Laser flash photolysis study of Nb2O5/g-C3N4 heterostructures as efficient photocatalyst for molecular H2 evolution

Improvements of visible light activity, slow recombination rate, stability, and efficiency are major challenges facing photocatalyst technologies today. Utilizing heterostructures of g-C3N4 (bandgap ∼2.7eV) with Nb2O5 (bandgap ∼3.4eV) as an alternative materials for the first time, we tried to overcome such challenges in this work. Heterostructures of Nb2O5/g-C3N4 have been synthesized via hydrothermal technique. And then a time-resolved laser flash photolysis of those heterostructures has been analyzed, focusing on seeking how to improve photocatalytic efficiency for molecular hydrogen (H2) evolution. The transient absorption spectra and the lifetime of charge carriers at different wavelengths have been observed for Nb2O5/g-C3N4, where g-C3N4 was used for a control. The role of hole scavenger (methanol) has also been investigated for the purpose of boosting charge trapping and H2 evolution. The long lifetime of Nb2O5/g-C3N4 heterostructures (6.54165 μs) compared to g-C3N4 (3.1651897 μs) has successfully supported the increased H2 evolution of 75 mmol/h.g. An enhancement in the rate of H2 evolution (160 mmol/h.g) in the presence of methanol has been confirmed. This study not only deepens our understanding of the role of scavenger, but also enables a rigorous quantification of the recombination rate crucial for photocatalytic applications in relation with efficient H2 production.


Introduction
Photocatalysis is one of the most intriguing energy conversion processes that employ renewable and eco-friendly solar energy to generate ideal fuels and resolve the problem of photocatalytic degradation [1,2]. Hence, photocatalysis has gained a growing interest in connection with its applicability in solar energy conversion including water-splitting, photocatalytic H 2 production, photoreduction of CO 2 , photodegradation of organic contaminants, and organic photosynthesis [3][4][5][6]. For photocatalytic H 2 reactions, few metal oxides have been explored but due to their large band gaps and inappropriate band potentials, they are usually not ideal for H 2 generation or other photocatalytic applications. Some other oxides, such as CuO and Cu 2 O, have favourable band gaps but have not been preferred because of their less chemical stability and less light absorption in visible region [7][8][9]. On account of this, scientists are trying to engineer photocatalysts which are cost-effective, efficient, and chemically stable, as well as possessing suitable band potentials [9].
On the other hand, niobium pentoxide (Nb 2 O 5 ) has been proven as one of the favourite materials for heterostructures preparation with g-C 3 N 4 . It belongs to n-type semiconductors with a bandgap between 3.1 and 3.4 eV, suitable for photocatalytic reactions and energy storage [15,16]. Moreover, it exhibits highly acidic properties and water-tolerant characteristics required for photoelectrochemical devices and photocatalysis. To make this material active on visible light with a high efficiency, its heterostructures such as Nb 2 O 5 /ZnO, and Nb 2 O 5 /Cu₂O [17][18][19] have been considered in previous works. Along this line, we prepared Nb 2 O 5 /g-C 3 N 4 heterostructures which exhibit high efficiency for molecular H 2 generation.
Herein, a hydrothermal synthetic method was used in synthesizing Nb 2 O 5 /g-C 3 N 4 heterostructures. Compared to bare Nb 2 O 5 and g-C 3 N 4 , the heterostructure of Nb 2 O 5 /g-C 3 N 4 can produce high molecular H 2 , which is 75 mmol/h.g without methanol and 160 mmol/h. g with methanol. This efficient photocatalytic performance is owing to appropriate interfacial contact, which promotes the rapid photogeneration of electron-hole pairs at the Nb 2 O 5 /g-C 3 N 4 . The laser flash photolysis (LFP) technique was used as an extension to observe the trapping and lifetime of charge carriers to increase the generation of molecular H 2 . The investigation of Nb 2 O 5 /g-C 3 N 4 LFP has not been carried out so far to the best of our knowledge.
In the LFP technique, Nb 2 O 5 /g-C 3 N 4 was excited by a nanosecond pulsed laser for the photogeneration of active species, i.e., electron-hole pairs were monitored through transient absorption spectra. The utilization of laser pulses in this field initiated the development of transient species in flash photolysis experiments. And this is typically followed by a change in the original optical characteristics of the tested material, which was recognized by using a flash lamp. Trapping of charge carriers was observed in the presence and absence of a-holes scavenger. Trapped-hole peaks are often seen at lower wavelengths ranging from 420 to 450 nm, whereas trapped electron peaks are usually found at higher wavelengths ranging from 700 to 760 nm. However, no substantial peak could be detected on the transient spectra between 400 and 600 nm [20]. A significant increase of hole trapping was observed in the presence of a-holes scavenger, while the range of electron trapping remains the same. Additionally, a unified model equation is used to explain the decay curves of the observed transient absorption signal. Then the lifetime of Nb 2 O 5 /g-C 3 N 4 was calculated and compared with that of g-C 3 N 4 to show its usefulness for photocatalysts associated with the molecular H 2 evolution.

Experimental setup
Hydrothermal synthesis was used to prepare Nb 2 O 5 and Nb 2 O 5 /g-C 3 N 4 heterostructures, while a conventional method was employed for g-C 3 N 4 preparation as reported in our previous work [1].

Preparation of Nb 2 O 5
Nb 2 O 5 has been synthesized through the oxidant-peroxo route. 0.5 ml of HNO 3 was diluted by 20 ml of deionized water, followed by adding 1g of NbCl 5 during continuous stirring. The solution became milky white containing a few white and colourless precipitates at the bottom. Then 20 ml of H 2 O 2 was added drop-wise which gradually changed the color of the solution to yellow, indicating the formation of niobium peroxo-complex (NPC): [Nb(O 2 ) 4 ] 3− . The prepared solution was transferred into an autoclave and was kept in the oven for 24 h at 200 • C. After cooling down naturally, the prepared powder was separated through centrifugation for 30 min and thoroughly washed with deionized water. The obtained product was dried in a drying oven at 90 • C for 12 h. Afterwards, the Nb 2 O 5 powder was annealed at 200 • C for 2 h.

Preparation of heterostructures of Nb 2 O 5 /g-C 3 N 4
The preparation route of g-C 3 N 4 has been provided in the supporting information. 0.04g of g-C 3 N 4 was added to the prepared NPC solution for heterostructures. The obtained solution was placed in a suitable autoclave and heated for 24 h at 200 • C in an oven. After cooling down, the powder was separated from the liquid by centrifugation for 10 min, followed by drying in the oven. This approach was employed to make heterostructures with g-C 3 N 4 . The schematic illustration of the whole synthesis has been provided in the supporting information for clarification.

Incident photon-to-current efficiency (IPCE)
IPCE measurements can give an important insight into the contribution of photocatalysts in the conversion of incident photons into charge carriers. In this experiment, the light powder at each photo energy is measured and calculated to the IPCE value from the photocurrent according to equation (1).
Where J is the photocurrent density (mA⋅cm − 2 ), P mono is the light power density (mW⋅cm − 2 ) at λ. λ-wavelength of illuminating monochromatic photons.

Laser flash photolysis (LFP)
Nitrogen was purged on each sample for 25 min to remove air. The mixture of Nb 2 O 5 /g-C 3 N 4 was excited by an excimer Laser (LPX200) provided by Lambda physic and the pulse duration was 20 ns. The excitation wavelength for the first sample was 680 nm and for the second sample 700 nm, whereas its laser energy per pulse was 2-10 mJ. The illumination area of the laser beam and analyzing light from the sample was 0.6 × 0.196 cm 2 . The change of reflectance was monitored through the LFP spectrometer LKS 80 from Applied Photophysics. Xenon arc lamp 150W was used as a light source. When the light was fallen on the Nb 2 O 5 /g-C 3 N 4 sample, then the photogenerated charge carriers were produced. The reflected light was collected by the monochromator before it was sent to the photomultiplier detector (Hamamatsu R928 photomultiplier) to produce current. This change of reflectance is then recorded by the oscilloscope. The oscilloscope captured the transient with better sensitivity by adjusting the input voltage. The best part of the diffuse reflectance LFP technique is that the results obtained can be directly compared with the results of the photocatalytic experiments, although powdered samples are used in both processes. Experimental conditions for various time scales are provided in Table 1.

Photocatalytic molecular hydrogen (H 2 ) formation
The molecular hydrogen (H 2 ) evolution process was conducted using the prepared photocatalysts in the presence of a scavenger and without a scavenger. Methanol was used as a hole scavenger. In a typical experiment, 0.02 g of the prepared photocatalysts were mixed with 40 ml of deionized water. With the addition of methanol, a suspension containing 10 vol% of a-holes scavenger was obtained. The prepared solutions were degassed with Ar for 30 min and then stirred for a further 25 min. The photoreactor was then put under a solar-simulating Xenon light source (1000 W, 1.5 G) for 8 h. A homemade cooling system was used to keep the temperature constant. A gas chromatograph was used to measure the evolved H 2 after every 1 h.

Material characterizations
The as-synthesized materials XRD data was recorded by a Bruker (D8 Advance) analyzer using Cu Kα (α = 0.15406 nm) radiation. The morphologies were analyzed using scanning electron microscopy (JEOL JSM-6700F, JEOL, Tokyo, Japan) equipped with a Lower Secondary Electron Image (LEI) detector. Additionally, the samples were characterised by transmission electron microscopy (TEM) using a 200 kV FEI TecnaiG2 F20 microscope. Micrographs were obtained in two modes: bright field (BF) and selected area electron diffraction (SAED). The UV-vis absorption spectrum was measured using a Varian Cary. Shimadzu 8A (Shimadzu, Kyoto, Japan) gas chromatograph equipped with a TCD detector and a 5 Å molecular sieve packed column; Ar as the carrier gas was used.

Structural and morphological analysis
Structural analysis was conducted by XRD for g-C 3 N 4 , Nb 2 O 5 before annealing (BA) and after annealing (AA) and Nb 2 O 5 /g-C 3 N 4 as shown in Fig. 1. Their corresponding Miller indices were also provided in Fig. 1. XRD measurements for Nb 2 O 5 -BA exhibited a combination of niobic acid and amorphous Nb 2 O 5 , therefore, slight distortion from the original crystalline structure was observed. While after annealing at 200 • C, the peaks were well-matched with Nb 2 O 5 orthorhombic phase (T-Nb 2 O 5 ): ICCDS#30-0873 (a = 6.17 Å, b = 29.97 Å and c = 3.93 Å). The lattice constants were calculated by using equation (2) and were found in good agreement with the T-Nb 2 O 5 (a = 5.96 Å, b = 29.26 Å and c = 3.89 Å) except 'a' was found smaller. This change possibly be due to changed morphology.
XRD of g-C 3 N 4 shows two peaks observed at 13.25 • and 27.8 • that can be indexed as (100) and (002), respectively. The (100)   diffraction plane corresponds to the characteristic interplanar system and the (002) plane corresponds to inter-layer structural packing. XRD pattern of Nb 2 O 5 exhibited mild and broad diffraction peaks indicating its amorphous behaviour. In comparison to pure Nb 2 O 5 and g-C 3 N 4 , Nb 2 O 5 /g-C 3 N 4 composites showed a combination of both characteristic peaks. High characteristic peak of g-C 3 N 4 (002), suggesting a major pairing between g-C 3 N 4 and Nb 2 O 5 . Energy-dispersive X-ray spectroscopy (EDX) analysis of the Nb 2 O 5 /g-C 3 N 4 heterostructures is given in Figure S1 (a-b) of the supplementary material, which confirmed the existence of the elements Nb, C, O, and N for the prepared sample. Moreover, it confirmed the Nb 2 O 5 formation for amorphous. EDX measurements revealed that Nb 2 O 5 agglomerates interact strongly with g-C 3 N 4 . The FTIR spectra of g-C 3 N 4 , Nb 2 O 5 and Nb 2 O 5 /g-C 3 N 4 were shown in Fig. 2. The spectrum of pure g-C 3 N 4 presented all bands described in the literature [21][22][23] with intensive bands at 808 cm − 1 , 886 cm − 1 and in the range of 1122-1632 cm − 1 , which are assigned to out-of-plane bending vibration of heptazine rings, deformation mode of N-H bonds and stretching modes of C-N heterocycles, respectively. These are also known as g-C 3 N 4 heterocycles which has been marked in Fig. 2(a) Table 2. The FTIR spectra of the Nb 2 O 5 /g-C 3 N 4 heterostructures exhibited characteristic peaks of both phases as marked in Fig. 2 Table 3).
SEM images of as-synthesized Nb 2 O 5 and the Nb 2 O 5 /g-C 3 N 4 heterostructures are presented in Fig. 3(a-b), respectively. The SEM image of g-C 3 N 4 is provided in Fig. S2 of the supporting information. The hierarchical nanospheres of Nb 2 O 5 and Nb 2 O 5 /g-C 3 N 4 heterostructures could be observed. A TEM analysis of Nb 2 O 5 /g-C 3 N 4 was performed as shown in Fig. 3(c-d). Lattice fringes were found for Nb 2 O 5 and g-C 3 N 4 as shown in the HRTEM image of Fig. 3 (d). SAED pattern has also been provided in Fig. 3 (c-inset). The BET surface areas of Nb 2 O 5 , g-C 3 N 4 , Nb 2 O 5 /g-C 3 N 4 before annealing and Nb 2 O 5 /g-C 3 N 4 after annealing at 200 • C have been provided in Table S1. The specific surface area (SSA) decreased after heterostructure formation with g-C 3 N 4 . Annealing processes further reduce the SSA which indicates a decrease in the surface defects.

UV-vis and IPCE spectroscopy
Bandgap energies and corresponding absorption spectra of g-C 3 N 4 , Nb 2 O 5 , and Nb 2 O 5 /g-C 3 N 4 were shown in Fig. 4. The bandgap energies were estimated by the absorption coefficient (inset) of Tauc's plot as shown in Fig. 4 (a). g-C 3 N 4 and Nb 2 O 5 possess indirect transition band gaps, i.e., 2.56 eV and 3.1 eV, respectively. In addition, the bandgap of Nb 2 O 5 /g-C 3 N 4 heterostructures was red-shifted to 2.52 eV. The observed band gap helped to increase the photocatalytic efficiency of the prepared heterostructures. IPCE (Incident Photon to Current Conversion Efficiency) spectra of as-prepared electrodes were measured in electrolyte without any sacrificial reagents and conducted with a standard two electrodes in a short circuit. The corresponding IPCE spectra of bare Nb 2 O 5 , g-C 3 N 4 , Nb 2 O 5 /g-C 3 N 4 and FTO are presented in Fig. 4 (b). Nb 2 O 5 /g-C 3 N 4 exhibited the highest IPCE values along the scanned wavelength in the UV region in comparison to Nb 2 O 5 and g-C 3 N 4 . The reduced recombination rate of photoinduced electrons and holes provides a meaningful contribution to an increment of IPCE for the Nb 2 O 5 /g-C 3 N 4 electrode. Its photoactivity is also observed in the visible light region from 400 nm to 450 nm, which IPCE value is higher than that of Nb 2 O 5 . These results indicated that the introduction of g-C 3 N 4 not only highly enhances the photoelectric conversion ability of Nb 2 O 5 /g-C 3 N 4 in the UV light region, but also extends its light capture into the visible light region.

Charge carriers trapping in Nb 2 O 5 /g-C 3 N 4
Nb 2 O 5 /g-C 3 N 4 bandgap is 2.52 eV and by IPCE spectra the maximum was observed at 350 nm, therefore, the irradiated laser pulse of λ ex = 320 nm was used. The transient absorption spectra of Nb 2 O 5 /g-C 3 N 4 were observed for different wavelengths ranging between 400 and 800 nm against ΔJ. The experiment was performed under inert conditions (N 2 environment) and the recombination of pure charge carrier kinetics was observed with and without a hole-scavenger (methanol). After a full spectrum scan ranging from 400 to 850 nm two different values of time was used to investigate charges in the transient spectra, as shown in Fig. 5. Without methanol, the transient absorption has shown a random pattern as shown in Fig. 5(a). With methanol, the transient absorption strongly increases at 680 nm and 720 nm, while decreasing at 425 nm as shown in Fig. 5 (b). This strategy employed that the trapped holes have transient absorption peaks around 420-450 nm and peaks of trapped electrons ranging from 700 to 760 nm at higher wavelengths. Therefore, we further observed the transient absorption spectra at 680 nm and 720 nm to calculate the lifetime of electrons which are possibly taking part in the increased photocatalytic efficiency for H 2 generation.

Transient absorption kinetics and second-order fit on transient signals
To observe the transients in the oscilloscope, the Beer-Lambert law was applied by using equation (3). Herein, photogenerated species were detected by the fraction of transmitted light.
Where I 0 is the incident intensity and I t is the fraction of light measured after interaction with the powder sample. Fig. 6 represents the transient absorption kinetics at (a) 680 nm and (b) 700 nm in an N 2 -methanol atmosphere for Nb 2 O 5 /g-C 3 N 4 after 0.078 μs in the 10 μs measurements. The full transient signal can be seen in the inset of Fig. 6(a-b). In the beginning, there is a negative peak associated with scattering from UV lasers, which was used as the source of excitation. The 40 Ω terminal resistance enables a short period of rising and allows the signal to be easily detected after 0.078 μs. The 5 kΩ high resistance used for the 20 μs measurements eventually led to a much longer detector rise time. Before kinetic analysis, the first 0.08 μs of observations were removed as they did not reflect the transient signal.
The data point was reduced to 897 for the easy processing of the transient signal by removing 103 points. The sample reflecting changes was determined by equation (4).
As predicted that LFP was used to observe the recombination of charge carriers corresponding to second-order kinetics [22]. Many Table 3 Results of nonlinear second-order kinetic fit at 680 and 720 nm for Nb 2 O 5 /g-C 3 N 4 after excitation with 2-10 mJ/cm 2 .  scientists worked on this kinetics, but Tamaki et al. [23] are the ones who provided the detailed kinetic analysis regarding second-order processes. However, they focused only on signals ranging from picosecond (ps) to nanosecond (ns). In the present work, a second-order kinetic fit was used, and the signal ranges have been taken from μs-ns. A second-order fit function that we have used is provided in equation (5) [24]. The fitted model has been shown in Fig. 6(c-d).
Where Y is the second-order adsorption, A is the altitude at which transient signals are detected, k is the recombination constant, B is the residual signal after fitting, X is the varying parameter used for fitting and t is the time taken by the curve of decay (μs-fs). By using the above equation, charge carrier signals were fitted with the obtained parameters by taking three different percentage change values (ΔJ %) at 680 nm and 720 nm. This model explains the enhanced performance of the composite material in photocatalytic hydrogen production experiments.   we have calculated the lifetime of g-C 3 N 4 . After plotting the values, we plotted a graph between these values and then integrated this graph by clicking on the analysis, mathematics and then integrated the function. After performing all these steps correctly, we finally generated the area A 1 = 3.79384980 × 10 − 10 which was taken in the numerator of equation (5). Then we plotted another graph by doing the same process as mentioned above and calculated the area A 2 = 1.1986165 × 10 − 06 which was taken in the denominator of equation (5). By dividing A 1 by A 2 , the average lifetime of the decay curve was calculated as 3.1651897 μs. Similarly, the average lifetime of Nb 2 O 5 /g-C 3 N 4 was calculated as 6.54165μs by using the A 1 = 3.88509 × 10 − 13 , A 2 = 5.939 × 10 − 08 .

Photocatalytic evolution of molecular H 2
Detailed research was conducted to determine the effect of various factors on the rate of molecular hydrogen evolution. Initially, photocatalytic production of molecular H 2 for P25, g-C 3 N 4 , Nb 2 O 5 , and Nb 2 O 5 /g-C 3 N 4 was performed with and without methanol. Fig. 7(a-b) exhibits the evolved H 2 , while Fig. 7(c-d) the overall comparison and total rate of H 2 evolution with and without methanol. Heterostructures prepared with and without methanol show a higher molecular H 2 than P25, g-C 3 N 4 , and Nb 2 O 5 . An increase in the H 2 production rate in the presence of methanol is evidence of efficient interfacial charge separation. Nb 2 O 5 /g-C 3 N 4 heterostructures generated the highest hydrogen production rate (22.75 mmol/h g) due to their wide surface area and suitable band positions. Thus,  Nb 2 O 5 /g-C 3 N 4 heterostructures with enhanced photocatalytic activity were produced by loading g-C 3 N 4 as an electron-hole separator.

Photocatalytic mechanism of g-C 3 N 4 /Nb 2 O 5
Direct Z-scheme charge transfer favoured the band positions of Nb 2 O 5 and g-C 3 N 4 . As shown in Fig. 8, under solar light electrons on the surface of the Nb 2 O 5 conduction band (CB) moved to the g-C 3 N 4 valence band (VB) and then recombined with holes. Electrons on the CB of g-C 3 N 4 react with H + to generate molecular H 2 . The holes on the VB of Nb 2 O 5 can undergo possible oxidation reactions. Compared to pure g-C 3 N 4 , the Nb 2 O 5 /g-C 3 N 4 heterostructures exhibited higher photocatalytic activity for molecular hydrogen production due to the longer lifetime of photogenerated charges. These results indicated the generation of efficient heterojunctions between the g-C 3 N 4 and Nb 2 O 5 phases, which promoted charge carrier transfer and enhanced the lifetimes of the photogenerated charges. The proposed Z-Scheme photocatalysis mechanism has been provided in Fig. 8. The calculated band edges by Mott-Schottky analysis have been provided in Fig. S4.

Conclusions
Hydrothermal synthesis was employed to produce Z-scheme Nb 2 O 5 /g-C 3 N 4 heterostructures, which reported the high molecular H 2 evolution with and without a hole scavenger as compared to pure g-C 3 N 4 and Nb 2 O 5 . These heterostructures exhibited higher photocatalytic activity due to the longer lifetime of photogenerated charge carriers, where the lifetime was calculated through the Laser Transient Absorption Spectra. This study was conducted first time for Nb 2 O 5 /g-C 3 N 4 -prepared heterostructures. The contribution of hole scavenger for possible charge trapping and increased H 2 evolution was demonstrated. The laser flash photolysis study in this work showed that the hole and electron trapping occur in the wavelength ranges ~440 nm and ~750 nm in the absence of methanol. But, in the presence of methanol, hole trapping strongly increased whereas electron trapping remains the same. A second-order kinetic fit was applied to describe the recombination of charge carriers. Our model well predicts the charge carrier signal in Nb 2 O 5 /g-C 3 N 4 that was seen in the nanosecond to microsecond time scale after being excited by a laser flash. Meanwhile average lifetimes of both g-C 3 N 4 (~3.1651897 μs) and Nb 2 O 5 /g-C 3 N 4 (~6.54165 μs) were shown, we observed a higher rate of molecular H 2 evolution for Nb 2 O 5 /g-C 3 N 4 heterostructures.