Interferometric Signatures of Single Molecules

In 2001, the first observation of interferometric signatures of single impurity molecules embedded in a solid matrix has been reported [22]. This investigation has been performed in cooperation with the SMS group @ ETH Zurich (laboratory of Prof. U.P.Wild).

It was believed before that the resonant emission of single molecules cannot be observed directly with acceptable signal-to-noise ratio (SNR) due to the background of much stronger excitation laser light. However, in this work an excellent SNR was achieved by using an interferometric technique, which gives us a new tool for fundamental studies.

A thin naphthalene crystal, doped with a very small concentration (~10^-9 mol/l) of terrylene molecules, was used as a sample. This impurity system, unlike many others host-guest combinations used in single-molecule spectroscopy (SMS), is very stable: virtually no spectral diffusion can be observed at low temperatures on the level of single-molecule spectra. Another words, once finding a sharp spectral line of a single terrylene molecule, the researcher can excite this molecule with a stable narrow-line single-frequency laser for hours, being almost sure that the molecule does not change its resonance frequency due to a matrix-induced spectral jump.

The sample was placed in an optical cryostat filled with superfluid liquid helium at a temperature of 1.6 K (see Fig. 1). A microscope objective L₁ was focusing the light of a single-frequency dye laser with λ ~ 574 nm into a Gaussian spot of about 1.6 μm in diameter, slightly behind the surface of the sample. Due to a very low concentration of the impurity, only about 10 terrylene molecules are located in this spot of laser excitation. But these molecules can be detected only if the laser frequency is tuned into resonance with their purely electronic line - in this case such a molecule is excited and starts to emit fluorescence, which can be detected by a photosensor.

Fig.1 Part of the experimental setup.

L₁ is the microscope objective, L₂ - a lens; d is the distance between the crystal surface and the single molecule SM under study. The emission of SM (shown in gray) is collected and collimated by L₁. BS labels a beamsplitter.

The normal procedure in SMS is to filter out the excitation laser light (together with the resonance fluorescence of the molecule) and to monitor the non-resonance red-shifted fluorescence light. The homogeneous width of the purely electronic line of a terrylene molecule is about 50 MHz, but its central frequency can vary due to a slightly different environment. At higher concentrations of the impurity molecules their lines form a broader line - a so-called inhomogeneous band. In a naphthalene crystal the inhomogeneous width of the purely electronic line of terrylene is about 1000 times broader than the homogeneous one. This means that, to find the number of terrylene molecules in the excitation spot and to determine their spectral positions in the inhomogeneous band, one have to scan the laser frequency over the inhomogeneous band while recording the intensity of the red-shifted fluorescence. Such excitation spectra contain several sharp peaks, each one corresponding to a single terrylene molecule.

As shown in Fig. 1, the emission of a single molecule under study (the one in resonance with the laser excitation) is collected by the same microscope objective, which is used to focus the laser light. Such a confocal configuration allows to very well restrict the collected light to the volume that is excited by the laser, thus decreasing the collection of unwanted stray light. After being reflected by a beam splitter, the emission is focused by the lens L₂ into a 10 μm diameter aperture (not shown in Fig. 1); the light that passes the aperture is focused onto a sensitive photomultiplier (PM), working in the regime of photon counting.

In order to register the resonance component of the single-molecule emission, no red-pass filters were used in front of the PM, thus also allowing the registration of the laser light, scattered and reflected from the optical elements. The purpose of the 10 μm aperture is to reject the most of the laser light, leaving only a small part reflected from a small spot on the crystal surface, thus decreasing the phase inhomogeneities caused by optical imperfections.

Although the laser light reflected from the surface of the crystal was 2500 times stronger than the resonance emission of the single molecule, it was still possible to observe the interference of these two emissions due to their coherence (in this case the ratio of the electromagnetic field amplitudes is of primary importance rather than the ratio of the light intensities, and the first one was only ).

By scanning the laser frequency over the resonance frequency of a chosen single molecule, changes of up to 4% were observed in the count rate of the PM (see Fig. 2). These changes can very well be described as a result of an interference of the single-molecule resonance fluorescence and of the laser light reflected from the crystal surface. It turns out that the different shapes of obtained spectra for different terrylene molecules can be very well described by different phase shifts between the two coherent waves, thus reflecting different distances of the molecules from the surface of the crystal. The phase shift between molecules b and c in Fig. 2 corresponds to a relative displacement of only 14 nm, which is a much higher precision than the best axial accuracy of the single-molecule localization achieved so far with a far-field optics (100 nm).

The conformation was also found for the theoretical prediction, according to which at higher excitation powers the single-molecule resonance fluorescence consists of two (coherent and incoherent) parts.

Because of its lower sensitivity (compared to the conventional technique of SMS) to a high-intensity background emission, the technique described could be applied for detection of single impurity molecules with lower luminescence quantum yield.

This study also rised the prospectives of single-molecule holography experiments.

Fig.2. Three examples of the PM count rate relative changes (dots) as a function of the laser frequency detuning in the vicinity of the electronic transition resonance frequency for three different terrylene single molecules. The laser power was 2 nW. The phase shifts are determined by fitting to the interference equation [22]. The data are normalized by the base-line countrate.