Active-Region Design of Mid-Infrared Quantum Cascade Lasers via Machine Learning

We present a machine-learning (ML) approach for designing the active regions (ARs) of mid-infrared (IR) quantum cascade lasers (QCLs). This complex process, which conventionally may consume weeks to months to obtain an AR structure meeting a set of desired performance metrics, can be accomplished wi...

Full description

Saved in:
Bibliographic Details
Main Authors: Hu, Y., Suri, S., Kirch, J. D., Knipfer, B., Jacobs, S., Nair, S. K., Zhou, Z., Yu, Z., Botez, D., Mawst, L. J.
Format: Conference Proceeding
Language:English
Subjects:
active region
Inverse problems
Machine learning
Machine learning algorithms
Manuals
Measurement
mid-infrared
Prediction algorithms
QCL performance metrics
quantum cascade laser
tandem network
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a machine-learning (ML) approach for designing the active regions (ARs) of mid-infrared (IR) quantum cascade lasers (QCLs). This complex process, which conventionally may consume weeks to months to obtain an AR structure meeting a set of desired performance metrics, can be accomplished within a few seconds. Challenging tasks such as manual wavefunction identification, and the inverse problem of mapping performance metrics to AR structure are tackled through the application of ML algorithms. Initial training was performed on a large data set of over 100,000 AR structures and their corresponding performance metrics. Preliminary analysis suggests a low error of 2-15% in predicting various metrics.
ISSN:2575-274X
DOI:10.1109/IPC57732.2023.10360786