The inherently slow kinetics of water dissociation during the alkaline hydrogen evolution reaction (HER) severely restrict the performance of non-noble metal catalysts. Herein, a NiMoO4 catalyst featuring electron-deficient Lewis acidic Ni sites, induced by asymmetric covalency competition in the Ni–O–Mo bridging structure, is developed to promote the adsorption, activation, and dissociation of water in alkaline HER. Experiments and calculations confirm that the stronger Mo–O covalency versus Ni–O in NiMoO4 drives electron redistribution from Ni to Mo via the bridging oxygen, rendering the Ni sites electron-deficient and significantly enhancing their Lewis acidity. With their vacant d orbitals, the Ni sites with enhanced Lewis acidity are capable of accepting lone-pair electrons from the oxygen atom of water molecule, thereby strengthening the adsorption and enrichment of interfacial H2O. This coordination simultaneously promotes the conversion and activation of interfacial water and weakens the H–OH bond, significantly lowering the water dissociation free energy. Therefore, NiMoO4 delivers a low overpotential of 317 mV at 100 mA cm−2, and the assembled electrolyzer requires only 1.79 V to achieve 500 mA cm−2 while maintaining robust durability for over 500 h. This work establishes a covalency competition mechanism to modulate Lewis acidity in bimetallic oxide bridging structures, providing an effective strategy for developing efficient electrocatalysts for alkaline HER.
Milling force signals are widely recognized as critical information sources for characterizing machining states, evaluating workpiece surface quality, and monitoring tool operating conditions. However, during high-speed acquisition processes, random missing segments may be introduced due to sensor failures, network transmission interruptions, or equipment buffering limitations, which consequently degrade subsequent feature extraction, process state assessment, and quality prediction. To overcome this limitation, a surface-quality-constrained milling force recovery method, termed Milling Force Recovery (MFR), is proposed. The core idea of MFR is to decompose intact milling force signals into periodic and residual components according to their distinct physical origins and subsequently recover the missing components based on their inherent characteristics. The periodic components are reconstructed by exploiting the intrinsic autocorrelation characteristics of milling force signals, whereas the residual components are recovered through surface-quality-constrained reconstruction using surface features spatially and temporally aligned with the corresponding missing segments. Specifically, intact milling force signals are first decomposed using variational mode decomposition (VMD) assisted by initial center frequency scanning to identify dominant periodic frequencies. A bidirectional SARIMA-based signal extension strategy is then introduced to mitigate boundary distortion induced by band-pass filtering, enabling reliable separation of periodic and residual components. Only local surface regions and surface features with low correlation to the target missing segments are utilized, thereby improving the physical consistency and reconstruction accuracy of the proposed method. Subsequently, component-specific neural network modules are established to capture the distinct characteristics of different signal components. The periodic components are reconstructed using a channel-independent autocorrelation prediction module and further refined through a surface-topography-based correction curve, whereas the residual components are recovered using a surface-quality-guided conditional variational autoencoder (CVAE) integrated with a conditional attention mechanism. Finally, a cross-component weighted loss function is introduced to alleviate error accumulation during signal reconstruction. Experimental results demonstrate that the proposed MFR method can effectively recover long-duration missing milling force signals. Across three datasets, the proposed method achieves an MSE of 0.0062 and an R2 value of 0.9657 in long-duration signal reconstruction experiments.
Thermal contact resistance (TCR) at the gas diffusion layer-catalyst layer (GDL-CL) interface significantly affects heat transfer, water distribution, and electrochemical performance in proton exchange membrane fuel cells (PEMFCs). However, previous studies have mainly focused on bipolar plate-related interfaces, while the coupled effects of GDL-CL interfacial TCR, electrode structures, and advanced flow-field configurations remain unclear. In this study, a three-dimensional two-phase non-isothermal computational fluid dynamics (CFD) model of a PEMFC with a bamboo-shaped (BS) cathode flow field is developed to investigate the effects of GDL-CL TCR, GDL thickness, and catalyst layer (CL) thickness on cell performance. The results demonstrate that the BS flow field enhances oxygen transport, liquid-water removal, and heat transfer, improving maximum power density by 2.81% under zero-TCR conditions and up to 3.97% under high-TCR conditions compared with the parallel flow field. Furthermore, within the investigated GDL thickness range of 0.1-0.3 mm, the preferred configuration shifts from 0.2 mm under low-TCR conditions to 0.1 mm under high-TCR conditions. Increasing CL thickness from 0.01 mm to 0.03 mm improves cell performance and TCR tolerance within the investigated range. These findings clarify the coupling between interfacial thermal resistance, flow-field structure, and electrode design for improved PEMFC performance.
Two dimensional nanomaterials have exhibit promising performance in membrane science, especially hexagonal boron nitride nanosheets (BNNSs) with inherent chemical inertness. Here, BNNSs membranes with compact and well-organized stacking texture were successfully prepared via vacuum-assisted filtration approach using chemo-mechanical exfoliated and functionalized BNNSs as matrix. The BNNSs with an average lateral dimension of ~100 nm have few layers structure and abundant hydroxyl and amino groups. The viscosity of the solvent and thickness of membrane are important factors affecting the permeability performance. Importantly, the BNNSs membranes exhibit fascinating filtration and separation performance for various pollutants in water, including abusive organic dyes, heavy metal ion and drugs. Specifically, the membrane with a typical thickness of 30 μm shows the water flux up to 2.1 L⋅m−2⋅min−1⋅bar−1 and 99.87% removal rate for tetracycline. These performances are not only caused by simple physical interception of BNNSs constructed nanochannels, but also by certain intermolecular chemical interactions between functional groups on BNNSs and pollutant molecules. This work would provide experimental foundation and analytical experience for the practical application of BNNSs membranes in the field of complex water purification.
