Effect of multiply charged ions on the performance and beam characteristics in annularand cylindrical type Hall thruster plasmasHolak Kim, Youbong Lim, Wonho Choe, and Jongho Seon Citation: Applied Physics Letters 105, 144104 (2014); doi: 10.1063/1.4897948 View online: http://dx.doi.org/10.1063/1.4897948 View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/105/14?ver=pdfcov Published by the AIP Publishing Articles you may be interested in An axially propagating two-stream instability in the Hall thruster plasma Phys. Plasmas 21, 072116 (2014); 10.1063/1.4890025 Effect of the annular region on the performance of a cylindrical Hall plasma thruster Phys. Plasmas 20, 023507 (2013); 10.1063/1.4793741 Electron cross-field transport in a low power cylindrical Hall thruster Phys. Plasmas 11, 4922 (2004); 10.1063/1.1791639 Plasma flow and plasma–wall transition in Hall thruster channel Phys. Plasmas 8, 5315 (2001); 10.1063/1.1421370 One-dimensional model of the plasma flow in a Hall thruster Phys. Plasmas 8, 3058 (2001); 10.1063/1.1371519

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27

Effect of multiply charged ions on the performance and beam characteristicsin annular and cylindrical type Hall thruster plasmas

Holak Kim,1 Youbong Lim,1 Wonho Choe,1,a) and Jongho Seon2

1Department of Physics, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Yuseong-gu,Daejeon 305-701, Republic of Korea2Department of Space Science and Astronomy, Kyung Hee University, 1732 Deokyoungdaero, Giheung-gu,Yongin-si, Gyeonggi-do 446-701, Republic of Korea

(Received 3 September 2014; accepted 29 September 2014; published online 10 October 2014)

Plasma plume and thruster performance characteristics associated with multiply charged ions in a

cylindrical type Hall thruster (CHT) and an annular type Hall thruster are compared under identical

conditions such as channel diameter, channel depth, propellant mass flow rate. A high propellant

utilization in a CHT is caused by a high ionization rate, which brings about large multiply charged

ions. Ion currents and utilizations are much different due to the presence of multiply charged ions.

A high multiply charged ion fraction and a high ionization rate in the CHT result in a higher

specific impulse, thrust, and discharge current. VC 2014 AIP Publishing LLC.

[http://dx.doi.org/10.1063/1.4897948]

Low-power Hall thrusters have, recently, received much

attention as one of the most promising electric propulsion

systems for space applications, particularly in conjunction

with advanced space missions such as formation flying and

micro-spacecraft constellation.1 The Hall thrusters have been

developed to a relatively high level of efficiency from 45%

to 55% in the power range from 0.5 to 5 kW.2 However, the

scaling down to low-power Hall thrusters with high effi-

ciency presents many difficulties in optimizing the magnetic

field, particle losses in the discharge channel, and channel

wall erosion, etc.2–6 In contrast to the conventional annular

type Hall thruster (AHT), cylindrical type Hall thrusters

(CHT) have mainly been studied for scaling down of the

Hall thruster for low power.3 The size reduction of the

thruster for low power inevitably involves an augmented

effect of the plasma-facing surface, and a high volume-to-

surface ratio of the discharge channel is obtained by remov-

ing the inner magnetic core in the CHT. Otherwise, the

operation principle of the CHT is basically similar to that of

the AHT.7,8 Both the high magnetic mirror ratio and the vol-

ume-to-surface ratio in the CHT are beneficial for better

electron confinement by a magnetic mirror effect and a

reduced particle loss effect at the dielectric channel wall.

Previous studies also demonstrated unusually high propellant

utilization and enhanced electron transport compared to

AHTs.7–10 As will be discussed in more detail below, we

found the existence of a high fraction of multiply charged Xe

ions, such as Xe2þ and Xe3þ in CHT plasmas, which is

closely related to performance characteristics, such as a high

specific impulse. In this paper, we describe the overall

thruster performance and Xe beam characteristics associated

with multiply charged ions in the AHT and CHT under iden-

tical conditions, including the diameter and depth of the dis-

charge channel and Xe flow rate.

Illustrated in Fig. 1 are schematics of the CHT and the

AHT used for the experiment, showing an anode with a gas

distributor, two electromagnetic coils, and a boron nitride ce-

ramic channel for each thruster. The outer diameter (50 mm)

and channel depth (24 mm) are identical for both thrusters,

but the inner core is retracted in the CHT as noted above.

The radial (Br) and axial magnetic fields (Bz) are also shown

in the figure. For the CHT, the currents in the two coils flow

in the same direction. The radial magnetic field lines along

the outer channel surface are mostly concentrated near the

channel exit for the AHT, but distributed over a relatively

wide region in the channel for the CHT.

The thruster was mounted on a thruster stand that is a

simple pendulum type with two pivots, and the displacement

of the stand calibrated by small weights was measured as

thrust using a laser and a position-sensitive detector combi-

nation. Experiments were carried out in a 3 m long and 1.5 m

diameter vacuum vessel, and the operation pressure was

maintained at 33 lTorr (for Xe gas) at a total Xe flow rate of

8 sccm. The anode mass flow rate was fixed at 7 sccm for

both thrusters. A commercial hollow cathode (Heatwave

HWPES-250) was used as an external neutralizer, and the

keeper current and Xe mass flow rate were kept at 1.58 A

and 1 sccm, respectively. Plume characteristics were studied

by a Faraday probe and an E�B probe. The Faraday probe

measures the angular distribution of the ion current density

and the total ion current collected by a commercial picoam-

meter (KEITHLEY 6485). The Faraday probe was installed

on the rotation stage with a radius of 48 cm centered at the

thruster exit, which can be rotated from �100� to þ100�

with respect to the thruster axis. An E�B probe is a velocity

filter that selects ions satisfying the Lorentz force equation

by balancing the perpendicular electromagnetic forces.11,12

It consists of an entrance collimator, a velocity filter, an exit

collimator, and a collector. Both collimators, made of stain-

less steel, are 70 mm in length and 4 mm in diameter. In the

velocity filter section, whose length is 140 mm, a magnetic

field is provided by two permanent magnets and its strength

a)Author to whom correspondence should be addressed. Electronic mail:

wchoe@kaist.ac.kr

0003-6951/2014/105(14)/144104/5/$30.00 VC 2014 AIP Publishing LLC105, 144104-1

APPLIED PHYSICS LETTERS 105, 144104 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27

is 0.23 T at the center of the filter body. An electric field per-

pendicular to the magnetic field is established by a pair of

aluminum plates separated by 10 mm. The casing of the

E�B probe is electrically grounded and its inside is kept

under vacuum via several small through holes. The E�B

probe was placed at 64 cm from the thruster exit on the

thruster axis.13

Shown in Fig. 2 are plume pictures and angular distribu-

tions of the ion current density for the AHT and CHT,

respectively. Both the plume pictures and the ion current

density profiles indicate that the angular distribution of the

ion beam is narrower for the AHT than for the CHT. In addi-

tion, the ion current density for the AHT and CHT increases

mainly near the thruster axis as the anode voltage is raised

FIG. 1. A schematic diagram and mag-

netic field profiles of (a) the AHT and

(b) the CHT. (c) Radial magnetic field

strength and (d) axial magnetic field

strength of the CHT and AHT along

the outer channel surface.

FIG. 2. Pictures of the plasma plume

of (a) the AHT and (b) the CHT.

Angular distribution of the ion current

density of (c) the AHT and (b) the

CHT at anode voltages from 220 V to

340 V with intervals of 20 V.

144104-2 Kim et al. Appl. Phys. Lett. 105, 144104 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27

from 220 V to 340 V. This large plume angle for the CHT

could be attributed to its distinct magnetic field topology and

the corresponding electric field normal to the magnetic

field.4,5,14 Because the broad plume angle is related to the

wall erosion and the thrust loss, various ways to reduce the

plasma plume angle have been investigated.8

The discharge current Id, ion current Ii, electron current

Ie, and propellant utilization Up under various voltages are

plotted in Fig. 3. Ion current Ii was obtained by angular inte-

gration of the measured ion current density, and Ie¼ Id � Ii.

At the same Xe mass flow rate, the CHT has higher Ie and Ii,

which could imply that the ionization rate of Xe is higher for

the CHT. The propellant utilization Up (¼MIi=e _m, where M,

_m, and e are the mass of a Xe atom, Xe mass flow rate, and

electron charge, respectively), which stands for ionization ef-

ficiency, is also higher for the CHT by a factor of 1.6–1.8

than for the AHT and well exceeds unity at anode voltages

from 220 V to 340 V as shown in Fig. 3(d). This large Up

exceeding unity for the CHT was reported previously9,10 and

was explained by the predicted presence of multiply charged

ions that were generated due to the increased ion residence

time inside the discharge channel. The high ionization rate

of the CHT also enables the thruster not only to operate at

low _m but also to keep a high Id even at the same _m in com-

parison with the AHT.

In order to experimentally investigate the multiply

charged ions in detail, the ion species fractions were meas-

ured at several different anode voltages. As illustrated in Fig.

4(a), the presence of multiply charged ions in the CHT is

clearly demonstrated in the E�B spectra that have three

prominent peaks corresponding to Xeþ, Xe2þ, and Xe3þ,

respectively. The figure shows the much higher normalized

peak current for Xe2þ in the CHT than in the AHT and the

distinct manifestation of the Xe3þ peak in the CHT. This

result can explain the reason Up> 1 for the CHT.

The plume attenuation as a result of charge exchange

(CEX) collisions can occur due to background neutrals

existing between the thruster exit and the E�B probe. This

effect was taken into account through the simplified CEX

correction model,15 which is a way to correct the attenuation

fraction between ions and background neutrals. The attenua-

tion fraction of each charge state can be written as

j

j0

� �Xeþ¼ e�n0r1z; r1 ¼ 87:3� 13:6 log Vdð Þ; (1)

j

j0

� �Xe2þ¼ e�n0r2z; r2 ¼ 45:7� 8:9 log 2Vdð Þ; (2)

j

j0

� �Xe3þ¼ e�n0r3z; r3 ¼ 16:9� 3:0 log 3Vdð Þ; (3)

where j is the E�B collector current, j0 is the current at the

thruster exit, ðj=j0ÞXekþ is the attenuation fraction, rk is the

CEX cross section for Xekþ (k¼ 1, 2, 3), n0 is the neutral

density, and z and Vd are the distance from the thruster exit

to the E�B probe and the discharge voltage of 300 V,

respectively. Here, the collision is regarded as a symmetric

reaction between an ion and a background neutral, i.e.,

XeþþXe ! XeþXeþ and Xe2þþXe ! XeþXe2þ. In

addition, Xe4þ and Xe5þ ions are neglected because their

fractions are considered to be small compared with Xe2þ and

Xe3þ and the calculation of the CEX correction for Xe4þ

and Xe5þ is complicated. The calculated attenuation frac-

tions for Xeþ, Xe2þ, and Xe3þ are 0.88, 0.95, and 0.98,

respectively. We define the ion current fraction Xk as

Xk ¼ Ik=Ii ¼ Ik=P3

k¼1 Ik ¼ nkZ3=2k =

P3k¼1 nkZ

3=2k , where nk,

Zk, and Ik are the number density, the charge state, and the

current of Xekþ (k¼ 1, 2, 3) ions, and Ii ¼P3

k¼1 Ik. The frac-

tions Xk were calculated by the triangle fitting method15 that

takes each triangle area as the amount of collected current at

each charge state, and are plotted in Figs. 4(b) and 4(c). The

fraction of the multiply charged ions X2 þ X3, which is the

sum of the current fractions of Xe2þ and Xe3þ, is 40%–52%

for the CHT and 10%–17% for the AHT. This large fraction

FIG. 3. (a) Discharge current, (b) ion

current, (c) electron current, and (d)

propellant utilization in relation to an-

ode voltages.

144104-3 Kim et al. Appl. Phys. Lett. 105, 144104 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27

for the CHT could be caused by the existence of slow ions

that have a long residence time inside the discharge chan-

nel.4 Since the large fraction of the multiply charged ions is

strongly linked to both the high thrust density and channel

erosion12 due to their high energy, the study on multiply

charged ions according to their fraction is important for the

thruster performance.

The overall performances of the thrusters, which are the

specific impulse Isp (¼T= _mg ¼ vi _mi= _mg, where vi, _mi, and gare the ion velocity, the ion mass flow rate, and the gravita-

tional acceleration, respectively), thrust T, and anode effi-

ciency g (¼T2=2 _mIdVa, where Va is the anode voltage), are

plotted in Fig. 5. Although the anode efficiency g is only

slightly (about 7%) higher for the CHT, Isp and T of the CHT

are much (about 40%) higher than those of the AHT at

Va¼ 300 V. At the same Va and _m for both thrusters, the

much higher Isp and T for the CHT are considered to be

attributed to the multiply charged ions and the high ioniza-

tion rate. In order to further understand the effect of Xe2þ

and Xe3þ in Isp and T, we define the effective ion speed

vi; eff ¼P3

k¼1 Nkvk=P3

k¼1 Nk and the effective ion mass flow

rate _mi; eff ¼P3

k¼1ðM _NkÞ ¼ MP3

k¼1ðAnkvkÞ, where vk and

A are the speed of the Xekþ ion and the hemispherical area,

respectively, and Nk is the number of Xekþ ions. Then,

the thrust and specific impulse can be expressed as

T ¼ vi; eff _mi; eff and Isp ¼ vi; eff _mi; eff= _mg. By assuming that

the ion current fractions measured at the thruster axis repre-

sents the most probable value in the entire plume, we use vk

obtained from the E�B spectra and nk from Xk, i.e.,

nk ¼ CXk=Z3=2k , where C is a constant. The calculated vi; eff

and _mi; eff for the CHT are 1.10 and 1.34 times higher,

respectively, than those for the AHT, which suggests that the

FIG. 4. (a) Normalized collector current versus bias potential of the E�B

probe. Ion current fraction of (b) the AHT and (b) the CHT corrected by

CEX collisions as a function of the anode voltage.

FIG. 5. (a) Specific impulse, (b) thrust, and (c) anode efficiency versus the

anode voltage for the AHT and CHT.

144104-4 Kim et al. Appl. Phys. Lett. 105, 144104 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27

high Isp of the CHT is more influenced by _mi; eff than vi; eff .

By following the calculation, both Isp and T of the CHT are

higher by 1.47 times than that of the AHT, which is

consistent with the measurement shown in Fig. 5. In addi-

tion, at the same anode voltage, the anode efficiency g is

about 3%–4% higher for the CHT due to the high Ii as shown

in Fig. 5(c).

On the other hand, another aspect is that the high ioniza-

tion rate and high multiply charged ion fraction in the CHT

give rise to a high discharge current Id that raises the power

consumption, and the broad plume angle causes the less effi-

cient use of thrust. It is also noted that a higher magnetic field

is required in the CHT to produce a thrust value similar to that

in the AHT in our experiment. By taking the coil power into

account, the thrust efficiencies (gt ¼ T2=2 _mðPa þ PcoilÞ,where Pa ¼ IdVa and Pcoil is the coil power) for AHT and

CHT are 34% and 30%, respectively, at the same anode

power of 300 W.

The high ionization rate and multiply charged ion frac-

tion have merits and demerits depending on limitations in ei-

ther the power or the propellant consumption. When there is

a limitation of the power consumption, AHTs show higher

efficiency, but if the limitation is in the propellant consump-

tion, CHTs have higher propellant efficiency.

In summary, performance and plume characteristics for

AHT and CHT were investigated under identical operating

conditions, such as channel diameter, channel depth, and Xe

flow rate. The ion current distribution was broader for the

CHT due to its distinct magnetic field lines forming equipo-

tential surfaces. The propellant utilization of the CHT

exceeded unity because of the high multiply charged ion frac-

tion and high ionization rate. The thrust and specific impulse

were much higher for the CHT at the same anode voltage

because of the higher _mi; eff and vi; eff . However, the large frac-

tion of multiply charged ions also caused the high Id that

raised the power consumption. Therefore, both thrusters can

be selectively used for low power or low propellant missions

depending on the criteria for each situation.

The authors thank Dr. Mihui Seo for her substantial

support. This work was partly supported by the Space Core

Technology Program (Grant No. 2014M1A3A3A02034510)

through the National Research Foundation of Korea funded

by the Ministry of Science, ICT and Future Planning. This

work was also partly supported by the Korea Institute of

Materials Science (KIMS) (Grant No. 10043470) funded by

the Ministry of Trade, Industry and Energy.

1J. Mueller, C. Marrese, J. Ziemer, A. Green, E. H. Yang, M. Mojarradi, T.

Johnson, V. White, D. Bame, R. Wirz, M. Tajmar, V. Hruby, M. Gamero-

Casta~no, J. Schein, and R. Reinicke, “JPL micro-thrust propulsion

activities,” Paper No. AIAA-2002-5714, Nanotech, Houston, USA, 2002.2K. D. Diamant, J. E. Pollard, Y. Raitses, and N. J. Fisch, IEEE Trans.

Plasma Sci. 38, 1052 (2010).3A. Smirnov, Y. Raitses, and N. J. Fisch, Phys. Plasmas 11, 4922 (2004).4A. Smirnov, Y. Raitses, and N. J. Fisch, J. Appl. Phys. 95, 2283 (2004).5M. Seo, J. Lee, J. Seon, H. J. Lee, and W. Choe, Appl. Phys. Lett. 103,

133501 (2013).6M. Rubanovich, V. Balabanov, A. Kapulkin, and E. Behar, in Proceedingsof the 33rd International Electric Propulsion Conference, GeorgeWashington University, Washington, D.C., USA, IEPC (2013), p. 235.

7M. Seo, J. Lee, J. Seon, H. J. Lee, and W. Choe, Phys. Plasmas 20, 023507

(2013).8Y. Raitses, A. Smirnov, and N. J. Fisch, Appl. Phys. Lett. 90, 221502

(2007).9J. Lee, M. Seo, J. Seon, H. J. Lee, and W. Choe, Appl. Phys. Lett. 99,

131505 (2011).10A. Smirnov, Y. Raitses, and N. J. Fisch, Phys. Plasmas 14, 057106 (2007).11S.-W. Kim, Ph.D. Dissertation, University of Michigan, 1999.12R. R. Hofer, Ph.D. Dissertation, University of Michigan, 2004.13Y. Lim, H. Kim, W. Choe, S. H. Lee, J. Seon, and H. J. Lee, Phys.

Plasmas 21, 103502 (2014).14B. E. Beal and A. D. Gallimore, Phys. Plasmas 12, 123503 (2005).15R. Shastry, R. R. Hofer, B. M. Reid, and A. D. Gallimore, Rev. Sci.

Instrum. 80, 063502 (2009).

144104-5 Kim et al. Appl. Phys. Lett. 105, 144104 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.145 On: Fri, 10 Oct 2014 14:54:27