Xiaolin Zheng

While high-entropy perovskite oxides have recently emerged as promising hosts for exsolution-enabled catalysts and electrodes, a systematic understanding of how high-entropy compositions influence exsolution remains limited. Here, we compare Ni exsolution in a simpler perovskite oxide, (La 0.6 Sr 0.4 ) 0.95 (Co 0.19 Fe 0.76 Ni 0.05 )O 3−δ (LSCF-5Ni), and two high-entropy perovskite oxides, (La 0.2 Sr 0.2 Ca 0.2 Nd 0.2 Y 0.2 ) 0.95 (Co 0.19 Fe 0.76 Ni 0.05 )O 3−δ (CaNdY-5Ni) and (La 0.2 Sr 0.2 Ba 0.2 Nd 0.2 Y 0.2 ) 0.95 (Co 0.19 Fe 0.76 Ni 0.05 )O 3−δ (BaNdY-5Ni). The experiment reveals that the exsolved nanoparticle number density follows the order LSCF-5Ni < CaNdY-5Ni < BaNdY-5Ni, demonstrating that high-entropy configurations can enhance exsolution. To understand this trend, we develop a Monte Carlo-based modeling framework that combines a machine-learned interatomic potential to simulate representative atomic configurations and statistically evaluate possible exsolution pathways. The results show that high-entropy configurations with greater variations in A-site cation sizes (and thus greater lattice distortions) can broaden distributions of Ni migration and reduction energies, thereby creating more thermodynamically favorable exsolution pathways.

High-entropy perovskite oxides have emerged as promising electrode materials for solid oxide electrolyzers. However, their compositional complexity makes the formation of oxygen vacancies, which influence properties such as oxygen ionic conductivity and thermal expansion, challenging to predict. Here, we experimentally measure changes in oxygen vacancy concentration for fourteen perovskite oxides with high and low-entropy A-site compositions, finding a dependence on cation size variance in addition to divalent cation fraction. Atomistic simulations using a machine-learned universal interatomic potential reveal cation size mismatches broaden a distribution of vacancy formation energies, shown through statistical thermodynamics to shift bulk formation thermodynamics. Treating oxygen vacancies statistically enables accurate predictions of oxygen vacancy formation compared to traditional models. Practically, increasing the size variance between A-site cations reduces the temperature sensitivity of oxygen vacancy concentrations, making it key for tuning critical properties. More broadly, this study demonstrates statistical treatment of oxygen vacancies is essential for understanding high-entropy perovskite oxides.

Designing electrocatalysts with high activity, selectivity, and stability for the two-electron water oxidation reaction (2e – WOR) to produce H 2 O 2 remains challenging. Although many single-component electrocatalysts have been investigated, their performance has plateaued. Here, we demonstrate an effective strategy for transforming an inert metal oxide such as rutile TiO 2 into an active electrocatalyst by supporting it with a metallic layer that incorporates very mobile charge carriers. Experimentally, we show that a Pt support layer dramatically enhances the performance of TiO 2, yielding higher current densities, lower onset potentials, and markedly improved Faradaic efficiency for producing H 2 O 2 . Our density functional theory calculations reveal the mechanistic origin of this enhancement, which lies in the presence of free-electron charge carriers on the support required for 2e – WOR. As a result, the Pt support causes the crucial OH* intermediate to bind more tightly to the active site than Pt-free TiO 2, leading to improved activity and selectivity of thin TiO 2 (110) overlayers for producing H 2 O 2 . Through a comparative analysis with a nonenhancing Ag support via examination of the density of states, we pinpoint the electronic cause as the hybridization between Ti 3 d and Pt 5 d orbitals, which creates available 5 d -band electrons near the Fermi level. This 5 d -band availability and higher work function difference allow for a more favorable charge transfer to the OH* reaction intermediate. Our findings establish the electronic character of the buried support as a key design parameter for creating highly active and selective heterostructure catalysts for H 2 O 2 synthesis.

High entropy spinel oxides provide a versatile platform for electrocatalysis because multiple metal cations can be incorporated into a single crystalline lattice, enabling tunable electronic structures. However, controlling how these cations distribute between tetrahedral and octahedral coordination sites remains a major challenge, limiting rational catalyst design. Here, we modulate cation coordination site occupancy between tetrahedral and octahedral sites in a Co–Fe–Cr–Mn–Ni framework by introducing a sixth cation (Zn, Ga, Mg, or Al) with distinct site preference energies. Using density functional theory, synchrotron X-ray absorption spectroscopy, and magnetic circular dichroism, we demonstrate that Zn preferentially occupies tetrahedral sites, driving increased octahedral occupancy of cobalt. This redistribution increases the population of octahedrally coordinated cobalt in mixed oxidation states, enhances electrical conductivity, and improves oxygen evolution reaction activity. Our findings establish coordination site occupancy as a critical design parameter, providing a strategic pathway for tailoring multicomponent spinel electrocatalysts with optimized performance. High entropy spinel oxides host multiple metals in one crystal structure with tunable catalytic properties, but controlling coordination site occupancy of metal cations is difficult. Here, the authors exploit site preference energy differences to direct metal occupancy, boosting conductivity and oxygen evolution activity.

Exsolution of metal nanoparticles (NPs) is a powerful strategy for creating strongly attached catalysts for various energy-related applications, but the conventional exsolution methods based on thermal reduction are typically conducted at high temperatures for several hours to achieve sufficient NPs formation. Here, we introduce a new flame exsolution method that is exceptionally simple and rapid for decorating perovskite oxide surfaces with a high density of metal NPs under ambient conditions. By exposing perovskite oxides to a controlled fuel-rich methane-air flame (equivalence ratio, Φ = 1.45), we produced dense Ni NPs (∼500–550 μm –2 ) in just 1 min on both (La 0.6 Sr 0.4 )(Co 0.2 Fe 0.8 Ni 0.05 )O 3-x (LSCF-5Ni) thin films and La 0.43 Ca 0.37 Ti 0.94 Ni 0.06 O 3-x (LCTN) pellets. We further demonstrated that similar exsolution results were obtained for both perovskites using a common propane torch within 20 s, demonstrating the method’s easy accessibility. This work establishes an ultrafast and transformative pathway to produce high-density metal NPs on perovskite oxides, overcoming the primary limitations of conventional exsolution techniques.

Developing robust, high-performance anode materials is essential for advancing solid oxide electrolysis cells (O-SOECs) as a key technology for high-efficiency hydrogen production. Among these materials, La 1-x Sr x Co y Fe 1-y O 3-δ (LSCF) is widely used as a mixed ionic-electronic conducting anode due to its favorable oxygen transport properties and electrochemical activity under reducing conditions [1]. A key challenge limiting the long-term performance of LSCF is the degradation of oxygen transport kinetics under operating conditions. In particular, the evolution of oxygen vacancies, changes in defect populations, and Sr surface segregation all influence ionic conductivity and ultimately the electrochemical performance and durability of O-SOECs. Understanding oxygen migration mechanisms is therefore central to improving anode functionality, as ionic transport is directly tied to the efficiency of the oxygen evolution reaction (OER) at high operating temperatures. In this work, we employ density functional theory with Hubbard U corrections (DFT+U) to investigate the formation energies of key point defects, including oxygen vacancies and interstitials, and to quantify oxygen diffusion barriers in both rhombohedral and cubic phases (as shown on figure) of LSCF using climbing image nudged elastic band (CI-NEB) calculations. We also explore Sr migration pathways and examine their coupling with oxygen vacancy dynamics to identify how they contribute to long-term degradation phenomena such as SrZrO 3 formation at the electrolyte interface [2]. We further explore high-entropy oxides (HEOs) with A-site substitutions in perovskite (ABO 3 ) structures as next-generation anode materials for O-SOECs. These high entropy perovskite oxides (HEPOs)exploit compositional complexity to suppress Sr segregation and improve structural stability and electrochemical performance. Experimental studies have demonstrated that HEPOs, such as (La 0.2 Pr 0.2 Nd 0.2 Sm 0.2 Sr 0.2 )MnO 3-δ , exhibit superior resistance to Sr surface enrichment and phase decomposition compared to conventional LSCF [3]. This is particularly important for mitigating degradation mechanisms such as Cr poisoning, where volatile chromium species, originating from stainless steel interconnects, can react with segregated Sr to form insulating phases at the electrode surface.Compositions like (La 0.2 Sr 0.2 Pr 0.2 Y 0.2 Ba 0.2 )Co 0.2 Fe 0.8 O 3-δ have shown excellent resistance to Cr poisoning, maintaining stable polarization resistance during extended high-temperature operation [4]. By combining first-principles modeling with emerging experimental results, this study establishes a foundational understanding of oxygen migration in both conventional and high-entropy anode materials. These insights directly inform the rational design of next-generation O-SOEC electrodes with improved ionic transport properties and long-term operational stability. Future work will focus on validating predicted migration pathways and degradation mechanisms through electrochemical impedance spectroscopy and high-temperature structural analysis. References: [1] S. Zhang, Mater. Today Commun. 42 , 111280 (2025). [2] D. The, J. Power Sources 275 , 901 (2015). [3] Y. Yang, J. Power Sources 482 , 228959 (2021). [4] Z. Li, ACS Appl. Mater. Interfaces 14 , 24363 (2022). Acknowledgements: This work is funded by the Department of Energy, Office of Energy Efficiency and Renewable Energy, project “Developing High-Entropy Materials as Superior Alternative Electrodes for Long-lasting Oxide-Conducting Oxide Electrolysis Cells (O-SOECs).” Figure 1

Boron (B) offers high gravimetric and volumetric energy densities, making it an attractive solid fuel for energetic applications. However, boron is hard to ignite and burns slowly and incompletely due to the presence of surface B 2 O 3 , which has a low melting point but a high boiling temperature. Recently, a new class of two-dimensional materials known as MXene (Ti 3 C 2 ) has emerged, exhibiting characteristics that could potentially enhance boron combustion, but this potential has not been previously explored. Herein, we experimentally investigate the ignition and combustion performance of boron particles, Ti 3 C 2 nanosheets, and an 80 wt.% B/Ti 3 C 2 mixture. We find that the addition of Ti 3 C 2 nanosheets enhances both the ignition and combustion properties of boron particles. Specifically, Schlieren images of CO 2 laser ignition experiments show that the B/Ti 3 C 2 mixture has a similar ignition delay time as Ti 3 C 2 but is shorter than boron, and the mixture produces more gaseous products, indicating more oxidation. Bomb calorimetry measurements show that the B/Ti 3 C 2 mixture’s heat of combustion is greater than the linear sum of its components, suggesting a favorable interaction between Ti 3 C 2 and boron. Similarly, differential scanning calorimetry shows that the mixture releases more heat overall and has lower onset temperatures than pure boron oxidation. Variable-temperature X-ray diffraction analysis of B/Ti 3 C 2 mixture shows the formation of anatase and rutile TiO 2 , TiF 2 , B 2 O 3 , and various mixed metal oxides at elevated temperatures due to reactions between boron and MXene or its oxidation products. In conclusion, these results demonstrate that Ti 3 C 2 nanosheets, and potentially other MXenes, are effective additives for promoting boron combustion, leading to easier ignition and increased combustion efficiency.

INTRODUCTION: Pancreatic cancer (PC) is one of the deadliest malignancies worldwide, with a low five-year survival rate of less than 10%. Transforming growth factor β regulator 4 (TBRG4) is differentially expressed in PC tissues, but its specific functions and regulatory role in PC have not been clarified. METHODS: role of TBRG4. RESULTS: assays verified that TBRG4 silencing inhibited tumorigenesis and TGF-β/smad3 signaling. CONCLUSION: The silencing of TBRG4 inhibits PC cell invasion, migration, EMT, and tumorigenesis by inactivating TGF-β/smad3 signaling.

Flame Treatment A rapid and scalable high-temperature flame technique is developed to enable surface reconstruction of sulfide semiconductor photoelectrode for enhanced photoelectrochemical(PEC) performances. In article number 2403164, Sung Sik Shin, Xiaolin Zheng, Sun Cho, and their team demonstrate rapid surface reconstruction of Indium sulfide (In2S3) photoanode using the flame process. The flame process transforms the In2S3 surface into a 5 nm-thin In2O3 layer and generates bulk sulfur vacancies simultaneously, leading to a remarkable PEC water oxidation photocurrent density of 8.5 mA cm−2 and iodide oxidation photocurrent density of 7.3 mA cm−2.

The two-electron water oxidation reaction (2e – WOR) has drawn growing attention as a potential method of producing hydrogen peroxide (H 2 O 2 ) on-site through electrochemical water splitting. Nevertheless, despite ongoing efforts to identify more effective electrocatalysts for 2e – WOR, it remains a challenge to simultaneously achieve good activity, selectivity, and stability with a single catalyst material. In the field of thermocatalysis, metal oxide supports have been used to tune the catalytic properties of the supported metal catalysts. Inspired by this support effect, herein, we explore using the inverse structure─a metal-supported metal oxide bilayer─as the electrocatalyst for 2e – WOR. The metal underlayer modifies the top metal oxide catalytic properties by forming a Mott–Schottky junction, where the built-in potential influences the electron transport. We investigated several semiconducting metal oxide catalysts and different metal support materials and observed a strong metal substrate effect on the metal oxide catalytic activity. We found that a thin layer of indium tin oxide (ITO) supported by a layer of platinum (Pt) shows the best catalytic ability toward 2e – WOR, which is significantly more active and stable than the unsupported ITO catalyst, and its H 2 O 2 production rates are also greater than most reported single component electrocatalysts in the literature. The method of metal underlayer provides a pathway to create future catalysts for this electrochemical reaction.

The traditional ammonia synthesis process of the Haber–Bosch method must be carried out at high temperatures and high pressure with energy-intensive consumption and carbon emission. As a green and low-energy alternative technology, photocatalytic ammonia synthesis technology has received increasing attention. Recent advances in the fabrication of TiO 2 -based nanocatalysts for the photocatalytic N 2 reduction reaction (NRR) under mild conditions are reviewed and summarized. In this paper, we compare and analyze various methods for detecting low ammonia concentrations with advanced detection techniques in the field. In order to enhance the activity of TiO 2 -based nanocatalysts for photocatalytic NRR, various methods are summarized such as molecular structure engineering, interfacial modulation engineering, crystallographic engineering, surface modulation engineering, and morphology engineering. Besides, the mechanism of the photocatalytic ammonia synthesis was analytically proven. A comparison for the advantages of the TiO 2 -based nanocatalyst synthesis process is presented and discussed, as well as its future potential applications in NRR.

Salvia miltiorrhiza Bge. (S. miltiorrhiza) is an important herbal medicine, and the researches mainly focused on the roots of S. miltiorrhiza rather than the above-ground parts. Therefore, the detailed research on the above-ground parts of S. miltiorrhiza is still unclear. This study aimed to elucidate the constituents of the above-ground parts of S. miltiorrhiza extract (SME) and evaluate its effect on endothelial cell injury. For this purpose, the constituents of SME were identified by ultra performance liquid chromatography tandem mass spectrometry, and 41 compounds were determined, including terpenes, phenylpropanoids, flavonoids, sterols and phenolic acids. The effect of SME on endothelial cell was evaluated using isoproterenol (ISO)-treated C57BL/6 mice and hypoxia-treated HUVECs, and the results revealed that SME could improve endothelial cell injury in ISO-induced myocardial infarction model and hypoxia-induced HUVECs injury. In this process, SME could inhibit reactive oxygen species production, decrease the levels of malondialdehyde, lactic dehydrogenase, pyruvic acid, intracellular iron ions and sugar, and increase the levels of glutathione. Meanwhile, SME could improve the abnormal expressions of apoptosis-, glycolysis- and ferroptosis-related proteins. Taken together, our data demonstrated that SME treatment could alleviate endothelial cell injury by regulating HIF-1α mediated glycolysis and ferroptosis signals.

Recommender Systems (RSs) are widely applied for navigating information, and Collaborative Filtering (CF) is one of prominent recommendation techniques due to the advantages of domain independence and easy interpretation. Among the numerous CF methods, Variational Autoencoders (VAE), benefiting from modeling in a probabilitistic way, stands out in capturing user preferences through representation learning. Despite the superiority, VAE-based CF models still suffer from two challenging problems: (1) Exposure bias: models in training state are narrowly exposed to a limited, biased sample of data, leading to a skewed understanding of users' true preferences; (2) Posterior collapse: models excessively simplify the learned latent variable distributions, generating na"ive representations that are unable to encapsulate the complex data patterns and thereby resulting improper recommendations. In this paper, we propose a Debiased and Representation-enhanced Variational AutoEncoder (DR-VAE) framework for collaborative recommendations. Specifically, for exposure bias problem, DR-VAE incorporates a Debiasing Estimator, mitigating the impact of exposure bias. For poster collapse issue, DR-VAE innovatively introduces a Flow-based Representation Enhancement module, ensuring us to encapsulate complex data patterns by fitting complex and intricate posterior distributions directly. We provide experimental validations over four datasets to substantiate the efficacy of our DR-VAE framework.

• Automotive bumper energy absorber combining optimized lattice structures and multi-material blends. • Requiring both low-speed collision performance and high-speed pedestrian impact protection. • 576 finite element simulations, exploring various combinations of lattice configurations, material blends, and material placements. • Optimized design and material combination providing a 68 % increase in energy absorption while achieving a 70% decrease in peak force. This study proposes a novel design for automotive bumper using optimized lattice structures and multi-materials to balance low-speed collision and high-speed pedestrian impact performance. Different blends of 20 % carbon fiber-reinforced acrylonitrile butadiene styrene with thermoplastic polyurethane were used to tailor material properties. The energy absorber features lattice structures with customized mechanical responses, created by varying the incline angle θ from 0 to 180°. We conducted 576 finite element simulations on a half-scale model to optimize energy absorption and stiffness, leading to 66 optimized designs that met both low-speed and high-speed impact criteria. Two sub-scale optimized energy absorbers with different peak forces—both meeting low-speed impact requirements—were 3D printed and validated through drop-weight testing. The one with lower peak stress demonstrated a more compliant response, exhibiting approximately 90 % lower initial peak force and an increase in energy absorption of around 33 % (from 24 J to 32 J). Compared to the baseline triangular lattice, the optimized absorber increased energy absorption by 68 % from (19 J to 32 J) and reduced peak stress by 70 %. It also showed near-complete recovery with minimal fractures, making it suitable for repeated use. This design improves safety while offering a lightweight, durable, and cost-effective bumper system.

High Resolution Image Download MS PowerPoint Slide High-entropy metal oxides (HEOs) have recently received growing attention for broad energy conversion and storage applications due to their tunable properties. HEOs typically involve the combination of multiple metal cations in a single oxide lattice, thus bringing distinctive structures, controllable elemental composition, and tunable functional properties. Many synthesis methods for HEOs have been reported, such as solid-state reactions and carbon thermal shock methods. These methods frequently are energy-intensive or require relatively expensive heating equipment. In this work, we report an ultrafast photoflash synthesis method for HEO nanoparticles on diverse substrates. The energy input is provided by a commercial Xe photoflash unit, which triggers exothermic reactions to convert metal salt precursors to HEO nanoparticles within tens of milliseconds. The formation of HEO nanoparticles is attributed to the ultrafast heating (∼10 6 K/s) and cooling (∼10 5 K/s) rates of the photoflash and overall high temperature (>1000 K) during the ultrafast synthesis process. When the synthesized CoNiFeCrMn oxide (HEO) is tested as an oxygen evolution reaction electrocatalyst, it shows similar activity to similar materials prepared by other methods. We believe this photoflash synthesis provides a simple method for many others to synthesize diverse HEOs and explore their properties and potential applications.

This Digital image correlation method is a non-contact measurement method that achieves deformation measurement by tracking speckle patterns that cover the measured surface and deform with the measured surface. The quality of speckle patterns has a significant direct impact on measurement accuracy of DIC. A global evaluation method for speckle pattern quality using a simple and easy-to-calculate parameter called two-dimensional entropy is proposed in this paper. To verify the effectiveness of the evaluation method, comparative experiments are conducted on displacement measurement accuracy using three types of speckle. The results indicate that the larger the two-dimensional entropy of the speckle, the higher the displacement measurement accuracy. In order to further validate the proposed evaluation method, measurement experiments are performed using the same type of speckle patterns with different two-dimensional entropies, and the similar conclusions are obtained. These results prove that the two-dimensional entropy of speckle pattern can be effectively used for speckle pattern quality evaluation.

Objective: Despite N 6 ‐methyladenosine (m 6 A) being closely involved in various pathophysiological processes, its potential role in liver injury is largely unknown. We designed the current research to study the potential role of fat mass and obesity‐associated protein (FTO), an m 6 A demethylase, on hepatic ischemia‐reperfusion injury (IRI). Methods: Wild‐type mice injected with an adeno‐associated virus carrying fat mass and obesity‐associated protein (AAV‐FTO) or adeno‐associated virus carrying green fluorescent protein (GFP) (AAV‐GFP) were subjected to a hepatic IRI model in vivo. Hematoxylin–eosin staining was performed to observe IRI. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was used to observe the cell apoptosis. Reverse transcription polymerase chain reaction (RT‐PCR) was used to observe the expression of FTO. The protein levels of FTO, apoptosis, or autophagy‐associated signaling proteins were detected by western blot. Reactive oxygen species (ROS) levels were determined by flow cytometry, and immunohistochemistry was used to detect the FTO and LC3‐II expression. For in vitro experiments, cultured hepatocytes were subjected to hypoxia/reoxygenation (H/R) stimulation. Monodansylcadaverine (MDC) staining was used to visualize autophagic vesicles. Results: In the present study, we showed that FTO was involved in hepatic IRI, apoptosis, and autophagy. Specifically, the expression level of FTO was significantly reduced in the hepatic IRI. Besides, increasing FTO expression (AAV‐FTO) ameliorated the hepatic IRI in animal models, accompanied by decreased apoptosis and autophagy. Furthermore, the FTO inhibitor (FB23‐2) aggravated autophagy in hepatocytes upon H/R‐induced damage. Conclusion: FTO could act as a protective effector during hepatic IRI, associated with decreased apoptosis and autophagy. FTO‐mediated m 6 A demethylation modification may be an important therapeutic target for hepatic IRI.

Additional experimental evidence was provided to demonstrate the growth of the [001]-textured BiVO 4 on polycrystalline substrates.

Highly energetic boron (B) particles embedded in hydroxyl-terminated polybutadiene (HTPB) thermosetting polymers represent stable solid-state fuel. Laser-heating of levitated B/HTPB and pure HTPB particles in a controlled atmosphere revealed spontaneous ignition of B/HTPB in air, allowing for examination of the exclusive roles of boron. These ignition events are probed in situ via simultaneous spectroscopic diagnostics: Raman and infrared spectroscopy, temporally resolved high-speed optical and infrared cameras, and ultraviolet–visible (UV–vis) spectroscopy. The emission spectra unravel two stages of the B/HTPB ignition─the exoergic ignition of boron followed by HTPB combustion. It was found that HTPB readily absorbs the energy from the irradiating carbon dioxide (CO 2 ) laser but efficiently transfers that thermal energy to the densely arranged boron particles due to the lower heat capacity of the latter. This transferred energy causes a surge in temperature for the boron particles, leading to ignition (in an oxygen environment) in B/HTPB, unlike the case with HTPB alone. The accumulated energy from the second stage of boron ignition triggers the decomposition of HTPB in conjunction with hydrogen abstraction to produce radical precursors via boron oxides (BO and BO 2 )─the key emitting intermediates detected. Along with conventional combustion products such as carbon dioxide (CO 2 ) and water (H 2 O), the formation of partially oxidized gaseous products such as methanol (CH 3 OH) and methyl vinyl ether have also been detected as a tracer of diverse oxidation events, suggesting a complex oxidation chemistry within HTPB and overall depict crucial insights for its use as a solid rocket fuel.

Over-the-air (OTA) attacks, such as flooding and jamming, significantly compromise the availability and reliability of radio access networks, especially with Quality of Service (QoS)-aware scheduling. Despite extensive research aimed at enhancing network performance and efficiency, the impact of OTA attacks on QoS performance has been insufficiently addressed. This paper conducts a thorough analysis of how OTA attacks influence the network performance. By leveraging a multi-class M/M/1 queuing model with selection and feedback mechanisms, the complex relationship between attack strategies and prioritized network scheduling is analyzed. Expressions for the delay and throughput are derived based on the result of the stationary distribution of Continuous Time Markov Chain (CTMC) model. Finally, numerical and simulation results are demonstrated to validate the theoretical analysis and to analyze the impact of OTA attack on the scheduling performance. Results show that the attacker can compromise the QoS of low-priority traffic by blocking high-priority traffic. Moreover, the QoS of high-priority traffic can be compromised by jamming low-priority traffic.

Federated Learning (FL) employs a training approach to address scenarios where users' data cannot be shared across clients. Achieving fairness in FL is imperative since training data in FL is inherently geographically distributed among diverse user groups. Existing research on fairness predominantly assumes access to the entire training data, making direct transfer to FL challenging. However, the limited existing research on fairness in FL does not effectively address two key challenges, i.e., (CH1) Current methods fail to deal with the inconsistency between fair optimization results obtained with surrogate functions and fair classification results. (CH2) Directly aggregating local fair models does not always yield a globally fair model due to non Identical and Independent data Distributions (non-IID) among clients. To address these challenges, we propose a Wasserstein Fair Federated Learning framework, namely WassFFed. To tackle CH1, we ensure that the outputs of local models, rather than the loss calculated with surrogate functions or classification results with a threshold, remain independent of various user groups. To resolve CH2, we employ a Wasserstein barycenter calculation of all local models' outputs for each user group, bringing local model outputs closer to the global output distribution to ensure consistency between the global model and local models. We conduct extensive experiments on three real-world datasets, demonstrating that WassFFed outperforms existing approaches in striking a balance between accuracy and fairness.

Ischemia/reperfusion (I/R) injury with severe cell death is a major complication involved in liver transplantation and resection. The identification of key regulators improving hepatocyte activity may provide potential strategies to clinically resolve I/R-induced injury. N6-methyladenosine (m6A) RNA modification is essential for tissue homeostasis and pathogenesis. However, the potential involvement of m6A in the regulation of hepatocyte activity and liver injury has not been fully explored. In the present study, we found that hepatocyte AlkB homolog H5 (ALKBH5) levels were decreased both in vivo and in vitro I/R models. Hepatocyte-specific ALKBH5 overexpression effectively attenuated I/R-induced liver necrosis and improved cell proliferation in mice. Mechanistically, ALKBH5-mediated m6A demethylation improved the mRNA stability of YTH N6-methyladenosine RNA-binding protein 1 (YTHDF1), thereby increasing its expression, which consequently promoted the translation of Yes-associated protein (YAP). In conclusion, ALKBH5 is a regulator of hepatic I/R injury that improves hepatocyte repair and proliferation by maintaining YTHDF1 stability and YAP content. The ALKBH5–m6A–YTHDF1–YAP axis represents promising therapeutic targets for hepatic I/R injury to improve the prognosis of liver surgery.

Federated learning (FL) is a promising machine learning paradigm that collaborates with client models to capture global knowledge. However, deploying FL models in real-world scenarios remains unreliable due to the coexistence of in-distribution data and unexpected out-of-distribution (OOD) data, such as covariate-shift and semantic-shift data. Current FL researches typically address either covariate-shift data through OOD generalization or semantic-shift data via OOD detection, overlooking the simultaneous occurrence of various OOD shifts. In this work, we propose FOOGD, a method that estimates the probability density of each client and obtains reliable global distribution as guidance for the subsequent FL process. Firstly, SM3D in FOOGD estimates score model for arbitrary distributions without prior constraints, and detects semantic-shift data powerfully. Then SAG in FOOGD provides invariant yet diverse knowledge for both local covariate-shift generalization and client performance generalization. In empirical validations, FOOGD significantly enjoys three main advantages: (1) reliably estimating non-normalized decentralized distributions, (2) detecting semantic shift data via score values, and (3) generalizing to covariate-shift data by regularizing feature extractor. The prejoct is open in https://github.com/XeniaLLL/FOOGD-main.git.

BACKGROUND: The non-exercise estimated cardiorespiratory fitness (eCRF) has been recognized as an important predictor of mortality among the general population. This study sought to evaluate the relationship between eCRF and mortality from all causes, cardiovascular disease (CVD), and cancer in hypertensive adults. METHODS: We included 27,437 adults with hypertension from the National Health and Nutrition Examination Survey (NHANES) III and 10 NHANES cycles from 1999 to 2018. Multivariate Cox proportional hazard models were used to assess the hazard ratios and 95% confidence intervals (CIs) of eCRF for mortality. RESULTS: A total of 8,023 deaths were recorded throughout a median 8.6-year follow-up, including 2,338 from CVD, and 1,761 from cancer. The eCRF with per 1 metabolic equivalent increase was linked to decreased risk of all-cause (adjusted HR 0.78, 95% CI: 0.75-0.81) and CVD mortality (adjusted HR 0.79, 95% CI: 0.74-0.84), rather than cancer mortality (adjusted HR 0.94, 95% CI: 0.86-1.03). Moreover, a stronger protective effect of eCRF was observed for females (HR 0.66 (95% CI: 0.62-0.72) versus HR 0.78 (95% CI: 0.73-0.83), Pinteraction < 0.001 for all-cause mortality; HR 0.70 (95% CI: 0.61-0.80;) versus HR 0.82 (95% CI: 0.73-0.92), Pinteraction = 0.026 for CVD mortality) compared with males. Findings did not significantly differ in subgroup analyses and sensitivity analyses. CONCLUSIONS: Among adults with hypertension, eCRF was inversely related to all-cause and CVD mortality, but not cancer mortality. A significant interaction effect existed between sex and eCRF. Further studies are needed to verify this association in different populations.

Solid composites comprising boron (B) and hydroxyl-terminated polybutadiene (HTPB) are attractive potential fuels for air-breathing propulsion systems. However, a notable challenge for B/HTPB composites is the sluggish oxidation kinetics of B that hinders the combustion efficiency. Addition of fluoropolymers in the composites has been found to improve the combustion efficiency but reduce the mechanical performance. Herein, we covalently grafted fluoroalkyl side chains onto HTPB (F-HTPB) to introduce fluorocarbon to B/HTPB composites without macroscopic phase separation and compared with composites using hydrocarbon grafted HTPB, pristine HTPB, and blended fluorocarbon. We found that B/HTPB composites using F-HTPB exhibited the best combined mechanical performance and flame propagation rate compared with the other formulations. These findings demonstrate that functionalizing HTPB with fluorocarbon side chains is a promising strategy to improve the mechanical and combustion properties of solid composite fuels.

The two-electron water oxidation reaction (WOR) has drawn recent attention as an efficient, simple method of producing on-site hydrogen peroxide through electrochemical water splitting. However, finding electrocatalysts with high activity and stability in the reaction’s harsh oxidizing environment is challenging. This is made especially difficult due to the competing four-electron and one-electron WOR, which necessitate a catalyst material with high selectivity towards the hydrogen peroxide production pathway. Here, we utilize metal/metal oxide coupling interactions to tune metal oxide catalysts to be active for 2 e - WOR via the creation of a Mott-Schottky junction. Metal/metal oxide interactions have been well studied for supported metal catalysts, but their behavior has been less investigated for catalysts where the metal oxide provides the active sites, called inverse catalysts. In this case, the difference in work functions between the metal support and metal oxide semiconductor catalyst drives electron transfer between the two, facilitating electronic interactions which can modulate the binding energy of reaction intermediates at the interface and thus tune the catalytic ability of the metal oxide. In this work, we investigate a variety of metal oxide catalyst and metal support materials and find that a thin layer of indium tin oxide (ITO) supported by a layer of platinum (Pt) has excellent catalytic ability towards 2 e - WOR. The activity, selectivity, and stability of the ITO/Pt catalyst shows significant improvement over the unsupported ITO catalyst and yields hydrogen peroxide production rates greater than those reported in literature using other standard catalysts for this reaction. This method of catalyst engineering also provides a future pathway to create new catalysts for this electrochemical reaction. Figure 1

Rehmannia glutinosa Libosch, Achyranthes bidentata Bl. (A. bidentata), Dioscorea opposita Thunb, and Chrysanthemum morifolium Ramat (C. morifolium) are known as the ‘Four Huaiqing Chinese Medicine’ in China, which are used as materials for functional foods. In this paper, the constituents of Four Huaiqing Chinese Medicine were identified by UPLC-Q-TOF-MS/MS, and flavones and aromatic compounds are mainly responsible for these herbs. Moreover, C. morifolium exhibited the most significant effect in cobalt chloride-induced HUVECs injury, which could decrease cell apoptosis and the overproduction of ROS, lactic dehydrogenase (LD) and pyruvic acid, and increase the migration capacity of cells. Meanwhile, A. bidentata exhibited the most significant effect in isoproterenol-induced H9C2 cell injury, which could decrease the levels of ROS overproduction, BNP, NO, LD and pyruvic acid. Western blot revealed that C. morifolium and A. bidentata also could decrease the levels of bax/bcl-2 ratio, cleaved caspase-3, cytochrome c, HIF-1ɑ, GLUT1, HKII and PFKFB3, respectively.

Federated learning achieves effective performance in modeling decentralized data. In practice, client data are not well-labeled, which makes it potential for federated unsupervised learning (FUSL) with non-IID data. However, the performance of existing FUSL methods suffers from insufficient representations, i.e., (1) representation collapse entanglement among local and global models, and (2) inconsistent representation spaces among local models. The former indicates that representation collapse in local model will subsequently impact the global model and other local models. The latter means that clients model data representation with inconsistent parameters due to the deficiency of supervision signals. In this work, we propose FedU<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> which enhances generating uniform and unified representation in FUSL with non-IID data. Specifically, FedU<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> consists of flexible uniform regularizer (FUR) and efficient unified aggregator (EUA). FUR in each client avoids representation collapse via dispersing samples uniformly, and EUA in server promotes unified representation by constraining consistent client model updating. To extensively validate the performance of FedU<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, we conduct both cross-device and cross-silo evaluation experiments on two benchmark datasets, i.e., CIFAR10 and CIFAR100.

The emergence of ChatGPT marks the arrival of the large language model (LLM) era. While LLMs demonstrate their power in a variety of fields, they also raise serious privacy concerns as the users' queries are sent to the model provider. On the other side, deploying the LLM on the user's device will also leak all the model data. Existing methods based on secure multiparty computation (MPC) managed to protect both the privacy of the model parameters and user queries. However, they require gigabytes of data transfer and several minutes to generate just one token, making them impractical for most real-world applications. To improve the efficiency of private LLM inference, we propose PermLLM, which accelerates the evaluation of non-linear functions using secure random permutation. Along with the optimized secret sharing protocols and homomorphic encryption, PermLLM achieves two-party private inference of the ChatGLM-6B model at the speed of around 3s/token, under a realistic network setting (10ms RTT and 1Gbps bandwidth), which is magnitudes faster than existing MPC solutions.

Abstract Surface reconstruction, reorganizing the surface atoms or structure, is a promising strategy to manipulate materials' electrical, electrochemical, and surface catalytic properties. Herein, a rapid surface reconstruction of indium sulfide (In 2 S 3 ) is demonstrated via a high‐temperature flame treatment to improve its charge collection properties. The flame process selectively transforms the In 2 S 3 surface into a diffusionless In 2 O 3 layer with high crystallinity. Additionally, it controllably generates bulk sulfur vacancies within a few seconds, leading to surface‐reconstructed In 2 S 3 (sr‐In 2 S 3 ). When using those sr‐In 2 S 3 as photoanode for photoelectrochemical water splitting devices, these dual functions of surface In 2 O 3 /bulk In 2 S 3 reduce the charge recombination in the surface and bulk region, thus improving photocurrent density and stability. With optimized surface reconstruction, the sr‐In 2 S 3 photoanode demonstrates a significant photocurrent density of 8.5 mA cm −2 at 1.23 V versus a reversible hydrogen electrode (RHE), marking a 2.5‐fold increase compared to pristine In 2 S 3 (3.5 mA cm −2 ). More importantly, the sr‐In 2 S 3 photoanode exhibits an impressive photocurrent density of 7.3 mA cm −2 at 0.6 V versus RHE for iodide oxidation reaction. A practical and scalable surface reconstruction is also showcased via flame treatment. This work provides new insights for surface reconstruction engineering in sulfide‐based semiconductors, making a breakthrough in developing efficient solar‐fuel energy devices.

Abstract Phonon polaritons, the hybrid quasiparticles resulting from the coupling of photons and lattice vibrations, have gained significant attention in the field of layered van der Waals heterostructures. Particular interest has been paid to hetero‐bicrystals composed of molybdenum oxide (MoO 3 ) and hexagonal boron nitride (hBN), which feature polariton dispersion tailorable via avoided polariton mode crossings. In this work, the polariton eigenmodes in MoO 3 ‐hBN hetero‐bicrystals self‐assembled on ultrasmooth gold are systematically studied using synchrotron infrared nanospectroscopy. It is experimentally demonstrated that the spectral gap in bicrystal dispersion and corresponding regimes of negative refraction can be tuned by material layer thickness, and these results are quantitatively matched with a simple analytic model. Polaritonic cavity modes and polariton propagation along “forbidden” directions are also investigated in microscale bicrystals, which arise from the finite in‐plane dimension of the synthesized MoO 3 micro‐ribbons. The findings shed light on the unique dispersion properties of polaritons in van der Waals heterostructures and pave the way for applications leveraging deeply sub‐wavelength mid‐infrared light‐matter interactions.

Roughly, 50% of primary energy worldwide is rejected as waste heat over a wide range of temperatures. Waste heat above 573 K has the highest Carnot potential (&amp;gt;50%) to be converted to electricity due to higher Carnot efficiency. Thermoelectric (TE) materials have gained significant attention as potential candidates for efficient thermal energy conversion devices. Silicon nanowires (SiNWs) are promising materials for TE devices due to their unique electrical and thermal properties. In this study, we report the successful fabrication of high-quality double-sided SiNW arrays using advanced techniques. We engineered the double-sided structure to increase the surface area and the number of TE junctions, enhancing TE energy conversion efficiency. We also employed non-agglomeration wire tip engineering to ensure uniformity of the SiNWs and designed effective Ohmic contacts to improve overall TE efficiency. Additionally, we post-doped the double-sided SiNW arrays to achieve high electrical conductivity. Our results showed a significant improvement in the TE performance of the SiNW array devices, with a maximum figure-of-merit (ZT) value of 0.24 at 700 K, fabricated from the single SiNW with ZT of 0.71 at 700 K in our previous work [Yang et al., Nat. Commun. 12(1), 3926(2021)].

Abstract Alkaline water splitting electrocatalysts have been studied for decades; however, many difficulties remain for commercialization, such as sluggish hydrogen evolution reaction (HER) kinetics and poor catalytic stability. Herein, by mimicking the bulk‐heterojunction morphology of conventional organic solar cells, a uniform 10 nm scale nanocube is reported that consists of subnanometer‐scale heterointerfaces between transition metal phosphides and oxides, which serves as an alkaline water splitting electrocatalyst; showing great performance and stability toward HER and oxygen evolution reaction (OER). Interestingly, the nanocube electrocatalyst reveals acid/alkaline independency from the synergistic effect of electrochemical HER (cobalt phosphide) and thermochemical water dissociation (cobalt oxide). From the spray coating process, nanocube electrocatalyst spreads uniformly on large scale (≈6.6 × 5.6 cm 2 ) and is applied to alkaline water electrolyzers, stably delivering 600 mA cm −2 current for &gt;100 h. The photovoltaic‐electrochemical (PV‐EC) system, including silicon PV cells, achieves 11.5% solar‐to‐hydrogen (STH) efficiency stably for &gt;100 h.