Effect of vacancies in monolayer MoS₂ on electronic properties of Mo–MoS₂ contacts

Abstract

Revealing the influence of intrinsic defects in monolayer MoS₂ on the electronic nature of metal–MoS₂ contacts is particularly critical for their practical use as nanoelectronic devices. This work presents a systematic study toward electronic properties of Mo metal contacts to monolayer MoS₂ with vacancies by using first-principles calculations based on density functional theory. Upon Mo- and S-vacancy formation in monolayer MoS₂, both the height and the width of the tunnel barrier between Mo metal and monolayer MoS₂ are decreased. Additionally, the Schottky barrier of 0.1 eV for the perfect Mo–MoS₂ top contact is reduced to zero for defective ones. The partial density of states near the Fermi level of defective Mo–MoS₂ top contacts is strengthened and electron densities at the interface of defective Mo–MoS₂ top contacts are increased compared with those of the perfect one, suggesting that Mo- and S-vacancies in monolayer MoS₂ have the possibility to improve the electron injection efficiency. Mo-vacancies in monolayer MoS₂ are beneficial to get high quality p-type Mo–MoS₂ contacts, whereas S-vacancies in monolayer MoS₂ are favorable to achieve high quality n-type Mo–MoS₂ contacts. Our findings provide important insights into future design and fabrication of nanoelectronic devices with monolayer MoS₂.

---

1. Introduction

Monolayer transition-metal dichalcogenides (mTMD), a family of 2D semiconductor layers arranged in a hexagonal lattice, have drawn tremendous attention as promising channel materials for digital electronic applications due to their nonzero band gap, small thicknesses, and pristine interfaces without out-of-plane dangling bonds.^1–10 Among various mTMD materials, monolayer MoS₂ has emerged because of its atomically thickness of ∼7 Å per layer,¹¹ considerable band gap of 1.8 eV,¹² planar nature, and pristine surfaces. Most recently, monolayer MoS₂ has been used to construct field-effect transistors (FETs), which can offer lower power consumption than classical transistors.^1,13 New phototransistor based on monolayer MoS₂ has been demonstrated to have a better photoresponsivity as compared with the graphene-based device.¹⁴ Moreover, FET based biosensors with mTMD semiconductor as the channel material have been fabricated and exhibit highly advantageous over all other nanomaterial-based FET biosensors.¹⁵

However, low-resistance metal contacts to monolayer MoS₂ remain a critical issue for its transistor applications because several factors, such as large band gap, pristine surfaces and lack of proper doping approach, may mask the innate exceptional electronic and magnetic properties of monolayer MoS₂.^16,17 In order to overcome this issue, many studies have been performed to reduce the tunnel barrier and Schottky barrier in metal–MoS₂ contacts.^16–20 Popov et al.²⁰ have studied Ti–MoS₂ and Au–MoS₂ top contacts by density functional theory, indicating that the most common contact Au metal is rather inefficient for electron injection into monolayer MoS₂. Kang et al.¹⁶ have evaluated In, Ti, Au, Pd, and Mo, contacts to monolayer MoS₂ by density functional theory calculations, implying that Ti and Mo have great potential to form favorable n-type top contacts to monolayer MoS₂. Nevertheless, the Schottky barrier for Ti–MoS₂ contact is about 0.33 eV,¹⁷ which is still very high. Mo–MoS₂ contact has ultra-low Schottky barrier of 0.1 eV at source/drain-channel junction, and high-performance FETs based on Mo–MoS₂ contact have been demonstrated.¹⁷ Hence, Mo has been proposed as a promising contact metal to monolayer MoS₂.

Vacancy defects were found to exist in monolayer MoS₂ when monolayer MoS₂ was prepared through sonochemical deposition²¹ and exfoliated method.²² Several literatures have reported the effect of vacancies on properties of monolayer MoS₂. Ataca et al.^23,24 have calculated the formation energy of neutral vacancies in monolayer MoS₂ and studied the influence of vacancies on magnetic properties of monolayer MoS₂, implying that vacancy creation appears to be a promising way to extend the applications of MoS₂. The formation energies of charged vacancies in monolayer MoS₂ under different atmospheric conditions have been investigated.^25,26 Feng et al.²⁷ have indicated that structural, electronic, and optical properties of monolayer MoS₂ depend greatly on its intrinsic vacancies.

It should be noted that vacancy defects in monolayer MoS₂ not only influence the properties of monolayer MoS₂ but also affect the interfacial and electrical properties of metal–MoS₂ contacts. However, to the best of our knowledge, effects of vacancies in monolayer MoS₂ on electronic structure and electronic properties of Mo–MoS₂ contact are not well understood yet. It is well known that the knowledge of electronic properties of Mo–MoS₂ contact is very important for the practical applications of monolayer MoS₂ as well as for the designing and analyzing of optoelectronic devices. Therefore, this work is focused on investigating the effect of vacancies in monolayer MoS₂ on electronic structure and electronic properties of Mo–MoS₂ contacts using first-principles calculations.

2. Computational details

In the present calculations, the exchange correlation of the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional as implemented in CASTEP code²⁸ was employed. In order to consider the van der Waals interactions in TMD materials, DFT-D2 is adopted in this work, where the potential is described via a simple pairwise force field and is optimized for popular DFT functionals.²⁹ The electron–ion interactions were described by norm-conserving Troullier–Martins pseudopotentials³⁰ with partial core corrections. The plane-wave cutoff energy was set to be 200 Ry after extensive convergence analysis. The Brillouin-zone of Mo–MoS₂ top contact region was performed over the 8 × 8 × 1 k-point grids using the Monkhorst–Pack method,³¹ where the self-consistent convergence of the total energy is 1.0 × 10⁻⁶ eV per atom. Conjugate gradient scheme was used to relax supercell until the component of the forces on each atom was less than 0.01 eV Å⁻¹.

Mo–MoS₂ top contact was modeled by a supercell slab, which is periodic in the x and y directions and separated by 20 Å vacuum region in the z direction to minimize the interactions between adjacent image cells. The supercell slab of Mo–MoS₂ top contact contains 4 × 4 unit cells of monolayer MoS₂ and the close-packed surface of Mo (001) extending to the 6^th layer, which is the most probable orientation to be found in experiments. The supercell geometry of perfect Mo–MoS₂ top contact is shown in Fig. 1(a), and geometries of defective Mo–MoS₂ top contacts with single Mo and S vacancy are presented in Fig. 1(b) and (c), respectively. When optimizing these models by using conjugate gradient technique, atoms except the 3^rd to 6^th layers of Mo metal from the interface were allowed to relax so as to evaluate the effect of interfacial layers.¹⁷ Although in real situations, the contact metals consist of many layers, the situation restricted to 6 layers of metal atoms is thick enough to accurately model the electronic properties of metal–MoS₂ contact^17,20,32 because the obtained results do not change appreciably beyond this thickness. A similar approach had been successfully used to characterize mTMD and their contacts to metal electrodes.^20,27

Fig. 1 The supercell geometry of Mo–MoS₂ top contacts. (a) Bottom view and side view of perfect Mo–MoS₂ top contact, (b) bottom view of defective Mo–MoS₂ top contact with Mo vacancy in monolayer MoS₂ and amplified sketch of monolayer MoS₂ with Mo vacancy, and (c) bottom view of defective Mo–MoS₂ top contact with S vacancy in monolayer MoS₂ and amplified sketch of monolayer MoS₂ with S vacancy.

3. Results and discussion

3.1 Tunnel barriers

The tunnel barrier between a metal and mTMD is characterized by its height and width, which are evaluated by the effective tunnel barrier height (Φ_TB,eff) and physical separation (d_p), respectively. The Φ_TB,eff is defined as the minimum effective potential (V_eff) difference between the Mo–MoS₂ interface and monolayer MoS₂. According to Kohn–Sham equations³³ and Kang et al.¹⁶ reports, the effective potential of an electron (V_eff(n)) represents the electron interaction with other electrons and external electrostatic field, and V_eff(n) can be expressed by^16,33

V_eff(n) = V_KS = V_H(n) + V_xc(n) + V_ext(n) (1)

where V_H(n) is the mean-field electrostatic interaction, V_xc(n) is the exchange–correlation potential, V_ext(n) represents external electrostatic interactions, and V_KS is the Kohn–Sham potential in Kohn–Sham equations. Fig. 2(a), as an example, shows the V_eff versus position on z axis for perfect Mo–MoS₂ top contact and denotes the Φ_TB,eff of perfect Mo–MoS₂ top contact. To evaluate the tunnel barrier height, the Φ_TB,eff of perfect and defective Mo–MoS₂ top contacts are plotted in Fig. 2(b). Obviously, Φ_TB,eff value of perfect Mo–MoS₂ top contact is about 0.12 eV, which is lower than those of Ti–MoS₂ (0.45 eV (ref. 20)), Au–MoS₂ (1.03 eV (ref. 20)) and Pd–MoS₂ (0.15 eV (ref. 16 and 32)) top contacts, implying that perfect Mo–MoS₂ top contact has higher carrier injection efficiency than Ti–, Au– and Pd–MoS₂ top contacts.¹⁶ When a Mo- or S-vacancy is formed in monolayer MoS₂, defective Mo–MoS₂ top contacts have negligible Φ_TB,eff value of 0.01 eV, indicating that carrier injection efficiency of defective Mo–MoS₂ top contacts is further improved. It should be mentioned that these estimates of Φ_TB,eff can be affected by a sizeable Self-Interaction Error (SIE),^34–37 owing to the use of the non-SIE free PBE functional. The SIE can be estimated to contribute an additional 0.18–0.27 eV to the calculated barriers.

Fig. 2 (a) Minimum effective potential height (V_eff) versus z position for perfect Mo–MoS₂ top contact. The effective tunnel barrier height (Φ_TB,eff) can be defined as the V_eff difference between the Mo–MoS₂ interface and monolayer MoS₂. Color key is Mo metal = blue and monolayer MoS₂ = green. (b) The Φ_TB,eff and physical separation d_S–Mo for perfect and defective Mo–MoS₂ top contacts.

In order to evaluate the tunnel barrier width, physical separations of perfect and defective Mo–MoS₂ contacts are calculated and presented in Fig. 2(b). The physical separation d_S–Mo and average separation d_Mo–Mo are defined as shown in Fig. 1(a). It is clear from Fig. 2(b) that d_S–Mo value of perfect Mo–MoS₂ contact is about 1.34 Å, which is consistent with previous theoretical value (1.27 Å (ref. 16)) and smaller than those of other metal–MoS₂ top contacts (2–3 Å).^20,32 This small physical separation is lower than the sum of S and Mo covalent radii.²⁰ Furthermore, the average separation d_Mo–Mo between the bottom Mo layer and the Mo layer in monolayer MoS₂ is about 3.02 Å, which is shorter than those of Ti–MoS₂ (3.57 Å) and Au–MoS₂ (4.21 Å)²⁰ top contacts. The small d_S–Mo and d_Mo–Mo indicate strong orbital overlaps and thin tunnel barrier.

When Mo-vacancy is formed in monolayer MoS₂, d_S–Mo value of defective Mo–MoS₂ top contact is reduced to 1.29 Å as shown in Fig. 2(b). There are two factors may lead to the decrease of the d_S–Mo value. On one hand, S ions surrounding Mo-vacancy undergo an outward relaxation.²⁷ On the other hand, dangling bonds of S ions, which induced by Mo-vacancy, can rebind with the Mo atoms in bottom Mo layer to shorten d_S–Mo. When S-vacancy is formed in monolayer MoS₂, d_S–Mo value of defective Mo–MoS₂ top contact is decreased to 1.25 Å because the Mo atoms around S-vacancy move toward to the vacancy.^27,38 Moreover, the average separation d_Mo–Mo between the bottom Mo layer and the Mo layer in monolayer MoS₂ is reduced from 3.02 to 2.92 Å, which is close to the diameter of Mo atom. Therefore, compared with perfect Mo–MoS₂ contact, defective Mo–MoS₂ top contacts are favorable to get stronger orbital overlaps and thus thinner tunnel barrier.

3.2 Density of states

The band structure and PDOS of monolayer MoS₂ are shown in Fig. 3(a) and (b), respectively. It can be seen that an obvious band gap of 1.78 eV is observed for monolayer MoS₂, which is consistent with the theoretical values of 1.73 eV (ref. 39) and 1.90 eV (ref. 40) obtained using GGA, but lower than the values of 2.12–2.78 eV obtained by HSE06 (ref. 41 and 42) and GW^23,43 (see ESI Table 1†). The experimental value of 1.80 eV (ref. 12) may be rather an optical band gap because of the experimental methods used and the approximations applied in the treatment of the data. Hence, the band gap of monolayer MoS₂ obtained by the higher-level functionals of HSE06 (ref. 41 and 42) and GW^23,43 may be more close to the true experimental data. This is consistent with the expectation that the PBE Kohn–Sham gap usually underestimates the true band gap.

Fig. 3 (a) Band structure of intrinsic monolayer MoS₂, (b) PDOS of intrinsic monolayer MoS₂, (c) PDOS of perfect Mo–MoS₂ top contact, (d) PDOS of defective Mo–MoS₂ top contact with Mo-vacancy in monolayer MoS₂, (e) PDOS of defective Mo–MoS₂ top contact with S-vacancy in monolayer MoS₂.

Fig. 3(c) presents the PDOS of perfect Mo–MoS₂ top contact. It is obvious from Fig. 3(c) that the band gap vanishes, indicating a metallic contact between Mo metal and monolayer MoS₂. According to the previous reports,^16,20 the n-type or p-type can be determined from the PDOS of monolayer MoS₂ in the Mo–MoS₂ contact. If the position of Fermi level (E_F) is shifted upwards the original conduction bands (E_c) of monolayer MoS₂, indicating that monolayer MoS₂ is doped n type. In contrast, if E_F is close to the original valence bands (E_v) of monolayer MoS₂, showing that monolayer MoS₂ is doped p type. In Fig. 3(c), the E_F of perfect Mo–MoS₂ top contact is shifted upwards, to 0.25 eV above the bottom of conduction bands of monolayer MoS₂, suggesting that monolayer MoS₂ is doped n-type by Mo. The shift of E_F is due to the fact that doping causes significant distortion to the band structure around the band gap. In order to comparatively study the PDOS of monolayer MoS₂ under different conditions, it is necessary to lineup the band structures of intrinsic monolayer MoS₂ and monolayer MoS₂ in Mo–MoS₂ top contacts. For this purpose, the valence bands maximum (E_VBM) of Mo–MoS₂ top contacts can be obtained by the following equation:^44,45

E_VBM = E_VBM(intrinsic) + V_av(interface) − V_av(intrinsic) (2)

where E_VBM(intrinsic) is the valence bands maximum of intrinsic monolayer MoS₂, V_av(intrinsic) and V_av(interface) represent the average potential of intrinsic monolayer MoS₂ and monolayer MoS₂ in Mo–MoS₂ top contacts, respectively. As a result, the band structures of Mo–MoS₂ top contacts are obtained (see ESI Fig. 1†). The borders of the valence and conduction bands of intrinsic monolayer MoS₂ are marked as vertical dot lines in the PDOS of Mo–MoS₂ top-contacts, as shown in Fig. 3(c). The dash line in Fig. 3(c) represents the E_F of the Mo–MoS₂ contact system. According to the previous reports,^16,46 the n-type (p-type) Schottky barrier is the difference between the bottom of conduction bands (the top of the valence bands) of intrinsic monolayer MoS₂ and the E_F of the Mo–MoS₂ top contact, as exhibited in Fig. 3(c). Like In–, Ti–, and Au–MoS₂ top contacts,^16,20,47 the E_F is pinned near the original conduction bands of intrinsic monolayer MoS₂. The Schottky barrier of perfect Mo–MoS₂ top contact is about 0.1 eV, which is in good agreement with previous theoretical results (0.13 eV,¹⁶ 0.1 eV (ref. 17)) and lower than those of Ti–MoS₂ (0.33 eV,^17,32 0.35 eV (ref. 16)), Au–MoS₂ (0.62 eV (ref. 16 and 32)) and Pd–MoS₂ (0.90 eV (ref. 16 and 32)) top contacts despite metal Mo has a high work function. In addition, high PDOS spread all over the original band gap of intrinsic monolayer MoS₂ implies the formation of Ohmic contact between Mo metal and monolayer MoS₂, which is consistent with previous theoretical and experimental report for Mo–MoS₂ top contact.^17,32 Furthermore, the broadening of the peaks in the PDOS near E_F reflects the formation of delocalized states with low effective electron mass allowing more electrons to be transferred between the metal and the mTMD layer.^20,32

PDOS of defective Mo–MoS₂ top contacts with single Mo and S vacancy are presented in Fig. 3(d) and (e), respectively. In the case of Mo-vacancy, as shown in Fig. 3(d), a high PDOS near E_F eliminates the band gap, indicating that Mo-vacancy has no effect on Ohmic contact and the metallic character of Mo–MoS₂ system. In addition, PDOS near E_F of defective Mo–MoS₂ top contact are much higher than those of perfect one because the dangling bonds of S atoms surrounding Mo-vacancy can form covalent bonding with Mo atoms in the bottom Mo layer, which almost like the covalent bonding formation for Mo–MoS₂ side contact.¹⁶ The higher PDOS near E_F suggests the lower effective carrier mass and the higher efficiency of carrier transport. Hence, upon Mo-vacancy forming in monolayer MoS₂, the Schottky barrier of defective Mo–MoS₂ top contact vanishes. Moreover, in contrast to perfect Mo–MoS₂ top contact, the E_F of defective Mo–MoS₂ top contact is shifted downwards, to 0.17 eV under the top of valence bands of monolayer MoS₂. Therefore, Mo-vacancy in monolayer MoS₂ is beneficial to achieve high quality p-type Mo–MoS₂ top contact.

In the case of S-vacancy, as shown in Fig. 3(e), high PDOS near E_F are also observed. Previous studies^27,38 have shown that S-vacancy in monolayer MoS₂ can induce defective states in the band gap. Thus, it can be found that some new peaks of PDOS appear near E_F due to strong hybridization of d orbitals between Mo metal and monolayer MoS₂. Additionally, the Schottky barrier of defective Mo–MoS₂ top contact disappears, implying the efficiency of electron transport is further improved. Similar to perfect Mo–MoS₂ top contact, the E_F of defective Mo–MoS₂ top contact is shifted upwards, to 0.62 eV above the bottom of conduction bands of intrinsic monolayer MoS₂. It should be noted that the position of the E_F of defective Mo–MoS₂ top contact is higher than that of perfect one. Therefore, in contrast to Mo-vacancy, S-vacancy is beneficial to achieve high quality n-type Mo–MoS₂ top contact.

3.3 Electron density

Average electron density of perfect and defective Mo–MoS₂ top contacts are calculated and shown in Fig. 4. The minimum x–y plane average electron density of Mo–MoS₂ interface and monolayer MoS₂ is marked ρ_i and ρ_m, respectively. Usually, high electron density at the interface of metal–mTMD contacts allows sufficient injection of charge into mTMD layer.²⁰ In Fig. 4(a), the ρ_i value of perfect Mo–MoS₂ top contact is about 0.042 bohr⁻³, which is higher than those of other metal–MoS₂ top contacts (0.013–0.033 bohr⁻³),^16,20 implying that Mo metal has advantage to achieve strong orbital overlaps with monolayer MoS₂, leading to low contact resistant and high electron injection efficiency.³² The previous studies^17,27 have indicated that strong covalent bonds are formed at the interface of perfect Mo–MoS₂ top contact. From Fig. 4(a), the ρ_i value of 0.042 bohr⁻³ is close to the ρ_m value of 0.051 bohr⁻³, indicating the formation of covalent bonds at the interface. This result is consistent with the previous reports.^17,27

Fig. 4 Average electron density value in the x–y planes normal to the z axis: (a) perfect Mo–MoS₂ top contact, (b) defective Mo–MoS₂ top contact with Mo-vacancy, (c) defective Mo–MoS₂ top contact with S-vacancy. Ball with a virtual edge in the panels represents Mo- or S-vacancy in monolayer MoS₂. The ρ_i and ρ_m in each panel indicate the minimum x–y plane average electron density at interface and monolayer MoS₂, respectively (in units of bohr⁻³).

Average electron densities of defective Mo–MoS₂ top contacts with single Mo and S vacancy are presented in Fig. 4(b) and (c), respectively. It can be seen that the average electron densities at the interface region of defective Mo–MoS₂ top contacts become higher compared with those in perfect Mo–MoS₂ system, showing that the covalent bonds between the Mo metal and monolayer MoS₂ are enhanced due to stronger orbital overlap, which are consistent with the above PDOS analysis. Consequently, Mo- and S-vacancy in monolayer MoS₂ have possibility to decrease the contact resistant and improve the electron injection efficiency. In addition, the ρ_i value of 0.047 bohr⁻³ for Mo–MoS₂ top contact with S-vacancy is slightly larger than the ρ_i value of 0.044 bohr⁻³ for Mo–MoS₂ top contact with Mo-vacancy, indicating that S-vacancy is more favorable to improve the electronic transport of Mo–MoS₂ top contacts. After introduction of Mo- or S-vacancy in monolayer MoS₂, the difference between ρ_i and ρ_m for defective Mo–MoS₂ top contacts is decreased.

3.4 Mulliken population

The electron density can also be confirmed by Mulliken population analysis. The calculated Mulliken populations of atoms in Mo metal layer and monolayer MoS₂ for perfect and defective Mo–MoS₂ top contacts are listed in Table 1. For perfect Mo–MoS₂ top contact, 0.042 e of Mo atoms in metal layer are transferred to interfacial S atoms in top sulfur layer, suggesting strong interactions between the Mo and S atoms. Simultaneously, interfacial S atoms accept 0.079 e from Mo atoms in monolayer MoS₂. As a result, the electron density at the Mo–MoS₂ interface region is lower than that of Mo–S bond in monolayer MoS₂. For defective Mo–MoS₂ top contacts, the electrons transferred from Mo atoms in metal layer to interfacial S atoms are increased whereas the electrons moved from Mo atoms in monolayer MoS₂ to interfacial S atoms are decreased compared with those in perfect Mo–MoS₂ top contact, as shown in Table 1. Accordingly, the electron density at the Mo–MoS₂ interface region increases whereas the electron density of Mo–S bond in monolayer MoS₂ reduces. Therefore, the difference between ρ_i and ρ_m for defective Mo–MoS₂ top contacts is decreased, which is consistent with above electron density analysis.

Table 1 Mulliken population of Mo atoms in metal layer as well as Mulliken population of Mo and S atoms in monolayer MoS₂ for perfect and defective Mo–MoS₂ top contacts. The atoms' serial number is indexed in Fig. 1(b) and (c)

	Mo metal layer	MoII	MoIV	SI	SIII
Perfect	5.958	5.921	5.921	6.073	6.073
Mo-vacancy	5.946	5.940	—	6.082	—
S-vacancy	5.943	—	5.941	—	6.119

---

4. Conclusion

The effects of intrinsic vacancies in monolayer MoS₂ on electronic structure and electronic properties of Mo–MoS₂ top contacts have been investigated using the first-principles plane-wave pseudopotential method based on density functional theory. Tunnel barrier, Schottky barrier, and electron density of perfect and defective Mo–MoS₂ top contacts were analyzed. Results show that the height and width of the tunnel barrier of Mo–MoS₂ top contacts are decreased when Mo- or S-vacancy is formed in monolayer MoS₂. Additionally, Schottky barriers are found to be 0.1 and 0 eV for perfect and defective Mo–MoS₂ top contacts, respectively. PDOS near Fermi level of defective Mo–MoS₂ top contacts are much higher than those of perfect one, implying the lower effective carrier mass for defective Mo–MoS₂ top contacts. Average electron density at interface of defective Mo–MoS₂ top contacts are increased compared with those in perfect Mo–MoS₂ top contact, showing that contact resistant and electron injection efficiency are further improved by vacancies. Moreover, Mo-vacancy in monolayer MoS₂ exhibits p-type Mo–MoS₂ top contact, whereas S-vacancy in monolayer MoS₂ shows n-type Mo–MoS₂ top contact.

Acknowledgements

We acknowledge the National Natural Science Foundation of China under grant no. 61376091, the National Aerospace Science Foundation of China under grant no. 2014ZF53070, the Fundamental Research Funds for the Central Universities under grant no. 3102014JCQ01033 and the 111 Project under grant no. B08040.

References

1. B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti and A. Kis, Nat. Nanotechnol., 2011, 6, 147.
2. H. S. Song, S. L. Li, L. Gao, Y. Xu, K. Ueno and J. Tang, et al., Nanoscale, 2013, 5, 9666.
3. B. Radisavljevic, M. B. Whitwick and A. Kis, Appl. Phys. Lett., 2012, 101, 043103.
4. H. Fang, S. Chuang, T. C. Chang, K. Takei, T. Takahashi and A. Javey, Nano Lett., 2012, 12, 3788.
5. J. Khan, C. M. Nolen, D. Teweldebrhan, D. Wickramaratne, R. K. Lake and A. A. Balandin, Appl. Phys. Lett., 2012, 100, 043109.
6. J. Mann, Q. Ma, P. M. Odenthal, M. Isarraaraz, D. Le and E. Preciado, et al., Adv. Mater., 2014, 26, 1399.
7. W. S. Hwang, M. Remskar, R. Yan, V. Protasenko, K. Tahy, S. D. Chae, P. Zhao, A. Konar, H. Xing, A. Seabaugh and D. Jena, et al., Appl. Phys. Lett., 2012, 101, 013107.
8. J. W. Jiang, Nanoscale, 2014, 6, 8326.
9. X. Liu, M. I. B. Utama, J. Lin, X. Gong, J. Zhang and Y. Zhao, Nano Lett., 2014, 14, 2419.
10. Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Coleman and M. S. Strano, Nat. Nanotechnol., 2012, 7, 699.
11. M. M. Benameur, B. Radisavljevic, J. S. Heron, H. Berger and A. Kis, Nanotechnology, 2011, 22, 125706.
12. K. F. Mak, C. Lee, J. Hone, J. Shan and T. F. Heinz, Phys. Rev. Lett., 2010, 105, 136805.
13. W. Bao, X. Cai, D. Kim, K. Sridhara and M. S. Fuhrer, Appl. Phys. Lett., 2013, 102, 042104.
14. Z. Yin, H. Li, H. Li, L. Jiang, Y. Shi and Y. Sun, et al., ACS Nano, 2012, 6, 74.
15. D. Sarkar, W. Liu, X. Xie, A. Anselmo, S. Mitragotri and K. Banerjee, ACS Nano, 2014, 8, 3992.
16. J. Kang, W. Liu, D. Sarkar, D. Jena and K. Banerjee, Phys. Rev. X, 2014, 4, 031005.
17. J. Kang, W. Liu and K. Banerjee, Appl. Phys. Lett., 2014, 104, 093106.
18. S. Das, H. Y. Chen, A. V. Penumatcha and J. Appenzeller, Nano Lett., 2013, 13, 100.
19. C. Gong, C. Huang, J. Miller, L. Cheng, Y. Hao and D. Cobden, et al., ACS Nano, 2013, 7, 11350.
20. I. Popov, G. Seifert and D. Tománek, Phys. Rev. Lett., 2012, 108, 156802.
21. N. A. Dhas and K. S. Suslick, J. Am. Chem. Soc., 2005, 127, 2368.
22. S. McDonnell, R. Addou, C. Buie, R. M. Wallace and C. L. Hinkle, ACS Nano, 2014, 8, 2880.
23. C. Ataca and S. Ciraci, J. Phys. Chem. C, 2011, 115, 13303.
24. C. Ataca, H. Sahin, E. Akturk and S. Ciraci, J. Phys. Chem. C, 2011, 115, 3934.
25. L. P. Feng, J. Su, S. Chen and Z. T. Liu, Mater. Chem. Phys., 2014, 148, 5.
26. D. Liu, Y. Guo, L. Fang and J. Robertson, Appl. Phys. Lett., 2013, 103, 183113.
27. L. P. Feng, J. Su and Z. T. Liu, J. Alloys Compd., 2014, 613, 122.
28. M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717.
29. T. Bučko, J. Hafner, S. Lebègue and J. G. Ángyán, J. Phys. Chem. A, 2010, 114, 11814.
30. N. Troullier and J. L. Martins, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 43, 1993.
31. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188.
32. J. Kang, D. Sarkar, W. Liu, D. Jena and K. Banerjee, IEEE International Electron Devices Meeting, 2012, pp. 407–410.
33. W. Kohn and L. J. Sham, Phys. Rev., 1965, 140, A1133.
34. J. Robertson, K. Xiong and S. J. Clark, Thin Solid Films, 2006, 496, 1–7.
35. J. J. Liu, X. L. Fu, S. F. Chen and Y. F. Zhu, Appl. Phys. Lett., 2011, 99, 191903.
36. R. Gillen and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84, 035125.
37. W. Li, C. F. J. Walther, A. Kuc and T. Heine, J. Chem. Theory Comput., 2013, 9, 2950–2958.
38. W. Zhou, X. Zou, S. Najmaei, Z. Liu, Y. Shi and J. Kong, et al., Nano Lett., 2013, 13, 2615.
39. S. Lebègue and O. Eriksson, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 115409.
40. A. Kumar and P. K. Ahluwalia, Tunable Electronic and Dielectric Properties of Molybdenum Disulfide, Springer International Publishing, 2014, vol. 21, p. 53.
41. Y. Jing, X. Tan, Z. Zhou and P. Shen, J. Mater. Chem. A, 2014, 2, 16892.
42. A. Kumar and P. K. Ahluwalia, Phys. B, 2013, 419, 66–75.
43. Q. Yue, J. Kang, Z. Shao, X. Zhang, S. Chang and G. Wang, Phys. Lett. A, 2012, 376, 1166.
44. A. Garcia and J. E. Northrup, Phys. Rev. Lett., 1995, 74, 1131.
45. S. Poykko, M. J. Puska and R. M. Nieminen, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 53, 3813.
46. M. Bokdam, G. Brocks, M. I. Katsnelson and P. J. Kelly, 2014, arXiv preprint arXiv:1401.6440.
47. W. Liu, J. Kang, D. Sarkar, Y. Khatami, D. Jena and K. Banerjee, Nano Lett., 2013, 13, 1983.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra15218b

---