2.1.

Chemicals and Solutions

Fluorescein-labeled fluorospheres (cat F8823, $2\text{-}\mu \mathrm{m}$ diameter and cat F8803, $0.1\text{-}\mu \mathrm{m}$ diameter) were from Thermo Fisher (Waltham, Massachusetts). They were supplied at $3.6\times {10}^{13}/\mathrm{mL}$ and used at $20\times$ dilution. All the other solvents were from Sigma (St. Louis, Missouri).

2.2.

Preparation of Coverslips

A 50-nm-thick layer of gold was deposited on the high refractive index (sapphire) coverglasses from Olympus by EMF Corp. (Ithaca, New York). About 5-nm layer of silica was deposited on top of gold to protect it from oxidation. About 2-nm chromium undercoat was used as an adhesive background.

2.3.

Sample Preparation

About $40\text{-}\mu \mathrm{L}$ suspension of fluorospheres were placed on metal side of gold coverslips and covered (to reduce evaporation) with #1 22-mm square Corning (Corning, New York) coverslips. For control experiments, fluorospheres were placed on a glass rectangular coverslip (Corning #1 $24\times 60\mathrm{mm}$).

2.4.

Microscopic Measurements

The schematic diagram of the RK microscope is shown in Fig. 1. The beam of 488-nm light from a Melles Griot (Model 85-BCF-020-112) SPSS laser is carried by a single-mode optical fiber to a beam expander. The expanded beam (1 cm in diameter) is incident perpendicularly from above on a sample that is placed a gold coverslip that is resting on the moveable stage of an inverted Olympus IX71 microscope. The fluorescent light is emitted at SPCE angle and is collected by the objective (Olympus Apo $100\times$, 1.65 NA). The beam splitter inside the microscope directs the beam either to an electron multiplier CCD camera operating at $-65°\mathrm{C}$ (EMImage, Hamamatsu, Japan, pixel size $16\times 16\mu \mathrm{m}$) attached to the side port of the microscope, or is projected onto a tube lens, which focuses it at the conjugate image plane of the microscope. The tip of an optical fiber (whose core acts as a confocal aperture) is inserted at this plane. An Avalanche photodiode (Perkin-Elmer SPCM-AQR-15-FC) collects light emerging from the aperture. The TTL signal from the diodes is fed to a correlator (Flex02-08D, Correlator Inc., Bridgewater, New Jersey), which acts as a data logger. The data are transferred to a PC that computes the time course of fluorescence, autocorrelation function (ACF), and histograms.

Fig. 1

RK microscope and the principle of reverse Kretschmann measurements. Fluorescence from fluorophores near the gold surface produces surface plasmons in the metal. Fluorescent light couples with the plasmons to emerge at the bottom of the coverslip at a (SPCE angle; which is smaller than SPR angle). Only the fluorescence from molecules excited within $\sim 10$ to $\sim 100\mathrm{nm}$ of the metal layer can penetrate the layer via plasmon resonance. Excitation light also gives rise to fluorescence from molecules outside this layer, but the emission is unable to penetrate the metal layer and enter the objective. This provides excellent background rejection. The fiber diameter was $3.5\mu \mathrm{m}$, laser intensity was $50\mu \mathrm{W}$, the intensity at the exit from single-mode fiber was $24\mu \mathrm{W}$, the intensity entering objective was $3\mu \mathrm{W}$, and the intensity at the exit from $60\times$ objective was $0.9\mu \mathrm{W}$.

3.1.

Reverse Kretschmann Signal

Suspension of $2\text{-}\mu \mathrm{m}$ diameter microspheres was diluted 10 times to $3.6\times {10}^{12}\text{spheres}/\mathrm{mL}$. The spheres were placed on a coverslip coated with gold. Figure 2(a) shows an image of the spheres on the gold-coated slide. Figure 2(b) shows a typical trace of intensity fluctuations. The intensities were measured in $160\text{-}\mu \mathrm{s}$ intervals for 30 s. The fluctuations are caused by spheres entering and leaving the detection volume. As the thickness of the detection layer is 35 nm (see below), much smaller than the lateral dimension of diffraction limited laser beam ($\sim 0.5\mu \mathrm{m}$), the fluctuations in the intensity are contributed mostly by spheres translating in the $Z$-dimension. Figure 1(c) is the ACF of fluctuations.

Fig. 2

(a) $2\text{-}\mu \mathrm{m}$ microspheres in RK. 1.65 NA $\times 100$ Olympus objective, sapphire substrate, refractive index of $\text{immersion oil}=1.78$. (b) The intensity fluctuations caused by diffusion of $2\text{-}\mu \mathrm{m}$ spheres through detection volume in RK experiment on gold substrate. (c) A typical ACF for $2\text{-}\mu \mathrm{m}$ nanospheres. Fluorospheres were diluted 100 times to $3.6\times {10}^{11}\text{spheres}/\mathrm{mL}$.

It should be noted that TIRF correlation function has a pronounced “tail,”⁸ which is most likely caused by particles sticking to the glass. In contrast, the RK ACF has no “tail.” The particles also probably stick, but they are invisible because of metal quenching.

3.2.

Background Rejection

Figure 3(a) shows that only fluorescence signal that enters metal layer through near-field interactions is able to penetrate the metal film. All near-field radiation enters the gold by coupling with surface plasmons and penetrates the metal surface to emerge at an SPCE angle. All far-field radiation is reflected by the metal film. Figure 3(b) compares the signal from 0.26-mM Rhodamine 6G on glass coverslip with the same dye on gold-coated coverslip. The background fluorescence is reduced by a factor of $>7000$. Figure 3(c) shows the dependence of signal on the concentration of the dye on the gold surface. The large background rejection applies only to the case when fluorescence originates 35 nm above the coverslip surface.

Fig. 3

(a) The illustration shows how the contribution of the background is reduced. (b) Fluorescent signal from 0.26-mM Rhodamine 6G (in the absence of fluorospheres) on glass coverslip (top panel) is compared with fluorescence of the same concentration of dye on a gold-coated coverslip (bottom panel). Dye was illuminated by 543-nm laser and fluorescence was viewed through Olympus TRITC Filter Cube Set (#67-007). The detector settings were identical in (a) and (b). (c) Concentration dependence of rhodamine 110. The decrease in fluorescence at the highest concentration is due to inner filter and concentration quenching. The red dot indicates fluorescence of the coverslip.

It should be noted that the method could also be applied in bulk (i.e., not in a microscope), where the suppression of background noise can be enhanced by taking advantage of directional and highly polarized character of SPCE signal.

3.3.

Biological Samples

A typical image of cell monolayer of astrocytes is shown in Fig. 4. Astrocytes provide structural and metabolic support for brain neurons. Astrocytes transiently expressing GFP were cultured for 24 to 48 h before imaging. GFP binds nonspecifically to the body of a cell and cell membrane and to astrocyte processes.

Fig. 4

Images of human fetal astrocytes transfected with GFP. (a), (b) Live astrocytes on glass, viewed through $\times 20$ objective. Micrographs were taken of each treatment at $400\times$ original magnification using an Eclipse Ti-300 (Nikon) equipped with a black and white Luca R (Andor). Images were contrast enhanced; (c), (d) RK image of astrocytes on sapphire-gold coverslips viewed through Olympus FITC Filter Cube set (#67-004) contrast NOT enhanced. Average intensity of the background at the upper left corner: 35. (e), (f) TIRF images of astrocytes on glass coverslip, the same objective, and the same settings of the video camera as in (c), (d). Viewed through Olympus FITC Filter Cube set (#67-004). Average intensity of the background at upper left corner: 45 contrast NOT enhanced.

The background rejection applies also to multilayers of cells. However, in the case when a sample consists of monolayers of cells, rejection is smaller because the background fluorescence is mostly contributed by the autofluorescence of glass and fluorescent glow of neighboring cells.

3.4.

Reduced Thickness of Detection Layer

The emission into the objective in RK originates from a volume that is thinner than in TIRF. Figure 5 (based on calculations in Ref. 9) shows that emission into the objective peaks around 35 nm. Also, in RK the fluorescence is quenched near the gold surface. The detected emission comes from a thin detection volume slightly above the metal surface. The rejection of the emission of fluorophores near the metal surface should be better for small fluorophores on the proximal surface of a cell than for spherical fluorescent beads employed here. In TIRF, there is no such rejection of emission near the surface.

Fig. 5

Thickness of detection layer in RK. The model of Ref. 8 was used to estimate the thickness. The theoretical approach to fitting the curves is the same as used in Ref. 9. It is an extension of Hassler et al theory.¹¹ It uses linear combinations of exponentials axially and a Gaussian laterally. The maximum intensity at emission into the sapphire prism is at $\sim 35\mathrm{nm}$. It is assumed that the exciting field is continuous in time and weak enough such that the time between excitations is much longer than the time between excitation and emission of the fluorophore. This means that the average number of photons emitted per unit time and therefore the average total emitted power does not depend on the lifetime; they only depends on the average time between excitations. The refractive indices of the metals used in the calculations are interpolated from Ref. 12. Inset: the same curve fitted with a single exponential. $1/e$ is 200 nm. Normal incidence. Metal is 50-nm layer of gold.

The thickness could not be determined experimentally by varying coating thickness and observing molecules immobilized or evaporated on this coating. Application of thin coating on top of gold results in an emission at several angles in the glass substrate, with either $s$ or $p$ polarization. This complex pattern of directional and polarized emission appears to be due to optical waveguide effects occurring when the sample thickness becomes comparable with the emission wavelength.^13,14 For example when layer is 290-nm thick, SPCE is observed at two angles, with different $s$ or $p$ polarization for each angle. Even more complex pattern is observed thicker films. The multiple rings of SPCE and the unusual $s$-polarized emission are consistent with the expected waveguide modes in the silver–PVA composite film.