Real-time tunable lasing from plasmonic nanocavity arrays

Plasmon lasers can support ultrasmall mode confinement and ultrafast dynamics with device feature sizes below the diffraction limit. However, most plasmon-based nanolasers rely on solid gain materials (inorganic semiconducting nanowire or organic dye in a solid matrix) that preclude the possibility of dynamic tuning. Here we report an approach to achieve real-time, tunable lattice plasmon lasing based on arrays of gold nanoparticles and liquid gain materials. Optically pumped arrays of gold nanoparticles surrounded by liquid dye molecules exhibit lasing emission that can be tuned as a function of the dielectric environment. Wavelength-dependent time-resolved experiments show distinct lifetime characteristics below and above the lasing threshold. By integrating gold nanoparticle arrays within microfluidic channels and flowing in liquid gain materials with different refractive indices, we achieve dynamic tuning of the plasmon lasing wavelength. Tunable lattice plasmon lasers offer prospects to enhance and detect weak physical and chemical processes on the nanoscale in real time.

A charge-coupled device (CCD) beam profiler was used to map the far-field pattern of the emitted lasing signal at different distances from 5 cm to 7 cm to determine spatial coherence.
The beam only slightly expanded over the distances measured, indicating high directionality of the emission. The beam extended along the high-symmetry grating direction due to off-angle ASE. dissolved in DMSO had a lifetime of ~ 0.9 ns on a glass substrate. For lifetimes shorter than the IRF, the system will return a curve similar to the IRF. To extract the lifetime, we fit the data and deconvolved with the measured IRF. In order to measure the decay time at a given wavelength, the signal was first dispersed by a spectrometer grating and filtered by an exit slit. The wavelength-dependent decay time data performed on the IR-140-DMSO-only as control showed only spontaneous emission 860 nm to 880 nm with a ~ 0.9 ns lifetime. , which was one order of magnitude longer than the lifetime of the lasing mode at the same pump intensity. When the pump intensity was above lasing threshold, at 0.143 mJ/cm 2 , and above ASE threshold, at 0.238 mJ/cm 2 , for example, the ASE lifetime was further reduced to 21 ps and 17 ps. We also measured the decay time of dye molecule emission from 860 nm to 890 nm at a pump intensity of 0.154 mJ/cm 2 to study the wavelength-dependence of the lifetime above threshold. A broad emission peak from ca. 870 nm to 890 nm exhibited a reduction in lifetime (~ 20-250 ps). Note: the lasing mode at 865 nm was also resolved in the scan because our objective (NA = 0.14) was located at 34 mm from the surface and had a large field view (~ 17°); even with the small lasing beam divergence (~1.5°), the objective was still able to collect lasing signal due to the spatial distribution of the lasing emission from the ~ 0.2 cm 2 excitation area.

Supplementary Note 1: Numerical Simulations of Lasing Emission
We used a semi quantum framework reported recently 3,4 to simulate the interaction between the electromagnetic fields and gain medium. The dye molecules (active medium) were described as a four-level system (Supplementary Figure 5)  In these equations, we coupled the molecular polarizations and to the electromagnetic field and neglected the polarizations from other transitions because they were assumed to be very fast (10 fs). The time evolution of the populations N i , and the population differences Δ , were calculated from the rate equations: The time evolution of each quantum state density is governed by spontaneous decay processes / , and stimulated processes ( ⋅ , ⁄ ). The field involved in Supplementary Equations 3-6 is the total field and accounts for the effects of any local optical intensity (e.g., plasmon enhanced near field) on the dynamics of the population densities.
The approach based on Supplementary Equations 1-6 is self-consistent in fields and populations. We also included the modification to the spontaneous decay rate arising from the Purcell factor using an approach described previously. 4

Supplementary Note 2: Calculation of Gain and Loss
The quantum yield of IR-140 dye molecules in a uniform environment at room temperature is -1 +  nr -1 ) = 16%, where  r0 and  nr are the intrinsic radiative and non-radiative decay lifetimes. 6 The radiative decay from a Purcell-enhanced spontaneous emission rate (F/ r0 ), where F is the Purcell enhancement factor can be estimated according to the following equations: 3 From the TCSPC lifetime measurement, we determined the slow time constant  uncoupled to be 715 ps and measured fast components of 21 ps (approaching threshold) and 12 ps (above threshold, limited by the resolution of TCSPC). Since stimulated emission induces fast decay, we used 21 ps as the fast time constant coupled to avoid overestimation of the Purcell factor. We estimate that the Purcell factor F is ca. 200, in agreement with our previous work on Au NP arrays and solid gain media determining F from transient absorption (TA) measurements below threshold. 3 Note that TA is a pump-probe technique where the excited state population is monitored, and so dynamics of the energy transfer between dye molecules and the surface plasmons is indirectly determined. This F value (200) will be similar for other substrates and IR-140 in other solvents in this work. The cavity loss per length is given by ωn/cQ, where n is the refractive index and Q is the quality factor. 7 For Au NPs arrays in different index environments, the loss was estimated to be 4580-4920 cm -1 using the quality factors (Q = 210-230) of the lattice plasmons measured in experiments. Using the emission cross section reported in the reference, 8 the gain was estimated to be Nσ e = 360 cm -1 (for 1 mM), 9 which cannot compensate the loss. With Au NPs arrays, however, the emission cross section σ e ' = Fσ e 10 can be enhanced by Purcell factor. With an F