Application of machine learning for predicting the millimetre-wave and sub-millimetre-wave characteristics of avalanche transit time sources

IF 2.5 4区工程技术 Q3 ENGINEERING, ELECTRICAL & ELECTRONIC

Journal of Computational Electronics Pub Date : 2025-07-07 DOI:10.1007/s10825-025-02382-7

Prerona Sanyal, Sneha Ray, Aritra Acharyya, Arndam Biswas, Rudra Sankar Dhar

{"title":"Application of machine learning for predicting the millimetre-wave and sub-millimetre-wave characteristics of avalanche transit time sources","authors":"Prerona Sanyal, Sneha Ray, Aritra Acharyya, Arndam Biswas, Rudra Sankar Dhar","doi":"10.1007/s10825-025-02382-7","DOIUrl":null,"url":null,"abstract":"<div><p>We explore the application of artificial neural networks (ANNs) for predicting the millimetre-wave (mm-wave) and sub-millimetre-wave (sub-mm-wave) characteristics of double-drift region (DDR) Si IMPATT diodes. The proposed ANN models predict key parameters such as DC, large-signal (L-S) performance, and avalanche noise characteristics across frequencies ranging from 94 to 500 GHz. A dataset derived from self-consistent quantum drift–diffusion (SCQDD) simulations is used to train the ANN models, which accurately capture the influence of structural, doping, and biasing variations. The ANN models showed a significant reduction in computational time, predicting device characteristics in just 4.4–15% of the time required by SCQDD simulations, while maintaining high accuracy. The mean square error (MSE) between ANN predictions and SCQDD simulations for breakdown voltage and power output was observed to be in the order of 10<sup>−3</sup> Unit<sup>2</sup>, indicating excellent predictive performance. The models were validated against experimental data, showing strong agreement in terms of power output, efficiency, and noise characteristics. This work demonstrates that machine learning can effectively replace traditional time-intensive simulations, making it a promising approach for the rapid design and optimization of high-frequency semiconductor devices.</p></div>","PeriodicalId":620,"journal":{"name":"Journal of Computational Electronics","volume":"24 5","pages":""},"PeriodicalIF":2.5000,"publicationDate":"2025-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Computational Electronics","FirstCategoryId":"5","ListUrlMain":"https://link.springer.com/article/10.1007/s10825-025-02382-7","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}

引用次数: 0

Abstract

We explore the application of artificial neural networks (ANNs) for predicting the millimetre-wave (mm-wave) and sub-millimetre-wave (sub-mm-wave) characteristics of double-drift region (DDR) Si IMPATT diodes. The proposed ANN models predict key parameters such as DC, large-signal (L-S) performance, and avalanche noise characteristics across frequencies ranging from 94 to 500 GHz. A dataset derived from self-consistent quantum drift–diffusion (SCQDD) simulations is used to train the ANN models, which accurately capture the influence of structural, doping, and biasing variations. The ANN models showed a significant reduction in computational time, predicting device characteristics in just 4.4–15% of the time required by SCQDD simulations, while maintaining high accuracy. The mean square error (MSE) between ANN predictions and SCQDD simulations for breakdown voltage and power output was observed to be in the order of 10⁻³ Unit², indicating excellent predictive performance. The models were validated against experimental data, showing strong agreement in terms of power output, efficiency, and noise characteristics. This work demonstrates that machine learning can effectively replace traditional time-intensive simulations, making it a promising approach for the rapid design and optimization of high-frequency semiconductor devices.

Abstract Image

查看原文本刊更多论文

机器学习在雪崩传输时间源毫米波和亚毫米波特性预测中的应用

我们探索了人工神经网络（ann）在预测双漂移区（DDR） Si IMPATT二极管毫米波（mm-wave）和亚毫米波（sub-mm-wave）特性中的应用。提出的人工神经网络模型预测了关键参数，如直流、大信号（L-S）性能和雪崩噪声特性，频率范围为94至500 GHz。来自自洽量子漂移-扩散（SCQDD）模拟的数据集用于训练人工神经网络模型，该模型准确捕获了结构，掺杂和偏倚变化的影响。ANN模型的计算时间显著减少，预测设备特性的时间仅为SCQDD模拟所需时间的4.4-15%，同时保持了较高的准确性。对于击穿电压和功率输出，ANN预测结果与SCQDD模拟结果的均方误差（MSE）为10−3 Unit2，表明ANN预测结果具有良好的预测性能。根据实验数据对模型进行了验证，在功率输出、效率和噪声特性方面显示出很强的一致性。这项工作表明，机器学习可以有效地取代传统的时间密集型模拟，使其成为高频半导体器件快速设计和优化的一种有前途的方法。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Computational Electronics ENGINEERING, ELECTRICAL & ELECTRONIC-PHYSICS, APPLIED

CiteScore

4.50

自引率

4.80%

发文量

142

审稿时长

>12 weeks

期刊介绍： he Journal of Computational Electronics brings together research on all aspects of modeling and simulation of modern electronics. This includes optical, electronic, mechanical, and quantum mechanical aspects, as well as research on the underlying mathematical algorithms and computational details. The related areas of energy conversion/storage and of molecular and biological systems, in which the thrust is on the charge transport, electronic, mechanical, and optical properties, are also covered. In particular, we encourage manuscripts dealing with device simulation; with optical and optoelectronic systems and photonics; with energy storage (e.g. batteries, fuel cells) and harvesting (e.g. photovoltaic), with simulation of circuits, VLSI layout, logic and architecture (based on, for example, CMOS devices, quantum-cellular automata, QBITs, or single-electron transistors); with electromagnetic simulations (such as microwave electronics and components); or with molecular and biological systems. However, in all these cases, the submitted manuscripts should explicitly address the electronic properties of the relevant systems, materials, or devices and/or present novel contributions to the physical models, computational strategies, or numerical algorithms.