The development of clean and renewable energy resources has been one of the hot subjects in recent decades due to the global energy crisis and environmental issues [1, 2]. Semiconductor mediated photocatalysis to produce H2 is a promising method to solve this problem [3-6]. In the past decades, many efforts have been made to find suitable semiconductors for this purpose, such as TiO2 [7-9], CdS [10-14], TaON [15, 16]. Recently, metal-free nanocarbon-based photocatalysts have attracted increasing interest due to their following advantages: low cost, environmental friendly, and good stability [17-21]. Among them, graphene quantum dots (GQDs) have been suggested as a promising alternative to the conventional semiconductor quantum dots because of their unique properties such as high water-solubility, low toxicity, excellent biocompatibility, and high photostability [22-24]. Moreover, the band gaps of GQDs can be flexibly tuned by controlling the particle size and ligand species [25, 26].
So far, bare metal-free GQDs have not been explored as a good photocatalyst for H2 production under visible light, which is probably due to its short excitation lifetime, leading to fast electron-hole recombination, and thus poor photocatalytic efficiency [27, 28]. The GQDs coupled with an efficient cocatalyst can probably solve this problem because of the existing fact of improved charge transfer from a semiconductor to a cocatalyst [29, 30]. Traditionally, noble metals (such as Pt, Ru, and Pd) are regarded as the best cocatalysts for photocatalytic H2 evolution when integrated with a semiconductor [31, 32]. However, their high costs make it less attractive for practical application. In our previous studies [33-36], we found that metal phosphides could serve as good cocatalysts to enhance photocatalytic H2 production. Among them, Ni2P nanoparticles (NPs) were proved to be the most promising candidate as cocatalyst considering its good electrical conductivity and simple synthesis process [37, 38].
Inspired by our previous studies, herein we report the use of water-soluble OH-functionalized graphene quantum dots (OH-GQDs) as the photosensitizer coupled with noble-metal-free Ni2P NPs as the cocatalyst for visible-light driven H2 production. Under optimal conditions, the highest photocatalytic H2 evolution rate can reach 1567 µmol·h–1·g–1, which is ~94 times higher than that of bare OH-GQDs, suggesting that metal-free photosensitizer OH-GQDs coupled with noble-metal-free Ni2P cocatalyst is efficient for photocatalytic H2 evolution.
All chemicals, including Pyrene (C16H10, 98.0%), Sodium hydroxide (NaOH, 96.0%), nitric acid (HNO3, 65~68%), ethylenediamine (C2H4(NH2)2, 99.0%), nickel nitrate hexahydrate (Ni(NO3)2·6H2O, 98.0%), triethanolamine (TEOA, 99%), ascorbic acid (VC, 99.7%), methanol (CH3OH, 99.5%), sodium sulfide nonahydrate (Na2S·9H2O, 99.0%), and anhydrous sodium sulfate (Na2SO3, 97.0%), were commercially available (Aldrich and Innochem) and used without further purification.
OH-GQDs were prepared according to a previous report with some modifications [39]. In a typical procedure, 2.0 g pyrene was nitrated into trinitropyrene in 160 mL hot HNO3 at 80 ℃ under refluxing and stirring for 12 h. After cooling to room temperature, the trinitropyrene was collected by centrifugation. The resulting yellow precipitates were dispersed in 30 mL NaOH solution (0.2 mol/L) by ultrasonication for 150 min. The suspension was transferred to a Teflon-lined autoclave (50 mL) and heated for 10 h at 200 ℃. After cooling to room temperature, the product containing water-soluble OH-GQDs was filtered through a 0.22 mm microporous membrane and then dialysed in a dialysis bag for 48 h. After that, the purified black OH-GQDs were dried at 80 ℃ under vacuum.
Ni2P NPs were synthesized according to our previous study [35]. Typically, a certain amount of Ni(NO3)2·6H2O and yellow phosphorus (1:5) were dissolved in ethylenediamine and stirred for 30 min. Then, the mixed solution was transferred to a 50-mL Teflon-lined, stainless-steel autoclave and maintained at 140 ℃ for 12 h. After cooling to room temperature, the product was collected and washed with benzene, ethanol, and distilled water three times each.
The crystal structures of the OH-GQDs and Ni2P were characterized by powder X-ray diffraction (XRD, D/m/ax-TTR Ⅲ) using graphite monochromatized Cu Kα radiation of 0.154178 nm, operating at 40 kV and 200 mA. The scanning range was from 10° to 70° (2θ) at the speed of 10° min–1. Scanning electron microscopy (SEM) was performed on JSM-6700F. Transmission electron microscopy (TEM) images was collected on a JEM-2010 electron microscope, operated at an acceleration voltage of 200 kV. The chemical composition was analyzed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250 X-ray photoelectron spectrometer). Ultraviolet-visible (UV-vis) diffuse reflection spectra were investigated by using an UV-visible spectrophotometer (SOLID 3700 UV-vis spectrometer) over the range of 300 to 800 nm. The photoluminescence (PL) spectra were recorded on a FluoroMax-4 spectrometer at ambient temperature. Fourier transformed-Infrared spectroscopy (FTIR) spectra of dried samples were studied on a Thermo Fisher Scientific instrument (Nicolet iS10).
Photocatalytic H2 evolution experiments were carried out in a 50 mL round-bottom flask at room temperature. A 300 W Xe-lamp equipped with a 420 nm cut-off filter was used to provide the visible light irradiation. In a typical run, 6 mg of the photocatalyst and 1.2 mg of the cocatalyst were dispersed in a mixed solution containing 18 mL water and 2 mL TEOA. Prior to irradiation, the solution was bubbled with high purity nitrogen for 15 min to remove the air inside. After that, 5 mL of methane was injected into the flask as the internal standard. The amount of produced H2 was calculated by a gas chromatograph (GC) equipped with a thermal conductivity detector (TCD) detector. For long-term irradiation, 20 mg of the photocatalyst and 4 mg of the cocatalyst were dispersed in a mixed solution containing 45 mL water and 5 mL TEOA and a 250 mL flask were used.
The apparent quantum yield (AQY) was measured using a 300 W Xe-lamp equipped with a 420 nm (±5 nm) band-pass cut filter. The light intensity was approximately 6.8 mW cm–2 and the illuminated area for the reactor was approximately 27 cm2. The AQY was calculated based on the following equation (1):
The synthesis process of the OH-GQDs is shown in the Fig. 1 [39]. Typically, pyrene was nitrated in hot HNO3 solution, followed by hydrothermal treatment in NaOH aqueous solution, and the as-prepared solution containing OH-GQDs was filtered. The filtrate was then dialysed in a dialysis bag for 48 h and dried at 80 ℃ under vacuum. The OH-GQDs solution emits strong green fluorescence when irradiated under 365 nm UV light.
The structure of the OH-GQDs was characterized by XRD. The characteristic feature of graphite with a (002) layer spacing is clearly observed (Fig. 2(a)), which is consistent with that of bulk graphite (0.33 nm) [40]. Moreover, Fig. 2(b) shows the FTIR spectra of bare OH-GQDs. The peak at 1618 cm–1 can be assigned to the C=C bond stretch, and a strong, rather broad vibration at 3448 cm–1 is attributed to the O–H bonds [41]. Noticeably, a peak at 1271 cm–1 is also observed, corresponding to the vibration of C–OH [42, 43]. All these results confirmed that the OH-GQDs were successfully synthesized. Furthermore, the surface chemical composition of the OH-GQDs was also studied by the EDX spectra. Fig. 2(c) shows the existence of C, O, and Cu elements. It should be noted that Cu element should be from the Cu grid substrate. In addition, the morphology and the particle size of the OH-GQDs were investigated by TEM. As shown in Fig. 2(d), the OH-GQDs particles are well dispersed with uniform lateral sizes at 3.6 ± 0.5 nm. In addition, the XRD pattern of the as-prepared Ni2P cocatalyst shows good crystallinity (Fig. S1(a)). Fig. S1(b) is the SEM image of pure Ni2P, in which small nanoparticles aggregated together with an average size of ~300 nm were observed. The TEM image of Ni2P/OH-GQDs reveals that OH-GQDs are distributed randomly on Ni2P (Fig. 3).
The light absorption of the OH-GQDs covers almost the entire visible light region, which make it very interesting for photocatalysis. The UV-vis absorption spectrum of the OH-GQDs sample is shown in Fig. 4(a). As can be seen, the absorption spectrum of the OH-GQDs have two pronounced peaks, maximized at about 350 and 490 nm, with an optical absorption edge extended to ~750 nm [44]. Its excellent absorbance in the visible region demonstrates great potential to generate H2 under visible light [45]. Moreover, PL spectra of the OH-GQDs were measured under different excitation wavelengths. As shown in Fig. 4(b), when the excitation wavelength changes from 315 to 390 nm, the PL spectra have nearly the same features and show a strong emission peak at ~540 nm, suggesting a band gap of about 2.3 eV [46].
To identify the surface chemical compositions and valence states of the photocatalyst and cocatalyst, XPS was used to measure the OH-GQDs and Ni2P samples. The high resolution C 1s XPS spectrum shows four peaks at the binding energy (BE) of 284.8, 285.7, 288.4, and 289.8 eV (Fig. 5(a)), which are ascribed to the strong signals of C=C, C–O, C–OH and C=O, respectively [39, 47, 48]. Meanwhile, the high-resolution O 1s XPS spectrum reveals the presence of O–H at 531.6 and 533.1 eV (Fig. 5(b)). The peak at 531.6 eV belongs to OH-GQDs [39] and the signal at 533.1 eV is assigned to the absorbed water [49, 50]. The XPS results further confirmed the successful synthesis of OH-functionalized GQDs. Moreover, the XPS spectrum of Ni 2p (Fig. 5(c)) shows two main typical peaks (857.1 and 874.8 eV) with satellite peaks (marked as "Sat.") and a very weak peak at 853.3 eV is attributed to Niδ+ of Ni2P [51]. The peaks at the BE of 857.1 and 874.8 eV, along with their satellite peaks, correspond to the Ni2+ originating from surface oxidation state. Fig. 5(d) gives the high resolution P 2p XPS spectrum. The peaks at the BE of 129.3 and 130.2 eV can be ascribed to the reduced phosphorus in the form of metal phosphides, while the peak at 134.1 eV can be assigned to oxidized P species due to air exposure [52]. The XPS results further proved the formation of Ni2P with surface partially oxidized to phosphate.
Based on the above characterizations and analysis, the photocatalytic H2 production experiments of the present OH-GQDs coupled with Ni2P cocatalyst (Ni2P/OH-GQDs) were further evaluated in aqueous solution under visible light (λ > 420 nm) [53]. Firstly, the effect of different sacrificial agents was investigated (0.75 mol/L ascorbic acid, pH = 4.2; 0.75 mol/L methanol; 0.75 mol/L TEOA; 0.31 mol/L Na2S/0.44 mol/L Na2SO3). The results show that only the TEOA solution can produce apparent H2, indicating that the weak alkaline condition might be favorable for photocatalytic H2 production using Ni2P/OH-GQDs (Fig. 6(a)). In order to explore the reason why only the TEOA solution can produce apparent H2, the PL spectra were measured. The concentrations of OH-GQDs for four different sacrificial agents are the same. As shown in Fig. S2(a), the emission intensity of Ni2P/OH-GQDs (black line) was significantly quenched when TEOA was added (red line), indicating that the electron-hole recombination can be efficiently suppressed in the presence of TEOA. However, for methanol and Na2S/Na2SO3, the PL intensities only slightly decreased (Fig. S2(b) and (c)). Interestingly, in the presence of VC (Figure S2d), the PL spectrum has been interrupted and damaged, which is probably caused by the VC solution. Based on the PL results, we can conclude that TEOA is the most efficient electron donor, which can quickly capture photo-generated holes and thus promoting the H2 evolution reaction.
Fig. 6(b) displays the rate of H2 production using bare OH-GQDs, bare Ni2P, Ni2P/OH-GQDs, and 1 wt% Pt/OH-GQDs samples in 0.75 mol/L TEOA solution [54]. As can be seen, Ni2P itself is not active for H2 evolution under visible light. Moreover, as for bare OH-GQDs, trace H2 production was detected, although it has a strong absorption band in visible region. The poor activity of pure OH-GQDs could be ascribed to the fast recombination of photo-generated electron-hole pairs [55]. Interestingly, it was found that an appropriate amount of Ni2P can significantly enhance the photocatalytic activity for H2 production using Ni2P/OH-GQDs sample. Noticeably, the rate of H2 evolution initially increased with the increasing amounts of Ni2P and then decreased with more amount of Ni2P. The highest photocatalytic H2 evolution rate of ~1567 µmol·h–1·g–1 was obtained when 20 wt% Ni2P was used, which is comparable to the activity of 1 wt% Pt/OH-GQDs (~1683 µmol·h–1·g–1). When the weight ratio of Ni2P reached more than 20 wt%, the H2 evolution rate decreased (~1000 µmol·h–1·g–1 for 30 wt%), probably resulting from the fact that more Ni2P per unit area can shied the visible light absorption of OH-GQDs, leading to the decrease of photons absorbed. Interestingly, the present photocatalytic reaction system also showed good photocatalytic activity when the reaction solution was exposed to air (Fig. 6(c)). All the above results suggest the importance of noble-metal-free Ni2P cocatalyst for the improvement of H2 evolution over OH-GQDs. To the best of our knowledge, this is the first time to explore the photocatalytic activity of OH-GQDs for H2 production coupled with a noble-metal-free cocatalyst.
The apparent quantum yield (AQY) for H2 evolution over the 20 wt% Ni2P/OH-GQDs photocatalyst was performed in a system containing 0.75 mol/L TEOA under a monochromatic light at 420 nm (±5 nm). As shown in Fig. 6(d), the AQY increases with the irradiation time, especially in the initial stage, which may be due to an induction period and the dissolved H2 in the solution [35]. H2 was not detected until the sample was irradiated for 2 h. The AQY is only 0.2% for the second hour and then reaches to 0.64% with a H2 evolution rate up to ~428.5 µmol·h–1·g–1 after 8 h of irradiation. The average AQY is 0.46% in the evaluated time. Meanwhile, the stability of the Ni2P/OH-GQDs under a 300 W Xe lamp with a 420 nm (±5 nm) band-pass filter was also studied (Fig. 7(a)). No significant decrease for photocatalytic H2 production was observed. After 28 h of irradiation, a total amount of ~199.8 µmol H2 was produced at a constant rate of ~440 µmol·h–1·g–1, suggesting excellent durability of the Ni2P/OH-GQDs for photocatalytic H2 production under visible light.
In order to explore the reaction mechanism for photocatalytic H2 production over Ni2P/OH-GQDs, the Mott-Schottky plot and PL measurements were examined to study the photoinduced electron transfer behavior. Fig. 7(b) shows the Mott-Schottky plot of OH-GQDs in a H2SO4 solution (1.0 mol/L) and the positive slope indicates n-type conductivity [56]. Moreover, the Fermi levels (EFn) of the OH-GQDs could be gained from the X-intercept by extrapolating the M-S plot to 1/C2= 0, and the values for the sample is about –0.52 V vs. NHE, which is consistent with previous reports [27, 28]. Therefore, the conduction band of the OH-GQDs is more negative than –0.52 V vs. NHE (EFn). In the previous studies, the photo-excited carriers can be swiftly transferred from the CB of CdS (–0.5 V vs. the NHE) to Ni2P, which acts as an excellent cocatalyst [57, 58]. Since the CB edge of OH-GQDs is more negative than that of CdS, the photo-excited electrons on OH-GQDs could be transferred to Ni2P through the Ni2P/OH-GQDs interface. The emission intensity was obviously quenched upon adding Ni2P to the OH-GQDs solution (Fig. 7(c)), indicating Ni2P can suppress the electron-hole recombination [59, 60]. According to the aforementioned results, a probable reaction mechanism for the enhanced photocatalytic activity of the Ni2P/OH-GQDs is proposed. As shown in Scheme 1, OH-GQDs (Eg = 2.3 eV) can effectively absorb visible light to excite electrons located at the valence band (VB) to the conduction band (CB), accompanied by the creation of photogenerated electron-hole pairs [61]. In addition, Ni2P shows metallic character and can not absorb visible light to produce electron-hole pairs [62]. Due to the introduction of Ni2P on OH-GQDs, a typical metal-semiconductor interface was formed. Then, the photo-generated electrons in the CB of OH-GQDs rapidly moved to the surface of the Ni2P NPs and the holes could be consumed by TEOA, and thus promoting separation of the electron-hole pairs and suppressing their recombination. Finally, the electrons on the surface of Ni2P will catalyze the reduction of protons to produce H2, resulting in the improved photocatalytic H2 production activity.
In summary, we studied the use of OH-GQDs as a metal-free photosensitizer for photocatalytic H2 evolution under visible light. The photocatalytic activity can be significantly enhanced by the addition of noble-metal-free Ni2P cocatalyst. Under optimal conditions, the photocatalyst sample exhibited a maximum H2 evolution rate of 1567 µmol·h–1·g–1, which is ~94 times higher than that of bare OH-GQDs and comparable to the catalytic activity of 1 wt% Pt/OH-GQDs sample (1683 µmol·h–1·g–1). All the results indicate that Ni2P can efficiently promote the separation of photo-generated charge carriers and highly enhance photocatalytic H2 evolution of OH-GQDs. This present study demonstrates that the OH-GQDs is an efficient metal-free catalyst for photocatalytic H2 evolution coupled with a suitable cocatalyst.
2018,
Vol. 39 Issue (11): 1753-1761 
PDF
2018,
Vol. 39 Issue (11): 1753-1761 PDF