《科学》（20240329出版）一周论文导读

Science, 29 MAR 2024, VOL 383, ISSUE 6690

《科学》2024年3月29日，第383卷，6690期

物理学Physics

Ultrafast Kapitza-Dirac effect

超快卡皮查—狄拉克效应

▲ 作者：Kang Lin, Sebastian Eckart, Hao Liang, Alexander Hartung, Sina Jacob, Qinying Ji, et al.

▲ 链接：

https://www.science.org/doi/10.1126/science.adn1555

▲ 摘要：

类似于光通过材料光栅的光学衍射，当电子被驻波光场衍射时，就会发生卡皮查—狄拉克效应。在最初的描述中，这种效应与时间无关。

研究组将卡皮查—狄拉克效应扩展到时域。通过跟踪60飞秒（其中1飞秒= 10?15秒）驻波脉冲衍射的脉冲电子波包在泵浦探测方案中的时空演化，观察到了随时间变化的衍射模式。

观察到的模式条纹间距与传统卡皮查—狄拉克效应产生的条纹间距不同。通过利用这种时间分辨衍射方案，研究组可获得自由电子相位特性的时间演化，并有望成像离子势和电子退相干。

▲ Abstract：

Similar to the optical diffraction of light passing through a material grating, the Kapitza-Dirac effect occurs when an electron is diffracted by a standing light wave. In its original description, the effect is time independent. Here, we extended the Kapitza-Dirac effect to the time domain. By tracking the spatiotemporal evolution of a pulsed electron wave packet diffracted by a 60-femtosecond (where one femtosecond = 10?15 seconds) standing wave pulse in a pump-probe scheme, we observed time-dependent diffraction patterns. The fringe spacing in the observed pattern differs from that generated by the conventional Kapitza-Dirac effect. By exploiting this time-resolved diffraction scheme, we can access the time evolution of the phase properties of a free electron and potentially image ionic potentials and electronic decoherences.

材料科学Materials Science

A hyperelastic hydrogel with an ultralarge reversible biaxial strain

具有超大可逆双轴应变的超弹性水凝胶

▲ 作者：Lili Chen, Zhekai Jin, Wenwen Feng, Lin Sun, Hao Xu & Chao Wang

▲ 链接：

https://www.science.org/doi/10.1126/science.adh3632

▲ 摘要：

超弹性材料对大应变表现出非线性弹性响应，而水凝胶由于交联不均匀和交联间链段有限，通常具有较低的弹性范围。

研究组通过引入可逆的珍珠项链结构（珠子由绳子连接）开发了一种具有更宽弹性范围的超弹性水凝胶。亚纳米微珠在循环机械应变下能高效展开并重新折叠；因此，水凝胶在被拉伸至面应变超过10000%后可迅速恢复。

此外，水凝胶可从轻微机械损伤（如针扎和割伤）中快速愈合。这些进展使该离子水凝胶成为多功能气动夹具材料的理想选择，可同时提供超大抓握范围、自我感知能力和快速愈合能力。

▲ Abstract：

Hyperelastic materials exhibit a nonlinear elastic response to large strains, whereas hydrogels typically possess a low elastic range due to the nonuniform cross-linking and limited chain segments among cross-links. We developed a hyperelastic hydrogel that possesses a broader elastic range by introducing a reversible pearl-necklace structure, in which beads are connected by strings. The subnanometric beads can efficiently unfold and refold under cyclic mechanical strains; thus, the hydrogel can rapidly recover after being stretched to an areal strain of more than 10,000%. Additionally, the hydrogel can quickly heal from minor mechanical damages such as needle punctures and cuts. These advancements make our ionic hydrogels ideal for multifunctional pneumatic gripper materials; they simultaneously offer an ultralarge gripping range, self-sensing capabilities, and fast healing abilities.

化学Chemistry

Biodegradable ferroelectric molecular crystal with large piezoelectric response

具有大压电响应的可生物降解铁电分子晶体

▲ 作者：Han-Yue Zhang, Yuan-Yuan Tang, Zhu-Xiao Gu, Peng Wang, Xiao-Gang Chen, Hui-Peng Lv, et al.

▲ 链接：

https://www.science.org/doi/10.1126/science.adj1946

▲ 摘要：

瞬态可植入压电材料是生物传感、药物传递、组织再生、抗菌和肿瘤治疗的理想材料。为了在人体中使用，这种压电材料必须表现出灵活性、生物相容性和生物可降解性。这些要求对传统的无机压电氧化物和压电聚合物而言颇具挑战。

研究组发现分子晶体HOCH2(CF2)3CH2OH [2,2,3,3,4,4—六氟戊烷—1,5—二醇（HFPD）]具有高压电性，无需极化处理的大压电系数d33约为138 pC/N，压电电压常数g33约为2450 × 10?3 Vm/N，具有良好的生物细胞相容性，在生理环境中表现出良好的生物降解性和生物安全性。

HFPD可与聚乙烯醇复合，以形成d33为34.3 pC/N的柔性压电薄膜。该材料证明了分子晶体具有吸引人的压电特性，且有望应用于瞬态可植入机电设备。

▲ Abstract：

Transient implantable piezoelectric materials are desirable for biosensing, drug delivery, tissue regeneration, and antimicrobial and tumor therapy. For use in the human body, they must show flexibility, biocompatibility, and biodegradability. These requirements are challenging for conventional inorganic piezoelectric oxides and piezoelectric polymers. We discovered high piezoelectricity in a molecular crystal HOCH2(CF2)3CH2OH [2,2,3,3,4,4-hexafluoropentane-1,5-diol (HFPD)] with a large piezoelectric coefficient d33 of ~138 picocoulombs per newton and piezoelectric voltage constant g33 of ~2450 ×10?3?volt-meters per newton under no poling conditions, which also exhibits good biocompatibility toward biological cells and desirable biodegradation and biosafety in physiological environments. HFPD can be composite with polyvinyl alcohol to form flexible piezoelectric films with a d33 of 34.3 picocoulombs per newton. Our material demonstrates the ability for molecular crystals to have attractive piezoelectric properties and should be of interest for applications in transient implantable electromechanical devices.

人工智能Artificial Intelligence

Mechanism for feature learning in neural networks and backpropagation-free machine learning models

神经网络中的特征学习机制与无反向传播机器学习模型

▲ 作者：Adityanarayanan Radhakrishnan, Daniel Beaglehole, Parthe Pandit & Mikhail Belkin

▲ 链接：

https://www.science.org/doi/10.1126/science.adi5639

▲ 摘要：

了解神经网络如何学习数据中的特征或相关模式以进行预测，对于其在技术和科学应用中的可靠使用至关重要。

在这项工作中，研究组提出了一种统一的数学机制，称为平均梯度外积（AGOP），以表征神经网络中的特征学习。他们提供了AGOP可捕获各种神经网络架构所学习特征的经验证据，包括基于转换器的语言模型、卷积网络、多层感知器和递归神经网络。

此外，还证明了无反向传播的AGOP在机器学习模型（如内核机器）中实现了特征学习，而先验不能识别特定于任务的特征。总而言之，研究组建立了一个基本机制，可在神经网络中捕获特征学习，并在通用机器学习模型中实现特征学习。

▲ Abstract：

Understanding how neural networks learn features, or relevant patterns in data, for prediction is necessary for their reliable use in technological and scientific applications. In this work, we presented a unifying mathematical mechanism, known as average gradient outer product (AGOP), that characterized feature learning in neural networks. We provided empirical evidence that AGOP captured features learned by various neural network architectures, including transformer-based language models, convolutional networks, multilayer perceptrons, and recurrent neural networks. Moreover, we demonstrated that AGOP, which is backpropagation-free, enabled feature learning in machine learning models, such as kernel machines, that a priori could not identify task-specific features. Overall, we established a fundamental mechanism that captured feature learning in neural networks and enabled feature learning in general machine learning models.

地球科学Earth Science

Quantifying methane emissions from United States landfills

量化美国垃圾填埋场的甲烷排放

▲ 作者：Daniel H. Cusworth, Riley M. Duren, Alana K. Ayasse, Ralph Jiorle, Katherine Howell, Andrew Aubrey, et al.

▲ 链接：

https://www.science.org/doi/10.1126/science.adi7735

▲ 摘要：

来自固体废物的甲烷排放可能占全球人为预算的很大一部分，但很少有全面研究来评估存量假设。

研究组使用空中成像光谱仪对2016年至2022年间美国18个州数百个大型垃圾填埋场的排放量进行了量化。该研究覆盖了美国20%的开放垃圾填埋场，代表了对废物部门甲烷点源最系统的测量研究。

研究组在大多数（52%）站点检测到显著的点源甲烷排放，其中许多站点的排放持续多次检测（数周至数年）。将这些数据与15个垃圾填埋场的独立同期原位航空观测数据进行比较，并建立了良好的一致性。

该研究结果表明，在气候变化减缓政策的背景下，有必要对垃圾填埋场甲烷排放进行长期的天气尺度监测。

▲ Abstract：

Methane emissions from solid waste may represent a substantial fraction of the global anthropogenic budget, but few comprehensive studies exist to assess inventory assumptions. We quantified emissions at hundreds of large landfills across 18 states in the United States between 2016 and 2022 using airborne imaging spectrometers. Spanning 20% of open United States landfills, this represents the most systematic measurement-based study of methane point sources of the waste sector. We detected significant point source emissions at a majority (52%) of these sites, many with emissions persisting over multiple revisits (weeks to years). We compared these against independent contemporaneous in situ airborne observations at 15 landfills and established good agreement. Our findings indicate a need for long-term, synoptic-scale monitoring of landfill emissions in the context of climate change mitigation policy.

公共卫生Public Health

Design of a SARS-CoV-2 papain-like protease inhibitor with antiviral efficacy in a mouse model

在小鼠模型中设计具有抗病毒功效的SARS-CoV-2木瓜酶样蛋白酶抑制剂

▲ 作者：Bin Tan, Xiaoming Zhang, Ahmadullah Ansari, Prakash Jadhav, Haozhou Tan, Kan Li, et al.

▲ 链接：

https://www.science.org/doi/10.1126/science.adm9724

▲ 摘要：

SARS-CoV-2变异株和耐药突变体的出现要求使用更多的口服抗病毒药物。SARS-CoV-2木瓜酶样蛋白酶（PLpro）是一个有前景但颇具挑战性的药物靶点。

研究组设计并合成了85种非共价PLpro抑制剂，其与最近发现的泛素结合位点和S4亚位点附近的已知BL2凹槽袋相结合。先导物抑制PLpro的抑制常数Ki值为13.2至88.2纳摩尔。八种先导物的PLpro共晶结构揭示了其相互作用模式。

体内先导Jun12682抑制SARS-CoV-2及其变异株，包括耐尼马特瑞韦株，EC50在0.44至2.02微摩尔之间。在SARS-CoV-2感染小鼠模型中，口服Jun12682可改善生存，减少肺部病毒载量和病变，这表明PLpro抑制剂是颇有前景的口服SARS-CoV-2抗病毒候选药物。

▲ Abstract：

The emergence of SARS-CoV-2 variants and drug-resistant mutants calls for additional oral antivirals. The SARS-CoV-2 papain-like protease (PLpro) is a promising but challenging drug target. We designed and synthesized 85 noncovalent PLpro inhibitors that bind to a recently discovered ubiquitin binding site and the known BL2 groove pocket near the S4 subsite. Leads inhibited PLpro with the inhibitory constant Ki values from 13.2 to 88.2 nanomolar. The co-crystal structures of PLpro with eight leads revealed their interaction modes. The in vivo lead Jun12682 inhibited SARS-CoV-2 and its variants, including nirmatrelvir-resistant strains with EC50 from 0.44 to 2.02 micromolar. Oral treatment with Jun12682 improved survival and reduced lung viral loads and lesions in a SARS-CoV-2 infection mouse model, suggesting that PLpro inhibitors are promising oral SARS-CoV-2 antiviral candidates.