Pb doped ZnO nanoparticles for the sorption of Reactive Black 5 textile azo dye

ANOVA

Analysis of variance

BET

Brunauer, Emmett and Teller

CCD

Central composite design

Df

Degrees of freedom

LOF

Lack of fit

MS

Mean square

NPs

Nanoparticles

pH_pzc

Point of zero charge

RB5

Reactive Black 5

SS

Sum of squares

TiO₂-NPs

Titanium dioxide nanoparticles

ZnO:Fe₂O₃

Iron oxide doped zinc oxide

ZnO:Pb

Lead doped zinc oxide nanoparticles

ZnO:V₃%

Vanadium doped zinc oxide

Rapid industrialization has a detrimental effect on the environment with water pollution becoming a major societal problem due to its effects on the ecosystem (Santhosh et al. 2016). The textile industry is a major polluter because of the large volumes of heavily contaminated wastewater that it rejects to the environment (Holkar et al. 2016). The composition of the textile effluent is a complex mixture of hazardous chemical compounds including unfixed dyes, acids, alkalis, hydrogen peroxide, starch, surfactant, dispersing agents, and metals. Despite being toxic to the aquatic and non-aquatic organisms and that they can cause cancer in humans, azo dyes are the most used dyes (around 70% of all dyes) (Salehi et al. 2016). Azo dyes (-n=N-) have complex aromatic molecular structures that resist biodegradation making them persist for long times in the environment (Heibati et al. 2014). The direct discharge of textile effluent into the aquatic environment, thus, poses environmental and health problems (Gita et al. 2017). Several techniques have been studied for the removal of azo dyes including coagulation–flocculation (Verma et al. 2012; Dassi et al. 2017), advanced oxidation (Wong et al. 2015) and membrane filtration (Riera-Torres et al. 2010). Despite showing promising results, these methods have certain defects such as sludge production, high treatment costs and membrane fouling (Crini & Lichtfouse 2019). Adsorption has also been widely studied for dye removal and it appears to be an economical, ecological and sustainable technique for dye removal (Kumar et al. 2013). Many materials show a high adsorption efficiency of textile dyes, such as activated carbon (Singh et al. 2003), chitosan (Monvisade & Siriphannon 2009) and geopolymer (Maleki et al. 2018). Recently, nanotechnology has been successfully integrated in adsorption processes (nanoadorbent), membrane processes (membrane composite) and photocatalysis (Qu et al. 2013). The sorption using metal oxide nanoparticles (ZnO, Fe₂O₃, TiO₂) has been widely studied, as one of the most advanced materials used for the textile wastewater treatment (Gehrke et al. 2015). Zinc oxide (ZnO) nanoparticles (NPs) have been used as advanced adsorption materials in the wastewater treatment (Zhang et al. 2016) due to their specific properties such as the presence of active sites at the surface, chemical stability with high efficiency, low cost and low toxicity (Khezami et al. 2017a). It was reported that the incorporation of metals such as Fe, Pb, Ni, Cu, Ag and Al in zinc oxide NPs enhances the adsorption capacity and the photocatalytic activity (Bordbar et al. 2018; Ebrahimi et al. 2019). Farrokhi et al. (2014) showed the ZnO–Fe₃O₄ NPs exhibit an efficient removal of azo dye Reactive Black 5 (RB5) with maximum sorption capacity of 22.1 mg/L. Lei et al. (2017) reported that ZnO-Al₂O₃ was an efficient adsorbent for the removal of Congo red dye. In this study, novel lead doped zinc oxide NPs were successfully synthesized by the sol-gel technique to remove an anionic textile dye (RB5) from aqueous solution using a batch sorption system. A central composite (CCD) design with response surface methodology was used to understand the interaction between the most influencing parameters (pH, the dose of NPs and contact time) in the sorption and to determine the optimal conditions of the process. Kinetic and isotherm models were used to study the sorption mechanism of RB5 on ZnO:Pb NPs.

RB5 azo dye in powder form at purity of ≥50% was purchased from Sigma Aldrich, USA (molecular formula C₂₆H₂₁N₅Na₄O₁₉S₆, molecular weight = 991.82 g/mol). Zinc acetate dehydrate (Zn(CH₃COO)₂.2H₂O) 99.99% was used as a zinc precursor, lead(II) nitrate (N₂O₆Pb) 99.99% as doping element source, methanol and ethyl alcohol (EtOH) as solvent and co-solvent respectively, and sodium hydroxide (NaOH) and hydrochloric acid (HCl) as pH adjustment solutions. All products were purchased from Sigma Aldrich, Germany and used as-received without further purification.

Lead-doped ZnO nanocrystals (ZnO:Pb) were prepared using sol–gel method following a method presented in literature (El Mir et al. 2008). Briefly, 16 g of precursor zinc acetate dehydrate was added to 112 mL of methanol and stirred at room temperature for 15 min using a magnetic stirrer. Adequate quantities of lead precursor (0.496 and 0.971 g), corresponding to 2 and 4% atomic percent calculated relative to the sum of atomic masses of Zn and Pb ([Pb]/([Zn] + [Pb]), were added to the solution, which was stirred for another 15 min. The solution was then placed in an autoclave (250 °C) and dried under supercritical conditions of EtOH.

The ZnO:Pb NPs were characterized with Fourier transform infrared spectroscopy (FTIR) (PerkinElmer, USA) in order to identify the main functional groups present on their surface. The spectral resolution of the FTIR is 1 cm⁻¹ measured between 4,000 and 400 cm⁻¹. X-ray diffraction (XRD) pattern was recorded with a Bruker D8 Advance X-ray diffractometer using Cu Kα (λ= 1.541 Å) radiation in the diffraction angle range of 10–100°. The surface morphology of ZnO:Pb nanopowders was characterized by field emission scanning electron microscopy (FE-SEM, Hitachi, S-4800, Japan). Energy dispersive X-ray (EDX, Zeiss EVO HD15) technique was also used to characterize the elemental composition of the adsorbent. The BET (Brunauer–Emmett–Teller) specific surface area, pore volume (micropore volume and total pore volume) and pore size distribution of the nanopowders were determined by nitrogen N₂ adsorption/desorption isotherm at 77 K (Micromerities, ASAP-2020).

The pH of point of zero charge (pH_pzc) of ZnO:Pb (2 and 4%) was measured by the pH drift method (Mustafa et al. 2013). A 5 mM NaCl solution was prepared and its pH was adjusted by adding HCl or NaOH to be in the range 2–12. An amount of 20 mg of ZnO:Pb 2% or 4% was added to 20 mL of the NaCl solution at an initial pH_i in an Erlenmeyer flask stirred at 250 rpm at room temperature for 24 h (Kulkarni et al. 2016). The final pH (pH_f) was measured and a plot of pH_f vs pH_i was made. The point at which the curve crosses the line pH_f = pH_i is the pH_pzc.

The analysis of variance (ANOVA) and the statistical analysis of the effect of individual linear, quadratic and interaction terms on the response were performed using STATISTICA 10 software. The significance of all terms in the polynomial model was assessed statistically by computing the F-value and the P-value at a probability level of 0.05.

FTIR spectra of Pb doped ZnO NPs are shown in Figure 1(a). It depicts a sequence of absorption bands in the range of 500 and 4,000 cm⁻¹. The absorption peaks at 852, 684 and 592 cm⁻¹ revealed the presence of Pb-O stretching vibration (Borhade et al. 2013; Sathya & Pushpanathan 2018). The two peaks of absorption, 1,400 and 1,581 cm⁻¹, could be attributed to precursor functional groups and ZnO surface-related reaction by-products (Kataria & Garg 2017). The stretching vibration peak of the C-H bond was observed at 3,093 cm⁻¹ (Mousavi et al. 2015).

Figure 1(b) depicts the XRD diffractogram of ZnO:Pb nanopowder. The diffraction peaks at 2θ = 31.75°, 34.23°, 36.05°, 47.39°, 56.35°, 62.65°, 67.80°, 68.2° and 69.4° which can be attributed to the (100), (002), (101), (102), (110), (103), (200), (112) and (201) planes correspond to ZnO hexagonal wurtzite structure, in agreement with the JCPDS card No. 36-1451 (Kataria & Garg 2017). No characteristic peaks related to Pb metal can be detected, which indicates the successful incorporation of Pb in the ZnO.

The SEM images illustrated in Figure 2(a) and 2(c) exhibited spherical particles with average particle size less than 100 nm. The EDX result (Figure 2(b) and 2(d)) of NPs indicated the existence of 65.7% of zinc and 2.2% of lead for ZnO:Pb 2% and 63.6% of zinc and 4.3% of lead for ZnO:Pb 4%, which confirms the theoretical calculation of NPs composition. The surface area and the pore volume of the ZnO:Pb 2 and 4% were investigated using nitrogen gas adsorption–desorption method based on BET theory. The doping of ZnO by Pb has an influence on the specific surface area of the nanopowder. The total pore volume was found to be 0.177 cm³/g and 0.197 cm³/g respectively for ZnO:Pb 2% and ZnO:Pb 4%. The average of pore width was 145 and 123 Å for ZnO:Pb 2% and ZnO:Pb 4% respectively.

The incorporation of Pb in zinc oxide NPs increases the specific surface area to 18.8 m²/g for ZnO:Pb 2% and 20.8 m²/g for ZnO:Pb 4%, which are higher than other NPs like ZnO, ZnO:Fe₂O₃ and ZnO:V₃% whose BET surface area ranged from 4 to 9 m²/g (Table 2).

Figure 3 presents the pH_pzc of ZnO:Pb 2% and ZnO:Pb 4% NPs. The figure shows that both ZnO:Pb 2 and 4% have the same pH_pzc value of 7.5. At pH less than pH_pzc, the surface of the ZnO:Pb NPs is positively charged and, due to electrostatic interactions, the adsorption of the anionic dye is thus preferred (Farrokhi et al. 2014). However, when the pH is greater than pH_pzc, the surface of the NPs is negatively charged and, due to electrostatic repulsion forces, the adsorption efficiency of RB5 decreases (Samadi et al. 2015).

The central composite design was used to estimate the influence of three factors: pH, dose of NPs and contact time in the range of pH 3.6–10.3, NPs dose (g/L) 1.32–2.67 and contact time (min) 46.4–113.6. The experimental and predicted removals (%) of RB5 for the 20 experiments are presented in Table 3.

The ANOVA presented in Table 4, gives useful information about the significance of the variables and their interaction based on F and P values at 95% confidence level. The significance of the model terms was investigated by P-value. A P-value less than 0.05 in the ANOVA table indicates that the interactions between the significant factors have a good influence on the response. Lack of fit (LOF) compares the experimental error of the replicate measurements. The non-significance (F-value >0.05) of lack of fit indicate the model was appropriate to predict the adsorption percentage within the investigated range of variables (Dil et al. 2019). The F-value of LOF for the present study (Table 4) indicates a good concordance between the experimental and the predicted RB5 removal efficiencies. The high values of the correlation coefficients (R² = 0.97 for ZnO:Pb 2% and R² = 0.95 for ZnO:Pb 4%) show a high significance of the model.

Equations (4) and (5) emphasize the interaction effect of the factors on the removal of RB5 by ZnO:Pb 2% and ZnO:Pb 4%. A positive impact of a factor means an improvement in the response when the factor level increases, while a negative impact means that the response is not improved with increasing the factor levels.

The three-dimensional response surface plots (Figure 5) show the interaction effect of the three variables, pH, dose of NPs and contact time, on the efficiency removal (%) of RB5. The pH of the solution is an important parameter during the dye sorption process. The effect of pH on the removal of RB5 was examined between pH 3.6 and 10.3 (Figure 5(a), 5(d), 5(c) and 5(f)). RB5 is an anionic dye, which exists in the aqueous solution in the form of negatively charged ions. As a charged species, the adsorption degree on the surface of the adsorbent is mainly influenced by the surface charge of the adsorbent, which in turn is influenced by the pH of the solution (Bazrafshan et al. 2015). According to the results of pH_pzc of ZnO:Pb 2% and ZnO:Pb 4% NPs, at pH below the pH_pzc (7.5) of the adsorbent the surface of the sorbent is positively charged and able to attract the negatively charged dye molecule from the solution. Electrostatic attraction exists between the positively charged surface of the adsorbent and the negative charge of the RB5 textile dye (due to the presence of - groups). The maximum removal efficiency of RB5 was observed at neutral pH for both of the NPs. The highest RB5 removal efficiencies (more than 80%) were observed when increasing the contact times and the dose of NPs simultaneously (Figure 5(b) and 5(e)). As expected, the dye removal percentage increased when the dose of adsorbent increased. This can be explained by the increase of active sites on the surface of the nanoadorbent. Similar results were observed in the literature for the removal of RB5 by ZnO-Fe₂O₃ NPs (Farrokhi et al. 2014) and RB5 by iron oxide NPs (Chang & Shih 2018). The optimal conditions of color removal using ZnO:Pb 2% and ZnO:Pb 4% NPs were pH 7, adsorbent dose 2 g/L and contact time of 79 min with maximum color removal of 79.4 and 98.1% respectively.

The maximum color removal efficiency for RB5 dye by ZnO:Pb 2% and ZnO:Pb 4% corresponding to the optimal experimental conditions of the independent variables were determined by using the desirability function on the STATISTICA 10.0.0 software (Table 5). The optimum conditions were found by the model to be pH 7, amount of NPs 2 g/L and time 79 min, giving an experimental maximum RB5 removal of 79.4 and 98.1% for ZnO:Pb 2% and ZnO:Pb 4% respectively. In order to confirm the model adequacy, confirmatory experiments were conducted with these optimal parameters. Results gave a percentage RB5 removal of 80.6 and 97.1% for ZnO:Pb 2% and ZnO:Pb 4%. The percentage error difference between the experimental and predicted value was low (<1.5%) which validated the findings of response surface optimization and indicates that the proposed model is adequate for obtaining the optimum value in the range of the studied parameters.

The effect of contact time on color removal was studied at pH = 7 using 30 mg/L of RB5, shown in Figure (6). The result shows that at 15 min of contact time the color removal percentage was 65.2 and 65.9% for ZnO:Pb 2 and 4% respectively. The sorption process was fast and after 90 min of contact time ZnO:Pb 2 and 4% achieved 79 and 98% maximum percentage removal of the dye respectively.

The experimental data were fitted by the pseudo-first-order, pseudo-second-order and intra-particle diffusion models plotted in Figure (7) and the values of the models' rate constants are given in Table 6. The pseudo-second-order model showed excellent fit of the adsorption data for both adsorbents with R² values of 0.998 and 0.999 for ZnO:Pb 2% and ZnO:Pb 4% respectively. The pseudo-second-order model makes it possible to characterize the adsorption of RB5 onto ZnO:Pb NPs.

Langmuir, Freundlich and Dubinin–Radushkevich isotherm models were examined to describe adsorption equilibrium and to study the interaction between adsorbent and adsorbate species and adsorption capacity of adsorbent.

R_L value indicates the adsorption nature as irreversible if R_L = 0, favorable if 0 < R_L < 1, linear if R_L = 1 and unfavorable if R_L > 1.

The data of the Dubinin–Radushkevich isotherm were obtained by the plot of log q_e versus ε² as in Figure 8(c). Dubinin–Radushkevich isotherm parameters K and q_s were determined by linear regression, and R² values are listed in Table 7.

Table 7 summarizes the Langmuir, Freundlich and Dubinin–Radushkevich isotherm constant parameters and the correlation coefficients calculated for the sorption of RB5 onto ZnO:Pb 2 and 4%. The determination coefficients (R²) for Langmuir, Freundlich and Dubinin–Radushkevich models were respectively 0.980, 0.930 and 0.968 for ZnO:Pb 2% and 0.994, 0.982 and 0.818 for ZnO:Pb 4%. A high value of R² was obtained by the Langmuir model. Also the R_L constant of the Langmuir isotherm for ZnO:Pb 2% and ZnO:Pb 4%, for different concentrations of RB5 (20–80 mg/L), was between 0 and 1, which indicates the favorable adsorption of RB5 onto ZnO:Pb NPs under the conditions used in this study. The result shows that the Langmuir isotherm model provided a good fit to the sorption equilibrium data. A similar isotherm model was observed in the adsorption of RB5 on modified dolomite (Ziane et al. 2018). The maximum monolayer adsorption capacities of ZnO:Pb 2% and ZnO:Pb 4% are respectively 27.85 mg/g and 29.1 mg/g, which are higher than results found using only ZnO for adsorption of RB5 (3.64 mg/g) (Saeed 2013).

A comparison of maximum adsorption capacity on ZnO:Pb (2 and 4%) NPs with other adsorbents for RB5 dye removal is listed in Table 8. The ZnO:Pb adsorbent has relatively higher maximum adsorption capacity toward RB5 dye than most of the listed adsorbents.

The isotherm models describe how the adsorbent interacts with the adsorbate. Figure 9 presents the experimental isotherms of sorption of RB5 onto ZnO:Pb 2% and ZnO:Pb 4%. The sorption of RB5 onto the NPs ZnO:Pb 2% and ZnO:Pb 4% were Class Langmuir (L); this class proposes, at low concentration in solution, a downward concavity which reflects a decrease in free sites as the adsorption progresses. This phenomenon occurs when the attractive forces between the adsorbed molecules are low. It is often observed when the molecules are adsorbed flat, which minimizes their lateral attraction. It can also appear when the molecules are adsorbed vertically and when the adsorption competition between the solvent and the solute is weak (Foo & Hameed 2010).

Pb doped ZnO showed a good efficiency in the sorption of the synthetic textile dye RB5. The influence of experimental parameters on RB5 removal percentage was investigated by an experimental design methodology allowing the optimization of pH, NPs dose and contact time. The maximum color removal was 79.4% for ZnO:Pb 2% and 98.1% for ZnO:Pb 4% obtained at pH 7 for dose of NPs of 2 g/L and 79 min as contact time. The equilibrium and kinetic studies were investigated and the results show that the pseudo-second-order model presented excellent fit of the adsorption data. The Langmuir monolayer model was found to best fit the isotherm data. The maximum adsorption capacities of ZnO:Pb 2% and ZnO:Pb 4%, according to the Langmuir isotherm model, were 27.85 and 29.1 mg/g respectively. The kinetic and isotherm sorption parameters were compared to other published studies and the magnitude of their values highlights the benefits of ZnO:Pb NPs for textile effluent treatment.

We thank Water Research and Technologies Center, Borj Cedria-Tunisia, European Membrane Institute Montpellier-France and Faculty of Sciences of Gabes-Tunisia for financial support of this work.

On behalf of all authors, the corresponding author states that there is no conflict of interest.

All relevant data are included in the paper or its Supplementary Information.