Highly efficient adsorption and removal of Chrysoidine Y from aqueous solution by magnetic graphene oxide nanocomposite

3.1.1 Morphology of GO and GO–Fe₃O₄

To investigate the morphology, TEM images of GO and GO–Fe₃O₄ were taken. The typical morphology of graphene oxide is shown in Fig. 2. As can be seen, the sheet-like, corrugated morphology exists, indicating the characteristics of single-layer GO sheets. There are small magnetite nanoparticles with an average size of 10 nm, uniformly decorate the surface of GO sheets. These results prove effectively that Fe₃O₄ nanoparticles are successfully modified on the surface of GO sheets.

Close

Figure 2

TEM images of GO–Fe₃O₄ (the resolution (a, c) ×200 nm, (b) ×50 nm) and GO (c).

Export to PPT

3.1.2 Crystal structure of GO and GO–Fe₃O₄

XRD measurement was employed to investigate the phase and crystalline structure of the synthesized samples. The XRD patterns of graphite powder (C), GO and GO–Fe₃O₄ are shown in Fig. 3. It is observed that the XRD pattern of graphite powder shows the peaks at 2θ values of 26° (0 0 2) and 54° (0 0 4), while that of GO has peaks at 2θ values of 11° (0 0 1) and 42° (1 0 0). The XRD pattern of GO–Fe₃O₄ displays six characteristic peaks of Fe₃O₄ at about 2θ = 30.1°, 35.5°, 43.1°, 53.4°, 57.0° and 62.6°, corresponding to magnetite’s indices (2 2 0), (3 1 1), (4 0 0), (4 2 2), (5 1 1) and (4 4 0) respectively. These results suggest that graphite powder is successfully oxidized to GO, and Fe₃O₄ nanoparticles are successfully attached to the surface of GO sheets. Compared with GO, the peak intensity of GO–Fe₃O₄ at 2θ = 11° is obviously reduced, and the peak at 2θ = 42° totally disappears. It indicates that GO–Fe₃O₄ has been exfoliated largely, and more disordered stacking and less agglomerated graphene sheets consist in the nanocomposite (Wu et al., 2010).

Close

Figure 3

X-ray diffraction patterns of graphite powder (C), graphene oxide (GO) and magnetic graphene oxide nanocomposite (GO–Fe₃O₄).

Export to PPT

3.1.3 Surface functional group of GO and GO–Fe₃O₄

To confirm the attachment of Fe₃O₄ nanoparticles on the GO nanosheet surface, FTIR spectra of GO and GO–Fe₃O₄ were examined and the results are shown in Fig. 4. The absorption peaks at around 1714 cm⁻¹, 1616 cm⁻¹, 1365 cm⁻¹, 1045 cm⁻¹, which are attributed to the stretching vibration of C⚌O, stretching C⚌C, bending vibration of C—OH, and stretching vibration of C—O, respectively, indicate that the surface of GO possessed some oxygen-containing functional groups (Saleh, 2011; Saleh et al., 2011), and the characteristic peak presented at 586 cm⁻¹ confirms the Fe—O vibration. Several characteristic peaks of GO and Fe₃O₄ can be observed, confirming the successful oxidation of graphite and Fe₃O₄ nanoparticles being successfully decorated on GO sheets, which is in accord with the results of XRD and TEM. It has been reported that the intensities of all the FTIR peaks of GO correlated to the oxygen containing groups decreased dramatically when GO was reduced (Dutta et al., 2013). Being similar with this previous results, the peak of GO at 1714 cm⁻¹ decreases when GO is decorated with Fe₃O₄ nanoparticles in this study, which demonstrates that the GO is reduced partially by the adsorption of Fe₃O₄ nanoparticles on it. Furthermore, owing to the sorbent’s abundant oxygen-containing groups such as epoxy, carboxyl and hydroxyl groups, the protonated Chrysoidine Y molecules could be adsorbed onto GO–Fe₃O₄ through the anionic-cationic interaction.

Close

Figure 4

FTIR spectra of GO and GO–Fe₃O₄.

Export to PPT

3.1.4 Thermo-gravimetric analysis results

The content of oxygen-containing groups on the surface of GO and GO–Fe₃O₄, which determined the adsorption capacity of GO–Fe₃O₄, could be determined with the TGA technique via oxidative decomposition which was carried out in nitrogen atmosphere from room temperature to 800 °C at a rate of 10 °C min⁻¹. Fig. 5 shows the representative TGA curves of GO and GO–Fe₃O₄. As shown in Fig. 5, there are three weight loss processes. Firstly, a slight weight loss that occurred below 100 °C can be ascribed to the evaporation of adsorbed water molecules. Secondly, the marked weight loss occurs from 100 to 250 °C, which can be assigned to the removal of labile oxygen-containing functional groups such as epoxy, carboxyl and hydroxyl vapors from the sorbent. Finally, a slight weight loss appeared from 250 to 800 °C is caused by the decomposition of the carbon skeleton (Wang et al., 2010). Moreover, comparing the two curves, the weight losses of GO and GO–Fe₃O₄ at 100–250 °C are about 36.72% and 27.04%, respectively. These results indicate that both GO and GO–Fe₃O₄ have abundant oxygen-containing groups on their surfaces, and Fe₃O₄ nanoparticles have replaced some oxygen-containing groups.

Close

Figure 5

TGA diagram of GO and GO–Fe₃O₄.

Export to PPT

3.1.5 Magnetic results

Fig. 6a shows a hysteresis loop measured at room temperature. As expected, the GO–Fe₃O₄ exhibits no remanence or coercivity at room temperature, revealing its superparamagnetic characteristics (Ge et al., 2008). Because of the microscale of Fe₃O₄ being modified on GO, the saturation magnetization of GO–Fe₃O₄ is 8 emu g⁻¹, but it is sufficient for magnetic separation by an ordinary magnet. Fig. 6b shows photograph of GO–Fe₃O₄ adsorption and the separation under an external magnetic field. The GO–Fe₃O₄ dispersed in aqueous solution can be easily collected by a handle magnet within 1–2 min and the aqueous solution becomes colorless, and GO–Fe₃O₄ can be readily re-dispersed with slight shake after removing the magnetic field. The results reveal the good magnetic separation characteristics of GO–Fe₃O₄, suggesting GO–Fe₃O₄ can be used as a magnetic adsorbent to remove some organic contaminants from aqueous solutions.

Close

Figure 6a

Hysteresis loop of GO–Fe₃O₄ nanocomposite at room temperature.

Export to PPT

Close

Figure 6b

Photograph of CY solution (left), the solution after adsorption by GO–Fe₃O₄ and separation with a magnet (right).

Export to PPT

3.1.6 Surface areas and absorption spectrum of GO and GO–Fe₃O₄

The specific surface areas of samples were measured by N₂ adsorption-desorption isotherms. Both GO and GO–Fe₃O₄ show typical “type IV” N₂ adsorption-desorption isotherms (Fig. 7), and the specific surface areas of GO and GO–Fe₃O₄ are 22.11 and 3.312 m² g⁻¹, respectively.

Close

Figure 7

Nitrogen adsorption–desorption isotherms of GO and GO–Fe₃O₄.

Export to PPT

Fig. 8 shows the absorption spectra of GO and GO–Fe₃O₄. GO displays a characteristic peak at 228 nm corresponding to π → π^∗ transitions of aromatic C—C bonds, while for GO–Fe₃O₄, the absorption peak blueshifts to 224 nm, indicating that Fe₃O₄ particles are covalently attached onto GO surface (Chen et al., 2013).

Close

Figure 8

Absorption spectra of GO and GO–Fe₃O₄.

Export to PPT

3.2.1 Effect of initial pH on adsorption

Since acidity not only affects the solution chemistry of contaminants (i.e. hydrolysis, redox reaction, polymerization and coordination) but also has a strong influence on the ionic state of functional groups on the surface of adsorbent, and the initial pH of the solution is one of the most important parameters in the adsorption of environmental contaminants in aqueous solution. In this work, the adsorption quantity and removal percentage of CY by GO–Fe₃O₄ were detected with pH values ranging from 3 to 11. As presented in Fig. 9, the adsorption quantity of GO–Fe₃O₄ increases from 154.61 to 193.94 mg g⁻¹ and removal percentage increases from 77.30% to 96.97% when the initial pH is varied from 3 to 6, and keeps constant over the pH range of 6–8. With the pH value increased from 8 to 11, the adsorption capacity and removal percentage decrease. Considering the pH values of water in nature, the experiments in this study were performed at pH 7.0.

Close

Figure 9

Effect of solution pH on the adsorption of CY by GO–Fe₃O₄ (initial CY concentration: 200 mg L⁻¹; contact time: 60 min).

Export to PPT

3.2.2 Adsorption kinetics

The adsorption kinetics information has a significant practical value. Rapid reaction time could facilitate smaller reactor volumes and ensure efficiency and economy. The adsorption kinetics was determined to understand the adsorption property of GO–Fe₃O₄. Kinetics experiments were carried out by adding 5.0 mg of new prepared GO–Fe₃O₄ to 5.0 mL aqueous solution containing 200, 300 and 400 mg L⁻¹ CY at room temperature. The effect of adsorption time on the removal of CY is shown in Fig. 10. The result displays that a fast adsorption process occurs during the first few minutes and then reaches equilibrium. It is observed that the adsorption equilibrium time is about 30 min for all the concentrations, which attributes to the adequate free adsorptive sites and a high concentration gradient. This rapid adsorption indicates that the adsorption occurs mainly on the surface of adsorbent.

Close

Figure 10

Effect of adsorption time on the adsorption capacity for CY onto GO–Fe₃O₄ nanocomposite at different initial CY concentration (200 mg L⁻¹, 300 mg L⁻¹ and 400 mg L⁻¹) at room temperature; dosage of GO–Fe₃O₄: 5.0 mg.

Export to PPT

In order to understand the adsorption kinetics, four different kinetic models were used to fit the experimental data. They are the Pseudo-first-order (Ho and McKay, 1998), Pseudo-second-order (Ho and McKay, 1999), Intraparticle diffusion and Elovich models (Chien and Clayton, 1980; Hao et al., 2010). The Pseudo-first-order model rate equation is expressed as follows: $$\ln (q_e-q_t)=\ln q_e-k_{1}t$$ where q_e is the amount of dye adsorbed by adsorbent at equilibrium (mg g⁻¹), q_t is the amount of dye adsorbed at time t (mg g⁻¹), and k₁ is the rate constant (min⁻¹). By plotting ln(q_e − q_t) versus t, the k₁ and the calculated equilibrium sorption capacities q_e,cal could be obtained from the slope and intercept of plots in Fig. 11(a), respectively.

Close

Figure 11

The fitting of different kinetic models for CY adsorption onto GO–Fe₃O₄ at different initial concentration (200 mg L⁻¹, 300 mg L⁻¹ and 400 mg L⁻¹) at room temperature ((a) Pseudo-first-order, (b) Pseudo-second-order, (c) Intraparticle diffusion and (d) Elovich).

Export to PPT

The Pseudo-second-order model is given by the following equation (Chen et al., 2010): $$\frac{t}{q_t}=\frac{1}{k_2{q}_e^2}+\frac{t}{q_e}$$ where k₂ is the Pseudo-second-order model rate constant (g mg⁻¹ min⁻¹) and q_e is the sorption capacity at equilibrium. k₂ and q_e can be determined respectively from the intercept and slope of the plot which is obtained by plotting t/q_t versus t (Fig. 11(b)).

The Intraparticle diffusion rate equation is provided as follows: $$q_t=k_i{t}^{1/2}+C$$ where C is the constant related to the thickness of the boundary layer (mg g⁻¹), and k_i is the intraparticle diffusion rate constant (mg g⁻¹ min^−1/2). The adsorption process is controlled only by intraparticle diffusion when the plot of q_t versus t^1/2 gives a straight line. However, if the data exhibit multi-linear plots, two or more parameters influence the adsorption process. Based on the equation, the values of q_t correlated linearly with values of t^1/2 and the rate constant k_i could be directly evaluated from the slope of the regression line (Fig. 11(c)) before reaching equilibrium. The beginning straight line represents macropore diffusion and the second one depicts micropore diffusion. Significantly, the C values obtained from the intercept in this work are not zero, indicating that intraparticle diffusion may not be the only controlling factor in determining the kinetics of the adsorption process (Hameed, 2008).

Elovich model rate equation is indicated as $$q_t=\frac{1}{\beta }\ln (\alpha \beta )+\frac{1}{\beta }\ln$$

It is used to describe chemical adsorption on highly heterogeneous adsorbents, where α (g g⁻¹ min⁻²) and β (g mg⁻¹ min⁻¹) are the equilibrium rate constants. These constants can be obtained from the slope and intercept of initial part of plot which is shown in Fig. 11(d) before reaching equilibrium.

From the parameters listed in Table 1, it suggests that the adsorption kinetics follows the Pseudo-second-order kinetic model much better than the other models because of its higher correlation coefficients (R²), which also indicates that the adsorption is a fast process. Based on the Pseudo-second-order model (Vadivelan and Kumar, 2005), it can be concluded that the adsorption of CY onto GO–Fe₃O₄ is a chemical adsorption, and more than one step may be involved in the adsorption process.

3.2.3 Adsorption isotherms

To further understand the adsorption mechanism and quantitatively describe adsorption capacities, the adsorption data were analyzed by the Langmuir and Freundlich isotherm models (Allen et al., 2004).

The Langmuir isotherm assumes the monolayer coverage of adsorbate over a homogeneous adsorbent surface. It has found successful application in many monolayer adsorption processes. This isotherm is depicted by the following equation: $$\frac{C_e}{q_e}=\frac{C_e}{q_m}+\frac{1}{K_L{q}_m}$$ where q_e and C_e are the adsorption quantity (mg g⁻¹) and concentration of the adsorbate at equilibrium (mg L⁻¹), while q_m and K_L represent adsorption capacity of adsorbent (mg g⁻¹) and the Langmuir adsorption equilibrium constant (L mg⁻¹). The results are shown in Table 2, and the adsorption capacities (q_m) of GO–Fe₃O₄ at 20 °C and 50 °C are 344.83 and 285.71 mg g⁻¹, respectively. Compared with some previous adsorbents (Table 3), the adsorption capacity of GO–Fe₃O₄ is much higher, indicating that GO–Fe₃O₄ is an ideal adsorbent for the removal of CY from aqueous solution. Moreover, it is noticed that the adsorption capacity decreases when the temperature increases from 20 °C to 50 °C, suggesting that this adsorption is exothermic.

The type of Langmuir isotherm can be used to predict whether the adsorption is favorable or unfavorable in terms of either the equilibrium parameter or a dimensionless constant separation factor R_L, which is defined by the following equation: $$R_L=\frac{1}{1+K_L{C}_o}$$ Here C_o (mg L⁻¹) is the initial concentration of adsorbate. The R_L indicates the type of isotherm to be irreversible (R_L = 0), favorable (0 < R_L < 1), linear (R_L = 1) or unfavorable (R_L > 1). The R_L values in this work are in the range of 0.024–0.31, which indicates a favorable adsorption between CY and GO–Fe₃O₄.

The Freundlich model is applied to monolayer adsorption but not a saturation-type isotherm. It is an empirical equation assuming that the adsorption process takes place on heterogeneous surfaces. The Freundlich adsorption equation is given as follows (Luo and Zhang, 2009): $$\ln q_e=\frac{1}{n}\ln C_e+\ln K_F$$ where K_F (mg g⁻¹) is the Freundlich constant indicating adsorption capacity, and n is an empirical parameter related to the intensity of adsorption. The value of n varies with the heterogeneity of the adsorbent and for favorable adsorption process the value of n should lie in the range of 1–10. It is illustrated that the value of n in the range 2–10 represents good, 1–2 moderately difficult, and less than 1 poor adsorption (Hao et al., 2010). From Table 2, the values of n at 20 and 50 °C are 4.8685 and 4.7939, showing a good adsorption. Moreover, the R² values at 20 °C for the Langmuir and Freundlich models in Table 2 are 0.9971 and 0.9662, which demonstrates that the Langmuir model fits the adsorption data better than the Freundlich one, and states that the adsorption of CY onto GO–Fe₃O₄ may be a monolayer coverage process.

3.2.4 Thermodynamic parameters

Considering the practical application, thermodynamic parameters should be detected in order to determine which process will take place spontaneously. Thermodynamic parameters including Gibbs free energy (△G⁰), enthalpy change (△H⁰) and change in entropy (△S⁰) are obtained according to the following formulas: $$\mathrm{\Delta }{G}^0=-\mathit{RT}\ln K_L$$ $$\mathrm{\Delta }{G}^0=\mathrm{\Delta }{H}^0-T\mathrm{\Delta }{S}^0$$

Here R is the gas constant, T is the absolute temperature, and K_L (q_e/C_e) is Langmuir constant at various temperatures when the concentration terms are expressed in L/mol. The values of △H⁰ and △S⁰ were calculated from the intercept and slop of the plot of △G⁰ vs. T. All these thermodynamic parameters are listed in Table 4.

The negative values of △G⁰ at different temperatures indicate spontaneity and feasibility of the adsorption process. Moreover, the values of △G⁰ in Table 4 are less than 10 kJ/mol, which suggest that the adsorption process is physisorption (Li et al., 2014). From the minus value of △H⁰, it further reveals that the adsorption reaction is exothermic and the adsorption process is energetically stable, which is also supported by the decrease in the mass of adsorbed CY with the increase in temperature. The negative value of △S⁰ confirms the decreasing randomness at the solid-solution interface during the adsorption of CY onto GO–Fe₃O₄. These results are in accordance with these previous results (Dong et al., 2014; Li et al., 2014; Shi et al., 2014).