DOE Center for  

Predictive Control of Plasma Kinetics:  

Multi‐Phase and Bounded Systems 

 

8th Annual Meeting 

June 1‐2, 2017 Bethesda North Marriott Hotel and Conference Center 

Bethesda, MD 

2

Participating Institutions 

We gratefully acknowledge the funding from  

The U.S. Department of Energy Office of Science  

Fusion Energy Sciences Program 

Grant # DE‐SC0001939 

   

3

Schedule Thursday, June 1, 2017 7:45 – 8:00 am

Registration

8:00 – 8:15 Mark J. Kushner (University of Michigan)

Introduction to the Annual Meeting: The Role of the Center in the Future of Low Temperature Plasmas

Abstract: p. 10

8:15 –9:55 Session I. Plasma-Surface Interactions: Solid and

Liquid Moderator: Sophia Gershman (Princeton Plasma Physics Laboratory)

Abstract:

8:15 – 8:40 Gottlieb Oehrlein (University of Maryland) Cold Atmospheric Plasma/Polymer Treatment as Model System for Elucidating Complex Plasma-Surface Interactions (PSI) at Atmospheric Pressure

p. 11

8:40 – 9:05 Peter Bruggeman (University of Minnesota) Water Vapor Plasma Kinetics: Non-equilibrium Kinetics, Radical Production and Transport

p. 12

9:05 – 9:30 David Graves (University of California-Berkeley) Physico-chemical Dynamics of Atmospheric Pressure Plasma-liquid Interactions

p. 13

9:30– 9:55 John E. Foster (University of Michigan) Progress Towards Understanding Physical Processes Prevailing at the Plasma Liquid Interface

p. 14

9:55 - 11:00 am

Coffee Break Poster Session I

4

Thursday, June 1, 2017 11:00 am – 12:40 pm

Session II. Particles, Aerosols, Specialty Sources Moderator: Emi Kawamura (University of California-Berkeley)

Abstract:

11:00 – 11:25

Steven Girshick (University of Minnesota) Nanoparticles in Low-Temperature Plasmas Are Not All Negatively Charged

p. 15

11:25 – 11:50

Mark J. Kushner (University of Michigan) Interaction Between Atmospheric Pressure Plasmas and Liquid Micro-droplets

p. 16

11:50 – 12:15

Yevgeny Raitses (Princeton Plasma Physics Laboratory) Generation of Energetic Electrons in a Microplasma at Moderate Pressures

p. 17

12:15 – 12:40

Brandon Smith (University of Michigan) Development of a GPU-Accelerated Poisson Solver for Hybrid Fluid/Kinetic Plasma Simulations

p. 18

12:40 – 1:40 pm

Lunch (on your own)

5

Thursday, June 1, 2017 1:40 – 3:45 pm

Session III. Kinetics and Diagnostics Moderator: Yangyang Fu (Michigan State University)

Abstract:

1:40 – 2:05 Eray Aydil (University of Minnesota) Plasma Diagnostics and Modeling of Lithium-containing Plasmas

p. 19

2:05 – 2:30 Igor Adamovich (Ohio State University) Electric Field Measurements in Ns Pulse Discharges in Atmospheric Pressure Air

p. 20

2:30 – 2:55 Vladimir Kolobov (CFDRC/University of Alabama at Huntsville) Dynamic Discharges

p. 21

2:55 – 3:20 Igor Kaganovich (Princeton Plasma Physics Laboratory) Kinetic Modeling of Non-Equilibrium High Pressure Plasmas for Modern Applications

p. 22

3:20 – 3:45 Valery Godyak (University of Michigan) Plasma Density Perturbation by Microwave Probes

p. 23

3:45 - 5:00 pm

Coffee Break Poster Session II

5:15 – 5:30 pm

Group Photo

6

Friday, June 2, 2017 8:00 – 9:40 am

Session IV. Atmospheric Pressure Plasmas Moderator: Marien Simeni Simeni (Ohio State University)

Abstract:

8:00 – 8:25 Vincent Donnelly (University of Houston) VUV to near IR Optical Emission Excitation Mechanisms in Atmospheric Pressure He Discharges into Open Air

p. 24

8:25 – 8:50 Michael Lieberman (University of California-Berkeley) Kinetic Instability in Water-Containing Atmospheric Pressure Discharges

p. 25

8:50 – 9:15 Ed Barnat (Sandia National Labs) Interrogating High Pressure, Highly Collisional Plasma Environments With Ultrafast Laser Diagnostics

p. 26

9:15 – 9:40 John Verboncoeur (Michigan State University) Dynamics in Strongly Driven High Pressure Reactive Plasmas

p. 27

9:40 - 10:00 am

Coffee break

7

Friday, June 2, 2017 10:00 – 11:15 am

Session V. Students, Post-docs, Visitors Session Moderator: Daniel Elg (University of California-Berkeley)

Abstract:

10:00 – 10:15 Emi Kawamura (University of California-Berkeley) Particle-in-cell Simulations of Water-Containing Atmospheric Pressure Plasmas

p. 28

10:15 – 10:30 Andrew Fierro (Sandia National Laboratories) Developing a Kinetic Approach to Radiation Transport and Its Interaction in He/N2 Ionization Waves

p. 29

10:30 – 10:45 Janis Lai (University of Michigan) Argon in 2-D Bubble Test Cell for Studying Active Species Transport Across Plasma-liquid Interface

p. 30

10:45 – 11:00 Yuchen Luo (University of Minnesota ) Validation of the Kinetics Mechanism of High Electron Density Argon-Water Plasma Discharge

p. 31

11:00 – 11:15 Heng Guo (Princeton Plasma Physics Laboratory) Electron Kinetics in Afterglow Plasma

p. 32

11:15 am – noon

Group Discussion Moderator: Mark J. Kushner

8

Poster Session I Abstract:

1 Xiang Chen (PPPL) An Investigation into Non-adiabatic Behavior of Magnetic Moment in a Mirror Traps

p. 33

2 Kenneth Engeling (University of Michigan) Time-Evolution of Microdischarges in Packed Bed Reactors with Varying Media

p. 34

3 Juliusz Kruszelnicki (University of Michigan) Impact of Catalysts on Discharge Evolution and Incident Surface Fluxes in a 2-dimensional Packed Bed Reactor

p. 35

4 Santosh Kumar Kondeti (University of Minnesota) Absolute OH Density Measurements in an RF Driven Atmospheric Pressure Plasma Jet with a Substrate Below by Laser Induced Fluorescence

p. 36

5 Yangyang Fu (Michigan State University) Investigation on the Effect of Forbidden Processes on Similarity Law in Gas Discharges at High Pressure Based on a Kinetic Global Model

p. 37

6 Heng Guo (PPPL) Electron Kinetics in Afterglow Plasma

p. 32

7 Yuchen Luo (University of Minnesota) Validation of the Kinetics Mechanism of High Electron Density Argon-Water Plasma Discharge

p. 31

8 Alexander Khrabrov (PPPL) The Pashen Curve at High Voltage and Low Pressure

p. 38

9 Gaurav Nayak (University of Minnesota) Gas and Liquid Phase Plasma-Bio Interactions: Role of Reactive Nitrogen Species

p. 39

10 Tam Nguyen (University of Houston) Optical Emission Diagnostics of a Non-equilibrium Helium Plasma Jet at 1 atm in Ambient Air

p. 40

11 Andrew Powis (PPPL) Implementation of a Multigrid Poisson Solver For Massively Parallel Two and Three-Dimensional PIC Simulations of Low-Temperature Plasma Devices

p. 41

12 Marien Simeni Simeni (Ohio State University) Measurements of Electric Field in Ns Pulse Discharges in Atmospheric Pressure Air by ps 4-Wave Mixing

p. 42

9

Poster Session II Abstract

1 Vladislav Vekselman (PPPL) Experimental and Theoretical Study of the Carbon Arc: Identification of Plasma Properties in the Region of Nanotube Synthesis

p. 43

2 Daniel Elg (UC-Berkeley ) TEMPO Production by O Atoms in Plasma-Liquid Interactions Driven by Spatio-Temporally Varying Atmospheric Pressure Plasma Jet

p. 44

3 Janez Krek (Michigan State University) Rapid Modeling of Kinetic Reactive Plasma Dynamic Using the Kinetic Global Model Framework

p. 45

4 Andrew Fierro (Sandia National Laboratories) Developing a Kinetic Approach to Radiation Transport and Its Interaction in He/N2 Ionization Waves

p. 29

5 Janis Lai (University of Michigan) Argon in 2-D Bubble Test Cell for Studying Active Species Transport Across Plasma-liquid Interface

p. 30

6 Steven Lanham (University of Michigan) Instabilities at Startup of Pulsed Electronegative Inductively Coupled Plasmas

p. 46

7 Sophia Gershman (PPPL) Plasma Characterization of Microhollow Anode and Microhollow Cathode Discharges at Moderate Pressures

p. 47

8 Pingshan Luan (University of Maryland) A Case Study of Plasma-Surface Interactions at Atmospheric Pressure: Polystyrene Treatment Using an RF Plasma Jet

p. 48

9 Emi Kawamura (University of California-Berkeley) Particle-in-cell Simulations of Water-Containing Atmospheric Pressure Plasmas

p. 28

10 Andrew Knoll (University of Maryland) On the Variation of the Activation Energy of Polymer Etching by Cold Atmospheric Plasma (CAP) Sources Under Well-Defined Conditions

p. 49

11 Toshisato Ono (University of Minnesota) Plasma Diagnostics in Air Plasmas Containing Water Droplets

p. 50

12 Alexander Khrabry (PPPL) 2D Simulations of the Carbon Arc Discharge for Synthesis of Nanotubes

p. 51

10

Abstracts

Introduction to the Annual Meeting: The Role of the Center in the Future of Low Temperature Plasmas

Mark J. Kushner

University of Michigan, Electrical Engineering and Computer Science, Ann Arbor, MI 48109-2122 (mjkush@umich.edu)

The Department of Energy Center for the Predictive Control of Plasma Kinetics: Multi-

Phase and Bounded Systems was established in the Fall of 2009 in response to recommendations of the National Research Council Plasma 2010 Decadal Study to improve research opportunities in low temperature plasmas (LTPs). Now, nearing the end of Year 8, the Center has made tremendous progress in addressing the fundamental science of low temperature plasmas across an impressive range of pressures, from sub-mTorr to liquid densities, and translating those science advances to societal benefit. Over the past 2 years, two major reports affecting the future of LTP science have been produced by DOE Office of Fusion Energy Science (OFES), Frontiers of Plasma Science, and the National Science Foundation, Enabling a Future Based on Renewable Electricity through Plasma Chemistry. (See Fig. 1.) The recommendations of these reports will be briefly reviewed, and the role of the Center in implementing those visions will be discussed.

Figure 1 – Recently released reports discussing priorities in low temperature plasmas. (left) NSF Non-Equilibrium Chemistry. (right) DOE Plasma Frontiers.

11

Cold Atmospheric Plasma/Polymer Treatment as Model System for Elucidating Complex Plasma-Surface Interactions (PSI) at Atmospheric Pressure

P. Luan(a), A. Knoll(a), V. S. S. K. Kondeti(b), P. J. Bruggeman(b) and G. S. Oehrlein(a)

(a) University of Maryland, College Park (oehrlein@umd.edu) (b) University of Minnesota, Twin Cities (pbruggem@umn.edu)

Cold atmospheric plasma (CAP) sources produce chemically reactive species that effectively

modify material surfaces which enables numerous new applications. While impressive source characterization results have been published, less is known about the interaction mechanisms of these sources with surfaces, e.g. of cells, biomolecules, and polymers. Due to the large quantity and strong spatial/temporal gradients in density of reactive species, the study of material surface interactions with such sources is often convoluted by interactive and synergistic effects. To build correlation between gas phase species and material surface responses, a model system with well-characterized CAP sources, selected model polymers and highly-controlled environments was adopted for obtaining mechanistic insights and understanding the physics of plasma-surface interaction (PSI) at atmospheric pressure.

Etching and surface modification of polymers with an RF jet [1] were studied using in-situ ellipsometry and other surface characterization techniques [2]. The effect of treatment distance, angle, feed and environment gas composition as well as substrate temperature on PSI was studied. Using measured gas phase density data of atomic O and OH radicals [2], we calculated the reaction rate between the flux of incident O atoms or OH radicals and the flux of etched C off polymer surface. For Ar/O2 and Ar/H2O plasma, we found that the polymer etch rate decayed exponentially with treatment distance, whereas the surface modification showed a maximum. A surface reaction model based on the competition of two types of plasma species, i.e. an oxidizing species modifying the surface by Langmuir adsorption and an etchant species that reduces the resulting surface oxidation, can reproduce the observed trends (see figure 1(a)). The resulting surface coverage θ is a function of both etchant species flux and modification species flux: θ increases with the flux of modification species and decreases with the flux of etchant species. The apparent activation energy (Ea) of the etching reaction was estimated by measuring etch rate versus substrate temperature as shown in figure 1(b). The variation of Ea with target material, feed gas composition and treatment distance points to additional, locally varying species fluxes, that give rise to the observed behavior.

The authors gratefully acknowledge financial support from US Department of Energy (DE-SC0001939) and the National Science Foundation (PHY-1415353). References [1] W. V. Gaens, P. J. Bruggeman, A. Bogaerts, New Journal of Physics 16, 063054 (2014). [2] P. Luan, A. J. Knoll, H. Wang, V. S. S. K. Kondeti, P. J. Bruggeman, and G. S. Oehrlein, J. Phys. D. Appl. Phys. 50, 03LT02 (2017); V. S. S. K. Kondeti, P. Luan, Y. Luo, G. S. Oehrlein and P. J. Bruggeman (this meeting).

Figure 1 – (a) Effect etchant species (atomic O) and modification species (Γmod) fluxes on PS surface oxidation (measured by XPS). (b) Apparent activation energy changes with treatment distance and feed gas composition.

12

Water Vapor Plasma Kinetics: Non-equilibrium Kinetics, Radical Production and Transport

Y. Luo(a), V.S.S.K. Kondeti(a), Y. Du(a), A.M. Lietz(b), S. Yatom(a), P. Luan(c), G. Oehrlein(c), M.J. Kushner(b) and P.J. Bruggeman(a)

(a) Department of Mechanical Engineering, University of Minnesota, 111 Church Street SE, Minneapolis, MN 55455, USA (pbruggem@umn.edu)

(b) Department of Electrical Engineering and Computer Science, University of Michigan, 1301 Beal Avenue, Ann Arbor, MI 48109, USA (mjkush@umich.edu)

(c) Department of Materials Science and Engineering, University of Maryland, Energy Research Facility, 8279 Paint Branch Drive, College Park, MD 20742 (oehrlein@umd.edu )

Gas phase non-equilibrium plasmas containing water vapor are of growing interest for many applications. Nonetheless, the plasma kinetics in the presence of water is poorly understood [1].

We performed detailed measurements of the OH, H and electron densities and gas temperature in a spark-like filamentary nanosecond pulsed discharge to validate a water vapor reaction set. The reaction set is implemented in Global-Kin [2]. We found that the model is capable of predicting the general trends in reactive species but has some challenges in predicting the species of lower concentrations in view of many competing reactions. These results show for the first time that we have a reasonable understanding of ionic recombination reactions in water containing plasmas that strongly contribute to radical formation in high electron density plasmas.

In addition, we have measured the OH density in plasma jets impinging on a substrate. As the jet is generated in Ar in a background gas of air or N2, we found that the OH density is strongly impacted by the gas mixing in the effluent. Spatial gradients in the gas composition strongly impact the OH density distribution and can enhance or impede the transport of OH towards a substrate. These findings have important consequences for material processing [3,4].

The above species densities have been measured by laser induced fluorescence, a technique not available in all plasma labs. We developed and validated a new approach to measure absolute OH densities by emission spectroscopy based on the variation of self-absorption of rotational emission lines with different rotational numbers. An example of a spectrum with a fitted modelled spectrum is shown in Figure 1. This new method also allows assessing the non-equilibrium rotational population distribution of the excited state and could in theory be used to measure the electron temperature in plasmas with a low electron density.

References [1] P. Bruggeman and C. Leys, J. Phys. D: Appl. Phys. 42, 053001 (2009). [2] A. M. Lietz and M. J. Kushner, J. Phys. D: Appl. Phys. 49, 425204 (2016). [3] P. Luan, A. J. Knoll, et al., J. Phys. D. Appl. Phys., vol. 50, no. 3, p. 03LT02, 2017. [4] P. Luan, A. J. Knoll, V. S. S. K. Kondeti , et al (this meeting).

Figure 1 – Fitting of a self-absorbed OH(A-X) emission spectrum showing the possibility to determine absolute OH densities by relative emission spectroscopy. The spectrum also shows non-equilibrium effects in the rotational distribution of the excited state.

13

Physico-chemical Dynamics of Atmospheric Pressure Plasma-liquid Interactions

D.B. Graves, Z. Xiong, D.T. Elg and I-W. Yang

University of California at Berkeley (graves@berkeley.edu, daniel.elg@berkeley.edu, zilan@berkeley.edu, happynight13@berkeley.edu)

Non-thermal atmospheric pressure plasmas remain poorly understood due to their often dynamic multi-mode character and other non-linear characteristics. [1,2] Although this class of plasma is currently of interest for a variety of applications, the present study focuses on their fundamental properties. One key set of questions revolve around the relationship between the physics of air plasmas and their interactions with liquid surfaces. One aspect of this interaction is plasma production of reactive oxygen and nitrogen species (RONS) and how these plasma-generated species interact with liquid surfaces. [3] Other aspects include plasma stability, role of external surfaces and electrical properties and voltage-current characteristics. The differences between air plasmas and rare gas jets, both operating at atmospheric pressure and in air environments, are not well understood. For example, the proper treatment of plasma-liquid interactions are generally unknown, in part because the physics of the plasma-liquid interface has been little studied. The liquid may act as a pure dielectric, a pure conductor or something intermediate between the two. The nature of charge transfer between the gas phase plasma amd the liquid is another key unresolved question. [1,2]

We report on a set of studies that aim to better characterize the electrical interactions between kHz powered rare gas jets and pulsed dc air discharges as they interact with adjacent surfaces, both solid and liquid. These discharges can be strongly affected by capacitive coupling to nearby surfaces. The physical dynamics of atmospheric pressure plasmas are complex, and can lead to multiple mode transitions. In addition, we explore the role of nanosecond pulsed power supplies on the plasma physico-chemical properties, in the presence and absence of liquid surfaces. References [1] R. Gopalakrishnan, E. Kawamura, A.J. Lichtenberg, M.A. Lieberman and D. B. Graves, J. Phys. D, 49, 295205 (2016). [2] A. Lindsay et al., J. Phys. D: Appl. Phys. 49 235204, (2016). [3] M. Hefny, C. Pattyn, P. Lukes, and J. Benedikt, J. Phys. D. 49, 404002 (2016).

Figure 1 – Image of point-to-water surface discharge in air. The plasma-water interface and liquid electrical properties can control key aspects of the plasma.

14

Progress Towards Understanding Physical Processes Prevailing at the Plasma Liquid Interface

John E. Foster(a), Janis Lai 2(a), Yao Kovach(a), and Maria C. Garcia(b)

(a) University of Michigan (jefoster@umich.edu) (b) Universidad de Cordoba (fa1gamam@uco.es)

 

The interaction of plasma with liquid water occurs at a phase boundary layer, which includes gas, a water vapor layer, and the liquid itself. A host of physical and chemical processes are active at this interface making it a rich multiphase physics problem. These processes ultimately give rise to changes in the bulk liquid. Such induced changes are the basis for a number of emerging technologies and applications such as plasma-based water treatment and plasma medicine. The nature of the physical processes and ensuing chemistry that “activates” the liquid water, which is believed to originate at the interface, is not well understood. Ongoing experimental and computational efforts however are making progress towards the formulation of a consistent picture of the role of the plasma liquid interface in driving chemistry in solution. Here we survey the current state of understanding regarding the interfacial region including electrohydraulic forces that can lead to fluid dynamical effects resulting in enhanced radical distribution as well as chemistry driven by direct plasma interaction with liquid water. In particular, we review recent results from single trapped bubble and 2-D bubble studies that have yielded insight into mechanisms of radical transport into solution as a function of discharge type present in the bubble.[1] Complex mass transport and induced chemistry generated in DC atmospheric pressure glows with liquid electrode resulting from plasma self-organization as depicted in Figure 1 is also not well understood. Insight into the physical mechanism underlying both self-organization and its role in radical transport in these systems as inferred from recent experiments is also discussed. The implications of these findings, the understanding gaps along with measurement and modeling needs for continued progress, and the connection of this understanding with technologies with a plasma liquid underpinning are also commented upon.

References [1] J. Foster and J. Lai, IEEE Trans. Plasma Science, 44, 1127 (2016).

Figure 1 – Self-organization pattern on liquid anode .

15

Nanoparticles in Low-Temperature Plasmas Are Not All Negatively Charged

Meenakshi Mamunuru(a), Roman Le Picard(b), Yukinori Sakiyama(a) and Steven L. Girshick(c)

(a) Lam Research Corporation, Tualatin, OR (Meenakshi.Mamunuru@lamresearch.com, Yukinori.Sakiyama@lamresearch.com) (b) Lam Research Corporation, Fremont, CA (Romain.LePicard@lamresearch.com)

(c) Department of Mechanical Engineering, University of Minnesota, Minneapolis, MN (slg@umn.edu)

Although particles in nonthermal dusty plasmas tend to charge negatively, a population of non-negative particles may still exist. These particles are not electrostatically trapped in the plasma, and thus can diffuse to walls. Additionally, they can facilitate coagulation. In a collaboration between the University of Minnesota and Lam Research Corporation, we conducted Monte Carlo charging simulations to explore the effects of several parameters on the fraction of particles that are not negatively charged [1]. These simulations accounted for two effects not considered by the orbital motion limited theory of particle charging [2]: single-particle charge limits, which were implemented by calculating electron tunneling currents from particles [3-6]; and the increase in ion currents to particles caused by charge-exchange collisions that occur within a particle’s capture radius [6-7]. The effects of several parameters were considered, including particle size, in the range 1 to 10 nm; pressure, ranging from 0.1 to 10 Torr; electron temperature, from 1 to 5 eV; positive ion temperature, from 300 to 700 K; plasma electronegativity, characterized in terms of n / ne ranging from 1 to 1000; and particle material, either SiO2 or Si. Within this parameter space, higher non-negative particle fractions are associated with smaller particle size, higher pressure, lower electron temperature, lower positive ion temperature, and higher electronegativity. Additionally, materials with lower electron affinities, such as SiO2 (1.0 eV), have higher non-negative particle fractions than materials with lower electron affinities, such as Si (4.05 eV). Fig. 1 shows that pressure has a strong effect on the non-negative particle fraction, because increasing pressure causes incoming ions to be more likely to experience a charge-exchange collision within a particle’s capture radius.

References [1] M. Mamunuru, R. Le Picard, Y. Sakiyama and S. L. Girshick, Plasma Chem. Plasma Process. 37, 701 (2017). [2] J. E. Allen, Phys. Scripta 45, 497 (1992). [3] A. Gallagher, Phys. Rev. E 62, 2690 (2000). [4] R. Le Picard and S. L. Girshick, J. Phys. D 49, 095201 (2016). [5] L. C. J. Heijmans, F. M. J. H. Wetering and S. Nijdam, J. Phys. D 49, 388001 (2016). [6] R. Le Picard and S. L. Girshick, J. Phys. D 49, 388002 (2016). [7] S. A. Khrapak et al., Phys. Rev. E 72, 10 (2005). [8] M. Gatti and U. Kortshagen, Phys. Rev. E 78, 046402 (2008).

Figure 1 – Effect of pressure and particle diameter Dp on fraction of SiO2 particles that are non-negative. Conditions: Te = 2 eV, Ti = 300 K, n+/ne = 10.

16

Interaction Between Atmospheric Pressure Plasmas and Liquid Micro-droplets Amanda M. Lietz, Juliusz Kruszelnicki and Mark J. Kushner

University of Michigan, Ann Arbor, MI, 48109-21122 USA (mjkush@umich.edu)

Plasma activation of liquids, and water in particular, is typically rate limited by transport processes. The plasma produced species must first transport to and cross the plasma-liquid interface, after which the solvated species themselves react and transport within the liquid. This sequence slows the rate of activation of the liquid and makes selective activation difficult. Liquid droplets immersed in the plasma relax the requirement for transport of plasma species to the interface. The larger surface to volume ratio of the droplets then enables increased transport efficiency of reactive species in the liquid. Unlike macroscopically large plasma liquid interfaces, the high curvature of droplets coupled with a high dielectric constant opens the possibility of electrical droplet-plasma interactions through the polarization of the droplets and the resulting electric field enhancement in the gas phase.

In this paper, results will be discussed from first-principles modelling of water micro-droplets immersed in atmospheric pressure plasmas (APPs). Two modelling platforms were used in this study. Global_Kin addresses two regions, gas and liquid, in a global-kinetics framework.[1] nonPDPSIM, addresses the formation and propagation of ionization waves in APPs, and the 2-phase transport of charged and neutral species.[2] The water-reaction mechanism, described in [1] accounts for ion, photon and neutral activation of the water from the gas phase, the latter limited by Henry’s law equilibrium at the surface of the droplet, and subsequent in-liquid chemistry.

Examples of results from the study are shown in Fig. 1 for a DBD sustained in humid air (N2/O2/H2O = 78/21/1) in a 2 mm gap bounded by a cathode overlaid with a dielectric (ε=4.0) and the ground electrode. A 40 μm diameter water droplet is at the center of the gap. A 100 μm water layer covers the grounded electrode for comparison to the droplet. The droplet is polarized in the applied electric field, resulting in intense enhancement of the electric field at the poles of the droplet. When the DBD streamer intersects the droplet, there is a secondary avalanche at the top of the droplet, which produces a high density of reactants at the surface, demonstrated here by the OH density. The large surface to volume ratio of the droplet enables a large rate of solvation of gas phase species into the water. For species having a low Henry’s law constant, such as O3, the droplet quickly fills up to its equilibrium density, nearly independent of droplet size. For species having a high Henry’s law constant, such as HNOx, the droplet continues to solvate gas phase species for the duration of its residence time in the plasma. This results in small droplets become highly acidic. Some amount of control of droplet activation can be achieved by droplets size and residence time in the plasma. References [1] W. Tian, A. M. Lietz and M. J. Kushner, Plasma Sources Sci. Technol. 25, 055020 (2016). [2] A. M. Lietz and M. J. Kushner, J. Phys. D: Appl. Phys. 49, 425204 (2016).

Figure 1 – Fig. 1: (a) DBD computational do-main and electron density at mid-pulse surrounding a 40 m water droplet. (b) Density of OH near the liquid droplet as the ionization wave passes by

17

Generation of Energetic Electrons in a Microplasma at Moderate Pressures

Yevgeny Raitses and Sophia Gershman

Princeton Plasma Physics Laboratory (yraitses@pppl.gov, sgershma@pppl.gov)

Microplasmas have emerged over the last decade as an important direction for plasma development, providing moderate ( 1 torr) and atmospheric pressure discharges at modest power input, that may someday be exploited in applications for catalysis, reforming processes, remediation, manufacturing etc. [1,2]. Plasma kinetics and plasma-surface interactions are key, but poorly understood, processes in microplasmas, and therefore, our research is aimed to study these processes and explore their control through the use of, for example, the anode sheath phenomenon [3,4].

In recent microplasma experiments, we demonstrated that non-local electron energy distribution in a micro hollow anode configuration (MHA) makes it possible to produce and control the fraction of relatively high energy electrons outside of the discharge region in areas important for plasma applications [4]. Calculations showed that at the operating pressure of 1-10 torr, the energy relaxation distance in MHA is ~0.1 – 1 cm, to the order of or greater than characteristic dimensions of the micro discharge and the mean free path of ~ 10-2 – 10-1 cm. Therefore, MHA is expected to have a non-local electron energy distribution (EEDF) and can be exploited as a source of energetic electrons that could be extracted and/or controlled outside of the discharge region [4]. Compared to a more traditional micro hollow cathode (MHC) configuration, MHA discharge had higher electron energy and at least an order of magnitude lower electron densities (~1010 cm-3) than MHC. Measurements outside the cavity of the MHA using electrostatic probes and Optical Emission Spectroscopy indicated the presence of 12 – 15 eV electrons in N2 gas at 3 torr [5]. A hybrid MHC-MHA configuration (Fig. 1-a) with a biased substrate as a third electrode exploits the promise of better control of EEDF outside of the micro discharge. Recent measurements demonstrated two different plasmas with different electron temperatures and densities produced at each side of the hybrid arrangement (Fig. 1-b). Following Ref. [4], we implemented an additional biased electrode/substrate and showed how it adds a flexibility in controlling plasma-surface interaction by extracting electrons of various energies from the discharge. We will discuss the effects of this electrode on plasma properties in the plume from this hybrid discharge and near the substrate.

References [1] D. Mariotti and M. Sankaran,J. Phys. D: Appl. Phys. 44, 174023 (2011). [2] E. Neyts, Plasma Chem. Plasma Process 36,185 (2016). [3] V. I. Kolobov and A. S. Metel, J. Phys. D: Appl. Phys. 48, 233001 (2015). [4] A. S. Mustafaev, V. I. Demidov, I. Kaganovich, S. F. Adams, M. E. Koepke, and A. Grabovskiy, Rev. Sci. Instrum. 83, 103502 (2012). [5] S. Gershman and Y. Raitses, submitted to Plasma Sources Science and Technology (2017).

Figure 1 – a. PPPL microdischarge set up and 1 – b. The electron temperature above the substrate for two substrate positions.

18

Development of a GPU-Accelerated Poisson Solver for Hybrid Fluid/Kinetic Plasma Simulations

Brandon D. Smith(a), Iain D. Boyd(a) and Vladimir Kolobov(b)

(a) University of Michigan (bradenis@umich.edu; iainboyd@umich.edu) (b) CFD Research Corporation (vik@cfdrc.com)

The propagation of streamers in atmospheric pressure electrical discharges is of interest in plasma remediation of toxic gases, ozone production, and functionalization of surfaces. The air (or other gases) in these non-pristine environments is often contaminated with particles or aerosols having sizes of tenths to hundreds of micrometers. For aerosol diameters in the range of 0.1 to 10 µm, the small length scale of the particle gives rise to a relatively large local Knudsen number (0.01 to 1.0 for atmospheric plasma) despite the small local mean free path, indicating that kinetic phenomena are dominant. This Project aims to develop a hybrid computational methodology that combines the numerical efficiency of a fluid approach with the localized physical accuracy of a kinetic method to address such multi-scale plasma physics problems. Similar hybrid fluid/kinetic methods have been developed successfully by Boyd et al. [1,2] and by Kolobov et al. [3] for nonequilibrium neutral gas flows. However, modeling plasma flows in a similar manner introduces new complications, such as electron kinetics and induced electric/magnetic fields, that require the solution of additional equations and greater computational power.

To address the challenges of modeling these multi-scale plasma flows, we have been developing a hybrid fluid/kinetic model for heterogeneous computing systems consisting of typical multi-core CPUs and high-performance graphics processing units (GPUs). This model is built in the Unified Flow Solver (UFS), a modular, hybrid code for simulating mixed continuum/rarefied gas and plasma flows. [4–6] Figure 1 shows an example UFS solution of an axisymmetric plasma arc forming in argon gas at a pressure of 6 Torr. Although UFS already contains GPU-accelerated kinetic flow solvers (including direct Boltzmann and Direct Simulation Monte Carlo), the solution of the Poisson equation remains a significant bottleneck, particularly in regions where the electron kinetics—and the Debye length—must be resolved. Hence, we are developing a library of GPU-accelerated linear algebra solvers to enable complete solution of multi-scale plasma flows on heterogeneous computing systems with UFS. Based on the observed speedup of the GPU-accelerated kinetic solvers [6], we expect the GPU-based Poisson solvers to perform about 10–30 times faster than comparable solvers on a single CPU core.

References [1] Q. Sun, I.D. Boyd, and G.V. Candler, J. Comput. Phys. 194, 256 (2004). [2] T. E. Schwartzentruber, L. C. Scalabrin, and I. D. Boyd, J. Comput. Phys. 225, 1159 (2007). [3] V.I. Kolobov, R.R. Arslanbekov, V.V. Aristov, A.A. Frolova, and S.A. Zabelok, J. Comput. Phys. 223, 589 (2007). [4] V.I. Kolobov and R.R. Arslanbekov, J. Comput. Phys. 231, 839 (2012). [5] V.I. Kolobov, R.R. Arslanbekov, V.V. Aristov, A.A. Frolova, S.A. Zabelok, J. Comput. Phys. 223, 2 (2007). [6] S. Zabelok, R. Arslanbekov, V. Kolobov, J. Comput. Phys. 303, 15 (2015).

Figure 1 – UFS solution of a plasma arc forming between a charged plate (410 V potential) and a grounded rod in argon gas at a pressure of 6 Torr. Plotted contours are taken at t=7.1 µs.

19

Plasma Diagnostics and Modeling of Lithium-containing Plasmas

Toshisato Ono(a), Eray Aydil(b) and Uwe Kortshagen(a)

(a) Dept. of Mechanical Engineering, University of Minnesota, Minneapolis (onoxx12@umn.edu, kortshagen@umn.edu)

(b) Dept. of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis (aydil@umn.edu)

Thin film deposition from chemically reactive plasmas offers intriguing basic plasma

physics due to the complex multi-component nature of the plasmas involved. In this study, we focused on argon plasmas containing a lithium (Li) – silicon (Si) precursor [lithium bis (trimethylsilyl) amide (LiHMDS)] used for the plasma enhanced chemical vapor deposition (PECVD) of LixSiy thin films, which are attracting attention as novel materials for lithium-ion batteries [1]. The Li-to-Si ratio is a crucial property of these films.

In this study, LixSiy films were deposited using a capacitively coupled argon plasma maintained in a 2.54 cm diameter 25 cm long cylindrical tube. Two 2 cm 3 cm copper parallel plates serve as electrodes and were powered by a 13.56 MHz radio frequency power supply (Figure 1). The chamber pressure was maintained at ~340 mTorr and typical rf power was 5 W. We studied the axial uniformity of the plasma using several techniques. First, to study the axial variation of the Li-to-Si ratio, films were deposited on 1 cm 10 cm molybdenum strips attached to the inner wall of the tube. Film characterization showed that the Li-to-Si ratio along the plasma was highly varied from from 3:1 (Li:Si) in the upstream of the tube to 1:1 in the downstream. We hypothesized that the axial variation of the Li-to-Si ratio is due to the large differences in the neutral and ionic diffusivities of Li and Si species.

To test this hypothesis, a simple one-dimensional reaction-transport model was developed. Axial excitation and ionization rates were determined based on electron temperature axial profiles measured using optical emission spectroscopy in conjunction with actinometry. The analysis shows that both the electron temperature variation as well as the differences between the Li and Si atoms cause the axial variation of the Li-to-Si ratio. Plasma regions with low electron temperature ionize more Li than Si because Li has a lower ionization potential. The smaller mass of Li also leads to a high diffusion rate of Li to the substrate. Axial depletion of the Li results in lower Li incorporation into films deposited downstream of the tube, yielding a smaller Li:Si ratio.

Reference [1] T. D. Hatchard and J. R. Dahn, Journal of The Electrochemical Society, 151 (6) A838-A842 (2004)

Figure 1 - Sketch of the deposition reactor showing the quartz tube, electrodes powered by an RF power supply, and gas handling lines with a precursor bubbler and exhaust.

20

Electric Field Measurements in Ns Pulse Discharges in Atmospheric Pressure Air

Marien Simeni Simeni(a), Benjamin Goldberg(b), Kraig Frederickson(a) and I.V. Adamovich(a)

(a) Ohio State University (adamovich.1@osu.edu) (b) Princeton University (b.goldberg@princeton.edu )

Temporally and spatially resolved electric field is measured in a nanosecond pulse discharge in atmospheric air, sustained between a razor edge high-voltage electrode and a plane grounded electrode covered by a thin quartz dielectric plate. The electric field is measured by picosecond four-wave mixing in a collinear phase-matching geometry, with time resolution of approximately 2 ns, using an absolute calibration provided by measurements of a known electrostatic electric field. During the discharge pulse, the electric field follows the applied voltage until “forward” breakdown occurs, after which the field in the plasma is significantly reduced due to charge separation. When the applied voltage is reduced, the field in the plasma reverses direction and increases again, until the weak “reverse” breakdown occurs, producing a secondary transient reduction in the electric field. After the pulse, the field is gradually reduced on a microsecond time scale, likely due to residual surface charge neutralization by transport of opposite polarity charges from the plasma. Spatially resolved electric field measurements show that the discharge develops as a surface ionization wave. Significant surface charge accumulation on the dielectric surface is detected near the end of the discharge pulse. Spatially resolved measurements of electric field vector components demonstrate that the vertical electric field in the surface ionization wave peaks ahead of the horizontal electric field. Behind the wave, the vertical field remains low, near the detection limit, while the horizontal field is gradually reduced to near the detection limit at the discharge center plane. Additional electric field measurements have been done in a ns pulse discharge sustained between the sane razor edge electrode and distilled water surface, demonstrating behavior qualitatively similar to the discharge over quartz dielectric, including forward and reverse breakdown, plasma seld-shielding, electric field reversal during the discharge pulse, and surface ionization wave formation. Both peak electric field and the extent of surface ionization propagation over the liquid surface are lower compared to that over the quartz surface. The results demonstrate significant potential of the present technique for electric field measurements in atmospheric pressure discharges in air, potentially with sub-nanosecond time resolution, providing quantitative insight into kinetics and transport of highly transient air plasmas. The present data can also be used to assess predictive capability of kinetic models of these plasmas.

Figure 1 – Time-resolved electric field in the negative polarity pulse discharge over quartz dielectric, ~ 100 μm below high-voltage electrode.

Figure 2 – Time-resolved electric field in the negative polarity pulse discharge over distilled water, ~ 100 μm below high-voltage electrode.

21

Dynamic Discharges

Vladimir Kolobov (a,c),Valery Godyak (b) and Robert Arslanbekov (c)

(a) University of Alabama in Huntsville (vladimir.kolobov@uah.edu) (b) University of Michigan (egodyak@comcast.net) (c) CFD Research Corporation (rra@cfdrc.com)

Fundamental properties of partially ionized plasmas are determined by the large difference of electron mass and the masses of ions and neutrals. Due to this difference, the characteristic time scale for the ion transport a is much larger than the electron energy relaxation time . Since a >> , three regimes of plasma operation can be distinguished with respect to the angular frequency of the field maintaining plasma: quasi-static (a<1), dynamic (-1 > >a

-1), and high-frequency ( >1). In our presentation, we will discuss

specifics of these three regimes. On the ion time scale, the electron kinetics is adiabatic, i.e. the Electron Energy Distribution function (EEDF) depends on instantaneous value/distribution of the electric field. Under non-local conditions, the EEDF is a function of total electron energy (kinetic + potential), whereas the potential energy is time-dependent. In quasi-static regimes, plasma density may change drastically. In dynamic regimes, the value and spatial distribution of plasma density vary only slightly over the field period.

The first example we consider in detail is an Inductively Coupled Plasma (ICP) at low driving frequencies [1]. Fig. 1 (a) shows an ICP with ferrite cores forming a closed magnetic path. Fig. 1 (b,c) illustrate calculated dependencies of electron temperature and plasma density at different driving frequencies in Argon gas at pressure of 2 Torr and coil current of 0.1 A. It is seen that a highly nonlinear behavior is observed at low frequencies. We will discuss peculiarities of ICP operation in the high-frequency, quasi-static and dynamic regimes.

The second example we consider is plasma expansion into vacuum or ambient gas. This expansion also occurs at the ion time scale and is characterized by ion acceleration and adiabatic cooling of electrons. We will show examples of simulations for collisionless and collisional regimes and discuss kinetic and gas dynamic effects [2].

References [1] V.Kolobov and V Godyak, Plasma Sources Sci. Technol., submitted (2017). [2] V.I. Kolobov, R R Arslanbekov, A Anders, “Simulations of Explosive Electron Emission in Cathodic Arcs”, Final Report for DoE Phase I SBIR project, March 2017.

Figure 1 – Schematic of ICP with ferrite cores (a) and calculated time dependencies of electron temperature (b) and plasma density (c) on the axis.

22

Kinetic Modeling of Non-Equilibrium High Pressure Plasmas for Modern Applications

Igor D. Kaganovich, Andrew T. Powis, Alexander V. Khrabrov, Johan Carlsson

Princeton Plasma Physics Laboratory, Princeton, New Jersey 08543 (apowis@pppl.gov)

We have modified and substantially upgraded two multi-dimensional high performance particle-in-cell (PIC) codes (EDIPIC and PPPL-modified LSP) [1]. The new parallel codes can effectively use many thousands of processors and allow us to perform 2-3D simulations of discharges, where kinetic and collective effects are important, see e.g. [2,3,4]. We study several modern applications where multi-dimensional kinetic effects are important and can be addressed by our PIC simulations. One of the applications is micro-discharges, where chemistry can be effectively controlled by presence of large population of accelerated electrons, either due to sheath electric field or high current densities in the narrow aperture [5,6]. The second application is atomic layer etching and/or deposition, where control of etched deposited layers is required on atomic level. For this purpose ion and electron velocity distribution functions have to be controlled to very high precision [7,8]. Another application of interest is magnetized DC discharges, which are used for high-power electrical switches [9] at high-pressure (few Torr and higher), plasma deposition and electric propulsion for lower pressure [10,11,12]. GE developed a plasma switch that can greatly increase performance of the electric grid [9]. However, a lot of underlining physics needs to be further investigated. Increased understanding will help achieving reliable operation of the switch and possibly other novel devices. The plasma inside the switch operates in an unusual mode of a magnetized, cold-cathode discharge plasma. The voltage between the electrodes in this mode is very low by the standards of cold-cathode plasmas (~80V), and sputtering is low [9]. The mode looks like a small bright spot on the cathode surface that moves in a circular motion at high speed. We plan to develop a physical model of this phenomena supported by 2-3D PIC simulations. Initial simulations with PPPL-LSP have been able to simulate magnetron discharge in realistic geometry of magnetic lines as shown in Fig.1.

References

[1] J. Carlsson, et al., Plasma Sources Sci. Technol. 26, 014003 (2016). [2] I.D. Kaganovich and D. Sydorenko, Phys. Plasmas 23, 112116 (2016). [3] D. Sydorenko, I.D. Kaganovich and P. Ventzek, Phys. Plasmas 23, 122119 (2016). [4] D. Sydorenko, I. Kaganovich, Y. Raitses, and A. Smolyakov, Phys. Rev. Lett. 103, 145004 (2009). [5] A. S. Mustafaev, et al., Phys. Plasmas 21, 053508 (2014). [6] R. L. Stenzel, et al., Phys. Plasmas 18, 012104 (2011). [7,8] H. Wang, et al., Plasma Sources Sci. Technol. 26 024001, and 024002 (2017). [9] T. J. Sommerer, “Back to the Future: How a plasma mode might change the electric power grid”, Frontiers of Plasma Science Workshops, U.S. DOE Office of Fusion Energy Sciences (2016). [10] A. Anders, “Self-organization in ExB discharges” ibid. [11] A. Smolyakov, “Turbulence, anomalous transport and structures in low temperature Hall plasmas with ExB drift “ ibid. [12] S. Baalrud and I.D. Kaganovich, “Plasma Theory: Role and Recent Trends” in “2017 Plasma Roadmap” to be published in J. Phys. D: Appl. Phys. (2017).

Figure 1 – Density of electrons created through impact ionization in PPPL-LSP simulation of the magnetron, which is part of high-voltage plasma switch. The magnetic field lines are overlaid. Cathode on the bottom and anode on top [9].

23

Plasma Density Perturbation by Microwave Probes

Valery Godyak(a) and Natalia Sternberg(b)

(a) University of Michigan (egodyak@comcast.net) (b) Clark University (nsternberg@clarku.edu)

A variety of microwave probes (MWP) have been recently proposed as new tools for plasma

diagnostics. [1] Microwave probe diagnostics are based on the resonance response in absorption, or reflection spectrum of some electro-dynamic structure (probe) immersed into plasma. Depending on the probe structure and on the particular resonance mode, the probe resonance frequency, r is some modeled function of the plasma frequency, pe, which corresponds to the local plasma density, r = r(pe, Te ).

Recent analyses of MWP probe models, [2] has shown that they are built with too simplified and unrealistic assumptions about plasma uniformity, sheath capacitance evaluation, cold plasma permittivity, and a Maxwellian EEDF. The assumption of plasma uniformity around the MWP is the most serious and common drawback of all MWP models, since plasma depletion around the probe has the same scale as the area of microwave field localization. This results in an essential underestimation of the plasma density inferred from MWP measurement.

In this presentation, we give an analysis of plasma perturbations by MWPs for arbitrary collisionality. The results were obtained by solving numerically a set of fluid equations for neutral plasma with cold ions, taking into account ion inertia and nonlinear ion friction force with ρ/i between 0 and 100, thus covering practically all ranges of interest of MWP applications. Here, i is the ion mean free path, and ρ=a+s is the position of the plasma boundary near the spherical probe, where a is the probe radius and s is the probe sheath width. In addition, an analytical solution was found for the collisionless case, α=0.

The normalized plasma density distributions around a spherical MWP for different ratios of ρ/i are shown in Fig. 1. Note that with growing gas pressure, p i

1, the plasma depletion caused by the probe is also growing, and it is localized at the distance (1-2) within the plasma boundary. The normalized plasma density, h, at the plasma-sheath interface, (x = 1) changes from h = 0.606 in the collisionless case (<< 1) to hfor The microwave field, E x2 shown in Fig. 1, makes it clear that the plasma-field interaction in MWP is localized in the area of strongly depleted plasma. Thus, existing MWP theories with uniform plasma should result in plasma densities significantly less than in unperturbed plasma.

References [1] D. W. Kim, S. J. You, J. H. Kim, H. Y. Chang and W. Y. Oh, Plasma Sources Sci. Technol. 25, 035026 (2016). [2] V. A. Godyak, Comments on Plasma Diagnostics with Microwave Probes, submitted to Phys. Plasmas (2017).

Figure 1 – Distribution of plasma density and microwave E-field around a spherical MWP.

24

VUV to Near IR Optical Emission Excitation Mechanisms in Atmospheric Pressure He Discharges into Open Air

Vincent M. Donnelly, Tam Nguyen and Demetre J. Economou

University of Houston (vmdonnelly@uh.edu, tamtintx@yahoo.com, economou@uh.edu)

We have investigated mechanisms for production of optical emission from reactive species in He plasmas sustained in a small quartz tube surrounded by a grounded and an rf-powered (200 kHz) electrode. The plasma emerges into open air and impinges on a “substrate” (either a MgF2 window or a quartz prism). Time resolved and time-averaged emission was recorded either (a) along the discharge axis, (b) through the discharge tube perpendicular to the discharge axis and between the electrodes, or (c) near the substrate surface.

Near the surface (40o), He 706 nm emission was only observed in a narrow spike near the peak positive voltage, when the discharge splays onto the substrate located 1.2 cm away from the tube exit. When observed along the discharge axis (0o), emission peaked broadly at the peak positive and negative voltages and dipped to zero intensity only near the voltage zero point crossings, reflecting the time dependent excitation by electron impact both inside and outside the discharge tube (at + voltages), and solely inside the tube (at - voltages).

N2 emission (Fig. 1a) is excited by the plasma exiting the discharge. Along the discharge axis no N2 emission is found at negative voltages (no N2 is in the He feed gas), while at positive voltages, emission rises in a manner similar to He emission, but falls much more abruptly, as the plasma near the surface rapidly extinguishes. This is evident from a comparison of the decay rates observed at 0o and 40o in Fig. 1a, just past the peak positive voltage. Emission from N2

+ along the discharge tube (Fig. 1b – red points) behaves very differently from N2 or He. Emission along the axis, though peaking near the peak positive voltage, is detected throughout the RF cycle. This indicates that emission is excited by species other than electrons, namely He metastables. A shift of the emission peak with respect to the peak excitation at peak negative voltage is observed and can be used to derive a lifetime = 2.5 s from tan = e where e is twice the angular frequency (2200kHz). A similar lifetime of 2.7 s is determined from the degree of modulation. Nearly equal lifetimes are observed for O, H and OH emissions. is ascribed to the destruction rate of Hemetastables. Finally, long-time dependences of emission of OH near the surface, during removal of water layers from the substrate surface exposed to the plasma jet will be presented. It appears thermal desorption dominates other processes such as ion, electron, or photon-stimulated desorption.

0 1000 2000 3000 4000 5000-4

-2

0

2

4

Vol

tag

e (k

V)

Time (ns)

V (kV)

0o

40o

0

500

1000

Inte

nsi

ty

(a)

0 1000 2000 3000 4000 5000-4

-2

0

2

4

Vol

tage

(kV

)

Time (ns)

V (kV)

0o

40o

(b)

0

200

400

600

800

1000

Inte

nsity

Figure 1 - RF Voltage waveform and time-resolved optical emission along the discharge axis (0o) and near the surface (40o). (a) N2 (337 nm) and (b) N2

+ (391 nm).

25

Kinetic Instability in Water-Containing Atmospheric Pressure Discharges

M.A. Lieberman, E. Kawamura and A.J. Lichtenberg

EECS Dept, University of California, Berkeley, CA 94720 (lieber@eecs.berkeley.edu)

Water-containing APP’s have wide energy and biomedical applications and, unusually, heavy water cluster ions such as H13O6

+, play a dominant role in their discharge dynamics. Discharge reproducibility can be disrupted by plasma instability. PIC simulations of a 1 mm gap, atmospheric pressure He/2%H2O capacitive discharge show an ionization instability resulting in standing striations in the bulk plasma [1]. Theory indicates that the striations are due to the unusually high electron-ion recombination rate coefficient Krec of the heavy water clusters. Striations are not seen in discharges such as He/N2 that do not form high mass positive ion clusters with high Krec’s. The theory indicates that instability is induced by non-local electron kinetics, by forming a spatially-varying high energy tail of the electron energy distribution that causes the ionization rate coefficient to decrease with increasing reduced electric field (E/N) in the instability regime. The spectrum of unstable wavelengths is limited at short wavelengths by axial diffusion and at long wavelengths by boundaries or electron locality.

We have extended the gap size in the PIC simulations to 2 and 4 mm and have examined both rf (Fig. 1) and dc-driven (Fig. 2) discharges [2]. Wider gap discharges tend to be more unstable as they can accommodate a wider range of wavelengths. Furthermore, the mixture of excited modes in the wider gaps leads to non-sinusoidal spatial oscillations, as seen in the figures.

The striation theory predicts that Ar/H2O APP’s should be more unstable than He/H2O APP’s, due to the reduced (stabilizing) diffusion rates in the heavier mass noble gas. We recently developed a hybrid model calculation of rf-driven Ar/H2O APP’s [3], which will enable us to choose an appropriate simplified set of species and reactions for a PIC simulation study of striations in this system. The hybrid simulations incorporate 47 species with positive ion clusters up to H21O10

+. An isothermal plug-flow model is used, with the gas temperature calculated self-consistently from the input power. The cluster density distributions are determined, and we find that the higher mass cluster densities decrease rapidly with increasing gas temperature. A cluster dynamics analytic model has been developed and solved to determine the cluster density distributions, which is in good agreement with the hybrid simulation results.

References [1] E. Kawamura, M.A. Lieberman and A.J. Lichtenberg, Plasma Sources Sci. Technol. 25, 054009 (2016). [2] E. Kawamura, M.A. Lieberman and A.J. Lichtenberg, to appear in J. Phys. D: Appl. Phys. (2017). [3] A. Tavant and M.A. Lieberman, J. Phys. D: Appl. Phys. 49, 465201 (2016).

Figure 1- RF 27 MHz, 0.23 A/cm2, He/2%H2O.

Figure 2 - DC 0.23 A/cm2, i=0.15, He/2%H2O.

26

Interrogating High Pressure, Highly Collisional Plasma Environments With Ultrafast Laser Diagnostics

Edward V. Barnat(a), Andrew Fierro(a), Ben Yee(a), Matt Hopkins(a) and Peter Bruggeman(b)

(a) Sandia National Laboratories (evbarna@sandia.gov) (b) University of Minnesota (pbruggem@umn.edu)

Diagnostics play a key role in assessing our

understanding of processes that occur in low-temperature plasmas by benchmarking predictive capabilities as well as through discovering otherwise unexpected behaviors. As the perceived landscape of low-temperature plasma science evolve and challenges become more complex (high densities, shorter lifetimes, more reaction pathways), a broad range of diagnostic capabilities are needed to provide a sufficiently complete picture of the plasma. Therefore, new methods need to be developed and made available to facilitate research efforts of the low-temperature plasma community.

First, continued implementation and demonstration of laser-collision induced fluorescence (LCIF) generated in atmospheric pressure helium environments is presented [1,2]. As collision times are observed to be fast (~ 10 ns), ultrashort pulse laser excitation (< 100 fs) of the 23S to 33P (388.9 nm) is utilized to initiate the LCIF process. Both neutral induced and electron induced components of the LCIF are observed in helium afterglow plasma as the reduced electric field (E/N) is tuned from < 0.1 Td to over 5 Td. Under the discharge conditions presented in this study (640 Torr He), the lower limit of electron density detection is ~ 1012 e/cm3. Spatial profiles of the 23S helium metastable and electrons are presented as functions of E/N to demonstrate the spatial resolving capabilities of the LCIF method. Recent results were presented in an invited fast-track communication [3]. Examples of the LCIF method applied to point-point and point-plane discharges (Figure 1) are presented to demonstrate the potential use of the diagnostic platform.

Finally, a brief survey of diagnostic methods that demonstrate promise to further understanding in low-temperature plasma science will be discussed. Emphasis will be placed on methods capable of interrogating surface and near surface regions of a plasma-surface interface.

References [1] E. V. Barnat and K. Frederickson, “Two-dimensional mapping of electron densities and temperatures using laser-collisional induced fluorescence”, Plasma Sources Sci. Technol. 19, 055015 (2010). [2] E. V. Barnat and B. R. Weatherford, “Two dimensional laser-collision induced fluorescence in low-pressure-argon discharges”, Plasma Sources Sci. Technol. 24, 055024 (2015). [3] E. V. Barnat and A. Fierro, “Ultrafast laser-collision induced fluorescence in atmospheric pressure plasma”, J. Phys. D: Appl. Phys. 50, 14LT01 (2017).

Figure 1 – Application of the LCIF method to a point to dielectric plane dischage in 640 Torr of Helium.

27

Dynamics in Strongly Driven High Pressure Reactive Plasmas

John Verboncoeur, Guy Parsey, Yangyang Fu, and Janez Krek

Michigan State University (johnv@egr.msu.edu)

Intense, strongly driven high pressure reactive plasmas provide a rich set of plasma physics different from the more traditional low temperature plasma regime. The high densities, large number of species and multifold chemistry pathways result in a daunting modeling task for most kinetic methods.

The Kinetic Global Modeling framework (KGMf) is an open source spatially independent model able to represent complex plasma chemistry among large numbers of species using energy-dependent reaction cross sections convolved with a specified distribution function. [1] The model comprises a set of spatially independent continuity equations representing each species, coupled by cross-terms representing reactions for sources and sinks. The kinetic electron energy distribution function, typically obtained using a particle-in-cell Monte Carlo code for the same parameter regime, is scaled based on solving an electron energy equation. This scaling is analogous to scaling the temperature for a Maxwellian or Druyvestein distribution.

The KGMf has been used to study the early rate of breakdown in high voltage systems, including air breakdown. [2-3] In the present period, the model was benchmarked for an argon plasma, showing excellent agreement using the same assumptions. [4] A stimulated emission model enabled modeling a rare gas laser in pulsed operation, demonstrating the impact of modifying the electron energy distribution on performance. [5] Response of the Ar excited state densities to optical pumping is shown in Fig. 1. The KGMf was also used to study the scaling effects of plasmas across a wide range of parameters, and in particular the validity of the pd scaling law, including with/without forbidden processes at low and high pressures. [6-7]

References [1] S. K. Nam and J. P. Verboncoeur, Appl. Phys. Lett. 92, 231502 (2008). [2] S. K. Nam and J. P. Verboncoeur, Comp. Phys. Comm. 180, 628 (2009). [3] S. K. Nam and J. P. Verboncoeur, Phys. Rev. Lett. 103, 055004 (2009). [4] G. Parsey, Y. Guclu, J. Verboncoeur, and A. Christlieb, 41st IEEE ICOPS, Washington, DC, 2P-37 (2014). [5] G. Parsey, J. Verboncoeur, and A. Christlieb, Bull. Am. Phys. Soc.61, BP10-17 (2016). [6] Y. Fu, H. Luo, X. Zou, and X. Wang, IEEE Trans. Plasma Sci. 42, 1544 (2014). [7] Y. Fu, H. Luo, X. Zou, and X. Wang, Plasma Sources Sci. Technol. 23, 065035 (2014).

Figure 1 – Ar species densities involved in an optically pumped 2% Ar:He system equilibrating from initial conditions. The model is spatially invariant, driven at 2.56 GHz, 132 W/cm2, 1 ns pulses with 10 kHz rep rate. Gas temperature was 500 K.

28

Particle-in-cell Simulations of Water-Containing Atmospheric Pressure Plasmas

E. Kawamura, M.A. Lieberman and A.J. Lichtenberg

EECS Dept, University of California, Berkeley, CA 94720 (kawamura@eecs.berkeley.edu)

Narrow gap atmospheric pressure plasmas (APPs) have wide ranging energy and biomedical applications. Common feedstock gases are helium and argon with a trace molecular gas admixture. Water containing APPs are of particular interest as H2O is present in many applications. Reproducible discharges require stability, but kinetic particle-in-cell (PIC) simulations of narrow gap He/2%H2O APPs showed an ionization instability resulting in standing striations in the bulk plasma [1,2]. The instability is induced by a combination of low ion mobility and high bulk recombination rates ( Krecn0), where Krec is the electron-ion recombination rate and n0 is the bulk electron density. Water-containing APPs tend to form high mass positive ion clusters with high Krec, and are thus more likely to exhibit striations. These ionization instabilities are due to non-local electron kinetics and would not be observed in commonly used APP fluid simulations.

The striation theory [1] also predicts that water-containing APPs with argon rather than helium feed stock gas should be more unstable due to the reduced (stabilizing) ion mobility in the heavier mass argon. A recently developed hybrid global model calculation of Ar/H2O discharges [3] showed that for a 1.5 mm gap Ar/1%H2O APP with input current 500 A/m2 at 27.12 MHz and wall temperature 300 K, the resulting discharge gas temperature is 400 K, and the dominant species are electrons, H9O4+ and H5O3- ions. This enabled us to choose an appropriate simplified set of species and reactions for a PIC simulation study of striations in this system. In figure 1, we show the PIC results for species densities. Compared to the He/2%H2O system [1], striations are observed at much lower discharge power (~7 times lower) and electron density (~5 times lower) in argon compared to helium.

We have also begun PIC simulations of water-containing APPs at a low frequency of 50 kHz. This introduces a new phenomenon of time-varying ion densities, which profoundly affects the striation formation (see figure 2).

References [1] E. Kawamura, M.A. Lieberman and A.J. Lichtenberg, Plasma Sources Sci. Technol. 25, 054009 (2016). [2] E. Kawamura, M.A. Lieberman and A.J. Lichtenberg, J. Phys. D: Appl. Phys. 50, 145204 (2017). [3] A. Tavant and M.A. Lieberman, J. Phys. D: Appl. Phys. 49, 465201 (2016).

Figure 2 - Contour plot of electron density vs. x and =t for a 1 mm gap He/2%H2O discharge at 1 atm and 300 K driven with current J=0.23 A/cm2 at 50 kHz.

Figure 1 - Species densities vs. x for a 1.5 mm gap Ar/1%H2O discharge at 1 atm and 400 K driven with 500 A/m2 at 27.12 MHz.

29

Developing a Kinetic Approach to Radiation Transport and Its Interaction in He/N2 Ionization Waves

Andrew Fierro, Ed Barnat, Chris Moore, and Matthew M. Hopkins

Sandia National Laboratories, Albuquerque NM, 87185 (asfierr@sandia.gov)

Recently, work has been carried out at Sandia National Laboratories in developing a fully kinetic approach to radiation transport in the low-temperature plasma code Aleph [1]. Aleph utilizes particle-in-cell (PIC) in combination with the Direct Simulation Monte Carlo (DSMC) approach to simulate plasma phenomena. Using a discrete approach, photons are treated as simulation particles with a finite wavelength and may interact with the atomic or molecular species (photoionization) as well as simulation boundaries (photoemission). In an initial example, a Townsend discharge is investigated in pure helium gas and operated near the self-sustaining threshold. The self-absorption of resonance transitions is included in the model and is a key piece to properly simulate radiation transport to the electrode surfaces.

The photon model has been previously verified by comparing time-varying excited species densities against an analytic solution for a single transition and successfully demonstrates a cascading effect for excited nitrogen species [2]. Current work is focused on reproducing emission spectra from simplified experimental plasma discharges (Figure 1a) and understanding the important kinetics that lead to non-Boltzmann emission line ratios. Comparison of experimental and simulated data for a one-dimensional helium discharge (p = 30 torr) indicates that incorporating self-absorption is important for correctly reproducing measured line intensity ratios.

Lastly, the photon model is currently being incorporated into three-dimensional simulations to investigate the effect of photoionization in a nitrogen discharge with a 10% mixture of helium gas (Figure 1b). The resonance transitions of helium are capable of direct photoionization of the nitrogen ground state and thus could be a source of new electrons for plasma development.

Although the discrete approach is promising for including photon effects into kinetic simulations because it reduces the number of assumptions normally made, there exists many challenges in such a detailed approach which are only magnified when moving to higher dimensionality. These challenges are open questions and a current focus for both ongoing and future work. This work was supported by the Office of Fusion Energy Science at the U.S. Department of Energy under contracts DE-AC04-94SL85000 and DE-SC0001939.

References [1] A. Fierro, C. Moore, B. Scheiner, B. Yee, M. Hopkins, J. Phys. D: Appl. Phys., 2017. [2] A. Fierro, J. Stephens, S. Beeson, J. Dickens, A. Neuber, Phys. Plasmas, 2016.

Figure 1 – (a) Ongoing comparisons between experimental and synthesized emission spectra. (b) Application of the photon framework to quantifying the effect of photoionization in a nitrogen/helium discharge.

30

Argon in 2-D Bubble Test Cell for Studying Active Species Transport Across Plasma-liquid Interface

Janis C. Lai and John E. Foster

University of Michigan (janislai@umich.edu)

Low temperature atmospheric plasma is currently being investigated for use in water purification and plasma medicine. Given that the treatment targets are often liquid, e.g. water, blood, understanding plasma-liquid interactions is crucial in effective delivery of plasma-derived active species. Though gas-phased plasma is well understood, its transport through the interface between the gas-phased plasma and liquid-phase target liquid is not. Additionally, studies have shown that the interface is active, where chemical reactions initiated by streamers can be sources for in situ active species generation.

A 2-D plasma-in-liquid apparatus was used in previous work to study the plasma-liquid interface region. [1] Using chemical probes such as methyl orange and starch-KI solutions, propagation of plasma-derived reactive species, such as hydrogen peroxide, into bulk liquid was measured. Streamers were shown to initiate flow that aided in mixing and active species propagation, though the streamer-initiated chemical reactions at the interface are not well understood. In this study, argon was used in the 2D apparatus instead of air. Since air is a composite gas containing high percentage of nitrogen and oxygen, various reactive oxygen and nitrogen species (RONS) are generated and competitively reacted with the colorimetric reagents. Using only argon as feed gas limits the species generated in gas-phase plasma, by comparing the generation and propagation of chemical fronts in the bulk liquid under both modes of plasma discharges: streamer-discharge and micro-discharge, the chemical and physical reactions at the interface can be probed.

References [1] J. Foster and J. Lai, IEEE. Trans. Plasma Science. 44, 1127-1136 (2016).

Figure 1 – Chemical front measured in 2D plasma-in-liquid apparatus filled with methyl orange solution with an air bubble.

31

Validation of the Kinetics Mechanism of High Electron Density Argon-Water Plasma Discharge

Y. Luo(a), A. Lietz(b), V.S.S.K. Kondeti(a), S. Yatom(a), M.J. Kushner(b)and P.J. Bruggeman(a)

(a) Department of Mechanical Engineering, University of Minnesota, 111 Church Street SE, Minneapolis, MN 55455, USA (pbruggem@umn.edu)

(b) University of Michigan, 1301 Beal Avenue, Ann Arbor, MI 48109, USA (mjkush@umich.edu)

Although the plasma kinetics of He-H2O has been developed and partially validated for a diffuse atmospheric pressure glow discharge [1], kinetics models of more complicated high electron density filamentary water containing discharges in which reactions with ions become important have not been studied in detail. In this work, we studied the global kinetics model of an Ar+0.26% H2O plasma with OH density measured by laser induced fluorescence (LIF) and the H density measured by two-photon absorption LIF (TaLIF). The dominant production and consumption mechanism of H and OH species have been obtained through the global model study and validated by the absolute values of number densities and temporal profiles of the species.

We implemented a 0-D chemical kinetics model in Global-Kin [2] and a detailed comparison between the model and the experiment. Figure 1 shows the comparison of the simulated and experimentally measured H and OH density as a function of time. While the model predicts the absolute values of the H and OH densities in an accuracy manner (approximately within a factor 2-3), there remain obvious discrepancies between the model and the experiment. The absolute number density and trailing edge of the H density match exceptionally well with the experimental data and the consumption mechanism is due to radical-radical recombination reactions. The production mechanisms of H/OH and the consumption mechanism of OH have been obtained through a modelling study and involve electron dissociation of water and electron-ion recombination reactions. The significantly lower OH density compared with H is the reason of electron induced dissociation of OH during the discharge pulse and enhanced recombination of OH due to the large H and O density.

The rising edge of measured H and OH may be influenced by the spatial resolution of the experiment. In addition, clear two dimensional transport effects have been found for OH radicals on microsecond timescales which have not been included in the model. Effects of air impurities, local depletion of water at the position of the discharge filament, gas heating and transport on the OH kinetics will also be discussed. The model is capable of predicting the general trends in reactive species but seems to have some challenges in predicting the species of lower concentrations (such as OH) in view of many competing reactions.

References [1] C.A.Vasko, D.X. Liu, et al, Plasma Chem. Plasma Process, 34 (5), 1081-1099 (2014). [2] A. M. Lietz and M. J. Kushner, J. Phys. D: Appl. Phys. 49, 425204 (2016).

Figure 1 - Comparison of the modelled H and OH density in a nanosecond pulsed Ar + 0.26% H2O discharge. The measured densities are obtained by TaLIF and LIF respectively.

32

Electron Kinetics in Afterglow Plasma

Alexander V. Khrabrov(a), Heng Guo (a, b) and Igor D. Kaganovich(a)

(a) Princeton Plasma Physics Laboratory, Princeton University, Princeton, New Jersey 08543, USA (hguo@pppl.gov)

(b) Department of Engineering Physics, Tsinghua University, Beijing 100086, P. R. China

In low-temperature and low-pressure discharge plasmas, electron energy relaxation length can exceed the size of the plasma and the electron velocity distribution function (EVDF) is non-Maxwellian, e.g. with a depleted tail [1]. Such distributions also possess non-local dependence on the self-consistent electric field. Under these conditions, kinetic treatment is required to study collisions and transport phenomena. Predictions based on Maxwellian EDVF, e.g. fluid models, disagree with experimental results [2]. A PIC simulation of a collisionless plasma decaying between two absorbing walls shows that the mean kinetic energy for the direction normal to the wall decreases, whereas the kinetic energy in for velocity components parallel to the wall remains nearly constant during plasma decay, resulting in anisotropy. The EVDF also develops a loss cone due to the escape of those electrons with

||| (0) |w e f f> - . The sheath potential

eventually becomes small compared to the initial electron temperature [3]. In afterglow plasma, where atoms or molecules in long-lived metastable states are present, the Penning ionization process creates suprathermal electrons. The sheath potential may increase or decrease depending on the flux of these electrons compared with the ion flux [4]. In this work, we further advance the model of [3] by considering the formation and decay of metastable atoms which produce fast electrons, with a goal to present a detailed description of the development of the anisotropy in EVDF (illustrated in Fig. 1) for afterglow plasmas.

References: [1] L.D. Tsendin, Plasma Sources Sci. Technol. 4, 200 (1995). [2] Y. Raitses, et al., Phys. Plasmas 13, 014502 (2006). [3] A. V. Khrabrov, and I. D. Kaganovich, The 3rd Annual Meeting of Plasma Science Center for Control of Plasma Kinetics (2012). [4] V. I. Demidov, C. A. Dejoseph, Jr. and A. A. Kudravtsev, Phys. Rev. Lett. 95, 215002 (2005).

Figure 1- EVDF as a function of energy corresponding to velocity components parallel (x) and perpendicular (y) to the sheath electric field between two planar walls [3]. The energies are normalized by the temperature of the initial Maxwellian distribution.

33

An Investigation into Non-adiabatic Behavior of Magnetic Moment in a Mirror Trap

Xiang Chen(a,b), Alexander Khrabrov(b), and Igor Kaganovich(b)

(a) Ghent University (Xiang.Chen@Ugent.be) (b) Princeton Plasma Physics Laboratory (ikaganov@pppl.gov)

For electrons in magnetized plasmas it is widely assumed that the first adiabatic invariant, defined to the lowest order as the magnetic momentμ v /2B, is conserved with high accuracy (except in the vicinity of lines or surfaces where B=0). However, in many real cases the adiabatic conditions are not fulfilled. For v large enough or B small enough, the Larmor radius rmv /qB may become comparable to the field-line radius of curvature, and as a result, μ cannot be regarded quasi-invariant. A good amount of analytical work, e.g. [1]-[4], was performed to reveal the universality as well as the nonlinear and chaotic nature of such non-adiabatic behavior. Numerical simulations further demonstrated the significance of non-adiabatic motion.

This project aims to reveal the relation between loss rate in a mirror trap and non-adiabatic behavior. Due to the non-adiabatic evolution ofμ, the loss rate of charged particles is enhanced compared to the case where loss is only caused by collisions. The non-conservation of μ provides a collision-less mechanism to scatter particles into the loss cone in the velocity space.

Figure 1 serves to illustrate the loss mechanism and Figure 2 presents a simulation result of an electron moving in a mirror trap under certain initial conditions, to verify the possibility of it entering the loss cone and escaping due non conservation of adiabatic momentum.

Figure 1 -- The charged particles stay on the energy-conserved surface in phase space undergoing elastic collisions with neutral atoms until they fall into loss cone after that they are lost to the wall.

Figure 2 –Variation of magnetic moment in the PPPL mirror device [5], E 2keV,r a, φ 0, θ 8° θ 7°. The electron is confined at first and after multiple bounces it escapes the mirror trap at around 900Tmax. The notation refers to the parameters of a MNX mirror device.

References [1] R. J. Hastie, J. B. Taylor, and F. A. Haas, Annals of Physics 41, 302 (1967). [2] J. B. Taylor, Phys. Fluids 7, 767 (1964). [3] V. M . Balebanov, and N. N. Semashko, Nucl. Fusion 7, 207 (1967). [4] D. C. Delcourt, R. F. Martin, and F. Alem, Geophys. Res. Lett. 21, 1543 (1994). [5] http://mnx.pppl.gov/pages/mnx_facility.html

34

Time-Evolution of Microdischarges in Packed Bed Reactors with Varying Media

Kenneth W. Engeling, John E. Foster, Juliusz Kruszelnicki and Mark J. Kushner

University of Michigan, Ann Arbor, MI, 48109-21122 USA (kenengel@umich.edu)

Formation and propagation of plasmas in packed bed reactors (PBRs) is not well

understood. Recent interest into application of these technologies has led to a closer investigation of time-evolution of plasmas in PBRs. Understanding transient effects could allow for optimization in operating conditions and increasing energy efficiency, selectivity and throughput. By manipulating the dielectric constant of the media, localized field enhancement effects change the plasma discharge mechanisms. A 2-dimensional setup has been designed to mimic and simulate the geometries of a full PBR. The simulated PBR consists of a pin-to-planar electrode configuration in which the dielectric rods are setup in a hexagonal configuration as seen in Fig. 1 (c). The investigated media were quartz (ε/ε0=3.8) and zirconia (ε/ε0=26.6). Dielectric polarization leads to increases in electric fields between then rods. Plasma then forms in these areas, first as a filamentary microdischarge and then transitioning to a surface ionization wave. Time-resolved imaging has been performed using simultaneously-triggered ICCD camera system and power supplies. Discharges due to both: an AC and a pulsed DC power source was investigated. Microscopic and macroscopic phenomena were studied using lenses with different foci. The differing lenses provided small scale, local plasma characteristics between dielectrics (Fig 1(a-b)) and the other allowing visuals of the whole media (Fig. 1(c-d)). In Fig. 1(a) shows a microfilamentary discharge moving from rod 5 to rod 4 and in (b), at the surface of rod 4, one can notice the slight build-up of plasma formation that is transitioning into a surface ionization wave. Fig.1 (c-d) shows the wide lens image of the setup and dielectric rods 1-7 with (d) showing the start of plasma formation as it favors moving towards the dielectric material. The understanding of the differences between the DC, pulsed supply and the AC driven supply grant insight in order to further packed bed reactor technologies. The DC discharge had locally shown a microdischarge transitioning into a surface ionization wave that was repeated on a large scale. The difference in dielectric constant had however caused a more wave like plasma formation with the quartz than the more intense, localized discharge of the zirconia. The AC discharge is expected to give more intense discharge as well as effect the overall wave-like formation as seen in the quartz due to the nature of the power supply discharge.

Figure 1 – Plasma properties in a PBR with DC, nanopulsed supply. (a) 5 ns exposure image with microfilamentary discharge (MD) formation, (b) exposure over next 5 ns resulting in buildup of surface plasma, (c) Initial setup of wide-lens with DC, ns pulsed discharge (d) plasma begins forming, starting at the anode.

35

Impact of Catalysts on Discharge Evolution and Incident Surface Fluxes in a 2-dimensional Packed Bed Reactor

Juliusz Kruszelnicki, Kenneth W Engeling, John E. Foster and Mark J. Kushner

University of Michigan, Ann Arbor, MI, 48109-21122 USA (jkrusze@umich.edu)

Plasma-based remediation of pollutants from gases has recently garnered increased attention.[1] The presence of highly energetic and reactive species within the bulk plasma can be utilized to assist in CO2 splitting, methane production, and breakdown of acids. These effects can be optimized through the judiscious combination of plasmas and catalysts. Metallic catalysts can be imbedded on the surface of dielectric materials and provide additional reaction pathways. A plasma packed bed reactor is an example of such a system. Plasma forms between dielectric beads (rods in 2D), allowing for gas re-processing. If catalytic particles are additionally imbedded on the surface of the dielectric, synergetic increases in reaction rates may occur.[2] To gain insight into this processes, electric discharges between dielectric rods with imbedded metallic particles were simulated using the 2D plasma hydrodynamics code, nonPDPSIM.[3] We report on discharge evolution and reactive fluxe to surfaces while varying presence and location of the catalysts,.

The first geometry investigated is full-scale and includes seven dielectric rods between two coplanar electrodes, 1 cm apart [Fig. 1(a)] which allows for the investigation of macroscopic discharge evolution through a packed bed lattice. A second geometry is scaled-down [Fig. 1 (b)] to focus on detailed examination of surface fluxes. In both cases, humid air (N2/O2/H2O = 78/21/1) was the fill gas, and a negative polarity, nanosecond pulse was used.

We found that the presence of metallic particles can significantly alter the structure of the discharges. Micro-discharges are more likely to form near catalysts, leading to high fluxes of charged, excited and reactive species to surfaces. Heavy species with highest fluxes were oxygen and nitrogen cations. These increased fluxes can, in turn, lead to high rates of heating of the metal, and therefore an increase in its reactivity. Catalysts were found to impede the formation of surface ionization waves (SIWs). SIWs are highly energetic transient phenomena and their distortion may lead to significant alteration of the PBR’s selectivity and energy efficiency.

References

[1] H. H. Kim et al. Plasma Chem Plasma Process. 36, 1, 45-72 (2016). [2] K. van Laer and A. Bogaerts, Plasma Sources Sci. Technol. 25, 015002 (2016). [3] S. A. Norberg, E. Johnsen and M. J. Kushner, Plasma Sources Sci. Technol. 24, 035002 (2015).

Figure 1 – Plasma properties in a PBR with catalysts. (a) Electron density in full-scale lattice at t=17 ns, (b) Electron density in scaled-down geometry interacting with catalysts, (c) resulting collapsed local electric field.

36

Absolute OH Density Measurements in an RF Driven Atmospheric Pressure Plasma Jet with a Substrate Below by Laser Induced Fluorescence

V. S. S. K. Kondeti(a), P. Luan(b), Y. Luo(a), G. S. Oehrlein(b) and P. J. Bruggeman(a)

(a) Department of Mechanical Engineering, University of Minnesota, 111 Church Street SE, Minneapolis, MN 55455 (pbruggem@umn.edu)

(b) Department of Materials Science and Engineering, University of Maryland, Energy Research Facility (Bldg. #223), 8279 Paint Branch Drive, College Park, MD 20742 (oehrlein@umd.edu)

An atmospheric pressure plasma jet (APPJ) was recently used to study the etching of polymethyl methacrylate and polystyrene [1,2]. This APPJ induced fast polymer etching while having mild to minimal chemical modification of the polymer. In the presence of oxygen, the variation in the etching rate of the polymers closely resembled the trends of the measured atomic oxygen density in the plasma effluent. When the composition of O2: H2O ratio was varied, a reduction in the etching rate was found for intermediate concentrations that can be attributed to the mutual quenching of OH and O [2]. However in pure H2O, very similar to pure O2, a significant etching occurred as well.

To understand this interesting etching behavior and link it with the underpinning plasma kinetics, we performed detailed 3-dimensional measurement of absolute OH density in the RF APPJ in N2, O2 and air environments by laser induced fluorescence (LIF) and assess the effect of gas composition on the OH production.

The APPJ is modulated at 20 kHz. The plasma is on for 10 µs and off for 40 µs. The OH density (Figure 1) in the core of the plasma builds up till 10 µs and falls near the end of the full cycle of 50 µs. The relative OH density is highest close to the jet nozzle and reduces with increasing distance from the nozzle. The variation of the OH density as a function of time close to the substrate is small which enables the estimation of the OH flux that reaches the substrate by measuring the spatial OH distribution at one time point.

The gas temperature reached a maximum of 430 K at the plasma jet nozzle and 360 K close to the substrate location, thus making it suitable for material processing. The OH density close to the plasma jet nozzle was found to be of the order of 1020 m-3. It is shown that the OH density distribution is strongly impacted by the gas composition. This is not only due to a change in plasma kinetics but also a change in discharge morphology.

We will present detailed OH density distributions in the region between the APPJ and the substrate. The flux to the substrate will be estimated from this data and correlated to the etching depth to deduce OH etching rates.

References [1] A. J. Knoll, P. Luan, et al., Plasma Process. Polym., vol. 13, no. 11, pp. 1067–1077, 2016. [2] P. Luan, A. J. Knoll, et al., J. Phys. D. Appl. Phys., vol. 50, no. 3, p. 03LT02, 2017. [3] T. Verreycken, R. Mensink, et al., Plasma Sources Sci. Technol., vol. 22, p. 1, 2013.

Figure 1 – Relative OH density at the center of the plasma jet at 0.5 mm and 2.7 mm from substrate. Distance between plasma jet nozzle and substrate is 4 mm. Inset shows the two dimensional OH profile at the start of an RF pulse.

37

Investigation on the Effect of Forbidden Processes on Similarity Law in Gas Discharges at High Pressure Based on a Kinetic Global Model

Yangyang Fu, Guy Parsey, John P. Verboncoeur, and Andrew J. Christlieb

Michigan State University (fuyangya, parseygu, johnv, andrewch@msu.edu)

The similarity law is investigated in the presence of so-called forbidden processes [1], such as three-body collisions and stepwise ionizations. Based on recent interest in microdischarges, such as the lab-on-a-chip concept, the applicability of similarity law to sub-millimeter and even nano-scale discharges [2] is important. In our previous studies, the validity of the similarity law was confirmed in discharges at low pressure [3, 4]. Recently, the effect of forbidden processes on the similarity law in discharges at high pressure has been investigated using the kinetic global model framework (KGMf).

In the KGMf model, ground state argon atoms, electrons, positive ions, molecular ions, and fourteen excited levels (4s and 4p) of argon are considered. The geometrically similar volume increases from micron scale (500μm) to millimeter scale (2.5mm) and the pressure decreases from 760Torr to 152Torr, correspondently, keeping the product of gas pressure and linear dimension unchanged. The steady-state electron and ion densities are obtained with the forbidden processes included and excluded in the model, respectively. It is shown that with the forbidden processes included, the normalized density relations are below those predicted from the similarity law. Without the forbidden processes, the parameter relations are in good agreement with the similarity law predictions.

The reason for this phenomenon is that the contributions to the particle balance from allowed and forbidden processes differs by the power on the ratios of the scale-up factor k, which causes cumulative deviations from the similarity relations. This study seeks modifications of similarity relations in discharges at high pressure when the effect from the forbidden processes cannot be generally neglected.

*Work supported by the AFOSR and a DOE Plasma Science Center Grant.

References

[1] G. A. Mesyats, Phys.-Usp. 49, 1045 (2006). [2] D. Janasek, J. Franzke, and A. Manz, Nature 442, 374 (2006). [3] Y. Fu, H. Luo, X. Zou, and X. Wang, IEEE Trans. Plasma Sci. 42, 1544 (2014). [4] Y. Fu, H. Luo, X. Zou, and X. Wang, Plasma Sources Sci. Technol. 23, 065035 (2014).

 

Figure 1 – Fundamental physical processes of argon discharge in the KGMf model.

1 2 3 4 50

5

10

15

20

25

30

Nor

ma

lize

d e

lect

ron

de

nsi

ty

k

Similarity relation KGMf with forbidden processes KGMf without forbidden processes

 Figure 2 – Normalized electron density from the KGMf model including and excluding the forbiden processes.

38

The Pashen Curve at High Voltage and Low Pressure A. V. Khrabrov(a), Liang Xu(a,b), I. D. Kaganovich(a), and T. J. Sommerer(c)

(a) Princeton Plasma Physics Laboratory, Princeton, NJ USA (akhrabro@pppl.gov) (b) CAS Key Laboratory of Geospace Environment, USTC, Hefei, China (hfxuliang@gmail.com)

(c) GE Global Research, Naskayana, NY USA (timothy.sommerer@ge.com)

This work continues our research into ionization breakdown at low pressure with applied voltage in the range of several hundred kiloVolt, i.e. for the left branch of the Paschen curve. Numerical simulation work in this area dates back to 1960s, although backscattering of ions and fast neutrals at the cathode had not been accounted for until recently. [1-3], [4] It is also necessary to correctly account for anisotropic scattering of all species. [5] Previously, we reported results based on a newly developed physical and numerical (particle-in-cell/Monte Carlo collision) model of the breakdown process in helium. An example of a Paschen curve for helium, based on simulations of the discharge, is shown in Fig. 1.

Further progress has been made in refining the breakdown model (in terms of physics as well as the input experimental data) and also in interpreting the results. We discuss the role of ionization by accelerated ions and by fast neutral atoms produced in charge exchange, including the reflected flux of fast neutrals from the cathode. Although the secondary electrons are in the runaway regime and their self-multiplication is small, an energy-relaxed trapped electron population exists in a high-voltage Townsend discharge due to the backscattering of fast electrons at the anode surface. These electrons do contribute to the ionization of the gas, in a process that requires a non-local description.

A reduced “multi-beam” semi-analytical model of discharge in the breakdown state has been developed to illustrate the essential physics.

References

[1] K. D. Granzow and G. W. McClure, Phys. Rev. 125, 1792 (1962). [2] D. Bhasavanich and A. B. Parker, Proc. Roy. Soc. A, 358, 385 (1978). [3] E. J. Lauer, S. S. Yu, and D. M. Cox, Phys. Rev. A 23, 2250 (1981). [4] P. Hartmann and Z. Donko´, Plasma Sources Sci. Technol. 9, 183 (2000). [5] A. V. Khrabrov and I. D. Kaganovich, Phys. Plasmas 19, 093511 (2012).

Figure 1 – The low-pressure branch of the Paschen curve for helium as predicted by PIC simulations

39

Gas and Liquid Phase Plasma-Bio Interactions: Role of Reactive Nitrogen Species

Gaurav Nayak(a), Hamada A. Aboubakr(b), Sagar M. Goyal(b) and Peter J. Bruggeman(a)

(a) University of Minnesota, College of Science and Engineering (nayak025@umn.edu) (b) University of Minnesota, College of Veterinary Medicine

Cold atmospheric pressure plasmas (CAPs) have emerged as one of the most effective tools in surface disinfection and bio-decontamination due to low gas temperature and high reactivity, which involves production of numerous reactive oxygen and nitrogen species (RONS) [1]. Consequently, understanding the detailed underlying plasma-bio-interaction mechanisms becomes challenging.

The focus of this study is to assess the effect of reactive nitrogen species (RNS) on biological substrates in low-temperature air plasmas, by studying plasma inactivation of a virus as a simple model. A flow-through reactor consisting of a 2D array of micro-discharges [2] operated in atmospheric pressure dry air and Ar + 20% O2, producing the same ozone density (3.5×1021 m-3) is used. Using dry air and Ar + 20% O2 eliminates the production of OH and H2O2 in the gas phase, and the difference in the reactive species produced is the presence of RNS in the afterglow. The discharge is generated in micro-cavities which strongly enhances the power density compared to the average power density in a volume DBD. This enhances the RNS production compared to O3 generation. The small gas residence time prevents ozone poisoning leading to the simultaneous production of RONS. The inactivation of the virus is performed in the afterglow of the effluent so the virus is only exposed to long-lived species (O3 and RNS). The virus inactivation is also compared in both gas phase (virus coated on metal surface) as well as liquid phase (virus suspended in solution).

Figure 1 shows the gas-phase treatment in air leading to complete virus inactivation. However, the Ar + 20% O2 plasma leads to incomplete inactivation, suggesting a strong contribution of RNS for virus inactivation.

Liquid-phase treatment of virus suspended in deionized water by dry air and Ar + 20% O2 plasma is also shown in Figure 1. The dry air treatment of virus leads to a significant larger inactivation compared to the Ar + 20% O2 treatment, which corresponds well with significant drop in pH. The measured concentration of NO2

-, which is 5 times larger than the estimated O3 concentration in the deionized water, correlates well with the different inactivation efficacy for 1 and 40 cm exposure distance. The results suggest an inactivation pathway by acidified nitrites (NO2

-) in the liquid phase and most likely NO2-enhanced inactivation in gas phase.

References [1] D. B. Graves, J. Phys. D: Appl. Phys. 45 263001 (2012). [2] G. Nayak, Y. Du, R. Brandenburg and P. J. Bruggeman, Plasma Sources Sci. Technol. 26 35001 (2017).

Figure 1 – Inactivation of virus with dry air and Ar + 20% O2 plasma producing the same amount of ozone. The length scale is the distance between the array and the virus sample. The gas phase and liquid phase treatment is for 3 and 10 minutes, respectively.

Control Dry Air Ar+20%O2

Dry Air Dry Air Ar+20%O2

0

1

2

3

4

5

6

7

1 cm40 cm1 cm13.5 cm

Vir

us T

iter

(lo

g 10 T

CID

50/1

00

L e

luen

t)

13.5 cm

Gas-Phase Liquid-Phase

40

Optical Emission Diagnostics of a Non-equilibrium Helium Plasma Jet at 1 atm in Ambient Air

Tam Nguyen, Demetre J. Economou and Vincent M. Donnelly

University of Houston (tamtintx@yahoo.com, chee1m5@central.uh.edu, vmdonnelly@uh.edu)

Applications of atmospheric pressure

plasma jets involve the jet impinging on the substrate surface, resulting in fluxes of neutral radicals, charged species and photons to the surface. These fluxes are responsible for the chemical reactions and other phenomena that modify the surface. Therefore, knowledge of the identities and relative abundance of these reactive species is important for applications of the plasma jet. We studied a He 200 kHz rf plasma jet emerging into open air from a quartz tube wrapped by a grounded and an rf-powered (typically 200 kHz) electrode. The jet impinged on a dielectric substrate (either a MgF2 window sealed to a vacuum UV spectrometer, or the flat face of a semicircular fused silica prism). Time-resolved and time-averaged emissions were recorded through the substrate either along the discharge axis, or at a steep angle to isolate emission close to the surface. Emissions from He, N2, N2

+, H, O, OH and NO were observed.

Time-resolved emission (Fig. 1) was observed close to the surface only during a brief period near to just past the peak in the positive applied rf voltage. No emission was observed during the negative voltage with the exception of a weak emission from N2(C

3uB3g) just prior to the peak negative voltage. Emissions along the discharge axis from impurities mixing into the He flow just outside the end of the tube are dominated by dissociative excitation via He metastables (He*). Axial emission from N2

+ is also produced by collisions with He* (i.e. Penning ionization of N2). Unlike electron-impact excited emission from N2 and He, emissions produced by collisions of He* with N2, O2 and H2O are only modulated to a small degree during the rf period, and are shifted in phase with respect to the peak positive and negative voltages. He emission along the axis is mainly from inside the discharge, while emission from H, O and OH (not shown) are coming entirely or mostly (for O) from just outside the discharge tube. N2 emission along the axis appears to originate along the entire length of the discharge jet that extends to the surface during positive voltage.

0 1000 2000 3000 4000 5000-4

-2

0

2

4

Vo

ltage

(kV

)

Time (ns)

0

200

400

600

800

1000 HeN

2

N2

+

O

Inte

nsity

a)

0 1000 2000 3000 4000 5000-4

-2

0

2

4

Vol

tage

(kV

)

Time (ns)

0

200

400

600

800

1000

1200b) HeN

2

N2

+

OH

Inte

nsity

Figure 1 - Time-resolved emission in a He atmospheric pressure plasma into open air. a) Emission within <1 mm from the surface, showing a Sharp peak near the most positive applied voltage. b) Emission integrated along the axis of the discharge.

41

Implementation of a Multigrid Poisson Solver for Massively Parallel Two and Three-dimensional PIC Simulations of Low-temperature Plasma Devices

Andrew T. Powis, Johan Carlsson and Igor D. Kaganovich

Princeton Plasma Physics Laboratory, Princeton, New Jersey 08543 (apowis@pppl.gov) Algebraic multigrid is regarded as a state of the art technique for solving large linear problems

representative of elliptical partial differential equations [1]. The technique exhibits efficient scaling on massively parallel systems, order N scaling with grid size, and operates based on the algebraic structure of the matrix rather than the problem geometry. This makes it ideal for dealing with large, complex geometries and boundary conditions.

Our current work includes implementing an algebraic multigrid Poisson solver into the Large Scale Plasma (LSP) particle-in-cell code [2]. The standard LSP suite includes a successive-over-relaxation technique for solving a finite difference approximation to Poisson’s equation. We coupled LSP with the Portable, Extensible Toolkit for Scientific Computation (PETSc) [3]. For smaller simulations, this allowed us to take advantage of direct methods such as PETSc’s native LU decomposition solver. PETSc itself can be coupled with the HYPRE package, which allowed us to utilize the BoomerAMG algebraic multigrid solver [4].

The multigrid solver was benchmarked against the direct methods and was found to be accurate to within the prescribed tolerance. Scaling tests were performed in two and three dimensional cartesian geometry, simulating a box of quasi-neutral thermal plasma with periodic-conducting boundary conditions on 100 million cells. The simulation was found to scale near perfectly from 192 to 4224 cores. In general, it was found that the solver exhibited near perfect scaling when each core received at least ~50,000 cells.

This enhanced code is being applied to study several low-temperature plasma technologies, including anomalous transport in closed drift ExB devices [5], control of the kinetic behaviour in micro-discharges [6] and magnetized discharges for use in a plasma switch [7].

References [1] Stüben, Klaus. "A review of algebraic multigrid." Journal of Computational and Applied Mathematics 128.1 (2001): 281-309. [2] Clark R and Hughes T 2005 User Manual for the Commercial Software LSP (Santa Barbara, CA: Mission Research Corporation) [3] Balay, Satish, et al. "PETSc users manual revision 3.3." Computer Science Division, Argonne National Laboratory, Argonne, IL (2012). [4] Yang, Ulrike Meier. "BoomerAMG: a parallel algebraic multigrid solver and preconditioner." Applied Numerical Mathematics 41.1 (2002): 155-177. [5] Janes, G. S., and R. S. Lowder. "Anomalous Electron Diffusion and Ion Acceleration in a Low‐Density Plasma." The Physics of Fluids 9.6 (1966): 1115-1123. [6] S. Gershman, Y. Raitses, K. Paradinas, B. Koel, “Reactive microplasma discharge for in-situ study of surface modification”, the 69th Gaseous Electronics Conference, October, 2016, Bochum, Germany. [7] Timothy J. Sommerer, “Back to the Future: How a plasma mode might change the electric power grid”, Frontiers of Plasma Science Workshops, U.S. DOE Office of Fusion Energy Sciences.

Figure 1 – Actual vs theoretical scaling of step time against number of cores for LSP with HYPRE BoomerAMG, simulating a 2D box of thermal plasma with 100 million cells and periodic-conducting boundary conditions. Note that the theory curve is scaled to the first data point.

42

Measurements of Electric Field in Ns Pulse Discharges in Atmospheric Pressure Air by ps 4-Wave Mixing

Marien Simeni Simeni(a), Edmond Baratte(ab), Kraig Frederickson(a) and I.V. Adamovich(a)

(a) Ohio State University, adamovich.1@osu.edu (b) Laboratoire EM2C, Ecole Centrale Paris

Picosecond four-wave mixing is used to measure temporally and spatially resolved electric field in a nanosecond pulse dielectric discharge sustained in room air over distilled water and in an atmospheric pressure hydrogen diffusion flame. Measurements of the electric field, and more precisely the reduced electric field (E/N) in the plasma is critical for determination of rate coefficients of electron impact processes in the plasma, as well as for quantifying energy partition in the electric discharge among different molecular energy modes. The four-wave mixing measurements are performed using a collinear phase matching geometry, with nitrogen used as the probe species, at temporal resolution of about 1 ns. Absolute calibration is performed by measurements at known electrostatic electric fields. In the flame experiments, the discharge is sustained between two stainless steel cylindrical electrodes, each placed in a quartz «sleeve», which greatly improves plasma uniformity. Whereas for measurements of electric field over distilled water, the discharge was sustained between a razor edge high-voltage electrode and a plane grounded electrode (layer of distilled water). In this case, the results demonstrate electric field offset on the discharge center plane before the discharge pulse due to surface charge accumulation on the dielectric from the weaker, opposite polarity pre-pulse. During the discharge pulse, the electric field follows the applied voltage until “forward” breakdown occurs, after which the field in the plasma is significantly reduced due to charge separation. When the applied voltage is reduced, the field in the plasma reverses direction and increases again, until the weak “reverse” breakdown occurs, producing a secondary transient reduction in the electric field. After the pulse, the field is gradually reduced on a microsecond time scale, likely due to residual surface charge neutralization by transport of opposite polarity charges from the plasma. Electric field measurements in a flame, which is a high-temperature environment, are more challenging because the four-wave mixing signal is proportional to the square of the difference between the populations of N2

ground vibrational level (v=0) and first excited vibrational level (v=1). At high temperatures, the total number density is reduced, thus reducing absolute vibrational level populations of N2. Also, the signal is reduced further due to a wider distribution of N2 molecules over multiple rotational levels at higher temperatures, while the present four-wave mixing diagnostics is using spectrally narrow output of a ps laser and a high-pressure Raman cell, providing access only to a few N2 rotational levels. Because of this, the four-wave mixing signal in the flame is lower by about two orders of magnitude compared to the signal generated in room temperature air plasma. Preliminary experiments demonstrated four-wave mixing signal generated by the electric field in the flame, following ns pulse discharge breakdown. The electric field in the flame before breakdown is estimated using the analytical expression of electrostatic field between two cylinders surrounded by dielectrics. After breakdown, four-wave mixing signals at known electrostatic fields are employed.

Figure 1 – Time-resolved electric field in the negative polarity pulse discharge over distilled water, ~ 100 μm below high-voltage electrode.

43

Experimental and Theoretical Study of the Carbon Arc: Identification of Plasma Properties in the Region of Nanotube Synthesis

Vladislav Vekselman, Alexander Khrabryi, Igor Kaganovich, and Yevgeny Raitses

Princeton Plasma Physics Laboratory, Princeton NJ 08542 (vekselman@pppl.gov, akhrabry@pppl.gov, ikaganov@pppl.gov, yraitses@pppl.gov)

A carbon arc for nanomaterial synthesis was comprehensively studied using spectroscopic

techniques and electrical measurements and modeled by specially modified computationally fluid dynamic (CFD) code ANSYS. The arc was operated at near atmospheric pressure of helium background gas. Under such conditions, the carbon arc plasma is generated and sustained by ablation of the graphite anode. The carbon feedstock synthesis form carbon nanomaterials at the periphery of the arc. We performed experimental study and CFD modeling to fully characterize plasma and carbon composition in the synthesis region that is important for understanding of synthesis of carbon nanomaterials by the arc method.

We applied a planar laser induced fluorescence (LIF) diagnostic to obtain instantaneous distribution of C2 in carbon arc [1]. In addition, the arc was characterized by optical emission spectroscopy (OES) and fast filtered imaging. Moreover, arc current distribution was measured using segmented electrodes [2]. Measurements of arc current density profile with segmented electrodes revealed i) dimensions of the hot arc core where most of the arc current flows and ii) a cooler arc periphery where synthesis of nanostructures occurs [1,2]. Measurements and simulations show that the main carbon component in the hot arc core is C, while in the cooler synthesis region is C2 (Figure 1), which is a key pre-cursor for synthesis of carbon nanostructures. Therefore, future theories of synthesis need to be reformulated to account for this finding. A hollow profile structure of C2 is preserved even in arc rapid motion over electrode observed at high current, where enhanced anode ablation is observed [3]. Measurements of the voltage drop in the arc confirms hypothesis that the enhanced ablation occur due to transition of the anode sheath from electron-repelling at low arc currents to electron-attractive at high currents [4,5].

This work was supported by US DOE Contract No. DE-AC02-09CH11466.

References [1] V. Vekselman, M. Feurer, T. Huang, B. Stratton and Y. Raitses, submitted to Plasma Sources Scie. and Technol. (2017). [2] Y-W. Yeh, Y. Raitses, and N. Yao, Carbon 105, 490 (2016). [3] S. Gershman and Y. Raitses, J. Phys. D: Appl. Phys. 49, 345201 (2016). [4] A. J. Fetterman, Y. Raitses, and M. Keidar, Carbon 46, 1322 (2008). [5] V. A. Nemchinsky and Y. Raitses, Plasma Sources Scie. and Technol. 25, 035003 (2016).

Figure 1 - Spatial distribution of C2 molecules: (a) number density obtained in simulation, (b) distribution of C2 measured by planar laser induced fluorescence (LIF).

44

TEMPO production by O Atoms in Plasma-Liquid Interactions Driven by Spatio-Temporally Varying Atmospheric Pressure Plasma Jet

Daniel T. Elg, I-Wei Yang and David B. Graves

Department of Chemical & Biomolecular Engineering, University of California – Berkeley (daniel.elg@berkeley.edu)

Non-thermal atmospheric pressure plasmas enable plasma treatment of surfaces without requiring a low-pressure environment. These plasmas are currently of interest for, among other things, their biomedical applications, many of which are enabled by production of reactive oxygen and nitrogen species (RONS). Plasma-liquid interactions are especially important due to the high amounts of water in biological materials. One method to quantify plasma-liquid interactions is to dissolve a reactant into the liquid which, when exposed to plasma-created RONS, forms a measurable product. In particular, the oxidation of the spin trap 2,2,6,6-tetramethylpiperidine (TEMP) to 2,2,6,6-tetramethylpiperidine-1oxyl (TEMPO) has been used to track trends in reactive oxygen species. TEMPO is a stable radical, and its concentration can be quantified with electron paramagnetic spectroscopy (EPR), allowing for a quantitative measurement of certain plasma-liquid interactions. TEMPO has historically been used to track singlet oxygen (1O2) in biological systems; however, plasmas produce a plethora of reactive species, and the production of TEMPO has not been shown to be selective to any particular one of these. Previous research has narrowed the field of potential TEMPO-producing species to O, 1O2, and O3.

In this paper, experiments are shown that separate the TEMPO contributions from each species. The device used is a He/O2 jet in a chamber with a controlled environment to eliminate the effect of variable atmospheric humidity. TEMP was mixed to a 100mM concentration in a 1mL cup of water before being exposed to plasma at various treatment distances; each exposure lasted for 5 minutes. After each experiment, the TEMPO concentration in the plasma-exposed solution was measured with EPR.

Varying the O2 flow in the jet with an air environment showed that increasing the O2 flow raised the O3 concentration (as measured with Fourier Transform Infrared Spectroscopy); however, the TEMPO yield was relatively low (10μM). Meanwhile, experiments with no O2 in the jet produced no O3 but resulted in a high TEMPO yield (90μM). Thus, the contribution of O3 to TEMPO production is minor, and large amounts of TEMPO are produced by non-O3 species originating in the entrainment.

With no O2 in the jet, the applied voltage and treatment distance were varied. These experiments were carried out once with a TEMP solution and once with a solution containing TEMP and NaN3, a 1O2 scavenger. Results of the voltage experiment are shown in Fig. 1. No significant difference was observed with and without NaN3, indicating that, in this jet, TEMPO is produced not by 1O2 but by O. Thus, the use of TEMP oxidation as a selective quantifier of O atom-liquid interactions is enabled, allowing for tracking of plasma-liquid interaction trends caused by this strong oxidizer. References [1] Y. Gorbanev, D. O’Connell, and V. Chechik, Chem. - A Eur. J., 22 (10), 3496-3505 (2016).

Figure 1 – The lack of difference between TEMPO concentrations produced with and without NaN3 shows that, in this jet, TEMPO production is caused by O, rather than by 1O2.

45

Rapid Modeling of Kinetic Reactive Plasma Dynamic Using the Kinetic Global Model Framework

Janez Krek, Guy Parsey and John Verboncoeur

Michigan State University (krek@msu.edu, parseygu@msu.edu, johnv@msu.edu)

Using computer simulations to describe systems that include spatial variation, many species, and complex reaction mechanisms requires a tremendous amount of time and computer resources. Volume-averaged global models reduce the geometrical complexity of simulated models, enabling users to rapidly explore the influence of chemical pathways in such systems. The Kinetic Global Model framework (KGMf), a Python-based framework developed at Michigan State University, offers users great flexibility in defining various system parameters, such as power input, included species and reactions, and an arbitrarily defined electron energy distribution function (EEDF). After being verified with an argon microwave discharge simulation and validated against classical gas phase methane combustion experiment, the KGMf was used to simulate multi-phase chemistry of plasma assisted combustion (microwave assisted jet plane and nanosecond pulse discharge).[1] The KGMf was also modified with two-way averaged intracavity intensity model to simulate rare-gas metastable-laser kinetic behavior. [2]

As shown by PIC-MCC simulations, though a Maxwellian EEDF is a good assumption for pressures up to 50 Torr, a better approximation of the EEDF is required for processes near to atmospheric pressure.[3] Presently, the KGMf pre-computes reaction rates using a parameterized static EEDF definition in advance of each run. Application of a self-consistent evaluation of the EEDF gives better insight to processes in the simulated model, at the cost of some computational time. As the evaluation is computationally intensive, the frequency of the EEDF update may be specified manually, or based on computed parameters, to preserve the speed of the KGMf.

Reaction rate updates are performed via an EEDFsolver class that acts as an interface between the KGMf and a specific implementation of the EEDF evaluation. This allows different implementations of evaluators to be compared. Current implementation of the EEDFsolver class uses BOLOS, an open-source Boltzmann equation solver written completely in Python, but it can be replaced with any other implementation.[4] The ability to compute a dynamic kinetic EEDF will enable rapid study of reactive plasma dynamics, including plasmas strongly driven by complicated voltage waveforms.

References

[1] G. Parsey, Y. Güçlü, J. Verboncoeur and A. Christlieb, ICOPS, doi: 10.1109/PLASMA.2013.6634762 (2013)

[2] G. Parsey, J. Verboncoeur, A. Christlieb and Y. Güçlü, ICOPS, doi: 10.1109/PLASMA.2015.7179543 (2015)

[3] S. K. Nam, and J. P. Verboncoeur, Comput. Phys. Comm., 180 (2009), p. 628 [4] A. Luque, https://pypi.python.org/pypi/bolos (2004)

Figure 1. – The comparison of the EEDF obtained by PIC-MCC and global model (from [4], fig. 3) at 760 Torr.

46

Instabilities at Startup of Pulsed Electronegative Inductively Coupled Plasmas

Steven J. Lanham(a) and Mark J. Kushner(b)

(a) Dept. of Chemical Engineering (sjlanham@umich.edu) (b) Dept. of Electrical Engineering and Computer Science (mjkush@umich.edu)

University of Michigan, Ann Arbor, MI 48109-2122 USA

Pulsing the power applied to an inductively coupled plasma (ICP) has beneficial aspects, such as decreasing average ion energy, allowing extraction of negative ions, and as additional means to customize reactive flux to materials [1,2]. The motivation for pulsing is that reactive species form and decay at different rates in the plasma, and so applying variable power can allow operation in a unique, repeated transient mode. Predicting this behavior is challenging since the process heavily depends on the processing gas, operating conditions, and the reactor geometry. Another complication that occurs when using highly electronegative gas mixtures is that the electron density at the beginning of the power pulse may be orders of magnitude smaller than during the pulse. Modeling pulsing in these highly attaching plasmas inherently implies resolving the transition between capacitive (E-mode) and inductive (H-mode) power deposition during the startup of a pulsed period [3]. This transition also depends on the antenna configuration and matching network.

In this work, the Hybrid Plasma Equipment Model (HPEM), a 2-dimensional simulator, was used to model the dynamics of pulsed Cl2 inductively coupled plasmas [4]. Specifically, this work focuses on resolving the initial power on period during a pulse, while retaining the ability to follow pulse to pulse variation. For example, the spatial distribution of electron density is shown in Fig. 1 for a 15 mTorr Cl2 plasma a few rf periods after the power is turned. The initially small electron density results in the majority of the power being capacitively coupled during the startup transient. Spikes of high temperature electrons form due to oscillations of the initially thick sheaths with electrostatic coupling from the antenna, which then transit into the plasma bulk. The finite spacing of the coils produces local peaks in ionization which individually propagate outwards. These phenomena will be discussed as a function of process parameters (coil frequency, antenna configuration) as well as how the initial power on period can affect the entire pulsed transient.

References [1] S. Banna et al. J. Vac. Sci. Technol. A Vacuum, Surfaces, Film. 30, 40801 (2012). [2] D. J. Economou, J. Phys. D. Appl. Phys. 47, 303001 (2014). [3] M. Zaka-ul-Islam, Phys. Plasmas 23, 113505 (2016). [4] M. J. Kushner, J. Phys. D. Appl. Phys. 42, 194013 (2009).

Figure 1 – Computed electron density a few rf periods after the power is turned on for a 15 mTorr, Cl2 plasma (3.9×107 – 3.9×109 cm-3).

47

Plasma Characterization of Microhollow Anode and Microhollow Cathode Discharges at Moderate Pressures

Sophia Gershman and Yevgeny Raitses

Princeton Plasma Physics Laboratory (sgershma@pppl.gov, yraitses@pppl.gov)

Micro discharges at moderate (> 1 torr) and atmospheric pressure are being investigated for a multitude of applications from surface treatment to pollution remediation. Various microdischarges producing non-equilibrium microplasmas are being investigated for use in these chemical processes. A dc microdischarge with a hollow cathode (MHC) geometry is widely used due to the fact that it maintains the nonthermal properties at relatively high plasma densities [1, 2]. This configuration has been investigated for applications that need relatively high plasma density without high power consumption. In a dc microdischarge configuration with a hollow anode (MHA), a stable, -highly non-equilibrium, and non-local microplasma can be formed with lower density and higher electron energy than the MHC configuration [3,4]. The non-locality of the EEDF is important for high pressure applications because it allows for possible control of electron energy in regions outside of the millimeter or submillimeter size plasma sources, for example close to the surface of a catalyst used in dry reforming or for surface modification.

We have designed and characterized MHC and MHA configurations in a dc regime (Figure 1a) [5]. The discharges were stable in a pressure range of 2 – 10 torr with N2, CO2, and CH4/CO2 as operating gases, with currents < 1 mA (MHA) and up to 10 mA (MHC). The optical emission spectroscopy, electrical, and probe measurements revealed higher electron energies in the MHA (12 – 15 eV) as compared to the MHC (9 – 10 eV) discharge with (Figures 1b). From these measurements, it appears that at high discharge voltages the MHA arrangement makes beam-like electrons possible. Estimations show that within the conditions of these experiments, the beam electron stopping distance is commensurate with the size of the discharge [5]. In experiments with the MHA, high energy electrons were indeed detected outside of the source indicating on the non-local nature of the EEDF.

References: [1] D. Mariotti and M. Sankaran,J. Phys. D: Appl. Phys. 44, 174023 (2011). [2] E. Neyts, Plasma Chem. Plasma Process 36,185 (2016). [3] A. S. Mustafaev, V. I. Demidov, I. Kaganovich, S. F. Adams, M. E. Koepke, and A. Grabovskiy, Rev. Sci. Instrum. 83, 103502 (2012). [4] V. I. Kolobov and A. S. Metel, J. Phys. D: Appl. Phys. 48, 233001 (2015). [5] S. Gershman and Y. Raitses, submitted to Plasma Sources Science and Technology (2017).

Figure 1 A dc microdischarge with cylindrical and flat electodes operated in the hollow anode (MHA) and hollow cathode (MHC) configurations: a) shematic setup, b) Langmuir probe measurements: the logarithmic plot shows a higher electron temperature for the MHA confguration.

‐5

‐4.5

‐4

‐3.5

‐3

‐2.5

‐2

‐40 ‐30 ‐20 ‐10 0 10 20

Ln(I), arb. u

nits

Probe Voltage, V

Logarithmic Plot of Probe Results

MHA

MHC

‐4.5

‐4

‐3.5

‐3

‐2.5

‐2

‐1.5

‐1

‐0.5

0

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

‐40 ‐20 0 20 40 60

LN(I), arb. u

nits

Curren

t, x10^‐7 A

Probe Voltage, V

MHA Probe Results 

MHA LN(I)

MHA Probe V(I)

a)

b)

48

A Case Study of Plasma-Surface Interactions at Atmospheric Pressure: Polystyrene Treatment Using an RF Plasma Jet

P. Luan(a), A. J. Knoll(a), V. S. S. K. Kondeti (b), P. J. Bruggeman(b) and G. S. Oehrlein(a)

(a) University of Maryland, College Park (oehrlein@umd.edu) (b) University of Minnesota, Twin Cities (pbruggem@umn.edu)

Cold atmospheric pressure plasma jets (APPJs) are able to generate a long but narrow plume which contains chemically reactive species useful for material processing and biomedical applications. Fundamental understanding of how material surfaces interact with these reactive species is required for both scientific and engineering reasons. Previously we studied the interaction of a well-characterized radio frequency (RF) APPJ with model polymers by avoiding direct contact with the plume, and we found a linear response between incident O flux and etched C flux, with the etching reaction probability of O atoms to carbon atoms in the range of 10-4 to 10-3 [1].

In this work we extended our studies of plasma-surface interaction (PSI) to the full downstream of the RF jet vicinity. A number of plasma processing parameters, such as treatment distance and angle, feed and environment gaseous composition, were investigated by evaluating both thickness change and surface chemical composition of polystyrene (PS) after treatment. The role of different plasma species on polymer surface was compared.

We found that for both Ar/O2 and Ar/H2O the etch rate of PS decayed exponentially with treatment distance (4 mm – 20 mm), whereas surface oxidation increased to a maximum and then decreased. This is shown for Ar/H2O in N2 and air environments in figure 1(a). Both the exponential decay constant and oxidation maximum varied with feed gas and environment gas compositions due to changes in the density profile of gas phase reactive species. A surface reaction model based on Langmuir adsorption can explain the difference in trends and the lower modification rate at high etch rate. We find that the degree of surface oxidation θ is a function of both etchant species flux and modification species flux: θ monotonically increases with the flux of modification species and decreases with the flux of etchant species (figure 1(b)). The reaction rate between incident atomic O (OH radicals) flux and etched C flux leaving the polymer surface was estimated by comparing O atoms (OH radicals) density profile measured with the same RF jet. Additionally, the apparent activation energy (Ea) of etching reactions was measured by varying substrate temperature. We found that Ea of Ar/O2 plasma (0.10 – 0.13 eV) is higher than for Ar/H2O plasma (0.02 eV).

* We gratefully acknowledge helpful discussion with D. B. Graves at UC Berkeley and funding from US Department of Energy (DE-SC0001939) and National Science Foundation (PHY-1415353). References [1] P. Luan, A. J. Knoll, H. Wang, V. S. S. K. Kondeti, P. J. Bruggeman, and G. S. Oehrlein, J. Phys. D. Appl. Phys. 50, 03LT02 (2017).

Figure 1 – (a) Spatial profiles of etching depth and surface O composition of PS treated by Ar/H2O plasma in air and N2 environment. (b) Effect of calculated etchant species (atomic O) and model predicted modification specis (Γmod) on the surface modification rate of PS shown as surface O composition (measured by XPS).

49

On the Variation of the Activation Energy of Polymer Etching by Cold Atmospheric Plasma (CAP) Sources Under Well-Defined Conditions

A. J. Knoll(a), P. Luan(a), V. S. S. K. Kondeti (b), P. J. Bruggeman(b)

and G.S. Oehrlein(a)

(a) University of Maryland, College Park (oehrlein@umd.edu) (b) University of Minnesota, Twin Cities (pbruggem@umn.edu)

Cold atmospheric plasma (CAP) sources are important sources of reactive chemical species

that can used to deactivate bacteria and biomolecules or modify surfaces under mild conditions, leading to use in numerous applications. The plasma source type and a multitude of treatment conditions can lead to different effects for a given surface caused by plasma-surface interactions. Our previous studies have shown that direct plume-surface interaction and VUV photons can play important roles in surface modification under certain conditions. A well characterized RF jet has also been shown to have a linear dependence of etching on the atomic oxygen flux to the surface [1]. In this study we examine in more detail the effect of varying substrate temperature and other details on polymer etching. We also compared a surface microdischarge (SMD) source which operates without noble gas flow and very effectively modifies polymer surfaces without etching. We found that polymer etch rate increases with temperature. Sample heating due to plasma exposure is insignificant and cannot explain etching effects. An Arrhenius equation is used to fit the temperature dependence of etch rate and yields apparent activation energies. Figure 1 shows how this apparent activation energy increases for 1% O2 admixture with increasing distance and is significantly lower for 1% H2O, for which is does not change as significantly with distance. From low pressure plasma treatments, photoresist polymers tend to have an activation energy of around 0.5 eV for pure oxygen [2]. Additionally, we investigated the effect of gas flow rate from the RF jet which we found previously is critical to the diffusion of ambient gas species into the plasma plume. The SMD source causes a small amount of etching while providing a large amount of surface modification which is very different from the behavior of the RF jet. We investigate the effect of ambient gas chemistry on the SMD treatment of polymer at various temperatures. The apparent activation energy for SMD treatments is on the order of 0.03 eV and with very little total etching. The observed differences in activation energy provide evidence of strongly changing plasma-surface interaction conditions over small APPJ/surface separations. In the case of Ar/O2, other reactive species may assist O atoms in the etching of these polymers, e.g. charged particles and photons, and change the rate of chemical etching.

The authors gratefully acknowledge financial support by US Department of Energy (DE-SC0001939). References [1] P. Luan, A. J. Knoll, H. Wang, V. S. S. K. Kondeti, P. J. Bruggeman, and G. S. Oehrlein, J. Phys. D. Appl. Phys. 50, 03LT02 (2017). [2] D'Agostino, R., Flamm, D. L., & Auciello, O. (1990). Plasma Deposition, Treatment, and Etching of Polymers: The Treatment and Etching of Polymers). Burlington: Elsevier Science.

Figure 1 – Dependence of apparent activation energy of polymer etching on the distance of the RF jet to the sample and feed gas chemistry.

50

Plasma Diagnostics in Air Plasmas Containing Water Droplets

Toshisato Ono(a), Eray S. Aydil(b) and Uwe R. Kortshagen(a)

(a) Dept. of Mechanical Engineering, University of Minnesota (onoxx012@umn.edu, korts001@umn.edu)

(b) Dept. of Chemical Engineering and Materials Science, University of Minnesota (aydil@umn.edu)

Atmospheric pressure air plasmas are often used to generate reactive oxygen and nitrogen species (RONS). Roles of RONS, especially nitric oxide (NO), in plant biology have been well studied by biologists. Atmospheric pressure plasmas are emerging as RONS sources [1] with potential applications in irrigation with plasma-treated water for enhanced agriculture. Consequently plasma-liquid interactions are receiving increased research attention. Of particular interest is the distance that reactive species diffuse into liquid before they react. Calculations by Lindsay et al. suggest most of the RONS are depleted within first tens of microns, while OH diffusion lengths are even smaller (~10 nm). [2] In this study, we conducted experiments on air plasmas containing water droplets produced by a liquid atomizer. Large surface area of the droplets enhances surface reactions and this spray configuration is particularly suitable for large-scale water treatment.

The reactor configuration used in our experiments is shown in Figure 1. A liquid atomizer made of stainless steel acts as the cathode and is positioned at the top of the reactor. The atomizer nozzle is made of Teflon. The anode is a tungsten ring electrode embedded in the Teflon as shown in Figure 1. The air plasma is generated between the two electrodes by powering the anode with a high voltage AC power supply (5-15 W). Air at high velocities (25 SLM through a 3 mm nozzle) is fed from the

top and merges with deionized water creating water droplets ranging 20-80 m. These water droplets are injected into and interact with the plasma.

Optical emission spectroscopy (OES) was used to detect emission from species created in the plasma. Figure 2 shows emission spectra taken 5 mm below the nozzle exit. The emissions detected at this location are from N2 (C-B) (band maximum at 377 nm), NH at 336 nm, O at 777 nm and NO (200-400 nm band). [3] Strong emission from Fe is due to the erosion of the electrode. This result confirmed that RONS were generated under humid and high-speed air flow environment, and that high power plasma dissociated N2 leading to a production of reactive nitrogen species.

References [1] D. B. Graves, J. Phys. D: Appl. Phys. 45, 31-34 (2012). [2] A. Lindsay, C. Anderson, E. Slikboer, S. Shannon and D. B. Graves, J. Phys. D: Appl. Phys. 48, 10-12 (2015). [3] P. Bruggeman, D. Schram, M. Á. González, R. Rego, M. G. Kong and C. Leys, Plasma Sources Sci. Technol. 18, 3-5 (2009).

Figure 1 – Sketch of the water plasma spray reactor showing the teflon nozzle, tungsten and stanless steel electrode powered by an AC power supply.

Figure 2 – A typical optical emission spectrum from an air plasma containing water droplets at two different powers.

51

2D Simulations of the Carbon Arc Discharge for Synthesis of Nanotubes

Alexander Khrabry, Andrei Khodak and Igor Kaganovich

Princeton Plasma Physics Laboratory, Princeton NJ 08542 (akhrabry@pppl.gov)

Self-consistent model of atmospheric pressure carbon arc discharge in helium atmosphere was developed in the framework of the nanoparticle synthesis project [1] and implemented into the 3D CFD-code ANSYS CFX. In order to simulate the arc discharge, the CFX code was customized to incorporate processes of heat transfer and current flow in electrodes, ablation of the anode, carbon deposition at the cathode and near-electrode space charge sheathes. The model also takes into account radiation from electrodes surfaces (including mutual radiation), Joule heating of electrodes, thermal resistance of the deposit at the cathode. Note that the radiation from electrodes constitutes a large fraction of the total heat flux. Plasma model accounts for non-equilibrium conditions in the arc and solves separate transport equations for electron, ion and gas species.

Special numerical procedure for coupling of the plasma current, emission current, ion current, sheath voltage drop, heat fluxes at plasma-electrode interfaces was developed and implemented into the ANSYS CFX code. To account for ablation/deposition a new boundary condition was developed that can switch between negative and positive sheathes at anode, which enhances the more common approach used for example in Ref. [2].

For benchmarking of the sheath model and plasma transport coefficients, 1D computations were performed and compared with the results of previous numerical studies [3]. Results of 2D computations were also validated by comparison with experimental data, and rather good agreement was observed.

One of the most important values characterizing synthesis process in the carbon arc discharge is the rate of carbon ablation from the anode and its deposition on the cathode. Figure 1 displays variation of ablation and deposition rates with the arc current. Computations are performed for the same conditions as in experiments [4] (helium pressure 0.66 atm., cylindrical cathode and anode 12 and 6 mm in diameter, 1.5mm inter-electrode gap). As can be seen from the figure, the computational results and experimental data show the similar trend and are close to each other.

This work was supported by the U.S. Department of Energy. References [1] http://nano.pppl.gov [2] M. Kundrapu and M. Keidar, "Numerical simulation of carbon arc discharge for nanoparticle synthesis", Phys. Plasmas 19, 073510 (2012). [3] N. Almeida et al., “Unified modelling of near-cathode plasma layers in high-pressure arc discharges”, J. Phys. D: Appl. Phys. 41 245201 (2008). [4] V. Vekselman et al., “Complex structure of the carbon arc discharge for synthesis of nanotubes”, accepted to Plasma Sources Science and Technology (2017).     

Figure 1- Ablation (a, b) and deposition (a) rates as a function of the arc current. Experimental data are taken from Ref. [4]

52

Tentative List of Participants

Last Name First Name Institution Email Adamovich Igor Ohio State University adamovich.1@osu.edu Aydil Eray University of Minnesota aydil@umn.edu Baratte Edmond Ohio State University Barnat Ed Sandia National Labs evbarna@sandia.gov Bolton Curt U.S. Dept. of Energy Curt.Bolton@science.doe.gov Boyd Iain University of Michigan iainboyd@umich.edu Bruggeman Peter University of Minnesota pbruggem@umn.edu Chen Xiang Princeton Plasma Physics Lab Xiang.Chen@Ugent.be;

coshxsinhx11@gmail.com Donnelly Vince University of Houston vmdonnelly@uh.edu Economou Demetre University of Houston chee1m5@central.uh.edu Efthimion Philip Princeton Plasma Physics Lab pefthimi@pppl.gov Elg Daniel Univ. California-Berkeley daniel.elg@berkeley.edu Engeling Kenneth University of Michigan kenengel@umich.edu Fierro Andrew Sandia National Labs asfierr@sandia.gov Finnegan Sean U.S. Dept. of Energy Sean.Finnegan@science.doe.gov Foster John University of Michigan jefoster@umich.edu Fu Yangyang Michigan State University fuyangyang2020@gmail.com Gershman Sophia Princeton Plasma Physics Lab sgershma@pppl.gov;

sgershman@arisenj.org Girshick Steven University of Minnesota slg@me.umn.edu Glen Crystal Sandia National Labs Godyak Valery University of Michigan egodyak@comcast.net Graves David Univ. California-Berkeley graves@berkeley.edu Guo Heng Princeton Plasma Physics Lab hguo@pppl.gov Hebner Greg Sandia National Labs gahebne@sandia.gov Kaganovich Igor Princeton Plasma Physics Lab ikaganov@pppl.gov Kawamura Emi Univ. California-Berkeley emi119@comcast.net;

kawamura@eecs.berkeley.edu Khrabrov Alexander Princeton Plasma Physics Lab akhrabro@pppl.gov Khrabry Alexander Princeton Plasma Physics Lab akhrabry@pppl.gov Kleinkonradt Marcel University of Maryland mkleinko@umd.edu Knoll Andrew University of Maryland knollaj@gmail.com Kolobov Vladimir CFDRC/University of

Alabama at Huntsville vik@cfdrc.com; vladimir.kolobov@cfdrc.com

Kondeti Santosh Kumar

University of Minnesota sivas006@umn.edu

Kortshagen Uwe University of Minnesota uk@me.umn.edu Krek Janez Michigan State University krek@msu.edu Kruszelnicki Juliusz University of Michigan jkrusze@umich.edu Kushner Mark J. University of Michigan mjkush@umich.edu Lai Janis University of Michigan janislai@umich.edu Lanham Steven University of Michigan sjlanham@umich.edu Lee Brad Samsung

53

Last Name First Name Institution Email

Li Chen University of Maryland chenli@umd.edu Lieberman Mike Univ. California-Berkeley lieber@eecs.berkeley.edu;

liebermn@pacbell.net Lin Kang-Yi University of Maryland kxl259@umd.edu Luan Pingshan University of Maryland luan.pingshan@gmail.com Luo Yuchen University of Minnesota luoxx364@umn.edu Nayak Gaurav University of Minnesota nayak025@umn.edu Nguyen Tam University of Houston tamtintx@yahoo.com Oehrlein Gottlieb University of Maryland oehrlein@umd.edu Ono Toshisato University of Minnesota onoxx012@umn.edu Podder Nirmol U.S. Dept. of Energy Nirmol.Podder@science.doe.gov Powis Andrew Princeton Plasma Physics Lab apowis@pppl.gov Pranda Adam University of Maryland apranda@umd.edu Raitses Yevgeny Princeton Plasma Physics Lab yraitses@pppl.gov Shafi Leith University of Maryland Simeni Simeni

Marien Ohio State University simenisimeni.1@osu.edu

Smith Brandon University of Michigan bradenis@umich.edu Synakowski Edmund U.S. Dept. of Energy Ed.Synakowski@science.doe.gov Thomas Edward Auburn University etjr@physics.Auburn.EDU Van Dam James U.S. Dept. of Energy James.VanDam@science.doe.govVekselman Vladislav Princeton Plasma Physics Lab vekselman@pppl.gov Verboncoeur John Michigan State University johnv@msu.edu