One step synthesis of high-efficiency AgBr–Br–g-C3N4 composite catalysts for photocatalytic H2O2 production via two channel pathway

In this work, a two-component modified AgBr–Br–g-C3N4 composite catalyst with outstanding photocatalytic H2O2 production ability is synthesized. XRD, UV-Vis, N2 adsorption, TEM, XPS, EPR and PL were used to characterize the obtained catalysts. The as-prepared AgBr–Br–g-C3N4 composite catalyst shows the highest H2O2 equilibrium concentration of 3.9 mmol L−1, which is 7.8 and 19.5 times higher than that of GCN and AgBr. A “two channel pathway” is proposed for this reaction system which causes the remarkably promoted H2O2 production ability. In addition, compared with another two-component modified catalyst, Ag–AgBr–g-C3N4, AgBr–Br–g-C3N4 composite catalyst displays the higher photocatalytic H2O2 production ability and stability.


Introduction
Hydrogen peroxide (H 2 O 2 ) is a highly efficient and green oxidant because it has a high content of active oxygen (47% w/w) and results in only H 2 O as a by-product. 1,2 Industrially, H 2 O 2 is produced by the anthraquinone method, in which energy consumption is high because of the multistep hydrogenation and oxidation reactions. Thus this method is unsuitable for the current new concept of being "green, energy saving, and environmentally friendly" in the chemical industry. Recently, the direct synthesis of H 2 O 2 from H 2 and O 2 gases has been widely studied using noble metals as catalysts. 3,4 It is considered to be an alternative and green chemical process. However, this method presents a potential explosion risk from H 2 /O 2 mixtures. In contrast, photocatalytic H 2 O 2 production requires only water, oxygen and visible light through two-electron reduction from the conduction band (reaction (1)). However, the H 2 O 2 can be decomposed by reduction with e À , which has caused the H 2 O 2 production rate of this photocatalytic reduction method to be unsatisfactory to date (reaction (2)).
Graphite phase carbon nitride (g-C 3 N 4 ), as the darling of the catalytic community in recent years, has special physical and chemical properties, excellent chemical stability and adjustable electronic structure. 5,6 Since its conduction band potential (À1.3 V) is more negative than the reduction potential of O 2 / H 2 O 2 (0.695 V), g-C 3 N 4 could reduce O 2 to H 2 O 2 under visible light thermodynamically. 7 However, g-C 3 N 4 also suffers from many disadvantages, such as the low visible-light utilization efficiency, high recombination rate and small BET surface area, which limit its practical application.
AgX (X ¼ Cl, Br, I) is a kind of common photocatalyst. They have the moderate energy level which is very appropriate to combine with g-C 3 N 4 for building heterojunction to resolve these disadvantages mentioned above. [8][9][10][11][12][13][14] Feng et al. prepared g-C 3 N 4 /AgBr nanocomposite photocatalyst via a protonation pretreatment method. 8 They found the efficient combination of g-C 3 N 4 and AgBr leads to Z-scheme charge transfer in the composite system, and the photocatalytic activity is therefore enhanced signicantly. Liu et al. prepared AgI@g-C 3 N 4 hybrid core@shell structure by ultrasonication/chemisorption method. 9 They suggested that the improved photocatalytic performance is due to synergistic effects at the interface of AgI and g-C 3 N 4 which can effectively accelerate the charge separation and reinforce the photostability of hybrid composite. In addition of building heterojunction with silver halide, halogen doping and Ag loading are also efficient method to promote the visible-light utilization efficiency and separation rate of electrons-holes. Lan et al. prepared Br doped g-C 3 N 4 for photoredox water splitting via one-pot co-condensation of urea with NH 4 Br. 15 The optimal sample shows more than two times higher H 2 evolution rates than pure g-C 3 N 4 under visible light irradiation, with high stability during the prolonged photocatalytic operation. Bu et al. prepared Ag loaded mesoporous g-C 3 N 4 with high photoelectric conversion performance. 16 They suggested that modifying mesoporous g-C 3 N 4 with Ag increases the conductivity and lowers the energy barrier of the interface reactions, thus enhances the separation efficiency of photogenerated electron-hole pairs.
Besides, two-component modication, such as Ag-AgX, are also used to promote the catalytic performance of g-C 3 N 4 . 17, 18 Wang et al. prepared a high-efficiency g-C 3 N 4 /Ag/AgCl plasmonic photocatalyst via a facile solvothermal method. 17 The surface plasmon resonance effect of the Ag nanoparticles, the polarization eld of AgCl and the g-C 3 N 4 /Ag/AgCl heterojunction all result in the improved photocatalytic degradation performance. Chen et al. prepared plasmonic photocatalyst Ag/ AgBr/g-C 3 N 4 by in situ ionic-liquid-assisted synthesis. 18 The enhanced photocatalytic activity is assigned to the extended light-absorption range and efficient charge separation caused by the surface plasmon resonance effect of Ag 0 and the wellmatched overlapping band-structure between Ag/AgBr and g-C 3 N 4 . Interestingly, till now, few literature concerning silver halide-halogen-g-C 3 N 4 catalyst is reported. F and I have been applied to dope g-C 3 N 4 . It was found that F dopant give much less promotional effect on the optical absorption of g-C 3 N 4 than I. 19 This is because the valence electrons in I atom with much less electronegativity are more delocalized to interact with the p electron system of g-C 3 N 4 . Such an extended conjugation system gives rise to the red-shi of the optical absorption of g-C 3 N 4 . However, the overlarge size of I atom is thermodynamically and geometrically difficult to dope into g-C 3 N 4 to form stable structure. To this end, Br in the middle of F and I is a recommended choice to modify g-C 3 N 4 . In this work, AgBr-Br-g-C 3 N 4 composite catalyst is prepared via one-pot synthesis. The as-prepared AgBr-Br-g-C 3 N 4 composite catalyst is applied for photocatalytic H 2 O 2 production under visible light irradiation. The effects of modication on the structural property, optical property, and photocatalytic performance of catalysts are discussed in detail. The possible reaction mechanism is proposed.

Experimental
Preparation and characterization 4 g of dicyandiamide and desired amount of AgNO 3 were dissolved into 40 mL deionized water under stirring to obtain solution A. Then, NH 4 Br solution was added dropwise into solution A (molar ratio Ag/Br ¼ 1). The obtained suspension was stirred for 2 h, and heated to 80 C to remove the water. The solid was dried at 80 C and followed by milling and annealing at 550 C for 2 h (at a rate of 5 C min À1 ). The obtained catalyst was denoted as AgBr/GCN(x), where x stands for the mass ratio of AgBr/dicyandiamide. Neat AgBr was prepared according to the method mentioned above in the absence of dicyandiamide. When excess NH 4 Br was added into solution A and followed the same procedure mentioned above, the prepared catalyst was denoted as AgBr/Br(y)-GCN(x), where y stands for the molar ratio of Ag/Br.
The XRD patterns of the samples were recorded on a Rigaku D/max-2400 instrument using Cu-Ka radiation (l ¼ 1.54Å). UV-Vis spectroscopy was carried out on a UV-Vis spectrophotometer (JASCO V-550) using BaSO 4 as the reectance sample. TEM images were taken on a Philips Tecnai G220 model microscope.
Nitrogen adsorption was measured at À196 C on a Micromeritics 2010 analyzer. The BET surface area (S BET ) was calculated based on the adsorption isotherm. Elemental analysis was performed using a vario EL cube from Elementar Analysensysteme GmbH. ICP was performed on a Perkin-Elmer Optima 3300DV apparatus. The XPS results were obtained on a Thermo Escalab 250 XPS system. Al Ka radiation was used as the excitation source. The electron paramagnetic resonance (EPR) was determined with a Bruker ESR 300E, using the radical scavenger dimethyl pyridine N-oxide (DMPO). The photoluminescence (PL) spectra were measured with a uorospectrophotometer (FP-6300) using Xe lamp as the excitation source.

Photocatalytic reaction
The photocatalytic H 2 O 2 production ability of the samples was evaluated by the reduction of molecular oxygen. For these experiments, 0.2 g of photocatalyst was added to 200 mL of deionized water. The suspension was dispersed using an ultrasonicator for 10 min. During the photoreaction under visible light irradiation, the suspension was exposed to a 250 W high-pressure sodium lamp with main emission from 400 to 800 nm, and O 2 was bubbled at 80 mL min À1 through the solution. The UV light portion of the sodium lamp was ltered by a 0.5 M NaNO 2 solution. All runs were conducted at ambient pressure and 30 C. At given time intervals, 5 mL aliquots of the suspension were collected and immediately centrifuged to separate the liquid samples from the solid catalyst. The H 2 O 2 concentration was analyzed by the normal iodometric method. 20,21 Results and discussion Fig. 1 shows the XRD patterns of GCN, AgBr and as-prepared composite catalysts. The characteristic peak of GCN around 27.5 could be clearly identied, which is attributed to the typical (0 0 2) interlayer-stacking peak corresponds to an interlayer distance of d ¼ 0.33 nm. The peak at 13.1 represents in-plane structural packing motif with a d value of 0.675 nm. 22 For AgBr, the sample shows several diffraction peaks at 31.0 , 44.4 , 55.1 and 64.6 , which are assigned to the (2 0 0), (2 2 0), (2 2 2) and (4 0 0) planes of AgBr crystal (JCPDS le: 6-438). 8,18 In the case of as-prepared composite catalysts, the peak position does not shi but the intensity for AgBr (GCN) decreases (increases) gradually with increasing the amount of GCN, as shown in Fig. 1a. It is shown in Fig. 1b that, when excess Br was  The morphologies of the representative samples were examined by TEM analysis. As shown in Fig. 3a, GCN shows sheet-like structure with no regular morphology. The morphology for AgBr is nanorod, as shown in Fig. 3b. In the case of AgBr/Br(1 : 4)-GCN(1 : 3) (Fig. 3c), both plate-like structural GCN and AgBr nanorod are observed, conrming the presence of both GCN and AgBr. The two-dimensional ordering of GCN is very weak and hard to nd the lattice fringe in HRTEM image (Fig. 3d). This is consistent with previous work. 18 However, the clear lattice fringe is observed for AgBr, very close to the (400) crystal face with the d ¼ 0.143 nm (Fig. 3d). This tight coupling is favorable for the charge transfer between GCN and AgBr and promotes the separation rate of electron-hole pairs. XPS spectra are used to investigate the structure of the asprepared catalysts. In Fig. 4a, the spectra of three catalysts in C 1s region can be tted with two contributions which located at 284.6 and 288.6 eV. The sharp peak around 284.6 eV is attributed to the pure graphitic species in the CN matrix. The peak with binding energy of 288.6 eV indicates the presence of sp 2 C atoms bonded to aliphatic amine (-NH 2 or -NH-) in the aromatic rings. [24][25][26] For N 1s region (Fig. 4b), three contributions located at 398.3, 399 and 400.5 eV were assigned to the sp 2 hybridized aromatic nitrogen atoms bonded to carbon atoms (C-N]C), tertiary nitrogen N-(C) 3 groups linking structural   motif or amino groups carrying hydrogen ((C) 2 -N-H) in connection with structural defects and incomplete condensation, and nitrogen atoms bonded three carbon atoms in the aromatic cycles. 27,28 For AgBr/GCN(1 : 3) and AgBr/Br(1 : 4)-GCN(1 : 3), the obvious shis to higher binding energies are observed in N 1s region compared with that of neat GCN. This is probably due to the change of chemical environment aer coupling with AgBr.
In Fig. 4c, AgBr displays Ag 3d spectrum with two peaks at 366.8 eV and 372.8 eV, corresponding to the binding energies of Ag 3d 5/2 and Ag 3d 3/2 of Ag + in AgBr, respectively. 29 In Br 3d region (Fig. 4d), the peaks at $67.5 and 68.3 eV for AgBr are assigned to the Br 3d 5/2 and Br 3d 3/2 of Br À state. 30  . This binding energy should be assigned to the C-Br bond, which conrming that Br is doped into g-C 3 N 4 lattice. 31 The VB XPS spectra were employed to determine the electronic structure (Fig. 4e). It is obvious to see that the VB potentials for GCN and AgBr locate at +1.32 and +2.15 V. They are very close to the previous results. 5,18 Combined with the UV-Vis results, the energy position of CB for GCN and AgBr locate at À1.43 and À0.38 V respectively. Obviously, the band structures of the two components are well-matched with each other. This facilitates the formation of heterojunction for charge transfer. Fig. 5 shows the PL spectra of as-prepared catalysts under air atmosphere using excitation at 255 nm. For GCN (Fig. 5a), broad PL band around 460 nm is observed with the energy of light approximately equal to the band gap of g-C 3 N 4 . AgBr exhibits several emission peaks which intensities are higher than that of GCN. In the case of as-prepared heterojunction catalysts, the PL spectra show the similar shape to that of GCN, whereas the intensities are obviously decreased. AgBr/GCN(1 : 3) shows the lowest PL intensity, hinting its most effective separation rate of electrons and holes. This is reasonable because, with this GCN/ AgBr mass ratio, GCN and AgBr have the approximate S BET (9.5 and 13.2 m 2 g À1 for GCN and AgBr, the yield for GCN is approximately 50 wt%). They can contact with each other as much as possible, leading to the formation of the maximum area of the heterojunction.
Room temperature electron paramagnetic resonance (EPR) was used to investigate the electronic property of the asprepared catalysts. As shown in Fig. 6, GCN shows almost no EPR signal. Aer coupling with AgBr, the sample also displays no EPR signal for AgBr /GCN(1 : 3). However, AgBr/Br(1 : 4)-GCN(1 : 3) shows one single Lorentzian line centering at a g ¼ 2.003, being originated from unpaired electrons on p-conjugated g-C 3 N 4 aromatic rings aer Br doping. This is probably due to that the delocalization of the valance electron of Br in the g-C 3 N 4 conjugation system can widen the band distribution, which improves the charge migration. 15 This is consistent with the PL result. Fig. 7 displays the photocatalytic H 2 O 2 production ability over as-prepared catalysts. It is clearly seen that the H 2 O 2 concentration of as-prepared catalyst increases with time for about 5 h when it reaches a constant level. This level corresponds to the steady-state where the rate of H 2 O 2 production is equal to the rate of decomposition. 32 In Fig. 7a, GCN and AgBr display the H 2 O 2 concentration of 0.5 and 0.2 mmol L À1 . In the case of as-prepared heterojunction catalyst, the H 2 O 2 production ability promotes obviously. Aer Br doping, the photocatalytic performance of as-prepared samples further increase, as shown in Fig. 7b. This is due to that the introduction of heteroatoms into the p-conjugated g-C 3 N 4 can accelerate the charge carriers transfer rate and thus restrain the recombination of electron and hole. AgBr/Br(1 : 4)-GCN(1 : 3) shows the highest H 2 O 2 equilibrium concentration of 3.9 mmol L À1 , which is 7.8 and 19.5 times higher than that of GCN and AgBr. The excess Br doping causes the decreased H 2 O 2 production ability of AgBr/Br(1 : 8)-GCN(1 : 3). This is probably due to that the excess doping Br acts as recombination sites to accelerate the recombination of electrons and holes, which is consistent with PL result.
Besides the O 2 reduction to form H 2 O 2 , another channel to produce H 2 O 2 is reported by Dong et al. 33 It is reported that photogenerated holes can oxidize OH À to $OH thermodynamically, as shown in reaction (3). Two $OH can form H 2 O 2 though combination with each other, as shown in reaction (4). In order to investigate the reaction mechanism of as-prepared heterojunction catalyst, the inuence of various scavengers on the H 2 O 2 production ability is carried out and shown in Fig. 7d. AgNO 3 and EDTA-2Na are used as the electrons (e À ) and hole (h + ) scavenger, respectively. 34 When AgNO 3 is added to trap the electrons, H 2 O 2 is still formed with the concentration of 0.15 mmol L À1 over AgBr. It is known that the redox potential for $OH/OH À is +1.99 V. 35 Whereas the VB of AgBr is +2.15 V. The VB holes in AgBr is positive enough to generate $OH. Therefore, H 2 O 2 should be produced by reaction (4). When EDTA-2Na is added to trap the holes, although the utilization rate of electrons is promoted, the H 2 O 2 equilibrium concentration is only increased to 0.45 mmol L À1 over AgBr. This is probably due to the poor reduction ability of CB electrons over AgBr. In the case of GCN, the CB and VB positions are À1.43 V and +1.32 V, respectively. Thus, when AgNO 3 is added, no H 2 O 2 is produced due to that the VB holes in g-C 3 N 4 are not positive enough to generate $OH. Without any doubt, the H 2 O 2 production ability of g-C 3 N 4 is obviously promoted by adding EDTA-2Na to trap the holes.
In general, there are two typical working mechanisms for heterojunction catalyst, double charge transfer mechanism and Z-scheme mechanism, as shown in Fig. 8a. 36 (1 : 3) is promoted to 3.2 mmol L À1 when EDTA-2Na is added. Based on the above results, it is deduced that not double charge transfer mechanism but Z-scheme mechanism with "two channel pathway" is proposed. Under visible light irradiation, the photogenerated electron-hole pairs are formed in both components. The electrons in the CB of AgBr combine with the holes in the VB of GCN at the interface of the heterojunction. Therefore, the holes tend to stay in the VB of AgBr and the electrons accumulate in the CB of GCN, leading to the enhanced separation rate of electron-hole pairs. The CB electrons in GCN can reduce O 2 to form H 2 O 2 , as well as the VB holes in AgBr can oxidize OH À to form $OH, which subsequently react with each other to form H 2 O 2 . Such "two channel pathway" causes the remarkably promoted H 2 O 2 production ability. When AgNO 3 (or EDTA-2Na) is added to trap the electrons (or holes), the separation efficiency of catalyst is promoted, leading to the enhanced H 2 O 2 production ability (Fig. 7d).
In order to prove the advance of this AgBr-Br-GCN heterojunction catalyst, Ag-AgBr-CN was prepared according to the previous work. 18   much lower than that of AgBr/Br(1 : 4)-GCN(1 : 3). In addition, the H 2 O 2 production ability for AgBr/Br(1 : 4)-GCN(1 : 3) keeps stable aer 48 h reaction. Whereas, the activity for Ag-AgBr-CN decreases obviously with increasing the reaction time, hinting its poor photocatalytic stability. The element analysis and ICP results indicate that the content of each element in the AgBr/ Br(1 : 4)-GCN(1 : 3) remains almost unchanged before and aer the reaction. However, for Ag-AgBr-CN, the Ag content decreases obviously (Br content does not change). This indicates the metal silver is not sturdy on the catalyst surface, which is probably lost during the reaction.

Conclusions
In this work, a two-component modied AgBr-Br-g-C 3 N 4 composite catalyst with outstanding photocatalytic H 2 O 2 production ability is synthesized. Modication with Br and AgBr does not change the structural property of g-C 3 N 4 but decreases the band gap and increases the visible light absorption. The formation of heterojunction with AgBr and introduction of heteroatoms Br into the p-conjugated g-C 3 N 4 can accelerate the charge carriers transfer rate and thus restrain the recombination of electron and hole. AgBr/Br(1 : 4)-GCN(1 : 3) shows the highest H 2 O 2 equilibrium concentration of 3.9 mmol L À1 , which is 7.8 and 19.5 times higher than that of GCN and AgBr. According to "Z-scheme" mechanism, not only the CB electrons of GCN reduce O 2 to form H 2 O 2 , but the VB holes in AgBr can oxidize OH À to form $OH, which subsequently react with each other to form H 2 O 2 . Such "two channel pathway" causes the remarkably promoted H 2 O 2 production ability. In addition, compared with another two-component modied catalyst, Ag-AgBr-g-C 3 N 4 , AgBr/Br(1 : 4)-GCN(1 : 3) displays the higher photocatalytic H 2 O 2 production ability and stability.

Conflicts of interest
There are no conicts to declare.