Ghostscript wrapper for C:\TEMP\AcademicPress.pdf

P1: ZBU Final Pages

Qu: 00, 00, 00, 00

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

Robert G. Jahn
Edgar Y. Choueiri

Princeton University

I. Conceptual Organization and History of the Field

II. Electrothermal Propulsion

III. Electrostatic Propulsion
IV. Electromagnetic Propulsion

V. Systems Considerations

VI. Applications

GLOSSARY

Arcjet

Device that heats a propellant stream by passing

a high-current electrical arc through it, before the pro-
pellant is expanded through a downstream nozzle.

Hall effect

Conduction of electric current perpendicular

to an applied electric field in a superimposed magnetic
field.

Inductive thruster

Device that heats a propellant stream

by means of an inductive discharge before the propel-
lant is expanded through a downstream nozzle.

Ion thruster

Device that accelerates propellant ions by

an electrostatic field.

Magnetoplasmadynamic thruster

Device that acceler-

ates a propellant plasma by an internal or external mag-
netic field acting on an internal arc current.

Plasma

Heavily ionized state of matter, usually gaseous,

composed of ions, electrons, and neutral atoms or
molecules, that has sufficient electrical conductivity to
carry substantial current and to react to electric and
magnetic body forces.

Resistojet

Device that heats a propellant stream by pass-

ing it through a resistively heated chamber before
the propellant is expanded through a downstream
nozzle.

Thrust

Unbalanced internal force exerted on a rocket

during expulsion of its propellant mass.

THE SCIENCE AND TECHNOLOGY

of electric

propulsion (EP) encompass a broad variety of strate-
gies for achieving very high exhaust velocities in order
to reduce the total propellant burden and corresponding
launch mass of present and future space transportation
systems. These techniques group broadly into three cat-
egories: electrothermal propulsion, wherein the propel-
lant is electrically heated, then expanded thermodynami-
cally through a nozzle; electrostatic propulsion, wherein
ionized propellant particles are accelerated through
an electric field; and electromagnetic propulsion, wherein
current driven through a propellant plasma interacts
with an internal or external magnetic field to provide a

Encyclopedia of Physical Science and Technology, Third Edition, Volume 5

2002 by Academic Press. All rights of reproduction in any form reserved.

125

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

126

Electric Propulsion

stream-wise body force. Such systems can produce a range
of exhaust velocities and payload mass fractions an order
of magnitude higher than that of the most advanced chem-
ical rockets, which can thereby enable or substantially
enhance many attractive space missions. The attainable
thrust densities (thrust per unit exhaust area) of these sys-
tems are much lower, however, which predicates longer

ight times and more complex mission trajectories. In ad-

dition, these systems require space-borne electric power
supplies of low speci

c mass and high reliability, inter-

faced with suitable power processing equipment. Opti-
mization of EP systems thus involves multidimensional
trade-offs among mission objectives, propellant and power
plant mass, trip time, internal and external environmental
factors, and overall system reliability. An enduring inter-
national program of research and development of viable
electric thrusters has been in progress for several decades,
and over the past few years this has led to the increas-
ing use of a number of EP systems on commercial and
governmental spacecraft. Meanwhile, yet more advanced
EP concepts have matured to high credibility for future
mission applications.

I. CONCEPTUAL ORGANIZATION

AND HISTORY OF THE FIELD

A. Motivation

The stimulus for development of electrically driven
space propulsion systems is nothing less fundamen-
tal than Newton

’

s laws of dynamics. Since a rocket-

propelled spacecraft in free

ight derives its only

acceleration from discharge of propellant mass, its equa-
tion of motion follows directly from conservation of
the total momentum of the spacecraft and its exhaust
stream:

(1)

where

is the mass of the spacecraft at any given time,

its acceleration vector,

the velocity vector of the ex-

haust jet relative to the spacecraft, and

the rate of change

of spacecraft mass due to propellant-mass expulsion. The
product

is called the thrust of the rocket,

, and for

most purposes can be treated as if it were an external force
applied to the spacecraft. Its integral over any given thrust-
ing time is usually termed the impulse,

, and the ratio of

the magnitude of

to the rate of expulsion of propellant

in units of sea-level weight,

, has historically been la-

beled the speci

c impulse,

. If

is constant

over a given period of thrust, the spacecraft achieves an
increment in its velocity,

, which depends linearly on

and logarithmically on the amount of propellant mass

expended:

(2)

where

and

are the total spacecraft mass at the start

and completion of the acceleration period. Conversely, the
deliverable mass fraction,

, is a negative exponen-

tial in the scalar ratio

υ/υ

−

υ/υ

(3)

Inclusion of signi

cant gravitational or drag forces on

the

ight of the spacecraft adds appropriate terms to Eq. (1)

and considerably complicates its integration, but it is still
possible to retain relation (3), provided that

is now

regarded as a more generalized

“

characteristic velocity

increment,

”

indicative of the energetic dif

culty of the

particular mission or maneuver. However represented, the
salient point is simply that if the spacecraft is to deliver a
signi

cant portion of its initial mass to its destination, the

rocket exhaust speed must be comparable to this charac-
teristic velocity increment. Clearly, for missions of large

, the burden of thrust generation must shift from high

rates of ejection of propellant mass to high relative exhaust
velocities. Unfortunately, conventional chemical rockets,
whether liquid or solid, monopropellant or bipropellant,
are fundamentally limited by their available combustion
reaction energies and heat transfer tolerances to exhaust
speeds of a few thousand meters per second, whereas
many attractive space missions entail characteristic ve-
locity increments at least an order of magnitude higher.
Thus, some fundamentally di

erent concept for the accel-

eration of propellant mass that circumvents the intrinsic
limitations of chemical thermodynamic expansion is re-
quired. Into this breech step the family of electric propul-
sion possibilities.

B. Conceptual Subdivision

So that propellant exhaust speeds in the range above
10,000 m/sec desirable for interplanetary

ight and other

high-energy missions can be obtained, processes basically
different from nozzled expansion of a chemically reacting

ow must be invoked. More intense forms of propellant

heating may be employed, provided that the walls of the
rocket chamber and nozzle are protected from excessive
heat transfer. Alternatively, the thermal expansion route
may be bypassed completely by direct application of suit-
able body forces to accelerate the propellant stream. Either
of these options is most reasonably accomplished by elec-
trical means, which constitute the technology of electric
propulsion.

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

127

Historically, conceptually, and pragmatically, this

eld

has tended to subdivide into three categories:

Electrothermal propulsion

, wherein the propellant is

heated by some electrical process, then expanded
through a suitable nozzle

Electrostatic propulsion

, wherein the propellant is

accelerated by direct application of electrostatic
forces to ionized particles

Electromagnetic propulsion

, wherein the propellant is

accelerated under the combined action of electric and
magnetic

elds

Over their periods of development, each of these ap-
proaches has spawned its own array of technical special-
ties and subspecialties, its own balance sheet of advantages
and limitations, and its own cadres of proponents and de-
tractors, but in serious assessment, each has validly qual-
i

ed for particular niches of application, many of which

do not seriously overlap. Throughout the history of EP de-
velopment, the original subdivision of the

eld into elec-

trothermal, electrostatic, and electromagnetic systems has
remained useful, and this subdivision will be respected
through the balance of this article. It should be recog-
nized, however, that in virtually all practical systems, two
or even all three of these processes function in some con-
cert to accelerate, channel, and expand the propellant

ow,

and in many cases it is the ef

cacy of this cooperation that

determines the utility of any given device.

The exhaust velocities attainable by these methods,

especially the latter two, are more than adequate for
many large-velocity-increment missions beyond the vi-
able chemical range. Indeed, some restraint of their

capability may be required because of their associated

“

power supply penalty.

”

Clearly, each of these concepts

entails two functional components: the thruster itself and
an electric power supply to drive it. The latter adds mass,

, to the composite propulsion system in some propor-

tion to the power level of operation,

, which in turn scales

with the square of the exhaust velocity:

(4)

where

is the speci

c mass of the power supply (mass

per unit power), and

is the ef

ciency with which the

thruster converts its input power to thrust power,

Since the requisite propellant mass scales inversely with

, it follows that for any given mission requirement,

there is an optimum

that minimizes the sum of the

propellant mass and that of the requisite power supply.
Relation (4) also emphasizes the importance of utilizing
power systems of low speci

c mass and thrusters of high

conversion ef

ciency. Overlaid on all this is the evident

necessity for impeccable reliability of both components
of the system over long periods of unattended operation
in the space environment.

C. History of Effort

The attractiveness of EP for a broad variety of space trans-
portation applications was recognized by the patriarch of
modern rocketry, Robert H. Goddard, as early as 1906. His
Russian counterpart, Konstantin Tsiolkovskiy, proposed
similar concepts in 1911, as did the German Hermann
Oberth in his classic book on space

ight in 1929 and the

British team of Shepherd and Cleaver in 1949. But the

rst

systematic and tutorial assessment of EP systems should
be attributed to Ernst Stuhlinger, whose book

Ion Propul-

sion for Space Flight

nicely summarizes his seminal stud-

ies of the 1950s.

The rapid acceleration of the U.S. space ambitions in

the 1960s drove with it the

rst coordinated research and

development programs explicitly addressing EP technol-
ogy. In its earliest phase, this e

ort drew heavily on reser-

voirs of past experience in other areas of physical science
and engineering that had employed similar electrother-
mal, electrostatic, and electromagnetic concepts to their
own purposes, such as arc-heated wind tunnels and weld-
ing practice, cathode ray tubes and mass-spectroscopic
ion sources, and magnetohydrodynamic channel

ows and

railguns. From these transposed technologies blossomed
a signi

cant new component of the burgeoning space in-

dustry that concerned itself not only with the development
of viable electric thrusters, but also with the provision
of suitable electric power supplies and power condition-
ing equipment, major ground test facilities, and sophisti-
cated mission analyses of a smorgasbord of potential space
applications.

Following a sizable number of experimental

ight tests,

EP entered its era of commercial application in the early
1980s, as resistojets became common options for sta-
tion keeping and attitude control on tens of commercial
spacecraft. In the early 1990s, electrothermal arcjets were
adopted for north

–

south station keeping (NSSK) of many

communication satellites in geosynchronous earth orbit
(GEO). The year 1994 saw the

rst use of electrostatic

ion thrusters for the NSSK of commercial satellites, and
the year 1998 their application on a planetary NASA mis-
sion. Although Hall thrusters have been used on Soviet
and Russian spacecraft since the mid-1970s, and there
have been a few applications of pulsed plasma thrusters,
electromagnetic thrusters are only now entering their era
of application on Western commercial spacecraft. In to-
tal, the number of electrically propelled spacecraft has
gone from single digits in the 1960s to double digits in
the 1970s and 1980s and has reached the triple-digit mark

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

128

Electric Propulsion

in the late 1990s. A recent emphasis in research and de-
velopment has been the scaling down, in both physical
size and power level (

100 W), of many EP concepts

for future applications on micro-spacecraft. At the other
extreme, the prospect of energetic missions

—

with large

cargo and piloted payloads

—

to the planets, which stand to

bene

t most from EP, remains futuristic until the required

high power levels (100 kW and above) become available
in space.

II. ELECTROTHERMAL PROPULSION

A. Overview

Electrothermal propulsion comprises all techniques
whereby the propellant is electrically heated in some
chamber and then expanded through a suitable nozzle to
convert its thermal energy to a directed stream that deliv-
ers reactive thrust power to the vehicle. Three subclasses
of this family may be denoted in terms of the physical
details of the propellant heating:

Resistojets

, wherein heat is transferred to the

propellant from some solid surface, such as the
chamber wall or a heater coil

Arcjets

, wherein the propellant is heated by an

electric arc driven through it

Inductively and radiatively heated devices

, wherein

some form of electrodeless discharge or
high-frequency radiation heats the

Each of these strategies relieves some of the intrinsic

limitations of the chemical rocket in the sense that the pro-
pellant species may be selected for its propitious physical
properties independently of any combustion chemistry, but
heat transfer constraints and frozen

ow losses (losses due

to unrecuperated energy

“

frozen

”

in the internal modes

and dissociation of the molecules) remain endemic.

The gross performance of any electrothermal thruster

can be crudely forecast by means of a rudimentary one-
dimensional energy argument that limits the exhaust speed
of the

ow from a fully expanded nozzle to

≤

where

is the speci

c heat at constant pressure per unit

mass of the propellant and

is the maximum tolerable

chamber temperature. Propellants of the lowest molecular
weight thus seem preferable, and indeed hydrogen might
at

rst glance appear optimum, but in practice its frozen

ow propensities and dif

culty of storage compromise

its attractiveness. More complex molecular gases such as
ammonia and hydrazine, which dissociate into fairly low
ef

ective molecular weights and high speci

c heat gas

mixtures in the chamber, are currently more popular, but in

these cases also, frozen

ow kinetics in the nozzle remain

important to performance.

B. Resistojets

In the resistojet subclass of devices, chamber temperature
is necessarily limited by the materials of the walls and/or
heater coils to some 3000

◦

K or less, and hence the ex-

haust velocities, even with equilibrated hydrogen, cannot
exceed 10,000 m/sec, which is nonetheless a factor of two
or three beyond that of the best chemical rockets. In con-
temporary practice, lower performance but more readily
space storable propellants, such as hydrazine and ammo-
nia, along with biowaste gases such as water vapor and
carbon dioxide, are more commonly employed because
of their overall system advantages.

Beyond the frozen

ow kinetics, the major practical

challenge facing resistojet technology is retaining the in-
tegrity of the insulator and heater surfaces at the very high
temperatures the concept demands, while still minimiz-
ing the viscous and radiative heat losses that further de-
crease thruster ef

ciency. Since the mid-1960s, many con-

gurations of resistojet have been conceived, researched,

and developed to optimize these processes, and a few,
such as the

ight-ready module shown schematically and

in the photograph in Fig. 1, have evolved to practical
space thrusters and been deployed on suitable missions.
A typical resistojet uses catalytically decomposed hy-
drazine as its propellant and achieves an exhaust velocity
of 3500 m/sec and a thrust of 0.3 N at an ef

ciency of 80%

when operating at a power level of 750 W.

From a system point of view, resistojets are particularly

attractive because they readily lend themselves to inte-
gration with previously developed and commonly used
propellant storage and

ow management systems for hy-

drazine monopropellant thrusters. Another advantage is
their low operational voltage, which, unlike that in other
EP systems, does not require complex power processing.
For these reasons, and the fact that satellites in GEO often
have excess electrical power, resistojets were among the

rst EP options to be used for the NSSK of communica-

tion satellites. While the earliest use of resistojets in space
dates to 1965 (the Air Force Vela satellites), their adop-
tion on commercial spacecraft did not start until the 1980
launch of the

rst satellites in the INTELSAT-V series. A

more recent application has been for orbit insertion, atti-
tude control, and deorbit of LEO satellites, including the
72 satellites in the Iridium constellation.

C. Arcjets

If an electrothermal thruster is to attain exhaust speeds
substantially higher than 10,000 m/sec, interior portions

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

129

FIGURE 1

Photograph and schematic of a flight-ready hydrazine

resistojet. [Courtesy of the Primex Corporation.]

of the propellant

ow through the heating chamber must

reach temperatures as high as 10,000

◦

K, while being re-

strained from direct contact with the chamber and nozzle
walls. Thus, steep radial gradients in temperature must be
sustained, which renders the entire

ow pattern explicitly

two-dimensional. The most effective and straightforward
means for achieving such pro

les is by passing an elec-

tric arc directly through the chamber in some appropriate
geometry. Figure 2 shows a diagram and a photograph of
a prototypical thruster of this class, commonly called an
electrothermal arcjet. Direct currents of tens or hundreds
of amperes are passed through the gas

ow between an

upstream conical cathode and a downstream annular an-
ode integral to the exhaust nozzle, generating a tightly
constricted arc column that reaches temperatures of sev-
eral tens of thousands of degrees on its axis. The incom-
ing propellant is usually injected tangentially, then swirls
around, along, and through this arc, expanding in the an-
ode/nozzle to average velocities of tens of thousands of
meters per second. Properly designed and operated, the

chamber and nozzle walls remain tolerably cool under the
steep radial gradients, and even the arc attachment regions
on the cathode and anode are somewhat protected by the
electrode sheath processes, even though the cathode tip
must reach incandescent temperatures to provide the req-
uisite thermionic emission of electron current.

Analytical models of thrusters of this type usually repre-

sent the arc in three segments: a cathode fall region, which
functions to heat the cathode tip and extract electrons from
it; an arc column, wherein ohmic heating hboxsustains the
necessary ionization against interior recombination and
radiation losses; and an anode fall region, wherein the arc
terminates in a diffuse annular attachment on the diverg-
ing nozzle wall, depositing thermal electron energy into
the body of the thruster. Heating of the propellant actually
occurs in two important modes: by direct passage of a core
portion of the

ow through the arc itself, and by conduc-

tion and convection to the outer

ow from the chamber and

nozzle walls, which themselves have been heated by radi-
ation from the arc column and by the anode attachment.
This latter, regenerative component rescues the ef

ciency

of the thruster somewhat from the detrimental frozen

losses associated with the failure of the hottest portion of
the core

ow to recover much of the energy invested in its

ionization and dissociation. Aside from these frozen

losses, the ef

ciency also suffers from viscous effects,

nonuniform heat addition across the

ow, and heat depo-

sition in the near-electrode regions due to voltage drops
in the electrode sheaths.

Arcjets on contemporary operational

ights typically

use catalytically decomposed hydrazine as propellant and
operate at a power level of about 1.5 kW with an ex-
haust velocity between 5000 and 6000 m/sec and an ef

ciency up to 40%. While ammonia, by virtue of its lower
molecular mass, can offer an exhaust velocity as high
as 9000 m/sec at the same power levels, the associated
complexity of the mass feeding system favors the use of

FIGURE 2

Photograph and schematic of a 1.5-kW arcjet in

operation. [Courtesy of the Primex Corporation.]

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

130

Electric Propulsion

hydrazine. Since these arcjets operate at a voltage of about
100 V, which is generally higher than the spacecraft bus
voltage, dedicated power processing units, whose mass
can exceed that of the dry propulsion system, are required.

Starting with the

rst of the Telstar-4 series of GEO

communication satellites launched in 1993, hydrazine ar-
cjets have quickly gained acceptance as viable propulsion
options for NSSK. They represent the second evolution-
ary step, after resistojets, in the use of EP systems and
offer substantial propellant mass savings over all previ-
ous monopropellant propulsion options. Although a re-
cent test

ight of a 30-kW ammonia arcjet (on the Air

Force ESEX spacecraft) has demonstrated the potential of
this higher power class of electrothermal propulsion for
more thrust-intensive missions such as orbit transfers and
primary propulsion maneuvers, the dif

culty of providing

such high power in space, combined with the lifetime-
limiting problems of electrode erosion and whiskering
have so far delayed such applications.

D. Inductively and Radiatively Heated Devices

The most vulnerable elements of direct current arcjets are
the electrodes that transmit the high currents from the ex-
ternal circuit to the arc plasma, and their erosion ultimately
limits the operational lifetime of these thrusters. In ef-
forts to alleviate this basic problem, a number of more ex-
otic concepts for electrothermal propulsion have been pro-
posed and implemented, wherein the propellant is ionized
and heated by means of some form of electrodeless dis-
charge. These have varied widely in power levels, geome-
tries, propellant types, and densities and have utilized ap-
plied frequencies ranging from low radio frequency (RF)
to the microwave bands. In all cases, the strategy is to heat
the free electron component of the ionized propellant by
means of an applied oscillating electromagnetic

eld and

then to rely either on ambipolar di

usion to direct the ions

and neutrals along an appropriate exhaust channel (induc-
tive thrusters) or on collisional and radiative heating of
the neutral component by a sustained plasma upstream
of the throat of a diverging nozzle (microwave thrusters).
Devices of this class are thus hybrid electrothermal and
electrostatic and, indeed, since some of them also employ
magnetic

elds to con

ne and direct the

ow, may actually

embody all three classes of interaction.

Early enthusiasm for this class of accelerators was nec-

essarily tempered by the relatively low ef

ciency of RF

and microwave power generation technologies of that
time, which would have transcribed into intolerably mas-
sive space power supplies. More recent advances in solid-
state power processing have revived some of these con-
cepts, although none has yet been

ight-tested. The most

mature of these concepts is currently a microwave elec-

trothermal thruster that operates with hydrogen, nitro-
gen, or ammonia at exhaust velocities ranging between
4000 m/sec and 12,000 m/sec and ef

ciencies as high as

60%, excluding the ef

ciency of the microwave source.

The microwave electrothermal thruster seems particularly
amenable to scaling to low powers. While most of the
recent development has been at the kilowatt level, scaled-
down prototypes operating ef

ciently at 100 W and below

have also been developed.

III. ELECTROSTATIC PROPULSION

A. Basic Elements

The fundamental thermal limitations on attainable exhaust
speeds and lifetimes associated with the heating and ex-
pansion processes of electrothermal accelerators can be
categorically circumvented if the propellant is directly ac-
celerated by an external body force. The simplest such
device, in concept, is the ion thruster, wherein a beam of
atomic ions is accelerated by a suitable electric

eld and

subsequently neutralized by an equal

ux of free electrons.

The essential elements of such a thruster are sketched in
Fig. 3, where a collisionless stream of positive atomic ions,
liberated from some source, is accelerated by an electro-
static

eld established between the source surface and a

suitable permeable grid. Downstream of this region, elec-
trons from another source join the ion beam to produce
a stream of zero net charge, which exits the accelerator
at a speed determined not only by the net potential drop
between the ion source and the plane of effective neu-
tralization, but also by the charge-to-mass ratio of the ion
species employed.

A quick calculation, based on reasonable electrode di-

mensions, manageable applied voltages, and available ion
charge-to-mass ratios, indicates that extremely high ex-
haust speeds, well in excess of 10

m/sec, are readily

achievable. Indeed, given the power supply mass penalty,
which scales strongly with the exhaust velocity accord-
ing to Eq. (4), these devices tend to optimize their thrust
ef

ciency at too high an exhaust velocity for most near-

earth and interplanetary mission applications. A more
troublesome drawback, however, is that regardless of the
particular electrode con

gurations and propellant species

employed, thrusters of this class are severely limited in
their attainable thrust density by space-charge distortions
of the applied electric

eld pattern. Speci

cally, it can

readily be shown that the maximum ion current density
that can be sustained through a one-dimensional acceler-
ation gap

across which is applied a voltage

(5)

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

131

FIGURE 3

Ion thruster schematic.

where

is the dielectric permittivity constant and

is the ion charge-to-mass ratio, all in mks units. It then
follows that the maximum thrust density of the emitted
beam depends only on

j M

(6)

where

is the area of the exhaust jet, while the corre-

sponding exhaust speed depends only on

and

q V

(7)

The thrust density and thrust power density that can be
conveyed on the exhaust beam for attainable values of

, and

thus compute to rather small values, on

the order of a few newtons per square meter and 10

per square meter, respectively, at best. On the positive
side, the attainable thrust ef

ciency is essentially limited

only by the energetic cost of preparing the individual ions,
which should be a small fraction of their exhaust kinetic
energy. System optimization, therefore, involves a some-
what complex multidimensional trade-off among the ex-
haust speed, thrust density, ef

ciency, and power system

speci

c mass, for any given mission application.

B. Ion Thruster Technology

1. Ion Sources

In practice, the most amenable propellants for electro-
static thrusters have proven to be cesium, mercury, argon,
krypton, and most commonly xenon, and many possible
sources of such ions of the requisite ef

ciency, reliabil-

ity, and uniformity have been conceived and developed.
Of these, only three, the electron bombardment discharge

source, the cesium

–

tungsten surface contact ionization

source, and one form of RF discharge source, have sur-
vived to application.

The essential elements of the bombardment sources are

some form of cylindrical discharge chamber containing
a centerline cathode that emits electrons, a surrounding
anode shell, and a permeating azimuthal and radial mag-
netic

eld that constrains the electrons to gyrate within the

chamber long enough to ionize the injected propellant gas
and to direct it, once ionized, to extractor and accelera-
tor grids downstream. One contemporary implementation
of such a chamber is shown in Fig. 4. This particular de-
vice employs a hollow cathode electron source, wherein
is sustained a secondary discharge that facilitates electron
emission from the interior walls of the cathode cavity. The
magnetic

eld permeating the entire chamber is provided

by three ring magnets, empirically con

gured to estab-

lish a grossly diverging but doubly cusped

eld pattern

that optimizes the discharge for ionization and ion ex-
traction purposes. The magnitude of this

eld is adjusted

in concert with the anode

–

cathode voltage differential to

maximize the ionization ef

ciency and discharge stabil-

ity while minimizing the production of doubly charged
ions, which would be out of focus in the accelerator gap
and thus tend to erode the grids through high-energy sput-
tering. Typical values for xenon and mercury propellants
would be in the regimes of 0. 25 T and 30 V, respectively.
Slightly different chamber con

gurations and

eld values

have also been used successfully.

Contact ion sources rely on the difference between the

electronic work function of a metallic surface and the ion-
ization potential of alkali vapors to ionize the latter on
contact with the former. Very few metal

–

alkali combina-

tions have this requisite positive voltage differential, and
of these the combination of tungsten and cesium provides

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

132

Electric Propulsion

FIGURE 4

Ring-cup ion thruster. [Courtesy of Colorado State University and NASA Glenn Research Center.]

the largest. The most common implementation has been
to force hot cesium vapor through a porous tungsten wafer
to enhance surface contact, but problems in degradation of
wafer porosity and recondensation of cesium vapor have
tended to compromise the ionization ef

ciency and life-

time of these sources.

The RF ionization sources currently favored in Western

Europe are similar in principle and con

guration to the

U.S. electron bombardment sources, except that the dis-
charge is inductively driven RF rather than directly cou-
pled dc. While the ef

ciency and lifetime of these RF

sources seem competitive, they entail the complication of
RF modules in their power processing equipment. In Japan
another cathodeless ion thruster concept has been devel-
oped that uses a microwave source to create and sustain
the plasma through electron cyclotron resonance (ECR)
and offers some system and lifetime advantages.

2. Accelerator Grids

In virtually all classes of ion thruster, the positive ions are
extracted from the source and accelerated downstream by
a system of grids con

gured to achieve the desired ex-

haust velocity with minimum beam impingement. In U.S.
bombardment engines, for example, a double grid con

uration is usually dished downstream as shown in Fig. 4

to improve its mechanical and thermal stability against
distortion. The upstream grid is maintained at a higher
positive potential than required by the desired exhaust
speed in order to enhance the ion extraction process and in-
crease the space-charge limited current density that can be
sustained. The downstream grid then reduces the exhaust
plane potential to the desired value. This

“

accel

–

decel

”

scheme has the advantages of higher beam density at a
given net voltage and of reducing electron backstreaming
from the neutralized beam downstream.

The grid perforations are con

gured analytically and

empirically to focus the ion stream into an array of beam-
lets that pass through with minimum impingement. In this
process, the downstream surface of the discharge plasma
in the chamber acts as a third electrode, and since this
contour is not independent of the discharge characteris-
tics and applied grid voltages, it can be a source of some
instability. Further complications are introduced by the
small fractions of double ions or neutrals that

nd their

way into the beam and are henceforth out of focus and
free to bombard the grid surfaces.

3. Neutralizers

If the ion beam emerging from the downstream electrode
is not to stall on its own interior potential pro

le, it must

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

133

be electrostatically neutralized within a very few units of
grid spacing. This is typically achieved by provision of a

ux of electrons, usually from another hollow cathode dis-

charge, which fortuitously mix effectively within the ion
beam by means of a variety of microscopic and macro-
scopic internal scattering processes. Once so neutralized,
this plasma constitutes a downstream

“

virtual electrode

”

that completes the axial potential pattern.

4. System Aspects and Application History

Although they are technically the most complex EP sys-
tems, ion engines like those outlined above and shown
in Fig. 5 have been the most thoroughly developed and
tested of all EP devices. Their appeal stems primarily from
their maturity, demonstrated long lifetime (

20,000 hr),

relatively low beam divergence (

20 deg) and high ef

ciency (65%) at a useful exhaust velocity (30,000 m/sec),
and power levels between 200 and 4000 W. These advan-
tages are somewhat offset by low thrust density, system
complexity, and high-voltage requirements which trans-
late into power processing unit (PPU) speci

c masses

as high as 10 kg/kW. Their space-worthiness has been
demonstrated by more than a dozen U.S. and Soviet

ight

tests, starting in 1962, which have provided guidance and
validation to ground-based research and development ef-
forts that have led to the optimization of their designs and
materials.

The

rst operational use of ion thrusters occurred in

1994 on the Japanese ETS-6 and COMETS satellites, for
which four 12-cm ion engines provided NSSK propulsion.
This was followed in 1997 by PAS-5, which inaugurated
the

rst U.S. commercial satellite bus to rely on ion propul-

sion for GEO station keeping.

FIGURE 5

A 30-cm bombardment ion engine. [Courtesy of

Hughes Aircraft Co.]

While ion engines compete well with other EP op-

tions for near-earth applications, their ability to operate
ef

ciently and reliably at even higher exhaust velocities

makes them ideally suited for energetic (i.e., high

)

deep-space missions, where long thrusting times can be
tolerated. In 1998, NASA

’

s Deep Space 1 became the

rst

interplanetary mission to bene

t from ion propulsion. On

its way to its encounter with asteroid Braille, the space-
craft used a xenon bombardment ion propulsion system
to provide the required

over 1800 hr of thrust, while

consuming only 12 kg of propellant and demonstrating in-
space performance within 1% of that measured in ground
tests. Various commercial and scienti

c missions using

ion propulsion are slated for launch in the next few years,
including the world

’

rst sample-and-return attempt from

an asteroid by the Japanese MUSES-C spacecraft.

C. Other Electrostatic Propulsion Concepts

Many of the complexities of ion bombardment sources,
multibeam focusing grids, electromagnets, and other sub-
systems of ion thrusters can be bypassed altogether if only
minute thrust levels are needed. Creating a high electric

eld concentration at the lips of a capillary slit, as shown

in Fig. 6, allows direct ionization from the liquid phase of
a metal to be achieved by

eld emission, and the resulting

ion beam can be accelerated electrostatically to very high
velocities. Field emission electric propulsion (FEEP) de-
vices of this kind have evolved in Europe since the late
1970s and have unique characteristics and advantages. In
a typical FEEP device, cesium propellant from a small
reservoir is allowed to wet the inside of a 1-

m capillary

channel and form a free surface between the blade-edge
lips of the emitter. An electric

eld of a few kilovolts is

FIGURE 6

Schematic of a

eld emission electric propulsion

(FEEP) device.

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

134

Electric Propulsion

applied between the emitter and an accelerator electrode,
and the resulting electric

eld concentrations at the slit

edges form protruding cusps, or

“

Taylor cones,

”

on the

edge of the liquid by means of the competition between
the electrostatic forces and surface tension. When the elec-
tric

eld reaches

eld emission levels (10

V/m), these

points become local ion emission sites. The extracted and
accelerated ion beams are subsequently neutralized by in-
jecting electrons from an appropriate source. For a typical
extraction voltage of 10 kV, the ion exhaust velocity is in
excess of 100,000 m/sec, the ef

ciency close to 100%, and

the thrust-to-power ratio about 16

N/W. Although FEEP

thrusters with thrust levels as high as 5 mN have been de-
veloped by the European Space Agency, near-term appli-
cations are for missions requiring small and precise thrust.
While no such devices have yet

own, a number of mis-

sions are planned in the United States and Europe, includ-
ing systems associated with space-borne interferometers
for detection of gravitational waves, and missions requir-
ing

ne pointing and formation

ying of micro-spacecraft.

Since these FEEP devices are operated with cesium be-
cause of its high atomic mass, low ionization potential, low
melting point (28.4

◦

C), and good wetting capabilities, a

number of practical problems related to spacecraft plume
interactions and propellant contamination will need to be
resolved.

Another simple electrostatic thruster concept that has

the advantages over FEEP of higher thrust-to-power ratios
and the use of more benign propellants is the colloidal
thruster. It employs similar physical processes, except
that nonmetallic liquids are used and sub-micron-sized
charged particles (colloids) are produced and accelerated.
This yields speci

c impulses more compatible with near-

earth missions. Much of the work on colloidal thrusters
was carried out in the 1960s and identi

ed limitations

on the achievable charge-to-mass ratios and the unifor-
mity of the the charge-to-mass distributions. The for-
mer transcribes to excessively large voltages (hundreds
of kilovolts) to attain the desirable exhaust velocities
(10,000 m/sec) and the latter results in large beam diver-
gences. Nonetheless, more recent research in Russia and in
the United States, driven by the advent of micro-spacecraft
missions, has returned colloidal electrostatic thrusters to
the arsenal of electric micropropulsion options.

IV. ELECTROMAGNETIC PROPULSION

A. Basic Concept

The third category of EP relies on the interaction of an
electric current pattern driven through a conducting pro-
pellant stream with a magnetic

eld permeating the same

FIGURE 7

Crossed-

eld electromagnetic thruster schematic.

region to provide the accelerating body force. Such sys-
tems can produce exhaust speeds considerably higher than
those of the electrothermal devices, and thrust densities
much larger than those of the electrostatic thrusters, but
are phenomenologically more complex and analytically
less tractable than either of these alternatives. The essence
of an electromagnetic thruster is sketched in Fig. 7, where
some electrically conducting

uid, usually a highly ion-

ized gas, is subjected to an electric

eld

and a mag-

netic

eld

, perpendicular to each other and to the

uid

velocity

. The current density

driven by the electric

eld interacts with

to provide a streamwise body force

that accelerates the

uid along the channel.

The process may alternatively be represented from a par-
ticulate point of view in terms of the mean trajectories
of the current-carrying electrons, which, in attempting to
follow the electric

eld, are turned downstream by the

magnetic

eld, transmitting their streamwise momentum

to the heavy particles in the stream by collisions and/or by
microscopic polarization

elds. It is important to note that

in either representation, the working

uid, although highly

ionized, is macroscopically neutral, hence not constrained
in its mass

ow density by space-charge limitations as in

the electrostatic accelerators.

B. Varieties

Unlike the electrothermal or electrostatic classes, which
offer only a few practical con

gurations, electromagnetic

acceleration presents myriad possibilities for implemen-
tation. The applied

elds and internal currents may be

steady, pulsed, or alternating over a broad range of fre-
quencies; the

elds may be externally applied or induced

by the current patterns; and a broad variety of propel-
lant types, including liquids and solids, may be employed,
along with a host of channel geometries; electrode and
insulator con

gurations; means of injecting, ionizing, and

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

135

ejecting the propellants; and modes of delivery of the req-
uisite electrical power. As a result of ongoing research
and development since the late 1950s, a huge number of
such possible permutations of the electromagnetic propul-
sion concept have been studied, both experimentally and
theoretically, but only a few have survived the gauntlet
of requisite ef

ciency, reliability, range of performance,

and system compatibility to retain enduring technolog-
ical promise. Of these, most advanced are the steady
or quasi-steady magnetoplasmadynamic (MPD) thrusters,
the Hall-current accelerators, and the pulsed plasma
devices.

C. Magnetoplasmadynamic Thrusters

1. Operating Principles

As sketched in Fig. 8, the magnetoplasmadynamic thruster
(MPDT) is characterized by a coaxial geometry consti-
tuted by a central cathode, an annular anode, and some
form of interelectrode insulator. Gaseous propellants are
introduced into the upstream portion of the channel,
whereafter they are ionized by passage through an intense,
azimuthally uniform electric arc standing in the interelec-
trode gap. If the arc current is high enough, its associated
azimuthal magnetic

eld is suf

cient to exert the desired

axial and radial body forces on the propellant

ow, directly

accelerating it downstream and compressing it toward the
centerline into an extremely hot plasma just beyond the
cathode tip. Subsequent expansion of this plasma, along
with the direct axial acceleration, yields the requisite ex-
haust velocity.

Theoretically, these self-

eld accelerators can be rep-

resented in relatively simple continuum plasmadynamic
form or in more elaborate three-

uid plasma kinetic for-

mulations. Irrespective of their interior details, electro-
magnetic tensor analysis yields a generic thrust relation:

(8)

FIGURE 8

Magnetoplasmadynamic thruster (MPDT) schematic.

where

is the total thrust,

the vacuum magnetic per-

meability,

the total arc current,

and

the effective

arc attachment radii on the anode and cathode, and

parameter slightly less than unity that depends on the

ner

details of the current attachment patterns on the electrodes.
Note that this relation, which generally agrees with exper-
iments, is independent of the mass

ow rate and any other

properties of the propellant, thus the exhaust velocity must
scale as the ratio

. An important nondimensional

scaling parameter,

, which arises in several other empiri-

cal and theoretical contexts, including the onset of severe
erosion and the appearance of various modes of plasma
waves and instabilities, is calculated from the following
formula:

≡








)








(9)

where the propellant is represented by its ionization poten-
tial

and atomic mass

. Nominal MPDT operation is

achieved at

while stable and low-erosion operation

is typically limited to

ξ <

2. The lifetime of such devices

is limited by component erosion, most notably evapora-
tion of cathode material, and the performance is bounded
by losses associated with ionization and thermal energy
frozen in the

ow as well as losses in the electrode sheaths.

The modeling and control of many loss mechanisms are
complicated by the presence of plasma turbulence due to
current-driven microinstabilities which can cause exces-
sive ionization and heating.

In space applications, the optimum power range of op-

eration will be delimited on the low side by the need to
ionize the propellant fully and to keep the electrode losses
relatively small compared with the thrust power. The up-
per power limit will be set either by tolerable erosion rates
and plasma instabilities, or by the realities of the overall
system, including the available space power source and as-
sociated heat rejection equipment. At megawatt power lev-
els and corresponding propellant

ow rates, both ground

testing and space testing of these highly promising devices
present formidable technological and economic problems.
Indeed, at present there are no U.S. facilities capable of
long-term megawatt operation of steady-state MPDTs.

2. Present and Projected Capabilities

The MPDT has demonstrated its capability of provid-
ing speci

c impulses in the range of 1500

–

8000 sec with

thrust ef

ciencies exceeding 40%. High ef

ciency (above

30%) is typically reached only at high power levels (above
100 kW); consequently, the steady-state version of the
MPDT is regarded as a high-power propulsion option.

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

136

Electric Propulsion

When the thruster is operated below 200 kW, the self-
induced magnetic

eld becomes only marginally suf

cient

to provide the desired body force, and external

elds are

frequently added to enhance performance in this range.
However, in its megawatt versions, the self-

eld MPDT

has the unique capability, among all developed electric
thrusters, of processing very high power levels in a simple,
compact, and robust device that can produce thrust den-
sities as high as 10

N/m

. These features have rendered

the steady-state MPDT particularly attractive for energetic
deep-space missions requiring high thrust levels, such as
piloted and cargo missions to Mars and the outer planets,
as well as for nearer-term orbit raising missions.

In addition to the present unavailability of high power

in space, the cathode erosion rates of the steady-state
MPDT (which can be as high as 0. 2

g/C), have slowed

the evolution of steady-state MPDTs toward

ight ap-

plications. A version of the steady-state MPDT, called
the lithium Lorentz-force accelerator (Li-LFA), shown in
Fig. 9, uses a multichannel hollow cathode and lithium
propellant to substantially reduce the cathode erosion
problem while signi

cantly raising the thrust ef

ciency

at moderately high power levels. For example, a 200-kW
Li-LFA has demonstrated essentially erosion-free opera-
tion over 500 hr of steady thrusting at 12.5 N, 4000 sec

, and 48% ef

ciency. Since no other electric thruster

has yet shown such a high power processing capability,
the Li-LFA is at the forefront of propulsion options for
nuclear-powered deep-space exploration and heavy cargo
missions to the outer planets.

FIGURE 9

A 100-kW-class, lithium MPDT or Lorentz-force

accelerator.

In order to bene

t from the advantages of MPD propul-

sion on today

’

s power-limited spacecraft, the MPDT can

also be operated in a quasi-steady (QS) pulsed mode using

at-top high-current pulses long enough (

350

sec) for

a steady-state current pattern to dominate the acceleration
process. The QS-MPDT can thus bene

t from the high

ciency associated with the instantaneous high power,

while drawing low steady-state power from the spacecraft
bus. This approach was adopted in the

rst MPDT to

as a propulsion system, a 1-kW-class QS-MPDT that op-
erated successfully in 1996 onboard the Japanese Space
Flyer Unit. (Previous MPDT space

ight tests by Japan

and the Russia in 1975, 1977, 1980, and 1983 were purely
experimental.)

While no present operational spacecraft employ MPD

propulsion systems, ongoing research and development
activities in Russia and the United States on the Li-LFA,
and in Europe and Japan on the gas-fed MPDT, aim at fur-
ther improving the performance and lifetime of the steady-
state MPDT to a level that meets near-future advanced
propulsion needs.

D. Hall Thrusters

1. Operating Principles

If any electromagnetic accelerator is operated at low
enough plasma density or high enough magnetic

eld, the

current driven through it will divert from strict alignment
with the applied electric

eld to acquire a component in

the

direction

—

a form of the well-known

“

Hall ef-

fect

”

that derives from the ability of the current-carrying

electrons to execute signi

cant portions of their cycloidal

motions in the crossed

elds before transferring their mo-

mentum to the heavy particles. In extreme cases, this Hall-
current component can totally dominate the conduction or

“

Lorentz

”

component.

Low-density Hall-current accelerators exploit this ef-

fect by providing channel and

eld geometries that lock

the plasma electrons into a nearly collisionless cross-
stream drift, which leaves the positive ions free to
accelerate downstream under a component of the ap-
plied electric

eld. In a sense, such devices are hybrid

electrostatic

–

electromagnetic accelerators with space-

charge neutralization automatically provided by the back-
ground of drifting electrons. Because the magnetic

elds

in these devices are externally supplied, and because the
mass

ow densities are intrinsically low, these thrusters

optimize their performance at considerably lower powers
than those of the self-

eld MPD devices.

2. Evolution and Present Capabilities

Coaxial Hall plasma accelerators were optimized in the
Former Russia during the late 1960s to the late 1990s,

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

137

where they attained ef

ciencies above 50%. Some of the

original work on Hall thrusters was also conducted in
the United States in the early and mid-1960s, but interest
in that accelerator waned in favor of more extensive ion
thruster development, until a vigorous revival of interest
began in the United States, Europe, and Japan in 1991.

Today

’

s Hall thrusters are sometimes referred to as

“

closed-electron-drift

”

devices, given the azimuthal drift

of electrons that is common to all present variants of such
thrusters. The most common versions are the stationary
plasma thruster (SPT) (also termed the

“

magnetic layer

thruster

”

) and the anode layer thruster (ALT). The former

differs from the latter by its extended channel, the use
of insulator chamber walls, and the extent of the quasi-
neutral acceleration region, but both rely on the same ba-
sic principles for ionizing and accelerating the propellant.
A schematic of a Hall thruster of the SPT type is shown in
Fig. 10. Electrons from the cathode enter the chamber and
are subjected to an azimuthal drift in the crossed radial
magnetic and axial electric

elds, wherein they undergo

ionizing collisions with the neutral propellant atoms (typ-
ically xenon) injected through the anode. While the mag-

FIGURE 10

Schematic of a Hall thruster with an extended insulator channel (stationary plasma thruster, or SPT),

showing the external cathode, the internal anode, the radial magnetic

eld, and typical particle trajectories.

netic

eld strength is suf

cient to lock the electrons in an

azimuthal drift, it does not signi

cantly affect the trajec-

tory of the ions, which are directly accelerated by the axial
electric

eld. An axial electron

ux equal to that of the ions

reaches the anode due to a cross

eld mobility that often

exceeds classical values, and the same

ux of electrons

is available from the cathode to neutralize the exhausted
ions. Quasi-neutrality is thus maintained throughout the
chamber and exhaust beam, and consequently no space-
charge limitation is imposed on the acceleration, which al-
lows relatively high thrust densities compared with those
of conventional electrostatic propulsion devices. Nomi-
nal operating conditions of a common

ight module (e.g.,

the Russian SPT-100) operating with xenon are a 2- to
5-mg/sec mass

ow rate; a 200- to 300-V applied voltage,

yielding a plasma exhaust velocity of 16,000 m/sec; and
a thrust of 40

–

80 mN, at ef

ciencies of about 50%.

3. Applications and Flight History

Hall accelerators of the closed-drift type are at present
the most commonly used plasma thrusters. Since 1972,

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

138

Electric Propulsion

more than 110 Hall thrusters have been

own on Russian

spacecraft, and more than 52 thrusters remain in opera-
tion. They have also been used as plasma sources in active
space experiments. In view of their high speci

c impulse,

relatively high ef

ciency, and high thrust density, Hall

thrusters continue to be developed by industry and gov-
ernment for purposes of orbit insertion, attitude control,
and drag compensation. Because of the stringent require-
ments for trouble-free operation for many thousands of
hours, efforts to improve performance and lifetime are
currently under way in the United States, Russia, Europe,
and Japan. These include efforts to lower beam divergence
(which is typically between 30 and 40 degrees), to re-
duce electromagnetic interference (due to various types
of plasma oscillations), to lower erosion rates, and to in-
crease thrust ef

ciencies. Strides toward more ef

cient,

compact, and lightweight PPUs (with speci

c masses as

low as 5 kg/kW) are also being made.

E. Pulsed Plasma Thrusters

The power conservation strategy underlying quasi-steady
operation of MPDTs can be carried further by employ-
ing small power systems to drive plasma thrusters in
short (

10-

sec) bursts of high instantaneous power

(

10 MW). The energy involved is typically stored in

FIGURE 11

Various pulsed plasma thruster (PPT) con

gurations.

capacitor banks or inductor coils, then delivered rapidly
to the electrodes by some form of high-speed switch.

When a gas-fed pulsed plasma thruster (GF-PPT) is op-

erated in a predominantly electromagnetic mode, the ac-
celeration is achieved by

“

snowplow

”

action of a current

sheet driven by its self-induced Lorentz body force. Vari-
ous geometries that had been used in early thermonuclear
fusion experiments, including coaxial guns and linear and
theta pinch discharges (Fig. 11), have been modi

ed into

propulsion con

gurations.

In the early and middle 1960s, GF-PPT systems were

developed that had ef

ciencies above 20% at speci

c im-

pulses near 5000 sec, using some 65 J of stored energy per
pulse, but since the main focus of EP research at that time
was on developing

primary

propulsion systems, GF-PPT

research and development ebbed by the end of that decade
due the need for massive capacitors and the energy and
lifetime requirements of the fast-acting valves required to
ensure high mass utilization ef

ciency.

However, one derivative of this concept survived in the

form of the simple ablative pulsed plasma thruster (APPT),
which promised to solve the mass utilization problem
without the use of valves, and to save system mass by shed-
ding the complex mass storage and control systems of their
gas-fed counterparts. In the APPT, shown schematically
in Fig. 12, the surface of a polymer block (most commonly

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

139

FIGURE 12

Schematic of an ablative pulsed plasma thruster

(APPT).

on) is successively eroded by intermittent arc pulses

driven across its exposed face, and the ablated material is
accelerated by a combination of thermal expansion and
self-

eld electromagnetic forces.

The APPT has the claim of being the

rst EP system

to operate in orbit, when the 1964 Soviet Zond-2 space-
craft used six Te

on APPTs for sun pointing control. The

United States followed in 1968 with the LES-6 satellite,
which used four APPTs for east

–

west station keeping

(EWSK). Since then, APPTs have had a sporadic history
of application and, except for a small number of exper-
imental suborbital and orbital tests by the United States
and China, they have been used on only a series of

U.S. Navy satellites launched in the late 1970s and early
1980s. It was not until the mid-1990s, in the context of
power-limited small satellites, that APPT research and
development were rekindled. Improved capacitor tech-
nology, combined with the simplicity of the APPT and
its propellant storage and feed system, and its capabil-
ity of providing small and precise impulses at high spe-
ci

c impulse and arbitrarily low spacecraft power, made

it suitable for many attitude-control chores on power-
limited small satellites. A small but growing number of
upcoming U.S. missions using APPTs have been planned,
which will use

ight-ready modules such as that shown in

Fig. 13.

Two of the most severe def

ciencies of APPTs, namely

their low ef

ciency (

15%) at low pulse energies and

spacecraft contamination by the polymer products in the
plume, have spurred some revival of their gas-fed pro-
genitors. GF-PPTs have the advantages of compatibility
with a wide range of propellants, cleaner exhaust, and
a wider scalability of performance. Recent advances in
low-inductance, high-frequency, and high-current pulsing
technologies have relieved the low mass utilization ef

ciency problem that plagued the 1960s prototypes. Further
improvements of these devices will depend on a better un-
derstanding of the nature and scaling of the complex dis-

FIGURE 13

ight-ready Te

on ablative pulsed plasma thruster

(APPT) module using two thrusters, positioned on the opposite
ends of the thrust axis. [Courtesy of the Primex Corporation.]

sipative mechanisms in such unsteady

ows, and on the

the ability to control and abate the canting and instability
of the accelerating current sheets.

F. Inductive Thrusters

As with the electrothermal arcjets, the most vulnerable ele-
ments of all the electromagnetic thrusters described above
are the electrode surfaces. Hence there has been some on-
going interest in a variety of inductive possibilities that
require no electrodes directly exposed to the intense dis-
charge environment. Again, most of these concepts have
been transposed from other technologies, including pulsed
inductive discharges, traveling wave accelerators of var-
ious classes, RF fringe-

eld accelerators, and cyclotron

resonance devices. All involve inherently unsteady

ows

over a wide range of operating frequencies, and all must
trade off the absence of electrode erosion against generally
poorer coupling ef

ciency between the external circuitry

and the accelerating propellant plasma. Like the inductive
electrothermal machines, their overall systems suffer from
the more complex and massive power processing equip-
ment needed to drive them, although some of this disad-
vantage has been ameliorated by recent improvements in
solid-state electronic technology, so that their future may
be somewhat brighter than their past.

V. SYSTEMS CONSIDERATIONS

A. Power Conditioning

Almost all electric thrusters, except resistojets, operate
at voltages larger than those provided by the standard

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

140

Electric Propulsion

bus of solar-powered spacecraft. This necessitates the use
of power processing subsystems that transpose the pri-
mary power from the space-borne source to the requisite
voltages, currents, and duty cycles of the given thrusters.
The PPU can have much in

uence over the overall ef

ciency, reliability, and mass of the total propulsion system.
Aside from voltage conditioning, each class of thruster
presents its own demands on the PPU. Many electrother-
mal arcs, for example, can develop negative slopes in
portions of their voltage

–

current characteristics that are

tantamount to negative impedance for the power source
and must be suitably ballasted if the arc is to be prop-
erly controlled. Ion thrusters require a broad range of
electrode voltages and currents for their discharge an-
odes, hollow cathodes, and accelerating grids and must be
protected against high-voltage shorts and insulator break-
downs. High-power MPD accelerators require huge cur-
rents at relatively low voltages, and their unsteady versions
present an additional overlay of energy storage and pro-
cessing requirements. Almost all electric thrusters require
some ignition mechanism that inevitably complicates the
power package.

Only the most mature thruster concepts have had signif-

icant

ight-quali

ed PPU development. For a hydrazine

arcjet, whose typical operating voltage peaks at about
100 V, the PPU is about 91% ef

cient and has a speci

mass of about 2.5 kg/kW. A photograph of four

ight-

ready arcjets and their PPU is shown in Fig. 14. The
PPU penalty becomes worse for higher-voltage devices
such as the xenon Hall thruster and the xenon-ion engine,
whose PPUs have ef

ciencies of 93% and 88%, respec-

tively, and speci

c masses as high as 10 kg/kW. Almost all

electric thruster PPUs developed to date are for use with
solar panel power sources and would need to be recon-
sidered when nuclear and other higher-power sources are
deployed.

FIGURE 14

Four

ight-ready arcjets with their power processing

unit (PPU). [Courtesy of the Primex Corporation.]

B. Primary Power Sources

All this power conditioning technology must relate to a
primary source that is itself reliable, compatible, and of
suf

ciently low speci

c mass to function in the given space

application. Although various nuclear, chemical, and solar
thermal conversion cycles have been studied as potential
spacecraft power sources, practically all present satellites
rely exclusively on solar panels and batteries for power
sources. Consequently, progress in photovoltaic cell tech-
nology is critical to the continued growth of EP appli-
cations on near-earth spacecraft. Present standard silicon
solar cells cost about $1500/W and have a power density
of about 140 W/m

, which corresponds to a speci

c power

of 40 W/array kg. The most likely near-term improvement
on these are gallium arsenide (GaAs) cells, which have
been proven to yield 220 W/m

. When such solar cells

are augmented with aluminized Mylar concentrators, they
promise speci

c powers of 100 W/array kg, which should

greatly enhance the proliferation of EP systems. Another
critical improvement in solar power technology is the de-
crease of solar cell degradation from accumulated envi-
ronmental radiation dose. This is particularly crucial for
the EP missions (such as orbit raising) which require sub-
stantially longer transfer times than those of the impulsive
maneuvers effected by high-thrust chemical rockets.

While for many near-earth missions of relatively mod-

est

solar or chemical energy sources may be viable,

for the interplanetary, megawatt scale of operations, nu-
clear power systems of major dimensions will inevitably
be required, which will further complicate the total system
with attendant environmental and safety hazards.

VI. APPLICATIONS

Although the primary motivation for development of
space-worthy EP systems is the conservation of propel-
lant mass for missions of large characteristic velocity in-
crements, electric thrusters offer a number of attractive
secondary operational bene

ts, including precision and

variability of thrust levels and impulse increments, gener-
ous shutdown and restart capabilities, and the use of chem-
ically passive propellants. Their major limitations are the
need for sophisticated external power sources, very low
to modest thrust density capabilities, and little empirical
experience with unattended operation in the space envi-
ronment. All these characteristics serve to circumscribe
the classes of missions for which EP may reasonably be
considered.

Obviously, the limited thrust densities predicate thrust-

to-mass ratios for electric systems that are not propi-
tious for rapid maneuvers in strong gravitational

elds.

There are no launch or ascent

–

descent capabilities near

P1: ZBU Final Pages

Encyclopedia of Physical Science and Technology

EN005C-201

June 15, 2001

20:23

Electric Propulsion

141

planetary surfaces, and even outer orbit transfer exercises
can be performed only very slowly over gentle spiral tra-
jectories. Thus, near-planet applications will be limited
to those attitude-control, station-keeping, drag-reduction,
and modest orbit-changing functions (such as orbit phase
changes in LEO constellations) where the minuteness and
precision of thrust, propellant conservation, and long life-
time give them superiority over chemical options. Many
such applications have been recognized and evaluated, and
in several cases appropriate electric systems have been, or
soon will be, deployed.

In the domain of interplanetary

ight, however, EP of-

fers much more substantial advantages over chemical sys-
tems, which extend in several important cases to enabling
missions that simply could not be performed by means
of any other reasonably projected propulsion technology.
These include heavy cargo and/or piloted missions to Mars
and the outer planets and many unpiloted probes beyond
the solar system and out of the ecliptic plane.

Comparison of electric and chemical systems for any

ambitious mission, piloted or unpiloted, reveals that the
basic dynamical distinction is between essentially impul-
sive thrust increments provided by the latter and much
more protracted, lower-level thrust pro

les necessarily,

and in some cases advantageously, provided by the former.
Analytical optimization of extended variable thrust trajec-
tories for interplanetary transportation is a complex task,
far from fully rendered in contemporary mission analy-
ses. Superimposed on the strictly dynamical aspects are a
host of systemic considerations, such as the importance of

ight time to crew, other payload considerations, internal

and external environmental hazards, utility of the power
supply at destination, fraction of payload to be returned,
secondary mission goals in

ight, human and structural

compatibility with

ight maneuvers such as aerobraking

and swingby, and in-

ight adjustment, service, and emer-

gency return capabilities. Since many of these, especially
those involving human factors, currently have inadequate
fundamental databases and theoretical representations, the
composite mission assessments are shaky, at best, and
much more sophisticated analyses will be required before
de

nitive mission projections.

Finally, as already mentioned, before any such ambi-

tious EP missions can seriously be contemplated, nonso-
lar alternatives for high-power sources in space must be
developed. For mostly political reasons, plans for deploy-
ment of nuclear high-power sources in space have so far
failed to materialize, and consequently the use of electric
thrusters for primary propulsion in energetic missions has
had a cyclical history of false starts and disappointments.
Indeed, the recent vigorous rejuvenation of the

eld of

electric propulsion can be attributed, at least in part, to
a conscious shift in emphasis away from the high-power
missions envisaged during the 1960s and 1970s toward
the less ambitious but more realistic power-limited small
satellites of today. Now that many EP systems have en-
tered the mainstream of astronautic technology, their role
in helping to expand human ambition beyond the inner
part of the solar system, although still dependent on the
hitherto unrealized development of high-power sources,
is perhaps on more credible ground.

SEE ALSO THE FOLLOWING ARTICLES

IQUID

OCKET

ROPELLANTS

•

AMJETS AND

CRAM

JETS

•

OCKET

OTORS

, H

YBRID

•

OCKET

OTORS

IQUID

•

OCKET

OTORS

, S

OLID

•

OLID

ROPELLANTS

•

PACECRAFT

HEMICAL

ROPULSION

•

PACECRAFT

YNAMICS

BIBLIOGRAPHY

Jahn, R. G. (1968).

“

Physics of Electric Propulsion,

”

McGraw-Hill, New

York.

Stuhlinger, E. (1964).

“

Ion Propulsion for Space Flight,

”

McGraw-Hill,

New York.

A special issue of the

Journal of Propulsion and Power

, Vol. 14,

No. 6, 1998, contains review articles on general EP, EP for solar system
exploration, ion thruster development, PPTs, Hall thrusters, MPDTs,
high-power arcjets, and

eld emission micropropulsion.