Technology requirements for a sustainable society lead to the surge of clean and renewable energy research in the 21st century. Compared with other chemical fuels, hydrogen (H2) as a new type of clean energy exhibits the highest specific energy of combustion [1-4]. The photocatalytic H2 generation from water splitting based semiconductor technology is proved to be a sustainable solution to resolve the current environmental and energy issues [5-7]. Various semiconductor-based photocatalysts (such as TiO2 [8], ZnO [9], CdS [10], and g-C3N4 [11] etc.) have been developed for the photocatalytic H2-evolution application. Among these reported semiconductors, g-C3N4, as a new class of nonmetallic organic polymer semiconductor, has attracted huge interest due to inimitable layer structure, suitable energy structure (2.7 eV), simple fabrication and outstanding stability [12-14]. However, single g-C3N4 demonstrates low activity owing to the poor utilization of photoexcited charge carriers caused by the weak van der Waals force [15].
A great deal of approaches have been proposed to improve the photocatalytic performance of g-C3N4, for example, liquid exfoliation of bulk g-C3N4, surface modification, elemental doping, construction of heterojunction with other semiconductors or loading of co-catalysts [16-22]. Heteroatoms doping (nonmetal elements such as P, S and N etc.) can tune energy bands of semiconductor at atomic level, which not only improves light absorption, but boost photogenerated-carrier separation [23-26]. For example, Guo et al. [27] employed phosphorus acid to fabricate P-doped g-C3N4 with hexagonal tubular morphology; Ran et al. [28] prepared P-doped g-C3N4 using 2-aminoethylphosphonic acid and melamine as P source and g-C3N4 precursor, respectively; and Bellardita and coauthors [29] synthesized P-doped g-C3N4 using NH4H2PO4 as P source. For enhancing the photocatalytic activity, cocatalyst is considered more efficient to speed up the transportation an separation of photoinduced carriers [30, 31]. Usually, the photocatalytic cocatalysts could be divided in two types. The first type of cocatalyst is called photogenerated electron-transfer (usually noble metals, Pt, Pd and Au) acting as electron sink and the second type of cocatalyst is called hole-transfer (MoO3, MnO2 and CoO etc.) [32, 33]. Keller et al. [34] fabricated an Au/TiO2-g-C3N4 nanocomposite photocatalyst which can achieve to produce hydrogen utilizing ultra low amounts of sacrificial agents under the action of the Au electron cocatalyst. Kang et al. [35] confirmed that CoO/g-C3N4 heterojunction exhibited higher overall water splitting performance due to the presence of CoO as a hole cocatalyst. More importantly, Xing et al. [36] have fabricated a novel MnOx@CdS/CoP catalyst in which MnOx act as a hole-transfer cocatalyst and CoP act as an electron-transfer cocatalyst, thus accelerating the photocatalytic activity by spatially separating the reduction surface. Yang et al. [37] reported excellent photocatalytic hydrogen evolution by spatial separation of dual cocatalysts (Au and CoP) to accelerate charge transfer in 3D ordered microporous g-C3N4. From previously reported results, it can be suggested that simultaneous modification of g-C3N4 by dual cocatalysts containing electron-transfer and hole-transfer cocatalysts may accelerate the transportation and separation of photogenerated carriers more efficiently, resulting in higher photocatalytic activity.
Herein, we employed low-temperature phosphating method to fabricate Co(Ⅱ)-modified P-doped g-C3N4 nanosheets via simultaneous doping of P atom and coupling of Co(Ⅱ) as hole cocatalyst. In the process of photocatalytic hydrogen production, Pt nanoparticles (NPs) as electron cocatalyst could be deposited on the Co(Ⅱ)-modified P-doped g-C3N4 nanosheets. Furthermore, the synergistic effect of Pt and Co(Ⅱ) cocatalysts for the enhanced photocatalytic performance of P-doped g-C3N4 photocatalyst is studied in detail. It has been supposed that the P-doping could optimize the band energy of g-C3N4 nanosheets to increase visible-light absorption, while Pt electron-transfer cocatalyst and Co(Ⅱ) hole-transfer cocatalyst could accelerate the transportation and separation of photoinduced carriers to improve the photocatalytic activity.
Bulk g-C3N4 powders were fabricated using the method of thermal polymerization of urea in accordance to previous literature [38]. The bulk g-C3N4 powder (100 mg) was ultrasonically dissolved in 200 mL deionized H2O to obtain g-C3N4 nanosheets solution for further use. The schematic illustration for fabrication of Co(Ⅱ)/PCN composites is shown in Scheme 1. Firstly, a certain amount of Co(NO3)2·6H2O was added into deionized water (200 mL) containing 100 mg of g-C3N4 nanosheets and then the mixture solution was stirred for another 2 h to result in the adsorption equilibrium of cobalt ions. Then the mixture was collected by centrifugation and washed several times with deionized water and ethanol, followed by drying at 60 ℃ for 12 h. Subsequently, 50 mg of Co(Ⅱ)/g-C3N4 nanosheets and 500 mg of NaH2PO2 powder were put into the closed porcelain boat, respectively. Afterwards, the samples were put in a quartz tube of the furnace, maintained at 300 ℃ for 2 h with a heating rate of 2 ℃ /min in an inert atmosphere (Ar flow). After cooling to room temperature, the obtained powders were collected, which were labeled as Co(Ⅱ)/PCN-x (x = 0.5, 2.0, 4.0, 6.0, 8.0), where x indicates the weight percentage of Co(Ⅱ) in the g-C3N4 nanosheets. P doped g-C3N4 nanosheets were fabricated by the same method except the Co(Ⅱ) addition and labeled as PCN.
The phase, morphology, composition of the as-synthesized materials were characterized by a Ultima Ⅳ diffractometer (Rigaku) with Cu Kα radiation (λ = 1.5406 Å), a field-emission transmission electron microscopy (Tecnai G2 F30 S-TWIN, FE-TEM), and a ray photoelectron spectroscopy (XPS) spectra (G Scientific ESCALAB Mark Ⅱ spectrometer), respectively. The optical properties and the electron spin resonance experiments were characterized by a UV-vis diffuse reflection spectroscopy (DRS, Shimadzu UV2600) and a X-band frequency on a Bruker A300 Spectrometer, respectively. The photocurrent, EIS and Mott-Schottky plots were explored by a CHI 660E electrochemical workstation and the photoluminescence spectra was characterized on a QuantaMasterTM 40.
50 mg of samples were ultrasonically suspended in a 100 mL of glass reaction cell containing 80 mL of H2O with 10 vol% triethanolamine and 400 μL of H2PtCl6·6H2O (10 g/L, theoretically depositing 3.0 wt.% on the catalyst). Then the cell was kept under a 300 W Xe lamp with a cutoff filter (λ > 420 nm). The amount of H2 was tested by a gas chromatograph equipped with a thermal conductivity detector (TCD) (GC-2014, Shimadzu Co., Japan). The external quantum efficiencies (EQE) of optimized Co(Ⅱ)/PCN were calculated via the following equation:
The crystal structures and phase compositions of the as-prepared g-C3N4 nanosheets and Co(Ⅱ)/PCN-2 were studied by XRD as shown in Fig. 1(a). Two characteristic peaks are existed in the XRD patterns of both g-C3N4 nanosheets and Co(Ⅱ)/PCN-2. The (100) peak at 13.1° and (002) peak at 27.5° are assigned to the in-plane structural packing motif of tris-triazine units and the interlayer stacking of aromatic segments [39]. The XRD pattern of the Co(Ⅱ)/PCN is similar to that of g-C3N4 nanosheets. The presence of Co(Ⅱ) and confirmation of P doping are verified by XPS and TEM-EDX. The XPS survey spectra of g-C3N4 nanosheets and Co(Ⅱ)/PCN-2 are shown in Fig. 1(b). As expected, two peaks of Co and P appear in the spectrum of Co(Ⅱ)/PCN-2 except for the peaks of C and N elements in both g-C3N4 nanosheets and Co(Ⅱ)/PCN-2. Figs. 1(c)‒(f) show high-resolution XPS spectra of individual elements. The dominant peak in the high-resolution C 1s spectra located at 284.8 eV is assigned to the C–C species of g-C3N4 nanosheets, while the peak at 288.0 eV is originated from the sp2-hybridized C in N=C–N [40, 41]. The spectrum of N 1s signal is deconvoluted into four peaks centered at 398.7, 399.6, 401.0, and 404.5 eV, which are attributed to the triazine rings (C=N–C) of g-C3N4, N–(C)3 groups, C–N–H and C=N conjugated structures, respectively [42]. Obviously, the binding energy of Co(Ⅱ)/PCN displayed in the high-resolution C and N XPS spectra shows a significant positive shift, revealing the reduction of electron density in g-C3N4 due to the incorporation of dopant (P) and the Co(Ⅱ) deposition. The high-resolution Co 2p XPS spectrum of Co(Ⅱ)/PCN-2 exhibits two strong peak located at 780.3 and 795.7 eV, respectively, which are ascribed to Co 2p1/2 and Co 2p3/2, respectively, manifesting the existence of the Co(Ⅱ) in the heterojunction [43]. Considering a low-temperature hydrolysis process in this study, the Co(Ⅱ) cocatalyst may be in the amorphous CoOOH-like structure, similar to the reported phenomenon of the other transition metal (such as Fe elements) [44]. The high-resolution P 2p XPS spectrum exhibits a strong peak at 133.4 eV attributing to P‒N coordination, revealing that P atoms may substitute the C atom in the triazine rings of the g-C3N4 and thus form P–N bonds [45]. The results of the XRD and XPS confirm the successful preparation of coupling of Co(Ⅱ) with P-doped g-C3N4.
The atomic force microscope (AFM) was employed to investigate the thickness of as-prepared samples. Fig. 2 indicates that bulk g-C3N4 has a thickness of 15–20 nm, while the thickness of g-C3N4 nanosheets ranges from 5 to 10 nm, manifesting that the bulk g-C3N4 can exfoliate into few-layered structures after ultrasonic exfoliation process. Such ultrathin g-C3N4 nanosheets were utilized to further realize P doping and heterojunction construction.
Furthermore, TEM and STEM-EDX elemental mapping were conducted to observe the morphologies of pure g-C3N4 nanosheets and Co(Ⅱ)/PCN-2. As observed in Fig. 3(a), g-C3N4 nanosheets have small thin flat irregular two-dimensional shape. TEM image of Co(Ⅱ)/PCN-2 (Fig. 3(b)) shows that phosphorus doping and Co(Ⅱ) loading did not affect the morphology of g-C3N4 nanosheets. STEM-EDX elemental mappings (Fig. 3(c)) confirm the existence and uniform distributions of C, N, Co, P elements. Based on XRD, XPS, TEM and EDX analysis, it can be concluded that P atoms are doped into g-C3N4 nanosheets and Co(Ⅱ) exists on the surface of the P-doped g-C3N4 nanosheets.
The visible light photocatalytic H2 production activity over pure g-C3N4 nanosheets, PCN and Co(Ⅱ)/PCN were evaluated. As shown in Figs. 4(a) and (b), pure g-C3N4 nanosheets have relatively low hydrogen production rate (89.2 μmol·g−1·h−1). However, the hydrogen production of PCN (155.2 μmol·g−1·h−1) exhibits 1.7 folds higher than that of pristine g-C3N4 nanosheets. Generally, P doping into g-C3N4 nanosheets can introduce an impurity level which can accelerate the separation of photogenerated carriers and result in the improvement of the photocatalytic activity. Also Co(Ⅱ)/PCN with different modified amounts of Co(Ⅱ) shows the improved photocatalytic properties in comparison to pure g-C3N4 nanosheets. Upon increasing the loading amount of Co(Ⅱ) from 0.5 wt% to 2 wt%, the Co(Ⅱ)/PCN samples show continuously increasing photocatalytic activity. However, When the loading amount of Co(Ⅱ) was further increased from 2.0 wt% to 6.0 wt%, the photocatalytic H2 activity decrease, which can be ascribed to the shielding effect of Co(Ⅱ). The Co(Ⅱ)/PCN-2 sample shows the highest photocatalytic performance (774 μmol·g−1·h−1), exhibiting 8.6 folds higher than that of pure g-C3N4 nanosheets. This indicates that the optimized Co(Ⅱ) loading can effectively enhance the water-reduced H2-production activity.
The cyclic stability of the Co(Ⅱ)/PCN-2 sample was also studied. The composite exhibits almost stable H2-evolution rate with a little decline after four cycles, revealing that Co(Ⅱ)/PCN-2 sample exhibits reliable stability for the photocatalytic H2-evolution, as shown in Fig. 4(c). The external quantum efficiencies (EQE) at different wavelengths of Co(Ⅱ)/PCN-2 were measured, and the EQE at λ = 402 nm is determined to be 2.76% (Fig. 4(d)).
Strong light-harvest capability, high separating efficiency of photoinduced carriers, as well as large surface reactions efficiency are important determining factors for excellent photocatalytic H2-evolution property. The UV-vis DRS of g-C3N4 nanosheets and Co(Ⅱ)/PCN-2 composite shown in Fig. 5(a) demonstrated that an enhancement of light absorption over g-C3N4 is achieved after P doping and Co(Ⅱ) decoration, and thus the Co(Ⅱ)/PCN-2 has the better visible light absorption ability than pure g-C3N4 nanosheets. Kubelk-Munk function is employed to calculate the bandgap of g-C3N4, which is determined to be 2.66 eV (Fig. 5(b)). As shown in Figs. 5(c)‒(d), both g-C3N4 nanosheets and Co(Ⅱ)/PCN-2 show positive slopes at different frequencies, indicating the classic n-type nature of semiconductor whether or not modified. According to the Mott-Schottky plots shown in Fig. 5(c) and the equation of the Efb = EAg/AgCl + 0.059 × pH + EAg/AgCl* [46], the hydrogen electrode potentials (NHE) of g-C3N4 nanosheets and Co(Ⅱ)/PCN-2 are determined as −0.67 and −0.53 V versus NHE, which are consistent with reported literatures [47]. Obviously, Co(Ⅱ)/PCN-2 shows more positive flat-band potential than g-C3N4 nanosheets, which might be caused by the lower overpotential as well as splendid electronic conductivity of Co(Ⅱ) [48, 49].
The photocurrent-time curves (i-t) of Co(Ⅱ)/PCN-2 working electrode were measured to examine the separation efficiency of photo-excited carriers. Apparently, the Co(Ⅱ)/PCN-2 exhibits stronger photocurrent density than that of the g-C3N4 nanosheets (Fig. 6(a)), implying Co(Ⅱ)/PCN-2 exhibits faster separation of photo-generated electron-holes [50, 51]. Also, Co(Ⅱ)/PCN-2 possesses a smaller arc than that of g-C3N4 nanosheets (Fig. 6(b)), which is the indication of rapid transportation rate of photogenerated electrons occurs due to P-doping and Co(Ⅱ) decorated nanosheets surface [52]. The photoluminescence (PL) measurement had also been applied to study the separation of the photo-excited carriers. Fig. 6(c) shows the PL emission spectra of Co(Ⅱ)/PCN-2 and g-C3N4 nanosheets both located at ~450 nm. The intensity of PL emission peak for Co(Ⅱ)/PCN-2 is lower than that of g-C3N4 nanosheets, indicating faster charge separation ability over the Co(Ⅱ)/PCN-2 [53]. Time-resolved photoluminescence measurements were employed to explore the lifetime of photogenerated carriers. The PL intensities, corresponding pre-exponential factors and mean lifetimes of g-C3N4 and Co(Ⅱ)/PCN-2 are shown in Fig. 6(d), a smaller lifetime (0.41 ns) is observed for the Co(Ⅱ)/PCN-2 compared with the g-C3N4 nanosheets (1.07 ns), revealing that Co(Ⅱ)/PCN-2 has the faster transportation of photogenerated electrons in case of Co(Ⅱ)/PCN-2 [54]. The results of PL measurement are well consistent with the results of photocurrent and EIS measurements, The improved separation efficiency of the photo-generated carriers of Co(Ⅱ)/PCN composite can enhance the photocatalytic activity.
The generation of the e− and h+ of g-C3N4 and Co(Ⅱ)/PCN-2 were characterized by the electron spin resonance (ESR). Usually, reduced ESR signal intensity indicates faster transfer rate of electrons from semiconductor to the spin label [55, 56]. As shown in Figs. 7(a) and (b), whatever for the TEMPO-e− and TEMPO-h+, both of the signal intensities of the g-C3N4 and Co(Ⅱ)/PCN-2 aqueous solution decreased within 5 min, however, more attenuation could be observed in the Co(Ⅱ)/PCN-2. The above results demonstrate that, under light irradiation, Co(Ⅱ)/PCN-2 composite excites more active electrons and holes compared to g-C3N4 nanosheets, which also means P doping and Co(Ⅱ) decorated nanosheets surface could effectively improve the reactivity of the photoinduced e− and h+. It is evident from the above results that Co(Ⅱ)/PCN-2 sample demonstrates better separation of the photo-generated e− and h+.
The XRD patterns of the Co(Ⅱ)/PCN-2 sample before and after the photocatalytic reaction are shown in Fig. 8(a), There is no obvious difference between Co(Ⅱ)/PCN-2 and Pt/Co(Ⅱ)/PCN due to loading of a small amount of Pt on the Co(Ⅱ)/PCN (only 3.0 wt% relative to Co(Ⅱ)/PCN). The TEM image of the Co(Ⅱ)/PCN-2 sample after the photocatalytic reaction is shown in Fig. 8(b), Some nanoparticles of diameter 5 nm were found on the Co(Ⅱ)/PCN-2 sample, which were not present previously. Furthermore, interplanar spacing of 0.23 nm corresponding to (111) facets of Pt was determined in HRTEM image [57]. It is assumed that under the irradiation of the Xe lamp, the PtCl62− is reduced to Pt by the incoming the electron from the conduction band of P-doped g-C3N4 nanosheets and acting as electron-transfer cocatalyst to promote half-reaction of hydrogen production [58].
The proposed mechanism for photocatalytic hydrogen production over Co(Ⅱ)/PCN in the light of the above discussion is shown in Fig. 9. The g-C3N4 nanosheets could be excited to generate e− and h+ under light irradiation. The dopant "P" in the g-C3N4 nanosheets could provide new channels to mediate electron migration and restrain the recombination of electrons and holes. In case of Co(Ⅱ) decorated P-doped g-C3N4 nanosheets, Co(Ⅱ) captures photogenerated hole to form Co(Ⅲ), and the new-formative Co(Ⅲ) will be reduced to Co(Ⅱ) by oxidizing sacrificial agent (TEOA) due to the high oxidative capacity of Co(Ⅲ) [59]. Herein, the Co(Ⅱ) plays role of h+-transfer cocatalyst to capture h+ in present photocatalytic system. Furthermore, the photogenerated-e− will first reduce PtCl62− to Pt according photoreduction deposition. After the formation of Pt, it can act as the electron-transfer cocatalysts to take control of the photogenerated-e− in order to provide the active sites for hydrogen production. Therefore, with the synergistic effect of Co(Ⅱ) as a hole cocatalyst and Pt as electron cocatalyst, electrons and holes demonstrated the directional shunting, which could further accelerate the migration of photogenerated electrons to platinum in order to join the photocatalytic reduction process for hydrogen evolution. It can be concluded that the transportation and separation of photogenerated carriers were accelerated to the greatest extent over the Pt/Co(Ⅱ)/PCN photocatalyst. The enhancement of photocatalytic performance of Co(Ⅱ)/PCN-2 can be ascribed to the synergistic effects of Co(Ⅱ), Pt deposition and P doping.
In summary, we employed low-temperature phosphating method to fabricate Co(Ⅱ)-modified P-doped g-C3N4 nanosheets, which simultaneously realized the P-doping and Co (Ⅱ) as hole cocatalyst co-modification, and Pt is deposited as electron cocatalyst for the photocatalytic H2 generation. The results of XRD, XPS, TEM and STEM confirmed the successful preparation of the Co(Ⅱ)/PCN samples, in which P element is doped into g-C3N4 nanosheets and Co(Ⅱ) is decorated on the surface of P-doped g-C3N4 nanosheets successfully. The optimized Co(Ⅱ)/PCN sample exhibited the highest photocatalytic hydrogen evolution (774 μmol·g−1·h−1) by using Pt as electron transfer cocatalyst, which is ~8.6 fold higher than pristine g-C3N4 nanosheets (89.2 μmol·g−1·h−1). The enhancement in photocatalytic activity is resulted from the faster separation of photogenerated e− and h+ caused by the synergistic effect of Co(Ⅱ) (hole transfer cocatalyst), Pt (electron transfer cocatalyst) and P doping (modulate the electronic structure of g-C3N4) of the Co(Ⅱ)/PCN heterojunction. This work offers a novel insight into the designing of excellent photocatalysts consisting of both hole-transfer and electron-transfer cocatalysts for energy conversion application.
2020,
Vol. 41 Issue (1): 72-81 DOI: S1872-2067(19)63430-3 
PDF
2020,
Vol. 41 Issue (1): 72-81 DOI: S1872-2067(19)63430-3 PDF