AI Science Tracker
EN

Learned reaction models guide improved yields for three ester hydrogenations

University of Cambridge; Université Paris Cité; University of St Andrews · 1 May 2023 · materials

Discovery

Learned reaction models guide improved yields for three ester hydrogenations

Source check: Source checked. Not independent scientific confirmation.

Finding: Learned models of ruthenium-catalysed ester hydrogenation guided laboratory condition choices for four previously studied substrates. Three experiments gave higher alcohol yields than the corresponding literature conditions; a fourth failed despite a high predicted yield. The represented result is a bounded improvement of known reactions, rather than a new catalyst, transformation or universally accurate yield predictor.

Technology: Gaussian-process regression and comparative learned reaction models

Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.

AI contribution
Learned regressors map substrate/catalyst descriptors and reaction conditions to measured yield and guide new condition choices subsequently tested in the laboratory.
Test and result
The journal paper reports measured alcohol yields of 96%,63% and 99% compared with corresponding literature values of 3%,0% and 32%. The fourth experiment yielded below 1% despite an 84% model prediction.
Limitation / still unknown
An earlier matching preprint,10.26434/chemrxiv-2022-t1zsn, is dated January 10,2022; its full original body was not authenticated, so this record uses the documented May 1,2023 journal version and does not claim first public priority. The substrates and catalysts were known; improvements concern these conditions, not new chemical entities or general catalyst superiority. Prediction errors and high-yield-biased data limit generalization.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://doi.org/10.1039/d3dd00029j · https://research-repository.st-andrews.ac.uk/bitstream/10023/27623/1/Mishra_2023_DD_PredictingRuthenium_CC.pdf

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Full primary author-deposited journal PDF front, learned-model Methods, experimental validation and Table2; exact matching preprint relationship and current Crossref status. Journal-version result represented; original preprint result content not asserted..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse materials discoveries.