ABSTRACT
AI has been widely applied in scientific scenarios, such as robots performing chemical synthetic actions to free researchers from monotonous experimental procedures. However, there exists a gap between human-readable natural language descriptions and machine-executable instructions, of which the former are typically in numerous chemical articles, and the latter are currently compiled manually by experts. We apply the latest technology of pre-trained models and achieve automatic transcription between descriptions and instructions. We design a concise and comprehensive schema of instructions and construct an open-source human-annotated dataset consisting of 3950 description-instruction pairs, with 9.2 operations in each instruction on average. We further propose knowledgeable pre-trained transcription models enhanced by multi-grained chemical knowledge. The performance of recent popular models and products showing great capability in automatic writing (e.g., ChatGPT) has also been explored. Experiments prove that our system improves the instruction compilation efficiency of researchers by at least 42%, and can generate fluent academic paragraphs of synthetic descriptions when given instructions, showing the great potential of pre-trained models in improving human productivity.
ABSTRACT
In metal hydride-catalyzed alkene hydrofunctionalization reactions via hydrogen atom transfer, simple carbonyl groups have been well-recognized as good somophiles at the carbon for C-C bond formation. Here we report an alternative pathway exploring the carbonyl as an O-nucleophile to make new C-O bonds during the CoH-catalyzed oxidative cyclization of alkenyl aldehydes. This reaction provides a rapid, mild, modular, and stereoselective (up to >20:1) entry to saturated O-heterocycles via nucleophilic trapping of an in situ-formed oxocarbenium intermediate. The key to overriding the carbonyl's innate somophilicity was found to be promoting the formation of organocobalt species and suppressing the radical exchange.
