Nuclear Propulsion

IAA COMMISSION III
SG 2 – NUCLEAR SPACE POWER AND

PROPULSION
M. Auweter-Kurtz
C. Bruno
D. Fearn
H. Kurtz
T.J. Lawrence
R.X. Lenard
List of contents
Introduction 8
1. Physics of Nuclear Propulsion – An Introduction 11

1.1. ABSTRACT 11
1.2. Introduction 11
1.3. Fundamental Physics 12
1.3.1. Forces 12
1.4. Propulsion 19
1.4.3. Power 23
1.4.4. Mass 25
1.5. Nuclear Propulsion Strategies 27
1.5.1. Nuclear Thermal Rockets (NTR) 27
1.5.2. Nuclear Electric Propulsion (NEP) 31
1.5.3. A Comparison between Chemical and NTR/NEP Isp 32
1.6. Massless (Photonic) Propulsion 33
1.7. Conclusions 34
1.8. References 35
2. Nulcear Thermal Rocket Propulsion Systems 38
2.1. ABSTRACT 38
2.3. System Configuration and Operation 41
2.4. Particle-Bed Reactor 46
2.4.1. CERMET 47
2.5. Safety 49
2.6. MagOrion and Mini-MagOrion 51
2.7. Conclusions 53
2.8. References 54
3. The application of ion thrusters to high thrust, high specific impulse
nuclear-electric missions 57
3.1. ABSTRACT 57
3.3. Background 62
3.3.1. Space Nuclear Programmes 62
2
3.3.2. Advantages of Electric Propulsion 63
3.3.3. Propulsion System Parameters 64
3.3.4. Propulsion Technology Review 66
3.3.4.1. Gridded Ion Engines 66
3.3.4.2. The Hall-Effect Thruster 69
3.3.4.3. Magnetoplasmadynamic (MPD) Thrusters 71
3.3.4.4. Variable Specific Impulse Magnetoplasma Rocket (VASIMR) 72
3.4. Gridded Ion Engines; Current Devices 73
3.4.1. Ionization Mechanisms 73
3.4.1.1. Radiofrequency Discharge Thrusters 73
3.4.1.2. Radiofrequency Discharge Thrusters Operating at High Frequencies. 75
3.4.1.3. Kaufman-Type Direct Current Discharge Thrusters 76
3.4.1.4. Magneto-electrostatic Containment (MESC) DC Thrusters 77
3.4.2. Current Gridded Thrusters 79
3.4.3. Summary of Capabilities of Existing Gridded Thrusters 90
3.5. The Scaling Process 93
3.5.1. Grid Design Options 93
3.5.2. Exhaust Velocity and SI 96
3.5.3. Thrust and Perveance 97
3.5.4. Current, Power and Thrust Densities 98
3.5.5. The Discharge Chamber 99
3.6. High SI, High Power Operation 100
3.6.1. Parametric Variations 100
3.6.2. Operation at 100s of kW Level 101
3.6.2.1. Increase of Thruster Diameter 101
3.6.2.2. Increase of SI 102
3.6.2.3. Increase of Perveance 103
3.6.2.4. Technology Readiness 104
3.6.3. Operation at the MW Level 105
3.6.3.1. Specific Designs at the MW Level 107
3.6.3.2. Peak Thruster Performance 107
3.6.3.3. Propellant Selection 108
3.6.3.4. Parameter Ranges 109
3.6.4. Systems Considerations 110
3
3.7. Conclusions 111
3.8. References 114
3.9. List of Symbols and Acronyms 123
4. HIGH POWER AND HIGH THRUST DENSITY ELECTRIC
PROPULSION FOR IN-SPACE TRANSPORTATION 125
4.1. Abstract 125
4.3. Thermal Arcjets 126
4.3.1. Operational Principle 126
4.3.1.1. Efficiency. 127
4.3.1.2. Discharge Voltage. 129
4.3.2. Theory and Numerics 129
4.3.3. System Description 130
4.3.4. Influence of Propellants 131
4.3.5. Lifetime Limiting Factors 132
4.3.6. Qualification Advantages 133
4.3.7. Existing Arcjet Thruster Technologies 133
4.3.7.1. USA 133
4.3.7.2. Germany 133
4.4. Magnetoplasmadynamic Thruster 134
4.4.1. Self-Field (SF-MPD) Thrusters 135
4.4.1.1. Importance of Electrode Design. 141
4.4.1.2. Electrode Voltage Drops and Loss Distribution. 142
4.4.1.3. Electrode Erosion. 144
4.4.1.4. Electrode Losses and Cooling. 145
4.4.2. Magnetoplasmadynamic Thruster with Coaxial Applied Field 148
4.4.2.1. Acceleration Mechanism. 148
4.4.2.2. Hall Parameter. 149
4.4.2.3. Current Distribution 149
4.4.2.4. Movements of Charged Particles in E and B Fields. 150
4.4.2.5. Rotational Frequency of the Plasma. 151
4.4.2.6. Thrust. 152
4.4.2.7. Experimental Evidence. 153
4.4.2.8. Applicable Propellants. 155
4
4.4.2.9. Numerical Simulation. 155
4.4.2.10. Thruster Developments. 156
4.4.2.11. Necessary Test Facilities 158
4.5. High Power Hybrid Thruster Concept ATTILA 159
4.6. Summary 160
4.7. References 161
5. A review of reactor configurations for space nuclear electric
propulsion and surface power considerations 167
5.1. ABSTRACT 167
5.2. Reactor introduction 167
5.3. Reactor REQUIREMENTS 168
5.4. Reactor And Power Conversion - Mass Comparison 169
5.4.1. Earlier Work and Reactor Descriptions 170
5.4.2. Static Model Development and Validation 172
5.4.2.1. Model Basis and Overview 172
5.4.2.2. Specific Design Anchor Points – Level Field. 173
5.4.2.3. Model Validation and Results. 181
5.4.3. A Simple Shield Model 183
5.4.4. Shield Parametrics 186
5.5. Reactor operation startup 189
5.6. Summary 190
5.7. References 191
6. Nuclear Safety, legal aspects and policy recommendations 192
6.1. ABSTRACT 192
6.2. Finding 1: Nuclear power and energy have significant roles in space
exploration now, and the future for nuclear power has exceptional
potential for future space exploration activities. 192
6.3. Finding 2: In order for the great potential advantages of nuclear
propulsion to be realized, it must be perceived by a majority of the
population to be safe. 193
5
6.4. Finding 3: Existing policies and procedures are generally adequate to
account for requirements of public safety and environmental
compliance. Some recommendations will assist in clarifying the
meaning of some of these procedures, principles and policies to aid in
the space systems engineering process. 194
6.4.1. History, Perspectives and Objectives 194
6.4.2. Safety Guidelines and Implementation. 195
6.5. Finding 4: The existing design, fabrication and test process, including
safety analyses is adequate for addressing all non-launch related
safety and environmental issues for a space nuclear reactor system;
launch and space related protocols must be developed. 201
6.5.1. Recommendation 16: The Space Nuclear Reactor Program Should
Concentrate on Major Post-Shipment Activity and Accident
Categories 202
6.5.2. Arrival at Launch Site: Possible Scenarios 203
6.5.3. Spacecraft Final Assembly and Checkout: Possible Scenarios 203
6.5.4. Launch Preparation and Countdown: Possible Scenarios 204
6.5.5. Early Launch Phase: Possible Scenarios 205
6.5.6. Late Launch Phase: Possible Scenarios 205
6.5.7. Finding 5: The Safety and Operations Phase for NEP or NTP Systems
Should be Developed so as to Maximize Possible Scenarios for Space
Nuclear Reactor Employment 206
6.5.8. Recommendation 17: Hazards and Risks Definitions Should be
Presented in a Consistent Fashion Throughout the Space Nuclear
Reactor Program 207
6.5.9. Scenario Bins From Cassini FSAR Format 208
6.5.10. Conversion of Hazard Categories From Cassini FSAR Format To
Hazards Table Format 209
6.5.11. Failure Events That Could Result In Release Of Radiation/Nuclear
Material 210
6.6. Finding 7: Safety Assessment for Additional Risks Posed By Lunar
and Mars Base Mission Scenarios Indicate that Space Reactor
Systems can be used Safely and Effectively on the Surfaces of Other
Celestial Bodies 211
6
6.6.1. Finding 8: A definable set of hazards for a surface nuclear reactor
power system can be delineated and risks can be effectively mitigated 212
6.6.2. Primary Differences Between Cargo and Robotic Payload NEP system
and a Moon or Mars Base Power System that can Impact Safety 213
6.6.2.1. Potential Safety Issues Associated with the Mars Base Power System 213
6.6.2.2. Initiating Events for Mars Base Power System 214
6.7. Finding 9: There appears to be no reason that a space nuclear reactor
power system cannot be safely deployed and operated on the surface
of another world while maintaining standards of planetary protection. 216
6.8. Finding 10: A Space Reactor System Enables Effective of Design
Options in Mitigating Potential Radiation Releases 216
6.9. Finding 11: A Transparent and Systematically Traceable Space
System Safety Test and Analysis Program Must be Conducted to
Ensure Crew and Public Safety 216
6.9.1. Safety Testing 217
6.9.2. Propellant Explosion and Fire Tests 217
6.9.3. Impact Tests 218
6.9.4. Reentry Tests 219
6.9.5. Launch Environment Tests 219
6.9.6. Criticality Tests 220
6.10. Finding 11: Prior Space Reactor Programs Expended Resources on
Destructive Disassembly Testing for Low-Probability Incidents –
System Level Testing Should be Reserved for More Likely Scenraios 220
6.10.1. Safety Analysis 221
6.10.2. Neutronics 221
6.10.3. Shielding 221
6.10.4. Fires and Explosions 221
6.10.5. Intentional Disassembly 221
6.10.6. Reentry Analysis 222
6.10.7. Coupled Impact/Reactivity Analyses 222
6.10.8. Risk/Consequence Analysis 223
6.10.9. Other Analysis 224
7
6.11. Finding 19: An Integrated Approach for Performance and Safety
Analysis and Testing is Critical to a Cost-Effective Development
Program 225
6.11.1. Safety Program Organization 225
6.12. Finding 20: The Ultimate Objective of All Programmatic Activities is
to Obtain Launch Approval for the Space Reactor System – the
Program Should be Structured to Attain the Goal. 227
6.13. Summary and Recommendations 229
6.14. References 230
7. Appendix 232
8
Introduction
This Study was prepared for Commission 3 (Space Technology and System Development) of the
International Academy of Astronautics (IAA) with the intention of providing, in one place, a
summary of technical and safety information about nuclear power and propulsion systems for space
applications. By agreement, emphasis was to be put on future systems and future interplanetary
missions, including manned missions. This study was suggested to IAA by M. Pouliquen (of
SNECMA, now Safran-SNECMA) and after being proposed, was accepted by Commission 3 as
Study No. 3.2. After a prolonged incubation and discussion period, the Study Group 3.2 members
(M. Auweter-Kurtz and H. Kurtz, C.Bruno, D. Fearn, T. J. Lawrence and R.X.Lenard) issued a first
draft in April 2005. This draft was submitted to a review panel chosen by the Academy.
What follows is the final version of the Final Report, after incorporating all suggestions and
comments made by the IAA panel. The authors wish therefore to acknowledge all peer reviewers
for their time, commitment and comments, sometimes very detailed and always useful.
The structure of this Report is as follows: Chapter 1 contains fundamental physics of nuclear energy
and sets the stage for its applications in space propulsion and power generation. The two main
classes of propulsion systems that in space may take advantage of nuclear energy (thermal and
electric rockets) are briefly introduced. Chapter 2 discusses nuclear thermal propulsion, where
energy is transferred directly to a propellant as in chemical rockets. Chapter 3 and 4 deal with the
second class of systems, where thrust is produced by converting nuclear power into electric power.
Chapter 3 illustrates current and projected capability of the first type of nuclear-powered
electrostatic thrusters, that based on Coulomb acceleration (ion thrusters). Electrodynamic
thrusters, those using the Lorentz force to accelerate plasma with a self-created or applied magnetic
field, and arcjets are discussed in Chapter 4. Chapter 5 focuses on the architecture of nuclear
reactors for in-space electric propulsion and for power generation, for instance on the surface of the
Moon or Mars. Chapter 6 discusses legal/legislative aspects of operating devices based on nuclear
power. An Appendix is dedicated to the issue of radiation hazards, including a primer on
fundamentals of fission radiation and doses.
Not included in this report are Radioactive Thermionic Generators (RTG). In fact, they are the
subject of an excellent AIAA paper by G.L. Bennett presented in June 2006*. They have been
extensively used whenever solar panels were impractical and are relatively well known, therefore
representative of current technology rather than future. Besides, their power output is orders of
magnitude lower than needed for the practical (fast) interplanetary missions of interest to this study.
Pulsed nuclear propulsion, i.e., propulsion by miniature nuclear explosions, recently revived and
currently simulated by means of plasma zeta-pinch devices, is concisely summarized in Chapter 2
but not discussed in detail. Its origin, the US Orion Project, is reviewed in recent literature, as is its
resurrection in the form of the Magnetic Orion (MagOrion) and the more recent MiniMagOrion.
The members of this study group regret the lack of contributions from the former USSR.
Substantive information (in Russian) on NTR became finally available through the good will and
efforts of Russian colleagues only at the time of this writing, and in the interest of time it was
decided to set it aside, waiting for a more propitious opportunity to make it known in an English
translation.
________________________________________________________________________________
* Bennett, G.L., (2006), Space Nuclear power: Opening the Final Frontier”, AIAA paper 2006-
4191, presented at the 4th Int. Energy Conversion Engin. Conf., San Diego, CA, 26-29 June 2006.
9
By definition, a report to the IAA does not represent a position by the IAA, nor wants to take a
specific posture vis a vis the issues discussed. By and large the members of this study group
intended to paint as accurate as possible a picture of this emerging, or re-emerging, technology and
of its potential for future space exploration. Within their limited resources (mainly, their available
time and energy) they hope their effort will be appreciated. Reviewers chosen by IAA made sure
the intention above was indeed clear in this Report, and added or suggested further material: these
authors wish to acknowledge their time and effort, as well as that of Prof. Gorshkov and Prof.
Popov, who, after much effort, provided historical and current information about the Russian space
propulsion program.
As this study was progressing, growing company responsibilities and health considerations
eventually forced M. Pouliquen to renounce his Chairmanship in 2003. It fell on this writer to keep
this study on the steady course Marcel started. A second sad event was the sudden demise of David
Fearn at the end of August 2007, while this team was waiting for the decision of AIAA concerning
publication. It is a pity Dave is not with us to see the oucome of our joint work. I want to
acknowledge all Commission 3 members who gently prodded me during this study, and Marcel, for
his intuition, common sense, and calm leadership. Finally, I want to acknowledge the dedication,
time and patience of Ing. Giuseppina De Flora in assemblying this report and making sure it was as
intended.
Fall 2007
Claudio Bruno
Study Group members:
M. Marcel Pouliquen, SAFRAN/SNECMA, France

Prof. Claudio Bruno, University of Rome “La Sapienza“, Italy
Prof. Monika Auweter-Kurtz, Universitaet Stuttgart, IRS, Germany
Dr. David Fearn, EP Solutions, UK
Dr. Helmuth Kurtz, Universitaet Stuttgart, IRS, Germany
Col.. Timothy J. Lawrence, Professor, US Air Force Academy, USA
Dr. Roger X. Lenard, IOSTAR and Sandia National Laboratories, USA
IAA Commission 3 members:
Dr. Hans Hoffman,
Dr. Horst Rauck,
Dr. Andre’ Van Gaver, ESA-ESTEC, The Netherlands
10
1. Physics of Nuclear Propulsion – An Introduction
1.1. ABSTRACT
With renewed interest in nuclear propulsion due to the JIMO missions and the Space Exploration
Initiative in the US, there is also a need for propulsion engineers to revisit the basic physics
associated to nuclear propulsion. Accordingly, starting from the description of the fundamental
three forces, the corresponding sources of energy, from gravitational to nuclear, are discussed and
their energy density ranked, showing that nuclear propulsion is the only practical alternative to
chemical for fast, deep space and interplanetary exploration. Depending on specific missions, the
nuclear engines mass consumption (specific impulse) may be modulated, and their thermal material
issues may be resolved by addition of inert mass (propellant) to the products of nuclear reactions.
Generally, relativistic effects must be accounted for in calculating performance of nuclear
propulsion systems: depending on mode of utilizing their nuclear energy source, they may produce
jet exhaust speeds non-negligibly small with respect to the speed of light. Choice between direct
thrust generation, as in nuclear thermal propulsion, or indirect, as in nuclear-electric propulsion, are
discussed, and examples of past experience with NTR in the US and the former Soviet Union are
reported, indicating this propulsion technology is viable as it is. Finally, nuclear so-called
‘massless’ propulsion based on photonics is also illustrated, showing that it too has specific impulse
limitations, due to the relativistic mass conversion into energy.
1.2. Introduction
The very idea of using nuclear energy for propulsion goes back to Goddard, who mused in
1906-7 about using radium as a source, and realized, correctly, its power was insufficient. In
Europe, Esnault-Pelterie concluded that nuclear energy was indispensable for space travel in a
conference he held in France in 1912 [FAS, 2005], so the concept of nuclear propulsion (NP) is not
a new topic, going back a century ago.The main purpose of this chapter is to illustrate characteristic
features of nuclear propulsion (NP) from the viewpoint of basic physics, and to orders of
magnitude. At a time when nuclear propulsion is again being revived, e.g., to power JIMO
missions, and its current technology and societal issues are discussed, e.g., see [Aerospace America,
2004], it is probably useful to review its fundamental principles. References on history of NP in the
US and in the Soviet Union may be found in Chapter 2 dealing with Nuclear Thermal Propulsion.
Fusion propulsion, a still conceptual form of NP, is in [Kammash, 1995]. Radiation and its
biological effects are discussed in [Del Rossi and Bruno, 2004]. A concise review of fission reactors
technology is included in Chapter 2 by T.J. Lawrence and again in Chapter 5 by R.X. Lenard. The
physics of pulsed nuclear propulsion, e.g., propulsion by means of nuclear explosions [Schmidt et
al, 2002; Dyson, 1979; Dyson, 2002] is briefly mentioned in this Chapter, although acceptance by
the public of this concept seems, at this time, remote.
Most, or all, of the ideas or concepts below were clear to researchers working at LASL,
Aerojet, General Electric or the Kurchatov Institute in the ‘60s. However, this is not necessarily so
for the majority of propulsion scientists and engineers, as this author realized during an Energy
Conversion Conference held in Istanbul in 2004. Besides, misconceptions (and fears) still exist that
may cloud understanding the basics of NP.
The second goal of this chapter is to explain why NP is physically inevitable if we want to
explore our Solar System and, perhaps, beyond.
11
1.3. Fundamental Physics
1.3.1. Forces
There are only three forces in nature according to the Standard Model of modern physics.
This might change in the years to come, e.g., if the nature of dark energy, or vacuum energy
fluctuations will be fully understood. However, potential new developments will strengthen, if
anything, the considerations below.
The three forces are the gravitational force; the so called electro-weak force (unified from
electrodynamic and weak by the work done with large particle accelerators since the ‘60s, and
standing ongoing theoretical scrutiny and experimental tests); and the nuclear force binding
nucleons and holding the nucleus together.
Gravitation is well known for its macroscopic effects, much less at the fundamental level. Of
the two remaining forces, the electroweak is responsible for Coulomb attraction and repulsion
among charges, thus for stable structures composed of electrons and nuclei (the atoms), and therfore
for the existence of bonds among atoms and molecules. The third force is the nuclear, or ‘strong’,
force. It acts at very short distances (of order of the nucleons size, 10-13cm, or 1 fermi), pulling
together neutrons and protons and allowing nuclei to exist. This force is the strongest known.
Quantum mechanics has been very effective in clarifying the nature of the last two forces, while
cannot ‘explain’ gravitation.
A force field, in modern sense, is the result of exchange of mediating particles (the action at
a distance postulated by Newton is not part of modern physics), and is always associated with
potential energy. It is this potential energy that, released and converted into different forms, can
eventually produce the macroscopic force called thrust.
The physics assumed in what follows is based on Newton’s Third Law (Action is equal to,
but has opposite sign of, Reaction) and on Relativity. Both were and are still questioned, e.g., see
[Cornille, 1999], but their validity will not be discussed or challenged here.
1.3.2. Potential Energy
The Theory of Special Relativity [Einstein, 1916; Lang, 1999] shows that formulating
velocity (and momentum) as a 4-component vector rather than via the usual three components
results in the invariance of the square of its magnitude. The fourth component added by relativity
theory is c dt/ds, where c is the speed of light and ds is the ‘distance’ between two ‘events’ [Harwit,
1973, Chapter 5]. This distance generalizes the concept of classical distance between two points in
space by including also the effect of motion in time. The new resulting space-time is called
Minkowski’s space. In this new space there is no privileged frame of reference; the laws of physics
(the Maxwell equations of electromagnetism in particular) are no longer invariant in this space for
classical Galilean transformations (where x = x’ + Vt’, with V the constant relative velocity
between the two frames of reference x’y’z’t’ and xyzt), but only for Lorentz transformations,
where, for instance
x = ( x ' + Vt ) (1 − V 2
c2 )
and similarly for y and z, and where
12
(
t = t ' + V ( x' c2 ) ) (1 − V 2
c2 )
Since the momentum p = m0 V (1 − V 2 c 2 ) , with V now a four-component vector, and

r
r
energy E = p c , the total energy of a rest mass m0 (‘at rest’ means with zero velocity in the frame
of reference used) becomes
Total energy = m0 c 2 (1 − V 2
c2 ) (1)
For a mass at rest V = 0, and equation (1) states then that its total energy, m0c2 , is only
potential energy, a famous result implying equivalence between mass and energy to a factor c2. If
V/c is nonzero, but still << 1, Taylor expansion of equation (1) yields
Total energy = m0c2 + ½ m0V2 + …
showing that at V << c energy may be separated into potential energy and kinetic energy.
More generally, for any V/c, the Theory of Special Relativity [Einstein, 1916; Lang, 1999] states
that total energy may be written as the sum of potential and kinetic energy (KE) as
Total energy = mc2 + KE (1a)
where, until 1948, Einstein called m the relativistic mass, that is, the rest mass, mo, divided
by 1 − (V / c) 2 ; c is the speed of light, about 3 x 108 m/s. In Newtonian mechanics mass and
energy are separate quantities, each independently ‘conserved’ in any isolated system; in Einstein
mechanics they are interchangeable, c being simply a constant. What is conserved is the total
energy: if kinetic energy increases, it must be at the expense of potential. Because c2 is ‘large’ on
the human scale, the energy potentially available is also ‘large’, if it can be tapped. Equation (1a)
states the maximum energy available, just as does thermodynamics in classical chemistry. What
fraction, α, can be actually extracted will depend on its kinetics, i.e., on the specific physical steps
of conversion.
Rest mass and relativistic mass coincide for V << c, a condition invariably met in chemical
propulsion. However, electric thrusters can accelerate propellant to exhaust V = Ve far larger than
chemical rockets. Future Isp of order 105 s correspond to V/c of order 1%, where relativistic effects
begin to be appreciable. A self-consistent description of relativistic gasdynamics, as well as that of
continuum mechanics, is still not available: ‘interpretations’ are a source of lively discussions. For
this and other reasons, only mass will be specified as either “rest” or “relativistic” in what follows.
The actual potential energy of a mass m of reactant(s), of any type (chemical, fissile,
fusible,…) is therefore the fraction α mc2 :
Reactant(s) potential energy = α m c2 (2)
where the type of conversion determines the actual value of α and thus the energy available
per unit reactant(s) mass.
Through a sequence of steps (that is, nuclear kinetics) potential energy in equation (2)
becomes microscopic kinetic energy of products (e.g., chemical species, photons, or nuclides,
neutrons, photons,). This kinetic energy may be the result of several conversion steps (for instance,
thermal to electric), or even straightforward thermodynamic expansion, followed by collisions
between products and solid or magnetic boundaries. Whatever the conversion path taken, to
13
produce thrust the end result must become orderly motion of particles ejected at exhaust speed Ve
(≡ V). At some stage during this chain of steps, ‘inert’ (non-reacting) mass, Mp, may be added or
present. One very good reason is to moderate temperature, but just as important for certain
applications is to increase momentum of the total mass ejected, that is, propulsion thrust.
How much mass converts into energy? The answer depends on the kind of potential energy
tapped. Different forcefields (potentials) are characterized by very different α:
Gravitational energy (according to Newtonian theory, not General Relativity) is associated,

for instance, to the force F = G m1 m2 r/r3 between two masses at distance r. Mutual attraction force
can perform work, so the potential energy E can be obtained from ∇E = -F, that is, E = G m1m2/r.
This shows that if masses were pointlike, the energy would tend to ∞ when r → 0. In fact detecting
changes of F with r needs very sensitive instrumentation: G = 6.67 x 10-11 m3 kg-1 s-1, so per pair of
unit masses at unit distance, the gravitational energy is exactly G, a very small number (this specific
energy is about 10-18 that released by combustion of 1 kg of H2 with O2). The gravitational potential
is detectable only when the product m1 m2 is large enough to compensate for the very small ratio
G/r. Typically, this is the case when massive bodies (stars and planets) are involved: e.g., near
Earth, its mass being of order 1024kg.
Instead of force between on-board masses, spacecrafts have used very successfully the
potential associated to celestial bodies, in gravity-assist maneuvres (planet swing-bys): multiple
swing-bys in the Solar System can achieve ΔV of order 25-30 km/s.
In terms of General Relativity, gravitation is due to curvature of space-time. Accordingly,
‘warping’ space-time would produce substantial, although speculative, propulsion effects.
The mass change Δm in converting completely gravitational potential into some other form
of energy is found by imposing
E = G m1m2/r = Δm c2
Hence α =Δm/m1, per unit m2 and distance, is of order G/c2 ~ 10-27, an exceedingly small
fraction. This too shows that on-board gravitational energy is impractical as a means of energy
generation and propulsion.
The electro-weak force potential drives nearly all aerospace propulsion systems. It is
associated with the Coulomb force F = k q1q2/r2 between two electric charges q1 and q2. In the
International System of units the constant k = 8.98 x 109 C -2 m-2, much larger than G. The constant
k is the force, in N, between two pointlike unit charges set at 1 m distance. The potential E is in this
case k q1q2/r. The Coulomb force is responsible for the existence of atoms and molecules, so re-
arranging their structures (in chemistry this is called reacting) may release energy, because bonds
(forces) and their potentials change. Macroscopically, this potential is known as chemical potential:
in propulsion applications its decrease corresponds to the heat release increasing temperature and
pressure in a combustion chamber. Chemical rockets use this type of energy.
The fraction α of the mass converted into energy by chemical reactions is much larger than
in gravitation, but still very small. Its value depends on the specific reaction, but it may be
approximately gauged from calculating the Δm/m of an ideal H atom composed of a proton and an
electron (r ~10-10 m, mH ~ 1.6 x 10-27 kg, q1 = q2 = 1.6 x 10-19 C) by writing
Δm/m = E c2/mH
that predicts α ~10-8. In fact bonds among atoms are weaker than between the electron and
the proton of a H atom, and in fact reaction (combustion) between H2 and O2 has α of the of order
1.5 x 10-10 . Coulomb forces are strong on a human scale, as suggested by the value of the constant k
14
and confirmed by experiments with electrostatics. However, their potential energy is limited by the
charge of the electron, ‘only’ 1.6 x 10-19 C. Atoms may contain many electrons (actually, A, with A
= atomic number), but their potential energy is the sum of attraction and repulsion among all
electrons and nuclei, including those of nearby (colliding) atoms. Whatever the atom, its atomic
number A is less than 100, so potential energy would be, at most, 102 that associated to a single
positive and negative charge pair interacting at collision distance.
Because of the importance of electrons at the dawn of modern physics, particle energy is
measured in eV. An electron of energy equal to 1 eV (1.602 x 10-19 J) has KE of order 11,300 K. A
hydrogen atom of 1 eV energy has velocity lower by the ratio m p m e , where the p and e
subscripts stay for proton and electron. In chemical reactions even such temperature is unusual:
energy or potentials associated to electro-weak forces are < O(1) eV. Only ionization reactions may
reach higher energies: for instance, ionizing hydrogen needs 13.8 eV.
Thus, electroweak forces have practical potential at most of order O(10) eV only. Equation
(2) quantifies the percentage mass change in chemical reactions, ~ 10-10. For this and also for
historical reasons, it is very inconvenient to use equation (2) in thermochemistry, i.e., keeping track
of energies by means of the mass change (‘mass defect’, if energy is being released). It is far more
practical to follow the evolution of the sum of all microscopic KE of particles (i.e., translational,
rotational, vibrational, electronic…). Macroscopically, this sum becomes the more familiar free
energy, internal energy, enthalpy and so on. This is what actually one does when using standard
thermodynamics, unaware that total mass also changes, although very slightly. In thermodynamics
what is being converted is only KE, for instance microscopic into macroscopic. The total mass in
standard thermodynamics stays constant (“mass is conserved”).
In fact, equation (1a) says this is not so: in exothermic chemical reactions mass decreases. In
fact, one could rightly say that it is the slight mass ‘defect’ in rocket combustion (of order 10-10 per
unit mass of propellants) that, converted into microscopic energy, heats the combustion products
and is responsible for thrust.
Chemical propulsion is very effective to produce thrust in a relatively simple way. In fact,
combustion dates back to the Paleolithic. The main drawback of chemical propulsion is ‘low’ Isp.
Why is that so? Aside from its units, Isp is equal to the exhaust velocity Ve for model ideal
expansions, defined here as isentropic, 1-D and reaching zero pressure. The exhaust velocity is
limited because it cannot be much different from that reached by molecules and atoms in their
random motion inside the combustion chamber. Inside the chamber the heat release forms
molecules possessing large translational, rotational and vibrational energy (all of them internal
energy E), and very little bulk (flow) velocity V. When the hot products expand in the nozzle,
collisions force all molecules to reach statistically the same orderly flow velocity, Ve at the exit, at
the expense of internal energy, that is, Ve = (2 E/m)1/2 if one ignores relativistic effects. The ratio
E/m is the energy density J. Even when burning the combination H2/O2, J cannot reach above 107
J/kg. J is also the microscopic KE/unit mass, or temperature, of particles present in the chamber. So
chemical Isp is limited intrinsically by the J of the electro-weak force. Note that it is the ratio E/m
that counts, not E: for instance, in LOX/LH2 rockets the ratio E/m is higher than in combustion
between HCN and O2, which has a much higher adiabatic flame temperature (about 5400 K), but
that has the much heavier N2 and CO as main products.
Since ΔE = Δm c2 , energy density J of order 107 means α of order 10-10, very small in terms
of the mass consumption predicted by Newton’s second law as embodied in Tsiolkovski’s law
linking mass expenditure to specific impulse (Ve). Indeed, it is much too small to enable practical
interplanetary missions, i.e., missions within reasonable times and mass consumption. Here a
reasonable time is purposely defined for human missions ≈ 1 year (to prevent unwanted
physiological effects due to microgravity and radiation), and for probes 3-4 years: longer unmanned
15
missions tend to ‘lose’ public attention and financial support, sometimes even in their proposal
phase. In terms of mass, reasonable depends on the specific mission, a reasonable cap being linked
to total cost as a fraction of order, say, 0.1% of the GNP. An α of order 10-10 is in fact too small to
enable reaching LEO for less than O(103 - 104) $/kg of payload. Research on HEDM propellants,
such as N8, metallic H, metastable helium, mixtures of nanoparticles, or even less exotic molecules
[Davenas et al, 2000; Agrawal, J.P., 1998] may well increase the energy available per unit mass,
perhaps by a factor of order ten. However, α will become only of order 10-9.
The nuclear potential is associated to the ‘strong’ nuclear force acting among nucleons. It is
this force that binds them together and prevents the nucleus from disintegrating spontaneously due
to Coulombic repulsion among protons. Coulombic repulsion is long-range, while the nuclear force
is very short range: its intensity is about 102 that of Coulomb, but only when inter-nucleon distance
is of order 1 nucleon size, or 1 fermi.
This force plays the same role as the chemical bond in thermochemistry, but at the scale of
nuclei instead of that of atoms and molecules.
The curve of binding energy (Figure 1) shows the energy, in MeV, associated to this force,
averaged over the nucleons of each nucleus (each A mass number). In fact, nucleons are bound to
other nucleons each with a slightly differently energy, just as electrons to the nucleus of an atom.
Note that the binding energy is of order MeV, not eV: so, it is an easy guess that the energy, or
potential, available in nuclear transformations, will be about 106 larger than the electro-weak
potential.
Figure 1 Average binding energy per nucleon, MeV, as a function of the mass number (from Mukhin, 1983)
The binding energy is actually the negative of the potential energy associated to the nuclear
force. The larger the binding energy, the deeper the potential well. In this respect, its maximum near
A = 60 means that near that A the energy well of those nuclei is the deepest and the binding force
strongest. Before WW II this fact suggested to O. Hahn, L. Meitner and others the possibility of
extracting energy from nuclear reactions. The particular shape of the binding curve can be
intuitively justified: on the left A is low, and the few nucleons imply binding energy is also low; on
the right A is large: Coulomb repulsion grows with A, weakening the nuclear force. The maximum
must be in between.
Binding energy, with its sign changed, plays the same role as the formation enthalpy for
chemical species. If nuclei with high A (to the right of the energy peak) are broken up into several
lower-A nuclei, for instance, by collisions with neutrons (fission), the lower-A nuclei produced ‘go
deeper’ into the well, and the energy difference is released. For example, breaking an hypothetical
16
A = 240 nucleus, composed of 240 nucleons, each with average energy 7.5 MeV, into four nuclei
(‘fragments’) with A = 60, each containing 60 nucleons of average 8.5 MeV energy, releases (240
x 7.5 - 4 x 60 x 8.5) = -240 MeV, all this energy from a single nucleus.
Likewise, low-A nuclei (to the left of the energy peak) may collide and form a higher-A
nucleus (fusion). Note that in this case the energy change could be much larger than in fission: for
instance, the binding energy of A = 1 (hydrogen) is much lower that that at A = 4 (helium), so the
net energy release per nucleon is larger.
Thus in both fission and fusion the end result is a stronger binding force per nucleon. In this
context, a much-used and useful analogy is with the surface tension of a liquid drop [Mukhin,
1983].
In nuclear reactions, original nucleus(i) plus other particles, such as neutrons, are the
‘reactants’. The total mass of the ‘products’ formed (ligher or heavier nuclei, plus energetic
particles) may decrease or increase. If mass decreases, as stated by equation (1a), energy is released.
Examples of nuclear reactions are 235U or 239Pu reacting with neutrons and splitting into
smaller A fragments (fission), and four H ions colliding and forming He, a higher A element
(fusion). Fusion processes inside stars drive the photon flux that, once captured, may power electric
thrusters. So-called “solar power” is kinetic energy of products of nuclear reactions.
Besides fission and fusion, a third form of nuclear energy is available from nuclei of certain
isotopes. They have the same atomic and mass numbers of more common nuclei, that is, the same
number of electrons, protons and neutrons, but the spatial arrangement of nucleons inside their
nucleus is different. In fact, nucleons may be spatially arranged in more than a single configuration.
Only the nucleus in its minimum energy configuration is stable; others may exist, but are potentially
unstable. When these nuclei relax to their stable configuration, either naturally or artificially, excess
potential energy is released, typically as high energy photons (X- and gamma-rays). Because no
nucleons are involved, the α of this process is lower than in fission, of order 10-7 (for instance, in
the case of the 180mTa isotope, α is 2 x 10-7). Natural isotopes with this peculiarity are called
“nuclear isomers”, by analogy with chemical isomers, and a ‘m’ suffix is then attached to their
mass number A. The energy available from these isomers is about 103 larger that of chemical
reactions but 103 smaller than that of fission.
The ultimate form of nuclear energy release occurs when matter and antimatter, for instance
a proton and an antiproton, are ‘fused’. This is called annihilation; both masses disappear,
becoming energy according to Einstein’s law, and α = 1.
Once α is known, the product α c2 is the energy/unit mass, or energy density, J. Table I
reports α and J of energy conversion processes. Data on fission are for 235U. Among many, fusion
data are for reaction between deuterium (D) and tritium (T), both hydrogen isotopes containing 1
and 2 neutrons, respectively.
17
(*) per pair of 1-kg masses at 1-m distance.
Table I Forces, conversion fractions and energy densities
No nuclear process can produce α between about 4 x 10-3 and the theoretical 1.00 of
annihilation: nuclear reactions transform mass into energy with α “limited” to a few tenths of
percent. The reason is the also limited average binding energy. Mass converts into energy when
new bonds forming among nucleons are stronger than the old ones. Since the average bond strength
per nucleon cannot be more than about 8 MeV, from equation (1a) α cannot be greater than about 4
x 10-3. Higher α would be possible only if binding energy differences among nuclei were higher
than those in Figure 1, and that is not the case in the Universe we know.
Note that no nucleon is annihilated in fission/fusion reactions. Annihilation is routinely
observed in particle accelerators, has been and is still investigated as an energy source, but
unfortunately remains for the time being a conceptual means of propulsion.
Comparing energies in Table I, that associated to the nuclear potential via fission or fusion is
about 106 to 107 times larger than chemical. The J from metastable nuclear isomers is smaller, but
still of order 103 larger than chemical.
Thus, at our stage of knowledge, besides being the only physical alternative to chemical,
nuclear energy is the only means of altering the prospectives of future propulsion in the practical
sense already mentioned.
This statement is not meant to imply NP is the panacea propulsion solution; other propulsion
systems have been recently proposed based on sound physics, such as solar or magnetic sails,
electromagnetic mass drivers and others. However, all these new developments suffer at present
from severe limitations in thrust available or power required, so their range of applications is
likewise limited. In terms of availability, prior expertise and know-how, and overall performance,
NP is a most attractive option for a broader class of interplanetary missions, including those with a
human crew. It is because of this, and other, reasons that NP is among technologies given higher
priority by the EU space industry [ASD report, 2005].
18
1.4. Propulsion
In the Standard Model, where gravitational mass equals inertial mass, Newton’s third law is
the only way to produce propulsive force (but some claim [e.g. Cornille, 1999] the electroweak
Lorentz force capable of accelerating plasma in applied-magnetic field thrusters, does not ‘obey’
this law).Thus, relative to a frame of reference fixed to the propulsion system, something must be
accelerated in the direction opposite to that of motion.
That something must possess momentum, if not mass. Chemical rockets eject matter
accelerated by chemical heat release, and collimated by the solid walls of a nozzle, but propulsion
by change of momentum of massless particles is also conceivable (see Section 5). The difference in
momenta ‘before and after’ the acceleration is the engineers’ thrust, F.
1.4.1. Specific Impulse
The process from potential energy to exhaust jet energy (also called thrust energy) can be
divided in three stages.
The first stage is the conversion of potential energy into KE of particles. This is microscopic
energy; in chemical rockets it corresponds to the creation of translation, vibration, rotation and
electronically excited species. In nuclear rockets this stage forms nuclear products such as fission
fragments, or fused particles, nucleons, and photons (radiation). Energies vary from 160-170 MeV
in the case of FF, to 5-15 MeV for nucleons and photons (gamma-rays, mostly), and O(1) MeV for
fused alpha particles.
In the second stage the nonequilibrium microscopic KE is redistributed among particles and
tends towards statistical equilibrium. In chemical rockets this stage corresponds to an increase of
roto-translational energy at the expense of the much higher electronic and vibrational excitation. In
NTR or NEP rockets this stage corresponds to the energy exchange between microscopic KE of
products (neutrons, fission fragments, fused particles and photons) and a medium where they
thermalize. This medium may be fuel rods, gaseous or liquid propellant, a working fluid, or
products of the nuclear reaction(s) themselves.
Finally, in the third stage the thermally equilibrated medium, with high energy density, is
exploited in some way to produce thrust. This stage may take different forms, according to the
specific propulsion concept chosen.
Whatever the potential energy tapped, energy is conserved through each stage. Propulsion
concepts differentiate only at the third stage, when deciding what to do with the microscopic kinetic
energy acquired by the medium.
What is the ideal velocity V (or Isp) of products after the potential energy α mc2 has been
thermalized? In general, the answer requires a relativistic energy balance. A truly rigorous balance
should include the contribution of massless particles, like photons (hν) and neutrinos, to the KE of
products. A little simplified, here KE will include only mass contributions. Assuming no losses, the
energy balance imposes the sum of potential and KE be conserved going from reactants (at V = 0)
to products, eventually exhausted at velocity Ve = V:
1 mo (1 − α )V 2 1 MpoV 2
mo c = (1 − α )mo c +
2 2
+ (3)
2 V2 2 V2
1− 2 1− 2
c c
19
where m0 and Mp0 are the rest masses of reactant and of inert propellant possibly added,
respectively. On the LHS energy is only potential energy: the reactant is still. Rearranging this
equation, a preliminary result is
(4)
showing that in the limit α →1 (that is, if all reactants are converted into energy, as in
matter-antimatter annihilation) and when no inert mass Mp is added, the exhaust velocity V tends
to the speed of light c. If inert mass is present or added, the limit velocity is less than c, as shown by
the complete solution
V2 2
= (5)
c2 2
1+ +1
A
with
2α 2
A≡ 2
(6)
⎡ Mpo ⎤
(1 − α ) 2 ⎢1 + ⎥
⎣ mo (1 − α ) ⎦
The ratio Mp/[m(1-α)] might be interpreted as the ‘dilution ratio’ of products; in the
following, the ratio Mp0 /m0 will be called μ.
The V/c ratio is plotted in Figure 2 for three different μ = Mp0/m0 ratios (1000, 10,000 and
100,000) and also for the special case Mp0 = 0. For clarity, the V/c curves for Mp0 ≠ 0 have been
magnified by a factor 10.
Figure 2 Velocity acquired by non-reacted plus inert mass as a function of percentage α of mass converted into
energy. The three lower curves are multiplied by a factor 10 for clarity. Mp0 is the rest mass of (inert) propellant
added.
20
Equation (6) may be written as
Isp = V = c f(μ, α) (6bis)
Where f(μ, α) ≡ 2 { [ (1 + 2/A ) + 1]}

showing that the exhaust velocity V, i.e., the ideal Isp when using its true physical units, is
essentially the speed of light (the limit speed according to Relativity) scaled down by the specific
mass conversion process (that is, by α), and by the addition of inert mass Mp0 (that is, by μ). So,
only a tiny portion of the curves in Figure 2 , those closest to the left (plus the α = 1 special case of
annihilation) are in fact relevant to energy conversion and propulsion. In essence equations (6) or
(6bis) tell that “in principle” the exhaust speed may reach up to the speed of light, c. The function f
may be interpreted as an attenuation factor: if fuel mass is is not completely converted into energy
(i.e., if α ≠ 1), or if inert mass Mp is added to the conversion products, the function f is < 1, and the
specific impulse will be < c.
What is the meaning of the special case Mp0 = 0? Mp0 = 0 (top curve) means that the
potential energy is completely converted to kinetic energy of products and products only: products
are ejected ‘as formed’, with KE from O(1) to O(100) MeV, and V is the highest possible at each α.
Such strategy has been proposed at the Lawrence Livermore National Laboratories to maximize Isp
in fission engines. Figure 2 also shows that with α typical of fission or fusion the Isp is a tiny
fraction of c. However, since c = 3 x 108 m/s, even a tiny fraction can result in Isp vastly larger than
for chemical rockets.
Note that if V << c relativistic mass and rest mass may be assumed the same; then the energy
balance in equation (1a) reduces to the well known expression
Isp = 2J (7)
Equation (6) shows why nuclear power is critical for future space missions. The square-root
dependence tells that modest increases of J, for instance, using HEDM, have a minor effect on Isp.
Only by increasing J by orders of magnitude Isp may increase significantly. This is the case of
nuclear energy: see Table I.
1.4.2. Thrust
Thrust, F, specific impulse (ideally V), and thrust power, P, satisfy the two relationships
F Isp = P
F = Isp dm/dt
where dm/dt is the total mass consumption, equal to the sum of the mass flowrate of
unreacted fuel (if ejected), and of that of the inert mass.
The reactor power, PR, and the thrust power, P, are related by
P = PR ηtot = α m& 0 c2 ηtot
21
with ηtot the total efficiency of energy conversion. Hence
F = (P dm/dt)1/2
Note that thrust scales with the square-root of power. This may sound trivial, but is a key
factor in nuclear propulsion, because costs grow with reactor power (size) more rapidly than in
chemical rockets. With the definitions given,
& ]}1/2
F = {α m& 0 c2 ηtot [z(1 – α) m& 0 + M (8)
P0
In equation (8) z is a factor either one or zero depending on whether or not unreacted fuel is
ejected.
Equation (8) may be rewritten in terms of the ratio μ:
12
F = α ⋅ m& 0 ⋅ c ⋅ ηtot ⋅ ⎡ z ⋅ (1-α ) 1- ( V c ) +μ 1- ( V c ) ⎤
2 2
(9)
⎣⎢ ⎦⎥
In the square brackets the first term is generally much smaller than the second (μ >> 1) so
that
12
F ≈ α ⋅ m& 0 ⋅ c ⋅ ηtot ⋅ ⎡μ 1- ( V c ) ⎤
2
(10)
⎢⎣ ⎥⎦
As anticipated in Section 1.3.1, equation (9) tells that if only fission or fusion fragments are
ejected (μ = 0) the thrust is of order μ smaller than that in a rocket where temperature is
moderated by adding mass, unless the fuel consumption (and power) is made μ times larger.
Since μ may be of order 103 to 106, the implications are significant.
Similarly to what seen in discussing the specific impulse in Section 1.3.1, Equation (10) may
be interpreted as specifying thrust composed of a “limit” thrust, α ⋅ m& 0 ⋅ c , scaled down by an
efficiency coefficient and scaled up by the addition of mass. Adding mass lowers Isp and raises F as
the same time and, approximately, with the same square-root dependence. If inert mass is not added
(µ = 0) the maximum thrust possible is limited by α ⋅ m& 0 ⋅ c , just as the Isp is also limited by c (see
Section 1.3.1); however, unlike the Isp, this ideal thrust increases if inert mass Mp is added. An
aeronautical engineer will find an analogy with turbofan engines, where the extra, unburnt air
passed through the fan contributes to thrust. In the case on NP adding mass, as seen, reduces at the
same time the specific impulse.
The nondimensional thrust Φ = F/ ⎡⎣ α ⋅ m& 0 ⋅ c ⋅ ηtot ⎤⎦ is plotted vs. μ in Figure 3. It shows
how the thrust F, for a given conversion process and for a given power level, grows with the
addition of inert mass. Quite intuitively, the effect of adding mass is enhanced by increasing α and
increasing power (that is, increasing fuel rate of consumption).
22
Figure 3 Effect of inert propellant addition to nondimensional thrust Φ ( case of fission, α = 9.1 x 10-4 )
The scaling with α in equation (10) implies that no matter what the nuclear source,
metastable nuclei, fission or fusion, the effect on thrust is slight. Only annihilation makes a
significant difference. Note that equation (9) predicts that for annihilation (α = 1) no thrust is
possible unless inert is added, because reactants are completely converted into energy (that is,
massless products). However, pure energy may still produce thrust, see Section 1.1.5. The influence
of ηtot on thrust is small.
Thrust may be written also as in standard rocketry textbooks, to show its dependence on Isp
(or V). Using equation (5bis),
F = ( ρe A e ) ⋅ c2 ⎡f 2 1-f 2 ⎤ (11)
⎣ ⎦
depends essentially on V2 for V/c << 1. Equation (10) may be interpreted as composed of a
“limit” thrust, (ρeAe) c2, corresponding to a rocket with Isp = c, scaled down by the function of V/c
and μ in square brackets, always << 1. As V/c grows, relativistic effects become more and more
significant, and F grows more rapidly than V2 (but so does the mass of the spacecraft, according to
Relativity).
1.4.3. Power
Power shapes size, complexity, and cost of nuclear reactors. In fact, at least conceptually,
nuclear reactors are not power limited: they are limited by the operating temperature of fuel and
structures, see Chapters 2 and 5. If nuclear-generated heat could be removed sufficiently fast,
nominal power could increase manifold above common practice. The key word is “sufficiently
fast”: during the Chernobyl plant accident the 1600 MW reactor involved was being operated at
only 200 MW. The sudden plunging of control rods, with the cooling system deliberately turned
off, raised power by a factor 100 in 4 seconds. That is, in four seconds the power became twelve
times the reactor nominal power. This is quite unlike chemical rockets, where more propellants
must be fed to the engine to increase power. Designing compact GW-class nuclear systems is
23
mostly a heat transfer problem, not a nuclear physics problem. Of course, even with adequate
materials and cooling, greater power means greater fuel consumption, see equation (1a).
Reactor power PR is 1/ηtot of the thrust power P defined as
P = Isp F (12)
From equation (10) thrust power can be written
P = ( ρ e A e ) ⋅ c3 ⎡ f 3 1-f 2 ⎤ (13)
⎣ ⎦
showing that for V << c power scales with V3, as expected. Similarly to thrust, as V/c Æ 1
power grows faster and faster because of relativistic effects.
When power is fixed (by size, materials or other reasons), the hyperbola P = constant on the
(Isp, F) plane (see Figure 4) illustrates graphically the power dilemma. Nuclear propulsion can raise
Isp by large factors with respect to chemical, but at fixed reactor power thrust must decrease by the
same factors, that is, as 1/Isp. Low propellant consumption goes together with low thrust, and one
must choose between either low consumption or low thrust.
Figure 4 Thrust vs. Isp at constant power P (total efficiency ηtot = 0.8 assumed) (Andrenucci, 2004)
Generally speaking, one of the purposes of NP, and a ‘must do’ for manned missions, is to
shorten mission times. So, spacecrafts must accelerate at a significant fraction of, say, 1g for longer
than the ten or so minutes of chemical rockets. This strategy shapes power requirements sharply.
For instance, a 1-ton spacecraft accelerating at 10-2 g needs a 100 N thruster. With Isp = 10,000 s (≈
105 m/s) the thrust power is of order 10 MW. This figure gives pause to any NP system designer.
Note that this figure is thrust power, not reactor power. In nuclear electric propulsion it does not
include total efficiency (see Section 1.1.4). It does not include power to run auxiliary systems and
subcomponents either. In fact, in nuclear thermal rockets turbopumps power might be non-
negligible, although nowhere near that of most LRE, not only because propellant consumption
24
scales as 1/Isp, but also because there is no need to reach high chamber pressures (in space an
expansion ratio of order ten is acceptable).
1.4.4. Mass
Deriving rules to predict even preliminary mass budgets is beyond the scope of the present
work. All items, engine, tanks, structure, auxiliary systems (some critical, like turbopumps) and,
finally, propellant, are too intricately linked. For this reason only fuel and propellant (inert) mass
will be discussed here.
The fuel consumption rate, by definition, is
m& 0 = PR J = PR (α c )2
(14)
while that of inert propellant depends on Isp (or α) and on mission time, i.e., acceleration and
distance, in turn a function of thrust and Isp. The total instantaneous consumption is
dm
dt
& = P
= m& 0 + M 0 R ( α c ) (1+μ )
2
And total mass consumed by propulsion over the total mission time, t, is
∫ (dm / dt )dt
0
which sometimes is added to the dry mass of the spacecraft and assumed a design constraint.
Doing so is not necessarily a wise choice, even though it looks the easier way to constrain total cost
[Czysz and Bruno, 2000].
Inert mass is an option in nuclear propulsion: in principle, a nuclear rocket could eject only
reaction products (μ = 0). However, moderating the nuclear reaction, and cooling may be
accomplished by a fluid, and is natural to think of it as of the propellant. In fact, if no inert mass is
added (Mp = 0: thrust provided only by the momentum of the reaction products) the temperature of
the mass ejected is the microscopic KE acquired at the end of the first stage, of order 5-20 MeV per
particle in fusion, and correspondently higher (~ 160 MeV) for the heavier products of fission.
Without appropriate measures, e.g., magnetic confinement, such temperatures cannot be withstood
by current or future materials: the forces keeping structural materials together are electro-weak, thus
with potential of order eV, not MeV.
The temperature may be lowered by adding inert mass Mp0 to products. How much mass to
add depends on maximum allowable temperature and on the specific heat of the inert, CpM, and
fuel, Cpm. Assuming V << c to simplify the reasoning, the ratio μ = Mp/m limiting the maximum
temperature rise, ΔT, is equal, from a simple energy balance and equation(3), to
μ = [αc2 /ΔT – (1 – α) Cpm]/CpM (15)
Since αc2 >> (1 – α) CpM ΔT for any reasonable structural temperature,
μ ≈ αc2 / (CpM ΔT)>> 1 (15bis)
25
that is, the inert mass fraction is proportional to the ratio between energy density of fuel and
enthalpy density of the inert, or to the ratio between potential energy of fuel and macroscopic KE of
the inert. Equations (15) and (15bis) tell that constraining maximum temperature is costly in terms
of mass addition: Mp >> m. For instance, adding inert hydrogen to a fission rocket (α = 9.1 x 10-4, J
= 8.2 x 1013J/kg) to limit its temperature rise to ≈ 2500 K requires μ ≈ 2 x 103, or carrying two
thousand times more hydrogen than fissionable fuel.
The need for inert worsens with increasing α. In fact, equation (14bis), at fixed ΔT, implies
the product
μ /α ≈ constant
that is, the more efficient the mass conversion process, the larger the inert mass moderating
temperature. The extreme case is annihilation, requiring two million times more hydrogen than fuel,
in this case proton-antiproton pairs. Such propulsion systems would be characterized by extremely
large tanks, size and weight. Thus ways to thermally confine the exhaust products is a primary
requirement for high energy density (high α) propulsion, in that it may allow higher temperatures
than bearable by structural materials.
It is clear that adding inert mass lowers Isp, see Figure 2. However, at fixed power, adding
mass raises thrust, that may be tailored to a specific ìmission. The temperature constraint explains
why solid core nuclear thermal rockets tested in the past had Isp no higher than 900 s, as shown
later.
The issues raised by balancing Isp vs. thrust at fixed power in planning interplanetary
missions can be appreciated by calculating their effect on propellant mass and ∆V. These questions
are seldom relevant to chemical propulsion, where ‘thrust applied for a very short time’ is the real
variable controlling ∆V, not power. These issues will become critical for future nuclear electric
propulsion, where continuous thrust may have to be applied for months or even years. A simple
quantitative analysis of these issues (without relativistic effects) involves the following equations:
P = Isp F Thrust power
& = F Isp
M Isp definition (V << c assumed)
P
Mp = F tacc / Isp mass of propellant consumed at constant M & after a time tacc under
P
power
dacc = ½ a (tacc)2 distance traveled at constant acceleration a
∆V = a tacc ∆V acquired after a time tacc
F=Ma Newton’s law. M is the spacecraft mass, assumed >> Mp
This set of equations model propulsion needs and performance for interplanetary missions.
They have been simplified to obtain a fast analytical solution. Part of the trajectory, dacc, is assumed
at constant acceleration a = P / (Isp M), positive or negative. Spacecraft mass and power (M and P)
may be assumed as input. If d is the distance to the final destination the dacc should be ≤ d/2.
However, dacc may turn out to be greater than d/2 when acceleration is modest (modest thrust and
power). In this case the spacecraft must spiral (for instance, around a planet), until reaching the ∆V
for planetary escape. At that moment the spacecraft can start the trans-planetary leg of its trajectory.
Solving for time, mass of propellant and ∆V [Czysz and Bruno, 2006],
26
tacc = (2dacc Isp M / P)1/2
Mp = (2dacc P M / Isp3 )1/2
∆V = [2dacc P / ( Isp M )]1/2 (16)
Equation set (16) shows propellant mass decreases rapidly with increasing Isp. However, the
effect of Isp on acceleration time and on ∆V is detrimental: at fixed power, time of acceleration (trip
duration) stretches and ∆V shrinks when Isp increases. This may be rather unsettling, unless one
realizes the consequences of keeping power constant.
It is instructive to apply the solution set (15) to a nominal Earth to Mars mission (minimum
distance about 1.5 x 108 km), starting from a hypothetical 0.7 MW ion engine with Isp = 4000 s and
a spacecraft of mass M = 100 ton. Sequential attempts are in Table II:
dacc (km) P (MW) tacc (days) ΔV (km/s) Mp (ton)

107 0.7 1157 2 5
8⋅107 0.7 3450 6 15
8⋅107 70 33 54 135
Table II Trajectories and performance of NEP as a function of power
This exercise shows that finding a practical solution is no easy task, even with all
simplifications made. The third attempt provides a reasonably fast mission, at the expense of much
higher power than initially assumed, and violating the Mp << M constraint. While a solution seems
within reach, xenon requirements will likely exceed current xenon worldwide production (about 59
t/year), unless Isp may be of order ten times higher than assumed in this example.
Note that a 100 ton spacecraft is likely a minimum for an interplanetary manned mission.
The preliminary conclusion is that, for certain ambitious missions currently being discussed, short-
term ion engine technology (Isp in the 4,000 to 10,000 s range), is insufficient to produce a
practical trajectory, defined as a trajectory that is both fast and cheap [Czysz and Bruno, 2006].
Only Isp of order ten times those now available (that is, 40,000 s) can provide a satisfactory
practical propulsion solution in the sense just defined. This means much more powerful nuclear
reactors than those envisaged, for instance, for JIMO missions [Randolph and Polk, 2004].
It is important to have a fast trajectory not only for technical and physiological reasons
(should the mission be manned), but also because among the many motivations of space exploration
there are also curiosity and adventure. Both are lived vicariously by the public, that is, the taxpayer.
Public enthusiasm cannot be maintained for a mission where ‘nothing happens’ for many months,
let alone years.
1.5. Nuclear Propulsion Strategies
After potential energy has gone through the second conversion stage of Section 1.3.1, the
choice is between two strategies: Thermal, or Electric. Each has variants, briefly discussed below.
1.5.1. Nuclear Thermal Rockets (NTR)
NTR are the subject of Chapter 2. Here only a brief summary of their features and
performance is given, by way of introduction to Chapter 2.
27
The NTR strategy is straightforward: it consists of using heat released during the second
stage of energy conversion to pressurize and eject mass. This mass may be only reaction products
(μ = 0, fission or fusion fragments), or inert Mp (z = 0, μ≠ 0), or both (z = 1, μ ≠ 0). In the last two
cases the ratio μ may have to be of order 103 to 106 to increase thrust and to moderate temperatures.
In the third stage of conversion microscopic energy becomes macroscopic kinetic energy of the
propellant jet. This third stage occurs via collisions, e.g., with the solid walls of a conventional
nozzle, or by magnetic confinement inside a magnetic nozzle (this solution is mandatory at extreme
thermal loads, and requires an electrically conductive propellant).
From this viewpoint NTR work just like chemical rockets: hence the definition “nuclear
thermal rockets”. In NTR the macroscopic KE of the jet issuing from the nozzle must be of the
same order of the microscopic KE acquired by particles at the end of the first or second stage
(which one depends on whether only reaction products are ejected, or inert Mp is also added, see
Figure 2. In either case thrust power P differs little from reactor power PR: nearly all of the energy
from the reactor ends as kinetic energy of products and inert.
Most NTR experience is from solid-core fission types [Gunn and Ehresman, 2003]. They are
characterized by μ of order 103 and z = 0 (unburnt fuel stays inside fuel rods). Their most appealing
feature is bulk power density of order 1 MW/liter. During the US ROVER program the Los Alamos
Laboratories (LASL) built the Phoebus II reactor, that was tested at 4 GW for 12.5 minutes, and had
a power density about 1.3 MW/liter (see Figure 5). On similar solid core NTR the Isp measured was
of order 880 to 900 s.
Figure 5 The Phoebus IIA solid-core nuclear reactor on its Los Alamos test stand (Dewar, 2004)
One might ask why Isp was almost twice that of LRE if temperatures were comparable or
lower. The answer is, in a chemical rocket the exhaust V (ideally, Isp) is close to the mean
molecular speed, (8kT/πm)1/2, inside the combustion chamber. Although the same is true in a solid
core NTR, its propellant of choice, hydrogen, has molecular weight 2 instead of 9 or 10 typical of
the combustion products of the best [LOX/LH2] chemical rockets. Accordingly, the exhaust velocity
is higher by the factor 9 2 to 10 2 , about two. So, Isp in vacuo may be in the 900 s range.
28
Variations on the NTR theme skip the second stage of energy conversion. In gas core rocket
concepts the fuel may be gaseous 235U fissioning at 8,000 - 20,000 K [Howe et al, 1998]. In fission
fragments (= FF) rockets, the FF from a solid fuel thermalize directly inside the propellant.
The gas core strategy depends on neutronics, e.g., on the collisional cross section between
neutrons and nuclei at high temperature. In solid core reactors the thinner the fissioning fuel, the
larger the volume of propellant injected to moderate fission kinetics and thermalize. Not much
information on gas-phase neutronics exists, but see [Koroteev, 2002]. LASL has been active in the
gas core rocket for two decades, a sign that its neutronics is viable. Either alone or with inert
addition, fission products may be ejected at speeds that at 10,000 K, for instance, should yield Isp of
order 1500 s at least, and with substantial thrust, see equations (9-10). Isp is not especially high
because 235U and other high molecular weight products may be ejected together with hydrogen.
There are sub-variations of the gas-core nuclear rocket: the cycle may be open (propellant
and FF are ejected as they are heated, raising the average molecula weight), or closed (the fuel is
confined in a transparent chamber, and heats low molecular weight propellant by radiative HT, see
[Howe et al, 1998]). The complexity of such schemes has not prevented testing of components
hardware.
In the FF engine [Ronen, 2000], or Rubbia engine [Augelli et al, 2000], hydrogen propellant
flows in a chamber coated with a thin layer of fuel (see Figure 6). Half of the FF from the coating
are injected and thermalize directly inside the propellant. Many solid core nuclear reactor problems
are bypassed in this way. This concept is capable of heating to much higher temperatures than
tolerable by solid core reactors, i.e., to 6,000 to 8,000 K. The Isp may reach 2,400 to 2,800 s. A rare
242m
Am isotope is the fuel of choice for the FF/Rubbia engine, chiefly because its neutron collision
section drops towards zero with increasing temperature. This feature prevents loss of coolant
accidents.
Figure 6 Conceptual scheme of a Rubbia-type nuclear rocket engine
29
The radical approach to fission NTR is of course eliminating inert propellant (μ = 0, z = 1),
as proposed by LLNL. The rocket may eject only the reaction products, their high momentum
‘undiluted’ by inert mass addition. Predictably, thrust will be small, see equation (7). For reference,
a fission reactor (α = 9.1 x 10-4) consumes fuel at a rate m& 0 ≈ 1.21 x 10-5 kg/s of fuel per thermal
GW. E.g., the 4.1 GW Phoebus IIA reactor had m& 0 ≈ 0.05 g/s. If operated at μ = 0 its thrust would
have been 40 N instead of the planned 440,000 N available by adding hydrogen propellant. The
tests indicate in fact that μ was 2.4 x 106 and Isp only about 900 s.
Fusion concepts based on magnetic mirror confinement are supposed also operated in this
mode. The products are mostly fused alpha particles and unburnt fuel (e.g., D-T). Cooling would be
a major issue, but Isp could range within 104 to 105 s, as predicted by equation (5) but also
depending on fusion engineering strategy. In any event, the implication is that to produce F in the
104 N range this type of fusion rockets must be massive, compensating the low thrust density with
sheer power and size. A discussion of future fusion rockets based on Open Magnetic Mirror
Confinement technology may be found in [Romanelli et al, 2005].
There is a third, radical way of exploiting nuclear energy for propulsion: repeated nuclear
explosions astern of a spacecraft (pulsed nuclear propulsion is a fitting name suggested by Schmidt
et al [2002]). Hardly conceivably now, this method was proposed and investigated in the ‘50s by
Freeman Dyson [Dyson, 1979; Dyson, 2002] and Ted Taylor, a fission bomb physicist, for single
stage to orbit (SSTO) launchers. A concise history of this project is in [Flora, 2002]; basic
propulsion aspects are discussed in [Schmidt et al, 2002].
This unusual propulsion technique was suggested by the results of thermo-nuclear bomb
testing on Eniwetok, after teams examining the ground in the aftermath of the explosion noticed that
the graphite-coated metal spheres hung some 30 feet from ground zero were left practically
unscathed. Until then it was assumed that nothing could survive a close nuclear explosion. In fact,
later testing and analyses showed ablation of a plate by the intense radiative environment could
protect an underlying structure. Suitably sized and reinforced, what was then called a ‘thrust plate’
could indeed receive and survive the force due to shocked matter accelerated by an exploding
atomic bomb and its radiation. Radiation from the fireball contributes to the force, for instance by
ablating the coating deposited on the thrust plate (e.g., a polymer, or grease), the momentum of the
ablating products ejected working just as a rocket jet exhaust. Much of the information concerning
this area of ablation and its physics is still classified today, but calculations and tests done with high
explosives confirmed in 1959 the concept was viable, particularly so for massive spacecrafts. Such
spacecraft must include also a shock absorber to protect the crew. In the ‘50s the nuclear test ban
was not in existence, so F. Dyson and the physicists working on this project (Project Orion),
envisaged taking off from ground and accelerating to orbital speeds all by sequential atomic
explosions, probably the first modern SSTO proposal. Orion was eventually designed for a
spaceship large enough to do a grand tour of the planets (as far as Saturn) lasting about one year.
The estimated mass of the spaceship for such mission was of order 10,000 ton. Specific impulse and
thrust calculations showed both could be much higher than with chemical propulsion, in particular
Isp of order 104 to 106 sec were theoretically predicted. Limitations to thrust were due to maximum
structural stresses, but also to the maximum acceleration tolerable by the crew.
As there was no military application in sight, because of potential opposition by the public,
and certainly that of then Secretary of Defence McNamara, Project Orion was cancelled.
A ‘revisited’ Orion (‘MiniMag Orion’) has been recently revived as a space propulsion
system by replacing atomic bombs with miniature nuclear explosions; among motivations is that of
reducing the mass of the spacecraft that must host this type of propulsion system. Ground testing is
carried on by substituting high intensity electro-magnetic energy pulses (theta pinch-accelerated
plasma jets) for nuclear mini-explosions. One of the actors in this not widely publicized program is
the Andrews Space and Technology company, based in Seattle, Washington. According to its chief
scientist, Dr. Dana Andrews, already in 2000 the Isp measured was greater than 1,000 sec. The
30
thrust impulse should be substantial, unlike that of any NEP thruster, because the instantaneous
power is much larger than possible by any nuclear reactor. Lack of detailed information prevents
saying more about this recent approach to pulsed nuclear propulsion; it looks suited for powering
long interplanetary missions, as it is capable of combining the best of the two classes of NP,
namely, the large thrust of NTR and the high Isp of NEP.
1.5.2. Nuclear Electric Propulsion (NEP)
NEP systems are the answer to an old question: are there ways to raise Isp over that of solid
core NTR? The answer is yes, and carries a price. It consists in converting fuel potential energy into
electricity, just as in nuclear utility powerplants. This strategy involves an extra step, in which
fission fragments heat a working fluid (not a propellant to be accelerated). Through a
thermodynamic cycle, or other means (e.g., thermionics) the fluid may produce mechanical power
and then electricity via an electric generator; see [Bidault et al, 2004] for a detailed analysis of this
issue in the context of the comparison between Solar-Electric Propulsion and NEP for a Mars
mission. Conceptually, in fact, there are also more direct ways of producing ‘large’ electric power
(e.g., > 100 kW), such as MHD generators; see, for instance, [Smith and Anghaie, 2004].
Once produced, electrical power can feed an electric thruster, either electrostatic (ion), or
Magneto-Hydro-Dynamic (Hall, PPT, MPD,…). Electrostatic thrusters, discussed in Chapter 3, rely
on the eectro-weak Coulomb force to accelerate ions; the MHD thrusters described in Chapter 4 use
instead the Lorentz force due to the simultaneous presence of electric and a magnetic fields.
Electric thrusters are capable of much higher Isp than any NTR, because gas acceleration is
not constrained by thermodynamics, that is, by a high T and low T cycle, but driven directly by the
Coulomb or Lorentz forces acting on charges. This peculiarity implies gas must be ionized; so at
high temperature the thermal part of its energy may also be exploited as in any conventional
chemical rocket.
The price of the NEP strategy is low overall efficiency ηtot.This efficiency is the product of
the thermal-to-electric energy conversion efficiency, η, times that (electric-to-propellant kinetic
energy) of the electric thruster, ηE. So the reactor power PR must be 1/ηtot the thrust power P.
Experience with large space power is nil, but η may be estimated of order 50% at most with
thermodynamic cycles. ηE are of order 70% for MW-class thrusters. This means that ηtot might be
only about 35%. A 1-GW of thrust power needs a reactor almost three times bigger.
Efficiency notwithstanding, experience in solar-powered electric propulsion has already

shown that Isp can be raised by a factor three above the 900 – 1000 s of NTR. Commercial ion
thrusters [Lawrence et al, 1995; Fearn, this Chapter 3] are routinely delivering 3000 s now, and
4000 s in the near future. Higher Isp, up to 8,000 – 10,000 s, are envisaged for the NEXIS and
HIPEP nuclear ion thrusters planned for the JIMO mission [Randolph and Polk, 2004; Oleson and
Katz, 2003]. MPD thrusters are capable of even higher Isp, see Chapter 4. In fact, with this strategy,
the ideal exhaust velocity V is limited only by relativistic effects.
Most of the drawbacks of NEP are due to its low ηtot. Wasted power must be somehow
disposed of. In space this may be done only by space radiators. They add extra mass to the
propulsion system: a range of specific radiator mass is 0.01 to 0.15 kg/kW for modern space
radiator concepts operated at moderate temperature (~ 800 K). Ground-based radiators and using
corrosive metals, such as Li, produce typically much larger estimates. So the price to pay for NEP
is not only a reactor with PR > P, but also a heavier propulsion system. The bulk power density of
NEP (power/unit mass of system) may be 10-2 times that of NTR, as estimated by NASA. These
31
considerations are not meant to disparage NEP systems, but simply to emphasize that practical
missions using NEP will need adequate nuclear reactor power.
Mixed (“hybrid”) NTR/NEP systems are also possible, e.g., see [Dujarric et al, 2000; Czysz
and Bruno, 2006] for details.
1.5.3. A Comparison between Chemical and NTR/NEP Isp
Comparing different sytems may yield different results depending on the choice of criteria.
Here a quick comparison is based on specific total impulse, Itot,s a mission-linked parameter defined
as
Itot,s ≡ Isp toperation /(Mp + m)
where toperation is the mission “engine-on” total time.
Therefore Itot,s = Isp/(dm/dt), and since P = F Isp and F = Isp (dm/dt),
Itot,s = Isp3/P = Isp3 ηtot/PR
The dimensions of Itot,s , distance per unit mass, classify it a fuel economy parameter, similar
to that for cars (miles/gallon, or km/liter). The dependence on Isp3 is to be expected but still
noteworthy.
The Itot,s normalized with respect to that of the best chemical propellants (LOX/LH2) will be
called the performance index, I, shown in Table III:
Table III Performance index, I, of some propulsion systems (fixed PR)
The price of high I is low absolute thrust, F Isp = P/ηR. Thus choosing the propulsion system
for an actual future mission should be based on solving a set of equations similar to the set (15),
although far more sophisticated.
32
1.6. Massless (Photonic) Propulsion
Because of their potential Isp and public attention, it may be of interest to discuss nuclear
propulsion systems not ejecting mass. These systems were dubbed “massless”, although this word
needs to be qualified.
“Massless” propulsion was proposed by E. Saenger in Germany [Saenger, 1956]. His
proposal was then popularized in the mid- and late ‘50s in a number of European magazines (this
author read it while in high school).
Saenger assumed annihilation reactions (protons p – antiprotons p-, thus α = 1) as the
ultimate energy source maximizing J. Annihilation produces radiation in the form of high energy
photons (but also muons, pions and neutrinos). Photons travel at the maximum speed known, c.
When collimated in a beam by a parabolic mirror hosting the energy source in its focus, photon
emission becomes the spacecraft propulsion system. It works as such because photons emitted from
the source invert their momenta when are reflected by the mirror: the spacecraft recoil is its thrust.
If P is the thermal power, Ep = hν the energy assumed uniformly imparted to all photons,
and if all photons are perfectly reflected and collimated, the flux Φ of photons producing thrust F is
Φ = P/Ep (photons/s)
and since the momentum of photons is hν/c = Ep/c, the thrust is
F = P/c (17)
Note that in equation (17) the contribution of massive particles (pions, muons,…) has been
neglected.
This thrust does not depend on photon wavelength, only on power. Equation (17) is intuitive
when observing that the ‘exhaust’ V of photons is exactly c. Alternative explanations may be cast
also in terms of the Poynting vector (electro-magnetic pressure), or based on E = mc2: that is, a
photon of energy E = hν has a ‘virtual’ mass hν/c2, so ejecting photons is the same as ejecting
particles of such mass (this mass is ≈ 9 x 10-29 kg for a hν = 5 MeV photon, a typical fission-
emitted gamma-ray).
A perfectly collimating mirror and an ideal reactor do not exist; however, photonic thrust
may still be produced with some efficiency (to be determined). The photon momentum concept
completes the discussion of thrust when α = 1 initiated in Section 1.3.2. It helps also in
understanding the principles of inertial confinement fusion and of pulsed nuclear propulsion.
A year later, Saenger [Saenger, 1957] proposed gas-phase fission (α = 9.1 x 10-4) as energy
source, since matter-antimatter reactions seemed too far into the future. Performance estimated was
surprisingly realistic. In [Saenger, 1956] the author mentions as ‘exhaust’ velocity, or Isp, 250
km/s, not c as one would expect based on the fact that photons move at light speed. The reason is
that, in calculating Isp, Saenger included also the contribution from heavy fission fragments,
assumed exhausting from a duct coaxial with the mirror axis. Isp is the average of two very
different exhaust speeds, heavily weighed towards that of fission fragments. As a footnote, he
added his estimate of Isp = 2.25 x 106 m/s if fusion, rather than fission, was the energy source. The
ROM of both these estimates is, in fact, correct.
Quite indipendently, C. Rubbia rediscovered photonic propulsion after looking for ways of
improving his fission fragments rocket concept. He proposed to heat a surface using a fission
reactor. This surface should radiate as much as possible as a black body. Rubbia’s concept, just like
Saenger’s, envisages a paraboloid mirror with the reactor at its focus. The paraboloid is much
33
deeper than in Saenger’s drawings, to capture as much as possible of the energy radiated
isotropically from the reactor. The power required by an ideal 26m-diameter mirror radiating at
about 3000 °C (about 0.25 eV) is about 3 GW, and thrust is 10 N.
This concept was presented at a European Science Foundation workshop held in Rome in
May 2002, and to the AAAF Propulsion Conference held in June of the same year in Versailles
[Szames, 2002; Rubbia, 2002]. It has been evaluated under an ESA contract.
It must be noted that even ‘massless’ photonic propulsion consumes mass: the fuel mass m0
fissioned. However, precisely because of the E = mc2 relationship between energy and mass, the Isp
calculated by using for d(m)/dt only the fraction αm converted into energy is exactly c, the photon
‘exhaust’ speed. However, not all fuel converts into pure radiation energy, only its fraction α. Short
of recovering it and reprocessing it, the remainder fraction (1 - α ) is effectively lost to propulsion.
In a solid-core fission reactor this fraction stays inside the fuel rods and cannot even work as inert
propellant (z = 0).
Therefore, accounting for fuel consumption, the specific impulse becomes Isp = α c, a lower
but still respectable 2.73 x 105 m/s using fission. Fusion has a higher α = (3 - 4) x 10-3, and Isp
would be in the 1 x 106 m/s range. If the fusion rocket is of the magnetic mirror type the fuel not
fused is ejected, μ is non-zero, total thrust is larger, but ideal Isp is lower. Apparently, the effect on
Isp of unburnt fuel was not mentioned by C.Rubbia.
Thus consumption of nuclear fuel must be included when estimating performance of

‘massless’ photonic rockets (and of more conventional nuclear propulsion systems as well, for very
long missions). The inherently low thrustof NEP means engine-on times may be very long. Since
mass consumption is PR/J, see equation (13), it depends of course on α.
Using fission, PR/J is about 1.21 x 10-5 kg/s per GW of reactor power. A single day of
operation consumes 1.45 kg of fuel per GW. A 3-GW system, e.g., like that proposed by Rubbia,
would consume more than13 ton of nuclear fuel over a 10-year mission, a reasonable time
yardstick, because the ideal F is only 10 N.
1.7. Conclusions
The very concept and the fundamental physics of NP has been established almost a century
ago, and current technology is already capable of implementing it in propulsion systems. These may
be NTR, where thrust may be very large, at the expense of a moderate Isp, or NEP, where the
reverse is typical. Mixed-mode operation appears also feasible,albeit still unproven. Choice of
system depends on mission; for manned interplanetary travel either advanced (future) NTR or NEP
are available now or will be in the next 10 to 15 years with reasonable investment in some areas
such as materials. In both cases the power required by fission reactors must be substantially higher,
by a factor roughly ten to a hundred, than that of some interplanetary, NP-based missions being
considered at this time. This conclusion is also relevant to manned and unmanned fast missions, the
only ones ensuring continuing interest and financial support from the public.
In this context, any encouraging technical considerations and their offshoots or conclusions
must be tempered by the unwanted NP effects, mainly radiation and its associated popular fears.
Short of conveying timely and correct information, and involving the public in future decisions
concerning NP, the common perception of anything nuclear, right or wrong it may be, will affect
adversely future choices, and prevent exploiting its potential.
34
Claudio Bruno,
Department of Mechanics and Aeronautics, University of Rome “La Sapienza”, Via Eudossiana 18,
00184 Roma Italy
Claudio.Bruno@uniroma1.it
1.8. References
Aerospace America, 2004: see roundtable discussion on NP at: http://boss.streamos.com/

wmedia/federal/aiaa/aiaa081004.wvx
ASD Report (2005), “Space R&T priorities for Europe – Recommendations, Technology Priorities,
Technology Roadmaps prepared by the European space manufacturing industry”,
www.eurospace.com.
Andrenucci, M., (2004), “Prospective Needs and Technology Options for High Power Devices”,
paper presented at the Internat. Symp. on Energy Conversion Fundamentals, 21-25 June 2004,
Istanbul. Available from Alta@alta-space.com.
Augelli, M., Bignami, G., Bruno, C., Calligarich, E., De Maria, G., Mulas, M., Musso, C.,
Pellizzoni, A., Piperno, W., Piva, R., Procacci, B., Rosa-Clot, M., and Rubbia, C., (1999), “Report
of the Working Group on a Preliminary Assessment of a New Fission Fragment Heated Propulsion
Concept and its Applicability to Manned Missions to the Planet Mars (Project 242)”, ASI Internal
Report, Roma, March 15, 1999 (Proprietary).
Auweter-Kurtz, M., and Kurtz, H., (2002), “High Power and High Thrust Density Electric
Propulsion for In-Space Transportation”, in: Proceedings of the International Workshop
“Technology and System Options Towards Megawatt Level Electric Propulsion”, June 9-10, 2003,
Lerici, Italy. Available from alta@alta-space.com. CD-ROM only.
Agrawal, J.P., (1998), “Recent Trends in High Energy Materials”, Progress in Energy and
Combustion Sciences, Vol. 24, pp. 1 – 30.
Bidault, C., Bond, R., and Sweet, D., (2004), “Assessment of Electric Propulsion Systems For
Exploration Missions: Comparison Between Solar-Electric and Nuclear-Electric Propulsion
Systems”, AURORA Final Report to ESA-ESTEC, July 15, 2004.
Cornille, P., (1999), “Review of the application of Newton’s third law in physics”, Progress in
Energy and Combustion Sciences, Vol. 25, pp. 161-210.
Czysz, P. A., and Bruno, C., (2000), “ Interaction of the Propulsion System and System Parameters
Determines the Design Space Available for Solution”, paper IAF-00-S5.07, presented at the 51st
International Astronautical Congress, 2-6 Ocober 2000, Rio de Janeiro.
Czysz, P. A., and Bruno, C., (2006), “ Future Spacecraft Propulsion Systems”, Springer-Praxis,
London, Chapter 7 (in print).
Davenas, A., Boury, D., Calabro, M., D’Andrea, B., and Mc Donald, A.J., (2000), “Solid
propulsion for Space Applications: A Roadmap”, Paper IAA.3.3.02, presented at the 51st
International Astronautical Congress, 2-6 October 2000, Rio de Janeiro.
35
Del Rossi, A., and Bruno, C., (2004), “Safety Aspects in Nuclear Space Propulsion”, IAC paper
IAC-04-R.4/S.7.07, presented at the 55th International Astronautical Congress, Vancouver, Oct. 4-8,
2004.
Dewar, J. A., (2004), “To the End of the Solar System – The Story of the Nuclear Rocket”, The
University Press of Kentucky, Lexington KY.
Dujarric, C., Fratacci, G., and Valentian, D., (2000), “Hybridisation of Chemical, Nucleothermal
and Electric Rocket propulsion principles: A Possible Way to Increase Rocket Specific Impulse?”,
paper IAF-00-S.6.02, presented at the 51st International Astronautical Congress, 2-6 Ocober 2000,
Rio de Janeiro.
Dyson, F. (1979), “Disturbing the Universe”, Harper and Row, New York, Chapter 10.
Dyson, G., (2002), “Project Orion”, Allen Lane – The Penguin Press, London.
Einstein, A., (1916), “Uber die spezielle und allgemeine Relativitaetstheorie

(gemeinvestaendlich),”, Newton-Compton Publishers, Rome, Sections 11 to 16.
FAS (Federation of Atomic Scientists), (2005), http://www.fas.org/nuke/space/c02early.htm
Flora, M. (2005), “Project Orion”, www.islandone.org/Propulsion/ProjectOrion.html
Howe, S.D., DeVolder, B., Thode, L., and Zerkle, D., (1998), “Reducing the Risk to Mars: the Gas
Core Nuclear Rocket”, in: Space Technology and Applications International Forum-1998, ed. by
Mohamed S. El-Genk, Publication CP-420, The American Institute of Physics, New York, p. 1138.
Kammash, T., (1995), “Principles of Fusion Energy Utilization in Space Propulsion”, in: Fusion
Energy in Space Propulsion, ed. by T. Kammash, AIAA Progress. In Astron. and Aeronautics.
Series, Vol. 167, AIAA, Washington, Chapter 1.
Koroteev, A. S., editor, (2002), “Rocket Engines and Powerplants Based on Gas-core Nuclear
Reactor”, Mashinostroenie Publ. House, Moscow. (in Russian).
Lang , J.R., (1999), “Astrophysical formulae: a compendium for the astronomer, astrophysicist and
physicist”, 3rd ed., Springer-Verlag, Berlin, Section 5.5.
Mukhin, K,N., (1983), “Experimental Nuclear Physics – Vol I Physics of Atomic Nucleus”, Mir
Publishers, Moscow, Ch. 2.
Oleson, S., and Katz, I., (2003), “Electric Propulsion for Project Prometheus”, AIAA Paper 2003-
5279, presented at the 39th AIAA/ASME/SAE/ASEE Joint Propulsion Meeting, 20-23 July 2003,
Huntsville, AL.
Randolph, T.M., and Polk Jr., J.E., (2004), “An Overview of the Nuclear Electric Xenon Ion
System (NEXIS) Activity”, AIAA paper 2004-3450, presented at the 40th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference, 11-14 July 2004, Ft. Lauderdale, Florida.
Romanelli, F., Bruno, C., and Regnoli, G., (2005), “Assessment of Open Magnetic Fusion for Space
Propulsion”, ESTEC Contract 18853/05/NL/MV Final Report, ESTEC, Noordwijk, September 20,
2005.
36
Ronen, Y., (2000), in: Nucl. Instr. and Meth. in Phys. Res. A, Vol. 455, pp. 442-451). See also
http://rense.co./general6/earthom.htm.
Rubbia, C. (2002), “ A Nuclear Propulsion Concept”, paper presented at Session 24, 6th
International Symposium on Porpulsion for Space Transportation of the XXI Century, Versailles,
14-17 May 2002. Available from AAAF, secr.exec@aaaf.asso.fr
Saenger, E., (1956), “Die Erreichbarkeit der Fixsterne”, in: Rendiconti del VII Congresso
Internazionale Astronautico, Associazione Italiana Razzi, (Proceedings of the VII International
Astronautical Congress), Rome, pp. 97-113. Also in: Mitteilungen der Landesgruppe Nordbayern
der DGRR vom 13.05.1958.
Saenger, E., (1957), “Zur Flugmechanik der Photonenraketen”, Astron. Acta, Vol. 3, pp. 89-99.
Schmidt, G.R., Bonometti, J.A., and Irvine, C.A., (2002), “Project Orion and Future Prospects for
Nuclear Pulsed Propulsion”, J. Prop. and Power, Vol. 18, No. 3, May-June 2002, pp. 497-504.
Smith, B., and Anghaie, S., (2004), “Gas Core Reactor with Magnetohydrodynamic Power System
and Cascading Power Cycle”, Nuclear Technology, Vol. 145, No. 3, pp.311-318.
Szames, A., (2002), “La fusée photonique à l’épreuve de la critique”, Air et Cosmos, No. 1850, July
5, 2002, pp. 34-35.
. quel che segue qui sotto in giallo si puo’ probabilmente cancellare
NASA (2005a), www.nasa.gov/vision/space/travelinginspace/25aug_plasticspaceships.html
NASA (2005b), www.radiationshielding.nasa.gov
Casali, D., and Bruno, C., (2004), “Superconducting Materials Applied to Electric Propulsion”, J.
Spacecraft and Rockets, Vol. 41, No. 4, pp. 671-676.
Gafarov, A.A., Gorshkov, O.A., Rozhdestvensky, N.M., Kudriashov, V.A., Skryabin, M.I.,
Bachmanov, M.M., and Fedotov, G.G., (2004), “Conceptual Project of the Interplanetary
Spacecxraft with Nuclear Power System and Eectric Propulsion System for Radar Sounding of Ice
Sheet of Europa, Jupiter satellite”, paper IAC-04-R.4-S.7.02, presented at the 55th International
Astronautical Congress (IAC), Vancouver, Oct. 4-8, 2004.
37
2. Nulcear Thermal Rocket Propulsion Systems
2.1. ABSTRACT
Ideas for using nuclear energy for space propulsion in thermal form (nuclear thermal rockets,
NTR) began shortly after the first controlled nuclear chain reaction in 1942. Starting in the late
1940s, several development programs were pursued by the United States Air Force, the Atomic
Energy Commission (now the US Department of Energy), and NACA (later: the National
Aeronautics and Space Administration) and they will be summarized here. This chapter does not
want to revisit history or deal with the engineering challenges and technologies of NTR investigated
in the past or proposed. In fact, a number of excellent reviews have already appeared, for instance,
on the history of the ROVER and NERVA programs that developed nuclear thermal rockets (NTR)
in the US [Gunn, 2001; Gunn and Ehresman, 2003; Dewar, 2004] and in the USSR [Goldin et al,
1991; Rachuk, 1996; Ponomarev-Stepnoy, 1999; Demyanko et al, 2001; Koniukov et al, 2004].
Engineering and science of NTR have been treated extensively in [Bussard and DeLauer, 1958; Hill
and Peterson, 1970; Lawrence et al, 1995; Koroteev, 2002].
Some of the systems developed from these programs are still in use today. To use nuclear
power for space propulsion, a propellant is heated in a suitable nuclear reactor to create hot, high-
pressure gas which is expanded through a nozzle. Nuclear reactors, at the simplest level, are heat
sources; they can heat a propellant directly (nuclear thermal) or create electricity (nuclear electric).
The resultant high thrust and high specific impulse enhance or enable missions which may not be
feasible using conventional chemical rocket engines.
2.2. Introduction
Most rockets are thermally driven gas devices in which energy is added in the form of heat.
This heat energy ejects propellant from the engine, giving us the required momentum exchange or
thrust. Energy can come from any number of sources. In chemical propulsion, the propellant
releases energy through combustion. In a nuclear rocket, the propellant heats up when energy
releases from the controlled fission of uranium or other fissionable material, see Chapter 1.
Fission involves the absorption of neutrons in a fuel material such as uranium. This
absorption excites the uranium atom until it splits into fragments and releases, on average, two new
nuclei and one to three free neutrons. The fission fragments have high kinetic energy from the
release of nuclear binding energy. This energy becomes thermal energy through collisions and
interactions with other atoms. The neutrons also give up kinetic energy and slow down so they can
be absorbed into the other fuel material. This process occurs more readily in lighter materials such
as carbon, hydrogen, and beryllium because of their cross sections. If each fission results in one
other fission event, the core is said to be critical. Neutrons can either be absorbed by other engine
materials or can leak from the reactor. The neutrons that leak out are lost from the cycle. Two or
three neutrons are usually released in each fission event to ensure that at least one is absorbed by the
fuel and causes another fission event.
38
Figure 7 The Fission Chain Reaction [Angelo and Buden, 1985]
The thermal energy produced from fission transfers to the coolant or propellant. For nuclear
rockets, we refer to the solid uranium as the fuel and to the gas, such as hydrogen or ammonia, as
the coolant or propellant. Conduction across the fuel material and convection into the coolant can
heat the coolant gas to high temperatures (3000 K), limited only by the requirement to keep the fuel
system below the fuel’s melting point. The following sections explain these concepts of fission and
heat transfer in more detail.
We can envision the nuclear rocket as a simple cold-gas thruster with a heat source added.
As the propulsion system “fires” to generate thrust, acceleration, and velocity change, it consumes
large quantities of propellant. To compare the efficiency of these different systems, we use specific
impulse. To increase specific impulse, the gas must have either a higher exit temperature or lower
propellant molecular mass. Nuclear propulsion offers an advantage over chemical systems because
we can choose the propellant with the lowest molecular mass. We can still impart large quantities of
thermal energy to get high exit velocity without worrying about the combustion properties.
Nuclear-fission rockets can have a specific impulse double that of chemical rockets. To get
this advantage, we usually choose a lightweight gas, such as hydrogen, as the reactor
coolant/propellant, but we can use higher-density propellants, such as methane, whenever storage
volume is limited. Figure 2 shows fission rockets can produce high thrust levels (low specific mass)
with good specific impulse. Having high specific impulse, high thrust, and high thrust-to-weight
ratio is a tremendous advantage for a propulsion system. Systems with high specific impulse but
high specific mass and low thrust, such as those using electric propulsion, require trip times of
hundreds of days to go from low-Earth orbit (LEO) to geosynchronous-Earth orbit (GEO). But
nuclear-propulsion systems need only hours.
39
Figure 8 Performance of Propulsion Systems. Fission rockets have better specific impulse than chemical rockets at
equivalent thrust-to-weight and specific mass ratios [NASA, 1990].
Mars mission—A manned mission to Mars has various ∆v requirements. A long mission
(hundreds of days) requires a ∆v of approximately 3.5 km/s from LEO to Mars. A 40-day transfer
from LEO to Mars requires 85 km/s. For a longer-duration mission, galactic radiation makes such
space travel hazardous. In addition, humans suffer physical and mental difficulties in a constant
free-fall environment (microgravity), so we must achieve the shortest possible trip time. Figure 9
assesses total radiation exposure in relation to trip time for hypothetical Mars missions using
chemical and nuclear propulsion. For both cases, the stay time on Mars’s surface is 30 days. But
because of the shorter length of the mission, we actually get reduced radiation exposure by using
nuclear propulsion, as compared with a mission using a conventional chemical rocket.
Figure 9 Comparison of Radiation Exposure for Nuclear and Chemical Systems. Nuclear systems can reduce overall
radiation exposure by reducing the trip duration. This shorter trip time occurs because the high specific impulse of
nuclear rockets allows higher ∆V’s for a given mission [Sager, 1993]. A rem (from roentgen equivalent man) is a
measure of radiation dosage based on the type of radiation an individual receives.
Nuclear propulsion systems can use any working fluid as a propellant and reactor coolant.
Hydrogen, ammonia, methane, octane, carbon dioxide, water, and nitrogen have been considered as
propellants. Because of their higher molecular mass, specific impulse is lower than for hydrogen.
However, these working fluids offer advantages for long missions, in which storability is an issue,
or for interplanetary missions, in which propellants could be acquired from a planet, moon, or
asteroid.
40
2.3. System Configuration and Operation
A nuclear rocket’s configuration is similar to that of a chemical system, except that it

requires a nuclear reactor as a heat source. Figure 10 shows a typical nuclear propulsion system,
which consists of a propellant tank, shielding, feed system (configured depending on type of engine
cycle), reactor, and nozzle. The tank and feed system are very similar to those in chemical systems.
The chemical rocket requires an oxidizer and fuel (each with its own feed system) for combustion to
a hot gas. The nuclear rocket works the same way, except that only one propellant feeds through the
core, where the reactor heats it to produce thrust. The main difference is engine control because the
gas’s thermodynamic conditions are coupled to the reactor and its unique requirements.
Figure 10 Schematic of a Nuclear Rocket. A nuclear rocket operates as a monopropellant liquid system, with the
nuclear reactor as a heat source.
Figure 11 shows a schematic of a reactor for nuclear propulsion. The reactor is complicated
by various components needed to keep the fission reaction under control. Let us discuss the major
reactor components that differ from those of a chemical system.
Figure 11 Schematic of a Reactor in a Nuclear Propulsion System. All elements work together to control the neutron
population and the energy level of individual neutrons. This control makes sure we have a reliable source of heat
energy ([NASA, 1990]).
41
Radial reflector—On the outside of the core is a radial reflector. To have a controlled chain
reaction (neutrons produced = neutrons used) and to reduce the size of the core, a reflector reflects
neutrons produced in the chain reaction back into the core. We must prevent them from leaking out
of the system in a large enough quantity to destroy the neutron balance and cause the reactor to shut
itself down. The reflector is usually made of beryllium.
Reactor pressure vessel—The reactor vessel is needed to maintain reactor pressure (3 MPa–8
MPa). It is made of aluminum or composite material to withstand the high radiation, heat flux, and
pressures from the reactor. The vessel may require cooling to support the heat flux in some reactor
designs.
Moderator—Reactors are said to be either thermal or fast, depending on the neutron energy
with which most of the fissions take place. In a thermal reactor, most of the fissions are caused by
neutrons having an energy less than 1 eV. Most neutrons produced from a nuclear-fission reaction
have energies much higher than 1 eV. To slow the neutrons down, we use a moderator assembly
made of a material with a low atomic mass (beryllium, plastics, lithium hydride, graphite). In a fast
reactor, the energy range in which most of the fissions take place is much wider, extending from
100 keV to the top range of the fission spectrum (15 MeV). If we wish to build a fast reactor, we
avoid light elements (moderating materials) and have no moderator. Some reactors effectively mix
moderating material with fuel material to limit the system’s size and mass.
Fuel-element assembly — The fuel-element assembly contains the uranium fuel and
propellant/coolant flow channels. The fuel produces the heat to be transferred to the propellant
flowing past the fuel. Different configurations of a fuel element can take advantage of surface area
to better transfer heat and to make sure some kind of barrier contains the fission products. The
reflector, control rods, and moderator are placed around the fuel to maintain the proper flow and
control of neutrons.
Control rods or drums — The control rods or drums contain materials (usually boron) that
absorb neutrons to decrease the neutron population. The rods control the reaction rate and can shut
the reactor down. This material is known as a “poison” because it lowers the number of fission
reactions when inserted in the core. The rods are dispersed around the core to ensure the neutron
population can be properly controlled and adjusted to meet engine power level requirements. The
control rods can be inserted into the reactor axially or rotated. For the axial insertion, the depth of
the rods controls the amount of neutrons captured. For the rotation insertion, one side of the rod
contains boron, whereas the other side contains beryllium. When the boron side is rotated into
place, neutrons are absorbed. When the beryllium side is rotated, neutrons are reflected back into
the core.
Coolant flow paths—Coolant piping cools components and provides the propellant gas
needed to generate thrust from the reactor. We usually want propellant to be completely vaporized
before it enters the reactor core.
Now that we basically understand most of the components associated with the reactor, let us
look at the operation of the core. The core is placed on the launch pad with the control poison fully
inserted. With the control poison in this position, the core produces no power and has negligible
radioactivity (only the natural radioactivity of the fuel). This condition allows workers to handle the
reactor with no protective shielding. Once the mission requires thrust, the control poison is
withdrawn (either lifted vertically or rotated outward) and a neutron source is put into the reactor to
provide the initial neutrons for fissioning. With the control poison withdrawn, the fissioning causes
the thermal power to increase exponentially to the desired power level. When the poison is
42
removed, the feed system immediately supplies gas to the core for cooling and thrust production.
When the reactor reaches the desired full-power level, the control poison’s position is adjusted to
keep the power at a steady-state (number of neutrons produced = number of neutrons used). This is
a delicate balance.
At mission’s end, the control poisons are inserted back into the core and the power decays
exponentially. But the reactor needs to cool down for some time, depending on how long the core
was at full power, and it may or may not need active cooling. This cooling may be necessary
because of delayed neutrons and residual heat production that result from radioactive decay of by-
products from the fission process (“fission products”).
Concepts
The following section discusses nuclear systems that have been developed or proposed for
propulsion applications in space, focusing (for lack of detailed information in English) on US
activities. Many other possibilities exist (18 concepts proposed by NASA [1990]), but we believe
these are the most likely candidates for near-term missions.
NERVA Derivative or Enabler
The NERVA (Nuclear Engine for Rocket Vehicle Applications) is a modified design of the
reactor used in the NERVA program (NERVA-1), see [Gunn, 2001; Gunn and Ehresman, 2003;
Dewar, 2004].
The NERVA program started in 1947 under the U.S. Air Force to design a reactor that could
propel intercontinental ballistic missiles (ICBMs). In 1958, NASA took control of NERVA as part
of their space-exploration program. NASA ran the program until 1972, achieving:
- 28 full-power tests with restarts

- Up to 30-minute test duration, total life 90 minutes
- Reactor sizes ranging from 300 MW to 200,000 MW
- Design of a hydrogen flow system
- Development of a way to contain affluents (propellant exhaust gases) for safe ground
testing (used only in later tests)
- Development of a high-temperature fuel (5500 F)
- Solution to the problem of fuel erosion from hot hydrogen
- Specific impulse levels as high as 835 seconds
- Thrust levels as high as 890,000 N
- Thrust to weight ratios of 3 to 4
In all, 23 tests were conducted at the Nevada Test Site, Nuclear Rocket Development
Station, NRDS. Figure 12 shows one of the NERVA Reactors undergoing testing.
Born from the need for carrying the heavy thermo-nuclear warhead of the early ‘50s, the US
nuclear thermal rocket (NTR) program found itself without a military mission as soon as smaller
warheads and larger ICBM came on line. Re-orienting its goals towards space applications took
time, while budget limitations at the time of the Vietnam war, and preliminary studies on what was
to become the Shuttle suggested in the early ‘70s that NTR were no longer needed and in any case
too expensive [Dewar, 2004]. This spelled the demise of large power NTR in the US, while work
was continuing in the USSR, where the RD 0410 and RD 0411 NTR were in fact developed and
tested. However, after the US reduced the scope and magnitude of its NTR work, even in the USSR
funding and activities declined, see for instance [Demyanko et al, 2001; Koniukov et al, 2004]. At
43
this time the only full-size NTR experience is therefore based on the NERVA-derived US engines
(the NRX family built by Westinghouse with Aerojet), and on the two USSR-built engines, for
which detailed information (although still in Russian) is slowly becoming available.
Figure 12 Test of the XE nuclear rocket engine at the Engine Test Stand at the Nevada Test Site. The engine had a
thrust of 75,000 pounds and was down firing mounted under a 75,000 gallon tank of liquid hydrogen. The exhaust
fired down into a "steam trench." The ETS was constructed only of aluminum to reduce the amount of long lived
radioactivity in structural materials induced by the neutrons coming from the engine.
The Table below, from the Report shown in Figure 13, shows a brief ROVER and NERVA
testing history. A concise account of their problems and technical developments is reported in
[Gunn and Ehresmann, 2003], while [Dewar, 2004] includes the history and politics of the ROVER
project. The technical reports concerning the rocket prototypes developed under NERVA by
industry may be found in [Westinghouse, 1972].
Reactor and Engine Systems Tests

Name Date
Phoebus 1B (one power test) Feb 1967
Phoebus 2 (cold flow tests) Jul 1967 - Aug 1967
NRX-A6 (one power test)* Dec 1967
XECF (cold flow tests) Feb 1968 - Apr 1968
Phoebus-2A (three power tests) Jun 1968 - Jul 1968
Pewee-1 (two power tests) Nov 1968 - Dec 1968
XE (28 starts) Dec 1968 - Aug 1969
*Operated 60 minutes at full power (1,100MW)
44
Figure 13 Handling of experimental rocket reactors at the NRDS in Nevada was accomplished by rail. Here the
reactor XECF is being moved to the static test site.
The reactor developed for the NERVA program in the US contains approximately 300
hexagonally shaped, graphite fuel elements in which uranium-carbide fuel particles coated with
pyrocarbon are disbursed (see Figure 14). The fuel particles provide the heat source while the
graphite matrix serves as the moderator and structural component of the fuel elements. Niobium-
carbide or zirconium-carbide coatings protect the surfaces of the fuel elements from the hydrogen
propellant. Twelve rotary drums in the radial reflector control the core. The drums have boron
plates which rotate in toward the core or out toward the perimeter, as required, to absorb and control
the neutron population. A shield containing boron carbide, aluminum, and titanium hydride is at the
top of the reactor. This shield limits nuclear-radiation heating of the engine assembly’s nonnuclear
components, including the propellant in the storage tank.
45
Figure 14 NERVA Fuel Element. This figure shows how individual fuel elements are configured and nested
together with additional elements for structural support [NASA, 1990].
2.4. Particle-Bed Reactor
The particle-bed reactor (PBR) (see Figure 15) is a US reactor concept originally developed
by the Air Force, and designed to provide a core with a high power density (by increasing the
propellant’s temperature and the fuel’s surface area) and a nuclear rocket engine with high thrust-to-
weight. The core consists of a number of fuel particles (Figure 9) packed in a bed and surrounded
by hexagonal moderator blocks arrayed in a cylindrical assembly (Figure 15). Its distinguishing
feature is that the hydrogen propellant directly cools small (200–500-mm diameter), coated,
particulate fuel spheres. The uranium-carbide fuel (coated with graphite buffer layers and an outer
layer of zirconium hydride) is packed between two concentric, porous cylinders called frits, which
confine the fuel but allow coolant flow. These small, annular fuel elements rest in a cylindrical
moderator block. Candidate materials for the moderator block are beryllium or lithium hydride.
Coolant flow moves radially inward through the cold frit, the packed bed, and the hot frit. At the
same time, it moves axially out through the inner annular channel and then expands through the
nozzle to produce thrust.
Figure 15 Configuration of a Particle-bed Reactor. This illustration shows the details of a) a fuel particle; b) a fuel
element; c) flow through the moderator block to the fuel bed; d) an assembled engine [NASA, 1990 and Maise,
1995].
46
The PBR’s advantages come from its high specific impulse, thrust, and thrust-to-weight ratio
(~40:1). This performance enables missions that the NERVA, with its lower thrust-to-weight (~5:1)
cannot do. The PBR has undergone several small proof-of-concept tests but has not had a full-scale
engine test. Because the PBR and NERVA have had the most money invested in them, they
represent near-term options. Most of the PBR development was done as part of the Air Force Space
Nuclear Thermal Propulsion program (SNTP).
This program lasted until 1993, with a budget of about $40 million per year. It intended to
design a particle-bed reactor for various US Air Force missions, but it ended having achieved:
- Fuel tests to temperatures as high as 3000 K (NERVA only reached 2650 K)

- Nuclear tests of single fuel elements
- Criticality experiments for a prototype 1000-MW core
- Tests of thermal hydraulics in multiple fuel elements at high power densities (40
MW/liter as compared to about 2 MW/liter for NERVA)
- Various mission designs (detailed analysis of a stage for a mission to Mars)
- Verification of computer codes for simulating reactor physics
2.4.1. CERMET
The USAF-developed CERMET-core nuclear rocket (see Figure 16) uses a fast fissioning
spectrum (greater than 1 MeV) compared to thermal reactors that slow down neutrons to fission
energies of less than 1eV. Therefore, it does not need a moderator. The CERMET has a lower
thrust-to-weight than the particle bed and has not been tested as extensively as the NERVA-type
engines. Fuel tests show the CERMET-type fuel is much more robust than that for either NERVA
or PBR. This feature makes CERMET attractive for applications such as a reusable orbital-transfer
vehicle (OTV), for which we may need as many as 50 burns.
The reactor consists of hexagonal fuel elements similar to those in the NERVA design,
except that the fuel is made of uranium-dioxide particles imbedded in a tungsten or
tungsten/rhenium matrix. The advantages of CERMET fuel are its potential for very long operating
life (more than 40 hours), ability to restart, handling of temperature cycling at high temperature, and
greater compatibility between the fuel and hydrogen coolant (resulting in high fuel integrity and
retention of fission products). However, the metal matrix can increase system mass because of
competition for neutron absorption in uranium and tungsten. Using an axial, two-zone fuel element
reduces system mass. In this two-zone concept, the fuel loading in the upper (low-temperature) half
of the core uses a molybdenum/urania matrix configuration. The lower (high-temperature) part of
the core uses a tungsten-rhenium/urania matrix configuration. This concept reduces the system’s
mass and gives it a thrust-to-weight ratio of 5.3:1—slightly better than NERVA. The system, shown
in Fig. 10, consists of a CERMET core surrounded by a cooled pressure-containment shell. A
neutron reflector and reactivity-control assembly mounts to the outside of the pressure vessel. So
far, tests of the CERMET have checked only the fuel—up to 1900 K for 10,000 hours.
47
Figure 16 Configuration of the CERMET Engine. (a) shows details of the reactor (b) shows a split view of the
engine configuration, with interior details on the left side and exterior details on the right side [Bhattacharyya, et al.,
1988].
Of the reactors discussed here, NERVA-1 (the flight engine developed in the early 1970s)
offers the lowest performance in terms of specific impulse. But it has had the most money invested
in it and was ready for a flight before the program ended. The PBR offers the highest performance
for a solid-core design. The CERMET may be a good design to pursue if reusability becomes an
issue because its fuel lasts longer than other types of cores investigated. Table IV shows the
characteristics of the three engine concepts.
48
NERVA Particle bed CERMET
Power (MW) 1570 1945 2000
Thrust (N) 334.061 333.617 445.267
Propellant H2 H2 H2
Fuel element Solid rod Porous particle bed Solid rod
Maximum propellant temperature (K) 2361 3200 2507
Isp (s) 825 971 930
Chamber pressure (MPa) 3.102 6.893 4.136
Nozzle expansion ratio 100 125 120
Engine mass (kg) 10138 1705 9091
Total shield mass (kg) 1590 1590 1590
Engine thrust/weight (no shield) 3.4 20.0 5.0
Table IV Comparison Of Possible Near-Term Concepts for Reactors. The engine using a particle-bed reactor has
higher performance and is lighter than the NERVA, but NERVA is more developed [Clark, 1991]. CERMET is a
possible fast-reactor concept that we can also reuse.
Advanced concepts were studied and also partly explored using component hardware since
the NERVA times. Most of them are indeed studies, but should nevertheless be cited. Liquid- and
Gas-core fission reactors were investigated (see also Chapter 1, Section 1.4.1, and [Koroteev,
2002]. In USSR a gas-core reactor, the IVG-1 was in fact partially tested. Partial testing of the two
gas-core reactor concepts proposed took place also at the Los Alamos Scientific Laboratories
(LASL) in the US, see [Czysz and Bruno, 2006]. LASL is still continuing work in this area, albeit at
a low funding level. The main issues of gas-core fission are the heat transfer between fission
fragments and the hydrogen propellant and how to limit fissioning fuel leakage.
Recently, work at Aerojet has produced a mixed Nuclear-Chemical propulsion concept

(LANTR, Liquid Oxygen Augmented NTR). For relatively short thrust bursts, liquid oxygen is
injected downstream of the throat of a hydrogen propellant NTR. The supersonic combustion
between hydrogen and oxygen can substantially boost the engine thrust (at the expense of the Isp),
see [Czysz and Bruno, 2006] for a summary of this work. Interesting applications of LANTR
include liftoff from planetary bodies, emergency maneuvering, and asteroid intercepting, to name a
few. The LANTR concept has been tested using as NP-ersatz a conventional liquid H2 rocket
engine, and has demonstrated thrust may be augmented by 40-60%.
2.5. Safety
Nuclear energy is an unique discovery since safety concerns were addressed from the
beginning of the discovery. Safety is paramount in any nuclear program developed for space. The
US Navy Naval Reactor development program has shown that nuclear energy can be developed
with safety as a concern and a zero nuclear accident record. One of the attractive features of a
nuclear space program, is that the reactor is inert and can be handled like any typical payload (as
long as U235 is the fuel) until it is turned on in space. It represents a zero radiological hazard until
operation, and then the exposure is based upon the amount of time it operates.
There are a few myths on the use of a nuclear rocket in space. There is n UN treaty violation
for nuclear power or propulsion applications, just nuclear weapons. RTG’s have flown plutonium
49
238 safely for many missions. Despite being toxic both chemically and because of its ionizing
radiation, plutonium is far from being 'the most toxic substance on earth' or so hazardous that 'a
speck can kill'. There are substances in daily use that, per unit of mass, have equal or greater
chemical toxicity (arsenic, cyanide, caffeine) and radiotoxicity (smoke detectors). Isotopes of
plutonium such as Pu-238 characteristically give off short-range alpha particles, helium nuclei that
usually travel no more than about three inches in air. There are three principal routes by which
plutonium can reach human beings:
-Ingestion (not a significant hazard)

-Contamination of open wounds,
-Inhalation
Workers at US nuclear weapons facilities have come in contact with or inhaled plutonium.
Intensive health checks of these people have revealed no serious consequences.
A nuclear reactor in space often requires a radiation shield to protect the crew, payload, or
other spacecraft equipment sensitive to radiation. The unit that describes the damaging effects of
radiation is the rem:
rem = absorbed radiation dose x quality factor
The absorbed radiation dose is (see Appendix 7) traditionally defined in terms of the rad
(radiation absorbed dose). One rad equals the amount of radiation required to cause the absorption
of 100 ergs (1 joule = 107 ergs) per gram of material. Therefore, the “rad” is the amount of energy
imparted to a component or person. Experimental data has shown that different types of radiation
produce different effects. So we add the quality factor to account for the effect of a particular type
of radiation. Figure 17 shows the biological consequences of acute, short-term radiation effects
from whole-body exposure to gamma radiation.
To put these numbers into context, we can list some typical exposures:
-Natural radioactive material in the bones—0.034 rem/year

-Flight in an aircraft—0.001 rem/hr at a 9-km altitude
-Chest x-ray (lung dose)—0.01 rem
-Living one year in Houston, Texas—0.25 rem
-Living one year in Denver, Colorado—0.35 rem (higher elevation allows more
cosmic radiation)
-Working in a nuclear power plant for 1 year—less than 0.5 rem
-Skylab astronauts on 84-day mission—18 rem
-90-day space station mission—16 rem (NASA estimate)
-Van Allen belts—16 rem/year
-Cosmic radiation outside of the Van Allen belts—19 rem/year
-A single solar flare with moderate energy protons—303 rem/year
-A properly shielded nuclear space engine—10 rem/year
Acute Irradiation
Level (rem) Acute Somatic Effect

15–25 Subtle reductions in white-blood-cell counts; not generally apparent from
exposure for one person unless a blood sample was taken before the exposure
50 Reduction in white-blood-cell count after exposure; the count returns to

normal in a few weeks
75 10% chance of nausea
50
100 10% chance of temporary hair loss
200 90% chance of radiation sickness; moderate depression of white-blood-cell

fractions
400–500 50% chance of death within 30 days without extensive medical treatment
>600 Lethal to most people in 3 to 30 days; even with extensive medical treatment,
death is likely within a few months from infection and hemorrhage
>10,000 Lethal within 24 hours from damage to central nervous system
Figure 17. Acute Radiation Effects From Whole-Body Exposure to Gamma Radiation. As the radiation dose
increases, the biological effect increases correspondingly. These somatic effects appear after exposure to acute doses
over short times rather than over longer periods, during which the cells can repair some damage.
Criteria General Public Occupational Workers Astronauts

30-day limit 0.04 rem 0.4 rem 150 rem
annual limit 0.50 rem 5 rem 300 rem
career limit N/A 5 x (age in years – 18) rem 600 rem
accident limit 25 rem 100 rem N/A
acute limit N/A N/A 10 rem
Table V Allowable Limits for Skin Dosage. The US National Committee on Radiation Protection and Measurement
has agreed to these allowables.
Table V shows allowable skin radiation dosages, as given by the US National Committee on
Radiation Protection and Measurement. Many factors influence the geometry, composition, and
mass of the radiation shield, including:
-Size and nature of the power source

-Type of radiation
-Configuration of the spacecraft or platform and its payload (radiation flux level
decreases by a factor of 1/distance2 from the radiation source)
-Generic operational procedures and requirements for the mission
-Length of the mission
-Total permitted levels of radiation dosage
As one key see, the key factor to reduce radiation exposure, is to have a high energy system
like a nuclear thermal rocket that can get the payload to the destination in the fastest way possible.
2.6. MagOrion and Mini-MagOrion
This technology goes back to the Orion project of the ‘50s and ‘60s. The Orion concept was based
on nuclear explosions for propelling a spacecraft. In essence, small atomic bombs were to be
released and detonated sequentially aft of a large spacecraft; the plasma from the explosion would
periodically impinge on a massive thruster plate, ablating its surface and pushing it forward. The
plate was designed to be connected to the spacecraft by means of shock absorbers, to soften and
reduce the acceleration pulses the crew would feel. This concept was proven viable, and its physics,
51
particularly that of ablation, was especially fascinating. Its history and technical issues is described
in [Schmidt et al, 2002; Dyson, 2002].
The Orion concept was revisited in the late ‘90s, with a major change: the pushing effect of a
nuclear explosion plasma, that in the original Orion took place because of ablation of the pusher
plate, was replaced by creating a Lorentz force in a superconducting (SC) ring. The force was the
result of the interaction between the plasma released by a miniature atomic explosion, and the
magnetic field. Calculations indicated that Isp of order 104s could be obtained with a 2-km dia.
ring, with a thrust/mass ratio between 0.2 to 10 [Ewig, 2003]. This concept was called the
“MagOrion” concept.
For practical reasons the SC material was to be a High Temperature SC, e.g., a YBCO ceramic. Its
brittleness, and the size of the ring suggested to scale down MagOrion to a more feasible concept,
Mini-MagOrion, or MMO, that is briefly presented below.
Technically, MMO is based on the following elements:
● unconventional fissile fuel, e.g., 245Cm

● Criticality reached by impulsive compression by a plasma Z-pinch, also moderating the yield
● Single large ring replaced by multiple small coils girding a MHD nozzle
Curium was chosen based on its neutronics, conversion fraction α (see Chapter 1), and also
availability.
Operation of the MMO may be schematically described as follows: current is made to flow along a
electrically conducting conical Mylar sheet, from the periphery toward the apex, where the mylar is
vaporized by the increasingly high current density and becomes plasma. This plasma pinches
(implodes) the nuclear fuel near the apex, compressing it impulsively so that it becomes critical and
fissions. Energy yield predicted is in the 50-500 GJ range per fission event.
The fuel is in the form of hollow 245Cm spheres surrounded by a neutron-reflecting Be layer. Upon
fissioning, the ionized debris and the current flowing in the superconducting rings bracing the
large, diverging B field inside the magnetic nozzle, create the Lorentz force pushing the spacecraft.
Isp depends on the energy yield almost logarithmically;
52
Figure 18
Figure 18 [Ewig, 2003] shows the payload mass fraction achievable as a function of Isp.
The MMO team examined all key aspects of this concept and found them workable, albeit with
many engineering problems still to be solved. A baseline concept ship, designed for a 100 ton
payload, has an initial mass about 730 ton. Details are reported in [Ewig, 2003].
In summary, provided all engineering problems may be solved, MMO is a futuristic but feasible
concept. Mass of the baseline vehicle mentioned would command a very high price for orbiting and
assembling all its components, but probably no higher than R, D&E costs. As with all advanced
propulsion technologies for interplanetary travel, the crucial step is reaching economically LEO;
however, once in orbit, MMO looks a radical twist to the general area of nuclear propulsion.
2.7. Conclusions
In the US President Bush signed into law a 2005 $16.2 billion budget for NASA, an
increase of about $822 million from 2004. This last is a package funding the entire Project
Prometheus program at $430 million and provides $10 million for nuclear thermal propulsion at the
Marshall Space Flight Center. In Greek mythology, Prometheus was the wisest of the Titans; he
gave the gift of fire to humankind (the name Prometheus means in Greek 'forethought’). Although
now given a lower priority, NASA is still developing plans for an ambitious mission to orbit three
planet-sized moons of Jupiter -- Callisto, Ganymede and Europa -- which may harbor vast oceans
beneath their icy surfaces. The mission, called the Jupiter Icy Moons Orbiter, would orbit each of
these moons for extensive investigations of their makeup, their history and their potential for
sustaining life. NASA's Galileo spacecraft found evidence for these subsurface oceans, a finding
that ranks among the major scientific discoveries of the Space Age. We hope that the ideas
53
considered in this Chapter will reinforce the case for nuclear propulsion, in particular nuclear
thermal propulsion, and support the revival of this technology.
Lt Col Timothy J. Lawrence

U.S. Air Force Academy
Department of Astronautics
2354 Fairchild Drive Suite 1M147
USAFA, Colorado, 80840 USA
Timothy.lawrence@usafa.af.mil
2.8. References
Angelo, Joseph and Buden, David. 1985. Space Nuclear Power. Orbit Book Company, a division of
Krieger Publishing Company, Malabar, Florida.
Benedict, M. and Pigford, T.H. 1957. Nuclear Chemical Engineering. McGraw-Hill Book
Company, New York, NY.
Bussard, R.W., and DeLauer, R.D., (1958), “Nuclear Rocket Propulsion”, McGraw-Hill Book
Company, Inc., New York.
Bhattacharyya, Samit et al. 1988. CERMET Reactor Orbit Transfer Vehicle Concept. USAF Report
AFAL-TR-88-033. Edwards AFB, CA: U.S. Air Force Astronautics Laboratory.
Clark, John S. 1991. An Historical Collection of Papers on Nuclear Thermal Propulsion. American
Institute of Aeronautics and Astronautics, Washington DC.
Czysz, P. A., and Bruno, C., (2006), “ Future Spacecraft Propulsion Systems”, Springer-Praxis,
London, Chapter 7 (in print).
Demyanko, Yu. G., Koniukov, G.V., Koroteev, A.S., Kuz’min, E.P., and Pavel’ev, (2001),
“Nuclear Rocket Engines” (in Russian), Norma Inform Publisher, Moscow. Chapter 1 contains a
short history of the nuclear rocket engine (“ARD” in Russian). Reactors developed are discussed in
Chapter 3.
Ewig, R., “Mini-MagOrion Final Report”, (2003)
Goldin, A. Ya., Koroteev, A.S., Semyonov, V.F., Konopatov, A.D., Pavshuk, V.A., and
Ponomarev-Stepnoy, N.N., (1991), “Development of Nuclear Rocket Engines in the USSR”, paper
presented at the AIAA/NASA/OAI Conference on Advanced Space Exploration Initiative (SEI)
Technologies, San Diego, September 4-6, 1991.
Gunn, S., (2001), “Nuclear propulsion – an historical perspective”, Space Policy, Vol. 17, Number
4, pp. 291-298.
Gunn, S.V., and Ehresman, C.M., (2003), “The Space Propulsion Technology Base Established
Four Decades Ago for the Thermal Nuclear Rocket is Ready for Current Applications”, AIAA
paper 2003-4590, presented at the 39th AIAA/ASME/SAE/ASEE Joint Propulsion Conference,
Huntsville, Alabama, 20-23 July 2003.
54
Dewar J. A. (2004), ”To the End of the Solar System: The Story of the Nuclear Rocket”, The
University Press of Kentucky, Lexington, KY.
Drake, M.K. 1970. Data Formats and Procedures for the ENDF Neutron Cross Section Library.
BNL-50274 (T-601). Brookhaven National Laboratory, Brookhaven, NY.
Gunn, S. V., and Ehresman, C.M., (2003), “The Space Propulsion Technology Base Established
Four Decades Ago for the Thermal Nuclear Rocket is Ready for Current Applications”, Paper
AIAA 2003-4590 presented at the 39th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, 20-
23 July 2003, Huntsville, Alabama.
Henry, Allan F. 1986. Nuclear Reactor Analysis. MIT Press, Boston, MA.
Hill, P.G., and Peterson, C.R., (1970), “Mechanics and Thermodynamics of Propulsion”, Addison-
Wesley Publishing Co., Reading, Mass., Chapter 15.
Knief, Rolland Allen. 1992. Nuclear Engineering: Theory and Technology of Nuclear Power.
Hemisphere Publishing, Washington, DC.
Koniukov, G.V., Petrov, A.I., Popov, S.A., Rachuk, V.S., Belogurov, Y.I., Mamontov, Yu.I., Fedik,
I.I., D’yakov, Ye..K., Mogil’ny, I.A., Konovalov, V.A., Raskach, F.L., and Zakarkin, I.I., (2004),
“Prototype of Atomic Rocket- IRGIT Reactor” (in Russian), Atomic Energy, Vol. 97, No. 3,
pp.173-177.
Koroteev, A. S., editor, (2002), “Rocket Engines and Powerplants Based on Gas-core Nuclear
Reactor”, Mashinostroenie Publ. House, Moscow. (in Russian).
Lawrence, T.J., Witter, J.K., and Humble, R.W., (1995), “Nuclear Rocket Propulsion Systems”, in:
“Space Propulsion Analysis and Design”, edited by R.W. Humble, G.N. Henry and W.J. Larson,
The McGraw-Hill Companies, Inc., New York, Ch. 8.
Los Alamos National Laboratory. 1986. Monte Carlo Neutron/Photon Transport Code. Los Alamos,
New Mexico.
Ludewig, Hans. 1993. Summary of Particle Bed Reactor Designs for the Space Nuclear Thermal
Propulsion Program. Brookhaven National Laboratory Report BNL-52408, Brookhaven, NY.
Maise, George, and Hans Ludewig. 1995. Brookhaven National Laboratory. Private
communication.
Ma, I. 1983. Materials for Nuclear Applications. McGraw Hill, New York.
NASA. 1990. NASA/DOD/DOE Nuclear Thermal Propulsion Workshop Notebook. NASA Lewis
Research Center (now: NASA Glenn RC), Cleveland OH.
Ponomarev-Stepnoy, N.N., Talyzin, V.M., Pavshuk, V.A., Putko, V. Ya., Konovalov, V.A.,
Raskach, F.L., Ulasevich, V.K., Smetannukov, V.P., Kolganov, V.D., Fedik, I.I., D’yakov, Ye. K.,
Mogil’ny, I.A., Rachuk, V.S., and Mamontov, Yu.I., (1999), “Rabotyi po Otiechiestvennogo ARD”
(in Russian), Atomic Energy, Vol. 86, No. 4, pp. 296-302. Note: this paper has a picture of the
famous “three K” (Korolev, Kurchatov and Keldysh). These scientists worked together at nuclear
thermal rockets development in USSR.
55
Rachuk, V.S., Belogurov, A.I., Grigorenko, L.N., and Mamontov, Yu. I., (1996), “Russian
Investigations in the Area of Nuclear Rocket Engines (NRE) Research International Programs“,
paper presented at the 5th International Symposium on Propulsion for Space Transportation, 22- 24
May 1996, Paris.
Westinghouse (1972), “Technical Summary Report of the NERVA Program”, Volumes I – VI,
Westinghouse Astronuclear Laboratory Publication WANL TNR-230, Pittsburgh, Pa.
Witter, Jonathan Keay. 1993. Modeling for the Simulation and Control of Nuclear Reactor Rocket
Systems. PhD Thesis, Massachusetts Institute of Technology, Boston, MA.
56
3. The application of ion thrusters to high thrust, high specific impulse nuclear-electric
missions
3.1. ABSTRACT
This report forms one part of a study by members of the IAA to examine the merits of nuclear
electric propulsion (NEP) for challenging deep space missions requiring a high value of velocity
increment, ΔV. The need for a large value of ΔV, typically of many km/s or even tens of km/s,
eliminates the use of chemical propulsion owing to the large mass penalty incurred by employing
this traditional technology. This mass penalty exists because the effectiveness of space propulsion
depends critically upon the exhaust velocity achieved by the engine utilised, or its specific impulse
(SI); basically, the higher the SI the smaller the propellant load required to complete the mission.
Whereas the best conventional chemical system, using liquid oxygen and liquid hydrogen, is limited
to an SI of about 470 s, a typical gridded ion engine (GIE) can readily yield 3500 to 6000 s, and
much higher values are being realised in preparation for future interplanetary missions. In
experimental work, values approaching 30,000 s have been reported.
It is significant that many missions which have been studied in depth require values of SI of
between 5000 and 10,000 s. These are clearly within the range that can be provided by GIEs, but
are currently well beyond those appropriate to Hall-effect thrusters (HETs) and arcjets. Moreover,
it is unlikely that typical existing magnetoplasmadynamic (MPD) thrusters can reach the higher of
these values. It is also worth noting that the exhaust velocity of a GIE is determined only by the
characteristics of the ion extraction grid system used, and that, in principle, any desired value can be
achieved merely by altering the applied voltages. Thus the GIE has the potential to be tuned exactly
to provide the required SI, thereby optimising the mission characteristics. No other relatively high
thrust EP system has this ability.
It should also be noted in passing that no electric propulsion system can match the high thrust
provided by almost any chemical engine. Thus chemical systems cannot be disregarded in designing
the missions and spacecraft of interest, since they will still be necessary for manoeuvres in which
high thrust is needed for specific purposes. An example would be to achieve planetary capture from
a situation in which the approach velocity is substantial and there is little remaining time in which
to conduct this manoeuvre.
This report therefore reviews in some depth the ways in which GIE systems might cover the power
range from a few tens of kW to several MW. By consideration of the relevant scaling relationships,
which have been validated up to about 30 kW, it is concluded that the same basic concepts will
suffice for this entire range of power levels. This conclusion is certainly valid for the Kaufman-
type of direct current (DC) discharge thruster, for the 1 MHz-type of radiofrequency ionisation
thruster (RIT), and for electron cyclotron resonance (ECR) ionisation thrusters which do not use
permanent magnets. Unfortunately, the severe temperature constraints on high field permanent
magnets suggest that the cusp-field type of DC thruster cannot operate at the very high power levels
considered here.
It is shown that the design process is aided considerably by the separation of the ion production and
extraction/acceleration regions in gridded thrusters. Thus the ion beam parameters can be deduced
without reference to the ionisation mechanism employed to produce the plasma from which the ions
are extracted. Consequently, the two regions of the thruster can be designed separately, which is a
simplifying benefit only available to gridded devices.
It is concluded that the required power density can be achieved and exceeded using GIEs, by
operating their grid systems at very high perveance and by raising the SI to values which are
significantly above those commonly employed at present. The latter ensures that most of the energy
supplied is used in accelerating the beam ions, thereby allowing extremely high values of electrical
efficiency to be achieved; in the limit, this parameter exceeds 99%. With propellant utilisation
efficiencies of above 85%, the overall thruster efficiency becomes in excess of 84%.
57
For the lower power applications, normal twin- or triple-grid configurations are entirely adequate if
higher perveance operation is adopted and if the SI can be increased above current values.
However, the limit here is due to the penetration of the plasma sheath at the innermost grid into the
discharge chamber plasma, which increases as the electric field becomes larger or the plasma
density lower. This greater curvature of the sheath, which effectively emits the beam ions, then
causes direct impingement of high energy ions onto the outer grids, severely limiting lifetime. With
this configuration, the maximum ion extraction potential is likely to be about 5 kV. If Xe propellant
is used, this gives an SI of approximately 7000 s, depending upon the propellant utilisation
efficiency achieved. A total of about 100 kW power consumption can then be realised with a small
array of thrusters.
As the applied potentials increase, a more complex but satisfactory alternative is to employ a 4-grid
system with both a greater perveance and providing an increased SI. In all cases, additional
research and development are needed, since this regime has not been properly explored to date. The
lifetime implications are of particular interest and may dictate the way forward.
Thus the operational envelope can be massively extended by use of the 4-grid configuration. There
is considerable documentation concerning such systems in the controlled thermonuclear research
(CTR) community, where ion accelerators of the size required, or smaller, have been constructed
which produce MW beams at 70 or 80 keV. This is possible because the 4-grid arrangement
permits the ion extraction process to be separated from the main acceleration region, and the
limitation of the sheath penetration no longer applies. Thus the extraction part of the system can be
designed to operate at near maximum perveance, and the subsequent further acceleration of the ions
can be dealt with independently in the design process. An additional advantage of this
configuration is the very low beam divergence typically found. This can be less than 1° at an
acceleration potential of 70 kV.
It is thus concluded that well understood thruster technology, when combined with the 4-grid
configuration based on that utilised in CTR ion injection machines, will permit MW power levels to
be achieved. Thus a relatively small array of thrusters, with beam diameters not exceeding 40 cm,
will be able to consume the several MW, although the SI, using Xe propellant, is likely to be
somewhat above 10,000 s. With this arrangement, the power density can reach 4.5 kW/cm2 and the
thrust density 30 mN/cm2 and, if required, the specific impulse can attain 30,000 s with Xe. If
higher values of SI are required, the utilisation of lower atomic mass propellants will permit this to
be achieved, with an ultimate limit using hydrogen compounds of about 150,000 s.
As a specific example, an array of nine 40 cm beam diameter thrusters using Xe propellant and
operating at 10 kV, but with an ion extraction potential of 5 kV, will consume 7.4 MW if the
perveance is limited to 50 % of its maximum value. The total beam current will be 702 A, the
thrust 120N, the SI 11,100 s, and the power density 644 W/cm2. If full perveance operation can be
achieved while maintaining long life, the power consumption becomes 14.7 MW.
To summarise, the design for a 20 to 50 kW thruster can be extrapolated directly from current
capabilities, taking advantage of the higher perveance values made possible by introducing
improved grid technologies. However, at the higher power level it would be advantageous to utilise
a 4-grid configuration. Thus an array of selected existing devices could probably consume up to
100 kW or so, although none of these thrusters are space qualified and they all require life-testing.
The work required to meet a multi-100 kW level, utilising the 4-grid configuration, would be much
more extensive, so would require more time and greater funding. A multi-MW system can be
implemented using extended scaling relationships, but there is no precedent for such an exercise.
Thus this complete process, encompassing design, development, performance testing,
environmental testing and life-testing, would require to be organised and funded. Bearing in mind
the need to co-ordinate such a development very closely with that of the power source, work on
both systems should be conducted in parallel, with very close collaboration at all times. This co-
operation is desirable because it is envisaged that the bulk of the power consumed by the thrusters
would be supplied directly from the source, with no further power conditioning.
58
Finally, it should be noted that the performance- and life-testing of large, high power thrusters
requires the use of major facilities with substantial pumping speeds, which are costly to build and to
operate. The provision of such facilities must precede the development of the thrusters.
59
3.2. Introduction
Deep space missions are characterised by the velocity increment, ∆V, required to accomplish
them. As this parameter becomes greater, the conventional rockets necessary to achieve the stated
objectives become larger and more expensive, with the result that very few missions of this kind
can be mounted, even by the countries and agencies with the largest budgets.
An excellent example of this problem is provided by the initial proposal by NASA/JPL to
send a spacecraft on a fast fly-by of Pluto [Staehle, 1992]. In this, a Titan 4 launcher aided by a
Centaur upper stage was required to deliver into an interplanetary trajectory a spacecraft with total
mass of only 160 kg. The payload was just 7 kg, with a power availability of 6 W, and the mission
time to Pluto was 7 to 10 years. Despite this time and the high cost, with the launch vehicle priced
in the range $250M to $400M in 1992, the data acquisition and scientific return would have been
very limited. The fly-by time over the last 100,000 km would have been only 1.7 hr, and crossing
the orbit of Charon, Pluto’s satellite, would have taken just 30 min. Clearly, a cost-benefit analysis
of such a mission would not produce encouraging results.
In order to reduce costs to affordable levels, the prime objective must be to reduce overall
initial spacecraft mass, so that smaller and less expensive launch vehicles can be employed.
Bearing in mind the very large values of ΔV required for the missions under consideration, and thus
the dominance of the propulsion system in determining total mass, this offers the greatest scope for
improvement. Indeed, this system, including the propellant, can account for 50 to 90% of the launch
mass. This is the primary role of electric thrusters, which can replace chemical engines in this role
with great advantage.
Thus one of the most exciting and challenging applications of electric propulsion (EP) is to
the exploration of the far reaches of the solar system and, perhaps eventually, of the nearest stars.
This process has begun, with the flight of NASA’s Deep Space 1 (DS-1) spacecraft [Rayman,
2000], which was propelled by a gridded ion engine (GIE) and completed an extended mission to
both an asteroid and a comet. Other exciting projects are currently underway, such as ESA’s
SMART-1 technology demonstrator mission to the moon [De Cara, 2004], in which a Hall-effect
thruster (HET) was used to transfer the spacecraft from a geostationary transfer orbit (GTO) to
lunar orbit, which was acquired during November 2004. Even more ambitious is Japan’s Muses-C
mission to return a sample from asteroid 1998 SF36 [Kawaguchi, 2004]. This spacecraft was
launched on 9 May 2003 by an M-5 rocket, and the velocity increment to be provided by the
electric propulsion system (EPS) will be about 3.7 km/s. The EPS consists of 4 small
radiofrequency (RF) ionisation GIEs, each producing a maximum thrust of 8.1 mN. However,
despite these impressive successes, as the energy required increases, the mass of propellant
becomes larger if present thrusters are used, resulting in a heavier spacecraft, a longer mission
duration, and a higher cost. Thus large increases in specific impulse (SI), and thus in exhaust
velocity, ve, are required.
This challenge has been recognised by NASA by commencing a search for completely new
ways of accomplishing space travel [Millis, 1999]. This, the “Breakthrough Propulsion Physics
Programme”, has examined many interesting suggestions, some involving extensions to the existing
laws of physics. However, no positive results offering a near-term solution to the problem outlined
above have so far emerged. It is therefore necessary, at the moment, to assume that such solutions,
if any, are probably decades away from implementation. It is therefore prudent to consider these
advanced missions employing extensions of existing technologies.
However, a superficial analysis of the missions in question indicates that the energy required
is vastly in excess of that available from solar arrays or any other non-nuclear substitute. In
addition, a large proportion of the most challenging missions involve travel to vast distances from
the sun, where the solar radiation flux is very low indeed. Thus nuclear power is essential; in most
recent studies, this is initially assumed to be derived from a fission reactor. It is therefore necessary
60
to consider whether electric thrusters can consume power of the order of 100 kW to several MWs,
since this level is necessary to achieve many important scientific objectives.
A cursory examination of the electric propulsion (EP) devices at present available reveals
that only the gridded ion thruster offers a way of increasing ve by orders of magnitude, while
providing thrust levels of the order of Newtons at this power level. Such values of these parameters
would be necessary for relatively economical missions beyond the regions of the solar system that
have already been explored. Lower values of SI, coupled with larger thrusts, would also be
available and would be suitable for manned missions to Mars, amongst other objectives. It should
be noted, however, that the HET is a strong competitor in the latter cases, since high power
consumption and correspondingly large thrusts have been demonstrated at values of SI up to about
3000 s. Power levels reaching 50 to 100 kW have been achieved in both Russia [6] and the USA
[Manzella, 2002], with NASA’s 457M thruster attaining a thrust of almost 4 N at 100 kW.
It should also be pointed out that many mission analyses have assumed the use of
magnetoplasmadynamic (MPD) thrusters, since they offer exceptionally high thrust density,
coupled with moderately large values of SI. However, the enormous current densities required to
achieve good performance lead to various life-limiting erosion processes. Solutions to the resulting
problems are being sought, but there is no guarantee that they will be found, bearing in mind the
need for thruster lifetimes of the order of 15,000 to 30,000 hours.
Various extremely challenging missions are now deemed to be of increasing importance, and
include Pluto and its satellite Charon [Rodgers, 2001], the Kuiper belt (at 30 to 100 astronomical
units (AU) distance) [Henry, 1999], the heliopause [Kluever, 1997] (at about 100 AU), the
gravitational lens focus of the Sun [West, 1999] (at 550 AU), and ultimately the nearest stars.
Preliminary theoretical work [12,13] has revealed that a combination of a reduced atomic mass
propellant, an improved ion extraction grid system, and much increased extraction potentials, could
conceivably lead to values of power consumption, SI and thrust suitable for such advanced
missions. At the extreme, values of SI above 150,000 s might be achieved, permitting various
missions to be undertaken which are currently regarded as nearly impossible. It is also worth
mentioning in this context that experimental studies with simplified grid systems have verified that
values of SI of 30,000 s are entirely feasible [Wilbur, 2004].
This report reviews the capabilities and limitations of gridded ion thruster technology in the
context of high energy mission objectives, with emphasis on reaching MW power levels at much
increased values of SI. This analysis is based on the performances of existing thrusters, coupled
with simple theory, so does not represent a major departure from current practice. It is shown that
the required performance characteristics can be attained by the use of high ion extraction potentials,
coupled with a comparatively minor redesign of grid systems. If found necessary, low atomic mass
propellants can be used to provide any further increase in SI deemed to be needed.
Although relevant analyses have been reported previously [15,16], they have been entirely
theoretical in nature and do not appear to have considered the practical limitations likely to be
encountered. The predictions of the present report include some discussion of these limitations,
which are due primarily to the properties of the materials employed in constructing ion thrusters and
the internal erosion processes.
61
3.3. Background
3.3.1. Space Nuclear Programmes
It has long been realised that high energy deep space missions cannot be accomplished using
solar arrays to provide the necessary power, unless very long flights can be accepted, together with
only brief fly-bys of the associated target bodies and extremely large and costly launch vehicles.
Such options cannot be regarded as cost-effective in any way. As a consequence, various
programmes have been commenced in several countries to develop suitable reactors for space use
[Reese, 1983], but only those of the Soviet Union have reached operational status [Reese, 1983].
The most highly developed version of the latter is the Topaz-2 reactor [Polansky, 2003], which was
intended to be flown in conjunction with US and UK ion thrusters in the NEPSTP (Nuclear Electric
Propulsion Space Test Program) orbit-raising technology demonstration mission [Cameron, 1993],
until funding for this was withdrawn. Similarly, cancellation was the eventual fate of the US SP-
100 reactor programme [Truscello, 1992].
Assuming, however, that nuclear power sources in the 100 kW to multi-MW range will
eventually become available, it is clearly necessary to establish whether appropriate electric
thrusters can be provided in a commensurate timeframe. It is the purpose of this report to show that
this can be accomplished in the case of GIEs, provided that high values of SI are acceptable (5000 s
and above); the latter is, in fact, a mandatory requirement of very many challenging missions. In
circumstances where high SI is not suitable, HETs provide a fully adequate alternative.
Unfortunately, it is not recommended that MPD thrusters be selected under the present
circumstances, owing to their development status, which leaves lifetime as an issue which has not
been resolved.
It is also relevant to point out that devices utilising power at high potential provide an
additional advantage. This is that power transmission losses are much reduced, a fact which is
important in the case of nuclear sources, where the reactor and its associated conversion equipment
are located at a considerable distance from the main body of the spacecraft. This latter feature,
depicted below in Figure 19 for the proposed NEPSTP spacecraft, is necessary to reduce the
radiation flux reaching the payload from the reactor.
Figure 19 Schematic diagram of NEPSTP spacecraft.
This section of the report provides additional clarification of the benefits of utilising EP for
high energy missions, and also discusses the principles underlying the operation of the three
technologies mentioned to date; the GIE, HET and MPD thruster. While the bulk of the report
deals with the GIE, some further brief mention will be made of high power HETs. Further
discussion of MPD devices is provided elsewhere.
62
3.3.2. Advantages of Electric Propulsion
It has long been recognised that the achievement of higher SI results in a reduction of the
propellant mass required to accomplish a given mission. This has usually been one of the
objectives of research into new and improved propulsion systems, including those in the EP
category. It is well established that electric thrusters can provide values of SI an order of magnitude
greater than those of chemical engines, so they are clearly candidates for deep space missions
requiring large velocity increments. Indeed, if orbit-raising [3,22,23] is included in the mission plan
to achieve greater economy, EP becomes critical to success.
These conclusions follow directly from Newton’s Laws of Motion, since it can be shown, by
equating the instantaneous rate of change of momentum to the force applied to a spacecraft by its
propulsion system, that
Mo
ΔV = veff log e ( Mo / Mf ) = veff log e (1)
Mo − ΔM
where veff is the effective exhaust velocity, taking losses of propellant into account, Mo and
Mf are the masses of the spacecraft at the beginning and end of the manoeuvre, and ΔM = Mo - Mf is
the mass of propellant consumed.
This, the so-called rocket equation, illustrates the importance of achieving a high value of
veff, assuming that adequate power is available. For a given value of ΔV, a low exhaust velocity must
be compensated for by a large increase in ΔM, and therefore in the size and cost of the vehicle. It is
this dependence of ΔV on the logarithmic term in the above equation for constant veff which has
provided the incentive for EP development.
An example of the advantages of EP operating at high SI is given below in Figure 20, in
which different propulsion technologies are compared for a mission to Pluto. This analysis, by
Widman et al [Widman, 1993], assumed the availability of advanced nuclear thermal propulsion
(NTP) and nuclear electric propulsion (NEP) systems. The initial spacecraft mass was taken to be
20 tonnes and this was assumed to be launched into a circular low Earth orbit (LEO) at an altitude
of only 200 km. The final orbit around Pluto was circular, with an altitude of 100 km. A long coast
period was assumed during the journey, which was possibly imposed by doubts concerning
propulsion system durability.
Figure 20 Payload mass in Pluto orbit, as a function of mission duration for three propulsion systems [Widman,
1993].
63
The results, in which the NEP and NTP systems are compared against a cryogenic chemical
alternative, show very clearly the superiority of the EP option, due to the high SI achieved. This was
assumed to be 8000 s, using argon propellant, with a thruster total efficiency of 77%. In
comparison, the NTP system was assumed to operate at below 1000 s and the performance of the
chemical option was taken to be typical of current technology. It should be noted that these ion
thruster performance values are readily achievable in the laboratory at present, and that there are no
reasons to suppose that practical, flight qualified thrusters cannot be developed. It will be shown
later that the consumption of the 310 kW assumed in this study is entirely feasible, if high SI
operation is required. It was thus concluded that only the NEP system can provide a large payload,
and this is also possible in a relatively short time. Indeed, a very reasonable mission can be
achieved in under 10 years, whereas the other two options need at least double that time, and then
provide an extremely poor performance.
The latter point emphasises that, in such ventures, the time taken to acquire useful data must
be sufficiently short to maintain interest within the associated scientific community and funding
agencies. Therefore such missions must normally be productive within a decade at most, implying
a very high maximum velocity. This, in turn, demands a relatively high thrust, but the need to
minimise propellant consumption also necessitates a large value of SI. These two requirements
dictate the need for a very substantial power supply, which can be met at present only by nuclear
fission technology.
Bearing in mind the fact that spacecraft in this mission category which employ chemical
propulsion gain most, in not all, of their initial velocity from the launch vehicle, the proposal to use
EP to reduce total mass must involve a fundamental change in strategy. Rather than boost the
spacecraft to an initial very high velocity using the launch vehicle upper stages, only a modest
velocity beyond that needed to escape from the Earth’s gravitational field is necessary. The
remaining velocity increment can then be provided over a long period of time, usually many
thousands of hours, by operating the EP system. Further large reductions in launch vehicle size are
possible through using the EP system for spiral orbit-raising from LEO [Fearn, 1980] or from GTO,
as has been demonstrated by ESA’s SMART-1 spacecraft [De Cara, 2004]. Of course,
disadvantages also accompany such strategies; these include the extended mission duration and the
additional radiation damage within the Earth’s radiation belts. However, as has been demonstrated
by SMART-1 and other recent missions which have encountered the radiation belts, modern
electronics and the associated design techniques are fully capable of withstanding successfully
extremely high radiation doses.
3.3.3. Propulsion System Parameters
In all missions, it is recognised that the SI and thrust attainable are crucial to success, and
that they together determine the power required. Of course, these two parameters can be traded off
against each other in most situations, where the power is defined by other factors. This can be
shown by recalling that the thrust, T, of an EP device is given by
T = m& veff (2)
where is the rate of flow of propellant. It will be shown later that the power consumed, P,
is approximately that in the exhaust beam in the types of device considered here, so that
P = 0.5m& veff 2 (3)
64
As the SI, denoted by Isp, is defined as the ratio of thrust to the rate of consumption of
propellant in units of sea-level weight, using Equ 2
T veff
Isp = = (4)
m& go go
where go is the acceleration due to gravity at sea level. Combining Equs 3 and 4 gives
2P
T= (5)
Ispgo
It is clear from Equ 5 that, for a given power supply, a higher thrust can be achieved only at
the expense of a reduced SI. Such a compromise is usually necessary in devising the optimum
strategy for a particular mission, hence the important need for thrusters able to operate over wide
ranges of these parameters. Thus it might be necessary to reduce the SI in order to achieve a shorter
mission duration through the provision of a greater thrust, although the mass of propellant
consumed will then be larger. It can be concluded that in all cases a detailed examination of the
overall mission is needed, taking into account the major factors of time, likely maximum power,
and launch capability, so that an optimum SI can be predicted.
As already described, the achievement of the higher values of SI at constant power must be
accompanied by a reduction of thrust, in accordance with Equ 5. There is, however, another
parameter which can assist in achieving high values of ΔV, and this is the thrusting duration.
Clearly, with a nuclear power source, this can be very long, reaching many years, although thruster
lifetime will then become a severe limitation. As in previous studies [8,26], this problem can be
resolved in principle by incorporating additional thrusters into the spacecraft, which will be used
sequentially. This is not a particularly heavy option, since they are of relatively low mass. The
power conditioning units (PCUs), which are much heavier, can be switched between these thrusters.
As an example of what can be achieved by use of an EP system with high SI, Figure 21
shows the initial mass, Mo, in LEO for a Pluto rendezvous mission as a function of ve, for a range of
masses delivered to the planet, which are denoted by the parameter Ms. The equivalent ion beam
accelerating voltage, VB, is also shown for xenon propellant. This graph illustrates clearly the
benefits of increased SI, and should be compared toFigure 20.
Figure 21 Mass in LEO required to place a specified payload, Ms, into Pluto orbit as a function of ve.
Despite the inevitable compromise between thrust and SI, the overall mission requirements
dictate that the SI must increase rapidly as the objectives become more demanding. This is
65
ilustrated in Table 1 for a variety of increasingly challenging objectives, in which minimum
acceptable values are shown, together with the associated velocity increments. Although these
values are somewhat arbitrary, and depend on a number of assumptions, they confirm this important
and challenging trend.
It should be noted that those values given in Table VI which are appropriate to the inner and
outer planets can be achieved with current technology, although a considerable increase in SI would
be of benefit in both cases. For example, the BepiColombo mission [Novara, 2001] to Mercury is at
present based on gridded thrusters operating at about 4500 s. This need for higher values is greater
if the launch is to LEO or to GTO, with an electrically-propelled orbit-raising manoeuvre then
required to achieve escape velocity. Similarly, the values of ∆V and SI required increase again if it
is necessary to rendezvous with and orbit the target body.
Mission Typical Minimum

ΔV (km/s) SI (s)
Nearby planets 5-7 3000
Outer planets 10-15 3000-5000
100-1000 AU 100 10,000
10,000 AU 1000 100,000
Interstellar 30,000 3 × 106
Table VI Values of velocity increment and SI for a variety of missions employing EP
3.3.4. Propulsion Technology Review

The three propulsion technologies which could, in theory, meet the power, thrust and SI
requirements of challenging NEP missions are summarised below. As mentioned previously, the
GIE is the main topic of this report, with some limited additional information also included below
concerning the HET. The MPD thruster is dealt with in more detail elsewhere.
3.3.4.1. Gridded Ion Engines

A schematic diagram of a gridded ion engine is shown in Figure 22. This consists of a
cylindrical discharge chamber which is closed at one end and has a set of perforated and accurately
aligned grids at the other. The propellant, in gaseous form, is ionised by an electric discharge in this
chamber, forming a moderately dense plasma. The positive ions in the plasma diffuse towards the
grids and are extracted and accelerated by the potentials applied to them. The space charge of the
emerging ion beam is neutralised by electrons provided by an external cathode, termed the
neutraliser.
66
Figure 22 Schematic diagram of a gridded ion engine.
This discharge can be direct current (DC), between an axial hollow cathode located in the
centre of the closed end of the chamber, and a concentric anode. The ionisation efficiency is
enhanced by the application of a magnetic field to the discharge chamber. A relatively weak
divergent axial field is used in the case of the Kaufman-type thruster [27-29], whereas much
stronger more localised fields are used in the alternative cusp-field geometry device [Beattie, 1993],
as used on the Boeing/Hughes HS-601 and HS-702 communications satellite platforms [Ocampo,
1998]. In another variant, a radiofrequency (RF) plasma is formed within an insulating discharge
chamber, using an external coil energised by an oscillator [32,33]. Alternatively, a microwave
source of RF energy can be employed [4,34], usually in conjunction with an auxiliary magnetic
field.
In all cases, accurately controlled flows of propellant gas are mandatory; in most present ion
propulsion systems this gas is xenon, although any other chemically inert gas is also acceptable.
These flows are produced by an external control system which, in turn, is fed at constant pressure
by a regulator connected to the storage tank via appropriate valves. Separate flows are needed to the
discharge chamber, neutraliser and, if one is used, to the cathode.
It is normal to use a triple-grid system on thrusters of this type. This operates in an
accelerate/decelerate mode to enhance the throughput of ions at the voltages usually employed. In
this scheme, the thruster body is at the high potential appropriate to the ion beam velocity required.
For example, if xenon ions are to be accelerated to a velocity of 40 km/s, which is appropriate for a
value of SI of about 3500 s, a potential of 1.1 kV is needed. This is thus the potential of the inner or
screen grid and of the thruster body. The next grid, the accel, is at a negative potential of perhaps -
250 V, to provide focusing of the beamlets and to enable the required current to be extracted.
Deceleration to space potential then follows, via a low negative voltage applied to the outer decel
grid and the coupling of the beam to the space plasma.
The neutralisation process involves the production of a weak external plasma by means of a
DC discharge between the neutraliser cathode and an adjacent keeper electrode; this is also closely
coupled to the space plasma. Electrons are extracted from this plasma by the ion beam as necessary
to neutralise its space charge. This is a natural automatic process involving no active control.
It should be emphasised here that the spacecraft potential is determined ultimately by its
interaction with the space plasma and that the difference in potential would normally be very small,
assuming that ion beam neutralisation is fully effective. Despite the enormous differences in plasma
conditions which can exist between the external environment and the plasma produced by the
propulsion system, the very great mobility of the electrons and the rapidity with which energy
equalisation between different electron populations occurs ensure that potential differences will
remain small.
67
Of course, the characteristics of the space plasma vary enormously over very many orders of
magnitude, according to the location of the spacecraft. If it is in low Earth orbit, where the residual
gas is ionised by solar radiation, this will constitute a moderately dense plasma, albeit with quite a
low electron temperature. Moving to geostationary orbit, the density will be far less, but there will
be a considerable contribution from the energetic solar wind, causing all temperatures to be much
greater than at low altitude; these are the electron, ion and neutral gas temperatures, which are not
necessarily identical owing to the collision particle processes which occur. Then in interplanetary
space the influence of the Earth’s atmosphere is absent and the plasma characteristics are dominated
by the solar wind, which decreases in density according to the inverse square law as distance from
the sun increases.
As an example of the variability of these plasmas, the solar wind in the vicinity of the Earth
has a velocity in the range 300 to 800 km/s [Baker, 1999], which can exceed 1000 km/s when solar
flares occur. Overall, these natural plasmas have densities which vary from less than 0.01 to 1000
particles per cm3, temperatures from less than 1 eV to more than 1000 eV, and particle energies
from less than 1 eV to many MeV [Barne, 1986]. To add to the complexity of analysing any
particular spacecraft utilising ion thrusters in interplanetary space, the plasma produced by the
thrusters close to the spacecraft is likely to be at least 3 orders of magnitude more dense than the
natural plasma. However, as implied above, interactions are generally benign.
To return to the description of the thruster, a more important function of the outer grid is to
minimise the erosion damage caused to the accel grid due to bombardment by low energy charge-
exchange ions [Fearn, 1993]. These ions are generated by interactions between beam ions and the
neutral gas atoms which escape from the discharge chamber and which also originate from the
neutraliser. The low energy ions are attracted to the accel grid by its negative potential, so shielding
by the decel grid diverts many of them, reducing the resulting damage. Of course, if this erosion
problem is not of concern for a particular mission, or is circumvented by other means, the decel grid
can be omitted, reducing complexity and also cost.
Several different thrusters in this category have flown on a variety of spacecraft [2,4,31,36-
38] and they and others are appropriately qualified for operational use. Including experimental
devices, grid diameters vary from 5 cm up to 65 cm, with thrusts from less than 1 mN up to nearly 1
N. Power consumption ranges from a few tens of Watts up to more than 10 kW. Typical electrical
efficiencies are 70 to 90%, and the SI is usually in the range 2500 to 5000 s, although both higher
and lower values are possible.
It should finally be mentioned that a major advantage of the GIE is that the plasma
production mechanism, the ion extraction and acceleration, and the ion beam neutralisation are
separate processes. They can be treated as such when designing a new thruster, greatly simplifying
what might otherwise be a complex procedure. This separation of these functions also applies to
thruster operation. For example, it is normal to change the ion accelerating potential, and thus the
SI, without altering the discharge characteristics in any way; this is not possible with most other EP
devices. As will be shown later, a further separation of functions is possible if a 4-grid
configuration is adopted for ultra-high SI operation. In that case, the ion extraction and acceleration
processes can be dealt with independently [12,13].
68
3.3.4.2. The Hall-Effect Thruster
The Hall-effect thruster (HET) [Kim, 1998], also known as the stationary plasma thruster
(SPT), is an ingenious device, originating in Russia in the 1950s and 1960s, in which the ions are
accelerated entirely within the discharge plasma. This type of engine has flown on numerous
Russian and Soviet spacecraft with considerable success [Gorshkov, 1998, Bober, 1993], and has
now been adopted by many European and American organisations [Cadiou, 2003, Jacobson, 2004],
primarily for application to communications satellites [Garnero, 2003]. This technology’s first
Western success has been to transfer the SMART-1 spacecraft from its initial GTO to lunar orbit
[De Cara, 2004], utilising Snecma’s PPS-1350G thruster [Dumazert, 2003].
An HET is shown schematically in Figure 23a, with the high power 457M thruster in Figure
23b. The typical HET consists of an annular discharge chamber made from ceramic material, with a
ring-shaped hollow anode situated at its closed end. The propellant gas, again usually xenon, is
injected through this anode, and the correct application of a magnetic field is critical to success.
The magnetic circuit provides an almost radial field across the annular discharge chamber, but with
very careful shaping along the axial direction to give high efficiency. The field is produced by 4 to
8 external solenoids bolted between a front circular polepiece and a ferromagnetic backplate. An
axial polepiece is also bolted to the backplate; in some thrusters, this is surrounded by another coil.
(a) (b)
Figure 23 Hall-effect thrusters: (a) Schematic diagram (b;) NASA 457M high power thruster (NASA photo [7])].
The discharge current is provided by an external hollow cathode, which is often duplicated
for redundancy. The electrons emitted by the cathode are attracted into the annular discharge
chamber by the anode, which is at a potential of typically 200 to 350 V, although values of up to 1
kV have been tried experimentally [Gorshkov, 2002]. However, they cannot penetrate far into the
annulus due to the radial magnetic field, and spiral around the field lines, colliding with propellant
gas atoms and causing ionisation to occur. The resulting axial current interacts with the magnetic
field, generating an azimuthal Hall current, hence the name of the device.
The anode potential appears mainly across the relatively thin plasma layer in which
ionisation occurs and the Hall current flows azimuthally. The ions are accelerated by the electric
field in this region, attaining velocities of the order of 16 to 18 km/s, with an anode potential of 300
V. As they have mean free paths and cyclotron radii much larger than the channel dimensions, this
acceleration is mainly in the axial direction, but the beam divergence is relatively large, at 40 to 50
69
deg. Non-axial ions impact the edges of the downstream end of the discharge chamber, causing
erosion and limiting life. Neutralisation of the ion beam is accomplished automatically, electrons
being provided by the external cathode as required.
Although very effective, the HET operates in a relatively noisy mode, with violent
oscillations of the discharge current often observed [Randolph, 1994]. This feature generates a need
for careful isolation from the spacecraft power system. Lifetime is limited to about 6000 to 8000
hours by the erosion mechanism mentioned above, but this is adequate for many missions of
interest. Throttling over a modest range is possible, but the large ion beam divergence can cause
problems with thruster-spacecraft integration.
HETs with nominal diameters of 5, 7 and 10 cm and thrusts of up to more than 80 mN have
been flown on more than 60 occasions in Soviet and Russian missions [Gorshkov, 1998, Bober,
1993]. No failures have been reported, but operational times have generally been of the order of
hundreds of hours. Much larger devices are under development in both Russia [6, Bober, 1993] and
the USA [Manzella, 2002], with thrusts that reach above 1 N. For example, the experimental NASA
457M shown in Figure 23b has operated at up to 72 kW input power.
As with most propulsion devices, efficiency and SI increase with size and power. However,
the SI available from thrusters with qualified status, or near to this, is limited to about 1600 to 1800
s. Numerous mission analyses suggest that this is inadequate for the majority of deep space
missions. So, although this technology has been used successfully by ESA on the SMART-1
demonstrator mission [3,22], it would not appear at present to be a serious contender for most future
projects requiring a large ΔV, unless success is achieved in developing the larger thrust devices with
high SI to qualified status.
It should also be mentioned that another approach to achieving high SI is the dual-stage
device [Jolivet, 2003, Prioul, 2004], in which the plasma production and acceleration processes are
separated within the same thruster. This work has preliminary experimental status at present, and it
is not possible to predict whether it will be successful. In any case, it would seem to be very
unlikely that values of SI much above about 3500 to 4000 s will be achieved by this means.
Another development, which has already achieved success, is to move the acceleration
region within the thruster along the axis towards the exit plane, so that the high energy ions are less
likely to impact on the ceramic wall [Sankovic, 1994, Semenkin, 1999]. This increases lifetime, but
the associated drawback is a greater ion beam divergence. In this form, the thruster has the generic
name of thruster with anode layer (TAL).
As regards the suitability of HET/TAL technology for high thrust, high SI NEP missions,
many recent studies have confirmed that values of SI well above 5000 s are necessary. Indeed, as
will be shown later, the more challenging missions are likely to require values in the range 10,000
to 14,000 s. There is no prospect of HET/TAL devices reaching this regime using xenon propellant,
since the most advanced experimental thrusters currently operate at the 3000 to 4000 s level. As an
example, the Keldysh Research Center’s X-85M thruster [Gorshkov, 2002] is designed to operate at
up to 1 kV. At 750 V its SI is reported to be 3200 s, which is itself a considerable achievement.
The best that can be expected at 1 kV is about 3800 s, and lifetime remains a serious issue at such
high anode potentials.
These and other results have been obtained using the standard propellant, xenon. As with
most types of EP technology, the substitution of a propellant with a lower atomic mass will
automatically increase the recorded SI at a given accelerating potential, via a square root
relationship discussed later. It is thus no surprise that an SI of 4700 s has been obtained at NASA
Glenn using krypton in the new 400M thruster at 1.05 kV anode potential [Jacobson, 2004]; the
power input was 43 kW and the total efficiency was 54%. Earlier, the 457M operated at an SI of
4500 s, with 1 kV applied and a discharge (not total) efficiency of 64%. However, the 8000 s
reported earlier by TsNIIMASH, with a dual-stage device and using bismuth is much harder to
understand and requires verification [Manzella, 2002]; 800 s would be more likely with this high
atomic mass propellant.
70
High thrust, coupled with high power consumption, is a different issue, and presents no
problems, provided that adequate lifetime can be assured. For example, several years ago the Fakel
SPT-290 [6] was operating successfully at 1.5 N, with a power consumption of 30 kW and an SI of
3000 s. The NASA 457M thruster (Figure 23b) later reached 50 kW, then 72 kW [Manzella, 2002],
with an SI of up to 3250 s. A thrust of up to 2.9 N was measured. More recently, NASA’s high
thrust HET programme has achieved a power consumption of 64 kW with the 400M thruster, using
krypton propellant and an accelerating potential of 1.1 kV [Jacobson, 2004]. However, durability
testing is still being discussed in terms of 200 hours of operation, which is a factor of at least 30
away from times of interest for real missions.
A further problem with high power operation concerns thruster temperatures. Unlike the
GIE, in which the bulk of the power is deposited directly into the ion beam, in the HET it is all fed
into the discharge chamber, which naturally becomes very hot indeed. With power levels of the
order of 50 kW, the 457 M thruster takes more than 4 hours to reach thermal equilibrium, at which
time many of the temperatures are well above 300°C, with the actual discharge chamber ceramics in
the region of 600°C to 800°C [Jacobson, 2004]. This adds to the problem of ensuring adequate
durability.
It can be concluded from the above discussion that the HET offers little promise of being
able to provide the values of SI required for the most challenging interplanetary missions using
nuclear power. While efforts to increase the SI are yielding some success, the values achieved do
not approach those of interest. However, the power levels and thrusts currently being attained are
relevant to certain missions in which SI and propellant mass have to be sacrificed in order to
accomplish manoeuvres within specified periods of time. Thus large thrusters, as exemplified by
the 457M, could be utilised in such circumstances where a reactor might be producing a few
hundred kW. However, they would still probably not be suitable in the multi-MW regime, since too
many would have to be mounted on the spacecraft to utilise this power. As will be shown later, this
restriction does not apply to the advanced GIE operating at high SI.
3.3.4.3. Magnetoplasmadynamic (MPD) Thrusters

In the longer term, much higher thrusts and thrust densities will be achieved by using MPD
thrusters. These are currently pulsed devices, in which the discharge current rises to kA levels and
well beyond, perhaps to many 100s of kA, so that the electromagnetic forces expelling the plasma
from the discharge vessel dominate all other effects [Andrenucci, 2001]. Within the same power
category are high power arcjets, in which the exhaust is further heated by an RF discharge; the
ATTILLA device is a good example of this technology [Böhrk, 2004]. However, these concepts
have at present an experimental or developmental status, and practical devices with long life are
unlikely to appear for at least a decade [Fearn, 1998]. A further problem to be overcome if viable
MPD applications are to materialise is that of the power supply. Present versions utilise capacitors,
which are massive and bulky, and a great deal of effort will be needed to reduce their mass and
volume to acceptable levels, unless it is found practicable to operate continuously.
A typical MPD thruster, shown schematically in Figure 24, consists of a tungsten cathode at
one end of a cylindrical or conical discharge chamber made of a refractory insulating material,
usually a ceramic. A ring-shaped anode is situated at the other end of this vessel. A very low
inductance source of high voltage, such as a capacitor, is applied to these electrodes as a pulsed
valve admits a puff of propellant gas into the cathode end of the chamber. With low inductance, the
discharge current rises very rapidly and the instantaneous power dissipation is often in the 100s of
kW to multi-MW regime. In order to match more precisely the current and gas flow profiles, the
current source is frequently a series of capacitors and inductances, forming a transmission line with
a time constant of the order of a few ms.
71
Figure 24 Schematic diagram of a conical MPD thruster.
Although Paschen breakdown can initiate the discharge, an auxiliary process may be used,
such as a separate high voltage spark or a laser pulse. Once the current flows, it pinches onto the
axis of the anode. The interaction between the azimuthal magnetic field, Bθ, around the pinched
axial current and the radial current, jr, to the anode produces a very large expulsive force jr ∧ Bθ,
which can provide instantaneous thrusts of the order 1 to 100 N. This is attractive, as is the
associated high thrust density, but many difficult problems must be overcome before such devices
can be considered for practical applications. In the meantime, the most promising option for low-
cost interplanetary exploration is the gridded ion thruster.
3.3.4.4. Variable Specific Impulse Magnetoplasma Rocket (VASIMR)
This device, devised and pioneered by astronaut Franklin Chang-Diaz [Chang Diaz, 1997],
has been under development since 1994, with NASA funding, but is very complex, so it is still at
the laboratory investigation stage. However, it promises to provide eventually a very high thrust
density and SI, with an electrical efficiency of 50%. The SI has been predicted to reach 30,000 s
[Batishchev, 2003] with propellant gases having low atomic masses, and will be readily variable
over a wide range.
The basic concept is similar to the magnetic “bottle” plasma confinement concepts which
were widely investigated over several decades in the context of thermonuclear fusion programmes
[Glassatone, 1960]. In the VASIMR device, three magnetic confinement regions are created. In the
first, a helicon discharge is used to generate the source plasma, the ions then drift into a volume in
which they are more tightly confined by a much larger magnetic field. There, they are heated by the
ion cyclotron resonance mechanism, configured as a downstream propagating wave, until they
emerge into a diverging magnetic nozzle. The variable SI feature is achieved, at constant input
power, by changing both the propellant flow rate and the division of power between the helicon
discharge and the subsequent ion heating. A high flow rate, with most of the power supplied to the
helicon discharge, results in a relatively low SI. As flow is progressively reduced and more power is
diverted to ion heating, the SI increases.
The most recent laboratory version [Glover, 2004], designated VX-10, is based on a 9 cm
diameter helicon discharge tube, with an axial magnetic field of about 0.05 T and an input power of
up to 3.5 kW. The plasma, of density of about 1011 cm-3, is then heated by the ion cyclotron wave
in a carefully shaped magnetic containment field of up to about 0.3 T. The heating power in this
version is up to 3 kW and the complete device is about 1.5 m long. The SI achieved, using
deuterium propellant and 1.5 kW heating power, was between 5000 and 12,000 s. With helium, a
72
target-type thrust balance indicated a thrust of 1.4 mN using a heating power of 3 kW. The overall
efficiency has not been given, but a simple calculation indicates a value of 1.8% if the SI reached
12,000 s with helium.
Although it has been predicted that this device can in future be operated at the MW level,
and an upgrade to 50 kW is underway [Glover, 2004], a comparison with Table 2 of this report
suggests that it is not currently a viable competitor for relatively near-term missions. It should also
be noted that the high value of SI achieved is a consequence of the use of a propellant with a low
atomic mass, and that it is also easy to change the SI of a gridded thruster over a very wide range,
merely by altering the potentials applied to the grids.
3.4. Gridded Ion Engines; Current Devices
A short summary of the operating principles of GIEs is provided in section 2.4.1 of this
report. However, as these devices are the most likely to be employed for high energy deep space
missions powered by nuclear reactors, and least in the near to medium term, further information
concerning them is given below. As the different categories of thruster can be identified by the
various ionization mechanisms utilized, these are covered qualitatively in more detail. In addition,
brief descriptions are given of the higher power devices currently available; these are either space
qualified or have shown promise in development or in laboratory tests.
3.4.1. Ionization Mechanisms

As mentioned previously, the major feature which differentiates between the several
varieties of GIE now available is the ionization mechanism employed. Four main mechanisms are
to be found; these are RF ionization at a frequency of about 1 MHz, electron cyclotron resonance
(ECR) RF ionization, with frequencies in the range 0.1 to 5 GHz, and DC discharge ionisation in a
weak or a strong magnetic field. In all cases, the resulting ions are extracted and accelerated by a
system of carefully aligned multi-aperture grids, and the space-charge of the resulting ion beam is
neutralised by electrons from a hollow cathode discharge plasma.
Since there is nothing to differentiate in a generic way between the ion beams generated by
these thrusters, as far as mission analysis and thruster-spacecraft interactions are concerned there is
little to choose between them. Nevertheless, other issues, including lifetime, reliability, efficiency
and cost, may dictate that one type be preferred to another. For this reason, more detailed
explanations of their modes of operation are given below.
3.4.1.1. Radiofrequency Discharge Thrusters

The RF thruster has been developed in Germany over many decades [Loeb,1970, Groh,
1989, Leiter, 2000], and has reached an advanced status in its 10 cm beam diameter form, which is
designated RIT-10 [Bassner, 1991]. In this context, the acronym RIT is the radiofrequency
ionisation thruster. It was flown successfully on ESA’s Eureca spacecraft [Bassner, 1994], although
the test was terminated prematurely by a wiring fault. More recently, it has been employed very
successfully for emergency orbit-raising of ESA’s Artemis communications satellite [Notarantonio,
2003], following the partial failure of the Ariane 5 launch vehicle, although its primary task on this
spacecraft is north-south station-keeping (NSSK). Further developments of this device [Leiter,
2002], using different grid materials and designs, show significant performance improvements, but
they are not yet qualified. This ionisation principle has also been used in several other thrusters of
larger size [32,33,56], extending to 35 cm beam diameter, and all could become available for
applications within the medium-term, provided that adequate funding was provided.
73
In this type of thruster, shown schematically in Figure 25, ionisation of the propellant gas,
which is usually xenon, is accomplished by winding an RF coil on the outside of the discharge
chamber, which is made from quartz or alumina. The coil, when fed from an RF generator operating
at about 1 MHz, produces a rapidly varying magnetic field within the discharge chamber. This, in
turn, induces an electric field which is adequate to cause ionisation to occur, resulting in the
formation of a dense plasma. The initial electrons required to start this process diffuse in through
the grid system from the neutraliser plasma, which is initiated as the first stage of the start-up
process. Thrust levels are readily adjusted by altering the plasma density in the discharge by
changing the RF power level and the propellant flow rate. The SI is typically about 3300 s, but
higher values are immediately available by increasing the beam acceleration potential
Figure 25 Schematic diagram of an RF gridded ion engine
It is notable that the power supplies required are very similar to those utilised by the DC
discharge thrusters, as will be confirmed later. The supplies used to extract and accelerate the ion
beam are identical in function and configuration in the two cases, as are those employed to provide
the neutralising electrons. The only difference, as implied by the names of the thrusters, is in the
production of the plasma in the discharge chamber. In the RF thruster the energy source consists of
an RF current with a frequency of about 1 MHz, whereas in the DC device a current regulated DC
discharge is used.
A component not mentioned previously is also indicated in Figure 25. This is the electrical
isolator situated between the propellant flow control unit (FCU) and the discharge chamber. This is
necessary because the feed system is at spacecraft potential, whereas the discharge chamber is
connected to the positive terminal of the beam supply, which is typically at 1 to 1.5 kV. These
isolators are very simple yet effective devices, requiring no power and no active control. They
operate by requiring the propellant gas to pass through a porous medium with a very high surface
area. This causes any ion-electron pairs formed in the high applied electric field to recombine
immediately, thereby inhibiting breakdown
Of course, there is no absolute guarantee that RF thrusters, or any other type of device, will
operate without failure. Indeed, two failures have been recorded to date, both owing to problems
not associated in any way with the fundamental principles of this technology.
The first occurred in the Eureca mission [Bassner, 1994], in which a RIT-10 thruster was
flown on the European Retrievable Carrier, deployed into orbit from the Space Shuttle. After more
74
than 200 hours of operation, the thruster ceased working, apparently due to a failure in the RF
circuit. After retrieval of the spacecraft by the Shuttle, laboratory examination showed that a
soldered joint had failed; this was clearly a problem associated with design, manufacture or quality
assurance, not with the thruster itself. Indeed, the thruster was found to be in perfect condition.
The second failure occurred on the Artemis mission [Notarantonio, 2003], when the
neutraliser of one of the two RIT-10 thrusters failed to operate, although the thruster itself was
subsequently operated successfully using the neutraliser on one of the T5/UK-10 thrusters mounted
on the same spacecraft. The probable reason for the failure was found to be damage caused to the
neutraliser cathode by vibration during launch. Again, this was attributable to detailed design,
manufacture or quality assurance, and the thruster itself was not at fault.
3.4.1.2. Radiofrequency Discharge Thrusters Operating at High Frequencies.

An alternative to the RF ionisation process used in the RIT series of thrusters is to raise the
frequency considerably and also apply a steady magnetic field to the discharge chamber. This
concept relies upon the gyration of the electrons within the discharge chamber around the magnetic
field lines, thus improving their chances of suffering an ionising collision with a neutral gas atom.
If the cyclotron and RF frequencies are equal, a resonance condition has been created which aids in
the effective absorption of the input energy, further enhancing overall efficiency. The latter is the
electron cyclotron resonance (ECR) discharge, as used by the Japanese team responsible for the
thrusters employed on the Muses-C asteroid sample return spacecraft [Kawaguchi, 2004,Toki,
2003, Kawaguchi, 1999].
It is notable that this thruster also employs the ECR principle within the neutraliser cathode.
However, although this eliminates the need for a source of low work function material [Fearn,
1993], it does not remove the major life-limiting factor of sputtering by relatively high energy ions.
A small Italian thruster, under development by Laben/Proel [Capacci, 2003], utilises this
magnetic enhancement principle. Termed the “radiofrequency with magnetic field thruster (RMT)”,
it is similar in concept to the RIT series, except that the plasma production process employs RF
power in the very high frequency (VHF) range, together with an axially symmetric magnetic field
of about 100 G. This device, shown schematically in Figure 26, relies upon the increased ionisation
rate caused by the cyclotron motion of the electrons, but without resonance. Using a variable
magnetic field component provided by a solenoid to optimise the discharge, it enables low thrusts to
be achieved while maintaining good efficiency.
Figure 26 Schematic diagram of the high frequency RF gridded ion engine (Laben diagram).
75
In Figure 26, both permanent magnets and the solenoid are shown; Ic is the current supplied
to the latter, for which a power supply is needed. The ion beam current is denoted by Ib, and the
currents to the accel and decel grids by Ia and Id respectively. The beam accelerating potential,
which is applied to the screen grid, is Vs, and the accel grid potential Va. The current, Is, provided
by the screen power supply is almost identical to Ib. The RF input power is denoted by WRF and the
propellant flow rate to the discharge chamber by m& . The neutraliser cathode is not shown in this
figure; it is very similar to those depicted in Figure 22 and Figure 25 and requires the same power
supplies.
3.4.1.3. Kaufman-Type Direct Current Discharge Thrusters

Thrusters of this general type are under development in the UK and Russia, and discontinued
work in this field was undertaken in Germany. All devices now use xenon propellant, although any
other gas which does not interact chemically with thruster components can also be employed. The
UK and Russian types are similar, in that they employ the Kaufman ionisation principle [Kaufman,
1961]. In this (Figure 27), the propellant gas is ionised in a DC discharge between an axial hollow
cathode [Fearn, 1993] and a downstream cylindrical anode. A divergent magnetic field, generated
by solenoids external to the discharge chamber, is provided to prevent the primary electrons from
reaching the anode directly. They are obliged to spiral around the field lines, thereby enhancing
their chances of suffering ionising collisions with neutral gas atoms. Since the hollow cathode
requires a propellant flow for its operation [Fearn, 1993], two feeds are usually employed to the
discharge chamber. However, small thrusters can often be operated successfully with only a single
flow, which is passed through the cathode.
Figure 27 Schematic diagram of Kaufman-type gridded ion engine.
High propellant utilisation efficiency is assured by causing the primary electrons from the
cathode to acquire the correct energy to maximise the probability of ionisation within the discharge
chamber. These electrons emerge from the cathode orifice [Fearn, 1993] into the coupling plasma,
where they have a relatively low energy/temperature of 1 to 2 eV. Since they are attracted by the
potential of the anode, which is typically at 30 to 40 V above cathode potential, they travel through
the annular gap between the inner magnetic polepiece and the circular baffle disc to enter the main
discharge chamber. However, the fringing magnetic field across this gap causes an impedance to
their flow, resulting in an azimuthal Hall current and a well-defined energy increase, which can be
76
adjusted at will by altering the magnetic field. Thus the operating parameters of the thruster can be
set as required to optimise performance at any required thrust.
It should be noted that there is no thruster parameter which intrinsically provides a measure
of propellant flow rate, although indications can be derived from the voltages at constant current of
both the main discharge and neutraliser cathode keepers, and by the anode potential for constant
cathode conditions, discharge current and magnetic field. However, so many parameters interact in
determining the performance of a thruster that it is far from easy to deduce unambiguously what any
of the three flow rates might be under a given set of conditions. Thus, in diagnosing faults and in
assessing absolute performance the use of flowmeters [Edwards, 1999] is highly recommended. As
they are small, of low mass and consume little power, their insertion into a propellant fed system
carries few significant overheads. In addition, should a flowmeter fail, it will not jeopardise the
mission, since all thrusters up to the present have been flown without them.
The power conditioning system required by a Kaufman-type thruster is as indicated in Figure
22, with the addition of the current regulated discharge supply and the heater and keeper supplies
needed for the hollow cathode used to provide the ionising electrons. Although the positive
terminal of the beam supply is often connected to the anode, an alternative arrangement is
sometimes employed, by making this connection to the thruster body. In that case the beam current
passes through the anode supply, but the ion accelerating potential is increased by the plasma
potential in the discharge chamber.
All current gridded thrusters of this type have roughly comparable performances, when
measured in terms of SI and the various thruster efficiencies. The typical SI has, until recently,
been in the region of 3000 s to 4500 s, but greater values are now being considered for very high
energy interplanetary missions [Fearn, 2000]. The available thrust covers the very wide range of a
few mN to more than 300 mN. In addition to the thruster design variations implied by this range,
some differences due to individual performance features can be significant. For example, flexibility
in operation is enhanced by employing a separately controllable propellant feed to the hollow
cathode, but this might not be necessary if no throttling was required.
3.4.1.4. Magneto-electrostatic Containment (MESC) DC Thrusters

The MESC, or cusp-field thruster [Beattie, 1991, Beattie, 1991, Beattie, 1976], is very
similar in most respects to the Kaufman device, apart from using a very high value of magnetic field
for plasma containment, and relying upon the difference between the discharge plasma and cathode
potentials to accelerate the primary electrons emitted by the cathode. The diagram shown in Figure
28is based on the very successful NSTAR thruster [Christensen, 1999] flown on the Deep Space 1
(DS-1) interplanetary spacecraft [Raymann, 2000, Brophy, 2002]. It indicates the positions of the
permanent ring magnets, often made of samarium-cobalt, which provide the magnetic field. This
field is located closer to the discharge chamber walls than in the Kaufman-type thruster and is of the
order of kGauss near the magnets. A separate anode is not normally used, since the complete
discharge chamber wall usually performs this function. The hollow cathode is of the same type as
used in most Kaufman-type thrusters.
77
Figure 28 Schematic diagram of the MESC concept, based on the NSTAR thruster.
In operation, the primary electrons emitted by the cathode are accelerated from the vicinity
of the keeper electrode into the main plasma, gaining energy in the process; this energy is
determined approximately by the difference between the keeper and anode potentials. These
electrons are constrained to orbit around the magnetic field lines, as in the Kaufman thruster, only
diffusing to the anode via collisions with other particles. As the field is much stronger close to the
magnets in this device, the reflection of the electrons there is very effective, and this allows the
discharge power to be lower for a given thrust; typical values of the discharge power to beam
current ratio are 150 and 250 W/A, respectively, in the two types of device. As indicated in Figure
28, the close proximity of the most important region of the magnetic field to the discharge chamber
wall can allow the thruster to be designed with a conical shape, although in some cases a
hemispherical alternative has been selected. It will also be noted that outward (convex) dishing of
the grids was chosen in the case illustrated, and that an electrical isolator was fitted in the propellant
flow path to the neutralizer.
Although the electrical efficiency can be very high, the thruster cannot be controlled as
readily as the Kaufman-type using a variable magnetic field, so the active throttling range at
sustained high efficiency is somewhat lower. Nevertheless, extensive operational experience with
this type of thruster has been gained owing to its use in the Boeing/Hughes HS-601 and HS-702
communications satellites for north-south station-keeping [Anzel, 1998] and, in the latter case, for
the final phase of orbit-raising [Ocampo, 1998] and for east-west station-keeping.
Although these commercial applications have been very successful, with more than 100
thrusters in orbit and over 125,000 hours of operation achieved [Chein, 2005], it is understood that
failures have been encountered in the case of the XIPS-13 thrusters. Unfortunately, no information
concerning the severity of these rumoured failures or of their cause has been published, and
representatives of the company (now L3 Communications) have not been at liberty to provide any
information. Of course, it is possible to surmise what may have occurred, based on published
knowledge of the various sub-systems involved, but this would not be productive here, bearing in
mind that no evidence whatsoever is available.
78
3.4.2. Current Gridded Thrusters
The thrusters listed and described below are in the high power category, where “high” in this
context means the use of an input of 1 kW or greater. This, of course, is very much below the
values required for NEP applications, by orders of magnitude, but these devices provide the starting
point for future developments to much higher power levels.
The one exception to this rule is the T5 thruster [Fearn, 1998], which is the first to be
described. This is included because its design involved the formulation of scaling relationships
which have been successfully employed subsequently in the development of the much larger UK-25
[7 Latham, 19903] and T6 [Fallace, 1999] devices. These scaling relationships provide the basis on
which much of the analysis in this report is founded
a. The T5 Thruster
This UK thruster [Fearn, 1998, Gray, 1997], which was flown on the Artemis
communications spacecraft [Notarantonio, 2003,75,], is shown in Figure 29. It was developed
jointly [Fearn, 1993 - Fearn, 1991] by RAE/DRA/DERA/QinetiQ, the Culham Laboratory, and
Matra Marconi Space (MMS) UK Ltd, now Astrium UK/EADS. As with the RIT-10, this version
has operated in the orbit-raising and NSSK roles on Artemis, as part of the UK-10 ion propulsion
system (IPS). In addition, the same basic thruster was selected for the now cancelled Nuclear
Electric Propulsion Space Test Program (NEPSTP) mission [Cameron, 1993, Bythrow, 1993] by
Johns Hopkins University, and a specially developed version has also been chosen for the drag
compensation propulsion task on ESA’s GOCE spacecraft [Bassner, 2000].
Figure 29 Photograph of T5 Kaufman-type thruster (QinetiQ photo).
The T5 GIE has several special features which permit very precise throttling [Mundy, 1997]
over an extremely wide range, 0.3 to 30 mN, and give a low ion beam divergence of less than 15
deg under most conditions. These features are the use of solenoids (see Figure 27) to generate the
magnetic field applied to the discharge chamber, independently controlled propellant flows to the
cathode, the neutraliser and the discharge chamber, and inward (concave) dishing of the grid
system.
When operated at high SI, with values of VB reaching beyond 2 kV, the T5 thruster provides
much enhanced performance [Martin, 1988], with the thrust extending to 71 mN and the SI
approaching 4500 s. The input power under these conditions reaches nearly 2 kW, but temperatures
do not become excessive, because the bulk of this power is then fed into the ion beam. It should be
emphasised that no mechanical changes to the thruster are needed to permit this operating regime to
be explored. This represents a considerable achievement for a 10 cm beam diameter device. It is
this latter accomplishment, as well as the verified scaling relationships, which render the design
features of this device of interest for high power missions.
79
b. The RIT-15
Recent development of the RIT-15 commenced in the late 1980s [Groh, 1976] when it
became clear that the increasing mass and lifetime of communications satellites necessitated a
higher thrust level for NSSK than could be provided by the RIT-10. Scaled from the RIT-10 and an
early version initially investigated at the University of Giessen in the 1970s [Freisinger, 1976],
initial concepts [Groh, 1998] utilised a variety of flat grid configurations, but all providing a beam
diameter of between 13 and 14.2 cm; an engineering model is shown in Figure 30. The main
design objective was a thrust level in the region of 50 mN and an SI exceeding 4000 s.
Figure 30 Photograph of early version of the RIT-15 thruster (University of Giessen photo).
Later versions [Leiter, 1999] utilised carbon-carbon grid technology for greater lifetime, and
the shape of the discharge chamber was modified from cylindrical to hemispherical to reduce the
internal energy losses caused by recombination on the walls. The latter idea reduced the discharge
chamber power losses by up to 23%. Two variants were investigated. The RIT-15LP was aimed at
relatively low power missions, for which a moderate specific power was required; 25.5 to 26.5
W/mN was achieved at an SI of 2900 to 3600 and a thrust of up to 50 mN. The other, the RIT-15S,
was intended for high SI applications, with beam accelerating potentials in excess of 2 kV.
However, as the RIT-15LP has been more extensively documented, this version is included later in
Table 2.
c. SERT II
It is interesting to recall that ion propulsion was thought to be ready for actual scientific and
commercial applications in the mid-1960s, with the final confirmation of this status expected from
NASA’s SERT II mission, which was launched in February 1970. As anticipated, this mission was
the first to demonstrate conclusively in space that GIEs will work as designed for very long periods
of time, producing the calculated thrust, without causing thruster-spacecraft interaction problems.
This mission included two 15 cm beam diameter Kaufman-type thrusters [Kerslake, 1971]
developed by the NASA Lewis Research Centre in the 1960s. Using mercury propellant, these
thrusters each provided a nominal thrust of 29 mN with an SI of 4770 s. The latter has still to be
exceeded in space. The mission was not without difficulties, which were largely predicted before
launch, but managed to continue until all propellant was exhausted. This took 11.5 years from the
launch in Feb 1970 [Kerslake, 1981]. A photograph of the spacecraft showing the two thrusters is
reproduced below in Figure 31.
80
Figure 31 The two Kaufman-type thrusters mounted on the SERT II spacecraft (NASA photo).
It was known before launch that there was severe charge-exchange erosion of the grid
system of this type of thruster, and the solutions to this difficulty had been determined. However,
there was no time in which to implement them before launch, so a relatively short operational
lifetime was envisaged from the outset. One thruster did, indeed, fail from this problem early in the
flight, but the operation of its discharge chamber continued to be demonstrated, thereby showing
excellent durability of all components apart from the grid system. The other engine continued to
operate reasonably satisfactorily until all the mercury propellant was used, giving an aggregate
thruster running time of nearly 4000 h.
d. The RIT-XT/RIT-22
This 20.8 cm beam diameter RF gridded device [Leiter, 2003- Leiter, 2002] has a long
heritage, including the RIT-10 [Bassner, 1991], RIT-10 Evo [Killinger, 2000] and the RIT-15
[Leiter, 1999] thrusters. It is shown in diagrammatic form in Figure 32 and a photograph is
presented in Figure 33. It has now been extensively characterised at thrust levels of interest for the
BepiColombo mission [Novara, 2001], including an investigation of its throttling capabilities. It
was originally designed as a 100 mN-class device, with a published design thrust extending to
around 150 mN, but this has now exceeded 200 mN [Leiter et al., 2003]. Values of SI of up to 6420
s have been reported, with a beam accelerating potential of 3 kV and an input power of 8 kW.
It will be noted from Figure 32 that the hemispherical discharge chamber of the RIT-15 has
evolved to a conical design to further reduce energy losses due to electron-ion recombination. The
carbon-carbon grid design has benefited considerably from the experience gained from the RIT-10
Evo, RIT-15 and ESA-XX [Bassner, 1999] thrusters, and now provides a very high perveance, with
a potentially long lifetime. As it nears qualified status, it has been renamed the RIT-22, having
been previously known as the RIT-XT.
81
Figure 32 Diagrammatic representation of the RIT-XT thruster (Astrium GmbH diagram).
Figure 33 Photograph of the RIT-XT thruster (Astrium GmbH photo).
e. The T6 Kaufman-Type Thruster

This is a 22 cm beam diameter DC gridded thruster [Wallace, 1999, Huddlestone, 2004]
originally developed for NSSK on large geostationary communications satellites, but with the
inherent capability to provide high thrust and SI for interplanetary missions. An engineering model
is shown in Figure 34. It was designed using basic scaling data derived from the T5 [Harbour,
1973, Fearn, 1991] and UK-25 [Latham, 1990] programmes, with a nominal thrust of 150 mN for
NSSK and 200 mN for deep space missions, and an SI of between 3000 and 5000 s. A smooth,
wide throttling range is an intrinsic capability of this concept [Mundy, 1997], which also provides
the low beam divergence and very stable thrust vector direction found with the T5 design [Pollard,
1995, Mundy, 1997].
The programme has proceeded well beyond the characterisation of the thruster’s
performance, vibration resistance and beam parameters, and has entered the long duration test
phase; this is at 200 mN to meet the requirements of the BepiColombo mission to Mercury
[Novara, 2001]. The thruster uses either molybdenum or carbon grids, the latter being employed
when a need for considerably enhanced total impulse has to be met, as in the case of the
BepiColombo application. A further potential application is to the station-keeping of the new
European communications satellite platform, AlphaBus [Lsyzyk, 2003], for which a high SI is a
significant advantage
82
Figure 34 T6 engineering model ion thruster (QinetiQ photo).
f. The ESA-XX RF Thruster

This 25 cm beam diameter RF ionisation gridded thruster [Bassner, 1999] was developed in
Germany (DASA/MBB and Giessen University), the UK (AEA Technology) and Italy
(Proel/Laben), under ESA sponsorship. The respective responsibilities were the discharge chamber,
the grids and the neutraliser. The aim was to combine the best features available at the time in
European gridded thrusters, with basic development data being taken from the RIT-35 discharge
chamber [56], the UK-25 grid system [Latham, 1990], and the RIT-10 neutraliser [Bassner, 1991].
A sectional diagram is shown in Figure 35, from which the heritage is clear, with the discharge
chamber resembling that of the RIT-35, and the inwardly dished grids those of the UK-25, T5
(Figure 29) and T6 (Figure 36) thrusters. The neutraliser (not shown in Figure 35) is an enlarged
version of that of the RIT-10 thruster.
Figure 35 Diagrammatic representation of the ESA-XX RF thruster (Astrium GmbH diagram).
Using Xe propellant, the nominal thrust is 200 mN, the demonstrated thrust range is 10 to
200 mN, and the SI at a maximum power of 8.5 kW is 5400 s, so this device clearly meets the basic
requirements of many possible deep space missions.
83
g. XIPS-25 MESC Thruster
The XIPS-25 thruster, initially designed by Hughes prior to the development of the smaller
XIPS-13 [Beattie, 1991], is a high power and thrust device [Beattie, 1985] intended for the NSSK
role [Anzel, 1998] on the HS-702 series of communications satellites. It is also employed for the
later stages of orbit emplacement [Ocampo, 1998], thereby saving large quantities of chemical
propellant. It was further developed, with a larger diameter, as the NSTAR thruster [Christensen,
1999] so successfully used on the DS-1 mission [Raymann, 2000, Brophy, 2002].
This 25 cm beam diameter design recorded values of discharge chamber efficiency of about
110 W/A at an uncorrected propellant utilisation efficiency of 80%, rising to 140 W/A at nearly
95%. This outstanding performance was partly a result of the large volume of the device, which
kept losses low. By comparison, the smaller XIPS-13 thruster operates typically at about 220 to
240 W/A at 80% utilisation efficiency, which is very similar to Kaufman thruster values. However,
the anode voltage is much less than in a Kaufman thruster, being typically at 28 to 30 V. This
increases discharge chamber and screen grid lifetime, and values as high as 25,000 hr have been
predicted for this grid.
The 25 cm thruster was successfully operated for 4350 hr, including 3850 on-off cycles,
using a flight-type propellant feed system and a breadboard model power conditioners [Beattie,
1985]. This experience, in which significant performance degradation occurred, led to the doubling
of the accel grid thickness to increase lifetime. Perhaps surprisingly, the opportunity to bias the
decel grid to enhance further overall lifetime was not taken; this is now standard practice on the T5
thruster. With the thicker accel grid, a lifetime in excess of 12,000 hr is predicted. As mentioned
above, many flight qualified thrusters have been deployed with great effect on several
Boeing/Hughes HS-702 communications satellites; a photograph of a flight thruster is shown
below in Figure 36.
Figure 36 Photograph of XIPS-25 MESC ion thruster (Boeing/Hughes photo).
h. The UK-25 Kaufman-Type Thruster

This 25 cm beam diameter DC gridded thruster [Latham, 1990], shown in Figure 37, was
originally designed for interplanetary missions at more than 200 mN thrust, with the cancelled
Comet Nucleus Sample Return Mission [Melita, 1986] being the initial objective. It was based on
scaling relationships [Harbour, 1973, Wells, 1973] derived during work on the earlier T4A and T5
thrusters, and initial laboratory tests confirmed the validity of this approach. It was established that
the performance envelope is very wide, extending to at least 320 mN, with values of SI approaching
5000 s and an input power of nearly 10 kW. The programme reached engineering model status, but
no long duration testing was undertaken.
84
Figure 37 UK-25 engineering model ion thruster with earth screen removed (Culham Laboratory photo).
For the purposes of this report, data for the UK-25 taken at a thrust of 250 mN were assumed
to be typical, noting that the device can be throttled over a very wide range, certainly up to 320 mN,
and that a gridded thruster has no “best” operating point. In passing, it should be recalled that the
lifetime of a gridded thruster falls as the thrust is increased, owing mainly to greater grid erosion.
Consequently, the choice of 250 mN will ensure a longer life than would be achieved at the highest
values of thrust. Clearly, this is somewhat arbitrary, but it allows a reasonably accurate impression
to be gained of the overall characteristics of this thruster. At a higher thrust, the overall performance
would be better, but the lifetime would be somewhat reduced.
i. NSTAR MESC Thruster

The 30 cm diameter NSTAR thruster [Christensen, 1999] flown on the Deep Space 1
interplanetary spacecraft [Raymann, 2000, Brophy, 2002] has been remarkably successful, and has
exceeded its original objectives. It is shown mounted in the rear of the spacecraft in Figure 38.
Developed jointly by NASA Lewis/Glenn and Hughes, it is considerably de-rated from the earlier
30 cm thrusters [Beattie, 1990], with a maximum thrust of 92.6 mN at an input power of 2.3 kW
and a wide throttling range. A successful fly-by of the asteroid 1992 KD was completed on 29 July
1999 and the spacecraft then went on to observe from a relatively close distance the nucleus of the
comet Borelly [Brophy, 2002]. The thruster operating time in space amounted to more than 16,000
hours, and a ground test subsequently exceeded 30,000 hours [Sengupta, 2004].
Figure 38 NSTAR thruster mounted in rear of Deep Space 1 spacecraft (NASA photo).
Following this success a derivative thruster has been designed [Soulas, 2001] with an
enhanced performance. The aim is to extend the performance envelope to 5 kW, with the capability
to “process” 400 kg of xenon. This is not a particularly difficult objective, since thruster operation
85
has been demonstrated at beyond 10 kW, with values of SI at up to 4500 s. To accomplish the
necessary total impulse, a change to titanium or carbon grids is being studied; the former have been
tested to 4.6 kW and 172 mN thrust. Even under the new operating conditions, the thruster is very
considerably de-rated, so there should be no difficulty in meeting the revised requirements
j. NASA 30 cm Beam Diameter Thrusters

In the context of large ion thrusters, mention should be made of the very extensive
experience gained in the USA over nearly four decades of research and development effort on a
wide variety of 30 cm beam diameter devices. These have used both the Kaufman-type and cusp-
field discharge chamber configurations and the work has included the accumulation of a great deal
of life-test data with mercury and xenon propellant. Many life-tests of the 30 cm beam diameter
Kaufman-type thruster developed jointly by NASA Lewis and Hughes have been reported,
including close to 10,000 hours with the J-Series device [Beattie, 1990]. In other work, this thruster
exceeded input power levels of 10 kW [Patterson, 1988].
Despite these successes, the US programme was eventually diverted to become concentrated
exclusively on the cusp-field geometry device [Beattie, 1991], with a variety of thrusters being
developed to flight status. Nevertheless, the performances achieved with the earlier devices remain
of interest, so the J-series and 900-series are represented later in Table 2, using Xe and Hg
propellant, respectively.
As mentioned earlier, the Hughes 30 cm MESC thruster is also capable of a much greater
performance than indicated by the NSTAR thruster [Christensen, 1999] or its derivative [Soulas,
2001]. It has, in fact exceeded 10 kW by a considerable margin, reaching a thrust of 364 mN at an
SI of 5160 s.
k. The RIT-35 RF Thruster

This is an older 35 cm beam diameter RF gridded thruster, which has been the subject of
many years of work at Giessen University; a schematic diagram is shown in Figure 39. It showed
very good potential in tests in the 1970s using Hg propellant [Groh, 1979], and should not be
disregarded since excellent results were obtained later using Xe and redesigned extraction grids
[Groh, 1989, Groh, 1988]. Although estimated thrusts extended to 500 mN with an input power of
18 kW and an SI of about 5000 s using Hg, experimental data reached only 250 mN. There the
power input was about 7 kW. With Xe, testing attained only 104 mN, but this is not a true
reflection of the thruster’s capabilities.
86
Figure 39 Diagrammatic representation of the RIT-35 thruster (University of Giessen diagram)
The above diagram of the thruster represents the version using Hg propellant and flat grids.
Later developments employed outwardly dished grids to enhance thermal stability, since carbon-
carbon technology was not available at the time to provide a superior capability. As this grid
technology is inferior to that of the latest RIT thrusters, high perveance could be obtained only by
operating the accel grid at - 1700 V, with a major impact on potential life.
l. NAL/NASDA/Toshiba BBM-2 MESC Thruster

This Japanese consortium has developed a very promising DC gridded thruster of 35 cm
beam diameter [Hayakawa, 2000,Tahara, 1999], and has also investigated experimentally larger
devices. The target for this work was a thrust of 150 to 180 mN at an SI of 3500 s. The input
power was specified to be in the region of 4 kW and ring cusp MESC geometry was selected for the
discharge chamber. A very ambitious objective was set for the lifetime, 30,000 hours. A recently
reported long duration test reached 2659 hours at 150 mN, with an SI of 3518 sec, the latter being
partly due to the excellent utilisation efficiency achieved, which was 90%.
m. Tokyo Metropolitan Institute of Technology 30 cm MESC Thruster

This development [Komurasaki, 2000] is supported by Melco and again exploits the ring
cusp discharge chamber principle. The thruster has demonstrated a very high standard of
performance, reaching 153 mN at an SI of 3710 s and with an input power of 3.5 kW. As with the
BBM-2, the utilisation efficiency is very impressive, as indicated by the modest beam accelerating
potential of 1 kV, with values of around 93 to 94% being reported. The total thruster efficiency is
very close to 80%, which is amongst the best available. However, this is a laboratory device, with
no indication of lifetime and with no attempt reported of qualification activities.
n. NEXT MESC Thruster

This is an advanced, larger diameter version of the NSTAR thruster, with the title “NASA’s
evolutionary xenon thruster” (NEXT) [Patterson, 2002 – 2003, -Soulas, 2002]. It has a beam
diameter of 40 cm and an initial design input power of up to 10 kW, with a throttling range of 8:1.
However, these values were later modified to 6.9 kW and 11:1 [Hoskins, 2004]. The maximum SI
was originally specified as 4040 s, with a total efficiency of greater than 68%, but later values are
greater than 4100 s and 70%. Following extensive development effort, the mass is now to be less
than 1.8 kg/kW and the total impulse very large, with a throughput of Xe of more than 270 kg
[Hoskins, 2004]. Drawings of two views of the thruster are shown in Figure 40.
Figure 40 Two views of the NEXT thruster (NASA diagrams).
87
Work has reached the engineering model and preliminary endurance testing stage, with
additional research concentrating on the grid design [Soulas, 2002]. Three engineering model
thrusters were constructed, and one of these was tested to 2038 hours, at an average SI of 4110 s
and total efficiency of 69.4% [Soulas, 2004]. The mean thrust at the maximum power point of 6.9
kW was 237 mN. Although the objectives are ambitious, particularly concerning the grid design,
the limitations of the cusp-field concept are clearly illustrated by the severe fall in SI as the thrust is
reduced. In one set of tests, the thruster operated at 4060 s at 6.9 kW, but only 2300 s at 1.1 kW.
The efficiency fell from 69% to 51%.
o. NEXIS MESC Thruster

For the more challenging interplanetary missions, probably using nuclear power sources,
NASA have decided to invest in the development of ion thrusters much larger than NEXT. The
more conventional of these is the NEXIS device [Randolph, 2004], where this acronym is derived
from “nuclear electric xenon ion system”. This is a 65 cm discharge chamber diameter thruster,
again employing the MESC principle, and with a configuration similar to those of NSTAR
[Christensen, 1999] and NEXT [Patterson, 2002- Soulas, 2004] GIEs. However, the exit from the
discharge chamber is masked down to a diameter of 57 cm, so that grid erosion problems caused by
low plasma densities at the periphery are avoided. The initial mission for which it is intended is the
Jupiter Icy Moons Orbiter (JIMO) [Prockter, 2004, Oleson, 2004].
Development is being led by JPL, and the objectives include a power consumption of 20
kW, an SI of 7500 s using xenon propellant, with a beam accelerating potential of 7 to 8 kV, and a
total efficiency of more than 78%. It is hoped to achieve a total Xe throughput of no less than 2000
kg by combining a new reservoir hollow cathode, a new carbon-carbon grid configuration, and the
flat plasma density profile permitted by masking the outer region of the grids.
The operating thruster is shown below in Figure 41. This photograph was taken when the
power consumption was 27 kW, the total ion extraction potential 7 kV, and the thrust 517 mN. The
SI was 8700 s, the propellant utilisation efficiency 94% and the total efficiency 81%. Thus all the
goals have been surpassed, although durability has yet to be demonstrated.; an operating time of 5
to 10 years has been specified by NASA.
Clearly, it is impossible in a meaningful programme to demonstrate a thruster lifetime of 5 to
10 years. Bearing in mind that it is usually necessary to include a margin of 50% in a lifetest
programme, the thruster would have to be tested for between 7.5 and 15 years. The cost of this
would be prohibitive, the programme would be extremely protracted, and a failure well into the test,
for any reason, would greatly extend the total time required even further. With a programme of this
sort, a conventional lifetest is not in any way viable.
It is now assumed, with programmes of this sort, that qualification must be carried out by
calculation, assuming a good understanding of the fundamental life-limiting factors. The latter
must be determined by means of a detailed research programme, which must include experimental
validation of the main mechanisms which limit lifetime. However, full reliance cannot be placed on
such a process, at least for early missions, so additional redundant thrusters should be carried. The
latter requirement is not too difficult to implement because the masses of such thrusters are
relatively small in comparison with the masses of other parts of the system. As an example, the
mass of each T5/UK-10 thruster carried on the Artemis spacecraft was about 1.8 kg, whereas each
power conditioner weighed 11.7 kg [Notarantonio, 2003].
88
Figure 41 The NEXIS thruster operating at 27 kW (NASA photo).
p. HiPEP MESC Thruster

One of the most innovative ideas is the configuration of the HiPEP thruster, which has a
rectangular discharge chamber and grids. This acronym is derived from the full name, “high power
electric propulsion”. The initial application for this technology is likely to be the JIMO mission
[Prockter, 2004, Oleson, 2004], for which a total power consumption of 100 to 250 kW is envisaged
[Oleson, 2004], with an SI of 6000 to 9000 s and a total thruster efficiency of greater than 65%.
The operating time is specified as 14 years, which is equivalent to a propellant throughput of 100
kg/kW. The aim for an individual thruster is a power consumption of 20 to 50 kW.
The 91 cm × 41 cm rectangular shape [Elliott, 2004] was selected to simplify the scaling
process, both of the grids and of the discharge chamber. Two approaches to the discharge chamber
ionisation are being studied; one is the conventional DC discharge in an MESC magnetic field
configuration using a hollow cathode electron source. The other is the ECR mechanism, with an
applied frequency of either 2.45 GHz or 5.85 GHz. Both are illustrated in Figure 42, in which the
DC thruster is operating with an input power of 34 kW and the ECR variant at 13 kW. Clearly, the
DC variant offers the best prospects at present, but ignoring the influence of the discharge mode on
potential lifetime.
Figure 42 HiPEP thrusters operating at 13 kW (ECR discharge) and 34 kW (DC discharge) (NASA photos).
To date [Elliott, 2004], the DC variant has operated at up to 40 kW, providing a thrust of 670
mN at an SI of 9620 s. Under these conditions, the beam accelerating potential was 7 kV, the
propellant utilisation efficiency about 92.5%, and the discharge power slightly over 1 kW. The total
thruster efficiency was 80%, and throttling down to 9.7 kW and 240 mN was demonstrated. These
results were obtained with curved titanium grids, but flat pyrolytic carbon grids [Williams, 2004]
represent the next stage in the development process; these have now been tested.
q. Very Large Diameter Thrusters

89
Over several decades, NASA has shown interest in very large diameter thrusters for a variety
of missions involving large and heavy spacecraft, usually requiring extremely high values of ΔV.
As examples, at the beginning of their EP programme, NASA Lewis investigated the possibility of
building a 1.5 m diameter GIE [Nakanishi, 1967]. Although this was constructed, the performance
achieved did not meet expectations and the project was discontinued. The difficulties associated
with such large grids was found later to still be present at 50 cm diameter [Patterson, 1987], leading
to the conclusion that the 40 cm of the NEXT engine may be the maximum possible with present
technology. However, as reported above, successful results have been reported from the larger
NEXIS and HiPEP devices.
In this context, it should also be mentioned that NASA Glenn has commenced a programme
to develop a very high power, high SI thruster for precursor interstellar missions [Patterson, 2000].
Owing to the difficulties inherent in manufacturing and operating very large diameter grid systems,
this employs a 76 cm diameter discharge chamber fitted with three separate 30 cm diameter grid
sets, as shown in Figure 43. The SI is expected to reach 14,000 s using Kr propellant. The
specification includes a maximum beam accelerating potential of 13 kV, a thrust of 112 mN at an
input power of 10.3 kW, and a total efficiency of 76%. Initial results confirm that, using Xe, the
efficiency is lower than in the NSTAR thruster; this was expected because the ion extraction area is
much less than the total discharge chamber cross sectional area (Figure 43). The SI has covered the
range 1800 to 3550 s, with a maximum input power of 4.4 kW and a thrust of 154 mN.
Complete thruster Front view when operating with Xe

Figure 43 Photographs of NASA’s high SI thruster (NASA photo).
3.4.3. Summary of Capabilities of Existing Gridded Thrusters

Selected performance data from the thrusters described above are summarised in Table 2
below. This includes all known thrusters with a power consumption above 1 kW, excluding some
devices which are at an extremely early stage in their development. That status is denoted by the
heading “S”, and the references quoted are examples only.
90
Thruster Pro Beam Beam Typica Thrus Thrust Pow Power Specif S Ref
p Dia Curre l SI (s) t Density er Densit ic
(cm) nt (A) (mN) (mN/cm (kW) y Power
2
) (W/cm2 (W/mN
) )
T5 (high SI) Xe 10 1.05 4334 71 0.90 2.22 28.3 31.3 E 76
RIT-10 Evo Xe 8.7 0.85 3700 35 0.59 0.98 16.5 28.0 Q 87
RIT-15LP Xe 14.2 0.90 3400 50 0.32 1.3 8.2 26.0 E 84
SERT II Hg 15 0.25 4770 29 0.16 0.91 5.1 31.4 F 37
RIT-XT Xe 21 2.41 6419 218 0.63 8.06 23.3 37.0 Q 33
T6 Xe 22 3.38 4650 230 0.61 7.05 18.6 30.7 Q 120
ESA-XX Xe 25 3.25 3500 240 0.49 8.45 17.2 35.2 E 89
XIPS-25 Xe 25 4.0 4338 245 0.50 6.8 13.9 27.8 F 121
UK-25 Xe 25 4.45 4500 316 0.64 9.5 19.4 30.1 E 73
NSTAR Xe 30 1.76 3100 90 0.13 2.33 3.3 25.9 F 70
NSTAR Xe 30 2.7 5000 172 0.24 4.6 6.5 26.7 Q 99
Derivative
NASA J-Series Xe 30 4.0 3310 201 0.28 5.1 7.2 25.4 E 122
NASA 900- Hg 30 2.0 3000 128 0.18 2.6 3.7 20.3 E 123
Series
Hughes MESC Xe 30 5.6 5160 364 0.51 11.1 15.7 30.5 E 121
RIT-35 Xe 33.5 1.95 3195 104 0.12 2.97 3.4 28.6 L 56
RIT-35 Hg 33.5 2.45 4343 250 0.28 7.5 8.5 30.0 L 56
BBM-2 Xe 35 2.88 3518 150 0.16 3.3 3.4 22.0 E 124
Tokyo MESC Xe 30 2.5 3710 153 0.22 3.5 5.0 22.9 L 105
NEXT (design) Xe 40 5.8 4400 380 0.30 10.0 8.0 26.3 E 125
NEXT (test) Xe 40 3.52 4110 237 0.19 6.9 5.5 29.1 E 110
NEXIS Xe 57 4.0 7500 415 0.16 20 7.8 48.2 L 111
HiPEP (DC) Xe 91 × 5.5 9620 670 0.18 39.3 10.5 58.7 L 114
41
NASA Kr 3× 2.3 14,000 345 0.16 30.0 14.1 87.0 L 119
Interstellar 30 dia
(predicted)
Table VII Summary of high power thruster capabilities.
Before commenting on the thruster characteristics presented in Table 2, it should be

mentioned that three propellants are included; these are Hg, Xe and Kr. These represent three
phases of the decades-long gridded ion thruster development period. Before explaining the broad
features of this, it should be pointed out that any material which can be a vapour or pure gas at a
convenient discharge chamber temperature can be used as a propellant, provided that it does not
chemically attack any part of the thruster. It is also desirable, but not essential, that it will not attack
chemically any part of the spacecraft. This broad range of possible propellants is discussed later in
section 3.5.1, including Figure 50, and section 3.5.3.3.
In the early days of research into gridded thrusters, it was initially decided that the highest
possible thrust should be achieved for a given beam current and power. Thus the atomic mass
(AM) of the propellant should be as large as possible, and Hg was selected, since its AM is 200.
This was very satisfactory and the flight of two thrusters for 11.5 years on the SERT-2 spacecraft
[Kerslake, 1971] did not indicate that this propellant was other than totally acceptable. However,
personnel and spacecraft safety considerations later led all agencies in this field to change to an
inert gas, Xe, with an AM of 131. Although not so good as Hg, from the point of view of thrust for
a given beam current, it performs satisfactorily, and nearly all present programmes rely on it.
However, with the possible need for much higher values of SI, programmes with that
requirement are now moving towards gasses with lighter AM, such as Ar and Kr. Although much
more difficult to store on board a spacecraft for a long period of time, they offer the required
performance enhancement, as is specified for the NASA Interstellar effort [Patterson, 2000].
91
S = status; E = engineering model, F = flight model, L = laboratory model, Q =
qualification model.
It is clear from Table 2 that all thruster technologies are capable of scaling to higher power
levels, and that the general requirements of many NEP missions can be satisfied. A power input of
10 kW has been reached or exceeded by all types, with the HiPEP concept attaining 40 kW. The
basic conclusion is that it is possible to design large thrusters which consume power levels of
interest; how far this process can continue is largely dictated by the ability to design grid systems
with adequate durability and resistance to the launch environment.
Similarly, twin and triple-grid ion extraction systems have been shown to be entirely viable
at much higher potentials than usually applied, with voltages of up to 7 kV routinely used. Indeed,
laboratory experiments have been conducted successfully at up to 30 kV [Wilbur, 2004]. As a
result, values of SI of close to 10,000 s have been reported for the HiPEP thruster, using Xe
propellant, and it is predicted that 14,000 s will be achieved with the NASA interstellar precursor
thruster with Kr. However, under these circumstances the power fed into the beam is much higher
than with low SI, so the power-to-thrust ratio becomes much greater; it reaches 87 W/mN in the
case of the interstellar precursor thruster. This trend cannot be avoided and must be accepted, so
this ratio is no longer of use as a performance indicator.
Thrust levels are also seen to rise with the dimensions of the thruster and with power
consumption, as expected, reaching 670 mN in the case of the HiPEP device. However, this is not
impressive when compared to some of the smaller devices; for example, the 25 cm beam diameter
UK-25 attained 316 mN under test with only a modest value of SI [Latham, 1990], and the smaller
T6 [Huddlestone, 2004, Wallace, 2003] and RIT-XT [Leiter, 2003, Leiter et al., 2003] routinely
operate at above 200 mN. In this context, the parameter of importance is the thrust density, since
this determines how much thruster mounting area will be required on a spacecraft to produce a
specified thrust. This area is sometimes very restricted and can thus become a primary design
parameter [Fearn, 2004].
It is interesting to note in Table 2 that all the advanced large thrusters operate at very modest
values of thrust density, typically below 0.2 mN/cm2. This follows the trend set very early in the
GIE development process by the SERT II thrusters [Kerslake, 1971], which achieved 0.16 mN/cm2,
a trend continued by the NSTAR GIE [Christensen, 1999] with 0.13 mN/cm2. Even the NSTAR
Derivative [Soulas, 2001] only extends this value to 0.24 mN/cm2. While the reason for these low
values is the need to ensure long grid life, what can be accomplished by a qualified device is
indicated by the XIPS-25 [Beattie, 1990], which is in commercial service on many communications
satellites [Christensen, 2004]. This reaches 0.5 mN/cm2. This value is surpassed by two thrusters
in the qualification phase of their development for missions of very long duration, which include the
challenging BepiColombo mission to Mercury [Novara, 2001]; these are the T6 [Huddlestone,
2004, Wallace, 2003] and RIT-XT [Leiter., 2003, Leiter et al., 2003], with typical thrust densities of
0.61 and 0.63 mN/cm2, respectively.
What can be accomplished, at moderate SI, is indicated by the high power version of the T5
GIE [Fearn, 1993], with a maximum thrust density of 0.90 mN/cm2, at an SI of 4334 s. It should be
noted that if the SI was increased at constant beam current density, this thrust density would be
much larger.
Thus it is thus clear that the beam current density is of prime importance in achieving a high
thrust density. The values for the thrusters listed in Table 2 vary by more than an order of
magnitude, with the lowest being recorded for the larger devices; this is surprising, because they
are operated at higher values of SI, therefore larger extraction potentials. It will be shown later that
such conditions are conducive to attaining higher values of beam current density, but this has
certainly not been achieved. The values derived from Table 2 vary from 1.1 mA/cm2 for the
interstellar precursor thruster and 1.6 mA/cm2 for NEXIS, through 2.5 mA/cm2 for NSTAR,
reaching 14.3 mA/cm2 for the RIT-10 Evo and 13.4 mA/cm2 for the T5. The value for T6 is 8.9
92
mA/cm2. It is therefore certain that the performances of the larger thrusters in Table 2 can be
increased very considerably, with close to an order of magnitude being feasible.
This conclusion is amplified by an examination of power density, which should increase
dramatically with SI, if the physical principles utilised in designing these devices remain the same
and independent of dimensions. This is certainly not the case, with the power densities for all the
newly developed large GIEs being well under the values achieved by the older, smaller thrusters.
The highest recorded power density is that of the T5 GIE, with 28.3 W/cm2, with the RIT-XT
reaching 23.3 W/cm2. Many other relatively low SI thrusters exceed 15 W/cm2, whereas the best of
the more modern high SI devices, the HiPEP, achieves experimentally only 10.5 W/cm2, and that is
at an SI of nearly 10,000 s. There is, very clearly, considerable room for improvement here. That
this can be realised is shown later in this report.
3.5. The Scaling Process

The scaling relationships which primarily determine the performance of a gridded thruster
concern the ion extraction and acceleration grid system. In evaluating the capabilities of any grid
system it is assumed that the required flux of ions to the screen grid (Figure 22) can be supplied
from the discharge chamber plasma under all circumstances. The scaling of this plasma is
reasonably well understood for both RF and DC thrusters. It should be emphasised that the
separation of the process for producing the ions from that for extracting and accelerating them is a
major advantage, which is not found in most other EP devices. A further advantage of this concept
is that the important performance parameters can be dealt with separately to first order. Thus,
although there is, for example, an impact on thrust if the SI is modified by altering either the
extraction potential or atomic mass of the propellant, this can be compensated for precisely by a
change of beam current.
A much more complex problem is to design an extraction system to meet specified lifetime
requirements, although the ion optics models now available permit reasonably accurate estimates to
be made of total erosion rates and of the locations of the dominant erosion sites. However, an
important limitation of this modelling activity concerns the lack of accurate primary data regarding
sputtering yields as functions of energy and angle, and also charge-exchange cross sections as
functions of particle energy. In most thrusters, there is also inadequate information concerning the
way in which the plasma density and electron temperature profiles in the discharge chamber vary
with operating conditions; such information is costly to obtain in a comprehensive way.
3.5.1. Grid Design Options

In all present thrusters the ion extraction system is of either twin- or triple-grid
configuration, as depicted in Figure 44. Such grids operate in an accelerate/decelerate mode to
enhance the throughput of ions. In this scheme, the thruster body and screen grid are at the high
potential, VB, necessary to achieve the ion beam velocity required. For example, if Xe ions are to be
emitted at 40 km/s (SI ~ 3500 s), a potential of 1.1 kV is needed. The accel grid is at a negative
potential, Vac, of perhaps - 250 V, to provide focusing of the beamlets, to enable the required current
to be extracted, and to prevent electron back-streaming into the discharge chamber from the
external plasma. Deceleration to space potential then follows; with the triple-grid variant this is via
a less negative voltage, Vdec, applied to the decel grid.
93
Figure 44 Schematic diagrams of twin and triple-grid ion extraction systems.
To raise the ion velocity, ve, and thus the SI, it is merely necessary to increase VB as
appropriate. However, this causes a greater penetration of the inter-grid electric field into the
discharge chamber plasma. Some penetration, as depicted in Figure 44, is desirable, since this
results in a curved plasma sheath which both increases the ion emission area and aids focusing.
Unfortunately, the curvature becomes severe at high voltages, influencing the ion trajectories
adversely and causing direct impingement on the outer grids.
This situation is depicted in Figure 45, in which the sheath positions for a moderate and a
high electric field, E, are shown. Direct ion impingement on the accel grid is evident in the latter
case. Moreover, the plasma number density, n, in the discharge chamber must also be taken into
account, since the penetration of the sheath increases as n falls. This effect is illustrated in Figure
45 for the grid configuration of the T5 thruster [Bond, 1997]. This problem is relevant to the use of
a wide throttling range, or if the radial plasma density distribution is strongly peaked. However, in
the latter case the dimensions of the screen grid apertures can be matched to this distribution to
alleviate adverse effects [Bond, 1997].
Figure 45 Triple-grid configuration, indicating the effects of the sheath shape on ion trajectories.
94
Figure 46 Plasma sheath penetration as a function of number density for the T5 thruster.
Thus it is clear that high values of E and/or low values of n allow the sheath to penetrate
deeply into the plasma. The resulting curvature of the surface from which the ions are extracted
causes many of the trajectories to diverge from the desired paraxial direction, and to impact upon
the accel or decel grids. The erosion that this causes severely limits lifetime and cannot be tolerated
for any but short missions.
This problem is avoided by the use of 4 grids, since the extraction field is defined by the first
two of these (Figure 47). Most of the ion acceleration then takes place between the second and
third grids, where the greater part of the applied voltage appears. Thus the ion extraction process
can remain constant no matter what final velocity and SI are required. This configuration originated
in the CTR research community in order to produce very high energy particles for injection into
fusion machines [Martin, 1984, Okumura, 1980]. Plasma sources fitted with such grid systems are
used to accelerate ions, usually hydrogen, to energies that can exceed 100 keV. They are then
passed through a charge-exchange gas cell to produce the required high energy beams of neutral
atoms. These are necessary to penetrate the large magnetic fields within fusion devices.
Figure 47 Schematic diagram of 4-grid system.
Thus in the 4-grid system, the ion extraction process is separated from any subsequent
acceleration. The ions are extracted from the plasma by the first two grids, which can operate at a
potential difference providing an acceptable field penetration into the discharge chamber plasma,
and direct impingement onto the downstream grids is avoided. The bulk of the acceleration then
occurs between the second and third grids, which can be widely separated to accommodate the
potential applied, which can be very large. The negative third grid acts in the same way as the accel
grid in the conventional two or three electrode system, as does the 4th grid. It should be noted that
the potential of the third grid in Figure 47, -1 to –2 kV, applies to the large grid apertures often
found in CTR devices. A more realistic value for a typical ion thruster would be –250 V.
95
3.5.2. Exhaust Velocity and SI
From the point of view of many missions, the most important parameter is ve, which
determines the value of SI achieved. This is given by
2eVB
ve = (6)
mi
where e is the charge on an electron, mi is the ion mass, and the ions are assumed to be
singly charged. Thus ve is determined by VB and by mi. In practice, VB is the potential of the body
of the thruster and of the screen grid, which is typically 1 to 2.5 kV in present designs. The total
accelerating potential is larger than this because Vac ∼ -200 to -500 V. With triple-grids, the decel
grid is at Vdec ∼ -50 V and decelerates the ions. Further deceleration occurs within the space
plasma.
Although ve is given by Equ 6, in reality it is reduced to an effective value, veff, determined
by the total propellant utilisation efficiency, ηmt, where this is the fraction of the total propellant
flow rate, m& t , accelerated into the ion beam. Thus veff = ηmtve . If IB is the ion beam current,
IBmi 1
ηmt = (7)
e m& t
As shown before in Equ 4, the SI, denoted by Isp, can be defined as the ratio of thrust, T, to
total rate of use of propellant, in units of sea-level weight. Thus
T
Isp = (8)
m& tgo
where m& t is now used rather than m& to differentiate between it and the other propellant
flows into a thruster. Using veff = ηmtve and Equs 6 and 7,
veff IB 2VBmi
Isp = = (9)
go m& tgo e
Thus it is possible to calculate the SI with good accuracy without using a thrust balance.
With screen to accel grid separations of the order of 0.5 to 1 mm, operation at the few kV
level is satisfactory, providing values of ve for Xe of up to about 65 km/s, and giving an SI of
between 2800 and 5500 s [Latham, 1990]. With greater separations and suitable insulator designs,
higher values of VB can be used, so that, from Equ 9, enhanced values of SI are feasible until the
plasma sheath curvature demands a change to a 4-grid system.
A further benefit arising from an increased extraction potential is a gain in the perveance or
extraction efficiency. This causes more of the ions drifting towards the grids within the discharge
chamber to emerge into the beam, increasing IB and T, and also power consumption. Moreover, an
associated improvement in ηmt reduces the rate of grid erosion. As an example of what can be
achieved, the values of ve and Isp attained using a relatively low atomic mass propellant, argon, are
plotted in Figure 48 for potentials of up to 40 kV. The latter voltage (which would necessitate the
use of a 4-grid system) gives a velocity of over 400 km/s and an SI of about 40,000 s, assuming ηmt
= 0.88.
96
Figure 48 Exhaust velocity and SI for argon as functions of accelerating potential.
Equ 6 shows that ve is dependent upon the propellant employed, as demonstrated by the high
values for argon plotted in Figure 48. Xe was originally chosen, as was mercury in earlier times,
for its high atomic mass; this is beneficial because it provides an excellent momentum transfer and
thrust for a given beam current. However, for ultra-high SI missions, a lower mi may be required.
The best that can be achieved is to use hydrogen, but this presents storage difficulties, since very
high pressures or extreme cryogenic temperatures are required. Nevertheless, suitable compounds
of hydrogen exist with average atomic masses when completely dissociated of 4 to 5 atomic mass
units (AMUs) [Fearn, 2000]. Although dissociation must occur within the discharge chamber, the
power required is not excessive, and the SI can be increased by a factor of about 5 compared to Xe.
3.5.3. Thrust and Perveance

As mentioned previously, the thrust and thrust density achievable are also important,
especially for missions where a time constraint exists. In such cases, if power is limited, it may be
necessary to forego the advantages of a very high SI in order to achieve a particular level of thrust
determined by mission requirements. In addition, low thrust densities may not be permissible
because adequate volume and surface area may not then be available for the propulsion system
[Fearn, 2004].
Thrust is related to the other parameters by
IBmive
T = IB 2VBmi =
e (10)
e
Thus, T will increase with VB, but decrease if mi is reduced. Compensation for this can be
obtained by extracting a larger IB, which requires a higher plasma density in the discharge chamber
and thus a greater discharge power. The limit is the ability of the grids to pass this increased
current. This is determined by their perveance, which can be defined [Harbour, 1973] as
1/ 2
IB 4εo ⎛ 2e ⎞ ATg
= ⎜ ⎟ (11)
VT3 / 2 9 ⎝ mi ⎠ d2
97
where VT = (VB + Vac ) , εo is the dielectric constant of free space, d is the ion acceleration
distance, A is the area of the grids, and Tg is the effective transparency of the screen grid. This leads
to the definition of a perveance parameter, Pg, which is
IBmi1 / 2 d 2 4εo 2e 3 / 2
Pg = = VT (12)
ATg 9
As can be seen in Figure 49, a logarithmic plot of Pg against VT is linear; here IB is in amps,
mi is in AMU, d is in mm, and A is in m2. Experimental points from the T5 [Fearn, 1993, Martin,
1988] and UK-25 [Latham, 1990] Kaufman-type thrusters were used to provide the indicated trends
and are consistent with the expected 3/2 power law.
Figure 49 Values of Pg as a function of VT.
From Equ 12, IB will fall if grids of a given design have to be separated by a greater distance
to accommodate higher potentials. However, increased ion currents are readily accessible at the
enhanced values of VT discussed here.
The data in Figure 49 fall on two lines, labelled “peak performance” and “Artemis
conditions”. The former represents the maximum perveance case, where lifetime is likely to be
limited. The “Artemis” line applies to the conditions required for long life, as specified for the T5
thruster for operational use on the Artemis communications satellite [Gray, 1997], for which a
lifetime of close to 15,000 hours was required.
3.5.4. Current, Power and Thrust Densities

The maximum current density, JB, emitted by the grids is, from Equ 11,
1/ 2
4εo ⎛ 2e ⎞ Tg 3 / 2
JB = ⎜ ⎟ VT (13)
9 ⎝ mi ⎠ d2
Assuming that the bulk of the power supplied to the thruster is used to accelerate the beam,
and thus operation is at high SI, the power density, Pd, is found by multiplying JB by VB, so that
1/ 2
4εo ⎛ 2e ⎞ Tg 3 / 2
Pd = ⎜ ⎟ VT VB (14)
9 ⎝ mi ⎠ d2
98
From Equ 10, the maximum thrust density, Td, is
1/ 2
⎛ 2miVB ⎞
Td = JB⎜ ⎟ (15)
⎝ e ⎠
where JB is given by Equ 13.
3.5.5. The Discharge Chamber

The grid relationships discussed above are applicable to all gridded thrusters. The only
difference between the detailed design for one thruster and another concerns the plasma density
distribution at the screen grid. This determines the radial ion beam density distribution and
influences the maximum and mean values of all parameters. Thus the only part of the scaling
process which is dependant upon the type of thruster is that applicable to the discharge chamber.
Of course, a main parameter for scaling the thrust is the effective area of the grid system.
This can in principle be increased as necessary, keeping the maximum current density (usually on
the axis) constant. This is accomplished by raising m& t and the discharge power in the same ratio as
the specified increase in IB. Second order effects include the improvement in ionisation efficiency
as the size increases, owing to the internal surface area being dependant upon the square of the
radius, whereas the volume has a cubic relationship.
With the RIT-XT/RIT-22 type of RF thruster [Leiter, 2004], the skin effect will cause the
ionisation rate to depend on radius, so that, for the larger devices, it may peak somewhere between
the centre line and the wall. The plasma density distribution, and thus the beam current density,
might therefore be “hollow”. This can be avoided by careful design, although it does not adversely
effect performance in most cases. The scaling of ECR devices is similar, although more account
must be taken of the skin effect due to the use of frequencies which can reach above 5 GHz. Thus
the geometries of both the waveguide/antenna combination and of the discharge chamber are more
significant than at 1 MHz.
Owing to the more complex geometry, scaling Kaufman-type DC thrusters to different
diameters is more complicated than for RF devices. However, reasonably accurate scaling
relationships exist [Harbour, 1973,Fearn, 1997] which were successfully employed in designing the
UK-25 [Latham, 1990] and T6 thrusters [Huddlestone, 2004]. Indeed, these designs worked well
“off the drawing board”, thereby validating the scaling process which starts by considering the need
to ensure that the mean flux density of ions to the screen grid remains constant, and that there is no
change in the radial plasma density distribution. This leads to quantification of the ionisation rate,
then to the current of primary electrons required, assuming that their energy remains constant to
first order. The next step is to consider the conditions in the annular gap between the baffle disc
and the inner polepiece (Figure 27), in which the primary electrons gain their energy. This
determines the dimensions of this region, the cathode operating conditions, and the magnetic field.
However, the latter is changed during optimisation and throttling, so only a broad specification is
required at this stage.
The scaling of an MESC thruster follows similar principles, although the magnetic field
value is much greater than in the Kaufman device, and its geometry is different (Figure 28).
However, the ionisation process and the magnetic containment principles are the same, so the
physical situation in the two devices is similar, and the ultimate requirements are essentially
identical. However, there is no direct control of the energy of the primary electrons, which is
determined by the difference between the cathode and discharge chamber plasma potentials, which
are not easy to regulate.
99
3.6. High SI, High Power Operation
3.6.1. Parametric Variations

To illustrate the influence of atomic mass (AM) and VB on performance at high SI, the
maximum thrust density is plotted against AM in Figure 50 for values of VB of between 1.5 and 70
kV and for constant IB. The use of a 4-grid system is assumed, with a constant ion extraction
potential, except in the 1.5 kV case. These results are based on the use of Equ 10 and on data from
the T5 thruster operating with Xe at maximum perveance, at a thrust of 64 mN [Martin, 1988].
Under these conditions, VT = 2 kV, VB = 1500 V, IB = 1.0 A, Vac = -500 V, Isp = 4071 s, and the total
input power, PT, is 1930 W. This is clearly an exceptional performance for a 10 cm beam diameter
thruster, although it should be noted that the performance envelope for this device extends beyond
this point to more than 70 mN.
Figure 50 Thrust density as a function of atomic (ion) mass and net ion accelerating potential.
As anticipated from Equ 10, at any value of VB and constant IB the thrust density falls with
decreasing ion mass, with a square root relationship. While this trend can be counteracted by
increasing IB, a higher perveance may then be necessary. At constant VB, this can be achieved only
by making the accel grid more negative, with adverse lifetime implications, or by closing the inter-
grid gap (see Equ 11).
Also as indicated by Equ 10 and Figure 50, at any value of AM the thrust density increases
with VB, again with a square root relationship. As a consequence, the rate at which gains are made
in thrust density diminish as VB is increased, so the greatest values are to be achieved by combining
high values of VB and mi with maximum perveance. The latter implies a small inter-grid gap and a
large value of Vac, neither of which are conducive to long life. Thus a compromise will always be
required.
The values of Td given in Figure 50 extend to well beyond those quoted for existing thrusters
in Table 2. Indeed, they reach about 7 mN/cm2 for Hg propellant, which is more than an order of
magnitude greater than found in most present devices. This confirms, as stated previously, that
100
thrust densities of interest for many future high power missions can be generated by increasing the
ion energy. It also confirms that the objectives in developing the NEXIS [Randolph, 2004], HiPEP
[Elliott, 2004] and interstellar precursor [Patterson, 2000] thrusters are far from ambitious; much
more can be achieved with devices of this generation, bearing in mind their dimensions and
intended values of SI.
3.6.2. Operation at 100s of kW Level

As a more specific example of the parametric variations which are predicted by the scaling
relationships discussed above, the case was considered of a 30 cm beam diameter thruster operating
under the conditions appropriate to the T5 engine at 71 mN thrust [Martin, 1988], where the SI was
4334 s. For the purpose of this illustration, it was assumed that this thruster will be run at a higher
SI to ensure that the grids are not at their perveance limit. Thus a nominal value of 5000 s was
deemed to be more appropriate. The beam current density is then 13.4 mA/cm2, giving a beam
current with 30 cm diameter grids of 9.47 A. Assuming the use of Xe, the thrust density is then 0.9
mN/cm2 and the power density 28.3 W/cm2, giving a thrust of 636 mN and a power consumption of
exactly 20 kW. Such a device would be appropriate for a 100 kW application, assuming that a
cluster of 5 or more mounted on the spacecraft would be acceptable.
3.6.2.1.Increase of Thruster Diameter

To cater for different applications in this power regime, Figure 51 illustrates how the thrust
varies with diameter, when operating under these high perveance conditions. Also plotted are de-
rated values, derived from T5 data at 64 mN; lifetime would be enhanced in this case. For
completeness, these plots extend to 50 cm diameter, although the development of satisfactory grid
systems of this size will be difficult, owing to problems associated with maintaining the very small
gaps required in an environment which includes severe vibration at launch and high but variable
temperatures during operation. Indeed, the largest diameter thruster which has successfully
overcome these problems inherent to space qualification to date is the 30 cm NSTAR device
[Christensen, 1999] of the DS-1 spacecraft [Raymann, 2000, Brophy, 2002].
2
Full Thrust De-rated Thrust
1.5
Thrust (N)
0.5
0
10 20 30 40 50
Thruster Diameter (cm)
Figure 51 Thrust as a function of thruster diameter.
101
The useful thrust for relevant missions is within the range 0.5 to 1.7 N, which is certainly
appropriate to many applications. However, as will be shown below, it can be increased very
considerably by operating at higher values of SI than the 5000 s assumed here. The equivalent
power consumption is shown as a function of diameter in Figure 52. This is also consistent with
clustered operation at the 100 kW level; indeed, just two 50 cm thrusters would be needed at 100
kW.
60
Full Thrust De-rated Thrust
40
Power (kW)
20
0
10 20 30 40 50
Thruster Diameter (cm)
Figure 52 Power consumption as a function of thruster diameter.
3.6.2.2. Increase of SI
Thrust density values scaled directly from the T5 thruster when operating at 64 mN were
presented in Figure 50, as functions of net accelerating potential and propellant atomic mass,
assuming that the ion extraction potential remains constant at 2 kV. This operating point was
selected to provide some de-rating from the maximum performance achieved, 71 mN. As a
consequence of the assumption concerning the extraction potential, all lines in Figure 50, apart from
that for 1.5 kV, are appropriate to a 4-grid system.
As an example of what might be achieved at a higher value of SI than considered above, a
net accelerating potential of 5 kV was assumed, while keeping the extraction potential at 2 kV. As
stated earlier, this requires the use of the 4-grid concept. For Xe, this yields about 1.5 mN/cm2 and,
for a 40 cm diameter thruster, which will provide a thrust of about 1.87 N at an SI of 7500 s,
assuming that ηmt is 85% and VB = 5 kV. The power consumption will be around 86 kW, assuming
that there is no increase in ionisation efficiency in the discharge chamber.
With a perhaps more realistic diameter of 30 cm, these figures become 1.05 N and 48.6 kW.
Such a performance is certainly relevant to a 50 to 300 kW requirement and suggests that
appropriate improvements to existing technologies will permit such objectives to be attained with
very few thrusters, if necessary. Indeed, it would seem to be feasible to consume 300 kW using just
6 thrusters of 30 cm diameter, operating with just a modest increase in SI above current maximum
values.
As stated above, this conclusion was reached assuming an ion extraction potential limited to
2 kV. While larger values have been investigated only rarely up to the present, considerably higher
values have already been achieved using a triple- or twin-grid systems [Randolph, 2004, Elliott,
2004, Patterson, 2000], and Wilbur has reported using up to 30 kV successfully [Wilbur, 2004].
However, the design of fully practical devices operating at such high potentials and thus perveance
values must assume that the discharge chamber plasma is dense and that the radial density
102
distribution is not excessively peaked. From Equ 11, such a change will increase the available
beam current, power consumption and thrust dramatically, via the (VT 3 / 2 ) relationship in this
equation. However, as mentioned previously, existing advanced thrusters (Table 2) do not appear
to have been designed to take advantage of this.
Taking again the operating parameters of the T5 thruster at 64 mN thrust and extending them
to higher values of SI by means of Equs 13 to 15, the results obtained are plotted in Figure 53.
They show clearly the major impact of the improved perveance achieved by increasing the potential
between the screen and accel grids. With the triple- or twin-grid system this increase is caused
automatically if the SI is raised, since VB, and thus VT, must then be higher. It can be seen that the
beam current can increase to as much as 41 A for a 30 cm diameter thruster at 5 kV, assuming that
the discharge chamber can provide the necessary flux of ions to the screen grid. The corresponding
thrust will be 4.8 N and the power consumption 264 kW. While these figures may appear to lack
realism in view of the need to achieve also an appropriate lifetime, they do confirm that
performance parameters consistent with a requirement to consume 100s of kW are achievable with
an extension of existing technology.
6 300
Thrust Power
5
Input Power (kW)

4 200
Thrust (N)
2 100
0 0
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Net Ion Accelerating Potential (kV)
Figure 53Thrust and power consumption as a function of VB for a 30 cm beam diameter thruster.
As an example, if there is a need to achieve 300 kW with perhaps 6 thrusters, these results
show that this will be possible with any value of SI above about 5200 s, remembering that it is not
necessary to operate at maximum beam current. Thus, as an example, if 7000 s is needed for a
particular application which will use 6 thrusters, the value of VB required is around 4.5 kV, but the
thruster can then be de-rated from its maximum performance by a factor of 4, to reduce the power
consumption from 200 kW to 50 kW. This will normally be extremely beneficial as regards
lifetime. However, note should be taken of the resulting reduction of the plasma density in the
discharge chamber and of the consequential further penetration of the plasma sheath at the screen
grid in the upstream direction. This must not be permitted to extend so far that lifetime is
compromised due to direct ion impingement on the accel or decel grids (Figure 45).
3.6.2.3. Increase of Perveance

The analysis presented above has illustrated the major effects which can be obtained by
increasing the perveance through the process of employing a higher value of SI. Basically, to
increase the value of SI achieved, the ion accelerating voltage, VB, applied to the grids must be
larger and this also improves the ion extraction capabilities of the grid system, or its perveance (see
Equ 12).
103
Unfortunately, owing to mission constraints, it might not be permissible in practice to
operate at a larger SI, so this avenue of improvement (ie the use of a higher acceleration voltage)
might not be available. However, the perveance of twin- and triple-grid systems is also dependent
upon the accel grid potential, Vac, so this can also be changed to achieve the desired results. Of
course, more negative values of Vac will accelerate the damage caused by charge-exchange ions,
thereby reducing lifetime. This is because the charge-exchange ions created between and outside the
grids gain an energy close to Vac in being accelerated back to the grid system. The higher the
negative value of this voltage, the more damage they cause.
This was originally a difficulty which could not be overcome easily, but the advent of carbon
grid technology has allowed much more negative potentials to be used. As the earlier RIT RF
thrusters [Loeb, 1970 - Bassner, 1994] operated successfully with potentials as negative as about -
1.5 kV, without the benefit of carbon grids, it is likely that the – 2 kV will be acceptable now.
To illustrate the results of examining a range of values of Vac, calculations were performed
for the intermediate SI case of 5260 s, for which VB = 2.5 kV assuming the use of a 30 cm beam
diameter thruster In this assessment, the value of Vac was varied from -250 V to -2000 V. The
results are presented below in Figure 54, from which it can be seen that the thrust covers the wide
range from 1.2 to 2.5 N and the power consumption from 47 to 98 kW. To take one specific
operating point, the thrust is 1.4 N and the power consumption 53 kW if Vac = -500 V. These are,
of course, maximum values for the conditions under consideration, and the thruster can be operated
at any reasonable lesser ones.
3 100
Power Consumption (kW)

Thrust Power
2.5 80
Thrust (N)
2 60
1.5 40
1 20
0 0.5 1 1.5 2
Negative Accel Grid Potential (kV)
Figure 54 Thrust and power consumption as a function of Vac for a 30 cm beam diameter thruster operating with
VB = 2.5 kV.
This assessment illustrates the strong influence of grid system perveance, or ion extracting
capability, on the maximum performance of a gridded thruster, for given ion beam diameter and SI.
These and the earlier results show that there are several ways in which high performance can be
achieved, and suggest that present devices, modified in appropriate ways, can probably meet a
multi-100s of kW power consumption requirement, even if a mission demands that very few
thrusters be flown and that long lifetime must therefore be achieved. It is also clear that, in this
case, there is no need to utilise very high values of SI, unless they are advantageous for saving
propellant mass.
3.6.2.4. Technology Readiness

The analysis presented in Table 2 suggests that present thrusters cannot meet the high power
end of a 50 to 100s of kW range requirement. However, an array of certain existing devices could
104
probably consume up to 100 kW or so, although none of these thrusters are space qualified and they
all require life-testing. Thus they are not yet ready for this application and considerable funding
would be needed to reach qualified status; in this, it should be noted that the life-testing of large,
high power thrusters requires the use of major facilities with substantial pumping speeds, which are
costly to build and to operate
The higher end of this power range can be met by the further development of existing
thrusters, utilising modified grid systems and by operating at higher SI and/or perveance than
normal. The latter can be achieved solely by making the accel grid more negative, although the
increase of SI has the same effect if twin- or triple-grids are used. The more complex alternative is
to employ a 4-grid system with both a greater perveance and providing an increased SI. In all cases,
additional research and development are needed, since this regime has not been properly explored
to date. The lifetime implications are of particular interest and may dictate the way forward.
3.6.3. Operation at the MW Level

From the above assessment, it is necessary to employ the 4-grid system to bring a large
thruster into the 100s of kW to MW class. To establish what might be possible, the maximum
performance of a 30 cm beam diameter thruster was again evaluated, using Xe propellant and the
scaling relationships in Section 4. This diameter is the largest for which full confidence exists
regarding tolerance to launch vibration and lifetime. In these calculations, the screen and accel
grids were separated by 1 mm, and a 5 kV ion extraction potential was applied. However, only
50% of maximum grid perveance (Equ 12) was assumed to avoid any direct ion beam impingement
on the downstream grids (Figure 44 and Figure 45), and VB was taken to be 7 to 70 kV and ηmt to be
90%.
The results, presented in Table 3, also include the case when full perveance is assumed. It
can be seen that values of PT in the specified range can be readily achieved, provided that the SI is
allowed to rise to well beyond current values. As these high values of PT result from depositing
most of the energy supplied directly into the beam, the discharge power required remains relatively
modest in all cases. Thus a high thruster temperature should not be a major problem.
Net Beam Accelerating Potential (kV)

7 10 15 20 30 50 70
Maximum Thrust (N) 5.3 7.5 9.2 10.6 13.1 16.8 19.9
Thrust Density (mN/cm2) 8.8 10.6 13.0 15.0 18.5 23.8 28.1
Max Beam Current (A) 43.9 43.9 43.9 43.9 43.9 43.9 43.9
Input Power at 50% Perveance 0.32 0.46 0.69 0.91 1.37 2.28 3.19
(MW)
SI (s) 9300 11,100 13,600 15,700 19,300 24,900 29,400
Power Density (W/cm2) 451 644 966 1288 1933 3221 4509
Input Power at Full Perveance 0.64 0.92 1.37 1.82 2.74 4.56 6.38
(MW)
Table VIII Predicted performance of a 30 cm beam diameter thruster utilising Xe propellant, a 4-grid extraction
system and beam potentials of up to 70 kV.
It is pertinent at this point to refer to earlier information from the controlled thermonuclear
reaction (CTR) field to confirm that the calculations presented above have some foundation in fact.
Of course, these CTR devices, listed in Table 4, do not use Xe propellant, since only hydrogen is of
interest in this field. However the ion optics principles remain the same. A further difference lies
in the operating times, since CTR injection machines are required to produce high energy ions for
only a few seconds on each activation; 10 s appears to be one of the longest times mentioned in the
literature [Okumura, 1980]. Thus lifetime is not addressed, whereas it is of critical importance in
105
the space field. Finally, high efficiency plasma production processes are of no significance in CTR
machines, since waste heat is readily removed. This is not possible in space, so much more
attention must be paid to the minimisation of losses and to thermal design.
Authors Grid Size Grid Open Beam Beam Beam Beam Current
[Reference] (cm) Form Area Energy Current Power Diverg. Density
Ratio (%) (keV) (A) (kW) (deg) (mA/cm2)
Okumura 10 dia Flat 31 70 4-7 280-490 1.4 170-190
[131]
Ohara* 12 dia Flat 40 75 15 1125 0.6 133
[134]
Menon 18 dia Dished 51 35-65 7-20 245-1300 2 27-79
[135]
Martin** 40 × 18 Flat 40 80 60 4800 0.4 200
[130]
*Design Study ** Production devices from the UKAEA Culham Laboratory
Table IX Characteristics of CTR machine ion beam accelerators.
The individual power consumptions given in Table 4 are very high, rising to nearly 5 MW
for the 40 cm × 18 cm device [Martin, 1984] developed by the Culham Laboratory. With a beam
energy of 80 kV, the SI is exceptionally large at 400,000 s, but this confirms that MW operation is
entirely feasible. The thrust is 2.4 N and the power density 6670 W/cm2, which is well above all the
values quoted in Table XI. The equivalent value for the 12 cm diameter device is even larger, at
9947 W/cm2. Similarly, the thrust densities, derived using Equ 10, are 8.1 and 5.2 mN/ cm2, which
are the highest experimental values yet recorded, as far as is known. For comparison, the largest
published value for a gridded ion engine is for high SI operation of the T5 thruster [Martin, 1988],
at 0.9 mN/ cm2. The predictions in Table 3 are generally greater at 8.8 to 28 mN/cm2, owing to the
use of Xe propellant, rather than hydrogen.
Since the open area ratio usually achieved in ion thrusters is well in excess of the 30 to 50%
of these CTR devices, a much improved extraction efficiency should be expected. This will lead to
enhanced current, thrust and power densities compared to those indicated by Table 4. However, it
should be noted that the calculations which gave the data in Table 3 assumed an open area ratio
typical of present thrusters.
The beam divergence values in Table 4 are more than an order of magnitude below those
typical of existing thrusters [Fearn, 1993], and suggest that grid impingement by primary ions may
not be a problem. In addition, the variation of 0.4 to 2 deg indicates that further benefits may
accrue from modelling work aimed at determining the best design solution. It should be noted that
there has been considerable progress in the ion optics modelling field since the devices listed in
Table 4 were designed in the late 1970s and early 1980s.
Recent work by Wilbur [Wilbur, 2004] has confirmed the viability of high voltage operation
in a thruster environment, using a twin-grid configuration. With a screen grid aperture diameter of
29 mm, the results with argon suggest that a beamlet energy of 2.3 kW with a 30 kV acceleration
potential should be achieved. The thrust from a single beamlet is then 12 mN, the SI is 27,000 s,
and the grid lifetime was estimated to be 30,000 hours. The power and thrust densities were 350
W/cm2 and 1.8 mN/cm2, respectively. These are remarkable results, bearing in mind the limitations
of the twin-grid extraction system.
It can thus be concluded that MW operation of a single ion thruster is possible, provided that
the SI can be permitted to exceed 10,000 s using Xe propellant. Of course, if a heavier atom is
utilised, such as Hg, the requirements can be met at lower values of SI. It is also worth noting that
the operation of two different EP technologies with different values of SI can enable any
106
intermediate performance to be attained. This concept has recently been studied for a combination
of gridded thrusters and Hall-effect thrusters [Chesta, 2003].
3.6.3.1. Specific Designs at the MW Level

In order to be of some use in actual mission analyses, three design studies were undertaken
at the 0.5, 1 and 1.5 MW power levels, assuming the thruster beam diameter to be 40 cm in each
case. Arbitrary beam accelerating potentials of 6, 18 and 12 kV, respectively, were assumed, as
examples of what can be achieved. The analysis was identical to that providing the data presented
in Table 3, but proceeded further into the details of the three thrusters and their operating
parameters. The results are given below in Table 5, in which the mass has been estimated from
current large devices. It is clear from this example that the aim of consuming 1 MW or more at
high SI can be achieved.
Parameter Power Input/Thruster (kW)

500 1500 1500
Propellant Xe Xe Xe
Grid diameter (cm) 40 40 40
No of grids (carbon) 4 4 4
Total propellant utilisation efficiency (%) 90 90 90
(c)
Total SI (s) (c) 8617 14,925 12,187
Thrust (N) 10.3 18.2 22.2
Ion beam current (A) 80.3 (a) 82.3 (a) 122.7 (b)
Beam acceleration (screen grid) potential 6000 18,000 12,000
(V)
Extraction grid potential (V) 1000 13,000 7000
Accel grid potential (V) -500 -500 -500
Decel grid potential (V) -50 -50 -50
Total propellant flow rate (mg/s) (c) 121.4 124.4 185.4
Total efficiency (%)(c) 86.7 88.9 88.3
Power/thrust (W/mN) 48.7 82.4 67.7
Thruster mass (kg) 32 35 35
Table X High power thruster design parameters.
Notes:
a. Assumes grid system operating at 50% of maximum perveance.
b. Assumes grid system operating at 75% of maximum perveance.
c. Includes allowances for neutraliser power and propellant flow rate.
3.6.3.2. Peak Thruster Performance

In the context of this study, an attempt was made to establish the maximum performance
which might be extracted from a given thruster, using a propellant with a low average atomic mass
of about 4.5 AMU. A 22 cm beam diameter Kaufman-type design was selected, so that a
comparison could be made with existing devices of this size, such as the T6 and RIT-XT (see Table
XI). The aim was to extend the SI as far as possible by utilising the very high values VB of noted in
107
Table 4, bearing in mind that these are, in general, measured parameters from actual laboratory
devices. An arbitrary value of 70 kV was adopted. The main limitation was assumed to be the
temperature of the thruster, since the insulation of the windings of the solenoids is currently unable
to exceed about 600° C; thus 500° C was selected as the maximum temperature.
In parallel, a similar exercise was conducted for the 10 cm diameter, equivalent to the T5 and
RIT-10 Evo thrusters (see Table XI). No upper temperature was assumed in this case; it was
calculated from the discharge power and an estimate of the mean emissivity of the discharge
chamber and other external materials. The results for both cases are presented below in Table 6.
Thruster beam diameter (cm) 10 22

Discharge chamber temperature (deg 790 500
C)
Ion extraction potential (kV) 5 5
Net beam accelerating potential (kV) 70 70
Beam current density (mA/cm2) 150 34
Anode current (A) 77 85
Beam current (A) 11.8 13
Perveance factor, Pg 4550 580
Thrust (N) 0.95 1.05
Input power (kW) 828 913
Total propellant flow rate (mg/s) 0.65 0.714
Discharge power (kW) 2.7 3.0
Beam divergence (deg) < 2.0 < 2.0
Exhaust velocity (m/s) 1.7 × 106 1.7 × 106
Propellant utilisation efficiency (%) 85 85
Electrical efficiency (%) 99.7 99.7
Specific impulse (s) 150,000 150,000
Total impulse in 10 000 hours (N/s) 3.4 × 107 3.8 × 107
Table XI Peak performances of 10 and 22 cm beam diameter thrusters.
It can be seen that both devices approach the MW level, despite their small size. However, it
is not suggested that thrusters as small as this should be developed to met the requirements of this
study, or that accelerating potentials as large as 70 kV would normally be employed. This exercise
merely puts the previous estimates into context, and shows that they are probably reasonable and
will eventually be fully viable.
It should be noted that the high discharge chamber temperatures are not of any great concern
for Kaufman-type DC thrusters, the RIT-type RF thrusters or the ECR devices, which do not utilise
permanent magnets. In each case, the limit is generally set by the insulation on electrical wiring,
either that required for solenoids or for RF field generation. However, the MESC thruster employs
permanent magnets (see Figure 38) which produce fields of the order of several kGauss. These
cannot withstand temperatures much higher than about 300° C, so the ability of this technology to
perform at these very high power levels is extremely doubtful. In this context, it will be observed
that the high energy CTR devices avoid permanent magnets for this reason.
3.6.3.3. Propellant Selection

Although Xe is the preferred propellant for most of the present applications of interest,
others are of relevance, especially as the supply of Xe is limited and its price will increase as
108
demand rises. In general, a propellant should be selected according to the SI required and the
available power level. As a general rule, and according to Equ 9, a high SI requires a low atomic
mass, and a large thrust at a given power level benefits from a much increased atomic mass.
Any material which can be vaporised and which does not react chemically with the thruster
or spacecraft constructional materials can be used. If chemical reactions occur with the cathode(s),
a separate inert propellant can be employed to feed these components, although this dual propellant
supply represents an additional complexity which would best be avoided. Cryogenic or high
pressure storage of gases is also to be avoided, since both incur a large mass penalties.
The ideal propellant is stored as a liquid at spacecraft temperature, to achieve a good storage
density under a low pressure. Compounds and mixtures are acceptable in principle. Such a liquid
would be fed to the thrusters via vaporisers, which would operate ideally at about 300 deg C.
Dissociation of compounds would occur in the discharge, and possibly in the vaporiser. Using this
principle, the average atomic mass can be reduced to about 4.5 AMU, as indicated below in Table
XII, if very high values of SI are required.
Compound Formula Melting Boiling Point Mean

Point (deg C) (deg C) Atomic Mass
(AMU)
Cyclohexane C6H12 6.5 80.7 4.7
Diethylamine C4H11N -50 55.5 4.6
Heptane C7H16 -90.6 98.4 4.3
Hexane C6H14 -95.3 68.7 4.3
Octane C8H18 -56.8 125.7 4.4
Table XII Possible low average atomic mass propellants.
3.6.3.4.Parameter Ranges
It was thought to be a useful exercise to attempt to suggest likely ranges for parameters of
interest to this study. These are listed below in Table XIII. They are derived from participation in a
number of other studies relevant to the present one, but must be regarded as tentative only.
• The major parameters which determine the characteristics of most missions are as follows:
• The power available from the reactor (or solar array), and possibly its change with time.
• The velocity increment and the total impulse required.
• The total time allowed for the mission.
• The thrusting time allowed.
• The thrust required to ensure that the mission can be accomplished in the above time.
• The maximum propellant load, which is usually determined by a trade-off between
numerous other parameters, and will often result in the selection of an optimum SI.
Consideration of these parameters will lead to a need to set values for many others, some of
which are tabulated below. If those requirements are then outside the ranges indicated, a review of
the complete design process will be necessary, since a conservative approach is preferable. For
example, if it is found that the thruster required will have a diameter of, say, 50 cm, this should be
regarded as impractical and the analysis should be repeated to determine whether a smaller
dimension might be acceptable. Similarly, a very high perveance should be discounted, since it
could lead to an impractically small spacing between the screen and accel grids, which would not be
maintained in the high temperature operating environment.
109
Parameter Lower Upper Comments
Limit Limit
Beam diameter (cm) ~3 ~ 40 Upper limit due to grid
distortion
Beam accelerating potential 10 70,000 Values > 5 kV need 4 grids
(V)
Beam current density - 150 Determined by perveance
(mA/cm2)
Grid perveance factor, Pg - 4500 Limit determined by T5
thruster data
Thrust density (mN/cm2) - 30 Determined by perveance
Power density (W/cm2) - 4500
Specific impulse (s) 2000 150,000 Determined by max voltage,
assumed to be 70 kV, and ion
mass
Mean ion mass (AMU) ~ 4.5 ~ 200 Minimum value is hydrogen
compound, maximum is
mercury
Thruster temperature for - ~ 800 Determined by solenoid and
Kaufman-type, RIT-type and RF coils, also wiring
ECR thrusters (deg C)
Thruster temperature for ~ 300 Determined by permanent
MESC thrusters (deg C) magnets
Table XIII Possible parameter limits.
3.6.4. Systems Considerations

In the context of this study, a number of systems considerations require attention,
particularly owing to the very high power levels envisaged. The first of these is the in-rush current
on applying the ion accelerating potentials to the grids of the thruster and, indeed, on first turning
on the discharge. In the latter case it is simple to ramp up the discharge current in a completely
controlled manner, selecting the speed at which this is done to match the capabilities of the power
generation system. This applies no matter what form of ionisation mechanism is selected. In the
case of the grids, this process also determines the rate at which ions are able to be extracted, and
therefore totally controls the way in which the beam current increases. Again, this can be matched
exactly to the power system and transient effects can be avoided without difficulty.
In many missions a significant throttling capability will be required, in order to match the
power consumption of the thrusters to the output of the power system, or to meet specific trajectory
requirements. In the case of an array of thrusters, this capability also provides a thrust vectoring
opportunity which can be of considerable value. This throttling capability is readily available with
all types of gridded thrusters, unless they have been designed to exclude it, as is the case with the
XIPS-13 [Beattie, 1993], for example. The best performance is available from the Kaufman-type
device using solenoids to produce the discharge chamber magnetic field and from the RIT-type of
RF thruster. A throttling ratio of 200:1 has been recorded with the former technology [Mundy,
1997], but with certain lifetime implications; 10:1 is easily achievable with both concepts.
Lifetime is clearly of concern for missions requiring a very large total impulse. To illustrate
overall capability, several thrusters have achieved very long operating durations in test facilities and
in space. The most encouraging example is that of DS-1, which reached in excess of 30,000 hours
110
in the laboratory and 16,000 hours in space [Sengupta, 2004]. The Muses-C thrusters were tested to
18,000 hours [Kuninaka, 2000] and the operation of the Artemis version of the RIT-10 at ESTEC
passed the 20,000 hours mark [Tartz, 2003]. These examples show that the technology has an
inherently long-life capability, but it is extremely costly to demonstrate this. Lifetime is mainly
limited by sputtering damage to the discharge chamber and grids, and this will be more severe at
high power, high plasma density and high thrust density. However, this process can be controlled
by design, the selection of materials, and the careful choice of the operating point. Unfortunately,
modelling is not, in general, sufficiently accurate to be used as a substitute for life-testing, so the
cost of this activity cannot at present be avoided.
A recommendation is therefore that modelling capabilities be improved so that they can be
used to predict lifetime with greater confidence than at present. In association with this advance,
more effort needs to be deployed to measure the fundamental physical data that are needed to make
realistic modelling predictions; these include sputtering yields and charge-exchange cross sections.
However, no matter what progress is made in this field, life-testing will remain necessary. To
circumvent the cost of this, mission designers should also consider the possibility of accepting a
shorter guaranteed life and flying more thrusters as a compensation, in order to achieve the required
total impulse. A study of this option might well suggest that it provides significant financial
benefits when dealing with very large, high power systems.
Since MW-level thrusters will require several kWs of discharge power (see Table I), and this
must be radiated to space, the correct thermal design of the installation will be an important
consideration. In general, the thrusters must be mounted externally and be thermally isolated from
the spacecraft. Provided that this is accomplished, the only additional thermal effects will concern
the removal of the waste heat from any power conditioning and control electronics. It is assumed
that this will be a relatively minor problem, since the bulk of the energy consumed will be supplied
directly from the source to the ion accelerating grids, without any additional power conditioning
circuitry; this option should be selected where possible in future design studies.
In the context of EP systems, the issue of spacecraft charging is often raised. It must
therefore be emphasised that the evidence from every flight that has taken place, using all types of
electric thruster, is that the existence of the exhaust plasma couples the spacecraft very effectively
to space potential. Thus there is no spacecraft charging effect whatsoever. Indeed, a neutralised ion
beam couples the spacecraft to space plasma potential usually to within 10 V. The residual concern
is that charging effects may occur if the neutraliser of the propulsion system fails. While this may
possibly be the case, a suitable sensing circuit will enable the EP system to be turned off sufficiently
rapidly to ensure that no damage occurs.
3.7. Conclusions
This report has reviewed the ways in which gridded ion thruster systems might cover the
power range from tens of kW to several MW, with the aim of performing a very wide variety of
challenging interplanetary missions utilising nuclear power sources. It has been shown that
reasonable extensions of current technologies should permit this to be accomplished.
By consideration of the relevant scaling relationships, which have been validated up to about
40 kW, it has been concluded that the same basic concepts will suffice for the entire range of power
levels and applications specified. This conclusion is certainly valid for the Kaufman-type of direct
current discharge thruster, for the RIT-type of radiofrequency ionisation device, and for ECR
ionisation thrusters which do not use permanent magnets. Unfortunately, the severe temperature
constraints on high field permanent magnets suggest that the cusp-field, MESC type of thruster
cannot operate at the very high power levels considered here, unless de-rated to very low thrust
density.
111
The design process is aided considerably by the separation of the ion production and
extraction/acceleration regions in gridded thrusters. Thus the ion beam parameters can be deduced
without reference to the ionisation mechanism employed to produce the plasma from which the ions
are extracted. Consequently, the two regions of the thruster can be designed separately, which is a
simplifying benefit only available to gridded devices. Similarly, the ion beam neutralisation process
can be treated independently.
It has been concluded that the required power density can be achieved and exceeded using
gridded ion thrusters, by operating their grid systems at very high perveance and by raising the
specific impulse to values which are significantly above those commonly employed at present. The
latter ensures that most of the energy supplied is used in accelerating the beam ions, thereby
allowing extremely high values of electrical efficiency to be achieved; in the limit, this parameter
exceeds 99%. With propellant utilisation efficiencies of above 85%, the overall thruster efficiency
becomes in excess of 84%.
For the applications at the lower end of the above power range, normal twin- or triple-grid
configurations may suffice, especially if higher perveance operation is adopted and if the specific
impulse can be increased above current values. However, the limit here is due to the penetration of
the plasma sheath at the screen grid into the discharge chamber plasma, which increases as the
electric field becomes larger or the plasma density lower. This greater curvature of the sheath,
which effectively emits the beam ions, then causes direct impingement onto the accel and decel
grids, severely limiting lifetime. With this configuration, the maximum extraction potential is likely
to be about 5 kV. If the accel potential is -500 V and if Xe is used, this gives an SI of
approximately 7000 s, depending upon the propellant utilisation efficiency achieved. About 100
kW power consumption can then be realised with a small array of thrusters.
The power range can be extended to several 100 kW by the further development of existing
thrusters, utilising modified grid systems and by operating at higher SI and/or perveance than
normal. While the latter can be achieved solely by making the accel grid more negative, a more
complex but satisfactory alternative is to employ a 4-grid system with both a greater perveance and
providing an increased SI. In all cases, additional research and development are needed, since this
regime has not been properly explored to date. The lifetime implications are of particular interest
and may dictate the way forward.
Thus the operational envelope can be massively extended by use of the 4-grid configuration.
There is considerable documentation concerning such systems in the CTR community, where ion
accelerators of the size required, or smaller, have been constructed which produce MW beams at 70
or 80 keV. This is possible because the 4-grid arrangement permits the ion extraction process to be
separated from the main acceleration region, and the limitation of the sheath penetration no longer
applies. Thus the extraction part of the system can be designed to operate at near maximum
perveance, and the subsequent further acceleration of the ions can be dealt with independently in the
design process. An additional advantage of this configuration is the very low beam divergence
typically found. This can be less than 1° at an acceleration potential of 70 kV.
It is thus concluded that well understood thruster technology, when combined with the 4-grid
configuration based on that utilised in CTR ion injection machines, will permit MW power levels to
be achieved. Thus a relatively small array of thrusters, with beam diameters not exceeding 40 cm,
will be able to consume many MWs, although the SI, using Xe propellant, is likely to be somewhat
above 10,000 s. With this arrangement, the power density can reach 4.5 kW/cm2 and the thrust
density 30 mN/cm2 and, if required, the specific impulse can attain 30,000 s with Xe. If higher
values of SI are required, the utilisation of lower atomic mass propellants will permit this to be
achieved, with an ultimate limit using hydrogen compounds of about 150,000 s.
As a specific example, an array of nine 40 cm beam diameter thrusters using Xe propellant
and operating at 10 kV, but with an ion extraction potential of 5 kV, will consume 7.4 MW if the
perveance is limited to 50 % of its maximum value. The total beam current will be 702 A, the
112
thrust 120N, the SI 11,100 s, and the power density 644 W/cm2. If full perveance operation can be
achieved while maintaining long life, the power consumption is 14.7 MW.
To summarise, the design for a 20 to 50 kW thruster can be extrapolated directly from
current capabilities, taking advantage of the higher perveance values made possible by introducing
improved grid technologies. However, at the higher power level it would be advantageous to utilise
a 4-grid configuration. Thus an array of certain existing devices could probably consume up to 100
kW or so, although none of these thrusters are space qualified and they all require life-testing. Thus
they are not yet ready for this application and considerable funding would be needed to reach
qualified status. The work required to meet a higher power level, utilising the 4-grid configuration,
would be much more extensive, so would require more time and greater funding.
A multi-MW system can be implemented using extended scaling relationships, but there is
no precedent for such an exercise. Thus this complete process, encompassing design, development,
performance testing, environmental testing and life-testing, would require to be organised and
funded. Bearing in mind the need to co-ordinate such a development with that of the power source,
work on both systems should be conducted in parallel, with very close collaboration at all times.
This co-operation is desirable because it is envisaged that the bulk of the power consumed by the
thrusters would be supplied directly from the source, with no further power conditioning.
For very high SI operation, the most significant challenge is to prove that the 4-grid ion
extraction and acceleration mechanism can be made to function successfully and to provide long
life. Although not simple, this objective is tenable, because of the prior experience of the CTR
community.
The scale of the endeavour is indicated by looking at the values of the most important
parameters included in Table 2, accepting the T6 values as representative of a thruster approaching
qualified status, and comparing them with the ultimate objectives in Table 3. The numbers are
somewhat daunting, but should be achievable, mainly because the additional power is to be fed
directly into the ion beam, and will not appear within the thruster. With Xe propellant, the aim is to
improve the SI by a factor of between 2 and 6.3. The corresponding increase in power density is by
a factor of between 24 and 240, and the thrust density will, as a consequence, increase by a factor of
between 14 and 46.
Although it may seem to be overly optimistic to predict that such extrapolations are feasible,
it should be recalled that this process is aided very considerably by three factors. The first is that
the separate processes occurring within a gridded thruster can be dealt with independently, to first
order. Thus the discharge chamber, the grid system, and the neutraliser can be designed and tested
independently, at least initially. The second factor is that the mathematical design processes
applicable to the grid system are well understood and the ion optics and erosion modelling
capabilities are excellent; these are critical to success. The third factor, mentioned above, is that the
huge increase in power contemplated in this study is fed directly into the ion beam.
Finally, it should be noted that the performance- and life-testing of large, high power
thrusters requires the use of major facilities with substantial pumping speeds, which are costly to
build and to operate. The provision of such facilities must precede the development of the thrusters.
Dr. D.G. Fearn

EP Solutions
23 Bowenhurst Road,
Church Crookham,
Fleet,
Hants,
GU52 6HS,
UK
113
(44)-(0)1252-615924
dg.fearn@virgin.net
3.8. References
Note:
AIAA is the American Institute of Aeronautics and Astronautics
IAF is the International Astronautical Federation
IEPC is the International Electric Propulsion Conference
ISU is the International Space University
Staehle, R L, Carraway, J B, Salvo, C G, Terrile, R J, Weinstein, S S and Hansen, E, “Exploration

of Pluto: search for applicable small satellite technology”, Proc 6th Annual AIAA/Utah State
University Conf on Small Satellites, Utah, 21-24 September 1992.
Rayman, M D, Varghese, P, Lehman, D H, and Livesay, L L, “Results from the Deep Space 1
technology validation mission”, Acta Astronautica, 47, 475-488, (2000).
de Cara, D M and Estublier, D, “SMART-1; an analysis of flight data”, IAF Paper IAC-04-S.4.02,
(October 2004).
Kawaguchi, J, Fujiwara, A and Uesugi, “The ion engines cruise operation and the Earth swingby of’
Hayabusa’ (Muses-C)”, IAF Paper IAC-04-Q.5.02, (October 2004).
Millis, M G, “NASA breakthrough propulsion physics program”, Acta Astronautica, 44, 2-4, 175-
182, (1999).
Arhipov, B, Vinogradov, V, Kozubsky, K, Kudriavtsev, S and Murashko, “Development and

application of electric thrusters at EDB “Fakel””, IAF Paper IAF-99-S.4.02, (October 1999).
Manzella, D H, Jankovsky, R S and Hofer, R R, “Laboratory model 50 kW Hall thruster”, AIAA
Paper 2002-3676, (July 2002).
Rodgers, D and Brophy, J, “Ion propulsion technology for fast missions to Pluto”, IEPC Paper 01-
176, (October 2001).
Henry, P K, Terrile, R J, Maddock, R W and Sobel, H R, “Exploring the Kuiper Belt: an extended
Pluto mission”, Acta Astronautica, 44, 2-4, 85-90, (1999).
Kluever, C A, “Heliospheric boundary exploration using ion propulsion spacecraft”, J Spacecraft

and Rockets, 34, 3, 365-371, (May-June 1997).
West, J L, “Design issues for a mission to exploit the gravitational lensing effect at 550 AU”, Acta
Astronautica, 44, 2-4, 99-107, (1999).
Fearn, D G, “The ultimate performance of gridded ion thrusters for interstellar missions”, Proc
STAIF-2000 Forum, Albuquerque, NM, 30 January-3 February, 2000.
114
Fearn, D G, “The application of gridded ion thrusters to high thrust, high specific impulse nuclear-
electric missions”, IAF Paper IAC-04-R.4/S.7.09, (October 2004).
Wilbur, P, “Limits on high specific impulse ion thruster operation”, AIAA Paper 2004-4106, (July
2004).
Martin, A R, Bond, A and Bond R A, “Ultimate performance limits and mission capabilities of
advanced ion thrusters”, IEPC Paper 88-084, (October 1988).
Hanley, G M, “Satellite power systems (SPS) concept definition study. Volume IV.
Transportation analysis”, NASA CR-3321, (September 1980).
Loeb, H W and Popov, G A, Eds, “Study Report. Advanced interplanetary missions using nuclear-
electric propulsion”, University of Giessen, Germany, (June 1995).
Reese, R T and Vick, C P, “Soviet nuclear powered satellites”, J Brit Interplan Soc, 36, 457-462,
(1983).
Polansky, G F, Schmidt, G L, Reynolds, E L, Schaefer, E D, Ogloblin, B and Bocharov, A, "A plan

to flight qualify a Russian space nuclear reactor for launch by the United States", AIAA Paper 93-
1788, (1993).
Cameron, G E and Herbert, G A, “System engineering of a nuclear electric propulsion testbed

spacecraft”, AIAA Paper 93-1789, (1993).
Truscello, V C, “SP-100 power system”, Proc 9th Symposium on “Space Nuclear Power Systems”,
Albuquerque, NM, (January 1992).
Racca, G D, Foing, B H and Rathsman, P, “An overview of the status of the SMART-1 mission”,
IAF Paper IAA-99-IAA.11.2.0, (October 1999).
Fearn, D G, "The use of ion thrusters for orbit raising", J Brit Interplan Soc, 33, 129-137, (1980).
Widman, F W, North, D M, Cross, E H and Burr, A D, “Evaluation of nuclear propulsion for solar
system exploration missions”, AIAA Paper 93-1952, (1993).
Novara, M, “The BepiColombo ESA Cornerstone mission to Mercury”, IAF Paper IAF-01-Q.2.02,
(October 2001).
Fearn D G, "The Ulysses mission: the ion propulsion alternative", J Brit Interplan Soc, 49, 181-192,
(1996).
Fearn, D G and Smith, P, “A review of UK ion propulsion - a maturing technology”, IAF Paper
IAF-98-S.4.01, (Sept/Oct 1998).
Wallace, N C, Mundy, D H and Fearn, D G, “A review of the development of the T6 ion thruster
system”, IAF Paper IAF-99-S.4.08, (October 1999).
Ozaki, T, Nishida, E, Gotoh, Y, Tsujihata, A and Kajiwara, K, “Development status of 20 mN

xenon ion thruster”, AIAA Paper 2000-3277, (July 2000).
Beattie, J R, Williams, J D and Robson, R R, “Flight qualification of an 18-mN xenon ion thruster”,
IEPC Paper 93-106, (1993).
115
Ocampo, C, “Geostationary orbit transfer using electric propulsion for the Hughes HS-702
satellite”, IAF Paper IAF-98-A.1.08, (September/October 1998).
Groh, K H, Leiter, H J and Loeb, H W, “Design and performance of the new RF-ion thruster RIT
15”, AIAA Paper 98-3344, (1998).
Leiter, H, Bassner, H, Frölich, T, Kukies, R, Müller, H and Killinger, R, “Development and
performance of the advanced radio frequency ion thruster RIT_XT”, IEPC Paper 2003-115, (March
2003).
Pessana, M, Noci, G, Matucci, A and Ricciardi, A, “Low thrust ion propulsion for orbit
maintenance and drag-free control of the Gravity Explorer Satellite”, Proc Second European
Spacecraft Propulsion Conf, Noordwijk, Holland, 27-29 May 1997. ESA SP-398..
Fearn, D G, “Ion thruster lifetime limitations imposed by sputtering processes”, IEPC Paper 93-177,
(1993).
Notarantonio, A, Killinger, R and Amorosi, L, “Ion propulsion system saves Artemis satellite”, IAF
Paper IAC-03-S.4.01, (September/October 2003).
Kerslake, W R, Goldman, R G and Nieberding, W C, “SERT II: mission, thruster performance, and
in-flight thrust measurements”, J Spacecraft and Rockets, 8, 3, 213-224, (March 1971).
Shimada, S et al (8 authors), “Ion engine system development of ETS-VI”, IEPC Paper 91-145,
(October 1991).
Kim, V, "Main physical features and processes determining the performance of stationary plasma
thrusters", J Propulsion and Power, 14, 5, 736-743, (1998).
Gorshkov, O A et al, “Activity of Russian Space Agency in the field of electric propulsion”, AIAA
Paper 98-3176, (1998).
Bober, A et al, "Development and application of electric propulsion thrusters in Russia", IEPC
Paper 93-001, (1993).
Cadiou, A, “An overview of the CNES electric propulsion programme”, IEPC Paper 2003-169,
(March 2003).
Jacobson, D T, Manzella, D H, Hofer, R R and Peterson, P Y, “NASA’s 2004 Hall thruster

program”, AIAA Paper 2004-3600, (July 2004).
Garnero, P, “Astra 1K and Stentor plasma propulsion subsystem experience”, AIAA Paper 2003-
4547, (July 2003).
Dumazert, P, Marchandise, F, Prioul, M, Darnon, F and Jolivet, L, “PPS 1350-G qualification

status”, IEPC Paper 03-270, (March 2003).
Gorshkov, O A and Koroteev, A S, “New trends in Hall-effect thruster development”, Proc 6th
International Symposium on “Propulsion for space transportation of the XXIst Century”, Versailles,
13-17 May 2002.
116
Randolph, T, Fischer, G, Kahn, J, Kaufman, H, Zhurin, V, Kozubsky, K and Kim, V, “The
mitigation of discharge oscillations in the stationary plasma thruster”, AIAA Paper 94-2857, (June
1994).
Jolivet, L, Secheresse, O, Kharchevnikov, V K, Bugrova, A and Morozov, A I, “Development and

testing of a two-stage Hall effect thruster”, IEPC Paper 2003-290, (March 2003).
Prioul, M et al (17 authors), “Double stage HET development activities at SNECMA”, Session
4SE4, Proc 4th International Spacecraft Propulsion Conference, Sardinia, 2-4 June 2004.
Sankovic, J M, Haag, T W and Manzella, D H, “Operating characteristics of the Russian D-55

thruster with anode layer”, AIAA Paper 94-3011, (June 1994).
Semenkin, A et al (8 authors), “Development program and preliminary test results of the TAL-110
thruster”, AIAA Paper 99-2279, (June 1999).
Andrenucci, M, Paganucci, F, Frazzetta, M, La Motta, G and Schianchi, G, “Scale effects on the

performance of MPD thrusters”, IEPC Paper 01-123, (October 2001).
Böhrk, H, Laure, S and Auweter-Kurtz, M, “Feasibility study of application of ATTILLA to in-

space propulsion for piloted missions”, IAF Paper IAC-04-S.4.03, (October 2004).
Fearn, D G and Cohen, R B, "The future development of non-chemical propulsion", IAF Paper
IAA-98-IAA.3.3.05, (1998).
Loeb, H W, “Recent works on radio frequency ion thrusters”, AIAA Paper 70-1102, (August-
September 1970).
Groh, K H and Loeb, H W, “State-of-the-art of radio-frequency ion thrusters”, AIAA Paper 89-
2381, (July 1989).
Leiter, H J, Loeb, H Wand Schartner, K-H, "The RIT-15 ion engines -a survey of the present state
of radio frequency ion thruster technology and its future potentiality", Proc Third International
ESA/CNES Conference on Spacecraft Propulsion, Cannes, 10-13 October 2000; pp 423-432, ESA
SP-465, (December 2000).
Bassner, H, Berg, H-P and Kukies, R, “RITA development and fabrication for the Artemis
satellite”, IEPC Paper 91-057, (October 1991).
Bassner, H et al, “Flight test and ground investigation results of the RITA experiment on Eureca”,
AIAA Paper 94-2848, (June 1994).
Leiter, H J and Killinger, R, “Radio frequency ion thruster RIT for space propulsion”, Proc ISU
Short Course on Space propulsion, Versailles, (13 May 2002).
Toki, K, “Flight readiness of the microwave ion engine system for Muses-C mission”, IEPC Paper
03-098, (March 2003).
Kawaguchi, J & Uesugi, T, “Technology development status of the Muses-C sample and return
project”, IAF paper IAF-99-IAA.11.2.02, (October 1999).
117
Fearn, D G and Patterson, S W, “Characterisation of the high current hollow cathode for the T6 ion
thruster”, AIAA Paper 98-3346, (July 1998).
Capacci, M, Matticari, G, Noci, G E, Severi, A, Ricciardi, A and Svelto, F, "Radiofrequency with

magnetic field ion thruster (RMT): review of the engineering phase accomplished under ASI
contract”, IEPC Paper 03-289, (March 2003).
Kaufman, H R, "An ion rocket with an electron bombardment ion source", NASA TN-585, (1961).
Edwards, C H, Mundy, D H and Clark, S D, “The development of gas flow control equipment at
DERA”, IAF Paper IAF-99-S.4.09, (October 1999).
Fearn, D G, “The possible application of ion propulsion to precursor interstellar missions”, AIAA
Paper 2000-3415, (July 2000).
Beattie, J R and Robson, R R, "18-mN xenon ion propulsion subsystem", IEPC Paper 91- 010,
(October 1991).
Beattie, J R and Wilbur, P J, “Cusped magnetic field mercury ion thruster”, AIAA Paper 76-1011,
(November 1976).
Christensen, J A, Benson, G, Bond, T A, Gallagher, J T and Matranga, M, “The NSTAR ion

propulsion subsystem for DS1”, AIAA Paper 99-2972, (June 1999).
Brophy, J R, Brinza, D E, Polk, J E, Henry, M D and Sengupta, A, “The DS1 hyper-extended

mission”, AIAA Paper 2002-3673, (July 2002).
Anzel, B, “Stationkeeping the Hughes HS 702 satellite with a xenon ion propulsion system”, IAF
Paper IAF-98-A.1.09, (Sept/Oct 1998).
Latham, P M, Martin, A R and Bond, A, "Design, manufacture and performance of the UK- 25
engineering model thruster", AIAA Paper 90-2541, (1990).
Gray, H L, “Development of ion propulsion systems”, GEC Rev, 12, 3, 154-168, (1997).
Oppenhaeuser, G, van Holtz, L and Bird, A, “The Artemis mission – ESA’s latest communication
satellite”, IAF Paper IAF-01-M.1.08, (October 2001).
Fearn, D G, Martin, A R and Smith, P “Ion propulsion development in the UK”, IAF Paper IAF-93-
S.5.490, (October 1993).
Harbour, P J, et al, “Physical processes affecting the design and performance of ion thrusters with
particular reference to the 10 cm, RAE/Culham T4 thruster,” AIAA Paper 73-1112, (1973).
Fearn, D.G., "The UK-10 ion propulsion system: a technology for improving the cost effectiveness
of communications spacecraft", IEPC Paper 91-009, (1991).
Bythrow, P F, Mauk, B H and Gatsonis, N A, "Operational experiments and thruster performance

plan for the nuclear electric propulsion space test program (NEPSTP)", AIAA Paper 93-1896, (June
1993).
Bassner, H, Killinger, R, Marx, M, Kukies, R, Aguirre, M, Edwards, C and Harman, H-P, "Ion
propulsion for drag compensation of GOCE", AIAA Paper 2000-3417, (July 2000).
118
Mundy, D H and Fearn, D G, "Throttling the T5 ion engine over a wide thrust range", AIAA Paper
97-3196, (1997).
Martin, A R and Latham, P M, “High thrust operation of the UK-10 rare gas ion thruster (T4A)”,
IEPC Paper 88-062, (October 1988).
Freisinger, J, Reinek, S, Rumpf, E and Loeb, H W, “RIT 15 – laboratory prototype of a 20 mN
NSSK engine”, AIAA Paper 76-1038, (1976).
Leiter, H J, Loeb, H W and Schartner, K-H, "RIT15S and RIT15LP – the development of high
performance mission optimised ion thrusters”, AIAA Paper 99-2444, (June 1999).
Kerslake, W R, "SERT -II thrusters -still ticking after eleven years", AIAA Paper 81-1539, (July
1981).
Leiter, H J, Killinger, R, Bassner, H, Müller, J, Kukies, R and Fröhlich, T, “Evaluation of the

performance of the advanced 200 mN radio frequency ion thruster RIT-XT”, AIAA Paper 2002-
3836, (July 2002).
Killinger, R, Bassner and H, Müller, J, “Development of an high performance RF-ion thruster”,

AIAA Paper 2000-2445, (July 2000).
Leiter, H J, Killinger, R, Müller, J, Kukies, R, Frölich, T, “Extended performance evaluation of

EADS ST’s 200 mN radio frequency ion thruster”, AIAA Paper 2003-5010, (July 2003).
Bassner, H, Bond, R, Thompson, V, Harmann, H-P and Groh, K, "Development and performance
testing of the ESA-XX ion thruster", AIAA Paper 99-2441, (June 1999).
Huddlestone, J et al, “An overview of the T6 gridded ion propulsion system pre-development
activities for Alpha-Bus”, Proc 4th International Spacecraft Propulsion Conference, Sardinia, 2-4
June 2004.
Pollard, J E and Welle, R P, “Thrust vector measurements with the T5 ion engine”, AIAA Paper 95-
2829, (1995).
Mundy, D H, “Factors affecting the beam divergence of a T5 ion engine”, IEPC Paper 97-095,
(August 1997).
Lsyzyk, M, “Propulsion on @Bus platform”, IEPC Paper 03-170, (March 2003).
Beattie, J R, Matossian, J N, Poeschel, R L, Rogers, W P and Martinelli, R, "Xenon ion propulsion

subsystem", AIAA Paper 85-2012, (September/October 1985).
Melita, O (ed), “The comet nucleus sample return mission”, ESA SP-249, (December 1986).
Wells, A A, Harrison, M F A, White, B M and Harbour, P J, “Laws for scaling electron

bombardment, thrusters”, Proc Conf on “Electric propulsion of space vehicles”, Culham
Laboratory, UK, (April 1973). IEE Publication 100, pp 250-257, (1973).
Beattie, J R and Matossian, J N, “High-power xenon ion thrusters”, AIAA Paper 90-2540, (July
1990).
119
Sengupta, A, Brophy, J, Anderson, J, Garner, C, Banks, B and Sovey, J, “An overview of the results
from the 30,000 hr life test of Deep Space 1 flight spare ion engine”, AIAA Paper 2004-3608, (July
2004).
Soulas, G C, “Improving the total impulse of the NSTAR ion thruster with thick-accelerator-grid
ion optics”, IEPC Paper 01-081, (October 2001).
Patterson, M J and Rawlin, V K, "Performance of 10-kW class xenon ion thrusters", AIAA Paper
88-2914, (July 1988).
Groh, K H et al, “RIT 35 – design and application of the Giessen primary propulsion engine”,
AIAA Paper 79-2053, (1979).
Groh, K H, Loeb, H W, Mueller, J, Schmidt, W and Schuetz, B, “RIT 35 RF-ion thruster – design
and performance, J Brit Interplan Soc, 41, 209-214, (1988).
Hayakawa, Y, Miyazaki, K, Kitamura, S, Yoshida, H, Yamamoto, Y and Akai, K, "Endurance test

of a 35-cm xenon ion thruster", AIAA Paper 3530, (July 2000).
Tahara, H and Nishida, M, "Overview of electric propulsion activity in Japan", AIAA Paper 99-
2159, (June 1999).
Komurasaki, K, Arakawa, Y and Takegahara, H, "An overview of electric and advanced propulsion
activities in Japan", Proc Third International ESA/CNES Conference on Spacecraft Propulsion,
Cannes, 10-13 October 2000; pp 27-39, ESA SP-465, (December 2000).
Patterson, M J, Foster, J E, Haag, T W, Rawlin, V K, Soulas, G C and Roman, R F, “NEXT:

NASA’s evolutionary xenon thruster”, AIAA Paper 2002-3832, (July 2002).
Patterson, M, “Engine development status for NEXT: NASA’s evolutionary xenon thruster”, IEPC
Paper 2003-163, (March 2003).
Soulas, G C, Haag, T W and Patterson, M J, “Performance evaluation of 40 cm ion optics for the
NEXT ion engine”, AIAA Paper 2002-3834, (July 2002).
Hoskins, W A et al (7 authors), “Development of a prototype model ion thruster for the NEXT
system”, AIAA Paper 2004-4111, (July 2004).
Soulas, G, Kamhawi, H, Patterson, M and Gardner, M, “NEXT ion engine 2000h wear test results”,
AIAA Paper 2004-3791, (July 2004).
Randolph, T M and Polk, J E, “An overview of the nuclear electric xenon ion system (NEXIS)
activity”, AIAA Paper 2004-3450, (July 2004).
Prockter, L et al, “The Jupiter Icy Moons Orbiter: an opportunity for unprecedented exploration of
the Galilean satellites”, IAF Paper IAC-04-IAA.3.6.4.02, (October 2004).
Oleson, S R, “Electric propulsion technology development for the Jupiter Icy Moons Orbiter
project”, AIAA Paper 2004-3449, (July 2004).
120
Elliott, F W, Foster, J E and Patterson, M J, “An overview of the high power electric propulsion
(HiPEP) project”, AIAA Paper 2004-3453, (July 2004).
Williams, G W et al, “Performance characterisation of HiPEP ion optics”, AIAA Paper 2004-3627,
(July 2004).
Foster, J E et al (9 authors), “The high power electric propulsion (HiPEP) ion thruster”, AIAA
Paper 2004-3812, (July 2004).
Nakanishi, S and Pawlik, E V, “Experimental investigation of a 1.5-meter diameter Kaufman

thrustor”, AIAA Paper 67-725, (September 1967).
Patterson, M J, “50 cm ion thruster research”, AIAA Paper 87-1034, (May 1987).
Patterson, M J, Roman, R F and Foster, J E, “Ion engine development for interstellar precursor
missions”, AIAA Paper 2000-3811, (July 2000).
Wallace, N C, Space Department, QinetiQ, Farnborough, UK, private communication, (June 2003).
Beattie, J R and Matossian, J N, “High-power xenon ion thrusters”, AIAA Paper 90-2540, (July
1990).
Aston, G, Brophy, J R, Garner, C E, Pless, L C, Owens, A G and Toomath, R L, “Operating

characteristics of a 10 kW xenon ion propulsion module”, AIAA Paper 87-1006, (May 1987).
Masek, T D, Poeschel, R L, Collett, C R and Schnelker, D E, “Evolution and status of the 30-cm
engineering model ion thruster”, AIAA Paper 76-1006, (November 1976).
Kitamura, S, Hayakawa, Y, Miyazaki, K, Yoshida, H, Kajiwara, K and Akai, K, “Research and

development of a 150-mN xenon ion thruster”, IEPC Paper 2003-165, (March 2003).
Soulas, G C, “Design and performance of 40 cm ion optics”, IEPC Paper 01-090, (October 2001).
Wilbur, P, “Limits on high specific impulse ion thruster operation”, AIAA Paper 2004-4106, (July
2004).
Fearn, D G, “The manned exploration of Mars using nuclear electric propulsion: a gradual,
minimum cost strategy”, Session 4NM2, Proc 4th International Spacecraft Propulsion Conference,
Sardinia, 2-4 June 2004.
Christensen, J, “Boeing EDD electric propulsion programs overview”, AIAA Paper 2004-3967,
(July 2004).
Bond, R A, Fearn, D G, Wallace, N C and Mundy, D H, “The optimisation of the UK-10 ion
thruster extraction grid system”, IEPC Paper 97-138, (1997).
Martin, A R, “High power beams for neutral injection heating”, Vacuum, 34, 1-2, 17-24, (1984).
Okumura, Y, et al, “Quasi-dc extraction of 70 keV, 5 A ion beam,” Rev Sci Instrum, 51, 728-734,
(1980).
Leiter, H et al, “RIT-22 – the ion thruster of the RITA propulsion system for commercial and
scientific applications”, Proc 4th International Spacecraft Propulsion Conference, Sardinia, 2-4 June
2004.
121
Fearn, D G, “The design of a versatile ion thruster for large communications spacecraft”, IAF Paper
IAF-97-S.3.01, (October 1997).
Ohara, Y, “Numerical simulation for design of a two-stage acceleration system in a megawatt

power ion source,” J Appl Phys, 49, 4711-4717, (1978).
Menon, M M, et al, “Power transmission characteristics of a two-stage multiaperture neutral beam

source,” Rev Sci Instrum, 51, 1163-1167, (1980).
Chesta, E, et al (22 authors), “Flexible variable-specific-impulse electric propulsion systems for

planetary missions”, Proc 5th Internal Conf on “Low-cost planetary missions”, ESTEC, 24-26
September, 2003; ESA SP-542.
Kuninaka, H, Funaki, I, Nishiyama, K, Shimizu, Y and Toki, K, “Result of 18,000-hour endurance

test on microwave discharge ion thruster”, AIAA Paper 2000-3276, (July 2000).
Tartz, M, Hartmann, E and Neumann, H, “Study of extraction grid erosion in a long-time ion
thruster operational test”, IEPC Paper 03-176, (March 2003).
Baker, D N, “Critical issues in space plasma physics”, Phys of Plasmas, 6, 5, 1700-1707, (May
1999).
Barne, S J, McComas, D J, Young, D T and Belian, R D, “Diagnostics of space plasmas”, Rev Sci
Instrum, 57, 8, 1711-1716, (August 1986).
Wang, J, Brinza, D and Young, M, “Modeling ion propulsion plasma environment for Deep Space
1”, AIAA Paper 2000-3528, (July 2000).
Chang Diaz, F R, “Research status of the Variable Specific impulse Magnetoplasma Rocket”,
Bulletin of APS, 42, 2057, (1997).
Batishchev, O and Molvig, K, “VASIMR thruster operation with hydrogen and helium gas
propellants”, IEPC Paper 2003-191, (March 2003).
Glasstone, S and Lovberg, R H, “Controlled thermonuclear reactions”, van Nostrand, (1960).
Glover, T W et al (7 authors), “Ion cyclotron heating results in the VASIMR VX-10”, AIAA Paper
2004-3639, (July 2004).
Wallace, N C and Gray, H, “Artemis Anomaly Investigations and Lessons Learnt”, Proc European
Workshop on Electric Propulsion Flight Experiences, Villa Marigola, Lerici, Italy, 1 to 2
September 2005.
Chein, K-R et al (6 authors), “L-3 Communications ETI electric propulsion overview”, IEPC Paper
2005-315, (October/November 2005).
122
3.9. List of Symbols and Acronyms
Acronyms
AM Atomic mass
AU Astronomical unit
CTR Controlled thermonuclear reaction
DC Direct current
DS-1 Deep Space 1 (mission)
ECR Electron cyclotron resonance
EP Electric propulsion
EPS Electric propulsion system
FCU Flow control unit
GIE Gridded ion engine
GTO Geostationary transfer orbit
HET Hall-effect thruster
HiPEP High power electric propulsion
IPS Ion propulsion system
JPL Jet Propulsion Laboratory
LEO Low Earth orbit
MESC Magneto-electrostatic containment
MMS Matra Marconi Space (Ltd)
MPD Magnetoplasmadynamic
NEP Nuclear electric propulsion
NEPSTP Nuclear Electric Space Test Program
NEXISNuclear electric xenon ion system
NEXT NASA evolutionary xenon thruster
NSSK North-south station-keeping
NTP Nuclear thermal propulsion
PCU Power conditioning unit
RF Radio frequency
RIT Radiofrequency ionisation thruster
RMT Radiofrequency with magnetic field thruster
SI Specific impulse
SPT Stationary plasma thruster
TAL Thruster with anode layer
VHF Very high frequency
XIPS Xenon ion propulsion system
accel Accelerator grid

decel Decelerator grid
Symbols
A Active grid area

Bθ Azimuthal magnetic field around MPD thruster’s axial discharge current
E Electric field
123
Ia Accel grid current in RMT
IB Ion beam current
Ib Ion beam current in RMT
Ic Coil current in RMT
Id Decel grid current in RMT
Is Screen grid current in RMT
Isp Specific impulse
JB Beam current density
Mf Spacecraft mass at the end of a manoeuvre
Mo Spacecraft mass at the beginning of a manoeuvre
Ms Mass delivered to Pluto orbit
Pg Perveance parameter
P Total power fed into a thruster
Pd Power density (per unit grid area)
PT Total power supplied to the thruster
R Grid radius
T Thrust
Td Thrust density
Tg Screen grid transparency
Va Accel grid potential in RMT
Vac Accel grid potential
VB Ion beam accelerating potential
Vdec Decel grid potential
Vs Screen grid (beam) potential in RMT
VT Total ion accelerating potential
WRF RF power supplied to RMT
d True acceleration length for ions in the grid system

e Charge on an electron
go Acceleration due to gravity at sea level
jr Radial current density in MPD thruster
m& Propellant flow rate (also to discharge chamber of RMT)
m& t Total propellant flow rate
mi Ion mass
n Plasma number density
ve Exhaust velocity of beam ions
veff Effective exhaust velocity
∆M Propellant mass required for a specified manoeuvre

∆V Mission velocity increment
εo Dielectric constant of free space

ηe Thruster electrical efficiency
ηm Uncorrected propellant utilisation efficiency
ηmt Total propellant utilisation efficiency
ηt Total thruster efficiency
124
4. HIGH POWER AND HIGH THRUST DENSITY ELECTRIC PROPULSION FOR IN-
SPACE TRANSPORTATION
4.1. Abstract
Mission studies for a manned Mars mission clearly show that with chemical propulsion round trip
times are too long for the safety and well being of the crew, and that propellant mass is very high,
requiring large spacecraft and increasing substantially orbital lifting cost. For this kind of mission,
as well as for reasonably fast interplanetary travel, stationary electric thrusters are becoming a very
appealing option, mainly because of their specific impulse, now at least an order of magnitude
larger than for chemical rockets. Work in progress worldwide implies that specific impulse is bound
to grow by a similar factor in the years to come. Although single electric thrusters still cannot reach
thrust comparable to chemical or nuclear thermal rockets, clustering is a step in that direction,
provided thrust density (thrust per unit exit area) may be sufficiently high. From this viewpoint, two
classes of electric thrusters appear to have significant potential: thermal arcjets, and MPD self-field
and applied-field thrusters. Most recently, hybrid thrusters (mixed-mode thermal arcjets feeding a
MPD section) have been developed in the US, Russia and in Germany, at the Institute of Space
Systems (IRS) of the Universität Stuttgart. Based on a brief review of the state of the art of
stationary high power thermal and MPD arcjets and hybrid thrusters, future development needs and
trends are presented, showing that thrust measured may already be in the O(10) N range with power
above 100 kW. Increasing efficiency through regenerative cooling, hybrid designs and modular
assembly may enable this class of electric thrusters to reach 100 N with power of order of a few
MW.
4.2. Introduction
Mission scenarios such as building a scientific outpost on the Moon or human exploration of
Mars require new propulsion systems for in-space transportation of heavy payloads [Schmidt et al,
2002]. A thrust level of at least 100 N and a specific impulse level of 30 km/s are of central
importance to increase the payload capacity and shorten the trip time as much as possible. The
propulsion system mass should be as lightweight as possible.
Therefore, both thrust density and efficiency of the propulsion device should be high.
Nuclear and solar thermal propulsion offer very high thrust densities but are still weak in their
specific enthalpy level and thus specific impulse, see Chapter 1, which is clearly below 10 km/s (see
Figure 55). Ion and Hall-ion thrusters offer both the required exit velocity level but their thrust
density is still relatively low.
Promising in-space propulsion candidates for heavy payloads are currently thermal arcjet
thrusters with which an exit velocity of 20 km/s at 100 kW and a thrust density of more than 2100
N/m2 can already be achieved. This technology is in an advanced developmental stage; low power
devices have been implemented in commercial applications with hydrazine as propellant and are
exceptionally reliable. The highest thrusts and thrust densities reached to date have been achieved
with MPD self-field thrusters. However, the effective exit velocity is still limited to 15 km/s.
Although the achievable thrust density is an order of magnitude lower, applied-field MPD thrusters
should still be considered for this type of application because it is possible to achieve very high exit
velocities. Hybrid thrusters are also an interesting development, but they have not been investigated
extensively as much yet. The hybrid concept ATTILA, in which a thermal arcjet thruster is
combined with a second inductive stage, is remarkable for its high thrust densities. A significant
125
increase in the effective exit velocities of these devices compared to the thermal arcjet thrusters can
be expected.
All of these high power, high thrust density thrusters are still at a relatively rudimentary
development stage and are far from being able to be used in space. Up to now, mostly basic
investigations have been performed and some tools for the construction have been developed. A
systematic investigation and optimization of these technologies is still outstanding. Appropriate
developmental test facilities are necessary in some cases. Following is a detailed description of the
developmental stage of these stationary operated thrusters and of future development needs.
Thrust Exhaust Nozzle exit Thrust

F [N] velocity area density
ce [km/s] A [m2] [N/ m2]
Saturn V (1 stage S-IC, 7 600 000 <5 10.752 706 839
_Rocketdyne-F-1)
Solar Thermal Rocket 25 <8 0.002 12 732
Nuclear Thermal Rocket (NERVA) 340 000 <8 11.282 30 138
Thermal Arcjet (HIPARC) IRS 6 20 0.003 2 122
MPD Self-Field (DT2) IRS 28 <15 0.008 3 565
MPD Applied-Field (T) 4.50 42 0.020 225
Ion Propulsion (DS1) <60 0.071 1.302
0.092
Hall-Ion Thruster (SPT100) 0.10 <30 0.008 12.732
Figure 55 Performance data of a solar thermal rocket [Shoji, 1984], a nuclear thermal rocket NERVA [Farbman,
1991], a high power thermal arcjet HIPARC [Auweter-Kurtz et al., 1998], a self-field MPD thruster DT2 [Wegmann,
1994], an applied-field MPD thrusters [Tikhonov et al., 1995], ion propulsion system DS1 [Christensen et al., 1998],
a Hall ion thruster SPT 100 [Archipov et al., 1995] and two hybrid concepts ATTILA [Laure et al., 2003] and
VASIMIR [Chavers et al., 2003] with the Saturn V data as reference.
4.3. Thermal Arcjets

In the last two decades, electrothermal arcjets have been investigated in a wide power range
(0.1 kW to 100 kW) for a wide range of uses [Kurtz, 2000], Error! Reference source not
found.Curran, 1993]. Low power arcjet thrusters from General Dynamics (Primex), USA, have
been used since 1993 as North-South-Station-Keeping (NSSK) thrusters with hydrazine as
propellant Error! Reference source not found. PRIMEX Aerospace Company, 1999], and were
foreseen as primary propulsion for the AMSAT P3 satellite with ammonia as propellant [Rihele,
1999]. The medium power ammonia arcjet thruster on the USAF Argos satellite, the ESEX
experiment [Bromaghin, 1999], also of Primex origin, proved the feasibility of such a thruster for
orbit transfer. In the area of high electric power up to approximately 100 kW, an operable
laboratory model for hydrogen as propellant has been developed [Auweter-Kurtz et al., 1998] and
the potential for this kind of technology has been investigated.
4.3.1. Operational Principle

In an arcjet thruster, electric power from an external power source is used to heat a
propellant gas by means of an electric arc. The heated gas expands through a converging-diverging
126
nozzle to produce thrust as shown in Figure 56. The thruster is axi-symmetric, with a central
cathode normally made of a tungsten alloy (1-2% ThO2) and with the nozzle, also made of
refractory materials, switched as anode. The gas flows through this inner-electrode region and the
arc is heated by Ohmic heat and is at least partially dissociated and ionized.
The arc originates at the cathode tip (typically about 1 to several mm², depending on the arc
power) and is stabilized by the nozzle throat which is built in the form of a ‘constrictor’ (its length
to diameter ratio is typically 1 to 2). The arc is forced by the gas flow through the constrictor to
attach downstream at the beginning of the diverging part of the nozzle. Arc attachment in the
converging part is unfavourable and has to be avoided (“low voltage mode”). A diffuse arc
attachment at the anode minimizes anode erosion and maximizes the heat transfer to the propellant.
The attachment zone moves downstream in the nozzle with rising mass flow and/or rising discharge
current. Heating of the gas takes place mainly in the core zone of the arc; the surrounding cold gas
cools and therefore protects the nozzle throat preventing the arc from attaching there.
Using a simplified one-dimensional model, the exhaust speed ce and hence the specific
impulse Isp of an ideal thermal expansion through a DeLaval-nozzle is proportional to
T
I sp ~
M
with T the plenum temperature of the gas, and M the average molecular mass. [Aweter-
Kurtz, 1992] (This is strictly true only for ideal gases. The effect of increasing temperature on
chemistry, electronic excitation or ionization for non-ideal gases will change the energy available
for propulsion). This means that higher specific impulses are obtained by increasing the temperature
and decreasing the molecular mass of the gas. The upper limits on the temperature are determined
by the heat transfer from the gas to the thruster body and the body’s materials. The practical lower
limit on the molecular mass M is dissociated (atomic) hydrogen, with 1 AMU (atomic mass unit).
This is the main reason that hydrogen and gases containing a high fraction of hydrogen (like
ammonia or hydrazine) are the most common propellants.
During expansion a large fraction of the thermal energy of the propellant gas is converted to
kinetic energy. With the exception of high power arcjets (>50kW), electromagnetic forces do not
play any role in generating the thrust. But also at a level of some hundred kW this portion is very
low. The energy which is deposited in the arc plasma by Ohmic heating is converted to internal,
kinetic, and chemical energy. It is distributed to the surrounding cold propellant gas through heat
transfer and diffusion, and mainly in the constrictor.
The average gas velocity ce of arcjets is high with respect to chemical rockets (up to 20000
m/s at 100 kW), but due to the high gas temperatures and the low mass flow rates the Reynolds
numbers are low (100 - 1000). This means there is an appreciable amount of viscous losses,
especially with small thrusters. The plume of an arcjet shows a pronounced velocity profile with
high core velocities and low border velocities. Other losses are thermal losses (heat deposition into
the thruster body, especially cathode and anode, that has to be removed by radiation).
4.3.1.1. Efficiency.
The main losses in arcjet propulsion are frozen flow losses: since the residence time of the
propellant in the nozzle is in the order of 1 µs, the energy comprised in dissociation, ionization and
excitation cannot be completely recovered in the nozzle as kinetic energy of the exhaust jet.
Therefore, the efficiency η of arcjets with molecular gases is restricted to η< 50%.
127
Figure 56 Operation principle of an arcjet.
The parameter which determines the Isp is the specific power of the gas, defined as the
electric power divided by the mass flow, or P / m& . The higher this specific power, the higher the
specific impulse also, see Chapter 1. The amount of specific power which can be sustained by a
thruster body is a function of the size and material: small thrusters are limited to about 150 MJ/kg,
large thrusters to > 700 MJ/kg (for H2). This P / m& parameter correlates with the performance
parameter Isp; this correlation is only relatively weakly affected by geometry changes in the
thruster. Also the quantity thrust/power, T/P, is correlated by this performance parameter. They are
connected by the efficiency, defined as thrust-power over input-power:
T2 ce2 c ⎛T ⎞
η= = = e ⎜ ⎟
2 m& P ⎛ P⎞ 2 ⎝ P⎠
2⎜ ⎟
⎝m⎠&
At fixed P / m& an increase in Isp will increase the efficiency, enabling an increased propellant
mass saving. Operation at fixed Isp and higher m& will lower P / m& and hence increase the lifetime of
the thruster.
As can be seen from the equation above, by increasing the specific power P / m& of an arcjet
to increase its Isp the gas temperature and hence the frozen flow losses are also increased. This leads
to the unfavorable effect that with increasing Isp the efficiency will inherently decrease, as shown
for example in Figure 57.
There are only few means to increase the efficiency:
• Regenerative cooling: With regenerative cooling, the cold propellant is made to pass through the
hottest parts of the thruster. Part of the thermal losses can be recovered as increased propellant
enthalpy, and the efficiency is raised. The gain in η naturally depends on the heat capacity of the
propellant, i.e. on its specific heat cp and mass flow, and on the ΔT. From the propellant
prospective, H2 is the best suited. Its cp is more than five times higher than that of NH3 at the
same temperature T = 1000 K. An extra advantage of regenerative cooling consists in reducing
the extreme heat stresses near the arc attachment area and nozzle throat, and hence prolonging
the lifetime of the thruster.
• Bi-exit or dual-cone nozzle: Primex first suggested the bi-exit anode [Butler, 1993] (Figure 58) to
recover some of the frozen flow losses. The idea is to provide an area of relatively high gas
density and sufficient time to achieve some recombination. This could be accomplished by a
dual-cone anode; that is, the diverging part of the nozzle is divided into two regions, the first (of
length Lc in Figure 58) with a slowly varying area allowing some recombination, the second a
conventional diverging nozzle for rapid expansion. Since this first section bears higher heat
loads, it is preferably combined with regenerative cooling. A similar approach was presented by
Aston. [Aston, 1993].
128
40 200 mg/s, W
Thrust Efficiency (%)

300 mg/s, W
35 150 mg/s, R
200 mg/s, R
300 mg/s, R
30
25
20
400 800 1200 1600 2000
Specific Impulse (s)
Figure 57 Exemplarily efficiency/Isp curve for HIPARC, water-cooled (W) and radiation-cooled (R) version, with
hydrogen as propellant. [Auweter-Kurtz et al., 1998]
Figure 58 Bi-exit nozzle principle [Aston, 2003].
4.3.1.2. Discharge Voltage.

For introducing power (current × voltage) into the arc, the current should be as low as
possible to alleviate the heat load to the electrodes, meaning the discharge voltage should be high.
An arcjet thruster system is normally ‘current-controlled’. The discharge voltage results from and is
determined by several factors:
• propellant: monoatomic gases with low dissociation potential such as argon show the lowest
voltage at constant current, H2 the highest
• mass flow: the discharge voltage rises with rising mass flow
• constrictor length: the length of the constrictor, i.e. the length of the arc, determines the voltage
drop due to the Ohmic losses
• cooling system: regenerative cooling allows significantly higher voltage at constant P / m& (the
preheated gases increase the plenum pressure and hence the discharge voltage)
• constrictor diameter: at constant mass flow and current the voltage is lower with larger
constrictor diameter (lower plenum pressure)
4.3.2. Theory and Numerics

Although the design of an arcjet thruster as illustrated in Figure 56 seems simple, the flow
and heating processes involved are very complex. Sophisticated numerical analysis and design tools
129
are needed for proper physical modeling and performance assessment. The theoretical treatment of
arcjets and the numerical prediction of their behavior is an ongoing task, mainly in the academic
environment of many countries such as the USA, Japan, Germany and Italy.
The standards required for accurate predictions are high. Recent simulations refinements are
concentrating on non-equilibrium chemistry, transport coefficients and electrode models. The goal
is to calculate all of the arcjet system characteristics and its performance using only global input
parameters, such as propellant type, mass flow and discharge current, with geometry imposing a
second set of conditions. The most efficient computational models have been developed for
hydrogen. Examples of these are Miller [Miller, 1993], Gölz [Aweter-Kurtz, 1998], Butler [Butler,
1993] and Fujita, [Fujita, 1996] which see for details.
Nitrogen as propellant, both experimental and numerical, is considered for basic
investigations mainly in Japan by Kyushu University. [Kuchiishi, 1999] In nitrogen/hydrogen
mixtures simulating hydrazine for low power arcjets, the standard reached is almost as high as with
hydrogen alone; the results are very encouraging. [Megli, 1996]
4.3.3. System Description

Figure 59 shows a simplified electrical scheme of an arcjet system. It consists of the thruster,
the power supply system and the propellant feed system.
Figure 59 Simplified arcjet thruster system with current and propellant circuits.
The power supply system is a key component in operational arcjet systems. It has to assure
reliable arc initiation and a stable steady state operation. The arc is ignited by a series of high
voltage (>2 kV) pulses across the arcjet electrodes. This generates sparks (via a so-called Paschen
breakthrough) in the propellant gas, leading the way to the main discharge. The power control unit
(PCU) must take over the incipient discharge by means of a fast control unit, maintain the discharge
stable, and withstand or correct its possible instabilities, mainly caused by erratic arc attachments.
Since the voltage/current characteristic of an arcjet has the typically negative slope of arc discharge,
see Figure 24 for the water-cooled case, this leads to demands on the PCU. Normally the power
supply is current controlled. As shown in Figure 59, the high power ignition circuit is normally
integrated in the power supply, and uses inductive coupling. In solar-powered satellites, the power
conversion from the nominal satellite bus DC is done by a pulse-width modulated PCU, resulting in
a current controlled output.
In the last few years, mainly by the application of modern digital electronics, immense
progress has been made in reducing the mass of power supplies for low power arcjets [Aweter-
Kurtz, 2002] leading to PCU efficiencies between 90 and 95%.
130
140
Voltage [V]
120
100
300 mg/s, R 300 mg/s, W
200 mg/s, R 200 mg/s, W
150 mg/s, R
80
0 200 400 600 800 1000
Current [A]
Figure 60 Typical voltage/current characteristics of a 100-kW hydrogen arcjet thruster; (R) radiation-cooled, (W)
water-cooled. [Aweter-Kurtz, 1998]
4.3.4. Influence of Propellants

Arcjets can operate on a variety of propellants. As discussed above, the molecular mass of
the propellant should be as low as possible. On the other hand, the propellant and/or the dissociated
products may not react chemically with the thruster body and electrode materials. This excludes all
compounds containing either oxygen or carbon. For use on the International Space Station (ISS)
H2O-containing propellant mixtures show some development potential. A laboratory model has
already been developed. [Nentwig, 1999] Compounds containing elements which could be
deposited on the spacecraft should also be ruled out. This restricts the preferred choice to H2 and
hydrogen/nitrogen compounds. Another advantage of H2 is its higher discharge voltage, which
helps to relay power to the propellant at low discharge currents.
Hydrazine (in laboratories normally replaced by simulated hydrazine, that is N2 + 2H2) is
already in use for low power arcjet applications. Hydrazine has to be gasified and decomposed in a
reactor yielding a mixture of hydrogen, nitrogen and a small amount of ammonia. The Isp is limited
from 500 s for small arcjets (<1kW) to more than 650 s for medium-size arcjets (ca 2 kW). [Lichon,
1993] Efficiencies lie between 30% at high Isp and 40% at Isp ≈ 500 s. No experimental data for high
power thrusters have been measured, but one can extrapolate to an Isp in the order of 900 s at power
levels > 30 kW, corresponding to specific power values P / m& in the 150 kJ/g range. One additional
advantage of hydrazine as propellant for arcjets is that in case of arcjet malfunctions (i.e. no arc
ignition), the thruster can still act as a monopropellant thruster, although at lower Isp.
Ammonia represents space qualified technology already used on ESEX and ATOS satellites
and on resistojets in Russia at the NIIEM Institute. [Trifonov, 1995] The Isp is limited from 500 s
for small arcjets (< 1 kW) to about 900 s extrapolated for medium-size arcjets (> 30 kW).
Efficiencies lie in the 30-40% range. It should be emphasized that the data above are for non-
regeneratively cooled thrusters. With an appropriate design embodying regenerative cooling some
improvements could be expected.
For both hydrazine and ammonia propellants, the Isp range for high power arcjets must still
be established, which is very important for assessing the limits of arcjet propulsion.
Hydrogen is the best choice for arcjets in terms of Isp because of its low atomic mass and
good thermophysical properties. Because of frozen flow losses, efficiencies are low, about 25 to
40%, when using hydrogen, but they can be improved by proper design, and especially by including
regenerative cooling. The maximum Isp depends on the size of the thruster, ranging from about 1000
s with1 kW and 1500 s with 10 kW to over 2000 s with 100 kW, see Figure 61. The data for the
ESEX thruster (Isp = 786s, η = 27%, P / m& =105 kJ/g) fit well into the Isp curve of Figure 61. A
131
disadvantage of hydrogen is its problematic storage for long missions, because it is a cryogenic
liquid. However, for high power missions this issue seems to be now of lesser concern.
2000
Specific Impulse (s)

1600
200 mg/s, W
1200
300 mg/s, W
150 mg/s, R
800 200 mg/s, R
300 mg/s, R
400
0 200 400 600 800 1000
Specific Input Power (MJ/kg)
Figure 61 Specific impulse vs. specific power for a radiation-cooled (R) and a water-cooled (W) high power arcjet
thruster [Aweter-Kurtz, 1998]
By using hydrogen propellant and solar-powered electric thrusters on orbit transfer vehicles
(EOTV) a tank fraction of 0.15 has been estimated [Curran, 19993] which would yield mass savings
of 43% for a 2000-kg payload on a LEO to GEO transfer mission. This result applies to an Isp of
1200 s, a 30-kW engine and 200 d transfer time. The system conceived in [Curran, 1993] does away
with long term cryogenic storage requirements by utilizing hydrogen boil-off cooling.
Mono-atomic gases are in principle preferable in terms of efficiency, as there are no
dissociation losses. However, only helium seems of interest, because of its low atomic weight. But
results achieved to date by The Aerospace Corporation and Stanford University [Walker, 1998
Welle, 1997] have been disappointing, both in terms of low Isp and unstable operation. In addition,
the storage of liquid He is difficult and expensive. Only seeding helium with hydrogen (or vice-
versa) could offer advantages [Welle, 1997 Rybakov 2002] and could possibly result in stable arcjet
behavior, but other issues related to He as propellant, such as unstable discharge, still remain
unsolved.
4.3.5. Lifetime Limiting Factors

At high power, the operational lifetime of an arcjet thruster is limited mainly by the
degradation of the cathode. Cathode tip erosion is a major problem because it can lead to severe
ignition problems. The arc attaches on the tip of the cathode rod and (ideally) is sustained in a stable
discharge by thermionically emitted electrons. That requires a cathode tip temperature >2000° C.
During ignition, when the cathode is still cold, arc attachment may cause local melting on the tip,
leading to high erosion by dislodging cathode material, a “spitting“ process observed in the
laboratory. Cathode erosion increases the discharge voltage by increasing the inner electrode gap,
and can drive discharge instabilities due to the loss of cathode tip symmetry and therefore erratic arc
attachment. Erosion increases with tip temperature, so the cathode geometry (i.e. diameter) has to
be designed carefully and must not be thermally overloaded by excessive specific power P / m& .
Cathode erosion is especially sensitive to oxidizing gases: the propellant must be very clean
and may not contain any oxygen or humidity. The maximum arcjet specific input power P / m& is
primarily limited by the maximum heat load the nozzle throat (the constrictor) can withstand.
132
Therefore, adequate cooling (i.e. regenerative cooling) and good refractory materials should be
used.
4.3.6. Qualification Advantages

In comparison with other electric propulsion systems, such as ion, Hall-ion, and applied-field
MPD thrusters, the qualification requirements of arcjets are much less demanding. The PCU is
simpler. Beside the ignition pulse generator, only one power circuit is needed, and that at relatively
low (80-200 V) voltage. The physics of arcjets allow representative operation at tank pressure in the
range of 1-10 Pa, allowing the use of roots-pump vacuum systems, and a factor ~103 less
demanding than with ion thrusters or SPT (Solid Propellant Thrusters). Due to the high thrust levels
achievable with arcjets, the accumulated test time is also about a factor 10 smaller than with the
above mentioned thruster types.
4.3.7. Existing Arcjet Thruster Technologies

This section describes ongoing activities concerning high power arcjet thruster technologies
and available arcjet systems worldwide.
High (to medium) power arcjet (10 to 100 kW) research has been restricted to only a few
places - JPL and Primex in the US and IRS in Germany - due to the lack of funding and testing
facilities (sufficient large pump systems) elsewhere.
4.3.7.1. USA
Work on arcjet propulsion started here in the late 1950s and 1960s and resulted in quite
remarkable laboratory thrusters. [G.L. Cann 1997] After a pause of roughly twenty years due to a
lack of power and missions, work was resumed by government agencies such as NASA, JPL,
USAF, SDIO etc. Part of the task was soon handed over to industry, mainly RRC - Rocket Research
Company (later renamed PRIMEX and now General Dynamics and TRW). It should be
emphasized, however, that the years of industrial development were accompanied by either direct
scientific support from the government agencies mentioned above or through sponsorship of
universities.
4.3.7.2. Germany
High power arcjet development is currently performed only in Germany at IRS. The high
power arcjet HIPARC has been designed and investigated to establish operation in the 100 kW
range and to produce first performance maps. With its large roots-pump system capable of
volumetric flowrates > 250 000 m³/h at 10 Pa, IRS is well equipped for high power, high mass
flowrate testing. Of the two HIPARC versions built, one was water-cooled one for testing
diagnostics and geometry changes, while the other was radiation-cooled, see Figure 62. Although in
no way optimized, it yielded Isp over 2000 s at an efficiency of 28 % (see Figure 57). [Aweter-
Kurtz, 1998]
133
Figure 62 Schematic drawing of the 100 kW HIPARC-R thruster; constrictor diameter 4 mm. [Aweter-Kurtz, 1998]
In Figure 60 the current/voltage characteristics for the radiation-cooled and the water-cooled
HIPARC thrusters are shown. In the case of the radiation-cooled HIPARC the voltage is
independent of the current over a wide range. It does not decrease, as is the case with low and
medium power arcjets, and with water-cooling. This is a clear sign that hydrogen is almost
completely dissociated and ionized, as also predicted by calculations [Aweter-Kurtz, 2002], and that
the magnetic acceleration due to the self-induced magnetic field does not play a significant role. In
addition, the efficiency of the radiation-cooled HIPARC is independent of the specific impulse over
a wide range.
It is to be expected that a thruster optimization using regenerative cooling and a bi-conical
nozzle design will allow for a significant increase of the effective exit velocity. Intensive
investigations of the potential of regenerative cooling were carried out at IRS with the medium
power arcjet MARC in the 5-12 kW region. [M. Riehle 1999] In particular, numeric design tools
were developed and tested for optimizing the design of arcjets. Figure 27 shows that specific
impulse and, especially, efficiency increase during the transition from the purely radiation-cooled
MARC2 to the regenerative device (Figure 38).
Figure 63 Efficiency gain with regenerative cooling; Figure 64 Cross section of the MARC 4 thruster
hydrogen, 5 kW class; MARC 2 radiation cooled. head. [Hammer 1997]
[Hammer 1997]
4.4. Magnetoplasmadynamic Thruster
The main operating principle of MPD thrusters is to use also electromagnetic forces (the
Lorentz force) to accelerate the propellant. Ohmic heating and the resulting thermal thrust portion
play a subordinate role, but often one that should not be ignored.
Generally, two main types of MPD systems may be defined:
134
self-field MPD thrusters
applied-field MPD thrusters.
MPD applied-field thrusters are always operated in a stationary mode. Self-field thrusters
have three operating modes: continuous, quasi-stationary and non-stationary. Because from today’s
perspective of space transportation high average thrust can only be achieved with stationary MPD
self-field accelerators, only this type of thruster will be considered.
When designing MPD thrusters, one must be aware that only charged particles can be
accelerated by electromagnetic force. Therefore the propellant should be almost completely ionized
in order to achieve high efficiency. Recombination is undesirable, anywhere in the entire thruster.
To this purpose the gas temperatures in MPD devices are higher and the pressure is lower than in
thermal arcjet devices. Pressure in the energy deposition (ionization) chamber is typically between
5⋅10-4 bar and 0.5 bar depending on geometry, power level and mass flow.
4.4.1. Self-Field (SF-MPD) Thrusters

The SF-MPD thrusters (Figure 65) have been much more extensively investigated
experimentally than the other thrusters discussed here, both in continuous and in pulsed quasi-
steady operation, and with many propellants. The principle of operation should be clear from Figure
65, showing how the self-induced circumferential (azimuthal) magnetic field Bθ , interacting with
the applied meridian currents produce meridian forces on the plasma that are normal to the current,
and directed mostly axially and radially inward. These forces, called blowing and pumping by
Maecker [Maecker, 1955] result in direct (electrodynamic) acceleration, and in a pressure (“pinch”
pressure) and temperature rise, driving also a thermodynamic expansion and acceleration,
respectively. [Jahn, 1968] The pinch pressure may also enhance nozzle recombination through the
increase of collision rate, scaling with the cube of pressure.
Figure 65 Operating principle of the self-field MPD thruster.
Of the thruster types discussed here, the SF-MPD thrusters have the lowest arc voltages for a
given power level, because the self-induced magnetic fields are relatively weak unless very high
currents (that is, tens of kiloamps) are applied. Typical values (for nozzle-type thrusters with argon
propellant) are ≈ 50-70 V for a ≈ 200 kW continuous thruster. Since the electrode loss voltages
change much less rapidly with the current or power level, the thermal efficiencies of these thrusters
135
improve with increasing power. Because of the high currents required, the SF-MPD thrusters also
have the most severe electrode erosion and cooling problems.
For in-space transportation missions considered here only continuous mode thrusters make
sense, pulsed and quasi-steady pulsed thrusters are disregarded in the following, despite they have
been investigated in many laboratories over decades. By directly comparing geometrically identical
thrusters in continuous operating mode to quasi-steady (q-s) pulsed mode [Aweter-Kurtz, 1994], it
was shown that results of q-s thrusters could not be extrapolated to the steady operation case, and
therefore that most q-s results are irrelevant to designing real high power and high thrust missions.
The phenomenon of self-magnetic plasma acceleration was first investigated by Maecker [
[Maecker, 1955], Wienecke [Wienecke, 1955] and others in Germany (ca. 1955) who studied
velocity distributions in the cathode jets of carbon arc lamps. A simple integral formula by Maecker
[Maecker, 1955] gives the electromagnetic thrust fairly precisely for any thruster geometry and
assumed, or measured, current distribution on the cathode and anode. In this way the contribution of
the electromagnetic forces to the average axial exhaust velocity could be found in the form
μ0 I 2 −7 I
2
c EM = f ( geom) = (1.85 ÷ 3.05) ⋅10
4 π m& m&
where cEM is the electromagnetically produced exhaust velocity, I is the total electric current,
m& the propellant mass flow rate, f(geom) is a function of the electrode geometry and of the radial
current distribution on the electrodes. [Jahn, 1968] The numerical values given in this equation are
typical for thruster models tested so far. It is important that the purely theoretical Maecker integral
formula for the thrust and the resulting electromagnetic contribution to the average axial component
of the exit velocity - which are well verified experimentally - are independent of any assumed
model, or rotational (circumferential) uniformity, of the discharge. Any arc spokes and/or anode
spot attachments have no influence on the result. These details of the arc form and spatial
distribution, however, affect the exit velocity distribution, the resulting thrust direction and the
efficiency of electromagnetic thrust production. Of course, models generally used to evaluate the
factor f(geom) are simple, and herein lie some uncertainties.
Especially if the thruster is nozzle shaped, a non-negligible electrothermal thrust adds to the
purely MPD thrust. The thrust vs. current curves for various argon mass flow rates for the IRS DT-2
thruster, Figure 66, are shown in Figure 67. Also plotted in this graph is the pure electromagnetic
thrust, TM, calculated according to [Maecker, 1955 Jahn, 1968]
μ 0 2 ⎛ ra 3 ⎞
TM = I ⎜⎜ ln + ⎟⎟ .
4π ⎝ rc 4 ⎠
Tm,max and Tm,min are calculated with cathode radius rc and for the maximum and minimum possible
anode radius ra. It can be seen that at lower current levels the electrothermal part of the thrust is
prevalent and only with rising current levels the total thrust does approach the gradient angle of the
magnetic thrust.
136
Figure 66 Nozzle type MPD thruster DT (DT2 with 24 Figure 67 Thrust vs. current curves for the DT2 thruster
mm throat , DT6 with 36 mm). at various mass flow rates.
Empirically, it has been known for some time that the current-to-gas-mass flow ratio I 2 / m&
has a limit for each thruster and propellant where stable operation of the arc becomes impracticable.
Beyond that so-called “onset” value, pronounced arc voltage fluctuations, severe anode spots and
anode erosion, increased anode losses and thruster efficiency drop are observed. Attempts to exceed
the onset limit ( I 2 / m& ) crit are foiled by erosion of the anode, exposed insulators, and perhaps
cathodes. This erosion apparently supplies the additional mass flow needed in front of the anode to
limit I 2 / m& , as demonstrated e.g. by Suzuki et al. [Suzuki, 1978]
The great importance of the onset or critical I 2 / m& values is clear since practically useful
thruster efficiencies (say above ≈ 30%) can (if at all with a given thruster) be achieved only with
each propellant at the highest specific impulse (or I 2 / m& ) values presently attainable with that
propellant, as will be seen.
For SF-MPD thrusters, Hügel [Hügel, 1980] has correlated the onset ( I 2 / m& ) ratios from
different sources and thrusters with different propellants against their atomic weights which gave
roughly
⎛ I 2 M a1 / 2 ⎞ ⎡ As ⎤
⎜⎜ ⎟⎟ ≈ (15 ÷ 33) ⋅1010 ⎢ kg ⎥
⎝ m& ⎠ crit ⎣ ⎦
or
⎡m⎤
(c EM ) crit ≈ (3 ÷ 10) ⋅10 4 Ma −1 / 2 ⎢⎣ s ⎥⎦
where Ma is the dimensionless atomic weight of the propellant, and the empirical constants
in these equations depend on the thruster geometry and the power level. These results agree with
those of Malliaris [Malliaris, 1971]. Both researchers confirmed these data with noble gases as
propellants. With molecular gases, however, this dependence could not be verified at IRS [Aweter-
Kurtz, 1998].
Pulsed SF thruster data from Princeton University and the University of Tokyo report
electromagnetic velocities by a factor of 2x larger than the upper value of the above equation for
argon, or ≈ 33 km/s, but they could not be verified with steady-state thrusters.
Hügel [Hügel, 1980] calculated the current and gas density distributions for some of these
thrusters using several simplified flow models and empirical electron temperatures. He showed that
near the onset, the ion density at the anode approaches zero due to the pinch (or radial inward)
137
component of the electromagnetic force distribution. He concluded that at the onset points the ion
densities at the anode become too low to neutralize the electron current there, assuming singly-
ionized plasma. This explanation is immediately plausible from the fact that the ratio of the
magnetic pinch pressure rise (ΔpM) to the mean gas pressure in the jet pj is proportional to the
current/mass flow ratio parameter:
Δ pM I2
~
pj m&
as may be verified by simple approximations.

However, there appear to also be other theoretical approaches for arriving at critical values
2
of I / m& . Cann [Mooer, 1965] arrives at this from his minimum arc voltage principle. Schrade
[Schrade, 1991] can show from his arc stability criterion that instability must occur above certain
I 2 / m& values.
King [King, 1981]has approximated the MPD accelerator flow with a one-dimensional
channel theory and from this has estimated values of the Hall parameter in MPD accelerators. He
ascribes the onset limit of I 2 / m& and the observed current distributions to excessively large Hall
parameter values (approaching ≈ 10) as previously suggested by Rudolph et al. [Rudolph, 1980]
Incidentally, the Hügel model would also indicate increasing Hall parameter near the anode surface
as a result of the decreased plasma density there. The induced or “back-emf” (δER = - ceBθ) which
was neglected in the Hügel model would add to this effect, as pointed out by Lawless [Lawless,
1983] and others. Wagner [Wagner part 1 e 2, 1998] showed in a detailed analysis that these macro-
instabilities could be caused by drift and gradient driven instabilities. The cause of the instabilities
found was the space charge or diode instability which in its nonlinear development could lead to so-
called current chopping instabilities.
All these different simplified analysis models, which are not contradictory to each other,
predict some of the experimentally observed trends, but they cannot give absolute limits of the
performance of thrusters. The discharge behavior near the onset, e.g. the arc attachment on the face
of the anode, is not analyzable with either the Hügel or the King model. But these models suggest
ways in which the limiting values of I 2 / m& , hence the achievable specific impulses with each
propellant and generally also the efficiencies could be improved. They are:
a) injecting neutral gas near the anode to increase the gas and ion density there,
b) extending the cathode to achieve (if possible) radial current flow to eliminate the radial
(“pinch”) pressure gradient,
c) lengthening the thruster anodes to reduce current densities and Ohmic heating and thereby
improve the thruster efficiencies.
Cold (neutral) gas injection near the anode surface has been used in pulsed self-field
thrusters by the Princeton University group [Boyle, 1976] and by the University of Tokyo group
[Kuriki, 1981], in each case apparently with success in improving the specific impulse or onset
velocity limit. With the IRS steady-state DT thrusters no significant influence of anode propellant
injection could be found. An improvement in the onset behavior could be achieved by a very
precisely symmetrically adjusted thruster: The onset of oscillations did not shift, but without the
formation of anode spots.
Extended cathodes and long cylindrical anodes have been designed and operated by IRS in
steady-state mode, the ZT thruster series, (and q-s pulsed by the Princeton University group [King,
1981]), to counteract the above mentioned anode starvation due to pinch forces which would not
occur with a purely radial current flow. With these MPD devices, which should have a mainly
radial current distribution and therefore low discharge voltage, an attempt was made to achieve a
138
maximum relation of magneto-plasma-dynamic thrust to thermal thrust and a low pinch effect at
high power levels.
The cylindrical ZT3, shown in Figure 68, consists of three anode segments, two neutral
segments and a backplate, all out of water-cooled copper, and the thoriated tungsten cathode. The
discharge chamber diameter is 70 mm, its length 150 mm. The current through each anode segment
can be measured separately. The cathode was designed to counterbalance thermal damage effects
encountered at high current levels (see below). In the back the anode diameter is 40 mm, and is
subsequently reduced to a 10 mm radius at the tip.
Figure 68 Scheme of the ZT3 thruster of the IRS.
In Figure 69 and Figure 70 experimental results of the ZT3 thruster with a mass flow rate of
2 g/s argon are presented. The voltage dependence is compared to the results of the nozzle shaped
thruster DT6 with 36-mm throat diameter, both running at the same mass flow rate. This
comparison shows that the voltage of the cylindrical thruster is low and increases only slowly with
the current, while for the nozzle shaped thruster it is a factor two to four times higher and rises
steeply. At ca 7450 A plasma instabilities occur for the DT6, also recognizable by the additional
ascent of the voltage. For the ZT3 thruster, no indication of an occurrence of instabilities could be
detected up to 12700 A where an I 2 / m& value of more than 8 × 1010 A2s/kg was reached, whereas
for the nozzle type MPD thruster a critical I 2 / m& value of ca 2.7 × 1010 A2s/kg was found, with
argon as propellant with all thrusters of the DT series.
With the ZT3 the thrust increases with current and reaches ca 10 N at 12.7 kA, Figure 70.
No higher values could be reached since with even higher current levels severe anode cooling
problems were experienced which prevented the thruster from being run to its I 2 / m& limits. This is
caused by a current concentration at the “anode edge”, verified also by numerical calculations
[Boie, 2000], Figure 71.
139
Figure 69 Voltage-current characteristics of the ZT3 and Figure 70 Thrust vs. current curve for the ZT3 thruster.
DT6 thrusters.
Figure 71 Calculated current distribution, 10 000A.
Through such geometric variations, found mostly empirically, but partly also by model
calculations [Heiermann, 2002], considerable performance improvements have been achieved. It
has been shown that the onset limits and thereby also the achievable specific impulse and efficiency
values depend very strongly on the thruster geometry, the propellant used and on the absolute
thruster size, the power level and operating regime relative to the thruster size.
The dependence of the performance on the propellant has not been as thoroughly explored
as that of the geometry. Very roughly, the maximum exhaust velocities still appear to vary like the
Alfven velocities, or like Ma-O.5 at least with noble gases as propellants, and maximum efficiencies
at the maximum velocities are expected to be relatively insensitive to the atomic weight of the
propellant.
In summary, continuous water-cooled self-field thrusters have, under reliable test conditions
(adequate vacuum), reached efficiencies with argon of up to ≈ 25% at Isp values of about 1400 s, at
power levels up to ≈ 500 kW. The low efficiencies are due to low arc voltages, compared to those
of thermal arcjet thrusters and low thermal efficiencies (55 to 80%). Radiation-cooled SF thrusters
should have higher thermal efficiencies but are mostly of lower power level.
To investigate the influence of hot anodes on the operation behavior of steady-state SF-
MPD, such as anode voltage drop and critical I 2 / m& values, the MPD thruster HAT (“Hot Anode
Thruster”) was built at IRS, geometrically relative similar to the DT2, Figure 72. Its anode is made
of thoriated tungsten and partly coated with TaC to increase the emissivity. Figure 73 shows that the
arc voltage levels are clearly below those of the DT2 thruster with water-cooled anode. Also the
critical I 2 / m& seems to be higher, but this could also be an effect of the yet slightly different
140
geometry which will be checked in the future. (The critical I 2 / m& with the HAT was not explored
in order to avoid a destruction of the expensive anode).
Figure 72 Configuration of the hot anode thruster HAT Figure 73 Voltage vs. Current curve of the HAT
of IRS. compared to the DT2 thruster.
For pulsed SF-MPD thrusters at power levels of 3 to 6 MW, efficiencies into the mid-thirties
at Isp levels of 2000 to 3000 s and correspondingly much higher voltages and thermal efficiencies
have been reported for argon, but it is very doubtful if these values could be reached by continuous
running SF-MPD thrusters.
4.4.1.1. Importance of Electrode Design.

An understanding of the electrode arc attachment regions is essential for successful thruster
design and operation in view of the following problems:
electrode erosion, which is critical both for thruster life and for spacecraft surface
contamination;
transition from uniform to spot attachment on the anodes (“onset” effect) which limits the
performance (Isp and η) and causes both excessive erosion and electromagnetic noise (EMI);
electrode losses, which critically affect thruster efficiency and cooling requirements (i.e.
thruster weight).
Erosion of cathodes is not an issue from the viewpoint of thruster life if the propellant is
purified from oxidizing components (see below), because the limits imposed by spacecraft
contamination cannot be stated generally. Anode erosion may be manageable, as long as uniform
(spotless) attachment is achieved.
The formation of macrospots on the electrodes, i.e. the shift from diffuse and more or less
even arc distribution (which for the cathode may include an even distribution of microspots) to
localized concentrated arc attachments is, of course, akin to the arc filament formation in the
midstream region but influenced by the specific boundary conditions at each of the electrodes. It
must therefore be discussed separately for each electrode type. In MPD thrusters, macrospots cause
excessive (destructive) erosion on each of the electrodes.
Electrode losses (notably on the anode) dominate the efficiency limits of small MPD
thrusters, where the electrode losses can take up 50 to 60% of the overall voltage drop. For larger
thrusters (MW regime) this loss is probably less dominant (extrapolated from pulsed SF-MPD
data), but it does strictly limit the current densities on anodes if simple radiation cooling is required.
141
4.4.1.2. Electrode Voltage Drops and Loss Distribution.
In MPD thruster designs, as in many other arc devices (such as short-arc noble gas and
carbon arc lamps, welding arcs, etc.) a major portion of the ionization takes place at or near the
cathode. In these cases, there is therefore a relatively large potential drop in the arc near the
cathode, much of which must, however, be attributed to ionization (frozen flow loss) rather than as
a cathode loss. The plasma thus formed is blown as a cathode jet toward the anode, where there is
usually no mechanism or need for additional ion generation if the ion density there is sufficient to
neutralize the imposed electron current. As long as this is satisfied, the anode potential drop can be
quite small (even negative). In spite of this, the overall losses at the anode can be very large, as will
be shown. Another energy loss which must be invested at the cathode is the electron work function
Φ (≈ 4 to 5 V for most metals), which appears as an additional heat load at the anode and is
frequently accounted for as an anode loss, though actually it should be considered as a usually
unrecoverable “cost” of running an arc.
MPD Thruster Cathode Phenomenology. Cathodes made of refractory metals (W, Ta) are at
present the only ones known to be potentially suitable for larger MPD thrusters, where the so-called
high current regime from a few hundred ampere into the high kiloamperes and gas pressures at the
cathode from a few hundred millibar on down is applicable.
Cathodes of continuous thrusters can operate, in their design range of conditions, with an
apparently spot-free diffuse arc attachment covering a fairly large area. To reach this optimal (from
the viewpoint of erosion) condition, the cathode must be of the right material (e.g. thoriated
tungsten), be able to reach the required temperature (2,600 - 3,300 K) over a sufficient area
(implying adequate but not excessive cooling) and be surrounded with sufficient gas density for the
imposed current. Under all other conditions, such as continuous thrusters during warm-up, cathodes
with inadequate gas pressure or design or “aged” material (thoria depletion) as well as cold
cathodes, pulsed thrusters operate in some form of spot mode. This involves local melting and
substantial vaporization of cathode material, as will be discussed.
The conditions for diffuse cathode attachments (“thermionic cathodes”) in MPD thrusters
are most comprehensively treated by Hügel and Krülle [Hügel, 1969] together with Cann and
Harder [Cann, 1964]and Goodfellow [Goodfellow, 1996]. [Hügel, 1969] also examines the possible
effects of the Bθ self- field on macrospots of cylindrical cathodes.
With diffuse cathode attachments Hügel and Krülle [Hügel, 1969] obtained current densities
to 2,000 A/cm2 with argon and to 5,000 A/cm2 with hydrogen at 100 mm Hg and 1,000 A, at Ts ≈
2,800 to 2,900 K and 3,200 to 3,300 K, respectively. The net heat fluxes into the cathode ranged
from ( 0.6 to 2.5 kW/cm2 for argon and 1.7 to 4.0 kW/cm2 for hydrogen, the largest values always
at the highest pressure (( 100 mm Hg). Note that the limit for re-radiation at the melting point is 1
kW/cm2, so that generally a larger area than the arc attachment must radiate to cool the cathode.
Thus, thoriated tungsten cathodes can operate in the spot-free thermionic mode, with current
densities up to at least 2 kA/cm2 with minimal losses, at ( 2,800 K, the highest still acceptable
surface temperature from the viewpoint of vapor pressure.
Much higher current densities (to at least 5 kA/cm2) are possible, but at excessive
temperatures for long life (i.e. above 3,000 K) because of high sublimation rates. Aging of the
material (presumably thoria depletion) was found to be a serious problem in some cases. [Malliaris,
1967]
Thruster anode phenomenology. In a long arc without axial flow, there is a small ion drift
(i.e. positive ion current) away from the anode. These ions must be replaced at the anode to provide
plasma (space charge) neutralization and thus allow the main electron current to reach the anode. To
produce these ions, either from ambient neutral atoms or from anode material, an anode arc
contraction and fall zone sets itself up having, at the lower densities, a voltage close to the
ionization potential of the substance to be ionized. Normally an anode spot develops, from which
142
anode material is evaporated and ionized, because in practically all (except alkali vapor) arcs the
anode metals are more easily ionized than other arc gases. These anode contractions, while less
concentrated than those at cathode spots, produce self-field anode jets (similar to those from
cathodes), which increase the ion flow and ion current contribution away from the anode above the
normal ion drift of an arc column. If they occur in SF-MPD thrusters the anode will be destroyed.
Spot-free anode functioning is, however, possible in SF-MPD thrusters. There the plasma is
generated mostly near the cathode and in some cases also in the midstream discharge and is blown
toward the anode surface. Under these conditions there is normally no need for ion generation near
the anode and the anode potential drop becomes zero or even negative as predicted by the classical
anode theory for this case. [Hall, 1964]
For such MPD thrusters (with rear cathode), Hügel [Bez, 1956] has delineated three regimes
of spot-free anode operation. The first regime is characterized by a sufficient supply of ions for
neutralization and replacement of these drifting (or blowing) away, and a sufficiently large natural
diffusion of electrons toward the anode to provide the current drawn. If the normal electron
diffusion exceeds the electron current drawn by the external circuit, the anode surface, if it is
“cold”, becomes slightly negative relative to the plasma, similar to a cold floating probe in a hot
plasma. (If it is hot, it will re-emit some electrons.) Depending on the electron temperature this
could be 3 to 5 V negative, though a cool gas wall boundary layer will also somewhat impede this
electron flow.
As the current is increased at constant mass flow rate (or I 2 / m& is increased by any path), a
second regime is reached through the increased magnetic pinch effect and increased relative current
density, i.e. closer approach to “the onset of instabilities” conditions. This second regime described
by Hügel is one in which the natural electron diffusion no longer supplies the demanded current
density to the anode, though the ion supply to the anode is still sufficient. This anode regime exists
for the thruster type and operating regime treated here, where the ion supply carried by the jet to the
anode exceeds the demand there longer (with increasing I 2 / m& ) than the natural electron diffusion
satisfies the required electron current. In this regime, a positive voltage drop develops near the
anode, increasing linearly with increasing I, from 0 to ≈ 6 to 8.5 V for argon.
This results primarily in increased electron drift current through potential gradient plus
small contributions through ion generation (δne) and temperature increase(δTe). Since in this regime
the increase in anode voltage is proportional to the increase in current density (δje) imposed, this
voltage (loss) rise should be reducible by increasing the anode area exposed to the plasma flow.
With further increase in the current/mass flow parameter ( I 2 / m& ), a third regime of
theoretically possible spot-free attachments is reached in these thrusters, where now also the rate of
ion flow toward the anode becomes insufficient. In this regime, which Hügel calls that of the
genuine or classical anode fall, the anode potential must instantly increase to the potential drop
required for ion generation.
Using classical anode theories [Bez, 1956 Engel, 1941] and the “field” ionization process
appropriate for the low density regime, Hügel calculated an anode fall voltage of 12.4 V for singly
ionized argon and slightly lower values (10.7 V and 8.8 V) for xenon and krypton. Double
ionization requires roughly twice these values.
Since all practical electrode materials have lower ionization potentials (below 8 V), a diffuse
anode attachment of this last type is unstable and will break down to anode spots with anode
vaporization, the normal operating mode of are anodes described initially which results in a lower
anode potential. Also the third regime just described, where the plasma density near the anode
approaches zero, is in fact close to or at the critical I 2 / m& point where probably various other arc
instabilities appear in addition to the anode spot formation. The beginning of this third ( i.e. the
classical) anode operating regime is therefore (also in Hügel's designation) the onset point, at least
for these thrusters where “anode starvation” due to the pinch pressure gradient is dominant.
This regime has to be avoided in SF-MPD thruster operation in any case.
143
4.4.1.3. Electrode Erosion.
The limits of allowable electrode erosion will first of all be set by thruster life requirements.
Much lower erosion limits could be imposed by spacecraft surface contamination if the eroded
material vapor can end up in the form of slow ions (e.g. by charge exchange) in the spacecraft
vicinity and deposit itself on charged insulator surfaces of the spacecraft. This latter erosion limit is
difficult to estimate in a general way, but all condensable vapors near a spacecraft constitute a
surface contamination hazard.
The allowable erosion, from the viewpoint of thruster life, depends on the life requirements
- typically 1,000 to 10,000 hours for near-earth missions - and the thruster design, type and size.
Some typical numbers for a 10-N continuous self-field thruster (4kA, 1 g/s argon) for 3,000 hours
life (consuming 4.3 . 1010 C, 11,000 kg propellant) would be:
cathode (W) anode (Mo)
allowable erosion mass loss mer, (g) 50 250
mer/Coulomb, (g/C ) 1.2 . 10-3 5.8 . 10-3
mer/propellant mass 4.6 . 10-6 2.3 . 10-5
This corresponds to a few millimeters per year (or the order of 10-7 mm/s) material loss from
the faces of the electrodes exposed to the arc for that size and type of thruster. Note that this is
about or less than the normal vaporization rate of tungsten at 3,000 K, and 1/2,000th of that rate at
the melting point.
The possible erosion mechanisms for both cathodes and anodes are
• macro particle ejection (spitting) and massive vaporization (boiling) from molten pools
under macrospots,
• localized vaporization (boiling) at microspots,
• vaporization from larger molten surface areas (without boiling or spots),
• sputtering due to ion impact (normally negligible),
• continuous surface vapor loss due to excessive average surface temperature without melting.
With brittle metals (e.g. tungsten) there is, in addition, the possibility of cracking due to
thermal fatigue.
To prevent substantial overall vaporization at excessive average surface temperatures, the
average electrode temperatures must be controlled by adequate cooling. Typical average
temperature limits for three electrode metals, as dictated by their vapor pressures, are ca 2800 K for
tungsten, 2600 K for tantal and 2200 K for molybdenum. Generally, the average electrode
temperatures should be held below these given for, say 1 mm surface loss per year to allow for local
overheating. This should be readily achievable with anodes, but with cathodes the temperature
given is at the limit for effective thermionic emission, so that the arc may concentrate itself and
produce higher surface temperatures in spite of radiation cooling.
This leaves the various kinds of spots as the major sources of erosion beside the sublimation
of material due to the high temperature. But with duly constructed cathodes, spots should only
occur during ignition.
Cathode erosion. In a steady-state MPD device one has to discern between two operation regimes:
• The ignition phase, which lasts typically about one second (depending on the ignition
current and cathode mass and design)
• The steady operating phase, which lasts (at least in the laboratory) between several minutes
to several hours.
The ignition phase is characterized by a cold cathode and a highly instable, spotty arc
attachment. These arc spots jump irregularly over the cathode surface, causing melting with
relatively high erosion and leaving “craters” of sizes ranging from some microns to a tenth of a
144
millimeter. This starting phase ends when the cathode is hot enough for the thermionic emission to
support the current demand. The measured averaged erosion during this phase is ca. 13 μg/s
[Aweter-Kurtz, 1993] and is insensitive to the propellant gas.
Contrary to the erosion during the ignition phase, the erosion in the thermionic phase
[Aweter-Kurtz, 1993]is dependant to the propellant: nitrogen yields the lowest (ca. 0.5 ng/C),
Hydrogen the highest (ca. 2.2 ng/C) with argon in between (ca. 1.5 ng/C). It showed that the steady
state erosion is very sensitive to the grade of the propellant: the cited values were reached with high
grade gases, which were further cleaned from oxygen and humidity. With normal uncleaned
welding argon, the erosion yields were about a factor 20 higher.
With a pure thermionic emission, erosion is dominated by the sublimation rate s, which can
be calculated with the Dushman equation to
m& subl M pv (T ) ⎡ kg ⎤
s (T ) = = ⎢C ⎥
I 2π ℜ T j ⎣ ⎦
with pv as the vapor pressure.

This sublimation rate yields, with a cathode temperature of 3000 K and a current density of
j = 5 × 106 A/m2, 2 ng/C, which corresponds well with the experiment.
A serious life limiting process in cathodes could be the depletion of thoria or other additives.
This is difficult to detect in cathode sections after, say 50 hours of life and may require long life
tests for accurate determination and special provisions to counteract, if possible.
Anode erosion. Anode erosion in SF-MPD thrusters can vary over wide ranges, but there
are only few accurate data available. Continuous self-field thrusters must be operated at currents
some safe margin below the onset of instabilities or anode voltage rise regime, to avoid rapid
destruction. At the first sign of any onset phenomena like spots, foot points and/or fluctuations, the
current has instantly to be turned off or reduced. With diffuse anode attachment the anode erosion
rates, if any, are so small as to be difficult to detect. On the contrary, on cooled copper anodes same
grey deposits (presumably of tungsten cathode vapor) are observed, which would tend to build up
rather than erode the anodes. With hot anodes, no erosion investigations are known.
In summary, the erosion of tungsten cathodes appears acceptable in the thermionic regime,
as long as the additive (thoria, etc.) remains present at the surface. For cooler cathodes (typical of
present pulsed thrusters) and for frequent cold starts, cathode erosion still is excessive.
Anode erosion appears manageable with diffuse attachment (i.e. far from onset conditions)
and possibly also with spots in pulsed and/or axial applied-field thrusters, using high melting
metals.
4.4.1.4. Electrode Losses and Cooling.

The electrode losses are hard to separate out precisely because of the complicated and
interlacing energy transfer mechanisms which occur at the electrodes. However, without attempting
to resolve all these still open questions, we will attempt to put down a few orders of magnitude on
the net heat loads and draw conclusions on limiting permissible current densities and cooling
capabilities.
Cathodes. For typical thermionic cathodes of continuous thrusters, the net heat loads have
been given [Hügel, 1969] as 1 to 5 W/A and 0.7 to 2 W/A for argon and hydrogen, respectively,
with the lower values applying to the higher current densities. Many other propellants fall
145
somewhere in between. The net heat flux values ranged from 0.6 to 2.5 kW/cm2 for argon and 1.7 to
4.0 kW/cm2 for hydrogen, depending on pressure and current density.
Radiation cooling at the allowable temperature for long life (2,800 K, with emissivity 0.35
for tungsten) removes only about 120 W/cm2. Therefore, purely radiation-cooled cathodes in
continuous operation can be loaded only to about 50 to 100 A/cm2, a factor of 10 to 100 lower than
the possible current densities, according to [Hügel, 1969]. In a radiation cooled thruster, the cathode
will receive same additional radiative heat load from the anode and other hot parts of the thruster. If
higher current densities than those given above are required (or unavoidable), additional cooling
must be provided. With a few propellants (e.g. H2, NH3, Li) this can be done regeneratively, see
below. Otherwise, heat removal by simple conduction or (for larger thrusters) by a liquid metal heat
pipe must be provided.
In any case, cathode cooling does not appear to present insurmountable problems if
macrospots can be avoided.
At high current densities within the cathode, the Ohmic heating gives rise to severe
problems. Cracks releasing molten material from the interior near to the rear end of the cathode,
behind the arc attachment zone. Figure 74 depicts a cathode of the DT2 thruster with a diameter of
16 mm which failed at a current level of 6500 A.
Figure 74 Damaged cathode of DT2 (16 mm (, 6500A)
Investigations [Aweter-Kurtz, 1993] lead to the conclusion that at the high temperatures
inside the cathode the thoria lumps together, forming spheres of ThO2, which reduce the conducting
cross section of the cathode and hence further rising the temperature due to Ohmic heating until the
thoria evaporates under destruction of the cathode (This was the reason for the tapered cathode of
the ZT3 thruster, Figure 68). In summary, the cathode has to be designed very carefully to avoid
overheating in its shaft and to minimize the erosion.
Anode. The anode net energy input equation in simplified form as given by Hügel
(assuming the total current being carried by the electrons) is:
QA = I (VA + 5/2 k Te + (0+ Vconv )
which must be removed by cooling. The first three terms on the right are energies brought in
by the electrons, while Vconv is the normal (heavy particle) convective heat transfer, and radiation
from gas and cathode is neglected.
For liquid cooled copper anodes of continuous thrusters, the anode fall voltage VA is ( 0 to
10 V, the upper value being applicable near the onset of instabilities. With (5/2)kTe ≈ 3.5 V, Φ0 =
4.4 V and Vconv ≈ 2 V, the net anode heat loads run from about 9 to 20 V times the current, where
146
the higher value applies to the interesting performance points. With optimized high performance
thrusters, the values can run considerably higher.
If the anode is to be radiation-cooled, then the average heat flux values and thus the average
current densities are quite limited. At the maximum allowable temperature for long life, tungsten
can radiate only about 120 W/cm2 and molybdenum less than half of that. The effective emissivity
of the anode can be increased by coatings or by fins on the metal which are more effective the
smaller the thruster. Still, except for very small thrusters, the current densities on radiation-cooled
anodes will be limited to the order of ≈ 30 A/cm2 on the inside surface, assuming 20 V anode loss.
Regenerative cooling of anodes (at specific impulse values in the MPD regime) is possible
only with hydrogen or with alkali metal propellants, notably lithium. In the latter case, the ideal
anode operating mode would be evaporation and ionization of the propellant on the porous or
wetted anode surface, resulting in increased ion current fraction, reduced anode fall and utilization
of part of the anode loss energy. Otherwise, where higher current densities are required or desirable,
additional anode cooling on larger thrusters can be achieved with built-in heatpipes or with a liquid
metal convective cooling loop.
Propellant Choice. The propellant choice affects the thruster, the system, the spacecraft, the
atmosphere environment and the operating logistics. If the thermal thrust portion is not taken into
account, the thrust to power ratio seemed to favor heavy propellants like argon, but also lighter
mixtures like N2H4 and NH3 have been taken into account, which also are storable and may be
logistically (N2H4) the most desirable with still fair thrust to power ratios. Further thruster tests and
system/mission studies are needed for final choices.
Liquid metals such as Li, K, In and others, have great advantages concerning storage,
cooling capacities, ionization potential but will probably be ruled out because of contamination of
the spacecraft by condensing and depositing of propellant on cool parts of the spacecraft.
The regenerative cooling capability is minimal for most propellants in the MPD regime,
except possibly for lithium and hydrogen. Lithium could remove ΔhRC ≈ 2.5 . 107 J/kg during
vaporization and heating, and hydrogen about 2.9 x 107 J/kg during heating (by ΔT ≈ 2000 K)
alone. Assuming for example η = 0.30 and ce = 30 km/s, the ratio
regen. cooling power 2η Δ hRC

=
input power ce2
would approach ≈ 2 to 3% for these two propellants. For continuous MPD thrusters of the
250 kW class, this regenerative cooling capability would be too small to be significant. For an MPD
in the 10 km/s ce regime with > 40% efficiency, the above ratio becomes > 23%, which with proper
design can be adequate for an almost full regenerative cooling.
Most other propellants have small to negligible regenerative cooling capability for the MPD
regime.
For storage weight and volume, the highly cryogenic, low density propellant hydrogen is of
course the worst. Hydrogen may be acceptable for missions requiring more or less continuous
propellant use from the beginning, sufficient that the boil-off keeps the rest cold enough and can be
used as propellant. In the longer future, hydrogen may be produced in space from water (where the
oxygen is needed also) or from same other compound.
The acceptability of cryogenic propellants generally will be very dependent on missions and
on future space storage facilities.
For possible propellant sharing with high thrust systems, and of course for emergency thrust
with insufficient or no electric power available, hydrazine stands out. Hydrogen also can give
emergency thrust with still reasonable specific impulse.
147
4.4.2. Magnetoplasmadynamic Thruster with Coaxial Applied Field
In this type of applied-field thruster, the electrodes are arranged coaxially like in an MPD
self-field thruster; there is a nozzle-shaped (Figure 74) and a cylindrical (Figure 82) type. By means
of a coil or with permanent magnets, an axial, diverging magnetic field Bf is produced as shown in
Figure 75. Here the magnetic field interacts with the induced, azimuthal flow which can be
explained with the Hall effect.
Figure 75 Principle of an applied-field thruster with a coaxial field.
4.4.2.1. Acceleration Mechanism.

The starting point for a discussion of the acceleration mechanism in applied-field devices with a
coaxial field is Ohm’s Law for plasmas. Neglecting the grad pe term results in
r
[ r r r
] ω τ r r
j = σ E + v × B − e e j × B . (1)
B
[ ]
The z-component Bfz of the applied magnetic field causes a ring current jθ in the direction of
θ due to the radial current density component jr according to the Hall term in Ohm’s Law. This ring
current, also called Hall current, in turn interacts with the magnetic filed and produces with the
component Bfr an accelerating force in the direction of the flow and together with the component
Bfz causes the pressure to rise on the thruster axis (pinch effect).
The formation of these Hall currents is therefore a prerequisite for propellant acceleration in
these devices. The ring current can only form if the individual particles seldom collide. A
characteristic parameter for these thrusters is therefore the so-called Hall parameter ωeτe; here ωe is
the gyration frequency and τe the mean free flight time of the electrons. The Hall parameter is
therefore the ratio of the gyration frequency of the electrons to their collision frequency. The higher
148
the gyration frequency and the lower the collision frequency, the better an azimuthal current can
form. That is why the Hall parameter must be as large as possible; in any case much larger than one.
4.4.2.2. Hall Parameter.

The dependence of the Hall parameter on the strength of the magnetic field, the pressure and
the temperature is easy to estimate. The gyration frequency of the electrons is connected to the
magnetic field by:
eB
ωe = , (2)
me
the collision frequency can be calculated as [Aweter-Kurtz, 1992]:
n e2
νe = e . (3)
σ me
Taking the equation pe = nekTe into consideration, the following results
pe e 2
νe = . (4)
k Te σ me
Because the electrical conductivity is proportional Te3/2 [Aweter-Kurtz, 1992], the result for the Hall
parameter is
σ B B Te5 / 2
ωe τ e = ∝ . (5)
e ne pe
This means that large magnetic fields, the lowest possible pressures and high temperatures
are necessary for forming azimuthal flows.
Estimates using this equation confirm that Hall parameters of several 103 can be anticipated;
however, measurements showed results between 3 and 5. [Aweter-Kurtz, 1997 Patrick, 1992] This
discrepancy can be explained by the instabilities that hinder the formation of the azimuthal flows.
This anomalous diffusion of electrons across the magnetic field [Janes, 1966] must be
seriously taken into account in an assessment of AF-MPD capabilities. (This is, after all, also true
for the SPT.) There are a number of consequences. For example, the magnetic nozzle is not ideally
effective (compare [Mikellides, 1995]), so that the plume mass is not fully contained within
boundaries given by the ‘anode magnet lines,’ that is, applied field lines passing through the anode
at its widest diameter.
4.4.2.3. Current Distribution

The magnetic field in coaxial applied-field thrusters is so large that due to the Hall effect, the
electrical conductivity is clearly reduced perpendicular to the magnetic field which results in the
transport of the current being hindered
σo σo
σ⊥ = ≅ . (6)
1 + (ω eτ e )
2
(ω eτ e )2
Therefore the current stream lines are extended far outside the thruster, as for example
measured by Schock [Shock, 1974] and shown in Figure 76. The influence of the applied magnetic
field is therefore also called “magnetic nozzle”.
149
Figure 76 Axial current distribution in a concentric applied-field thruster X9 from DLR Stuttgart. [Shock, 1974]
nΦo: lines of constant magnetic flow.
.
I = 80 A / m = 7 mg/s Ar
3.0 X 16
2
Thrust / current [mN/A]
B = 0.6 Vs / m
2.5 . .o
I = 60 A / m = 6 mg/s Ar mA ~
. ~ 0.85
m
2.0 X 13
. 2
I = 200 A / m = 40 mg/s Ar B = 0.2 Vs / m
1.5 o .
mA
. ~~ 0.5
m
1.0
.
I = 50 A / m = 10 mg/s Ar
0.5
0.0
0.1 0.5 1.0 5.0 10.0
Tank pressure [Pa]
Figure 77 Thrust per current as function of the tank pressure [Krüller, 1972] for experimental thruster X13 and
preflight model X16. [Krüller, 1975]
The discharge current, whose distribution is a function of ωeτe (= σB/ene), bulges out far
downstream as B/ne increases. This fact has been experimentally [Fradkin, 1973 Connolly, 1970
sovevy, 1991 Schock, 1972] and computationally [Krüller, 1972 Krüller, 1974 Tanaka, 1988]
verified. Figure 76 shows measured current distributions at different magnetic field strengths.
According to the preceding dependency, the same applies when density decreases. This suggests the
important influence of ambient pressure as demonstrated in Figure 77, where specific thrust (F/I) is
shown to increase as the tank pressure and, therefore, the gas density, is reduced. [Connolly, 1970]
At higher values of the critical parameter, saturation may occur. Still, environmental influence on
the overall process (participation of ambient gas) is given, leading to an uncertainty of thrust and
particularly Isp determination. Even with condensable propellants this effect can not be avoided
reliably because of the wide extension of the plume and the normally moderate dimensions of the
vacuum facility.
4.4.2.4. Movements of Charged Particles in E and B Fields.

In order to understand the behavior of an applied-field thruster, especially plasma rotation,
one must first consider the movement of charge carriers in E and B fields.
The equation of motion for charged particles is:
d r
dt
v =
q r r r
m
E + v × B (7) ( )
150
r r
In order to find a solution for this equation for the case of E perpendicular to B , the
following ansatz can be made: r r
r r E×B
v = v1 + (8)
2
r r B
Because E and B are constant in time, with Eq. (8) the following is true:
) ( )
r r r r
q ⎡r E × B × B⎤
dv1
= ( r r
⎢ E + v1 × B + ⎥ (9)
dt m⎣ B2 ⎦
Upon solving the double cross product, the result is:
( )
r r r r r r
q ⎡r r r r B⋅B B E⋅B ⎤
dv1
= ( )
⎢ E + v1 × B − E 2 + ⎥
dt m⎣ B B2 ⎦
and after considering the scalar products which appear, this remains:
r
dv1
dt
=
m
[
q r r
]
v1 × B (10)
The result here is a gyration movement with the gyration velocity ∇.
According to Eq. (8) and Eq. (10), the movement of the particles consists of a constant drift
with velocity vd and a circular movement. The drift is independent of the polarity sign of the charge
carriers; thus electrons and rionsr drift in the same direction and so it is therefore a plasma flow.
r E×B r r E
vd = and in the special case of E ⊥ B is v d = (11)
2 B
r B r r r
If E is not perpendicular to B , E has to be divided into a component Es perpendicular and a
r r r r
component E p parallel to B . The same as above then applies to the part Es : E p on the other hand
leads to a particle acceleration parallel to the magnetic field.
4.4.2.5. Rotational Frequency of the Plasma.

The components of the electrical field which are perpendicular to the magnetic field cause a
r r
plasma flow in the direction E s × B . For the applied-field thruster with a coaxial magnetic field this
means that the plasma begins to rotate. The size of the moment of momentum D can be calculated
as a closed integral without knowing the current density distribution. [Cann, 1966] If the radial
component of the magnetic field in the exit cross section of the thruster can be neglected, the
following results for a cathode are assumed to be punctual:
1 2
D = B I rA (12)
2
with rA as the anode radius.
If one considers the plasma as a rigid body, the moment of momentum can, on the other
hand, be estimated with
rA
1
D = ∫ ω r 2 dm& = m& ω rA . (13)
2
0
2
By equating the right side of this equation, one gets the following result for the rotational
frequency of the plasma
IB
ω = . (14)
m&
In several experiments, circulating current spokes were observed and their frequency was
measured. It was determined that Eq. (14) is a good reproduction of the dependency of the
151
rotational frequency of the plasma on the current, on the strength of the magnetic field and on the
propellant capacity. [Cann, 1966 Malliaris, 1968]
At DLR Stuttgart these results were compared with their own measurements and then
compiled (Figure 78 and [Maisenhälder, 1969]).
Figure 78 Rotational frequency of the plasma as a function of IB/(2π m

& ) [Maisenhälder, 1969]
From Figure 78 it follows that the rotational frequency is proportional to the expression In
Bk/ m& , whereby the exponents n and k lie between .5 and 1.
In addition it was determined that an explicit dependency on the molecular weight of the
propellant exists. Malliaris [Malliaris , 1968] showed that the rotational frequency is by a good
approximation inversely proportional to the root of the molecular weight. In this way one gets the
equation
I n Bk
ω ∝ with 0,5 < n, k <1 . (15)
m& M
4.4.2.6. Thrust.
The thrust of an applied-field MPD device increases significantly with the magnetic field
(Figure 83). It is not possible to derive a simple equation for this dependency as with self-field
thrusters. The thrust in an MPD thruster with a coaxial applied field essentially has four sources:
-The heating of the gas and expansion through a nozzle (dynamic effect on the material
walls like an arcjet). ->Ftherm.
-The discharge current crossing the magnetic field simultaneously results in an azimuthal
force component that puts the plasma into rotation, which is considered an important source of
energy addition. This energy can be partly converted into thrust energy. -> Fswirl
-As the discharge current crosses the applied magnetic field, azimuthal currents are induced
that yield axial and radial Lorentz (j x B) forces, of which the axial component directly accelerates
the plasma while the radial component confines the plasma and builds up a pressure hill,
respectively. This energy again can be partly converted into thrust energy. -> FHall
-The interaction between the radial component of the primary current and the induced
azimuthal magnetic field gives a self-magnetic acceleration. This plays only a subordinate role in
applied-field thrusters. -> Fself
152
The total thrust can be calculated as the sum of these thrust components Ftot = FHall+ Fswirl +
Ftherm + Fself
Whether this simple algorithm is applicable depends, however, on the definition of thrust
portions. Quantifying each portion as being generated by total conversion of attributed energy
added (e.g. Fswirl =∫ ( jr Bz uθ )dV ) to useful (axial) velocity, the following dependency is obtained
V
under simplifying assumptions [Sasoh, 1994 Arakawa, 1992]:
F = [( Fhall + Fself ) / 2] + {[( Fhall + Fself ) / 2]2 + F 2 swirl + F 2therm }1 / 2 . (16)

Using this formula, one can estimate the order of magnitude of the maximum effect
produced by each mechanism. Which mechanism prevails in thrust production depends on the
selection of operating parameters [Arakawa, 1992 Bishop, 1971] and thruster geometry. In Table
3.1 below, a comparison is made between the electromagnetic thrust portions for different cases of
B, I, and m& that proves this statement. There is evidence that Joule heating and swirl production are
energetically most relevant in the AF-MPD, with a substantial portion of the conversion taking
place in the magnetic nozzle.
The discharge current-equivalent mass flow m& equ = (mi / e) I (single ionization) where mi
is the ion mass, is far from being reached in known experiments [Krüller, 1971 Arakawa, 1992
Fradkin, 1970 Krüller, 1967] except in rare cases of light propellants. In the SPT, in contrast,
currents are low and voltages comparatively high, so that the m& equ requirement is met even with
heavy propellants, whereas Isp is still reasonably high at the expense of low thrust density.
In the ideal case of energy gain and conversion under special consideration of rotational
energy, voltage U (with a consequence to thrust F) should relate quadratically with the magnetic
field strength B as a result of back electromotive force (EMF) produced by azimuthal velocities.
This dependency is not found in the experimental results [Myers, 1993], see also Eq. (17), which is
proof that those energy processes are connected with inherent losses. There are different hypotheses
used to explain non-ideal performance. One is that plasma viscosity plays an important role,
hindering development of ideal azimuthal velocities by friction on outer (anode) walls, with
velocity conversion in addition being limited by anomalous conductivity effects. [Mikellides, 1995]
Plasma turbulence and anomalous diffusion certainly have a substantial influence, which leads to
very moderate effective ωeτe [Krüller, 1972 Krüller, 1974] as explained in section 4.1.2.
Thruster B [T] & [kg/s]

m I [A] Fhall [N] Fswirl [N] Fself [N]
-7 -2 -2
Univ. Tokyo 0.10 9 10 200 4.4 10 1.4 10 1 10-4
Los Alamos 0.19 2.5 10-5 350 1.8 10-1 6.9 10-1 9 10-3
Univ. Osaka 0.075 2.75 10-3 15000 4.0 22.0 22.0
Table XIV Operation conditions and calculated thrust components of some AF-MPD thrusters [Sasoh, 1991]
4.4.2.7. Experimental Evidence.

Thrust and discharge voltage generally tend to rise with the strength of the magnetic field
and the discharge current [Krüller, 1974]
153
F and Utot ∼ I 0.8-1 B0 0.5-1 ; (17)
see Figure 79 for thrust. Propellant mass flow has a more complex influence, whereas thrust
is hardly affected, i.e. only the aerodynamic part, voltage tends to go down as mass is increased, the
effect depending on whether the mass is fed in the vicinity of the cathode or the anode (flow
fractions m & A , see Figure 80). Thrust seems more dependent on m
& K, m & K , whereas voltage is more
influenced by m & A as shown in Figure 80. The role of m & A is not, or at least not primarily, to carry
part of the current as an ion current, but to guarantee a certain charge carrier density in the vicinity
of the anode, thus reducing anode losses and eventual strong deviation from discharge axisymmetry
[Maisenhälder, 1969]. Figure 81 gives anode losses as a function of mass distribution [Shall. 1972].
Figure 79 Thrust as a function of discharge current

and magnetic field strength (elevated power of
laboratory type X-9, 100 mg/s argon, pa = 3.0 Pa).
[Krüller, 1998]
Figure 80 Voltage as a function of anode and cathode

mass fractions (X-16, argon [Krüller, 1975]).
154
Figure 81 Relative anode loss as a function of
distribution of supplied mass [Shall, 1972], thruster
X13. Krüller, 1975]
As a consequence, for AF-MPD acceleration to be effective, at least a strong magnetic field

(B) of adequate shape, an optimal degree of ionization to ensure good coupling of mass to
electromagnetic effects, and moderate particle density in the discharge region are required.
4.4.2.8. Applicable Propellants.

The magnetic thrust of an applied-field thruster essentially depends on the formation of Hall
currents and for this a Hall parameter as large as possible is necessary. The Hall parameter is,
however, as shown above, independent of the propellant. Therefore, it imposes no additional criteria
on the propellant choice.
In view of the need for a high degree of ionization at moderate losses, certain propellants
suitable for AF-MPD engines can be considered. These are preferably easy to ionize monatomic
elements (with preferred materials underscored): noble gases (He, Ne, Ar, (Xe)) and alkali metals
[Fradkin, 1973] (Li [Fradkin, 1973 Fradkin, 1980 Tikhonov, 1995] (Na, K)), but H2, N2, and N-H
combinations have also been used, e.g., [Arakawa, 1992 Myers, 1993 Polk, 1992 Sasoh, 1988],
which suggests the capability of operating parallel to auxiliary chemical systems. The application of
alkali metals is combined with considerable feed system problems and the danger of S/C
contamination; the attainable tank pressure, however, is low in laboratory tests using condensable
propellants that have a self pumping effect. [Fradkin, 1980] Which propellant is optimal depends on
system considerations; this question may still be regarded as open.
Although metallic propellants can offer high efficiency, they are not favored today for fear
of contamination problems. Moreover, they produce only a small thermal thrust portion due to high
atomic mass. Considering the acceleration mechanism, a lightweight propellant appears preferable
for Isp gain. In addition, optimization studies have shown that with lithium the exit velocity is too
high for many missions. [Aweter-Kurtz, 2001]
4.4.2.9. Numerical Simulation.

The development of an efficient simulation process is of great importance for the
optimization of HF-MPD thrusters and should therefore be started. AF-MPD thrust depends on
plasma parameters and their distributions, which are a priori unknown, in contrast to
155
electromagnetic thrust of the self-magnetic MPD thruster, which is a function of independent
parameters. Consequently, computations are difficult, as a number of far-reaching assumptions have
to be made. [Mikellides, 1995 Krüller, 1972 Krüller, 1975 Tanaka, 1988 sasoh, 1991 Turchi, 1997
Thomas, 1991]
4.4.2.10. Thruster Developments.

Applied-field thrusters with coaxial fields were developed in the USA, Germany, and the
USSR in the 1960s and 1970s for power levels between 1kw-30kW The strength of the magnetic
field was between .1 T and several T. Averaged exit velocities of approximately 40 km/s at
efficiencies of up to 40% were achieved (the power to produce the magnetic field is not included in
the calculation). Most of the development projects were terminated in the middle of the 1970s, with
the exception of the USSR where limited work has been continued.
The typical average data for laboratory devices run at different locations are:
discharge current (I) = 100-200 (-1500) A; applied field (maximum) (B0) = 0.05-1.0Vs/m2;
propellant mass flow ( m& ) = 5-50 mg/s Ar, Li (He, Kr, Xe, H2, N2, NH3); ambient pressure (pa) =
10-0.05 (- 10-4) Pa; discharge voltage (U) = (50 -) 100-150 V; thrust (F) = 200-2000 mN; specific
impulse (Isp) = F/ m& = 15 - >35 km/s; and thrust efficiency (ηT) = F/(2 m& U-I) = <20 - 40%.
Efficiencies are reported as tending to be low for gaseous propellants, in particular argon (which has
been used for the majority of tests), in contrast to alkali propellants, especially lithium [Sovevy,
1991 Myers, 1993]. [Arakawa, 1969] and [Arakawa, 1969], however, report excellent results using
H2 and He. The propellant’s atomic mass seems to play an important role, and so does the magnetic
field strength that must exceed a certain limit in order to effectively support performance [Krüller,
1972 Arakawa1992] (see Figure 75). As stated previously, it is the overall operating conditions that
determine performance, and consequently, efficiency.
Data imply that a power range of 10-100 kW is possible, with operational voltage at an
unproblematic level. The devices work in a stationary mode. There are, however, voltage
oscillations (on the order of ± 5 - 10%) due to destabilizing effects typical for plasma accelerators.
This noise has to be considered with respect to its potential system impact.
The radiation cooled applied-field accelerator DFVLR-X16 developed at DLR in the early
70s, in which the magnetic field is produced by means of a coil, is from today’s point of view one
of the best devices. At the same time, NASA was examining a thruster with a superconductive
magnet. Both of these devices and the results achieved are discussed here as examples of this
technology. A more detailed description of the experiments on the AF-MPD thrusters is given by
Krülle. [Krüller, 1998]
Today, applied-field thrusters are once again being investigated in Japan, the USA and
Germany.
Thruster X-16 from DLR Stuttgart. The thruster which reached the most advanced stage of
development was X-16 (Figure 84) from DLR Stuttgart. Then in the mid 1970s the development of
plasma thrusters was discontinued because of a lack of missions and insufficient qualification
facilities. X-16 is equipped with a radiation-cooled anode and a hollow cathode. This device was
able to achieve a thrust of 251 mN with a magnetic field of 0.6 T at a current of 80 A and a voltage
of 145 V with 7 mg/s argon mass flow rate [Krüller, 1974]. This is equal to an effective exit
velocity of approximately 36 km/s at an efficiency of 38.8%.
156
Figure 82 Applied-field thruster X-16 from DLR Stuttgart. [Krüller, 1974
A similar device, X-13 from DLR Stuttgart, was used to investigate the dependency of the thrust on
the magnetic induction [Kurtz, 1971]. In Figure 83 the distinct influence of the magnetic field on
the thrust is apparent.
Figure 83 Thrust dependent on magnetic induction with X-13. [Kurtz, 1971]
NASA thruster with superconductive magnet. Among others, a radiation-cooled thruster

with a superconductive magnet was tested at the NASA Lewis Research Center (Figure 83).
One can see from Figure 75 that compared to self-field thrusters, the efficiency can only be
significantly improved when the strength of the magnetic fields is distinctly larger than 0.1 T. This
means that the effort of producing the magnetic field, which is not taken into account in calculating
the efficiency, seems to be only justifiable with stronger magnetic fields. Currently JPL (NASA Jet
Propulsion Laboratory) is building a test stand for a Russian applied-field thrusters (see below) with
lithium as propellant. This test stand will enable pre-development for thrusters up to several 100 kW
to be carried out.
157
Helium
Magnetic Field [T]

Superconductive Magnet
Dewar
Coil
Thrust Efficiency [%]

Thermal Insulation
Dewar
Magnatic Core
Tungsten
Propellant
Cathode
Graphite
Eff. Exhaust Velocity [km/s]
Figure 84 Thruster with a superconductive magnet. Figure 85

[Seikel, 1982] Dependence of the efficiency on the effective exit
velocity with different magnetic fields using the device
from Figure 83. [Seikel, 1982
Thruster developments in Russia. For several decades the Moscow Aviation Institute has
been investigating applied-field thrusters with lithium as propellant up to a performance level of
approximately 150 kW. A sketch of this device is shown in Figure 86. The cathode consists of a
bundle of hollow cathodes. The tungsten anode is radiation-cooled. At relatively low magnetic field
strength, (0.075 < Bf < 0.09 T), a thrust of ca 4.5 N was achieved at a current of ca 2.9 kA and a
mass flow rate of 107 mg/s. This is equivalent to an effective exhaust velocity of ca 42 km/s. The
efficiency was ca 48%. [Tikhonov, 1995]
Figure 86 Applied-field thruster with lithium as propellant up to ca 150 kW.
4.4.2.11. Necessary Test Facilities

When testing applied-field thrusters in a laboratory, one must keep in mind that compared to
self-field thrusters the current distribution extends quite far outside of the device (Figure 87).
However, interaction with the wall of the test container must be precluded. This means that the tank
158
must be made of non-magnetic steel. Due to the interaction of the discharge with the residual gas,
surrounding gas may be sucked in near the anode if the tank pressure is too high. Therefore,
pressures lower than 10-3 mbar are required (see Figure 87). Currently there are no facilities
available for testing thrusters at high electric power levels (higher than 100 kW). At JPL a test stand
for lithium thrusters is being built. However, it cannot be used for noble gas propellants due to
insufficient pump capacity. At IRS a new test stand has been built which is suitable for thrusters up
to 20 kW.
Measurement Krülle Ku,

40 mg/s A,
0.3 Tesla
[N]
Thrust
4 HIGH POWER HYBRID THRUSTER CONCEPT ATTILA

Tank pressure [mbar]
Figure 87 Thrust dependency on tank pressure in applied-field devices [Bishop, 1971 Krülle, 1974].
4.5. High Power Hybrid Thruster Concept ATTILA

Among the hybrid thrusters currently being investigated, only the concept ATTILA
promises a thrust density high enough so that high thrusts can be expected. The high power hybrid
thruster ATTILA (see Figure 88), which is under construction at IRS, consists of two stages: a
thermal arcjet section is followed by an induction stage. The plasma jet generated by the thermal
arcjet section, which is operated in a power range between 50 to about 100 kW, is injected into a
ceramic tube which is surrounded by a magnetic coil. In this stage additional energy is added by
inductive heating prior to the expansion in the nozzle (see Figure 89). [Laure, 2001] In the inductive
second stage the energy is coupled near the surface owing to the skin-effect. In this way the
temperature of the thermal arcjet’s cold gas casing can be increased considerably and ultimately a
significantly higher exit velocity can be achieved. In preliminary tests this effect has already been
proven by means of a baffle plate for force measurements. The initial experiments with the hybrid
plasma generator were very promising. They proved that in the second stage a significant increase
of the specific power can be achieved. Two experiments were carried out, each with 17 kW total
power. In the first test, the thermal arcjet was operated with 17 kW without the second stage and
with 0.9 N measured at the baffle plate. By dividing the power, 10 KW in the arcjet section and 7
kW in the inductive stage, the force increases by more than 20% to 1.1 N using the same mass flow
rate. Currently a baffle plate device for measuring the thrust is being developed further and
calibrated at IRS with the help of existing thrust balances. Tests with the current ATTILA
configuration are planned up to 200 kW total power. The optimal power splitting, maximal thrust
and specific impulse for different propellants will be investigated in the near future. The test stand
has already been built. At present a laboratory model is being developed.
159
Thermal arcjet Induction heating stage
Figure 88 Hybrid plasma generator
Figure 89 ATTILA operating in arcjet mode only. Figure 90 ATTILA operating in arcjet mode with
superposed induction heating.
4.6. Summary
Promising candidates for heavy payloads in space propulsion are different plasma
propulsion systems. Electrothermal arcjets cover a wide range of input power, from several 100 W
to several 100 kW [Aweter-Kurtz, 2002]. Low power devices have been implemented in
commercial applications with hydrazine as propellant because of its rather evolutionary step in
specific impulse compared to the well established chemical thrusters and its simplicity, reliability
and low system mass compared to other electric propulsion systems. This technology has been
investigated so far up to 100 kW power on a laboratory level and the results are very promising.
With hydrogen as propellant exit velocities could be achieved over 20 km/s and there is hope to
further increase this value by increasing the power level and optimizing the design [Aweter-Kurtz,
1998]. High thrust levels of 25 N have already been achieved in the laboratory with stationary self-
field MPD thrusters with argon as propellant [Aweter-Kurtz, 2001], but so far plasma instabilities
limit the exit velocity to 15 km/s for various investigated propellants including hydrogen. These
instabilities have been intensively investigated and there is hope to increase specific impulse by
changing the thruster geometry [Heiermann, 2002 Wagner, 1994]. High exit velocities of 40 km/s
160
and 4.5 N thrust have been achieved with an applied filed MPD device with lithium as propellant
[Tikhonov, 1995]. But there are some disadvantages: thrust density is lower by one decade
compared to the other two plasma sources, the performance is not demonstrated so far with non-
contaminating propellants in stationary operation at high power levels and much lower tank
pressure is required for the investigation of applied-field devices. A hybrid thruster, consisting of an
electrothermal arcjet as a first stage combined with an inductive RF-stage offers the possibility to
not only increase the exit velocity but also the thrust density. Preliminary laboratory tests look
promising.
Monika Auweter-Kurtz and Helmut Kurtz

e-mail: ep@irs.uni-stuttgart.de
Institute of Space Systems, Universität Stuttgart
Stuttgart, Germany
4.7. References
Schmidt, T., Seboldt, W., Auweter-Kurtz, M. “Propulsion Options for Manned Mars Missions”,
16_333_P, 6th International Symposium, Propulsion for Space Transportation of the XXIst
Century, Versailles, Frankreich, May 2002.
Shoji, J.M., “Solar Rocket Component Study”, AFRPL-TR-84-057, 1984.
Farbman, G., “Upgraded Nerva Systems: Enabler Nuclear System”, in: NASA Conference
Publication 10079, Nuclear Thermal Propulsion, Cleveland 1991.
Auweter-Kurtz, M., Gölz, T., Habiger, H., Hammer, F., Kurtz, H., Riehle, M., Sleziona, C., “High
Power Hydrogen Arcjet Thrusters”, Journal of Propulsion and Power, Vol. 14, No. 5, September-
October 1998, pp. 764-773.
Wegmann, T. “Experimentelle Untersuchung kontinuierlich betriebener magnetoplasma-
dynamischer Eigenfeldtriebwerke”, Dissertation, Institut für Raumfahrtsysteme, Universität
Stuttgart, 1994.
Tikhonov, V., Semenikhin, S., Brophy, J.R., Polk, J.E., “The Experimental Performances of the 100
kW Li MPD Thruster with an External Magnetic Field”, Proc. Of the 24th IEPC, Moscow, Russia,
1995
Christensen, J.A., et al., “Design and Fabrication of a Flight Model 2.3 kW Ion Thruster for the
Deep Space 1 Mission”, AIAA-98-3327, 34th AIAA/ASME/SAE/ASEE Joint Propulsion
Conference, Reston, VA, July 1998.
Laure, S., Boehrk, H., Auweter-Kurtz, M., “Design Optimization of the Hybrid Plasma Generator
System – ATTILA”, 4th Int. Symp. On Applied Plasma Science, Kyoto, Japan, Sept. 2003.
Chavers, G., Chang-Diaz, F., “Momentum and Heat Flux Measurements in the Exhaust of
VASIMR Using Helium Propellant”, IEPC-28, Proc. of the 28th IEPC, Toulouse, France, March
2003.
Archipov, B:A., Bober, A.S., Gnizdorr, R.Y., et al., “The Results of 7000-hour SPT 100 Life
Testing”, IEPC-95-39, Proc. of the 24th IEPC, Moscow, Russia, 1995
Kurtz, H., “Arcjet Thrusters”, Final Report – Propulsion 2000, IB-00-06a, Institut für
Raumfahrtsysteme, Universität Stuttgart, 2000.
Curran, F., Caveny, L., “Hydrogen Arcjet Technology Status”, IEPC-93-215, Seattle, WA, 1993.
161
PRIMEX Aerospace Company, Space Systems Capabilities & Technologies, 99-H-2512A,
Redmond, WA, 1999.
Riehle, M., Kurtz, H. L., Auweter-Kurtz, M., “Performance Evaluation of Regeneratively Cooled 1,
10 & 100 kW Arcjets”, IEPC-99-027, Kitakyushu, Japan, 1999.
Bromaghin, D. et al., “An Overview of the On-Orbit Results from the Electric Propulsion Space
Experiment (ESEX)”, IEPC-99-182, Kitakyushu, Japan, 1999.
Auweter-Kurtz, M., Lichtbogenabtriebe für Weltraumaufgaben, B. G. Teubner, Stuttgart, 1992.
Butler, G., Cassady, R., Hoskins, W., King, D., Kull, A., “Performance of Advanced Concept
Hydrogen Arcjet Anodes”, IEPC-93-211, Seattle, WA, 1993.
Aston, G., Aston, M.B. “Integrated Design Arcjets for High Performance”, IEPC-97-090,
Cleveland, OH, 1997.
Miller, S., Martinez-Sanchez, M., “Nonequilibrium Numerical Simulation of Radiation-Cooled
Arcjet Thrusters”, IEPC-93-218, Seattle, WA, 1993.
Butler, G., Kull, A., King, D., “Numerical Simulation of Hydrogen Arcjet Performance”, IEPC-93-
249, Seattle, WA, 1993.
Fujita, K., “Performance Computation of a Low Power Hydrogen Arcjet”, AIAA Paper 96-3183,
July 1996.
Kuchi-ishi, S., Nishida, M., “Thermo-Chemical Nonequilibrium Modeling for a Nitrogen Arcjet
Thruster”, IEPC-99-029, Kitakyushu, Japan, 1999.
Megli, T., Krier, H., Burton, R., “Plasmadynamics Model for Nonequilibrium Processes in N2/H2
Arcjets”, Journal of Thermophysics and Heat Transfer, Vol. 10, No. 4, 1996, pp.554-562.
Auweter-Kurtz, M., Kurtz, H., Schöttle, U., “Future Directions for Electrothermal Arcjet
Propulsion”, 8th International Workshop on Combustion and Propulsion, Rocket Propulsion:
Present and Future, Pozzuoli, Italy, June 2002.
Nentwig, G., et al., “Experimental and Theoretical Investigation on an Arc Heated Water Steam
Plasma”, 14th International Symposium on Plasma Chemistry, Prague, Czech Republic, 1999, pp.
2393-2398.
Lichon, P., Sankovic, J., “Development and Demonstration of a 600 s Mission Average Arcjet”,
IEPC-93-087, Seattle, WA, 1993.
Trifonov, Yu., Khodenko, V., Salikhov, R., Avatinyan, G., Rylov, Yu., “VNIIEM Activity in the
Field of EPE”, IEPC-95-010, Moscow ,1995.
Walker, Q., Hargus, W., Capelli, M., “Characterization of a Low Power Helium Arcjet”, Joint
Propulsion Conference, Cleveland, OH, June 1998.
Welle, R., “Space Propulsion Applications of Helium Arcjets”, AIAA Paper 97-0794, 1997.
Rybakov, A., Auweter-Kurtz, M., Kurtz, H., “Investigations of a 1 kW Class Arcjet Thruster with
Helium as Propellant“, AIAA-2002-3659, 38th AIAA/ASME/SAE/ASEE Joint Propulsion
Conference, Indianapolis, IN, July 2002.
G.L. Cann, State of the Art of Electromagnetic and Electrothermal Propulsion, Technion Inc.,
Report 07-040, prepared for Rockwell Int. Space Division, Irvine, CA, July 1977.
M. Riehle, H. L. Kurtz, M. Auweter-Kurtz, Performance Evaluation of Regeneratively Cooled 1, 10
& 100 kW Arcjets, IEPC-99-027, Kitakyushu, Japan, Oct. 1999.
162
Hammer, F., Riehle, M., Auweter-Kurtz, M., Kurtz, H., Schöttle, U., “Arcjet Development
Capabilities at the IRS: Experiment and Analysis”, SP-398, Proc. of the Second European
Spacecraft Propulsion Conference, Noordwijk, The Netherlands, May 1997, pp-693-700.
Maecker, H., “Plasmaströmungen in Lichtbögen infolge eigenmagnetischer Kompression”,
Zeitschrift für Physik, Bd. 141, 1955, pp. 198 – 216.
Jahn, R.G.,Physics of Electric Propulsion, McGraw Hill Book Co., New York, 1968.
Auweter-Kurtz, M., Boie, C., Kaeppeler, H.J., Kurtz, H.L., Schrade, H.O., Sleziona, P.C., Wagner,
H.P., Wegmann, T., “Magnetoplasmadynamic Thrusters: Design Criteria and Numerical
Simulation”, Int. J. of Applied Electromagnetics in Materials 4 (1994) pp. 383 – 401.
Wienecke, R., Zeitschrift für Physik, Bd. 143, pp. 118-140, 1955.
Suzuki, H., Kuriki, K., “Effect of Insulating Anode Hood on MPD Arcjet Performance”, AIAA
Paper 78-652, AIAA/DGLR 13th Intern. Electric Prop. Conf., San Diego, CA, April 1978.
Hügel, H., “Zur Funktionsweise der Anode im Eigenfeldbeschleuniger”, DFVLR-FB 80-30, 1980.
Malliaris, A.C., John, R.R., Garrison, R.L., Libby, D.R., “Quasi-Steady MPO Propulsion at High
Power”, NASA CR-111872, 1971.
Auweter-Kurtz, M., Kurtz, H.L., Merke, W.D., Schrade, H.O., Sleziona, P.C., “Self Field MPD
Thruster Investigations”, Final Report, Grant No. N 00014-87-G-0119 Plasma Thruster
Development, for ONR, 1988 (IRS-88-P-10).
Moore, R.A., Cann, G.L., Galagher, L.R., “High Specifie Impulse Thermal Are Jet Thruster
Technology”, Report SO90-Phase Final. Electro-Optical Systems, Pasadena, CA, June 1965. Also
available as AFAPL-TR-6S-48.
Schrade, H.O., Rösgen, T., Wegmann, T., “The Onset Phenomena Explained by Run-Away Joule
Heating”, IEPC-91-022, 22nd IEPC, Viareggio, Italy, 1991.
King, D.Q., “Magnetoplasmadynamic Channel Flow for Design of Coaxial MPD Thrusters”, Ph.D.
thesis, Department of Mechanical and Aerospace Engineering, Princeton University, 1981, also
available as MAE Report 1552.
Rudolph, L.K., “The MPD Thruster Onset Current Performance Limitation”, Ph.D. thesis, Nov.
1980, Princeton University, Princeton.
Lawless, J.L., Subramanian, V.V. “A Theory of Onset in MPD Thrusters”, AFOSR Report 83-033,
1983.
Boyle, M.J., Clark, K.E., John, R.G., “Flowfield Characteristics and Performance Limitations of
Quasi-Steady Magnetoplasmadynamic Accelerators”, AIAA Journal, Val. 14, No. 7, July 1976.
Kuriki, K., Suzuki, H., “Quasi-Steady MPD Arcjet with Anode Gas Injection”, in: Electric
Propulsion and its Applications to Space Missions, Vol. 79, edited by R.C. Finke, Progress in
Astronautics, AIAA, New York, 1981.
Heiermann, J. “Ein Finite-Volumen-Verfahren zur Lösung magnetoplasmadynamischer
Erhaltungsgleichungen”, Dissertation, Institut für Raumfahrtsysteme, Universität Stuttgart, 2002.
Hügel, H. and Krülle, G., “Phänomenologie und Energiebilanz von Lichtbogenkathoden bei
niedrigen Drucken und hohen Stromstärken”, Beiträge aus der Plasmaphysik, Bd. 9, pp. 87-116,
1969.
Cann, G., Harder, R.L., “Study of Electrode Attachment Regions in High Current Gaseous
Discharges”, Electro Optical Systems Report 3240-F, AF-AEDC-TDR, 1964.
163
Goodfellow, K.D., “A Theoretical and Experimental Investigation of Cathodes Processes in Electric
Thrusters”, Ph. D. thesis, University of Southern California, Los Angeles, 1966. Also: JPL internal
Doc. JPL D-13969, 1996.
Malliaris, A.C., “Phenomena in the Cathode Region of an MPD Accelerator”, AIAA Paper 67-47,
1967.
Hall, L.S., "Thermoelectric Effects in a Joule-Heated Plasma", Phys. Fluids 7 (1964), pp. 425-427.
Bez, W., Höcker, K.H., Mayser, B., “Der Anodenfall in Niederdruckentladungen”, Annalen der
Physik 18, 1956, pp. 335-344.
V. Engel, A., “A Theory of the Anode Fall in Flow Discharges”, Phil. Mag. 32, 1941, pp. 417-427.
Auweter-Kurtz, M., Glocker, B., Kurtz, H.L., Loesener, O., Schrade, H.O., Tubanos, N., Wegmann,
T., Willer, D., Polk, J.E., “Cathode Phenomena in Plasma Thrusters“, Journal of Propulsion and
Power, Vol. 9, No. 6, Nov.-Dec. 1993, pp. 882-888.
Wagner, H.P., Kaeppeler, H.J., Auweter-Kurtz, M., “Instabilities in MPD thruster flows: 1. Space
charge instabilities in unbounded and inhomogeneous plasmas”, J. Phys. D: Appl. Phys., 31,(1998)
pp. 519-528.
Wagner, H.P., Kaeppeler, H.J., Auweter-Kurtz, M., “Instabilities in MPD thruster flows: 2.
Investigation of drift and gradient driven instabilities using multi-fluid plasma flow models”, J.
Phys. D: Appl. Phys., 31,(1998) pp. 529-541.
Boie, C., “Numerische Simulation magnetoplasmadynamischer Eigenfeldtriebwerke mit
hochauflösenden adaptiven Verfahren”, Shaker Verlag, Dissertation, Institut für Raumfahrtsysteme,
Universität Stuttgart, 2000.
Auweter-Kurtz, M., Lecture manuscript: “Elektrische Raumfahrtantriebe”, Institut für
Raumfahrtsysteme, Universität Stuttgart, 1997.
Patrick, R.M., Janes, G.S., Pre-print, AFOSR Orbit-Raising Propulsion Planning Session, Orlando,
FL, 1982.
Janes, G.S. and Lowder, A.S., “Anomalous Electron Diffusion and Ion Acceleration in a Low
Density Plasma,” Physics of Fluids Vol. 9, No. 6, 1966, p. 1115.
Mikellides, P.G., Turchi, P.J., Roderick, N.F., “Theoretical Model for Applied Field MPD
Thrusters”, AIAA-95-2676, July 1995.
Schock, W., “Zur Verteilung der elektrischen Stromdichte in Magneto-Plasma-Dynamischen
(MPD)-Beschleunigern”, Dissertation, Universität Stuttgart, 1974.
Kruelle, G., “Theoretical Treatment of Current, Mass Flow, and Related Distributions in MPD
Plumes,” AIAA Paper 72-501, 1972.
Kruelle, G. and Zeyfang, E., “Preliminary Conclusions of Continuous Applied Field
Electromagnetic Thruster Research at DFVLR,” AIAA Paper 75-417, 1975.
Fradkin, D.B. and Roehling, D.J., “Thrust Stand Performance Measurements of a Lithium Fuelled
Applied Field MPD Thruster,” Proceedings of the 13th Symposium on Eng. Asp. MHD, University
of Mississippi, 1973.
Connolly, D.J. and Sovie, R.J., “Effect on Background Pressure on Magnetoplasmadynamic
Thruster Operation,” Journal of Spacecraft and Rockets Vol. 7, No. 3, 1970, p. 255.
Sovey, J.S., Mantenieks, M.A., “Performance and Lifetime Assessment of Magnetoplasmadynamic
Arc Thruster Technology”, Journal of Propulsion and Power, Vol. 7, No. 1, 1991, pp. 71-83.
164
Schock, W., “Diagnostics and Interpretation of the Electric Current Distribution in an MPD
Thruster,” AIAA Paper 72-498, 1972.
Krülle, G., “Zur Dynamik des axialsymmetrischen magnetoplasmadynamischen Beschleunigers
(MPD-Triebwerk) mit überlagertem Magnetfeld”, Dissertation, Technische Universität München,
1974.
Tanaka, M. and Kimura, I., “Current Distribution and Plasma Acceleration in MPD Arcjets with
Applied Magnetic Fields,” Journal of Propulsion Vol. 4, No. 5, 1988, p. 428.
Cann, G.L., Harder, R.L., Moore, R.A., Lenn, P.D., “Hall Current Accelerator”, NASA CR-54705,
1966.
Malliaris, A.C., “Oscillations in an MPD Accelerator”, AIAA Journal, Vol. 6, No. 8, S. 1575-1577,
1968.
Maisenhälder, F., Ungerer, E. and Kruelle, G., “Some Experiments on the ‘Spoke’ Phenomenon in
MPD Operation,” AIAA Journal, Vol. 7, No. 2, 1969, pp. 348-350.
Sasoh, A., “Simple Formulation of Magnetoplasmadynamic Acceleration”, Physics of Plasma, Vol.
1, 1994, pp. 464-469.
Arakawa, Y. and Sasoh, A., “Electromagnetic Effects in an Applied Field Magnetoplasmadynamic
Thruster,” Journal of Propulsion & Power Vol. 8, No. 1, 1992, p. 98.
Bishop, A.R., Connolly, D.J. and Seikel, G.R., “Test of Permanent Magnet and Superconducting
Magnet MPD Thrusters,” AIAA Paper 71-696, 1971.
Fradkin, D.B., Blackstock, A.W. et al., “Experiments Using a 25-kW Hollow Cathode Lithium
Vapor MPD Arcjet,” AIAA Journal, Vol. 8, No. 5, 1970, pp. 886.
Kruelle, G., “Characteristics and Local Analysis of MPD Thruster Operation,” AIAA Paper 67-672,
1967.
Myers, R.M., “Applied Field MPD Thruster Performance with Hydrogen and Argon Propellants,”
Journal of Propulsion & Power, Vol. 9, No. 5, 1993, p. 781.
Sasoh, A. and Arakawa, Y., “Thrust Formula for an Applied-Field MPD Thruster Derived from
Energy Conservation Equation,” IEPC 91-062, 1991.
Schall, W., “Influence of Magnetic Fields on Anode Losses in MPD Arcs,” AIAA Paper 72-502,
1972.
Krülle, G., Auweter-Kurtz, M., Sasoh, A., “Technology and Application Aspects of Applied Field
Magnetoplasmadynamic Propulsion”, Journal of Propulsion and Power, Vol. 14, No. 5, 1998, pp.
754-763.
Tikhonov, V, Semenikhin, S., Brophy, J.R. and Polk, J.E., “The Experimental Performances of the
100kW Li MPD Thruster with an External Magnetic Field”, Proceedings of the 24th International
Electric Propulsion Conference, Moscow, Russia, 1995, pp. 718-724.
Polk, J.E. and Goodfellow, K.D., “Ammonia Arcjet Behavior in a Cyclic Endurance Test at 10
kW,” IAF Paper 92-0612, Oct. 1992.
Sasoh, A., Solem, A. and Arakawa, Y., “10 kW Steady-State MPD Thruster,” Journal of the
Faculty of Engineering, University of Tokyo (B), Vol. XXXIX, No. 3, 1988, pp. 275-296.
Auweter-Kurtz, M., Kurtz, H., “Optimization of Electric Thrusters for Primary Propulsion Based on
the Rocket Equation”, J. of Propulsion and Power, Vol. 19, No. 3, May-June 2003, pp. 406-423.
Turchi, P.J., “The Effect of Magnetic Nozzle Configurations on Plasma Thrusters”, NASA CR
202341, May 1997.
165
Thomas, H., “A Numerical Simulation of Axisymmetric, Steady Plasma Flow with Self-Induced
and Applied magnetic Fields,” Ph.D. Thesis, University of Tennessee, 1991.
Arakawa, Y. and Sasoh, A., “Steady State Permanent Magnet Magnetoplasmadynamic Thruster,”
Journal of Propulsion, Vol. 5, No. 3, 1989, pp. 301-304.
Kurtz, H.L., “Integrale Messungen an einem axialsymmetrischen elektromagnetischen
Plasmabeschleuniger”, Diplomarbeit, Institut für Raumfahrtsysteme, Universität Stuttgart, 1971.
Seikel, G.R., York, T.M., Condit, W.C., “Roles for Magnetic Thrusters in Orbit-Raising Missions”,
SeiTec-Report 8203, SeiTec, Cleveland, OH, 1982.
Krülle, G., “Continuous MPDA Development Status”, Proceedings of the 3rd European Electric
Propulsion Conference, DGLR Fachbuchreihe, Band 5, Köln, 1974.
Laure, S., Heiermann, J., Auweter-Kurtz, M., Kurtz, H., “ATTILA – Adjustable Throttle Inductively
Afterburning Arc Jet”, IEPC-01-142, Pasadena, CA, 2001.
Wagner, H., “Theoretische Untersuchung drift- und gradientengetriebener Instabilitäten in MPD-
Triebwerksströmungen”, Dissertation, Institut für Raumfahrtsysteme, Universität Stuttgart, 1994.
166
5. A review of reactor configurations for space nuclear electric propulsion and surface power
considerations
5.1. ABSTRACT
Contained herein are results of a trade study (for Northrop-Grumman Space Technology)
that was performed for three reactor types coupled to three power conversion types for the Jupiter
Icy Moons Orbiter (JIMO) program of the National Aeronautics and Space Administration (NASA)
Lipinski, et. al . These results provide a conceptual benchmark for considering what reactors might
play a significant role for space power and in what power regimes each has either virtues or
liabilities. The three reactor types are: a Liquid-Metal Reactor (LMR), a Heat Pipe Reactor (HPR)
and a Gas Cooled Reactor (GCR). The three conversion types are Brayton, Stirling, and
Thermoelectric. As will be noted, a considerable range of information is necessary to be considered.
A key feature to the analyses conducted was that all concepts could be made equally safe. The
practical consequence of this condition was to place emphasis on such items as technical maturity,
cost, risk and fabricability. The approach is first to determine the mass of the three reactor types
when coupled to a specific power conversion system as a function of the thermal power of the
reactor. This is the dominant measure of performance. Then the various technical issues and
difficulties are reviewed for each reactor type. This is the dominant measure of schedule and cost.
From these analyses an understanding of the tradeoffs among the three reactor types can be
obtained.
5.2. Reactor introduction

The development and safe and effective implementation of the fission reactor module is a
key goal of any space nuclear power project. In addition to providing a significant level of power
reliably for over a number of years, perhaps ten years, the reactor module must satisfy numerous
requirements regarding nuclear safety and integration with the rest of the spacecraft while
simultaneously minimizing both programmatic risk and system mass.
It has been recognized from the earliest days of the fission era that the application of nuclear
fission energy to space power and propulsion needs to provide a number of advantages. The very
high energy density achievable with fissile fuel leads to compact, high power, long-lived systems
that have great advantages in space use. In recognition of these advantages, a number of programs
have been undertaken in the U.S. and in the former USSR to capitalize on these technologies for
space use. These programs have produced a wide array of technologies, many of them space-
qualified, that provide an excellent base on which to build a future space nuclear program. In fact,
approximately 35 space reactors have been designed and operated in space – all have been liquid-
metal cooled reactor concepts.
The SNAP program in the U.S. resulted in the launch and space demonstration of the SNAP-
10A reactor in 1965. This was United States’ first and only launch of a nuclear reactor into space.
SNAP-10A produced 43 kW of thermal power (“kWt”), was cooled with a potassium-sodium
(NaK) eutectic, and had a peak coolant temperature of about 820 K. It operated for 40 days before
an electrical fault on the Agena upper stage triggered an automatic shutdown. The Russian program
resulted in 33 reactors launched in space and provided clear demonstration of the feasibility of
using nuclear fission systems in space. The highest power level achieved was ~ 6 kW of electrical
power (kWe) for a period of under 1 year.
These programs and others, such as the SP-100 program, the ROVER/NERVA, Project
Timberwind programs and the TOPAZ program, created high temperature fuels, (UN, carbide fuels)
167
materials and other technologies that form the nucleus of any future space nuclear power
development program. In addition, a large number of study and evaluation programs funded by the
US DOE, NASA, and DoD have resulted in numerous plausible designs for a range of missions and
suggested many mission enhancing or enabling features of nuclear systems.
The proposed JIMO mission, and other robust space exploration related missions require a
significant extrapolation from the demonstrated fission reactor capability in space. A power level of
200 kWt – 2000 kWt is envisioned for an operational life of around 10 years. The required coolant
temperatures could be as high as 1350K. This is a challenging task. Fortunately, there has been
considerable progress in ground-based and sea-based nuclear reactor development and
implementation over the past four decades. This coupled with the above mentioned space nuclear
program database gives us a wealth of information and conceptual design options to choose among
as a starting point for the JIMO space reactor development.
This next series of sections describe the following aspects of the reactor module design,
development and implementation:
1. Top level requirements for the reactor module
2. A comparison of system masses for the three reactor types
3. Reactor operations and dynamic behavior
4. Discussion of nuclear safety issues and approaches to handling them
5. Technology readiness levels and overall assessment
6. Interface considerations with the rest of the spacecraft, and a
7. Summary and conclusions.
5.3. Reactor REQUIREMENTS

After extracting initial JIMO requirements, analytical efforts approached the space reactor
and power conversion system design process from the following perspective:
1. Nuclear safety must be assured for all mission phases
2. Power requirements: 50 to 300 kWe
3. Lifetime requirements: Capability of achieving mission with EP
4. Reactor fueling: Optimal – before or at launch site
5. Reactor integration: Optimal – before or t launch site
6. Launch assumptions – Several launch vehicle configurations considered
In order to perform the trade studies in a consistent manner, the high-level requirements
were translated to a set of detailed requirements that are necessary for design and trade studies.
Table XV, below, presents the detailed set of requirements. A number of the numerical values of
the requirements necessary for a detailed design were to be determined at this stage of the work, but
will need to be settled prior to start of serious designs.
A cumulative lifetime of 10 full power years was considered necessary to meet the projected
mission life and required duty cycle of the reactor. Several constraints on the reactor performance
dictated a set of key requirements, see Table XV. They are:
1. Operational Constraint: The need to remain critical at end of life (EOL), accounting
for uncertainty, sets the key requirements of keff (the reactivity coefficient: see also Section 5.3.2.2)
at EOL at 1.000 (+0.005 – 0.00).
2. Safety Constraint: In order to ensure sub criticality under the worst credible accident
scenario, the reactors are designed to have a keff = 0.985 under those conditions. Accident
conditions included immersion in water or wet sand and burial in dry sand. The keff value of 0.985
is appropriately conservative given the low probability of the event, the detailed analyses that are
performed to satisfy the requirement, and the test program that is planned to confirm the level of
sub criticality for simulated submersion configurations. It has been determined that reactor impacts
may result in considerable reactor core deformation, leading to large reactivity swings. These may
be more safety-significant than a simple ‘intact impact with original core geometry’ into either
water, wet sand or dry sand.
168
3. Resource Constraint: Past designs and the various reactor systems have used
different fissile fuel enrichment values. In order to ensure a level playing field for all reactor
configurations studied, the fissile enrichment of the fuel was set at 93 percent.
4. Technology Constraint: Recognizing the interest in avoiding significant fuel
qualification, the peak fuel burn-up was set at 4 (atom) percent. This is well below the peak value
demonstrated for SP-100. The data at the higher burn-up values, however, were sparse, and for
reasons of conservatism, the 4 percent value was chosen for our design. It is noted that at the low
power densities anticipated, larger diameter pins could be effective in reducing system mass without
loss of performance. A trade-off analysis will need to be performed to decide if the mass savings
warrants the extra fuel qualification expense; for the purpose of early studies, reliance on the SP-
100 fuel database has been assumed, and a 4 percent burn-up constraint imposed.
Category Requirement Spec Value

Reactivity Lifetime / Mission assurance 10 full power years
Provide Starting Function
Keff (EOL) 1.000+0.005/-0.000
Keff (BOL) = Keff (EOL) + Burnup reactivity + Temperature reactivity + Margins TBD
Safety Keff for maximum credible accident bounding case, i.e., reflectors gone, wet sand < 0.985
surrounding the reactor, internals flooded with pure water, fuel pin internals not
flooded
Assured safety for all mission phases TBD
Minimum shutdown element speed TBD
Safety monitoring parameters TBD
Control Operational for 10 yrs at full power with one stuck control element
Maximum control element speed TBD
Autonomous control system
Leakage control strategy (avoid in-core rods)
Max rod worth of one element TBD
Keff of all elements in least reactive configuration < 0.9
Control monitoring parameters N flux, T, P, position
Fuel Peak Central Temperature (UN) <1450K (TBR)
Peak Clad Temperature (UN) <1375K (TBR)
Enrichment 93%
Burnup <4%
Operational Power coefficient over mission life Negative, all powers
Number of scheduled shutdowns TBD
Restart time after unscheduled shutdown <TBD days
Shield Lifetime dose at 30 meters <25 krad
Lifetime fluence at 30 meters <1011 n/cm2
Shadow angle 5o x10o
Table XV Reactor Requirements
The other requirements reflect the standard constraints on shutdown reactivity, negative
power coefficient over entire power range, dose levels at payload, etc. It is anticipated that the
reactor will not need to be shut down (and subsequently restarted) once it attains initial criticality.
Finally, the results of transient accident analysis will develop requirements on the rate of
movements of the control blades (moderating fission) in order to ensure safety.
5.4. Reactor And Power Conversion - Mass Comparison

A model for reactor module mass was developed and used to compare the reactor types,
perform parameter studies of power and shield design, and later incorporated in the overall power
train model.
A trade study was performed for three reactor types: 1) liquid-metal cooled reactor (LMR),
2) heat pipe-cooled reactor (HPR), and 3) gas-cooled reactor (GCR). This trade study took into
169
consideration coupling these reactors to three conversion systems: 1) Brayton, 2) Stirling, and 3)
thermoelectric (TE). A trade study should consider performance, cost, schedule, and risk. This
section concentrates on performance; later sections describe technical readiness, ease of testing,
versatility, scalability, development concerns, and other characteristics, all of which influence cost,
schedule, and risk. Since the trip time to Jupiter is dictated most strongly by the thrust to mass ratio
of the vehicle, and since thrust scales with the electrical power, the dominant measure of
performance is the system dry mass for a given electrical power level available to the thrusters.
5.4.1. Earlier Work and Reactor Descriptions

Past studies have compared various combinations of reactor type and conversion type. The
trades have traditionally looked at performance first (as indicated by system mass for a given
power) and then tempered the conclusions with qualitative features such as versatility, growth
potential, or technological readiness. During the early days of the SP-100 program (1983),
Rockwell performed a technology assessment followed by a power system trade study. That study
concluded that the gas-cooled reactor directly-coupled with a Brayton power conversion system had
the highest overall figure of merit for the originally specified requirement set. But when the
requirements and weightings were adjusted, the liquid metal cooled reactor (LMR), with lithium
coolant, and a fairly high-temperature Brayton system was the combination selected, with an LMR
and thermoelectric option as backup. Later the LMR-thermoelectric combination became the basis
for the SP-100 program. The requirement change that led to this was a very strong emphasis on
minimum mass, and an additional emphasis on ability of the reactor to drive various conversion
systems that might be developed later.
The more recent DOE SPFT program report form the NASA/DOE Interagency Space
Reactor Power System Concept Screening Committee concluded that the preferred combination of
reactor and power conversion system changed as the power regime and allowed development time
changedi. The regime consistent with the JIMO requirements was called Level 3 and called for 100
to 250 kWe, coolant exit temperature >1200K, launch by 2015, power system specific mass of 25 to
35 kg/kWe, and scalability to multi-megawatts. For these conditions, the plausible combinations
were reduced to 1) LMR with Brayton, 2) LMR with Rankine, and 3) GCR with Brayton (direct
drive). Figure 91 shows the summary result of voting by the committee members. There was clearly
a strong difference between the favored three combinations and the remaining options, The
committee did not offer GCR driving Rankine, Stirling, or thermo-electric as favorable options.
The present study follows upon the earlier work and focuses on LMRs, HPRs, and GCRs.
Configurations for these three reactor types were developed in the DOE SPFT program and these
were chosen as the basis for the comparisons in this study. Alternative configurations (such as the
SP-100 Generic Flight System for the LMR, or the NERVA-derived GCR) were also considered but
did not appear as attractive for this mission set.
170
10
Number of Affirmative Votes

9
8
7
6
5
4
3
2
1
CCB Ran Stir TE TI CCB Ran Stir TE CCB Ran Stir TE
LMR HPR GCR

Figure 91 Results for Level-3 system comparison from DOE-SPFT program screening study
A brief overview of the three reactor types is shown in Figure 93. All three reactor types use
uranium nitride (UN) fuel pellets with a Nb1Zr clad containing a Re (Rhenium) liner. It is also
feasible for these reactors to use UO2 fuel, but its lower density and thermal conductivity make UO2
less attractive. The LMR uses pumped lithium to remove the heat from the reactor to a heat
exchanger where it is transferred to the power conversion system. The HPR uses sodium-filled heat
pipes to transfer the heat from the reactor to a heat exchanger. The GCR uses a flowing helium-
xenon gas mixture to transfer the heat from the reactor to the power conversion system. With a
Brayton conversion system, the GCR does not use an intermediate heat exchanger. Rather the
Brayton working fluid passes through the core and the reactor heat is transferred directly to the gas
from the fuel pins.
All three systems control the reactor with external neutron reflectors rather than internal
neutron absorbers, similar to the SNAP-10A approach, as opposed to the SP-100 approach. For
larger reactor systems, internal control rods will be needed, and so they may be needed for
thermoelectric options, which tend to have low conversion efficiency. This difference in control
approach did not have a strong effect on the trade comparisons.
A detailed discussion of the various characteristics, pros, and cons of the three systems will
be discussed in a later section. But briefly, the LMR draws upon extensive LMR experience with
NaK and Na reactors (albeit at lower temperatures), is compact because of the excellent heat-
transfer capability of liquid metals, and is versatile in working with different conversion systems.
The HPR draws upon extensive design work and electrically heated testing, has passive heat
removal from the core via numerously redundant heat pipes, and has an accessible core (no pressure
vessel). The GCR draws upon the High Temperature Test Reactor experience (in Japan) with
prototypic temperature and pressure levels, has no freeze-thaw issues and few material interaction
concerns (because of the inert-gas coolant), and in the direct-drive configuration employ all- super
alloy outer boundaries (which simplifies testing and integration).
171
Nb1Zr structures Nb1Zr structures Superalloy
Li coolant Na heat pipes structures
HeXe coolant
HX
Gas
Separator
Induction No
Pump & TEM Heat
Radiator Pump & Exch.
Radiator No
Pumps
Radiation No
Shield Radiation Pumps
Heatpipe
Heat Exch. Shield
Inert gas
Reactor Reactor coolant
& BeO & BeO
Reflector Reflector
Peak clad: 1372 K Peak clad: 1240/1305 K
Peak clad: 1295 K
Peak fuel: 1250/1375 K
Peak fuel: 1392 K Peak fuel: 1322 K Thin pin/ Fat pin
Lithium Metal Cooled Heat-Pipe Cooled Gas Cooled

Figure 92. Reactor Options (All with External Control Elements)
5.4.2. Static Model Development and Validation
5.4.2.1. Model Basis and Overview

In order to compare the three reactor types across a broad range of conditions that included
various power levels, power conversion schemes, and shielding requirements, a simplified model
for the reactor module mass was developed. The reactor module was defined as including the
reactor element (sometimes called the core and reflectors), radiation shield, primary heat transport
system, reactor instrumentation and controls, re-entry shield, and superstructure. A model was
developed to estimate the mass of all these components as a function of reactor type, power
conversion type, reactor thermal power, radiation dose requirements, shield half angle, and other
parameters. The heat exchanger between the reactor and the power conversion system was included
in the primary heat transport system for the Brayton system, but not for the Stirling or
thermoelectric systems (because those systems include the heat exchanger as an integral part of
their hardware).
The reactor model was based on the RSMASS-D model by Al Marshall [Marshall, 1997].
This model takes into account criticality, fuel burn-up, heat transfer, and numerous other constraints
in estimating the mass of space reactor. It was developed for LMRs, but it was extended easily to
GCRs and HPRs by a judicious choice of input parameters. The shield model within RSMASS-D
was replaced with a similar model that could be solved in a closed form rather than iteratively.
RSMASS does not include the primary heat transport system, so a model for its mass was inserted.
Updated values for the instrumentation and controls, and for the re-entry shield, from the DOE
Special Purpose Fission Technology (SPFT) program (as reported in the DOE space reactor Design
Data Packages provided to this contract) were used to update the models for those items. Finally,
the superstructure was assigned and estimated mass of 5 percent of the other masses.
Table XVI lists the dominant input and outputs that can be passed between the reactor
module model and the other modules of the overall system model. Additional input parameters may
be found within the model. Many of these additional parameters are dependent on the type of
reactor and on the type of power conversion system to which the reactor is connected.
172
Primary Inputs Mass of reactor element (kg)
Power conversion type Mass of radiation shield (kg)
Reactor type Mass of Primary Heat Transport (kg)
Reactor Thermal Power (MW) Mass of I&C (kg)
Time at full power (yr) Mass of reentry shield (kg)
Number of full-power heat exchangers Mass of super-structure (kg)
Desired lifetime n fluence (n/cm^2) Mass of the Reactor Module (kg)
Desired lifetime gamma dose (rad) Electrical power for pumps (kW)
Distance from reactor to payload (m)
Time averaged xenon propellant mass (kg)
Xenon propellant ave diameter (m)
Shield cone half angle, x axis (degrees)
Shield cone half angle, x axis (degrees)
Table XVI Dominant Input and Outputs in the Reactor Module Model
5.4.2.2. Specific Design Anchor Points – Level Field.

This model includes several internal adjustable coefficients. Most of these are based on
actual physical parameters, but the fine-tuning of these parameters depends on the assumptions
made for operational and safety requirements. In order to anchor this general mass model, a set of
specific designs was developed for the three reactor types at two different power levels, 400 and
1500 kWt. These designs were offshoots of the SPFT DDP designs, but were adjusted to have
consistent modelling assumptions, because the mass of a reactor system depends on the design
assumptions and constraints.
For this study, the dominant assumption or constraint was that all the reactors would be
designed to be equally safe, to first order. As will be discussed in later sections, the dominant
hazard involving the reactor is release of radiation to the public during a launch accident. Since a
space reactor is not very radioactive before it is operated, this hazard is greatly mitigated by
designing the reactor so that it remains sub-critical for all credible launch accidents. Previous
studies have shown that one of the most challenging credible launch accident conditions for a
reactor to remain sub-critical under water immersion conditions is for all the internal coolant
channels to be flooded with pure water and the outside to be surrounded by wet sand. Fuel
compaction under impact can be even more challenging. Water slows down (thermalizes) the
neutrons within the reactor and makes them more prone to induce fission; flooding the core with
pure water maximizes this effect without adding absorption by water impurities (such as are
actually present in sea-water or pond water). Surrounding the reactor with wet sand rather than
water is a worse case since wet sand reflects the neutrons back into the core better and without as
much absorption.
The scenario assumed was that the reactor falls into the ocean after a failed launch event and
either buries itself in the sand in the ocean bottom, or is covered over by sand after settling to the
bottom. For this trade study, the internal geometry is assumed to remain intact for all reactor types
considered. Experiments during the SNAP-10A development showed that external reflectors are
easily stripped off during most impact scenarios with water, so it was assumed for all these cases
that the external reflectors were removed. The reactor would be launched with the external
reflectors locked in their most open configuration, so this also was an expedient means to remove
fine differences in the amount of openness of the reflectors. Any neutron-absorbing block that
might be placed in the reflector gap for launch also was assumed to be stripped off.
A final design requirement was that the reactor remains operable to the end of life. This
means that the core must remain critical in spite of the burn-up of some of the uranium fuel, and in
spite of the reactivity reduction that occurs when the reactor heats from room temperature to
operating temperature. So, specifically, the design condition for reactivity (or “k-effective”, or “k-
173
eff”) for the accident condition of water-flooded and surrounded by wet sand was taken to be k-eff
= 0.985 for all reactor types (k-eff = 1 is critical; k-eff = 0.985 gives some margin; A k-eff change
of 0.015 is substantial for a reactor, but not excessive). The exact value of the level of sub-critically
is subject to some debate. However, we believe that our chosen value is the optimum number based
on conservatism and design and is adequate for a trade study since it would be applied to all the
reactors being considered. The reactivity loss during heat-up (or “temperature defect”) was
determined to be about 1.5 percent, and this amount was added to the calculated fuel burn-up for a
given power level, amount of fuel, and trip time (in full-power years) to determine the required k-
eff at start-up when cold. Typically, a reactivity swing needed between the wet-sand, water-flooded
condition and beginning of life cold-start-up condition was about 5 percent.
A very significant difference between the reactors designed during the DOE SPFT program
and the SP-100 program was the ability to meet these conditions without the use of internal safety
rods to absorb neutrons before start-up. This simplifies the design and eliminates penetrations of
‘thimbles’ into the hot reactor core and drive shafts through the radiation shield into the core center,
see Chapter 2. This design approach was used for all reactor types in order to be consistent (all three
types could also use in-core safety rods if desired or needed). Finally, the same reflector type was
used for all three reactor types. This consisted of sliding reflectors made of BeO to maximize
reflection with some neutron production and also to withstand high temperature (BeO was actually
slightly better than just Be).
Summarizing the design conditions for the anchoring points, the following criteria were
imposed:
9 Uranium enrichment is 93.15 percent U-235 (which is generally more available than
97 percent enriched fuel used for SP-100)
9 External sliding reflectors for control with BeO as the reflecting material
9 Reflectors extending axially to the mid-plane of the top and bottom axial BeO
reflectors within the pins, which were approximately 4 cm long
9 Pressure drop of He/Xe gas flowing through reactor or reactor heat exchanger is
about 3.5 percent of the absolute pressure (for the Brayton conversion system)
9 Reactivity at start-up while cold is 1.015 plus projected burn up for 10 full-power
years
9 Reactivity at beginning of life when submerged under wet sand with no reflectors
and flooded internally with water (and inside heat pipes but not inside the fuel pins)
is 0.985
9 Pressure vessel creeps less than 2 percent in 10 years at operating temperature and
pressure
9 Clad creeps less than 2 percent in 10 years at operating temperature and pressure
9 Peak clad temperature less than 1400K.
For all reactor concepts, the particular version of the Sandia code called FEPSIM was used
to model the gas flow pressure drop through the core or heat exchanger, the peak clad temperature,
and the vessel and clad creep. The number of fuel pins, size of fuel pins, size of coolant passages,
and pitch of pins were varied until a near minimum mass was achieved consistent with the
requirements listed above. To expedite meeting the reactivity requirement, the coolant volume
fraction in the core was adjusted to be about 16 percent, because experience with neutronics
calculations has shown that void fractions much higher than this causes excessive problems keeping
the reactor sub-critical under water immersion (and smaller void fractions result in excessive
pressure drops). An additional constraint that arose recently is the need to keep the turbo-machinery
shaft speed at about 45000 rpm or less. This imposes limitations on the gas pressure and helium
content. The GCR design was adjusted to meet this requirement also. The impact on the HPR and
LMR is expected to be smaller than for the GCR, provided the pressure is around 200 psi or greater.
174
The LMR version of FEPSIM was used to tally masses for the LMR case, but was not used
to optimize the heat transfer and pressure drops in the flowing lithium. Rather, the various sizing
runs listed in the LMR Design Data Package, developed during the DOE SPFT program by using
the Lockheed Martin code COROPT from the SP-100 program, were used and extrapolated to 1500
kWt. The extrapolated reactor geometry was then adjusted to meet the reactivity requirements.
The HPR version of FEPSIM was used to tally masses for the HPR case, but it does not
handle the heat transfer limitations of the HPR (which are a combination of axial limitations within
the heat pipes, and transverse conduction for a loss-of-heat-pipe scenario). However, there were
comments on scaling in the HPR Design Data Package that were used to estimate the geometry for
the 1500 kWt case. The extrapolated reactor geometry was then adjusted to meet the reactivity
requirements.
Table XVII shows a summary of the point-design reactor geometries and masses for the
GCR, HPR, and LMR at 400 and 1500 kWt using the SP-100 sized fuel pins. The GCR masses are
very high. That is because in this figure the fuel pins are forced to be close to the SP-100 pin
diameters. The SP-100 pins were sized for liquid metal-cooled reactors. Liquid metal has a
considerably higher density and higher thermal conductivity than a helium-xenon gas mixture and
so a LMR does not need as thick a coolant boundary around the pins as the GCR does. So a GCR
with SP-100 sized pins will have a larger total diameter than an LMR. The shield also would be
larger and heavier. That is why a GCR has higher system mass with SP-100 sized fuel pins. In
addition, any reactor that uses larger pins will have a larger fuel-pellet to clad ratio (if the clad
thickness stays the same), consequently, the reactor can be made smaller in diameter because of less
dilution of the uranium. The DOE SPFT team realized this and adjusted the pin size upward to a
more appropriate value. The baseline case in the DOE SPFT DDP for GCR has 313 pins with a pin
diameter of 1.06 cm for the 400-kWt case. With that pin diameter, the reactor module mass is 1567
kg.
Reactor Type GCR GCR HPR HPR LMR LMR

Enrichment 93% 93% 93% 93% 93% 93%
Thermal power, Brayton cycle (kW) 400 1500 400 1500 400 1500
Inlet pressure (MPa) 1.50 3.00 1.50 1.50 1.50 1.50
Number of fuel pins 757 757 381 610 673 727
Diameter of UN pellet (cm) 0.72 0.80 0.79 0.79 0.63 0.61
Diameter of pressure vessel (cm) 38.67 45.11 24.00 30.30 24.95 26.55
Outer diam of radial reflector (cm) 56.67 63.11 45.18 52.25 44.60 42.98
Reflector material BeO BeO BeO BeO BeO BeO
Length of active fuel in reactor (cm) 52.00 52.00 56.00 56.00 43.00 56.00
Length of the pressure vessel (cm) 104.17 110.61 64.00 64.00 65.09 97.04
Mass of the UN pellets (kg) 219 273 142 228 121 164
Reactor (kg) 867 1040 554 795 420 586
Radiation shield (kg) 739 1155 515 610 444 672
Primary heat transport system (kg) 41 58 228.0 328.3.5 148.9 428.8
Reactor I&C (kg) 140 140 140 140 140 140
Re-entry shield (kg) 65 80 57 57 40 50
Superstructure (kg) 82 113 65 87 51 84
Reactor module (kg) 1935 2586 1559 2018 1244 1961
Table XVII Summary of the Point-Design Reactor Geometries and Masses with SP-100-Sized Fuel Pins
This round of analyses showed that the GCR will pay a substantial mass penalty if it is
forced to use fuel pins that were sized for a liquid-metal cooled reactor. Larger pins are better. This
makes better use of the high thermal conductivity of the UN fuel and allows a smaller void fraction,
which helps maintain submerged sub-criticality. It was later determined that the use of larger pins
also reduces the mass of the LMR (while still meeting the other system requirements). The HPR
also could likely be improved by this change at low powers, but at higher powers it might run up
against another constraint. For the failed-heat pipe scenario, the HPR depends on thermal
175
conduction of heat through the pin and along the clad to transport heat from the fuel pins
surrounding the failed heat pipe to adjacent modules. Larger pins might exacerbate this problem and
lead to unacceptable clad temperatures and thermal stresses.
The UN fuel pellet diameter for the SP-100 General Flight Unit design was 6.48 mm. The
DDP LMR design assumed this UN pellet diameter, but the DDP designs for the HPR and GCR
assumed larger diameters, so already there was some assumed deviation from the established SP-
100 designs. Further mass reductions could be achieved with even larger diameters, as much as 2 or
3 times the SP-100 size. This raises the question of the validation level of the larger pin sizes.
Fortunately, UN fuel is very good at retaining fission gases without much damage. Comparison of
micrographs of indicated UO2 fuels (1 atom percent burn-up) and UN fuel (6 atom percent burn-up)
shows marked superiority of the UN fuel. This comparison suggests that the large diameter pins
should perform well, especially at the expected burn-up of 4% or less. No matter what diameter
pellet is chosen (even the established SP-100 size), it will need to be re-validated by an irradiation
test, because the fabrication line that makes those pellets will not be the same one that produced the
pellets that were tested (the original process line was dismantled at the end of the program). This
leaves open the possibility of validating two or more pellet sizes in parallel (the database for the SP-
100 baseline configuration with Nb1Zr clad and Re liner is fairly limited). It is recognized that this
approach will increase the fuel re-qualification time and cost, but it is anticipated that the
perturbation will not be very large.
Figure 93 shows a comparison of the cross sections of the original 313-pin GCR DDP
design and the smaller lower-mass version that uses only 85 pins with larger diameters. The larger-
diameter fuel (1.67 cm diameter) is smaller than the SNAP-10A fuel (3.1 cm diameter). SNAP-10A
used UZrH fuel rather than UN. Larger pins can be attractive from a handling and monitoring point
of view.
FEPSIM-GCR is an excel-based model that can be used to calculate the reactor
characteristics of temperature, pressure drop, material creep, system masses, fuel burn-up, etc. But
separate MCNP neutron transport calculations must be performed to determine k-eff for every
proposed geometry. This was awkward and severely limited the number of designs that could be
considered. So numerous MCNP reactor criticality runs were performed to systematically vary the
key GCR parameters and establish a data set of k-eff for various GCR configurations. From this set,
we determined a multi-variable quadratic formula for k-eff. This formula was inserted into FEPSIM
and with it we were able to create plots of reactor module mass vs. number of pins in the reactor, as
well as fuel pin diameter, Reynolds number, and peak clad temperature.
85-Pin GCR with larger pins

313-Pin DDP GCR but smaller pressure vessel
Figure 93 Comparison of Original GCR DDP and the Lighter-weight Option with Larger Pins
176
Figure 94 through Figure 96 show the results of parameter scans for a 400-kWe direct-drive
gas-cooled reactor system. Figure 94 shows the UN pellet diameter vs. number of pins. With fewer
pins, the pellet must have a larger diameter in order to have enough U-235 to go critical. Figure 97
shows the resulting peak clad temperature. SP-100 clad has been validated up to 1400K, and there
were some in-reactor tests run at 1500K clad temperature without failure. The figure shows that
acceptable clad temperatures are assured for high system pressures. But for lower pressures (where
the heat transfer is not as good), one must move to a greater number of smaller diameter pins.
Figure 96 shows the Reynolds numbers. The Re number approaches laminar regime for a large
number of pins and small resulting channels. Figure 97 shows the resulting reactor module masses
(including shield, controls, etc.). The large red circles show the cases that keep the clad temperature
at around 1300K (to leave 100K of margin) for lower pressures. Figure 98 shows how the reactor
module mass scales with pressure in the reactor. The red circles indicate the points that keep the
clad temperature below 1300K. It is not too sensitive until the mass approaches a cliff at about 150
psi. Figure 99 and Figure 100 show similar plots for various reactor power levels.
Table XVIII shows how the GCR system parameters vary for different pin diameters and
number of pins.
25 1 1=300:600 psi
Baseline: 400-kW
300:600 psia, 9 cm refl 2 2=150:300 psi
900-1150 K, 0.03 dp/p 3 3=100:200 psi
20 10% He mass frac 4 4=80:160 psi
25 krad, 1011 n/cm2 5 5=99% He
Diameter of UN (mm)
at 30 m, 7.5o shield 6 6=0.02 dp/p

SP-100
15
UN
10
5
SP-100 UN = 6.5 mm
SP-100 clad = 7.5 mm
0
0 200 400 600 800
No. of Pins
Figure 94 Diameter of UN Fuel Pellet vs. Number of Pins
177
1550
Baseline: 400-kW 1 1=300:600 psi
1500 300:600 psia, 9 cm refl 2 2=150:300 psi
900-1150 K, 0.03 dp/p 3 3=100:200 psi
1450
10% He mass frac 4 4=80:160 psi
1400 25 krad, 1011 n/cm2 5 5=99% He
Peak Clad (K)
at 30 m, 7.5o shield 6 6=0.02 dp/p

1350
1300
1250
1200
SP-100 clad validated at
1150 1400 K, tests successful
at 1500 K
1100
0 200 400 600 800
No. of Pins
Figure 95 Peak Clad Temperature vs. Number of Pins
14000
300:600 psia, 9 cm refl 2 2=150:300 psi
12000 3 3=100:200 psi
900-1150 K, 0.03 dp/p
10% He mass frac 4 4=80:160 psi
10000 25 krad, 1011 n/cm2 5 5=99% He
6 6=0.02 dp/p
at 30 m, 7.5o shield
Re Number
8000
6000
4000
2000
0
0 200 400 600 800
No. of Pins
Figure 96 Reynolds Number vs. Number of Pins
178
2400
300:600 psia, 9 cm refl 2 2=150:300 psi
2200 900-1150 K, 0.03 dp/p 3 3=100:200 psi
4
Reactor Module (kg)
10% He mass frac 4=80:160 psi

25 krad, 1011 n/cm2 5 5=99% He
2000
6
at 30 m, 7.5o shield 6=0.02 dp/p
1800
1600
No internal
1400 control rods
1200
0 200 400 600 800
No. of Pins
Figure 97 Reactor Module Mass vs. Number of Pins
2400 595 Baseline: 400-kW

451 9 cm refl
2200 313 900-1150 K, 0.03 dp/p
199
Reactor Module (kg)
10% He mass frac

85
2000 Mod 25 krad, 1011 n/cm 2
1800
1600
1400
No internal control
rods
1200
0 100 200 300 400 500 600 700
Reactor Pressure (psia)
Figure 98 Reactor Module Mass vs. Number of Pins and Pressure
179
1500
Baseline: 400-kW 1 1=400 kW
1450 300:600 psia, 9 cm refl 2 2=600 kW
900-1150 K, 0.03 dp/p 3 3=800 kW
1400 10% He mass frac
4 4=1000 kW
25 krad, 1011 n/cm2
5 5=1200 kW
Peak Clad (K)
1350 at 30 m, 7.5o shield
1300
1250
1200
SP-100 clad validated at
1150 1400 K, tests successful
at 1500 K
1100
0 200 400 600 800
No. of Pins
Figure 99 Peak Clad Temperature vs. Number of Pins and Power
2400
No internal control 1 1=400 kW
2200 rods 2 2=600 kW
3 3=800 kW
Reactor Module (kg)
4 4=1000 kW
2000 5 5=1200 kW
1800
Baseline: 400-kW
1600 300:600 psia, 9 cm refl
900-1150 K, 0.03 dp/p
10% He mass frac
1400 25 krad, 1011 n/cm2
1200
0 200 400 600 800 1000 1200
No. of Pins
Figure 100 Reactor Module Mass vs. Number of Pins and Power
180
757-pin 313-pin 85-pin
Thermal Power (kW) 400 400 400
He Fraction (mass %) 10.0 10.0 10.0
Pressure (MPa) 2.06 2.06 2.06
Pellet Diameter (mm) 7.0 10.2 17.8
Clad Diameter (mm) 9.2 12.7 21.1
Channel Width (mm) 0.64 0.92 1.56
Coolant Volume Fraction (%) 16.7 18.0 19.3
Pressure Drop Fraction (%) 3.0 3.0 3.0
Pressure Vessel Diameter (m) 0.373 0.332 0.290
Reflectors Diameter (m) 0.555 0.514 0.472
Gas-Clad Temp Diff (K) 61.7 101.5 209.9
Peak Clad Temperature (K) 1200.9 1234.0 1324.4
Peak Fuel Temperature (K) 1213.4 1257.5 1387.1
Reactor Module Mass (kg) 1870 1620 1377
Table XVIII Summary of GCR Parameters for Different Fuel Pin Sizes
5.4.2.3. Model Validation and Results.

Historic SP-100 design data were obtained to help validate the model for the LMR. In
addition, detailed designs at 400 and 1500 kWt were developed using the DOE SPFT designs as
starting points to anchor the model at those power levels for all three reactor types. Figure 101
shows a comparison of the reactor model for the LMR with the historical SP-100 design data. Only
the reactor is compared since often the mass of the primary heat transport system was not reported
in the literature, and the mass of the shield is strongly dependent on the radiation design
requirement and shield angle. Also shown are the four design points listed in the SPFT Design Data
Package for the LMR.
Figure 102 shows the reactor module mass (reactor, shield, heat transport, I&C, reentry
shield, and superstructure) vs. reactor thermal power for the three reactor types, assuming a Brayton
power conversion system; that is, the reactor heat transport system includes a heat exchanger
suitable for passing the heat to the working fluid of the Brayton system for the LMR and HPR. The
GCR is directly coupled to the Brayton working fluid. Also included are the detailed design points
for the three systems at 400 and 1500 kWt. For all of these systems, there is assumed to be only one
full-power heat exchanger (or equivalent). To support multiple rotating units, valving is used to
connect the backup rotating unit flow into the heat exchanger. This minimizes the mass impact of
using spare units. This is not the approach that we finally adopted; rather, we went to three separate
heat exchangers. So that results in higher masses than what is shown here, except for the GCR
direct drive system.
The module mass increases slowly with power for all systems up to about 1 MW. The
reactor mass is essentially fixed since it is initially criticality limited, and the slow increase is due to
increasing heat transfer requirements. Beyond 1 MW the reactor is burn-up limited. The
requirement of 10 full-power years of reactor operation causes about 5 percent burn-up at about 1.5
MWt. This is about the limit of the fuel-validated database. For higher powers, the reactor fuel mass
must increase linearly with power to keep the burnup from exceeding the demonstrated 4 percent (6
percent has been demonstrated in fuel with W clad). This results in a nearly linear increase in
module mass with power. Demonstrating higher burnup capability in the fuel could push this
limitation back. But one would still need to consider whether the resulting reactivity decrease with
time could be accommodated with external control elements.
181
10000
Model, 10-yr life
10-yr
Model, 3-yr life
Reactor Mass (kg)

DDP
SP-100 Based Designs
Effect of
1000 Burnup
3-yr
LMR, Reactor Only

100
0.1 1 10
Reactor Thermal Power (MW)
Figure 101 Comparison of Reactor Element Masses for Historic SP-100 Published Calculations, SPFT DDP
Numbers, and Present Model
10000
Full Reactor Module (Brayton HX)
Reactor Module Mass (kg)
(Rx+Shield+PHTS+I&C+Aerosh+SS)
10-yr life, 7.5 deg half ang, no int rods
Model, GCR-Big Pins
Model, HPR
Model, LMR-Small-Pins
Model, GCR-BP-NbHX
Model, GCR-SP
Pt Design, GCR-BP
Pt Design, HPR
Pt Design, LMR
DDP, GCR
DDP, HPR
External
Control
Add 10% "effective mass" for
Rods
electric power to LMR and GCR-
HX
1000
0.1 1 10
Figure 102 Comparison of Reactor Module Masses for Brayton Conversion and Various Configuration Options.
Figure 103 and Figure 104 show similar results for the Stirling and thermoelectric
conversion systems. The GCR uses large pins in these plots; for small pins options, consult previous
plots and adjust accordingly.
182
10000

Full Reactor Module (Stirling HX)
Model, GCR, Big Pins

Model, HPR
Model LMR, Small Pins
1000
0.1 1 10
Figure 103 Comparison of Reactor Module Masses for Stirling Conversion and Various Configuration Options
10000
Full Reactor Module (TE HX)
Model, GCR, Big Pins

Model, HPR
Model LMR, Small Pins
1000
0.1 1 10
Figure 104 Comparison of Reactor Module Masses for Thermoelectric Conversion and Various Configuration
Options
5.4.3. A Simple Shield Model

The Monte Carlo reactor and shield neutronics code MCNP was run numerous times for
various configurations of shield materials. Figure 105 shows the fundamental configuration. From
this data base a simple formula was developed to model the amount of gamma dose and neutron
fluence attenuation that would result from a given shield configuration. This, combined with mass
formulae based on geometry, yields a model for shield mass vs. radiation attenuation and cone
angle for the zone of protection. The coefficients are summarized in Table XIX. They can be used
to estimate the thicknesses of the Be, W, and LiH.
183
LiH
Be
Reactor Core
Reflectors
Figure 105 Shield Configuration for the Dose Model
Knowing the effective attenuation coefficient allows development of a simple shield model
Material Neutron Attenuation Gamma Attenuation Density
Be 5.8cm 35cm 1.850 g/cc
W 5.2cm 2.0cm 19.30 g/cc
LiH 5.6cm 33cm 0.757 g/cc
Table XIX Attenuation Lengths (1/e) for Fast Neutron Fluence and Gamma Dose
The attenuation lengths (where the flux falls by a factor of “e”), combined with the fact that
the flux falls of with distance squared, yields a simple formula for the fluence and dose behind a
conical shield if there are no other structures outside the shadow of that shield:
⎛ − LBe − LLiH − LW ⎞ Pr t i
Fn = An exp⎜⎜ + + ⎟⎟ 2 (1)
λ .
⎝ nBe λ nLiH λ nW ⎠ L p
⎛ − LBe − LLiH − LW ⎞ Pr t i
Dγ = Aγ exp⎜ + + ⎟ (2)
⎜ λ. λ λ ⎟ 2
⎝ γBe γLiH γW ⎠ L p
where An and Aγ are the fluence and dose coefficients (8.1E8 n/J and 0.30 rad cm2 / J); λnBe,
λnLiH, and λnWe are the attenuation lengths (1/e) for neutrons in Be, LiH, and W; λγBe, λγLiH and λγW
are the attenuation lengths for gammas in Be, LiH, and W; Pr is the reactor power; ti is the reactor
operation time; and Lp is the distance from the reactor to the payload.
If LBe is fixed and known and the desired neutron fluence and gamma dose are specified,
then these equations can be inverted to determine the length of LiH and W needed. Let Cn and Cγ be
defined as follows to include the attenuation effect of the Be and of distance, plus the conversion
from dose to full-power-years of the reactor:
⎛ − LBe ⎞ Pr t i
C n = An exp⎜⎜ ⎟⎟ 2 (3)
λ .
⎝ nBe ⎠ L p
⎛ − LBe ⎞ Pr t i
Cγ = Aγ exp⎜ ⎟ (4)
⎜ λ. ⎟ L2
⎝ γBe ⎠ p
Then the needed thicknesses of LiH and W are:
184
⎛ ⎛D ⎞⎞
⎜ − λ ln⎛⎜ Fn ⎞
⎟⎟ + λγW ln⎜ γ ⎟⎟
⎜ nW ⎜⎝ C n ⎠
⎜C ⎟⎟
⎝ ⎝ γ ⎠⎠
LLiH = λ nLiH λγLiH (5)
λnW λγLiH − λγW λnLiH
⎛ ⎛ Fn ⎞ ⎛D ⎞⎞
⎜λ ⎜ ⎟⎟ − λγLiH ln⎜ γ ⎟⎟
nLiH ln ⎜ ⎜C ⎟⎟
⎜ ⎝ Cn ⎠
⎝ ⎝ γ ⎠⎠
LW = λ nW λγW
λ nW λγLiH − λγW λ nLiH
(6)
From this information, and the assumed angle of divergence, a first estimate of the mass of
the shield can be made. Additional correction factors can be made for tapering the outer edges of
the W or LiH, but this requires guidance from 2-D modeling.
More MCNP neutronic calculations were run to explore the dose around the spacecraft. To
simplify things in these early stages, no credit was taken for the various structures in the spacecraft
and science package body. The shielding benefit of spacecraft parts will not be too strong since
neutrons and 1-MeV gammas from the reactor penetrate materials fairly readily. It was determined
that beyond the shield the lifetime neutron fluence fell of with the inverse of distance from the
reactor center to the second power, and that gamma dose fell off with the inverse of distance to the
2.3 power (because of view factors and the tapered gamma shield). It was also determined that at
the edges of the shadow projected by the shield the dose and fluence changed strongly with radial
position, and the rate of change was approximately exponential with radial distance (or angle). With
these insights, a simple radiation dose model was developed. Figure 106 shows the fall off of dose
with distance from the reactor for a 400-kWt reactor. Figure 107 shows a plot for lifetime gamma
dose (10 full-power years). Figure 108 shows the neutron fluence map. The maps show that the
radiation shield reduces the neutron fluence by a factor of 30,000, and the gamma dose by a factor
of 100.
1.E+08
1.E+07 n Fluence
2 6 2
1/z (10 n/cm )
1.E+06
Gamma Dose
1.E+05
(rads) 1/z2.3
J. D. Tucker
1.E+04
1 10 100
Distance from Reactor Center Along Axis (m)
Figure 106 Dose Fall-Off with Distance
Log of Neutron Fluence (n/cm2)

15
Distance from Axis (m)
10
1016
1015 10o shield
5 angle cone
1017
1011 n/cm2
0 1012 1011
L
3 10 20 30
Figure 107 Map of Lifetime Gamma Doses
185
Distance from Axis (m) 15
107
10
10o shield
106
5 angle cone
105
108
25 krads
0
3 10 20 30
Distance from Reactor (m)
Figure 108 Map of Lifetime Neutron Fluences
5.4.4. Shield Parametrics

The shield model was converted to Excel and included in the reactor module system model.
The model was exercised to determined trends. Figure 109shows the shield mass vs. cone half angle
for 800-kW thermal reactors with Brayton conversion. Each reactor concept exhibits various
virtues or liabilities associated with its particular design. The GCR is slightly larger, so the shield
mass is somewhat heavier. The reverse is true for the LMR. The HPR requires that the heat
exchangers be between the reactor and shield. This increases the distance between the reactor and
shield necessitating a larger shield in order to achieve the desired shielded protection area. Moving
sensitive payloads closer to the shield makes the shield thicker and more mass. Using an elliptical
shield (with half the angle transversely) reduces the mass slightly. Shield mass reduces when
reducing the gamma dose, see Figure 110.
186
2000
Radiation Shield (Brayton HX) L=10 m
Radiation Shield Mass (kg)

10-yr life, 25 krad & 1011 n/cm 2 at 30 m,
800 kWt
1500
GCR
HPR
LMR
GCR, L=10 m
1000 GCR, Elliptical
500
Elliptical Shield
0
0 5 10 15 20
Shield Half Angle (deg)
Includeseffect
Includes effectof
ofHX
HXhelping
helpingto
toattenuate
attenuateradiation
radiation
Assumesone
Assumes one100%
100%HXHXwith
withvalving
valvingto
tospare
spareconverter
converter
HPRhas
HPR haslargest
largestshield
shieldtotoaccommodate
accommodateHX HXbetween
betweenreactor
reactor&&shield
shield
Figure 109 Shield Mass for Various Configurations
Xenon or ammonia, both electric thruster propellants, can be used to augment the reactor
radiation shield. A simple model was developed to determine the integrated dose or fluence vs. time
for a shield augmented by a tank of propellant in which the propellant was consumed at a constant
rate. Figure 111 shows the resulting dose either with or without such an augmentation. The
horizontal axis is in full-power years so it is assumed that the propellant consumption rate is
proportional to the system power. The case is for an 800-kWt reactor, a 30-m boom, and 8000 kg of
xenon initially in a 2-m diameter cylindrical tank. The xenon attenuation coefficient was taken from
the shield model. The blue lines are for the augmented shield, and the magenta curves are for a
shield without using propellant. The two fixed shields were adjusted so that the final doses were 25
krad and 1011 n/cm2 at the end of 10 full-power years. The propellant consumption rate was
assumed to be such that 400 kg of xenon would remain at the end of 10 years (reserve). Under these
assumptions, the fixed shield would be 158 kg less mass for the augmented case.
187
800
Radiation Shield (Brayton HX)
Radiation Shield Mass (kg)

750 10-yr life, 25 krad & 1011 n/cm 2 at 30 m,
25 krad 800 kWt, GCR, 10-deg half ang
700
Reducinggamma
Reducing gammadosedoserequirement
requirementcan
can
650 save shield mass
save shield mass
600 Benefitreduced
Benefit reducedwhen
whenstructure
structureor
orXe
Xe
propellantshielding
propellant shieldingaccounted
accountedfor
for
550
Xe=0 kg
500
Xe=1000 kg
450
in 2-m diameter tank
400
0.E+00 2.E+05 4.E+05 6.E+05 8.E+05 1.E+06
Gamma Dose (rad)
Figure 110 Shield Mass for Different Dose Limits
1.E+11
1.E+10
o
1.E+09 800 kWt, 30 m boom, 10 half-ang
(n/cm ) or γ (rads)
(n/cm2)
8000 kg Xe in 2-m diam cyl tank
1.E+08
400 kg reserve at end
1.E+07
1.E+06 Shield mass savings = 690-532 = 158 kg
2
1.E+05
1.E+04 γ (rads)
1.E+03
1.E+02
0 2 4 6 8 10
Full-power years
Figure 111 Time Dependent Dose and Fluence Behind Xenon Propellant Tank
The attenuation coefficient for neutrons and gammas through ammonia propellant was
determined via MCNP runs. These coefficients were used to determine the integrated dose for
shields augmented with ammonia shielding. The results are shown in Figure 112 below. The mass
savings in the fixed shield for this case is 237 kg. The extremely low neutron fluences at early times
probably will not be achieved in reality because of leakage and scattering paths, but the final results
at high fluences are approximately correct.
188
1.E+12
(n/cm2)
1.E+10
(n/cm ) or γ (rads)
o
800 kWt, 30 m boom, 10 half-ang
1.E+08
8000 kg NH3 in 2-m diam cyl tank
1.E+06 400 kg reserve at end
2
1.E+04
γ (rads)
1.E+02
Shield mass savings = 690-453 = 237 kg
1.E+00
0 2 4 6 8 10
Full-power years
Figure 112 Time Dependent Dose and behind Ammonia Propellant Tank
5.5. Reactor operation startup

Accurate modeling of reactor transient behavior during operations and off-nominal
conditions is essential to obtain a detailed understanding of performance. This, in turn, is important
in ensuring public safety and mission assurance. Dynamic modeling also will lead to the design of
safe and reliable modes of operation. Development of a dynamic model was performed using Next
Generation of Space Transportation (NGST) Internal Research and Development funds; the details
of that development are reported in a separate report on those IR&D efforts. What is reported here
is the application of that model to the three reactor types to help in the trade study comparisons.
Detailed results for dynamic operations are not shown here, however, Figure 113 shows the results
of detailed start-up models on the time required to achieve full power operations with each reactor
concept, assuming a Brayton conversion system.
The dynamic modeling suggests that the Brayton system can start up very rapidly if the
structures are pre-heated with a temperature difference between the turbine inlet structures (e.g., the
reactor, or reactor heat exchanger and ducts, and the radiator heat exchanger). Operational
experience with Brayton units (such as auxiliary power systems) supports this suggestion. The
characteristic time scale is that needed to establish a pressure difference with the compressor and to
start cooling the gas locally. If the structures are pre-heated, it is expected that the Brayton system
will become self-sustaining in less than a minute of motoring the compressor (via the alternator).
So, the proposed start-up sequence for the reactor and turbomachinery is as follows:
• Bring the reactor to a critical state but at a very low power level such that any heat
generated can be removed by conduction and radiated to space. Nominally this is
less than a kilowatt of thermal power.
• For an LMR, use this low power to help melt the Li in the system. This must be done
slowly, over a period of hours, to prevent damage of the piping and vessel during
melt expansion.
• Slowly raise the reactor temperature to suitable temperature for Brayton startup. This
is around 500K or 600K and let the reactor and heat exchanger reach thermal
equilibrium. Again, the reactor power is fairly low at this point, nominally about a
kilowatt.
• Motor the Brayton compressor (via the alternator) to begin Brayton flow and
simultaneously start the flow in the radiator coolant pumps. Keep the load to the
PMAD (Power Management And Distribution) as low as possible at first, and adjust
189
as needed to keep the turbomachinery from over-spinning. The turbomachinery
should become self-sustaining in a minute or less. If the system fails to start, stop
motoring, recharge the start-up batteries via the solar arrays and try again. Increase
reactor power and temperature if needed.
GCR:
Start & check Achieve Pre-heat core Motor Brayton Increase reactor and
reactor controller criticality with fission to compressor to start Brayton power to first
and diagnostics: 700 K self-sustaining flow check point
0:00-2:00 2:00-4:00 4:00-6:00 6:00-6:02 6:02-6:30
HPR:
Start & check Achieve Melt heat pipes, traverse Motor Brayton Increase reactor &
reactor controller criticality sonic regime, and pre- compressor to start Brayton power to
and diagnostics: heat core to 1000 K self-sustaining flow first check point
0:00-2:00 2:00-4:00 4:00-10:00 10:00-10:02 10:02-10:30
LMR:
Start & check Achieve Partially melt Increase reactor Motor Brayton Start ALIP with Increase reactor
reactor criticality Li loops power and temp to compressor to Brayton power andand Brayton
controller and electrically or 1000 K to start TEM start self- increase reactor power to first
diagnostics: with reactor pumps & continue Li sustaining flow power to match check point
0:00-2:00 2:00-4:00 4:00-10:00 melt 10:00-13:00 13:00-13:02 13:02-14:00 14:00-14:30
Figure 113 Approximate Startup Timelines
5.6. Summary
Based upon detailed work for the JIMO program, it is likely that any of the reactor
configurations can be engineered. Depending upon the concept evaluation figures of merit, e.g., the
things that evaluators deem to be important, one concept can be easily selected above another. It
should be noted that no space reactor in this thermal or electric power range has been ever operated
before; consequently, it would appear prudent to select a concept with the fewest unresolved risks in
order to achieve operations within a feasible timeframe. For this reason, the GCR appears to hold
an advantage. It should also be noted that it was straightforward to meet all known safety criteria,
although there remain some uncertainties regarding impact criticality. Achieving most safety
requirements necessitated a modest increase in mass for the reactor system. However, to put this in
context, it may be estimated that the reactor and power conversion system could weigh perhaps
2200 kg. This is to be compared to a nearly equivalent mass for the radiator and associated
plumbing, and to a very significant mass for electrical power conditioning equipment and electric
thrusters. It would not be unreasonable to force the reactor design team to work very hard to
achieve a 200 kg mass reduction, while simply changing the radiator topology could reduce mass by
several hundreds to perhaps a thousand kilograms. Ultimately, this becomes a very involved
systems engineering process, the results of which may be highly dependent upon the end
application (mission).
There appears to be no reason why a nuclear reactor based power system in the 100-1000
kWe class could not be built, deployed and operated. A more significant issue will be maintaining
the necessary funding support to carry the project through fruition. Although not prescient, after the
initial draft of this report was written, NASA terminated the JIMO. This occurred for a variety of
190
reasons, but most likely was the large demand upon agency funding for projects such a Shuttle
Return to Flight, ISS assembly complete, Shuttle retirement, Hubble repair, on-going exploration
demands, the Crew Exploration Vehicle and the Crew Launch Vehicle. It is important to note that
several commercial firms are interested in pursuing nuclear electric propulsion for commercially
viable missions.
5.7. References
Greene, S, et.al., “Report on the Special Purpose Fission Technology program”, Oak Ridge
National Laboratories
Lipinski, R. J. Wright, S.A., Lenard, R.X., Suo-Antilla, A.J., Vernon, M. E.,

Marshall, A.C., Jablonski, J.A., and Helmick, P.H., “CRADA SC03/01673 Final Report:
Space Reactor Trade Studies and Conceptual Design”, Sandia National Laboratories, November 23,
2004.
Lipinski, R. J. Wright, S.A., Lenard, R.X., Metzinger, K.M., Nygren, R., Youchison, D.L.,
Viswanathan, S., Jablonski, J.A., Helmick, P.H., Beard, S.G., Humberstone, M., Rollston, L.R.,
Potter, D. L., Young M.F., Vehar, D., Berry, D. and Hanson, D. “CRADA SC03/01670 Final
Report: “IRAD Activities for Jupiter Icy Moons Orbiter”, Sandia National Laboratories, November
23, 2004.
Marshall, A.C., (1997), “RSMASS-D Models: An Improved Method For Estimating Reactor And
Shield Mass For Space Reactor Applications”, Report SAND91-7826, Sandia National
Laboratories, Albuquerque, NM 87185.
191
6. Nuclear Safety, legal aspects and policy recommendations
6.1. ABSTRACT
This chapter is divided into several sections. The first section covers a series of findings and
develops a set of recommendations for operations of space reactor systems in a safe,
environmentally compliant fashion. The second section develops a generic set of hazard scenarios
that might be experienced by a space nuclear system with emphasis on different methods under
which such a system could be engaged, such as surface power, in-space nuclear electric or nuclear
thermal propulsion. The third section develops these into test and analysis efforts that would likely
be conducted. Risk areas, with concepts generated with regard to frequency and consequences of
potential accident events should be a high priority items and should be started with the initial design
concept efforts. The fourth section identifies what probable technology limits might be experienced
by nuclear propulsion systems and the exploration limitations these technology restrictions might
impose.
6.2. Finding 1: Nuclear power and energy have significant roles in space exploration now,
and the future for nuclear power has exceptional potential for future space exploration
activities.
Many missions have been carried out the very existence of which is owed to the availability
of nuclear energy. There is little doubt that the information from Galileo, Cassini, Magellan, and
the Pioneer probes has been of inestimable scientific and even social value. In the case of the
Pioneers probes, a whole new regime of data has been opened by the so-called anomalous
acceleration effect. These probes all enjoy the use of the nearly inexhaustible energy of nuclear
fuel, most useful in areas deficient in solar flux, but utilitarian under other extremes as well. While
solar energy has played a major role in the development of space, there are certain areas where the
utility of solar power has achieved its limit, for example in the cases of Mars Exploration Rovers,
(MERs). These devices are performing magnificently, but their utility is limited by the available
power and sensitivity to dust, hence, the next generation of rovers, the Mars Science Laboratories,
(MSLs) will employ small Radio isotope thermoelectric Generators (RTGs) as a source of
electricity and thermal energy in order to significantly expand their utility envelope. Further, the
Space Exploration Initiative announced by President Bush will require surface nuclear power,
almost certainly nuclear electric propulsion (NEP), and potentially nuclear thermal propulsion
(NTP). NASA’s Project Prometheus, although with reduced funding, is completing Phase A
reaching toward a NEP spacecraft capable of navigating and orbiting the icy moons of the Jovian
system.
The Space Exploration Program, for example, may involve the employment of surface
nuclear power systems for enabling an exploration-rich lunar or Mars exploration program due to
the surfeit of solar energy in many locations, but also as a cost-effectiveness gesture. The many
competing factors in any exploration enterprise typically are traded against cost of the project. The
Human and Robotic Space Exploration effort in the U.S. is hampered in the early years by a set of
infrastructural and exploration-related budgetary demands, that include Shuttle Return to Flight, the
Crew Exploration Vehicle and its Return to the Moon Initiative, International Space Station
Assembly Complete, Hubble Robotic Repair, the on-going Mars Robotic Exploration initiative as
well as a large array of NASA Center support requirements. Thus, only a small amount of funding
192
will be available to support new exploration venues. What should the investment profile be, what
are the technical aggregates with the highest leverage in terms of safety and cost? How are these
factors to be weighed? These overarching systems architecture issues can enable the
implementation of an exploration-rich venue if correct investments in power technologies are made.
The seminal question is: why nuclear power systems for an exploration-rich venue? The answer
itself is many-faceted, but a few examples can be fashioned in the main. We have many competent
analyses indicating that the cost of a given crewed lunar mission can be reduced by ~50% (uplift
costs) if in-situ resources are used for only replacing the return propellant oxygen. This reduction is
possible because for a cryogenic system, approximately 85% of the propellant mass is oxygen, and
the mass necessary to get the payload off the Moon and on the return leg to Earth must be carried
from the surface of the Earth all the way to the lunar surface. The mass multiplier is significant,
yet, it could be even more compelling for Mars transportation. This considerable reduction applies
only to round-trip payloads, and for one-way missions other avenues must be investigated. A
second area is that of replenishing life support gases such as oxygen and nitrogen. Oxygen is used
at the rate of almost 1 kg per day per crew member, not including airlock or EVA loses. A crew of 6
on a 600 day mission consumes substantial, ~3600 kg oxygen. Additionally, each crew member
consumes approximately 3-4 kg of food and substantial quantities of water each day. These
resources must either be carried to the target world or generated in-situ. At the Moon and Mars,
copious quantities of water are bound up in various mineral formations; on Mars, subsurface water
may be abundant. In all cases, whether due to propellant production, food generation, life support
replenishment or normal system losses, large amounts of power is essential if for no other reason
than cost reduction. Thus, it appears that nuclear power must be used if cost-effective surface
exploration is to become a reality [Lipinski et al., 2005].
The utility of nuclear electric propulsion has been analyzed for cost reduction of commercial
satellite transportation and more recently for support of the lunar Mars exploration project. The use
of a commercially viable NEP system could reduce costs for transporting the cargo segment by
~50%; this is as significant a cost reduction for one-way payloads as the use of ISRU lunar oxygen
is for round-trip payloads. Nuclear thermal propulsion has similar advantages in crew safety by
dramatically reducing transportation time to Mars for a given Initial Mass in Low Earth Orbit
(IMLEO), a commonly employed metric for cost, since one of the most costly segments of the
exploration enterprise is in Earth-to-Orbit transportation.
6.3. Finding 2: In order for the great potential advantages of nuclear propulsion to be
realized, it must be perceived by a majority of the population to be safe.
The utility of nuclear power is almost axiomatic, however, nuclear power has a slightly
blemished reputation; only a very small fraction of this impeachment is deserved or justified. These
blemishes on an otherwise impressive history have created a perception amongst a small group of
anti-nuclear aficionados that all nuclear power in space is either unsafe or has some clandestine
military function. As with the introduction of many new technologies, nuclear power has
experienced a learning curve. Early in its infancy, the totality of nuclear system hazards were not
well understood, and some operational, administrative, design bases, and implementation
procedures did not appreciate the potential risks associated with its early use. Over time, and, in
some respects, as a consequence of some high-profile accidents, more well-defined procedures
began to be developed. The United Nations Committee On the Peaceful Uses of Outer Space
(COPOUS) generated the the statement of principle [Principle 3, COPOUS 2005] that has codified
the world body’s consensus on this subject. The United States has a well-established set of
regulations and policy guidelines including NSC/PD-25 (National Security Council Memorandum,
1977), the National Environmental Policy Act and the Interagency Nuclear Safety Review Panel
193
(INSRP) as mechanisms to ensure from a design basis and from an independent administrative and
operational safety review that the space systems are safe. Typically, the INSRP employs an
approach where are credible accidents are analyzed and the risks enumerated. Other international
Treaties and Principles also apply, including The Outer Space Treaty [United Nations, 1967],
United States Space Policy [National Security Council Memorandum, 1996], Return of Astronauts
and Objects from Outer Space [United Nations, 1968], The Registration Convention [United
Nations, 1975], the Non Proliferation Treaty [United Nations, 1995], and the Liability Treaty
[United Nations, 1972]. All of these have some bearing on what a state’s party can or cannot do
with respect to their activities in outer space. Some, such as the Outer Space and Liability Treaty
are actual Treaties, and in the U.S. carry force of law when ratified by the United States Senate,
some, such as the Astronaut Return are codified as agreements, others, such as the Principle 3 are
only conventions and are not ratified, and are therefore considered as guidelines. Finally, still other
documents, such as the nefarious Moon Treaty have only a few signatories and are almost never
recognized. are Non-credible incidents with high potential consequence are also assessed to
ascertain if some consequence or frequency limiting factor can be applied to the mission without
compromising science objectives.
The policies in the U.S have led to a very impressive array of safe, highly successful RTG
missions. While not all of the UN’s COPOUS principles are met by the letter, they are met in the
spirit of the US policy approach. For this reason, it is useful to review the applicable policies and
principles.
6.4. Finding 3: Existing policies and procedures are generally adequate to account for
requirements of public safety and environmental compliance. Some recommendations will
assist in clarifying the meaning of some of these procedures, principles and policies to aid in
the space systems engineering process.
6.4.1. History, Perspectives and Objectives

A total of 44 U.S. radioisotope power sources were deployed into space between 1961 and
1997. The US has experienced a few failures with RTGs, however some of the plutonium sources
have been recovered and the radioisotope containment system has worked as designed, preventing
the release of the radioisotope fuel. The United States also launched the SNAP-10A reactor into
space in 1965. Although the United States has carried out a number of space reactor programs since
1965, none of these subsequent programs led to a space-deployed reactor system and none
completed the launch safety review process. The former Soviet Union (SU) developed and flew
approximately 38 nuclear reactors for a variety of defense-related programs. Two of these missions
resulted in an accidental reentry due to failure of the end-of-life boost rocket engine. The former
SU also deployed a number of RTGs, one of which ended in a failure to achieve desired orbit,
namely the Mars 96 probe. The final impact site for this mission is not known, although it is
conjectured to have landed either in Chile or the south Pacific Ocean. In total, there have been
many more RTG missions than reactor missions world wide. As a consequence, particularly in the
West, the safety approach is not as well established for space reactors as it is for space-based
radioisotope sources. Space nuclear safety considerations for reactor systems differ significantly
from those for radioisotope sources.
The principal safety issue for a radioisotope source is the possibility of accidental release of
the highly radioactive fuel. For reactors, however, the radioactive inventory at launch is typically
very small and the principle safety concern is the possibility of an inadvertent criticality. The low
194
radioactive inventory is a consequence of the requirement that no pre-flight operation of the flight
reactor at substantial power should be accomplished, consequently, no fission product inventory
will be contained in the fuel at launch. Zero-power critical operation is allowed to calibrate the
reactor, if necessary. Zero-power operations produce effectively no fission product inventory.
Safety considerations for space reactors are also very different from those for terrestrial
reactors. For terrestrial reactors, much attention is given to the possibility of fission product and
actinide release into Earth’s biosphere following a postulated accident. A space reactor does not, in
general, pose a threat to Earth’s biosphere after deployment in high orbit. Operations of a reactor in
lower Earth orbits do present a risk, and mitigating effects should be followed in cases where the
reactor may be operated in lower Earth orbits. Additionally, a significant safety advantage can
result if space reactors are required to remain sub-critical (with no previous high-power operation)
until safely deployed in space. On the other hand, terrestrial reactors have not been subject to
postulated propellant fires and explosions or high-speed impact, although these scenarios are now
being investigated as the potential for terrorist threats has increased.
It is imperative that whatever complexion the space nuclear mission may assume, the
program makes a clear distinction between space-related safety considerations and other important
considerations. Here the scope of space nuclear safety is defined to include protection of the public,
workers, astronauts, and Earth’s environment from radiation or radioisotopes produced by or
contained in the space reactor system or components following the delivery of the reactor system to
the launch processing location. Although in the U.S., the Interagency Nuclear Safety Review Panel
(INSRP) does not address non-radiological safety issues for space nuclear systems, these issues are
included in this discussion as well as an additional point to be considered in the promulgation of the
benefits of space nuclear systems. Reactor system operational failures must be prevented to assure
mission success; however, an operational failure that does not threaten Earth’s biosphere or its
inhabitants is not considered a safety issue. In the recent past, several highly visible Mars probe
failures have led to a mistaken focus of mission success and safety, this must be avoided, because it
will lead to completely unnecessary and highly expensive mission assurance features being raised to
safety proportions. Safeguarding special nuclear materials and threats to the space environment,
and if surface nuclear systems are deployed, adequate precautions to protect the celestial surface are
also important considerations that must be addressed by the program. Some of these issues are
discussed later. Nonetheless, neither safeguards nor space environment protection issues are or
should be included as safety issues. It is necessary to make these distinctions to assure that safety
standards are not applied to programmatic manifestations that do not directly affect safety and to
assure that all safety activities are focused on safety issues.
Great care must be used to assure that the nuclear safety program addresses the unique
safety issues and approaches appropriate for space reactors. In addition to the safety guidance
provided by the mission sponsor, basic guidance has been established by an interagency Nuclear
Safety Policy Working Group (NSPWG) [Marshall et al., 1993] that can be adapted to many space
reactor programs. The original NSPWG guidelines were developed for a piloted nuclear propulsion
mission to Mars; consequently, the guidelines have been somewhat modified to apply to the overall
space exploration mission, including non-crewed civil space and commercial space missions that
might eventuate. These guidelines, summarized here, will serve as an overarching guideline for
nuclear safety planning for the JIMO space reactor system.
6.4.2. Safety Guidelines and Implementation.

The following recommendations should be considered to effectively implement a nuclear
reactor safety program that is perceived by the public to be rigorous. Some of these are codified by
the Nuclear Safety Policy working Group, but are modified in some ways in these recommendations
to be more transparent and utilitarian to the space system architect. Others have been added to
clarify more adequately parameters that might aid in mission design. These recommendations are
195
primarily for space reactors since there is widespread acceptance of RTG safety requirements and
policies.
• Recommendation 1: The reactor shall not be operated prior to space deployment, except for
low-power testing on the ground, from which negligible radioactivity is produced
The purpose for low-power testing is to assure that the reactor critical configuration is
within design and safety limits and to calibrate the control system. This guidance is already part of
the existing operating framework.
• Recommendation 2: The reactor system shall be designed to remain shut down prior to the
system achieving its planned orbit.
This guideline is part of the presently proposed operating protocols. Administratively, dual-
fault-tolerant criticality inhibits will be used for the reactor system. These can take various forms,
including positive control drive lock-outs and reflector drive blocks, for example. Consideration
should be given to providing some form of authenticated, two-man control logic with encryption
key firewalls to prevent hacking into the systems’ control scheme. Further, the same two-man
authentication logic ought to also be applied to overall spacecraft commands so that no potential for
unauthorized commands to change orbits, reenter or cause collision potential can be performed.
• Recommendation 3: Inadvertent criticality shall be precluded for credible accident

conditions. Inadvertent criticality can be assured through passive features (poisons and
safety rods) or through active disassembly as part of the launch system’s Flight Termination
System.
Design requirements and design features must be implemented to meet this criterion.
During a launch, a number of accident scenarios appear to be credible. The most credible accident
is a launch system failure. These occur at a rate of ~0.05-.1 of launches. During a launch system
accident, details discussed later, the reactor power system can be subjected to fireballs,
overpressures, propellant fires, impacts with hard surfaces, and immersion in water or burial in wet
or dry sand. Each of these possibilities must be adequately analyzed and designed so that the
reactor remains sub-critical during these insults. A preliminary analysis suggests the possibility that
for some highly contrived compaction accidents the reactor may quickly pass through a
supercritical condition with no significant consequences. If further analysis demonstrates that
momentary criticality does not compromise safety, the basic guidelines may be modified to include
this possibility.
Another, more important approach to resolving both safety and safeguards related concerns
would be to install conventional dismantlement approaches to the reactor system as part of the
launch vehicle’s Flight Termination System (FTS). This concept was investigated fairly thoroughly
by Sandia National Laboratories as part of the Jupiter Icy Moons Orbiter program for the Naval
Reactors Prime Contract Team. The FTS has a demonstrated reliability of 0.99999995 for the first
stage and 0.99999992 for the second stage. The concept here is that the reactor would be
energetically, but not in a nuclear sense, dismantled as part of the Flight Termination System
operations for a failed launch. It should be noted that launch vehicles are already considered a very
hazardous practice, and the FTS is designed to preclude a highly hazardous system from impacting
areas where any damage could result. If the reactor is split into 2 or more fragments, (depending
upon the amount of uranium present), it simply cannot become critical. This approach would most
likely address safeguards concerns, since there is insufficient highly enriched uranium to provide an
attractive target for diversion. Active disassembly eliminates any feasible concerns that the reactor
could become critical during a launch accident. It is important to understand that the unique
features of a reactor, (near zero radioactive source term at launch) enables this feature. This
196
approach is not applicable to RTG’s, which contain a very high source term at launch, and every
effort must be made to prevent the spread of the radioactive heat source.
• Recommendation 4: Radiological release from the spacecraft during normal operation shall
have an insignificant effect on Earth.
No credible mechanism has been identified that can result in radiological consequences on
Earth as a result of operations in space. This safety guideline has been expanded to include no
radiological release as a consequence of normal startup, operations, or shutdown whether the
reactor is in space or on the surface of another celestial body. Consideration to expanding this to
include contamination of other celestial bodies should be discussed.
• Recommendation 5: The consequence on Earth of a radiological release from an accident in

space shall be insignificant
Mission operation and mission planning for the proposed space program shall be performed
to analyze credible accident scenarios that could result in a radiological release to Earth from an
accident in space. The approach to this requirement is to minimize the potential for unplanned
reentries to an extremely low probability through orbit selection and highly reliable spacecraft
propulsion, emergency orbit raising capabilities and the spacecraft’s attitude control system. In
order to assist the spacecraft design process, consideration to stating definitive frequency of failure,
for example, that the probability of an unplanned reentry be < 10-7 would dramatically assist in the
space system design process. Such a minimal probability is likely adequate for all cases except
those involving the most grave of consequences. It is also necessary to carefully plan flybys past the
Moon on the departure leg since the Moon’s gravitational influence could dramatically perturb the
spacecraft’s trajectory. The latter condition is primarily a consideration for nuclear electric
propulsion systems with attendant low acceleration levels.
• Recommendation 6: Safe disposal of spent nuclear systems shall be explicitly included in

mission planning.
Until the present time, all space nuclear reactor systems have been operated in Earth orbit.
The Space Exploration program will present the possibility of space reactor systems departing Earth
orbit on a variety of mission scenarios, and also with sample return missions, the potential for
nuclear systems to reenter Earth orbit from lunar or interplanetary space missions. Because of the
experience with the former Soviet Union “Cosmos”satellites reactor systems leaking cesium and
NaK coolant resulting in substantive debris formation, disposal outside Earth orbits should be
considered if possible. Deep space disposal should be accompanied by detailed orbital analyses to
ensure that Earth orbit crossings are highly unlikely. Detailed orbit analyses should be performed
to ensure that gravitational anomalies from other planetary bodies, the moons of any planet that the
nuclear power system is orbiting, and the tidal forces of the celestial body being orbited or flown by
would not force the departure of the space reactor system out of from its orbit or depart significantly
from its intended flyby trajectory.
• Recommendation 6a: Disposal beyond Earth orbit in non-Earth crossing orbits should
be considered where feasible
• Recommendation 6b: The probability of Earth vicinity return of any disposed asset
should be of a timeframe consistent with the decay of fission products to about the level
of the actinides. This terminology is more well-defined in recommendation 6c.
197
• Recommendation: 6c: The total body burden source term of the in-space decayed fission
products should not exceed the approximate actinide body burden of the reactor.
The term: “ to about the level of the actinides”, contained in UN Principle 3 is an important
guideline, but is a vague recommendation; we recommend that the TEDE (Total Effective Dose
Equivalent) burden of the fission products be no greater than the reactor’s initial actinide TEDE
inventory.
• Recommendation 6d: Disposal by active dismantlement should be considered as a mission

option
• Recommendation 7: Planned or Intentional Earth reentry shall be precluded from mission

profiles at this time.
Planned Earth reentries of a space nuclear reactor system are proscribed from mission
planning at this time. The benefits of a planned reentry must clearly outweigh the associated risks.
• Recommendation 8: Both the probability and the consequences of an inadvertent reentry

shall be made as low as reasonably achievable.
Careful design, detailed deterministic analysis, risk analysis, and testing will be conducted
to meet this guideline. As a consideration, defense-in-depth to preclude any high probability
reentry should be applied. The inadvertent reentry condition applies primarily to either a fresh
reactor, whose reentry represents no radiological, but a safeguards consequence, or a previously
operated reactor, which represents little safeguards consequence, but potentially high radiological
consequence. In order to minimize risk, the probability for inadvertent reentry should be
minimized. The probability of inadvertent reentry is a function of many factors, including operating
orbit, mission scenarios, potential for MMOD damage, and defense-in-depth. The following
recommendations should be used as a guide to designers and mission planners:
• Recommendation 8a: A reactor that is operated in or above a Sufficiently High Orbit

(SHO) [approximately 800 km] should be configured to remain in that orbit for
approximately the lifetime of the fission products to decay in accordance with
recommendation 6c. If such an orbital lifetime cannot be achieved through spacecraft
configuration control then a reliable system (Πf<10-5) should be installed to boost the
orbit to an orbit whose lifetime is consistent with recommendation 6c.
• Recommendation 8b: A reactor power system operated below a SHO but above the 50%
of the SHO lifetime shall have a highly reliable emergency orbit raising system whose
integrated probability of failure, Πf, such that Πf<10-.6
• Recommendation 8c: A reactor power system operated below a 50% SHO lifetime, but
above 5% SHO lifetime shall have a highly reliable emergency orbit raising system
whose integrated probability of failure, Πf, is such that Πf<10-.7. This reliability level
can be generated through multiple independent systems
• Recommendation 8d: Routine operation of a space nuclear power system below 300 km
should be avoided if operationally feasible, periodic operations to 300 km may be
justifiable if risk is well controlled
• Recommendation 9: For inadvertent reentry through the atmosphere, the reactor shall
be analyzed to ascertain the highest likelihood reentry condition. It is desirable for the
198
reactor to remain essentially intact or shall result in essentially full dispersal of radioactive
materials at high altitude unless this consideration results in lower reliability or
compromises safety or operations in other ways.
This criterion may be required to cover a multitude of design, administrative and operational
considerations. The primary concern is of a space reactor utilizing a highly enriched uranium core
reentering and becoming a safeguards issue. This consideration, post-9/11, dominates almost every
other administrative and operational concern. Many different approaches are being considered to
address this issue, including the use of lower enrichment uranium reactors for space missions in
order to make the reactor less attractive regarding diversion. Of course, the longer the reactor has
operated, the less attractive as a diversion item it becomes, on the other hand, it represents a greater
safety hazard. The employment of active disassembly to guarantee dispersal at high altitudes
should be a serious consideration, since this approach would eliminate the uncertainties associated
with the reentry of the reactor as compromised by mission specific equipment. Studies of this
approach have been made in the past, (see Recommendation 11). Under such scenarios where
safeguards is not an issue, other requirements could render this condition of substantially lower
importance.
• Recommendation 10: The reactor shall remain sub-critical throughout an inadvertent

reentry and Earth impact.
This design requirement is design specific, but is an important space nuclear system design
requirement. A preliminary analysis suggests the possibility that for some highly contrived
compaction accidents the reactor may quickly pass through a supercritical condition with no
significant consequences. If further analysis demonstrates that momentary criticality does not
compromise safety, the basic guidelines may be modified to include this possibility. Again,
energetic disassembly as part of the Flight Termination System should be considered as a design
option. (See Recommendation 11).
• Recommendation 11: For any credible Earth impact scenario, radioactivity should be
confined to a local area to limit radiological consequences. A variety of passive, active or
combination of features may be employed to achieve this result.
In order to meet mission design requirements, accidental reentries of a nuclear reactor

should be kept to ~10-7 or below if the system contains a fission product inventory that is
substantially greater than the TEDE of the contained at launch actinides, If the reactor remains
intact throughout the reentry, care should be applied to limit the spread of radioactivity. There exist
a number of approaches to meeting the above criteria along with launch accident sub-criticality and
other major safety and nuclear safeguards concerns. The NERVA program postulated the explosive
disassembly, into small fragments, of the NERVA rocket engine, should a hot reentry be a mission
potential [Wackerle et al., 1962]. From a safety and safeguard perspective, the envelope of an
intentional disassembly should be considered seriously for all phases of flight. During the launch
phase, active disassembly precludes unintended criticality no mater how prone to criticality the
reactor might be in a given accident outcome condition. Further, a highly dispersed core represents
no safeguards threat because the enriched uranium is too highly diluted in the environment. During
reentry conditions the JIMO program agonized over how to ensure an intact reentry without
jeopardizing other aspects of launch safety or compromising operational performance. Further, it
became apparent that there was little control over the initial reentry configuration. These
uncertainties combined to result in a staggeringly large number of possible reentry accident
initiating conditions, each with a dramatically different and highly unpredictable outcome
condition. The problem was how to comply with recommendations 10 and 11 in a rigorous fashion.
199
The independent assertion that active disassembly is a feasible option was a watershed moment in
the initial concept design cycle. A high altitude explosive disassembly could likely guarantee
virtual complete dispersal into very small fragments that represent little to no public risk. In
essence, a completely independent variable has not been introduced into the design parameter
space. Design tools for explosive packages have achieved a high degree of fidelity, and test
facilities exist to validate the code results, consequently, the system can be designed to result in any
degree of dispersal of radioactive material from a reentering reactor system that might be required.
• Recommendation 12: For surface power reactor systems, the reactor shall not be operated
at significant power prior to landing on the surface of another celestial body unless the
mission planning cannot be readily accomplished otherwise
Surface nuclear reactor power systems may be landed on celestial objects where human
presence is extant or planned. To preclude potential for exposure to radiation or contamination by
fission products or other isotopes, the surface power reactor should not be operated prior to landing.
• Recommendation 12a: Surface power reactors where humans are or will be present
should be adequately shielded prior to operation so that radiation dose to individuals
routinely habiting the site is ~ 1 mrem/hr. This is well below the annual dose limit to
astronauts of 50 Rem/year, but still leaves margin for other sources of dose.
[Approximately 2m lunar regolith shielding should be sufficient for most cases, 3m
thickness is entirely adequate]. The space program should also thoroughly review the
data available on radiation hormesis to ascertain what optimum shielding levels might
be for space systems.
• Recommendation 12b: Surface power reactors should be located remotely from the site
of routine habitation to enhance shielding and to preclude adverse consequences of a
landing accident at the reactor site that releases radiation. [A nominal distance of 1 km
has been used in most analyses].
• Recommendation 12c: A rover-mounted radiation sensor should be included in a

regolith mover that maps the radiation fields at low reactor power levels and
supplementary shielding added to locations of inadequate neutron or gamma shielding.
• Recommendation 12d: A well-delineated no-occupancy zone should be enforced for

habitats, rovers or EVAs that will prevent inadvertent incursion into a reactor generated
radiation area.
• Recommendation 12e: If a nuclear reactor system is operated on the surface of another

celestial body where humans are not or will not be present, operations and end-of-life
disposal consideration should minimize the insult to the celestial system where feasible.
• Recommendation 12f: Initial surface nuclear reactors should be made as reliable a

possible, however, if maintenance is required, the a combination of robotic,
telepresence, or direct human intervention may be necessary, however, dose levels
should apply the ALARA principle during such situations
• Recommendation 13Any in-space nuclear power or propulsion system shall not enter within
a keep-out zone of any inhabited structure so that astronauts, crew or workers will not be
exposed to an enhanced radiation environment.
200
The above is intended as an administrative control over space nuclear systems operations. If
close proximity to a radiation source is unavoidable for abbreviated periods of time, then specific
approval will be required by the mission commander or other authority.
• Recommendation 14: Any space nuclear system returning to Earth orbit shall have positive
two-man, authenticated control established over the reactor and spacecraft with secure
communications links established. Any space nuclear system returning to Earth orbit shall
have all main and backup propulsion systems operational and there shall be no system or
component failures that could compromise a highly reliable SHO or above maintenance and
control capability.
Recommendation 15: Disposal of a space nuclear system that is operating on the surface of
another world shall be decommissioned and entombed to adequately comply with planetary
protection objectives.
6.5. Finding 4: The existing design, fabrication and test process, including safety analyses is
adequate for addressing all non-launch related safety and environmental issues for a space
nuclear reactor system; launch and space related protocols must be developed.
The terrestrial nuclear power industry in the West is well established, with well-understood
design, fabrication, test and transportation procedures. During recent studies and analyses, it was
determined that most, if not all of the design, fabrication test and transportation to the launch site
scenarios are adequately covered by existing rules, regulations, policies and requirements.
However, there exist new areas of risk once the reactor system has arrived at the launch site. For
this reason, we here investigate potential post-shipment scenarios and perform and initial screening
as to probability and consequence of that scenario. Potential post-shipment accidents that could lead
to the release of radiation or nuclear material are addressed in this section. The basic safety issues
for a space reactor mission are fresh fuel dispersal and accidental criticality. The probability of
accidental criticality is expected to be very small, and with active dispersal mechanisms can be
made virtually non-existent through all flight phases; nonetheless, the potential consequences of an
inadvertent criticality are expected to be far more significant than accidental fresh fuel dispersal,
consequently, the active dispersal approach is preferred. This, of course. Will result in dispersal of
fresh fuel during a launch phase accident. Although dispersal of fresh fuel does not entail a high
risk, the issue must be addressed in detail. Scenarios relating to radiologically-hot fuel dispersal are
not expected to be credible; however, these scenarios must be studied to assure that their probability
is extremely low, and that active dispersal techniques are included in the scenario analysis. The
sequences will be divided into pre-launch, a series of launch and deployment scenarios and post-
deployment situations including surface activities. Since credible accident scenarios for the launch
and deployment into Earth orbit are similar for both nuclear electric and nuclear thermal propulsion,
as well as for surface nuclear power systems, all will be treated in an equivalent way through these
phases of flight. Surface power systems will be differentiated during the landing and operations
phases.
201
6.5.1. Recommendation 16: The Space Nuclear Reactor Program Should Concentrate
on Major Post-Shipment Activity and Accident Categories
Figure 1 depicts the general sequence of events for a space system once it has arrived at
Cape Canaveral Air Force Station (CCAFS) or other launch site.
• Move to processing facility

Arrival at Launch
• Removal from shipping container
Site
• System installation at facility
Spacecraft Final • Post-arrival checkout

Assembly and • Fuel reactor (if applicable)
Checkout • Integrate reactor system
• Integrated system checkout
Launch • Integration with launch vehicle

Preparation & • Integrated launch stack checkouts
Countdown • Countdown operations
• Ignition
Early Launch • Liftoff
Phase • Clear tower and launch pad
Late Launch • Ascent to orbit

Phase • Placement in orbit
• Reactor startup, operation

Operational • Spiral out of orbit
Phase • Thrust to Jovian system, capture
• Jovian system maneuvers/operations
• Terminate operations/mission
Figure 114 Sequence of Events at the Launch Site
For these sequences, the principal types of possible accidents can be categorized as follows:
Fresh Fuel Dispersal:

9 Impact accidents with fuel or reactor assembly
9 Explosion-driven fuel dispersal
9 Overheating in fire environment
Accidental Criticality:
9 Compaction accidents
9 Flooding/reflection accidents
9 Movement of control/safety elements
9 Fuel loading accident
The scenarios for these postulated accident categories are discussed by mission phase in the
following.
202
6.5.2. Arrival at Launch Site: Possible Scenarios
After the reactor system arrives at the launch site the system must be moved to a processing
facility, removed from the shipping container, and installed at the facility for checkout and post-
arrival integration and testing. Possible accident scenarios that could lead to fuel dispersal or
accidental criticality during these manipulations are summarized in the following:
9 Transportation accident
9 Dropping while handling with crane, forklift, or manual handling
9 Impact by vehicle or heavy equipment, such as a fork lift or falling heavy
equipment
9 Storage facility or nearby vehicle explosion or fire
9 Terrorist activities
9 Impacts as a result of natural disasters (tornados, hurricanes, earthquakes)
Accidental Criticality: Compaction accidents

9 Dropping while handling with crane or forklift
equipment
Accidental Criticality: Flooding/reflection accidents

9 Fuel falls into moderating fluid as a result of transportation or handling accident or
natural disaster
9 Assembled unsealed reactor accidentally flooded by moderating fluid
9 Impact accident breaches pressure vessel followed by flooding
9 Accidental Criticality: Movement of control/safety elements
9 Impact as a result of transportation or handling accident or natural disaster causes
movement of control/safety elements
9 Explosion or fire causes movement of control/safety elements
6.5.3. Spacecraft Final Assembly and Checkout: Possible Scenarios

Possible scenarios for this phase are similar to the previous phase. Additional scenarios must
be considered if fuel loading takes place at the launch site.

equipment

9 Dropping while handling with crane or forklift
203
9 Impact by vehicle or heavy equipment during assembly
Accidental Criticality: Reactor fuel loading accident (If performed at launch site)
9 Improper fuel loading
9 Fuel elements dropped on fully loaded reactor
9 Control element withdrawn beyond critical during 1/μ (neutron lifetime
measurements in approach to critical experiments) measurements or criticality
testing
Accidental Criticality: Flooding/reflection or compaction accidents

9 Assembled unsealed reactor accidentally flooded by moderating fluid (if not sealed
during preceding phase).
9 Impact accident breaches pressure vessel followed by flooding
9 Reactor dropped or heavy load falls on reactor
Accidental Criticality: Movement of control/safety elements

9 Assembly mistake causing movement of control/safety elements
9 Explosion or fire causes movement of control/safety elements
9 Spurious or unintentional signal causing control or safety element movement
6.5.4. Launch Preparation and Countdown: Possible Scenarios

In addition to many of the scenarios present for the previous phases, propellant fire and
explosion scenarios must be considered for launch preparation.

equipment
9 Propellant explosion impact with launch pad
9 Propellant explosion shrapnel impact
9 Propellant fire, pressure vessel and fuel integrity loss, dispersal of fresh fuel
9 Propellant fire, collapse of launch vehicle, impact with launch pad

9 Reactor dropped or heavy load falls on reactor

9 Propellant explosion impact with launch pad, breach pressure vessel, subsequent
flooding by water or propellant
9 Propellant fire, collapse of launch vehicle, impact with launch pad, breach pressure
vessel, subsequent flooding by water or propellant
204
9 Impact with launch pad causing movement of control/safety elements
9 Propellant explosion, subsequent impact, or fragment impact causes movement of
control/safety elements
9 Propellant fire, collapsing impact causes movement of control/safety elements
9 Propellant fire, thermally induced movement of control/safety elements
9 Other types of explosions or fires causing movement of control/safety elements
9 Spurious or unintentional signal causing control or safety element movement
6.5.5. Early Launch Phase: Possible Scenarios

The early launch phase (from the US) begins with launch vehicle ignition and ends when the
projected impact point passes from land to ocean water.

9 Propellant fire, pressure vessel and fuel integrity loss, dispersal of fresh fuel
9 Launch Vehicle fall-back or tip-over impact
9 Command destruct and impact
9 Launch failure, sand impact

9 Launch Vehicle fall-back or tip-over impact
9 Command destruct and impact
9 Launch failure, sand impact, water impact

9 Propellant explosion impact with launch pad, breach pressure vessel, subsequent
9 Propellant fire, collapse of launch vehicle, impact with launch pad, breach pressure
vessel, subsequent flooding by water or propellant
9 Launch Vehicle fall-back or tip-over impact, breach pressure vessel, subsequent

9 Impact with launch pad causing movement of control/safety elements
9 Propellant explosion, subsequent impact, or fragment impact causes movement of
control/safety elements
6.5.6. Late Launch Phase: Possible Scenarios

This phase begins after clearing land and includes ascent as well as placement into orbit.

9 Deployment failure, ocean impact
205
9 Deployment failure, land impact
9 Deployment failure, reentry heating dispersal

9 Deployment failure, ocean impact, subsequent flooding by ocean water
9 Deployment failure, land impact, subsequent flooding by water
9 Deployment failure, land or water impact compaction reactivity increase

9 Deployment failure, impact, causing movement of control/safety elements
6.5.7. Finding 5: The Safety and Operations Phase for NEP or NTP Systems Should be
Developed so as to Maximize Possible Scenarios for Space Nuclear Reactor
Employment
Flight operations include reactor startup, rise to full power operation, electric propulsion or
direct thermal propulsion, departure from Earth orbit, in-flight operations, celestial system capture,
transfer trajectories, maneuvering, operations, return trajectories (if performed), Earth orbit capture
(if performed), subsequent mission performance and final disposal and decommissioning. Since all
activities subsequent to Earth orbit represent effectively no risk to Earth’s biosphere unless a return
trajectory is involved, we will not regard them as a safety issue. The possible scenarios for NEP are
shown in Figure 2, below, while operational scenarios for NTP are shown in Figure 115, below.
Single Multiple
Return Spiral Return Spirals
No Return
Departure
Spiral
Above SHO
C3>0 Below SHO
Deployment
Beyond SHO Disposal

Launch
Beyond Earth
Orbit
3
Figure 115 Possible Operational Modes for NEP
The distinction between the two systems is inherent in the difference between high thrust
and low thrust systems. Generally, NTP systems are considered high thrust and NEP systems are
206
considered low thrust systems. For purposes of this distinction, any total space system with a
thrust-to-weight ratio >10-1 g can be considered high thrust, in that the time over which ΔV added to
perform the thrusting maneuver is small compared to the time of an orbit. However, at this time, no
known electric propulsion systems fall into a high thrust category. It is also possible for an NTP
system to operate below 10-1 g, consequently, in order to reduce gravity losses, multiple perigee
burns may be necessary. These distinctions are reflected in Figure 116.
6.5.8. Recommendation 17: Hazards and Risks Definitions Should be Presented in a

Consistent Fashion Throughout the Space Nuclear Reactor Program
Ultimately, independent safety reviewers and the public are the consumers of safety and
environmental compliance documentation; therefore these documents should have a common
format that is readily comprehensible. To the reader familiar with the Cassini Final Safety Report
Analysis (FSAR) format, we have initially listed our definitions in the Cassini FSAR format. For
example, the Cassini FSAR has apportioned Pad/Post-launch credible scenarios into the major event
bins and configurations shown below. While this division is an acceptable approach, it is not
completely commensurate with the Hazards Tables that have been generally defined, however, we
employ this format as a transition to the hazards tables format that have been used thus far. First,
we define vehicle codes that represent the various phases of vehicle assembly and operations
through payload deployment.
Single Multiple
Return Burns Return Thrust
No Return
Departure
Burn(s)
Above SHO
C3>0 Below SHO
Deployment
Beyond SHO Disposal

Launch
Beyond Earth
Orbit
4/25/2001 3
Figure 116 Possible NTP Operational Modes
207
JIMO System Configuration Code Definitions
Vehicle JIMO System Configuration
Code
-2 Launch Vehicle, No Propellants (First Stage Main First Stage Strap-ons (if
used), Upper Stage(s), SP/NEP/NTP* Vehicle (Fueled) separate from launch
vehicle no Payload Fairing)
-1 Launch Vehicle, No Propellants (First Stage Main First Stage Strap-ons (if
used), Upper Stage(s), SP/NEP/NTP Vehicle (Fueled) separate from launch
vehicle in Payload Fairing)
0 Launch Vehicle, No Propellants (First Stage Main First Stage Strap-ons (if
used), Upper Stage(s), SP/NEP/NTP Vehicle (Fueled, Payload Fairing)
1 Launch Vehicle, Fully Fueled (First Stage Main First Stage Strap-ons (if used),
Upper Stage(s), SP/NEP/NTP Vehicle (Fueled, Payload Fairing)
2 Launch Vehicle, Fully Fueled (First Stage, Upper Stage, SP/NEP/NTP Vehicle
(Fueled), Payload Fairing), Post Strap-on separation
4 SP/NEP/NTP Spacecraft (Fueled, First and Upper Stage, Payload Fairing
Separation)
5 SP/NEP/NTP spacecraft and Upper Stage only
6 SP/NEP/NTP spacecraft Only
*SP= Surface Power, NEP= Nuclear Electric Propulsion, NTP=Nuclear Thermal
Propulsion
Figure 117 Launch Vehicle and Spacecraft Configuration Codes
These represent the major contributors of risk to the public.
Pad/Post-launch
Credible Scenarios
Accidental Launch Intentional Launch Trajectory

Vehicle Destruct Vehicle Destruct Failures
Figure 118 Division of Scenarios
6.5.9. Scenario Bins From Cassini FSAR Format

From the above generic binning, we establish a set of failures that can appear to result in
release of radiation or nuclear material as shown in the chart below (Figure 119). The failures and
formats are those used in the Cassini FSAR. While this particular division of activities will not be
used in our final hazards assessment format, it is indicative of our process. This chart relates the
above activity bins with launch vehicle configurations so that a credible risk analysis can be
performed.
Radiation or Nuclear Material Release Scenarios in Cassini FSAR Format

Scenario Bin System Configuration
2 1
208
Accidental Spacecraft/Launch
Vehicle Failure
Intentional Spacecraft/Launch
Vehicle Destruct
Trajectory Failure
Figure 119 Failure Modes in Cassini Format for JIMO Configurations
From the above, certain credible events can occur that could result in the release of radiation
or nuclear material. From the hazards tables previously submitted, we take the general category
series LV-XX-YY (covering nuclear spacecraft integration with launch vehicle to cover all pad
accidents and scenarios including accidental destruct under configuration 0 or 1.
Under the general hazards tables category series LV (mentioned above) and LA (Launch,
ascent and on-orbit activities prior to reactor startup), vehicle configurations 1 through 6 are
germane under both accidental and intentional launch vehicle failure/destruct, and trajectory failures
are covered. The hazards tables cover additional scenarios, including on-orbit and in-space
activities. We prefer covering activity bins in the hazards tables format, because the end scenario is
reasonably independent of the pathway to the failure analysis pathway, although the probabilities
associated with the failure may be different in the case of accidental versus intentional launch
vehicle destruct. Also, the scenarios as listed above are not independent, in view of the fact that
when a trajectory failure occurs, once the failure propagates beyond a given point, either an auto-
destruct or a manual destruct is initiated. Further, trajectory failures will likely have markedly
different failure probabilities for the launch, ascent and orbital deployment phase versus the on-
orbit/in-space portion of the activities. Therefore, we re-write the above table in the hazards format
formulation.
6.5.10. Conversion of Hazard Categories From Cassini FSAR Format To Hazards

Table Format
This alternative binning process results in fewer bins to assess and appears to document a
more operationally realistic approach. These scenarios/hazard bins represent JIMO system
activities until the vehicle is in orbit around the Jovian system. From a safety perspective, we stop
the hazards analysis process at that point, because the JIMO system cannot return to Earth in any
credible scenario. In each of the above scenarios, a category of failures can occur. These are listed
below with respect to launch vehicle/spacecraft configuration.
209
Conversion of Cassini FSAR to Hazards Tables Format
Scenario/Hazard System Configuration
Bin -2 -1 0 1 2 3 4
LV-XX-YY: JIMO X X X X
Integration with
Launch Vehicle
LA-XX-YY: X X X X
Launch, Ascent
and Orbit Insertion
Prior to Reactor
Startup
OO-XX-YY: On-
orbit and In-space
Activities after
reactor startup
Table 20 Hazards Table and SP/NEP/NTP Space System Configuration Codes
6.5.11. Failure Events That Could Result In Release Of Radiation/Nuclear Material

For each of the hazard bins, we now categorize a set of failures that could result in a release
of radiation or nuclear material. In these cases, the initiating events are listed in a nomenclature of
LV-XX-YY, where the XX is a two-letter designator for such incidents as Inadvertent Criticality, or
Radiation Exposure or Fission Product/Isotope Release, and the YY is the numerical identifier of
independent initiators.
SP/NEP/NTP System Configuration versus Pre-launch Scenarios

LV-XX-YY: System Configuration
JIMO -2 -1 0 1 2 3 4
Integration with
Launch Vehicle
Failure Mode
Reactor hard X X X X
surface impact
Reactor water X
impact
Reactor ground X X X X
impact
Reactor dry sand
impact
Reactor wet
sand impact
On-pad X
explosion
JIMO propellant X X X X
explosion
210
Accidental X X X X
criticality
Table XXI Credible Scenarios During SP/NEP/NTP Spacecraft/Launch Vehicle Integration Phase
JIMO System Configuration versus Pre-launch Scenarios

LA-XX-YY: System Configuration
Launch, Ascent and -2 -1 0 1 2 3 4
On-orbit Activities
Prior to Reactor
Startup Failure
Mode
Reactor hard X X X X
surface impact
Reactor water X X X X
impact
Reactor ground X X X X
impact
Reactor dry sand X X X X
impact
Reactor wet sand X X X X
impact
JIMO propellant X X X X
explosion
Accidental X X X X
criticality
Sub-orbital reentry X X X X
Orbital reentry X
Table XXII Credible Scenarios During Launch, Ascent
During integration activities with the launch vehicle, only configurations 0 and 1 are
feasible, and certain impact scenarios are eliminated due to CCAFS restrictions and limitations on
where integration activities are performed. Each of the above scenarios will have a group of
initiating events that must be analyzed separately.
6.6. Finding 7: Safety Assessment for Additional Risks Posed By Lunar and Mars Base
Mission Scenarios Indicate that Space Reactor Systems can be used Safely and Effectively on
the Surfaces of Other Celestial Bodies
Table XXIII identifies hazards by activity for the Lunar and Mars Base Power System. The
scenarios could apply in an equivalent way to other celestial bodies. Although the nuclear power
subsystem from a NEP system could be used, including the reactor, controls, payload shield,
primary heat transport and power conversion system in a fashion essentially unchanged, there are
several additions, primarily to compensate for the impact of lunar or Mars dust and materials as
well as atmospheric constituents on exposed high temperature materials. Also, there are numerous
other activities that can result in problems. There are differences between the lunar base power
211
systems in that while the lunar base system may only need to be singly contained, or with a super-
alloy system, simply contained to prevent lunar dust from contacting reactor moving parts, while on
Mars the system may (depending upon materials the system is made from) be hermetically sealed.
According to design considerations, the power system is to be located approximately 1 km from the
central base complex to reduce potential radiation levels. Routine human activity, rover activity
and regularly scheduled take and landing activities will occur. These activities represent initiating
events for safety incidents. The selected groups form areas that are new types of activities based on
an enumeration of potential hazardous actions.
9 Recommendation 18: Serious and competent design studies for the use of a nuclear
reactor for base-load power at a lunar and Mars base should begin immediately to
ascertain the power levels and duty factors that might be required for such a system.
There are many studies present about power requirements for a lunar and Mars base. In
order for the human exploration of space to become viable, it must be more cost-effective as well as
scientifically productive. The use of in-situ resources for life support closure and propellant
production can well lead to dramatic cost reductions while maintaining a high level of science
return. However, such practices will require substantial levels of surface power in locations where
solar energy is not the most viable candidate. Determining the requirements for a nuclear power
system so competent design studies can process is a necessity.
6.6.1. Finding 8: A definable set of hazards for a surface nuclear reactor power
system can be delineated and risks can be effectively mitigated
Hazard Identification Matrix for Mars Base Scenarios

Mars Base Mars Base Mars Base Mars Base Decommissio
Power NPS Startup NPS Standby/ n-ing/
System Operations Dormancy Disposal
Landing
Fission X X X X X
Product/
Isotope
Release
Contaminatio X X X
n Of
personnel
Radiation X X X
Exposure
Inadvertent X X X X X
Criticality
Inadvertent X
Reentry
Damage to X X X X
Spacecraft
Damage to X
Facilities
Table XXIII Lunar or Mars Base Nuclear Power Station Hazards
212
6.6.2. Primary Differences Between Cargo and Robotic Payload NEP system and a
Moon or Mars Base Power System that can Impact Safety
• Routine human activity in the local area
• Routine crew or robotic roveactivity in the local area
• A portion of the reactor shield is comprised of Mars soils via burial at the power station
location, the quality and uniformity of the lunar regolith or Mars soils used shielding is
unknown and could be highly variable, impacting the shielding quality factor
• Regularly scheduled takeoffs and landing events to support the Mars base
• A vertical mast and radiator boom assembly to eliminate waste heat on the Moon and the
incorporation of a forced Mars convection radiator for power production on Mars
6.6.2.1. Potential Safety Issues Associated with the Mars Base Power System
Once the Moon or Mars base power system has safely landed on the surface of the desired
celestial bodies, it represents no hazards to Earth or its biosphere. All hazards are localized to the
immediate environs of the Moon or Mars, or any other world where the nuclear power system is
located, activities that may be conducted by humans in the vicinity of the power system, activities
of crewed or remotely controlled rovers in the vicinity of the power system, and potential for
impacts from routine takeoffs and landings of Moon or Mars ascent/descent vehicles. For purposes
of this assessment, a safety issue is only some activity, risk or consequence that actually represents
an increase in concerns respecting the Moon’s or Mars’ future use by humanity or the potential risk
to persons or to astronauts working at the base, in rovers or at other locations on the Mars surface.
To this end, the above table is modified to include risk incorporating the aforementioned restriction.
Probabilities for some accidents are increased due to the potential failure of the landing
system. Some consideration should be given to the failure probability of Moon or Mars landers.
The success rate for lunar landings is quite high, > 95%, however, to date the Mars entry and
landing sequence failure rate is ~50%.
The conditional probabilities in Figure 11 should be referenced as follows. A green stop-
light should be interpreted that the hazard will probably occur at most once during the lifetime of
the Moon base or Mars base power system. A yellow stop-light means this activity will probably
occur at least once during the lifetime of the Mars base power system. An orange stop-light means
this hazard could occur more than once in the lifetime of a Mars base power system. Since routine
landing and take-off activities will be occurring, potential for a mis-guided landing or an aborted
takeoff that impact the Moon or Mars nuclear power system are at least feasible, although of
extremely low probability. The more remote the power system is located, the more remote is this
possibility, however, the more challenging the setup of the power station
213
Probability of an Initiating Event Occurring for Mars Base Power System Scenarios
Mars Base Mars Base Mars Base Mars Base Decommis
Power NPS Startup NPS Standby/ sioning/
System Operations Dormancy Disposal
Landing
Fission Product/
Isotope Release
Contamination
Of personnel
Radiation
Exposure
Inadvertent
Criticality
Inadvertent
Reentry
Damage to
Spacecraft
Damage to
Facilities
Failure or event occurs at most once during lifetime of system
Failure or event occurs at least once during lifetime of system
Failure or event occurs more than once during lifetime of system or for single events a high
probability of his failure occurring
Table XXIV Probability of an Initiating Scenario for Different Hazard Categories
6.6.2.2. Initiating Events for Mars Base Power System

Potential initiating events associated with the Mars power system include
• Fission product/isotope release
o Rover or ascent/descent vehicle accident ruptures pressure vessel and fuel
cladding
o Reactor over-temperature, or power transient, ruptures fuel cladding and
breaches pressure vessel/ducting
o Loss of coolant accident
o Loss of flow accident
o Unauthorized human activities
o Reactor power system landing vehicle accident
o Corrosion of reactor
o Erosion of components by sand or dust
• Contamination of Personnel/Mars Environment pressure vesse land fuel cladding by
the Moon or Mars environment
o Core pressure vessel and fuel cladding breach
o Mars power system landing incident
o Improper decommissioning
• Radiation Exposure
214
o Crewed operations in and around operating reactor facility
o Unplanned ascent/descent operations within 1 km of Mars power system in
greater does rate zones
o Shielding breach by rover, ascent/descent vehicle or unauthorized human
activity near Mars power system
o Radiation shine from Mars soil or atmosphere
o Inadvertent contact with improperly decommissioned reactor
• Inadvertent Criticality
o Unplanned control input
o Landing incident with core compaction
o Violation of criticality limitations
• Inadvertent reentry (on Mars surface)
o Mars power system lands in wrong location
o Mars power system crashes after unplanned reentry
• Damage to Spacecraft (Mars power system) by external source
o Hard landing or crash
o Rover impact
o Another landing spacecraft or an aborted takeoff land on Mars power system
o Severing of power cables by EVA or rover activity
• Damage to facilities (other than Mars base power system)
o Cooling fan blades break up and hits Mars base or personnel
Based on the above scenarios, there are feasible conditions where astronauts or personnel
could be exposed to radiation or become contaminated. EVA situations would presume the
astronaut or crewmember to be in a space suit, therefore, the contamination would likely be able to
be removed in a straightforward fashion, although special provisions would have to be made for
such an event. Radiation exposure is a more serious concern, since an astronaut or crewmember
will experience increasing exposure as the person gets closer to the power station within the 1 km
keep-out zone. It would appear prudent to have some form of audible and visual alarm that would
be activated whenever personnel get within 1 km of the power station while it is operating, to
adequately warn the individual or group that radiation levels are hazardous within this zone. Since
there is a possibility that a crewed lander could land accidentally within the keep-out zone, this
radiation exposure condition could result in some serious consequences if the reactor is not shut
down and the lander is close to an operating reactor.
Mars power system decommissioning is an important issue from a long-term radiation
protection perspective. Unlike the lunar counterpart, where no wind or water is present, on Mars, it
is now known that there exists significant dust storms and there is a very high likelihood of at least
stagnant water. This water is expected to be a richly briny solution, (at least in some locations) with
sulfate and bromide salts. This poses a longer term environmental compatibility problem with
extended entombment than the lunar system would likely pose. Proper entombment will allow the
system to remain in place with sufficient shielding to minimize radiation dose to any individual who
enter the area. Even with the potential for chemical attack during entombment, due to the relatively
short half-life of most of the fission products, there should be no unusual safety concerns with long-
term disposal on the surface of either the Moon or Mars.
215
6.7. Finding 9: There appears to be no reason that a space nuclear reactor power system
cannot be safely deployed and operated on the surface of another world while maintaining
standards of planetary protection.
A space nuclear power system appears to be adaptable in a technical and risk-abatement
fashion for use in crewed activities on the Mars surface. However, there could be substantial
changes to a design adapted for in-space operations and other design modifications may be
necessary. Some of these changes are from a safety perspective, others from an operational
perspective. The primary assumption in this assessment is that radiation levels postulated in the
shielding analysis area can, in fact, be achieved employing Lunar regolith or Mars soils, and that the
logistics of drilling a hole into which the modified reactor can fit, or that regolith or Mars soils can
be piled around the reactor to enable a high quality radiation shield and that the requisite precision
landing and deployments can be performed. All of this must be accomplished on Mars without
benefit of human presence, whereas on the lunar surface, it envisioned that crew members will
already be in place to assist in the preparation of the lunar surface for the power system. Depending
on the deployment architecture, it is possible that lunar deployment could occur without human
presence as well. The primary risks with either the Moon or Mars nuclear power station would
appear to be procedures violations of some form. The violations would most likely include
violation of the radiation area exclusion boundaries by EVA crewmembers or crewed rovers. Other
risks such as reactivity excursions would likely be extremely rare given the autonomous nature of
the reactor control system. With a properly designed reactor power system, overall risks should be
quite small.
6.8. Finding 10: A Space Reactor System Enables Effective of Design Options in Mitigating
Potential Radiation Releases
The International Academy of Astronautics (IAA) recommends that a space nuclear
missions place paramount importance on safety. The inclusion of careful safety planning at an early
stage and throughout all mission activities is vital to assure that the space nuclear system mission
profiles will be conducted in a manner that clearly demonstrates an impeccable safety program. The
guiding philosophy for the space nuclear power and propulsion program is to maintain a high
profile for safety throughout concept development and design such that a safety mentality is
established as an integral aspect of design and operations. Inherent in a safety-based effort is the
imperative to periodically review operations and administrative considerations to ascertain if any of
these might result in diminished risk.
6.9. Finding 11: A Transparent and Systematically Traceable Space System Safety Test
and Analysis Program Must be Conducted to Ensure Crew and Public Safety
This section discusses the safety test and analysis program for the space reactor system.
Although the focus is on nuclear safety associated with the launch of the JIMO reactor system,
some discussion is provided on the additional safety testing required for the ground reactor test
program. Safety activities associated with the conduct of the reactor ground test need to be
integrated into the overall safety plan; nonetheless, safety requirements, testing, analysis, and
implementation for the ground test must be clearly differentiated from the safety associated with the
launch of the space reactor system
216
6.9.1. Safety Testing
Zero-power critical testing will be performed during the fueling operation. Following zero-
power testing, the reactor will be shutdown and will not be taken critical again until it is placed in
its intended operational orbit. This approach presents a major safety advantage in that no safety
impact on Earth’s biosphere or its inhabitants is predicted to result from any postulated reactor
system accident after deployment in high orbit. A careful analysis is required to assure that
postulated operational accidents do not present safety issues; nonetheless, this basic approach limits
reactor system radiological safety considerations to postulated accidents that could lead to
accidental criticality or lead to the dispersal of essentially fresh nuclear fuel. The radioactive
content of the essentially fresh uranium fuel is very low, however, and is not expected to present a
major radiological risk. Non-radiological safety issues associated with the reactor system include
non-radiological consequences of liquid metal fires and explosions and the dispersal of toxic
materials, such as beryllium and beryllium compounds.
Given this limitation on types of postulated accidents resulting in safety issues, safety-
testing requirements for a space reactor are quite limited. Safety testing for the space reactor
systems falls potentially into two major categories: System Disruption Tests and Criticality Tests,
as shown in Figure 12, and possible complete ground test of the reactor power system.
Testing Validating Safety Margins

System Disruption Tests Criticality Tests
Propellant explosions Normal configurations
Propellant fires Water flooded configuration
Shrapnel Impacts from explosions Compacted geometries (including water
immersion)
Reactor system impact with surfaces or Various reflector configurations
objects
Reentry effects Sand reflection
Normal launch environments Reconfigured core geometries
Impact launch environments Compacted or reconfigured geometries
post launch environment
Explosive disassembly of reactor core Test bechmarked code predictions using
surrogate materials for uranium
Table XXV Possible Types of Safety Tests
At this stage it is difficult to predict if all of the subcategories of testing will be required, and
the detailed requirements for these tests are difficult to forecast. Nonetheless, some basic features
of the required safety tests are outlined here. The test facilities at Sandia National Laboratories are
used in the following as examples of the kind of facilities that are available for safety testing;
however, the proposed test plan will examine all possible test facilities and select those facilities
best suited to the safety test program.
6.9.2. Propellant Explosion and Fire Tests

The range of possible environments resulting from a postulated propellant fire or explosion
is likely to be very large. Detailed computer modeling will be used to determine the effect of these
environments on the reactor system. The principal concern is the possibility of reactor system
reconfiguration that could result in an inadvertent criticality. The possibility of dispersion of nuclear
fuel materials and toxic substances will be examined as well. Some propellant fire and explosion
217
testing will be required to validate computer models and to provide parameters required by these
models. More accurate equations-of-state for propellants may be required. These tests will require a
mockup of the reactor system, reactor components, or both. Mockups will include relevant design
features and either actual materials or surrogate materials. Analysis of the design and potential
propellant fire and explosion environments may identify specific issues that require additional
testing. Facilities for conducting fire and explosion tests are well established. Figure 120 shows a
nuclear container fire test at the Coyote Canyon Test Facility.
Figure 120 Nuclear container fire testing at the Coyote Canyon test facility (at SNL)
6.9.3. Impact Tests

Fragments produced and propelled by a propellant explosion may impact the reactor system.
Although liquid-fueled rockets (such as the Delta-IV/H) exhibit more benign fire and explosion
scenarios that solid rocket systems, realistic environments must be modeled. Further, in an
international space exploration program, it is considered feasible that a space nuclear system will be
launched on a system containing solid propellants. A launch abort or failure to achieve orbit could
result in high-speed impact of the reactor system with solid surfaces or water. Other possible
incidents, such as dropping the reactor system during mating with the launch vehicle, can result in
reactor impact. The consequences of these postulated accidents will be assessed using computer
modeling. Actual impact testing using a mockup reactor with simulated fuel will benchmark the
codes. As for propellant fire and explosion assessments, some shrapnel or fragment impact tests
will be required to validate computer models and to provide parameters required by these models.
Also, an analysis may identify specific issues that require additional testing. Some of the presently
existing facilities for impact testing include: drop test facilities, pull-down test facilities (Figure 14),
rocket sled impact facilities (Figure 15), shock test facilities, water impact facilities, and large
centrifuge facilities.
218
Figure 121 Nuclear container pull-down impact testing (at SNL)
Figure 122 Nuclear container rocket sled impact test facility at SNL
6.9.4. Reentry Tests

The geometry of the reentry shield (if used) and the materials of potentially exposed
components have been well studied. Established computer analysis tools will be used to examine
postulated reentry accidents. If more detailed analysis identifies a need for testing, wind tunnel
facilities can be used to perform the required tests.
6.9.5. Launch Environment Tests

For a normal launch, the rector system will be subject to dynamic g-loads, staging shocks,
and vibration. Testing will be required to assure that safety systems will not be compromised by
these environmental insults. Shaker tables and shock test facilities will be required to verify the
integrity of safety features for the space reactor system for the expected launch environment.
219
6.9.6. Criticality Tests
The neutronic analysis methods needed to determine the reactor multiplication factor for
possible accident geometries are well established. Typically, advanced transport theory or Monte
Carlo computer codes are used to determine whether core reconfiguration, core flooding, or other
possible environments can lead to an inadvertent reactor criticality. However, these computer codes
require accurate neutron cross sections (neutron cross sections are related to the probability of
neutron interactions with specific materials). Inaccuracies in neutron cross-sections for some reactor
materials may introduce uncertainties into the neutronic analysis. Thus, critical tests may be
required to reduce uncertainties resulting from cross section inaccuracies or unusual geometries.
Criticality testing mockups should use the actual reactor materials and use dimensions and
geometries similar to those expected for the actual flight reactor system. Criticality mockups are
only tested at very low power levels (often called zero-power testing). As a consequence, many
details of the actual reactor design (coolant system, pressure boundaries) are not required for the
criticality mockup. The simplified design of the mockup permits relatively inexpensive, early
verification of cross sections and analysis methods. If criticality predictions do not accurately
predict test findings, alternative cross sections may be required or the analysis predictions may be
benchmarked to the results obtained from criticality testing. The use of a criticality mockup also
permits rapid change-out of materials and components such that a variety of postulated accident
configurations, geometries, and environments can be studied. Some alternative configurations may
include reflector reconfigurations, alternative reflecting materials (e.g., wet sand), water flooding,
and distorted or compacted geometries. Criticality testing capabilities are less widespread today
than they were several decades ago. Criticality testing requires not only the appropriate equipment
and expertise, it also requires a facility approved for carrying out a fairly wide scope of criticality
tests. Several of the national laboratories and other facilities should be capable of performing the
required criticality tests.
6.10. Finding 11: Prior Space Reactor Programs Expended Resources on Destructive
Disassembly Testing for Low-Probability Incidents – System Level Testing Should be
Reserved for More Likely Scenraios
Safety testing and analysis should be integrated into planning for the reactor system ground
test program. Sharing of facilities and test results will optimize program resources and minimize
development time. For example, some of the critical facility test results obtained for the flight
reactor will be essential to reactor designers and will be applicable to the ground reactor safety
program. Reactivity temperature coefficient measurements obtained for ground test safety
assessments are also needed for understanding the performance of the space reactor system. Reentry
safety analysis and (possibly) testing can provide information needed for safeguards issues relating
to recovery of the nuclear fuel. Testing of the ground reactor cooling system can provide needed
performance information for the flight reactor. Other types of tests and analysis may be required
for reliability, mission success, operational performance, safeguards, and space environmental
protection. These testing considerations should be included in an overall test plan.
The early space reactor programs postulated reactor accidents, such as core-flooding
accidents, that resulted in reactivity excursions sufficiently rapid to cause fuel vaporization and
destructive disassembly. Much effort was directed toward the study of reactivity induced
destructive disassembly accidents for space reactors. Full-scale disruptive disassembly tests were
carried out for both the SNAP-10A and the NERVA reactors. In retrospect, however, the emphasis
placed on reactivity-induced disassembly accidents is questionable. Scenarios required to induce
accidents of this type were not highly plausible, and if the space nuclear system meets the
220
aforementioned safety requirements and recommendations, they should be precluded by the design.
Much of the emphasis on reactivity-induced destructive disassembly accidents was motivated by the
absence of a requirement for sub-criticality during postulated accident scenarios for SNAP-10A.
This type of testing is should not be justified for the presently understood space nuclear missions;
consequently, reactivity-induced destructive disassembly tests are not recommended.
6.10.1. Safety Analysis

Safety analysis for the JIMO space reactor program will be extensive. Test-validated safety
analysis is essential to the overall safety assessment and flight approval, and plays a key role in the
guidance of the safety program. For the most part, existing computer models will be used to carry
out the safety analysis. Some modifications to these tools and development of additional analysis
tools may be required to address the special needs of the JIMO program. Although the test program
discussed in the preceding will be required to validate much of the analysis, many of these analysis
tools have been validated by previous testing. Some of the important types of safety analysis
planned for the space reactor mission include:
6.10.2. Neutronics
Well-established Monte Carlo and transport theory computer codes (e.g., MCNP –Monte
Carlo N-Particle Transport code) will be used to explore the reactivity effects of various reactor
configurations and environments for postulated accident scenarios.
6.10.3. Shielding
Monte Carlo and transport theory computer codes will be used to determine possible
radiation exposure doses resulting from postulated inadvertent criticality scenarios.
6.10.4. Fires and Explosions

A number of established computer codes can be used to perform an analysis of postulated
propellant fire (e.g., the Sandia Fireball Code) and explosion environments. Structural response
codes will be used to determine system response to explosions and the consequences of fragments
impact on the reactor system. Some code modifications may be required to couple the system
thermal and structural response analysis with the predictions of the postulated accident
environments, including possible sequential or multi-source events
6.10.5. Intentional Disassembly

A number of codes will be necessary to characterize intentional disassembly to prevent accidental
criticality and to ensure dispersal in the event of an accidental reentry. These codes include CTHii,
PRONTO3D/SPH (Single Particle Hydrodynamic), and under the latest version for the advanced
simulation multiple processor efforts, ALEGRA. These will be used in conjunction with reactivity
codes such as MCNP to ensure that the reactor will not become critical during an intentional
disassembly sequence.
221
6.10.6. Reentry Analysis
Trajectory analysis tools are well developed. For example, the Trajectory Simulation and
Analysis Program (TSAP) is well written and documented. The Heating and Analysis Done
Interactively (BLUNTY) code can be used to perform reentry-heating calculations and to generate
output in a format suitable for standard thermal response codes.
6.10.7. Coupled Impact/Reactivity Analyses

A number of established impact analysis computer codes are available to analyze impact of
complex geometries. An example of a recent impact is presented here. Sandia National Laboratories
recently carried out an analysis of a postulated vertical drop of a JIMO space reactor from 100 m
onto a hard surface, resulting in a 100-m/s impact. The analysis was carried out under Internal
Research and Development funding from NGST [Lipinski et al., 2004] using the PRONTO-
3D/SPH hydrodynamic code. Some preliminary results are shown in Figure 123. Two cases are
shown, one with the lower grid plate weakly attached to the vessel, and one with it firmly attached
to the vessel. The analysis indicates that the frozen lithium absorbs some of the shock; nonetheless,
significant distortion of the reactor vessel is predicted. When the grid plate is attached to the vessel,
impact causes the fuel pins to splay outward slightly; consequently, attaching the grid firmly to the
vessel may aid in the reduction of reactor criticality. In any case, wide swings in reactivity were
noted, (Figure 124) Placing a wedge at the base of the lower head might help drive the pins further
apart for this impact scenario. The predicted distorted geometry will be used in an MCNP
neutronics analysis to determine the effect on reactivity [Lenard, 2005]. Although it is not possible
to assess every possible accident scenario, analysis of this type can provide useful insights and
envelope the realm of possible accident conditions.
3-D hydrodynamic modeling can predict distortion levels of the core for accident scenarios
Figure 123 Impact of bare reactor on concrete at 44 m/s (100-m drop)
222
k e ff a n d ra d ia l d e fo rm a tio n
fo r fu lly re fle c te d 4 2 1 p in re a c to r
0 .0 4 5 0
k e ff 0 .0 4 0 0
1 .0 3 5 ra d ia l d e fo rm a tio n
0 .0 3 5 0
0 .0 3 0 0
a v e ra g e ra d ia
d e fo rm a tio n (c
0 .0 2 5 0
1 .0 2 5
k e ff
0 .0 2 0 0
0 .0 1 5 0
1 .0 1 5 0 .0 1 0 0
0 .0 0 5 0
0 .0 0 0 0
1 .0 0 5 -0 .0 0 5 0
0 .0 0 0 0 0 .0 0 2 0 0 .0 0 4 0 0 .0 0 6 0 0 .0 0 8 0 0 .0 1 0 0
T im e (s )
Figure 124 Induced Reactivity Swings During 100 m/s Impact
6.10.8. Risk/Consequence Analysis

Established methods will be used to perform a risk/consequence analysis for the JIMO space
reactor-mission. Part of the analysis will provide source terms for various postulated accident
scenarios. The analysis will also examine the possibility of atmospheric dispersion of radioactive
materials, possible inhalation and ingestion of radioactive materials, environmental contamination,
cloud shine and ground shine, and projected collective doses for various scenarios. Risks and
projected consequences associated with various scenarios will be tabulated and discussed. Event
tree and fault tree methods will be used to guide and elucidate the analysis and to identify issues or
components that require further study or modifications to reduce mission safety risk. This analysis
will take a graded approach as shown in Figure 125 [Lenard, 2005]. Such an approach precludes
involving large effort on scenarios with minimal risk and concentrates effort on areas of high risk.
Simple hazard analysis Detailed quantitative analysis

Figure 125 Graded Risks Assessment and Analysis Approach
223
6.10.9. Other Analysis
Other types of analysis will be used to assess and guide the safety program for the ground
reactor test and to provide performance predictions. Example analyses include: reactor kinetics and
burn up analysis, fission product inventory predictions, reactor cooling analysis, and predictions of
system response to postulated accidents. These analyses, however, do not directly address safety as
applied to the launch of the JIMO reactor system.
• Recommendation 18: The Space Nuclear Reactor Program Should Prioritize Test and
Analysis Programs Based on Engineering Judgment of Failure Probabilities
The proposed safety test and analysis program should not proceed independent of failure
scenario probabilities. While launch vehicle data-books may not be available, a basis for what
likely failures might include can be based on historic data and engineering judgment. To that end,
the IAA study team has identified potential failure scenarios and has performed an initial screening
of events and matched realistic probabilities of failure (based on the Cassini FSAR in some cases)
coupled with probability of a release of nuclear material, fission products, or an inadvertent
criticality. The potential failure probabilities are matched with different phases of flight from the
time the reactor module arrives at the launch site until reactor startup occurs. The results are shown
in Table XXVI through Table XXIX. [Lenard et al., 2004]
(t < 0 seconds)
Failure and Initiators Estimated Probability of Occurrence
Initiators Accident Accident Release of Induced Release of
Type Fresh Fuel Criticality* Fission
Products
Hoisting Dropped 10-5 <10-5 <10-6 <10-6
Failure Reactor
Spurious Accidental ~10-9 N/A ~10-9 <10-9
Signal Startup
Spurious Command ~10-9 ~10-9 ~10-9 ~10-9
Signal Destruct
Propellant Propellant 10-4 < 10-4 <10-6 <10-6
Spill or Leak Fire/Explosi
on
*Includes subsequent reactor surface impact and immersion in water or wet sand
Table XXVI Accident Scenario Probability Estimates for Pre-launch
(t <20 seconds)
Products
Fall-back or Launch pad < 10-6 < 10-6 < 10-6 < 10-6
Tip-over Impact
Trajectory Command 6.6 X10-4 ~1* <10-7** <10-7**
Failure destruct
In-flight Land ~ 10-4 ~ 10-4 < 10-7** < 10-6
explosion Impact
*Assumes external environment reactivity enhancement, impact, or control failure
** Assumes explosive disassembly
Table XXVII Accident Scenario Probability Estimates for Early Launch Phase
224
(t = 20 – 4800 seconds)
Products
-2 -7**
Trajectory Command ~ 10 ~1 <10 <10-7**
Failure destruct
Suborbital Land ~ 10-4 ~1 <10-7** <10-7**
Reentry Impact
Ocean ~ 10-3 ~1 <10-7** ~1**
Impact
Orbital Reentry Land or ~ 10-7*** ~1 <10-7** ~1**
Ocean
Impact
*Assumes external environment reactivity enhancement, impact, or control failure
** Assumes explosive disassembly and complete dispersal
*** Requirement
Table XXVIII Accident Probability Estimates for Late Ascent Phase and Orbit Circularization
(t > 4800 seconds)

Type Fresh Fuel Criticality Fission
Products
-7 -7
Trajectory Reentry < 10 ~1** < 10 ~1**
Failure
Fly-by Reentry < 10-7 ~1** <10-7 ~1**
Trajectory
Malfunction
** Assumes explosive disassembly and complete dispersal
Table XXIX Accident Scenario Probability Estimates for Reactor Operational Phase
6.11. Finding 19: An Integrated Approach for Performance and Safety Analysis and Testing
is Critical to a Cost-Effective Development Program
This section describes a “recommended approach for integrated performance and safety
analysis, compliant with Atomic Energy Act requirements. We interpret this to mean following a
launch approval process similar to that used by previous launches of radioisotopic power systems,
following all DOE orders and EPA rules for safe handling of radioactive materials, and integrating
these processes with the traditional launch process. An integrated safety program will be
established to address all mission related safety issues in a coordinated manner. In the following,
the safety program organizational structure is discussed along with a discussion of the basic steps
required to secure launch approval.
6.11.1. Safety Program Organization

The space nuclear reactor safety program will assure that the mission will be conducted
safely and that launch approval will be obtained. The safety program will foster a safety culture
among all program participants. This philosophy makes safety everyone’s responsibility, rather than
225
assuming that all safety issues are only the safety team’s concern. Making safety an integral aspect
of the program also enhances safety communication and teamwork. A preliminary safety program
has been initiated during the Phase-A design effort and will continue through system development
and deployment. Addressing safety issues during system development ensures that the design can
either be changed to eliminate the hazard or appropriate controls can be incorporated. An example
of this is the launch-accident impact safety effort that was started during Task 2. This has enabled
us to ascertain methods to internally mount the reactor that enhance safety during hard surface
impacts. The designed-in safety approach should be required by space reactor system standard
practice and should ensure a cost-effective safety approach by dictating that safety is not
implemented post-design.
An integrated organizational structure of the space nuclear reactor safety program will need
to be developed jointly with all the relevant parties. An Integrated Safety Program Manager will be
responsible for assuring the integration and proper functioning of all program safety activities. The
Program Manager is the recipient for safety reporting and is the ultimate authority within the
program for all safety decisions. Traditionally, an Independent Nuclear Safety Team is established
and reports directly to the Program Manager to provide independent assurance that nuclear safety
issues are properly addressed. Where possible, the analysis and testing program that is associated
with system validation and performance is employed to support safety analysis or to verify safety
functions and safety margins.
A Reactor Design-Safety Team performs the primary safety functions and provides safety
findings to the Reactor Project Manager and the Integrated Safety Program Manager. In order to
provide an additional measure of independence, the Design-Safety Team is sometimes funded by
and reports to a line organization different from the Reactor Design Team. The Reactor Design-
Safety Team functions include:
• Interfacing with the Project and Program Offices on safety maters

• Development of a Safety Program Plan
• Identifying applicable safety guidance and requirements
• Identifying safety documentation and approval requirements
• Interpretation of safety guidance
• Development of detailed safety specifications
• Contributing to safety component design
• Performance of in-depth safety analysis
• Performance of safety testing
• Development of safety procedures
• Preparation and Safety Analysis Reports for the Program Office
• Presentation of safety findings to INSRP and defense of findings
• Interfacing with design team and quality assurance team on safety matters
• Conducting internal safety audits and reviews
• Preparation of much of the EIS
• Preparation of other safety documentation (e.g., for Range Safety)
• Serving as public interface on safety matters.
A comprehensive, ongoing operational safety program is an essential aspect of the

integrated safety program. The operational safety program is needed to assure safe operation for the
protection of workers, the public, and the environment. The operational safety program is conducted
in compliance with DOE orders relating to radiation safety, occupational safety, and environmental
protection. The operational safety program includes an operational safety analysis, a quality
assurance program, a radiation protection plan and radiation protection program, operational
reviews and surveys, a contractor self-appraisal program, provisions for DOE reviews surveys and
226
approvals, and readiness reviews for each specific facility. A separate safety program will be
established, in compliance with appropriate DOE orders, for the Ground Reactor Test Program.
DOE orders and DOT regulations will be followed for packaging and transport of special nuclear
materials and nuclear systems.
6.12. Finding 20: The Ultimate Objective of All Programmatic Activities is to Obtain
Launch Approval for the Space Reactor System – the Program Should be Structured to
Attain the Goal.
A number of review processes have addressed launches of Radioisotope Thermoelectric
Generators (RTGs) in the past and one would expect essentially identical processes to be used to for
review and approval of a space reactor launch. The principal review processes in the past have
included the Interagency Nuclear Safety Review Panel (INSRP) as well as reviews for the
environmental protection, range safety, range operations, and orbital safety. Other safety review and
approval processes include approvals for transportation of nuclear materials to the launch site and
launch site safety approval. An approximate sequence for the INSRP review and approval process is
presented in Figure 21. The figure also presents the approximate sequence for the Environmental
Impact Statement (EIS), range safety, range operations, and orbital safety. For any U.S. space
mission involving the use of nuclear energy systems with significant quantities of radioactive or
fissile material, INSRP review is required and launch approval must be obtained from the Office of
the President.
The launch decision is based on a consideration of the projected benefits and risks of the
mission and on a safety analysis and evaluation that includes both the program’s analysis and an
independent characterization of mission safety by an Interagency Nuclear Safety Review Panel. The
process includes several increasingly detailed Safety Analysis Reports (SARs) prepared by the
JIMO program office and participating contractors, review and evaluation of those SARs by
INSRP, and characterization of the mission radiological risks by INSRP in a Safety Evaluation
Report (SER). The SER is provided to the Office of Science and Technology Policy (OSTP) and the
National Security Council (NSC) within the Office of the President. The Office of the President
grants (or denies) final approval for launch.
227
Draft
PSAR FSAR FSAR OSTP NSC
SER Office of the President
INSRP
Process INSRP Review DOE, DoD, NASA
Operational Presidential
Program Program and Operational Documents Directives Approval
Introduction
Mission
Range OPS Process Approval
FSAR LAUNCH
required Safety
MSPSP* Review
Range Safety Process EIS
required * Missile System Prelaunch Safety Package
Orbit Operational Safety and Readiness Reviews Test

Operations
Orbital Safety Process Space
Safety
Approval
Notice Draft
of Final Record of
EIS EIS Decision
Intent
EIS Process
0 1st year 2nd year 3rd year 4th year 5th year
Figure 126 Notional Timeline of Events for Launch Approval
The National Environmental Policy Act (NEPA) EIS process is entirely separate from the
INSRP review. Typically, the NEPA process is completed early in the program cycle, preliminary
data and analysis may be used. In addition, the NEPA process differs in that it involves the public
and a large number of federal, state, and local agencies. The EIS provides a general understanding
of how the action will be implemented, and what potential environmental impacts could result. It is
also designed to inform decision makers and the public of a reasonable range of alternatives that are
compatible with the purpose and need of the action. These alternatives must be compared with the
proposed action in terms of their potential environmental impacts. An EIS for the JIMO nuclear
mission must provide a detailed analysis of the impact of a normal launch on the air, water, plants,
etc. of the launch area. It must also discuss a reasonable range of postulated accidents their potential
impact.
An emergency response plan is also required. Emergency preparedness plans are required
for all sites and situations involving the utilization, handling, transporting, or storing special nuclear
materials. Contingency plans and emergency response resources are also required for the launch
aborts and for other postulated scenarios that could result in reentry of the system. Unique features
of launch abort and reentry emergency response includes plans for nuclear source or nuclear system
recovery and long-term post-contingency assessments and recovery.
228
• Recommendation 19: The US Interagency Nuclear Safety Review Panel (INSRP) process
has ensured a rich legacy of safety radioisotope launches – the same (for US launches) or a
similar practice (for foreign launches) should be adopted for space reactors
•
A number of documents have been needed in the past as part of the launch approval process
for RTGs; the list needed for a reactor launch might be very similar. The list of important safety
documentation for RTG launches which has traditionally been provided by the mission Program
Office includes:
• Launch Vehicle Data Book

• Space Nuclear Safety Documentation
• Safety Program Plan
o Flight Safety Requirements Documents
o Safety Analysis Reports (PSAR, Draft FSAR, and FSAR)
o Radiological Protection Plan
o Emergency Preparedness and Response Plan
o Safety Analysis Report for Packaging (SARP)
o Ground Safety Analysis Report (GSAR)
• Range Documentation
o Program Introduction Plan (PI)
o Statement of Capabilities Document (SOC)
o Program Requirements Document (PRD)
o Operational Requirements (OR)
o Operational Directive (OD)
o Missile System Pre-launch Safety Package (MSPSP)
• Orbital Safety Documentation
o Test Operational Risk Assessment (TORA)
o Preliminary Hazards List (PHL)
o Test Operations Space Safety Approval (TOSSA)
o Orbital Readiness Space Safety Review (ORSSR)
• Environmental Protection Documentation
o Notice of Intent
o Environmental Impact Statement (Draft and Final)
o Notice of Availability
6.13. Summary and Recommendations

The IAA staunchly endorses the use of reliable and safe space and surface nuclear power
and propulsion systems in areas where they can provide distinct advantages. The IAA recognizes
there are a number of venues under which space nuclear systems may be implemented. There are
commercial space transportation and asset recovery missions that appear to be profitable only under
implementation with space nuclear power. There are science missions such as the Jupiter Icy
Moons Orbiter program where large payload, massive payload radiation protection (from the Jovian
radiation field) and data transmission requirements are enabled by space nuclear power and
propulsion. Human transportation to Mars is made demonstrably more safe by nuclear thermal
propulsion, and using nuclear electric power on the surface of the Moon and Mars can dramatically
reduce the costs of human exploration programs.
229
Other systems concept have recently surfaced where nuclear fission power systems, either
through advanced technologies in an electric propulsion mode or through such concepts as
MiniMag Orion, coupled with macro-particle accelerators, can enable one-way accelerations to .1c
to .15c (c is the speed of light) and retain the ability to rendezvous with star systems 4-5 light years
distant within a human lifespan. This is an astounding although futuristic concept. However, the
fact that known extensions to existing and operational technologies can provide the basis for initial
interstellar rendezvous flights is significant, and nuclear energy is the enabling capability.
It is for the above reasons, to mention a few, that the IAA has reviewed the wide body of
safety and policy literature and has made some recommendations to improve their utility to the
space systems design community where appropriate. The existing safety and environmental
compliance process employed in the United States appears appropriate for other countries as
modified to meet their own internal requirements, while still remaining in compliance with
international norms.
We have reviewed the issue regarding use of reactors in Earth orbit and find that the
abnormally large advantages to using Isp > 4,000 s to accelerate large payloads from lower Earth
orbits to their destination, and to rapidly and cost-effectively transport humans to the Moon or Mars
cannot be dismissed or ignored. To this end, we have clarified some ambiguities in the meaning of
decay times, hence the implications of a Sufficiently High Orbit (in UN parlance), and we further
have provided some proposed guidance regarding the design approaches to what highly reliable
orbit raising systems might mean. While the term highly reliable might have been employed to
allow some states parties to create their own design approaches, consistent with internal criteria, we
note that ambiguous requirements do not enhance public confidence. Nuclear systems are not like
other technologies, at least not in the opinion of the public, that in many cases has been highly
sensitized by anti-nuclear activists whose agenda is not to enhance the use of nuclear power
systems, regardless of the operational or scientific justification. To cope properly with this issue,
where, in many cases, objective truth and physical reality have been overtaken by rhetoric and
ideology, we cannot overstate the importance of a rigorous, transparent and highly safe space
reactor program.
6.14. References
Greene, S.R., Smith, B., Nesmith, B., Bhattachryya, S., Houts, M., Marshall, A.C., Mason, L.,
Poston, D., Weitzberg, A., and Wright, S., „Summary of an Interagency NASA/DOE Review of
Space Reactor Power System Cocnepts“ ORNL/SR/LTR-2003/001, April 2003.
Lipinski, R. J. Wright, S.A., Lenard, R.X., Suo-Antilla, A.J., Vernon, M. E., Marshall, A.C.,
Jablonski, J.A., and Helmick, P.H., “CRADA SC03/01673 Final Report: Space Reactor Trade
Studies and Conceptual Design”, Sandia National Laboratories, November 23, 2004.
COPOUS „Principles Relating to the Peaceful Uses of Nuclear Power Systems in Outer Space“
Resolution 47/67., December 14, 1992,
http://www.unvienna.org/SpaceLaw/Gares/html/gares_47_0067.html.
National Security Council Memorandum NSC/Presidential Directive-25, Presidential Directive:

Science and Technology Experiments with Possible Large Scale Adverse Environmental Effects
and Launch of Nuclear Systems into Outer Space, December 14, 1977.
230
United Nations. “Treaty on the Principles Governing the Activities of Stats in the Exploration and
Use of Outer Space, Including the Moon and Other Celestial Bodies“, October 10, 1967,
www.state.gov/t/trt/5181.htm
National Security Council Memorandum. Presidential Decision Directive PDD-NSC-49/NSTL-48

National Space Policy of 1996. September 14, 1996
United Nations. A“Agreement on the Rescue of Astronauts and the Return of Astronauts and the
Return of ObjectsLaunched into Outer Space“, 3 December 1968,
www.unoosa.unvienna.org/SpaceLaw/rescuetxt.htm .
United Nations “Convention on Registration of Objects Launched into Outer Space,“ 14 January
1975, www.unoosa.unvienna.org/SpaceLaw//SORregistry/registxt.htm
United Nations. “Treaty on the Non Proliferation of Nuclear Weapons, (NPT),“ Ratified 1970,
extended indefinitely 11 May 1995, http://document2.un.org/wmd/not/index.htm
United Nations. “Convention on the International Liability for Damage Caused by Outer Space
Objects“, 29 March 1972, www.unoosa.unvienna.org/SpaceLaw/liabilitytxt.htm
231
7. Appendix
7.A. Radioactivity, Doses and Risks in Nuclear Propulsion(by A. Del

Rossi and C. Bruno)
7.A.1. INTRODUCTION
Radiation and nuclear are words that tend to spread fear among people. Even in highly
technologically developed countries, the public has little or no knowledge of radiation, and when
they do it usually associates it with weapons, accidents, fallout and cancer. Only specialists know
about natural background exposure or about medical use of radiation. In this context the use of
nuclear energy for rockets may encounter strong resistance.
The purpose of this Appendix is to inform the non specialist about what radiation and dose
are, about effects of radiation on humans and about sources of radiation, including estimates of the
dose from nuclear propulsion systems.
7.A.2. RADIOACTIVITY
Radioactivity is the process undergone by unstable nuclei (radionuclides), as well as nuclei

in excited states, causing spontaneous changes, or transformations, in composition and/or internal
energy of the nucleus. This means that radioactivity may change a chemical element into another,
releasing or absorbing energy in the process. The most common transformations are three: alpha
decay, beta decay and gamma decay.
7.A.2.1 Alpha Decay
In alpha decay the nucleus of an element with mass number A1 and atomic number Z1 emits
an alpha particle. Alpha particles are made of two protons and two neutrons, that is, a Helium
nucleus. The original nucleus is replaced by a new nucleus whose mass number A2 is equal to A1-4
and atomic number Z2 is Z1-2, and an alpha particle.
For instance, 222Rn (ARn=222, ZRn=86) decays into 218Po, meaning that the nucleus of 222Rn
emits an alpha particle (Aα=4, Zα=2), leaving as remainder a nucleus whose mass number is 218
(222-4) and atomic number (86-2) = 84, that is, 218Po.
The mass (energy) of the parent nucleus must exceed the sum of the masses (energies) of the
daughter nucleus and alpha particle emitted. The condition for α-decay to occur can be expressed as
follows [1]:
M (A, Z) > M (A-4, Z-2) + M (He4)
232
7.A.2.2 Beta Decay
Beta decay is the spontaneous transformation of an unstable nucleus into a new nucleus with
charge differing by ΔZ = ±1, because of the emission of an electron (β- decay) or a positron (β+
decay) or the capture of an electron (e-capture)
In the first case (β- decay) one of the neutrons of the nucleus emits an elctron and becomes a
proton. The mass number A does not change, while the new nucleus has an atomic number higher
by 1.
Tritium (3H, often symbolized by a T), AT = 3 ZT = 1, β- decays into 3He, AHe = 3 ZHe = 2,
meaning that one of the two neutrons of the tritium nucleus emits an electron and becomes a proton;
the mass number does not change, i.e., AT = AHe, while the positive charge of the new nucleus
increases by 1, ZHe = ZT + 1.
The energy condition is that the mass (energy) of the parent nucleus is higher than the sum
of the masses (energies) of the daughter nucleus and the electron, and is expressed by [1]:
M (A, Z) > M (A, Z+1) + me
In the β+ decay the unstable nucleus emits a positron (i.e., a positive electron). The β+ decay
can be treated as the transformation of a proton into a neutron, because also in this case the parent
nucleus and the daughter nucleus have the same mass number A, while the atomic number of the
daughter Z is lower by 1. The proton mass is lower than the neutron mass (energy). The
transformation of the proton into a neutron is possible since the proton is bonded to a nucleus and
the excess energy to become a neutron is supplied by the nucleus itself. The energy condition can be
expressed in analogy with the β- case [1]
M (A, Z) > M (A, Z-1) + me+

11
C, AC = 11 ZC = 6, decays β+ into 11B, AB = 11 ZB = 5, and the missing charge of Boron-
11 is that of the positron emitted.
The third type of beta decay is the electron capture: it consists in the capture of an electron
by a nucleus from its own electron shell. For heavy nuclei with the K-shell close to the nucleus, this
phenomenon (also defined K-capture) is quite common; captures from L shell (L-capture), M shell
(M-capture), etc. have also been observed. After the capture, the nucleus has the same mass number
A, but its atomic number Z decreases by 1: the electron captured and one of the protons of the
nucleus become a neutron in the daughter nucleus.
For instance, 7Be, ABe = 7 ZBe = 4, after capturing an electron from its K shell, becomes 7Li,
ALi = 7 ZLi = 3; the mass number does not change: ABe = ALi = 7, while the atomic number Z of the
lithium is lower by 1.
The mass (energy) condition is that the sum of the masses (energies) of the captured electron
and the parent nucleus is higher than the mass (energy) of the daughter nucleus [1].
M (A, Z) < M (A, Z+1) + me
Because of the vacancy created in the electron shell, there is the transition of one of the shell
electrons to that vacancy, accompanied by the photon emission, in the X-ray band.
233
7.A.2.3 Gamma Rays
Unstable nuclei going from an excited energy state down to a less energetic, eventually
stable, state can emit energy quanta in the γ rays wavelength (10-8 ≥ λ ≥ 2*10-11 cm). There can be
single transitions, where the nucleus goes directly from an excited state to the ground, or stable,
state following the emission of a single γ quantum, or there can be multiple transitions, i.e., a
cascade of transitions bringing the nucleus to the ground state and involving multiple emissions of γ
quanta. The energy of the γ quantum emitted is determined by the difference in energy of the two
energy levels between which the transition has occurred.
There are different mechanisms responsible for exciting nuclei and leading to gamma
radiation. In fact, quite commonly alpha and beta decays can leave the nucleus in an excited state.
An alpha decay is usually followed by the emission of low energy γ-quanta (< 0.5 MeV), while
after a beta decay higher γ-quanta are emitted (energy up to 2-2.5 MeV) [1].
7.A.3. RADIATION AND DOSE QUANTITIES AND UNITS
An ad hoc set of quantities and related units required to describe radiation decay and its
effects has been introduced since the effects of nuclear radiation were discovered and gradually
understood. A list of them follows.
7.A.3.1 Activity (Bq)
Given any radiation decay (α, β, γ, etc.), the activity of an element is the rate at which any
and all transitions (i.e., emission of α, β, γ rays) occur. A radionuclide has an activity of 1
Becquerel (Bq), when it undergoes one transition per second. An older unit, going back to M.me
Curie, is the Curie (Ci), equivalent to 3.7 * 1010 transitions per second. This is ‘the quantity of
emanation in equilibrium of 1 gram of Radium’ (M.me Curie said: ‘la quantité d'émanation en
équilibre avec un gramme de radium’), that is that quantity of Radon-222 in equilibrium with one
gram of its parent Radium-226 [2]. It is worth to note here that not only for activity but also for all
other quantities both SI units and old ones, partly deriving from the CGS system, are currently used.
1 Bq = 1 transition/second
1 Ci = 3.7 * 1010 Bq
Activity is not a synonymous of power or energy and has nothing to do with the effects of
radiation on matter, living or not.
7.A.3.2 Half Life (s)
The half life is the time period over which half the nuclei of a given radionuclide decay. The
half life, depending on the radionuclide considered, varies 214 Po has 164 μs). An example may
help: 214Pb has a half life of 26.8 min, and this means that if there are N nuclei of 214Pb at time zero,
after 26.8 minutes there will be N/2 nuclei (the N/2 left have become 214Bi because of beta decay),
after 53.6 minutes there will be N/4 nuclei of 214Pb (3/4N have become 214Bi) and so on.
234
7.A.3.3 Absorbed Dose D (Gy)
When radiation passes through matter it releases energy. The absorbed dose is the energy
deposited by radiation inside matter per mass unit. Its SI unit is the Gray (Gy), equivalent to 1
Joule deposited per kilogram of absorbing target material (1 J/kg). The older unit is the RAD
(Radiation Absorbed Dose), defined as the deposition of 100 erg per gram [3].
1 Gy = 100 rad
7.A.3.4 Equivalent Dose H (Sv)
Biological effects caused by radiation are not only dependent upon the dose absorbed (Gy)
but also, and above all, by the kind of radiation. ‘Sparsely’ ionizing radiations such as gamma, x-
ray or beta rays are less effective in damaging than ‘densely’ ionizing radiation such as alpha
particles or fission fragments. To account for this difference, a weighing factor dependent on the
kind of radiation and energy has been introduced. The weighing factor goes from 1 (for photons or
electrons) up to 20 (for alpha particles), and is dimensionless (see Table XXX) [3].
Radiation and Energy Weighing Factor, wr

Photons, all energy 1
Electrons, all energy 1
Neutrons, < 10 KeV 5
10 – 100 KeV 10
100 KeV – 2 MeV 20
2 MeV – 20 MeV 10
> 20 MeV 5
see also figure 1
Protons, all 1
Protons, (not recoil) > 2 MeV 5
Alpha particles, all energy 20
Fission fragment, all energy 20
Heavy nuclei, all energy 20
Table XXX Weighing factors for different types of radiation
Figure 127 Weighing factor for neutrons
The sum of the total radiation doses D combined with the proper weighing factor wr gives
the equivalent dose H [3]:
235
H = ∑ wr * D
Since wr is dimensionless, the equivalent dose H has the same dimensions as the absorbed
dose D, i.e., Joule per kilogram. Its SI unit is the Sievert (Sv). The older unit is the REM (or rem)
(Roentgen Equivalent Man), whereby
1 Sv = 100 rem
7.A.3.5 Effective Dose E (Sv)
Consequences of radiation on human body depend on the particular organ or tissue hit by
radiation, as different organs have different responses to radiation exposure. This is the reason why
another weighing factor (wT) must be introduced (see Table XXXI) [3]:
Organ or Tissue Weighing Factor, WT

Gonads 0.20
Red bone marrow 0.12
Colon 0.12
Lung 0.12
Stomach 0.12
Bladder 0.05
Breast 0.05
Liver 0.05
Oesophagus 0.05
Thyroid 0.05
Skin 0.01
Bone surfaces 0.01
Remainder 0.05
Table XXXI Weighing factors for tissue/organ
The sum of the equivalent dose D with the tissue weighing factor gives the effective dose E
[3]:
E = ∑ wT * H
The dimensions of the effective dose are the same as absorbed dose and equivalent dose,
Joule per kilogram. Its SI unit is the same as that of the equivalent dose: Sievert.
7.A.3.6 Collective Dose (man Sv)
Absorbed, equivalent and effective dose apply to individuals or average individuals. In order
to assess the dose received by a group or population, it is more practical to introduce the collective
dose. It is obtained by summing up the individual doses of each person of the group considered. Its
SI unit is man Sv. A collective dose of 1000 man Sv corresponds to 1000 people receiving each 1
mSv or 10 people 100 mSv. This quantity is defined for a specific source of radiation or for a
specific practice causing exposure, and is convenient when considering nuclear accidents [4].
236
7.A.3.7 Dose Commitment (Sv)
Some events, such as weapon tests, release radioactivity directly into the environment and
cause a continuous exposure over a long time period, that may include several generations. To
account for the dose committed to a typical, though hypothetical, individual from a given moment
into the future, the so-called “dose commitment” has been defined. This quantity is the integral over
a specified time period (typically 250 or 10000 years) of the average dose rate, per person, to a
specified group (even the whole world population) after the event considered. Its SI unit is the
Sievert (Sv) [4]. If an event delivers a dose commitment of 1.4 mSv for 250 years, a hypothetical
individual, born at the moment of the event and died 250 years old, would receive a dose of 1.4
mSv during his entire life.
7.A.4. EFFECTS OF IONIZING RADIATION
Ionizing radiation interacts with matter changing the state of atoms and molecules. In cells
there are two types of consequences after radiation interaction: the cell may die or may be modified.
These two different outcomes have different implications for the whole body: in fact, there can be
deterministic and stochastic effects.
7.A.4.1 Deterministic Effects
Radiation may kill cells of a tissue or organ. If the numbers of cells killed is low, the tissue
keeps functioning without serious consequences. If the number of cells killed is sufficiently large,
the tissue is harmed and may lose its function; eventually, the tissue or even the organ itself may
die.
It is clear that an increasing number of dead cells causes more and more serious damage to the
tissue. The specifics of the outcome depend on the fact that cell depletion is a dynamic process, in
competition with proliferation of unaffected cells. If the loss of cells is low it can be quickly
compensated by repopulation (with no damage, or short time effects). If the loss is large there is a
drastic non compensated reduction of tissue cells (resulting in serious damage and/or death). The
fraction of cells killed depend on dose: therefore the severity of effects depends on dose as well.
These effects are defined deterministic, and have dose thresholds.
Some deterministic effects are: temporary or permanent sterility, depression of the blood-
forming system, skin reddening, desquamation, skin loss, lens inflammation, cataract. A peculiar
case of deterministic effect is the radiation syndrome from acute and whole body irradiation. If the
dose is high enough, the strong cell depletion in vital organs (blood-forming organs, gastro-
intestinal tract etc.) causes death. An acute whole body exposure dose between 3 and 5 Gy, without
any specific medical treatment, causes the death of 50% of the people exposed.
Table XXXII shows some thresholds for deterministic effects. The thresholds, like all
thresholds for deterministic effects, apply to people in normal health [4].
237
Deterministic Effect Threshold, Gy
Male temporary sterility
acute exposure 0.15
chronic exposure (per year) 0.4
Male permanent sterility
acute exposure 3.5-6
chronic exposure (per year) 2
Female permanent sterility
single exposure 2.5-6
chronic exposure (per year) 0.2
Depression of blood formation
acute bone marrow exposure 0.5
long-term exposure (per year) 0.4
Lens opacities (sparsely ionizing radiation) 2-10
Lens opacities (densely ionizing radiation) 1-2
Lens opacities (chronic exposure to sparsely ioniz. rad. per year) 0.15
Dry skin desquamation (3 weeks after exposure) 3-5
Moist desquamation (blistering after 1 month) 20
Tissue necrosis 50
Table XXXII Threshold for deterministic effects
7.A.4.2 Stochastic Effect
If a cell is not directly killed by radiation but somehow modified, the outcome will be
different from those included among deterministic effects. In-vitro cellular research shows that
damage from radiation to deoxyribonucleic acid (DNA) causes the most of detrimental effects.
There are two mechanisms by which radiation may damage DNA: direct or indirect interaction.
In the first case ionizing radiation directly damages a gene, in the second radiation produces
active chemical radicals near the DNA. The radicals may diffuse and interact with DNA, inducing
chemical changes. Very efficient mechanisms exist (enzyme actions) that repair DNA, whatever the
cause of harm. For instance, if only one of the two symmetric strands forming the DNA is damaged,
information on the other strand makes the repair process highly probable and successful, though
not always error free. It is this repair process that is activated and energized by radiation that is
the basis for a field of radiation effects called radiation hormesis, to be discussed later. Radiation
hormesis, is actually beneficial to organisms as will be discussed later. If both strands are damaged
at the same location, information is lost forever: the repair process is more difficult and genetic
changes are likely. Such changes are defined genetic mutations. The very nature of this
damage/repair process causes effects that are random and statistical, therefore called stochastic.
Stochastic effects can be somatic (i.e., cancer induction), that is, they occur in the individual
exposed, or hereditary, when the cells damaged are those whose function is to transmit genetic
information to offspring. There is increasing evidence that below a certain dose repair process is
highly effective, reversing even effects of chemical oxides, peroxides and super oxides within cells.
This process is in direct opposition to the linear no threshold concept. However, since stochastic
effects may have no dose lower bound, there is no threshold in this case [4].
7.A.4.2.1 Radiation Induced Cancer
There is substantive evidence that almost all cancers originate from a single cell. However,
single changes in the cell genetic code are usually insufficient to initiate a cancer. Several cell
mutations (two to seven) are required in carcinogenesis to go from pre-neoplasia (cancer
precursors) to cancer. Radiation may act at several stages of this process, but it seems to have a
major role in the initial conversion of the cell to a pre-neoplastic state. A pre-neoplastic cell is
238
immersed in an environment of normal cells, which tend to suppress and constrain pre-neoplastic
properties. Overcoming these constraints results in a cancer.
Cancer may be triggered by many factors such as smoke, chemical agents and others, and it is
therefore impossible to determine whether radiation is the cause of a particular type of cancer or
not. The only way to ascertain a correlation between radiation and cancer induction is statistical.
Epidemiology is the study of the distribution of diseases among people, and is regrettably still an
observational rather then experimental science: bias or confounding factors are highly probable.
In the present context, the so-called Life Span Study (LSS), an ad-hoc study of survivors of
Hiroshima and Nagasaki, has produced a significant amount of data on effects of exposure to
radiation on humans. Studies of people partially exposed to radiation due to medical investigations
or treatments are another source of data, together with information available from studies of
occupational exposures, i.e., in the Mayak facility in Russia, and the Chernobyl accident (see
Appendix 7.C???5].
From a general point of view, linear (or linear-quadratic) no-threshold dose response is to be
expected, even though for certain cancers and at low doses correlations are less reliable.
Some interesting results are those for solid cancers obtained by LSS, where EER (Excess
Relative Risk) (Figure 127) and EAR (Excess Absolute Risk) (see Figure 128) are estimated. EER
and EAR represent the increased cancer rate in an exposed group relative to an unexposed group,
measured on relative and absolute scales. An EER of 1 corresponds to a doubling of the cancer rate.
EAR may be expressed as the number of excess cases of cancers per, for example, 10000 persons.
They can be expressed per unit dose or per a specific dose (i.e., 1 Sv) [5].
7.A.4.2.2 Radiation Hormesis Considerations
The Hiroshima/Nagasaki data clearly dispute the Linear No-Threshold hypothesis as shown in
Figure 128.
Figure 128 ERR at 1 Sv
239
Figure 129 EAR at 1 Sv
Figure 130
Radiation Hormesis Data From Hiroshima and Nagasaki Survivors
The above data indicate that the actual Leukaemia deaths do not follow the LNT hypothesis, whose
intersection at zero dose is shown in the overlay line. The problem with the LNT hypothesis is that
it ignores realities of dose in other circumstances. This, when coupled with a no de Minimis dose
can become very problematic. For example, it is well known that 100 aspirin in a single dose can
kill. If we follow the LNT concept, it means that a person taking one aspirin a week for 100 weeks
will die, or out of 100 persons taking a single aspirin, 1 will die [16]. ∟ However there exists
more data beyond Hiroshima and Nagasaki, data exist for U.S Naval Nuclear Shipyard workers as
shown in Figure 129.
240
Figure 131 U.S Nuclear Naval Shipyard Workers (Matanowski, 1991)
From Figure 129, we note that the control group of 20,419 unexposed workers, and the nuclear
naval shipyard workers indicated a reduced mortality rate at a given age than did unexposed
workers. The same effect is observed to be true for U.S. nuclear weapons workers. It should be a
priority for the IAA to establish the importance to cost and risk reduction to adopt the radiation
hormesis approach to radiation protection.
7.A.4.2.3 Hereditary Effects
No radiation-induced hereditary disease has been demonstrated in humans so far. However

ionizing radiation is recognized as mutagenic (i.e., capable of changing hereditary traits).
Experiments on plants and animals have clearly shown that radiation may cause genetic effects, and
there is no reason to believe that humans are an exception.
It has been estimated that, for a population exposed to radiation in one generation, the risk,
expressed as number of cases per million people per Gy, in the progeny of the first post-radiation
generation is: 750-1500 autosomal-dominant and X-linked diseases; 250-1200 chronic multi-
factorial diseases; and 2000 congenital abnormalities. The total radiation-induced cases are 3000-
4700 per Gy per million and constitute 0.41-0.67% of the total 738,000 cases per million [6].
7.A.5. SOURCES OF RADIATION EXPOSURE
Radiation to which humans are exposed comes from various sources. It can be natural
radiation, or may be produced by human activities.
7.A.5.1 Natural Radiation Exposure
Natural radiation, also defined background radiation, has always existed in nature. Life on
Earth has developed and keeps proliferating in a naturally radioactive environment. There are
different sources of background radiation, responsible for either internal or external exposure.
Table XXXIII shows the doses from natural sources. The worldwide annual effective dose is
2.4 mSv and, considering a 5.3 billion world population, the collective dose is 13*106 man Sv [5].
241
7.A.5.1.1 Cosmic Rays
Cosmic rays are a source of external exposure. They can be divided into primary and
secondary radiation. Primary radiation can be further divided, depending on its origin, into galactic
and solar, the second being less significant on the Earth ground. Outside the Earth atmosphere the
main component of cosmic radiation are positively charged particles, mostly protons, of energy
between 102 and 105 MeV; they constitute the so-called primary radiation (galactic and solar).
When these particles approach Earth they are deflected by the terrestrial magnetic field according to
their momenta. In their travel toward the ground, primary radiation particles interact with the
atmosphere producing many particles such as electrons, photons, mesons, protons and neutrons:
these are called the secondary radiation.
Secondary radiation particles themselves can interact with the atmosphere, or decay,
producing so-called avalanche ionization: from a single-particle starting event, up to 108 particles
can be generated. In the atmosphere below 20 km cosmic radiation is constituted almost exclusively
of its secondary component [7]. Typical range of effective dose per person per year is 0.3-1.0 mSv,
with average effective dose ~0.4 mSv [5]. For locations high above the sea level very large doses
are receive, e.g., in La Paz – Bolivia (3600 m) the average dose due to cosmic rays is 2.02 mSv per
year. Flying at 8000 m altitude results in a dose rate of 2.8 μSv h-1 [7]. Cosmic rays are an
important and serious consideration in interplanetary travel [17].
7.A.5.1.2 Terrestrial Radiation
Inside the Earth there are radionuclides whose half life (T1/2) is comparable with Earth’s age.
In fact the Earth’s core is still hot thanks to the energy released by radionuclides in their
decay processes. The most significant for dose computation are 40K (T1/2=1.28*109 y), 232Th
(T1/2=1.41*1010 y), 238 U (T1/2=4.47*109 y); of secondary importance are 87Rb (T1/2=4.7*1010 y) and
235
U (T1/2=7.04*108 y). Most radionuclides belong to one of the three families of Uranium, Thorium
and Actinium (see Figure 130Figure 132) [7]. In all three families Radon (Rn) is formed. Radon
appearance is the clearest evidence that Earth crust is radioactive. Terrestrial radiation can be
responsible for internal or external exposure.
Figure 132 Uranium-238 decay chain
242
Figure 133 Thorium-232 decay chain Figure 134 Uranium-235 decay chain
243
7.A.5.1.2.1 External Exposure from Terrestrial Radiation
External exposure to gamma rays from natural radionuclides can occur both outdoor, since
radionuclides are present in the Earth crust, and indoor, as they may be present in construction
material, or seep through building foundations. Combining outdoors and indoor exposure, for a
person spending 80% of time indoor, a range of 0.3-0.6 mSv per person per year is typical.
Worldwide-averaged annual effective exposure is estimated ~0.5 mSv [5].
244
7.A.5.1.2.2 Internal Exposure from Terrestrial Radiation
Potassium isotopes are present in the human body with a weight percentage 0.18%; the
isotope 40K has an isotopic abundance 1.18*10-4, and its main decay mechanism is beta. The annual
dose from 40K is estimated 0.165 mSv. Some isotopes (the most significant 210Pb and 210Po) can be
ingested through food and water. The typical range of the annual effective dose is 0.2-0.8 mSv, but
higher values are detected in South America (due to large quantity of 210Po present in ‘yerba mate’,
an herb used in drinks) and arctic and sub-arctic areas (where 210Po and 210Pb tend to accumulate in
moose meat). The worldwide-averaged annual effective dose is 0.3 mSv
Some radioisotopes may be inhaled, the most significant in this case being 222Rn and, much
less importantly, 210Po (smoking 10 cigarettes a day doubles 210Po introduction). Typical range of
inhaled dose is 0.2-10 mSv. The range is so wide because the contribution is mainly given by radon
and is contribution depends on its indoor accumulation. The worldwide-averaged annual effective
dose due to inhalation is 1.2 mSv. The summary of background radiation sources is in Table
XXXIII [5].
Source Worldwide average annual effective dose (mSv) Typical range (mSv)
External Exposure
Cosmic rays 0.4 0.3-1.0
Terrestrial gamma rays 0.5 0.3-0.6
Internal Exposure
Inhalation (mainly radon) 1.2 0.2-10
Ingestion 0.3 0.2-0.8
Total 2.4 1-10
Table XXXIII Mean dose value for background radiation
7.A.5.2 Medical Radiation Exposure
Ionising radiation for medical purposes, both in diagnosis and in treatment, is widely used. It
must be noted that most of these procedures are carried out in countries where only one quarter of
the world population lives. World health care has been divided into four qualitative levels,
depending on the number of physicians available.
Diagnostic exposures are characterised by low doses to individuals, while therapeutic
exposure is usually much larger. High doses are used to treat diseases, especially cancer. The
number of diagnostic procedures is much larger than treatment procedures (the ratio is about 450
to1): this is due to the widespread use of X-rays (they contribute to 78% of collective dose).
The worldwide-averaged annual effective dose is 0.4 mSv, the total collective dose estimated
is 2500*106 man Sv. Table XXXIV shows effective doses reported for each health care level [5].
Table XXXV [7] shows the effective dose for some diagnostic examinations.
245
Health care level Population per physician Annual number of examinations Average annual effective
per 1000 persons dose to population (mSv)
I <1000 920 1.2
II 1000-3000 150 0.14
III 3000-10000 20 0.02
IV >10000 <20 <0.02
Worldwide average 330 0.4
Table XXXIV Average dose from medical use
Examination Effective dose per examination (mSv)

Chest radiography 0.14
Mammography 0.5
Angiography 12
Urography 3.7
Dental 0.03
Table XXXV Doses from some examinations
7.A.5.3 Exposure from Atmospheric Nuclear Testing
Until the Treaty Banning Nuclear Weapon Tests in the Atmosphere, in Outer Space, and
Under Water, signed in Moscow on August, 5th 1963, almost all nuclear explosions to test weapons
(fissions and fusions) were carried out in the northern hemisphere atmosphere. For instance, in the
former Soviet Union at Semipalatinsk, in Kazakhstan, 456 tests were carried out between 1949 and
1989 [12]; after the treaty almost all tests have been conducted underground. The two time periods
of most intense atmospheric tests are 1952-1958 and 1961-1962 (see Figure 130). The total number
of atmospheric tests is 543 and the total yield estimated is 440 Megatons (189 Megatons from
fission) [5].
Figure 135 Number of weapons tests
The total collective effective dose resulting from weapon tests to date is 3*107 man Sv; 7*106
man Sv will be delivered within the first 250 years (until 2200) the remainders, due to the long life
of the C14 radionuclide produced, in the next 10000 years. The annual average effective dose varies
both with time (decreasing thanks to the ban treaty) and with location: in the northern hemisphere
the dose is higher than in the southern. The average effective dose estimated for the year 1999 is
5.87 μSv in the northern hemisphere, 2.68 μSv in the southern and 5.51 globally (see Table
XXXVI).
246
Average Annual Effective Dose μSv
Year Northern Hemisphere Southern Hemisphere World
1945 0.64 0.57
1955 16.8 3.34 15.4
1965 48.7 11.7 44.6
1975 14.8 5.01 13.7
1985 8.98 2.78 8.30
1995 6.61 2.55 6.20
1996 6.42 2.57 5.97
1997 6.23 2.59 5.85
1998 6.05 2.63 5.63
1999 5.87 2.68 5.51
1945-1999 1076 328 994
1999-2099 264 157 253
2099-2199 63 53 62
2200-_ 2181 2180 2181
1945-_ 3580 2720 3490
Table XXXVI Doses from weapons tests
7.A.5.4 Nuclear Power Generation by Utilities
Today about 17% of electricity produced worldwide, i.e. about 250 gigawatt, is from nuclear
power plants. Assuming that this practice continues over the next 100 years, the maximum
collective dose can be estimated from the cumulative dose over the period of practice. The
normalized 100-year collective dose is 6 man Sv per gigawatt and per year. The annual dose is 1500
man Sv (6 x 250), resulting in a maximum annual dose per person of 0.2 μSv [5].
7.A.5.5 Exposure from Major Accidents
There have been accidents in using nuclear energy or radioactive elements. In medical and
diagnostic practice accidents may occur (a few hundreds of all types each year), and usually have
serious consequence. The probability that any member of the public be involved is, however, very
small, and by and large, the consequences do not affect the public.
Weapons production and transportation have resulted in several accidents, but the collective
dose committed is small. The two most serious accidents in nuclear weapons production were at
Kyshtym, in the former USSR, and at the Windscale plant at Sellafield (UK), both in 1957. The first
accident caused a collective dose of 2500 man Sv over the next 30 years. The Sellafield accident
caused a total collective dose in Europe (including England) of about 2000 man Sv.
The two most important accidents in power plants were those at Three Mile Island and
Chernobyl, although the Chernobyl installation produced energy only as a byproduct, the plant
being chiefly a plutonium-producing facility, and although what happened can hardly be defined an
accident. At Three Mile Island the containment system, missing at Chernobyl, prevented a large
amount of fission fragments from spreading in the environment: the total collective effective dose
was ≤ 40 man Sv, with the maximum dose to nearby individuals ≤ 1 mSv. The Chernobyl accident
(see its detailed account in Appendix 7.B), had much more serious consequences. It caused the
death of 30 people among the rescue workers within a few weeks, and 1800 cases of thyroid cancer
in the children exposed; no other health impact has been detected up to the year 2000. The
247
worldwide average annual effective dose per person due to the Chernobyl accident, estimated for
the year 2000, is 0.002 mSv, down from its peak 0.04 mSv in 1986 [5].
7.A.5.6 Occupational Exposure
There are jobs in which workers are routinely exposed to radiation, both because of man-
made sources (i.e., medical practice, people employed in nuclear fuel cycle facilities etc.) and
because of enhanced levels of natural radiation (i.e., airplane crews flying at a height of 8000 m
receive a dose of 2.8 μSv per hour). This kind of exposure does not affect other members of the
public, but it is interesting to see the dose (see Table XXXVII) that these workers receive in order
to have a better understanding of the issue [5]:
Source/practice Number of monitored workers Average annual effective dose

(thousands) (mSv)
Man-made sources
Nuclear fuel cycle (including uranium mining) 800 1.8
Industrial uses of radiation 700 0.5
Defence activities 420 0.2
Medical uses of radiation 2320 0.3
Education/veterinary 360 0.1
Total from man-made sources 4600 0.6
Enhanced natural sources
Air travel (crew) 250 3.0
Mining (other than coal) 760 2.7
Coal mining 3910 0.7
Mineral processing 300 1.0
Above ground workplaces (radon) 1250 4.8
Total from natural sources 6500 1.8
Table XXXVII Occupational exposure
7.A.5.7 Exposure from Future Nuclear Propulsion Systems
A new source of dose could in principle result due to future nuclear propulsion systems.
Rubbia’s engine and MITEE are two of the most promising systems: an assessment of the dose
committed to the public arising from their use is necessary in order to show the impact they could
have.
To set to rest a very old misconception, there is literally no way a nuclear reactor, whether
for power generation or propulsion, could trigger a nuclear explosion: the reason is the impossibility
of reaching the proper conditions of confinement time and critical mass.
However, what could happen is that because of coolant loss, or other reasons, ‘runaway’
fission in a reactor can heat too much the reactor core, eventually melting it down. This effect is
called a Loss of Coolant Accident, or LOCA). When this happens (it did in the case of Chernobyl),
high temperature chemical reactions can occur, especially if water or graphite moderators are
present. Water could be dissociated by the high temperatures, producing hydrogen and oxygen and
possibly burning or exploding, and graphite could burn in an oxygen or hydrogen atmosphere.
Besides, excessive heat release rates may also cause explosions simply due to rapid thermal
expansion of the nuclear ‘fuel’ or other reactor material. LOCAs are most serious in nuclear
reactors. In absence of a containment structure, radionuclides from the core can be ejected by the
chemical or thermal explosion and contaminate the nearby environment.
248
This said, it should be clear that this type of accident is in fact due to chemistry, not fission (the use
by the popular press of the word ‘nuclear explosion’ in this context is due to ignorance and is
misleading). In space, it is highly unlikely that there will be the necessary constituents available to
cause the spectacular failures such as those seen in Chernobyl.
To test the effects of an actual meltdown due to runaway fission, during the NERVA program a test
was performed at Los Alamos in which a KIWI nuclear reactor was deliberately let to explode by
excluding the cooling system (this was the so-called KIWI-TNT test). The results are reported in
[13-14]. The reactor was totally destroyed, but contamination was limited to a relatively small area
(of order 100 m). After clearing appropriately the site from debris, activities could be resumed. This
test did much to allay fears that a NERVA-type core meltdown and explosion could in any way
produce a large scale catastrophe. A nuclear rocket reactor must be inherently far smaller than that
in power plants, so the outcome of the KIWI-TNT test is not surprising.
A specific and more serious concern in propulsion applications, a nuclear reactor must be orbited,
i.e., lifted through the Earth atmosphere, perform its interplanetary mission starting from LEO or
MEO, and (possibly) be parked again in Earth orbit at the end of its mission. The question is, what
could happen during each of these three legs.
Any reactor will contain fissile fuel, of order O(1) to O(10) kg depending on fuel type. Of course no
reactor will be operated while being lifted off, but the danger exists of an accident, such as that of
the “Challenger”, in which the conventional launcher could explode, damaging the reactor to be
orbited and spreading fissile material from the damaged reactor stored in the payload bay either in
the atmosphere or on the ground.
During the interplanetary trajectory instead any accident would not affect Earth. The most
dangerous occurrence would be if the reactor, containing all the [new] radionuclides produced
during operation in space, would for some reason re-enter the Earth atmosphere accidentally: in
fact, no space agency is considering re-entering nuclear reactors, so that such event would have to
be unplanned, unwanted and therefore accidental.
The consequences would be of spreading many radionuclides families in the atmosphere, at a height
that can be estimated roughly between 40 and 10 km, at the peak of aerodynamic heating. The total
mass of radionuclides spread would be approximately the same of the original fuel, i.e., O(1) to
O(10) kg. Additional contamination would come from secondary radioactivity, that is, induced in
the reactor structural materials.
Insofar the actual consequences, this event is similar to what happens during an atomic explosion in
the atmosphere, where fissionable material and bomb structure are vaporized and released into the
atmosphere. Data from atmospheric atomic tests exist that can be effectively estimate these effects.
In any event, the quantity of radionuclides in an atomic explosion is many times larger that in any
nuclear reactor at this time envisaged for space missions; accordingly, radioactive contamination is
expected to be smaller.
7.A.5.7.1 Nuclear Accidents in a Rubbia’s Engine
Like all other nuclear propulsion concepts, the Rubbia’s engine is not projected to fission
whilst in the atmosphere. The dose to the public to would be the highest in a hypothetical accidental
re-entry, for instance at the end of a Mars mission. For each kg of Americium loaded, it is estimated
that the total collective dose committed for the following 250 years be 9.5 man Sv. The individual
dose commitment over the following 250 years would be 1.8*10-6 mSv. In the case of an
Americium stockpile of 15 kg, typical of a manned Mars mission using the Rubbia’s engine, the
total collective dose committed for the first 250 years would be 140 man Sv, while the individual
value would be 3*10-5 mSv [8].
249
In addition, the fuel considered in Rubbia engine (Am-242m) was purposely chosen because of its
neutron cross section sharply decreasing with temperature. This feature means that any runaway
fission in Am 242m would automatically stop above a certain temperature, and the reactor regime
brought back to a stable state.
7.A.5.7.2 Nuclear Accidents in a MITEE Engine
Also in the case of MITEE the most catastrophic accident could be the total destruction of the
vehicle accidentally re-entering in the atmosphere after a mission. Like Rubbia’s Engine, MITEE is
planned not to fission while in the Earth atmosphere, so that also in this case, a prompt criticality
accident (explosion caused by overheating) would have less considerable consequences than the
total destruction for a chemical explosion or unwanted re-entry in the atmosphere after returning
from a mission to Mars. The average dose commitment over the following 250 years would be
about 1.6 * 10-8 mSv for each kg of Uranium loaded, and for a typical MITEE configuration the
average dose commitment for 250 years would be therefore about 4*10-7 mSv [4, 9, 10, 11].
7.A.5.7.3 Safety in Ground Testing of Future Nuclear Rockets
A key worry in planning Nuclear Propulsion revolves around the issue of ground testing. In the
past, KIWI, NERVA and PHOEBUS were all tested at Los Alamos in the open air.
A recent book by Dewar recounts details of those tests and the safety measures employed; it
suffices to say here that no accidents involving loss of life or damage to people took place during
the entire US program [13, 14].
Nevertheless, planning future ground tests is a concern. However, at least in the case of the type of
reactor envisaged by C. Rubbia and investigated by the Italian Space Agency ASI under the Project
P 242, the following considerations apply.
The Rubbia engine is modular, each module being a self-standing generator of hot hydrogen gas.
About 30 to 40 modules compose the engine.
For a manned Mars mission the thrust F required is of order 103 N, while the specific impulse Isp is
of order 2500 s. Comparison with the NERVA thermal engine tested at Los Alamos (F = 334,000
N, Isp = 825 s, mass flow rate = 40 kg/s) shows that the single module of the Rubbia engine to be
tested in an appropriate test facility will process a mass flow rate of hydrogen of order 2.5 g/s. So,
the scale factor between a module of Rubbia engine and NERVA is about 16,000. The amount of
hydrogen, and therefore of fission fragments deposited inside the hydrogen used as propellant, will
be exceedingly small.
As a consequence, testing a single module of the Rubbia engine may be performed in a closed loop,
and this appears also feasible for all nuclear rockets of comparable thrust, therefore also MITEE, or
NER. In fact ways of efficiently separating fission fragments from hydrogen have already been
described in the Final Report of ASI P 242 Phase A. Closed loop tests can be performed in any
reasonably self contained facility and building, thus ensuring no radiation may escape.
7.A.5.8 Comparison of Exposures
The doses received by an individual from the main different sources in year 2000 are
summarised in Table 9. Their values are given in annual per caput effective dose (mSv). The values
are averaged, meaning that there are significant variations in exposure to individuals, depending on
location, diet, personal habits and so forth.
The largest contribution to total dose is from the natural background: 2.4 mSv, but typical
values may range from 1 up to 10 mSv, with large groups of population receiving a dose of 10-20
250
mSv. The second most important source, 0.4 mSv, from the medical use of radiation. It has an
increasing trend, thanks to increasingly available medical radiation facilities. The third cause is the
fallout from past weapons tests; i.e., 0.005 mSv. The value has been decreasing thanks to the Treaty
Banning Nuclear Weapon Tests, the maximum value being reached in 1963, when it was 7% of the
natural background. Other man-made sources, like the Chernobyl accident and nuclear power
production, are much smaller, 0.002 mSv and 0.0002 mSv, respectively.
Source Worldwide annual Range or trend

per caput dose (mSv)
Natural background 2.4 Typical range 1-10 mSv.
Sizeable population also 10-20 mSv
Diagnostic medical use 0.4 Typical range 0.04-1.0 mSv at lowest and highest
level of health care
Atmospheric nuclear testing 0.005 Has decreased from a maximum of 0.15 mSv in
1963. Higher in northern hemisphere and lower
in southern hemisphere
Chernobyl accident 0.002 Has decreased from a maximum of 0.04 mSv in
1986. Higher in locations near the accident area
Nuclear power production 0.0002 Has increased with expansion of plants but
decreased thanks to improved practice
Table XXXVIII Annual per caput doses in the year 2000
7.A.6. CONCLUSIONS
The individual dose commitments for 250 years arising from a rather improbable total ‘crash’
of Rubbia’s engine, 1.8*10-6 mSv, and MITEE, 1.6*10-8 mSv, are insignificant compared to other
sources of exposure. Should the Rubbia’s engine ‘crash’, a hypothetical individual born in the year
of crash and dying at age 250, would have received all along his life a dose of 3*10-5 mSv, much
lower than the dose imparted by a dental examination (0.03 mSv); the same would be true for a
MITEE accident of the same type. The average dose from natural background to each individual is
2.4 mSv in one single year. Table 10 shows contribution to dose compared to other sources.
The contribution to individual average dose from the crash of Rubbia’s engine or MITEE is
therefore not a reason of concern to public health.
Source Effective Dose/Dose Commitment (mSv) Comment

Rubbia’s Engine Accident → 1.8*10-6 Dose committed for 250 years
Catastrophic LEO Re-entry (per kg fuel)
MITEE Accident → 1.6*10-8 Dose committed for 250 years
Catastrophic LEO Re-entry (per kg fuel)
Natural Background 2.4 Average effective dose in 1 year
Dental x-ray examination 0.03 Average effective dose from a
single examination
Flying at 8 km for 10 hours 2.8*10-8 1 hour gives 2.8*10 μSv
Table XXXIX Comparison of doses from different sources
As a general conclusion, we note that nuclear thermal propulsion and nuclear electric propulsion are
exceptional enablers to space exploration, and that doses from nuclear reactors in space is
251
insignificant compared to the space environment itself. In fact, the reduced trip times from the use
of nuclear propulsion will enhance safety due to reduced exposure to zero gravity and the cosmic
ray environment of space. We note that there is extensive data supporting the radiation hormesis
approach to radiation health effects as opposed to the linear no threshold hypothesis should guide
the radiation safety doctrine for space nuclear power and propulsion systems, including ground
testing. It is likely that no other single event will assist the technical development of nuclear power
and propulsion than the reduction in costs and regulatory burdens generated by eliminating the
linear no threshold hypothesis as the basis for radiation health effects.
REFERENCES
1. K.N. Mukhin, Physics of Atomic Nucleus Volume I. Mir Publishers, Moscow,1987

2. http://physics.nist.gov/GenInt/Curie/1913.html
3. ICRP Publication 60. 1990 Recommendations of the International Commission on
Radiological Protection. Annals of the ICRP – Volume 21 n. 1-3, 1991
4. UNSCEAR (1993). United Nations Scientific Committee on the Effects of Atomic
Radiation. Sources and Effects of Ionizing Radiation. UNSCEAR 1993 Report to the General
Assembly, with Scientific Annexes. United Nations, New York.
Radiation. Sources and Effects of Ionizing Radiation. UNSCEAR 2000 Report to the General
Assembly, with Scientific Annexes. United Nations, New York.
Radiation. Hereditary Effects of Radiation. UNSCEAR 2001 Report to the General Assembly,
with Scientific Annexes. United Nations, New York.
7. G. Galli and C. Mancini, Esposizione alla radioattività ambientale, Ingegneria Nucleare e
Tecnologie Energetiche, 38, n.1-4, Jan-Aug. 1996
8. C. Rubbia, Fission Fragments Heating for Space Propulsion. European Organization for
Nuclear Research. Geneva 2000
9. J. Powell, G. Maise and J. Paniagua, MITEE: An Ultra Lightweight Nuclear Engine for New
and Unique Planetary Science and Exploration Missions. IAF-98-R.1.01 49th International
Astronautical Congress, Sept. 28-Oct. 2, 1998, Melbourne, Australia
10. J. Powell, G. Maise and J. Paniagua, The MITEE Family of Compact, Ultra Lightweight
Nuclear Thermal Propulsion Engines for Planetary Space Exploration. IAF-99-5.6.03 50th
International Astronautical Congress, Oct. 4-8, 1999, Amsterdam, The Netherlands
11. G. Maise, Personal communication
12. http://www.nato.int/science/e/grants
13. J. A. Dewar, The Story of The Nuclear Rocket: Lessons for The Future. IAC-02-IAA.2.4.06
rd
53 International Astronautical Congress-The World Space Congress, Oct. 10-19, 2002,
Houston, Texas, USA
14. J. A. Dewar, To the End of the Solar System: The Story of the Nuclear Rocket, The
University Press of Kentucky, Lexington, KY, USA, 2002
15. H.A. Klein, The Science of Measurement: A Historical Survey, Dover, New York, Ch. 49.
16. Frunze, J.W., ISU Engineering, October 28, 1999.
17. Parker, E.N., “Shielding Space Travellers,” Scientific American, V. 294, No. 3 March 2006.
252
7.B. The Chernobyl Accident - A Detailed Accounty A. Del Rossi and C.
Bruno
7.B.1. INTRODUCTION
Nuclear Propulsion is one of the most promising propulsive systems. One of the major
obstacles it encounters is the fear that people feel when hearing about ‘nuclear’ in any of its
application such as propulsion or energy production. This fear is born mainly after Chernobyl
disaster on April 26, 1986 in Ukraine, formerly in the USSR. The accident renewed worries about
the use of nuclear sources. The goal of the present appendx is to explain causes and show
consequences of the event, in order to clarify ideas about use of nuclear power in space application
as well.
7.B.2. REACTOR DESIGN
The reactor used in Chernobyl was a soviet-designed 925 MWe RBMK (see Figure 136).
RBMK reactors are pressurized water reactors [1]. The main purpose of the reactor was to produce
weapon grade Plutonium, not energy.
Figure 136 RMBK
In RMBK reactors individual fuel channels are used. Ordinary water is the coolant and
graphite the neutrons moderator. The combination of graphite moderator and water coolant is found
in no other power reactor. The RMBK differs substantially from most power reactor designs as it
was intended mainly for plutonium and secondarily for power production. The design features of
the reactor caused instability at low power. This was due primarily to control rod design and a
positive void coefficient [1].
253
7.B.2.1 Void Coefficient
In a water cooled reactor steam may accumulate and form pockets, called voids. If excess
voids are created, the operation of the reactor is disturbed both because steam does not cool as well
as water, and because liquid water is a neutron moderator and absorber, while steam is not. Void
coefficient is said positive if excess steam increases power generation, negative if it leads to a
decrease in power. When the void coefficient is positive, the power can rise very rapidly because
any power increase generates more steam, which in turn leads to a further increase in power. Such
increases are very difficult to control. Most of the world’s operating reactors have negative void
coefficients.
Another serious flaw of the Chernobyl reactor was the lack of a containment structure. The
Three Miles Island accident in the US damaged seriously the core reactor, but almost all fission
products were retained by the containment system [1].
7.B.3. CHERNOBYL EVENTS
On April 25, during a routine shut down, a non scheduled and ad hoc test to determine how
long turbines would spin and supply power following a loss of the main electrical power supply was
performed. Similar tests had already been carried out at Chernobyl and other plants, despite the fact
that these reactors were known to be very unstable at low power settings. This test had nothing to
do with reactor operation, purpose or maintenance, and was done at the instigation of a single
individual.
What follows is a reconstruction of the sequence of events. It is still incomplete, as many
details of the accident were, and still are, missing [1].
April 25: Prelude

01:06 The scheduled shutdown of the reactor started. Gradual lowering of the power level began
03:47 Lowering of reactor power halted at 1600 MW(thermal).
14:00 The Emergency Core Cooling System (ECCS) was isolated (part of the test procedure) to
prevent it from interrupting the test later. This means that all safety devices were
deliberately made inoperative, in order to test the spinning down time of the turbines.
14:00 The power was supposed to be lowered further; however, the controller of the electricity
grid in Kiev requested the reactor operator to keep supplying electricity to enable demand to
be met. Consequently, the reactor power level was maintained at 1600 MW(t) and the
experiment was delayed.
Without this delay, the test would have been conducted during the day shift (not at night).
23:10 Power reduction recommenced.
24:00 Shift change took place.
April 26: Preparation for the test

00:05 Power level had been decreased to 720 MW(t) and continued to be reduced.
It is now recognised that the safe operating level for a pre-accident configuration RBMK
was about 700 MW(t) because of the positive void coefficient mentioned.
00:28 Power level was now 500 MW(t).
Control was transferred from the local to the automatic regulating system. Either the
operator failed to give the `hold power at required level' signal or the regulating system
failed to respond to this signal. This led to an unexpected fall in power, which rapidly
dropped to 30 MW(t).
254
00:32 (approximate time). In response to the power drop, the operator retracted a number of
control rods in an attempt to restore the power level.
Station safety procedures required that approval of the chief engineer be obtained to operate
the reactor with fewer than the effective equivalent of 26 control rods. It is estimated that
there were less than this number remaining in the reactor at this time.
01:00 The reactor power had risen to 200 MW(t).
01:03 An additional pump was switched on in the left hand cooling circuit in order to increase the
water flow to the core (this was part of the turbine test).
01:07 An additional pump was switched on in the right hand cooling circuit (this was also part of
the test).
Operation of additional pumps removed heat from the core more quickly. This reduced the
water level in the steam separator.
01:15 Automatic trip systems to the steam separator were deactivated by the operator to permit
continuing operation of the reactor.
01:18 Operator increased feed water flow in an attempt to address the problems in the cooling
system.
01:19 Some manual control rods were withdrawn to increase power and raise the temperature and
pressure in the steam separator.
Operating policy required that a minimum effective equivalent of 15 manual control rods be
inserted in the reactor at all times. At this point it is likely that the number of manual rods
was reduced to less than this (probably eight). However, automatic control rods were in
place, thereby increasing the total number.
01:21:40 Feed water flow rate reduced to below normal by the operator to stabilise steam separator
water level, decreasing heat removal from the core.
01:22:10 Spontaneous generation of steam in the core began.
01:22:45 Indications received by the operator, although abnormal, gave the appearance that the
reactor was stable.
The test
01:23:04 Turbine feed valves closed to start the turbine coasting test. This was the beginning of the
actual spin down test.
01:23:10 Automatic control rods withdrawn from the core. An approximately 10 second withdrawal
was the normal response to compensate for a decrease in the reactivity following the
closing of the turbine feed valves.
Usually this decrease is caused by an increase in pressure in the cooling system and a
consequent decrease in the quantity of steam in the core. The expected decrease in steam
quantity did not occur due to reduced feedwater to the core.
01:23:21 Steam generation increased to a point where, owing to the reactor's positive void
coefficient, a further increase of steam generation would lead to a rapid increase in power.
01:23:35 Steam in the core begins to increase uncontrollably.
01:23:40 The emergency button (AZ-5) was pressed by the operator. Control rods started to enter
the core.
The insertion of the rods from the top concentrated all of the reactivity in the bottom of the
core.
01:23:44 Reactor power rose to a peak of about 100 times the design value.
01:23:45 Fuel pellets started to shatter, reacting with the cooling water to produce a pulse of high
pressure in the fuel channels.
01:23:49 Fuel channels ruptured.
01:24 Two explosions occurred. One was a steam explosion; the other resulted from the
expansion of fuel
255
The explosions lifted the pile cap, allowing the entry of air. The air reacted with the
graphite moderator blocks to form carbon monoxide. This flammable gas ignited and a
reactor fire resulted.
Thereafter, over nine days, some 8 of the 140 tonnes of fuel, which contained plutonium and
other highly radioactive materials (fission products), were ejected from the reactor along with a
portion of the graphite moderator, which was also radioactive. These materials were scattered
around the site. In addition, caesium and iodine vapours were released both by the explosion and
during the subsequent fire.
7.B.4. CONSEQUENCES
The UNSCEAR report to UN in 2000 [2] reported the consequences of this accident that can
be considered the most serious ever happened in nuclear industry. The radionuclides released from
the reactor that caused exposure were mainly iodine-131, caesium-134 and caesium-137.
Average doses were about 100 mSv for the 240,000 recovery operation workers, 30 mSv for
the 116,000 people evacuated and 10 mSv during the first decade after the accident to those who
kept on dwelling in contaminated areas. Maximum doses may be an order of magnitude higher.
Of the 600 workers present on April 26, 134 received high doses (0.7-13.4 Gy) and suffered
from radiation sickness, 28 of these died within the first three months and another 2 soon after.
During 1986-87, about 200,000 recovery workers received doses between 0.01 and 0.5 Gy.
This cohort (group) will be followed by UNSCEAR for further studies on radiation exposure
consequence.
The number of thyroid cancer in persons exposed (children), 1800, is considerably greater
than usual.
Apart from the increase in thyroid cancer mentioned, no other increase in cancer incidence
or mortality has been observed so far (up to the year 2000). The risk of leukaemia (leukaemia is the
first cancer to appear after radiation exposure), even among recovery workers, does not appear to be
elevated. However there were strong psychological reactions to the accident, due, as a matter of
fact, to fear of radiation.
There is a tendency to attribute the increase in the rates of cancer to Chernobyl, but it is
worth to note that even before the accident, the cancer rates were increasing in the affected areas,
and more generally this was an increasing mortality in most areas of the former USSR. This fact
must be considered when interpreting studies on the Chernobyl accident.
7.B.5. CONCLUSIONS
The Chernobyl accident was caused by three kinds of factors:
• A flawed design of the RMBK reactor itself: positive void coefficient, no containment facility
• Serious errors made by plant operators, who were minimally trained
• Planning the turbine test itself, and, of paramount significance the deliberate switching off of the
ECCS security system to perform an unnecessary test
REFERENCES
1. World Nuclear Association Web Site http://www.world-nuclear.org
256
2. UNSCEAR (2000). United Nations Scientific Committee on the Effects of Atomic Radiation.
Sources and Effects of Ionizing Radiation. UNSCEAR 2000 Report to the General Assembly,
with Scientific Annexes. United Nations, New York
i
Greene, S.R., Smith, B., Nesmith, B., Bhattachryya, S., Houts, M., Marshall, A.C., Mason, L., Poston, D., Weitzberg,
A., and Wright, S., „Summary of an Interagency NASA/DOE Review of Space Reactor Power System Cocnepts“
ORNL/SR/LTR-2003/001, April 2003.
ii
CTH is a large-scale structural deformation code originally developed by Sandia National Laboratories. While still
used by Sandia and other laboratories, it is no longer supported in favor of more advanced structural codes such as
PRONTO3D and ALEGRA.
257

Nuclear Propulsion

Uploaded by

Copyright:

Available Formats

Nuclear Propulsion

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nuclear Propulsion

Uploaded by

Copyright:

Available Formats

IAA COMMISSION III

SG 2 – NUCLEAR SPACE POWER AND

1. Physics of Nuclear Propulsion – An Introduction 11

Study Group members:

M. Marcel Pouliquen, SAFRAN/SNECMA, France

IAA Commission 3 members:

Dr. Hans Hoffman,

Dr. Horst Rauck,

Dr. Andre’ Van Gaver, ESA-ESTEC, The Netherlands

1.3.2. Potential Energy

and similarly for y and z, and where

Since the momentum p = m0 V (1 − V 2 c 2 ) , with V now a four-component vector, and

Total energy = m0c2 + ½ m0V2 + …

Total energy = mc2 + KE (1a)

Reactant(s) potential energy = α m c2 (2)

Gravitational energy (according to Newtonian theory, not General Relativity) is associated,

1.4.1. Specific Impulse

Isp = V = c f(μ, α) (6bis)

Where f(μ, α) ≡ 2 { [ (1 + 2/A ) + 1]}

P = PR ηtot = α m& 0 c2 ηtot

Equation (8) may be rewritten in terms of the ratio μ:

Reactor power PR is 1/ηtot of the thrust power P defined as

From equation (10) thrust power can be written

μ = [αc2 /ΔT – (1 – α) Cpm]/CpM (15)

Since αc2 >> (1 – α) CpM ΔT for any reasonable structural temperature,

μ ≈ αc2 / (CpM ΔT)>> 1 (15bis)

P = Isp F Thrust power

∆V = a tacc ∆V acquired after a time tacc

F=Ma Newton’s law. M is the spacecraft mass, assumed >> Mp

dacc (km) P (MW) tacc (days) ΔV (km/s) Mp (ton)

1.5. Nuclear Propulsion Strategies

1.5.1. Nuclear Thermal Rockets (NTR)

Figure 6 Conceptual scheme of a Rubbia-type nuclear rocket engine

1.5.2. Nuclear Electric Propulsion (NEP)

Efficiency notwithstanding, experience in solar-powered electric propulsion has already

1.5.3. A Comparison between Chemical and NTR/NEP Isp

Itot,s ≡ Isp toperation /(Mp + m)

where toperation is the mission “engine-on” total time.

Therefore Itot,s = Isp/(dm/dt), and since P = F Isp and F = Isp (dm/dt),

Itot,s = Isp3/P = Isp3 ηtot/PR

Table III Performance index, I, of some propulsion systems (fixed PR)

and since the momentum of photons is hν/c = Ep/c, the thrust is

Thus consumption of nuclear fuel must be included when estimating performance of

Aerospace America, 2004: see roundtable discussion on NP at: http://boss.streamos.com/

Einstein, A., (1916), “Uber die spezielle und allgemeine Relativitaetstheorie

FAS (Federation of Atomic Scientists), (2005), http://www.fas.org/nuke/space/c02early.htm

Flora, M. (2005), “Project Orion”, www.islandone.org/Propulsion/ProjectOrion.html

. quel che segue qui sotto in giallo si puo’ probabilmente cancellare

NASA (2005a), www.nasa.gov/vision/space/travelinginspace/25aug_plasticspaceships.html

NASA (2005b), www.radiationshielding.nasa.gov

A nuclear rocket’s configuration is similar to that of a chemical system, except that it

NERVA Derivative or Enabler

- 28 full-power tests with restarts

Reactor and Engine Systems Tests

*Operated 60 minutes at full power (1,100MW)

2.4. Particle-Bed Reactor

- Fuel tests to temperatures as high as 3000 K (NERVA only reached 2650 K)

Recently, work at Aerojet has produced a mixed Nuclear-Chemical propulsion concept

-Ingestion (not a significant hazard)