Cooperative Energy Transfer Controls the Spontaneous Emission Rate Beyond Field Enhancement Limits

Quantum emitters located in proximity to a metal nanostructure individually transfer their energy via near-field excitation of surface plasmons. The energy transfer process increases the spontaneous emission (SE) rate due to plasmon-enhanced local field. Here, we demonstrate a significant acceleration of the quantum emitter SE rate in a plasmonic nanocavity due to cooperative energy transfer (CET) from plasmon-correlated emitters. Using an integrated plasmonic nanocavity, we realize up to sixfold enhancement in the emission rate of emitters coupled to the same nanocavity on top of the plasmonic enhancement of the local density of states. The radiated power spectrum retains the plasmon resonance central frequency and line shape, with the peak amplitude proportional to the number of excited emitters indicating that the observed cooperative SE is distinct from superradiance. Plasmon-assisted CET offers unprecedented control over the SE rate and allows us to dynamically control the spontaneous emission rate at room temperature which can enable SE rate based optical modulators.

Figure 1

(a) An excited QE coupled to plasmonic resonator non-radiatively transfers its energy, at a rate ${\mathrm{\Gamma }}^{\mathrm{ET}}$ to the plasmon mode, which radiates it away. (b) An ensemble of QEs coupled to a resonant plasmon mode transfer their energy to it cooperatively at a rate ${\mathrm{\Gamma }}_c^{\mathrm{ET}}$ that is the sum of individual rates [9].

Figure 2

(a) SEM image of plasmonic nanocavity (PNC) array (scale $\mathrm{bar}=5\mu \mathrm{m}$). (b) SEM image of a cross section of a single PNC that was cut using focused ion beam FIB (scale $\mathrm{bar}=100\mathrm{nm}$). (c) Schematic of the nanopillar PNC. The quantum dots (QDs) are spin coated on a polymeric scaffold, then an Au layer is deposited. (d) Schematic of a cross section of a single nanopillar. Incident light excites QDs that, subsequently, transfer their energy to excite localized surface plasmons (LSPs) which decay into a photon. (e) Measured scattering for PNC array; the resonance maximum was determined by fitting the data with a Lorentzian function. The measured resonance closely agrees with the calculated absorption and scattering presented in (f). (g) The photoluminescence of QDs spin coated on an Au film vs QDs incorporated in a single PNC.

Figure 3

(a) Measured time-resolved photoluminescence for five different excitation intensities for the PNC (top) and the reference Au film (bottom). The SE lifetime is intensity dependent only for the PNC. (b) The fitted SE rate fast component (black spheres) and slow component (red spheres) for the PNC (Top) and for the reference Au film (Bottom). (c) Reversible, dynamic control over SE rate. The fast and slow SE rate components vary by modifying $N$. The SE rate is linearly proportional to the excitation intensity.

Figure 4

(a) The ratio of the measured fast ${\mathrm{\Gamma }}^{\mathrm{fast}}$ and slow ${\mathrm{\Gamma }}^{\text{slow}}$SE rates for QDs on the reference Au film and inside the PNC. The ratio ${\mathrm{\Gamma }}^{\mathrm{fast}}(I)/{\mathrm{\Gamma }}^{\text{slow}}(I)$ is $\sim 3$. Inset: schematic of the decay process of charged biexcitons and charged excitons. (b) The rate ${\mathrm{\Gamma }}^{\mathrm{fast}}$ is calculated from experimental rate ${\mathrm{\Gamma }}^{\text{slow}}$ by assuming that the slope of the SE rate vs intensity curve is proportional to the energy transfer rate of individual QD, as predicted by Eq. (2). (c) The photoluminescence as a function of excitation intensity shows that the emission spectrum retains the plasmon line shape as the peak emission wavelength is $\sim 631\mathrm{nm}$.