Feasibility and Statistical Analysis of Sulfanilic Acid Degradation in a Batch Photo-Fenton Process

Abstract

Sulfanilic acid (SA) is a representative sulfonated aromatic amine commonly found in industrial effluents, posing significant risks to both human health and the ecosystem. Efficient and cost-effective treatment of SA-containing wastewater is crucial for sustainable environmental management. This study investigates the performance of the photo-Fenton process in degrading SA-containing wastewater. Three process variables are selected to study their effects on percent total organic carbon (%TOC) removal and final pH (pH_Final): initial total organic carbon concentration (TOC₀) (150–250 mg/L), Fe²⁺ concentration (15–85 mg/L), and H₂O₂ concentration (1000–1500 mg/L). A combination of response surface methodology (RSM) and Box-Behnken design (BBD) is applied to examine both the individual and interactive effects of these variables. A total of 15 experimental trials are conducted, with the center point repeated three times. The results indicate significant interaction effects between Fe²⁺ and H₂O₂ concentrations on %TOC removal, while the interaction between TOC₀ and H₂O₂ concentration notably influences pH_Final. The optimal operating parameters to maximize %TOC removal within 45 min of operation are determined as a TOC₀ of 54.2 mg/L, an Fe²⁺ catalyst concentration of 33.7 mg/L, and an H₂O₂ concentration of 1403 mg/L. Under these conditions, the predicted %TOC removal and pH_Final were 89.2% and 2.93, respectively, which confirmed through validation experiments. Additionally, a correlation between pH_Final, TOC₀, and final TOC concentration (TOC_Final) is observed, leading to the development of a linear model capable of predicting TOC_Final based on TOC₀ and ${\mathrm{p}\mathrm{H}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}$ within the experimental space. The latter finding facilitates online monitoring of the process progress.

1. Introduction

Sulfonated aromatic amines are common pollutants in pharmaceutical and textile industry wastewater effluent, posing significant threats to human health and the environment due to their toxic, allergenic, mutagenic, and carcinogenic properties [1,2,3]. Sulfanilic acid (SA, 4-aminobenzenesulfonic acid) is a typical sulfonated aromatic amine, serving as a common intermediate in industries related to dyes, optical brighteners, pesticides, and sulfonamide antibiotics [4,5,6,7]. Substantial quantities of SA are introduced into water bodies during production and application. In addition, its relatively high water-solubility poses a challenge for efficient removal [7]. Therefore, developing an effective method for remediating SA-contaminated wastewater is crucial for the sustainable development of the industrial sector.

Aerobic biological treatment is the most cost-effective and environmentally friendly process [8]. However, the high chemical stability and low permeability of SA through bacterial membranes make it biorecalcitrant. Adsorption is commonly applied to remediate water containing non-degradable compounds [8]. Activated carbon, a typical adsorbent for sulfonated aromatic amines, demonstrates low adsorption capacity for SA due to its high water-solubility [9]. Additionally, challenges related to activated carbon regeneration and the generation of secondary pollutants hinder its widespread application [8].

Currently, advanced oxidation processes (AOPs) have received considerable attention for their ability to degrade complex organic compounds while minimizing negative impacts on ecosystems and public health. The degradation of organic pollutants during AOPs primarily occurs through the formation of hydroxyl radicals (HO^•) from hydrogen peroxide (H₂O₂). As a strong oxidizing species, HO^• radicals attack organic compounds non-selectively, transforming them into less harmful by-products such as water, carbon dioxide, and inorganic salts [2,10,11,12,13,14]. AOPs are promising technologies for treating persistent wastewater [15]. Various AOPs have been applied for the degradation of sulfonated aromatic amine compounds, including Fenton [16], Photo-Fenton [17,18], (Photo)Electro-Fenton [2], and Photocatalysis [19].

Research conducted by Khankhasaeva et al. (2020) demonstrated that heterogeneous catalytic photo-oxidation achieved a 100% conversion rate for sulfonated aromatic amine wastewater [17]. However, the total organic carbon (TOC) removal was limited to 52.3%, indicating relatively low mineralization efficiency [17]. Another study applied the Electro/Photoelectro-Fenton process to SA wastewater [2], achieving 95–98% TOC removal. However, the system complexity and its high maintenance demand due to the frequent electrode passivation, limits its real-world applicability. Limited research has been conducted on SA degradation using AOPs. Among them, the Fenton reaction is one of the most extensively studied. The Fenton reaction is initiated by ferrous ion (Fe²⁺) and H₂O₂, generating HO^• radicals and ferric ion (Fe³⁺). When pH < 3, this process becomes autocatalytic due to the regeneration of Fe²⁺ from Fe³⁺ through its reaction with H₂O₂, known as the Fenton-like reaction, outlined below [20]:

$$\mathrm{F}{\mathrm{e}}^{2+}+{\mathrm{H}}_2{\mathrm{O}}_2\to \mathrm{F}{\mathrm{e}}^{3+}+{\mathrm{H}\mathrm{O}}^{•}+{\mathrm{O}\mathrm{H}}^{-}$$

(1)

$$\mathrm{F}{\mathrm{e}}^{3+}+{\mathrm{H}}_2{\mathrm{O}}_2\to \mathrm{F}{\mathrm{e}}^{2+}+{\mathrm{H}\mathrm{O}}_2^{•}+{\mathrm{H}}^{+}$$

(2)

It is important to note that the Fenton-like reaction is significantly slower than the Fenton reaction, leading to the accumulation of Fe³⁺ species and limiting the overall reaction. Additionally, the produced hydroperoxyl radicals (${\mathrm{H}\mathrm{O}}_2^{•}$) exhibit lower oxidation power toward organic pollutants [21]. This obstacle can be addressed by introducing Ultraviolet (UV) radiation, which stimulates the reduction of Fe³⁺ to Fe²⁺ while simultaneously generating additional HO^• radicals, as described below [21]:

$$\mathrm{F}{\mathrm{e}}^{3+}+{\mathrm{H}}_2\mathrm{O}\stackrel{\mathrm{h}\mathsf{\nu }}{\to }\mathrm{F}{\mathrm{e}}^{2+}+{\mathrm{H}}^{+}+{\mathrm{H}\mathrm{O}}^{•}$$

(3)

Moreover, UV can decompose H₂O₂ through direct photolysis, as described in equation below, contributing to the formation of HO^• radicals [13,22]:

$${\mathrm{H}}_2{\mathrm{O}}_2\stackrel{\mathrm{h}\mathsf{\nu }}{\to }2\mathrm{H}{\mathrm{O}}^{•}$$

(4)

where h and $\mathsf{\nu }$ represent Planck’s constant and the light frequency, respectively. In comparison, the photo-Fenton process can operate with a lower concentration of Fe salts, reducing the amount of toxic metal released into the water and minimizing Fe sludge production. Additionally, the shorter degradation time enhances the process’s feasibility in terms of energy demand. The primary limitation of the photo-Fenton process is its sensitivity to pH, with an optimal range of 2.8–3.5 [15,23]. However, the acidic nature of SA aqueous solutions, with a natural pH of approximately 3 in the studied range (50–250 mg/L), compensates for the need for pH adjustment, further simplifying the process.

The objective of this research is to investigate the feasibility of the photo-Fenton process for degrading SA-containing wastewater without pH adjustment. The effects of various process variables, including initial TOC concentration (TOC₀) (mg/L), Fe²⁺ concentration (mg/L), and H₂O₂ concentration (mg/L), on percentage of TOC removal (%TOC) and final pH (pH_Final) are examined. The optimum operating conditions for maximizing %TOC removal within the studied space are determined. Furthermore, a statistical expression is proposed to predict the final TOC concentration (TOC_Final) based on pH_Final and TOC₀ within the experimental space. A combination of response surface methodology (RSM) and a Box–Behnken design (BBD) is employed to develop the statistical model, optimize operating conditions, and perform statistical analysis, which is achieved using Design-Expert software (Stat-Ease, Version 22.0.1).

2. Materials and Methods

2.1. Materials

Analytical-grade iron (II) sulfate heptahydrate (FeSO₄·7H₂O, 99%+, Sigma-Aldrich, Oakville, ON, Canada), TOC standard solution, (1000 µg/mL, Specpure, Thermo Scientific Chemicals, Barrington, IL, USA), pH buffer kit (Ward’s Science, Rochester, NY, USA), H₂O₂ (30% stabilized, ACS, VWR Chemicals BDH, Rouses Point, NY, USA), and reagent-grade SA (98%+, Thermo Scientific Chemicals, Mississauga, ON, Canada) were obtained from VWR Canada. For pH adjustments, sulfuric acid (H₂SO₄, 99%, Anachemia, Radnor, PA, USA) and sodium hydroxide (NaOH, 97%+, VWR Chemicals BDH, Mississauga, ON, Canada) were purchased from VWR Canada. The sample filtration is performed through a 0.45-micron non-sterile PVDF hydrophilic syringe filter (SF-PVDF-4513-S, Globe Scientific, Mahwah, NJ, USA). Phosphoric acid (H₃PO₄, ACS reagent, ≥85 wt%, JT Baker Chemicals, Phillipsburg, NJ, USA) was diluted in distilled water to prepare a 21% solution by volume. This solution is sparged into a sample to remove inorganic carbon (IC) and purgeable organic carbon (POC) before TOC analysis.

2.2. Preparation of Synthetic SA-Contained Wastewater

The concentration of SA in wastewater can vary considerably across different contaminated sources. Additionally, real process effluents exhibit varying concentrations of SA over time due to process disturbances, production fluctuations, and seasonal changes, even when an average concentration is observed. Therefore, the robustness of the model within a reasonable concentration range is important. Preliminary experiments have shown that the initial SA concentration plays a significant role in the overall degradation process, which has also been reported and examined as a key operational parameter in previous AOP studies [2,7,9,24]. In this study, the selected SA concentration levels are 50, 150, and 250 mg/L in terms of TOC concentration. This range was chosen to simulate typical SA-containing wastewater [1,7,8,25,26]. SA-containing wastewater was synthesized on-site by dissolving the proper amount of SA in distilled water. The TOC of the synthesized wastewater was confirmed using a TOC analyzer, along with its pH.

2.3. Analytical Method

2.3.1. pH Measurement

The pH measurements in this experiment were conducted using a potentiometric pH meter (Model HQ2200, Hatch, London, ON, Canada). A three-point calibration was performed daily using pH 4, 7, and 10 buffer solutions to ensure measurement accuracy and quality.

2.3.2. TOC Analysis

The degradation level of the SA solution was assessed through %TOC removal. The TOC content of all samples was determined using a combustion-type TOC analyzer (Model Apollo 9000, Teledyne Tekmar, Mason, OH, USA) operated under the direct method. In this method, IC is removed by adding 21% (v/v) H₃PO₄ to the sample, after which the acidified sample is purged with purified gas to eliminate dissolved inorganic carbon. The remaining organic carbon is quantified by oxidizing the purged sample in a high-temperature combustion furnace using Grade 4.5 oxygen. The resulting oxidized gases are transported by a carrier gas (O₂, Grade 4.5) through a condenser and a semi-permeable Nafion tube to remove water vapor. The dehydrated gas is then passed through a corrosive scrubber to remove any halides. Finally, the carbon dioxide (CO₂) produced during oxidation is detected by a nondispersive infrared (NDIR) detector. The NDIR detector measures the CO₂ concentration, and the results are converted into TOC readings, providing an accurate evaluation of the organic carbon content in the sample with a range of 4–25,000 mgC/L. The %TOC removal for each trial was calculated using the following equation:

$$\mathrm{\%}\mathrm{T}\mathrm{O}\mathrm{C} \mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}={\displaystyle \frac{{\mathrm{T}\mathrm{O}\mathrm{C}}_0-{\mathrm{T}\mathrm{O}\mathrm{C}}_{\mathrm{t}}}{{\mathrm{T}\mathrm{O}\mathrm{C}}_0}}\times 100$$

(5)

where TOC₀ and TOC_t represent the initial TOC concentration and TOC concentration at any time interval that is tested.

2.4. Design of Experiments and Analysis of Data

The RSM combined with a BBD was used to statistically evaluate the effects of process variables on the %TOC removal and pH change. Moreover, the potential of using TOC₀ concentration and pH_Final as an indicator of final TOC content is investigated. The required number of experiments using BBD is estimated as follow [27]:

$$\mathrm{N}=2\mathrm{k}\left(\mathrm{k}-1\right)+{\mathrm{c}}_{\mathrm{p}}$$

(6)

where N is the total number of experiments, k represents the number of factors and c_p denotes the number of central points. The independent variables and their studied levels are listed in

Table 1. Independent variables along with their corresponding symbol, coded, and actual range utilized in BBD.

Independent Variable	Representative Symbol	Coded Level and Actual Range
		−1	0	1
TOC0 Concentration (mg/L)	${\mathrm{x}}_1$	50	150	250
Fe2+ Concentration (mg/L)	${\mathrm{x}}_2$	15	50	85
H2O2 Concentration (mg/L)	${\mathrm{x}}_3$	500	1000	1500

.

In total, 15 trials were designed with the central point repeated 3 times. The details of each experiment are shown in

Table 2. BBD for three-factor three-level experiments alongside the observed and predicted %TOC removal and pH_Final for each run. x₁, x₂, and x₃ are TOC₀, Fe²⁺, and H₂O₂ concentrations, respectively.

Run	Independent Coded Variable			%TOC Removal		pHFinal
	x1	x2	x3	Observed y1	Predicted Y1	Observed y2	Predicted Y2
1	0	1	1	65.7	59.7	2.40	2.30
2	−1	0	1	87.8	93.0	2.94	2.94
3	0	0	0	61.0	64.1	2.26	2.29
4	1	1	0	23.2	24.7	2.14	2.16
5	−1	0	−1	86.1	81.5	2.94	2.9
6	0	1	−1	8.62	12.4	2.45	2.43
7	−1	−1	0	88.0	86.5	3.01	2.99
8	0	0	0	67.1	64.0	2.31	2.29
9	1	0	1	36.5	41.1	2.01	2.05
10	0	−1	1	37.6	33.9	2.39	2.41
11	0	−1	−1	19.2	25.3	2.49	2.54
12	1	0	−1	2.01	3.23	2.34	2.34
13	1	−1	0	5.36	4.53	2.27	2.21
14	−1	1	0	78.5	79.3	2.78	2.83
15	0	0	0	63.8	64.0	2.30	2.29

.

The recorded data on %TOC removal and pH_Final with respect to the process variables at 45 min of reaction time were fitted into a second-order model given by the following equation:

$$\mathrm{Y}={\mathrm{b}}_0+\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{b}}_{\mathrm{i}}{\mathrm{x}}_{\mathrm{i}}+\sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{b}}_{\mathrm{ii}}{{\mathrm{x}}_{\mathrm{i}}}^2+\sum _{\mathrm{i}=1}^{\mathrm{n}-1}\sum _{\mathrm{j}=2}^{\mathrm{n}}{\mathrm{b}}_{\mathrm{ij}}{\mathrm{x}}_{\mathrm{i}}{\mathrm{x}}_{\mathrm{j}}+\mathrm{e}$$

(7)

where Y represents the predicted response, n is the number of variables, and e is the residual term, indicating the deviation between the predicted and actual values. b₀, b_i, b_ii, and b_ij are the intercept regression coefficient, linear regression coefficient, quadratic regression coefficient, and interaction regression coefficient, respectively. The independent variables are represented by x_i and x_j. The coefficients are determined using the least squares method, which minimizes the sum of squared differences between the observed and predicted values. Design-Expert (Stat-Ease, Version 22.0.1) was used in the design of the experiment (DOE) to create a model, perform Analysis of Variance (ANOVA), and optimize the operating parameters.

2.5. Experimental Setup and Procedure

The structure of the photo-Fenton experimental setup is illustrated in Figure 1. The lab-scale batch photoreactor consists of a cylindrical glass beaker with an effective volume of 2000 mL. A germicidal lamp (Model PLS9W/TUV, LSE Lighting, Southampton, PA, USA) with a peak wavelength of 254 nm is positioned at the center of the photoreactor to ensure a homogeneous mixture. The same lamp was used within a two-week period, and any potential changes in the lamp’s intensity or wavelength over this period were considered negligible. The specifications of the lamp are listed in

Table 3. Characteristics of the UV-C lamp.

Parameter	Value
Lamp wavelength	UV-C, Large emission band centred at 254 nm
Lamp wattage	9 W
Lamp voltage	60 V
Length of lamp (base face)	129 mm
Model manufacturer	PLS9W/TUV
Manufacturer	LSE Lighting

. The photoreactor is submerged in an isothermal water bath to maintain a temperature of 25 °C during the reaction, with the reaction temperature monitored using a thermometer. Adequate agitation is provided by a magnetic stirrer to ensure uniform mixing. The entire setup is wrapped in aluminum foil to enhance UVC irradiation and minimize external disturbances. The photoreactor lid contained several ventilation holes to release any gas formed during the reaction, and the entire setup was placed under a fume hood. The stirrer geometry, rotational speed, and lamp positioning were not included as variables in this study and were kept constant across all experiments to ensure controlled operating conditions. Given the small scale of the laboratory setup, the system was assumed to behave as an ideal homogeneous batch photoreactor, with no significant mass transfer limitations arising from mixing or other fixed parameters. Therefore, the effects of the key process variables, including TOC₀, Fe²⁺, and H₂O₂ concentrations, were investigated, as these were identified from the literature and preliminary experiments as the most influential factors for evaluating the feasibility of the proposed process [2,16,17,28].

In each experiment listed in Table 2, 2 L of SA solution was prepared by dissolving the required amount of solid SA into distilled water. The prepared solution was tested for its TOC content and pH to confirm the desired quality of the stock solution. The solution was then loaded into the reactor, and the submerged UV lamp was activated and allowed to warm up for 30 min.

To confirm the necessity of the synergistic effect of UV/Fe²⁺/H₂O₂ in degrading SA, several preliminary experiments were conducted as listed in

Table 4. Summary of %TOC removal from preliminary experiments along with their conditions, TOC₀, Fe²⁺, and H₂O₂ concentrations.

Experimental Condition	TOC0 (mg/L)	Fe2+ (mg/L)	H2O2 (mg/L)	%TOC Removal
UV	150	N/A	N/A	N/A
UV/H2O2	150	N/A	1150	N/A
Fenton	150	46.8	1150	18.3
photo-Fenton	150	46.8	1150	78.8

. These experiments consist of different combinations of experimental conditions, including UV irradiation alone, UV/H₂O₂, Fenton (Fe²⁺/H₂O₂) and photo-Fenton process (UV/Fe²⁺/H₂O₂). The result indicated that UV irradiation alone or in combination with H₂O₂ was insufficient for degrading the wastewater, as the TOC removal was negligible. In contrast, the Fenton and photo-Fenton processes achieved TOC removals of 18.3% and 78.8%, respectively, demonstrating the significant enhancement of the Fenton process when combined with UV irradiation. These results highlight the necessity of the combined UV/Fe²⁺/H₂O₂ system to achieve efficient SA degradation.

To initiate the photo-Fenton experiments, the corresponding dosages of Fe²⁺ catalyst and H₂O₂ were added to the photoreactor after the warm-up period. The preliminary estimation of H₂O₂ concentration (mg/L) range was calculated based on the stoichiometric ratio, under the fundamental assumption of complete SA mineralization, as described in equation below:

$${\mathrm{C}}_6{{\mathrm{H}}_7\mathrm{N}\mathrm{O}}_3\mathrm{S}+{16\mathrm{H}}_2{\mathrm{O}}_2\to 6\mathrm{C}{\mathrm{O}}_2+{\displaystyle \frac{1}{2}}{\mathrm{N}\mathrm{H}}_4^{+}+{\displaystyle \frac{1}{2}}{\mathrm{N}\mathrm{O}}_3^{-}+{\mathrm{S}\mathrm{O}}_4^{2-}+{\displaystyle \frac{35}{2}}{\mathrm{H}}_2\mathrm{O}+2{\mathrm{H}}^{+}$$

(8)

The stoichiometric ratio of 16 determined in this study falls within the range of 15–18 reported in previous studies involving similar pollutants [16,17,28]. The preliminary estimation of H₂O₂ dosage was further adjusted with preliminary experiments, and concentrations of 500, 1000, and 1500 mg/L were selected for subsequent tests.

The experimental Fe²⁺ concentration range, 15–85 mg/L, was selected as based on previous studies [16,28,29,30] and validated through preliminary experiments. Subsequently, the corresponding amounts of FeSO₄·7H₂O were calculated according to the molar equivalent of Fe²⁺ required. To minimize the risk of catalyst overdosing, in addition to the Fe²⁺ concentrations, their corresponding Fe²⁺/TOC₀ ratios reported in the literature were calculated and presented in

Table 5. Summary of Fe²⁺ concentrations and Fe²⁺/TOC₀ ratios reported in the literature and the present study.

Fe2+ Concentration (mg/L)	Fe2+/TOC0 (w/w)	References
66.0–72.0	1.01–3.45	[16]
52.5–210	0.198–14.2	[28]
55.9	7.75	[29]
2.50–15.0	0.0260–0.633	[30]
15.0–85.0	0.060–1.70	Present Study

. These ratios were also considered in determining the Fe²⁺ concentration range used in this study. It is noteworthy that two relatively high values reported in the open literature correspond to Fe²⁺/TOC₀ ratios of 14.2 and 7.75 [28,29]. This discrepancy arises because Pariente et al. (2008) employed a heterogeneous catalyst [28], while Baran et al. (2009) performed photocatalysis in the absence of H₂O₂ [29]; neither study investigated the photo-Fenton process as applied in the present work.

Based on the results of preliminary experiments, samples were collected at 45, 60, 90, and 120 min of reaction time. Each sample was immediately tested for its pH, followed by quenching with 5 N NaOH to adjust the pH above 11. The quenching process ensures that the reaction is terminated, and the iron content is sufficiently precipitated. The quenched samples were then filtered through a 0.45-micron non-sterile PVDF hydrophilic syringe filter. The filtered samples were subjected to TOC analysis.

3. Results and Discussion

3.1. Data-Driven Prediction Model and ANOVA

One of the main objectives of this research is to predict the %TOC removal and pH_Final in response to different process variables, in the experimental domain. Therefore, %TOC removal (Y₁) and pH_Final (Y₂) were selected as the two key response variables, with respect to three process variables: TOC₀ concentration (x₁), Fe²⁺ concentration (x₂), and H₂O₂ concentration (x₃). The data collected at 45 min from each trial were fitted into quadratic models, which are presented in equations below:

$$\begin{gathered} {\mathrm{Y}}_1=64.0-34.2{\mathrm{x}}_1+3.23{\mathrm{x}}_2+13.7{\mathrm{x}}_3+6.84{\mathrm{x}}_1{\mathrm{x}}_2+8.19{\mathrm{x}}_1{\mathrm{x}}_3+9.67{\mathrm{x}}_2{\mathrm{x}}_3 \\ +2.53{{\mathrm{x}}_1}^2-17.8{{\mathrm{x}}_2}^2-13.4{{\mathrm{x}}_3}^2 \end{gathered}$$

(9)

$$\begin{gathered} {\mathrm{Y}}_2=2.29-0.364{\mathrm{x}}_1-0.0538{\mathrm{x}}_2-0.0650{\mathrm{x}}_3+0.0250{\mathrm{x}}_1{\mathrm{x}}_2-0.0825{\mathrm{x}}_1{\mathrm{x}}_3 \\ +0.00250{\mathrm{x}}_2{\mathrm{x}}_3+0.198{{\mathrm{x}}_1}^2-0.0625{{\mathrm{x}}_2}^2+0.0700{{\mathrm{x}}_3}^2 \end{gathered}$$

(10)

The 45 min interval was selected for modeling because this reaction time demonstrated the highest energy efficiency, as discussed later in this section.

All independent variables are represented by their coded value instead of the actual value to diminish the impact of the scale on the statistical model. The quality of the models can be visualized in Figure 2 and Figure 3 for %TOC removal and pH_Final response, respectively. The predicted values from the model were plotted against the corresponding experimentally determined values. The close alignment of data points with the 45-degree line indicates the precision and consistency of the model in predicting process responses.

The selection of 45 min operation time for the process was based on the calculated energy efficiency ratio (%/min) at different reaction times. This ratio, defined as the %TOC removal per unit of irradiation time (min), is shown in the equation below:

$$\mathrm{E}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y} \mathrm{E}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y} \mathrm{R}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\left({\displaystyle \frac{\mathrm{\%}}{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)={\displaystyle \frac{\mathrm{\%}\mathrm{T}\mathrm{O}\mathrm{C} \mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l}}{\mathrm{I}\mathrm{r}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{T}\mathrm{i}\mathrm{m}\mathrm{e}}}$$

(11)

The energy efficiency at different time intervals is presented in

Table 6. The energy efficiency ratio $\left({\displaystyle \frac{\mathrm{\%}}{\mathrm{m}\mathrm{i}\mathrm{n}}}\right)$ after 45, 60, 90, and 120 min reaction time.

Run	120 min	90 min	60 min	45 min	Optimum Time
1	0.801	1.07	1.45	1.46	45 min
2	0.792	1.05	1.52	1.95	45 min
3	0.788	0.962	1.37	1.36	60 min
4	0.302	0.358	0.514	0.516	45 min
5	0.806	1.05	1.48	1.91	45 min
6	0.173	0.156	0.152	0.192	45 min
7	0.823	1.10	1.58	1.96	45 min
8	0.779	1.01	1.37	1.49	45 min
9	0.623	0.801	0.738	0.810	45 min
10	0.790	0.800	0.648	0.836	45 min
11	0.243	0.306	0.354	0.427	45 min
12	0.176	0.128	0.157	0.045	120 min
13	0.183	0.166	0.146	0.119	120 min
14	0.803	1.06	1.50	1.74	45 min
15	0.786	1.02	1.38	1.42	45 min

.

In most trials, energy efficiency ratio at 45 min outperformed other time intervals, indicating that 45 min is the optimal operating time. Furthermore, solving the optimization model at 45 min resulted in a %TOC removal around 85.3%, demonstrating the effectiveness of SA mineralization within this time frame. In cases where higher %TOC removal is required, the operation time can be extended; however, this comes with a trade-off in cost-effectiveness due to reduced energy efficiency at longer durations.

ANOVA was applied to evaluate the model’s effectiveness. The ANOVA results for %TOC removal and pH_Final response models are presented in

Table 7. ANOVA results for the developed second-order model predicting %TOC removal in the photo-Fenton process for the degradation of SA.

		Sum of Squares	df	Mean Squares	F-Value	p-Value
TOC removal model (Y1)		1.359 × 104	9	1.510 × 103	33.71	0.0006	significant
${\mathrm{x}}_1$		9.319 × 103	1	9.329 × 103	2.083 × 102	<0.0001	significant
${\mathrm{x}}_2$		83.46	1	83.46	1.860	0.2305
${\mathrm{x}}_3$		1.559 × 103	1	1.559 × 103	34.80	0.0020	significant
${\mathrm{x}}_1{\mathrm{x}}_2$		1.871 × 102	1	1.871 × 102	4.180	0.0964
${\mathrm{x}}_1{\mathrm{x}}_3$		2.685 × 102	1	2.685 × 102	5.990	0.0581
${\mathrm{x}}_2{\mathrm{x}}_3$		3.740 × 102	1	3.740 × 102	8.350	0.0342	significant
${{\mathrm{x}}_1}^2$		23.55	1	23.55	0.5257	0.5009
${{\mathrm{x}}_2}^2$		1.164 × 103	1	1.164 × 103	25.99	0.0038	significant
${{\mathrm{x}}_3}^2$		6.659 × 102	1	6.659 × 102	14.87	0.0119	significant
Residual		2.240 × 102	5	44.80
Corrected total SS		13,814.08	14
R2		0.9840
Adjusted R2		0.9550
Predicted R2		0.7590
Adeq Precision		17.61

and

Table 8. ANOVA results for the developed second-order model predicting final pH value in the photo-Fenton process for the degradation of SA.

		Sum of Squares	df	Mean Squares	F-Value	p-Value
pH response model (Y2)		1.310	9	0.1453	42.29	0.0003	significant
${\mathrm{x}}_1$		1.060	1	1.060	3.082 × 102	<0.0001	significant
${\mathrm{x}}_2$		2.310 × 10−2	1	2.310 × 10−2	6.730	0.0486	significant
${\mathrm{x}}_3$		3.380 × 10−2	1	3.380 × 10−2	9.840	0.0258	significant
${\mathrm{x}}_1{\mathrm{x}}_2$		2.500 × 10−3	1	2.500 × 10−3	0.728	0.4326
${\mathrm{x}}_1{\mathrm{x}}_3$		2.720 × 10−2	1	2.720 × 10−2	7.930	0.0373	significant
${\mathrm{x}}_2{\mathrm{x}}_3$		0	1	0	7.300 × 10−3	0.9353
${{\mathrm{x}}_1}^2$		0.144	1	0.144	41.93	0.0013	significant
${{\mathrm{x}}_2}^2$		1.440 × 10−2	1	1.440 × 10−2	4.200	0.0958
${{\mathrm{x}}_3}^2$		1.810 × 10−2	1	1.810 × 10−2	5.270	0.0702
Residual		1.720 × 10−2	5	3.400 × 10−3
Corrected total SS		1.32	14
R2		0.9870
Adjusted R2		0.9637
Predicted R2		0.8071
Adeq Precision		19.77

, respectively. The R² values of the quadratic models for %TOC removal and pH_Final response are 0.984 and 0.987, respectively, demonstrating the effectiveness of the second-order models in explaining most data points. The small differences (<0.2) between predicted and adjusted R² for each expression highlight the model’s accuracy and consistency in predicting responses within the experimental range. The significance of different parameters is represented by the p-value, where a p-value < 0.05 indicates statistical significance, and a p-value > 0.05 is considered insignificant. It was observed that TOC₀, Fe²⁺ and H₂O₂ concentrations are statistically significant to pH_Final while only TOC₀ and H₂O₂ concentrations were significant to %TOC removal.

3.2. Interactive Effects of Fe²⁺ and H₂O₂ on %TOC Removal and pH_Final

Three-dimensional response surface plots of the process variables were generated based on the mathematical model across the entire experimental space. These plots are generated by holding one process variable constant while varying the other two, allowing for clearer visualization and a better understanding of how the process behaves across the experimental range.

Figure 4a and b illustrates the interaction effects of Fe²⁺ and H₂O₂ concentrations on %TOC removal and pH_Final, with the TOC₀ held constant at its mid-range value of 150 mg/L. Figure 4a demonstrates a curvature trend in %TOC removal as both Fe²⁺ and H₂O₂ concentrations increase. Specifically, increasing the H₂O₂ concentration from 500 to 1500 mg/L across the full range of Fe²⁺ concentrations result in a rising %TOC removal, peaking near the mid-range value of 1000 mg/L. Beyond this point, further increases in H₂O₂ concentration leads to a decline in removal efficiency. This behavior is primarily attributed to the scavenging effect of HO^• radicals at high H₂O₂ concentrations [31,32,33]. At lower H₂O₂ concentrations (500–1100 mg/L), the added H₂O₂ reacts under UV irradiation and in the presence of Fe²⁺ to generate HO^• radicals, which effectively oxidize and mineralize SA, thereby improving %TOC removal. However, at elevated H₂O₂ levels, excess H₂O₂ begins to scavenge the HO^• radicals, as shown in Equation (12), resulting in the formation of ${\mathrm{H}\mathrm{O}}_2^{•}$ instead. These ${\mathrm{H}\mathrm{O}}_2^{•}$ radicals possess a lower oxidation-reduction potential than HO^• radicals, thus reducing the overall degradation efficiency [34].

$${\mathrm{H}}_2{\mathrm{O}}_2+{\mathrm{H}\mathrm{O}}^{•}\to {\mathrm{H}\mathrm{O}}_2^{•}+{\mathrm{H}}_2\mathrm{O}$$

(12)

$$\mathrm{R}\mathrm{H}+{\mathrm{H}\mathrm{O}}_2^{•}\to {\mathrm{R}}^{•}+{\mathrm{H}}_2{\mathrm{O}}_2$$

(13)

where RH represents the organic pollutant and ${\mathrm{R}}^{•}$ is the organic radical produced from its reaction with ${\mathrm{H}\mathrm{O}}_2^{•}$. A similar trend is observed with increasing Fe²⁺ dosage. For instance, when H₂O₂ concentration and TOC₀ are fixed at their mid-range values, %TOC removal increases from 40.0% to 49.0% as Fe²⁺ concentration rises from 15 to 49 mg/L. However, beyond 49 mg/L, removal decreases to 41.0%. As noted earlier, Fe²⁺ catalyzes the reaction via Equations (2) and (3). The rise in Fe²⁺ concentration from 15 to 49 mg/L enhances catalytic activity and HO^• radical production, resulting in higher TOC removal. The distinctive behavior of the system after 49 mg/L of Fe²⁺ can be explained by two factors. One plausible explanation is that excessive iron leads to the scavenging effect of HO^• radicals by Fe²⁺ ions, as described below [35,36,37]:

$${\mathrm{F}\mathrm{e}}^{2+}+{\mathrm{H}\mathrm{O}}^{•}\to {\mathrm{Fe}}^{3+}+{\mathrm{OH}}^{-}$$

(14)

At high Fe²⁺ concentrations, excess Fe²⁺ ions scavenge HO^• radicals, thereby reducing the number of HO^• radicals available for the degradation of organic compounds. Similar scavenging effects have been reported in other studies [38,39,40,41]. Another contributing factor is that surplus Fe²⁺ ions form excess free radicals and Fe³⁺ ions, leading to a higher concentration of metal complexes due to complexation reactions [42]. These metal complexes contribute to a more persistent color change, will be explained in the next section, ultimately reducing the permeability of UV light and limiting its efficiency in generating HO^• radicals. Moreover, the complexes also hinder the regeneration of Fe²⁺ in Equation (3), resulting in lower catalytic activity and fewer HO^• radicals produced [42].

Figure 4b demonstrates a reversed behavior of pH_Final compared to %TOC removal, where increases in Fe²⁺ and H₂O₂ concentrations slightly decrease pH_Final, followed by a slight increase, showing an overall concave-down trend. This pattern can be attributed to the formation of acidic intermediates and final products, primarily carboxylic acids and H₂SO₄, during the decomposition of SA. Initially, enhanced degradation of SA at higher catalyst and H₂O₂ concentrations results in the generation of more H₂SO₄ and carboxylic acids. The slight increase in pH observed in Figure 4b, upon the introduction of excess Fe²⁺ and H₂O₂, occurs due to the scavenging effect of these reagents. This scavenging reduces the amount of HO^• radicals formed, thereby limiting degradation and lowering the concentration of acidic compounds in the medium. Although it has been suggested that the sulfate heptahydrate catalyst used in this experiment may affect pH, its influence was found to be negligible based on experimental observations. The p-value of 0.0342 in Table 7 indicates that the interaction between Fe²⁺ and H₂O₂ concentrations significantly impacts %TOC removal. In contrast, their interaction is statistically insignificant for pH_Final, with a p-value of 0.935.

3.3. Interactive Effects of TOC₀ and H₂O₂ on %TOC Removal and pH_Final

Figure 5a,b demonstrates the interaction effects of TOC₀ concentration and H₂O₂ concentration on %TOC removal and pH_Final, with Fe²⁺ dosage maintained at its mid-range value (50 mg/L). Figure 5a and b shows that, regardless of the H₂O₂ concentration, increasing TOC₀ leads to a significant reduction in both %TOC removal and pH_Final. At the lowest H₂O₂ dosage of 500 mg/L, %TOC removal drops from 86.0% to 5.00% as the TOC₀ concentration increases from 50 to 250 mg/L. This trend is attributed to the increased demand for HO^• radicals and the extended reaction time required to mineralize higher concentrations of SA. Correspondingly, it is observed that extending the experiment duration to 120 min enables most trials to achieve over 90% TOC removal, as shown in

Table 9. TOC₀ concentration, %TOC removal, and pH_Final at 45 and 120 min reaction time.

Run	TOC0	%TOC Removal (120 min)	%TOC Removal (45 min)	pHFinal (120 min)	pHFinal (45 min)	pHInitial
1	150	96.1	65.7	2.68	2.36	2.93
2	50	95.1	87.8	2.87	2.94	3.36
3	150	94.5	61.0	2.57	2.26	2.92
4	250	24.9	23.2	2.25	2.14	2.76
5	50	96.8	86.1	2.83	2.94	3.31
6	150	20.8	8.62	2.50	2.45	2.94
7	50	98.7	88.0	3.11	3.01	3.33
8	150	93.5	67.1	2.59	2.31	2.92
9	250	74.7	36.5	2.20	2.01	2.76
10	150	94.8	37.6	2.55	2.39	2.97
11	150	29.2	19.2	2.50	2.49	2.97
12	250	21.2	2.01	2.57	2.34	2.75
13	250	21.9	5.36	2.22	2.27	2.74
14	50	96.4	78.4	2.73	2.78	3.34
15	150	94.3	63.8	2.65	2.30	2.92

.

TOC₀ reflects the concentration of SA. At higher SA levels, more UV-C light is absorbed by the dissolved SA, limiting the UV-C available to be absorbed by H₂O₂. This leads to fewer HO^• radicals being generated, reducing the overall process efficiency. Notably, during the experiments, a visible color change from dark black to brown and eventually to transparent was observed, primarily associated with TOC₀. This color change is demonstrated in Figure 6.

According to Mijangos et al. (2006), this change is linked to the formation of dihydroxylated rings, which produce color-intense compounds such as benzoquinone [42]. Reactions involving these rings and their quinones further contribute to the color, even at low concentrations [42]. Additionally, colored metal complexes can form as Fe³⁺ ions interact with benzene derivatives, intensifying the observed color [42]. The extent of color change is largely influenced by the initial pollutant concentration, with minimal impact from the oxidant and catalyst [42]. The color, thus, can serve as an indicator of the oxidation level, as color-intense intermediates are progressively degraded into less colored compounds as the reactions proceed [42]. These mechanisms are consistent with the commonly reported intermediates in the degradation of SA through AOP-based methods, including hydroquinone, p-benzoquinone, and aliphatic carboxylic acids, along with their corresponding degradation pathways [2,16,19,28]. Although these color correlations are preliminary, they provide valuable insights that can serve as a basis for the proposed process mechanism, which can be further studied, analyzed, and evaluated in future research. The presence of these intermediates depends on the degree of conversion. According to the general pathway proposed in the literature [2,16,43,44], HO^• radicals initially attack the benzenic ring of SA, leading to the release of $\mathrm{N}{\mathrm{H}}_4^{+}$, $\mathrm{N}{\mathrm{O}}_3^{-}$, and $\mathrm{S}{\mathrm{O}}_4^{2-}$, and resulting in dihydroxylation of the aromatic ring to form hydroquinone. Hydroquinone is subsequently oxidized to p-benzoquinone, which then undergoes ring cleavage to produce aliphatic carboxylic acids. These aliphatic acids represent the ultimate organic intermediates that are further oxidized to inorganic end products, primarily CO₂. The proposed degradation pathway is illustrated in Figure 7.

In this study, it is observed that the duration of the dark phase is mainly influenced by the initial SA concentration. Higher TOC₀ leads to a longer dark phase (~5–25 min), decreasing UV light efficiency due to reduced transparency. In contrast, lower TOC₀ shows a shorter dark phase (<30 s), enabling more efficient utilization of UV light in producing HO^• radicals. Notably, even at the highest TOC₀, a 27.0% removal is achieved, demonstrating the feasibility of the photo-Fenton process for SA degradation. At the minimum TOC₀ (50 mg/L), increasing H₂O₂ results in a uniform curvature in %TOC removal, with values of 86.1% and 86.4% at the lowest and highest H₂O₂ dosages, respectively. Conversely, at the highest TOC₀ (250 mg/L), %TOC removal plateaus as H₂O₂ increases. This may be due to relative excess H₂O₂ at low TOC₀ concentration, leading to scavenging effects discussed in Section 3.2, while high TOC₀ demands more oxidant to reach peak efficiency.

Similar pattern is observed in Figure 5b, where pH decreases across all H₂O₂ concentrations as TOC₀ increases. At higher TOC₀ loads, more SA is present, resulting in a lower initial pH, as shown in Table 9. As degradation progresses, the reduction in SA concentration leads to a further decrease in pH. This is attributed to the formation of acidic organic intermediates and inorganic final products, mainly carboxylic acids and H₂SO₄, during SA degradation. During the experiments, a rapid pH drop is observed at the beginning, followed by a gradual rise once a certain level of %TOC removal is reached. Eventually, the pH stabilizes at a value higher than the initial pH. One possible explanation is that, as the reaction proceeds, most SA is first converted into more acidic organic intermediates, most likely benzenesulfonic acid as reported in [16,45]. These intermediates are then further broken down into less acidic organics as carboxylic acid [2,16], and inorganics such as CO₂ and H₂O, and partially into H₂SO₄, as suggested by Equation (8), as the solution approaches complete mineralization, the pH of the solution stabilize as the concentration of carboxylic acid decreases to zero. This behavior aligns with the observation that pH inversion typically occurs when TOC removal reaches 20.0–35.0%. In cases of high SA concentration (e.g., 250 mg/L) combined with relatively low H₂O₂ concentration (e.g., 500 mg/L), this inversion may not occur due to the continuous generation of acidic intermediates. Opposite to the trend observed in %TOC removal, Figure 5b shows that increasing H₂O₂ concentration results in an upward curvature in pH_Final. This is because the scavenging effect of H₂O₂ leads to variations in the amount of acidic intermediates and final products generated, as discussed previously in this section. The curvature is more pronounced at lower H₂O₂ concentrations, aligning with the trend observed in Figure 5a. According to Table 8, the interaction between TOC₀ and H₂O₂ concentration significantly influences pH_Final, with a p-value of 0.037. In contrast, their interaction is statistically insignificant for %TOC removal, with a p-value of 0.058. Notably, the p-value presented in Table 7 indicates that TOC₀ concentration, as an individual process variable, has the most significant impact on %TOC removal, with the lowest p-value of <0.0001. Similarly, TOC₀ concentration and its second-order term are statistically significant for pH_Final shown in Table 8, with p-values of <0.0001 and 0.0013, respectively. This suggests a higher-order correlation between TOC₀ and pH_Final.

3.4. Interactive Effects of TOC₀ and Fe²⁺ on %TOC Removal and pH_Final

Figure 8a and b presents the interaction effects of TOC₀ concentration and Fe²⁺ dosage on %TOC removal and pH_Final, with the H₂O₂ concentration held at its mid-range value of 1000 mg/L. The curvature behavior observed is a result of the scavenging effect of excessive Fe²⁺ catalyst, as discussed in previous sections, which leads to a nonlinear variation in both %TOC removal and pH_Final with increasing Fe²⁺ concentration. Noteworthy that the curvature shown in Figure 8b exhibits a high degree of homogeneity: the difference in %TOC removal between the two extreme Fe²⁺ concentrations (15 mg/L and 85 mg/L) is very similar at both the lowest and highest TOC₀ concentrations (50 mg/L and 250 mg/L), with values of 9.52% and 9.24%, respectively. This supports the earlier observation that the reduction in %TOC removal due to color change is primarily governed by TOC₀ concentration rather than Fe²⁺ dosage. Based on Table 7 and Table 8, the interactive effect between TOC₀ concentration and Fe²⁺ concentration is not significant for %TOC removal and pH_Final, with p-values of 0.096 and 0.433, respectively.

3.5. Potential Use of TOC₀ and pH_Final as Indicators for TOC_Final

TOC is the primary parameter in this study, as it serves as a direct indicator of the degree of mineralization. However, due to the inherent limitations of TOC measurement techniques, real-time monitoring is not feasible [46]. This constraint poses a significant challenge for the practical implementation of the studied process, particularly in AOPs, where fluctuations in operational conditions and external disturbances can significantly influence process performance. The findings presented in Section 3.2 and Section 3.3 indicate a strong correlation between pH and TOC concentration within the studied experimental range. This correlation suggests the possibility of developing a statistical predictive model to estimate TOC_Final based on pH_Final and TOC₀ concentration. Utilizing the collected experimental data, a linear regression model has been established to predict TOC_Final, as described in equation below:

$${\mathrm{T}\mathrm{O}\mathrm{C}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}(\mathrm{m}\mathrm{g}/\mathrm{L})=-643.0+1.710{\mathrm{T}\mathrm{O}\mathrm{C}}_0+195.2{\mathrm{p}\mathrm{H}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}$$

(15)

Excel (Version 2108) was employed to develop the first order model applying least squares method. The R² value of 0.943, as shown in

Table 10. ANOVA results for the linear model predicting TOC_Final according to ${\mathrm{T}\mathrm{O}\mathrm{C}}_0$ and ${\mathrm{p}\mathrm{H}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}$.

		Sum of Squares	df	Mean Squares	F-Value	p-Value
TOCFinal mg/L (Y)		9.064 × 104	2	4.532 × 104	99.16	<0.0001	significant
${\mathrm{T}\mathrm{O}\mathrm{C}}_0$		4.715 × 104	1	4.715 × 104	1.032 × 102	<0.0001	significant
${\mathrm{p}\mathrm{H}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}$		1.013 × 104	1	1.013 × 104	22.17	0.0005	significant
Residual		5.484 × 103	12	4.570 × 102
Corrected total SS		9.611 × 104	14
R2		0.9429
Adjusted R2		0.9334
Predicted R2		0.9197
Adeq Precision		26.85

, indicates a strong correlation pH_Final, TOC₀, and TOC_Final. Additionally, the accuracy and consistency of this linear model are confirmed by the difference between the predicted and adjusted R² values being less than 0.2. The p-values of TOC₀ and ${\mathrm{p}\mathrm{H}}_{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}$ are <0.0001 and 0.0005, respectively, indicating their statistical significance.

The actual TOC_Final versus their anticipated values using Equation (14) were plotted, as shown in Figure 9. The consistency of data points around 45$°$ line verified the accuracy and effectiveness of the model.

3.6. Optimization of Operating Conditions

In this study, the optimization objective is to maximize the %TOC removal by adjusting three process variables. Equation (9) is regarded as the objective function to maximize %TOC removal within the experimental range. The optimization solutions were generated using Design Expert (version 22.0.1) and based on the desirability criterion, the solution shown in

Table 11. Calculated optimum operation conditions to maximize the %TOC removal. Predicted and observed value for process response of %TOC removal and pH_Final along with process variables of TOC₀ concentration in mg/L (x₁), Fe²⁺ concentration in mg/L (x₂), and H₂O₂ concentration in mg/L (x₃).

Optimum Solution			Predicted Process Responses		Observed Process Responses
${\mathrm{x}}_1$	${\mathrm{x}}_2$	${\mathrm{x}}_3$	%TOC Removal	pHFinal	%TOC Removal	pHFinal
54.2	33.7	1403	89.2	2.93	85.3	2.92

was selected. The forecasted values of %TOC removal and pH_Final under the proposed optimal conditions aligned with the results from the validation experiment, with a percent error of less than ± 5.00%. At the optimum conditions, the %TOC removal and pH_Final response were determined as 85.3 percent and 2.92, respectively, yielding a percent error of 4.39% and 0.057 for %TOC removal and pH, in comparison to the anticipated values.

The work done by Kalinski et al. (2014), which employed the UV/H₂O₂ process to degrade SA, reported an optimal H₂O₂-to-SA molar ratio of 122 as shown in Table 1 [47]. In comparison, the molar ratio identified in this study for optimum solution is 54.7, indicating a significantly lower H₂O₂ demand in the photo-Fenton process and suggesting improved cost-effectiveness. Additionally, the optimal Fe²⁺/TOC₀ mass ratio determined in this study is compared with values reported in the literature listed in

Table 12. Summary of optimum Fe²⁺/TOC₀ and H₂O₂/TOC₀ ratios reported in the literature and the present study.

Optimum Fe2+/TOC0 (w/w)	References
0.279–1.12	[16]
4.65	[17]
0.152	[30]
0.619	Present Study
Optimum H2O2/TOC0 (mol/mol)	References
122	[47]
54.7	Present Study

, falling within the documented range of 0.152–4.65 on the lower side. Considering that Khankhasaeva et al. (2020) applied a heterogeneous catalyst [17], the lower optimum Fe²⁺/TOC₀ ratio observed in this study is reasonable since the Fe²⁺ catalyst is introduced in salt form to make a homogeneous system; in comparison, heterogeneous systems mass transfer is limited by factors such as particle size and porosity.

Intermediates are important indicators of the toxicity of the treated influent. As discussed in Section 3.3, various intermediates are formed throughout the degradation process, with their concentrations varying according to the extent of conversion. A study by El-Ghenymy et al. (2012) applying a Fenton-based process to SA demonstrated that, under optimal conditions, the concentrations of hydroquinone and p-benzoquinone peaked at approximately 5 and 3 min, respectively, and decreased to nearly zero after about 8 min with no significant change in the TOC of the solution [2]. Beyond this point, aliphatic carboxylic acids became the dominant intermediates, and their concentrations continued to increase until approximately 20% TOC removal was achieved [2]. This observation aligns with the color transition we observed between 5 and 25 min and the pH reversion occurring at TOC removals between 20 and 35%. Therefore, it is proposed that the predominant species present after achieving the optimum TOC removal (89.19%) are aliphatic carboxylic acids, which have been reported to exhibit lower aquatic toxicity toward Daphnia magna compared to SA, as summarized in

Table 13. Summary of EC₅₀ toxicity for SA, oxalic acid and formic acid.

Compound	Toxicity EC50 (mg/L)	References
SA	23	[48]
Oxalic acid	162.2	[49]
Formic acid	365	[50]

.

4. Conclusions

This study investigated the feasibility of the photo-Fenton process for degrading SA-rich wastewater without prior pH adjustment, with a primary focus on maximizing %TOC removal. A RSM approach, combined with the BBD, was employed to evaluate both the individual and interactive effects of TOC₀ concentration, Fe²⁺ concentration, and H₂O₂ concentration on the process efficiency. The statistical model, developed using experimental data and validated using ANOVA, was utilized as an objective function to optimize process parameters with the aim of maximizing %TOC removal within the experimental space. The optimal conditions were identified as 54.2 mg/L for TOC₀ concentration, 33.7 mg/L for Fe²⁺ concentration, and 1403 mg/L for H₂O₂ concentration, yielding a predicted %TOC removal of 89.2%. These conditions were confirmed through a validation experiment, demonstrating a variance of less than 5% between predicted and observed values. Compared to reported UV/H₂O₂ systems in the literature, the optimal parameters determined in this study require significantly lower H₂O₂ input, highlighting the improved cost-effectiveness of the photo-Fenton process. Additionally, a linear correlation model was developed within the experimental range to predict TOC_Final based on TOC₀ concentration and pH_Final, which will be helpful toward the online monitoring of process progress.

The findings from this study confirm the effectiveness of the photo-Fenton process in remediating SA-contaminated wastewater and highlight its potential as a viable treatment or pre-treatment method for such industrial effluents. The next recommended step of this work is to evaluate the potential application of the photo-Fenton process as a pretreatment method to enhance the biodegradability of SA, particularly in terms of biochemical oxygen demand. Furthermore, the collection and analysis of gas emissions during larger-scale experiments are suggested to support the development of a mechanistic model. Finally, LC–MS or HPLC analyses of experimental samples are recommended for investigating the reaction mechanism, identification and toxicity assessment of intermediate compounds formed during the process, and evaluation of reaction rate constants.