Text Mining and Machine Learning-Based Data Extraction for Environmental Pollutants (PPDs and PPDs-Q): Machine Learning-Enabled Global Aqueous Concentration Compilation and Potential Biotic Risk Implications

p-Phenylenediamine antioxidants (ppDs) and their quinone derivatives (ppDs-Q) are key additives in rubber products with strong toxicity, persistence, and increasing aquatic concentrations, though global data scarcity hinders risk assessment. Thisstudy addressed limitations of manual data extraction by using a python-based toolkit integrating ocR and Spacy neural networks to eficiently extract key information (concentrations,locations, media) from unstructured literature, compiling globaldata via Web of science, scopus, and pubMed. ppDs/ppDs-Q show significant concentration differences across global aquaticmedia: in artificial water, 6ppD in road runoff reaches 80019,000 ng/L (Seattle, UsA) and 907 ng/L (china's Greater Bay Area)with lower levels in wastewater effluents; in natural water, 6ppD-Q varies regionally (2003,500 ng/L in Us rivers, 290890 ng/lin Canadian rivers, 0.2611.3 ng/L in chinese rivers); snowmelt water shows high 6ppD.Q (19.0 g/L in seattle, 367 ng/L aver.age in Canadian cold cities,. species sensitivity follows patterns: salmoniformes are most sensitive to 6PPD/6PPD-Q (LC/EC5O 0.001 mg/L), followed by echinoderms/mollusks, with lower sensitivity in lower trophic organisms. Coupled analysis showsUS river 6PpD-Q exceeds ultra-sensitive organisms' thresholds, chinese rivers pose subchronic risks to benthic organisms,and widespread detection in human fluids indicates continuous exposure. This "ocR-neural network" framework resolvesmanual extraction bottienecks, provlding an extensible paradigm for other bolutants, wnle conclusons support regulationsecological protection,and risk management for PPDs.