Spectral variation of fluorescence lifetime near single metal nanoparticles

Abstract

We explore the spectral dependence of fluorescence enhancement and the associated lifetime modification of fluorescent molecules coupled to single metal nanoparticles. Fluorescence lifetime imaging microscopy and single-particle dark-field spectroscopy are combined to correlate the dependence of fluorescence lifetime reduction on the spectral overlap between the fluorescence emission and the localised surface plasmon (LSP) spectra of individual gold nanoparticles. A maximum lifetime reduction is observed when the fluorescence and LSP resonances coincide, with good agreement provided by numerical simulations. The explicit comparison between experiment and simulation, that we obtain, offers an insight into the spectral engineering of LSP mediated fluorescence and may lead to optimized application in sensing and biomedicine.

Figure 1. Sample layout and experimentally measured dark-field and fluorescence maps.

(a) Sketch of the sample cross-section. (b) Dark-field light scattering image of GNPs embedded in a Nile Red-doped PMMA film and (c) the respective fluorescence intensity distribution; scale bars are 2 μm. (d) Fluorescence intensity and (e) LSPR modified lifetime of Nile Red in the vicinity of a single GNP; scale bars are 100 nm.

Figure 2. Experimental spectral and time dependencies.

(a) Dark field scattering spectra from three single GNPs (solid lines) and fluorescence spectrum of the Nile Red-doped PMMA film (dashed line). (b) Fluorescence decays of Nile Red at the locations of the GNPs in (a); the decay for a reference sample is also plotted for comparison (black line). The strong red-shift identified in LSPR spectrum 1, and its associated longer fluorescence decay, represent a rare outlier measurement, which might result from a small nanoparticle aggregate or a rare non-spherical nanoparticle.

Figure 3. Lifetime modification in the presence of resonant GNPs.

(a) The fluorescence lifetime reduction τ_m/τ₀ as a function of LSPR peak position; the Nile Red fluorescence spectrum from a thin film is also shown (red line). LSPR peak positions are binned together by 6 nm intervals, with the average value of the modified lifetime plotted for each bin (error bars show the standard deviation for each bin). (b) The calculated modified lifetime shows a strong dependence on the spectral overlap between LSPR and Nile Red emission (dashed line denotes the emission wavelength of the radiating dipole used in the simulations).

Figure 4. Samples and the experimental setup.

(a) SEM image of single GNP on a glass coverslip. (b) Absorption and emission spectra of a Nile Red-doped PMMA film. (c) Schematic of FLIM and dark-field experimental setup. (The schematic was drawn by J.L.).

Figure 5. Fluorescence lifetime analysis.

Top: Typical fluorescence decay fitted with a bi-exponential curve (A0 = 916, τ₀ = 3.8 ns, A_m = 1186 and τ_m = 0.98 ns in the notations of Eq. 2). Bottom: residuals from the fitting.