Polarized Stimulated Emission of 2D Ensembles of Plasmonic Nanolasers

Abstract

Plasmonic nanolasers produce coherent light with wavelengths on a scale similar to their own or larger. In the past decade they have attracted intense interest, particularly from the emerging areas of integrated photonic circuits and biomedicine. Despite these capabilities, plasmonic nanolasers are still not completely understood, and this lack of understanding leads to confusing them with spasers and random lasers. Herein, the operation of pure spaser‐based plasmonic nanolaser arrays is presented. For this, a monolayer of silver nanoparticles (NPs) affixed to a dielectric surface and covered with a fluorescent polymethyl methacrylate (PMMA)–coumarin solid composite is investigated. The input–output characteristic measured for the composites on a bare substrate (without Ag NPs) reveals that the emission at pump pulse energies above 2.4 mJ (at 355 nm excitation wavelength) stops growing, and instead is inhibited by saturation. In contrast, in such structures with Ag NPs an additional emission band pops up over a fluorescence background. It has a spectral width order of units of nanometers and its intensity grows faster than at lower pump pulse energies, revealing a nonlinear dependence of the input–output characteristic. The spaser‐based lasing observed is completely linearly polarized and clearly directed as 45° from the substrate.

Theory of lasing in a two-dimensional array of plasmonic nanolasers

“…Again one observes both positive and negative threshold values for different h, however, now the character of lasing (conventional versus lasing without inversion) is inverted for the same values of h. This implies also that lasing occurs in either p or s polarization that agrees with the observations reported in Ref. [16]. In both cases lasing is possible in a limited angular range adjacent to the grazing propagation.…”

“…On the one hand, this assumption corresponds to the experimental conditions realized in Ref. [16]. On the other hand, it allows one to neglect the reflections from the dielectrics which are unimportant for the considered mechanism.…”

“…Taking as an example the coumarin 481 dye used in Ref. [16] and assuming its relative concentration in a solvent in the range 10 −4 − 10 −2 , one finds the number density of active molecules in the range N ≈ 3 × 10 17 − 3 × 10 19 cm −3 . Assuming that the population inversion is maximal (w = 1) and taking µ ≈ 5 D [24] and γ ⊥ ≈ 10 13 s −1 [12], one obtains η ≈ 0.002 − 0.2.…”

“…In the case where ǫ 1 ≈ ǫ 2 ≈ ǫ h and the dielectric slab does not contain MNPs the reflection coefficients r s j and r p j are close to zero (see the Appendix) and the same is true for the quantities φs and φp . This makes the laser generation impossible that was observed in experiments [16].…”

“…In the present paper, we report on a dramatic impact which the backaction has on the laser operation in a twodimensional array of metal nanoparticles (MNPs) in contact with molecular gain medium. We theoretically investigate lasing in a model system which resembles the * bordo@sdu.dk configuration realized in recent experiments [16] where ppolarized laser emission was observed. We demonstrate that the lasing condition is determined not just by the population inversion in gain medium as in conventional lasers, but also by the averaged phase factor which stems from the backaction of the MNPs layer.…”