Enhancement of Photocatalytic H2-Evolution Kinetics through the Dual Cocatalyst Activity of Ni2P-NiS-Decorated g-C3N4 Heterojunctions

Zhuonan Lei , Xinyi Ma , Xiaoyun Hu , Jun Fan ,＊, Enzhou Liu ,＊

1 School of Chemical Engineering/Xi’an Key Laboratory of Special Energy Materials, Northwest University, Xi’an 710069, China.

2 School of Physics, Northwest University, Xi’an 710069, China.

Abstract: With rapid industrialization, issues pertaining to the environment and energy have become an alarming concern. Photocatalytic water splitting is considered one of the most promising green technologies capable of resolving these issues, as it can convert solar energy into chemical energy and have a positive impact on the realization of “carbon neutrality”. Current research focuses on the development of highly efficient catalysts to improve the photocatalytic H2-production activity. Transition metal phosphides and sulfides are often used as photocatalysts owing to their low H2-evolution overpotential and excellent electrical conductivity. Among them, Ni2P and NiS have generally been used independently during photocatalytic H2 production; however, it is necessary to study the synergistic effect when they are combined as a dual cocatalyst. In this work, we successfully prepared a Ni2P-NiS dual cocatalyst for the first time via a simple hydrothermal method using red phosphorus (RP) and thioacetamide (C2H5NS) as the sources of P and S. Ni2P-NiS was then introduced to the surface of g-C3N4 nanosheets through solvent evaporation to create a Ni2P-NiS/g-C3N4 heterojunction. Furthermore, X-ray diffraction (XRD), scanning electron microscope (SEM), transmission electron microscope (TEM), ultraviolet-visible spectrophotometry (UV-Vis), X-ray photoelectron spectroscopy (XPS),photoluminescence (PL), linear sweep voltammetry (LSV), Mott-Schottky (M-S), and electrochemical impedance spectroscopy (EIS) were used to reveal the crystal structures, morphologies, element compositions, and photoelectric characteristics of the samples; thus, it was demonstrated that Ni2P-NiS was successfully deposited on the surface of g-C3N4 and that together they exhibited better activity than their monomers. Moreover, the optimized 15% Ni2P-NiS/g-C3N4 composite exhibits a H2 generation rate of 6892.7 μmol·g-1·h-1, which is about 46.1, 7.5 and 4.4 times higher than that of g-C3N4 (150 μmol·g-1·h-1), 15% NiS/g-C3N4 (914.5 μmol·g-1·h-1), and 15% Ni2P/g-C3N4 (1565.9 μmol·g-1·h-1),respectively. In addition, photoelectric performance tests show that Ni2P-NiS/g-C3N4 has a stronger photocurrent intensity,smaller charge-transfer resistance, more positive H2-evolution overpotential, and better charge-separation ability than the individual components (i.e., Ni2P and NiS), suggesting that the coexistence of Ni2P and NiS can further boost the activity of g-C3N4 during H2 evolution compared with their monomers. This is mainly due to the Schottky barrier effect between Ni2P-NiS nanoparticles and g-C3N4 nanosheets, which can greatly promote charge separation and charge transfer at their interface. Additionally, Ni2P-NiS can reduce the H2-evolution overpotential, leading to the increased surface kinetics of H2 evolution. This work offers a promising approach to obtaining a highly active and stable noble-metal-free dual cocatalyst for photocatalytic H2 production.

Key Words: Ni2P-NiS; g-C3N4; Cocatalyst; Schottky barrier; Photocatalytic water splitting

1 Introduction

Nowdays, energy and environmental issues from the consumption of fossil resources have become two major problems for the sustainable development of the world. It is urgent to exploit new energy sources to solve above problems1,2.“Zero carbonization of fuel” will play an important role in the future energy systems. Hydrogen (H2) is considered as an ideal candidate to replace fossil fuels for its pollution-free and highenergy density features3. Photocatalytic H2evolution by water splitting can convert solar energy into chemical fuels, which is a promising way to produce H2through a cost-efficient and environment-friendly ways4-8.

Since 19729, many photocatalysts have been explored for water splitting, however, the development of more excellent catalysts is still a hot spot. For the traditional catalysts, TiO2can only harvest ultraviolet light, leading to a lower energy conversion efficiency10,11. Although CdS with narrow band gap can absorb visible light, it usually suffers from poor photostability12,13. Since 200914, g-C3N4has been widely investigated for its superiorities, such as suitable bandgap (about 2.7 eV), good chemical stability, simple preparation process,large specific surface area and so on15-18. However, its H2evolution efficiency is seriously blocked to the quick recombination of electrons and holes induced by the incident light. Therefore, various strategies were used to facilitate the H2evolution process over g-C3N4, including morphology control19,20, element doping21,22, cocatalysts loading23,24and heterojunction construction25,26,etc. Among them, cocatalysts loading is an efficient way by increasing the active sites,inhibiting the charge recombination and accelerating the REDOX kinetics27. The most effective cocatalysts are noble metals nanoparticles, including Pt28, Au29, Ag30, however, the high cost makes their large-scale application impossible. Hence,it is of significance to develop low-cost and high-efficiency cocatalysts in the view of economic.

Recently, transition metal phosphides and sulfides have become a hot spot due to their excellent chemical stability and conductivity, which are considered as ideal substitution of noble metals31-33. For instance, Luet al.found that highly dispersed Ni2P QDs over g-C3N4can effective promote the photoelectric activity of g-C3N434. Wanget al.used NiS cocatalyst to promote the H2evolution activity of g-C3N4, and the maximum H2production rate over the ideal sample reached 244 μmol?g-1?h-135.Further investigations showed that the transition metal thiophosphates exhibited better H2production activity36. Zhanget al.reported that porous iron phosphosulfide (FePS3) prepared by one-step sulfurphosphidation method exhibited a relative higher H2production rate of 305.6 μmol?g-1?h-137. Zhanget al.prepared two different of copper phosphosulfide (Cu3P|S and CuS|P) by the solid-state reaction and the wet chemical processes respectively, the photocatalytic HER (Hydrogen Evolution Reaction) activities are dramatically enhanced over both Cu3P|S and CuS|P, the H2evolution rates are 2085 μmol?g-1?h-1and 976 μmol?g-1?h-1, respectively38. However, the preparation processes of transition metal phosphorsulfides are slightly complex, which usually requires tubular furnace,harmful reagents and higher temperature.

Inspired by the works above, if transition metal phosphates and transition metal sulfides can be used as cocatalyst at the same time, the advantages of them might trigger a positive synergistic effect. For example, Xiaoet al. obtained NiS/Ni2P heterostructure on carbon cloth (NiS/Ni2P/CC) by a three-step method with only an overpotential of 111 mV for the HER,which is superior to NiS and Ni2P alone39. Xuet al. found that NiS/Ni2P nanoparticles is an efficient and durable dualfunctional cocatalyst for the amorphous Ni(OH)2during the photoelectric catalytic overall water splitting40. Despite these advances, photocatalytic H2production over Ni2P/NiS cocatalysts-related system has never been reported.

In this work, the Ni2P-NiS nanoparticle was prepared through a hydrothermal approach using thioacetamide (C2H5NS) and red phosphorus (RP) as the S and P sources respectively, and then Ni2P-NiS/g-C3N4heterojunction was obtained by deposition of Ni2P-NiS on the g-C3N4using a solvent evaporation process. The as-prepared 15% (w, mass fraction) Ni2P-NiS/g-C3N4exhibits a superior H2production activity than 15% Ni2P/g-C3N4and 15%NiS/g-C3N4. According to a series of analysis, the Ni2P-NiS can restrain the charge recombination of g-C3N4, and decrease H2evolution overpotential, sequentially a faster H2evolution kinetics.

2 Experimental

2.1 Synthesis of g-C3N4 nanosheets

g-C3N4nanosheet was prepared using a simple hightemperature polymerization process followed by a ultrasonic treatment. In detail, first, 20 g of urea was heated at 550 °C for 2 h in a microwave muffle oven with tin foil wrapped tightly.After cooling down, the sample was gathered to acquire the bulk g-C3N4. Second, 2 g of obtained bulk g-C3N4was put into 1000 mL of water and treated under ultrasonic. Finally, the g-C3N4nanosheet with light-yellow color was obtained after centrifugation, washing and drying at 60 °C.

2.2 Preparation of Ni2P-NiS nanoparticles

Ni2P-NiS nanoparticles were prepared by a one-step hydrothermal method. Typically, 2 mmol of NiCl2·6H2O, 1.5 mmol of RP (the purification of RP is shown in Supporting Information) and 0.5 mmol of C2H5NS were added into 50 mL of water. Subsequently, the pH of the suspension was adjusted by adding 300 μL of 10 mol·L-1NaOH, after magnetic stirring,the mixture solution was sealed in a Teflon-lined stainless-steel autoclave and treated at 140 °C for 12 h. Finally, Ni2P-NiS nanoparticles was acquired by centrifugal washing and dry at 60 °C for 8 h.

Moreover, pure Ni2P nanoparticles was fabricated by a similar hydrothermal reaction. In detail, 2.5 mmol of NiCl2·6H2O, 25 mmol of RP were dispersed in 50 mL of water under stirring,which was sealed and treated at 140 °C for 12 h.

For pure NiS nanoparticles, 1 mmol of Ni(CH3COO)2·4H2O was dispersed in 50 mL of ethyl alcohol, after stirring for 30 min and ultrasonic treatment for 20 min, then 1 mmol of C2H5NS was dispersed in the above solution under stirring, finally the above suspension was heated at 200 °C for 12 h in a sealed condition.

2.3 Preparation of Ni2P-NiS/g-C3N4 composites

A simple solvent evaporation process was employed to synthesis of Ni2P-NiS/g-C3N4heterojunction, as shown in Fig.1, 100 mg g-C3N4nanosheet and an appropriate Ni2P-NiS were added into 10 mL of ethyl alcohol respectively, after 2 h ultrasonic treatment, they were mixed and stirred at 60 °C to obtain gray powder by fully evaporation of ethyl alcohol.Herein,x% Ni2P-NiS/g-C3N4composites (x= 3, 5, 10, 15 and 20) were prepared by changing the loading amount of Ni2P-NiS,xrepresents the experimental mass ratio of g-C3N4and Ni2PNiS.

2.4 Characterization

More information about the XRD, SEM, TEM, XPS, UV-Vis,PL, LSV, M-S, EISetc. tests are listed in Supporting Information.

3 Results and discussion

3.1 Structure and morphology

Fig. 2a presents the XRD of the pure Ni2P, NiS and Ni2P-NiS nanoparticles. The signals at 40.7°, 44.6°, 47.4° and 54.2° belong to (111), (201), (210) and (300) crystal planes of Ni2P (PDF 74-1385), respectively41. For NiS, the signals at 18.4°, 32.2°, 35.7°,40.5°, 48.8°, 52.6° are well-indexed to (110), (300), (021), (211),(131) and (401) crystal planes of NiS (PDF 86-2281)42. In addition, the XRD pattern of Ni2P-NiS is consistent with the standard cards of Ni2P and NiS. The above results indicate that Ni2P-NiS has been obtainedviathe solvothermal approach in our work. In addition, two peaks at 27.4° and 13.0° are observed in Fig. 2b, which are belonging to the (100) and (002) planes of g-C3N4(PDF 87-1526), respectively43. It is clearly that the signals of g-C3N4are preserved after loading Ni2P-NiS, indicating g-C3N4is not destroyed in the process of solvent evaporation. The characteristic peaks of Ni2P-NiS, Ni2P and NiS are obviously observed in the composites too, indicating Ni2P-NiS, Ni2P and NiS are introduced to the surface of g-C3N4, respectively.

In Fig. 3, the peak at 810 cm-1belongs to the bending vibration of the tri-s-triazine ring. The peaks from 1000 cm-1to 1800 cm-1result from the typical stretching vibration of C―N and C＝N bonds. And the wide signals from 3000 cm-1to 3400 cm-1is attributed to the ―NH stretching and ―OH from the surface-adsorbed H2O44,45. In addition, there is no obvious difference among the spectra of 15% Ni2P/g-C3N4, 15% NiS/g-C3N4, 15% Ni2P-NiS/g-C3N4composites and pure g-C3N4,implying the basic chemical structure of g-C3N4is reserved,which agrees with the above XRD results.

Fig. 4 shows the SEM image of the samples. In Fig. 4a, Ni2PNiS particles presents an irregular shape and aggregates together.However, Ni2P-NiS is tightly attached with g-C3N4nanosheet in Fig. 4b, indicating Ni2P-NiS is introduced to the g-C3N4.Furthermore, the corresponding SEM element mappings of 15%Ni2P-NiS/g-C3N4are presented in Fig. 4c, and the evenly distribution of each element further indicates that Ni2P-NiS are decorated on the g-C3N4uniformly. The EDS spectra in Fig. 4d shows the weight and atom percentage of each element. In addition, only a little Ni2P-NiS is introduced to the g-C3N4during the physical recombination process, while most of Ni2PNiS enters into the solution, therefore P, S, and Ni make up only～3% of the total mass.

To further reveal the surface morphology and structure of the samples, TEM investigation was performed on Ni2P-NiS and 15% Ni2P-NiS/g-C3N4. As shown in Fig. 5a, Ni2P-NiS nanoparticles is dispersed evenly. In Fig. 5b, the spaces of 0.192 nm and 0.222 nm correspond to the (210) plane of Ni2P and (211)plane of NiS, they are also observed in the XRD patterns in Fig. 246,47. From Fig. 5c-e, it can be seen that Ni2P-NiS nanoparticles is closely contacted with g-C3N4. Besides, as seen in Fig. 5e, the (211) crystal plane of NiS and the (210) crystal plane of Ni2P are also corresponding to the spaces of 0.222 nm and 0.192 nm respectively.

The surface chemical state of the elements was analyzed using XPS. All the spectra are adjusted according to the C 1ssignal at 284.8 eV. Besides, the signals of C 1s, N 1s, O1s, Ni 2p, P 2pand S 2pin 15% Ni2P-NiS/g-C3N4can all be detected as exhibited in the survey spectrum (Fig. 6a). The O element might originate from the surface adsorbed CO2and H2O. As displayed in Fig. 6b, C 1shigh resolution spectrum could be divided into two peaks at 284.80 and 288.34 eV, they correspond to thesp2C―C andsp2N―C＝N bonds in g-C3N4respectively. For N 1sspectra in Fig. 6c, three peaks could be resolved at 398.83,400.27 and 401.32 eV, which belong to thesp2hybridized aromatic N atoms of C＝N―C, the tertiary N atom bonded to C of N―(C)3and the terminal amino function groups of C―N―H, respectively48. After introducing Ni2P-NiS, the signals of C 1sand N 1sshift toward high energy obviously,indicating there is a powerful internal forces between g-C3N4and Ni2P-NiS49. Fig. 6d exhibits the Ni 2pXPS spectrum, the two peaks at 874.38 and 856.52 eV are identified as Ni 2p1/2and Ni 2p3/2, and another two peaks at 880.01 and 862.36 eV result from the satellite peaks of Ni 2p1/2and Ni 2p3/2. What’s more, the peak at 852.76 eV can be attributed to the Niδ+(0 ＜δ＜ 2) species50,51.For the P 2p(Fig. 6e), the peak at 134.29 eV in P 2pspectra results from the POXspecies for the oxidation in air40. For the S 2pspectra in Fig. 6f, the signals at 163.42 and 162.11 eV are attributed to S 2p1/2and S 2p3/252, and the peak at 169.06 eV belongs to S ＝O due to the oxidation of surface sulfide.Compared with the signals of Ni2P-NiS, the peaks of 15% Ni2PNiS/g-C3N4shift toward lower binding energy.

3.2 Photocatalytic experiment

Fig. 7 shows the activity of the samples during H2production in 20% (volume faction) TEOA aqueous solution (TEOA is the sacrificial agent) under 300 W Xe-lamp irradiation. As displayed in Fig. 7a, g-C3N4possesses a low H2production activity, after introducing Ni2P-NiS, the H2production rate of g-C3N4first obviously enhances with the increase the mass ratio of Ni2P-NiS,and then it decreases for the shielding effect from the overloaded Ni2P-NiS, which usually has excellent light harvesting ability53.In addition, 15% Ni2P-NiS/g-C3N4achieves a highest H2production rate of 6892.7 μmol·g-1·h-1, which is about 46.1 times higher than that of g-C3N4(150 μmol·g-1·h-1). Besides, the rate of 15% Ni2P-NiS/g-C3N4is approximately 4.4 and 7.5 times higher than those of 15% Ni2P/g-C3N4and 15% NiS/g-C3N4respectively (Fig. 7b). It’s obvious that Ni2P-NiS is an excellent H2evolution dual cocatalyst.

Fig. 8 N2 adsorption-desorption isotherm curves with corresponding SBET.

3.3 N2 adsorption-desorption isotherm

N2absorption-desorption isotherms can reflect the microstructure and theSBETof the catalysts. All the curves share H3 adsorption hysteresis loops, resulting from the accumulation of sheet-like g-C3N4in Fig. 854. TheSBETof 15% Ni2P-NiS/g-C3N4is 64.38 m2·g-1, it is lower than theSBETof pure g-C3N4(99.99 m2·g-1), but larger than theSBETof Ni2P-NiS (47.72 m2·g-1), which may result from the hole blocking effect of Ni2PNiS nanoparticles after loading on the surface of g-C3N455. This indicates that theSBETis not the key factor to determine the activity of the samples, but the active site from Ni2P-NiS.

3.4 Photoelectric properties

Besides the structural features of the catalyst itself, the optical and electrical properties of the catalyst are also important factors to the activity of the catalysts. Photoelectric properties tests were conducted to further reveal of the HER activity of Ni2P-NiS/g-C3N4. The UV-Vis diffuse reflectance spectroscopy is exhibited in Fig. 9a, the g-C3N4absorption edge of is approximately 455 nm. NiS, Ni2P and Ni2P-NiS cocatalysts exhibit excellent light absorption ability in the whole regions for its black color. That is why the absorption of the composites is better than the g-C3N4.It strongly proved that the introduction of cocatalysts such as NiS, Ni2P and Ni2P-NiS can affect the absorption capacity of g-C3N4. Furthermore, on the basis of the Mott-Schottky plots in Fig. 9b, g-C3N4is a n-type semiconductor owing to the positive slope56. Meanwhile, its flat band potentials (Efb) is -1.24 Vvs.SCE according to the horizontal intercept. In Fig. 9c,d, theEFof Ni2P and NiS are -0.85 V, -0.96 V respectively. According to Fig. 9e, the bandgap of g-C3N4is calculated to be 2.67 eV based on the Kubelka-Munk plots. Furthermore, it is known to all that the flat band of n-type semiconductor is approximately more positive 0.2 eV than its CB position. On the basis of the formula(ENHE=ESCE+ 0.24, standard hydrogen electrode conversion),the CB of g-C3N4is -1.2 Vvs. NHE. Therefore, the VB of g-C3N4is 1.47 V base on the formula (ECB=EVB-Eg)correspondingly.

Fig. 10a exhibits theI-tcurves of the samples, compared with 15% Ni2P/g-C3N4and 15% NiS/g-C3N4, the photocurrent intensities of 15% Ni2P-NiS/g-C3N4has a slight enhancement under the same bias voltage, indicating the superiority conductivity of Ni2P-NiS/g-C3N4. The electrochemical impedance spectroscopy (EIS) of samples is described in Fig. 10b,the arc radius of semicircle of EIS can reflect the resistance of the systems, the smaller radius means the smaller resistance57.Obviously, the radius of the composites is smaller compared with pure g-C3N4, suggesting a smaller charge transfer resistance of the composites, and further illustrating that Ni2P-NiS can facilitate the charge transfer. To investigate the surface reduction ability of protons and the kinetics of H2evolution of the catalyst,the HER polarization curves were measured. As shown in Fig.10c, Ni2P-NiS can obviously enhance the H2evolution reduction ability of the sample. The corresponding potentials of g-C3N4,15% Ni2P/g-C3N4, 15% NiS/g-C3N4and 15% Ni2P-NiS/g-C3N4are -1.9, -1.71, -1.71 and -1.69 Vvs. NHE under 1 mA·cm-2current density, respectively. It is apparently that the H2evolution overpotential of the composites is lower than that of g-C3N4, that is more conducive to promote the H2evolution reaction kinetics. Finally, photoluminescence (PL) spectra is shown in Fig. 10d, the emission peak around 460 nm belongs to the light emission from the charge recombination58.Furthermore, the PL intensity of 15% Ni2P-NiS/g-C3N4is significantly reduced in comparison to the g-C3N4, 15% Ni2P/g-C3N4and 15% NiS/g-C3N4, suggesting an effectively charge separation after loading Ni2P-NiS.

3.5 Analysis of photocatalytic mechanism

Based on the analysis above, the charge transfer route between Ni2P-NiS and g-C3N4is proposed preliminary in Fig. 11. Before g-C3N4and Ni2P-NiS contacting, the Fermi level (EF) of g-C3N4is negative than H2evolution potential (EH+|H2), while Ni2P-NiS is more positive. Therefore, it is impossible for Ni2P-NiS to produce H2from the view of thermodynamics. The electrons can transfer from g-C3N4to Ni2P-NiS until reaching an equilibrium after contacting, resulting in an internal electric field between them with a Schottky barrier formed59. After light excitation, a large number of electrons on the CB of g-C3N4can shift to Ni2PNiS in an irreversible way due to the existence of Schottky barrier, furthermore, the electrons in Ni2P-NiS are able to quickly reduce H+to H2. While TEOA can quickly consume the holes on VB. On the one hand, the Schottky barrier can effectively restrain the charge recombination, on the other hand,Ni2P-NiS can reduce the H2evolution barrier to facilitate the H2evolution kinetics. Thus, Ni2P-NiS as dual cocatalyst can effectively promote the H2production performance of g-C3N4.

4 Conclusions

In this work, Ni2P-NiS dual cocatalyst was successfully fabricatedviaa facile hydrothermal method, subsequently, it was introduced to the surface of g-C3N4through a solvent evaporation way. The 15% Ni2P-NiS/g-C3N4shows an excellent H2production rate of 6892.7 μmol·g-1·h-1, that is about 46.1, 7.5 and 4.4 times higher than those of g-C3N4, 15% NiS/g-C3N4and 15% Ni2P/g-C3N4, respectively. The investigation demonstrates that dual cocatalyst modification is a good approach to improve the activity of g-C3N4compared with the single ones. The introduction of Ni2P-NiS not only can suppress the charge recombination and improve the electric conductivity, but also can lower H2evolution overpotential, leading to an excellent HER activity. This work sheds light on the fabrication of noblemetal-free dual cocatalyst to improve the photocatalytic H2evolution of a semiconductor.

Supporting Information:available free of chargeviathe internet at http://www.whxb.pku.edu.cn.