This study describes the preparation of graphitic carbon nitride (g-C₃N₄), hematite (α-Fe₂O₃), and their g-C₃N₄/α-Fe₂O₃ heterostructure for the photocatalytic removal of methyl orange (MO) under visible light illumination. The facile hydrothermal approach was utilized for the preparation of the nanomaterials. Powder X-ray diffraction (XRD), Scanning electron microscopy (SEM), Energy dispersive X-ray (EDX), and Brunauer–Emmett–Teller (BET) were carried out to study the physiochemical and optoelectronic properties of all the synthesized photocatalysts. Based on the X-ray photoelectron spectroscopy (XPS) and UV-visible diffuse reflectance (DRS) results, an energy level diagram vs. SHE was established. The acquired results indicated that the nanocomposite exhibited a type-II heterojunction and degraded the MO dye by 97%. The degradation ability of the nanocomposite was higher than that of pristine g-C₃N₄ (41%) and α-Fe₂O₃ (30%) photocatalysts under 300 min of light irradiation. The formation of a type-II heterostructure with desirable band alignment and band edge positions for efficient interfacial charge carrier separation along with a larger specific surface area was collectively responsible for the higher photocatalytic efficiency of the g-C₃N₄/α-Fe₂O₃ nanocomposite. The mechanism of the nanocomposite was also studied through results obtained from UV-vis and XPS analyses. A reactive species trapping experiment confirmed the involvement of the superoxide radical anion (O₂^•−) as the key reactive oxygen species for MO removal. The degradation kinetics were also monitored, and the reaction was observed to be pseudo-first order. Moreover, the sustainability of the photocatalyst was also investigated.

The UF membrane with a narrow and uniform pore size distribution and a low tendency to foul has significant applications in wastewater treatment. A major hindrance in the preparation of the UF membrane with these features is the lack of a scalable and economical membrane fabrication method. Herein, we devise a new strategy to prepare a high-quality polyvinylidene fluoride/polymethyl acrylate/cellulose acetate (PVDF/PMMA/CA) blend UF membrane via a combination of the etching mechanism with the traditional Loeb–Sourirajan (L-S) phase inversion method. Different concentrations of silicon dioxide (SiO₂) nanoparticles (NP) in the membrane matrix were etched by using a 0.2 M hydrofluoric acid (HF) solution in a coagulation bath. This strategy provided the membrane with unique features along with a narrow and uniform pore size distribution (0.030 ± 0.005 μm). The etched membrane exhibits an increase of 2.3 times in pure water flux (PWF) and of 6.5 times in permeate flux(PF), with a slight decrease in rejection ratio (93.2% vs. 97%) when compared to than that of the un-etched membrane. Moreover, this membrane displayed outstanding antifouling ability, i.e., a flux recovery ratio (FRR) of 97% for 1000 mg/L bovine serum albumin (BSA) solution, a low irreversible fouling ratio of 0.5%, and highly enhanced hydrophilicity due to the formation of pores/voids throughout the membrane structure. The aforementioned features of the etched membrane indicate that the proposed method of etching SiO₂ NP in membrane matrix has a great potential to improve the structure and separation efficiency of a PVDF/PMMA/CA blend membrane.

In this paper, the structure optimized support vector machine (SVM) model was applied to predict the membrane permeate flux during dead-end microfiltration of activated sludge suspensions from sequencing batch reactor (SBR) with different experimental samples. The membrane permeate flux was considered as a function of mixed liquor suspended, temperature, dissolved oxygen, hydraulic retention time, transmembrane pressure, and operating time. Excellent agreements between the predicted values of SVM model and the experimental data demonstrated that SVM model has sufficient prediction accuracy. Furthermore, the results showed that the predicted values of SVM model agreed well with experimental data at different experimental samples in comparison with back propagation artificial neural network (BP-ANN) model. From the simulation results, the conclusion can be derived that SVM model outperforms BP-ANN model when the experimental samples sizes are small.