Superior Electrocatalytic Activity of MoS₂-Graphene as Superlattice

Abstract

Evidence by selected area diffraction patterns shows the successful preparation of large area (cm × cm) MoS₂/graphene heterojunctions in coincidence of the MoS₂ and graphene hexagons (superlattice). The electrodes of MoS₂/graphene in superlattice configuration show improved catalytic activity for H₂ and O₂ evolution with smaller overpotential of +0.34 V for the overall water splitting when compared with analogous MoS₂/graphene heterojunction with random stacking.

1. Introduction

There is considerable interest in developing noble metal-free electrodes that could efficiently perform water splitting due to the current change from fossil fuels to renewable electricity [1,2,3,4]. In this context, it has been reported that MoS₂ could be an efficient electrocatalyst for the oxygen evolution reaction (OER) replacing IrO₂ and Pt [5,6]. MoS₂ also shows electrocatalytic activity for hydrogen evolution reaction (HER) [7,8,9] and, therefore, it could be an ideal material to perform both redox processes of overall water splitting.

Regarding the use of MoS₂ as electrocatalyst, it has been shown that the assembly of this chalcogenide with graphene improves its performance as electrodes [10,11,12]. Graphene introduces electrical conductivity favouring electron transfer from the catalytic site on MoS₂ to the external circuit. For this reason, there has been considerable interest in developing different procedures for the preparation of MoS₂/graphene heterojunctions to be used in electrolysis [13,14,15]

In this context, a procedure consisting in the one-step formation of MoS₂ strongly interacting with graphene has been recently reported. The process is based on the pyrolysis of (NH₄)₂MoS₄ and chitosan as precursors of MoS₂ and N-doped defective graphene, respectively [16]. One of the advantages of this procedure was that chitosan allows for the formation of nanometric films on arbitrary, non-conductive substrates, such as quartz or ceramics, which, after transformation on (N)G, becomes electrically conductive and can be used as electrode [17,18]. Also related to the present study is a recent report from us on the formation of large area boron nitride/graphene heterojunctions with superlattice configuration [19].

Since both boron nitride and graphene are two-dimensional (2D) materials with very similar lattice parameters, the term superlattice refers to the coincidence of the hexagons of boron nitride overlapping those of graphene. Fundamental studies have shown that superlattice configuration of 2D materials heterojunctions can give an assembly with tunable electrical conductivity in contrast to the random heterojunctions of the two materials [20,21]. This modulation of the electron mobility through the graphene layer by the underlying boron nitride could be reflected in some unique properties of the superlattice heterojunction, particularly those that are related to electrochemistry and electrocatalysis [22,23,24]. In view of these precedents, it would be of interest to determine the electrocatalytic behaviour of MoS₂/graphene heterojunction in superlattice configuration, particularly when considering the well-known activity of MoS₂ for important reactions, such as HER, OER, and oxygen reduction reaction (ORR) processes [5].

This manuscript reports a novel procedure for the preparation of large area (cm × cm) films of MoS₂/graphene heterojunction in superlattice configuration. It will be shown that this fl-MoS₂/graphene material (fl meaning few layers) exhibits much lower onset potential for HER and OER than an analogous electrode prepared with random MoS₂/G heterojunction, illustrating the advantages of lattice matching to improve the electrochemical performance.

2. Materials and Methods

2.1. Methods-Exfoliation MoS₂ and Preparation of MoS₂/fl-G

Molybdenum sulfide exfoliation was carried out while using polystyrene as an exfoliating agent (ALDRICH). The polystyrene was dissolved in dichloromethane at a concentration of 3 mg/mL and powdered Molybdenum sulfide was mixed at a concentration of 1.5 mg/mL. The sample was sonicated while using a Sonic tip (FisherbrandTM Model 705 at 50% 700W for 5 ho per sample with a on/off sequence consisting in 1 s off and 1 s pulse and an ice bath was used to prevent solvent evaporation). After sonication, the dispersion was centrifuged at 1500 rpm for 45 min (Hettich Zentrifugen EBA 21(Hettich, Westphalia, Germany)).

The supernatant was preconcentrated and deposited on quartz films by rotation coating on 2 × 2 cm² quartz substrate to prepare the films (APT-POLOS spin-coater(SPS-Europe B.V., Putten, The Netherlands): 4000 rpm, 30 s). The pyrolysis of polystyrene treatment was performed using an electric oven and using the following heating program: heating at 5 °C/min at 900 °C for 2 h.

For the preparation of films with random configuration, commercial MoS₂ (Aldrich, St. Louis, Missouri, USA) at a concentration of 1.5 mg/mL and commercial graphene with an area of 700 m² (STREM CHEMICALS, Newburyport, MA, USA) at a concentration of 7.5 mg/mL were added to a 30 mg/mL polystyrene solution (Aldrich, St. Louis, MO, USA) in dichloromethane and the suspension submitted to ultrasounds using a Sonic tip at 700 W for 5 h. After sonication, the dispersion was centrifuged at 1500 rpm for 46 min. and the supernatant used to prepare films of randomly configured MoS₂/G heterojunction following the same procedure indicated above.

2.2. Methods-Catalytic Measurements

Electrochemical measurements were made in a conventional three electrode electrochemical cell while using a Pt disc pseudo-reference electrode and a Pt wire auxiliary electrode that accompanies the graphene modified glassy carbon (GCE) working electrode (BAS, MF 2012, area geometric 0.071 cm²). Voltammetric experiments were carried out with a potentiostatic device CH 920c (Cambria Scientific, Llwynhendy, Llanelli, Wales, UK), using H₂SO₄ 1.0 M saturated with air as an electrolyte. The working electrodes were the MoS₂/fl-G films that were deposited on graphite. The graphite was deposited on the quartz films using the automatic carbon coater (sputter coater BALTEC_SCD005 (BAL-TEC AG, Schalksmühle, Germany) with carbon evaporation supplied BALTEC-CEA035).

3. Results

Upper and bottom layers of S atoms sandwiching and internal Mo IV layer constitute the crystal structure of MoS₂. The geometrical arrangement of the three layers is such that top views define hexagons with alternating edges S and Mo atoms that are located at different planes. These hexagons are similar in size to those of G and, therefore, a superlattice configuration can be possible for MoS₂-G heterojunctions when there is a coincidence of the two hexagonal arrangements. Figure 1 illustrates an ideal model of MoS₂ structure and graphene layer assembled in the superlattice configuration.

There are precedents in the literature showing the possibility to obtain MoS₂/Graphene heterojunction with lattice matching geometry [25,26]. It was the leading hypothesis of the present study that the adaptation of this method could also serve to obtain MoS₂/G superlattice since it has been previously reported the preparation of large area films of BN/G heterojunctions as superlattice [19].

Scheme 1 illustrates the procedure followed in the present study. As it can be seen there, the process starts with commercial MoS₂ crystals that are subsequently exfoliated in viscous halogenated solvent while using polystyrene as promoter. After exfoliation, the residual bulk MoS₂ crystals can be removed from the single and few layers MoS₂ sheets by decanting the supernatant. Figure 2a presents AFM (Atomic Force Microscope) images of films of polystyrene containing MoS₂. As expected in view of the plastic characteristic of polystyrene, these films were smooth and have a thickness about 200 nm (see profile in panel C of Figure 2). The presence of MoS₂ in these films was not apparent from AFM at this moment due to the thickness. In a subsequent step, these films of polystyrene embedding MoS₂ were submitted to pyrolysis in the absence of oxygen at 900 °C. These conditions have been previously reported to transform polystyrene into thin films of few layers defective of graphene [27]. One of the main advantages of the present procedure is its reproducibility and the possibility to prepare large surface areas.

A considerable shrinkage in the film thickness accompanies this transformation, as it can be seen in Figure 2b, where films of about 4 nm thickness corresponding to graphene can be seen (Figure 2d). The frontal image of MoS₂/G film shows also the presence of MoS₂ particles on this continuous graphene film. Measurement by AFM of a statistically relevant number of MoS₂ particles on graphene indicate that the lateral size distribution ranges from 20 to 140 nm with an average about 80 nm and the thickness of the MoS₂ is from 2 to 8 nm, with an average about 2. Figure S1 in the supporting information illustrates the corresponding histograms of the lateral dimension and height. When considering that the interlayer distance of MoS₂ is 0.615 nm, this average thickness corresponds to less than four MoS₂ layers.

The SEM images of these films show a smooth surface with low apparent roughness and without cracks or pinholes. Figure 3a shows one of these representative SEM images. The TEM of these films were obtained by scratching a bit of these films. As an example, Figure 3b shows a representative TEM image of MoS₂/G film after detachment from the quartz support.

The composition of MoS₂ was confirmed by EDS elemental mapping, determining an overlap of Mo and S in 1:2 atomic ratio. Figure S2 in the supporting information provides a summary of the Mo, S, O (from graphene defects), and C elemental distribution for a representative TEM image. These TEM images show a graphene layer larger than hundreds of nm having dark spots corresponding to smaller MoS₂ particles of size from 20–140 nm. High resolution TEM images show the expected hexagonal arrangement characteristic of graphene and MoS₂ (Figure 3c). Importantly, selected area electron diffraction at every point of the TEM image shows bright hexagonal spots corresponding to graphene. Coincident with those spots, there were also other points due to MoS₂. The coincidence of the electron diffraction patterns between two 2D materials (Figure 3d) is taken as the best evidence for the superlattice configuration.

It is similarly proposed here that the existing MoS₂ sheets resulting from bulk MoS₂ exfoliation are templating during the pyrolysis the formation of the nascent graphene layers in such a way that the growing graphene is replicating the existing MoS₂ sheet resulting in lattice matching throughout the heterojunction, as in the precedent reporting the BN/G superlattice heterojunction.

XRD and Raman spectroscopy convincingly evidence the presence of both components, graphene and MoS₂, in the heterojunction. Figure S3 in the supporting information shows the XRD pattern recorded for MoS₂/G. This XRD shows a sharp peak at 2θ 14.60 corresponding to the 0.0.2 facet of MoS₂. Other peaks of much lesser intensity correspond to other MoS₂ planes. In addition to the peaks of MoS₂, a broad diffraction band at 24° due to few layer defective graphene is also recorded [12].

Figure 4 shows the corresponding Raman for the MoS₂/G heterojunction. At high wavenumbers, the expected broad peaks at 2840, 1540, and 1355 cm⁻¹ corresponding to the 2D, G, and D peaks characteristic of defective graphenes are recorded. Two very sharp lines at 383 and 408 cm⁻¹ that are characteristic of MoS₂ are also seen. No changes in the position of the MoS₂ or G bands is observed, indicating that the superlattice configuration does not alter the lattice constants of MoS₂ or G, in agreement with the lattice match. The difference in wavenumber of these two peaks indicates the single layer (18 cm⁻¹) or few layers structure of MoS₂ (25 cm⁻¹). In the present case, the difference between the A_1g and E_2g vibration modes of MoS₂ was 24 cm⁻¹, which agrees with the case of few layers MoS₂. This configuration of MoS₂ as few layers platelets is also in accordance with the previously commented AFM measurements of MoS₂ particles that were supported on graphene.

Catalytic Activity

The main purpose of this study is to determine the possibility of preparing large area films of superlattice MoS₂/graphene (s-MoS₂/G, s- meaning superlattice and corresponding to the material prepared, according to Scheme 1) heterojunctions that are suitable for electrocatalytic characterization and compare their electrochemical properties with those of a random MoS₂/graphene (r-MoS₂/G, r-meaning random) heterojunction, as stated in the introduction. Aimed at this purpose, an additional sample of r-MoS₂/G was prepared by introducing commercial preformed graphene during the process of exfoliation of bulk commercially available MoS₂ (Aldrich). Supporting information provides characterization data of r-MoS₂/G sample used for the comparison, including a set of TEM images and EDS analysis with the corresponding elemental mapping. (Figure S6) Importantly, selected area diffraction patterns show the spots corresponding to hexagonal patterns of MoS₂ and graphene not overlapped and clearly distinguishable for each 2D material, as expected for samples lacking the superlattice configuration (Figure S5, frame d).

Electrochemical measurements were performed with a conventional three-electrode single cell with Pt as auxiliary electrode accompanying the MoS₂/G electrode and a reference electrode. The quartz plates used for preparation of electrodes were previously modified by subliming carbon before MoS₂/G film preparation, thus improving the electrical contacts, in order to increase the electrical conductivity of MoS₂/G electrodes. Prior controls with sublimed carbon quartz plates lacking MoS₂/G electrocatalyst did not exhibit any peak in cyclic voltammetry. Experiments were carried out for air-saturated, 1.0 M aqueous H₂SO₄ solution as electrolyte. Figure 5 shows the cyclic voltammetry corresponding to s-MoS₂/G and r-MoS₂/G under these conditions n. As it can be seen there, for s-MoS₂/G a cathodic peak at −0.8 V vs. Ag/AgCl corresponding to ORR is recorded before a step peak corresponding to HER with onset being determined by extrapolation of the current density plot at −0.70 V. In the anodic region, the s-MoS₂/G electrode exhibits the expected OER process with an onset potential of 0.87 V. For comparison,

Table 1. Electrochemical data for the catalytic HER process of reference electrodes in comparison to s-MoS₂/G. From linear potential scan voltammograms of material-modified graphene modified glassy carbon (GCE) electrodes in contact with 0.50 M H₂SO₄; potential scan rate 10 mV s⁻¹. The exchange current density (jo) was calculated from Tafel graphs using the current extrapolation method.

Material	Eonset (V vs. RHE)	Tafel Slope (mV per Decade)	Jo (mA cm−2)
Pt/C	0.0	30	7.1 × 10−1
MoS2	−0.24	101	9.1 × 10−4
s-MoS2/G	−0.60	54	7.9 × 10−3

compiles reported electrochemical data [28,29] for Pt/C and MoS₂ electrodes together with the new data herein obtained for s-MoS₂/G.

4. Conclusions

In summary, it has been shown that the exfoliation of bulk MoS₂ crystals using polystyrene as promoter, followed by film casting and pyrolysis is a suitable procedure for obtaining few layers MoS₂ deposited on few layers defective graphene in superlattice configuration. The procedure allows for obtaining large area films that are suitable for electrochemical characterization. By comparing the performance of two MoS₂/G electrodes, one with lattice matching and another with random configuration in the heterojunction, it has been observed that the electrocatalytic activity of MoS₂/G improves significantly in the superlattice configuration. This probably reflects the favourable orbital overlapping and electron migration in the MoS₂/G heterojunction when the two lattices match one on top of the other. The present results show the far-reaching potential of superlattice assembly of 2D heterojunctions for application in electrocatalysis.