Quantum Optoelectronics

Abstract:

The Kerr gate technique is used to time-resolve the gain in an In0.02Ga0.98N/In0.16Ga0.84N multiple quantum well sample. A new way of analyzing the data in such a variable stripe length method gain experiment is used to analyze both the time-integrated and time-resolved spectra. We confirm that the stripe length dependence of the gain in the multiple quantum wells under nanosecond excitation is caused by the change of the chemical potential along the excited stripe due to the interaction of the carrier and photon densities, and the gain threshold density is estimated. A trial function assuming a Lorentzian line shape for the stripe length dependence of the gain is compared with the edge emission intensity. This is found to fit very well with our data, even beyond the saturation region. Furthermore, we have extended the investigation to examine the dynamics of the emission and gain. These measurements suggest that the photoexcited carriers must localize (possibly at indium-rich sites) before strong stimulated emission is seen. © 2002 Elsevier Science B.V. All rights reserved.

Abstract:

The Kerr gate technique is used to time-resolve the gain in an In0.02Ga0.98N/In0.16Ga0.84N multiple quantum well sample. A new way of analyzing the data in such a variable stripe length method gain experiment is used to analyze both the time-integrated and time-resolved spectra. We confirm that the stripe length dependence of the gain in the multiple quantum wells under nanosecond excitation is caused by the change of the chemical potential along the excited stripe due to the interaction of the carrier and photon densities, and the gain threshold density is estimated. A trial function assuming a Lorentzian line shape for the stripe length dependence of the gain is compared with the edge emission intensity. This is found to fit very well with our data, even beyond the saturation region. Furthermore, we have extended the investigation to examine the dynamics of the emission and gain. These measurements suggest that the photoexcited carriers must localize (possibly at indium-rich sites) before strong stimulated emission is seen. (C) 2002 Elsevier Science B.V. All rights reserved.

Abstract:

Spectrally resolved and time-integrated four-wave mixing are used to measure the polarization dependence of biexcitonic signals and quantum beats between two-A-exciton (XAXAZ*) and A-biexciton (XAXA) states in a high-quality GaN epilayer. Mixed beats with two periods are observed: the first beating period corresponds to the energy splitting between XAXA*and XAXA; the second period corresponds to beating between A excitons (XA) and donor bound excitons (D^OX). We also measure the polarization-dependent B-biexciton (XBXB) signal. The effective masses for the A and B holes are deduced from the binding energy.

Abstract:

Non-thermal carrier states at early times are studied using femtosecond pump-probe spectroscopy in GaN. After the residual chirp on the continuum probe is removed, the normalized difference spectra (NDS) for different probe energies are synchronized, recovering the full time resolution of our laser pulse (120 fs). Our Monte-Carlo simulation agree well with the unchirped NDS spectrum, which shows the development of the carrier distribution at early times, where phonon satellites are seen, together with a strong non-thermal electron distribution in the region of the LO-phonon energy arising from the remarkably strong electron-LO phonon interaction. Employing a new technique which involves the integration of the normalized NDS multiplied by the corresponding energy, a measure of the mean energy of the carriers in non-thermal states is obtained. By comparing the time-dependent energy loss with the theoretical energy loss rate, we estimate the effective temperature of the phonon modes as well as the population of phonons. Our Monte-Carlo model agrees well with this data, and confirms the dominance of hot phonon effects at early times. © 2002 Elsevier Science B.V. All rights reserved.