Metal-organic frameworks (MOFs) have emerged as promising platforms for developing high-performance oxygen evolution reaction (OER) electrocatalysts. However, their inherently poor conductivity and sluggish kinetics significantly limit OER activity, which hinders practical applications. Herein, we demonstrate a novel strategy to enhance the OER performance of Ni-btc MOFs through the spatial confinement effect of accordion-like MXene. The optimized Ni-btc/Ti3C2Tx catalyst achieves an exceptional OER performance with an overpotential of 225 mV at 10 mA cm-2 and stability for 144 h under a high current density of 1000 mA cm-2, in sharp contrast to pristine Ni-btc (309 mV at 10 mA cm-2). In situ electrochemical impedance spectroscopy (EIS) analyses reveal that the spatial confinement effect induced by MXene optimizes the adsorption kinetics of reaction intermediates, accelerates electron transfer dynamics, and enhances mass transport during OER. This study demonstrates that spatial confinement strategies can effectively improve the reaction kinetics in OER processes, thereby enhancing catalytic activity.

Accurate identification of chiral enantiomers is of critical importance in pharmaceutical sciences, as mirror-image isomers often exhibit distinct pharmacological activities and toxicological profiles. Surface-enhanced Raman spectroscopy (SERS) is a powerful tool for discriminating between chiral molecules due to its ability to provide highly specific molecular vibrational fingerprints. However, its practical application is frequently hindered by significant signal variability and limited detection sensitivity. To address these persistent challenges, a novel SERS substrate is developed through the coffee-ring effect (CRE)-driven self-assembly of chiral plasmonic gold-nanorod@silver (AuNR@Ag) core–shell nanostructures. Chiral AuNR@Ag nanostructures are synthesized by immobilizing L-/D-cysteine (Cys) onto AuNR surfaces, followed by the controlled deposition of a silver shell. These nanoparticles subsequently self-assemble into dense, uniform, ring-shaped architectures during droplet evaporation, which generates highly reproducible and intense electromagnetic hotspots. As a proof-of-concept, this platform achieves sensitive enantiomeric discrimination of naproxen with a significant 3- to 4-fold SERS signal contrast and high reproducibility, enabling the quantitative determination of enantiomeric excess (ee). Furthermore, this CRE-driven self-assembly strategy provides a promising and versatile platform for accurate enantioselective sensing of chiral drug enantiomers.

Metallic phase molybdenum disulfide (1T-MoS2), a kind of metallic phase, possesses higher carrier concentration and electronic conduction potential, which shows significant application in the fields of electrochemical energy storage. This work systematically exhibits the dual-driving mechanism through Co doping and plasma treatment and the optimization of electrochemical performance for 1T-MoS2 as electrodes materials. Due to the differences in electronic structure and the mismatching of atomic radii between Co and Mo atoms, the electron cloud distribution inside 1T-MoS2 is regulated through precise doping of Co atoms, which leads to the stability of the 1T phase improving and electron transport channels constructing. Meanwhile, eliminating impurities and creating defects for 1T-MoS2 are implemented by the plasma treatment, which promotes the uniform diffusion of Co element to alleviate the stress in lattices. A dual-driving optimization system with synergistic effect is realized. As a result, Co(1%)-D-Mo0.64W0.36S2–30 W electrode has the maximum specific capacitance (232.2 F g−1), which is higher than that of the original Mo0.64W0.36S2 (100 F g−1) and Co(1%)-D- Mo0.64W0.36S2 (133.8 F g−1) without plasma treatment. After 10,000 cycles at 15 A g−1, 93.9% of the initial capacitance is retained. Similarly, for LIBs, the Co(1%)-D-Mo0.64W0.36S2–30 W achieves an initial specific capacity of 1087.1 mAh g−1 and retains a specific capacity of 490.6 mAh g−1 after 600 cycles. This work provides a scalable strategy to modify layered sulfides through plasma power regulation, and provides a new insight into the design of high-performance electrode materials for dual-mode energy storage systems.

Accurate, nondestructive, and high-throughput identification of sorghum varieties is essential for intelligent agriculture and industrial quality control. However, single-modality approaches based on either RGB imaging or near-infrared spectroscopy (NIRS) often suffer from limited robustness when inter-variety differences are subtle. In this study, a synchronized NIRS-RGB multimodal dataset was established for six representative sorghum varieties, including 14,400 paired kernel samples (2400 samples per variety). Each sample contained a NIR spectrum covering 901.467-1706.76 nm with 256 variables and a high-resolution RGB image. An end-to-end Multi-modal Attention Network (MAN) was proposed to perform deep feature fusion through bi-directional cross-attention, enabling adaptive interactions between spectral physicochemical signatures and visual appearance cues. Under an 8:2 split with five-fold cross-validation on the training set, MAN with a Swin Transformer visual branch achieved an accuracy of 0.9888 and an F1-score of 0.9889, outperforming the best NIRS-only baseline by 8.40 percentage points in accuracy and the best RGB-only baseline by 6.09 percentage points. Ablation results showed that cross-attention was the major contributor to performance improvement, while bi-directional interaction and spectral SE attention further enhanced discriminability. Furthermore, independent three-day replicate validation achieved a pooled accuracy of 0.9758 and an F1-score of 0.9755, indicating good generalizability across acquisition days. The results demonstrate that the proposed MAN effectively exploits complementary NIRS and RGB information, providing a robust and nondestructive solution for fine-grained sorghum variety identification in high-throughput inspection scenarios.

For wide-bandgap perovskite solar cells (WBG PSCs), significant non-radiative recombination at the self-assembled monolayers (SAM)/perovskite interface, along with high bulk defect densities, severely limit device performance and stability. Hence, defect suppression is critical to boost the fill factor (FF), representing a key challenge in developing high-performance devices. Herein, an innovative bilayer SAM (BL-SAM) interface engineering strategy is proposed via sequentially depositing [4-(3,6-dimethyl-9H-carbazol-9-yl) butyl] phosphonic acid (Me-4PACz) and (4-(3,11-dimethoxy-7H-dibenzo[c,g]carbazol-7-yl) butyl) phosphonic acid (MeO-4PADCB) as hole selective layers (HSLs). Theoretical and experimental evidence demonstrate that MeO-4PADCB not only passivates the defects caused by incomplete coverage at buried surface, but also interacts with the perovskite lattice. The BL-SAM strategy operates by orchestrating three synergistic effects: interfacial vacancy passivation, energy level alignment optimization, and residual lattice strain release. Consequently, the champion power conversion efficiency (PCE) of 20.35% is obtained with an exceptional FF of 85.18% and an open-circuit voltage (V OC) of 1.33 V for 1.79 eV PSCs, representing one of the highest performance values reported in this category. The general applicability of this strategy is validated in 1.85 eV PSCs, delivering a PCE of 19.24% and an outstanding V OC of 1.39 V. This work offers a versatile strategy for achieving high-performance and stable PSCs.

Microbial fuel cells (MFCs) offer a sustainable solution for clean energy generation and organic wastewater treatment. However, the resource utilization of waste MFCs has received limited attention. Here, we effectively repurposed components of waste MFCs to enhance practical performance and environmental treatment. First, using the waste MFC reactors can obtain a fast start-up time and save 42 % of the time compared to new MFC reactors. Microbial community analysis revealed that the waste MFC reactor walls/residual electrolytes harbored a more evenly distributed microbial community, enriched with genera such as Pseudomonas that could secrete electron shuttles to facilitate extracellular electron transfer and might have been responsible for the rapid startup of the MFCs using the waste reactors. The gene predictions indicated that the anode biofilms were enriched with more anaerobic bacteria, but many of the gene functions and modules were less abundant. Furthermore, the waste anolyte proved to be a valuable inoculum source for the newly built MFCs; the waste catholyte demonstrated efficacy as a reducing agent for Cr(VI), and more than 88 % of the Cr(VI) was reduced. These findings provide a paradigm for the resource utilization of waste MFCs and can be generalized to the whole microbial electrochemical system.

As the primary raw material for Chinese Baijiu, the variety of sorghum grains critically influences the taste, flavor, and yield of the Baijiu. This study introduces a method based on near-infrared spectroscopy (NIRS) and decision-level fusion (DLF) for high-throughput, nondestructive, and precise discrimination of sorghum grain varieties. The method leverages the relationship between key sorghum components (e.g., tannin content and amylopectin content) and variety categories. By fusing the spectral prediction results from NIR quantitative models for single-grain tannin content, single-grain amylopectin content, and a basic qualitative model for single-grain variety category, a DLF model with enhanced variety discrimination performance can be developed for high-throughput detection. This method was validated using more than 20 000 sorghum grains from 82 different sources. Spectral acquisition and analysis were performed on a self-developed high-throughput single-grain detection device. Results show that, under various machine learning algorithm optimizations, the proposed DLF model improved prediction accuracy by 0.09%-7.82% and prediction F1-score by 0.09%-9.77% compared to the basic sorghum variety discrimination model. This demonstrates the effectiveness of the proposed method and its potential in the quality control of raw materials for Baijiu production.

Rapid detection of pumpkin quality is of great significance for pumpkin production and breeding. In this study, hyperspectral imaging technology was utilized to facilitate the rapid detection of moisture content, starch content, and sensory quality in pumpkins, as well as to investigate their distribution within the pumpkin. The hyperspectral imaging data acquired from pumpkin slices was extracted and averaged. The models for moisture content and starch content in pumpkin built under Multiple Scatter Correction (MSC) pretreatment and Ridge regression proved to be the best ones, whose determination coefficients for cross-validation (R2cv) were 0.968 and 0.869, and the root mean square error for cross-validation (RMSEcv) were 1.142 and 0.365, respectively. Based on the moisture and starch values of pumpkin slices predicted by these models, the sensory quality scores of pumpkin slices can be further estimated. The sensory quality evaluation equation of pumpkin has a correlation of 0.934 to the sensory quality score of pumpkin obtained from the cooking experiment. Additionally, distribution maps summarizing the moisture, starch, and sensory quality of the pumpkin slices were generated, which could well reflect the spatial distribution characteristics of pumpkin quality indexes.

BACKGROUND:It is a common issue that the prediction results of near-infrared (NIR) calibration models are sensitive to changes in measurement conditions. Many existing chemometric algorithms focus on optimizing the accuracy of NIR quantitative models. However, to ensure that a NIR model remains effective across a broader range of working conditions and minimizes the need for recalibration, a more robust model-though not necessarily the most accurate one-may sometimes be a more suitable choice. RESULTS:This study aims to develop a simple method, termed the external calibration-assisted screening (ECA) method, to help analysts identify quantitative models with the highest robustness. In this approach, a model constructed under previous conditions is externally calibrated using samples collected under new conditions. Based on cross-validation and external calibration results, a new metric (PrRMSE) is introduced to assess the model's robustness. By adjusting the modeling parameters, the most robust model can be selected. This method can be integrated with competitive adaptive reweighted sampling (CARS) for model optimization. Using a set of lab-measured rice flour data and two publicly available corn datasets as test samples, we compared the proposed method (ECCARS, external calibration-assisted CARS) with control methods (CARS method). The results demonstrate that, compared with the CARS method, models selected by ECCARS achieved reductions in root mean square error of 12.15 %-725 % for calibration and 27.63 % to 482.00 %for validation under varying measurement conditions, indicating a marked enhancement in robustness. SIGNIFICANCE AND NOVELTY:This study introduces a method for evaluating and selecting NIR quantitative models based on robustness. The proposed ECA method has the potential to be integrated with additional chemometric techniques or strategies to reduce barriers to NIRS adoption and enhance its effectiveness.

Zero-dimensional (0D) organic indium halides have been emerged as promising broadband light emitters with wide application prospects, but most of the present halides suffer from low photoluminescence quantum yield (PLQY), and a high-power excitation light source is needed to obtain desirable performance. In this work, we elaborately select appropriate organic cations as crystal structural engineering and obtained a series of highly efficient 0D indium bromides. Under UV light excitation, these 0D indium halides display broadband yellow light emissions (550-600 nm) with near-unity PLQYs, which represents one of the highest values in all the previously reported indium halides. Benefiting from successful nanoscale engineering, highly luminescent inks based on these 0D indium halides are facilely prepared by dispersing nanocrystals into various organic solvents. The luminescent ink can be utilized to print various anticounterfeiting patterns, which displays photoreversible switching with visible/invisible transformation under the alternating irradiation of UV and visible light. Furthermore, white light emitting diodes can be fabricated with high color rendering index above 90 by using these 0D halides as down-conversion phosphors. This work not only promotes the development of indium halides but also significantly broadens the application in solid-state illumination and anticounterfeiting, etc.

Emitter and transparent conductive oxide (TCO) films are the critical functional layers of extremely promising silicon heterojunction (SHJ) solar cells. Here, p-type nanocrystalline silicon oxide (nc-SiOx:H(p+)) are employed as the emitter, replacing the widely used nanocrystalline silicon. The nc-SiOx:H shows a mixed-phase structural characteristic of nanocrystalline silicon grains and amorphous silicon oxide, in which the former spans the whole emitter, facilitating the carrier collection. A variety of TCO films, including Ce, Sn, or Hf doped and undoped indium oxides, are optimized for the nc-SiOx:H(p+) emitter. Film quality, work function, and bandgap states of the TCO films affect the contact resistivity of TCO/nc-SiOx:H(p+) and the solar cell performance. Using Ce doped indium oxide (ICO) with high mobility and certain bandgap states as the TCO layers, an efficiency of 26.29% and a high fill factor (FF) of 86.21% are achieved on the champion bifacial SHJ solar cells.

Winograd algorithm powerfully accelerates Convolutional Neural Networks. However, for backward-filter convolution (BFC), existing implementations often struggle to achieve both high throughput and low memory usage, due to challenges from large filters and small outputs. We propose WinRS, a fast, memory-efficient, and flexible BFC algorithm. WinRS reduces N-D large filters into 1D formats and precisely splits them to match the fastest kernels. These fully-fused kernels execute BFC in on-chip memory with tiny workspace, leveraging the superior acceleration potential of 1D Winograd. WinRS adaptively balances workloads into an optimal number of block groups, maximizing hardware utilization in small-output cases. When ported to FP16 on Tensor Cores, WinRS achieves 3.27x throughput of its FP32 CUDA-Core version. In experiments, WinRS achieves 1.05x to 4.7x speedup over cuDNN GEMM using comparable workspace; WinRS uses less than 4% workspace of cuDNN FFT and Winograd, and exhibits higher throughput with memory- and FLOP-bound workloads.

Sorghum is a primary raw material for Chinese Baijiu production. Rapid and accurate identification of sorghum varieties is crucial for quality control in the Baijiu industry. Near-infrared spectroscopy (NIRS) offers a promising nondestructive solution, though conventional discriminant models-relying solely on category labels-lack interpretability. This study proposes a novel classification method integrating NIRS with decision-level fusion (DLF). The approach involves developing quantitative NIRS models for sorghum tannin and amylopectin contents, building a label-based sorghum variety discriminant model, and fusing predictions from these three models via DLF to calculate the new category prediction values. Experiments on 81 sorghum flour samples from diverse sources demonstrated that DLF generally improved model discrimination performance. Specifically, partial least squares discriminant analysis (PLS-DA) model preprocessed with standard normal variate (SNV) + competitive adaptive reweighted sampling (CARS) and optimized by DLF achieved an accuracy of 0.938, while the CARS-PLS-DA model with DLF reached an accuracy of 0.951. The proposed method effectively bridges quantitative and qualitative analysis, enhancing model interpretability while improving discriminatory power.

The design of transition metal oxides (TMOs)-based sensing materials for the electrochemical detection of heavy metal ions in water environment is critically important. In this study, we propose an economical and straightforward strategy to synthesize a hollow flower-like NiO@Co3O4 heterojunction catalyst via in situ calcination of NiCo-layered double hydroxide (NiCo-LDH) hollow nanocages, using ZIF-67 as a template. A series of independent experiments including scanning electron microscopy, transmission electron microscopy, and X-ray photoelectron spectroscopy were performed to characterize the NiO@Co3O4 material. Furthermore, the electrochemical performance of NiO@Co3O4 was evaluated through cyclic voltammetry and electrochemical impedance spectroscopy. According to the experiment characterizations, the NiO@Co3O4 heterojunction exhibits a larger specific surface area, improved electrical conductivity, and stronger valence cycle capability between Ni (II)/(III) and Co(II)/(III). When used as a modified electrode-sensitive interface at the glassy carbon electrode (NiO@Co3O4/GCE) with Nafion-assisted, it can provide a sensitivity of 86.20 mu A mu M-1 and a theoretical detection limit of 12.66 nM for the electrochemical detection of Pb(II) within a concentration range of 0.05 -1.0 mu M in an acetate buffer solution (ABS, pH 5.0). Furthermore, the engineered hollow flower-like NiO@Co3O4 sensor demonstrated exceptional long-term stability, anti-interference capability, selectivity, and reproducibility. These characteristics indicate that NiO@Co3O4 is an auspicious material for detecting Pb(II) in actual water samples.

The study of sodium content in plants is crucial for the improvement of saline-alkali soil. Existing metal element detection methods pose challenges because they are complicated and time-consuming. In this study, we propose a quantitative detection model, FusionNet, that integrates Laser-Induced Breakdown Spectroscopy (LIBS) and Near-Infrared Hyperspectral Imaging (NIR-HSI) to realize the detection of Na element content in sorghum roots. To address the small-sample dataset, A Generative Adversarial Network (GAN) was employed to increase the diversity of the samples. The results indicated that data augmentation effectively enhanced the diversity of the original dataset and improved model performance. The modeling results from the FusionNet network achieved R2cv of 0.9915 and RMSECV of 0.7418, while R2p and RMSEP were 0.9808 and 0.6693. Compared to training with LIBS data alone, FusionNet achieved improvements of 4.94 % in R2cv and 5.61 % in R2p. This study provides a new method for detecting metal elements in plants.

Recent development of artificial neural networks (ANNs) and inverse design methods have demonstrated their prospective significance for planar diffractive lens design, with a plethora of optical lenses designed for wavelengths ranging from visible to Thz wavelengths. However, previous research to design planner diffractive lenses only considers the maximum intensity in the focus area or its derivatives as the optimization function, leaving the intensity outside the focus area unconsidered. We proposed and investigated a two-dimensional (2D) physics-driven ANN method assisted by the negative Pearson correlation coefficient (NPCC) to design microlenses with varied focusing distances, which takes the entire 2D intensity distribution at the focus plane as an optimization function. Taking advantage of 3D two-photon nanolithographic technology, sub-micrometer thickness microlenses with varied focusing lengths are designed and fabricated, achieving an average focusing efficiency of around 35%, and an average focusing spot size of about 1 µm. Furthermore, a microlens array (19 by 19 microlenses with a total size of 4 mm2) with a curved focusing plane was fabricated and integrated into a CMOS sensor, achieving direct object imaging under incoherent white light illumination. Our results demonstrate that the NPCC is a very useful optimization function for designing planar diffractive lenses, and the use of NPCC in ANNs is of great potential for the future design of functional diffractive optical elements in optics and nanophotonics.

In recent years, the Factor Graph Optimization (FGO) algorithm has gained a great attention in the field of integrated navigation owing to its better positioning performance than the traditional filter-based approaches. However, the practical application of the FGO algorithm remains challenging due to its significant computational complexity and processing time consumption, especially for the case of limited storage and computation resources. In order to overcome the problem, we first conduct a thorough analysis of the factor graph model for the Global Navigation Satellite System/Inertial Navigation System (GNSS/INS) integrated navigation. Then, based on the Incremental Smoothing and Mapping (iSAM), an Optimized iSAM (OiSAM) algorithm is proposed to efficiently solve the optimization problem in FGO, with reducing computational load and required memory resources. For the re-linearization problem, we propose a novel Adaptive Joint Sliding Window Re-linearization (A-JSWR) algorithm combining periodic and on-demand re-linearization to further improve the efficiency of OiSAM. Finally, the OiSAM-FGO method utilizing OiSAM and A-JSWR is presented for the GNSS/INS integrated navigation. The experiments on real-world datasets demonstrated that the OiSAM-FGO can reduce the time consumption of the optimization procedure by up to 52.24%, while achieving a performance equivalent to that of the State-of-the-Art (SOTA) FGO method and superior to the Extended Kalman Filter (EKF) method.

Near-infrared phosphor conversion light-emitting diode (NIR pc-LED) is a widely used ideal NIR light source due to its unique advantages of low cost, energy saving, compactness and long operational lifetime. However, there is currently a scarcity of appropriate phosphor for NIR light sources utilized in spectral detection, with emission peak larger than 850 nm and steady spectrum output of ultra broadband spectra. Here, a broadband Cr3+ activated LiScSnO4 (LSS) phosphor (lambda ex-max = 460 nm) matching the pump band of commercial blue LED chips has been developed, which shows a long wavelength NIR emission peaking at 900 nm. Cr3+ simultaneously replace the Wyckoff site co-occupied by Sc and Sn in LSS. Due to the difference of local environment of the specific crystal field, two different wide emission bands are caused. These two broad emission bands both contribute to the broadband luminescence with a full width at half maximum (FWHM) of 227 nm. Benefiting from the crystallography consistent Wyckoff position occupation of Cr3+ ions, the spectral distribution of LSS: Cr3+ exhibit excellent stability over a wide temperature span (80-423 K). The broadband NIR pc-LED prototype based on LSS:Cr3+ shows great potential in NIR spectrum detection, night vision imaging applications.

Accurate classification of sorghum varieties is crucial to the production and processing of liquor with sorghum as raw materials. Hyperspectral imaging (HSI) has the potential to achieve this goal quickly and nondestructively. This study proposes a novel algorithm, an improved Principal Component Analysis combined with SpectrumImage-Convolutional Neural Network (PCA-SICNN), which can combine spectral features and image features of HSI data, to enhance the accuracy of variety identification of sorghum seeds. To verify the effect of this algorithm, hyperspectral imaging data (939-1700 nm) of 13,200 sorghum seeds from 6 varieties were collected. The principal component analysis (PCA) was employed to select 20-dimension images from the origin hyperspectral imaging data. Spectrum-Image-Convolutional Neural Network (SICNN) extracts the spectral and image features of sorghum in the network and then fuses the features. By fully learning the HSI data features of sorghum, it achieves the classification of sorghum varieties. The results demonstrate that PCA-SICNN achieves an accuracy on the training set and on the test set reaches 98.67 % and 98.64 %, respectively. Compared with other control methods, the prediction accuracy of the PCA-SICNN model increased by at least 1.10 %. These results suggest the potential for the method to be widely applied in the production and processing of sorghum.

Sorghum is an important crop, and the quality of sorghum of the same variety from different geographic origins varies greatly. This study focuses on HongYingZi sorghum from five distinct origins, employing a combination of hyperspectral imaging (HSI) technology and machine learning algorithms to investigate methods for classifying sorghum origin. Multiplicative scatter correction and the Savizkg-Golay algorithms were used to preprocess HSI data, and the characteristic wavelengths were screened by the successive projections algorithm (SPA). Based on AdaBoost, ExtraTreesClassifier, Gradient Boosting, Decision Tree, and Random Forest algorithms, classification models based hyperspectral data were established respectively, and validation experiments were conducted. The results show that for the full-band spectra, the ExtraTreesClassifier algorithm has the highest accuracy; the average accuracy on the training set and test set were 0.9925 and 0.9854, respectively. The classification results were visualized and analyzed using Python. The results highlight the effectiveness of HSI combined with machine learning algorithms in achieving nondestructive detection of sorghum origin within the same variety. This study provides a precise method for rapid and nondestructive determination of sorghum origin.