National Science Review最新文献

Affordable high-resolution cameras and state-of-the-art computer vision techniques have led to the emergence of various vision-based tactile sensors. However, current vision-based tactile sensors mainly depend on geometric optics or marker tracking for tactile assessments, resulting in limited performance. To solve this dilemma, we introduce optical interference patterns as the visual representation of tactile information for flexible tactile sensors. We propose a novel tactile perception method and its corresponding sensor, combining structural colors from flexible blazed gratings with deep learning. The richer structural colors and finer data processing foster the tactile estimation performance. The proposed sensor has an overall normal force magnitude accuracy of 6 mN, a planar resolution of 79 μm and a contact-depth resolution of 25 μm. This work presents a promising tactile method that combines wave optics, soft materials and machine learning. It performs well in tactile measurement, and can be expanded into multiple sensing fields.

Wetlands in the Qinghai-Tibet Plateau are a unique and fragile ecosystem undergoing rapid changes. We show two unique patterns of mercury (Hg) accumulation in wetland sediments. One is the 'surface peak' in monsoon-controlled regions and the other is the 'subsurface peak' in westerly-controlled regions. The former is attributed to the combined effects of increasing anthropogenic emissions and climate-induced changes in the cryosphere and wetland hydrology in the last 100-150 years. The climate changes in westerly-controlled regions in the last 50-70 years led to a fluctuation in hydrology and Hg peak in the sediment subsurface. The increase in legacy Hg input from soil erosion has largely enhanced the Hg accumulation rate in wetlands since the 1950s, especially in the proglacial wetlands. We highlight that accelerated glacier melting and permafrost thawing caused by global warming have altered geomorphology and hydrology, and affected Hg transport and accumulation in wetlands.

It remains challenging to design efficient bifunctional semiconductor materials in organic photovoltaic and photodetector devices. Here, we report a butterfly-shaped molecule, named WD-6, which exhibits low energy disorder and small reorganization energy due to its enhanced molecular rigidity and unique assembly with strong intermolecular interaction. The binary photovoltaic device based on PM6:WD-6 achieved an efficiency of 18.41%. Notably, an efficiency of 19.42% was achieved for the ternary device based on PM6:BTP-eC9:WD-6. Moreover, the photodetection device based on WD-6 demonstrated an ultrafast response speed (205 ns response time at λ of 820 nm) and a high cutoff frequency of -3 dB (2.45 MHz), surpassing the values of most commercial Si photodiodes. Based on these findings, we showcased an application of the WD-6-based photodetection device in high-speed optical communication. These results offer valuable insights into the design of organic semiconductor materials capable of simultaneously exhibiting high photovoltaic and photodetective performance.

Heterogeneous catalysts for parahydrogen-induced polarization (HET-PHIP) would be useful for producing highly sensitive contrasting agents for magnetic resonance imaging (MRI) in the liquid phase, as they can be removed by simple filtration. Although homogeneous hydrogenation catalysts are highly efficient for PHIP, their sensitivity decreases when anchored on porous supports due to slow substrate diffusion to the active sites and rapid depolarization within the channels. To address this challenge, we explored 2D metal-organic layers (MOLs) as supports for active Rh complexes with diverse phosphine ligands and tunable hydrogenation activities, taking advantage of the accessible active sites and chemical adaptability of the MOLs. By adjusting the electronic properties of phosphines, TPP-MOL-Rh-dppb (TPP = tris(4-carboxylphenyl)phosphine), featuring a κ ₂-connected di(phosphine) ligand, generated hyperpolarized styrene achieving an over-2400-fold signal enhancement and a polarization level of 20% for ¹H in methanol-d ₄ solution. The TPP-MOL-Rh-dppb effectively inherited the high efficiency and pairwise addition of its homogenous catalyst while maintaining the heterogeneity of MOLs. This work demonstrates the potential of 2D phosphine-functionalized MOLs as heterogeneous solid support for HET-PHIP.

Recently, the multimodal large language model (MLLM) represented by GPT-4V has been a new rising research hotspot, which uses powerful large language models (LLMs) as a brain to perform multimodal tasks. The surprising emergent capabilities of the MLLM, such as writing stories based on images and optical character recognition-free math reasoning, are rare in traditional multimodal methods, suggesting a potential path to artificial general intelligence. To this end, both academia and industry have endeavored to develop MLLMs that can compete with or even outperform GPT-4V, pushing the limit of research at a surprising speed. In this paper, we aim to trace and summarize the recent progress of MLLMs. First, we present the basic formulation of the MLLM and delineate its related concepts, including architecture, training strategy and data, as well as evaluation. Then, we introduce research topics about how MLLMs can be extended to support more granularity, modalities, languages and scenarios. We continue with multimodal hallucination and extended techniques, including multimodal in-context learning, multimodal chain of thought and LLM-aided visual reasoning. To conclude the paper, we discuss existing challenges and point out promising research directions.

Aqueous zinc batteries offer promising prospects for large-scale energy storage, yet their application is limited by undesired side reactions at the electrode/electrolyte interface. Here, we report a universal approach for the in situ building of an electrode/electrolyte interphase (EEI) layer on both the cathode and the anode through the self-polymerization of electrolyte additives. In an exemplified Zn||V₂O₅·nH₂O cell, we reveal that the glutamate additive undergoes radical-initiated electro-polymerization on the cathode and polycondensation on the anode, yielding polyglutamic acid-dominated EEI layers on both electrodes. These EEI layers effectively mitigate undesired interfacial side reactions while enhancing reaction kinetics, enabling Zn||V₂O₅·nH₂O cells to achieve a high capacity of 387 mAh g^-1 at 0.2 A g^-1 and maintain >96.3% capacity retention after 1500 cycles at 1 A g^-1. Moreover, this interphase-forming additive exhibits broad applicability to varied cathode materials, encompassing VS₂, VS₄, VO₂, α-MnO₂, β-MnO₂ and δ-MnO₂. The methodology of utilizing self-polymerizable electrolyte additives to construct robust EEI layers opens a novel pathway in interphase engineering for electrode stabilization in aqueous batteries.

The ability to rapidly recognize basic facial emotions (e.g. fear) is crucial for social interactions and adaptive functioning. To date, the origin of facial-emotion-recognition ability remains equivocal. Using a classical twin design in humans, we found a clear dissection of low and high spatial frequencies (LSF and HSF) in facial emotion perception: whereas genetic factors contributed to individual variation in LSF processing, HSF processing is largely shaped by environmental effects. Furthermore, the ability to recognize facial emotions of LSF content genetically correlated with the function of the amygdala. Crucially, single-unit recording of the amygdala in macaques further revealed the dissociation between LSF and HSF processing in facial emotion perception, indicating the existence of an evolutionarily conserved mechanism. This cross-species study enhances insights into the neurobiological dual-route model (subcortical vs. cortical) of emotion perception and illuminates the origin and the functional development of the emotional brain in primates.

In the face of advancements in microrobotics, intelligent control and precision medicine, artificial muscle actuation systems must meet demands for precise control, high stability, environmental adaptability and high integration miniaturization. Carbon materials, being lightweight, strong and highly conductive and flexible, show great potential for artificial muscles. Inspired by the butterfly's proboscis, we have developed a carbon-based artificial muscle, hydrogen-substituted graphdiyne muscle (HsGDY-M), fabricated efficiently using an emerging hydrogen-substituted graphdiyne (HsGDY) film with an asymmetrical surface structure. This muscle features reversible, rapid and continuously adjustable deformation capabilities similar to the butterfly's proboscis, triggered by the conversion of carbon bonds. The size of the HsGDY-M can be tuned by changing the HsGDY film width from ∼1 cm to 100 μm. Our research demonstrates HsGDY-M's stability and adaptability, maintaining performance at temperatures as low as -25°C. This artificial muscle was successfully integrated into a robotic mechanical arm, allowing it to swiftly adjust its posture and lift objects up to 11 times its own weight. Its beneficial responsiveness is transferable, enabling the transformation of 'inert' objects like copper foil into actuators via surface bonding. Because of its super sensitive and rapid deformation, HsGDY-M was applied to create a real-time tracking system for human finger bending movements, achieving real-time simulation and large-hand-to-small-hand control. Our study indicates that HsGDY-M holds significant promise for advancing smart robotics and precision medicine.