Fig. 1. Cavity and sample structures. (a) Schematic of open optical microcavity. The top mesa consists of a 20  μm×40 μm glass window area, a gold-coated planar region, and lenses. (b) Microscope image of top mirror: gold-coated mesa of 100  μm×100  μm. The lens structures have uniform depth of 300 nm but different diameters of 3–6 μm. (c) Microscope image of MoSe2-WSe2 HBL on bottom DBR. (d) PL spectra of iX at 3.5 K measured through the planar window (black) and a 6 μm lens (red) that shows discretized transverse lens modes. The inset shows the sketch of type-II band alignment.

Fig. 2. Tunable radiative dynamics of iX via low-Q longitudinal modes. (a) Cavity detuning effects: the longitudinal modes enhance the PL intensity as tuned through the iX emission band. (b) Several TRPL traces of iX at cavity detuning between 40 V and 60 V. The inset shows the accelerated spontaneous emission in the enlarged temporal domain of 1–2 ns. (c) Tuning of the lifetime: the fitted faster decay process τ2 oscillates with a period of 40 V, corresponding to cavity length variation ∼400  nm. (d) Transfer matrix simulation of the EM field intensity in the glass-DBR cavity as a function of the cavity gap. The intensity can change periodically up to 23 μm gapped cavity (Q-factor ∼24) to perturb iX. (e) FDTD simulation of the angle-dependent mode intensity distribution in far-field. The fitting curve (red) under the shade area (magenta) represents the cutoff of a Gaussian-shaped function (FWHM=4.5°). The dotted line highlighting the 1/e2 of maximum is substantially higher than the intensities distributed at the high-angle flanks, justifying the selection on the cutoff. Left inset: schematic of FDTD simulation box. The dipole is placed in the middle of a 5 μm×5 μm HBL on a DBR mirror and the detector is set 1.3 μm above the surface to simulate far-field emission pattern. Refractive indices are taken for the wavelength of the observed mode. Right inset: corresponding mode distribution in quasi-particle momentum space. The k-vectors are converted using k∥=2π sin θ/λ. (f) Linear fit of the measured Γtot as a function of FP calculated from Eq. (1).

Fig. 3. Tunable radiative dynamics of iX via discretized high-Q Tamm-plasmon modes. (a) Stacked PL spectra of the cavity scan using a 6 μm gold-coated lens. The 10th and 11th sets of discretized modes are tuned through iX emission. The white dashed lines represent a spectral window of 870–890 nm confined by filters for following TRPL measurements. (b) TRPL measurements for emission energies at E1, E2, and E3 marked in (a). The TRPL shows clearly faster decay as the lens modes enter the spectral window. The inset shows the calculated Purcell factor by Lumerical methods.

Fig. 4. Second-harmonic generation measurements.

Fig. 5. (a) Three-level system, where the intralayer exciton state |X⟩, interlayer exciton state |iX⟩, and the crystal ground state |0⟩ are considered. (b)–(f) Bi-exponential fit of the TRPL traces in Fig. 2(b) with a fixed slow decay time of τ1=12  ns.

Fig. 6. Data on HBL 2. (a) Microscope image of HBL 2. (b) Static PL measured through the mesa glass window. The emission is centered at 940 nm. (c) Stacked PL spectra by cavity detuning of the glass window. The anti-node and node are seen around 20 V and 50 V DC voltage, respectively. (d)–(g) TRPL traces measured under 50 V, 40 V, 30 V, and 22 V DC voltages. The bi-exponential fit and residual are also shown. (h) Fitting results (τ2=2.8±0.22  ns) of the oscillating bright iX decay time, with τ1=20.8  ns as the fixed fitting parameter.

Fig. 7. Sketch of general system geometry in the transfer matrix approach. The system is made up of layers of refractive index ni and thickness δi. Outside the structure we assume homogeneous media with refractive indices n0 and nN+1. Red arrows indicate propagation directions of the individual light-mode components in the case of forward emission through the structure.

Fig. 8. Refraction and reflection of light incident on a planar interface between two media at angle θ1. εs and εp± indicate the polarization vectors for s- and p-polarized light.

Fig. 9. Modes at emitter position as functions of angle of incidence (θ) and the cavity length for both directions and polarizations. Horizontal lines indicate the line cut shown in Fig. 10.

Fig. 10. Modes at the fixed cavity length (d=22.5  μm) as functions of angle of incidence (θ) for all combination of polarizations (s,p) and directions (F,B). Mode functions for both polarizations coincide for normal incidence (θ=0°).