Surface plasmon polaritons (SPPs) are electromagnetic waves travelling along the surface of metals accompagnied by charge oscillations in the metal. They can be generated, for example, by electrons passing inelastically from a metallic tip to a metallic substrate in a photon scanning tunneling microscope (PSTM) [1]. Plasmons generated near the STM tip apex can transform into a free electromagnetic wave (light wave) which is then detected at some macroscopic distance from the tunnel junction by a photon detector. The control of dc voltage and the electron current in the tunnel junction allows experiments with plasmons very different from the laser-induced plasmon generation in an apertureless SNOM.

By placing the STM tip above individual nano-objects on a surface it becomes possible to excite them with the STM current and to record spectra of the emitted light. Here, the plasmons couple the local emission to the light field which is detected by a spectrometer or a single photon counter (e.g. avalanche photo detector). Nano-objects to be studied in this way can be nanocrystals built of organic molecules with extended electronic pi-systems, single luminescent molecules, semiconductor quantum dots, and quantum wires. Standard measurements of the low-temperature STM can in addition provide highly resolved images of the studied nano object and spatially resolved differential conductance spectra.

Let us focus on three topics of current interest: Spectroscopic identification of the light emitter, localization of the emission process, and spectroscopic features due to the small size of the emitters. For this we look at the data from two systems [2,3] (see schematic figures above)
(a) in UHV prepared pentacene nanocrystals adsorbed on an ultrathin KCl layer on Au(111) and
(b) CdSe nanowires deposited from solution on a gold film.

Spectroscopic identification of the emitter is possible by assignment of the recorded spectra. In the case of pentacene nanocrystals [2] the emission line (red spectra above) is assigned to a pentacene exciton due to its significant spectral red-shift with respect to the emission from the single molecule transition. The light emission appears thus as a collective feature - probably of the entire nano crystal. We find that the pentacene emission (red spectra) is well reproducible and much narrower than the spectra of plasmon polaritons (green spectra) which were recorded with the STM tip positioned at some distance from the pentacene crystal. Plasmon spectra depend on tip-shape and on the metals used as tip and substrate (see labels and schematic drawings at the top of the figure). Moreover, we find that the confinement of the excitation inside the pentacene crystals which might lead to a spectral blue-shift has no measurable effect on the wavelength of the emission. While the crystal thicknesses varied from 1.5nm to 4nm and their lateral extension from 10 nm to more than 50 nm the systematic spectral shift is less than 10 meV.

If we now turn to the CdSe nanowires [3] we find that there indeed, size is an important parameter for the position of the emission line (top panel in the figure above). The diameter of each wire (18nm, 12nm, 9nm) is obtained from the apparent height in the topographic scan. When the peak photon energies are ploted (lower graph) and extrapolated (red line) towards infinite wire diameter we find the known low-temperature emission energy of the CdSe zincblende structure. Interestingly, due to growth conditions this type of structure represents only a small fraction of the nanowire while most of its length exhibits wurtzite structure. This suggests that only small sections of the wire participate in the emission of light which is indeed seen in photon maps showing bright and dark regions on the CdSe wire (see [3], not shown here).
We discussed so far the characterization of the emitting object and the emission process. It will also be interesting to target in the future the excitation process, which can vary strongly for different types of samples. Other topics of current interest are organic layers composed of two different types of organic molecules and the shift of the wavelength of the emitted light which can occur due to the very strong electric fields (of the order of 1V/nm) in the tunnel gap of the STM.

References:

[1]   K. Kuhnke, A. Kabakchiev, W. Stiepany, F. Zinser, R. Vogelgesang, and K. Kern ; Rev. Sci. Instr. 81, 113102 (2010)
[2]   A. Kabakchiev, K. Kuhnke, T. Lutz, and K. Kern; ChemPhysChem  11 (2010) 3412
[3]   T. Lutz, A. Kabakchiev, T. Dufaux, C. Wolpert, Z. Wang, M. Burghard, K. Kuhnke, and K. Kern ; Small 7 (2011) 2396