INVESTIGATING THE EFFECTS OF INTERFACE ROUGHNESS IN INDIUM GALLIUM ARSENIDE/ALUMINUM INDIUM ARSENIDE QUANTUM CASCADE LASERS

Abstract

Interface roughness in Quantum Cascade (QC) lasers has been reported to influence intersubband absorption, intersubband broadening, and various transport mechanisms [1]. Here, we investigate the effects of interface roughness in InGaAs/AlInAs material on broadening in particular. The hypothesis is that broadening increases with the number of barriers in a QC laser design; absorption measurements of electrons from the ground energy state to the second energy state for a given sample should dictate how broadening changes with the number of interfaces. Two sets of samples were measured, each set containing three samples. For samples P448 (no barrier), P449 (one barrier), and P450 (two barriers), a Lorentzian and linear fit to the absorption spectra reveal a width of approximately 136 cm-1, 139 cm-1, and 139 cm-1 for each sample, respectively. The first set of samples corresponds to an energy difference E2 – E1 of around 160 meV. For samples P482 (no barrier), P483 (three barriers), and P484 (five barriers), a Lorentzian and linear fit to the absorption spectra convey a width of approximately 146 cm-1, 160 cm-1, and 298 cm-1 for each sample, respectively. This second set of samples corresponds to an energy difference E2 – E1 of about 200 meV. All absorption spectra were obtained using a onepass transmission setup, which will be elaborated upon later on in this work. A plot of number of barriers versus width for each sample reveals that for the first set (Samples P448, P449, and P450), width does not necessarily increase with increasing number of barriers in each design. However, this trend seems to be evident for the second set (Samples P482, P483, and P484). Another plot of expected energy difference E2 – E1 versus measured energy difference E2 – E1 illustrates that these values are not very close for the first set of samples, but they are closer for the second set of samples. It has been speculated that calibration may have drifted for the first set of growth, which may have onset the deviations in the expected energy difference versus the measured energy difference. Future work on this project includes designing new sets of samples with varying E2 – E1 energy differences and number of barriers, growing the samples, and obtaining absorption spectra measurements.