Carbon molecular sieves (CMSs) offer a promising alternative to energy-intensive cryogenic distillation for gas separations such as C3H6/C3H8. However, they suffer from physical aging that reduces gas permeability and from a trade-off between size-sieving capability and mechanical stability. Herein, we address both challenges by engineering robust nanoporous carbons via an integrated low-temperature carbonization and vapor-phase infiltration (VPI) strategy. A 6FDA-DAM polyimide is carbonized at 500 degrees C to form mechanically resilient carbons and then subjected to VPI treatment to generate sub-nanometer aluminum oxyfluoride onto internal pore walls. The inorganic coating rigidifies carbon nanostructures, enhancing their resistance to physical aging, and narrows pore sizes, thereby increasing size-sieving ability. 6-Cycle VPI treatment of a CMS increases C3H6/C3H8 selectivity from 16 to 43 with stable separation properties over 90 days, outperforming most leading CMSs and polymers. The VPI process enables atom-by-atom tuning of micropores and provides a scalable route to enhance membrane separation efficiency.

Organic semiconducting materials generally have large exciton binding energy due to low dielectric constants, necessitating a donor:acceptor interfacial charge transfer state (CTS) for effective exciton dissociation. Precisely regulating the energy levels of donors and acceptors can effectively modulate CTS energetics. In this work, we develop a strategy for decoupled regulation of molecular energy levels by enhancing short-range intramolecular charge transfer (ICT) in non-fullerene acceptors (NFAs). We design two NFAs (B-2TPIC and T-2TPIC) incorporating strong electron-donating thieno[3,2-b]pyrrole and strong electron-accepting 3-(1,1-dicyanomethylene)-5,6-difluoro-1-indanone (IC2F) units. Density functional theory calculations reveal that pronounced electron density redistribution induces strong short-range ICT between thieno[3,2-b]pyrrole and IC2F units, leading to a considerable upshift of the lowest unoccupied molecular orbital (LUMO) level and small migration of the highest occupied molecular orbital (HOMO) level with decoupling HOMO/LUMO regulation. Both B-2TPIC and T-2TPIC show LUMO levels higher than -3.8 eV, HOMO levels deeper than -5.5 eV, and strong visible and near-infrared absorption. When these NFAs are blended with the polymer donor PM6 and used as photocatalysts for hydrogen evolution, the PM6:B-2TPIC nanoparticles exhibit faster hole transfer and weaker charge recombination, resulting in an average hydrogen evolution rate of 267.4 mmol g-1 h-1 at a low concentration of 6.67 & micro;g mL-1 for 4 h under AM 1.5G simulated sunlight (100 mW cm-2), higher than that of the PM6:T-2TPIC counterpart (165.1 mmol g-1 h-1).

Electronic coherence underpins the control of quantum character in molecular systems; however, such control is fundamentally constrained by ultrafast decoherence arising from significant vibronic relaxation, which critically shapes carrier transport and functionality in optoelectronic materials. Here, we employ the two-dimensional electronic spectroscopy (2DES) based on the coherent loop-based integrated modulating and beam splitting system to investigate intersystem crossing dynamics in a molecular thin film of Pt^II complex aggregates (4H) at room temperature. Our results reveal that the electronic coherence transfer from singlet to triplet excited states and persists for up to 680 fs with the decay lifetime of 228 fs, substantially exceeding typical electronic dephasing times. These findings demonstrate the existence of long-lived electronic coherence in a solid-state molecular system, offering new prospects for coherent control in optoelectronic devices and emerging quantum technologies.

Graph neural networks (GNNs) can effectively model structural information of graphs, making them widely used in knowledge graph (KG) reasoning. However, existing studies on the expressive power of GNNs mainly focuses on simple single-relation graphs, and there is still insufficient discussion on the power of GNN to express logical rules in KGs. How to enhance the logical expressive power of GNNs is still a key issue. Motivated by this, we propose Path-Neighbor enhanced GNN (PN-GNN), a method to enhance the logical expressive power of GNN by aggregating node-neighbor embeddings on the reasoning path. First, we analyze the logical expressive power of existing GNN-based methods and point out the shortcomings of the expressive power of these methods. Then, we theoretically investigate the logical expressive power of PN-GNN, showing that it not only has strictly stronger expressive power than C-GNN but also that its (k+1)-hop logical expressiveness is strictly superior to that of k-hop. Finally, we evaluate the logical expressive power of PN-GNN on six synthetic datasets and two real-world datasets. Both theoretical analysis and extensive experiments confirm that PN-GNN enhances the expressive power of logical rules without compromising generalization, as evidenced by its competitive performance in KG reasoning tasks.

Knowledge Graph Embedding (KGE) maps entities and relations into continuous vector spaces to facilitate link prediction tasks. Given the inherent inability of knowledge graphs to directly supply high-quality negative samples with multi-level difficulty, existing methods typically rely on post-sampling assessment strategies, which lack controllable generation of difficulty-calibrated negatives tailored to diverse KGE training requirements. To address these challenges, we propose ConDNS, a novel conditional diffusion-based negative sampling method for knowledge graph embedding. By adjusting the diffusion timestep, our model achieves dynamic difficulty modulation of synthetic negatives through global entity-relation information utilization. This enables generation of semantically valid samples that synergistically integrate with conventional samples, thereby overcoming single-strategy sampling bottlenecks and establishing a multiscale difficulty configuration. Experiments demonstrate that ConDNS achieves state-of-the-art performance across multiple benchmarks with minimal synthetic samples while functioning as a plug-and-play module compatible with mainstream KGE architectures. Source code is available at: https://github.com/zrj-wang/ConDNS.

Perovskite light emitting diodes (PeLEDs) are highly promising for next-generation displays owing to their exceptional emissive properties. However, their complex fabrication, often relying on interleaved solution processing and vacuum deposition, hinders scalable mass production. Here, we present a strategy for high-performance, all-vacuum-deposited PeLEDs. By incorporating vacuum-sublimed nitrogen-containing additives, guanidinium bromide, we effectively suppress intrinsic bulk defects in perovskite films and mitigate nonradiative recombination. Furthermore, an ultrathin, vacuum-deposited alkali metal halide top layer is introduced to heal the defective interface by compensating for the loss of bromide during thermal evaporation and improving crystallinity. Integrating these advanced perovskite films with vacuum-sublimed organic transporting layers resulted in PeLEDs demonstrating a maximum luminance exceeding 164,000 cd/m2 and external quantum efficiencies (EQEs) of 14.83%. Our EQE approaches the current vacuum-device record, while the luminance represents a more than 5-fold improvement over most of the previously reported all-vacuum-deposited devices, highlighting a significant leap toward industrially viable PeLED displays.

We present a two-dimensional electronic spectroscopy (2DES) platform driven by a novel Coherent Loop-based Integrated Modulating and Beamsplitting System (CLIMBS). Coupled with an octave-spanning multiple-plate continuum (MPC) source, CLIMBS enables broadband, phase-coherent measurements with attosecond-level time delay precision. Its Sagnac-inspired, nearly common-path geometry provides exceptional long-term phase stability without active feedback, eliminating beam walk-off and preserving beam pointing during delay scans. Delay calibration using spectrally resolved interferometric fringes yielded a wedge angle in excellent agreement with the designed geometry, confirming precise, linear coherence time control. The MPC technique generates broadband excitation pulses spanning 550–980 nm and temporally compressed to 3.7 fs. This bright, few-cycle source enables simultaneous interrogation of widely separated electronic and vibronic transitions, with high temporal and spectral resolution, allowing 2DES to capture vibronic cross peaks, energy-transfer pathways, and undistorted ground-state bleaching (GB), stimulated emission (SE), and excited-state absorption (ESA) features across a broad spectral window. System performance was benchmarked on chlorophyll-a in methanol, where the excitation bandwidth fully covers the Q_x and Q_y bands, ensuring distortion-free spectra. The nearly collinear configuration of CLIMBS eliminates beam walk-off during delay scanning, supports ultrabroadband few-cycle 2DES enabled by the high-brightness MPC source, and maintains attosecond-level phase stability, providing a simple and robust platform for high-fidelity multidimensional spectroscopy.

Hole-conductor-free, printable mesoscopic perovskite solar cells with carbon electrodes present a viable strategy for industrial manufacturing of photovoltaics with low cost. However, depositing perovskite absorber within their intricate triple-layer (TiO2/ZrO2/carbon) mesoporous scaffold makes it challenging to regulate the crystallization process, leading to limited crystal quality and high defect density of perovskite, resulting in performance loss. Moreover, the adopted TiO2 electron transport layer (ETL) with insufficient electrical properties and high surface defect density restricts the carrier injection at the ETL/perovskite interface and exacerbates recombination loss. Herein, two sulfonamide additives, including sulfanilamide (SA) and sulfaguanidine (SG), are introduced into the perovskite precursor. Both additives present strong interaction capability with the key components, such as Pb2+ and formamidinium (FA+) of the halide perovskite, thus regulating the crystallization processes. Meanwhile, they passivate surface defects and enhance the electrical property of TiO2, thus promoting carrier injection at the ETL/perovskite interface. With their concurrent modulation in crystallization and interface, SA and SG improve the device power conversion efficiencies to 19.84% and 21.50% from 18.02%. Meanwhile, the better-performing SG device retains 90% of its initial PCE after 530 h of maximum power point tracking at 55 degrees C +/- 5 degrees C under 1-sun illumination.

Carbon-based, fully printable hole-transport-layer-free mesoscopic perovskite solar cells (p-MPSCs) offer a low-cost, scalable photovoltaic technology, yet suffer from efficiency losses due to non-radiative recombination at electron-selective interface. Here, we design geometrically asymmetric zwitterions, featuring a bulky charge-dispersed triphenylphosphonium cation and a small charge-concentrated sulfonate anion linked by an alkyl chain, to weaken intermolecular and intramolecular charge attraction, thereby constructing a permanent dipole interlayer that mitigates such recombination. By extending the alkyl spacer and introducing methyl substituents on triphenylphosphonium, the optimized zwitterion, 4-(tri-p-tolylphosphonio)butane-1-sulfonate (4MePS), achieves an ultrahigh dipole moment of 14.92 Debye. 4MePS strongly interacts with both perovskite and TiO2 to form surface dipoles, drastically lowering their work functions by 0.42 and 0.54 eV, respectively. Comprehensive characterizations confirm that 4MePS suppresses non-radiative recombination and accelerates charge extraction in p-MPSCs. Consequently, 4MePS-treated p-MPSCs deliver a champion power conversion efficiency (PCE) of 23.3% (vs. 21.8% for control) and a minimodule efficiency of 20.2% over 57.3 cm2, among the highest reported for p-MPSCs. Encapsulated devices retain 90% of their initial PCE after 1200 hours of maximum power point tracking under 1‑sun illumination at 55 ± 5 °C. This work establishes a charge density-asymmetric molecular design strategy for engineering interfacial dipoles toward high-performance perovskite devices.

A systematic screening and evaluation framework integrating COSMO-RS prediction, quantum chemical (QC) calculations, molecular dynamics (MD) simulations, and process simulation was developed to identify efficient ionic liquids (ILs) extractants for the separation of the heptane (HEP)/ethyl acetate (EA) azeotrope via liquid–liquid extraction. COSMO-RS was employed to screen candidate ILs based on selectivity and solvent power, revealing a trade-off between separation purity and extraction capacity. To balance these competing factors, a mixed ionic liquid (MIL) composed of [EMIM][BF4] and [EMIM][TF2N] was proposed, and the optimal molar ratio (3:1) was identified. The extraction performance of the MIL was further validated through liquid–liquid equilibrium experiments and correlated using the UNIFAC-Lei model. QC calculations and MD simulations were then conducted to elucidate the microscopic separation mechanism, demonstrating that electrostatic interactions dominate EA–IL interactions, with cations contributing more significantly than anions, while dispersion interactions enhance solvent capacity. Process simulations combined with heat integration confirmed that the MIL-based process exhibits superior economic, energy, and environmental performance compared with single-IL systems. The strong consistency among molecular simulations, experimental data, and process evaluation demonstrates that the proposed framework is an effective tool for designing IL-based extraction systems for azeotropic separations.

Multipodal self-assembled monolayers (SAMs) composed of π-conjugated aromatic units have emerged as highly promising hole-transporting layers in perovskite solar cells (PSCs) due to their excellent interfacial anchoring and scalable fabrication. However, it remains a formidable challenge to develop high-performance tetrapodal SAMs that can simultaneously achieve ideal surface coverage, enhanced device efficiency, and improved long-term stability. Herein, we rationally designed and synthesized two novel pyrene-centered tetrapodal SAMs, 4PACz-Py-C2 and 4PACz-Py-C4. Spacer engineering has been employed to regulate molecular packing, facilitate hole extraction, and suppress non-radiative recombination. The optimized PSCs based on 4PACz-Py-C2 achieved a power conversion efficiency (PCE) of 26.61% and maintained 90.1% of their initial PCE after 2000 h of maximum power point (MPP) tracking, demonstrating outstanding commercialization potential. Moreover, perovskite solar modules (aperture area: 21 cm2) and 1.68 eV wide-bandgap perovskite devices based on 4PACz-Py-C2 yielded high PCEs of 23.52% and 23.48%, respectively. This work exhibits significant application value in the development of novel multipodal SAMs for enhancing the efficiency and stability of PSCs.

Recommendation actively selects information for users, yet it persistently face data sparsity and cold-start problems. The incorporation of knowledge graph as side information has demonstrated effectiveness in mitigating these issues, leading to the development of knowledge-aware recommendation. Existing methods often use graph augmentation by constructing other knowledge views from the original knowledge graph (KG) to address external noise (e.g., erroneous triplets) and information overload (e.g., redundant data), yet they seldom consider the knowledge incompleteness (e.g., inherent missing facts and long-tail relation sparsity), which leads to decreased recommendation performance. Unlike noise (which introduces distortions) and information overload (which causes selection inefficiency), knowledge incompleteness stems from structural gaps in the graph that hinder semantic connectivity. To address this challenge, we propose RKGRec, a relation-guided conditional diffusion framework that generates a relation-centric auxiliary KG to alleviate knowledge incompleteness for recommendation. The model primarily consists of three core modules tailored for knowledge-aware recommendation: (1) a relation-attention network that captures multi-hop entity-relation patterns to obtain knowledge embeddings. (2) a relation-guided conditional diffusion model that strategically refines knowledge graph through controlled noise injection by the forward process and relation-guided denoising by the reverse process. (3) a joint prediction and optimization module that jointly trains recommendation and knowledge graph generation. Experimental results show RKGRec outperforms baselines across multiple datasets, particularly achieving both comprehensive leading predictive accuracy and competitive predictive diversity. The model also demonstrates robust performance in cold-start users, long-tail items, interaction noise, and knowledge graph noise or sparsity conditions.

Carbon molecular sieves (CMSs) contain ultramicropores with strong molecular sieving ability, and their pore sizes and distribution are usually determined by the structures of polymer precursors. However, it remains elusive to design polymer architectures that achieve the desired ultramicropores and thus high O2/N2 selectivity. Herein, we demonstrate that a highly permeable polyimide (6FDA-DAM) can be cross-linked by thermolabile phosphoric acid (PA) prior to carbonization at 700 °C or below to achieve the desired ultramicropores. PA forms hydrogen bonds with 6FDA-DAM and can be degraded at lower temperatures to control free volume. For example, doping 6FDA-DAM with PA at a doping level of 0.25 followed by carbonization at 700 °C enhances O2 permeability from 130 to 300 Barrer and O2/N2 selectivity from 3.4 to 11, surpassing Robeson’s upper bound. Our approach provides an efficient and versatile method to engineer ultramicropores in CMSs to enhance molecular sieving ability while retaining gas permeability and mechanical properties.

High quality light-absorbing layers with matched band gaps of sub-cells are crucial for achieving high power conversion efficiency (PCE) in the perovskite/organic tandem solar cells (PO-TSCs). In this work, we systematically optimized the band gaps of wide-bandgap perovskite layer (1.73-1.85 eV) and narrow-bandgap organic active layer (1.34-1.38 eV). An asymmetric small-molecule acceptor, namely SY2 (two F atoms and two Cl atoms), is introduced into the PM6:BTP-eC9 blend to enhance the light absorption, form the fibril network morphology, and facilitate exciton dissociation and transport. By precisely integrating a 1.80 eV perovskite sub-cell and a 1.34 eV ternary organic sub-cell, we achieved a champion PCE of 25.47% for the PO-TSCs based on the well-matched short-circuit current density. Besides, the optimized devices exhibited outstanding long-term stability. Our findings highlight the importance of band gap matching between front and near sub-cells in reducing recombination loss for high-performance tandem solar cells.

Hole transport layer-free printable mesoscopic perovskite solar cells (p-MPSCs) employing carbon electrodes offer cost-effective fabrication but face efficiency limitations due to suboptimal charge transport in the TiO2-based mesoporous electron transport layer (mp-ETL). Here, we develop a TiO2@SnO2 bilayer mp-ETL for p-MPSCs and obtain encouraging performance enhancement. By performing tailored chemical bath deposition of the preformed triple mesoporous scaffold of TiO2 ETL/ZrO2 spacer/carbon electrode rather than the mp-TiO2 alone, the conformal SnO2 coating is formed without experiencing high-temperature annealing suffering, thus circumventing associated electronic property degradation. This approach enables selective conformal SnO2 deposition exclusively on mp-TiO2, preventing the formation of undesired current leakage pathway in the spacer. Notably, intentional SnO2 incorporation in the carbon electrode shows no detrimental effects. The conformal SnO2 coating successfully improves interfacial energy alignment, suppresses non-radiative recombination, and boosts electron transport. The resulting TiO2@SnO2 p-MPSCs achieve a well improved champion power conversion efficiency (PCE) of 22.5

The development of cost-effective and highly-efficient photoelectrodes (PE) is critical for advancing solar driven photoelectrochemical (PEC) hydrogen (H2) production technologies. For enhancing PEC performance, it appears promising to combine plasmonic nanoparticles with active catalysts. Here we report a facile and powerful PE comprised of Au nanoparticles (NP) decorated on NbS2 nanoflakes (NF), directly grown on the carbon fiber paper. In this heterostructure, Au NP acts as plasmonic antennas that intensify local electromagnetic fields strengthening light matter interactions, NbS2 NF functions as a highly conductive semimetal scaffold offering abundant active HER sites, and the junction of Au NP/NbS2 NF enables efficient charge separation and fast electron transfer from NbS2 to Au, thereby synergistically accelerating HER kinetics under illumination. Consequently, this Au NP/NbS2 NF heterostructure, for the first time, demonstrates remarkable PEC performance under illumination, such as a 6-fold increase in the exchange current density, a 1.75-fold reduction in overpotential, and a 2.45-fold yield in the H2 production, respectively. Based on the heterostructure of Au NP/NbS2 NF, this work establishes a strategy of constructing a cost-effective and highly-efficient PE and providing valuable insights into the rational design of next generation PE for the sustainable hydrogen production.

Ultraviolet (UV) light-induced degradation at the buried interface poses a significant challenge to the long-term stability of perovskite solar cells (PSCs), resulting in substantial efficiency losses and hindering their commercialization. Here, we developed two simple deuterated self-assembled monolayers (SAMs), 2DPh-4PACz and 1DPh-4PACz. In particular, 2DPh-4PACz, featuring double deuterophenyl groups as pi-conjugated extension units, demonstrates enhanced intrinsic UV stability, improved hole-extraction capability, and effective protection of the perovskite film against UV exposure, while simultaneously improving film quality. As a result, PSCs incorporating 2DPh-4PACz achieved a power conversion efficiency (PCE) of 26.34% (certified 25.9%) and retained 96.9% of their initial PCE after 240 h of continuous UV irradiation, representing the best UV light stability reported to date. Additionally, extensive UV-aging experiments were conducted comparing with Ph-4PACz, confirming deuteration of SAMs as an effective strategy to improve UV resistivity. Moreover, these devices maintained 92.8% of their initial PCE after over 900 h of thermal aging at 85 degrees C, and 73% after more than 1380 h at 80% relative humidity (RH). This deuterophenyl groups design strategy with pi-conjugated extension offers a promising molecular design route for next-generation SAMs in high-performance, durable PSCs.image

Predicting future anomalies in Industrial Control System (ICS) sensor networks is critical for preventing equipment failures, minimizing downtime, and ensuring operational safety. Temporal Knowledge Graphs (TKGs) offer a structured framework to model evolving causal relationships among sensors, actuators, and processes within these infrastructures. However, state-of-the-art subgraph-based TKG extrapolation methods are susceptible to spurious correlations introduced by noisy data and lack the ability to capture essential multi-hop relational semantics required to model complex fault propagation chains in ICS networks. To address this limitation, we propose CARE (Causality-Aware Recursive Encoding), a framework for robust TKG reasoning in ICS settings. CARE integrates causal subgraphs to isolate structurally essential dependencies for sensor anomaly prediction, while filtering non-causal features such as temporary sensor correlations without genuine links. A subgraph-aware recursive encoder, inspired by path aggregation algorithms, captures multi-hop path semantics through transitively aware message passing. This design enables the modeling of long-range dependencies in fault propagation, where primary sensor anomalies propagate through interconnected control loops. Applied to ICS sensor anomaly prediction, CARE outperforms state-of-the-art methods in evaluations on real-world sensor data. Moreover, by providing traceable causal paths, CARE supports interpretability for proactive root cause diagnosis, achieving a 3.44% MRR improvement on ICEWS18 and superior performance on ICS-specific datasets.

The integration of crystallographic control into solution-processed perovskite films remains a challenge for efficient light emission, as disordered optical dipoles fundamentally limit photon extraction, a bottleneck constraining both classical and quantum planar optoelectronic devices. Here, we address this by developing an in situ formation strategy for oriented quasi-2D perovskite nanosheets within films via ligand-engineered crystallization. By designing and orchestrating steric hindrance and π-π interactions of ligands, we direct the crystallization kinetics to yield regular face-on nanosheets exhibiting enhanced horizontal transition dipole moment orientation compared to conventional isotropic films. The in situ architectural control also elevates both the photoluminescence quantum yield beyond 90% and carrier mobility comparable to 3D perovskite levels. These synergies enable perovskite light-emitting diodes (PeLEDs) with an external quantum efficiency (EQE) of 31.2% for pure-red emission at 635 nm, comparing favorably to other pure-red PeLEDs. Concurrently, the peak luminance and operational stability of the in situ nanosheet PeLEDs exhibit significant improvements.