Available online at www.sciencedirect.com

Planetary and Space Science 50 (2002) 1355–1360

www.elsevier.com/locate/pss

Plasma diagnostics and simulation for the SMART-1 mission

M. Tajmara ;∗, J. Gonz,aleza, G. Saccocciaa, G. Nocib, H. Laaksoc

aElectric Propulsion Section, Directorate of Technology and Operation Support, ESA/ESTEC, P.O. Box 299, 2200 AG Noordwijk, The NetherlandsbLABEN S.p.A. Proel Tecnologie Division, Viale Machiavelli 31, 50125 Firenze, Italy

cFinnish Meteorological Institute, P.O. Box 503, 00101 Helsinki, Finland

Accepted 27 June 2002

Abstract

SMART-1 is a technology demonstrator for using primary electric propulsion on interplanetary spacecraft. Hence, studying of theinteraction of the plasma emitted by the thruster with the environment and the spacecraft is one of the top priorities during the mission. Twoexperiments (Electronic propulsion diagnostic package and Spacecraft potential, electron and dust experiment) are available to measurethe electron densities and temperatures as well as wave electric 5elds during the operation of the electric propulsion thruster. Additionally,a retarding potential analyser, a quartz microbalance and a solar-cell sample will analyse data from slow charge-exchange ions which are apotential contamination source. ESTEC is developing a 3D particle-in-cell model in order to study the spacecraft/environment interactionson SMART-1 and interpret the measurements. In the present paper, we will review the contamination e:ects associated with electricpropulsion and how the plasma sensors cover them. We further present preliminary results from the numerical simulation and show howthe <ight data will be used to validate the modelling code. A successful validation of the simulation will support future interplanetary andcommercial missions featuring electric propulsion to reduce the risk of contamination and interference with on board instruments.? 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Electric propulsion; Spacecraft interaction; Charge exchange plasma

1. Introduction

SMART-1 is the 5rst of the Small Missions for AdvancedResearch in Technology of the ESA Horizons 2000 scienti5cprogramme. The mission is dedicated to the testing of newtechnologies for preparing future cornerstone missions, us-ing solar-electric propulsion in deep space. SMART-1 willbe placed in orbit around the moon using a Hall thruster(PPS-1350) with a maximum thrust of 70 mN built by SEP(Valentian and Maslennikov, 1997; Lyszyk et al., 1999). Itwill be launched in 2003 as an Ariane 5 cyclade-like auxi-lary payload.

This will be the 5rst time of primary electric propulsionon a European spacecraft. Hence, the evaluation of the Hallthruster impact on the spacecraft and its instruments is oneof the primary scienti5c objectives (Racca et al., 1999). Ad-dition to primary beam ions, electric propulsion thrusters

∗ Corresponding author. Space Propulsion Section, Austrian Re-search Centers Seibersdorf, Austria. Tel.: +43-50550-3142; fax:+43-50550-2813.

E-mail address: martin.tajmar@arcs.ac.at (M. Tajmar).

create a low-energy charge-exchange ion environment. Thedistribution of these ions is strongly a:ected by the potentialdistribution near the spacecraft being a potential contam-ination source for instruments and solar arrays. Althoughcharge-exchange plasma interactions have been a subject ofextensive experimental and theoretical studies (e.g. Carruth,1981; Tajmar et al., 1999; Wang et al., 1999), there havebeen few comprehensive in-<ight investigations due to thelack of <ight opportunities. The 5rst interplanetary space-craft using solar electric propulsion is deep space one (Wanget al., 1999) using the NSTAR ion engine. SMART-1 willbe the 5rst interplanetary <ight using a Hall thruster.

Two payload experiments (EPDP, SPEDE) are dedi-cated to measure the ambient plasma variables during theoperation of the Hall thruster. ESTEC is currently de-veloping modelling tools to predict and help to interpretinstrument data to study spacecraft/environment interac-tions on SMART-1. This paper will present an overviewof the plasma diagnostics and preliminary modelling re-sults to assess possible contamination issues. A successfulvalidation of the simulation with in-<ight data will supportfuture interplanetary and commercial missions featuring

0032-0633/02/$ - see front matter ? 2002 Elsevier Science Ltd. All rights reserved.PII: S0032 -0633(02)00129 -0

1356 M. Tajmar et al. / Planetary and Space Science 50 (2002) 1355–1360

electric propulsion to reduce the risk of contamination andinterference with on board instruments.

2. Spacecraft plasma sensors overview

SMART-1 is a cube spacecraft with the dimensions 1:15×1:15 × 1 m and two solar arrays stretching out from twoopposite sides giving a total length of 8 m. A schematic loca-tion of the thruster and the electric propulsion related instru-ments is shown in Fig. 1. The PPS-1350 Hall e:ect thrusterwill operate at an speci5c impulse of 1640 s delivering amaximum thrust of 70 mN using Xenon gas as propellant.

During operation, the thruster emits an ion beam with adivergence of about 40◦. Typical operating parameters aresummarised in Table 1. Electrons from an external cathodeact as a neutraliser creating a quasi-neutral plasma. Al-though the propellant eHciency for these type of thrustersexceed 95%, the neutral density is comparable to thebeam ion density due to the much lower thermal velocities(400 m=s) compared to the ion velocities gained due to theacceleration potential of 350 V (22; 500 m=s). Resonantcharge-exchange collisions between the ionised and neutraleJuents create slow ions that can be distributed around thespacecraft following the potential distribution in the vicinityof the spacecraft.

Part of these charge-exchange ions will <ow back tothe surface causing variations in the spacecraft’s <oatingpotential. Depending on the impact energy, back<ow ionscan also cause sputtering on the surface. As a spiral or-bit raising phase in the Earth’s magnetosphere is part ofthe mission which also in<uences the <oating potential, thecharge-exchange ion e:ects will di:er along the trajectory.

Several plasma sensors onboard SMART-1 will charac-terise the ambient plasma and the e:ects on the low-energycharge-exchange plasma emitted by the Hall thruster. In thissection, we will give an overview of the instruments in-volved in the evaluation of the spacecraft/environment in-teractions related to electric propulsion:

2.1. Electric propulsion diagnostic package

This package consists of four instruments outside the pri-mary ion beam (Langmuir probe, retarding potential anal-yser, solar cell sample, and quartz crystal microbalance)aiming at characterising the charge-exchange ion environ-ment around the spacecraft (Capacci et al., 1999). The in-strument requirements are summarised in Table 2.

2.1.1. Langmuir probeThe spherical Langmuir probe is located 55 cm next to

the Hall thruster (Fig. 2). This sensor provides informa-tion about the plasma potential, the electron density and thetemperature, respectively. The charge-exchange ion trajec-tories are determined by the potential distribution createdby space charge e:ects around the spacecraft. Outside the

Fig. 1. Schematic instrument location on SMART-1.

Table 1PPS-1350 performance parameters

Parameters PPS-1350 thruster

Thrust 70 mNVoltage 350 VCurrent 3:8 AMass <ow rate 4:2 mg=sSpeci5c impulse 1640 sPower 1350 WTotal eHciency 51%Divergence angle 42◦

Table 2EPDP instrument requirements

Physical parameters Instrument requirements

Plasma density 1013–1014 m−3

Ion energy 0–400 eVElectron temperature 1–5 eV (focus on 1.7–3:5 eV)Plasma potential −150–100 VIon current density 0.001–1 mA=cm2

(focus on 0.002–0:05 mA=cm2)Deposition 0–0:44 mg=cm2

Fig. 2. EPDP location.

primary beam ions, the potential is composed of the spacecharges from the charge-exchange ions and the neutraliserelectrons. Almost all studies up to now (Oh and Hastings,1996; VanGilder and Boyd, 1998; VanGilder et al., 1999)

M. Tajmar et al. / Planetary and Space Science 50 (2002) 1355–1360 1357

treat the Hall thruster beam as a quasi-neutral plasma andassume a Boltzmann energy distribution of the electrons.

Hence, the Langmuir probe data provides useful informa-tion on the electron behaviour and therefore on the potentialand charge-exchange distribution near the thruster.

2.1.2. Retarding potential analyserThis sensor is located next to the Langmuir probe on

the same probe assembly (Fig. 2). The retarding poten-tial analyser (RPA) measures the ion energy and currentdensity distribution passing through a grid structure. Thecharge-exchange ion energy is of crucial importance to pre-dict sputtering phenomena on the spacecraft surface. Onlyif the energy is above a certain material dependent thresh-old (Aluminium: 68 eV, Silicon 150 eV, Rosenberg andWehner, 1962), sputtering occurs. The predicted energy isin the order of several tens of volt, below such thresholds.However, extremely negative <oating potentials as acquiredduring eclipses (up to −1000 V or more, Olsen, 1981) willchange the energy distribution which is diHcult to simulatenumerically. This will not only e:ect sputtering, but alsoa change in the ion distribution leading to possible con-tamination of other parts than the thruster area, e.g. to thesolar arrays or even instruments. A detailed analysis of thecharge-exchange energy distribution at di:erent <oating po-tential conditions will give valuable answers to this problem.

2.1.3. Solar cell sampleA solar cell sample will be mounted on the −X panel

of SMART-1 to study possible degradation due to the op-eration of the electric propulsion system (Fig. 1). If powerlosses are observed, they also provide information of thecharge-exchange ion density and energy related to sputter-ing of the solar cell’s cover glass causing the degradation.The presence of charge-exchange ions in this location is,however, unlikely. As the simulation results in Section 4show, the charge-exchange ions are expected to expand ra-dially from the primary ion beam. The only possible mech-anism to attract ions to the solar panel location is a changein the ambient potential structure or large di:erent <oatingpotential conditions depending on the orbit and eclipses.

Hence, the analysis of these data is crucially linked tothe Langmuir probe and the SPEDE sensors which provideinformation about the ambient plasma environment and thespacecraft potential, as shown later.

2.1.4. Quartz crystal microbalanceThis sensor is located next to the solar cell sample to mon-

itor possible deposition of propellant ions during thrusteroperation (Fig. 1). Deposition is especially important for op-tical instruments like cameras. As already mentioned above,the presence of charge-exchange ions at this location is un-likely. Similar measurements in deep space one indicated thepresence of an charge-exchange ion <ux to a Langmuir probeon the opposite side of the thruster (Wang et al., 1999).

Fig. 3. SPEDE location.

However, changes in the <oating potential were not mon-itored. The <ux was orders of magnitude below the solarwind <ux and occurred only at certain high thrust levelconditions. If sensor data appears during thruster operation,quartz crystal microbalance (QCM) data will also contributeto a better understanding of the interaction between thecharge-exchange ions and the ambient plasma environment.

2.2. Spacecraft potential, electron and dust experiment

The SPEDE experiment consists of two electric sensors ofcylindrical shape mounted on the ends of two 60 cm booms(Fig. 3). Each sensor can work either in a Langmuir (LP)mode or in an electric 5eld (EF) mode.

When operated in an EF mode, the sensor is current-biased,and both the spacecraft potential and wave electric 5eldscan be monitored. As already pointed out, large variationsin the spacecraft potential a:ect the charge-exchange ionsdistribution. These measurements will aid the analysis ofpossible contamination detected by the solar cell sampleand the QCM. Also, gas molecules absorped on the space-craft will later be slowly desorped, resulting in enhancedplasma wave activity (Laakso et al., 2002).

Using the potential measurement of an EF sensor and theelectron temperature from the EPDP Langmuir probe, wecan even estimate the charge-exchange ion density, assum-ing a Boltzmann energy distribution of the neutralising elec-trons.

In an LP mode, the sensor is voltage-biased in orderto monitor the variation of the electron <ux. An increaseof the electron <ux would also indicate the presence ofcharge-exchange ions in a quasi-neutral plasma.

3. Spacecraft/environment modelling for SMART-1

Both EPDP and SPEDE perform single-point measure-ments of the charge-exchange ion environment produced bythe Hall thruster. Therefore, a numerical model is necessaryto predict the whole plasma environment around the space-craft and to interpret and relate the obtained measurements.

Hall thrusters have been extensively tested in vacuumchambers and abundant data is available to verify numericalmodels with them (Manzella, 1994; Manzella and SPankovic,1995; Kim et al., 1996; King, 1998; Perot et al., 1996).

1358 M. Tajmar et al. / Planetary and Space Science 50 (2002) 1355–1360

Several models have been developed (Oh, 1997; VanGilderand Boyd, 1998; VanGilder et al., 1999) obtaining a goodagreement between measurements and simulation.

All models use a hybrid particle-in-cell (PIC) algorithm(Birdsall and Langdon, 1991), treating ions and neutrals astest particles and the neutraliser electrons as a <uid assuminga Boltzmann energy distribution. Most models are 2D ax-isymmetrical (VanGilder and Boyd, 1998; VanGilder et al.,1999), only the model from Oh and Hastings (1996) en-ables 3D and real spacecraft geometries. However, all mod-els up to now did not account for the real physical di-mensions of the instruments like the Langmuir probe orthe retarding potential analyser with which they comparedtheir simulation data. This is especially necessary if largegradients appear in the data as shown by measurements(Kim et al., 1996).

ESTEC is developing a 3D hybrid PIC code capable ofpredicting the plasma environment of a Hall thruster in realspacecraft geometries with virtual plasma sensors of spec-i5ed dimensions. The code has been tested against severalmeasurements and is applied to the SMART-1 spacecraftfeaturing virtual EPDP and SPEDE instruments. This sectionwill give a brief overview of the model, a detailed descrip-tion will be presented in a subsequent publication (Tajmaret al., 2000).

3.1. Numerical model

The code is based on a PIC-MCC code (Tajmar andWang, 2000; Birdsall, 1991) used to track ions and neutralsacross the simulation domain and applies a Monte-Carlomethod to compute charge-exchange collisions betweenthem. The plume emitted from the Hall thruster is assumedto be quasi-neutral. Furthermore, the neutraliser electronsare treated as a collisionless <uid neglecting magnetic 5elde:ects. This approach is similar to previous successfulmodelling approaches.

Usual PIC methods need to compute the potential using aPoisson equation solver. However, in a quasi-neutral plasma,the ion density can be used to obtain the potential once theelectron temperature and far-5eld electron density is known

ni ≈ ne = ne; ref exp(e�kTe

): (1)

Instead of using a reference electron density, a referencepotential can be set at the thruster’s exit plane. In this simu-lation, a 5xed electron temperature of 4 eV and a referencepotential of 20 V have been set as suggested by measure-ments from Haas (Haas and Gallimore, 1999) scaling downhis 5nding to a 1:3 kW Hall thruster as utilised in SMART-1.

10% of the ion density at the thruster’s exit plane areassumed to originate from double-charged Xenon ions(Manzella and SPankovic, 1995). Using the thrust T and thetotal mass <ow dm=dt through the anode, we can express themass <ow and velocity of each specie. Not all propellant is<owing through the anode, approximately 10% is directed

Fig. 4. Ion beam velocity vectors at the thruster’s exit.

through the cathode contributing to a large extent to theneutral propellant environment in front of the thruster.

The ions which are injected are randomly distributed be-tween the inner and outer radius of the Hall thruster (Fig. 4)with a divergence angle linearly varying with the radial po-sition as indicated by measurements from Manzella (1994).As these measurements and work by King (1998) shows,the high-energy ions are directed outwards of the centre ofthe thruster. Hence, we assume a smaller divergence angleat the left boundary than at the right boundary as shown inFig. 4. The neutrals are placed randomly along the anode andcathode exit following a standard cosine distribution. Addi-tionally for ground testing comparisons, the vacuum cham-ber pressure is considered to contribute to a homogenousneutral propellant background density at room temperature.

A Monte Carlo method is used to compute thecharge-exchange collisions between the ion and neutralbeam. Therefore, a probability of collision P is added ateach time step dt to an ion particle

P = 1 − exp[ − vrelative CEXnn(x; y; z)dt]; (2)

where vrelative is the velocity di:erence between the ion andneutral beam, nn the neutral density at the particle’s posi-tion (including the vacuum chamber background density),and CEX the collision cross section which is derived fromquantum theory expressed by

CEX(vrelative) = (k1 ln vrelative + k2)2; (3)

where k1 =−0:882× 10−10 s and k2 = 1:513× 10−9 m forXe+–Xe(Rapp and Francis, 1962) and k1=−2:704×10−10 sand k2=3:6×10−9 m for Xe2+–Xe (Oh and Hastings, 1996)collisions, respectively. If P is greater than a random num-ber Rn between 0 and 1, a collision occurred and the mo-mentum is exchanged with the closest neutral particle nextto the ion particle.

All particles are moved along a homogenous grid struc-ture using a standard leap-frog algorithm. The cell size isnot limited by the Debye length as in general PIC simula-tions due to the assumption of a quasi-neutral plasma. How-ever, if we put a plasma sensor in the simulation domain,the grid must at least resolve the probe diameter. Typicalsimulation parameters are a grid size of 100 × 100 × 100and up to 1,500,000 computer particles. The computations

M. Tajmar et al. / Planetary and Space Science 50 (2002) 1355–1360 1359

Fig. 5. y–z Plot of ion density through middle of SMART-1 spacecraft.

last at least 1 day on a Silicon Graphics Power ChallengeXL workstation.

4. Simulation of the SMART-1 plasma environment

In this section, we will show some initial modellingresults of the SMART-1 plasma environment due to the op-eration of the Hall thruster. Fig. 5 plots the ion density on ax–z plane through the middle of the thruster and theSMART-1 spacecraft including the solar arrays. The initialbeam divergence of the primary beam ions is clearly evi-dent. In the middle of the thruster, the ion density reaches avalue of 1 × 1017 m−3. The charge-exchange ions radiallyleave the beam creating an ion density 5 orders of mag-nitude less than at the thruster’s exit. Most important, wesee that the ion’s space charge is not suHcient to expandthe ion beam down in the direction of the solar arrays. Thisshows that the operation of the Hall thruster does not causecontamination to the spacecraft other than at the top surfacewhere the thruster is located.

This simulation has been computed assuming an initialspacecraft <oating potential of 0 V. As already mentionedabove, fast transients in the potential can in<uence the dis-tribution of the charge-exchange ions. However, due to ourinitial assumption of a quasi-neutral plasma used to derivethe potential distribution, charged surfaces cannot be treatedin this code. This would require to solve the potential at ev-ery time step using the charge density on the grid and theboundary conditions on the surface which is computationallyvery expensive (Tajmar et al., 2000). If <ight data may re-

Fig. 6. Virtual EPDP-RPA sensor.

veal such fast potential transients and accordingly an impactof charge-exchange ions at sensors outside the top surface.In that case, a new modelling approach would be neces-sary to avoid quasi-neutrality. Since code veri5cations withground test data was successful using the quasi-neutralityassumption (Oh and Hastings, 1996), the present numericalmodel does not address problems related to fast spacecraft<oating potential transients.

The ions will form a space charge potential hump in frontof the thruster which will de<ect the charge-exchange ions.Virtual plasma sensors were implemented in the simulationto compare with ground tests and in-<ight measurements.

1360 M. Tajmar et al. / Planetary and Space Science 50 (2002) 1355–1360

As an example, Fig. 6 shows the data of the virtual RPAsensor from the EPDP next to the ion beam. It is a functionof the potential distribution and maximum potential builtup in front of the thruster. The data shows a peak around20 eV going down to a maximum of 35 eV. The 20 V peakcorresponds to the potential hump in front of the thruster(Tajmar et al., 2000). These data suggest no sputtering onthe spacecraft surface (Aluminium has a sputter thresholdof 68 eV).

5. Conclusion

SMART-1 will be the 5rst interplanetary spacecraft usinga Hall thruster. Two plasma experiments (EPDP, SPEDE)will assess the change of the plasma environment aroundthe spacecraft before, during and after the operation of thethruster. A 3D PIC-MCC code has been developed to studyspacecraft/environment interactions related to the SMART-1mission. The code is capable of simulating the Hall thrusteron the spacecraft geometry including virtual sensors to sim-ulate the plasma instruments.

Preliminary modelling data assuming a quasi-neutralplasma suggests that the ion beam will not in<uence parts ofthe spacecraft other than the top surface where the thrusteris located. The peak energy of the charge-exchange ions<owing back to the surface was found to be 20 eV. This iswell below the sputter yield threshold of Aluminium, thematerial of the spacecraft surfaces.

A successful validation of the model with in-<ight datawill provide mission designers with a very powerful tool tostudy spacecraft/environment interactions for Hall thrusters.

Acknowledgements

We acknowledge many helpful discussions with B.Foing (Project Scientist of SMART-1), A. Hilgersand D. Estublier. This work was carried out during aYoung-Graduate-Traineeship of M. Tajmar at ESTEC.

References

Birdsall, C.K., 1991. Particle-in-cell charged-particle simulations, plusMonte Carlo collisions with neutral atoms, PIC-MCC. IEEE Trans.Plasma Sci. 19 (2), 65–85.

Birdsall, C.K., Langdon, A.B., 1991. Plasma Physics via ComputerSimulation. Institute of Physics Publishing, New York.

Capacci, M., Matticari, G., Noci, G., Severi, A., Borie, D., 1999.An electric propulsion diagnostic package for the characterizationof the plasma thruster/spacecraft interactions on STENTORsatellite. Proceedings of the AIAA Joint Propulsion Conference,AIAA-99-2277.

Carruth, M. (Ed.), 1981. Experimental and Analytical Evaluation of IonThruster/Spacecraft Interactions. JPL Publication, pp. 80–92.

Haas, J.M., Gallimore, A.D., 1999. Characterization of the internal plasmastructure of a 5 kW Hall thruster. Proceedings of the InternationalElectric Propulsion Conference, Kitakyshu, Japan, IEPC-99-078.

Kim, S.W., Foster, J.E., Gallimore, A.D., 1996. Very near 5eld plumestudy of a 1:35 kW SPT-100. Proceedings of the AIAA JointPropulsion Conference, Lake Buenavista, FL, AIAA-96-2972.

King, L.B., 1998. Transport-property and mass spectral measurements inthe plasma exhaust plume of a Hall-e:ect space propulsion system.Ph.D. Thesis, University of Michigan.

Laakso, H., Grard, R., Janhunen, P., Trotignon, J.-G., 2002. Plasma andwave phenomena induced by neutral gas releases in the solar wind.Ann. Geophys. 20 (1), 1–11.

Lyszyk, M., Klinger, E., et al., 1999. Quali5cation status of the PPS1350plasma thruster. Proceedings of the AIAA Joint Propulsion Conference,Los Angeles, CA, AIAA-99-2278.

Manzella, D.H., 1994. Stationary plasma thruster ion velocity distribution.Proceedings of the AIAA Joint Propulsion Conference, Indianapolis,IN, AIAA-94-3141.

Manzella, D.H., SPankovic, J.M., 1995. Hall thruster ion beamcharacterization. Proceedings of the AIAA Joint PropulsionConference, San Diego, CA, AIAA-95-2927.

Oh, D.Y., Hastings, D., 1996. Experimental veri5cation of a PIC-DSMCmodel for Hall thruster plumes. Proceedings of the AIAA JointPropulsion Conference, Lake Buenavista, FL, AIAA-96-3196.

Olsen, R.C., 1981. Modi5cation of spacecraft potentials by plasmaemission. J. Spacecraft 18 (5), 462–469.

Perot, C., Gascon, N., Bechu, et al., 1996. Characterization of a laboratoryHall thruster with electrical probes and comparisons with a 2Dhybrid PIC-MCC model. Proceedings of the AIAA Joint PropulsionConference, Los Angeles, CA, AIAA-99-2716.

Racca, G.D., Foing, B.H., Rathsman, P., 1999. An overviewon the status of the SMART-1 mission. Proceedings of theInternational Astronautical Congress, Amsterdam, Netherlands,IAA-99-IAA.11.2.09.

Rapp, D., Francis, W.E., 1962. Charge exchange between gaseous ionsand atoms. J. Chem. Phys. 37 (11), 2631–2645.

Rosenberg, D., Wehner, G.K., 1962. Sputtering yields for low energyHe+-, Kr+-, and Xe+-ion bombardment. J. Appl. Phys. 33 (5),1842–1845.

Tajmar, M., Wang, J., 2000. 3D numerical simulation of5eld-emission-electric-propulsion neutralization. J. Propulsion Power16 (3), 536–544.

Tajmar, M., RVudenauer, F., Fehringer, M., 1999. Back<ow contaminationof indium liquid-metal ion emitters (LMIE): numerical simulations.Proceedings of the International Electric Propulsion Conference,Kitakyushu, Japan, IEPC-99-070.

Tajmar, M., Gonzales, J., Hilgers, A., 2000. Modelling ofspacecraft-environment interactions on SMART-1. Proceedings of theAIAA Joint Propulsion Conference, Huntsville, AL, AIAA-2000-3526.

Valentian, D., Maslennikov, N., 1997. The PPS 1350 program.Proceedings of the International Electric Propulsion Conference,Cleveland, OH, IEPC-97-134.

VanGilder, D.B., Boyd, I., 1998. Particle simulations of the SPT-100plume. Proceedings of the AIAA Joint Propulsion Conference,Cleveland, OH, AIAA-98-3797.

VanGilder, D.B., Keidar, M., Boyd, I., 1999. Modeling Hall thrusterplumes using particle methods. Proceedings of the AIAA JointPropulsion Conference, Los Angeles, CA, AIAA-99-2294.

Wang, J., Brinza, D.E., Polk, J.E., Henry, M.D., 1999. Deep space oneinvestigations of ion propulsion plasma environment: overview andinitial results. Proceedings of the AIAA Joint Propulsion Conference,Los Angeles, CA, AIAA 99-2971.