Advanced modeling and optimization of rhodamine B photodegradation by BiOCl/Bi₂Mo₃O₁₂ nanoparticles for sustainable water treatment
Year: 2026
Type: article
Abstract: Dye-polluted industrial wastewater is toxic due to organic compounds and endangers the habitat of living organisms. This study determined the optimal process conditions using experimental design, modeling, and optimization, and predicted the process for removing Rhodamine B (RhB) dye from water via photodegradation with BiOCl/Bi2Mo3O12 under irradiation. This study presents a novel integrated framework that simultaneously combines photocatalyst surface roughness characterization, statistical correlation analysis (PKS), response surface methodology (RSM), and advanced machine learning (ML) models to improve prediction and process understanding. Parameters such as RhB concentration, pH, and polyvinylpyrrolidone (PVP) concentration were analyzed to evaluate RhB removal from water. Optimum conditions for dye removal under irradiation were a RhB concentration of 10 mg/L, a PVP concentration of 0.1 g, and a pH of 3, yielding a maximum efficiency of 99.65%, as predicted by the response surface method (RSM). At the same time, a feedforward neural network (FNN) indicated that the RhB concentration was a significant factor, with a sensitivity of -0.9243. This was further supported by Pearson, Kendall, and Spearman correlation analyses, as well as by machine learning (ML) models [Random Forest, Gradient Boosting, XGBoost, Autoencoder, and Deep Neural Network (DNN)], trained on a dataset comprising 27,681 observations, which achieved 95.4–99.7% prediction accuracy for dye removal during photodegradation. The surface structure characteristics of the BiOCl/Bi2Mo3O12 prepared with 0.1 g PVP were 0.75 (Ra/Rq), 2.98 (Rz/Ra), and 1.41 (Rv/Rp), with an arithmetic average surface roughness of 35.2 nm, demonstrating that topology-modulated surface roughness is a key controlling factor in photocatalytic performance. The performance of the BiOCl/Bi2Mo3O12 heterogeneous nanoparticles was evaluated for the photocatalysis of Rhodamine B dye. Efficient removal of Rhodamine B dye was demonstrated in both acidic and neutral environments within a short time. Surface characteristics of BiOCl/Bi2Mo3O12 heterogeneous nanoparticles were determined and designated as influential. Machine Learning models provided a better description of Rhodamine B dye photodegradation prediction.
Source: Discover Materials
Institutions: Semnan University, Durban University of Technology
Language: English
Field: Energy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,590
OpenAlex ID: W7212292684
FWCI:
Cites:
Cited by:
Related to: