Intensification of Azo Dye Removal Rate in the Presence of Immobilized TiO 2 Nanoparticles and Inorganic Anions under UVC Irradiation : Optimization by Response Surface Methodology

Wastewaters contain inorganic anions that affect the removal rate of organic pollutants. The present study aims to optimize the effects of inorganic anions such as NO 3 , Cl, HCO 3 , and H 2 PO 4 − on the removal rate of an organic pollutant in the presence of immobilized TiO 2 nanoparticles using response surface methodology (RSM). C.I. Acid Red 17 (AR17) was used as a model organic pollutant. Thirty experiments were required to study the effects of anions in various concentrations. The results indicate that the addition of NO 3 − and HCO 3 − ions intensifies the removal rate of AR17. The results of the analysis of variance (ANOVA) showed a high coefficient of determination value (R = 0.9866 and Radj. 2 = 0.9822). The results indicate that RSM is a suitable method for modeling and optimizing the process. The results prove that in the presence of and NO 3 − and HCO 3 − ions especially in the combination situation the removal rate of AR17 is enhanced considerably. An important synergy effect was observed in the combination of NO 3 − and HCO 3 − ions, so that AR17 removal percent under the optimized RSM conditions was considerably more than the sum of removal percent when these ions are used individually.


Introduction
Producing many products such as clothes, leather accessories, and furniture, needs synthetic dyes.Most of these dyes are toxic and potentially carcinogenic in nature.The discharge of their effluents into the environment is harmful for human and all living beings.Therefore, their removal from industrial effluents is considered as a major concern and responsibility [1][2][3][4].Azo dyes are the largest group of commercial dyes [5].Among the various treatment methods, advanced oxidation processes (AOPs) have been reported as a promising and practical method [2,6,7].Heterogeneous photocatalysis is one of the AOPs and has attracted extensive attention of many researchers.This method is based on producing photoexcited electrons (e − ) in the conduction band and positive-charged holes (h + ) in the valence band of an oxide semiconductor like TiO 2 under ultraviolet (UV) light.The valence band holes are powerful oxidants, and the conduction band electrons are great reductants.Reactions include oxidation of organic pollutants by photogenerated positive holes (h + ) or by reactive oxygen species (HO • , HOO • , and O 2 •− ) formed on the TiO 2 surface under UV light irradiation [8][9][10].
Dye-containing wastewaters have inorganic anions such as nitrate, chloride, bicarbonate, sulfate, and dihydrogenphosphate that impress removal rate of pollutants.Most of the investigations have reported inhibitory effects of inorganic anions on the removal rate of pollutants in the AOPs [11][12][13][14][15].
Response surface methodology (RSM) is a statistical method to reduce the number of the experiments.Mathematical and statistical techniques are applied in this method to analyze the effect of the independent variables (inorganic anions concentration) on a specific dependent variable (removal percent).This method gives linear interaction and quadratic effects of the factors and is useful for the optimization of process [16,17].In this paper, NaNO 3 , NaHCO 3 , NaH 2 PO 4 , and NaCl were used as additives, and the effects of these anions on the removal of the model dye pollutant in the presence of immobilized TiO 2 nanoparticles on a glass plate were investigated and optimized via using response surface methodology.were obtained from Merck Co. and were used without further purification.N,N-dimethyl-p-nitroso-aniline was purchased from Fluka (Switzerland).

Immobilization of TiO 2 -P25 on Glass
Plates.Heat attachment method was used to prepare the immobilized TiO 2 -P25 on glass plate (1.5 × 18 cm 2 ) [18].To obtain immobilized TiO 2 , a suspension containing 10 g L −1 TiO 2 -P25 in double distilled water was prepared.Then, its pH was adjusted to about 1.5 by HNO 3 .The prepared suspension was sonicated in an ultrasonic bath (Elma, Windaus T460/H, Germany) under frequency of 35 kHz for 30 min in order to improve the dispersion of TiO 2 -P25 in water.Glass plate was treated with a diluted HF solution (5% v/v) in order to make a rough surface for better contact with TiO 2 nanoparticles.The treated glass plate was washed with double distilled water several times.Next, sonicated TiO 2 suspension was poured on glass plate for 20 min, then removed from the suspension, and then placed in an oven at 150 ∘ C.After drying, the glass plate was heated at 500 ∘ C in a furnace for 2 h and then was thoroughly washed with double distilled water to remove the poorly attached or free TiO 2 particles [19,20].The deposition process was carried out twice to increase the loaded TiO 2 on the surface of glass plate.Typically, the deposited TiO 2 is 0.26 mg cm −2 .The characteristics of the TiO 2 nanoparticles immobilized on the glass plate were determined using an atomic force microscopy (AFM) carried out by DME (Danish Micro Engineering A/S).[22,23].The CCD combines a two-level full/fractional factorial design with additional star points and minimum one point at the center of the experimental position.The required experiment numbers in the CCD is computed using the following equation:

Photoreactor and Light
where k is the number of factors and  0 is the replication numbers of central point.In this design, all factors are studied at five levels [22,24].In addition, in the CCD the variable parameters, that is,   , were coded as   .For this purpose, the following relationship was used:  where  0 is the value of   in the central point and X represents the step change [25,26].In the present study, the concentrations of NO 3 − , HCO 3 − , H 2 PO 4 − , and Cl − ions were determined in terms of four main factors in order to evaluate the effects of these inorganic anions on the removal efficiency of AR17.Totally, 30 experiments were carried out including six replications in the central point.The experimental ranges and the levels of the independent variables for AR17 removal are provided in Table 2.The experimental data were analyzed using DX7 software.

Central Composite Design
Model.The present study made use of Central Composite Design (CCD) for the optimization of AR17 removal rate in the presence of various concentrations of inorganic anions.The structure of this design and the results of the measurement of AR17 removal percent after 5 min irradiation are shown in Table 4.In order to correlate the dependent and independent variables, the following second-order polynomial response equation was used: where  is AR17 removal percent after 5 min irradiation (response variable),   is regression coefficient for linear effects,   is regression coefficient for squared effects,   is regression coefficient for interaction effects, and   is the coded experimental level of the variables.According to this design, an empirical reduced quadratic relationship with significant model terms between the removal efficiency of AR17 (response, ) and inorganic ions concentration (variable parameters) was obtained using the following formula: where  1 ,  2 , and  3 represent the NO 3 − , HCO 3 − , and H 2 PO 4 − concentrations, respectively.The results show that a quadratic response surface model fits the process.The predicted removal efficiencies by (4) are provided in Table 4.The results show that there are perfect agreements between the experimental and predicted values of removal efficiency.According to (4), the effective parameters in this process include NO The measurement of the residuals is a method to evaluate the adequacy of the model where the residuals represent the differences between the observed and the predicted response values.Normal probability plot is a suitable graphical method to know about the normality of the residuals.As shown in Figure 2, the inclination of the residuals is towards normal distribution.Figure 2 shows that the residuals in the plot rise and fall randomly around the central line.These two plots revealed that the model is adequate for explaining the process under study.
ANOVA was used to study the significance and adequacy of the model.The correlation coefficient ( 2 ) value is a standard value to know how much variability in the observed response values can be explained by the experimental factors and their interactions.The results show a high value of coefficient of determination ( 2 = 0.9866 and  adj. 2 = 0.9822).This implies that 98.66% of the variations for AR17 removal efficiency can be explained by the independent variables in the represented model.However, 1.34% of variation cannot be explained by the model.Table 5 shows the results of the reduced quadratic response surface model fitting in the form of ANOVA for the removal of AR17.The variation of the model and the experimental error constitute the total variation of the results.ANOVA shows whether the variation of the model is significant in comparison with the variation of residual error or not.For this purpose, the ratio between the mean square of the model and the residual error (F-value) was used [16,17,28,29].The obtained F-value (230.70) is obviously greater than the F-value for LOF (8.96) which confirms the appropriateness of the model.

Effect of Variables as Response Surface and Contour Plots.
Three-dimensional (3D) surfaces and two-dimensional (2D) contour plots, based on the model polynomial function, were drawn to investigate the effects of the inorganic anions within the range considered.Figure 3 shows the effect of NO 3 − and HCO 3 − concentration on the AR17 removal efficiency () for H 2 PO 4 − and Cl − concentration of 60 mM.Apparently, the removal efficiency increased as NO 3 − and HCO 3 − concentration increased.However, the effect of HCO 3 − was less than the effect of NO 3 − .In order to gain approximately 80% removal efficiency, the concentration of these ions should increase up to 100 mM.Thus, it is practical to achieve the maximum removal efficiency by providing the maximum concentration of these ions.As can be seen in Figure 3, increasing the removal efficiency of the process due to increasing HCO 3 − concentration is less significant in comparison to the increasing of NO 3 − concentration which indicates that the increasing effect of NO 3 − concentration had more contribution than the increasing effect of HCO 3 − concentration.Figure 4 shows the effect of NO concentration had less contribution than the increasing effect of NO 3 − concentration.Figure 5 shows the effect of HCO 3 − and H 2 PO 4 − concentration on the removal efficiency () for NO 3 − and Cl − concentration of 60 mM.It is apparent that the removal efficiency increased with increasing HCO 3 − concentration and decreased with increasing H 2 PO 4 − concentration.Hence, the maximum removal efficiency is achieved by providing the maximum HCO 3 − and the minimum H 2 PO 4 − concentration.The effects of aforementioned anions on the removal efficiency of AR17 in the presence of immobilized TiO 2 under UV-C irradiation can be explained as follows.
Apparently, the two possible effects of inorganic ions on the photocatalytic reactions are the increase of the ionic strength of the reaction medium and the scavenging of the reactive species such as hydroxyl radicals [30].Scavenging of reactive species by inorganic anions causes an inhibitory effect in the photocatalytic activity.The increase of the ionic strength has positive effect on enhancing the photocatalytic activity.The enhancement of the removal efficiency with increasing of NO 3 − and HCO 3 − ions concentration in the fixed-bed system indicates that the reinforcement of the ionic strength of the solution in the immobilized system is more significant than their inhibitory effects.The increase of ionic strength of solution and charge transfer in fixed-bed system reduce the mass transfer limitations in this medium and increase the surface contacts between the pollutant and catalyst surface, resulting in suitable medium for removal of AR17 [13].
In the literature, three various roles for NO 3 − ions in the heterogeneous photocatalysis process have been reported.The reduction of UV light absorption by TiO 2 due to NO 3 − ions inner filter affects, and adsorption of NO 3 − ions onto TiO 2 surface causes the decreasing of adsorption and the contact of substrate with active sites on the catalyst surface and finally the production of hydroxyl radials according to the following equations [13,31,32]: In an immobilized catalyst surface, which has been tested in this study, adsorption process can be neglected due to low accessible active sites.In other words, in fixed-bed system under UV-C irradiation the increase of ionic strength of the solution and especially producing hydroxyl radials in the bulk solution according to ( 5)-( 8) are main reasons that lead to a very high removal efficiency in the presence of NO 3 − ions.Selvam et al. [33] indicated that the presence of NO reacts with hydroxyl radical to form SO 4 •− ions which can accelerate the reaction, and no hydroxyl radical quenching was reported for NO 3 − ion.Zhang et al. [13] reported that the photocatalytic degradation of reactive Brilliant Orange K-R in the packed-bed reactor increased at higher concentrations of NO 3 − ions.This effect is attributed to the formation of hydroxyl radicals from interaction of NO 3 − with UV light, and the excess free NO 3 − would also enhance charge transfer in the packed-bed reactor.
The main role of bicarbonate ions on the AOPs is that these ions are scavenger for hydroxyl radicals according to (9) [34]: Bicarbonate ions can have an inhibitory or enhancing effect on the photocatalytic reactions depending on the oxidizability of substances.In the case of easily oxidizable substances, bicarbonate ions can have enhancing effect on the removal rate.It means that degradation of these substances which do not need strong oxidizing conditions can be supplied by CO  Blue 74 and indicated that the degradation efficiency of the process was enhanced with bicarbonate anion during the first 80 min of the process.This can be explained by the fact that bicarbonate ions can react with hydroxyl radicals to produce carbonate radicals, which are weak oxidizing agent [36].The substitution of hydroxyl radicals with carbonate radicals could enhance the removal rate though less reactivity of carbonate radicals since hydroxyl radicals undergo radicalradical recombination 275 times higher than that the reaction of carbonate radicals with itself [37].Bekbölet et al. [38] reported that the presence of phosphate ions strongly inhibited the color removal rate.• [11,12,39].Previous researches showed that Cl − ions had inhibitory effects on the photocatalytic activity in slurry systems and packed-bed reactor with scavenging of hydroxyl radicals [11][12][13]39].Wang et al. [40] reported that the Cl − ion is strongly adsorbed on the TiO 2 surface at pH 3 and reduces the removal  rate whereas at neutral or alkaline conditions, the addition of the Cl − ion did not influence the removal rate.Since attraction of the Cl − ion by TiOH 2 + and repulsion of it by TiO − can occur at acidic and basic pH, respectively.In the present research due to the proximity of solution pH to neutral pH, this ion did not influence the removal rate.radical but no perhydroxyl and superoxide radicals [41].

Optimization of the Removal Efficiency Using RSM in the
As shown in Figure 6, the absorbance of RNO at 440 nm decreased faster in the optimized conditions.This result shows that in the optimized conditions high amount of hydroxyl radicals is produced; therefore the absorption peak of RNO at  max = 440 nm decreased very faster in the presence of [NO

Conclusion
Removal of organic pollutants in the fixed-bed system under UV-C irradiation can be accelerated by inorganic anions such as NO 3 − and HCO 3 − ions.The efficiency of the process depends on the inorganic anions types and concentration.The results of this study indicate that CCD methodology could be suitable for optimizing of removal process in the presence of various concentrations of inorganic anions.The NO 3 − concentration had the greatest effect and contribution in the AR17 removal whereas Cl − ions did not have a significant effect on that.The optimal values of the NO 3 − , HCO 3 − , and H 2 PO 4 − concentrations were found to be 120, 120, and 0 mM, respectively.Under this condition, the highest AR17 removal percent was obtained.An important synergy effect was observed in the combination of NO 3 − and HCO 3 − ions for the removal of pollutant.Analysis of variance (ANOVA) shows a high coefficient of determination ( 2 = 0.9866 and  adj. 2 = 0.9822), proving an adequate adjustment of the second-order regression model with the experimental data.In addition, it was indicated that the residuals followed a normal distribution.
Figure1shows two-and three-dimensional AFM pictures of TiO 2 -P25 nanoparticles immobilized on glass plate for a projected area of 5 × 5 m 2 .The results of AFM (Table3) allow making a morphological interpretation of the catalyst surface.The results indicate the formation of a surface with root mean square (RMS) and average roughness of 37.1 and 28.9 nm, respectively.The surface skewness (Ssk) describes the asymmetry of the height distribution histogram.Ssk = 0 indicates a symmetric height distribution like a Gaussian.The surface kurtosis (Sku) is another parameter used to describe the surface topography.For Gaussian height distribution, Sku approaches 3.0 whereas the lower values indicate broader height distribution [27].The values of Ssk and Sku for the prepared catalyst surface indicate a fairly symmetric height distribution of Gaussian.

3 −
concentration, HCO 3 − concentration, H 2 PO 4 − concentration, mutual effects of NO 3 − and HCO 3 − concentration, NO 3 − and H 2 PO 4 − concentration, HCO 3 − and H 2 PO 4 − concentration, and the square effect of NO 3 − concentration.Furthermore, according to the model, Cl − concentration was considered as an ineffective factor.

Figure 2 :
Figure 2: Residual plots for removal of AR17 in the presence of various concentrations of inorganic anions.

6 AFigure 4 :
Figure 4: Response surface (a) and contour plots (b) of the removal efficiency (%) as a function of NO 3 − and H 2 PO 4 − concentration.

Figure 5 :
Figure 5: Response surface (a) and contour plots (b) of the removal efficiency (%) as a function of HCO 3 − and H 2 PO 4 − concentration.

Table 2 :
Experimental ranges and levels of the independent test variables.

Table 3 :
Morphological information of TiO 2 -P25 nanoparticles immobilized on glass plate.

Table 4 :
The four-factor central composite design matrix and the value of the response function (%).

Table 5 :
Analysis of variance (ANOVA) for fit of removal efficiency from central composite design.
3 − and H 2 PO 4 − concentration on the removal efficiency () for HCO 3 − and Cl − concentration of 60 mM.As it is shown in this figure, the removal efficiency increased with increasing NO 3 − concentration and decreased with increasing H 2 PO 4 − concentration.It is obvious that in order to achieve the maximum removal efficiency, NO 3 − concentration should be maximized and H 2 PO 4 − concentration should be minimized.The results indicate that the reduction effect of H 2 PO 4 − 3 •− radicals [34].Aleboyeh et al. [35] studied the effect of dyeing auxiliaries on UV/H 2 O 2 removal of Acid International Journal of Photoenergy The strong inhibition of H 2 PO 4 − ion was mainly attributed to the coagulation role of NaH 2 PO 4 .However, in the present study, the coagulation effect should not be an important phenomenon because TiO 2 was immobilized on glass plate.The hindering effect of H 2 PO 4 − can be explained by the competitive adsorption of H 2 PO 4 − with substance onto the surface of immobilized TiO 2 .Furthermore, H 2 PO 4 − can react with h + and hydroxyl radicals to form a less reactive species, H 2 PO 4