Introduction

With the rapid development of social economy and the continuous growth of the population, people’s demand for energy is increasing [1, 2]. Long-term use of traditional fossil fuels will accelerate the energy crisis, climate change, and environmental pollution. Therefore, it is an irresistible trend to develop clean, efficient, and renewable alternative energy, such as wind energy, hydrogen energy, solar energy, and biomass energy [1,2,3]. However, much renewable energy cannot be effectively utilized because of their instability, intermittence, and regionality. As an energy carrier, hydrogen has the advantages of high energy density, non-toxicity, wide range of sources, environmentally friendliness, and so on, which is considered to be one of the most promising renewable clean energy alternatives to fossil fuels [4, 5]. The development of hydrogen production technology is of great practical significance to relieve energy crisis and environmental problems, and realize the sustainable development of energy. Electrochemical water splitting is recognized to be the most promising way to achieve sustainable energy development and zero emission in the current hydrogen production technologies [6, 7]. However, because of its slow reaction kinetics for HER, it is necessary to design and develop effective electrocatalysts that can reduce HER overpotential [8]. At present, noble metal platinum-based catalysts exhibit outstanding HER catalytic activity. Unfortunately, such catalysts have the disadvantages of limited earth reserves and high price, which limit their large-scale industrial applications [9]. Therefore, it is urgent to develop non-noble metal–based catalysts with abundant reserves, low cost, high efficiency, and stability, which is very important to realize the industrial application of hydrogen production by electrochemical water splitting.

At present, the most studied non-noble metal HER catalysts are transition metal compounds, such as sulfides [10], selenides [11, 12], and phosphides [13,14,15,16]. Especially, transition metal phosphides have been widely studied as HER catalysts in recent years because of their abundant reserves, low cost, good conductivity, and stable catalytic activity. But they show higher overpotential, slower charge transfer, and poorer conductivity compared with noble metal platinum-based catalysts. So, a variety of modification methods have been proposed, such as morphological controlling [17, 18], compounding [19, 20], and doping [21,22,23,24,25,26]. Sumboja et al. [18] synthesized the alveolar sac–like morphology of cobalt phosphide (Co-P I) and randomly shaped nanoparticles of cobalt phosphide (Co-P II). Compared with Co-P II, Co-P I displayed better catalytic activity in acidic and alkaline media, which was due to that its unique structure could increase the catalytic active area, contributing to enhance HER activity. Ge et al. [21] reported that well-crystallized molybdenum phosphide was dispersed on graphene carbon nanosheets co-doped with N and P (MoP/NPG). The results showed that the coupled interfaces and synergistic effect between MoP and graphene substrate could promote charge transfer kinetics and improve its catalytic activity. Gao et al. [27] revealed that rare-earth elemental Ce was doped into CoP to improve HER performance by modulating electronic structures and decreasing adsorption free energy of hydrogen. The above modification methods have significantly improved the HER performance of catalysts, but the insufficient active sites are still the key to affect catalytic activity. In recent years, constructing heterojunction by interface engineering has been recognized as an effective method to further improve the HER electrocatalytic performance of materials [28,29,30,31]. Boppella et al. [32] synthesized two-dimensional cobalt phosphide/nickel cobalt phosphide (CoP/NiCoP) heterojunction nanosheets supported by nitrogen-doped carbon. The collective effects of electronic structure engineering and strong interfacial coupling between CoP and NiCoP in heterojunction improved the reaction kinetics and catalytic performance of HER. Yan et al. [33] successfully prepared Ni2P-Ni5P4 heterostructure nanosheet arrays by simply tuning the reaction temperatures. The abundant catalytic active sites, large surface areas, highly conductive support of carbon cloth substrates, and unique free-standing arrays could enhance kinetics and electrocatalytic performances for the HER, oxygen evolution reaction (OER) and overall water splitting.

These studies reveal that porous nickel phosphides have good conductivity and large surface area, which can provide abundant active sites and accelerate electron transportation and charges transfer [34, 35]. Constructing a heterostructure through different nickel phosphide phases can effectively increase electroactive sites, contributing to improve catalytic activity [33]. At the same time, the three-dimensional porous nickel foam substrate possesses large surface area, high structural porosity, and fast electron transport channels, which are beneficial for electron transport and mass transfer. Wang et al. [36] prepared a self-supported Ni5P4-Ni2P nanosheet (NS) array by directly phosphating nickel foam with phosphorus vapor, which showed better electrocatalytic activity and long-term persistence for HER in acidic medium. Herein, we constructed a self-supporting Ni2P-Ni12P5/NF heterostructure catalyst supported on three-dimensional porous nickel foam substrate through hydrothermal reaction and low-temperature phosphorization. The synergy between the Ni2P-Ni12P5 heterostructure and the three-dimensional porous nickel foam substrate effectively enhanced the HER electrocatalytic activity and stability of the catalyst.

Experimental section

Materials

Hydrochloric acid, acetone, ethanol, and urea were supplied by Sinopharm Chemical Reagent Co., Ltd. Red phosphorus and nickel nitrate hexahydrate were obtained from Shanghai Zhongqin Chemical Reagent Co., Ltd. Sodium hypophosphite monohydrate and ammonium fluoride were obtained from Yantai Shuangshuang Chemical Co., Ltd. Potassium hydroxide was purchased from Taicang Hushi Reagent Co., Ltd. Nickel chloride hexahydrate and sodium lauryl sulfate were produced by Shanghai Titan Technology Co., Ltd, and Shanghai Aladdin Biochemical Technology Co., Ltd., respectively. All chemicals utilized were of analytical grade and were used as supplied without any further purification.

Material preparation

Nickel nitrate hexahydrate (2 mmol), ammonium fluoride (6 mmol), and urea (10 mmol) were added to 30 mL of deionized water, forming a green homogeneous mixture after ultrasonicating for 1 h. The above mixture and acid-washed nickel foams (NF) were introduced into a 50 mL high-pressure reactor and kept at 120 °C for 6 h in the oven. The obtained solid was washed with ethanol and deionized water, and then placed in a vacuum oven at 60 °C to completely dry to obtain a Ni(OH)F/NF precursor [37]. Ni(OH)F/NF was placed in a tube furnace and then heated to 450 °C for 2 h at a heating rate of 5 °C min−1 in air atmosphere to obtain NiO/NF. The subsequent phosphorization was carried out in Ar atmosphere by gradually heating sodium hypophosphite monohydrate and NiO/NF to 350 °C at a heating rate of 3 °C min−1 for 2 h. The product was naturally cooled to room temperature, washed with ethanol and deionized water alternately for several times, and dried in a vacuum oven at 60 °C to obtain the material Ni2P-Ni12P5/NF. The preparation process of the Ni2P-Ni12P5/NF is shown in Fig. 1. The preparation process of Ni2P was given in the Supporting Information.

Fig. 1
figure 1

Preparation process of Ni2P-Ni12P5/NF

Full size image

Material characterizations

X-ray diffraction (XRD) tests were performed on a D/Max 2400 powder diffractometer to analyze crystal phases of the as-synthesized materials. Scanning electron microscopy (SEM) was conducted on JSM-6701F to characterize the morphology of the materials. Transmission electron microscopy (TEM) and energy-dispersive X-ray (EDX) spectrometry measurements were carried out on TF20 to characterize lattice fringes and chemical elements. X-ray photoelectron spectroscopy (XPS) characterization was performed on a PHI 5702 XPS instrument to analyze the chemical compositions and valence states of the materials.

Electrochemical measurements

All the electrochemical measurements were conducted on a three-electrode electrochemical cell by an Autolab PGSTAT128N electrochemical workstation. Ni2P-Ni12P5/NF and other samples as controls were used as the working electrode. Ag/AgCl and Hg/HgO were used as the reference electrodes in acidic and alkaline electrolytes, respectively. The graphite rod was used as counter electrode. Electrochemical tests were performed in 0.5 M H2SO4 and 1.0 M KOH electrolytes, respectively. The presented potentials in this work were all converted to reversible hydrogen electrode (RHE) via the equation: E (RHE) = E (Ag/AgCl) + 0.059 pH + 0.197 (in acidic electrolyte) and E (RHE) = E (Hg/HgO) + 0.059 pH + 0.098 (in alkaline electrolyte). All of the polarization curves were recorded using linear sweep voltammetry (LSV). Electrochemical impedance spectroscopy (EIS) measurements were performed at the corresponding open-circuit potential to the electrode. The frequency range was 50 kHz–0.01 Hz. The charge-transfer resistance (Rct) was calculated by the diameter of the semicircular arc in the Nyquist plots. The double-layer capacitance (Cdl) values were determined by performing cyclic voltammetry (CV) measurements at different scanning rates of 30 ~ 150 mV s−1 under a non-Faradaic potential range. The Faradaic efficiency (FE) was evaluated by comparing the theoretical value with the amount of H2 obtained from the experiment based on the current density of 10 mA cm−2. The amount of produced H2 was recorded by the drainage method. The calculation formula is FE (%) = (m × n × F)/(I × t). m and n represent the number of hydrogen production (mol) and electrons transferred by a hydrogen molecule, respectively. F represents Faraday’s constant (96,485 C mol−1). I and t represent current (A) and time (s), respectively.

Results and discussion

XRD tests were used to identify the crystal structure of materials. Figure 2a shows the XRD patterns of the as-synthesized Ni2P and Ni2P-Ni12P5. The diffraction peaks of Ni2P-Ni12P5 can be well indexed to Ni2P (PDF#74–1385) and Ni12P5 (PDF#74–1381). The peaks at 40.7, 44.6, 47.4, 54.3, and 55.1° are indexed to (111), (201), (210), (300), and (211) crystal planes of Ni2P, respectively. The diffraction peaks appeared at 48.9, 46.9, and 38.4° are corresponding to (312), (240), and (112) crystal planes of Ni12P5, respectively. This indicates that two kinds of nickel phosphide were successfully prepared by hydrothermal reaction and the followed low-temperature phosphorization. The diffraction peaks of the pure Ni2P are also well indexed to Ni2P (PDF#74–1385). The XRD pattern of NiO was given in Fig. S1a (Supporting Information).

Fig. 2
figure 2

The characterizations of Ni2P-Ni12P5/NF. a XRD patterns, b SEM image, c TEM image, d HRTEM image, e–h HAADF-STEM image and EDX elemental mappings

Full size image

Figure 2b shows the SEM image of Ni2P-Ni12P5/NF. It can be seen that Ni2P-Ni12P5/NF is characterized with a three-dimensional porous structure composed of nanoparticles and nanosheets. As shown in Fig. S1b, the precursor NiO/NF covered on nickel foams with a sheet structure composed of nanoparticles. NiO/NF was phosphorized with PH3 gas that was generated by in-situ thermal decomposition of sodium hypophosphite to form Ni2P-Ni12P5/NF with a three-dimensional porous structure of nanoparticles and nanosheets. This porous structure is capable not only of increasing catalytic active areas of the material, but also of contributing to improve HER performance. TEM image in Fig. 2c reveals that Ni2P-Ni12P5 possesses a three-dimensional porous structure composed of nanoparticles and nanosheets, in agreement with the results of its SEM image (Fig. 2b). The high-resolution TEM was further performed to confirm the boundary of Ni2P and Ni12P5, as shown in Fig. 2d. We can see that the lattice fringes with distances of 0.523 and 0.343 nm correspond to (100) and (001) planes of Ni2P (PDF#74–1385), respectively, and the lattice fringe with distance of 0.615 nm corresponds to the (110) plane of Ni12P5 (PDF#74–1381), which are consistent with the XRD results. The boundary of Ni2P and Ni12P5 can be clearly observed, proving the successful preparation of the Ni2P-Ni12P5 heterostructure. The element composition and distribution of the material were analyzed by the EDX attached to the TEM instrument, as shown in Fig. 2e–h. It demonstrates the existence of Ni, P, and O elements in the material. The oxygen element may be caused by the slight surface oxidation of the material exposed in air.

Fig. 3
figure 3

aXPS survey spectrum of Ni2P-Ni12P5,b Ni 2p, and c P 2p spectra of Ni2P-Ni12P5 and Ni2P.

Full size image

The surface chemical compositions and valence of the Ni2P-Ni12P5 heterostructure were investigated by XPS. The XPS survey spectrum of Ni2P-Ni12P5 shows the presence of C, O, Ni, and P elements in the sample (Fig. 3a), which is basically consistent with the results of EDX. Ni 2p and P 2p spectra of Ni2P-Ni12P5 and Ni2P are shown in Fig. 3 b and c, respectively. In Fig. 3b, the two peaks of Ni2P-Ni12P5 located at 870.5 and 853.1 eV correspond to the Ni 2p1/2 and Ni 2p3/2 of \({\mathrm{Ni}}^{\delta +}\) species, and the two peaks appeared at 874.6 and 856.4 eV correspond to the Ni 2p1/2 and Ni 2p3/2 of Nix+ species, respectively. Nix+ species are assigned to nickel oxides due to the surface oxidation of the catalyst exposed in air, which is very common for the transition-metal phosphides[32, 35, 37]. Figure 3c shows that the two peaks of Ni2P-Ni12P5 at the binding energies of 129.4 and 130.1 eV are ascribed to P 2p3/2 and P 2p1/2, respectively. The peak appeared at 133.4 eV may be P-O species due to the superficial oxidation of the catalyst exposed in air [38, 39]. Compared with Ni2P, the peaks of Ni 2p in Ni2P-Ni12P5 shift towards higher binding energy, while the peaks of P 2p shift towards lower binding energy, suggesting that the electron interaction between Ni2P and Ni12P5 leads to the charge redistribution at the coupling interface [32, 40, 41].

Fig. 4
figure 4

HER performance of different samples in 0.5 M H2SO4 solution. a LSV curves, b corresponding Tafel slopes, c Nyquist plots, the inset shows Nyquist plots magnified, d electrochemical double-layer capacitance, and e LSV curves of the Ni2P-Ni12P5/NF before and after 2000 CV cycles; the inset shows the chronopotentiometric curve at a current density of 10 mA cm−2 for 20 h. f Actual H2 production versus theoretically calculated quantities for Ni2P-Ni12P5/NF in 0.5 M H2SO4 under constant current density of 10 mA cm−2

Full size image

In order to evaluate the HER electrocatalytic performance of Ni2P-Ni12P5/NF, the electrochemical tests were performed in 0.5 M H2SO4 and 1.0 M KOH using a standard three-electrode cell. Linear sweep voltammetry (LSV) plots in 0.5 M H2SO4 and 1.0 M KOH are shown in Figs. 4 and 5, respectively. The catalytic activities of NiO/NF, Ni2P/NF, and Pt/C/NF were also investigated for comparison. All of the polarization curves of the samples above were collected without iR correction. As shown in Fig. 4a, the overpotential of Ni2P-Ni12P5/NF is only 124 mV at a current density of 10 mA cm−2 that is obviously much lower than those of Ni2P/NF (308 mV) and NiO/NF (534 mV), exhibiting better HER catalytic activity. It may be due to the fact that heterostructure for Ni2P-Ni12P5/NF could provide more catalytic active sites, promoting the HER electrocatalytic performance of the material.

Fig. 5
figure 5

HER performance of different samples in 1.0 M KOH solution. a LSV curves, b corresponding Tafel slopes, c Nyquist plots, d electrochemical double-layer capacitance, e LSV curves of the Ni2P-Ni12P5/NF before and after 2000 CV cycles; the inset shows the chronopotentiometric curve at a current of 10 mA cm−2 for 20 h. f Actual H2 production versus theoretically calculated quantities for Ni2P-Ni12P5/NF in 1.0 M KOH under constant current density of 10 mA cm−2

Full size image

The electrocatalytic kinetics and mechanism for HER were investigated by the Tafel slope obtained via fitting the linear regions of Tafel plot, which is an important parameter to evaluate catalytic activity of the catalysts. According to the HER mechanism for electrolysis of water, the theoretical values of Tafel slope for Volmer, Heyrovsky, and Tafel reaction as the rate-limiting step are 120, 40, and 30 mV dec−1, respectively [42]. Figure 4b shows that the Tafel slope of Ni2P-Ni12P5/NF is 84.1 mV dec−1, indicating that the HER process undergoes Volmer-Heyrovsky mechanism [42, 43]. In addition, it is clearly observed that the Tafel slope of Ni2P-Ni12P5/NF is much smaller than those of Ni2P/NF (207.8 mV dec−1) and NiO/NF (273.4 mV dec−1). It demonstrates that the HER kinetics on the Ni2P-Ni12P5/NF is faster than the contrast samples, displaying good HER activity. The HER activity of Ni2P-Ni12P5/NF is better than that of most reported transition metal phosphide catalysts, as shown in Table S1.

Electrochemical impedance spectroscopy (EIS) is an effective analytical technique to study the catalytic reaction kinetics at electrode/electrolyte interface, which can disclose the charge-transfer properties [7, 44]. The charge-transfer resistance (Rct) was calculated through the diameter of the semicircular arc in the Nyquist plots, indicating the charge-transfer kinetics [13]. Figure 4c shows the Nyquist plots of the Ni2P-Ni12P5/NF, Ni2P/NF, and NiO/NF materials in the frequency range from 50 kHz to 0.01 Hz. It can be found that Ni2P-Ni12P5/NF shows a much smaller semicircle than those of Ni2P/NF and NiO/NF. The lower Rct of the Ni2P-Ni12P5/NF demonstrates its higher intrinsic catalytic activity and the faster charge-transfer rate at the electrode/electrolyte interface during the HER process [7, 45].

The electrochemically active surface area (ECSA) reveals the HER catalytic activity of catalysts, which is estimated by measuring the double-layer capacitance (Cdl) because of their positively proportional correlation [46,47,48]. The Cdl values were obtained by performing cyclic voltammetry (CV) measurements at different scanning rates of 30–150 mV s−1 under a non-Faradaic potential range in 0.5 M H2SO4 (Fig. S2). As shown in Fig. 4d, Ni2P-Ni12P5/NF exhibits a Cdl value of 13.9 mF cm−2, which is much higher than those of Ni2P/NF (0.42 mF cm−2) and NiO/NF (0.059 mF cm−2). The high Cdl value of the Ni2P-Ni12P5/NF heterostructure indicates the presence of abundant catalytic active sites, contributing to improve HER activity.

HER catalysts are required to possess not only good catalytic activity but also good stability. In practical applications, good stability means maintaining good catalytic activity for enough long time. So stability is an important index to evaluate the quality of the catalysts. The cycling stability of Ni2P-Ni12P5/NF was examined by continuous CV with a potential scan from − 0.2 to − 0.8 V at a scan rate of 50 mV s−1 for 2000 cycles. From the LSV curves in Fig. 4e, it can be seen that Ni2P-Ni12P5/NF shows negligible degradation after 2000 CV cycles, demonstrating its superior cyclic stability in acidic media during HER process. In addition, the long-term stability of Ni2P-Ni12P5/NF for HER was measured in 0.5 M H2SO4 using chronopotentiometry at controlled current density of 10 mA cm−2 for 20 h. As displayed in the inset of Fig. 4e, Ni2P-Ni12P5/NF exhibits a stable potential response for HER without significant degradation after continuous testing for 20 h, revealing preferable long-term stability.

Furthermore, the Faradaic efficiency (FE) of generated H2 for the Ni2P-Ni12P5/NF heterostructure was evaluated by comparing the theoretical value with the amount of H2 obtained from the experiment based on the current density of 10 mA cm−2, as shown in Fig. 4f. The amount of produced H2 was recorded by the drainage method and the experimental value is very close to the theoretical value.

Figure 5 shows the HER tests of different catalysts in 1.0 M KOH solution. It can be seen from Fig. 5a and b that the overpotential of Ni2P-Ni12P5/NF is 206 mV at a current density of 10 mA cm−2 and Tafel slope is 109.6 mV dec−1, which are obviously lower than those of Ni2P/NF (275 mV, 218.3 mV dec−1) and NiO/NF (428 mV, 194.1 mV dec−1), displaying better HER catalytic activity in alkaline medium. As shown in Fig. 5c, the Rct value of Ni2P-Ni12P5/NF is much lower than those of Ni2P/NF and NiO/NF, exhibiting its faster charge-transfer rate for HER in alkaline medium. The Cdl values were obtained by CV tests at different scanning rates of 30–150 mV s−1 in 1.0 M KOH solution (Fig. S3). As displayed in Fig. 5d, the Ni2P-Ni12P5/NF shows a higher Cdl value (0.96 mF cm−2) than Ni2P/NF (0.29 mF cm−2) and NiO/NF (0.22 mF cm−2), indicating its good HER activity. The stability of the Ni2P-Ni12P5/NF catalyst for HER was also measured in 1.0 M KOH. Figure 5e shows that the polarization curve obtained after 2000 cycles of the CV test displays slight degradation. The chronopotentiometry curve from the inset in Fig. 5e reveals that Ni2P-Ni12P5/NF maintains good activity after continuous testing for 20 h, displaying better long-term durability in alkaline medium. Figure 5f shows that the FE of produced H2 for Ni2P-Ni12P5/NF in 1.0 M KOH at current density of 10 mA cm−2 is almost 100%.

Based on the above electrochemical analysis results, the as-synthesized Ni2P-Ni12P5/NF heterostructure displays better HER electrocatalytic activity and outstanding stability, especially in acidic medium, which can be ascribed to its unique structure and composition. First, the unique heterostructure formed between Ni2P and Ni12P5 leads to a strong electronic interaction at the interface, thereby modulating the electronic structure [49, 50], optimizing the adsorption of hydrogen, increasing the number of active sites, and accelerating the charge transfer rate [33, 51, 52]. It is beneficial to improve HER electrocatalytic activity of the catalyst. Second, the three-dimensional porous nickel foam substrate with large surface area is beneficial for not only generating highly exposed catalytically active sites, but also providing many transport channels for accelerating interfacial charge transfer. In addition, the synergistic effect between the Ni2P-Ni12P5 heterostructure and nickel foam substrate effectively improves the HER electrocatalytic activity and stability of the as-synthesized catalyst.

Conclusions

In summary, to design an effective HER electrocatalyst, a three-dimensional porous self-supporting Ni2P-Ni12P5/NF heterostructure was successfully synthesized by hydrothermal reaction and low-temperature phosphorization. The as-synthesized Ni2P-Ni12P5/NF catalyst exhibits better catalytic activity and stability for HER, especially in acidic medium. It only requires the low overpotential of 124 mV to drive a current density of 10 mA cm−2 with a Tafel slope of 84.1 mV dec−1 in 0.5 M H2SO4, showing better electrocatalytic performance than Ni2P/NF. The synergistic effect between Ni2P-Ni12P5 with a unique heterostructure and nickel foam substrate with three-dimensional porous structure contributes to the improvement of HER catalytic activity and the stability of the catalyst.