| Machine learning-assisted metronidazole antibiotic removal from polluted water by Bi2WO6/graphene oxide nanophotocatalyst |
| کد مقاله : 1290-IWWA |
| نویسندگان |
|
رسول گردان شکن1، آریا گردان شکن2، شکیبا عربیان3، شایان عبائی4، امید توکلی *5 1مدیر عامل شرکت مهندسی تحقیقات بسپار بهینه اصفهان رئیس هیئت مدیره شرکت بین المللی انبوهش آریای اصفهان 2دانشجوی دکتری مهندسی شیمی دانشگاه تهران مدیرعامل شرکت بین المللی انبوهش آریای اصفهان 3دانشجوی دکتری مهندسی شیمی دانشگاه صنعتی اصفهان مدیر تحقیق و توسعه شرکت بین المللی انبوهش آریای اصفهان 4دانشجوی کارشناسی ارشد دانشگاه تهران 5هیئت علمی گروه مهندسی شیمی دانشگاه تهران |
| چکیده مقاله |
| Metronidazole (MTZ), a widely prescribed nitroimidazole antibiotic, is an emerging water pollutant whose persistence and capacity to promote antibiotic resistance demand effective treatment technologies. Herein, a Bi2WO6/graphene oxide (GO) nanophotocatalyst was synthesized by a facile hydrothermal route with GO loadings of 0, 0.5, 1, 1.5 and 2 wt%. X-ray diffraction confirmed the orthorhombic russellite phase of Bi2WO6 in all composites, while scanning electron microscopy revealed Bi2WO6 nanoplates intimately anchored on wrinkled GO nanosheets, forming a heterojunction. The composite containing 1 wt% GO exhibited the highest photocatalytic activity. A full factorial design (three factors, three levels, 27 runs) was employed to investigate the effects of initial MTZ concentration (5–10 mg/L), photocatalyst dosage (0.4–0.8 g/L) and light intensity (50–150 W/m2). Removal efficiency increased with light intensity and catalyst dosage but decreased with increasing initial concentration, reaching a maximum experimental value of 99.4%. A feed-forward artificial neural network with a single hidden layer (5–9 neurons) was trained on the experimental data; the 6-neuron architecture was selected and achieved a validation coefficient of determination of 0.9815. Coupling this surrogate model with a real-coded genetic algorithm identified the optimal operating conditions as 5 mg/L MTZ, 0.8 g/L catalyst and 150 W/m2. Garson's analysis ranked catalyst dosage (37.2%) and light intensity (36.2%) as the dominant variables, followed by initial concentration (26.7%). Scavenger experiments confirmed that hydroxyl radicals and photogenerated holes govern mineralization, whereas superoxide plays a negligible role. |
| کلیدواژه ها |
| nanomaterials, water treatment, photocatalyst, heterojunction, artificial-intelligence, semiconductors |
| وضعیت: پذیرفته شده برای ارائه شفاهی |
