Hybridized Exciton-Photon-Phonon States in a Transition Metal Dichalcogenide van der Waals Heterostructure Microcavity

Excitons in atomically thin transition-metal dichalcogenides (TMDs) have been established as an attractive platform to explore polaritonic physics, owing to their enormous binding energies and giant oscillator strength. Basic spectral features of exciton polaritons in TMD microcavities, thus far, were conventionally explained via two-coupled-oscillator models. This ignores, however, the impact of phonons on the polariton energy structure. Here we establish and quantify the threefold coupling between excitons, cavity photons, and phonons. For this purpose, we employ energy-momentum-resolved photoluminescence and spatially resolved coherent two-dimensional spectroscopy to investigate the spectral properties of a high-quality-factor microcavity with an embedded ${\mathrm{WSe}}_2$ van der Waals heterostructure at room temperature. Our approach reveals a rich multibranch structure which thus far has not been captured in previous experiments. Simulation of the data reveals hybridized exciton-photon-phonon states, providing new physical insight into the exciton polariton system based on layered TMDs.

Figure 1

(a) Schematic plot of the vibronic polariton model. (b) Simulated polariton dispersion curves under two-coupled-oscillator model. (c),(d) Simulated polariton dispersion curves using vibronic polariton model under the assumption that the coupling terms with phonon-dressed photon states are deactivated (c) and activated (d). The dashed purple lines indicate the dispersion relation of the pure exciton.

Figure 2

(a) Microscope image of the ${\mathrm{WSe}}_2/\mathrm{h}\text{−}\mathrm{BN}/{\mathrm{WSe}}_2$ structure. Insets: Map of PL intensity (left); schematic diagram of the microcavity with embedded ${\mathrm{WSe}}_2/\mathrm{h}\text{−}\mathrm{BN}/{\mathrm{WSe}}_2$ heterostructure (right). (b) Schematic diagram of 2D microspectroscopy. (c) Detection scheme of momentum-resolved PL spectroscopy. Abbreviations: Fourier-space plane (FSP), real-space plane (RSP), and avalanche photodiode (APD).

Figure 3

(a) Energy-momentum-resolved PL map (for negative $k_{\Vert }$, see Supplemental Material [20] for the full dataset). (b) Simulated polariton dispersion relations using the vibronic polariton model. The horizontal dotted line indicates the energy of pure excitons. (c) Retrieved energy positions (squares) of five bands as functions of $T$, and linear fitting results (solid lines). The error bars are evaluated as the standard deviation of the fluctuations calculated from the differences between the measured values (squares) and linear fit results (solid lines) at each $T$. Colored shaded areas in (b) and (c) indicate the total confidence ranges of experimentally determined energy positions from 2D spectroscopy data evaluated over all $T$, to facilitate comparison with the theoretical dispersion curves. (d) Linear slopes of the five fits from (c), with error bars depicting 95% confidence bounds from the fit.

Figure 4

(a) Feynman pathways corresponding to rephasing and nonrephasing 2D spectra. (b),(c) Bottom: Measured rephasing (b) and nonrephasing (c) 2D spectrum (absolute value) at $T=50\mathrm{fs}$. Top: Slices through the 2D spectra along the diagonal lines (squares) and the fitting results (black solid curves) using five Gaussian functions (colored dashed curves). Abbreviations: Ground-state bleach (GSB) and stimulated emission (SE).