Machine Learning for Real-Time Diagnostics of Cold Atmospheric Plasma Sources

Gidon, Dogan; Pei, Xiaobing; Bonzanini, Angelo D.; Graves, David B.; Mesbah, Ali

doi:10.1109/trpms.2019.2910220

Cited by 47 publications

(50 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nowadays, predictive modeling with machine learning is emerging as a ubiquitous technology, and its scope and application to material science and specifically to material design is receiving significant research attention. There are reported works on using machine learning algorithms to model plasma surface interactions occurring in a given target [118][119][120]. Employing machine-learning-based models not only helps to predict a given outcome, but it also opens possibilities to find the optimal and desired parameter combinations that would be ideal for the deposition process.…”

Section: Discussionmentioning

confidence: 99%

Atmospheric Pressure Plasma Deposition of TiO2: A Review

et al. 2020

View full text Add to dashboard Cite

Atmospheric pressure plasma (APP) deposition techniques are useful today because of their simplicity and their time and cost savings, particularly for growth of oxide films. Among the oxide materials, titanium dioxide (TiO2) has a wide range of applications in electronics, solar cells, and photocatalysis, which has made it an extremely popular research topic for decades. Here, we provide an overview of non-thermal APP deposition techniques for TiO2 thin film, some historical background, and some very recent findings and developments. First, we define non-thermal plasma, and then we describe the advantages of APP deposition. In addition, we explain the importance of TiO2 and then describe briefly the three deposition techniques used to date. We also compare the structural, electronic, and optical properties of TiO2 films deposited by different APP methods. Lastly, we examine the status of current research related to the effects of such deposition parameters as plasma power, feed gas, bias voltage, gas flow rate, and substrate temperature on the deposition rate, crystal phase, and other film properties. The examples given cover the most common APP deposition techniques for TiO2 growth to understand their advantages for specific applications. In addition, we discuss the important challenges that APP deposition is facing in this rapidly growing field.

show abstract

Section: Discussionmentioning

confidence: 99%

Atmospheric Pressure Plasma Deposition of TiO2: A Review

et al. 2020

View full text Add to dashboard Cite

show abstract

“…the label) can take various forms, including discrete components in classification methods, real-valued components in regression methods, or a mixture of discrete and real-valued components. In plasma applications, output can range from chemical, physical, and electrical properties of a target surface [4][5][6] to plasma properties such as degrees of molecular gas dissociation, plasma density, electron energy, neutral species rotational and vibrational temperature [6], or energetic and angular distribution of sputtered particles [7]. Input features for these ML applications in a plasma context can include, for example, optical emission spectra, current-voltage signals, electro-acoustic emission measurements, laser-induced fluorescence data, mass spectrometry data, and visual imaging.…”

Section: Machine Learningmentioning

confidence: 99%

“…We envision that ML will become indispensable for addressing major science and technological challenges in NEPs in the years ahead. [4,22], learning inexpensive surrogate models from theoretical simulation data [7], plasma dose quantification Selection of relevant input features for building simpler models from data [5] Diagnostics Inference of plasma and surface properties from spectral data [6,10,32], Extraction of latent information from measurements [46][47][48], Detection of abnormal drifts and variabilities [6] Process control…”

Section: Machine Learning For Process Control Of Nepmentioning

confidence: 99%

Machine learning for modeling, diagnostics, and control of non-equilibrium plasmas

Mesbah¹,

Graves²

2019

J. Phys. D: Appl. Phys.

Self Cite

101

View full text Add to dashboard Cite

Machine learning (ML) is a set of computational tools that can analyze and utilize large amounts of data for many different purposes. Recent breakthroughs in ML and artificial intelligence largely enabled by advances in computing power and parallel computing present cross-disciplinary research opportunities to exploit some of these techniques in the field of non-equilibrium plasma (NEP) studies. This paper presents our perspectives on how ML can potentially transform modeling and simulation, real-time monitoring, and control of NEP.

show abstract

“…[21] Studies have utilized machine learning for various purposes such as gas-phase analysis, [22,23] material characterization, [24] and process monitoring and control. [25,26] In addition, our previous study introduced the algorithm of artificial neural network (ANN) for plasma OES analysis to predict the conductivity of aqueous solution in which plasma is ignited. [27] The results show that employing ANN greatly improves the prediction accuracy of the solution conductivity, with the mean square error (MSE) being three orders of magnitude lower.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning with Explainable Artificial Intelligence Vision for Characterization of Solution Conductivity Using Optical Emission Spectroscopy of Plasma in Aqueous Solution

Wang

Hsu

2021

Plasma Processes & Polymers

View full text Add to dashboard Cite

This study presents an explainable artificial intelligence (XAI) vision for optical emission spectroscopy (OES) of plasma in aqueous solution. We aim to characterize the plasma and OES with XAI. Trained with 18000 spectra, a multilayer artificial neural network (ANN) model accurately predicted the solution conductivity. Local interpretable model-agnostics explanations (LIME), an XAI method, interpreted the model through perturbing spectral features and fitting the feature contribution with a linear model. LIME showed that OH, H γ , and H β emission lines were critical to the model, differing from the lines typically selected by humans. The results demonstrated that machine captured the spectral features neglected by humans. We believe using XAI for plasma OES analysis impacts the fields of plasma and analytical chemistry.

show abstract

Machine Learning for Real-Time Diagnostics of Cold Atmospheric Plasma Sources

Cited by 47 publications

References 47 publications

Atmospheric Pressure Plasma Deposition of TiO2: A Review

Atmospheric Pressure Plasma Deposition of TiO2: A Review

Machine learning for modeling, diagnostics, and control of non-equilibrium plasmas

Machine Learning with Explainable Artificial Intelligence Vision for Characterization of Solution Conductivity Using Optical Emission Spectroscopy of Plasma in Aqueous Solution

Contact Info

Product

Resources

About