Skip to content
Open access

Sustainable W–BiVO₄ Nanocatalysts for Visible-light Removal of Reactive Black 5: Physicochemical Characterization and Multiresponse Optimization

Sep 2026 · Journal of Environmental Nanotechnology · 0 citations

Abstract

This study reports the green synthesis of pristine BiVO₄ and tungsten-doped BiVO₄ (W–BiVO₄) nanocatalysts using dragon-fruit-peel extract for the visible-light degradation of Reactive Black 5 (RB5). The novelty lies in integrating waste-derived green synthesis, multiresponse statistical optimization, machine-learning prediction, and kinetic evaluation within a single RB5-treatment framework. The W–BiVO₄ nanoparticles exhibited irregular rod-like and ellipsoidal morphologies with a mean particle size of 48.6 ± 10.4 nm. SEM–EDX confirmed the presence of Bi, V, O, and W, although this analysis alone does not conclusively establish substitutional incorporation of W into the BiVO₄ lattice. The synthesized BiVO₄ exhibited an estimated optical band gap of 2.55 eV, supporting visible-light activation. The photocatalytic performance of W–BiVO₄ was investigated using a 29-run Box–Behnken design by varying initial RB5 concentration, catalyst dosage, initial pH, and W-doping level. The highest experimental performance was 93.40% RB5 removal with an apparent pseudo-first-order rate constant of 0.01947 min⁻¹ at 20 mg L⁻¹ RB5, 1.20 g L⁻¹ catalyst dosage, pH 6, and 3 wt.% W doping. The quadratic RSM models were statistically significant and achieved R² values of 0.9991 for RB5 removal efficiency and 0.9831 for the apparent pseudo-first-order rate constant. Multiresponse optimization predicted 93.51% RB5 removal and an apparent pseudo-first-order rate constant of 0.02003 min⁻¹ at an initial RB5 concentration of 20.68 mg L⁻¹, W–BiVO₄ dosage of 1.162 g L⁻¹, initial pH of 5.30, and W-doping level of 3.44 wt.%, with an overall desirability of 1.000. Among six machine-learning models evaluated using leave-one-out cross-validation, Gaussian process regression provided the best prediction of RB5 removal efficiency (R² = 0.991), whereas support vector regression provided the best prediction of the apparent pseudo-first-order rate constant (R² = 0.924). These findings demonstrate the potential of combining green W–BiVO₄ synthesis, experimental optimization, and machine-learning prediction for visible-light treatment of RB5-containing wastewater.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.