Next Article in Journal
Mechanical Behaviour of the Rotating Target SORGENTINA-RF
Previous Article in Journal
Life-Cycle Greenhouse Gas (GHG) Emissions Calculation for Urban Rail Transit Systems: The Case of Pernambuco Metro
Previous Article in Special Issue
Design of Pyrolysis Plant for Waste Methyl Ethyl Ketone from the Polarizer Manufacturing Process
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 13
Issue 15
10.3390/app13158966
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Effect of Textural Properties on the Degradation of Bisphenol from Industrial Wastewater Effluent in a Photocatalytic Reactor: A Modeling Approach

by
May Ali Alsaffar
1,*,
Mohamed Abdel Rahman Abdel Ghany
1,
Alyaa K. Mageed
1,
Adnan A. AbdulRazak
1,
Jamal Manee Ali
1,
Khalid A. Sukkar
1 and
Bamidele Victor Ayodele
2,3,*
1
Department of Chemical Engineering, University of Technology-Iraq, Baghdad 10066, Iraq
2
Department of Chemical Engineering, Universiti Teknologi PETRONAS, Seri Iskandar 32610, Perak, Malaysia
3
Center of Contaminant Control & Utilization (CenCou), Institute of Contaminant Management for Oil and Gas, Universiti Teknologi PETRONAS, Seri Iskandar 32610, Perak, Malaysia
*
Authors to whom correspondence should be addressed.
Appl. Sci. 2023, 13(15), 8966; https://doi.org/10.3390/app13158966
Submission received: 29 May 2023 / Revised: 24 June 2023 / Accepted: 6 July 2023 / Published: 4 August 2023
(This article belongs to the Special Issue Advances in Waste Treatment and Material Recycling)
Browse Figures
Versions Notes

Abstract

:
Conventional treatment methods such as chlorination and ozonation have been proven not to be effective in eliminating and degrading contaminants such as Bisphenol A (BPA) from wastewater. Hence, the degradation of BPA using a photocatalytic reactor has received a lot of attention recently. In this study, a model-based approach using a multilayer perceptron neural network (MLPNN) coupled with back-propagation, as well as support vector machine regression coupled with cubic kernel function (CSVMR) and Gaussian process regression (EQGPR) coupled with exponential quadratic kernel function, were employed to model the relationship between the textural properties such as pore volume (Vp), pore diameter (Vd), crystallite size, and specific surface area (SBET) of erbium- and iron-modified TiO2 photocatalysts in degrading BPA. Parametric analysis revealed that effective degradation of the Bisphenol up to 90% could be achieved using photocatalysts having textural properties of 150 m2/g, 8 nm, 7 nm, and 0.36 cm3/g for SBET, crystallite size, particle diameter, and pore volume, respectively. Fifteen architectures of the MPLNN models were tested to determine the best in terms of predictability of BPA degradation. The performance of each of the MLPNN models was measured using the coefficient of determination (R2) and root mean squared errors (RMSE). The MLPNN architecture comprised of 4 input layers, 14 hidden neurons, and 3 output layers displayed the best performance with R2 of 0.902 and 0.996 for training and testing. The 4-14-3 MLPNN robustly predicted the BPA degradation with an R2 of 0.921 and RMSE of 4.02, which is an indication that a nonlinear relationship exists between the textural properties of the modified TiO2 and the degradation of the BPA. The CSVRM did not show impressive performance as indicated by the R2 of 0.397. Therefore, appropriately modifying the textural properties of the TiO2 will significantly influence the BPA degradability.

1. Introduction

It is becoming increasingly common to find Bisphenol A [BPA] in the environment [1]. Globally, BPA is manufactured at a rate of three million tonnes annually [2,3]. It has wide application as a starting material for the production of epoxy and polycarbonate resins [4]. However, its production has been characterized by environmental pollution [5,6]. All environmental compartments, including air, water, and soil, are often contaminated by the indiscriminate disposal of BPA due to its widespread usage [7,8]. Pre- or postconsumer sources of BPA contamination dominate the environmental landscape [9]. The huge amount of BPA in the environment can be traced to wastes generated during manufacturing processes and transportation, as well as the wastes generated when BPA-based items are used [10,11]. BPA leaching into water is worrisome since water pipelines are covered in polymers like epoxy resins [12]. BPA leaching has been demonstrated to increase when hot water passes through these pipes [13].
BPA exposure is harmful to all living creatures, including humans [14]. One of the harmful impacts of BPA is endocrine disruption in humans [15]. Human exposure to high concentrations of BPA has been reported to be one of the causes of cancer and epigenetic alterations [16,17,18]. Furthermore, exposure to BPA during pregnancy has been related to autism [19]. Besides humans, aquatic lives have been reported to be highly affected by exposure to high concentrations of BPA. A study by Shirdel et al. [20] has shown that 100 mg/L of BPA produced oxidative damage in Atlantic salmon kidney cells. Long-term exposure, however, induced immune gene transcription. Fish gill tissue is also affected by 50 mg/L. BPA exposure during pregnancy affects invertebrates such as mice [20]. Exposure to BPA has also been reported to be detrimental to plants [21]. By reducing overall chlorophyll concentration in soybean seedlings, BPA inhibited photosynthesis [22].
The removal of BPA from various sources being contaminated has been accomplished by a variety of methods, including physical, chemical, and biological processes [23,24]. However, the ability to mineralize BPA makes photocatalytic degradation with light irradiation superior to the other methods [25]. To achieve an effective photodegradation of BPA in contaminated water, a highly efficient photocatalyst is required [26]. One of such highly effective photocatalyst is TiO2. TiO2 has received a great deal of interest in the degradation of various contaminants in the last few years due to its high photosensitivity [27]. As a low-cost photocatalyst, it has a substantial excitation binding energy and low threshold power for optical pumping, making it an ideal photocatalyst to degrade organic molecules in an aqueous solution [28]. The photocatalytic activity of TiO2 with a wide band gap of 3.2 eV is possible under ultraviolet and rarely edges out to the visible spectrum, though careful attempts have been made to change it. Doping TiO2 with nonmetals such as carbon and nitrogen as well as metals such as vanadium, chromium, and iron has been developed to achieve this purpose [29]. Three dopant groups, namely, carbon, iron, and erbium, have been researched extensively [30,31,32]. Nevertheless, it is unclear how the textural properties of the TiO2 photocatalyst influence its photocatalytic activity during the degradation of BPA. Authors such as Parida et al. [33] and Parayil et al. [34] reported that the modulation of TiO2 textural properties significantly influences its photocatalytic activities during hydrogen production. A data-driven approach can be employed to understand the nonlinear relationship between the textural properties of the TiO2 photocatalysts and the tendency to degrade the organic pollutant from the wastewater. Several studies reported in the literature have proven the robustness of data-driven approaches in modeling photocatalytic processes [35]. Ayodele et al. [36] employed 20 neural network architectures in modeling the photodegradation of chloramphenicol, phenol, azo dye, gaseous styrene, and methylene blue present in different sources of wastewater. The ANN algorithm with R2 > 0.9 was reported to have the best prediction of organic pollutant degradation. Liu et al. [37] employed machine learning algorithms such as linear regression, Random Forest, XGBoost, and LightGBM to model an accelerated design strategy of photocatalytic degradation activity prediction of the doped TiO2 photocatalyst. Based on the analysis of the results, the LightGBM showed outstanding performance in predicting the photocatalytic degradation activity of doped TiO2. An extensive review by Bhagat et al. [38] revealed that several machine learning techniques have been used for modeling the photocatalytic degradation of dye in wastewater. However, no study has investigated the effect of the textural properties of photocatalysts on the degradation of organic pollutants in wastewater. In this study, a modeling approach using a multilayer perceptron back-propagation neural network was employed to investigate the relationship between the textural properties of an erbium- and iron-modified TiO2 photocatalyst and the performance of the photocatalysts in the degradation of the BPA.

2. Data Acquisition and Model Configuration

The dataset used for testing the model configuration consist of various combinations of parameters representing the textural properties of the photocatalysts. These parameters include the pore diameter, the BET-specific surface area, the pore volume, and the crystallite size. The various combinations of the dataset were obtained using a central composite experimental design as detailed by Hou et al. [39]. Based on the CCD formulation, the Er and Fe amounts used for doping the TiO2 photocatalysts were varied together with the calcination temperature for the photocatalysts. The textural properties such as the BET-specific surface area, the photocatalyst particle size, and the pore volumes for each of the formulations were measured as detailed by Hou et al. [39]. The crystallite size for each of the combinations was obtained from the XRD analysis. The details of the textural data obtained for each of the photocatalyst formulations are reported by Hou et al. [40]. A typical experimental set-up shown in Figure 1 is often employed for contaminant photocatalytic degradation [40]. As reported by Hou et al. [39], the photocatalyst activity of Fe/Er-TiO2 in degrading the BPA from industrial wastewater was investigated in a photoreactor. Before the authors initiated the photocatalytic experiment, a 200 mL solution containing 200 mg photocatalysts and 10 mg/L of BPA was ultrasonically mixed for 5 min and magnetically stirred for 60 min. The photocatalytic degradation reaction was initiated by turning on the light source. After the photodegradation reaction, the solution was filtered using cellulose acetate syringe membrane filters and analyzed with an HPLC. Photocatalytic degradation of BPA was also examined in terms of photocatalyst dosage, light sources, initial BPA concentration, and aqueous matrix anions.
The configuration employed for this study consists of a multilayer perceptron neural network with a back-propagation algorithm (MLPNN), cubic support vector machine regression (SVMR), and exponential quadratic Gaussian process regression (EQGPR). The MLPNN, SVMR, and EQGPR were employed to model the relationship between the textural properties and the performance of modified TiO2 in degrading the BPA [41]. The MLPNN consists of three layers, namely, input, output, and hidden. The input layer receives the processed signal. The output layer performs tasks like prediction and categorization. One or more hidden layers between the input and output layers make up the MLP’s true computational engine. An MLP is a feed-forward network with data flowing from the input to the output layer. The MLP’s neurons are taught using back-propagation learning. For problems that cannot be solved linearly, MLPs can approximate continuous functions. In this study, several configurations of the MLPNN were tested to determine the architecture with the best performance. The support vector machine approach for binary response variables served as inspiration for the development of support vector regression [42]. The algorithm’s fundamental notion is to only make use of residuals whose absolute values are less than a certain threshold. Finding a well-fitting hyperplane in a kernel-induced feature space with acceptable generalization performance utilizing the original features is the basic idea behind the SVMR [43]. The CSVMR model was trained with the help of fitrsvm in the MATLAB environment to achieve higher levels of accuracy while working with low- to medium-dimensional datasets. In a GPR model, the anticipated value of the target variable is specified by the covariance function, which describes how the expected value varies when values are transformed throughout the input space [44]. In this case, a quadratic exponential covariance function with automated relevance determination was used. ARD refers to the inclusion of a length scale for each feature within the covariance function, which may be inspected after training to identify the relative relevance of that feature to prediction.
The coefficient of determination (R2) and root mean square error (RMSE) were employed to measure the performance of the MLPNN, CSVMR, and EQGPR in predicting the degradation of BPA [45]. The R2 was employed to determine how a change in one variable may be explained by a change in another. Invariably, the R2 is a measure of how much of the variation in y is explained by the x-variables. There is a range from 0 to 1. The closer the R2 to 1, the better the performance of the model. The RMSE is the standard deviation of the residuals. To understand how far data points are from a regression line, the residuals are examined. RMSE is a measure of how spread out these residuals are. The lower the RMSE, the better the predictability of the model. The predictability of the MLPNN is a determinant of how well the relationship between the textural properties of the modified TiO2 and the activity was modeled.

Data Visualization and Trend Analysis

In this study, the input variables are mostly the textural properties of the erbium- and iron-modified TiO2 photocatalyst. The texture properties examined include the specific surface area (SBET), the pore volume (Vp), the pore diameter (dp), and the crystallite size. The target variable was measured as a function of the photodegradation of the BPA. Three-dimensional plots shown in Figure 2 were employed to visualize the relationship between the input variables and the target variables. It can be seen in Figure 2a that there is a nonlinear interaction between the photocatalyst-specific surface area, the crystallite size, and the BPA degradation. A high BPA degradation potential can be observed at an SBET range of 120–140 m2/g and crystallite size of 4–8 nm, which is consistent with that reported in the literature [46]. Besides using TiO2-based photocatalysts, photocatalysts such as cobalt ferrite have been reported to photodegrade BPA in water at a high rate [47]. In addition, Figure 2b also illustrates the visualization of the interaction between the photocatalyst pore diameter, the crystallite size, and the BPA degradation. Also, a nonlinear interaction exists between the photocatalyst pore diameter, the crystallite size, and the BPA degradation, resulting in a high BPA degradation at a pore diameter range of 5–7 nm and crystallite size of 4–8 nm. Figure 2c depicts the visualization of the interaction relationship between the photocatalyst pore diameter, crystallite size, and high BPA degradation. It can be seen that a nonlinear interaction relationship also exists between the photocatalyst pore diameter and crystallite size, resulting in high BPA degradation. In Figure 2d, a nonlinear interaction relationship can be observed between the photocatalyst pore diameter, the pore volume, and the BPA degradation, which also resulted in a high BPA degradation. The three-dimensional plots show that the photocatalyst’s pore volume, the pore diameter, the pore volume, and the crystallite size significantly influence the BPA degradation. The effect of textural properties of an alumina-supported Ni catalyst for hydrogen production via methane decomposition has been reported by Ahmed et al. [48]. The results revealed that the textural properties of Ni/Al2O3 significantly influence the yield of hydrogen produced from methane decomposition. Similarly, González-Castaño et al. [49] reported the effect of textual properties in Ni-Fe catalysts used for CO2 methanation. The study established that there is a relationship between the surface area and the rate of CO2 methanation.

3. Performance Analysis of the Data-Driven Models

Several configurations of the MLPNN representing the input layer, the hidden layer, and the output layer were tested in modeling the relationship between the textural properties and the BPA degradation from the wastewater. The performance of each of the MLPNN configurations as a function of the RMSE and R2 is depicted in Figure 3. The artificial neuron in the hidden layer of the MLPNN was varied from 1 to 15 to determine the best configuration. As shown in Figure 3a, the RMSE obtained varies with the changes in the number of artificial neurons in the hidden neuron. Similarly, the R2 values also vary with changes in the artificial neurons in the hidden layer. The MLPNN with the configuration of 4 input nodes, 13 artificial neurons, and 3 output nodes displayed the best performance since the RMSE values are 4.2 and 4.6 for training and testing, respectively. R2 of 0.902 and 0.992 were obtained for the 4-13-3 MLPNN configuration during training and testing, respectively.
Using the best MLPNN configuration of 4-13-3, the performance of the model was tested as a function of the predictability of the targeted variable. As shown in Figure 4, the targeted variables to be tested were the BPA degradation and the rate of degradation of the BPA. The predictability of the MLPNN is a function of how well the nonlinear relationship between the input variables and the targeted variables can be trained and learned by the model. Figure 4 depicts the prediction of the BPA degradation and the rate of BPA degradation from the wastewater across the data points. It can be seen that at each data point, the actual values of the BPA degradation are consistent with MLPNN predicted values (Figure 4a). Similarly, the actual rate of BPA degradation at each data point is consistent with the MLPNN predicted values as shown in Figure 4b. A similar trend is observed for predicting the initial concentration of the BPA as shown in Figure 4c. This is an indication that the MLPNN model is robust in training and learning the nonlinear relationship between the input variables and the targets, resulting in a good prediction of the BPA degradation as well as the rate of BPA degradation. The robustness of the MLPNN model in predicting the BPA degradation and the rate of degradation can be attributed to the inbuilt features such as the back-propagation algorithms that help to minimize prediction errors [50]. The back-propagation algorithm helps the MLPNN model to improve its accuracy of the model [51]. With the help of the back-propagation algorithm, gradient descent is computed as a function of the weight [52]. Hence, the targeted outputs are compared with the predicted output, and the MLPNN network is positioned to adjust the associated connection weights with the hidden neurons in other to minimize the prediction errors [53,54]. The performance of the MLPNN in the prediction of the targeted output is consistent with the literature. The performance of MLPNN in modeling hydrogen production by catalytic steam methane reforming has been reported by Ayodele et al. [55]. The study revealed that the MLPNN robustly predicted hydrogen production with an R2 of 0.988 and RMSE of 0.437. Similarly, nanofluid viscosity has been accurately predicted using MLPNN with an R2 of 0.999. The MLPNN modeled the nonlinear relationship between the input parameters which include temperature, nanoparticle size, density, volume fraction and base fluid viscosity, and the nanofluid viscosity. Al-Haiqi et al. [56] reported a robust prediction of phenol degradation from wastewater using MLPNN. The MLPNN model trained using the Bayesian Regularization algorithm was capable of learning the nonlinear relationship between the irradiation time, initial phenol concentration, pH, photocatalyst dosage, and phenol degradation. The predictability of the MLPNN was evidenced by the RMSE of 1.126 and R2 of 0.999.
The performance analysis of the EQGPR model as a comparison between the actual and the predicted BPA degradation, actual and predicted BPA degradation rate, and actual and the predicted rate of initial BPA concentration is depicted in Figure 5. Figure 5a shows that the actual values of BPA degradation slightly agree with the predicted values using EQGPR at each data point. As indicated by R2 of 0.763 and RMSE of 15.819, the EQGPR is not robust enough to generalize the interrelationship between the textural properties and the BPA degradation efficiency. Similarly, in Figure 5b,c, there is a wide disparity between the actual and the predicted rate of BPA degradation from industrial wastewater. This is an indication that the EQGPR model is not robust enough to model the interconnectivity between the textural properties of the catalyst, the BPA degradation, and the initial concentration of the BPA. Although the EQGPR model did not display an impressive performance in modeling the prediction of the BPA degradation, the rate of BPA degradation, and the initial BPA concentration, studies have shown that the EQGPR is effective in modeling other processes. Zeng et al. [57] demonstrated the application of the GPR model in the prediction of building electricity usage. After comparing the observed real data to the predicted results, the suggested algorithms’ dependability and efficiency were shown to be conclusive. Data prediction accuracy with average deviations below 15% and minimal computation time are shown to be the sweet spot for which GPR may provide reasonable prediction on the energy usage of office buildings. The performance of the GPR could be improved by using different kernel functions. The kernel function is a technique for taking raw data and transforming it into the desired shape for processing.
Figure 6 depicts the performance of the CSVMR model in the form of a comparison between the actual and the predicted BPA degradation, the actual and predicted BPA degradation rate, and the actual and the predicted rate of initial BPA concentration. It can be seen that there is a wide disparity between the actual and the predicted values as indicated in Figure 6a–c. This can further be confirmed by the R2 of 0.397 and the RMSE values of 19.591, which is an indication that the CSVMR model is not robust enough to generalize the modeling of the relationship between the textural properties and the BPA degradation efficiency, the BPA degradation rate, and the initial BPA concentration. Nevertheless, the SVMR models have been successfully applied in modeling other processes. Huang et al. [58] demonstrated the robustness of the SVMR in modeling the heat exchanger performance in cryogenic oscillating flow conditions. With an R2 of 0.922 and a maximum error of 12.4%, the SVMR model displayed an accurate prediction. Heat transfer characteristics in cryogenic oscillating flow may be processed using the SVMR model, and the findings show that it is both an economical and precise approach. As a recommendation for further studies, the improvement of the SVMR used in this study can further be investigated using different types of kernel functions.

4. Comparison of the Best Model with the Literature

The MLPNN model with an R2 of 0.992 and RMSE of 4.21 displayed the best performance in the modeling of relationship between the photocatalyst textural properties and the activity during degradation of BPA. The MLPNN robustness in predicting the BPA degradation can be attributed to its capability to model complex nonlinear process and its tendency to make quick prediction after training. The performance of the MLPNN model compared with that of CatBoot, LightGBM, BPANN, CGCCN-MF-ANN, and AdaBoost reported in literature for modeling the photodegradation of various organic pollutants is summarized in Table 1. With an R2 of 0.992, the MLPNN models outperformed models like CatBoost, LightGBM, CGCCN-MF-ANN, and AdaBoost with R2 of 0.990, 0.928, 0.746, and 0.878, respectively. However, the MLPNN model displayed a slightly lower performance compared to the BPANN, which had an R2 of 0.999. The differences in the performance can be attributed to the nature of the network configuration, the training algorithms, and the datasets used for the modeling process.

5. Conclusions

This study examined the robustness of applying MLPNN to model the nonlinear relationship between various textural properties such as pore diameter, pore volume, specific surface area, crystallite size, and BPA degradation from wastewater. The parametric analysis and visualization show that there are nonlinear relationships between the textural properties of the photocatalysts and their tendency to photodegrade the BPA in the water. Data-driven models, namely, MLPNN, CSVM, and RQGPR, have been employed to model the relationship between textural properties and the degradation efficiency of BPA in industrial wastewater. Various configurations of the MLPNN, CSVM, and RQGPR were trained and tested to determine the one that best models the relationship between the input parameters and the targeted output. The MLPNN configuration with 4 input nodes, 13 artificial neurons in the hidden layer, and 3 output nodes in the output layer displayed the best performance, with R2 of 0.902 for training and 0.992 for testing, while RMSE values of 4.21 and 4.66 were obtained for the training and the testing, respectively. The 4-13-3 MLPNN trained using the Levenberg–Marquardt algorithm was robust in predicting the BPA degradation, the rate of BPA degradation, and the initial concentration of BPA in the wastewater due to its inherent ability to model complex nonlinear processes. Neither the CSVM nor RQGPR showed impressive performance as indicated by the low R2 values. Hence, the MLPNN algorithms could be subsequently employed to improve and optimize the performance of the photodegradation process.

Author Contributions

Conceptualization, M.A.A. and B.V.A.; methodology, B.V.A.; software, B.V.A.; validation, M.A.R.A.G., A.K.M. and A.A.A.; formal analysis, J.M.A.; investigation, K.A.S.; resources, M.A.A.; data curation, A.K.M.; writing—original draft preparation, B.V.A.; writing—review and editing, M.A.A., A.K.M. and A.A.A.; visualization, J.M.A. and K.A.S.; supervision, M.A.A. and B.V.A.; project administration, M.A.A. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

No new data were generated in this research.

Acknowledgments

The authors acknowledge the support of Department of Chemical Engineering University of Technology Iraq.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Diao, P.; Chen, Q.; Wang, R.; Sun, D.; Cai, Z.; Wu, H.; Duan, S. Phenolic endocrine-disrupting compounds in the Pearl River Estuary: Occurrence, bioaccumulation and risk assessment. Sci. Total Environ. 2017, 584–585, 1100–1107. [Google Scholar] [CrossRef] [PubMed]
  2. Liu, Y.; Gao, W. Photodegradation of Endocrine Disrupting Chemicals by ZnO Nanorod Arrays. Mol. Cryst. Liq. Cryst. 2014, 603, 194–201. [Google Scholar] [CrossRef]
  3. Sánchez-Avila, J.; Bonet, J.; Velasco, G.; Lacorte, S. Determination and occurrence of phthalates, alkylphenols, bisphenol A, PBDEs, PCBs and PAHs in an industrial sewage grid discharging to a Municipal Wastewater Treatment Plant. Sci. Total Environ. 2009, 407, 4157–4167. [Google Scholar] [CrossRef] [PubMed]
  4. Chiang, L.-F.; Doong, R.-A. Cu–TiO2 nanorods with enhanced ultraviolet- and visible-light photoactivity for bisphenol A degradation. J. Hazard. Mater. 2014, 277, 84–92. [Google Scholar] [CrossRef] [PubMed]
  5. Azizullah, A.; Khan, S.; Gao, G.; Gao, K. The interplay between bisphenol A and algae—A review. J. King Saud Univ.-Sci. 2022, 34, 102050. [Google Scholar] [CrossRef]
  6. Al-Qadri, A.A.Q.; Drmosh, Q.A.; Onaizi, S.A. Enhancement of bisphenol a removal from wastewater via the covalent functionalization of graphene oxide with short amine molecules. Case Stud. Chem. Environ. Eng. 2022, 6, 100233. [Google Scholar] [CrossRef]
  7. Yang, X.; Zhu, J.; Yang, D.; Zhang, J.; Guo, Y.; Zhong, X.; Kong, J.; Gu, J. High-efficiency improvement of thermal conductivities for epoxy composites from synthesized liquid crystal epoxy followed by doping BN fillers. Compos. Part B Eng. 2020, 185, 107784. [Google Scholar] [CrossRef]
  8. Sheikh Asadi, A.M.S.; Malakootian, M.; Kowsari, E.; Alidadi, H. Ionic liquid-assisted sol-gel synthesis of Fe2O3-TiO2 for enhanced photocatalytic degradation of bisphenol a under UV illumination: Modeling and optimization using response surface methodology. Optik 2020, 204, 164229. [Google Scholar] [CrossRef]
  9. Ncube, A.; Mtetwa, S.; Bukhari, M.; Fiorentino, G.; Passaro, R. Circular Economy and Green Chemistry: The Need for Radical Innovative Approaches in the Design for New Products. Energies 2023, 16, 1752. [Google Scholar] [CrossRef]
  10. Dueñas-Moreno, J.; Mora, A.; Cervantes-Avilés, P.; Mahlknecht, J. Groundwater contamination pathways of phthalates and bisphenol A: Origin, characteristics, transport, and fate—A review. Environ. Int. 2022, 170, 107550. [Google Scholar] [CrossRef]
  11. Wang, X.; Nag, R.; Brunton, N.P.; Siddique, M.A.B.; Harrison, S.M.; Monahan, F.J.; Cummins, E. Human health risk assessment of bisphenol A (BPA) through meat products. Environ. Res. 2022, 213, 113734. [Google Scholar] [CrossRef] [PubMed]
  12. Cantoni, B.; Cappello Riguzzi, A.; Turolla, A.; Antonelli, M. Bisphenol A leaching from epoxy resins in the drinking water distribution networks as human health risk determinant. Sci. Total Environ. 2021, 783, 146908. [Google Scholar] [CrossRef]
  13. Mohammadi, A.; Dobaradaran, S.; Schmidt, T.C.; Malakootian, M.; Spitz, J. Emerging contaminants migration from pipes used in drinking water distribution systems: A review of the scientific literature. Environ. Sci. Pollut. Res. 2022, 29, 75134–75160. [Google Scholar] [CrossRef] [PubMed]
  14. Penserini, L.; Cantoni, B.; Gabrielli, M.; Sezenna, E.; Saponaro, S.; Antonelli, M. An integrated human health risk assessment framework for alkylphenols due to drinking water and crops’ food consumption. Chemosphere 2023, 325, 138259. [Google Scholar] [CrossRef]
  15. Mustieles, V.; Arrebola, J.-P.; Porta, M. From old pollutants to the regulation of bisphenol A: Lessons learned for health promotion and disease prevention. Prev. Med. 2023, 169, 107460. [Google Scholar] [CrossRef] [PubMed]
  16. Postnikova, L.A.; Patkin, E.L. The possible effect of lactoferrin on the epigenetic characteristics of early mammalian embryos exposed to bisphenol A. Birth Defects Res. 2022, 114, 1199–1209. [Google Scholar] [CrossRef]
  17. He, H.; Li, X.; Shen, J.; Bai, S.; Li, C.; Shi, H. Bisphenol A exposure causes testicular toxicity by targeting DPY30-mediated post-translational modification of PI3K/AKT signaling in mice. Ecotoxicol. Environ. Saf. 2022, 243, 113996. [Google Scholar] [CrossRef]
  18. Li, Z.; Ren, Y.; Li, X.; Wang, W. KDM2A interacts with estrogen receptor α to promote bisphenol A and S-induced breast cancer cell proliferation by repressing TET2 expression. Ecotoxicol. Environ. Saf. 2023, 262, 115132. [Google Scholar] [CrossRef]
  19. Casas, M.; Forns, J.; Martínez, D.; Avella-García, C.; Valvi, D.; Ballesteros-Gómez, A.; Luque, N.; Rubio, S.; Julvez, J.; Sunyer, J.; et al. Exposure to bisphenol A during pregnancy and child neuropsychological development in the INMA-Sabadell cohort. Environ. Res. 2015, 142, 671–679. [Google Scholar] [CrossRef]
  20. Shirdel, I.; Kalbassi, M.R.; Esmaeilbeigi, M.; Tinoush, B. Disruptive effects of nonylphenol on reproductive hormones, antioxidant enzymes, and histology of liver, kidney and gonads in Caspian trout smolts. Comp. Biochem. Physiol. Part C Toxicol. Pharmacol. 2020, 232, 108756. [Google Scholar] [CrossRef]
  21. Liang, J.; Li, Y.; Xie, P.; Liu, C.; Yu, L.; Ma, X. Dualistic effects of bisphenol A on growth, photosynthetic and oxidative stress of duckweed (Lemna minor). Environ. Sci. Pollut. Res. 2022, 29, 87717–87729. [Google Scholar] [CrossRef] [PubMed]
  22. Mg, R.; Girigoswami, A. Bisphenol A-an Overview on its Effect on Health and Environment. Biointerface Res. Appl. Chem. 2021, 12, 105–119. [Google Scholar] [CrossRef]
  23. Bhatnagar, A.; Anastopoulos, I. Adsorptive removal of bisphenol A (BPA) from aqueous solution: A review. Chemosphere 2017, 168, 885–902. [Google Scholar] [CrossRef] [PubMed]
  24. Ohore, O.E.; Zhang, S. Endocrine disrupting effects of bisphenol A exposure and recent advances on its removal by water treatment systems. A review. Sci. Afr. 2019, 5, e00135. [Google Scholar] [CrossRef]
  25. Kataria, N.; Bhushan, D.; Gupta, R.; Rajendran, S.; Teo, M.Y.M.; Khoo, K.S. Current progress in treatment technologies for plastic waste (bisphenol A) in aquatic environment: Occurrence, toxicity and remediation mechanisms. Environ. Pollut. 2022, 315, 120319. [Google Scholar] [CrossRef]
  26. Tai, X.H.; Lai, C.W.; Yang, T.C.K.; Johan, M.R.; Lee, K.M.; Chen, C.-Y.; Juan, J.C. Highly effective removal of volatile organic pollutants with p-n heterojunction photoreduced graphene oxide-TiO2 photocatalyst. J. Environ. Chem. Eng. 2022, 10, 107304. [Google Scholar] [CrossRef]
  27. Xiao, C.; Wang, L.; Zhou, Q.; Huang, X. Hazards of bisphenol A (BPA) exposure: A systematic review of plant toxicology studies. J. Hazard. Mater. 2020, 384, 121488. [Google Scholar] [CrossRef]
  28. Bassim, S.; Mageed, A.K.; AbdulRazak, A.A.; Al-Sheikh, F. Photodegradation of Methylene Blue with Aid of Green Synthesis of CuO/TiO2 Nanoparticles from Extract of Citrus Aurantium Juice. Bull. Chem. React. Eng. Catal. 2023, 18, 1–16. [Google Scholar] [CrossRef]
  29. Nur, A.S.M.; Sultana, M.; Mondal, A.; Islam, S.; Robel, F.N.; Islam, A.; Sumi, M.S.A. A review on the development of elemental and codoped TiO2 photocatalysts for enhanced dye degradation under UV–vis irradiation. J. Water Process. Eng. 2022, 47, 102728. [Google Scholar] [CrossRef]
  30. Hou, D.; Feng, L.; Zhang, J.; Dong, S.; Zhou, D.; Lim, T.-T. Preparation, characterization and performance of a novel visible light responsive spherical activated carbon-supported and Er3+:YFeO3-doped TiO2 photocatalyst. J. Hazard. Mater. 2012, 199–200, 301–308. [Google Scholar] [CrossRef]
  31. Singh, K.; Harish, S.; Kristy, A.P.; Shivani, V.; Archana, J.; Navaneethan, M.; Shimomura, M.; Hayakawa, Y. Erbium doped TiO2 interconnected mesoporous spheres as an efficient visible light catalyst for photocatalytic applications. Appl. Surf. Sci. 2018, 449, 755–763. [Google Scholar] [CrossRef]
  32. Martins, P.M.; Gomez, V.; Lopes, A.C.; Tavares, C.J.; Botelho, G.; Irusta, S.; Lanceros-Mendez, S. Improving Photocatalytic Performance and Recyclability by Development of Er-Doped and Er/Pr-Codoped TiO2/Poly(vinylidene difluoride)–Trifluoroethylene Composite Membranes. J. Phys. Chem. C 2014, 118, 27944–27953. [Google Scholar] [CrossRef]
  33. Parida, K.; Satpathy, M.; Mohapatra, L. Incorporation of Fe3+ into Mg/Al layered double hydroxide framework: Effects on textural properties and photocatalytic activity for H2 generation. J. Mater. Chem. 2012, 22, 7350–7357. [Google Scholar] [CrossRef]
  34. Parayil, S.K.; Psota, R.J.; Koodali, R.T. Modulating the textural properties and photocatalytic hydrogen production activity of TiO2 by high temperature supercritical drying. Int. J. Hydrogen Energy 2013, 38, 10215–10225. [Google Scholar] [CrossRef]
  35. Mageed, A.K.; Abdel Ghany, M.A.; Alsaffar, M.A.; Ali, J.M. Modeling the Effect of Magnesia Nanoparticles on CO Hydrogenation to Light Olefins in a Continuous Flow Reactor Using Fine Gaussian Support Vector Machine. Asian J. Water Environ. Pollut. 2022, 19, 73–79. [Google Scholar] [CrossRef]
  36. Ayodele, B.V.; Alsaffar, M.A.; Mustapa, S.I.; Cheng, C.K.; Witoon, T. Modeling the effect of process parameters on the photocatalytic degradation of organic pollutants using artificial neural networks. Process. Saf. Environ. Prot. 2021, 145, 120–132. [Google Scholar] [CrossRef]
  37. Liu, Q.; Pan, K.; Lu, Y.; Wei, W.; Wang, S.; Du, W.; Ding, Z.; Zhou, Y. Data-driven for accelerated design strategy of photocatalytic degradation activity prediction of doped TiO2 photocatalyst. J. Water Process. Eng. 2022, 49, 103126. [Google Scholar] [CrossRef]
  38. Bhagat, S.K.; Pilario, K.E.; Babalola, O.E.; Tiyasha, T.; Yaqub, M.; Onu, C.E.; Pyrgaki, K.; Falah, M.W.; Jawad, A.H.; Yaseen, D.A.; et al. Comprehensive review on machine learning methodologies for modeling dye removal processes in wastewater. J. Clean. Prod. 2023, 385, 135522. [Google Scholar] [CrossRef]
  39. Hou, D.; Goei, R.; Wang, X.; Wang, P.; Lim, T.-T. Preparation of carbon-sensitized and Fe–Er codoped TiO2 with response surface methodology for bisphenol A photocatalytic degradation under visible-light irradiation. Appl. Catal. B Environ. 2012, 126, 121–133. [Google Scholar] [CrossRef]
  40. Ayodele, B.V.; Ghazali, A.A.; Yassin, M.Y.M.; Abdullah, S. Optimization of hydrogen production by photocatalytic steam methane reforming over lanthanum modified Titanium (IV) oxide using response surface methodology. Int. J. Hydrogen Energy 2019, 44, 20700–20710. [Google Scholar] [CrossRef]
  41. Benimam, H.; Si-Moussa, C.; Laidi, M.; Hanini, S. Modeling the activity coefficient at infinite dilution of water in ionic liquids using artificial neural networks and support vector machines. Neural Comput. Appl. 2020, 32, 8635–8653. [Google Scholar] [CrossRef]
  42. Mageed, A.K.; Shnain, Z.Y.; Mahdi, G.S. Modeling the effect of non-linear process parameters on the prediction of hydrogen production by steam reforming of bio-oil and glycerol using artificial neural network. Int. J. Energy Res. 2020, 44, 10523–10537. [Google Scholar] [CrossRef]
  43. Alwasiti, A.A.; Shnain, Z.Y.; Abid, M.F.; Abdul Razak, A.A.; Abdulhussein, B.A.; Mahdi, G.S. Experimental and numerical study on the degradation of mefenamic acid in a synthetic wastewater. IOP Conf. Ser. Earth Environ. Sci. 2021, 779, 012073. [Google Scholar] [CrossRef]
  44. Shnain, Z.Y.; Alwaiti, A.A.; Rashed, M.K.; Shakor, Z.M. Experimental and Data-driven approach of investigating the effect of parameters on the fluid flow characteristic of nanosilica enhanced two phase flow in pipeline. Alex. Eng. J. 2022, 61, 1159–1170. [Google Scholar] [CrossRef]
  45. Katsumata, H.; Kawabe, S.; Kaneco, S.; Suzuki, T.; Ohta, K. Degradation of bisphenol A in water by the photo-Fenton reaction. J. Photochem. Photobiol. A Chem. 2004, 162, 297–305. [Google Scholar] [CrossRef]
  46. Cai, C.; Kang, S.; Xie, X.; Liao, C.; Duan, X.; Dionysiou, D.D. Efficient degradation of bisphenol A in water by heterogeneous activation of peroxymonosulfate using highly active cobalt ferrite nanoparticles. J. Hazard. Mater. 2020, 399, 122979. [Google Scholar] [CrossRef]
  47. Ahmed, W.; Noor El-Din, M.R.; Aboul-Enein, A.A.; Awadallah, A.E. Effect of textural properties of alumina support on the catalytic performance of Ni/Al2O3 catalysts for hydrogen production via methane decomposition. J. Nat. Gas Sci. Eng. 2015, 25, 359–366. [Google Scholar] [CrossRef]
  48. Ahmed, Y.; Yaakob, Z.; Akhtar, P.; Sopian, K. Production of biogas and performance evaluation of existing treatment processes in palm oil mill effluent (POME). Renew. Sustain. Energy Rev. 2015, 42, 1260–1278. [Google Scholar] [CrossRef]
  49. González-Castaño, M.; de Miguel, J.C.N.; Penkova, A.; Centeno, M.A.; Odriozola, J.A.; Arellano-Garcia, H. Ni/YMnO3 perovskite catalyst for CO2 methanation. Appl. Mater. Today 2021, 23, 101055. [Google Scholar] [CrossRef]
  50. Zeng, Y.-R.; Zeng, Y.; Choi, B.; Wang, L. Multifactor-influenced energy consumption forecasting using enhanced back-propagation neural network. Energy 2017, 127, 381–396. [Google Scholar] [CrossRef]
  51. Sahu, S.; Shera, S.S.; Banik, R.M. Artificial Neural Network Modeling to Predict the Non-Linearity in Reaction Conditions of Cholesterol Oxidase from Streptomyces olivaceus MTCC 6820. J. Biosci. Med. 2019, 7, 14–24. [Google Scholar] [CrossRef] [Green Version]
  52. Goldt, S.; Advani, M.S.; Saxe, A.M.; Krzakala, F.; Zdeborová, L. Dynamics of stochastic gradient descent for two-layer neural networks in the teacher–student setup*. J. Stat. Mech. Theory Exp. 2020, 2020, 124010. [Google Scholar] [CrossRef] [PubMed]
  53. Alsaffar, M.A.; Ayodele, B.V.; Mustapa, S.I. Scavenging carbon deposition on alumina supported cobalt catalyst during renewable hydrogen-rich syngas production by methane dry reforming using artificial intelligence modeling technique. J. Clean. Prod. 2020, 247, 119168. [Google Scholar] [CrossRef]
  54. Bahadar, A.; Kanthasamy, R.; Sait, H.H.; Zwawi, M.; Algarni, M.; Ayodele, B.V.; Cheng, C.K.; Wei, L.J. Elucidating the effect of process parameters on the production of hydrogen-rich syngas by biomass and coal Co-gasification techniques: A multi-criteria modeling approach. Chemosphere 2022, 287, 132052. [Google Scholar] [CrossRef]
  55. Ayodele, B.V.; Mustapa, S.I.; Kanthasamy, R.; Zwawi, M.; Cheng, C.K. Modeling the prediction of hydrogen production by co-gasification of plastic and rubber wastes using machine learning algorithms. Int. J. Energy Res. 2021, 45, 9580–9594. [Google Scholar] [CrossRef]
  56. Al Haiqi, O.; Nour, A.H.; Ayodele, B.V.; Bargaa, R. Bayesian Regularization-Trained Multi-layer Perceptron Neural Network Predictive Modelling of Phenol Degradation using ZnO/Fe2O3 photocatalyst. J. Phys. Conf. Ser. 2020, 1529, 052058. [Google Scholar] [CrossRef]
  57. Zeng, A.; Ho, H.; Yu, Y. Prediction of building electricity usage using Gaussian Process Regression. J. Build. Eng. 2020, 28, 101054. [Google Scholar] [CrossRef]
  58. Huang, J.; Jin, T.; Liang, M.; Chen, H. Prediction of heat exchanger performance in cryogenic oscillating flow conditions by support vector machine. Appl. Therm. Eng. 2021, 182, 116053. [Google Scholar] [CrossRef]
  59. Jaffari, Z.H.; Abbas, A.; Lam, S.-M.; Park, S.; Chon, K.; Kim, E.-S.; Cho, K.H. Machine learning approaches to predict the photocatalytic performance of bismuth ferrite-based materials in the removal of malachite green. J. Hazard. Mater. 2023, 442, 130031. [Google Scholar] [CrossRef]
  60. Ayodele, B.V.; Alsaffar, M.A.; Mustapa, S.I.; Vo, D.-V.N. Back-propagation neural networks modeling of photocatalytic degradation of organic pollutants using TiO2-based photocatalysts. J. Chem. Technol. Biotechnol. 2020, 95, 2730–2749. [Google Scholar] [CrossRef]
  61. Jiang, Z.; Hu, J.; Tong, M.; Samia, A.C.; Zhang, H.; Yu, X. A Novel Machine Learning Model to Predict the Photo-Degradation Performance of Different Photocatalysts on a Variety of Water Contaminants. Catalysts 2021, 11, 1107. [Google Scholar] [CrossRef]
  62. Navidpour, A.H.; Hosseinzadeh, A.; Huang, Z.; Li, D.; Zhou, J.L. Application of machine learning algorithms in predicting the photocatalytic degradation of perfluorooctanoic acid. Catal. Rev. 2022, 1–26. [Google Scholar] [CrossRef]
Applsci 13 08966 g001
Figure 1. Setup illustrating the photocatalytic degradation of the Bisphenol from industrial wastewater effluent.
Figure 1. Setup illustrating the photocatalytic degradation of the Bisphenol from industrial wastewater effluent.
Applsci 13 08966 g001
Applsci 13 08966 g002
Figure 2. Interaction effect of (a) specific surface area and crystallite size, (b) pore diameter and crystallite size, (c) pore volume and crystallite size, and (d) pore volume and particle diameter on the BPA degradation.
Figure 2. Interaction effect of (a) specific surface area and crystallite size, (b) pore diameter and crystallite size, (c) pore volume and crystallite size, and (d) pore volume and particle diameter on the BPA degradation.
Applsci 13 08966 g002
Applsci 13 08966 g003
Figure 3. (a) The RMSE of each of the MLPNN configurations; (b) the R2 of each of the MLPNN configurations.
Figure 3. (a) The RMSE of each of the MLPNN configurations; (b) the R2 of each of the MLPNN configurations.
Applsci 13 08966 g003
Applsci 13 08966 g004aApplsci 13 08966 g004b
Figure 4. (a) Comparison between the actual and the MLPNN predicted BPA degradation; (b) comparison between the actual and the MLPNN predicted BPA degradation rate; (c) comparison between the actual and the MLPNN predicted rate of initial BPA concentration.
Figure 4. (a) Comparison between the actual and the MLPNN predicted BPA degradation; (b) comparison between the actual and the MLPNN predicted BPA degradation rate; (c) comparison between the actual and the MLPNN predicted rate of initial BPA concentration.
Applsci 13 08966 g004aApplsci 13 08966 g004b
Applsci 13 08966 g005
Figure 5. (a) Comparison between the actual and the EQGPR predicted BPA degradation; (b) comparison between the actual and the EQGPR predicted BPA degradation rate; (c) comparison between the actual and the EQGPR predicted rate of initial BPA concentration.
Figure 5. (a) Comparison between the actual and the EQGPR predicted BPA degradation; (b) comparison between the actual and the EQGPR predicted BPA degradation rate; (c) comparison between the actual and the EQGPR predicted rate of initial BPA concentration.
Applsci 13 08966 g005
Applsci 13 08966 g006aApplsci 13 08966 g006b
Figure 6. (a) Comparison between the actual and the CSVMR predicted BPA degradation; (b) comparison between the actual and the CSVMR predicted BPA degradation rate; (c) comparison between the actual and the CSVMR predicted rate of initial BPA concentration.
Figure 6. (a) Comparison between the actual and the CSVMR predicted BPA degradation; (b) comparison between the actual and the CSVMR predicted BPA degradation rate; (c) comparison between the actual and the CSVMR predicted rate of initial BPA concentration.
Applsci 13 08966 g006aApplsci 13 08966 g006b
Table 1. Comparison of the best model with literature.
Table 1. Comparison of the best model with literature.
ProcessBest ModelPerformanceReference
Photodegradation of BPAMLPNNR2 = 0.992
RMSE = 4.21		This study
Photodegradation of malachite greenCatBoostR2 = 0.990, RMSE = 1.34[59]
Photodegradation of organic pollutantsLightGBMR2 = 0.928, RMSE = 0.194[37]
Photodegradation of organic pollutantsBPANNR2 = 0.999, RMSE = N. R[60]
Photodegradation of organic pollutantsCGCCN-MF-ANNR2 = 0.746 and RMSE of 0.293[61]
Photodegradation of perfluorooctanoic acidAdaBoostR2 = 0.878 and RMSE = 10.33[62]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Alsaffar, M.A.; Ghany, M.A.R.A.; Mageed, A.K.; AbdulRazak, A.A.; Ali, J.M.; Sukkar, K.A.; Ayodele, B.V. Effect of Textural Properties on the Degradation of Bisphenol from Industrial Wastewater Effluent in a Photocatalytic Reactor: A Modeling Approach. Appl. Sci. 2023, 13, 8966. https://doi.org/10.3390/app13158966

AMA Style

Alsaffar MA, Ghany MARA, Mageed AK, AbdulRazak AA, Ali JM, Sukkar KA, Ayodele BV. Effect of Textural Properties on the Degradation of Bisphenol from Industrial Wastewater Effluent in a Photocatalytic Reactor: A Modeling Approach. Applied Sciences. 2023; 13(15):8966. https://doi.org/10.3390/app13158966

Chicago/Turabian Style

Alsaffar, May Ali, Mohamed Abdel Rahman Abdel Ghany, Alyaa K. Mageed, Adnan A. AbdulRazak, Jamal Manee Ali, Khalid A. Sukkar, and Bamidele Victor Ayodele. 2023. "Effect of Textural Properties on the Degradation of Bisphenol from Industrial Wastewater Effluent in a Photocatalytic Reactor: A Modeling Approach" Applied Sciences 13, no. 15: 8966. https://doi.org/10.3390/app13158966

APA Style

Alsaffar, M. A., Ghany, M. A. R. A., Mageed, A. K., AbdulRazak, A. A., Ali, J. M., Sukkar, K. A., & Ayodele, B. V. (2023). Effect of Textural Properties on the Degradation of Bisphenol from Industrial Wastewater Effluent in a Photocatalytic Reactor: A Modeling Approach. Applied Sciences, 13(15), 8966. https://doi.org/10.3390/app13158966

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop