2D Molybdenum Disulphide (2D-MoS2) Modified Electrodes Explored towards the Oxygen Reduction Reaction

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains. Accepted Manuscript Nanoscale

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal's standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

Introduction
The effects of anthropogenic induced climate change are beginning to be realised on both a local and global scale, which has created a demand for the development and implementation of new "clean" methods of energy generation. 1-7 Replacing the typical combustion of fossil fuels (FF) with the utilisation of hydrogen fuel cells in the worlds energy economy could dramatically decrease the production of anthropogenic greenhouse emissions. 8,9 The most widely used fuel cell is the proton exchange membrane (PEM) fuel cell (also known as a polymer electrolyte membrane fuel cell) which is potentially viable in a vast number of applications, from vehicles to combined heat and power units. 10,11 Their implementation is advantageous over typical FF engines due to their zero carbon emissions and ability to undergo long periods of inactivity without detrimental implications on energy output. 10 The reason why they are not currently a viable alternative to FF engines in the majority of applications is a greater cost per unit energy. 10 Resultantly, there is a need to lower the cost of energy production associated with fuel cells. This can be done via lowering the cost of a PEM fuel cell's fuel, typically H 2 , as well as increasing the energy output per unit of fuel utilised. It is therefore essential that research producing alternative cheaper electrocatalysts in order to increase the efficiency of PEM fuel cell energy generation is performed.
The essential reactions which allow a fuel cell to produce a current are the hydrogen oxidation reaction (HOR) and the oxygen reduction reaction (ORR). [12][13][14] The HOR occurs on the anode and typically has a negligible overpotential, whilst the ORR occurs at the cathode and has a large kinetic inhibition given the strong (di)oxygen double bond resulting in a large energy input to initiate the reaction. 12,15 This results in the ORR being the rate determining step in the production of output energy from the initial H 2 fuel source. Taking this into account, by reducing the overpotential at which ORR occurs at the cathode the process will be "more energetically favourable" and it is possible to make a significant increase within fuel cell efficiency. 16,17 Ideally this reaction combines O 2 (typically atmospheric, in the case of PEM fuel cells) with hydrogen in order to produce H 2 O; however, the reaction mechanism is dependent upon the pH of the electrode material and/or electrolyte used. 18 The ORR has proven to be problematic in fuel cells due to membrane degradation and electrode fouling which occurs when the electrode utilised reduces O 2 via a 2 electron pathway (see below) resulting in the unfavourable production of H 2 O 2 . [16][17][18]21 PEM fuel cell degradation via H 2 O 2 induced electrode fouling is the predominate factor in limiting the lifespan this PEM fuel cell, potentially limiting the voltage output by up to 50% as a result of cathode corrosion (causing slow ORR kinetics). 19 The exact mechanism for H 2 O 2 poisoning of the cathode is unclear with direct 20 and indirect 21 attack mechanisms proposed in the literature. The ORR processes in alkaline and acidic media are as follows: 22,23 Acidic media: In order to avoid the production of H 2 O 2 it is essential that an effective electrocatalyst is used so that a direct and more efficient 4-electron pathway is favourable, producing only water as the product. Platinum (Pt) is typically implemented as an electrocatalyst for the ORR as the ORR reaction mechanism occurs via the desirable 4 electron pathway which produces the favourable product H 2 O. 13 However, the use of Pt on a global industrial scale as an electrode material within PEM fuel cells has numerous real world limitations, such as its high cost and the relative global scarcity. 24 Clearly, finding a cheap, non-polluting and widely available alternative to Pt to be used as a catalyst for the ORR, 20, 25 but yet, is capable of matching the ORR onset potential observed when Pt is a clear research goal.
In order to try and achieve this goal, researchers have investigated the electrocatalytic activity of various 2D materials towards the ORR. 23 The electrochemical properties of 2D-MoS 2 are anisotropic in nature, with the basal plane of the 2D-MoS 2 being relatively inert and the exposed edges being reported as the active sites of electron transfer. 1, 29,30 Resultantly, highly defected sheets of 2D-MoS 2 have a greater catalytic activity due to the larger number of exposed edges. 31 Interestingly in its bulk form, MoS 2 exhibits poor electrochemical activity due to a low ratio of exposed edge to basal planes. 3

Experimental section
All chemicals used were of analytical grade and used as received from Sigma-Aldrich without any further purification. All solutions were prepared with deionised water of resistivity not less than 18.2 MΩ cm. The sulfuric acid solutions utilised are of the highest possible grade available from Sigma-Aldrich (99.999%, double distilled for trace metal analysis). The sulfuric acid (0.1 M) solution used to explore the HER was vigorously degassed prior to electrochemical measurements with high purity, oxygen free nitrogen. All ORR measurements were performed in 0.1 M sulfuric acid was oxygenated and subject to rigorous bubbling of 100% medical grade oxygen for one hour, resulting in a 0.9 mM concentration of oxygen, assuming this to be a completely saturated solution at room temperature which is common practice in the literature. 22,23 Where ORR onset potentials are denoted within the manuscript, note that this is defined as the potential at which the current initially deviates from the background current by a value of 25 µA cm -2 , thus signifying the commencement of the faradaic current associated with the ORR redox reaction.
Electrochemical measurements were performed using an Ivium Compactstat TM (Netherlands) potentiostat. Measurements were carried out using a typical three electrode system with a Pt wire counter electrode and a saturated calomel electrode (SCE) as the reference electrode. The working electrodes used were as follows: an edge plane pyrolytic graphite (EPPG) (Le Carbone, Ltd. Sussex, UK) electrode, which was machined into a 4.9 mm diameter, with the disc face parallel with the edge plane as required from a slab of highly ordered pyrolytic graphite (HOPG); a glassy carbon (GC) electrode (3 mm diameter, BAS, USA); a boron-doped diamond (BDD) electrode (3 mm diameter, BAS, USA); a Pt electrode (3 mm diameter, BAS, USA); and screen-printed graphite electrodes (SPE), which have a 3 mm diameter working electrode. The SPEs were fabricated in-house with an appropriate stencil using a DEK 248 screen-printing machine (DEK, Weymouth, U.K.). 37 These electrodes have been used extensively in previous studies. 1, [38][39][40][41] For their fabrication, first, a carbon-graphite ink formulation (product code C2000802P2; Gwent Electronic Materials Ltd., U.K.) was screenprinted onto a polyester (Autostat, 250 µm thickness) flexible film (denoted throughout as standard-SPE); This layer was cured in a fan oven at 60 °C for 30 minutes. Next, a silver/silver chloride reference electrode was included by screen-printing Ag/AgCl paste (product code C2040308D2; Gwent Electronic Materials Ltd., U.K.) onto the polyester substrates and a second curing step was undertaken where the electrodes were heated at 60 °C for 30 minutes. Finally, a dielectric paste (product code D2070423D5; Gwent Electronic Materials Ltd., U.K.) was then printed onto the polyester substrate to cover the connections. After a final curing at 60 °C for 30 minutes these SPEs are ready to be used. These SPEs have been reported previously and shown to exhibit a heterogeneous electron transfer (HET) rate constant, k o , of ca. 10 -3 cm s -1 , as measured using the [Ru(NH 3 ) 6 ] 3+/2+ redox probe. 40,[42][43][44][45] For the purpose of this work, electrochemical experiments were performed using the working electrode of the SPEs only and external reference and counter electrodes were implemented as detailed earlier to allow a direct comparison between all the utilised electrodes as well as with academic literature.
The 2D-MoS 2 was commercially procured from 'Graphene Supermarket' (Reading, MA, USA). 46 The 2D-MoS 2 nanosheets have a reported purity of >99% and are dispersed in ethanol at a concentration of 18 mg L -1 . 46 Our previous work has implemented extinction spectroscopy (ESI Figure 1) to determine the lateral length and number of 2D-MoS 2 nanosheets in our commercially sourced sample which are found to correspond to 61.5 nm and an average of 3 (2.89) monolayers per nanosheet, respectively. 1, 46 The modification of each electrode was carried out using a drop casting approach, where an aliquot of the 2D-MoS 2 suspension was deposited onto the desired supporting electrode surface using a micropipette. 42 This deposition was allowed to dry for 5 minutes (at 35 o C) to ensure complete ethanol evaporation. Finally, the electrode was allowed to cool to ambient temperature, after which the process was repeated until the desired mass was deposited onto the surface at which point the electrode was ready to be used.
Where specific masses of modification are donated within the paper (i.e. ng cm -2 ), note that this value represents the quantity/mass of 2D-MoS 2 that will be present over the averaged area specified and this does not stipulate that an even spread/distribution of monolayer 2D-MoS 2 is present. Rather, the reader should be aware that in reality it is likely that there are areas of multilayer, bilayer and indeed monolayer 2D-MoS 2 randomly distributed across the electrode surface. 47 Interested readers are directed to ESI Figure 2, which shows how different masses of 2D-MoS 2 distribute across the surface of a SPE. Essentially, the values reported represent the mass of 2D-MoS 2 deposited respective to the area of the electrode utilised.
An Agilent 8453 UV-visible Spectroscopy System (equipped with a tungsten lamp assembly, G1315A, 8453 for absorption between 250 nm and 1500 nm and a deuterium lamp, 2140-0605 for absorption between 200 nm and 400 nm) was used to obtain the absorption spectroscopy. The absorption spectrum was analysed using UV-Visible ChemStation software.
Scanning electron microscope (SEM) images and surface element analysis were obtained using a JEOL JSM-5600LV model SEM equipped with an energy-dispersive X-ray microanalysis (EDS) package. Transmission electron microscopy (TEM) images were obtained using a 200 kV primary beam under conventional bright-field conditions. The 2D-hBN sample was dispersed onto a holey-carbon film supported on a 300 mesh Cu TEM grid. Raman Spectroscopy was performed using a 'Renishaw InVia' spectrometer equipped with a confocal microscope (×50 objective) and an argon laser (514.3 nm excitation). Measurements were performed at a very low laser power level (0.8 mW) to avoid any heating effects. X-ray diffraction (XRD) was performed using an "X'pert powder PANalytical model" with a copper source of K α radiation (of 1.54 Å) and K ß radiation (of 1.39 Å), using a thin sheet of nickel with an absorption edge of 1.49 Å to absorb K ß radiation. The Omega was set to 3.00 and the 2θ range was set between 10 and 100 2θ in correspondence with literature. 48 Additionally, to ensure well defined peaks, an exposure of 100 seconds per 2θ step was implemented for all the above analysis. 2D-MoS 2 was utilised after deposition onto a sterilised glass slide (coated with excess 2D-MoS 2 in ethanol then allowed to dry) or a silicon wafer where appropriate. The X-ray photoelectron spectroscopy (XPS) data was acquired using a bespoke ultra-high vacuum system fitted with a Specs GmbH Focus 500 monochromated Al Kα X-ray source, Specs GmbH Phoibos 150 mm mean radius hemispherical analyser with 9-channeltron detection, and a Specs GmbH FG20 charge neutralising electron gun. 49 Survey spectra were acquired over the binding energy range 1100 -0 eV using a pass energy of 50 eV and high resolution scans were made over the C 1s and O 1s lines using a pass energy of 20 eV. Under these conditions the full width at half maximum of the Ag 3d 5/2 reference line is ca. 0.7 eV. In each case, the analysis was an area-average over a region approximately 1.4 mm in diameter on the sample surface, using the 7 mm diameter aperture and lens magnification of ×5. The energy scale of the instrument is calibrated according to ISO 15472, and the intensity scale is calibrated using an in-house method traceable to the UK National Physical Laboratory. 50 Data were quantified using Scofield cross sections corrected for the energy dependencies of the electron attenuation lengths and the instrument transmission. 51 Data interpretation was carried out using CasaXPS software v2.3.16. 52

Characterisation of the commercially obtained 2D-MoS 2
Extensive physiochemical characterisation of the 2D-MoS 2 has been previously conducted and reported, 1 including: Raman spectroscopy, EDS, SEM, TEM, UV-Vis spectroscopy, XRD and XPS. Full characterisation is presented in the ESI and is summarised below for convenience. Despite some aggregation, which is the case for all 2D materials, upon close inspection a lateral grain size of ca. 100 -400 nm is evident. UV-Vis (ESI Figure 1) indicates that the lateral length and stacking number of the 2D-MoS 2 corresponds to 61. 53 This implies that upon deposition of 2D-MoS 2 , utilised herein, onto the supporting electrode materials that the structural model is likely that of re-assembly, with few-layer nanosheets forming as bulk.
Thus the 2D-MoS 2 utilised in this work has been fully characterised and revealed high quality, few layer sheets of MoS 2 , which are next implemented towards the ORR.

Catalytic activity of 2D-MoS 2 towards the ORR at an assigned coverage
Previous work focused on using 2D-MoS 2 as an electrocatalyst for the HER and showed 2D-MoS 2 to be electroactive when immobilised on carbon based electrode substrates. 1 It was therefore essential to benchmark the electrochemical activity of the 2D-MoS 2 when electrically wired using BDD, EPPG, GC and SPEs and explored in degassed 0.1M H 2 SO 4 . This was to ensure that no electroactivity was observed in the region of a linear sweep voltammogram (LSV) where the ORR is expected to occur, as this would convolute the interpretation of the ORR, the results of which can be observed in ESI Figure 10. EPPG, GC and SPE respectively. All of which are significantly more electronegative than that of the Pt's ORR peak and onset potential of + 0.46 and + 0.13 V respectively. The lack of an observable oxygen reduction peak for the BDD electrode (whilst using an acidic electrolyte) corresponds with previous literature. 22 Yano et al. 54 suggest that for ORR to be initiated at a BDD electrode it must first undergo a pre-treatment step at + 1.4 V vs. (Ag/AgCl). 22 This pretreatment step serves to oxidise the sp 2 hybridised carbon species, the likely location for the sp 2 species being the grain boundaries of the sp 3 diamond structure. 19 The oxidised sp 2 species subsequently mediate the ORR. ORR to occur when compared against the bare/unmodified electrodes. Thus, there has been a reduction in the reactions activation energy to a potential that is closer to the value obtained at the unmodified Pt electrode (ca. + 0.46 V). The above data implies that 2D-MoS 2 is an effective electrocatalyst for the ORR when electrically wired with various carbon based electrodes.

Electrocatalytic activity of 2D-MoS 2 towards the ORR at differing coverages
Previous work utilising 2D-MoS 2 as an electrocatalyst for the HER revealed that there is an optimal immobilised mass, where the structure of said material has the highest ratio of active edge planes to comparatively inert basal planes. 1 We therefore investigated the effect of altering the immobilised mass of 2D-MoS 2 onto the carbon based electrodes upon the ORR. Figure 2 shows It is apparent from the above discussion (and inspection of Figure 2 endeavour to vary the mass of 2D MoS 2 utilised in order to deconvolute its optimum electrocatalytic activity. It also proves that 2D-MoS 2 is a promising catalyst that could be utilised to increase the efficiency and energy output of hydrogen fuel cells, thereby making them a more viable alternative to FF combustion as a method of energy generation.

Tafel assessment of the reaction pathway mechanism
It is evident from above that immobilisation of 2D-MoS 2 onto a carbon based electrode substrate reduces the ORR onset and peak potential. Next, consideration was given to the question of whether 2D-MoS 2 once, immobilised onto the carbon based electrodes demonstrated preferential selectivity for the ORR to occur via the desirable 4 electron pathway (producing H 2 O) or the 2 electron pathway (producing H 2 O 2 , which is detrimental to PEM fuel cells). 19 Tafel analysis is a common approach employed within the literature to deduce the number of electrons involved in the ORR electrochemical mechanism. 59 Initially, a plot of ln (I) vs. E p (V) was considered for each of the four carbon based electrodes (see ESI Table 3 and ESI Figure 11) and for each mass of 2D-MoS 2 modification.
This was performed via analysis of the voltammograms depicting the ORR (which were utilised to produce Figure 2) and using the following equation 2.21, respectively, for n involved in their ORR mechanism on a bare/unmodified electrode, followed by a slight increase to a maximum value of 2.63 at 762 ng cm -2 for EPPG and 2.51 at 1009 ng cm -2 for GC. A gradual decrease is then observed with greater masses of immobilisation until EPPG has a 2 electron process at 2533 ng cm -2 and GC has a 1.56 electron process at 2533 ng cm -2 . BDD remains relatively stable in the ORR reaction mechanism between 2 to 2.5 n involved for a range of modifications between 1009 to 2533 ng cm -2 , there appears to be a slight decrease with greater masses of 2D-MoS 2 immobilisation, however it is of little significance.
The results above show that for bare/unmodified and 2D-MoS 2 wired BDD, EPPG and GC the n involved never exceeds n = 3 which suggest that H 2 O 2 is the major product of the reaction occurring rather than the desired H 2 O. It can therefore be assumed that whilst 2D-MoS 2 lowers the ORR onset and peak potential for BDD, EPPG and GC electrodes, it has a minor effect upon the reaction mechanism taking place.
Note, of the carbon based electrodes utilised within this study, GC had the lowest number of electrons involved in its ORR reaction mechanism, this raises the question of why it is the commonly used electrode within the literature as it is clearly the least effective at enabling the desirable 4 electron ORR mechanism. Future studies should use a range of bare/unmodified carbon based electrodes, which exhibit different HET kinetics resulting in unique interactions between the supporting carbon based electrode and any deposited material, breaking from the convention of solely using GC; this will help establish the true electrocatalytic activity of a given material.
SPEs show the highest initial n involved in the ORR reaction mechanism at 2.67 for a bare/unmodified electrode (this corresponds to the literature). 60 From 256 to 1009 ng cm -2 of 2D-MoS 2 immobilised on a SPE's surface there is an increase in the n involved in the ORR reaction mechanism to 3.96. Greater than 1009 ng cm -2 masses of 2D-MoS 2 modification result in a decrease in the n involved until n is 2.64 at 2533 ng cm -2 . Unlike BDD, EPPG and GC electrodes, it is clear that 2D-MoS 2 , once deposited onto a SPE, not only results in a significant decrease in the ORR onset and peak position but also in a beneficial change in the ORR reaction mechanism from ca. 2 to a 4 electron process. Indicating that the major product of the ORR is the desired H 2 O and not the detrimental H 2 O 2 . The reason for 2D-MoS 2 altering the n involved for SPE and not for BDD, EPPG and GC is likely due to the SPEs having "rougher" surfaces, resulting in the 2D-MoS 2 once deposited exhibiting structural/electronic orientations not capable on the "smoother" surface of BDD, EPPG and GC. 60 A comparison was made between the surface topography of BDD, EPPG, GC and SPE using white light profilometry (a ZeGage 3D Optical Surface Profiler, produced by Zygo, was utilised for this). The surface of a SPE was observed to be significantly rougher, with a root mean squared value of the heights over the whole surface (SQ) of 1904.9 nm, than that of BDD, EPPG and GC which had values of 7.5, 26.1 and 15.9 nm respectively (See ESI Figure 12). Next, it was necessary to determine whether  Table 2) and allowed the ORR to occur via a 3.4 electron pathway both of which are significantly less than that of the unpolished (rougher) alternative. We infer that the increased catalytic behaviours, observed for a rougher surface electrode are due to the unique structural/electronic orientations which are formed once 2D-MoS 2 is immobilised onto an SPE. Resulting in an exposure of larger numbers of active edge plane sites/edge planelike defects than their BDD, EPPG and GC counterparts, thereby, offering a greater catalytic prospective. Future studies should consider which supporting material they employ as the results observed above show that this has a significant effect upon the deposited material's structure and electron transfer kinetics.
Whilst other studies have managed to produce a 4 electron pathway using alkaline conditions, (such as Suresh et al. 35 ) we believe, given that all previous studies utilising MoS 2 materials towards the ORR are shown in Table 1, that this report is the first to observe the ORR occur via the 4 electron pathway (thus producing H 2 O rather than H 2 O 2 ) in acidic conditions using an 2D-MoS 2 based electrocatalytic material on a carbon based substrate (SPEs). Clearly, these results are of significant importance as it is acidic conditions found within a PEM fuel cell, thusly making the results of this study highly applicable to real world industry.
This work clearly indicates that there is an optimal/critical mass, which we determine to be ca. 1009 ng cm -2 for SPEs, whereby there is the largest average n (4) involved in the ORR reaction mechanism as well as a significant improvement in the ORR onset and peak potential.
Subsequent studies within the literature which use 2D-MoS 2 should consider using a range of differing loadings/modifications in order to deconvolute the true/optimal electrocatalytic performance of a given electrocatalyst. The findings of this study have clear implications that are applicable when using any 2D material.

Conclusions
This study sought to break from the conventions found within the literature when 2D-MoS 2 materials are explored towards the ORR, of solely using GC as a supporting electrode, using only one mass of the electrocatalytic material to modify the supporting electrode and using KOH as the electrolyte.
Our investigations implemented a range of diligent control experiments. Rather than solely using GC as a supporting electrode we employed BDD, EPPG, GC and SPE's. The ORR onset was reduced to ca. + 0.1 V for EPPG, GC and SPEs at a 2D-MoS 2 1524 ng cm -2 modification, which is far closer to Pt at + 0.46 V compared to the bare/unmodified EPPG, GC and SPE counterparts. BDD was observed to have an ORR onset potential of -0.03 V at 2D-MoS 2 1524 ng cm -2 modification. Using a range of 2D-MoS 2 modification masses rather than one set mass allowed us to observe once a critical mass of 2D-MoS 2 had been achieved (in this case ca. 1009 ng cm 2 ). At this critical mass, there is optimal catalytic activity, after which the catalytic benefits plateau with additional masses of 2D-MoS 2 immobilisation. This is as a result of the structure of 2D-MoS 2 at the critical mass exposing the largest ratio of electroactive edge planes after which there the structure is that of bulk MoS 2 . 0.1 M H 2 SO 4 was utilised as an electrolyte for all the experiments described herein, unlike previous studies which used KOH.
Performing the experiments in an acidic electrolyte resembles the conditions that PEM fuel cells operate, making the observations presented herein highly applicable to industry.
SPEs were the only carbon based electrode found to allow the ORR to occur via the desirable 4 electron pathway (producing H 2 O rather than H 2 O 2 ) at 2D-MoS 2 (ca. 1009 ng cm 2 ). This is likely as a result of the structurally rougher SPE surfaces allowing for unique 2D-MoS 2 structural/electronic orientations, where larger numbers of active edge planes are exposed, which are not possible on the "smoother" BDD, EPPG and GC electrodes. Whilst other reports have managed to produce a 4 electron process we believe that this report is the first to observe the ORR to occur via a 4 electron process in acidic conditions using a 2D-MoS 2 based electrocatalyst material on a carbon based substrate. There is no reason why the findings of this study would not be applicable to other 2D materials, this opens up new avenues of research where the surface roughness of a supporting electrode could be altered allowing 2D materials to exhibit unique and unreported structural/electronic orientations and electrochemical behaviours.
By straying from these literature conventions we de-convoluted the true electrochemical behaviour of 2D-MoS 2 and revealed SPEs as a valid alternative to GC for research purposes and for Pt in real world fuel cell applications. SPEs are significantly cheaper, adaptable and mass producible when compared to Pt and other carbon based electrodes examined herein, whilst upon modification with an optimal mass of 2D-MoS 2, exhibit preferential electrocatalytic activity towards the ORR.