3.1 Characterization of the nZVI-Pd/AC
The XRD patterns of the nZVI-AC and nZVI-Pd/AC (Fig. 2) show five characteristic peaks at 44.34°, 51.67°, 76.26°, 92.79° and 98.63°, corresponding to (111), (200), (220), (311) and (222) standard lattice plane of standard Fe0 samples (JCPDS no. 52–0513), but the peak positions do not precisely match the five characteristic peaks. This may be due to the inclusion of carbon atoms in the grains. According to the Bragg equation, the addition of carbon atoms causes the unit cell to shrink, and these five characteristic peaks move in a larger angle. And no peaks of Pd or other other (hydrogen) oxides are observed in Fig. 2 (a), which may because the energy spectrum of Fe is strong enough to mask the weak signal of Pd. And the surface of the material has very little (hydrogen) oxide deposition, and its content is below the detection limit of XRD (Li et al. 2017a, Wu et al. 2018). As shown in Fig. 2 (b), comparing the XRD patterns of iron-supported and iron-supported palladium, it can be concluded that the peak value of the iron-supported palladium electrode sheet is smaller than that of the iron-supported electrode sheet, which may be due to the reduction of Pd2+ by Fe0 for reducing the crystallinity of Fe0. Besides, there is not any visible Pd peak on the entire iron-bearing palladium electrode pad. So Pd has a slight lead over the whole electrode sheet with a relatively uniform distribution. On the basis of XRD results, the mean crystal size of nZVI-Pd/AC was calculated by the Scherrer diffraction formula to be 49.82 nm.
The micrograph of activated carbon is shown in Fig. 3 (a). The structural clusters of activated carbon can be observed. The photomicrograph of activated carbon electrode with nZVI is presented in Fig. 3 (b). Compared with the activated carbon, the iron-loaded activated carbon is noticeably smaller, and the particle size distribution is uniform in the range of 40 ~ 80nm. The nanometer nZVI-Pd/AC electrode as Fig. 3 (c), the particle distribution has a uniform, clean-cut, particle size of about 50 nm, and with local agglomeration. According to the multi-point BET method, the BET surface area of the prepared nZVI/Pd nanostructures is 444.43m2 g − 1, which is much larger than the specific surface area of 76.9 m2 g− 1 observed by Frost (Frost et al. 2010) et al. This is beneficial to the adsorption and removal of pollutants by nZVI/Pd-AC.
The chemical composition of the prepared activated carbon-supported nZVI/Pd bimetal was determined by EDS analysis (Fig. 4). The molar ratio of Fe to Pd was just 14.9:1, compared with the 1000:1 molar ratio of Fe to Pd in the deposition solution. The palladium content increased markedly, and the reason may be the existence of a reaction: Fe0 + Pd2+=Pd + Fe2+, which increased the palladium content, and the iron content was reduced compared with the electrode sheet without Pd loading. It is consistent with the XRD results. The 14.9:1 molar ratio of Fe to Pd also confirms the reason for the lack of characteristic peaks of Pd in XRD.
XPS was performed to analyze the chemical composition and oxidation state of representative nZVI-Pd/AC surface species. From the nZVI-Pd/AC wide XPS scan spectrum (Fig. 5 (a)), the Fe 2p, Pd 3d, O 1s and C 1s peaks corresponding to the binding energies of Fe, Pd, O and C were identified respectively. As shown in the Fig. 5 (b), the Fe 2p spectrum deconvolution of nZVI-Pd/AC corresponds to Fe0 (706.7 eV (Li &Zhang 2006)), Fe2O3 (710.9 eV (Li et al. 2016), 724.5 eV) and FeOOH (713.8 eV). The weak signal displayed at 706.7 eV corresponds to the binding energy of 2P3/2 of iron, reveals the presence of Fe0 in the nanoparticles. The Pd 3d spectrum of nZVI-Pd/AC deconvolution has two double peaks. For nZVI-Pd/AC (Fig. 5 (c)), the doublet near 336 eV and 341.3 eV (Li et al. 2018) is attributed to the metal Pd0, while the other pair near 336.8 eV and 342.4 eV (Shi et al. 2016) is related to the oxidation state Pd (II). Since PdO, Fe2O3 and FeOOH have similar characteristic peak positions in this region, there is further investigation of the oxygen state of the scanning O 1s (Fig. 5 (d)). Deconvolution of the O 1s spectrum can be observed four peaks at 529.7 eV, 530.1 eV, 531.1 eV and 531.78 eV, which is commensurate with PdO, Fe2O3, FeOOH and H2O respectively. According to the 2.31:1 ratio of the peak areas of Fe2O3 to FeOOH, it can be obtained that the content of Fe2O3 is higher than FeOOH, which is consistent with the peak area ratio 2.66:1 collected from the deconvolution of the Fe 2p spectrum. The sample showed the presence of simple substance and oxide while iron and palladium.
3.2 Effect of initial pH of MB degradation
The pH of the solution is important for the reduction of MB by nZVI/Pd-AC. It affects the chemical properties of the solution and the complex surface charge. Previous studies (Xi et al. 2011) reported that the reduction and removal of MB is strongly affected by pH. Figure 6 depicts the effect of different initial pH on the removal of 30 mg L− 1MB dye by nZVI/Pd-AC. The results show that the removal effect of MB by nZVI/Pd-AC is better under weakly acidic conditions, and the maximum decomposition efficiency is observed to be 95.76% at pH = 5.0. The removal effect is not good under extremely acidic or alkaline conditions. at pH < 4 or pH > 10, MB removal efficiency is significantly reduced. Under strong acid conditions, excessive H+ on the surface of the nZVI/Pd-AC electrode sheet will form an electrostatic repulsion with MB, a cationic dye (Hamdy et al. 2018), which prevents the MB from contacting the Fe0 active site. Under weakly acidic conditions, a small amount of H+ ions will enhance the corrosion of Fe0 nanoparticles and promote the generation of hydrogen atoms (Yang &Lee 2005), thereby destroying the chromophore of MB molecules. Under alkaline conditions, with the increase of OH− concentration, OH− and Fe0 form FeOOH, which reduces the active surface sites of the material (Shu et al. 2007), and ultimately leads to a decrease in the decomposition efficiency of MB.
3.3 Effect of the initial concentration of MB
The effect of the initial concentration of MB on the decolorization by nZVI/Pd-AC was studied, as shown in Fig. 7 (a). The initial pH was 5.0 ± 0.5, and the final pH after nZVI/Pd-AC treatment was 6.8 ± 0.5. When the concentration of MB increased from 20 mg L− 1 to 60 mg L− 1, the removal rate increased first and then decreased, and the reaction rate was also highest at the dye concentration of 30 mg L− 1. The possible explanation is that a low initial concentration the contact probability of MB molecules with nZVI-Pd/AC is small, causing a slow reaction rate and removal rate, when the concentration is excessive, the surface Fe0 is covered by the oxide layer, so the MB molecules cannot contact Fe0 (Fan et al. 2009, Wang et al. 2013). In addition, the high dye concentration will cause competitive adsorption and reduction between dye molecules and nZVI/Pd, and the removal rate will also be reduced.
According to literature reports, the chemical kinetics of vat dyes in solution systems can be described by first-order models or pseudo first-order kinetic model (Lin et al. 2015, Nairat et al. 2015), and is usually modeled by a simple Langmuir-Hinshelwood type rate equation. The model can be described below:
The k1 (min− 1) is the rate constant of the quasi-first order reaction, and C (mg L− 1) is the concentration of MB in the solution. The reaction in the contaminant removal system is a surface-mediated process. If the active surface of Fe0 is considered unchanged, the reaction can be regarded as quasi-first order kinetics (Bhaumik et al. 2017, Xiong et al. 2007). Table 1 lists the fitted model parameters. The higher R2 value indicates that the removal of MB is more in line with the pseudo-first-order kinetic modulus. In addition, the degradation rate constant increased with the MB concentration and then decreased, indicating that the reaction rate was not only related to the concentration of pollutants, but also to the active surface site of Fe0 (Zeng et al. 2017).
Table 1
Pseudo-first-order rate constants k1 of reductive degradation at various MB concentration
Initial MB concentration (mg L− 1)
|
K1 (min− 1)
|
R2
|
20
|
0.0163
|
0.9907
|
30
|
0.0179
|
0.9706
|
40
|
0.0119
|
0.9734
|
50
|
0.0107
|
0.9771
|
60
|
0.0093
|
0.9779
|
3.4 Effect of temperature on MB removal
Figure 8 shows the effect of temperature on MB removal. Reacting for 180 min, the MB removal rate increases with increasing temperature. When the solution temperature is 50℃, the MB removal rate reaches a maximum of 98.48%. When the temperature rises from 30℃ to 50℃, the removal efficiency gradually increases. When the temperature changes from 25℃ to 50℃, the removal efficiency in the first 90 min decreases slightly, but it slightly rises finally. The possible reason is the higher temperatures provided more energy for the reaction and accelerated molecule mobility, Which enhanced diffusion and transfer of MB into reactive sites of nZVI/Pd-AC(Zhu et al. 2018). This made MB more accessible to be adsorbed and reduced by nZVI/Pd-AC. Thus, nZVI/Pd-AC achieved higher MB removal efficiency at higher temperatures.
3.5 The effect of NaCl on the removal of MB
Figure 9 shows the effect of NaCl on MB removal. It can be seen that NaCl has a negative effect on the removal of MB. With the increase of NaCl concentration, the removal rate of MB decreases first and then increases, but the increase was not obvious. The removal rate of 0.04 mol L− 1 NaCl at 180 min was lower than that of 0 mol L− 1 NaCl. The content of NaCl can hinder the reduction of MB by nZVI-Pd/AC. This may be due to the addition of Cl−, which inhibits the ionization of MB molecules (Burkinshaw &Salihu 2019b, a), reduces the adsorption of MB on the surface of nZVI/Pd-AC, and thus reduces the reduction and degradation of MB.
3.6 MB removal efficiency of the recycled nZVI-Pd/AC
The reusability/recycling efficiency of catalytic materials is an essential factor in evaluating their cost effectiveness. Figure 10 shows the repeated service life of nZVI-Pd/AC by using 30 mg L− 1 MB solution with a pH of 5.0 for 3 h of reaction. As the number of reactions increased, its MB removal effect decreased linearly. The MB removal rate was 95.7% in the first batch. In the second and third batches, it decreased to 81.7% and 66.2%, respectively. When the number of consecutive reactions is more than 3, the removal rate was less than 60%.
Due to the poor reusability of the nZVI-Pd/AC material, regeneration experiments were carried out. It’s observed that the removal efficiency of the redeposited nZVI-Pd/AC remained at 88% after the fourth reaction. After the fifth regeneration, it decreased by 4.51%, indicating that the prepared nZVI-Pd/AC has a good reproducibility. In the second and third reactions, the MB removal rate increased slightly because new nZVI/Pd nanoparticles were deposited during the redeposition process. The decrease in the MB removal rate after the fifth reaction may be related to the destruction of the composite’s pore shape. As a result, the surface area of nZVI-Pd/AC is reduced, lessening the removal rate of MB.
3.7 GC–MS analysis of the MB reductive degradation products
To study the possible degradation path of MB removal by nZVI/Pd-AC, the MB solution after 3 hours of the reaction was used for GC-MS analysis, and the intermediate products produced were characterized in detail. Figure 11 (a) shows the intermediate products of MB decomposition at different m/z. The fragment ions in the mass spectrum are mostly possible molecular ion peaks. The possible reasons are analyzed: (1) The difference of each intermediate product in the concentration and physical and chemical properties affects its response value in mass spectrometry. (2) In mass spectrometry detection, ion collision may also occur, forming a new ion peak. Combined with the analysis of existing literature (Huang et al. 2010, Zhou et al. 2020), after treatment there are mainly 2-amino-5-dimethylamino-benzenesulfonic acid (m/z = 213) and N,N-dimethylaniline (m/z = 121).
The degradation path of MB dye molecules is speculated from the intermediate products, as shown in Fig. 11 (b). Under the reduction of Fe0 and the catalytic effect of Pd, MB first breaks the S-C bond and C-N bond to produce intermediate products N,N-dimethylaniline and 2-amino-5-dimethylamino-benzenesulfonic acid. The cleavage of dimethylamino, sulfonic acid, and amino groups eventually degrades into small molecules such as carbon dioxide and water.