A Novel Synthesis of Rod-Shape BaNiSn-Graphene Decorated TiO2 Composite as a Ternary Photocatalyst to Improve Visible-Light Driven H2 Evolution with Lactic Acid and TEA

A novel rod-shape BaNiSn-Graphene oxide decorated TiO2 composite (BaNiSn-GT) has been synthesized using a simple ultrasonic method to enhance the visible-light-driven H2 evolution with cationic scavengers. The unique structure between the interfaces of BaNiSn-Graphene and TiO2 provides graphene oxide of contact and excellent electron transfer for H2 evolution activity. The BaNiSn-GT ternary photocatalyst exhibits relatively high photocatalytic activity with a hydrogen evolution rate of 1012 μmol/g during 4 h. On the other hand, BaNiSn-GT composite exhibited significantly higher hydrogen evolution rates of 870 μmol/g with TEA scavenger and 730 μmol/g with LA scavenger during 1 h, respectively. Moreover, the higher photocurrent density of BaNiSn-GT is correlated with electron–hole recombination, providing evidence for its inhibition, which leads to a longer lifetime of carriers produced by photoelectrons. The mechanism of the photocatalytic H2 evolution of BaNiSn-GT based on a full physicochemical characterization was proposed. This study provides new insight into the efficient hydrogen-evolution of graphene-based photocatalysts.


Introduction
The evolution of photocatalytic H 2 is an interesting way to generate H 2 and a graphene oxide on way to overcome the threat of an energy crisis. Despite the many efforts that have been devoted to the development of a highly efficient photocatalyst, the practical application of the H 2 evolution rate is still unsatisfactory, this is mainly due to the fast recombination of photogenerated carriers. In one of the newest approaches, heterojunction construction is widely adopted due to its efficiency in facilitating the separation of photogenerated electron-hole pairs and flexibility in material selection [1]. The development of efficient semiconductor composites utilizing visible light for renewable energy production is highly desirable. In many studies to date, many semiconductor materials are used to synthesize photocatalysts that are most suitable and efficient for H 2 production under solar irradiation, such as metal oxides, metal selenides, metal sulfides, metal-organic frameworks, and carbon-based materials commonly used in water photolysis [2][3][4][5]. Graphene oxide (GO) has recently attracted considerable attention in the field of energy conversion and environmental improvement due to its distinct advantages such as high conductivity, environmental friendliness, excellent solubility, and chemical inertness [6]. In particular, the formation of band gap energy of graphene oxide can be tuned in a wide energy range by controlling the size or functionality of the energy region, so it can be applied as a semiconductor with narrow gap energy for photocatalytic application.
In addition, it has been reported that the combination of GO and TiO 2 shows improved photocatalytic performance in the field of H 2 evolution and decomposition of pollutants [7,8]. GO, a photocatalyst in the graphene oxide-TiO 2 1 3 heterojunction, generally acts as a photoabsorbent or photosensitive agent to maximize photoelectron emission, and acts as an electron acceptor to receive photoelectrons generated from TiO 2 to induce efficient charge separation. However, despite its practical superiority, photocatalytic activity for graphene oxide-TiO 2 binary photocatalysts under visible light irradiation was not significantly improved for these effects due to the limited ability of graphene oxide to absorb visible light. [9,10]. To date, many semiconductors with wide visible-light utilization have been explored for hydrogen generation including CdS, ZrO 2 , BaTi 4 O 9 , ZnNb 2 O 6 ternary ZnIn 2 S 4 materials, and polymer composite. However, their wide application is restricted by many factors, including toxicity, photo-corrosion behavior, and poor photocatalytic efficiency. The construction of a non-toxic, inexpensive, and visible-light responsive photocatalyst for hydrogen evolution remains an urgent and challenging topic. In attempting to overcome these limitations, we try to synthesize multi-component metal sulfide with a tunable band gap and band gap alignment by controlling the composition. Recently, Mariia et al. also reports on the integration of CsPbBr 3 on g-C 3 N 4 nanosheets to construct composite photocatalysts [11]. These works have revealed that a welldesigned chemical bonded interface in the heterojunction is beneficial for the photocatalytic process. They found that the chemical bonded interface with extensive areas facilitates the migration of carriers across the junction. Our previous work, BaCuZnS-G-TiO 2 (BCZS-G-T) showed that Ba, and Zn containing compound with graphene oxide has high hydrogen evolution efficiency under visible-light irradiation for H 2 production reaction [12]. For a single conductive or semiconductor material, it is hard to fulfill all demands like cost effective, low bandgap effect and efficient photocatalytic hydrogen evolution. Thus, our study point is to make efficient short-rod type catalyst to for hydrogen evolution [10,13,14]. Other issue is implementing graphene oxide with other TiO 2 , due to the low band gap of graphene oxide and good catalytic activity of TiO 2 . The preparations of graphene oxide based composite nano-composites are convenient and the obtained products are usually more stable than isolated particles. Most importantly, the intimately bonded graphene oxide/TiO 2 surface promotes the separation of the photogenerated electrons and holes, and absorbs a wide range of incident light, thus enhances the photocatalytic performance [15,16].
In this study, an ultrasonication method was used to synthesize BaNiSn-graphene oxide with a mesoporous TiO 2 (BaNiSn-GT) serving as a co-catalyst for H 2 evolution. We also investigate the photoelectrochemical properties of pure BaNiSn, BaNiSn-G, and BaNiSn-GT, as well as the effects of the most relevant sacrificial hole scavenger on the photocatalytic H 2 evolution performance. Compared to pure BaNiSn and BaNiSn loaded with graphene oxide, the BaNiSn-GT composite exhibits remarkably improved performance in the evolution of photocatalytic H 2 . As the loading amount of BaNiSn-graphene oxide with TiO 2 increased, the H 2 generation rate under visible light irradiation (θ > 420 nm) increased by about 2.5 times compared to the pure BaNiSn sample. Improved photocatalytic activity can be attributed to the synergistic effect of BaNiSn and graphene oxide-TiO 2 , where graphene oxide acts as an electron acceptor and medium, while BaNiSn provides an active site for H 2 production. We hope that this work can contribute to the design and construction of highly efficient visible lightresponsive photocatalysts for sustainable energy harvesting and transition.

Synthesis of Metal Alloy (BaNiSn)-Graphene-TiO 2 (BaNiSn-GT) Photocatalyst
The BaNiSn-GT photocatalyst was synthesized using the ultrasonic method. A three-step combined method was used to synthesize this photocatalyst. First, the BaNiSn (BNS) photocatalyst had to be synthesized. In this preparation, 3.75 g and 4.45 g of Barium chloride (Solution A) and Nickel (II) nitrate hexahydrate (Solution B) were individually dissolved in 50 ml water. Then, the A and B solutions were consistently mixed under vigorous stirring for 1 h, which we named the AB solution. We took 3.50 g Tin (IV) Chloride Pentahydrate (Solution C) contained in 50 ml Hydrochloric Acid and followed the stirring process for 30 min at 60 °C. After that, Solution C was mixed dropwise (two drops/ sec) into the previously prepared AB solution. Then, the solution (ABC solution) was sealed with aluminum foil and stirred overnight at 80 °C. Finally, the ultrasonic process was followed for 5 h with 70 Hz amplitude at 90 °C. Then, the precipitation had to be washed several times with DI water and dried at 100 °C for 12 h, followed by calcination at 550 °C. This collected sample was named BaNiSn. Empirical results were obtained for the quantification of metal compounds from previous studies [12,17], and the mixing ratio was determined based on these results.
Next, we followed the same process to synthesize BaNiSn-G. To synthesize graphene by the oxidation of graphite powder, we used the Hummer-Offeman's method [18]. First, 15 g of natural graphite and 450 ml of sulfuric acid (H 2 SO 4 ) were mixed with de-ionized water and agitated for an hour at 0 °C in an ice bath. Then, the solution was removed from the ice bath, and potassium permanganate (45 g, KMnO 4 ) was slowly added into the solution, which was then continuously agitated at 35 °C. After turning to a dim brownish color, the container was sealed and kept at 1 °C with agitation for 30 min. Then hydrogen peroxide (H 2 O 2 ) was added drop wise within 5 min. The precipitation was required to wash with acetone and hydrochloric acid (HCl, 10%) several times, and it was heat-treated at 90 °C for 12 h until it formed into a graphite oxide powder. At this point, graphite oxide powder was dissolved in 300 mL DI water, agitated for 30 min, and sonicated for 2 h (using Ultrasonic Processor, VCX 750). Finally, the prepared solution was purified and washed with hot water several times. Then, the collected amount of solution was put in an oven for 6 h to obtain graphene oxide powder. The 1.0 g graphene oxide was dissolved in 60 ml of water and 40 ml ethanol, 94.0%, and sonicated for 2 h. After that, the solution was supposed to be combined with the previously made ABC solution and stirred for 1 h at 100 °C, then sonicated for 5 h with 70 Hz amplitude at 90 °C. The collected mixture was washed with DI water several times and dried at 160 °C for 8 h. It was then calcinated at 600 °C for 3 h. The collected sample was named BaNiSn-G. At the final step, 1.8 ml of titanium (IV) isopropoxide, 97.0%, was treated with 4 ml of acetic acid and 30 ml of ethanol, 94.0%. The BaNiSn-G solution was prepared following the same process. Then, the TiO 2 solution was added dropwise into the BaNiSn-G solution and stirred for 1 h. One more time of sonication was conducted for 3 h under the same conditions. Then, the precipitation was washed several times and dried at 160 °C for 12 h. Finally, we calcinated the sample at 600 °C for 3 h. The collected sample was names BaNiSn-GT.

Photocatalytic H 2 Production and Scavengers Experiments
In the first step, 50 mg of each of the three types of samples (BaNiSn, BaNiSn-G, and BaNiSn-G) was dispersed in 90 ml of water, and then 10 ml of triethanolamine (TEOA) and an aqueous chloroplatinic acid solution (H 2 PtCl 6 ⋅6H 2 O, 0.04 g/l) were added. The solution was stirred for 0.5 h while maintaining a vacuum state. 30 W of LED lamps (Fawoo, Lumidas-H, Korea, λ ≥ 420 nm) were used as a light source, and the reaction system was maintained at 6 °C for the entire process. The production of H 2 was determined by the handmade instrumental setting.
For hydrogen evolution, we made a handmade set up through U-shape tube with ml/cm scaling. And volume to mol conversion calculation process has been used to calculate the H 2 and O 2 amount. The schematic diagram of the Hydrogen evolution measurement process is shown in Scheme 1. Scavenger's experiments were performed using aqueous solutions containing 0.1 M Na 2 SO 4 and one of the following hole scavengers: 20 vol.% methanol (MeOH) (pH = 7), 20 vol.% isopropanol (i-PrOH) (pH = 7), Lactic acid (LA), and Triethanolamine (TEA) were used as a reference under the same conditions.

Instrumental Characterization
The morphology of the samples was examined by Transmission Electron Microscopy (TEM) (JSM5600 JEOL, Japan) (Hitachi HT7700, operated at 100 kV). The crystal structure, binding energy, and chemical composition of the composite sample were performed using X-ray diffraction patterns (XRD, Mini Flex600 diffractometer) and X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Fisher Scientific) as AlKa X-ray radiation sources. We used a UV-vis spectrophotometer (SIMADZU UV-2600) to obtain UV-vis diffuse reflectance spectra in the range of 200-800 nm. Photoluminescence spectra (PL) were examined using a Fluorescence Spectrophotometer (Hitachi F-7000) with an emission wavelength region of (280-500) nm at RT.

Structural and Surface Morphology
The X-ray diffraction (XRD) patterns of BaNiSn, BaNiSn-G, and BaNiSn-GT co-catalyst are displayed in Fig. 1 (a).
According to a previous report [19,20], the graphene exhibits a strong diffraction peak at a 2θ value of 26.3°. The XRD pattern of the BaNiSn sample showed characteristic diffraction peaks of (101), (510), (110), (004), and (220)  3 and BaNiSn 3 very well, respectively. Furthermore, compared to the XRD pattern of BaNiSn, In the case of the BaNiSn-G sample, some BaNiO 2 and BaSnO 3 disappeared or showed weak diffraction line peaks, and the BaNiSn 3 peak of the metal component was mainly strong. It is believed that some unstable BaNiO 2 and BaSnO 3 compounds disappeared during the graphene-ultrasonic-hydrothermal reaction. The formation of these crystals showed a similar phenomenon in BaNiSn-GT. In BaNiSn-GT, TiO 2 and BaNiSn 3 mainly showed strong peaks. Even in this sample, BaNiO 2 and BaSnO 3 disappeared or showed weak diffraction line peaks. After loading graphene oxide and TiO 2 , these peaks become less pronounced, especially at higher angle, as a result of the reduced diffraction contrast between walls and graphene oxide layer after the incorporation of TiO 2 particles into the layers. EDX element analysis was performed on the BaNiSn-GT sample. The analysis results for this are shown in b and c of Fig. 1. From this result, it can be seen that C occupies the largest content. In addition, the content is shown in the order of O and Ba. In addition to these, Ti, Ni, and Sn were additionally found.
Moreover, the surface morphologies of the synthesized BaNiSn, BaNiSn-G, and BaNiSn-GT composites were measured by SEM and TEM images, as shown in Fig. 2a-f. High temperature condition has been used in the synthesis process for the formation of rod like structure. Meanwhile, the interconnected architecture of the BaNiSn-GT showed a smaller particle size than that of TiO 2 , which can improve the active size. Figure 2 (d) shows the formation of BaNiSn nanorods with a size of almost 500 nm in aggregated form. Moreover, graphene thin layers could not be clearly observed the BaNiSn surface. Figure 2e-f depicts the TEM image of BaNiSn-GT, which reveals the formation of nanorods that is nicely dispersed on the graphene oxide sheets. The low loading of graphene oxide and TiO 2 can be reasonable for unchanged rod shape structure. The high temperature treatment produced framework shrinkage and the samplesmaintained rod like structure.

Chemical Composition
By analyzing X-ray photoelectron spectroscopy (XPS), the interfacial interaction among BaNiSn, graphene oxide, and TiO 2 compounds was studied, and the surface composition and chemical state of BaNiSn-GT were analyzed as shown in Fig. 3a-f. Figure 3a shows high-resolution XPS spectrum irradiation of Ba3d, Ni2p, Sn3d, O1s, C1s, and Ti2p. The main peak of Ni2p is divided into six sub-peaks by deconvolution operation [21]. Two spin split peaks located at 853.3 and 870.2 eV demonstrate the presence of Ni 2+ . Furthermore, the double peaks located at 861.4 and 879.6 eV imply the presence of corresponding satellite peaks (Ni2p 3/2 and Ni2p 1/2 ) (Fig. 3b). Peaks of 780.3 and 795.6 eV having a division value of 15.3 eV are all assigned to Ba 2+ (Fig. 3c) Also, in Fig. 3d, peaks at 486.8 and 495.4 eV are all assigned to Sn 4+ . Figure 3e shows two binding energy peaks of C1s at 285.5 and 289.5 eV confirming C=C and C=O or O-C=C, respectively. As can be seen from Fig. 3f, the Ti 4+ -O bond of the TiO 2 lattice is typical of the photoelectron peaks of Ti2p 3/2 and Ti2p 1/2 around 458.7 and 464.4 eV, respectively [22][23][24].
The results can confirm the structure of the BaNiSn-GT composite. In the overall XPS survey results, the intensity of the peak was significantly different, and the quantitative relationship of the element can be seen from this intensity. The intensity of C1s was the largest, and after that, the intensity of the peak gradually decreases in the order of O1s and Ni2p. Also, it can be found that Ti2p, Ba3d, and Sn3d show small peak intensity and occupy a small amount quantitatively. The peaks of O1s consisted of four peaks mainly occurring at 530.50 eV of O-Me,  (Fig. 3g). Table 1 presents the positions of the peak and the FWHM values.

Raman Analysis
Spectra of BaNiSn, BaNiSn-G and BaNiSn-GT measured by Raman spectroscopy are shown in 4, and the main Raman shift peaks at 394 cm −1 (B 1g ), 517 cm −1 (A 1g ), and 638 cm −1 (E g ) are due to the characteristic peaks of titanium dioxide nanotubes [25]. The two weak peaks at positions 1350 and 1600 cm −1 are due to the D-band and G-band derived from graphene, respectively. According to graphene's double resonance theory, 1387 cm −1 with a high D-band frequency results from disturbance of the connected hexagonal ring base plane [17]. In addition, the general E 12g and A 1g vibration modes of BaNiSn showed active modes of 425 and 3425 cm −1 , respectively, which is believed to be due to metal compounds. In our experiments, the blue shift of D-band and G-band (approximately 30 cm −1 ) is likely due to the combined action of quantum confinement effects and the strong interaction between the three compounds of BaNiSn, TiO 2 and GO (Fig. 4).

Optical Performance
Light absorbance spectroscopy has been analyzed to confirm the light response of the samples. From the Fig. 5a, it is apparent that the pure BaNiSn has less light absorption ability. After implemented the graphene oxide and TiO 2, light response increased. Thus BaNiSn-G, BaNiSn-GT have shown stable peak in the more visible light region. Furthermore, to   4 Raman Spectra of BaNiSn, BaNiSn-G and BaNiSn-GT study the optical absorbances of the BaNiSn, BaNiSn-G, and BaNiSn-GT composites, ultraviolet-visible diffuse reflectance spectroscopy was conducted (Fig. 5b). The bandgaps of the samples were evaluated by the Kubelkae-Munk function [17] where α, h, v, Eg, and C 1 are absorption coefficients, plank constants, optical frequencies, bandgap energy, and constants, respectively. Reflectivity was converted into absorbance data, and for this transformation, the Kubelkaie-Munk theory was used to obtain it. The photoabsorption increased significantly in the range of 2.3-1.1 eV. Pure BaNiSn showed 2.05 eV, and the BaNiSn-GT composite presents a band gap of 1.10 eV, which was lower than pure BaNiSn. Low band gap energy values are advantageous for the production of H 2 evolution under visible light irradiation.

Possible Photocatalytic Mechanisms (Mott-Schottky Plots)
Theoretical analysis and experimental tests were used simultaneously to verify the successful formation of the Z-system in the composite heterojunction structure. Figure 6 shows a Mott-Shottky plot of BaNiSn, BaNiSn-G, and BaNiSn-GT using a saturated calomel electrode as a reference electrode, and may measure a flat band potential of a material from the plot. In the n-type semiconductor, the flat band potential may be approximated by the position of the energy region of the conductive band CB. From the measured results, the CBs of BaNiSn, BaNiSn-G and BaNiSn-GT are located at −0.6, −0.42, −0.31 V (vs NHE, pH = 7), respectively. Based on the band gap data obtained from the UV-Vis diffusion reflectance spectrum, the valence band positions (VB) of GO and TiO 2 are located at + 2.4 and + 3.2 V, respectively.

Photonic Spectroscopic Analysis (PL, EIS Spectra, and Photocurrent)
Close interface contact between Junction bonded semiconductors is advantageous for the separation of electrons and holes produced by irradiated light that can be determined by photo-electrochemical measurements. The PL spectrum is described as some combination of photoelectrons and holes received by the semiconductor. The higher the PL strength of the photocatalytic semiconductor composite compound, the slower the recombination and the shorter the electron lifetime. Figure 7a shows PL spectra of TiO 2 , GO, BaNiSn, BaNiSn-G, and BaNiSn-G at an excited wavelength of 550 nm. The emission peak of the BaNiSn-G shows strength similar to that of BaNiSn-GT in the range of 450 to 650 nm. However, BaNiSn-GT showed the lowest peak intensity, and the PL intensity of BaNiSn showed the strongest curve. When BaNiSn-GT was incorporated, the PL strength of BaNiSn decreased dramatically, suggesting a longer lifetime of photo-generated electron-hole pairs [2]. As is well known, TiO 2 and GO are basic substances. The results of PL analysis on these can be seen from many documents [12,26]. Our previous work [12] shows that TiO 2 exhibits a high PL curve due to the large bandgap energy value. This shows the effect of the photo-reaction effect late due to the late electron transfer reaction. However, the graphene complex compound presented in this study shows a low PL curve because the electron transfer reaction proceeds easily and quickly.
Junction-bonded composites have been studied by electrochemical impedance spectroscopy (EIS). From the Nyquist plot (Fig. 7b) it can be clearly observed that the circle radius of BaNiSn-GT is small in the BaNiSn and BaNiSn-G samples, indicating an increase in conductivity and improved charge separation ability after addition of graphene oxide. These results further confirm the higher efficiency of the BaNiSn-GT composite, as shown by photocurrent and photocatalytic experiments, which exhibit the smallest electron transfer resistance at the semiconductor/ electrolyte interface. GO is a good conductive material in complex compound studies [12,27,28]. When the metal compound forms a complex with GO, the metallic complex has good photo-electron transfer effect due to the synergy effect. In the results of previous studies [28], many results can be seen. As shown in Fig. 7b, the GO composite shows the effect of improving the conductivity.
The photocurrent density of the composite is the key to enhancing the performance of the photocatalyst. The instantaneous transient photocurrent measurement of BaNiSn-GT under visible light irradiation is shown in Fig. 7c. A higher photocurrent value corresponds to a higher photogenerative charge density, resulting in a more necessary charge for the photooxidation-reduction reaction at the BaNiSn-GT heterojunction. BaNiSn-GT obtained the strongest reaction value of 13,800 μA cm −2 . After four cycles of intermittent on/off irradiation, the stability of the graphene oxide rod photoelectric response of BaNiSn-GT was demonstrated through a nearly decay-free photocurrent curve.

H 2 evolution Analysis and Scavenger Effect
The activity of the prepared photocatalyst for hydrogen production was conducted under visible light irradiation for Radical capture experiments using complex photocatalysts and scavengers were also conducted. Na 2 SO 4 , methanol, ethanol, lactic acid (LA), and triethanolamine (TEA) were adopted as scavengers, respectively. As shown in Fig. 8b, the addition of TEA has the highest effect on the degradation efficiency, whereas the addition of Na 2 SO 4 , methanol, and ethanol having the same concentration and volume leads to almost 50% inhibition effect of the degradation efficiency. Thus, h + and ·OH are essential for the evolution of photocatalytic hydrogen, and in this reaction, ·OH has a greater effect than h + . Also, as shown in Fig. 9, the VB position of TiO 2 is much closer to the CB position of BaNiSn-G at the hetero interface. The distance between VB and CB position reduced due to zero band gap of graphene oxide. Addition of graphene oxide and TiO 2 produced a mid bandgap and hinder the recombination (electron to hole conversion). Thus, the electron lifetime increase and enhance the photocatalytic activity. During light excitation, the photoinduced electrons on the CB of BaNiSn-G can be spatially described as their movement between CB and VB to create a photo-generating hole on the VB. This is an irrefutable solid Z-system charge transfer process over heterogeneous interfaces. The role of BaNiSn in this photoelectron transfer reaction is high hydrogen production by controlling band gap energy. In the photocatalytic experiment, BaNiSn produced an oxide radical to produce a metal-electron conversion. Hydrogen molecules can be dissociated through adsorption on the surface of the metal catalyst. In the next step, the hydrogen atom moves from the metal catalyst to the non-metal support or adsorbent. Here, the sacrificial scavenger acted as an electron donor that played an intermediate role. In the experiment, the sacrificial scavenger released electrons, which atoms covered holes in the band and increased hole-electron conversion, which is an active site for efficient photocatalytic activity of the photocatalyst. The addition of an auxiliary sacrificial agent also increases the amount of hydrogen evolution.
The variation of the band gap energy and the location of the band gap energy due to the heterojunction effect of the photocatalysts used are shown in Fig. 9. From these results, it is shown that when metal alloy and metal oxide are bonded to graphene, the band gap energy is significantly reduced. In addition, the modified band gap energy has a significant impact on the photocatalytic effect. Figure 9 is presented based on the results shown in Table 2.
The results of previous studies on hydrogen production studies with several types of photocatalysts are shown in Table 3. Since they have different research conditions such as reaction time and the presence or absence of scavengers, quantitative comparison is problematic. However, from these studies, the accessibility of reactor design and commercialization can be considered in the future.

Conclusions
In summary, the unique structure of the novel rod-shape BaNiSn-GT co-catalyst was successfully prepared using an ultrasonic method. Photochemical measurements proved that BaNiSn, graphene oxide, and TiO 2 act as photo-sensitive agents to react to visible light reactions. The Junction-bonded BaNiSn-GT complex exhibits enhanced photocatalytic H 2 production activity and stability under visible light irradiation due to the combination of extended visible light absorption capacity, long life of photogenerated  The ability of the selected sacrificial agent to clean the photogenic holes was evident in the order of Na 2 SO 4 , methanol, ethanol, LA, and TEA solutions. The BaNiSn-GT ternary photocatalyst exhibits relatively high photocatalytic activity with a hydrogen evolution rate of 1012 μmol/g during 4 h. The presence of the sacrificial agent also results in a negative change in the starting potential of the photocurrent, which can be largely explained by the different redox potentials of the additive. This work not only provides an innovative strategy for synthesizing heterojunction ternary hybrid nanocomposites, but also provides new insights to understand the essential role of BaNiSn and TiO 2 in graphene oxide-based nanohybrids.