Abstract

Ammonia removal from synthetic wastewater was studied through a photocatalytic degradation process under UV light. In this study, TiO2/C3N4 was synthesized through a simple method of preparing g-C3N4 through the pyrolysis of melamine then adding it to TiO2. On the other hand, ZnO/C3N4 composite was prepared by a deposition–precipitation technique. The composites were described by Fourier transform infrared spectroscopy (FTIR), X-ray powder diffraction (XRD) and scanning electron microscope (SEM). Response surface methodology (RSM) has been utilized to model variables using Minitab 18. Calculated values of degradation efficiency were in good agreement with experimental values (R2 = 0.97 and Adj–R2 = 0.91). The influence of parameters over ammonia initial concentration (10–50 ppm), catalyst dosage (0.2–1.5 g), light intensity (6–30 W) and stirring speed (100–500 rpm) on ammonia removal percentage was investigated, and their main and interaction contribution was examined. The optimum conditions of the degradation were observed at a dosage of 1 g/L and initial concentration of ammonia 10 ppm for UV intensity irradiation with 24 W lamps. It was concluded that the photocatalytic degradation of the ammonia solution, after 50 min of UV irradiation, can reach percentages of 46%, and 52% using the catalysts TiO2/g-C3N4 and ZnO/g-C3N4, respectively.

INTRODUCTION

Ammonia is a major aquatic pollutant which accelerates the eutrophication and causes an increase of oxygen demand. The permitted level of ammonia is beneath 1 ppm (Luo et al. 2015). It was concluded that when ammonia was released in water sources, fish could experience ammonia poisoning (Gupta et al. 2015).

The photocatalytic oxidation of NH3 is viewed as a promising approach to detoxify animal waste containing effluents by a singular walk handle. The fundamental examinations go back to the 1980s, when the photodecomposition of vaporous ammonia on ultraviolet (UV) illuminated anatase TiO2 particles was first considered by Mozzanega etal. (Altomare et al. 2014). The illumination of a semiconductor photocatalyst with UV radiation activates the catalyst, establishing a redox environment in the aqueous solution. TiO2 can be activated using UV illumination with a wavelength up to 387.5 nm. In the meantime, solar irradiation starts at a wavelength of about 300 nm. However, the content of UV in sunlight is only 3−4%. Consequently, it limits the use of sunlight as an energy source (Dong et al. 2012).

TiO2 was utilized to break N-H bonds and study the instrument of NH3 deterioration (Zhou et al. 2016). A few studies have demonstrated that the photocatalytic method can destroy N-H bonds (Shavisi et al. 2013; Altomare et al. 2014).

Some investigations demonstrated that TiO2 addresses the most promising photocatalyst for such a response, since the oxidizing species productively photograph delivered on its surface can oxidize NH3, generally into nitrite and nitrate particles (Mikami et al. 2010), which, however, are also noxious. Consequently, the most alluring approach to completely remove nitrogen-containing wastes is the selective photocatalytic oxidation of NH3 into innocuous N2. Therefore, the mechanism of NH3 photocatalytic oxidation has to be examined by taking into account the distribution of the final reaction products (Mikami et al. 2010). Recently, some of the researchers methodically investigated the impact different experimental conditions have on the photocatalytic abatement of NH3 in aqueous suspensions and on the selectivity toward its oxidation products (Altomare et al. 2014).The deposition of metal nanoparticles on TiO2 was investigated, focusing on the impacts of the type/amount of metal on the reaction pathways and product selectivity, even by coupling ammonia abatement with hydrogen production (Yuzawa et al. 2012) or with NO reduction. The response has been researched using both TiO2 and different semiconductor materials as photocatalysts (Shavisi et al. 2013), also with the trial of tailoring them toward NH3 selective conversion into innocuous N2 (Mikami et al. 2010; Altomare et al. 2014).

Among the utilized photocatalysts, ZnO has demonstrated a high capability in photocatalytic degradation as a result of its remarkable band crevice vitality and ease (Xiong et al. 2009). Zinc oxide (ZnO) is an important photocatalyst due to its similar band gap energy and required band gap sites compared with titanium dioxide (TiO2). The main factor affecting the photocatalytic activity of ZnO is the quick recombination of charge carriers. In order to improve the photocatalytic activity; ZnO was doped by metals and metal oxides (Achouri et al. 2014); these doping composites broaden the absorption spectrum of this semiconductor toward the visible-light region, as new energy levels are formed between its valence and conduction bands (Xiong et al. 2009).

Graphitic carbon nitride (g-C3N4) has high activity for the photodegradation of pollutants in waste water with the benefits of high thermal and chemical stability. Graphitic carbon nitride (g-C3N4) is investigated as a sensitizer candidate possessing several advantages towards other photocatalysts; and shows a clear decrease in band gap values for g-C3N4 TiO2 systems (Li et al. 2009; Miranda et al. 2013).

For these reasons, the combination of ZnO or TiO2 with C3N4 may be an excellent photocatalyst to achieve an enhanced charge separation in electron-transfer processes. Our research on TiO2 coupled with g-C3N4, and ZnO coupled with g-C3N4 were carried out to upgrade the photocatalytic activity for the ammonia degradation in water samples that simulated those from the effluent from the fertilizer industry.

In this study, we used the response surface methodology (RSM) for the experimental condition optimization. The use of factorial designs leads to optimized parameters with a minimum set of runs and also to the possibility of obtaining a polynomial expression that describes the process yield (Bali 2004; Oliveira et al. 2006).

MATERIALS AND METHODS

Melamine, titanium dioxide, zinc chloride, sodium carbonate, zinc nitrate, sodium hydroxide and soluble starch were utilized as precursors for the synthesis of photocatalysts composites. All chemicals were utilized as acquired without further purification.

Synthesis of g-C3N4

The g-C3N4 photocatalyst was incorporated by heating melamine powder. Five grams of melamine powder was placed into an alumina pot crucible heated in a muffle furnace at 520 °C for 4 h, g-C3N4 was obtained.

TiO2 coupled with g-C3N4 photocatalyst preparation

Graphitic carbon nitride-–TiO2 composites were accomplished by a simple impregnation method. In a typical technique, 0.25 g of g-C3N4 and 10 g of TiO2 were dissolved into methanol and sonicated independently for 30 min. At that point the two slurries were blended and stirred at room temperature for 24 h. Later, the composite photocatalysts were acquired by evaporating the methanol in a rotary evaporator at 80 °C.

ZnO coupled with g-C3N4 photocatalyst preparation

The precursors of C3N4/ZnO photocatalysts were submitted to a deposition–precipitation technique at room temperature, in which 0.074 g of melamine was dissolved in 25 mL of ZnCl2 (0.5 mol/L) solution in a 250 mL beaker, and the suspension was mixed for 20 min. At that point, 25 mL of (0.5 mol/L) solution was added drop wise into the above suspension and stirred magnetically for 30 min. Subsequently, the mixture was filtered, washed with deionized water many times, and dried at 60 °C for 24 h. The precursor of C3N4/ZnO photocatalyst in the C3N4 ratio of 5.0 wt.% was obtained.

Characterization

The surface functional groups and structure were investigated by Fourier transform infrared (FTIR) spectroscopy. The FTIR spectra of the TiO2-C3N4 and ZnO-C3N4 mixtures were recorded between 500 and 4,000 cm−1.

The X-ray diffraction (XRD) patterns of the samples were measured to confirm the prepared composite formations.

A scanning electron microscope (SEM) was employed to visualize sample morphology. In the present study, the samples prepared were analyzed by using SEM JEOL JSM 6,360 L and were taken at 20,000 and 35,000 magnification.

Photocatalytic experiment

The glass reactor (inner volume: 2 L) was utilized for the photodegradation of ammonia from synthetic waste water samples. The system was illuminated with a 12 W lamp with a peak light intensity at 365 nm, covering the glass reactor. The initial pH was measured and adjusted to the desired value (pH = 10 using HCl and/or NaOH solution), this is because the number of OH ions increases when pH increases, and more OH ions generated result in promoting the degradation rate of ammonia nitrogen. Also, there are two forms of ammonia nitrogen present in water: NH3 and NH4+. The ratio of NH3 molecules increases when the pH increases in the solution (Luo et al. 2015). The entire system was protected by a black case during the reaction to obstruct outside light. Catalyst granules were stirred in 2.0 L of ammonia waste water solutions and exposed to the UV source. The temperature of the suspension was kept at 20 ± 8 °C, and the irradiation time was 50 min. A needle-type probe was inserted in the reactor to withdraw samples. The liquid sample (<5.0 mL) was collected in a vial wrapped in aluminum foil to reduce interference from indoor fluorescent light before the analysis. The concentration of residual ammonia was determined by simple titration against sulphoric acid.

Central composite design

Central composite design (CCD) was utilized for optimization of four factors (initial ammonia concentration, irradiation intensity, stirring speed and catalyst dosage) to gather information about their impact. The number of conducted experiments was calculated as follows:  
formula
(1)
where k is the number of variables and nc is the number of central points. The coded levels and the natural values of variables shown in Table 1 were initial ammonia concentration (X1), catalyst dosage (g/L) (X2), irradiation intensity (Watt) (X3) and stirring speed (rpm) (X4) that were coded as Xi according to the following equation:  
formula
(2)
where Xo is the value of Xi at the central point, and ΔXi is the step change.
Table 1

Design matrix for the central composite designs

Factors Low (–1) Central (0) High (+1) α + α 
Initial concentration ppm 20 30 40 10 50 
Dosage (g) 0.2 0.5 0.8 0.1 
Light intensity 12 18 24 30 
Stirring speed (rpm) 200 300 400 100 500 
Factors Low (–1) Central (0) High (+1) α + α 
Initial concentration ppm 20 30 40 10 50 
Dosage (g) 0.2 0.5 0.8 0.1 
Light intensity 12 18 24 30 
Stirring speed (rpm) 200 300 400 100 500 
The removal percentage dependency to all variables can be represented as follows:  
formula
(3)
where y is the response, β0, βi, βii are the regression coefficients of variables for intercept, linear, quadratic and interaction terms, respectively. Xi and Xj are the independent variables and ɛ is the residual term. The Minitab software (version 18) was used for data processing. The analysis of variance (ANOVA) was performed to justify the significance and adequacy of the developed regression model according to the determination coefficient (R2).

RESULTS AND DISCUSSION

Characterization of photo catalyst

FTIR analysis

The FTIR spectra of composites are shown in Figures 1 and 2. For the TiO2-C3N4 mixture, a few peaks relating to TiO2 are noticed. The broad band centered at 500–600 cm–1 is likely due to the vibration of the Ti–O bonds in the TiO2 lattice and the peaks in the 500–1,000 cm–1 region are allotted to the vibrations of the Ti–O and Ti–O–Ti framework bonds (Komatsu 2001a). Some strong bands in the 1,242–1,639 cm–1 area are the stretching modes of CN heterocyclic (Komatsu 2001a, 2001b, 2001c). The two absorption peaks at 1,300–1,412 and 1,529–1,639 cm–1 are appointed, individually, to C(sp2)–N (1,325 cm–1) and C(sp2) = N (1,639 cm–1) stretching modes in a graphite-type structure; such a band is not allowed in the FTIR spectrum of pure graphite single crystals. The broad peaks allocated at 3,100–3,600 cm–1 are assigned to vibrations of hydroxyl groups (Salavati-Niasari et al. 2011).

Figure 1

FTIR analysis of TiO2-C3N4.

Figure 1

FTIR analysis of TiO2-C3N4.

Figure 2

FTIR analysis of ZnO-C3N4.

Figure 2

FTIR analysis of ZnO-C3N4.

For the ZnO-C3N4 mixture, Figure 2 demonstrates that the peak at 1,632.2 cm–1 is related to Zn–O stretching, while the peak noticed at 3,446.3 cm–1 may be because of O–H stretching assigned to the water adsorption on the metal surface. The sharp peak positioned at 471.6 cm–1 is attributed to the Zn–O stretching bonds. The IR bands in the region of 1,700–600 cm–1 correspond to C = O, C–O and C–H vibrations respectively (Xiong et al. 2009). Several strong bands in the 1,242–1,639 cm–1 region are the stretching modes of CN heterocyclic (Komatsu 2001a, 2001b, 2001c). The two absorption peaks at 1,300–1,412 cm–1 and 1,529–1,639 cm–1 are allocated, respectively, to C(sp2)–N (1,387 cm–1) and C(sp2) = N (1,632 cm–1) stretching modes in a graphite-type structure (such a band is forbidden in the FTIR spectrum of pure graphite single crystals). The broad peaks appearing at 3,100–3,600 cm–1 are assigned to vibrations of hydroxyl groups (Salavati-Niasari et al. 2011).

SEM analysis

The surface morphology of TiO2-C3N4 and ZnO-C3N4 are shown in Figures 3 and 4. Figure 3 shows that the sample of TiO2-C3N4 appeared to have aggregated particles that contained many smaller crystals and well-crystallized C3N4 nanostructures that the TiO2 aggregates covering the C3N4 particles.

Figure 3

SEM analysis of TiO2-C3N4 with different magnification factor.

Figure 3

SEM analysis of TiO2-C3N4 with different magnification factor.

Figure 4

SEM analysis of ZnO-C3N4 with different magnification factor.

Figure 4

SEM analysis of ZnO-C3N4 with different magnification factor.

Figure 4 shows the sample of ZnO-C3N4. It is clearly seen that the ZnO-C3N4 composite appears to be spherical in shape with a smooth surface like agglomerate, which is composed of many small spherical nanoparticles which may be because the melamine plays a conglomeration role in the chemical precipitation process. It can be inferred that ZnO is uniformly distributed on the surface of the sphere-like composite, which favors the formation of heterojunction and results in an increase in the specific surface areas, which is beneficial to the photocatalytic activity. Particle size also plays a critical role in the photocatalytic activity which was found to increase with increasing particle size for spherical diameters smaller than 200 nm (Amano et al. 2013).

XRD analysis

Figures 5 and 6 show the XRD for TiO2-C3N4 and ZnO-C3N4 mixtures. For TiO2-C3N4 composite, it can be seen from Figure 5 that the peaks appeared at 2θ values, the diffraction peak at 2θ with 25.3, 38.3, 48, 54, 62, 74, 76 and 83° corresponds to the crystal planes of (101), (004), (200), (105), (204), (213), (107), (301) indicating the formation of anatase phase of TiO2. The peaks of the graph are in good agreement with the literature (Akarsu et al. 2006; Jiang et al. 2011). The location of the peaks was contrasted to values in literature and the existence of titanium dioxide particles was affirmed. Also, the figure shows that a strong peak at 2θ value with 26° corresponds to the crystal planes of (110), indicating the formation of α-C3N4.

Figure 5

XRD analysis of TiO2-C3N4.

Figure 5

XRD analysis of TiO2-C3N4.

Figure 6

XRD analysis of ZnO-C3N4.

Figure 6

XRD analysis of ZnO-C3N4.

For ZnO-C3N4 composite, it can be seen from Figure 6 that the peaks appeared at 2θ values, the diffraction peak at 2θ with 31.74, 36.83, 47.62, 56.7, 62.7, and 68° are observed corresponding to (100), (101), (102), (110), (103), and (112) crystal planes indicating the formation of ZnO, the peaks of the graph are in good agreement with the literature (Alwan et al. 2015). Also, all peaks are indexed according to the hexagonal phase of ZnO. There are strong peaks at 2θ values, the diffraction peak at 2θ with 32, 34, and 36° appeared to correspond with (321) crystal planes indicating the formation of Zn3N2 (Jiangyan et al. 2012).

Experimental design and optimization

The impact of four variables on ammonia removal from aqueous solutions was researched by conduction of a total of 31 runs (see Table 2); their results were analyzed by ANOVA to obtain an empirical equation that can predict the real behavior of the adsorption system.

Table 2

The experimental design with variables response

Run no. Initial concentration (ppm) Dosage (g) Light intensity Stirring speed (rpm) % Removal (observed) % Removal (predicted) 
10 12 300 44.3 42.46 
20 12 300 36.4 37.17 
30 18 300 35.1359 35.14 
40 12 300 30.1 30.87 
50 12 300 28.6 29.85 
10 0.2 12 300 20.3 20.54 
10 0.5 300 29.4451 29.45 
10 0.8 12 300 34.9 39.34 
10 1.5 12 300 43.6 43.38 
10 20 0.2 12 300 15.2 16.38 
11 20 0.5 12 300 31.1 27.13 
12 20 0.8 12 300 31.3 34.33 
13 20 1.5 12 300 37.1 37.38 
14 30 0.2 12 300 12.7 13.64 
15 30 0.5 12 300 28.5 23.96 
16 30 0.8 12 300 29.9 30.75 
17 30 1.5 12 300 32.3 32.81 
18 40 0.2 12 300 10.7 12.33 
19 40 0.5 12 300 25.6 22.23 
20 40 0.8 12 300 25.9 28.59 
21 40 1.5 12 300 29.7 29.67 
22 50 0.2 12 300 10.7 12.44 
23 50 0.5 12 300 25.6 21.92 
24 50 0.8 12 300 25.9 27.86 
25 50 1.5 12 300 29.7 27.95 
26 20 24 300 36 36.00 
27 10 24 300 50.9 50.90 
28 10 12 100 45.8 47.25 
29 10 12 200 47.2 43.33 
30 10 30 400 60.3174 60.32 
31 10 12 500 50.4 49.92 
Run no. Initial concentration (ppm) Dosage (g) Light intensity Stirring speed (rpm) % Removal (observed) % Removal (predicted) 
10 12 300 44.3 42.46 
20 12 300 36.4 37.17 
30 18 300 35.1359 35.14 
40 12 300 30.1 30.87 
50 12 300 28.6 29.85 
10 0.2 12 300 20.3 20.54 
10 0.5 300 29.4451 29.45 
10 0.8 12 300 34.9 39.34 
10 1.5 12 300 43.6 43.38 
10 20 0.2 12 300 15.2 16.38 
11 20 0.5 12 300 31.1 27.13 
12 20 0.8 12 300 31.3 34.33 
13 20 1.5 12 300 37.1 37.38 
14 30 0.2 12 300 12.7 13.64 
15 30 0.5 12 300 28.5 23.96 
16 30 0.8 12 300 29.9 30.75 
17 30 1.5 12 300 32.3 32.81 
18 40 0.2 12 300 10.7 12.33 
19 40 0.5 12 300 25.6 22.23 
20 40 0.8 12 300 25.9 28.59 
21 40 1.5 12 300 29.7 29.67 
22 50 0.2 12 300 10.7 12.44 
23 50 0.5 12 300 25.6 21.92 
24 50 0.8 12 300 25.9 27.86 
25 50 1.5 12 300 29.7 27.95 
26 20 24 300 36 36.00 
27 10 24 300 50.9 50.90 
28 10 12 100 45.8 47.25 
29 10 12 200 47.2 43.33 
30 10 30 400 60.3174 60.32 
31 10 12 500 50.4 49.92 
Analyses of experimental results gave the following empirical relationship:  
formula
(4)

The determination coefficient and residuals of the ANOVA in Table 3 were used to check the statistical adequacy of the model; in which the F test indicates the relationship between the mean square and the residual error of the model.

Table 3

Analysis of variance (ANOVA)

Source DF Adj SS Adj MS F-value P-value 
Model 12 4052.97 337.747 44.40 0.000 
Linear 936.04 234.011 30.76 0.000 
Initial concentration (ppm) 111.14 111.135 14.61 0.001 
Dosage (g) 39.90 39.904 5.25 0.034 
Light intensity 26.41 26.413 3.47 0.079 
Stirring speed (rpm) 36.19 36.193 4.76 0.043 
Square 605.98 151.496 19.91 0.000 
Initial concentration (ppm)*Initial concentration (ppm) 32.07 32.073 4.22 0.055 
Dosage (g)*Dosage (g) 329.16 329.162 43.27 0.000 
Light intensity*Light intensity 22.94 22.938 3.02 0.100 
Stirring speed (rpm)*Stirring speed (rpm) 48.95 48.951 6.43 0.021 
2-Way interaction 93.23 23.307 3.06 0.043 
Initial concentration (ppm)*Dosage (g) 18.61 18.609 2.45 0.135 
Initial concentration (ppm)*Light intensity 42.57 42.573 5.60 0.029 
Dosage (g)*Light intensity 23.23 23.226 3.05 0.098 
Light intensity*Stirring speed (rpm) 31.53 31.529 4.14 0.057 
Lack of fit 10 89.138 8.914 1.7497 0.254792 
Error 18 136.94 7.608   
Total 30 4189.90    
Source DF Adj SS Adj MS F-value P-value 
Model 12 4052.97 337.747 44.40 0.000 
Linear 936.04 234.011 30.76 0.000 
Initial concentration (ppm) 111.14 111.135 14.61 0.001 
Dosage (g) 39.90 39.904 5.25 0.034 
Light intensity 26.41 26.413 3.47 0.079 
Stirring speed (rpm) 36.19 36.193 4.76 0.043 
Square 605.98 151.496 19.91 0.000 
Initial concentration (ppm)*Initial concentration (ppm) 32.07 32.073 4.22 0.055 
Dosage (g)*Dosage (g) 329.16 329.162 43.27 0.000 
Light intensity*Light intensity 22.94 22.938 3.02 0.100 
Stirring speed (rpm)*Stirring speed (rpm) 48.95 48.951 6.43 0.021 
2-Way interaction 93.23 23.307 3.06 0.043 
Initial concentration (ppm)*Dosage (g) 18.61 18.609 2.45 0.135 
Initial concentration (ppm)*Light intensity 42.57 42.573 5.60 0.029 
Dosage (g)*Light intensity 23.23 23.226 3.05 0.098 
Light intensity*Stirring speed (rpm) 31.53 31.529 4.14 0.057 
Lack of fit 10 89.138 8.914 1.7497 0.254792 
Error 18 136.94 7.608   
Total 30 4189.90    
R-sq R-sq(adj) 
2.75820 96.73% 94.55% 
R-sq R-sq(adj) 
2.75820 96.73% 94.55% 

The fitness of the model was also indicated by its high R2 value (0. 967) and adjusted R2 value of 0.946. According to a P-value of 0.05, it is concluded that initial concentration dosage stirring speed and quadratic dosage are important terms. The plot of measured versus calculated values of removal (%) demonstrate a solid match (Figure 7) with R2 of 0.967. This implies that 967% of the variations for percent removal are explained by the independent variables and the presence of a linear relationship between them with high correlation coefficient shows the normal distribution of error, the great applicability of model for clarification of exploratory information. These plots are required to check the ordinariness supposition in the fitted model. Adjusted R2 is also a measure of goodness of a fit. Here, the Adj–R2 value (0.946) was very close to the corresponding R2 value.

Figure 7

The actual data versus predicted data for removal of ammonia.

Figure 7

The actual data versus predicted data for removal of ammonia.

Table 4 shows the regression results of the predicted response surface quadratic model for the ammonia removal by TiO2/C3N4 in the form of ANOVA. ANOVA is required to test the significance of the model. ANOVA is used to show whether the variation from the model is significant or not. This is performed by F-value. If the model is a good predictor of the experimental values, the F-value should be greater than the tabulated value of F at a level of significance α. The F-value obtained, 44.4, is clearly greater than the tabulated F (2.34 at 95% significance) confirming the adequacy of the model fits.

Table 4

Regression coefficients

Term Coef SE coef T-value P-value 
Constant 26.28 3.23 8.13 0.000 
Initial concentration (ppm) −15.49 4.05 −3.82 0.001 
Dosage (g) 40.3 17.6 2.29 0.034 
Light intensity −24.9 13.4 −1.86 0.079 
Stirring speed (rpm) 17.77 8.15 2.18 0.043 
Initial concentration (ppm)*Initial concentration (ppm) 2.85 1.39 2.05 0.055 
Dosage (g)*Dosage (g) −8.31 1.26 −6.58 0.000 
Light intensity*Light intensity −28.9 16.6 −1.74 0.100 
Stirring speed (rpm)*Stirring speed (rpm) 6.12 2.41 2.54 0.021 
Initial concentration (ppm)*Dosage (g) −1.83 1.17 −1.56 0.135 
Initial concentration (ppm)*Light intensity −19.22 8.12 −2.37 0.029 
Dosage (g)*Light intensity 61.4 35.1 1.75 0.098 
Light intensity*Stirring speed (rpm) 32.9 16.2 2.04 0.057 
Term Coef SE coef T-value P-value 
Constant 26.28 3.23 8.13 0.000 
Initial concentration (ppm) −15.49 4.05 −3.82 0.001 
Dosage (g) 40.3 17.6 2.29 0.034 
Light intensity −24.9 13.4 −1.86 0.079 
Stirring speed (rpm) 17.77 8.15 2.18 0.043 
Initial concentration (ppm)*Initial concentration (ppm) 2.85 1.39 2.05 0.055 
Dosage (g)*Dosage (g) −8.31 1.26 −6.58 0.000 
Light intensity*Light intensity −28.9 16.6 −1.74 0.100 
Stirring speed (rpm)*Stirring speed (rpm) 6.12 2.41 2.54 0.021 
Initial concentration (ppm)*Dosage (g) −1.83 1.17 −1.56 0.135 
Initial concentration (ppm)*Light intensity −19.22 8.12 −2.37 0.029 
Dosage (g)*Light intensity 61.4 35.1 1.75 0.098 
Light intensity*Stirring speed (rpm) 32.9 16.2 2.04 0.057 

Photocatalytic degradation of ammonia

Effects of ammonia initial concentrations

Figure 8 demonstrates the impact of ammonia concentration on the removal percentage against time at constant stirring speed (300 rpm), time (50 min), light intensity (12 W), and dosage of catalyst (0.2 g/L).

Figure 8

Effect of ammonia initial concentration on the degradation of ammonia using TiO2/C3N4 at stirring speed = 300 rpm, time = 50 min, light intensity = 12 W, and dosage of catalyst = 0.2 g/L.

Figure 8

Effect of ammonia initial concentration on the degradation of ammonia using TiO2/C3N4 at stirring speed = 300 rpm, time = 50 min, light intensity = 12 W, and dosage of catalyst = 0.2 g/L.

It can be concluded that the degradation rate was high at the beginning of the photocatalytic reaction, and then it became low at the last 10 min. The rate of removal also decreases with the increasing concentration of ammonia in the simulated water. The removal at its initial concentration of 10 mg/L reached up to the maximum value of removal percentage.

From previous investigations, the reasons for this may be due to the fact that at lower ammonia concentration the photocatalytic oxidation was mainly governed by the adsorption of NH4 + on TiO2/C3N4, any increase in initial concentration led to an increase in the amount of adsorption of ammonia. On the other hand, at very high concentrations the titanium dioxide surface becomes saturated owing to reaching adsorption/desorption equilibrium, and the photonic efficiency reduces, thus leading to catalyst deactivation.

Effect of catalyst dosage

The impact of the dosage of TiO2/g-C3N4 composite on the photocatalytic degradation of ammonia was examined at a constant stirring speed (300 rpm), time (50 min), light intensity (12 W), and concentration of 10 ppm. The results in Figure 9 show that the optimal dosage of TiO2/g-C3N4 composite was found to be 1 g/L. A possible reason is that when the initial dosage of TiO2/g-C3N4 composite powder is increased, the amount of involved photolysis reaction of TiO2/g-C3N4 composite particles is increased, and the degradation rate is also increased. However, higher catalyst amounts make solid particles hinder the UV light, and light scattering increases (not enough light reaches the surface). Hence, the photo degradation efficiency is reduced (Altomare et al. 2014).

Figure 9

Effect of catalyst dosage on degradation of ammonia at stirring speed = 300 rpm, time = 50 min, light intensity = 12 W, and concentration = 10 ppm.

Figure 9

Effect of catalyst dosage on degradation of ammonia at stirring speed = 300 rpm, time = 50 min, light intensity = 12 W, and concentration = 10 ppm.

Effect of light intensity

As shown in Figure 10, by utilizing the optimum dosage and optimum ammonia initial concentration (at a constant stirring speed (300 rpm), time (50 min), catalyst dosage1 g/L, and concentration of 10 ppm), it was seen that while when increasing the light intensity a remarkable rate of removal is noticed, due to the huge increment of the light propagation across the reactor. The UV irradiation produces the photons required for the electron transfer from the valence band to the conduction band of a semiconductor photocatalyst, and the energy of the photon is dependent on the light intensity. Also, as more radiations fall on the catalyst surface and the rate of removal increases, the more hydroxyl radicals are produced. The enhancement of removal rate is due to an increase in hydroxyl radical concentration, hence the increase in the light intensity increases the removal (Nasser et al. 2009).

Figure 10

Effect of light intensity on the degradation of ammonia using TiO2/C3N4 at stirring speed = 300 rpm, time = 50 min, catalyst dosage = 1 g/L, and concentration = 10 ppm.

Figure 10

Effect of light intensity on the degradation of ammonia using TiO2/C3N4 at stirring speed = 300 rpm, time = 50 min, catalyst dosage = 1 g/L, and concentration = 10 ppm.

Effect of stirring speed on the percentage removal of ammonia

Changing the speed of mixing at time 50 min, catalyst dosage1 g/L, light intensity (12 W) and concentration of 10 ppm, an increase in the removal rate was observed as demonstrated in Figure 11. Photocatalysis is controlled by two steps, mass-transfer and chemical reaction. The mass transfer step is affected by the mixing speed. So, increasing of the mixing speed leads to a higher mass transfer and a high degradation rate. Also, increasing the mixing speed can promote oxygen transfer on the liquid phase (Merabet et al. 2009), which leads to an increase in the degradation rate.

Figure 11

Effect of stirring speed on the degradation of ammonia using TiO2/C3N4 at light intensity = 12 W, time = 50 min, catalyst dosage = 1 g/L, and concentration = 10 ppm.

Figure 11

Effect of stirring speed on the degradation of ammonia using TiO2/C3N4 at light intensity = 12 W, time = 50 min, catalyst dosage = 1 g/L, and concentration = 10 ppm.

Comparison of photocatalytic activity of the two composites

To understand the efficiency of different composites, trials were carried out at fixed conditions of initial concentration, catalyst weight, light intensity and stirring speed. The efficiency of TiO2/g-C3N4 and ZnO/g-C3N4 were examined and the results are presented in Figure 12. The results clearly indicate that ZnO/g-C3N4 was found to be the most active in the degradation of ammonia. This could be because the recombination of photo excited electrons in the large particles is slower than that in the small particles which concludes that the fast recombination in small particles occurred on the surface which led to an increase in the photocatalytic activity. Doping with ions and heterojunction coupling has been reported to enhance separation of the electron-hole pair and was found to reduce recombination and therefore enhance the interfacial charge transfer efficiency which leads to an improvement in the photocatalytic activity (Pelaez et al. 2012).

Figure 12

Comparison of efficiency of TiO2/C3N4 and ZnO/C3N4 composites on the degradation of ammonia at stirring speed = 300 rpm, time = 50 min, catalyst dosage = 1 g/L, light intensity = 12 W and concentration = 10 ppm.

Figure 12

Comparison of efficiency of TiO2/C3N4 and ZnO/C3N4 composites on the degradation of ammonia at stirring speed = 300 rpm, time = 50 min, catalyst dosage = 1 g/L, light intensity = 12 W and concentration = 10 ppm.

Response surface methodology

The three-dimensional response surface plots give well-recognized knowledge about the main effect of four variables (Figure 13). To study the impact of the initial concentration on the removal efficiency, some experiments with 0.2–1.5 g of adsorbent at light intensity ranging from 6 to 30 W with a stirring speed ranging from 100 to 500 rpm were designed. The results in Figure 13(a) indicate that raising the catalyst dosage causes a significant improvement in the removal efficiency which emerged from an increase in available surface and the presence of active surface area of the catalyst surfaces that enables ammonia adsorption. However, increasing ammonia initial concentration is joined by a decrease in the removal efficiency. To concentrate the effect of the initial concentration on the adsorption efficiency, a few runs with concentrations of 10–50 mg L–1 adsorbent at a light intensity of 12 W and stirring speed of 100–500 rpm were designed and the results are displayed in Figure 13(b).

Figure 13

Response surfaces for the ammonia removal: (a) initial concentration–dosage; (b) initial concentration–stirring speed; (c) dosage–stirring speed.

Figure 13

Response surfaces for the ammonia removal: (a) initial concentration–dosage; (b) initial concentration–stirring speed; (c) dosage–stirring speed.

The noticed decreases in removal percentage at higher initial concentrations arise from the lower ratio of vacant sites on catalyst for ammonia molecules that are competing for binding to the surface. This means that there are not enough sites for all molecules in high ammonia concentrations. As shown in Figure 13(c), the more catalyst quantity associated with high adsorption efficiency and increased stirring speed leads to better and higher removal efficiency.

CONCLUSIONS

Different photocatalysts (ZnO/g-C3N4, TiO2/g-C3N4) were prepared using different precursors. The composites were characterized by Fourier transform infrared spectroscopy (FTIR), scanning electron microscope (SEM) and X-ray powder diffraction (XRD). Their efficiency was researched by the degradation of an ammonia solution. The use of RSM based on CCD was applied for optimization of parameters such as initial concentration, catalyst dosage, light intensity and stirring speed. The results indicate that the photocatalytic degradation of the ammonia solution, after 50 min of UV irradiation, can reach percentages of 46%, and 52% using the catalysts TiO2/g-C3N4 and ZnO/g-C3N4, respectively, which suggests that ZnO/g-C3N4 has a higher degradation efficiency.

ACKNOWLEDGEMENTS

This work was financially supported by the Petrochemical Engineering Department, Pharos University, Alexandria, Egypt.

REFERENCES

REFERENCES
Achouri
,
F.
,
Corbel
,
S.
,
Aboulaich
,
A.
,
Balan
,
L.
,
Ghrabi
,
A.
,
Said
,
M. B.
&
Schneider
,
R.
2014
Aqueous synthesis and enhanced photocatalytic activity of ZnO/Fe2O3 heterostructures
.
J. Phys. Chem. Solids
75
(
10
),
1081
1186
.
Akarsu
,
M.
,
Asiltürk
,
M.
,
Sayilkan
,
F.
,
Kiraz
,
N.
,
Arpac
,
E.
&
Sayilkan
,
H.
2006
A novel approach to the hydrothermal synthesis of anatase titania nanoparticles and the photocatalytic degradation of rhodamine B
.
Turk. J. Chem.
30
,
333
343
.
Altomare
,
M.
,
Dozzia
,
M. V.
,
Chiarelloa
,
G. L.
,
Paolab
,
A. D.
,
Palmisanob
,
L.
&
Selli
,
E.
2014
High activity of brookite TiO2 nanoparticles in the photocatalytic abatement of ammonia in water
.
Catal. Today
252
,
184
189
.
Alwan
,
R. M.
,
Kadhim
,
Q. A.
,
Sahan
,
K. M.
,
Ali
,
R. A.
,
Mahdi
,
R. J.
,
Kassim
,
N. A.
&
Jassim
,
A. N.
2015
Synthesis of zinc oxide nanoparticles via sol–gel route and their characterization
.
Nanosci. Nanotechnol
.
5
(
1
),
1
6
.
Amano
,
F.
,
Ishinaga
,
E.
&
Yamakata
,
A.
2013
Effect of particle size on the photocatalytic activity of WO3 particles for water oxidation
.
J. Phys. Chem. C
117
(
44
),
22584
22590
.
Dong
,
S. S.
,
Zhang
,
J. B.
,
Gao
,
L. L.
,
Wang
,
Y. L.
&
Zhou
,
D. D.
2012
Preparation of spherical activated carbon supported and Er3+:YAlO3 doped TiO2 photocatalyst for methyl orange degradation under visible light
.
Trans. Nonferrous Met. Soc. China
22
,
2477
2483
.
Gupta
,
V. K.
,
Sadegh
,
H.
,
Yari
,
M.
,
Ghoshekandi
,
R. S.
,
Maazinejad
,
B.
&
Chahardori
,
M.
2015
Removal of ammonium ions from wastewater: a short review in development of efficient method
.
Glob. J. Environ. Sci. Manage
.
1
(
2
),
149
158
.
Jiang
,
H. B.
,
Cuan
,
Q.
,
Wen
,
C. Z.
,
Xing
,
J.
,
Wu
,
D.
,
Gong
,
X.-Q.
,
Li
,
C.
&
Yang
,
H. G.
2011
Anatase TiO2 crystals with exposed high-index facets
.
Angew. Chem. Int. Edn.
50
(
16
),
3764
3768
.
Li
,
X. F.
,
Zhang
,
J.
&
Shen
,
L. H.
2009
Preparation and characterization of graphitic carbon nitride through pyrolysis of melamine
.
Appl. Phys. A Mater. Sci. Proc.
94
,
387
392
.
Luo
,
X.
,
Yan
,
Q.
,
Wang
,
C.
,
Luo
,
C.
,
Zhou
,
N.
&
Jian
,
C.
2015
Treatment of ammonia nitrogen wastewater in low concentration by two-stage ozonization
.
Int. J. Environ. Res
.
12
(
9
),
11975
11987
.
Miranda
,
C.
,
Mansilla
,
H.
,
Yánez
,
J.
,
Obregόn
,
S.
&
Colόn
,
G.
2013
Improved photocatalytic activity of g-C3N4/TiO2 composites prepared by a simple impregnation method
.
J. Photochem. Photobiol. A Chem.
253
,
16
21
.
Nasser
,
M.
,
Behnajady
,
A. M.
,
Oskui
,
J.
&
Reza
,
M.
2009
Investigation of the efficiency of ZnO photocatalyst in the removal of p-nitrophenol from contaminated water, Iran
.
J. Chem. Chem. Eng.
28
(
1
),
49
55
.
Oliveira
,
R.
,
Almedia
,
M. F.
,
Santos
,
L.
&
Madeira
,
L. M.
2006
Experimental design of 2,4-dichlorophenol oxidation by Fenton's reaction
.
Ind. Eng. Chem. Res.
45
,
1266
1276
.
Pelaez
,
M.
,
Nolan
,
N. T.
,
Pillai
,
S. C.
,
Seery
,
M. K.
,
Falaras
,
P.
,
Kontos
,
A. G.
&
Entezari
,
M. H.
2012
A review on the visible light active titanium dioxide photocatalysts for environmental applications
.
Appl. Catal. B
125
,
331
349
.
Shavisi
,
Y.
,
Sharifnia
,
S.
,
Hosseini
,
S. N.
&
Khadivi
,
M. A.
2013
Application of TiO2/perlite photocatalysis for degradation of ammonia in wastewater
.
J. Ind. Eng. Chem
.
20
(
1
),
278
283
.
Xiong
,
H.-M.
,
Shchukin
,
D. G.
,
Möhwald
,
H. X. Y.
&
Xia
,
Y. Y.
2009
Sonochemical synthesis of highly luminescent zinc oxide nanoparticles doped with magnesium (II)
.
Angew. Chem. Int. Edit.
48
,
2727
2731
.
Zhou
,
Y.
,
Xiao
,
B.
,
Liu
,
S.-Q.
,
Meng
,
Z.
,
Chen
,
Z.-G.
,
Zou
,
C.-Y.
,
Liu
,
C.-B.
,
Chen
,
F.
&
Zhou
,
X.
2016
Photo-Fenton degradation of ammonia via a manganese–iron double-active component catalyst of graphene–manganese ferrite under visible light
.
Chem. Eng. J.
283
,
266
275
.