Optical Experiments on Semiconductors

1. Absorption

Figure 2.1. Detection by transmission.

Figure 2.2. GaAs Absorption vs. Energy.

Physics Department, University of Oxford

One method of determining the properties of a given sample, especially a semiconductor's bandgap, is to look at its photon absorption at varying energies of light. In a basic absorption experiment, light is incident on one side of the sample with a detector on the opposite side to measure the intensity of the light exiting the sample in comparison to the incident beam (see Figure 2.1). Absorption does not occur until the light is of a sufficient energy to allow the electron to jump the energy gap between the valence and conduction bands. Nonradiative transitions between energy levels are also possible if the energy is carried away by phonons (vibrations in the material's lattice). /or an absorption vs. energy graph for GaAs, see Figure 2.2. The sharp rise in the graph near 1.51 eV is indicative of the energy gap in GaAs. Only light that will cause transitions to excited states will be absorbed. It can then be concluded that the sample is transparent light with energy less than ~1.51 eV.

2. Reflection

Figure 2.3. Detection by reflection.

Figure 2.4. GaAs Reflectance vs. Energy.

Electronic Archive, Ioffe Physico-Technical Institute

Similar to absorption measurements are reflection experiments. Light is again incident on the sample, but the detector is situated on the same side as the light source and positioned at the same angle with respect to the surface (see Figure 2.3). The energy of the photons is then varied while the reflected light intensity is monitored. See Figure 2.4 for a reflectance vs. energy graph for GaAs. Analysis of a reflectance graph is useful in determining the band structure of the semiconductor. Changes to the band structure (and transitions) due to various doping levels can be understood by comparing reflectance graphs for each concentration of doping.

3. Photoluminescence

Figure 2.5. Detection by photoluminescence.

Figure 2.6. Photoluminescence (solid curves) of 3 GaAs single quantum wells at 300K.

K. Fujiwara,. N. Tsukada, and T. Nakayama. "Observation of free excitons in room-temperature photoluminescence of GaAs/AIGaAs single quantum wells." Appl. Phys. Lett. 53, 675-677.  

Photoluminescence is a useful experiment for the study of semiconductors as it may be used to determine its band gap. Light of a fixed wavelength is absorbed by electrons in the sample. The energy is radiated in all directions as the electrons drop to a lower energy state. Part of the emitted light is focused by a lens and fed into a spectrometer (see Figure 2.5). The relative intensity of the emitted light is measured as the wavelength analyzed by the spectrometer is varied. P.L. differs from absorption and reflection since it measures the light that is reradiated by the sample at various energies instead of catching the main reflected beam. A typical graph of emitted light intensity as a function of incident light wavelength is shown for GaAs in Figure 2.6.

4. Photoluminescence Excitation (P.L.E.)

Figure 2.7. GaAs P.L.E. vs. energy.

A. Babinski, et. al. "Photoluminescence excitation spectroscopy of InAs/GaAs quantum dots in high magnetic field." Physica E 22, 603-606.

P.L.E. uses the same experimental set-up as photoluminescence (see Figure 2.7), except the wavelength of the spectrometer is set to measure a fixed wavelength (usually the one corresponding to the energy gap) while the energy of the incident light is varied. In Figure 2.2 the first "bump" in the intensity is due to the formation of excitons, which form at a lower energy than that required for an electron to jump to the conduction band.

5. Kerr Rotation

Figure 2.8. Detection by Kerr Rotation.

Figure 2.9. GaAs electron spin resonance detected via Kerr Rotation.

For this experiment, linearly polarized light is incident on a magnetized material. The magnetization must have a component that is parallel to the direction of propagation of the light for the effect to be observed. The plane of polarization of the reflected light is different from that of the incident light after interacting with the sample as measured by a balanced detector. It differs from the Faraday effect since it measures the reflected light rather than the transmitted light (see Figure 2.8). Kerr rotation is useful for the detection of the coherence of electron spin since the changing spin will cause changes in the polarization of the incident light that can be measured over any time interval (see Figure 2.9).