Boundary activated hydrogen evolution reaction on monolayer MoS2

Recently, monolayer molybdenum disulphide (MoS2) has emerged as a promising and non–precious electrocatalyst for hydrogen evolution reaction. However, its performance is largely limited by the low density and poor reactivity of active sites within its basal plane. Here, we report that domain boundaries in the basal plane of monolayer MoS2 can greatly enhance its hydrogen evolution reaction performance by serving as active sites. Two types of effective domain boundaries, the 2H-2H domain boundaries and the 2H-1T phase boundaries, were investigated. Superior hydrogen evolution reaction catalytic activity, long-term stability and universality in both acidic and alkaline conditions were achieved based on a multi-hierarchy design of these two types of domain boundaries. We further demonstrate that such superior catalysts are feasible at a large scale by applying this multi-hierarchy design of domain boundaries to wafer-scale monolayer MoS2 films.


Device fabrication procedure
The MoS2 electrochemical devices were fabricated using as-grown type-I and type-III monolayer MoS2 on Si/SiO2 substrates. PMMA was first spin-coated onto the Si/SiO2 surface at 4000 rpm for 60 seconds. The samples were then prebaked at 180°C for 90 seconds. Then, e-beam lithography (EBL) was conducted to pattern gold contact, followed by development, metal deposition, and lift-off processes. Ti/Au (3nm/40nm) was deposited as metal electrodes using an e-beam evaporator at a deposition rate of 0.7 Å/s under high vacuum conditions (10 -7 Torr). Note that type-III MoS2 sample need to perform one more step that the films should be patterned to ribbons using EBL and oxygen plasma etching. Thus each ribbon can be tested separately for subsequent comparison ( Fig S2). Finally, to eliminate the influence of electrolyte and to avoid electrochemical reactions on the metal electrodes, another layer of PMMA was then deposited on the MoS2 device with spin coating. A smaller window that only exposes the basal plane was opened by EBL. The device measurements were performed using electrochemical workstation (Autolab PGSTAT 302N) in a self-developed probe station system (Fig S2), platinum wire and Ag/AgCl electrodes are used as counter and reference electrode, respectively.

Hydrogenation of heterophase MoS2
To hydrogenate the heterophase MoS2, the heterophase MoS2 samples were directly immersed in hydrogen plasma at 0.3 Torr with 30W input powers. The hydrogen plasma was ignited a RF coil with the frequency of 13.56 MHz in our remote-plasma enhanced chemical vapour deposition (rPECVD) system. The duration of plasma treatment was set at 5 min with a continuous Hydrogen flow rate of 100 sccm and vacuum pumping.
The treatment was carried out at room temperature.

Local electronic probe of hydrogen adsorption of heterophase MoS2
Scanning tunneling microscopy/spectroscopy (STM/STS) was used to obtain the hydrogenation of heterophase MoS2. STM experiments were performed ex situ in a combined nc-AFM/STM system (Createc, Germany) at 77 K with base pressure <7×10-9 Pa. Electrochemically etched W-tips were cleaned by alternative annealing and sputtering before the experiments, and further by controlled field-emission and voltagepulse procedures during the scanning. Bias voltage refers to the sample voltage with respect to the tip. All of the STM topographic images were obtained in constant-current mode. The scanning tunnelling spectroscopy (STS) dI/dV spectra were acquired using lock-in detection of the tunnelling current by adding a 5 mVrms modulation at 481 Hz to the sample bias.

Growth of graphene films on pristine monolayer MoS2
Growth of nanographene was carried out in our homemade remote-plasma enhanced chemical vapour deposition (rPECVD) system. [3][4][5] Inductively coupled plasma is generated by a RF coil with the frequency of 13.56 MHz. The precursor is methane (flow rate: 30sccm) that can be dissolved into radicals in plasma for direct growth.
Continuous graphene films were directly deposited on monolayer MoS2 supported by SiO2/Si substrate. The growth temperature is ~450℃ which may not affect the quality of most monolayer MoS2. During deposition, the pressure in chamber was kept constant at ~0.2 Torr.
The grain size of as-grown graphene is much smaller than grown by CVD method, at nm scale. The AFM and SEM images of as-grown graphene on monolayer MoS2 is shown in Fig. S6. As the growth time is 2h, the grain size of graphene varies from a few to tens of nanometers and the thickness of graphene/MoS2 films is around 1.9 nm.

Fabrication process of heterophase-MoS2/graphene/Au catalytic electrode
The heterophase-MoS2/graphene (Gr)/Au electrode were fabricated based on a direct peel-off process (Fig. S5). Firstly, graphene was prepared on as-grown monolayer MoS2 as mentioned above. The graphene is about 2 layers. Then 20nm-thick gold film was deposited on the surface by e-beam evaporation. After deposition, a thermal release tape was sticked to the surface smoothly and then peeled off. Thus, the Au/Gr/MoS2 films were separated from SiO2/Si substrate. This is resulted from the competition of binding energy (f) between interfaces, in the present case, fAu-graphene and fgraphene-MoS2 lager than fMoS2-Substrate. 6 Therefore, we achieved the three layer structures with the MoS2 on the top. Finally, we use Argon plasma treatment to obtain heterophase MoS2 aforementioned.

Raman characterization
The Raman measurement was performed with the excitation laser line of 532 nm using a Horiba Jobin Yvon Lab RAM HR-Evolution Raman system in ambient air environment. The power of the excitation laser line was kept well below 5 mW to avoid damage of MoS2. The Raman scattering was collected by an Olympus 100 × objective (N.A. =0.9).
The Raman spectra of as-grown MoS2 on Silicon substrate, pristine-MoS2/nanographene/Au and Au substrate were shown in Fig. S6. Two typical Raman peaks of MoS2 (A1g at ≈ 404.6 cm -1 and E2g at ≈384.6 cm −1 ) were observed with a frequency difference of ≈20cm −1 for the as-grown MoS2 on silicon substrate, suggesting the monolayer feature of 2H-MoS2. 6 Intriguingly, a blue shift (≈2 cm −1 ) of the E2g peak for MoS2/nanographene/Au (≈382.6 cm −1 ) is observed compared with that of MoS2/Si, and both A1g and E2g peaks broaden together with the decrease of intensity. This is due to perturbations to the perfect 2H-phase lattice, resulting in poorly defined vibrational mode energies, 7 suggesting that the peel-off process changes the vibrational energies of the lattice.

X-ray photoelectron spectroscopy (XPS) characterization
XPS characterization was carried out to confirm the as-grown MoS2/Gr films on SiO2 using Kratos Analytical Axis Ultra system. The peaks around 229.1 and 232.3 eV correspond to the Mo4 +3 d5/2 and Mo 4+ 3d3/2 components in 2H-MoS2, as shown in Fig.   s13a. And Fig. S13b shows the binding energies of 161.9 and 163.1 eV corresponding to the S 2p3/2 and 2p1/2 of MoS2, respectively. 8 The C 1s peak (284.7eV) confirms graphene C-C network (Fig. S13c).

Growth of wafer-scale monolayer 2H-MoS2
The wafer-scale MoS2 growth was performed in a three-temperature-zone chemical vapor deposition (CVD) chamber. S (Alfa Aesar, 99.9%, 4 g) and MoO3 (Alfa Aesar, 99.999%, 50 mg) powders, loaded in two separate inner tubes, were used as sources and placed at zone-I and zone-II, respectively, and 4 in. sapphire wafers were loaded in zone-III as the substrates. During the growth, the two inner tubes were flowed with Ar (gas flow rate 100 sccm) and Ar/O2 (gas flow rate 75/3 sccm) as carrying gases, respectively. During the growth, the temperatures for the S source, MoO3 source, and wafer substrate are 115, 540, and 900 °C, respectively. For a typical growth, the growth duration is ∼40 min, and the pressure in the growth chamber is ∼1Torr.

Computational details
First-principles calculations based on Density Functional Theory (DFT) were carried out by using the Vienna Ab initio Simulation Package (VASP). 9,10 The interactions between valence electrons and ions were treated with the projector-augmented wave (PAW) method. 11 The exchange-correlation interactions were described by generalized gradient approximation (GGA) 12 with the Perdew−Burke−Ernzerhof (PBE) functional. 13 The electron wave functions were expanded in a plane-wave basis set with cutoff energy of 520 eV. The convergence criterion for residual force on each atom during structure relaxation was set to 0.02 eV/Å. and the geometries were relaxed to minimize the total energy of the system until a precision of 10 -4 eV was reached.

Hydrogen adsorption free energy
We study 4 kinds of grain boundaries of the MoS2 monolayer , including 21.8°-tilt composed of 5-7rings , 60°-tilt composed 4-4 rings , 4-8rings and 6-8rings as shown in Fig. 3. The periodic structure model of grain boundaries is complicated to build, so we used different sizes of supercell and k-point mesh to simulate grain boundaries. All structure are fully relaxed and the vacuum spaces in all supercells were larger than 12 Å above the MoS2 plane to avoid any artificial interaction.
We used a supercell to simulate the 1T and 2H phase boundaries. As is well known,   (111) surface, which is in agreement with the result before. Moreover, the supercells for the calculations of ΔGH* on pristine 2H-phase and pristine 1T-phase have the same sizes as for the calculations on phase boundaries. The grain boundaries with different structures also adsorbed one H on one S atom after structural optimization.
In our calculations, the size of the supercell is limited due to the limitations in computing power and crystalline geometry. We had to include two boundaries in one model and the distance between the two boundaries is limited to ~1nm. However, the trend revealed by the present model is general. The 2H-phase is too inert to adsorb H and the 1T-phase is too active to adsorb H atoms moderately as required by HER. At the phase boundaries, due to the atomic relaxations, the S atoms from both phases can become active sites. In this case, the active sites are on the basal plane and not limited to the edges of MoS2 and the density of the active sites can be significantly increased.

Atomic structure with S-vacancies
Each unit cell consists of 16 Mo atoms and 32 S atoms on the surface. Assuming that S-vacancies are only formed on the surface. With the %S-vacancy being defined as (number of S-vacancies)/ (number of total S atoms).