Year: 2026
Type: preprint
Abstract: IRexp is a collection of experimental infrared band lists (cm⁻¹ peak positions) mined from open chemistry literature, with author-reported ¹H/¹³C NMR strings and resolved structures where available. The research corpus holds 121,233 records from the PMC Open Access Subset and the Chemotion repository, including 57,646 structure-linked entries and 39,118 IR + ¹H + ¹³C + structure quadruples. The public dataset of record contains 88,519 CC-BY/CC0 records released under CC-BY-4.0, with ShareAlike and non-commercial records distributed separately under their source licences. Every record carries a source identifier and a per-record licence stamp. Technical validation covers automated transcription, recall-proxy, and consistency audits. IRexp targets multimodal training, retrieval, and structure-elucidation tools that need structured, attributable experimental peak lists.
Source: ChemRxiv
Language: English
Subfield: Analytical Chemistry
Field: Chemistry
Domain: Physical Sciences
Open Access status: green
Awards: G80479029, G9000845080
OpenAlex ID: W7215039864