Polarized Optical Gain and Polarization-Narrowing of Heavily Oxidized Porous Silicon

We report on a polarization-sensitive optical gain in a blue-emitting $\mathrm{S}\mathrm{i}/{\mathrm{S}\mathrm{i}\mathrm{O}}_2$ nanocrystalline system having a high degree of emission polarization memory. This system can show a positive optical gain or optical loss depending on the polarization state of the pump and emitted light. Under optical gain conditions, the degree of polarization of the amplified spontaneous emission increases with the pumping fluence. This effect has been attributed to an increase in the stimulated emission efficiency occurring for the linearly polarized emission component characterized by high photon occupation numbers (stimulating photon flux). This finding is independently supported by other experimental observations. The occurrence of polarization dependent stimulated emission strongly indicates the relevance of morphological effects in light emission from ultrasmall elongated silicon nanostructures.

Figure 1

Sketch of the VSL method with linearly polarized excitation (polarization labels are $V$ and $Z$). Edge emission ($I_{\mathrm{A}\mathrm{S}\mathrm{E}}$) is measured for two orthogonal polarization directions, labeled as $V$ and $H$. The thick two-sided arrow indicates the directions of displacement of the movable slit which define the excited stripe of length $\ell$. The sample is S1.

Figure 2

PL peak emission intensity versus the illuminated stripe length for various polarization configurations. The numbers shown in the frames are the gain coefficient ($g$) deduced by the one-dimensional amplifier fit of the data. The observation wavelength was centered at 460 nm. The sample was S1. The excitation fluence was $111\mathrm{m}\mathrm{J}{\mathrm{c}\mathrm{m}}^{-2}$. The first (second) letter of the label of each panel indicates the pumping (detection) polarization.

Figure 3

Integrated edge emission intensity versus the incident fluence for various polarization configurations and for a long stripe length (2.5 mm). The numbers on the various curves are the exponents deduced by a power law fit of the data. The observation wavelength was centered at 450 nm for $V$ excitation and at 460 nm for $Z$ excitation (the observation bandwidth was 90  nm). The sample studied was S1. The first (second) letter of the label of each panel indicates the pumping (detection) polarization.

Figure 4

The DOP of edge photoluminescence for a long (2.5 mm) stripe length, as a function of pumping fluence for vertically polarized (open circles, $V$-exc) and horizontally polarized (solid circles, $Z$-exc) excitation.

Figure 5

The DOP of photoluminescence as a function of pumping fluence, under edge excitation and front collection. The excitation beam was linearly polarized (labels $V$ and $H$ follow the convention displayed in Fig. 1). The experimental setup is sketched in the top left corner, where focusing, collecting, and polarizing/analyzing optics are not shown for clarity. The PL signal was collected at 90° to the sample plane at the edge border. The sample was the freestanding sample S2.