Exciton states in monolayer MoSe: impact on interband transitions
Abstract
We combine linear and nonlinear optical spectroscopy at 4 K with ab initio calculations to study the electronic bandstructure of MoSe monolayers. In 1photon photoluminescence excitation (PLE) and reflectivity we measure a separation between the A and Bexciton emission of 220 meV. In 2photon PLE we detect for the A and Bexciton the 2p state 180 meV above the respective 1s state. In second harmonic generation (SHG) spectroscopy we record an enhancement by more than 2 orders of magnitude of the SHG signal at resonances of the charged exciton and the 1s and 2p neutral A and Bexciton. Our postDensity Functional Theory calculations show in the conduction band along the direction a local minimum that is energetically and in kspace close to the global minimum at the Kpoint. This has a potentially strong impact on the polarization and energy of the excitonic states that govern the interband transitions and marks an important difference to MoS and WSe monolayers.
I Introduction
Monolayers (MLs) of the transition metal dichalcogenides (TMDCs) MoS, MoSe, WS and WSe (abbreviated MX) are semiconductors with a direct bandgap in the visible region Mak et al. (2010); Splendiani et al. (2010); Zhang et al. (2014). MX MLs are promising materials for (opto)electronics Geim and Grigorieva (2013); Butler et al. (2013); Zhao et al. (2013); Mak et al. (2010); Splendiani et al. (2010); Radisavljevic et al. (2011); Ross et al. (2014), nonlinear optics Kumar et al. (2013); Zeng et al. (2013); Li et al. (2013); Yin et al. (2014); Malard et al. (2013); Trolle et al. (2014); Wang et al. (2015a); Grüning and Attaccalite (2014) and for exploring electron kvalley physics Xiao et al. (2010); Mak et al. (2014); Xu et al. (2014). The MX ML materials share common characteristics: (i) Their optical properties are dominated by excitons, strongly Coulombbound electron hole pairs Cheiwchanchamnangij and
Lambrecht (2012); Komsa and Krasheninnikov (2012); Ross et al. (2013); Song and Dery (2013); He et al. (2014); Ugeda et al. (2014); Chernikov et al. (2014); Ye et al. (2014); Wang et al. (2015a); Klots et al. (2014); Zhu et al. (2014); Hanbicki et al. (2015); Hill et al. (2015). (ii) In these TMDC MLs crystal inversion symmetry breaking together with the strong spinorbit (SO) interaction leads to a coupling of carrier spin and kspace valley physics, i.e., the circular polarization ( or ) of the absorbed or emitted photon can be directly associated with selective carrier excitation in one of the two nonequivalent K valleys (K or K, respectively) Xiao et al. (2012); Cao et al. (2012); Mak et al. (2012); Sallen et al. (2012); Kioseoglou et al. (2012); Jones et al. (2013); Mak et al. (2014).
Using nonresonant laser excitation in ML MoS Mak et al. (2012); Zeng et al. (2012); Kioseoglou et al. (2012); Sallen et al. (2012), WSe Jones et al. (2013); Wang et al. (2015a) and WS Zhu et al. (2014) high values () for the circular polarization of the stationary photoluminescence (PL) corresponding to successful valley polarization have been reported. In contrast, the polarization reported for the promising material MoSe Singh et al. (2014); Kumar et al. (2014); Ross et al. (2013) under similar experimental conditions is surprisingly very low () MacNeill et al. (2015); Li et al. (2014); Wang et al. (2015b). Experiments combing PL excitation (PLE) and time resolved PL suggest either ultra fast polarization relaxation in MoSe in the subpicosecond range or inefficient optical polarization generation due to anomalies in the bandstructure to be at the origin of this low valley polarization Wang et al. (2015b).
Our target is to investigate how an eventual bandstructure anomaly (local extremum), can influence the nature of the optical transitions. To this aim we combine linear and nonlinear optical spectroscopy at T=4 K with calculations beyond standard Density Functional Theory (DFT) to study the electronic bandstructure of MoSe MLs. In 2photon PLE we detect 180 meV above the 1s state the 2p state of the Aexciton, well separated in energy from the 1s Bexciton emission recorded in both reflectivity and 1photon PLE. The energy position of the neutral and charged exciton transitions is determined using second harmonic generation (SHG) spectroscopy, as the SHG signal is enhanced by more than 2 orders at these particular resonances, important for applications in nonlinear optics.
The strong excitonic effects that dominate the 1 and 2photon spectroscopy results in the investigated MoSe MLs have important consequences not just in terms of the transition energy Ugeda et al. (2014). Due to the small exciton Bohr radius, the extension in kspace is considerable and the exciton wavefunction will include contributions from states far away from the Kpoint of the Brillouin zone where the direct free carrier (electronic) bandgap of the MX MLs is situated. The position in energy and in kspace of the adjacent local minima in the conduction band (CB) and maxima in the valence band (VB), respectively can be very different from one MX material to another Kormányos et al. (2013); Kormanyos et al. (2015). Our DFTGW calculations show in the CB of ML MoSe a local minimum that is energetically and in kspace close to the global minimum at the Kpoint Bradley et al. (2015). We evaluate how the proximity of this local minimum along the direction can influence the polarization and energy of the excitonic states that govern the interband transitions, marking an important difference compared to MoS and WSe monolayers.
Ii Spectroscopy of Exciton States
Similar to the hydrogen model, an electron hole pair in the TMDC ML interacts through an attractive Coulomb potential and will form below the gap a series of excitonic Rydberglike states with definite parity, where states are of even, and states are of odd parity (n is an integer) Kli (). The exciton binding energy of the order of 0.5 eV in ML MoSe can be estimated by determining the free carrier bandgap in scanning tunnelling spectroscopy (STS) and subtracting the neutral exciton energy (corresponds to the 1s Aexciton state), see Ugeda et al. (2014). Here we aim to uncover in addition to the 1s exciton ground state also higher excited exciton states, which were shown to dominate the linear and nonlinear optical response in MoS and WSe MLs Wang et al. (2015a); Carvalho et al. (2015). In Fig. 1b we see a typical timeintegrated PL spectrum at T=4 K for ML MoSe with two prominent peaks (see Appendix for information on samples and the experimental setup). The higher energy peak (FWHM=10 meV) at 1.667 eV has previously been attributed to the neutral Aexciton X Ross et al. (2013); Li et al. (2014); MacNeill et al. (2015). At 1.633 eV we record the trion emission (T) corresponding to a binding energy of 34 meV, in agreement with previous measurements Ross et al. (2013); Li et al. (2014); MacNeill et al. (2015); Wang et al. (2015b).
ii.1 OnePhoton PLE experiments
Here we detect the intensity of the X PL emission as a function of the laser energy, shown in Fig. 1b. We find a clear maximum at 1.885 eV, which we attribute to the 1s Bexciton state. In order to confirm this, we have performed reflectivity measurements using a white light source. We find in reflectivity two prominent transitions in Fig. 1c, one at 1.667 eV, which corresponds exactly to the neutral Aexciton emission energy measured in PL. This indicates negligible localization of excitons in this sample (no Stokes shift). The second transition in reflectivity at 1.885 eV gives the position of the 1s Bexciton. The measured difference between 1s A and Bexciton is meV determined from both PL and reflectivity. From our DFT calculations (see section III for details) we obtain a spin splitting of meV in the valence band. This is in very good agreement with the measured value of 180 meV from angleresolved photoemission spectroscopy (ARPES) measurements at T=40 K Zhang et al. (2014). The CB spin splitting is predicted to have the opposite sign compared to the VB, as sketched in Fig. 1a Liu et al. (2013); Kosmider et al. (2013); Kormanyos et al. (2015). We obtain from our DFT calculations a CB spin splitting of meV. The measured difference of 220 meV is in close agreement with the calculated energy difference taking into account the VB and CB spin splitting of 183+30=213 meV. In addition, we find in DFT calculations in section III that the effective masses in CB and VB are slightly different for different spin states, and hence also the exciton binding energies of A and Bexcitons are expected to differ and contribute to the measured value of .
ii.2 TwoPhoton PLE experiments
An experimental signature of excited exciton states, in analogy with the and states of the hydrogen atom, have not been reported yet for MoSe MLs, to the best of our knowledge. The excited exciton states and the exact exciton binding energy in MX MLs are currently debated in the literature for WS Chernikov et al. (2014); Ye et al. (2014); Zhu et al. (2014); Hanbicki et al. (2015), MoS Klots et al. (2014); Hill et al. (2015) and WSe He et al. (2014); Wang et al. (2015a); Hanbicki et al. (2015), with differences in the reported for the same material by factor of 2, so additional experiments are important. In our 2photon PLE measurements we can directly address exciton states with psymmetry and not with ssymmetry as in the 1photon PLE and reflectivity. In the 2photon PLE of Fig. 1b we find a well defined peak at 1.844 eV, which we assign to the 2p Aexciton transition, not obscured by the 1s Bexciton, which is parity forbidden here. We measure an energy difference 1s2p of 180 meV. Using a binding energy of the order of eV Ugeda et al. (2014) in a simple 2D hydrogen model Kli (), the energy separation 1s2s,2p is expected to be meV, much larger than our measured value of 180 meV. This shows for ML MoSe a strong deviation from a simple hydrogenic series, observed also in ML WS and WSe, very likely due to a strong variation of the effective dielectric constant as a function of the spatial extension of the exciton state Cudazzo et al. (2011); Deslippe et al. (2009). Interestingly we observe a second, clear maximum at 2.06 eV, about 175 meV above the Bexciton 1s state. We tentatively assign this peak in 2photon PLE to the 2p Bexciton state.
ii.3 Second Harmonic Generation Spectroscopy
In addition to the 2photon PLE signal we can also plot the intensity of the SHG signal as a function of Laser energy in Fig. 2. The SHG spectroscopy results in this ML material without crystal inversion centre also show clear resonances at the exciton energies. Interestingly we observe a strong SHG signal at the transition energy of the charged exciton (trion). This hints at a substantial density of states of this complex.
The SHG signal at the 1s Bexciton is about 120 times higher than in the region between 1.7 and 1.8 eV. This finding shows that excitonic resonances dominate the nonlinear optical response of ML MoSe. The SHG signal is nonzero for all laser energies, the SHG peaks outside the exciton resonances between 1.7 and 1.8 eV are not clearly visible on the linear scale used in Fig. 2. We confirm the energy position of the 1s and 2p exciton levels A and B. Note that resonances of 1s exciton states in SHG are forbidden if strict electric dipole selection rules apply. The prominent 1s features in our SHG spectrum in Fig. 2 can come from the interplay of electric dipole with magnetic dipole transitions Wang et al. (2015a); Lafrentz et al. (2013). In addition strict electric dipole selection rules could be slightly relaxed if the overall symmetry of the crystal is lowered by extrinsic effects such as the substrate.
Iii Electronic BandStructure calculations
In the experiments on ML MoSe we find sharp excitonic features that are well defined as for the related ML material WSe Wang et al. (2015a). Despite similar exciton binding energy and excited state spectrum in ML WSe and ML MoSe the optical valley polarization using nonresonant laser excitation can be generated in the former but not in the latter Li et al. (2014); MacNeill et al. (2015); Wang et al. (2015b). One reason could be an anomaly in the bandstructure, which we aim to uncover in band structure calculations for MoSe. For comparison and to validate our computational approach we have performed in parallel calculations with the more thoroughly studied ML MoS. The bandstructure of ML MoSe and MoS can be compared in Fig. 3, where striking differences in both valence and conduction band appear. We extract the effective carrier masses, the band gaps and the exciton binding energies. We discuss the most important features for optical transitions in detail, in particular the competition between direct and indirect exciton states in MoSe MLs.
iii.1 QuasiParticle Band Structure
The comparison of quasiparticle band structure calculated by DFT (see Appendix for computational details) of ML MoS and MoSe provides interesting insights: when replacing S with Se, the spinsplitting is enhanced. For MoS the valence and conduction state the spinsplittings at the K point are meV and meV, respectively. For ML MoSe we find meV and meV. The direct band gap values (neglecting excitonic effects) at K, depicted by arrows in the left panels of Fig. 3, are 2.13 and 2.31 eV for MoSe and MoS, respectively.
The latter value is lower than previous theoretical studies as energies from 2.41 to 2.97 eV have been reported Komsa and Krasheninnikov (2012); Cheiwchanchamnangij and
Lambrecht (2012); Ramasubramaniam (2012); MolinaSánchez
et al. (2013); Qiu et al. (2013); Shi et al. (2013); Klots et al. (2014). It is important to take into account that the uncertainty in this type of band gap calculations is in the hundreds of meV range, depending on the computational settings. The obtained results critically depend on the choice of the methods used (selfconsistent or partially selfconsistent scheme), the number of unoccupied states included, the vacuum height and the kpoint sampling.
Interestingly using exactly the same computational settings, our MoSe direct band gap estimate is closer to recent reports: 2.33 eV for a calculation in Ref.Horzum et al., 2013, when Ugeda et al Ugeda et al. (2014) propose 2.26 eV for a free standing MoSe ML. On the experimental side, using STS techniques the band gap of MoSe ML is 2.18 0.04 eV Bradley et al. (2015) but it includes substrate screening effects.
If one compares the full bandstructure of the two material systems another striking difference appears: in MoS ML, the topmost valence bands in and K are only separated by 65 meV. This feature appears to be essential in the mechanism of direct to indirect band gap transition and its evolution with layer thickness (1ML, 2ML …) Jin et al. (2013) and is expected to have a strong impact on the optically generated valley polarization as a function of laser energy Kormányos et al. (2013); Lagarde et al. (2014). The same trend for the gap evolution with layer thickness applies to MoSe Zhang et al. (2014), albeit with a much larger to K energy separation of 485 meV for 1ML.
For ML MoSe our calculations show a remarkable anomaly in the CB, which could play a key role for the valley polarization dynamics: In ML MoSe there is a small energy difference of only 46 meV between the CB minimum in K and in , as can be clearly seen in Fig. 3. The corresponding value for MoS is 345 meV and therefore not in competition with direct optical transitions at the K points. We note another important difference: The position along the K line for MoS lies exactly at the coordinates. In contrast, for MoSe it is slightly shifted towards the K point. Interestingly the spinup and spindown minima are not positioned at the same kvalue. Note that the appearance of this local minimum has also been predicted recently in Bradley et al. (2015).
In the following we discuss how the proximity of the CB minimum at can potentially influence the optical transitions. Excitons with large binding energies in ML MoSe have a very small Bohr radius and as a result extend in kspace well beyond the Kpoint (see Fig. 3). To allow for a more quantitative discussion, we need to extract the effective carrier masses from our calculated band structure. At VB and CB extrema, effective masses for holes and electrons can be extracted with a simple quadratic fitting procedure. The extracted mass values will depend on the selected kvalue range around the extrema. For consistency, we have used the interval , with being defined as , where is the Bohr radius of the exciton ground state in a WannierMott picture of electronhole pair Ye et al. (2014). The computed effective mass values are given in Table 1. For MoS, the and values are only slightly different from previous studies Cheiwchanchamnangij and
Lambrecht (2012); Ramasubramaniam (2012); Shi et al. (2013), the difference being reduced in the electronhole pair effective mass values. There is a certain spread in values, especially for the MoS hole effective mass with the value provided in Ref. Qiu et al., 2013, or for MoSe values Horzum et al. (2013). We obtain smaller values, but we attribute these differences to the choice of the extension in kspace around the extrema. Indeed, if one takes a fitting interval of length the effective masses are significantly increased by 30% in both material systems. Interestingly spinup and spindown excitons have different effective masses at K: 0.21 vs 0.24 for and respectively for MoS, 0.25 and 0.28 in MoSe, due to larger values of the spindown and , in both systems. This confirms that the exciton binding energies for A and Bexcitons should be slightly different, which will in turn influence the energy difference between the corresponding optical transitions measured in Fig. 1.
On top of calculations, exciton binding energies , given in Table 1, have been extracted from the imaginary part of transverse dielectric constant after solving the BetheSalpeter Equation (BSE). Our 0.58 eV estimate is in reasonable agreement with previous theoretical studies for the MoS Aexciton binding energy: values from 0.55 Hüser et al. (2013), around 0.6 Shi et al. (2013); Klots et al. (2014), or around 0.9 Cheiwchanchamnangij and
Lambrecht (2012); Qiu et al. (2013) and up to 1.1 eVRamasubramaniam (2012); Komsa and Krasheninnikov (2012) have been reported. Please note the computational details are extremely different for each of these values. It is therefore difficult to extract precise trends. Alternatively, the spread in values should rather be used as a reasonable error bar for this type of calculations. For MoSe theoretical values for of 0.9 Ramasubramaniam (2012), 0.78 Komsa and Krasheninnikov (2012) and 0.65 eV Ugeda et al. (2014) can be found in the recent literature. This indicates that the calculated exciton binding energy in ML MoSe is smaller than in MoS, as observed experimentally.
in  eV  eV    Å  

MoS  2.31  0.58  0.40  0.46  4.47  2.77 
MoSe  2.13  0.51  0.49  0.52  5.17  2.40 

iii.2 Exciton States
Our target is now to estimate how far the exciton state extends in kspace around the K point. This allows us to estimate the possible impact of the proximity of the local CB minimum at on the optical transitions and the comparatively weak valley polarization in ML MoSe. This order of magnitude discussion is graphically represented in the right column of Fig. 3. Starting from the standard 2D WannierMott model Kli (), the exciton ground state energy writes simply as:
(1) 
with . An estimate of the corresponding exciton Bohr radius is thus given by
(2) 
Combining these equations allows for the calculation of the relative dielectric constant, knowing the binding energy and the reduced mass of the electronhole pair, extracted from +BSE
calculations. It yields values of 4.47 and 5.17 for MoS and MoSe, respectively. The main differences from previous theoretical results arise from our choice of a 2D model, the use of our own computed and .
These values are usually larger than the estimates found in Ref. Cheiwchanchamnangij and
Lambrecht, 2012; Ramasubramaniam, 2012 using (3.44) or by our own direct calculation of the static dielectric constant at the DFT level including local field effect (2.61) for MoS. Using the same approximation we find 2.80 for MoSe in comparison with 5.17 of Table 1.
Continuing our discussion based on a simple hydrogenlike model, the exciton groundstate wave function can be written as:
(3) 
where and is the coordinate of the relative electronhole motion. The corresponding Fourier transform of this ground state provides a rough estimation of how far around the K point the exciton state is spread in kspace.
(4) 
Our representation of for ML MoSe in Fig. 3 suggest the following scenario: The close proximity of the conduction band minimum could be at the origin of the low valley polarization degree of the Aexciton 1s state luminescence, as contributions to the exciton wavefunctions away from the K points do not obey the strict chiral valley selections rules. It is important to underline in Fig. 3 the striking difference with ML MoS, where the exciton state is mainly build with K electronic states as the contribution from the states around the point at much higher energy is negligible.
Iv Conclusions
We probe the exciton states in ML MoSe in 1 and 2photon PLE. We see a clear signature of the 2p state of the A and Bexciton about 180 meV above the respective 1s exciton state in 2photon PLE. Our postDFT calculations reveal in the MoSe ML conduction band a local minimum at the point only 46 meV above the global minimum at the Kpoints. As excitons with large binding energies ( eV) involve kstates far away from the Kpoint, the contribution of states around the point to optical transitions is possible. As here the chiral optical selections rules are not applicable, these states could contribute to the overall low valley polarization reported for MoSe Li et al. (2014); MacNeill et al. (2015); Wang et al. (2015b). To further test this hypothesis, experiments that modify the band structure via strain tuning He et al. (2013); Zhu et al. (2013) would provide useful information. Also a systematic study of valley polarization in MoSSe alloy monolayers Mann et al. (2014) will help clarifying, why MoS shows strong and MoSe shows very weak valley polarization in nonresonant PL experiments.
Acknowledgements
We thank Misha Glazov for fruitful discussions and acknowledge funding from ERC Grant No. 306719, ANR MoS2ValleyControl and Programme Investissements d’Avenir ANR11IDEX000202, reference ANR10LABX0037NEXT. I. C. Gerber thanks the CALMIP initiative for the generous allocation of computational times, through the project p0812, as well as the GENCICINES and GENCICCRT for the grant x2014096649. I. C. Gerber also acknowledges the CNRS for financial support.
V Appendix
v.1 Samples and Experimental Setup
MoSe ML flakes are obtained by micromechanical cleavage of a bulk MoSe crystal on an SiO/Si substrates using viscoelastic stamping CastellanosGomez et al. (2014). The ML region is identified by optical contrast and very clearly in PL spectroscopy. Experiments at T=4 K are carried out in a confocal microscope optimized for polarized PL experiments Wang et al. (2014). The MoSe ML is excited by picosecond pulses generated by a tunable frequencydoubled optical parametric oscillator (OPO) synchronously pumped by a modelocked Ti:Sa laser. The typical pulse and spectral width are 1.6 ps and 3 meV respectively; the repetition rate is 80 MHz. The laser average power is tunable from 2 to 200 W. The detection spot diameter is m, i.e. considerably smaller than the ML diameter. For time integrated experiments, the PL emission is dispersed in a spectrometer and detected with a SiCCD camera.
v.2 Computational Details
The atomic structures, the quasiparticle band structures and optical spectra are obtained from DFT calculations using the VASP package Kresse and Hafner (1993); Kresse and Furthmüller (1996). PBE functional Perdew et al. (1996) is used as approximation of the exchangecorrelation term. It uses the planeaugmented wave scheme Blöchl (1994); Kresse and Joubert (1999) to treat core electrons. Fourteen and six electrons for Mo and S, Se respectively are explicitly included in the valence. All atoms are allowed to relax with a force convergence criterion below eV/Å. After primitive cell relaxation, the optimized lattice parameters are 3.22 and 3.32 Å for MoS and MoSe respectively, these values being in good agreement ( 1%) with previous studiesRamasubramaniam (2012); Horzum et al. (2013) and slightly larger than the bulk experimental values. To sample the Brillouin zone a grid of 12121 kpoints has been used, in conjunction with a vacuum height of 17 Å, to take benefit of error’s cancellation in the band gap estimates Hüser et al. (2013). This provides exciton binding energies in reasonable agreement with experiments as suggested in different works Klots et al. (2014); MolinaSánchez et al. (2013). A gaussian smearing with a width of 0.05 eV is used for partial occupancies, when a tight electronic minimization tolerance of eV is set to determine with a good precision the corresponding derivative of the orbitals with respect to needed in quasiparticle band structure calculations. Spinorbit coupling was also included nonselfconsistently to determine eigenvalues and wave functions as input for the fullfrequencydependent calculations Shishkin and Kresse (2006) performed at the level. The total number of states included in the procedure is set to 600, after a careful check of the direct band gap convergence smaller than 0.1 eV. We have used the WANNIER90 code Mostofi et al. (2008) and the VASP2WANNIER90 interface Franchini et al. (2012) to interpolate the band structures on a finer grid. Optical absorption spectra have been calculated using BetheSalpeter Equation in the TammDancoff approximation, including the six highest valence bands and the eight lowest conduction bands Wang et al. (2015a).
References
 Mak et al. (2010) K. F. Mak, C. Lee, J. Hone, J. Shan, and T. F. Heinz, Phys. Rev. Lett. 105, 136805 (2010).
 Splendiani et al. (2010) A. Splendiani, L. Sun, Y. Zhang, T. Li, J. Kim, C.Y. Chim, G. Galli, and F. Wang, Nano Letters 10, 1271 (2010).
 Zhang et al. (2014) Y. Zhang, T.R. Chang, B. Zhou, Y.T. Cui, H. Yan, Z. Liu, F. Schmitt, J. Lee, R. Moore, Y. Chen, et al., Nature Nanotechnology 9, 111 (2014).
 Geim and Grigorieva (2013) A. K. Geim and I. V. Grigorieva, Nature 499, 419 (2013).
 Butler et al. (2013) S. Z. Butler, S. M. Hollen, L. Cao, Y. Cui, J. A. Gupta, H. R. Gutiérrez, T. F. Heinz, S. S. Hong, J. Huang, A. F. Ismach, et al., ACS Nano 7, 2898 (2013).
 Zhao et al. (2013) W. Zhao, Z. Ghorannevis, L. Chu, M. Toh, C. Kloc, P.H. Tan, and G. Eda, ACS Nano 7, 791 (2013).
 Radisavljevic et al. (2011) B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti, and A. Kis, Nature. Nanotech. 6, 147 (2011).
 Ross et al. (2014) J. S. Ross, P. Klement, A. M. Jones, N. J. Ghimire, J. Yan, D. G. Mandrus, T. Taniguchi, K. Watanabe, K. Kitamura, W. Yao, et al., Nature Nanotechnology 9, 268 (2014).
 Kumar et al. (2013) N. Kumar, S. Najmaei, Q. Cui, F. Ceballos, P. M. Ajayan, J. Lou, and H. Zhao, Phys. Rev. B 87, 161403 (2013).
 Zeng et al. (2013) H. Zeng, G.B. Liu, J. Dai, Y. Yan, B. Zhu, R. He, L. Xie, S. Xu, X. Chen, W. Yao, et al., Scientific Reports 3, 1608 (2013).
 Li et al. (2013) Y. Li, Y. Rao, K. F. Mak, Y. You, S. Wang, C. R. Dean, and T. F. Heinz, Nano Letters 13, 3329 (2013).
 Yin et al. (2014) X. Yin, Z. Ye, D. A. Chenet, Y. Ye, K. O’Brien, J. C. Hone, and X. Zhang, Science 344, 488 (2014).
 Malard et al. (2013) L. M. Malard, T. V. Alencar, A. P. M. Barboza, K. F. Mak, and A. M. de Paula, Phys. Rev. B 87, 201401 (2013).
 Trolle et al. (2014) M. L. Trolle, G. Seifert, and T. G. Pedersen, Phys. Rev. B 89, 235410 (2014).
 Wang et al. (2015a) G. Wang, X. Marie, I. Gerber, T. Amand, D. Lagarde, L. Bouet, M. Vidal, A. Balocchi, and B. Urbaszek, Phys. Rev. Lett. 114, 097403 (2015a).
 Grüning and Attaccalite (2014) M. Grüning and C. Attaccalite, Phys. Rev. B 89, 081102 (2014).
 Xiao et al. (2010) D. Xiao, M.C. Chang, and Q. Niu, Rev. Mod. Phys. 82, 1959 (2010).
 Mak et al. (2014) K. F. Mak, K. L. McGill, J. Park, and P. L. McEuen, Science 344, 1489 (2014).
 Xu et al. (2014) X. Xu, D. Xiao, T. F. Heinz, and W. Yao, Nature Physics 10, 343 (2014).
 Cheiwchanchamnangij and Lambrecht (2012) T. Cheiwchanchamnangij and W. R. L. Lambrecht, Phys. Rev. B 85, 205302 (2012).
 Komsa and Krasheninnikov (2012) H.P. Komsa and A. V. Krasheninnikov, Phys. Rev. B 86, 241201 (2012).
 Ross et al. (2013) J. S. Ross, S. Wu, H. Yu, N. J. Ghimire, A. M. Jones, G. Aivazian, J. Yan, D. G. Mandrus, D. Xiao, W. Yao, et al., Nature Communications 4, 1474 (2013).
 Song and Dery (2013) Y. Song and H. Dery, Phys. Rev. Lett. 111, 026601 (2013).
 He et al. (2014) K. He, N. Kumar, L. Zhao, Z. Wang, K. F. Mak, H. Zhao, and J. Shan, Phys. Rev. Lett. 113, 026803 (2014).
 Ugeda et al. (2014) M. M. Ugeda, A. J. Bradley, S.F. Shi, F. H. da Jornada, Y. Zhang, D. Y. Qiu, S.K. Mo, Z. Hussain, Z.X. Shen, F. Wang, et al., Nature Materials 13, 1091 (2014).
 Chernikov et al. (2014) A. Chernikov, T. C. Berkelbach, H. M. Hill, A. Rigosi, Y. Li, O. B. Aslan, D. R. Reichman, M. S. Hybertsen, and T. F. Heinz, Phys. Rev. Lett. 113, 076802 (2014).
 Ye et al. (2014) Z. Ye, T. Cao, K. O’Brien, H. Zhu, X. Yin, Y. Wang, S. G. Louie, and X. Zhang, Nature 513, 214 (2014).
 Klots et al. (2014) A. R. Klots, A. K. M. Newaz, B. Wang, D. Prasai, H. Krzyzanowska, D. Caudel, N. J. Ghimire, J. Yan, B. L. Ivanov, K. A. Velizhanin, et al., Scientific Reports 4, 6608 (2014).
 Zhu et al. (2014) B. Zhu, X. Chen, and X. Cui, ArXiv eprints (2014), eprint 1403.5108.
 Hanbicki et al. (2015) A. Hanbicki, M. Currie, G. Kioseoglou, A. Friedman, and B. Jonker, Solid State Communications 203, 16 (2015).
 Hill et al. (2015) H. M. Hill, A. F. Rigosi, C. Roquelet, A. Chernikov, T. C. Berkelbach, D. R. Reichman, M. S. Hybertsen, L. E. Brus, and T. F. Heinz, Nano Letters Article ASAP, null (2015).
 Xiao et al. (2012) D. Xiao, G.B. Liu, W. Feng, X. Xu, and W. Yao, Phys. Rev. Lett. 108, 196802 (2012).
 Cao et al. (2012) T. Cao, G. Wang, W. Han, H. Ye, C. Zhu, J. Shi, Q. Niu, P. Tan, E. Wang, B. Liu, et al., Nature Communications 3, 887 (2012).
 Mak et al. (2012) K. F. Mak, K. He, J. Shan, and T. F. Heinz, Nat. Nanotechnol. 7, 494 (2012).
 Sallen et al. (2012) G. Sallen, L. Bouet, X. Marie, G. Wang, C. R. Zhu, W. P. Han, Y. Lu, P. H. Tan, T. Amand, B. L. Liu, et al., Phys. Rev. B 86, 081301 (2012).
 Kioseoglou et al. (2012) G. Kioseoglou, A. T. Hanbicki, M. Currie, A. L. Friedman, D. Gunlycke, and B. T. Jonker, Applied Physics Letters 101, 221907 (pages 4) (2012).
 Jones et al. (2013) A. M. Jones, H. Yu, N. J. Ghimire, S. Wu, G. Aivazian, J. S. Ross, B. Zhao, J. Yan, D. G. Mandrus, D. Xiao, et al., Nat. Nanotechnol. 8, 634 (2013).
 Zeng et al. (2012) H. Zeng, J. Dai, W. Yao, D. Xiao, and X. Cui, Nat. Nanotechnol. 7, 490 (2012).
 Zhu et al. (2014) B. Zhu, H. Zeng, J. Dai, Z. Gong, and X. Cui, Proceedings of the National Academy of Sciences 111, 11606 (2014).
 Singh et al. (2014) A. Singh, G. Moody, S. Wu, Y. Wu, N. J. Ghimire, J. Yan, D. G. Mandrus, X. Xu, and X. Li, Phys. Rev. Lett. 112, 216804 (2014).
 Kumar et al. (2014) N. Kumar, J. He, D. He, Y. Wang, and H. Zhao, Nanoscale 6, 12690 (2014).
 MacNeill et al. (2015) D. MacNeill, C. Heikes, K. F. Mak, Z. Anderson, A. Kormányos, V. Zólyomi, J. Park, and D. C. Ralph, Phys. Rev. Lett. 114, 037401 (2015).
 Li et al. (2014) Y. Li, J. Ludwig, T. Low, A. Chernikov, X. Cui, G. Arefe, Y. D. Kim, A. M. van der Zande, A. Rigosi, H. M. Hill, et al., Phys. Rev. Lett. 113, 266804 (2014).
 Wang et al. (2015b) G. Wang, E. Palleau, T. Amand, S. Tongay, X. Marie, and B. Urbaszek, Applied Physics Letters 106, 112101 (2015b).
 Kormányos et al. (2013) A. Kormányos, V. Zólyomi, N. D. Drummond, P. Rakyta, G. Burkard, and V. I. Fal’ko, Phys. Rev. B 88, 045416 (2013).
 Kormanyos et al. (2015) A. Kormanyos, G. Burkard, M. Gmitra, J. Fabian, V. Zolyomi, N. D. Drummond, and V. Fal’ko, 2D Materials 2, 022001 (2015).
 Bradley et al. (2015) A. J. Bradley, M. M. Ugeda, F. H. da Jornada, D. Y. Qiu, W. Ruan, Y. Zhang, S. Wickenburg, A. Riss, J. Lu, S.K. Mo, et al., Nano Letters 15, 2594 (2015).
 (48) C. F. Klingshirn, Semiconductor Optics (SpringerVerlag, Berlin, Germany) 2006.
 Carvalho et al. (2015) B. R. Carvalho, L. M. Malard, J. M. Alves, C. Fantini, and M. A. Pimenta, Phys. Rev. Lett. 114, 136403 (2015).
 Liu et al. (2013) G.B. Liu, W.Y. Shan, Y. Yao, W. Yao, and D. Xiao, Phys. Rev. B 88, 085433 (2013).
 Kosmider et al. (2013) K. Kosmider, J. W. González, and J. FernándezRossier, Phys. Rev. B 88, 245436 (2013).
 Cudazzo et al. (2011) P. Cudazzo, I. V. Tokatly, and A. Rubio, Phys. Rev. B 84, 085406 (2011).
 Deslippe et al. (2009) J. Deslippe, M. Dipoppa, D. Prendergast, M. V. O. Moutinho, R. B. Capaz, and S. G. Louie, Nano Letters 9, 1330 (2009).
 Lafrentz et al. (2013) M. Lafrentz, D. Brunne, A. V. Rodina, V. V. Pavlov, R. V. Pisarev, D. R. Yakovlev, A. Bakin, and M. Bayer, Phys. Rev. B 88, 235207 (2013).
 Ramasubramaniam (2012) A. Ramasubramaniam, Phys. Rev. B 86, 115409 (2012).
 MolinaSánchez et al. (2013) A. MolinaSánchez, D. Sangalli, K. Hummer, A. Marini, and L. Wirtz, Physical Review B 88, 045412 (2013).
 Qiu et al. (2013) D. Y. Qiu, F. H. da Jornada, and S. G. Louie, Phys. Rev. Lett. 111, 216805 (2013).
 Shi et al. (2013) H. Shi, H. Pan, Y.W. Zhang, and B. I. Yakobson, Phys. Rev. B 87, 155304 (2013).
 Horzum et al. (2013) S. Horzum, H. Sahin, S. Cahangirov, P. Cudazzo, A. Rubio, T. Serin, and F. M. Peeters, Phys. Rev. B 87, 125415 (2013).
 Jin et al. (2013) W. Jin, P.C. Yeh, N. Zaki, D. Zhang, J. T. Sadowski, A. AlMahboob, A. M. van der Zande, D. A. Chenet, J. I. Dadap, I. P. Herman, et al., Phys. Rev. Lett. 111, 106801 (2013).
 Lagarde et al. (2014) D. Lagarde, L. Bouet, X. Marie, C. R. Zhu, B. L. Liu, T. Amand, P. H. Tan, and B. Urbaszek, Phys. Rev. Lett. 112, 047401 (2014).
 Hüser et al. (2013) F. Hüser, T. Olsen, and K. S. Thygesen, Phys. Rev. B 88, 245309 (2013).
 He et al. (2013) K. He, C. Poole, K. F. Mak, and J. Shan, Nano letters 13, 2931 (2013).
 Zhu et al. (2013) C. R. Zhu, G. Wang, B. L. Liu, X. Marie, X. F. Qiao, X. Zhang, X. X. Wu, H. Fan, P. H. Tan, T. Amand, et al., Phys. Rev. B 88, 121301 (2013).
 Mann et al. (2014) J. Mann, Q. Ma, P. M. Odenthal, M. Isarraraz, D. Le, E. Preciado, D. Barroso, K. Yamaguchi, G. von Son Palacio, A. Nguyen, et al., Advanced Materials 26, 1399 (2014).
 CastellanosGomez et al. (2014) A. CastellanosGomez, M. Buscema, R. Molenaar, V. Singh, L. Janssen, H. S. J. van der Zant, and G. A. Steele, 2D Materials 1, 011002 (2014).
 Wang et al. (2014) G. Wang, L. Bouet, D. Lagarde, M. Vidal, A. Balocchi, T. Amand, X. Marie, and B. Urbaszek, Phys. Rev. B 90, 075413 (2014).
 Kresse and Hafner (1993) G. Kresse and J. Hafner, Phys. Rev. B 47, 558 (1993).
 Kresse and Furthmüller (1996) G. Kresse and J. Furthmüller, Comput. Mater. Sci. 6, 15 (1996).
 Perdew et al. (1996) J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996).
 Blöchl (1994) P. E. Blöchl, Phys. Rev. B 50, 17953 (1994).
 Kresse and Joubert (1999) G. Kresse and D. Joubert, Physical Review B 59, 1758 (1999).
 Shishkin and Kresse (2006) M. Shishkin and G. Kresse, Phys. Rev. B 74, 035101 (2006).
 Mostofi et al. (2008) A. A. Mostofi, J. R. Yates, Y.S. Lee, I. Souza, D. Vanderbilt, and N. Marzari, Computer Physics Communications 178, 685 (2008).
 Franchini et al. (2012) C. Franchini, R. Kováčik, M. Marsman, S. Sathyanarayana Murthy, J. He, C. Ederer, and G. Kresse, Journal of Physics: Condensed Matter 24, 235602 (2012).