ACS Publications最新文献

P2-type layered oxides, particularly those based on Mn/Fe/Ni, are promising cathode materials for sodium-ion batteries but have structural instability caused by phase transitions during sodiation/desodiation. We show that the phase behavior of the P2-type material Na_0.66Mn_2/3Fe_1/3–xNi_x series (x = 0, 0.02, 0.05, and 1/3) is strongly influenced by the Fe/Mn/Ni cationic ratio, which is directly coupled to the Mn³⁺/Mn⁴⁺ ratio. Powder X-ray diffraction reveals a systematic variation in the lattice parameter as a function of cation distribution, reflecting the corresponding changes in cation radii. Furthermore, the two binary systems exhibit different structural behaviors, with the Ni/Mn system forming a superstructure, whereas the Fe/Mn system does not exhibit such a phenomenon. The progressive narrowing of the linewidth determined by electron paramagnetic resonance indicates that the proportion of Mn³⁺ cations decreases as Ni²⁺ substitution in the P2 lattice increases. In situ XRD between 1.5 and 4.2 V vs Na/Na⁺ shows that, in the absence or at low amounts of Ni²⁺ (x < 0.05), the P2 phase coexists with an undesired, Jahn–Teller-distorted orthorhombic P′2 phase. Notably, no high-voltage structural transformations to O2 or OP4/Z phases are observed above 4.0 V. At x ≥ 0.05, the Mn³⁺ fraction is sufficiently reduced to suppress the formation of the P′2 phase in the low-voltage regime during discharge. These findings underscore the structural stabilization of Fe/Mn-based P2-type layered oxides achieved through minimal Ni substitution.

The supramolecular organization of aromatic bioactive molecules in aqueous media critically governs their physicochemical accessibility and biological function, yet strategies to redirect their native crystalline assemblies into alternative colloidal states remain limited. Here, we show that CPUL1, a phenazine-derived anticancer molecule, undergoes strong self-association in water to form extended, ordered sheet-like microstructures. Screening of structurally related aromatic coassemblers identified an oligoether-functionalized partner (Carrier 1) that uniquely induces a crystalline-to-colloidal transition-defined primarily by the loss of sharp long-range diffraction features and the formation of finite dispersed nanoparticles, rather than implying a uniquely defined amorphous molecular structure-whereas closely related analogues fail to prevent macroscopic aggregation. Combined spectroscopic, microscopic, diffraction, and simulation analyses are consistent with a model in which aromatic association perturbs the native packing of CPUL1 while solvophilic oligoether chains stabilize the resulting colloidal state. This supramolecular reorganization produces stronger intracellular CPUL1-associated fluorescence and enhanced anticancer activity relative to CPUL1 alone. Together, these results provide a proof-of-concept that structurally matched coassembly can access an alternative assembly state of an aggregation-prone aromatic bioactive molecule and modify its cellular performance. Because only the CPUL1/Carrier 1 pair has been examined and the closely related Carriers 2 and 3 were ineffective, the generality of this strategy for other aromatic bioactive molecules and co-assembling partners remains to be established.

Decellularized extracellular matrix (dECM)-based biomaterials have emerged as promising therapeutic platforms for spinal cord injury (SCI) repair due to their ability to provide regenerative biochemical cues. In this study, we developed an injectable, electroconductive, and self-healing hydrogel for potential application in SCI repair. Injectability enables minimally invasive delivery, electroconductivity facilitates electrical communication that may promote neural regeneration, and self-healing properties allow restoration of the hydrogel network following shear stress during injection. The self-healing hydrogel was formulated using Schiff-base chemistry between aldehyde-functionalized pectin (PEC-ALD) and amine-functionalized chitosan (CS). Poly(3,4-ethylenedioxythiophene) (PEDOT) derivatives were incorporated to enhance electrical conductivity and support neuronal differentiation and axonal outgrowth. In addition, decellularized human spinal cord (dSC), prepared using a modified Hudson protocol, was incorporated to provide tissue-specific pro-regenerative cues. The resulting PEC-ALD-CS-PEDOT-dSC hydrogels exhibited dynamic self-healing behavior characteristic of native spinal cord tissue. Under high oscillatory strain (1000%), the hydrogel network structure temporarily dissociated, while recovery of the crosslinked network occurred upon reduction to low strain (1%). The hydrogels demonstrated a storage modulus of approximately 600 Pa, consistent with the mechanical properties favorable for neuronal growth. Incorporation of PEDOT derivatives resulted in an electrical conductivity of approximately 5.6 mS/cm. Schiff base-mediated crosslinking enabled rapid gelation within 20 min at 37 °C. Furthermore, the hydrogels exhibited shear-thinning behavior, supporting their suitability for injectable delivery. Biological evaluation demonstrated immunomodulatory potential, as RAW 264.7 macrophages encapsulated within the hydrogel exhibited greater CD206 expression relative to CD86 expression, even under lipopolysaccharide-induced inflammatory conditions. Additionally, the hydrogels supported proliferation of human SH-SY5Y neuroblastoma cells, indicating cytocompatibility and potential to support neural regeneration. Overall, this study presents a multifunctional self-healing and electroconductive hydrogel platform incorporating decellularized spinal cord matrix for potential minimally invasive treatment of SCI.

Accurate prediction of RNA tertiary structure remains a central challenge in structural biology, yet existing computational approaches face fundamental limitations arising from the ruggedness of the RNA free-energy landscape and the scarcity of experimentally resolved structures. Here we present RNADaemon, a framework that formulates RNA tertiary structure prediction as a Markov Decision Process and trains a Soft Actor-Critic agent to apply adaptive external force-field biases that steer all-atom molecular dynamics toward native-like conformations to minimize the eRMSD metric. Trained on 479 monomeric RNA structures and evaluated on 14 test cases from CASP16, RNA-Puzzles, and the PDB, the agent trains stably, approaches SimRNA and FARFAR2 in median RMSD, and recovers the global fold of most reference structures. However, performance remains markedly inferior to that of AlphaFold 3 and trRosettaRNA. We identify the scalar Bellman error as a fundamental information bottleneck that currently limits the performance of the agent. To our knowledge, this is the first application of continuous-action deep reinforcement learning (RL) with all-atom force fields to RNA tertiary structure prediction. This study establishes a benchmark for future comparisons and provides diagnostic insights for RL-guided RNA structure modeling.

Rising global temperatures are driving demand for air conditioning, while leakage of high-GWP refrigerants further accelerates warming. In addition, the transition to low-GWP alternatives generates large volumes of waste refrigerants, which are currently treated by energy-intensive thermal destruction processes that often lead to the loss or downgrading of fluorine resources. Although dehydrofluorination offers a promising route to convert hydrofluorocarbon (HFC) refrigerants into low-GWP alternatives, conventional approaches are often limited by catalyst deactivation and degradation of target products caused by hydrogen fluoride (HF) as a by-product. Here, we show a practical continuous-flow process that overcomes these barriers by using alkali metal fluoride catalysts and silica as a sacrificial reagent to capture HF irreversibly as gaseous silicon tetrafluoride. This system enables the direct upcycling of legacy refrigerants into sustainable chemical feedstocks, including low-GWP hydrofluoroolefins (HFOs) and key monomers for fluoropolymers. Furthermore, selective conversion of the HFC-125 component within the near-azeotropic refrigerant blend R410A alters the boiling-point distribution, providing a potential route to HFC-32 recovery. Waste glass from end-of-life solar panels can also serve as the silica source under the reaction conditions.

This work reports the solvent-directed synthesis, structures,andchiroptical properties of enantiomeric Zn(II) metal–organiccages (MOCs) and coordination polymers (CPs) assembled from identicalsubcomponents under the same conditions. Varying the DMF/MeOH ratioselectively produces either a discrete cubic cage (1-R/S) or a chain polymer (2-R/S). Both materials exhibitblue emission, with the MOCs demonstrating higher solid-state quantumyields. MOCs 1-R/S in poly(methyl methacrylate) filmsexhibit clear circularly polarized luminescence (CPL) (|g_lum| ≈ 4.0 × 10^–4), whereasCPs 2-R/S display very weak CPL signals (|g_lum| ≈ 1.0 × 10^–5). Thiswork highlights that the supramolecular structure determines the CPLperformance of excited-state chirality.

Lewis structures constitute a foundational language inchemicaleducation, mechanistic reasoning, synthesis, and modern cheminformatics.With the accelerating shift toward data-driven and AI-assisted molecularscience, the ability to rapidly generate Lewis-like representations,including RDKit-parsable SMILES, has become essential for large-scaleanalysis and predictive modeling. While mature tools exist for organicmolecules, transition-metal complexes (TMCs) remain systematicallyunderserved owing to their inherently diverse coordination patterns.To address this gap, we develop TMC-LewisGen, an automated tool forgenerating Lewis-like structures of TMCs directly from Cartesian coordinates.The task is formulated as a discrete variable optimization problem,where a comprehensive set of explicit chemical rules defines the feasiblesolution space and integer linear programming (ILP) identifies a globallyoptimal assignment of bond orders, lone pairs, and formal charges,followed by postprocessing corrections. The final outputs includeLewis-like structures, RDKit-parsable SMILES, and interpretive bondingreports framed within the covalent bond classification (CBC) formalism,in which ligands are categorized as L-, X-, or Z- type donors. Benchmarkingon nearly 29,000 TMCs demonstrates robust performance, with validSMILES generated for 100% of the structures and subsecond ILP solutiontimes in most cases. Representative examples, including Werner-typecomplexes, η-coordinated π-ligands, and metal–carbenemotifs, demonstrate broad applicability. The framework supports large-scalestructural annotation, automated chemical interpretation, and data-drivenmodeling of TMCs. A freely accessible web server for 3D TMC analysisand visualization is available at https://tmc-lewis-structure.streamlit.app/.

Hydrogen-assisted degradation of metals is governed by the local hydrogen population available at near-surface lattice sites and defects. However, atomistic mechanical simulations commonly employ hydrogen concentrations that are not representative of realistic exposure conditions. In this study, ReaxFF-based reactive molecular dynamics is used to translate sustained CH₄ and CH₄ + H₂O exposure on a Ni(211) surface under an accelerated closed-cell protocol at 1000–1350 K into exposure-conditioned hydrogen availability descriptors. Three independent trajectories are performed for each exposure-temperature condition to quantify the normalized Ni-associated H inventory, hydrogen partitioning between adsorbed and subsurface regions, and localization across CN-defined step and terrace regions. A key finding is a dry–wet inversion in hydrogen accessibility. Although CH₄ + H₂O exposure produces a larger normalized Ni-associated H inventory (θ_H = 0.67–0.77), 71.6–86.1% of that inventory is primarily O-bound, while less than 3.5% of the geometrically defined near-surface H population lies within the subsurface window. In contrast, CH₄-only exposure yields a smaller normalized Ni-associated H inventory (θ_H = 0.088–0.447) but substantially greater subsurface partitioning, reaching 12.3% at 1350 K. These results show that the surface with the largest normalized Ni-associated H inventory does not necessarily have the largest geometric subsurface fraction. Furthermore, CH₄ + H₂O suppresses carbon accumulation, sustains an oxygen-containing surface state, and maintains higher per-site hydrogen localization around CN-defined step regions. Additional Ni(111) and laterally doubled Ni(211) checks retain the wet-over-dry H areal-density ordering and the complementary dry-over-wet subsurface-fraction ordering at 1250 K. The resulting protocol-specific descriptors provide candidate inputs for controlled downstream simulations of hydrogen transport, trapping, and degradation.

Due to its unique electronic structure and suitable active sites, Cu is a promising catalyst for CO₂ reduction. In this study, a series of Cu_x-Ga_0.9-In_0.1 oxides was synthesized via co-precipitation by varying Cu²⁺ concentration. X-ray diffraction and X-ray photoelectron spectroscopy reveal that Cu exists as CuO, Ga forms Ga₂O₃ due to its strong oxophilicity, and a small amount of In is uniformly dispersed as In₂O₃ within the CuO lattice, forming a composite catalyst. Combined density functional theory calculations and X-ray absorption fine structure spectroscopic analyses indicate Cu as the primary active site. The introduction of Ga and In enhances electron transfer with Cu, strengthening intermediate adsorption and accelerating product conversion, thereby improving catalytic performance for the CO₂ reduction reaction. When the molar ratio of divalent to trivalent ions n(M²⁺):n(M³⁺) is 5:1, the catalyst achieves a total current density of 28.4 mA·cm^–2, a CO Faradaic efficiency of 98.54%, and good stability.

The acid dissociation constants (pK_a) and the protonation states that they determine govern solubility,permeability, and protein–ligand binding, making their accurateprediction essential in drug discovery. We present AcepK_a, an application on the PlayMolecule AI platform thatimplements the Uni-pK_a framework, whichcouples statistical mechanics with representation learning. Ratherthan treating pK_a as a scalar regressiontarget, AcepK_a models the complete protonationensemble, enforcing thermodynamic consistency across coupled ionizationsites. The application is built on an independently retrained Uni-Molbackbone that matches state-of-the-art accuracy on standard publicbenchmarks. We further describe three engineering contributions: AceConfgen,a GPU-accelerated conformer generator approximately 7 times fasterthan other GPU implementations and more than an order of magnitudefaster than multithreaded RDKit; a streamlined inference engine thatprotonates molecules directly; and a 3D-aware mode that applies predictedprotonation states to bound ligand poses. AcepK_a supports library-scale prediction and provides a validated,ready-to-use implementation of this methodology, available at https://open.playmolecule.org.