Tailoring spontaneous infrared emission of HgTe quantum dots with laser-printed plasmonic arrays

Chemically synthesized near-infrared to mid-infrared (IR) colloidal quantum dots (QDs) offer a promising platform for the realization of devices including emitters, detectors, security, and sensor systems. However, at longer wavelengths, the quantum yield of such QDs decreases as the radiative emission rate drops following Fermi’s golden rule, while non-radiative recombination channels compete with light emission. Control over the radiative and non-radiative channels of the IR-emitting QDs is crucially important to improve the performance of IR-range devices. Here, we demonstrate strong enhancement of the spontaneous emission rate of near- to mid-IR HgTe QDs coupled to periodically arranged plasmonic nanoantennas, in the form of nanobumps, produced on the surface of glass-supported Au films via ablation-free direct femtosecond laser printing. The enhancement is achieved by simultaneous radiative coupling of the emission that spectrally matches the first-order lattice resonance of the arrays, as well as more efficient photoluminescence excitation provided by coupling of the pump radiation to the local surface plasmon resonances of the isolated nanoantennas. Moreover, coupling of the HgTe QDs to the lattice plasmons reduces the influence of non-radiative decay losses mediated by the formation of polarons formed between QD surface-trapped carriers and the IR absorption bands of dodecanethiol used as a ligand on the QDs, allowing us to improve the shape of the emission spectrum through a reduction in the spectral dip related to this ligand coupling. Considering the ease of the chemical synthesis and processing of the HgTe QDs combined with the scalability of the direct laser fabrication of nanoantennas with tailored plasmonic responses, our results provide an important step towards the design of IR-range devices for various applications.

In this section we demonstrate tunability of the first-order lattice plasmon resonance (FLPR) of the nanobump array via tailoring the nanobump geometry as well as via array pitch dimensions p. Two representative series of FTIR reflection spectra demonstrate red-shift of the FLPR spectral position caused by either an increase of the geometric size of the nanobumps tailored by increasing the applied pulse energy in the process of their laser printing (left-side column) as well as by increasing the array pitch p for the fixed size of the nanobumps in the array (fixed pulse energy applied; rightside column). In the first case, the red shift of the FLPR is caused by increased "effective" array period (or plasmon running distance) accumulated on the curved surface of the nanobumps. Figure S1: Series of FTIR reflection spectra showing the ability to tailor a spectral position of the FLPR via either a nanobump geometry tuned by applied pulse energy (left-side column) or an array pitch at fixed nanobump size (right column). FTIR spectra are vertically shifted for better representation. Shaded areas on both series indicate the FLPR tuning range. Two inset SEM images show the FIB cross-section central cut through the nanobumps produced at the lowest (purple) and the highest (green) pulse energy shown on the left-side column. Three spectra on the right-side column shown as red dotted curves correspond to arrays used in the manuscript as samples A, B and C.
To further illustrate this point, we provide two SEM images showing the FIB cross-section central cut through the nanobumps produced at the lowest (purple curve) and the highest (green curve) pulse energy. Careful analysis of the geometrical size of both structures indicate that the smallest nanobumps increase the actual array pitch by ≈ 300 nm, while the largest ones -by 700 nm, which is in satisfactory agreement with the experimentally observed spectral shift of the FLPR. From the other hand, for the fixed size of the nanobumps, an increase of the array pitch also results in a scalable red shift of the FLPR position. The provided results also indicate that the efficiency of coupling to the surface plasmon waves depends on both the nanobump geometry and their spacing. Indeed, smaller or too separated nanobumps provide weaker coupling efficiency. Notably, maximal coupling efficiency, which can be assessed by the FLPR amplitude on the FTIR spectrum, was about 40% for large-scale nanobump arrays produced using direct laser printing.

Spectral position of the plasmonic lattice resonances
As it was shown in refs. [1,2], nanobump arrays support first-order lattice plasmon resonances at a wavelength determined by the plasmon running length (or effective array period). The latter comprises the slopes of the nanobump sidewalls, plus the remaining flat surface between the structures, and so it is greater than the array period (see inset in Fig. 1b). At normal incidence, when the plasmon wavelength (where λ0 is the wavelength in free space, εm and εs are the complex permittivities of the metal layer and of the superstrate), is equal to the effective period, the plasmons interfere in phase and a resonance is observed.  The second harmonic of this resonance corresponds to the wavelength when the effective period of the nanobump array equals to double the plasmon wavelength. This resonance can be tailored by geometry and arrangement of the nanobumps in a similar way to fit the pump laser wavelength. This could allow for more efficient excitation, albeit this resonance has much weaker amplitude due to a stronger plasmon dissipation in the visible spectral range.
We performed 3D FDTD calculations to reveal the spectral range where the isolated nanobumps can contribute to enhanced electromagnetic fields via excitation of LSPRs. For these calculations, we considered an isolated nanobump having the geometry shown in Fig. 1b, which was excited from the top by a broadband linearly polarized total-field scattered-field source. Computational volume was limited by perfectly matched layers. The back-scattered spectrum normalized over the excitation spectrum and averaged over the upper hemisphere collection angles is shown in Fig. S3 revealing multiple LSPRs at wavelengths below 1100 nm. This spectrum is also shown in an inverted view in Fig.  1c of the main manuscript indicating good correlation with the experimental results and explaining the broadband dip in reflection in the visible and near-IR spectral ranges. Figure S3: Back-scattered spectrum of the isolated Au nanobump normalized over the excitation spectrum and averaged over the upper hemisphere collection angles.

Calibration of the spectral shift of the first-order lattice resonance versus the thickness of the capping dielectric layer
To assess the spectral shift of the main lattice band of the Au nanobump array (centered at λ0) which is caused by the deposition of the HgTe QD layer, we have performed a series of calibration experiments using an amorphous silicon (α-Si) film, deposited above several similar plasmonic nanobump arrays using a commercial e-beam evaporation system equipped with a calibrated quartz microbalance. This material has averaged bulk refractive indices of nα−Si ≈ 3.47 in the near-IR spectral range [3], which is expected to be close to the bulk refractive index of the HgTe QDs, according to the available experimental data. The experimentally measured relative spectral shift of the main lattice band ∆λ/λ0 as a function of the thickness of α-Si capping layer, d, is presented in Fig. S4.   S5. The schematic drawing of FLS920P spectrometer system used in this study. "Laser" -2 W, 880 nm solid state laser for steady-state measurements or 670 nm, 62.8 ps pulse width pulsed laser for timeresolved measurements; "LP" -custom manufactured laser port for PL excitation by external laser; "L"focusing laser lens, f = 150 mm; "S" -sample on XYZ-translation and rotation stage for its precise adjusting with collection optics; "SM" -spherical mirror for emission collection; "M" -plane mirror; "EM" -emission monochromator; NIR PMT -nitrogen-cooled photomultiplier tube used for timeresolved measurements (spectral range 550-1630 nm); InSb -nitrogen-cooled solid state detector used for PL steady state measurements (spectral range 1200-4500 nm).