Liquid biopsy shows promise for early cancer detection due to its non-invasive nature and insights into tumor biology. Although Surface-enhanced Raman spectroscopy (SERS) offers high sensitivity and specificity for serum analysis, its practical application is often limited by spectral complexity, background fluorescence, and noise, which degrade classification performance and model interpretability. In this study, we integrate wavelet transform with convolutional neural networks (CNNs) to decompose spectra into multiscale frequency components. By effectively separating mid-frequency Raman features from low-frequency fluorescence and highfrequency noise, our proposed method achieves a high accuracy of 99.0 % (+/- 0.7 %) in 10-fold crossvalidation (10-CV). Combination of Gradient-weighted Class Activation Mapping (Grad-CAM) and spectral difference analysis further reveals key spectral intervals for the CNN decision process, providing insights into potential biomarkers. Overall, this paper demonstrates that combining the wavelet transform with CNNs not only improves the classification accuracy but also enhances the transparency of our spectral-based machine learning models.

The fundamental nature of the ionic thermoelectric (i-TE) effect-whether it represents genuine thermoelectricity or merely an analogy-remains unresolved. Existing devices exploit ion migration within the electrolyte but lack interfacial contributions, and no experimental evidence of the Peltier effect in i-TE materials has been reported. A central question thus arises: if such a Peltier effect exists, do the Seebeck, Peltier, and Thomson coefficients of i-TE systems satisfy the Kelvin relations? Here, we systematically formulate the three thermoelectric effects in the i-TE context and provide the experimental observation of an ionic Peltier effect, whose cooling power exceeds that of conventional electrolytes by more than 2 orders of magnitude. Furthermore, we introduce an electromotive-force-based method to determine thermoelectric coefficients without direct heat-flux measurements, thereby enabling experimental verification of the Kelvin relations. These results resolve fundamental questions of the i-TE effect and establish its position within thermoelectricity.

2D materials like graphene are renowned for their exceptional thermal and electrical properties, yet their performance can be significantly altered by structural irregularities such as wrinkles. While previous studies have reported modulation of thermal conductivity (κ) and electrical resistance (R) in wrinkled graphene, the results are often inconsistent or even contradictory, primarily due to challenges in experimentally disentangling geometric distortion from lattice strain. Here, a nearly zero-strain wrinkling strategy is introduced for bilayer graphene (BLG) and uncover a strikingly inverse anisotropic relationship: thermal conductivity perpendicular to the wrinkles (κ⊥) is lower than that parallel to the wrinkles (κ∥), whereas electrical resistance exhibits the opposite trend, with R⊥ lower than R∥, highlighting the decoupling of thermal and electrical transport in wrinkled graphene. Atomistic simulations reveal that this behavior arises from phonon mode hybridization induced by out-of-plane geometric perturbations, which decelerates heat-carrying phonon modes across the wrinkles and modifies electron-phonon scattering, thereby governing both the thermal conductivity and phonon-limited electrical resistance. This work advances the understanding of energy carrier transport in wrinkled 2D materials and provides new insights into directionally modulating heat and charge flow in advanced electronic devices.

The sprinkler serves as critical equipment for reducing steam temperature and pressure in the pressurizers of nuclear power plants, and the even distribution of spray droplets predominantly governs the heat transfer efficiency. This study involved the design and fabrication of four straight swirling sprinklers, accompanied by the establishment of a performance testing platform. Experimental results indicate that the pressure drop across these sprinklers is consistently related to jet velocity, exhibiting a power-law correlation with an exponent exceeding 2. Additionally, the Gaussian curve peak fitting technique was applied to model the radial distribution curve, effectively revealing the interactive synergistic influence of swirling and axial flows within the sprinkler. This approach has introduced a novel methodology for investigating the uniformity of sprinkler distribution. Further analysis of the scale-up effects associated with varying swirling channel area ratios revealed that an increase in this ratio diminishes the magnification effect.

The growing ability of pathogens and tumor cells to evade immune surveillance underscores the urgent need for new vaccine platforms that harness diverse biological mechanisms. Logistical constraints associated with cold-chain transport further limit vaccine accessibility, particularly in resource-limited settings. Migrasomes—specialized organelles produced during cell migration—are inherently stable and enriched with immune-modulating molecules. To overcome the low yield of natural migrasomes, we engineered migrasome-like vesicles (eMigrasomes) using hypotonic shock combined with cytoskeletal disruption to promote vesicle formation. eMigrasome biogenesis depends on core migrasome machinery and recapitulates the biophysical and molecular features of native migrasomes while achieving higher production efficiency. In murine models, eMigrasomes loaded with a model antigen elicited potent antibody responses and retained structural integrity and immunogenicity at room temperature. Moreover, eMigrasomes displaying the SARS-CoV-2 Spike protein induced strong humoral responses and conferred protection against viral challenge in mice. These results establish eMigrasomes as an innovative, thermally stable, and broadly applicable vaccine platform derived from migrasome biology.

Traditional Line-Commutated Converter HighVoltage Direct Current (LCC-HVDC) systems are prone to commutation failure due to grid faults or disturbances, which affects the safe and stable operation of the power grid. In contrast, IGCT-based HVDC can forcibly turn off the current to resist commutation failure. Based on the ADPSS simulation software, The Lingbao HVDC model is modified to establish an IGCT-based HVDC electromagnetic transient model. By setting single-phase and three-phase faults, the simulation waveforms before and after the modification are compared for fault analysis, verifying the commutation failure resistance and fault ride-through capability of the modified HVDC model. Finally, it is proposed that the control strategy of IGCT-based HVDC can be optimized in the future to explore control strategies that better adapt to grid characteristics.

Background: Rapid and reliable bacterial detection is very important for both disease diagnosis and bacterial research, but the speed of bacterial detection is greatly affected by the fact that traditional detection methods require bacterial culture first. Therefore, it is eager to develop a single bacterium detection method that do not require culture. Raman spectroscopy is a promising approach to identify single bacterium, but it still faces two challenges: how to precisely control individual bacterium, and how to improve signal-to-noise ratio in the Raman measurement. Results: To address these problems, we propose a Raman-microfluidic single bacterium identification and sorting method. A bi-directional flow controllable microfluidic system was designed to control bacteria movement, while a special S-shaped channel and a large laser spot ensured that the Raman spectrum of a single bacteria can be detected stably and with high quality on the move. The accuracy and repeatability of this detection system was verified, and a preliminary identification and sorting of bacterial drug resistance can be realized. In addition, Raman spectra of dozens of bacteria were obtained through this detection system, and based on these Raman spectra, multiple bacteria were identified. Their compositional and structural characteristics were further analyzed by the Raman features, which provided a reliable data base for future bacterial studies. Significance and novelty: We built a Raman-Microfluidic system that can realize high-quality single bacterium identification and sorting. The Raman spectrum of a single bacterium can be detected stably and with high quality on the move. And a Raman spectral database of several bacterial species was preliminary established, which laid the foundation for the rapid bacterial identification. The establishment of our systems and databases is of great significance for rapid diagnosis of diseases and dynamic research on drug resistance.

Currently, Wi-Fi-based indoor localization methods have been proven to be promising due to their low deployment cost. However, the overhead of constructing and maintaining maps remains a bottleneck for the widespread deployment of Wi-Fi-based indoor localization methods. In this article, we propose a novel combined Wi-Fi and vision to construct and maintain maps. This method consists of three parts (including constructing the logarithmic distance path loss (LDPL) model using an improved whale optimization algorithm (IWOA), a novel fusion localization module (called LDPL-PF), and a lightweight threshold-based map maintenance model). Specifically, the LDPL model based on IWOA can first construct high-quality maps with limited data. Then, LDPL-PF localization method is used to determine the user’s location based on the map. Finally, a feedback mechanism is introduced to achieve map maintenance automatically. The localization results and the collected data are fed into a multidecision mechanism to build a feedback network for long-term map maintenance. Extensive experimental results show that our proposed method has good accuracy and stability with state-of-the-art methods.

The huge thermopower observed in the thermodiffusion mechanism of ionic thermoelectric (i-TE) materials has led researchers to conceive of the upgrading of thermoelectric technology. However, the intermittent power generation in the capacitor mode has been a major hindrance to achieving optimal performance. This work proposes a conveyor mode for continuous i-TE conversion in mixed ion-electron-conducting i-TE material with an ionic circuit. In this conveyor mode, ion-electronic friction serves as the link between ions and electrons, enabling the thermally diffused ions to convey electrons to power the load, and persistent ionic transport, owing to the ionic circuit, ensures continuous power generation. Experiments demonstrate continuous power generation and significant improvements of power density in the conveyor mode. Theoretical analysis shows that the conveyor mode is competitive to not only the capacitor mode but also a general electronic thermoelectric conversion. Our study points out a direction for the development of i-TE technology.

Surface-enhanced Raman spectroscopy (SERS) has been used in Raman-based metabolomics to provide abundant molecular fingerprint information in situ with extremely high sensitivity, without damaging the sample. However, poor reproducibility, caused by the randomness of the adsorption sites, and the short-range effect of SERS have hindered the development of SERS in metabolomics, resulting in very few SERS reference databases for small-molecule metabolites. In this work, our previously proposed large laser spot-swift mapping SERS method was adopted for the measurement of 24 commercially available metabolite standards, to provide reproducible and reliable references for Raman-based metabolomics study. Among these 24 metabolites, 22 contained no Raman data in PubChem. Other than the SERS spectra data, we extracted and explained the molecular vibration information of these metabolites, and combined with the density functional theory (DFT) calculations, we provided a new possibility for the fast Raman recognition of small-molecule metabolites. Accordingly, a large laser spot-swift mapping SERS database of metabolites in human serum was initially established, which contained not only the original spectral data but also other detailed feature information regarding the Raman peaks. With continuous accumulation, this database could play a promising role in Raman-based metabolomics and other Raman-related research.

Liquid biopsy offers promise for the diagnosis of malignant tumors or their precancerous lesions at the subclinical stage, which is crucial for improving the survival rate of cancer. Label-free surface-enhanced Raman spectroscopy (SERS) has become an emerging detection method in liquid biopsy. The accuracy of SERS-based diagnosis relies on both precise measurement and effective data analysis. In our previous study, a large laser spot SERS method was proposed for the precise detection of complex liquids. In this work, a logical Raman data analysis method was developed, to address the lack of Raman characteristics and biological significance in current SERS-based diagnosis data analysis. The measured Raman data is classified to obtain a weighting matrix of classification criteria, and this matrix is used to establish a material correspondence with the Raman spectra, extract key Raman peaks, determine corresponding biomolecules and biological processes, thus enabling a logical analysis of the data. This method has been used to analyze the SERS data of lung cancer, it can not only show exceptional performance in diagnosing lung cancer and differentiating small cell lung cancer, but also obtain the Raman diagnostic and classification criteria with clear numerical values and reasonable biological significance, demonstrating its great potential value in precise biosensing.

With the development of power devices towards high integration and power, the electrical and thermal stresses per unit area become more concentrated and intensified. The impact of cross-plane strain resulting from inverse piezoelectric effect and thermal expansion on power devices cannot be ignored. Cross-plane strain has a substantial influence on the thermal properties of GaN. However, the research on the influence of cross-plane strain on the thermal conductivity of GaN has not been reported in the literature. Based on the first-principles calculation method and the phonon Boltzmann transport equation, the influence of cross-plane strain on the thermal conductivity and phonon characteristics of the GaN lattice is systematically studied in this study. The thermal conductivity of GaN has anisotropy and increases significantly with the decrease in temperature. The thermal conductivity of GaN at room temperature under the free state is calculated to be 257 and 275 W/(mK) for in-plane (k(perpendicular to)) and cross-plane (k(parallel to)) directions, respectively. Compared with previous theoretical reports, our calculation results are more consistent with the existing experiment values. Under the state of cross-plane strain, the lattice thermal conductivity changes remarkably. In detail, the average thermal conductivity at room temperature decreased by 35 % under a 5 % cross-plane tensile strain state, while it increased by 11.5 % under a 5 % cross-plane compressive strain state. According to the calculation results, the influence of cross-plane on lattice thermal conductivity is mainly due to the change in phonon lifetime. The analysis of two mechanisms of phonon lifetime suppression indicates that the cross-plane strain will significantly change the frequency of the high-frequency optical acoustic branch. This result leads to a change in the phonon scattering process and thus affects the phonon lifetime. Besides, the anisotropy of thermal conductivity changes under different strain values, which may be due to the weakened piezoelectric polarization effect induced by strain.

The accuracy of existing ultra-wideband (UWB) range-based indoor localization methods is generally degraded due to the non-line-of-sight (NLOS) situations where a serve bias in UWB range measurements is unavoidable. In this article, we first propose a two-stage NLOS detection method to detect line-of-sight (LOS)-measured distances in mixed LOS/NLOS indoor environments. Then, a high-accuracy UWB/IMU/Odometer integrated localization system is presented using an adaptive multi-algorithm localization framework based on the number of detected LOS-measured distances. Specifically, under conditions of one or two LOS-measured distances, an improved adaptive EKF positioning algorithm (IAEKF) is proposed. Compared with the traditional extended Kalman filter (EKF)-based fusion scheme, the weight function of innovation is exploited to adaptively estimate the measurement noise covariance matrices and further reduce the influence of the changing measurement noise in NLOS conditions. For three or more LOS-measured ranges, a novel tightly coupled fusion factor graph framework is developed. To further improve the performance of seamless positioning in transaction areas, a strong constraint of trajectory smoothness is designed and added to the factor graph framework using the weight value of IMU/odometer measurements. The experimental results show that the proposed localization system achieves an average localization error of 0.227 m, which surmounts UWB range-based and integrated methods in LOS/NLOS mixed environments.

AbstractThermoelectrics converting heat and electricity directly attract broad attentions. To enhance the thermoelectric figure of merit, zT, one of the key points is to decouple the carrier-phonon transport. Here, we propose an entropy engineering strategy to realize the carrier-phonon decoupling in the typical SrTiO3-based perovskite thermoelectrics. By high-entropy design, the lattice thermal conductivity could be reduced nearly to the amorphous limit, 1.25 W m−1 K−1. Simultaneously, entropy engineering can tune the Ti displacement, improving the weighted mobility to 65 cm2 V−1 s−1. Such carrier-phonon decoupling behaviors enable the greatly enhanced μW/κL of ~5.2 × 103 cm3 K J−1 V−1. The measured maximum zT of 0.24 at 488 K and the estimated zT of ~0.8 at 1173 K in (Sr0.2Ba0.2Ca0.2Pb0.2La0.2)TiO3 film are among the best of n-type thermoelectric oxides. These results reveal that the entropy engineering may be a promising strategy to decouple the carrier-phonon transport and achieve higher zT in thermoelectrics.

As a potential alternative to fossil fuel, hydrogen has attracted much attention due to its renewable and environmentally friendly properties. Systems built for hydrogen-production and hydrogen-application tend to be larger, more integrated and more complex. In order to more efficiently design the vital components of hydrogen energy systems, accurate estimations of the thermodynamic and thermophysical properties of hydrogen-containing mixtures involved is essential. In this study, we introduced methods typically for calculating the thermodynamic and thermophysical properties of H2/CO2/CO/CH4/H2O mixtures, and established the technical database covering a wide range of mole fractions, pressures and temperatures. Moreover, a user-friendly software integrating all the calculation methods called H2MixThermoDatabase is compiled, whose code has been made available on the GitHub page for other researchers to effectually facilitate the further developments of hydrogen energy system.

A storage-based distributed fixed-time frequency synchronization method is developed to enhance the frequency and transient stability of smart grid. The multi-agent cyber-physical model of power system is the foundation of the proposed control method which utilizes data on the relative angle and frequency between the local agent and its neighbor agents. The control method mainly consists of frequency control based on energy storage system (ESS), P- $\omega $ droop control of synchronous generator (SG), and P- $\theta $ droop control of converter-based generator (CBG), which is designed by the backstepping method and Lyapunov theory. In addition, to hasten frequency synchronization, the nonlinear voltage control which is compatible with frequency control is presented. Meanwhile, the control method is not only applied to the homogeneous power systems only include SGs, but also deals with the challenge of heterogeneous bus dynamics introduced by the coexistence of SGs and CBGs. The distributed fixed-time frequency control can achieve self-protecting from denial-of-service (DoS) attacks by transforming into decentralized control. Comparative simulations show that the designed storage-based frequency synchronization method is preferable in improving the transient stability and resilience of smart grid.

In view of the need for stability analysis of large- scale new energy cluster connecting to large-scale power grid system, the existing electromagnetic transient simulation tools have certain limitations. At present, there is a lack of simulation platform that can uniformly establish large-scale new energy and large-scale power grid models. Based on two different simulation tools, RTLAB and CloudPSS, a large-scale new energy cluster and large-scale power grid electromagnetic transient co- simulation platform is built. A new energy station equivalence method based on single machine representation method is proposed, and the accuracy of the equivalence method is verified by comparing with detailed station model, which provides a basis for the modeling of large-scale new energy clusters. On the other hand, the electromechanical - electromagnetic model conversion tool based on CloudPSS is introduced, which provides a method for large-scale power grid modeling. At last, the decoupling principle and hardware implementation scheme of co-simulation model based on RTLAB and cloudPSS are presented and the comparison of simulation results verifies the accuracy of the simulation platform.

Active neutral-point-clamped (ANPC) topology three-level converter is widely used in medium- high voltage new energy power stations. Aiming at the shortcomings of existing electromagnetic transient modeling methods in model accuracy, computational efficiency and simulation scale in small step real-time simulation, an efficient electromagnetic transient simulation model was proposed. Based on the associated discrete circuit of the ANPC three-level, the model constructed a multi-port Norton equivalent circuit to replace the bridge arm for the electromagnetic transient simulation of the network. This model achieved node dimension reduction and simulation acceleration without loss of modeling accuracy. An ANPC three-level converter simulation system based on the equivalent model was built in MATLAB / Simulink platform. Compared with the detailed model simulation results, the equivalent model can accurately simulate the steady-state and transient operating conditions, and has obvious simulation speed-up performance.

In this paper, oriented to the fault types in the electromagnetic transient simulation of grid operation mode calculation, the program analysis method of the fault file that describes the fault is studied, and the data structure used to describe the electromagnetic transient fault in the program is designed. For complex faults that consider the randomness of the fault location of the fault element and the action of the circuit breaker, research the electromagnetic network topology transformation method covering chain operations such as node splitting, component splitting, switching action, and fault setting. Research and develop electromagnetic transient simulation fault handling programs that support faults of bus, line, transformer, and converter, etc. Finally the correctness of the program is verified.