High-Resolution Imaging of Single Fluorescent Molecules with the Optical Near-Field of a Metal Tip

We show that a concentration of light at a metal tip allows near-field optical imaging of single fluorescent dye molecules at very high resolution, despite strong quenching effects. Details as small as 10 nm were observed in the fluorescence patterns of single Cy-3 dyes bound to the termini of DNA. Data evaluation by model fitting determines the positions of the dyes to an accuracy even better than 1 nm and also yields their 3D orientation. The metal tip simultaneously provides high-resolution topographic imaging complementing the optical signal for a detailed surface examination.

Figure 1

Schematic drawing of the TOA probe with the DNA sample. The TOA probe consists of a sharp metal tip mounted on a conventional aperture probe. The aperture probe is made up of a tapered glass fiber coated laterally with metal, which confines the light to the glass except at the aperture in the front end. The sample consists of DNA with covalently bound Cy-3 molecules, prepared on mica.

Figure 2

Fluorescence image of single Cy-3 dye molecules, which appear mostly as double maxima. Scale bar: 200 nm. The inset on the left shows a zoomed image of a dye molecule together with a section along the line (three lines average). The maximum on the right has a value of only 10 nm for the FWHM. A bleaching event from one scan line (oriented vertically) to the next one is displayed enlarged in the inset on the right side. Scale bars for the insets: 25 nm.

Figure 3

Field below the TOA tip approximated by a dipole field. The strength and direction of the field is indicated by the length and orientation of the short lines. The excitation of a dye molecule is proportional to the squared field component parallel to the dye's absorption dipole moment. It depends strongly on orientation and position of the dye relative to the tip.

Figure 4

Fluorescence patterns of differently tilted dye molecules. Upper row: Measurements (scan lines are oriented vertically, also in Figs. 5, 6). Lower row: Patterns calculated with parameters fitted to the measurements. Resulting tilt angles out of the sample plane are 0°, 14°, 20°, 49°, and 68°. Fitted tip radius: 22 nm. Image sizes: 117 nm. The patterns are broader than in Figs. 2, 5, 6 due to a blunter tip.

Figure 5

DNA with Cy-3 labeled termini on mica and corresponding data modeling. (a) Topography together with calculated positions of the analyzed dye molecules marked by “$+$.” (b) Fluorescence. (c) Positions as in (a), together with tilt angles (upper numbers) and azimuth angles (lower numbers) as resulting from our first approximation fits shown in (d). The fitted radius of the tip is 12 nm. Good fits are displayed in green; problematic fits due to, e.g., blinking or bleaching of the dyes are marked in yellow. The data in (a) are flattened (from each line, a fitted parabola has been subtracted). Scale bar: 200 nm.

Figure 6

Fluorescence pattern of two dyes located close to each other. (a) Experimental data. (b) Best fit when assuming two dyes for the encircled pattern, resulting distance 12 nm. (c) Difference between data (a) and fit (b). Outside the circle enclosing the pattern in question, a single dye molecule is visible. Image size: $300\times 300{\mathrm{n}\mathrm{m}}^2$.