Site Li

Electrocatalytic hydrogen evolution reaction (HER) holds promise in the renewable clean energy scheme. Crystalline Au and Ag are, however, poor in catalyzing HER, and the ligands on colloidal nanoparticles are generally another disadvantage. Herein, we report a thiolate (SR)-protected Au₃₆Ag₂(SR)₁₈ nanocluster with low coverage of ligands and a core composed of three icosahedral (I_h) units for catalyzing HER efficiently. This trimeric structure, together with the monomeric I_h Au₂₅(SR)₁₈^– and dimeric I_h Au₃₈(SR)₂₄, constitutes a unique series, providing an opportunity for revealing the correlation between the catalytic properties and the catalyst’s structure. The Au₃₆Ag₂(SR)₁₈ surprisingly exhibits high catalytic activity at lower overpotentials for HER due to its low ligand-to-metal ratio, low-coordinated Au atoms and unfilled superatomic orbitals. The current density of Au₃₆Ag₂(SR)₁₈ at −0.3 V vs RHE is 3.8 and 5.1 times that of Au₂₅(SR)₁₈^– and Au₃₈(SR)₂₄, respectively. Density functional theory (DFT) calculations reveal lower hydrogen binding energy and higher electron affinity of Au₃₆Ag₂(SR)₁₈ for an energetically feasible HER pathway. Our findings provide a new strategy for constructing highly active catalysts from inert metals by pursuing atomically precise nanoclusters and controlling their geometrical and electronic structures.

Ligand effects are of major interest in catalytic reactions owing to their potential critical role in determining the reaction activity and selectivity. Herein, we report ligand effects in the CO₂ electrochemical reduction reaction at the atomic level with three unique Au₂₅ nanoclusters comprising the same kernel but different protecting ligands (–XR, where X = S or Se, and R represents the carbon tail). It is observed that a change in the carbon tail shows no obvious impact on the catalytic selectivity and activity, but the anchoring atom (X = S or Se) strongly affects the electrocatalytic selectivity. Specifically, the S site acts as the active site and sustains CO selectivity, while the Se site shows a higher tendency of hydrogen evolution. Density functional theory (DFT) calculations reveal that the energy penalty associated with the *COOH formation is lower on the S site by 0.26 eV compared to that on the Se site. Additionally, the formation energy of the product (*CO) is lower on the sulfur-based Au nanocluster by 0.43 eV. We attribute these energetic differences to the higher electron density on the sulfur sites of the Au nanocluster, resulting in a modified bonding character of the reaction intermediates that reduce the energetic penalty for the *COOH and *CO formation. Overall, this work demonstrates that S/Se atoms at the metal–ligand interface can play an important role in determining the overall electrocatalytic performance of Au nanoclusters.

Wind is one of the most important sources of green energy, but the current technology for harvesting wind energy is only effective when the wind speed is beyond 3.5–4.0 m s⁻¹. This is mainly due to the limitation that the electromagnetic generator works best at high frequency. This means that light breezes cannot reach the wind velocity threshold of current wind turbines. Here, a high‐performance triboelectric nanogenerator (TENG) for efficiently harvesting energy from an ambient gentle wind, especially for speeds below 3 m s⁻¹ is reported, by taking advantage of the relative high efficiency of TENGs at low‐frequency. Attributed to the multiplied‐frequency vibration of ultra‐stretchable and perforated electrodes, an average output of 20 mW m⁻³ can be achieved with inlet wind speed of 0.7 m s⁻¹, while an average energy conversion efficiency of 7.8% at wind speed of 2.5 m s⁻¹ is reached. A self‐charging power package is developed and the applicability of the TENG in various light breezes is demonstrated. This work demonstrates the advantages of TENG technology for breeze energy exploitation and proposes an effective supplementary approach for current employed wind turbines and micro energy structure.

Semantic segmentation (SS) is an important perception manner for self-driving cars and robotics, which classifies each pixel into a pre-determined class. The widely-used cross entropy (CE) loss-based deep networks has achieved significant progress w.r.t. the mean Intersection-over Union (mIoU). However, the cross entropy loss can not take the different importance of each class in an self-driving system into account. For example, pedestrians in the image should be much more important than the surrounding buildings when make a decisions in the driving, so their segmentation results are expected to be as accurate as possible. In this paper, we propose to incorporate the importance-aware inter-class correlation in a Wasserstein training framework by configuring its ground distance matrix. The ground distance matrix can be pre-defined following a priori in a specific task, and the previous importance-ignored methods can be the particular cases. From an optimization perspective, we also extend our ground metric to a linear, convex or concave increasing function w.r.t. pre-defined ground distance. We evaluate our method on CamVid and Cityscapes datasets with different backbones (SegNet, ENet, FCN and Deeplab) in a plug and play fashion. In our extenssive experiments, Wasserstein loss demonstrates superior segmentation performance on the predefined critical classes for safe-driving.

The labels in medical diagnosis task are usually discrete and successively distributed. For example, the Diabetic Retinopathy Diagnosis (DR) involves five health risk levels: no DR (0), mild DR (1), moderate DR (2), severe DR (3) and proliferative DR (4). This labeling system is common for medical disease. Previous methods usually construct a multi-binary-classification task or propose some re-parameter schemes in the output unit. In this paper, we target on this task from the perspective of loss function. More specifically, the Wasserstein distance is utilized as an alternative, explicitly incorporating the inter-class correlations by pre-defining its ground metric. Then, the ground metric which serves as a linear, convex or concave increasing function w.r.t. the Euclidean distance in a line is explored from an optimization perspective. Meanwhile, this paper also proposes of constructing the smoothed target labels that model the inlier and outlier noises by using a unimodal-uniform mixture distribution. Different from the one-hot setting, the smoothed label endues the computation of Wasserstein distance with more challenging features. With either one-hot or smoothed target label, this paper systematically concludes the practical closed-form solution. We evaluate our method on several medical diagnosis tasks (e.g., Diabetic Retinopathy and Ultrasound Breast dataset) and achieve state-of-the-art performance.

We consider the problem of comparing the similarity of image sets with variable-quantity, quality and un-ordered heterogeneous images. We use feature restructuring to exploit the correlations of both inner$\&$inter-set images. Specifically, the residual self-attention can effectively restructure the features using the other features within a set to emphasize the discriminative images and eliminate the redundancy. Then, a sparse/collaborative learning-based dependency-guided representation scheme reconstructs the probe features conditional to the gallery features in order to adaptively align the two sets. This enables our framework to be compatible with both verification and open-set identification. We show that the parametric self-attention network and non-parametric dictionary learning can be trained end-to-end by a unified alternative optimization scheme, and that the full framework is permutation-invariant. In the numerical experiments we conducted, our method achieves top performance on competitive image set/video-based face recognition and person re-identification benchmarks.

This mini-Review summarizes the fusion growth patterns of metal nanoclusters based upon M₄, M₁₃ and M₁₄ building blocks.