Covalent Bisfunctionalization of Two-Dimensional Molybdenum Disulfide.

Covalent functionalization of two-dimensional molybdenum disulfide (2D MoS2 ) holds great promise in developing robust organic-MoS2 hybrid structures. Herein, for the first time, we demonstrate an approach to building up a bisfunctionalized MoS2 hybrid structure through successively reacting activated MoS2 with alkyl iodide and aryl diazonium salts. This approach can be utilized to modify both colloidal and substrate supported MoS2 nanosheets. We have discovered that compared to the adducts formed through the reactions of MoS2 with diazonium salts, those formed through the reactions of MoS2 with alkyl iodides display higher reactivity towards further reactions with electrophiles. We are convinced that our systematic study on the formation and reactivity of covalently functionalized MoS2 hybrids will provide some practical guidance on multi-angle tailoring of the properties of 2D MoS2 for various potential applications.

Introduction

Two‐dimensional transition metal dichalcogenides (2D TMDs) are in the spotlight of nanomaterial community in the last decade since they have shown great potential in a variety of areas such as electronic/optoelectronic devices,[ , , ] catalysis,[ , , ] sensing,[ , , ] and biomedical applications.[ , , ] A lot of research projects has been concentrated on the development of methods to controllably produce TMD nanosheets.[ , , , , ] These methods enable the production of 2D TMDs with customizable size, thickness, and spatial orientations, encouraging the deep investigation of fundamental physics and chemistry of TMD nanosheets. In the meantime, the research topics on chemical functionalization of 2D TMDs, through either covalent or non‐covalent approaches, have flourished.[ , , , , ] This has enabled further tailoring of properties, control over processability, and improved versatility of TMD based material applications, facilitating the fulfillment of the potential of 2D TMDs.

As a prototype of 2D TMDs, 2D MoS2 is extensively used as a model system to explore the functionalization chemistry of TMDs. Compared to non‐covalent functionalization, covalent functionalization shows particular advantages in terms of the robustness of formed conjugates, the efficiency of electronic communication at the functional groups/MoS2 interface, and the versatility endowed by the variety of functional groups.[ , , ] Recently, Knirsch et al. and Voiry et al. have demonstrated protocols to covalently functionalize chemically exfoliated MoS2 (ce‐MoS2) nanosheets using either diazonium salts or organohalides (‐I or ‐Br). By virtue of these two approaches, a library of covalently functionalized MoS2‐organic/inorganic hybrid structures have been generated.[ , , ] Among all of these reported covalent hybrids, almost all of the materials are monofunctionalized, in which only one type of functional group is grafted onto MoS2. One step further, the incorporation of multiple and diverse functionalities into 2D MoS2 would enable the integration of multiple functionalities of different surface addends into one device, opening the opportunity to achieve a smart nanomachine with tailor‐made properties and integrated functions. In this respect, the development of such a functionalized MoS2 hybrid structure bearing two or multiple covalently tethered functionalities is highly desired.

Additionally, we note that the reactions of MoS2 nanosheets with organohalides have shown some distinct phenomena compared to those with diazonium salts. For example, the functionalization of 2D MoS2 with organohalides has only been achieved so far using 1T‐phase ce‐MoS2 nanosheets but not 2H‐MoS2 (unless with a catalyst), whereas the arylation of 2D MoS2 using diazonium salts can be realized by either 1T‐ phase ce‐MoS2 or 2H‐phase MoS2 nanosheets (prepared by mechanical exfoliation or chemical vapor deposition). Secondly, the treatment of negatively charged 1T‐phase ce‐MoS2 with organohalides leads to partial charge neutralization, while the treatment of ce‐MoS2 with diazonium salts leads to complete charge neutralization. Many aspects regarding the chemistry of covalent functionalization of MoS2 nanosheets using these two types of electrophiles are not fully understood yet.

In this study, we demonstrate an approach to construct a bisfunctionalized MoS2 hybrid structure bearing both alkyl and aryl groups. These two types of functional groups were covalently tethered onto MoS2 through the successive reaction of chemically activated MoS2 with alkyl iodides and aryl diazonium salts. By adjusting the functionalization sequence and carefully analyzing the reaction intermediates and final adducts using a series of spectroscopic and microscopic techniques, we shed light on the fundamental aspects of covalent functionalization chemistry of 2D MoS2.

Results and Discussion

Synthesis of bisfunctionalized MoS2

The bisfunctionalized 2D MoS2 was synthesized using a stepwise approach, starting with the preparation of MoS2 nanosheets via chemical exfoliation. Specifically, bulk MoS2 powder was initially reacted with n‐BuLi followed by the hydration and sonication of the intercalated compound in water, as detailed in previous literature.[ , ] Then, the freshly prepared and purified ce‐MoS2 was reacted with iodohexane (electrophile A) in a manner analogous to the reported procedure, giving the alkylated intermediate product A‐MoS2. After purification, A‐MoS2 was re‐dispersed in isopropanol and reacted with 4‐bromobenzenediazonium tetrafluoroborate (electrophile B) to form the final product AB‐MoS2. Alternatively, the two‐step functionalization was also performed by following the reverse order, in which ce‐MoS2 was firstly reacted with 4‐bromobenzenediazonium tetrafluoroborate (electrophile B), yielding the arylated intermediate product B‐MoS2. The subsequent reaction of B‐MoS2 with iodohexane (electrophile A) gave the product BA‐MoS2 (Scheme 1, see the Supporting Information for details).

Scheme 1

Illustration of the functionalization sequence I and II. Reagent A: iodohexane. Reagent B: 4‐bromobenzenediazonium tetrafluoroborate. Sequence I, alkylation followed by arylation, leads to the formation of bisfunctionalized adduct AB‐MoS2. Sequence II, arylation followed by alkylation, can only give the monofunctionalized product BA‐MoS2. The structural differences between B‐MoS2 and BA‐MoS2 are illustrated in the following context.

Characterization

To identify the nature and the content of functional groups, the as‐prepared ce‐MoS2, A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2 were analyzed by thermogravimetric analysis coupled with mass spectrometry (TGA‐MS). ce‐MoS2 (Figure S1a) displays a gradual thermal decomposition starting from 200 °C with a total mass loss of 5 % at T=700 °C, which was attributed to the degradation of MoS2 itself. In comparison, TGA‐MS profile of A‐MoS2 (Figure S1b) shows a total mass loss of 8 % over the same temperature range with the maximum ion current detected at 275 °C. The major gaseous products were identified with the m/z of 55, 43, and 41, corresponding to the hexyl group related fragments C4H7 +, C3H7 +, and C3H5 +, respectively, suggesting that hexyl groups are present in A‐MoS2. Thermolysis of B‐MoS2 (Figure S1c) reveals a mass loss of 20 % with the maximum ion current detected at 410 °C. The dominant ion currents are found to be the fragments C6H5 + (m/z=77) and C6H6 (m/z=78) associated with the aryl moiety, suggesting the presence of phenyl groups in B‐MoS2.

Interestingly, the final hybrid product from sequence I, AB‐MoS2 (Figure 1 a), exhibits a mass loss of 19 %, much higher than that of the intermediate product A‐MoS2 (8 %), suggesting that a higher content of functional groups are present in AB‐MoS2 compared to A‐MoS2. The major gaseous products from thermal decomposition of AB‐MoS2 were detected with m/z=43 and 77, corresponding to the characteristic fragments of hexyl group and phenyl group, respectively. This result demonstrates the successful introduction of both hexyl and phenyl groups in AB‐MoS2. In contrast, the final product from sequence II, BA‐MoS2 (Figure 1 b), shows a total mass loss of 17 %, slightly lower than that of the intermediate product B‐MoS2 (20 %). In addition, the dominant ion current was detected as the phenyl group related fragment C6H5 + (m/z=77), suggesting that phenyl groups remain present in BA‐MoS2. However, the hexyl group related fragment (C3H5 +, m/z=43) was detected in a small amount and at low temperature region. This implies that only a trace amount of hexyl group related species is present in BA‐MoS2, which was presumably physisorbed on the MoS2 surface.

Figure 1

TGA‐MS profiles (black) of AB‐MoS2 (a) and BA‐MoS2 (b). The ion currents in AB‐MoS2 and BA‐MoS2 are the fragment C3H7 + (m/z 43, orange trace) derived from the decomposed hexyl group and the fragment C6H5 + (m/z 77, cyan trace) associated with the aryl group. The small amount of fragment C3H7 + detected in BA‐MoS2 at low temperature range originates from the detachment of physisorbed alkyl groups.

The Fourier‐transform infrared spectroscopy (FT‐IR) measurement of AB‐MoS2 (Figure S2, red trace) demonstrates strong absorption peaks in the range of 3000–2840 cm−1 and 1760–1000 cm−1, which are attributed to vibrational features of alkyl chain and aromatic ring, respectively; whereas BA‐MoS2 (Figure S2, blue trace) shows almost identical features to the arylated intermediate B‐MoS2 (Figure S2, dark cyan trace). The alkyl chain related peaks with very low intensity were discernable in the FT‐IR spectra of BA‐MoS2 (Figure S2, blue trace), which were presumably derived from the physisorbed alkyl moieties. This is in a good agreement with TGA‐MS results.

To further investigate the chemical state and quantify the elemental composition of ce‐MoS2 and functionalized materials, high resolution X‐ray photoelectron spectroscopy (XPS) analysis was carried out. The elemental analysis of survey spectra (see Table S1 for the detailed elemental composition) shows the presence of Br (Br 3d, 70 eV) in B‐MoS2, AB‐MoS2, and BA‐MoS2, suggesting the successful incorporation of bromophenyl functional groups in these three samples. No N related signals were detected in functionalized MoS2, confirming that the unreacted 4‐bromobenzenediazonium salts are not present in the functionalized samples and the functionalization reaction likely proceeds via the substitution of the diazo groups. Further comparison of BA‐MoS2 with B‐MoS2 reveals a negligible change of C and Br content (Table S2), suggesting that the degree of functionalization is likely the same. Additionally, the atomic percentages of C and Br were found to significantly increase in AB‐MoS2 relative to A‐MoS2, indicating the successful introduction of bromophenyl group in AB‐MoS2 through the second step functionalization. This coincides with our TGA‐MS results.

Fitting the Mo 3d core level spectra by deconvolution allows us to determine the concentration of 1T‐ and 2H‐phase in each sample. The Mo 3d core level spectra of ce‐MoS2 and functionalized MoS2 can be fitted with three sets of doublets, which are attributed to the Mo 3d3/2 and Mo 3d5/2 components of 2H‐MoS2, 1T‐MoS2, and MoO3 (Figure 2 a–e). The content of 1T‐ phase was found to be 19 % in ce‐MoS2, whereas this was approximately 60–70 % in the functionalized MoS2. The lower concentration of 1T‐phase in ce‐MoS2 can be ascribed to aging effects during the sample delivery and storage, which has been proved to cause phase rearrangement from 1T to 2H.[ , , ] In comparison, the large proportion of 1T‐phase in the functionalized MoS2 suggests that the 1T‐phase was largely preserved after functionalization. Furthermore, the continuing functionalization of the intermediate did not alter the concentration of 1T‐phase, which is evidenced by the similar content of 1T‐phase in AB‐MoS2 relative to A‐MoS2 and likewise in BA‐MoS2 relative to B‐MoS2 (Table S2, content of 1T‐phase in all the samples). This is important, because it demonstrates that the concentration of 1T‐phase can be stabilized against the conversion back to 2H‐phase through functionalization.

Figure 2

XPS core level spectra of Mo 3d (a–e) and S 2p (f–j) for ce‐MoS2, A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2. The orange component in (a–e) can be attributed to the S 2s peak, which heavily overlaps with Mo 3d components. The green components in (f) and (g) can be assigned to oxidized sulfur species.

We performed DFT calculations (see supporting information for details) in order to get a better understanding of the phase composition in functionalized MoS2. Starting with 2×2 1T‐MoS2 and 2H‐MoS2 unit cells and placing one functional group 2 Å above the MoS2 layer, this results in a 25 % coverage of one side of the layer. For 1T‐MoS2 (octahedral coordination), the calculations indicate a phase transition to a distorted octahedral coordination (also referred to as 1T′‐phase). This suggests that the 1T‐phase, which is thermodynamically metastable, is stabilized upon covalent functionalization, and the slight lattice distortion helps to maintain the octahedral coordination in functionalized MoS2 structures. This result corroborates our observations in XPS measurements. For 2H‐MoS2 (trigonal prismatic coordination), the structure is preserved when adding the functional group. These effects are observed independently of the type of attached functional groups.

To determine the degree of functionalization, we closely inspected the S 2p core level spectra. In addition to the typical doublet feature (yellow components, Figure 2 f–j) corresponding to the S 2p3/2 (161.9 eV) and S 2p1/2 (163.1 eV) components of MoS2, the S 2p spectra of A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2 display a shoulder at 164.45±0.02 eV (purple components, Figure 2 f–j), which is usually assigned to the C‐S species derived from covalent functionalization.[ , ] The degree of covalent functionalization (functional groups per S), calculated based on the peak area underneath of this C‐S component relative to the total area of S 2p, was found to be 24 %, 32 %, 31 %, and 31 % for A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2, respectively.

Compared to A‐MoS2, AB‐MoS2 shows a higher degree of functionalization, which is consistent with the TGA‐MS analysis. The higher content of functional groups in AB‐MoS2 relative to A‐MoS2 is attributed to the successful introduction of aryl functional groups through the second step functionalization which is further seen based on the Br concentration in the survey spectra (Table S1). In comparison, BA‐MoS2 displays almost same degree of functionalization to B‐MoS2. This is consistent with the observation in XPS survey analysis. Interestingly, the maximum degree of functionalization after two‐step successive functionalization reaches about 30 % regardless of the reaction sequences. In addition, the functionalization of activated MoS2 using organoiodides leads to a lower degree of functionalization compared to that using diazonium salts, indicating that the latter is more efficient.

For further exploration of the chemical identity and the binding sites of functional groups, A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2 were investigated using solid state 13C (126 MHz) magic angle spinning nuclear magnetic resonance spectroscopy (MAS NMR, Figure 3 and S3). The 13C NMR spectrum of iodohexane (Figure S3a) displays six singlet signals in the range of 5–35 ppm, corresponding to the aliphatic carbons of hexyl chain. In particular, the characteristic signal at δ=7.27 ppm corresponds to the α‐C connected to iodine. In comparison, A‐MoS2 (Figure S3b) shows three sets of widened signals in the range of 10–35 ppm, demonstrating the presence of aliphatic carbons associated with the hexyl chain. The dramatically broadening of signals in the solid state sample probably hinges on the anisotropy, orientation or magnetic susceptibility effects arising from the charged nanosheets.[ , ] Most importantly, the signal related to the α‐C is absent from the original chemical shift region, suggesting that this α‐C was involved in the functionalization reaction between iodohexane and MoS2. The α‐C related peak is shifted downfield and thus, jointly with other aliphatic carbons, constitute the prominent signals in A‐MoS2. This downfield shift of α‐C also suggests a de‐shielding effect, which was presumably caused by the formation of bonds between α‐C atoms and more electronegative atoms (in this case the sulfur atoms) through coupling iodohexane to MoS2. The formation of C−S bonds is in accord with our results in the XPS S2p core level spectra, corroborating that hexyl groups were tethered to the S atoms of MoS2 in A‐MoS2. Similar observations have been reported in other alkyl functionalized MoS2 systems.

Figure 3

13C solid‐state NMR spectra of A‐MoS2 (a), B‐MoS2 (b), AB‐MoS2 (c), and BA‐MoS2 (d). All the samples were spun at 14 kHz except for BA‐MoS2, which was spun at 10 kHz.

Analogous analysis was also performed on 4‐bromothiophenol reference (Figure S3c) and B‐MoS2 (Figure S3d). In this case, the characteristic signal corresponding to the aromatic carbons of the bromophenyl group was observed in B‐MoS2, verifying the presence of these groups. The resonances of aromatic carbons in B‐MoS2 are centered at the similar chemical shift to that of bromothiophenol, suggesting that a similar electronic environment of aromatic ring is maintained by the functional groups in B‐MoS2 relative to bromothiophenol. We only observed a slight downfield shift of δ‐C (‐CHaromatic‐Br) in B‐MoS2, which was likely due to the variation of C‐S bond strength at the interface of MoS2 and bromophenyl functional groups compared to that between thiol group (SH) and bromophenyl moiety in bromothiophenol. Nevertheless, the solid state NMR analysis confirmed the covalent attachment of alkyl and aryl functional groups in A‐MoS2 and B‐MoS2, respectively.

Further comparison of the 13C solid state NMR spectra for A‐MoS2, B‐MoS2, and AB‐MoS2 (Figure 3 a–c) reveals the characteristic shifts corresponding to the aliphatic and aromatic carbons, demonstrating that both hexyl and bromophenyl functional groups were covalently bound to the S‐atoms of MoS2 in AB‐MoS2. In contrast, BA‐MoS2 (Figure 3 d) shows similar features to B‐MoS2, where the major signal is derived from the covalently tethered aromatic carbon moieties. It is noted that a few peaks with very low intensities can be distinguished, which show characteristic features of iodohexane, for example, α‐C (‐CH2‐I) was detected, but not the covalently tethered hexyl chain, suggesting the presence of trace amounts of physisorbed iodohexane in BA‐MoS2, in good agreement with our TGA‐MS and FT‐IR results.

The morphology of MoS2 nanosheets before and after functionalization was examined using scanning electron microscopy (SEM) at different magnification scales. The SEM images (Figure 4 a and f) of ce‐MoS2 show characteristic layered structures with diverse lateral sizes and well‐defined edges. The features of the nanoflakes are well preserved in A‐MoS2 (Figure 4 b and g) with relatively more disrupted edges compared to ce‐MoS2. This can be a sign of chemical modification via the reaction with iodohexane. The morphology of B‐MoS2 (Figure 4 c and h) is strikingly different from either ce‐MoS2 or A‐MoS2, where instead of a layered structure, amorphous chunks were observed. This is likely attributed to the formation of oligomer during the functionalization reaction, in which the excess aryl radicals can either attach to the surface tethered aryl groups, forming the oligomer layers, or react with each other via biaryl coupling to form aryl oligomers prior to covalently bonding to the MoS2 surface. These side reactions have been reported in the reactions between diazonium salts and graphene.[ , ] The dramatic change in morphology of B‐MoS2 indicates a transformative change in the surface properties upon functionalization using diazonium salts. Despite the fact that covalent functionalization of MoS2 nanosheets using diazonium salts has been extensively applied in recent years,[ , , ] this is the first time that covalent functionalization induced morphology change is reported and outlined. Our finding can provide a new perspective to fundamentally understand the dramatic variation in the properties of materials functionalized using diazonium chemistry, when compared to their parental MoS2 nanosheets. This effect can also be distinctly observed when comparing the morphology of intermediate product A‐MoS2 (Figure 4 b and g) with bisfunctionalized product AB‐MoS2 (Figure 4 d and i). The layered structure shown in A‐MoS2 was transformed into a fragmented and petal‐like structure in AB‐MoS2 after the functionalization with diazonium salts.

Figure 4

SEM images of ce‐MoS2 (a), A‐MoS2 (b), B‐MoS2 c), AB‐MoS2 (d), and BA‐MoS2 (e) and corresponding views (f–j) under higher magnification.

Further inspection of the SEM images of BA‐MoS2 (Figure 5 e and j) reveals features of the nanoflakes with much smaller lateral sizes compared to ce‐MoS2. Our previous TGA‐MS, FT‐IR and XPS results have indicated that the second‐step reaction of B‐MoS2 with 1‐iodohexane failed to covalently attach alkyl groups to MoS2. The degree of covalent functionalization remained same in B‐MoS2 and BA‐MoS2. We then deduced that the recovery of layered structure in BA‐MoS2 was possibly attributed to the removal of amorphous oligomer layers in highly functionalized B‐MoS2.

Figure 5

Extinction spectra of ce‐MoS2 (grey), A‐MoS2 (orange), B‐MoS2 (cyan), AB‐MoS2 (red), and BA‐MoS2 (blue). The dashed lines indicate the bands discussed in the following context. The spectra were normalized to the maximum intensity for comparison.

We then performed energy dispersive X‐ray (EDS) mapping of all of the samples (Table S3 for detailed elemental composition). In addition to the uniformly distributed Mo and S signals, the Br signal was detected in B‐MoS2, AB‐MoS2, and BA‐MoS2, verifying the presence of bromophenyl functional groups in these samples. Moreover, the atomic percentage of Br slightly dropped in BA‐MoS2 compared to B‐MoS2, which could be due to the removal of Br‐containing moieties (e.g. oligomers) during the second step treatment. These observations show good agreement with our XPS results.

Optical properties

Having determined the chemical nature of the functionalized MoS2, we then set out to monitor how the functionalization affects the optical properties of MoS2 nanosheets. The extinction spectra (Figure 5) of ce‐MoS2 and functionalized MoS2 display intense bands in the ultra‐violet region except for B‐MoS2 (dark cyan trace in Figure 5), in which the excitonic transitions were virtually quenched. This is possibly due to the complete change in morphology of B‐MoS2 in comparison to the others. Interestingly, the excitonic transition bands were recovered in BA‐MoS2 (blue trace in Figure 5), which was not surprising as the recuperation of MoS2 layered structure after the second step reaction was observed in our SEM study (Figure 5 e and j). Similar observations have been reported for other MoS2 based materials, which were functionalized using diazonium chemistry as well.

To closely inspect the peak positions of the bands in the ultra‐violet region, we plotted the second derivative of extinction spectra (Figure S4a). Three distinct bands can be identified in ce‐MoS2: band (I) at 246 nm, band (II) at 317 nm, and band (III) at 417 nm. These bands are all preserved in A‐MoS2 and AB‐MoS2, while only band (II) and (III) are present in B‐MoS2 and BA‐MoS2. The disappearance of band (I) in B‐MoS2 and BA‐MoS2 demonstrates the dramatic change in electronic properties in these two samples. The position of band (III) is located in the visible region and close to where C exciton of MoS2 is, thus is indicative of the presence of 2H‐phase in all samples, which is consistent with our XPS observations. In addition, band (III) is up‐shifted in all functionalized samples compared to ce‐MoS2, which could be ascribed to increased scattering effects arising from re‐aggregation.[ , , ] Interestingly, in comparison to ce‐MoS2, we observed down‐shift of band (II) in A‐MoS2, AB‐MoS2, and BA‐MoS2, but up‐shift in B‐MoS2. The morphology of B‐MoS2 is totally different from the others, which brings some complexity to analyzing the change of electronic bands of B‐MoS2. Herein, we will only discuss the shift of band (II) with respect to the other samples (Figure S4b, red plot). The magnitude of the down‐shift follows the order: A‐MoS2 < AB‐MoS2 < BA‐MoS2, indicating that the more bromophenyl groups tethered on the surface, the more significant the change imparted onto the electronic structure of MoS2.

To further elucidate the influence of covalent functionalization on the optical properties of MoS2, the dispersions of ce‐MoS2, A‐MoS2, B‐MoS2, AB‐MoS2, and BA‐MoS2 were deposited on Si/SiO2 wafers via spin‐coating to form thin films and measured by scanning Raman spectroscopy under resonant excitation (λ=633 nm) at room temperature (Figure 6 a). The average spectrum calculated from 121 recorded single point spectra was analyzed for each sample. The Raman spectrum of ce‐MoS2 shows characteristic peaks of MoS2 at 377, 406, and 450 cm−1, corresponding to E1 2g, A1g, and 2LA(M), respectively. In addition, a series of features at lower frequencies was detected, among which the features at 150, 223 and 326 cm−1 are assigned to 1T‐phase related J1–J3 modes, and the peak at 188 cm−1 is identified as the two‐phonon difference combination mode A1g(M)‐LA(M). The Raman spectra of the functionalized samples display almost all of the key features of MoS2 with decreasing peak intensity and an increase in peak widths compared to ce‐MoS2, suggesting reduced MoS2 crystallinity after functionalization. There is no significant shift of the frequencies of characteristic Raman modes of MoS2 after functionalization. In addition, the J1 and J2 modes are still present in the functionalized samples, indicating the presence of at least partial 1T‐ or 1T′‐phase in the functionalized samples. These observations are consistent with our XPS, UV/Vis, and DFT calculation results.

Figure 6

Raman spectra a) of ce‐MoS2 (grey), A‐MoS2 (orange), B‐MoS2 (cyan), AB‐MoS2 (red), and BA‐MoS2 (blue): the arrows show the characteristic Raman scattering of ce‐MoS2; the dashed block highlights the emerging peak at 210 cm−1, the peaks marked with an asterisk are Raman modes of the underlying Si/SiO2 substrate. The sharp peak at around 270 cm−1 in ce‐MoS2 and A‐MoS2 is a spike signal. Plot of the intensity of the phonon mode at 210 cm−1 normalized to the E1 2g mode (I 210/I 377) as a function of the degree of functionalization (b). Plot of the intensity of the phonon mode at 210 cm−1 normalized to the A1g mode (I 210/I 406) as a function of the degree of functionalization (c). The dashed lines in (b) and (c) are guides for the eyes. For the intensity ratios shown here, the amplitudes of the Raman peaks were used.

Noticeably, a phonon mode at around 210 cm−1 strongly increases in intensity in all the functionalized MoS2 samples (Figure 6 a). This peak has been observed in many functionalized MoS2 structures,[ , , ] yet it has not been investigated in most cases. The position of this peak is very close to the first‐order scattering of the LA phonon, which is usually activated by structural disorder such as sulfur vacancies and surface adatoms.[ , , , ] A study reported by Roy et al. has pointed out that for single layer MoS2, the peak intensity of LA mode relative to one of first‐order Raman modes (E1 2g or A1g), is proportional to the density of defects. Inspired by their study, we plotted the intensity of this emerging peak at 210 cm−1 relative to the characteristic E1 2g mode (I 210/I 377) and A1g mode (I 210/I 406), as a function of the degree of functionalization (Figure 6 b and 7 c). The intensity ratios, I 210/I 377 and I 210/I 406, increase with increasing degree of covalent functionalization. This correlation indicates that the activation of the phonon mode at 210 cm−1 is associated with functionalization, i.e., the 210 cm−1 mode may be used as an indicator to estimate the degree of covalent functionalization of MoS2.

Figure 7

AFM images of pristine me‐MoS2@Si/SiO2 (a), A‐MoS2@Si/SiO2 (b), B‐MoS2@Si/SiO2 (c), AB‐MoS2@Si/SiO2 (d), and BA‐MoS2@Si/SiO2 (e) and the corresponding height profiles (f). The height profiles along the red marker in each image were plotted.

We also performed temperature dependent Raman measurements on AB‐MoS2 and BA‐MoS2 powders to monitor the Raman spectra change over the course of the thermolysis of samples. The sample powders were placed on Si/SiO2 wafers, flattened by glass slides and then subjected to Raman measurements under inert atmosphere. For AB‐MoS2 (Figure S5a), with increasing the temperature, new peaks at 817, 283, and 119 cm−1 gradually appear, indicating the formation of MoO3. When the temperature rises above 400 °C, the E1 2g and A1g modes of MoS2 completely vanish, leaving the spectrum dominated by the features of MoO3, suggesting a complete structural transformation upon annealing. Interestingly, the phonon mode at 210 cm−1 is preserved over the entire thermolysis process. The intensity ratio of the 210 cm−1 mode relative to the characteristic E1 2g mode (I 210/I 377) and A1g mode (I 210/I 406), against temperature (Figure S5b and S5c) reveals two temperature regions: (1) for T<300 °C, the intensity ratios slightly increase till reaching a local maximum at around T=200 °C, and then decrease to the original value. (2) For T>300 °C, the intensity ratios increase. Analogous results were also obtained for BA‐MoS2 (Figure S6). Several events, such as the detachment of surface tethered functional groups, phase transition from 1T to 2H‐MoS2, and the oxidation of MoS2 to MoO3 may take place simultaneously or subsequently with increasing temperature. A detailed study on how these events contribute to the evolution of the Raman spectra will provide a better understanding, however is beyond the scope of this work.

Functionalization of mechanically exfoliated MoS2

We have shown above that the optical properties of MoS2 can be fine‐tuned in bulk through step‐wise covalent functionalization where both sides of the MoS2 nanoflakes are accessible to electrophiles during the process. Given such a possibility of constructing a bisfunctionalized MoS2 hybrid structure, we have examined whether this strategy can be utilized on substrate supported MoS2 surfaces. To this end, we have used mechanically exfoliated MoS2 (me‐MoS2) deposited on a Si/SiO2 substrate. Pristine me‐MoS2 on Si/SiO2 was initially activated by immersing the wafer in n‐BuLi under inert atmosphere. After rinsed with anhydrous hexane, the activated MoS2 wafer was soaked in the solution of the first electrophile (under inert conditions) followed by a thorough wash with organic solvents and water to give the monofunctionalized MoS2. This wafer was subsequently treated with the second electrophile without further activation, yielding the final adducts (see Supporting Information for details). The topographic information and the thickness profiles of pristine me‐MoS2@Si/SiO2, A‐MoS2@Si/SiO2, B‐MoS2@Si/SiO2, AB‐MoS2@Si/SiO2, and BA‐MoS2@Si/SiO2 flakes were monitored using atomic force microscopy (AFM, Figure 7). The AFM image of pristine me‐MoS2 (Figure 7 a) shows a very smooth and flat topography with a lateral dimension of about 2–6 μm, and a thickness of about 4–8 nm, corresponding to 4–8 layers. In comparison, the AFM images of functionalized MoS2 display a rough and corrugated topography. In particular, the protrusions ranging from 10–20 nm are clearly observed in the AFM images of B‐MoS2@Si/SiO2, AB‐MoS2@Si/SiO2, and BA‐MoS2@Si/SiO2. The mechanism of the formation of protrusions on MoS2 nanosheets functionalized using diazonium salts has been discussed in detail in another study, where the generation of chain‐like protrusions in those samples can be indicative of the covalent attachment of aryl functional groups via diazonium chemistry. Comparing the AFM images of B‐MoS2@Si/SiO2 and BA‐MoS2@Si/SiO2 reveals significant fragmentation of nano‐flakes in BA‐MoS2@Si/SiO2 (Figure 7 e). This phenomenon was also observed in the SEM image of bulk functionalized sample (Figure 5 e, BA‐MoS2). On the basis of these results, we urge the caution on maintaining an intact MoS2 crystal structure while seeking a high degree of functionalization in future studies. The Raman spectra of pristine me‐MoS2@Si/SiO2, A‐MoS2@Si/SiO2, B‐MoS2@Si/SiO2, AB‐MoS2@Si/SiO2, and BA‐MoS2@Si/SiO2 were recorded straightforward under ambient conditions, while the Raman spectrum of activated MoS2 was recorded under inert atmosphere (Figure S7). Importantly, the Raman mode at 210 cm−1, which we have previously ascribed to a functionalization related phonon mode, was also detected in all the functionalized samples, confirming the successful covalent functionalization of me‐MoS2.

The key to form bisfunctionalized MoS2

We have thoroughly characterized the functionalized intermediates and final adducts from two functionalization sequences. TGA‐MS, FT‐IR, and solid state NMR have shown the presence of hexyl and bromophenyl groups in AB‐MoS2 but predominantly the bromophenyl group in BA‐MoS2. XPS has proved that the functional groups were covalently tethered on the surface of MoS2 in all the functionalized samples. The successful covalent modification is also reflected by the prominent feature at 210 cm−1 in Raman spectra of functionalized samples (both chemically and mechanically exfoliated MoS2). SEM‐EDS corroborates that not only the chemical composition, but also the morphology of AB‐MoS2 is different from BA‐MoS2. These results collectively demonstrate that varying the covalent functionalization sequence can result in different hybrid structures. The bisfunctionalized structure can only be achieved through sequence I, alkylation followed by arylation.

To explain why the sequence II, arylation followed by alkylation, cannot lead to the formation of bisfunctionalized product, we set out to evaluate the difference of the number of available binding sites and the reactivity in the two intermediate products: A‐MoS2 and B‐MoS2. Our XPS results have shown that the maximum surface coverage (the degree of covalent functionalization) is approximately 30 % for both mono‐ and bisfunctionalized products. In this regard, A‐MoS2 (degree of functionalization=21 %) is still rich in binding sites, whereas in B‐MoS2 (degree of functionalization=32 %), the binding sites are already saturated. Therefore, further introducing hexyl groups onto surface of B‐MoS2 to form the bisfunctionalized product would require the replacement of some surface tethered bromophenyl groups (B‐) with hexyl groups (A‐). However, the possibility to displace the bromophenyl groups in B‐MoS2 with hexyl groups without any other activation treatment seems fairly low. This is in line with preliminary results from our DFT calculation, where the calculated surface bonding distance between functional groups and MoS2 basal plane in B‐MoS2 is smaller than that in A‐MoS2. Thus it is reasonable to assume that the bromophenyl groups are more strongly bound to MoS2 than the hexyl groups.

Secondly, the alkylation of MoS2 using alkyl iodides requires negative charge on nanoflakes, however, our previous study and the reference experiment (Figure S8) have proven that the negative charges of ce‐MoS2 were completely quenched after the functionalization with diazonium salts. To prove the lack of negative charge is the limiting factor for the second step covalent functionalization of B‐MoS2, we performed a control reaction by reacting re‐charged B‐MoS2 with 1‐iodohexane. To this end, the absolutely dried B‐MoS2 was initially reduced using Na/K alloy, then reacted with 1‐iodohexane, yielding the product B‐ac‐A‐MoS2 (“ac” denotes activation, see Supporting Information for details). The TGA‐MS profile of B‐ac‐A‐MoS2 (Figure S9) showed the major fragments associated with hexyl groups and phenyl groups, indicating that the both functional groups were attached in B‐ac‐A‐MoS2. Comparing the preparation method of B‐ac‐A‐MoS2 with that of BA‐MoS2 revealed the reductive activation of B‐MoS2 with Na/K alloy was the crucial step to ensure the second step reaction with 1‐iodohexane and the formation of bisfunctionalized product. Therefore, the lack of negative charges in B‐MoS2 was likely the key reason that impeded the second step addition in the original sequence II.

Conclusion

In conclusion, we have demonstrated an approach to build up a bisfunctionalized MoS2 hybrid structure through successively reacting activated MoS2 with alkyl iodides and diazonium salts. This approach can be utilized to modify both colloidal and substrate supported MoS2 nanosheets. We have discovered distinct reactivity of MoS2 functionalized using alkyl iodide and diazonium salt and only the former allows for subsequent reactions with other electrophiles. By adjusting the reaction sequence and carefully analyzing the reaction intermediates and final adducts, we are able to obtain a series of functionalized MoS2 with varied degrees of functionalization. Raman spectra show a relation between the degree of functionalization and the relative peak intensity of a phonon mode at 210 cm−1, which can potentially serve as an indicator to estimate the degree of functionalization. We anticipate that our systematic study on the formation and reactivity of covalently functionalized MoS2 hybrids would encourage the development of new nanomaterials endowed with multiple and customizable functions.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Click here for additional data file.