Superior Electrocatalytic Activity of MoS2-Graphene as Superlattice

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.


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 MoS 2 could be an efficient electrocatalyst for the oxygen evolution reaction (OER) replacing IrO 2 and Pt [5,6]. MoS 2 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 MoS 2 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 MoS 2 to the external circuit. For this reason, there has been considerable interest in developing different procedures for the preparation of MoS 2 /graphene heterojunctions to be used in electrolysis [13][14][15] In this context, a procedure consisting in the one-step formation of MoS 2 strongly interacting with graphene has been recently reported. The process is based on the pyrolysis of (NH 4 ) 2 MoS 4 and chitosan as precursors of MoS 2 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 Nanomaterials 2020, 10 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 MoS 2 /graphene heterojunction in superlattice configuration, particularly when considering the well-known activity of MoS 2 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 MoS 2 /graphene heterojunction in superlattice configuration. It will be shown that this fl-MoS 2 /graphene material (fl meaning few layers) exhibits much lower onset potential for HER and OER than an analogous electrode prepared with random MoS 2 /G heterojunction, illustrating the advantages of lattice matching to improve the electrochemical performance.

Methods-Exfoliation MoS 2 and Preparation of MoS 2 /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 cm 2 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 MoS 2 (Aldrich, St. Louis, Missouri, USA) at a concentration of 1.5 mg/mL and commercial graphene with an area of 700 m 2 (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 MoS 2 /G heterojunction following the same procedure indicated above.

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 cm 2 ). Voltammetric experiments were carried out with a potentiostatic device CH 920c (Cambria Scientific, Llwynhendy, Llanelli, Wales, UK), using H 2 SO 4 1.0 M saturated with air as an electrolyte. The working electrodes were the MoS 2 /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).

Results
Upper and bottom layers of S atoms sandwiching and internal Mo IV layer constitute the crystal structure of MoS 2 . 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 MoS 2 -G Nanomaterials 2020, 10, 839 3 of 9 heterojunctions when there is a coincidence of the two hexagonal arrangements. Figure 1 illustrates an ideal model of MoS 2 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 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.  There are precedents in the literature showing the possibility to obtain MoS 2 /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 MoS 2 /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 MoS 2 crystals that are subsequently exfoliated in viscous halogenated solvent while using polystyrene as promoter. After exfoliation, the residual bulk MoS 2 crystals can be removed from the single and few layers MoS 2 sheets by decanting the supernatant. Figure 2a presents AFM (Atomic Force Microscope) images of films of polystyrene containing MoS 2 . 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 MoS 2 in these films was not apparent from AFM at this moment due to the thickness. In a subsequent step, these films of polystyrene embedding MoS 2 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. 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 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 MoS 2 /G film shows also the presence of MoS 2 particles on this continuous graphene film. Measurement by AFM of a statistically relevant number of MoS 2 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 MoS 2 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 MoS 2 is 0.615 nm, this average thickness corresponds to less than four MoS 2 layers. Nanomaterials 2020, 10, x FOR PEER REVIEW 4 of 9 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 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 MoS 2 /G film after detachment from the quartz support.
The composition of MoS 2 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 MoS 2 particles of size from 20-140 nm. High resolution TEM images show the expected hexagonal arrangement characteristic of graphene and MoS 2 (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 MoS 2 . The coincidence of the electron diffraction patterns between two 2D materials (Figure 3d) is taken as the best evidence for the superlattice configuration. Nanomaterials 2020, 10, x FOR PEER REVIEW 5 of 9 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]. It is similarly proposed here that the existing MoS 2 sheets resulting from bulk MoS 2 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 MoS 2 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 MoS 2 , in the heterojunction. Figure S3 in the supporting information shows the XRD pattern recorded for MoS 2 /G. This XRD shows a sharp peak at 2θ 14.60 corresponding to the 0.0.2 facet of MoS 2 . Other peaks of much lesser intensity correspond to other MoS 2 planes. In addition to the peaks of MoS 2 , a broad diffraction band at 24 • due to few layer defective graphene is also recorded [12]. Figure 4 shows the corresponding Raman for the MoS 2 /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 MoS 2 are also seen. No changes in the position of the MoS 2 or G bands is observed, indicating that the superlattice configuration does not alter the lattice constants of MoS 2 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 MoS 2 (25 cm −1 ). In the present case, the difference between the A 1g and E 2g vibration modes of MoS 2 was 24 cm −1 , which agrees with the case of few layers MoS 2 . This configuration of MoS 2 as few layers platelets is also in accordance with the previously commented AFM measurements of MoS 2 particles that were supported on graphene.
Nanomaterials 2020, 10, x FOR PEER REVIEW 6 of 9 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 A1g and E2g 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

Catalytic Activity
The main purpose of this study is to determine the possibility of preparing large area films of superlattice MoS 2 /graphene (s-MoS 2 /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 MoS 2 /graphene (r-MoS 2 /G, r-meaning random) heterojunction, as stated in the introduction. Aimed at this purpose, an additional sample of r-MoS 2 /G was prepared by introducing commercial preformed graphene during the process of exfoliation of bulk commercially available MoS 2 (Aldrich). Supporting information provides characterization data of r-MoS 2 /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 MoS 2 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 MoS 2 /G electrode and a reference electrode. The quartz plates used for preparation of electrodes were previously modified by subliming carbon before MoS 2 /G film preparation, thus improving the electrical contacts, in order to increase the electrical conductivity of MoS 2 /G electrodes. Prior controls with sublimed carbon quartz plates lacking MoS 2 /G electrocatalyst did not exhibit any peak in cyclic voltammetry. Experiments were carried out for air-saturated, 1.0 M aqueous H 2 SO 4 solution as electrolyte. Figure 5 shows the cyclic voltammetry corresponding to s-MoS 2 /G and r-MoS 2 /G under these conditions n. As it can be seen there, for s-MoS 2 /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-MoS 2 /G electrode exhibits the expected OER process with an onset potential of 0.87 V. For comparison, Table 1 compiles reported electrochemical data [28,29] for Pt/C and MoS 2 electrodes together with the new data herein obtained for s-MoS 2 /G. 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 compiles reported electrochemical data [28,29] for Pt/C and MoS2 electrodes together with the new data herein obtained for s-MoS2/G.  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.

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. The peaks corresponding to hydrogen evolution reaction (HER), oxygen evolution reaction (OER), and oxygen reduction reaction (ORR) have been indicated in the plot, as well as the corresponding onset potentials. Table 1. Electrochemical data for the catalytic HER process of reference electrodes in comparison to s-MoS 2 /G. From linear potential scan voltammograms of material-modified graphene modified glassy carbon (GCE) electrodes in contact with 0.50 M H 2 SO 4 ; potential scan rate 10 mV s −1 . The exchange current density (jo) was calculated from Tafel graphs using the current extrapolation method.

Conclusions
In summary, it has been shown that the exfoliation of bulk MoS 2 crystals using polystyrene as promoter, followed by film casting and pyrolysis is a suitable procedure for obtaining few layers MoS 2 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 MoS 2 /G electrodes, one with lattice matching and another with random configuration in the heterojunction, it has been observed that the electrocatalytic activity of MoS 2 /G improves significantly in the superlattice configuration. This probably reflects the favourable orbital overlapping and electron migration in the MoS 2 /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.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2079-4991/10/5/839/s1, Figure S1: Lateral size (a) and height (b) of MoS 2 nanoparticles present on MoS 2 /G, Figure S2: EDX analysis of S, Mo, C and O for the MoS 2 /G particles shown in the TEM image (left frame); Figure S3: XRD spectrum of MoS 2 /G; Figure S4: Frontal view (a and b) and height profiles (c and d) along the white lines of r-MoS 2 /polystyrene (a and c) and r-MoS 2 /G (b and d); Figure S5: TEM at different magnifications and selected area electron diffraction of r-MoS 2 /G; Figure S6: TEM image (left) and EDS analysis of C, S and Mo for r-MoS 2 /G.

Author Contributions:
The research was performed with the contribution of all the authors. The concept of the study was developed by A.P., and H.G. Sample preparation and characterization was performed by A.R.-P. and Electrochemical measurements were carried out by A.R.-P and A.D.-C. Drafting of the manuscript was performed by H.G. and A.P. All authors have read and agreed to the published version of the manuscript.