3.1 Structure characterization
The powder XRD patterns of the g-C3N4, MoS2 and MoS2/g-C3N4 nanocomposites were shown in Fig. 1. In the typical XRD pattern, all diffraction peaks of g-C3N4 correspond to the standard card (JCPDS 87-1526). At 13.2 °, the peak value corresponds to the (100) crystal planes, caused by the periodic arrangement of stacking units between layers of g-C3N4. The peak value at 27.5 ° corresponds to the (002) crystal planes, which is the characteristic peak of the accumulation of graphite like layered structure of g-C3N4 (Lin et al., 2019). The diffraction peaks of pure MoS2 at 15.1 °, 32.7 °, 35.2 °, 38.8 °, 44.2 ° and 57.3 ° belongs to the (002), (100), (102), (103), (006) and (110) crystal planes of hexagonal MoS2 (JCPDS 37-1492), respectively (Ali et al., 2019). From this typical XRD patterns of MoS2/g-C3N4 nanocomposites, we can see the peak of MoS2 clearly, but the peak value of g-C3N4 is very small, which may be due to the nanocomposites contain the relatively low content of g-C3N4. The other possible reason of weak intensity of (002) peak of g-C3N4 may coincides or interfere with the (100) crystal planes of MoS2 peak. Nevertheless, the diffraction peak intensity of MoS2 decreases with the addition of g-C3N4, especially at 15.1 ° peak. This indicates that the addition of g-C3N4 further limits the accumulation of the molybdenum disulfide layer (Y. X. Chen et al., 2017).
XPS analysis further determined the chemical state and element composition of the MoS2/g-C3N4 nanocomposites. Figure 2a illustrated the XPS survey spectrum confirmed C, N, Mo, and S elements exist in MoS2/g-C3N4 nanocomposites. Figure 2b represents the spectra of C1s, where 284.1 eV corresponds to the C-C in aromatic rings, 285.8 eV corresponds to N-C = N and 288.0 eV corresponds to C-NH2 (Yi et al., 2017). Figure 2c suggests that the peaks of N 1s spectra at 398.6 eV, 399.1 eV and 403.8 eV belong to C = N-C, N-(C)3 and N–H structures, respectively (Cao et al., 2017). In addition, the two peaks in the Mo 3d spectrum at around 228.95 eV and 232.13 eV belong to Mo 3d5/2 and Mo 3d3/2 in Fig. 2d, respectively. There was another peak of 226.01 eV in this area which is part of S 2s. The spectrum of S 2p in Fig. 2e shows two peaks: 161.87 eV and 163.20 eV were due to s 2p3/2 and s 2p1/2, respectively. These studies confirmed the successful preparation of MoS2/g-C3N4 nanocomposites.
The morphology and structures of pure g-C3N4, pure MoS2 and MoS2/g-C3N4 nanocomposites were directly analyzed by SEM and TEM. Figure 3 (a-b) FESEM analysis confirmed that the as prepared g-C3N4 possessed the hollow sphere morphology, and Fig. 3a shows that the precursor MCA of g-C3N4 is spherical-like. Figure 3b shows the g-C3N4 formed after calcination of MCA, which is hollow and spherical. Figure 3c shows pure MoS2, which is flower-like. Figure 3d indicates that the morphology of the MoS2/g-C3N4 nanocomposites basically keeps the appearance of MoS2. Figure 3e shows the HRTEM micrograph of MoS2/g-C3N4 nanocomposites. The figure shows that MoS2 is flower like structure, and g-C3N4 is coated by MoS2. Figure 3f shows the presence of well-defined lattice fringes of 0.62 nm, this can be attributed to the (002) crystal plane of MoS2 (Monga et al., 2020).
Figure S1 showed the N2 adsorption desorption isotherm. Since there was almost no adsorption limitation at the high-pressure stage, the adsorption isotherm of MoS2/g-C3N4 nanocomposites was determined to be type-IV isotherm with H3 type hysteresis loops, were closely related to the capillary condensation phenomenon of slit pores generated by the stacking of lamellar particles (Liu et al., 2020; Zhu et al., 2020). The pore size of MoS2/g-C3N4 nanocomposites was mainly between 2–17 nm, this indicates the composite was mesoporous. It can be seen in Table S1, the BET surface area, average pore size and pore volume of MoS2/g-C3N4 nanocomposites were 70.656 m2/g, 15.4494 nm and 0.2729 cm3/g, respectively. These sufficient parameters indicated that mesoporous MoS2/g-C3N4 nanocomposites have excellent adsorption properties for MB.
3.2 Experiment majorization
3.2.1 Effect of pH value
The pH value has great influence on the ionization degree of MB molecule and the surface charge of adsorbent in the solution (I et al., 2020). For the sake of studying the effect of pH value on MB adsorption, we set the pH value from 1 to 11. In Fig. S2, the MB adsorption efficiency of MoS2/g-C3N4 nanocomposites increased with the pH value in the range of 1 to 11. And the surface charge of MoS2/g-C3N4 nanocomposites was shown in Fig. S3. At relatively low pH, the proton ions in the solution may compete with the active sites of cationic dye MB and MoS2/g-C3N4 nanocomposites, which inhibits the adsorption of MB on synthetic adsorbent. With the increase of pH, the surface charge of MoS2/g-C3N4 nanocomposites became lower, resulting in stronger electrostatic attraction between MB and adsorbents. The results showed that the higher the pH value was, the higher the adsorption efficiency was. Fig. S4 showed that the MoS2/g-C3N4 nanocomposites have higher adsorption efficiency when pH = 7, compared with g-C3N4 and MoS2.
3.2.2 Effect of adsorbent dosage
By adding different amounts of MoS2/g-C3N4 nanocomposites into MB solution, the effects of adsorption dose on adsorption efficiency and adsorption capacity were tested. Fig. S5 illustrated when the adsorbent Dosage increases, the adsorption efficiency of MB first increases, and then almost remains unchanged, the equilibrium adsorption capacity of MB decreased from 251.2mg/g to 99.7mg/g. This was due to the equilibrium state of the adsorption process, the adsorbent can no longer adsorb MB. It is obvious that a high MB adsorption rate can be achieved even with a small amount of adsorbent. Considering the removal efficiency and practicability, we selected 10mg adsorbent for the following experiments.
3.2.3 Effect of MB concentration
The adsorption capacity and adsorption efficiency of the nanocomposites were tested by adding different concentrations of MB solution. Fig. S6 showed that when MB concentration increased from 20mg/L to 100mg/L, the equilibrium adsorption capacity increased from 98.8mg/g to 227.4mg/g. It may be that with the increase of MB concentration, the adsorbate in solution has stronger driving force to overcome the mass transfer resistance between solution and adsorbent (L. Zhao et al., 2020). But with the increase of MB concentration, the adsorption rate of MB reduced from 98.80–45.48%, because at a certain concentration, the active center of the adsorbent reached saturation.
3.2.4 Effect of adsorption temperature
In this work, the temperatures were set at 15 ℃, 25 ℃, 35 ℃ and 45 ℃, respectively. Fig. S7 showed that temperature had obvious influence on MB adsorption. With the increase of temperature from 15 ℃ to 45 ℃, the adsorption efficiency rose from 86.84–98.17%. Because when the temperature increased, the molecular Brownian motion will accelerate, and low temperature will usually reduce the diffusion rate of MB in the nanocomposites (Fu et al., 2017). Considering the adsorption efficiency and practicability, the following experiments were carried out at 25 ℃.
3.2.5 Recyclability of MoS2/g-C3N4 for the MB adsorption
The recyclability and stability of adsorbents can greatly improve efficiency and reduce cost. In this study, MB was desorbed with 0.1mol/L NaOH. Fig. S8 showed the MB adsorption efficiency of MoS2/g-C3N4 nanocomposites for five consecutive cycles. The results illustrated that the adsorption efficiency of MB decreased gradually because the adsorbate could not be completely desorbed and occupied some active adsorption sites. The adsorption efficiency of the MoS2/g-C3N4 nanocomposite for MB was still 73.58% after five cycles, which showed that the adsorbent has good reusability.
3.3 Adsorption kinetics
Figure 4a showed the effect of contact time on adsorption capacity. With the increase of contact time, the adsorption capacity increased rapidly from the beginning to slowly until the adsorption equilibrium reached 60 minutes. In this study, pseudo-first-order, pseudo-second-order kinetics and intra-particle diffusion models were used to analyze the adsorption process (Tehrani & Zare-Dorabei, 2016). The calculation formula of the pseudo-first-order kinetics model was as follows:
The calculation formula of pseudo-second-order model was as follows:
The calculation formula of intra-particle diffusion model was as follows:
Where qe (mg/g) is described above, qt (mg/g) represent the adsorption capacity at any contact time, and t (min) is the adsorption time. k1, k2 and k3 are the pseudo-first-order, pseudo-second-order and intra-particle diffusion model rate constants respectively. C (mg/g) is the intercept obtained by fitting the model of intra-particle diffusion.
The linear fitting results of the two kinetic models were shown in Fig. 4. It illustrated that the pseudo-second-order model (Fig. 4c) is better than the pseudo-first-order model (Fig. 4b) in describing the process of adsorption kinetics, which is proved by the higher R2 (R2 = 0.9999) value. In addition, the calculated results of the two kinetic models were shown in the Table S2. Theoretical equilibrium adsorption capacity qe (cal) of the pseudo-second-order model is more consistent with the experimental data qe (exp), which indicates that the adsorption of MB on MoS2/g-C3N4 nanocomposites conforms to the pseudo-second-order model, and the adsorption process is mainly controlled by chemical adsorption (Fang et al., 2018; Guan et al., 2017).
The model of intra-particle diffusion after fitting is shown in (Fig. 4d). The graph was nonlinear, which proved that intra-particle diffusion was not the only factor limiting particle diffusion. There are three processes for MB adsorption. The first stage was from the solution to the surface of MoS2/g-C3N4 nanocomposites. The second stage was the diffusion of MB in MoS2/g-C3N4 nanocomposites. The third stage was the final balance stage. The above analysis showed that the intra-particle diffusion and surface adsorption occur simultaneously, which affects the adsorption of MB significantly.
3.4 Adsorption isotherm
Figure S9 showed the MB adsorption isotherms of MoS2/g-C3N4 nanocomposite at different temperatures. The experimental data were fitted into Langmuir model for monolayer adsorption (Eq. (6)) and Friedrich model of multilayer adsorption (Eq. (7)) (Gunture, Kaushik, et al., 2020). They can be expressed as:
where Ce (mg/L) was the equilibrium concentration of MB, qm (mg/g) was the theoretical maximum adsorption capacity, KL (L/mg) and KF were Langmuir constant and Friedrich constant, respectively. n was the constant related to the surface inhomogeneity of adsorbent (L. Zhao et al., 2020).
The fitting results of the two isotherm models were shown in Fig. 5. The datas in Table 1 showed that R2 values in the Langmuir model (Fig. 5a) were bigger than those in the Friedrich model (Fig. 5b). Therefore, the adsorption of MB on MoS2/g-C3N4 nanocomposites followed the Langmuir isotherm model. At 45 ℃, qmax is 278.4 mg/g, which indicates that MoS2/g-C3N4 nanocomposites has higher MB adsorption effect. The comparison of qmax of different adsorbents is given in Table S3. Therefore, the adsorption capacity of MoS2/g-C3N4 nanocomposites for MB was similar or higher than that of other commonly used adsorbents.
We use the separation coefficient RL to evaluate the feasibility of adsorption by the following equation:
Where C0 (mg/L) and KL (L/mg) had explained above. As shown in Table S4, we observed that 0 < RL < 1, it indicated that the process of MB adsorption was favorable (Mohammadnejad et al., 2018).
Table 1. Modeling parameters of adsorption isotherm calculated by Langmuir model and Freundlich model.
3.5 Adsorption thermodynamics
At different temperatures, we used various thermodynamic parameters to understand the adsorption process. The thermodynamic parameters can be obtained by the following expression:
Where KL is the equilibrium distribution coefficient, R is the molar gas constant (8.314 J/mol − 1·K − 1), T (K) is the temperature, ΔGθ is the Gibbs free energy change, ΔHθ is the enthalpy change, ΔSθ is entropy change. Figure 6 showed the plot of ln KL against 1/T. From Table 2 we can see that the value of ΔGθ was negative, indicating that the adsorption process was a spontaneous process at different temperatures, and the increase in temperature was conducive to the adsorption of MB. The value of ΔHθ was positive, indicating that the adsorption was an endothermic process. In addition, the value of ΔSθ was positive which indicated that the disorder of the solid-liquid interface increases during the adsorption process.
Table 2. Thermodynamic parameters for adsorption of MB by MoS2/g-C3N4 nanocomposites.
3.6 Ecological assessment of treated wastewater
In order to evaluate the ecological characteristics of the treated MB solution, the untreated MB solution and the treated solution were used for the culture of wheat and chickpea seeds, respectively . The growth of germinated wheat and chickpea seeds was analyzed in the next 15 days. As shown in Fig. 7a, the average length of wheat seedlings grown with MB solution is 14.7 cm, while that is 17.2 cm under treated solution. As shown in Fig. 7b, the average length of chickpea grown with MB solution is 21.4 cm, and that with treated solution is 43.6 cm. The results showed that MB solution had strong inhibition on root and bud germination of both plants. Compared with untreated MB solution, the treated solution shows a great improvement role for growth of root and bud parts of wheat and chickpea.
3.7 Adsorption mechanism
Figure S10 showed the possible adsorption mechanism of MoS2/g-C3N4. From the above that the adsorption of MB by MoS2/g-C3N4 belonged to Langmuir monolayer adsorption model. The adsorption rate decreased with the increase of adsorption capacity. From the molecular structure of MB and MoS2/g-C3N4, there was an electrostatic interaction between MB and MoS2/g-C3N4 due to that MB is a cationic dye, while MoS2/g-C3N4 is negatively charged in the solution (Fig. S3). As shown in Figure S10, MB and MoS2/g-C3N4 both contain aromatic rings and possible stronger π-π interaction. In addition, both MoS2/g-C3N4 and MB contain nitrogen atoms, and hydrogen bonds may be formed between them. The above analysis indicated that the adsorption of MB was mainly driven by electrostatic interactions, together with π-π interaction and hydrogen bond.