Funneling propagating photons into single molecules

Two decades ago it was commonly believed that single molecules were too small to be seen using an optical microscope. Today, single-molecule fluorescence microscopy has become a standard tool in many biology laboratories. For example, the diffusion or directed motion of proteins labeled with a dye molecule can be detected and tracked by the fluorescent emission of the dye. An advantage of this approach is its specificity; each dye molecule emits at a particular wavelength, enabling the excitation light to be filtered out spectrally. Moreover, one can even count individual photons of fluorescence on a zeroed background. However, organic dye molecules typically used in these processes are prone to photobleaching after ∼1min of illumination. The technique also relies on the high fluorescent quantum yield of the target. Recently, there has been increasing interest in detecting single molecules that do not fluoresce. Here, we image single molecules directly in transmission mode using optics with very high numerical apertures.

Our general approach to detecting molecules and nano-particles without using fluorescence is similar to a standard absorption experiment.^1–6 That is, we examine the effect of a single nano-object on the intensity of a light beam. If the object absorbs or scatters the incident light, the intensity in the forward direction is diminished. The strength of the interaction and degree of attenuation is usually expressed in terms of the ratio σ:A, where σ is the optical cross-section of the object and A is the area where light was confined. Despite the general perception that σ can be very small, calculations tell us that σ≈λ²/2 for an atom or molecule suspended in a vacuum. This is comparable to tight beam focus of 200–250nm when operating at typical visible wavelengths (λ=450–650nm) and using a high numerical aperture (1.4). Unfortunately, for molecules embedded in solids, σ is reduced because of coupling with the thermal bath of the surrounding matrix. This can be avoided by conducting measurements at temperatures below 2K in liquid helium. However, in such cryogenic conditions, the diffraction limit is difficult to achieve and high-quality imaging is impaired.

At room temperature and in conventional microscopes, oils are used in high-end imaging as immersion media to increase the refractive index and, therefore, the numerical aperture. For measurements at 2K, we emulated this using a solid immersion medium made of a half-ball lens (i.e., solid-immersion lens) of cubic zirconia. The advantage of this approach is that, if the part of the sample to be studied lies at the center of the half-ball lens, the rays remain radial and unbroken. We then use a simple second lens—with fairly low numerical aperture—to reach the diffraction limit of a few hundred square nanometers. Figure 1 shows ∼20% attenuation of a laser beam by a single molecule of dibenzanthanthrene recorded in a cryostat.⁴ Such strong coherent interactions between an incoming photon with a molecule allow optical signal processing at the nanometer and quantum scales. For example, we have shown that a single molecule can act as a nanoscopic transistor that can attenuate or amplify an incoming stream of photons.³ Furthermore, a molecule can apply a substantial phase shift on a laser beam.^{4, 5}

Figure 1. The spectrum shows nearly 20% extinction of power from a laser beam incident on a single molecule of dibenzanthanthrene at a temperature of 1.5K.

The origin of the extinction signal is the interference between the incident light and coherent scattering by the molecule. The theoretical limit of this interaction depends on both the spatial mode of the incident beam and degree of focusing by the lenses. Theory shows that full extinction in the forward direction and perfect reflection of the incident light can be possible if one matches the spatial modes of the scattered and incident fields.² However, thus far this has not been achieved in the laboratory and represents an ongoing challenge for us and others.

Parallel to our cryogenic work, we exploited this tight focusing of light to detect single nanoparticles and molecules at room temperature.^{5, 6} For nanoparticles, the volume and refractive index contrast against the surrounding medium determines their polarizability and scattering cross-sections. Recently, we succeeded in detecting and imaging single molecules in extinction mode at room temperature (see Figure 2).^5,6 The difficulty of working at ambient conditions was not the imaging optics. In fact, we used commercial microscope objectives with numerical apertures ∼1.4. The challenge instead lies in the cross-section of the molecule, which is approximately one million times smaller than at liquid helium temperature. This ‘shrinking’ is caused by thermal agitations of the matrix at room temperature, which perturb the phase of the molecular dipole moment. As a result, the extinction signal is easily lost in the laser noise and other fluctuations.

Figure 2. A raster scan absorption image of a single molecule of terrylene diimide at room temperature. The absorption signal is of the order of one per million. T: Temperature.

We successfully eliminated laser fluctuations by splitting the excitation beam to a reference beam—sent directly to a photodiode—and signal beam that first traversed the sample and then landed on a second photodiode. We subtracted the outputs of the two photodiodes to account for laser intensity fluctuations. By index-matching the substrate with the immersion oil, we eliminated sample fluctuations caused by nanoscopic substrate roughness. Finally, we subtracted an image containing the molecule from an image without it to remove any residual fluctuations caused by the sample or setup. This demonstration of single-molecule sensitivity in extinction mode creates opportunities for detection and spectroscopy of nanomaterials that do not fluoresce. In particular, biological entities such as proteins and viruses can be detected and imaged without the need for labeling.

In summary, we devised methods for the strong focusing of light into single molecules and nanoparticles. Our ability to reach such strong interactions between photons and quantum emitters has significant implications for the fast-moving fields of quantum information processing and engineering. We are currently working to replace laser beams with single photons emitted by individual molecules. This will allow transfer of individual photons between molecules over large, macroscopic distances. One future direction will be to miniaturize single-photon sources and single-molecule optical elements along nanoguides on a chip to process quantum optical signals. Another very promising application lies in label-free biosensing.

I am indebted to many talented doctoral students, postdoctoral fellows, and senior scientists who have made our work at The Swiss Federal Institute of Technology (ETH) Zurich, Switzerland possible over the years: M. Agio, M. Celebrano, Y. Chassagneux, I. Gerhardt, S. Götzinger, J. Hwang, P. Kukura, R. Lettow, N. Mojarad, M. Pototschnig, A. Renn, G. Wrigge, and G. Zumofen.