Constructing g-C₃N₄/SnO₂ S-scheme heterojunctions for efficient photocatalytic NO removal and low NO₂ generation

er g -H d, yen Ch vers , 32 Ho Chi Minh City University of Technology (HUTECH), 475A Dien Bien Phu Street, Binh Thanh District, Ho Chi Minh City, 700000, Viet Nam high efficiency [6,7]. However, the photocatalytic process generates cause human diseases such as airway inflammation, pulmonary edema, or even death [8,9]. Therefore, to apply photocatalytic technology in the experimental condition, a desirable photocatalyst NO2, (ii) be stable r a wide range of paration method. nging to realize a e of poor perfor- O2, and lack of an 12]. llent candidate for erformance in NO n large scales, effi- ostability [13e15]. However, its wide bandgap is a significant barrier to the widespread while, current studies have reported graphitic carbon nitride (g- C3N4) as an excellent supporter of visible-light-driven photocatalysis [17,18]. The advantages of g-C3N4 come from its strong covalent CeN bond, large specific surface area, sustainability, and straightforward fabrication. Unfortunately, g-C3N4 generates NO2 during the NO removal under the light because of its poor photo-induced oxidiz- ability [15,19]. Thus, the combination of SnO2 and g-C3N4 could * Corresponding author. E-mail address: cmthi@hutech.edu.vn (C.M. Thi). Contents lists available at ScienceDirect Journal of Science: Advanc .e l Journal of Science: Advanced Materials and Devices 6 (2021) 551e559Peer review under responsibility of Vietnam National University, Hanoi.or converts NO to NO2, which is much more toxic than NO and can use of SnO2 for visible-light-driven photocatalysis [12,16]. Mean-1. Introduction Nitric oxide (NO), resulting from fossil fuel combustion-based technologies, has become a severe problem over the last few years. This gas causes acid rain, photochemical smog, and plays a key role in inducing particulate pollution [1e3]. Currently, comprehensive approaches for removing NO pollution include a wide range of chemical, biological, electrochemical, photochemical, and photocatalytic technologies [4,5]. Among these approaches, photocatalysis is highly preferred because of its low energy con- sumption, ease of synthesis and operating conditions, low cost, and for NO removal should: (i) generate very little upon light irradiation, (iii) be able to work ove light, and (iv) have a scalable and simple pre Despite significant progress, it remains challe photocatalyst that meets these criteria becaus mance under visible light, a high generation of N appropriate design for scalable production [10e Among the semiconductors, SnO2 is an exce removing NO by photocatalysis due to its high p degradation, many active sites, quick synthesis o cient reduction of NO2 generation, and photArticle history: Received 2 May 2021 Received in revised form 13 July 2021 Accepted 20 July 2021 Available online 28 July 2021 Keywords: S-Scheme photocatalyst g-C3N4/SnO2 NO removal NO2 generation Visible-light-driven photocatalysishttps://doi.org/10.1016/j.jsamd.2021.07.005 2468-2179/© 2021 The Authors. Publishing services b ( facile fabrication of g-C3N4/SnO2 S-scheme heterojunctions for photocatalytic removal of NO under visible light is reported. Optical and electrochemical investigations indicate the formation of these heterojunctions that enable bending at the interface of g-C3N4 and SnO2 and give rise to an efficient separation. A high photocatalytic 500-ppb NO removal performance of 35% and low NO2 generation of 2% are realized after 30 min of visible light irradiation upon the g-C3N4/SnO2 heterojunction with 30% of g- C3N4 addition. In contrast, the bare g-C3N4 extensively produces NO2 greater than 12% compared to 30% from the g-C3N4 sample. This study also shows that the g-C3N4/SnO2 heterojunction is a stable catalyst system and superoxide radicals play a crucial role in the photocatalytic NO removal. Since the prepa- ration of the g-C3N4/SnO2 heterojunction reported in this work is straightforward, it can potentially enable the preparation of highly robust visible-light-driven photocatalysts to remove NO pollution. © 2021 The Authors. Publishing services by Elsevier B.V. on behalf of Vietnam National University, Hanoi. This is an open access article under the CC BY license ( r t i c l e i n f o a b s t r a c tOriginal Article Constructing g-C3N4/SnO2 S-scheme het photocatalytic NO removal and low NO2 Pham Van Viet a, b, Hoang-Phuong Nguyen c, Hong Le Viet Hai a, b, Minh-Thuan Pham d, Sheng-Jie You a Faculty of Materials Science and Technology, University of Science, VNU-HCM, 227 Ngu b Vietnam National University-Ho Chi Minh City, Linh Trung Ward, Thu Duc District, Ho c Faculty of Environment e Natural Resources and Climate Change, Ho Chi Minh City Uni Ward, Tan Phu District, Ho Chi Minh City, 700000, Viet Nam d Department of Environmental Engineering, Chung Yuan Christian University, Taoyuan e journal homepage: wwwy Elsevier B.V. on behalf of Vietnamojunctions for efficient eneration uy Tran a, b, Dai-Phat Bui a, b, Cao Minh Thi e, * Van Cu Street, District 5, Ho Chi Minh City, 700000, Viet Nam i Minh City, 700000, Viet Nam ity of Food Industry (HUFI), 140 Le Trong Tan Street, Tay Thanh 023, Taiwan ed Materials and Devices sevier .com/locate/ jsamdNational University, Hanoi. This is an open access article under the CC BY license an aextend the photoactive range of both materials and improve the lifetime of photogenerated electronehole pairs [20]. Given the characteristic band alignment of SnO2 and g-C3N4, it is generally accepted that SnO2 is an oxidation reaction-based photocatalyst, whereas g-C3N4 is a reduction reaction-based photocatalyst [18,21]. It is thus intriguing to form an S-scheme interface between SnO2 and Scheme 1. Schematic illustration of SnO2 NPs by a hydrothermal process, g-C3N4 by individual materials. P. Van Viet, H.-P. Nguyen, H.-H. Tran et al.g-C3N4, which is presently an outstanding approach to the con- struction of heterojunction photocatalysts [21,22]. This is especially true since the synthesis of SnO2 nanomaterials leads to the formation of oxygen vacancy sites in the structure such that the band gap of SnO2 decreases while active sites for the chemisorption of single NO2 molecules are created on the SnO2 surface [23]. This property im- proves the photoactivity of a g-C3N4/SnO2 S-scheme heterojunction and helps decrease NO2 emission after the reaction process. Despite this great potential, a study on heterojunctions comprised of SnO2 and g-C3N4 for degrading NO with low NO2 generation under visible light has not been thoroughly reported. Therefore, this work reports the fabrication of a heterojunction between SnO2 nanoparticles (NPs) and g-C3N4 by physical mixing and demonstrates promising photocatalytic removal of NO under visible light. In addition, the interfacial charge transfer between g- C3N4 and SnO2 in the S-scheme during photocatalytic NO removal is determined by experimental evidence that includes electrochemical analysis and trapping test results during scavenger presence in the photocatalytic reaction. Furthermore, the mass production of SnO2 NPs via a hydrothermal process and the abundance of g-C3N4 could enable scalable manufacturing of high performance photocatalysts that solve the pressing problem of NO pollution. 2. Experimental 2.1. Preparation of g-C3N4/SnO2 heterojunction The preparation of the g-C3N4/SnO2 heterojunctions begins by a mixing process that is illustrated in Scheme 1. SnO2 NPs were prepared using a simple hydrothermal process according to the 552protocol in our previous studies [12,24]. g-C3N4 was prepared by annealing melamine at 500
C for 2 h. To prepare the g-C3N4/SnO2 heterojunctions, a total of 0.2 g of material was prepared by varying the amount of g-C3N4. Hereafter, the g-C3N4/SnO2 heterojunction will be referred to as x CNS in figures with x indicating the per- centage of g-C3N4. The powder mixture was then dispersed in nnealing step, and the SnO2/g-C3N4 heterojunction via a mixing process of both the Journal of Science: Advanced Materials and Devices 6 (2021) 551e55960 mL of ethanol solvent for 30 minwith vigorous stirring followed by 30 min of sonication. Next, the mixture was further stirred at a temperature of 60
C for 30 min. The well-mixed solution is then heated at 60
C in an oven to dry out the ethanol. The resulting powder was dried for another 20 min at 100
C to obtain the g- C3N4/SnO2 composites. 2.2. Characterizations The phase components of the materials were determined by X- ray diffraction (XRD) patterns on an Advance Bruker X-ray diffractometer with Cu Ka radiation (l ¼ 1.5406 Å, 40 kV, 40 mA) and a step size of 0.02
in the range of 10
e80
. The structural vi- brations of the materials were observed by Fourier-transform infrared spectroscopy (FTIR) where the powder samples were mixed with KBr followed by a moulding process to obtain a round pill shape (99wt% of KBr). Thematerials' morphologywas observed via transmission electron microscopy images taken on a JEM 2100, JEOL, Japan. The optical properties of the materials were deter- mined by diffuse reflectance spectroscopy (DRS) using a JASCO-V- 670 spectrometer equipped with an integrated sphere. Photo- luminescence spectra (PL) were measured using a fluorescence spectrometer (Horiba Jobin-Yvon Nanolog, excitation source wavelength of 325 nm). 2.3. Photoelectrochemical measurement MotteSchottky plots were recorded with an electrochemical workstation (BioLogic SP-240). A semiconductor film, a Pt wire, and a Pt wire in Ag/AgCl (3 M KCl, 0.21 V vs. NHE) were used as the P. Van Viet, H.-P. Nguyen, H.-H. Tran et al. Journal of Science: Advanced Materials and Devices 6 (2021) 551e559action chamber for 1 h to reach adsorption/desorption equilibrium in the dark. Finally, the xenon lamp was turned on to start the photocatalytic reaction. The NO and NO2 concentrations were digitally monitored every minute by a chemiluminescence NOx analyser (Sabio 6042) with a sampling rate of 0.6 L min1. The conversion (e) of NO to NO2 on the materials was calculated by Eq. (1) below: εNO to NO2 ¼ CNO2 CNO (1) where CNO2 (ppb) is the concentration of NO2, CNO (ppb) is the concentration of NO at time t (min). The photocatalytic reaction kinetics were quantified by the LangmuireHinshelwood (L-H) model. The L-H equation is given by Eq. (2), which shows the relationship between the rate constant k (min1), time t, concentration Ct, and initial concentration Co as [25]: ln
C C0  ¼  kt (2) The efficiency of the photons in the photocatalytic reactions was evaluated by calculating the apparent quantum efficiency (AQE, %) as shown in Eq. (3) [26,27]: fapp¼ NA ðt 0 ðCo  CtÞVt PhotonFlux IrradiationArea t  1000M  100% (3) where NA (mol1) is Avogadro's number, (C0) is the initial con- centration of NO, (Ct) is the concentration of NO at time t (min), Vt (L min1) is the flow rate of NO, and M (g mol1) is the molar mass of NO. The photocatalytic test shows that the photon flux was 2.72 1019 cm2 min1 and the irradiation areawas about 78.5 cm2 for our 12-cm diameter Petri dish. After the first photocatalytic run, the Petri dish containing the photocatalyst was dispersed in DI, followed by sonication, and dried at 60
C until the water vanished. Cycling runs for the pho- tocatalytic activity of NO over the sample are repeatedly performed under the same conditions. The reaction mechanisms of the photocatalysts were investi- gated by the trapping experiment. In this test, the photocatalystworking electrode, counter electrode, and reference electrode, respectively. A solar simulator (ABET Low Cost) was used as the light source for the test. The working electrode was prepared as follows: first, 10 mg of sample was dispersed in 10 mL of 10% polyvinyl alcohol solution; then, the mixture was sonicated before coating on an ITO substrate with a size of 1 cm  1 cm; finally, the substrate was kept at room temperature for 12 h. 2.4. Photocatalytic activity test The photocatalytic performance of the materials was evaluated for photocatalytic NO degradation under visible light. First, 0.2 g of the photocatalyst was dispersed in 25 mL of DI with sonication assistance. Second, this suspension was coated on a 12-cm diam- eter Petri dish, followed by a drying step at 60
C for 24 h. The photocatalyst-coated Petri dish was then placed in a photocatalytic chamber (30 cm length 15 cmwide 10 cm height), containing a glass window and a xenon lamp (OSRAM, 300W) 15 cm above. The mass flow controller (Environics Series 4000) mixed 15 mL of NO 100 ppb and 2985 mL of zero air (Sabio 1001 Generator). The airstream was flowed through the surface of the sample in the re-was mixed with 1 wt% of one of the scavengers (K2Cr2O7, KI, 553benzoquinone (BQ), or terephthalic) to inactivate the e , hþ, O2 , or OH, respectively. Finally, the test was prepared with the same protocol as the photocatalytic test. 3. Results and discussion 3.1. Structural properties of as-prepared materials Structural information of the SnO2 NPs, g-C3N4, and 30% g-C3N4/ SnO2 heterojunction by XRD and FTIR are shown in Fig. 1. The phase components of the materials were determined by XRD patterns (Fig. 1a). Typical diffraction peaks of SnO2 are observed with high crystallinity at 26.7, 34.1, 38.0, 52.1, and 65.2
, which are attributed to the (110), (101), (200), (211), and (112) planes, respectively [JCPDS 41e1445]. Typical diffraction peaks of g-C3N4 are observed at 13.1 and 27.7
, which are attributed to the (100) and (002) planes of the tri-s-triazine and the characteristic stacking, respectively [JCPDS 87e1526]. These typical diffraction peaks of SnO2 and g- C3N4 also present in the XRD pattern of the 30% g-C3N4/SnO2 het- erojunction. However, the diffraction peak at 2q ¼ 13.1
has a much weaker intensity than the peak at 2q ¼ 27.7
. Therefore, when combination g-C3N4 with 30 wt% in the CNS composite, the diffraction peak at 2q ¼ 13.1
barely appeared. Noticeably, an apparent expansion is observed for the typical peak at 26.7
in the XRD pattern of the heterojunction because of two strong diffrac- tions from (002) g-C3N4 and (110) SnO2 (see the red box for clarity). Furthermore, vibration signals of both SnO2 NPs and g-C3N4 from FTIR profiles support the formation of the proposed heterojunction (Fig. 1b). In particular, the FTIR profile of g-C3N4/SnO2 presents peaks at 621 cm1 that is attributed to SneO bonds from SnO2 and the well-known band cantered around 3000e3600 cm1 is attributed to the stretching vibrations of hydroxyl groups [28]. Peaks at 810, 1242e1576, 1638, and 3000 cm1 are attributed to the characteristic vibrations of triazine and ring structures from g-C3N4 [29]. In addi- tion, the macroscopic colour of the samples (see inserted digital photos) indicates a structural change in the material. 3.2. Morphology observation The morphology of the bare materials and the heterojunction are observed using HRTEM images (Fig. 2). The as-prepared SnO2 NPs have a spherical shape with an average diameter of about 2 nm (Fig. 2a). Meanwhile, g-C3N4 presents a typical bulk-morphology (Fig. 2b). Upon being combined via the mixing step, the g-C3N4/ SnO2 has a combinative morphology that includes NPs and thick layers percolating into each other (Fig. 2c). Furthermore, lattice fringes with d-spacing values of 0.326 nm and 0.249 nm could be attributed to the (110) and (101) planes of SnO2, respectively (Fig. 2d). Thus, we confirm the successful preparation of a com- posite between g-C3N4 and SnO2 NPs. 3.3. Band structure and optical properties The optical properties of the as-prepared materials are studied using diffuse reflectance spectroscopy (DRS) followed by the esti- mate of their bandgaps by Tauc plots (Fig. 3aec). The absorption edges in the materials present essential information on the photo- response, which reflects the photocatalyst efficiency of the mate- rials. The SnO2 NPs retain their UV-induced response because of the wide bandgap with an absorption edge below 400 nm. Meanwhile, g-C3N4 shows a great response in the visible light range with an absorption edge around 450 nm. The g-C3N4/SnO2 composite pre- sents a shift of the absorption edge toward the visible light range. The bandgap of the resulting g-C3N4/SnO2 composite is estimated P. Van Viet, H.-P. Nguyen, H.-H. Tran et al. Journal of Science: Advanced Materials and Devices 6 (2021) 551e559for two possible cases, i.e., direct and indirect bandgaps. Both cases unambiguously show a narrowed bandgap compared to the bare material. This result strongly suggests a heterojunction between SnO2 NPs and g-C3N4 is formed, leading to the observed reduction in bandgap. This change indicates that the bandgap of the SnO2 NPs was likely modified to create a more suitable band alignment, which will eventually form an interface with g-C3N4. Current un- derstanding suggests that the combination of SnO2 and g-C3N4 could lead to the formation of a so-called S-scheme heterojunction, Fig. 1. (a) XRD patterns and (b) FTIR spectra of the SnO2 NPs, g-C3N4, and the 30% g-C3N4/S materials taken by a digital camera. Fig. 2. Morphology of the materials observed by TEM images: (a) SnO2 NPs after the hydrot resulting 30% g-C3N4/SnO2 heterojunction, and (d) at a high-resolution to measure d-spaci 554which is extremely favourable for photocatalytic applications [21,30e32]. MotteSchottky behaviours are examined to indicate the alignment of the band structure of the proposed heterojunction. The conduction band (CB)/lowest unoccupied molecular orbital (LUMO) potential of the SnO2 NPs and g-C3N4 is þ0.20 V and 0.93 V, respectively (Fig. 3d). Thus, g-C3N4 has a higher CB than the SnO2 NPs, and this trend is the same for the highest occupied molecular orbital (HOMO) energy levels. Also, the nO2 heterojunction. Inset photos show the difference in the macroscopic colour of the hermal synthesis, (b) g-C3N4 after the annealing of melamine at 500
C for 2 h, (c) the ng values. s, g- Ps, greported work function of g-C3N4 (f ¼ 4:39 eV) is smaller than that of the SnO2 NPs (f ¼ 4:9 eV) [33e35]. The MotteSchottky behaviour of the g-C3N4/SnO2 heterojunction presents a shift in energy level that agrees well with the proposed S-scheme model. Accordingly, the CB of SnO2 changed from 0.20 V to 0.15 V, whereas the HOMO of g-C3N4 shifted by 0.05 V (0.93 V to 0.88 V). In the heterojunction, electrons from the HOMO of g- C3N4 move to that of the SnO2 NPs and form an internal layer at the interface. This interesting behaviour enables bending at the interface of g-C3N4 and SnO2, which likely gives rise to efficient Fig. 3. Optical properties of the as-prepared materials: (a) absorption spectra of SnO2 NP (c) SnO2 NPs to the bandgap of 30% g-C3N4/SnO2 NPs; (d) MotteSchottky plots of SnO2 N materials before and upon being combined. P. Van Viet, H.-P. Nguyen, H.-H. Tran et al.separation. We thus propose a band energy structure of the in- dividual materials and the resulting heterojunction after forma- tion of the S-shaped interface (Fig. 3e). Such an interface promises to considerably boost photocatalytic performance because of its prolonged electronehole lifetime. 3.4. Photocatalytic activity of the as-prepared materials Various photocatalytic NO degradation experiments under visible light were conducted to evaluate the performance of the as- prepared materials (Fig. 4). Under irradiation, the blank sample (without photocatalyst) does not reduce NO, which indicates no self-reduction/leaching of NO in the system. The SnO2 NP sample shows a limited ability in degrading NO, mainly resulting from surface defects, because of its wide bandgap [12,14]. Meanwhile, g- C3N4 shows a moderate degradation against N