Room-Temperature Production of Nanocrystalline Molybdenum Disulfide (MoS2) at the Liquid-Liquid Interface.

Scalable synthesis of 2D materials is a prerequisite for their commercial exploitation. Here, a novel method of producing nanocrystalline molybdenum disulfide (MoS2) at the liquid-liquid interface is demonstrated by decomposing a molecular precursor (tetrakis(N,N-diethyldithiocarbamato) molybdenum(IV)) in an organic solvent. The decomposition occurs over a few hours at room temperature without stirring or the addition of any surfactants, producing MoS2 which can be isolated onto substrates of choice. The formation of MoS2 at the liquid-liquid interface can be accelerated by the inclusion of hydroxide ions in the aqueous phase, which we propose to act as a catalyst. The precursor concentration was varied to minimize MoS2 thickness, and the organic solvent was chosen to optimize the speed and quality of formation. The kinetics of the MoS2 formation has been investigated, and a reaction mechanism has been proposed. The synthesis method is, to the best of our knowledge, the first reported room-temperature synthesis of transition-metal dichalcogenides, offering a potential solution to scalable 2D material production.

Introduction

Two-dimensional (2D) materials have been the focus of significant scientific interest since the discovery of graphene.[1] Having a catalogue of structurally related 2D materials, each with their own characteristic properties, and combining them with atomic precision, in heterostructures for example, generates countless opportunities for applications with 2D materials.[2] One complementary 2D material to graphene is the transition-metal dichalcogenide (TMD) molybdenum disulfide (MoS2). MoS2 displays semiconducting or metallic behavior, dependent on the polytype, which means that it has interesting applications in electronics energy storage and as an electrocatalyst for hydrogen evolution.[3] Few-layer MoS2 exhibits a high specific surface area (180–240 m2 g–1)[4] and excellent charge carrier mobility at room temperature (148 cm2 V–1 s–1) compared with the bulk material.[5] In addition, monolayer hexagonal MoS2 is a direct band gap semiconductor (∼1.9 eV), with strong photoluminescence, whereas its bulk counterpart has a smaller indirect band gap (∼1.2 eV).[6]

Developing a production method capable of generating inexpensive, large-area pristine films of 2D materials is one of the biggest current barriers to commercial exploitation of this class of materials. Large-scale production of 2D materials is generally achieved by chemical vapor deposition (CVD) or liquid-phase exfoliation (LPE). CVD requires dedicated high-vacuum equipment, reactive gas feeds, and the need to remove the underlying growth substrate. LPE of either natural or synthetic materials, albeit inexpensive, produces a polydisperse product with varying material thickness and lateral sizes. In aqueous exfoliation, surfactants are typically included to increase the stability and concentration of the exfoliated material.[7,8] A recently proposed alternative to these two production methods is liquid–liquid (L/L) interfacial assembly of 2D materials, where thin films assemble at the interface between two immiscible liquids. This is driven by a reduction in the total free energy of the system as a result of the reduction in the organic/water interfacial area.[9−13]

2D material growth at the L/L interface is a simple and inexpensive technique that can be achieved with a variety of liquids. Theoretically, the size of the films produced is only limited by the area of the interface. In addition, the formation of a film at the interface does not depend on the substrate and use of a surfactant. Assembly of solids at the L/L interface is well-researched and can be traced back to the work of Faraday,[14] yet assembly of graphene at the L/L interface was only first demonstrated in 2009[15] as an approach to generate a highly ordered monolayer film from pre-exfoliated 2D materials. Graphene[16,17] and graphene oxide[18−20] monolayers have been produced at the L/L interface, while graphene-based supercapacitors[21,22] and TMD photoelectrochemical cells[23] have been produced using L/L interfacial films. Nanomaterials have also been prepared at the L/L interface using bottom-up syntheses. Starting with molecular precursors, a variety of nanostructured materials have been formed including graphene[24] and transition-metal chalcogenides (TMCs) such as CdS.[25] However, there has not been, to the best of our knowledge, any prior reports of a bottom-up synthesis of MoS2 at the L/L interface.

In this report, the assembly of MoS2 at the liquid–liquid interface from the decomposition of a molybdenum coordination complex in an organic solvent was investigated. It has previously been shown that thin films of metal chalcogenides can be produced from the decomposition of molecular precursors at elevated temperatures[26−28] and by tribological stress.[29−31] Here, the production of MoS2 thin films at the liquid–liquid interface at temperatures as low as 298 K is demonstrated for the first time using a single source precursor. A variety of techniques including Raman spectroscopy, powder X-ray diffraction (pXRD), scanning electron microscopy (SEM) combined with energy-dispersive X-ray (EDX) spectroscopy, transmission electron microscopy (TEM), atomic force microscopy (AFM), and X-ray photoelectron spectroscopy (XPS) were used to determine the thickness, chemical composition, and surface quality. A Galvani potential was applied across the interface to give further insight into the reaction mechanism, and the reaction kinetics were also determined from UV–visible absorption spectra. The reaction conditions were optimized by varying concentration, temperature, pH, starting precursor, and organic solvent, which, in combination with mass spectrometry and UV–visible absorption spectroscopy, enabled a reaction mechanism to be elucidated.

Experimental Section

General

1,2-Dichlorobenzene (DCB) (99%) was purchased from Sigma Aldrich, toluene (99%) was purchased from Alfa Aesar, and both were purified by vacuum distillation to minimize the amount of trace impurities in film formation. Sodium sulfide (anhydrous) was purchased from Fluorochem, and sodium hydroxide was purchased from Fluka. Ultrapure water (18.2 MΩ cm, Milli-Q Direct 8) was used in all studies. A sodium alkyl sulfate-containing surfactant, Teepol 610, was purchased from Sigma-Aldrich.

Synthesis of Tetrakis(diethyldithiocarbamate) Molybdenum (IV) (Mo[Et2DTC]4)

The synthesis of the molecular precursor MoL4 was achieved using the method of Decoster et al.[32] Briefly, molybdenum hexacarbonyl was heated under reflux in acetone in the presence of tetraethylthiuram disulfide and held under reflux for 2 h. The mixture was then allowed to cool, and the black microcrystalline powder was isolated by suction filtration and washed with pentane (4.5 g, 59%). ES-TOF+m/z 690 [M + H]+. Anal calcd for C20H40N4S8Mo.H2O: C, 33.98; H, 5.99; N, 7.92; S 36.28; Mo, 13.57%; found: C, 33.79; H, 5.65; N, 7.81; S 36.22; Mo, 13.28%. FTIR (solid) νmax (cm–1): 2970(w), 2930 (w), 2870 (w), 1510 (m), 1440 (m), 1420 (sh), 1380 (sh), 1350 (m), 1300 (sh), 1270 (m), 1250 (sh), 1200 (m), 1150 (m), 1090 (sh), 1070 (m), 1020 (sh), 990 (m), 970 (sh), 950 (sh), 900 (m), 850 (m), 810 (sh), 780 (m), 730 (sh), 710 (w), 660 (m), 610 (m), 580 (m), 560 (m), 490 (w), 470 (sh), 430 (w), 400 (w). The IR spectrum was consistent with the previous literature.[28,33]

Liquid–Liquid Interface Assembly

A 10 mM solution of NaOH or Na2S in water was layered with 2.5 × 10–4 M MoL4 in a variety of different solvents including DCB and toluene in a glass vial. The vial was left at room temperature or heated to the required temperature for each experiment in an oil bath. At 75 °C, the organic phase changes color to pale yellow after approximately 30 min, eventually turning colorless at low concentrations. Upon removal from heat, the organic phase turns cloudy. After leaving to stand at room temperature for approximately 24 h, a film visible to the eye forms between the two liquid phases; a film is also present between the organic phase and the glass vial (Figure ). The film is made up of small flakes, probably held together by weak van der Waals forces and is therefore difficult to remove. Silicon wafers with a 300 nm-thick SiO2 surface layer were dipped at an angle across the interface and washed with organic solvent, water, and acetone before being vacuum-dried for 24 h at 900 mbar and 40 °C. As film transfer onto a substrate is dependent on surface energy, various techniques were used to aid effective transfer. A 200 ppm Teepol surfactant was added to the top phase after the film was removed from the L/L interface to minimize film breakup when passing through the liquid–air interface. Substrates were also treated with dichlorodimethylsilane to increase hydrophobicity, facilitating the removal process.

Figure 1

Assembly of a thin film at the liquid–liquid interface from the decomposition of MoL4 in organic solvent (lower phase) in the presence of a 10 mM solution of NaOH in water. (A) Starting liquid–liquid setup in a 2 cm-diameter vial and (B) two-phase reaction on completion where a film has formed at the interface; (C) zoomed image of the thin film formed at the interface between the two liquids.

Instrumentation

Raman spectroscopy was carried out with a Renishaw inVia confocal Raman microscope. The samples were analyzed using a 523 nm laser at a power of 1 mW with a 100× objective lens, and a grating of 1800 mm–1 was used to achieve a spectral resolution of 1 cm–1. Raman area maps were completed, taking a spectrum every 0.9 μm2, for a high-resolution image. Powder X-ray diffraction (pXRD) analysis was carried out with a Bruker D8 Discover diffractometer. Diffraction patterns were recorded on glass substrates using a Cu Kα radiation source (1.54178 Å) and operating at 40 kV and 30 mA while spinning at 0.5 s–1 between 10 and 70°.

An FEI Quanta 650 FEG SEM was used for scanning electron microscopy (SEM) with an accelerating voltage of 5 kV and using secondary electron detection. Energy-dispersive X-ray (EDX) spectroscopy was carried out at 25 kV. Samples were prepared, as discussed above, on Si/SiO2 wafers, and conductive carbon tape was used to reduce the charging of the sample during imaging. Transmission electron microscopy (TEM) analysis was performed using an FEI Tecnai G20 operating at an accelerating voltage of 200 kV, with selected area electron diffraction (SAED) performed using a 0.75 μm effective diameter SAED aperture. Atomic force microscopy (AFM) was carried out using a Bruker Multimode 8, equipped with a silicon nitride tip in contact mode at a scan rate of 1 Hz.

X-ray photoelectron spectroscopy (XPS) was carried out in a Specs UHV system. The photoejected electrons were excited using a monochromatic Al Kα (1486.6 eV) source operating at 120 W and analyzed using a Phiobos 150 electron energy analyzer operating at 30 eV pass energy. An electron flood gun was used during analysis (20 μA at 2 eV) due to the insulating nature of the SiO2 substrates used, and spectra were charge-corrected to the C 1s line at 284.8 eV. All spectra were analyzed using AAnalyzer and CasaXPS software packages, utilizing a linear background and Voigt functions to fit recorded spectra.

Absorption spectra of MoL4 in DCB solution were measured using a Mikropack DH-1000-BAL UV–vis–NIR spectrometer. A concentration of 2.5 × 10–4 M MoL4 was used in a quartz cuvette. Initial rate calculations were carried out using a Cary 60 UV–visible spectrometer where 2 mL of MoL4 in DCB and 1 mL of Na2S in H2O were layered within a quartz cuvette to allow rapid in situ absorbance spectra of the organic phase to be taken. Absorbance spectra were recorded every ∼3 min for ≤20 min to monitor the initial rate of reaction. Cyclic voltammetry was carried out on an Autolab potentiostat between −0.1 and 0.6 V at a scan rate of 200 mV s–1. Low-resolution mass spectrometry was performed using a Waters SQ detector attached to a Waters Acquity UPLC. High-resolution mass measurements were performed using a Thermo Q-Exactive Plus (ORBITRAP) attached to a Thermo Ultimate 3000 HPLC. Both instruments used an electrospray ion source in positive and negative polarity.

Results

Characterization of MoL4

The absorbance of MoL4 in DCB was measured and gave a characteristic spectrum (Figure S3). The UV–visible absorption spectrum measured is consistent with that reported for molybdenum dithiocarbamate complexes.[33] The peaks between 330 and 590 nm were attributed to low-energy ligand-to-metal charge-transfer (LMCT) bands. Electron paramagnetic resonance (EPR) suggests the presence of Mo5+ in the MoL4 precursor (Figure S4), which suggests that some oxidation may occur in solution, the extent of which cannot be quantified from this data alone.

Thermogravimetric analysis (TGA) was completed under nitrogen and air (Figure S5). Decomposition in nitrogen occurs in three steps as has previously been reported.[34] In both the air and nitrogen cases, the precursor initially follows a similar decomposition: first at ∼100 °C losing 5% of its weight in the air case and 4% in the N2 case, corresponding to the loss of ethylene (4%), and then a second step loss at ∼130 °C of 57% in air and 57% in nitrogen, corresponding to the loss of 2(S2CNEt2) and SCNEt (56%). In a nitrogen atmosphere, the weight loss step at ∼400 °C corresponds approximately to that of MoS2 (25% experimental vs 23% calculated). In air, the third decomposition step corresponds to formation of MoS2 followed by oxidation to MoO3 above 300 °C[35] where the loss is equal to 21% (cf. 20% calculated). We attribute the fourth decomposition step to volatilization of MoO3 as the weight loss tends to become zero. Overall, the TGA suggests either MoS2 or MoO3 can be produced from the starting precursor, which has recently been experimentally confirmed by Lewis and co-workers.[36]

Characterization of MoS2

Raman spectroscopy was used to characterize the few-layer MoS2. The main features of interest in the spectrum are the in-plane E2g and the out-of-plane A1g phonons (Figure ). The difference in the Raman shift of the phonon resonances observed at 381 cm–1 (E2g) and 406 cm–1 (A1g) is consistent with the literature for multilayer MoS2 (>4 layers).[37] The ratio of the intensity of the A1g and E2g phonon resonances is suggestive of a (002) stacking orientation with respect to the incident photons.[38] Large-area flakes >4 μm in length were observed from Raman spectral mapping, and the intensity maps for both peaks are in agreement. Corresponding bright-field and dark-field optical images are shown in Figure S6.

Figure 2

(a) Raman spectrum of MoS2 produced at the L/L interface by reaction assembly. (1) MoL4 in DCB and Na2S in H2O; (2) MoL4 in DCB and H2O; (3) MoL4 in DCB and NaOH in H2O; and (4) MoL4 in toluene and NaOH in H2O. Raman map of peak intensity for MoL4 in DCB and NaOH in H2O at (b) 381 and (c) 406 cm–. Scale bars: 4 μm. (d) pXRD pattern for MoS2 on glass substrate showing the reflection of the (002) plane of MoS2.

pXRD confirmed that crystalline MoS2 was produced at the L/L interface, with Bragg reflections indexed to 2H-MoS2 (Figures S7 and S8). The reflection at 2θ = 14.44° (Figure d) corresponds to the (002) plane in 2H-MoS2 with a d spacing of 6.13 Å, which is consistent with the prior literature (6.15 Å).[39] The preferred orientation of growth in the (002) plane is consistent with the Raman data (vide supra). In addition, the large full width at half-maximum (FWHM) value of 0.45° is consistent with exfoliated MoS2 as the (002) reflection of bulk MoS2 has a much narrower FWHM (0.05°) under the same instrumental settings.[40] This suggests that the material grown at the liquid–liquid interface has nano- to microscale crystalline domains.

Secondary electron scanning electron microscopy (SEM) revealed the morphology of the MoS2 films and the lateral size of the flakes. Both nanoparticles and micron-scale particles were found as shown in Figure . Particle formation varied in size from smaller nanoparticles to large microparticles >10 μm in lateral size. Figure a,b reveals how particles arrange together to form larger microparticles and thin films. In all cases, the observed crystals exhibited typical hexagonal facets associated with 2H-MoS2.

Figure 3

Characterization of MoS2 by electron microscopy. (a) SEM image of an agglomerated MoS2 microparticle comprising smaller nano- and microparticles on a Si/SiO2 substrate. (b) SEM image of a large-area thin film comprising MoS2 nano- and microparticles on Si/SiO2. (c) TEM image of a MoS2 microparticle on a holey carbon grid. The dashed yellow circle indicates the location from which the SAED pattern in (d) was acquired. (d) SAED pattern from the region indicated in (c) with the (110) and (100) spots highlighted in red and blue, respectively.

Energy-dispersive X-ray (EDX) spectroscopic mapping was also used to demonstrate the spatial distribution of the elements found within the films. EDX spectroscopic mapping confirmed that the particles are MoS2 (Figure S9 and Table S2) and there was a strong agreement with the EDX map and the corresponding Raman map (Figure S6). Adventitious carbon is observed in the EDX spectra (Figure S9) alongside oxygen and silicon emission from the underlying Si/SiO2 substrate. Integration of the Kα peak for molybdenum (due to the overlap of the Lα peak) and the Kα peak for sulfur yielded elemental analysis consistent with MoS2 (Mo = 5.6 ± 0.8 atom %; S = 9.1 ± 1.3 atom %; Mo/S ratio is 1: 1.6).

Characterization of a MoS2 flake by transmission electron microscopy (TEM) was performed (Figure c,d). The hexagonal symmetry in the SAED pattern matches the expected symmetry of thin, monocrystalline MoS2 viewed along the [001] zone axis, and the indexed (100) and (110) spots relate to a d spacing of 0.28 and 0.16 nm, respectively, agreeing well with literature values.[41] Additional spots, for example, at 3.6 nm–1 from the [001] zone axis, are due to additional smaller misorientated flakes stacked on the larger single crystal. The TEM data agree well with the Raman, pXRD, and SEM results, strongly indicating that 2H-MoS2 has been produced.

Atomic force microscopy (AFM) was used to determine the film topology. The nature of transfer of the film to a silicon substrate meant that the structure of the film was disturbed. However, the thickness of individual flakes that comprise the film’s surface can be determined, which has been used to ascertain the flake thickness distribution. Figure and Figure S10 demonstrate that similar to SEM analysis, a variety of flakes form from larger-area crystallites >1 μm in length to smaller few-layered nanosheets. The histogram (Figure b) shows that ∼32% of the flakes imaged have a height between 35 and 40 nm and ∼31% of the sample has a thickness between 5 and 15 nm, some of these thinner nanosheets are mapped in Figure c. The correlation between lateral flake size and thickness for this methodology is consistent with that of LPE.[42] However, it is worth noting that this thickness versus flake size distribution has been achieved without any energy input and with minimal optimization. Therefore, there is potential to have significantly enhanced lateral characteristics for a given thickness compared with traditional methodologies due to the thickness confinement of the assembly process and the mild processing conditions used.

Figure 4

(a) AFM image of large-area MoS2 crystallites and few-layered nanosheets on a silicon wafer grown at the L/L interface. Scale bar: 300 nm. (b) Thickness distribution profile as a percentage of voxel height (N = 262,144) and (c) thickness profiles of specific smaller flakes labeled in (a).

X-ray photoelectron spectroscopy (XPS) was used to characterize any changes in the material surface compared with the bulk crystals. XPS results (Figure ) suggest that the surface of the MoS2 crystallites is predominantly oxidized at the surface of the material. However, due to the small amount of material present on the sample compared with the size of the beam, the stoichiometry cannot be accurately determined. Figure displays narrow window XPS data for the Mo 3d/S 2s region of the sample. Peak fitting analysis indicates a strong doublet peak with the main component at 232.2 eV and a separation of 3.15 eV, consistent with molybdenum oxide, MoO3, as seen in previous studies.[43] In addition to the main doublet, a smaller doublet feature was noted, seen as a shoulder feature at lower binding energy. Peak fitting analysis of these features indicates a position of 229.7 eV, which is associated with the Mo4+ 3d5/2 feature. In addition, the Mo 3d spectrum contains a weak S 2s photoelectron peak at 226.9 eV, suggestive of the presence of 2H-MoS2 species.[44] To further confirm the presence of MoS2, a reference MoS2 sample was analyzed immediately following the sample of interest. When overlaid, the spectral features within the reference Mo 3d/S 2s spectrum are in agreement with the lower binding energy doublet features seen within Figure , which gives further confidence to the presence of MoS2 from other characterization techniques.

Figure 5

XPS spectrum of Mo 3d/S 2s region. A main MoOx feature is noted in addition to small shoulders at lower binding energy consistent with the presence of MoS2.

As MoO3 has characteristic diffraction patterns for XRD and TEM SAED, which are not seen during characterization, we can confidently propose that the presence of MoO3 in the XPS is a result of surface oxidation. The penetration depth for Al Kα is ∼6 nm,[36] and therefore from XPS analysis, it is suggested that the top ∼5 layers have been oxidized, which is higher than typical exfoliated MoS2 in aqueous containing solvents.[45] Therefore, it is suggested that in order to reduce surface oxidation for future experiments, all solvents should be degassed prior to use and reactions should be carried out under an inert atmosphere.

Reaction Kinetics

Over time, a film forms at the liquid–liquid interface, and the organic phase turns from a deep red color to pale yellow, which is indicative of a vanishing charge-transfer band associated with the depletion of MoL4 as the reaction progresses. The reaction rate for film formation could therefore be determined from absorption spectroscopy of the components within the organic phase. The molar absorptivity coefficient (ε) of MoL4 at 460 nm as calculated using the Beer–Lambert law was 3210 M–1 cm–1 (Figure S11), consistent with the literature.[33] To determine the order of reaction with respect to [MoL4], [Na2S] in the aqueous phase was kept constant. [MoL4] was varied, and the reaction was monitored to calculate the initial rate, ν0 in M s–1 (Table S3). Once ν0 was calculated, a plot of log(ν0) versus log[MoL4] yielded a slope equal to the order of reaction (Figure S12). Hence, the rate of decomposition of MoL4 (r with units mol dm–3 s–1) and with respect to time (t), in this case, was calculated to be first order and can be written as followswhere k is a unimolecular rate constant with units of s–1. The integrated rate law for this reaction is thereforewhere [MoL4]0 is the concentration of MoL4 at t = 0.

The concentration of the starting complex was optimized to achieve a thin film at the L/L interface. By using the area of a MoS2 unit cell, an approximate concentration was calculated, which would yield, in theory, a film of nanometer thickness for the given vessel dimensions, assuming that the film does not only occupy the L/L interface but also the organic liquid/glass interface. Experimentally, we observed that the majority of the film aggregated to the L/L interface and there was an uneven distribution of flakes; however, it did allow a starting concentration to be chosen that was then further optimized to yield a sufficiently thin film at the L/L interface. The reaction was initially carried out at room temperature, but the rate of decomposition of the precursor and resulting film formation was slow. Therefore the temperature was increased to 75 °C, and the rate constant was measured (Figure S13). In addition, the effect of pH was measured as it was believed from previous results that increasing the alkalinity of the aqueous phase increases the rate of decomposition. It can be seen from Table that as expected, increasing the temperature increases the rate of reaction. A 50 °C increase in temperature causes an increase of approximately a factor of 10 in the rate. Changing the pH also has a significant impact on the rate, and as expected, an increase in the pH increases the rate of reaction. However, it is worth noting that the quality of the film at the L/L interface was adversely affected by increasing the temperature, and therefore even though a faster rate of precursor decomposition was observed, the reaction was preferably performed at room temperature to optimize material quality. At elevated temperatures, the size of the flakes appeared to be smaller, and the quantity of MoS2 produced was less, suggesting further decomposition or that kinetically favorable side reactions occurred at increased temperatures. In addition, the reaction was also affected by the organic solvent, with DCB comparably producing the highest quality product in the least time. The use of toluene produced a different UV–vis spectrum, which was attributed to formation of the neutral Mo4+ complex, as opposed to the Mo5+ variant produced in DCB, when compared to the literature.[33]

Table 1

Rate Constants Determined under Different Conditions To Optimize the Reaction

solvent	temperature (°C)	pH	rate constant (s–1)
DCB/H2O	25	12	1.37 × 10–5
DCB/H2O	50	12	6.37 × 10–5
DCB/H2O	75	12	1.27 × 10–4
DCB/H2O	50	3	1.22 × 10–5
DCB/H2O	50	7	4.46 × 10–5
DCB	50	N/A	1.33 × 10–5

The rate constant was also calculated for the decomposition of MoL4 in a single-phase experiment (Figure S14) in DCB. Not only is the rate lower compared with a biphasic system, but the L/L interface facilitates the thin film formation. In the single phase case, MoS2 precipitates from the organic phase and aggregates together, creating thicker, less ordered MoS2 flakes. Using an Arrhenius plot, it was also possible to elucidate the activation energy of the reaction from data in Table (Figure S15).

Discussion

Raman spectroscopy showed that MoS2 flakes were formed at the L/L interface when Na2S and NaOH were present in the aqueous phase and without any additives in the aqueous phase. However, the rate of precursor decomposition and film formation was noticeably slower without the presence of Na2S or NaOH. As Na2S reacts with water in a two-step reaction to produce OH– ions, it is proposed that both the molybdenum and sulfur sources of the reaction come from the starting precursor and the alkalinity of water acting as a catalyst facilitates the reaction.

In addition to spectroscopic information, a Galvani potential difference was applied to determine if any charge-transfer reactions (e.g., ion transfer, heterogeneous electron transfer, or proton-coupled electron transfer (PCET)) were occurring at the interface between two immiscible electrolyte solutions (ITIESs).[46] The L/L electrochemical cell was prepared with and without sodium sulfide, and the cyclic voltammetry (CV) can be seen in Figure S16. Our CV experiments suggest that there is no evidence of sulfide transfer into the organic phase when a potential is applied, which was expected due to the hydrophobicity of the sulfide ion. Therefore, it can be concluded that the sulfide does not partition into the organic solvent, which is at odds with conclusions drawn in previous work.[25]

Analysis of the DCB solution post-reaction by electrospray mass spectrometry yielded peaks at mass-to-charge ratios of m/z 335 and m/z 319 peaks corresponding to the [M + K]+ and [M + Na]+ adducts of ionized disulfiram (DSF, C10H20N2S4; Figure S17). In addition, UV–visible absorption spectroscopy was used to confirm that the absorbance maxima of DSF at the predicted end concentration matched the absorbance of the organic solvent at the end of reaction (Figure S18). This suggests that disulfiram is a product of the reaction and that there may be two-electron transfer steps involved at some point within the reaction.

Base-catalyzed decomposition of dithiocarbamates has previously been proposed as a pathway to form the corresponding metal sulfide materials.[47,48] Within these proposals, two dithiocarbamates were suggested to produce DSF and additional electrons, which could be further oxidized by air to form an isothiocyanate and elemental sulfur. Hence, it is proposed that a similar reaction here would proceed to produce MoS2. The TGA profiles of MoL4 suggest that ethylene can dissociate easily from the dithiocarbamate. It is also suggested that the excess electrons produced from the formation of DSF could convert any oxidized metal centers to Mo4+. Therefore, we propose the above reaction scheme (Scheme ) for the formation of MoS2 at the L/L interface, which is similar to the mechanisms suggested for the decomposition of other molybdenum coordination complexes prior to this study.[29,31]

Scheme 1

Proposed Reaction Scheme for the Decomposition of MoL4 in Organic Solution To Form a Film of MoS2 at the L/L Interface

Conclusions

We have shown for the first time that a molybdenum coordination complex decomposes to produce a thin film of few-layer MoS2, which to the best of our knowledge is the first reported bottom-up room-temperature synthesis route to TMDs. These films form at the interface between two liquids, and the film formation can be optimized by changing the concentration, increasing the temperature, or increasing the alkalinity in the aqueous phase. Raman spectroscopy confirms the formation of MoS2, and AFM demonstrates a flake thickness as low as 4 nm. SEM images show a variety of different MoS2 formations including microstructures made from many nanoparticles and microparticles as large as 10 μm. pXRD and TEM results confirm that the films are crystalline and hexagonal in nature, and XPS gives insight into the surface state of the crystallites. In addition, a greater understanding of the nature of the reaction was achieved by applying a Galvani potential across the interface, and as a result, a potential reaction mechanism has been suggested based on the information gathered.

Overall, the liquid–liquid interface is a very exciting alternative to traditional methods for the production of 2D films. Films of nanometer thickness have been produced by confining their formation to the L/L interface. As a room-temperature production method, it could serve as a scalable route to 2D material production for use in a variety of applications such as energy storage devices and electronics. Larger flake sizes have been generated using this method than those from traditional liquid-phase exfoliation, and the production technique is a lower-cost alternative to CVD. We have demonstrated this method for MoS2; however, this technique should be generic and therefore transferrable to other 2D materials.