The hydrogen evolution reaction (HER) presents an environmentally friendly, efficient, and cost-effective way to produce hydrogen fuel. Recently, two-dimensional (2D) transition metal dichalcogenides (TMDCs) have emerged as a fascinating class of HER materials. Their atomically thin nature acts as an interface with the possibility of maximum exposure of nearly all active sites to the reaction environment. However, for most non-metallic TMDCs, as exemplified by 2H-MoS₂, 1T-PtSe₂, and 1T′-ReSe₂, their basal planes are usually HER-inert except for minor edges. Based on their rich chemistry, significant progress has been recently made to activate the basal planes for HER. This review provides an overview of the current activating strategies. They can be classified into internal and external regulations, depending on whether the pristine structure is altered or not. The former directly creates more active sites through doping, vacancies, phases, grain boundaries (GBs), and superlattices. On the other hand, the latter indirectly promotes the activity utilizing external methods such as the electric field, strain, substrates, and heterostructures. We discuss each strategy's principles and characteristics and highlight their enhanced catalytic activities. We finally provide personal perspectives on the challenges and opportunities in this emerging field, including a joint strategy, a phase-engineering strategy, electrical–electrocatalytic coupling, as well as the potential of “single-atom-layer” catalysis.

Shallow feed-forward networks are incapable of addressing complex tasks such as natural language processing that require learning of temporal signals. To address these requirements, we need deep neuromorphic architectures with recurrent connections such as deep recurrent neural networks. However, the training of such networks demand very high precision of weights, excellent conductance linearity and low write-noise- not satisfied by current memristive implementations. Inspired from optogenetics, here we report a neuromorphic computing platform comprised of photo-excitable neuristors capable of in-memory computations across 980 addressable states with a high signal-to-noise ratio of 77. The large linear dynamic range, low write noise and selective excitability allows high fidelity opto-electronic transfer of weights with a two-shot write scheme, while electrical in-memory inference provides energy efficiency. This method enables implementing a memristive deep recurrent neural network with twelve trainable layers with more than a million parameters to recognize spoken commands with >90% accuracy.

Sulfonyl and sulfonimidoyl fluorides are versatile substrates in organic synthesis and medicinal chemistry. However, they have been exclusively used as S(VI)⁺ electrophiles for defluorinative ligations. Converting sulfonyl and sulfonimidoyl fluorides to S(VI) radicals is challenging and underexplored due to the strong bond dissociation energy of S^VI−F and high reduction potentials, but once achieved would enable dramatically expanded synthetic utility and downstream applications. In this report, we disclose a general platform to address this issue through cooperative organosuperbase activation and photoredox catalysis. Vinyl sulfones and sulfoximines are obtained with excellent E selectivity under mild conditions by coupling reactions with alkenes. The synthetic utility of this method in the preparation of functional polymers and dyes is also demonstrated.