Preparation and characterizations of multi-heterojunction interface electrocatalysts. The formation process of multi-heterojunction interface electrocatalysts is schematically illustrated in Fig 1g. 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 catalysts were synthesized by a three-step procedure. First, 1T0.41-MoS2 catalyst was obtained on carbon cloth (CC) by acid-induced hydrothermal approach at 200 oC for 12 h (see details in “Methods” section). The as-obtained 1T0.41-MoS2 catalyst shows a large number of microspheres (Supplementary Fig. 1b-d) with a narrow diameter distribution of 2.0 ~ 4.0 µm distributed uniformly on the surface of CC substrate. Flower-shaped MoS2 microspheres are consisted of many aligned 1T0.41-MoS2 nanosheets, on which the Ni(OH)2 nanoparticles were then electro-deposited (see details in “Methods” section). 1T0.41-MoS2@Ni(OH)2 material inherited its morphology from spherical MoS2. Subsequently, 1T0.41-MoS2@Ni(OH)2 material was loaded into a quartz tube mixed with red phosphorus or sulfur powder and sealed by oxyacetylene flame. Finally, these were heated to 600 oC for the reaction with red phosphorus or sulfur to synthesize 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 catalysts, respectively (Supplementary Fig. 2, 3). As to 1T0.81-MoS2@Ni2P catalyst, the MoS2 microspheres are very rough, on which there distributes many random Ni2P nanoparticles (Supplementary Fig. 3). It is because that the 1T/2H mixed-phase and heterojunction-interface structure reduces the adhesion of the gas-solid interface and facilitates releasing hydrogen from the catalyst surface, which is essential for enhancing HER.29
Next, the phase composition and crystal properties of 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 were obtained by X-ray diffraction (XRD) and Raman spectroscopy. There are some obvious characteristic diffraction peaks of 14.3°, 33.4°, and 59.2° (Supplementary Fig. 4a), which can be ascribed to 2Hphase-MoS2 according to the JCPDS card number #37-1429. However, the XRD peak of 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 located at 2θ ≈ 28.8° can be indexed as the (004) peak of 1Tphase-MoS2, which indicates that 1T- and 2H-mixed phases were successfully hydrothermally synthesized.35 The other characteristic peaks (2θ ≈ 31.3°, 35.2°, 38.8°, 44.9°, and 53.3°) demonstrate that the 1T0.72-MoS2@NiS2 is a hybrid of NiS2 (JCPDS#11-0099), which verifies the presence of NiS2 nanoparticles. Similarly, as to 1T0.81-MoS2@Ni2P catalyst, its XRD results also showed the presence of Ni2P nanoparticles (JCPDS#21-0590) on the 1T0.41-MoS2 surface. Raman spectroscopy showed E2g1 and A1g vibrational bands at 376.2 and 402.9 cm-1 peaks typical for 2Hphase-MoS2.36 J1, J2 and J3 vibrations at 147.3, 235.4 and 335.2 cm-1 are characteristic for 1Tphase-MoS237 (Supplementary Fig. 4b). These results prove that the 1T phase of MoS2 is formed by the hydrothermal reaction induced by organic acids (e.g., citric acid).35 1T0.72-MoS2@NiS2 or 1T0.81-MoS2@Ni2P demonstrated three characteristic peaks of 1Tphase-MoS2 and the two characteristic peaks (E2g1 and A1g) of 2Hphase-MoS2. Additionally, they showed a vibrational peak (437.3 cm-1) of Ni-S32 or three vibrational peaks (216.2 cm-1, 249.7 cm-1, and 269.5 cm-1) of Ni-P.31 More importantly, the E2g1 and A1g vibrations of 1T0.72-MoS2@NiS2 at 382.2 and 408.1 cm-1 were red-shifted by 6.0 and 5.2 cm-1, respectively (Supplementary Fig. 4b). Thus, NiS2 nanoparticles are between the 1T0.41-MoS2 layers. Similarly, the E2g1 and A1g peaks for the 1T0.81-MoS2@Ni2P catalyst slightly red-shifted by 7.3 and 3.0 cm-1, respectively. These results confirm that rich multi-heterojunction interface edges active sites catalysts were successfully synthesized.
Electronic structure characterizations of 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P catalysts. To further identify the surface electronic structure of multi-heterogeneous interface catalysts, we applied the high-resolution transmission electron microscopy (HRTEM) to assess the morphology and crystal structures of 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 catalysts. Supplementary Fig. 5a, b shows the typical low-magnification image of the 1T0.72-MoS2@NiS2 on the Cu grid, which confirms the flower-like nanosphere morphologies of 1T0.72-MoS2@NiS2. TEM and corresponding elemental distribution map obtained for the 1T0.72-MoS2@NiS2 sample demonstrated uniformly distributed Mo, Ni, and S (Supplementary Fig. 5c-c4). As revealed by the HRTEM image (Fig. 2a-c and Supplementary Fig. 5e, f), NiS2 nanoparticles are decorated on MoS2 nano-sheets edge. The HRTEM of the 1T0.72-MoS2@NiS2 catalyst clearly shows the crystal lattice of 0.25 nm, referring to the NiS2 (210). Interestingly, Fig. 2a shows the HRTEM image of 1T0.72-MoS2@NiS2 flower-like nano-sheets, in which there demonstrates the lattice fringes perpendicularly to the electron beam direction circled by blood-color, justifying the S defect (Fig. 2c). The trigonal lattice in the yellow circle (Fig. 2c) implies the presence of 1T phase MoS2, while the hexagonal lattice in the blue circle (Fig. 2c) suggests the presence of 2H phase MoS2. The above-described results further confirm the successful preparation of the 1T0.72-MoS2@NiS2 multi-heterojunction interface catalyst. The anion is changed to be P to produce 1T0.81-MoS2@Ni2P multi-heterojunction interface catalyst by phosphorus vapor thermal treatment. Supplementary Fig. 6a, b displays the morphologies of 1T0.81-MoS2@Ni2P catalyst, overlapping nanosheets with many embedded particles can be clearly identified. There is an obvious alternation of 1T and 2H phases, and a large number of defects (Fig. 2f). As shown in Supplementary Fig. 6c, there are the distributions of Mo, Ni, S, and P over the whole 1T0.81-MoS2@Ni2P catalyst, verifying that Ni2P nanoparticles are encapsulated by MoS2 edges. The interplanar spacings, equal to 0.62 and 0.22 nm, match (002) and (111) interplanar distances of MoS2 and Ni2P, respectively (Fig. 2d, and e). Similarly, Fig. 2e, f displays two amplified HRTEM images truncated from Fig. 2d. Fig. 2f demonstrates some hexagonal and trigonal lattice areas of semi-conductor 2Hphase- and metallic 1Tphase-MoS2, respectively. The HRTEM results further confirm the successful preparation of the 1T0.81-MoS2@Ni2P multi-heterojunction interface catalyst.
Next, we performed XPS measurement to assess the elemental valence states of all the as-synthesized samples (Fig. 2g-i and Supplementary Fig. 7). Full XPS spectrum for 1T0.72-MoS2@NiS2 electrode (Supplementary Fig. 7a) showed that atomic ratios of Mo, S and Ni were equal to 13.96%, 36.96% and 4.39 %, respectively, and close to that measured by HRTEM elemental mapping (∼14.30 %, 35.87 %, and 4.76 %). Mo 3d spectra obtained for the multi-heterojunction interfaces of the 1T0.72-MoS2@NiS2, 1T0.81-MoS2@Ni2P, 1T0.41-MoS2 and 2Hphase-MoS2 electrodes showed Mo 3d3/2 and 3d5/2 peaks of the 2H phase at 233.2 and 229.9 eV, and of the 1T phase at 232.4 and 229.1 eV in 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P, respectively. These results suggest Mo4+ presence (Fig. 2g). Interestingly, the 1T phase contents in the 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 samples (equal to 81% and 72%, respectively) were higher than the 41% value observed for the 1T0.41-MoS2 phase. Thus, phosphorus or sulfur implantation further facilitates the phase transformation of 1Tphase-MoS2.22,27 The S 2p spectra also displayed similar results (Fig. 2h). All as-prepared electrodes also have two new doublet peaks of 161.9 eV and 163.2 eV, indicating that these belong to the characteristic peaks of 1Tphase-MoS2, further confirming 2H → 1T phase transformation.22 Ni 2p spectrum showed peaks at 855.4 and 872.9 eV corresponding to Ni 2p1/2 and 2p3/2, respectively (Fig. 2i) and two satellite peaks, suggesting the existence of Ni2+ 32 in the 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P samples. Both peaks were slightly shifted (by 0.5 eV) towards higher binding energies, suggesting that there is an interaction between NiS2 (or Ni2P) and MoS2 via as-formed hetero-structures. For the 1T0.81-MoS2@Ni2P catalyst, the XPS survey spectrum displays that the atomic ratio of Mo, S, P and Ni are 19.29%, 32.7%, 8.23%, 8.23% (Supplementary Fig. 7a). Also, as shown in Supplementary Fig. 7b, there are two doublets of P 2p peaks (129.4 eV, 130.2 eV), which further confirms the formation of Ni2P. These results indicate the successful synthesis of multi-heterogeneous-interface catalysts.
Electrocatalytic HER performances in alkaline and acidic media. 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P electrodes exhibited attractive multi-heterogeneous interface edges, plentiful active sites and abundant mass transfer and gas release channels and are expected to be used as very effective and stable catalysts for H2 production. First, we analyzed HER activities (in 1.0 M KOH) of the electrodes containing these electrodes. The overpotentials of the electrodes containing 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P were 128 and 168 mV at 10 mA/cm2, respectively (see linear sweep voltammetry (LSV) results in Fig. 3a). These values are close to the Pt/C electrode potential equal to 84 mV. For the reaction kinetics analysis, we adjusted the Tafel slopes of these electrodes using the Tafel equation 38,39 and obtained the smallest slopes equal to 68 and 79 mV/dec for the electrodes containing 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P, respectively (Supplementary Fig. 8a). These values are even closer to the corresponding slope of the Pt/C electrode (equal to 56 mV/dec). Thus, electrodes containing 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P as active materials exhibit the fastest HER processes and better reactivity, which is attributed to the multi-heterogeneous interface effect, a large number of defects, and a higher proportion of 1Tphase-MoS2. Next, we evaluated the long-term cycling stability of the as-prepared electrodes using the chronopotentiometry technique at 10 and 30 mA/cm2, respectively. The electrodes containing 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P were very robust and exhibited negligible damping after 16 h measurement (Supplementary Fig. 8b), and the LSV curves measured before and after the long-term tests are almost the same (Supplementary Fig. 8c), demonstrating excellent long-term stability. Supplementary Fig. 8d lists the overpotential values for the 20.0 wt % Pt/C, 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P electrodes in 1.0 M KOH at various current densities. 1T0.72-MoS2@NiS2 electrodes exhibited lower overpotential. Generally, low overpotential and Tafel slope values demonstrated the superior HER catalytic activities, which was the case for our 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P electrodes. 1T0.72-MoS2@NiS2 containing electrode has such excellent HER activity comparable to those of as-reported Mo-based materials (Fig. 3b) and composites and various representative catalysts25,26,28,31-34 (Supplementary Table 1). Thus, 1T0.72-MoS2@NiS2 electrode is a catalyst with the best HER activity in alkaline solutions.
To obtain the electrochemically active area (ECSA) of the 1T0.72-MoS2@NiS2 and 1T0.81-MoS2@Ni2P electrodes, the double-layer capacitance (Cdl) was calculated because the two values are proportional to each other. Therefore, we tested their cyclic voltammetry (CV) by continuously increasing scanning speed (Supplementary Fig. 9a-c) in order to obtain the CV curve of the electrode materials in the non-Faraday region (-0.2-0.4 V). Then, as shown in Supplementary Fig. 9d, the Cdl was calculated from the plot slope (slope = 2Cdl) between current-density difference (∆j) (0.15 V vs. RHE) and scan rate. The 1T0.72-MoS2@NiS2 electrodes possessed the highest Cdl value (Cdl = 359.7 mF/cm2), suggesting a multi-heterogeneous interface could be effectively enhanced conductivity and exposed more active sites of as-prepared electrodes. We recorded the electrochemical impedance spectra (EIS). The corresponding Nyquist (Supplementary Fig. 10) of the 1T0.72-MoS2@NiS2 electrode showed the lowest value for the charge transfer resistance (Rct). Thus, it possessed very favorable charge transfer kinetics.
Next, we also studied the HER performance of all the as-prepared electrodes in 0.5 M H2SO4 (Fig. 3c). The HER catalytic performance of the electrodes containing 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 were significantly improved their HER activities according to the LSV data: their overpotential values at 10 mA/cm2 were as low as 38.9 and 98.5 mV, respectively, which is lower than the values for the electrodes containing 1T0.41-MoS2@Ni(OH)2 (236 mV), 1T0.41-MoS2 (389 mV), 1Tphase-MoS2 (392 mV), and 2Hphase-MoS2 (354 mV). The Tafel slopes for the electrodes containing 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 were 41 and 42 mV/dec (Supplementary Fig. 11a). These values were lower than the values obtained for electrodes containing 1T0.41-MoS2 (65 mV/dec), 1Tphase-MoS2 (76 mV/dec), and 2Hphase-MoS2 (71 mV/dec) and were close to the electrode based on 20 wt% Pt/C (38 mV/dec). It is probably because, in the acidic environment, the H2 desorption is the limiting step because H+ are abundant. The 1T0.81-MoS2@Ni2P electrode had a weaker adsorption capacity toward Hads so it exhibits a better catalytic effect than 2Hphase-MoS2.40 Meanwhile, compared to the other electrodes, 1T0.81-MoS2@Ni2P also has a higher ECSA because it has a larger Cdl (Cdl = 106.15 mF/cm2, Supplementary Fig. 12) and, as a result, more catalytical sites, which significantly contributed to the overall activity. Furthermore, 1T0.81-MoS2@Ni2P also possesses a much smaller Rct, in contrast to other electrodes at 50 mV overpotential vs. RHE (Supplementary Fig. 13), revealing satisfied electron transport and good catalytic kinetics, which leads to high activity and low Tafel slope. Supplementary Fig. 11b shows that at 10 and 45 mA/cm2, 1T0.81-MoS2@Ni2P and 1T0.72-MoS2@NiS2 electrodes were very durable and possesses negligible damping after 16 h measurement, which displays excellent long-term stability. In addition, even after 16 h of a chronoamperometric stability test of the electrodes, the current density remains above 95% (Supplementary Fig. 11c), and there is only a slight deviation for the LSV recorded after the stability test, indicating that as-prepared electrodes have very good stability in an acidic environment. As to 20.0 wt % Pt/C, 1T0.72-MoS2@NiS2, and 1T0.81-MoS2@Ni2P electrodes in 0.5 M H2SO4, Supplementary Fig. 11d shows overpotentials vs. various current densities. 1T0.81-MoS2@Ni2P exhibits lower overpotential. We also compared the overpotentials (at 10 mA/cm2 in acidic medium) and Tafel slopes with previously excellent Mo-based electrocatalysts27,31,41-44 (Fig. 3d and Supplementary Table 2). Catalytic HER performance of 1T0.81-MoS2@Ni2P is also superior. Based on the above results, 1T0.81-MoS2@Ni2P multi-heterogeneous interface catalyst shows the remarkable intrinsic HER activities in acidic medium mainly attributed to multi-heterointerface interface edges active sites.
Theoretical calculation and mechanisms analysis of the surface electronic structure and HER activation energy for the as-prepared electrocatalysts. To explain the distinguished synergistic effect of 1T0.72-MoS2@NiS2 (or 1T0.81-MoS2@Ni2P) multi-heterogeneous interface catalysts, Density functional theory (DFT) calculations were also performed. Model building and computational parameters can be seen in the “Methods” section. Firstly, the interfacial electron interaction was investigated. The charge difference images (Fig. 4a, b and Supplementary Fig. 14) reveal the charge transfer from 1T0.41-MoS2 to the Ni2S or/and Ni2P interface, and the introduction of 1T phase is more conducive to charge transfer from MoS2 to NiS2 or Ni2P interface, which significantly increase the interface electron concentration and thus improve its activity. To better understand the surface electronic structure reconfiguration of MoS2 through coordinated phase transition and interface regulation in theory, the band structure and density of states (DOS) of bare NiS2, Ni2P, 2Hphase-MoS2, 1Tphase-MoS2, 2Hphase-MoS2@NiS2, 2Hphase-MoS2@Ni2P, 1Tphase-MoS2@NiS2, and 1Tphase-MoS2@Ni2P (Fig. 4c-e and Supplementary Fig. 15) obtained using the hybrid DFT-HSE06 exchange–correlation functional, which is presented in the Supplementary Information. The calculation results show that the bare NiS2 exhibits typical semiconductor characteristics (Fig. 4c), with a narrow bandgap equal to 0.68 eV (Supplementary Fig. 15a). The band structure of 1Tphase-MoS2 (Fig. 4d) and 1Tphase-MoS2@NiS2 (Fig. 4e) exhibited a certain zero bandgap, indicating a complete transition from the semiconductor phase (0.91 eV, Supplementary Fig. 18b) to the metallic phase (0 eV) with improved conductivities.27 Notably, the intensity of PDOS of 1Tphase-MoS2@NiS2 was higher than that of 1Tphase-MoS2 and NiS2 at the Fermi level (Supplementary Fig. 15). Thus, the electron mobility of the 1Tphase-MoS2@NiS2 catalysts was more favorable for the efficient charge transfer, which agrees consistent with the EIS test results.45 Moreover, the PDOS results imply that the NiS2 interface hybrid generates some new interface electronic states in 1Tphase-MoS2 (Supplementary Fig. 15c), which was very likely because of hybridization of the d-orbital of Mo and an empty d-orbital of Ni. Thus, higher HER activity of 1Tphase-MoS2@NiS2 in comparison to 1Tphase-MoS2 agrees with the Fermi level DOS (Fig. 4d. e). Thus, the actual electrochemical performance would show even faster conductivity and charge transfer kinetics.
The HER effect was mainly studied by a three-state diagram containing original H+ and intermediate H* states and 1/2 H2 formation 30,46. However, the energy of the intermediate state H*(ΔGH*) is a critical indicator of the ability of hydrogen evolution.47 To reveal further the relationship of HER activity of catalysts with phase structure and heterojunction-interface, we used DFT to calculate the ΔGH* for HER on 2Hphase-MoS2, 1Tphase-MoS2, 2Hphase-MoS2@NiS2, 2Hphase-MoS2@Ni2P, 1Tphase-MoS2@NiS2, and 1Tphase-MoS2@Ni2P catalysts with partially multi-heterojunction interface modification. Fig. 4f displays the calculated free energy diagram on the most stable energy of the 2H phase, 1T phase, 2Hphase-MoS2@Ni2P, 2Hphase-MoS2@NiS2, 1Tphase-MoS2@NiS2 and 1Tphase-MoS2@Ni2P catalysts (Supplementary Fig. 15). For 2Hphase-MoS2, the ΔGH* is very positive (1.18 eV), indicating that there is a strong interaction between H* and 2Hphase-MoS2, showing poor HER reaction kinetics. The introduction of the 1T-phase into MoS2 can obviously increase the value of ΔGH* to – 0.36 eV, implying promoted HER activity compared to 2Hphase-MoS2. However, constructing multi-heterointerface interface edges active sites with NiS2 (1Tphase-MoS2@NiS2) would lead to the ΔGH* value equal to almost 0 eV (- 0.17 eV). For comparison, ΔGH* for the 2Hphase-MoS2@NiS2 was equal to 0.74 eV. The reason is that H* adsorbed on the surface of 2Hphase-MoS2 bounds to Mo atoms, and strong Mo-H strength and poor conductivity. However, H* can be absorbed not only by the 1Tphase-MoS2@NiS2 surface. Ni atoms possess empty d orbitals capable of binding H atoms, thereby weakening the Mo-H strength. More importantly, the introduction of the 1T-phase not only increases its electrical conductivity but also creates abundant active sites at the multi-heterojunction interface edges, which synergistically promote HER activity (Fig. 4g). Thus, our work demonstrates a novel and efficient design to create multi-heterogeneous interfacial electrocatalysts without noble metal materials and with excellent HER activity.