Photo-catalytic Ozonation for Degrading Terephthalic Acid in Aqueous Environment

Background & Aims of the Study: Terfetalic Acid (TPA ) is produced in large quantities and used in various industries. Besides, TPA is among the main sources of water pollution in industrialized countries. TPA photo-degradation process was performed in a Circulating Fluidized Bed Reactor (CFBR) by one Ultraviolet type A (UV-A) lamp and ozone generator with MnFe2O4/Willemite photo-catalyst. Materials and Methods: In this research, the nanoparticles of MnFe2O4 and Willemite were synthesized by co-precipitation reactions and wet mixing method, respectively. Then MnFe2O4/ Willemite was synthesized by the immobilization of MnFe2O4 on Willemite by mechanical method. Full factorial experimental design with 4 factors, including the pH, the initial concentration of TPA, the amount of Catalyst (Cat.), and O3 dosage was used for modeling and optimizing the process. Results: In the optimal conditions, the amounts of pH, TPA, Cat., and O3 were obtained equal to 9, 20 ppm, 1.5 g/L, and 2.17 mg/h, respectively. In these conditions, degradation efficiency was obtained to be 98.2695%, and decomposition kinetics was determined as pseudo-first-order with Kapp=0.2707 min-1, kLH=3.729 ppm min-1, and kadd=0.051 ppm-1. Conclusion: Comparing experiments results in different processes, such as UV, UV-Cat, O3, UVO 3, Cat-O3, and UV-Cat-O3 revealed that photo-catalytic ozonation (O3/MnFe2O4/Willemite) in the presence of UV for degradation of TPA in an aqueous environment, present the higher efficiency.


Introduction
upplying drinking water, energy, and food are the major future challenges; more importantly, water significantly influences the quality and quantity of energy, environment, and food [1]. Aromatic organic compounds, like Terephthalic Acid (TPA), are produced in high volume and are among the main sources of water pollution in industrialized countries.
TPA is widely used in numerous industries for producing polymers, such as Polyethylene Terephthalate (PET), polybutylene terephthalate, polyester, and polyamide [2]. Additionally, it is used for producing carriers in paint, plasticizer, drug, metal-organic framework, filler in some military smoke grenades, as well as deuterated terephthalate and related polymer [3,4]. Due to the very large-scale production of this substance in the world, there is always some of this substance in the environment. Moreover, according to research, this substance presents carcinogenic properties in rats [5] and reproductive toxicity in male mice [6].
There exist numerous methods to remove pollutants from water and wastewater, such as coagulation and flocculation, filtration, biological purification by microorganisms, reverse osmosis, Electro-Dialysis Reverse (EDR), ion exchange resins, adsorption, evaporation, and Advanced Oxidation Processes (AOP) [7][8][9]. Researchers found that each of these methods has disadvantages and advantages [7][8][9]. Implementing Advanced Nanoparticles (NPs) in AOP provides special opportunities for the recovery of wastewater. Specifications, such as the elimination of trace, toxic contaminants, microbial pathogens, resistant pollutants, the conversion of the wide range of organic materials to carbon dioxide and water, the low cost of equipment, the high rate of reactions, less energy consumption, and less human toxicity are benefits, i.e., likely to result from AOP [7,10]. Additionally, what has motivated the researchers to choose the AOP method is the ability of photo-catalyst to integrate the NPs with multifunction water purification systems. Moreover, using NPs has been developed in the field of water treatment [11].
Therefore, MnFe 2 O 4 was applied as the main catalyst to exploit its properties. The MnFe 2 O 4 was immobilized on a base to increase the specific surface area, raise the strength of the catalyst, facilitate its separation on large scale, reuse the catalyst, and reduce the cost of the operation [11,12]. Therefore, Willemite was selected as the basis, because of its excellent chemical and thermal stability, water resistance, and cost-effectiveness [13].
Statistical design techniques may be used to model and optimize the process. In the statistical design experiments, the factors involved in tests at their respective levels were simultaneously varied. One of the most widely used experimental design techniques is the full factorial method; this approach simultaneously investigates the effect of more than one factor in the response. Two-level full factorial designs of experiments are the most common patterns in which each factor is experimentally tested for only two levels [14].
Therefore, in this research, the degradation of TPA, as an aromatic pollutant was investigated by O 3 /MnFe 2 O 4 / Willemite, as photo-catalytic ozonation under the UV Irradiation process. The effects of pH, the initial concentration of TPA, the amount of catalyst, and O 3 dosage for higher degradation of TPA were investigated using a full factorial experimental design. Additionally, a kinetic study was developed using the experimental results attained in this investigation.
Numerous studies have been accomplished to eliminate TPA (Table 1) by photo-catalysts, such as TiO 2 , Fe, ZnO, and TiO 2 /AC [3,[15][16][17][18][19][20][21]. The process of removing contaminants by these photo-catalysts is not only time-consuming but also the mineralization of TPA and intermediates fail to occur absolutely. Furthermore, catalyst separation is a difficult task. The magnetic properties of MnFe 2 O 4 /Willemite allow the photo-catalyst to be separated and retrieved. It has certain characteristics, such as chemical stability, low cost, and very low bandgap energy [12].
In this study, the TPA removal process was examined by photo-catalytic methods (MnFe 2 O 4 /Willemite) in a Circulating Fluidized Bed Reactor (CFBR) by one UV-A lamp and ozone generator. To compare different processes, the TPA degradation experiments were performed by UV, UV-Cat., O 3 , UV-O 3 , Cat.-O 3 , and UV-Cat.-O 3 , separately. Full factorial experimental design with 4 factors was used for modeling and optimizing the process. In the optimal conditions, the amounts of pH, TPA, Cat., and O 3 were obtained. The relevant results suggested that the UV-Cat.-O 3 process has higher efficiency. Finally, the kinetic and reproducibility of photocatalytic was considered in optimum conditions.

Materials and Methods
Most required chemical reagents were provided from Sigma Aldrich and Merck Company. Agilent 8453 UV-visible S   99 [21] spectrophotometer was employed to appoint the presence of TPA in a sample and the percentage of degradation in aqueous media. In the degradation pollutant process, one UV-A lamp with a power of 15 W and ozone generator was obtained from Philips and Nabbzist Company (Iran), respectively. The decomposition of TPA was performed in a Circulating Fluidized Bed Reactor (CFBR) with rotary current and an effective volume of 1 liter. The structure and chemical properties of TPA are illustrated in Table 2.
The process of photocatalytic decomposition of TPA was performed by Circulating Fluidized Bed Reactor (CFBR) with rotary current and effective volume of 1 liter ( Figure 1). One UV-A lamp with a power of 15 W (Philips) was directly placed in the reactor so that the liquid flow revolves around the lamp. This reactor was continuous for ozone and batch for MnFe 2 O 4 /Willemite and TPA. An ozone generator from Nabbzist Company (Iran) was used in this process. A pressure capsule was applied to produce the pure oxygen for passing in the ozone generator. The photo-reactor was developed with a water-flow jacket joined to a Thermo-Bath (ALB64 model, FINEPCR Korean Company) for controlling the temperature at 25ºC in all experiments. The pH was measured by pH meter F-71 HORIBA (Germany).

Catalyst preparation
The preparation of Willemite was performed according to previous research [13,22]. Therefore, 0.1 moles of (ZnSO 4 .7H 2 O) was dissolved in 100 mL of deionized water. NaOH solution (0.2 moles NaOH/100ml H 2 O) was added drop-wise to zinc sulfate solution and stirred at 80°C. The solution was gradually cooled to ambient temperature. The precipitated Zn(OH) 2 was separated from the solution by centrifugation and dried at 100°C for 12 h. Moreover, 2 moles of Zn(OH) 2 and 1 mole of silica high purity powder were added to 200 mL water and stirred at 80°C for 4 h. Finally, the precipitate was separated by centrifuge, dried at 80°C, and calcined at 1050°C (Figure 2).
Based on previous research [23,24], 0.2 moles of (Fe(NO 3 ) 3 .9H 2 O) and 0.1 moles of (Mn(NO 3 ) 2 .4H 2 O) were completely dissolved in the 200 mL of distilled water. Then, urea solution (0.4 moles of urea in 250 mL water), as the precipitating agent, was added to the solution containing iron nitrate and manganese nitrate, drop  (Figure 2).
The experimental design was used to study the effects of simultaneously changing the values of variables on the removal of TPA from an aqueous solution. Additionally, the main factor influencing the process, the interaction manner of variables on each other, and the percentage of removal of TPA were determined by the experimental design method.
The designed experiment was created in Minitab 17 using a full factorial design [14]. Operational factors, containing an initial TPA concentration, ozone dosage, the amount of photo-catalyst, and pH are presented in Table  3. Operation factors and factor levels being influential on the response of the TPA degradation were confirmed by primary experiments. Other parameters, such as temperature, light wavelength, light intensity, and stirring rate were constant in the process. Corresponding to Equation 1, nineteenth runs (N) were designed based on 4 factors (k), 2 of factor levels, 3 of center points (C0), and without replicates of the corner points (r).
The randomization technique was selected to balance the effect of irrelevant or uncontrollable conditions and estimate the intrinsic variation. Besides, DOE was performed by coding factors corresponding to the design matrix.
Based on experimental data, there were significant relationships between the response (TPA degradation) and 4 input factors (Equation 2). New results will easily predict based on mathematical Equation 2. A regression analysis model was simplified by eliminating insignificant terms at P≤0.05. The adequacy of the model was inspected by examining residuals distribution and its independence. The main effect plot and interaction plot were employed to predict possible interactions. Finally, the response optimizer was applied to recognize the factor settings that optimize a TPA degradation percentage, cost, the sensitivity of factors, and desirability. Even- At any run, the TPA solution to a specific concentration, a particular pH, and a known amount of MnFe 2 O 4 / Willemite photo-catalyst was mixed in the tank. This solution was circulated in the reactor for 30 min in the absence of UV-A radiation to reach equilibrium. After homogenization, the radiation was started, and simultaneously, O 3 gas, created by an ozone generator, was injected from the bottom of the reactor by a spongy diffuser. The concentration of ozone in the gas phase was determined by the iodometric method, using 2% neutral buffered potassium iodide and sodium thiosulfate as a titrant [25]. The required samples were obtained at specified intervals, and the TPA concentration was measured by spectroscopy at 240 nm. The acidity was adjusted at the beginning of the experiments with dilute sodium hy-

Results
In the process of the mechanism of photo-degradation of TPA, MnFe 2 O 4 semiconductor supported the Willemite absorbs photons from light ( Figure 3). The photons which have energy equal (or more than) the photocatalyst band gaps, transfer electrons from the Valence Band (VB) to the Conduction Band CB). In this manner, holes and electrons are created in valance and conductive bonds, respectively. Then, the generated electrons  (1) Furthermore, hydroxyl radicals are created by the radiation of UV on ozone and hydrogen peroxide in solution (Reaction 6 and 7).

(10) OH•+R• → R ox
The percentages of TPA degradation for 19 sequential experiments regarding the influence of factors, the levels, and the combination of factors were applied for the execution of full factorial design ( Table 4). The comparison of P-value of the main factors and interaction factors, choosing α-level of 0.05 indicated that TPA, pH, O 3 , Cat., pH×TPA, pH×O 3 , TPA×Cat., TPA×O 3 , Cat.× O 3 , and pH×TPA×O 3 were statistically significant. In other words, the mentioned terms were effective with %95 probability. Moreover, the P-value of the lack of Fits (0.071) was more than the selected α-value (0.05), reflecting that the model accurately fitted the data ( Table 5).
The R 2 value emphasized that the model can explain 99.97% of the variance in TPA degradation yield. The ad-justed R 2 was measured as 99.92 %. Both values signify that the model fitted the data well. R 2 prediction was computed as 99.68% for the new observations. The standard deviation of the residuals (S) was equal to 0.46656. This value reflected the standard distance of data from the regression line.

The regression coefficients were obtained by Equation 4
based on un-coded data. In this pattern, Deg% is the percentage of TPA degradation yield or response; pH, Cat., TPA, and O 3 are factors; Equation 4 represents the relationship between the percentage of TPA degradation and predictor variables. In this equation, the overall mean score was equal to 68.097, and the absolute value of the coefficient expresses the relative strength of each factor: Deg%=68.097+15.404pH-4.170TPA+5.923Cat.+2.97O 3 -2820pH×TPA+2.724p The normal plot of the standardized effects ( Figure  4A) shows that all the main effects were significant. Besides, AB (pH by TPA conc.) and AD (pH by O3), as the interaction effects, presented the largest effect on the response, and BC (TPA by O3) provided the slightest effect on response. Furthermore, the normal plot proves that A, C, D, AD, BC, and BD positively impacted the response. In other words, the TPA degradation yields in-crease when the mentioned effects are altered from the low-level to the high levels.
Pareto chart of the standardized effects ( Figure 4B) displays the ranking of effects; there exist 10 significant effects on the TPA elimination processes.
Residual plots were considered to check the assumptions of normality, constant variance, and randomness in the final model ( Figure 5). The normal probability plot ( Figure 5A) illustrates that the data were normally distributed, and outliers did not exist in the data. Ad-  ditionally, residuals versus fitted values ( Figure 5B) established no outlets, and the normality assumption of residuals was valid. In the histogram ( Figure 5C), there was no distance between bars; thus, it signifies that outliers did not exist in the data. Residuals versus the order of the data ( Figure 5D), specified no systematic effects in the data due to the time or order of data collection. The structure of Figure 5D acknowledges no correlation between the residuals. Thus, all mentioned interpretations confirmed that the ANOVA data (Table 5) were reliable and the model can predict TPA degradation yield within +/-0.4665 with a 95% confidence level.   The settings that maximize degradation of TPA were determined by target parameters (goal=100 i.e., target, lower boundary=39.2, & upper boundary=100) in response optimizer for each response. The best-operating conditions which maximized the sensitivity of factors and desirability were obtained equal to 9, 20 ppm, 1.5 g/L, and 2.17 mg/h for pH, TPA, Cat., and O 3 , in sequence. The predicted value for photo-catalyst efficiency and desirability, based on the best factor levels, equaled 98.2695% and 0.97168, respectively.
The magnitude of the main effect for pH indicated the greater effect of it, compared to the other variables. Besides, the main effects plot proves that pH, Cat., and O 3 positively affected the response. Furthermore, TPA adversely impacted the response ( Figure 6A). The interaction plot for the percentage of TPA degradation ( Figure  6B) manifested several interactions, as lines are not parallel to each other.
The plot indicated the increase in the degradation of TPA (Deg%) while the amount of TPA changes from 50 ppm to 20 ppm depending on the extent of ozone. While ozone is low (dashed line), the change in degradation is more than the time when ozone is high (straight line). Additionally, enhanced degradation of TPA (Deg%), as moving from the low to the high level of catalyst, is greater while the ozone is lower than (straight line) the time when it is high (dashed line).
Moreover, the interaction plot indicates that the increase in the degradation of TPA (Deg%) when the pH changes from low to high depending on the amount of TPA and ozone. When the amount of TPA is high, the alternation in degradation is less than low TPA. Moreover, when O 3 dosage is high, the change in degradation is more than low O 3 conditions. To investigate the kinetics of TPA degradation, the experiments were performed by synthesized photo-catalysts in optimum conditions. Then, experimental data were fitted with Equation 5. Where A0 and A are the maximum absorptions at maximum absorption wavelength before light exposure and the maximum absorption at time t, respectively. Moreover, k represents the rate constant.
The kinetic equations ( To gather more evidence, a modified Langmuir-Hinshelwood (L-H) equation was applied (Equation 6). Where k add , k LH , and k are the apparent Langmuir adsorption constant, the apparent maximum photo-catalytic degradation rate, and the pseudo-first-order kinetic constant (derived from Equation 4), respectively. C0

Discussion
The contamination of water by aromatic compounds, like TPA, is among the greatest problems. Therefore, in this research, an applied, affordable, and efficient method was presented for the degradation of TPA. The high yield of the TPA photo-degradation process applying O 3 / MnFe 2 O 4 /Willemite as photo-catalytic ozonation under UV irradiation process corroborated the high efficiency of the employed method.
The study data disclosed that DOE, using full factorial design, is the best technique to ascertain the efficiency of O 3 /MnFe 2 O 4 /Willemite, to define influential factors, to minimize response variability, and to affect the uncontrollable variables, to reduce reaction time and total cost; consequently, to optimize response. In this regard, the TPA degradation process was successfully modeled employing a full factorial methodology. of the obtained ANOVA data were reliable; thus, the model can predict TPA degradation yield within +/-0.46656 and a 95% confidence level. The model with R 2 of 99.97%, Adj. R 2 of 99.92% and Pred. R 2 of 99.68% fitted the data well. Additionally, according to optimized results, the best level for operation factors of pH, TPA, Cat., and O 3 was obtained equal to 9, 20 ppm, 1.5 g/L, and 2.17 mg/h, in sequence. According to optimal conditions, the results of testing the catalyst indicated that the model is suitable and the photo-catalyst is efficient. In addition, the operation factors provided the highest sensitivity due to the efficiency of 98.2695% and the desirability of 0.97168.

Conclusion
TPA decomposition kinetics was determined pseudofirst-order. Furthermore, the repeatability of the results of the 5 tests of TPA degradation using the depreciated photo-catalyst of MnFe 2 O 4 /Willemite and its constant value indicated that the photo-catalyst is considerably reusable and economic. Therefore, based on the evidence, implementing this catalyst is recommended to TPA pro-ducer companies. Moreover, developing studies and using MnFe 2 O 4 /Willemite photo-catalyst in the removal of other aromatic pollutants are recommended.

Compliance with ethical guidelines
This article is a meta-analysis with no human or animal sample.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-profit sectors.