BAL FENTON DATA SET LORETT 2026

This dataset contains 31,595 patent documents retrieved through systematic data mining on Google Patents Public Datasets (patents-public-data.patents.publications) via Google BigQuery, covering the period 2010–2026. The dataset was generated as part of a Freedom to Operate (FTO) analysis for a technology based on the oxidation of dimercaprol (British Anti-Lewisite, BAL) via a Fenton-type process for the therapeutic and localized generation of hydrogen sulfide (H₂S). Documents are organized into three thematic blocks: (1) therapeutic applications of H₂S and gasotransmitters (8,392 documents), (2) medical and synthetic uses of dimercaprol/BAL (606 documents), and (3) Fenton-type reactions and thiol-oxidative processes generating small molecules including H₂S (22,597 documents). Retrieval was performed using multilingual Boolean queries (English, Spanish, French, German, Chinese, Japanese) applied to title, abstract, and claims fields. Each record includes publication number, publication date, country code, patent family ID, title, abstract, IPC classification, CPC classification, and assignees. The dataset was processed and deduplicated using Python (pandas), resulting in 31,595 unique records across 22,514 patent families. It serves as the primary input for the KNIME-based depuration and IPC filtering workflow described in the associated methodology. The SQL script used for data extraction is provided as supplementary material.本数据集通过谷歌BigQuery平台，针对谷歌专利公开数据集（Google Patents Public Datasets，patents-public-data.patents.publications）开展系统性数据挖掘，共检索得到31,595件专利文献，覆盖2010至2026年的时间范围。本数据集是为一项基于芬顿型工艺（Fenton-type process）氧化二巯基丙醇（dimercaprol，又名英国抗路易氏剂，British Anti-Lewisite，BAL）以实现治疗性局部生成硫化氢（hydrogen sulfide, H₂S）的技术，开展自由实施（Freedom to Operate, FTO）分析所生成的配套成果。 文献按照三大主题模块进行组织：(1) 硫化氢（hydrogen sulfide, H₂S）及气体递质的治疗应用（共计8,392件文献）；(2) 二巯基丙醇/BAL的医疗与合成用途（共计606件文献）；(3) 芬顿型反应及硫醇氧化过程（可生成包括硫化氢在内的小分子物质，共计22,597件文献）。检索阶段采用多语言布尔查询（覆盖英语、西班牙语、法语、德语、中文、日语），针对专利的标题、摘要和权利要求字段开展检索。每条记录包含公开号、公开日期、国家代码、专利家族ID、标题、摘要、国际专利分类（International Patent Classification, IPC）号、协同专利分类（Cooperative Patent Classification, CPC）号以及专利权人信息。 本数据集通过Python（pandas库）完成去重处理，最终得到覆盖22,514个专利家族的31,595条唯一记录。本数据集作为核心输入，用于配套方法学中基于KNIME的净化与IPC分类过滤工作流。数据提取所用的SQL脚本已作为补充材料提供。