Bridging TCAD and AI: Its Application to Semiconductor Design

Jeong, Changwook; Myung, Sanghoon; Huh, In; Choi, Byung-Soon; Kim, Jin-Woo; Jang, Hyung Wook; Lee, Hojoon; Park, Daeyoung; Lee, Kyuhun; Jang, Wonik; Ryu, Jisu; Cha, Moon-Hyun; Choe, Jae Myung; Shim, Mun‐Bo; Kim, Dae Sin

doi:10.1109/ted.2021.3093844

Cited by 37 publications

(13 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data-driven approaches are used to characterize defects automatically by assigning classes to wafer map patters, morphology, and chemical spectra [568,569], as well as detecting and triggering auto-clean routine to improve productivity by minimizing the failures caused by these defects. In addition to process challenges at the unit process level, data-driven approaches can be used for optimizing the entire process flow, allowing engineers to study the sensitivity of a particular layer and build appropriate trade-offs to achieve their desired product [570].…”

Section: B Data-driven Approaches For Plasma-assistedmentioning

confidence: 99%

2022 Review of Data-Driven Plasma Science

Archibald¹,

Asif²,

Becker³

et al. 2022

Preprint

View full text Add to dashboard Cite

Data science and technology offer transformative tools and methods to science. This review article highlights latest development and progress in the interdisciplinary field of data-driven plasma science (DDPS). A large amount of data and machine learning algorithms go hand in hand. Most plasma data, whether experimental, observational or computational, are generated or collected by machines today. It is now becoming impractical for humans to analyze all the data manually. Therefore, it is imperative to train machines to analyze and interpret (eventually) such data as intelligently as humans but far more efficiently in quantity. Despite the recent impressive progress in applications of data science to plasma science and technology, the emerging field of DDPS is still in its infancy. Fueled by some of the most challenging problems such as fusion energy, plasmaprocessing of materials, and fundamental understanding of the universe through observable plasma phenomena, it is expected that DDPS continues to benefit significantly from the interdisciplinary marriage between plasma science and data science into the foreseeable future. CONTENTSI. Introduction 6 II. Fundamental Data Science 6 A. Introduction 6 B. Data Reduction/Compression 7 C. Dimensional reduction and sparse modeling 8 D. ML-enhanced modeling and simulation 9 E. ML Hardware and integration with models 10 F. Workflow Automation 11 G. Uncertainty Quantification 12 H. Visualization and Data Understanding 13 I. ML Control Theory 15 III. Basic Plasma Physics and Laboratory Experiments A. Introduction B. Spectroscopy, imaging and tomography C. Sparse measurement and noise D. Synthetic instruments and data E. Experimental data visualization F. High-rep rate laser experiments G. Charged particle beams H. Control and Optimisation of Plasma Accelerator Experiments I. Dusty and complex plasmas J. Physics and machine learning K. Challenges and outlook 5 IV. Magnetic Confinement Fusion 36 A. Introduction 36 B. Data-Driven Physics Models 36 C. Optimizing experimental workflows with data-driven methods 37 D. Diagnostics and Fusion Data Streams 38 E. Prediction of Tokamak Disruption 39 F. Surrogate models of fusion plasma 40 G. Magnetic Fusion Energy Data Challenges and Solutions 41 H. Data Science for extreme scale simulation 43 I. Challenges and outlook 44 V. Inertial confinement fusion and high-energy-density physics 45 A. Introduction 45 B. Representation learning for multimodal data 45 C. Transfer learning for simulation and experimen 47 D. Uncertainty quantification and Bayesian inference 47 E. High-performance computing and simulation acceleration 49 F. Design exploration and optimization 49 G. Self-driving experimental facilities 51 H. Challenges and outlook 52 VI. Space and astronomical plasmas 52 A. Introduction 52 B. Space and ground instruments 53 C. Space weather prediction 55 D. Transfer learning to improve historic data 56 E. Surrogate models of fluid closures using machine learning 56 F. Magnetic reconnection 58 G. Challenges and outlook 60 VII. Plasma tec...

show abstract

Section: B Data-driven Approaches For Plasma-assistedmentioning

confidence: 99%

2022 Review of Data-Driven Plasma Science

Archibald¹,

Asif²,

Becker³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…However, due to the complexity of parameter adjustment and the uncertainty of simulation results, it always spends much time on device simulation, which is an unavoidable problem. With the development of artificial intelligence technology, the data-driven model appeared in the context of the integration of machine learning technology and TCAD, and can make full use of data to solve and optimize scientific problems [12] . Meanwhile, TCAD simulation is based on numerical calculation, and a series of scientific semiconductor physical theories and formulas are used for relevant calculation [11] .…”

Section: Introductionmentioning

confidence: 99%

The performance prediction model of NMOSFET based on BP neural network

Feng

2023

Third International Conference on Sensors and Information Technology (ICSI 2023)

View full text Add to dashboard Cite

A neural network prediction model for NMOFET device is proposed in this paper by using BP algorithm in machine learning technology and Silvaco TCAD simulation tool, so as to improve the efficiency of actual simulation work and provide a new method for the performance research of NMOSFET. In the process of modeling, the substrate bias, the impurity concentration of substrate, the thickness of oxide and the threshold voltage adjustment implant doping concentration are regarded as independent variable, and substituted into the simulation software for simulation. The performance parameters of NMOSFET are obtained as the dependent variable, such as the threshold voltage, the maximum transconductance, the subthreshold slope and the Ion/Ioff ratio. All the simulation results are used to assisted the training and prediction of BP neural network. The results of experiment show that the BP neural network prediction model with Levenberg-Marquardt algorithm can well predict the performance of NMOSFET devices. The determination coefficient of all models is above 0.99, and the mean absolute percentage error of all models is below 1.5%.

show abstract

“…Few works have been reported in the past that attempted to analyze the effects of fluctuations in GAA NS MOSFETs using device simulation [17], [18]. However, despite the high accuracy of device simulation, it is a very time-consuming process and desired alternative time-efficient methods [19].…”

Section: Introductionmentioning

confidence: 99%

A Machine Learning Approach to Modeling Intrinsic Parameter Fluctuation of Gate-All-Around Si Nanosheet MOSFETs

Butola

Kola

2022

IEEE Access

View full text Add to dashboard Cite

The sensitivity of semiconductor devices to any microscopic perturbation is increasing with the continuous shrinking of device technology. Even the small fluctuations have become more acute for highly scaled nano-devices. Therefore, these fluctuations need to be addressed extensively in order to continue further device scaling. In this paper, we mainly focus on three intrinsic parameter fluctuation sources, work function fluctuation (WKF), random dopant fluctuation (RDF), and interface trap fluctuation (ITF) for gate-all-around (GAA) silicon nanosheet (NS) MOSFETs. Generally, the effect of these fluctuations is analyzed using a time-consuming device simulation process. A machine learning (ML) based powerful and efficient artificial neural network (ANN) model is used to accelerate this process. Firstly, the effects of fluctuation sources are analyzed individually by using the ANN model and results have been presented that show the WKF variations dominate the threshold voltage (VTH), off-state current (IOFF), and on-state current (ION) variations among other fluctuation sources. Next, we examine the combined effect of all three fluctuation sources. It is crucial because considering only one fluctuation can result in unexpected variations due to other fluctuations present in the device. Therefore, the ANN model is used to estimate the combined effects as well. The results show that the proposed model predicts the outputs with an R 2 -score of 99% and an error rate of less than 1%. In addition, the ML is also utilized to calculate the permutation importance of input variables as a measure to investigate the contribution of each fluctuation source.

show abstract

Bridging TCAD and AI: Its Application to Semiconductor Design

Cited by 37 publications

References 29 publications

2022 Review of Data-Driven Plasma Science

2022 Review of Data-Driven Plasma Science

The performance prediction model of NMOSFET based on BP neural network

A Machine Learning Approach to Modeling Intrinsic Parameter Fluctuation of Gate-All-Around Si Nanosheet MOSFETs

Contact Info

Product

Resources

About