Ammonia (NH3) is the second largest product by volume of the Haber-Bosch process and is used for feedstock and fertilizers [1]. However, as a major nitrogen-containing pollutant, NH3 is also a nutrient source that can promote eutrophication and algal growth in natural waters [2]. Excessive amounts of NH3 in the environment can exert harmful effects on human health [3]. NH3 attacks the human respiratory system, skin, and eyes, and exposure to high concentrations (>300 ppm) may cause death [4, 5].
The discharge of NH3 from wastewater has become an urgent challenge. Various conventional technologies, including biological processes, chemical precipitation, advanced oxidation processes, ion exchange, air stripping, adsorption, and the use of membranes, are applied to remove NH3 from industrial wastewater [6-11]. However, these treatment methods are generally focused on large chemical systems and are often energetically and operationally intensive.
Photocatalytic degradation of NH3 [2, 12-17] has been proposed as a practical method to decontaminate wastewater for renewable use because of its low cost and ability to rapidly disinfect and purify water. Many researchers have found that titanium dioxide (TiO2) and TiO2-based materials, which are the primary photocatalysts explored for NH3 degradation, can completely destroy N-H bonds [18-23]. However, TiO2-based materials can only utilize approximately 4% of the available solar energy, including 43% of visible light, because of the wide band gap of TiO2. Thus, the development of novel photocatalysts that can decompose NH3 under visible-light irradiation is necessary.
BiFeO3, a perovskite-type photocatalyst, has attracted considerable attention because of its narrow band-gap energy (2.1 eV) [24], high chemical stability [25], and simultaneous presence of ferroelectric and magnetic order parameters [26]. In addition to potential electronic and magnetic applications, BiFeO3 powders have been used as a new visible-light photocatalyst [27]. Many researchers have utilized BiFeO3 as a high-efficiency photocatalyst to degrade various organic pollutants under ultraviolet (UV)-visible or visible-light irradiation [28-33].
Recently, we have fabricated photocatalysts based on graphene- manganese ferrite (rG-MnFe2O4) and activated carbon bonded to nickel ferrite (AC-NiFe2O4). These photocatalysts achieved high performance in the photocatalytic degradation of NH3 in the presence of hydrogen peroxide under visible-light irradiation [16, 17]. However, manufacturing hydrogen peroxide consumes energy and/or other chemical resources. Therefore, the development of magnetically separable graphene- based photocatalysts with high catalytic activity in the absence of hydrogen peroxide is warranted.
In this study, BiFeO3 nanocrystals were deposited on graphene sheets via a one-step hydrothermal method. BiFeO3 nanoparticles possess weak magnetic properties. This weak magnetism makes the rG-BiFeO3 composite magnetically separable in a suspension system and thus conveniently renewable and reusable. The rG-BiFeO3 composite material was tested for the first time as a highly active catalyst for NH3 degradation under visible-light irradiation in the absence of hydrogen peroxide. The optimum operational parameters for NH3 degradation were also determined.
Ferric chloride hexahydrate (FeCl3·6H2O) and sodium hydroxide (NaOH) were purchased from Tianjin Damao Chemical Factory. Bismuth nitrate pentahydrate (Bi(NO3)3·5H2O), ammonium sulfate ((NH4)2SO4) and ammonium chloride (NH4Cl) were obtained from Nanjing Chemical Reagent Co., Ltd. Graphite powder with an average particle size of 30 μm was acquired from Shanghai Colloid Chemical Plant. All reagents used were of analytical grade and applied without further purification. All solutions were prepared with 18.2 MΩ cm deionized Milli-Q water.
Bi(NO3)3·5H2O (4.8507 g, 0.01 mol) and FeCl3·6H2O (2.7030 g, 0.01 mol) were separately dissolved in 15.0 mL of deionized water. The solutions were then stirred together to form Solution A. The final pH was adjusted to approximately 12 by adding NaOH. Graphene oxide (GO) was synthesized via a modified Hummers method as previously described [34]. The as-synthesized GO (125.2 mg, ~4% of the BiFeO3 mass) was dispersed in 10.0 mL of deionized water with an ultrasonic vibrator, and the resultant solution was denoted Solution B. Solution A was added dropwise to Solution B under stirring. NaOH (2.4 g, 0.06 mol) was dissolved in 10.0 mL of deionized water. This solution was added dropwise to the mixed suspension solution described above under continuous stirring. Deionized water was also added to the suspension to obtain a final volume of 60 mL. Afterward, the suspension solution was transferred to a 100-mL Teflon-lined stainless-steel autoclave, which was subsequently sealed and maintained at 180 ℃ for 10 h. The solution was cooled to room temperature and filtered to obtain rG-BiFeO3 precipitates. The products were rinsed three times with water to remove excess NaOH and other electrolytes. A magnetic powder was obtained after sintering at 200 ℃ for 4 h. This weakly magnetic powder was used for the characterization and photocatalytic tests. Pure BiFeO3 was similarly prepared for comparison.
X-ray diffraction (XRD) was performed with an X’Pert-Pro MPD X-ray diffractometer (Panalytical, Netherlands). The X-ray source emitted Cu Kα radiation with a wavelength of 0.154 nm at a tube voltage of 40 kV and a tube current of 40 mA. Morphological observations were conducted with a transmission electron microscope (Tecnai G220; FEI, USA). The rG-BiFeO3 and GO powders were dispersed in water by an ultrasonication device, placed on carbon-coated copper grids, and dried under ambient conditions prior to transmission electron microscopy (TEM). Brunauer-Emmett-Teller (BET) surface areas of the photocatalysts were measured by N2 adsorption-desorption on a Micromeritics ASAP 2200. Prior to measurement, the samples were pretreated at 150 ℃ under vacuum (1.33 Pa) for 2 h. A Fourier transform infrared (FT-IR) spectrophotometer (Spectrum BX; PerkinElmer Ltd., USA) was used to characterize group vibrations at an optical resolution of 4 cm−1. Raman spectrum measurements were performed on a Renishaw inVia Reflex Raman Microprobe. UV-vis diffuse reflectance spectroscopy was performed with a Hitachi U-3010 UV-vis spectrometer. The mulls of rG-BiFeO3 were supported on KBr plates.
Photocatalytic experiments for NH3 degradation were conducted under visible-light irradiation (λ > 400 nm). A 300-W UV-visible lamp (OSRAM, Germany) was used as a light source. Photocatalytic degradation of NH3 solution (50.0 mg/L) was performed in a 100-mL beaker at room temperature (25 ± 2 ℃). The distance between the lamp and the test solution was approximately 10 cm, and the wall of the beaker was shielded from surrounding light by aluminum foil. Visible light was allowed to pass through a λ > 400 nm cut-off filter that covered the window of the beaker; this filter absorbed UV light and allowed visible light (λ > 400 nm) to pass through. About 50 mL of solution was used in a typical photocatalytic experiment. The NH3 solutions were prepared according to the desired concentrations, and 0.20 g of the rG-BiFeO3 catalyst was used for the photocatalytic experiments. NaHCO3-Na2CO3 buffer (0.1 mol/L) was used to control the pH of the test solutions. At predetermined time intervals, approximately 3 mL of the sample was extracted with a 5-mL plastic syringe and then passed through a 0.45-μm membrane filter to a clean and dried glass tube for further analysis. All photocatalytic degradation experiments were performed in triplicate.
A double-beam TU-1901 spectrophotometer was used to determine the concentration of NH3 by reaction with Nessler’s reagent during the photocatalytic process [35]. Nessler’s reagent is an alkaline solution of dipotassium tetraiodomercurate (II). This reagent was prepared by dissolving 10 g of HgI2 and 7 g of KI in water, adding to NaOH solution (16 g of NaOH in 50 mL of water), and then diluting with deionized water to 100 mL. The reagent was stored in dark bottles and diluted appropriately before analysis. NH3 reacts with this reagent to yield colored solutions via Reaction (1). As the absorbance of the solutions showed a maximum value at 392 nm, absorbance was measured at this wavelength for analysis.
In Reaction (1), [HgI4]2− was yellow, and HgO·Hg(NH2)I was brown.
Fig. 1 presents the XRD patterns of BiFeO3 and the 4% rG-BiFeO3 hybrid catalyst (where 4% is the raw-material mass ratio of GO to BiFeO3). The diffraction peaks obtained can be indexed as perovskite-type BiFeO3 (JCPDS Card No. 20-0169). The peaks at 2 θ values of 22.43°, 31.90°, 39.52°, 45.76°, 51.30°, and 57.07° can be indexed to the (012), (104), (202), (024), (116), and (300) crystal planes of spinel BiFeO3, respectively. The diffraction peak of rG-BiFeO3 was shifted toward a lower 2θ value, and the intensity of the diffraction peaks for the composites was reduced compared with pure BiFeO3, which can be attributed to the intercalation of GO. No impurity peak was found in the rG-BiFeO3 hybrid.
The average diameter (D ) of the as-synthesized rG-BiFeO3 particles was calculated as 18.5 nm from the Debye-Scherrer equation, D = K λ /(W cos θ ), at a diffraction angle of 31.90° (2 θ ), where W is the breadth of the observed diffraction peak at its half height, K is the so-called shape factor (usually approximately 0.89), and λ is the wavelength of the X-ray source used (0.154 nm in our measurement). The D value derived from the Debye-Scherrer equation is consistent with that obtained from the TEM observations.
FT-IR spectroscopy was conducted to characterize the chemical structures of GO, BiFeO3, and rG-BiFeO3, and hence the changes in the first two after the impregnation of GO with BiFeO3 (Fig. 2). The various surface functional groups of GO are clearly evident in Fig. 2(1), including the C-O stretching vibration at 1051 cm−1 and the C=O (carboxyl) stretching vibration at 1726 cm−1, which were attributed to the typical functional groups of GO (e.g., -O-, -COOH, and -OH) [36-38]. The peak at 1618 cm−1 can be assigned to the C=C skeletal vibration of unoxidized graphitic domains and/or the H-O-H bending band of the adsorbed H2O [39, 40].
Fig. 2(3) clearly shows that almost all the characteristic peaks of the oxygen-containing groups (C=O, O-H, C-OH, and C-O-C) disappeared for rG-BiFeO3, suggesting that the GO in the composite was reduced to graphene. The adsorption by rG-BiFeO3 around 1559 cm−1 can be assigned to the stretching vibrations of the unoxidized carbon backbone [41]. The band at 634 cm−1 can be attributed to the Bi-O vibration in the BiO6 octahedra [42], and the band at around 811 cm−1 was due to the presence of traces of trapped NO3− ions in the composite nanoparticle [43].
It exhibits characteristic peaks for BiFeO3 at 441 cm−1 (Fe-O stretching vibrations of the FeO6 octahedra) and 548 cm−1 (O-Fe-O bending vibrations of the FeO4 tetrahedra) [44, 45]. After being combined with rG, the peak positions of BiFeO3 remained unchanged in the FT-IR spectrum of rG-BiFeO3, but the corresponding intensities changed considerably, especially for rG-BiFeO3 (Fig. 2). This is further evidence of the possible formation of chemical bonding between BiFeO3 and rG [46].
The Raman spectra of BiFeO3 and rG-BiFeO3 are shown in Fig. 3. The G and D bands of rG-BiFeO3 (4%) were observed at 1598 and 1361 cm−1, respectively. In previous studies, Raman peaks at 1580 and 1350 cm−1 were assigned to the G and D bands of GO [47]. Compared with GO, the blue shift in the G and D bands of rG-BiFeO3 indicated that the GO was reduced to graphene during the hydrothermal process [48].
Fig. 4 shows the TEM images of GO, BiFeO3, and rG-BiFeO3. It reveals the wrinkled surface texture of the exfoliated GO sheets. BiFeO3 nanoparticles were sparsely deposited on the two-dimensional rG. These nanoparticles evidently tended to accumulate along the edge of the rG (Fig. 4(c)), similarly to the previously reported BFO249/rG (4.5) composites [46]. The diameter of the particles was about 18.5 nm.
Fig. 5 shows the degradation curves of NH3-N without catalyst under visible-light irradiation; with rG-BiFeO3 (4%) as the catalyst under visible-light irradiation, or in the dark; and with BiFeO3 as the catalyst under visible-light irradiation. It shows that the photodegradation rate of NH3-N after visible-light irradiation without catalyst for 8 h was 5.9%. The photodegradation rate after reaction with rG-BiFeO3 (4%) as a catalyst for 8 h in the dark was 9.1%. This percentage corresponds to the amount of NH3 adsorbed on the rG-BiFeO3 catalyst. After visible- light irradiation with pure BiFeO3 as a catalyst, the degradation rate was 57.5%, indicating that pure BiFeO3 is photocatalytically active and can be used as a photocatalyst under visible light. Notably, the rG-BiFeO3 (4%) hybrid catalyst under the same conditions (visible-light irradiation for 8 h) achieved 91.2% degradation of NH3. Therefore, the combination of BiFeO3 and graphene resulted in exceptionally high catalytic activity for the degradation of the contaminant NH3. A possible cause of this significant enhancement in photoactivity was a powerful synergistic effect of the combination of BiFeO3 and the rG sheets, which promotes efficient separation of electrons (e−) and holes (h+). In this mechanism, the separated e− migrated along the surface of the rG, and the e− on the surfaces of both rG and BiFeO3 were trapped by the adsorbed O2 to form superoxide anion radicals (·O2−) [46]. Meanwhile, the holes were scavenged by the adsorbed water to form hydroxyl radicals. Finally, the holes, superoxide anion radicals, and hydroxyl radical directly oxidized NH3 adsorbed on the active sites of the rG-BiFeO3.
A series of rG-BiFeO3 nanocomposite photocatalysts with different graphene contents were prepared, and their photocatalytic activities under visible light were evaluated by the degradation of NH3 in aqueous solutions. As shown in Fig. 6, the degradation ratios initially increased with increasing mass concentration of rG in the hybrid catalysts, reaching a maximum value of 91.2% with 4.0% rG. The significant enhancement in photo-activity under visible-light irradiation was connected with the reduction of graphene oxide. Therefore, increasing the content of rG facilitated the transfer of photo- generated electrons in BiFeO3 from the conduction band to the reduced GO, effectively preventing the direct recombination of electrons and holes. In addition, the BET analysis revealed that while the specific surface area of BiFeO3 was 21.0 m2/g, that of the rG-BiFeO3 (4%) composite was much higher, 48.6 m2/g. The much greater BET surface area of the rG-BiFeO3 is likely to have provided more active adsorption sites and photocatalytic reaction centers than BiFeO3
However, when the rG contents were increased to 6.0% and 8.0%, the degradation ratios slightly decreased in comparison with the 4.0% composite, contrary to our expectations. This decrease may have been caused by a reduction in the quantity of NH3 adsorbed on BiFeO3. Although increasing the rG content in the hybrid catalyst improved the absorption of visible light, an overload of rG may have inhibited the adsorption of NH3 on the surface of the BiFeO3 catalyst by decreasing the availability of active sites, i.e., transition metal atoms exposed on the surface of the hybrid catalyst. This phenomenon would have led to lower activity, which is consistent with a previous study of the rG-MnFe2O4 photocatalyst [16]. Therefore, 4% was the ideal mass ratio, and the 4% rG-BiFeO3 catalyst was used in all subsequent tests.
The pH value of the solution plays an important role in the photocatalytic degradation of organic pollutants [50, 51]. The effect of pH on the photodegradation of NH3 was studied in the pH range of 6-11. As shown in Fig. 7, the degradation of NH3 reached 2.99%, 3.57%, 14.81%, 73.62%, 82.11%, and 91.96% at pH = 6, 7, 8, 9, 10, and 11, respectively, after 8 h of irradiation. The degradation ratio of NH3 therefore increased continually with increasing solution pH. This trend is ascribed to the protonation state of NH3 [52]. As the p K a of NH4+ is 9.3, this species is predominantly protonated in solutions with pH ＜ 9.3. The presence of NH4+ ions limits the adsorption of NH3 on the BiFeO3 surface. Hydroxyl radicals (·OH) can oxidize NH3 but not NH4+ [53]. Therefore, NH3 is decomposed more rapidly in solutions with higher pH, as the fraction of NH4+ ions is smaller.
The influence of the catalyst dosage on the heterogeneous photocatalytic degradation of NH3 by rG-BiFeO3 is illustrated in Fig. 8. Degradation ratios of 78.76%, 83.3%, 92.7%, and 86.0% were achieved when the dosages of the rG-BiFeO3 catalyst were 0.1, 0.15, 0.2, and 0.25 g, respectively, after 8 h. A large enhancement in the degradation ratio of NH3 was observed when the catalyst dosage was increased from 0.1 to 0.2 g, because the increase in the amount of the rG-BiFeO3 catalyst increased the presence of active sites on the catalyst surface as well as the generation of free hydroxyl radicals [54]. However, when the dosage of the catalyst was increased from 0.20 g to 0.25 g, the decomposition rate of NH3 decreased from 92.7% to 86.0%, because higher dosages of catalyst also increased the turbidity of the reaction solution, thus reducing the intensity of the visible- light irradiation available for the photocatalytic reaction [55].
Fig. 9 shows the effect of the NH3 concentration on the degradation rate. Solutions of NH3 at various concentrations, i.e., 25.0, 50.0, 75.0, and 100.0 mg/L NH3-N, were degraded with 0.2 g of rG-BiFeO3 catalyst at pH = 11, under visible-light irradiation. Study of the degradation kinetics showed that the photocatalytic degradation of NH3-N followed pseudo-first-order kinetic behavior. As shown in the inset of Fig. 9, the average reaction rate constant (k ) was calculated to be 0.2949 min−1 and the average value of the linear correlation coefficient was 0.99381.
The stability and reusability of a catalyst are key factors for its application, and must be investigated from the economic and environmental points of view. Fig. 10 shows that there was no obvious change in the photocatalytic activity of the recycled rG-BiFeO3 catalyst after seven cycles under visible-light irradiation. Moreover, the catalyst could be easily separated from the reaction solution by simple precipitation to allow its reuse. This indicates that the catalyst is highly stable and is potentially suitable for practical applications.
Based on the literature, two different reaction pathways exist to oxidize the NH3 adsorbed on the catalyst surface to form the final products [56, 57]. The first pathway, in which N2 is formed as the final product, produces a series of NH2, NH, and N2H x + y (x + y = 0, 1, 2) intermediates. The second pathway yields NO2− and NO3− ions as the final products through a HONH2 intermediate. These two reaction pathways proceed as follows.
Reaction pathway 1:
Reaction pathway 2:
In UV-vis absorption spectrophotometry, the absorbance peaks of NO2− (nitrite) and NO3− (nitrate) ions appear at λ = 206 and 211 nm [58], respectively, if those ions are present. Therefore, the absorbance detected at those wavelengths reveals whether NO2− and NO3− ions formed during the reaction in our system. As illustrated in Fig. 11, consecutive measurements did not detect any absorbance in the wavelength range of 200-230 nm during the photocatalytic process, thereby excluding the possibility of the second reaction pathway, which yields these ions as the final products.
As indicated earlier, rG-BiFeO3 showed the highest photocatalytic activity for NH3 degradation under visible light. Compared with that of the BiFeO3 material, the diffuse reflectance spectrum of the rG-BiFeO3 shows a red shift and an enhanced absorption of visible light, indicating that the rG-BiFeO3 catalyst can utilize visible light more effectively than pure BiFeO3 for the photocatalytic degradation of NH3 (Fig. 12).
The transient photocurrent responses of BiFeO3 and rG-BiFeO3 electrodes were also recorded via several on-off cycles of irradiation. Fig. 13 shows the photocurrent transient responses under visible light and in the dark for these samples, where the photocurrent was measured at 0.8 V vs. SCE in Na2SO4 aqueous solution (0.1 mol/L). The photocurrent of the rG-BiFeO3 was much higher than that of the BiFeO3, indicating that the separation efficiency of photo-generated electrons and holes was much improved as a result of the electronic interaction between BiFeO3 nanoparticles and graphene sheets.
Based on previous research [59], the most plausible mechanism for the degradation of NH3 under visible-light irradiation is presented in Scheme 1. Photo-generated electrons and holes within the pure BiFeO3 either took part in redox reactions at the surface or recombined. However, in rG-BiFeO3, the photo- generated electrons transferred to graphene. Thus, the possibility of the recombination of electron-hole pairs decreased, enhancing NH3 degradation. Meanwhile, O2 molecules adsorbed on the surface of graphene and BiFeO3 could capture e− to form ·O2− radicals, while the holes reacted with the adsorbed water to form hydroxyl radicals. Finally, the holes, superoxide anion radicals, and hydroxyl radicals directly oxidized NH3 adsorbed on the surface of the rG-BiFeO3. This photocatalytic mechanism can be summarized as follows:
Active-species trapping experiments were carried out to identify the main oxidizing agent in the photocatalytic reaction process. Isopropanol (1.0 mL), disodium ethylenediamine tetraacetate (EDTA-2Na) (10 mg), and nitrogen gas were introduced as scavengers for the various reactive species, i.e., hydroxyl radicals (·OH), holes (hv), and superoxide anion radicals (·O2−), respectively. As shown in Fig. 14, the addition of isopropanol greatly reduced the photodegradation rate of NH3, whereas the addition of EDTA-2Na and purging with nitrogen gas had relatively weak effects on the photodegradation of NH3. Therefore, we can conclude that hydroxyl radicals played the role of the main oxidant in the rG-BiFeO3 system.
A novel heterogeneous photocatalyst consisting of graphene- supported BiFeO3 was successfully prepared using a one-step hydrothermal method. The rG-BiFeO3 hybrid photocatalyst effectively degraded NH3 into N2 under visible-light irradiation. Analysis of the reaction mechanism of the catalyst showed that its highly efficient photoactivity could be ascribed to the powerful synergistic effect of the combination of BiFeO3 and the rG sheets. Holes, superoxide anion radicals, and hydroxyl radicals, arising from the interaction between rG and BiFeO3, oxidized NH3 directly to N2. It was found that the NH3 removal efficiency increased with increasing pH. The optimal efficiency of the photocatalytic degradation process reached 91.2% at pH = 11 and a catalyst dosage of 0.2 g. Cycling and stability tests showed that the rG-BiFeO3 catalyst was highly stable through multiple cycles, and easily separable after each cycle for subsequent reuse, suggesting its potential utility in the degradation of NH3.
2017,
Vol. 38 Issue (1): 20-28 
PDF
2017,
Vol. 38 Issue (1): 20-28 PDF