1 Introduction

To address the threats of the energy crisis and environmental pollution from the increasing utilization of fossil fuels, great efforts have been devoted to exploring clean and renewable energy sources [1, 2]. Hydrogen, due to its high energy density, non-polluting and renewable features, has long been advocated as an excellent alternative to fossil fuels [3,4,5]. Electrocatalytic water splitting is considered as the most promising and desirable technology by researchers and entrepreneurs for hydrogen production [6,7,8]. So far, noble-metal-based materials (such as Pt, Ir and Ru-based catalysts) have been extensively investigated for hydrogen production. Unfortunately, the scarcity and high price limit their broad applications [9,30].

Inspired by the previous results, this work reports the synthesis of B, N dual-doped carbon layer encapsulated on Mo2C nanocrystals (Mo2C@BNC) via a facile one-pot pyrolysis method for alkaline HER. The introduced B atom in the carbon layer is aimed to optimize the carbon shell electronic structures for H2O adsorption. As expected, the prepared Mo2C@BNC catalyst displays remarkable electrocatalytic activity towards HER under alkaline conditions. More importantly, the catalytic activity is even higher than that of the commercial 10% Pt/C catalyst at large current density. Theoretical calculations reveal that the H2O could be decomposed spontaneously over the introduced B sites, and the defective C atoms in the dual-doped carbon layer provide the best H binding sites. The synergistic effect between the non-metal sites (B and C) on the carbon shell optimize the rates of Volmer and Heyrovsky reaction, simultaneously, leading to the superb performance of the catalyst.

2 Experimental Methods

2.1 Experimental Reagents

Ammonium molybdate tetrahydrate ((NH4)6Mo7O24·4H2O), dicyandiamide (C2H4N4), boric acid (H3BO3), KOH are analytical grade and without further purification. Millipore water (resistivity: ∼ 18 MΩ cm) was used in all experiments.

2.2 Synthesis of Mo2C@BNC

Mo2C@BNC was synthesized via two steps. First, 1 g of C2H4N4 was dissolved in 10 mL of deionized water, stirred for 15 min, then added 50 mg of (NH4)6Mo7O24·4H2O with 5 mg of H3BO3, heated and stirred until a white crystalline powder was formed. The white crystalline powder was collected and ground in an agate mortar for 15 min. In the second step, the powder was heated under an H2/Ar (5:95) atmosphere at 500 °C for 30 min with a heating rate of 5 °C min−1, then heated at 800 °C for another 6 h with a heating rate of 10 °C min−1.

2.3 Synthesis of Mo2C@NC

The Mo2C@NC was prepared using the same method as the Mo2C@BNC except that the H3BO3 was not added.

2.4 Synthesis of Mo2C@C

50 mg of (NH4)6Mo7O24·4H2O was mixed with 1 g of glucose and ground in an agate mortar for 15 min, then the homogeneous mixture was placed in a quartz tube and heated under an Ar atmosphere at 800 °C for 6 h with a heating rate of 5 °C min−1.

2.5 Materials Characterizations

The phase structures of all samples were identified by Shimadzu X-ray diffraction (XRD)-7000 powder X-ray diffractometer (Cu Ka). The morphologies and structures of the products were characterized by Tecnai G2 F30 S-TWIN transmission electron microscopy (TEM). X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo ESCALAB250**s show that the Mo, C, N and B are homogeneously distributed (Fig. 2d). The Energy Dispersive X-Ray Spectroscopy (EDX) results show the do** contents of B and N are 10.32 and 8.42 wt% (Table S1), respectively, confirming the successful introducing of B and N elements into the catalyst. During the one-pot pyrolysis, the dicyandiamide not only serves as a nitrogen source to generate N-doped carbon shell that encapsulates the Mo2C nanocrystal, but also regulates the electronic structure of the catalyst surface and promotes the conductivity and stability of the catalyst. As control, when using glucose as the carbon source, the prepared Mo2C@C fails to form a regular morphology and exists in an amorphous material instead (Fig. S1).

Fig. 1
figure 1

Schematic of synthesis N, B dual-doped Mo2C nanocrystals (Mo2C@BNC)

Full size image
Fig. 2
figure 2

Structural characterization of the materials. a, b TEM images of the Mo2C@BNC, inset of b: the particle size distribution of the Mo2C nanocrystals. c HRTEM images of the Mo2C@BNC. d Elemental map** of the Mo2C@BNC. e XRD pattern of the Mo2C@C, Mo2C@NC, and Mo2C@BNC. f Raman spectra of the Mo2C@NC and Mo2C@BNC. g FT-IR spectra of the Mo2C@NC and Mo2C@BNC

Full size image

XRD, FT-IR and Raman patterns were further employed to explore the detailed structure of the Mo2C@BNC. As shown in Fig. 2e, the peaks centered at 33.6°, 37.4°, 39.5°, 52.1°, 61.6° and 74.8° correspond to the (100), (002), (101), (102), (110) and (112) crystal planes of the hexagonal β-Mo2C (PDF#35-0787), respectively. No other peak can be observed in the Mo2C@BNC XRD pattern, indicating the boron and nitrogen mainly introduced into the carbon host of the Mo2C crystals. In the Raman spectra (Fig. 2f), vibration peaks around at 649.8, 818.7 and 991.1 cm−1 are fingerprint bands characteristic, confirming the formation of the Mo2C phase [34]. Additional peaks located at 1348 and 1595 cm−1 are attributed to the characteristic D and G bands of carbon materials, respectively. The AreaD/AreaG ratio increases from 2.2 of the Mo2C@NC to 2.4 of the Mo2C@BNC, implying that the disorder of the graphite structure increases after the introduction of the extra B atoms [35]. The FT-IR spectrum of the Mo2C@BNC sample (Fig. 2g) shows a peak centered at 734 cm−1, corresponding to the out-of-plane vibration of the hexagonal B–N bond [36], which is not found in the Mo2C@NC spectrum. Besides, the peaks located at 1209.3 and 1297.6 cm−1 belong to the characteristic C−B bond and C−N bond, respectively [37, 38]. These results confirm the unique structure of the Mo2C@BNC, in which the Mo2C nanocrystals are encapsulated by an ultrathin B, N dual-doped carbon layer.

XPS analysis reveals the chemical states and electronic structures of the samples. In the Mo 3d spectra (Fig. 3a), the peaks at 228.9 and 231.98 eV correspond to the binding energies of Mo2+, which are typical characteristics of the Mo2C. The peaks at 229.89 and 233.26 eV correspond to the binding energies of Mo4+, while the peaks at 232.4 and 235.98 eV correspond to the binding energies of Mo6+, respectively, which are assigned to Mo species that oxidized in the preparation process [39, 40]. Compared with that of the Mo2C@NC sample, the Mo 3d XPS spectrum of the Mo2C@BNC presents a positive shift of 0.12 eV. It indicates the enhanced electron delocalization of the interior Mo2C after boron introduction, which facilitates the charge redistribution of the exterior BNC shell. In the N 1s spectrum (Fig. 3b), four peaks at 395.2, 398.4, 399.6 and 401.5 eV are observed, which can be assigned to the binding energies of Mo 3p, pyridinic N, pyrrolic N and graphitic N, respectively, further demonstrating the successful do** of N into the carbon layer. An additional peak is observed at 397.8 eV in the N 1s spectrum after B do**, which is attributed to the B-N binding energy [41]. In this work, according to the XPS spectrum, the pyridinic N content of the Mo2C@BNC is 38.97%, far greater than that of the Mo2C@NC (24.50%), implying that the electron structure of Mo2C@BNC has been greatly modified (Table S2). This means that the do** of B decorated the electronic structure of the exterior carbon shell. In the C 1s spectrum (Figs. 3c and S2), strong peaks can be observed for the C-C species at 284.8 eV, other peaks at 283.7, 286.4 and 288.3 eV correspond to the C-B, C-N and C-N-O binding energies [42]. The B 1s spectrum (Fig. 3d) can be observed with three peaks located at 189.8, 191.5 and 192.8 eV, corresponding to the B-C, B-N and B-O binding energies, respectively [43]. The co-existence of the B-C and B-N species indicates the partial replacement of C atoms with B atoms in the carbon layer. It’s also affirmed that partial electrons in the Mo2C@BNC are transferred from Mo to the carbon layer after the introduction of element B, and the dual-doped B and N elements enables the activation of the adjacent C atom due to electron restructuring, facilitating the hydrogen evolution reaction [44].

Fig. 3
figure 3

XPS spectra of Mo2C@NC and Mo2C@BNC: a Mo 3d. b N 1s. c C 1s XPS spectrum of Mo2C@BNC. d B 1s XPS spectrum of Mo2C@BNC

Full size image

3.2 Electrocatalytic Performance

The HER activities of the prepared Mo2C@BNC catalyst were determined in 1 M KOH. The catalytic performances of the Mo2C@NC, Mo2C@C and commercial 10% Pt/C catalysts were also measured in the same condition as control. As shown in Fig. 4a, the designed Mo2C@BNC catalyst displays an excellent HER catalytic activity with an over potential of 99 mV to deliver the current density of 10 mA cm−2, which is significantly lower than those of the Mo2C@NC (144 mV) and Mo2C@C (259 mV), indicating that B do** can substantially enhance the HER electrocatalytic activity. We have also investigated the catalytic activity of the Mo2C@BNC catalysts with various B do** concentrations, provided an optimized B do** level in this study (Fig. S3). Moreover, the Mo2C@BNC shows a robust catalytic activity at large current density that even superior to the commercial 10% Pt/C catalyst (Fig. 4a, c). The overpotential is 168 mV for Mo2C@BNC to achieve 100 mA cm−2, 20 mV less than that of the commercial 10% Pt/C (188 mV), demonstrating its great potential in industrial applications. The performance of the Mo2C@BNC surpasses most state-of-the-art reported Mo2C-based materials (Table S3), addressing one of the best catalytic activities among Mo2C-based catalysts ever reported.

Fig. 4
figure 4

HER electrocatalytic properties of the Mo2C@BNC catalyst. a LSV curves of Mo2C@C, Mo2C@NC, Mo2C@BNC and Pt/C in 1 M KOH electrolyte. b Tafel slopes. c Overpotentials at 10 and 100 mA cm−2. d EIS curves. e The polarization curves normalized by ECSA. f Constant potential V–t curves of Mo2C@BNC

Full size image

Furthermore, the Tafel slope is used to study the catalytic kinetics during the reaction process [45]. HER usually comprises two possible steps in alkaline media, as shown in the following equation:

$${\text{H}}_{2} {\text{O}} + {\text{e}}^- \to {\text{H}}^* + {\text{OH}}^- \;\;\;\;\left( {\text{Volmer reaction}} \right)$$
$${\text{H}}^* + {\text{H}}_{2} {\text{O}} + {\text{e}}^- \to {\text{H}}_{2} + {\text{OH}}^- \quad ({\text{Heyrovsky reaction}})$$

or

$${\text{H}}^* + {\text{ H}}^* \to {\text{H}}_{2} \quad ({\text{Tafel reaction}})$$

where H* refers to the H atom adsorbed on the surface of the catalysts. HER is separated into the water-cleaving (Volmer reaction) and the hydrogen coupling (Tafel reaction or Heyrovsky reaction) steps, resulting in two reaction pathways, namely the Volmer–Heyrovsky or Volmer–Tafel mechanism. In general, the rate-determining step (RDS) can be understood by the Tafel slope, which reflects the electron transfer capacity between the H* and the catalyst. The Tafel slope of 30 mV dec−1 shows that the RDS is Tafel step and the reaction follows the Volmer–Tafel pathway. The Tafel slope of 40 mV dec−1 indicates that the Heyrovsky step is the RDS and the reaction follows the Volmer–Heyrovsky pathway. The Tafel slope of 118 mV dec−1 means that Volmer step is the RDS and the reaction may follow the Volmer–Heyrovsky or Volmer–Tafel pathway [46]. As shown in Fig. 4b, the calculated Tafel slope of the Mo2C@BNC for HER is 58.1 mV dec−1, smaller than those of the Mo2C@NC (71.3 mV dec−1) and Mo2C@C (88.5 mV dec−1), demonstrating the superior kinetics of the Mo2C@BNC among the catalysts. The Tafel slope value of the Mo2C@BNC also implies that the reaction proceeds through a Volmer–Heyrovsky pathway, during which the Heyrovsky step is the rate-determining step. Meanwhile, we further confirm the superior adsorption of Mo2C@BNC on H by testing HER performance in acidic solution (Fig. S4). Besides, the EIS results (Fig. 4d) measured at − 0.2 VRHE confirm that the Mo2C@BNC has the smallest charge-transfer resistance in the reaction.

The electrochemically active surface area (ECSA) of the catalyst is measured to further evaluate the effect of surface area on the electrochemical performance. ECSA is proportional to the electrochemical double-layer capacitance (Cdl) of the material (Fig. S5). The Mo2C@BNC shows a great ECSA value, indicating that the B do** can expose more active sites and contribute to the HER catalytic activity [43]. Besides, the polarization curves were normalized by ECSA (JECSA) to explore the effect of intrinsic activity, as shown in Fig. 4e. Mo2C@BNC exhibits the highest JECSA among all the samples, demonstrating the intrinsically enhanced activity. The stability of the Mo2C@BNC was tested by chronopotentiometry at a constant density of 10 mA cm−2 (Fig. 4f). There is a tiny increment of potential during the 12 h reaction, and no significant change in the morphology and structure of the catalyst after the durability test (Fig. S6), demonstrating the favorable stability of the catalyst. This is attributed to the fact that the carbon layer is wrapped with Mo2C nanocrystals, which can effectively prevent the electrolyte from corroding the catalyst and greatly maintain the catalytic activity [47].

3.3 Mechanism Analysis of the Enhanced Catalytic Performance

DFT calculations were performed to gain atomic insights into the synergistic effect of B do** on the HER activity. Figure 5a–b revealed that H2O is unfavorable for the adsorption on the Mo2C@NC substrate. However, the introduced B endows the superior ability of H2O adsorption and the following spontaneous decomposition on the Mo2C@BNC. This is confirmed by the hydrophilicity test of the samples. Mo2C@BNC possessed the best hydrophilicity with a water contact angle of 20.4°, which was smaller than that of Mo2C@NC (25.41°, Fig. S7), confirming the Mo2C@BNC exhibits better hydrophilicity than the Mo2C@NC. During this process, the reaction intermediates H* and OH* are prone to be adsorbed on the defective C site and B site, respectively. To identify the detailed roles of the dopants in the reaction, we calculated the possible adsorption of H* and OH* species on the theoretical models. As shown in Fig. 5c, the free energy for H* adsorption (ΔGH*) on Mo2C@C surface is evaluated to be 1.87 eV, which is quite unfavorable for the hydrogen evolution reaction (Fig. S8a). The ΔGH* on Mo2C@NC surface is 1.7 eV, implying that the N dopants could adjusted the electron distribution of external carbon layer, thus slightly improving the free energy of H* adsorption (Fig. S9a). In contrast, the Mo2C@BNC displays a Pt-like H* adsorption energy of − 0.085 eV (the C site) after B do**, which is an optimal state close to a thermo-neutral value. Furthermore, H* free energies on different adsorption sites for the Mo2C@BNC surface were calculated to identify the active site in the reaction (the insert in Fig. 5d). The chosen theoretical model for the Mo2C@BNC is shown in the inset of Fig. 5d, in which the C site adjacent to B (Fig. S9b) is the most effective site for H* adsorption (Fig. 5d). The significantly improved adsorption ability is attributed to the Co-introduction of B and N. When the electrons transfer from Mo to the carbon layer, due to the different electronegativity (B < C < N), the electrons will be redistributed between those non-metallic atoms, which can effectively activate the C atoms, making it a near-zero H* adsorption Gibbs free energy.

Fig. 5
figure 5

DFT calculation results of the HER process on the catalyst surfaces. H2O cleaving process on the surfaces of a Mo2C@NC and b Mo2C@BNC. c OH* and H* adsorption energies for the Mo2C@NC and Mo2C@BNC. d Calculated H* free energy (ΔGH*) for different sites. Yellow: C atom, blue: N atom, pink: B atom, cyan: Mo atom. (Color figure online)

Full size image

Furthermore, the do** of B also strengthens the adsorption energy of OH* (Fig. 5c). The best adsorption site for OH* of the Mo2C@C and Mo2C@NC sample is the C site that near the H* adsorption site (Figs. S8b and S9c). After B do**, the electron-deficient B atoms adsorbs water by coordinating with the lone pair of electrons, thus weakening the O−H bond to accelerate water dissociation. The best reaction center on the Mo2C@BNC surface becomes the B site (Fig. S9d). This is evidenced by the distance between the O atom from H2O and the active sites. It is calculated that the distance between the O atom and the adsorbed C site (dO-C) is 1.71 Å for Mo2C@C and 2.186 Å for Mo2C@NC. The value changes to 1.457 Å for the distance between the O atom and the adsorbed B site (dO-B) for Mo2C@BNC. As a result, the adsorption energy of the OH* increases from − 4.0 and − 3.74 eV over the primary C site to − 4.54 eV over the B site. The enhanced adsorption energy of H* and OH* after the introduction of B is more favorable for H2O decomposition, which is clarified from the changes of density of projection state (DOS) of the Mo2C@BNC and Mo2C@NC samples (Fig. S10). Compared with the B-free sample, the H*-C and OH*-B interactions are strengthened in the B-do** sample. In short, DFT calculations confirm that the B do** greatly enhances the Volmer reaction in the HER process, making water cleaves into H* and OH*. At the same time, the inert C atom is activated together with the B and N atoms and facilitated the Heyrovsky step. The synergistic effect between these non-metal B site and C site over the carbon shell thus greatly improving the HER activity.

4 Conclusions

In summary, we have designed a Mo2C@BNC catalyst that N, B dual-doped ultrathin carbon shell encapsulated on Mo2C nanocrystals. The do** of N and B elements coupled with defects on the surface of the carbon layer, generates multiple active centers, which significantly enhances HER activity compared to the pristine samples. This catalyst delivers a current density of 10 mA cm−1 with a very low over-potential of 99 mV. More importantly, the catalytic activity even exceeds that of the commercial Pt/C catalyst at large current density, addressing one of the best non-noble metal based HER catalyst in alkaline solutions. Theoretical calculations reveal that the introduction of B can trigger the spontaneous decomposition of H2O and facilitate the Volmer reaction, simultaneously activating the adjacent C atom, resulting in a near-zero regulation of ΔGH*. This paper demonstrates the importance of multiple active centers concerning the alkaline HER activity and provides fresh insight to explore the acceleration of water dissociation via the effect of boron.