Intensity Dependent Cyclotron Resonance in a Gallium Arsenide-Aluminum Gallium Arsenide Two-Dimensional Electron Gas

Linear cyclotron resonance measurements on two -dimensional electron gas systems have provided important information about the carrier density, mean scattering time and effective mass. The technological applications of GaAS/AlGaAs heterostructures require an understanding of the system's behavior under extreme conditions. Luminescence experiments under intense dc electric fields have revealed hot electron behavior manifested in a decrease of the sample's mobility. In this dissertation techniques are developed to study the intensity dependence of the carrier density, mean scattering time and effective mass through analysis of cyclotron resonance transmission lineshapes. A high-power superfluorescent far infrared laser provides intensity levels in excess of 1 MW/cm ^{it 2} at several discrete frequencies. GaAs/AlGaAs samples with densities between 0.9-3.7 times 10^ {11} cm^{it 2} have been studied at frequencies between 8.2 cm^{-1} and 110.3 cm^{-1}. Cyclotron resonance transmission lineshapes are obtained by scanning the magnetic field at fixed laser frequency. The carrier density, mean scattering time and effective mass are obtained by fitting the lineshapes to the calculated transmission. For all samples and at every frequency, the intensity dependence of the fit parameters is similar: the carrier density remains nearly constant, while the mean relaxation time decreases and the effective mass increases at intensity levels above ~2 W/cm^{it 2}. The observed behavior for the mean scattering time is surprisingly similar to the one obtained from the dc electric field luminescence experiments even though an ac electric field excites the system in the cyclotron resonance experiments. Our results are not consistent with previous reports of saturation of a multi-level system. A simple multi -level system model which includes stimulated emission and absorption processes together with spontaneous emission processes fails to reproduce the experimental observations. By comparing the intensity dependent and lattice temperature induced resonance frequency shifts we conclude that at high frequency (110.3 cm^{-1} ) the carriers are not heated enough to populate the n = 2 Landau level. Hence, the observed changes can not be attributed to hot electron behavior. Instead, since the heating does not follow the resonant shape of the absorption, a non-resonant process must be important.