Chemical vapor deposition of clean and pure MoS₂ crystals by the inhibition of MoO_3−x intermediates

Abstract

Molybdenum disulfide (MoS₂) synthesized via chemical vapor deposition (CVD) is commonly accompanied by some intermediate products in the form of MoO_3−x, resulting in MoS₂–MoO_3−x hybrids with diverse structures. In this study, we show that through the rational design of heating evaporation rate of the precursors of sulfur and MoO₃ separately in a CVD system with double temperature zones, the intermediate product of MoO_3−x involved in the CVD growth of MOS₂ (generally in the form of quadrilateral MoO₂/MoS₂ composites and/or stacks of quadrilateral MoO₂/triangular MoS₂) can be inhibited largely. On this basis, the appropriate sulfur sublimation versus the evaporated MoO₃ powder is revealed to be the key factor for the inhibition of intermediates, obtaining high-quality MoS₂ crystals with as high as >90% cleanness/purity. This study will assist our understanding of the fundamental CVD growth mechanisms of transition metal dichalcogenides (TMDCs) and provides a facile direction for the controlled synthesis of high-quality MoS₂ and other TMDCs with or without intermediate hybrids for applications.

---

1 Introduction

As a typical transition metal dichalcogenide (TMDC), layered MoS₂ has attracted extensive attention due to its unique material properties and various potential applications in nanoelectronics^1–5 and optoelectronics^6,7etc. Simultaneously, numerous efforts have also been devoted to the synthesis of MoS₂ films, notably via chemical vapor deposition (CVD),^8–13 which can produce large-area MoS₂ layers with decent quality as a counterpart with respect to the mechanical exfoliation method. In general, the CVD process employed for the synthesis of MoS₂ involve the use of solid sulfur and molybdenum oxide powder (e.g., MoO₃) as the precursors, and it has been shown that controlling the reaction conditions in the CVD growth is of significance for obtaining large-area and high-quality MoS₂ layers.^14,15 There have been many attempts to control the growth processes by changing the relative position of the growth substrates and the precursor,^14,16 or by regulating the carrier gas flow^17,18 and growth temperature.¹⁹ Despite the enormously experimental progress, some processes involved in the formation of clean and pure MoS₂ layers during such a CVD reaction between the precursors is still unclear. For instance, there remains a lack of mechanistic understanding of the factors that control the sulfurized purity and micro/nano-hybrid structures of the grown MoS₂. Indeed, it is a common problem with the CVD growth of MoS₂ that the finally grown products would be accompanied by some contaminants or intermediates of MoO_3−x, thus making the occurrence of both pure MoS₂ layers and hybrid MoS₂–MoO_3−x structures (such as core–shell structured MoO₂–MoS₂).^19,20 Hence, it is highly desirable to reveal the mechanisms on how to controllably grow clean and pure MoS₂ crystals, and MoS₂–MoO_3−x hybrids, respectively, for different synthesis target.

Here, in this study, we present an optimized growth direction for synthesizing uniform, pure and clean MoS₂ crystals in large scale by the inhibition of the intermediate products of MoO_3−x involved in the CVD growth of MoS₂. Through the improvement of the heating sublimation rate of sulfur at a certain evaporated MoO₃ atmosphere near the growth substrates, the formation of composites having mixed structures of MoO_3−x–MoS₂ (mainly MoO₂–MoS₂) could be inhibited largely, and then pure and clean MoS₂ crystals in large scale were obtained on the Si/SiO₂ substrates. Further, their electrical performance was measured, which showed a decent field effect transistor (FET) carrier mobility and on/off ratio. Under the guidance of analyzing the CVD reaction processes and kinetics between sulfur and MoO₃ precursor, we are able to attribute this inhibition of the intermediates to a relatively high atmosphere ratio between sulfur and MoO₃/MoO_3−x vapor, which would result in the complete transformation of the thermally decomposed MoO_3−x vapor into MoS₂ within a stable and supersaturation of sulfur atmosphere even at fast kinetic reactions at high temperatures. In addition, a series of vertical stacking structures of the layered MoO₂–MoS₂ hybrids could be obtained as well by controlling the sulfur atmosphere with respect to the MoO₃ vapor. This study reveals a CVD reactive mechanism for completely converting MoO₃/MoO_3−x into MoS₂ and will provide a direction for enabling both the synthesis of clean and pure MoS₂ crystals, and vertically structured MO₂–MoS₂ hybrids for potential applications. More important, it will assist in our understanding of the fundamental CVD growth mechanisms of TMDCs for optimizing synthesis.

2 Experimental section

2.1 CVD growth of pure, clean MoS₂ and MoO_3−x–MoS₂ hybrids

The experiment adopts the tube furnace with a double temperature zone, comprising of low temperature heating zone and high temperature heating zone. Sulfur powder (20 mg, 99.98%, Sigma-Aldrich) was loaded in a separate quartz boat located at the low temperature heating zone, which is 29 cm away from the MoO₃ (4 mg, 99.97%, Sigma-Aldrich) containing boat. Growth substrates were silicon with a 300 nm layer of SiO₂. Substrates were cleaned via ultra-sonication in acetone and IPA for 15 min, respectively, followed by N₂ drying. After that, the cleaned silicon substrates (300 nm-SiO₂/Si) were placed upside down on the quartz boat directly above the MoO₃ source. The schematic for the experiment is shown in Fig. 1a. Before starting the CVD growth, the CVD system was pumped down to ∼0.1 Pa, and then the quartz tube was filled with Ar gas and pumped again. This was repeated three times to exhaust the air and impurities in the quartz tube as far as possible. For the CVD growth, the furnaces were heated up to 120 °C and 880 °C in 20 min, respectively, under atmospheric pressure with 200 sccm Ar flowing. Then, the first furnace was heated at a varied rate (1, 5, 8 °C min⁻¹) for 10 min to the desired temperature, while the second furnace was held at 880 °C for 10 min growth. After the growth, the system was cooled down to the room temperature by simply opening the furnaces.

Fig. 1 (a) Schematic of the CVD system that is divided by low and high temperature zones. (b) Schematic illustration of the CVD synthesis of (c) MoS₂–MoO_3−x composites and (d) clean, pure MoS₂ crystals on SiO₂/Si, respectively, using S and MoO₃ precursors.

2.2 Direct synthesis of MoO_3−x intermediate products without S precursor

Here, only the second furnace was used for the synthesis of MoO_3−x intermediate products. 4 mg MoO₃ powder was put in the quartz boat, which was covered by a growth substrate. The furnace was then heated up to 880 °C within 20 min at 200 sccm Ar flowing under atmospheric pressure. After holding the temperature for 10 min, the furnace was cooled to the room temperature by opening the furnace.

2.3 Characterization methods

The features of the as-grown pure and clean MoS₂ layers and MoS₂–MoO_3−x composites were characterized via optical microscopy, scanning electron microscopy (SEM, JEOL, JSM-6390A at 15 kV), X-ray diffraction (XRD, Bruker, D8 Advance with a Cu Kα X-ray source), and Raman spectroscopy (HORIBA microscope under ambient conditions using a 514 nm excitation laser source).

2.4 Monolayer clean MoS₂-based device fabrication

Monolayer MoS₂-based FETs on SiO₂ were defined and patterned via electron beam lithography (EBL), followed by the electron beam evaporation of Ti/Au (5 nm/45 nm) and lift-off. Devices were electrically tested using a Keithley 4200 semiconductor parameter analyzer at room temperature at ambient condition.

3 Results and discussion

Fig. 1a shows the main configurations of the CVD system used for the growth of MoS₂. As shown, the samples were grown by CVD with solid S and MoO₃ powders in an Ar atmospheric pressure. The growth substrate (SiO₂/Si) faced down to MoO₃ powder. In theory, the powder-based CVD growth of MoS₂ would undergo a series of thermal-dynamics and chemical processes, which can be denoted by these chemical reaction equations:²¹

2MoO₃ + xS → 2MoO_3−x + xSO₂ (1)

2MoO_3−x + (7 − x)S → 2MoS₂ + (3 − x)SO₂ (2)

Eqn (1) and (2) above are combined into a final equation:

2MoO₃ + 7S → 2MoS₂ + 3SO₂ (3)

These equations show that the quantity matching between the evaporated MoO₃ and sublimated S is significant for the formation of MoS₂. As shown in Fig. 1b–d, the reaction environment with low S super-saturation (S-poor) would induce some intermediate products MoO_3−x derived from MoO₃, e.g. MoO₂, forming MoO_3−x–MoS₂ composites (Fig. 1c) as a final product. On the other hand, the reaction environment with high S super-saturation (S-rich) would instead inhibit the formation of MoO_3−x intermediates and grow clean and pure MoS₂ crystals (Fig. 1d).

Therefore, here, two independent furnaces were used to precisely control the temperature of S and MoO₃ with a specific heating rate. This is important for the control over the sublimation of S and the evaporation of MoS₃, respectively, and then determination of the CVD products. In detail, the temperature for S and MoO₃ powder was ramped to 120 °C (slightly higher than the sulfur melting point of 112.8 °C) and 880 °C in 20 min, respectively. This temperature ramp would ensure the full evaporation of solid MoO₃ and provide an active environment with high diffusivity of the precursor and increased number of the active MoO_3−x intermediates.²² The followed growth time was set to be 10 min at 880 °C; meanwhile, the S powder was further heated up in varied ramping rates (1, 5, 8 °C min⁻¹) during 10 min growth time, as shown in Fig. 2a–c. The varied ramping rate would produce different quantity of the sublimated S vapor, which would cause different CVD growth products. For instance, the ramp rate of 1 °C min⁻¹ (Fig. 2a) would produce a S-poor environment with low S concentration, resulting in an insufficient sulfurization of the MoO_3−x intermediates, finally forming MoO_3−x–MoS₂ composites (see Fig. 2d and g). As shown in Fig. 2d and g, quadrilateral MoO_3−x–MoS₂ composites with different nucleation densities could be found in the substrate center and edge due to the location effect of the precursor diffusion.^23–25 When the heating ramp rate of S came to 5 °C min⁻¹ (Fig. 2b), the CVD products would change to MoO_3−x–MoS₂ stacks (Fig. 2c and h), showing quadrilateral MoO_3−x stacked with triangular MoS₂, as shown in the inset of Fig. 2c and h. Further increasing the heating ramp rate of S to 8 °C min⁻¹ (Fig. 2c), almost clean and pure MoS₂ crystals would be obtained in both of the substrate center (Fig. 2f) and edges (Fig. 2i), just with different layers.

Fig. 2 Temperature ramping processes of the MoO₃ and S precursors as a function of time with different heating rates of the S during growth: (a) 1 °C min⁻¹, (b) 5 °C min⁻¹, and (c) 8 °C min⁻¹. (d–i) Optical images of as-grown (d and g) MoO_3−x–MoS₂ composites, (e and h) stacks of quadrilateral MoO_3−x/triangular MoS₂, and (f and i) clean triangular MoS₂ grains at the center and edge of the SiO₂/Si substrate using different S heating rates ((d and g) 1 °C min⁻¹, (e and h) 5 °C min⁻¹, and (f and i) 8 °C min⁻¹).

Fig. 3a–c show the deposited pattern transition of the sample as the heating rate of S increases gradually in Fig. 2. Heating S with a low sublimation rate results in an insufficient S reduction atmosphere for MoO₃, which allows the evaporated MoO₃ and MoO_3−x preferentially nucleate and grow on the substrate in kind of quadrilateral shapes. When almost all the MoO₃ precursor was consumed, the S vapor was transported to the surface of the quadrilateral MoO_3−x, sulfurization of MoO_3−x would occur and MoS₂ layers would be formed on the top (schematically shown in Fig. 3a). It should be noted that the MoS₂ layers on top would hinder the further reaction between MoO_3−x and S vapor, forming quadrilateral MoO_3−x–MoS₂ composites as experimentally shown above (Fig. 2d and g). With the increase in the S heating rate, triangular MoS₂ sheet can be formed and deposited on top of quadrilateral MoO_3−x, forming stacks of quadrilateral MoO_3−x and triangular MoS₂ (schematically shown in Fig. 3b). With further rapid increase in the S heating rate, the appropriate ratio between S and MoO_3−x vapor would be meted, and the evaporated MoO_3−x would be completely transformed to be clean and pure MoS₂ crystals, as schematically shown in Fig. 3c. Alternatively, the increased S ramping rate during the growth time would produce different quantities of sublimated S vapor and enable a transition of the reaction environment from being S-poor to S-rich, which induce the corresponding change of products from the MoO_3−x–MoS₂ composites to stacks of quadrilateral MoO_3−x/triangular MoS₂, then ultimately to clean and pure triangular MoS₂ (Fig. 3a–c).

Fig. 3 (a–c) Schematic evolution of the deposited patterns of sample prepared by gradually increasing the heating rate of S during the growth from (a) 1 °C min⁻¹ to (b) 5 °C min⁻¹, and finally to (c) 8 °C min⁻¹. (d) XRD patterns of the samples prepared by gradually increasing the heating rate of S during the growth from 1 °C min⁻¹ to (e) 5 °C min⁻¹, and finally to (f) 8 °C min⁻¹. (g) Raman spectra of deposited samples corresponding to the pattern in (a)–(c). (h) The fitted curves of the Raman intensity ratio of I_MoO2/I_MoS2 at ∼494 and ∼381 cm⁻¹ and the corresponding cleanness with respect to the heating rate of S during the growth.

In order to conform what MoO_3−x exactly is, Raman spectroscopy and X-ray diffraction (XRD) were employed for measuring these grown samples. Fig. 3d–g is the typical XRD pattern and Raman spectra for the as-grown quadrilateral MoO_3−x–MoS₂ composites, stacks of quadrilateral MoO_3−x/triangular MoS₂, and clean, pure triangular MoS₂ crystals (Fig. 2d–i), which corresponds to 1, 5, 8 °C min⁻¹ heating rates of S and the configuration patterns in Fig. 3a–c, respectively. As shown in Fig. 3d–f, all the XRD patterns show characteristic diffraction peaks at ∼14.0° (002), ∼33.0° (102), and ∼60.7° (112), confirming the existence of the MoS₂ phase (JCPDS No. 75-1539)^26,27 in all the samples. Particularly, there are obvious diffraction peaks (∼26.0° (011), ∼37.3° (020), and ∼53.5° (022)) present in both the samples of quadrilateral MoO_3−x–MoS₂ composites (Fig. 3d) and stacks of quadrilateral MoO_3−x/triangular MoS₂ (Fig. 3e), which are ascribed to the MoO₂ phase (JCPDS No. 86-0135).^26,27 On the other hand, the typical Raman spectra (Fig. 3g) demonstrate obvious Raman characteristics of E_2g¹ and A_1g for MoS₂,^28,29 which are assigned to the in-plane vibration of molybdenum and sulfur atoms (E_2g¹), and the out-of-plane vibration of sulfur atoms (A_1g),³⁰ respectively. It is necessary to note that the difference between E_2g¹ and A_1g (Δk = A_1g − E_2g¹), which can be employed to discriminate the layers number of MoS₂,^28,29,31,32 shows differences for each here. As shown, Δk for the sample grown at the 1, 5, and 8 °C min⁻¹ heating rates of S is 27, 23, and 21 cm⁻¹ (also see the summarized results in Table S1†), respectively, which demonstrates that the layers number of MoS₂ here are changed from multi-layer to mono-layer²⁸ with the increase in the S heating rate from 1, to 5 and finally to 8 °C min⁻¹. This is due to a different reaction mechanism, where a high S heating rate with an appropriate ratio between S and MoO_3−x vapor would react in a thermodynamically stable process²² for the diffusion of Mo and S atoms to the energetically favourable sites, and forming thin MoS₂ layer. On the contrary, a low S heating rate with insufficient S to MoO_3−x vapor would cause the formation and deposition of MoO_3−x intermediates with sulfurized thick top layers of MoS₂. More importantly, there are many other peaks appearing in the Raman spectra of the samples of quadrilateral MoO_3−x–MoS₂ composites (located at 128, 204, 230, 344, 361, 494, 566, 741 cm⁻¹) and stacks of quadrilateral MoO_3−x/triangular MoS₂ (located at 124, 204, 227, 494, 566, 742 cm⁻¹). These peaks are assigned to the Raman characteristics of MoO₂ (ref. 9, 13, 33) and/or MoOS₂,³⁴ which are absent in the Raman spectrum for the sample of clean and pure MoS₂ crystals prepared at the rapid heating rate of S (8 °C min⁻¹) during the growth. All the results from Raman characterization are summarized in the Table S1 in the ESI.† These XRD and Raman features confirm that the general MoO_3−x intermediates generated here during the growth are mainly MoO₂ and/or MoOS₂ due to a stepwise sulfurization to convert MoO₃ into MoS₂.³⁴ Further, based on these Raman features, the Raman intensity ratio of I_MoO2/I_MoS2 at ∼494 and ∼381 cm⁻¹ is plotted and fitted (Fig. 3h black curve), showing a decreasing tendency of MoO₂ (increase tendency of MoS₂) with the increase in the heating rate of S. Simultaneously, the corresponding cleanness/purity of the as-prepared samples can be plotted and fitted as well with the increase of heating rate of S during the growth. As shown in Fig. 3h (red curve), the cleanness/purity of the sample prepared increases with the increase in the heating rate of S during the growth. The cleanness/purity of the sample prepared at the 8 °C min⁻¹ heating rate of S during the growth can be as high as 90% (Fig. 3h). Indeed, all of these XRD and Raman results indicate that an appropriate ratio between sublimated S and evaporated MoS₃ during the growth is the key factor for the inhibition of MoO_3−x intermediates and determination of the final products with clean and pure MoS₂ crystals.

In addition, an S-free CVD process with only MoO₃ precursor (the setup and growth processes can be seen in Fig. S1a and b†) was conducted to indirectly explore the existence of MoO_3−x intermediates. The optical (Fig. S2a†) and SEM (Fig. S2b†) images show that the quadrilateral and striped shapes of the products are grown and distributed on the substrate after growth. Further, the corresponding Raman spectrum of the obtained product show the Raman characteristic peaks (128, 230, 343, 361, 494, 566, and 742 cm⁻¹) of MoO₂ (Fig. S2c†), which are almost the same with that of the quadrilateral MoO_3−x–MoS₂ composites grown at 1 °C min⁻¹ heating rate of S, as we discussed above. This indicates that here the as-prepared sample without S was mainly MoO₂. In particular, after the post-sulfurization of the sample, MoO₂ could be transformed into quadrilateral MoO₂–MoS₂ composites as well (Fig. S3a and b†), which was confirmed by the presence of Raman A_1g and E_2g¹ features of MoS₂ in their Raman spectra (Fig. S3c†). This agrees well with the growth mechanism we proposed with S-poor environment: the MoO₃ precursor would be consumed first and transform into MoO_3−x (mainly MoO₂) in an S-poor environment, followed by sulfurization on the top layers, forming MoO_3−x–MoS₂ hybrids.

Some larger clean and pure MoS₂ crystals were chosen for the detailed Raman spectroscopy and AFM characterizations (Fig. 4). As shown, two characteristic Raman vibration modes E_2g¹ and A_1g with a frequency difference Δk of 19 cm⁻¹ can be seen in the spectrum in Fig. 4a. This frequency difference Δk is slightly changed compared with the Raman data (21 cm⁻¹) in Fig. 3g due to the sample difference. This matches well with the frequency difference of the CVD-grown monolayer MoS₂ in a previous work.³⁵ The full width at half-maximum (FWHM) of the E_2g¹ peak is 3.8 cm⁻¹, which is close to that of the exfoliated monolayer MoS₂, 3.7 cm⁻¹, suggesting good crystalline quality of the clean and pure monolayer crystals here. We also performed Raman mapping on a corner for the large triangular MoS₂ crystal (inset of Fig. 4a) by plotting the Raman intensity map of the A_1g (Fig. 4b) and E_2g¹ (Fig. S4†) peaks to show the characteristic uniformity of a large area. Fig. 4c and d give a typical AFM measurement, showing the triangular topography (Fig. 4c) with a thickness of the edge ∼0.8 nm (Fig. 4d). This layer thickness is in the range of a single-layer MoS₂ film on the bare substrates.^24,36

Fig. 4 (a) Raman spectrum of the clean and pure MoS₂ crystals. The inset is the corresponding optical image of the clean and pure MoS₂ crystals. (b) The A_1g intensity Raman map corresponding to the corner area of the clean and pure MoS₂ crystals shown in the upper right of panel (a). (c) AFM topography image and (d) the correlated height profile taken across the dotted white line in (c), respectively, for a clean and pure MoS₂ crystals.

To further evaluate the quality of the as-obtained clean and pure MoS₂, the back-gated field-effect transistor (FET) devices were directly constructed and fabricated based on these monolayer clean and pure MoS₂ crystals. Their fabrications were based on previous reports,^37–39 and briefly described in the experimental section. Fig. 5a is the optical image of the fabricated FET devices, which is also schematically shown in Fig. 5b, showing the structural configuration on the SiO₂/Si substrate using Ti/Au as the source-drain (S-D) electrodes. A typical device numbered with ① is optically shown with high-magnification in the Fig. 5c. As shown, the ratio between specific channel length and width (L/W) is ∼8 μm/13 μm.

Fig. 5 (a) Optical image of the FET devices based on as-grown clean and pure MoS₂ crystals. (b) Schematic of the FET device configuration based on the MoS₂ layer. (c) The high-magnification optical image of the FET device numbered with ① in (a). (d) Output characteristics (I_ds–V_ds) of the FET device with respect to the back-gate voltage at 0, 20, 30, and 40 V, respectively. (e) Transfer characteristics curves (I_ds–V_gate) of the FET device at 5 V D–S voltage.

The electrical transport measurements of the fabricated devices were conducted under ambient conditions. Fig. 5d is the typical output characteristics (I_ds–V_ds) of the FET device at various gate voltage V_bg (0, 20, 30, and 40 V) under ambient conditions, and the nonlinear output characteristics suggest the existence of a small Schottky barrier between the Ti/Au metal contacts and MoS₂ layer.⁴⁰ The transfer characteristics (I_ds–V_gate) of the MoS₂ device at various S-D voltages are shown in Fig. S5.† As shown, the I_ds value increases monotonically with the increase in V_gate, indicating an n-type semiconducting behavior. Specifically, the transfer curve at the 5 V S-D voltage (V_ds = 5 V) is plotted in Fig. 5e (red curve), and the FET mobility could be extracted based on the linear region of the transfer curve according to the formula, μ_FE = L_chg_m/W_chC_oxV_ds, where L_ch, and W_ch are the device channel length and width, respectively, g_m is the transconductance of MoS₂ that can be extracted by performing a linear fit in the I_ds–V_gate curve, V_ds is the S-D voltage, C_ox is the gate capacitance of SiO₂ dielectrics. This yields electron mobility of ∼4 cm⁻² V⁻¹ s⁻¹ with an observed current modulation I_on/I_off ratio of ∼10⁶ by plotting the I_ds–V_gate curve on a logarithmic scale (black curve in Fig. 5e). The electron mobilities of all the devices we measured are in the range of 1–12 cm⁻² V⁻¹ s⁻¹. These results demonstrate that here the as-obtained clean and pure MoS₂ crystals possess decent quality with large scale. It is necessary to note that the mobility could be improved by measuring the device under certain conditions (or by other treatments) such as low-temperature, vacuum, and high-k top gate dielectric treatment and interface engineering.^12,41,42

4 Conclusions

In conclusion, we have successfully realized the growth of clean and pure MoS₂ crystals by the inhibitation of MoO_3−x intermediates in a powder-based CVD method. Through rational design of the heating rate of S from low S supersaturating to a high S supersaturating environment during the growth process, a series of CVD products transformed from the quadrilateral MoO_3−x/MoS₂ composites to stacks of quadrilateral MoO_3−x/triangular MoS₂, and finally to clean, pure triangular MoS₂ crystals could be obtained. Various characterization techniques confirmed that the intermediate of MoO_3−x is mainly MoO₂, which could be inhibited largely by appropriate ratio between the sublimated S and evaporated MoO₃ vapor. Further, the electrical transport measurements indicate that the as-synthesized clean and pure MoS₂ crystals have high quality with decent FET electron mobility. This work will assist in our understanding of the fundamental CVD growth mechanisms of TMDCs and provide a growth-optimized direction for the synthesis of clean and pure TMDCs crystals.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Birong Luo acknowledges financial support from Natural Science Foundation of Tianjin (20JCYBJC00350), the 13th Tianjin Thousand Youth Talents Plan Project (ZX0471800601) and generous start-up funding from Tianjin Normal University (YJRL201904).

References

1. H. Zeng, J. Dai, W. Yao, D. Xiao and X. Cui, Nat. Nanotechnol., 2012, 7, 490–493.
2. M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh and H. Zhang, Nat. Chem., 2013, 5, 263–275.
3. M. Belete, S. Kataria, U. Koch, M. Kruth, C. Engelhard, J. Mayer, O. Engstrom and M. C. Lemme, ACS Appl. Nano Mater., 2018, 1, 6197–6204.
4. H. Y. Chang, M. N. Yogeesh, R. Ghosh, A. Rai, A. Sanne, S. Yang, N. Lu, S. K. Banerjee and D. Akinwande, Adv. Mater., 2016, 28, 1818–1823.
5. Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman and M. S. Strano, Nat. Nanotechnol., 2012, 7, 699–712.
6. P. Yang, Z. Zhang, M. Sun, F. Lin, T. Cheng, J. Shi, C. Xie, Y. Shi, S. Jiang, Y. Huan, P. Liu, F. Ding, C. Xiong, D. Xie and Y. Zhang, ACS Nano, 2019, 13, 3649–3658.
7. E. Singh, P. Singh, K. S. Kim, G. Y. Yeom and H. S. Nalwa, ACS Appl. Mater. Interfaces, 2019, 11, 11061–11105.
8. Y.-H. Lee, X.-Q. Zhang, W. Zhang, M.-T. Chang, C.-T. Lin, K.-D. Chang, Y.-C. Yu, J. T.-W. Wang, C.-S. Chang, L.-J. Li and T.-W. Lin, Adv. Mater., 2012, 24, 2320–2325.
9. H. F. Liu, S. L. Wong and D. Z. Chi, Chem. Vap. Deposition, 2015, 21, 241–259.
10. G. H. Han, N. J. Kybert, C. H. Naylor, B. S. Lee, J. Ping, J. H. Park, J. Kang, S. Y. Lee, Y. H. Lee, R. Agarwal and A. T. C. Johnson, Nat. Commun., 2015, 6, 6128.
11. V. Kranthi Kumar, S. Dhar, T. H. Choudhury, S. A. Shivashankar and S. Raghavan, Nanoscale, 2015, 7, 7802–7810.
12. J. Chen, W. Tang, B. Tian, B. Liu, X. Zhao, Y. Liu, T. Ren, W. Liu, D. Geng, H. Y. Jeong, H. S. Shin, W. Zhou and K. P. Loh, Adv. Sci., 2016, 3, 1500033.
13. X. Wang, H. Feng, Y. Wu and L. Jiao, J. Am. Chem. Soc., 2013, 135, 5304–5307.
14. A. Govind Rajan, J. H. Warner, D. Blankschtein and M. S. Strano, ACS Nano, 2016, 10, 4330–4344.
15. Y. H. Lee, X. Q. Zhang, W. Zhang, M. T. Chang, C. T. Lin, K. D. Chang, Y. C. Yu, J. T. Wang, C. S. Chang, L. J. Li and T. W. Lin, Adv. Mater., 2012, 24, 2320–2325.
16. S. Wang, M. Pacios, H. Bhaskaran and J. H. Warner, Nanotechnology, 2016, 27, 085604.
17. M. A. Nikpay, S. Z. Mortazavi, A. Reyhani and S. M. Elahi, Opt. Quantum Electron., 2018, 50, 252.
18. H. Liu, Y. Zhu, Q. Meng, X. Lu, S. Kong, Z. Huang, P. Jiang and X. Bao, Nano Res., 2017, 10, 643–651.
19. T. Park, M. Leem, H. Lee, W. Ahn, H. Kim, J. Kim, E. Lee, Y.-H. Kim and H. Kim, J. Phys. Chem. C, 2017, 121, 27693–27699.
20. S. S. Withanage, H. Kalita, H. S. Chung, T. Roy, Y. Jung and S. I. Khondaker, ACS Omega, 2018, 3, 18943–18949.
21. N. K. Perkgoz and M. Bay, Nano-Micro Lett., 2016, 8, 70–79.
22. J. Lee, S. Pak, P. Giraud, Y.-W. Lee, Y. Cho, J. Hong, A.-R. Jang, H.-S. Chung, W.-K. Hong, H. Y. Jeong, H. S. Shin, L. G. Occhipinti, S. M. Morris, S. Cha, J. I. Sohn and J. M. Kim, Adv. Mater., 2017, 29, 1702206.
23. A. M. van der Zande, P. Y. Huang, D. A. Chenet, T. C. Berkelbach, Y. You, G.-H. Lee, T. F. Heinz, D. R. Reichman, D. A. Muller and J. C. Hone, Nat. Mater., 2013, 12, 554–561.
24. S. Wang, Y. Rong, Y. Fan, M. Pacios, H. Bhaskaran, K. He and J. H. Warner, Chem. Mater., 2014, 26, 6371–6379.
25. H. Schmidt, S. Wang, L. Chu, M. Toh, R. Kumar, W. Zhao, A. H. Castro Neto, J. Martin, S. Adam, B. Özyilmaz and G. Eda, Nano Lett., 2014, 14, 1909–1913.
26. Y. Huang, X. Xie, Y. Zhang, J. Ding, L. Liu, Y. Fan, H. Lv, Y. Liu and Q. Cai, Appl. Surf. Sci., 2020, 520, 146340.
27. J. Zhou, N.-S. Xu, S.-Z. Deng, J. Chen, J.-C. She and Z.-L. Wang, Adv. Mater., 2003, 15, 1835–1840.
28. H. Li, Q. Zhang, C. C. R. Yap, B. K. Tay, T. H. T. Edwin, A. Olivier and D. Baillargeat, Adv. Funct. Mater., 2012, 22, 1385–1390.
29. S. Najmaei, Z. Liu, P. M. Ajayan and J. Lou, Appl. Phys. Lett., 2012, 100, 013106.
30. P. Tonndorf, R. Schmidt, P. Bottger, X. Zhang, J. Borner, A. Liebig, M. Albrecht, C. Kloc, O. Gordan, D. R. T. Zahn, S. M. de Vasconcellos and R. Bratschitsch, Opt. Express, 2013, 21, 4908–4916.
31. M. Boukhicha, M. Calandra, M.-A. Measson, O. Lancry and A. Shukla, Phys. Rev. B: Condens. Matter Mater. Phys., 2013, 87, 195316.
32. F. Bonaccorso, P.-H. Tan and A. C. Ferrari, ACS Nano, 2013, 7, 1838–1844.
33. M. A. Camacho-López, L. Escobar-Alarcón, M. Picquart, R. Arroyo, G. Córdoba and E. Haro-Poniatowski, Opt. Mater., 2011, 33, 480–484.
34. J. V. Pondick, J. M. Woods, J. Xing, Y. Zhou and J. J. Cha, ACS Appl. Nano Mater., 2018, 1, 5655–5661.
35. N. Kumar, R. Tomar, N. Wadehra, M. M. Devi, B. Prakash and S. Chakraverty, Cryst. Res. Technol., 2018, 53, 1800002.
36. C. Lee, H. Yan, L. E. Brus, T. F. Heinz, J. Hone and S. Ryu, ACS Nano, 2010, 4, 2695–2700.
37. B. Luo, B. Chen, L. Meng, D. Geng, H. Liu, J. Xu, Z. Zhang, H. Zhang, L. Peng and L. He, Adv. Mater., 2014, 26, 3218–3224.
38. B. R. Luo, J. M. Caridad, P. R. Whelan, J. D. Thomsen, D. M. A. Mackenzie, A. G. Cabo, S. K. Mahatha, M. Bianchi, P. Hofmann, P. U. Jepsen, P. Boggild and T. J. Booth, 2D Mater., 2017, 4, 045017.
39. J. Duan, X. Wang, X. Lai, G. Li, K. Watanabe, T. Taniguchi, M. Zebarjadi and E. Y. Andrei, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 14272–14276.
40. Y.-H. Lee, L. Yu, H. Wang, W. Fang, X. Ling, Y. Shi, C.-T. Lin, J.-K. Huang, M.-T. Chang, C.-S. Chang, M. Dresselhaus, T. Palacios, L.-J. Li and J. Kong, Nano Lett., 2013, 13, 1852–1857.
41. W. Chen, J. Zhao, J. Zhang, L. Gu, Z. Yang, X. Li, H. Yu, X. Zhu, R. Yang, D. Shi, X. Lin, J. Guo, X. Bai and G. Zhang, J. Am. Chem. Soc., 2015, 137, 15632–15635.
42. S. Das, H.-Y. Chen, A. V. Penumatcha and J. Appenzeller, Nano Lett., 2013, 13, 100–105.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0ce01354d

---