We prepared transparent polymer films doped with π-conjugated organic dyes around a multimode silica fiber and observed very narrow fluorescence peaks as compared with the fluorescence in solution. The peak position showed no dependence on the excitation wavelength, indicating that it could not be explained by a whispering gallery mode in a medium with broad optical gain. The peaks can be explained by amplified spontaneous emissions (ASE) because the intensity depended linearly on the excitation intensity with a threshold. When the dye-doped film was formed on the portion of a fiber with the clad etched out and was excited by the polarized laser propagating inside the fiber core, we observed ASE peaks not of the dyes but of the optical fiber itself, suggesting the possibility that the dyes were oriented with their transition moment pointing parallel to the film surface. These fiber ASE peaks shifted to longer wavelength when we varied the excitation wavelength to shorter wavelength, which clearly ruled out the possibility of silicate Raman scattering as the origin.

Self-organized organic dye particles of micrometer and submicrometer size were prepared by utilizing a wetting/dewetting process of polar solvent on a hydrophilic glass substrate. The near-field scanning optical microscopy successfully identified near-field excited near-field fluorescence from single particles, however, the majority of the small particles with diameters around 2 µm or less did not show fluorescence under near-field observation. In contrast, far-field fluorescence, when excited by a polarized evanescent field, was observed, with the intensity depending on the excitation polarization, indicating that molecules' transition moment within dye particles was oriented parallel to the substrate surface. Single particle fluorescence spectrum consistently showed an identical sharp peak with a large redshift, indicating that the particles were composed of identical dye aggregates similar to J-aggregates. These observations suggest that the near-field at the probe tip was polarized parallel to the probe axis. Another observation, that molecules were oriented in a similar direction among adjacent particles, suggests that the dewetting process contributed to the alignment of the molecular orientation among adjacent particles, which further proves that the present specimen was formed by a self-organizing mechanism.