1 Introduction

Exploring renewable carbon-free energy alternatives is one of the most promising pathways for alleviating the energy and environmental crisis [1, 2]. Hydrogen has been considered as one of the most promising clean energy carriers because of its high energy density and no pollutant product [3]. Moreover, electrochemical water splitting is a highly efficient sustainable hydrogen production route [4]. However, the corresponding hydrogen production process always requires excellent catalysts to achieve fast kinetics and lower the overpotential for HER. Up to now, although platinum (Pt) and other Pt-group metals exhibit the best catalytic activity for HER, the scarcity and high cost impede their widespread applications [5]. Thus, there is still an urgent demand for developing earth-abundant catalysts to replace these noble metal catalysts for effective HER.

Recently, a large variety of non-precious metal candidates, including transition metal carbides [6,7,8,9], phosphides [10,11,12,13], chalcogenides [14,15,16,17,18,19,20] and so on, have been investigated and shown striking HER performances. Among these alternatives, tungsten disulfide (WS2) has received persistent interest for its high activity, robust chemical stability, and earth-abundant resources [21]. However, both theoretical and experimental studies revealed that the edge state possesses a lower hydrogen adsorption Gibbs free energy, which means that HER performance mainly arises from the edge site, while the large-area basal planes are catalytically inert and useless for HER [22,23,24]. Moreover, as a semiconductor material, the activity of WS2 is primarily limited by its low electrical conductivity, which restricts charge transfer kinetics for HER [25]. Considering the above two factors, there are generally two routes to enhance the catalytic activity of WS2. One is to increase the density of the exposed active sites. Among large amounts of methods, structural designing is one of the most effective strategies for increasing the number of active sites [26, 27]. In our previous work [28], the WS2 nanosheets (NSs) with loosely stacked layers were successfully obtained by a mechanical activation strategy. This special nanostructure provides highly exposed rims and edges for HER. Cheng et al. [29] synthesized the ultrathin WS2 nanoflakes by the high-temperature solution-phase method. The obtained catalyst possessed abundant edges and the ultrathin thickness. By directly vulcanizing the WO3 nanosheets, Shang et al. [30] fabricated a WS2/WO3 heterostructure which could expose abundant active sites for HER. The other route is to improve the electrical conductivity. Sun et al. [31] synthesized the N-doped WS2 nanosheets by one-step sol-gel process and found that N doping in WS2 might be an effective way to improve the intrinsic conductivity of WS2. Duan et al. [32] fabricated 3D WS2 nanolayers@heteroatom-doped graphene films via a vacuum-filtration process, and the conductive network of graphene sheets greatly accelerates the charge transfer kinetics. Despite these developments, the HER activity of WS2 is still much lower than that of Pt. Thus, it is still a big challenge to further improve the HER performance of WS2.

Herein, the flower-like WS2 nanosheets were synthesized via a facile one-pot hydrothermal method. This unique nanostructure not only can greatly increase the edge sites and defects, resulting in a high density of exposed active sites, but also can make the acidic electrolyte easily accessible to the surface of WS2 and accelerate the electron transfer rate. Moreover, the HER performance of WS2 is improved by controlling the reaction temperature. As a result, the optimized WS2 NSs exhibit a good HER activity in an acidic solution with a small Tafel slope of 70 mV dec−1 and a good stability.

2 Experimental

2.1 Materials

Sodium tungstate dihydrate (Na2WO4·2H2O), sodium hypophosphite (NaH2PO2·H2O), thioacetamide (CH3CSNH2) and bulk WS2 were purchased from Aladdin. The Pt/C catalyst (20 wt.%) was purchased from the Johnson Matthey and Nafion solution (5 wt.%) was purchased from the DuPont.

2.2 Synthesis of WS2

Typically, 0.99 g Na2WO4·2H2O, 0.32 g NaH2PO2·H2O and 1.13 g CH3CSNH2 were dissolved into 50 mL distilled water. After being stirred to form a transparent solution, 0.75 mL concentrated HCl was dropped into the solution and continuously stirred for 30 min. Then, the solution was transferred into a 100 mL Teflon-lined stainless-steel autoclave, and heated in an oven maintained at 180–220 °C for 24 h. After being naturally cooled to room temperature, the as-prepared samples were obtained by centrifugation, washed by absolute ethanol and deionized water for several times, and dried at 80 °C overnight.

2.3 Preparation of working electrodes

In a typical procedure, 3 mg of the catalyst was added into the solution containing 80 μL Nafion solution (5 wt.%), 0.2 mL absolute ethanol and 0.8 mL deionized water. After being sonicated for 30 min, 5 mL of the formed homogeneous slurry was dropped onto a smooth glassy carbon electrode (GCE) of 3 mm in diameter, generating a working electrode with a loading of 213 μg cm−2.

2.4 Physical measurements

The powder XRD patterns were recorded using the Mini Flex 600 system, and the Raman spectroscopy was obtained on the LabRAMHR-800 (French company HRIBA). The transmission electron microscopy (TEM) images and the scanning electron microscopy (SEM) images were acquired on the Tecnai G2 F20 and the FEI Sirion 200, respectively. The X-ray photoelectron spectroscopy (XPS) was conducted on the ESCALAB 250Xi and the correspond binding energy values were corrected via the C1s.

2.5 Electrochemical measurements

A CHI660E workstation (CH Instruments, China) with a standard three-electrode system was used to perform all the electrochemical tests. The modified GCE was used as the working electrode, and the carbon rod and the saturated Hg/Hg2Cl2 electrode (SCE) were served as the counter electrode and the reference electrode, respectively. The linear sweep voltammetry (LSV) was carried out in 0.5 mol L−1 H2SO4 using a sweep rate of 2 mV s−1 from 0.1 to − 0.4 V (vs. SCE). The long-term cycling test was performed via the cyclic voltammetry (CV) measurements with a sweep rate of 50 mV s−1 from 0.1 to − 0.4 V for 2000 cycles. In addition, the electrochemical impedance spectroscopy (EIS) was conducted between the frequency range of 106 Hz and 1 Hz with a sinusoidal perturbation of 5 mV at an overpotential of − 300 mV.

2.6 Electrochemical measurements

The electrochemically capacitance measurements were conducted by cyclic voltammograms from 0.00 to 0.20 V with various scan rates (20, 40, 60, 80, 100, 120, 140 mV s−1). The capacitive currents were measured in a potential where no faradic processes were observed, i.e., at 0.1 V versus RHE. According to the previous report [14], the specific capacitance, a flat standard with 1 cm2 of the real surface area, is approximately 60 μF cm−2. Thus, the electrochemical active surface area can be calculated by the following Eq. (1):

$$A_{\text{ECSA}} = \frac{\text{electrochemical capacitance}}{{60\;\upmu{\text{F}}\;{\text{cm}}^{ - 2} \;({\text{per cm}}_{\text{ECSA}}^{2} )}}$$
(1)

3 Results and discussion

3.1 Microstructural characterization

The XRD pattern of WS2 NSs is shown in Fig. 1a. The peaks at 14.2°, 32.1°, 35.4°, 43.4° and 56.9° can be assigned to the (002), (100), (102), (006), and (110) planes of WS2 (JCPDS No. 08-0237), respectively. There are no major impurities detected, suggesting the high purity of the obtained WS2. Besides, the XRD pattern of bulk WS2 is also given for a comparison, which is well consistent with the standard WS2. Interestingly, compared with the bulk WS2, all peaks of WS2 NSs are weaker and broader, indicating a poor crystallinity and disordered structure [33]. Raman spectroscopy was conducted to further confirm the structure of WS2 NSs. As shown in Fig. 1b, there are two pronounced peaks at ~ 50 and 420 cm−1, which are assignable to the E 12g modes and the A1g modes of 2H-WS2, respectively [34, 35]. Compared with the bulk WS2, the Raman peaks of WS2 NSs show slight shifts, which can be attributed to the possible changes in the bond polarization [36].

Fig. 1
figure 1

a XRD patterns and b Raman spectra of the WS2 NSs and bulk WS2

Full size image

The morphologies of as-prepared WS2 and bulk WS2 are shown in Fig. 2. It can be seen from Fig. 2a, d that the as-prepared WS2 exhibits a flower-like morphology, which is consisted of a large number of loosely assembled thin nanosheets. The high-resolution TEM (HRTEM) image (Fig. 2e) of the WS2 NSs reveals that the interlayer spacing of the WS2 (002) planes is ~ 0.62 nm. In Fig. 2b, the pristine bulk WS2 shows large and thick bulk morphology. The HRTEM image (Fig. 2c) suggests that the basal plane of the bulk WS2 possesses high crystallinity and no defects. As for the basal plane of WS2 NSs (Fig. 2f), many defects (the yellow selection in Fig. 2f) can be observed, which may be beneficial for HER as possible active sites.

Fig. 2
figure 2

a SEM image of WS2 NSs; b, c TEM and HRTEM images of bulk WS2; d–f TEM and HRTEM images of WS2 NSs

Full size image

To get insights into the surface chemical valence states of WS2 NSs, the XPS analysis was further performed (Fig. 3). As shown in Fig. 3a, the high-resolution W 4f spectra can be fitted into two doublets. The doublets located at binding energies of 38.2 eV (W 4f5/2) and 36.1 eV (W 4f7/2) are originated from W6+ in WO3, which can be ascribed to surface oxidation of WS2 [25, 30, 37]. The peaks located at 34.4 eV and 32.3 eV are assigned to 4f5/2 and 4f7/2 of W4+, respectively, which are consistent with the presence of WS2 [38, 39]. Correspondingly, the doublets at around 163.0 eV and 161.8 eV (Fig. 3b) are indexed to 2p3/2 and 2p1/2 of S, respectively, which can be attributed to S2− species in WS2 [40, 41].

Fig. 3
figure 3

XPS spectra of WS2 NSs: a W 4f and b S 2p

Full size image

3.2 Insights into the HER kinetics

The electrochemical performance of the bulk WS2 and WS2 NSs for HER was detected by the LSV measurement in 0.5 mol L−1 H2SO4. The polarization curves of different catalysts are shown in Fig. 4a, and the key electrochemical results are also displayed in Table 1. The commercial Pt/C catalyst (20 wt.%, Johnson Matthey) was also measured as a reference. The bulk WS2 exhibits a poor HER performance with a high onset potential (the current density reaches 1 mA cm−2) of − 396 mV, and low current density (j−400) of 1.1 mA cm−2 at a specific overpotential of − 400 mV. In comparison, the WS2 NSs exhibits a good HER activity with a low onset potential of − 306 mV, and a high j−400 of 8.5 mA cm−2, which is ~ 7 times higher than that of the bulk WS2.

Fig. 4
figure 4

a Polarization curves, b Tafel slopes and c Nyquist plots of the WS2 NSs and the bulk WS2; d Polarization curves of the WS2 NSs and after 5000 cycles

Full size image
Table 1 Summary of information over various WS2 catalysts
Full size table

To further study the catalytic activity of the WS2 catalysts, the Tafel slopes were extracted from the corresponding LSV curves through the Tafel equation (η = blog j + a, where a is the onset potential and b is the Tafel slope) (Fig. 4b). The WS2 NSs exhibit a Tafel slope of 70 mV dec−1, which is remarkably smaller than that of bulk WS2 (112 mV dec−1), indicating a more efficient HER process [42, 43]. Generally, in acid media, it is accepted that the HER process involves two steps. The electrochemical adsorption of hydrogen (Volmer reaction, H+ + M + e− → M − H*, where M is the catalysts) is the first step. While, the desorption of adsorbed hydrogen is the second step, which includes the electrochemical desorption (Heyrovsky reaction, M − H* + H+ + e− → M + H2) or the chemical desorption (Tafel reaction, M − H* + M − H* → 2M + H2). According to the classic theory, the Tafel slope can be used to evaluate the rate-determining step in the HER process [44,45,46]. Correspondingly, the Tafel slope values will reach 118, 29 and 39 mV dec−1 if the Volmer, Tafel, and Heyrovsky reactions are the rate-determining steps, respectively. The small Tafel slope of 70 mV dec−1 denotes that the WS2 NSs undergo a mixed Volmer-Heyrovsky mechanism and the Heyrovsky reaction acts as the rate-determining step.

To get insights into electrode kinetics of the HER process, the EIS measurements were performed at a constant voltage of − 300 mV. A simplified equivalent circuit (inserted in Fig. 4c) was utilized to fit the Nyquist plots. The charge transfer resistance (Rct) can be determined from the semicircle. The Rct value of WS2 NSs (244 Ω) is much smaller than that of bulk WS2 (5231 Ω), suggesting a faster electron transfer rate and superior kinetics of WS2 NSs in HER process [47]. The small Rct value of WS2 NSs can be attributed the sheet-like nanostructure, which can make the acidic electrolyte easily accessible to the surface of WS2, resulting in a faster electron transfer rate [48, 49]. On the contrary, the bulk structure of bulk WS2 slows down the electron transfer rate, thus, leading to a large Rct value.

Besides, the stability of catalysts is an important indicator for evaluating the electrocatalytic performance. The long-term cycling measurement of the WS2 NSs was adopted using the CV. As shown in Fig. 5a, only a negligible shift of the current density can be observed after scanning for 2000 cycles, implying the good catalytic stability of WS2 NSs. The time-dependent current density curve (I–t curve) was also obtained (Fig. 5b). Only a slight degradation of the current density over 50 h further proves its good stability. The SEM and TEM images of the WS2 NSs after scanning for 2000 cycles are shown in Fig. 6, and there are few changes in the morphology after the stability test.

Fig. 5
figure 5

a Polarization curves of the WS2 NSs and after 2000 cycles; b I–t curve of the WS2 NSs for 50 h

Full size image
Fig. 6
figure 6

a SEM image and b TEM image of WS2 NSs after scanning for 2000 cycles

Full size image

The ECSA of WS2 catalysts was estimated by testing the double-layer capacitance (Cdl) at non-faradaic potentials (Fig. 7). A higher ECSA value means a larger density of active sites for WS2. Figure 7c is the measured current density plotted as a function of the scan rate, in which the Cdl value is equivalent to the slope value. A direct comparison of these two samples can be obtained because the ECSA value is proportional to the Cdl value. One can see that the Cdl value of the WS2 NSs is 11.2 mF cm−2, whereas the Cdl value of the bulk WS2 is 0.2 mF cm−2. Accordingly, the WS2 NSs exhibits a high ECSA value of 190 cm 2ECSA ~ 57 times as high as that of the bulk WS2 (3.33 cm 2ECSA ), suggesting that the WS2 NSs possess plenty of edge sites and defects that will greatly increase the density of active sites.

Fig. 7
figure 7

Electrochemical capacitance measurements: a WS2 NSs and b bulk WS2; c Measured capacitive current densities (at 100 mV vs. RHE) plotted as a function of the scan rate for the WS2 NSs and bulk WS2

Full size image

Due to the hydrothermal reaction being sensitive to the preparation temperature, the effect of the preparation temperature is also been discussed. All the as-prepared WS2 NSs are named as WS2-XXX for convenience, where XXX represents the preparation temperature in the unit of °C. In Fig. 8a, one can see that the HER activity of WS2 NSs firstly increases with the increase of the preparation temperature and reaches the maximum at 200 °C, and then decreases at higher temperatures, which means that 200 °C is the most suitable temperature for the formation of the high-efficiency HER catalyst in WS2 NSs. Similar regulations can also be found in Tafel and Nyquist plots (Fig. 8b, c). In general, as the preparation temperature increases, the crystallinity of the as-prepared catalysts is improved, leading to a decrease in the number of defects and a corresponding decrease in the number of active sites. While, on the other hand, a higher crystallinity of the WS2 NSs could result in a better conductivity [50]. Thus, we believe that a balance is achieved between the number of active sites and conductivity at 200 °C, resulting in the best HER performance.

Fig. 8
figure 8

a Polarization curves, b Tafel slopes and c Nyquist plots of the WS2 NSs obtained at various temperatures

Full size image

As a result, the good HER activity of the optimized WS2-200 can be ascribed to a sheet-like structure, which not only can greatly increase the number of exposed active sites, but also can make the acidic electrolyte easily accessible to the surface of WS2 and accelerate the electron transfer rate. Besides, the appropriate preparation temperature is also important, which makes the WS2-200 possess a relatively large number of active sites and high conductivity, in the same time.

4 Conclusion

In summary, a flower-like WS2 catalyst is successfully synthesized via a facile one-pot hydrothermal method. The unique nanostructure can greatly increase the edge sites and defects, resulting in a high density of exposed active sites. Compared with bulk WS2, the WS2 NSs show a high ECSA value of 190 cm 2ECSA , ~ 7 times as high as that of bulk WS2; and a high j−400 of 8.5 mA cm−2, ~ 7 times higher than bulk WS2 and a lower Tafel slope of 70 mV dec−1. Moreover, the good stability also makes it a promising alternative to replace precious-metal HER catalysts. Besides, the effect of preparation temperature is also discussed, and 200 °C is found to be the best preparation temperature for WS2 NSs.