Strained Indium Arsenide/gallium Arsenide Layers for Quantum Cascade Laser Design Using Genetic Algorithm PDF Download

Author: David W. Mueller
Publisher:
ISBN:
Category :
Languages : en
Pages : 134

Get Book Here

Book Description
Achieving high power, continuous wave, room temperature operation of midinfrared (3-5 [micrometer]) lasers is difficult due to the effects of auger recombination in band-to-band (type I) designs. Intersubband laser designs (type II) such as quantum cascade lasers reduce the effects of recombination, increasing efficiency and have advantages in large tunability of wavelength ranges (3 [micrometer] -25 [micrometer] and into the THz spectrum). Highly efficient quantum cascade laser designs are typically used in lasers designed for [greater than]5 [micrometer] wavelength operation due to the small offset of conduction band energy [delta]Ec in lattice matched materials. Some promising material systems have been used to achieve high-power output in the first atmospheric window (3-5 [micrometer]) but still suffer from low efficiency due to the lack of electron confinement. Larger [delta]Ec is attainable through the use of strained (lattice mis-matched) materials such as InGaAs/AlGaAs on GaAs. However, this material system has limitations on the traditional (100) crystal orientation due to the large strain and low critical thickness, hc. The necessity for controlled two-dimensional, optical quality layer growth limits the amount of strain incorporation due to defect formation in highly lattice mis-matched layers. The material systems used in this study are GaAs (100) and (111)B, AlGaAs, and (Ga)InAs. In the initial stage of research, I found that pseudomorphic growth of highly strained InAs layers on GaAs (111)B is possible. However, the growth window is very narrow and necessitates precise control over growth temperature and anion overpressure to achieve optical quality layers. As a result, a second stage of research explores the design space made available by this finding by using genetic algorithm based design and simulation of devices with a Schrödinger-Poisson solver - nextnano. This genetic algorithm is designed to rank candidate solutions on four important objectives for achieving novel QCL designs: operation between 3-5 [micrometer], Einj alignment just below E3, E2−E1 [almost equal to] ELO for optimal scattering for depopulation of carriers at the E2 energy level, and a gain metric [tau]3(1 − [tau]2/[tau]32). This approach was generally successful at finding unique designs, which have promise to work. However, since the algorithm ranked candidates based on only these four objectives, several interesting designs trends emerged. Many of the designs are far from the classic QCL structure and may not emit, however the trends help reveal the important design criteria. Discussion of this approach is compared to the traditional design philosophy and suggestions for improvement are centered around incorporating more objectives to guide the algorithm’s search. While improvements to the search mechanism are certainly necessary, several candidate solutions emerged and show great promise toward the goal of using this novel surface index and material system for efficient quantum cascade lasers that operate in the first atmospheric window.