Introduction

Pigments and dyes are widely used in various industries and have become an important part of daily life (Coruh and Elevli 2014; Ertugay and Acar 2014; Bhatt et al. 2012; Tunc et al. 2012). Among the various types of the dye, the azo dyes are large class of synthetic organic dyes used in various industries. It has been estimated that azo dye accounts for approximately 60–70% use in food and textile industries (Tunc et al. 2012; Salehi et al. 2016). The effluent with azo dyes into the aqueous systems is of severe concern due to its hazardous effects on human activities and for aquatic ecosystems (Puvaneswari et al. 2006; Sawant and Cho 2016). Hence, it is important to treat azo dye wastewater before discharge into the aquatic bodies or environment. There are several traditional methods already in use to remove organic synthetic dyes from wastewaters, such as physical (Asgher and Bhatti 2012; Nasuha et al. 2010), biological (Jamal et al. 2012), electrochemical (Ghalwa et al. 2012), and chemical processes (Hua et al. 2013). However, these all techniques have some advantages and limitations. The advanced oxidation treatment processes (AOPs) have received a great attention since they are able to deal with various types of dye pollutants and many other organics in wastewater (Anouzla et al. 2009). ZnO is a well-known material utilized in the degradation of organic pollutants in aqueous system (Parsa et al. 2014; Anouzla et al. 2009). It is widely used due to the property of narrow band gap (~3.37 eV), non-toxicity, low cost of manufacturing, and high initial rate of photocatalysis (Parsa et al. 2014). However, it is observed active under ultraviolet spectrum, which limits its potential application in environmental remediation (Li et al. 2013a). Thus, it is crucial to tailor the structure and morphology of ZnO through suitable strategies in order to alter its bulk surface properties. A suitable material as an additive or dopant has been reported as a feasible way to tailor the photocatalytic performance of ZnO under visible spectrum (Shahmoradi et al. 2010). Copper (II) oxide is a p-type semiconductor used to improve the photocatalytic efficiencies of wide band gap semiconductors such as TiO2, ZnO, and SnO2 (Huang et al. 2013, Mageshwari et al. 2015). The CuO–ZnO-mixed metal oxide nanoparticles are more efficient than the pure ZnO, and this is because there is condensed recombination of generated electron–hole pairs in the coupled system (Witoon et al. 2013). Coupling between two semiconductors with suitable band locations, such as ZnO and CuO, can extend the ZnO photoapproachable range to visible light and shift the photocatalytic procedure to the solar light range; (Saravanan et al. 2013; Gajendiran and Rajendran 2014). There are many reports on the synthesis of ZnO–CuO nanoparticles such as co-precipitation (Gajendiran and Rajendran 2014), thermal decomposition (Saravanan et al. 2013), sol–gel (Habibi and Rahmati 2015), hydrothermal (Li et al. 2013b), and photodeposition methods (Wang et al. 2011). Among all these methods, visible light-induced hydrothermal growth technique for synthesis of nanostructures composite is a promising and well-known synthesis method due to its characteristics including mild temperature synthesis, accurate control of composition, and controllable particle size (Maleki and Shahmoradi 2012). The hydrothermal process has several advantages over other synthesis methods including better nucleation control, simple equipment, catalyst-free growth, one-step synthesis without high-temperature calcination, and milling. In addition, the synthesized material showed a low level of aggregation and ability to obtain amorphous phase, less hazardous by-products, no need for post-treatment (Zazouli et al. 2015; Moghaddam et al. 2011).

Traditional approaches for evaluating photodegradation process are dependent on the change of one independent variable parameter such as dosage of catalyst, pH of solution, concentration of dye, while keeping all other variables constant (Shojaeimehr et al. 2014; Zazouli et al. 2015). Thus, traditional approaches resulted in additional consumption of chemicals for evaluating each parameter and eventually increased time and cost of study (Zazouli et al. 2015; Shojaeimehr et al. 2014; Moghaddam et al. 2011). In attempt to solve this issue, statistical programs such as RSM have been utilized to design of experiments and parameters optimization of process control. (Ramakrishna and Susmita 2012). The main functionality of RSM is to build the relations between numerous independent variables and responses (Danbaba et al. 2015).

RSM has been effectively utilized in various fields such as adsorption, Fenton’s oxidation, electro-coagulation, and photocatalytic decolonization (Bhaumik et al. 2013; Bhatti et al. 2011; Witek-Krowiak et al. 2014; Khataee et al. 2010; Wu et al. 2012; Khataee et al. 2011). The present study investigates the effectiveness of modeling of RSM and to optimize the influencing factors on DB71 degradation using RSM approach.

Materials and methods

Materials

CuO–ZnO, trimethylamine, and NaOH were purchased from Merck, Germany. Direct Blue 71 dye was bought from AlvanSabet Co., Iran. After dilution, the specific concentration was used as a model pollutant from textile industry. Table S1 (Supplementary data) shows the chemical structure and some characteristics properties of Direct Blue 71.

Experimental methods

CuO–ZnO nanocomposites were synthesized under a mild hydrothermal condition (temperature = 150 °C, pressure = autogenous, time = 12 h). ZnO of three different concentrations (1, 2, and 3 mol), by keeping 1 mol of CuO and 10 ml of 1 mol NaOH, was added into a Teflon reactor (V fill = 10 ml). A fixed concentration (1 mL) of trimethylamine was added to the above mentioned mixture, and the mixture was stirred vigorously for a few minutes. The Teflon reactor was then fixed into an autoclave. Then autoclave was sealed and reacted in an oven at 150 °C for 12 h. After 12 h of reaction, the autoclave reactor was kept at room temperature to air-quenched. The as-synthesized nanocomposite from the Teflon reactor was moved to a pre-cleaned beaker for several time washing with distilled water and then allowed to settle down. After filtration the residue was dried at room temperature and stored prior to further use.

Photocatalytic degradation of DB71

The dye degradation efficiencies of the as-synthesized composites materials were evaluated by the photocatalytic degradation of DB71 under visible light. The tests were performed by reacting different amounts of CuO–ZnO composite (0.5–2.5 g/L) in 200 ml of the DB71 solution (20–60 mg/L) at various pH levels (3–11). The initial pH of dye solution was adjusted using a pH meter with addition of 0.1 N NaOH or HCl. Prior to exposure to natural sunlight, the DB71 solution was stirred for 30 min in dark to achieve the equilibrium between the composite and dye molecules. Subsequently, the solution with CuO–ZnO nanocomposites was illuminated under sunlight. At each scheduled time intervals, 5-ml dye solution sample was taken out from the reactor. Then, each solution sample was centrifuged for 10 min at 5000 rpm. Then the separated solutions were analyzed by UV–Vis spectrophotometer at 687 nm. The degradation efficiency removal (R%) of the CuO–ZnO nanocomposite on the photodegradation of DB71 dye was calculated using Eq. 1.

$${\text{Degradation}}\;{\text{efficiency}}\;\left( {R\% } \right) = \left( {A_{0} - A} \right)/A_{0} \times 100$$
(1)

where A 0 is the initial concentration (mg/L) of dye and A is the concentration of dye at time t (Salehi et al. 2014).

Experimental design and data analysis

The central composite design (CCD)-based response surface methodology (RSM) is widely accepted for investigational design, modeling, and optimization. In this study, CCD statistical software design expert version 7 was used to optimize the several operational parameters for the degradation of DB71 by CuO–ZnO composite under natural sunlight. The total 31 runs were considered at 5 levels containing of 16 factorial points, 8 axial points (α = 2), and seven replications at the center point (Table 1). The experiments were conducted in a random. The degradation percentage of DB71 as dependent variable, and other variable such as dosage of catalyst, pH of solution, contact time, and concentration of dye were chosen as independent variables in experiments.

Table 1 Experimental range and levels of independent process variables
Full size table

After completion of design of experiment, the photodegradation experiments of DB71 dye were carried out to obtain an appropriate model. The obtained data were examined by quadratic models. The general form of the quadratic models is shown as follow (Eq. 2):

$$\eta = b_{0} + \sum\limits_{i = 1}^{n} {b_{i} x_{i} + } \sum\limits_{i = 1}^{n} {b_{ii} x_{{_{i} }}^{2} + \sum\limits_{i = 1}^{n - 1} {\sum\limits_{i = 1}^{n} {b_{ij} x_{i} x_{j} } } }$$
(2)

where η is the predicted response variable, b 0 is the coefficient constant, b i is the linear coefficient, b ii is the quadratic coefficient, b ij is the interaction coefficient, and x i and x j are the coded values of the parameters (Salehi et al. 2014).

Results and discussion

Characterization of CuO–ZnO nanocomposite

The structural and textural properties of CuO–ZnO nanocomposite were investigated by various characterization techniques. Figure 1a represents the X-ray diffraction pattern of CuO–ZnO nanocomposite. A series of (100), (002), (101), and (102) reflections at 2θ = 31.8, 34.5, 36.3, and 47.61°, were observed in the ZnO sample, which is identical with arrangement with the wurtzite structure (JCPDS file no. 36–1451). In addition, the peaks at 2θ value of 35.5° and 38.7° belong to the CuO peaks could be easily noticed on the XRD. Peaks appears in the pattern confirms the presence of patterns of ZnO and CuO both materials in nanocomposite with and changed lattice of ZnO. The average particle size of CuO–ZnO was calculated by the Scherer formula (Eq. 3). The average size was calculated as 36 nm.

$$D_{\text{Scherrer}} = \frac{k\lambda }{{\beta \;{ \cos }\theta }}$$
(3)

where D represents the average crystallite size, λ for the radiation wavelength (1.5418 Å), k is related to the crystallite shape (k = 0.089), β is the peak width at half maximum, and θ is the Bragg diffraction angle.

Fig. 1
figure 1

a XRD pattern of CuO–ZnO nanocomposite, b SEM image, and c EDAX analysis of CuO–ZnO nanocomposite

Full size image

The Fourier transform infrared spectroscopy (FTIR) analysis was performed to study the significance of CuO on the surface chemistry of CuO–ZnO nanocomposite and also to determine the presence of functional group impurity due to surface modifier. Figure S1 (Supplementary data) shows the FTIR profile of ZnO and CuO–ZnO nanocomposite. The spectra observed at 540 cm−1 represent the Zn–O, whereas the peak at 950 cm−1 represents the Cu–O stretching frequencies. In FTIR spectra, no new functional group peak was observed on the surface of CuO–ZnO nanocomposite indicates the whole detachment of surface modifier by several cycle of washing.

The surface morphology analysis of CuO–ZnO nanocomposite was performed by scanning electron microscopy (SEM) studies. Figure 1[B] reveals that the nanocomposites synthesized have different size and heterogeneous morphology; among them hexagonal and spherical nanocomposites are more obvious. Moreover, no aggregation was observed. Energy-dispersive spectrometry (EDAX) analysis of CuO–ZnO is suitable for confirming elemental analysis. As shown from Fig. 1c, the occurrence of the Zn, O, and Cu elements suggested the high purity of nanocomposite.

DB71 degradation studies

Effect of CuO–ZnO nanocomposite

To assess the consequence of the alteration of ZnO nanoparticles with CuO, the degradation proficiency test was carried out with reagent-grade ZnO and the synthesized CuO–ZnO nanocomposite. It was observed that the CuO–ZnO nanocomposite showed higher photocatalytic activity than reagent-grade ZnO (Fig. 2a). About 97% of the DB71 was degraded by CuO–ZnO after 180 min of exposure to natural sunlight, whereas only 7% of DB71 was degraded by reagent-grade ZnO under the identical conditions. The higher photocatalytic performance of CuO–ZnO nanocomposite can be explained by enhanced quantum efficiency of visible light absorption (Shahmoradi et al. 2012). Panneerselvam and coworkers evaluated the absorbance bands for simple ZnO and coupled semiconductor CuO–ZnO composite. The absorbance onset was at about 385 nm for unmodified ZnO particles, whereas CuO–ZnO composite displays blueshift. Additionally, a broad absorption band seen around 550–800 nm might be attributed due to the presence of Cu2+ species in the prepared CuO–ZnO nanocomposites (Sathishkumar et al. 2011). Hence, reagent-grade ZnO was not included as a variable for further photodegradation studies.

Fig. 2
figure 2

a Comparison of degradation of DB71 in the presence of reagent-grade ZnO and CuO–ZnO nanocomposite. (DB71 concentration 20 mg/L, pH = 7, contact time = 150 min, and catalyst dosage = 1.5 g/L), b Effect of CuO–ZnO molar ratio in the presence of 1.5 g/L of catalyst, an initial dye concentration of 20 mg/L and PH = 5 for 25 min

Full size image

Influence of CuO–ZnO molar ratio

The catalytic activities of the samples with variable molar ratios (1:1, 1:2, and 1:3) of CuO–ZnO nanocomposites were used for the degradations of DB71, and the observed results are presented in Fig. 2b. It can be observed from the obtained results that, the sample with molar ratio 1:3 of CuO–ZnO shows the highest photocatalytic activity toward DB71 degradations. Therefore, the entire photodegradation experiments were conducted using 1:3 CuO–ZnO nanocomposite.

Modeling and optimization of DB71 degradation

The investigational results of DB71 photodegradation using as-synthesized CuO–ZnO nanocomposite were examined through response surface methodology to attain an empirical model. Based on these consequences, the association between the degradation of DB71 and the independent variables parameters was achieved and stated by the second-order polynomial equation (Eq. 4):

$$\begin{aligned} R & = \, \left( { - 35.97} \right) + \left( {3.668X_{1} } \right) + \left( {63.023X_{2} } \right) + \left( {0.6833X_{3} } \right) + \left( {0.91328X_{4} } \right) + \left( { - 0.5906X_{1} X_{2} } \right) + \left( { - 0.01005X_{1} X_{3} } \right) \\ & \quad + \left( {0.023X_{1} X_{4} } \right) + \left( {0.0693X_{2} X_{3} } \right) + \left( { + 0.0218X_{1} X_{4} } \right) + \left( { - 0.2594X_{1}^{2} } \right) + \left( { - 15.86X_{2}^{2} } \right) + \left( { - 1.1429X_{3}^{2} } \right) + \left( { - \, 0.022304X_{4}^{2} } \right) \\ \end{aligned}$$
(4)

where R is response degradation percent, X 1, X 2, X3, and X 4 are the corresponding independent parameters such as pH, catalyst amount (g/l), reaction time (min), and DB71 initial concentration (mg/L), respectively.

Table 2 shows the outcomes of ANOVA study of the quadratic models. The significance of each coefficient was calculated by the F test and p value. The values of Prob > F less than 0.05 and more than 0.1000 indicate the model terms are not significant and significant, respectively. In Table 3, the terms are significant for X 1, X 2, and X 3, but those are not significant for X 1 X 2, X 2 X 3, X 2 X 4, X 3 X 4 according to Prob > F values. The “lack of fit F value” 0.2600 for the model obtained implies the deficiency of fit insignificancy, and the model is adequate. Acceptable precision is calculated for the signal-to-noise ratio, and a ratio larger than 4 is generally essential. In this research, precision obtained was 24.78 for the degradation percentage of DB71. The ratio determined was greater than 4, and the representative for the model can be used to navigate the design space (Moghaddam et al. 2011).

Table 2 ANOVA of the quadratic model of photo catalyst degradation of DB71 dye using CuO–ZnO nanocomposite
Full size table
Table 3 Optimum values of the process parameter for maximum efficiency
Full size table

Moreover, ANOVA indicated that the coefficient of determination (R 2) and adjusted R 2 were 0.9786 and 0.9585, respectively. The values of coefficient of determination and the adjusted R 2 are succeeded to estimate the fitting of the model. The obtained R 2 in DB71 degradation using CuO–ZnO nanocomposite indicated a stronger association between the factors and predicted results.

Figure 3 represents the connection between experimental and projected value for degradation of DB71 using CuO–ZnO composite. The obtained results showed that the experimental and predicted values for dye degradation were found higher, and the present model well fits the experimental data. Therefore, the model could be used to calculate the dye degradation in the investigational range of variables.

Fig. 3
figure 3

A plot of the predicted versus the experimental degradation efficiency

Full size image

Figure 4 represents the plots of residuals against normal probability for DB71 degradation. As shown in plot, obtained data show a linear relationship, which indicates distributed normal distribution of residuals. Thus, it is concluded that the prediction of the experimental results obtained from the developed quadratic model is satisfactory in the parameterization of the degradation of DB71.

Fig. 4
figure 4

A plot of the normal probability of the raw residuals

Full size image

Response surface plotting for estimation of operating variables

Influence of initial solution pH

The consequence of the pH on the DB71 degradation efficacy results is shown in Fig. 5, at a composite dosage of 1.5 g/L and early dye concentration ~20 mg/L. As shown in plot, the degradation efficiency (%) of DB71 increases from pH 3–7 and then reduces at the higher pH 11. The increase in pH from 3.0 to 7.0 resulted in an increase in DB71 degradation efficiency from 46.8 to 70.2%, and the efficiency decreases from 70.2 to 62% as the pH increases from 7 to 11. Such phenomenon may be observed due to the change in electrostatic magnetism or revulsion among DB71 molecules and the nanocomposite. The surface charge of nanocomposite has a significant effect on the adsorption and dissociation of DB71. The pHpZC (point of zero charge) of ZnO was already reported as 9 ± 0.3 (Shibin et al. 2014), which suggests that the nanocomposite surface must be positively charged at the pH of 3–7 and negative at the pH of 7–11. Given the fact that DB71 is an anionic dye, it has a high efficiency to be adsorbed onto the nanocomposite surface when the pH is <pHZPC. Therefore, the degradation efficiency of DB71 increases accordingly when pH is <pHZPC (Mohaghegh et al. 2014; Nilamadanthai et al. 2013).

Fig. 5
figure 5

Response surface plot and counter plot of the DB71 as a function of initial pH and reaction time (CuO–ZnO = 1.5 g/L, initial dye concentration = 20 mg/L)

Full size image

Influence of catalyst amount

Figure 6 shows the influence of nanocomposite amount on the degradation efficiency of DB71 concentration of 20 mg/L and a contact time of 150 min. As shown in profile the degradation productivity increases as the CuO–ZnO dosage increases from 0.50 to 2.0 g/L and then decreases as the dosage increases from 2 g/L to the maximum value of 2.5 g/L. These results showed that the increase in composite dosage improved the degradation efficiency of the dye by increasing the number of active sites for the catalytic events. Moreover, the degradation efficacy of the dye decreases due to the interruption of light by the aqueous suspension and agglomeration in nanocomposite when the loading of nanocomposite exceeds the limiting value (2 g/L). Additionally, at a higher level of catalyst dosage, the aggregation of catalyst causes a reduction in the active surface sites and increases the scattering of light, which eventually lead to the inhibition of the photon absorption of CuO–ZnO nanocomposite (Senthilraja et al. 2014; Harraz et al. 2014).

Fig. 6
figure 6

Response surface plot and counter plot of the DB71 as a function of initial catalyst dosage (g/L) and Initial pH, contact time = 150 min, initial dye concentration = 20 mg/L

Full size image

Influence of the dye concentration

The consequence of initial dye concentration on the photodegradation efficacy is shown in Fig. 7 at a contact time of 150 min and an initial pH of 8. The effect of concentration plot showed that the degradation efficiency decreased from 73 to 22% by enhancing the initial dye concentration from 20 to 60 mg/L. This was observed due to the presence of hydroxyl radical (OH) on the composite surface and the probability of hydroxyl radicals interacting with the dye molecule. Additionally, the decrease in degradation efficiency at higher dye concentration could be due to the reduction in the path length of photons entering the solution when the solution was intensively colored at higher dye concentration. Thus, fewer photons reached the nanocomposite surface (Harraz et al. 2014; Kaur et al. 2013).

Fig. 7
figure 7

Response surface plot and counter plot of the DB71 degradation as a function of initial dye concentration and initial catalyst dosage, pH = 7, reaction time = 150 min

Full size image

Process optimization

It is important to control the optimized circumstances of photodegradation in direction to enhance the degradation efficacy using developed mathematical model. Hence, the experimental variable parameters were set to the values in the considered range, while the response was set to a maximum level. In this approach, the optimal conditions for dye degradation using CuO–ZnO nanocomposite were determined to be contact time of 177.13 min, initial solution pH of 3.93, dye concentration of 24.34 mg/L, and a dosage of catalyst of 1.85 g/L. Under these conditions, a high degradation efficiency of 89.58% was achieved (Table 3).

Conclusion

In this research, the synthesis, characterization, modeling, and optimization of CuO–ZnO nanocomposite for the degradation DB71 was investigated. The prepared CuO–ZnO nanocomposite was characterized using XRD, FTIR, SEM, and EDS. The CCD was effectively applied for the design of experiment and analysis of results. The obtained correlation coefficient R 2 was found 0.9786 for DB71, indicating that the experimental data fit well with the modeling data obtained from the as-developed quadratic models. The influence of solution pH, dosage of catalyst, contact time, and dye concentration at optimal experimental conditions was evaluated. The optimum photodegradation efficiency (89.58%) was found at an initial pH of 3.93, dosage of catalyst of 1.85 g/L, dye concentration of 20.34 mg/L, and reaction time of 177.13 min. The investigational values were found promising with the predicted values from the developed quadratic model. This study sheds light on the potential applications of response surface methodology in the modeling and optimization of photodegradation of aromatic dyes.