LG

Large language models now enable chemistry experts to extract structured data from voluminous literatures. Here we present a human-centered workflow that lets chemistry experts explore, curate, and formalize LLM-extracted data without programming barriers. This work establishes a practical paradigm for collaboration between humans and artificial intelligence and urges chemistry experts to work directly with the extracted data to safeguard human judgment, chemical intuition, and critical thinking at the center of discovery.

Abstract

Large language models (LLMs) hold considerable promise for large-scale data extraction from scientific literatures for catalyst design and practical optimization. Yet, turning such outputs into reliable, formalized chemical knowledge would heavily rely on domain expertise rather than end-to-end automation. Herein, we present a human-in-the-loop workflow integrating LLM-facilitated structured data extraction with iterative, expert-guided curation and analysis. As a proof of concept, we take single-atom catalysts (SACs) for advanced oxidation processes (AOPs) as an example, enabling efficient data extraction, rigorous curation, and statistically driven interpretation. Thus, we uncover the key correlations among metal types, coordination environments, reaction substances, and catalytic performance, providing deeper mechanism insights into SAC-driven AOPs. In contrast to fully automated, end-to-end models, our approach relies on human-driven optimization at multiple stages, and underscores human insight as central to understand LLM outputs. By introducing human-driven prompt refinement, model comparison, and expert-led analysis, our method ensures that human cognition remains central to interpreting LLM outputs and converting structured data into reliable scientific knowledge. Our work addresses the limitations inherent in fully automated, end-to-end methodologies and effectively bridges the gap between structured outputs and catalytically meaningful insights.

A series of cationic phosphonium-palladium(0) complexes were synthesized using an ambiphilic phosphorus-based ligand. A strong Z-type interaction was identified by DFT calculations (NBO, ETS-NOCV, and QTAIM analyses), surpassing the strength of classical L-type P─Pd bonds. The pronounced Pd(0) stabilization facilitates mild P─C coupling reaction, showcasing the effect of Z-type ligands on reductive elimination in transition metal mediated transformations.

Abstract

Phosphorus-based ligands are prototypical L-type donors. Here, we introduce a Lewis acidic phosphonium cation that engages in unprecedented Z-type coordination to palladium(0). The resulting Pd→P(V) interactions are structurally unique and stronger than conventional L-type Pd─phosphine bonds, as revealed by NMR spectroscopy, x-ray crystallography, and quantum chemical analyses (NBO, ETS-NOCV, and QTAIM). It enables the formation of rare L₂Z-type Pd(0) complexes with distinctive optical and electronic properties. A notably mild P─C bond forming reaction suggests an overlooked role of Z-type interactions for facilitating reductive elimination pathways.

A lithium iodide-mediated electrophile–electrophile ring opening of azabicyclo[1.1.0]butanes (ABBs) furnishes densely substituted azetidines. Leveraging ABB ring strain under mild, base-free conditions, the method enables sequential electrophilic incorporation, displays broad scope and operational simplicity, and is amenable to late-stage functionalization.

Abstract

Azetidines are privileged nitrogen heterocycles in medicinal chemistry; however, current synthetic methodologies utilizing azabicyclo[1.1.0]butanes (ABBs) predominantly rely on classical nucleophile–electrophile or radical-based approaches. Here, we report an unprecedented electrophile–electrophile ring-opening strategy enabled by lithium iodide-mediated activation of ABBs, offering direct and versatile access to densely substituted azetidines through sequential electrophilic incorporation. This new reactivity exploits the inherent ring strain of ABBs under mild, base-free, room-temperature conditions, thereby eliminating the necessity for harsh reaction environments. Mechanistic studies and control experiments unequivocally establish the pivotal role of lithium iodide in ring-opening and generating an in situ enolate intermediate, facilitating efficient bisfunctionalization via electrophilic trapping. The operational simplicity, extensive substrate scope, and remarkable compatibility with late-stage functionalization significantly enhance the synthetic versatility and modularity of ABB-derived azetidines, presenting a powerful approach for rapidly assembling complex molecules containing azetidines.

We report a radical umpolung strategy for regioselective difunctionalization of azabicyclo[1.1.0]butanes and bicyclo[1.1.0]butanes using sulfinate-derived electrophilic sulfonyl radicals, enabling ring-opening and subsequent functionalization with acyl derivatives, aldehydes, or CO₂ under mild conditions

Abstract

Azetidines and cyclobutanes are increasingly valued as potent bioisosteres of pyridines and benzenes in medicinal chemistry. Herein, we report a radical umpolung strategy for the regioselective difunctionalization of azabicyclo[1.1.0]butanes (ABBs) and bicyclo[1.1.0]butanes (BCBs) that exhibits complementary regioselectivity to conventional polar strain-release methods. This approach uses photocatalytically generated electrophilic sulfonyl radicals from readily available sulfinates to selectively add to nitrogen in ABBs and electron-rich sites of BCBs, triggering strain-release ring-opening. The resulting radical intermediates are subsequently captured through two pathways: N-heterocyclic carbene (NHC)-catalyzed radical–radical cross-coupling enables efficient acylation, while single-electron reduction generates carbanions capable of nucleophilic addition to electrophiles such as CO₂ and aldehydes. The umpolung reactivity of this protocol enhances synthetic versatility by accommodating diverse azetidine functionalities under mild conditions to afford densely functionalized azetidines and cyclobutanes that are difficult to access through existing methods.

At present, industrial production methods, such as the Haber–Bosch (H–B) process using N₂ and D₂, require harsh reaction conditions and complexity equipment, leading to a high production cost of ND₃. This work demonstrates a two-step relay route to produce ND₃ by using air and deuterium oxide (D₂O) as raw materials, including plasma-driven air-to-NO_x conversion and electrocatalytic NO_x ^–-to-ND₃ conversion, which can be directly driven by volatile green energy.

Abstract

Deuterated ammonia (ND₃) exhibits growing market demand in the fields of chemical analysis, pharmaceutical industry and semiconductor manufacturing. Currently, industrial production of ND₃ relies on harsh conditions and complex processes, leading to high production cost and security risk. Herein, we propose a sustainable relay strategy to produce ND₃ by using air and deuterium oxide (D₂O) as raw materials, including plasma-driven air-to-NO_x conversion and electrocatalytic NO_x ^–-to-ND₃ conversion. The insufficient supply of reactive deuterium (*D) leads to sluggish kinetics of electrocatalytic deuterium reaction. The well-designed F modified cobalt (F–Co) catalyst exhibits a remarkable yield of 0.75 mmol h⁻¹ cm⁻² and a Faradaic efficiency of 80.43% for ND₃ at 200 mA cm⁻². The combined results of characterizations reveal that fluorine (F) atom can boost D₂O dissociation and suppress competing deuterium evolution reaction, thereby providing abundant *D for deuteration reaction. Notably, a pilot-scale demonstration system, consisting of non-thermal plasma, flow electrolyzer, air stripping and ammonia absorber, is constructed to produce practicable ND₃ solution (2.8 wt%) with ∼21.45 mmol h⁻¹ ND₃ production capability by using air and D₂O as sources.

An electrochemical α-C─H functionalization of nitramines enables the synthesis of molecules containing bifunctional energetic heterocycles with promising properties. A telescoped, HNO₃-free sequence involving nitration and azolation steps offers a safer, modular, and scalable platform for the synthesis of energetic compounds.

Abstract

The synthesis of energetic materials (EMs) often involves hazardous reagents and harsh conditions, raising safety and environmental concerns. We herein present an electrochemical method for the ⍺-C─H azolation of nitramines, enabling the integration of nitramines and various nitrogen-rich azoles as dual energetic components within the same molecule. To enhance the practicality of the overall synthesis, we developed a tandem two-step process that transforms free amines into nitramines using stable and readily available reagents, which was complemented by subsequent electrochemical azolation to complete a streamlined, scalable preparation of bifunctional energetic compounds. Finally, a continuous flow system was employed to further improve the practicality of the electrosynthetic method, which substantially reduced electrolyte usage and increased productivity. Computational and experimental data revealed that the introduction of azoles, particularly those with additional nitro substituents, improves the energy density and thermal stability of nitramines. This work provides a proof of concept that the reported electrochemical azolation reaction may not only offer a safer and more sustainable alternative to traditional approaches for energetic material synthesis, but it will also provide a platform for the discovery of novel compounds with favorable energetic properties.

A series of proton-resistant TEMPO catholytes were developed via N-acetylamino bridging and heterocycle grafting, enabling efficient charge redistribution and exceptional redox stability. The optimized DMA-TEMPO derivative delivers near-quantitative capacity retention (99.98% over 560 cycles), offering a robust molecular design strategy for durable AORFBs.

Abstract

TEMPO is a widely studied catholyte for aqueous organic redox flow batteries (AORFBs) but suffers from proton-induced ring-opening degradation when its solubility is enhanced via hydrophilic substitution at the 4-position, leading to structural failure and rapid capacity fade. To address this issue, five TEMPO derivatives were synthesized through N-acetylamino bridging and nitrogen-containing heterocycle grafting strategy. Combined analyses using atomic dipole moment-corrected Hirshfeld (ADCH) charges, Fukui functions, and linear ion trap mass spectrometry (LTQ-XL) reveal that aromatic heterocycle functionalization enables favorable charge redistribution during redox cycling, enhancing both redox kinetics and molecular stability. In particular, dimethylaminopyridine-functionalized TEMPO (DMA-TEMPO) exhibits enhanced π-conjugation and basicity, which suppresses proton-driven ring-opening and significantly improves structural resilience. 1 M DMA-TEMPO catholyte delivers exceptional cycling performance, retaining 99.98% of its capacity after 560 cycles, while 2 M system maintains 97% capacity over 100 cycles. Compared to its structural analog 1 M PA-TEMPO, the cycle life is improved 18-fold. This study offers a robust molecular design strategy for developing proton-resistant catholytes, advancing the practical deployment of long-lasting AORFBs for grid-scale energy storage.

Nature Chemistry, Published online: 22 September 2025; doi:10.1038/s41557-025-01948-z

Synthetic methods that are applicable to a broad range of substrates are sought after, owing to their utility in industrial settings. This minireview describes considerations associated with how chemists define and identify general methods, especially with the emergence of modern analytical, high-throughput, and data science tools in chemistry, and gives the reader an overview of workflows that have been used to expedite this pursuit.

Abstract

The term “generality” has recently been popularized in synthetic chemistry, owing largely to the increasing use of high-throughput technology for producing vast quantities of data and the emergence of data science tools to plan and interpret these experiments. Despite this, the term has not been clearly defined, and there is no standardized approach toward developing a method with a diverse (general) scope. This minireview will examine different emerging strategies toward achieving generality using selected examples and aims to give the reader an overview of modern workflows that have been used to expedite this pursuit.

This review outlines major advances in the design, execution, analysis, and data management phases of high-throughput experimentation (HTE). The limitations and potential opportunities of applying modern HTE to organic synthesis are highlighted.

Abstract

High-throughput experimentation (HTE), the miniaturization and parallelization of reactions, is a valuable tool for accelerating diverse compound library generation, optimizing reaction conditions, and enabling data collection for machine learning (ML) applications. When applied to organic synthesis and methodology, HTE still poses various challenges due to the diverse workflows and reagents required, motivating advancements in reaction design, execution, analysis, and data management. To address these limitations, cutting-edge technologies, automation, and artificial intelligence (AI) have been implemented to standardize protocols, enhance reproducibility, and improve efficiency. Additionally, strategies to reduce bias and promote serendipitous discoveries have further strengthened HTE's impact. This review highlights recent advances at every stage of the HTE workflow, including the development of customized workflows, diverse analysis, and improved data management practices for greater accessibility and shareability. Furthermore, we examine the current state of the field, outstanding challenges, and future directions toward transforming HTE into a fully integrated, flexible, and democratized platform that drives innovation in organic synthesis.