Antonia
Kagkoura
a,
Mario
Pelaez-Fernandez
b,
Raul
Arenal
bcd and
Nikos
Tagmatarchis
*a
aTheoretical and Physical Chemistry Institute, National Hellenic Research Foundation, 48 Vassileos Constantinou Avenue, 11635 Athens, Greece. E-mail: tagmatar@eie.gr
bLaboratorio de Microscopias Avanzadas, Instituto de Nanociencia de Aragon, Universidad de Zaragoza, 50018 Zaragoza, Spain
cARAID Foundation, 50018 Zaragoza, Spain
dInstituto de Ciencias de Materiales de Aragon, CSIC-U. de Zaragoza, Calle Pedro Cerbuna 12, 50009 Zaragoza, Spain
First published on 31st January 2019
A facile route for the preparation of molybdenum disulfide (MoS2) and tungsten disulfide (WS2), uniformly deposited onto sulfur-doped graphene (SG), is reported. The realization of the SG/MoS2 and SG/WS2 heterostructured hybrids was accomplished by employing microwave irradiation for the thermal decomposition of ammonium tetrathiomolybdate and tetrathiotungstate, respectively, in the presence of SG. Two different weight ratios between SG and the inorganic species were used, namely 3:1 and 1:1, yielding SG/MoS2 (3:1), SG/MoS2 (1:1), SG/WS2 (3:1) and SG/WS2 (1:1). SG and all newly developed hybrid materials were characterized by ATR-IR and Raman spectroscopy, TGA, HR-TEM and EELS. The electrocatalytic activity of the SG/MoS2 and SG/WS2 heterostructured hybrids was examined against the hydrogen evolution reaction (HER) and it was found that the presence of SG not only significantly improved the catalytic activity of MoS2 and WS2 but also made it comparable to that of commercial Pt/C. Specifically, hybrids containing higher amounts of SG, namely SG/MoS2 (3:1) and SG/WS2 (3:1), exhibited extremely low onset overpotentials of 26 and 140 mV vs. RHE, respectively. The latter results highlighted the beneficial role of SG as a substrate for immobilizing MoS2 and WS2 and stressed its significance for achieving optimum electrocatalytic performance toward the HER. Finally, examination of the Tafel slopes as extracted from the electrocatalytic polarization curves, manifested the adsorption of hydrogen as the rate-limiting step for SG/MoS2 (3:1), while for SG/WS2 (3:1) the electrochemical desorption of adsorbed hydrogen atoms to generate hydrogen was revealed to be the rate-limiting step.
In general, TMDs are two-dimensional layered materials composed of stacks of atomic metal layers sandwiched by chalcogen layers, in which the individual chalcogen–metal–chalcogen layers weakly interact each other. These materials are considered to be the inorganic analogues of graphene, as the weak van der Waals forces between the layers can be easily overcome when exfoliated from the bulk, leading to few or even monolayered sheets. Molybdenum and tungsten disulfide, MoS2 and WS2, respectively, as typical examples of TMDs, have been exfoliated from the bulk by diverse wet chemistry approaches, mainly assisted by sonication,1–5via a top-down approach. The exfoliated TMDs exhibit dramatically different properties from the bulk material,6 and their novel physical and electronic properties7,8 make them suitable for energy related applications.9
Recently, TMDs have started to be used as effective materials for the HER, owing to their appealing electrocatalytic properties and based on the improvement of the energy conversion efficiency through harvesting a higher current density at a lower overpotential.10–18 The HER activity of TMDs is directly related with exposed edges, in contrast to the catalytically inert basal planes.19 Hence, in order to improve the electrocatalytic performance of TMDs, the density of active edges should be increased. This can be achieved by edge and/or defect engineering, based on unsaturated chalcogen atoms at the edges15,20–22 and/or by promoting the electron transport efficiency between the electrode and the electrocatalyst. Furthermore, the population of defect sites can be adjusted by adopting a synthetic bottom-up approach of TMDs coupled with the presence of an additional component playing the role of the electrocatalyst substrate.23 Specifically, incorporating TMDs on supports with high surface area can provide more active edge sites for electrocatalysis, while by employing highly conductive supports, such as graphene, fast electron transport is guaranteed.
Graphene, due to its large surface area, excellent electrical conductivity and high chemical stability, has been widely employed as supporting material for electrocatalysts in general24 and for TMDs in particular.18,25–32 In the same context, introduction of heteroatoms within the sp2 hybridized carbon network of graphene alters its electrical properties, resulting in enhanced electrocatalytic activity. The enhanced catalytic activity of graphene-doped materials is attributed to the electronegativity difference between carbon and the doping element, which polarizes the adjacent carbon atoms in the graphene lattice, hence potentially facilitating the HER. For example, this was true when N-doped graphene/MoS2 nanocomposites were prepared and found to exhibit high catalytic activity for the HER.33,34 However, beyond N-doped graphene and the aforementioned results, the examination of electrocatalytic properties of other heteroatom-doped graphene materials as substrates for TMDs has yet to be realized. Particularly, focusing on sulfur-doped graphene (SG), an n-type doping effect in graphene, analogous to that observed in nitrogen-doped graphene, exists.35 However, since the difference in electronegativity between S and C as compared to that between N and C is small, an alternative mechanism for improved electrocatalytic activity is prevalent in SG.36 Briefly, incorporation of sulfur within the skeleton of graphene modifies the electronic structure of the material by inducing a non-uniform spin density distribution, which derives from the mismatch of the outermost orbitals of sulfur and carbon atoms, being responsible, along with the charge density, for the SG intrinsic electrocatalytic activity.37 Considering all the above points, the development of hybrid heterostructures, incorporating sulfur-doped graphene sheets and TMDs, as electrocatalysts is absolutely necessary and surely deserves examination.
The current study goes beyond the state of the art, by in situ growing MoS2 and WS2 onto SG, aimed at the development of efficient HER heterostructured electrocatalysts. Herein, based on the microwave-assisted thermal decomposition of the inorganic precursor species of MoS2 and WS2, namely (NH4)2MoS4 and (NH4)2WS4, respectively, in the presence of SG as obtained using a facile and low-cost route, the preparation of SG/MoS2 and SG/WS2, for two different weight ratios between the inorganic species and the SG, was accomplished. The newly developed heterostructures were characterized by Raman spectroscopy, while transmission electron microscopy (TEM) imaging allowed their morphological evaluation and electron energy loss spectroscopy (EELS) provided necessary information of the elements present in these structures. Furthermore, the SG/MoS2 and SG/WS2 heterostructured hybrids were tested towards the HER and it was found that the presence of SG not only significantly improved the catalytic activity of MoS2 and WS2 but also made it comparable to that of the commercial Pt/C catalyst. The exceptional electrocatalytic functioning of SG/MoS2 and SG/WS2 was attributed to the following reasons: (a) the electronegativity difference between carbon and sulfur inducing n-type behaviour in S-doped graphene by polarizing the adjacent carbon atoms to sulfur in the graphene lattice and facilitating the HER, (b) the enhanced population of defect sites in TMDs, as a result of the bottom-up approach employed, (c) the intimate contact of SG with TMDs, (d) the uniform immobilization of TMDs onto the SG surface, in the absence of any organic/surfactant species, and (e) the synergetic effect between SG and TMDs.
Fig. 1 Illustration of the preparation of sulfur-doped graphene (SG) with Lawesson's reagent and fabrication of SG/MoS2 and SG/WS2 hererostructures via microwave irradiation. |
Direct evidence for the realization of SG was obtained by vibrational spectroscopy and thermogravimetric analysis. In more detail, the ATR-IR spectrum of SG shows a broad band at 1080 cm−1 corresponding to C–S vibrations (Fig. 2a), while bands related to oxygen species owing to the starting material graphene oxide and attributed to carbonyl stretching vibrations of carboxylic acid moieties at 1720 cm−1 and to ether groups at 1220 and 1050 cm−1 disappeared or were significantly reduced. Furthermore, the Raman spectrum of SG was changed as compared to that of graphene oxide. Evidently, the sp2 related G-band was shifted to lower frequencies by 8 cm−1, at 1593 cm−1 (Fig. 2b), as compared to that attributed to graphene oxide, indicating that incorporation of sulfur within the lattice of graphene results in n-doping for SG.39 Moreover, the ID/IG ratio was found to increase for SG as compared to that for graphene oxide. This is due to the increase of the defectiveness attributed to etching of graphene sheets, with the simultaneous formation of new and/or more graphene domains with continuous sp2 hybridization, resulting from a reduction process that also takes place along with the sulfur-doping upon treatment with Lawesson's reagent.38 The latter was further confirmed by thermogravimetric analysis (TGA), which revealed a 50% reduction on the mass loss observed from the thermal decomposition of SG as compared to the value registered for graphene oxide. In Fig. 2c the TGA graphs of SG and graphene oxide are presented and compared. Evidently, a 10% mass loss was observed in the temperature range of 230–420 °C under an inert atmosphere for SG, as opposed to the 20% for graphene oxide in the temperature range 150–260 °C. The shift to a higher decomposition temperature for SG as opposed to that for graphene oxide is justified by considering that Lawesson's reagent not only incorporated sulfur species within the graphene lattice but also reduced the oxygenated moieties, restoring to some extent the graphene network by forming islands with a continuous sp2 structure. The mass loss occurrig above 420 °C is attributed to the thermal decomposition of sp3 defects created at sites nearby where sulfur doping took place.
Focusing on SG/MoS2 and SG/WS2, Raman spectroscopy revealed characteristic modes due to both components, regardless of the weight ratio of 3:1 or 1:1 between the SG and the inorganic precursor species for MoS2 and WS2 employed for the preparation of the hybrids. This is to say that, for SG/MoS2, bands at 379 and 405 cm−1 (Fig. 3a), corresponding to the in-plane E12g and out-of-plane A1g vibrational modes of MoS2, respectively,1,40,41 and for SG/WS2 at 352 and 415 cm−1 (Fig. 3b), corresponding to the E12g and A1g modes of WS2, respectively,42 were observed, while in both spectra the D and G bands due to SG were present at 1352 and 1593 cm−1, respectively. The ID/IG ratio for SG/MoS2 and SG/WS2 was found almost unaltered, ca. 1.08 and 1.12 respectively, as compared to that registered for SG, demonstrating that the microwave irradiation conditions employed for the growth and incorporation of MoS2 and WS2 did not cause any serious damage to the framework of SG.
Fig. 3 Raman spectra (514 nm) of (a) SG/MoS2 (3:1) and (b) SG/WS2 (3:1) heterostructured hybrid materials. |
Thermogravimetric analysis (TGA) assays, performed under a nitrogen atmosphere, allowed the estimation of the amount of MoS2 and WS2 within the hybrid materials. Specifically, SG/MoS2 (3:1) and SG/MoS2 (1:1) were found to thermally lose 22 and 26% of mass, respectively, in the temperature range of 200–800 °C (ESI, Fig. S1a†). Based on the latter observation and considering that SG is also thermally labile, showing around 10% mass loss in the temperature range of 230–420 °C and another 6% up to 800 °C (Fig. 2), while intact MoS2 shows a continuous mass loss over the whole temperature range (ESI, Fig. S1a†) due to the presence of defect sites, the rough content of MoS2 within SG/MoS2 (3:1) and SG/MoS2 (1:1) was estimated to be 37 and 66%, respectively. Similarly, from the TGA profiles for SG/WS2 (3:1) and SG/WS2 (1:1) (ESI, Fig. S1b†), while taking into account the overall 16% mass loss due to SG, the content of WS2 in the two hybrids was estimated to be 35 and 64%, respectively.
Fig. 4 shows HR-TEM images and EEL spectra of intact MoS2 and WS2 as well as of SG/MoS2 and SG/WS2 hybrids for the two different weight ratios 3:1 and 1:1 screened. Intact MoS2 and WS2 showed different nanostructures with different crystallographic orientations, aggregated in clusters with size in tenths of nm. This behavior can be clearly seen in the FFT diffraction pattern of MoS2 (inset of Fig. 4a), where several defined spots can be observed. As for the SG/MoS2 and SG/WS2, the TEM analysis showed a variation in contrast, which is believed to be related to the presence of MoS2 and WS2, respectively, on SG – see for example the crumpled structures in Fig. 4d. The EELS results, shown in Fig. S2a, S2b1 and S2b2† validated the presence of MoS2 and WS2 in SG, which is coherent with the rest of the microscopy and spectroscopic studies performed. In addition, SEM imaging of SG (ESI, Fig. S3†) together with SEM and EDX analysis for MoS2, SG/MoS2 and SG/WS2 (ESI, Fig. S4 and S5†) and TEM imaging (ESI, Fig. S6†) further confirmed the presence of MoS2 and WS2 in all the hybrid materials.
Next, the electrocatalytic activity of SG/MoS2 and SG/WS2 toward the HER was examined by performing linear sweep voltammetry measurements with a rotating-disc working glassy carbon electrode in a standard three-electrode glass cell at a scan rate of 5 mV s−1 in nitrogen saturated 0.5 M aqueous sulfuric acid. Polarization curves for the SG/MoS2 (3:1) and SG/MoS2 (1:1) hybrids, accordingly denoted for the different weight ratios of SG and the inorganic precursor species for MoS2 employed for the preparation, together with those for individual MoS2 and SG along with that for commercially available Pt/C as a reference are shown in Fig. 5a. Evidently, for a given potential, the cathodic current increased for SG/MoS2 as compared to that for individual MoS2 and SG. The onset overpotential for SG/MoS2 (1:1) was registered at −0.175 V vs. RHE, and shifted by 0.215 V to a more positive potential as compared to that for individual MoS2 appearing at −0.390 V vs. RHE. Notably, for the electrocatalyst with a higher amount of SG, namely SG/MoS2 (3:1), the onset overpotential is similar to that of commercially available Pt/C, appearing at −0.026 V vs. RHE (Fig. 5a). The same trend was identified for SG/WS2, in which the heterostructured hybrid with the higher amount of SG, namely SG/WS2 (3:1), showed lower overpotential, ca. −0.140 V vs. RHE as compared not only with that of the individual WS2, ca. −0.460 V, but also with that of SG/WS2 (1:1), ca. −0.390 V (Fig. 5b). An overall graph showing the onset overpotential for all examined materials is presented in Fig. 5c. The aforementioned results not only highlight the beneficial role of SG as the substrate for uniformly immobilizing MoS2 and WS2, but also more importantly stress the importance of the amount of SG relative to MoS2 and WS2 for achieving optimum electrocatalytic performance toward the HER. Overall, a synergetic effect attributed to both the conductive nature of SG in intimate contact with MoS2 and WS2, allowing effective charge transport, and the presence of active edge sites in MoS2 and WS2, leads to more active electrocatalytic behavior towards the HER with comparable performance with that of platinum. Focusing on the best performing hybrids as electrocatalysts for the HER, namely SG/MoS2 (3:1) and SG/WS2 (3:1), bubbles of hydrogen were observed to evolve at a cathodic current density of 0.6 and 0.3 mA cm−2, respectively, with enhanced rates at around −0.290 and −0.240 V, respectively. Considering that the amount of cathodic current density is proportional to the amount of hydrogen evolved and since it is very common to compare electrocatalysts for the HER against overpotentials at a cathodic current density of 10 mA cm−2, the corresponding tabulated graph for all the examined materials, shown in Fig. 5d, nicely demonstrates the highest activity for SG/MoS2 (3:1) and SG/WS2 (3:1), with the former performing best at lower overpotentials (cf.Fig. 5c).
In order to obtain meaningful insights and characterize the charge transport efficiency and efficacy of the electrocatalytic HER, Tafel slopes for SG/MoS2 (3:1), SG/MoS2 (1:1), SG/WS2 (3:1) and SG/WS2 (1:1) were extracted from the LSV polarization curves and presented in Fig. 6a and b, respectively. In addition, the tabulated Tafel values are given in Fig. 6c. Generally, the Tafel slope is an inherent property of the electrocatalyst that is determined from the rate-limiting step of the HER. Hence, analysis of the Tafel plot data aids in the elucidation of the HER mechanism and more importantly in the identification of the rate-limiting step of the reaction. With all this in mind, the dominant mechanism of hydrogen evolution for the materials tested is interpreted by considering initial adsorption of a proton onto the electrode surface via a reduction process, according to Volmer adsorption – eqn (1). Next, bonding of the adsorbed hydrogen with a proton and electron transfer from the electrode surface takes place, according to Heyrovsky desorption – eqn (2). Alternatively, recombination of two hydrogen atoms adsorbed on the electrode surface occurs, according to Tafel desorption – eqn (3).
H3O+ + e− → Hads + H2O | (1) |
Hads + H3O+ + e− → H2 + H2O | (2) |
Hads + Hads → H2 | (3) |
Individual MoS2, WS2 and SG exhibited relatively high Tafel slopes, 271, 246 and 153 mV dec−1, respectively. Realization of the SG/MoS2 and SG/WS2 heterostructured hybrids, regardless of the relative amount of SG versus the transition metal precursor employed, i.e. 3:1 and 1:1, caused a drop in the corresponding Tafel value (Fig. 6). Since a smaller Tafel slope implies that for the generation of an equivalent current only a lower overpotential is required, the electrocatalytic activity of MoS2 and WS2 was improved, particularly for the SG/MoS2 (3:1) and SG/WS2 (3:1) systems, which showed Tafel slopes of 152 and 53 mV dec−1, respectively. Again, the improvement in charge transport is facilitated by the good electrical contact between SG and TMDs, while the observed Tafel values manifest that the adsorption of hydrogen onto the modified electrode is the rate-limiting step for the SG/MoS2 (3:1) according to eqn (1), while for the SG/WS2 (3:1) the electrochemical desorption of adsorbed hydrogen atoms onto the modified electrode to generate hydrogen is the rate-limiting step, according to eqn (2).
In order to further understand the improved electrocatalytic activity of the SG/MoS2 and SG/WS2 hybrid materials, the electrochemically active surface area (ECSA) was calculated according to the equation ECSA = Cdl/Cs, where Cdl is the electrochemical double-layer capacitance and Cs is the specific capacitance of a flat surface with 1 cm2 real surface area with a value assumed to be 40 μF cm−2 for the flat electrode.43 To this end, measuring cyclic voltammograms in a non-faradaic region at different scan rates of 50, 100, 200, 300, 400 and 500 mV s−1 (ESI, Fig. S7†), allowed the estimation of the ECSA value from the Cdl by plotting the Δj = (ja − jc) at 0.1 V vs. RHE as a function of the scan rate according to the equation Cdl = d(Δj)/2dVb.43 Evidently, the ECSA values of SG/MoS2 (3:1) and SG/WS2 (3:1) were significantly higher than those of SG/MoS2 (1:1) and SG/WS2 (1:1), respectively, as shown in Table 1. These higher ECSA values indicate more effective accessibility of the active sites of the SG/MoS2 (3:1) and SG/WS2 (3:1) hybrid materials, similar to recent reports.44
Catalyst | Onset overpotential (V vs. RHE) | Overpotential (V vs. RHE) at −10 mA cm−2 | Tafel slope (mV dec−1) | ECSA (cm2) |
---|---|---|---|---|
a After 1000 cycles. | ||||
MoS2 | 0.39 | 0.63 | 271 | ∼2.25 |
MoS2a | 0.39 | 0.69 | 240 | — |
SG/MoS2 (3:1) | 0.026 | 0.29 | 152 | ∼15.4 |
SG/MoS2 (3:1)a | 0.033 | 0.30 | 160 | — |
SG/MoS2 (1:1) | 0.175 | 0.37 | 124 | ∼6.9 |
SG/MoS2 (1:1)a | 0.175 | 0.41 | 111 | — |
WS2 | 0.46 | 0.65 | 246 | ∼9.75 |
WS2a | 0.47 | 0.68 | 213 | — |
SG/WS2 (3:1) | 0.14 | 0.25 | 53 | ∼20 |
SG/WS2 (3:1)a | 0.15 | 0.26 | 56 | — |
SG/WS2 (1:1) | 0.39 | 0.55 | 111 | ∼13.87 |
SG/WS2 (1:1)a | 0.39 | 0.57 | 113 | — |
SG | 0.68 | 0.85 | 153 | ∼7.25 |
SGa | 0.68 | 0.87 | 157 | — |
Pt/C | 0.026 | 0.09 | 50 | — |
Pt/Ca | 0.054 | 0.096 | 50 | — |
Finally, the long-term stability of the hybrid materials was evaluated. Durability studies as an important way to assess the electrocatalytic activity of the materials were performed (ESI, Fig. S8†) and it was found that all the tested materials exhibited high stability after cycling continuously for 1000 cycles, with negligible loss of cathodic current. In Table 1, the various HER parameters of all the screened materials, before and after 1000 cycles, are summarized.
Footnote |
† Electronic supplementary information (ESI) available: Images and spectra. See DOI: 10.1039/c8na00130h |
This journal is © The Royal Society of Chemistry 2019 |