Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms
Abstract As an essential block in optical communication systems, silicon (Si) Mach–Zehnder modulators (MZMs) are approaching the limits of possible performance for high-speed applications. However, due to a large number of design parameters and the complex simulation of these devices, achieving high...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-09-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-41558-8 |
_version_ | 1797559663162556416 |
---|---|
author | Romulo Aparecido de Paula Ivan Aldaya Tiago Sutili Rafael C. Figueiredo Julian L. Pita Yesica R. R. Bustamante |
author_facet | Romulo Aparecido de Paula Ivan Aldaya Tiago Sutili Rafael C. Figueiredo Julian L. Pita Yesica R. R. Bustamante |
author_sort | Romulo Aparecido de Paula |
collection | DOAJ |
description | Abstract As an essential block in optical communication systems, silicon (Si) Mach–Zehnder modulators (MZMs) are approaching the limits of possible performance for high-speed applications. However, due to a large number of design parameters and the complex simulation of these devices, achieving high-performance configuration employing conventional optimization methods result in prohibitively long times and use of resources. Here, we propose a design methodology based on artificial neural networks and heuristic optimization that significantly reduces the complexity of the optimization process. First, we implemented a deep neural network model to substitute the 3D electromagnetic simulation of a Si-based MZM, whereas subsequently, this model is used to estimate the figure of merit within the heuristic optimizer, which, in our case, is the differential evolution algorithm. By applying this method to CMOS-compatible MZMs, we find new optimized configurations in terms of electro-optical bandwidth, insertion loss, and half-wave voltage. In particular, we achieve configurations of MZMs with a $$40~\text {GHz}$$ 40 GHz bandwidth and a driving voltage of $$6.25~\text {V}$$ 6.25 V , or, alternatively, $$47.5~\text {GHz}$$ 47.5 GHz with a driving voltage of $$8~\text {V}$$ 8 V . Furthermore, the faster simulation allowed optimizing MZM subject to different constraints, which permits us to explore the possible performance boundary of this type of MZMs. |
first_indexed | 2024-03-10T17:48:31Z |
format | Article |
id | doaj.art-07c7c62b906c4530ba2f4ceabe9de431 |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-03-10T17:48:31Z |
publishDate | 2023-09-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-07c7c62b906c4530ba2f4ceabe9de4312023-11-20T09:25:36ZengNature PortfolioScientific Reports2045-23222023-09-0113111210.1038/s41598-023-41558-8Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithmsRomulo Aparecido de Paula0Ivan Aldaya1Tiago Sutili2Rafael C. Figueiredo3Julian L. Pita4Yesica R. R. Bustamante5Center for Advanced and Sustainable Technologies, State University of Sao Paulo (UNESP)Center for Advanced and Sustainable Technologies, State University of Sao Paulo (UNESP)Centre for Research and Development in Telecommunications (CPQD)Centre for Research and Development in Telecommunications (CPQD)Department of Electrical Engineering, École de Technologie Supérieure (ÉTS)Centre for Research and Development in Telecommunications (CPQD)Abstract As an essential block in optical communication systems, silicon (Si) Mach–Zehnder modulators (MZMs) are approaching the limits of possible performance for high-speed applications. However, due to a large number of design parameters and the complex simulation of these devices, achieving high-performance configuration employing conventional optimization methods result in prohibitively long times and use of resources. Here, we propose a design methodology based on artificial neural networks and heuristic optimization that significantly reduces the complexity of the optimization process. First, we implemented a deep neural network model to substitute the 3D electromagnetic simulation of a Si-based MZM, whereas subsequently, this model is used to estimate the figure of merit within the heuristic optimizer, which, in our case, is the differential evolution algorithm. By applying this method to CMOS-compatible MZMs, we find new optimized configurations in terms of electro-optical bandwidth, insertion loss, and half-wave voltage. In particular, we achieve configurations of MZMs with a $$40~\text {GHz}$$ 40 GHz bandwidth and a driving voltage of $$6.25~\text {V}$$ 6.25 V , or, alternatively, $$47.5~\text {GHz}$$ 47.5 GHz with a driving voltage of $$8~\text {V}$$ 8 V . Furthermore, the faster simulation allowed optimizing MZM subject to different constraints, which permits us to explore the possible performance boundary of this type of MZMs.https://doi.org/10.1038/s41598-023-41558-8 |
spellingShingle | Romulo Aparecido de Paula Ivan Aldaya Tiago Sutili Rafael C. Figueiredo Julian L. Pita Yesica R. R. Bustamante Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms Scientific Reports |
title | Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms |
title_full | Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms |
title_fullStr | Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms |
title_full_unstemmed | Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms |
title_short | Design of a silicon Mach–Zehnder modulator via deep learning and evolutionary algorithms |
title_sort | design of a silicon mach zehnder modulator via deep learning and evolutionary algorithms |
url | https://doi.org/10.1038/s41598-023-41558-8 |
work_keys_str_mv | AT romuloaparecidodepaula designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms AT ivanaldaya designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms AT tiagosutili designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms AT rafaelcfigueiredo designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms AT julianlpita designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms AT yesicarrbustamante designofasiliconmachzehndermodulatorviadeeplearningandevolutionaryalgorithms |