The discharge of wastewater containing organic pollutants leads to serious environmental hazards. Phenolic compounds are one of the most toxic industrial pollutants. They are more degradation-resistant than other organic dyes, such as methyl orange (MO) and rhodamine B (RhB) [1, 2]. Therefore, attention is paid to developing effective removal methods for phenolic compounds in industrial wastewater. An advanced oxidation process (AOP) has attracted attention. TiO2 is a promising photocatalystwidely applied in the degradation of organic pollutants. However, the wide band gap (3.2 eV) photocatalyst can be only driven by ultraviolet light, which limits its application due to the low utilization of solar energy [3, 4]. To utilize solar energy effectively, researchers consideredtwoaspects: modification of TiO2 to extend its response to the visible light region [4-6] and development of new photocatalysts with a narrow band gap [7-9].
Bismuth ferrite (BiFeO3) is a potential photocatalyst due to its narrow band gap (2.2-2.5 eV), chemical stability and low cost [10]. However, the photocatalytic performance of pure BiFeO3 is poor due to the high recombination rate of photogenerated electron-hole pairs, and impurity phases are formed easily during its preparation, which restrict its practical application in the photocatalytic field [11]. So far, the modifications of BiFeO3 are focused on the preparation method [12], noble metal deposition [13], and the construction of a heterojunction [10]. The photocatalytic performance of BiFeO3 has been improved to some degree, while a simple and low cost method is always required. It is known that ion doping as a simple and effective method to enhance the photocatalytic performance by introducing an impurity level or tailoring the band gap and promoting the separation of photo-generated electrons and holes. However, studies on ion-doped BiFeO3 are mainly concentrated on the enhancement of optical properties [14], multiferroic properties [15], ferromagnetism [16] and photovoltaics [17]. Studies on the photocatalytic degradation of refractory organics are few. There are scarce reports on the study of the photocatalytic reaction mechanism of La-doped BiFeO3, effect of lanthanum on the phase, surface oxygen species and photocatalytic activity for phenol degradation. Therefore, it is necessary to systematically explore the influence of La3+ on the photocatalytic behavior of BiFeO3 and the reaction mechanism.
There are no reports on La-doped BiFeO3 photocatalysts prepared by the sol-gel method for phenol degradation. Here, we first prepared a series of photocatalysts Bi1-xLaxFeO3 (mole fraction x = 0, 0.10, 0.15, 0.20) by a one-step sol-gel method using citric acid as complexing agent. The effects of lanthanum doping on the phase, surface oxygen speciesand photocatalytic activity were studied. A reaction mechanism was proposed.
All of the reagents were analytical grade and used without further purification. Lanthanum nitrate hexahydrate (La(NO3)3· 6H2O) and citric acid were obtained from Tianjin Fengchuan Chemical Reagent Technology Co., Ltd, China. Bismuth nitrate pentahydrate (Bi(NO3)3·5H2O), iron nitrate nonahydrate (Fe(NO3)3·9H2O), phenol (C6H5OH) and nitric acid (HNO3, 65 wt%) were obtained from Sinopharm Chemical Reagent Co., Ltd, China. Distilled water was used throughout this study.
The Bi1-xLaxFeO3 samples were prepared by a one-step sol-gel process. Stoichiometric amounts of La(NO3)3·6H2O, Bi(NO3)3·5H2O and Fe(NO3)3·9H2O were dissolved in 100 mL deionized water mixed with 10 mL HNO3 and an appropriate amount of citric acid. The solution in a thermostatic water bath at 70 °C was stirred with a magnet until the solution became a viscous crimson gel. An orange red xerogel was obtained after the viscous gel was dried in an oven at 100 °C under air atmosphere. The xerogel was calcined at 500 °C for 2 h and then at 600 °C for 1 h in a muffle furnace with a heating rate of 5 °C/min. Brick red Bi1-xLaxFeO3 samples were obtained. Undoped BiFeO3 powders were prepared as in the above but without adding La(NO3)3·6H2O.
X-ray powder diffraction (XRD) patterns were recorded with an Empyrean X-ray diffractometer (Panalytical, Holland) using Cu Kα radiation (λ = 0.1542 nm, scanning step = 0.0667°/s), scanning over the range of 2θ = 10°-60° and operating at 40 kV and 40 mA. Scanning electron microscopy (SEM) used an S-4800 microscope (Hitachi, Japan) with acceleration voltage of 20 kV. Transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HR-TEM) analyses were carried out on a G2F20S-TWI (Tecnai, American) at 200 kV. X-ray photoelectron spectroscopy (XPS) was performed on an ESCALAB 250 spectrometer (Thermo, American). The binding energy positions were calibrated against C 1s at 284.6 eV. The optical property was characterized by UV-Vis diffuse reflectance spectroscopy on a UV-3600 Spectrophotometer (Shimadzu, Japan) over the wavelength range of 200 to 800 nm. BaSO4 was used as a reflectance standard. Fluorescence emission (PL) spectra were measured on a FLS920-type fluorescence spectrophotometer (Edinburgh Instruments, England) with light source of 450 W xenon lamp and excitation wavelength of 375 nm. Chemical oxygen demand (COD) was determined by a 5B-3F instrument (Lianghua, China), and COD removal rate was calculated by the equation: COD removal rate = (CODinitial - CODfinal)/CODinitial × 100%, where CODinitial is the COD of the phenol solution after the dark reaction and CODfinal is that after irradiation.
The photocatalytic performance of the catalysts was evaluated by the photodegradation of phenol in aqueous solution. The photocatalytic experiments were carried out on a CEL-LAB500 multibit photochemical reaction instrument (AULTT, China). A 300 W halogen lamp was used to simulate solar irradiation. A constant room temperature was maintained by replenishing the cooling water during the reaction. In a typical photocatalytic reaction, 30 mg photocatalyst was added into a 30 mL phenol aqueous solution (20 mg/L) with air ventilation and magnetic stirring. Before the illumination was turned on, the catalyst was placed in the dark for 30 min to reach adsorption equilibrium. The sampling during the photocatalytic reaction was taken every 30 min and the suspension was irradiated for 180 min. The phenol concentration was measured by an ultraviolet spectrophotometer (UV-7504PC) at the wavelength of 270 nm after centrifugation. Phenol degradation (D) was calculated by formula D = (A0 - At)/A0 × 100%, where A0 is the absorbance of the phenol solution before the light source was turned on, and At was after irradiation. A blank test of phenol solution irradiated without any catalyst was also performed for comparison.
In this experimental procedure, terephthalic acid (TA) as a probe molecule readily reacts with •OH to produce a highly fluorescent product, 2-hydroxyterephthalic acid (excitation, λexc = 315 nm; emission, λem = 425 nm) [18]. The 15% La-doped BiFeO3 sample (0.1 g) was dispersed in 100 mL of a 0.5 mmol/L TA aqueous solution with 2 mmol/L concentration of NaOH at room temperature. The resulting solution was magnetically stirred and the suspension was exposed to simulated sunlight. At 10 min intervals, the suspension was sampled and centrifuged to measure the maximum PL intensity using a fluorescence spectrophotometer with an excitation wavelength of 315 nm.
The phase composition of the samples is exhibited in Fig. 1. For the BiFeO3 sample without La doping, not only the rhombohedral phase of BiFeO3 (JCPDS No. 86-1518) but also Bi25FeO40 (JCPDS No. 02-2190) at 28° and 33° were detected. Along with the increase of lanthanum doping, the characteristic peaks of Bi25FeO40 disappeared. However, when the mole fraction of lanthanum was increased to 20%, the characteristic peaks of Bi25FeO40 appeared again. This was derived from the tilt and twist of FeO6 octahedra and the collapse of the structure around the Bi atoms resulting from the smaller radius of the La ions [19]. This phenomenon indicated that an appropriate amount of La3+ doping can suppress the generation of the impurity Bi25FeO40 phase. Enlarged views of the strongest peak are shown in the inset of Fig. 1. It clearly can be seen that the (104) diffraction peak shifted to a higher angle with the increase of La content, suggesting that the La ions were incorporated into the perovskite structure. More importantly, it was also observed that the (110) and (104) diffraction peaks of BiFeO3 were more split compared to those of Bi1-xLaxFeO3, further demonstrating that La ions were successfully doped into the BiFeO3 lattice [20].
Table 1 shows the cell parameters of the catalysts computed by the X’Highscore plus software. The average crystallite size was calculated with the full width at half maximum using the (110) plane of BiFeO3 peaks using the Scherrer formula [21]. The crystallite size decreased from 79.2 nm (La = 0%) to 46.1 nm (La = 20%) with an increase of La3+ dopant concentration, implying that the lanthanum ion could restrict the grain size of BiFeO3. Besides, the radius of the lanthanum atom (RLa3+ = 0.116 nm) is smaller than that of the bismuth atom (RBi3+ = 0.117 nm), so the shrinkage of the cell volume further proved the incorporation of lanthanum atoms into the crystal lattice of BiFeO3.
The SEM images of the samples are shown in Fig. 2. Obviously, the particle size of undoped BiFeO3 was different from 15% La-doped BiFeO3. Undoped BiFeO3 was composed of a number of elongated and massive particles with the size of 210-400 nm, which adhered to each other. Besides, the morphology of 15% La-doped BiFeO3 exhibited irregular spherical particles with the size of 60-150 nm. Clearly, the particle size of the 15% La-doped BiFeO3 sample was smaller than that of undoped BiFeO3, which further illustrates that the incorporation of La3+ could decrease the catalyst particle size. The smaller size is beneficial to the transfer of photo-generated electrons and the enhancement of the photocatalytic performance [22].
The TEM images of undoped BiFeO3 and 15% La-doped BiFeO3 are shown in Fig. 3. It can be seen that the particle size of undoped BiFeO3 was larger than that of the La-doped BiFeO3 catalyst, which is in agreement with the SEM results. The fringe interval with a distance of 0.40 nm corresponded to the (012) crystal plane of undoped BiFeO3 as shown in Fig. 3(a), while the interplanar distance of 0.38 nm matched the same (012) crystallographic plane of 15% La-doped BiFeO3. The smaller interplanar spacing indicated that La3+ was partly incorporated into the structure of BiFeO3 [23].
Energy dispersive spectroscopy (EDS) analysis was carried out to examine the element distribution of the 15% La-doped BiFeO3 catalyst. This is shown in Fig. 4. Clearly, the rare earth lanthanum atoms were highly dispersed among the bismuth and iron atoms. Moreover, Fig. 4(e) further confirmed that the atomic ratio of La:Bi:Fe was 7.16:42.84:50.00, which indicated that the atomic ratio of La/(La + Bi) was 0.1432, close to the 0.15 in 15% La-doped BiFeO3 catalyst. Therefore, from this analysis results, it is reasonable to believe that rare earth La atoms were successfully incorporated into BiFeO3 and dispersed uniformly.
The XPS spectra of BiFeO3 and La-doped BiFeO3 were further used to investigate the chemical composition and surface state. These are shown in Fig. 5. Fig. 5(a) shows the survey spectra of the undoped BiFeO3 and 15% La-doped BiFeO3 catalysts. Clearly, the spectrum of the La-doped BiFeO3 was almost the same as that of undoped BiFeO3 except for the weaker spectral features of La 3d. Fig. 5(b), (c) and Fig. 5(d) show the high resolution La 3d, Bi 4f and Fe 2p XPS spectra of the samples, respectively. For the spectra of La 3d, it can be seen that two binding energy peaks appeared at 835.1 and 851.6 eV, corresponding to La 3d5/2 and La 3d3/2, respectively, which were attributed to La3+ [24]. Besides, as shown in Fig. 5(c), the binding energies at 158.8 and 163.9 eV were Bi 4f7/2 and Bi 4f5/2, respectively, which were assigned to Bi3+ [25]. Fig. 5(d) shows the high resolution spectra of Fe 2p. For undoped BiFeO3, the peaks corresponding to the Fe 2p3/2 and Fe 2p1/2 of Fe3+ appeared at 714.5 and 728.0 eV, respectively. The Fe 2p3/2 spectra of La-doped BiFeO3 were shifted towards lower binding energy, which was near 710 eV, which suggested the presence of Fe2+ [26, 27]. The slight shift of the XPS spectra of Bi 4f and Fe 2p indicated that the chemical environment of Fe and Bi in BiFeO3 was changed because of La doping. Notably, the presence of Fe ions of various valence states will be conducive to the formation of more oxygen vacancies [28, 29], which is beneficial to the surface adsorption on BiFeO3 of organic and oxygen species.
A photocatalytic reaction involves a series of redox reactions and the surface oxygen species play an important role in the photocatalytic process, so the O 1s XPS spectra were analyzed. This is depicted in Fig. 6. All the O 1s XPS peaks can be well-fitted by three Lorentzian-Gaussian lines, which indicated that there were three kinds of surface oxygen species. The binding energy at 529-530 eV was due to lattice oxygen atoms (Olat). The main peak at 531eV can be assigned to hydroxyl oxygen (OOH-). The binding energy of 532-533 eV was assigned to surface adsorbed oxygen (OO2) [30, 31]. The peak fitting and analysis data of the three oxygen species are shown in Table 2. It can be seen that the hydroxyl oxygen of the La-doped BiFeO3 was more than that of undoped BiFeO3, especially that the surface hydroxyl oxygen content of 15% La-doped BiFeO3 reached 51.89%. In a photocatalytic reaction, a higher content of surface hydroxyl groups can produce more reactive oxygen species (hydroxyl radical, •OH), which is beneficial to the photocatalytic reaction [32].
To examine the optical properties of the photocatalysts, the UV-Vis DRS and PL spectra were investigated. These are shown in Fig. 7. As can be seen in Fig. 7(a), all the samples exhibited a wide absorption ranging from 300 to 600 nm. The optical intensities of the La-doped BiFeO3 were enhanced and the absorption edges have an obvious red shift to different degrees. Especially, the 15% La-doped BiFeO3 exhibited the strongest absorption property. The band gaps of the samples were further studied using the Kubelka-Munk (K-M) formula [33]. The plot of [F(R∞)hv]2 versus hv (photon energy) is in the inset in Fig. 7(a). The band gaps of the BiFeO3 and Bi1-xLaxFeO3 (x = 0.10, 0.15, 0.20) samples were estimated to be 2.10, 2.15, 2.06, and 2.07 eV, respectively, suggesting that lanthanum doping narrowed the band gap. The changes of band gaps may be attributed to the formation of an impurity level due to the incorporation of La [34], which will be beneficial for the photocatalytic performance.
PL conducted at the excitation wavelength of 375 nm can provide information about surface defects, oxygen vacancies and electron-hole recombination. As shown in Fig. 7(b), 0% and 15% La-doped BiFeO3 show the obvious PL signals at the wavelength of 510 nm, which was attributed to the electronic transition from the conduction band of BiFeO3 to its valance band [10]. Generally speaking, a lower PL intensity indicates a lower recombination of charge carriers. The lower PL intensity of 15% La-doped BiFeO3 suggested the better separation of the charge carriers, and thus interfacial charge-transfer efficiency was enhanced after the incorporation of lanthanum.
The photocatalytic performance of the samples was evaluated by the degradation of phenol (20 mg/L) under simulated solar light irradiation (300 W halogen lamp) for 180 min. The results are illustrated in Fig. 8. As shown in Fig. 8(a), the photocatalytic performance of BiFeO3 was enhanced by the incorporation of an appropriate amount of lanthanum. Especially, the degradation rate of 15% La-doped BiFeO3 reached 96%, which was much higher than that of undoped BiFeO3. Moreover, the COD measurement was further conducted to investigate the mineralization degree of phenol, which was quantified by the COD removal rate. This is shown in Fig. 8(b). The COD removal rate of 15% La-doped BiFeO3 was 81.53%, which was much higher than that of undoped BiFeO3 (14.81%). Obviously, the higher COD removal rate indicated the higher mineralization rate of the organic material in the photocatalytic degradation. In addition, the plot of ln (C0/Ct) versus time demonstrated that the photocatalytic reaction followed first order kinetics. The apparent rate constants were obtained from Fig. 8(c). Furthermore, the phenol degradation rates of the La-doped catalysts were clearly higher than that of the undoped sample.
The stability of the catalyst is an important issue in the photo-degradation reaction, so the stability of the 15% La-doped BiFeO3 catalyst was investigated. As shown in Fig. 8(d), 15% La-doped BiFeO3 exhibited excellent photocatalytic stability in the degradation of phenol. At the third recycle test, the degradation rate of phenol was still 89.47%, which was similar to that of the first run (95.06%). Besides, Fig. 8(d) shows the XRD patterns of the fresh and used 15% La-doped BiFeO3 catalyst. There was no diffraction peak of an impurity phase that can be detected. The difference was the intensity of the corresponding peaks, which further confirmed the stability of the 15% La-doped BiFeO3 photocatalyst.
In the photocatalytic process, the active species mainly involve the hole (h+), hydroxyl radical (•OH) and superoxide radical (O•−2). So what active species is the key in the photo-degradation reaction? Free radical capture experiments were conducted by adding different active species scavengers using phenol photo-degradation as the model reaction. Briefly, the phenol photo-degradation was repeated by adding benzoquinone (BQ) as a superoxide radical scavenger (O2•-), tert-butyl alcohol (TBA) as a hydroxyl radical scavenger (•OH) and ethylene diamine tetraacetic (EDTA) as a hole (h+) scavenger [35-37]. As shown in Fig. 9(a), the photo-degradation of phenol was inhibited when EDTA and TBA were added, while there was no obvious photo-degradation reduction with the addition of BQ. This gave the evidence that the degradation of phenol was dominated by the oxidation reaction of hydroxyl radicals and direct hole oxidation.
To probe the generation of the •OH radical in the simulated sunlight irradiated suspension of the La-doped BiFeO3 sample, the photoluminescence (PL) spectra with terephthalic acid (TA) as a probe molecule were carried out [38]. A maximum intensity peak in the PL spectra was observed at 425 nm with an excitation wavelength of 315 nm as shown in Fig. 9(b). It was also observed that the PL intensity increased gradually with increasing irradiation time, showing a markedly enhanced production of 2-hydroxyterephthalic acid. This demonstrated that •OH was indeed generated on the irradiated surface of La-doped BiFeO3. The formation of •OH on the surface of the La-doped BiFeO3 photocatalyst would lead to enhanced photocatalytic activity for the degradation of phenol under simulated sunlight irradiation.
The generation of the active species is closely related to the potential energy of the conduction band (CB) and valence band (VB) of the semiconductor. The potentials of the CB and the VB edges of BiFeO3 were obtained using Mulliken electronegativity theory [37, 39]. This is shown as follow:
where EVB is the VB potential, X is the electronegativity of the semiconductor (X = (χ(A)aχ(B)bχ(C)c)1/(a+b+c), χ is the electronegativity of the elements), Ee is the standard hydrogen electrode potential (~4.5 eV), Eg is the semiconductor band gap obtained from Fig. 6(a), and ECB is the CB potential. From the above, the VB potential of the BiFeO3 semiconductor photocatalyst was calculated to be 2.55 eV and the CB potential was 0.45 eV.
When solar light irradiates La-doped BiFeO3, BiFeO3 is activated to produce h+ and e- as described in Eq. (1). The electrons at the CB of BiFeO3 cannot reduce O2 to O•-2 because the CB edge potential of BiFeO3 (0.45 eV vs. NHE) is more positive than the standard redox potential Eθ (O2/O•-2) (-0.33 eV vs. NHE), which is consistent with the result of the free radical capture test. So the superoxide radicals are not the active species. However, the CB edge potential of BiFeO3 is more negative than the standard redox potential Eθ (O2/H2O2) (0.685 eV vs. NHE) [40], suggesting that oxygen adsorbed on the surface of the composite semiconductor can react with two electrons to form H2O2 as shown in Eq. (4). Subsequently, H2O2 combines with one electron to further form •OH as described in Eq. (5) and Eq. (6). •OH has a strong oxidation characteristic to participate in the photocatalytic reaction as shown in Eq. (8). Furthermore, the VB edge potential of BiFeO3 (2.55 eV vs. NHE) is more positive than the standard redox potential of Eθ (•OH/OH-) (1.99 eV vs. NHE) [31], suggesting that the accumulated holes on the VB of BiFeO3 can oxidize OHads- to form ·OH as shown in Eq. (7). The XPS results showed a large amount of hydroxyl oxygen adsorbed on the surface of the La-doped BiFeO3, which provided a reliable basis for the inference. Some of the holes may be directly involved in the oxidation of organic compounds according to the above free radical capture experiment. Moreover, for the substitution lanthanum, the band gap was narrowed from 2.10 to 2.06 eV and a dopant energy level (La3+ impurity level) was formed below the conduction to enhance the absorption of visible light [41]. Consequently, the transmission distance of the charge carriers would be shortened, which is beneficial to the transfer of electrons to promote the photodegradation of phenol [42]. Based on the above discussion, a photocatalytic degradation mechanism was proposed as shown in Fig. 10. The phenol degradation is as follows:
We prepared BiFeO3 perovskite catalysts with different lanthanum dopant concentration by the sol-gel method using citric acid as a complexing agent. Incorporation of lanthanum suppressed the formation of Bi25FeO40 and improved the photocatalytic performance of BiFeO3. The photocatalytic activity of the 15% La-doped BiFeO3 catalyst was 3.4 times that of undoped BiFeO3 in the degradation of phenol under simulated sunlight irradiation for 180 min. The COD removal rate reached 81.53%. The smaller particles, stronger light absorption, lower electron-hole recombination rate and more hydroxyl groups on the catalyst surface were responsible for the excellent catalytic activity of 15% La-doped BiFeO3.
2016,
Vol. 37 Issue (8): 1283-1292 
PDF (1188 KB)
2016,
Vol. 37 Issue (8): 1283-1292 PDF (1188 KB)