1/6) embedded in potassium metal. This newly discovered thermodynamically stable potassiophilic crystal, termed "KBT", is fcc antifluorite with K2Te archetype. Symmetric cells achieve 880 h of cycling at 0.5 mA cm-2 and 0.5 mAh cm-2. Potassium metal battery (KMB) with Prussian blue (PB) cathode in carbonate electrolyte retains 80% capacity after 200 cycles at 1C. In ether-based electrolyte with organic cathode, it achieves 80% retention after 900 cycles at 2C. Combined synchrotron X-ray nano-tomography, cryogenic-focused ion beam microscopy (Cryo-FIB) and sputter-down X-ray photoelectron spectroscopy (XPS) demonstrate uniform electrodeposits, versus baseline of potassium filaments intermixed with pores and coarse SEI. Binary K3Bi-K and K2Te-K intermetallic supports also provide improved electrochemical performance, albeit to lesser extent. Multiscale simulation provides insight into role of support structure in adatom energetics, film nucleation, early-stage SEI morphology and interfacial stability.

Molecular dynamics simulations demand an unprecedented combination of accuracy and scalability to tackle grand challenges in catalysis and materials design. To bridge this gap, we present AlphaNet, a local-frame-based equivariant model that simultaneously improves computational efficiency and predictive precision for interatomic interactions. By constructing equivariant local frames with learnable geometric transitions, AlphaNet encodes atomic environments with enhanced representational capacity, achieving state-of-the-art accuracy in energy and force predictions. Extensive benchmarks on large-scale datasets spanning molecular reactions, crystal stability, and surface catalysis (Matbench Discovery and OC2M) demonstrate its superior performance over existing neural network interatomic potentials while ensuring scalability across diverse system sizes with varying types of interatomic interactions. The synergy of accuracy, efficiency, and transferability positions AlphaNet as a transformative tool for modeling multiscale phenomena, decoding dynamics in catalysis and functional interfaces, with direct implications for accelerating the discovery of complex molecular systems and functional materials.

Catalyst deactivation by surface bound intermediates is a persistent challenge in plasma catalysis, often limiting conversion, product selectivity, and energy efficiency in chemical processes. In this study, we identify surface oxygen accumulation (O*) as the dominant deactivation mechanism in plasma assisted nitrous oxide (N2O) decomposition and present a direct strategy to mitigate it. Using polycrystalline Cu/Al2O3 and Al2O3 catalysts under nonthermal plasma conditions, we show that Cu deactivates rapidly due to O* poisoning, as confirmed by time-resolved experiments and X-ray photoelectron spectroscopy (XPS). To counteract this, we introduce hydrogen based co-reactants (H2 and CH4) that remove surface O* by forming H2O and CO2, thereby regenerating active sites and restoring catalytic activity. This co-reactant approach enables efficient N2O decomposition at ambient temperature and pressure, conditions where traditional thermal catalysis typically requires elevated temperatures (>= 300 degrees C) and expensive noble metals. Co-reactant addition not only enhances conversion and halves energy cost, but also, when using CH4, produces value added C2 hydrocarbons. Density functional theory calculations support these findings, revealing that co-reactants open up more favorable reaction pathways for O* removal. These results highlight how plasma catalysis creates a multidimensional design space where plasma conditions, catalyst surface chemistry, and gas phase composition can be co-optimized. This flexibility enables the use of weakly binding, earth abundant catalysts like Cu, which are otherwise inactive under mild thermal conditions. These findings demonstrate how plasma assistance can augment heterogeneous catalysis, offering alternative avenues for efficient, low temperature transformations in both environmental remediation and synthetic applications.

New anodic electrocatalysts with high performance and cost-effectiveness at large current densities help advance the emerging anion exchange membrane water electrolyzer (AEMWE) technology. To this end, a ruthenium (Ru) single atoms and sulfur (S) anions dual-doped NiFe layered double hydroxides (Ru-S-NiFe LDH) catalyst is reported with remarkably low alkaline oxygen evolution reaction (OER) overpotentials, high mass activities and prolonged stabilities at high current densities. Inspiringly, the AEMWE performance on Ru-S-NiFe LDH is also superior to the NiFe LDH. In-depth mechanism investigations reveal that Ru single atoms not only act as the highly active sites, but also facilitate the conductivity of NiFe LDH. Meanwhile, S anions accelerate the electrochemical reconstruction of NiFe LDH to OER-active NiFeOOH and alleviate the over-oxidation issue on Ru active sites. Benefiting from these, Ru-S-NiFe LDH shows significantly enhanced OER activity and stability. Theoretical calculations further validate the decreased OER free energy difference brought about by the Ru single atoms and S anions dual-doping. This study offers a proof-of-concept that the noble metal single atoms and anions dual-doping is a feasible strategy to construct the promising 3d transition metal-based electrocatalysts toward the practical alkaline water electrolyzer.

Metal nanocrystals (NCs) show utility in a variety of applications due to their unique structure-dependent properties. Isolating these structure-property relationships is crucial for NC design, but heterogeneities present in NC ensembles as well as limitations in NC characterization strategies complicate this goal. Herein, we describe the various types of intraparticle and interparticle heterogeneities common to NC ensembles and then provide a detailed description and comparison of single-particle techniques that can be used to characterize these different heterogeneities. Case studies then showcase the use of multimodal characterization approaches, where multiple, primarily single-NC techniques are used in tandem to provide new insights into metal NC structure-property relationships. We conclude with a critique of single-NC techniques that motivates the development of new high-throughput and high-resolution single-NC characterization approaches as well as computational tools, with a proposed workflow outlined to accelerate NC design and discovery.

High-valent iron-oxo species (FeIV═O) is a fascinating enzymatic agent with excellent anti-interference abilities in various oxidation processes. However, selective and high-yield production of FeIV═O remains challenging. Herein, Fe diatomic pairs are rationally fabricated with an assisted S bridge to tune their neighbor distances and increase their loading to 11.8 wt.%. This geometry regulated the d-band center of Fe atoms, favoring their bonding with the terminal and hydroxyl O sites of peroxymonosulfate (PMS) via heterolytic cleavage of O─O, improving the PMS utilization (70%), and selective generation of FeIV═O (>90%) at a high yield (63% of PMS) offers competitive performance against state-of-the-art catalysts. These continuous reactions in a fabricated device and technol-economic assessment further verified the catalyst with impressive long-term activity and scale-up potential for sustainable water treatment. Altogether, this heteroatom-bridge strategy of diatomic pairs constitutes a promising platform for selective and efficient synthesis of high-valent metal-oxo species.

Machine-learning algorithms have been proposed to capture electrostatic interactions by using effective partial charges. These algorithms often rely on a pretrained model for partial charge prediction using density functional theory-calculated partial charges as references, which introduces complexity to the force field model. The accuracy of the trained model also depends on the reliability of charge partition methods, which can be dependent on the specific system and methodology employed. In this study, we propose an atom-centered neural network (ANN) algorithm that eliminates the need for reference charges. Our algorithm requires only a single NN model for each element to obtain both atomic energy and charges. These atomic charges are then employed to compute electrostatic energies using the Ewald summation algorithm. Subsequently, the force field model is trained on total energy and forces, with the inclusion of electrostatic energy. To evaluate the performance of our algorithm, we conducted tests on three benchmark systems, including a Ge slab with an O adatom system, a TiO2 crystalline system, and a Pd-O nanoparticle system. Our results demonstrate reasonably accurate predictions of partial charges and electrostatic interactions. This algorithm provides a self-consistent charge prediction strategy and possibilities for robust and reliable modeling of electrostatic interactions in machine-learning potentials.

Due to their unique advantages, single atoms and clusters of transition metals are expected to achieve a breakthrough in catalytic activity, but large-scale production of active materials remains a challenge. In this work, a simple solvent-free one-step annealing method is developed and applied to construct diatomic and cluster active sites in activated carbon by utilizing the strong anchoring ability of phenanthroline to metal ions, which can be scaled for mass productions. Benefiting from the synergy between the different metals, the obtained sub-nano-bimetallic atom-cluster catalysts (FeNiAC -NC) exhibit high oxygen reduction reactions (ORR) activity (E1/2 = 0.936 V vs. RHE) and a small ORR/oxygen evolution reaction (OER) potential gap of only 0.594 V. An in-house pouch Zn-air battery is assembled using an FeNiAC -NC catalyst, which demonstrates a stability of 1000 h, outperforming previous reports. The existence of clusters and their effects on catalytic activity is analyzed by density functional theory calculations to reveal the chemistry of nano-bimetallic atom-cluster catalysts.