APPLICATIONS OF STATISTICAL ANALYSIS METHODS ON PHOTONIC DEVICES

Deng, Xiangrui

Permalink

https://hdl.handle.net/2142/124774 Copy

Description

Title

APPLICATIONS OF STATISTICAL ANALYSIS METHODS ON PHOTONIC DEVICES

Author(s)

Deng, Xiangrui

Issue Date

2023-12-01

Keyword(s)

Volumetric Optical Integrated Circuit Elements (VOICE); Image Analysis; Optical Microscope; Edge Detection; Statistical Learning

Abstract

We apply statistical learning to predict device parameters based on the analysis of images obtained through device characterization. Previously, we fabricated and characterized volumetric optical integrated circuit elements (VOICE). Compared to traditional planar photonic devices, VOICE can be in principle more compact and denser, but introduce difficulties in characterization because the devices are in the volume of a wafer. Therefore, we would like to seek a fast and economical way to evaluate device performances. In this work, we develop statistical methods in Python to understand device performance, including the reflection and transmission spectra of VOICE devices. We implemented Python programs to: (1) balance out the non-uniform illumination of optical images and (2) standardize the format of each optical image. We parametrize 3D optical images and introduce support vector machine and gradient boosting models to classify devices and predict the figure of merit of each device. This work's broader impact is automating the process of characterizing novel photonic devices in a more costefficient way, without manually measuring the performance of each individual device.

Type of Resource

text

Language

eng

Owning Collections