Large Excitonic Reflectivity of Monolayer ${\mathrm{MoSe}}_2$ Encapsulated in Hexagonal Boron Nitride

We demonstrate that a single layer of ${\mathrm{MoSe}}_2$ encapsulated by hexagonal boron nitride can act as an electrically switchable mirror at cryogenic temperatures, reflecting up to 85% of incident light at the excitonic resonance. This high reflectance is a direct consequence of the excellent coherence properties of excitons in this atomically thin semiconductor. We show that the ${\mathrm{MoSe}}_2$ monolayer exhibits power-and wavelength-dependent nonlinearities that stem from exciton-based lattice heating in the case of continuous-wave excitation and exciton-exciton interactions when fast, pulsed laser excitation is used.

Figure 1

Electrically switchable, atomically thin mirror. (a) Schematic of the experimental setup. A ${\mathrm{MoSe}}_2$ monolayer encapsulated between two $h\text{−}\mathrm{BN}$ layers is placed on ${\mathrm{SiO}}_2$ $(285\mathrm{nm})/\mathrm{Si}$ substrates. Platinum (Pt) electrodes are used to contact the ${\mathrm{MoSe}}_2$ monolayer, while Si is used as back gate. The incident light resonant with the exciton transitions in ${\mathrm{MoSe}}_2$ can be strongly reflected. The free carrier density in ${\mathrm{MoSe}}_2$ can be tuned with a gate voltage applied to the Si back gate. (b) Optical image of a monolayer ${\mathrm{MoSe}}_2$ device $M1$. The dashed white and blue lines show the outline of ${\mathrm{MoSe}}_2$ and a Pt contact, respectively. (c),(d) Reflection images of $M1$ under 750-nm resonant laser illumination with a gate voltage of (c) $=-20\mathrm{V}$ and (d) 30 V at $T=4\mathrm{K}$. The fully encapsulated ${\mathrm{MoSe}}_2$ region shows substantial reflection, while the part that is not fully encapsulated is not as reflective. (e) Reflection spectra of another monolayer ${\mathrm{MoSe}}_2$ device $M2$ as a function of gate bias at 4 K. Reflection due to charged excitons can be observed in the electron-doped regime. To obtain absolute reflectance, we normalize the reflected intensity using the measured reflectance of the bare substrate and metal electrodes.

Figure 2

Temperature-dependent reflectance spectra of a ${\mathrm{MoSe}}_2$ monolayer. (a) Temperature-dependent reflection spectra of the monolayer ${\mathrm{MoSe}}_2$ device $M2$. The line colors correspond to different temperatures listed in (b). (Inset) Schematic of interference between the reflection from the ${\mathrm{MoSe}}_2$ monolayer and the reflection of the $\mathrm{BN}/\mathrm{SiO}2/\mathrm{Si}$ substrate. (b) Reflectance spectra of the ${\mathrm{MoSe}}_2$ monolayer as a function of temperature obtained from the fit. (c) Temperature dependence of peak reflectance and total linewidth of the ${\mathrm{MoSe}}_2$ monolayer in $M2$. (d) Radiative, nonradiative, and pure dephasing linewidths of the same monolayer as a function of temperature.

Figure 3

Nonlinearity of the reflection from a ${\mathrm{MoSe}}_2$ monolayer at $T=4\mathrm{K}$ under cw resonant excitation. (a) Reflectance of the monolayer ${\mathrm{MoSe}}_2$ device $M2$ as a function of the cw laser power ($I_p$) at different detuning wavelengths, as indicated in the inset. (b) Nonlinear reflectance predicted by the thermal model [34] at different detuning wavelengths. (Inset) Schematic of the optical bistability with red detuning. Here, $n_s$ and $I_s$ denote the photon density and the incident laser peak power when the energy shift of the exciton transition is equal to the radiative linewidth.

Figure 4

Nonlinearity of the optical response at $T=4\mathrm{K}$ under pulsed-laser excitation. (a) Reflection as a function of peak power ($I_p$) of the picosecond laser excitation (6 ps full width at half maximum), at the wavelengths indicated in the inset. (b),(c) Two-dimensional plots of (b) experimental and (c) theoretical reflectance at various excitation wavelengths as a function of laser peak power.