Photocatalytic Activity of Aeroxide TiO 2 Sensitized by Natural Dye Extracted from Mangosteen Peel

: Natural dye sensitizers are environment-friendly and inexpensive substances that could be used for photocatalytic decontamination of organic pollutants. In this study, a natural dye extracted from mangosteen peel, containing a signiﬁcant amount of anthocyanin dye, has been successfully employed to sensitize aeroxide TiO 2 to lower its bandgap, thereby making the process visible sunlight-driven. We have demonstrated the photocatalytic activity of mangosteen dye-sensitized-TiO 2 (MS-TiO 2 ) under visible solar light by studying the degradation of methylene blue (MB), a well-studied model compound. A multivariate parametric study was performed using factorial design methodology with three factors—pH, MS-TiO 2 dosage, and visible light intensity. The study indicated that pH and MS-TiO 2 dosage are the two most dominant factors for MB degradation under visible solar light. The kinetic rate constant and adsorption equilibrium constant were determined, and a Langmuir-Hinshelwood-type equation was proposed to describe MB degradation on MS-TiO 2 under visible solar light. Apparent quantum yield was also reported for the MS-TiO 2 photocatalyst at optimum experimental conditions.


Introduction
In an era where freshwater scarcity is turning into an environmental systemic risk, degradation and complete mineralization of organic contaminants are of immense significance. In the past few decades, photocatalysis using TiO 2 as a photocatalyst has been extensively studied for achieving this purpose due to its effectiveness in addressing a vast range of organics. High stability, corrosion resistance, non-toxicity, excellent optical transparency, and low cost of TiO 2 makes it one of the most widely used photocatalysts for water decontamination [1]. To date, it has been vastly studied for the removal of toxic organic pollutants [2][3][4], metal ion reduction [5], and degradation of emerging pollutants [1]. However, the wide bandgap of TiO 2 (~3.2 eV) makes the photocatalysis processes UV light-driven only [6][7][8]. As a result, only a small portion of the entire solar spectrum can be utilized for this purpose, thereby limiting the process of expensive UV irradiation. This necessitates the modification of conventional TiO 2 photocatalyst to lower its bandgap to visible light wavelengths.
Lowering bandgaps of photocatalysts by sensitizing it with a dye molecule is a technology that has been widely investigated for dye-sensitized solar cells (DSSC) for the past few years [7,[9][10][11]. The same technology has also been applied for photocatalytic hydrogen generation using dye-sensitized TiO 2 /Pt photocatalyst [12][13][14]. Hence, this provides an opportunity to use this dye sensitization technology for photocatalytic water decontamination processes as well [7,15,16]. In this sensitization process, the dye molecule gets excited by absorbing visible solar light, and an electron is transferred from its highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) [7]. This excited electron finally gets transferred to the conduction band of the semiconductor

•
To extract natural dye from mangosteen peels and prepare a novel TiO 2 -based natural dye-sensitized photocatalyst.
• Demonstrate the photocatalytic activity of MS-TiO 2 under visible solar light by studying the degradation of methylene blue (MB), a well-studied model compound.

•
Investigate the photocatalytic activity of MS-TiO 2 under varying experimental conditions such as solution pH, visible solar light intensity, and photocatalyst dosage-a statistical approach was adopted to understand the complete and combined effect of three independent variables on the photocatalyst efficiency. • Apparent quantum yield was calculated for MS-TiO 2 and compared the photocatalysis efficiency with literature values.

Anthocyanin Content in Mangosteen Peel
Anthocyanin dye undertakes reversible structural transformations with a change in pH demonstrated by significantly different absorbance spectra. The colored oxonium form dominates at pH 1.0 and the colorless hemiketal form at pH 4.5. The pH differential method is based on this reaction and permits accurate and rapid measurement of the total anthocyanins, even in the presence of polymerized degraded pigments and other interfering compounds [32]. In our case, we determined approximately 28 mg anthocyanin content per 100 g of dry mangosteen peel (mg cyanidin-3-glucoside/100 g). Anthocyanin content in our extracted mangosteen peel-based dye was very similar to literature values [33,34].

Characterization of the MS-TiO 2 Photocatalyst
A detailed characterization study of MS-TiO 2 was performed using XRD, SEM, EDX, and UV-Vis spectroscopy. From X-ray diffraction (XRD) studies of aeroxide TiO 2 and MS-TiO 2 (Figure 1), it was evident that there was no extra peak observed for MS-TiO 2 . Only the peak areas intensified for MS-TiO 2 , as compared to that of aeroxide TiO 2 , essentially signifying a change in surface structure in MS-TiO 2 compared to the latter. For further understanding of the morphology of MS-TiO 2 , scanning electron microscopy (SEM) studies were performed on both aeroxide TiO 2 and MS-TiO 2 at different levels of magnification. The following figure (Figure 2a,b) shows the SEM picture of aeroxide TiO 2 and MS-TiO 2 , respectively, at a magnification of 10,000 X. From these two figures, the morphological change in MS-TiO 2 is clearly observed and is more prominent than that of aeroxide TiO 2 . The surface for MS-TiO 2 became more compact with reduced pores than that of aeroxide TiO 2 . This change in the surface morphology of MS-TiO 2 compared to that of aeroxideTiO 2 signifies the adsorption of the dye on aeroxide TiO 2 surface, changing its surface properties.
Catalysts 2020, 10, x FOR PEER REVIEW 3 of 16 • Demonstrate the photocatalytic activity of MS-TiO2 under visible solar light by studying the degradation of methylene blue (MB), a well-studied model compound.

•
Investigate the photocatalytic activity of MS-TiO2 under varying experimental conditions such as solution pH, visible solar light intensity, and photocatalyst dosage-a statistical approach was adopted to understand the complete and combined effect of three independent variables on the photocatalyst efficiency.
• Apparent quantum yield was calculated for MS-TiO2 and compared the photocatalysis efficiency with literature values.

Anthocyanin Content in Mangosteen Peel
Anthocyanin dye undertakes reversible structural transformations with a change in pH demonstrated by significantly different absorbance spectra. The colored oxonium form dominates at pH 1.0 and the colorless hemiketal form at pH 4.5. The pH differential method is based on this reaction and permits accurate and rapid measurement of the total anthocyanins, even in the presence of polymerized degraded pigments and other interfering compounds [32]. In our case, we determined approximately 28 mg anthocyanin content per 100 g of dry mangosteen peel (mg cyanidin-3-glucoside/100 g). Anthocyanin content in our extracted mangosteen peel-based dye was very similar to literature values [33,34].

Characterization of the MS-TiO2 Photocatalyst
A detailed characterization study of MS-TiO2 was performed using XRD, SEM, EDX, and UV-Vis spectroscopy. From X-ray diffraction (XRD) studies of aeroxide TiO2 and MS-TiO2 (Figure 1), it was evident that there was no extra peak observed for MS-TiO2. Only the peak areas intensified for MS-TiO2, as compared to that of aeroxide TiO2, essentially signifying a change in surface structure in MS-TiO2 compared to the latter. For further understanding of the morphology of MS-TiO2, scanning electron microscopy (SEM) studies were performed on both aeroxide TiO2 and MS-TiO2 at different levels of magnification. The following figure (Figure 2a,b) shows the SEM picture of aeroxide TiO2 and MS-TiO2, respectively, at a magnification of 10,000 X. From these two figures, the morphological change in MS-TiO2 is clearly observed and is more prominent than that of aeroxide TiO2. The surface for MS-TiO2 became more compact with reduced pores than that of aeroxide TiO2. This change in the surface morphology of MS-TiO2 compared to that of aeroxideTiO2 signifies the adsorption of the dye on aeroxide TiO2 surface, changing its surface properties.   The energy-dispersive X-ray spectroscopy (EDX) study gives further information on the composition of both the catalysts. The result obtained from EDX is shown in the following table (Table  1 and S1). From this study, it is evident that the carbon weight % increased significantly more in MS-TiO2 than in aeroxide TiO2, which indicates the adsorption of the mangosteen dye onto the TiO2 surface. Finally, the bandgap of MS-TiO2 was measured using a UV-Vis spectrophotometer. As mentioned earlier, the large bandgap of anatase TiO2 (~3.2 eV) makes the photocatalytic process UV light-driven, where anatase TiO2 is being used as a photocatalyst [6][7][8]. However, for MS-TiO2, the bandgap was calculated to be ~2.95 eV (Supporting Information S2). The bandgap was calculated from the absorption wavelength determined by the UV-Vis spectrophotometer. Hence, it is evident that the lowering of bandgap was achieved by the mangosteen dye-sensitizing TiO2 molecule with the natural dye (anthocyanin dye), which drives the photocatalytic process toward visible sunlight.

Evaluation of the Photocatalytic Activity of MS-TiO2 under Visible Solar Light
Aeroxide TiO2 did not show any photocatalytic activity for MB degradation under visible solar light. This is because of TiO2′s bandgap of 3.2 eV, which did not allow for electron-hole pair formation under low-energy visible light, whereas MS-TiO2 could harvest visible solar light as evident from the MB degradation ( Figure 3). The main reason behind such photocatalytic reactions with visible photons can be explained with a dye sensitization mechanism. Dye-sensitized photodegradation of MB under only visible light was initiated through excitation of the dye molecule from its ground state to the excited state, which finally facilitated electron transfer to the conduction band of the semiconductor (TiO2). The oxidized dye molecule (dye ⊕ ) could interact with the pollutant (MB), or water to return to its ground state. The energy-dispersive X-ray spectroscopy (EDX) study gives further information on the composition of both the catalysts. The result obtained from EDX is shown in the following table (Table 1 and S1). From this study, it is evident that the carbon weight % increased significantly more in MS-TiO 2 than in aeroxide TiO 2 , which indicates the adsorption of the mangosteen dye onto the TiO 2 surface. Finally, the bandgap of MS-TiO 2 was measured using a UV-Vis spectrophotometer. As mentioned earlier, the large bandgap of anatase TiO 2 (~3.2 eV) makes the photocatalytic process UV light-driven, where anatase TiO 2 is being used as a photocatalyst [6][7][8]. However, for MS-TiO 2 , the bandgap was calculated to be~2.95 eV (Supporting Information S2). The bandgap was calculated from the absorption wavelength determined by the UV-Vis spectrophotometer. Hence, it is evident that the lowering of bandgap was achieved by the mangosteen dye-sensitizing TiO 2 molecule with the natural dye (anthocyanin dye), which drives the photocatalytic process toward visible sunlight.

Evaluation of the Photocatalytic Activity of MS-TiO 2 under Visible Solar Light
Aeroxide TiO 2 did not show any photocatalytic activity for MB degradation under visible solar light. This is because of TiO 2 s bandgap of 3.2 eV, which did not allow for electron-hole pair formation under low-energy visible light, whereas MS-TiO 2 could harvest visible solar light as evident from the MB degradation ( Figure 3). The main reason behind such photocatalytic reactions with visible photons can be explained with a dye sensitization mechanism. Dye-sensitized photodegradation of MB under only visible light was initiated through excitation of the dye molecule from its ground state to the excited state, which finally facilitated electron transfer to the conduction band of the semiconductor (TiO 2 ). The oxidized dye molecule (dye ⊕ ) could interact with the pollutant (MB), or water to return to its ground state.

Methylene Blue Photodegradation Kinetics and Apparent Quantum Yield
The photocatalytic degradation kinetics of numerous organic substrates have been studied in terms of Langmuir−Hinshelwood (L−H) rate equations [4,35]. Mostly in the literature, the L−H equation has been applied to describe the initial rate of degradation at time zero as a function of the initial substrate concentration. Chowdhury et al. incorporated an irradiation term in the L−H equation for their eosin Y-sensitized phenol degradation study, which made it very useful for the design, scale-up, and optimization of photocatalytic reactors [35]. We followed a similar approach for the MS-sensitized photodegradation of methylene blue, which suggests a kinetic rate equation in the form of a modified L−H rate equation, given by the expression: where Kapp is the apparent rate constant, KA is the adsorption equilibrium constant, I is the light intensity, β is a constant, and CMB is the methylene blue concentration. The superscript 0 indicates the initial concentration. The modified L−H model fitted very well with the experimental data, as reported in Figure 4. Kinetics constants are reported in Table 2.

Methylene Blue Photodegradation Kinetics and Apparent Quantum Yield
The photocatalytic degradation kinetics of numerous organic substrates have been studied in terms of Langmuir−Hinshelwood (L−H) rate equations [4,35]. Mostly in the literature, the L−H equation has been applied to describe the initial rate of degradation at time zero as a function of the initial substrate concentration. Chowdhury et al. incorporated an irradiation term in the L−H equation for their eosin Y-sensitized phenol degradation study, which made it very useful for the design, scale-up, and optimization of photocatalytic reactors [35]. We followed a similar approach for the MS-sensitized photodegradation of methylene blue, which suggests a kinetic rate equation in the form of a modified L−H rate equation, given by the expression: where K app is the apparent rate constant, K A is the adsorption equilibrium constant, I is the light intensity, β is a constant, and C MB is the methylene blue concentration. The superscript 0 indicates the initial concentration. The modified L−H model fitted very well with the experimental data, as reported in Figure 4. Kinetics constants are reported in Table 2.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 16  We also calculated the apparent quantum yield (ϕapp) to express the efficiency of the MS-  We also calculated the apparent quantum yield (φ app ) to express the efficiency of the MS-sensitized photocatalytic degradation processes. The photocatalytic degradation of MB could be expressed as follows: So, at an average light intensity (60 mW/cm 2 ), if only 420-650 nm range of wavelengths were considered, apparent quantum yields of 0.03% can be achieved.

Parametric Study of Photocatalytic Degradation
Univariate parametric studies were carried out to determine the effective experimental ranges of different reaction parameters for MS-sensitized TiO 2 -based photocatalysis. The parameters chosen for this study were the initial concentration of MB, incident visible light intensity, and MS-TiO 2 dosage. In each case, the % degradation of MB was measured using Equation (4). The details of all the univariate parametric studies are given in the Supporting Information (S3).

Statistical Analysis and Design of Experiments
To study the combined effects of different reaction parameters, a full factorial two-level three-factor (2 3 ) design of experiments was performed as detailed in Table 6 in Section 3.6. Since the effect of different reaction parameters cannot be compared if they are in different units, hence, there is a clear need to normalize all the parameters to a single unit. Thus, the variables were used in coded values for the modeling procedure. The normalization of the parameters was done by the following equation: where X is the value of a reaction parameter, X mid is the midpoint of the highest and lowest value of the parameter, and range is the difference between the highest and lowest value of the parameter. This is the only way to fit all the reaction parameters to a single polynomial equation.
The results obtained in the factorial design study are summarized in the following table (Table 3). Here, Y has been averaged for the duplicate runs. An empirical second-order polynomial equation was derived from the experimental design shown in Table 3 using a Minitab 17 software. The equation is given below. Y = 25.13 + 6.887A − 0.575B + 10.01C + 0.0875AB + 4.075AC − 0.4375BC + 0.225ABC (4) Here, Y represents the % degradation of MB following Equation (7), and A, B, and C represent the coded values of reaction parameters, as described in Table 6. From this model equation, it can be inferred that the independent variable A, C, and the interaction of A and C are highly significant since the values of their coefficients are considerably higher than the values of the other coefficients.
From the values of the coefficients of the model equation (Equation (4)), it can be said that the order of effect of independent variables on MB degradation is catalyst dosage (C) > pH (A) > light intensity (B). The effect of B is almost negligible as compared to A and C. Hence, following further regression analysis, the final model equation was derived as described below.
The R 2 value predicted for this regression model (Equation (5)) is 0.9965, which is in good agreement with the adjusted R 2 value of 0.9938. Moreover, the R 2 value for this model is very close to 1. Both two facts indicate that this model equation can be effectively used for the degradation of MB under the experimental conditions.
The following figure (Figure 5) also shows a perfect fit (R 2 > 98%) of the predicted responses with that of the experimental responses, which further validates the predictability of this model for MB degradation. The data used for the validation of this model are given in the Supporting Information (S4). The following figures (Figure 6a,b) show the response surface plot and two-dimensional contour plot, respectively, showing the combined effect of MS-TiO 2 dosage (C) and pH (A) on the degradation of MB (Y). From Figure 6a,b, it is evident that as the MS-TiO 2 dosage and pH is increased, the degradation of MB also increases under the experimental conditions.

Sensitivity Study
Finally, a sensitivity study was performed to determine the sensitivity of the reaction parameters toward the % degradation of MB (Y). The results are shown in the following table (Table 4). From this

Sensitivity Study
Finally, a sensitivity study was performed to determine the sensitivity of the reaction parameters toward the % degradation of MB (Y). The results are shown in the following table (Table 4). From this table (Table 4), it is observed that MS-TiO 2 dosage is the most sensitive reaction parameter where 10% decrease of MS-TiO 2 dosage can cause 47% decrease in Y. However, an increase in MS-TiO 2 dosage does not affect Y to that extent. The effect of pH becomes more pronounced when it is decreased by 10%, which causes the degradation of MB to drop to around 23%. The effect of change in visible light intensity is less sensitive when it is increased. Y remains almost unchanged with increasing visible light intensity. However, Y is decreased to around 20% when the visible light intensity is reduced. Lastly, the following table (Table 5) shows the degradation of MB under different photocatalytic processes. From this table (Table 5), it is evident that this method is the most environmentally friendly approach for degrading MB with considerably low reaction time and a high degradation rate.

Materials
Mangosteen fruit was procured from the local market. This fruit was selected for a dark purple colour. TiO 2 (Aeroxide P25: 80-20% anatase to rutile) was bought from Evonik Industries. Acetone was purchased from Sigma-Aldrich Company. MB (99% pure) was purchased from EM Science, Merck KGaA, Darmstadt, Germany. Ultra-pure water (18.2 MΩ·cm) was prepared from an in-house EASYPure RODI system (Thermo Scientific, Markham, ON, Canada). All chemicals used were analytical grade and used without any further purification.

Instruments
Solar simulator: A solar simulator was used to generate sunlight for photocatalysis. The solar simulator model (SS1KW, Sciencetech, London, ON, Canada) has the following specifications: 1000 W Xe arc lamp and an Air Mass (AM) 1.5 G filter; the simulator is capable of producing indistinguishable simulated 1 Sun irradiance of 100 mW/cm 2 at full power. To generate only visible light, a UV cut-off filter (Omega Optical, Brattleboro, VT, USA: λ ≥ 420 nm) was used. UV Spectrophotometer: A UV-3600 UV-VIS-NIR spectrophotometer from Mandel Scientific was used for MB analysis.

Preparation of Mangosteen Dye Extract
Peels from mangosteen cut into small pieces were air-dried for two days and stored in the dark for future use. 2 g of dried peels were slowly stirred overnight, under dark condition, in 200 mL of 1:1 acetone: water solvent. It was then centrifuged and vacuum filtered to get the mangosteen dye extract (Figure 7). The dye extract is orange in color. The solution was then filtered and transferred into a dark color glass bottle for further use.
Catalysts 2020, 10, x FOR PEER REVIEW 11 of 16 EASYPure RODI system (Thermo Scientific, Markham, ON, Canada). All chemicals used were analytical grade and used without any further purification.

Instruments
Solar simulator: A solar simulator was used to generate sunlight for photocatalysis. The solar simulator model (SS1KW, Sciencetech, London, ON, Canada) has the following specifications: 1000 W Xe arc lamp and an Air Mass (AM) 1.5 G filter; the simulator is capable of producing indistinguishable simulated 1 Sun irradiance of 100 mW/cm 2 at full power. To generate only visible light, a UV cut-off filter (Omega Optical, Brattleboro, VT, USA: λ ≥ 420 nm) was used. UV Spectrophotometer: A UV-3600 UV-VIS-NIR spectrophotometer from Mandel Scientific was used for MB analysis.

Preparation of Mangosteen Dye Extract
Peels from mangosteen cut into small pieces were air-dried for two days and stored in the dark for future use. 2 g of dried peels were slowly stirred overnight, under dark condition, in 200 mL of 1:1 acetone: water solvent. It was then centrifuged and vacuum filtered to get the mangosteen dye extract (Figure 7). The dye extract is orange in color. The solution was then filtered and transferred into a dark color glass bottle for further use.

Determination of Total Anthocyanins Content (TAC) in the Mangosteen Peel
A well-known pH differential method was used to determine the anthocyanin content of extracted mangosteen dye [32]. The anthocyanin dye concentration was expressed as cyanidin-3glucoside equivalents as follows: where A = (A520 nm-A700 nm) pH 1.0-(A520 nm-A700 nm) pH 4.5; MW (molecular weight of cyanidin-3-

Determination of Total Anthocyanins Content (TAC) in the Mangosteen Peel
A well-known pH differential method was used to determine the anthocyanin content of extracted mangosteen dye [32]. The anthocyanin dye concentration was expressed as cyanidin-3-glucoside equivalents as follows: where A = (A 520 nm -A 700 nm ) pH 1.0-(A 520 nm -A 700 nm ) pH 4.5; MW (molecular weight of cyanidin-3-glucoside) = 449.2 g/mol; DF is dilution factor, L is the cell path length for absorbance measurement in cm; ε (molar extinction coefficient of cyanidin-3-glucoside) = 26,900 L/mol/cm; and 1000 is the conversion factor from g to mg. Finally, the TAC was expressed as mg of cyanindin-3-glucoside equivalents per 100 g of dry mangosteen peel sample.

Preparation of Mangosteen Dye-Sensitized TiO 2 (MS-TiO 2 )
In the prepared mangosteen dye extract, 2 g of aeroxide TiO 2 powder was added and slowly stirred overnight under a dark condition. Next, it was centrifuged and vacuum dried at 70 • C for 2 h. It was then ground in a mortar-and-pestle to prepare the photocatalyst (MS-TiO 2 ). The white TiO 2 powder changed its color to yellowish-gray after treating it with dye extract (Figure 8). The prepared photocatalyst was then stored in the dark and in airtight vials for further uses.
Catalysts 2020, 10, x FOR PEER REVIEW 12 of 16 conversion factor from g to mg. Finally, the TAC was expressed as mg of cyanindin-3-glucoside equivalents per 100 g of dry mangosteen peel sample.

Preparation of Mangosteen Dye-Sensitized TiO2 (MS-TiO2)
In the prepared mangosteen dye extract, 2 g of aeroxide TiO2 powder was added and slowly stirred overnight under a dark condition. Next, it was centrifuged and vacuum dried at 70 °C for 2 h. It was then ground in a mortar-and-pestle to prepare the photocatalyst (MS-TiO2). The white TiO2 powder changed its color to yellowish-gray after treating it with dye extract (Figure 8). The prepared photocatalyst was then stored in the dark and in airtight vials for further uses.

Characterization of MS-TiO2 Photocatalyst
The bandgap of MS-TiO2 was determined using the UV-spectrophotometer to verify the bandgap shift of MS-TiO2 as compared to that of aeroxide TiO2 toward the visible range from the UV range. Chemical compositions of MS-TiO2 and aeroxide TiO2 were determined by the EDX study to confirm the adsorption of the mangosteen dye on the TiO2 surface. Finally, the change in the surface structure and morphology of MS-TiO2 compared to that of aeroxide TiO2 was analyzed through XRD followed by SEM at different magnifications.

Dye-Sensitized Photocatalysis Experiment
All the photocatalysis reactions were carried out in a Pyrex glass reactor (11 cm diameter, 6.3 cm height) with a flat window at the top for illumination, placed over a magnetic stirrer. As mentioned above (Section 3.2), solar light was irradiated by a solar simulator, where a UV cut-off filter (λ ≥ 420 nm) was used so that only the visible part of sunlight could be utilized. The volume of the reaction mixture was 150 mL, and all the reactions were carried out with constant aeration using an air pump. Dilute sodium hydroxide, and hydrochloric acid were used to adjust the pH of the reaction mixture. The reaction mixture (i.e., MB aqueous solution and MS-TiO2) was first stirred in the dark for 2 hrs to ascertain adsorption equilibrium and then irradiated with visible sunlight for an additional 2 hrs. Samples were collected and analyzed in a UV-Vis spectrophotometer after filtering through 0.45 μm

Characterization of MS-TiO 2 Photocatalyst
The bandgap of MS-TiO 2 was determined using the UV-spectrophotometer to verify the bandgap shift of MS-TiO 2 as compared to that of aeroxide TiO 2 toward the visible range from the UV range. Chemical compositions of MS-TiO 2 and aeroxide TiO 2 were determined by the EDX study to confirm the adsorption of the mangosteen dye on the TiO 2 surface. Finally, the change in the surface structure and morphology of MS-TiO 2 compared to that of aeroxide TiO 2 was analyzed through XRD followed by SEM at different magnifications.

Dye-Sensitized Photocatalysis Experiment
All the photocatalysis reactions were carried out in a Pyrex glass reactor (11 cm diameter, 6.3 cm height) with a flat window at the top for illumination, placed over a magnetic stirrer. As mentioned above (Section 3.2), solar light was irradiated by a solar simulator, where a UV cut-off filter (λ ≥ 420 nm) was used so that only the visible part of sunlight could be utilized. The volume of the reaction mixture was 150 mL, and all the reactions were carried out with constant aeration using an air pump. Dilute sodium hydroxide, and hydrochloric acid were used to adjust the pH of the reaction mixture. The reaction mixture (i.e., MB aqueous solution and MS-TiO 2 ) was first stirred in the dark for 2 h to ascertain adsorption equilibrium and then irradiated with visible sunlight for an additional 2 h. Samples were collected and analyzed in a UV-Vis spectrophotometer after filtering through 0.45 µm membrane filters. The concentration of MB was determined by UV-Vis spectrophotometer at λ max 664 nm [38][39][40]. Percent degradation of MB was calculated using the following equation [31]: where C 0 and C t are the concentration of MB at time 0 and t, respectively.

Experimental Design and Statistical Analysis
For the photocatalytic degradation of MB, first, univariate parametric studies were performed to determine the most effective reaction conditions. The parameters studied were the initial concentration of MB, MS-TiO 2 dosage, and visible light intensity. Based on this single factor study, the most feasible reaction conditions were determined for MB degradation.
Depending on the single factor study, the initial concentration of MB for further studies was fixed at a definite value. For the multivariate parametric study, three selected reaction parameters, e.g., pH, visible light intensity, and MS-TiO 2 dosage, were further studied using factorial design of experiments. A two-level three factorial design of experiment was followed, comprising of eight experiments and their duplicates, thereby amounting to 16 sets of experiments. This statistical approach was adopted to understand the complete and combined effect of three independent variables on the response, i.e., % degradation of MB. The ranges and levels of independent variables are shown in the following table (Table 6). Finally, the results obtained from 16 sets of experiments were fitted to an empirical quadratic polynomial model for three parameters in the form of the following equation: Y = m 0 + m 1 A + m 2 B + m 3 C + m 12 AB + m 23 BC + m 13 AC (8) In this equation (Equation (8)), Y is the response, i.e., % degradation of MB; m 0 is the intercept; m 1 , m 2 , and m 3 are coefficients of independent variables; m 12 , m 23, and m 13 are coefficients of interaction terms; and A, B, and C are three independent variables as shown in Table 6. The multivariate regression analysis was performed using Minitab 17 software. Followed by the regression analysis, a sensitivity study was performed to determine the relative sensitivities of reaction parameters on the response under experimental conditions.

Conclusions
From this study, we can conclude that the dyes extracted from mangosteen peel can be successfully utilized for sensitizing TiO 2 , thereby replacing heavy metal and synthetic organic dyes, and the produced novel photocatalyst can be effectively used for visible-solar light-driven photocatalytic degradation of organic pollutants. The extracted natural dye from mangosteen peel contained a significant amount of anthocyanin dye (28 mg anthocyanin content per 100 g of dry mangosteen peel) that was responsible for the sensitization of aeroxide TiO 2 , resulting in a lower bandgap photocatalyst (MS-TiO 2 ) compared to bare aeroxide TiO 2 . SEM and EDX studies also indicated the mangosteen dye adsorption on aeroxide TiO 2 and the subsequent morphological change of the MS-TiO 2 photocatalyst. MS-TiO 2 could harvest visible solar light, as evident from the MB degradation. The visible light-driven photodegradation kinetics of MB fitted well with a modified Langmuir−Hinshelwood rate equation where the apparent rate constant was 2.291 × 10 −6 1/min. The apparent quantum yield for such photocatalysis was estimated as 0.03% at 60 mW/cm 2 , considering a 420-650 nm wavelength range. Although the novel MS-TiO 2 photocatalyst showed low apparent quantum yield, the efficiency is either comparable or more competent for photocatalysis of MB under solar, visible, and UV light with various photocatalysts as mentioned in Table 6. A full factorial two-level design of experiment was successfully employed here to study the degradation of MB by solar light-driven photocatalysis using natural dye (mangosteen-based dye)-sensitized TiO 2 as the photocatalyst. The process shows a significantly high MB removal under the experimental conditions. From the regression model developed by Minitab17, MS-TiO 2 dosage was found to be the most influential factor for MB degradation under experimental conditions, followed by pH. The model was found to fit very well with experimental data, as confirmed by the high R 2 value and the validation experiments.