Catalytic performances of nZVI@BCs catalysts
Fig. 1 exhibited the adsorption and degradation of BPA in various systems. As displayed in Fig. 1(a), BCs, nZVI@BCs and nZVI@Cs provided about 55.7%, 43.1% and 23.5% of BPA adsorption, respectively. Remarkably, when PS was introduced, the removal of BPA reached 100% in nZVI@BCs/PS and nZVI@Cs/PS systems, whereas only 61.1% of BPA removal was achieved in the BCs/PS system. TOC result further demonstrated that nZVI@BCs possessed outstanding mineralization ability, up to 63.2% after 90 min (Fig. 1(b)). Fig. 1(a) and Fig. 1(b) illustrated that nZVI@BCs possessed the best BPA removal performance and mineralization ability among all of the systems. And the fastest apparent rate constant (kobs) of BPA removal in various systems was shown in Fig. S2. Removal of BPA in our system could undergo two phases: a fast stage (first stage) and a subsequent slow phase (second stage). In the nZVI@BCs/PS system, approximately 24% of BPA is rapidly removal in the first min, and 76% of BPA is removed in the following 39 min.
3D EEM fluorescence spectroscopy was further employed to confirm efficient degradation of BPA. Fig 2(a) showed that a distinct fluorescence peak was identified from the initial samples. The peak centered at Ex/Em wavelengths of 275/305 nm may be ascribed to the BPA, because Lin et al. once reported that the fluorescence peak of phenol was located at Ex/Em wavelengths of 260–280/280–320 nm (Lin et al. 2019). As shown in Fig 2(b), the fluorescence intensity decreased to some extent after 30 min of adsorption. When PS was added, the fluorescence intensities of BPA gradually weakened as reaction proceeded and another fluorescence peak located at Ex/Em wavelengths of 285/395 nm rapidly strengthened and then weakened, which may be the intermediate products of the BPA (Fig 2(c) and Fig 2(d)). Further prolonging reaction time to 180 min, the fluorescence signals of both BPA and its intermediate products gradually weakened, which indicated that BPA was mostly destroyed or mineralized.
Characterization
As illustrated in Fig. 3, the SEM micrographs revealed that nZVI@BCs had a spherical porous structure and nZVI was dispersed on the surface of BCs with a relatively uniform distribution, which could be also confirmed by TEM (Fig. 4). Moreover, EDX elemental mapping images exhibited the uniform distribution of C, O, B and Fe elements within the carbon frameworks (Fig. 3 (c)) and the atoms contents of C, O, B and Fe were 81.83%, 6.85%, 9.69, and 1.62%, respectively (Fig. S4). Fig. S5 (a) showed XRD pattern for nZVI@BCs and the characteristic peaks of the pristine nZVI@BCs powder at 44.7º, 65.3º, and 82.5º matched well with the standard patterns of Fe0 (PDF#06-0696) (Li et al. 2015), which indicated that nZVI was doped into the framework of the BCs. FT-IR spectra of the nZVI@BCs was shown in Fig. S5 (b). The peaks at 3400 cm−1, 1590 cm−1, and 1103 cm−1 of the nZVI@BCs were assigned to the hydroxyl groups (–OH), carbonyl/carboxyl (C=O) and C–O–C stretching vibration modes. The peaks at 800–400 cm−1 were associated with the Fe–O, which indicated that Fe was attached to the surface by bonding with oxygen-containing groups (Li et al. 2020, Liu et al. 2010).
XPS survey spectra of nZVI@BCs before and after reaction were shown in Fig. S6. There were four distinct peaks, which were assigned to C1s, O1s, B 1s and Fe 2p, respectively. It's worth noting that the oxygen content increased from 20.44% to 29.84%, indicating the surface oxidation of nZVI@BCs during catalytic reaction, which may because the surface of nZVI@BCs was surrounded with large amount of radicals during PS activation process, resulting in surface oxidation (Duan et al. 2015, Yang et al. 2019). Fig. 5 (a) and (e) indicated that C 1s atom was deconvoluted into five peaks at 283.11 eV (the sp2 carbon bonds), 284.15 eV (C-H), 285.15 eV (C-OH or C-O-C), 287.90 eV (C=O) and 290.40 eV (O-C=O) , respectively (Zhou et al. 2019). The corresponding O1s high resolution scans were also displayed in Fig. 5 (b) and (f), the nZVI@BCs mainly fit into three individual peaks located at around 530.42 eV, 531.81 eV and 535.63 eV, which were ascribed to C-O, C=O and Fe-O (Zhou et al. 2020). The increased propotion of Fe-O (from 7.5 to 36.75%) before and after reaction was attributed to the oxidization of the surface Fe0. The two predominant peaks at 188.13 eV and 192.30 eV are ascribed to the interfacial suboxide B and B2O3, respectively (Fig. 5 (c) and (g)). For the Fe 2p, as shown in Fig. 5 (d), the peak at 705.93 eV was assigned to Fe0 (Wu et al. 2020), which corroborated the impregnation and reduction of Fe2+ to nZVI on the porous surface of BCs. However, after reaction in the nZVI@BCs/PS system, the binding energy of Fe0 almost disappeared (Fig. 5 (h)), which may also because the surface oxidation during catalytic reaction. The result was consistent with the Fig. 5 (f) results.
Mechanism investigation
BPA adsorption kinetics and isotherms
The adsorptive capacity of nZVI@BCs would notably affect the efficiency of BPA removal. As shown in Fig. 6, the adsorption kinetics experiments and adsorption isotherm experiments were carried out in this study. Through experiments, it was found that the adsorption of BPA by nZVI@BCs basically reached equilibrium within 30 min, therefore, the reaction time for the adsorption experiments were set at 30 min. The fitting of the kinetic model and its equilibrium data were shown in Fig. 6 and Table 1. As described in Fig. 6 and Table 1, the kinetics of BPA adsorption on nZVI@BCs might be simulated well using a pseudo-second-order kinetics model ( R2 > 0.9999, which is closer to 1.0 compared to the first-order kinetic model and the intra particle diffusion model), indicating a chemisorption process. The adsorption isotherm model fitting and its data was shown in Fig. 6 and Table 2. The linear form R2 of Langmuir model is closer to 1.0 than that of Freundlich model, indicating that the adsorption of BPA by nZVI@BCs occurs on a uniform surface with approximately equal energy in a single layer.
Table 1. Kinetic parameters for BPA adsorption onto nZVI@BCs.
Adsorption kinetics
|
Pseudo-first-order
|
Pseudo-second-order
|
Intra-particle diffusion
|
k1 (g/mg min)
|
R2
|
k2 (g/mg min)
|
R2
|
kid (mg/g min0.5)
|
C (mg/g)
|
R2
|
nZVI@BCs
|
0.0074
|
0.3584
|
0.5718
|
0.9999
|
0.0425
|
2.9513
|
0.6684
|
Table 2. Adsorption isotherm parameters for BPA adsorption onto nZVI@BCs.
Adsorption isotherm
|
Langmuir
|
Freundlich
|
qm/(mg/g)
|
KL/(L/mg)
|
R2
|
KF/(mg/g)
|
n
|
R2
|
nZVI@BCs
|
19.685
|
0.927
|
0.9868
|
9.685
|
3.2755
|
0.9744
|
Identifying Main Active Species
EPR spectra in nZVI@BCs/PS system using DMPO for trapping •OH, SO4•− and O2•− were shown in Fig. 7(a). No signals were detected for sole PS in aqueous solution and in methanol solution. When nZVI@BCs was added, characteristic signals of DMPO-SO4 and DMPO-OH appeared in the aqueous solution (Zhou et al. 2019, Zhou et al. 2020) and six characteristic peaks of DMPO-HO2 was also obtained in the methanol solvent (Wu et al. 2020, Zhou et al. 2020), confirming the formation of •OH, SO4•− and O2•− during the reaction. Herein, different radical quenching reactions based on the radical scavengers (Ethanol, TBA and Chloroform) were conducted. Ethanol is a forceful quencher for both •OH and SO4•− (k(Ethanol/•OH) = 1.2 ~ 2.8 × 109 M-1s-1, k(Phenol/SO4•−) = 1.6 ~ 7.7 × 107 M-1s-1), while TBA could only quench •OH in aqueous solutions (k(TBA/•OH) = 3.8 ~ 7.6 × 108 M-1s-1, k(TBA/SO4•−) < 4.0 ~ 9.1 × 105 M-1s-1) (Liang &Su 2009). And trichloromethane is a powerful inhibitor for O2•− (k(CHCl3/O2•−) < 106 M-1s-1) (Yang et al. 2019). As Fig. 7(b) showed, BPA removal obviously reduced when ethanol, TBA and chloroform were added. The BPA removal rates decreased 28.92%, 21.51% and 5.54% in the presence of 10 mM ethanol, TBA and chloroform, and 42.14%, 28.19% and 13.99% in the presence of 100 mM ethanol, TBA and chloroform, which demonstrated that all of SO4•−, •OH and O2•− existed in nZVI@BCs/PS system.
Electrochemical analysis
In the previous studies (Cheng et al. 2022, Yi et al. 2022), the electron transfer pathways were investigated on nonradical induced organic pollutants oxidation, and the equilibrium values of the open circuit voltage could reflect the potential of [carbon]*. As displayed in Fig. S8, the open circuit potentials of [nZVI@BCs]* and [nZVI@Cs]* were measured to evaluate the electron transfer process. Above all, the nZVI@BCs and nZVI@Cs were added into the reaction mixture, respectively. The nZVI@BCs had an open circuit voltage of approximately 0.25 V (vs. NHE), much greater than that of nZVI@Cs (averagely 0.02 V). In addition, when the PS was added into the reaction mixture the equilibrium potential raised further and the potential of [nZVI@BCs]* was raised to around 0.80 V (0.31 V for nZVI@Cs). The measures of open circuit voltages indicated that when PS was added into the reaction mixture, the potential of the [nZVI@BCs]* complexes was sharply elevated, then boosting the BPA removal rate via an electron transfer pathway.
As discussed above, the proposed radical and nonradical mechanism in our system was present in Fig. 8. On the one hand, nZVI@BCs as a reaction catalyst could react with PS to generate reactive species (SO4•−, •OH and O2•−) (Eqs. (1) to (7)) (Huo et al. 2019, Pulicharla et al. 2018, Yang et al. 2019), which could accelerate the reduction of BPA. On the other hand, PS could be bound onto the surface of nZVI@BCs to form metastable complexes ([nZVI@BCs]*) with a high oxidation potential toward BPA oxidation.
Fe0 + S2O82- → Fe2+ + SO42− (1)
Fe0 + 2H+ + O2 → Fe2+ + H2O2 (2)
Fe0 + 2H+ + H2O2 → Fe2+ + 2H2O (3)
Fe2+ + S2O82- → Fe3+ + SO4•− + SO42− (4)
SO4•− + H2O → •OH + H+ + SO42− (5)
•OH + H2O2 → O2•− + H+ + H2O (6)
Fe2+ + O2 → Fe3+ + O2•− (7)
BPA removal in different conditions
To study the practical application of the nZVI@BCs/PS system, the effect of nZVI@BCs dosage, PS dosage, initial pH, water common anions and various water matrix on the removal of BPA was studied, as shown in Fig. 9. The BPA removal ability with different nZVI@BCs dosages was shown in Fig. 9(a). With increasing nZVI@BCs dosage, both adsorption and degradation of BPA were enhanced. Similar phenomenon was observed that the enhanced BPA removal via increasing PS dosage (Fig. S9 and Table S5). Nevertheless, further increasing PS dosage did not result in a significant removal, which may because the available active sites on the nZVI@BCs surface were mostly occupied by PS (Chen et al. 2019). BPA removal in terms of initial pH was shown in Fig. 9(c) and it’s found that the initial pH could affect the adsorption and oxidation processes. Fig. S10 and Table S6 showed that when reaction solutions were adjusted to be more alkaline, the kobs were first decrease, and then increase apparently. It is because that the activation of PS can be facilitated in alkaline conditions (Duan et al. 2016). The related pH variation during the experiments was also recorded in the Fig. S11. The solution pH drastically decreased upon PS addition, which was attributable to the strong acidity of PS (Wu et al. 2018) and the formation of acid intermediates and H+ (Eqs. (8) and (9)) (Tang et al. 2018).
S2O82− + 2H2O → 2SO42− + HO2− + 3H+ (8)
S2O82− + HO2− → SO42− + SO4•− + O2•− + H+ (9)
BPA removal with respect to different anions (Cl− and HCO3−) in nZVI@BCs/PS system was shown in Fig. 9(c). As the Cl− levels were increased from 0 mM to 10 mM, distinct increases in the k were observed, the k of first stage were increased from 0.4508 to 2.7074 min-1, which may because excessive Cl− with negative charge could also donate electrons to PS, generating superabundant active chlorine species and sulfate radicals via Eqs. (10) to (13) participating in the BPA degradation process (Luo et al. 2019, Tang et al. 2018). However, BPA removal in the background of different levels of HCO3− showed the different phenomenon, a significantly inhibition effect of both adsorption and degradation of BPA were observed when dosing HCO3− (Fig. 9(c), Fig. S12 and Table S7), and the inhibition was enhanced with increasing HCO3− concentration. Previous studies (Ghauch &Tuqan 2012, Zhang et al. 2016) showed that HCO3− was an effective scavenger for SO4•− and •OH (Eqs. (14) to (18)), which could lead to an inhibitory effect on BPA removal. What’s more, humic acid is also common water component in the actual water, so the effect on BPA removal was investigated. Fig. 9(e) illustrated that in the presence of humic acid, only 5.8% of BPA removal decreased, which showed insignificant inhibition effect on BPA removal.
To evaluate the availability of nZVI@BCs as the catalyst in real applications, three kinds of water matrices, including tap water (RW), Jiangan river water (JAW) and Mingyuan laker water (MYW) were used as the reaction medium to perform the BPA removal experiment. Table S10 showed the characteristics of three real water samples, which indicated a near neutral pH and the dissolved organic matters and the inorganic species existing in three water samples. Fig. 9(d) illustrated that the removal efficiencies in RW, JAW and MYW samples were 85.47%, 74.40% and 76.95%, respectively, which demonstrated that the real water samples could affect the effectiveness of the oxidative process. But these also suggested the application potential of nZVI@BCs in practical water treatment.
SO4•− + Cl− → SO42− + Cl• (10)
•OH + Cl• → ClOH•− (11)
ClOH•− + H+ → Cl• + H2O (12)
Cl• + Cl− → Cl2•− (13)
SO4•− + HCO3− → SO42− + •HCO3 (14)
SO4•− + CO32− → SO42− + •CO3− (15)
HCO3− + •OH → •CO3− + H2O (16)
CO32− + •OH → •CO3− + OH− (17)
•HCO3 → H+ + •CO3− (18)