Nuclear Batteries and Radioisotopes

Lecture Notes in Energy 56
Mark Prelas
Matthew Boraas
Fernando De La Torre Aguilar
John-David Seelig
Modeste Tchakoua Tchouaso
Denis Wisniewski
Nuclear
Batteries and
Radioisotopes
Lecture Notes in Energy
Volume 56
Lecture Notes in Energy (LNE) is a series that reports on new developments in the
study of energy: from science and engineering to the analysis of energy policy. The
series’ scope includes but is not limited to, renewable and green energy, nuclear,
fossil fuels and carbon capture, energy systems, energy storage and harvesting,
batteries and fuel cells, power systems, energy efficiency, energy in buildings,
energy policy, as well as energy-related topics in economics, management and
transportation. Books published in LNE are original and timely and bridge between
advanced textbooks and the forefront of research. Readers of LNE include
postgraduate students and non-specialist researchers wishing to gain an accessible
introduction to a field of research as well as professionals and researchers with a
need for an up-to-date reference book on a well-defined topic. The series publishes
single- and multi-authored volumes as well as advanced textbooks.
More information about this series at http://www.springer.com/series/8874

Mark Prelas Matthew Boraas
•

John-David Seelig Modeste Tchakoua Tchouaso
•
Denis Wisniewski
Nuclear Batteries
and Radioisotopes
123
Mark Prelas John-David Seelig
Nuclear Science and Engineering Institute Nuclear Science and Engineering Institute
University of Missouri University of Missouri
Columbia, MO Columbia, MO
USA USA
Matthew Boraas Modeste Tchakoua Tchouaso

USA USA
Fernando De La Torre Aguilar Denis Wisniewski

USA USA
ISSN 2195-1284 ISSN 2195-1292 (electronic)

Lecture Notes in Energy
ISBN 978-3-319-41723-3 ISBN 978-3-319-41724-0 (eBook)
DOI 10.1007/978-3-319-41724-0
Library of Congress Control Number: 2016945768
© Springer International Publishing Switzerland 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
As a gesture to the fading concepts of
academic integrity and excellence, the
authors wish to dedicate this book to the
students, alums, staff, and faculty of the
Nuclear Science and Engineering Institute
(NSEI) at the University of Missouri. Its
history began in 1957 when the university
offered its first course in nuclear engineering
through the chemical engineering
department. The university began hiring
dedicated nuclear engineering faculty who
then designed and built the University of
Missouri Research Reactor. Unlike other
programs of the era where the reactor
remained in the nuclear engineering
department, the research reactor
management structure was moved to the
University of Missouri System where it was a
shared resource for the system’s four
campuses. In 1964 NSEI’s first Masters of
Science student graduated and the university
formally established a degree program in
nuclear engineering. It offered MS and PhDs
in Nuclear Engineering with a traditional
emphasis and a health physics emphasis. In
1980, it developed one of the first and most
prestigious nationally accredited programs in
medical physics through its offering of MS
and PhD degrees with medical physics
emphasis. NSEI’s approximately 600 alumni
have served in government, the military,
industry, hospitals, the health professions and
academia in the US and internationally with
great honor. Many NSEI alumni have walked
the halls of power serving as corporate
leaders, entrepreneurs, a Chairman of the
Nuclear Regulatory Commission, an
Assistant Secretary of Defense, Generals,
academic program chairs, deans, respected
faculty, medial physicists at prestigious
hospitals such as MD Anderson, Washington
University, and many more of the greatest
hospitals in the US. A significant fraction
of the PhDs awarded in the nuclear
engineering community nationally graduated
from NSEI and its predecessor (the nuclear
engineering program at MU). This book has
five authors from the last group of NSEI
graduate students.
Remember: NSEI, 1958–2012

Preface
Nuclear Batteries and Radioisotopes and its sister book, Nuclear-Pumped Lasers
(published by Springer in 2016), examine the direct conversion of nuclear energy
into other useful energy forms. What these two books demonstrate is that all types
of nuclear energy conversion have common features. These include the principles
of the interaction of ionizing radiation with matter (covered as a chapter in
Nuclear-Pumped Lasers), the matching up of the range of the ionizing radiation
with the scale length of the transducer, the concept of reaction rates, the concept of
power density, as well as many other shared principles. Another shared common-
ality between the two texts is that both were originated from graduate level courses.
Nuclear Batteries and Radioisotopes was developed from a graduate level course
of the same title taught by Prof. Prelas during the fall semester 2015. Five of the
students who participated in this course assisted in the writing of this text as
authors.
Nuclear Batteries and Radioisotopes was undertaken to provide researchers and
students in the field a resource which can be used as a tool to identify the critical
engineering issues in nuclear battery design. This book expands upon the funda-
mental aspects of radiation transport and the physics of transducers. Ionizing
radiation transports through matter with a characteristic scale length Called the
range. Transducers also have basic scale lengths over which they absorb power and
convert that power into another useful form. The matchup between the transducer
scale length and the range of the ionizing radiation is one of the dominant factors in
determining the system efficiency of the nuclear battery design. The complexities of
nuclear battery design are difficult to comprehend even for skilled practitioners. The
field seems to gain a new life and renewed interest every 20 years or so which so
happens to coincide with generational changes in the scientific community. The
driving force of this cycle of renewed interest is the enormity of amount of energy
that is stored in radioisotopes. Often times the lessons learned from the previous
cycles are lost in the enthusiasm of the current cycle. More so, it is the complex
nature of the designs that make the lessons from previous works hard to interpret.
vii
viii Preface
The goal of this book is to help simplify the factors which make the design of
nuclear batteries so complex. In Chap. 1, the physical parameters of the radioiso-
tope sources themselves are explored. Compounds can be formed which contain the
radioisotope atoms. Chemically, there is a compound which maximizes the atomic
number density of the radioisotope. This compound then has the highest possible
concentration of the radioisotope and thus has the maximum thermal power density
that is feasible for that radioisotope. By taking the inverse of the maximum thermal
power density, the minimum volume of the compound able to produce 1 W of
power can be found. This minimum volume provides the reader a feel for the scale
of a device fueled by a specific radioisotope.
Chapter 2 looks at the radioisotopes and where they come from or how they can
be made. There are two fundamental sources of radioisotopes: naturally occurring;
and man-made. In choosing a radioisotope for a nuclear battery design, the designer
will have questions. Are the isotopes readily available? How much do the isotopes
cost? These two fundamental questions are addressed.
Chapter 3 focuses on the ranges of ionizing radiation in matter as well as the
scale length of the transducers of interest. The scale length match up of the ionizing
radiation range to the scale of the transducer is an important issue. If the matchup is
poor, the efficiency of the nuclear battery will be low. The power deposition effi-
ciency in the transducer is dependent upon the scale length matchup between the
charged particles ranges and the transducer.
Chapter 4 covers the atomic dilution factor, which is the ratio of the number of
radioisotope atoms in the volume of the nuclear battery versus the number of
radioisotope atoms in the same volume of the compound which maximizes the
atomic density of radioisotope. The way the source is interfaced with the transducer
in the nuclear battery design will determine how large the dilution factor will be.
A surface source interface will have large dilution factors relative to volume
interfaces. However, even volume interfaces will have some level of dilution factor.
The dilution factor is a measure of how much the power density in a nuclear battery
will be diminished from the optimum possible value.
Transducer and system efficiency of the nuclear battery is covered in Chap. 5.
Specific examples of transducers are discussed and the reader is given a method-
ology of making such determinations for new transducer concepts beyond those
that are already known. In addition the effects of radiation damage on components
of a nuclear battery are examined. The goal of this chapter is to consider the
integrated nuclear battery system. A section on problems and issues in nuclear
battery literature is included. The purpose of this section is to provide the reader
with ability to evaluate reported systems in the literature and to develop the nec-
essary skills to use the tools developed in prior chapters to accurately determine the
validity of claims that are made.
In Chap. 6, a host of projected applications are discussed and the capability of
nuclear battery systems is examined to determine the viability of batteries to meet
the challenges presented by the specific applications. A brief discussion of the
transition from radioisotopes to fission reactors is discussed which examines the
benefits of the control of the power density which fission reactors offer.
Preface ix
Appendix A has valuable data on the range of ionizing radiation from viable
radioisotopes that may be used for applications in nuclear batteries. The range data
helps the reader understand the scale length of the ionizing radiation ranges.
Appendix B has formulas which represent the energy spectra of beta particles
emitted from key radioisotopes.
Appendix C looks at advanced theoretical concepts for the purpose of presenting
methodologies for the reader to be able to use the tools presented in the book to
tackle never-seen-before designs and to understand the potential of the design.
The concepts from the chapters of this text demonstrate the sophistication and
subtleties of nuclear battery design. There are numerous difficulties involved in
integrating these concepts into a nuclear battery system and at least in part is a
reason why this technology seems to be enthusiastically rediscovered every gen-
eration with high expectations solely based on the energy storage potential of
radioisotopes. It is the humble hope of the authors that this text will help the reader
to focus on the fundamental engineering limitations which underlie nuclear battery
design and will temper expectations of the technology. It is only through a full
understanding of limitations that real breakthroughs can be made.
Columbia, USA Mark Prelas

Matthew Boraas
John-David Seelig
Modeste Tchakoua Tchouaso
Denis Wisniewski
Contents
1 Introduction to Nuclear Batteries and Radioisotopes . . . . . . . . . . . . 1

1.1 Fundamental Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Nuclear Battery Design Considerations . . . . . . . . . . . . . . . . . . . 9
1.2.1 Surface Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.2.2 Volume Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.3 Products from Ionizing Radiation: Heat and Ion Pairs. . . . . . . . . 17
1.4 Geometrical Considerations in the Interface of an Isotropic
Radiation Source to a Transducer . . . . . . . . . . . . . . . . . . . . . . 28
1.5 Methodology for Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2 Radioisotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.1 Existing Radioisotope Supplies . . . . . . . . . . . . . . . . . . . . . . . . 39
2.1.1 Primordial Radioisotopes . . . . . . . . . . . . . . . . . . . . . . 40
2.1.2 Cosmogenic Radioisotopes . . . . . . . . . . . . . . . . . . . . . 41
2.1.3 Manmade Radioisotopes . . . . . . . . . . . . . . . . . . . . . . . 45
2.2 Radioisotope Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.2.1 Separation from Spent Fuel . . . . . . . . . . . . . . . . . . . . . 50
2.2.2 Separation from Natural Decay Chains . . . . . . . . . . . . . 57
2.2.3 Production by Neutron Capture in a Reactor . . . . . . . . . 63
2.2.4 Production by Accelerator. . . . . . . . . . . . . . . . . . . . . . 64
2.3 Cost of Radioisotopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.3.1 Cost of Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.3.2 Cost of Neutron Capture . . . . . . . . . . . . . . . . . . . . . . . 69
2.3.3 Cost of Accelerator . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.4 Other Factors Influencing Cost . . . . . . . . . . . . . . . . . . . . . . . . 74
2.4.1 Safety. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.4.2 Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.4.3 Liquidity of Capital (Cash) . . . . . . . . . . . . . . . . . . . . . 75
2.5 Isotopes Produced from the Manhattan Project . . . . . . . . . . . . . 75
xi
xii Contents
2.6 Mixed Oxide Fuel Fabrication Facility (MOX FFF) . . . . . . . . . . 75

2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3 Interactions of Ionizing Radiation with Matter
and Direct Energy Conversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.1 Ionizing Radiation Types and Ranges. . . . . . . . . . . . . . . . . . . . 81
3.1.1 Fission Fragments . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.1.2 Alpha Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.1.3 Beta Particles and Positrons . . . . . . . . . . . . . . . . . . . . 90
3.1.4 Shielding Considerations. . . . . . . . . . . . . . . . . . . . . . . 92
3.1.5 Rules of Thumb and Their Limitations . . . . . . . . . . . . . 95
3.1.6 The Limitations of Average Beta Energy . . . . . . . . . . . 97
3.1.7 What Types of Radiation Work Best with Nuclear
Batteries and Why . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.2 Types of Transducers Used in Nuclear Batteries . . . . . . . . . . . . 102
3.2.1 Ion Pair Based. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.2.2 Schottky Barriers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.2.3 Direct Charge Collection. . . . . . . . . . . . . . . . . . . . . . . 116
3.2.4 Indirect. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
3.2.5 Solid-State Emitter and PV . . . . . . . . . . . . . . . . . . . . . 149
3.2.6 Hybrid Solid-State Emitter . . . . . . . . . . . . . . . . . . . . . 155
3.2.7 Heat Based . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
4 Power Density Dilution Due to the Interface of the Isotope
with the Transducer . . . . . . . . . . . . . . . . . . . . . . ...... . . . . . . . 177
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . ...... . . . . . . . 178
4.2 Phase of the Radioisotope. . . . . . . . . . . . . . . ...... . . . . . . . 180
4.2.1 Radioisotope in Solid Phase . . . . . . . ...... . . . . . . . 180
4.2.2 Radioisotope in Liquid Phase . . . . . . ...... . . . . . . . 182
4.2.3 Radioisotope in Gaseous Phase . . . . . ...... . . . . . . . 183
4.2.4 Gaseous-like Radioisotopes . . . . . . . . ...... . . . . . . . 187
4.3 Phase of the Transducer . . . . . . . . . . . . . . . . ...... . . . . . . . 189
4.3.1 Solid Phase Transducer. . . . . . . . . . . ...... . . . . . . . 189
4.3.2 Liquid Phase Transducer. . . . . . . . . . ...... . . . . . . . 189
4.3.3 Gas Phase Transducer . . . . . . . . . . . ...... . . . . . . . 192
4.3.4 Plasma Phase Transducer . . . . . . . . . ...... . . . . . . . 192
4.4 Surface Interface . . . . . . . . . . . . . . . . . . . . . ...... . . . . . . . 192
4.4.1 Methods of Forming Surface Sources . ...... . . . . . . . 193
4.4.2 Electroplating, Painting and Baking . . ...... . . . . . . . 195
Contents xiii
4.4.3 Evaporation and Sputtering . . . . . . . . . . . . . . . . . . . . . 195

4.4.4 Implanting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
4.5 Charged Particle Escape Probability from Surface Sources . . . . . 196
4.6 Scale Length Matching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
4.6.1 Scale Length of Ionizing Radiation . . . . . . . . . . . . . . . 198
4.6.2 Scale Length of Transducer. . . . . . . . . . . . . . . . . . . . . 205
4.7 Geometrical Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
4.8 Power Density Dilution Factors for Surface Interfaces . . . . . . . . 215
4.9 Power Density Dilution Factors for Volume Interfaces . . . . . . . . 216
4.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
5 Efficiency Limitations for Various Nuclear
Battery Configurations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
5.1 Basics of Nuclear Battery Design. . . . . . . . . . . . . . . . . . . . . . . 221
5.1.1 Transducer Efficiencies . . . . . . . . . . . . . . . . . . . . . . . . 222
5.1.2 Direct Charge Nuclear Battery (DCNB) . . . . . . . . . . . . 230
5.2 Radiation Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
5.3 Health and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.3.1 Nuclear Regulatory Commission Rules
and Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
5.3.2 Po-210 Poisoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
5.3.3 NASA RTG Safety . . . . . . . . . . . . . . . . . . . . . . . . . . 245
5.4 System Efficiencies and Power Density. . . . . . . . . . . . . . . . . . . 252
5.4.1 Alphavoltaics Analysis . . . . . . . . . . . . . . . . . . . . . . . . 253
5.4.2 Betavoltaics Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 258
5.4.3 PIDEC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
5.5 Analysis of Problems in Nuclear Battery Literature . . . . . . . . . . 269
5.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
6 Potential Applications for Nuclear Batteries . . . . . . . . . . . . . . . . . . 285
6.1 Successful Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
6.1.1 Pacemakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6.1.2 Deep Space Probes . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6.1.3 Curiosity Rover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
6.1.4 Remote Power Applications . . . . . . . . . . . . . . . . . . . . 289
6.1.5 Other . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
6.2 Military Missions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
6.2.1 Light Weight Portable Nuclear Batteries . . . . . . . . . . . . 292
6.2.2 MEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
xiv Contents
6.3 MEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

6.3.1 Drones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
6.3.2 Nano-Power Systems . . . . . . . . . . . . . . . . . . . . . . . . . 299
6.3.3 Fission Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
Appendix A: Range Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Appendix B: Beta Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Appendix C: Theoretical Nuclear Battery Design Concepts . . . . . . . . . . 337
Appendix D: Ranges for Alpha Emitters . . . . . . . . . . . . . . . . . . . . . . . 343
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Chapter 1
Introduction to Nuclear Batteries
and Radioisotopes
Abstract This chapter provides the reader with background and fundamental
information on the subject of nuclear batteries. The approach used in this chapter is
to describe the characteristics of a nuclear battery relying on easy to understand
physical properties. For example, a commonly used descriptive parameter is the
maximum power density. However, a more intuitive parameter is to use the inverse
of maximum power density and define a quantity of minimum volume per Watt for
a radioisotope volume source or minimum surface area per Watt for a radioisotope
surface source. This approach gives the reader a feel for actual dimensional limi-
tations of the technology as well as other constraints.
Keywords Energy conversion Scaling Surface source Volume source

Interfaces
Nuclear batteries have long been thought of as potential long-lived small power
supplies for host of critical applications. The quest for a viable nuclear battery
began soon after the discovery of radiation in the early 1900s [1] and continues
today because of one factor: the potential for a long battery lifetime. The reasons
that a viable micro-battery has yet to materialize can be explained. This chapter
introduces the reader to some fundamental concepts which will provide the foun-
dation for understanding the difficulties in using radioisotopes with various types of
nuclear battery energy conversion schemes. There are many competing types of
nuclear batteries: thermoelectric, thermophotoelectric, direct charge collection,
thermionic, scintillation intermediate, alphavoltaics, and betavoltaics. These battery
types depend on ionizing radiation for heat production (e.g., thermoelectric, ther-
mophotoelectric, and thermionic), for the production of ions and excited states (e.g.,
alphavoltaic, betavoltaic, and scintillation intermediate) or the conversion of kinetic
energy into potential energy contained in electrostatic fields (e.g., direct charge
collection). For the past 40 years the dominant nuclear battery technology has been
the radioisotope thermoelectric generator, or RTG, which converts the decay heat of
radioisotopes into electricity through the Seebeck effect [2–4]. RTGs have been
deployed in numerous deep space missions [5] and their demonstrated success
© Springer International Publishing Switzerland 2016 1

M. Prelas et al., Nuclear Batteries and Radioisotopes,
Lecture Notes in Energy 56, DOI 10.1007/978-3-319-41724-0_1
2 1 Introduction to Nuclear Batteries and Radioisotopes
makes the RTG the benchmark against which other nuclear battery technologies are
measured. The most advanced RTG is the Multi-Mission Radioisotope
Thermoelectric Generator (MMRTG). The MMRTG has a mass of 45 kg, a length
of 0.66 m and a diameter of 0.63 m. The unit produces 2 kW thermal and 125 W
electric at an efficiency of 6.25 % with a power density of 2.8 W/kg.
To adapt the advantages of nuclear battery technology for use in the ever-smaller
devices which are in development, attempts have been made to both miniaturize
nuclear batteries and improve their total energy conversion efficiency. This has
produced a variety of miniature nuclear battery concepts. They use transducers that
are very different than the Seebeck effect.
All nuclear battery systems share many of the same design considerations, but
the additional goals of increased efficiency and smaller size introduce additional
caveats to the design process. The performance of any nuclear battery technology is
ultimately determined by the physics of radioisotope(s), radiation transport, and
energy conversion transducers. The specific energy density (J/kg) of radioisotopes
is intrinsically higher than chemical energy sources by many orders of magnitude
due to the energetics of nuclear decay. However, the appropriateness of a
radioisotope source for a given battery power application also depends on the
specific power density (W/kg). Within a certain radioactive decay type (alpha, beta,
fission, etc.), the energy density varies inversely with the half-life of the isotope; the
shorter the half-life, the higher the power density. This fundamental principle
causes the two properties desired of a nuclear battery, long shelf-life and high
power density, to be contradictory.
Another design consideration specific to nuclear batteries is that the scale lengths
of the system components need to be ‘well-matched’. Within the context of this
book, the range of a given particle in a specific material is referred to as the
transport scale length of the radiation (kRadTr); the relevant physical dimension of
the energy conversion volume in the transducer is referred to as the scale length of
the transducer (Ltrans). These two scale lengths, kRadTr and Ltrans, should be
approximately equal. This fundamental principle dominates the efficiency of a
nuclear battery. A ‘well-matched’ system has a higher maximum theoretical effi-
ciency, while a system which is not ‘well-matched’ has a lower maximum theo-
retical efficiency. Achieving ‘well-matched’ scale lengths is one of the primary
challenges encountered in nuclear battery design.
Variables that influence kRadTr include: the mass, charge, angular distribution
and energy distribution of the source particles; the atomic number, density, and
ionization potential of the target material; and the mechanisms through which the
particle interacts with the target. These collectively cause kRadTr to vary greatly
among radioisotopes even for the same target material. The factors which determine
Ltrans include the energy conversion mechanism of the battery, the mechanical and
electrical properties of the target material, and the effect of radiation damage on the
target. The final factor is of essential importance to nuclear batteries as described in
the literature.
1 Introduction to Nuclear Batteries and Radioisotopes 3
In contrast, the RTG does not have issues with scale-length matching; the sheer
size of the RTG ensures that all of the radioisotope energy is deposited within the
transducer and converted to heat.
The challenges in matching the range of the radiation to the scale length of
potential transducers is the first of many factors that needs to be evaluated in
nuclear battery design.
1.1 Fundamental Concepts
The interest in radioisotopes as a possible source of energy is founded on the

immense amount of energy stored in a radioactive isotope, which can be as large as
2 109 J/gm (Table 1.1). Even though the stored energy is enormous as compared
to chemical batteries (Table 1.2), the radioisotope releases this energy at a rate
based upon its half-life (t1/2) which is defined as the time period at which half of the
number of radioactive atoms present in an arbitrary sample decay. The process of
radioactive decay is statistical. The time dependent number of isotopes (N(t)) in a
sample can be modeled using a rate equation:
dNðtÞ
¼ kNðtÞ ð1:1Þ
dt
where k is a decay constant and t is time in seconds.

The solution to Eq. (1.1) is,
NðtÞ ¼ Nð0Þekt ð1:2Þ
where N(0) is the initial number of radioisotope nuclei in the sample.

Dividing Eq. (1.2) by N(0), the fraction of remaining isotopes in a sample can be
found. Half-life can be found by solving,
N ðtÞ
¼ ekt ð1:3Þ
Nð0Þ
When t is equal to the half-life, half the radioactive particles in the sample
remain. So, by evaluating Eq. (1.3) for the value of the half-life, it is found that,

N t1=2
¼ 0:5 ¼ ekt1=2 ð1:4Þ
Nð0Þ
Solving Eq. (1.4),

4
Table 1.1 Properties of some possible isotopes for nuclear batteries: Power is calculated for beta and alpha particles only
Nuclide Decay Half Decay bmax (MeV) Other emission k (s−1) Molar mass Atoms/gm Activity Power Energy Power W/Ci Energy
energy life (Units in MeV) (decay/s)/gm W/gm stored stored
(MeV) (year) (J/gm) (J/Ci)
H-3 0.019 12.43 Beta 0.01861 1.782 3.0160492 1.997 x 3.559 x 1014 0.323538 1.97 108 3.749 x 10-5 2.10 104
10-9 1023
Ar-39 0.565 269 Beta 0.565 8.169 x 39.948 1.508 x 1.232 x 1012 0.037108845 4.51 108 0.001114933 1.32 107
10-11 1022
Ar-42 0.6 32.9 Beta 0.6 6.679 x 41.963046 1.435 x 9.586 x 1012 0.306735936 4.44 108 0.001184 1.77 106
10-10 1022
Co-60 2.824 5.2713 Beta and c 0.318 strong c at 1.17 4.169 x 59.933819 1.005 x 4.189 x 1013 0.710417655 1.46 109 0.00062752 1.29 106
(99 %) and 1.33 10-9 1022
(0.12 %)
Kr-85 0.67 10.755 Beta 0.67 0.4 % yield of 2.043 x 84.91253 7.092 x 1.449 x 1013 0.517808194 2.45 108 0.001318029 6.27 105
99.6 %, 0.514 c 10-9 1021
0.15
0.4 %
Sr-90 0.546 28.77 Beta 0.546 Yr -90 7.638 x 89.9077 6.698 x 5.116 x 1012 0.148981453 9.77 108 0.005578613 7.30 106
daughter with 10-10 1021
b decay energy
of 2.281 and
half life of
2.67 days
Ru-106 0.039 1.0234 Beta 0.039 2.147 x 105.907329 5.686 x 1.221 x 1014 0.25396244 1.14 107 0.00007696 3.47 103
10-8 1021
Cd-113m 0.58 14.1 Beta 0.58 1.559 x 112.9044 5.334 x 8.313 x 1012 0.257142774 1.60 108 0.001144533 7.11 105
10-9 1021
Sb-125 0.767 2.73 Beta 0.302 (40 %), Lots of c in the 8.049 x 124.905254 4.821 x 3.881 x 1013 0.478708387 1.91 108 0.000456393 1.82 105
0.622 (14 %) 0.5 range, yields 10-9 1021
and of 5 −20 %
0.13 (18 %)
Cs-134 2.058 2.061 Beta 0.662 97 % yield c at 1.066 x 133.90672 4.497 x 4.795 x 1013 1.265751647 4.78 108 0.000976682 3.69 105
71 %, 0.6–0.8 10-8 1021
0.089
28 %
(continued)
1 Introduction to Nuclear Batteries and Radioisotopes
Table 1.1 (continued)
Cs-137 1.175 30.1 Beta 1.176 c at 0.6617 7.301 x 136.907104 4.399 x 3.211 x 1012 0.095403085 2.67 108 0.001099206 3.08 106
6.5 %, (93.5 %) 10−10 1021
0.514
93.5
Pm-146 1.542 5.52 ec (66 %). 0.795 33 % yield of 3.981 x 145.914696 4.127 x 1.643 x 1013 0.23685592 1.12 108 0.000533392 2.52 105
Beta (34 %) 0.747 c 10−9 1021
Pm-147 0.225 2.624 Beta 0.225 8.375 x 146.915139 4.099 x 3.433 x 1013 0.411934768 4.76 107 0.000444 5.30 104
10−9 1021
Sm-151 0.076 90 Beta 0.076 2.442 x 150.919932 3.990 x 9.743 x 1011 0.003949124 1.57 107 0.000149973 5.95 105
10−10 1021
Eu-152 1.822 13.54 ec (72.1 %), 1.818 lots of c in 0.1– 1.623 x 151.921745 3.964 x 6.433 x 1012 0.174034676 1.04 108 0.001000918 5.60 105
beta (27.9 %) 0.3 range 10−9 1021
Eu-154 1.969 8.592 Beta 1.845 (10 %), lots of c 2.556 x 153.922979 3.912 x 1.001 x 1013 0.247022865 3.98 108 0.000913579 1.47 106
(99.98 %), ec 0.571 (36.3 %) 10−9 1021
(0.02 %) and 0.249
(28.59 %)
Eu-155 0.253 4.67 Beta 0.147 c 0.086 (30 %) 4.706 x 154.922893 3.887 x 1.829 x 1013 0.16702465 5.08 107 0.00033786 1.03 105
(47.5 %), and 0.105 10−9 1021
0.166 (25 %), (21 %)
0.192 (8 %)
and
0.253 (17.6 %)
Gd-148 3.182 74.6 Alpha 3.182 a 2.946 x 147.918115 4.071 x 1.199 x 1012 0.610572936 2.01 109 0.01883744 6.19 107
10−10 1021
Tm-171 0.096 1.92 Beta 0.0964 (98 %) c 0.0667 1.1445 x 170.936429 3.523 x 4.032 x 1013 0.204439908 1.75 107 0.000187597 1.60 104
and (0.14 %) 10−8 1021
0.0297 (2 %)
Os-194 0.097 6 Beta 0.0143 Lots of c in 3.662 x 193.965182 3.104 x 1.137 x 1013 0.039759437 1.56 107 0.000129372 5.06 104
(0.12 %), 0.01–0.08 range 10−9 1021
0.0535 (76 %)
and
0.0966 (24 %)
5
(continued)
6

Tl-204 0.763 3.78 Beta (97.1 %), 0.763 5.813 x 203.973864 2.952 x 1.712 x 1013 0.678193708 1.13 108 0.001461989 2.44 105
ec (2.90 %) 10−9 1021
Pb-210 0.063 22.29 Beta (100 %), 0.0169 (84 %) 4 % c 0.046. 9.859 x 209.984189 2.867 x 2.827 x 1012 2.4035348 2.37 109 0.01052 3.10 107
alpha and 0.0635 10−10 1021
(1.9 10−6 %) (16 %)
Po-208 5.216 2.8979 Alpha 100 % 5.115 a 7.583 x 207.981246 2.895 x 2.196 x 1013 17.96948512 2.34 109 0.030279528 3.94 106
(99.9958 %), and very low 10−9 1021
ec (0.0042 %) level of c from
0.291 to 0.861
Po-210 5.305 0.379 Alpha 5.305 a 5.7983 x 209.982874 2.867 x 1.663 x 1014 141.1431705 2.36 109 0.0314056 5.25 105
(100 %), c (100 %), 0.803 10−9 1021
(0.0011 %) c (0.0011 %)
Ra-228 0.046 5.75 Beta 0.0128 (30 %), Very low 3.822 x 228.03107 2.640 x 1.009 x 1013 0.015405793 6.28 106 5.648 x 10-5 2.30 104
0.0257 (20 %), percentage of 10−9 1021
0.0392 (40 %), low energy c
0.0396 (10 %)
Ac-227 0.044 21.773 Beta (98.6 %), 0.02 (10 %), Main a 4.953 1.009 x 227.027752 2.653 x 2.653 x 1012 2.127061926 5.95 106 0.000427968 8.22 104
alpha (1.38 %) 0.0355 (35 %) (47.7 %) and 10−9 1021
and 0.0448 4.940 (39.6 %)
(54 %) with very low
percentage of
0.1–0.24 c
Th-228 5.52 1.9131 Alpha a mainly 5.340 1.1487 x 228.028741 2.641 x 3.034 x 1013 26.05395081 2.26 109 0.0326784 2.76 106
(27.2 %) and 10−8 1021
5.423 (72.2 %)
with about
0.25 % c at
0.216
(continued)
U-232 5.414 68.9 Alpha a mainly 5.263 3.189 x 232.037156 2.595 x 8.278 x 1011 0.700087964 2.18 109 0.031293454 9.73 107
(31.55 %) and 10−10 1021
5.32 (68.15 %)
with low
percentage of
0.1 to 0.3 c
Pu-236 5.867 2.857 Alpha a mainly 5.721 7.692 x 236.046058 2.551 x 1.962 x 1013 18.03219635 2.32 109 0.034000066 4.37 106
(100 %), fis (30.56 %) and 10−9 1021
(1.3E−7 %) 5.768 (69.26 %)
Pu-238 5.593 87.74 Alpha a mainly 5.456 2.505 x 238.04956 2.53 x 6.336 x 1011 0.5555868 2.19 109 0.03311056 1.28 108
(100 %), fis (28.98 %) and 10−10 1021
(1.85E−7 %) 5.499 (70.91 %)
Pu-241 0.021 14.35 Beta 0.02082 a mainly 4.853 1.531 x 241.056851 2.50 x 3.826 x 1012 2.855800841 8.13 106 0.000108753 7.86 104
(99.998 %), (12.2 %), 4.896 10−9 1021
alpha (83.2 %)
(0.00245)
Am-241 5.638 432.2 Alpha a mainly 5.442 5.084 x 241.056829 2.498 x 1.27 x 1011 0.108572645 2.18 109 0.031626327 6.36 108
(100 %), fis (13 %) and 10−11 1021
(4.3E−10 %) 5.485 (84.5 %)
with 0.05954
(35.9 %)
gammas
Cm-243 6.168 29.1 Alpha a mainly 5.742 7.552 x 243.061389 2.478 x 1.871 x 1012 1.647617418 2.37 109 0.03248839 4.68 107
(99.71 %), ec (11.5 %), 5.785 10−10 1021
(0.29 %), fis (72.9 %), 5.992
(5.3E−9 %) (5.7 %) and
6.058 (4.7 %)
with about 20 %
gammas in 0.2–
0.3 range
(continued)
7
8

Cm244 5.902 18.1 Alpha a mainly 5.762 1.214 x 244.062753 2.467 x 2.996 x 1012 2.777536879 2.26 109 0.034305524 2.79 107
(100 %), fis (23.6 %), 5.805 10−9 1021
(1.37E−4 %) (76.4 %) with
low percentage
of gammas
Bk248 5.793 9 Alpha 5.793 a 2.442 x 248.073086 2.428 x 5.927 x 1012 5.49387705 2.18 109 0.03429456 1.36 107
10−9 1021
Cf250 6.128 13.07 Alpha a at 6.0304 1.681 x 250.076406 2.408 x 4.049 x 1012 3.890798482 2.29 109 0.035528482 2.09 107
(99.923 %), fis (84.6 %) and 10−9 1021
(0.0775 %) 5.989 (15.1 %)
and gamma at
0.04285
(0.014 %)
Cf252 6.217 2.645 Alpha Fission and a 8.308 x 252.081626 2.389 x 1.985 x 1013 18.78859265 2.23 109 0.035028006 4.16 106
(96.908 %),fis mainly at 10−9 1021
(3.092 %) 6.0758 (15.7 %)
and 6.118
(84.2 %) and
gamma between
0.043 and 0.155
(0.015 %)
Es-252 6.739 1.292 Alpha a mainly at 1.701 x 252.082979 2.389 x 4.063 x 1013 30.86044068 1.91 109 0.028101681 1.74 106
(76.4 %), ec 6.5762 (13.6 %) 10−8 1021
(24.2 %), beta and 6.632
(0.01 %) (80.2 %) and
gamma between
0.043 and 0.924
(25 %)
1.1 Fundamental Concepts 9
Table 1.2 Energy per gram Storage medium Energy per gram (J/gm)
(J/gm) stored in typical
chemical batteries and in the Alkali metal 422.6
Pu-238 radioisotope (used in Carbon-Zinc 130.7
the Radioisotope Thermoionic Ni-Cad 117.8
Generator) NiMH 288
Li Ion 460
Pu-238 2.19 109
kt ¼ ln ð0:5Þ ð1:5Þ
The rate constant can be found from Eq. (1.5).
ln ð0:5Þ 0:693
k¼ ¼ ð1:6Þ
t1=2 t1=2
Equations (1.1) and (1.6) are important to nuclear science. The product of k and
N(t) is called the activity (A(t)). The activity represents the number of nuclei that
decay in a sample over 1 s.
AðtÞ ¼ kNðtÞ ð1:7Þ
1.2 Nuclear Battery Design Considerations
Nuclear battery research has gone through spurts which seem to ebb and flow from
one generation to the next. For example, in the 1950s the concept of a betavoltaic
cell using silicon and germanium semiconductors was developed by RCA [6, 7].
The RCA cell used Sr-90, had an efficiency of 0.2 %, and also had a short lifetime
caused by radiation damage [8]. The semiconductor based betavoltaic cell concept
was resurrected in the 1960s [9–11], the 1970s [12], the 1990s [13], and the 2000s
[14]. Even though there were some optimistic claims made in quite a few of these
reports, the problems of low efficiency and radiation damage that were part of the
original conclusions in the 1950s were eventually recognized as persistent problems
in all of the studies. Unfortunately, resurrections of problematic concepts are an all
too common trend experienced in many different fields of research.
There are pitfalls in the physics of nuclear batteries which lead to limitations in
the underlying technologies. Understanding the limitations begin by asking
appropriate questions:
• What isotope should be chosen and why?
• How should the isotope be embedded and why?
• How should the embedded isotope be interfaced to the transducer (e.g., a surface
or volume interface)?
• What type of ionizing radiation does the isotope emit?

• What is the energy (or energy spectra) of the ionizing radiation?
• What is the range of the ionizing radiation emitted from the isotope in the
material that the isotope is embedded in?
• What is the range of the ionizing radiation emitted from the isotope in the
materials that make up the cell, especially the transducer?
• What is the scale length of the transducer?
• Is there a good match between the range of the ionizing radiation and the scale
length of the transducer?
• What fraction of the isotropic ionizing radiation intersects with the transducer?
• What type of transducer is used?
• What is the potential power density of the battery?
• What are the effects of radiation damage on the transducer?
Each of these questions represents a specific parameter that affects the design of
a nuclear battery. There are complex dependencies and interdependencies in these
parameters that make the design of a nuclear battery difficult. Often times when one
parameter is varied to solve one problem, it tends to create an equal or larger
problem elsewhere. This perhaps is the best explanation as to why nuclear battery
concepts keep being recycled from one generation to the next. In this book, the
complex nature of the interrelationships among the parameters which govern nu-
clear battery design will be discussed, with the goal of providing the reader the
background to better sort through the design issues.
One of the simplest limitations on the battery is in the capability of the isotope to
produce power. In the following example, a nuclear battery that uses a tritium
radioisotope interfaced to a transducer is considered. The concept of a surface
interface and a volume interface are introduced and will be examined. The design
process in building a nuclear battery can be compared to peeling an onion layer by
layer. The first layer of this complex process is in understanding the limitations of
the chosen radioisotope.
Radioisotope decay rates (as modeled by Eqs. 1.1–1.7) can be used to calculate
the power output from a specific quantity of the isotope. In this first example,
tritium is the chosen radioisotope. Tritium has a half-life of 12.33 years with a
maximum beta (b) energy of 19 keV (otherwise known as Q value which is the
energy released in the reaction). It decays by the reaction,
T ! 31 He2 þ b þ m þ 19 keV ð1:8Þ
where the general form for a nucleus is represented by AN Xz (X is the nucleus, A is

the atomic mass, N is the number of neutrons, and Z is the number of protons in the
nucleus, governed by the relationship A = N + Z), m is an antineutrino and the Q
value is 19 keV.
The antineutrino and the beta particle share the 19 keV energy and obey con-
servation of momentum and conservation of energy principles. A rule of thumb that
has been used for estimating the average beta energy is to multiply the Q value by
1.2 Nuclear Battery Design Considerations 11
1/3. There are errors involved in using this rule of thumb, as discussed in Ref. [1],
due to the fact that isotopes emit beta particles in a spectrum of energies [15]. For
the purposes of the following calculation the error of using this rule of thumb,
which can be as high as 10 % for some isotopes [15], are not considered. Thus, the
estimated average energy ðE Þ carried by one beta particle is,
¼ 19 keV ¼ 6:33 keV

E ð1:9Þ
3
A Curie (represented by the symbol Ci) is a unit of radioactivity which repre-

sents the amount of radioisotope that emits radiation at a rate of 3.7 1010 decays
per second. It is a commonly used as a measure for radioactivity. As a first step, the
amount of power emitted by one Ci of tritium will be calculated. Also, the energy
stored in one Ci of tritium will be estimated. A calculation is made for the power
emitted by the isotope at the beginning of its insertion into the experiment or device
(P(0) W/Ci). This value is found by multiplying the average energy of a beta
particle emitted in the reaction by the total number of decays per second for one Ci
of the isotope plus a conversion factor (to convert keV into Joules),
W keV Decay eV J
Pð0Þ ¼ 6:33 3:7 1010 1000 1:6 1019
Ci Decay s - Ci keV eV
ð1:10Þ
5 W
¼ 3:75 10
Ci
The power will decay with time. This time dependent power decay is needed to
calculate the energy stored in one Ci of tritium. The first step is to find the decay
constant (in inverse seconds) using Eq. (1.6),
0:693
k¼ ¼ 1:782 109 s1 ð1:11Þ
d 24 hr 3600 s
12:33yr 365 yr d hr
The time dependent power decay is directly proportional to the amount of iso-
tope present and it follows Eq. (1.3).
W W W 9 1
PðtÞ ¼ Pð0Þekt ¼ 3:75 105 e1:78210 s t ð1:12Þ
Ci Ci Ci
After about five half-lives (tf), the power decays to about 3 % of the initial
power. For tritium, five half-lives is,
d hr s
tf ¼ 5 12:33 yr 365 24 3600 ¼ 1:94 109 s 61 yr ð1:13Þ
yr d hr
The energy stored in one Ci of the isotope (Etotal/Ci) is estimated (to within 3 %)
by integrating Eq. (1.12) from 0 to tf,
Ztf
W 1:782109 s1 t J
Etotal=Ci ¼ 3:75 105 e dt ¼ 2:04 104 ð1:14Þ
Ci Ci
0
Using a similar procedure to the one used in Eqs. (1.12–1.14), the power emitted
from a gram of any isotope can be calculated. The power emitted from a gram of an
isotope at the beginning of the insertion of the isotope into the experiment or device
(P(0) W/gm), is the average energy emitted per decay times the decay constant
times the number of atoms in a gram of isotope. The number of atoms is found by
dividing the atoms per mole by the atomic mass (A in grams per mole where for
tritium A = 3) of the material. So,
W keV 6:02 1023 atoms

Pð0Þ ¼ 6:33 1:782 109 s1 gm
mol
gm Decay 3 mol
eV J ð1:15Þ
1000 1:6 1019
keV eV
W
¼ 0:363
gm
and,
W W W 1:782109 s1 t
PðtÞ ¼ Pð0Þekt ¼ 0:363 e ð1:16Þ
gm gm gm
The energy stored in one gm of the isotope (Etotal/gm) is estimated (to within
3 %) by integrating Eq. (1.16) from 0 to tf,
Ztf
W 1:782109 s1 t J
Etotal=gm ¼ 0:363 e dt ¼ 1:97 108 ð1:17Þ
gm gm
0
Similar calculations have been made for the list of viable isotopes. Table 1.1 is a
compilation of these parameters for viable isotopes that may be useful for nuclear
battery applications. As can be seen, radioisotopes store an enormous amount of
energy. This fact explains why there has been such a long-term (dating back
100 years) interest in nuclear batteries. One of the complicating factors is that the
isotope emits particles at a relatively slow rate. This fact accounts for one of the
limitations of nuclear batteries—low power density.
Tritium was chosen for the above example because it is mentioned frequently as
a desirable isotope for use in nuclear batteries due to the low range of the tritium
betas in various materials and its correspondingly low rate of radiation damage to
the material. In considering the properties of tritium, an interesting story emerges.
Due to the reaction’s low Q value of 19 keV, the power emitted from one Ci of
tritium is about 37 µW. This is one of the lowest values of power output per Ci in
the list of viable isotopes from Table 1.1. Table 1.1 also shows that 1 g of tritium
gives off 0.3625 W. In this example developed here it is both necessary and
instructive to discuss the simplest and most atomically dense compound that tritium
can form—tritium oxide (T2O) which has a density of 1.215 g/cm3.
The density of the tritium in T2O is 0.331 g/cm3. This is calculated by dividing
the atomic mass of two tritium atoms (six atomic mass units, also known as six
AMU) and dividing by the atomic mass of T2O (22 AMU) to determine the fraction
of the mass of tritium atoms in a T2O molecule, which is 0.2727. The power density
in T2O is found by multiplying the density of tritium in T2O (0.331 g-tritium/cm3)
times the Watts per gram of tritium (0.323538 W/g-tritium) which results in a value
of 0.107 W/cm3 of T2O.
The next step requires that the transport properties of betas emitted by tritium be
considered. It is very important that thickness of the liquid T2O layer on the surface
of the transducer does not exceed the range of the beta particle in the liquid T2O.
Otherwise the bulk of the beta energy will be deposited in the T2O layer. It is well
known that low energy beta particles (such as those emitted by tritium) have a short
range. Tritium is commonly mentioned as a desirable isotope for nuclear batteries
because of its low Q value. The reasoning behind this is that radiation damage in
semiconductors is a function of the energy of the ionizing radiation and tritium
emits a very low energy beta. Unfortunately, this also means that the power density
will be low because it too is a function of the energy of the ionizing radiation. The
range of the tritium beta will also be small. The range of beta particles with an
average energy of 5.7 keV was calculated using MCNP6 [16] and is about
0.61875 µm in liquid T2O (see Appendix A). As discussed in one reference [15],
there is an error that is introduced by using the average beta energy rather than
using the full beta spectrum from an emitter. The use of average beta energy in
transport calculations slightly underestimates the range of the beta. However, the
error does not significantly impact this example problem because the transducer is
assumed to be ideal. Thus, the average beta energy can be used for this calculation
with minimal error for the properties that are derived here. Below, the two fun-
damental ways that a radioisotope can be mated to a transducer, for both surface
and volume interfaces, are discussed. The concepts of the theoretical absolute
minimum nuclear Battery surface Area per Watt (BAWmin) and of the theoretical
absolute minimum nuclear Battery Volume per Watt (BVWmin) for various ra-
dioisotopes are introduced at this point. Both the BAWmin and the BVWmin are the
most wildly optimistic values for these quantities that are possible. The reason that
these are wildly optimistic begins with ion or electron transport. As ions or elec-
trons move through matter, they lose energy continuously. That means that as they
transport through the material in which the isotope is embedded, the ions or elec-
trons will slow down and deposit some energy in this material. For a surface
interface this is an energy loss. If the trajectory of the ion or electron moves through
any material other than the transducer, this is an energy loss. The wildly optimistic
assumption made in finding BAWmin and BVWmin is that there is no ion or
electron transport energy loss.
BAWmin and BVWmin provide the reader with an image of what the device scale
might be. This is important because device scale is often times not factored into the
discussions of nuclear battery concepts. As will be shown, this concept of scale is a
critical piece in understanding the limitations of the technology. The properties of
BAWmin and BVWmin are related to the surface power density and the power
density. The inverse of BAWmin is the theoretical maximum surface power density
and the inverse of BVWmin is the theoretical maximum volume power density. In
calculating these parameters it is important that the reader completely understand
the underlying assumptions:
• First, the compound containing the radioisotope has the maximum atomic
density possible for the isotope;
• Second, there are no energy losses from the ion or electron as it moves out of the
compound from which the radioisotope is embedded into the transducer.
• Third, the transport properties of the ionizing radiation into the transducer are
perfect;
• Fourth, the efficiency of the transducer is 100 %;
• Fifth, for a volume source the transducer and radioisotope interface is ideal and
seamless.
It is also important that the reader understand that real systems are going to be
much larger than the ideal systems that these assumptions lead to.
1.2.1 Surface Interface
The most common type of interface between an isotope and a transducer is a surface
interface where the isotope is coated on the surface of the transducer. This is typical
for betavoltaic and alphavoltaic nuclear battery designs where a p-n junction in a
semiconductor is used as the transducer. The limitations of the isotope in a surface
interface configuration can be delineated with some basic calculations. Making the
following assumptions, BAWmin (which is a scaling factor for a surface coupled
nuclear battery) can be found:
1. All of the beta particles emitted intersect the transducer (a Beta Particle Energy
Collection Efficiency—BPECE—of 100 %).
2. The transducer is 100 % efficient (an ideal transducer).
If a 0.619 µm film of liquid T2O covers the surface of an ideal transducer, the
BAWmin for power production from a liquid T2O coated nuclear battery is
approximately 150,740 cm2/W. The fact is that even though 150,740 cm2 is large,
there is no such thing as a 100 % efficient BPEC and there is no such thing as an
ideal transducer. Once the assumptions of ideal beta transport and an ideal trans-
ducer are discarded and real values are used, this will drastically increase the
surface area per Watt of a surface interfaced nuclear battery. One can call the
BAWmin a Perpetual Motion Limit (PML) for a surface interface where tritiated
water coats the surface of a transducer.
The PML is only part of the story. There are other important factors that erode
the efficiency of beta particle transport (or for that matter any other type of radia-
tion), such as the fact that radiation is emitted isotropically. Thus, half of particles
(or photons) have trajectories in the wrong direction. The other half must travel
along different pathways (dependent upon the angle of emission) and will lose
energy as they travel through the T2O and experience self-absorption before
reaching the transducer. A second issue is that nature always seems to create a
mismatch between the range of a beta particle and the scale length of the transducer
(this is true for other ionizing radiation types as well). This mismatch will reduce
efficiency of the nuclear battery significantly. Finally, no transducer is 100 %
efficient. The type of transducer governs efficiency and there are many types. This is
why a nuclear battery that uses a surface interface will have a surface area sig-
nificantly higher than the BAWmin limit. Any report of a battery that has a surface
area below the BAWmin limit is thus a physical impossibility and deserves to be
called a perpetual motion device. The arguments made thus far do not even take
into account radiation damage to the transducer and other critical components.
Radiation damage in most cases will severely decrease the device lifetime as well as
the transducer’s performance over time.
A BAWmin can be calculated for any isotope that is part of any compound or is
embedded in any type of material coated on any type of transducer. It should be
noted that the inverse of the BAWmin is the maximum value of Watts produced per
square centimeter or the maximum surface energy density (W/cm2).
Several examples from the literature demonstrate the degree of deviation that
tritium nuclear batteries have from BAWmin. These very large deviations are due to
the inefficiency of beta particle transport from the source layer (where tritium is
present) to the transducer and the inefficiency of the transducer. One such study
used a tritiated amorphous silicon nip drift junction (in order to maximize beta
particle transport to the transducer) with an ultrathin (5 nm) metal contact layer
[17]. The device produced an output power per square centimeter of 259 nW/cm2.
The BAW for this experimental cell was 3,861,000 cm2/W which is 25.61 times
larger than BAWmin (150,749 cm2/W). A second study used tritium-titanium and
tritium gas sources with a Al0.35Ga0.65As transducer [18]. The reported power was
0.024 lW/cm2 for tritium-titanium and 0.55 lW/cm2 for tritium gas. These values
lead to a BAW of 41,670,000 cm2/W for tritium-titanium and 1,818,000 cm2/W for
tritium gas. Both of these values are far beyond the BAWmin limit of 150,749 cm2/
W. In the case of tritium-titanium it is 276.4 times the limit and in the case of
tritium gas it is 12.06 times the limit. The lesson is that in the real world, a BAW
that is hundreds or even thousands of times greater than the BAWmin limit is typical
for any isotope and transducer combination. In essence, the ratio of BAW/BAWmin
or BVW/BVWmin is a measure of the device efficiency.
From a practical point of view, having an ideal cell with a BAW that approaches
BAWmin (150,749 cm2/W) still leads to a very large device. Even the most opti-
mistic schemes involving the stacking of thin cells will lead to large-scale devices
which are not suitable for projected Micro-Electro-Mechanical Systems (MEMS)

power supplies (where the desired power output is a milliWatt or better).
1.2.2 Volume Interface
A volume interface is the most efficient type of interface. This can be demonstrated
by assuming an ideal system in which T2O is interfaced with an ideal transducer
dissolved in the tritium oxide. Realizing that the power produced by the T2O is
0.107 W/cm3, the absolute minimum limit for the nuclear Battery Volume per Watt
(BVWmin) is the inverse of the power generated per cubic centimeter of T2O
(1/0.107). Thus, the BVWmin for an ideal system is 9.33 cm3 W−1. One can call
BVWmin a PML for a battery based on the volume interface between tritiated water
and a transducer. Like the case for the surface source battery, there are many
inefficiencies which will force the battery volume to be well above the BVWmin.
These inefficiencies come from the less than perfect beta transport properties, the
less than ideal transducer properties and a likely poor match between the scale of
the transducer and the range of the particle.
Volume interfaces are able to produce the smallest possible nuclear battery
package and are able to produce the highest possible energy density for any
combination of isotope and transducer. The BVWmin can be calculated for all
combinations of isotopes and transducer. The BVM is a tool, used to evaluate what
is feasible and what is not feasible. It provides an anchor to the real world. For
example, if a report claims that a nuclear battery is able to produce 1 W of power in
a smaller volume than the BVWmin limit, then the battery is a perpetual motion
device. It also can be useful for evaluating reports on cells that produce a fraction of
a Watt. Suppose that a paper reports that a 1-lW battery is made with T2O
embedded in some transducer with a volume of 0.000002 cm3. Does this system
obey the BVMmin limit? Divide 0.000002 cc by 0.000001 W and compare this
value to the BVMmin limit. As can be seen in Eq. (1.18), this cell violates the
BVMmin limit.
0:000002
¼ 2\BVMmin ¼ 9:33 ð1:18Þ
0:000001
The above example demonstrates several important lessons:

• First, that matching the range of a particle in the material in which it is
embedded is very important.
• Second, the interface between the material in which the radioisotope is
embedded and the transducer is critical.
• Third, the efficiency of the transducer, which converts the energy from the
ionizing radiation into something useful, is a very significant consideration.
In evaluating a nuclear battery concept, the first thing that a reader should do is
to identify whether the source is coupled to the transducer through the surface of the
transducer or within the volume of the transducer. This one simple step will tell the
reader much about the nuclear battery.
• First, if surface interfaced, the battery will have a scale length that is consid-
erably larger than a volume interfaced battery.
• Second, a surface interface will have a lower transport efficiency from the source
to the transducer than a volume interfaced battery.
• Third, a surface interfaced battery will tend to have a lower overall efficiency
than a volume interfaced battery.
Often times researchers will claim to have made a major breakthrough in nuclear
battery research without consideration of the fundamental lessons described above.
It is important that the reader use basic tools such as these to scrutinize claims. One
goal of this text is to provide the reader with the background to ask the right
questions and the tools to find answers.
Pertinent observations can be made about the suitability of a particular
radioisotope for use in a nuclear battery based on information contained in
Table 1.3.
• Observation 1. The longer the half-life, the larger BAWmin and BVWmin will
trend.
• Observation 2. The smaller the energy of the ionizing radiation, the larger
BAWmin and BVWmin will trend.
• Observation 3. Alpha emitters will have a much smaller BAWmin and BVWmin
than beta emitters.
• Observation 4. The higher the atomic density of the compound that the isotope
is embedded in, the lower BAWmin and BVWmin will trend.
1.3 Products from Ionizing Radiation: Heat and Ion Pairs
Ionizing radiation interacts with matter by producing ionization and heat. The
fraction of energy deposited in matter by ionizing radiation that goes into ion pair
production and the fraction that produces heat have been measured in gases and
solids.
The slowing down of ions and electrons in matter depends upon the density of
electrons in the stopping material. Through Coulombic interactions, high energy
ions transfer their energy to an electron bound by an atom in the target material.
This electron is then kicked out of its orbit through the energy transfer. It has
sufficient energy to create additional ionizations along its path through the target
material via the formation of secondary electrons. The secondary electrons are born
with sufficient energy to create tertiary electrons, and tertiary electrons have
Table 1.3 The theoretical absolute minimum nuclear battery surface area per Watt (BAWmin) and the theoretical absolute minimum nuclear battery volume
18
per Watt (BVWmin) for various radioisotopes are shown

Nuclide Decay Half Decay bmax (MeV) Power Compound Density Range (lm) BAWmin BVWmin
energy life W/gm gm/cm3 (cm2/W) (cm3/W)
(MeV) (year)
H-3 0.01861 12.43 Beta 0.01861 0.323538 T2O 1.215 0.61875 150,749 9.328
Ar-39 0.565 269 Beta 0.565 0.03710885 Liquid Ar 1.4 85.5 2,251 19.25
Ar-42 0.6 32.9 Beta 0.6 0.30673594 Liquid Ar 1.4 107.2 217.4 2.33
Co-60 2.824 5.271 Beta and c 0.318 0.71041766 Metal 8.9 3.12 506.4 0.158
Kr-85 0.67 10.76 Beta 0.67 99.6 %, 0.15 0.51780819 Liquid Kr 2.413 49.45 161.8 0.80
0.4 %
Sr-90 0.546 28.77 Beta 0.546 0.14898145 Metal 2.64 27 & 234 108.7 2.543
Ru-106 0.039 1.023 Beta 0.039 0.25396244 Metal 1.53 0.4715 54,590 2.574
Cd-113m 0.58 14.1 Beta 0.58 0.25714277 Metal 8.69 6.8 658.1 0.4475
Sb-125 0.767 2.73 Beta 0.302 (40 %), 0.622 0.47870839 Metal 6.684 3.333 & 282.8 0.3125
(14 %) and 0.13 11.05
(18 %)
Cs-134 2.058 2.061 Beta 0.662 71 %, 0.089 1.26575165 Metal 1.93 33.15 & 123.5 0.4093
28 % 1.615
Cs-137 1.175 30.1 Beta 1.176 6.5 %, 0.514 0.09540309 Metal 1.93 88 & 26.35 617.2. 5.431
93.5
Pm-146 1.542 5.52 ec (66 %). Beta 0.795 0.23685592 Metal 7.26 9.8 593.4 0.5815
(34 %)
Pm-147 0.225 2.624 Beta 0.225 0.41193477 Metal 7.26 1.2 2,787 0.3344
Sm-151 0.076 90 Beta 0.076 0.00394912 Metal 7.54 0.216 1,555,000 33.58
Eu-152 1.822 13.54 ec (72.1 %), 0.696 (13.6 %), 1.457 0.17403468 Metal 5.259 12.65 & 255.7 1.093
beta (27.9 %) (8.4 %), 385 (2.5 %), 42.75
0.176 (1.8 %)
(continued)
(MeV) (year)
Eu-154 1.969 8.592 Beta (99.98 %), 1.845 (10 %), 0.571 0.24702287 Metal 5.259 60.3 & 1.92 127.7 0.7698
ec (0.02 %) (36.3 %) and 0.249 & 8.8
(28.59 %)
Eu-155 0.253 4.67 Beta 0.147 (47.5 %), 0.166 0.16702465 Metal 5.259 0.672 & 5,932 1.139
(25 %), 0.192 (8 %) 1.275 & 1.92
and 0.253 (17.6 %) & 0.99&
0.7905
Gd-148 3.182 74.6 Alpha 0.61057294 Metal 7.90 8.44 245.6 0.2073
Tm-171 0.096 1.92 Beta 0.0964 (98 %) and 0.20443991 Metal 9.321 0.27 & 0.045 19,440 0.5248
1.3 Products from Ionizing Radiation …
0.0297 (2 %)
Os-194 0.097 6 Beta 0.0143 (0.12 %), 0.03975944 Metal 22.57 0.065 & 1,675,000 12.06
0.0535 (76 %) and 0.072
0.0966 (24 %)
Tl-204 0.763 3.78 Beta (97.1 %), 0.763 0.67819371 Metal 11.85 5.4 230.4 0.1244
ec (2.90 %)
Pb-210 0.063 22.29 Beta (100 %), 0.0169 (84 %) and 2.4035348 Metal 11.342 0.02 & 0.117 113.3 0.212
alpha 0.0635 (16 %) &
(1.9 10−6 %) 18.7
Po-208 5.216 2.898 Alpha 17.9694851 Metal 9.32 18.2 3.2 0.005971
(99.9958 %), ec
(0.0042 %)
Po-210 5.305 0.379 Alpha (100 %), 141.143171 Metal 9.32 18.7 0.41 0.0007602
c (0.0011 %)
(continued)
19
20
(MeV) (year)
Ra-228 0.046 5.75 Beta 0.0128 (30 %), 0.0257 0.01540579 Metal 5.5 0.12 & 0.035 916,700 11.00
(20 %), 0.0392 &0.0925
(40 %), 0.0396 (10 %)
Ac-227 0.044 21.77 Beta (98.6 %), 0.02 (10 %), 0.0355 0.031 Metal 10.07 0.07 & 0.05 442.9 0.0031
alpha (1.38 %) (35 %) and 0.0448 & 0.025
(54 %)
Th-228 5.52 1.913 Alpha 26.0539508 Metal 11.72 15.4 2.12 0.003275
U-232 5.414 68.9 Alpha 0.70008796 Metal 18.95 9.55 78.9 0.07538
Pu-236 5.867 2.857 Alpha (100 %), 18.0321964 Metal 19.86 8.33 3.35 0.002792
fis (1.3E−7 %)
Pu-238 5.593 87.74 Alpha (100 %), 0.5555868 Metal 19.86 7.71 117.6 0.09064
fis (1.85E−7 %)
Pu-241 0.021 14.35 Beta 0.02082 0.0112 Metal 19.86 0.015 376 0.000564
(99.998 %),
alpha (0.00245)
Am-241 5.638 432.2 Alpha (100 %), 0.10857265 Metal 13.69 11.2 600.7 0.6728
fis (4.3E−10 %)
Cm-243 6.168 29.1 Alpha 1.64761742 Metal 13.52 12.8 35.1 0.04489
(99.71 %), ec
(0.29 %), fis
(5.3E−9 %)
Cm244 5.902 18.1 Alpha (100 %), 2.77753688 Metal 13.52 12 22.2 0.02663
fis (1.37E−4 %)
(continued)
(MeV) (year)
Bk248 5.793 9 Alpha 5.49387705 Metal 14.78 14.5 8.5 0.01232
Cf250 6.128 13.07 Alpha 3.89079848 Metal 15.1 15.5 11 0.01702
(99.923 %), fis
(0.0775 %)
Cf252 6.217 2.645 Alpha 18.7885927 Metal 15.1 15.5 2.27 0.003525
(96.908 %), fis
(3.092 %)
Es-252 6.739 1.292 Alpha (76.4 %), 30.8604407 Metal 5.24 16.5 3.75 0.006184
ec (24.2 %),
1.3 Products from Ionizing Radiation …
beta (0.01 %)
The inverse of BAWmin is the theoretical maximum surface power density and the inverse of BVWmin is the theoretical maximum power density. In
calculating these parameters it is assumed that the compound containing the radioisotope has the maximum atomic density possible for the isotope, the
transport properties of the ionizing radiation into the transducer are perfect, the efficiency of the transducer is 100 %, and for a volume source the transducer
and radioisotope interface is ideal and seamless
21
sufficient energy to form quaternary electrons, and so on. It is the formation of these
higher order electrons from the primary event which dominate the formation of ion
pair production in radiation interaction with matter [19]. These higher order
interactions makes it extremely difficult to model electron-hole pair (or ion pair)
production using available transport codes like MCNP, GEANT4, etc. These
transport models are capable of producing results on the energy deposited in a
material as the photon, electron or ion passes through the material. With the
knowledge of energy deposition, it is possible to rely on the use of the W value to
find the number of electron-hole pairs or ion pairs produced. W values, which are
defined as the average energies required to form an electron-hole pair or an ion pair,
are experimentally measured and have been reliably used in designing calibrated
nuclear detectors for decades. The W value has been measured for specific gases,
gas mixtures and semiconductor materials (Tables 1.3 and 1.4). All forms of ion-
izing radiation (gamma rays, neutrons, betas, ions, etc.) have similar W values for
any given material that they interact with (e.g., the W value for gammas, betas [20],
ions [21], and neutrons interacting with helium will be similar). The first ionization
potential for each material is given in Table 1.3 as well as the ratio of the ionization
potential to W-value. Another important point to be made is that the W-value is
independent of the distance that the ionizing radiation travels before losing energy.
The W-value is simply a measure of how much energy it takes to form an ion pair in
a specific medium.
In addition to forming ion pairs, alpha particles such as those in Table 1.3 can
lose energy without creating ion pairs. This can occur when an electron receives
enough energy by Coulomb interactions to jump to a higher energy state but not
Table 1.4 The average energy required to produce ion pairs in various gases [21, 23]
Gas Energy per Ion Pair, First ionization Fraction of energy used
W (eV) potential (eV) in ionization (I/W)
H2 36.3 15.6 0.43
He (pure) 43 24.5 0.58
N2 36.5 15.5 0.42
O2 32.5 12.5 0.38
Air 35.0
Ne (pure) 36.8 21.5 0.58
Ar 26.4 15.7 0.59
Kr 24.1 13.9 0.58
Xe 21.9 12.1 0.55
CH4 30 14.5 0.48
C2H4 29 10.5 0.36
CO 34 14.3 0.42
CO2 34
CS2 26 10.4 0.40
NH3 39 10.8 0.28
1.3 Products from Ionizing Radiation … 23
sufficiently high to cause ionization. The W-value represents the average energy
required to create an ion pair and part of the inefficiency in producing ions is in the
production of non-ionized energy states. Non-ionizing energy losses go into the
creation of excited states and can be accounted for. A rare gas, for example, has a
first electronic state that is a metastable state. This state is typically illustrated with
the symbol “*” (e.g. He*). A W*-value is the average energy required to produce a
metastable state such as a single He* state in helium gas. For He, the W*-value is
about 90 eV/He* [22]. In a gaseous system, some of this energy will go into the
production of photons through spontaneous emission. Some of the energy ends up
as kinetic energy which raises the temperature of the gas. If the photons created in
the process are absorbed by the walls of the containment vessel or by the gas, then
eventually heat is produced. Basically, the interaction of radiation with matter will
have three products: heat; excited state production in the material; and ion pair
production in the material. If the system does not exploit the excited state pro-
duction or the ion pair production the energy in these two products, the excited
states and ions will interact with the material and eventually produce heat.
Ions and electrons have a very high effective temperature (Teff). The relationship
between energy and temperature is given by the relationship:
Teff ¼ E=kB ð1:19Þ
where E is the energy of the ion or electron in eV, kB is Boltzmann’s constant

8.617 10−5 eV K−1.
If the ion has an energy of 5,307,000 eV (from Po-210), the effective temper-
ature from Eq. (1.19) is 6.16 1010 K. Thus if the energy conversion method
exploits the excited state products and/or the ion pair products of the nuclear
reaction then the high effective temperature has an ideal Carnot cycle efficiency of
nearly 100 %. No known cycle approaches Carnot efficiency and neither will
ion-pair based cycles. However, it is intriguing to imagine how it might be possible
to use the effective high temperatures of ions and electrons to develop an efficient
direct energy conversion cycle. The cycles that are known will be discussed.
Radiation interactions with a solid have some similarities to radiation interaction
with gases. Electron-hole pairs are created as well as heat. The fraction of energy
that goes into electron-hole formation depends on the W value and the band-gap
energy of the material. In Table 1.5 some common semiconductor materials are
shown along with their relevant properties. As described above for gaseous mate-
rials, the mean ionization energy required to form one electron-hole pair in a solid is
the W-value. The ratio of the band-gap energy (Eg) to the W value is the effective
efficiency for producing electron-hole pairs through the interaction of radiation with
matter. As can be seen in the last column of Table 1.5, the electron-hole pair
production efficiency has considerable variation from one material to another.
Diamond has the highest at 0.442. Thus, when ionizing radiation interacts with
diamond, 44.2 % of the energy goes into electron-hole pair production. 55.8 % of
the energy essentially goes into heat production. If nothing is done to use the
24
Table 1.5 Properties for some common semiconductor materials which are useful for direct nuclear energy conversion [24]
Material Minimum Electron drift Fano Density Atomic Molar Displacement Mean ionization Eg/W
band-gap (Eg) mobility (l) (cm2/ factor (q) mass density energy (Ed) (eV) energy (W) (eV)
(eV) V-s) (F) (g/cm3) (g/mol) (mol/cm3)
Silicon 1.12 1450 0.115 2.329 28.1 0.0829 *19 3.63 0.308
Germanium 0.68 3900 0.13 5.323 72.6 0.0733 30 2.96 0.23
Gallium 1.42 8500 0.1 5.317 144.6 0.0368 10 4.13 0.344
arsenide
Silicon 2.9 400 0.09 3.22 40.1 0.0803 28 6.88 0.421
carbide
Gallium 3.39 1000 – 6.15 83.7 0.0735 24 8.9 0.381
nitride
Diamond 5.48 1800 0.08 3.515 12 0.293 43 12.4 0.442
Fig. 1.1 Energy conversion flow chart for radiation sources
electron-hole pairs that are being produced, they will recombine and the energy
eventually is transformed into heat by various processes.
Figure 1.1 illustrates how the two basic forms of energy produced by ionizing
radiation as it streams through matter (heat and ion pairs) can be used in energy
conversion. Both heat and ion pairs can be used in energy conversion processes and
these processes are illustrated as branches in the figure. The radiation source is
shown as a circle. The radioisotope can be embedded in matter in any of its four
possible phases:
1. Gas,
2. Solid,
3. Liquid, or
4. Plasma.
Likewise, the transducer can be made of matter in any of its possible four phases.
There are 16 possible interface combinations between the radioisotope and trans-
ducer (Table 1.6).
In an energy conversion system, the radiation first interacts with matter in a
solid, liquid, gaseous or plasma phase and then branches into heat production or ion
pair production. The ion pairs can be used with different transducers to produce
useful products, such as electricity, laser light, or chemicals [26].
Figure 1.1 shows typical combinations of radiation sources with transducers that
have been used. In Branch 1, the ions recombine and eventually produce heat. This
Table 1.6 Potential types of Case State of Fuel State of transducer

radioisotope and transducer
interface [25] 1 Solid Solid
2 Solid Liquid
3 Solid Gas
4 Solid Plasma
5 Liquid Solid
6 Liquid Liquid
7 Liquid Gas
8 Liquid Plasma
9 Gas Solid
10 Gas Liquid
11 Gas Gas
12 Gas Plasma
13 Plasma Solid
14 Plasma Liquid
15 Plasma Gas
16 Plasma Plasma
heat is then combined with the heat initially produced by the interactions between
the radiation source and the material. This heat can then be used to produce elec-
tricity in a variety of ways.
• The most common use is in the steam cycle of a commercial nuclear power
plant. However, there are other uses for the heat such as the Radioisotope
Thermoelectric Generator or RTG [5] which produces electricity from the
Seebeck effect.
• Another potential use for the heat is in thermionic energy conversion [27], in
which a hot electrode emits electrons due to the low work function of the
electrode material. The electrons are collected over a potential barrier by a cool
electrode thus producing a current flow.
• A different approach is thermophotovoltaics [28], in which a hot material emits
grey body radiation which is harvested by photovoltaic cells. These cells have
an appropriate band-gap for maximizing power production from the grey body
radiation.
• The Alkali Metal Thermal to Electric Converter (AMTEC) [29] is based on a
thermally regenerative electrochemical cell.
• It is also feasible to use heat to drive a carbon dioxide laser [25, 30].
Branch 2 utilizes only the energy from the ion pairs produced in a solid
transducer.
• This is typical of an alphavoltaic cell [31], where the electron-hole pairs are
separated by the voltage generated in the depletion region in a p-n junction of a
semiconductor material to produce a current flow.
• It is also used by a betavoltaic cell [17], where the electron-hole pairs are
separated by the voltage generated in the depletion region in a p-n junction of a
semiconductor material to produce a current flow.
• This process is also used in a reciprocating cantilever [32], where charge is
collected from the source on a thin cantilever and when sufficient charge is built
up the electrostatic force between the tip of the cantilever and the source causes
the cantilever to bend and touch thus driving a current in a repetitive process.
Branch 3 utilizes only the energy from the electron-hole pair production in a
solid to produce a laser.
• An example is through the interaction of ions with wide direct band-gap
materials which use the radiation source to produce the electron-hole pairs.
These electron-hole pairs then recombine and form the laser’s photons [33].
Branch 4 utilizes the photons from the recombination of electron-hole pairs in a
direct band-gap semiconductor to drive a photolytic laser.
• Examples are the Photon Intermediate Direct Energy Conversion (PIDEC)
system using a solid-state fluorescence medium to drive a photolytic laser [22].
Branch 5 utilizes the photons from the creation of excited states and ion pairs in
a gaseous system to excite a photolytic laser.
• Examples are the PIDEC system used to drive the iodine, xenon fluoride, and
Nd3+- doped glass or crystal lasers [22].
Branch 6 utilizes the energy from the excited states and ion pairs to directly drive
a gaseous nuclear-pumped laser.
• Examples include the bulk of the gaseous nuclear-pumped lasers directly excited
by nuclear reactions [26].
Branch 7 utilizes the energy from the production of excited states and ion pairs
for chemical production through radiolysis.
• Examples include a variety of radiolytic reactions with water to produce
hydrogen and CO2 to produce CO [26].
The branches which only utilize the energy from ion pair production will
inherently be limited to a maximum theoretical efficiency of 40–50 % since the heat
produced in the interaction of radiation with matter is wasted. As will be discussed,
process inefficiencies for systems represented by each of these branches will
determine the actual efficiency.
Branch 1 uses radiation for heat production. Branch 2 uses the production of
charged species in a solid to generate a current flow. Branch 3 uses the production
of charged species in a solid to produce laser photons. Branch 4 uses the production
of charged species in a solid to produce photons which are used to produce elec-
tricity from photovoltaic (PV) cells. Branch 5 uses the production of charged
species in a gas to produce photons which then interact with photovoltaic (PV) cells
to produce electricity [15].
The name nuclear battery comes from two facts.
1. “Nuclear” is used because nuclear energy is converted to ion-pairs or heat that is
used to directly to produce electricity.
2. The term “battery” comes from the energy stored by the radioisotopes.
1.4 Geometrical Considerations in the Interface

of an Isotropic Radiation Source to a Transducer
An isotropic point source emits ionizing radiation with equal probability in any
direction (see Fig. 1.2). A point source is a fundamental concept since any shape or
volume can be constructed from a combination of point sources (albeit the number
could approach infinity depending on the assumption of the point source volume).
As seen in Fig. 1.3, with a surface interface the isotropic emission of the ionizing
radiation has a limited probability of intersecting the transducer. In contrast,
Fig. 1.4 shows how a radioisotope embedded within the volume of the transducer
deposits the energy from the ionizing radiation with a high probability in the
transducer volume. Thus, the most efficient means of interfacing a radioisotope with
a transducer is with a volume interface. A second important consideration in
examining a nuclear battery design is to determine if the device uses a surface
interface or a volume interface. This indicates whether or not the device is
extracting the maximum amount of energy from the ionizing radiation that is
Fig. 1.2 Shown is an

illustration of a point source
where the particles that are
emitted are isotropic (that is
having an equal probability of
being emitted in any
direction)
1.4 Geometrical Considerations in the Interface … 29
Fig. 1.3 An illustration of how a radioisotope embedded in a compound is interfaced to a

transducer by surface area. Here, the probability of the ionizing radiation source depositing energy
in the transducer is dependent upon the trajectory of the radiation and geometrical factors
involving the volume of the transducer
Fig. 1.4 An illustration of

how a radioisotope embedded
in a compound is interfaced to
a transducer by volume. Here,
the probability of the ionizing
radiation source depositing
energy in the transducer is
high because the path of the
ionizing radiation is within
the volume of the transducer
emitted by the isotope. Surface type interfaces typically do not maximize the energy
absorption in the transducer.
Once the geometrical factors in the coupling between the radiation source and
the transducer are known, the next important issue centers on the energy deposition
in the transducer. It is possible to begin this analysis by looking at the trajectory of a
random particle (Fig. 1.5). As can be seen in the figure, the particle first travels
through the medium in which the radioisotope is embedded (a path length of Δl1).
Second, the particle passes through an inert material layer that makes up the first
part of the transducer structure (a path length of Δl2). Third, the particle travels
though the energy conversion layer within the transducer (a path length of Δl3). The
path of a beta particle will not be a straight line since it interacts with electrons in
the material and can transfer some or all of its energy through Coulombic collisions
with the electrons bound to the nuclei of the atoms that make up the compound.
Thus electrons will follow a random path [25, 34] as shown in Fig. 1.6. Thus,
electrons are complicated to track.
Ions (such as alpha particles and fission fragments) travel in a straight line
because they are much more massive then the electrons in the atoms that make up
Fig. 1.5 Shown is the path length of a single energetic ion emitted by the radioisotope with a
trajectory through a nuclear battery’s transducer. The device uses a surface interface between the
source and the transducer. The path length through the material in which the isotope is embedded
is Δl1. The path length through the inert part of the transducer is Δl2. The path length through the
energy conversion layer of the transducer is Δl3. Each particle emitted by the source will have a
different trajectory and path lengths Δl1, Δl2 and Δl3
the compound. An ion will not be deterred from its path by scattering interactions
with electrons. Thus, the electron density of each compound governs the energy
deposited by the ion in the material. Each material has a characteristic electron
density. When the ion travels through various layers of materials the energy
deposition rate in the specific layer will be dependent on the path length traveled
and the electron density of the compound. For example, in Fig. 1.5, the path lengths
for an ion moving through the three different layers are Δl1, Δl2 and Δl3. The energy
deposited by the ion into each of these three regions has to be calculated in order to
find the total energy deposited in the transducer by that single particle.
Consider that a radioisotope will emit 3 1010 particles per second per Curie.
To model the huge number of particles being emitted, their pathways through the
device and the total energy deposited in the transducer, a Monte Carlo based
computer code is typically used such as MCNP6 [16] or Geant4 [35].
As can be surmised from the discussions on particle emission, the random
emission angle of the particles from a radioisotope source will lead to design issues
and device inefficiencies.
The magnitude of the inefficiencies can be demonstrated by a simple example.
A sphere made of silicon carbide (SiC) has an americium-241 point source at its
center. As shown in Table 1.1, it emits an alpha particle with an energy of
approximately 5.485 MeV. A thin shell transducer (a SiC p-n junction) is placed
Fig. 1.6 Shown are the paths of beta particles emitted by a point source in the middle of a silicon
carbide sphere as modeled by GEANT4 [35]. The electrons follow a random walk. The straight
green lines are bremsstrahlung emissions [34]
within the sphere at radius Δl2 from the point source and has a shell thickness of Δl3
(Fig. 1.7). The alpha particles will stream through the undoped silicon carbide
material losing energy until it intersects with the thin transducer layer (the silicon
carbide p-n junction). Alpha particles will then travel through the transducer losing
energy. The question that must be answered is:
• What fraction of the total energy contained in the particles that are emitted by
the source is deposited in the most important component of the device, the
transducer?
Going about finding the energy deposited in the thin layer of the transducer
material shown in Fig. 1.7 requires the use of a computer code such as SRIM [36].
SRIM is designed to model an ion beam penetrating a target. The spherical
geometry set up shown in Fig. 1.7 has a point source of nearly monoenergetic alpha
particles being emitted isotropically. Due to the spherical symmetry of the problem,
all of the alpha particles will have almost the same path length albeit in different
directions. Figure 1.8 is the plot of the energy loss from alpha particles as a function
of distance as calculated by SRIM. Any alpha particle emitted by the point source
will stream through two regions and lose energy. The first region is the inert silicon
Fig. 1.7 Shown is an

illustration of an
Am-241point source in the
center of a silicon carbide
sphere which emits alpha
particles at an energy of
5.485 MeV. The alpha
particles stream from the
point source isotropically.
Due to the spherical
symmetry of the geometrical
configuration, the path length
for each alpha particle will be
the same
Fig. 1.8 Shown is a model of

energy loss of alpha particles
streaming through the SiC
sphere shown in Fig. 1.7
using SRIM [36]. The
transducer is a 1 lm thick
depletion layer of a silicon
carbide p-n junction. The
amount of energy deposited in
the transducer is the area
under the curve for the
transducer. As can be seen,
the bulk of the alpha energy is
deposited in the inert silicon
carbide material
carbide material in path length Dl2. The second region is the transducer with a path
length of Dl3. The area under the curve of region 2 along the path length Dl3
(Atransducer) is the amount of alpha energy that is deposited in the transducer. The
energy deposition efficiency is found by dividing Atransducer by the total area under
the curve consisting of the combined path lengths Dl2 + Dl3 (Atotal) as shown in
Eq. (1.20).
Atranducer
g¼ ð1:20Þ
Atotal
This example is very simple when compared to real designs where there are many
possible complex geometries for the source and transducer, as well as many ways in
which a source can be coupled to the transducer. In addition, if the source is a beta
emitter, the complexity of the problem multiplies significantly since beta particles
random walk their way through the material. Last but not least, the energy con-
version efficiency of the chosen transducer is an added consideration. So, the actual
operational efficiency of a nuclear battery consists of the transport efficiency of the
ionizing radiation getting through the various layers of the battery and depositing
energy in the transducer ηtr, the ionization efficiency which is the fraction of
deposited energy that creates ion pairs (for ion based energy conversion methods)
ηion, and the efficiency of the transducer to convert the ions produced in its volume
to the desired product (e.g., electrical current with a driving potential) ηec.
gsystem ¼ gtr gion gec ð1:21Þ
1.5 Methodology for Analysis
In Sects. 1.1 and 1.2, some basic tools for nuclear battery analysis have been
introduced. When analyzing a nuclear battery, it is very important that the reader
understand the fundamental principles behind the building blocks used in the design
of the system.
• First, consider the energy conversion method that is used. Is it dependent upon
heat or ion production? If it is dependent upon heat, then virtually all of the
energy in a radioisotope can be used to produce heat (minus the energy in
products which do not interact with matter such as neutrinos). If it is dependent
upon ion production (or ions plus excited states) than only about 40–50 % of the
energy from the radioisotope can be used. Something is wrong with a report if it
claims that their battery exceeds these limits.
• Second, determine if the battery uses a surface interface between the radioiso-
tope and the transducer or a volume interface in which the radioisotope is
embedded in the transducer volume.
• Third, authors will always divulge the radioisotope that is used in the battery.
This knowledge provides a great deal of information. From the radioisotope the
reader will know the BAWmin and the BVWmin. From the inverse of BAWmin
and BVWmin, the reader will know the absolute theoretical maximum possible
surface power density (W/cm2) and the absolute theoretical maximum power
density (W/cm3).
• Fourth, use common sense. Understand that the limits discussed above are
wildly optimistic. Realize that the radiation transport efficiency is typically far
from ideal and that the transducer efficiency will not be very close to being ideal.
Tools will be developed in later chapters to help the reader understand how to
analyze these parameters.
• Fifth, is the battery practical? There are limited supplies of isotopes available
from spent nuclear fuel and natural sources. Also, the costs of manufacturing
isotopes from high flux nuclear reactors or accelerators are very high, as will be
discussed in Chap. 2.
1.6 Summary
A nuclear battery needs a power source and a means of extracting energy from the
power source (e.g., a transducer). Ionizing radiation from radioisotopes is used as
the power source. Nuclear batteries are complex and their feasibility depends upon
many factors such as:
• the type of ionizing radiation (ions or electrons),
• the radioisotope,
• the material that the radioisotope is embedded in,
• the range of the ionizing radiation in the materials that make up the nuclear
battery,
• how the ionizing radiation source is interfaced to the transducer,
• how the transducer uses the energy that is deposited in it from the ionizing
radiation,
• the efficiency of the transducer,
• radiation damage to various components of the nuclear battery, such as the
transducer,
• and the means of extracting the energy from the transducer.
In subsequent chapters each of these issues and their impact on the nuclear
battery will be discussed.
Problems
1. Calculate the decay constants for tritium, krypton-85, strontium-90,
gadolinium-148, polonium-210, plutonium-238 and americium-241. Then, use
your answers to determine the number of particles in a 1 Ci pure isotope
source.
2. Suppose a report claims that a 20 lW battery is made with T2O embedded in
some transducer, with a volume of 0.000161 cm3. Does this system obey the
BVMmin limit? What if the battery produced 7 lW and had a volume of
0.00008 cm3?
3. Suppose that the BAWmin limit of a particular isotope is 125,000 cm2/W.
A paper claims they have made a battery with a surface area of 0.65 cm2 which
produces 4 lW. Based on BAWmin, is this battery possible?
4. Find the charged particle power output from 10 Ci of tritium, krypton-85,
strontium-90, gadolinium-148, polonium-210, plutonium-238 and
americium-241.
1.6 Summary 35
5. Plot the charged particle power decay curve from zero to 80 years for 10 Ci of
tritium, krypton-85, strontium-90, gadolinium-148, polonium-210,
plutonium-238 and americium-241.
6. What is maximum possible surface power density (W/cm2) for tritium,
krypton-85, strontium-90, gadolinium-148, polonium-210, plutonium-238 and
americium-241?
7. What is maximum possible power density (W/cm3) for tritium, krypton-85,
strontium-90, gadolinium-148, polonium-210, plutonium-238 and
americium-241?
8. Find two examples of nuclear batteries in the literature. Find their reported
BAW or BVW. How do they compare to BAWmin and BVWmin?
9. Find three examples of nuclear batteries in the literature. Analyze the batteries
and determine whether or not the reported data violate any of the principles
discussed in this chapter.
10. What are reasonable efficiencies for surface coupled and volume coupled
batteries?
11. Estimate the mass of tritium, krypton-85, strontium-90, gadolinium-148,
polonium-210, plutonium-238 and americium-241 that would be needed to
build a 100 W nuclear battery.
12. Estimate the mass of tritium, krypton-85, strontium-90, gadolinium-148,
polonium-210, plutonium-238 and americium-241 that would be needed to
build a 1,000 W nuclear battery.
13. What isotope do you think is the best one for use in a nuclear battery? Why?
14. Estimate the efficiency of the battery design described in the text which uses an
Am-241 point source in a sphere.
15. Download SRIM and make it operational with your computer. Generate a curve
like Fig. 1.8 for an Am-241 point source in a diamond sphere.
16. Generate a curve like Fig. 1.8 for an Am-241 point source in a gallium nitride
sphere using SRIM.
17. How would you optimize a nuclear battery design?
References
1. Moseley HGJ, Harling J (1913) The attainment of high potentials by the use of radium. Proc R
Soc (Lond) A 88:471
2. National Research Council Radioisotope Power Systems Committee (2009) Radioisotope
power systems: an imperative for maintaining US leadership in space exploration. National
Academies Press
3. Ritz F, Peterson CE (2004) Multi-mission radioisotope thermoelectric generator (MMRTG)
program overview. In: Aerospace conference, 2004. proceedings. 2004 IEEE, 2004, vol 5,
p. 2957
4. Blanke BC, Birden JH, Jordan KC, Murphy EL (1962) Nuclear battery-thermocouple type
summary report, 16th ed, USA E Commission, US Department of Commerce
5. Schmidt GR, Sutliff TJ, Dudzinski LA (2011) Radioisotope power: a key technology for deep
space exploration. In: Singh PN (ed) Radioisotopes—applications in physical sciences.
INTECH
6. Rappaport P (1954) The electron-voltaic effect in p-n junctions induced by beta-particle
bombardment. Phys Rev 93:246–247
7. Rappaport P (1956) Radioactive battery employing intrinsic semiconductor. USA Patent
2,745,973
8. Adams TE (2011) Status of betavoltaic power sources for nano and micro power applications.
https://www.navalengineers.org/SiteCollectionDocuments/2011%20Proceedings%
20Documents/GDDS2011/Adams.pdf
9. Friedrich-Wilhelm D, Jurgen S (1966) Semiconductor device. USA Patent 3257570 A, 21
June 1966
10. Walter P, Weddell JB (1967) Semiconductor battery. USA Patent 3304445 A, 14 Feb 1967
11. Knight RD (1967) Nuclear battery. USA Patent 3344289 A, 26 Sept 1967
12. ACJ, GBI, OLC, SSE (1972) Nuclear battery. USA Patent 3706893 A, 19 Dec 1972
13. Brown PM, Herda PG (2001) Isotopic semiconductor batteries. USA Patent 6238812 B1, 29
May 1998
14. Eiting CJ, Krishnamoorthy V, Rodgers S, George T, Robertson JD, Brockman J (2006)
Demonstration of a radiation resistant, high efficiency SiC betavoltaic. Appl Phys Lett
88:064101-064101-3
15. Prelas MA, Weaver CL, Watermann ML, Lukosi ED, Schott RJ, Wisniewski DA (2014) A
review of nuclear batteries. Prog Nucl Energy 75:117–148, Aug 2014
16. LANL (2014) MCNPX. https://mcnpx.lanl.gov/
17. S Deus (2000) Tritium-powered betavoltaic cells based on amorphous silicon. In: Photovoltaic
Specialists conference, 2000. Conference record of the twenty-eighth IEEE, 2000, pp 1246–
1249
18. Andreev V, Kevetsky A, Kaiinovsky V, Khvostikov V, Larionov V, Rumyantsev V et al
(2000) Tritium-powered betacells based on Alx > Ga1-xAs. In: Photovoltaic specialists
conference, 2000. conference record of the twenty-eighth IEEE, 2000, pp 1253–1256
19. Guyot J, Miley G, Verdeyen J (1972) Application of a two-region heavy charged particle
model to Noble-gas plasmas induced by nuclear radiations. Nucl Sci Eng 48:373–386
20. Jesse WP (1958) Absolute Energy to Produce an Ion Pair in Various Gases by Beta Particles
from 35S. Phys Rev 109:2002–2004
21. Friedlander G (1981) Nuclear and radiochemistry. Wiley, New York, NY
22. Prelas MA, Boody FP, Miley GH, Kunze JF (1988) Nuclear driven flashlamps. Laser Part
Beams 6:25–62
23. Friedländer G, Kennedy JW (1955) Nuclear and radiochemistry. John Wiley
24. Wrbanek JD, Wrbanek SY, Fralick GC, Chen L-Y (2007) Micro-fabricated solid-state
radiation detectors for active personal dosimetry. NASA/TM 214674
25. Prelas M (2016) Nuclear-pumped lasers, 1st edn. Springer International Publishing
26. Prelas MA, Loyalka SK (1981) A review of the utilization of energetic ions for the production
of excited atomic and molecular states and chemical synthesis. Prog Nucl Energy 8:35–52
27. Hatsopoulos GN, Gyftopoulos EP (1973) Thermionic energy conversion, vol 1. The MIT
Press, Processes and Devices
28. Nelson RE (2003) A brief history of thermophotovoltaic development. Semicond Sci Technol
18:S141–S143
29. Hunt TK, Weber N, Cole T (1981) High efficiency thermoelectric conversion with Beta”-
Alumina Electrolytes, the sodium heat engine. Solid State Ionics 5:263–266
30. Fein ME, Verdeyen JT, Cherrington BE (1969) A thermally pumped CO2 laser. Appl Phys
Lett 14:337–340
31. Department of Energy, “Summary of Plutonium-238 Production Alternatives Analysis Final
Report. I. N. Laboratory, DOE, 2013
32. Duggirala R, Li H, Lal A (2008) High efficiency radioisotope energy conversion using
reciprocating electromechanical converters with integrated betavoltaics. Appl Phys Lett 92
References 37
33. Watermann ML, Prelas MA (2013) Integrated solid-state nuclear pumped laser/reactor design
for asteroid redirection. Transactions of the American Nuclear Society
34. Oh K, Prelas MA, Rothenberger JB, Lukosi ED, Jeong J, Montenegro DE et al (2012)
Theoretical maximum efficiencies of optimized slab and spherical betavoltaic systems utilizing
Sulfur-35, Strontium-90, and Yttrium-90. Nucl Technol 179:9
35. CERN (2015) Geant4. http://geant4.cern.ch/index.shtml
36. Ziegler JF, Ziegler MD, Biersack JP (2010) SRIM–The stopping and range of ions in matter
(2010). Nucl Instrum Methods Phys Res, Sect B 268:1818–1823
Chapter 2
Radioisotopes
Abstract Radioisotopes are of critical importance in nuclear batteries due to their role
as the battery’s power source. Characteristics of radioisotopes such as type of radiation
emitted, decay energy of its radiation, and half-life all influence their usefulness. The
type of radiation will determine whether or not there is a good match between the
range of radiation and the scale length of the transducer, the decay energy of the
radiation determines the effective power density of the source, and the half-life
determines the effective lifetime of the nuclear battery. Because of this, special
attention must be given during the radionuclide selection process so that it fits the
desired design criteria. While simple in concept, the limitations enforced by the
physical properties of the substance being used represent an extremely significant
barrier to the production of nuclear batteries. Another significant barrier to the pro-
duction of nuclear batteries arises from limitations in the isotope supply and cost of
production. These limitations can be surprising, as a common misconception is that
radioisotope supplies are at least abundant, even if they cannot be cheaply produced.
For nuclear batteries, supply is an issue because of the large amount of isotopes that
would be needed to produce significant power and overcome transducer inefficiencies.
For example, a good transducer can only get an average of 1 mW of power per Ci of
radionuclide. Production is also an issue because of the cost of producing radionuclide
amounts large enough to produce significant power in the nuclear battery. In this
chapter, the reader receives a more detailed introduction to radioisotopes, which
includes an analysis of origin, existing supplies, production methods, associated
production costs, and some basic isotope selection complications.

Keywords Radioisotopes supplies Radioisotope production Radioisotope costs
2.1 Existing Radioisotope Supplies
Radioactive elements existed in the Earth since its formation and can be found in
air, water, and soil. Their origin is both terrestrial and extraterrestrial. They can be
classified into three groups (Fig. 2.1): primordial, cosmogenic, and manmade.

40 2 Radioisotopes
Fig. 2.1 Types of

radioisotope sources
Primordial radionuclides are the ones created in Earth’s formation and their
radioactive decay products, cosmogenic radionuclides are the ones produced by
cosmic-ray interactions, and manmade radionuclides are those produced by humans
after the discoveries of radioactivity and nuclear fission [1]. As both primordial and
cosmogenic radioisotopes have a natural origin, they are called Naturally Occurring
Radioactive Materials (NORM).
2.1.1 Primordial Radioisotopes
Some of the primordial radionuclides that were created during the Earth’s formation
have survived until now due to their long half-lives [2]. One example is U-238,
whose half-life is 4.468 109 years. This time is very close to the age of the Earth,
calculated to be around 4.5 109 years [3]. Some of those long-lived radioisotopes
undergo a series of successive radioactive decays, generating radioisotopes with
shorter half-lives. Those series of successive radioactive decays are known as decay
chains, these chains stop with the production of a stable isotope.
The three existing natural decay chains are the uranium, thorium, and actinium
series. Each name originates from the parent radionuclide, which has a much longer
half-life than the other radioisotopes in the series. This condition generates secular
equilibrium, which is when the activity of the daughter isotope equals the activity of
the parent isotope after approximately eight half-lives of the daughter. Taking
advantage of this condition, the amount of daughter radioisotopes can be estimated
if the amount of the parent radioisotope is known. Uranium and thorium are the two
2.1 Existing Radioisotope Supplies 41
parent radioisotopes in the three decay chains. Assuming world uranium reserves of
about 7,900,000 t and world thorium reserves of 5,385,000 t [4, 5], an estimated
inventory of the radioisotopes pertaining to the three decay chains was calculated
(Table 2.1). The table was calculated assuming secular equilibrium and an undis-
turbed sample, which means no loss of gaseous radioisotopes like Radon. The
Nuclear data for Table 2.1 was obtained from the National Nuclear Data Center site
of Brookhaven National Laboratory [6], and the activities of the daughter were
calculated considering the branching probabilities of the decay chains.
One additional decay chain, called the neptunium series, has a parent radioiso-
tope of Np-237. It is no longer found in nature since the half-life of Np-237 is 3
three orders of magnitude less than the age of the Earth. However, this series can be
produced artificially through Pu-241, which decays to Np-237.
There are also primordial radionuclides, present in nature, that do not form part of
decay chains. They have very long half-lives and their decay product (daughter) is a
stable isotope. To have an idea of the amount of the non-series primordial radionu-
clides, an estimation was done and is shown along with half-lives, relative abundances,
and types of radiation in Table 2.2. The supplies were calculated using information
about the abundance of elements in the Earth’s crust (EC) and in the sea from the
Handbook of Chemistry and Physics [7]. Also, it was assumed that the mass of the
Earth’s crust is 2.36 1022 kg, the mass of the hydrosphere is 1.664 1021 kg, and
seawater is about 97 % of the hydrosphere [7]. Sm-149 and Dy-156 were not included
in the table because some references consider them radioisotopes [1], but others [6]
label them as stable isotopes. This might be due to fact that it is very difficult to
determine if an isotope is radioactive when it has a very long half-life. As improved
experiments (greater sensitivity and lower background) become available, the number
of confirmed non-series radionuclides may increase [8].
K-40 is the most abundant radioisotope in Table 2.2, which is not surprising
because it is the eighth most abundant element in both the Earth’s crust and the sea.
If we could make 1 kW nuclear batteries using the total amount of K-40 available,
that would result in about 423 million batteries. However, it is impossible and
infeasible to extract or separate such high quantities of a radioisotope. Further, none
of the non-series primordial radioisotopes meet the proper characteristics to be used
in nuclear batteries. It may be economical to use primordial radionuclides only if
large quantities are not needed due to extraction difficulties.
2.1.2 Cosmogenic Radioisotopes
Cosmogenic radioisotopes are created by the interaction of cosmic-ray particles

with Earth’s soil or atmosphere. The flux of cosmic-ray particles is made up of
energetic particles (mostly protons with energies of 1–100 MeV) striking Earth’s
atmosphere. They originate from both the sun and other galaxies in the universe [1].
Most of the atmospheric radionuclide generation takes place in the stratosphere [9]
with a higher production around the magnetic poles and lower in the regions near
42 2 Radioisotopes
Table 2.1 Estimated world supply of natural decay chains’ radioisotopes

Series Isotope Half-life World supply Total activity Type of
(year) (kg) (Ci) radiation
Uranium U-238 4.47E+09 7.843E+09 2.636E+06 a, c
Th-234 6.60E−02 1.139E−01 2.636E+06 b, c
Pa-234 m 2.20E−06 3.796E−06 2.631E+06 b, c
U-234 2.46E+05 4.237E+05 2.636E+06 a, c
Th-230 7.54E+04 1.279E+05 2.636E+06 a, c
Ra-226 1.60E+03 2.667E+03 2.636E+06 a, c
Rn-222 1.05E−02 1.714E−02 2.636E+06 a, c
Po-218 5.89E−06 9.469E−06 2.636E+06 a
Pb-214 5.10E−05 8.039E−05 2.636E+06 b, c
Bi-214 3.78E−05 5.971E−05 2.636E+06 b, a
Po-214 5.21E−12 8.214E−12 2.636E+06 a, c
Pb-210 2.22E+01 3.437E+01 2.636E+06 b, c
Bi-210 1.37E−02 2.124E−02 2.636E+06 b, a
Po-210 3.79E−01 5.865E−01 2.636E+06 a, c
At-218 4.75E−08 1.528E−11 5.270E+02 a
Pa-234 7.64E−04 4.349E−06 8.692E+03
Actinium U-235 7.04E+08 5.691E+07 1.214E+05 a, c
Th-231 2.91E−03 2.284E−04 1.214E+05 b, c
Pa-231 3.28E+04 2.570E+03 1.214E+05 a, c
Ac-227 2.18E+01 1.679E+00 1.214E+05 b, a, c
Th-227 5.11E−02 3.889E−03 1.197E+05 a, c
Ra-223 3.13E−02 2.370E−03 1.214E+05 a, c
Rn-219 1.25E−07 9.333E−09 1.214E+05 a, c
Po-215 5.64E−11 4.121E−12 1.214E+05 a, b
Pb-211 6.86E−05 4.918E−06 1.214E+05 b, c
Bi-211 4.07E−06 2.915E−07 1.214E+05 a, b, c
Tl-207 9.07E−06 6.356E−07 1.211E+05 b, c
Po-211 1.64E−08 3.278E−12 3.397E+02 a, c
Fr-223 4.18E−05 4.372E−08 1.675E+03 b, a, c
Thorium Th-232 1.400E+10 5.385E+09 5.927E+05 a, c
Ra-228 5.750E+00 2.174E+00 5.927E+05 b
Ac-228 7.016E−04 2.652E−04 5.927E+05 b, c
Th-228 1.912E+00 7.227E−01 5.927E+05 a, c
Ra-224 9.944E−03 3.693E−03 5.927E+05 a, c
Rn-220 1.762E−06 6.427E−07 5.927E+05 a, c
Po-216 4.595E−09 1.646E−09 5.927E+05 a
Pb-212 1.214E−03 4.266E−04 5.927E+05 b, c
Bi-212 1.151E−04 4.046E−05 5.927E+05 a, b, c
Po-212 9.475E−15 2.134E−15 3.797E+05 a
Tl-208 5.805E−06 7.191E−07 2.129E+05 b, c
Table 2.2 Estimated world supply of non-series primordial radioisotopes
Radio-isotope Relative Half-life Supply Earth’s crust Supply sea Total activity EC Total activity sea Type of
abundance (years) (kg) (kg) (Ci) (Ci) radiation
K-40 0.0117 1.248E+09 5.771E+16 7.327E+13 4.228E+14 5.368E+11 b, e
V-50 0.25 2.10E+17 7.080E+15 1.009E+10 2.366E+05 3.371E−01 e, b
Rb-87 27.85 4.81E+10 5.915E+17 5.394E+13 5.144E+13 4.691E+09 b
2.1 Existing Radioisotope Supplies
In-115 96.67 4.41E+14 5.704E+15 3.121E+13 4.027E+07 2.203E+05 b

Te-123 0.87 9.20E+16 2.053E+11 n/a 6.253E+00 n/a e
La-138 0.089 1.02E+11 8.192E+14 4.884E+06 2.067E+10 1.232E+02 e, b
Ce-142 11.7 5.00E+16 1.836E+17 2.266E+08 9.370E+06 1.156E−02 b
Nd-144 23.8 2.29E+15 2.331E+17 1.076E+09 2.523E+08 1.164E+00 a
Sm-147 15.1 1.06E+11 2.512E+16 1.097E+08 5.635E+11 2.460E+03 a
Sm-148 11.35 7.00E+15 1.888E+16 8.244E+07 6.414E+06 2.800E−02 a
Gd-152 0.205 1.08E+14 3.000E+14 2.316E+06 6.314E+06 4.876E−02 a
Hf-174 0.163 2.00E+15 1.154E+14 1.842E+07 1.156E+05 1.844E−02 a
Lu-176 2.588 3.76E+10 4.886E+14 6.266E+06 2.655E+10 3.405E+02 b
Re-187 62.93 4.33E+10 1.040E+13 4.063E+09 4.609E+08 1.801E+05 b
Pt-190 0.0127 6.50E+11 1.499E+10 n/a 4.225E+04 n/a a
Pb-204 1.4 1.40E+17 4.626E+15 6.779E+08 5.701E+04 8.355E−03 a
43
44 2 Radioisotopes
the equator. The half-lives of these radionuclides are relatively short, which means
that they are present in Earth today due to a continuous process of production,
decay, and transport. The distribution of these radionuclides is done by air currents.
As such, most isotopes coagulate to aerosol particles that could become conden-
sation nuclei for the formation of clouds, and getting deposited in Earth’s surface by
precipitation. Cosmogenic radionuclides are also produced in rocks but with a
smaller rate, depending on the location, size, and exposure history of the rocks [1].
Two examples of cosmogenic radioisotopes are tritium (H-3) and radiocarbon
(C-14). The average inventory of cosmogenic isotopes in the environment is shown
in Table 2.3, contrasting the values of two references [1, 9]. Among these
radioisotopes only tritium is a possible candidate to be used in nuclear batteries. Its
worldwide steady state inventory is estimated to be 35 106 Ci, and if we look at
the power that one can extract from tritium in Table 1.1, the worldwide inventory
would produce 1312.5 W. Note that in a nuclear battery this amount of power
would be reduced further depending on the efficiency of the battery.
Table 2.3 Cosmogenic isotope inventory

Radioisotope Half-life Inventory (MCi) Type of radiation
O’Brien Masarik
Be-10 1.39E+06 years 3.5E+00 2.5E+00 b
Al-26 7.17E+05 years 5.6E−03 1.6E−03 e
Cl-36 3.01E+05 years 1.2E−01 2.7E−01 b, e
Kr-81 229,000 years – 4.9E−05 e
C-14 5700 years 2.4E+02 3.0E+02 b
H-3 12.32 years 3.5E+01 3.5E+01 b
Na-22 2.603 years 5.0E−03 – e
S-35 87.37 days 8.0E−02 – b
Be-7 53.24 days 7.7E+00 4.9E+00 e
P-33 25.35 days 4.5E−02 – b
P-32 14.26 days 4.7E−02 – b
Mg-28 20.92 h 2.3E−03 – b
Na-24 15 h 9.7E−03 – b
S-38 170.3 min 2.0E−03 – b
Si-31 157.3 min 2.2E−02 – e
F-18 109.77 min 4.2E−03 – e
Cl-39 56.2 min 1.3E−01 – b
Cl-38 37.24 min 8.0E−02 – b
Cl-34m 32.00 min 5.4E−03 – e, IT
Al-29 6.56 min 1.4E−03 – B
S-37 5.05 min 5.5E−03 – b
Ne-24 3.38 min 7.2E−04 – b
P-30 2.45 min 5.8E−03 – e
Al-28 2.24 min 2.0E−02 – b
2.1.3 Manmade Radioisotopes
Manmade radioisotopes are generated by the production of nuclear fuels (separation

and enrichment of uranium or thorium from natural decay chains), the reprocessing
of nuclear fuels (separation of radioisotopes from spent fuel), nuclear power
reactors (production of radioisotopes by neutron capture), particle accelerators, and
nuclear weapon explosions. The first four production methods are discussed with
more detail in Sect. 2.2.
Nuclear reactors use as fuel mostly uranium enriched with the radioisotope
U-235. It is obtained from mining, milling, and refining uranium ores. In the
process of fabricating nuclear fuel, the gaseous radioisotope Rn-222 could be
released. U-235 is utilized because it has a high thermal neutron cross section for
fission, which means that when U-235 absorbs a thermal neutron there is a high
probability that it splits into two fission fragments and liberates some neutrons and
energy. Nuclear reactors take advantage of this to create a controlled chain reaction
in which the neutrons released generate more fissions and the heat produced is
transformed into electricity. The fuel usually consists of nuclear fuel pellets of
uranium dioxide (UO2), Fig. 2.2, surrounded by a cladding made of a Zirconium
alloy [10]. Under normal operation, most of the fission fragments created remain
trapped in the nuclear fuel cladding. These fragments, of about 800 different iso-
topic species, have atomic masses that range from 72 to 160 amu. They are
radioactive on the neutron rich side of stability (decay by b−), the most probable
fission products have masses of 94 and 140 [6, 11, 12]. Figure 2.3 show the fission
fragment yields per atomic mass of radioisotope for thermal and fast neutrons. Note
that these are the direct yields and that many of these radioisotopes undergo further
Fig. 2.2 Nuclear fuel pellets [13]

46 2 Radioisotopes
Fig. 2.3 U-235 fission yields

for high- and low-energy
(thermal) incident neutrons
[11]
beta decays until they reach a stable isotope. Among the isotopes created in the
operation of a nuclear reactor, some of them are stable and others have very short
half-lives. Around fifty of them live longer than 25 min, and these are the ones that
could be dangerous if released during a nuclear accident [1]. In previous nuclear
accidents, significant amounts of I-131 and Cs-137 have been released to the
environment. However, in normal operation, a very small portion of these
radioisotopes leak through the cladding and are released to the moderator (usually
light water). However, not all the radioisotopes are generated from fission frag-
ments in a nuclear reactor. The interaction of the neutron flux with the structure
materials also produces many radioisotopes such as Cr-51, Mn-54, Fe-55, Fe-59,
and Co-60 [1]. Moreover, part of the fuel undergoes neutron capture instead of
fission, producing heavier radioisotopes.
Once the fuel is used, a fraction of U-235 and most of the U-238 remains
unchanged. Some Pu-239 is created by neutron absorption in U-238. The two
radioisotopes of uranium, as well as the plutonium, can be reused to fabricate new
fuel. This is called reprocessing. When the nuclear fuel is reprocessed, the
radioisotopes created as fission fragments can be separated using different processes
discussed in Sect. 2.2. Although it seems a very good option to extend the supply of
nuclear fuel and separate useful radioisotopes, it is costly and controversial due to
proliferation issues with fissile radioisotopes that could be used to produce weapons
like Pu-239 [14].
Once separated from spent fuel, these radioisotopes can be used in different
applications like radiopharmaceuticals, medical imaging, tracers, radiation steril-
ization of food, elimination of insect pests, measuring thickness of materials, remote
power sources (RTGs, and nuclear batteries), among others [12].
Worldwide inventories for these radioisotopes are difficult to obtain because the
detailed information needed about the nuclear reactors is usually kept confidential
and so is not readily available. However, efforts have been made to estimate the
world inventory of particular elements and radioisotopes. Some of these, such as
plutonium, Am-241, and Kr-85, are useful for nuclear batteries. For example,
Pu-238 has been an important radioisotope for space programs since it has been
used as a source of power for RTGs. The total world estimated supply of the
element (Pu) by the end of 2014 was 2,627 tonnes. Of this, 2,388 tonnes corre-
sponded to civil nuclear reactors (2,113 being irradiated in working nuclear fuel,
and 275 in spent fuel), while 240 tonnes corresponded to military stocks [15]. Thus,
the most readily available Pu for potential use in nuclear batteries are the 275 tonnes
in spent fuel. Two potential Pu sources for nuclear batteries are Pu-238 and Pu-241,
which have an estimated concentration of total Pu in typical PWR burned fuel of
1.3 and 4.7 %, respectively [16]. Therefore, there are about 3.57 tonnes of Pu-238
and 12.92 tonnes of Pu-241 worldwide in spent fuel. The amount of Americium that
has been produced by the end of 2004 was estimated to be 87 tonnes, with about 68
tonnes being Am-241, a decay product of Pu-241 [17]. In the case of Kr-85,
Ahlswede et al. [18] have estimated the global atmospheric content of this
radioisotope to be approximately 5400 PBq. This activity corresponds to the
amount of Kr-85 that has been released to the atmosphere mainly by nuclear
reprocessing facilities, which are the major contributor, and other small sources like
nuclear power reactors, naval reactors and isotope production plants. It does not
consider the amount of Kr-85 in fuel that has not been reprocessed and thus not
released to the atmosphere. Table 2.4 shows the estimated world inventories of
Kr-85, Pu-238, and Am-241. The cost of separating these isotopes from spent fuel
was considered by the National Research Council [19]. The conclusion of this
review was that the feasible approach is to use isotopes that have already been
separated in the weapons program to produce Pu-238. The available separated
inventory of Pu-238 within the National Laboratories in the United States is about
39 kg and the estimated Np-237 inventory is 300 kg [20]. Pu-238 is created from
the 237Np(n, c)238Np ! 238Pu + b reaction. It is feasible to produce Pu-238 from
the separated inventory of Np-237 using high neutron flux reactors such as the
Table 2.4 Estimated world supplies of Kr-85, Pu-238, Pu-241, and Am-241
Radioisotope Half-life Year of Total activity World supply
(year) estimation (Ci) (kg)
Kr-85‡ 10.752 2009 1.46E+08 3.73E+02
Pu-238 87.7 End of 2014 6.11E+07 3.57E+03
Pu-241 14.325 End of 2014 1.34E+09 1.29E+04
Am-241 432.6 End of 2004 2.33E+08 6.80E+04
‡
Global atmospheric content
48 2 Radioisotopes
Table 2.5 Average information of PWR and BWR reactors in the U.S
Reactor First month of Original thermal Number of Initial mass of
type operation power (MWt) assemblies U (kg)
PWR Sep-80 2908.9 184 424
BWR Jan-79 2799.9 755 183
Advanced Test Reactor at Idaho National Laboratory or the High Flux Isotope
Reactor at Oak Ridge National Laboratory [19, 21] to produce about 5 kg of
Pu-238 per year. The cost of building the production capability is estimated to be 77
million dollars US [19, 20]. The cost of Pu-238 per kg is about 8 million dollars US.
Even though the inventory of unseparated Np-237 and Pu-238 worldwide from
commercial reactor spent fuel is significant (see Table 2.4) [17], the cost of spent
fuel processing is too high to be considered feasible.
Although the estimation of the global inventory has been done only for a few
radioisotopes, it is possible to make estimations about the radioisotopes contained
in U.S. used nuclear fuel because more information in available. A report by Oak
Ridge National Laboratory estimates that as of 2011, there are 67,600 metric tons of
heavy metal (MTHM) of commercial used nuclear fuel [22], with 43,900 MTHM
correspond to PWR and 23,700 MTHM to BWR. Assuming 70 % (PWR) and
57 % (BWR) as the weight percentages of Uranium found initially in fuel
assemblies, the weights corresponding only to spent fuel are: 30,790 tonnes
(PWR) and 13,580 tonnes (BWR). Using these masses and the average charac-
teristics of PWRs and BWs, an estimation of the radioisotope inventories can be
made using the methodology proposed by Yancey and Tsvetkov to analyze the
spent nuclear fuel [23]. Information representing average reactors is introduced in
ORIGEN-ARP (Scale software package) to develop a plant model and estimate the
masses of the radioisotopes contained in a used fuel assembly. Once the contents of
an “average” PWR and BWR assemblies are obtained, the result is multiplied by
the number of assemblies calculated dividing the total mass of spent fuel by the
initial mass of U per assembly. The average information about PWR and BWR
reactors used in ORIGEN-ARP is shown in Table 2.5 [22]. The estimation of
medium half-life radioisotope inventory in spent fuel is shown in Table 2.6. By the
end of 2009, there were about 240,000 MTHM of spent fuel worldwide, with
64,500 MTHM corresponding to the U.S. [24], therefore the U.S. spent fuel rep-
resents about 27 % of the global amount. Likewise, data in Table 2.6 could rep-
resent about 27 % of the global radioisotope supply. This is a gross estimation,
because information about U.S. reactors is not necessarily representative of the
foreign reactors, which in some cases have different types of fuel, power, etc.
However, these numbers provide an idea on how much radioisotope activity might
be available by extracting radioisotopes from spent fuel.
2.2 Radioisotope Production 49
Table 2.6 Estimation of Radioisotope Half-life Total activity Supply

radioisotope supply from (year) (Ci) (g)
spent fuel in the U.S.
including potential useful Cs-137 30.08 4.62E+09 5.31E+07
isotopes for nuclear batteries Sr-90 28.9 2.98E+09 2.11E+07
Cm-244 18.1 3.03E+08 3.74E+06
Kr-85 10.752 1.69E+08 4.29E+05
Am-241 432.6 1.58E+08 4.61E+07
Eu-154 8.6 1.01E+08 3.73E+05
Pm-147 2.6234 4.05E+07 4.37E+04
Sm-151 90 1.88E+07 7.16E+05
Cs-134 2.065 1.53E+07 1.18E+04
Eu-155 4.753 1.10E+07 2.23E+04
Sb-125 2.76 3.23E+06 3.08E+03
Cm-243 29.1 1.32E+06 2.56E+04
Sn-121 m 43.9 7.31E+05 1.36E+04
Eu-152 13.528 1.34E+05 7.59E+02
Ru-106 1.02 3.92E+04 1.18E+01
Cd-113 m 14.1 1.15E+04 5.14E+01
Th-228 1.912 1.53E+03 1.86E+00
U-232 68.9 1.51E+03 6.82E+01
Pm-146 5.53 1.33E+03 2.99E+00
Pu-236 2.858 2.84E+02 5.43E−01
Cf-250 13.08 2.39E+01 2.19E−01
Tm-171 1.92 4.80E+00 4.40E−03
Cf-252 2.645 1.12E+00 2.08E−03
Ac-227 21.772 9.10E−02 1.26E−03
Po-210 0.379 8.71E−04 1.94E−07
Pb-210 22.2 8.71E−04 1.14E−05
Ra-228 5.75 1.13E−05 4.14E−08
2.2 Radioisotope Production
Due to the limited quantities of radioisotopes available from Naturally Occurring

Radioactive Materials and human stockpiles, acquiring significant amounts of many
isotopes from existing supplies is unattainable. In order to obtain these substances
in quantities sufficient for use, some manner of radioisotope production must be
pursued. Unfortunately however, production of these isotopes remains a scientifi-
cally complex process. This complexity places a significant barrier on researchers
attempting to develop new radioisotope separation methods. Currently, there are
only four general methods available for the production of these substances: sepa-
ration from nuclear reactor spent fuel, separation from natural decay chains, pro-
duction of radioisotopes through neutron capture in a nuclear reactor, and
radioisotope production via the use of a particle accelerator.
50 2 Radioisotopes
2.2.1 Separation from Spent Fuel
The first method of radioisotope production is the separation of isotopes from the
spent fuel of a nuclear reactor. In this method, the spent fuel is transported to a
reprocessing facility, where it is undergoes a series of chemical and/or physical
processes. Through these processes, different substances are isolated and removed,
leaving the desired chemical elements as products. These chemical elements can
then be further processed and refined to yield the proper concentration of the needed
radioisotope. Once the process is complete, these recycled isotopes can be used in
many applications, including nuclear batteries.
Unfortunately, one of the biggest scientific concerns with this method (in regards
to nuclear batteries) is the composition of the isotopes present in spent fuel. Most
nuclear reactors in the United States are Light Water Reactors (LWR’s). These
reactors produce spent fuel that consists of approximately 96 % uranium and less
than 1 % plutonium [25]. The remaining 3 % consists of a group of hundreds of
radioisotopes known as “fission products,” which were produced during the reac-
tor’s operation. It can therefore be concluded that, if a chemical element other than
uranium or plutonium is desired for a nuclear battery’s operation, an extremely high
amount of separation and refinement would be required. This would be a limiting
factor in the rate of production. In addition, this problem would also affect the cost
of radioisotopes produced using the method, which is an issue discussed later in the
chapter.
As was mentioned previously, these concerns apply primarily to scientists
interested in the research and development of nuclear batteries. Since this was not
the original intention of spent fuel reprocessing, many methods and variations of
spent fuel reprocessing have been developed. Before termination of civil spent fuel
reprocessing in the United States, the nation employed three different processes to
recycle spent fuel: the bismuth phosphate process, the reduction and oxidation
(REDOX) process, and the plutonium uranium redox extraction (PUREX) process.
2.2.1.1 Bismuth Phosphate Process
The first method of radioisotope separation from spent fuel, known as the bismuth
phosphate process, originated in the early 1940s from the Metallurgical Laboratory
at the University of Chicago. At the time, the University of Chicago was evaluating
several chemical separation processes for use in the Manhattan Project [26], a secret
military program to design and construct the first nuclear bomb. After evaluation of
the potential processes was completed, it was concluded that the bismuth phosphate
process was the best choice for separation of plutonium radioisotopes. Three large
bismuth phosphate plants were subsequently built by The Manhattan District in
Hanford, Washington. These plants were called T, B, and U, but were often
nicknamed “canyons” or “Queen Mary’s” because of their massive size (see
Fig. 2.4). The fundamental concept behind these bismuth phosphate plants was to
Fig. 2.4 T Plant (left) and U Plant (right) in Hanford, Washington [27]. “Canyons” such as these
could be larger than 800 feet long, 85 feet wide, and 100 feet high
take advantage of different plutonium (Pu) valence states to recycle the spent fuel.
In the Pu4+ valence state, plutonium has the ability to co–precipitate with bismuth
phosphate (BiPO4). However, in the Pu6+ valence state, plutonium loses this
capacity. By manipulating plutonium through this difference in valence states, the
bismuth phosphate process produced an end product that consisted of a plutonium
nitrate solution.
The first step in the bismuth phosphate process was to remove the aluminum
reactor fuel cladding. This was done by submerging the spent fuel in a boiling
sodium hydroxide and sodium nitrite solution. The bare fuel, containing mostly
uranium and small amounts of plutonium and other fission products, was then
dissolved in a concentrated aqueous nitric acid solution. This solution was stabi-
lized with sulfuric acid to stop the uranium from precipitating [28]. To acquire the
proper valence state, nitrous acid was added to reduce the plutonium. Bismuth
phosphate was added next, in order to co-precipitate with the plutonium. The
plutonium and bismuth phosphate were isolated through the use of centrifuges, and
then washed [29]. This complete cycle would then be repeated multiple times in
order to purify the product. After the cycles had completed, the plutonium solution
would be further decontaminated and concentrated via a lanthanum fluoride carrier,
shrinking the volume of the solution from about three hundred and thirty gallons
down to 8 gal [26]. This volume reduction transformed the plutonium solution into
a paste. Further processing of the paste turned it into plutonium metal. For this
process, the waste stream contained the uranium and most of the fission products
(see Fig. 2.5).
Unfortunately, there were several significant problems with the bismuth phos-
phate process. First, the spent fuel had to be processed in batches [30]. A second
major issue was the loss of several useful products in the waste stream, such as the
uranium. One more problem was the low reprocessing rate. On average, a bismuth
phosphate plant could only reprocess about one metric ton of fuel per day [26].
Another disadvantage was product quantity: One metric ton of bare fuel produced
only 2.5 kg of plutonium product. Additionally, that same ton of bare fuel created
about ten thousand gallons of liquid waste. This resulted in approximately 1.5
52 2 Radioisotopes
Fig. 2.5 The bismuth phosphate process [26]
million gallons of contaminated wastewater being discharged into the ground each
day [31]. Of course, in addition to all these problems, there were issues with the
reprocessing plants themselves.
In order to accommodate the necessary radiation shielding for the spent fuel, the
T, B, and U reprocessing plants were extremely large. The main buildings (can-
yons) were over 800 feet long, 85 feet wide, and 100 feet high. Each facility utilized
six-foot thick concrete walls to shield workers from the radioactivity. Every plant
was divided into at least 20 sections, known as “process cells,” with removable
shield covers to separate them. Direct exposure to the fuel or the processing
equipment was extremely hazardous due to the radiation dose received, and could
reach fatal exposure levels in under a minute [31]. To minimize such exposure, the
facilities used overhead cranes and manipulators operated by remote access. In
order to successfully operate this equipment, each facility had shielded operating
galleries that ran the length of the building. These contained electrical equipment,
control equipment, and pipes, as well as the plant workers themselves. Due to the
dangers, a closed-circuit television system was installed to allow equipment oper-
ators to see inside the canyons while remaining safe from radiation. Each facility
used a ventilation system that passed outside air into the personnel areas, and
then drew it into the processing area before it was filtered and sent out via a tall
stack [26].
2.2.1.2 REDOX Process
As discussed previously, the bismuth phosphate reprocessing method had some

significant disadvantages. During the 1940s, difficulties in obtaining uranium made
the loss of the uranium in the waste stream the main concern. In order to stop this
loss and recover the uranium, the US developed the Reduction and Oxidation
(REDOX) fuel separation process [31]. A pilot plant was tested at Oak Ridge
National Laboratory in 1948–1949, followed swiftly by a full scale reprocessing
facility built at Hanford in 1951 [30]. This facility, known as the “S Plant,” was
significantly less massive than its bismuth phosphate predecessors, measuring a
mere 470 feet long by 160 feet wide (Fig. 2.6). During its peak operating efficiency,
the S Plant could process up to twelve metric tons of spent fuel each day. Over the
course of its operational lifetime, the S Plant at Hanford processed roughly 24,000
metric tons of spent fuel. The facility was closed in 1967 [27].
The REDOX method was a critical step forward for the nuclear community in
the development of spent fuel reprocessing. Unlike with the bismuth phosphate
process, the REDOX technique was a countercurrent, continuous flow solvent
extraction process. This eliminated the need to process the spent fuel in batches.
Further, the REDOX process produced uranium and plutonium as products. This
was accomplished by first removing the reactor fuel cladding in a nitric acid, similar
to the bismuth phosphate process. The aqueous solution from this step was then
mixed with an organic solvent called Hexone (methyl isobutyl ketone), which is
immiscible with water. During the mixing, the uranium and plutonium nitrates
Fig. 2.6 REDOX Plant in Hanford, Washington [27]

54 2 Radioisotopes
would get extracted from the aqueous solution and transferred into the Hexone. Due
to its immiscibility, the Hexone could be easily separated from the aqueous solu-
tion. The uranium and plutonium were then chemically reduced and separated.
These products were then concentrated and sent to other facilities to be refined and
used [26].
In comparison to the bismuth phosphate process, the REDOX technique defi-
nitely had both advantages and disadvantages. One obvious advantage of the
REDOX method over the bismuth phosphate process was the ability to extract
uranium. Another advantage was the consolidation in the physical plant size
required while simultaneously increasing the number of plutonium processing
operations that could be done inside the facility. Unfortunately, the REDOX pro-
cess had some serious disadvantages as well. Waste from the plant was far less than
ideal: While it was smaller in volume than the waste produced by previous
methods, it contained more chemicals and was substantially hotter when it was
discharged. Also of major concern was the use of Hexone, as it is volatile and
explosive [27]. At only 69 °F, this organic solvent reaches its flash point—the
lowest temperature in which the liquid can evaporate enough fluid into the sur-
rounding air to create an ignitable gas. Such a low flash point necessitated the entire
REDOX reprocessing operation be conducted in an inert atmosphere [31].
2.2.1.3 PUREX Process
Another technique to separate specific radioisotopes from spent fuel is called the
Plutonium Uranium Recovery by Extraction (PUREX) process. The PUREX pro-
cess was developed by Knolls Atomic Power Laboratory in Schenectady, New
York during the early 1950s. Part of the driving force for PUREX development was
that the reprocessing technique was significantly safer than the REDOX process: As
it didn’t incorporate Hexone, there was no need for an inert atmosphere or for
concern about a Hexone based explosion. After development, the method was
tested at Oak Ridge National Laboratory in Tennessee from 1950–1952 [30]. It was
then adopted by the Savannah River Site in South Carolina. During November of
1954, the first PUREX reprocessing facility was opened at Savannah River in order
to recover the plutonium that had been residing in the spent fuel from five reactors
on site [26]. By January 1956, Hanford had constructed their own PUREX
reprocessing facility and starting operations (see Fig. 2.7). At over 1000 feet long,
60 feet wide, and 100 feet high, the Hanford PUREX reprocessing facility was the
largest chemical processing plant on the site [32].
While similar to the REDOX method in many respects, the PUREX process
allowed the countercurrent, continuous flow solvent extraction process to be con-
ducted under a considerably larger amount of safety. It also allowed additional
radioisotopes to be extracted. Just like with the REDOX technique, the first step in
the PUREX process involved removing the reactor fuel cladding through the use of
a nitric acid. The aqueous solution was then mixed with the organic solvent Tributyl
phosphate (TBP) [33]. Uranium, plutonium, and other desired radioisotopes (such
Fig. 2.7 PUREX Plant in Hanford, Washington [27]. The PUREX “canyon” was over 1000 feet
long, 60 feet wide, and 100 feet high [32]
as neptunium) got extracted from the aqueous solution and transferred into the TBP
through valence state manipulation. These products would then be concentrated in
another organic solvent such as paraffin, where they would be subjected to a
chemical scrubbing via dilute nitric acid. Both the solvent extraction and scrubbing
steps would be repeated multiple times in order to separate and purify the products.
Each final product would be in the form of a nitrate [31]. A diagram of the PUREX
process is shown in Fig. 2.8.
For the nuclear community, the PUREX process definitely proved itself to be
superior to the REDOX technique in almost all respects. As was previously dis-
cussed, the PUREX process didn’t incorporate Hexone, which required an inert
atmosphere inside the facility and created concern about a Hexone based explosion.
The organic solvent used instead, known as TBP, has a much higher flash point
(almost 300 °F). This made it possible to use a natural atmosphere with no
explosion concerns. Another factor in favor of PUREX was its capability to recover
multiple isotopes. Lastly, as might be expected from the previously mentioned
benefits, the PUREX process had an higher efficiency than earlier methods [26].
Even though the PUREX process was a significant improvement over the
REDOX method, the plutonium product produced with the new technique created a
serious concern. Due to its central role in the making of nuclear weapons, the
plutonium product quickly became a nuclear weapons proliferation issue. In 1976,
President Gerald Ford indefinitely suspended the commercial reprocessing and
recycling of plutonium [34]. This directly affected the PUREX method. By the end
of April of 1977, President Jimmy Carter banned all civilian reactor nuclear fuel
reprocessing in the United States [35].
2.2.1.4 Other Processes
Fortunately, while civil spent fuel reprocessing in the United States was terminated
after the creation of the PUREX process, development in this area has continued.
56 2 Radioisotopes
Fig. 2.8 The PUREX Process [26]
This has resulted in numerous different chemical reprocessing methods and their
variations. For example, one of the variations to the PUREX process is the TRUEX
reprocessing method, which attempts to isolate transuranic elements.
Another PUREX reprocessing variation is the DIAMEX-SANEX process, which
removes long-lived radioisotopes (like americium and curium) and then separates
them into different products. The COEX process extracts a mix containing both
uranium and plutonium, as well as a pure uranium stream. Lastly, the GANEX
technique extracts a mix containing both uranium and plutonium, similar to COEX,
except it also has the ability to separate some of the actinides. Many more examples
can be given, however the main concept and issues associated with chemical spent
fuel reprocessing remain.
In an attempt to avoid the complications associated with chemical spent fuel
reprocessing, some organizations are investigating reprocessing alternatives. One
such method worth mentioning is pyroprocessing, currently being investigated by
Argonne National Laboratories in Chicago. Unlike other methods, this technique
utilizes a process called electrorefining to separate radioisotopes. Similar to
Fig. 2.9 Pyroprocessing Schematic [36]
electroplating, electrorefining involves attaching the spent fuel, already processed

into a metallic form, to an anode suspended in a molten salt bath. An electric current
is applied across the bath, and the fuel elements dissolve. Recyclable radioisotopes,
such as uranium and other actinides, collect on the cathode. These elements are then
sent to the cathode processor for further refinement, where any remaining salt is
returned to the bath. A schematic of the entire process is shown in Fig. 2.9.
2.2.2 Separation from Natural Decay Chains
Another method of radioisotope production involves taking advantage of the three

radioisotope decay chains found naturally in the environment. These three decay
chains are called the uranium, actinium, and thorium series, and have initial iso-
topes consisting of uranium-238, uranium-235, and thorium-232, respectively.
Usually, these decay chains are approximated as being in “secular equilibrium,” a
state in which the activity of each radionuclide in the series is approximately equal.
To satisfy the criteria for this definition, each decay chain must have an initial
radioisotope with a much longer half-life than any of its following decay products.
Additionally, a long time period must have elapsed to allow for the growth of the
subsequent radioisotopes. A good approximation for this span is about ten half-lives
of the longest living decay product [37].
During each decay process the original radioisotope, known as the “parent,”
decays through either alpha or beta particle emission. With the decay particle
ejected, the remaining radioisotope is of the form of the new nuclide, known as the
“daughter,” This daughter product is more stable than the parent, although it can
still decay further. A few of these daughter isotopes are strong gamma emitters. In
the context of decay chains, however, gamma emission is not in the same category
as an alpha or beta emission. This is because it is the release of excess energy, and
by itself does not transform the radionuclide it was emitted from [37].
58 2 Radioisotopes
As was mentioned previously, the first decay chain, known as the uranium
series, has an initial parent consisting of U-238. This radioisotope has a half-life of
4.5 billion years. Its decay daughter is Th-234, which is achieved through the
emission of an alpha particle. The Th-234 has a half-life of 24 days, during which it
emits a beta particle and transforms into Pa-243 m. Pa-243 m is a metastable state,
with a half-life less than 1.25 min (75 s). As a parent, Pa-243 m emits a beta,
returning the nuclide to an isotope of uranium (U-234). The nuclide then undergoes
five consecutive alpha decays. These decays consist of U-234 (with a half-life of
240,000 years) to Th-230, Th-230 (with a half-life of 77,000 years) to Ra-226,
Ra-226 (with a half-life of 1,600 years) to Rn-222, Rn-222 (with a half-life of
3.8 days) to Po-218, and Po-218 (with a half-life of 3.1 min) to Pb-214. This
radioisotope has a half-life of 27 min and has the decay daughter Bi-214, achieved
by beta emission. The bismuth nuclide, having a half-life of 20 min, quickly decays
via another beta particle, transforming itself into Po-214. Due to an exceedingly
tiny half-life of only 160 ls, the Po-214 decays almost instantaneously to Pb-210
through the emission of an alpha particle. Over its 22 year half-life, Pb-210 emits a
beta particle, converting the nuclide into Bi-210. Bi-210, with a 5 day half-life,
decays via beta emission to Po-210. Lastly, Po-210 emits an alpha particle during
its 140 day half-life, transforming into Pb-206. Pb-206 is stable and does not decay
further. This entire sequence can be seen in Fig. 2.10.
The actinium series, which is the second decay chain, is shown in Fig. 2.11. Its
initial parent is U-235. U-235 is a long-lived isotope, with a half-life of 700 million
years. Its decay daughter is Th-231, which is reached through the emission of an
alpha particle. The Th-231 has a half-life of 26 h, during which it emits a beta
particle and transforms into Pa-231. Pa-231, with a 33,000 year half-life, decays via
alpha emission to Ac-227. Interestingly, Ac-227 emits both alpha and beta particles
during its 22 year half-life, creating two separate decay paths. Approximately 99 %
of the time, Ac-227 decays via a beta particle, yielding Th-227. The Th-227 has a
half-life of 19 days, during which time an alpha particle is emitted, resulting in
Ra-223. In the other 1 % of the time, Ac-227 decays via an alpha particle, yielding
Fr-223. Fr-223, with a 22 min half-life, emits a beta particle, also resulting in
Ra-223. The nuclide then undergoes three consecutive alpha decays. These decays
consist of Ra-223 (with a half-life of 11 days) to Rn-219, Rn-219 (with a half-life
of 4 s) to Po-215, and Po-215 (with a half-life of 1.8 ms) to Pb-211. This
radioisotope has a half-life of 36 min and has the decay daughter Bi-211, achieved
by beta emission. The bismuth nuclide, having a half-life of 2.1 min, quickly
decays via an alpha particle, transforming itself into Tl-207. Lastly, Tl-207 emits a
beta particle during its 4.8 min half-life, transforming into Pb-207. Pb-207 is stable
and does not decay further.
Last is the thorium series, which has an initial parent consisting of Th-232. This
radioisotope has a half-life of 14 billion years. Its decay daughter is Ra-228, which
is achieved through the emission of an alpha particle. The Ra-228 has a half-life of
5.8 years, during which it emits a beta and turns into Ac-228. Ac-228, with a
half-life of 6.1 h, releases a beta particle, returning the nuclide to an isotope of
thorium (Th-228). The nuclide then undergoes four consecutive alpha decays.
Fig. 2.10 The uranium decay series [37]
These decays consist of Th-228 (with a half-life of 1.9 years) to Ra-224, Ra-224
(with a half-life of 3.7 days) to Rn-320, Rn-320 (with a half-life of 56 s) to Po-216,
and Po-216 (with a half-life of 0.15 s) to Pb-212. This radioisotope has a half-life of
11 h and has the decay daughter Bi-212, achieved by beta emission. Similar to the
actinium in the actinium series, Bi-212 emits both alpha and beta particles during its
61 min half-life, creating two separate decay paths. Roughly 64 % of the time,
Bi-212 yields Po-212 through beta particle decay. The Po-212 has an almost
instantaneous half-life of only 310 ns, during which time an alpha particle is
emitted, resulting in Pb-208. In the other 36 % of the time, Bi-212 emits an alpha
60 2 Radioisotopes
Fig. 2.11 The actinium decay series [37]
particle, giving Tl-208. Tl-208, with a half-life of 3.1 min, emits a beta particle, also
resulting in Pb-208. Pb-208 is stable and does not decay further. A diagram of the
thorium decay series is shown in Fig. 2.12.
One factor that can significantly disrupt the secular equilibrium of uranium,
actinium, and thorium decay chains is whether or not specific radionuclides have
been extracted from them. If extraction of a particular isotope has occurred to a
considerable extent, its absence will inhibit the formation of its daughter isotopes,
effectively severing the decay chain. The remaining daughter isotope sub-series can
then be considered a separate (and smaller) series, with one of the long-lived
radioisotopes from the original decay chain as its initial parent. For example, if
U-238 was extracted from the uranium series in Fig. 2.10, the immediate daughter
Fig. 2.12 The thorium decay series [37]
product (Th-234) would not continue to form. With relatively short half-lives, both
the remaining Th-234 and its daughter Pa-243m would decay out. This would leave
the long-lived radioisotope U-234 as the initial parent for a smaller chain, consisting
of the subsequent daughters of U-234 originally included in the uranium series.
Similarly, if U-234 was extracted from the uranium series, the Th-230 daughter
would cease forming. However, since Th-230 itself is a long-lived radioisotope, it
would become the initial parent for the remaining isotopes in the decay chain. Other
nuclides in the uranium series that can be considered initial parents for smaller
chains are Ra-226 and Pb-210 [37].
In determining the initial parent for a smaller series, nuclides with half-lives of
less than a year are usually not considered. This is because these radioisotopes are
62 2 Radioisotopes
heavily dependent on the isotope supply generated by their parent nuclide, and so
reestablish equilibrium with their parent (decay out) within several years [37].
Applying this consideration to the actinium series, it can be determined that it
contains the initial parents U-235, Pa-231, and Ac-227. Similarly, applying this
consideration to the thorium series reveals the initial parents Th-232, Ra-228, and
Th-228.
Once an isotope from one of the natural decay chains has been selected for
extraction, a chemical separation process is applied to remove the nuclide from the
other radioisotopes in the corresponding decay series and manipulate it into a useful
form. The exact details of this process depend on the properties of the radioisotope
being separated. In the case of uranium, crushed uranium ore (UO3) is leached with
sulfuric acid (H2SO4). The resulting oxidation reaction is [38]:
UO3 þ 2H þ ! UO22 þ þ H2 O ð2:1Þ
UO22 þ þ 3SO2 4
4 ! UO2 ðSO4 Þ3 ð2:2Þ
The uranium solution is then recovered using a resin/polymer ion exchange or

liquid ion exchange solvent extraction system. To continue processing, the uranium
is stripped from the ion exchange resin/polymer with either a strong acid or chloride
solution, or with a nitrate solution in a semi-continuous cycle. In the chemical
equation below, an acid is used. The “R” is an alkyl (hydrocarbon) grouping with a
single covalent bond [38]:
2R3 N þ H2 SO4 ! ðR3 NHÞ2 SO4 ð2:3Þ
2ðR3 NHÞ2 SO4 þ 2UO2 ðSO4 Þ4

3 ! ðR3 NHÞ4 UO2 ðSO4 Þ3 þ 2SO4
2
ð2:4Þ
After the previous reaction, any cation impurities are removed from the solution
by the use of sulfuric acid, and anion impurities are removed from the solution by
utilizing gaseous ammonia. An ammonium sulfate solution is then introduced,
removing the alkyl (hydrocarbon) groupings from the uranium. Once this is com-
pleted, gaseous ammonia (2NH3) is added to the solution. This is done to neutralize
the solution and create ammonium diuranate [38]:
ðR3 NHÞ4 UO2 ðSO4 Þ3 þ 2ðNH4 Þ2 SO4 ! 4RN þ ðNH4 Þ4 UO2 ðSO4 Þ3 þ 2H2 SO4
ð2:5Þ
2NH3 þ UO2 ðSO4 Þ4

3 ! ðNH4 Þ2 U2 O7 þ 4SO4
2
ð2:6Þ
Lastly, the ammonium diuranate is processed. Any remaining water is removed

from the solution, and then the ammonium diuranate is then dried at a high tem-
perature [38]. This converts the product to uranium oxide (U3O8), which contains
approximately 85 % uranium by mass. U3O8, more commonly known as yellow-
cake uranium, is then sold to customers.
2.2.3 Production by Neutron Capture in a Reactor
A third method, which is commonly used today, is the production of radioisotopes

through neutron capture in a nuclear reactor. There are still a few national labs in
the United States with this capability. The two current typical methods of neutron
exposure to a sample are beam port exposure and pneumatic tube. It is also possible
to have a loading dock which remains a stationary area during operation above the
reactor, such as demonstrated in the Oak Ridge National Laboratory diagram
(Fig. 2.13).
Each research reactor has very different properties associated with the key fac-
tors in irradiation (Table 2.7). These key factors include neutron flux, operation
time, and volume available for sample exposure. Also important to note is that extra
“dead time” occurs due to sample handling procedures and sample loading. This
Fig. 2.13 Oakridge National Lab diagram of Low-Intensity Testing Reactor. The horizontal beam
holes connect to beam ports and the large facilities near the active lattice show closer exposure
areas
64 2 Radioisotopes
Table 2.7 Properties of reactor facilities by laboratory

Reactor Fast flux Thermal flux Operational time Max volume
name (n cm−2 s−1) (n cm−2 s−1) (d year−1) (cm−3)
MIT—**RR 1.2 1014 6.0 1013 168 460
MURR 6.0 1014 6.0 1014 180 350
INL **ATR 5.0 10 14
1.0 10 15
180–195 15,400*
ONRL 5.0 1014 1.5 1014 270–290 14.5 (f) 3.2 (th)
Reactors
*Note that INL has an operation port that has a tremendous available volume relative to pneumatic
tubes present in other operations. **RR is the abbreviation of Research Reactor, as MURR is
University of Missouri Research Reactor and ATR is Advanced Test Reactor [39–41]
dead time can extend up to 3 or 4 times the exposure time for pneumatic tubes or be
as low as 20 % for beam ports/staging areas.
The operational time above is gross of down time, the actual transitioning of
material between runs. Thus, effective exposure could be as low as 5 % of the
reactor operating time or as high as 70 %. Future calculations will involve an
assumption of 50 %.
2.2.4 Production by Accelerator
The last method of radioisotope production utilizes a particle accelerator. In these

accelerators, beams of charged particles (such as protons) are accelerated by an
external magnetic field (see Fig. 2.14). When the appropriate energy level has been
reached, these particles are then deliberately collided with target isotope atoms.
During the collision, the atom absorbs the particle, transforming itself into a new
isotope while releasing subatomic particles. After sufficient irradiation, the target
isotope can be removed from the accelerator, and the desired radionuclide can be
processed out for use.
One example of particle accelerator based radioisotope production is the
Brookhaven Linac (linear accelerator) Isotope Producer, otherwise known as BLIP.
This accelerator is located at Brookhaven National Laboratory (BNL) in Upton,
New York, and operates by accelerating protons. Since opening in 1972, BLIP has
received two upgrades: One in 1986 and the second in 1996 [43]. As shown in
Fig. 2.15, BLIP works by diverting some of the particle beam produced by a
previous linac. The length of BLIP from the diverging point from the previous linac
to the target at the end is 98.4 feet. Target isotopes are formed into round “discs”
and are placed inside a vertical 16 inch diameter containment shaft that is filled with
water. They are then lowered down approximately 30 feet until they are in line with
the incoming particle beam from the linac. Depending on particle beam strength and
disc thickness, up to eight target isotopes can be inserted at a time [43]. Currently,
BLIP can utilize proton beams with energies of up to 200 MeV, as well as up to 110
µA intensity [44].
Fig. 2.14 A linear accelerator (LINAC) at Brookhaven National Laboratory [42]
Fig. 2.15 BLIP Beam Line Schematic [42]

66 2 Radioisotopes
Using BLIP, Brookhaven National Laboratory has the capability to produce

several radioisotopes. This is illustrated in Table 2.8, which shows the radioisotope
production capability for the year 2012. As can be seen from the table, obtaining
certain isotopes (such as Be-7 or Fe-52) required placing a special order with BNL.
Other radionuclides, such as Ni-63, were readily in stock. Some isotopes were
produced monthly (such as Cu-67 and Y-86), while others (such as As-73 and
Tc-95 m) were not scheduled to be produced that year. One interesting note is that,
for the 2012 radioisotope production year, BLIP was only run from January through
July [44]. This means that the longer lived nuclides were available all year, while
the supply of shorter lived nuclides quickly died out after BLIP was shut down for
the year.
After the desired radioisotopes are produced, they still must be separated from
any coproduced isotopes, as well as other target materials. Fortunately, BNL has
nine hot cells, eight radiochemistry development laboratories, an instrumentation
laboratory for radionuclide and chemical assays, and radioactive waste handling
and storage facilities that all enable it to do the required isotope post-processing
[43]. In these facilities, BNL employs several radiochemical processing techniques.
These methods include ion exchange chromatography, solvent extraction, distilla-
tion, and precipitation [43].
A second example of radioisotope production via a particle accelerator is
Niowave, a super conducting linear accelerator company located in Lansing,
Michigan. Unlike with BLIP (which uses protons), the Niowave super conducting
linac accelerates electrons. It operates at only 4 K (−452 °F) and can produce
Table 2.8 BLIP radioisotope Isotope Half-life Primary Production

production capability in 2012 usage frequency
[44]
Be-7 53.3 days Research Special order
Mg-28 20.9 h Research Special order
Fe-52 8.3 h Research Special order
Fe-55 2.73 years Commercial In stock
Ni-63 100.1 years Commercial In stock
Zn-65 243.8 days Commercial In stock
Cu-67 61.9 h Research Monthly
Ge-68 270.8 days Commercial Monthly
As-73 80.3 days Research Not scheduled
Sr-82 25.4 days Commercial Monthly
Rb-83 86.2 days Research In stock
Y-86 14.7 days Research Monthly
Y-88 106.6 days Commercial Not scheduled
Tc-95 m 61 days Research Not scheduled
Tc-96 4.3 days Research Not scheduled
electrons from 20 to 40 MeV, with the particle beam power ranging from 50 to
100 kW in intensity [45]. In 2014, construction began for a linac facility specifically
designed for radioisotope production. When completed in 2015, Niowave intends to
pursue production of medical radionuclides Mo-99, Cu-67, and Ac-225. It also plans
on investigating production of industrial isotopes Mn-54 and Zn-65 [45].
2.3 Cost of Radioisotopes
There are many factors that impact the cost of radioisotope production discussed in
this section. The Nuclear Energy Agency OECD 2010 report reviewed production
costs of Mo-99/and Tc-99 m. Encompassing all factors can be dissolved down into
a per unit cost by the equation:

It þ O&M þ Dt
Rt ð1 þ rÞt
Pa ¼ ð2:7Þ
at
Rt ð1 þ rÞt
In this notation, a, is the particular radioisotope to be produced; ‘t’ is a set

particular time, and is summed over all effective times; ‘I’ is the investment costs,
typically taken at t = 0 but will include interest payments on loans; D is the
decommissioning cost of any facility; O&M the operating and maintenance costs at
a given time; and ‘r’ is the interest rate (to be considered constant, but it also can be
dynamic in time). The following are key factors influencing these costs from a per
unit basis. Production rates and particular stoichiometry for chemical/nuclear pro-
cesses are proprietary to each industry, and thus cannot be presented in a reliable
fashion. Instead, this section will conclude with previous production costs as a
reference point for future work.
2.3.1 Cost of Separation
Separation steps exist in every method of source material creation for nuclear
batteries. In neutron capture processes, source material is purified typically using an
acid and a liquid-liquid extraction step [46]. Accelerator targets are processed in a
similar manner. Of particular interest though would be the anticipated cost of a
liquid-liquid and liquid-solid extraction processing plants. These are typical designs
similar to expected equipment used to harvest material from spent nuclear fuel and
convert it to source material.
Separation costs are the composite of capital equipment necessary and pro-
cessing materials needed. The primary processes involved are: medium based
68 2 Radioisotopes
liquid-liquid extraction, centrifuge separation, dissolving and mixing, distillation,

and drying. While the discussion present is a cost guide to extraction of radioactive
material from spent nuclear fuel, pieces of the overall process can be extracted to
specific methods such as PUREX.
2.3.1.1 Cost Analysis
The first consideration in cost analysis is consultation of reactants and recyclable

chemicals. While exact stoichiometry and chemical engineering principles are pro-
prietary to companies such as AREVA and their operations in France, by studying
operations at the La Hague facility or the US nuclear weapons reprocessing facility
at Hanford, WA, one gathers typical chemical reactants used (Table 2.9).
Typically, there are multiple physical modifications steps needed to achieve
desired conditions for chemical reactions, thus the discussion begins with
Dissolution, Mixing, and Drying Capital Costs.
2.3.1.2 Dissolution, Mixing and Drying Equipment
Dissolution, mixing, and drying are the bracket steps to major processes of distil-
lation, liquid-liquid extraction, and centrifuging. Typically, manufacturing of cus-
tom capital equipment takes place. But there are major items associated with these
processes; equipment such as mixers, dryers, containers, and extraction columns.
A facility may need multiple iterations of similar equipment, which cuts down on
manufacturing costs (Table 2.10).
Table 2.9 Typical process materials cost for PUREX

Reactant name Quantity Unit Costs ($)
Nitric acid 15 L 382.5
Sodium carbonate 12 kg 492.7
H3PO4—85 % wt 25 kg 234
Sodium bismuthate 100 g 276.5
Sodium dichromate 5 kg 556
Zirconium 100 g 82.7
Cerium in ammonium cerium nitrate 500 g 267.5
Hexone (methyl isobutyl ketone) *98 % 100 ml 30
Sodium hydroxide *97 % 50 kg 971
Tributyl phosphate—TBP 500 ml 174
Kerosene 3.78541 L 3
Gallium 1 kg 339
Note these prices were sourced to be minimums from multiple online and retail sources. They are
listed by reactant name, the quantity and unit for the associated cost [47–50]
2.3 Cost of Radioisotopes 69
Table 2.10 The following are industrial quotes listed as averages from major producers and users
of these industrial equipment (Flour, Sulzer, URS)
Device Cost estimate ($k)
Industrial dryer 5.00
Packed extraction columns (ECP) 1,000.00
Mixer settler column 2,500.00
Organic solvent nanofiltration 3,000.00
2.3.1.3 Sunk Costs Considerations
Referring to the equation at the start of the section it is important to note that
investment costs must be considered at time zero. There are instances where these
costs can get out of control, resulting in the death of a project. Additionally, in the
early studies of nuclear fuel separation the following situation occurred resulting in
the destruction of the facility and need for a full redesign.
Nuclear fuel separation occurs via the metal being dissolved in a significant high
concentration of acid solutions in preparation for liquid-liquid extraction. Slight
oxidation of the metal can occur in long periods of storage. Direct mixing with the
TBP had resulted in disaster when the storage container was stainless steel. Via
these disasters it is now known that a polymer lining is required in the storage
vessel before the combination step. It prevents the violent redux reaction of any
oxidized metal in the acid interacting with “TPB,” and removes the possibility of
explosive gas generation during storage [51].
2.3.2 Cost of Neutron Capture
Projecting the cost of neutron capture is a multi-step process, but can be simplified
by dividing the total annual operating and fuel costs by the potential amount of
neutrons captured by a given reference material.
C Cop þ Cfuel
¼ ð2:8Þ
n n
Note the above formula requires costs to be taken on the same time basis. This
investigation takes into account annual operational and fuel costs. Referencing one
of the authors, research paper, typical fuel cost for research reactors is roughly
$20 M per year. Additionally, a typical reactor will need a minimum of 50 personnel
for operations and primary safety precautions. The median salary in the industry is
roughly $70,000 while the average is closer to $100,000; the average will be used
for conservatism in the estimate [52, 53]. Thus in a year a typical research reactor
would cost roughly $25 M annually for operations and fuel. Typically neutrons
available are given in terms of available flux u, so total neutrons:
70 2 Radioisotopes
Table 2.11 Note that is table represents cost per flux and based on parameters given in
Table 2.10
Neutron costs Fast (10−14/(n cm−2) Thermal (10−14/(n cm−2)
MIT-RR $2.87 $5.74
MURR $6.43 $0.53
INL-ATR $0.64 $0.32
ONRL-Reactors $1.44 $0.65
n0 ¼ u Dt ð2:9Þ
Note that Δt is exposure time of the material during operations. Assuming

downtime of 50 % the cost estimates from published fluxes of test reactors around
the country is shown in Table 2.11.
The number of neutrons though should be more formally analyzed as neutrons
are captured. This requires significant number of assumptions and potentially some
complexity. First, the neutrons need to be accounted for in terms of thermal and fast
neutrons, which have different capture cross sections based on target materials.
Second, one is required to take into account decay constants, exposure time, capture
cross section, and other physical parameters. Given the complexity, it is easier to
find the cost assuming some average parameters for materials and saturation level is
reached, removing the specifics of decay effects. To be conservative the decay
effects can be coupled with associated downtime of experiments and exposure.
1
ncap ¼ uf ;prod rf þ uth;prod rth q V Dt ð2:10Þ
2
Note that sometimes the chambers available for thermal and fast production have
different volumes, V; where that factor would be distributed. Also density, q, is
taken into account, but this equation has the capture cross section in per gram.
Typically cross sections are listed on a per atom basis, so multiplying by
Avagadro’s number NA and dividing by molecular weight should be multiplied into
the factors. Doing so results in a cost per neutron captured for the test reactors.
Note significant assumptions about the density and capture cross section of
materials are taken into account. The table assumes a material with a density of
8 g cm−3, a molecular weight of 80 g mol−1 and fast cross section of 1 barn
(10−24 cm2 atom−1) and thermal cross section of 5 barns.
Unfortunately, most production methods for nuclear batteries would like mate-
rial on the order of grams and not atoms. This requires to multiply the cost by the
amount of neutron captures needed to produce a gram of radioactive material,
which is on the order of 1023. Suddenly, the costs above reach the 100s–10s of
millions per gram of source material produced, from a fuel cost basis. For mass
production, beam port operations will be required and only the lower costs pre-
sented would be applicable. INL-ATR thermal and MURR thermal neutron pro-
duction are based on beam port size calculations (Table 2.12).
Table 2.12 Table is cost per neutron captured, assuming neutrons captured independently of
other operations, and a factor of 2 is divided to conserve this operation
Neutron costs Fast 10−17/n Thermal 10−17/n
MIT-RR $103.68 $41.47
MURR $232.25 $3.87
INL-ATR $23.23 $2.32
ONRL-Reactors $51.85 $4.71
As an example of how to go about calculating the cost of isotope production

using neutron capture consider the case of Ni-63. It has a half-life of 101 years and
emits a beta particle with a maximum energy of 65.9 keV. The long half-life makes
it promising for nuclear power battery applications—where long shelf-life is
desired. Nickel-63 is produced via neutron capture of Nickel-62 (63Ni (n,c) 63Ni).
The isotopic composition and neutron cross sections of natural nickel are shown in
Table 2.13. The large neutron thermal absorption cross section (14.5 barns) as well
as the resonance integral (6.6 barns) of Ni-62 increases the likelihood for neutron
absorption to occur. A large fraction of natural nickel is Ni-58. To increase the
specific activity of Ni-63, Ni-62 is enriched as high as possible and is irradiated
with a high neutron flux.
The activity of the radioisotope produced during or at the end of a neutron
capture reaction is calculated from the equation below
dN 0
¼ nvract NT kN 0 ð2:11Þ
dt
where NT is the total number of atoms present in the target, nv is the neutron flux =
u, ract is the activation cross-section, N 0 is the number of activated atoms, A atomic
weight of the target, and kN 0 is the decay rate of the product.
The activity can be obtained by integrating the equation above resulting in the
equation below.

A ¼ kN ¼ ract uNT 1 ekt ð2:12Þ
Table 2.13 Percentage composition of natural nickel isotope

Isotope Ni-58 Ni-60 Ni-61 Ni-62 Ni-64
% abundance 68.07 26.22 1.14 3.63 0.93
Thermal neutron cross section (barns) 4.6 2.96 2.5 14.5 1.52
Resonance Integral (barns) 2.2 1.5 1.5 6.6 0.98
72 2 Radioisotopes
The above equation does not take into account the following which will reduce
the calculated activity:
• Flux depression due to adjacent samples in the reactor especially when such
samples are high neutron absorbers,
• Burn-up of the target material with time,
• Destruction of the product nucleus due to subsequent neutron capture,
• Self-shielding effect of the target,
• Power variation in the reactor, and
• Target encapsulation.
A final time. t0 will elapse after irradiation and just before counting. Hence,
Eq. 2.12 reduces to
0
A ¼ ract uNT 1 ekt ekt ð2:13Þ
The quantity (1 − e−kt) is called the saturation term. The equation shows that the
activity grows exponentially and reaches a saturated value limited by the flux in the
reactor. After sufficient period of irradiation (t t1/2) the activity approaches the
value
Asat ¼ ract uNT ð2:14Þ
ASat is called the saturation activity. For relatively short irradiation time com-
parable to the half-life of the isotope, the saturated term approaches kt and the
activity varies linearly with time.
The neutron flux at the flux trap at the University of Missouri Research Reactor
(MURR) is 6.0 1014 n cm−2 s−1. There are many positions within beam
port of this reactor that are designated for experimental work. Targets could easily
be inserted in and out of this positions during reactor operation. The flux at the H1
position is 1.0 1014 n cm−2 s−1 [54]. Table 2.14 shows the nuclear reactions and
calculated specific activity for Ni-63. To minimize the content of Ni-58 in the
target, the target is made of mostly Ni-62. The small fraction of Ni-58 in the target
would burn in the reactor to produce a small percentage of Ni-59. Figure 2.16
shows the specific activity, defined as the activity per unit mass, for the production
of Ni-63 at the H1 position at different irradiation times. The maximum specific
Table 2.14 Nuclear reactions and specific activity calculation for Nickel-63 at different
irradiation times
Reaction Isotope abundance Cross section (barns) Activity at
1.0 1014 n cm−2 s−1
(mCi/g)
1 week 1 month 1 year
62
Ni (n, c) 63
Ni 3.63 % 14.5 1.98 7.92 94.7
Fig. 2.16 Calculated production yields of Nickel-63 by neutron irradiation of Nickel-62 targets as
a function of irradiation time at a thermal neutron flux of 1:0 1014 n cm−2 s−1
activity achieved is 13.7 Ci/g, obtained after a few half-lives. Due to the long
half-life of Nickel-63, the amount of Ni-63 produced that decays before counting is
negligible.
Secondary reactions and products are: (1) 58Ni (n, c) 59Ni [t1/2 =7.6104 years,
abundance = 68.08%, r = 4.6 barns], (2) 58Ni (n,a) 55Fe [t1/2 = 2.73 years r 1
mb], (3) 62Ni (n,a) 59Fe [t1/2 = 2.73 years, r = 0.002 mb], and (4) 64Ni (n,c) 65Ni
[t1/2 = 2.51 h, abundance = 0.9265, r = 1.52 b]
Ni-63 also has an appreciable cross section (24 barns) and is activated in reactors
to produce Ni-64. This limits the specific activity of the Ni-63 produced. Generally,
10 g of Nickel-62 powder is encapsulated in an aluminum container with diameter
of 4 mm and a height of 2 mm (50.27 mm2 area) prior to irradiation. Irradiation is
done between 8 and 12 weeks [46]. The cost per neutron at MURR has been
indicated Table 2.12. Based on this value, the cost of producing 1 Ci/g of Ni-63 is
estimated to be US$4252. The cost of neutrons generally accounts for a small
fraction of the total production cost.
2.3.3 Cost of Accelerator
Accelerators are used to produce medical isotopes. Medical isotopes are necessary
for health related diagnostic imaging and therapy. For these medical related
applications, isotopes can be priced at a premium. The diagnostic imaging appli-
cation typically uses short lived isotopes. The therapeutic isotopes are longer lived.
Medical isotopes are created by charged particle reactions (protons, deuterons,
helium or neutrons) or a by a photonuclear process with a target material.
74 2 Radioisotopes
Table 2.15 Cost of Cost per Ci Cost per gram

manufacturing Mo-99 using
various processes Reactor $1,010.00 $484,000,000
Photo-nuclear (enriched target) $356.67 $171,000,000
Photo-nuclear (natural target) $2,893.33 1,387,000,000
Dedicated cyclotron $1,216.67 583,000,000
Multipurpose cyclotron $766.67 367,400,000
To illustrate the costs of producing medical isotopes, one of the most studied,
Mo-99, is used as an example. Mo-99 is one of the more prominent medical
isotopes and has been widely studied [55, 56]. The processes studied included the
use of low enriched uranium in a high flux nuclear reactor, the use of Mo-100(c,n)
Mo-99 reaction using an accelerator, or a neutron flux of 1014 n cm−2 s−1 provided
either by a multipurpose 30 MeV cyclotron or a dedicated cyclotron were costed
out as shown in Table 2.15.
The high cost of producing a medical radioisotope is acceptable for medical
procedures where 10s of mCi of the isotope are needed and the cost premium is
worth it for the lifesaving benefits. However, these types of costs are not affordable
for nuclear batteries.
New technologies are being developed based on linear accelerators [57]. There
are some developments in superconducting linear accelerators. These new
approaches and new developments will not have a significant impact on price.
2.4 Other Factors Influencing Cost
2.4.1 Safety
The costs associated with safety are discussed in further detail in Sect. 5.3.1 of
Chap. 5 where regulation based on particular nuclides is discussed. Typically, there
will be resident experts whose salaries are in the range of $90,000–$300,000 based
on qualification and licenses attributed to that individual. Additionally, special
precautions such as transportation effects are required.
2.4.2 Software
It typically is important to consult the industry standards of modeling tools and

regulatory code. A leader in chemical engineering process software is Aspen
Technologies, and their license can range from $5,000 to $20,000 annually per user
depending on detail of specifications needed (quote from call). In the case of
nuclear operations, the consent and royalties to LANL, for use and modification of
MCNP can be in excess of $50,000, depending on size of the company.
2.4 Other Factors Influencing Cost 75
2.4.3 Liquidity of Capital (Cash)
A big factor recognized in the initial equation is the current interest rate, ‘r’.
Historically lending of this scale requires that significant interest costs are at a
significant premium to central bank rates. Currently, the world is seeing record low
interest rates, where the 10 year treasury is going for 1.7 %. The premium for a
capital project of this magnitude and risk would typically go for slightly higher rates
than that of owning a home, currently 4.3 %. Many companies will issue con-
vertible debt or preferred stock to fund these operations, and yields can be as costly
as 10 %. Additionally, should circumstances change or problems come up, it would
become necessary to borrow and refinance the loans. In a tight market, noticed in
2008, such costs could be as high as 17 %. Most industrial models will anticipate
between 6 and 8 % interest costs, but timing certainly plays an effect in project
costs [58].
2.5 Isotopes Produced from the Manhattan Project
The technologies and capabilities available for nuclear battery source material are
firmly established, but they are extremely costly, difficult to organize, and filled
with regulation. Due to the risks, there are virtually no profitable uses for nuclear
energy outside the scale of reactors or premium of medicine. During the beginning
of the nuclear era there were a lot of parts of the Manhattan Project and recycling at
Hannover, WA was one of those parts. The costs of nuclear fuel separation were a
part of the technological development and were assumed by the U.S. government as
a means to an end for producing weapons grade plutonium. While this progressed
the science, it did not provide good insight into creating a business model for
operating profitable sourcing of other radioisotopes [59].
2.6 Mixed Oxide Fuel Fabrication Facility (MOX FFF)
In 1999 the US government approved joint operation of a facility to reprocess spent

nuclear fuel at the Savannah River Site in South Carolina. After 16 years and
roughly $8.5B the project is still ‘underway’ as certain engineering failures, dis-
putes over regulations, and other things have impeded progress. The project has
documented over 20 million safe work hours and will be able to process 3.5 metric
tons of weapons grade plutonium into a mixed oxide fuel (MOX). MOX fuel
contains plutonium and uranium and can be put into many typical reactor designs
used today, with the only major change being timing of delayed neutrons. The
76 2 Radioisotopes
facility will have the capability to produce over 150 MOX assemblies a year, and
anticipated to finally be online in 2020. All in costs, not including operation will be
on the order of $17B, but was initially projected at $2-5B [59–61].
2.7 Summary
There are limited sources of radioisotopes. Isotopes can come from natural sources
and from man-made sources (fission and through other nuclear reactions based on
neutron capture, high energy interactions with charged particles or photonuclear).
The issue with radioisotopes is the quantity of suitable isotopes that are available
from natural sources or fission. These types of radioisotopes have to be separated
from other materials which contribute to the cost of the isotopes. If the isotope must
be made from nuclear reactions, then the cost of the neutron source, high energy
charged particle source or high energy photon source becomes a significant factor.
Availability and cost of isotopes are a problem for the development of nuclear
batteries.
Problems
1. How long would it take for the world’s supply of natural Th-234 to reach one
gram (Note it is being replenished by a parent isotope)? How about Ac-227?
Ra-228?
2. What would be the parent isotopes for sub-series created by extracting the
following nuclides: U-238, Th-230, Ra-226, Pa-231, Ac-227, Ra-228, and
Th-228?
3. Suppose a sub-series with a parent isotope of U-234 was discovered. What
radioisotope would likely have been extracted? What if the parent isotope was
Th-228?
4. Which radioisotopes mentioned in this chapter could be used in a nuclear
battery?
5. Is Pb-210 a viable isotope for nuclear batteries? Explain your reasoning.
6. Is it feasible to make Pu-238 from the existing inventory of Np-237? Explain
your reasoning.
7. Is Am-241 a possible replacement for Pu-238 in a Radioisotope Thermoelectric
Generator?
8. How much Kr-85 does a 3 GW thermal commercial reactor with a 90 %
capacity factor produce in one year?
9. Estimate how much it costs to produce a gram of Ar-39.
10. Estimate how much it costs to produce a gram of Gd-148.
References 77
References
1. Masarik J (2009) Chapter 1 Origin and Distribution of Radionuclides in the Continental

Environment. In: Klaus F (ed) Radioactivity in the environment, vol 16. Elsevier, pp 1–25
2. Asimov I (1953) Naturally occurring radioisotopes. J Chem Educ 30:398, 1953/08/01
3. Dalrymple GB (2001) The age of the Earth in the twentieth century: a problem (mostly)
solved. Geol Soc Lond Spec Publ 190:205–221
review of nuclear batteries. Prog Nucl Energ 75:117–148
5. Vance R (2014) Uranium 2014: resources, production and demand. NEA News 26
6. Brookhaven National Laboratory (11/09/2015) Interactive Chart of the Nuclides. http://www.
nndc.bnl.gov/chart/
7. Haynes WM, Lide DR, Bruno TJ, CRC handbook of chemistry and physics: a ready-reference
book of chemical and physical data, 96th edn
8. Baum EM, Ernesti MC, Knox HD, Miller TR, Watson AM (2010) Knolls Atomic Power
Laboratory. Nuclides and isotopes: chart of the nuclides, 17th edn. KAPL: BECHTEL
9. O’Brien K (1979) Secular variations in the production of cosmogenic isotopes in the Earth’s
atmosphere. J Geophys Res: Space Phys (1978–2012) 84:423–431
10. Glasstone S, Sesonske A (1994) Nuclear reactor engineering, 4th edn. Chapman & Hall, New
York, NY
11. DOE-HDBK-1019/1-93 (1993) Nuclear physics and reactor theory. Department of Energy,
Washington DC
12. Murray RL (1981) Understanding radioactive waste PNL-3570; Other: ON: DE82007628
United States 10.2172/5155221 Other: ON: DE82007628 Thu Sept 22 07:33:03 EDT
2011NTIS, PC A06/MF A01.PNNL; EDB-82-141946 English
13. Nuclear Regulatory Commission (2016) Nuclear fuel pellets. https://www.flickr.com/photos/
nrcgov/15420174614/
14. Von Hippel FN (2001) Plutonium and reprocessing of spent nuclear fuel. Science 293:2397–
2398
15. Albright D, Kelleher-Vergantini S (2015) Plutonium and highly enriched uranium inventories,
2015 [Text]. http://isis-online.org/isis-reports/detail/plutonium-and-highly-enriched-uranium-
inventories-2015/17
16. Albright D, Berkout F, Walker W (1996) Plutonium and highly enriched uranium. Oxford
University Press, Oxford
17. Albright D, Kramer K (2005) Neptunium 237 and Americium: World Inventories and
Proliferation Concerns. In: ISIS Document Collection, I. f. S. a. I. Security. ISIS
18. Ahlswede J, Hebel S, Ross JO, Schoetter R, Kalinowski MB (2013) Update and improvement
of the global krypton-85 emission inventory. J Environ Radioact 115:34–42
19. National_Research_Council_Radioisotope_Power_Systems_Committee (2009). Radioisotope
Power Systems: An Imperative for Maintaining US Leadership in Space Exploration: National
Academies Press
20. Department_of_Energy (2005) Draft EIS for the Proposed Consolidation of Nuclear
Operations Related to Production of Radioisotope Power Systems, S. a. T. Office of
Nuclear Energy. Washington DC: DOE
21. Lastres O, Chandler D, Jarrell JJ, Maldonado GI (2011) Studies of Plutonium-238 production
at the high flux isotope reactor. Trans Am Nucl Soc 104:716–718
22. Wagner JC, Peterson JL, Mueller D, Gehin JC, Worrall A, Taiwo T, et al (2012)
Categorization of used nuclear fuel inventory in support of a comprehensive national nuclear
fuel cycle strategy. Oak Ridge National Laboratory (ORNL)
23. Yancey K, Tsvetkov PV (2014) Quantification of U.S. spent fuel inventories in nuclear waste
management. Ann Nucl Energ 72:277–285
24. Feiveson H, Mian Z, Ramana MV, Hippel FV (2011) Managing spent fuel from nuclear power
reactors. International Panel on Fissile Materials
78 2 Radioisotopes
25. World Nuclear Association (2015) Processing of used nuclear fuel. http://www.world-nuclear.
org/info/nuclear-fuel-cycle/fuel-recycling/processing-of-used-nuclear-fuel/
26. U.S. Linking legacies: connecting the Cold War nuclear weapons production processes to their
environmental consequences. [Washington, DC]: U.S. Dept. of Energy, Office of
Environmental Management, 1997
27. DOE Richland Operations Office (2015) Projects and facilities. http://www.hanford.gov/page.
cfm/ProjectsFacilities
28. Schwantes JM, Sweet LE (2011) Contaminants of the bismuth phosphate process as signifiers
of nuclear reprocessing history. Pacific Northwest National Laboratory
29. Serne RJ, Lindberg MJ, Jones TE, Schaef HT, Krupka KM (2007) Laboratory-Scale Bismuth
Phosphate Extraction Process Simulation to Track Fate of Fission Products. Pacific Northwest
National Laboratory
30. Todd T (2008) Spent nuclear fuel reprocessing. In: Nuclear regulatory commission seminar,
Rockville, MD
31. Prelas MA, Peck MS (2005) Nonproliferation issues for weapons of mass destruction. CRC
Press, Boca Raton
32. Oregon Department of Energy (2009) Hanford Cleanup: The First 20 Years. Oregon
Department of Energy, Salem, OR
33. Zabunoglu OH, Ozdemir L (2005) Purex co-processing of spent LWR fuels: Flow sheet. Ann
Nucl Energ 32:151–162
34. Ford GR (1976) Statement on nuclear policy. http://www.presidency.ucsb.edu/ws/?pid=
6561#axzz1zILTm1BT
35. Carter J (1977) Nuclear power policy statement on decisions reached following a review.
http://www.presidency.ucsb.edu/ws/?pid=7316
36. Williamson M, Pyroprocessing technologies: recycling used nuclear fuel for a sustainable
energy future. Argonne National Laboratory, US Department of Energy
37. Peterson J, MacDonell M, Haroun L, Monette F (2007) Radiological and chemical fact sheets
to support health risk analyses for contaminated areas. Argonne National Laboratory
38. World Nuclear Association (2016) Uranium mining overview. http://www.world-nuclear.org/
information-library/nuclear-fuel-cycle/mining-of-uranium/uranium-mining-overview.aspx
39. Carpenter D, Koshe G, Hu L-W (2012) MITR User’s Guide. Massachusetts Institute of
Technology
40. George K (1962) The Oak Ridge Research Reactor (ORR), the Low-Intensity Testing Reactor
(LITR), and the Oak Ridge Graphite Reactor (OGR) as experimental facilities, USAEC Report
ORNL-TM-279. Oak Ridge National Laboratory, Oak Ridge
41. Schultz C, Campbell J (2011) Advanced Test Reactor—Meeting U.S. nuclear energy research
challenges. I. N. Laboratory
42. Mausner L, BNL Radioisotope Research & Production Program. US Department of Energy,
Brookhaven National Laboratory
43. Mausner L, Isotope Production at High Energy. US Department of Energy, Brookhaven
National Laboratory
44. Srivastava S, Mausner L (2012) Radioisotope production. https://www.bnl.gov/cad/Isotope_
Distribution/Isodistoff.asp
45. Niowave (2014) Medical and industrial radioisotope production. Niowave
46. IAEA (2003) Manual for reactor produced radioisotopes. International Atomic Energy Agency
47. Alibaba. http://www.alibaba.com/
48. Ebay. http://www.ebay.com
49. Sigma-Aldrich. http://www.sigmaaldrich.com/united-states.html
50. Grainger. http://www.grainger.com/
51. Groenier W (1972) Calculation of the transient behavior of a dilute-purex solvent extraction
process having application to the reprocessing of LMFBR fuels. Oak Ridge National Lab,
Tenn
52. Monster Jobs. http://www.monster.com/
53. Bureau of Labor Statistics. http://www.bls.gov/
References 79
54. Lecours MJ, Prelas MA, Gunn S, Edwards C, Schlapper G (1982) Design, construction, and
testing of a nuclear-pumping facility at the University of Missouri Research Reactor. Rev Sci
Instrum 53:952–959
55. C. o. M. I. P. W. H. E. U. N. a. R. S. B. D. o. E. a. L. S. N. R. Council (2009) Medical isotope
production without highly enriched uranium. Washington, ED: National Research Council
56. Updegraff D, Hoedl SA (2013) Nuclear medicine without nuclear reactors or uranium
enrichment. Center for Science, Technology, and Security Policy American Association for
the Advancement of Science, Washington, DC 13:2013
57. Harvey JT, Isensee GH, Messina GP, Moffatt SD (2011) Domestic Production of Mo-99.
Presented at the 2011 Mo-99 Topical Meeting, Santa Fe, New Mexico
58. Theodore L (2013) Chemical engineering: the essential reference. McGraw Hill Professional
59. Schneider M (2015) Independent review on escalating MOX plant costs and DOE reversal of
MOX option for surplus plutonium. http://fissilematerials.org/blog/2015/05/independent_
review_on_esc.html
60. MFF Facility (2012, 2016) MOX fuel fabrication facility. http://www.moxproject.com/about/
61. Rot PDW Group (2014) Analysis of surplus weapon‐grade plutonium disposition options. US
Department of Energy
Chapter 3
Interactions of Ionizing Radiation
with Matter and Direct Energy
Conversion
Abstract Ionizing radiation is a broad term which refers to the fact that different
types of radiation will create ion pairs in matter. Ionizing radiation includes ions
(e.g., fission fragments and alpha particles), beta particles, gamma rays, x-rays, and
neutrons. Radioisotopes emit ionizing radiation and are viewed as the primary
power source for nuclear batteries. This chapter will explore various radioisotope
sources and their properties. The transducers which can be used in concert with
radioisotope sources will be discussed.
Keywords Radiation interactions Range Transducers
3.1 Ionizing Radiation Types and Ranges
Each type of ionizing radiation source has a characteristic range. Consider a

material in the solid phase, for example. Swift heavy ions such as fission fragments
and alpha particles will deposit their energy within a solid over a distance of
micrometers. Electrons deposit their energy over a range of millimeters. Particles
which possess high energy and either no rest mass or no net charge, such as gamma
rays and neutrons, deposit their energy over a range of meters.
3.1.1 Fission Fragments
The shortest transport scale lengths are for ions, and the most massive ions are the
fragments produced by fission. Fission commonly occurs through spontaneous de-
cay of a heavy atom like californium-252, which releases fast neutron energy and
fission fragments. The neutron energy and fission yield spectra are shown in
Figs. 3.1 and 3.2 respectively. Products of a spontaneous fission event are shown in
Eq. (3.1) where ffl is the light fission fragment, ffh is the heavy fission fragment, m is
the statistical average number of prompt fission neutrons, nfast, released during

82 3 Interactions of Ionizing Radiation with Matter …
Fig. 3.1 Energy spectrum of

neutrons produced by the
spontaneous fission of Cf-252
[1]
Fig. 3.2 Spontaneous fission

yields of Cf-252 [3]
fission and are emitted with a typical fast neutron distribution [1, 2]. Fission can also
be stimulated by neutron capture, whereby a nucleus absorbs an incident neutron,
becomes unstable, and breaks apart. An example of fission initiated through the
interaction of thermal neutrons with a fissile material, such as uranium-235, is shown
3.1 Ionizing Radiation Types and Ranges 83
in Eq. (3.2); where nth is a thermal neutron with energy on the order of 25 MeV.
Thermal fission also releases fast neutrons and fission fragments; the neutron energy
distribution and bimodal fission yield distribution of U-235 are shown in Figs. 3.3
and 3.4, respectively. The average energy produced by particles released in the
Fig. 3.3 Neutrons energy

spectrum produced by the
thermal fission of U-235 [4]
Fig. 3.4 U-235 fission yields

for high- and low-energy
(thermal) incident neutrons
[4]
Table 3.1 Statistical Radiation Energy in

distribution of energy released MeV
in the fission of U-235 [4]
Kinetic energy of fission fragments 167
Fission neutrons 5
Prompt gamma rays 5
Delayed gamma rays from fission 6
fragments
Capture gamma ray energy 10
Beta particles from fission fragments 7
Neutrinos 10
Total energy 210
fission of U-235, including neutrons, gamma rays, beta particles and neutrinos, is
shown in Table 3.1.
252
98 Cf ! ffl þ ffh þ m nfast ð3:1Þ
92 U þ nth ! 236
92 U ! ffl þ ffh þ m nfast ð3:2Þ
235
Kinetic energy for each fission fragments depends on the mass of the fragments
as shown in Eqs. (3.3) and (3.4), where KEffl is the kinetic energy of the light
fission fragment, KEffh is the kinetic energy of the heavy fission fragment, KEff is
the total kinetic energy of both fission fragments, mh is the mass of the heavy fission
fragment and ml is the mass of the light fission fragment. The linear energy transfer
of fission fragments and other swift heavy ions can be calculated using the
Bethe-Bloch formula.
mh
KEffl ¼ KEff ð3:3Þ
mh þ ml
ml
KEffh ¼ KEff ð3:4Þ
mh þ ml
For example, consider the specific fission reaction of U-235 shown in Eq. (3.5) that
produces La-147 and Br-87. The kinetic energies of the fission fragments are calcu-
lated in Eqs. (3.6) and (3.7), respectively, and the energy from the fission reaction
products are shown in Table 3.2. As stated earlier, the ranges of fission fragments in
matter are very short due to their mass and charge. Ranges for the two fission frag-
ments used in this example are pictured in Figs. 3.5 and 3.6. The bromine-87 atom has
a range of 6.29 lm in uranium metal. The spatial energy distributions of both fission
fragments within the material are shown in Figs. 3.7 and 3.8, respectively.
92 U þ nth ! 147
57 La þ 35 Br þ 2nfast þ Q(195 MeV) ð3:5Þ
235 87
Table 3.2 Distribution of Radiation Energy in

energy released during the MeV
fission of U-235 which yields
the specific fission fragments Kinetic energy of fission fragments 162
La-147 and Br-87 [5] Fission neutrons 6
Prompt gamma rays 6
Delayed gamma rays from fission 5
fragments
Beta particles from fission fragments 5
Neutrinos 11
Total energy 195
Table 3.1 is an overall statistical yield for fission while this table
is specific to a single fission reaction
Fig. 3.5 SRIM2011 model showing ion distribution in uranium of 101 MeV bromine-87 ions
slowing down [6]. The plot ordinate has units (Atoms/cm3)/(Atoms/cm2). By multiplying with ion
dose (units of Atoms/cm2 of bromine-87), the ordinate converts to a density distribution of Br-87
with units of (Atoms/cm3). The ion source originates from the left side so the two dimensional
plane indicates the depth and width of the ion distribution
Fig. 3.6 SRIM2011 model showing final ion distribution of 60 MeV lanthanum-147 ions
transported through uranium metal [6]. The plot ordinate has units (Atoms/cm3)/(Atoms/cm2). By
multiplying with ion dose (units of Atoms/cm2 of La-147), the ordinate converts to a density
distribution of La-147 with units of (Atoms/cm3). The ion source originates from the left side so
the two dimensional plane indicates the depth and width of the ion distribution
87
KELa147 ¼ 162 ¼ 60:23 MeV ð3:6Þ
147 þ 87
147
KEBr87 ¼ 162 ¼ 101:77 MeV ð3:7Þ
147 þ 87
The fission reaction shown in Eq. (3.8) is the consequence of the interaction of a
fast neutron (energy greater than 1 MeV) and U-238. Fast fission of U-238, for
example, provides a large part of the explosive yield in a thermonuclear weapon.
The energy distribution of the fast fission products is similar to that of products in
thermal fission.
Fig. 3.7 SRIM2011 model

showing target ionization in
uranium metal by 101 MeV
bromine-87 ions [6]. The left
ordinate is energy loss (eV/
Å), the right ordinate is the
number of recoil atoms

of target ionization in uranium
metal by 60 MeV
lanthanum-147 ion [6]. The
left ordinate is energy loss
(eV/Å), the right ordinate is
the number of recoil atoms
92 U þ nfast ! 239
92 U ! ffh þ ffl þ m nfast ð3:8Þ
238
3.1.2 Alpha Particles
Alpha-emitting radioisotopes which are appropriate for use in a nuclear battery are
described in Table 3.3. Polonium-210 is used an example here (Eq. 3.9).
Table 3.3 Potential a sources for nuclear batteries

Nuclide Decay energy Half life Other emissions Production
(MeV) (Years) (MeV, %) reactions
Gd-148 3.182 74.6 N/A Sm-147(a,3n)
Eu-151(p,4n)
Po-208 5.216 2.8979 b+ 0.3783 Bi-209(d, 3n)
(0.00223 %) Bi-209(p,2n)
Po-210 5.305 0.379 c 0.803 (0.0011 %) Natural source
Th-228 5.52 1.9131 a 5.340 (27.2 %) Natural source
5.423 (72.2 %)
c 0.216 (0.25 %)
U-232 5.414 68.9 a 5.263 (31.55 %) Pa-232(b)
5.32 (68.15 %) Th-232(a, 4n)
c 0.1–0.3 (low %)
Pu-236 5.867 2.857 a 5.721 (30.56 %) Np-236(b)
5.768 (69.26 %) U-235(a, 3n)
Pu-238 5.593 87.74 a 5.456 (28.98 %) Np-238(b)
5.499 (70.91 %) Np-237(n, c)
Am-241 5.638 432.2 a 5.442 (13 %) Pu-241(b)
5.485 (84.5 %)
c 0.05954 (35.9 %)
Cm-243 6.168 29.1 a 5.742 (11.5 %) Multiple-n capture
5.785 (72.9 %) U-238, Pu-239
5.992 (5.7 %)
6.058 (4.7 %)
c 0.2–0.3 (20 %)
Cm-244 5.902 18.1 a 5.762 (23.6 %) Multiple-n capture
5.805 (76.4 %) U-238, Pu-239,
c Low percentage Am-243
Bk-248 5.793 9 Cm-246(a,pn)
Cf-250 6.128 13.07 a 6.0304 (84.6 %) Multiple-n capture
5.989 (15.1 %) U-238, Pu-239,
c 0.04285 (0.014 %) Cm-244
Cf-252 6.217 2.645 SF FF (3.092 %) Multiple n capture
a 6.0758 (15.7 %) U-238, Pu-239,
6.118 (84.2 %) Cm-244
c 0.043–0.155
(0.015 %)
Es-252 6.739 1.292 a 6.5762 (13.6 %) Bk-249(a,n)
6.632 (80.2 %) Cf-252(d,2n)
c 0.043–0.924
(25 %)
The criteria used in identifying these isotopes is based on a half-life between 0.379 and 100 years.
Other emissions are shown such as gamma emission (for which additional shielding would be
needed) [7]
t1=2 ¼138:376 d
210
84 Po126 ! 206
82 Pb124 þ 42 He2 þ ð5:305 MeVÞ ð3:9Þ
Alpha particles are swift heavy ions whose interactions with matter are governed
by the Bethe-Bloch stopping power equation. The range of an alpha particle will be
greater than the range of a fission fragment in uranium metal due to its lower charge
and mass. This means that the alpha particles shown in Fig. 3.9 will travel a
distance of 9.32 lm, as opposed to 4.22 lm for heavy fission fragments and
6.29 lm for light fission fragments. The ionization produced by an alpha particle
along its path in a solid will follow a classical Bragg curve with a Bragg peak (see
Fig. 3.10), whereas a fission fragment has no Bragg peak (see Figs. 3.7 and 3.8),
due to the highly changing linear energy transfer of fission fragments as it picks up
electrons during the slowing down process. Further, the range of any charged
particle is a function of the electron density of the stopping material, such that less
dense materials provide a lower stopping power than higher density materials. For
Fig. 3.9 SRIM2011 model showing final ion distribution of 5.3 MeV alpha particles in uranium
metal [6]. The plot ordinate has units (Atoms/cm3)/(Atoms/cm2). By multiplying with ion dose
(units of Atoms/cm2 of He-4), the ordinate converts to a density distribution of He-4 with units of
(Atoms/cm3). The ion source originates from the left side so the two dimensional plane indicates
the depth and width of the ion distribution

showing target ionization
produced in uranium metal by
5.3 MeV alpha particles [6].
The left ordinate is energy
loss (eV/Å), the right ordinate
is the number of recoil atoms
example, the range of 5 MeV alpha particles in air is 40.6 mm (as compared to
9.32 lm in uranium metal). Therefore, it is often instructive to talk about ranges in
terms of areal density, which is the linear range divided by the density of the
material. The areal density is independent of density changes of the absorbing
materials and is similar for similar-Z materials (see the next section for more
discussion on this concept).
3.1.3 Beta Particles and Positrons
An isotope which produces an electron and antineutrino is a b− emitter, while an

isotope which produces a positron and neutrino is a b+ emitter. Energetic electrons
transfer energy to the electrons of the target material via Coulomb scattering and
Bremsstrahlung emission as calculated by the modified Bethe formula. Beta-
emitting radioisotopes which have suitable half-lives are shown in Table 3.4.
If an isotope instead emits a positron, the b+ will encounter some electron in
orbit around an atom. The two will then mutually annihilate to produce two
energetic gamma rays. These gamma rays then interact with matter, using mech-
anisms which are very different from those of electrons or other charged particles.
Compared to high velocity heavy ions, the path of an electron in matter is
complicated. Because the incident electron has a mass equal to that of the electrons
in the target, the electron undergoes significant scattering and follows the random
walk-like path shown in Figs. 3.11 and 3.12.
Table 3.4 Potential b− sources for nuclear batteries

Nuclide Decay Half life bmax (MeV) Other emissions Production
energy (Years) (Units in MeV) method
(MeV)
H-3 0.019 12.33 0.019 N/A Li-6(n, a)
Ar-39 0.565 269 0.565 N/A Ar-38(n, c)
KCl(n, c)
Ar-42 0.6 32.9 0.6 N/A Ar-40(n, c)
Ar-41(n, c)
Co-60 2.824 5.2713 0.318 c 1.17 (99 %) Co-59(n, c)
1.33 (0.12 %)
Kr-85 0.67 10.755 0.67 (99.6 %) c 0.514 (0.4 %) Fission product
0.15 (0.4 %)
Sr-90 0.546 28.77 0.546 bmax: 2.281 Fission product
(Y-90, daughter)
Ru-106 0.039 1.0234 0.039 N/A Fission product
Cd-113m 0.58 14.1 0.58 N/A Cd-112(n, c)
Cd-113(n, n)
Sb-125 0.767 2.73 0.7667 c 0.5 (5–20 %) Sn-124(n, c)
Cs-134 2.058 2.061 0.662 (71 %) c 0.6–0.8 (97 %) Cs-133(n, c)
0.089 (28 %)
Cs-137 1.175 30.1 1.176 (6.5 %) c 0.6617 (93.5 %) Fission Product
0.514 (93.5)
Pm-146 1.542 5.52 0.795 c 0.747 (33 %) Nd-146(p,n)
Nd-148(p,3n)
Pm-147 0.225 2.624 0.225 N/A Nd-146(n, c)
Sm-151 0.076 90 0.076 N/A Fission product
Eu-152 1.822 13.54 1.818 c 0.1–0.3 Eu-151(n, c)
Eu-154 1.969 8.592 1.845 (10 %) c 0.123 (38 %), 0.248 Eu-153(n, c)
0.571 (36.3 %) (7 %), 0.593 (6 %),
0.249 0.724 (21 %), 0.759
(28.59 %) (5 %), 0.876
(12 %), 1.0 (31 %),
1.278 (37 %)
Eu-155 0.253 4.67 0.147 (47.5 %) c 0.086 (30 %) Sm-154(n, c)
0.166 (25 %) 0.105 (21 %)
0.192 (8 %)
0.253 (17.6 %)
Tm-171 0.096 1.92 0.0964 (98 %) c 0.0667 (0.14 %) Er-170(n, c)
0.0297 (2 %)
Os-194 0.097 6 0.0143 c 0.01–0.08 Os-192(n, c)
(0.12 %) Os-193(n, c)
0.0535 (76 %)
0.0966 (24 %)
Tl-204 0.763 3.78 0.763 N/A Tl-203(n, c)
(continued)

Nuclide Decay Half life bmax (MeV) Other emissions Production
energy (Years) (Units in MeV) method
(MeV)
Pb-210 0.063 22.29 0.0169 (84 %) c 0.046 (4 %) Natural source
0.0635 (16 %)
Ra-228 0.046 5.75 0.0128 (30 %) c low E (low %) Natural source
0.0257 (20 %)
0.0392 (40 %)
0.0396 (10 %)
Ac-227 0.044 21.773 0.02 (10 %) a 4.953 (47.7 %) Ra-226(n, c)
0.0355 (35 %) 4.940 (39.6 %)
0.0448 (54 %) c 0.1 to 0.24 c
Pu-241 0.021 14.35 0.02082 a 4.853 (12.2 %) Multiple-n
4.896 (83.2 %) capture
U-238, Pu-239
The criteria used in identifying these isotopes is based on a half-life between 1 and 269 years.
Other emissions are shown such as gamma emission (for which additional shielding would be
needed) [7]
Fig. 3.11 GEANT4 simulation of Sr-90 Beta Decay into SiC of Slab, showing beta particle tracks
(random walk path) and bremsstrahlung photons (straight lines) [8]
3.1.4 Shielding Considerations
It should also be noted that most alpha and beta emitters considered in nuclear
battery technology do not emit significant levels of gamma rays due to potential
shielding concerns. For instance, if Co-60 is utilized in a beta-based nuclear battery,
Fig. 3.12 GEANT4 simulation of Sr-90 Beta Decay into SiC of Spherical Model, showing beta
particle tracks (random walk path) and bremsstrahlung photons (straight lines) [8]
then for 1 mW of power, assuming a 100 % conversion efficiency and complete

escape of the high energy gamma rays, would require 1.76 Ci. The associated high
energy gamma ray radiation (99 % of the decays will emit 1.17 MeV gamma rays)
from this large activity limits its suitability in many situations where radiation
effects to surrounding materials (e.g. electronics) and personnel is of importance.
Unshielded, the dose-rate at 20 cm from the source is 57.1 Rem/h. If a 20 cm lead
shield surrounds the device, the dose rate drops to 0.000569 Rem/h at the surface of
the lead shield. This is a particularly bad problem for microscale nuclear batteries,
where the shielding required to reduce the gamma-ray flux to acceptable levels
oftentimes severely reduces the overall energy density (We/kg) of the battery, which
also increases the battery footprint as a consequence.
An important point to make is that if the isotope has gamma emissions along
with charged particle emissions, there will be a source of penetrating radiation to
contend with. Depending on the gamma ray energy and probability of emission, the
dose rates can be significant. For example, krypton-85 emits a 514 keV gamma ray
0.4 % of the time [9–11]. Even though the 0.4 % seems small, it can add up to a
significant dose. A good example is based on an indirect nuclear battery developed
by one of the authors. The device was simple in that the krypton-85 gas operated as
both the source and transducer. The device drew commercial interest and two large
companies partnered with the author to develop the device. Beta particles emitted
from the isotope caused ionization and excitation of the krypton gas. The excited
atoms of krypton quickly form the krypton excimer molecule which then decays to
an unbound ground state. The excimer emission is narrow (with a peak at 149 nm
with a Full Width at Half Maximum of 10 nm). The excimer fluorescence is then
captured by the photovoltaic cells on the walls of the pressure vessel. The theo-
retical efficiency of this method is over 20 %. The device drew commercial interest
because the simplest configuration used a spherical pressure vessel filled with Kr-85
gas. The author and his collaborators designed a device for a space based mission
that required 1,000 W of power. The device design concept required an activity of
1,000,000 Curies of krypton gas. The atomic density of krypton in the sphere is,
Að0Þ ¼ kNð0Þ ð3:10Þ
where k = 0.693/t½ = 0.693/(10.755 year 365 day/year 24 h/day 3600

s/h) = 2.043 10−9 s−1.
decays
106 Ci 3:7 1010 sCi
Nð0Þ ¼ ¼ 1:811 1025 atoms ð3:11Þ
2:043 109 s1
There are 2.68 1019 atoms (or molecules) cm−3 in one atmosphere of gas at
Standard Temperature and Pressure (STP) and this number is used in the calculation
to find the pressure of 1,000,000 Ci of Kr-85 confined in the sphere. The calcu-
lation begins by assuming that the gas pressure in the sphere is low enough to be
governed by the idea gas law. The number of atoms (or molecules) in the sphere at
a pressure of one atmosphere (Natm) is found by multiplying the sphere volume by
the constant (Volumesphere 2.68 1019). The pressure of Kr-85 in the sphere can
be estimated from the relationship PKr85 = N(0)/Natm. This calculation shows that
the pressure of 1,000,000 Ci of Kr-85 in the 1000 cm3 sphere is 676 atmospheres.
From an engineering point of view, neither the size of the sphere nor the gas
pressure are significant challenges.
The sphere (radius of 6.2 cm) is small enough to be approximated as a point
source in a shielding calculation. Using Radpro Calculator [12] to estimate the
dose-rate at one meter from the bare sphere, the dose was found to be about 12.27
Rem/h. This dose-rate was determined to be unacceptable by the industrial partner
due to radiation sensitive electronics in the package that was going to be deployed
along with the power supply. In order to reduce the dose rate to an acceptable level
at 1 m, the device needed a shield. Several shield thicknesses were examined and an
8 cm thick lead shield reduced the dose-rate to a workable level of 24 lRem/h. The
mass of the 8 cm thick lead shield around the sphere is about 124.6 kg. This
indicates that the device would have a mass to power ratio of approximately
0.1246 kg/W which is extremely good compared to mass to power ratio of an RTG
(*5 kg/W).
3.1.5 Rules of Thumb and Their Limitations
The range R of electrons can be estimated using rules of thumb. For example,
Eqs. (3.12) and (3.13) can be used to estimate the range of electrons in air. On the
whole, rules of thumb such as these are useful for radiation protection considera-
tions but, as discussed below, the usefulness of any rule of thumb quickly breaks
down when applied to nuclear battery systems.
Rair ðftÞ 12 ft=MeV ð3:12Þ

R kg=m2 ¼ RðmÞ q kg=m3 ð3:13Þ
An ion pair which is produced from the interaction of beta particles with matter
includes an electron which is kicked out of orbit during the Coulomb interaction;
this electron is referred to as a secondary electron. Secondary electrons typically
have kinetic energies in the keV range [13]. The secondary electrons create tertiary
electrons through ionization and the tertiary electrons can interact to create qua-
ternary, quinary, or higher-order electrons. Because of their equal mass and charge,
electrons can transfer their full energy to a target electron through Coulomb
interactions. Nonetheless, the electron energy distribution in an electron beam
excited plasma [14] is very similar to the spectra created by light ion bombardment
(Fig. 3.13).
For example, the maximum energy of the beta particle produced by the decay of
phosphorus-32 (32P) is 1.71 MeV. According to Eq. (3.12) the range of the beta
particle in air is 1.71 12 = 20.5 ft 21 ft. For health physics professionals, this
number would be used as an estimate for a safe distance from the source.
There are also similar conservative calculations used by professionals to estimate
shielding thicknesses. The concept of areal density is the basis of one such method.
It can also be used to conservatively estimate the range of a beta particle in a target
material. The areal density is the density of a target material per square centimeter.
It is related to the electron density of a given material which is dependent upon the
atomic density and atomic number of that material. Since beta particles interact with
the electrons which make up the atom, it also can be used to estimate the amount of
energy lost by the beta particle as a function of distance (µm−1). Materials with
higher electron density have more electrons for the beta particle to interact with and
thus will stop the electron in a shorter distance. The areal density is calculated for a
specific material using Eq. (3.13), where q is the mass density of the material in
g/cm3. In many texts, the areal density is also referred to as ‘the range’ and given
the symbol R. The reader must sort out how the term ‘range’ is used in each
reference. There are rules of thumb for areal density which are shown in
Eqs. (3.14–3.18), where Emax is the maximum beta energy in MeV, and areal
density is given in g/cm2 (and uses the symbol R).
Fig. 3.13 Nascent and

asymptotic electron energy
distributions for 1-MeV alpha
particle bombardment of
helium [13]

Emax 2 R g=cm2 for 1 Emax 4 MeV ð3:14Þ
Emax 1:92 R0:725 for R 0:3 g=cm2 ð3:15Þ

R g=cm2 0:407 ðEmax Þ1:38 for Emax 0:8 MeV ð3:16Þ
Emax 1:85 R þ 0:245 for R 0:3 g=cm2 ð3:17Þ

R g=cm2 0:542 Emax 0:133 for Emax 0:8 MeV ð3:18Þ
Models which calculate energy deposition using rules of thumb to find areal density,
and therefore particle ranges, will introduce significant errors if used in nuclear battery
calculations. The rules of thumb presented above overestimate particle range in order to
ensure adequate radiation protection. For example, consider the beta emissions of
90 90
sulfur-35 35
16 S , strontium-90 38 Sr , and yttrium-90 39 Y . The maximum energy of
the beta particle emitted by 35S is 0.167 MeV, which results in a required areal
density of 0.407(0.167)1.38 0.034 (g/cm2) from Eq. (3.16). For a silicon carbide
target with density q = 3.210 g/cm3, the estimated range would be R(cm) R
(g/cm2)/q 0.034/3.210 0.0106 cm. The maximum energy of a beta particle
emitted by 90Sr is 0.546 MeV, which has an estimated areal density of 0.407
(0.546)1.38 0.177 (g/cm2), and the estimated range is R(cm) R(g/cm2)/q
0.177/3.210 0.0551 cm. The 2.28 MeV maximum energy of the beta particle
emitted by 90Y is considerably higher than the two previous examples. Yttrium-90
has an areal density of 2.28/2 1.14 (g/cm2) to stop the particles, which equates to a
range in silicon carbide of R(cm) R(g/cm2)/q 1.14/3.21 0.355 cm.
When two rules of thumb cover the same energy space, the significant errors
between them are obvious. For example, Eq. (3.18) covers beta energies greater
than 0.8 MeV and Eq. (3.14) covers beta energies between 1 and 4 MeV. If
Eq. (3.18) is used to estimate the range for a Sr-90 beta, it yields 0.546546
2.280–0.133 1.103 (g/cm2) and a consequent range in cm of approximately equal
to R (m) R (g/cm2)/q 1.103/3.210 0.344 cm. The difference between the
two rules of thumb (Eqs. 3.14 and 3.18) is about 11 µm, an order of magnitude
larger than the transducer scale length of some microscale nuclear battery designs.
Rules of thumb should never be used to design a nuclear battery.
3.1.6 The Limitations of Average Beta Energy
In addition to the rules of thumb discussed above, there are other simplifications that
may misrepresent energy deposition and, therefore, device efficiency. One common,
but incorrect, assumption is that all beta particles are emitted with an energy of 1/3
bmax [15]. This simplifies the calculation of energy deposition versus depth curve for
the device, but at the cost of verisimilitude. The ionization profile produced by a true
spectrum of beta particles is significantly different from the results of either the
simple rules of thumb described above or the 1/3 bmax assumption would indicate.
A second fatal simplification arises when it is assumed that the ionizing particles are
not emitted isotropically. The lack of accurate transport models creates shifts in the
energy deposition profile within the energy transducer, so a higher fraction of the
incident energy is deposited deeper within the device, which is not the case [16].
Realistic modeling of these sources is essential to accurately represent energy
deposition within the nuclear battery. Three relevant beta-decay reactions with low,
medium, and high-energy are shown in Eqs. (3.19–3.21).

17 Cl þ b þ
! 35 m þ 167:47 keV ð3:19Þ
35
16 S

39 Y þ b þ
! 90 m þ 546 keV ð3:20Þ
90
38 Sr
Table 3.5 Characteristics of common beta-emitting radioisotopes

Isotope Half-life Max Average energy % Daughter
energy 1/3 bmax rule Spectrum Difference isotope
(keV) (keV)
S-35 87.51 days 167.47 keV 55.8 53.1 +5 Cl-35
Sr-90 28.8 years 546 keV 182 167 +9 Y-90
Y-90 2.67 days 2.28 MeV 760 945 −20 Zr-90
The average energy is calculated using the 1/3 bmax rule and using a full spectrum analysis. The
differences in the average energy are substantial for high energy beta sources [7]

40 Zr þ b þ
! 90 m þ 2:28 MeV ð3:21Þ
90
39 Y
Table 3.5 shows pertinent data for the three beta emitters, including half-life, the
average beta energy according to the commonly-used rule of thumb (1/3 bmax), and
the average beta energy calculated using the beta spectrum. As can be seen in
Fig. 3.14, the sulfur-35 beta spectrum intensity continuously increases as energy
decreases, the beta spectrum intensity of the medium-energy strontium-90 emitter
tends to flatten out at low energies, and the high energy yttrium-90 beta spectrum
intensity has a distinct maxima and then drops as energy decreases. As can be seen
in Table 3.5, the differences between the average energy calculated by the 1/3 bmax
rule and the average energy calculated directly from the spectrum differs signifi-
cantly as the maximum energy of the beta particle increases. If the average beta
energy is calculated using the 1/3 bmax rule, the error inherent in using the 1/3 bmax
rule is propagated through the rest of the system calculations. These incorrect
average energies will then be used to calculate incorrect estimates of particle range
and stopping power. The 1/3 bmax rule should not be used for design calculations
and modeling of nuclear batteries.
Accurate range calculations are essential when designing a nuclear battery in
order to match the active region of the transducer (Ltrans) in the optimum position to
harvest the energy from the beta particle (kRadTr). To calculate the range of a beta
particle in matter, the full beta energy spectrum should be used in the model
(Fig. 3.14). Calculations which use the complete beta spectrum lead to the best
Fig. 3.14 Beta emission

energy spectra for S-35
(solid), Sr-90 (dashed), and
Y-90 (inset) [8]
possible estimation of energy deposition profiles. This is clearly shown in the

following example. The actual range of beta particles from S-35, Sr-90 and Y-90
decay have been calculated for a beam of beta particles hitting a slab and for a point
source in the center of a sphere [8]. These results are significantly different from
results that use the average beta energy calculated from the beta spectrum. This
further reinforces the premise that unacceptable inaccuracies come from the use of
any rules of thumb (Eqs. 3.14–3.18) when designing a nuclear battery. Results from
the rules of thumb are several orders of magnitude larger than calculations based on
the average beta energy (using the average based on the beta spectrum) or the full
beta spectrum. It is interesting that there is about a factor of 4 difference between
the range calculated from average beta energy and the range calculated with the full
beta spectrum, where the range for the full beta spectrum is greater.
Figures 3.15 and 3.16 represent the energy deposition as a function of distance
for a calculation based on the average beta energy and a calculation based on the
full beta spectrum, respectively. In the slab geometry, the beta particles were
simulated as monodirectional, normal to the stopping material, whereas the point
source in the sphere was modeled as isotropic. The results are notably different and
again reinforce why designs based on average beta energy have significant errors.
Looking at Fig. 3.14, the beta spectrum from each of the isotopes that are repre-
sented in this discussion, it is apparent that there is a significant distribution of low
energy beta particles which are emitted. By definition, the average beta energy is at
the point where the number of beta particles greater than the average energy equal
to the number of beta particles less than the average energy. For S-35, the low
energy beta population continually increases as the energy approaches zero. For
Sr-90, there is a slight peak at 0.08 MeV, but generally the population is flat at
lower energies. For Y-90, there is a well-defined maximum in the spectrum at
0.8 MeV. The differences between Figs. 3.15 and 3.16 are not so surprising when
the low energy betas in a full spectrum are taken into account.
In the calculations based on average beta energy shown in Fig. 3.15, there are
distinct peaks for both the slab and spherical geometries for S-35, Sr-90 and Y-90.
For S-35, the peaks occur at 6 lm for the slab geometry and at 10 lm for the
spherical geometry. For Sr-90, the peaks occur at 35 lm for the slab geometry and
at 65 lm for the spherical geometry. For Y-90, the peaks occur at 400 lm for the
slab geometry and at 600 lm for the spherical geometry. In contrast, Fig. 3.16
shows that low energy betas from the spectrum dominate where the energy is
deposited. Low energy betas will have a shorter range in the material than high
energy beta particles. So, the energy deposited per unit depth into the stopping
material for both S-35 and Sr-90 is highest near the surface of the stopping material
and decays exponentially with depth. For the high energy Y-90 beta particles there
is a distance where energy deposition peaks. For the slab case it peaks at 150 lm
and for the spherical case it peaks at about 300 lm. Another interesting observation
is that the maximum amount of energy that is deposited is substantially different
(about a factor of two lower) for the full spectrum calculation as opposed to the
average energy calculation.
Fig. 3.15 Simulated energy deposition based on the average beta energy versus distance in both
the monodirectional beta source incident on a slab (left) and an isotropic source at the center of a
spherical (right) geometries using GEANT4, PENELOPE, and MCNPX codes for a S-35, b Sr-90,
c Y-90. [8]
The implications of the observations above are significant. First of all, in using
average beta energy to calculate the location of maximum energy deposition, and
therefore the location of the transducer, there will be substantial errors in placement
of the depletion layer within the cell. A betavoltaic cell is a p-n junction in which
the p-type material and n-type material form a junction through compensation. This
region in the cell is called the depletion zone. As will be discussed, by adjusting the
density of p type impurities and n-type impurities, the depletion layer width will
change. Typically, the depletion layer width in a well-designed cell will be about
1 lm thick. If the betavoltaic cell is viewed as a box, then it is possible to locate the
Fig. 3.16 Simulated energy deposition based on the full beta energy spectrum versus distance for
spherical (right) models using GEANT4, PENELOPE, and MCNPX codes for a S-35, b Sr-90,
c Y-90 [8]
depletion zone within the boundaries of the box. The challenge is to deposit as
much of the energy from the source’s beta particles into the 1 lm thick depletion
layer as possible. Realizing that the slab model is idealized in that a monodirec-
tional beta particle beam strikes the cell normal to the surface and that the spherical
model is idealized by placing a point source at the center of the sphere where the
beta source is equidistant from a shell within the sphere, both models vastly over
Table 3.6 Range of beta Radioisotope Range in millimeters

particles in SiC based on the
Rule of Average Beta
rules of thumb from
thumb beta spectrum
Eqs. (3.13–3.16) compared to
results in Figs. 3.15 and 3.16 S-35 10.6 0.02 0.08
Sr-90 55.1 0.12 0.40
Y-90 344.0 1.6 3.00
The results from the beta spectrum are exact and this table shows
the magnitude of expected errors in calculations which use rules
of thumb or average beta energy [7]
predict the beta energy being deposited in any given layer. In a realistic device the
beta source will be isotropic, thus creating a much greater challenge for depositing
the energy of the beta particles in the thin depletion layer. In summary, by using
average beta energies in design calculations, significant errors are made in locating
the optimum position for the depletion layer. There is also a significant error in
calculating the energy transfer rate to the depletion layer (Table 3.6).
3.1.7 What Types of Radiation Work Best with Nuclear

Batteries and Why
The types of ionizing radiation that work well with nuclear battery technology
where the major goal is to reduce the mass per watt are alpha emitters and beta
emitters. Alpha emitters have the shortest ranges and can couple well to many types
of transducer technologies. Beta emitters have a much longer range than alpha
emitters as well as other complexities which make them a more difficult challenge
to use. In the next section, examples of types of transducers that can be used in
nuclear battery design are discussed to provide a more complete picture of how
alpha and beta emitters can be coupled to the transducer.
3.2 Types of Transducers Used in Nuclear Batteries
The basic operational principles of a nuclear battery can be classified as being based
on the generation of heat (Heat Based) or on the generation of ion pairs (Ion Pair
Based). Several examples of heat based and ion pair based nuclear batteries will be
explored.
3.2 Types of Transducers Used in Nuclear Batteries 103
3.2.1 Ion Pair Based
Alphavoltaic and betavoltaic cells are typical of semiconductor based energy

conversion. These types of cells have been studied since the 1950s [17, 18]. The
ionizing radiation creates electron-hole pairs in a semiconductor. The depletion
region, where the p-n junction forms, creates a local potential due to the com-
pensation of the n- and p-type regions, and this potential separates the electrons
from the holes. The depletion region is limited in thickness to the micrometer range
(*1 lm), as will be shown. If the range of the ionizing radiation in the solid is
greater than the depletion region, then the fraction of charged pairs produced that
the cell can harvest is low. Alpha particles have a range of about 20 lm in a solid.
The theoretical maximum efficiency of an alphavoltaic cell based on silicon carbide
was reported to be about 3 % [19]. Likewise, the range of a beta particle in a solid is
much higher, thus the corresponding theoretical maximum efficiency of a betavoltic
cell is around 1 % [8].
If the depletion layer can be extended beyond the 1 lm range, then the
achievable efficiency can be increased [16]. In this reference, tables are generated
for the energy transport efficiency (ηd) as a function of distance from the source for
idealized geometries in SiC (mono directional beam on a slab or a point source in a
sphere—this is an intentional overestimation of transport efficiency, since these
idealized geometries represent the most efficient means to transport radiation to the
depletion layer and all other designs will be less efficient. In other words, this logic
leads to a theoretical maximum). The transport efficiency is maximized by choosing
the thickness of the depletion layer and locating it at the peak of the energy
deposition curve (see Table 3.7) for an alpha particle. In several references [8, 16,
19] the depletion layer was chosen as 1 lm because the impurity level in a high
quality SiC wafer is typically on the order of 1 1016 atoms/cm3. To broaden the
depletion layer beyond 1 lm, as will be discussed, the dopant densities should be
less than 1 1016 atoms/cm3. However, the ability to reduce impurity densities
below 1 1016 atoms/cm3 is a technological challenge when this is about the same
impurity density level that a crystal will have when fabricated by typical methods.
One can try to compensate for high intrinsic donor or acceptor impurity levels in the
wafer by counter-doping with acceptor or donor impurities, respectively to attempt
to reduce effective dopant densities below 1 1016 atoms/cm3, but the net increase
in impurity density (dopant density plus counter dopant density) will result in
shorter carrier lifetimes and a reduction in the efficiency of the cell.
The major problem with a p-n diode is that the n- and p-type impurities in the
depletion region are displaced and the structure becomes more random due to
ionizing radiation. Thus, the cell ceases to function as a p-n junction relatively
quickly. The lifetime of a typical p-n junction beta or alpha voltaic cell is short even
at low dose rates. At high dose rates the cell’s lifetime could be on the order of
milliseconds [20] since radiation damage is a function of dose rate. However, this is
not the only problem that p-n diodes have. The other problem is a poor match
between the scale length of the radiation and the transducer.
Table 3.7 GEANT4 and SRIM/TRIM calculations for predicting energy deposition in a
Depletion Region of 1 µm thick for the slab and sphere models for a 5.307 MeV (e.g., Po-210)
alpha beam or point source respectively [19]
Range GEANT4 SRIM/TRIM
(µm) Sphere Slab Slab
Energy % r Energy % r Energy %
(keV) deposited (%) (keV) deposited (%) (keV) deposited
0–1 208 3.92 0.006 208 3.92 0.006 211 3.98
1–2 214 4.03 0.006 214 4.03 0.007 218 4.11
2–3 220 4.15 0.006 220 4.15 0.007 223 4.19
3–4 228 4.29 0.006 228 4.29 0.007 226 4.26
4–5 236 4.44 0.006 236 4.45 0.007 235 4.42
5–6 245 4.61 0.006 245 4.61 0.007 243 4.59
6–7 254 4.79 0.006 254 4.79 0.007 258 4.85
7–8 265 5.00 0.006 266 5.00 0.007 269 5.06
8–9 279 5.25 0.006 279 5.25 0.007 279 5.25
9–10 294 5.55 0.006 294 5.55 0.007 296 5.58
10–11 312 5.89 0.006 313 5.89 0.007 315 5.94
11–12 335 6.32 0.006 335 6.32 0.006 337 6.35
12–13 364 6.86 0.005 364 6.86 0.006 365 6.88
13–14 402 7.58 0.006 402 7.58 0.006 405 7.62
14–15 456 8.60 0.006 457 8.61 0.006 454 8.56
15–16 527 9.93 0.006 527 9.94 0.006 508 9.57
16–17 408 7.68 0.017 407 7.66 0.017 382 7.19
17–18 59 1.12 0.118 58 1.09 0.121 0 0.00
18–19 0 0.00 – 0 0.00 – 0 0.00
19–20 0 0.00 – 0 0.00 – 0 0.00
Total 5307 100.00 – 5307 100.00 – 5220 98.40
The ranges of beta particles do not match well with the depletion widths of p-n
diodes except at very low beta energies. Table 3.8 shows the energy deposition
characteristics in SiC, Xe gas and ZnSe for betas from tritium (average energy
5.45 keV), Ni-63 (average energy 17.20 keV), Sr-90 (average energy 196.03 keV)
and Y-90 (average energy 934 keV). As can be seen from the table, 99.95 % of the
beta energy is deposited in SiC within 1.71 lm for tritium, 15.7 lm for N-63,
474 lm for Sr-90 and 2873 lm for Y-90. This data shows that the scale length
match for a depletion width of the order of 1 lm is good for tritium but gets much
worse as the beta energy increases.
Large ions such as fission fragments have a range on the order of 5 lm. When
compared to the 1 lm depletion region, this range is a much better match of scale
lengths than that for most beta emitters (where betas at very low energy are on par
with 5 lm). In this case the theoretical maximum energy conversion can approach
Table 3.8 Energy deposition characteristics of the beta particles emitted by four radioisotopes in
various materials
Target Isotope Eexp Etot,M % Depth (in µm) of % Etot,M deposition
material (keV) (keV) Diff 25 % 50 % 75 % 99.95 %
SiC H-3 5.45 4.32 20.8 <0.1 0.128 0.292 1.71
Ni-63 17.20 14.08 18.2 <1 0.968 2.47 15.7
Sr-90 196.03 159.12 18.8 16.9 49.0 107.7 474
Y-90 934.40 772.79 17.3 136 366 756 2873
Xe H-3 5.45 3.72 31.7 29.7 83.7 190 1179
Ni-63 17.20 11.95 30.6 181 548 1329 8929
Sr-90 196.03 133.46 31.9 8170 22,880 49,909 247,437
Y-90 934.40 655.57 29.8 64,838 169,497 348,379 1,148,492
ZnS H-3 5.45 4.01 26.4 <0.1 0.0987 0.2361 1.3941
Ni-63 17.20 13.04 24.2 <1 <1 1.84 12.5
Sr-90 196.03 101.62 48.2 10.9 31.5 70.5 344
Y-90 934.40 476.51 49.0 86.97 236.4 498.9 2119
Eexp is the expected beta particle energy calculated using the spectral energy distribution [21]. Etot,
M is the average beta energy deposited in a sample (where Ltrans kRadTr) as calculated by
MCNPX. The % difference between the two methods is shown in column 5. The last four columns
describe the depth at which 25, 50, 75 and 99.95 % of Etot,M is deposited for the various beta
emitters and materials [7]
20 %. Unfortunately, the radiation damage to the junction will be even more rapid
with these larger fragments of higher energy because of large dose rates.
3.2.1.1 Efficiency of a Beta or Alpha Voltaic Cell Based on a Classic

P-N Junction
A junction is formed between n- and p-type semiconductor materials. At the

junction’s interface, the diffusion of charge carriers across the junction sets up a
space charge which in turn sets up an electrical potential and electrical field
(Fig. 3.17). The depletion region thus has an electrical field which separates
electron-hole pairs when they are created by the interaction of radiation with the
material. Outside of the depletion region, electron-hole pairs which are created by
the interaction of radiation with matter do not have electrical fields to cause their
separation. The most likely fate of the charge carriers beyond the depletion region is
that they will instead recombine or become trapped by defects. Some of the charge
carriers may drift into the depletion region and become part of the cell current
driven by the junction’s electric field.
Once an electrical field is in place, a driving potential for current flow exists
within the cell. The absolute efficiency (ηab) of an energy conversion device is
Fig. 3.17 Illustration of a p-n junction showing (from top to bottom): concentration profiles of n
and p type materials; junction interface with charge carriers compensation; diffusion of charge
carriers across the junction to produce space charge; the electrical potential; and electric field [7]
always defined as shown in (Eq. 3.22), where Pout is the output power from the
energy conversion device calculated by (3.23), Ptotal is the total power deposited in
the energy-transducing regions of the device, and Prad is the total radioisotope
power incident on the cell; Ptotal can be calculated from (3.24).
Pout
gab ¼ ð3:22Þ
Ptotal
Pout ðWÞ ¼ VðVoltsÞ IðAÞ ð3:23Þ

where V is cell potential in volts and I is the cell current in Amps.

Ptotal ðWÞ ¼ AðBqÞ Edis eV Bq1 1:6 1019 ðJ=eVÞ ð3:24Þ
where A is radioactive source activity, Edis is the energy of the particle produced by
the decay of the source.
To estimate the cell current I, one can find the maximum current that can be
created in the cell by the interaction of radiation with matter. The maximum current
will be dependent on the energy transport efficiency of the radiation (beta or alpha)
to the depletion zone. This can be expressed in a term which represents the fraction
of power that is deposited in the depletion zone and is a factor called ηd (Eq. 3.25).
This is typically calculated by using Monte Carlo based transport codes.
Power absorbed in depletion zone ¼ Pdpl ¼ Ptotal gd ð3:25Þ
The maximum rate of charge production in the depletion zone is then found
through the number of electron-hole pairs that are created per second in the
depletion zone. As discussed previously, Monte Carlo radiation transport codes do
not have the capability of modeling electron and ion motion in the media. The use
of the W (eV/ion pair) value, which is the average amount of energy that it takes to
make an electron-hole pair, in this calculation is necessary. Monte Carlo codes have
the capability of calculating the rate of energy absorbed spatially in a material. But,
as previously discussed, the spatial and temporal distribution of electrons and ions
should be in close proximity to the spatial energy deposition in a solid. Thus, a
reasonable estimate is to use the spatial rate of energy lost calculated by Monte
Carlo Codes to find the power density in the depletion zone and then use the power
density distribution to estimate the electron and ion density distribution. Therefore,
the number of electron-hole pairs created per second (Ne) in the depletion zone is,
Ptotal ðJ=sÞ gd
Ne ð# pairs=sÞ ¼ 6:25 1018 ðeV=JÞ ð3:26Þ
WðeV=ion pairÞ
Assuming that the cell has no losses due to traps, the production rate of
electron-hole pairs is proportional to the maximum ideal short circuit current (Jsc) in
the junction. Ideal short circuit current is equal to the production rate of
electron-hole pairs multiplied by the charge per electron (1.6 10−19 C):
Jsc ¼ Ne ð#pairs=sÞ 1:6 1019 ðC=pairÞ ð3:27Þ
Jsc ¼ Ptotal gd =W ð3:28Þ
The power out is related to the open circuit voltage (Voc) multiplied by the short
circuit current (Jsc) and the Fill Factor (FF). Thus, the maximum power out that a
p-n junction is capable of producing is shown in Eq. (3.29). In Eq. (3.30), Pmax is
the optimized power output of the cell and the FF is typically > 0.7 for high grade
solar cells.
Voc Ptotal gd FF
Pout ðWÞ ¼ Voc ðVÞ Jsc ðAÞ FF ¼ ð3:29Þ
W
Pmax
FF ¼ ð3:30Þ
Voc Jsc
Open Circuit Voltage (Voc) and the Driving Potential Efficiency
When electron-hole pairs are created in a solid-state material, any energy trans-
ferred to the electron greater than the band-gap energy (Eg) eventually ends up
becoming heat. Thus, the process of creating electron-hole pairs has a material
dependent efficiency (shown in Table 3.9). When creating an electron-hole pair, the
energy that is expended to create the charge carriers is Eg (eV). In a beta or alpha
voltaic cell, the electron-hole pair is driven by the junction voltage. The maximum
voltage is the open circuit voltage Voc. The maximum power produced by the cell is
thus the open circuit voltage times the short circuit current, Jsc. Current flow I (C/s)
is electron density qe (electrons/m3) times charge per electron times the drift
velocity u (m/s) times the surface area A (m2). This is very similar to a photovoltaic
solar cell, and therefore, for a nuclear battery the maximum possible current will be
equal to the number of electron-hole pairs created in the depletion zone per second,
assuming no recombination losses.
Using an alternate definition, one can start with the electron-hole production rate
through the interaction of the radiation with matter in Eq. (3.26) to define the short
circuit current, as in Eq. (3.28), and the maximum power out in Eq. (3.29). The
power expended to create the current flow (Pex) can be defined as the electron flow
rate times the energy used to form a single electron-hole pair (e.g., the band gap
(Eg) energy),
Pex ðWÞ ¼ Ne ðpair=sÞ Eg ðeV=pairÞ 1:6 1019 ðJ=eVÞ ð3:31Þ
The concept of the driving potential efficiency (ƞdp) was introduced by Oh, et al.
[19]; it originates from the relationship between the open circuit voltage and the
band gap of the cell material. The magnitude of the open circuit voltage is less than
or equal to the material band gap. A simple way to consider this relationship is that
the cell power output will be less than the power expended in creating the
electron-hole pairs. The ratio of the power out to the power expended is the driving
potential efficiency, as shown in Eqs. (3.32) and (3.33). The open circuit voltage
can then be represented as the product of the driving potential efficiency and the
band gap, shown in Eq. (3.34).
Table 3.9 Properties for some common semiconductor materials which are useful for direct nuclear energy conversion [22]
Material Minimum Electron drift Density Atomic Molar Displacement Mean ionization Klein Eg/
band-gap mobility (l) (q) mass density energy (Ed) energy (W eV) Formula* W
(Eg) (eV) (cm2/V-s) (g/cm3) (g/mol) (mol/cm3) (eV) [22] (W eV)
Silicon 1.12 1450 2.329 28.1 0.0829 *19 3.63 3.64 0.308
Germanium 0.68 3900 5.323 72.6 0.0733 30 2.96 2.404 0.23
Gallium 1.42 8500 5.317 144.6 0.0368 10 4.13 4.476 0.344
arsenide
Silicon 2.9 400 3.22 40.1 0.0803 28 6.88 8.62 0.421
3.2 Types of Transducers Used in Nuclear Batteries
carbide
Gallium 3.39 1000 6.15 83.7 0.0735 24 8.9 9.99 0.381
nitride
Diamond 5.48 1800 3.515 12 0.293 43 12.4 15.84 0.442
*Klein Formula for predicting W values for semiconductors (W = 2.8 Eg + 0.5 eV [23]). The predicted values from the Klein Formula are different than
the measured values especially at larger band-gap energies
109

Ne ðelectron=sÞ
Pout Voc ðVÞ 6:251018 ðelectron=cÞ
gdp ¼ ¼ ð3:32Þ
Pex Eg ðeV=pairÞ Ne ðpair=sÞ 1:6 1019 ðJ=eVÞ
Voc
gdp ¼ ð3:33Þ
Eg
Voc ¼ gdp Eg ð3:34Þ
So, the maximum efficiency can be written in terms of the power out and total
power of the emitter and can be calculated from Eq. (3.35). Substituting Eq. (3.32)
causes this expression to simplify to Eq. (3.36). Similarly, the use of the pair
production efficiency ηpp discussed by Oh, et al. [19], shown in Eq. (3.37), will
resolve Eq. (3.36) to the convenient form shown in Eq. (3.38).

FF
g ¼ Pout =Ptotal ¼ Voc ðVÞ gd ð3:35Þ
WðeV=pairÞ
FF
g ¼ gdp Eg gd ð3:36Þ
WðeV=pairÞ
Eg ðeVÞ
gpp ¼ ð3:37Þ
WðeV=pairÞ
g ¼ gdp gd gpp FF ð3:38Þ
The efficiency calculation requires one more parameter, the open circuit voltage.
Using the ideal PV cell-equivalent circuit in Fig. 3.18, the output current for an
ideal PV cell is related to the dark saturation current of the p-n junction (ID) and
radiation generated current (IL) by taking the nodal balance of the circuit in
Eq. (3.42).
Fig. 3.18 Ideal alpha or beta

voltaic cell equivalent circuit
The open circuit voltage is a function of the dark saturation current (I0) and the
radiation generated current (IL). The diode current is a function of the dark satu-
ration current in Eq. (3.39), where I0 is the reverse saturation current which is a
function of the material and temperature, q is the electron charge
(1.602 10−19 C), kB is Boltzmann’s constant (1.381 10−23 J/K), T is the cell
temperature in Kelvin and n is the shape factor (for an ideal cell n = 1). The open
circuit voltage occurs when the radiation generated current balances the diode
current. So setting I = 0 and using Eq. (3.40) for ID, one can find the open circuit
voltage as shown in Eq. (3.41).
I ¼ IL ID ð3:39Þ
qV
ID ¼ I0 enkB T 1 ð3:40Þ

nkB T IL
Voc ¼ ln þ1 ð3:41Þ
q I0
There is also a relationship between band gap energy and open circuit voltage.
The dark saturation current I0 depends on the charge carriers due to the temperature
of the cell T. The relationship between the dark saturation current and band-gap Eg
is shown in Eq. (3.42).
qEg
I0 ¼ D T 3 enkB T ð3:42Þ
As the band-gap increases, the dark saturation current will decrease. Thus, open
circuit voltage will increase for the ideal cell as band gap energy goes up. The
highest open circuit voltage achieved for a wide band-gap cell (e.g., diamond) was
2.6 V [24], which is equivalent to a driving potential efficiency (ƞdp) of 0.48.
Depletion Zone Width and Current
The collection probability is the probability that a carrier generated in a cell by the
interaction of radiation with matter in a region of the cell will be collected and
thereby contribute to the radiation generated current flow (IL). The carriers gener-
ated in the depletion region will have a collection probability of unity because the
electron-hole pairs are quickly separated by the electric field and eventually are
collected. Outside of the depletion zone, the collection probability decreases
because the electron-hole pairs must diffuse into the depletion region. If the distance
is more than one diffusion length away from the junction, the collection probability
is negligible (Fig. 3.19). For example, the diffusion length in SiC can vary from
0.07 lm to a few lm depending on material defects [25, 26].
In the determination of the width of the depletion zone, the conduction of charge
carriers is important, but even of more significance is the lifetime of the charge
Fig. 3.19 The collection probability for a linearly graded alpha or beta voltaic structure. The
collection probability is negligible beyond the diffusion length of carriers
carriers within the semiconducting material, define as the time from their creation or
injection to their subsequent collection by a true conductor (e.g., copper). The
lifetime of charge carriers (electrons and holes) in semiconductors are governed by
the traps that are present within the band structure. The consequence is that the
depletion region is limited to a very small thickness. This means that only the
energy that is deposited by ionizing radiation within the depletion region has a
collection efficiency of unity for power production. The depletion layer thus limits
the efficiency of the energy conversion system. Equations (3.43) through (3.45)
describe the pertinent properties of the semiconductor through the intrinsic carrier
concentration, ni, the voltage across the depletion region, Vbi, and the depletion
region width, W [27].

2pkB T 3=2 3=4 Eg
ni ¼ 2 mn mp exp ð3:43Þ
h2 2kB T
where kB is Boltzmann’s constant, T is the temperature of the semiconductor in

Kelvin, Eg is the band gap energy, and mp and mc are the effective masses of the
holes and electrons, respectively.

kB T Na Nd
Vbi ¼ ln ð3:44Þ
e n2i
where e is the unit electron charge, Na is the relative concentration of vacancies

available in the p-type region, Nd is the relative concentration of electrons in the
n-type region.

2es Vbi Na þ Nd 1=2
W¼ ð3:45Þ
e Na Nd
where es is the permittivity.

As can be seen in Eqs. (3.43–3.45), the potential barrier and the depletion region
width are based on the intrinsic semiconductor properties and the doping concen-
trations Na and Nd in the respective p- and n-type regions of the semiconductor.
Doping concentration is the primary means of varying the potential barrier and
depletion width. The equations show that by reducing the doping concentrations,
larger depletion widths are formed, but at a penalty of a lower potential barrier. The
potential approaches zero as the doping concentration product approaches the
square of the intrinsic carrier concentration and there is no depletion region. In
addition, the doping concentration must be greater than the injected charge from the
depletion region for the transport of the generated charges to be in the low injection
regime, and very low doping concentrations are not controllable in a
semiconductor.
In order to understand the depletion width, 4H-SiC is chosen as an example
material. For this semiconductor, es is about 10 [28], mn and mp were taken to be 1.2
me and 0.76 me, respectively, and the band gap is approximately 3.25 eV at room
temperature [29]. Using a temperature of 300 K in Eq. (3.43), the intrinsic doping
concentration for this material was found to be 9.5 10−9 cm−3. This information
is used with Eqs. (3.44) and (3.45) and both Nd and Na are varied between 1015 and
1020 cm−3 to calculate the range of depletion widths in silicon carbide (see
Fig. 3.20). It can be seen that the largest depletion region width in silicon carbide is
2.6 µm. This shows that the depletion width is the factor that limits the absorption
efficiency of ionizing radiation in any planar single p-n junction-based nuclear
battery system. It should be noted that in order to achieve a depletion width of
2.6 µm, the impurity levels have to be on the order of 0.1 ppm. Achieving sub-ppm
impurity levels is difficult for SiC. It is more reasonable to obtain impurity levels in
the 1–10 ppm range. This leads to the conclusion that a depletion width on the
order of 1 µm is approximately the maximum size that is feasible [19].
3.2.2 Schottky Barriers
The Schottky barrier is a potential energy barrier between a metal and a semi-
conductor that forms a diode. It depends on the differential work function between a
semiconductor and a metal contact. The surface interface between the metal and
semiconductor forms a depletion region that is effectively 100s of nanometers thick.
Thus the effective size of the energy conversion region is thin and not well matched
to the ranges of ionizing radiation. The theoretical maximum efficiency for Schottky
barrier structures is approximately 0.3 % for alphas and approximately 0.1 % for
low energy betas. Schottky barriers are less susceptible to radiation damage since
Fig. 3.20 Density plot indicating depletion region width for varying donor and acceptor density
concentrations in 4H-SiC [7]
the atomic layer bonding the metal to the semiconductor is thin and is thus less
prone to displacements caused by the radiation. Schottky barriers can be formed
between liquid semiconductors and metals. These types of Schottky diodes are not
as prone to radiation damage since the liquid constantly wets the interface.
Fission fragments have a better match (a range of about 5 lm) for the *100 nm
depletion region of a Schottky barrier diode. For fission reactions combined with a
Schottky barrier diode, the theoretical maximum efficiency approaches 3 %. This is
still not attractive enough for a large scale direct energy conversion system in
fission reactors.
A Schottky barrier is formed when two different materials with different work
functions are placed in direct contact with each other. A Fermi potential will form at
the material interface. Because this boundary between the two materials is sharp, a
Schottky barrier will have a much thinner depletion zone than a p-n junction. The
width of the depletion region for a Schottky barrier, W, is expressed by Sze and Ng
[30]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2es
W¼ ðVbi VA Þ ð3:46Þ
qN
Vbi
gdp ¼ ð3:47Þ
Eg
where Vbi is the built-in voltage in the Schottky contact, N is the dopant density, q is
the unit charge, es is the permittivity of the semiconductor, and VA is the applied
voltage across the junction in the forward bias. In the beta or alpha voltaic mode,
VA = 0. A typical barrier height for a SiC Schottky barrier is 1 V [31]. The
depletion width for a Ni/4H-SiC Schottky diode is given for several donor con-
centrations in Östlund [32]. With a dopant density of 1 1017 cm−3, the depletion
width of a Ni/4H-SiC Schottky diode is about 0.25 µm.
The efficiency of a Schottky barrier diode for nuclear energy conversion will be
significantly less than a p-n diode due to the smaller depletion width as well as a
lower potential barrier height (Vbi). The smaller depletion width will decrease the
transducer scale length and thus will decrease efficiency. The lower potential barrier
height will lower the driving potential efficiency in Eq. (3.46) where Vbi is sub-
stituted for Voc. The difference between Vbi and Voc can be substantial, Vbi = 1 V
and Voc = 2.04 V [33], which can reduce ηdp by 50 %.
3.2.2.1 Liquid Semiconductor Schottky Barrier
The use of a liquid semiconductor to mitigate radiation damage to a nuclear battery

was patented with a priority date of November 21, 2003 [34]. The premise of this
cell is that two plates with a liquid semiconductor (such as selenium) between them,
one forming a Schottky barrier and the other an Ohmic contact, can form a potential
barrier which will serve as the driving force for the collection of electron-hole pairs
created in the cell. The semiconducting property of a molten semiconducting metal,
such as germanium, is accompanied by a rapid temperature-dependent increase in
conductivity. Thus, the molten material has metallic properties. There exist, how-
ever, some materials which upon melting do not exhibit metallic conductivity but
retain their semiconducting properties. These materials are primarily the chalcogens
(oxygen, sulfur, selenium, tellurium and polonium). Selenium is particularly
interesting due to the decrease in conductivity as temperatures increases, thus
exhibiting semiconducting behavior at high temperatures. The preferred embodi-
ment of the cell uses selenium as the liquid semiconductor. There is a depletion
region that is formed in the cell by the Schottky barrier where the collection
efficiency will be close to 100 %, and there will be a rapid drop off in collection
efficiency a diffusion length or so beyond the depletion region. A test cell was built
by Global Technologies Inc. and operated at an efficiency of about 1 %. This
efficiency is greater than the expected efficiencies of solid-state Schottky cells.
However, the diffusion lengths in a liquid medium should be larger than those in a
solid, which in turn should lead to higher efficiencies than solid-state Schottky cells.
Thus, the 1 % efficiency reported is reasonable. The basic technology behind the
Global Technology Inc. liquid selenium Schottky cell has since been explored by
others using a beta source that is mixed with the liquid [35].
A theoretical study was published on a liquid nuclear battery by the Jet
Propulsion Laboratory. This study [36] used liquid gallium as the electrolyte in an
electrolytic cell. The driving voltage was the work function difference between the
anode and cathode (*1.6 V). The basis for the device is that the liquid gallium
would be a semi-metal (the conductivity of liquid gallium is linear with temperature
and is on the order of 26–46 lX-cm, which appears to be metallic—e.g., silver has
a resistivity at room temperature of 1.59 lX-cm). Furthermore, it was argued that
the gallium ion does not immediately recombine, thus leading to high current
collection efficiencies. Unfortunately, the ion and electron mobilities and carrier
lifetimes in the liquid metal were not discussed, making it difficult to evaluate the
study.
3.2.3 Direct Charge Collection
The Direct Charge Nuclear Battery (DCNB) is based on the conversion of the
kinetic energy of a charged particle generated by nuclear decay to potential energy
stored in an electric field. Several theories have been developed to describe the
operation of a DCNB [37, 38].
The following derivation is more fundamental than many of the other approa-
ches. It examines how the kinetic energy of the charged particle is converted to
stored energy in an electric field. The DCNB consists of a radioisotope source
which is the emitter of charged particles. The source is thin enough so as to allow
the bulk of the kinetic energy in the emitted particle to escape the layer of source
material. The DCNB also has a collector which is separated from the charged
particle source by a vacuum (a separation distance of l). The breakdown voltage in a
vacuum is about 1 105 V/cm [39]. The vacuum is void of atoms which will
interact with the charged particle and absorb some of its energy. Thus, the particle
will slow down due to the electric field and it will collide with the collector. The
charge will remain in the collector, contributing an additional charge to add to the
strength of the electrical field. This effectively converts the kinetic energy of the
charged particle into stored potential energy in the form of an electric field.
3.2.3.1 The Ideal Match Between the Electric Field and the Ionizing
Radiation
In an ideal system, the charged particle will travel perpendicular to the electric field
where it will slow down sufficiently so that its velocity is zero when it attaches to
the collector. Because there is no kinetic energy remaining in the charged particle
when it attaches to the collector in this ideal match-up, its energy is fully converted
into stored potential energy in the DCNB. The additional charge will incrementally
increase the electric field created by the separation of the charged particle from its
paired opposite (e.g., for a beta particle the paired opposite is an ion) trapped in the
radioisotope layer. If the match-up is less than ideal, the charged particle still has
kinetic energy when it collides with the collector. This remaining kinetic energy in
the particle generates heat and ionization (as described in Chap. 1), which effec-
tively heats the collector. In this case the excess kinetic energy is wasted.
No DCNB design is ideal, thus a significant amount of the kinetic energy in the
charged particles ejected by the radiation source will interact with matter and
generate heat. Thus there are significant inefficiencies which will be described in the
following text. These inefficiencies in a DCNB are associated with imperfect
geometries and other import factors.
• One factor is self-absorption, where energy is lost as the charged particle moves
through the source layer.
• There will also be losses associated with the charged particle emission being
isotropic. Some part of the velocity vector will be perpendicular to the electric
field and the energy associated with the perpendicular velocity component will
be a loss because that portion of the energy will eventually interact with matter
and will be converted into heat.
• There will be inefficiencies in the electron collections process such as electron
backscatter and electron emission from the collector material.
DCNB Design and Inefficiencies
Putting the above concepts into context, envision a process where the source (for
this construct it is assumed that the source is a beta emitter) is coated on a rect-
angular plate located a distance “l” from the collector is a parallel plate (see
Fig. 3.21). The space between the two plates is a vacuum.
Each particle emitted has an equal probability of being emitted at any angle of
emission. The particles which are collected must have a high enough kinetic energy
Fig. 3.21 An illustration of a parallel plate DCNB. The electron is emitted at an angle h. The
electric field is in the positive x direction which impedes the electron’s velocity vector in the
x direction due to the Lorentz force. (Note that if the charged particle were an ion, simply reverse
the field. The ion would experience a Lorentz force which impedes its velocity vector in the
x direction)
to overcome the electrical field created by the accumulated charge on the collector
plate. Looking at the charge build up process, the first particle which is emitted has
no electrical potential to overcome, so it will collide with the collection plate and
will interact with the material causing the particle to slow down. The transfer of
kinetic energy from the charged particle to the collector plate material results in the
collector plate gaining heat. Each charged particle that is collected will incremen-
tally build up the electric field. The electric field for a parallel plate capacitor is,
~ Q
E¼ ^x ð3:48Þ
Ae0
where Q is the total charge in Coulombs collected on the plate, A is the area of the
plate, e0 is the permittivity of free space (8.8542 10−12 Farads/m) and ^x is a unit
vector in the x direction.
The electrical potential between the parallel plates is,
jV j ¼ El ð3:49Þ
where jV j is the absolute value of the electrical potential in volts, E is the magnitude
of the electric field vector, l is the distance between the plates in meters.
To provide an idea of how much energy is lost during the charge buildup phase,
assume that the source of charged particles is a monoenergetic electron beam (of
energy 100 keV) moving in the positive x direction, striking a 0.1 m by 0.1 m plate
(a large enough area where edge effects considerations on the electric field are not
significant). As previously discussed, the vacuum breakdown voltage is
1 107 V/m. Thus a 0.01 m spacing (l) between the plates is sufficient to prohibit
a breakdown for a voltage of 1 105 V (i.e., the ideal matchup between the
electric field and the 100 keV particle). Using Eqs. (3.48) and (3.49), an equation
for the charge on the plate needed to produce 100 kV is,
Q ¼ Ex Ae0 ¼ 107 102 8:854 1012 ¼ 8:854 107 C ð3:50Þ
The number of electrons (Ne) needed to produce this charge is,
Ne ¼ 8:854 107 C=1:66 1019 ðC=electronÞ ¼ 5:53 1012 :
The buildup time (sb) to fully charge the plate for a source that produces
3.7 1010 electrons per second is
sb ¼ 5:5353 1012 electrons=3:77 1010 ¼ 150 s ¼ 2:5 min:
As charge is accumulated on the plate, each electron which hits the plate will
lose some of its kinetic energy to the plate material. The kinetic energy lost (KElost)
can be calculated by summing the lost energy for each electron. The kinetic energy
of the electron (KEe) is,
KEe ¼ 100 103 eV=electron 1:66 1019 J=eV ¼ 1:66 1014 J:
The energy which is converted to potential energy stored in the electric field is
eExl. As a charge is captured by the plate, the incremental field is increased by
e2l/Ae0. Thus KElost is found by the sum,
X 12
5:5310
KElost ¼ 1:6 1014 n 2:881 1027 0:5 ð3:51Þ
0
where,
e2 l
¼ 2:881 1027 :
Ae0
The ideal efficiency during the charge build up phase for converting kinetic
energy to stored electrical field energy for a mono-energetic, mono-directional
e-beam is *50 %.
If the DCNB depends on the radioisotope to build up the charge on the plate,
half of the kinetic energy in the particles is used to build up the electric field and the
rest is lost to heat. Thus the efficiency of converting kinetic energy into stored
potential energy in the electrical field for an ideal charge build up based system (ηb)
is about 50 %.
Some DCNB systems function by first building up a charge followed by a
discharge. These are dynamic direct charge nuclear batteries and are limited in
efficiency by the charge build-up phase. The efficiency of the dynamic systems are
primarily limited by the charge build-up process (ηb), the mismatch of the field
strength with the energy distribution of the particles, the transport efficiency of the
particles out of the coating, the angular distribution of the particles being emitted,
electron backscatter, and electron emission from the surface.
Some DCNB systems operate at steady-state by continually collecting charge
and bleeding it off at a rate equal to the collection rate. During steady-state oper-
ation, enough charge is accumulated on the plate to produce a static electric field in
which the charged particles emitted by the source slow down. The efficiency of the
steady-state systems is primarily limited by the mismatch of the field strength with
the energy distribution of the particles (ηED), the transport efficiency of the particles
out of the coating (ηtr), the effect of the angular distribution of the particles being
emitted (ηm ηf), electron backscatter (ηBS) and electron emission (ηEM) from the
surface. The impact of these efficiencies on the DCNB system efficiencies will be
discussed.
Thus, the system efficiency for the Dynamic DCNB is,
gDynamic ¼ gb gED gtr gm gf gBS gEM ð3:52Þ

and the system efficiency for a steady state DCNB is,
gSS ¼ gED gtr gm gf gBS gEM ð3:53Þ
The potential energy that is stored per unit volume by the electric field of the
parallel plate configuration is,
Energy 1
q~E ¼ ¼ e0 Ex2 ð3:54Þ
Volume 2
where e0 is the permittivity of free space, q~E is the energy density, and Volume is
plate surface area (A) times gap between the plates (l).
Mismatch of the Electric Field with the Energy Distribution of the Particles
After enough charge is trapped on the collector plate to produce an electric field
which can slow down the particles from the source, a good fraction of the particle’s
kinetic energy can be stored in the electric field. The process depends on the
Lorentz force to slow down the charged particle (Eq. 3.55).
~
F ¼ e~
E ð3:55Þ
where ~ E is the electric field.

The work in moving the charged particle from the source plate to the collector
plate is,
Z l
W¼ e~
E d~
x ¼ eEx l ð3:56Þ
0
If the particle velocity is sufficient to overcome the repulsive Lorentz force, it

will contribute to the cumulative charge on the collector plate and will incremen-
tally increase the electric field strength. It is important to realize that in the parallel
plate configuration only the velocity component in the x direction (Vx) is impeded
by the Lorentz force. Because of isotropic emission, the Vx component (Eqs. 3.57
and 3.58) will vary as a function of emission angle. The Vy component will not
contribute to the energy stored in the electrical field and the energy associated with
the y component of velocity is lost.
1
KEx ¼ mq Vx2 ð3:57Þ
2
where mq is the mass of the charged particle.

The x component of the velocity is,
Vx ¼ V sin h ð3:58Þ
Thus,
1
KEx ¼ mq V 2 sin2 h ¼ KE sin2 h ð3:59Þ
2
The fraction of energy which is stored as potential energy in the electrical field
for this particle (ηq) is the kinetic energy in the x component times the fraction of
this energy which is stored in the electric field (Eqs. 3.60 and 3.61).
KE sin2 h
gq ¼ g~E ¼ g~E sin2 h ð3:60Þ
KE
where,
8 eE l
< KE sinx 2 h ; for KE sin2 h [ eEx l
g~E ¼ 1; for KE sin2 h ¼ eEx l ð3:61Þ
:
0; for KE sin2 h\eEx l
ηq is an idealized form of the energy distribution factor ηED (meaning ηq > ηED)
since ηq is derived for a mono-energetic charged particle source. Beta particles are
emitted with a spectrum of energies as discussed in Appendix A, thus finding ηED
requires a more complex analysis. The spectral distribution of the beta particles
must be accounted for in Eq. (3.60) in the KE term. The term becomes a distri-
bution function for particle kinetic energy F(E).
As previously discussed, the energy distribution is only one inefficiency factor.
There are other factors which reduce the DCNB efficiency including isotropic
emission (ηfηm), energy losses from charged particles transporting out of source
materials (ηtr), effects from secondary electron emission (ηSEE), and electron
backscatter.
Effects of Angular Distribution
The angular distribution of the particles emitted by the source will reduce the
efficiency of the device. Primarily, the electric field acts on the velocity component
of the particle which is parallel to it due to the Lorentz force. The components of
velocity which are perpendicular to the electric field are not affected, and thus the
energy associated with the perpendicular components of velocity is lost. Another
effect that is important is the loss of energy due to the mismatch of the velocity
component with the field. Only particles with KEx = eExl are perfectly matched
with the field and convert kinetic energy to potential energy stored in the field. If
Fig. 3.22 Minimum angle of

emission (hm) where a particle
of kinetic energy (KE) is
stopped at the collection plate
and the charge is absorbed
onto the plate thus
contributing incrementally to
the electric field (Ex)
KEx > eExl, then some of the energy is stored as potential energy in the field (the
portion equal to eExl), while the remainder of the energy is lost. If KE x < eExl, then
none of the kinetic energy is converted to potential energy in the field. Based on
these observations, there should be an emission angle which optimizes the effi-
ciency of the DCNB. The operating voltage of a cell which has an isotropic source
will be lower than that of the cell with a mono-energetic, mono-directional source
because of the broad range of Vx due to the emission angle. The optimum operating
voltage can be analytically determined.
In order to understand the effect of isotropic emission, assume that the kinetic
energy of the particles are mono-energetic. An angle hm is defined which represents
the emission angle where the electrical field is able to reduce Vx to zero when it
reaches the collector plate. This emission angle is the point which determines the
operating voltage of the DCNB (see Fig. 3.22). The Vx of the particle emitted at hm
is,
Vx ¼ V sin hm ð3:62Þ
The kinetic energy of this particle is,
1
KEx ¼ mq V 2 sin2 hm ¼ KE sin2 hm ð3:63Þ
2
The electric field which reduces the particle’s Vx at an emission angle hm is

found by setting Eq. (3.56) equal to Eq. (3.63),
elEx ¼ KE sin2 hm ð3:64Þ
So,
KE sin2 hm
Ex ¼ ð3:65Þ
el
where KE is in electron volts, e is 1.6 10−19 C and l is the distance between the
plates. The field strength will be in V/m.
Any particle that is emitted at an angle less than hm is turned around by the
electric field before reaching the plate. The particles emitted at an angle equal to or
greater than hm are collected. However, as indicated by Eq. (3.61), particles which
are emitted at an angle greater than hm have excess energy (KE-elEx) which goes
into heating the collector plate. The fraction of kinetic energy that contributes to the
potential energy of the field for any particle which is emitted at an angle greater
than or equal to hm is,
KE sin2 hm
gm ¼ ¼ sin2 hm ð3:66Þ
KE
The fraction of particles (ηf) emitted by an isotropic source which have angles of
emission between hm and p − hm is,
p 2hm
gf ¼ ð3:67Þ
2p
So, the ideal efficiency of a parallel plate DCNB with a mono-energetic isotropic
source is,
p 2hm
gideal ¼ gf gm ¼ sin2 hm ð3:68Þ
2p
The ideal efficiency will be reduced if the particles are not mono-energetic (e.g.,
beta particle spectrum), if there is electron backscatter, and/or if there is electron
emission.
Using Eq. (3.68) and plotting ideal efficiency (ηideal) as a function of the min-
imum angle of emission (hm), the optimum efficiency for an ideal DCNB using an
isotropic mono-energetic source can be found (Fig. 3.23).
The ideal system efficiency does not account for energy distribution effects
(ηED), transport effects (ηtr), electron backscatter effects (ηBS), or electron emission
effects (ηEM). It simply sets a theoretical maximum efficiency for a steady-state
DCNB. In the case of an isotropic source emitting particles with energy of 100 keV
as shown in Fig. 3.23, the operating voltage will be maximized for an emission
angle of about 1 rad (or 57.3°) which is 70.8 kV. The ideal efficiency will be about
13.5 %.
Fig. 3.23 The ideal efficiency of a DCNB as a function of minimum angle of emission
3.2.3.2 Reciprocating Cantilever
A reciprocating cantilever nuclear battery is a form of a DCNB. The principle of

operation of a reciprocating cantilever is that a radioisotope source emits charge and
that charge is collected on a receiver (the cantilever material). The kinetic energy
from the charged particles is converted into potential energy stored in an electro-
magnetic field. The cantilever is a flexible column and the Lorentz force causes the
column to bend towards the plate where the positive ions accumulate. When the
column touches the plate, then current will flow. This is a dynamic charge build up
as described by Eq. (3.51). In a dynamic DCNB, the charge is built up to a point
where the field is strong enough to cause the cantilever to bend and touch the plate.
As a dynamic DCNB, the intrinsic efficiency is 50 %, meaning that half of the
energy is lost as heat and half is stored as potential energy in the electric field. The
intrinsic efficiency does not include transport efficiency of the particles out of the
coating (ηtr), nor does it include the efficiency of the transducer once a discharge is
initiated (ηdischarge), as well as other factors shown in Eq. (3.52) (i.e., the energy
distribution of the particles (ηED), the effect of the angular distribution of the
particles being emitted (ηmηf), electron backscatter (ηBS), and electron emission
(ηEM)). Figure 3.23 shows the idealized effect of the angular distribution of the
particles being emitted having a peak efficiency of approximately 13.5 %. Due to
the 50 % intrinsic efficiency of a dynamic DCNB, the idealized efficiency of a
reciprocating cantilever is 6.75 %* (ηtr ηED ηBS ηEM ηdischarge.) (Fig. 3.24).
Fig. 3.24 Illustration of a self-cantilever system: 1 Electrons emitted from the beta source are
collected on the cantilever. 2 Positive charges on the cantilever are due to charge transfer. The
electrostatic force pulls the cantilever towards the beta source. 3 Cantilever gets in contact with
source and charges are neutralized. 4 After charge neutralization, there is not enough attractive
force to keep the cantilever in contact with the radioisotope. 5 The spring force retracts the
cantilever to its original position and a new cycle begins
3.2.4 Indirect
NASA has successfully used thermionics with the Radioisotope Thermal Generator
(RTG) systems in a number of missions (e.g., Apollo-SNAP-27 generators,
Voyager, and Cassini). The energy resulting from the decay of radioisotopes starts
out as very high-grade energy: multi-MeV, heavy, and highly charged particles.
However, in the usual embodiment of radioisotope fuel in the RTG, as a solid, the
charged particles are absorbed in a very short distance (micrometers) and ther-
malized to moderate temperatures (in the range of 600 K). At this temperature, the
energy would be considered of significantly “lower grade,” because the efficiency
of the subsequent thermal energy conversion process is low. This latter process
converts the heat into a more usable high-grade form of energy, such as electricity.
Thermoelectric generation of electricity using the Seebeck (thermocouple) effect
has worked successfully in a number of power packages for satellites. Thermionic
production of electricity directly inside reactor cores is again being pursued for
space power applications.
In contrast, fluorescence produced from the interaction of charged particles from
the radioactive decay of radioactive isotopes with a flourescer medium has been
used to produce low intensity lights for remote applications, such as runway lights.
These devices involve relatively low power output. The fluorescence production
process has some advantages which can be exploited by nuclear battery technology.
Fluorescence can be transported away from the radiation source to a shielded
transducer. This indirect photon based process involves a two-step method for
converting the energy of the charged particles from nuclear reactions into a usable
energy form such as electricity, chemicals, or coherent light. In the first step, the
ions produced by nuclear reactions transfer their energy to an intermediate photon
generator—a fluorescer medium. This step in the process uses a nuclear-driven
fluorescer or NDF (a medium which produces incoherent narrow-band electro-
magnetic radiation). In the second step of the process, the intermediate photons
from the nuclear-driven fluorescer are absorbed by a material, which converts the
photons into a useful high-grade energy form. The medium which converts the
fluorescence into a useful product is called the transducer. The fluorescer can be in
the form of a solid, a liquid or a gas. The transducer can be of many types
depending on the desired product. Examples include: A photovoltaic cell for the
production of electricity; a chemical medium that is photolytically transformed to a
useful chemical (e.g., hydrogen); or a laser gain medium which produces laser light.
3.2.4.1 Gaseous Fluorescers
Gaseous fluorescers have been extensively studied as lighting sources. Examples of

these include fluorescent lights which use a mercury vapor discharge to generate
254 nm light. The 254 nm light output interacts with phosphors to produce the
visible spectrum that is used for lighting. Other examples include sodium lamps
which use sodium vapor to generate the yellowish light emitted by street lamps.
These two example fluorescent light sources are limited by self-absorption because
they depend upon line emission.
PIDEC and RECS
In gases, the energy goes into ionization and excitation. In typical gas mixtures, the
ionization and excitation produces excited states in the atoms or molecules that
make up the gas. For low power density, the bulk of the excitations will occur in
neutral atoms and molecules. The line emission will be dispersed over a broad
range of the spectra from the ultraviolet to the infrared. Optical emitters based on
line emission are subject to self-absorption and will only be optically thin over a
limited size and pressure. The efficiency of a specific line with a narrow frequency
is typically less than 1 %, which is generally the theoretical limitation of a single
line (with few exceptions). One such exception is mercury, which can have high
efficiencies at very low power density and low pressure (up to 70 %). Another
exception is sodium which can have high efficiencies at low power density and low
pressure (up to 40 %). Both mercury and sodium are not acceptable for use in a
nuclear battery for several reasons. The low power density requirement means that
the size of the fluorescer would have to be very large in order to get a reasonable
power output. Secondly, the efficiency of the light source would decrease as the size
of the lamp increases due to the self-absorption effects of line emitters (meaning the
lower energy state of the emitter is sufficiently dense as to absorb photons created
by the spontaneous emission of an upper state). Mercury and sodium lamps are
useful for fluorescer bulbs or street lights, respectively, because these lamps are
scaled to a relatively small size with a low power density to avoid self-absorption,
thus they are optically thin. As the lamp scales, it becomes optically thick and loses
its efficiency.
Spectral Considerations for Excimer Emitters
The term excimer is short for excited dimers. A dimer is a short lived dimeric or
heterodimeric molecule formed by two atoms, where at least one of the atoms has a
completely filled valence shell (such as a rare gas). For example, rare gases can’t
form molecules in the ground state, but can form a molecule in an electronic excited
state. Examples of excimer molecules include the rare gas excimers Ar2* (where the
“*” indicates an excited state), Kr2* and Xe2*, the rare gas-halide excimers ArF*,
KrF*, XeF*, ArCl*, KrCl*, XeCl*, etc. and a number of other excimer gas com-
binations [40]. When an excimer molecule decays by spontaneous emission, a
photon is given off by the bound excited molecular state to an unbound state where
the atoms making up the molecule become neutral and independent. In an excimer
gas mixture (a mixture of gases that form the excimer state), both ionization and
excitation contribute to the formation of the excimer states (with about 50 % effi-
ciency for photon production for rare gas excimers). Excimers thus depend upon the
formation of ions and metastable states by the interaction of ionizing radiation with
a gas. In an excimer forming gas, if the pressure of the gas is high enough (usually
greater than a half atmosphere), the formation of excimers is favored over formation
of atomic excited states (atomic excited states lead to undesirable line emissions
from the atom). Excimers emit in a very narrow wavelength range (±10 nm).
Additionally, excimers do not have a bound ground state and thus are not subject to
self-absorption. Thus, excimer gas mixtures remain optically thin (meaning no
self-absorption) over large size, large power density, and high pressure.
If the excimer is a rare gas excimer, like the xenon excimer, the bulk of the
energy that goes into the formation of xenon ions and xenon metastable states
proceeds into the formation of the xenon excimer state. From Table 3.9, it takes
21.9 eV to form an ion pair. The W* value for xenon metastable state formation
(42 eV per metastable state) is also known [40]. The theoretical maximum effi-
ciency for xenon excimer production (ηf) is the ratio of the xenon excimer photon
energy (7.2 eV) divided by the W value for ion pair production plus the ratio of the
xenon photon energy divided by the W* value for Xe metastable production shown
in Eq. (3.69).
7:2 7:2
gf ffi þ ¼ 0:5 ð3:69Þ
21:9 42
Thus the theoretical maximum efficiency of an energy conversion method using

xenon excimer formation is about 50 %. This value of 50 % is approximately
correct for all rare gas excimers.
Experimental measurements of excimer fluorescence efficiency are difficult due to
the fact that typical research used high flux nuclear reactors as the power source to
generate reasonable power densities in the excimer gas. One example is the calibrated
measurement of the excimer fluorescence efficiency from xenon gas [41]. The goal of
this experiment was to measure vacuum ultra violet (VUV) fluorescence production
efficiencies from a nuclear-pumped xenon excimer system. The experiment was
conducted using the F beam port of the University of Missouri Research Reactor
(also known as MURR) as a source of neutrons to pump the B10(n, a)Li7 reaction.
Several constraints had to be considered: (1) a vacuum was required along the light
path for the VUV radiation produced by excimer gas; (2) the detector needed to be
shielded from the radiation emitted from port opening so a turning mirror was nec-
essary; and (3) the production of activation products had to be minimized. The
constraints led to the experimental design shown in Fig. 3.25. The use of the F beam
port posed an interesting design problem. Primarily, the limited space of the port
(1.2 in. inner diameter) presented the most significant challenge. Additionally, it was
Fig. 3.25 Experimental position in the F beam port in the MURR reflector
necessary to measure the neutron flux distribution along the length of the active area
of the port since this data was not known.
The flux was measured by using a hollow 6061 aluminum alloy tube. A rod with
a very thin research grade gold wire wrapped around it was placed in the hollow
aluminum tube. Additionally, a copper-constantine thermocouple wire was moun-
ted at both ends and in the middle of the hollow tube. The thermocouple was used
to measure the temperature profile of the hollow tube.
The hollow aluminum alloy tube was placed in the active area of the assembly
and the optical light path tube and the reaction chamber assembly was vacuum
tested before being set in beam port F. A pump down procedure was developed
where the beam port tube was evacuated and filled with He. The optical path was
vacuum checked. The reactor was then brought up to a power level of 500 kW. The
temperature was measured and a profile indicating a gradient from the reaction
chamber nearest the reactor to the opposite end of 50°F was recorded (T = 130 °F
to T = 80.0 °F). The reactor power level was monitored as well as the exposure
time during the run. After the run, the gold wire was taken out and cut up into
pieces—0.5 cm long and the position of each piece in the reaction chamber was
recorded. Each piece was analyzed for the amount of the gold isotope with an
atomic mass of 198 (created by the thermal neutron activation reaction,
Au197(nthermal, c)Au198). The amount of Au198 is proportional to thermal neutron
flux. So the with the measured amount of Au198 in each piece of wire, the thermal
neutron flux profile was found.
The optical system’s major components were the reaction chamber, the vacuum
tight optical path, and the turning mirror assembly. Of the major components, the
reaction chamber and the turning mirror assembly were the most difficult to design.
Given the limited space of the beam port tube, a unique method of filling a
reaction chamber with a sample gas had to be developed. This step was accom-
plished by using a multiple seal on the end cap (see Fig. 3.26).
It was also necessary to develop a turning mirror assembly which could be
moved in the X-Y-Z direction and which maintained vacuum integrity. This task
was accomplished by using a bellows mount and a rotational feed through (see
Fig. 3.27).
The optical system required a preliminary test. An experiment was designed and
performed that used helium gas which has no possible activation products. In order
to calibrate an optical system, either a calibrated light source or a calibrated detector
is required. In this experiment, a calibrated Sampson type UV lamp was used to
calibrate the optical system. An additional experiment was run to find the time it
would take to measure a reasonably high resolution spectra from helium using a
strip chart recorder (note that a motor was used to move the grating in the Acton
vacuum ultra violet spectrometer). From this experiment, it was determined that a
spectra could be successfully taken within a 3 min period. The helium spectra that
was taken matched up very well with a helium spectra taken in previous work [42].
Of the excimer fluorescers, xenon has the most hazardous activation products.
These products include Xe125 and Xe133m. The activation products accumulated
Fig. 3.26 Reaction chamber used for fluorescence experiments
during a 5 min run, if released, were found to be within the reactor’s maximum
permissible concentration limits.
The xenon excimer experimental was run using 1 atm sample of Xe The effi-
ciency for the production of xenon excimer fluorescence at 172 nm was measured
and found to be 49 ± 10 %. This agrees well with the theoretical efficiency of
50 %. In general the nuclear-pumped excimer systems appear to operate very close
to their theoretical maximum efficiency [40]. It should also be noted that under no
circumstance should the excimer efficiency exceed the theoretical maximum
efficiency.
Fig. 3.27 Diagram of the vacuum ultra violet turning mirror assembly
Fig. 3.28 Energy structure of

the xenon molecule [44]
Effect of Impurities on Excimer Efficiency

Impurities will cause parasitic reactions which take energy away from the excimer
channel. The dominant reactions in forming the xenon excimer (Fig. 3.28) are 1, 7,
11, 13, 14, 22 and 36 from Table 3.10 [43].
Experimental data matches well with the model, as shown in Fig. 3.29. There are
parasitic reactions due to impurities which can channel energy away from the Xe*
and Xe+ states. These are:
Xe þ M ! Xe þ M ðor M þ Þ ð3:70Þ
where M is the impurity atom or molecule, M* is the excited state of the impurity
atom or molecule and M+ is the ionized state of the impurity atom or molecule.
Xe þ þ M ! Xe þ M þ ð3:71Þ
Even research grade purity rare gases have some impurity content, as shown in
Table 3.11. For example, if a chamber were filled with one atmosphere of xenon,
there are 2.44 1019 xenon atoms-cm−3. One part per million of an impurity will
represent 2.44 1013 impurity atoms or molecules-cm−3. These impurities will
siphon energy from the desired reactions that form excimers. A kinetic study
performed by one of the authors indicates that when impurities are at the ppm level,
the energy siphoned away from excimer formation can be significant when the
pumping power density is below 1 mW-cm−3. As the power density increases, the
siphoning effect of the impurity “burns out” and the percentage of energy siphoned
away from excimer formation decreases. The “burn out” effect is due to slower
Table 3.10 Reactions used in the Xe model [43]

No. Reaction Rate constant
(s−1, cm−3 s−1, or cm−6 s−1)
1 Xe* + 2Xe ! Xe2*(1R) + Xe 1.7 10−32
2 Xe** + Xe ! Xe* + Xe 1.0 10−10
3 Xe* + Xe ! Xe + Xe 3.5 10−15
4 Xe** + 2Xe ! Xe2** + Xe 1.0 10−31
5 Xe2** + Xe ! Xe* + 2Xe 1.0 10−11
6 Xe2*(1R) + Xe ! Xe2*(3R) + Xe 1.2 10−13
7 Xe* + Xe* ! Xe++Xe + e 2.0 10−9
8 Xe2*(1R) + Xe2*(1R) ! Xe2+ + 2Xe + e 5.0 10−10
9 Xe+ + 2Xe ! Xe2+ + Xe 2.5 10−31
10 Xe2+ + e ! Xe** + Xe 2.3 10−7
11 Xe2*(1R) ! 2Xe + hv 2.1 10−8
12 Xe** ! Xe* + hv 1.5 10−7
13 Xe2*(3R) + e ! Xe2*(1R) + e 1.8 10−7
14 Xe2*(1R) + e ! Xe2*(3R) + e 4.9 10−7
15 Xe2** + e ! Xe2*(3R) + e 5.0 10−7
16 Xe2** + e ! Xe2*(1R) + e 2.0 10−7
17 Xe2*(1R) + e ! 2Xe + e 1.3 10−9
18 Xe2+ + e ! Xe* + Xe 2.3 10−7
19 Xe* + 2Xe ! Xe2*(3R) + Xe 4.4 10−32
20 Xe2*(3R) + Xe2*(3R) ! Xe2+ + 2Xe + e 5.0 10−10
21 Xe2*(3R) + Xe2*(1R) ! Xe2+ + 2Xe + e 5.0 10−10
22 Xe2*(3R) ! 2Xe + hv 1.0 10−7
23 Xe2*(3R) + e ! 2Xe + e 1.3 10−9
24 Xe++Xe + e ! Xe* +Xe 1.0 10−26
25 Xe2+ + e + Xe ! Xe2*(1R) + Xe 1.0 10−26
26 Xe2*(1R) + e ! Xe2+ + 2e 5.00 10−9
27 Xe2*(3R) + e ! Xe2+ + 2e 5.0 10−9
28 Xe2+ + e + Xe ! Xe2*(3R) + Xe 1.0 10−26
29 Xe** + Xe ! Xe + Xe 1.0 10−15
30 Xe2*(3R) + Xe ! Xe2*(1R) + Xe 4.6 10−15
31 Xe** + e ! Xe* + e 8.0 10−7
32 Xe2*(3R) + e ! Xe2** + e 3.0 10−7
33 Xe** + Xe** ! Xe++Xe + e 5.0 10−10
34 Xe2** + Xe2** ! Xe2+ + 2Xe + e 5.0 10−10
35 Xe* + e!Xe + e 1.0 10−9
36 Xe++2e ! Xe* + e 1.0 10−20
37 Xe2** ! Xe* + Xe + hv 1.0 10−8
Fig. 3.29 Comparison of theoretical and experimental data for e-beam driven Xe2* fluorescence
efficiency at 3 atm. of Xe
Table 3.11 The impurities in research grade xenon gas (99.999 % pure) are shown
Contaminant Research Grade Xenon, 99.999 % pure (ppm)
Ar 1.0
CO2 1.0
CF4 0.5
H2 2.0
Kr 5.0
N2 2.0
O2 0.5
Total hydrocarbons 0.5
H2O 0.5
reaction rates that occur once an impurity ion is formed (like recombination or
further excitation). At power densities above 1 W-cm−3 the bulk of the energy
forms excimers. Thus, excimers lose efficiency due to siphoning at power densities
below 1 W-cm−3 because of interactions between rare gas metastable states or ion
states with low level impurities. However, at power densities above 1 W-cm−3,
excimers approach the theoretical maximum efficiencies of the gas because the
impurity ground states are depleted rapidly.
PIDEC
The pursuit of the production of electrical power from a nuclear light bulb was
suggested by one of the authors in 1981 [45]. The nuclear light bulb was based on
the Photon-Intermediate Direct Energy Conversion (PIDEC) process shown in
Fig. 3.30 Schematic diagram of the Photon-Intermediate Direct Energy Conversion (PIDEC)
process
Fig. 3.30, and was conceived as a method of energy conversion for a variety of
nuclear reactions including fission, advanced fuel fusion and radioisotope decay
[45–47]. Advanced-fuel high temperature fusion, however, will not be available in
the foreseeable future. Potential sources of ions from fission reactions and ra-
dioisotopes however, are viable near term energy sources. The nuclear-driven
fluorescer concept, which was initially suggested as a photolytic driver for a
nuclear-pumped laser in 1977 by the author, his fellow students and his mentor
George Miley, was the fundamental component of the nuclear light bulb [48]. This
concept makes feasible efficient ion-driven photon sources by removing the nuclear
source material from the solid, where it prevented transmission of the photons, and
moving it to the volume of the fluorescer gas by one of four methods; a gaseous
source, a source embedded in thin films of arbitrary shape and size or as an aerosol
source (Fig. 3.31) [40]. A critical component of the nuclear light bulb is that the
scale lengths of the solid sources are chosen small enough so that most of the
particulate energy escapes into the surrounding gas, creating a weak plasma which
produces photons. The coexistence of weak plasma with an aerosol introduces a
unique set of problems. Another critical aspect of the nuclear light bulb is that the
combination of the photon producing weak plasma and the aerosol source or thin
film source remains optically thin to allow the photons to transport out of the
system. As will be discussed, the nuclear light bulb can be made optically thin. The
nuclear light bulb in conjunction with radioisotopes led to the development of the
Radioisotope Energy Conversion System (RECS) [49], Fig. 3.32.
RECS utilizes a two-step method for directly converting the energy of the
charged particles from nuclear reactions into a usable energy form such as elec-
tricity, chemicals, or coherent light. In the first step the ions produced by nuclear
reactions transfer their energy to an intermediate photon generator—a fluorescer
medium. This portion of the RECS comprises what is called the nuclear-driven
Fig. 3.31 An illustration of the use of thin solid geometries which allow reaction products to
escape the solid matrix into a surrounding gas
fluorescer or NDF (a medium which produces incoherent narrow-band electro-

magnetic radiation). In the second step, the intermediate photons from the
nuclear-driven fluorescer are absorbed by a material, which converts the photons
into a useful high-grade energy form. This portion of the RECS is the photon
energy converter.
The efficiency of the two-step PIDEC process, used for the RECS has two major
advantages over thermal energy conversion, which is a many step process. These
advantages are: (1) that it is a direct process producing a high grade photon source
from a high grade ion source, thus avoiding the Carnot cycle efficiency limits
imposed by thermalization and (2) that it is mechanically simple, potentially leading
to more compact, more reliable, and less expensive systems.
The advantage of the two step process over a one-step direct energy conversion
process is that of feasibility. The scale length for the transport of the primary
high-grade ion source must match the geometrical scale of the energy converter.
Energetic ions have a transport length of micrometers while useful energy con-
verters, on the other hand, have a scale length of fractions of meters. For this reason,
direct conversion of nuclear energy has not previously been possible. What was
required (and provided by the nuclear light bulb concept) was an intermediate
high-level energy converter that can be intermingled with nuclear material on a
Fig. 3.32 Schematic diagram of the Radioisotope Energy Conversion System (RECS) which uses
the radioisotope in aerosol form
micrometer scale-length, but produces an energy form that can be transported to

meter scale-length direct converters producing useful output—a sort of “impedance
matching” for the scale length of energy forms. With PIDEC that scale length
matching medium is a fluorescing gas, the nuclear-driven fluorescer. The photons it
produces can be transported great distances, making it possible to couple them to
various energy conversion processes. Also, some conversion processes require
greater power densities than the primary energy sources can provide. PIDEC can
effectively implement optical concentrators, enabling achievement of the high
threshold power density required for a conversion process such as a photolytic
laser.
The RECS concept of high efficiency production of light from radioisotopes
coupled to wide band-gap photovoltaic cells makes the nuclear light bulb especially
useful for compact power supply applications.
Nuclear-Driven Fluorescers
Nuclear Driven Fluorescers (NDF) are photon producing media which directly use
the energy from ionizing radiation. Scintillator materials, for example, constitute a
medium which generates fluorescence. The reader may find references in the lit-
erature of the Former Soviet Union which refers to NDF media as scintillators [50].
Fluorescer media can be solids, liquids and gases [40].
The Ion Source

As shown in Fig. 3.32, one type of ion source can come from the decay of ra-
dioisotopes dispersed within a fluorescer gas. Effective dispersal is essential so that
the ions produced by the isotope decay deposit most of their kinetic energy in the
excimer gas rather than in the radioisotope material. There are several methods of
achieving the desired dispersal: gaseous radioisotopes [46], radioisotopes embed-
ded in thin films [51], microscopic aerosol of radioisotopes [47, 52, 53], or the atom
embedded in a C60 cage [54]. The efficiency of transport of the ion energy from the
radioisotope to the fluorescer medium varies with the scale length of the thin film or
aerosol (or in the case of a gaseous radioisotope the transport distance of the
particulate radiation in the gas), the chemical form of the radioisotope, and the
uniformity of the radioisotope density. The variation of ion energy transport effi-
ciency from a microsphere to the fluorescer medium with thin films and micro-
spheres is discussed in the work of Chung and Prelas [51]. Energy transport
efficiencies are about 60–70 % for reasonably designed thin films and 70–80 % for
reasonably designed microspheres. The average atomic density in the medium must
be on the order of 1 1019 particles cm−3, enough to achieve reasonable power
densities but not so great as to significantly degrade the transport of the fluorescence
through the aerosol. Combining the constraints of efficiency and optical trans-
parency determines scale length of the thin film or microspheres, as well as the
number density. For example, a microsphere diameter of 5 lm and number density
of 1 106 cm−3, which should not create significant absorption of the fluorescence
[53, 55], results in a fuel density of 0.63 mg cm−3, quite reasonable dimensions,
and a good average number density (3.9 1019 atoms cm−3) of the radioisotope. It
is also possible to enhance the average number density by coating the aerosol
particles with a thin reflective optical material. An alternative, which couples
essentially 100 % of the ion energy to the fluorescer medium, is a gaseous
radioisotope such as Kr-85 [9, 10].
Excimer Fluorescers
Excimer fluorescers are the most efficient optical radiators known and, because of
their unbound lower levels, do not self-absorb. They radiate in a single, relatively
narrow, band of wavelengths and are suitable for efficient photovoltaic energy
conversion. The intrinsic fluorescence efficiencies of rare-gas and rare-gas halide
excimers, based on standard W-value theory, are listed in column 2 of Table 3.12.
Achievable efficiencies should be near the intrinsic values at the power and electron
densities characteristic of nuclear reactions.
When combined with photovoltaic cells as the transducer, the combination of
excimer fluorescence and PV cells can have high intrinsic efficiencies. This is
shown in Table 3.13.
Experimental data on a number of excimer fluorescers show that the efficiencies
match W-value theory. In fact, one group has reported measuring a nuclear-driven
rare-gas excimer fluorescence efficiency higher than that predicted by W-value
Table 3.12 Excimer fluorescence efficiencies

Excimer k (nm) ηmax Excimer k (nm) ηmax
NeF* 108 0.43 I2 343 0.24
Ar2 129 0.50 XeF* 346 0.24
Kr2 147 0.47 Kr2F* 415 0.17
F2 158 0.44 Na2* 437 0.46
Xe2 172 0.48 HgI* 443 0.19
ArCl* 175 0.48 Li2* 459 0.42
KrI* 185 0.37 Hg2* 480 0.21
ArF* 193 0.35 HgBr* 502 0.17
KrBr* 206 0.33 KrO* 547 0.15
KrCl* 222 0.31 XeO* 547 0.15
KrF* 249 0.28 HgCl* 558 0.15
XeI* 253 0.37 K2* 575 0.42
Cl2 258 0.32 Rb2* 605 0.41
XeBr* 282 0.29 CdI* 655 0.13
Br2 290 0.29 Cs2* 713 0.37
XeCl* 308 0.27 CdBr* 811 0.10
Table 3.13 Efficiencies for selected rare-gas and rare-gas halide excimer fluorescers (ηf) with
well-matched wide-band-gap photovoltaic materials
Excimer ηf Ek Photovoltaic Band-gap ηpv = Eg/ ηie = ηpvηf
(eV) material energy (eV) Ek
Ar2* 0.5 9.6 AlN 6.2 0.645 0.324
Kr2* 0.47 8.4 AlN 6.2 0.789 0.345
0.47 8.4 Diamond 5.5 0.655 0.308
Xe2* 0.48 7.2 AlN 6.2 0.861 0.413
0.48 7.2 Diamond 5.5 0.764 0.367
ArF* 0.35 6.4 AlN 6.2 0.969 0.339
0.35 6.4 Diamond 5.5 0.859 0.301
KrBr* 0.33 6 Diamond 5.5 0.917 0.302
KrCl* 0.31 5.6 Diamond 5.5 0.982 0.304
Na2* 0.46 2.84 ZnSe 2.7 0.951 0.437
0.46 2.84 SiC 2.4 0.845 0.389
Li2* 0.42 2.7 CuAlSe2 2.6 0.963 0.404
0.42 2.7 SiC 2.4 0.889 0.373
The theoretical maximum intrinsic efficiency for the match of the given excimer spectra to the PV
cell band-gap (or maximum value for ηin) is ηpv, and the ion-to-electric efficiency, ηie
Ek (eV) is the average photon energy Eg is band-gap energy of the photovoltaic material
theory at 68 % [56]. However, this is clearly an outlier. All of the other experiments
with a variety of excitation sources (e.g. electrons, fission fragments, protons) and
particle densities have given fluorescence efficiency values near the theoretical
limits predicted by W-value theory [40]. The most efficient excimer fluorescers are
the rare-gas excimers.
The Photon Energy Converter
The key to the feasibility of the RECS is the photovoltaic Photon Energy Converter.
The common impression of photovoltaics is that they cannot be very efficient. This
misunderstanding comes from the fact that photovoltaics are most commonly
employed as “solar cells.” Solar cells are not very efficient, ranging from 10–20 %
for commercial units and reaching as high as about 25 % for laboratory cells.
However, the low efficiency is more due to the characteristics of the solar spectrum
than to the photovoltaic devices themselves, especially for the laboratory units with
efficiencies of *25 %. The problem with the solar spectrum is that it is very
broadband—its ratio of the average photon energy to the width (FWHM) of the
spectrum (Emean/DE) is about 1. This is good for color vision but quite bad for
efficient energy conversion. For excimers, however, this ratio is greater than 10.
Under these conditions photovoltaics have intrinsic efficiencies of 75–95 %.
Photovoltaic cells for use in photon-intermediate direct energy conversion of
electricity will require the development of a doped semiconductor material with a
band-gap that matches the UV radiation. With wide band-gap photovoltaic cells,
fusion ion driven fluorescence in a PIDEC system was studied [45, 57]. Fission ion
driven fluorescence sources in a PIDEC system were studied as well [58].
Photovoltaic Conversion of Narrowband Fluorescence

For a given spectrum, the efficiency of conversion is basically determined by the
variation of the irradiance with photon energy and by the substrate band-gap
energy, Eg, of the photovoltaic converter. Complete conversion (100 %) is not
possible because of the width of the solar spectrum. This leads to two competing
effects on the efficiency. The first effect is that the energy of all photons with
quantum energy hm < Eg is lost because they do not have sufficient energy to excite
electrons from the valence band to the conduction band. The power density lost in
this case is given by
Z Eg
Plost ¼ WðEÞdE ð3:72Þ
0
where W(E) is the irradiance in W/cm2-eV.

Thus, the lower the band-gap of the photovoltaic converter, the larger the
fraction of the total spectrum converted. Competing with this effect however is the
fact that, for the photons with quantum energy hm > Eg that do contribute, the
Fig. 3.33 Efficiency of photovoltaic conversion of broadband solar radiation using photovoltaics
with Eg = Emin. Data points are fraction of total number of photons with an energy greater than E
versus photon energy normalized to the average photon energy
Fig. 3.34 Efficiency of

photovoltaic conversion of
broadband solar radiation
using Photovoltaics with
Eg = Eav. Data points are
fraction of total number of
photons with an energy
greater than E versus photon
energy normalized to the
average photon energy
photon energy in excess of the band-gap energy is lost. Thus the maximum intrinsic
efficiency for photovoltaic conversion, assuming an ideal collection device, is
R1 E
Eg W ðEi Þ Egi dEi
gin ¼ R 1 ð3:73Þ
Eg W ðEi ÞdEi
where Ei is the impact ionization energy.

These two effects are shown graphically in Figs. 3.33 and 3.34 for broadband
(solar) and in Fig. 3.35 for narrowband (Xe2*) radiation. In Figs. 3.33 and 3.34 it
can be seen that for a broadband spectrum, such as the solar spectrum, there is no
value of Eg that results in most of the energy being converted. In Fig. 3.35, on the
Fig. 3.35 Efficiency of

photovoltaic conversion of
narrowband excimer radiation
using high bandgap
photovoltaics. Data points are
fraction of total number of
photons with an energy
greater than E versus photon
energy normalized to the
average photon energy
other hand, it can be seen that, for Eg just less than Emin, efficient energy conversion
takes place.
Typically, the effect of the details of the solar radiation spectrum on calculating
overall conversion efficiency is translated into a photon flux density, which then
relates to an ideal short circuit current density. This is convenient because it is a
good assumption that each photon absorbed and collected effectively causes one
electron to move around the circuit. Also, after each electron thermalizes (gives off
energy in excess of Eg to the lattice), it contributes the maximum energy amount Eg
to the overall process.
This contribution is conveniently modeled in the photovoltaic device using the
ideal Shockley model for the p-n junction. Using these concepts, the intrinsic
conversion efficiency can be written in the following terms,
R1
Eg Eg Nph ðEÞdE Eg N E [ Eg Eg geg
gin ¼ R 1 ¼ ¼ ð3:74Þ
Eg Nph ðEÞEdE
Emean Ntotal Emean
where Nph is the photon flux density in #/s-cm2-eV, N(E > Eg) is the photon flux in
the interval E > Eg, Ntot is the total photon flux, and ηeg is the fraction of photons
with E > Eg.
The solid lines in Figs. 3.36 and 3.37 show, respectively, the air mass 2 (AM2)
irradiance of the solar spectrum (W/cm2-eV) [11] and the irradiance for Xe2*, a
narrow-band fluorescer. For Xe2*, Emean/DE = 14, compared to a corresponding
value of 1.3 for the AM2 solar spectrum. For a narrow distribution one can have Eg/
Emean * 1 and still have ηEG * 1. A narrow-band spectra will consequently have
the highest intrinsic efficiency. The crosshatched areas in Figs. 3.36 and 3.37
represent, respectively, the power density converted for the solar spectrum and for
the Xe2* spectrum. The ratio of the cross hatched area to the total area in both cases
gives the intrinsic efficiency. The white area under each curve corresponds to
Fig. 3.36 The spectral irradiance of the solar spectrum (AM2). The cross hatched area under the
curve is the fraction of energy which is usable with a 1.1 eV bandgap photovoltaic cell
Fig. 3.37 The normalized irradiance of Xe2* excimer source. The cross hatched area under the
curve is the fraction of energy which is usable with a 6 eV bandgap photovoltaic cell
unconverted power. These figures illustrate that a large portion of the energy in a
broad band source, such as a solar source, is wasted while a much smaller fraction
of the energy in a narrow band source, Xe2*, is wasted.
Figure 3.38 is a plot of maximum efficiencies for a p-n junction converter versus
band-gap energy of the converter substrate material. Two plots are shown, one for
an AM2 solar spectrum and one for a Xe*2 spectrum. The previously derived
equations were used to calculate these curves. Also shown are vertical lines rep-
resenting the band-gap energies of the two materials theoretically predicted to
maximally convert these two spectra. Lines representing Si and diamond are shown
for comparison. The curve for the solar spectrum lies in between the two best
Fig. 3.38 The theoretical

maximum conversion
efficiency of photovoltaic
cells with various bandgap
energies using either a solar
spectrum or a Xe2* spectrum.
Band-gap energies associated
with Si, GaAs, diamond, and
AlN are shown
known curves of this type [40]. The approximately 30 % maximum is thought to be

an upper bound on the ability of a single material junction to convert the solar
spectrum. The highest known conversion efficiency for silicon, to date, has been
26 % obtained with a highly optimized MIS solar cell [40].
In contrast to the relatively low values for conversion of the solar spectrum, it
can be seen from Fig. 3.38 that efficiencies as high as 80 % can theoretically be
obtained using a p-n junction and converting the Xe*2 spectrum. The efficiencies in
Fig. 3.38 assume a fill factor of 1. A fill factor of one is an optimistic assumption.
However, a fill factor of 0.8 is reasonable [59].
Wide Band-Gap Photovoltaic Materials
Table 3.14 lists several potential wide band-gap materials. One material, diamond
(5.4 eV), has recently received intense study as a microelectronic device substrate.
However, the most efficient excimer UV radiation sources, the rare-gas excimers,
have larger photon energies (>7 eV) and hence the development of higher band-gap
photocells for these excimers would be desirable. Table 3.13 matches the more
efficient and desirable fluorescers from Table 3.12 to materials with appropriate
band-gaps from Table 3.14. The theoretical maximum intrinsic photovoltaic effi-
ciency (the ratio of the band-gap to the mean photon energy) ranges from 75 to
95 %) while the corresponding theoretical maximum efficiency for conversion of
ion energy to electrical energy (the product of the photovoltaic efficiency and the
fluorescence efficiency) ranges from 30 to 45 %.
The outlook for wide band-gap photovoltaic cells is hopeful since p-n junctions
are available in some materials that are commercially available (e.g., SiC and GaN)
and p-n junctions have been made in these materials. Progress has also been made
in forming p-n junctions in other wide band-gap materials (e.g., diamond) [24].
It is possible to find flourescers that have photon energies that are lower than rare
gas excimers. For example, rare-gas halide excimers have lower photon energies
(3.5 eV for XeF*, 5.0 eV for KrF*, and 6.4 eV for ArF*) and, while their
Table 3.14 Band-gap Material Band-Gap (eV) Bandgap energy (eV)

energies for various
semiconductors Si 1.1 1685
CdSe 1.44 1370
ZnSe 2.26 1510
3C-SiC 2.36 3103
6H-SiC 3.05 3103
4H-SiC 3.21 3103
GaN 3.4 2770
ZnS 3.54 1920
UO2 5.2 3150
Diamond 5.5 100*
AlN 6.02 2500
*Graphitization Temp
fluorescence efficiency may be lower than that of the rare-gas excimers, their
photon energy falls in the range of commercially available p-n junctions in some
semiconductor materials (Table 3.15).
PIDEC is an indirect energy conversion scheme which uses high efficiency
fluorescers to convert the energy from ionizing radiation to photons. These photons
are then transported to a transducer (photovoltaic cell where the band-gap energy is
matched to the photon energy) to produce electrical current. A great deal of work
has been developed around PIDEC in which different geometrical interfaces
between the ionizing radiation and an excimer emitter have been examined. PIDEC
was developed to overcome the limitations which were identified in the 1950s for
p-n junction based betavoltaics and alphavoltaics (e.g., radiation damage, short
lifetime, and low efficiency). PIDEC addresses some important issues; it is a way of
matching the scale length of the radiation source to the transducer (the fluorescer
can absorb all of the energy from the radiation source by adjusting size and pres-
sure). It has high theoretical conversion efficiencies and provides a means of
shielding sensitive transducers from radiation. Typical theoretical maximum system
efficiencies can approach the theoretical limitation of rare gas excimers fluorescers
as shown in Table 3.13 where ηf is the fluorescence efficiency, Ek (eV) is the
average photon energy and Eg is band-gap energy of the photovoltaic material. The
variation in efficiencies is based on design parameters (e.g., interface between the
fluorescer gas and radiation source, the system size, photon transport to the
transducer, the operational characteristic of the p-n junction and matchup of
transducer to photons). PIDEC was proposed for use with fusion [57], fission [58],
and conventional decay sources [46, 60, 61].
Table 3.15 Properties for some commercially available wide band-gap materials are shown in this Table (3.20)
146
Property GaP 3C-SiC 6H-SiC 4H-SiC GaN ZnO Diamond AlN BN

Crystal structure Zinc blende Zinc blende Wurtzite Wurtzite Wurtzite Wurtzite Diamond Wurtzite Zinc
(Cubic) (Cubic) blende
Group of symmetry T2d-F43 m T2d-F43 m C46v-P63mc C46v-P63mc C46v- C46v- O7h- C46v- T2d-F43m
P63mc P63mc Fd3m P63mc
Number of atoms per cm3 4.91022 8.91022 1.71023 9.61022
Debye temperature K 445 1200 1200 1300 600 1860 1150 1700
Density g cm−3 4.14 3.166 3.21 6.15 5.642 3.515 3.255 3.48
Dielectric const. (static) 11.1 9.72 9.66 9.66 8.9 8.75 5.7 9.14 7.1
Dielectric const. (high freq.) 9.11 6.52 6.52 6.52 5.35 3.75 4.84 4.5
Effective longitudinal electron 1.12mo 0.68mo 0.20mo 0.29mo 0.20mo 1.40 mo 0.4 mo 0.35mo
mass ml
Effective transverse electron 0.22mo 0.25mo 0.42mo 0.42mo 0.20mo 0.36 mo 0.24mo
mass mt
Effective heavy hole mass mh 0.79mo 1.4mo 2.12mo 3.53mo 0.37mo
Effective light hole mass mlp 0.14mo 0.3mo 0.70mo 3.53mo 0.150mo
Electron affinity eV 3.8 4.1 −0.070 0.6 4.5
Lattice constant angstrom 5.4505 4.3596 a = 3.0730 a = 3.0730 a = 3.189 a = 4.75 3.567 a = 3.11 3.6157
b = 10.053 b = 10.053 c = 5.186 c = 2.92 c = 4.98
Optical phonon energy MeV 51 102.8 104.2 104.2 91.2 160 99.2 130
Band-gap eV 2.26 2.26 3.0 3.3 3.5 3.37 5.47 6.2 6.2–6.4
Breakdown V MV/cm *1.1 *2 *3 *3 *3 1–10 1.2 2
Electron mobility cm2V−1s−1 250 1000 380 800 300 80 2200 300 200
Hole mobility cm2V−1s−1 150 50 40 140 350 2000 14 500
Melting Point C 1457 2830 2830 2830 2500 1977 4373 3273 2973
Thermal conductivity 1.1 4.9 4.9 4.9 1.3 0.54 20 2.85 7.4
Wcm−1C−1
3 Interactions of Ionizing Radiation with Matter …
Hardness mohs scale 5 9.2 9.2 9.2 4 10 9.5

Gas
If a gaseous volume source is used, then it is possible to scale a PIDEC cell to a

reasonable power density. In this section, an example based on a high pressure
Kr-85 is used. Kr-85 uses the beta particles it releases to self excite the gas and
produce excimer photons [9]. This example was chosen because it illustrates how
effective a volume source can be.
If the PIDEC cell volume is 1 l at 1000 atm, it will contain about 3450 g of
Kr-85.1 Kr-85 produces about 0.51 W/g (see Chap. 1). The absolute maximum
power density produced by 1000 atm of Kr-85 per cm3 is 0.51 W/g times 3.
45 g/cm3 which is 1.76 W/cm3. The total power deposited in the 1 l volume with
1000 atm of Kr-85 is about 1,760 W. Using the Kr-85 as an intermediate transducer
that produces fluorescence, the efficiency of converting the power deposited in the
cell into fluorescence is 50 %. If the PIDEC cell uses a SiC photovoltaic cell as a
transducer, the approximate efficiency of the photovoltaic cell in converting a Kr
excimer photon to electrical power is approximately 16 % [60]. If the SiC photo-
voltaic cells are perfectly coupled to the fluorescence source, the PIDEC cell will
produce about 140.8 W or a power density of about 0.1408 W/cm3. This is an
impressive power density, but as discussed in Chap. 2, there is a world-wide in-
ventory of *373,000 g of Kr-85 [62] which is enough for only 108 units.
Solid Sources that Can Mimic Gas

Solid radioisotopes can effectively mimic gaseous like behavior by mixing the solid
radioisotope in aerosol form with a gas or by embedding the solid radioisotope in
thin fibers or films that hang in the gas (Figs. 3.31 and 3.39). The average density of
radioisotopes in the mixed phase systems can be very high. These type of mixed
phase systems can approach the maximum possible power density of a true volume
source [10].
Solid-State
By focusing on the matching of scale lengths for the radiation transport process to
the transducer in configurations which limit radiation damage, improvement in
efficiency is possible. One example which might bear interesting results is the use of
solid-state fluorescers which have limited or no self-absorption of the photon
1
The mass of Kr is estimated using Van der Waals equation of state to describe the relationship
between pressure and temperature as shown in the equation below, where n is the number of moles
of Kr, a and b for Krypton are 0.5193 Pa m6/mol2 and 0.000106 m3/mol respectively, R is the
universal
h gasiconstant (8.3145 J/mol K) and T is temperature in Kelvin.
2
P þ a Vn 2 Vn b ¼ RT.
Fig. 3.39 Schematic diagram of a PIDEC nuclear battery which uses a radioisotope fuel in
aerosol form [10]
emission. This differs from the discussion of solid-state scintillator research (e.g.,
polymers, phosphors, liquid scintillators, etc.). Scintillators were developed for
radiation detection in situations where the dose rates were orders of magnitude
lower than nuclear batteries. In prior nuclear battery research the existing scintil-
lators were adopted for use in nuclear batteries. The self-absorption problems in
these types of line sources have not been adequately considered as the power
density in the scintillator scales up. The solid-state scintillator lifetime is thus short
with strong ionizing radiation sources [63–67].
The self-absorption issue was considered for solid-state emitters by focusing on
direct wide-band gap materials and excitons [68]. In this approach, the radiation
source is embedded in a diamond crystal. As the radiation interacts with the dia-
mond crystal lattice, the electron-hole pairs which form are weakly bound and
create an exciton (with a binding energy of *70 mV). Provided the crystal tem-
perature is on the order of liquid nitrogen temperatures, the electron-hole pairs
diffuse together and when they recombine, they do so directly without the need for a
phonon. The photon energy is about 5.1 eV. The photon will efficiently exit the
crystal and be transported to a photovoltaic cell that has a band-gap that matches
well with the photon energy (e.g., AlxGa1−xN band-gap can be tailored from 3.4 eV
(xAl = 0) to 6.2 eV (xAl = 1) [69]). The Solid-State Electrical Generator using
Radionuclide Induced Exciton Production (SEGRIEP) was conceived of for space
exploration. Due to diamond’s high thermal conductivity and the background
temperature of space being 4 K, it was shown that it is feasible to build a radiative
cooling system to maintain liquid nitrogen temperatures in the diamond crystal.
3.2.5 Solid-State Emitter and PV
Diamond is not a direct band-gap material but does have a bound exciton which can
be used like a direct band-gap emitter. However, there will be no self-absorption of
the exciton photon because the photon energy, 5.1 eV, is less than the band-gap of
diamond, 5.49 eV. The binding energy of the electron-hole pair that makes up the
exciton is 70 mV. This device will have temperature limitations that need to be
explored. The theoretical maximum efficiency for this configuration is 33 %.
An approach being studied by the authors that is similar to the SEGRIEP concept
is to use solid-state emitters based on high quality binary solid state crystals which
exhibit wide band-gaps and direct band-gap transitions. In a direct wide band-gap
binary material, the photon self-absorption and reabsorption processes are in bal-
ance until the photon escapes the solid. Loss processes such as luminescence
emission from the surface and Auger recombination can be limited by proper
design. The photon can escape through a loss cone that is coupled to a photovoltaic
cell transducer (Fig. 3.40). Ionizing radiation will create displacements in the
solid-state crystal. The rate of displacements will be on the order of 170 dis-
placements per ion fragment. The number of photons created by each fragment
(with estimated energy of 10 MeV) is on the order of 2 million (fraction of energy
into electron-hole pair formation (0.42) times the energy of fragment
(10,000,000 eV)/band-gap of semiconductor (2.2 eV for GaP)). Thus, the rate of
photon generation exceeds the rate of potential trap formation by a factor of 20,000.
The potential traps do continue to build up with time. However, if the device
operates at a temperature where self-annealing can occur (600–800 K), there will be
Fig. 3.40 The solid-state material interacts with radiation and produces an electron-hole pair. The
pair recombines and produces a photon. The photon is then reabsorbed to form another
electron-hole pair or to reflect off the surface. If an electron-hole pair is formed, it recombines and
produces a photon. The process is in balance with few other losses and continues until the photon
is lost through the loss cone into the PV cell. The theoretical maximum efficiency for this
configuration is 33 %
a rate at which point defects (displacements) are repaired at a sufficient rate to limit
the effects of displacements due to radiation damage. It is feasible to use this
balance of defect creation and defect repair to extend the lifetime of the solid-state
emitter. This device will still have radiation damage issues. Defects will be created
in the emitter and the traps that are formed can absorb photons and electrons. The
key to the extended lifetime of the emitter is through self-annealing to mitigate trap
formation. Displacement issues are more serious with binary materials such as, III–
V, than with diamond (used in the SEGRIEP concept). The physics of the process is
still being researched and refined.
3.2.5.1 Phosphors
A number of phosphors have been developed for use in CRT based color TV and
radiation detectors. The progress in materials is discussed.
Materials may emit light by incandescence (all atoms radiate) or luminescence
(where a few atoms called emission centers radiate) when excited by electron beams
or by other ionizing radiation (e.g., alphas, betas, neutrons or gamma rays). The
mechanism is based on the interaction of ionizing radiation with matter [70]. The
ionizing radiation interacts with matter by generating ions and excitation in the
medium. The ions eject highly energetic electrons (primary electrons) which are
further multiplied through collisions, creating secondary and higher order electrons.
These energetic electrons excite the light-emitting centers which create the photons.
The class of nuclear batteries discussed in this section uses solid phosphors.
When energetic electrons are incident on a solid surface, a small fraction of the
electrons are scattered and reflected while the bulk of the electrons continue into the
solid and create electron-hole pairs. The average energy (Eav) required to form an
electron-hole pair near the band edges is given for both the direct or indirect
band-gap by an empirical formula,
Eav ¼ 2:67Eg þ 0:87 ½eV ð3:75Þ
where Eg is the band-gap energy for either director or indirect band-gap materials.
Inorganic phosphors have defects in the crystal structure such as dopants (or
activators) or dislocations which create the emission center. The wavelength that is
emitted is material dependent. The majority of phosphors lose efficiency over time
due to several factors. The energy input into the phosphor can cause the activator to
undergo a change of its valence state, the crystal lattice degrades due to radiation
induced displacements, the activators can diffuse through the material, and the
surface can oxidize thus creating a layer which absorbs either the excitation particle
or the photons emitted by the phosphor.
Phosphors have been used in a number of applications such as lighting (e.g.,
fluorescence tubes and white light LEDs), phosphor thermometry (measuring
temperature), glow-in-the-dark toys, radioluminescence (luminescence from inter-
actions with ionizing radiation), electroluminescence (LCDs), and Cathode Ray
Table 3.16 Typical high Material Energy efficiency Peak wavelength (nm)
efficiency phosphors [71]
Zn2SiO4:Mn2+ 8 525
CaWO4:Pb: 3.4 425
ZnS:Ag,Cl 21 450
ZnS:Cu,Al 23, 17 530
Y2O2S:Eu3+ 13 626
Y2O3:Eu3+ 8.7 611
Gd2O2S:Tb3+ 15 544
CsI:Tl+ 11 Green
CaS:Ce3+ 22 Yellowish green
LaOBr:Tb3+ 20 544
Tubes (CRTs). Phosphors that have been developed for use in CRTs or radiolu-
minescence are typically targeted for nuclear battery research based on their
resistance to radiation damage. Some example phosphors with high efficiencies are
shown in Table 3.16.
The more efficient phosphors in Table 3.16 have been used in nuclear battery
research. For example, ZnS:Cu emits in the green wavelength [72] as shown in
Fig. 3.41, and Y2O2S:Eu emits in the red wavelength [73] as shown in Fig. 3.42.
A common cell design is based on rectangular geometries. An example of a cell
design which demonstrates the main issues involved is shown in Fig. 3.43. Here, a
thin coating of a beta emitting isotope is placed at the center of the cell. The isotope
layer has a thickness “d” which should be thin enough to allow as much of the
isotropic beta emission to exit the radioisotope layer and enter the phosphor layer.
Optimizing d requires the use of modeling which take into account the beta
emission spectrum (Appendix A) as well as the angle of emission. The cell is
symmetric so that beta particles with both up and downward vector components
Fig. 3.41 The emission

spectra for the ZnS:Cu
phosphor. The emission
ranges from about 400–
650 nm with a peak at
530 nm
Fig. 3.42 The emission

spectra for the Y2O2S:Eu
phosphor. The emission
ranges from about 570–
650 nm with a peak at
626 nm
Fig. 3.43 Diagram of an

indirect nuclear battery which
uses a Pm-147 beta source
and a ZnS:Cu phosphor which
interacts with the beta
particles and produces
photons. The photons then are
transported to photovoltaic
cells [73]
have a good probability of entering the phosphor. This typically requires that “d” is
less than the range of the beta particle in the Pm-147 layer (Rb). The beta particle
must then travel through the phosphor and interact. The phosphor layer has a
thickness “b” which needs to be thick enough to absorb the energy from the beta
particle. The interaction produces photons in the phosphor. The photon emission is
isotropic. The photons which are born with a vector in the direction towards the
Pm-147 layer are lost unless the design puts a thin reflective layer of aluminum on
both side of the Pm-147 layer. The other complicating factor is that the phosphor
layer is not optically thin, so there will be self-absorption of the photons. The
design of the phosphor layer is complicated in that the thickness of the phosphor
should be greater than the range of a beta particle in the phosphor (Rb) but less than
the mean free path of a photon in phosphor layer (kph). The grid has transparency
efficiency (ηGtr) which is defined as the fraction of photons which are transmitted
through the grid. The photons then enter the photovoltaic cell. This is somewhat
complicated in that the photons enter at the photovoltaic cell at angles which are
much different than those experienced by solar cells (Figs. 3.44 and 3.45).
The photovoltaic cell used in a nuclear battery is basically the same type of cell
used to capture solar energy. The photovoltaic cell has a band-gap of Eg, which
needs to be matched as closely as possible to the emission spectrum of the
Fig. 3.44 This is a

photovoltaic cell which
consists of a n-type layer, a
p-type layer and a depletion
area. Solar radiation is made
up of paraxial rays which
impinge upon the solar cell
and bounce between the
reflective layers until the
photon is absorbed. Solar cell
architecture is optimized so
that there is a high probability
that the photon is absorbed by
the depletion layer
Fig. 3.45 This is a

photovoltaic cell which
consists of a n-type layer, a
p-type layer and a depletion
area. Fluorescence from the
indirect battery phosphor
produces rays which impinge
upon the solar cell and bounce
between the reflective layers
until the photon is absorbed.
The architecture for an
indirect battery must be
designed so that the
probability that the photon is
absorbed by the depletion
layer is maximized
phosphor. As described in the previous section on gas fluorescers, since the

phosphor has a broad emission spectrum, there will be inefficiencies in the con-
version process. The cell is made by interfacing p-type material with n-type
material. A junction is formed between n- and p-type semiconductor materials. At

the junction’s interface, the diffusion of charge carriers across the junction sets up a
space charge which in turn sets up an electrical potential and electrical field
(Fig. 3.17). The depletion region thus has an electrical field which separates
electron-hole pairs when they are created by the interaction of radiation with the
material. Outside of the depletion region, electron-hole pairs which are created by
the interaction of radiation with matter do not have electrical fields to cause their
separation. The most likely fate of the charge carriers beyond the depletion region is
that they will instead recombine or become trapped by defects. Some of the charge
carriers may drift into the depletion region and become part of the cell current
driven by the junction’s electric field.
There is also another significant difference between the solar photon source and a
phosphor photon source. First, assume that photons are rays being emitted from a
source. A solar cell captures photons (operating as paraxial rays) from the sun as
shown in Fig. 3.44. The reason that the rays are paraxial is that the sun is an
isotropic photon emitter approximately 93 million miles from the earth. The earth
intersects only the photons which are emitted in a very small solid angle. The
photons which the earth intersects are essentially parallel rays or “paraxial.”
The paraxial rays enter a solar cell at the same angle. The cell is coated with an
antireflective material which allows the rays to enter with minimal reflection. It
moves through the cell and will eventually be absorbed. However, the challenge is
to absorb the photon in the depletion layer. Some photons are not absorbed by the
depletion layer and they are reflected by the interface because the antireflective
coating has a lower index of refraction than the photovoltaic cell materials. As light
travels from one medium with an index of refection of n1 to another medium with
lower index of refraction n2, the light will be refracted at 90° when it hits the
interface at an angle greater than a critical angle (hc). This is called total internal
reflection. The critical angle can be determined from Snell’s law,
n2
sinðhc Þ ¼ ð3:76Þ
n1
where n1 > n2.

Thus any photon that is not absorbed will reflect off the bottom surface. This
gives the photon one more pass at being absorbed in the depletion region where the
p-n junction resides. If the photon is absorbed in the region outside of the depletion
region, it will form an electron-hole pair. The electron-hole pair most likely will
recombine thus emitting an isotropic photon of energy Eg. This photon will travel
through the medium and hopefully be absorbed in the depletion layer. If not, and if
it is able to reflect off of the top or bottom surface and is absorbed outside the
depletion layer, it will again form an electron-hole pair and its most likely fate is to
recombine. This process of perpetually creating electron-hole pairs and through
recombination and creating a photon with energy equal to the band-gap will
practically extend the photon lifetime until it is absorbed in the depletion layer, at
which point the most likely fate of the electron-hole pair is to be separated by the
electric field thus creating a current.
The dimensions of the solar cell can be adjusted such that photons bounce
between the top and bottom surfaces until the photons are absorbed with a high
probability in the depletion layer.
Isotropic fluorescer sources such as phosphors enter the photovoltaic cells at
various angles based upon the angle of emission and the reflective property of the
waveguide which directs the photons to the PV cell. Due to the various angles at
which a photon can enter the cell, there will be additional inefficiencies because of
the wide variety of paths photons will take before they are absorbed or lost. The
probability of absorption in the depletion layer will be lower than that for paraxial
rays.
3.2.6 Hybrid Solid-State Emitter
A hybrid approach to addressing the self-absorption problem is to form

micro-bubbles with excimer gases in the solid-state material [74]. Micro-bubbles
can be formed at very high pressures in a solid-state material using ion implantation
(up to 4 GPa). At 4 GPa, the density of a xenon gas bubble is on the order of
4 g/cm3. The transport length of radiation in a high pressure xenon micro-bubble is
about 5 lm, about the scale length of the heavy fragment. As shown in Fig. 3.46,
the radioisotope can be coated on the cell’s surface. A series of micro-bubbles lie
between the radioisotope layer and the p-n junction. The particle from the
radioisotope is emitted isotropically and the micro-bubble serves as both a shield to
protect the junction as well as a photon source which emits at the excimer
Radioisotope Layer
N
K
UV
photons
P
Fig. 3.46 Option A: Micro bubble as a radiation shield as well as a way of converting the kinetic
energy of radiation into narrow band UV photons that are absorbed by the p-n junction
wavelength. The photons then resonate in the PV cell and are absorbed. Even at this
high density the issues of pressure broadening should not lead to losses and the
micro bubble should not self-absorb. Thus the cell will have a transducer scale
length compatible with the radiation source and with the PV cell. The advantages of
this approach are that the wide band-gap p-n structure will use a thin film with the
radioisotope coated or imbedded into the structure. Wide band-gap materials can
operate at high temperatures without efficiency loss and have high thermal con-
ductivities. The films can be stacked, which will allow for scaling of the power
source at relatively high power densities (see discussion on the limitations of nu-
clear battery power density). Problems with this mechanism do exist. Even though
it is well known that micro-bubbles form by ion implantation, the possibility of the
bubble delaminating the material may be a problem. The theoretical maximum
efficiency for this configuration is 20–30 %.
3.2.7 Heat Based
Nuclear energy conversion systems based on heat production can take advantage of
the fact that all of the energy content in ionizing radiation can be converted to heat,
unlike systems which are based on ions produced by ionizing radiation where at
most 50 % of the energy goes into ion production. Direct energy conversion sys-
tems based on heat production include the RTG (which uses the Seebeck effect),
thermoelectric energy conversion and thermionic energy conversion.
3.2.7.1 Seebeck Effect and RTG
The thermoelectric effect was first observed by Thomas Johann Seebeck in 1821. In
this experiment, a magnetic needle was deflected when it was positioned near a
circuit made of two different conductors. Seebeck attributed the observation to
temperature, but did not associate it with a current flow (which would produce a
magnetic field) in the junction due to temperature. He went on to study many types
of materials and saw variation of the effect. One of his choices of materials would
have produced thermoelectric efficiencies near 3 %. This was on par with steam
engine efficiencies during that time [75]. In 1834, Jean Charles Athanase Peltier
observed that when current flowed through a junction made of two different
materials, it caused the metal to either cool down or heat up depending on the
direction of the current flow. This discovery was furthered by the work of Emil
Lenz of the St. Petersburg Academy in 1838 when he demonstrated that water on a
bismuth-antimony junction could form ice when current was flowed one direction
or could melt ice when current flowed in the other direction.
William Thompson (also known as Lord Kelvin) reasoned that the Seebeck and
Peltier effects were related. Based on a thermodynamic analysis, Lord Kelvin
concluded that a third effect must exist in which heating and cooling could take
Fig. 3.47 The basic circuit for a thermocouple is shown. There are two materials (A and B) that
make up the thermoelectric junction. Point 5 is where the two materials are connected and are
attached to the heat source (Th). Point 3 is where material A is attached to the sink (Tc) and point 4
is where material B is attached to the sink. The voltage across points 1 and 2 is the Seebeck voltage
place in a homogeneous conductor when current is flowed in the direction of a

temperature gradient. This is called the Thompson effect.
In its most basic form, the thermoelectric effect occurs when two different metals
are connected between a high temperature side (Th) and a low temperature sink (Tc)
as shown in Fig. 3.47. The voltage between points 1 and 2 (V1,2) is the Seebeck
voltage. The Seebeck voltage is related to the temperature difference between the
heat source and the sink (Th − Tc) by the Seebeck coefficient (a) as seen in
Eq. (3.78).
DV ¼ V2 V1 ¼ V1;2 ¼ ðaA ab ÞðTh Tc Þ ¼ ðaA aB ÞDT ð3:77Þ
So,
dV1;2
¼ aA a B ð3:78Þ
dT
A basic thermoelectric generator has a heat source (at temperature Th) and a
thermal sink (at temperature Tc). There will be two semiconductor legs, one being a
p-type material and one being an n-type material as shown in Fig. 3.48. A load
(with resistance RL) is connected between the two legs as shown in the figure.
Fig. 3.48 Shown is a

thermoelectric generator
which consists of a heat
source, two legs, a thermal
sink and a load. The length of
the legs is x
The efficiency of a thermoelectric generator is defined as,
P0
g¼ ð3:79Þ
qh
where P0 is the electrical power out of the generator and qh is the thermal power
input.
The thermal power into the hot legs of the thermoelectric generator from the heat
source is dependent upon the thermal properties of the legs (namely the thermal
conduction which goes as KDT where K is the thermal conductivity of the leg
material). There will also be losses due to the current flow in the legs (this term goes
as ½I2R where R is the resistance of the leg materials and I is the current flow
through the legs). Additionally, there is a heat generated due to the Peltier effect
(aThI). Since electrons flow from the n-type leg to the p-type leg, the current flow
will have a minus sign. The term ½ comes from the fact that the Peltier effect causes
absorption of heat at the source while half of the Joule heat is returned to the source.
Thus,
1
qh ¼ KDT þ aTh I I 2 R ð3:80Þ
2
The output power of the thermoelectric generator is,
P0 ¼ I 2 RL ð3:81Þ
where RL is the load resistor.

The open circuit voltage (Voc) for the thermoelectric generator is the Seebeck
voltage,
Voc ¼ aDT ð3:82Þ
So, the current that flows through the generator is,
aDT
I¼ ð3:83Þ
R þ RL
Plugging Eqs. (3.81) and (3.82) into Eq. (3.80), the system efficiency becomes,
RL I 2
g¼ ð3:84Þ
KDT þ aTh I 12 I 2 R
The efficiency equation is very basic and more complex versions of this equation
exist in the literature.
In evaluating thermoelectric generators, an important grouping of properties of
the materials used is the figure of merit (Z),
Fig. 3.49 Figure of merit of p-type materials (a), figure of merit for n-type materials (b), and
doping concentration modifying the maximum ZT (c) [76]
a2
Z¼ ð3:85Þ
RK
The figure of merit (Z) is an oft reported value which is used to judge the
suitability of a material for a temperature range. Another value which is oft reported
is the product ZT, which is a dimensionless number and provides basically the same
type of information as Z as shown in Fig. 3.49.
RTGs are the “gold standard” in nuclear battery technology. They have had a
long history of successful missions as shown in Table 3.17. Some of these RTGs
are shown in Figs. 3.50, 3.51, 3.52, 3.53 and 3.54.
3.2.7.2 Thermoionics
The thermionic generator is basically made of two electrode surfaces (one an

emitter and the other a collector) as shown in Fig. 3.55. The emitter is a metal with
a known work function. In a metal, the electrons are free to move, but are bound to
the metal by forces of attraction between the electrons and ions in the lattice. The
Table 3.17 RTGs used in various NASA missions [76]

Mission RTG TE Destination Year Mission
length
Transit 4A SNAP-3B7(1) PbTe Earth orbit 1961 15
Transit 4B SNAP-3B8(1) PbTe Earth orbit 1962 9
Apollo 12 SNAP-27 RTG (1) PbTe Lunar surface 1969 8
Figure 3.50
Pioneer-10 SNAP-19 RTG (4) PbTe Outer planets 1972 34
Figure 3.51
Triad-01-1X SNAP-9A (1) PbTe Earth orbit 1972 15
Pioneer-11 SNAP-19 RTG (4) PbTe Outer Planets 1973 35
Figure 3.51
Viking 1 SNAP-19 RTG (2) PbTe Mars 1975 4
Figure 3.51
Viking 2 SNAP-19 RTG (2) PbTe Mars 1975 6
Figure 3.51
LES 8 MHW-RTG Si-Ge Earth orbit 1976 15
Figure 3.52
LES 9 MHW-RTG SiGe Earth orbit 1976 15
Figure 3.52
Voyager 1 MHW-RTG Si-Ge Outer planets 1977 31
Figure 3.52
Voyager 2 MHW-RTG Si-Ge Outer planets 1977 31
Figure 3.52
Galileo GPHS-RTG Si-Ge Outer planets 1989 14
Figure 3.53
Ulysses GPHS-RTG (1) Si-Ge Outer 1990 18
Figure 3.53 planets/sun
Cassini GPHS-RTG (3); RHU Si-Ge Outer planets 1997 11
Figure 3.53 (117)
New GPHS-RTG (1) Si-Ge Outer planets 2005 3
Horizons
Figure 3.53
MSL MMRTG PbTe Mars surface 2011 3
Figure 3.54
work function (u) is defined as the energy at which electrons overcome the force of
attraction.
When heat is applied to the emitter, the energy the free electrons will have a
Boltzmann’s energy distribution (Eq. 3.86). In Fig. 3.56, the function f(E) is plotted
as a function of energy. For this figure, a surface heated to 1,800 K has a work
function of 4.15 eV. The cross hatched area represents the electrons whose energies
exceed the work function and thus are emitted. One potential complication is a
space charge buildup, which adds to the potential barrier the electrons must over-
come. In this case, the electrons which are collected are those that have energy
greater than the work function and potential barrier.
Fig. 3.50 SNAP 27 reactor

used on the Apollo 12, 14, 15,
16 and 17 missions. The
power output was initially
70 W. Its design lifetime was
2 years [77]
rffiffiffiffi
E 1 3=2 k ET
f ðEÞ ¼ 2 e B ð3:86Þ
p kB T
The ideal current emitted by a heated surface is described by the

Richardson-Dushman equation,
k uT
J ¼ A1 T 2 e B ð3:87Þ
where A1 is a constant, T is temperature in Kelvin, u is the work function and kB is

the Boltzmann constant.
The Richard-Dushman equation represents the saturation current from an
emitting surface per unit area at a temperature T. The potential energy diagram of
thermionic emitter and collector configuration is shown in Fig. 3.57.
In Fig. 3.57, electrons emitted by the emitter must have velocities such that they
overcome the potential, d + uc + V − ue. This potential can be plugged into the
Richardson-Dushman Eq. (3.87) to find the emitter current,

eue eð d þ u c þ V þ u e Þ
Je ¼ A1 Te2 exp exp ð3:88Þ
kB Te kB T
where Ve ¼ uk þ ue ¼ due þ V
Using the definition for Ve, Eq. (3.88) becomes
Fig. 3.51 SNAP 19 RTG used in Pioneer and Viking Missions. The initial power was 40.3 W
electric for the Pioneer missions and 42.6 W electric for the Viking missions [77]
Fig. 3.52 The Multi-Hundred Watt (MHW) RTG was capable of producing 158 W electric. It
was used on the Voyager missions [77]
eVe
Je ¼ A1 Te2 exp ð3:89Þ
kB T e
The current from the collector to the emitter (or back current) comes from
electrons in the collector that overcome the potential barrier d. The back current is
calculated from the Richardson-Dushman Eq. (3.87),

euc ed
Jc ¼ A1 Tc2 exp exp ð3:90Þ
kB T c kB Tc
Fig. 3.53 The GPHS-RTG was used for the Ulysses and Cassini missions. Its power output was
292 W electric [77]
where Vc ¼ uc d

eVc
Jc ¼ A1 Tc2 exp ð3:91Þ
kB T c
The net current is,

eVe eVc
J ¼ Je Jc ¼ A1 Te2 exp A1 Tc2 exp ð3:92Þ
kB T e kB Tc
where J is a function of Te, Tc, d and uc.

The output voltage (V) is,
V ¼ Ve Vc ¼ ðue uc Þ þ ðuk dÞ ð3:93Þ
When a load resistor (RL) is added to the circuit, the voltage on the load (VL) is,
VL ¼ Ve Vc IRL
Fig. 3.54 The Multi-Mission radioisotope thermoelectric generator. The MMRTG will be used in
NASA mars science laboratory rover mission [77]
where I is the current flow.

The open circuit voltage occurs when the net current flow (J) is zero. Thus the
condition that Je = Jc gives,

eVe eVc
A1 Te2 exp ¼ A1 Tc exp
2
ð3:94Þ
kB Te kB Tc
Rearranging Eq. (3.94),

Te2 eVc eVe

¼ exp exp ð3:95Þ
Tce kB Tc kB Te
Fig. 3.55 The basic structure

of a thermionic generator is
shown. The generator consists
of an emitter electrode and a
collector electrode. The
emitter electrode is heated and
the collector electrode is
cooled to a sink
Fig. 3.56 The Boltzmann

distribution of electron energy
for a surface heated to
1,800 K. Shown is a line for a
material with a work function
of 4.15 eV. The electrons
with energies above 4.15 eV
are emitted from the surface
Equation (3.95) then becomes,

2
Te e Ve Vc
¼ exp ð3:96Þ
Tc kB Te Tc
Fig. 3.57 The potential

energy level diagram between
the emitter and collector for a
thermionic device. Ve is the
potential energy of the
emitter, Vc is the potential
energy of the collector, uk is
the retarding potential barrier
due to space charge buildup, d
is the retarding potential
barrier due to space charge
buildup on the collector side,
uc is the work function of the
collector, ue is the work
function for the emitter, V is
the operating voltage of the
cell, we is the electrical
potential for the emitter and
wc is the electrical potential
for the collector
Take the natural log of Eq. (3.96),

Te e Ve Vc
2 ln ¼ ð3:97Þ
Tc kB Te Tc
Realizing that Vc ¼ uc þ d and Vc ¼ d þ ue þ V,

Te e Te
2 ln ¼ ðd þ uc þ Voc Þ ð u c þ dÞ ð3:98Þ
Tc kB Te Tc
Solve for Voc,

2kB Te Te Te
Voc ¼ ln þ ðuc þ dÞ þ1 ð3:99Þ
e Tc Tc
The efficiency of the cell is the power out (V[Je − Jc]) divided by the power in
(qs, which is the heat supplied by the emitter).
V ð Je Jc Þ
g¼ ð3:100Þ
qs
Equation (3.100) does not take parasitic heat losses, space charge losses, or back
current losses into account.
Table 3.18 Various space based nuclear reactors developed over the years that operate with
various energy conversion mechanisms
SNAP SP-100 Romashka Bouk Topaz 1 USSR Topaz 2 USSR SAFE
10 US US USSR USSR 400
US
Date 1965 1992 1967 1977 1987 1992 2016
service
KWth 45.5 2000 40 *100 150 135 400
KWe 0.65 100 0.8 *5 10 6 100
Fuel U-ZrHx UN UC2 U-Mo UO2 UO2 UN
Converter TE TEc TE TE TI TI Bryton
Mass Kg 435 5422 455 *390 320 1061 512
Neutron Thermal Fast Fast Fast thermal Epithermal Fast
energy
Control Be Be Be Be Be Be Be
Coolant NaK Li None NaK NaK NaK Na
Core 585 1377 1900 NA 1600 1900 1020
Temp C
Emitter Monocrystal Monocrystal
Mo with W Mo with W
coating coating
Collector Polycrystalline Polycrystalline
Mo Mo
Efficiency 1.4 % 5% 2% *5 % 6.67 % 9.3 % 25 %
The Topaz reactors used thermionics
TE-Thermoelectric, TI-Thermionic, Brayton-Brayton Cycle
The performance of a thermionic device is dependent upon the operating tem-

perature and on the materials used for the emitter and collector. Table 3.18 is a
summary of nuclear reactors using the method of direct energy conversion. The
efficiencies achieved are consistent with the DCNB (which is the underlying
mechanism of a thermionic energy convertor) discussed in this chapter.
3.2.7.3 Thermophotovoltaics
Thermophotovoltaics are similar in principle to PIDEC [75, 78–85]. The main

differences are,
1. The photon spectrum is much broader. It is a black body spectrum coming off of
a heat source (obeying Planks radiation formula in Eq. 3.101), and
2. The photovoltaic cells have to have a low band-gap in order to maximize the
conversion of a black body spectrum to electricity since the emission for rea-
sonably hot emitters is in the infrared (Fig. 3.58) [75].
Fig. 3.58 Shown is the black body emission S (J/m3-nm) as a function of wavelength and
temperature
8phc 1
Sk ¼ ð3:101Þ
k exp hc 1
5
kkB T
where Sk is in J m−3 nm−1, h is Planck’s constant (6.62607 10−34 m2 kg s−1), c

is the speed of light (2.99799979 108 ms−1), kB is Boltzmann’s constant
(1.38 10−23 m2 kg s−2 K−1), k is wavelength in meters, and T is temperature of
the black body in Kelvin.
Wein’s displacement law predicts the peak wavelength of a black body at any
temperature and shown in Table 3.19.
b
kpeak ¼ ð3:102Þ
T
where b = 2.9 106 nm K.

Thermophotovoltaic materials should match up with the grey body spectrum
generated by the emitter. A diagram of a thermophotovoltaic cell is shown in
Fig. 3.59.
Table 3.19 Peak wavelength T (K) k peak (nm) hm (eV)

at various black body
temperatures 1000 2900.00 0.43
1200 2416.67 0.51
1400 2071.43 0.60
1600 1812.50 0.69
1800 1611.11 0.77
2000 1450.00 0.86
2200 1318.18 0.94
2400 1208.33 1.03
2600 1115.38 1.12
2800 1035.71 1.20
3000 966.67 1.29
3200 906.25 1.37
3400 852.94 1.46
3600 805.56 1.54
3800 763.16 1.63
4000 725.00 1.72
4200 690.48 1.80
4400 659.09 1.89
Fig. 3.59 Diagram of a thermophotovoltaic system. The emitter gives off grey body radiation
with an efficiency (ηrad). The radiation then passes through an optical filter and the fraction of
photons which can be absorbed by the photovoltaic cell has a spectral efficiency (ηspec). Emitted
radiation expands and a portion will reach the photovoltaic. This is described by the geometrical
factor (ηgeom). Once absorbed in the photovoltaic cell, the photons are converted into an electrical
current described by the photovoltaic efficiency (ηpv)
The thermophotovoltaic generator system efficiency is (ηsystem),
gsystem ¼ grad gspec ggeom gpv ð3:103Þ
Typical matchups between low band-gap materials (binary, ternary, or quater-

nary semiconductors) and emitter temperature are shown in Table 3.20.
Table 3.20 Thermophotovoltaic material properties and probable emitter operating temperature
[79]
Material Band-Gap Typical operating band-gap Probable operating grey body
Eg (eV) energy (eV) temperature (K)
InP 1.344 1.344 3100
InGaAsSb 0.5–0.6 0.55 1300
InGaAs 0.4–1.4 0.55 1300
InPAsSb 0.3–0.55 eV 0.5 1200
Ge 0.66 0.66 1500
GaSb 0.72 0.72 1899
Thermophotovoltaic systems have promise. Theoretically, efficiencies of 30+ %

have been predicted [83]. Practically, the system efficiencies range from about
0.04–24 % [80].
3.3 Summary
Ionizing radiation interacts with matter by creating ionization and excitation of the
atoms and/or molecules that make up the material. In an energy conversion system
which depends on the production of ions, there will be a reduction of efficiency due
to the capability of producing ions (35–50 %) based on the transducer material
properties. For systems which depend on heat, the ionization and excitation will
eventually be converted to heat.
The energy conversion devices which are ion based include alphavoltaics, be-
tavoltaics, Schottky barrier, direct charge collection and indirect systems such as
PIDEC. A few heat based direct energy conversion systems were also discussed
including the RTG, thermionics and thermophotovoltaics.
Problems
1. An electron is emitted from the source plate of a DCNB at angle h = 30°.
Assuming this angle exceeds hm, what percent of the particle’s initial kinetic
energy will reach the collector? If the minimum angle of emission was deter-
mined to be 25°, what was the ideal efficiency of the parallel plate DCNB?
2. Suppose the distance between the source and the collector is 1.5 cm. A beta
particle from Sr-90 is emitted off the surface of the source at the minimum
angle of emission of 60°. What is the electric field strength between the source
and collector? Calculate using the average energy determined by the 1/3 bmax
rule as well as the full spectrum.
3. An RTG utilizing the Seebeck effect is placed in an environment resulting in a
temperature gradient of 700 K between the radioisotope heat source and the
heat sink. If material A has a Seebeck coefficient of 7.5 µV/K and material B
has a Seebeck coefficient of −15 µV/K, what is the voltage produced?
4. Discuss the maximum efficiency limitations of ion pair based transducers.

5. Discuss the efficiency limitations of alphavoltaic cells.
6. Discuss the efficiency limitations of betavoltaic cells.
7. Discuss the efficiency limitations of Schottky barrier based nuclear batteries.
8. Discuss the efficiency limitations of direct charge collection nuclear batteries.
9. Discuss the efficiency limitations of a direct charge collection nuclear battery
that works on charge build up and discharge cycles.
10. Design a 1 W electric indirect nuclear battery based on Kr-85. What is the dose
rate to personnel for an unshielded nuclear battery with your design?
11. Discuss the efficiency limitations for an indirect nuclear battery using
solid-state isotopes?
12. Discuss the best wide band-gap semiconductors that can be used as photo-
voltaic cells for excimer fluorescence conversion.
13. What is the spectral matching efficiency for SiC, GaN, diamond and AlN with a
xenon excimer fluorescence source?
14. Are phosphor emitters more efficient than excimer emitters? Explain your
answer.
15. Explain the Seebeck effect and how it works.
16. Explain how thermionic systems work.
17. What are the strengths and the weaknesses of thermophotovoltaics?
References
1. Smith A, Fields P, Roberts J (1957) Spontaneous fission neutron spectrum of Cf252. Phys Rev
108:411–413
2. Nicodemus DB, Staub HH (1953) Fission neutron spectrum of U^{235}. Phys Rev 89:1288–
1290
3. Zeynalov S, Hambsch F-J, Obertstedt S (2011) Neutron emission in fission of ^252Cf(SF).
J Korean Phys Soc 59:1396
4. DOE-HDBK-1019/1–93 (1993) Nuclear physics and reactor theory. Department of Energy,
Washington, DC
5. Turner JE, Kelsey KA (1995) Atoms, radiation, and radiation protection. Wiley, New York
(2010). Nucl Instrum Methods Phys Res, Sect B 268:1818–1823
review of nuclear batteries. Prog Nucl Energ 75:117–148
9. Prelas M, Charlson E, Charlson E, Meese J, Popovici G, Stacy T (1993) Diamond photovoltaic
energy conversion. In: Yoshikawa M, Murakawa M, Tzeng Y, Yarbrough WA (eds) Second
international conference on the application of diamond films and related materials. MY Tokyo,
pp 5–12
10. Prelas M, Popovici G, Khasawinah S, Sung J (1995) Wide band-gap photovoltaics. In: Wide
band gap electronic materials. Springer, pp 463–474
11. Prelas MA, Hora HP (1994) Radioactivity-free efficient nuclear battery. Germany Patent
References 173
12. Radprocalculator (2015) Gamma emitter point source dose-rate with shielding. http://www.
radprocalculator.com/Gamma.aspx
13. Guyot J, Miley G, Verdeyen J (1972) Application of a two-region heavy charged particle
model to Noble-gas plasmas induced by nuclear radiations. Nucl Sci Eng 48:373–386
14. Giuliani JL, Petrov GM, Dasgupta A (2002) Electron energy deposition in an electron-beam
pumped KrF amplifier: Impact of the gas composition. J Appl Phys 92:1200–1206
15. Bernard S, Slaback Jr Lester A, Kent BB (1998) Handbook of health physics and radiological
health. Williams & Wilkins, Baltimore
16. Oh K (2011) Modeling and maximum theoretical efficiencies of linearly graded alphavoltaic
and betavoltaltaic cells. M.Sc., Nuclear Science and Engineering Institute, University of
Missouri, University of Missouri, Columbia
2,745,973, 1956
18. Anno JN (1962) A direct-energy conversion device using alpha particles. Nucl News 6
19. Oh K, Prelas MA, Lukosi ED, Rothenberger JB, Schott RJ, Weaver CL et al (2012) The
theoretical maximum efficiency for a linearly graded alphavoltaic nuclear battery. Nucl
Technol 179:7
20. Schott RJ (2012) Photon intermediate direct energy conversion using a Sr-90 beta source.
Nuclear Science and Engineering Institute, University of Missouri, PhD
21. Eckerman KF, Westfall RJ, Ryman JC, Cristy M (1994) Availability of nuclear decay data in
electronic form, including beta spectra not previously published. Health Phys 67:338–345
22. Wrbanek JD, Wrbanek SY, Fralick GC, Chen L-Y (2007) Micro-fabricated solid-state
radiation detectors for active personal dosimetry. NASA/TM 214674
23. Ravankar ST, Adams TE (2014) Advances in betavoltaic power sources. J. Energy Power
Sources 1:321–329
24. Popovici G, Melnikov A, Varichenko VV, Sung T, Prelas MA, Wilson RG et al (1997)
Diamond ultraviolet photovoltaic cell obtained by lithium and boron doping. J Appl Phys
81:2429
25. Doolittle WA, Rohatgi A, Ahrenkiel R, Levi D, Augustine G, Hopkins RH (1997)
Understanding the role of defects in limiting the minority carrier lifetime in sic. MRS Online
Proc Lib 483:null–null
26. Seely JF, Kjornrattanawanich B, Holland GE, Korde R (2005) Response of a SiC photodiode
to extreme ultraviolet through visible radiation. Opt Lett 30:3120–3122
27. Neamen DA (2003) Semiconductor physics and devices. McGraw Hill
28. Savtchouk A, Oborina E, Hoff A, Lagowski J (2004) Non-contact doping profiling in epitaxial
SiC. Mater Sci Forum 755–758
29. Huang M, Goldsman N, Chang C-H, Mayergoyz I, McGarrity JM, Woolard D (1998)
Determining 4H silicon carbide electronic properties through combined use of device
simulation and metal–semiconductor field-effect-transistor terminal characteristics. J Appl
Phys 84:2065–2070
30. Sze SM, Ng KK (2006) Physics of semiconductor devices. Wiley
31. Latreche A, Ouennoughi Z (2013) Modified Airy function method modelling of tunnelling
current for Schottky barrier diodes on silicon carbide. Semicond Sci Technol 28:105003
32. Östlund L (2011) Fabrication and characterization of micro and nano scale SiC UV
photodetectors. Student Thesis, Masters of Science, Royal_Institute_of_Technology.
Stockholm, p 74
33. Eiting CJ, Krishnamoorthy V, Rodgers S, George T, Robertson JD, Brockman J (2006)
Demonstration of a radiation resistant, high efficiency SiC betavoltaic. Appl Phys Lett
88:064101-064101-3
34. Tsang FY-H, Juergens TD, Harker YD, Kwok KS, Newman N, Ploger SA (2012) Nuclear
voltaic cell. Google Patents
35. Wacharasindhu T, Jae Wan K, Meier DE, Robertson JD (2009) Liquid-semiconductor-based
micro power source using radioisotope energy conversion. In: Solid-state sensors, actuators
and microsystems conference, 2009. TRANSDUCERS 2009. International, pp 656–659
36. Patel JU, Fleurial J-P, Snyder GJ (2006) Alpha-voltaic sources using liquid ga as conversion
medium. NASA: NASA Tech Briefs
37. Miley GH (1970) Direct conversion of nuclear radiation energy. American Nuclear Society
38. Galina Yakubova AK (2012) Nuclear Batteries with tritium and promethium-147 radioactive
sources: design, efficiency, application of tritium and pm-147 direct charge batteries, tritium
battery with solid dielectric. LAP Lambert Academic Publishing
39. Mulcahy MJ, Bolin PC (1971) In: Agency ARP (ed) High voltage breakdown study:
handbook of vacuum insulation. National Technical Information Services, Springfield, VA,
p 78
Beams 6:25–62
41. Prelas MA (1985) Excimer Research using nuclear-pumping facilities. National Science
Foundation: NSF, p 1–131
42. Lecours MJ, Prelas MA, Gunn S, Edwards C, Schlapper G (1982) Design, construction, and
testing of a nuclear-pumping facility at the University of Missouri Research Reactor. Rev Sci
Instrum 53:952–959
43. Chung AK, Prelas MA (1987) Sensitivity analysis of Xe2* excimer fluorescence generated
from charged particle excitation. Laser Part Beams 5:125–132
44. Klein MB (1974) In: Laboratories HR (ed) Waveguide gas lasers. NTIS
45. Prelas MA (1981) A potential UV fusion light bulb for energy conversion. In: Presented at the
23rd annual meeting division of plasma physics American physical society, New York, NY,
1981
47. Prelas MA (1981) In: Missouri U (ed) Notorized notes dust core reactor and laser. p 7
48. Boody EP, Prelas MA, Miley GH (1977) Nuclear generated excimer radiation for pumping
lasers. University of Illinois, Nuclear Engineering Department
49. Prelas MA, Boody FP, Zediker M (1984) A direct energy conversion technique based on an
aerosol core reactor concept. In: IEEE International conference on plasma science, p 8
50. Melnikov SP, Sizov AN, Sinyanskii AA, Miley GH (2015) Lasers with nuclear pumping.
Springer, New York, NY
51. Chung A, Prelas M (1984) Charged particle spectra from U-235 and B-10 micropellets and
slab coatings. Laser Part Beams 2:201–211
52. Prelas M, Boody F, Zediker M (1984) A direct energy conversion technique based on an
aerosol core reactor concept. IEEE Publication, p 8
53. Prelas MA, Boody FP, Zediker MS (1985) An aerosol core nuclear reactor for space-based
high energy/power nuclear-pumped lasers. In: El-Genk MS, Hoover M (eds) Space nuclear
power systems. Orbit Book Company
54. Mencin DJ, Prelas MA (1992) Gaseous like Uranium reactors at low temperatures using C60
cages. In: Proceedings of nuclear technologies for space exploration. American Nuclear
Society, Aug 1992
55. Guoxiang G, Prelas MA, Kunze JF (1986) Studies of an aerosol core reactor/laser’s critical
properties. In: Hora H, Miley GH (eds) Laser interaction and related plasma phenomena.
Springer, pp 603–611
56. Walters RA, Cox JD, Schneider RT (1980) Trans Am Nucl Soc 34:810
57. Prelas M, Charlson E (1989) Synergism in inertial confinement fusion: a total direct energy
conversion package. Lasers Part Beams 7:449–466
58. Prelas M, Charlson E, Boody F, Miley G (1990) Advanced nuclear energy conversion using a
two step photon intermediate technique. Prog Nucl Energy 23:223–240
59. Lee MYJJ, Simones MMP, Kennedy JC, Us H, Makarewicz MPF, Neher DJA et al (2014)
Thorium fuel cycle for a molten salt reactor: state of Missouri feasibility study. ASEE annual
conference. Indianappolis, IN, p 28
60. Schott RJ, Weaver CL, Prelas MA, Oh K, Rothenberger JB, Tompson RV et al (2013) Photon
intermediate direct energy conversion using a 90Sr beta source. Nucl Technol 181:5
References 175
61. Weaver CL (2012) PIDECa: photon intermediate direct energy conversion using the alpha
emitter polonium-210. PhD Nuclear Science and Engineering Institute, University of Missouri.
http://hdl.handle.net/10355/15908
62. Ahlswede J, Hebel S, Kalinowski MB, Roß JO, Update of the global krypton-85 emission
inventory. Carl-Friedrich-von-Weizsäcker-Zentrum für Naturwissenschaft und Friedensforschung
der Universität Hamburg
63. Steinfelds E, Tulenko J (2011) Isotopes and radiation: general-evaluation and verification of
durability and efficiency of components of photon assisted radioisotopic batteries. Trans Am
Nucl Soc 104:201
64. Steinfelds EV, Tulenko JS (2011) Development and testing of a nanotech nuclear battery for
powering MEMS devices. Nucl Technol 174:119–123
65. Steinfelds E, Tulenko JS (2009) Designs and performance assessments of photon assisted
radioisotopic energy sources. Trans Am Nucl Soc 100:672–674
66. Steinfelds E, Prelas M (2007) More sources and review of design for radioisotope energy
conversion systems. Trans Am Nucl Soc 96:811–812
67. Bower KE, Barbanel YA, Shreter YG, Bohnert GW (2002) Polymers, phosphors, and voltaics
for radioisotope microbatteries. CRC Press
68. Prelas MA, Sved J, Dann A, Jennings HJ, Mountford A (1999) Solid state electric generator
using radionuclide-induced exciton production, WO 1999036967 A1
69. Shan W, III JWA, Yu KM, Walukiewicz W, Haller E, Martin MC et al (1999) Dependence of
the fundamental band gap of AlxGa1-xN on alloy composition and pressure. J Appl Phys
85:8505–8507
70. Prelas MA (2016) Nuclear-pumped lasers: Springer International Publishing
71. Shonoya S, Yen WM (1999) Phosphor handbook. CRC Press, Boca Raton, FL
72. Xu Z-H, Tang X-B, Hong L, Liu Y-P, Chen D (2015) Structural effects of ZnS:Cu phosphor
layers on beta radioluminescence nuclear battery. J Radioanal Nucl Chem 303:2313–2320
73. Hong L, Tang X-B, Xu Z-H, Liu Y-P, Chen D (2014) Parameter optimization and experiment
verification for a beta radioluminescence nuclear battery. J Radioanal Nucl Chem 302:
701–707
74. Prelas MA (2013) Micro-scale power source, United States Patent 8552616, USA Patent
75. Angrist SW (1982) Direct energy conversion, 4th edn. Allyn and Bacon Inc, Boston
76. Nochetto H, Maddux JR, Taylor P (2013) High temperature thermoelectric materials for waste
heat regeneration. Army Research Laboratory, Adelphi, MD, pp 20783–1197
77. NASA (2016) Radioisotope power systems. https://solarsystem.nasa.gov/rps/rtg.cfm
78. Ferrari C, Melino F (2014) Thermo—Photo—Voltaic generator development. Energy
Procedia 45:150–159
79. Ferrari C, Melino F, Pinelli M, Spina PR (2014) Thermophotovoltaic energy conversion:
Analytical aspects, prototypes and experiences. Appl Energy 113:1717–1730
80. Ferrari C, Melino F, Pinelli M, Spina PR, Venturini M (2014) Overview and status of
thermophotovoltaic systems. Energy Procedia 45:160–169
81. Nam Y, Yeng XY, Lenert P, Bermel P, Celanovic I, Soljačić M et al (2014) Solar
thermophotovoltaic energy conversion systems with two-dimensional tantalum photonic
crystal absorbers and emitters. Sol Energy Mater Sol Cells 122:287–296
82. Bitnar B, Durisch W, Holzner R (2013) Thermophotovoltaics on the move to applications.
Appl Energy 105:430–438
83. Teofilo VL, Choong P, Chang J, Tseng YL, Ermer S (2008) Thermophotovoltaic energy
conversion for space. J Phys Chem C 112:7841–7845
84. Nelson RE (2003) A brief history of thermophotovoltaic development. Semicond Sci Technol
18:S141–S143
85. Robert EN (2003) A brief history of thermophotovoltaic development. Semicond Sci Technol
18:S141
Chapter 4
Power Density Dilution Due
to the Interface of the Isotope
with the Transducer
Abstract In chapter 4, definitions for various types of dilution factors for a nuclear
battery are discussed. The average atomic density of the radioisotope in a nuclear
battery cell is described and a relationship between the average atomic density and
dilution factor are derived. The dilution factor will impact the minimum scale and
the power density of the battery. It is an important parameter that is used in the
assessment of nuclear battery designs.

Keywords Radioisotope dilution factor Phase Geometry Scale length matching
The discussion of the interface between the isotope and the transducer begins with
the source. In Chap. 1, the maximum power density was calculated for a list of
viable nuclear battery isotopes in the compounds which maximizes the atomic
density of the radioisotope atoms per cubic centimeter. Depending on the interface,
the effective atomic density of the isotope will be diluted compared to the maximum
atomic density discussed in Chap. 1. There is no such thing as a perfect interface so
any viable interface will dilute the average isotope density. Thus, the power density,
which is dependent on the average isotope density, will decrease. The following
sections begin by discussing the interface type. An analysis is developed to show
how the effective power density is affected by the specifics of the interface design.
There is a decrease in the power density which is specific to the design of the
interface. The dilution factor can be determined in several ways:
• The atomic dilution factor (DFatomic) is calculated if the cell volume is known.
First, calculate the number of radioisotope atoms contained in the cell volume
(Ncell). Second, assume that the volume is completely filled with a compound
that has the maximum possible density of radioisotope atoms (Nmax). The
atomic dilution factor is, DFatomic = Ncell/Nmax.
• The power density of the radioisotope source in the nuclear battery can be
related to the power density available from the compound with the highest
known atomic density of the radioisotope (1/BVWmin), Psource = DFatomic/
BVWmin.

178 4 Power Density Dilution Due to the Interface …
• The power dilution factor (DFpower) is defined as the power density of the
nuclear battery divided by the power density the compound with the maximum
atomic density of radioisotope:
• DFpower = (Pout/Vcell)/(1/BVWmin) = Pout*BVWmin/Vcell (where Pout is the
power out of the cell and Vcell is the volume of the cell).
• The volume dilution factor (DFvolume) is the ratio of the volume of the source
material to the total volume of the cell (source plus transducer).
• The cell dilution factor (DFcell) is the product of atomic dilution factor and the
volume dilution factor: DFcell = DFatomic*DFvolume.
The utility of the dilution factors used is as follows. The efficiency of the nuclear
battery can be calculated from the atomic dilution factor, the cell volume and the
power dilution factor,
ηsystem = Pout/Psource = (DFpower*Vcell/BVWmin)/(DFatomic*Vcell/ BVWmin) =
(DFpower)/(DFatomic).
Depending on the interface, the power density dilution factor can be consider-
able and will lead to significant increases the in the device volume.
4.1 Introduction
The interface between the charged particle producing nuclear reactions and an
energy conversion medium (or transducer) are the key to any direct nuclear energy
conversion system. A detailed description of the potential interfaces between
charged particle producing nuclear reactions and transducers has been discussed in
the book “Nuclear-Pumped Lasers” [1]. There are three primary methods of
interfacing radioisotopes to nuclear battery transducers: (1) surface interface,
(2) volume interface and (3) multiphase interface (Fig. 4.1). As described in
Chap. 1 of this text, charged particles emitted from a surface source must first pass
through the volume of the source material before reaching the transducer (Fig. 4.2).
Here a particle is born at a random point inside of the radioisotope containing layer
and it is emitted isotopically (meaning it has a 50 % probability of being emitted
towards the surface and a 50 % probability of being emitted away from the surface).
The particles that are directed towards the surface of the radioisotope bearing layer
will travel a path length dependent on the angle of emission. The path length will
determine the fraction of energy which is deposited in the radioisotope bearing
layer. Assume that the particle exits the surface at the point (x,y,z). As shown in
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Fig. 4.2, this particle travels the path length r ¼ x2 þ y2 þ z2 . Finding the fraction
of energy lost to the radioisotope containing layer for a statistically significant
number of particle emissions is the first important design consideration for a battery
that uses a surface source. Once the particle escapes the source material, the second
important design consideration is how much of its energy is transferred to the
transducer. Finding the fraction of energy absorbed by the transducer for a statis-
tically significant number of particle emissions is the next step. The third important
4.1 Introduction 179
Fig. 4.1 The figure illustrates the various types of interfaces between the radioisotope source and
the transducer. The basic surface source is shown in a where the radioisotope is embedded in a thin
film which allows the charged particles to escape the film and interact with the medium. A basic
illustration of a volume source is shown in b where the radioisotope is embedded directly in the
medium. Illustrated in c is one possible multiphase configuration where small aerosol particles are
embedded in a gas thus allowing the charged particles to efficiently escape the aerosol to interact
with a gaseous background. This is a fundamental indirect nuclear battery in that the background
gas produces photons which are then transported to the energy conversion medium. a Surface
source. b Volume source. c Aerosol source
Fig. 4.2 A particle is emitted isotopically by a radioisotope atom that is embedded in a

radioisotope containing layer. The particle is emitted at an angle / with respect to the z-axis and
the angle h with respect to the x-axis. The particle travels along the path shown in the figure and
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
exits the layer after traveling a distance r ¼ x2 þ y2 þ z2 . The path length is directly related to
the energy that is lost in the material that the particle travels through
design consideration is how the configuration of the surface source and transducer
dilutes the energy density of the nuclear battery.
Chapter 4 focuses on the physics of source design, the efficiency of the design
and the impact of the source design on the battery’s power density (or how the
design dilutes the power density).
4.2 Phase of the Radioisotope
Radioisotopes can be in the solid, liquid or gaseous phase. There is one other
possible phase, a plasma. A plasma will not be considered here.
4.2.1 Radioisotope in Solid Phase
Typically the form of the radioisotope is in the solid phase. One example is
polonium-210 added to silver. The silver and Po-210 is typically rolled into a thin
film which is interfaced to the transducer as a thin film (Fig. 4.3). Methods for
Fig. 4.3 A thin film of radioisotope (e.g., film thickness for a representative alpha emitter based
on a mixture of silver and Po-210) is coated on a metal structural plate
4.2 Phase of the Radioisotope 181
creating thin films will be discussed in Sect. 4.3. A typical film type source only
exposes one side of the source to the transducer thus limiting the transport effi-
ciency of the charged particles to the transducer medium as described in Sect. 4.1.
In this configuration, half of the energy contained in the charged particles goes the
wrong direction. The remaining half of the energy is deposited in the source
material (self-absorption) and in the transducer medium.
It is instructive to examine the literature on fission fragment transport to under-
stand how various interfaces affect the transport efficiency in a nuclear battery. A thin
film fission source in which one side of the source is exposed to a transducer has a
transport efficiency on the order of 20–30 % [2, 3]. Here, the boron-10 source shown
in the references would be similar to an alpha emitting radioisotope. One possible
way to improve the transport efficiency is to use the emission of charged particles
from both sides of the source layer. This can only happen if the charged particle
emitting source is coated on a very thin membrane. One possible way to do this is to
use a thin material as a substrate from the charged particle emitting source such as
nano-carbon fibers woven into flat rectangular membrane (Fig. 4.4). The source
Fig. 4.4 A very thin membrane coated with a source (e.g., film thickness for a representative
alpha emitter based on a mixture of silver and Po-210) which allows charged particles to escape
from both the left side and right side to the transducer medium
Fig. 4.5 A thin fiber coated (e.g., fiber thickness for a representative alpha emitter based on a
mixture of silver and Po-210) or embedded (e.g., atoms of the Po-210 alpha emitter) with a
radioisotope that is suspended in the transducer
material can be coated on the membrane or embedded in the nano-fibers. In this

configuration, the charged particles from both sides of the layer can be transported to
a transducer medium thus improving the efficiency (the transport efficiency for a
membrane will be on the order of 50–60 %) [2, 4]. It is possible to further improve
the geometrical coupling by coating thin fibers or embedding the fibers with
radioisotope (Fig. 4.5). A thin fiber can improve the transport efficiency even more
(60–70 %) [4]. A spherical geometry (e.g., aerosol) is the optimal route to improving
the transport efficiency [5, 6]. If the radioisotope is in the form of an aerosol
(Fig. 4.6), the transport efficiency would be on the order of 70–80 % [2, 6].
4.2.2 Radioisotope in Liquid Phase
It is instructive to look at liquid fission reactor literature to understand the possi-

bilities. Some forms of uranium salt (uranium sulfate (U(SO4)2) and uranium nitrate
Fig. 4.6 A particle (e.g., particle radius for a representative alpha emitter based on a mixture of
silver and Po-210) in the form of an aerosol suspended in the transducer [5, 6]
(UO2(NO3)2)) can be dissolved in water to form an aqueous core reactor [7].

Uranium can also form liquids at high temperatures. An example being uranium
dissolved in a fluoride salt to form molten NaF-ZrF4-UF4 [8].
As discussed in Sect. 3.2.2.1, the use of a liquid semiconductor to mitigate
radiation damage to a nuclear battery was proposed [9]. As described in Chap. 3,
liquid selenium is a semiconductor. Global Technology Inc. developed a liquid
selenium Schottky cell as a transducer for converting the energy of ionizing radi-
ation into electricity. This technology has since been exploited by others using a
beta source that is mixed with the liquid selenium [10]. The reported efficiencies in
these studies were high. In Chap. 5, there is a discussion on why high efficiencies
are sometimes reported in nuclear battery literature.
Section 3.2.2.1 in Chap. 3 discusses a theoretical study on a liquid gallium
nuclear battery that was reported in which the liquid gallium was used an electrolyte
in an electrolytic cell [11].
4.2.3 Radioisotope in Gaseous Phase
Nuclear reactors in the gaseous phase have been studied and it is instructive to
examine this technology to understand how it might be applied to nuclear batteries.
Gaseous uranium, namely uranium hexafluoride (UF6), has been of interest as a
potential fuel for nuclear energy conversion (e.g., nuclear-pumped lasers) since the
1950s [12]. With a gaseous fuel, the fission fragment transport efficiency to the
gaseous medium would be 100 % if the transducer were part of the gas (Fig. 4.7).
Uranium hexafluoride is very corrosive and has a high self-absorption of photons
by the uranium hexafluoride molecule.
Fig. 4.7 Charged Particle transport from a gaseous radioisotope to a transducer in which the
isotope is embedded
Research on a nuclear battery based on a gaseous radioisotope was performed in

the early 1990s [13]. The focus of this work was a radioisotope, Kr-85, which
serves both as the emitter and a transducer. The interaction of beta particles emitted
from Kr-85 with the krypton gas leads to both ionization and excitation of the gas.
Ionized gases form weak plasmas which exhibits a low fraction of ions and a high
fraction of neutral species. Weak plasmas emit photons which can be used for
energy conversion. Examples of systems based on the formation of weak plasma
are various types of gas lasers (e.g., helium-neon laser) and fluorescent light bulbs.
In the case of Kr, the photons that are produced are predominantly excimer photons.
The excimer state is produced by either the ions generated by the interaction of the
charged particle with matter or by the excited states generated by the interactions of
the charged particle with matter. This leads to the efficient channeling of energy
from the charged particles into the excimer state (see reaction 11 in Table 4.1).
A sensitivity analysis of the reactions in Table 4.1 can be performed.
A sensitivity analysis for the production of the analogous Xe2 ð1RÞ excimer system
indicates that there are only 7 dominant reactions that impact the production of the
excimer state (Table 4.2). These same reactions are dominant in any of rare gas
excimers gases (e.g., Ne2 ð1RÞ; Ar2 ð1RÞ; Kr2 ð1RÞ; and Xe2 ð1RÞÞ. The reactions
incorporate both production and quenching mechanisms. The bulk of the energy
transferred from the ionizing radiation to the gas produces rare gas ion and meta-
stable states. The rare gas ion and metastable states have efficient channels to funnel
the energy to the rare gas excimer state through these 7 dominant reactions.
As discussed in Chap. 3, excimer is short for excited dimers. A dimer is a short
lived dimeric or heterodimeric molecule formed by two atoms where at least one of
the atoms has a completely filled valance shell (such as a rare gas). For example,
Table 4.1 Reactions used in the Kr model [14]

# Reaction Rate constant
(s−1, cm−3 s−1, or cm−6 s−1)
1 Kr* + 2Kr ! Kr2*(1R) + Kr 17 10−32
2 Kr** + Kr ! Kr* + Kr 100 10−10
3 Kr* + Kr ! Kr + Kr 35 10−15
4 Kr** + 2Kr ! Kr2** + Kr 10 10−31
5 Kr2** + Kr ! Kr* + 2Kr 10 10−11
6 Kr2*(1R) + Kr ! Kr2*(3R) + Kr 12 10−13
7 Kr* + Kr* ! Kr++Kr + e 20 10−9
8 Kr2*(1R) + Kr2*(1R) ! Kr+2 + 2Kr + e 50 10−10
9 Kr+ + 2Kr ! Kr+2 + Kr 25 10−31
10 Kr+2 + e ! Kr** + Kr 23 10−7
11 Kr2*(1R) ! 2Kr + hv 21 10−8
12 Kr** ! Kr* + hv 15 10−7
13 Kr2*(3R) + e ! Kr2*(1R) + e 18 10−7
14 Kr2*(1R) + e ! Kr2*(3R) + e 49 10−7
15 Kr2** + e ! Kr2*(3R) + e 50E-7
16 Kr2** + e ! Kr2*(1R) + e 20E-7
17 Kr2*(1R) + e ! 2Kr + e 13E-9
18 Kr+2 + e ! Kr* + Kr 23E-7
19 Kr* + 2Kr ! Kr2*(3R) + Kr 44 10−32
20 Kr2*(3R) + Kr2*(3R) ! Kr+2 + 2Kr + e 50 10−10
21 Kr2*(3R) + Kr2*(1R) ! Kr+2 + 2Kr + e 50 10−10
22 Kr2*(3R) ! 2Kr + hv 10 10−7
23 Kr2*(3R) + e ! 2Kr + e 1.3 10−9
24 Kr++Kr + e ! Kr* +Kr 10 10−26
25 Kr+2 + e + Kr ! Kr2*(1R) + Kr 10 10−26
26 Kr2*(1R) + e ! Kr+2 + 2e 50 10−9
27 Kr2*(3R) + e ! Kr+2 + 2e 50 10−9
28 Kr+2 + e + Kr ! Kr2*(3R) + Kr 10 10−26
29 Kr** + Kr ! Kr + Kr 10 10−15
30 Kr2*(3R) + Kr ! Kr2*(1R) + Kr 46 10−15
31 Kr** + e ! Kr* + e 80 10−7
32 Kr2*(3R) + e ! Kr2** + e 30 10−7
33 Kr** + Kr** ! Kr++Kr + e 50 10−10
34 Kr2** + Kr2** ! Kr+2 + 2Kr + e 50 10−10
35 Kr* + e!Kr + e 10 10−9
36 Kr++2e ! Kr* + e 10 10−20
37 Kr2** ! Kr* + Kr + hv 10 10−8
Table 4.2 The seven major Reaction number Reaction

reactions from the plasma
chemistry model in Table 4.2 1 Kr þ 2Kr ! Kr2 ð1RÞ þ Kr
that dominate excimer 7 Kr þ Kr ! Kr þ þ Kr þ e
production are shown [14] 11 Kr2 ð1RÞ ! 2Kr þ hm
13 Kr2 ð3RÞ þ e ! Kr2 ð1RÞ þ e
14 Kr2 ð1RÞ þ e ! Kr2 ð3RÞ þ e
22 Kr2 ð3RÞ ! 2Kr þ hm
36 Kr þ þ 2e ! Kr þ e
rare gases can’t form molecules in the ground state, but can form a molecule in an
electronic excited state (Fig. 4.8). Examples of excimer molecules include the rare
gas excimers Ar2* (where the “*” indicates an excited state), Kr2* and Xe2*, the
rare gas-halide excimers ArF*, KrF*, XeF*, ArCl*, KrCl*, XeCl* etc. and a
number of other excimer gas combinations [17]. When an excimer molecule decays
by spontaneous emission, a photon is given off by the bound excited molecular state
to an unbound state where the atoms making up the molecule become neutral and
independent. In an excimer gas mixture (a mixture of gases that form the excimer
state), both ionization and excitation contribute to the formation of the excimer
states (with about 50 % efficiency for photon production for rare gas excimers).
Excimers thus depend upon the formation of ions and metastable states by the
interaction of ionizing radiation with a gas. In an excimer forming gas, if the
pressure of the gas is high enough (usually greater than a half atmosphere), the
formation of excimers is favored over formation of atomic excited states (atomic
excited states lead to undesirable line emissions from the atom). Excimers emit in a
very narrow wavelength range (plus or minus 10 nm). Excimers do not have a
bound ground state and thus are not subject to self-absorption. Thus excimer gas
mixtures remain optically thin (meaning no self-absorption) over large size, large
power density and high pressure.
As previously discussed, if the excimer is a rare gas excimer, like the xenon
excimer, the bulk of the energy that goes into the formation of xenon ions and
xenon metastable states proceeds into the formation of the xenon excimer state.
From Table 4.3, it takes 21.9 eV to form an ion pair. The W* value for xenon
metastable state formation (42 eV per metastable state) is also known [17]. The
theoretical maximum efficiency for xenon excimer production (ηf) is the ratio of the
xenon excimer photon energy (7.2 eV) divided by the W value for ion pair pro-
duction plus the ratio of the xenon photon energy divided by the W* value for Xe
metastable production shown in Eq. 4.1 (Table 4.4).
7:2 7:2
gf þ ¼ 0:5 ð4:1Þ
21:9 42
Thus the theoretical maximum efficiency for producing xenon excimer fluores-
cence from ionizing radiation is about 50 % [25]. This value of 50 % is
Fig. 4.8 Energy level

diagram for the Xe excimer.
The xenon molecule is
weakly bound. When the
xenon molecule
spontaneously decays to the
lower level a photon is
emitted but as shown in the
figure is unbound. This means
that the energy is
predominantly will be
channeled through the
excimer path
Table 4.3 The average energy required to produce ion pairs in various gases [15, 16]
Gas Energy per Ion Energy per First Ionization Fraction of
Pair, W (eV) Metastable W*(eV) Potential (eV) energy
used in
Ionization
(I/W)
He (pure) 43 90 24.5 0.58
Ne (pure) 36.8 77 21.5 0.58
Ar 26.4 55.2 15.7 0.59
Kr 24.1 50.4 13.9 0.58
Xe 21.9 42 12.1 0.55
approximately correct for all rare gas excimers [17]. An indirect nuclear battery can
be constructed based on a first step in which excimer fluorescence is produced by the
interaction of the excimer gas with ionizing radiation [13, 26]. The second step is
transporting the excimer fluorescence to a photovoltaic cell. Finally the interaction of
the excimer fluorescence with the photovoltaic cell produces a voltage and current.
4.2.4 Gaseous-like Radioisotopes
In 1990, when uranofullerenes were discovered, the possibility of incorporating

them into a reactor as part of a gaseous-like core was proposed [27] (Fig. 4.9).
A uranofullerene is a carbon-60 cage with two uranium atoms trapped in the cage
Table 4.4 Theoretical and experimental efficiencies for rare-gas and rare-gas halide excimers [7]
Excimer k (nm) hm (eV) η theoretical η experimental Reference
Ar2 126 9.83 505 na [18]
ArF 193 6.42 33 na [18]
Ar2F 284 4.36 224 na [18]
Kr2 147 8.42 475 46.2 [19]
KrF 249 4.97 340 na [18]
Kr2F 415 2.98 168 na [18]
NeF 108 11.5 35 na
Xe2 172 7.2 477 68a [20]
7.2 45b [19]
7.2 39 ± 10 [21]
7.2 43c,61a [22]
45d,46e [22]
XeF 346 3.58 237 12.1 [23]
XeBr 282 4.39 291 15 [24]
XeI 252 4.91 372 na [18]
a
Fission fragments and gaseous phase reactant
b
Electron beam and gaseous phase reactant
c
Gamma rays and gaseous phase reactant
Fig. 4.9 Uranofullerene embedded in a laser medium showing a fission fragment escaping the
uranofullerene and reaction with the laser medium
[1, 27]. The uranofullerene is a large molecular-like structure which can behave like
a gas. Much is unknown about the properties of uranofullerenes but there is some
promise that the material will not be corrosive and may have limited self-absorption
problems. The fission transport efficiency of a uranofullerene is 100 %.

A gaseous-like core fission reactor was studied using uranofullerene as a fuel [27].
If a radioisotope is trapped by the fullerene, it would have a similar structure to a
uranofullerene. The fullerene with a trapped radioisotope can be embedded in a
transducer as a very unique nuclear battery design. It is possible to take isotopes
normally in the solid phase and trap them in a fullerene. The fullerene would have
gaseous-like properties. A mixture of an excimer gas transducer mixed with
fullerenes with trapped radioisotopes would have similarities to the Kr-85 concept
described above. It would effectively have transport efficiencies from the fullerene
to the transducer gas of nearly 100 %. The absorption properties of fullerenes with
trapped radiation isotopes is not known but large molecular structures such as this
may have acceptable absorption properties. The advantage of this concept is that it
is possible to use the most abundant radioisotopes that would normally be in the
solid phase in a high transport efficiency source configuration with gaseous-like
properties.
4.3 Phase of the Transducer
Transducers can be in the solid, liquid, gas or plasma phase. Examples of each
phase will be discussed.
4.3.1 Solid Phase Transducer
Transducers in the solid-state are the most common type of transducers. One of the
most common transducers is the p-n junction in semiconductors [28]. They are used
in betavoltaic cells and alphavoltaic cells. Other examples of solid-state transducers
are thermocouples used in RTGs, electrodes in direct charge nuclear batteries,
cantilevers, and so on.
4.3.2 Liquid Phase Transducer
There are few examples of liquid transducers. Water has been used as a transducer
for the generation of hydrogen through radiolysis [18]. When generating chemicals,
the efficiency of the reaction is measured by the G value (number of
molecules/100 eV deposited). Common reactions include the production of carbon
monoxide from carbon dioxide and the production of hydrogen from water
(Table 4.5).
The G value is related to the process efficiency. Consider that the chemical
product is a means of storing energy which can be released through a process such
190
Table 4.5 Radiolysis of water and Carbon Dioxide [18]

Parameter Yields
Reaction Radiation Dose rate eV/g-s Pressure Temperature Addatives Dose G Mol/100 eV Product
atm K MRad
Rad + CO2 ! CO + ½O2 5.5 MeV a 1 1018 0.2 300 None 0.1 CO
1.5 MeV Protons 1.3 1019 0.4 300 None 4.25 CO
4
5.5 MeV a 1.7 1016 0.8 300,453 1 % NO2 2 -175 4.4 CO

Fission fragments 1.3 1019 0.8 348 1 % NO2 100 9–10 CO
9 1020 1.2 348 None 10 CO
2 1014 13 540 None 1.4–7.3 CO
2 1020 13 540 0.8 %NO2 10 CO
Rad + H2O(l) ! H2 + ½O2 Fission fragments 2 1014 to 2 1020 Liquid Variable None 1.5–1.8 H2
Rad + H2O(v) ! H2 + ½O2 Gamma 2 1014 to 2 1020 Variable Variable None 0.5 H2
Power Density Dilution Due to the Interface …
4.3 Phase of the Transducer 191
as oxidation [29]. In the case of hydrogen, the energy stored in hydrogen molecule
is 2.512 eV (assuming it reacts with oxygen and forms a water molecule). A typical
G value in Table 4.5 for water radiolysis is 1.7 hydrogen molecules per 100 eV
deposited. Thus the effective efficiency is,

eV
2:512 H Molecules
gHydrogenRadiolysis ¼ GðH2 MoleculesÞ 2
¼ 0:0437 ð4:2Þ
100 eV
Thus for this specific process in which G = 1.7, the effective production effi-
ciency is 4.27 %. Radiolysis is not very efficient because there are many possible
back reactions which curtail the production efficiency of forming hydrogen
molecules.
There are transducers which consist of interfaced phases such as a combination
of liquid and solid. As previously discussed, one such transducer uses a liquid
semiconductor to mitigate radiation damage (a problem common in a solid trans-
ducer such as a betavoltaic where atoms are displaced in the crystal lattice that is
caused by interactions with ionizing radiation). As noted earlier, the liquid semi-
conductor battery was developed to minimize the effects of radiation damage [9].
The cell is formed by two metal plates that have a liquid semiconductor (such as
selenium) between them, one forming a Schottky barrier and an Ohmic contact.
This forms a potential barrier which serves as the driving force that separate
electron-hole pairs produced by the interaction ionizing radiation with matter and
creates current flow. The conductivity is related to rapid temperature changes.
Molten materials in general will have metallic properties. Recall that the chalcogens
(oxygen, sulfur, selenium, tellurium and polonium), do not retain metallic prop-
erties during melting but can become semiconductor like. Hence selenium is ideal
and particularly interesting as it exhibits semiconductor behavior at high temper-
atures, and selenium is the preferred embodiment of the cell. Additionally, the
depletion region that is formed in the cell by the Schottky barrier has a collection
efficiency that reaches 100 %, but will have a rapid drop off at a diffusion length or
so beyond the depletion region. A test cell was built by Global Technologies Inc.
and operated at an efficiency of about 1 %. The Schottky barrier is a potential
energy barrier between a metal and a semiconductor that forms a diode. It depends
on the differential work function between a semiconductor and a metal contact. The
surface interface between the metal and semiconductor forms a depletion region
that is effectively 100s of nanometers thick. Thus the effective size of the energy
conversion region is thin and not well matched to the ranges of ionizing radiation.
The theoretical maximum efficiency for Schottky barrier structures is approximately
0.3 % for alphas and approximately 0.1 % for low energy betas [28]. But the
diffusion lengths in a liquid medium should be larger than those in a solid which
should lead to higher efficiencies than solid-state Schottky cells. The 1 % efficiency
reported is feasible for an alpha source.
Another example of interfaced phases is a water based nuclear battery that was
reported which uses an interface between liquid water and a solid (Pt/TiO2) [30].
Fig. 4.10 A Kr-85 indirect

nuclear battery in which the
Kr serves as both the source
and the transducer. The beta
particles excite the Kr gas and
excimer photons are
produced. The photons
interact with wide band-gap
photovoltaic cells surrounding
the sphere and produce power
The Pt was sputtered on nanoporous TiO2 which was formed by anodizing and
oxidizing a thin Ti film. The Pt/TiO2 forms a 0.6 eV Schottky barrier. A solid Sr-90
source was used as the beta source. Beta particles produced by the source streamed
through water and then through the Pt/TiO2 Schottky barrier. This device will be
discussed in more detail in Chap. 5.
4.3.3 Gas Phase Transducer
It is difficult to find examples of gaseous transducers. One might mistakenly call the
Kr-85 based nuclear battery a gaseous transducer that produces excimer fluores-
cence [13]. However, this transducer is a better described as a weak plasma which
makes it a plasma phase transducer as discussed below.
4.3.4 Plasma Phase Transducer
An indirect battery which uses Kr-85 as both the source and the transducer is a
plasma phase transducer because the self-excitation of the Kr gas by the emitted
beta particles forms a weak plasma (Fig. 4.10). From this weak plasma, excimer
states are formed which produce excimer fluorescence [13].
4.4 Surface Interface
To completely characterize a surface source, the method of manufacturing the

source is important. The source must have adhesion to the structure on which it is
attached. The atomic makeup and geometry of the source needs to be known.
A great deal of work from nuclear-pumped laser research has been done in finding
4.4 Surface Interface 193
the probability that a charged particle generated by a nuclear reaction escapes the
material in which it is embedded [1]. The efficiencies that have been found for
various types of source designs are also applicable to nuclear battery source
designs. The source designs and efficiencies will be discussed as well in this
section.
4.4.1 Methods of Forming Surface Sources
One of the difficulties in following nuclear battery research literature is that most
studies fail to provide enough information for a reader to fully understand the
experiment/work and to properly interpret the results. The problems begin with a
complete description of the important variables. These variables include informa-
tion about the radionuclide source, the mix of materials which make up the source
(e.g., a fairly common practice for a surface source is to mix the radionuclide with
gold for example), how the source is manufactured, the dimensions of the source
and how the source is geometrically coupled to the cell. Radioisotopes are extre-
mely hazardous.
Firstly, even a small amount of a radioisotope can be deadly (e.g., one micro-
gram of Po-210 or 0.0045 Ci). Given that it will take about 1000 Ci of Po-210 to
make a 1 W nuclear battery, the huge amount of Po-210 that would be needed
represents a major problem. Secondly, creating a surface source from Po-210 brings
up a number of issues. The Po-210 must not flake off the surface. It must be
encapsulated with multiple barriers (a philosophy employed with nuclear fuel).
Thirdly, in forming the source wastage of Po-210 atoms must be kept at a mini-
mum. Any Po-210 atom not put into the source leads to higher costs and more
contamination.
4.4.1.1 Thin Foil
Commercial polonium-210 sources (Po-210 is an alpha emitter) mix the radioiso-

tope with silver at a 1:10 ratio to form the source material. This metal mixture is
then rolled into a thin foil. The foil must be thin enough to allow the alpha particles
to escape.
It is feasible to press the thin Po-210 source foil onto the surface of a semi-
conductor. But, this method of attachment will not comply with U.S. Nuclear
Regulatory Commission standards which require that a radioactive source of a
certain size to have three methods of encapsulation to be transported [31].
An example to consider is a 5 lm thick foil pressed onto a SiC p-n junction that is
1 cm by 1 cm. In order to fully describe the nuclear battery cell, it is important to
know the ratio of Po-210 to silver, the surface area of the p-n junction, the thickness of
the source (t1), the activity of the source, the thickness of the n-type layer (t2), the
depletion width (t3), and the thickness of the p-type layer (t4) as shown in Fig. 4.11.
Fig. 4.11 Layout of a typical

p-n junction based
alphavoltaic using a polonium
210 foil made of 10 %
polonium and 90 % silver
placed on a silicon carbide
p-n junction
The thicknesses are needed to model the path of alpha particle emissions from the
source into the depletion layer. The foil will absorb some of the energy of an alpha
particle. For example, commercially available encapsulated alpha sources typically
lose *10 % of the alpha particle energy in the source structure [32] (Fig. 4.12). It is
reasonable to assume that the Po-210 atoms are uniformly distributed in the foil. In
setting up a Monte Carlo transport model, the location of the atom which decays in the
foil, the time of decay, and the emission angle can be incorporated. Since alpha decays
are isotropic, there is an equal probability that the alpha particle will be emitted at any
Fig. 4.12 Typical design of a

Po-210 alpha source. The
polonium 210 foil is made of
10 % polonium and 90 %
silver
4.4 Surface Interface 195
possible solid angle. Thus, half of the emission trajectories are away from the surface
of the p-n junction. The Monte Carlo transport model would then follow the trajectory
of the alpha particle and determine its path and where electron-hole pairs are created.
Thus, a complete description of the geometry of the device is necessary. If an
experimental paper fails to provide the complete description, then the experiment is
not fully described. The reader can’t model the alpha or beta voltaic cell without
making potentially invalid assumptions about the missing variables, nor can the
experiment be duplicated. Most nuclear battery research papers are lacking in their
discussion of important physical properties of the cell [28].
4.4.2 Electroplating, Painting and Baking
Processes have been developed to form thin films of enriched uranium (U-235 enrich-
ment) on the interior surface of a cylinder used for a fission chamber. These processes
include electroplating or painting the uranium on the inner walls of the cylinder.
Additionally, boron-10 can be plated on the inner surface of a cylindrical
chamber. One process used for producing a boron-10 lined detector utilized a
colloid where boron-10 particles are mixed in oil. The oil is placed inside of the
cylinder. The cylinder is heated while rotating so that the colloid uniformly wets the
surface while being baked. The oil eventually evaporates and leaves behind a
residue made up of the boron-10 particles. These particles adhere to the walls and
form a fairly uniform and thin boron-10 layer.
Similar processes can be adopted for the formation of thin films of radioisotopes.
4.4.3 Evaporation and Sputtering
Fission chambers also can use evaporation. For this process, enriched uranium is
heated in a vacuum to a relatively high temperature such that the metallic uranium
evaporates. When the uranium atoms condense on a cool surface, they create a thin
film [33]. This same process can be used to deposit a thin film of radioisotope on a
nuclear battery structure. The problem with evaporation is that the atoms of the
evaporated material are emitted from the source isotropically. Unless the surface area
of the substrate is large enough to collect most of the atoms, there will substantial
losses of radioisotope. Another issue is adhesion of the radioisotope to the surface.
Sputtering deposition is a Physical Vapor Deposition (PVD) process [33].
Sputtering needs an energy source such as Radio Frequency (RF) to eject atoms
from a target material. The atoms come off the target with a wide energy distri-
bution (with maximum values of 10 eV). The energetic atoms travel in a straight
line in a vacuum and then collide with the surface of the substrate. Due to the
energetic interaction at the surface, thin films that are sputtered can adhere to the
surface better than evaporated films.
4.4.4 Implanting
Ion implantation uses an ion beam directed at a substrate to interact with the
substrate. Based on the energy of the ions, the depth at which the ion is implanted is
dependent upon the energy of the ion beam [33].
4.5 Charged Particle Escape Probability from Surface

Sources
The physics of charged particle emission from thin films in which nuclear reactions
take place have been extensively studied in the field of nuclear-pumped lasers [1].
There are strong similarities between nuclear-batteries and nuclear-pumped lasers.
Both technologies depend upon the production of ionizing radiation from nuclear
reactions in order to drive a transducer. Albeit, fission reactions have a much higher
reaction rate (or power density) than radioisotopes.
In developing an example which incorporates escape probability, consider a
5 lm thick uranium metal foil in the arrangement shown in Fig. 4.13. The foil,
when exposed to a large current of thermal neutrons undergoes fission reactions
(e.g., U235(n, mn)ff). Power densities from a fission foil source will be calculated
from first principles as an illustration of the methodology and its usefulness for
nuclear battery calculations [2]. The efficiency of a rectangular (or slab) foil can be
Fig. 4.13 A uranium layer

coated on a thin membrane is
used as the surface source (a
full cell) to drive a transducer.
The foil thickness is 5 lm and
the channel width, r, is
0.01 cm
4.5 Charged Particle Escape Probability … 197
Fig. 4.14 The efficiency of a

U-235 metal foil (made with
highly enriched uranium) on a
thick substrate (or a
“half-cell”) is shown along
with a spherical uranium
particle as a function of fuel
thickness or radius
found using Fig. 4.14. The data from this Figure is for a thin film coated on a thick
plate (or a “half-cell”). Figure 4.14 is also useful for finding efficiencies of a
structurally self-supported thin film (e.g., uranium coating on a membrane in which
fragments can escape from both sides then it is a “full-cell”). The uranium metal
foil’s midpoint is 2.5 lm. As can be seen in Fig. 4.14, the transport efficiency for a
2.5 lm thick “half-cell” is 25 %. The “full-cell” efficiency is double that of the
“half-cell” efficiency, thus it is 50 % for this example.
4.6 Scale Length Matching
Another design consideration specific to nuclear batteries not of the thermal type is
that the scale lengths of the system are ‘well-matched’ [28]. Within the context of
this discussion, the range of a given form of ionizing radiation in a specific material
is referred to as the transport scale length of the radiation (kRadTr); the relevant
physical dimension of the energy conversion volume in the transducer is referred to
as the scale length of the transducer (Ltrans). It is desirable that these two scale
lengths, kRadTr and Ltrans, should be approximately equal in order to maximize
efficiency. Scale length matching is the primary factor that dominates the efficiency
of a nuclear battery; ‘well-matched’ systems have a higher maximum theoretical
efficiency, while systems which are not ‘well-matched’ have a lower theoretical
efficiency. Achieving ‘well-matched’ scale lengths is one of the primary challenges
encountered by nuclear batteries in the literature because of the respective param-
eters which determine each scale length.
4.6.1 Scale Length of Ionizing Radiation
Variables that influence kRadTr include: the mass, charge, angular distribution and
energy distribution of the source particles; the atomic number, density, and ion-
ization potential of the target material; and the mechanisms through which the
particle interacts with the target. These collectively cause kRadTr to vary greatly
among radioisotopes even for the same target material. The factors which determine
Ltrans include the energy conversion mechanism of the battery, the mechanical and
electrical properties of the target material, and the effect of radiation damage on the
target. The final factor is of essential importance to miniature nuclear batteries in the
literature, as new designs demonstrate persistent problems with radiation damage.
In contrast, the RTG design is not concerned with scale-length matching; the
sheer size of the RTG ensures that all of the radioisotope energy is deposited within
the transducer and converted to heat. However, all of the aforementioned concerns
must also be accounted for in a RTG-type miniature nuclear battery design as well.
Each one of these factors represents a challenge in matching the range of the
radiation to the scale length of potential transducers. In order to better define these
challenges, Chap. 3 has examined the characteristics and ranges of the various types
of radiation sources, focusing on those relevant to nuclear batteries. Chapter 5 will
examine the characteristics and associated scale lengths of various transducers
which are appropriate for energy conversion. The principles for integrating radia-
tion sources into transducers are then discussed in Chap. 5. Sources of radioisotopes
and availability have been discussed in Chap. 2.
Note, ionizing radiation is a broad term which refers to the fact that different
types of radiation will create ion pairs in matter; the category includes ions (e.g.,
fission fragments and alpha particles), beta particles, gamma rays, x-rays, and
neutrons, and each has a characteristic ionizing range. To a solid surface, heavy
ions such as fission fragments and alpha particles will deposit their energy within
micrometers, electrons in contrast will deposit over a larger range of millimeters.
Finally, those particles with no rest mass or net charge can reach meters.
4.6 Scale Length Matching 199
Fig. 4.15 SRIM2011 [34] was used to model a beam of mono energetic (5.037 MeV) alpha
particles that enter a silicon target normal to its surface. Shown are the tracks of the alpha particles
in the silicon. The beam spread is very narrow. This illustrates the concept of Linear Energy
Transfer
4.6.1.1 Ion Scale Length
Because ions are much more massive than the electrons they transfer their energy
to, while slowing down in a material their path length is linear (known as Linear
Energy Transfer or LET). Shown in Fig. 4.15 are tracks for a beam of 5.307 MeV
alpha particles entering normal to the surface of a silicon plate. As can be seen, the
beam spread is very minimal. This is a good illustration of Linear Energy Transfer.
4.6.1.2 Beta Scale Length
Beta particles on the other hand do not travel in a linear path because their mass is
equivalent to the mass of the electrons which they interact with as they pass through
matter. Compared to ions, the path of an electron in matter is complicated. Electrons
undergo significant scattering and follows a random walk-like path as shown in
Figs. 4.16 and 4.17.
Beta particles are emitted isotropically, betas are also emitted in an energy
spectrum. Thus, beta particles are complicated in that the scattering and energy loss
through the source medium must be accounted for as well and their angular
emission and energy spectrum. Using simplifications used to model beta particle
emissions will lead to significant errors. One common, but incorrect, simplification
is the monoenergetic assumption. In this assumption, all beta particles have an
energy of 1/3 bmax (rule of thumb used for estimating the average beta energy) [36].
This simplifies the modeling of the spatial energy deposition for the device, but as
will be shown, it is not accurate. Many models published in the literature use this
simplification. The ionization profile produced by a true beta energy spectrum is
significantly different from the results that use the 1/3 bmax assumption.
Fig. 4.16 A GEANT4

simulation of an electron
beam (generated by Sr-90
beta decay) entering normal to
the surface of a slab of SiC
which shows the beta particle
tracks (random walk path) and
bremsstrahlung photons
(straight lines) [35]
Fig. 4.17 A GEANT4

simulation of a Sr-90 point
source in the center of a
sphere made of SiC. The beta
particle tracks produced in the
decay of Sr-90 are shown
(random walk path) and
bremsstrahlung photons
produced in the interaction are
shown (straight lines) [35]
A second fatal simplification arises when it is assumed that the beta particles
emitted by the source are not isotropic. This simplification also occurs in the
literature. This simplification leads to unacceptable errors [37]. Models need to
include two elements (isotropic emission and true beta energy spectrum). Three
beta-decay reactions are chosen as examples to represent low emission energy
(Eq. 4.3), medium energy emission (Eq. 4.4) and high-energy emission (Eq. 4.5).

17 Cl þ b þ
!35 m þ 167:47 keV ð4:3Þ
35
16 S

39 Y þ b þ
!90 m þ 546 keV ð4:4Þ
90
38 Sr

40 Zr þ b þ
!90 m þ 2:28 MeV ð4:5Þ
90
39 Y
Table 4.6 shows pertinent data for the three beta emitters, including half-life, the
average beta energy according to the commonly-used rule of thumb (1/3 bmax), and
the average beta energy calculated using the beta spectrum. As can be seen in
Fig. 4.14, the sulfur-35 beta spectrum intensity continuously increases as energy
decreases, the beta spectrum intensity of the medium-energy strontium-90 emitter
tends to flatten out at low energies, and the high energy yttrium-90 beta spectrum
intensity has a distinct maxima and then drops as energy decreases. As can be seen
in Table 4.6, the differences between the average energy calculated by the 1/3 bmax
rule and the average energy calculated directly from the spectrum differs signifi-
cantly as the maximum energy of the beta particle increases. If the average beta
energy is calculated using the 1/3 bmax rule, the error inherent in using the 1/3 bmax
rule is propagated through the rest of the system calculations. These incorrect
average energies will then be used to calculate incorrect estimates of particle range
and stopping power. The 1/3 bmax rule should not be used for design calculations
and modeling of nuclear batteries.
Accurate range calculations are essential when designing a nuclear battery in
order to match the active region of the transducer (Ltrans) in the optimum position to
harvest the energy from the beta particle (kRadTr). To calculate the range of a beta
particle in matter, the full beta energy spectrum should be used in the model
(Fig. 4.18). Calculations which use the complete beta spectrum lead to the best
possible estimation of energy deposition profiles. This is clearly shown in the
Table 4.6 Characteristics of common beta-emitting radioisotopes

Isotope Half-life Max Average energy % Daughter
energy 1/3 bmax spectrum difference isotope
rule
S-35 87.51 days 167.47 keV 55.8 keV 53.1 keV +5 Cl-35
Sr-90 28.8 years 546 keV 182 keV 167 keV +9 Y-90
Y-90 2.67 days 2.28 MeV 760 keV 945 keV −20 Zr-90
The average energy is calculated using the 1/3 bmax rule and using a full spectrum analysis. The
differences in the average energy are substantial for high energy beta sources
Fig. 4.18 Beta emission

energy spectra for S-35
(solid), Sr-90 (dashed), and
Y-90 (inset) [35]
following example. The actual range of beta particles from S-35, Sr-90 and Y-90
decay have been calculated for a beam of beta particles hitting a slab and for a point
source in the center of a sphere [35]. These results are significantly different from
results that use the average beta energy calculated from the beta spectrum. This
result reinforces the premise that unacceptable inaccuracies come from the use of
any rules of thumb when designing a nuclear battery. It is interesting that there is
about a factor of 4 difference between the beta range calculated from average beta
energy and the beta range calculated with the full beta spectrum, where the range
for the full beta spectrum is greater.
Figures 4.19 and 4.20 represent the energy deposition as a function of distance
for a calculation based on the average beta energy and a calculation based on the
full beta spectrum, respectively. In the slab geometry, the beta particles were
simulated as monodirectional, normal to the stopping material, whereas the point
source in the sphere was modeled as isotropic. The results are notably different and
again reinforce why designs based on average beta energy have significant errors.
Looking at Fig. 4.15, the beta spectrum from each of the isotopes that are repre-
sented in this discussion, it is apparent that there is a significant distribution of low
energy beta particles which are emitted. By definition, the average beta energy is at
the point where the number of beta particles greater than the average energy equal
to the number of beta particles less than the average energy. For S-35, the low
energy beta population continually increases as the energy approaches zero. For
Sr-90, there is a slight peak at 0.08 MeV, but generally the population is flat at
lower energies. For Y-90, there is a well-defined maximum in the spectrum at
0.8 MeV. The differences between Figs. 4.15 and 4.16 are not so surprising when
the low energy betas in a full spectrum are taken into account.
In the calculations based on average beta energy shown in Fig. 4.19, there are
distinct peaks for both the slab and spherical geometries for S-35, Sr-90 and Y-90.
For S-35, the peaks occur at 6 lm for the slab geometry and at 10 lm for the
spherical geometry. For Sr-90, the peaks occur at 35 lm for the slab geometry and
at 65 lm for the spherical geometry. For Y-90, the peaks occur at 400 lm for the
slab geometry and at 600 lm for the spherical geometry. In contrast, Fig. 4.20
shows that low energy betas from the spectrum dominate where the energy is
deposited. Low energy betas will have a shorter range in the material than high
Fig. 4.19 Simulated energy deposition based on the average beta energy versus distance in both
spherical (right) geometries using GEANT4, PENELOPE, and MCNPX codes for a S-35, b Sr-90,
c Y-90. [35]
energy beta particles. So, the energy deposited per unit depth into the stopping
material for both S-35 and Sr-90 is highest near the surface of the stopping material
and decays exponentially with depth. For the high energy Y-90 beta particles there
is a distance where energy deposition peaks. For the slab case it peaks at 150 lm
and for the spherical case it peaks at about 300 lm. This is demonstrated in contrast
of Figs. 4.19, 4.20 and Table 4.7.
Fig. 4.20 Simulated energy deposition based on the full beta energy spectrum versus distance for
spherical (right) models using GEANT4, PENELOPE, and MCNPX codes for a S-35, b Sr-90,
c Y-90
Another interesting observation is that the maximum amount of energy that is

deposited is substantially different (about a factor of two lower) for the full spec-
trum calculation as opposed to the average energy calculation.
Table 4.7 Range of beta Radioisotope Range in millimeters

particles in SiC based on the
Rule of thumb Average Beta
rules of thumb mentioned in
beta spectrum
Chap. 3 and compared to
results in Fig. 4.19 and 4.20 S-35 10.6 0.02 0.08
Sr-90 55.1 0.12 0.40
Y-90 344.0 1.6 3.00
The results from the beta spectrum are exact and this table shows
the magnitude of expected errors in calculations which use rules
of thumb or average beta energy
Fig. 4.21 The collection probability for a linearly graded alpha or beta voltaic structure. The
collection probability is negligible beyond the diffusion length of carriers
4.6.2 Scale Length of Transducer
Transducers also have a broad range of scale lengths. These scale lengths are
specific to the transducer material. Below is a discussion of scale lengths in various
transducers.
4.6.2.1 Scale Length of P-N Junctions
The implications of the observations above are significant. First of all, in using
average beta energy to calculate the location of maximum energy deposition, and
therefore the location of the transducer, there will be substantial errors in placement
of the depletion layer within the cell. A betavoltaic cell is a p-n junction in which
the p-type material and n-type material form a junction through compensation. This
region in the cell is called the depletion zone. As will be discussed, by adjusting the
density of p-type impurities and n-type impurities, the depletion layer width will
change.
The collection probability is the probability that a carrier generated in a cell by
the interaction of radiation with matter in a region of the cell will be collected and
thereby contribute to the radiation generated current flow (IL). The carriers gener-
ated in the depletion region will have a collection probability of unity because the
electron-hole pairs are quickly separated by the electric field and eventually are
collected. Outside of the depletion zone, the collection probability decreases
because the electron-hole pairs must diffuse into the depletion region. If the distance
is more than one diffusion length away from the junction, the collection probability
is negligible (Fig. 4.21). For example, the diffusion length in SiC can vary from
0.07 lm to a few lm depending on material defects [38, 39].
In the determination of the width of the depletion zone, the conduction of charge
carriers is important, but even of more significance is the lifetime of the charge
carriers within the semiconducting material from the time of its creation or injection
to their subsequent collection by a true conductor, such as copper. The lifetime of
charge carriers, electrons and holes, in semiconductors, is governed by the traps that
are present within the band structure. The depletion region is limited to very small
thicknesses. This means that only the nuclear energy that is deposited within the
depletion region has a collection efficiency of unity for power production and thus
limits the efficiency of the energy conversion system. Equations 4.6–4.8 describe
this property through the intrinsic carrier concentration, ni, the built-in potential
voltage across the depletion region, Vbi, and the depletion region width, W [40].

2pkB T 3=2 3=4 Eg
ni ¼ 2 mn mp exp ð4:6Þ
h2 2kB T

kB T Na Nd
Vbi ¼ ln ð4:7Þ
e n2i

2es Vbi Na þ Nd 1=2
W¼ ð4:8Þ
e Na Nd
In these equations kB is Boltzmann’s constant, T is the temperature of the

semiconductor in Kelvin, e is the unit electron charge, es is the permittivity, Eg is
the band gap energy, Na is the relative concentration of vacancies available in the
p-type region, Nd is the relative concentration of electrons in the n-type region, and
mp and mn are the effective masses of the holes and electrons, respectively. As can
be seen in these equations the critical variables used in determining the built-in
potential barrier and the depletion region width are the intrinsic semiconductor
properties and the doping concentrations Na and Nd in the respective p- and n-type
regions of the semiconductor. Doping is the means used to change the built in
potential barrier and depletion width. The equations show that lower doping con-
centrations produce larger depletion widths but lower built in potentials.
In addition, the built in potential approaches zero as the doping concentration

product approaches the square of the intrinsic carrier concentration, meaning that no
depletion region exists. It is important to note that the doping concentration must be
greater than the injected charge from the depletion region in order for the transport
of the generated charges to be in the low injection regime. Impurity control in the
growth process of semiconductors severely limits the low end of feasible doping
concentrations (where parts per million impurity levels in a solid is a doping density
of about 1 1016 atoms per cm3). This level of impurity control is difficult to
achieve.
The depletion width is calculated for a widely used semiconductor in nuclear
battery research, 4H-SiC [28]. For this semiconductor es is about 10 [41], mn and mp
were taken to be 1.2me and 0.76me, respectively, and the band gap is approximately
3.25 eV at room temperature [42]. Using a temperature of 300 K in Eq. 4.6, the
intrinsic doping concentration for this material was found to be 9.5 10−9 cm−3.
This data was used in Eqs. 4.7 and 4.8. Both Nd and Na were then varied between
1015 and 1020 cm−3 to demonstrate the possible depletion widths for silicon carbide
as shown in Fig. 4.22. From the figure it can be seen that the largest depletion
region width in silicon carbide is 2.6 µm. The depletion width is the scale length for
a planar single p-n junction transducer. In order to achieve a depletion width of
2.6 µm, the impurity levels have to be on the order of 0.1 ppm. Achieving sub-ppm
impurity levels is difficult for SiC. It is more reasonable to obtain impurity levels in
the 1 to 10 ppm range and thus achieve a depletion width on the order of 1 µm [42].
Fig. 4.22 Density plot indicating depletion region width for varying donor and acceptor density
concentrations in 4H-SiC [28]
A Schottky barrier is formed when two different materials with different work
functions are placed in direct contact with each other. A Fermi potential will form at
the material interface. Because this boundary between the two materials is sharp, a
Schottky barrier will have a much thinner depletion zone than a p-n junction. The
width of the depletion region for a Schottky barrier, W, is expressed by [33]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2e2
W¼ ðVbi VA Þ ð4:9Þ
qN
Vbi
gdp ¼ ð4:10Þ
Eg
where Vbi is the built-in voltage in the Schottky contact, N is the dopant density, q is
the unit charge, es is the permittivity of the semiconductor, and VA is the applied
voltage across the junction in the forward bias. In the beta or alpha voltaic mode,
VA = 0. A typical barrier height for a SiC Schottky barrier is 1 V [43]. The
depletion width for a Ni/4H-SiC Schottky diode is given for several donor con-
centrations in Östlund [44]. With a dopant density of 1 1017 cm−3, the depletion
width of a Ni/4H-SiC Schottky diode is about 0.25 µm.
The depletion width of a Schottky barrier diode transducer is significantly less
than the depletion width of a p-n diode transducer. The other factor to note is that
the Schottky barrier diode also has a lower potential barrier height (Vbi) than a p-n
junction. The smaller depletion width will decrease the transducer scale length and
thus will lead to larger mismatches between the ionizing radiation range and the
scale length of the transducer.
Typically, the depletion layer width in a well-designed p-n junction will be about
1 lm thick. In order to envision how scale length mismatches effect nuclear bat-
teries, consider a beta source coupled to a p-n junction. Envision the betavoltaic cell
as a box and the depletion zone within the boundaries of the box. The challenge is
to deposit as much of the energy from the source’s beta particles into the 1 lm thick
depletion layer as possible. Realizing that the slab model is idealized in that a
mono-directional beta particle beam strikes the cell normal to the surface and that
the spherical model is idealized by placing a point source at the center of the sphere
where the beta source is equidistant from a shell within the sphere, both models
vastly over predict the beta energy being deposited in any given layer. In a realistic
device the beta source will be isotropic, thus creating a much greater challenge for
depositing the energy of the beta particles in the thin depletion layer. In summary,
by using average beta energies in design calculations, significant errors are made in
locating the optimum position for the depletion layer. There is also a significant
error in calculating the energy transfer rate to the depletion layer.
4.7 Geometrical Considerations 209
4.7 Geometrical Considerations
The geometry of the source and its interface to the transducer is an important
consideration in nuclear battery design. Nuclear reactions emit ionizing radiation
isotropically (having equal probability of emitting in any direction) as seen in
Fig. 4.23. The point source shown in the figure is a fundamental concept since any
geometrical structure can be made up with vast quantities of point sources. The
complexities of particle transport from a geometrical structure to a transducer are
many as shown below:
• Each point source emits particles isotropically so unless the transducer sur-
rounds the point source, some of the particles will never intersect with the
transducer.
• The particles must pass through the geometrical source volume to reach the
transducer, thus the particle will lose energy to the material due to coulombic
interactions.
• The amount of parasitic energy lost in the source volume is directly related to
the path that a particle takes through the source volume—i.e., the distance the
particle travels in the source material.
• The amount of energy that is ultimately transferred to the transducer is related to
the shape of the source, the scale of the source, the shape of the transducer and
the scale of the transducer.
Fig. 4.23 An illustration of

the principle of isotropic
emission is shown. Here a
point source emits particles
isotropically. The point
source is a fundamental unit
since any geometrical volume
can be constructed with a vast
number of point sources [1]
There has been a body of literature which examines the issues of source shapes
and scales and transducer shapes and scales in the areas of nuclear-pumped lasers
[1, 45], radiation chemistry [18, 29] and nuclear energy conversion [18]. It is
instructive to examine some of the lessons learned in these fields and to apply these
lessons to nuclear battery design.
The curvature of the source surface and the transducer surface impacts the
amount of energy that can be deposited in the transducer. Experiments at the
University of Missouri-Columbia demonstrated that surface sources coated on the
inside of cylindrical tubes had much lower power densities than surface sources
coated on flat plates inside a rectangular geometry [3]. This problem was important
to nuclear-pumped laser design because the range of ions generated in boron-10 or
uranium-235 coatings on the inside of a cylinder lead to small tube radius. When
the radius is small, the cylindrical geometry deviates substantially from slab
geometries. This impacts the properties of nuclear-pumped laser oscillation on the
lowest order transverse electric and magnetic field (TEM00) mode. Nuclear batteries
have similar problems- namely getting power into the transducer.
In a cylindrical geometry, a nuclear reaction can take place at any point in the
source material (Fig. 4.24). The particle is emitted from point P and has equal
probability of being emitted at any angle. The center point of the cylinder is at point
O. The radius of curvature of the cylinder, plus the small probability that a particle
emitted is able to go through point O, is a significant limitation on the power
Fig. 4.24 Diagram defining

the basic coordinate systems
and the projection of PQ onto
the cylindrical plane for a
cylindrical tube with a coating
of a nuclear reactant
containing material on the
inner surface (color red in the
diagram) [46]
density deposited at point O. Conversely, in a Cartesian geometry with two planar

sources, there is a plane at the center. The probability that a particle intersects the
central plane is much higher.
If the nuclear reactant is boron-10, it will undergo a reaction with thermal
neutrons (10B(n,4He)7Li with an energy release of 2.79 MeV). In 92 % of the
reactions the ions are emitted in an excited state 0.44 MeV above ground state.
Thus in 92 % of the reactions the 4He ion has an energy of 1.495 MeV (call this ion
HeI) and the 7Li ion has an energy of 0.855 MeV (call this ion LiI). In the
remaining 8 % of the reactions the 4He ion has an energy of 1.780 MeV (call this
ion HeII) and the 7Li ion has an energy of 1.015 MeV (call this ion LiII). Consider
the rate at which ions emitted in the nuclear reactions deposit their energy in a
transducer (or spatial power deposition). The spatial power density distribution is
much flatter in a slab than it would be in a cylinder. The curvature effect does lead
to a higher particle flux when comparing similar points in a slab to a cylinder.
However, the power deposition in a cylinder can be represented as a constant
(K) times the probability a particle will pass through the flux surface times an effect
due to the curvature.
Cylinder: Pd ¼ KPcy ðr)R ð4:11aÞ
Slab: Pd ¼ KPslab ðrÞ ð4:11bÞ
where r is the position of the flux surface relative to the coating, Pd is power density
(W cm−3), Pcy(r) is the probability a particle will pass through the flux surface for a
cylindrical geometry, Pslab(r) is the probability a particle will pass through the flux
surface for a slab geometry and R is the radius of curvature (a constant). It is clear
that the function Pcy(r) decreases as r approaches zero since the flux surface area
decreases with r until it collapses to a line when r = 0.
In an example problem, the gas which fills the cylinder is 0.1 MPa of helium and
the boron film thickness is 1.7 10−4 cm. The cylinder is placed in a neutron flux
(to trigger the 10B(n,4He)7Li reaction). The power deposition for each emitted ion is
shown as a function of r in Fig. 4.25.
The transport efficiency of cylindrical 10B coatings at various thicknesses were
examined and it was found that with an optimized 10B film thickness of 1.7 10−4
cm, the calculated efficiency was a poor 9 % (Fig. 4.26). This efficiency is sig-
nificantly lower than the 15 % that the slab model predicted [46]. Two very
important observations are (a) the correspondingly lower efficiency of a cylinder as
compared to the slab, and (b) the steeper power deposition gradient of the cylin-
drical versus the slab geometry. The radius of curvature of the surface of the source
places an important role in efficiency and power deposition.
Since most radioisotopes from metals or compounds are in the solid-state, the
most common means for interfacing an isotope to a transducer is to use a surface
source (Fig. 4.27). As discussed above, curvature in surfaces do present a problem.
Solid materials can be embedded in structures which have different shapes.
Fig. 4.25 Contribution from

each of the ions to the total
power deposition for a 10B
coating and 0.1 MPa of
helium gas [46]
As described previously, one such shape could be micro spheroids suspended in a

transducer material (Fig. 4.28) where the transducer could be a gas, liquid or solid.
The radioisotope can also be embedded in the material. In Fig. 4.29, fibers are
suspended between two plates with the fibers embedded in a transducer material
(gas, liquid or solid). The fibers can be made by mixing the radioisotope with a
material from which the fiber is formed or the fibers can have radioisotopes directly
ion implanted into them.
A hybrid approach to addressing the self-absorption problem is to form
micro-bubbles with excimer gases in the solid-state material [47]. In this example,
Kr-85 is the radioisotope that is implanted. Micro-bubbles can be formed at very high
pressures in a solid-state material using ion implantation (up to 4 GPa). At 4 GPa, the
density of a krypton gas bubble is on the order of 4 g/cm3. The transport length of
radiation in a high pressure krypton micro bubble is about 5 micrometers, about the
scale length of a heavy ion. As shown in Fig. 4.30 the radioisotope micro-bubble lies
above a p-n junction transducer. The particle from the radioisotope is emitted
Fig. 4.26 Efficiency of a cylindrical geometry as a function of thicknesses of boron-10 coated

film on the inner surface of the cylinder
Fig. 4.27 Diagram of an alphavoltaic cell
isotropically and the micro-bubble serves as both a shield to protect the junction from
the ionizing radiation as well as a photon source which emits at the excimer wave-
length. The photons then bounce around in the PV cell structure until they are
absorbed in the p-n junction. Even at this high density the issues of pressure
broadening should not lead to losses and the micro-bubble should not self-absorb.
Thus the cell will have a transducer scale length compatible with the radiation source
and with the PV cell. The advantages of this approach are that the wide band-gap p-n
Fig. 4.28 A solid aerosol

particle, which contains the
radioisotope, can be
embedded in a transducer
material (solid, liquid or gas).
Here the aerosol is small
enough to allow a substantial
fraction of the energy in the
charged particles produced by
radioactive decay to escape
the aerosol and deposit their
energy in the transducer
Fig. 4.29 This figure shows

fibers which are suspended
between two plates. The
material surrounding the
fibers is the transducer. Here
the fiber is thin enough to
allow a substantial fraction of
the energy in the charged
particles produced by
radioactive decay to escape
the fiber and deposit their
energy in the transducer
Fig. 4.30 A radioisotope can be ion implanted into a transducer material to form a microbubble.
In this figure the radioisotope is Kr-85. Here the Kr-85 is a beta emitter and a transducer which
produces ultraviolet photons. The optical transparency of the microbubble is sufficient to allow the
photons to escape the bubble into the transducer
structure will use a thin film with the radioisotope coated or imbedded into the
structure. Wide band-gap materials can operate at high temperatures without effi-
ciency loss and have high thermal conductivities. The films can be stacked, which
will allow for scaling of power densities. Problems with this approach do exist. Even
though it is well known that micro-bubbles form by ion implantation, the possibility
of the bubble delaminating may be a problem [28].
There are probably more potential interfaces between fuels of various phases to
transducers of various phases that haven’t been discussed. Suffice it to say that
human imagination is the ultimate limitation.
4.8 Power Density Dilution Factors for Surface Interfaces
The nature of the fuel to transducer interface is that it absolutely will dilute the
theoretical maximum power density (1/ BVWmin). The simplest way to define a
dilution factor is to look at a few examples. Suppose that an alphavoltatic is
powered by Po-210. From Table 1.3 in Chap. 1, (1/BVWmin) for Po-210 in pure
metal form is 1315.44 (W cm−3). As described in Sect. 4.4.1.1, Po-210 is mixed
with silver at an atomic ratio of 1to10. The first step is to calculate the atomic
density in the radioisotope containing source material and compare this to the
atomic density of the radioisotope in the compound of the radioisotope which has
maximum atomic density. Call this the atomic dilution factor (DFatomic) . The
second step is to calculate the ratio of the volume of the source material to the total
volume of the cell (source plus transducer). Call this the volume dilution factor
(DFvolume). The dilution factor for the cell (DFcell) is the product of atomic dilution
factor and the volume dilution factor.
Assume that the silver plus Po-210 film has been rolled into a 5 lm film and is
put on a cell with the geometrical configuration shown in Fig. 4.23 (where
t1 = 5 lm, t2 = 10 lm, t3 = 1 lm and t4 = 10 lm). The thicknesses of these layers
are based on the transport properties of the 5.3 MeV alpha given off by Po-210
decay. The proper method of determining the thicknesses is to run a transport code
(such as GEANT4) and to make t1 thin enough so that self absorption of alpha
energy in the source material is minimized while maximizing the power density.
Also, the amount of power deposited in the transducer layer, t3, is optimized.
In calculating DFatomic, consider that the fuel containing layer is t1. In analyzing
this layer, realize that the atomic density of silver is 0.0586 10−24 atoms cm−3
with a mass density of 10.5 gm cm−3. With a 1 to 10 atomic ratio of Po-210 to
silver, there will be 0.00586 10−24 atoms cm−3 of Po-210 and 0.05274 10−24
atoms cm−3 of Ag. The mass density of the t1 layer will be approximately 11.5 gm
cm−3. Pure metallic polonium has an atomic density of 0.0265 10−24 atoms
cm−3. The atomic dilution factor for the film (DFatomic) is,

DFatomic ¼ 0:00586 1024 =0:0265 1024 ¼ 0:221 ð4:12Þ
The volume dilution factor is volume (DFvolume) of the fuel layer divided by the
total volume of the cell,
DFvolume ¼ t1 =ðt1 þ t2 þ t3 þ t4 Þ ¼ 5=26 ¼ 0:192 ð4:13Þ
The dilution factor of the cell is,
DFcell ¼ DFatomic DFvolume ¼ 0:221 0:192 ¼ 0:042 ð4:14Þ
This means that the effective power density (Pdeffective) available in the cell (not
considering transport and transducer inefficiencies) is,
Pdeffective ¼ ð1=BVWmin Þ DFcell ¼ 1; 315:44 0:042 ¼ 55:24 W cm3 ð4:15Þ
Note that the source power (Psource) is related to the effective power density by
multiplying it with cell volume: Psource = Pdeffective*Vcell.
4.9 Power Density Dilution Factors for Volume Interfaces
In calculating power density dilution factors for volume sources, the procedure
outlined in Sect. 4.8 can be used (namely find the atomic dilution factor and the
volume dilution factor). The Kr-85 battery example of Chap. 3 is reiterated here.
The Kr-85 gas is held in a vessel at 1000 atmospheres. The power density from the
beta particles released in the high pressure Kr-85 gas is 1.76 W cm−3. As a gas, the
Kr-85 atoms are simply pressurized in a chamber. There is no atomic dilution so
DFatomic = 1. Since the Kr-85 serves both as the radioisotope and as the transducer
there is no volume dilution factor so DFvolume = 1. In essence due to the unique
features of Kr-85, it represents a nearly ideal system where no dilution occurs.
4.9 Power Density Dilution Factors for Volume Interfaces 217
Another isotope which could serve both as the radioisotope as well as the transducer
is Ar-39.
It is possible to interface solids with a transducer by embedding the isotope in
the transducer material as a particle or an implanted ion. The method of calculating
dilution uses the same approach as Sect. 4.8; namely find the atomic dilution factor
and the volume dilution factor. Suppose that Pm-147 is ion implanted into a 25 lm
thick silicon block at a uniform density of 1 1019 atoms cm−3 in a 5 lm layer.
The atomic density of pure Pm is 0.4095 1022 atoms gm−1. The atomic dilution
factor is,
DFatomic ¼ 1 1019 =0:4095 1022 ¼ 0:002442 ð4:16Þ
The volume dilution factor is,
DFvolume ¼ 5=25 ¼ 0:2 ð4:17Þ
The cell dilution factor is,
DFcell ¼ 0:002442 0:2 ¼ 0:0004884 ð4:18Þ
The (1/BVWmin) for Pm-147 from Table 1.3 is 2.99 W cm−3. This means that
the effective power density (Pdeffective) available in the cell (not considering
transport and transducer efficiencies) is,
Pdeffective ¼ ð1=BVWmin Þ DFcell ¼ 2:99 0:0004884 ¼ 0:00146 W cm3 ð4:19Þ
4.10 Summary
Radioisotopes are usually in the solid or gas phase. In Chap. 1, the maximum
possible power density that ionizing radiation can have occurs in compounds which
have the highest possible atomic density of isotope. In order to interface the isotope
with a transducer, the mean free path of the ionizing radiation in the isotope
containing compound as well as the mean free path of the ionizing radiation in the
transducer needs to be considered. The isotopes need to be contained in a material
which bounds them so they are not able to migrate. The material has to have a scale
length which is on the order of the mean free path of the radiation so that the
optimum amount of energy contained in the radiation can be deposited in the
transducer. Based on the phase of the source material containing the radiation, the
interface can be a surface interface or a volume interface. Due to the nature of the
interface, the atomic density of the radioisotope will be diluted and there will be a
volume dilution factor. The dilution factor will reduce the available power density
from the isotope. Unfortunately, there are still other factors which affect the effi-
ciency of the nuclear battery.
Problems
1. Some transducers are made from semiconductor material and perhaps a rea-
sonable material simplification is doped Si. Assuming no self-adsorption cal-
culate the typical path length of 1 MeV electron (i), proton (ii), and neutron
(iii) into an infinite medium. In your calculations observe the number of
interactions and energy deposited over each interaction. Then comment on the
quality of this energy and the potential for the radiation to drive a current
through a circuit. (Hint: start with the simplification of pure Si material, then
add consequences of impurities)
2. Refer to Sect. 4.2.4 and comment on the probability of interference and
self-adsorption risks of a uranofullerene.
3. Calculate the above.
4. Calculate the same for a captured plutonium molecule in the fullerene.
5. What is the guiding equations of energy capture of a radioactive particle
embedded in its transducer?
6. Given manufacturing was no issued but took a significant period of time, what
would be the ideal radioisotope selection? What would be its shelf life and
resulting viability as a nuclear battery source?
7. Po-210 is mixed with silver at an atomic ratio of 1to10. Calculate the ratio
(DFatomic) of the atomic density in the radioisotope containing source material
to the atomic density of the radioisotope in the compound of the radioisotope
which has maximum atomic density. Assuming the volume dilution factor is
20 %, what is the dilution factor for the cell (DFcell)?
8. Using the tables in Chap. 1, look up the BVWmin value for Po-210. Use it and
the DFcell from the previous Po-210 problem to calculate the effective power
density available in the cell.
9. Several different geometries are under consideration for the development of a
nuclear battery. The options include a point source centered in a sphere, a point
source centered in the middle of a box, a surface source coated on the inside of
a hollow cylinder, and a surface source coated on a flat plate. In terms of
efficiency only, what is the best design? Justify your answer.
10. Explain the concept of scale length matching as applied to nuclear batteries.
11. Explain how the dilution factor increases the size of a nuclear battery.
References
1. Prelas MA (2016) Nuclear-Pumped Lasers. Springer International Publishing, Cham

3. Chung AK, Prelas MA (1984) The transport of heavy charged particles in a cylindrical
nuclear-pumped plasma. Nucl Sci Engg 86:267–274
4. Platzmann RL (1961) Total ionization in gases by high energy particles: an appraisal of our
understanding. Int J Appl Rad Isot 10:116
References 219
5. Prelas MA (2016) Title, unpublished|

6. Prelas M, Boody F (1982) Charged particle transport in Uranium Micropellets. In: presented at
the IEEE International Conference on Plasma Science, Ottawa, Ontario
7. Prelas MA, Boody FP, Miley GH, Kunze J (1988) Nuclear driven flashlamps. Laser Part
Beams 6:25–62
8. Lee MYJJ, Simones MMP, Kennedy JC, Us H, Makarewicz MPF, Neher DJA et al (2014)
Thorium fuel cycle for a molten salt reactor: State of Missouri feasibility study. ASEE Annu
Conference. IN, Indianappolis, p 28
9. Tsang FY-H, Juergens TD, Harker YD, Kwok KS, Newman N, Ploger SA (2012) Nuclear
voltaic cell, ed: Google Patents
10. Wacharasindhu T, Jae Wan K, Meier DE, Robertson JD (2009) Liquid-semiconductor-based
micro power source using radioisotope energy conversion. In: Solid-state sensors, actuators
and microsystems conference, 2009. TRANSDUCERS 2009. International, pp 656–659
11. Patel JU, Fleurial J-P, Snyder GJ (2006) Alpha-voltaic sources using liquid Ga as conversion
medium. ed. NASA: NASA Tech Briefs
12. Boody FP, Prelas MA, Anderson JH, Nagalingam SJS, Miley GH (1978) Progress in
nuclear-pumped lasers. In: Billman K (ed) Radiation energy conversion in space, vol 61. ed:
AIAA, pp. 379–410
energy conversion. In: Yoshikawa M, Murakawa M, Tzeng Y, Yarbrough WA (ed) Second
International Conference on the Application of Diamond Films and Related Materials. MY
Tokyo, pp 5–12
14. Chung AK, Prelas MA (1987) Sensitivity analysis of Xe2* excimer fluorescence generated
from charged particle excitation. Laser Part Beams 5:125–132
15. Friedländer G, Kennedy JW (1955) Nuclear and Radiochemistry. Wiley
16. Friedlander G (1981) Nuclear and radiochemistry. Wiley, New York
Beams 6:25–62
of excited atomic and molecular states and chemical synthesis. In: Progress in Nuclear Energy,
vol 8, pp 35–52
19. Eckstrom DJ, Lorents DC, Nakano HH, Rothem T, Betts JA, Lainhart ME (1979) The
Performance of Xe2* as a photolytic driver at low e-beam excitation rates. In: Topical Meeting
on Excimer Lasers
20. Walters RA, Cox JD, Schneider RT (1980) Trans Am Nucl Soc 34:810
21. Prelas MA (1985) Excimer Research Using the University of Missouri Research Reactor’s
Nuclear-Pumping Facility. National Science Foundation
22. Baldwin GC (1981) On vacuum ultraviolet light production by nuclear irra-diation of liquid
and gaseous Xenon. Unpublished Report, Los Alamos Na-tional Laboratory
23. Miley GH, Boody FP, Nagalingham SJS, Prelas MA (1978) Production of XeF(B-X) by
Nuclear-Pumping
24. Boody FP, Miley GH (unpublished) Title
band gap electronic materials. Springer, ed, pp 463–474
26. Prelas MA, Hora HP (1994) Radioactivity-free efficient nuclear battery. Germany Patent
27. Mencin DJ, Prelas MA (1992) Gaseous like Uranium reactors at low temperatures using C60
Cages. In: Proceedings of Nuclear Technologies for Space Exploration, American Nuclear
Society, August 1992
review of nuclear batteries. In: Progress in Nuclear Energy, vol 75, pp. 117–148, August 2014
29. Prelàs MA, Romero J, Pearson E (1982) A critical review of fusion systems for radiolytic
conversionof inorganics to gaseous fuels. Nucl Technol Fus 2:143
30. Kim H, Kwon JW (2014) Plasmon-assisted radiolytic energy conversion in aqueous solutions.
In: Nature science reports, vol. 4
31. U.S.N.R. Commission (1991) NRC Regulations 10 CFR Part 20, ed. US Nuclear Regulatory
Commission
32. Syed A (2012) Modeling the energy deposition of alpha particles emitted from Po-210 source
on Silicon Carbide for possible nuclear battery and laser pump applications. M.Sc., Nuclear
Science & Engineering Institute, University of Missouri—Columbia, Columbia, MO
33. Sze SM, Lee M-K (2012) Semiconductor devices: physics and technology, 3rd ed. Wiley
(2010). Nucl Instrum Methods Phys Res Sect B 268:1818–1823
36. Bernard S, Slaback Jr Lester A, Kent BB (1998) Handbook of health physics and radiological
health. Williams & Wilkins, ed: Baltimore
and betavoltaltaic cells. M.Sc., Nuclear Science & Engineering Institute, University of
Missouri, University of Missouri—Columbia
38. Doolittle WA, Rohatgi A, Ahrenkiel R, Levi D, Augustine G, Hopkins RH (1997)
Understanding the role of defects in limiting the minority carrier lifetime in Sic. In: MRS
Online Proceedings Library, vol. 483, pp. null-null
39. Seely JF, Kjornrattanawanich B, Holland GE, Korde R (2005) Response of a SiC photodiode
to extreme ultraviolet through visible radiation. Opt Lett 30:3120–3122
40. Neamen DA (2003) Semiconductor physics and devices. McGraw Hill
41. Savtchouk A, Oborina E, Hoff A, Lagowski J (2004) Non-contact doping profiling in epitaxial
SiC. In: Materials Science Forum, pp 755–758
42. Huang M, Goldsman N, Chang C-H, Mayergoyz I, McGarrity JM, Woolard D (1998)
Determining 4H silicon carbide electronic properties through combined use of device
simulation and metal–semiconductor field-effect-transistor terminal characteristics. J Appl
Phys 84:2065–2070
43. Latreche A, Ouennoughi Z (2013) Modified Airy function method modelling of tunnelling
current for Schottky barrier diodes on silicon carbide. Semicond Sci Technol 28:105003
44. Östlund L (2011) Fabrication and characterization of micro and nano scale SiC UV
Photodetectors. In Student Thesis, Masters of Science, Royal_Institute_of_Technology, Ed.,
ed. Stockholm, p. 74
45. Melnikov SP, Sizov AN, Sinyanskii AA, Miley GH (2015) Lasers with nuclear pumping.
Springer, New York
46. Chung AK, Perelas MA (1984) The transport of heavy charged particles in a cylindrical
nuclear-pumped plasma. Nucl. Sci. Eng. (United States) 86:3 Medium: X; Size: pp 267–274
47. Prelas MA (2013) Micro-scale power source, United States Patent 8552616. USA Patent
(2013)
Chapter 5
Efficiency Limitations for Various Nuclear
Battery Configurations
Abstract A nuclear battery can be viewed as a radiation source embedded in

various layers of materials with one of the layers being a transducer. The goal of
nuclear battery design is to deposit as much of the power produced by the source
into the transducer. Nuclear battery designs vary depending on the source, the
transducer and the method by which the source and transducer are interfaced. It is
this variability in design which can obfuscate the simple nature of the design. This
chapter’s focus is on the fundamental vision of nuclear battery design.
Keywords Transducer Efficiency Radiation Damage Safety Problems
5.1 Basics of Nuclear Battery Design
In Chaps. 1–4, the background needed to understand how ionizing radiation

released by the decay of radioisotopes is transported from the source layer to the
transducer layer was given. A general scheme for this transport is shown in Fig. 5.1
which is illustrative of many types of nuclear batteries. As can be seen from Chap.
1 to 4, the underlying principles of ionizing radiation transport is complex and is
dependent on the properties of the ionizing radiation (spectra of emitted particles,
half-life, etc.), the properties of the materials in the layers, the geometry of the
design and the type of interface between the source and the transducer. With regard
to Fig. 5.1, the number of layers of materials is dependent on the specific design of
the nuclear battery. The two most important layers are the source layer and the
transducer layer. The nuclear battery efficiency (ηNB) is equal to the transducer
efficiency (ηtransducer) times the fraction of power deposited in the transducer or
power deposition efficiency gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 Þ;

P4
gNB ¼ gtransducer ð5:1Þ
P1 þ P2 þ P3 þ P4

222 5 Efficiency Limitations for Various Nuclear Battery Configurations
Fig. 5.1 The basics of nuclear battery design. A nuclear battery is composed of layers of
materials. The two most important layers are the source and the transducer. The efficiency of a
nuclear battery (ηNB) is basically defined as the transducer efficiency (ηtransducer) times the fraction
of power deposited in the transducer or the power deposition efficiency
gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 Þ
In the previous chapters, the efficiency terms for various steps in the energy
conversion process for specific nuclear battery concepts that go into the calculation
of ηpd have been discussed. This chapter looks at the efficiency of the transducer
which is used in the nuclear battery concept.
5.1.1 Transducer Efficiencies
Once the energy is deposited in the transducer, there is an efficiency of converting

the deposited energy into electrical power (e.g., a nuclear battery). There are other
useful energy forms that a transducer can produce (e.g., optical power, chemicals
which store energy, heat, etc.). These will not be discussed in this text but may be
found in other sources [1, 2].
5.1.1.1 p-n Junctions
The efficiency of a p-n junction transducer can be viewed very simply. Begin by
defining the power absorbed in the junction (Pabsorbed). Of this absorbed power, the
fraction of power that is used to raise an electron from the ground state into the
conduction band of the cell is called the electron utilization efficiency
(ηElectronUtilization). The product of these two terms is related to the energy expen-
diture for producing electron-hole pairs. Each electron produced is then part of the
current flow in the circuit. There are inefficiencies which are not counted such as
absorption of electron by traps or recombination. These inefficiencies will be ignored
5.1 Basics of Nuclear Battery Design 223
for the time being so that an idealized cell model can be formulated. Once the
electron-hole pair is produced, the driving voltage of the cell needs to be described in
the model. Ultimately, if the potential energy of the junction were fully utilized, the
maximum achievable cell voltage is related to the band-gap energy (Vmax = Eg·UC,
where UC is a conversion factor of energy to voltage equal to 1). No photovoltaic
cell is perfect so the open circuit voltage (Voc) is less than Vmax. An efficiency called
the driving potential efficiency (ηdp = Voc/Vmax) is introduced which basically is a
measure of how far the circuit’s open circuit voltage is from the ideal voltage (Vmax).
Last but not least, the maximum power output (PMaxOut) for the ideal cell is the open
circuit voltage times the short circuit current (PMaxOut = Jsc·Voc). The actual power
out of the cell is Pactual = J·V (where J is the actual current of the cell and V is the
actual voltage of the cell). The fill factor (FF) is defined as FF = Pactual/PMaxOut. So
the general rule for finding the efficiency of any energy conversion scheme using a
photovoltaic based transducer can be defined as.
gtransducer ¼ gElectronUtilization gdp FF ð5:2Þ
Photons use the photoelectric effect to create electrons in the photovoltaic circuit.
This is a resonance effect and it minimizes randomness in the energy transfer
process. This is a beneficial effect. However, photon sources can have an angular
distribution and an energy distribution. Both an angular distribution and energy
distribution are randomizing effects that will reduce the efficiency of a photovoltaic
transducer. For example, for a solar cell, the sun’s rays are paraxial (virtually no
angular distribution) but have a broad energy distribution. The broad energy dis-
tribution is a detriment to the photovoltaic convertor because the band-gap of the
cell need to be optimized for the photon energy distribution which the sun produces.
Thus, single band-gap photovoltaic cells struggle with low efficiencies. There are
strategies for using multi-junction photovoltaic cells that is based on stacking cells
with different band-gap energies so that the effects of a broad photon energy dis-
tribution are mitigated. Multi-junction photovoltaic cells are more efficient but also
more expensive.
Thermophotovoltaics have both an angular photon distribution and a photon
energy distribution to contend with. This impacts the efficiency of the ther-
mophotovoltaic energy conversion process.
Charged particles interacting with a photovoltaic material are a highly ran-
domizing form of creating electrons in photovoltaic materials. The energy transfer
mechanisms involved with charged particle interactions with matter have a sig-
nificant impact on the efficiency of an alphavoltaic or betavoltaic. Additionally,
both alpha particles and beta particles are emitted isotropically and this has a
devastating effect on the alphavoltaic and betavoltaic cells. Beta particles have a
very broad energy distribution and this too has a devastating effect on the efficiency
of betavoltatic cells.
In the following sections, the transducer efficiency is described for various
energy conversion schemes based on the use of photovoltaic cells.
The Use of Ionizing Radiation from Nuclear Reactions to Produce

Electron-Hole Pairs in a p-n Junction
To estimate the cell current I generated by beta or alpha energy absorbed in a

photovoltaic cell, one can find the maximum current that can be created in the cell
by the interaction of radiation with matter. The maximum current will be dependent
on the energy transport efficiency of the radiation (beta or alpha) to the depletion
zone. This can be expressed in a term which represents the fraction of power that is
deposited in the depletion zone and is a factor called ηd (Eq. (5.3)). This is typically
calculated by using Monte Carlo based transport Codes.
Power absorbed in depletion zone ¼ Pdpl ¼ Ptotal gd ð5:3Þ
The maximum rate of charge production in the depletion zone is then considered
through the examination of the number of electron-hole pairs that are created per
second in the depletion zone. As discussed previously, Monte Carlo radiation
transport codes do not have the capability of modeling electron and ion motion in
the media. The use of the W (eV/ion pair) value, which is the amount of energy that
it takes to make on average an electron hole pair, is imperative. Monte Carlo codes
have the capability of calculating the rate of energy absorbed spatially in a material.
But, as previously discussed the spatial and temporal distribution of electrons and
ions should be in close proximity to the spatial energy deposition in a solid. Thus a
reasonable estimate is to use the spatial rate of energy lost calculated by Monte
Carlo Codes to find the power density in the depletion zone and then use the power
density distribution to estimate the electron and ion density distribution. Therefore
the number of electron-hole pairs created per second (Ne) in the depletion zone is,
Ptotal ½J/sgd
Ne ½# pairs=s ¼ 6:25 1018 ½eV/J ð5:4Þ
W ½eV/ion pair
Assuming that the cell has no losses due to traps, the production rate of
electron-hole pairs is proportional to the maximum ideal short circuit current (Jsc) in
the junction. Ideal short circuit current is equal to the production rate of
electron-hole pairs multiplied by the charge per electron (1.6 × 10−19 C):
Jsc ¼ Ne ½# pairs/s 1:6 1019 ½Coulomb/pair ð5:5Þ
Jsc ¼ Ptotal gd =W ð5:6Þ
The power out (Eq. (5.7)) is related to the open circuit voltage (Voc) multiplied
by the short circuit current (Jsc) and the Fill Factor (FF). The maximum power out
that a p-n junction is capable of producing, Pmax, is the optimized power output of
the cell [3]. The FF is typically > 0.7 for high grade solar cells (Eq. (5.8)).
Voc Ptotal gd FF
Pout ½W ¼ Voc ½V Jsc ½A FF ¼ ð5:7Þ
W
Pmax
FF ¼ ð5:8Þ
Voc Jsc
The efficiency of the transducer is defined as,
Pout
gtransducer ¼ ð5:9Þ
Pin
Pin is the energy deposited in the transducer which for a p-n junction is,
Pin ¼ Ptotal gd ð5:10Þ
So,
Voc FF
W
The concept of the driving potential efficiency (ƞdp) was introduced by Oh et al.
[4]; it originates from the relationship between the open circuit voltage and the band
gap of the cell material as described in Eq. (5.12). The magnitude of the open
circuit voltage is less than or equal to the material band gap. The open circuit
voltage can then be represented as the product of the driving potential efficiency and
the band gap, shown in Eq. (5.13).
Voc
gdp ¼ ð5:12Þ
Eg
Voc ¼ gdp Eg ð5:13Þ
So,
Eg gdp FF
W
The driving potential efficiency is a simplified means of saying that the open
circuit is related to the material band-gap. This relationship can be determined by
beginning with the ideal PV cell-equivalent circuit in Fig. 5.2, the output current for
an ideal PV cell is related to the dark saturation current of the p-n junction (ID) and
radiation generated current (IL) by taking the nodal balance of the circuit in
Eq. (5.15).
The open circuit voltage is a function of the dark saturation current (I0) and the
radiation generated current (IL). The diode current is a function of the dark satu-
ration current in Eq. (5.16), where I0 is the reverse saturation current which is a
Fig. 5.2 Ideal alpha or beta

voltaic cell equivalent circuit
function of the material and temperature, q is the electron charge

(1.602 × 10−19 C), kB is Boltzmann’s constant (1.381 × 10−23 J/K), T is the cell
temperature in Kelvin and n is the shape factor (for an ideal cell n = 1). The open
circuit voltage occurs when the radiation generated current balances the diode
current. So setting I = 0 and using Eq. (5.16) for ID, one can find the open circuit
voltage as shown in Eq. (5.17).
I ¼ IL ID ð5:15Þ

qV
ID ¼ I0 exp 1 ð5:16Þ
nkB T

nkB T IL
Voc ¼ ln þ1 ð5:17Þ
q I0
There is also a relationship between band gap energy and open circuit voltage.
The dark saturation current I0 depends on the charge carriers due to the temperature
of the cell T. The relationship between the dark saturation current and band-gap Eg
is shown in Eq. (5.18).

qEg
I0 ¼ D T 3 exp ð5:18Þ
nkB T
As the band-gap increases, the dark saturation current will decrease. Thus, open
circuit voltage will increase for the ideal cell as band gap energy goes up. The
highest open circuit voltage achieved for a wide band-gap cell (e.g., diamond) was
2.6 V [5] which is equivalent to a driving potential efficiency (ƞdp) of 0.48. The
implications of the driving potential efficiency on the open circuit voltage and
energy efficiency is that in order to achieve the highest possible power output the
driving potential efficiency needs to approach the value of one [3]. As discussed,
wide band-gap materials will have higher driving potential efficiencies so it may be
possible to have a value greater than 0.48. Other important parameters for diamond
are a band-gap energy of 5.48 eV and a W value of 12.4 eV (Table 3.9). The fill
factor (FF) is a more difficult parameter to quantify. High quality photovoltaic cells
used in solar energy conversion have a FF value between 0.7 and 0.8. Thus it is not
unreasonable to assume that diamond photovoltaic cells could achieve a fill factor
of 0.8. Thus the estimated transducer efficiency of a diamond p-n junction used in a
nuclear battery is,
Eg gdp FF 5:48 0:48 0:8

gtransducer ¼ ¼ ¼ 0:17 ð5:19Þ
W 12:4
Here Eg/W is the electron utilization efficiency (ηElectronUtilization) for

electron-hole pair production by the absorption of alpha or beta particle energy in a
p-n junction.
Temperature Effects in p-n Junctions
The operating temperature of a p-n junction will have a profound effect on the
efficiency of the device. The intrinsic carrier density in a semiconductor is a
function of temperature and high temperature does significantly reduce the effi-
ciency of a nuclear battery [6, 7].
5.1.1.2 The Use of Ionizing Radiation from Nuclear Reactions

to Produce Electron-Hole Pairs in a Schottky Barrier
Schottky Barrier transducers follow a similar analysis to the p-n junction problem
discussed in the prior section. The major difference is in the definition of Eg.
A Schottky Barrier will have a Schottky Barrier Height (Vsbh) which is a function of
the electron affinity (ξ) of the semiconductor, the work-function of the metal (φ),
and the band gap (Eg).
For a metalized p-type semiconductor the Schottky Barrier Height is,
p
Vsbh ¼ Eg þ n u ð5:20Þ
For a metalized n-type semiconductor the Schottky Barrier Height is,

p
Vsbh ¼un ð5:21Þ
A typical driving potential efficiency (ƞdp) for a Schottky Barrier cell is 0.6. If
the Schottky Barrier is made with aluminum on diamond, the work function for
aluminum is 4.08 eV (Table 5.1). Other important parameters for diamond are a
band-gap energy of 5.48 eV and an electron affinity (ξ) which is negative
(−0.07 eV). The fill factor (FF) is a more difficult parameter to quantify. High
quality photovoltaic cells used in solar energy conversion have a FF value between
0.7 and 0.8. Thus it is not unreasonable to assume that diamond photovoltaic cells
could achieve a fill factor of 0.8. Equation (5.19) is modified for a Schottky Barrier
Table 5.1 Work functions Metal Work Function (eV)

for various metals
Aluminum 4.08
Beryllium 5.0
Cadmium 4.07
Calcium 2.9
Cesium 2.1
Cobalt 5.0
Copper 4.7
Gold 5.1
Iron 4.5
Lead 4.14
Magnesium 3.68
Mercury 4.5
Nickel 5.01
Niobium 4.3
Potassium 2.3
Platinum 6.35
Selenium 5.11
Silver 4.26
Sodium 2.28
Uranium 3.6
Zinc 4.3
in Equation 5.22. The estimated transducer efficiency for a p-type (boron doped)
diamond Schottky Barrier photovoltaic cell metalized with aluminum is shown in
Eq. (5.23).
p
Vsbh gdp FF
W
p
Vsbh gdp FF ð5:48 0:07 4:08Þ 0:6 0:8
gtransducer ¼ ¼ ¼ 0:052 ð5:23Þ
W 12:4
p
Here Vsbh =W is the electron utilization efficiency (ηElectronUtilization) for
electron-hole pair production by the absorption of alpha or beta particle energy in a
Schottky barrier.
5.1.1.3 The Use of Photons to Produce Electron-Hole Pairs in a p-n

Junction
A photon is absorbed by a photovoltaic cell if its energy exceeds the band-gap

energy of the semiconductor (Fig. 5.3). There is an energy dependent photon
Fig. 5.3 Energy utilization

efficiency for the absorption
of a photon from a spectral
source in a photovoltaic cell
spectrum as discussed in Chap. 3 (e.g., Eq. (3.73)). The energy distribution func-
tion of the photon spectra is ϕ(E). The function has the property that the integration
of the function over all possible energies is equal to 1 (Eq. (5.24)). If the total
number of photons absorbed in the photovoltaic cell is Nph, then the total number of
photons which have sufficient energy to move an electron into the conduction band
is given by Eq. (5.25) and is defined as the short circuit current for an ideal
photovoltaic cell. This is simply a statement that each electron that is raised into the
conduction band will contribute to the current flowing in the photovoltaic cell. The
fraction of the energy stored in the absorbed photons that are used to produce
electrons can be calculated for an ideal cell by the maximum intrinsic efficiency
calculation is shown in Eq. (5.26) (discussed in Chap. 3).
Z 1
/ðEÞdE ¼ 1 ð5:24Þ
0
Z 1
Nph
Jsc ¼ /ðE ÞdE ð5:25Þ
6:25 1018 Eg
R1 E
E W ðE Þ E dE
g
gin ¼ Rg1 ð5:26Þ

Eg W ðE ÞdE
Here ηin is the electron utilization efficiency (ηElectronUtilization) for electron-hole

pair production by the absorption of photon energy in a p-n junction.
The maximum power output from an ideal photovoltaic cell is defined as the
open circuit voltage (Voc) times the short circuit current (Jsc) as shown in Eq. (5.27).
The power output from the cell will be the maximum power output times the fill
factor (FF) as shown in Eq. (5.28).
Pmax ¼ Voc Jsc ð5:27Þ
Pout ¼ Voc Jsc FF ð5:28Þ

The power input into the photovoltaic cell is,

Z 1
Pin ¼ Nph E/ðE ÞdE ð5:29Þ
0
The cell efficiency is equal to the power out divided by the power in.
Pout
gpv ¼ ð5:30Þ
Pin
Photons can be lost by reflection, absorption in defects and transmission.

Electrons can be lost by absorption in defects or recombination. These losses are
not taken into account in the above equations. However, ideal efficiencies can be
calculated and these calculations can be useful. Suppose that the photovoltaic cell is
made of diamond. If the photons entering the cell are mono energetic with energy
equal to Eg, then it is feasible to do an ideal calculation. If 1.6 × 1019 photons/s
enter the cell and using Eg = 5.48 eV, ƞdp = 0.48 and FF = 0.8, then Jsc = 1×1019
electrons/s. The power in is 1.6 × 1019 photons/s times 5.48 eV/photon which is
equal to 8.768 × 1019 eV/s. Multiply this number by 1.6 × 10−19 J/eV which
equals 14.03 Watts. The open circuit voltage is 5.48 times 0.48 which equals
2.63 V. The short circuit current is 2.56 A. So, the ideal cell efficiency is,
2:63 2:56 0:8

gpv ¼ ¼ 0:384 ð5:31Þ
14:03
This analysis also holds for thermophotovoltaic energy conversion. In this case
the spectrum will be very broad because of the grey body emitter.
In summary, the photoelectric effect is intrinsically more efficient at creating
electron-hole pairs than ionizing radiation.
5.1.2 Direct Charge Nuclear Battery (DCNB)
As discussed in Chap. 3, the ideal efficiency of a parallel plate DCNB with a

mono-energetic isotropic source is,
p 2hm
gidea ¼ gf gm ¼ sin2 hm ð5:32Þ
2p
A DCNB will not be ideal. The source of ionizing radiation will typically have
an energy distribution. This energy distribution will reduce the ideal efficiency as
modeled by Eq. (5.32). The effects of energy distribution can be modeled. If the
source of ionizing radiation has an energy distribution function f(E), then the theory
in Chap. 3 for an isotropic mono-energetic source can be modified. The modifi-
cation begins by assuming that there is an optimum electric field between the source
Fig. 5.4 Shown is a diagram for an emitter with uniform emission (which due to symmetry allows
the problem to be solved in two dimensions) and a collector which can collect particles with
emission angles between θc and π − θc (due to symmetry this problem can be solved using half of
the angular emission space, θc to π/2 and multiplying the result by 2). The particles emitted within
this angular emission space have sufficient velocity to overcome the built up electric field
!
plate and the collector plate ( E ). In Chap. 3, it is assumed that the source is
mono-energetic and that a particle which is emitted at an angle greater than or equal
to the critical angle (θc) will be collected. At the critical angle, the x component of
the particle velocity (Vx) is just sufficient to overcome the Lorentz force exerted by
the electric field so that the particle will reach the plate with its x component
velocity (Vx) just reaching zero (Fig. 5.4). Particles with an angle greater than zero
will be collected (with Vx > 0), but only a fraction of the particle’s energy is
converted to electric field potential energy. The fraction of energy that is converted
to electrical field potential energy is sin2 hm . Emitted particles will have an energy
distribution function f(E). The distribution function defines the probability that an
emitted particle from a nuclear reaction will have an energy between 0 and
Q (where Q is the exothermic energy of the nuclear reaction). Ec is defined as the
minimum particle energy that can reach the collector. Since the particles are emitted
isotropically, only particles with angles greater than the minimum angle (θmin,
which has a relationship to Ec as defined in Eq. (5.33)) can be collected.
E sin2 hmin ¼ Ec ð5:33Þ
So,
rffiffiffiffiffi!
1 Ec
hmin ¼ sin ð5:34Þ
E
Assuming that the source is spatially uniform, three dimensional symmetry in

the problem allows for use of a two dimensional solution. It also allows for using
half of the angular emission space and multiplying the solution by 2. The proba-
bility for an isotropic source that a particle is emitted in the range of angles within
an angular space dθ (where the angular space resides between θ − ½ dθ and θ + ½

dθ) is therefore,
dh
PðhÞ ¼ ð5:35Þ
2p
The efficiency of a DCNB which has both an angular emission distribution and
an energy distribution can be constructed from the distribution functions f(E) and P
(θ). An integral function can be constructed which finds the probability that par-
ticles emitted at specific angles have sufficient energy to impinge on the collector,
Z p=2
dh
P1 ¼ 2 ð5:36Þ
hmin 2p
where the factor of 2 comes from the symmetry of the problem about the angle π/2
where particles are emitted in the +x direction at angles 0 ≤ θ ≤ π,
The probability that a particle is emitted at energies between Ec to Q (energies
sufficient for the particle to impinge on the collector) is,
Z Q
P2 ¼ f ðEÞdE ð5:37Þ
Ec
To factor in that a particle at the appropriate energy is emitted at the correct

angle, multiply the argument in Eq. (5.37) by P1,
Z Q Z Q Z p=2
1
P3 ¼ P1f ðE ÞdE ¼ dhf ðEÞdE ð5:38Þ
Ec p Ec hmin
Solving the angular dependent integral in Eq. (5.38),

Z rffiffiffiffiffi!!
1 Q
p Ec
P3 ¼ sin1 f ðEÞdE ð5:39Þ
p Ec 2 E
The amount of particle energy which is converted to electrical potential energy is

Ec (in a DCNB device in equilibrium). The number of particles emitted by the
source is N. Multiply P3 by NEc to find the total energy stored as electrical potential
energy,
Z rffiffiffiffiffi!!
NEc Q
p Ec
Estored ¼ sin1 f ðEÞdE ð5:40Þ
p Ec 2 E
The total energy of the particles emitted by the source is,

Z Q
ETotalEmitted ¼ N f ðE ÞEdE ð5:41Þ
0
The ideal efficiency of a DCNB which has particles that have both an energy
distribution and an angular distribution is found by dividing Eq. (5.40) by
Eq. (5.41),
R qffiffiffiffi
Ec Q
p Ec
p
2 sin1 Ec
E f ðEÞdE
gDCNB ¼ RQ ð5:42Þ
0 f ðE ÞEdE
Equation (5.42) does not incorporate spatial information and the loss of energy
that the particles experience as they transport through the source material is not
accounted for. Energy losses as particles transport through the source material occur
due to self absorption (Fig. 5.5). The energy transport losses in the source material
will reduce efficiency from the ideal. The spatial issue makes the problem far more
complex. Equation (5.42) does have some utility in generating calculations for the
spatial problem by using a transport code to solve for the angular and energy
distributions of particles leaving the surface of the source.
As an example for using Eq. (5.42), assume that the DCNB uses a thin source of
Ni-63. The energy distribution function for the source (See Appendix B) is,
f ðEÞ ¼ 0:145723 exp½43:2635E ð5:43Þ
The thin Ni-63 source distribution function in Eq. (5.43) can used to solve
Eq. (5.42). The Q value of Ni-63 is 0.0626 MeV. The equation can be solved using
Ec as the variable (recall that energy is equal to Ve, so the operating voltage of the
Fig. 5.5 An illustration is given for self-absorption losses of charged particles transporting
through the source material at a point deep (left) in the source material and shallow (right) in the
source material. The deeper particle on the left is bound by the range of the charged particle in the
material (Rs,Max) and thus produces a narrower escape cone then the shallow particle on the right.
The term Rs is the distance traveled in the source material and RT is the range in the transducer
Fig. 5.6 Efficiency of an ideal DCNB using a thin Ni-63 radioisotope source as a function of the
critical x component energy Ec. The operating voltage of the battery is Vc = Ec·e. In this case the
optimum operating voltage is 25 kV
cell is Vc = Ec/e). As shown in Fig. 5.6, an ideal DCNB powered with a Ni-63
source reaches a peak efficiency of 19.5 % at Ec = 25 keV. Without considering
space charge effects and other inefficiencies, the optimum operating voltage of the
DCNB is 25 kV (Fig. 5.6).
The ideal charge collection model described in Eq. (5.42) is still very powerful.
It works for a broad variety of charge collection methods including magnetohy-
drodynamic and thermionic energy conversion. However, charge collection based
systems will deviate from this ideal model when any process broadens the particle
energy distribution being emitted from the source. Effects which randomize the
energy flow will result in a significantly reduced efficiency. For example, the
highest efficiency occurs when mono-energetic particles in a directed beam hit the
collector. When the particle emission is isotropic, this detracts significantly from the
efficiency of the charge collection energy conversion mechanism. If the particles
deviate from mono-energetic to a broad distribution of energies, this too will sig-
nificantly reduce the efficiency of the charge collection energy conversion mech-
anism. The combined random effects of isotropic emission (or angular distribution)
and a broad energy distribution will devastate the ultimate efficiency of a charge
based energy conversion device. Unfortunately nature is not kind in this regard.
Radioisotopes produce charged particle sources that are isotropic. Radioisotope
beta sources will have an energy distribution as well. Even though alpha sources
can have narrow energy distributions, when the source is embedded in a material,
alpha particle transport will broaden the energy distribution of the particles.
Other effects can erode the efficiency of the ideal charge collection model. These
include electron backscattering, electron emission from the collector due to part of
the charged particle kinetic energy interacting with the collector material, space
charge build up, loss of electrons flowing from the collector to the load due to
defects in the collector material, and potentially other loss mechanisms. Each of
these could have a slight impact on efficiency. If there are a number of small loss
mechanisms, this could lead to a significant cumulative effect.
5.2 Radiation Damage
Radiation damage limits the efficiency of some nuclear battery designs and the
effects get worse over time (due to the integrated dose). It is caused by ionizing
radiation and is generally associated with solid transducers. Solids are composed of
tightly bound atoms which on a microscopic level are either a crystalline solid (has
a regular geometrical pattern in the lattice which includes metals, semiconductors,
ice, etc.) or an amorphous solid (non-crystalline solid which includes glass, plastic,
gel, amorphous forms of silicon, amorphous forms of carbon, etc.). A liquid is made
of atoms or molecules held together by intermolecular bonds. Liquids have a
volume but no fixed shape in that they conform to the shape of the container in
which they are held. Liquids can be classified as being incompressible. Gases are
made up of atoms or molecules which are not bound. Gases are compressible. There
are three categories of gas: (1) noble gases (gases made from atoms); (2) elemental
gases (gases made from one type of atom such as oxygen and nitrogen); or
(3) compound gases (gases made from a various types of atoms such as carbon
monoxide, carbon dioxide, methane, etc.).
There are four basic effects from ionizing radiation which can cause damage:
(1) transmutation of nuclei into other nuclei; (2) atom displacement which damages
the structure of a material; (3) ionization; (4) and localized heating which can
change material properties. Ions cause the most damage to the atomic structure
solids due to their short range in materials and their ability to displace atoms.
Transmutation is caused by neutron capture and the transmuted atom is a defect in
the material. Displacements create vacancies which are harmful to the properties of
a semiconductor. Displacements result from Non Ionizing Energy Loss (NIEL).
Ionization is harmful to electronic circuits because the carriers it produces can break
down insulating structures, but in a p-n junction based energy conversion trans-
ducer it is the means that generates current.
When ionizing radiation interacts with a gas, it creates excitation and ionization.
In a noble gas, the ionic and excited states decay back to the ground state of the
noble gas. In an elemental gas, the ionic and excited states decay back to the ground
state of the molecule (such as the ground state in O2 or N2). In a compound gas,
when the ions or excited states decay it is possible to form new molecules (such as
CO2 forming CO and ½O2). Thus radiation damage of a sort can occur in a
compound gas in that the molecular structure of some of the original molecules can
be changed after being exposed to radiation.
When ionizing radiation interacts with a liquid, it will typically create excitation
and ionization of the atoms or molecules making up the liquid. Radiation damage of
a sort can occur in a liquid if the molecular structures of some of the original
molecules in the liquid are changed after being exposed to radiation.
When a solid is exposed to ionizing radiation, there is a higher probability of
radiation damage than in a gas or a liquid because the structure of the tightly bound
atoms which make up the solid can be changed. Radiation can create electron-hole
pairs (as it does in a semiconductor). Electron-hole pair production is not consid-
ered damaged in a p-n junction based transducer because it is responsible for the
current flow in the device or the electron-hole pair can recombine. The original
structure of the solid is not affected by current flow or recombination. Energy
deposited in the transducer can lead to Non Ionizing Energy Loss which can
generate recoil atoms and form vacancies. Vacancies create microscopic damage in
the lattice of the semiconductor.
If the radiation is a neutron, then there are two primary reactions which create
damage: (1) neutron capture leading to transmutation of the atom; and (2) a neutron
colliding with the nucleus of an atom which creates a recoil ion (Primary Knock-on
Atom-PKA) which leads to the formation of a Frankel pair (meaning the loss of an
atom in the structure creates a vacancy in the structure and the free atom ends up in
an interstitial position). Both reactions lead to a change in the original structure of
the solid.
NIEL has been a useful tool because it has been shown in many types of devices
that the damage predicted by calculating PKA matches experiment [8].
The minimum energy for neutrons or protons (which have a mass of 1 atomic
mass unit or amu) to create a Frankel pair in crystalline silicon is 110 eV. The
threshold energy for an electron to create a Frankel pair in silicon is 260 keV. The
difference in the two cases is that the electron is much lighter than either a neutron
or proton (0.0005486 amu versus 1 amu). Thus it is much more probable in an
elastic collision for a heavy particle to create a PKA. Figure 5.7 shows the calcu-
lation of NIEL for different ionizing radiations (electrons, muons, pions, neutrons,
protons, deuterons and helium ions) where a limited set of radiation damage data is
available. These calculations can explain the radiation damage behavior in certain
devices.
The effect of ion energy and mass can be modeled using SRIM/TRIM [9]. In
Fig. 5.8, the displacements (25 displacements per ion) created in a silicon crystal by
1 MeV protons is shown. As the proton energy increases to 10 MeV the dis-
placements per ion increases to 91 and the range increases as shown in Fig. 5.9.
A higher mass ion of the same energy will create more displacements (991 per ion)
over a much shorter range as shown in Fig. 5.10 for a 1 MeV carbon ion interacting
with a silicon substrate. In Fig. 5.11, as the energy of the carbon ion increases to
10 MeV, the range increases as well as the displacements per ion (1316).
Neutrons create ions by a recoil reaction. The recoil energy of a PKA can be
very high depending on the energy of the incident neutron. The recoil ion can knock
out other atoms from the lattice giving rise to a PKA cascade. As discussed in
Chap. 3, about half of the damage occurs when the recoil ion losses its final 10 keV
of energy. In a PKA cascade, it is possible to form clusters of Frankel pairs. A high
percentage of Frankel pairs recombine (about 90 %) with interstitials resulting in no
5.2 Radiation Damage 237
NIEL for Various Particles vs Energy
1.00E+00 Electrons
Muons
Neutrons
1.00E-01
Pions
Protons
Deuterons
NIEL (MeV cm2/g)
1.00E-02
Helium Ions
1.00E-03
1.00E-04
1.00E-05
1.00E-06
1.00E-01 1.00E+00 1.00E+01 1.00E+02 1.00E+03 1.00E+04
Energy (MeV)
Fig. 5.7 Theoretical values of NIEL in silicon for various ionizing radiation particles as a
function of particle energy. Experimental data show an apparent relationship between NIEL and
displacement damage [8]
Fig. 5.8 Total displacements in a silicon lattice created by 1 MeV protons [9]
Fig. 5.9 Displacements in a silicon lattice created by 10 MeV protons [9]
Fig. 5.10 Shown are displacements in a silicon lattice created by 1 MeV carbon ions [9]
5.2 Radiation Damage 239
Fig. 5.11 Shown are displacements in a silicon lattice created by 10 MeV carbon ions [9]
net damage to the crystal structure. Some of the clusters of vacancies can form
multi-vacancy defects and the interstitial atoms diffuse through the crystal and can
interact with other defects or impurity atoms to form stable complexes.
Electrons have a much different and not well understood mechanism for creating
Frankel pairs because of their small mass as compared to the mass of the nucleus of
the atoms with which they interact. A direct knock-out collision is not very probable
even though NIEL occurs. However an alternate view will be discussed which may
help explain the long term slow radiation damage rate in p-n junction based trans-
ducers exposed to low energy beta particles. As discussed in Chap. 2, electrons are
efficient in transporting energy to the bound electrons in atoms that make up the target
to create ions. Thus the electrons which form covalent bonds in a solid can be kicked
out. The probability that a covalent bond is broken on a specific atom at a given
instant in time by a beta is P (which is the power density that the betas deposit in the
material and the atomic density of the atoms in the crystal lattice). If the atom has “n”
covalent bonds, the probability that all of the covalent bonds are broken simulta-
neously is Pn. Once the covalent bonds which hold an atom in place are broken, the
atom can drift away and create a Frankel pair. The probability function does have a
dependence on the maximum beta energy (because it directly factors into the power
density) and is important to the concept of the damage energy threshold (Edth).
The radiation damage threshold energy (Edth) for the various semiconductors
used in betavoltaic energy conversion have been reported for Si (about 200 keV),
GaAs (about 225 keV) and Ge (about 350 keV) [10]. Many beta sources have
maximum energies which exceed Edth. Thus, researchers have focused on ra-
dioisotopes with lower beta energies such as tritium (maximum energy of 18 keV),
Ni-63 (maximum energy of 67 keV) and Pm-147 (maximum energy of 230 keV).
Even though low beta energies minimize radiation damage, these particles still have
enough energy to break bonds. Betavoltaic cells using low energy beta particles still
degrade over time. The reason being is that ionizing radiation above 1 keV has
enough energy to break covalent bonds in a solid. The probability function, P,
predicts that there is a probability that all of the bonds binding an atom to a lattice
can be broken simultaneously (Pn) which can create a Frankel pair. As power
density increases, Pn increases. As Pn increases damage in the crystal lattice
increases. Power density can be increased by using radioisotopes that emit higher
energy beta particles and/or by using radioisotopes with shorter half-lives.
The production of Frankel pairs is harmful for p-n junctions. Vacancies and
multi-vacancy defects can lead to traps which limit the lifetime of carriers. In binary
semiconductors, vacancies could be filled with the wrong atom thus causing a
structural defect in the crystal. Another consequence is that the p- and n- type
impurities in the depletion region can be reordered in a way that damages the
junction.
In using low energy beta sources, as discussed in Chap. 1, the cell will have a
low power density. In considering the dilution effect discussed in Chap. 4, the
power density of the cell will be further eroded. Low energy beta particles will
require that the source be thin and this leads to an unfavorable dilution factor.
5.3 Health and Safety
Radioisotopes are hazardous. As such, they are regulated by the Nuclear Regulatory
Commission. In this section, specific guidelines regarding the regulation of isotopes
of interest for nuclear batteries are discussed. In addition, a case study on the effects
of the pure alpha emitter Po-210 is included. Finally, safety studies by NASA on
the flight certified RTG which uses Pu-238 is considered.
5.3.1 Nuclear Regulatory Commission Rules

and Regulations
The primary markets for radioisotope production and transportation are medical,
power production, and waste handling. Much of the current regulation of nuclear
Isotopes are centered around radiation exposure to different personnel. The role of
5.3 Health and Safety 241
Table 5.2 Weighting factors Tissue (T) wT

relative to specific tissues
Gonads 0.25
Breast 0.15
Redbone Marrow 0.12
Lung 0.12
Thyroid 0.03
Bone Surface 0.03
Remainder/Skin 0.3
Wholebody 1.00
the individual and their associated clearance determines radiation limitations of

exposure. Additionally, dose is highly based on the type, energy and location of the
body being exposed. Magnification of these effects, in limit documentation is
referred to as the ‘Q’-value.
The primary regulatory agency is the U.S. Nuclear Regulatory Commission and
they have Regulations and Standard labeled 10 CFR. The Standards for Protection
Against Radiation is Part 20 of this 10 CFR, or 10 CFR 20. This regulation part
encompasses possession, transport, disposal, special conditions, operation, pro-
duction, and medical exposure of radioactive materials.
Given this formulation the effective dose, E is defined as the summation over
each tissue or organ equivalent measured in Grays or Rads.
E ¼ RT wT HT ð5:44Þ
Here, HT is the equivalent dose in tissue or organ, T, and wT is the tissue

weighting factor. Table 5.2 are the weighting factors relative to the whole body.
To find equivalent dose one must apply the associated weighting scaling factor
wR for each type of radiation, R. The resulting effective dose follows:
Ef ¼ RT wT RR wR DT;R ð5:45Þ
Here DT,R represents the average absorbed does in the tissue or organ by radi-
ation type R. The SI unit for effective does is Sievert (Sv) or joule per kilogram
(J/kg) and the American unit is Rem (Table 5.3).
For neutrons of known energy, the wR or Q-factor is in Table 5.4.
Note that 1 Sv is equivalent to 100 Rem, and limits are imposed on an annual
basis. The annual inhalation or digestion intake limit is 50 Rem or 0.5 Sv
Table 5.3 wR for various Radiation type (R) wR

radiation types
Xray, Gamma, Beta 1
Heavy Particle 20
Neutron Unknown 10
High Energy Proton 10
Table 5.4 Q-factor for various neutron energies

Neutron energy (MeV) Q-factor Fleunce per dose Equiv. (n cm−2 rem−1) 108
−8
2.50 × 10 2 9.8
1.00 × 10−7 2 9.8
1.00 × 10−6 2 8.1
1.00 × 10−5 2 8.1
1.00 × 10−4 2 8.4
1.00 × 10−3 2 9.8
1.00 × 10−2 2.5 10.1
1.00 × 10−1 7.5 1.7
5.00 × 10−1 11 0.39
1.00 11 0.27
2.50 9 0.29
5.00 8 0.23
7.00 7 0.24
1.00 × 101 6.5 0.24
1.40 × 101 7.5 0.17
2.00 × 101 8 0.16
4.00 × 101 7 0.14
6.00 × 101 5.5 0.16
1.00 × 102 4 0.2
2.00 × 102 3.5 0.19
3.00 × 102 3.5 0.16
4.00 × 102 3.5 0.14
equivalent to 1 ALI or annual limit intake. In the presence of radioactivity 2000 h is

considered for ALI exposure. Higher ALI will reduce the resultant limit of intake.
(Ex. 4000 h of work can be done if 25 Rem of exposure). Public exposure is 1/10th
the limits imposed upon workers. The maximums for workers based on location of
exposure are: body core, head, eye lens at 1.25 Rem each; extremities (hands/feet)
is 18.75 Rem total and Skin 7.5 Rem total.
Complexity arises from identifying transport restrictions, labelling, and limita-
tions of handling regarding radiations since concentrations are in terms of activity,
or decays per second. Activity can translate to mass via its definition of mass times
decay constant, λ or ln(2)/t1/2 (half life). Most limits are listed in Curies, 3.7 × 1010
decays a second or micro (10−6)-Curies (µCi). The limitation for liquid and air
concentrations associated with transportation, from 1981 Report (Table 5.5).
Additionally, there are regulations on special licensing with regards to handling
and production (Table 5.6).
Obviously the handling of nuclear fuel and weapons grade materials will require
extensive paper work, onsite regulation during transportation and handling. The
broad spectrum of radiation limits associated with every radio nuclide can be found
Table 5.5 Permissible concentrations of various isotopes in air and water

Permissible Conc. µCi/ml
Isotope Abbreviation Air Water
Cesium-137 Cs-137 2.00 × 10−09 2.00 × 10−05
Cerium-144 Ce-144 3.00 × × 10−10 1.00 × 10−05
Hydrogen-3 H-3 2.00 × 10−07 3.00 × 10−03
Iodine-129 I-129 2.00 × 10−11 6.00 × 10−08
Iodine-131 I-131 1.00 × 10−10 3.00 × 10−07
Krypton-85 Kr-85 3.00 × 10−07
Neptunium-237 Np-237 1.00 × 10−13 3.00 × 10−06
Plutonium-238 Pu-238 3.00 × 10−08 1.00 × 10−04
Plutonium-239 Pu-239 6.00 × 10−14 5.00 × 10−06
Radium-226 Ra-226 3.00 × 10−12 3.00 × 10−08
Radon-222 Rn-222 3.00 × 10−09
Strontium-90 Sr-90 3.00 × 10−11 3.00 × 10−07
Uranium-235 U-235 2.00 × 10−11 3.00 × 10−05
Uranium-238 U-238 3.00 × 10−12 4.00 × 10−05
Table 5.6 Quantities of Special licensing

various isotopes and the limits
Radio-nuclide Quantity in Ci
that require special licensing
Cs-137 1.0
Co-60 1.0
Au-198 100
I-131 1.0
Ir-192 10
Kr-85 1000
Pr-147 10
Tc-99 m 1000
in the Nuclear Regulatory Commission guidelines. There are a select few that have
medical and potential battery uses that are tracked, in Ci and tera-becquerel, TBq
(1012 decays per second). (Table 5.7)
5.3.2 Po-210 Poisoning
Radioisotopes in general can be dangerous to human health because of the radiation

emitted. When a person is contaminated, the amount of damage and the method to
detect the contamination depend on the type of decay of the radioisotope. For
example, alpha emitters are dangerous if inhaled, ingested, or injected by pene-
trating wounds, but they are not for external contamination because they are
Table 5.7 Radioisotopes and the limits for various catagories

Radioactive material Category 1 Category 1 Category 2 Category 2
(TBq) (Ci) (TBq) (Ci)
Actinium-227 20 540 0.2 5.4
Americium-241 60 1,600 0.6 16
Americium-241/Be 60 1,600 0.6 16
Californium-252 20 540 0.2 5.4
Cobalt-60 30 810 0.3 8.1
Curium-244 50 1,400 0.5 14
Cesium-137 100 2,700 1 27
Gadolinium-153 1,000 27,000 10 270
Iridium-192 80 2,200 0.8 22
Plutonium-238 60 1,600 0.6 16
Plutonium-239/Be 60 1,600 0.6 16
Polonium-210 60 1,600 0.6 16
Promethium-147 40,000 1,100,000 400 11,000
Radium-226 40 1,100 0.4 11
Selenium-75 200 5,400 2 54
Strontium-90 1,000 27,000 10 270
Thorium-228 20 540 0.2 5.4
Thorium-229 20 540 0.2 5.4
Thulium-170 20,000 540,000 200 5,400
Ytterbium-169 300 8,100 3 81
stopped by clothing or intact skin. If there is no external contamination, analysis of

biological samples and blood test are necessary because radiation detectors are not
able to discover alpha particles inside the body. Beta and gamma emitters represent
both an internal and external hazard, betas can penetrate a few millimeters into
tissue and cause severe burns and gammas pass through matter (skin and clothing)
resulting in a greater exposure [11].
One recent known case of radiation poisoning is associated with the death of
Alexander Litvinenko. He was a former Moscow anticorruption detective who had
immigrated to London in 2001 where he worked as a writer. On November 1, 2006
he presented to a hospital with severe gastrointestinal symptoms. As his symptoms
matched radiation poisoning, the physicians analyzed urine and blood samples with
gamma spectrometry but they didn’t found anything unusual. While doctors sear-
ched for other causes for the disease different from radiation poisoning (like toxins),
his health deteriorated rapidly suffering from hair loss, shortage of all types of red
and white blood cells, and platelets. Then, his urine was sent to Britain’s Atomic
Weapons Establishment for further analysis with specialized equipment, the result
showed significant amount of alpha particle radiation. He died on November 23.
This kind of toxicant inside a patient is difficult to detect when there is no external
contamination, advanced laboratory testing is needed to analyze biological samples.
As a consequence, several weeks elapsed until the realization that Po-210 had been
used as poison. [11, 12].
Po-210 is an alpha particle (5.3 MeV) emitter with a half-life of 138 days, it is
considered one of the most dangerous radioisotopes (more toxic than hydrogen
cyanide). In fact, less than a microgram of Polonium could have caused
Litvinenko’s Death. This high toxicity is due to its relatively easy distribution and
retention in organs and tissues which is facilitated by its chemical properties such as
solubility in water and the ability to form simple salts (e.g. chloride). It has a
biological half-life of 1–2 months being eliminated mainly by fecal route. When
orally administered, it is deposited in soft tissues with greater concentrations in
liver, spleen, and bone marrow. Also, it is more readily absorbed in blood than
other alpha emitting radioisotopes. Once ingested and distributed, it damages the
gut mucosa causing severe vomit and high fever. Moreover, it kills the cells lining
the gastrointestinal tract causing nausea, pain and gastrointestinal bleeding.
A possible treatment for this condition is aggressive fluid replacement, prevention
of infection, and decorporative countermeasures like chelation [11, 13].
5.3.3 NASA RTG Safety
The Radioisotope Thermal Generators (RTG) has been used as a power source for
space missions beginning with the launch of the 1961 Navy Transit 4A spacecraft.
The RTG has been flown on many space missions by the DOD and NASA. The
safety of an RTG which contains Pu-238 was a concern and a launch approval
process that was ordered by Presidential Directive/NSC-25 [14]. This directive
requires the close collaboration of different agencies that share a safety responsi-
bility for the mission. The first component is the development of a Safety Analysis
Report (SAR) which is the responsibility of the Department of Energy (DOE). With
regard to the RTG, the content of the SAR addresses the risk due to the release of
plutonium dioxide in case of an accident. The contents of the SAR include a safety
analysis of systems, such as the power source, the spacecraft, launch vehicles and
missions design. The National Aeronautics and Space Administration (NASA)
prepares a launch-vehicle data book, which is used by the DOE to develop the
Safety Analysis Report (SAR). The book outlines specific accident scenarios which
may occur during the pre-launch, launch and earth-orbit phases of the mission. An
independent Interagency Safety Review Panel (INSRP) is assembled consisting of
representatives from NASA, DOE, the Department of Defense (DOD), the
Environmental Protection Agency (EPA), and the Nuclear Regulatory Commission
(NRC). This group performs a safety/risk evaluation of the radiological risk and
documents their results in a Safety Evaluation Report (SER). The SER is evaluated
by the DOE and is used for additional analysis of the SAR. Based upon the SAR
and SER, further recommendations from the INSRP and other government agen-
cies, NASA submits the SER-SAR to the White House Office of Science and
Technology (OSTP) as part of the safety launch approval process.
The safety analysis process has evolved over time since the launch of SNAP-3B
in 1961 [15]. Based on the knowledge gained from the cumulative mission history,
the safety analysis of new missions has improved. The safety analysis process used
for the Cassini mission was based in part on the prior Galileo and Ulysses missions.
The Cassini mission was an international cooperation between NASA, the
European Space Agency, and the Italian Space Agency. The objective was to
conduct scientific exploration of Saturn and its moons. The mission was launched
on October 15, 1997 using the Titan IV/Centaur launch vehicle. The Titan launch
vehicle consisted of a core vehicle with two liquid propellants, an upper stage, a
payload fairing, and two Solid Rocket Motor Upgrades (SRMU). The spacecraft
was placed into a 6.7 year Venus-Venus-Earth-Jupiter-Gravity-Assist trajectory to
Saturn. The Cassini space vehicle consisted of a lower equipment module, an upper
equipment module, a propulsion module, and a high gain antenna. The two
equipment modules housed the external mounting of the magnetometer boom, and
the three General Purpose Heat Source (GPHS) RGTs that supplied power to the
space vehicle. The high gain and low gain antennas were used for data transmission
and reception. The space vehicle contained about 687 kg of scientific instrument
including the Huygens Probe System [16] (Fig. 5.12).
The Radioisotope Thermal Generators used in the Cassini mission consisted of
three basic parts: (1) a heat source, (2) a converter shell, and (3) a thermopile. The
heat source produced a nominal power of 4400 W thermal by radioactive decay of
plutonium fuel. The radioisotope was made of a mixture of plutonium dioxide with
82.2 weight percent of 238Pu. It was formed into a cylindrical, solid ceramic pellet
with diameter of 2.76 ± 0.02 cm. The thermopile converts some of the thermal
energy of the heat source into *300 W of direct current electric power through the
thermoelectric process. The outer shell provides containment and structural support
for the heat source and thermopile, and it rejects the surplus energy by using its
radiating fins.
The designed features of the GPHS-RTG include safety related components to
ensure the containment of fuel in all aspects of the mission and its preparation. The
GPHS-RTG has a total of eighty modules as shown in the Fig. 5.13. Each module
has five different elements: the fuel, the cladding, the graphite impact shell, the
carbon insulation, and the aeroshell. The modules have four plutonium dioxide
(PuO2) fuel pellets each with a thermal power output of about 62.5 W. The PuO2
fuel is robust, chemically stable, insoluble, and has a high melting point. The pellet
is encapsulated with iridium cladding and this is called a fueled clad [17].
Each GPHS module contains four fueled clads encapsulated within two cylindrical
Fine Weave Pierced Fabric (FWPF) containers known as Graphite Impact Shells
(GISs). Carbon materials provide thermal insulation to each GIS during operation
of the GPHS. The GISs protect the fuel clad from debris in case of an accident. The
aeroshell is the primary heat shield and serves as protection during reentry. The
RTGs was tested for performance as was the GPHS modules and the fuel clad [18,
19] (Fig. 5.14).
The safety analysis consisted of four steps: (1) accident definitions, (2) source
term determination, (3) consequence evaluation, and (4) risk analysis.
Fig. 5.12 Cassini space vehicle
Fig. 5.13 General purpose heat radioisotope thermal generator (GPHS-RTG)

Fig. 5.14 View of a general purpose heat source (GPHS) module [19]
The accident scenarios for the Cassini mission was prepared using the Titan IV
RTG safety databook (provided by Martin Marietta under contract with NASA),
and the environmental impact support system study. The fundamental scenarios for
an accident that may lead to plutonium release and the probability of occurrence
(failure or POF) during launch were developed. Meanwhile, the swingby defines
those accidents which occur after escape from earth orbit. The source term is
determined based on the response of the RTG to different accident scenarios con-
sidered at different portions of the launch. RTG response to postulated accidents
includes: threats to local environment such as blast, overpressure, and fragments
impacts. The response is determined using mathematical models. The models are
based on sets of test data that have been acquired in support of GPHS-RTG safety
analysis for earlier missions. Launch Accident Scenario Evaluation Program for the
Titan IV (LASEP-T) code was used for this evaluation. The LASEP-T is based on
Monte Carlo simulation because of the complexity and probabilistic nature of the
dependence of the RTG on dynamic variables and vehicle configuration. The
LASEP-T code is capable of evaluating the RTG response from initial event
(generally at altitude) to the final even (Earth impact). Based on results from this
model, source terms are defined which provide a statistical description of potential
release of PuO2, and the characteristics of such a release (mass, location, particle
size distribution, and particle density, etc.). The FALLBACK computer code was
used to calculate the probability of direct or near miss impact from solid propellant
fragments with an RTG. A modified LASEP-T was used to determine the potential
releases for this case. The SRMU Propellant Fallback Response Model (SRESP)
code simulated RTG response from mechanical impact and vaporization effects that
occur from solid propellant fires. A three-degree-of-freedom trajectory simulation
(3DMP) was used to model out-of-orbit accidents. The 3DMP code predicted the
trajectory of the RGT modules and their associated aerothermal environments in

case of accidents resulting from events such as vehicle failure, aerodynamics
heating, and loads which cause structural damage to the vehicle. The graphitic
aeroshell was modeled with the Reaction Kinematics and Ablation Program
(REKAP). The effect of an impact to fuel release was modeled using the LASEP-T
code. A NASA-JPL code calculated the total probability of short term impact
reentry during Earth Swingby while the probability of the flight dynamics of the
modules were predicted using a model developed by the Johns Hopkins University
Applied Physics Laboratory (JHU/APL). The JHU/APL model provided a 6-degree
of freedom (3 rotatory and 3 vibrational) analysis of the trajectory. Consequences
evaluation was based on the results obtained from the source term. The Space
Accident Radiological Release and Consequence (SPARRC) family of codes were
used to perform such evaluation. SPARRC consist of four different codes:
(1) Site-Specific Analysis of Transport and Dispersion of Radioactive Particles
(SATRAP) code which is use to model site specific transport and dispersal of
radioactive particles, (2) Global Transport and Dispersion of Radioactive Particles
(GEOTRAP) for modeling global transport and dispersal of radioactive particles,
(3) High Attitude Aerosol Dispersion (HIAD) Code for modeling high altitude
aerosol transport and dispersal, and (4) Particulate Dose (PARDOS), used for
modeling dose and health effects. The resulting outcome was a cumulative com-
plementary distribution function (CCDF) which provides a relative likelihood of
various consequences. These values and probability calculations were used to
quantify the risk [20].
Table 5.8 shows the mean 50 year value of collective dose, health effects, and
dry land contaminated above 0.2 μCi/m2 by accident cases which contribute sig-
nificantly to the risk for each mission segment. The contribution to total mission
risk for the pre-launch, late launch, and EGA Swingby mission segments are 2, 27,
55, and 16 %, respectively.
Accidents near the launch site (Phase 0 and 1) produced the lowest expected
health consequences from the launch. The combined risk from phase 0 and 1
contributes to about 29 % of the total mission risk. The greatest risk will result from
inhalation of re-suspended particles. The probability of inducing health effects from
exposure due to release is about 1 in 4 million and 1 in 10,000 in phase 0 and 1,
respectively. Accident 1.1 is the dominant risk contributor in this phase of the
mission. Ground contaminations greater than 0.2 μCi/m2 are expected to be from
large particles and will be restricted around the launch site. Phases 0–3 represent
risk from reentry accidents that results from late launch. The dominant contributor
to this mission segment results from direct inhalation during cloud passage.
The mean health consequences from reentry accidents is estimated to be about
0.0544. The probability of more than one health effects due to reentry accidents is
about 1 in 100,000. Accident scenario 5.2 is the dominant contributor to risk in this
phase. The highest mean health consequence is from the Earth Swingby scenarios.
Air releases at high attitude and soil impact releases from melted clad are the
dominant contributors in EGA accidents, contributing 32 and 54 %, respectively.
Table 5.8 Mean 50 year radiological consequence values for accident scenarios from the CASINO launch
250
Accident Accident Collective dose Collective dose Health Health effectsa Maximum Land
case description (person-rem)(without (person-rem)(with effectsa (withoutde-minimisb) individual area
de-minimis) de-minimisb) (without dose(rem) 0.2
de-minimis) μCi/m2
(km2)
0.0 On-pad explosion, 130 96 0.066 0.053 0.013 1.5
configuration 1
1.1 Total boost 160 110 0.081 0.05 0.015 1.8
vehicle destruct
(TBVD)
1.3 TBVD with 130 88 0.067 0.045 0.046 1.3
SRMU Aft
segment impact
1.10 SV/RTG impact 80 51 0.04 0.021 0.020 1.1
within PLF
1.13 Full stack intact 240 180 0.12 0.098 0.028 2.2
impact
3.1 Sub-orbital 8.4 7.5 4:2 103 3:8 103 0.37 0.028
reentry
5.1 CSDS (config. 5) 8.4 7.5 4:2 103 3:8 103 0.37 0.027
sub-orbital reentry
5.2 Orbital reentry 92 82 0.046 0.041 1.1 0.058
(nominal)
EGA EGA reentryc 1:7 105 1:1 105 90 60 1780 21
(short term)
a
Health Effects are incremental latent cancer fatalities
b
The de-minimis dose level is 4:2 103 sv (4:2 103 rem) per year
c
For 800 km swingby attitude
5 Efficiency Limitations for Various Nuclear Battery Configurations
The probability of more than one health effect due to reentry accidents is about 3 in
10 million. A total of 90 incremental cancers are calucalated for EGA reentry
accidents [20].
For all combined postulated accidents scenarios, the results show that the
radiological risk is very small. Using the estimated background radiation of
0.3 rem/year and the health effects estimate of 5.0 × 10−4 latent cancer
fatalities/rem, the risk from an individual developing cancer in a 50 year period is 1
in 133. This value is about 8 orders of magnitude greater than Cassini mission
individual risk which is about 1 in 50 billion [20].
The transportation system of the RTG (package and payload) was designed by
assessing the damage that can be done on an RTG when it is dropped from a height
of 30 feet while being transported on the containment vessel and impact limiter.
Packaging and transport of the radioisotope power sources is subjected to Title 10,
Code of Federal Regulations, “Part 71” (10 CFR71) [21].
Figure 5.15 shows the worst case scenario that could cause the most damage to
the generator. The event sequence proceeds from 5.15a and ends with 5.15d with
the heat source contained in the containment vessel but released from the generator
case.
Fig. 5.15 The Worst case transportation accident scenarios

The assessment concluded that in some instances, the RTG case could rupture,
thus potentially releasing heat source modules. However, no further breakage
would occur and the fuel itself would not be release to the interior of the inner
containment vessel.
5.4 System Efficiencies and Power Density
The system efficiency is dependent on the power deposition efficiency described in

Fig. 5.1, gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 Þ: The power deposition efficiency is an
energy transport problem in which a source volume made up of a large number of
isotropic point source emitters (radioisotope atoms) generates charged particles and
the energy deposition of the charged particles is calculated for each region of the
nuclear battery. The amount of power which is deposited in region 4, the trans-
ducer, is the critical value which determines the power deposition efficiency. The
second part of the problem is to determine the efficiency of the transducer (which
simply stated is the amount of useful output power from the transducer divided by
the power deposited in the transducer). The power density (Pd) of the nuclear
battery is the useful output power (Pout) from the transducer divided by the total
volume of the battery (where V1 is volume of region 1, V2 is volume of region 2, V3
is volume of region 3, V4 is volume of region 4 and Vtotal = V1 + V2 + V3 + V4),
Pd ¼ Pout =Vtotal ð5:46Þ
The power density (Pd) also has a relationship with the atomic dilution factor
(DFatomic). This relationship requires that effective power density, as described in
Chap. 4, (Pdeffective) is known or can be calculated. The effective power density is
calculated in a convoluted way using the theoretical absolute minimum nuclear
battery volume per Watt (BVWmin) as shown in Table 1.3. The reason for using
BVWmin is to continue with the discussion of Chap. 1 that the absolute minimum
volume of a nuclear battery is achieved by using the radioisotope containing
compound which has the highest possible atomic density of the radioisotope.
A device based on a compound with the highest possible atomic density of a
radioisotope would use a heat based transducer such as a thermoelectric. Other
nuclear battery concepts require that the atomic density of the radioisotope in the
volume of the nuclear battery be diluted. It is this dilution factor of the atomic
density which results in a lower power density from the alternative nuclear battery
concept as compared to a heat based nuclear battery. So this convoluted method of
calculation has a reason. It provides a means to directly compare the parameters of a
conceptual design with a heat based nuclear battery design.
The first step in this methodology is to find the power density of the radioisotope
source used in the conceptual nuclear battery design. This calculation uses the
relationship that the power density of the source is the atomic dilution factor
divided by BVWmin (Pdeffective = DFatomc/BVWmin). From Chap. 1, 1/BVWmin
5.4 System Efficiencies and Power Density 253
was defined as the power density from the compound which has the highest known
atomic density of the radioisotope. The atomic dilution factor, as discussed in
Chap. 4, is the ratio of the average atomic density of the radioisotope in the volume
of the conceptual nuclear battery divided by the atomic density in the compound
with the highest known atomic density of the radioisotope. In the next few sections,
some common nuclear battery/transducer combinations will be analyzed.
5.4.1 Alphavoltaics Analysis
The optimal geometrical configuration for a mono-energetic isotropic alpha source

is when its alpha source was placed at the center of a spherical PV cell (Fig. 5.16).
The optimum energy production will occur when the electron-hole pair formation is
maximized in the depletion region. Therefore, the maximum energy output will
occur when the specific ionization is a maximum within the 1 µm depletion region
of a linearly graded p-n junction transducer. GEANT4 was used for creating the
experimental model as well as for finding the maximum production of electron-hole
pairs in the depletion region.
The point source is equidistant from any spherical shell within the volume and
thus is able to model an idealized p-n junction based system. A p-n junction with a
planar geometry is more realistic. A model which takes out the angular dependence
of charged particle emission is to look at a directed beam of alpha particles normal
to the surface (Fig. 5.17).
Silicon Carbide (SiC) was selected because it is a wide band-gap (Eg ≈ 3.0 eV)
material which has some radiation resistance, and it also has a very low leakage
current. Thus, it has been considered a viable base material for nuclear battery
research [3]. There are two predominant crystal structures, 4H-SiC and 6H-SiC. In
this model, the 6H-SiC structure was used.
The alpha source chosen for this study was the alpha emitter, Polonium-210
(Po-210). It has a half-life of 138.38 days and is subject to less regulatory
Fig. 5.16 Optimized

geometrical cell using a
spherical geometry and a with
a point alpha source at the
center
Fig. 5.17 Slab model used

for benchmarking GEANT4
with SRIM/TRIM.
A mono-directional,
mono-energetic alpha beam
impinges on the slab target
constraints than Am-241 or Pu-238. It is also more widely available than Gd-148
[3]. Po-210 decays to Pb-206 through the emission of an alpha particle with an
energy of 5.307 MeV by the following reaction,
210
82 Po ! 206
82 Pb þ að5:307 MeVÞ ð5:47Þ
As alpha particles interact with the medium, low energy (*1 keV) secondary
electrons are created [18]. The secondary electrons then interact with the medium to
create tertiary and higher order electrons, and the electron yield is proportional to
the specific ionization of the material [19, 20]. The energy flow of secondary
products from alpha particle interactions follows the principle of detailed energy
balance. If the energy lost due to secondary ionizations from one cell into an
adjacent cell equals the energy lost from the adjacent cell to the original cell, then
the energy deposition in the cell will be equal to the energy deposited by the alpha
particles. Since the depletion region is at the Bragg peak, the assumption of the
principle of detailed energy balance overestimates the electron-hole production,
thus providing a theoretical maximum limit for electron-hole pair production.
The SRIM/TRIM results from the SiC slab model shows that the peak energy
deposition (Bragg curve peak) occurs at around 16 µm (Fig. 5.18). Table 5.9 shows
the deposited energies in each of the defined cells in models 1 and 2. The energy
deposition rate follows the typical Bragg curve which peaks near 16 µm. Beyond
the Bragg peak, the energy deposition drops sharply. In addition, the table shows
that the maximum energy deposition in a 1 µm layer occurs between 15 to 16 µm
from the source. The percentage of energy deposited is consistent for both the slab
and the spherical models, as expected. GEANT4 predicts that the energy deposited
in the 15–16 µm layer is *9.93 % for the spherical configuration and *9.94 %
for the slab configuration (Table 5.9). SRIM/TRIM’s value for the slab model
is *9.57 % which is consistent within a reasonable margin of error with the
GEANT4 simulations. Assuming that the depletion region is 1 µm wide, this data
Fig. 5.18 Alpha particle

energy deposition vs. distance
in the slab model using
SRIM/TRIM and in the slab
and sphere models using
GEANT4
Table 5.9 GEANT4 and SRIM/TRIM calculations predicting energy deposition in a depletion
region of 1 µm thick for the slab and sphere models
Range GEANT4 SRIM/TRIM
(µm) Sphere (Fig. 5.xx1) Slab (Fig. 5.xx2) Slab (Fig. 5.xx2)
Energy % Energy % Energy %
(keV) deposited (keV) deposited (keV) deposited
0–1 208 3.92 208 3.92 211 3.98
1–2 214 4.03 214 4.03 218 4.11
2–3 220 4.15 220 4.15 223 4.19
3–4 228 4.29 228 4.29 226 4.26
4–5 236 4.44 236 4.45 235 4.42
5–6 245 4.61 245 4.61 243 4.59
6–7 254 4.79 254 4.79 258 4.85
7–8 265 5.00 266 5.00 269 5.06
8–9 279 5.25 279 5.25 279 5.25
9–10 294 5.55 294 5.55 296 5.58
10–11 312 5.89 313 5.89 315 5.94
11–12 335 6.32 335 6.32 337 6.35
12–13 364 6.86 364 6.86 365 6.88
13–14 402 7.58 402 7.58 405 7.62
14–15 456 8.60 457 8.61 454 8.56
15–16 527 9.93 527 9.94 508 9.57
16–17 408 7.68 407 7.66 382 7.19
17–18 59 1.12 58 1.09 0 0.00
18–19 0 0.00 0 0.00 0 0.00
19–20 0 0.00 0 0.00 0 0.00
Total 5307 100.00 5307 100.00 5220 98.40
also represents the maximum energy transport efficiency (ηd), which will be used
later to define the absolute energy conversion efficiency.
The energy deposition from radiation is a very important parameter for nuclear
batteries [21]. However, there are more factors beyond the total energy deposition
to be considered when determining the efficiency of the nuclear battery. To develop
the expression for the absolute efficiency, the total power of the source needs to be
taken into account along with the energy conversion efficiency of the transducer.
The total power of a radioactive source is described by,
Ptot ¼ A Ea ð5:48Þ
where A is the decay rate of the alpha source and Eα is the energy of the emitted
alpha particle in the decay process (see Eq. (5.48)). The total number of
electron-hole pairs created per second in the depletion region is,
Pin Ptot A Ea
Jsc ¼ ¼ gd ¼ gd ð5:49Þ
W W W
where Pin is the power deposited into the depletion region, ηd is the maximum
energy transport efficiency from the source to the depletion region, and W is the W
value, or the average energy required to make an electron-hole pair in SiC. Note
that the total number of electron-hole pairs created per second in the depletion
region is also equal to the short circuit current of an alphavoltaic cell, analogous to
that used in the definition of the fill factor in photovoltaic cells.
The other factor in determining the fill factor is the open circuit voltage, defined
below, along with the fill factor formula. As shown in Eq. (5.50), the open circuit
voltage can be related to the band-gap of the material.
eVoc ¼ gdp Eg ð5:50Þ
Pmax
FF ¼ ð5:51Þ
Voc Jsc
In Eq. (5.50) Voc is the open circuit voltage, Egis the band gap energy of SiC, e is
the unit electric charge, and ηdp is the factor relating open circuit voltage to
band-gap which is called the driving potential efficiency. In Eq. (5.51), FF is the fill
factor and Pmax is the maximum power attainable from the alphavoltaic cell.
Another important factor that needs to be taken into account is the pair production
efficiency, ηpp, which describes the fraction of energy deposited by alpha particles
that goes into the production of electron-hole pairs.
Eg
gpp ¼ ð5:52Þ
W
The total efficiency of any system is defined as the ratio of the total output power
to the total input power. Using the definition of the fill factor the total efficiency of
an alphavoltaic cell is defined in Eq. 5.53.
Pmax Voc Jsc

gtot ¼ ¼ FF ð5:53Þ
Ptot Ptot
By utilizing Eqs. (5.48) and (5.49) along with applying the definitions of the
driving potential and pair production efficiencies defined in Eqs. (5.50) and (5.52)
the total efficiency of an alphavoltaic cell can be reduced to the form shown in
Eq. (5.54).
gtot ¼ FF gd gdp gpp ð5:54Þ
The theoretical maximum energy efficiency can now be determined by the

energy transport efficiency values calculated from the optimized slab and spherical
models. Common alphavoltaic cell structures will have the isotropic alpha source
randomly distributed on the surface, or in a volume in close proximity to the p-n
junction. Thus, the alpha emitters are distributed in such a way that the bulk of the
particles are not directed in the optimum direction or located at the optimum dis-
tance to the depletion region of the cell. Models 1 and 2 took this into account by
utilizing a mono-directional source so that all alpha particles have the optimum
trajectory and maximum energy deposition occurs within the depletion region.
Thus, the maximum possible energy transport efficiency, ηd, obtained from
Table 5.9 is 9.81 % (average value of all results). The pair production efficiency,
ηpp, for SiC is 42 % [22]. The driving potential efficiency for an optimized PV cell,
ηdp, will generally not exceed 50 %. Assuming that an alphavoltaic cell can produce
fill factors equal to that of high quality solar cells (FF * 0.8), then the theoretical
maximum efficiency of an alphavoltaic cell is about 1.68 % (using models 1 and 2
which have no angular emission dependence). However, one must take into account
the fact that radioisotopes are isotropic emitters. The angular dependence of an
isotropic source will significantly reduce the theoretical maximum efficiency. In
addition, the source must have some thickness. Thus there will be energy loses to
the source material as the alpha particle transports out of the source material into the
transducer material. This too will significantly reduce the theoretical maximum
efficiency of the cell.
A realistic alphavoltaic system was the thin foils example from Sects. 4.4.1.1
and 4.8. In Sect. 4.8, the cell dilution factor of a planer alphavoltaic with a Po-210
foil is DFcell = 0.042. The transport efficiency of depositing alpha particle energy
created by the Po-210 source into the transducer (depletion region) is 5.49 % for a
point source on the surface of a planar p-n junction [4] (this efficiency does not take
into account self absorption losses associated with a source that has thickness and
thus is called the thin source deposition efficiency-ηd-thinsource). The transducer
efficiency for a diamond p-n junction is about 40 % (Sect. 5.1.1). Thus the power
density out from the transducer (Pdout) will be the effective power density
(Pdeffective = 55.24 W cm−3 from Eq. (4.15)) times the transport efficiency
(ηtransport = 0.0549) times the transducer efficiency (ηtransducer = ηdpηppFF = 0.
5 × 0.44 × 0.8 = 0.176) [4],
Pdout ¼ Pdeffective gtransport gtransducer ¼ 55:24 0:0549 0:17 ¼ 0:534 W cm3

ð5:55Þ
The system efficiency of the nuclear battery using a thin source is,
gsystem ¼ gtransport gtransducer ¼ 0:0549 0:170 ¼ 0:0097 ð5:56Þ
5.4.2 Betavoltaics Analysis
There are some similarities between alphavoltaics and betavoltaics. The problem of
matching the range of the beta particle in the medium to the scale length of the
transducer used for energy conversion is the underlying problem for both radiation
types. However, due to the increased range of the higher energy beta particles, the
mismatch between the range of the radiation and the scale length of the transducer
is even more pronounced for betavoltaics. Additional inefficiency is attributed to the
broad energy spectrum of beta particle emission from the source, which makes it
very difficult to harvest energy by depositing energy in a relatively narrow physical
space (i.e., depletion region) in a betavoltaic. A study has shown that beta particles
of different energy have very different regions where maximum pair production rate
exists [22].
Monte Carlo simulations were used in calculating the energy deposition of beta
particles in the depletion region of a silicon carbide (SiC) betavoltaic cell along with
the corresponding theoretical efficiencies [22]. Three Monte Carlo codes were used
in the study in order to benchmark the results: GEANT4, PENELOPE, and
MCNPX. The transportation of beta particles from Y-90, Sr-90, and S-35 were
modeled using the beta particle spectrum emitted by the radioisotopes. The maxi-
mum theoretical energy deposition in both spherical and slab geometries were
found. The maximum energy deposited as a function of distance from position
where the radiation first enters the material was generated. From these calculations
the energy deposition efficiency (ηd) was found. Using the transport efficiency
(ηtransport) and the transducer efficiency (ηtransducer) [4], the maximum efficiencies
for the slab configuration model were approximately 1.54, 0.25, and 0.019 % for
S-35, Sr-90, and Y-90, respectively. The theoretical maximum efficiencies when
using the spherical configuration model were 1.10, 0.17, and 0.013 % for S-35,
Sr-90, and Y-90, respectively. The results show that as beta energy increases, the
mismatch between the range of the beta particle and the scale length of the p-n
junction transducer is worse.
Fig. 5.19 Shown is the GEANT4 simulation of Sr-90 beta decay into SiC with the beam entering
a slab (left) and a point source in the center of a sphere (right). The beta tracks are illustrated by the
dark lines and the bremsstrahlung photons are illustrated by the light lines
The slab model uses a directed electron beam with an energy spectrum char-
acteristic of the specific isotope (Fig. 5.19). The angular emission distribution is not
taken into account which significantly overestimates the energy deposition effi-
ciency. The spherical model puts a point source at the center of a sphere so that all
emitted particles are equidistant from a shell in the sphere (Fig. 5.19).
The spherical model does include some valuable insights on the angular dis-
tribution of the particles emitted from the point source. If the spherical shell sur-
rounding the point source was flattened out, the result would have the
characteristics of a planar source. The theoretical maximum efficiency of point beta
source in the center of a sphere is 1.1 % for S-35, 0.17 % for Sr-90 and 0.013 % for
Y-90 as compared to the theoretical maximum of 2.1 % for a Po-210 point source
in the middle of a sphere. It is also known that a point Po-210 isotropic alpha source
on the surface of a planar p-n junction has a thin source power deposition efficiency
of 5.49 % [4]. As a first approximation for the thin source power deposition effi-
ciency for a point beta source on the surface of a planar p-n junction assume that
there is a linear relationship for the alpha and beta particle behavior. The taking the
ratios of the theoretical maximum for a point source at the center of a sphere is
1.1/2.1 = 0.52 for S-35, 0.17/2.1 = 0.081 for Sr-90 and 0.013/2.1 = 0.0062 for
Y-90. Thus the estimated thin source power deposition efficiency (ηd-thinsource) is
0.52 × 5.49 % = 2.85 % for S-35, 0.081 × 5.49 = 0.44 % for Sr-90 and
0.0062 × 5.49 = 0.034 % for Y-90. These numbers are overestimations of the real
values. However, based on Eq. (5.56), the estimated system efficiencies for a thin
source betavoltaic should be about 0.0097 × 0.52 = 0.005 for S-35,
0.0097 × 0.081 = 0.00079 for Sr-90 and 0.0097 × 0.00620 = 0.00006. The point
is that betavoltaics are fundamentally low efficiency devices due to the mismatch of
the range of a beta particle and the size of a transducer. As the range of the beta
particle increases (as it will when the beta particle energy increases), the betavoltaic
efficiency decreases. Also, betavoltaic cells have a much lower efficiency than
alphavoltaic cells.
Another important observation that can be made is that the power density of a
betavoltaic is lower than that of an alphavoltaic in part due to the lower system
efficiency of a betavoltaic. In addition the power density of the betavoltaic cell will
get much worse as the beta energy decreases.
5.4.3 PIDEC Analysis
As an example of an almost ideal nuclear battery, a PIDEC system was discussed in

Chap. 3. This PIDEC nuclear battery is based on krypton-85 gas which is used both
as the radioisotope source and the transducer. The way that the battery works is the
beta particles emitted from Kr-85 ionize and excite the krypton gas. The Kr ions
and metastable states rapidly form the krypton excimer molecule which then decays
to an unbound ground state while emitting an excimer photon. The excimer
emission is narrow band (with a peak at 149 nm with a Full Width at Half
Maximum of *10 nm). Regardless of the emission angle of the excimer photons
they will they will travel unimpeded in the optically thin medium and intersect with
the photovoltaic cells on the walls of the pressure vessel where the photons are
absorbed (Fig. 5.20). In this nuclear battery configuration it can be safely assumed
that P4 =ðP1 þ P2 þ P3 þ P4 Þ 1: Also, there is no dilution effect because the
gaseous source and the transducer are the same material. Thus, virtually all of the
energy deposited in the gas by the beta particles goes into the production of Kr
excimer fluorescence. It is possible to design the photovoltaic cell so the bulk of the
photons which are absorbed end up in the cell’s depletion zone. The remaining
factors which determine the system efficiency (ηsystem) of the nuclear battery are the
fluorescence production efficiency (ηf), the spectrum match efficiency with
Fig. 5.20 Diagram of a

Kr-85 powered PIDEC
nuclear battery. The Kr-85
gas is both the radioisotope
source and transducer. It
produces excimer
fluorescence which then
interacts with a batter of
photovoltaic cells surrounding
the medium [23]
the band-gap energy of the semiconductor (ηin is defined in Chap. 3, Eq. (3.73)),
the driving potential efficiency (ηdp), and the fill factor (FF). Terms can be grouped
to define the transducer efficiency (ηtransducer = ηin ηdp FF) as shown in Eq. (5.57).
gsystem ¼ gf gtransducer ð5:57Þ
The maximum value of the spectral matching efficiency (ηpv) for various com-
binations of excimers sources and semiconductors is shown in Table 5.10.
From Table 5.10, the fluorescer efficiency for a Kr excimer fluorescer is 0.47. If
AlN is chosen as the semiconductor, the band-gap is 6.2 eV. The ideal transducer
efficiency (ηtransducer = ηinηdpFF) for an AlN photovoltaic cell, as discussed in a
previous example, is the maximum spectral and band-gap matching efficiency
(ηin ≈ ηpv = 0.789), the driving potential efficiency (ηdp ≈ 0.5) and the fill factor
Table 5.10 The theoretical maximum spectral matching efficiency (ηpv) and the efficiency for the
production of ion-to-electrical current (ηie) are shown in the table for selected rare-gas, rare-gas
halide and alkali metal excimer fluorescers
Excimer ηf Eλ Photovoltaic Band-gap energy ηpv = Eg/ ηie = ηpvηf
(eV) material Eg (eV) Eλ
Ar2* 0.5 9.6 AlN 6.2 0.645 0.324
Kr2* 0.47 8.4 AlN 6.2 0.789 0.345
0.47 8.4 Diamond 5.5 0.655 0.308
F2* 0.44 7.8 AlN 6.2 0.79 0.35
0.44 7.8 Diamond 5.5 0.71 0.31
Xe2* 0.48 7.2 AlN 6.2 0.861 0.413
0.48 7.2 Diamond 5.5 0.764 0.367
ArF* 0.35 6.4 AlN 6.2 0.969 0.339
0.35 6.4 Diamond 5.5 0.859 0.301
KrBr* 0.33 6 Diamond 5.5 0.917 0.302
KrCl* 0.31 5.6 Diamond 5.5 0.982 0.304
Na2* 0.46 2.84 ZnSe 2.7 0.951 0.437
0.46 2.84 SiC (3C) 2.3 0.810 0.373
Li2* 0.42 2.7 CuAlSe2 2.6 0.963 0.404
0.42 2.7 SiC (3C) 2.3 0.852 0.358
Hg2* 0.21 2.58 GaS 2.5 0.97 0.2
0.21 2.58 SiC 2.4 0.93 0.19
ArO* 0.11 2.27 GaP 2.2 0.97 0.105
0.11 2.27 GaAlAs 2.2 0.97 0.105
KrO* 0.13 2.27 GaP 2.2 0.97 0.125
0.13 2.27 GaAlAs 2.2 0.97 0.125
XeO* 0.15 2.27 GaP 2.2 0.97 0.145
0.15 2.27 GaAlAs 2.2 0.97 0.145
The fluorescence efficiency (ηf), the energy of the photon (Eλ) and the band gap of the photovoltaic
(Eg) are also shown
(FF ≈ 0.8). Thus the system efficiency can be found from Eq. (5.57)
(ηsystem = 0.47 × 0.789 × 0.5 × 0.8 ≈ 0.148).
The Kr-85 PIDEC nuclear battery drew commercial interest in the early 1990s
because of its simplicity, no dilution factor and because of its optimized utilization
of the deposited energy [23]. As described in Chap. 3, the author and his collab-
orators designed a device with a mission requirement of 1,322 Watts of thermal
power. The design required an activity of 1,000,000 Curies (based on the estimated
system efficiency) of krypton-85 gas. The first step in the calculation is to find the
atomic density of krypton-85. The activity of Kr-85 when the nuclear battery is first
filled is
Að0Þ ¼ kN ð0Þ ð5:58Þ
where
λ = 0.693/t½ = 0.693/(10.755 yr × 365 d/yr × 24 h/d × 3600 s/hr) =
2.043 × 10−9 s−1.
The number of Kr-85 atoms in 1,000,000 Ci is,
decays
106 Ci 3:7 10 10 sCi
Nð0Þ ¼ ¼ 1:920 1021 atoms ð5:59Þ
2:043 109 s1
Using the fact that there are 2.68 × 1019 atoms (or molecules) cm−3 in one
atmosphere of gas at Standard Temperature/Pressure (STP), the pressure of Kr-85
gas can be found for the nuclear battery container. The calculation begins by
assuming: (1) the volume is spherical (which is 5 cm3) for this example; and (2) the
gas pressure in the sphere is governed by the idea gas law. The number of atoms (or
molecules) in a sphere at a pressure of one atmosphere (Natm) is found by multi-
plying the sphere volume by the number of gas atoms (or molecules) in a one cubic
centimeter volume (Volumesphere cm3 × 2.68 × 1019 atoms cm−3). The pressure of
Kr-85 in the 1000 cm3 sphere is then found by using the relationship PKr85 = N(0)/
Natm. This calculation shows that the pressure of 1,000,000 Ci of Kr-85 in a
1000 cm3 volume is 675.8 atmospheres. The thermal power in one million curies of
Kr-85 is 1,322 W. From an engineering perspective, neither the size of the sphere
nor the gas pressure presents a significant challenge.
The sphere (having a radius of 6.02 cm) is small enough to be approximated as a
point source in a shielding calculation. Using the Radpro Calculator [24] to estimate
the dose-rate at one meter from the bare sphere, the dose was found to be about
12.71 Rem/hr. This dose-rate was determined to be unacceptable by the industrial
partners working with the author in the early 1990s, due to personnel exposure
issues and radiation sensitive electronics in the package that was planned to be
deployed along with the power supply. In order to reduce the dose rate to an
acceptable level at 1 m, the device required a shield. Several shield thicknesses
have been examined using the Radpro Calculator and it can be shown that 8 cm
thick lead shield reduced the dose-rate to an acceptable level of 24.3 μR/hr. The
mass of the shield is about 124 kg. This indicates that the device would have a mass
to power ratio of approximately 0.124 kg/(W-thermal). The radius of the sphere is
14.2 cm. The volume of the sphere and shield is 11,990 cm3. The power density is
calculated by dividing the thermal power output by the device volume
(1,322/11990 = 0.11 W cm−3). If aluminum nitride photovoltaic cells are inter-
faced to the Kr-85 source, the electrical power that can be produced is found by
multiplying 1,322 W-thermal times the transducer efficiency as calculated above of
0.148. The predicted electrical power output is 195.7 W-electric. From Table 5.1,
the maximum power output of liquid Kr per gm is 0.517808194 W gm−1 and the
density is 2.413 gm cm−3. The fundamental parameters for the Kr-85 cell are,
• Radioisotope Power Output = 1,322 W
• Cell Volume = 11990 cm3

• Power Dilution Factor = 1322= 11990½cm3 0:5178½W/gm2:413 gm/cm3
¼0:08825
The system efficiency is ηsystem = 0.148 with a battery power output of 195.7 W.
The cell power density is 195.7/11990 = 0.0163 W cm−3.
Kr-85 is one of the safest radioisotopes available. It is a gas and quickly dis-
perses in air if released. It is a rare gas and has virtually no biological half-life if it
enters the lungs. These benefits and numbers are attractive for a nuclear battery.
However, the weakness of the Kr-85 PIDEC nuclear battery is that the world’s
supply of Kr-85 is about 142,000,000 Ci [3]. Thus, there is enough Kr-85 to build
only 142 of these nuclear batteries.
Most of the other potential isotopes discussed in Chap. 2 are in the solid phase.
It is possible to use these more plentiful isotopes which are in the solid phase in the
PIDEC based nuclear battery. An aerosol interface between the solid-state isotope
and the excimer gas (Fig. 5.21) will have some efficiency benefits. Below a PIDEC
nuclear battery which uses an aerosol interface with an aerosol made up of highly
Fig. 5.21 A transducer gas/aerosol source interface in the PIDEC nuclear battery [25]
reflective Gd-148 spherical particles of 5 μm radius is discussed. Gd-148 is chosen

because it is an almost perfect alpha emitter with a particle of energy of 3.182 MeV
and a half-life of 74.6 years. Assume that the aerosol particles are pure Gd-148
metal spheroids with a density of 7.95 gm cm−3 coated with a thin film of aluminum
reflector material. The power output per gm from metallic Gd-148 is 0.61 W gm−1.
One of the highest efficiency excimers is xenon. Rare gas excimers have the highest
fluorescence efficiencies of the excimer family. The xenon excimer has the longest
wavelength of the rare-gas excimers at 172 nm. The longer wavelength is important
because the reflectivity of aluminum falls off at about 160 nm (aluminum is the best
reflector for the vacuum ultra violet). Of the viable photovoltaic materials that have
matchup with the xenon excimer spectrum, AlN (Eg = 6.2 eV) has the highest
band-gap energy.
The analysis of this system is similar to the Kr-85 based PIDEC system. In this
system, the choice of gas can be based on the best excimer and semiconductor
match as described above. The combination of Xe (ηf = 0.48) and an AlN photo-
voltaic cell (ηin ≈ ηpv = 0.861), which has a driving potential efficiency of 0.4 and a
fill factor of 0.8, is the best choice. This combination has a product of ηf ηtransducer
equal to 0.165. This represents a base efficiency for the system minus consideration
of the losses of photons transporting through the aerosol/gas mixture. To find the
system efficiency, the next step is to find the transport efficiency of the excimer
photons through the reflective aerosol particles/gas mixture. The problem involves a
weak dusty plasma excited by nuclear reactions. This concept was first studied as a
viable energy conversion system based on nuclear-driven flashlamps by one of the
authors in 1981 [26]. From this research it was determined that a weak dusty plasma
combination can be optically thin for a nuclear-driven flashlamp of reasonable
size [27].
The methodology of analyzing a reflective aerosol/weak plasma combination
was developed [28–30]. Following a diffusion model derived for reflective aerosol
particles by the author [1, 30], the transport of photons created in a weak plasma
(formed by charged particle emission in the reflective radioisotope aerosol particles
to the fluorescer gas) can be used to find the photon transport properties of the cell.
The development of a code based on transport theory is complex and difficult [31].
A model based on diffusion theory is feasible and was developed [32]. The fun-
damental aspect of the model is that the cross section of an aerosol particle can be
approximated by assuming that the particle is spherical. The total cross section of a
spherical particle is,
rt ¼ prp2 ð5:60Þ
where rp is the radius of an aerosol particle.

The particles have to be sized in the micrometer range based on the transport
range of the emitted ionizing radiation from the radioisotope. Part of the solution for
this problem is the calculation of the power transport efficiency from the charged
particles emitted by the radioisotope that forms the spherical pellets to the excimer
gas. The power deposited in the gas is then converted into photons through the
formation of excimers.
Photons moving through the aerosol/weak plasma mix and interact. The inter-
action of photons with an aerosol particle is modeled by finding the probability of a
scattering event when a photon strikes an aerosol particle. The probability of the
photon scattering off of the particle is the total reflection coefficient (Rtotal). This
total reflection coefficient is due only to the reflectivity of the thin aluminum film
coating on the particle (R0). Rayleigh scattering (an effect which occurs when the
photon wavelength is on the order of the particle size) is not a factor because the
optimum particle sizes are larger than the wavelength of the photons. Thus the
scattering the cross section is,
rs ¼ R0 rt ð5:61Þ
The absorption cross section is found by assuming that the photon either is
scattered off the particle or absorbed by the particle. Thus the absorption cross
section is,
ra ¼ ð1 R0 Þrt ð5:62Þ
Three assumptions are made about the behavior of photons in an environment

where particles are reflective in order to apply diffusion theory:
1. σs > σa
2. The intensity is a slowly varying function of position
3. Intensity does not vary with time
These assumptions closely resemble those of neutron diffusion theory used in
reactor physics. The governing equation for diffusion theory is,
U S
r2 U ¼ ð5:63Þ
L2 D

where L2 = D/Σa, Σa = Gσa, G = aerosol density (cm−3), D ¼ Rs = 3R2t and
S = photon source (photons cm−3 s−1).
In order to solve Eq. (5.63), the geometry needs to be chosen. In this case the
cell uses a spherical geometry (but other geometries can be used as well).
Equation (5.63) becomes,

1 d 2 dUðr Þ Uðr Þ S
2
r 2 ¼ ð5:64Þ
r dr dr L D
The boundary conditions come from the physics of the system. The photovoltaic
cells surrounding the sphere are assumed to be a strong photon absorber so the
photon interaction at the wall resembles the effects of a control rod in the analogous
neutron transport problem [33]. The boundary condition at the absorbing wall
assumes that no partial current travels backwards or an alternate view is that no

photon escapes the photovoltaic absorber on the wall.

Uðr Þ D dUðr Þ
J ðaÞ ¼ þ ¼0 ð5:65Þ
4 2 dr b
where b is the radius of the sphere.

The second boundary condition can be derived from the partial current condi-
tions at the center of the sphere. In order to avoid a singularity at the center, it is
assumed that a small reflective sphere is placed at the center of cell. The reflective
sphere has a radius “a” and a wall reflectivity of R1. Thus the boundary condition at
the inner sphere wall is

1 D dUð xÞ
Jaþ ¼ R1 Ja ; where J ðaÞ ¼ Uð xÞ ð5:66Þ
4 2 dx a
The power balance (discussed below) of the cell can also be used as the second
boundary condition or as a verification check. (Both the partial current boundary
condition and power balance boundary condition give the same results.)
There are three potential loss mechanisms for a photon as it moves through the
cell: (1) It is absorbed by an aerosol particle, (2) It is absorbed by a photovoltaic cell
on the wall at radius b, or (3) it is absorbed by the wall of the reflective sphere at
radius a. The photon source strength per unit volume is a function of the physics of
the alpha particles transporting out of the aerosol particles and into the gas. The
source function has the form,
S ¼ C1 Grp3 eArp ð5:67Þ
where C1 and A are constants related to the activity of the radioisotope and material
properties. A is a function of the mean free path of an alpha in the source (λl). C1 is a
function of the power deposited in the fluorescer gas by the alphas exiting the
source particle and the fluorescence production efficiency. Thus the term eArp
represents the transport efficiency of the alpha particle energy out of the source
particle [34]. The term C1 Grp3 (similar to Eq. 3.14 in Chap. 3 of Nuclear-Pumped
Lasers [1]) is the average power density deposited in the fluorescer gas from the
aerosol source times the efficiency of the production of fluorescence from the gas
divided by the energy per photon. This gives the number of photons produced per
cubic centimeter per second.
A photon balance equation can be written realizing that the following must be
true,
Photons Absorbed by Inner Sphere þ Photons Absorbed by Aerosol þ

ð5:68Þ
Photon Absorbed by Photovoltaics ¼ S Volume
with,

Loss to Photovoltaics ¼ J ðbÞ 4pb2 ð5:69Þ
where b is the outer radius of the spherical chamber.

Z b
Loss to Aerosol ¼ Ra Uðr Þ 4pr 2 dr ð5:70Þ
a

Loss to Inner Sphere ¼ J þ ðaÞ 4pa2 J ðaÞ 4pa2 ð5:71Þ
Thus,
Z
b 4p 3
J ðbÞ 4pb 2
þ Ra Uðr Þ 4pr 2 dr þ ð1 R1 ÞJ þ ðaÞ 4pa2 ¼ S b a3
a 3
ð5:72Þ
The fraction of energy lost to the PV cell (or coupling efficiency ηL) is,
Photons Absorbed by PV Cell J ðbÞð4pb2 Þ J ðbÞð3b2 Þ

gL ¼ ¼ 4p 3 ¼ ð5:73Þ
Total Photons Produced S 3 ð b a3 Þ Sð b3 a3 Þ
The fraction of energy absorbed by the aerosol particles (ηa) is,

Rb Rb
Photons Absorbed by Aerosol Ra Uðr Þð4pr 2 Þdr 3Ra a Uðr Þr 2 dr
gc ¼ ¼ a
¼
Total Photons Produced 3 ðb a Þ
S 4p 3 3 Sð b3 a3 Þ
ð5:74Þ
The fraction of energy absorbed by the inner sphere (ηw) is,
Photons Absorbed by Wall ð1 R1 ÞJ þ ðaÞð4pa2 Þ ð1 R1 ÞJ þ ðaÞð3a2 Þ

gw ¼ ¼ ¼
Total Photons Produced 3 ðb a Þ
S 4p 3 3 Sð b3 a3 Þ
ð5:75Þ
Solving Eq. (5.64), the photon intensity as a function of r is,
2 expðr=LÞ L expðr=LÞ SL2

Uðr Þ ¼ C1 þ C2 þ ð5:76Þ
r 2r D
Using the following parameters,

1. b = 10
2. a=1
3. rp = 0.0005
4. G = 1,000,000
5. R1 = R2 = 0.9
6. Effp = Exp[-rp/(9 × 10−4)]
From these input parameters,
• VolumePellet = (4/3) × π × r3p = 5.236 × 10−10 cm
• MassPellet = VolumePellet × Desityradioisotope = 4.163 × 10−9
• PowerPellet = MassPellet × Wattpergm = 2.54 × 10−9
• Powerdensity = G × PowerPellet = 0.00254
• FluorescerEff = 0.5
• S = Powerdensity × FluorescerEff
The following important terms were found using a Mathamatica model similar to
one that was discussed in the book Nuclear-Pumped Lasers [1],
• C1 = 0.00021645582571915947
• C2 = −0.0012388587796092745
• PV Absorption Efficiency = ηPVabsorption = 0.4065
• Percent Absorption in Aerosol = 0.593
• Percent Absorption in Inner Sphere Wall = 0.000881
• Radioisotope Power Output = 10.625 W
• Cell Volume = 4184.6 cm3
• Power Dilution Factor = 10:625
3 ¼ 00:0005236
4184:6½cm3 0:61½W/gm7:95 gm/cm
The system efficiency is ηsystem = ηf ηtransducer ηPVabsorption = 0.165 × 0.4065 =

0.067 with a battery power output of 0.71 W. The cell power density is
0.000171 W cm−3.
It is interesting to compare the performance of the two PIDEC battery designs,
one using a Kr-85 volume source and the other a reflective aerosol source made of
Gd-148. The power density of the Kr-85 battery is over 1181 times higher than the
Gd-148 battery. This is primarily due to the difference in the power dilution factors
for the two cells (*1,910 times). The conclusion from this comparison is that the
dilution factor, which is a result of the scale length mismatch between charged
particle range of the source and the size of the transducer, significantly reduces
nuclear battery performance.
5.5 Analysis of Problems in Nuclear Battery Literature 269
5.5 Analysis of Problems in Nuclear Battery Literature
The difficulty in following alpha and beta voltaic cell research literature is that most
studies fail to provide enough information for a reader to fully understand the
experiment and to properly interpret the results. The problems begin with a com-
plete description of the important variables. These variables include information
about the radionuclide source, the mix of materials which make up the source (e.g.,
a fairly common practice is to mix the radionuclide with gold for example), how the
source is geometrically coupled to the cell, the dimensions of the source, the p-n
junction material, and both the geometry and dimensions of the p-n junction. To
illustrate these factors in an example, suppose that device uses a polonium 210
source (Po-210 is an alpha emitter) which is mixed with silver at a 1 to 10 ratio.
This metal mix is then rolled into a foil. If a 5 μm foil of the Po-210 source material
is placed on top of a 1 cm by 1 cm p-n junction made of SiC, then in order to fully
describe the experiment it is important to know the ratio of Po-210 to silver, the
surface area of the p-n junction, the thickness of the source (t1), the activity of the
source, the thickness of the n-type layer (t2), the depletion width (t3), and the
thickness of the p-type layer (t4). The thicknesses are needed to model the path of
alpha particle emissions from the source into the depletion layer. The foil will
absorb some of the energy of an alpha particle. For example, commercially avail-
able encapsulated alpha sources typically lose *10 % of the alpha particle energy
in the source structure [35]. It is reasonable to assume that the Po-210 atoms are
uniformly distributed in the foil. In setting up a Monte Carlo transport model, the
location of the atom which decays in the foil, the time of decay, and the emission
angle can be incorporated. Since alpha decays are isotropic, there is an equal
probability that the alpha particle will be emitted at any possible solid angle. Thus
half of the emission trajectories are away from the surface of the p-n junction. The
Monte Carlo transport model would then follow the trajectory of the alpha particle
and determine its path and where electron-hole pairs are created. Thus a complete
description of the geometry of the device is necessary. If an experimental paper fails
to provide the complete description, then the experiment is not fully described. The
reader can’t model the alpha or beta voltaic cell without making potentially invalid
assumptions about the missing variables.
There are some basic methods which a reader can use to analyze nuclear battery
reports. Beginning with the fundamental concept presented in Fig. 5.1, a nuclear
battery is made up of layers. Only the power deposited in the transducer layer
produces useful power output. There is a power deposition efficiency (ηpd) which is
calculated by analyzing the energy transport properties of the ionizing radiation in
the various layers of materials (gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 ÞÞ. This is a complex
calculation which requires the use of a sophisticated transport code such as
GEANT4 or MCNP. As discussed in prior sections of this book, there must be a
good matchup between the range of the ionizing radiation and the scale length of
the transducer. A scale length matchup for a nuclear batter is analogous to an
impedance match in circuit analysis. A good scale length match means that the
power deposition efficiency is high or vice versa. There are studies in nuclear
battery research which report impossibly high efficiencies. When these studies are
examined in depth, it appears that there are three general explanations: (1) Errors;
(2) Misunderstandings; or (3) To gain attention.
A basic understanding of the parameters that make up the photovoltaic trans-
ducer efficiency, ηtransducer = ηdpηppFF, is a good starting point for the reader to
help identify any problems in a nuclear battery report. The pair production effi-
ciency, ηpp, when multiplied by the term P4 in Fig. 5.1, is proportional to the
number of electron-hole pairs produced in the transducer. Table 3.9 shows the pair
production efficiency for some typical semiconductors. For example the pair pro-
duction efficiency for silicon is 0.308, for germanium is 0.23, for silicon carbide is
0.421, for gallium nitride is 0.381 and for diamond is 0.442. As discussed in
Table 3.9, the energy required to produce an electron-hole pair (or W value) in a
semiconductor material can be estimated by the Klein formula,
W ¼ 2:8Eg þ 0:5 eV ð5:77Þ
The importance of the knowing the pair production efficiency is that the effi-
ciency of an alphavoltaic or betavoltaic can never exceed ηpp. If the reported system
efficiency exceeds ηpp, it is simply not plausible.
A p-n junction has other inherent inefficiencies which reduce the performance of
the cell. One of the performance reducing factors is the driving potential efficiency,
ηdp = V/Eg(where V is the photovoltaic cell’s operating voltage). The importance of
the driving potential efficiency can be understood by following an electron which
has been elevated from the valance band to the conduction band at an expenditure
of energy which is at least equal to the band-gap energy, Eg. The excess energy
which is used to elevate an electron to the conduction band is factored in the
definition of the pair production efficiency, ηpp. Thus from an energetics point of
view, the electron which has been created as part of an electron-hole pair has
available an energy of Eg. If the electron-hole pair recombines, the energy Eg is
dissipated in the process. If the electron becomes part of the current produced in the
p-n junction, the energy level Eg is its starting point. The Fermi energy level at the
interface between the p-type and n-type materials is where the diffusion of charge
carriers across the junction sets up a space charge which in turn sets up an electrical
field. This region where the electrical field is formed is called the depletion region.
If an electron-hole pair is formed in the depletion region, the electrical field causes
the electron-hole pair to separate. The potential energy of the electron within the
junction’s electric field is the voltage “V” times the elementary charge “e” (where
1e = 1.6 × 10−19 C) or Ve (which has the units of electron volts). It should be noted
that the actual operating voltage of the cell (V) is less than the open circuit voltage
(Voc). The driving potential efficiency can be defined as the available potential
energy of the electron within the junction (Ve) divided by the energy level of the
band-gap (Eg) or Ve/Eg. If the device were perfect, Ve would equal Eg which means
that the potential energy of the electron in the junction would exactly equal to the
energy expended to place the electron in the conduction band. As discussed, the
electric field created within the junction is not perfect thus there is a need to account
for this imperfection as a loss in the form of the driving potential efficiency.
In Chap. 3 it was discussed that the driving potential efficiency goes up with
band-gap energy. However, it is important to develop a feel for real driving
potential efficiencies. High quality silicon solar cells typically have a driving
potential efficiency of 0.5. This optimized driving potential efficiency comes from
the technology of perfecting silicon for the semiconductor industry and the effort of
perfecting silicon has taken many decades and hundreds of billions of dollars of
world-wide investment. Thus the driving potential efficiency of silicon has reached
a plateau. On the other hand wide band-gap materials are still in their infancy and
have a long way to go before reaching a plateau. Of the wide band-gap materials,
diamond has demonstrated the highest reported driving potential at about 0.48 [5].
The pair production and driving potential efficiencies do provide an additional
basis for the reader to analyze reported nuclear battery system efficiencies. If the
reported efficiency of the nuclear battery exceeds the product of ηppηdp, this is
simply not feasible.
One other important factor that governs the efficiency of a p-n junction is the Fill
Factor (FF). Adding the fill factor to the model, the transducer efficiency becomes,
ηtransducer = ηdpηppFF. The best silicon solar cell has a fill factor that approaches
0.8. This is the optimized limit for silicon and it is based on the highest purity,
defect free silicon photovoltaic cell available. Thus, the reader has one additional
tool to analyze the validity of a reported efficiency for a p-n junction based nuclear
battery. If the nuclear battery has a reported system efficiency that exceeds the
maximum possible transducer efficiency for a p-n junction based battery
(ηtransducer = ηdpηppFF), then the reported efficiency is impossible.
The maximum efficiency that p-n junction based nuclear battery can achieve
(ηsysmax) must take into account the fraction of power from the source which
is deposited in the transducer, gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 Þ: So, ηsysmax =
ηpdηtransducer. The reader now has a set of tools which can be used to evaluate
whether or not a reported p-n junction based nuclear battery is reasonable.
The above analysis for a p-n junction based nuclear battery can serve as a road
map for the reader to analyze other nuclear battery designs that use different
transducers. The simplest variation is a Schottky barrier. The main difference in the
analysis is in the driving potential efficiency. In a Schottky barrier the voltage is the
barrier voltage (Vb). Thus the driving potential efficiency, ηdp = Vbe/Eg changes.
The other important difference is the thickness of the Schottky barrier transducer
layer which will be at least an order of magnitude less than that of a p-n junction
transducer layer. This makes a huge difference in ηpd. Basically the reader can
follow the same analytical procedure in evaluating reported results for a Schottky
barrier based nuclear battery as was discussed for a p-n junction based nuclear
battery.
Looking at the factors which govern a single junction cell structure, the maxi-
mum transducer efficiency as well as the maximum system efficiency can be sur-
mised. The logic begins with the previously discussed pair production efficiency,
ηpp, which varies from 0.2 to 0.44 for solids. The highest value occurs in diamond.
For a gas, ηpp varies from 0.3 to 0.5. The highest value occurs in rare gases. As a
first cut, the reader should be very suspicious of any report of system efficiency
greater than 0.4 for a solid transducer. A system efficiency greater than ηpp can be
called the perpetual motion limit for the device. This limit comes from the fraction
of energy from ionizing radiation that can go into the production of electron-hole
pairs.
Other factors are important also. For a p-n junction or Schottky barrier based
nuclear battery, the factors of ηdp and FF set the efficiency limit of an ideal
transducer. As discussed in previous sections, ηdp is typically below 0.5 for a p-n
junction and below 0.25 for a Schottky barrier. The fill factor will be less than 0.8.
Thus maximum transducer efficiencies will be below 0.17 for solid state p-n
junction based nuclear batteries and below 0.08 for Schottky barrier based nuclear
batteries. Finally, the power deposition efficiency (ηdp) needs to be factored into the
efficiency considerations. Depending on the matchup between the range of the
ionizing radiation from the source and the scale length of the transducer, the power
deposition efficiency calculation is usually a complex transport problem. The values
of power deposition efficiency can be anywhere from 0.001 to 0.1. The maximum
system efficiency will be below 0.017 for a p-n junction based nuclear battery and
below 0.008 for a Schottky barrier based nuclear battery.
A nuclear battery design may claim new types of operational mechanisms. This
category of claim is difficult to interpret without a full understanding of the exact
mechanism. However, the reader does have some analytical tools available as
described in this text to help gain some insight into unusual designs and unknown
mechanisms.
1. Does the mechanism depend upon the production of heat or does it depend upon
the formation of ions?
2. If the mechanism depends upon ion production, then the pair production effi-
ciency (ηpp) sets the upper limit of the maximum possible system efficiency.
3. Of the power deposited in the transducer, consider the efficiency of converting
this power to a useful product such as electrical power (ηtransducer).
4. The transport efficiency (ηpd) of depositing power from ionizing radiation into
the transducer also must be found.
An example of a nuclear battery which includes claims of a new mechanism is the
water based Plasmon-assisted radiolytic energy conversion device [36]. The nuclear
battery uses a Sr-90 source mounted in the fluid. Beta particles stream through the
water and have sufficient range to reach a Schottky barrier made with a thin platinum
layer on top of a nanoporous TiO2 structure. The band-gap of the TiO2 is 5.2 eV. The
Schottky barrier height is 0.45 eV. Beta particles that reach the Schottky barrier from
the source interact with the TiO2 to produce electron-hole pairs. The driving potential
of the barrier height creates the current flow. It is hypothesized that aqueous electrons
(eaq) from the radiolysis of water reach the platinum and some of these aqueous
electrons are converted into electron-hole pairs in the platinum. It is claimed that this
effect adds additional electrons to the current flow and boosts the power output of the
cell. The reader should follow steps 1 to 4. From step 1, the reader will determine that
this nuclear battery depends on ion production. In step 2, the reader needs to
determine the pair production efficiency (ηpp). This determination has two parts. The
first part is to find the efficiency of electron-hole pair production in TiO2. This
requires that the W value for the material be found. There are no experimental
measurements of W for TiO2 so the reader should use the Klein formula described in
Chap. 3 (W = 2.8 Eg + 0.5). The Klein formula does present some concerns. As can
be seen in Table 3.9, the Klein formula tends to overestimate the W value for wide
band-gap materials. Based on the Klein formula, the W value is estimated as being 2.
8 × 5.2 + 0.5 = 15.06 eV. Thus the pair production efficiency (ηpp) for TiO2 is 5.
2/15.06 = 0.3453. The second part is the determination of the maximum efficiency
of converting aqueous electrons into electron-hole pairs which add to the current
e
flow in the Schottky barrier (gppaq ). Assuming that the mechanism for conversion of
aqueous electrons to electron-hole pairs is correct, the maximum efficiency for this
contribution to current can be calculated. The G value for the production of aqueous
electrons with beta particles of the energy range from Sr-90, as shown in Table 5.11,
is 2.63 [37, 38]. Recall that the definition of G is the number of species produced per
100 eV deposited. The 2.63 aqueous electrons are produced per 100 eV of energy
deposited in water by the beta particles. Assume that the diffusion of aqueous
electrons to the platinum electrode is perfect in order to calculate the maximum pair
production efficiency due to the Plasmon-assisted effect. The energy expenditure for
the production of an aqueous electron is 100/2.63 = 38.02 eV per aqueous electron.
If the Plasmon-assisted reaction is perfect in producing an electron-hole pair in the
e
Schottky barrier, the pair production efficiency for aqueous electron conversion (gppaq )
is the barrier height (or potential energy per electron added to the circuit) divided by
the energy expenditure per aqueous electron, 0.45/38.02 = 0.0118. The total pair
e
production efficiency is ηpp + gppaq = 0.3453 + 0.0118 = 0.3571. This value repre-
sents the absolute upper limit for the system efficiency. As discussed above, there are
other significant inefficiencies associated with Schottky barrier systems that reduce
the operation efficiency. Consider that the driving potential efficiency is ηdp = 0.
Table 5.11 Free radicle and molecular product yields in water for different types of ionizing
radiation [38]
Radiation Type G(−H2O) G G G G
(H2 + H2O2) (e−aq) (H) (OH)
X-Rays and fast electrons 4.08 1.13 2.63 0.55 2.72
(0.1 MeV < E < 20 MeV) (3 < pH < 13)
Alpha (12 MeV) 2.84 2.19 0.42 0.27 0.54
(pH 7)
Po Alpha (3 MeV) 3.62 3.02 0 0.60 0.50
(pH 0.46)
45/5.2 = 0.087 and the maximum FF * 0.8. The realistic transducer efficiency is
far less than the calculated maximum value, 0.3571 × 0.087 × 0.8 = 0.024. If the
power deposition efficiency, gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 Þ; is also considered, the
device efficiency will drop another order of magnitude or more.
There are other potential issues in the existing nuclear battery literature of which
the reader needs to be aware. The issues impact the interpretation of reported
experiments. One example is the method that is used to measure the power gen-
erated from a beta or alpha voltaic cell. A p-n junction has an operating voltage
(Vop) and current (Iop) which is dependent on the p-n junction’s material properties
(e.g., type of material, carrier lifetime, defect density, etc.). The product of the
operating voltage and current represents the power output from the cell. The sources
are typically of low activity (A < 1 mCi). This low activity leads to low power
output. Most reported data is on the order of pico-Watts to nano-Watts. So for
example, if a SiC cell has reported power output of one nano-Watt, it should have
an operating voltage on the order of 1.4 V. Thus the current output would be on the
order of 0.71 nA. A current level of 0.71 nA is difficult to measure and requires a
great deal of attention to the conditioning of the power lines and ground which
provides power to the device used to measure current (e.g., a picoammeter). The
conditioning is done to take out ripple effects that can cause errors in the
picoammeter readings. Picoammeters can also exhibit an electronic drift which can
lead to higher output signals at the beginning of a measurement which then
decreases with time. If the experimental measurement methods are not well
described, the reader would have to decide whether or not the authors took
appropriate precautions in making these critical measurements.
Another example of reported experiments in the literature has to do with the
definition of efficiency that is used. The standard method of reporting is to use the
absolute efficiency (ηab) of a cell which is defined as the power out of the cell
divided by the total power produced by the radioisotope source (Eq. (5.27)). Not all
researchers report absolute efficiency and sometimes it is not clear how the reported
efficiency is defined. One of the common efficiencies that the reader may encounter
in nuclear battery literature is the intrinsic efficiency (ηint), a term defined and used
in this book for illustrative purposes. The intrinsic efficiency is defined as the power
out (Pout) of the cell divided by the power deposited the depletion layer, Pdpl or
from Fig. 5.1 for the general cell, P4 (Eq. (5.52)). The intrinsic efficiency does not
take into account the power deposited in materials other than the transducer (e.g.,
gpd ¼ P4 =ðP1 þ P2 þ P3 þ P4 ÞÞ. The intrinsic efficiency will typically be an order
of magnitude or greater than the absolute efficiency.
Pout
gint ¼ ð5:78Þ
Pdpl
There are also issues in how an author chooses to calculate the power deposited
in the depletion layer. This calculation requires specific details about how the alpha
or beta source is constructed (an example of which is shown in Fig. 4.12), the exact
geometry of the source interface with the cell, and the critical dimensions of the
cell. Manufacturers of alpha or beta sources will provide data with regard to a
standard design (e.g., Eckert and Ziegler Isotope Product radioisotope sources).
However, the thin metal coatings can vary ±30 % for high precision sources or
±200 % for lower precision sources. The thickness of the metal coatings will
impact the energies of the alpha particles which are emitted from the source’s
surface. The emitted alpha particles will have a spectral distribution and an angular
distribution which is dependent upon the dimensions of the source. For example, a
Nuclespot Alpha Ionizer model P-2042 will have an average alpha energy of
4.5 MeV coming out of the surface of the ionizer even though the alpha energy
emitted from Po-210 reaction is 5.3 MeV [35]. The 0.8 MeV difference is due to
the self-absorption losses from the alpha particles as they transport through the
source. Models which do not include a full alpha spectrum or its angular distri-
bution produce differences in energy deposition which are substantially different
from the actual value. Beta sources are even more complex due to beta spectrum
[22, 39]. If the author relies on an all too commonly used approximation, namely
that the beta source is mono-energetic (i.e., the average beta energy is 1/3 βmax), the
error in the transport calculation will be substantial. Thus authors who report
intrinsic efficiencies need to provide the reader with complete details of the source,
a complete set of dimensions (e.g., layer thicknesses, areas, etc.), the geometry of
the source coupling to the cell, how Pdpl is calculated and complete data on the cell
design in order for the reader to understand if the calculation of the power deposited
in the depletion layer was done correctly.
Some authors have defined efficiency as the power out (Pout) divided by the
power absorbed in the device (Pabdev). So for the example cell shown in Fig. 4.11,
only the alpha particles which have trajectories that intersect the p-n junction are
counted. Thus 50 % of the particles going the wrong way are not counted. This
efficiency will be called the device efficiency ηdev, and is defined in Eq. (5.79) [3].
The device efficiency will be about twice as high as the absolute efficiency. Again,
authors can make some of the same errors in calculating the transport efficiency as
are made in calculating the intrinsic efficiency as described above.
Pout
gdev ¼ ð5:79Þ
Pabdev
In most cases the authors are not purposefully misleading the reader by reporting
the intrinsic efficiency and device efficiency instead of the absolute efficiency. From
the author’s point of view by taking out geometrical efficiencies in transporting
power to the transducer, the maximum potential of the nuclear battery design can be
presented. However, the reliance on the reader to fully understand the subtleties of
the intrinsic efficiency or the device efficiency is problematic.
Another important parameter to note is whether or not an external bias to the p-n
junction is used to increase the depletion width of the cell. If an external bias is
used, then the external voltage source can contribute current output. Often times the
authors do not account for this important contribution in current measurements and
thus will report efficiencies which are higher than they should be.
In order to make meaningful comparisons between nuclear battery designs, the
absolute efficiency of the devices are needed. The use of intrinsic efficiencies and
device efficiencies detracts the reader from the pertinent facts and leads to incorrect
conclusions.
A final but very important issue is that p-n junction based transducers are highly
susceptible to radiation damage which effects their operational lifetime adversely.
Ionizing radiation displaces atoms within the crystal lattice thus creating a vacancy.
The displaced atom is typically located at an interstitial site [40]. The rate of
displacements is directly proportional to the power density created by the interac-
tion of the radiation source with the material. For example, an alpha particle with
high Linear Energy Transfer (LET) can displace a couple of hundred atoms in the
crystal. The characteristics of p-n junctions degrade with radiation exposure, thus
limiting the lifetime of the junction. It has been argued that wide band-gap p-n
junctions are less susceptible to radiation, but even wide band-gap p-n junctions are
impacted by radiation damage since the binding energy of atoms in the crystal is
much lower than the energy of ionizing radiation. Some authors [41] provide time
dependent data on the power output from the cell which shows how radiation
damage impacts the device. Radiation damage could be an issue for transducers
other than photovoltaic cells. The reader needs to make this determination for the
specific type of transducer.
Table 5.12 presents a representative group of recent alpha- and betavoltaic cell
studies from the literature. As the table shows, important properties of the cell need
to be reported and it states whether or not these important dimensions were
reported. Because so many variables are incorporated in the radioisotope design, all
of these variables can be combined into a column titled “Is Radioisotope Design
Information Sufficient.”
In summary, Table 5.12 is a representation of articles in the literature. As shown
in the table, the papers generally do not provide all of the information which is
needed for the reader to fully understand the experiment and to properly interpret
the results. The complexity of reporting nuclear battery research is due to the nature
of radiation transport. Each type of radiation has a scale length (λRadtr) associated
with it which is energy and material specific. The efficiency of the energy con-
version scheme is dependent upon the similarity of scale lengths of the radiation
source and the transducer (Ltrans). Inherently, the scale length of alpha and beta
voltaic cells based on a linearly graded p-n junction (on the order of 1 μm), typical
of the nuclear batteries described in the literature, is poorly matched to the scale
length of the radiation source. Even in a favorable scenario in which the radiation
source is an alpha particle, the scale length in a solid is on the order of 20 μm which
is not very good. As previously discussed, beta particles have a much less favorable
scale length match.
Table 5.12 Summary of completeness of information provided by a representative set of nuclear battery papers
Paper Isotope Activity Is radioisotope Are dimensions of How is power measured Pout Reported Edep
design junction or device (is signal to noise ratio efficiency fraction
information described given) ηab, ηint or ηdev
Sufficient? adequate?
Andreev et al. Tritium No Activity Only 1D No doping I-V analysis Max: 0.55 Expect 3 % No
[42] & info; geometry concentrations μ W/cm2 discussion
Fe-55 Pdep 10-15
(mW/cm2) but
no 3D info
Lu et al. [43] Ni-63 33 µCi/mm2 Yes Yes J-V data Not stated 0.32 % No
Flicker et al. Pm-147 6.3 Ci, 6.8 Ci Yes Yes J-V data Not explicitly 0.4 %, 0.77 % No
[44] stated, given in a
plot (less than
10 µW)
Li et al. [45] Ni-63 0.12 mCi Yes Yes J-V analysis 4.08 nW/cm2 1.01 % No
Clarkson et al. Tritium Only states No Yes J-V data Not stated. I-V 0.22 % No
[46] samples were plots given

exposed to
tritium gas
Zaijun et al. Ni-63 11 mCi No No J-V data Not stated, plots <1 % No
[47, 48] given
Deus [49] Tritium Only states No No J-V data 129 nW max 2.3 % (initial) No
samples were initial
exposed to
tritium gas
Hang and Lal Ni-63 0.25 mCi No Yes I-V analysis 0.24 nW NA No
[50] (beta) electroplated
Hang and Lal Ni-63 1 mCi No Yes I-V analysis 0.32 nW NA No
[50] (beta) Exposed to
277
encapsulated
source
(continued)
278
Cress et al. [51] Po-210 0.35 mCi No- Yes J-V data 0.0504 μW ηd = 36.3 % No
(foil ηdev = 3.2 % “
information
issue)
Cress et al. [51] Am-241 1.12 μCi No Yes J-V data 0.0068 nW ηd = 0.8 % No
(Foil ηdev = 0.04 %
information
issue)
Duggirala et al. Ni-63 9 mCi No- Yes Electrical output of 22 22 ηd = 5.10 % No
[52] (foil nW + 750 uW nW + 750 μW
information piezoelectric power at
issue) calls it a 0.07 % duty cycle
thin film
source
Kavetskiy et al. Pm-147 1.5 × 1010 Bq No- Yes (direct charge VI NA ηd = 4.2 % No
[1, 53] and (foil radioisotope calculated,
Galina information battery) ηd = 3.5 %
Yakubova [54] issue) experimental
Galina information battery) ηd = 8 %
Galina information battery) ηd = 15 %
5 Efficiency Limitations for Various Nuclear Battery Configurations
(continued)
Rybicki, Am-241 5 mCi No No VI 0.015 μ ηd = 18 %. No
Vargas-Aburto, (alpha) W/cm2 at 0 h
and Uribe [41] 0.0085 μ
W/cm2 at 100 h
Patel et al. [55] Cm-244 1 Ci Yes No dimensions theoretical calculations, 20 mW ηd = 57 %. No
(alpha) specified. no actual tests
Sychov et al. Pu-238 300 mCi No- No Measured power output 21 μ ηdev = 0.11 %. No
[56] (alpha) W
Qiao et al. [57] Ni-63 No No Yes Measured power output 2.04 nW/cm2 ηdev = 0.50 % No
Schottky barrier
diode 1.46 micron
depletion width
Qiao et al. [57] Am-241 Am-241: No 1.46 micron Measured power output Am: 1.25 ηdev = 0.10 % No
0.018 depletion width nW/cm2
mCi/cm2
A paper should provide information on the isotope used, the detailed design of the isotope encapsulation, details on how the source is interfaced to the device, detail
design information on the device, details information on how power out is measured, details on how efficiencies are calculated and was external bias applied to increase
the depletion width
279
5.6 Summary
Nuclear batteries have physical limitations which have been discussed throughout
this text. First and foremost the device efficiencies are limited by matching the
range of the ionizing radiation in the materials which the battery is constructed and
the scale length of the transducer. The fraction of power deposited in the transducer
is a significant efficiency limiting factor.
In Chap. 1, the idea of using the concepts of the minimum Battery surface Area
per Watt (BAWmin) and the minimum Battery Volume per Watt (BVWmin) are
introduced. The purpose for introducing these concepts is twofold. First, in dis-
cussions of nuclear battery technology there is need for a common factor by which
different designs can be compared, and size is the most intuitive. The second is that
the source of ionizing particles can either be coated on a transducer (surface source)
or embedded in a volume (volume source). BAWmin provides a comparator of
surface type sources for various designs. BVWmin provides a comparator of volume
type sources for various designs. The inverse of BAWmin is the maximum possible
Surface Power Density (SPDmax) for a specific radioisotope (W cm−2) and the
inverse of BVWmin is the maximum Volume Power Density (VPDmax) for a specific
radioisotope (W cm−3). The values of BAWmin and BVWmin are based on com-
pounds that contain the specific radioisotope in which the atomic density (atoms
cm−3) of the isotope is a maximum. Thus no nuclear battery can have scale lengths
smaller than BAWmin or BVWmin or power density values greater than SPDmax or
VPDmax. In Chap. 1 an example of a device which uses a tritiated amorphous
silicon nip drift junction was discussed. The device produced a surface power
density of 259 nW cm−2. The Battery surface Area per Watt (BAW) for this battery
is 3,861,000 cm2 W−1. BAWmin for tritium is 150,749 cm2 W−1. Thus
BAW/BAWmin = 25.61 which shows that there are both geometrical design
problems and transducer efficiency issues with the concept. Some other important
aspects of a tritium based nuclear battery can be seen in this analysis. First and
foremost even a perfect tritium based nuclear battery (meaning no power dilution
factor and 100 % transducer efficiency) will be huge 150,749 cm2 W−1.
The Kr-85 battery discussed in Sect. 5.4.3 is an example of a battery design
which uses a volume source. The BVWmin for Kr-85 is 0.8 cm3 W−1. The battery in
Sect. 5.4.3 has a shield that takes up a lot of the cell volume. The BVW of the cell
is 61.3 cm3 W−1.
The purpose of Chap. 3 is to provide insight into how power is transported from
the source to the transducer. Nuclear sources have isotropic emission, thus design of
interfaces between the source and transducer that maximize power deposition in the
transducer is complex. The ionizing radiation must escape the source material with
minimal self absorption and then must reach the transducer (possibly having to pass
through non power producing layers) and deposit as much power as possible. BVW
or BAW both increase substantially when the transport of ionizing radiation is taken
into account. When ionizing radiation transport is accounted for, true dimensions of
devices come to light. Even though nuclear battery literature is fraught with claims
5.6 Summary 281
of miniature sizes (it should be noted that the power levels of the systems in these
claims are very low), the conclusion drawn from Chap. 3 is that nuclear batteries
with useful power levels are not small.
Chapter 4 introduces the concept of the atomic dilution factor and the power
dilution factor. Based on the interface between the source and transducer, the
average number of radioisotope atoms in the volume will be diluted. The atomic
dilution causes the device size to increase. Power dilution is the ratio of the power
out of the cell to the maximum thermal power feasible from a compound with the
highest known density of radioisotope atoms. Surface source configurations will
have very large atomic dilution factors as well as large power dilution factors.
Volume sources can improve atomic dilution factors and power dilution factors.
The dilution factors provide the reader another analytic tool for evaluating the
feasibility of nuclear battery concepts and validity of claims that are made about a
new nuclear battery technology.
The efficiency of the transducer is the final consideration. Once power is
absorbed by the transducer, Chap. 5 shows how the efficiency of producing useful
power is found.
The lessons from these chapters can be used to develop a feel for and a vision of
nuclear battery design and their realistic dimensions and efficiencies.
Problems
1. Figure 5.1 is a pictorial representation of how power is transferred from the
source to the transducer. Explain how the various layers impact the transport
efficiency of ionizing radiation energy to the transducer.
2. What efficiencies dominate a p-n junction transducer? How does temperature
impact the efficiency of the junction?
3. What is the difference between photon and ionizing radiation interactions with a
p-n junction transducer? How do these differences impact transducer
efficiencies?
4. Recreate Fig. 5.6 using Pm-147.
5. Recreate Fig. 5.6 using Sr-90.
6. How does radiation damage impact p-n junction transducers? Direct Charge
Nuclear Batteries? Indirect PIDEC type batteries?
7. How do health and safety concerns impact nuclear batteries?
8. Compare the ideal efficiencies of a DCNB for θm = 15°, 30°, 45°, 60°, and 75°.
What angle yields the highest efficiency? Why?
9. A radiation worker accidentally receives 60 Rads of x-ray radiation to her
whole body, as well as 30 Rads of x-ray radiation to the gonads and 15 Rads of
alpha exposure to the skin. What is her effective dose? The lethal dose for 50 %
of the population (LD50) is approximately 400 Rem. Did she exceed this limit?
10. A literature report claims to have developed a nuclear battery using the
radioisotope Sr-90. According to the report, the authors have successfully
constructed a surface source based battery with a total efficiency of 15 %. Is

this feasible? Justify your answer.
11. If you could select the ideal properties for a radioisotope in a nuclear battery,
what activity, energy, type of decay, and half-life would you choose? Consider
aspects like radiation damage, lifetime of the battery, power, and size.
12. Compare system efficiencies for alphavoltaics, betavoltaics and PIDEC.
13. Discuss the main problems that are found in the nuclear battery literature.
References
1. Prelas MA (2016) Nuclear-pumped lasers. Springer International Publishing

review of nuclear batteries. Prog Nucl Energy 75:117–148
4. Oh K, Prelas MA, Lukosi ED, Rothenberger JB, Schott RJ, Weaver CL et al (2012) The
theoretical maximum efficiency for a linearly graded alphavoltaic nuclear battery. Nucl
Technol 179:7
5. Popovici G, Melnikov A, Varichenko VV, Sung T, Prelas MA, Wilson RG et al (1997)
Diamond ultraviolet photovoltaic cell obtained by lithium and boron doping. J Appl Phys
81:2429
6. Xu Z-H, Tang X-B, Hong L, Liu Y-P, Chen D (2015) Structural effects of ZnS:Cu phosphor
layers on beta radioluminescence nuclear battery. J Radioanal Nucl Chem 303:2313–2320
7. Liu Y, Tang X, Xu Z, Hong L, Chen D (2014) Experimental and theoretical investigation of
temperature effects on an interbedded betavoltaic employing epitaxial Si and bidirectional
63Ni. Appl Radiat Isot 94:152–157
8. Claeys C, Simoen E (2002) Radiation effects in advanced semiconductor materials and
devices. Springer
9. Ziegler JF, Ziegler MD, Biersack JP (2010) SRIM–The stopping and range of ions in matter.
Nucl Instrum Methods Phys Res, Sect B 268:1818–1823
10. Revankar ST, Adams TE (2014) Advances in betavoltaic power sources. J Energy Power
Sources 1:321–329
11. McFee RB, Leikin JB (2009) Death by polonium-210: lessons learned from the murder of
former Soviet spy Alexander Litvinenko. Semin Diagn Pathol 61–67
12. McAllister JFO, Burger TJ, Carsen J, Gumbel P, Israely J, Zarakhovich Y (2006) The spy who
knew too much. Time 168:30
13. John H, Rich L, David L, Alan P, Bobby S (2007) Polonium-210 as a poison. J Radiol Prot
27:17
14. Carter PJ (1977) Presidential Directive/NSC-25, scientific or technological experiments with
possible large-scale adverse environmental effects and launch of nuclear systems into space.
http://www.jimmycarterlibrary.gov/documents/pddirectives/pres_directive.phtml
15. Bennett GL (1981) Overview of the U.S. flight safety process for space nuclear power. Nucl
Safety 22:423–434
16. Standley S (2006) Cassini-Huygens engineering operations at Saturn, P. Jet Propulsion
Laboratory, CA, National Aeronautics and Space Administration
17. Rinehart GH (2001) Design characteristics and fabrication of radioisotope heat sources for
space missions. Prog Nucl Energy 39:305–319
18. Kern D, Scharton T (2004) NASA handbook for spacecraft structural dynamics testing. In:
Environmental testing for space programmes, vol 558, NASA, pp 375–384
References 283
19. Rosenberg KE, Johnson S (2006) Assembly and testing of a radioisotope power system for the
New Horizons Spacecraft. In: 4th International energy conversion engineering conference and
exhibit (IECEC). San Diego, California, p 2006
20. Martin L (1998) AGPHS RTGs in support of the cassini RTG program. Final Technical
Report RR18. Philadelphia, PA
21. Commission NR (2004) 10 CFR Part 71. Compatibility with IAEA transportation standards
(TS-R-1) and other transportation safety amendments, vol 67, pp 21390–21484
energy conversion. In: Yoshikawa M, Murakawa M, Tzeng Y, Yarbrough WA (eds) Second
international conference on the application of diamond films and related materials. MY Tokyo,
pp 5–12
24. Radprocalculator (2015). Gamma emitter point source dose-rate with shielding. http://www.
radprocalculator.com/Gamma.aspx
band gap electronic materials. Springer, pp 463–474
26. Prelas MA Title, unpublished|
27. Prelas M, Boody F, Zediker M (1984) A direct energy conversion technique based on an
aerosol core reactor concept. IEEE Publication, p 8
28. Prelas MA, Boody FP, Zediker MS (1985) An aerosol core nuclear reactor for space-based
high energy/power nuclear-pumped lasers. In: El-Genk MS, Hoover M (eds) Space nuclear
power systems, Orbit Book Company
29. Guoxiang G, Prelas MA, Kunze JF (1986) Studies of an aerosol core reactor/laser’s critical
properties. In: Hora H, Miley GH (eds) Laser interaction and related plasma phenomena,
Springer, pp 603–611
30. Prelas MA, Boody FP, Kunze JF (1986) A compact aerosol core reactor/laser fueled with
reflective micropellets. In: Hora H, Miley GH (eds) Lasers and related plasma phenomena, vol
7. Plenum Press, New York, NY
31. Prelas MA, Boody FP, Zediker MS (1984) Study of the basic transport properties (charged
particle transport, fluorescence transport & coupling efficiency) of the photon intermediate
direct energy conversion technique. Research Gate
32. Prelas M, Kunze J, Boody F (1986) A compact aerosol core reactor/laser fueled with reflective
micropellets. In Hora H, Miley G (eds) Laser interaction and related plasma phenomena.
Springer US, pp 143–154
33. Henry AF (1975) Nuclear-reactor analysis. MIT Press
35. Syed A (2012) Modeling the energy deposition of alpha particles emitted from Po-210 source
on silicon carbide for possible nuclear battery and laser pump applications, M.Sc., Nuclear
Science & Engineering Institute, University of Missouri—Columbia, MO
36. Kim H, Kwon JW (2014) Plasmon-assisted radiolytic energy conversion in aqueous solutions.
Nat Sci Rep 4
37. Buck EC, Wittman RS, Skomurski FN, Cantrell KJ, McNamara BK, Soderquist CZ (2012)
Radiolysis process modeling results for scenarios: used fuel disposition. Pacific Notrhwest
National Laboratory
38. Spinks JWT, Woods RJ (1990) An introduction to radiation chemistry. John-Wiley and Sons,
Inc., New York, Toronto
and betavoltaltaic cells, M.Sc., Nuclear Science & Engineering Institute, University of
Missouri—Columbia
40. Ashburn P, Morgan DV (1974) The role of radiation damage on the current-voltage
characteristics of pn junctions. Solid-State Electronics 17:689–698
41. Rybicki G, Vargas-Aburto C, Uribe R (1996) Silicon carbide alphavoltaic battery, pp 93–96
42. Andreev V, Kevetsky A, Kaiinovsky V, Khvostikov V, Larionov V, Rumyantsev V et al.
(2000) Tritium-powered betacells based on Al xGa 1-xAs. In: Photovoltaic specialists
conference. Conference record of the twenty-eighth IEEE, pp 1253–1256
43. Lu M, Zhang G-g, Fu K, Yu G-h, Su D, Hu J-f (2011) Gallium nitride schottky betavoltaic
nuclear batteries. Energy Convers Manage 52:1955–1958
44. Flicker H, Loferski J, Elleman T (1964) Construction of a promethium-147 atomic battery.
IEEE Trans Electron Devices 11:2–8
45. Li X-Y, Ren Y, Chen X-J, Qiao D-Y, Yuan W-Z (2011) 63Ni schottky barrier nuclear battery
of 4H-SiC. J Radioanal Nucl Chem 287:173–176
46. Clarkson J, Sun W, Hirschman K, Gadeken L, Fauchet P (2007) Betavoltaic and photovoltaic
energy conversion in three-dimensional macroporous silicon diodes. Physica status solidi A
204:1536–1540
47. Zaijun C, Haisheng S, Yanfei L, Xuyuan C (2010) The design optimization for GaN-based
betavoltaic microbattery. In: 2010 5th IEEE international conference on Nano/micro
engineered and molecular systems (NEMS), pp 582–586
48. Cheng Z, San H, Li Y, Chen X (2010) The design optimization for GaN-based betavoltaic
microbattery. In: 2010 5th IEEE international conference on nano/micro engineered and
molecular systems (NEMS), pp 582–586
49. Deus S (2000) Tritium-powered betavoltaic cells based on amorphous silicon . In: Conference
record of the twenty-eighth IEEE, photovoltaic specialists conference, 2000, pp 1246–1249
50. Hang G, Lal A (2003) Nanopower betavoltaic microbatteries. In: 12th international conference
on transducers, solid-state sensors, actuators and microsystems 2003, vol 1, pp 36–39
51. Cress CD, Landi B, Raffaelle RP, Wilt DM (2006) InGaP alpha voltaic batteries: synthesis,
modeling, and radiation tolerance. J Appl Phys 100:114519–114700
52. Duggirala R, Li H, Lal A (2008) High efficiency radioisotope energy conversion using
reciprocating electromechanical converters with integrated betavoltaics. Appl Phys Lett 92
53. Kavetskiy A, Yakubova G, Yousaf SM, Bower K, Robertson JD, Garnov A (2011) Efficiency
of Pm-147 direct charge radioisotope battery. Appl Radiat Isot 69:744–748
54. Galina Yakubova AK (2012) Nuclear batteries with tritium and promethium-147 radioactive
sources: design, efficiency, application of tritium and pm-147 direct charge batteries, tritium
battery with solid dielectric. LAP Lambert Academic Publishing
55. Patel JU, Fleurial J-P, Snyder GJ (2006) Alpha-voltaic sources using liquid ga as conversion
medium. NASA Tech Briefs, NASA
56. Sychov M, Kavetsky A, Yakubova G, Walter G, Yousaf S, Lin Q et al. (2008) Alpha indirect
conversion radioisotope power source. Appl Radiat Isot 66:173–177
57. Qiao DY, Chen XJ, Ren Y, Yuan WZ (2011) A micro nuclear battery based on Sic schottky
barrier diode. J Microelectromech Syst 20:685–690
Chapter 6
Potential Applications for Nuclear
Batteries
Abstract The obvious attributes of radioisotopes are a large amount of energy

stored per unit mass and long life-time. These properties have been the primary
motivation in the development of nuclear batteries driven with radioisotopes. The
applications have changed over time based on the evolution of technology. In this
chapter, successful nuclear battery designs are discussed along with the applications
that they were designed for. In addition, the technologies which have evolved over
time requiring significant energy storage and long shelf lives are discussed.
Keywords Nuclear Battery Applications Military Space Health Fission

Reactors
6.1 Successful Applications
The commercial application of radioisotope powered nuclear batteries date back

over 60 years [1–3], with labs such as RCA engaged in the research. A Direct
Charge Nuclear Battery (DCNB) developed by Sandia in the 1960s produced a high
operating voltage (20 kV) with currents ranging from 1 × 10−10 to 1 × 10−9 amps
[4] using Kr-85 (Fig. 6.1). A 1.6 cm3 glass bulb was created by glass blowing.
The wall thickness of the Corning 7053 glass was 0.015 cm. The glass bulb was
filled with 80 psi of Kr gas containing 5 % Kr-85. Thus, the total activity of Kr-85
was 0.8 Ci. Several layers of graphite were coated on the outside of the bulb
to a thickness of 0.1 cm and painted with silver. A 0.05 cm thick layer of Ni was
then plated on the silver paint. This structure served as the collector. About 30 %
of the beta particles penetrated the glass wall and were collected on the graphite
collector.

286 6 Potential Applications for Nuclear Batteries
Fig. 6.1 A DCNB based on Kr-85 developed at Sandia in the 1960 [4]. This figure was reprinted
with permission from IEEE
6.1.1 Pacemakers
Several nuclear batteries were developed for cardiac pacemakers in the 1970s [5, 6].
Thermoelectric and betavoltaic transducers were used in the pacemaker nuclear
battery power source (see Table 6.1). Miniature thermoelectric nuclear batteries
require thermal insulation to minimize the heat loss as well as shielding to minimize
radiation exposure and effective encapsulation. As shown in Table 6.1, several
thermoelectric nuclear batteries for pacemakers were clinically tested. In Fig. 6.2, a
thermoelectric battery used in a Medtronic pacemaker is shown. The
McDonnell-Douglas Betacel 400 (Fig. 6.3) was based on stacked layers of silicon
betavoltaic cells with surface coupling to Pm-147. The design used both sides of the
Pm-147 layer and produced 400 μW of electrical power at a reported efficiency of
4 %. The Betacel 400 was used in clinical trials of pacemakers (Fig. 6.4). However,
after lithium batteries were developed, the market for nuclear batteries dried up.
6.1.2 Deep Space Probes
For space exploration beyond Mars, the available solar flux is not sufficient to
power probes with solar cells. To this end, the Radioisotope Thermal Generator
(RTG) was used with great success. As shown in Table 3.17, there were 26 mis-
sions which have used 45 RTG’s for power. RTGs are not particularly efficient,
only reaching 4–10 % efficiency, but they have been proven to be reliable and have
been extensively flight tested. The voyager probes, for instance, were launched
nearly 40 years ago and are still sending signals. In addition to electrical power,
heat generation is also used to keep various components of the probes heated.
Key to RTG technology is the isotope Pu-238. The National Research Council
concluded in its study of potential isotopes for nuclear batteries that the supply of
6.1 Successful Applications 287
Table 6.1 Summary of nuclear pacemaker clinical data [6]

Manufactured Nation Fuel Loading-gm Clinical Comments
cases
ARCO US Pu-238 0.410 42 Thermoelectric
Oxide
AtomCell (nuclear US Pu-238 0.120 Animal Thermoelectric
battery corp) Oxide trials
Atomic energy UK Pu-238 0.180 80 Thermoelectric
research Oxide
establishment
McDonnel Douglas US Pm-147 0.105 60 Betavoltaic
(BetaCel 400) Oxide
Gulf energy and env. US Pu-238 0.180 Animal Thermoelectric
sys. Oxide trials
CIT Alcatel France Pu-238/Sc 0.160 400 Thermoelectric
Alloy (first pacemaker)
Siemens AG W. Pu-238 0.200 None Thermoelectric
Germany Oxide
Syncal US Pu-238 None Thermoelectric
Oxide
radioisotopes is very limited. Only Pu-238 was deemed viable for NASA missions
[7]. Pu-238 is created from the 237Np(n, γ)238Np → 238Pu + β reaction. Neptunium
237 is produced in both commercial and military reactors by reactions with uranium
238 (238U(n,2n)237U → 237Np + β) and uranium 235 (235U (n, γ)236U (n,
γ)237U → 237Np + β). The available separated inventory of Pu-238 within the
National Laboratories in the United States is about 39 kg and the estimated Np-237
inventory is 300 kg [8]. It is feasible to produce Pu-238 from the separated
inventory of Np-237 using high neutron flux reactors such as the Advanced Test
Reactor at Idaho National Laboratory or the High Flux Isotope Reactor at Oak
Ridge National Laboratory [7, 9]. Doing so would produce about 5 kg of Pu-238
per year. Unfortunately, t [7, 9] to produce about 5 kg of Pu-238 per year. The cost
of building the production capability is estimated to be 77 million dollars US [7, 8].
The cost of Pu-238 per kg is about 8 million dollars US. The inventory of
unseparated Np-237 worldwide from commercial reactor spent fuel is estimated to
be 54,000 kg and from military reactors 1655 kg [10].
The European space program identified Am-241 as its isotope of choice. It can
be produced economically from the civilian fuel reprocessing program [11, 12].
Am-241 has a longer half-life than Pu-238 (433 years vs. 88 years) and thus more
mass is needed to produce the same amount of activity.
Fig. 6.2 A long lived thermoelectric nuclear battery fueled with Pu-238 oxide was used in a
Medtronic pacemaker device
Fig. 6.3 An illustration of the Betacel 400 (a 400 mW nuclear battery) which was used in cardiac
pacemakers in the 1970s [5] (printed with permission from Physics Today)
Fig. 6.4 A pacemaker made

by CCC del Uruguay powered
with a BetaCel 400 power
supply
6.1.3 Curiosity Rover
All previous rovers used solar cells and batteries for electrical power production.
These missions were also fairly small. The Curiosity Rover is as large as an SUV
and it contains advanced\analysis tools such as gas chromatography equipment.
A Multi Mission Radioisotope Thermoelectric Generator (MMRT) was used for the
mission and is about 6–7 % efficient. It carries 4.8 kg of Pu-238 dioxide fuel which
used eight General Purpose Heat Source (GPHS) modules. This configuration
produces 110 W of electrical power as well as heat. The source’s heat allows the
device to operate during the winter season. Also, the GPHS is able to provide heat
to maintain effective operating temperatures for the mission’s instruments and
systems.
6.1.4 Remote Power Applications
The Russians have used RTGs containing beta emitters such as Sr-90 for remote
RTG power sources for lighthouses and land-based navigation sites throughout
Russia [13]. It is believed that over 1000 RTG units were deployed in Russia.
Unfortunately, many of them were abandoned, and several have been found by
scavengers. A decommissioning process has been underway since 2002.
6.1.5 Other
While not technically nuclear batteries, radioisotope powered lighting sources have
been used for years [14]. Their most common arrangement is to mix a phosphor with a
radioisotope to produce visible light. Radium, for example, has been mixed with zinc
sulfide and a binder to produce a luminescent paint that was used on self-illuminated
clock dials (Fig. 6.5). Due to the hazards of using radium, tritium eventually replaced
it. Tritium has been used in gaseous form or has been bound to styrene to produce
luminescent paints. The tritium paints were used by the military for aircraft lumi-
nescent dials, mine field markers, and gunsight illumination. Tritium paints have been
used by industry for exit signs and lighting for aircraft landing strips An example of a
commercially available tritium light stick is shown in Fig. 6.6.
At one time, Sr-90 was used along with phosphors to produce luminescent watch
dials. A few watches with Sr-90 watch dials were produced in Europe and shipped
to the US. In addition, some dock marker lights were made with Sr-90. However,
due to the hazards of Sr-90, its use was discontinued.
Fig. 6.5 Luminescent watch

dial which uses tritium paint
Fig. 6.6 A 3 × 22.5 mm

tritium vial shown in daylight
(a) and total darkness (b). The
vial is filled with tritium paint
Pm-147 is a fission produced isotope which emits a low energy beta. Its half-life
is 2.5 years and it easily shielded. The isotope was adopted for use in the Apollo
Space Program to make markers which outlined the docking mechanism on the
moon orbiter and the moon lander.
Tritium and Kr-85 are gaseous isotopes which can be used to excite phosphors.
Tritium excited phosphors have been used to produce large taxiways or runway
distance at airports. They have also been used in exit signs in public buildings.
Kr-85 was used to make experimental luminescent markers as well.
6.2 Military Missions
There is a need for portable power supplies for various military missions. The
battlefield has grown more complex with new and better technologies including, but
not limited to: GPS, night vision, radios, smartphones, tablets, helmet mounted
screens, imaging devices, laser guided weapons, arrays of sensors, drones, robot
mules, and wireless devices for a digital battlefield. The common theme in these
advanced weapons is the need for electrical power. A modern solder is well
equipped to handle these electrical needs but is overburdened as a result. For
example, a modern solder must pack 16 lb of batteries to power the technology
which he/she carries. Even though systems are being developed such as vest power
managers and improved electronics efficiency, as technology advances, power will
always be the limiting factor.
6.2.1 Light Weight Portable Nuclear Batteries
The dream of having batteries which are light, portable and long lasting has driven
some interest in nuclear batteries for the battlefield. Unfortunately, dream and
reality are two different things. Chapter 1 discusses the size limitations of surface
sources and volume sources. In order for a nuclear battery to produce power levels
in the mW to 10s of watts range, the size of the battery becomes unmanageable.
Table 1.3 is reproduced here for the purposes of evaluating the scale of a nuclear
battery (Table 6.2). The smallest nuclear battery scale size will use a volume
source. The last column of Table 6.2 presents the scale size per watt of those
compounds which contain the maximum atomic density of the radioisotope. This
value does not take into account the interface of the isotope to the transducer or the
transducer efficiency. In order to introduce the transducer interface to the nuclear
battery, the reader can use the lessons from Chap. 4 to determine the appropriate
dilution factor. A further increase in size will be due to inefficiencies of the energy
conversion process as discussed in Chap. 5. Thus, the true size of the nuclear
battery will be BVMmax multiplied by the dilution factor and divided by the system
efficiency. As can be seen in discussions from Chaps. 4 and 5, the dilution factor
can be on the order of 100–10,000 and the typical system efficiencies on the order
of 0.1–5 %. This means that the dilution factor divided by the system efficiency can
range anywhere from 2000 to 10,000,000 depending on the nuclear battery design.
Thus BVM will be 2000 to 10,000,000 times BVMmin: a much larger package
indeed.
There are other factors that impact the size of the nuclear battery besides BVM.
The most important is shielding. As shown in Table 1.1, isotopes can also emit
gamma rays. In Chap. 5, an example of a Kr-85 PIDEC nuclear battery is discussed.
Even though Kr-85 is one of the safest isotopes, it also emits a 0.514 MeV gamma
ray in 0.4 % of the decays. This relatively low probability gamma ray emission
event still requires that an 8 cm lead shield be incorporated into the design to
protect personnel and electronics. The lead shield contributes the majority of the
mass to the battery design. Thus, if there are secondary gamma emissions for the
chosen isotope, a radiation shield will also increase BVM and the mass of the
battery.
Table 6.2 The theoretical absolute minimum nuclear battery surface area per Watt (BAWmin) and the theoretical absolute minimum nuclear battery volume
per Watt (BVWmin) for various radioisotopes are shown
Nuclide Decay energy Half Decay βmax (MeV) Power W/gm Compound Density Range BAWmin BVWmin
(MeV) life (yr) gm/cm3 (μm) (cm2/W) (cm3/W)
H-3 0.0186101861 12.4343 Beta 0.0186101861 0.323538323538 T2O 1.215 00.61875 150,150,749 99.32759
Ar-39 0.565 269 Beta 0.565 0.037108845 Liquid Ar 1.4 85.5 2251 19.25
Ar-42 0.6 32.9 Beta 0.6 0.306735936 Liquid Ar 1.4 107.2 217.4 2.33
Co-60 2.824 5.2713 Beta and γ 0.318 0.710417655 Metal 8.9 3.12 506.4 0.158
Kr-85 0.67 10.755 Beta. 0.67 MeV 0.517808194 Liquid Kr 2.413 49.45 161.8 0.80
99.6 %,
0.15 MeV
0.4 %
Sr-90 0.546 28.77 Beta 0.546 0.148981453 Metal 2.64 27 and 234 108.7 2.543
Ru-106 0.039 1.0234 Beta 0.039 0.25396244 Metal 1.53 0.4715 54,590 2.574
Cd-113 m 0.58 14.1 Beta 0.58 0.257142774 Metal 8.69 6.8 658.1 0.4475
Sb-125 0.767 2.73 Beta 0.302 MeV 0.478708387 Metal 6.684 3.333 and 282.8 0.3125
(40 %), 11.05
0.622 MeV
(14 %) and
0.13 MeV
(18 %)
Cs-134 2.058 2.061 Beta 0.662 MeV 1.265751647 Metal 1.93 33.15 and 123.5 0.4093
71 %, 1.615
0.089 MeV
28 %
Cs-137 1.175 30.1 Beta 1.176 MeV 0.095403085 Metal 1.93 88 and 617.2. 5431
6.5 %, 26.35
0.514 MeV
93.5
Pm-146 1.542 5.52 ec (66 %). 0.795 0.23685592 Metal 7.26 9.8 593.4 0.5815
Beta (34 %)
(continued)
293
294
Pm-147 0.225 2.624 beta 0.225 0.411934768 Metal 7.26 1.2 2787 0.3344
Sm-151 0.076 90 beta 0.076 0.003949124 Metal 7.54 0.216 1555,000 33.58
Eu-152 1.822 13.54 ec (72.1 %), 0.696 (13.6 %), 0.174034676 Metal 5.259 12.65 and 255.7 1.093
beta (27.9 %) 1.457 (8.4 %), 42.75
385 (2.5 %),
0.176 (1.8 %)
Eu-154 1.969 8.592 Beta 1.845 (10 %), 0.247022865 Metal 5.259 60.3 and 127.7 0.7698
(99.98 %), ec 0.571 (36.3 %) 1.92 and
(0.02 %) and 0.249 8.8
(28.59 %)
Eu-155 0.253 4.67 Beta 0.147 (47.5 %), 0.16702465 Metal 5.259 0.672 and 5932 1.139
0.166 (25 %), 1.275 and
0.192 (8 %) 1.92 and
and 0.253 0.99and
(17.6 %) 0.7905
Gd-148 3.182 74.6 Alpha 0.610572936 Metal 7.90 8.44 245.6 0.2073
Tm-171 0.096 1.92 Beta 0.0964 (98 %) 0.204439908 Metal 9.321 0.27 and 19,440 0.5248
and 0.0297 0.045
(2 %)
Os-194 0.097 6 Beta 0.0143 0.039759437 Metal 22.57 0.065 and 1675,000 12.06
(0.12 %), 0.072
0.0535 (76 %)
and 0.0966
(24 %)
Tl-204 0.763 3.78 Beta (97.1 %), 0.763 0.678193708 Metal 11.85 5.4 230.4 0.1244
ec (2.90 %)
Pb-210 0.063 22.29 Beta (100 %), 0.0169 (84 %) 2.4035 Metal 11.342 0.02 and 113.3 0.03668
alpha and 0.0635 0.117 and
(1.9 × 10−6 %) (16 %) 18.7
(continued)
6 Potential Applications for Nuclear Batteries
Po-208 5.216 2.8979 Alpha 17.96948512 Metal 9.32 18.2 3.2 0.005971
(99.9958 %),
ec (0.0042 %)
Po-210 5.305 0.379 Alpha 141.1431705 Metal 9.32 18.7 0.41 0.0007602
(100 %), γ
(0.0011 %)
Ra-228 0.046 5.75 beta 0.0128 (30 %), 0.015405793 Metal 5.5 0.12 and 916,700 11.00
0.0257 (20 %), 0.035
0.0392 (40 %), and0.0925
0.0396 (10 %)
Ac-227 0.044 21.773 Beta (98.6 %), 0.02 (10 %), 0.31183 Metal 10.07 0.07 and 442.9 3.207
alpha (1.38 %) 0.0355 (35 %) 0.05 and
and 0.0448 0.025
(54 %)
Th-228 5.52 1.9131 alpha 26.05395081 Metal 11.72 15.4 2.12 0.003275
U-232 5.414 68.9 alpha 0.700087964 Metal 18.95 9.55 78.9 0.07538
Pu-236 5.867 2.857 Alpha 18.03219635 Metal 19.86 8.33 3.35 0.002792
(100 %), fis
(1.3E-7 %)
Pu-238 5.593 87.74 Alpha 0.5555868 Metal 19.86 7.71 117.6 0.09064
(100 %), fis
(1.85E-7 %)
Pu-241 0.021 14.35 Beta 0.02082 0.0112 Metal 19.86 0.015 376 4.48
(99.998 %),
alpha
(0.00245)
Am-241 5.638 432.2 Alpha 0.108572645 Metal 13.69 11.2 600.7 0.6728
(100 %), fis
(4.3E-10 %)
(continued)
295
296
Cm-243 6.168 29.1 Alpha 1.647617418 Metal 13.52 12.8 35.1 0.04489
(99.71 %), ec
(0.29 %), fis
(5.3E-9 %)
Cm244 5.902 18.1 Alpha 2.777536879 Metal 13.52 12 22.2 0.02663
(100 %), fis
(1.37E-4 %)
Bk248 5.793 9 alpha 5.49387705 Metal 14.78 14.5 8.5 0.01232
Cf250 6.128 13.07 Alpha 3.890798482 Metal 15.1 15.5 11 0.01702
(99.923 %), fis
(0.0775 %)
Cf252 6.217 2.645 Alpha 18.78859265 Metal 15.1 15.5 2.27 0.003525
(96.908 %),fis
(3.092 %)
Es-252 6.739 1.292 Alpha 30.86044068 Metal 5.24 16.5 3.75 0.006184
(76.4 %), ec
(24.2 %), beta
(0.01 %)
The inverse of BAWmin is the theoretical maximum surface power density and the inverse of BVWmin is the theoretical maximum power density. In calculating these parameters it is assumed
that the compound containing the radioisotope has the maximum atomic density possible for the isotope, the transport properties of the ionizing radiation into the transducer are perfect, the
efficiency of the transducer is 100 %, and for a volume source the transducer and radioisotope interface is ideal and seamless
6.2 Military Missions 297
Another consideration which is discussed in Chap. 5 is the radiation safety

concerns for the isotope. Isotopes which are bone seekers (such as the actinides,
heavy metals, and Sr-90) are particularly dangerous. This leads to potential radia-
tion hazards that might occur if the nuclear battery is punctured by a bullet or an
explosion.
A final consideration is the cost of the radioisotope as discussed in Chap. 2. The
cost of the radioisotope is going to be less if it is available from spent nuclear fuel.
Still, the cost of spent fuel reprocessing is high. If the radioisotope has to be
produced in a reactor using neutron capture or produced with an accelerator, the
costs will be very high. The cost of the radioisotope can easily be the deal breaker
for any application.
It is possible to have a truck mounted power supply where batteries can be
recharged on the battlefield. This would have to be a fairly large power supply.
However, it is possible that the Kr-85 battery described in Chap. 5 would be viable.
It can produce enough power to be of interest for charging a large quantity of
batteries; it is light enough to be mounted on a vehicle in the battlefield; and it uses
a safe isotope so if the battery is punctured, it would not pose a significant risk to
personnel on the battlefield. The problem with Kr-85 is its availability. Kr-85 is
produced from fission with a yield of 0.2717 %. With a 3000 MWth the Kr-85
production rate is 1.193 × 1018 Kr-85 atoms/sec. This is a yearly production rate of
3.762 × 1025 Kr-85 atoms/yr. Thus the amount of Kr-85 stored in the core or
holding tanks of an operating commercial reactor is about 2.07 MCi/yr (Table 6.3).
There is Kr-85 inventory stored in the spent fuel of nuclear power plants which can
be estimated. First, the equilibrium value in the spent fuel over a long period of time.
needs to be analyzed. The equilibrium value for Kr-85 atoms is calculated by dividing
the radioisotope Source Rate by the Rate Constant. For a 3000 MWth plant, the Source
Rate/Rate Constant = 1.193 × 1018/2.04210−9 = 5.84 × 1026. Thus, 5.84 × 1026
Kr-85 atoms are stored in the spent fuel of a 3000 MWth commercial power plant. The
equilibrium activity is then, the number of Kr atoms * Rate Constant/3.7 × 1010
d/s-Ci = 5.84 × 1026*2.042 × 10−9 s−1/3.7 × 1010 d/s-Ci * 32 × 106 Ci per 3000
Table 6.3 Operating parameters for commercial nuclear power plants

Manufacturer MWt MWe Loops Pressurizer Reactor coolant Steam
pumps per loop generators
per loop
Westinghouse 450–3000 167–1000 1–4 1 1 1
Framatome 2700–3600 900–1300 3–4 1 1 1
Babcock and 2400–3000 800–1000 2 1 2 1
Wilcox
Combustion 2400–3600 800–1300 2 1 2 1
Engineering
ABB 3000 1000 4 1 1 1
Mitsubishi 3000 1000 4 1 1 1
MWth plant. The world’s total installed nuclear generated power is 380,000 MWe
which is about 1,140,000 MWth. The total Kr-85 inventory in the operating nuclear
power plants of the world is approximately 32 MCi*1,140,000/3000 MWth * 12,160
MCi of Kr-85.
6.2.2 MEMS
The current generation of sensors typically requires hundreds of milliwatts to

operate. Sensors do have a theoretical operating power limit between microwatts to
milliwatts and advanced sensors may eventually reach this limit. Sensors are used in
a multitude of military applications in the battlespace. Another tool in the bat-
tlespace is the cell phone. The power requirements of a cell phone are several Watts
of RF power in order to operate.
6.3 MEMS
MicroElectroMechanical Systems (MEMS, which are devices between 0.02 and

1 mm) and NanoElectroMechanical Systems (NEMS, which are less than
0.02 mm) are micro machines or nano machines that can perform a variety of
functions. Examples include accelerometers, gyroscopes, optical switches, pressure
sensors, piezoelectric sensors, bio-MEMS, etc. The scale of MEMS or NEMS is so
small that nuclear battery scale lengths do not match up well.
It is understandable that researchers dream of developing nuclear batteries for
small scale sensors as a permanent source of power. Given the potential shelf life of
such a battery, it could be operated continuously without maintenance for years.
The problem is that the ideal does not match up well with the physics of nuclear
battery scaling in reality.
6.3.1 Drones
Unmanned Aerial Vehicles (UAVs) have found many applications (both military
and civilian) and can be flown by a pilot on the ground, autonomously based on a
pre-programmed flight plan, or by a complex dynamic automation system [15].
These are relatively massive vehicles and have substantial power requirements. The
concept of Micro Aerial Vehicles (MAV) and Nano Areal Vehicles (NAV) shrinks
the size and mass of the unmanned aerial vehicle, thus reducing power requirements
as well as other benefits [16]. MAVs are small air vehicles that can operate over a
range of 10 km with a flight time of 1 h and a maximum takeoff weight of 5 kg.
NAVs on the other hand are air vehicles that can operate over a range of 1 km at an
6.3 MEMS 299
altitude of 100 m with a flight time of 1 h and a maximum takeoff weight of 25 g.

With a light weight and a long lived power source, the operational time of the
MAVs and NAVs could substantially increase. A 15 g NAV requires about 585
mW to power the propeller, about 500 mW to power an RF communication device
and 50 mW to power a camera. The total battery power would have to be on the
order of 1.135 W. The dream is that since radioisotopes can store as much as
2 × 109 J/gm (as compared to 460 J/gm for a Li-ion battery), a nuclear battery
should be an ideal power supply. This dream is challenged because radioisotopes
can only produce power at a level that is dependent on the half-life of the isotope
and this rate is constant and is not very high (as compared to a Li-ion battery or a
storage capacitor). As discussed in prior chapters, this is a substantial challenge for
a nuclear battery. It would require on the order of 5000–10,000 Ci of a radioisotope
(depending on the isotope and nuclear battery design). This shear amount of
radioisotope has significant safety concerns. If the radioisotope emits gamma rays,
it should make the NAV detectable using simple radiation detectors. Also, the size
and mass of a 1 W nuclear battery would be far larger than the 15 g take off mass of
the postulated NAV.
There may be tradeoffs for certain applications where the size, mass and radi-
ation safety issues are acceptable. One example is deep space missions as described
previously. However, there are potential terrestrial applications. For example, the
SNAP-7 RTG series nuclear batteries were developed for a various marine related
needs such as deep sea probes, buoys, etc. With advanced technology, autonomous
robotic probes powered with a long lived power source may be feasible.
6.3.2 Nano-Power Systems
There is an ongoing effort to develop “Nano-Power Systems” to produce nano-sized

power system components. Radioisotope powered batteries are of specific interest.
The goal of this effort is to develop power systems that have the following
characteristics:
• A power density of 1–100 mW/cm3
• A device lifetime of 10 years or greater
• Power production density of 1000 We/kg
In order for a nuclear battery to meet all three conditions, the radioisotope must
have a half-life on the order of 10 years. The isotopes listed in Table 6.2 can be
paired down by eliminating those isotopes which do not have a long enough
half-life and/or produce gamma rays which would require shielding which would
make condition 3 impossible to meet (Table 6.4). The maximum power density
(Pdmax) listed in this table is for a one gram mass of the compound which has the
maximum atomic density of the radioisotope possible. To meet condition 1, in
order to design a nuclear battery transducer interface, the radioisotope will have to
be diluted by the atomic density dilution factor (DFatomic) of the nuclear battery
Table 6.4 A list of isotopes that have a half-life around 10 years or greater (condition 2) and do not emit secondary gamma rays (from Table 1.1) is shown
300
Nuclide Decay Half Decay Power W/gm BVWmin Max Minimum Minimum
energy life Has to be > 1 (cm3/W) Power Value Value
(MeV) (yr) Density ηsystem/ ηsystem/
(W/cm3) DFatomic DFatomic
1 mW/cm3 100
mW/cm3
H-3 0.01861 12.43 Beta 0.32353832353 9.3275932759 0.3931393 2.28 × 10−3 2.28 × 10−1
Ar-39 0.565 269 Beta 0.037108845 19.25 0.052 1.92 × 10−2 1.92
Ar-42 0.6 32.9 Beta 0.306735936 2.33 0.43 2.33 × 10−3 2.33 × 10−1
Sr-90 0.546 28.77 Beta 0.148981453 2.543 0.39 2.54 × 10−3 2.54 × 10−1
Cd-113 m 0.58 14.1 Beta 0.257142774 0.4475 2.23 4.84 × 10−4 4.84 × 10−2
Sm-151 0.076 90 Beta 0.003949124 33.58 0.03 3.36 × 10−2 3.36
Gd-148 3.182 74.6 alpha 0.610572936 0.2073 4.82 2.07 × 10−4 2.07 × 10−2
Pb-210 0.063 22.29 Beta (100 %), alpha (1.9 × 10−6 %) 2.4035348 0.037 27.26 3.67 × 10−5 3.67 × 10−3
plus 5.305 MeV alpha from Po-210
Ac-227 0.044 21.773 Beta (98.6 %), alpha (1.38 %) 0.031 3.2 0.312 3.21 × 10−3 3.21 × 10−1
U-232 5.414 68.9 alpha 0.700087964 0.07538 13.27 7.54 × 10−5 7.54 × 10−3
Pu-238 5.593 87.74 Alpha (100 %), fis (1.85E-7 %) 0.5556 0.09064 11.03 9.06 × 10−5 9.06 × 10−3
Pu-241 0.021 14.35 Beta (99.998 %), alpha (0.00245) 0.0112 0.000564 0.2233 4.48 × 10−3 4.48 × 10−1
Am-241 5.638 432.2 Alpha (100 %), fis (4.3E-10 %) 01086 0.6728 1.49 6.73 × 10−4 6.73 × 10−2
Cm-243 6.168 29.1 Alpha (99.71 %), ec (0.29 %), fis 1.647617418 0.04489 22.28 4.49 × 10−5 4.49 × 10−3
(5.3E-9 %)
Cm244 5.902 18.1 Alpha (100 %), fis (1.37E-4 %) 2.777 0.02663 37.55 4.49 × 10−5 4.49 × 10−3
Bk248 5.793 9 alpha 5.49387705 0.01232 81.17 1.23 × 10−5 1.23 × 10−3
Cf250 6.128 13.07 Alpha (99.923 %), fis (0.0775 %) 3.890798482 0.01702 58.75 1.7 × 10−5 1.7x103
Those isotope which are not able to produce 1 W/gm or greater are shown the table because of condition 3
6.3 MEMS 301
design. It is also understood that the output power density will be reduced by the
system efficiency (ηsystem) of nuclear battery design. Thus it is possible to develop a
relationship between the desired power density (Pd1mw/cc or Pd100mW/cc) in condi-
tion 1 and Pdmax. This relationship is bounded as shown in Eq. 6.1.
Pdmax gsystem
0:1 0:001 ð6:1Þ
DFatomic
The upper limit and lower limit for ηsystem/DFatomic is shown in the last two
columns of Table 6.4. The importance of this relationship is that conditions 2 and 3
are determined by the properties of the isotope. Condition 1 is determined by the
design factors of the nuclear battery.
Condition 2 requires that the isotope have a reasonable half-life (about
10 years). Condition 3 implies that the mass of the nuclear battery has to be
minimized. In order to minimize the mass of a nuclear battery, isotopes which
produce gamma rays that require heavy shielding are not included in Table 6.4.
Condition 3 also requires that the isotope produce sufficient power density to
exceed 1 We/g. Any isotope which can’t even produce 1 W/g is shown in Table 6.4.
This leaves very few viable isotopes as seen in Table 6.4.
It is easier to see the significance of equation one by using the radioisotope
Pb-210 as an example. Taking the parameters from Table 6.4 for Pb-210, Eq. 6.1
becomes,
gsystem
0:1 27:26 0:001 ð6:2Þ
DFatomic
Or
gsystem
3:67 103 3:67 105 ð6:3Þ
DFatomic
where the smaller the values the better since a small value means the system can
have a higher dilution factor and a lower system efficiencies and still meet the
criteria.
Equation 6.3 can be evaluated by making some simple assumptions (based on
discussions from Chap. 4). If DFatomic = 1000, then Eq. 6.3 becomes: 3.
67 ≥ ηsystem ≥ 3.67 × 10−2. The upper limit is simply not possible to achieve since no
system efficiency can exceed 1. The lower limit of 3.67 × 10−2 is achievable since this
is the approximate value for the maximum system efficiency of some types of nuclear
battery designs (see Chap. 5). So it may be possible to meet the 1 mW/cm3 power
density in condition 1. However, the volume of the nuclear battery would have to be
1,000,000 cm3 in order to produce 1,000 We and it must weigh 1 kg. In order to meet
condition 3, the average mass density of the device would have to be about 1 µg-cm−3.
A nuclear battery design with this low of an average mass density is not feasible.
The requirements for nano-power systems are stringent. As shown in Table 6.4,
only a handful of alpha emitters can even come close to meeting these requirements.
In order for these isotopes to be used in nuclear battery designs that have a chance
of meeting the requirements, the nuclear battery would have to use a volume type
design, the atomic dilution factor would have to be low, and the system efficiency
high. These are ambitious goals considering the limitations of radioisotope/
transducer interfaces, the availability of isotopes, and the cost.
6.3.3 Fission Reactors
One of the major problems with using radioisotopes is that the power production is
constant and is based upon the half-life of the isotope. Another major problem with
radioisotopes is their limited power density. These were recognized early on as
limitations and the United States developed both Systems Nuclear Auxiliary Power
(SNAP) radioisotope thermoelectric generators and space nuclear reactors which
were flown by NASA in the 1960s. The odd numbered SNAPs were radioisotope
based and the even numbered SNAPs were nuclear reactors. The first of the reactors
was called the SNAP Experimental Reactor (SER). It utilized uranium zirconium
hydride fuel and a eutectic sodium-potassium alloy (NAK) as the coolant. It pro-
duced 50 kW thermal but did not have a transducer for the conversion of heat to
electricity. It achieved criticality in September 1959 and operate until December
1961. The basic concept of SER was used in SNAP-2 which was a 55 KWth heat
source, designed and built by Atomics International Division, North American
Aviation, Inc., to power a mercury Rankine cycle to produce 3.5 KWe [17, 18].
Soon after SNAP 2 was built, the development of SNAP 8 began as a joint project
between the Atomic Energy Commission and NASA. It was a 600 kWth reactor
which operated between 1963 and 1965. SNAP 10 A was the first reactor flown in
space. The thermal power level was 35 kW with an electrical power output of
0.5 kW. It used a thermoelectric transducer which was coupled to a hydride core
cooled with NaK [18]. It was designed to power an Agena D research satellite.
The SNAP 10 A reactor still orbits the earth and is predicted to re-enter the earth’s
atmosphere in about 4,000 years.
NASA, DOD and DOE initiated a tri-agency effort in the 1980s to develop the
SP-100 nuclear reactor for applications in the power range from 50 kWe to 1 MWe
[19]. The reactor was based on the use of heat pipes with a lithium coolant. The
project was not advanced to flight hardware development and was cancelled in
1994.
Los Alamos National Laboratory developed the Safe Affordable Fission Engine
(SAFE) concept for electricity production in space. The reactor fuel is uranium
nitride pins with a rhenium cladding. The pins surround a molybdenum-sodium
heat pipe that transports heat to a heat exchanger. The heat exchanger heats up
helium and the hot gas drives a Brayton power system [20].
The USSR launced about 31 Romashka nuclear reactors used in Radar Ocean
Reconnaissance Satellites on Cosmos missions. The Romashka reactors were a fast
6.3 MEMS 303
Table 6.5 List of space reactor power systems from the US and USSR [20]
SNAP SP-100 Romashka Bouk Topaz 1 Topaz 2 SAFE-400
10
Dates 1965 1992 1967 1977 1987 1992 2007
kWth 45.5 2000 40 *100 150 135 400
kWe 0.65 100 0.8 *5 5-10 6 100
Transducer TE TE TE TE TI TI TI
Fuel U-ZrHx UN UC2 U-Mo UO2 UO2 UN
Mass-kg 435 5422 455 *390 320 1061 512
Neutron Thermal Fast Fast Fast Thermal Thermal/ Fast
spectrum Epithermal
Coolant NaK Li None NaK NaK NaK Na
Core Temp. C 585 1377 1900 NA 1600 1900 1020
spectrum graphite reactor which used 90 % enriched uranium carbide fuel. The
Romashka reactor used thermoelectric converters to produce electricity. The USSR
then developed the Bouk fast reactor which produced 3 kW of power for about
4 months. Later the USSR flew a design, like the Cosmos-954 which re-entered the
earth over Canada in 1978, that used uranium-molybdenum fuel rods.
The Topaz I reactors were then designed and they used enriched uranium fuel
with a high temperature moderator with hydrogen and a thermionic converter. The
Topaz I reactors produced about 5 kWe of electrical power. It was flown in 1987 on
Cosmos 1818 and 1867. It was designed to operate for three to five years as a power
source for ocean surveillance.
A list of space reactor power systems developed by the US and USSR is shown
in Table 6.5.
6.4 Summary
Nuclear batteries have a long history. They have been used as power sources for
cardiac pacemakers in clinical trials in the 1970s but were phased out when lithium
ion batteries were developed. The military has had and continues to have an interest
in nuclear batteries for a number of applications. The problem is that nuclear battery
designs are difficult if not impossible to scale to micro and nano size scales due to
the scale length mismatch of the range of the charged particles produced in the
decay of radioisotopes and the scale of the transducer. This chapter has listed a
number of desired applications but has contrasted this desire with the scaling issues
of nuclear battery design in size, power, and power density. There are potential
niches where nuclear battery size, power, and power density are not hindrances to
the designers. These applications include areas such as deep space exploration and
underwater probes.
The reader has available a number of tools which can be used to evaluate
different nuclear battery technologies and the suitability of the design for specific
applications.
Problems
1. Why were nuclear batteries dropped for pacemaker use?
2. Suppose an automobile company proposes to eliminate the use of gasoline by
producing a line of cars powered by nuclear batteries. Is this feasible? Why or
why not? Are there any safety or security concerns? If so, what are they?
3. A mobile electronics manufacturer, interested in supplementing the amount of
power available to their gadgets, investigates placing nuclear batteries around
the devices’ circuitry. What considerations should be taken into account when
deciding whether or not to pursue this idea? Why?
4. Consider the remaining isotopes that are candidates for a “Nano-Power
System” nuclear battery. Are there any additional considerations which would
further eliminate them? Justify your answer.
5. Considering the actual problems in nuclear battery designs, are there other
applications not mentioned in this chapter which might use nuclear batteries?
6. Describe the Pioneer missions and the nuclear battery technology used in the
missions.
7. Describe the Curiosity Rover mission and the nuclear battery used in the
mission.
8. Will nuclear batteries find a role in military missions? Why or Why not?
9. Will a nuclear battery ever be useable for new drone technology?
10. Can nuclear battery technology ever meet the requirements of nano-power
systems?
11. Contrast nuclear battery technology to fission reactor technology.
12. What is the status of SNAP reactor technology?
13. What is the status of TOPAZ reactor technology?
14. Are there plans to use nuclear reactors in space?
References
1. Ehrenberg W, Chi-Shi L, West R (1951) The electron voltaic effect. In: Proceedings of the
physical society, Section A, vol 64, p 424
2. Rappaport P (1954) The electron-voltaic effect in $p-n$ junctions induced by beta-particle
bombardment. Phys Rev 93:246–247
2,745,973
4. Windle WF (1964) Microwatt radioisotope energy converters. IEEE Trans Aerosp 2:646–651
5. OlsenLC, Cabauy P, Elkind BJ (2012) Betavoltaic power sources. In: Physics today, pp 35–38
6. Huffman FN, Migliore JJ, Robinson WJ, Norman JC (1974) Radioisotope powered cardiac
pacemakers. Cardiovasc Dis 1:52–60
References 305
7. National_Research_Council_Radioisotope_Power_Systems_Committee, Radioisotope Power

Systems: An Imperative for Maintaining US Leadership in Space Exploration: National
Academies Press
8. Department_of_Energy, “Draft EIS for the Proposed Consolidation of Nuclear Operations
Related to Production of Radioisotope Power Systems,” S. a. T. Office of Nuclear Energy, Ed.,
ed. Washington DC: DOE, 2005
9. Lastres O, Chandler D, Jarrell JJ, Maldonado GI (2011) Studies of Plutonium-238 production
at the high flux isotope reactor. Trans. Am. Nucl. Soc. 104:716–718
10. Albright D, Kramer K (2005) Neptunium 237 and Americium: World Inventories and
Proliferation Concerns. In: ISIS Document Collection, I. f. S. a. I. Security, Ed., ed. ISIS
11. Ambrosi RM, Williams HR, Samara-Ratna P, Bannister NP, Vernon D, Crawford T et al
(2012) Development and testing of Americium-241 Radioisotope thermoelectric generator:
concept designs and breadboard system, presented at the Nuclear and Emerging Technologies
for Space. The Woodlands, TX
12. Williams HR, Ambrosi RM, Bannister NP, Samara-Ratna P, Sykes J (2012) A conceptual
spacecraft radioisotope thermoelectric and heating unit (RTHU). Int J Energy Res 36:1192–
1200
13. Fiskebeck P-E (2006) Utilization of spent radioisotope thermoelectric generators and
installation of solar cell technology as power source for Russian Lighthouses—Final
Report. In: Strand P, Sneve M, Pechkurov A (eds) Radiation and environmental safety in
North-West Russia: use of impact assessments and risk estimation. Springer, Dordrecht,
pp 85–88
14. Case FN, Remini WC (1980) Radioisotope powered light sources, presented at the
CONF-801157–1, United States
15. Pines DJ, Bohorquez F (2006) Challenges facing future micro-air-vehicle development.
J. Aircr. 43:290–305
16. Petricca L, Ohlckers P, Grinde C (2011) Micro- and nano-air vehicles: state of the art. Int J
Aerosp Engg 2011:17
17. Shure LI, Schwartz HJ (1965) Survey of electric power plants for space applications, presented
at the National Meeting of the American Institute of Chemical Engineers. Philidelphia,
Pennsylvania
18. Corliss WR (1965) SNAP Nuclear Space Reactors. ed: Atomic Energy Comisison
19. Sovie RJ (1987) SP-100 Advanced technology program, presented at the 22nd Intersociety
Energy Conversion Engineering Converence, Philadelphia
20. Association WN (2016) Nuclear reactors and radioisotopes for space. http://www.world-
nuclear.org/information-library/non-power-nuclear-applications/transport/nuclear-reactors-for-
space.aspx
Appendix A
Range Calculations
Throughout this book the range of electrons in matter are frequently referenced.
These range calculations are based on simulations of electrons traveling through
various materials using MCNP6. The range of electrons is a difficult matter.
Electrons have mass thousands of times less than the atoms in any material and are
therefore scattered in many directions as they give up their energy. This is in stark
contrast compared to heavy ions which mostly travel in straight paths. Heavy ions
only have dramatic changes in path when they directly strike a nucleus.
Electron energy deposition in materials is not a problem that can be solved
analytically. Instead they are primarily found with monte carlo simulations using
cross sections derived from quantum mechanical principles. These cross sections
are typically calculated from the dirac equation [1], the Hatree-Fock method [2, 3],
and those derived by Moliere [4–8].
The simulations used in this book are based off of a “Pencil Electron Beam”, or a
beam with a small or point-like source. This beam is directed at a given material,
the energy deposition is produced and ranges are calculated from there. The point of
these simulations is to show the maximum range a beta particle if it approached a
surface at the optimum angle, i.e. perfectly perpendicular to a surface.
In principle, if you could find a function D(r, h, z) which represented the energy
deposition in material at any point, you could derive a function which describes the
energy deposition in the axial direction in terms of z by integrating over r and h.
Z 1 Z 2p
ZðzÞ ¼ Dðr; h; zÞr dr dh ðA:1Þ
0 0
There is no way to produce a continuous solution such as the example above, so

in computer simulations one must consider how we can find our desired results. Our
calculations are based on taking a cylinder with an arbitrary radius large enough to
ensure no electrons escape through the outer surface. Then the cylinder is cut into
numerous slices, or cells, and the energy deposition is calculated in each cell. The
finer the size of each cell, the finer the detail of the picture. However, electron
simulations are very computationally expensive so increasing the number of cells

Lecture Notes in Energy 56, DOI 10.1007/978-3-319-41724-0
308 Appendix A: Range Calculations
can greatly increase computational time as well reduce statistical accuracy of the
results.
The two main values of interest in these calculations are the point of maximum
energy deposition and the furthest point in the z direction where 1 % of the total
remaining energy is deposited. These values are used to calculate what we call the
“tail to peak ratio” or t/p. This term is a way to demonstrate how far out that the
particle range is from the maximum energy deposition peak. It was found the t/p
ratio for each material was surprisingly close despite large differences in b energies.
The t/p ratio for each energy was within ±1 for many materials. It is proposed here
the t/p ratio could be used as a new rule of thumb for nuclear battery design.
Specifically for identifying the position of the source with respect to a transducer.
If we use the continuous example, we could define the tail to be at position
x where the following equation is satisfied
R1
ZðzÞdz
Rx1 ¼ 0:01 ðA:2Þ
0 ZðzÞdz
Merely stating the point where electrons deposits no energy is not sufficient. The
details of a model, such as the number of particles transported, can change where
the last cell the energy is deposited in. If we use the last 1 %, we can argue that the
final range where the energy left to be insignificant. Figure A.1 gives the axial
energy deposition of Y-90 in Sr-90 with the positions of the peak and tail identified
It must be noted that in some of the following tables the ranges of certain
energies are far less than 1 l. These results must be taken with a grain of salt. At
these ranges we approach the realm where size of the cells approximating the
medium as continuous is dubious. Instead it is asked for that the reader understand
how thin a film or layer would need to be to not waste significant amounts of
available energy from the source (Table A.1).
Peak
Tail
Fig. A.1 Axial energy deposition in Sr-90 from Y-90 bmax

Appendix A: Range Calculations 309
Table A.1 Beta particle ranges in encapsulating material

Isotope Branching ratio Energy type Peak depth (lm) Tail Depth (lm) t/p
H-3a 1 Max 4.959 7.809 1.6
Average 0.61875 0.99375 1.6
Ar-39 1 Max 348.75 1383.75 4.0
Average 85.5 347.7 4.1
Ar-42 1 Max 409.2 1623.6 4.0
Average 107.2 415.4 3.9
Co-60 1 Max 18.63 93.15 5.0
Average 3.12 13.91 4.5
Kr-85 1 Max 195.5 1105 5.7
Average 49.45 266.6 5.4
Sr-90 1 Max 118.75 650 5.5
Average 27 145.8 5.4
Y-90a 1 Max 780 3770 4.8
Average 234 1305 5.6
Ru-106 1 Max 3.325 18.025 5.4
Average 0.4715 1.9885 4.2
Cd-113 m 1 Max 34.2 201.6 5.9
Average 6.8 38.4 5.6
Cs-134 0.71 Max 156.8 1048.6 6.7
Average 33.15 204.75 6.2
0.28 Max 8.82 51.94 5.9
Average 1.615 8.67 5.4
Cs-137 0.06 Max 369 2214 6.0
Average 88 572 6.5
0.93 Max 122.4 761.6 6.2
Average 26.35 156.55 5.9
Sb-125 0.4 Max 18 107 5.9
Average 2.7 15.15 5.6
0.18 Max 5.1 29.4 5.8
Average 0.594 3.333 5.6
0.14 Max 48.6 291.6 6.0
Average 11.05 64.35 5.8
Pm-146 1 Max 50.25 338.35 6.7
Average 9.8 61.6 6.3
Pm-147 1 Max 9.6 58.2 6.1
Average 1.2 7.5 6.3
Sm-151 1 Max 1.608 10.1505 6.3
Average 0.216 1.152 5.3
(continued)
310 Appendix A: Range Calculations
Table A.1 (continued)

Eu-152 0.13 Max 56 389.2 7.0
Average 12.65 77.625 6.1
0.084 Max 150 990 6.6
Average 42.75 275.5 6.4
Eu-154 0.1 Max 212.5 1312.5 6.2
Average 60.3 395.3 6.6
0.28 Max 14 94 6.7
Average 1.92 12.36 6.4
0.36 Max 48 303 6.3
Average 8.8 55.55 6.3
Eu-155 0.022 Max 5.25 34.125 6.5
Average 0.672 3.948 5.9
0.077 Max 10.2 61.2 6.0
Average 1.275 7.5 5.9
0.18 Max 14.8 93.425 6.3
Average 1.92 12.24 6.4
0.26 Max 8 48 6.0
Average 0.99 5.775 5.8
0.47 Max 6.4 39.6 6.2
Average 0.7905 4.65 5.9
Tm-171 0.98 Max 1.8375 12.1275 6.6
Average 0.27 1.41 5.2
0.02 Max 0.3515 1.85 5.3
Average 0.045 0.22 4.9
Os-194 0.76 Max 0.304 1.976 6.5
Rule of Thumbb 0.065 0.35 5.4
0.24 Max 0.84 5.04 6.0
Average 0.072 0.387 5.4
Tl-204 1 Max 24.05 185 7.7
Average 5.4 39.15 7.3
Pb-210 0.8 Max 0.11875 0.625 5.3
Average 0.02 0.08 4.0
0.2 Max 0.77 5.06 6.6
Average 0.117 0.6045 5.2
Ra-228 0.4 and 0.1c Max 0.85 4.85 5.7
Average 0.12 0.594 5.0
0.3 Max 0.171 0.873 5.1
Rule of thumbb 0.035 0.175 5.0
0.2 Max 0.4505 2.544 5.6
Rule of thumbb 0.0925 0.481 5.2
(continued)
Appendix A: Range Calculations 311
Table A.1 (continued)

Ac-227 0.54 Max 0.508 3.1115 6.1
Average 0.07 0.375 5.4
0.35 Max 0.36 2.1375 5.9
Average 0.05 0.265 5.3
0.1 Max 0.153 0.873 5.7
Average 0.025 0.11 4.4
Pu-241 1 Max 0.09 0.505 5.6
Average 0.015 0.065 4.3
a
The encapsulating material for H-3 is T2O and for Y-90 is Sr-90
b
The average energy for these transitions could not be located. The rule of thumb, 1/3rd bmax, was
used instead
c
The energy of these transitions two transitions are so close it was not necessary to run simulations
for both
References
1. Thumm U, Norcross DW (1993) Angle-differential and momentum-transfer

cross sections for low-energy electron-Cs scattering. Phys Rev A 47:305–316
2. Bharadvaja A, Kaur S, Baluja KL (2015) Electron-impact cross sections of SiH2
using the R-matrix method at low energy. Phy Rev A—At, Mol, Opt Phys 91
3. Kwei CM, Hung CJ, Su P, Tung CJ (1999) Spatial distributions of elastically
backscattered electrons from copper and silver. J Phys D: Appl Phys 32:3122–
3127
4. Ikegami S (2013) A new screening length for small angle multiple scattering.
Nucl Instr Meth Phys Res, Sect B: Beam Interact Mater Atoms 311:14–19
5. Nakatsuka T, Okei K, Takahashi N (2013) Analytical derivation of higher-order
terms of Molière’s series and accuracy of Molière’s angular distribution of fast
charged particles. Nucl Instrum Methods Phys Res, Sect B: Beam Interact Mater
Atoms 311:60–70
6. Bethe HA (1953) Molière’s theory of multiple scattering. Phys Rev 89:1256–
1266
7. Bednyakov AA (2014) On the Molière theory of multiple scattering of charged
particles (1947–1948) and its critique in subsequent years. Phys Part Nucl
45:991–999
8. Borisov NM, Panin MP (2005) Generalized particle concept for adjoint monte
carlo calculations of coupled gamma-ray-electron-positron transport. Nucl Sci
Eng 150:284–298
Appendix B
Beta Spectra
The energies of electrons emitted from beta decay are not discrete. They come in
spectra which vary from transition to transition. The fact that beta decay has an
energy spectrum is critical when designing batteries.
A beta spectrum can be calculated using the following equations [1, 2]
1=2
NðWÞ ¼ FðZ; WÞ W 2 1 W ðW0 W Þ2 an ðWÞ ðB:1Þ

2pe R 2ðc0 1Þ jCðc0 þ imÞj2
pm
FðZ; WÞ ¼ ð1 þ c0 Þe ðB:2Þ
h jCð2c0 þ 1Þj2
where
h i aZW 1
c0 ¼ 1 ðaZÞ2 ; m¼ ; R ¼ aA1=3
pe 2
where F(Z, W) is the fermi correction factor, Z is the number of protons in the
daughter nucleus, W is the total electron energy, W0 is the maximum electron
energy, a is the fine structure constant (a 1/137), pe is the electron momentum,
A is the atomic number of the daughter nucleus, and C is the gamma function. The
term an is a shape factor, or forbiddeness correction, which varies for each tran-
sition. This last term is a complicated quantity which depends the total angular
momentum of both the father and daughter nucleus, nuclear matrix elements, and
Laguerre polynomials.
Such calculations are far too complicated for the needs of this text. We have
opted to use a simple curve fit using two different functions: a general nth order
polynomial, and a Gaussian times a general nth order polynomial.
X
n
Pn ðxÞ ¼ c0 þ c1 x þ c2 x2 þ þ cn1 xn1 þ cn xn ¼ ci xi ðB:3Þ
i¼0

Gn ðxÞ ¼ exp aðx bÞ2 Pn ðxÞ ðB:4Þ

314 Appendix B: Beta Spectra
Some distributions are not easily described by a single function, in this case a
piecewise function using (B.3) and (B.4) are used instead. The reader should take
note of how strange the shape of some spectra are, and consider our previous
discussions of why the rule of thumb is not a valid quantity for energy deposition
calculations.
The first table gives curve fit results for spectra that can be represented by a
single function. The second table is for data which requires more complicated fitted
functions.
All tabulated data can be found from Eckerman et al. [3] and Burrows [4].
Ar-39 5th order polynomial P5(x)

c0 = 0.0542975, c1 = 0.211152, c2 = −1.04579, c3 = 3.9536, c4 = −11.4275,
c5 = 10.6228
(continued)
Appendix B: Beta Spectra 315
(continued)
Co-60 5th order Gaussian polynomial G5(x)
a = 1.53819, b = 0.844349, c0 = 0.000292291, c1 = 0.000340777,
c2 = −0.00182501, c3 = 0.00256999, c4 = −0.00163961, c5 = 0.000386332
(continued)
(continued)
Kr-85 5th order polynomial P5(x)
c0 = 0.0751608, c1 = −0.0164363, c2 = −0.123639, c3 = 1.03957,
c4 = −3.66591, c5 = 3.0966
(continued)
(continued)
Sr-90 5th order polynomial P5(x)
P5(x) c0 = 0.0783593, c1 = −0.0525869, c2 = 0.0353627,
c3 = 0.939949, c4 = −6.84788, c5 = 8.15519
(continued)
(continued)
Y-90 6th order polynomial P6(x)
P6(x) c0 = 0.0371757, c1 = 0.0967542, c2 = −0.052831,
c3 = −0.0762372, c4 = 0.114233, c5 = −0.0599542,
c6 = 0.0108771
(continued)
(continued)
Ru-106 2nd order polynomial P2(x)
P2(x) c0 = 0.146473, c1 = −7.21746, c2 = 88.3578
(continued)
(continued)
Cd-113 m 2nd order gaussian polynomial G2(x)
G2(x) a = 1.75094, b = 0.981609, c0 = 0.493132,
c1 = −1.68291, c2 = 1.43633
(continued)
(continued)
Pm-146 5th order polynomial P5(x)
P5(x) c0 = 0.0442974, c1 = −0.216431, c2 = 1.14036,
c3 = −3.11062, c4 = 3.62793, c5 = −1.51221
(continued)
(continued)
Pm-147 3rd order polynomial P3(x)
P3(x) c0 = 0.125455, c1 = −0.701068, c2 = −1.09674,
c3 = 7.68007
(continued)
(continued)
Sm-151 3rd order polynomial P3(x)
P3(x) c0 = 0.14261 c1 = −3.39532 c2 = 15.9662
c3 = 52.4037
(continued)
(continued)
Eu-152 4th order polynomial P4(x)
P4(x) c0 = 0.0384366, c1 = 0.0531053, c2 = −0.0138342,
c3 = −0.0558261, c4 = 0.0225823
(continued)
(continued)
Tm-171 4th order polynomial P4(x)
P4(x) c0 = 0.147879, c1 = −3.32881, c2 = 32.0602,
c3 = −260.598, c4 = 1262.28
(continued)
(continued)
Os-194 5th order polynomial P5(x)
P5(x) c0 = 0.180259, c1 = −4.46917, c2 = 1.76161,
c3 = 1019.67, c4 = −12168.3, c5 = 44613
(continued)
(continued)
Tl-204 5th order polynomial P5(x)
P5(x) c0 = 0.101451, c1 = −0.126086, c2 = −0.260198,
c3 = 1.37329, c4 = −2.59362, c5 = 1.60531
(continued)
(continued)
Ra-228 2nd order polynomial P2(x)
P2(x) c0 = 0.147361, c1 = −7.40471, c2 = 92.6067
Some distributions simply cannot be fitted by our two functions. Particularly

where the distributions is very close to zero. In the following distributions, using
our basic functions to fit the entire spectrum results in equations which oscillate and
give negative values. Here we use piecewise functions made from Eqs. (B.3) and
(B.4). In some cases there are mid points where no order of polynomial give
sufficient accuracy, so for simplicity we use interpolation between each point.
It must be noted the following spectra plots are log plots. A linear plot would
show you nothing for values very close to zero.
Co-60 P5(x) − P3(x) − P3(x) − G4(x)

P5(x) c0 = 0.502678, c1 = −1.81922, c2 = 11.4158, c3 = −120.578, 0 < x < 0.26095
c4 = 409.44, c5 = −425.973
P3(x) c0 = 1.19148, c1 = −8.88462, c2 = 21.8461, c3 = −17.7167 0.26094 < x < 0.3355
P3(x) c0 = 0.0258329, c1 = −0.16217, c2 = 0.33741, c3 = −0.232179 0.3355 < x < 0.41005
G4(x) a = −1.11036, b = 1.68613, c0 = 0.00124666, 0.41005 < x < bmax
c1 = −0.00352486, c2 = 0.00429838, c3 = −0.00250657,
c4 = 0.000558856
(continued)
(continued)
Sb-125 P7(x) − P5(x)
P7(x) c0 = 0.383138, c1 = −2.33611, c2 = −68.8253, c3 = 1260.63, 0 < x < 0.29555
c4 = −9003.28, c5 = 32722, c6 = −60073, c7 = 44310.7
P5(x) c0 = 0.0280618, c1 = 0.0707625, c2 = −1.19134, c3 = 3.97343, 0.29555 < x < bmax
c4 = −5.65366, c5 = 2.99594
(continued)
(continued)
Cs-134 P1(x) − G3(x) − P1(x) − P1(x) − G6(x) − P1(x)
P1(x) c0 = 0.496753, c1 = −3.24645 0 < x < 0.10905
G3(x) a = 1.18212, b = 1.19126, c0 = 0.817923, c1 = −2.48238 0.10995 < x < 0.54515
c2 = 1.88652, c3 = 0.00033918
P1(x) c0 = 0.111159, c1 = −0.175544 0.54515 < x < 0.61785
P1(x) c0 = 0.0253043, c1 = −0.0365876 0.61785 < x < 0.69055
G6(x) a = 11.7436, b = 0.353791, c0 = 1.71261, c1 = −11.8098, 0.69055 < x < 1.34475
c2 = 33.8344, c3 = −51.5625, c4 = 44.1068, c5 = −20.0935,
c6 = 3.81241
P1(x) c0 = 6.52118 10−6, c1 = −4.55218 10−6 0.134475 < x < bmax
(continued)
(continued)
Cs-137 P5(x) − P9(x)
P5(x) c0 = 0.202055, c1 = −0.319891, c2 = 0.130764, c3 = 2.90727, 0 < x < 0.4986
c4 = −17.6648, c5 = 21.4826
P9(x) c0 = 33.3686, c1 = −370.918, c2 = 1818.72, c3 = −5162.83, 0.4986 < x < bmax
c4 = 9351.68, c5 = −11210.3, c6 = 8894.72, c7 = −4505.05,
c8 = 1321.87, c9 = −171.227
(continued)
(continued)
Eu-154 P6(x) − P7(x)
P6(x) c0 = 0.469017, c1 = −2.48171, c2 = 6.33741, c3 = −9.47392, 0 < x < 0.6915
c4 = 8.29434, c5 = −3.87991, c6 = 0.741933
P7(x) c0 = 1.352, c1 = −7.19861, c2 = 16.4171, c3 = −20.6283, 0.6915 < x < bmax
c4 = 15.4187, c5 = −6.86221, c6 = 1.6847, c7 = −0.176007
(continued)
(continued)
Eu-155 P6(x) − P5(x)
P6(x) c0 = 0.204388, c1 = −2.16722, c2 = 7.64514, c3 = −79.6122, 0 < x < 0.16645
c4 = 659.986, c5 = −1115.43, c6 = −2269.47
P5(x) c0 = 0.725355, c1 = −16.8712, c2 = 159.544, c3 = −759.329, 0.16645 < x < bmax
c4 = 1807.68, c5 = −1716.02
(continued)
(continued)
Pb-210 P7(x) − P5(x)
P7(x) c0 = 0.470392, c1 = −35.916, c2 = −6489.76, 0 < x < 0.32265
c3 = 1.62722 106, c4 = −1.76164 108,
c5 = 1.04837 1010, c6 = −3.20069 1011,
c7 = 3.87786 1012
P5(x) c0 = 0.0313845, c1 = −1.20417, c2 = 24.4462, c3 = −478.983, 0.32265 < x < bmax
c4 = 6065.21, c5 = −28545.7
The last spectrum is for Pu-241. Our data simply does not have enough points to
produce a satisfactory curve fit. In this case it is recommended to use a linear
interpolating function. In the following table we give the available data points and a
plot of a linear interpolation of the data.
The data points are given in pn = (energy [MeV], intensity)
p1 (10−3, 2.59 10−1) p2 (3 10−3, p3 (5 10−3,

2.11 10−1) 1.67 10−1)
p4 (7 10−3, p5 (8.9 10−3, p6 (1.27 10−2,
1.28 10−1) 9.38 10−2) 1.42 10−1)
p7 (2 10−2,
6.00 10−5)
References
1. Cross WG, Ing H, Freedman N (1983) A short atlas of beta-ray spectra. Phys
Med Biol 28:1251
2. Keefer G, Piekpe A (2004) Beta spectra for Ar-39, Kr-85, and Bi-210
3. Eckerman KF, Westfall RJ, Ryman JC, Cristy M (1994) Availability of nuclear
decay data in electronic form, including beta spectra not previously published.
Health Phys 67:338–345
4. Burrows TW (1988) Information analysis center report national nucliar data
center brookhaven national laboratory associated universities, inc. under con-
tract no. DE-AC02-76CK00016 with the
Appendix C
Theoretical Nuclear Battery Design
Concepts
As nuclear battery designs evolve, theoretical concepts will come forth. The reader
has developed skills which can help dissect them. Below are some theoretical
design concepts developed by one of the authors. An analysis procedure developed
in Chap. 5 will be used to evaluate the strengths and weaknesses of the concepts.
• Does the mechanism depend upon the production of heat or does it depend upon
the formation of ions?
• If the mechanism depends upon ion production, then the pair production effi-
• Of the power deposited in the transducer, consider the efficiency of converting
this power to a useful product such as electrical power ðgtransducer Þ.
• The transport efficiency (ηpd) of depositing power from ionizing radiation into
Solid-State Emitter and PV
Diamond is not a direct band-gap material but does have a bound exciton which can
be used like a direct band-gap emitter. There will be no self-absorption of the
exciton photon, because the photon energy (5.1 eV) is less than the band-gap of
diamond (5.49 eV). The binding energy of the electron-hole pair that makes up the
exciton is 70 mV. This device will have temperature limitations that need to be
explored. The theoretical maximum efficiency for this configuration is 33 %.
An approach being studied by the authors that is similar to the SEGRIEP concept
is to use solid-state emitters based on high quality binary solid state crystals which
exhibit wide band-gaps and direct band-gap transitions. In a direct wide band-gap
binary material the photon self-absorption and reabsorption processes are in balance
until the photon escapes the solid. Loss processes such as luminescence emission
from the surface and Auger recombination can be limited by proper design. The

338 Appendix C: Theoretical Nuclear Battery Design Concepts
Fig. C.1 The solid-state material interacts with radiation and produces an electron-hole pair. The
pair recombines and produces a photon. The photon is then reabsorbed to form another
electron-hole pair or to reflect off the surface. If an electron-hole pair is formed, it recombines and
produces a photon. The process is in balance with few other losses and continues until the photon
is lost through the loss cone into the PV cell. The theoretical maximum efficiency for this con-
figuration is 33 %
photon can escape through a loss cone that is coupled to a photovoltaic cell
transducer (Fig. C.1). Ionizing radiation will create displacements in the solid-state
crystal. The rate of displacements will be on the order of 170 displacements per ion
fragment. The number of photons create by each fragment (with estimated energy
of 10 MeV) is on the order of 2 million (fraction of energy into electron-hole pair
formation (0.42) times the energy of fragment (10,000,000 eV)/band-gap of
semiconductor (2.2 eV for GaP)). Thus the rate of photon generation exceeds the
rate of potential trap formation by a factor of 20,000. The potential traps do con-
tinue to build up with time. However, if the device operates at a temperature where
self-annealing can occur (600–800 K), there will be a point defects (displacements)
are repaired at a sufficient rate to limit the effects of displacements due to radiation
damage. It is feasible to use this balance of defect creation and defect repair to
extend the lifetime of the solid-state emitter. This device will still have radiation
damage issues. Defects will be created in the emitter and the traps that are formed
can absorb photons and electrons. The key to the extended lifetime of the emitter is
through self-annealing to mitigate trap formation. Displacement issues are more
serious with binary materials such as III–V, than with diamond (used in the
SEGRIEP concept). The physics of the process is still being researched and refined.
Appendix C: Theoretical Nuclear Battery Design Concepts 339
(i) Does the mechanism depend upon the production of heat or does it depend
upon the formation of ions.
The solid-state emitter and photovoltaic cell uses ionizing radiation to create
electron-hole pairs in a direct band-gap semiconductor.
(ii) If the mechanism depends upon ion production, then the pair production effi-
Depending upon the direct band-gap material that is used, the pair production
efficiency will vary. The value of pair production efficiency can be found in
Table 3.9 (e.g., 0.421 for SiC, 0.381 for GaN and 0.344 for GaAs).
(iii) Of the power deposited in the transducer, consider the efficiency of converting
In the solid-state emitter and photovoltaic cell, the spectral matching efficiency is
1 since the photon emitted exactly matches the band-gap of the photovoltaic cell.
The only remaining factors are the driving potential efficiency, which will be
around 0.5 (see Chap. 5), and a fill factor of about 0.8 (see Chap. 5). Thus, the
transducer efficiency is 0.4.
(iv) The transport efficiency (ηpd) of depositing power from ionizing radiation into
The transport efficiency depends upon the interface of the radioisotope to the
transducer. In this case, the most logical interface is a surface interface. If the
radioisotope is layered in between the solid-state interface, then in the best case the
transport efficiency will be about 0.3. The atomic dilution factor will be on the order
of 10.
(v) Radiation damage and other issues
Radiation damage occurs in semiconductor materials through the creation of
Frenkel pairs. In the case of displacement rates, there will be approximately 180
displacements per 5.3 MeV alpha particle stopped in GaN. This can be converted to
a vacancy production per eV deposited, 180 displacements/5.3 106 = 3.4 10−5
d/eV. The number of electron-hole pairs that will be created by a 5.3 MeV alpha
particle is is 5.3 106/8.9 = 5.96 105 (where the mean energy needed to create
an electron-hole pair in GaN is 8.9 eV from Table 3.9). This can be converted to
number of photons produced per eV deposited 5.96 105/5.3 106 = 0.112
photon/eV. About 90 % of the Frenkel pairs will recombine so the total number of
vacancies created per alpha particle is about 18. So the displacement rate becomes 3.
4 10−6 d/eV. If the temperature (T) of the semiconductor is elevated, more of the
vacancies will recombine and the rate at which this occurs is Rv(T) vacancy
recombinations/s. The cell uses a 5 lm thick Po-210 source for excitation in order to
minimize self absorption in the Po-210 layer. The GaN cell is 1 cm by 1 cm in area
and 20 lm thick. The Po-210 volume is 5 10−4 cm3. The power density per gm of
Po-210 is 1,315 W/cm3 so the power output from the Po-210 source is
1,315 W/cm3 * 5 10−4 cm3 = 0.6575 W. About 20 % of this power will be

transported into the GaN cell from a surface source. The cell volume is Vcell =
1 * 1 * 20 * 10−4 cm3. So the average power deposition in the cell 0.675/20 * 10−4
cm3 * 3,288 W/cm3 = 2.055 1022 eV/cm3. Thus the total number of photons
produced will be 2.055 1022 eV/s cm3 * 0.112 photons/eV = 2.3 1021
photons/s cm3 and the rate of Frenkel pair production rate will be, 2.
055 1022 eV/s cm3 * 3.4 10−6 d/eV = 6.986 1016 d/s cm3, the Frenkel
pair loss rate will be Rv(T)/Vcell. So, the net Frenkel pair production rate is,
Fpp ¼ 6:986 1016 d=s cm3 Rv ðTÞ=Vcell :
The buildup of Frenkel pairs will determine the lifetime of the cell. Postulating
that some critical density of Frenkel pairs (DFpcrit) exists where the cell performance
drops too much, the lifetime (sc) can be estimated from the following integral
equation,
Zsc

DFPcrit ¼ 6:986 1016 Rv ðTÞ=Vcell dt
0
DFpcrit
sc ¼
ð6:986 1016 Rv ðTÞ=Vcell Þ
The cell can be operated at elevated temperature in order for the Frenkel pair
recombination rate, R(T), is maximized.
If the Po-210 is sandwiched between two cells, the surface source interface
efficiency is 0.4. The efficiency of producing photons by the interaction of alpha
particles with the cell is 0.381. The photons created in the cell will transport through
the cell by reflecting off the cell surface and will have a life cycle that includes
absorption and re-emission (thus effectively increasing the photon lifetime until the
photon either enters the photovoltaic cell or is lost through an interaction with a
defect). The photon transport efficiency is 1. The spectral matching efficiency (ηin)
will be 1 since the photovoltaic cell is GaN. The driving potential efficiency (ηdp)
will be about 0.5, the fill factor (FF) will be about 0.8. Thus the estimated maximum
potential system efficiency is 0.4 * 0.381 * 1 * 1 * 0.5 * 0.8 = 0.061. The prob-
lems with the cell will be radiation damage in the cell and how it impacts photon
production and transport. The maximum power produced by the cell is
0.061 * 0.6575 Wth = 0.04 We. The maximum power density produced by the cell
will be 0.04 We/20 10−4 cm3 = 20 We/cm3.
Appendix C: Theoretical Nuclear Battery Design Concepts 341
Hybrid Solid-State Emitter
A hybrid approach to addressing the radiation damage and self-absorption problems is

to form microbubbles with excimer gases in the solid-state material [1]. Micro bubbles
can be formed at very high pressures in a solid-state material using ion implantation
(up to 4 GPa). At 4 GPa, the density of a xenon gas bubble is on the order of 4 g/cm3.
The transport length of radiation in a high pressure xenon micro bubble is about 5 lm,
about the scale length of the heavy fragment. As shown in Fig. C.2 the radioisotope
can be coated on the cell’s surface. A series of micro bubbles lie between the
radioisotope layer and the p-n junction. The particle from the radioisotope is emitted
isotropically and the micro bubble serves as both a shield to protect the junction as well
as a photon source which emits at the excimer wavelength. The photons then resonate
in the PV cell and are absorbed. Even at this high density the issues of pressure
broadening should not lead to losses and the micro bubble should not self-absorb.
Thus the cell will have a transducer scale length compatible with the radiation source
and with the PV cell. The advantages of this approach are that the wide band-gap p-n
structure will use a thin film with the radioisotope coated or imbedded into the
structure. Wide band-gap materials can operate at high temperatures without effi-
ciency loss and have high thermal conductivities. The films can be stacked, which will
allow for scaling of the power source at relatively high power densities (see discussion
on the limitations of nuclear battery power density). Problems with this mechanism do
exist. Even though it is well known that micro bubbles form by ion implantation, the
possibility of the bubble delaminating may be a problem. The theoretical maximum
efficiency for this configuration is 20–30 %.
(i) Does the mechanism depend upon the production of heat or does it depend
upon the formation of ions.
The mechanism is dependent on the formation of ion pairs.
(ii) If the mechanism depends upon ion production, then the pair production
efficiency (ηpp) sets the upper limit of the maximum possible system
efficiency.
Fig. C.2 Option A: Micro

bubble as a radiation shield as Radioisotope Layer
well as a way of converting
the kinetic energy of radiation N
into narrow band UV photons K
that are absorbed by the p-n
junction UV
photons
P
The microbubble is filled with an excimer gas which interacts with the ionizing
radiation. The ion pair production efficiency for rare gases is 0.5.
(iii) Of the power deposited in the transducer, consider the efficiency of converting
Excimer photons are created in the microbubble and then transport to a p-n
junction. The spectral matching efficiency (ηin) for a Kr-85 microbubble which uses
an aluminum nitride p-n junction is Eg/hm = 6.2/8 = 0.775. The driving potential
efficiency (ηdp) is 0.5 and the fill factor (FF) is 0.8. The maximum transducer
efficiency is 0.775 * 0.5 * 0.8 = 0.31.
(iv) The transport efficiency (ηpd) of depositing power from ionizing radiation into
Assuming a 10 µm radius microbubble filled with high density Kr-85 at liquid
density, a significant amount of the beta energy emitted by Kr-85 is retained within
the microbubble. The transport efficiency (ηpd) is estimated to be approximately 0.6.
(v) Radiation damage and other issues
The microbubble protects the p-n junction from radiation damage. The efficiency
of transporting photons from the microbubble to the p-n junction is about 0.9. The
maximum system efficiency is ηsystem = 0.6 * 0.31 * 0.5 = 0.093. Forming a
matrix of 10 µm diameter microbubbles in the semiconductor volume surrounding
the p-n junction is a complicated problem, but ion implantation is a means of
generating such a matrix. The problem with this design is the atomic dilution factor
(DFatomic). How far apart can the microbubbles be in order to maintain stability
within the materials structure? It is estimated that the optimum case leads to an
atomic dilution factor of about 500. This means that the average power density in
the cell is approximately Pdav = (1/BVWmin)/500 = 0.0025 W/cm3. The maximum
cell power output will be about 0.0025 * 0.155 = 0.00039 W/cm3.
Reference
1. Prelas MA (2013) Micro-scale power source, United States Patent 8552616.

USA Patent
Appendix D
Ranges for Alpha Emitters
The range of alpha particles emitted from a source in the source material is an
important consideration in determining the optimum thickness of the source
material. The range of an alpha particle in the source material is calculated using the
ion transport code SRIM [1]. The model uses a beam of monoenergetic alpha
particles entering the source material normal to its surface.
Figure D.1 is for metallic Gd-148 which emits a 3.182 MeV alpha particle. The
range of this alpha particle in metallic Gd is 8.44 lm.
Figure D.2 is for metallic Po-208 which emits a 5.115 MeV alpha particle. The
range of this alpha particle in metallic Po is 18.2 lm.
Figure D.3 is for metallic Po-210 which emits a 5.305 MeV alpha particle. The
range of this alpha particle in metallic Po is 18.7 lm.
Figure D.4 is for metallic Th-228 which emits a 5.34 MeV alpha particle. The
range of this alpha particle in metallic Th is 15.4 lm.
Figure D.5 is for metallic U-232 which emits a 5.263 MeV alpha particle. The
range of this alpha particle in metallic U is 9.55 lm.
Figure D.6 is for metallic Pu-236 which emits a 5.721 MeV alpha particle. The
range of this alpha particle in metallic Pu is 8.33 lm.
Figure D.7 is for metallic Pu-238 which emits a 5.456 MeV alpha particle. The
range of this alpha particle in metallic Pu is 7.71 lm.
Figure D.8 is for metallic Am-241 which emits a 5.442 MeV alpha particle. The
range of this alpha particle in metallic Am is 11.2 lm.
Figure D.9 is for metallic Cm-243 which emits a 5.742 MeV alpha particle. The
range of this alpha particle in metallic Cm is 12.8 lm.
Figure D.10 is for metallic Cm-244 which emits a 5.762 MeV alpha particle.
The range of this alpha particle in metallic Cm is 12 lm.
Figure D.11 is for metallic Bk-248 which emits a 5.793 MeV alpha particle. The
range of this alpha particle in metallic Bk is 14.5 lm.
Figure D.12 is for metallic Cf-250 which emits a 6.03 MeV alpha particle. The
range of this alpha particle in metallic Cf is 15.5 lm.

344 Appendix D: Ranges for Alpha Emitters
Fig. D.1 A plot of the range of a 3.182 MeV alpha particle in metallic Gd using the SRIM
transport code [1]
Fig. D.2 A plot of the range of a 5.115 MeV alpha particle in metallic Po using the SRIM
transport code [1]
Appendix D: Ranges for Alpha Emitters 345
Fig. D.3 A plot of the range

of a 5.305 MeV alpha particle
in metallic Po using the SRIM
transport code [1]

in metallic Th using the SRIM
transport code [1]

in metallic U using the SRIM
transport code [1]

in metallic Pu using the SRIM
transport code [1]
Fig. D.7 A plot of the range of a 5.456 MeV alpha particle in metallic Pu using the SRIM
transport code [1]
Fig. D.8 A plot of the range of a 5.442 MeV alpha particle in metallic Am using the SRIM
transport code [1]

in metallic Cm using the
SRIM transport code [1]
Fig. D.10 A plot of the

range of a 5.762 MeV alpha
particle in metallic Cm using
the SRIM transport code [1]

particle in metallic Bk using

particle in metallic Cf using
Reference
1. Ziegler JF, Ziegler MD, Biersack JP (2010) SRIM–The stopping and range of
ions in matter (2010). Nucl Instrum Meth Phys Res Sect B: Beam Interact Mater
At 268:1818–1823
Index
A Beta particles, 12, 14, 15, 32, 34, 58, 59, 81,
Absorbed power, 222 84, 90, 94, 95, 97, 99, 102, 104, 121, 147,
Accelerator, 35, 45, 49, 64–67, 73, 74, 299 151, 184, 192, 198, 200–203, 208, 216,
Actinium, 40, 57–60, 62 223, 237, 238, 258–261, 273, 280, 288
Activity, 10, 40, 47, 57, 71, 72, 194, 243, 267, Beta scale length, 200
269, 289, 299 Beta spectrum, 14, 98, 99, 102, 201, 205, 275,
Aerosol, 44, 135, 137, 138, 147, 179, 182, 183, 315
214, 249, 264, 265, 267–269 Betavoltaic, 15, 145, 171, 191, 223, 260, 270,
Alpha particles, 4, 23, 30, 31, 33, 58, 59, 81, 282, 288, 289
87, 89, 90, 96, 103, 193, 194, 198, 199, Betavoltaic analysis, 258
223, 234, 244, 253–257, 267, 269, 275, Betavoltaic cell, 10, 28, 100, 103, 172, 189,
276, 280, 340, 343 205, 208, 223, 259, 260, 276, 288
Alphavoltaic, 1, 15, 27, 103, 145, 171, 172, Bethe-Bloch stopping power equation, 89
189, 194, 213, 223, 258, 270, 282 Bethe formula, 90
Alphavoltaic analysis, 253 Bismuth phosphate process, 50–54
Alphavoltaic cell, 103, 223, 257, 260 BLIP, 64–66
Angular distribution, 2, 119, 121, 124, 198, Bragg curve, 89, 255
223, 233, 234, 260, 275 Bragg peak, 89, 255
Areal density, 90, 95, 96 BWR, 48
Atomic dilution factor (DFatomic), 177, 215,
253, 339, 342 C
Canyons, 50, 52
B Circuit current, 107, 108, 223, 224, 229, 230,
Baking, 195 257
Band-gap, 24, 27, 28, 108, 111, 137, 139–141, Conversion factor, 12, 223
144, 145, 148–150, 154, 168, 170, 192, Cosmogenic, 39, 41, 44
215, 223, 225–228, 254, 257, 261, 264, Cost of accelerator, 73
271, 273, 276, 337–339, 341 Cost of neutron capture, 69
Battery configuration, 261 Cost of separation, 67
Battery surface area per Watt (BAWmin), 14, Cost per neutron, 70, 73
19, 280, 295 Curiosity rover, 291, 306
Battery volume per Watt (BVWmin), 14, 17,
19, 253, 280, 295 D
Beta, 2, 13–15, 17, 23, 46, 58, 90, 93, 99, 101, Dark saturation current, 110, 111, 225, 226
103, 108, 112, 115, 150, 152, 183, 192, Decay, 1–3, 10–12, 35, 36, 39, 40, 44, 45, 47,
309, 315, 342 58, 61, 70, 71, 73, 81, 92, 94, 95, 99, 107,
Beta energy, 11, 14, 95, 97–102, 104, 201, 116, 125, 127, 138, 145, 187, 200, 202,
203–205, 208, 237, 259, 260, 275 214, 235, 256, 260, 269, 282, 294, 305, 315

352 Index
Decay chain, 40–42, 45, 49, 57, 58, 60, 61 F

Decay series, 59–62 Fill Factor (FF), 107, 223, 224, 226, 227, 229,
Deep space probes, 288 261, 271, 339, 340, 342
Depletion region, 27, 103–105, 111–115, 154, Fission fragments, 30, 45, 46, 81, 84, 85, 89,
191, 206, 207, 239, 253, 255–259, 271 104, 188, 198
Depletion zone, 100, 107, 108, 111, 114, 205, Fluorescer, 126, 127, 135–138, 144, 147, 265
206, 208, 224, 261 Fluorescer efficiency, 261
Dilution factor, 177, 178, 215–218, 239, 253, Frenkel pair, 339, 340
258, 261, 269, 281, 294, 301, 304, 342
Direct charge nuclear battery (DCNB), 116, G
117, 119, 121–124, 168, 171, 230, Gamma, 23, 57, 88, 93, 244, 294, 315
232–234, 282, 287, 116, 230, 287 Gamma rays, 23, 81, 84, 90, 92, 150, 188, 198,
Directed beam, 234, 253 294, 301–303
Direct energy conversion, 24, 28, 114, 136, Gaseous fluorescer, 126
145, 156, 168, 171 Gaseous phase, 180, 183, 188
Driving potential efficiency, 108, 111, 115,
223, 225–227, 257, 258, 261, 264, H
270–272, 274, 339, 340, 342 Half life, 243
Drones, 293, 300 Health, 73, 95, 239, 244, 249, 251, 282
Heat, 1, 3, 18, 24, 26, 28, 34, 45, 102, 108,
E 117, 124, 156–158, 160, 167, 171, 198,
Efficiency, 2, 10, 15–17, 22, 24, 28, 33–36, 44, 246, 247, 251, 253, 289, 291, 304, 337,
53, 55, 93, 94, 97, 103, 105, 108, 112–115, 338, 341
119, 121, 123–127, 130, 134, 136–142,
144, 145, 147, 150, 152, 158, 167, 170, I
171, 181, 186, 189, 191, 197, 206, 213, Ideal, 14, 15, 17, 24, 34, 107, 110, 116, 119,
222, 223, 225, 227, 230, 234, 252, 141, 161, 216, 223, 224, 226, 229, 230,
257–262, 264, 269–275, 280, 281, 288, 233, 234, 260, 282, 300
298, 304, 337, 339–342 Ideal cell, 16, 111, 223, 226, 229, 230
Efficiency of the transducer, 15, 22, 34, 124, Impurities, 62, 100, 103, 132, 206, 239
222, 225, 252, 281, 298, 339 Interface, 10, 11, 15–17, 22, 26, 29, 31, 34,
Electric field, 105, 106, 111, 116–122, 124, 106, 114, 145, 154, 177, 178, 181, 192,
154, 155, 171, 206, 230, 231, 271 208, 209, 217, 221, 264, 281, 298, 304,
Electron-hole pair, 23, 24, 26, 27, 103, 105, 339, 340
107, 108, 111, 115, 148–150, 154, 191, Inventory, 41, 44, 47, 48, 147, 289, 300
194, 206, 222, 224, 227–230, 236, 253, Ionizing radiation, 11, 14–18, 22–24, 26, 29,
255, 256, 270–273, 337–339 30, 34, 35, 81, 102, 103, 112, 113, 127,
Electrons, 14, 18, 24, 30, 66, 90, 95, 103, 112, 137, 145, 148, 150, 156, 183, 186, 187,
125, 150, 159, 160, 199, 206, 224, 237, 191, 196, 198, 209, 217, 221, 230, 235,
273, 309, 310, 338 236, 238, 270, 272, 273, 276, 280, 281,
Electron utilization efficiency, 222, 227–229 298, 337–339, 341, 342
Energy distribution, 2, 83, 84, 86, 95, 96, 105, Ionizing radiation products, 18
119, 121, 123, 124, 160, 196, 198, 223, Ion pairs, 18, 23, 26–28, 34, 102, 187, 198, 341
229–234 Ion scale length, 199
Energy spectrum, 82, 83, 201, 258, 259, 315 Ion source, 85, 86, 89, 136, 138
Escape probability, 196 Isotope, 2, 3, 12–15, 17, 18, 30, 31, 35, 40, 41,
Evaporation, 195 44, 46, 47, 49, 50, 57, 58, 62, 64, 66, 73,
Excimer, 94, 127–129, 132, 138, 140, 144, 76, 90, 93, 125, 177, 189, 202, 214, 217,
145, 147, 172, 184, 186, 187, 192, 213, 275, 289, 293, 298, 299, 301–304
260, 264, 265, 340–342 Isotropic, 11, 29, 99–102, 120, 122, 123, 152,
Excimer fluorescer, 138–140, 145, 261, 262 154, 194, 201, 202, 204, 209, 230, 231,
234, 252, 253, 257, 260, 270, 281
Index 353
L Nuclear fission, 40
Lighthouses, 291 Nuclear fuel reprocessing, 55
LINAC, 65, 66 Nuclear regulatory commission (NRC), 194,
Linear energy transfer (LET), 84, 89, 199, 276 239, 240, 246
Liquid phase, 182, 189
Liquid semiconductor, 114, 115, 183, 191 O
Ohmic contact, 115, 191
M Open circuit voltage, 107, 108, 110, 111, 158,
Manmade, 39, 40, 45 165, 223–226, 229, 230, 257, 271
Maximum power, 1, 22, 34, 107, 108, 177,
215, 223, 224, 229, 257, 263, 298, 301, 340 P
MEMS, 17, 300 Pacemakers, 288, 290, 305
Micro aerial vehicle (MAV), 300 Particle range, 96, 98, 201, 269, 310
Micro-bubbles, 155, 212 Path length, 30, 31, 33, 178, 179, 199, 218
Military missions, 293 Peltier effect, 156, 158
Multi mission radioisotope thermoelectric Perpetual motion limit (PML), 16, 272
generator (MMRTG), 291 Phosphors, 126, 148, 150, 151, 155, 292, 293
MURR, 64, 70–73, 128 Photon energy converter, 136, 140
Photon-intermediate direct energy conversion
N (PIDEC), 28, 135, 140
Nano areal vehicles (NAV), 300 Photons, 16, 24, 28, 92, 93, 126, 127, 135–137,
Nano power systems, 301, 303 140–142, 145, 149, 150, 152, 154–156,
NASA, 125, 165, 239, 245, 246, 248, 249, 304 170, 179, 184, 192, 200, 213, 215, 223,
NASA missions, 160, 289 229, 230, 260, 264–267, 337–342
NEMS, 300 Photovoltaic, 27, 28, 94, 108, 137–142, 144,
Neutron capture, 45, 46, 49, 63, 67, 69, 71, 72, 145, 147–149, 152–155, 168, 170, 187,
76, 82, 235, 236, 299 223, 224, 226, 228–230, 261, 262, 264,
Neutron flux, 46, 47, 63, 71–74, 129, 211, 289 266, 267, 270, 271, 276, 337–340
Neutrons, 11, 23, 45, 46, 63, 69, 70, 73, 75, 81, PIDEC analysis, 260
82, 128, 150, 196, 211, 236, 242 Plasma phase, 26, 189, 192
Nucleal fuel reprocessing, 45, 46 P-N junction, 15, 28, 31, 33, 100, 103, 106,
Nuclear, 1, 2, 10, 11, 14–19, 23, 25, 27, 29, 34, 107, 110, 113, 114, 142–145, 154, 155,
39, 41, 45, 46, 49, 55, 68, 75, 92, 97, 99, 189, 194, 208, 213, 222, 224, 225, 227,
102, 115, 126, 134–136, 138, 151, 152, 229, 236, 238, 253, 257–260, 269–272,
156, 159, 168, 178, 183, 192, 194, 198, 275, 276, 280, 341, 342
211, 222, 227, 231, 240, 252, 258, 268, P-N junction scale length, 205, 207, 208
280, 315 Poisoning, 245
Nuclear batteries, 1, 2, 4, 10, 13, 14, 16, 35, 39, Portable nuclear batteries, 294
41, 44, 46, 47, 50, 67, 70, 74, 81, 88, 91, Positrons, 90
98, 102, 119, 148, 150, 183, 189, 198, 201, Potential applications, 287
208, 210, 221, 239, 255, 264, 272, 280, Potential energy, 1, 113, 116, 119–121, 123,
281, 287–289, 292, 294, 300, 301, 305, 341 124, 161, 167, 191, 223, 231, 232, 271, 273
Nuclear battery design, 2, 3, 10, 11, 15, 29, 97, Power density, 2, 11, 14, 15, 22, 34, 36, 39,
102, 189, 198, 209, 210, 222, 235, 253, 107, 126, 127, 132, 137, 140, 147, 148,
272, 276, 281, 287, 294, 301, 303, 305, 156, 177, 178, 180, 186, 196, 211,
310, 337 215–217, 224, 237, 239, 252, 253, 258,
Nuclear battery efficiency, 221 263, 267, 269, 276, 280, 298, 301–304,
Nuclear battery literature, 183, 269, 274, 275, 339–342
281 Power deposited, 106, 147, 216, 221, 222, 252,
Nuclear-driven fluorescer (NDF), 126, 136, 256, 265, 267, 270, 273, 275, 280, 337,
137 339, 341
354 Index
Power dilution factor (DFpower), 178, 263, S

269, 281 Safety, 54, 69, 74, 239, 245–248, 299, 301
Primary Knock-On Atom, 236 Scale length, 2, 3, 11, 16, 18, 39, 81, 97, 103,
Primordial, 39–41, 43 104, 115, 135, 136, 138, 145, 147, 155,
Protons, 11, 41, 64, 66, 73, 140, 190, 236, 238, 198, 205, 208, 258, 259, 269, 270, 280, 341
239, 315 Scale length matching, 137, 198
PUREX, 50, 54, 55, 68 Schottky barrier, 113–115, 171, 191, 192, 208,
PWR, 47, 48 227, 228, 272–274
Pyroprocessing, 56, 57 Secular equilibrium, 40, 57, 60
Seebeck effect, 1, 2, 27, 156
R Semiconductor, 10, 14, 15, 23, 24, 27, 103,
Radiation, 2, 3, 11, 12, 16, 18, 24, 26, 28, 30, 105, 112, 113, 115, 140, 145, 149, 154,
39, 43, 81, 84, 85, 93, 95, 103, 111, 114, 157, 170, 183, 191, 206, 208, 227, 228,
117, 126, 128, 138, 140–142, 144, 145, 235, 236, 238, 261, 264, 270, 271, 338,
149, 154, 155, 170, 184, 186, 190, 192, 339, 342
198, 210, 212, 217, 224, 230, 235, 240, Shielding considerations, 92
244, 258, 270, 276, 280, 288, 337, 338, Short circuit current, 107, 142, 223, 229
340, 341 Solid phase, 81, 180, 189, 264
Radiation damage, 2, 10, 11, 13, 14, 16, 35, Spent nuclear fuel, 35, 48, 67, 75, 299
103, 105, 113, 115, 145, 147, 150, 151, Sputtering, 196
183, 191, 198, 235–237, 276, 338–340, 342 Successful applications, 287
Radiation generated current, 110, 111, 206, Sunk costs, 69
225 Supply, 39, 42, 43, 46–49, 62, 66, 94, 137,
Radioisotope, 339, 341 263, 264, 291, 299, 301
Radioisotope energy conversion system Surface interface, 14–16, 18, 34, 113, 178, 192,
(RECS), 135, 137 193, 215, 339
Radioisotope phase, 26 System efficiencies, 119, 145, 171, 252, 260,
Radioisotope powered lighting sources, 292 271, 294, 303, 337, 339–342
Radioisotope production, 49, 57, 64, 66, 240
Radioisotopes, 1, 2, 13, 14, 19, 29, 35, 39–41, T
44–50, 54, 56, 57, 60–63, 66, 75, 81, 87, Theoretical maximum energy efficiency, 257
98, 105, 135, 137, 138, 147, 178, 180, 187, Thermionic generator, 159, 166
189, 193, 196, 198, 211, 217, 221, 234, Thermionics, 125, 168, 171
238, 239, 244, 245, 259, 263, 287, 289, Thermoelectric generator, 2, 157, 158
295, 301, 304 Thermophotovoltaics, 27, 168, 223
Radioisotope thermoelectric generator (RTG), Thin foil, 193, 258
1, 27, 165, 291, 304 Thorium, 40, 45, 57, 58, 60, 62, 244
Radionuclide extraction, 60, 148 Transducer, 2, 3, 11, 14–17, 26, 29–31, 33–35,
Reciprocating cantilever nuclear battery, 124 81, 97, 98, 100, 102, 115, 126, 145, 147,
Recombination, 28, 108, 134, 149, 154, 222, 149, 156, 177–181, 183, 189, 192, 197,
230, 236, 337, 339, 340 198, 205, 208–210, 212, 214, 215, 217,
REDOX, 50, 53–55 221, 222, 227, 233, 235–237, 252, 253,
Remote power applications, 291 255, 258, 261, 264, 269–273, 310, 337, 339
Reserve, 41 Transducer efficiency, 34, 221, 222, 258, 261,
Reverse saturation current, 111, 225 263, 270–272, 274, 280, 294, 342
Richard-Dushman equation, 161 Transducer Phase, 26
Rules of thumb, 95, 97, 99, 102, 202, 205 Transducer scale length, 213, 341
Index 355
Transport scale length, 2, 198 V

Traps, 107, 112, 149, 206, 222, 224, 238, 338 Voltage, 27, 108, 111, 112, 115, 116, 122, 157,
Tritium, 11–14, 16, 17, 44, 104, 238, 277, 281, 164, 167, 187, 206, 208, 225, 226, 233,
292, 293 234, 257, 271, 272, 276, 287
Volume interface, 11, 14, 17, 29, 178, 216, 217
U
Unmanned aerial vehicles (UAVs), 300 W
Uranium, 40, 45, 46, 48, 50, 51, 53, 56–58, 60, Work function, 27, 113, 115, 159, 160, 167,
62, 74, 75, 82, 84, 87, 89, 90, 182, 183, 191, 208, 227
195, 196, 210, 228, 304 World supply, 42, 47
Uranofullerene, 187, 189

Nuclear Batteries and Radioisotopes

Uploaded by

Document Information

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 Batteries and Radioisotopes

Uploaded by

Copyright:

Available Formats

Lecture Notes in Energy 56

More information about this series at http://www.springer.com/series/8874

Fernando De La Torre Aguilar

Matthew Boraas Modeste Tchakoua Tchouaso

Fernando De La Torre Aguilar Denis Wisniewski

ISSN 2195-1284 ISSN 2195-1292 (electronic)

Library of Congress Control Number: 2016945768

© Springer International Publishing Switzerland 2016

Printed on acid-free paper

This Springer imprint is published by Springer Nature

Remember: NSEI, 1958–2012

Columbia, USA Mark Prelas

1 Introduction to Nuclear Batteries and Radioisotopes . . . . . . . . . . . . 1

2.6 Mixed Oxide Fuel Fabrication Facility (MOX FFF) . . . . . . . . . . 75

4.4.3 Evaporation and Sputtering . . . . . . . . . . . . . . . . . . . . . 195

6.3 MEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298

Appendix A: Range Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307

Appendix B: Beta Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Appendix C: Theoretical Nuclear Battery Design Concepts . . . . . . . . . . 337

Appendix D: Ranges for Alpha Emitters . . . . . . . . . . . . . . . . . . . . . . . 343

Keywords Energy conversion Scaling Surface source Volume source

© Springer International Publishing Switzerland 2016 1

1.1 Fundamental Concepts

The interest in radioisotopes as a possible source of energy is founded on the

where k is a decay constant and t is time in seconds.

NðtÞ ¼ Nð0Þekt ð1:2Þ

where N(0) is the initial number of radioisotope nuclei in the sample.

Solving Eq. (1.4),

Table 1.1 (continued)

Table 1.1 (continued)

kt ¼ ln ð0:5Þ ð1:5Þ

The rate constant can be found from Eq. (1.5).

AðtÞ ¼ kNðtÞ ð1:7Þ

1.2 Nuclear Battery Design Considerations

• What type of ionizing radiation does the isotope emit?

T ! 31 He2 þ b þ m þ 19 keV ð1:8Þ

where the general form for a nucleus is represented by AN Xz (X is the nucleus, A is

 ¼ 19 keV ¼ 6:33 keV

A Curie (represented by the symbol Ci) is a unit of radioactivity which repre-

W keV 6:02 1023 atoms

1.2.1 Surface Interface

which are not suitable for projected Micro-Electro-Mechanical Systems (MEMS)

1.2.2 Volume Interface

The above example demonstrates several important lessons:

1.3 Products from Ionizing Radiation: Heat and Ion Pairs

per Watt (BVWmin) for various radioisotopes are shown

Teff ¼ E=kB ð1:19Þ

where E is the energy of the ion or electron in eV, kB is Boltzmann’s constant

Fig. 1.1 Energy conversion flow chart for radiation sources

Table 1.6 Potential types of Case State of Fuel State of transducer

1.4 Geometrical Considerations in the Interface

Fig. 1.2 Shown is an

Fig. 1.3 An illustration of how a radioisotope embedded in a compound is interfaced to a

Fig. 1.4 An illustration of

Fig. 1.7 Shown is an

Fig. 1.8 Shown is a model of

gsystem ¼ gtr gion gec ð1:21Þ

1.5 Methodology for Analysis

2.1 Existing Radioisotope Supplies

© Springer International Publishing Switzerland 2016 39

Fig. 2.1 Types of

2.1.1 Primordial Radioisotopes

2.1.2 Cosmogenic Radioisotopes

Cosmogenic radioisotopes are created by the interaction of cosmic-ray particles

T ! 31 He2 þ b þ m þ 19 keV ð1:8Þ

¼ 19 keV ¼ 6:33 keV