High Performance Research Articles

High-performance direct borohydride-hydrogen peroxide fuel cell using Ni foam-supported Ni-Co nanoparticles as a noble-metal free anode electrocatalyst

The Ni-Co nanoparticles with various Ni/Co ratios are immobilized on a Ni foam (NF) template, coated with reduced graphene oxides (rGO), and used as a novel noble-metal-free anode electrocatalyst for enhanced performance direct borohydride fuel cells. The electrochemical active surface area (EASA) of Ni50Co50 /rGO/NF (1515 cm2) catalyst is 45.5 times larger than NF (33.3 cm2). A comprehensive study of direct borohydride-hydrogen peroxide fuel cell (DBHPFC) by Pt/C (0.5 mg cm−2) as a cathode and Ni50-Co50/rGO/NF as an anode is accomplished, so an open-circuit potential (OCP) of 1.90 V and the maximum power density of 309 mW cm−2 is obtained at 60 ◦C. These results show that electrocatalyst Ni50Co50/rGO/NF is a suitable candidate for use as an electrocatalyst in DBHPFC due to its low cost, ease of synthesis, excellent structural stability, and high catalytic performance.

Unraveling the mechanistic interplay between CO and CO2 hydrogenation over Ni, Co, and NiCo catalysts

Although CO and CO2 hydrogenation reactions have been extensively studied, there is still significant controversy regarding the mechanism of methane formation and the routes for byproducts, especially when both carbon sources are simultaneously present. This work combines kinetic, operando-spectroscopic, isotopic techniques and DFT calculations to elucidate the relationships between CO and CO2 hydrogenation over mono Ni, Co and bimetallic NiCo catalysts with similar metal dispersion. The bimetallic catalysts showed a slight synergistic effect for methane formation from CO, while from CO2 the effect was opposite, showing a negative shift of activity as compared with those observed on the monometallic catalysts. The kinetic analysis shows similar apparent order with respect to H2 (∼0.5) and an inverse secondary isotope kinetic effect (<1) for both methanation reactions, suggesting that CH4 formation proceeds via C-O bond breaking in an H*-assisted mechanism, where the rate-determining step was not shown to be sensitive to the carbon source. The anti-synergistic effect observed on the bimetallic catalysts during CO2 hydrogenation is explained by the formation of unreactive HCOO* species (spectator), which generate a lower density of methane intermediates. On the other hand, the formation of the undesired products, i.e., CO or CO2 during CO2 or CO hydrogenation, respectively, shows relevant differences since the CO formation during CO2 hydrogenation proceeds through the desorption of the carbonyl species adsorbed on weak sites, meanwhile the CO2 formation from CO hydrogenation is due to a minority route of direct dissociation of CO, allowing the rejection of O* with CO* to produce CO2. This study contributes to elucidate the routes that determine the selectivity and reactivity for CO and CO2 hydrogenation and their mechanistic relationship. This information is valuable for the rational design and development of new materials with high performance during hydrogenation processes.

Hybrid deep learning for design of nanophotonic quantum emitter lenses

Inverse design of nanophotonic structures has allowed unprecedented control over light. These design processes however are accompanied with challenges, such as their high sensitivity to initial conditions, computational expense, and complexity in integrating multiple design constraints. Machine learning approaches, however, show complementary strengths, allowing huge sample sets to be generated nearly instantaneously, and with transfer learning, allowing modifications in design parameters to be integrated with limited retraining. Herein we investigate a hybrid deep learning approach, leveraging the accuracy and performance of adjoint-based topology optimization to produce a high-quality training set for a convolutional generative network. We specifically explore this in the context of 3D nanophotonic lenses, used for focusing light between plane-waves and single-point, single-wavelength sources such as quantum emitters. We demonstrate that this combined approach allows higher performance than adjoint optimization alone when additional design constraints are applied; can generate large datasets (which further allows faster iterative training to be performed); and can utilize transfer learning to be retrained on new design parameters with very few new training samples. This process can be used for general nanophotonic design, and is particularly beneficial when a range of design parameters and constraints would need to be applied.

FDIA localization and classification detection in smart grids using multi-modal data and deep learning technique

False data injection attacks (FDIA) exploit the vulnerabilities of bad data detection in energy management systems to maliciously tamper with state estimation results, seriously jeopardizing the safe and reliable operation of power systems. In order to promptly detect FDIA, recent studies have used machine learning techniques to extract attack characteristics and detect FDIA based on changes in these characteristics. Current research predominantly focuses on detecting a single type of FDIA attack model and topology. However, the increasing diversification of FDIA construction methods and the variability of topological structures in power systems can reduce or even invalidate detection effectiveness. To cope with the abovementioned difficulties, this study introduces a multimodal deep learning detection model based on variational graph auto-encoders (VGAE), temporal convolutional networks (TCN), and gated recurrent units (GRU). The topological features of power systems are obtained through VGAE to adapt to the detection needs of various topological structures and performance enhancement. These features are then integrated with preprocessed measurement data via BDD to form multimodal data. Furthermore, this multimodal data is used to locate and classify FDIA with complete information, FDIA with incomplete information, and topology attack after applying a multi-label classification algorithm based on TCN-GRU for temporal feature extraction. Experiments carried out on the IEEE 14 and IEEE 118 bus systems show that the proposed method is robust and has high detection performance. The results indicate that the proposed detection method achieves AUC values that are, on average, 0.085, 0.145, and 0.04 higher than those of CNN, LSTM, and CNN-LSTM, respectively. Moreover, under various noise and attack intensities, recall, precision, F1 score, and RACC remain above 0.859, 0.877, 0.867, and 0.818, respectively, with a classification accuracy greater than 0.912. This study provides a unique perspective on detecting FDIA across various attack models and power system topologies.

SiC GAA MOSFET Concept for High Power Electronics Performance Evaluation through Advanced TCAD Simulations

In this work a groundbreaking SiC power MOSFET based on innovative vertical Gate All Around (GAA) concept is presented. Extensive TCAD simulations are performed to analyze the performance in forward as well as in reverse conditions for this new device concept. The proposed design has a target rating voltage of 1200V, suitable for e-mobility applications.

Promoting active hydrogen supply for kinetically matched tandem electrocatalytic nitrate reduction to ammonia

Electrocatalytic nitrate reduction (NO3RR) shows admirable potential for environmental remediation and producing valuable NH3. However, the catalyst is restricted by the kinetic mismatch between NO3--to-NO2- and NO2--to-NH3, which results in excess NO2- and poor NH3 selectivity. Herein, a kinetically matched tandem electrocatalytic strategy with tunable active hydrogen(*H) supply is designed. The combination of Cu2O and Co11(HPO3)8(OH)6 (CoPO) playing key roles in NO3RR by i) on-demand supply of *H from CoPO; ii) the electronically coupled interface between Cu2O and CoPO enhance *H transfer kinetics. The Cu2O-CoPO-2 exhibits high NH3 yield of 22 mg cm−2 h−1 with Faradaic efficiency of 95 % at −0.37 V vs. RHE. In situ characterizations indicate the dynamic equilibrium between *H production and consumption contributes to high NO3RR performance. The techno-economic analyses reveal the system is economically viable compared to Haber-Bosch (H-B) process, which benefits from industrial current densities and superior energy efficiency at low potentials.

Diagnostic accuracy of non-invasive cardiac imaging modalities in patients with a history of coronary artery disease: a meta-analysis

BackgroundThe diagnostic performance of non-invasive imaging techniques for detecting obstructive coronary artery disease (CAD) in patients with a history of myocardial infarction or percutaneous coronary intervention has not been comprehensively...

Synergic effects of silica aerogel (SA) particles on tensile behavior of cryo-conditioned epoxy: The role of particle morphology and mixing sequence

Cryogenic conditions are of crucial importance for gas storage tanks to provide increased storage density, reduced pressure requirements and minimized energy losses. Incorporation of silica aerogel (SA) particles as a nanostructured material with extremely low density and exceptional thermal insulating properties into epoxy matrices has the potential to facilitate the progress of high performance nanocomposite coatings for advanced cryogenic applications. The objective of this study is therefore to investigate the impacts of particle size, particle content, milling method and mixing sequence on toughness of the cryo-conditioned epoxy coatings. SA particles of varying size (50 μm to 300 μm) were prepared using two different techniques (planetary ball milling vs. 2-blade grinding). Nanocomposite samples were prepared using two distinct mixing sequences, i.e., adding particles (in 0, 0.5, 1 and 1.5%wt.) to epoxy followed by mixing with hardener or adding particles to epoxy/hardener (with 3 or 15 min delay). The nanocomposite samples were subjected to different cryogenic conditions (single 3-h immersion in liquid nitrogen vs. three consecutive 1-h immersions). Mechanical properties of the samples were evaluated by conducting fractography, dynamic mechanical thermal analysis (DMTA) and tensile tests. Large SA particles settled towards the bottom of nanocomposite and resulted in loss of mechanical properties at cryo-conditions. Mixing 1 wt% of optimized SA particles with epoxy/hardener system led to an enhanced tensile strength (26 %), stiffness (3 %) and toughness (71 %) by providing uniform cross-linking reactions and mitigating thermal shock effects at cryo-conditions. Toughening mechanisms included crack deflection, crack pinning, crack arrest and crack branching.

A new hybrid fixed and adaptive gains control algorithm to reduce power losses on LLCL filter-based renewable energy conversion systems with systematic parametrization using Grey Wolf optimizer

Currently, there is a transition in the energy matrix around the world, where traditional sources of energy generation are continually being replaced by energy generation systems based on renewable sources to mitigate the climate crisis. In this bias, this work presents the mathematical modeling of an LLCL filter, used to connect power generation systems based on renewable energy sources to the electrical grid, and presents a novel hybrid fixed-and adaptive gains control strategy for current injection into the grid using this system. The developed hybrid controller is composed of a proportional–integral controller and a direct robust adaptive controller. The first term of the controller guarantees the reference current, while the second term of the controller is used for disturbance rejection. Furthermore, a systematic procedure for the controller’s parametrization based on Grey Wolf Optimizer is also provided. The control of the current injected into the grid is carried out considering the LLCL filter without passive damping resistors in the filter structure to avoid power losses due to the passive filter elements. Additionally, the LLCL filter model considers minimal parasitic resistances to evaluate the controller’s performance and optimize it for the application of interest, aiming to maximize the system performance by ensuring a short transient regime due to the fast closed-loop system response. Simulation results indicate high performance of this optimized control strategy with small tracking error even considering grid impedance variations.

Evaluating the Geoeffectiveness of Interplanetary Coronal Mass Ejections: Insights from a Support Vector Machine Approach with SHAP Value Analysis

In this study, we compiled a data set of 510 interplanetary coronal mass ejections (ICME) events from 1996–2023 and trained a radial basis function support vector machine (RBF-SVM) model to investigate the geoeffectiveness of ICMEs and its dependence on the solar wind conditions observed at 1 au. The model demonstrates high performance in classifying geomagnetic storm intensities at specific Disturbance Storm Time thresholds and evaluating the geoeffectiveness of ICMEs. The model’s output was assessed using precision, recall, F1 score, and true skill statistics (TSS), complemented by stratified k-folds cross-validation for robustness. At the −50 nT threshold, the model achieves precisions of 0.84 and 0.93, recalls of 0.94 and 0.82, and corresponding F1 scores of 0.89 and 0.87 for the categories separated by this threshold, respectively. Overall accuracy is noted at 0.88, with a TSS of 0.76. Despite challenges at the −100 nT threshold due to data set imbalance and limited samples, the model maintains an overall accuracy of 0.87, with a TSS of 0.69, demonstrating the model’s ability to effectively handle imbalanced data. Physical insights were gained through model explanation with a SHapley Additive exPlanations (SHAP) value analysis, pinpointing the role of the southward magnetic field component in triggering geomagnetic storms, as well as the critical impact of shock-ICME combinations in intensifying these storms. The effective application of an SVM model with SHAP value analysis offers a way to understand and predict the geoeffectiveness of ICMEs. It also underscores the capability of a relatively simple machine learning model in predicting space weather and revealing the underlying physical mechanisms.

Honeycomb-like macroporous crosslinked chitosan assisted EDTA-intercalated Ca-Mg-Al layered hydrotalcite composite foams for efficient U(VI) biosorption

The biosorption is considered to be highly efficient for the separation of radionuclide from radioactive wastewater. Herein, the crosslinked chitosan assisted EDTA intercalated Ca-Mg-Al layered double hydroxides composite foam (CS-EDTA-LDH) was synthesized by combining EDTA intercalation and freeze-drying methods. The macroporous and ultralight properties of CS-EDTA-LDH facilitates its rapid adsorption and facile recovery, and the inorganic/organic incorporation can avoid pore collapse and provide numerous adsorption sites, while the EDTA intercalation can enhance the complex capture of U(VI). The CS-EDTA-LDH presents various functional groups (carboxyl, hydroxyl and amino groups) for U(VI) adsorption, and the adsorption capacity for U(VI) reached 272.3 mg/g at pH 5.0 and 298 K. The adsorption kinetics of U(VI) conformed to PSO equation, whereas the isotherms conformed to the Freundlich model, indicating heterogeneous adsorption with diffusion process as a rate-controlling step. The thermodynamic parameters indicate that U(VI) adsorption by CS-EDTA-LDH is endothermic and spontaneous in nature. The adsorption mechanism is related to the synergic complexation by multi-functional groups, ion exchange, and possible isomeric substitution. Overall, CS-EDTA-LDH could be a promising biosorbent for the cleanup of radioactive pollution due to its high performance for U(VI) adsorption and facile recovery.

Effects of coarse aggregates on 3D printability and mechanical properties of ultra high performance fiber reinforced concrete

3D printing of ultra high performance fiber reinforced concrete (UHPFRC) suffers from high shrinkage and poor interlayer properties. To mitigate these problems, this study develops new mixtures of UHPFRC materials containing coarse aggregates (CA), and critically examines their suitability for 3D printing (3DP). Totally 90 mould-cast and 3DP cylinder and beam specimens with different CA sizes (5–15 mm) and CA-binder ratios (0.3–0.5) were tested under compression and bending in three directions. The internal distribution and volume fractions of steel fibers and pores and the crack trajectories were characterized and analysed by micro X-ray CT scanned 3D images with 37 μm voxel resolution. The results show that adding more and bigger CAs into UHPFRC reduced the flowability but enhanced the buildability with desired extrudability achieved by adjusting 3D printing velocity. Although the compressive strength of the 3DP cylinders was 15–42 % lower than that of the mould-cast ones due to higher porosities, over 100 MPa strength was still achieved for all the 3DP cylinders with less than 10 % anisotropy. The CT images did not show evident interlayers and interlayer delamination under compression, indicating the new 3DP mixtures and printing parameters may have highly promoted cement hydration. The flexural strength of 3DP beams was 3–34 % higher than that of the cast ones, because most fibres were oriented in the printing or beam axis direction as represented by overall orientation indices calculated from CT images, thereby providing significant crack-bridging effects. The developed new mixtures are therefore well suited for 3D printing, particularly for fabrication of structural members with preferred fibre orientation, such as beams, slabs and shells.

Accelerometers-based position and time interval comparisons for predicting the behaviors of young bulls housed in a feedlot system

Animal behavior monitoring is an important tool for animal production. This behavior monitoring strategy can indicate the well-being and health of animals, which can lead to better productive performance. This study aimed to assess the most effective accelerometer attachment position (on either the halter or a neck collar) and data transmission time intervals (ranging from 6 to 600 s) for predicting behavioral patterns, including water and food intake frequencies, as well as other activities in young beef cattle bulls within a feedlot system. A range of machine learning algorithms were applied to satisfy the aims of the study, including the random forest, support vector machine, multilayer perceptron, and naive Bayes classifier algorithms. All studied models produced high performance metrics (above 0.90) when using both attachment positions, except for the models built using the naive Bayes classifier. Therefore, coupling accelerometers with collars is a more viable alternative for use on animals, as doing so is easier than applying accelerometers to halters. Utilizing a dataset with more observations (i.e., shorter time intervals) did not result in considerable improvements in the performance metrics of the trained models. Therefore, using datasets with fewer observations is more advantageous, as it can lead to decreased computational and temporal demands for model training, in addition to saving the battery of the device considered in this study.

Cognitive Intraindividual variability in injection drug use among Hispanic residents of Puerto Rico

Despite a high prevalence of injection drug use (IDU) in Puerto Rico, little is known about how it affects neuropsychological functioning in this population. Investigations of intra-individual variability (IIV) have alluded to its utility as a potential indicator of neural decline. The purpose of this study was to characterize IIV among Hispanic residents of Puerto Rico who engage in IDU. Injectors and non-injectors completed the Neuropsí Atención y Memoria battery. Measures of IIV were calculated for the overall test battery (OTB), the three battery indices, and three domains of attention, memory, and executive functioning. The injector group showed significantly greater IIV than the non-injector group on all measures (OTB, indices, and individual domains). Additionally, injectors showed significantly higher IIV in the domain of executive functioning compared to other cognitive domains and battery indices. In contrast, non-injectors did not show any significant within-group differences on any IIV measures. The higher performance variability observed in the IDU group suggests a negative influence of IDU on cognition, with executive functioning being more susceptible to these effects. These findings support the need for continued investigations into the clinical application of IIV for diagnostic and prognostic purposes in the Hispanic IDU population.

Efficient electrochemical removal of ammoniacal nitrogen from livestock wastewater: The role of the electrode material

Wastewater from livestock farms contains high concentrations of suspended solids, organic contaminants, and nitrogen compounds, such as ammoniacal nitrogen. Discharging livestock effluents into water bodies without appropriate treatment leads to severe environmental pollution. Compared to conventional treatment methods, electrochemical oxidation exhibits higher nitrogen removal efficiencies. In the present work, the electrochemical removal of ammoniacal nitrogen from real livestock wastewater was investigated through a lab-scale reactor. Preliminary experiments were carried out to investigate the effects of different anode materials, including boron-doped diamond and iridium/ruthenium-coated titanium, on the total nitrogen removal efficiency using synthetic wastewater. Boron-doped diamond, a well-known non-active electrode, allowed to obtain 63.7 ± 1.21 % of total nitrogen degradation efficiency. However, the iridium/ruthenium-coated titanium electrode, belonging to the class of active anodes, showed a higher performance, achieving 78.8 ± 0.76 % contaminant degradation. Coupling iridium/ruthenium-coated titanium anode with a stainless-steel cathode improved the performance of the system, achieving even 96.2 ± 2.73 % of total nitrogen removal. The optimized cell configuration was used to treat livestock wastewater, resulting in the degradation of 67.0 ± 2.25 % of total nitrogen and 37.3 ± 0.68 % of total organic carbon when sodium chloride was added. At the end of the process, the ammonium content was completely removed, and only 17.7 ± 0.51 % of the initial nitrogen turned into nitrate. The results show that the proposed system is a promising approach to treating livestock wastewater by coupling high contaminant removal efficiencies with low operational costs. Anyway, further studies on process optimization with an emphasis on power requirements and electrode costs need to be carried out.

High performance and stability of perovskite solar cells achieved through anti-solvent modulation of multi-orientation PbI2 and stress relief

The morphology of the pre-deposited PbI2 layer plays a crucial role in determining the quality of perovskite films in sequentially deposited perovskite solar cells. The main reason is that the PbI2 film is too dense, which makes it difficult for organic salts to fully enter and react with them. This work presents a novel solvent extraction post-treatment method for the construction of bilayer PbI2 films. The PCE of the prepared champion device reaches 23.8 %, which is higher than 21.6 % of the control device. We observed that the perovskite solar cells treated with EA retained over 90 % of their initial efficiency after being stored for approximately 700 h in a nitrogen atmosphere, and over 80 % in air, demonstrating superior stability compared to the control group.

Green and sustainable strategies to control scaling in industrial plants: investigation of the efficacy of Rosmarinus officinalis L. Extract against CaCO3 scale using experimental and theoretical approaches

ABSTRACT In recent years, plant extracts have attracted increased interest as green alternatives to conventional anti-scaling. This is because they contain a wide range of bioactive compounds with high performance against inorganic scale. Additionally, they are biodegradable and pose minimal risks to human health and ecosystems. The present study aimed to assess the protection offered by the Rosmarinus officinalis L. leaf extract for industrial plant surfaces against the CaCO3 scale. Before assessing the anti-scaling performance of the Rosmarinus extract, phytochemical characterisation was performed by quantitative assays and HPLC-DAD analysis. Subsequently, the inhibition potential of the extract was studied using the conductivity and LCEE tests at 25°C and TH = 40°f. In addition, SEM and XRD analysis were used to assess the effect of the extract on scale morphology and crystalline phases. Finally, DFT calculations and Monte Carlo simulation were carried out to enhance knowledge of the interaction between inhibitor molecules and CaCO3(104) and (110) surfaces and optimise [extract molecule – Ca]2+ complexes. Phytochemical analysis revealed the presence of several phenolic compounds (rosmarinic acid, vanillic acid, cinnamic acid, rutin, kaempferol, trans chalcone and quercetin). Further LCEE studies demonstrated the promising anti-scaling activity of the extract at an effective concentration of 54 mg/L. SEM micrographs and XRD diffractograms revealed a significant change in the morphology and phases of precipitated CaCO3 scales upon the addition of the inhibitor. In addition, the computational approach strongly supported the experimental results. These results underlined the Rosmarinus extract’s potential as a valuable green and sustainable scaling inhibitor source.

Study on the properties and components of solid-liquid products by co-pyrolysis of sludge and cotton stalk

The effects of sludge content and pyrolysis temperature on the pyrolysis process and product properties were studied through the co-pyrolysis of sludge and cotton stalk. Multi-objective fuzzy hierarchical analysis was used to analyze the multiphase properties of pyrolysis products, and describe the internal relationship between pyrolysis conditions and product properties. It has been proved using kinetic calculation that the mass ratio of sludge and cotton stalk is higher than 1:2 required the lower pyrolysis activation energy. According to the comparison between experimental and theoretical results, the addition of sludge is beneficial to increase the bio-oil yield. In terms of product performance, the addition of sludge and increase of pyrolysis temperature both can improve the specific surface area, pore volume of biochar and the pH value of bio-oil, reduce the total acid content, increase the total phenol content and optimize calorific value of bio-oil. The specific surface area of biochar, calorific value of bio-oil, total phenol content and pH were evaluated by multi-objective decision analysis method, the optimum process condition of sludge: cotton stalk = 2:1 at 600 ℃. Under the conditions, the specific surface area of biochar is 57.40 m2/g, the pH of the bio-oil is 8.06, the total acid content is 5.94 mg/g, the calorific value is 27.18 KJ/g and the total phenol content is 47.32 mg/g, respectively. Compared to the pyrolysis product of cotton stalks, the specific surface area is increased almost 20 times, the calorific value is increased by 13 %, the total acid content is decreased by 95 %, and the total phenol content is similar. This study proposed a preparation process of bio-oil with high performance, and provided a theoretical basis for the subsequent co-pyrolysis research.

Enhanced adsorption of Cr(VI) pollutants using CaFe2O4@rGO modified-biochar derived from mixture of black soldier flier exuviae and durian peel

To solve the problem of hexavalent chromium (Cr(VI))-polluted water sources, mixed black soldier flier exuviae (BSFE) and durian peel (DP)-derived biochar modified with CaFe2O4@rGO (BC-CFr) was synthesized to remove Cr(VI) from wastewater. Our findings indicated that the adsorption of Cr(VI) on BC-CFr reached a maximum at a pH of 2.0, biochar dosage of 0.3 g/100 mL, and initial Cr(VI) concentration of 50 mg/L, with a Cr(VI) adsorption capacity of 30.8 mg/g, which was approximately two times greater than that on pristine biochar (PBC). The characterization results showed that the as-prepared modified biochar was rich in surface functional groups, including amine groups (−NH2) and active sites. Cr(VI) adsorption followed the Elovich kinetic and Freundlich isotherm models, illustrating the adsorption process involved multilayer on the heterogeneous surface. The thermodynamic study confirmed the adsorption process was endothermic. The proposed adsorption mechanisms included adsorption-coupled reduction, electrostatic attraction, and pore filling. This study provided a simple method to prepare compositing biochar using abundantly available agricultural by-products that dealt with the dual problems of Cr(VI)-contaminated water and agricultural solid waste. Besides, the BC-CFr exhibited high desorption performance and recyclability over five consecutive adsorption-desorption cycles, suggesting the BC-CFr is a feasible nanocomposite biochar for effectively removing Cr(VI) from wastewater in practice.

SFT-SGAT: A semi-supervised fine-tuning self-supervised graph attention network for emotion recognition and consciousness detection

Emotional recognition is highly important in the field of brain-computer interfaces (BCIs). However, due to the individual variability in electroencephalogram (EEG) signals and the challenges in obtaining accurate emotional labels, traditional methods have shown poor performance in cross-subject emotion recognition. In this study, we propose a cross-subject EEG emotion recognition method based on a semi-supervised fine-tuning self-supervised graph attention network (SFT-SGAT). First, we model multi-channel EEG signals by constructing a graph structure that dynamically captures the spatiotemporal topological features of EEG signals. Second, we employ a self-supervised graph attention neural network to facilitate model training, mitigating the impact of signal noise on the model. Finally, a semi-supervised approach is used to fine-tune the model, enhancing its generalization ability in cross-subject classification. By combining supervised and unsupervised learning techniques, the SFT-SGAT maximizes the utility of limited labeled data in EEG emotion recognition tasks, thereby enhancing the model’s performance. Experiments based on leave-one-subject-out cross-validation demonstrate that SFT-SGAT achieves state-of-the-art cross-subject emotion recognition performance on the SEED and SEED-IV datasets, with accuracies of 92.04% and 82.76%, respectively. Furthermore, experiments conducted on a self-collected dataset comprising ten healthy subjects and eight patients with disorders of consciousness (DOCs) revealed that the SFT-SGAT attains high classification performance in healthy subjects (maximum accuracy of 95.84%) and was successfully applied to DOC patients, with four patients achieving emotion recognition accuracies exceeding 60%. The experiments demonstrate the effectiveness of the proposed SFT-SGAT model in cross-subject EEG emotion recognition and its potential for assessing levels of consciousness in patients with DOC.