Electromagnetic ion cyclotron resonance heating in the …have demonstrated single-pass electromagnetic ICRH in fast-ﬂowing laboratory plasma. 2. Experiment 2.1. The VASIMR engine

Available online at www.sciencedirect.com

www.elsevier.com/locate/asr

Advances in Space Research 42 (2008) 192–205

Electromagnetic ion cyclotron resonance heating in the VASIMR

E.A. Bering a,b,c,*, F.R. Chang-Dıaz c, J.P. Squire c, M. Brukardt a,c, T.W. Glover c,R.D. Bengtson d, V.T. Jacobson c, G.E. McCaskill c, L. Cassady a,c

a Physics Department, University of Houston, 617 Science and Research I, Houston, TX 77204-5005, USAb Department of Electrical and Computer Engineering, University of Houston, Houston, TX 77204, USA

c Ad Astra Rocket Company, Mail Code: ASPL, 2101 NASA Parkway, Houston, TX 77058, USAd Department of Physics, University of Texas at Austin, 1 University Station C1600, Austin, TX 78712-0264, USA

Received 1 November 2006; received in revised form 20 September 2007; accepted 23 September 2007

Abstract

Plasma physics has found an increasing range of practical industrial applications, including the development of electric spacecraftpropulsion systems. One of these systems, the Variable Specific Impulse Magnetoplasma Rocket (VASIMR) engine, both applies severalimportant physical processes occurring in the magnetosphere. These processes include the mechanisms involved in the ion accelerationand heating that occur in the Birkeland currents of an auroral arc system. Auroral current region processes that are simulated inVASIMR include lower hybrid heating, parallel electric field acceleration and ion cyclotron acceleration. This paper will focus on usinga physics demonstration model VASIMR to study ion cyclotron resonance heating (ICRH). The major purpose is to provide a VASIMRstatus report to the COSPAR community. The VASIMR uses a helicon antenna with up to 20 kW of power to generate plasma. Thisplasma is energized by an RF booster stage that uses left hand polarized slow mode waves launched from the high field side of the ioncyclotron resonance. The present setup for the booster uses 2–4 MHz waves with up to 20 kW of power. This process is similar to the ioncyclotron heating in tokamaks, but in the VASIMR the ions only pass through the resonance region once. The rapid absorption of ioncyclotron waves has been predicted in recent theoretical studies. These theoretical predictions have been supported with several indepen-dent measurements in this paper. The ICRH produced a substantial increase in ion velocity. Pitch angle distribution studies show thatthis increase takes place in the resonance region where the ion cyclotron frequency is equal to the frequency on the injected RF waves.Downstream of the resonance region the perpendicular velocity boost should be converted to axial flow velocity through the conserva-tion of the first adiabatic invariant as the magnetic field decreases in the exhaust region of the VASIMR. In deuterium plasma, 80% effi-cient absorption of 20 kW of ICRH input power has been achieved. No evidence for power limiting instabilities in the exhaust beam hasbeen observed.� 2007 COSPAR. Published by Elsevier Ltd. All rights reserved.

Keywords: Electric propulsion; Plasma rocket; ICRH; Auroral simulation

1. Introduction

1.1. Electric propulsion

The exploration of the solar system will be one of thedefining scientific tasks of the new century. One of the obvi-

0273-1177/$34.00 � 2007 COSPAR. Published by Elsevier Ltd. All rights rese

doi:10.1016/j.asr.2007.09.034

* Corresponding author. Address: Physics Department, University ofHouston, 617 Science and Research I, Houston, TX 77204-5005, USA.Tel.: +1 713 743 3543; fax: +1 713 743 3589.

E-mail address: [email protected] (E.A. Bering).

ous challenges faced by this enterprise is the scale size ofthe system under study, 1011–1014 m. Over distances on thisscale and given the performance of present day rockets, themission designer is faced with the choice of accepting multi-year or even decadal mission time lines, paying for enor-mous investment in rocket propellant compared to usefulpayload, or finding a way to improve the performance oftoday’s chemical rockets. For human space flight beyondEarth’s orbit, medical, psychological, and logistic consider-ations all dictate that drastic thruster improvement is the

rved.

mailto:[email protected]

E.A. Bering et al. / Advances in Space Research 42 (2008) 192–205 193

only choice that can be made. Even for robotic missionsbeyond Mars, mission time lines of years can be prohibitiveobstacles to success, meaning that improvements in deepspace sustainer engines are of importance to all phases ofsolar system exploration (Sankaran et al., 2003). Improve-ment in thruster performance can best be achieved by usingan external energy source to accelerate or heat the propel-lant (Chang-Dıaz, 2002; Bering et al., 2004b). This paperwill discuss an experimental investigation of the use ofion cyclotron resonance heating (ICRH) to provide anefficient method of electrode-less plasma acceleration inthe Variable Specific Impulse Magnetoplasma Rocket(VASIMR) engine.

Research on the VASIMR engine began in the late1970’s, as a spin-off from investigations on magnetic diver-tors for fusion technology (Chang-Dıaz and Fisher, 1982).A simplified schematic of the engine is shown in Fig. 1. TheVASIMR consists of three main sections: a helicon plasmasource, an ICRH plasma accelerator, and a magnetic noz-zle (Chang-Dıaz, 1997, 1998; Chang-Dıaz et al., 1999,2001, 2004; Bering et al., 2004b). Fig. 1 shows these threestages integrated with the necessary supporting systems.One key aspect of this concept is its electrode-less design,which makes it suitable for high power density and longcomponent life by reducing plasma erosion and othermaterials complications. The magnetic field ties the threestages together and, through the magnet assemblies, trans-mits the exhaust reaction forces that ultimately propel theship.

The plasma ions are accelerated in the second stage byion cyclotron resonance heating (ICRH), a well-knowntechnique, used extensively in magnetic confinement fusionresearch (Fisch, 1978; Golovato et al., 1988; Yasaka et al.,1988; Swanson, 1989; Stix, 1992). Owing to magnetic fieldlimitations on existing superconducting technology, the

Fig. 1. Simplified diag

system presently favors the light propellants; however,the helicon, as a stand-alone plasma generator can effi-ciently ionize heavier propellants such as Argon andXenon.

An important consideration involves the rapid absorp-tion of ion cyclotron waves by the high-speed plasma flow.This process differs from the ion cyclotron resonancemethod utilized in tokamak fusion plasmas as the particlesin VASIMR pass under the antenna only once (Beringet al., 2002a,b; Chavers and Chang-Dıaz, 2002; Chang-Dıaz et al., 2004). Sufficient ion cyclotron wave (ICW)absorption has nevertheless been predicted by recent theo-retical studies (Breizman and Arefiev, 2001), as well asobserved in VASIMR and previously reported in variousconferences and symposia.

Elimination of a magnetic bottle, a feature in the origi-nal VASIMR concept, was motivated by theoretical mod-eling of single-pass absorption of the ion cyclotron waveon a magnetic field gradient (Breizman and Arefiev,2001). While the cyclotron heating process in the confinedplasma of fusion experiments results in approximately ther-malized ion energy distributions, the non-linear absorptionof energy in the single-pass process results in a boost, ordisplacement of the ion kinetic energy distribution. Theions are ejected through the magnetic nozzle before thermalrelaxation occurs.

Natural processes in the auroral region may also exhibita related form of single-pass ICRH. ‘‘Ion conic’’ energeticion pitch angle distributions are frequently observed in theauroral regions of the Earth’s ionosphere and magneto-sphere (Shelley et al., 1976; Sharp et al., 1977, 1979; Mozeret al., 1977; Ghielmetti et al., 1978; Shelley, 1979; Collinet al., 1981). It is not relevant to list the entire range ofmodels that have been proposed to account for these obser-vations. Many models propose wave-driven transverse ion

ram of VASIMR.

194 E.A. Bering et al. / Advances in Space Research 42 (2008) 192–205

acceleration followed by adiabatic upwelling of the distri-bution (see Zintl et al., 1995; Andre et al., 1998; and refer-ences therein). Proposed driver wave modes include currentdriven electrostatic ion cyclotron (EIC) waves, (Kindel andKennel, 1971; Ungstrup et al., 1979; Ashour-Abdalla et al.,1981) and electromagnetic ion cyclotron waves (Changet al., 1986; Erlandson et al., 1988, 1994), among others.Other mechanisms proposed include interaction with anoblique double layer or dc potential structure (Borovsky,1984; McWilliams and Koslover, 1987). The fact that ionconics are commonly found on auroral field lines suggeststhat transverse ion acceleration is a ubiquitous process inauroral arcs (Andre et al., 1998). Space-borne observationsof narrow-band ion cyclotron waves with unambiguousspectral peaks near the ion cyclotron frequencies are rela-tively rare (Bering and Kelley, 1975a,b; Kelley et al.,1975; Kintner et al., 1978, 1979, 1991; Kintner, 1980; Ber-ing, 1983, 1984). Most studies have found that the mostcommon wave phenomenon found in association withtransverse ion acceleration is broad-band ELF noise (Bon-nell et al., 1996; Kintner et al., 1996, 2000a,b; Andre et al.,1998). All of these authors suggest that current driven EICwaves make up some or all of the broad-band ELF noise,but they are unable to prove it, even when wavelength mea-surements are available (Kintner et al., 2000a,b). The roleof inhomogeneities or shear in reducing the threshold forcurrent driven EIC instability is suggested as one solutionto this problem (Amatucci et al., 1998; Kintner et al.,2000b). EMIC waves appear to be associated with trans-verse ion acceleration �10% of the time (Erlandson et al.,1994; Andre et al., 1998).

In addition to the extensive body of work on the heatingof magnetic confinement fusion plasmas that was superfi-cially cited above, there is a 30 years body of theoreticaland laboratory work on transverse ion acceleration by cur-rent driven EIC modes (Kindel and Kennel, 1971; Correllet al., 1977; Cartier et al., 1986; McWilliams and Koslover,

Fig. 2. Physics o

1987; Zintl et al., 1995). All of these experiments have typ-ically used current driven EIC waves, parametric decay oflower hybrid waves, or other mode conversion process tolaunch the required wave field. Direct injection, which isused in VASIMR, requires good plasma coupling to theantenna in order to launch the waves with useful efficiency,as discussed below. Since the magnetospheric simulationexperiments have aimed at simulating EIC driven heatingand VASIMR uses EMIC waves, these prior results havelimited application to the VASIMR. What has been shownof relevance is that acceleration followed by adiabatic fold-ing is a viable mechanism for producing ion conics (Cartieret al., 1986; Zintl et al., 1995). However, the field ratiosattained were an order of magnitude smaller that used inthe VASIMR studies reported here.

VASIMR has a transverse ion acceleration stage orbooster that uses EMIC waves, followed by adiabaticexpansion. Simultaneous ambipolar acceleration is alsoobserved in the VASIMR exhaust plume that may be inter-preted as a large-scale double layer (Fruchtman, 2005).Thus, VASIMR results may be of interest to proponentsof more than one model of ion conic production. Thispaper presents the data from a series of experiments thathave demonstrated single-pass electromagnetic ICRH infast-flowing laboratory plasma.

2. Experiment

2.1. The VASIMR engine

The VASIMR engine has three major subsystems, theplasma generator stage, the RF ‘‘booster’’ stage and thenozzle, shown in Figs. 1 and 2 (Chang-Dıaz et al., 2001).A laboratory physics demonstrator experiment (VX-50),shown in Figs. 3 and 4, has been under development andtest at the NASA Johnson Space Center, for several years(Squire et al., 2000; Petro et al., 2002). The details of the

f VASIMR.

Fig. 3. VX-50 at JSC.

Fig. 4. VX-50 configuration.


engine and its design principles have been previouslyreported (Chang-Dıaz, 2001; Chang-Dıaz et al., 2004).The first stage is a helicon discharge that has been opti-mized for maximum power efficiency (lowest ionizationcost in eV/(electron-ion pair)) (Boswell, 1984; Chen,1991; Boswell and Chen, 1997; Boswell and Charles,2003). The next stage downstream is the heating system.Energy is fed to the system in the form of a circularly polar-ized rf signal tuned to the ion cyclotron frequency. ICRHheating has been chosen because it transfers energy directlyand largely to the ions, which maximizes the efficiency ofthe engine (Golovato et al., 1988; Yasaka et al., 1988). Inthe present small-scale test version, there is no mirrorchamber and the ions make one pass through the ICRH

antenna. The system also features a two-stage magneticnozzle, which accelerates the plasma particles by convert-ing their azimuthal energy into directed momentum. Thedetachment of the plume from the field takes place mainlyby the loss of adiabaticity and the rapid increase of thelocal plasma b, defined as the local ratio of the plasma pres-sure to the magnetic pressure.

The main VASIMR VX-50 vacuum chamber was a cyl-inder 1.8 m long and 35.6 cm in diameter. The VASIMRVX-50 exhaust flowed through a conical adapter sectioninto a 5 m3 exhaust reservoir. The magnetic field was gen-erated by four liquid nitrogen cooled 150 turn copper mag-nets, which could generate a magnetic induction of up to1.5 T. The high vacuum pumping system consisted of a


cryopump and two diffusion pumps with a combined totalcapacity of 5000 l/s.

Over the course of the investigations reported here, therehave been a series of improvements and upgrades to the hel-icon plasma source (Glover et al., 2004; Bering et al., 2005).The earliest experiments that will be presented were per-formed using a 5 cm diameter ‘‘Boswell’’ type double saddleantenna and 3 kW of rf power. Directionality was providedby use of a magnetic cusp configuration. Over the 3.5 yearsinterval spanned by these experiments, the helicon hasundergone a number of successive upgrades. The heliconplasma source in the VASIMR engine has been incremen-tally improved in three steps. First, antenna size, connectorand power supply improvements raised the available powerfrom 3 to 10 kW, still operating with Boswell antenna and amagnetic cusp. The factors limiting the application of rfpower to the helicon ionizing discharge had been the powerof the 25 MHz transmitter (3 kW), the voltage limits on thevacuum rf feed-through, the voltage limits of the rf matchingnetwork and the diameter of the discharge. Successive stepsto 9 cm increased the diameter of the helicon antenna. Thehelicon rf supply was completely rebuilt. A new high voltagepower supply was built, along with a new high power trans-mitter. A 10 kV rf vacuum feed-through was obtained andinstalled, along with a new high voltage matching network.These new components produced helium plasma operatingat 10 kW, with�4 times the ion flux achievable at 3 kW. Sec-ond, the transmitting antenna was changed from the Boswellconfiguration to a helical half-twist antenna. Finally, thedecision was taken to switch transmitters and power the hel-icon with our 100 kW transmitter, which can operate the hel-icon at 13.5 MHz. A high voltage (30 kV) water-cooledvacuum feed-through was manufactured to enable fullpower operation. In all configurations, the vacuum feed-through and helicon antennas were water cooled and capa-ble of steady state operation.

The ICRH antenna is a helical, double strap quarter-turnantenna configuration. It is polarized to launch left-handedslow mode waves (Bering et al., 2002b; Chang-Dıaz et al.,2003). The present configuration of the rf booster or ICRHsystem uses 1.5–3.0 MHz left hand polarized slow modewaves launched from the high field, over dense side of the res-onance (region 13 of the Clemmow-Mullaly-Allis (CMA)diagram, using the Stix (1992) notation). The antenna wascylindrically symmetric and concentric to the plasma andthe axis of the confining magnets. It was designed to launchelectromagnetic ion cyclotron waves along this axis, withkk � k?. The efficiency issues of this direct launching tech-nique (Zintl et al., 1995) are discussed in Section 2.3 below.The antenna was uncooled, which limits the pulse length totypically less than 0.5 s. The pulse power was limited bythe rf power source to 1.5 kW until 2006.

2.2. Diagnostics

Available plasma diagnostics include a triple probe, 32and 70 GHz density interferometers, a bolometer, a televi-

sion monitor, an H-a photometer, a spectrometer, neutralgas pressure and flow measurements, several griddedenergy analyzers (retarding potential analyzer or RPA)(Krassovsky, 1959; Whipple, 1959; Hanson and McKibbin,1961; Parker and Whipple, 1970; Hanson et al., 1970, 1981;Minami and Takeya, 1982; Chang-Dıaz et al., 2001; Beringet al., 2004a), a momentum flux probe, an emission probe,a directional, steerable RPA and other diagnostics (Squireet al., 1997). Reciprocating Langmuir probes (Chen andSekiguchi, 1965) and density interferometer are the pri-mary plasma diagnostics. The Langmuir probe measuresion current and temperature profiles and is calibrated bythe density interferometer. An array of thermocouples pro-vides a temperature map of the system. The Langmuirprobe has four molybdenum tips that are biased as a tripleprobe, with an extra tip for measuring electrostatic fluctu-ations (Chen and Sekiguchi, 1965).

2.2.1. Retarding potential analyzer (RPA)

Retarding potential analyzer (RPA) diagnostics havebeen installed to measure the accelerated ions. The presentRPA is a planar ion trap located �40 cm downstream fromthe plane of the triple and Mach probes, which correspondsto a factor of 8 reduction in the magnetic field strength.The grids are 49.2-wire/cm nickel mesh, spaced 1 mm apartwith Macor spacers. The opening aperture is 1 cm in diam-eter, usually centered on the plasma beam. A four-grid con-figuration is used, with entrance attenuator/collimator,electron suppressor, ion analyzer and secondary suppressorgrids. The entrance attenuator/collimator consists of a1 mm thick nickel plate drilled with a uniform rectangulargrid of.35 mm diameter holes spaced to give 1%transparency.

The interpretation of RPA output data in terms of ionenergy requires an accurate knowledge of plasma potential(Vp). When available, data from an rf compensated sweptLangmuir probe provided by Los Alamos National Labo-ratory (LANL) are used to determine Vp. When other Vp

data are not available, plasma potential is assumed to bethe value at which dI/dV first significantly exceeds 0 withICRH off, which usually agrees with the LANL probevalue within the error bars (±5 V). This value is typically�+30 to 50 V with respect to chamber ground. The opera-tor biases the body and entrance aperture of the RPA tothis value. The ion exhaust parameters are deduced fromthe raw data by means of least squares fits of drifting Max-wellians to the current–voltage data (Whipple, 1959; Par-ker and Whipple, 1970; Minami and Takeya, 1982;Hutchinson, 1987).

In this paper, the RPA data will be presented in severalformats, including the voltage derivative of the I–V charac-teristic, the one-D ion velocity distribution function, a pla-nar cut through the full ion velocity distribution function,and as derived parameters. The dI/dV plots (e.g. Fig. 7)show the smoothed, numerically calculated derivative withrespect to sweep voltage of the measured RPA current.Sweep voltage zero is set to plasma potential found using

Fig. 5. Circuit diagram of the ICRH antenna, indicating how the plasmaimpedance couples to the circuit.


the methods of the previous paragraph, for ICRH off con-ditions and all other parameters unchanged. Unless statedotherwise, 16 sweeps per shot of the RPA have been aver-aged to produce each figure. The presence of features in thedI/dV curves at retarding voltages less than the plasmapotential are the result of temporal fluctuations in the ionsaturation current, and largely serve to illustrate the risksin taking numerical derivatives of data. The ion velocitydistribution functions (e.g. Fig. 8) were found from thedI/dV curves by dividing by the energy and multiplyingby a calibration factor.

The RPA I–V characteristic data has been reduced byleast squares fitting the characteristic that would be pro-duced be a drifting Maxwellian to the data. This fit hasthree parameters. The three free parameters in these fitsare ion density, mean drift speed and the ion temperaturein the frame of reference moving with the beam. The tem-perature is found from these least squares fits, not fromtaking the slope of the logarithm of the data. The densityis calibrated by comparison with nearby Langmuir probesand is probably best understood as a relative measurement.The temperature and ion drift speed parameters dependmost strongly on the accuracy with which the retardingpotential is known. The absolute uncertainty of the sweepvoltage digitization with respect to chamber ground wasa few percent when digitizer calibration uncertainty, sweepisolator reduction ratio precision and related parametersare folded in. There are systematic uncertainties associatedwith the determinations of plasma potential, which are dis-cussed two paragraphs above. Plasma potential as deter-mined above is always subtracted prior to any otheranalysis.

The full ion velocity phase space distribution function ofthe ions can be obtained by scanning the RPA in pitchangle between otherwise identical shots, assuming cylindri-cal (gyrotropic) symmetry (Stenzel et al., 1982, 1983). Theangle step size was 10� for Figs. 10 and 11, and 5� from 0�to 60� and 10� thereafter for Figs. 13, 14, 17, and 18.

2.3. Plasma loading

The light ion VASIMR engine is best suited to highpower operation. The critical factor that limits efficiencyat low power is the efficiency of helicon stage and hencethe ionization cost (Carter et al., 2002). At high power,the ICRH stage is more important and hence ICRHantenna loading is more critical here. The need to obtaingood antenna coupling in order to achieve efficient directlaunching of ion cyclotron wave (ICW) has been noted inthe literature (Zintl et al., 1995). The load impedance ofthe antenna must be significantly larger than the impedanceof the transmission and matching network (see Fig. 5). Thecyclotron resonance frequency of helium in the availablemagnetic field is about 2 MHz, which means that onerequires plasma that is �8–10 cm in diameter before rea-sonable load impedances are obtained. The cost of produc-ing dense plasma of this diameter and thereby achieving

good plasma loading by an ICRH antenna has been oneof the reasons why most laboratory studies of auroraltransverse ion acceleration by electrostatic ion cyclotronwaves have used current driven instabilities or parametricdecay of lower hybrid waves to launch ion cyclotron wavefields used in their studies (Correll et al., 1977; Cartieret al., 1986; McWilliams and Koslover, 1987; Zintl et al.,1995; Amatucci et al., 1998).

The following calculation illustrates how plasma loadingis determined. The analysis begins with the relationshipthat:

VSWRplasma

VSWRvacuum

¼ Rp þ Rc

Rc

; ð1Þ

where VSWR is the Voltage Standing Wave Ratio. A net-work analyzer is used in place of the high power rf trans-mitter to measure the quality factor (Qc) of the antennacoupling circuit and to tune the circuit when no plasma ispresent. With measuring the impedance matching (Lm)and the antenna inductances (Lm and LA, respectively),we have:

Rc ¼xðLm þ LAÞ

Qc

� 0:24 X ð2Þ

and

Rp ¼ RcðVSWRplasma � 1Þ: ð3Þ

Thus, the coupling efficiency is:

gA ¼Rp

ðRp þ RcÞ¼ 0:89 ð4Þ

and the power radiated into the plasma is estimated as:

P plasma ¼ gAP ICRH ¼ 1:25 kW: ð5Þ

The initial deuterium loading data were taken when only4 kW of helicon power was available. Plasma loading wasgenerally maximum when f/fci > 1. The plasma loadingwas comparable to the circuit resistance. Helium datashowed similar behavior. More recent loading measure-ments have been made using a 20 kW helicon discharge.The measurements show very good loading, �2 X, whichimplies a corresponding high coupling efficiency, as shownin Fig. 6. This high loading impedance was obtained with asizable, �1 cm, plasma-antenna gap. Maximum in the

Fig. 6. Time history of the plasma loading of the ICRH antenna andantenna efficiency.

Fig. 7. Comparison of ion energy distributions; ICRH-on vs. -off, with8.3 kW of helicon power and 1.5 kW of ICRH, using 70 sccm ofdeuterium.

Fig. 8. Comparison of ion velocity distributions; ICRH-on vs. -off, with8.3 kW of helicon power and 1.5 kW of ICRH, using 70 sccm ofdeuterium.


loading occurred at f/fci = 0.95, indicating that the antennawas launching a more propagating wave than it was in thelow-density case.

3. Ion cyclotron resonance heating (ICRH)

3.1. Low density discharge, 1.5 kW ICRH

Until 2006, the RF booster or ICRH system used1.5 kW of 1.85 MHz left hand polarized slow mode waveslaunched from the high field, overdense side of the reso-nance (region 13 of the CMA diagram). An important con-sideration involves the rapid absorption of ion cyclotronwaves by the high-speed plasma flow. This process differsfrom the familiar ion cyclotron resonance utilized in toka-mak fusion plasmas as the particles in VASIMR pass underthe antenna only once. Sufficient ICW absorption has nev-ertheless been predicted by recent theoretical studies. Theexperiments presented below have supported these theoret-ical predictions with a number of independentmeasurements.

The effect of the ICRH heating is expected to be anincrease in the component of the ion velocity perpendicularto the magnetic field. This increase will take place in theresonance region, i.e. the location where the injected RFwave frequency is equal to the ion cyclotron frequency.Downstream of the resonance, this perpendicular heatingwill be converted into axial flow owing to the requirementthat the first adiabatic invariant of the particle motion beconserved as the magnetic field decreases (Figs. 7 and 8).Since the total ion flux is not expected to increase, this axialacceleration should be accompanied by a density decrease,as shown in Fig. 9. Furthermore, the particles should havea pitch angle distribution that does not peak at 0�. Instead,the distribution should peak at an angle that maps to a per-pendicular pitch angle at resonance (Fig. 10).

An example of the distribution functions mapped backto the resonance that have been observed during ICRHexperiments is shown in Fig. 11. During this particularexperiment, the ICRH transmitter was operated at1.85 MHz, at a power level of 1500 W, using helium. Thedata show a pronounced enhancement or jet of ions withv > 60 km/s at 90� mapped pitch angle. The data also showa depletion of ions with lower velocities.

3.2. Higher density discharge, 1.5 kW ICRH

The most recent improvements in the helicon dischargehave both increased the available power to 20 kW andreduced the energy cost per ion to �200 eV per ion pair.These improvements have combined to give nearly anorder of magnitude increase in total ion flux. Fig. 12 com-pares the ion energy distributions detected by the RPA in

Fig. 9. Time series of ion data taken during ICRH on shot, with 8.3 kW ofhelicon power and 1.5 kW of ICRH, using 70 sccm of deuterium.

Fig. 10. Ion velocity distribution function, ICRH-on, with 1.5 kW ofICRH using helium. The color bar legend reads ‘‘Log DistributionFunction, # s�3 m�-6’’. (For interpretation of the references in color in thisfigure legend, the reader is referred to the web version of this article.)


a 20 kW deuterium discharge showing the effect of 1.5 kWof ICRH. The figure shows clear evidence of ion acceler-ation by a mere 1.5 kW of ICRH. The observed energyincrease was 17 eV/ion, yielding plasma with more thantwice the bulk kinetic energy obtainable with the heliconalone.

Two dimensional color contour plots of a planar cutthrough the distribution function are shown in Fig. 13for ICRH-off and Fig. 14 for ICRH-on. There is a clear,visible difference between the two distributions, indicatingthat the ICRH has accelerated the entire plasma. Thereduction in the maximum value of the distribution func-tion in the ICRH-on case corresponds to a 0.55 densitydrop. This decrease is consistent with the results ofsimultaneous interferometer observations.

3.3. Higher density discharge, 20 kW ICRH

The final step in the progressive improvement processthat will be described in this paper was an order of magni-tude increase in the available ICRH power, up to 20 kW.The effect of this increase in ICRH power level was dra-matic and clear. Fig. 15 shows the derivative with respectto voltage of the current–voltage RPA data from threeshots with different ICRH power levels, 0, 8.3 and20 kW. The derivative is proportional to the ion energy dis-tribution and shows that 20 kW of ICRH adds approxi-mately 200 eV to the mean energy of the deuterium ionsin this configuration. The ion velocity phase space distribu-tion functions corresponding to Fig. 15 are shown inFig. 16. The ICRH-off distribution is an elegant exampleof a cold, Maxwellian ion distribution. The two ICRH-on distributions show that the plasma was drifting fasterand was hotter in the rest frame of the beam. This resultis somewhat in contrast to the 1.5 kW results, where thebeam remained cold in the beam rest frame. The 20 kWcurve in particular shows the presence of multiple Maxwel-lians with different drift speeds and temperatures. We takethis lower energy ‘‘filled in’’ character of the distributionfunction to be a consequence of inadequate vacuum pumpcapacity and high background neutral density. It shouldalso be noted that the RPA voltage sweep did not extendto high enough voltage (360 V) to stop the ion current com-pletely. The highest energy tail of the ion distributionextended to >350 eV energy.

The pitch angle distribution of the accelerated ions wasinvestigated using the same method previously used to pro-duce Fig. 13. The ICRH power was limited to 14 kW sothat the available sweep voltage stopped the entire popula-tion. The two dimensional ion velocity phase space distri-butions measured with 14 kW of ICRH and no ICRHare shown in Figs. 17 and 18. The presence of a high veloc-

Fig. 11. Difference in velocity distribution function between ICRH-on and ICRH-off conditions at resonance, with 1.5 kW of ICRH using helium. Thecolor bar legend reads ‘‘Difference in Distribution Function, s�3 m�-6’’. (For interpretation of the references in color in this figure legend, the reader isreferred to the web version of this article.)

Fig. 12. Comparison of ion energy distributions; ICRH-on vs. -off, with20 kW of helicon power and 1.5 kW of ICRH, using 400 sccm ofdeuterium.

Fig. 13. Ion velocity phase space distribution function, ICRH-on, with20 kW of helicon power and 1.5 kW of ICRH, using 400 sccm ofdeuterium. The color bar legend reads ‘‘Log Distribution Function, #s�3 m�-6’’. (For interpretation of the references in color in this figurelegend, the reader is referred to the web version of this article.)


ity, somewhat conical jet of plasma is clearly evident inFig. 17. As above, 90� pitch angle at the location of theICRH antenna maps to 10� pitch angle at the location ofthe RPA, which means that all the high velocity ionsobserved at greater pitch angles in Fig. 17 are indicativeof both the somewhat coarse collimation of the RPA andthe presence of ion-neutral scattering.

Typically, reduced RPA data figures from this experi-ment present all three of these parameters, plotted eitheras functions of time within a single shot (e.g. Fig. 7), oras functions of other parameters varied on a shot-to-shotbasis. Fig. 19 shows another example of a time series plot,showing the effect of a 100 ms burst of 14 kW of ICRHpower on deuterium plasma. From top to bottom, the

figure shows ion temperature in the frame of the beam,ion density, and ion drift speed. During the ICRH burst,the ion temperature increased from �1 to 35 eV, the den-sity dropped by a factor of 3 and the drift speed increasedby a factor of 4. The apparent increase in total flux seemsto indicate that the ICRH is acting to increase the ioniza-tion of the gas, which indicates that the helicon was notoptimized for this shot. The temperature shown is effec-tively the parallel temperature of the ions in the frame mov-ing with the beam. The parallel ion temperature increasewas much larger than the increases seen in previous work(Zintl et al., 1995). The apparent waste of energy in addingto the parallel heat content of the beam is not as severe as it

Fig. 15. Comparison of ion energy distributions for 0, 8.3, and 20 kW ofICRH power in deuterium plasma, input flow rate 450 sccm, and 20 kWhelicon power.

Fig. 16. Comparison of ion velocity distributions for the same three shotsas in Fig. 15.

Fig. 17. Ion velocity phase space distribution function, 14 kW ICRH-on,in deuterium plasma, input flow rate 450 sccm, and 20 kW helicon power.

Fig. 14. Ion velocity phase space distribution function, ICRH-off, with20 kW of helicon power and 1.5 kW of ICRH, using 400 sccm ofdeuterium. The color bar legend reads ‘‘Log Distribution Function, #s�3 m�-6’’. (For interpretation of the references in color in this figurelegend, the reader is referred to the web version of this article.)

Fig. 18. Ion velocity phase space distribution function, ICRH-off indeuterium plasma, input flow rate 450 sccm, and 20 kW helicon power.


might seem at first glance. At least two processes areresponsible. First, there is a relatively large amount ofion neutral scattering owing to the pumping problemalluded to above. Second, the ICRH process has a disper-sive component that depends on the incident gyrophase ofthe ions relative to the ICRH waveform. The amount ofenergy that an ion obtains in the resonance region depends,in part on this phase angle. This energy spread appears tothe least squares fit as an apparent parallel temperature.However, it does not represent waste heat content of thebeam. Finally, at high input power, non-linear broadeningof the resonance (Dum and Dupree, 1970) and changes ofthe phase of the ion orbits will act to increase the parallelheating rate of the ions. From a practical standpoint, thesmearing out of the resonance will act to allow a largerfraction of the ions to absorb energy from the wave field(Correll et al., 1977), which is a very desirable outcome.

Fig. 19. Time series of fit parameters obtained by least squares fitting adrifting Maxwellian representation to the raw RPA data during ICRHdeuterium plasma shots. From bottom to top, the panels show ion driftvelocity, ion density and ion temperature in the frame of the beam.

Fig. 20. A radial scan of the RPA, showing fit parameters obtained byleast squares fitting a drifting Maxwellian representation to the raw RPAdata during ICRH deuterium plasma shots as functions of radius from thecenter of the beam. From bottom to top, the panels show ion driftvelocity, ion density and ion temperature in the frame of the beam.


Fig. 20 shows the results from another day, scanning theRPA in radius with 14 kW of ICRH power applied. Thedata indicate that the plasma column had a roughly con-stant temperature and density out to a radius of �9 cm.The velocity curve shows a generally parabolic profile witha dip in the center. The dip in the center may indicate thatthe ICRH signal was fully absorbed prior to penetrating allthe way to the center of the plasma.

Fig. 21 shows the comparison of two scans of ICRHpower for two different gas inlet flow rates, 450 standardcubic centimeters per minute (sccm), which is plotted inred and appears as the top curve in the middle paneland the bottom curve in the other two, and 250 sccm,which is plotted in blue and is the other curve in eachpanel. The helicon power was reduced to 17 kW in thelower flow rate shot series. The purpose of this set ofexperiments was to explicitly demonstrate the variablespecific impulse capability of the VASIMR. The lower

flow rate shots produced an output that was �20% fas-ter, had a much higher parallel temperature than thehigh flow shots and were lower in density. These datacan be used to estimate the thrust and output power,by integrating across the plume profile shown inFig. 20 for each of the shots shown in Fig. 21. Theresults are shown in Fig. 22. The output power is virtu-ally the same in each case, whereas the lower flow pro-duces lower thrust and higher specific impulse, which isas expected, and which verifies the fundamental constantpower throttling premise of the VASIMR concept.

4. Conclusions

Substantial progress has been made in the develop-ment of the VASIMR engine. In the past 2 years, theoperating power level of the helicon discharge has beenincreased by nearly an order of magnitude, to 20–

Fig. 21. Scans of ICRH power for two different flow rates and heliconpower levels, showing fit parameters obtained by least squares fitting adrifting Maxwellian representation to the raw RPA data during ICRHdeuterium plasma shots as functions of ICRH power. From bottom totop, the panels show ion drift velocity, ion density and ion temperature inthe frame of the beam. The meaning of the two separate curves is discussedin the text.

Fig. 22. Scans of ICRH power for two different flow rates and heliconpower levels, showing parameters inferred from the fit parameters shownin the previous figure as functions of ICRH power. From bottom to top,the panels show output power, thrust and specific impulse. The meaning ofthe two separate curves is discussed in the text.


24 kW. The energy cost per ion pair has been reduced to�200 eV. Plasma loading of the ICRH antenna increaseswith increasing total ion flux and plasma density. Usingpresent maximum helicon discharge output, loadingimpedance in excess of 2 X is being obtained, which cor-responds to gA > 90%. Single-pass ICRH has been dem-onstrated repeatedly in a variety of configurations. Highpower (20 kW) ICRH has been achieved. gB > 80% hasbeen obtained with deuterium. Increases in the parallelion temperature observed at high ICRH power may beevidence for non-linear broadening of the absorption res-onance. No macroscopic power limiting beam instabili-ties were observed. Auroral process simulationexperiments are now possible in the VASIMR. The teaminvites interested modelers to contact us regarding possi-ble collaborative proposals.

Acknowledgments

NASA Johnson Space Center under Grant NAG 9-1524, the Texas Higher Education Coordinating Board un-der Advanced Technology Program project 003652-0464-1999 and the Ad Astra Rocket Company sponsored thisresearch.

References

Amatucci, W.E., Walker, D.N., Ganguli, G., Duncan, D., Antoniades, J.,Bowles, J.H., Gavrishchaka, V., Koepke, M.E. Velocity-shear-drivenion-cyclotron waves and associated transverse ion heating. Journal ofGeophysical Research 103 (A6), 11711–11724, 1998.

Andre, M., Norqvist, P., Andersson, L., Eliasson, L., Eriksson, A.L.,Blomberg, L., Erlandson, R., Waldemark, J. Ion energization mech-anisms at 1700 in the auroral region. Journal of Geophysical Research103 (A3), 4199–4222, 1998.


Ashour-Abdalla, M., Okuda, H., Cheng, C. Acceleration of heavy ions onauroral field lines. Geophysical Research Letters 8, 795–798, 1981.

Bering III, E.A. Apparent electrostatic ion cyclotron waves in the diffuseaurora. Geophysical Research Letters 10, 647–650, 1983.

Bering III, E.A. The plasma wave environment of an auroral arc, 1.,Electrostatic ion cyclotron waves in the diffuse aurora. Journal ofGeophysical Research 89, 1635–1649, 1984.

Bering III, E.A., Kelley, M.C. Observation of electrostatic ion cyclotronwaves at the boundary of an auroral arc. EOS, Transactions,American Geophysical Union 56, 173, 1975a.

Bering III, E.A., Kelley, M.C., Mozer, F.S. Observations of an intensefield aligned thermal ion flow and associated intense narrow bandelectric field oscillations. Journal of Geophysical Research 80 (34),4612–4620, 1975b.

Bering III, E.A., Brukardt, M., Chang-Dıaz, F.R., et al. Experimentalstudies of the exhaust plasma of the VASIMR engine. Presented at the40th AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV,2002a.

Bering III, E.A., Brukardt, M.S., Rodriguez W.A., et al. Ion dynamics andICRH heating in the exhaust plasma of the VASIMR engine.Presented at the 53rd International Astronautical Congress/The WorldSpace Congress, 10–19 Oct 2002, Houston, Texas, 2002b.

Bering III, E.A., Chang-Dıaz, F.R., Squire, J.P., et al. Velocity phasespace studies of ion dynamics in the VASIMR engine. Presented at the42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV,2004a.

Bering III, E.A., Chang-Dıaz, F.R., Squire, J.P. The use of RF waves inspace propulsion systems. Bulletin of Radio Science 310, 92–106,2004b.

Bering III, E.A., Squire, J.P., McCaskill, G., et al. Progress towardthe development of a 50 kW VASIMR engine. Presented at the43rd AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV,2005.

Bonnell, J., Kintner, P., Wahlund, J.E., Lynch, K., Arnoldy, R.Interferometric determination of broadband ELF wave phase velocitywith a region of transverse auroral ion acceleration. GeophysicalResearch Letters 23 (23), 3297–3300, 1996.

Borovsky, J.E. The production of ion conics by oblique double layers.Journal of Geophysical Research 89 (A4), 2251–2266, 1984.

Boswell, R.W. Very efficient plasma generation by whistler waves near thelower hybrid frequency. Plasma Physics and Controlled Fusion 26,1147, 1984.

Boswell, R.W., Charles, C. The helicon double layer thruster. Presented atthe 28th International Electric Propulsion Conference, IEPC 2003,Toulouse, France, 2003.

Boswell, R.W., Chen, F.F. Helicons – the early years. IEEE Transactionson Plasma Science 25 (6), 1229–1244, 1997.

Breizman, B.N., Arefiev, A.V. Single-pass ion cyclotron resonanceabsorption. Physics of Plasmas 8 (3), 907–915, 2001.

Carter, M.D., Baity III, F.W., Barber, G.C., et al. Comparing experi-ments with modeling for light ion helicon plasma sources. Physics ofPlasmas 9 (12), 5097–5110, 2002.

Cartier, S., D’Angelo, N., Merlino, R. A laboratory study of ionenergization by EIC waves and subsequent upstreaming along diverg-ing magnetic field line. Journal of Geophysical Research 91 (A7),8025–8033, 1986.

Chang-Dıaz, F.R. Research status of the variable specific impulsemagnetoplasma rocket. Presented at the 39th Annual Meeting of theDivision of Plasma Physics, 1997.

Chang-Dıaz, F.R. Research status of the variable specific impulsemagnetoplasma rocket. Presented at Open Systems, July 27–31,1998, Novosibirsk, Russia, 1998.

Chang-Dıaz, F.R. An overview of the VASIMR engine: high power spacepropulsion with RF plasma generation and heating. Presented at the14th Topical Conference on Radio Frequency Power in Plasmas, May7–9, 2001, Oxnard, CA, 2001.

Chang-Dıaz, F.R. Fast, power-rich space transportation, key to humanspace exploration and survival. Presented at the 53rd International

Astronautical Congress/The World Space Congress, 10–19 Oct 2002,Houston, Texas, 2002.

Chang-Dıaz, F.R., Fisher, J.L. A supersonic gas target for a bundledivertor plasma. Nuclear Fusion 22 (8), 1982.

Chang-Dıaz, F.R., Squire, J.P., Ilin, A.V., et al. The development of theVASIMR engine. Presented at the International Conference ofElectromagnetics in Advanced Space Applications, Sept. 13–17,1999, Torino, Italy, 1999.

Chang-Dıaz, F.R., Squire, J.P., Bering III, E.A., et al. The VASIMRengine approach to solar system exploration. Presented at the 39thAIAA Aerospace Sciences Meeting and Exhibit, Jan. 8–11, 2001,Reno, NV, 2001.

Chang-Dıaz, F.R., Squire, J.P., Ilin, A.V., et al., Early results of ICRHexperiments in VX-10. Bulletin of the American Physical Society,DPP03, RP1.138, 2003.

Chang-Dıaz, F.R., Squire, J.P., Glover, T., et al. The VASIMRengine: project status and recent accomplishments. Presented atthe 42nd Aerospace Sciences Meeting and Exhibit, Reno, NV,2004.

Chang, T., Crew, G.B., Hershkowitz, N., Jasperse, J.R., Retterer,J.M., Winningham, J.D. Transverse acceleration of oxygen ions byelectromagnetic ion cyclotron resonance with broad band left-handpolarized waves. Geophysical Research Letters 13 (7), 636–639,1986.

Chavers, D.G., Chang-Dıaz, F.R. Momentum flux measuring instrumentfor neutral and charged particle flows. Review of Scientific Instruments73 (10), 3500–3507, 2002.

Chen, F.F. Plasma ionization by helicon waves. Plasma Physics andControlled Fusion 33 (4), 339–364, 1991.

Chen, S.-L., Sekiguchi, T. Instantaneous direct-display system of plasmaparameters by means of triple probe. Journal of Applied Physics 36,2363, 1965.

Collin, H.L., Sharp, R.D., Shelley, E.G., et al. Some general character-istics of upflowing ion beams over the auroral zone and theirrelationship to auroral electrons. Journal of Geophysical Research86, 6820, 1981.

Correll, D.L., Boehmer, H., Rynn, N., Stern, R.A. Temporal evolution ofion temperatures in the presence of ion cyclotron instabilities. Physicsof Fluids 20 (5), 822–828, 1977.

Dum, C.T., Dupree, T.H. Nonlinear stabilization of high-frequencyinstabilities in a magnetic field. Physics of Fluids 13 (8), 2064–2081,1970.

Erlandson, R.E., Zanetti, L.J., Potemra, T.A., Andre, M., Matson, L.Observation of electromagnetic ion cyclotron waves and hot plasma inthe polar cusp. Geophysical Research Letters 15, 421, 1988.

Erlandson, R.E., Zanetti, L.J., Acuna, M.H., Eriksson, A.I., Eliasson, L.,Boehm, M., Blomberg, L.G. Freja observations of electromagnetic ioncyclotron ELF waves and transverse oxygen ion acceleration onauroral field lines. Geophysical Research Letters 21 (17), 1855–1858,1994.

Fisch, N.J. Confining a tokamak plasma with rf-driven currents. PhysicalReview Letters 41 (13), 873, 1978.

Fruchtman, A. Double layer and thrust. 47th Annual Meeting of theDivision of Plasma Physics, Bulletin of the American Physical Society,50, pp. 109, Denver, Colorado, 2005.

Ghielmetti, A.G., Johnson, R.G., Sharp, R.D., et al. The latitudinal,diurnal and altitudinal distributions of upward flowing energetic ionsof ionospheric origin. Geophysical Research Letters 5, 59, 1978.

Glover, T.W., Chang-Dıaz, F.R., Jacobson, V.T., et al. Ion cyclotronheating results in the VASIMR VX-10. Presented at the 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, FortLauderdale, FL, 2004.

Golovato, S.N., Brau, K., Casey, J., et al. Plasma production and heatingin a tandem mirror central cell by radio frequency waves in the ioncyclotron frequency range. Physics of Fluids 31 (12), 3744–3753, 1988.

Hanson, W.B., McKibbin, D.D. An ion-trap measurement of the ion-concentration profile above the F2 peak. Journal of GeophysicalResearch 66, 1667, 1961.


Hanson, W.B., Sanatani, S., Zuccaro, D., et al. Plasma measurementswith the retarding potential analyzer on OGO 6. Journal of Geophys-ical Research 75 (28), 5483–5501, 1970.

Hanson, W.B., Heelis, R.A., Power, R.A., et al. The retarding potentialanalyzer for Dynamics Explorer-B. Space Science Instruments 5, 503–510, 1981.

Hutchinson, I.H. Principles of Plasma Diagnostics. Cambridge UniversityPress, Cambridge, 1987.

Kelley, M.C., Bering III, E.A., Mozer, F.S. Evidence that the electrostaticion cyclotron instability is saturated by ion heating. Physics of Fluids18, 1590–1597, 1975.

Kindel, J.M., Kennel, C.F. Topside current instabilities. Journal ofGeophysical Research 76, 3055, 1971.

Kintner, P.M. On the distinction between electrostatic ion cyclotron wavesand ion cyclotron harmonic waves. Geophysical Research Letters 7,585, 1980.

Kintner, P.M., Bonnell, J., Arnoldy, R., Lynch, K., Pollock, C., Moore, T.SCIFER – transverse ion acceleration and plasma waves. GeophysicalResearch Letters 23, 1873–1876, 1996.

Kintner, P.M., Kelley, M.C., Mozer, F.S. Electrostatic hydrogen cyclo-tron waves near one earth radius altitude in the polar magnetosphere.Geophysical Research Letters 5, 139, 1978.

Kintner, P.M., Kelley, M.C., Sharp, R.D., et al. Simultaneous observa-tions of energetic (keV) upstreaming ions and electrostatic hydrogencyclotron waves. Journal of Geophysical Research 84, 7201–7212,1979.

Kintner, P.M., Scales, W., Vago, J., et al. Harmonic H+ gyrofrequencystructures in auroral hiss observed by high-altitude auroral soundingrockets. Journal of Geophysical Research 96 (A6), 9627–9638, 1991.

Kintner, P.M., Schuck, P.W., Franz, J.R. Spaced measurements andprogress in understanding space plasma waves. Physics of Plasmas 7(5), 2135–2141, 2000a.

Kintner, P.M., Franz, J., Schuck, P., Klatt, E. Interferometric coherencydetermination of wavelength or what are broadband ELF waves?Journal of Geophysical Research 105 (A9), 21237–21250, 2000b.

Krassovsky, V.I. Exploration of the upper atmosphere with the help of thethird Soviet sputnik. Proceedings of the Institute of Radio Engineers47, 289, 1959.

McWilliams, R., Koslover, R. Laboratory observation of ion conics byvelocity space tomography of a plasma. Physical Review Letters 58 (1),37–40, 1987.

Minami, S., Takeya, Y. Ion temperature determination in the ionosphereby retarding potential analyzer aboard sounding rocket. Journal ofGeophysical Research 87, 713, 1982.

Mozer, F.S., Carlson, C.W., Hudson, M.K., et al. Observations of pairedelectrostatic shocks in the polar shocks in the polar magnetosphere.Physical Review Letters 38, 292, 1977.

Parker, L.W., Whipple, E.C. Theory of spacecraft sheath structure,potential, and velocity effects on ion measurements by traps and massspectrometers. Journal of Geophysical Research 75, 4720–4733, 1970.

Petro, A., Chang-Dıaz, F.R., Ilin A.V., et al. Development of a spacestation-based flight experiment for the VASIMR magneto-plasmarocket. Presented at the 40th AIAA Aerospace Sciences Meeting andExhibit, 14–17 January 2002, Reno, NV, 2002.

Sankaran, K., Cassady, L., Kodys A.D., et al. A survey of propulsionoptions for cargo and piloted missions to Mars. Presented at theInternational Conference on New Trends in Astrodynamics, 20–22,2003.

Sharp, R.D., Johnson, R.G., Shelley, E.G. Observation of an ionosphericacceleration mechanism producing energetic (keV) ions primarilynormal to the geomagnetic field direction. Journal of GeophysicalResearch 82, 3324, 1977.

Sharp, R.D., Johnson, R.G., Shelley, E.G. Energetic particle measure-ments from within ionospheric structures responsible for auroralacceleration processes. Journal of Geophysical Research 84, 480, 1979.

Shelley, E.G. Heavy ions in the magnetosphere. Space Science Reviews 23,465, 1979.

Shelley, E.G., Sharp, R.D., Johnson, R.G. Satellite observations of an ionacceleration mechanism. Geophysical Research Letters 3 (11), 654,1976.

Squire, J.P., Chang-Dıaz, F.R., Bengtson, R., et al. A plasma diagnosticset for the study of a variable specific impulse magnetoplasma rocket.Presented at the APS Division of Plasma Physics Meeting, Pittsburg,PA, 1997.

Squire, J.P., Chang-Dıaz, F.R., Jacobson, V.T., et al. Helicon plasmainjector and ion cyclotron acceleration development in the VASIMRexperiment. Presented at the 36th AIAA/ASME/SAE/ASEE JointPropulsion Conference, July 17–19, 2000, Huntsville, Alabama, 2000.

Stenzel, R.L., Williams, R., Aguero, R., Kitazaki, K., Ling, A., McDon-ald, T., Spitzer, J. Novel directional ion energy analyzer. Review ofScientific Instruments 53 (7), 1027–1031, 1982.

Stenzel, R.L., Gekelman, W., Wild, N., Urrutia, J.M., Whelan, D.Directional velocity analyzer for measuring electron distributionfunctions in plasma. Review of Scientific Instruments 54, 1302–1310,1983.

Stix, T.H. Waves in Plasma. American Institute of Physics, New York,NY, 1992.

Swanson, D.G. Plasma Waves. Academic Press, Boston, 1989.Ungstrup, E., Klumpar, D.M., Heikilla, W.J. Heating of ions to

superthermal energies in the topside ionosphere by electrostatic ioncyclotron waves. Journal of Geophysical Research 84, 4289, 1979.

Whipple, E.C. The ion trap-results in Exploration of the upperatmosphere with the help of the third Soviet sputnik. Proceedings ofthe Institute of Radio Engineers 47, 2023, 1959.

Yasaka, Y., Majeski, R., Browning, J., et al. ICRF heating with modecontrol provided by a rotating field antenna. Nuclear Fusion 28, 1765,1988.

Zintl, M., McWilliams, R., Wolf, N. Transverse ion acceleration and ionconic formation in a divergent-field laboratory plasma. Physics ofPlasmas 2 (12), 4432–4441, 1995.

Documents

Electromagnetic ion cyclotron resonance heating in the …have demonstrated single-pass electromagnetic ICRH in fast-ﬂowing laboratory plasma. 2. Experiment 2.1. The VASIMR engine