A R T I C L E  I N F O		ABSTRACT
ORIGINAL ARTICLE		Introduction: The use of arsenic contaminated water can cause a variety of adverse health effects in humans. Therefore, it is essential to seek out a method to remove arsenic more efficiently. This study examined the amount of arsenic oxidation by response surface methodology (RSM) based on Box-Bencken design. Materials and Methods: In this study, oxidizing arsenite to arsenate was performed by activation of persulfate with UV and the optimal conditions determined using the RSM based on Box-Bencken design to evaluate the effects of independent variables on the response (arsenite oxidation efficiency) performance and to predict the best response rate. In this study, the effects of different parameters such as pH (3-11), concentration of persulfate (4-14 mM), and initial concentration of arsenic (0.1-0.9 mg/l) on process efficiency were investigated. The number of tests in this study was 45, and the oxidation rate was measured using the UV visible spectrophotometer (DR 6000) and the molybdate colorimetric method. Results: Increasing the concentration of arsenic increased oxidation. However, with increasing pH, the oxidation rate decreased and the highest oxidation rate at all concentrations was observed at pH 3. The value higher than R2 (0.934) indicated that the oxidation of arsenic (v) could be determined by this model. Conclusion: Arsenite is a highly toxic metal that is difficult to remove by conventional treatment methods, but a pre-treatment phase can convert arsenite into arsenate and facilitate the removal process. In this study, the use of UV-activated persulfate increased the efficiency of arsenic oxidation to 96%.
Article History: Received:25 May 2018 Accepted: 10 July 2018
*Corresponding Author: Ali Asghar Ebrahimi Email: ebrahimi20007@gmail.com Tel: +983531492273
Keywords: Persulfate Oxidation, Ultraviolet Radiation, Aqueous Solution, Arsenic.

Introduction
Access to safe drinking water is one of the basic physical, chemical, and biological needs of humans. High concentration of arsenic in groundwater is one of the main problems affecting the health of millions of people ^{1, 2}. Water-soluble arsenic in groundwater is used for drinking and irrigation purposes ³. Exposure to high levels of arsenic can lead to chronic toxicity and its symptoms include skin lesions (keratosis, melanosis, pigmentation, and blackleg), internal cancers, such as lung, liver, breast, and kidneys, and heart disease as well as growth anomalies in children ^{4, 5}. Due to the carcinogenicity of arsenic, World Health Organization lowered allowed arsenic levels in drinking water from 50 μg/l to 10 μg/l, but drinking water standards aim to reduce arsenic levels in water to less than parts-per million. Arsenite and arsenate are the forms of mineral arsenic, and the concern about drinking water supplies is due to these two forms of arsenic  ⁶. Various methods have been proposed for the removal of arsenic, such as coagulation and filtration, membrane processes such as reverse osmosis and electrodialysis, as well as absorption processes, but one step of oxidation maximizes the efficiency before implementation of any of the processes  ⁷.
United States Environmental Protection Agency (USEPA) suggests to convert trivalent arsenic (arsenite) to pentavalent arsenic (arsenate) and then proceed with the removal of arsenate to achieve a highly efficient arsenic removal from drinking water during a pre-treatment step before the main removal process (using an oxidizing chemical or aeration step)  ⁸. In recent years, the use of Advanced Oxidation Processes (AOPs) as an acceptable and efficient technology has always attracted the attention of researchers and operators of water and wastewater treatment plants. Meanwhile, hydrogen peroxide and ozone have been successfully used in laboratory and real scales, but the short life of oxidizing agent, low solubility, storage and displacement, and high cost are the disadvantages of this method. In recent years, a compound called persulfate with oxidation potential of 2.01 V has been identified. Persulfate is capable of oxidizing toxic and resistant organic compounds  ⁹. Some of the advantages of persulfate are low cost, high stability of the radical produced from it in different conditions, high solubility, solid form and, thus, ease of handling and storage. Regardless of these advantages, extensive studies on the use
of persulfates have shown that the degradation of organic materials by persulfate at room temperature is low and slow ^{10, 11}. Therefore, activation of persulfate is necessary to accelerate the oxidation process. The activation of persulfate, as an AOP, is performed with heat, UV light, and intermediate metals. The final product of the activation is sulfate radical with an oxidation potential of 6.2 V. The reaction below shows the thermal activation of persulfate ^{9, 12,13}.
S₂O₈^2- + heat or UV → SO₄^•-
The UV/persulfate process is one of the newest and most efficient treatment methods. In addition, SO₄^•- is almost neutral and is not considered as a contaminant as the USEPA ranked it as the second agent in the water standards with a maximum allowable concentration of 250 mg/l in drinking water  ¹⁴. The RSM is a method of optimization that uses a set of mathematical and statistical techniques to model issues.
This method evaluates the interactions of several factors in a different spectrum, and illustrates the region, where the results are in that spectrum, as a three-dimensional surface. In recent years, certain tools have been used to optimize analytical methods with significant advantages such as reducing the number of experiments, as well as reducing the use of reagents and laboratory works  ⁶. Given that the high concentrations of arsenic in groundwater in many countries, and no study has yet been conducted with the RSM, and also because trivalent arsenic is very toxic and hazardous, the aim of this study was to investigate oxidizing trivalent arsenic to pentavalent arsenic, which is the low-risk form of arsenic and is eliminated more easily by conventional treatment processes.
Materials and Methods
All chemicals used to prepare standard stock solutions were purchased from Merck Co., Germany ^{15, 16}. To adjust the pH, 0.1 M sulfuric acid and 0.1 M Sodium hydroxide were used ¹⁷. To prepare arsenite stock solution 100 ppm, sodium arsenite salt of 99% purity was dissolved in distilled water. To prepare this solution, we weighed 0.1734 g of sodium arsenite using a digital laboratory scale with precision 0.0001 g and then dissolved it in 1000 ml of distilled water. Under laboratory conditions, the concentration was synthetically kept at a range of 0.1-0.9 mg/l using deionized water (the water distilled two times) and adding sodium arsenite to it. Then, a plexiglas reactor with a useful volume of 1, equipped with a UVc lamp (15 W), was prepared to conduct the experiments . The system was used discontinuously, and during the reaction time, a mechanical stirrer was used to keep the solution homogeneous. Using the appropriate materials and equipment and changing the pH‌ by using a pH meter (HACH( , and the concentration of potassium persulfate, necessary samples were collected and analyzed. Finally, the amount of remained arsenic ​​was measured as the response variable, which is shown by C₂ in Table 1.

Measuring arsenate
Using UV visible spectrophotometer (DR6000, HACH), and and the molybdate colorimetric method, the following steps were performed:
Preparation of the reagent A
To prepare reagent 1, 13g of (NH₄) 6Mo₇O₂₄/4H₂O (/99%, Aldrich) were mixed with 9 M H₂SO₄ (98%, Prolabo) in a 500 ml volumetric flask. This reagent (called ‘‘reagent A’’) is stable for 2 months if stored in an opaque flask.
Preparation of ascorbic acid
Another solution of 10 /100 ml ascorbic acid (/99%, Aldrich) was prepared daily before use.
Measurement of As (V)  
After preparing the reagents and according to the instructions, 1 ml ascorbic acid solution and 2 ml reagent A are successively added to a 40 ml sample aliquot in a 50 ml volumetric flask, and the volume completed by de-ionised water. A blank is prepared according to the same procedure using the appropriate volume of de-ionised water. The analysis is carried out in 1 cm quartz cells of the UV‌ visible spectrophotometer ¹⁸, and the results for each of the 45 samples recorded.
The RSM based on the Box-Bencken design was used to evaluate the effect of independent variables on the performance of response (arsenite oxidation efficiency) and to predict the best response rate ¹⁹.
In this study, the effect of independent variable X₁ was considered to indicate initial concentration of arsenic (mg/l), X₂ to indicate pH, and X₃ to indicate the concentration of potassium persulfate (mM) that were used at three minimum, moderate, and maximum levels of -1, 0, and +1, respectively, with three factors and five central points to determine the percentage of the error sum of squares (Table 2).

Run Order	Factor1: Arsenic;C1	Factor2: pH	Factor3: S2O8	C2	R
1	0.5	7	9	0.37	75
2	0.9	11	9	0.39	43
3	0.1	11	9	0.08	21
4	0.9	7	4	0.42	47
5	0.5	7	9	0.36	72
6	0.9	7	4	0.41	46
7	0.1	7	4	0.07	34
8	0.5	3	4	0.42	84
9	0.9	7	14	0.46	51
10	0.5	11	14	0.34	68
11	0.1	7	14	0.06	44
12	0.5	7	9	0.37	74
13	0.5	11	4	0.32	65
14	0.5	7	9	0.34	68
15	0.5	3	4	0.42	84
16	0.5	7	9	0.33	66
17	0.5	7	9	0.38	76
18	0.1	7	4	0.06	41
19	0.1	7	4	0.06	43
20	0.5	7	9	0.34	68
21	0.5	3	4	0.45	90
22	0.9	3	9	0.64	71
23	0.5	11	14	0.33	66
24	0.9	3	9	0.65	72
25	0.5	7	9	0.36	72
26	0.5	11	14	0.29	58
27	0.5	11	4	0.3	60
28	0.1	11	9	0.08	18
29	0.5	7	9	0.31	63
30	0.9	11	9	0.38	42
31	0.1	3	9	0.02	78
32	0.9	7	4	0.39	43
33	0.1	7	14	0.06	35
34	0.1	11	9	0.07	26
35	0.1	3	9	0.21	79
36	0.5	11	4	0.29	58
37	0.1	7	14	0.05	46
38	0.5	3	14	0.48	96
39	0.9	3	9	0.76	85
40	0.5	3	14	0.47	95
41	0.1	3	9	0.02	82
42	0.5	3	14	0.48	96
43	0.9	7	14	0.36	40.5
44	0.9	7	14	0.35	38.8
45	0.9	11	9	0.33	37.2

Variable	Symbol	+1	0	-1
Arsenic concentration	C1	0.9	0.5	0.1
pH	pH	11	7	3
Persulfate concentration	S2O8	14	9	4

The total number of experiments in the Box-Bencken design was 45 tests. The model used in the RSM is generally a quadratic relationship. The data from the Box-Bencken design were used to determine the proportion of quadratic polynomial regression equations.
Y = β0 + β1x1 + β2x2 + β3x3 + β4x4 + β12x1x2 + β13x1x3 + β14x1x4 + β23x2x3 + β24x2x4 + β34x3x4 + β11 X1² + β22 X2² + β33 X3² + β44 X2²
Ethical issues
Ethical approval was obtained from the Ethics Committee of Shahid Sadoughi University of Medical sciences, Yazd, Iran (ID: IR.SSU.SPH.REC.1395.135).
Results
According to the RSM based on the Box-Bencken design and the model presented in Table 1, after adjusting the desired pH level, and then adding the specified amounts of potassium persulfate, as the oxidizing agent, they were exposed to the UV lamp for a specified period. The values of remained arsenic (V) (C₂), were measured as the arsenic (V) and response variable. Regarding the recording of the initial concentration of arsenic and calculating the final concentration of arsenic (V), oxidation efficiency was recorded at each step of the experiment. The results of the study are presented in Table 1. Based on the results of Table 1, it can be inferred that the lowest efficiency of arsenic oxidation by the UV/k2S2O8 process is 18% and the highest efficiency 96%. The following equation represents the experimental relationship between the experiment variables and the percentage of efficiency as encoded, and the result of the RSM is based on the Box-Bencken design:
Oxidation efficiency= 70.5 + (2.9 × X1) – (18.7 × X2) – (25 × X1²) + (9 × X2²) + (5.6 × X1 × X2)
As the above equation shows, a number of factors are positive and some others are negative. The negative values in the resulting equation indicate increased efficiency with decreasing the parameter and the positive values in the equation represent a direct correlation between the efficiency and the parameter. It should be noted that X₁ represents the initial concentration of arsenic (mg/l), X₂ pH, and X₃ the concentration of potassium persulfate. To determine the quality of the proposed polynomial model, the coefficient R₂ and adjusted R² were used. Analysis of variance (ANOVA) was used as the statistical method to analyze the responses ²⁰ . According to the results from ANOVA test (Table 3), the oxidation of arsenic by the UV/k₂ S₂O₈ process was well described by the model. It should also be noted that the effect of all studied parameters, except for persulfate, on the oxidation efficiency of arsenic is significant with regard to the statistically significant level of lower than 0.05. The high linear regression coefficient for arsenic oxidation by persulfate (R² = 0.934 and R² adjusted = 0.917) shows that the model worked appropriately and predicted well. The closer to 1 the R² value is, the greater the power of the model in describing response variations as a function of independent variables.
pH effect
The pH parameter is an important variable in the oxidation process due to its effect on the amount and type of produced radicals. For this purpose, the effects of pH 11, 9, 3 on the efficiency of the UV/persulfate process were investigated at concentrations 0.1, 0.5, and 0.9 mg/l of arsenic. 0.1 N NaOH and H₂SO₄ solutions were used to adjust the pH to the desired value. The findings are illustrated in Figure 2. According to Figure 2, the UV/persulfate process at pH 3 had the highest efficiency in removing arsenic.

Source	Mean Square	F -value	P-value
Corrected Model	1506.303	73.598	.000
Intercept	142344.997	6955.000	.000
Arsenic	201.898	9.865	.004
pH	8430.002	411.891	.000
S2O8	63.994	3.127	.087
Arsenic pH	387.603	18.938	.000
Arsenic S2O8	13.675	.668	.420
pH S2O8	33.333	1.629	.211
Arsenic pH S2O8	.	.	.
Error	20.467
Total
Corrected Total

Effect of Persulfate Concentration
To determine the optimum concentration of persulfate for the removal of arsenic using the UV/persulfate process, the concentrations 4, 9, and 14 mM of persulfate were examined. The results are illustrated in Figures 3 and 4. As it can be observed, the oxidation efficiency of arsenic in the UV/persulfate process does not undergo a specific trend with changing the persulfate concentration.
By increasing the concentration of persulfate, an excessive amount of sulfate ion is produced which acts as an interfering radical. According to the equations below, it is an agent for the conversion of sulfate radical to persulfate. On the other hand, persulfate radical reacts with persulfate; therefore, anion sulfate is produced, all of which lead to destruction of persulfate radical and reduce the removal efficiency.
 
      SO₄^{• -}  +  S₂O₄^{2 -}                SO4^{2  -}   +  S₂O₈^{• -}       
      SO₄^{• -}  +   SO₄^•                   S₂O₈^•-
 
In the Figure 2, the interaction effects and binary interactions of the studied factors were used to determine the amount of arsenic oxidation with pH at different concentrations of potassium persulfate and arsenite.

Figure 2: Interaction effects and binary interactions of the studied factors (arsenic and pH)
 

The effect of initial arsenic concentration
In order to investigate the effect of initial
arsenic concentration on the UV/persulfate process efficiency at pH 3, concentrations 0.1, 0.5, and 0.9 mg/l of arsenic were investigated and, as shown in Figures 3 and 4, the highest oxidation (96%) carried out at 0.5 mg/l concentration with 96% efficiency.

Figure 3: Interaction effects and binary interactions of studied factors (persulfate and pH)
 

Figure 4: Interactions effects and binary interactions of the studied factors (persulfate and arsenic)
 

1. Taheri E, Hatamzadeh M. Evaluation of the influence of conventional water coolers on drinking water quality. Int J Hydrogen Energy.2010 ; 1(4): 268-75 .
2. Aleseyyed SB, Norouzi H, Khodabakhshi M. Evaluation of heavy metals in drinking water resources in urban and rural areas of Hamadan province in 2016. Journal of Environmental Health and Sustainable Development. 2018;  3(1):  448-53.
3. 3.Bhattacharya P, Chatterjee D, Jacks G. Occurrence of arsenic-contaminated groundwater in alluvial aquifers from delta plains, eastern India: options for safe drinking water supply. International Journal of Water Resources Development. 1997; 13(1): 79-92.
4. Bellamri N, Morzadec C, Fardel O, et al. Arsenic and the immune system. Curr Opin Toxicol.  2018; 10(1): 60-8.
5. Sorlini S, Gialdini F, Stefan M. UV/H2O2 oxidation of arsenic and terbuthylazine in drinking water. Environmental Monitoring and Assessment. 2014; 186(2): 1311-6.
6. Zarei H, Mahvi AH, Nasseri S, et al. Modeling adsorption on fluoride and application of Box–Behnken design and response surface methodology for arsenic(V) removal from aqueous solution using nano-scale alumina on multi walled carbon nanotube. Iranian Journal of Health & Environment. 2015; 8(3): 309-22.
7. Chaudhry SA, Khan TA, Ali I. Zirconium oxide-coated sand based batch and column adsorptive removal of arsenic from water: Isotherm, kinetic and thermodynamic studies. Egyptian Journal of Petroleum. 2017; 26(2): 553-63.
8. Mokhtari SA, Gholami M, Shakerkhatibi  M, et al. The impact of the concentration parameters, pH, temperature and pressure to remove arsenic from drinking water using a reverse osmosis process. Journal of Ardabil University of Medical Sciences. 2010; 10(3): 261-69.
8. 9.Shih YJ, Putra WN, Huang YH, et al. Mineralization and deflourization of 2,2,3,3-tetrafluoro-1-propanol (TFP) by UV/persulfate oxidation and sequential adsorption. Chemosphere. 2012; 89(10): 1262-6.
8. 10.Mohiuddin M,  Yamamoto K. Low cost technique of arsenic removal from drinking water by coagulation using ferric chloride salt and alum. Water Science & Technology Water Supply. 2002; 2(2): 281-8.
11. Hussain I, Li M, Zhang Y, et al. Efficient oxidation of arsenic in aqueous solution using zero valent iron- activated persulfate process. J Environ Chem Eng. 2017; 5(4):  8-90.
11. 12.Yang S, Wang P, Yang X, et al. A novel advanced oxidation process to degrade organic pollutants in wastewater: Microwave-activated persulfate oxidation. Journal of Environmental Sciences. 2009; 21(9): 1175-80.
11. 13.Jing Y, Li L, Zhang Q, et al. Photocatalytic ozonation  of dimethyl phthalate over TiO2  prepared  by a hydrothermal method. Journal of Hazardous Materials. 2011; 189(1-2): 40-7.
11. 14.Neppolian B, Doronila A, Ashokkumar M. Sonochemical oxidation of arsenic(III) to arsenic(V) using potassium peroxydisulfate as an oxidizing agent. Water Research. 2010; 44(12): 3687-95.
11. 15.Zhao P, Guo Y, Zhang W, et al. Neurotoxicity induced by arsenic in Gallus Gallus : Regulation of oxidative stress and heat shock protein response. Chemosphere. 2017; 166(Supplement C): 238-45.
11. 16.Davoudi M, Ehrampoush M, Vakili T, et al. Antibacterial effects of hydrogen peroxide and silver composition on selected pathogenic enterobacteriaceae. International Journal of Environmental Health Engineering. 2012; 1(2): 45-8.
11. 17.Mafu LD, Mamba BB, Msagati TAM. Synthesis and characterization of ion imprinted polymeric adsorbents for the selective recognition and removal of arsenic and selenium in wastewater samples. Journal of Saudi Chemical Society. 2016; 20(5): 594-605.
11. 18.Lenoble V, Deluchat V, Serpaud B, et al. Arsenite oxidation and arsenate determination by the molybdene blue method. Talanta. 2003; 61(3): 267-76.
11. 19.Zhao C, Liu J, Li X , et al. Biosorption and bioaccumulation behavior of uranium on Bacillus sp. dwc-2: Investigation by Box-Behenken design method. Journal of Molecular Liquids. 2016; 221(Supplement C): 156-65.
11. 20.Hashemi H, Ebrahimi A, Mokhtari M, et al. Removal of PAHs and heavy metals in composting leachate using the anaerobic migrating blanket reactor (AMBR) process. Desalination and Water Treatment. 2016; 57(52): 24960-9.
21. Dutta PK, Pehkonen SO ,  Sharma VK, et al. Photocatalytic Oxidation of Arsenic(III): Evidence of Hydroxyl Radicals. Environmental Science  &  Technology.  2005; 39(6):1827-34.
21. 22.Saxena V, Sadoqi M, Shao J. Degradation kinetics of  Indocyanine Green in aqueous solution. Journal of Pharmaceutical Sciences. 2003; 92(10): 2090-7.
21. 23.Daneshvar N, Oladegaragoze A, Djafarzadeh N. Decolorization of basic dye solutions by electrocoagulation: An investigation of the effect of operational parameters. Journal of Hazardous Materials. 2006; 129(1-3): 116-22.
21. 24.Mollah MYA,  Pathak SR, Patil PK, et al. Treatment of orange II azo-dye by electrocoagulation (EC) technique in a continuous flow cell using sacrificial iron electrodes. Journal of Hazardous Materials. 2004; 109(1-3): 165-71.
21. 25.Panizza M, Cerisola G. Influence of anode material on the electrochemical oxidation of 2-naphthol: Part 2. bulk electrolysis experiments. Electrochimica Acta. 2004; 49(19): 3221-6.
21. 26.Jamshidi N, Torabian A, Azimi A,et al. Efficiency comparison of advanced photochemical oxidation technologies in phenol removal from aqueous solution in Iran. Journal of Water and Wastewater(parallel title ). 2010; 20(4): 9-24.
27. Zarrabi M, Samarghndi MR, Rahmani AR, et al. Kinetic study of acid red 18 and acid red 14 removal from aqueous solution using metallic iron. Journal of Health. 2010; 1(3): 31-40.
28. Piri R, Kermani M, Esrafili A. Using persulfate-based photochemical oxidation (UV/Na2S2O8) in eliminating 4-chlorophenol from aqueous solutions. Journal of Mazandaran University of Medical Sciences. 2017; 27(147): 358-70.