Etch Profile Prediction Model Using Convolutional Neural Network
Solid State Phenomena · 14.08.2023 · 346 · 236-243
DOI: 10.4028/p-c9ficg ↗Kurz gefasst
Fachartikel in der Zeitschrift Solid State Phenomena, online erschienen am 14.08.2023. Verfasst von Masahiro Tokuyama und 1 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 6 Literaturangaben.
Abstract (Originalsprache)
Etch profile control of the wafer surface is a key application for single-wafer wet process equipment. Wet etch processes are grouped into two types, either uniform flat etch profiles or specific non-flat etch profiles that are required for downstream processes. For both groups of etch profile it can consume time and resources to obtain the processing conditions to achieve the desired etch profile due to the complex interactions in the process. Etch profile prediction models can provide process engineers a valuable tool to identify processing conditions to get the desired etch profile in less time. In this paper, we introduce an etch profile prediction model using a Convolutional Neural Network [1] and validation of the prediction model against actual experimental data. We also investigated methods on how to select the comprehensive learning conditions and understand the relationship with prediction accuracy and the number of learning conditions. We found the choice of parameter type to describe a process condition can affect prediction accuracy. The prediction model reproduced the trend of etch profiles even when learned on a small dataset.
Zitieren
APA
Tokuyama, M., & Shinohara, K. (2023). Etch Profile Prediction Model Using Convolutional Neural Network. Solid State Phenomena, 346, 236-243. https://doi.org/10.4028/p-c9ficg
Vancouver
Tokuyama M, Shinohara K. Etch Profile Prediction Model Using Convolutional Neural Network. Solid State Phenomena. 2023;346:236-243. doi:10.4028/p-c9ficg
BibTeX
@article{Tokuyama2023,
author = {Tokuyama, Masahiro and Shinohara, Kensuke},
title = {Etch Profile Prediction Model Using Convolutional Neural Network},
journal = {Solid State Phenomena},
year = {2023},
volume = {346},
pages = {236--243},
doi = {10.4028/p-c9ficg}
}
EUVIMED