Magnetron-sputtered Polytetrafluoroethylene-stabilized Silver Nanoisland Surface for Surface-Enhanced Fluorescence.

Surface-enhanced fluorescence (SEF) requires the absorption/emission band of the fluorophore, the localized surface plasmon resonance (LSPR) of the nanostructure and the excitation wavelength to fall in the same (or very close) spectral range. In this paper, we monitor the SEF intensity and lifetime dependence of riboflavin (vitamin B2) adsorbed on a spacer-modified Ag substrate with respect to the thickness of the spacer. The substrates were formed by silver nanoislands deposited onto magnetron-sputtered polytetrafluoroethylene (ms-PTFE). The spacer was formed by the ms-PTFE layer with the thickness ranging from ~5 to 25 nm. The riboflavin dissolved in dimethylsulfoxide (DMSO) at a 10 µM concentration forms, at the ms-PTFE surface, a homogeneous layer of adsorbed molecules corresponding to a monomolecular layer. The microspectroscopic measurements of the adsorbed layer were performed through a sessile droplet; our study has shown the advantages and limitations of this approach. Time-resolved fluorescence enabled us to determine the enhanced fluorescence quantum yield due to the shortening of the radiative decay in the vicinity of the plasmonic surface. For the 5 nm ms-PTFE layer possessing the largest (estimated 4×) fluorescence enhancement, the quantum yield was increased 2.3×.


Surface density determined from the concentration decrease inside the droplet
The concentration decrease (c) inside the droplet caused by the adsorption leads to an average surface density of the adsorbed molecules (NA is the Avogadro's number):

Dependence of the fluorescence signal on the droplet volume
The intensity of the fluorescence signal was measured for droplets of different volumes on the substrate with a 5 nm overlay. To minimize the effect of the sessile droplet evaporation, we used the confocal microspectrometer Witec (the disadvantage, however, was that we could not take the obtained ratio of the volume to the surface signal into the analysis of the time-resolved measurement). Figure S3 shows the results for the dimethylsulfoxide (DMSO) solution. The data were fitted with a theoretical curve based on an assumption of a Langmuir adsorption isotherm, valid for the formation of the first monolayer [36]. The surface density depends then on the residual concentration of the adsorbate inside the droplet (c') as where 0 is the surface density of the first molecular monolayer. For a given coefficient  the 0 value was calculated so that  was 2.0 × 10 12 molecules/mm 2 for the 5 µ L droplet (c' = 5 µ M). The surface density and the corresponding concentration in the droplet volume are related as where c0 is the initial concentration of the solution. Relations S4 and S5 provide quadratic equations for  and c' as functions of the Langmuir coefficient  and the droplet volume (through the ratio  of the droplet volume and the surface contact area): The experimental data were least-square fitted by the theoretical dependence The free parameters of the fit were the value of the coefficient  and the two coefficients of proportionality between the fluorescence signal and the riboflavin surface density or the riboflavin concentration inside the droplet. The best fit was reached for the high values of . It can be seen from Figure S6 that in this case the surface density of a monomolecular is virtually reached for droplet volumes higher than 1 µ L. An analogous experiment for the riboflavin aqueous solution gave somewhat less quality data. Furthermore, the results of the fit did not satisfactorily match the experimental data, showing only that the signal from the volume should be substantially stronger than that from the surface.

Focus position inside the droplet
The focus position inside the droplet was calculated by a numerical integration of the beams across the objective area, assuming the Gaussian intensity profile (half of the maximum at the edge), followed by a determination of the intensity maximum along the vertical axis. The calculations were performed for the 5 µ L DMSO (71° contact angle) and 10 µ L water (104° contact angle) droplets (see Figure S4).

Theoretical dependence of fluorescence intensity on the objective displacement
The measured fluorescence is a superposition of the fluorescence signal from the surface (depicted by a rectangle) and the fluorescence from the solution above the surface. Due to the axial symmetry, the position of any point, from which the fluorescence signal is detected, can be sufficiently characterized by two coordinates, r and z (see Figure S5). Let ( , ) is a contribution of a single molecule in the r, z position to the total fluorescence signal. This quantity is given by the local spatial density of the exciting radiation, the fluorescence yield and the efficiency of the collecting optical system to get fluorescent radiation from this location to the detector. The total fluorescence signal from the droplet bulk is then where v is the droplet height and c , is the actual concentration of the fluorophore inside the droplet. Supposing the fluorescence from the molecules at the surface is enhanced by a factor , their contribution to the measured fluorescence signal is where  is the surface density.
The dependence of ( , ) on r (for a particular z) is unknown, but we can employ the commonly used Lorentzian function to describe the dependence of the signal on z, which means ∫ ( , ) 2 d = 0 1 + ( ) 2 (S10) for focus at the surface.
Assuming that the parameter  does not change when the focus is moved up to d, the intensity of the fluorescence signal is