Bulletin of the American Physical Society
65th Annual Gaseous Electronics Conference
Volume 57, Number 8
Monday–Friday, October 22–26, 2012; Austin, Texas
Session AM3: Workshop on Verification and Validation of Low-Temperature Plasma Simulations |
Hide Abstracts |
Chair: Mirko Vukovic, Tokyo Electron Limited, and Miles Turner, Dublin City University Room: Classroom 202 |
Monday, October 22, 2012 1:30PM - 2:00PM |
AM3.00001: Overview of Verification and Validation in Low Temperature Plasma Physics M.M. Turner, M. Vukovic Computational work has formed an essential part of low-temperature plasma physics research for many years. As in many other fields, there is a tendency towards more complex simulations, as physical models become more elaborate and computer hardware becomes more sophisticated. Inevitably, this means that computer codes have become more complex. Complex physical models combined with complex implementation leads to considerable difficulty in establishing the fidelity of simulation results, as the physical model may contain assumptions that have been inadvertently violated, and the computer code may contain implementation errors. One needs a methodology for finding such mistakes, and also distinguishing between these two categories of mistake. In modern parlance, accumulating evidence that simulation codes are correct is known as Verification, while the process of testing the physical model is called Validation. This paper will present an overview of some important modern ideas on how Verification and Validation should be carried out, and discuss the implications of these concepts for the practice of computer simulation in low-temperature plasma physics. [Preview Abstract] |
Monday, October 22, 2012 2:00PM - 2:15PM |
AM3.00002: Verification strategies for fluid-based plasma simulation models Shankar Mahadevan Verification is an essential aspect of computational code development for models based on partial differential equations. However, verification of plasma models is often conducted internally by authors of these programs and not openly discussed. Several professional research bodies including the IEEE, AIAA, ASME and others have formulated standards for verification and validation (V{\&}V) of computational software. This work focuses on verification, defined succinctly as \textit{determining whether the mathematical model is solved correctly}. As plasma fluid models share several aspects with the Navier-Stokes equations used in Computational Fluid Dynamics (CFD), the CFD verification process is used as a guide. Steps in the verification process: consistency checks, examination of iterative, spatial and temporal convergence, and comparison with exact solutions, are described with examples from plasma modeling. The Method of Manufactured Solutions (MMS), which has been used to verify complex systems of PDEs in solid and fluid mechanics, is introduced. An example of the application of MMS to a self-consistent plasma fluid model using the local mean energy approximation is presented. The strengths and weaknesses of the techniques presented in this work are discussed. [Preview Abstract] |
Monday, October 22, 2012 2:15PM - 2:30PM |
AM3.00003: Towards the numerical verification of plasma simulation codes Mirko Vukovic To aid in verification of existing and new plasma simulation codes, we propose a suite of standard simulation problems against which a new code would be compared with. Each standard problem provides a detailed input specifications and results in forms of tables of numeric values. The problems use an idealized and simplified reaction cross-section and rates set. The problems are designed to verify individual numerical components of plasma simulation codes and the overall plasma simulation. The issue of establishing a ``correct'' plasma simulation result will be discussed. In addition, we will discuss the portability of these problems: the problems should be specified in a manner that can be read by simulation codes written in different languages, and executed on different platforms. [Preview Abstract] |
Monday, October 22, 2012 2:30PM - 3:00PM |
AM3.00004: BREAK |
Monday, October 22, 2012 3:00PM - 3:15PM |
AM3.00005: Benchmark solutions for simulations of capacitively coupled discharges M.M. Turner, D. Eremin, T. Mussenbrock, A. Derzsi, Z. Donko Benchmarks are an important element of Verification and Validation strategies. Such strategies define a process for increasing confidence in the fidelity of computer simulations, with the aim of making confident predictions of physical behaviour under conditions of practical interest. Such confidence can be increased by developing benchmark solutions for representative conditions. A benchmark solution is a high quality solution that is accepted to be correct. In this paper, we describe an attempt to develop such solutions for capacitive discharges, and we show that a number of independently developed particle-in-cell simulations can reproduce the benchmark solutions. These solutions are useful not only for particle-in-cell simulations, but also for other kinds of plasma simulations. We will show comparisons of fluid model solutions with the benchmarks. [Preview Abstract] |
Monday, October 22, 2012 3:15PM - 3:30PM |
AM3.00006: Verification of Particle-in-Cell Codes for Low Temperature Plasma Physics Keith Cartwright A broad overview of verification procedures for computer simulation with an emphasis in low temperature plasma physics will be presented. To have a high degree of confidence in simulations one needs code verification, code validation, solution verification, and uncertainty quantification. Code verification is a set of test developed to uncover coding mistakes that affect the numerical solution. Code validation addresses the appropriateness of the model in reproducing experimental data. Solution verification is the estimation of the numerical errors that occur in every computer simulation including validation and verification. Uncertainty quantification is the characterization of the sensitivity of results to parameters and geometry used in the models. The verification of the simulation code is an important first step in increasing the credibility of the results from simulations. Methods for generating of good verification problem for low temperature plasma physics will be shown. This includes the use of analytic solutions as well as using the method of manufactured and nearby solutions. DC sheaths with both pure electrons and electron and ion sheaths is the verification problem that will be discussed in detail. [Preview Abstract] |
Monday, October 22, 2012 3:30PM - 3:45PM |
AM3.00007: Inductively-coupled plasmas in pure chlorine: comparison experiments/HPEM Jean-Paul Booth, Nishant Sirse, Yasmina Azamoum, Pascal Chabert Inductively-coupled plasmas in chlorine-based gas mixtures are widely used for etching of nanometric features in silicon for CMOS device manufacture. This system is also of considerable fundamental interest as an archetype of strongly electronegative plasmas in a simple gas, for which reliable techniques exist to measure the densities of all key species. As such, it is an ideal test-bed for comparison of simulations to experiment. We have developed a technique based on two-photon Laser-Induced Fluorescence to determine the absolute Cl atom density. The Cl surface recombination coefficient was determined from time-resolved measurements in the afterglow. Electron densities were determined by microwave hairpin resonator and EEDF's were measured by Langmuir probe. Whereas the HPEM results were in good agreement at lower pressures (below 10mTorr), electron densities are increasingly underestimated at higher pressures. The gas temperature was measured by Doppler-resolved Infra-red Laser Absorption spectroscopy of Ar metastable atoms (with a small fraction Ar added). At higher pressures the gas temperature was considerably underestimated by the model. The concomitant overestimation of the gas density is a major reason for the disagreement between model and experiment. [Preview Abstract] |
Monday, October 22, 2012 3:45PM - 4:00PM |
AM3.00008: Breakdown voltage calculations using PIC-DSMC Paul Crozier, Jeremiah Boerner, Matthew Hopkins, Christopher Moore, Lawrence Musson In general, modeling and simulation provide physical insight and enable extrapolative predictions beyond theory and experimental data. In the specific case of electrostatic discharges, modeling and simulation may enable extrapolative predictions of breakdown voltages and a better physical understanding of breakdown phenomena. Using our PIC-DSMC software, we compute breakdown voltages for molecular nitrogen gas and compare our results against Bolsig+ for simple 1D geometries. We further verify our breakdown voltage calculations for a simple 3D geometry. In these calculations, 25 different N$_{2}$ -- electron interactions are included and good agreement with Bolsig+ is observed. Our approach to computing breakdown voltages using PIC-DSMC software can be extended to the prediction of breakdown voltages in more difficult cases where experimental data may be unavailable and the Paschen equation assumptions are no longer valid, as in the cases of complex 3D geometries and microscale discharges. [Preview Abstract] |
Monday, October 22, 2012 4:00PM - 4:30PM |
AM3.00009: Wrap-up/Discussion |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700