Carbon Nanotubes Incorporated Z-Scheme Assembly of AgBr/TiO2 for Photocatalytic Hydrogen Production under Visible Light Irradiations

Photocatalytic H2 production is a promising strategy toward green energy and alternative to carbon-based fuels which are the root cause of global warming and pollution. In this study, carbon nanotubes (CNTs) incorporated Z-scheme assembly of AgBr/TiO2 was developed for photocatalytic H2 production under visible light irradiations. Synthesized photocatalysts were characterized through transmission electron microscope (TEM), X-ray photoelectron spectra (XPS), X-ray diffractometer (XRD), Fourier transform infrared (FTIR), photoluminescence spectra (PL), Brunauer Emmet-Teller(BET), and UV-vis spectroscopy analysis techniques. The composite photocatalysts exhibited a H2 production of 477 ppm which was three-folds higher than that produced by TiO2. The good performance was attributed to the strong interaction of three components and the reduced charge recombination, which was 89 and 56.3 times lower than the TiO2 and AgBr/TiO2. Furthermore, the role of surface acidic and basic groups was assessed and the photocatalytic results demonstrated the importance of surface functional groups. In addition, the composites exhibited stability and reusability for five consecutive cycles of reaction. Thus, improved performance of the photocatalyst was credited to the CNTs as an electron mediator, surface functional groups, higher surface area, enhanced charge separation and extended visible light absorption edge. This work provides new development of Z-scheme photocatalysts for sustainable H2 production.


Introduction
Photocatalytic H 2 production is an attractive approach to satisfy the worldwide energy requirements and to reduce the global warming. However, development of efficient photocatalyst with the ability to function under visible light irradiations and good charge separation is still a great challenge. State-of-the art Z-scheme photocatalysts offer a two-fold advantage of efficient charge separation and visible light solar energy utilization [1,2]. In Z-scheme photocatalytic system, compatible semiconductors with respect to their band positions establish efficient charge transfer mechanism which lessen the charge recombination. Further, the efficiency of charge separation is even higher in indirect Z scheme structure owing to the presence of electron mediators. The electron mediator expedites the process of charge transfer and reduce the recombination of electrons-holes pairs. These parted electrons and holes with high redox potential increase the efficiency of photocatalytic production [2,3].
The Z-scheme assembly of TiO 2 with compatible semiconductors is considered as a promising strategy to enhance the charge transfer and light absorption capability [4,5]. Silver bromide (AgBr) with an indirect bandgap energy of 2.6 eV have hardly been used in photocatalysis because of decomposition problem under light irradiations. However, the coupling of AgBr with suitable semiconductors like TiO 2 enhances the stability by developing heterojunction which fastens the transport of electrons [6] and prevent the degradation. The conduction band potential of AgBr is −0.30 eV allowing it to develop Z-scheme assembly with TiO 2 whose conduction band is −0.5 eV [7]. Asi et al. [8] studied the AgBr/TiO 2 photocatalysts for CO 2 conversion into CH 3 OH and reported the stable activity for continuous four cycles. Similarly, AgBr particles supported over amine functionalized rGO were reported by Shah et al. [6] for the decomposition of methylene blue and they noticed stable activity for continuous 12 cycles. Further, electron mediator plays an important role in the efficiency of indirect Z-scheme system. Noble metals such as Ag, Au, and Pt have been reported as electron mediators for Z-scheme photocatalysts [9][10][11]. However, noble metals-based Z-scheme systems are costly and require complex synthesis protocols.
Nowadays, there is an increasing trend to develop carbon-based electron mediators such as graphene and carbon nanotubes owing to their high electrical conductivity [12][13][14]. Numerous studies have reported the use of graphene as an electron mediator because of high charge carrier separation capability. Similarly, carbon nanotubes have been used as an electron mediator owing to extraordinary electrical, optical, and physical properties. Recently, Mohamad et al. [13] have reported the Z-scheme assembly of g-C 3 N 4 /CNTs/TiO 2 composites for the photocatalytic degradation of dyes. The composite exhibited almost six times higher degradation of phenol because of intimate surface contact between the two semiconductor through CNTs which reduced the charge recombination. Similarly, Boon-Joon et al. [3] reported CNTs as an electron mediator in Zn 0.5 Cd 0.5 S-MWCNT-TiO 2 nanocomposites for H 2 production. The CNTs efficiently transported the electrons from TiO 2 to Zn 0.5 Cd 0.5 S. Yet, there are few studies which focus on carbon nanotubes as an electron mediator. Further, effective bonding is essential to achieve the desired charge separation and transportation through an electron mediator. The strength of bonding depends on the surface functional groups that promote the interaction between semiconductors. The surface functional groups play a dynamic role in enhancing the interaction of reactants with surface of photocatalyst and photocatalytic stability. For example, Nasir et al. [5] reported that AgBr/rGO/TiO 2 exhibited higher efficiency in the basic medium rather than acidic. Acidic and basic groups can be introduced on the surface of the composite by functionalization. Additionally, the difficulty of pH control during the photoreaction experiment could also be avoided by functionalization of materials with different solvents which would be an added advantage to the stability of photocatalyst. However, there is limited study on coupling of CNTs as an electron mediator and the role of the functional groups in Z-scheme structures.
In this study, carbon nanotubes-incorporated Z-scheme assembly of AgBr/TiO 2 was studied. The AgBr/bCNTs/TiO 2 photocatalysts were synthesized by deposition of AgBr and TiO 2 on CNTs followed by reflux treatment to develop effective interfacial bonding. Obtained photocatalysts were calcined and analyzed through several characterization techniques including TEM, XRD, XPS, Raman, PL, UV-vis spectroscopy, N 2 adsorption and desorption technique. Performance of the AgBr/CNTs/TiO 2 was studied for H 2 production under visible light irradiations. Furthermore, role of functional groups was assessed by treating the CNTs with NaOH and NaOCl, and with H 2 SO 4 , HNO 3 and CH 3 COOH solvents and integrated with AgBr/TiO 2 . In the end, the mechanism of photocatalytic H 2 generation was proposed.

Preparation of Functionalized Carbon Nanotubes
The carbon nanotubes (CNTs) were dispersed into 1M NaOH and were boiled at 120 • C for 15 min. After that 30% NaOCl was added into dispersion followed by boiling for 2 h. The suspension was then cooled down to room temperature, filtered and neutralized by dispersing into the HCl solution. Obtained CNTs were washed with 0.1 M HCl and water to remove residual acid and base contents, followed by drying at 80 • C for 8 h. The product was denoted as bCNTs. Similarly, CNTs functionalization was carried out in acidic media containing 0.1 M H 2 SO 4 , 0.1 M HNO 3 , and 0.1 M CH 3 COOH at 120 • C for 4 h. The mixture was then filtered and washed several times with water to remove residual acid contents and denoted as aCNTs.

Preparation of AgBr/CNTs/TiO 2
The AgBr/bCNTs/TiO 2 was prepared by employing a two-steps approach which includes growth of AgBr over CNTs and bonding with TiO 2 particles followed by reflux. A particular amount of bCNTs were distributed in absolute ethanol with the help of sonication for 1 h. The desired quantity of CTAB was added to the bCNTs mixture. The required weight of AgNO 3 was dissolved in NH 3 solution and mixed into the above suspension. The obtained mixture was stirred for 1 h for the uniform deposition of AgBr nanoparticles on the surface of bCNTs. Simultaneously, the desired quantity of TiO 2 was distributed in absolute ethanol, sonicated and stirred for 1 h. The TiO 2 suspension was mixed with AgBr/bCNTs solution and the resulting mixture was shifted to round neck flask and refluxed for 2 h at 80 • C to strengthen the interfacial bonding among AgBr, bCNTs and TiO 2 . The obtained product was filtered, dried in an oven at 100 • C and calcined at 300 • C for 2 h. A similar method was repeated to synthesize the AgBr/aCNTs/TiO 2 , where AgBr/TiO 2 was synthesized using the same approach without the addition of the CNTs. Likewise, aCNTs/TiO 2 and bCNTs/TiO 2 were also prepared without the addition of AgNO 3 and CTAB.

Characterizations of Photocatalyst
The surface morphology of the synthesized photocatalysts was studied employing a transmission electron microscope (TEM HITACHI HT7700, Tokyo, Japan) on different magnifications. Crystalline structure photocatalysts were investigated using an X-ray diffractometer (XRD, PANalytical X'pert 3 powder, Malvern WR14 1XZ, U.K) with a diffraction angle range of 5 • to 90 • employing Cu α radiations (40 kV and 30 mA) at a scanning rate of 0.07878 • . The surface chemical composition was investigated using X-ray photoelectron spectra (XPS) (Thermo scientific K-alpha (Kα) spectrometer, East Grinstead, U.K). The XPS spectrometer was equipped with Al Kα radiations with corrected calibration binding energies against the C 1 of carbon fixed at 284.6 eV. Raman microscope (Lab RAM HR Evolution, HORIBA, Kyoto, Japan) with laser excitation at 325 nm in the range of 100-1000 cm −1 was used to record the PL spectra. The functional groups present on the surface of different photocatalysts were observed using Fourier transform infrared (FTIR) spectroscopy (Thermo Nicolet FTIR 6700, Thermo Scientific, Waltham, MA, USA) with wavelengths of 600 to 4000 cm −1 . Optical properties of photocatalysts were studied using UV-visible spectrometer (Agilent Cary technologies 100 UV-vis spectrometer Model G9821A, Santa Clara, CA USA) in the range of 200 to 800 nm. Bandgap energies of photocatalysts were estimated from Tauc plots of Kubelka-Munk Function versus photon energy.

Photocatalytic H 2 Production
The performance of the photocatalysts was evaluated for H 2 generation from water using a slurry reactor as reported in our previous study [5]. The glass photoreactor consisted of two chambers with 35 W HID Xe lamp (intensity 20 mWcm −2 , Philips, Selangor, Malaysia) as a source of light. Methanol was added as a sacrificial agent. Typically, 100 mg of photocatalyst was dispersed into a 20% methanol solution. Photoreactor setup was positioned on a magnetic stirrer for good dispersion of photocatalyst in the slurry. Prior to photoreaction, N 2 gas was purged for 1 h to remove gaseous impurities from the reactor and also used as a carrier gas at the flow rate of 20 mL/min. Photoreaction was triggered by switching on Xe lamp. Gaseous samples were collected after an interval of 1 h in gas collecting bags and H 2 was detected using Intelligent H 2 analyser (BROTIE Technology Co., Ltd., Haidian District, Beijing China).

Characterization of Photocatalysts
Crystallinity and phase structures of aCNTs, bCNTs, aCNTs/TiO 2 , bCNTs/TiO 2 , AgBr/aCNTs/TiO 2 and AgBr/bCNTs/TiO 2 photocatalysts were studied using XRD analysis as shown in Figure 1 [15]. The particles size was 44.79 nm measured from the prominent peak at 25.33 • . The pattern of both aCNTs/TiO 2 and bCNTs/TiO 2 did not show any peaks for CNTs owing to the very low quantity of CNTs. In AgBr/aCNTs/TiO 2 and AgBr/bCNTs/TiO 2 spectra, there were additional new peaks at 31.21 • and 44.54 • representing the AgBr nanoparticles with face-centred cubic structure with crystal planes of (200) and (220) crystal plane, respectively [16]. No peak was observed for silver (Ag • ) indicating the formation of only the AgBr/bCNTs/TiO 2 photocatalyst . Similarly, no peak was found for CNTs because of marginal quantity, even dispersion and shielding effect of TiO 2 peak at 2θ = 25.33 • [17]. These results showed the successful formation of AgBr/aCNTs/TiO 2 and AgBr/bCNTs/TiO 2 by the two-steps method.  Figure 2d,e demonstrate higher magnification images showing the unvarying distribution of TiO 2 , AgBr and multi-wall carbon nanotubes. Multiwall structure and thickness of each carbon nanotubes could be clearly observed. Deposition of TiO 2 and AgBr particles on the walls of CNTs can also be observed, which shows the strong integration of particles with CNTs. The strong bonding among particles would help in efficient charge separation across Z-scheme assembly of the photocatalysts and consequently higher H 2 production. Furthermore, Figure 2f shows selected area electron diffraction (SAED) image of AgBr/bCNTs/TiO 2 having several bright spots circles which shows the various crystalline planes of AgBr/bCNTs/TiO 2 . Additionally, TEM images showed that CNTs were uniformly dispersed and no carbon peak in the XRD analysis was observed which was due to the lower amount and even dispersion of CNTs. Surface chemical composition of AgBr/bCNTs/TiO 2 was investigated through XPS spectroscopy as shown in Figure 3. Survey spectra of AgBr/bCNTs/TiO 2 exhibited various peaks for silver (Ag), titanium (Ti), bromide (Br), carbon (C) and oxygen (O) as shown in Figure 3a and deconvoluted spectra of individual elements were further plotted to analyse the chemical species in depth as in Figure 3a-d. The Ag spectra presented two major peaks at 367.78 eV and 373.88 eV predicting the Ag3d 1/2 and Ag3d 3/2 matching to Ag +1 [18]. Valency of Ag was additionally confirmed from the spectra of Br which has peaks at 68.58 and 69.68 eV of Br3d 5/2 and Br3d 3/2 , corresponding to Br −1 in AgBr [19]. The Ti2p spectra has three different peaks at 459.28, 464.98 and 472.68 eV corresponding to Ti 4+ of Ti2p 3/2 and Ti 2p 1/2 , respectively [20]. Similarly, C1s spectra produced two peaks at 285.18 and 289.58 eV, respectively. The peaks at 289.58 eV represented the chemical bonding of carbon with oxygen in the form of C-O and C-O-C functional groups. On the other hand, an intense peak at 285.18 eV indicated the sp 2 characteristics of carbon nanotubes corresponding to C=C, C-C, and C-H bonding [21]. Importantly, the bigger area of peak at 285.18 eV showed that functional groups from the surface of CNTs were eliminated during heat treatment and calcination process. Likewise, O1s spectra of AgBr/bCNTs/TiO 2 demonstrated the main peak at 530.38 indicating the Ti-O-Ti, and Ti-O-C bonding [22]. Therefore, findings showed the strong bonding of AgBr, CNTs and TiO 2 .  Figure 4 shows the FTIR spectra of aCNTs, bCNTs, aCNTs/TiO 2 , bCNTs/TiO 2 , AgBr/aCNTs/TiO 2 , AgBr/bCNTs/TiO 2 . Spectra have several peaks at different region corresponding to various surface moieties that helped in the development of effective bonding in final composites. In aCNTs and bCNTs, peaks at 3300 cm −1 correspond to the OH functional groups. However, intensities of peaks are higher in aCNTs compared to bCNTs. Peaks appearing at 2923 cm −1 in bCNTs, and 2890 cm −1 in aCNTs represent the -CH stretching. Peaks at 2302 cm −1 in aCNTs and peaks at 2183 cm −1 in bCNTs showed the COOH bonding [23]. Intensity of COOH bonding is higher in bCNTs compared to aCNTs/TiO 2 . Similarly, bCNTs exhibited peaks at 1493, 1307, 1132 and 943 cm −1 in the C=C and -COH groups, whereas peaks were observed at 1556 and 1278 as well 1005 cm −1 for C=C and -COH groups in aCNTs [23]. Intensities of these groups are obviously higher in bCNTs which shows that it would help in more effective bonding in composites. In rest of the samples including aCNTs/TiO 2 and bCNTs/TiO 2 , AgBr/bCNTs/TiO 2 and AgBr/bCNTs/TiO 2 most of peaks disappeared because of thermal treatment during the synthesis at high temperature. Only dominant peaks near the 600 cm −1 was due to the Ti-O-Ti bonding. Photogenerated charge carriers trapping, separation, transportation and recombination was estimated by PL spectroscopy as shown in Figure 5. It is obvious that charge recombination intensity of AgBr/aCNTs/TiO 2 was higher than other photocatalysts owing to the poor bonding between the two semiconductors. The luminance intensity of TiO 2 compared to aCNTs/TiO 2 , bCNTs/TiO 2 and AgBr/bCNTs/TiO 2 was very high because of lower charge separation efficiency. The charge recombination intensity of AgBr/TiO 2 (8978 a.u) was lower compared to TiO 2 (14,541 a.u). This shows the better separation of charges in direct Z-scheme assembly of AgBr/TiO 2 which would be helpful for higher photocatalytic H 2 production. In addition, aCNTs/TiO 2 , bCNTs/TiO 2 , and AgBr/bCNTs/TiO 2 exhibited significantly lower intensities compared to that of aCNTs/TiO 2 , AgBr/aCNTs/TiO 2 , AgBr/TiO 2 and TiO 2 . Remarkable reduced charge recombination intensity indicated the important role of carbon nanotubes as an electron mediator in AgBr/bCNTs/TiO 2 photocatalysts [24]. Thus, improved charge separation was ascribed to multiwall carbon nanotubes as a solid-state electron mediator in AgBr/TiO 2 . The optical behaviour of photocatalyst shows the threshold wavelength of light photons required to initiate the photoreaction. Figure 6 shows the light absorption profile of photocatalysts examined by UV-vis spectrophotometer. All the photocatalysts showed higher absorption in the UV light spectrum owing to the transition of energized electrons from O 2p to Ti 3d of TiO 2 [25]. The aCNTs/TiO 2 and bCNTs/TiO 2 exhibited little extended light absorption edges compared to TiO 2 and also higher absorption coefficient. The AgBr/bCNTs/TiO 2 and AgBr/aCNTs/TiO 2 exhibited stretched light absorption in the visible light region as compared to bCNTs/TiO 2 and aCNTs/TiO 2 and TiO 2 showing that optical response was greatly enhanced. The light absorption edge of AgBr/bCNTs/TiO 2 is more stretched into the visible light region compared to the AgBr/aCNTs/TiO 2 which makes the former more efficient compared to the latter. This will help to produce greater number of photoelectrons and eventually yield of H 2 production. Moreover, the indirect bandgap energy of AgBr is 2.60 eV which helped to improve the optical response and excitation of photogenerated electrons from valence band (VB) to the conduction band (CB) in AgBr/CNTs/TiO 2 [26,27]. Thus, improved UV-Vis spectrum of AgBr/bCNTs/TiO 2 can be regarded as a superposition of AgBr and bCNTs/TiO 2 optical absorption spectra. Surface areas of TiO 2 , bCNTs/TiO 2 and AgBr/bCNTs/TiO 2 were estimated by BELSORP-mini as shown in Figure 7. The N 2 adsorption and desorption isotherms of TiO 2 , bCNTs/TiO 2 and AgBr/bCNTs/TiO 2 resembled type IV isotherm of IUPAC having hysteresis loop in the range 0.8-1 of P/P o and designated the mesoporous structures of photocatalysts. The surface area of TiO 2 was 45 m 2 which was increased to 49 m 2 because of the addition of bCNTs. However, the surface area of final composite, AgBr/bCNTs/TiO 2 , was 47 which was lower than bCNTs/TiO 2 . This decrease in the area was due to the lower surface area of AgBr, and small agglomeration of particles [28]. Further, bandgap structure of AgBr/bCNTs/TiO 2 photocatalyst were investigated from XPS spectra and Tauc plot of modified Kubelka-Munk function. To measure VB, XPS data were standardized with reference to carbon C1s position at 284.6 eV and were plotted as shown in Figure 8a. The value of VB was measured from the intersection of a tangent line drawn from the curve and straight lines from the initial point of curves. The measured value of VB was 2.42 eV, whereas reported VB potential of AgBr and TiO 2 are 2.30 eV and 2.7 eV, respectively [29]. Modified VB value indicated the stronger bonding among TiO 2 , CNTs and AgBr. Furthermore, Figure 8b shows the Tauc plot of modified Kubelka-Munk (KM) function vs. photoenergy/bandgap to estimate the bandgap energies of different photocatalysts. Bandgap energy of AgBr/bCNTs/TiO 2 was 2.62 eV which is lower than TiO 2 (3.2 eV). Lower bandgap energy of photocatalyst designated the higher expected light absorption and photocatalytic efficiency. The value of CB was calculated using Equation (1) given as follows.
where E g , E VB and E CB are the bandgap energy, VB and CB potentials of semiconductors. The E CB of AgBr/bCNTs/TiO 2 was found to be −0.2 eV, whereas stated E CB of AgBr and TiO 2 are −0.30 and −0.5 eV, respectively. Altered CB position of AgBr/bCNTs/TiO 2 was ascribed to the bCNTs and AgBr.

Photocatalytic H 2 Production
The photocatalytic performance of all the composites was assessed for H 2 generation from water using methanol as hole scavenger under visible light irradiations. Production of H 2 using different photocatalysts is shown in Figure 9. Pure TiO 2 produced 151 ppm of H 2 , which was lower than other photocatalysts owing to its large bandgap energy and severe charge recombination. The AgBr/TiO 2 photocatalyst exhibited 289 ppm of H 2 almost twice than benchmark TiO 2 . The AgBr/bCNTs/TiO 2 and AgBr/aCNTs/TiO 2 photocatalysts exhibited higher yield of 477 ppm and 376 ppm of H 2 which were greater than TiO 2 . The significant increase in yield of the H 2 was due to lower bandgap energy and efficient charge recombination. Also, incorporation of CNTs boosted the amount of H 2 production owing to higher surface area, electrical, and optical properties. Therefore, enhanced efficiency of AgBr/bCNTs/TiO 2 could be credited to efficient charge separation, light absorption, and indirect Z-scheme assembly developed between AgBr and TiO 2 through bridge of CNTs. Evidently, the optical behaviour of AgBr/bCNTs/TiO 2 was far better than TiO 2 which played a pivotal role in the enhancement of photocatalytic activity under visible light irradiations. Bandgap energy of AgBr/bCNTs/TiO 2 was 2.62 eV, lower than pure TiO 2 because of the incorporation of CNTs and AgBr. The modified optical response enabled the composites to work efficiently under visible light irradiations and led to 477 ppm of H 2 production. Moreover, the conduction band position of composite was −0.2 eV, which favoured the water splitting. It is due to the fact that the reduction potential for H 2 is 0 eV, whereas CB of AgBr/bCNTs/TiO 2 was −0.20 eV, higher than overall reduction potential required for water splitting and sufficient to produce the H 2 [30]. In addition, the dominant factor for the enhancement in H 2 production was efficient charge separation and transportation capability of photocatalyst. Charge recombination intensity of composites was 89 times lower than simple TiO 2 owing to the development of Z-scheme heterojunction. Incorporation of CNTs increased the surface area and enhanced the charge separation because of the excellent electrical conductivity [24]. Separated electrons and holes on the surface of photocatalyst carried out redox reaction and exhibited a remarkable improvement in H 2 production. The AgBr/aCNTs/TiO 2 charge separation was very poor compared to the AgBr/bCNTs/TiO 2 and led to the lower H 2 production. Further, the dual role of CNTs was very important owing to the generation of greater active sites because of the greater area, surface functional groups and higher diffusion capability. Diffusion of charge through tubular structure lowered the charge recombination and led to good separation and transportation [24,31,32].
Moreover, as an electron mediator, the activity of acid functionalized CNTs was quite different from the basic one. The H 2 production rate of acid-functionalised aCNTs-based photocatalyst was 376 ppm whereas base-functionalised bCNTs-based photocatalyst exhibited 477 ppm of H 2 , which was 1.27 times higher than the former. The difference of yield was attributed to functional groups attached on the surface of bCNTs. The functional moieties on the bCNTs surface affected the interfacial bonding between AgBr/TiO 2 and bCNTs. Strength of bonding was responsible for charge recombination [33]. Higher efficiency of the base-functionalised CNTs-based composites showed that there was uniform dispersion and strong bonding among CNTs, TiO 2 and AgBr which led to efficient charge separation and transportation. Quickly migrated electrons and holes attacked the H + and reduced it into H 2 . On the other hand, the comparatively lower yield of H 2 over aCNTs-based photocatalyst revealed that AgBr and TiO 2 interfacial bonding strength was weaker than the former which led to relatively lower production of solar fuel. Also, functional groups present on the surface of the photocatalysts help in redox reaction and intermediate formation which further enhance the H 2 production. In addition to the interfacial bonding strength, basic and acidic moieties on the surface of photocatalyst affected the dispersibility of the photocatalyst. It was visualised that dispersion of bCNTs-based photocatalyst was better and uniform as compared to aCNTs-based photocatalyst. Uniform dispersion of active photocatalyst particles enable the efficient harnessing of visible light irradiations and led to higher H 2 production. This can be further better explained by considering the affinity of AgBr particles. The AgBr particle is highly soluble in basic solution. The solubility of AgBr particles strongly affects the deposition process. When basic solvents were attached on the surface of the bCNTs, they promoted the interaction of depositing particles of AgBr on the surface of bCNTs. Uniform distribution and strong interaction of AgBr and bCNTs fortified the effective bonding of light-sensitive element with electron mediator. In the case of acid solvents, which is hydrophobic to the AgBr particles lead to poor dispersion and weaker interfacial bonding and eventually lower comparative efficiency of the photocatalyst. The difference in the behaviour of both the composites was due to the different nature of the functional groups present on the surface. Therefore, surface chemical species on electron mediator strongly influence the interfacial bonding and hence, yield of solar fuels [33][34][35].
In addition, effect of time was observed for the acid-and base-functionalised CNTs-based composites by performing the reaction for 5 h continuously as plotted in Figure 10. Both the photocatalysts demonstrated smooth production of H 2 . The yield was initially increased with time and then become almost constant. Further, recyclability analysis was performed for consecutive five cycles and photocatalyst exhibited almost the same yield of H 2 without losing activity as shown in Figure 11. Notable stability of photocatalyst was due to good interfacial bonding, generation and transportation of the photogenerated electrons, resulting in enhanced spatial charge separation [21,36]. The AgBr is highly sensitive to light and immediately decomposes to Ag and Br. Reason of decomposition could be the inability of generated charge carriers to move to other elements because of lower electrical conductivity or poor bonding. The stable activity was observed for five consecutive cycles owing to the smooth transfer of electrons across the electronic interface between AgBr and TiO 2 through bCNTs. Similarly, Xin et al. [37] reported that the activity of Ag/AgBr/GdVO 4 was stable without any significant loss for five consecutive cycles. Likewise, Xu et al. [38] used the AgBr/AgIn(MoO 4 ) 2 photocatalysts for the degradation of dyes and observed stable and smooth performance for six cycles. Thus, the stable photocatalytic performance of AgBr/bCNTs/TiO 2 was credited to the bCNTs as an electron mediator and stronger bonding because of the surface functional which enabled efficient separation of charges and reduced the recombination intensity. In addition, comparison was made with other studies reported in literature as shown in Table 1. It shows that photocatalytic H 2 in current study is comparable and even higher than many photocatalysts. However, true comparison cannot be developed till process parameters and reactor geometry are identical.   Water splitting involves complex redox reaction on the surface of photocatalyst and therefore it is crucial to understand the mechanism of H 2 production using photocatalysis. Figure 12 illustrates the mechanism of H 2 production over AgBr/bCNTs/TiO 2 . It could be seen that when light irradiations strike the surface of photocatalyst, electrons started moving from the VB to CB of AgBr. Photogenerated electrons jumped to the carbon nanotubes which were acting as a bridge in transferring the charge carriers from AgBr to TiO 2 . Separated electrons take part into redox reaction thereby reducing the H + into H 2 and holes left on AgBr attacked the adsorbed H 2 O molecules to oxidize it to H + and O 2 . Further, CB potentials of TiO 2 and AgBr are −0.5 and −0.3 eV, respectively, which determine the mechanism of transfer of electrons and hence redox reaction [7,46,47]. In AgBr/bCNTs/TiO 2 , AgBr is a light-sensitive element and emits electrons under visible light irradiations. These electrons are transferred to the TiO 2 for H 2 production else could recombine on the surface of AgBr leading to its decomposition and loss of activity. The CB of TiO 2 is higher than AgBr, therefore, the excited electrons move to the VB of TiO 2 only, resulting in the formation of Z-scheme assembly of AgBr/TiO 2 . As observed in PL spectra and H 2 production results, the efficiency of AgBr/TiO 2 was better than pure TiO 2 but considerably lower than AgBr/bCNTs/TiO 2 . These results indicated that the presence of CNTs was very important to enhance the efficiency of Z-scheme AgBr/TiO 2 because of its outstanding electrical conductivity as it reduced the charge recombination and also enhance the light absorption capability. Photocatalytic reduction of H + into H 2 depends upon the CB potential of a photocatalyst. For AgBr/bCNTs/TiO 2 , overall CB was −0.2 and VB was 2.42 which was sufficient to convert H 2 O into H 2 and O 2 . Further, the standard thermodynamics redox potential of H 2 is 0 eV, while theoretically redox potential of O 2 is 1.23 eV [48]. Similarly, individual CB positions of AgBr and TiO 2 are −0.30 and 0.5 eV, respectively [7,46,47]. Therefore, indirect Z-scheme assembly of AgBr/bCNTs/TiO 2 was an efficient photocatalyst for photocatalytic H 2 production form H 2 O.

Conclusions
Carbon nanotubes-incorporated Z-scheme assembly of AgBr/TiO 2 photocatalysts were successfully prepared by facile deposition and reflux method. The photocatalysts exhibited great potential for H 2 production with maximum yield of 477 ppm of H 2 . The performance of AgBr/bCNTs/TiO 2 was higher than AgBr/aCNTs/TiO 2 , AgBr/TiO 2 and TiO 2 , respectively. The charge recombination intensity of AgBr/bCNTs/TiO 2 was 89 times lower than TiO 2 and bandgap energy was 2.62 eV showing that properties of TiO 2 were greatly improved. Further, the photocatalyst exhibited stability for consecutive five cycles showing its potential for continuous H 2 production under visible light irradiations. Remarkable activity was ascribed to strong interfacial bonding, high surface area, reduced charge recombination and improved visible light response because of the incorporation of CNTs and surface functional groups. Thus, this study would be helpful to develop new Z-scheme photocatalyst with enhanced efficiency for various photocatalytic applications.