{
  "source": "ORD Dataset Wishlist",
  "snapshot_date": "2026-09-25",
  "ord_wishlist_status_at_snapshot": "Available",
  "ord_wishlist_card": "https://github.com/orgs/open-reaction-database/projects/11?pane=issue&itemId=126570009",
  "ord_data_issue": null,
  "title": "A self-optimised approach to synthesising DEHiBA for advanced nuclear reprocessing, exploiting the power of machine-learning",
  "doi": "10.1039/D3RE00357D",
  "journal": "Reaction Chem Eng",
  "publication_date": "2023-11-03",
  "authors": [
    "Thomas Leonard Shaw",
    "Adam D. Clayton",
    "Ricardo Labes",
    "Thomas M. Dixon",
    "Sarah L. Boyall",
    "Oliver J. Kershaw",
    "Richard A. Bourne",
    "Bruce C. Hanson"
  ],
  "article_open_access": {
    "status": "hybrid",
    "license": "cc-by"
  },
  "dataset_size_estimate": 300,
  "experiment_type": "flow",
  "ord_priority": "Medium",
  "ord_difficulty": "Normal",
  "source_data_kind": "si_pdf_tables",
  "smiles_in_source": "no",
  "needs_author_contact": false,
  "data_locations": [
    {
      "url": "https://doi.org/10.1039/D3RE00357D",
      "what": "Article and SI (per the board notes, ~300 self-optimisation experiments in SI PDF tables). The publisher page blocks automated access, so the SI files were not listed. The paper is not in Europe PMC and no DataCite deposit was found.",
      "license": "CC BY 3.0 (article, per Crossref; SI license not stated separately)"
    }
  ],
  "notes": "Reaction Chemistry & Engineering 2023/2024, DOI 10.1039/D3RE00357D. The article is CC BY 3.0 per Crossref, which covers the article only; the SI license is not stated separately. About 300 experiments come from machine-learning-driven self-optimisation in flow of the synthesis of DEHiBA (N,N-di(2-ethylhexyl)isobutyramide), an extractant for nuclear reprocessing. The data is in tables in the SI PDF and needs extraction. Only a few distinct chemicals are involved, so SMILES are easy to write by hand. No data repository deposit was found. The publisher site blocks automated access, so SI file names and formats were not verified. A machine-readable copy of the tables, if the authors can share one, would reduce transcription errors.",
  "verification_caveats": "Flow self-optimisation. Each experiment is a steady-state flow condition set (flow rates or residence time, temperature, stoichiometry) chosen by an ML optimiser, so it needs ORD flow conditions. Objectives may be several (e.g. yield plus throughput or cost) and must be mapped to the right outcomes. Transcribing PDF tables risks numeric errors, which need spot checks against the SI."
}