- Email: [email protected]
H2O2 process and enhanced mechanism analysis
Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata
Photocatalytic oxidation of NO over TiO2 -Graphene catalyst by UV/H2 O2 process and enhanced mechanism analysis Yanan Wang 1 , Shule Zhang ∗,1 , Yiqing Zeng, Man Ou, Qin Zhong ∗ School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, PR China
a r t i c l e
i n f o
Article history: Received 4 April 2016 Received in revised form 14 July 2016 Accepted 15 July 2016 Available online 18 July 2016 Keywords: Photocatalytic oxidation NO TiO2 -GR UV/H2 O2 Mechanism
a b s t r a c t This study aimed at investigating the photocatalytic oxidation (PCO) of NO under the UV/H2 O2 system over a series of TiO2 -Graphene (TiO2 -GR) catalysts, which was able to dramatically improve the PCO efficiency of NO compared with pure TiO2 , and the NO oxidative product was the stable nitrate. The electronic interfacial interaction between GR and TiO2 resulted in a negative shift of the CB of TiO2 evidenced by MS. The excellent conductivity of GR suppressed e− /h+ pairs recombination effectively, and GR as a kind of dispersant reduced the size of TiO2 , increased the active sites. These advantages offered e− and h+ more opportunities to participate in PCO of NO, resulting in significant improvement of the PCO efficiency. The effects of the active species involved in the photocatalytic process were also examined. Moreover, different tests were designed to further confirm the improved mechanism. Further investigations showed that the h+ could oxidize NO directly. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Nitrogen oxides (NOx ) can cause ozone depletion, photochemical smog and the acid rain [1,2]. Over the past few decades, atmospheric NOx concentrations have greatly increased with the development of society and industry [3]. The SCR of NOX is considered as the most effective method for the removal of NOX from stationary sources [4]. However, this reaction must be controlled at a certain temperature (320–400 ◦ C) [5]. That indicates SCR devices must be placed before the air pre-heater, where the high concentration of ash (e.g. K2 O, CaO and As2 O3 ) in the flue gas would reduce the performance and longevity of catalysts. Moreover, this technique using NH3 to reduce NO to N2 is not economical in terms of atoms. Therefore, development of a method, by which NOx is oxidized to nitrate species dissolved in water at a low temperature, overcoming the bottlenecks of the SCR technique, is important [6]. Advanced oxidation processes (AOPs) can produce free radicals with strong oxidation, such as • OH, • O2 − and HO2 • , which can simultaneously oxidize and remove multiple pollutants, such as NOX , H2 S, trace elements, and volatile organic compounds in recent years [7]. The UV/H2 O2 process, as a representative of the AOPs, has been widely studied for the water remediation. Since the process
∗ Corresponding authors. E-mail address: [email protected] (S. Zhang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.molcata.2016.07.029 1381-1169/© 2016 Elsevier B.V. All rights reserved.
can produce more • OH and • O2 − by photolysis of H2 O2 . Cooper et al. [8] first used • OH produced by UV decomposition of H2 O2 to oxidize NO from simulated flue gas. Liu et al. [9] developed a more applicable wet UV/H2 O2 to remove NO from simulated flue gas. The process is a promising way to take place of SCR. The low utilization ration of H2 O2 limited its wide application. Titanium dioxide (TiO2 ) reported first by Frank and Bard [10], has attracted enormous attention because of its high-efficiency, nontoxicity, low cost, and photochemical stability [11]. However, the PCO of TiO2 is low due to the rapid recombining of photogenerated e− /h+ pairs. Graphene discovered first in 2004 [12] has been a rising star on the horizon of materials science because of its superior electron conductivity, unique 2D structure with a high surface area, structural flexibility and chemical stability [13]. These unusual properties make it a suitable candidate to couple with TiO2 to improve the photocatalytic performance. In 2008, Williams et al. firstly reported the fabrication of TiO2 -GR nanocatalyst through in situ induced photocatalytic reduction of GO [14]. TiO2 -GR photocatalysts have attracted a lot of attention for diverse applications, such as H2 production from water splitting [15], hydrocarbon production from CO2 reduction [16], organic photosynthesis [17]. Accordingly, TiO2 -GR photocatalysts have the potential to be applied in photocatalytic oxidation (PCO) of NO in flue gas under the UV/H2 O2 system. However, few reviews focusing on applications and mechanism analysis of TiO2 -GR catalysts to the PCO of NO under UV/H2 O2 are reported.
340
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
In this study, we investigated the PCO of NO over a series of TiO2 GR catalysts under the UV/H2 O2 system. Compared with TiO2 , the TiO2 -GR catalysts exhibited higher PCO efficiency of NO under the UV/H2 O2 system. The reasons contributing to the high PCO of NO were discussed and corresponding mechanism was further analyzed. Moreover, a trapping experiment was conducted to examine the effects of the active species involved in the PCO of NO. This study could provide a new insight into the interaction between TiO2 and GR, as well as the mechanism on PCO of NO under UV/H2 O2 . 2. Experimental 2.1. Preparation of materials 2.1.1. Synthesis of TiO2 A sol-gel method was used to prepare TiO2 support. 0.064 mol of TBOT was added to 0.128 mol of acetylacetone under continuous stirring. Then the solution was diluted with 50 mL ethanol. After being stirred for 2 h at room temperature, the solution was heated at 60 ◦ C for 4 h in a water bath and dried at 120 ◦ C for 6 h. Then it was calcined at 450 ◦ C for 3 h. 2.1.2. Synthesis of TiO2 -GR The graphite oxide (GO) used to prepare the TiO2 -GR catalyst was synthesized using the improved GO synthesis method [18]. GO was ultrasonicated in deionized water and anhydrous ethanol solution. Then the TiO2 was added to the GO solution to prepare 2.5, 5, and 7.5 wt% TiO2 -GR catalysts (2.5, 5, and 7.5 wt% represent the mass fraction of GO). Then, a Teflon-sealed autoclave was filled with the homogeneous suspension up to 80% of the total volume, maintained at 150 ◦ C for 24 h, and cooled down to room temperature naturally. The resulting catalysts were washed by water, dried at 60 ◦ C in oven to get the final TiO2 -GR catalyst with different weight addition ratios of GR. These catalysts of TiO2 and GR are denoted as TiO2 -xGR (x = 2.5, 5, 7.5). 2.2. Characterizations The structural and chemical information for prepared samples were measured by X-ray diffraction (XRD, Cu K, Purkinjie XD-3), Raman spectroscopy (DRX, Thermo Fisher, USA), Transmission electron microscope (a Philips CM-10 at 80 kV and a CM-12 at 120 kV), UV–vis diffuse reflectance spectra (DRS, Shimadzu UV-2550), X-ray photoelectron spectroscopy (PHI-5000C ESCA system), Photoluminescence spectra (PL, He-Cd laser, Labram-HR800), Electron paramagnetic resonance (EPR, Bruker EMX-10/12-type spectrometer in the X-band). Produced ions in the solution were analyzed by Ion Chromatography (IC, Dionex ICS90). The temporal absorption spectra change of RhB was characterized by UV–vis Spectrophotometer (T6 New Century, Beijing Persee Co, Ltd.). All photoelectrochemical tests were performed on a CHI660D electrochemical workstation (Chenhua Instrument, Shanghai, China) with a conventional three-electrode cell. Platinum wire and saturated calomel electrode (SCE) were used as counter and reference electrodes, respectively. The working electrode was prepared by dip-coating powders onto fluorine-doped tin oxide (FTO) substrate with a size of 1 cm × 1 cm: 2 mg of catalyst was suspended in 5 mL ethanol to produce a slurry by ultrasonic dispersion for 30 min, which was then coated on a FTO glass electrode. After the films were dried under ambient conditions, and subsequently treated at 180 ◦ C for 2 h. The electrochemical impedance spectra (EIS) were carried out in the frequency ranging from 105 to 0.1 Hz in an aqueous solution of 0.5 M Na2 SO4 . The Mott-Schottky analysis was measured at potential ranging from −0.8 to 0.6 V, an amplitude of 5 × 10−3 V, and frequency of 1 × 103 Hz. Prior to and during all measurements,
the electrolyte (0.1 mM K3 [Fe(CN)6 ] in 1 M KCl) was purged with nitrogen. 2.3. Photocatalytic activity The PCO of TiO2 -GR was evaluated by removing NO in simulated flue gas. These samples were performed in a quartz tubular vessel (d = 1 cm, h = 15 cm) in a fixed-bed with continuous flow reactor at room temperature and atmospheric pressure. A 500 W Hg-lamp equipped with a visible cut off filter to provide ultraviolet was chosen as the ultraviolet light source. The catalyst (0.2 g) was loaded in the reactor parallelled with the Hg-lamp. The reactant gases included 400 ppm NO, 5% O2 and balanced N2 with a total flow rate of 100 mL min−1 . Prior to the light irradiation, adsorptiondesorption equilibrium between NO and photocatalysts had been reached. Meanwhile, 30% H2 O2 solution was injected into the reactor with a flow rate of 0.02 mL min−1 . The gas products (every 12 min reaction) were analyzed by an Ecom-JZKN flue gas analyzer (Germany). The NO conversion was defined as Eq. (1): NO Conversion =
NOinlet − NOoutlet × 100% NOinlet
(1)
3. Results and discussion 3.1. Photocatalytic activities The photocatalysts with different GR mass ratios were applied on the PCO of NO under ultraviolet light illumination to basically evaluate the potential ability for air purification (Fig. 1a). In the absence of UV radiation, the no-catalyst system showed low NO removal efficiency. In such a case, NO could only be removed by single oxidation of H2 O2 , which could be represented by the reaction Eq. (2) [19,20]: 2NO + 3 H2 O2 → 2HNO3 + 2 H2 O
(2)
When the UV light was introduced, the NO removal efficiency improved significantly. It could be attributed to the • OH free radicals produced by photolysis of H2 O2 , and NO could been removed by the oxidation of • OH free radical (3)–(7) [21]: H2 O2 + h → • OH
(3)
• OH
(4)
NO +
→ HNO2
NO + • OH → NO2
(5)
NO2 + • OH → HNO3
(6)
HNO2 + • OH → HNO3
(7)
Under UV light, for the TiO2 catalyst, the NO removal efficiency was higher. It is well known that TiO2 can promote the formation rate of • OH and • O2 − from O2 , H2 O (surface hydroxyl groups) and H2 O2 under the H2 O2 /UV system. This could be represented by reaction Eqs. (8)–(11) [22,23]. The reaction of conduction band is represented by reaction Eqs. (8)–(9). O2 + h + TiO2 → • O2 − H2 O2 + h + TiO2 →
• OH
E = −0.05 eV
E = 0.06 eV
(8) (9)
The reaction of valence band is given by reaction Eqs. (10)–(11). H2 O2 + h + TiO2 → • O2 − OH− + h + TiO2 → • OH
E = 1.00 eV E = 2.72 eV
(10) (11)
It is well known that • O2 − could oxidize NO according to reaction Eqs. (12)–(14) [21]. Therefore, • OH and • O2 − improved the NO
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
341
Fig. 1. (a) PCO efficiency of NO; (b) iron chromatography analysis of iron in solution after 6 h photocatalytic reaction.
removal efficiency by a synergistic effect of catalyst, H2 O2 and UV radiation. NO +
•O − 2
→ NO2
−
(12)
NO + • O2 − → NO2
(13)
NO + • O2 − → NO3 −
(14)
As expected, all of TiO2 -GR composites exhibited higher photocatalytic performance than that of TiO2 . Obviously, the NO removal efficiency of TiO2 -GR catalyst increased and then decreased with increasing of GR, the photocatalytic efficiency order of the various TiO2 -GR samples could be obtained as follows: TiO2 -5%GR (80.4%) > TiO2 -7.5%GR (69.5%) > TiO2 -2.5%GR (65.2%) > TiO2 (58.5%). It suggested that the composition ratio in TiO2 -GR was crucial to obtain an optimal synergistic effect to improve the PCO of NO. Furthermore, the differential value of NO removal efficiency with UV radiation was much higher than the one without UV radiation. Therefore, the GR could improve the synergistic effect of TiO2 -GR catalyst, H2 O2 and UV light. Additionally, a comparative experiment for the PCO of NO was carried out by TiO2 -5%activated charcoal (TiO2 -AC) and TiO2 -GR (5%) with the same photocatalytic reaction parameters in Fig. S1 (ESI†), in order to prove which one is better. TiO2 -5%GR had a better photoactivity due to unique structural and electronic properties of graphene. The TiO2 -5%GR was used for subsequent experiments and research. Additionally, no NO2 in the outlet gas was detected when we analyzed the gas products using an Ecom-JZKN flue gas analyzer. Ion Chromatography (IC) was used to clarify the composition of reaction products of experiments on the PCO of NO. After the experiment was completely conducted, 1 mL solution was removed from the collected solution (∼4 mL) in a volumetric flak, diluted by a factor of 200. Then the diluted solution was injected into the IC. The result of qualitative analysis, taking TiO2 -5%GR for an instance, is shown in Fig. 1b, which demonstrated that NO3 − existed in the solution. In addition, standard solutions were prepared and the corresponding IC was conducted, of which the fitting line is displayed in Fig. S2 (ESI†). Therefore, the concentration of NO3 − could be acquired using an external standard method, which is 17.56 mg L−1 . In order to confirm NO3 − was the main reaction product for removal of NO with the PCO behavior, the material balances for NO were carried out, just as Ding et al. reported [24], and the detailed calculate process was given in the supporting information. The results of IC analysis and material balance calculation showed that NO3 − was the main reaction product in the PCO of the NO process, which also had been demonstrated in previous report [25].
3.2. Structure and morphology characterization The XRD patterns of TiO2 -GR with different weight ratios are shown in Fig. 2a. It can be observed that the crystal structure of the samples obtained by varying the content of graphene can be indexed as anatase TiO2 (JCPDS No. 21-1272), which suggested the GR had little effect on crystal phase of TiO2 . Notably, the typical diffraction peak of the GR at 24.4◦ was not observed in the TiO2 GR catalyst, probably because the peak was shielded by the main peak of anatase TiO2 at 25.3◦ [26]. Moreover, the broad diffraction peaks of TiO2 -GR suggested a smaller crystallite size. It is widely accepted that a smaller particle size contributes to improvement of the catalytic performance [27]. The morphology and particle sizes of samples were visually investigated by TEM. As shown in Fig. 2b, it displays clearly the two dimensional structure of GR sheets with micrometers long wrinkles. The SAED pattern inset in Fig. 2b shows the perfect hexagonal arrangement of the C atoms in GR. Compared with TiO2 with an average size of 25 nm, it could be seen that the TiO2 size dramatically decreased with the addition of GO. The smaller particle diameter with an average size of 10 nm (Fig. 2c and d) was consistent with the XRD pattern. On the other hand, compared to the pure TiO2 particles, the TiO2 displayed a superduper dispersibility over the GR (Fig. 2d). The above results showed that the GR could be as dispersant to avoid agglomeration of TiO2 particles, resulting in improvement of the specific area. The increased surface area could provide more active sites and promote the separation of e− /h+ pairs during the photocatalytic reaction. The HRTEM in Fig. 2e shows lattice spacing of 0.34 nm corresponding to the (110) lattice planes of TiO2 . While a narrower spacing between the lattice planes is observed as about 0.32 nm from Fig. 2f for TiO2 -GR. As is reported, the atomic radius of C about 0.09 nm is smaller than the ionic radius of Ti4+ (0.06 nm) and O2− (0.14 nm). Thus, the lattice spacing becomes narrower when one ion is superseded by an element with smaller radius, which directly indicated the GR had been injected into the matrix of TiO2 , thus the interfacial interaction between GR and TiO2 was observed [28]. Fig. 3 displays the Raman spectroscopy of samples. For the TiO2 , five fundamental peaks located at 143, 196, 396, 516 and 640 cm−1 can be attributed to the vibration modes of anatase phase (Fig. 3a) [29], which was consistent with the XRD results. The Raman spectra of GO, GR and TiO2 -GR samples exhibited two peaks, denoted as the D peak (around 1340 cm−1 ) and the G peak (around 1585 cm−1 ). The D peak indicates the presence of defects, and the G peak indicates the sp2 -bonded carbon atoms in a twodimensional hexagonal graphitic layer [30]. The ID /IG intensity ratio usually is a measure of disorder degree and average size of the sp2 domains in graphite materials [31]. Compared with GO (ID /IG = 1.80), the increased ID /IG ratios were 1.85 and 1.88 for GR
342
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 2. (a) The XRD patterns, TEM images of (b) GR, (c) TiO2 , (d) TiO2 -GR, HRTEM images of (e)TiO2 , (f) TiO2 -GR.
and TiO2 -GR, respectively. Thus, this result suggested that the GO had been reduced and more defects had been formed [32]. In addition, the sharp peak of TiO2 -GR had a slightly blue-shifted from 143 cm−1 to 146 cm−1 (Fig. 3b). The blue shift can be contributed to a strong chemical interaction between GR and TiO2 , which was similar with the results of previous reported [33,34]. In fact, the surface Ti O C structure formed on TiO2 inevitably impairs the symmetry of Ti-O-Ti and thus changes the wavenumber of corresponded active Raman modes. Furthermore, it was also reported that oxygen vacancies, which could improve the photocatalytic activity, could lead to a considerable blue shift in the Raman band. For the formation of oxygen vacancies over TiO2 , it was accepted that oxygen vacancies could be generated by doping with lower valence ions[35]. However, in this investigation the valence state of C in Ti O C is equal to that of Ti, thus, the oxygen vacancies might be no increase over TiO2 -GR catalyst. Therefore, the improvement of photocatalytic activity of TiO2 -GR catalyst could be attributed to other mechanism rather than oxygen vacancies.
3.3. Optical absorption properties and interfacial interactions analysis The optical properties of the samples were probed with UV–vis diffuse reflection spectra (DRS), as shown in Fig. 4. Compared with the TiO2 , TiO2 -GR displayed an improved light absorption in both UV and visible light region, which could be ascribed to the reintroduction of the black-body properties of graphite-like materials [36]. Meanwhile, a red shift in the absorption edge of the TiO2 -GR catalyst was observed, indicating that the TiO2 and GR interacted with each other and this interaction could improve the light absorption and broaden the absorption range, which was important to the photocatalytic activity. The UV–vis spectra of the samples after Kubelka-Munk treatment (i.e., relationship of [␣h]1/2 versus photon energy) are shown in Fig. 4b. The roughly estimated bandgap value of TiO2 was 3.17 eV, which was close to the theoretical value (3.2 eV). While the band gap of TiO2 -GR was 3.01 eV. The obvious variation (0.16 eV) of the Eg value can be attributed to the chem-
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 3. Raman spectra of the GO, GR, TiO2 and TiO2 -GR.
343
GR, (2) the more chemisorbed oxygen decreased the density of the electronic cloud of TiO2 over TiO2 -GR. Interestingly, the binding energy of O␣ for TiO2 -GR was also significantly different from the one of TiO2 and GR. This might be attributed to that: (1) from the aspect of adsorption site for O2 : considering of the low content of GR and the low content of O␣ over TiO2 -GR catalyst (Fig. 5b), it can be confirmed that the O␣ was from TiO2 . (2) from the aspect of O␣ content, the O␣ content of TiO2 -GR was much higher than that of TiO2 . It can be attributed to the TiO2 with better dispersibility and smaller particle size over the TiO2 -GR that can improve the O2 adsorption. (3) from the aspect of binding energy: the binding energy of O␣ for TiO2 -GR was highest. This observation can be attributed to two possible reasons. The one was that GR with higher conduction decreased the electronic cloud O␣ over the TiO2 -GR catalyst. However, considering of the higher electronegativity of O2 , this reason was nearly impossible. The other reason was that the interaction between TiO2 and GR increased the reduction potential of CB to decrease the density of the electronic cloud of chemisorbed oxygen under the condition of XPS test. The more negative reduction potential can enhance the photocatalytic reactions Eqs. (8) and (9) to improve the NO oxidation. 3.4. Surface redox potential: Mott-Schottky
Fig. 4. (a) UV–vis diffuse reflection spectra patterns, (b) plots of (ahv)1/2 versus hv of the as-synthesized TiO2 and TiO2 −GR.
ical bonding between TiO2 and the specific sites of carbon, which meant the formation of Ti O C bond [37].Considering of the superior electron conductivity of the GR, thus, the charge transportation and separation via the chemical bonding between TiO2 and GR could enhance phtotcatalytic activity. The faster charge transportation and separation for the TiO2 -GR will be confirmed in Section 3.4. Fig. 5a shows XPS spectra for Ti2p of TiO2 and TiO2 -GR catalysts. The two peaks of TiO2 were observed at 464.1 and 458.5 eV, assigned to Ti4+ 2P1/2 and Ti4+ 2P3/2 , respectively. It is worth noting that, for the TiO2 -GR catalyst, a clear increase in binding energies was observed to shift to 465.1 eV and 459.5 eV. It indicated that the Ti of TiO2 -GR had a lower density of the electron cloud than that of TiO2 . Fig. 5b shows XPS spectra for O1s of GR, TiO2 and TiO2 -GR catalysts. Three kinds of surface oxygen species could be distinguished in the O1s spectra. The lower binding energy of 529.0–531.0 eV could be ascribed to the lattice oxygen (O) [38]. Two peaks at the higher binding energy side were observed due to chemisorbed oxygen (O␣), and surface oxygen by hydroxyl species and/or adsorbed water species (Oc ) [39]. Compared with that of pure TiO2 , the binding energy of O for TiO2 -GR was much higher, which was consistent with the observation of XPS spectra of Ti4+ 2p. Similarly, it could be suggested that the density of the electronic cloud for TiO2 species decreases by GR. This result might be attributed to that: (1) the electron of TiO2 can be conducted by
The surface redox reaction is a very important factor during the photocatalytic activity. While the reduction and oxidation abilities of e− and h+ are strongly depending on positions of CB and VB. In consideration of the GR changing the potential of CB and VB, thus, it is important to verify whether a shift of the CB or VB. The CB position could be calculated from the Mott-Schottky (MS) by using the impedance-potential method [40]. As displayed in Fig. 6, both TiO2 and TiO2 -GR photoelectrodes exhibited positive slopes, suggesting that the two samples were n-type semiconductors. In addition, with the plots being extrapolated to 1/C2 = 0, the values of Fermi Level (EF ) were estimated at −0.39 and −0.42 V vs. SCE (equivalent to −0.15 and −0.18 eV vs. NHE), respectively. It is accepted that the negative shift of EF in TiO2 -GR can increase the transfer of photogenerated charge carriers [41], which is agreement well with the observations of Section 3.3. It is generally known that the CB potential (ECB ) of n-type semiconductor is very close to (0–0.2 eV more negative) the EF [42]. Here, the voltage difference between the CB and the flat potential was set to 0.1 eV. Hence, the ECB of the TiO2 and TiO2 -GR were −0.25 and −0.28 eV vs. NHE, respectively. The decreasing CB resulted in a stronger reductive power of the photogenerated e− and enhanced the reaction of CB (Eqs. (8)–(9)). Accordingly, that improved PCO of NO. The band gap energies estimated from the intercept of the tangents to the plots of (ahn)2 vs. photon energy were 3.17 eV for TiO2 , 3.01 eV for TiO2 -GR, respectively. Both of VB positions were estimated by the following relationship: EVB = ECB + Eg , they were 2.92 and 2.73 eV. The VB maximum of TiO2 -GR was lower by 0.19 eV than that of TiO2 . As is known, the higher potential of VB is good for the PCO of pollutants. On the contrary, the PCO of TiO2 -GR was better than that of TiO2 , which might be ascribed to the more negative reduction potential of CB, thus, leading to a much stronger reductive power of the photogenerated e− . In addition, the high conductivity of GR reduced the recombination of e− /h+ pairs, so as to make the quantity of h+ dominant, which would be confirmed in the next section. Therefore, the slight decrease of VB potential had no great influence on photocatalytic reaction. 3.5. Efficient charge separation and transportation The electrochemical impedance spectroscopy (EIS) as a common electrochemical method has been widely used in evaluating the charge transfer efficiency. Fig. 7a presents Nyquist plots for TiO2
344
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 5. XPS spectrum for TiO2 and TiO2 -5%GR catalysts.
Fig. 6. Mott-Schottky plots of TiO2 and TiO2 -5%GR.
TiO2 , as shown in Fig. 7b, which indicated that the recombination of photogenerated charge carrier was inhibited by GR. Therefore, for the TiO2 -GR catalyst, the properties of longer-lived e− /h+ pairs and high photon efficiency can be improved by GR
3.6. Mechanism analysis
Fig. 7. (a) EIS Nyquist plots of TiO2 and TiO2 -GR under UV irradiation, (b) PL spectra of TiO2 and TiO2 -GR catalysts.
and TiO2 -GR-based electrodes under UV light. It is well-known that the smaller arc radius of Nyquist plot, the faster charge transfer of the test material [43]. As shown in Fig. 7a, the diameter of the arc radius of TiO2 -GR was smaller than that of TiO2 . It indicated that GR significantly enhanced the electron transfer, which could reduce the recombination of e− /h+ pairs. Photoluminescence (PL) spectra are used to characterize recombination processes and the rate of e− /h+ pairs in semiconductor particles [44]. In general, a lower recombination rate of e− /h+ pairs or a higher transfer of e− can result in a lower PL intensity. It is noted that the PL spectrum of the TiO2 -GR exhibited lower intensity in comparison with that of the
Additionally, we carried out a trapping experiment to intuitively understand the specific reactive species during the PCO of the NO process under UV irradiation over the TiO2 -5%GR photocatalysts. Five quartz tubes filled with 30 mL deionized water, 3 mL 30% H2 O2 solution was divided into above five groups (a–e). Furthermore, 15 mg TiO2 -5%GR was put into groups (a-d), respectively. Subsequently, several scavengers, isopropyl alcohol (IPA), ammonium oxalate (AO), and p-benzoquinone (BQ), which are known as effective • OH, h+ , and • O2 − scavengers, respectively, were added into group (b–d). Then NO was introduced, and all quartz tubes were exposed to UV irradiation under stirring conditions for 20 min. By centrifugation, the obtained solution was injected into IC. In addition, standard solutions were prepared and the corresponding IC was conducted, and the fitting line is displayed in Fig. S2 (ESI†). Therefore, the concentration of NO3 − could be acquired using an external standard method. Quantitative analysis shows that concentrations of NO3 − of the group (a)–(e) were 19.45, 16.76, 12.44, 8.28 and 3.25 mg/L−1 , respectively. It could be seen that OH, O2 and h+ were the important and efficient active species in the PCO of NO. The experimental result showed that the concentrations of NO3 − decreased in the order OA < BQ < IPA, obviously, indicating that h+ and • O2 − were the main oxidative species in the photocatalytic process.
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 8. EPR spectrum of GR, TiO2 and TiO2 -GR catalysts by 5% O2 adsorption with UV radiation for 2 h.
Fig. 9. Photocatalytic degradation of RhB under different reaction time (a: TiO2 -GR with UV, b: TiO2 -GR with UV and NO).
The generation of • O2 − is confirmed by the EPR technique further. It is clearly observed from Fig. 8 there was no signal observed in the GR. However, • O2 − - species were observed for both TiO2 and TiO2 -GR samples. Also the intensity of the characteristic peaks of • O2 − over TiO2 -GR was higher conspicuously than that of TiO2 , indicating that the ability to generate • O2 − was enhanced after coupling with GR. As is known, there was only one reaction (Eq. (8)) on CB in the presence of O2 . That confirmed the reaction on CB was
345
enhanced, which was consistent with the result of Section 3.4–3.5. Reactions of (Eqs. (8)–(9)) were also enhanced. As such, when e− on the CB was consumed, the recombination of e− and h+ thereby could be avoided to the maximum extent. So more h+ was consumed by H2 O, H2 O2 and other sacrificial reagent effectively (Eqs. (10)–(11)). The e− /h+ pairs were completely isolated in space so that they efficiently participated in the photocatalytic redox reactions, resulting in a higher photocatalytic activity. The position of VB of TiO2 had a negative shift, nevertheless, it is still more positive than that of the most reactant. The h+ possessing strong oxidation ability had a great affinity to capture electrons from reactant, thus, can oxidize pollutants directly. From the thermodynamics point of view, the h+ of VB could oxidize NO because the EVB (about 2.73 eV vs NHE) of TiO2 -GR was more positive than E (HNO3 /NO, 0.94 eV vs NHE). The temporal absorption spectra changes of RhB taking place different reaction time over the TiO2 GR catalyst are shown in Fig. 9. The tendency for the absorption maximum of the solution under different reaction time was similar. With the reaction time increasing, the concentration of RhB was lower. It is interesting to note that the absorption maximum of reaction under H2 O2 + NO was lower than that of the reaction under H2 O2 in every parallel time. Therefore, the NO could improve the RhB oxidation under the low concentration of RhB over the TiO2 -GR catalyst, rather than competing free radicals. The enhanced performance suggested the h+ was consumed by NO, which restrained the recombination of e− /h+ pairs. More electron transferred to the CB, improved the formation rate of • OH and • O2 − (Eqs. (8)–(9)), consequently, which improved the photocatalytic reaction of RhB. Accordingly, it is confirmed that NO could be oxidized in by h+ directly. Based on the above results, the enhanced PCO activity of NO is attributed to the high migration efficiency of photoinduced electrons and the negative shift in the CB of TiO2 -GR. According to investigations above, a reasonable mechanism of the TiO2 -GR photocatalyst is illustrated in Fig. 10 to explain the significantly enhanced PCO under the UV/H2 O2 system. 4. Conclusion In summary, the TiO2 -GR catalysts exhibited higher PCO of NO under UV/H2 O2 . TiO2 -5%GR was considered to be the optimum photocatalyst and its photocatalytic efficiency can reach up to 80.4%. For the TiO2 -GR catalysts, the enhanced PCO ability were attributed to the smaller particle size, the more negative reduction potential of CB and the high electron conductivity attributed to the presence of GR. A trapping experiment was also conducted to examine the effects of the active species involved in the PCO of NO,
Fig. 10. Schematic illustration of possible reaction mechanism of the PCO of NO on the TiO2 -GR.
346
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
illustrating that h+ played the dominant role among active species in the PCO of NO. Further investigations of mechanism showed that the h+ could oxidize NO directly. Acknowledgements This work was financially supported by the Assembly Foundation of The Industry and Information Ministry of the People’s Republic of China 2012 (543), the National Natural Science Foundation of China (51408309 and 51578288), Science and Technology Support Program of Jiangsu Province (BE2014713), Natural Science Foundation of Jiangsu Province (BK20140777), Industry-Academia Cooperation Innovation Fund Projects of Jiangsu Province (BY2014004-10), Science and technology project of Nanjing (201306012), Jiangsu Province Scientific and Technological Achievements into a Special Fund Project (BA2015062), A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education of Jiangsu Higher Education Institutions, Key Project of Chinese National Programs for Research and Development (2016YFC0203800). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molcata.2016.07. 029. References [1] G. Busca, M.A. Larrubia, L. Arrighi, G. Ramis, Catal. Today 107–108 (2005) 139–148. [2] V. Ramanathan, Y. Feng, Atmos. Environ. 43 (2009) 37–50. [3] K.C. Taylor, Catal. Rev. 35 (1993) 457–481. [4] D.A.P. Panagiotis, G. Smirniotis, B.S. Uphade, Angew. Chem. Int. Ed. 40 (2001) 2479–2482. [5] L.L. Guido Busca, Gianguido Ramis, Francesco Berti, Appl. Catal. B 18 (1998) 1–36. [6] G.H. Yao, K.T. Gui, F. Wang, Chem. Eng. Technol. 33 (2010) 1093–1098. [7] D. Alibegic, S. Tsuneda, A. Hirata, Chem. Eng. Sci. 56 (2001) 6195–6203. [8] C.D.C.C.A.C.I.L.P.M.M.C.M.P.d. Fernandez, Journal of Environmental Engineering, 128 (2002) 68–72. [9] Y. Liu, J. Zhang, C. Sheng, Y. Zhang, L. Zhao, Energy Fuels 24 (2010) 4925–4930. [10] A.J.B. Steven, N. Frank, ACS 99 (1977) 303–304.
[11] K.T.N. Negishi, T. Ibusuki, J. Mater. Sci. 33 (1998) 5789–5794. [12] A.K.G.K.S. Novoselov, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666–669. [13] P.B.R.R. Nair, A.N. Grigorenko, K.S. Novoselov, T.J. Booth, T. Stauber, N.M.R. Peres, A.K. Geim, Science 320 (2008) 1308. [14] G. Williams, B. Seger, P.V. Kamat, ACS Nano 2 (2008) 1487–1491. [15] X. Quanjun, Y. Jiaguo, J. Mietek, Nanoscale 3 (2011) 3670–3678. [16] T.L. Yu, B.K. Vijayan, O. Lyandres, K.A. Gray, M.C. Hersam, J. Phys. Chem. Lett. 3 (2012) 1760–1765. [17] J.S. Lee, K.H. You, C.B. Park, Adv. Mater. 24 (2012) 1084–1088. [18] D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L.B. Alemany, W. Lu, J.M. Tour, ACS Nano 4 (2010) 4806–4814. [19] S. Colle, J. Vanderschuren, D. Thomas, Chem. Eng. Sci. 60 (2005) 6472–6479. [20] J.L.D. Paiva, G.C. Kachan, Ind. Eng. Chem. Res. 37 (1998) 609–614. [21] (!!! INVALID CITATION !!!). [22] G. Isabel Xiaoye, T. Wenjie, N. Matthew, J.T. Yates, Science 333 (2011) 736–739. [23] T. Hirakawa, Y. Nosaka, Langmuir 18 (2002) 3247–3254. [24] J. Ding, Q. Zhong, S. Zhang, F. Song, Y. Bu, Chem. Eng. J. 243 (2014) 176–182. [25] H. Nie, O. Man, Z. Qin, S. Zhang, L. Yu, J. Hazard. Mater. 300 (2015) 598–606. [26] Y.J. Xu, Y. Zhuang, X. Fu, J. Phys. Chem. C 114 (2010) 2669–2676. [27] J. Chen, B. Lim, E.P. Lee, Y. Xia, Nano Today 4 (2009) 81–95. [28] M. Shang, W. Wang, L. Zhang, H. Xu, Mater. Chem. Phys. 120 (2010) 155–159. [29] P. Falaras, H.L. Goff, M.C. Bernard, A. Xagas, Sol. Energy Mater. Sol. Cells 64 (2000) 167–184. [30] G.A. Zickler, B. Smarsly, N. Gierlinger, H. Peterlik, O. Paris, Carbon 44 (2006) 3239–3246. [31] A. Omid, ACS Nano 4 (2010) 4174–4180. [32] G. Hui-Lin, W. Xian-Fei, Q. Qing-Yun, W. Feng-Bin, X. Xing-Hua, ACS Nano 3 (2009) 2653–2659. [33] S.D. Perera, R.G. Mariano, K. Vu, N. Nour, O. Seitz, Y. Chabal, K.J. Balkus, ACS Catal. 2 (2012) 949–956. [34] N. Li, G. Liu, C. Zhen, F. Li, L. Zhang, H.M. Cheng, Adv. Funct. Mater. 21 (2011) 1717–1722. [35] S. Zhang, X. Liu, Q. Zhong, Y. Yao, Catal. Commun. 25 (2012). [36] E. Gao, W. Wang, M. Shang, J. Xu, Phys. Chem. Chem. Phys. 13 (2011) 2887–2893. [37] H. Zhang, X. Lv, Y. Li, Y. Wang, J. Li, ACS Nano 4 (2010) 380–386. [38] S. Hamoudi, F.C¸. Larachi, A. Adnot, A. Sayari, J. Catal. 185 (1999) 333–344. [39] J.C. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Phys. Chem. Chem. Phys. 2 (2000) 1319–1324. [40] Y. Meidan, G. Jiaojiao, L. Yuekun, L. Changjian, L. Zhiqun, J. Am. Chem. Soc. 134 (2012) 15720–15723. [41] X. Cheng, H. Liu, Q. Chen, J. Li, W. Pu, Carbon 66 (2014) 450–458. [42] A. Ishikawa, T. Takata, J.N. Kondo, M. Hara, H. Kobayashi, K. Domen, J. Am. Chem. Soc. 124 (2002) 13547–13553. [43] W.H. Leng, Z. Zhang, J.Q. Zhang, C.N. Cao, J. Phys. Chem. B 109 (2005) 15008–15023. [44] H. Yamashita, Y. Ichihashi, S.G. Zhang, Y. Matsumura, Y. Souma, T. Tatsumi, M. Anpo, Appl. Surf. Sci. 121–122 (1997) 305–309.
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata
Photocatalytic oxidation of NO over TiO2 -Graphene catalyst by UV/H2 O2 process and enhanced mechanism analysis Yanan Wang 1 , Shule Zhang ∗,1 , Yiqing Zeng, Man Ou, Qin Zhong ∗ School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, PR China
a r t i c l e
i n f o
Article history: Received 4 April 2016 Received in revised form 14 July 2016 Accepted 15 July 2016 Available online 18 July 2016 Keywords: Photocatalytic oxidation NO TiO2 -GR UV/H2 O2 Mechanism
a b s t r a c t This study aimed at investigating the photocatalytic oxidation (PCO) of NO under the UV/H2 O2 system over a series of TiO2 -Graphene (TiO2 -GR) catalysts, which was able to dramatically improve the PCO efficiency of NO compared with pure TiO2 , and the NO oxidative product was the stable nitrate. The electronic interfacial interaction between GR and TiO2 resulted in a negative shift of the CB of TiO2 evidenced by MS. The excellent conductivity of GR suppressed e− /h+ pairs recombination effectively, and GR as a kind of dispersant reduced the size of TiO2 , increased the active sites. These advantages offered e− and h+ more opportunities to participate in PCO of NO, resulting in significant improvement of the PCO efficiency. The effects of the active species involved in the photocatalytic process were also examined. Moreover, different tests were designed to further confirm the improved mechanism. Further investigations showed that the h+ could oxidize NO directly. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Nitrogen oxides (NOx ) can cause ozone depletion, photochemical smog and the acid rain [1,2]. Over the past few decades, atmospheric NOx concentrations have greatly increased with the development of society and industry [3]. The SCR of NOX is considered as the most effective method for the removal of NOX from stationary sources [4]. However, this reaction must be controlled at a certain temperature (320–400 ◦ C) [5]. That indicates SCR devices must be placed before the air pre-heater, where the high concentration of ash (e.g. K2 O, CaO and As2 O3 ) in the flue gas would reduce the performance and longevity of catalysts. Moreover, this technique using NH3 to reduce NO to N2 is not economical in terms of atoms. Therefore, development of a method, by which NOx is oxidized to nitrate species dissolved in water at a low temperature, overcoming the bottlenecks of the SCR technique, is important [6]. Advanced oxidation processes (AOPs) can produce free radicals with strong oxidation, such as • OH, • O2 − and HO2 • , which can simultaneously oxidize and remove multiple pollutants, such as NOX , H2 S, trace elements, and volatile organic compounds in recent years [7]. The UV/H2 O2 process, as a representative of the AOPs, has been widely studied for the water remediation. Since the process
∗ Corresponding authors. E-mail address: [email protected] (S. Zhang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.molcata.2016.07.029 1381-1169/© 2016 Elsevier B.V. All rights reserved.
can produce more • OH and • O2 − by photolysis of H2 O2 . Cooper et al. [8] first used • OH produced by UV decomposition of H2 O2 to oxidize NO from simulated flue gas. Liu et al. [9] developed a more applicable wet UV/H2 O2 to remove NO from simulated flue gas. The process is a promising way to take place of SCR. The low utilization ration of H2 O2 limited its wide application. Titanium dioxide (TiO2 ) reported first by Frank and Bard [10], has attracted enormous attention because of its high-efficiency, nontoxicity, low cost, and photochemical stability [11]. However, the PCO of TiO2 is low due to the rapid recombining of photogenerated e− /h+ pairs. Graphene discovered first in 2004 [12] has been a rising star on the horizon of materials science because of its superior electron conductivity, unique 2D structure with a high surface area, structural flexibility and chemical stability [13]. These unusual properties make it a suitable candidate to couple with TiO2 to improve the photocatalytic performance. In 2008, Williams et al. firstly reported the fabrication of TiO2 -GR nanocatalyst through in situ induced photocatalytic reduction of GO [14]. TiO2 -GR photocatalysts have attracted a lot of attention for diverse applications, such as H2 production from water splitting [15], hydrocarbon production from CO2 reduction [16], organic photosynthesis [17]. Accordingly, TiO2 -GR photocatalysts have the potential to be applied in photocatalytic oxidation (PCO) of NO in flue gas under the UV/H2 O2 system. However, few reviews focusing on applications and mechanism analysis of TiO2 -GR catalysts to the PCO of NO under UV/H2 O2 are reported.
340
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
In this study, we investigated the PCO of NO over a series of TiO2 GR catalysts under the UV/H2 O2 system. Compared with TiO2 , the TiO2 -GR catalysts exhibited higher PCO efficiency of NO under the UV/H2 O2 system. The reasons contributing to the high PCO of NO were discussed and corresponding mechanism was further analyzed. Moreover, a trapping experiment was conducted to examine the effects of the active species involved in the PCO of NO. This study could provide a new insight into the interaction between TiO2 and GR, as well as the mechanism on PCO of NO under UV/H2 O2 . 2. Experimental 2.1. Preparation of materials 2.1.1. Synthesis of TiO2 A sol-gel method was used to prepare TiO2 support. 0.064 mol of TBOT was added to 0.128 mol of acetylacetone under continuous stirring. Then the solution was diluted with 50 mL ethanol. After being stirred for 2 h at room temperature, the solution was heated at 60 ◦ C for 4 h in a water bath and dried at 120 ◦ C for 6 h. Then it was calcined at 450 ◦ C for 3 h. 2.1.2. Synthesis of TiO2 -GR The graphite oxide (GO) used to prepare the TiO2 -GR catalyst was synthesized using the improved GO synthesis method [18]. GO was ultrasonicated in deionized water and anhydrous ethanol solution. Then the TiO2 was added to the GO solution to prepare 2.5, 5, and 7.5 wt% TiO2 -GR catalysts (2.5, 5, and 7.5 wt% represent the mass fraction of GO). Then, a Teflon-sealed autoclave was filled with the homogeneous suspension up to 80% of the total volume, maintained at 150 ◦ C for 24 h, and cooled down to room temperature naturally. The resulting catalysts were washed by water, dried at 60 ◦ C in oven to get the final TiO2 -GR catalyst with different weight addition ratios of GR. These catalysts of TiO2 and GR are denoted as TiO2 -xGR (x = 2.5, 5, 7.5). 2.2. Characterizations The structural and chemical information for prepared samples were measured by X-ray diffraction (XRD, Cu K, Purkinjie XD-3), Raman spectroscopy (DRX, Thermo Fisher, USA), Transmission electron microscope (a Philips CM-10 at 80 kV and a CM-12 at 120 kV), UV–vis diffuse reflectance spectra (DRS, Shimadzu UV-2550), X-ray photoelectron spectroscopy (PHI-5000C ESCA system), Photoluminescence spectra (PL, He-Cd laser, Labram-HR800), Electron paramagnetic resonance (EPR, Bruker EMX-10/12-type spectrometer in the X-band). Produced ions in the solution were analyzed by Ion Chromatography (IC, Dionex ICS90). The temporal absorption spectra change of RhB was characterized by UV–vis Spectrophotometer (T6 New Century, Beijing Persee Co, Ltd.). All photoelectrochemical tests were performed on a CHI660D electrochemical workstation (Chenhua Instrument, Shanghai, China) with a conventional three-electrode cell. Platinum wire and saturated calomel electrode (SCE) were used as counter and reference electrodes, respectively. The working electrode was prepared by dip-coating powders onto fluorine-doped tin oxide (FTO) substrate with a size of 1 cm × 1 cm: 2 mg of catalyst was suspended in 5 mL ethanol to produce a slurry by ultrasonic dispersion for 30 min, which was then coated on a FTO glass electrode. After the films were dried under ambient conditions, and subsequently treated at 180 ◦ C for 2 h. The electrochemical impedance spectra (EIS) were carried out in the frequency ranging from 105 to 0.1 Hz in an aqueous solution of 0.5 M Na2 SO4 . The Mott-Schottky analysis was measured at potential ranging from −0.8 to 0.6 V, an amplitude of 5 × 10−3 V, and frequency of 1 × 103 Hz. Prior to and during all measurements,
the electrolyte (0.1 mM K3 [Fe(CN)6 ] in 1 M KCl) was purged with nitrogen. 2.3. Photocatalytic activity The PCO of TiO2 -GR was evaluated by removing NO in simulated flue gas. These samples were performed in a quartz tubular vessel (d = 1 cm, h = 15 cm) in a fixed-bed with continuous flow reactor at room temperature and atmospheric pressure. A 500 W Hg-lamp equipped with a visible cut off filter to provide ultraviolet was chosen as the ultraviolet light source. The catalyst (0.2 g) was loaded in the reactor parallelled with the Hg-lamp. The reactant gases included 400 ppm NO, 5% O2 and balanced N2 with a total flow rate of 100 mL min−1 . Prior to the light irradiation, adsorptiondesorption equilibrium between NO and photocatalysts had been reached. Meanwhile, 30% H2 O2 solution was injected into the reactor with a flow rate of 0.02 mL min−1 . The gas products (every 12 min reaction) were analyzed by an Ecom-JZKN flue gas analyzer (Germany). The NO conversion was defined as Eq. (1): NO Conversion =
NOinlet − NOoutlet × 100% NOinlet
(1)
3. Results and discussion 3.1. Photocatalytic activities The photocatalysts with different GR mass ratios were applied on the PCO of NO under ultraviolet light illumination to basically evaluate the potential ability for air purification (Fig. 1a). In the absence of UV radiation, the no-catalyst system showed low NO removal efficiency. In such a case, NO could only be removed by single oxidation of H2 O2 , which could be represented by the reaction Eq. (2) [19,20]: 2NO + 3 H2 O2 → 2HNO3 + 2 H2 O
(2)
When the UV light was introduced, the NO removal efficiency improved significantly. It could be attributed to the • OH free radicals produced by photolysis of H2 O2 , and NO could been removed by the oxidation of • OH free radical (3)–(7) [21]: H2 O2 + h → • OH
(3)
• OH
(4)
NO +
→ HNO2
NO + • OH → NO2
(5)
NO2 + • OH → HNO3
(6)
HNO2 + • OH → HNO3
(7)
Under UV light, for the TiO2 catalyst, the NO removal efficiency was higher. It is well known that TiO2 can promote the formation rate of • OH and • O2 − from O2 , H2 O (surface hydroxyl groups) and H2 O2 under the H2 O2 /UV system. This could be represented by reaction Eqs. (8)–(11) [22,23]. The reaction of conduction band is represented by reaction Eqs. (8)–(9). O2 + h + TiO2 → • O2 − H2 O2 + h + TiO2 →
• OH
E = −0.05 eV
E = 0.06 eV
(8) (9)
The reaction of valence band is given by reaction Eqs. (10)–(11). H2 O2 + h + TiO2 → • O2 − OH− + h + TiO2 → • OH
E = 1.00 eV E = 2.72 eV
(10) (11)
It is well known that • O2 − could oxidize NO according to reaction Eqs. (12)–(14) [21]. Therefore, • OH and • O2 − improved the NO
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
341
Fig. 1. (a) PCO efficiency of NO; (b) iron chromatography analysis of iron in solution after 6 h photocatalytic reaction.
removal efficiency by a synergistic effect of catalyst, H2 O2 and UV radiation. NO +
•O − 2
→ NO2
−
(12)
NO + • O2 − → NO2
(13)
NO + • O2 − → NO3 −
(14)
As expected, all of TiO2 -GR composites exhibited higher photocatalytic performance than that of TiO2 . Obviously, the NO removal efficiency of TiO2 -GR catalyst increased and then decreased with increasing of GR, the photocatalytic efficiency order of the various TiO2 -GR samples could be obtained as follows: TiO2 -5%GR (80.4%) > TiO2 -7.5%GR (69.5%) > TiO2 -2.5%GR (65.2%) > TiO2 (58.5%). It suggested that the composition ratio in TiO2 -GR was crucial to obtain an optimal synergistic effect to improve the PCO of NO. Furthermore, the differential value of NO removal efficiency with UV radiation was much higher than the one without UV radiation. Therefore, the GR could improve the synergistic effect of TiO2 -GR catalyst, H2 O2 and UV light. Additionally, a comparative experiment for the PCO of NO was carried out by TiO2 -5%activated charcoal (TiO2 -AC) and TiO2 -GR (5%) with the same photocatalytic reaction parameters in Fig. S1 (ESI†), in order to prove which one is better. TiO2 -5%GR had a better photoactivity due to unique structural and electronic properties of graphene. The TiO2 -5%GR was used for subsequent experiments and research. Additionally, no NO2 in the outlet gas was detected when we analyzed the gas products using an Ecom-JZKN flue gas analyzer. Ion Chromatography (IC) was used to clarify the composition of reaction products of experiments on the PCO of NO. After the experiment was completely conducted, 1 mL solution was removed from the collected solution (∼4 mL) in a volumetric flak, diluted by a factor of 200. Then the diluted solution was injected into the IC. The result of qualitative analysis, taking TiO2 -5%GR for an instance, is shown in Fig. 1b, which demonstrated that NO3 − existed in the solution. In addition, standard solutions were prepared and the corresponding IC was conducted, of which the fitting line is displayed in Fig. S2 (ESI†). Therefore, the concentration of NO3 − could be acquired using an external standard method, which is 17.56 mg L−1 . In order to confirm NO3 − was the main reaction product for removal of NO with the PCO behavior, the material balances for NO were carried out, just as Ding et al. reported [24], and the detailed calculate process was given in the supporting information. The results of IC analysis and material balance calculation showed that NO3 − was the main reaction product in the PCO of the NO process, which also had been demonstrated in previous report [25].
3.2. Structure and morphology characterization The XRD patterns of TiO2 -GR with different weight ratios are shown in Fig. 2a. It can be observed that the crystal structure of the samples obtained by varying the content of graphene can be indexed as anatase TiO2 (JCPDS No. 21-1272), which suggested the GR had little effect on crystal phase of TiO2 . Notably, the typical diffraction peak of the GR at 24.4◦ was not observed in the TiO2 GR catalyst, probably because the peak was shielded by the main peak of anatase TiO2 at 25.3◦ [26]. Moreover, the broad diffraction peaks of TiO2 -GR suggested a smaller crystallite size. It is widely accepted that a smaller particle size contributes to improvement of the catalytic performance [27]. The morphology and particle sizes of samples were visually investigated by TEM. As shown in Fig. 2b, it displays clearly the two dimensional structure of GR sheets with micrometers long wrinkles. The SAED pattern inset in Fig. 2b shows the perfect hexagonal arrangement of the C atoms in GR. Compared with TiO2 with an average size of 25 nm, it could be seen that the TiO2 size dramatically decreased with the addition of GO. The smaller particle diameter with an average size of 10 nm (Fig. 2c and d) was consistent with the XRD pattern. On the other hand, compared to the pure TiO2 particles, the TiO2 displayed a superduper dispersibility over the GR (Fig. 2d). The above results showed that the GR could be as dispersant to avoid agglomeration of TiO2 particles, resulting in improvement of the specific area. The increased surface area could provide more active sites and promote the separation of e− /h+ pairs during the photocatalytic reaction. The HRTEM in Fig. 2e shows lattice spacing of 0.34 nm corresponding to the (110) lattice planes of TiO2 . While a narrower spacing between the lattice planes is observed as about 0.32 nm from Fig. 2f for TiO2 -GR. As is reported, the atomic radius of C about 0.09 nm is smaller than the ionic radius of Ti4+ (0.06 nm) and O2− (0.14 nm). Thus, the lattice spacing becomes narrower when one ion is superseded by an element with smaller radius, which directly indicated the GR had been injected into the matrix of TiO2 , thus the interfacial interaction between GR and TiO2 was observed [28]. Fig. 3 displays the Raman spectroscopy of samples. For the TiO2 , five fundamental peaks located at 143, 196, 396, 516 and 640 cm−1 can be attributed to the vibration modes of anatase phase (Fig. 3a) [29], which was consistent with the XRD results. The Raman spectra of GO, GR and TiO2 -GR samples exhibited two peaks, denoted as the D peak (around 1340 cm−1 ) and the G peak (around 1585 cm−1 ). The D peak indicates the presence of defects, and the G peak indicates the sp2 -bonded carbon atoms in a twodimensional hexagonal graphitic layer [30]. The ID /IG intensity ratio usually is a measure of disorder degree and average size of the sp2 domains in graphite materials [31]. Compared with GO (ID /IG = 1.80), the increased ID /IG ratios were 1.85 and 1.88 for GR
342
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 2. (a) The XRD patterns, TEM images of (b) GR, (c) TiO2 , (d) TiO2 -GR, HRTEM images of (e)TiO2 , (f) TiO2 -GR.
and TiO2 -GR, respectively. Thus, this result suggested that the GO had been reduced and more defects had been formed [32]. In addition, the sharp peak of TiO2 -GR had a slightly blue-shifted from 143 cm−1 to 146 cm−1 (Fig. 3b). The blue shift can be contributed to a strong chemical interaction between GR and TiO2 , which was similar with the results of previous reported [33,34]. In fact, the surface Ti O C structure formed on TiO2 inevitably impairs the symmetry of Ti-O-Ti and thus changes the wavenumber of corresponded active Raman modes. Furthermore, it was also reported that oxygen vacancies, which could improve the photocatalytic activity, could lead to a considerable blue shift in the Raman band. For the formation of oxygen vacancies over TiO2 , it was accepted that oxygen vacancies could be generated by doping with lower valence ions[35]. However, in this investigation the valence state of C in Ti O C is equal to that of Ti, thus, the oxygen vacancies might be no increase over TiO2 -GR catalyst. Therefore, the improvement of photocatalytic activity of TiO2 -GR catalyst could be attributed to other mechanism rather than oxygen vacancies.
3.3. Optical absorption properties and interfacial interactions analysis The optical properties of the samples were probed with UV–vis diffuse reflection spectra (DRS), as shown in Fig. 4. Compared with the TiO2 , TiO2 -GR displayed an improved light absorption in both UV and visible light region, which could be ascribed to the reintroduction of the black-body properties of graphite-like materials [36]. Meanwhile, a red shift in the absorption edge of the TiO2 -GR catalyst was observed, indicating that the TiO2 and GR interacted with each other and this interaction could improve the light absorption and broaden the absorption range, which was important to the photocatalytic activity. The UV–vis spectra of the samples after Kubelka-Munk treatment (i.e., relationship of [␣h]1/2 versus photon energy) are shown in Fig. 4b. The roughly estimated bandgap value of TiO2 was 3.17 eV, which was close to the theoretical value (3.2 eV). While the band gap of TiO2 -GR was 3.01 eV. The obvious variation (0.16 eV) of the Eg value can be attributed to the chem-
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 3. Raman spectra of the GO, GR, TiO2 and TiO2 -GR.
343
GR, (2) the more chemisorbed oxygen decreased the density of the electronic cloud of TiO2 over TiO2 -GR. Interestingly, the binding energy of O␣ for TiO2 -GR was also significantly different from the one of TiO2 and GR. This might be attributed to that: (1) from the aspect of adsorption site for O2 : considering of the low content of GR and the low content of O␣ over TiO2 -GR catalyst (Fig. 5b), it can be confirmed that the O␣ was from TiO2 . (2) from the aspect of O␣ content, the O␣ content of TiO2 -GR was much higher than that of TiO2 . It can be attributed to the TiO2 with better dispersibility and smaller particle size over the TiO2 -GR that can improve the O2 adsorption. (3) from the aspect of binding energy: the binding energy of O␣ for TiO2 -GR was highest. This observation can be attributed to two possible reasons. The one was that GR with higher conduction decreased the electronic cloud O␣ over the TiO2 -GR catalyst. However, considering of the higher electronegativity of O2 , this reason was nearly impossible. The other reason was that the interaction between TiO2 and GR increased the reduction potential of CB to decrease the density of the electronic cloud of chemisorbed oxygen under the condition of XPS test. The more negative reduction potential can enhance the photocatalytic reactions Eqs. (8) and (9) to improve the NO oxidation. 3.4. Surface redox potential: Mott-Schottky
Fig. 4. (a) UV–vis diffuse reflection spectra patterns, (b) plots of (ahv)1/2 versus hv of the as-synthesized TiO2 and TiO2 −GR.
ical bonding between TiO2 and the specific sites of carbon, which meant the formation of Ti O C bond [37].Considering of the superior electron conductivity of the GR, thus, the charge transportation and separation via the chemical bonding between TiO2 and GR could enhance phtotcatalytic activity. The faster charge transportation and separation for the TiO2 -GR will be confirmed in Section 3.4. Fig. 5a shows XPS spectra for Ti2p of TiO2 and TiO2 -GR catalysts. The two peaks of TiO2 were observed at 464.1 and 458.5 eV, assigned to Ti4+ 2P1/2 and Ti4+ 2P3/2 , respectively. It is worth noting that, for the TiO2 -GR catalyst, a clear increase in binding energies was observed to shift to 465.1 eV and 459.5 eV. It indicated that the Ti of TiO2 -GR had a lower density of the electron cloud than that of TiO2 . Fig. 5b shows XPS spectra for O1s of GR, TiO2 and TiO2 -GR catalysts. Three kinds of surface oxygen species could be distinguished in the O1s spectra. The lower binding energy of 529.0–531.0 eV could be ascribed to the lattice oxygen (O) [38]. Two peaks at the higher binding energy side were observed due to chemisorbed oxygen (O␣), and surface oxygen by hydroxyl species and/or adsorbed water species (Oc ) [39]. Compared with that of pure TiO2 , the binding energy of O for TiO2 -GR was much higher, which was consistent with the observation of XPS spectra of Ti4+ 2p. Similarly, it could be suggested that the density of the electronic cloud for TiO2 species decreases by GR. This result might be attributed to that: (1) the electron of TiO2 can be conducted by
The surface redox reaction is a very important factor during the photocatalytic activity. While the reduction and oxidation abilities of e− and h+ are strongly depending on positions of CB and VB. In consideration of the GR changing the potential of CB and VB, thus, it is important to verify whether a shift of the CB or VB. The CB position could be calculated from the Mott-Schottky (MS) by using the impedance-potential method [40]. As displayed in Fig. 6, both TiO2 and TiO2 -GR photoelectrodes exhibited positive slopes, suggesting that the two samples were n-type semiconductors. In addition, with the plots being extrapolated to 1/C2 = 0, the values of Fermi Level (EF ) were estimated at −0.39 and −0.42 V vs. SCE (equivalent to −0.15 and −0.18 eV vs. NHE), respectively. It is accepted that the negative shift of EF in TiO2 -GR can increase the transfer of photogenerated charge carriers [41], which is agreement well with the observations of Section 3.3. It is generally known that the CB potential (ECB ) of n-type semiconductor is very close to (0–0.2 eV more negative) the EF [42]. Here, the voltage difference between the CB and the flat potential was set to 0.1 eV. Hence, the ECB of the TiO2 and TiO2 -GR were −0.25 and −0.28 eV vs. NHE, respectively. The decreasing CB resulted in a stronger reductive power of the photogenerated e− and enhanced the reaction of CB (Eqs. (8)–(9)). Accordingly, that improved PCO of NO. The band gap energies estimated from the intercept of the tangents to the plots of (ahn)2 vs. photon energy were 3.17 eV for TiO2 , 3.01 eV for TiO2 -GR, respectively. Both of VB positions were estimated by the following relationship: EVB = ECB + Eg , they were 2.92 and 2.73 eV. The VB maximum of TiO2 -GR was lower by 0.19 eV than that of TiO2 . As is known, the higher potential of VB is good for the PCO of pollutants. On the contrary, the PCO of TiO2 -GR was better than that of TiO2 , which might be ascribed to the more negative reduction potential of CB, thus, leading to a much stronger reductive power of the photogenerated e− . In addition, the high conductivity of GR reduced the recombination of e− /h+ pairs, so as to make the quantity of h+ dominant, which would be confirmed in the next section. Therefore, the slight decrease of VB potential had no great influence on photocatalytic reaction. 3.5. Efficient charge separation and transportation The electrochemical impedance spectroscopy (EIS) as a common electrochemical method has been widely used in evaluating the charge transfer efficiency. Fig. 7a presents Nyquist plots for TiO2
344
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 5. XPS spectrum for TiO2 and TiO2 -5%GR catalysts.
Fig. 6. Mott-Schottky plots of TiO2 and TiO2 -5%GR.
TiO2 , as shown in Fig. 7b, which indicated that the recombination of photogenerated charge carrier was inhibited by GR. Therefore, for the TiO2 -GR catalyst, the properties of longer-lived e− /h+ pairs and high photon efficiency can be improved by GR
3.6. Mechanism analysis
Fig. 7. (a) EIS Nyquist plots of TiO2 and TiO2 -GR under UV irradiation, (b) PL spectra of TiO2 and TiO2 -GR catalysts.
and TiO2 -GR-based electrodes under UV light. It is well-known that the smaller arc radius of Nyquist plot, the faster charge transfer of the test material [43]. As shown in Fig. 7a, the diameter of the arc radius of TiO2 -GR was smaller than that of TiO2 . It indicated that GR significantly enhanced the electron transfer, which could reduce the recombination of e− /h+ pairs. Photoluminescence (PL) spectra are used to characterize recombination processes and the rate of e− /h+ pairs in semiconductor particles [44]. In general, a lower recombination rate of e− /h+ pairs or a higher transfer of e− can result in a lower PL intensity. It is noted that the PL spectrum of the TiO2 -GR exhibited lower intensity in comparison with that of the
Additionally, we carried out a trapping experiment to intuitively understand the specific reactive species during the PCO of the NO process under UV irradiation over the TiO2 -5%GR photocatalysts. Five quartz tubes filled with 30 mL deionized water, 3 mL 30% H2 O2 solution was divided into above five groups (a–e). Furthermore, 15 mg TiO2 -5%GR was put into groups (a-d), respectively. Subsequently, several scavengers, isopropyl alcohol (IPA), ammonium oxalate (AO), and p-benzoquinone (BQ), which are known as effective • OH, h+ , and • O2 − scavengers, respectively, were added into group (b–d). Then NO was introduced, and all quartz tubes were exposed to UV irradiation under stirring conditions for 20 min. By centrifugation, the obtained solution was injected into IC. In addition, standard solutions were prepared and the corresponding IC was conducted, and the fitting line is displayed in Fig. S2 (ESI†). Therefore, the concentration of NO3 − could be acquired using an external standard method. Quantitative analysis shows that concentrations of NO3 − of the group (a)–(e) were 19.45, 16.76, 12.44, 8.28 and 3.25 mg/L−1 , respectively. It could be seen that OH, O2 and h+ were the important and efficient active species in the PCO of NO. The experimental result showed that the concentrations of NO3 − decreased in the order OA < BQ < IPA, obviously, indicating that h+ and • O2 − were the main oxidative species in the photocatalytic process.
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
Fig. 8. EPR spectrum of GR, TiO2 and TiO2 -GR catalysts by 5% O2 adsorption with UV radiation for 2 h.
Fig. 9. Photocatalytic degradation of RhB under different reaction time (a: TiO2 -GR with UV, b: TiO2 -GR with UV and NO).
The generation of • O2 − is confirmed by the EPR technique further. It is clearly observed from Fig. 8 there was no signal observed in the GR. However, • O2 − - species were observed for both TiO2 and TiO2 -GR samples. Also the intensity of the characteristic peaks of • O2 − over TiO2 -GR was higher conspicuously than that of TiO2 , indicating that the ability to generate • O2 − was enhanced after coupling with GR. As is known, there was only one reaction (Eq. (8)) on CB in the presence of O2 . That confirmed the reaction on CB was
345
enhanced, which was consistent with the result of Section 3.4–3.5. Reactions of (Eqs. (8)–(9)) were also enhanced. As such, when e− on the CB was consumed, the recombination of e− and h+ thereby could be avoided to the maximum extent. So more h+ was consumed by H2 O, H2 O2 and other sacrificial reagent effectively (Eqs. (10)–(11)). The e− /h+ pairs were completely isolated in space so that they efficiently participated in the photocatalytic redox reactions, resulting in a higher photocatalytic activity. The position of VB of TiO2 had a negative shift, nevertheless, it is still more positive than that of the most reactant. The h+ possessing strong oxidation ability had a great affinity to capture electrons from reactant, thus, can oxidize pollutants directly. From the thermodynamics point of view, the h+ of VB could oxidize NO because the EVB (about 2.73 eV vs NHE) of TiO2 -GR was more positive than E (HNO3 /NO, 0.94 eV vs NHE). The temporal absorption spectra changes of RhB taking place different reaction time over the TiO2 GR catalyst are shown in Fig. 9. The tendency for the absorption maximum of the solution under different reaction time was similar. With the reaction time increasing, the concentration of RhB was lower. It is interesting to note that the absorption maximum of reaction under H2 O2 + NO was lower than that of the reaction under H2 O2 in every parallel time. Therefore, the NO could improve the RhB oxidation under the low concentration of RhB over the TiO2 -GR catalyst, rather than competing free radicals. The enhanced performance suggested the h+ was consumed by NO, which restrained the recombination of e− /h+ pairs. More electron transferred to the CB, improved the formation rate of • OH and • O2 − (Eqs. (8)–(9)), consequently, which improved the photocatalytic reaction of RhB. Accordingly, it is confirmed that NO could be oxidized in by h+ directly. Based on the above results, the enhanced PCO activity of NO is attributed to the high migration efficiency of photoinduced electrons and the negative shift in the CB of TiO2 -GR. According to investigations above, a reasonable mechanism of the TiO2 -GR photocatalyst is illustrated in Fig. 10 to explain the significantly enhanced PCO under the UV/H2 O2 system. 4. Conclusion In summary, the TiO2 -GR catalysts exhibited higher PCO of NO under UV/H2 O2 . TiO2 -5%GR was considered to be the optimum photocatalyst and its photocatalytic efficiency can reach up to 80.4%. For the TiO2 -GR catalysts, the enhanced PCO ability were attributed to the smaller particle size, the more negative reduction potential of CB and the high electron conductivity attributed to the presence of GR. A trapping experiment was also conducted to examine the effects of the active species involved in the PCO of NO,
Fig. 10. Schematic illustration of possible reaction mechanism of the PCO of NO on the TiO2 -GR.
346
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 339–346
illustrating that h+ played the dominant role among active species in the PCO of NO. Further investigations of mechanism showed that the h+ could oxidize NO directly. Acknowledgements This work was financially supported by the Assembly Foundation of The Industry and Information Ministry of the People’s Republic of China 2012 (543), the National Natural Science Foundation of China (51408309 and 51578288), Science and Technology Support Program of Jiangsu Province (BE2014713), Natural Science Foundation of Jiangsu Province (BK20140777), Industry-Academia Cooperation Innovation Fund Projects of Jiangsu Province (BY2014004-10), Science and technology project of Nanjing (201306012), Jiangsu Province Scientific and Technological Achievements into a Special Fund Project (BA2015062), A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education of Jiangsu Higher Education Institutions, Key Project of Chinese National Programs for Research and Development (2016YFC0203800). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molcata.2016.07. 029. References [1] G. Busca, M.A. Larrubia, L. Arrighi, G. Ramis, Catal. Today 107–108 (2005) 139–148. [2] V. Ramanathan, Y. Feng, Atmos. Environ. 43 (2009) 37–50. [3] K.C. Taylor, Catal. Rev. 35 (1993) 457–481. [4] D.A.P. Panagiotis, G. Smirniotis, B.S. Uphade, Angew. Chem. Int. Ed. 40 (2001) 2479–2482. [5] L.L. Guido Busca, Gianguido Ramis, Francesco Berti, Appl. Catal. B 18 (1998) 1–36. [6] G.H. Yao, K.T. Gui, F. Wang, Chem. Eng. Technol. 33 (2010) 1093–1098. [7] D. Alibegic, S. Tsuneda, A. Hirata, Chem. Eng. Sci. 56 (2001) 6195–6203. [8] C.D.C.C.A.C.I.L.P.M.M.C.M.P.d. Fernandez, Journal of Environmental Engineering, 128 (2002) 68–72. [9] Y. Liu, J. Zhang, C. Sheng, Y. Zhang, L. Zhao, Energy Fuels 24 (2010) 4925–4930. [10] A.J.B. Steven, N. Frank, ACS 99 (1977) 303–304.
[11] K.T.N. Negishi, T. Ibusuki, J. Mater. Sci. 33 (1998) 5789–5794. [12] A.K.G.K.S. Novoselov, S.V. Morozov, D. Jiang, Y. Zhang, S.V. Dubonos, I.V. Grigorieva, A.A. Firsov, Science 306 (2004) 666–669. [13] P.B.R.R. Nair, A.N. Grigorenko, K.S. Novoselov, T.J. Booth, T. Stauber, N.M.R. Peres, A.K. Geim, Science 320 (2008) 1308. [14] G. Williams, B. Seger, P.V. Kamat, ACS Nano 2 (2008) 1487–1491. [15] X. Quanjun, Y. Jiaguo, J. Mietek, Nanoscale 3 (2011) 3670–3678. [16] T.L. Yu, B.K. Vijayan, O. Lyandres, K.A. Gray, M.C. Hersam, J. Phys. Chem. Lett. 3 (2012) 1760–1765. [17] J.S. Lee, K.H. You, C.B. Park, Adv. Mater. 24 (2012) 1084–1088. [18] D.C. Marcano, D.V. Kosynkin, J.M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L.B. Alemany, W. Lu, J.M. Tour, ACS Nano 4 (2010) 4806–4814. [19] S. Colle, J. Vanderschuren, D. Thomas, Chem. Eng. Sci. 60 (2005) 6472–6479. [20] J.L.D. Paiva, G.C. Kachan, Ind. Eng. Chem. Res. 37 (1998) 609–614. [21] (!!! INVALID CITATION !!!). [22] G. Isabel Xiaoye, T. Wenjie, N. Matthew, J.T. Yates, Science 333 (2011) 736–739. [23] T. Hirakawa, Y. Nosaka, Langmuir 18 (2002) 3247–3254. [24] J. Ding, Q. Zhong, S. Zhang, F. Song, Y. Bu, Chem. Eng. J. 243 (2014) 176–182. [25] H. Nie, O. Man, Z. Qin, S. Zhang, L. Yu, J. Hazard. Mater. 300 (2015) 598–606. [26] Y.J. Xu, Y. Zhuang, X. Fu, J. Phys. Chem. C 114 (2010) 2669–2676. [27] J. Chen, B. Lim, E.P. Lee, Y. Xia, Nano Today 4 (2009) 81–95. [28] M. Shang, W. Wang, L. Zhang, H. Xu, Mater. Chem. Phys. 120 (2010) 155–159. [29] P. Falaras, H.L. Goff, M.C. Bernard, A. Xagas, Sol. Energy Mater. Sol. Cells 64 (2000) 167–184. [30] G.A. Zickler, B. Smarsly, N. Gierlinger, H. Peterlik, O. Paris, Carbon 44 (2006) 3239–3246. [31] A. Omid, ACS Nano 4 (2010) 4174–4180. [32] G. Hui-Lin, W. Xian-Fei, Q. Qing-Yun, W. Feng-Bin, X. Xing-Hua, ACS Nano 3 (2009) 2653–2659. [33] S.D. Perera, R.G. Mariano, K. Vu, N. Nour, O. Seitz, Y. Chabal, K.J. Balkus, ACS Catal. 2 (2012) 949–956. [34] N. Li, G. Liu, C. Zhen, F. Li, L. Zhang, H.M. Cheng, Adv. Funct. Mater. 21 (2011) 1717–1722. [35] S. Zhang, X. Liu, Q. Zhong, Y. Yao, Catal. Commun. 25 (2012). [36] E. Gao, W. Wang, M. Shang, J. Xu, Phys. Chem. Chem. Phys. 13 (2011) 2887–2893. [37] H. Zhang, X. Lv, Y. Li, Y. Wang, J. Li, ACS Nano 4 (2010) 380–386. [38] S. Hamoudi, F.C¸. Larachi, A. Adnot, A. Sayari, J. Catal. 185 (1999) 333–344. [39] J.C. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Phys. Chem. Chem. Phys. 2 (2000) 1319–1324. [40] Y. Meidan, G. Jiaojiao, L. Yuekun, L. Changjian, L. Zhiqun, J. Am. Chem. Soc. 134 (2012) 15720–15723. [41] X. Cheng, H. Liu, Q. Chen, J. Li, W. Pu, Carbon 66 (2014) 450–458. [42] A. Ishikawa, T. Takata, J.N. Kondo, M. Hara, H. Kobayashi, K. Domen, J. Am. Chem. Soc. 124 (2002) 13547–13553. [43] W.H. Leng, Z. Zhang, J.Q. Zhang, C.N. Cao, J. Phys. Chem. B 109 (2005) 15008–15023. [44] H. Yamashita, Y. Ichihashi, S.G. Zhang, Y. Matsumura, Y. Souma, T. Tatsumi, M. Anpo, Appl. Surf. Sci. 121–122 (1997) 305–309.