AFRPL TR-85-034 AD:

AD-A 160 734

Final Report
for the period
Antiproton Annihilation
1 April 1984 to
31 January
Propulsion
1985

September 1985 University of Dayton

Research Institute
Dayton, Ohio 45469

UDR-TR-85-55
F04611-83-C-0046
Author:
R. L. Forward Forward Unlimited
34 Carriage Square
Oxnard, CA 93030
Approved for Public Release RI-32901
Distribution unlimited. The AFRPL Technical Services Office has reviewed this report, and it is
releasable to the National Technical Information Service, where it will be available to the general
public, including foreign nationals.

preparedfor the: Air Force

Rocket Propulsion
Laboratory
Air Force Space Technology Center
Space Division, Air Force Systems Command
Edwards Air Force Base,
California 93523-5000

85 10 31 063
- -- ""7- . . , , . ... ,- -
 NOTICE
"When U.S. Government drawings, specifications, or other data
are used for any purposes other than a definitely related
government procurement operation, the Government thereby
incurs no responsibility or any obligation whatsoever, and the
fact that the Government may have formulated, furnished, or in
any other way supplied the said drawings, specifications, or
other data, is not to be regarded by implication or otherwise,
or in any manner licensing the holder or any other person or
corporation, or conveying any rights or permission to
manufacture, use, or sell any patented invention that may in
any way be related thereto."

This report was submitted by the University of Dayton Research

Institute, Dayton, Ohio 45469 USA under Contract Number
F04611-83-C-0046, Job Order Number 2308 M3 RC with the Air
Force Rocket Propulsion Laboratory, Edwards Air Force Base,
California 93523 USA. The report includes work done from
1 April 1984 to 31 January 1985. The research was performed
at the Air Force Rocket Propulsion Laboratory, Edwards Air
Force Base, California. The Principal Investigator was Dr.
Robert L. Forward. The Project Manager was Mr. Wayne E. Roe,
and the Task Manager was Dr. Franklin B. Mead, Jr.
This technical report is approved for release and distribution
in accordance with the distribution statement on the cover and
on the DD Form 1473.

WAYNE Ey ROE FRANKLIN B. MEAD

Research Coordinator Project Manager

FOR THE DIRECTOR

* JAMES !f. EDWARDS EDWARD L. IBBOTSON, Lt Col, USAF

. Chief, Plans and Programs Deputy Chief, Liquid Rocket Division
 Unclassified
SECURITY CLASSIFICATION OF THIS PAGE

REPORT DOCUMENTATION PAGE

is REPORT SECURITY CLASSIFICATION 1b. RESTRICTIVE MARKINGS
Unclassified
2.. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUT30NIAVAILABILITY OP REPORT
Approved for Public Release:
2b. OECLASSIFICATION/DOWNGRADING SCHEDULE Distribution is unlimited.

4. PERFORMING ORGANIZATION REPORT NUMbERIS) S. MONITORING ORGANIZATION REPORT NUMBERISI

UDR-TR-85-55 AFRPL-TR-85-034

6.&NAME OF PERFORMI1NG ORGANIZATION 5b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION

Forward Unlimited, subcontrac- (Ifappliwoblej Air Force Rocket Propulsion Lab.

tor to University of Da~yton ________ ________________________

6c. ADDRESS (City. Stats and ZIP COdeo 7b. ADDRESS (City, State and ZIP Code)

Oxnard, CA 93030/Dayton, Oil 45469 AFRPL/XRX, Stop 24

Edwards Air Force Base, CA 93523-5000
a.NAME OF FUNOING/SPONSORING

AFRPL
ORGANIZATION j
8 b. OFFICE SYMBO0L
(Ifapplicablei
9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
P04611-83-C-0046

8c. ADDRESS (City. State and ZIP Code) 10. SOURCE OF FUNDING NOS.

AFRPL/XRX, Stop 24 PROGRAM PROJECT TASK WORK UNIT

Edwards Air Force Base, CA 93523-5000ELMTN.O.O.0

11. TITLE iinclude Security Claeasfication,
ANTIPROTON ANNIHILATION PROPULSION (U) 62302F 2308 M3 RC
12. PERSONAL AUTHOR(S)

SForward, Robert L.
13&. TYPE OF REPORT 13b. TIME COVERED 1' AEOF REPORT (Yr..
Mo.. Day) 15, PAGE COUNT
FnlFROM 8.4/L4/L1 TO / 4.DT 85/9 212
16. SUPPLEMENTARY NOTATION

17. COSATI CODES it. SUBJECT TERMS lContinuer an ressere If necesssary aund identify by block numbers

21EL 03U UB A Antimatter Propulsion,* Antiproton Propulsion,* Advanced

2103Propulsion,
1.ABSTRACT (Con lnue on TEJritsEri ifneceuae-s and Identify by block number#

Antiproton annihilation propulsion is a new form of space propulsion, where milligrams of

antimatter are used to heat tons of reaction fluid to high temperatures. The hot reaction
fluid is exhausted from a nozzle to produce high thrust at high specific impulse. This
study was to determine the physical., engineering, and economic feasibility of antiproton
annihilation propulsion. The conclusion of the study is that antiproton propulsion is
feasible, but expensive.-Because the low mass of the antimatter fuel more than compensates
for its high price, comparative mission studies show that antimatter fuel can be cost
effective in space, where even normal chemical fuel is expensive because its mass must be
lifted into orbit before it can be used. Antiproton annihilation propulsion is mission
enabling, in that it allows missions to be performed that cannot be performed by any other
propulsion system. ..

20, OISTRIUUTION/AVAILAUILITY Of AISETRACT 21. ABSTRACT SECURITY CL.ASSIFICATION

UNCLASSIPIEO/UNLIMAI'EO 0SAME AS Fir C DYIC USERS C3 Unclassified

22m. NAME OP RESPONSIBLE INDIVIDUAL 22b TELEP111HONE NUM@ER 22C. OPPICI SYMBOL1101
IIRlvelsi Ame. Co~de
Wayne E. Rooe (605) 277-5206 XRX
DD FORM 1473,83 APR I OITION Of I jAN ?32ItOSOLIYE. t1,i,-1*.alFiaA
i/li ICURITY CLASSIP '1CATION Of 1T"IS Padk&
 TABLE OF CONTENTS

SECTION PAGE

INTRODUCTION AND SUMMARY 1

1 PRESENT ANTIPROTON PRODUCTION FACILITIES 5

1.1 BACKGROUND 5
1.2 ANTIPROTON PHYSICS AT CERN 7
1.3 ANTIPROTON PHYSICS AT FERMILAB 14
1.4 ANTIPROTON PHYSICS AT IHEP 17
1.5 COMPARISON OF ANTIPROTON PRODUCTION FACILITIES 20
1.6 EFFICIENCY OF PRESENT PRODUCTION FACILITIES 21

2 SOME ANTIPROTOR PRODUCTION FUNDAMENTALS 25

2.1 GENERATION OF ANTIPROTONS 25

2.2 ANGULAR CAPTURE BY MAGNETIC LENSES 29
2.3 MOMENTUM CAPTURE BY COLLECTING RINGS 31
2.4 ELECTRON COOLING OF ANTIPROTON BEAMS 32
2.5 STOCHASTIC COOLING OF ANTIPROTON BEAMS 34
2.6 STORING OF ANTIPROTONS 36

3 IMPROVING ANTIPROTON PRODUCTION EFFICIENCIES 41

3.1 ALTERNATE ANTIPROTON PRODUCTION TECHNIQUES 42

3.2 MAXIMIZING THE ANTIPROTON PRODUCTION RATE 44
3.3 IMPROVING TARGET EFFICIENCY 46
3.4 IMPROVING ACCELERATOR EFFICIENCY 47
"3.5 IMPROVING ANGULAR CAPTURE EFFICIENCY 49
3.6 IMPROVING MOMENTUM CAPTURE EFFICIENCY 52
3.7 SYSTEM EFFICIENCY ESTIMATES 54
3.8 CONCEPTUAL ANTIPROTON FACTORY 55
3.9 COST ESTIMATES 57

4 GENERAtTION OF ANTIHYDROGEN 59

4.1 CONVERSION OF ANTIPROTONS TO ATOMIC ANTIHYDROGEN 60

4.2 CONVERSION OF ANTIHYDROGEN ATOMS TO MOLECULES 64

5 SLOWING AND COOLING OF ANTIHYDROGEN 69

5.1 ELECTRONIC SLOWING AND COOLING OF IONS 69

5.2 LASER SLOWING AND COOLING OF NEUTRAL ANTIHYDROGEN 70 QiTALUTy

6) -

- ° -

~
.~ -. ~ ~ ~ ~. A' jC-•A '•sA "•-~, I
 .rw',r ~r C
- r771v rrY
7.r77 -- r.777- .. Z7 -77

6 TRAPPING AND STORING ANTIHYDROGEN 85

6.1 TRAPPING OF ATOMIC ANTIHYDROGEN 85

6.2 TRAPPING OF ANTIHYDROGEN 88
6.3 CONVERSION OF ANTIHYDROGEN GAS TO ICE 93
6.4 LEVITATION OF ANTIHYDROGEN ICE 94
6.5 ANTIHYDROGEN ICE BALL ENERGY BALANCE 98

7 ANTIPROTON ANNIHILATION PROPULSION 105

7.1 ANTIMATTER PROPULSION i05

7.2 EXTRACTION OF ANTIMATTER FROM STORAGE 107
7.3 PARTICLE PRODUCTION FROM ANTIPROTON ANNIHILATION 109
7.4 STOPPING OF PARTICLES 113
7.5 ANTIMATTER ROCKET ENGINE CONCEPTS 119

8 ANTIMATTER MISSION STUDIES 129

8.1 MINIMUM ANTIMATTER OPTIMIZATION 130

8.2 MINIMUM FUEL COST OPTIMIZATION 132
8.3 IMPOSSIBLE MISSIONS 134
8.4 COMPARATIVE COST STUDIES 135

9 CONCLUSIONS AND RECDONDATIONS 141

9.1 CONCLUSIONS 141

9.2 RECOMMENDED RESEARCH INVESTIGATIONS 141
9.3 RECOMMENDED ENGINEERING STUDIES 145
9.4 RECOMMENDED AFRPL IN-HOUSE RESEARCH PROJECT 147
9.5 REQUIRED FACILITIES 148
9.6 AREAS OF CONCERN 149

10 BIBLIOGRAPHY 151

10.1 PRODUCTION OF ANTIPROTONS AND ANTIHYDROGEN 151

10.2 SLOWING, COOLING, AND TRAPPING OF PARTICLES 160
10.3 ANTIMATTER ANNIHILATION AND PROPULSION 167

APPENDICES

A SOME USEFUL DATA AND CONVERSION FACTORS A-I

B COST COMPARISON OF CHEMICAL AND ANTIHYDROGEN

PROPULSION SYSTEMS FOR HIGH 4V MISSIONS B-i

C ANTIPROTON ANNIHILATION PROPULSION C-i

D MATERIAL ENTHALPY ASCENT/DESCENT (MEAD) MODULE D-I

iv

"•..... ~~~~~-------------------------..-----
_ • . - -.-- -..-- -..-.--.-.
.--------------.. * .*. .--..- ;•.•- -- . .- ,;•. :-. -. o
 (.>\ LIST OF FIGURES

Figure Title Page

1- 1 Producing, capturing, and storing antiprotons. 6
1- 2 Producing antiprotons at CERN. 8
1- 3 Antiproton production facility at Fermilab. 15
1- 4 Antiproton production facility at IHEP. 18
1- 5 Present antiproton capture efficiencies. 22

2- 1 Relative antiproton production rates

for hydrogen and heavy metal targets. 26
2- 2 Total antiproton production rates. 28
2- 3 Penning trap for antiproton ions. 38

3- 1 Total antiproton production rates and efficiencies. 45

3- 2 Antiproton production versus angular capture. 50
3- 3 Antiprotons versus momentum spread. 53
3- 4 Antiproton factory (one segment). 56

4- 1 Paul trap for positive and negative ions. 61

4- 2 Laser-enhanced antihydrogen formation. 62
4- 3 Molecular hydrogen rotational energy levels. 67
5- 1 Laser slowing of a neutral particle. 71
5- 2 Particle slowing using chirped laser frequency. 74
5- 3 Particle slowing using variable magnetic field. 76
5- 4 Pi-pulse deflection and cooling of atoms. 79
' 5- 5 Energy levels of atomic and molecular hydrogen. 82
5- 6 Tunable Lyman alpha source using four-wave mixing. 84
6- 1 Laser-magnet trap for polarized antihydrogen atoms. 87
6- 2 Trapping of atoms by resonance radiation pressure. 89
6- 3 Geometry of alternating-beam optical trap. 92
6- 4 Many paths from antiprotons to antihydrogen ice. 94
6- 5 Magnetostatic trap for antiparahydrogen ice. 96
6- 6 Active electric levitation of antihydrogen ice. 97
6- 7 Energy balance for antihydrogen ice ball. 99

7- 1 Schematic of an idealized antiproton rocket. 106

7- 2 Extraction of antiprotons from storage. 108
7- 3 Antiproton-proton annihilation particle spectra. i1
7- 4 Range of charged pions. 115
7- 5 Conceptual schematic for radiation shielding. 118
7- 6 First generation antimatter thermal rocket. 120
7- 7 Magnet assisted antimatter heated hydrogen rocket. 122
7- 8 High thrust magnetic containment antimatter rocket. 124
7- 9 High exhaust velocity antiproton rocket. 126
7-10 Magnetic nozzle for microexplosion plasmas. 128

8- 1 Chemical and antimatter rocket mass ratios. 133

8- 2 Relative fuel cost vs. mission velocity. 138
8- 3 Relative fuel cost vs. expanded mission velocity. 140

*K7
 LIST OF TABLES

Table Title Page

1-1 Comparison of Antiproton Production Facilities. 21
3-1 Antiproton Production Efficiencies. 55
7-1 Branching Ratios of pp Annihilation Products. 109
7-2 Range of Pions per 100 MeV of Energy. 114
7-3 Attenuation of Gamma Rays by Tungsten. 116
8-1 Mass Ratios for Difficult Missions. 134

vi
 INTRODUCTION AND SUMMARY

Antiproton annihilation propulsion is a new form of space

propulsion in which milligrams of antimatter are used to heat
tons of reaction fluid to high temperatures. The hot reaction
fluid is then exhausted from a nozzle to produce high thrust
at high specific impulse (1000 to 3500 s).
This study was contracted by the Air Force Rocket Ptopulsion
Laboratory to determine the physical, engineering, and
economic feasibility of antiproton annihilation propulsion.
The conclusion of the study is that antiproton annihilation
propulsion is feasible, but expensive.

Antimatter fuel has to be manufactured. It is probable that

the efficiency of production will always be low, so the price
will be high. Antimatter is a very lightweight form of fuel,
however, since the antimatter converts all of its mass to
energy upon annihilation with normal matter. Because its low
mass more than compensates for its high price, comparative
*u mission studies show that antimatter fuel can be cost
"effective in space, where even normal chemical fuel is
expensive because its mass must be lifted into low earth orbit
before it can be used.
For propulsion applications the antimatter should be in the
form of antiprotons. Unlike antielectrons (positrons), the
antiproton does not convert into gamma rays upon annihilation.
Instead, two-thirds of the annihilation energy is emitted as
charged particles (pions) whose kinetic energy can be
converted into thrust by interaction with a magnetic field
nozzle or a working fluid.
Antiproton annihilation propulsion is mission enabling, in
that it allows missions to be performed that cannot be
performed by any other propulsion system. The most striking
example of such an "impossible" mission is a simple sortie
mission that involves leaving an orbiting base, inspecting a
spacecraft in a counter-rotating orbit, then returning to base
a few hours later. This can be done with an antiproton
powered vehicle that has a mass ratio of 3:1. To carry out a
similar mission, a chemical rocket would have to have an
unachievable mass ratio of 500:1, while an electric propulsion
system would require days instead of hours to complete the
task.

=%e1

p-~ccC *~ * t -
Antiprotons are already being generated, captured, cooled, and
stored at a number of particle physics laboratories around the
world, albeit in small quantities. The rest of this report
discusses in detail the techniques for the efficient
generation, long-term storage, and effective utilization of
milligram quantities of antiprotons for space propulsion.

Since the fields of particle physics, laser physics, and

molecular beam physics are not included in the educational
background of the usual propulsion professional, this report
contains more than the usual amount of tutorial material and
a great number of bibliographic references. It is hoped that
this tutorial material will be useful to those attempting to
make antimatter propulsion a reality.
Section 1 describes the present facilities for the production
of antimatter in the form of antiprotons. The major producer
is the European Center for Nuclear Research (CERN) which has
been producing, capturing, and storing antiprotons in magnetic
storage rings for a number of years. Antiproton production
facilities are also under construction at the Fermi National
Accelerator Laboratory in the U.S. and at the Institute for
High Energy Physics in the USSR.
Antiprotons are produced by accelerating normal matter protons
to high speed and smashing them into a heavy metal target.
The interaction causes new particles to be made, among them a
small number of antiprotons. The antiprotons are focused with
a magnetic lens, captured in a collecting ring, then cooled
off and slowed down by other rings, and finally accumulated in
a storage ring. In the present facilities they are later
accelerated back up to high energies and smashed head-on into
normal protons to carry out high-energy particle physics
exper iments.
Sections 1.5 and 1.6 show that the production efficiencies of
the present machines are extremely low. Due to space, time,
and funding restrictions, only a small fraction (-0.1%) of the
antiprotons generated in the target are captured.
Section 2 is a tutorial summary of the present methods for
generating, capturing, cooling, and storing of antiprotons.
The present methods all have problems with efficiency. Some
of the problems are inherent in the physics, some are due to
the engineering limitations of the present designs, and some
are just due to a lack of time or money.
Section 3 is a detailed discussion of the limitations of the
present antiproton production techniques and methods for
improving the antiproton production efficiencies. If the
improvements were to be made, the antiproton production
efficiency in terms of number of antiprotons captured per

24..
incident protons hitting the target could be raised from the
present p/p=4xlO 7 at CERN and 3x10- 5 at Fermilab to a
production ratio of p/p=5xl0- 2 . Then, if the proton
accelerator were optimized for energy efficiency, the overall
energy efficiency could be raised to 2.5xi0- 4 . Although an
energy efficiency of 0.025% does not seem very efficient, it
is adequate to allow the production of antimatter at a cost of
10M$/mg, at which point antimatter becomes cost effective for
space propulsion.
The present methods for storing antiprotons are not suitable
for space propulsion. The storage rings are too massive and
the antimatter they can hold is too diffuse. Section 4
discusses the various techniques for adding a positron to the
antiproton to make antihydrogen atoms, then combining two
antihydrogen atoms to make an antihydrogen molecule.
Section 5 then shows how electromagnetic fields and laser
photons can be used to control, slow down, and cool
antihydrogen atoms and molecules.

Section 6 shows the same laser techniques can be used to stop

the antihydrogen molecules and put them in a trap. The laser
also cools the gas until it has the temperature of a
millidegree Kelvin. The supercooled gas is then turned into a
crystal of antihydrogen ice and the antihydrogen ice ball is
levitated in an electrostatic or magnetostatic trap until
ready ýor use. If the antihydrogen ice ball is kept below
2 K, its vapor pressure is so low (4x10- 1 8 To:r) that it can
be kept indefinitely. Section 6.5 contains a detailed energy
balance for the iceball assuming that some annihilation
reactions are going on inside the storage chamber. The amount
of energy deposited by the expected reactions is less than the
cooling to the chamber walls.

Section 7 goes into detail on the reaction products to be

expected from the annihilation of the antiprotons and how they
can be utilized to provide thrust. On the average there are 3
charged pions with an average kinetic energy of 250 MeV and
1.5 neutral pions that turn into 3 gamma rays with an average
energy of 200 MeV. The charged pions have a relatively short
lifetime, but in a properly designed engine, they last long
enough to transfer most of their kinetic energy into the
working fluid. The present estimates are that one third of
the annihilation energy from the antimatter fuel ends up as
kinetic energy in the vehicle. In a typical antiproton rocket
design where the working fluid was hydrogen gas, 1 mg of
antimatter was equivalent to 6 metric tons of propellant.
Section 8 discusses the effect of antiproton annihilation
propulsion on space mission design. First it is shown that no
matter what the mission characteristic velocity or the
antimatter rocket efficiency, the optimum mass ratio of an

3
antimatter rocket is never greater than 5:1. For most
missions near earth and in the solar system it is 2.5:1. This
contrasts strongly with chemically fueled missions, where mass
ratios are much greater.
Section 8.4 and Appendix B contain a comparative cost study of
a storable chemical fuel propulsion system, a liquid
oxygen/liquid hydrogen propulsion system, a nuclear thermal
hydrogen propulsion system, and an antiproton annihilation
propulsion system. Since hauling chemical fuel into low earth
orbit costs 5K$/kg or 5M$/T, it is shown that if antimatter
fuel costs 10M$/mg or less it is more cost effective than any
chemical propulsion system for any mission characteristic
velocity greater than 5 km/s. If the price of antimatter fuel
could be brought down to less than iM$/mg, then any mission in
the solar system, including a rendezvous mission to the rings
deep down in the gravity well of Saturn becomes possible.
Section 9 contains the basic conclusion that antiproton
annihilation propulsion is feasible, but expensive. It then
recommends a number of research and engineering studies that
need to be undertaken to verify that antimatter propulsion is
indeed feasible and to obtain a better estimate of the
antimatter production efficiencies and costs. Section 9.6 is
included for the skeptics. Here are listed those areas of
technology that are considered the weakest. These are the
areas where a "show stopper" 'ay lurk. If found and proven,
it would mean that antiproton annihilation propulsion is
either not possible or too difficult or expensive to pursue.
Section 10 contains a lengthy bibliography of all of the
pertinent papers in particle physics, nuclear physics, atomic
physics, laser physics, molecular beam physics, and antimatter
propulsion engineering that might be useful for someone
intending to work further in the field.

4
 SECTION 1

PRESEMT ANTIPROTON PRODUCTION FACILITIES

In this section we describe the present facilities for the

production of antimatter in the form of antiprotons. As we
will see, antiprotons are being made, collected, and stored
today. Thus, the production of antimatter is no longer a
question of technical feasibility, but a question of economic
feasibility.
The present methods for producing antimatter are highly
inefficient and extremely expensive, but they don't have to
be. Before we can start considering the use of antimatter for
propulsion, however, we will need to identify efficient
methods for making and storing antimatter that will produce
significant quantities of antimatter at a reasonable cost. To
start, let us see what is being done now in the production,
capture, and storage of antimatter.

1.1 BACKGROUND

Antimatter in the form of antiprotons is being made and stored

today, albeit in small quantities. The only known major
producer is the Organisation Europ~enne pour la Recherche
Nucl~aire (formerly tte Center for European Nuclear Research
or CERN) in Europe.l* Fermilab in the U.S. has started
construction of their antiproton facility and expects to be in
operation in 1985.1-2 In 1980, it was reported1-3 that the
Institute for High Energy Physics (IHEP) in the USSR was
constructing an antiproton production facility, but no further
information on construction progress has been published since.
In these facilities, the antiprotons are generated by sending
a high-energy beam of protons into a metal target as is shown
in Figure 1-1. When the relativistic protons strike the dense
metal nuclei, their kinetic energy, which is many times their
rest-mass energy, is converted into a spray of particles, some
of which are antiprotons. A magnetic field focuser and
selector separates the antiprotons from the resulting debris,
decelerates it, and directs it to a storage ring.
When the antiprotons are generated, they have a wide spread of
energies. Before they can be used further, it is necessary to
"cool" the beam so that all the antiprotons have the same
energy. Two techniques for reducing the spread in velocity
have been demonstrated. They are called electron cooling and
stochastic cooling and are described in Sections 2.3 and 2.4.

5
 RELATIVISTIC METAL
PROTON BEAMDEBR

100+10 MeV

* Fig. 1-1 Producing, capturingl, and storing antiprotons.

These cooled antiprotons could then go through another stage

of deceleration and cooling to bring TAGE
them ANTIPROTONS~
down to speeds V
suitable for capture and control by other techniques. The
present accelerator at CERN~ generates 3.5 GeV anti~pjotons
using a 26 GeV proton beam and has stored up to 10•
antiprotons In~their magnetic ring "racetrack" antiproton
accumulator. •
To give some scale to what has already been accomplished, 1012
antiprotons have a mass of 1.7 picograms. When this amount of
* antimatter is annihilated with an equivalent amount of normal
matter, it will release 300 joules, a significant quantity of
energy from an engineering viewpoint. To obtain this
"firecracker" amount of annihilation energy required the use
of multimillion dollar machines that used an enormous amount
of electric energy. Yet it is important to recognize that
scientists, working in basic physics, using research tools not
designed for the job, have produced and continue to produce
significant quantities of annihilation energy.

."

"INL.--ING
References:
1 1
' CERN Proton Synchrotron Staff, "The CERN PS complex as an
antiproton source," IEEE Trans. NS-30, 2039-2041 (1983).
1 2
. J. Peoples, "The Fermilab antiproton source," IEEE Trans.
NS-30, 1970-1975 (1983).
1
' 3 T.A. Vsevolozhskaya, B. Grishanov, Ya. Derbenev, N.
Dikansky, I. Meshkov, V. Parkhomchuk, D. Pesrikov, G.
Sil'vestrov, A. Skrinsky, "Antiproton source for the
accelerator-storage complex, UNK-IHEP," Fermilab Report FN-353
8000.00 (June 1981), a translation of INP Preprint 80-182
(December 1980).
1 4 D.B.Cline, C. Rubbia, and S. van der Meer, "The search for
intermediate vector bosons," Scientific Ax~erican 247, No. 3,
48-59 (March 1982).

1.2 ANTIPROTON PHYSICS AT CERN

The only facility in the world presently making significant
quantities of antiprotons is the European Organization for
Nuclear Research or CERN, outside Geneva, Switzerland. CERN
operates three large machines used for elementary particle
physics. These are the Proton Synchrotron (PS), the
Intersecting Storage Rings (ISR), and the 7 km circumference
Super Proton Synchrotron (SPS). There are also two smaller
machines used for antiproton collection and research, the
Antiproton Accumulator (AA) and the Low Energy Antiproton Ring
(LEAR). The CERN complex will grow in the future with the
addition of an Antiproton Collector (AC) and a Large Electron-
Positron (LEP) ring that will be 28 km in circumference.

1.2.1 Present Status of CERN facility.

The heart of the complex is the Proton Synchrotron, the
original CERN machine that has matured into a complex tool.
As is shown in Figure 1-2, protons are accelerated by the
Linear Accelerator (linac) to 50 MeV and injected into the
Booster which accelerates them to the near relativistic energy
of 0.8 GeV. The kinetic energy of the proton is now
comparable to its rest mass energy of 0.938 GeV. The Booster
then sends the protons to the Proton Synchrotron, which
further accelerates them up to 26 GeV. These 26 GeV protons
can be sent to the ISR or the SPS to carry out particle
physics experiments, or can be switched to a target area to
make antiprotons.
S.•.

S•.•-, .•- ,- • -• .- - .- • - • -• - - -• - - . .-.- - -. • ; o / . -• . 2,"L7- -

 "W'•7• " •S • • -- '

TO SPS
p ANTIPROTON COLLECTOR (1987) TO SPS - SUPER PROTON
SYNCH ROTRON
ANTIPROTON 270 GeV
ACCUMU LATOR
R'
RTURN 3.5 GeV/c
." TO PS.
: TARGET
LENS AAND
-* p

.. PS -
.. '." •PROTON
•. SYNCH ROTRON
.. 26 GeV

LINAC
50 MOV

!-' • LOW ENERGY ANTIPROTON RING

LEAR
100 - 600 MeV/c

Fig. 1-2 ,Producing antiprotons at CERN.

8
 Antiprotons are produced by focusing the 26 GeV protons down
%* to a 2 mm beam and inserting them into a 3 mm diameter, 11 cm
long copper wire target. The protons collide with the copper
nuclei in the wire and the kinetic energy of the collision
produces a spray of particles, mostly pions and kaons, some
protons and neutrons from the original nuclei, and an
occasional antiproton. The production rates are discussed
further in Section 2.1.
* The spectrum of the antipxotons is peaked around a momentum of
3.5 GeV/c (energy of 3 GeV). The antiprotons with that
momentum are focused by a short focal length pulsed magnetic
*. horn designed to capture a momentum bite of 1.5% at angles up
to 50 mrad. 1 -5 The focused antiproton beam is then
transported to the Antiproton Accumulator by a normal
quadrupole focusing channel. Although the AA was designed to
have an acceptance of 1001rmm-mrad (2 mm by 50 mrad), in
practice the acceptance has been found to only be 701r mm-mrad.
The antiprotons are injected into the outer half of the AA
ring where their momentum spread is reduced by a stochastic
precooling system to the point where they can be moved to the
inner part of the ring and deposited in the tail of the stack
of antiprotons accumulating there. About 7x10 6 antiprotons
are injected in each burst and are precooled before the next
burst arrives, 2.4 s later. 1 - 5 The antiprotons in the stack
"W-4 undergo further stochastic cooling and slowly build up into an
intense core of about 1011 antiprotons.
The scientists at CERN have discovered an effect which limits
the density of the antiprotons that can be stored and kept
cool in a single accumulator ring. The effect, called
intramodulation blow-up, is due to intrabeam scattering. It
is relate.d to the space charge limit in stationary collections
of ions and is independent of beam energy. The
intramodulation blow-up is an exponentially increasing
expansion of the beam which must be kept down by stochastic
cooling. With the present cooling system, the intrabeam
scattering expansion will equal the stochastic cooling
compression at a beam intensity of 6x10ll P. This effect must
be taken into account in the design of the accumulators and
coolers for an antiproton factory since it will limit the
number of antiprotons that can be held at one time in a
stochastic cooler before the antiproton ions must be
decelerated further or turned into neutral hydrogen.
. For particle physics experiments at high energies, the 3 GeV
*[ antiprotons are extracted from the AA and sent to the PS,
where they are accelerated up to 26 GeV. These high-energy
antiprotons are then sent to the SPS where they and an
oppositely directed beam of 26 GeV protons are simultaneously
accelerated up to as much as 270 GeV each and collided at

9
At

ýK'ý (t
center of mass energies of 540 GeV to produce new particles
such as the W and ZO vector bosons that are the carriers of
the weak force. (Finding these particles won the 1984 Nobel
Prize in Physics for S~mon van der Meer and Carlo Rubbia of
CERN).
For particle physics experiments at intermediate energies, the
26 GeV antiprotons from the PS are sent to one of the storage
rings of the ISR, where they undergo further cooling. Once
optimized, the transfer efficiency from the AA stack through
the PS to the ISR approaches 100%. The ISR has a good vacuum
system and has stored an 1.998+0.0025 ma beam of antiprotons
with no detectable loss for 55-hours, leading to an estimate
for the antiproton lifetime of greater than 30,000 hr.'- 6 One
experimental run on the ISR lasted for 2 weeks. For part 4 cle
physics experiments, the second ring of the ISR is filled with
protons,
where the and
two experiments are carried out intersect
at the eight
counter-circulating beams each regions
other.

For particle physics experiments at low energies, the 3 GeV

antiprotons are decelerated by the PS down to 200 MeV and sent
to the LEAR. Here they are decelerated to an energy of 50 MeV
and are further cooled by stochastic cooling. The antiprotons
are then used for various experiments such as the measurement
of the production rates for some of the more exotic products
of the proton-antiproton annihilation process and the X-rays
emitted by a protonic atom (a proton and antiproton orbiting
about each other just before aqn;hilation) as it drops from
one excited state to the next.j'"
A typical "fill" of the LEAR ring is 1010 antiprotons, and the
beam is used for a number of hours by many different
experiments. Many of the antiprotons, however, do not
interact with the experiments and as many as 109 of them end
up in the beam dump. These "dumped" antiprotons might be a
source for antiprotons needed for experiments on freezing and
storage of antihydrogen for propulsion.

References:
1 .5 R.
Billinge, "CERN's pp source," CERN Publication 84-09,
Proc. Fourth Topical Workshop on Proton-Antiproton Collider
Physics, Berne, 5-8 March 1984, pp. 357-364 (8 August 1984).
1. 6 p.j. Bryant, "Antiprotons in the ISR," IEEE Trans. Nuclear
Sci. NS-30, 2047-2049 (August 1983)
I' 7 K. Kilian, "Physics with antiprotons at LEAR," CERN
Publication 84-09, Proc. Fourth Topical Workshop on Proton-
Antiproton Collider Physics, Berne, 5-8 March 1984, pp. 324-
341 (8 August 1984).

10
 1.2.2 Future Plans at CEN
Since Fermilab has no present plans to decelerate the
antiprotons at their facility, the only source of antiprotons
for the next few years will be the Low Energy Antiproton Ring
(LEAR) at CERN. Thus it is important to know what the future
plans are for antiproton generation, collection, and storage.
The present CERN antiproton facility, although remarkable in
its present capabilities, falls short of the original
expectations of the designers. First, the CERN facility is
limited in its production capability by the energy limits of
the available proton accelerator, the Proton Synchrotron (PS).
The PS has an upper limit to its energy of about 28 GeV. For
producing antiprotons it is operated at a momentum of
26 GeV/c. This is not too far from the antiproton production
• threshold of 8.8 GeV/c. Thus, the initial production is 10
times lower than Fermilab, which will use 120 G-V protons, and
20 times lower than the energy optimum at a proton energy of
200 GeV. There is no reasonable way that the proton energy
can be increased, so CERN is stuck with the present proton
energy and its effect on the production rate.
The PS operates at a current intensity of l.2x101 3 protons for
each 2.4 s cycle. The cycle time per pulse is fixed by the
design parameters of the PS. Efforts are being made to
increase the current intensity to 2x10 1 3 protons/pulse. 1 . 8
The increased beam intensity will have an effect on the choice
of the target, since the present targets are being stressed
close to their limit.
To increase the brightness of the antiproton source, the CERN
engineers are looking at a number of modifications to the
present techniques. One is to reduce the radius of the
primary beam by focusing the protons onto the target with a
magnetic lens. For a fixed acceptance of the Antiproton
Accumulator in mm-mrad, this means that antiprotons can be
captured over a wider angular range.
The decreased beam size and increased current means that the
energy density on the target will be increased. The present
targets are copper, which have been found to be capable of
standing the present energy deposition rates. It may be
necessary to use tungsten targets although the yield per
incident proton is slightly less thanthe yield for copper.
An alternative would be to use rotating or liquid metal
targets.
A second modification being considered at CERN is to pass a
high current through or just outside the target to create an
azimuthal field in and around it. The antiprotons, instead of
spreading out in angle, will tend to follow along the magnetic
*,-
11
field lines until they reach the end of the target. This, in
effect, turns the antiproton source from a rod source to a
disk source.
A third modification is to decrease the focal length of the
magnetic lens following the target so as to be able to collect
antiprotons at larger production angles. Replacing the
present magnetic horn with a lithium lens 1 - is one
possibility being considered. An alternative possibility is
to use a non-linear magnetic lens that selectively collects
some particles at higher energies. Another, which is not
presently being considered by CERN, is to use an array of
lenses with longer focal lengths.
CERN has carried out calculations on a target system using a
prefocusing lithium lens 2 cm in diameter by 15 cm long and
pulsed at 350 kA, a current carrying copper target 1.8 mm
diameter by 17 cm long pulsed at 225 kA, followed by a
collecting lithium lens 4 cm in diameter by 15 cm long pulsed
at 800 kA. This system would multiply the present target
yield by a factor of 22.1.10 The output emittance, however,
would be too big to enter the present AA.
A major disappointment to CERN was that the beam acceptance of
the Antiproton Accumulator did not reach the design goals.
The CERN engineers will continue to look for the source of the
decrease in acceptance and attempt to correct it. Meanwhile
they will bypass the problem by adding an antiproton
preconditioning ring called the Antiproton Collector (AC or
ACOL). The AC will be optimized for collecting the
antiprotons, while the AA will be reworked to optimize it for
stacking and storing of the collected antiprotons.
The AA will be shut down in late 1986 and will be combined in
late 1987 with the AC. The AC will be designed to have twice
the transverse acceptance of the AA in both planes and four
times the momentum acceptance (6%).1.11 The AC will carry out
a phase space "compression" of the incoming antiproton pulse
using a combination of longitudinal and transverse stochastic
beam cooling and bunch manipulation in phase space.
The antiprotons arrive in 5 short bunches and it is possible
to exchange momentum spread against bunch length by using a
technique called bunch "rotation". This process uses a RF
cavity pulsed at 1 MV for a fraction of a millisecond. After
this bunch rotation, the bunches smear out into a continuous
beam and the transverse emittances will be cooled from 200W
down to 3w mm-mrad by fast horizontal and vertical stochastic
cooling systems. Longitudinal cooling will then be applie• 9
reduce the momentum spread of 6% by an order of magnitude.l.ý"

12
 The precooled beam will then be transferred to the AA. The
transfer should be highly efficient since the momentum spread
and the transverse emittances of the preconditioned beam are
now small enough to easily fit into the present AA
acceptances. The stochastic cooling systems in the AA will be
upgraded to handle the higher antiproton flux. Primarily this
means increased bandwidth (2-4 GHz) and increased power

I (10 kW) in the sensing and kicker circuits. This amount of

power is quite expensive in this frequency range. The AA will
also be modified by removing the precooling circuits and
movable shutters so that it will be a reliable machine for
cooling, stacking, and storage of the antiprotons until they
are needed for experiments.

References:
1 .8 E.
Jones, S. van der Meer, R. Rohner, J.C. Schnuriger, and
T.R. Sherwood, "Antiproton production and collection for the
CERN antiproton accumulator," IEEE Trans. Nuclear Sci. NS-30,
2778-2780 (August 1983).
1 9 B.F.
Bayanov, J.N. Petrov, G.I. Silvestrov, J.A.
MacLachlan, and G.L. Nicholls, "A lithium lens for axially
symmetric focusing of high energy particle beams," Nucl.
-Instr. & Methods, 190, 9-14 (1981).
1
" •IIB. Autin, "The future of the antiproton accumulator," pp.
573-582, Proc. Workshop on Proton-Antiproton Physics and the W
discovery," La Plagne (1983).
1"IlS. van der Meer, "Practical and foreseeable limitations in
usable luminosity for the collider," CERN Publication 83-04,
Proc. Third Topical Workshop on Proton-Antiproton Collider
Physics, Rome, 12-14 January 1983, pp. 555-561 (10 May 1983).

13
 "1.3 ANTIPROTON PHYSICS AT FERMILAB
The particle physics facilities at the Fermi National
Accelerator Laboratory consist of the Booster, the 400 GeV
"Main Ring, and the superconducting 1 TeV Tevatron in the same
tunnel as the Main Ring. To collect the antiprotons needed
for p3 collision experiments Fermilab is constructing a Target
Station followed by a Debuncher and an Accumulator shown in
* Figure i-3.1"12 The antiproton source will come on line in
1985 and be operational in 1986. There are no plans for
". decelerating the antiprotons to subrelativistic velocities.
In the Fermilab complex, every two seconds a batch of 2x10 1 2
protons in 82 rf bunches are accelerated by the Main Ring to
120 GeV, then rotated by rf pulsing to convert the long pulses
into short pulses. The short proton bunches strike a 5 cm
long tungsten target, producing 82 equally spaced short
antiproton bunches. The 8 GeV antiprotons near the peak of the
antiproton spectrum are collected by a lithium lens 2 cm in
diameter by 15 cm long, that is pulsed with a current of
500 kA to produce a magnetic gradient of 1000 T/m. The
focused antiprotons near 8 GeV are diverted 30 by a pulsed
dipole magnet and then transported to the Debuncher. The
combination of lens and Debuncher acceptance collects the
7x10 7 antiprotons near 8 GeV within a momentum spread of 3%
and beam emittances of 201Y mm-mrad in each plane.

In the Debuncher the antiproton bunches are rotated so that

the narrow pulse length and the large momentum spread have
been transformed into a small momentum sp:ead and a long pulse
length. After the rf manipulations, the h-rizontal and
vertical transverse emittances are stochastically cooled in
the Debuncher from 20 to 7ir mn-mrad d'iring the two seconds
. before the next antiprotons are to be injected.
The antiprotons are then extracted rr%. the Debuncher and
injected into the Accumulator. Successive batches are
accumulated by rf stacking each batch at the edge of the
stack. Between injection cycles, the stack is stochastically
cooled using a combination of longitudinal and transverse
cooling. Some antiprotons are lost during transfer and rf
stacking, and some diffuse away from the stack into the
chamber walls. Allowing for losses, 6x10 7 antiprotons are
stacked in each pulse. In 4 hours, the core will grow to
4.3xl0II antiprotons. During this time the transverse cooling
systems will have reduced the horizontal and vertical
emittances to 2W mm-mrad.

14

•-14
 DEBUNCHE R

ACCUMLATORPROTON
ACCUMULATOR BOOSTER
8Ge8 8GeV

ý INJECTION
LINE EXTRACTION
LINE

.p PRODUCTION MAIN RING AND TEVATRON

TARGET AND 10GV1 TeV
Li LENS

Fig. 1-3 Antiproton production facility at Fermilab.

.-4. 15
After the Accumulator is full, antiproton bunches of the
desired intensity are individually extracted from the core,
transferred to the Main Ring, accelerated to 150 GeV and
injected into the Tevatron for pf collision experiments (and,
it is hoped, some Nobel prizes in the late 1980's).
Future plans at Fermilab include adding momentum precooling to
the Debuncher, improving the stochastic cooling in the
Accumulator, improving the Main Ring extraction for antiproton
production, and installing intermediate energy electron
cooling in the Accumulator.
The cost of the antiproton generation, collection, and stotage
facility, including the modifications to the existing
accelerators, the pp interaction area, and overhead is 124 M$.
Of this amount, 62.5 M$ is for the building of the antiproton
source.1-3 The oveial cost of the 1 TeV Tevatron is
estimated at 300 M$.

References:
"1 2 Fermilab staff, Design Report: Tevatron 1 Project, Fermi
National Accelerator Laboratory, Batavia, Illinois (September
1984).
1. 1 3 j.p. Marriner, "The Fermilab pp collider," pp. 583-592,
Proc. Workshop on Proton-Antiproton Physics and the W
discovery," La Plagne (1983).
1*1 4Physics Today editors, "Fermilab's superconducting
synchrotron strives for 1 TeV," Search and Discovery Section,
Physics Today, 37, No. 3, 17-20 (March 1984).

16
 1.4 ANTIPROTON PHYSICS AT IHEP
High-energy particle physics in the USSR is carried out at the
Institute for High Energy Physics (IHEP) in Serpukhov,
Novosibirsk. The major machine is the U-70 proton synchrotron
with a maximum energy of 70 GeV with a beam intensity of
7x10 1 2 protons/cycle and a cycle duration of 7 s. This
machine has been used for particle physics experiments
including measurements of the production spectrum of hadrons
(including antiprotons) in the collision of 70 GeV protons
with heavy nuclei including Al, Be, Cu, C, Sn, and Pb. 1 - 1 5

The future plans for the particle physics research at IHEP are
to construct by 1989 an electron-positron collider (VLEPP)
with an initial center of mass energy of 300 GeV. This will
later be upgraded to 1000 GeV. By 1990 it is planned to have
one ring of the Accelerating and Storage Complex (UNK)
operating with protons and antiprotons with a center of mMs6
energy of 6000 GeV (6 TeV) and a 3xl0o3 /cm 2 's luminosity.
The UNK tunnel will be 19.3 km in circumference (6 km
diameter) and contains two rings. The first uses conventional
magnets and will accelerate the 70 GeV protons from the U-70
to 400 GeV. A second ring will use superconducting magnets
and provide energies from 400 to 3000 GeV. The 400 and
3000 GeV proton beams can be collided to provide a center of
mass energy of 2.2 TeV. A third superconducting intersecting
ring will be added later to provide proton-proton collisions
i" at 6 TeV withm
~1032/cm2-s..1 substantially higher luminosity of

The antiproton source for the UNK will use the 70 GeV protons
from the U-70 machine. A new booster for the U-70 will
increase the beam intensity to 5x101 3 protons/cycle. The
proton beam is focused down to 0.5 mm by a 100 kG lithium lens
5 mm in diameter Pnd 10 cm long. The peak of the antiproton
spectrum is about 5.5 GeV. The antiprotons from the target
are collected by a 170 kG lithium lens 2 cm in diameter by
15 cm long. This lens is able to collect antiprotons with a
linear angle of 0.1 rad (solid angle of 0.0314 sterrad).
As shown in Figure 1-4, the antiprotons go to the synchrotron-
decelerator where rotation of the antiproton bunches decreases
their momentum spread. This is followed by deceleration of
the antiprotons to 400 MeV. The antiprotons are then sent to
the cooler-accumulator where they are cooled and stored. The
cooler-accumulator is in a race-track configuration with two
half-rings with a radius of 40 m at the end and two straight
sections 100 m long. Electron beam cooling is used in the
cooler-accumulator. A 218 MeV electron beam is generated in
one end, travels around the track and is collected near the
source. With the source and the collector at the same
potential, most of the electron beam energy is recycled. 1 ' 1 8

17
 - ------.

p 70 GeV

PROTON
U-70 p"TARGET AND LENSES
SYNCHROTRON
P 5.5 GeV

SYNCH ROTRON
P DECELERATOR

p P 400 MeV
ANTIPROTON LOOP

ELECTRON LOOP

COOLEr,-ACCUMULATOR
STRAIGHT
COOLING
SECTIONS
(100 m)

e- e-

ELECTRON SOURCE - •.L, ELECTRON

COLLECTOR

ANTIPROTON
LOOP

Fig. 1-4 Antiproton production facility at IHEP.

18

= * *.
 The 400 MeV antiprotons are injected into the cooler-
'•C'.. travel
accumulator ring in the straight section where they coulomb
along with the 218 MeV electrons and are cooled by
interaction with the low temperature electron beam. At the
ends of the track, the antiprotons, with their greater
momentum, exit from the ring to be turned around by the
antiproton bending magnet channel and sent back in the other
straight section for further cooling.
There are several plans for future improvement of the
antiproton source.I'18 One concept increases the proton beam
intensity by stacking up 7 bunches of protons by sending
sequential bunches to delay lines of differing lengths. A
second concept involves decreasing the momentum spread of each
bunch and stacking four bunches in the length of one present
bunch. A third concept uses a multiple target system where
'the proton beam passes through several targets in succession.
The antiprotons produced in a target (and other targets
upstream) are refocused on the next target by a lithium lens.
Thus, the "image" of all the targets are superimposed. If
successful, these concepts and others could improye the
antiproton production rate by a factor of 140. *' The
production rate would then be 6x10 9 1/s or 300 ng/year.
It is interesting to note that there have been no significant
publications on antiproton sources by USSR scientists since
1981.

References:
I'I5v.v. Abramov, et al., "Production of hadrons with
transverse momentum 0.5-2.5 GeV/c in 70-GeV proton-nucleus
collisions," Soy. J. Nucl. Phys. 31, 343-346 (1980).
1- 1 6 B.E. Balakin and A.N. Skrinsky, "Project VLEPP," Akademiya
Nauk USSR, Vestnik, No. 3, 66-77 (1983).
1- 1 7 A.I. Ageyev, et al., "The IHEP accelerating and storage
complex (UUIK) status report," pp. 60-70, Proc. 11th Int. Conf.
High Energy Accelerators, Geneva (1980).
1 ' 1 8 T.A.
Vsevolozskaya, B. Grishanov, Ya. Derbenev, N.
Dikansky, I. Meshkov, V. Parkhomchuk, D. Pesrikov, G.
Sil'vestxov, A. Skrinsky, "Antiproton source for the
accelerator-storage complex, UNK-IHEP," Fermilab Report FN-353
8000.00 (June 1981), a translation of INP Preprint 80-182
. (December 1980).

19
1.5 COMPARISON OF ANTIPROTON PRODUCTION FACILITIES

The three antiproton production facilities are compared in

Table 1-1. Fermilab has the highest proton energy, while CERN
partially makes up for the lower production at its low proton
energy by having over five times as many protons per second
hitting the production target. CERN was the first to be
operational and had some problems meeting the initial goals
in angular and momentum capture. The numbers for Fermilab and
IHEP are still design goals and it will be interesting to see
how close the actual machine performance comes to the goals.
If the Fermilab and IHEP machines achieve their production and
capture goals, then the production rate for the number of
antiprotons captured for every incident proton will be 40 to
110 times that of the present CERN accomplishments. Of
course, with the additicn of the Antiproton Collector in 1987,
CERN hopes to close that production gap somewhat. Still, in
all the machines, the production rate is only a few parts in a
million antiprotons per proton. When we calculate the
efficiency in terms of energy rather than number of particles
the differences between the machine efficiencies become
smaller because of the higher energy in the initial protons in
the FNAL and IHEP machines.

The energy efficiency for the machines shown in the last line
of Table 1-1 is the ratio of the energy that would be obtained
by annihilat-n of the captured antiproton with a proton
(2mc 2 ), divided by the beam energy of all the protons that it
took to generate that one captured antiproton. As we can see,
the energy efficiency from proton beam energy to annihilation
energy varies from 3x10 8 for CERN to 1.3x10- 6 for IHEP. If
we then assume that a typical synchrotron has an energy
efficiency from ac power to beam power of 5% (or less), then
the present total energy efficiency for producing antiprotons
is only a few parts in a billion.

It is this extremely low production efficiency for the CERN

machine that has led many experts to categorically state that
the capture and storage of antiprotons for later use as an
energy source is pure science fiction and will never become
practical. The following sections discuss each of the various
inefficiencies in the present machine designs that lead to
this parts-per-billion total efficiency and outline what hopes
there are to improve the efficiencies in each area to bring
the total production efficiency up to some more reasonable
number, like a part in 10-4.

20
"" Table 1-1 Comparison of Antiproton Production Facilities.

CERN FNAL IHEP

(EUROPE) (USA) (USSR)
PROTON ENERGY (GeV) 26 120 70
PROTONS/CYCLE (x 1012) 13 2 7
CYCLE DURATION (sec) 2.4 2 7
ANTIPROTON ENERGY (GeV) 3.5 8 5.5
ANGULAR CAPTURE (mm - mrad) 70 7" 201r 60r
MOMENTUM CAPTURE (CP/P) 1.5% 3% 6%
ANTIPROTONS/PROTON (x 10-6) 0.4 30 46
E = 2 x ii ENERGY/p ENERGY (x 10"7) 0.3 4 13

1.6 EFFICIENCY OF PRESENT ANTIPROTON PRODUCTION FACILITIES

The capture efficiencies of the present antiproton facilities
are abysmally low. The present magnetic lenses and collector
rings capture only a very small fraction of the antiprotons
that are generated in the target. The reason for this is
sammarized by Figure 1.5 from a recent papejron the efficiency
of the proposed Fermilab collector system.
The upper part of the figure shows the total number of
antiprotons generated per GeV of antiproton momentum per
steradian of solid angle at the central portion of the
antiproton beam. Integrating the curve over the antiproton
momenta shows that each proton produces 7.7 antiprotons per
steradian. The half-width of the angular spectrum is unknown
since the angular spectrum has never been measured.
If the magnetic lens had a wide enough angular capture
(<100 mrad), then probably that would be enough to capture
nearily all the antiprotons. Their total number is estimated
at 4.xlO- 2 antiprotons per proton (see Section 2.1). The
estimated momentum spectrum of these antiprotons is the middle
curve in Figure 1-5.

21
d3 N NUIVBER OF ANTIPROTONS PER 120 GeV PROTON (5/p)
dPdnS

N = 7.7 /P "
10-1

10-2

1O- 3 N 0.047'Ip
dN
dP N = 0.014j/p
/ /_ • •--AP 3% (0.003 ster.)

10 4.- AIP = 0.25 GeV.

Nth = 1.8 x 10- 4 /p

(Nexp = 3.0 x 10- 5 i/p)

-5I I I
0 5 10 15 20 25 30
ANTIPROTON MOMENTUM P, GeV/c

Fig. 1-5 Present antiproton capture efficiencies.

22
 The Fermilab magnetic lens, as presently designed, is going to
"•. have an angular capture range of 30 mrad. Thus the spectrum
of antiprotons it can capture is shown by the bottom line of
Figure 1-5. When the 30 mrad curve is integrated over the
antiproton momenta we find a total of only 0.014 antiprotons
per proton in this narrow angular acceptance. This is only
30% of the total number of antiprotons generated by the
target.
Then, of this small angular spread the Fermilab collector-
debuncher is only able to capture those with a momentum
(velocity) spread of 3% or 0.27 GeV/c around the momentum
design center of 8.9 GeV. Even under ideal conditions, with
this momentum capture bandwidth, they would expect to capture
only about 1.8x10- 4 antiprotons per proton. This is only 1.3%
of the protons that were collected by the lens and only 0.4%
of those generated in the target. In practice, the Fermilab
group expects that after target losses, transfer losses, and
other losses they will only get 1/6th of the ideal figure and
are hoping to ultimately collect 3x10-5 1/p or 0.06% of the
antiprotons generated in the target.
It is obvious that to increase the efficiency of antiproton
production, one of the areas needing improvement is the
angular and especially the momentum capture efficiencies of
the collection system. We will discuss these further in
Section 3.

References:
1 "1 9 Hojvat,
C., and Van Ginneken, A., "Calculation of
Antiproton Yields for the Fermilab Antiproton Source," Nuclear
Instrumentation and Methods, Vol. 206, 1983, pp. 67-83.

23
 SECTION 2
SOME ANTIPROTON PRODUCTION FUNDAMENTALS

In this section we will expand upon some of the fundamental

* techniques for the generation, capture, cooling, and trapping
of antiprotons that were discussed briefly in Section 1.

2.1 GENERATION OF ANTIPROTONS

At present, the best known technique for producing antiprotons

is to send a beam of high-energy protons into a metal target.
The protons strike the protons and neutrons in the nuclei of
the target and various reactions occur which produce gamma
'ays, pions, kaons, and baryons, including antiprotons.
Although the absolute production rate of antiprotons is
unknown to roughly a factor of two, the relative production
rate of antivrotons to the total inelastic cross section is
known and is shown in Figure 2-1.2-1
The curve is a fit of a specific model to the available
experimental data for the relative cross section at the peak
in the forward direction. There are two curves, one for
protons hitting a hydrogen target and the other for a lead
target. The lead data is not significantly different from the
data for other heavy metal target nuclei. The total inelastic
cross section is dominated by the pion production. At 200 Gev
the ratio of the antiproton production cross-section to the
total inelastic cross-section is about 0.04. This low
production ratio is one of the major factors in the total
production efficiency and cannot be increased by engineering
improvements.
The best available data on the total antiproton production
rate is a 1973 paper 2 - 2 based on p-p reactions in colliding
beam experiments. Because the colliding beam experiments are
symmetric in the laboratory frame, all the emitted particles
have the same angular distributions, and the production ratios
do not have the angular biases that occur when measurements
are made using protons hitting a stationary target nucleus.

A typical measurement made at a center of mass squared energy

s=485 GeV 2 (equivalent to a proton hitting a stationary target
nucleon at an energy of 258 GeV), gave 3.56 W", 2.98 ,-,
3.5 v, 0.35 K+, 0.24 K, 1.28 p', and only 0.061 p. Thus
for every antiproton emitted, there are ab'ut 165 pions and 10
kaons.

25 ,,,VotS-PAGE

~SLANK
10-2

10-2

10-

10 100 1000 10,000

PROTON ENERGY, GeV

Fig. 2-1 Relative antiproton production rates

for hydrogen and heavy metal targets.

26
 A plot of the available data on the total antiproton
multiplicity of the reaction p-p is shown in Figure 2-2. The
dots are the available data points and the curve is a smooth
fit to the data. If we now take Figure 2-1, which shows the
difference between the antiproton production rates in hydrogen
and a heavy metal, we can use the ratio of p-Pb to p-p at each
energy to make an estimate for the total antiproton production
for protons striking a stationary heavy metal target. This is
the top curve in Figure 2-2. It must be emphasized that this
estimate is not very accurate. It could be off by as much as
a factor of two. Note that at 200 GeV, the total antiproton
production rate in L.lead target is 0.085 antiprotons per
incident proton. This is a factor of two less than an
estimate mae 3 in a previous report on antiproton
propulsion.2"

Now that we have an estimate of the total production, we can

estimate how well present antiproton collectors perform in
collecting the total number of antiprotons generated. If the
target is long, there will be some absorption of the
antiprotons in the target, but this is compensated somewhat by
additional antiprotons produced by secondary reactions of the
emitted particles from the first p-N collision with other
target nuclei. For example, the secondary reaction
antiprotons are 22% of the total antiproton f ux for a 6 cm
tungsten target bombarded by 120 GeV protons.1-l The
antiprotons come out of the target with a spread in angle and
momentum. As the energy of the incident proton beam is
increased, the total number of antiprotons produced increases,
and the a ,le into which they are concentrated becomes
smaller, making it easier to collect them, but the spread in
longitudinal momentum becomes larger, requiring that the
collector accept a larger momentum spread to capture the same
percentage of generated antiprotons.

References:
2.1C. Hojvat and A. van Ginneken, "Calculation of antiproton
yields for the Fermilab antiproton source," Nucl. Inst. &
Methods 206, 67-83 (1983).
2 . 2 M. Antinucci,
A. Bertin, P. Capiluppi, M. D'Agostino-Bruno,
A.M. Rossi, G. Vannini, G. Giacomelli, and A. Bussiere,
"Multiplicities of charged particles up to ISR energies,"
Lett. Nuovo Cimento 6, 121-127 (1973).
2 " 3 R.L.
Forward, AltErnate Propulsion Energy Sources, p. 1-8,
AFRPL-TR-83-067, Final Report on Contract F04611-83-C-0013,
Air Force Rocket Propulsion Lab, Edwards, CA 93523 (Dec 1983)

27

'."• -''-'', • ' " " " " . ' "- ."i.. ." ",••' " • •".•""","•".. " •..".........\•'•. ,'••
"" . ,• .•
' •.'•
• •• .'• ."•"• '•
 °''¾

.%~

17 - I 1 1 1 I I I1 I 1 I 1IIL1I

ESTIMATED
TOTAL p-Pb
z
0 MEASURED p-p
0
C 10-11
C-
2
I--
0

* 0
* 10-2

1--3

10 100 1000 10,000

PROTON ENERGY, GeV

Fig. 2-2 Total antiproton production rates.

28

• -. -."=''L-•-
-. -. • - • - -. •. -•- -• `• -• • > -•- °••- *. 'a - * - ~• - "- • •N*--- - • -
 .N=A_7.YWZ 't ~

2.2. ANGULAR CAPTURE BY MAGNETIC LENSES

We will now look at the problem of angular capture. The spray

of antiprotons coming from the target are focused by a
magnetic lens into the aperture of the collector ring that
will manipulate, cool, and accumulate the antiprotons.
The standard lens used in particle accelerators is a magnetic
quadrupole. It consists of four long magnetic pole pieces
wound with coils of alternating polarity just outside the beam
transport tube. The magnetic quadrupole focuses in one axis
while causing a slight defocusing in the orthogonal axis. A
series of quadrupoles of orthogonal orientation, however, can
produce a net focusing of the beam in all orientations. The
off-axis aberrations of a quadrupole magnetic lens are severe.
To obtain symmetric focusing, it is necessary to produce a
magnetic field along the beam with a radially increasing
gradient. So far, the only meth)ds found to achieve this
field configuration have required that the particle beam pass
through a portion of the current conductors.

Conceptually, the lithium lens is the simplest of the

symmetric lenses, although its development came later. These
lenses consist of long lithium cylinders placed directly in
the particle beam path that carry a large pulsed current from
one end to the other. Provided the diameter of the lithium
lens is not too much greater than the electromagnetic skin
depth at the frequencies corresponding to the pulse width, the
current distribution in the lithium cylinder is uniform. A
uniform current distribution gives an azimuthal magnetic field
within the conductor that increases linearly with radius,
giving a constant radial gradient.
A typical lens designed for focusing an 80 GeV proton beam to
a spot size of 0.1 mm radius is 5 mm in diameter, 10 cm long,
and is excited at a 13 Hz repetition rate by a 60 jus pulse of
160 kA to produce a magnetic field of 10 T (100 kG) and a
magnetic field gradient of 4000 T/m. 2 - 4 Because the beam must
go though the material in the lens, the magnetic design of the
lens is compromised by requirements that the beam not destroy
the lens and the lens not destroy the beam. This is
especially critical in the design of a lens for antiprotons.
For example, a longer lens needs less drive current but
absorbs more antiprotons.

Since the metal targets for antiproton production are also

long and thin, they can be pulsed with current to provide
self-focusing. 2 * 5 Although this has not been implemented yet
in any system, it is an obvious next step since the target has
to be in the beam anyway.

29
 The lenses presently in use at CERN, Fermilab, and elsewhere
are parabolic horns. 2 - 6 These are usually made of aluminum,
or beryllium copper, and consist of two elongated parabolic
horns connected at the center at the narrowed down
"mouthpiece". The wall thickness varies inversely with
distance from the axis and is 0.5 mm at the maximum radius at
the bell of the horn. The current runs from one horn rim
through the connection at the "mouthpieces" to the other horn
rim. There is no magnetic field inside the horn, while the
magnetic field outside the horn drops off as 1/r. The center
of the beam passes through the center axis of the horn, but
any off axis particles pass through the conductor of the horn
to enter the region containing the magnetic field. The
parabolic shape of the horn means that a particle a distance r
from the axis will travel a distance r2 outside the horn,
giving a net magnetic interaction that varies as r. 2 " This
constant magnetic gradient force then bends the particle back
to the focal point. Real (thick) lenses have horn shapes that
are not exactly parabolic and are different sizes at the entry
and exit ends.
After the antiprotons have been focused by the magnetic lens
they are diverted a few degrees by a pulsed dipole magnet that
selects negatively charged particles with energies near the
peak of the antiproton production spectrum. The antiprotons
not selected, the remnants of the incident protons, and the
other interaction products continue on towards the beam dump.
The selected antiprotons pass through a channel in the beam
dump into a transport line leading to the collector ring.

References:
2 " 4 B.F.
Bayanov, J.N. 1'etrov, G.I. Silvestrov, J.A.
"- MacLachlan, and G.L. Nicholls, "A lithium lens for axially
symmetric focusing of high-energy particle beams," Nucl.
Instr. & Methods, 190, 9-14 (1981).
2 . 5 B.
Autin, "The future of the antiproton accumulator," pp.
573-582, Proc. Workshop on Proton-Antiproton Physics and the W
discovery, La Plagne (1983).
2 - 6 B.F. Bayanov, A.D. Chernyakin, V.N. Karasyuc, G.I.
Sil'vestrov, T.A. Vsevolozhskaya, V.G. Volohov, G.S.
Willewald, "The antiproton target station on the basis of
lithium lenses," pp. 362-368, Proc. llth Int. Conf. High
Energy Acceleratozz, Geneva (1980).
2" 7 T.A. Vsevolozskaya and G.I. Sil'vestrov, "Optical
properties of fast parabolic lenses," Zh. Tekh. Fiz. 43, 61-70
(1973) [English translation Soy. Phys. Tech. Phys. 18, 38-43
(1973)].

30
2.3 MOMENTUM CAPTURE BY COLLECTING RINGS

As was pointed out in Section 1.6, one of the major reasons

"for the low efficiencies in the present antiproton production
facilities is the low momentum capture capabilities of the
present collecting rings. The efficiency of a collecting ring
is usually given in terms of dp/p or the momentum spread dp
accepted by the ring divided by the center momentum p of the
ring (usually the peak of the antiproton spectrum).
Siace the antiproton spectrum is presently unknown, and the
spectrum is usually modified by the magnetic lens, the use of
the present figure of merit is understandable, but it usually
implies a better capture efficiency for the ring than it
really delivers. A better figure of merit would be the
momentum spread accepted by the ring divided by the half-width
Dp of the antiproton spectrum. For example, at CERN,
dp/p=l.5%, but dp/Dp is significantly less at 1%. At
Fermilab, the quoted dp/p=3% for the capture portion of the
debun-her ring means that at a capture momentum of p=8.9 GeV/c
the moi.'entum bite is dp=0.25 GeV/c. Since the estimated half-
width oi the total antiproton spectrum is Dp=22 GeV/c, the
real capture efficiency is only dp/Dp=l.2%. Thus, though the
Fermilab collecting ring has a quoted momentum capture
efficiency that is twice that of CERN, the actual momentum
capture efficiency is only 20% better.

It is very difficult to build a ring that can accept a wide

spread in the momentum of the particles. Particles with
different momenta have different radii of curvature in the
magnetic field of the bending magnets. If the magnetic field
strength is adjusted so that the radius of curvature of a
particle is equal to the radius of curvature of the beam line,
then a particle with a different momentum will have a
different radius of curvature and hit the wall of the beam
line.
Conceptually, one could make a beam line that is much wider
than it is tall so that it could hold particles with a wider
spread in momentum. This is in fact done, but there is a
practical limit to the width imposed by a combination of
pressure forces on the top and bottom of the vacuum line, the
"amount of vacuum volume to be pumped, and the size of the
magnets. The present state-of-the art in collector rings is
represented by the IHEP Synchrotron Decelerator and the new
CERN Antiproton Collector, which both have a dp/p=6%.

31
2.4 ELECTRON COOLING OF ANTIPROTON BEAMS

In carrying out almost any experiments with beams of charged

particles, it is of fundamental importance to be able to
compress the beams, to decrease the size of the beam as well
as narrow the momentum spread of the particles in the beam in
both magnitude and direction. In other words, it would be
desirable to "cool" the flux of fast charged particles,
lowering their effective "temperature" about the average
motion of the beam.

In electron cooling of an antiproton beam, a beam of

monoenergetic electrons with a velocity equal to the average
velocity of the antiprotons is made to travel along with the
antiproton beam.2-8 Then, in the rest system of the
electron beam, the "hot" antiproton gas is within the "cold"
electron gas and is cooled in all directions as the result of
Coulomb collisions. As a result, the phase space of the
antiproton beam decreases in all degrees of freedom and the
beam is compressed. The compression continues in principle
until the antiproton temperature in the center-of-mass system
becomes equal to the electron temperature. Since the
antiprotons are much heavier than the electrons, the resultant
decrease in the velocity spread for the antiprotons goes as
the square root of the mass ratio. For an easily achieved
velocity spread ratio of dp/p=10- 3 for the electrons, the
resultant velocity spread for the antiprotons is 2x10- 5 .
Electron cooling has been demonstrated at CERN 2 - 9 and
Fermilab 2 -1 0 using protons and electrons. The recombination
reaction rate is small enough that this gives a good
simulation of electron cooling of antiprotons. At Fermilab a
200 MeV proton beam with an initial momentum spread of
dp/p=3xl0- 3 was cooled with a 110 key beam of electrons with a
energy spread of 1 eV transverse and 8x10- 5 eV longitudinal.
With 1 A of electrons, the protons reached a momentum spread
of dp/p=1xl0- 5 in 4 s of cooling. Transverse cooling by a
factor of 25 was also achieved in less than 20 s of cooling
time. 2* 1 0 Electron cooling can be made very efficient. The
electrons can be circulated in a ring and "cooled" using a
"wiggler" section 2 -1 1 or up to 98% of the electron energuyvan
be recovered using a well designed depressed collector.

Despite the experimental success of electron cooling, it is

not used much in practice. It was originally felt that
stochastic cooling would be most effective when the
antiprotons were at relativistic energies, while electron
cooling would be more effective at subrelativistic energies.
It has turned out that stochastic cooling has been found to be
effective at all energies of present interest.

32

* 't- *.'>"b b* A ~ *'

"Electron cooling has also been found to have catastrophic
failure modes. For this, as well as other reasons, stochastic
cooling is usually the choice for cooling of antiproton beams.
This may change as new machines such as the Extra Low ENergy
Antiproton (ELENA) ring are built to take the 5 MeV 2 13
antiprotons from LEAR and decelerate them down to 200 keV. "

References:
2 " 8 G.I.
Budker and A.N. Skrinsky, "Electron cooling and new
possibilities in elementary particle physics," Usp. Fiz. Nauk
124, 561-595 (1978) [English translation Soy. Phys. Usp. 21,
277-296 (1978).
2. 9 M. Bell, J. Chaney, H. Herr, F. Krienen, P. Moller-
Petersen, and G. Petrucci, "Electron cooling in ICE at CERN,"
Nuclear Instr. & Methods 190, 237-255 (1981).
2- 1 0 T. Ellison, W. Kells, V. Kerner, F. Mills, R. Peters, T.
Rathbun, D. Young, P.M. McIntyre, "Electron cooling and
accumulation of 200-MeV protons at Fermilab," IEEE Trans.
Nuclear Sci. NS-30, 2636-2638 (1983).
2. 1 1 H. Herr and C. Rubbia, "High energy cooling of protons and
antiprotons for the SPS collider," pp. 825-829, Proc. llth
Int. Conf. on High Energy Accelerators, Geneva, Switzerland
(1980).
2 . 1 2 L.
Hdtten, H. Poth, and A. Wolf, "The electron cooling
device for LEAR," pp. 605-618, Physics at LEAR with Low-Energy
Cooled Antiprotons, Workshop on Physics at LEAR with Low-
Energy Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May
1982, U. Gastaldi and R. Klapisch, ed., Plenum Press, NY
(1984).
H. Herr, "A small deceleration ring for extra low energy
antiprotons (ELENA)," pp. 633-642, Physics at LEAR with Low-
Energy Cooled Antiprotons, Workshop on Physics at LEAR with
Low-Energy Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May
1982, U. Gastaldi and R. Klapisch, ed., Plenum Press, NY
(1984).

33

V -
2.5 STOCHASTIC COOLING OF ANTIPROTON BEAMS

It is not possible to compress a beam of charged particles by

any known combination of external electromagnetic forces which
only act on the beam as a whole and which do not depend upon
the motion of the individual particles of the beam. By means
of focusing and acceleration in any combination one can only
change the shape of the phase space occupied by the beam
particles but cannot change its magnitude or increase the
phase density. This is a special case of Liouville's theorem,
that the density of the particles of a beam in six-dimensional
phase space (the space of generalized coordinates and
conjugate momenta) is a constant and is determined by the
initial conditions.
Stochastic cooling is the damping of the momentum spread and
the transverse oscillations of a particle beam by an
electronic feedback system. In its simplest form, a pickup
electrode detects the instantaneous transverse position or
momentum of a fluctuation in the average current of the
particles in a storage ring due to particle bunching. The
signal is amplified, delayed, and phase shifted, then sent
across the ring to a kicker which applies a force to the bunch
that is designed to damp out the fluctuation.
Stochastic cooling is not a violation of Liouville's theorem.
It works only if the number of particles available is small
and there is a lot of empty phase space. During stochastic
cooling, a "capsule" of phase space is formed around each one
of the particles. Liouville's theorem then applies to the
phase space inside those capsules. The capsules are then
moved around freely in the rest of the (empty) phase space and
brought closer together to achieve a denser packing. The
phase space around each particle is the same, except the empty
phase space has been removed. Although there is an apparent
violation of the Liouville theorem, there is no violation in
reality because the initial phase space was so sparsely
populated.
If there were only a single particle in the ring, it is
obvious that the transverse oscillations and momentum offset
of the particle could be damped using electronic feedback. In
stochastic cooling of many particles in a beam, the cooling of
a single particle is hampered by the presence of the other
particles in the dense cool core, which create a noise signal
(Schottky noise) that heats the single particle. For a
properly designed system, however, the net effect over many
turns is that cooling is achieved. 2 - 1 4
In addition to the Schottky noise created by the "cool"
particles in the beam, another major noise source is the
thermal noise generated by the resistive terminations in the

34
 "pickups and the noise figure of the preamplifiers. A series
of notch filters are used to protect the cool dense core of
the stacked particles from the broadband thermal noise. What
is typically done is to have two cooling systems, one of which
operates at low gain appropriate for cooling of the dense core
and another with the high gain needed to manipulate the newly
injected antiprotons in the low density part of the stack.
The high gain system has a number of notch filters which
prohibit power from being transmitted at th e harmonics of the
* revolution frequencies of the dense core. To decrease the
* thermal noise contribution of the pickup terminations, the
"termination resistors2 at CERN are now cooled to cryogenic
temperatures (18 K). 6

In the Antiproton Accumulator at CERN, the stochastic

precooler system has a signal bandwidth from 150-500 MHz,
192 pickups with 2 db noise figure preamplifiers, 50 power
amplifiers that operate at 2 kW total, and 200 kickers. They
are able to reduce the momentum spread from a dp/p of 1.5% to
0.17% in just 2 seconds. The horizontal and vertical
transverse cooling systems operate over a bandwidth from 1 to
2 GHz and in 15 to 30 minutes can •e~ce the transverse
emittance from 100w to 5w mm-mrad.K'1"
In the Fermilab Debuncher, the stochastic precooler is
designed for transverse cooling only, since the debunching
process will take care of the momentum compression. It will
operate over a bandwidth of 2 to 4 GHz, use 128 pairs of
pickups and kickers, driven by 8 microwave traveling wave
tubes operating at 500 W, to reduce the e~iiance from 20vr to
* <5Wmm-mrad in the 2 s system cycle time.
Stochastic cooling has a significant energy requirement. At
CERN the power requirement for the stocha5týJ cooling in the
Antiproton Accumulator is 1/4 mW per §/s.' Since the AA
handles l.4x10 6 p/s, the total power requirement on the
stochastic cooling electronics is 350 W of expensive microwave
power. Unlike decelerating a particle beam, where the
coherent energy of the beam can actually be extracted out of
the beam slowing apparatus as rf power, stochastic cooling is
working with the random, incoherent energy differences between
the particles and requires the input of energy. This is
because stochastic cooling uses an active damping technique in
which the energy difference between initial energy and the
final energy of the antiprotons in the beam is "cancelled" by
the insertion of an equal amount of energy from the
electronics. The need for extensive amounts of stochastic
cooling energy will be an important cost factor in the design
of a facility for producing large numbers of antiprotons.

35

-°• • .- • • •• • .- • . - °- -••- •--.-*

*-- c''*-
- •.% . -•'t-°. U •- - . - ""• -- -
 - T_

* References:
2"1 4 Fermilab staff, Design Report: Tevatron 1 Project, p. 4-13
to p. 4-16, Fermi National Accelerator Lab, Batavia, Illinois
(September 1984).
R.P. Johnson and J. Marriner, "Stochastic stacking without
filters," Fermilab 1 Note 226, Fermi National Accelerator Lab,
Batavia, Illinois (17 August 1982).
2- 1 6 G. Carron, R. Johnson, S. van der Meer, C. Taylor, and L.
Thorndahl, "Recent experience with antiproton cooling," IEEE
Trans. Nucl. Sci. NS-30, 2587-2589 (1983).
2.17S. van der Meer, "Stochastic cooling in the CERN
antiproton accumulator," IEEE Trans. Nuclear Science, NS-28,
1194-1998 (1981).
2 "1 8 R. Billinge, CERN, Switzerland (personal communication).

2.6 STORING OF ANTIPROTONS

Once the antiprotons have been slowed down to subrelativistic
energies they can be stored as ions for long periods of time
(many days) in various traps using a combination of electric,
magnetic, electrodynamic, gravity, and/or inertial fields.
The present technique is to build a magnetic storage ring
using a combination of inertial forces and static magnetic
fields that keep the antiproton ions circulating about the
ring. The low-energy antiproton ring (LEAR) at CERN is the
best example of this type of trap.
For antiproton annihilation propulsion this type of trap has
two major problems. Because of the magnets and vacuum system,
the present designs are bulky and heavy. It might be
possible, however, to produce a compact design with the
antiproton beam wound up into an endless spiral somewhat like
a continuously playing magnetic tape cartridge.
The other problem is the relatively low ion current the
present machines can carry. A phenomenon called intrabeam
scattering, which is the particle beam version of space charge
forces, puts a practical upper limit of about 1012 ions to the
number of antiproton ions that can be stored in present
machine designs. At about that number of ions the expansion
of the beam due to intrabeam scattering is just compensated by
the shrinking of the beam from the stochastic cooling systems.
Obtaining beam control forces from the stochastic cooling
process is very costly since it requires wide bandwidth
microwave power generators.

36
A number of experimenters have now proposed to decelerate the
antiproton ions down to almost zero velocity and put them into
a Penning trap. 2 -1 9 -2- 2 1 The Penning trap uses only static
electric and magnetic fields and can trap nonmagnetic charged
ions. 2 - 2 2 A properly constructed trap kept at cryogenic
temperatures is completely stable and can hold one or more
ions for long periods of time.

A Penning trap has demonstrated it can even hold antimatter.

A trap at the University of Washington has kept a single
positron trapped for a month. 2 -2 3 This indicates that the
number of residual normal matter gas molecules in a
cryogenically cooled ultrahigh vacuum Penning trap is
extremely small. How good the vacuum is, however, is unknown,
since the cross section for the annihilation of a positron
with the bound electron on a slowly moving atom is unknown.
The present estimates are that a UHV sealeý1 Penning trap at
5 K will have a vacuum of <<10-14 torr. u

As is shown in Figure 2-3, the Penning trap consists of a

small trapping region containing an axial magnetic field and
bounded by three electrodes. Therc is a central ring
electrode with a hyperbolic inner surface shape and two cap
electrodes, also with carefully machined hyperbolic shapes.
For the storage of negative ions (negatively charged
antiprotons), the ring electrode is positively charged and the
cap electrodes are negatively charged.
When a charged ion is inserted into the Penning trap, it has
energy and in general is moving both radially and axially.
The radial motion in the strong magnetic field is converted
into a tight circle that is much smaller than the dimensions
of the trap. This is the cyclotron motion shown in the
diagram above the drawing of the Penning trap in Figure 2-3.
The axial motion of the negatively charged antiprotons is
Lestrained by the repulsion of the negatively charged cap
electrodes. This sends the antiprotons back through the
center of the trap to the other electrode, causing the axial
oscillations shown in the upper diagram of Figure 2-3. The
positively charged ring electrode attracts the antiprotons
radially outward. But as the antiprotons attempt to drift
outward, their motion is curved by the magnetic field into the
large circle shown as the magnetron motion in Figure 2-3.
Thus, if the initial energy of the antiproton is small enough,
the antiproton is trapped.

37
 CYCLOTRON MOTION
/ ~MAGNETRON
MOTION
CL
RESISTOR

MAGNETIC FOIELD

Fig. 2-3 Penning trap for antiproton ions.

38
The motion of the antiproton between the electrodes causes
image voltages in the electrodes, which cause currents to flow
between the two cap electrodes, and between each of the cap
electrodes and the ring electrode. As shown in Figure 2-3, a
resistor placed between the cap electrode and the ring
electrode will dissipate these currents, extracting energy out
of the antiproton. The damping from this external resistor is
strong enough that positrons inserted into the trap 2.
immediately lose enough energy to stay in the trap. 2 2 4 This
approach is not suitable with externally injected antiprotons,
however, because the damping rates go inversely as the mass of
the trapped particle. It will be necessary to vary the
trapping voltages to trap the antiproton initially, then using
the damping resistor for further cooling.

If the resistor is kept at cryogenic temperatures, the amount

of random Johnson noise voltage that it generates is kept
small so that it doesn't excite the antiproton. Thus, if the
resistor is kept in a cryogenic bath, the energy in the
antiproton will be damped out over a period of a number of
seconds until its kinetiý gerq2Ay "s equivalent to the
temperature of the bath. 9I In this manner, the last
bit of kinetic energy can be extracted from the antiproton
without touching it.
"An alternative cooling method is to use an electron buffer
gas, trapped in the center of the same trap and cooled to
cryogenic temperatures via coupling to an external
circuit. 2 . 2 0 Antiprotons oscillating through this electron
buffer gas will scatter from the cold electrons and transfer
energy to them and thus to their external resistor damping
circuit. The cooling time is estimated to be 5 s.

The limit to the number of antiproton ions that can be stored

in such a trap is determined by the space charge limit.
Estimates of the number density vary from 1011 P/cc for 2 cm
radius trap2 - 1 9 to 1014 p/cc for a 50 cm radius trap.2"2 If
a trap that large could be built, it would hold about a
milligram of antiprotons.
The ability of a Penning trap to store antiproton ions for
long periods of time will be tested 6n the next few years.
Bolh the University of Washington -2 U and the Los Alamos
National Laboratory 2 . 2 1 are mounting efforts to trap one or
more of the slow antiprotons available at LEAR. It is
suspected that the Italian group will proceed with their plans
also.

39
*: The University of Washington program will use one of their
cryogenic Penning traps modified by thinning the center
portion of one of the cap electrodes down to about 0.25 mm.
This will maintain the ultrahigh vacuum capability that they
have demonstrated in their sealed traps. The thinned portion
of the electrode will act as a stopping foil for the medium
energy (100 keV) antiprotons. With the thickness adjusted to
equal the average range of the antiprotons in cap electrode
material, some of the antiprotons will emerge from the other
side with just the proper energy to be caught in the trap.
The intention of the University of Washington experiment is to
trap only a few antiprotons and measure the mass to high
precision. As a byproduct, however, the lifetime of the
antiproton in the trap will put a good upper limit to the
quality of the vacuum in the trap.
The LANL experiment is more ambitious. They are fabricating a
radio frequency quadrupole (RFQ) decelerator that will
decelerate the beam of antiprotons from LEAR down to the trap
energy and deposit a large number (up to i010 antiprotons) in
the trap.

References:
2. 1 9 N. Beverini, L. Bracci, V. Lagomarsino, G. Manuzio, R.
Parodi, and G. Torelli, "A Penning trap to store antiprotons,"
pp. 771-778, Physics at LEAR, U. Gastaldi and R. Klapisch,
ed., Plenum Press, NY (1984).
2 . 2 0 W.
Kells, G. Gabrielse, and K. Helmerson, "On achieving
cold antiprotons in a Penning trap," FERMILAB-Conf-84/68-E,
Fermi National Accelerator Lab, Batavia, Illinois (August
1984). [Preprint submitted to the IX Int. Conf. on Atomic
Physics, Seattle, Washington (23-27 July 1984).]
2- 2 1 L. Campbell, W.R. Gibbs, T. Goldman, D.B. Holtkamp, M.V.
Hynes, N.S.P. King, M.M. Nieto, A. Picklesimer, and T.P.
Wangler, "Basic research in atomic, nuclear and particle
physics," LA-UR-84-3572, Los Alamos National Lab, Los Alamos,
New Mexico (1984).
* 2. 2 2 p. Ekstrom and D. Wineland, "The isolated electron," Sci.
Am. 243, 105-121 (August 1980).
2 " 2 3 G. Gabrielse, personal communication (1985).
2 * 2 4 p.B.
Schwinberg, R.S. Van Dyck, Jr., and H.G. Dehmelt,
"New comparison of the positron and electron g factors," Phys.
Rev. Lett. 47, 1679-1682 (1981).

40

-- -o * - . - , - * . - * - - . . -.. . .
 - -~
-~ -~ -- - - - -. ,-

SECTION 3

IMPROVING ANTIPROTON PRODUCTION EFFICIENCIES

There are a number of obvious ways to improve the efficiency

of antiproton production over the present methods. It should
be realized that the present production facilities were
designed under a number of restrictions. They had to use
existing proton accelerators, fit onto the existing sites, and
not use up too much precious time on the main machine. Above
all, they had to be built as fast as possible so that some
laboratory in some other country didn't reach the next highest
interaction energy region first, find the next batch of new
"elementary" particle resonances, and win that year's Nobel
prize in Physics.
The original motivation behind the production of antimatter at
these laboratories was very pragmatic. It was not for the
study of the properties of antimatter, although some of that
work is now going on at CERN since the antiprotons are
available. The reason antimatter sources were built was to
reach a higher center-of-mass energy by colliding two particle
beams head-on. Beam-beam collisions produce much more
interaction energy than a proton beam hitting a stationary
target (see Appendix A). The present proton machines can be
used in a colliding beam mode with minor modifications if they
are loaded with two beams, one of protons and one of
antiprotons, orbiting in different directions.
To obtain the same results with two colliding proton beams
would have required buildiig a second proton ring that
intersects with the first one at a number of points (like the
ISR at CERN). It is interesting to note that the reason the
three different national facilities have tackled the very
difficult task of making, collecting, cooling, and storing
antimatter is that they realized it was cheaper to make an
antiproton facility for producing antimatter than to build
another proton ring.
A good part of the effort on this study contract was to
determine the reasons behind the low efficiencies of the
present facilities. Some of the low efficiencies are
inherent, such as the number of antiprotons per proton from a
target. Most of the other low efficiencies are just artifacts
of the particular choices forced on the particular facility,
and there are obvious ways to improve these efficiencies by
large factors. These will be discussed in the following
subsections.

41
 3.1 ALTERNATE ANTIPROTON PRODUCTION TECHNIQUES

A survey was undertaken to find alternate methods of producing

antiprotons other than by high-energy protons striking a heavy
nuclei target.

3.1.1 Colliding Particle Beaus

It is known that colliding proton beams produce
antiprotons. 3 -I Because the center of mass energy in a
colliding proton beam configuration is higher than in a proton
beam-target configuration, the energy efficiency for
production of antiprotons per proton is significantly higher
for the colliding beam case. Unless methods for significantly
improving the luminosity of the beams are found, however, the
absolute rate of production will remain small.

It should also be remembered that in the true colliding beam

case, the antiprotons are emitted isotropically, making them
more difficult to collect. This can be rectified somewhat by
making one beam more energetic than the other so the resulting
products come out in a directional spray.
The colliding beams do not have to be protons. Colliding
beams of positrons and electrons at 29 GeV center of mass
energy produce 10.7 pions, 1.4 kaons, 0.3 protons and 0.3
antiprotons per annihilation event. 3 "2 Again, this method
suffers from the same low luminosity and wide angular
distribution problems as the colliding proton beam technique.

To get around the low luminosity problem, which is basically

caused by space charge repulsion between the particles in the
beam, it has been proposed to use colliding beams of heavy
ions such as uranium with energies so high (2.5 TeV) that each
nucleon has 10 GeV or ten times its rest mass energy.
Because of the small charge to mass ratio of singly charged
uranium, calculations show that for a 100 m diameter colliding
* beam facility with an intersection area of 1 cm2 , the
production rate could be as high as 1018 P/s (1 gm/wk) without
. exceeding the space charge limit. Thus, the real limit in
"these machines is not space charge, but the available power.
There are problems with this concept because of the isotropic
distribution of the antiprotons and the large amounts of
nuclear debris from the U-U interactions that would have to be
filtered out. However, as design studies start on antiproton
factories, colliding heavy ions beams should definitely be one
of the candidate systems.

42

"." 42
"3.1.2 Laser Induced Pair Production

It has been proposed to make antiprotons by laser-induced pair

production.3"4 The origi al an iproton-proton production
threshold estimate of 101 W/cm for CO 2 las H radiation in
that pareg has since been increased to 3x10 3 W/cm by the
author. This is still far from the capabilities of even
the most optimistic laser designs. Perhaps self-focusing of
the laser light in plasmas 3 . 6 or other exotic effects might
bring the goal closer.
Even then, unless some unexpected resonance cross-section is
found near the proton-antiproton energy threshold, it is
expected that with laser production, just as is observed in
production of antiprotons with high-energy particles, for
every antiproton produced most of the energy will go into the
production of many positron-electron, pion, kaon and other
particle pairs.

References
3 .1 M.
Antinucci, A. Bertin, P. Capiluppl, M. D'Agostino-Bruno,
A.M. Rossi, G. Vannini, G. Giacomelli, and A. Bussiere,
"Multiplicities of charged particles up to ISR Energies,"
Lett. Nuovo Cimento 6, 121-127 (1972).
3.2H. Aihara and TPC Collaboration, "Charged hadron production
in e+e- annihilation at 29 GeV," LBL-17142 preprint, Lawrence
Berkeley Lab, Berkeley, California 94720 (December 1983)
[submitted to Physical Review Letters].
3. 3 G. Chapline, "Antimatter Breeders?" J. British
Interplanetary Soc., 35, 423-424 (1982).
3.4H. Hora, "Estimates for the efficient production of
antihydrogen by lasers of very high intensities," Opto-
Electronics 5, 491-501 (1973).
3 _5 H. Hora, personal communication (23 November 1983).
3 . 6 H.
Hora, "Theory of relativistic self-focusing of laser
radiation in plasmas," J. Opt. Soc. Am. 65, 882-886 (1975).

43
 3.2 M4AXIMIZING THE ANTIPROTON PRODUCTION RATE

Data on the antiproton production spectrum of high-energy

protons impacting heavy metal targets is available only for
small angles about the forward direction. These data are
sufficient for the design of the present antiproton collector
systems that only attempt to capture the antiprotons emitted
around the forward peak.
In order to design systems that will capture a higher
percentage of the antiprotons, it will be necessary to know
the antiproton spectrum as a function of angle and incident
proton energy over a greater angular spread. Such data does
not seem to exist and there are no present plans to make these
measurements, since *obtaining the data would require an
extensive amount of time on the large synchrotron machines.
The particle physics community prefers to use the machine time
to study issues more important to particle physics. As a
result of this lack of detailed knowledge of the spectrum, the
total number of antiprotons generated is also unknown (to
probably a factor of two).
In Section 2.1, Figure 2-2, an estimate was made of the total
antiproton production efficiency for protons hitting a heavy
metal target as a function of proton energy. This is the
upper curve in Figure 3-1. If we now take the upper curve and
divide it by the energy of the proton making the antiprotons,
we obtain the bottom curve. This is the energy efficiency for
producing antiprotons. Note that it has a broad peak around
200 GeV. Although the number of antiprotons produced
continues to increase as the incident proton energy is
increased, above 200 GeV the gain in production is not enough
to offset the increased proton energy required.

From Figure 3-1 we see that the maximum energy efficiency

production rate occurs for an incident proton energy of
200 GeV and is 0.085 antiprotons/proton. (There are roughly
5 K mesons, 50 pi mesons, and large numbers of positrons and
electrons produced for each antiproton generated.) This
antiproton production rate is 2 times the production at the
Fermilab energy of 120 GeV and 20 times the production at the
CERN energy of 26 GeV. It should be emphasized that the
curves in Figure 3-1 are based on sparse data and that actual
measurements of antiproton production spectra as a function of
angle and proton energy are needed before any major
engineering studies on antiproton production are done.

44

V.............................................. -*,-.--- .'. -* ", " .

V.

ESTIMATED
TOTAL p-Pb

0- 10EASURED p-p
IX 10-1 OPTIMUM

z
0.

* 0

10-

* 2 PRODUCTION
CERN ENERGY
EFFICIENCY

i 10- 1001 10001 il

10 10 100010,000
PROTON ENERGY, GeV

S.-Fig. 3-1 Total antiproton production rates and efficiencies.

45
 3.3 IMPROVING TARGET EFFICIENCY
The present targets for antiproton production are long wires
of beryllium, copper, or tungsten a few millimeters in
diameter and a length designed to optimize the tradeoff
between a longer length to maximize the proton interaction and
a shorter length to minimize the antiproton absorption. The
first improvement to the present targets would be to have them
carry a current so that the antiprotons produced would have a
tendency to stay near the target axis. This would in effect
convert the antiproton source from a rod source to a disk
source at the exit plane of the target. An alternative is to
break up the target into multiple targets and use magnetic
3
lenses between each section to refocus the antiprotons. "7

At Fermilab, designs have been carried out on rotating

targets. 3 8 The length of the tungsten targets at Fermilab
is only 5 cm, so a disk target will be of reasonable size and
can be rotated at high enough speeds to minimize the local
heating problem. At IHEP designs have been carried out on
targets made of a flat jet of mercury with a width of 0.5 to
3 mm and 60 cm long. 3 -9 The flowing mercury solves the target
cooling problem for the proton beam intensities expected at
IHEP. Beryllium foil windows are used to separate the mercury
target vacuum chamber from the main vacuum system.

For an antiproton factory, a major problem area needing

invention, materials experimentation, and study is the
development of a suitable target design. Rotating and liquid
targets are a potential solution to this problem, but further
work needs to be done to find the best target-lens
configuration for high power production.

References:
. MacLachlan, "Current carrying targets and multitarget
arrays for high luminosity secondary beams," FN-334, 8055.000,
Fermi National Accelerator Lab, Batavia, Illinois (April 1982).
3- 8 Fermilab staff, "Design Report: Tevatron 1 Project," p. 3-4
and Figures 3-6 and 3-7, Fermi National Accelerator Lab,
Batavia, Illinois (September 1984).
3 "9 B.F.
Bayanov, A.D. Chernyakin, V.N. Karasyuc, G.I.
Silvestrov, T.A. Vsevolozskaya, V.G. Volohov, G.S. Willewald,
"The antiproton target station on the basis of lithium
lenses," pp. 362-368, Proc. llth Int. Conf. High Energy
Acceleritors, Geneva (1980).

46

S *-* *- .c%,
S.. '. I* w
 3.4 IMPROVING ACCELERATOR EFFICIENCY

The present machines that are used for accelerating protons to

relativistic speeds are synchrotrons. The design of these
machines enables protons to be accelerated to very high
energies using a reasonable amount of hardware and site
acreage. The synchrotron is built in a large circle and the
protons travel around through the machine many times, gaining
energy with each cycle. The upper limit to the energy is
primarily determined by the maximum strength of the magnetic
fields in the bending magnets that keep the proton beam
turning in a circle. The synchrotron provides the particle
physicist with high-energy protons at a very precisely known
energy. It is an ideal tool for studying elementary particle
physics. The average current that the synchrotron can handle
is small, however, and the efficiency is only a few percent.

There is an alternate machine for producing high-energy

protons, that can handle high average currents and has high
energy efficiency. It is called the linear accelerator or
linac. As its name implies, the linear accelerator is a long,
straight machine. The protons only pass through it once. The
linac consists of a series of RF cavities which set up
electric fields that accelerate the protons, drift tubes to
isolate the protons when the RF fields reverse phase, and
alternating polarity quadrupole focusing magnets to keep the
proton beam from spreading.
"- By using the alternating gradient focusing concept, it has
-" become possible to accelerate very intense beams (up to one
"ampere) to very high energies. The energy limit is economic,
"not technical. It is known that machines can be built to
handle over 250 mA, since it has been demonstrated in the
first section of the linac injector at Fermilab, which is the
only section where current limitations would occur.
Acceleration to higher energies only requires more RF power.
Because of their high average current capability, linacs have
been considered for an unconventional type of nuclear reactor,
called an accelerator breeder or electrothermal reactor. The
linac is used to accelerate protons to about 1 GeV and the
beam is sent into a target made of unenriched or even depleted
uranium. The protons cause spallation reactions to occur,
which produce up-to 100 neutrons per incident 1 GeV proton.
The neutrons react with the 2 3 8 U to produce 23 9 Pu as well as
6 GeV of thermal energy, which can be used to make electricity
to help run the linac. Originally considered as an alternate
method of making plutonium for weapons, these types of
reactors received considerable study in the 1950's and 1960's,
but as the supply of uranium ore became more secure, effort on
the accelerator breeder stopp~di The last major meeting on
". ~--': the subject was held in 1977. '•

47
 What came out of the meeting was the realization that high
current linacs could be built and that with the availability
of new high efficiency (75%) klystrons, their AC "wallplug"
power to proton beam power efficiencies could exceed 50%. At
these efficiencies the thermal energy released in the reactor
could supply enough electrical power to keep the linac running
so that the reactor would be self-sustaining. The output of
the reactor would then be the plutonium. One specific design
of such a system would use 600 MW of electrical power to drive
a 50% efficient linac. The linac would then produce a 300 MW
proton beam. The proton beam would be sent into an
electronuclear reactor where it would release over 1500 MW of
thermal energy. The thermal energy would then be used to
operate a thermal power plant. At 40% efficiency, thp power
plant would produce the 600 MW of electricity needed to run
the linac, and the cycle would be closed.
The Chalk River, Canada linac program3 -I 0 has n studying
100% duty factor linacs, with the goal of producing a linac
capable of of 300 mA average current at 1 GeV (0.3 GW beam
power) for use in an accelerator breeder. They have several
designs for a proton ion source that are capable of delivering
a 300 mA beam of satisfactory emittance, so ion sources are
not a limitation.
The acceleration limit of a linac (the energy increase per
meter) is determined by the sparking limit in the cavity. The
sparking limit is inversely proportional to the wavelength.
Present machines operating at 200 MHz usually operate at
1 MeV/m, and there are designs for higher frequency machines
that will operate at 5 MeV/m.

A linac for an antiproton factory with a proton energy of

200 GeV would be 40 km long. This is a little longer than the
28 km LEP ring presently under construction at CERN and 1/5th
the size of the 200 km Supercollider being proposed as the
next large particle accelerator in the US. If run at a power
level of 10 GW, the proton current required would only be
50 mA, which is less than the 300 mA of the Chalk River
design. By skimming off the antiprotons and using the
remaining particles to run a closed electronuclear breeder
cycle, such a factory would require no outside power source
and would essentially be turning depleted uranium into
plutonium and antiprotons.

References:
3 "1 0 H.J.C.
Kouts, Chairman, Proceedings of an Information
Meeting on Accelerator Breeding, Brookhaven National
Laboratory, Upton, New York, 18-19 Jan 1977.

48

• .
 "3*3.5 IMPROVING ANGULAR CAPTURE EFFICIENCY

The present angular capture efficiencies of the magnetic lens

collection systems are already quite good, with up to 30% of
the antiprotons collected and directed into the aperture of
the collecting ring. The present designs, however, were
optimized for the antiproton energies expected at the
particular facility and the particular conditions in the
target area (such as power supply limitations, tunnel size,
etc.).
The estimated number of antiprotons collected as a function of
antiproton momentum has been calculated for both 30 mrad and
60 mrad angular acceptance at a number of incident proton beam
energies. 3 "1 1 These curves were then integrated to obtain an
estimate of the total number of antiprotons collected by the
two lens acceptances. These estimates are plotted in
Figure 3-2 along with the estimate of the total number of
antiprotons obtained from protons hitting heavy metal targets
S(see Figure 2-2).
As is shown in Figure 3-2, the limitation on angular
acceptance to 30 mrad means that only 55% of the antiprotons
generated are being captured by the lens at the higher
energies and even less at the lower energies. If the angular
capture could be increased to 60 mrad or more, then the
"* angular capture efficiency could reach a value of 80% or
greater.
Also shown in Figure 3-2 is the energy efficiency of the
antiproton production process. Each of the top three curves
was multiplied by the energy that would be released by the
annihilation of the antiproton (2mc 2 =1.876 GeV) and divided by
the energy of the incident proton to produce an energy
efficiency estimate. This energy efficiency estimate gives us
a rough guide to the choice of the optimum proton energy for
production. We can see from the bottom three curves that, if
we can capture all the antiprotons, the optimum energy for the
proton beam is 200 GeV. If our angular acceptance is limited
for some reason, it pays to go to slightly higher energies.
(This is intuitively obvious since at higher energies the
antiprotons are more concentrated in the forward direction.)

These "optimum energy" estimates, however, do not take into

account the fact that the antiproton longitudinal momentum is
spread over a wider band at higher energies. The best
operating point must take this and a number of other factors
into account, but it looks as though a proton beam energy from
100 to 500 GeV will give the best overall energy efficiency
for antiproton production. We will assume for now that
200 GeV is an optimum energy for the incident proton beam.

49
 ESTIMATED
- TOTAL

z ACCEPTANCE
0

I-- 10-1 ACCEPTANCE

2 S•. NUMBER OF

z ANTIPROTONS
PER PROTON'.
z
* 0
0
cc PRODUCTION
'j• 10-2 ENERGY EFFICIENCY
-N (2 mO 2)/E -

TOTAL
60Omrad

10-3 1 1 1 t ill , , , ,I II I'

t ' '
10 100 1000 10,000

Ep, PROTON ENERGY, GeV

Fig. 3-2 Antiproton production vs. angular capture.

50
 "All of the studies to date have assumed that only a single,
on-axis lens would be used to capture the antiprotons.
Because the antiprotons are being emitted over a wide angle,
this immediately leads to the requirement of a short focal
length for the lens so it can capture these wide angle
antiprotons. Research needs to be done on the feasibility and
comparative merits of an array of lenses. Since each lens has
to capture only the antiprotons in a small portion of the
emitted beam solid angle, the focal length requirements can be
relaxed. The support hardware for the lenses will cause
interception losses, however, and realistic tradeoff studies
need to be done between the number of lenses and the lens
design parameters. If a multiple lens approach looks
desirable, then invention is needed on low-loss devices for
separating the wide angle antiproton beam into multiple beams
to minimize the interception losses of the multiple lens
hardware.
Most present magnetic lens designs require the antiproton beam
to pass through the material of the lens. This causes
significant losses in the antiproton spectrum. This is not
true for the magnetic quadrupole lens, but it does not focus
well in all orientations. Invention is needed to develop new
lens designs with low loss and good focusing.
- Another method for construction of a magnetic lens similar to
that of the lithium lens would be to carry the current for the
lens in a cylinder of ionized plasma instead of lithium metal.
The problem of the current or beam heating up the lens would
be gone and it is likely that absorption of the antiprotons in
the lens would be less. A plasma lens would have its own
problems, such as all the various plasma instabilities that
would be driven by the high currents needed. The concept is
still in the preliminary idea stage at CERN and to date there
have been no publications concerning its feasibility.
The present single magnetic lens concepts have already
achieved angular capture efficiencies of 30% or greater and
there are many ideas for new lens concepts with greatly
improved performance. It is therefore reasonable to expect
that after modest investment in invention, engineering, and
testing, there should be new magnetic lens designs capable of
capture angles of 60 mrad even at higher antiproton momentum
levels and capture efficiencies of 85% or greater.

References:
3.11C. Hojvat and A. Van Ginneken, "Calculation of antiproton
yields for the Fermilab antiproton source," Nucl. Instr. &
Methods 206, 67-83 (1983).

51
 3.6 IMPROVING MOHENTUM CAPTURE EFFICIENCY

We now will look at the problem of momentum capture. When the

antiprotons come from the target, they not only have a wide
spread in angle, they also have a wide spread in momentum. It
is this wide spread in resultant antiproton momentum and the
difficulty of making an antiproton collection ring with a wide
momentum acceptance that leads to the extremely low
inefficiencies in present antiproton "factories".

For example, at CERN, the present momentum acceptance in e

Antiproton Accumulator is only dp/p=l.5%. For a peak momentum
of p=3.5 GeV/c, this translates into a momentum bite of only
dp=0.05GeV/c. Thus, only 1% of the 5 GeV/c half-width of the
antiproton momentum spectrum is captured.

If we assume that an antiproton factory has an incident proton

beam at 200 GeV energy and a 60 mrad angular acceptance, then
the flux of antiprotons per protqn per unit antiproton
momentum is shown in Figure 3-3. ' The antiproton flux
peaks at 12.5 GeV/c antiproton momentum and spreads from
1 GeV/c to 50 GeV/c, with a half-width of 22 GeV/c.
The present state-of-the-art in collection rings for
antiprotons is a momentum acceptance of about dp/p=6%. Even
if this could be raised to 8%, the momentum bite at an
antiproton momentum of p=12.5 GeV/c would still be only
dp=l GeV/c and would capture only 4.5% of the antiprotons.
The obvious solution is to build a multiplicity of rings.
They would be identical copies of each other, with the same
radius of curvature and sharing the same tunnel (since the
tunnel costs are a major portion of the expense of a ring).
Each ring would have the strength of the magnetic fields in
its bending magnets set at a different level to keep a
different antiproton momentum circulating through its vacuum
pipe. If each ring could handle a momentum bite of 1 GeV/c,
then as we see in Figure 3-3, 16 rings could capture 61% of
the antiproton momentum spectrum and 27 rings could capture
84% of the antiprotons.

Separating the antiproton beam into different momentum buckets

should not be too difficult. The magnetic focusing lens has
chromatic aberration. Each level of antiproton momentum is
focused at a different focal point. A string of diverter
magnets can deflect the different antiproton momenta in
different directions, where they can be channeled to the
proper collecting ring.

52

V............................
* . . .
 1i-2

12.5GeV ANTIPROTONS PRODUCED FROM

" ~200 GeV PROTONS INTO
_ Sz 60 mrad ACCEPTANCE
0
0 1 GeV -4.5 %
0r.
z 16•Ge - 61%
0• 10-3 %
cr. -2 GeV -84%,

. MOMENTUM ACCEPTANCE

0
•..

"" ~~~10-4IIII
0- 10 20 30 40 50

ANTIPROTON MOMENTUM, GeVic

.4

Fig. 3-3 Antiprotons vs. momentum spread.

53

'I
 ,I.•

3.7 SYSTEM EFFICIENCY ESTIMATES

From the previous sections we can see that there are a number
of ways to improve the efficiency of antiproton production
over the present techniques. The first obvious improvement is
to use a linear accelerator or other high-current, high-
efficiency machine to produce the protons instead of a
synchrotron. A linear accelerator with its energy efficiency
from ac mains to beam energy of 50% will be more than an order
of magnitude better than the present synchrotrons.
The next significant improvement is to use a higher proton
energy so that more antiprotons are produced in the target.
As is shown in Table 3-1, by going to 200 GeV, the number of
antiprotons produced can be increased by a factor of 20 from
the 3.8xi0- P/p at CERN to 8.5xi0- f/p.
The present magnetic lenses are relatively efficient in
capturing the resulting antiprotons in angle. Yet by
improving the design and going to multiple lens collectors, we
should be able to improve the angular capture efficiency by a
factor of 3 or 4 to 85%.

Another place where a significant improvement can be made is

in the momentum capture efficiency. The single ring
collectors at CERN, Fermilab, and IHEP are severely limited by
inherent difficulties with matching the magnetic and vacuum
pipe acceptances to the emittances of the lenses. With a
large enough tunnel and enough money to build copies of the
collectors, each tuned to accept a different momentum range,
it should be possible to improve the capture efficiencies
significantly. Whether it will be possible to go from the
present few percent to a desirable 85% is unknown. Studies of
the feasibility of separating a beam of antiprotons into
separate beams at different momenta are needed.
There are many losses as the antiproton beam is generated,
collected, and switched around from one device to another.
The handling efficiencies in the present facilities are not
bad, but improvement in this area is also needed.
As we can see from Table 3-1, if all these efficiencies can be
achieved, the total production efficiency of antiprotons can
be raised from the present 4xi0- 7 P/p at CERN by more than
five orders of magnitude to 5x10- 2 p'/p. With an efficient
linear accelerator replacing the low-efficiency synchrotron,
the energy efficiency can rise even more from its present
value of 1.4x10- 9 to 2.5xi0- 4 .

54
 •~~~t ; -- W.

Table 3-1 Antiproton Production Efficiencies

CERN Fermilab Goal

Incident Proton Energy (GeV) 26 120 200
Generation Efficiency (P/p) 0.4% 4.7% 8.5%
Angular Capcure Efficiency 20% 30% 85%
Momentum Capture Efficiency 1% 1.2% 85%
Handling Efficiencl 5% 18% 80%
Total Production Efficiency (p/p) 4x10- 7 3x10- 5 5x10 2

4
Overall Energy Efficiency 1.4x10- 9 2.5x10- 8 2.5x10-

3.8 CONCEPTUAL ANTIPROTON FACTORY

In Figure 3-4 is shown a conceptual design for an antiproton
factory which would utilize the technologies being developed
at CERN, Fermilab, and IHEP, but on a much larger scale and
with the design optimized for energy efficiency. First, the
proton accelerator should be a high-current rf linear
accelerator (linac) with a wallplug efficiency of 50%, rather
than the low-current, low-efficiency, but high-energy
resolution synchrotron preferred as a research tool by
particle physicists.
There would be more than one proton beam with each beam
operated at the optimum beam current for thp particular target
design chosen. Each proton beam would strike a metal target
and the resulting particles would be sorted by an array of
wide-angle collecting lenses to extract the antiprotons and
positrons. The positrons with the right energy would be
picked off and sent to the antihvy*rogen generator, while all
the antiprotons possible would be sorted by energy and sent to
a stack of stochastic coolers, each optimized for a particular
central antiproton momentum.
After stochastic cooling, the stack of beams at different
energies would go to a decelerator stack that would reduce all
the antiproton energies to the same subrelativistic energy
(200 MeV). The combined beam would then be sent to a
subrelativistic cooling ring using either stochastic or
electron cooling before being further decelerated and sent on
to the antihydrogen generator where the antiprotons are
*.combined with the positrons to make antihydrogen atoms.

55
 200 GeV PROTON BEAM TRE

(ONEOF ANYDOAO0.ENw

AND TRA GENEATO

~
Fig. ~ ~ iý A-4
(oOeLECmeTO.
eniro
+atr
DECELRATORARRAY

ANiYRG
1'
T A
BEAMLAER 56
 3.9 COST ESTIMATES

Cost estimates of major facilities requiring new technology

are notorious for their inaccuracy. It is possible, however,
to get a first cut at the operational cost of producing
antimatter in a prototype production facility by making some
assumptions about the energy efficiency of the production and
the cost of energy.

A milligram of antiproton fuel would be a useful quantity to

consider for antiproton propulsion since annihilation with a
milligram of protons would give an energy output equivalent to
the burning of 6 tons of liquid hydrogen/liquid oxggen rocket
fuel. One milligram of antiprotons contains 6x10 1. Thus,
if the antiproton-to-proton production ratic is 5%, the
factory would have to produce 1.2x1022rotons, each with
200 GeV of energy or a total of 3.8x10 IA J.
If the protons were accelerated with a linear accelerator with
an ac mains to beam power of 50%, then the linac would require
7.7x10 1 4 J of energy or in electrical power terms, 2x10 8 kW-
hr. If we had to buy that electrical energy at 0.05$/kW-hr,
then the cost of the antiprotons is found to be an optimistic
10M$/mg.
* Liquid oxygen and liquid hydrogen, being cryogenic fuels, are
expensive. But their production costs certainly are much less
than 10M$ for 6 metric tons. Thus, it is obvious that unless
the cost of producing antimatter comes down significantly,
antimatter will never be a cost competitive method of storing
energy on the surface of the earth where hydrogen and oxygen
are easily obtained. We shall see in Section 8, however, that
even at 10M$/mg, antimatter is a cost effective fuel in space.
In space, mass costs money, and the low mass of an antiproton
propulsion system makes it much more attractive than any
chemical propulsion system for the more difficult space
"missions.

57
 SECTION 4
GENERATION OF A*TIMYDIDGEN

In this section we will discuss methods for the generation of

antihydrogen atoms and molecules, given that a supply of
antiprotons is available as a result of the research discussed
in the previous sections. These investigations into the
generation of antihydrogen are ideally suited for relatively
inexpensive near-term experimental research efforts, since in
nearly every case, the research can be carried out using
normal protons, normal electrons, and normal hydrogen atoms,
"ions, and molecules. If performed at a university, the
research projects would be ideal thesis topics.
There is one caveat. The techniques used during the research
must not allow the protons, electrons, or hydrogen to come
into contact with other forms of matter. This includes
residual gas atoms in the vacuum system. To be a valid
simulation, the research should be carried out in ultra-high
vacuum (UHV) systems, not standard vacuum systems. In some
experiments, the UHV systems must be cryogenically cooled.
In each investigation, the effect of background gas on the
antimatter version of the experiment should be taken into
account. If the technique being studied allows for the loss
of protons, electrons, or hydrogen from the sample, the
effects of the radiation from the annihilation of the
equivalent antiprotons, positrons, and antihydrogen with the
walls should be calculated. As the investigations progress,
it may be found that the use of atomic and molecular hydrogen
ions, or excited states of either ions or neutral particles
may give better results than the use of neutral atoms or
molecules. Where possible, these alternate pathways will be
indicated in the following list of potential research
projects.
The research projects discussed in this section are relatively
independent of each other and can be carried out in any order.
The equipment needed is similar to that used in modern atomic
and molecular beam experinmnts. In many of the projects,
there will be a requirement for high-strength electric,
magnetic, and photon fields to produce significant interaction
rates. There will also be a need for sophisticated diagnostic
tools that can interrogate the state of individual particles
and abultrahigh vacuum to maintain the quality of the atomic
and molecular beams. Thus, it is essential that adequate
equipment funds be budgeted to provide the apparatus needed to
carry out the experiments.

59

PREVIOUS PAGE
JS BLAýNK
 4.1 CONVERSION OF ANTIPROTONS TO ATOMIC ANTIHYDROGEN

It will probably be found desirable to store the antiprotons

as some form of antihydrogen rather than as antiproton ions.
The first step is to convert the antiprotons into atomic
antihydrogen by adding a positron.

* 4.1) p- +e+ -- i +h%

Since the antiproton and positron attract each other, the

reaction will take place spontaneously. The system will emit
"photons as the electron jumps from one state to the next on
its way to the ground state. If there is a large velocity
difference between the antiproton and positron, however, the
cross section can be small and the conversion rate slow. It
may be found necessary to enhance the conversion rate by using
electromagnetic stimulation or possibly a catalyst made of
some form of antihydrogen.
If the antiprotons are obtained as trapped and cooled ions in
an ion trap, such as a Penning trap, then a straightforward
* approach would be to convert tne trap into one that would hold
and cool both positive and negative ions. A simple example is
the Paul trap. As is shown in Figure 4.1, a Paul trap has the
same electrode configuration as the Penning trap, but uses rf
electric fields instead of static electric and magnetic fields
to form the trapping region. A Paul trap could hold both
antiprotons and positrons and cool them until their relative
velocities were low enough that the conversion to hydrogen
takes place naturally.
There are a number of obvious research questions concerning
the recombination process that need to be investigated.
Efficient methods for injecting the antiprotons and positrons
and removing the neutral antihydrogen need to be found. The
effect of high ion densities on the trapping lifetime and the
recombination rate needs to be evaluated. The use of rf
modulation or laser radiation to enhance the various reactions
needs to be studied. Also, the effectiveness of magnetic
fields or laser beams to trap and manipulate the neutral
antihydrogen atoms and ions needs to be determined.
Then, the conversion rate of the neutral atomic hydrogen to
positively charged atomic antihydrogen ions, or antihydrogen
molecules with positive, negative, or neutral charge needs to
be measured and the efficacy of these antihydrogen atoms or
molecules in acting as a "catalyst" to enhance subsequent
conversions needs to be determined.

60

jr..
 i•: • CAPELECTRODi

RADIO FREQUENCY
FEDFORCES

ELECTRODF ta for positive an* ELECTRODE

"" :~CAP ELECTROE

Fig. 4-1 Paul trap for positive and negative ions.

For example, after the first antinydrogen atom is formed, it

may be desirable to deliberately enhance the addition of a
positron to it to form an atomic antihydrogen ion. The
antihydrogen ion can then be used in a charge exchange
reaction with an antiproton ion. In this manner, the
antihydrogen ion acts as a catalyst to breed new antihydrogen
atoms that act as catalysts for further reactions:

4.2) H+ e+ + h 1 -- H+ 2h1
4.3) H+p+h 2 -- H +H + 2hf 2

Research on the reaction given by Equation 4.3 for normal

hydrogen ions has been proposed as a method for neutralizing a
hydrogen particle beam.-' 1 Any results from that study of
hydrogen ions will have relevance to the antihydrogen problem.

61

"---------------------------------------
- - . 2 C L•ttXtr%.2C S % .. \' N* ' C'-x 2~ .' .' ' ' .'
If this research on conversion of trapped antiprotons to
antihydrogen shows promise, thought needs to be given to
methods for converting this "batch" process to a "flow"
process suitable for a high throughput antiproton "factory".
Instead of a trap, the antiprotons may be available in the
form of a low-energy, low-temperature beam. The research task
here is to combine this antiproton beam with a positron beam
and have it convert into an atomic antihydrogen beam. This
experiment has already been carried out at CERN using normal
matter during the process of studying the concept of
electronic cooling. (See Section 2.3 on Electron Cooling of
Antiprotons.) Electron cooling was first tested by mixing a
cold electron beam with a hot proton beam. During the tests
it was noted that some of the electrons recombined with the
protons to produce hydrogen atoms. 4 " The recombination rates
were low, however, so the concept.Qf,sing lasers to enhance
the recombination was developed.'* A conceptual system
that incorporates this concept is shown in Figure 4.2.

ANTIPROTON RING

BENDING
MAGNETS

DECELERATOR. PHOTON - 4 + IRROR

-4 P .-- +

LENS 'Z:_" LENS

SEPARATOR COMBINER
MAGNET FOCUSED MAGNET +
~LASER
BEAM•

POSITRON RING

MIRRORý MIRRORIE
LASER RING

Fig. 4.2 Laser-enhanced antihydrogen formation.

62
 S"->
The apparatus for laser.enhanced antihydrogen formation
consists of three storage rings coupled at a straight section.
One storage ring contains the antiprotons. This would be very
similar to the Low Energy Antiproton Ring (LEAR) at CERN and
would incorporate stochastic cooling to keep the velocity
spread of the antiprotons low. There would be a second ring
for the cooling and storage of the positrons. In the combined
straight section there would be magnets that would separate
the oppositely charged moving particles at one end and
recombine them at the other so they are moving in the same
direction with very low relative velocity. A laser, tuned to
the optimum frequency to enhance the recombination, would then
be sent up the combined beam to stimulate the recombination of
the antiproton and the positron into antihydrogen. The
neutral antihydrogen atoms would pass through the magnet
section without being bent and would exit as a neutral
antihydrogen atomic beam. The neutral antihydrogen beam would
then be manipulated using lasers as described in the section
on laser cooling and trapping.

In this apparatus, the basic reaction in Equation 4.1 is

stimulated by irradiation with light corresponding to a
transition from the continuum into a bound atomic state.

"4.4) p + e+ + h - H + 2hy

The proposed CERN experiment 4 . 3 will be carried out with

antiprotons in the LEAR storage ring. These antiprotons are
typically at an energy of 100 MeV. Although this is "low
energy" to a particle physicist, the velocities are high
enough that the frequency of the laser light coming up the
beams of antiprotons and positrons is significantly Doppler
shifted. The CERN experimenters will use visible light
lasers, but the light will be Doppler shifted to the
ultraviolet frequencies needed to stimulate the recombination.
At lower antiproton energies, coherent ultraviolet sources
such as those discussed in the section on cooling and trapping
of neutral particles will be needed.
The experiment proposal to produce and study antihydrogen has
not received a high rating from the CERN experiment evaluation
committee because it is not elementary particle physics. In
the probable view of the committee, it is not even a physics
experiment, it is a trivial chemistry experiment. The
recombination of an electron with a nucleus is a well
understood atomic process and the mere demonstration of the
process with antimatter is of marginal value. Thus, the
experiment will have to wait a number of years while more
important science questions are answered with the limited
supply of antiprotons available in LEAR.

63

."
Thus, to obtain any near term results on the controlled
production of antihydrogen, it will be necessary to design and
carry out experiments using normal matter (protons and
electrons). The objective of the research would be to
investigate the interaction of the protons with the electrons
under the influence of other protons, electrons, hydrogen
atoms, and electromagnetic fields, and to optimize the
conversion of the protons and electrons into atomic hydrogen.
The resultant hydrogen atoms should emerge from the process
under "control". That is, they should be stationary, or in a
reasonably collimated beam with a small velocity spread.

References:
4.ID.G. Steel, J.F. Lam, R.A. McFarlane, "Studies of laser
enhanced relativistic ion beam neutralization." White Paper,
Hughes Research Laboratozies, Mialibu, California (October
1983).
4 . 2 M.
Bell and J.S. Bell, "Capture of cooling electrons by
cool protons," Particle Accelerators, 12, 49-52 (1982).
4 3R. Neumann, H. Poth, A. Winnacker, and A. Wolf, "Laser-
enhanced electron-ion capture and antihydrogen formation," Z.
Phys. A, 313, 253-262 (1983).
H. Herr, D. Mohnl, and A. Winnacker, "Production of and
experimentation with antihydrogen at LEAR," pp. 659-676,
Physics at LEAR with Low-Energy Cooled Antiprotons, Workshop
on Physics at LEAR with Low-Energy Cooled Antiprotons, Erice,
Sicily, Italy, 9-16 May 1982, U. Gastaldi and R. Klapisch,
Ed., Plenum Press, NY (1984).

4.2 CONVERSION OF ANTIHYDROGEN ATOMS TO MOLECULES

Once we have atomic hydrogen in some controllable form, the
next step is to produce molecular hydrogen. Atomic hydrogen
easily converts to molecular hydrogen since the process
releases energy. A third body is needed, however, to carry
off the energy. Under normal laboratory conditions, the third
body is the wall or other atoms or molecules. We do not want
the antihydrogen to touch the walls or to have antihydrogen
atoms or molecules leaving the reaction with too much energy.
To keep control of the process, it will be necessary to supply
the "third body" in the form of one or more photons.

64
 4.2.1 Atomic to Molecular Hydrogen Reactions
In addition to neutral molecular hydrogen, both HJ, a molecule
of hydrogen missing an electron, and Hj, a molecule of
hydrogen with an additional electron are known to be stable,
so the antimatter analogs should be stable. Thus, there are
many reactions that could be investigated in this research
project. Some of them are listed below:

4.5) HO + HO + hV -- > H2 + nhy

4.6) HO + HO + h-> H + nhr + e-
4.7) H- + p+ +h -- >H 2 + nhV

4.8) H- + p+ + h -- >H + nh + e-
4.9) H- + H- + hif -- + nhV + e-

4.10) H- + H- + hV--> H2 + nhV + 2e-

4.11) H + H- +h -> H2 + nhY
4.12) HO + H + htf H2 + nhif + e-
4.13) H + H- + hV -- >H + nhl + 2e-

Some of these reactions involve the loss of positrons. This

may be an acceptable loss, depending upon the results of the
"investigations. For example, the availability of a positron
to act as a third body to aid in conservation of both energy
and momentum in the reaction may give these reactions a much
higher conversion rate than those that have only photons
* emitted after the formation of the molecule.

All of these processes have modifications where one or more of

*" the hydrogen atoms or the resultant hydrogen molecules are in
an excited state. It might be found advantageous to excite
one or both of the initial hydrogen atoms before the two are
brought near each other, rather than having to supply the
photon energy at the exact instant the two atoms approach each
other. Also, exciting one or both of the atoms to a very high
level to produce a "Rydberg" atom would increase the physical
size, thus increasing the cross-section.
By using polarized hydrogen atoms and polarized laser beams
combined with magnetic fields, it may be possible to produce
neutral hydrogen molecules in the desired ground level
parahydrogen state. If that research is successful, then
research on the conversion of hydrogen to parahydrogen would
-, ,~-.not be needed.

-Aa
65
4.2.2 Orthohydrogen and Parahydrogen Molecules

Molecular hydrogen is usually found as a mixture of two

different states. The ground state is called parahydrogen and
consists of two bound hydrogen atoms with the nuclear magnetic
moments or spins of the two proton nuclei pointing in opposite
directions (I=0). Since the two electrons in the ground state
are required by the Fermi exclusion principle to have their
spins oppositely directed (J=O), the atom has zero net
magnetic moment (M=0).
The first energy level of molecular hydrogen is created when
the spins of the two protons are aligned with each other
(I=1). This form, called orthohydrogen, has net magnetic
moment (M=l) and is highly metastable at low temperatures. It
is only 170.5 K (0.0147 eV) above the ground state and
conversion to the ground state can only take place when a
magnetic gradient perturbs the magnetic moments of the two
protons, causing one of them to flip over. The flipping can
be caused by the magnetic field from another orthohydrogen
molecule or by a magnetic catalyst such as nickel or iron.
The hydrogen molecule is a "homonuclear" molecule in that it
contains two identical nuclei. Since the molecule looks
identical if tne molecule is rotated so the two nuclei are
interchanged, the rules of quantum mechanics only allow
alternate energy levels to exist. As is shown in Figure 4-3,
for parahydrogen the allowed levels are the even rotational
states (J=0,2,4...), while for orthohydrogen it is the odd
nu•bered states (J=1,3,5...). 2 . 2 3 There are no restrictions
on the vibrational states.
Since the energy levels for both ortho and parahydrogen are
separated by two units of angular momentum, transitions
between two rotational states cannot be excited by a single
photon, which only carries one unit of angular momentum. Even
collisions cannot produce a transition between the rotational
states.
Transitions can be nroduced, however, by Raman scattering,
which involves two photons, the incoming photon and the
scattered photon. It also might be possible to produce two-
photon transitions using the nonlinear effects that arise when
the molecule is irradiated with intense beams of laser photons
at exactly half the transition energy. The orthohydrogen
molecule also has an electric quadrupole moment which can be
coupled into by a strong magnetic field.

66
 S:3 (23+1)
Rotational levels of 5 (11)
free Hydrogen
molecule

(9) I.

(7)

(5) 1 844.7K

509.9 K
1 (3) Yi M
YO 0 1_ - _17,0.5K
Para H2 Ortho H2
I=O II

Fig. 4-3 Molecular hydrogen rotational energy levels.

67

Y
 Because of the small energy difference between the ortho and
para states the orthohydrogen form is populated by thermal
.[•. excitation
temperature atconsists
elevatedof temperatures. Hydrogen gas parahydrogen
at room
75% orthohydrogen and 25%
since the first ortho state with J=l has three possible
"magnetic orientations (M=O,+l), while the para ground state
with J=0 has only one. When the mixture is rapidly cooled and
turned into a liquid or solid, this initial 3:1 mixture ratio
is "frozen in".

In the liquid or solid state, the orthohydrogen slowly

converts to parahydrogen, releasing energy. This energy must
be removed from the hydrogen or the liquid or solid will rise
in temperature and the hydrogen will evaporate. Since the
conversion process requires that two orthohydrogen molecules
approach each other closely enough that the magnetic field of
one causes the other to flip, the absolute rate of conversion
decreases as the density of the orthohydrogen molecules
decrease, so the process is never really complete. The
conversion rate in the solid state has been measured as
1.9%/hr or a l/e decay time of about 2 days. 4 . 5
At present, the method for producing pure parahydrogen from
the mixed form is to pass the liquid hydrogen over a magnetic
salt. The conversion process is quite rapid and very pure
samples of parahydrogen can be prepared that way. It will not
be possible to use matter in the form of magnetic salts to
convert antiorthohydrogen to antiparahydrogen, so it will be
necessary to find some other way to do it.
One approach is to prepare the antihydrogen initially in the
para state by controlling the molecular formation using
polarized beams of atoms and/or lasers combined with strong
applied electric and magnetic fields. Another is to convert
the orthohydrogen while it is still a gas using Raman or
nonlinear two-photon transitions. If nothing else works, one
could allow it to form as a low temperature mixed state solid
and cool the solid electronically for the number of days
needed for the conversion energy release to decrease to an
amount that can be handled by infrared radiation to the cold
walls of the containing vessel.

References:
I.F. Silvera, "The solid molecular hydrogens in the
condensed phase: Fundamentals and static properties," Rev.
Mod. Phys. 32, 393-452 (1980).

68

* I' "/ . . ' .. • '• . • ' ' . ~. .. . - v . .- . .. • : .• • .. .- . . . .. < • - . :• - - . .

 SECTION 5

SLOWING AND COOLING OF ANTIHYDROGEN

As was pointed out in Section 4, antihydrogen can either be

produced in a trap from stationary antiprotons and positrons,
or it can be produced "on the fly" in the form of an
antihydrogen beam. Slowing and cooling of the antihydrogen
will be necessary if the process of producing the antihydrogen
introduces some excursions in its direction and velocity.

5.1 ELECTRONIC SLOWING AND COOLING OF ANTIHYDROGEN IONS

If the resulting antihydrogen is a positive hydrogen ion with

an extra positron (or a negative molecular hydrogen ion with a
missing positron), then it can be guided by magnetic fields
and slowed by radio frequency fields or electrostatic
gradients in a process that is essentially the inverse of the
well known electrostatic and rf accelerator techniques.
A beam of atomic hydrogen ions can also be cooled by the
stochastic cooling technique where fluctuations in the beam
current are sensed and used to create feedback signals. When
these signals are applied to the beam by an electromagnetic
"kicker," they damp the fluctuations in the beam and produce
cooling. As discussed in Section 2.4 this technique has been
successfully demonstrated with protons and antiprotons
(although not with negative atomic hydrogen ions) at CERN and
Fermilab.
An alternate cooling technique for a beam of antihydrogen ions
is to combine them with a beam of co-propagating positrons
that have been generated with a very low spread in velocity.
Those ions that are moving too fast will be decelerated by
collisions with the positrons, and those ions that are moving
too slowly will be accelerated. As discussed in Section 2.3,
cooling of protons and antiprotons has been successfully
demonstrated at CERN using electrons.
Although electron beams can be used to cool antiprotons, since
electrons and antiprotons do not annihilate each other,
electron beams cannot be used to cool antihydrogen positive
ions. The electrons will annihilate with the positrons in the
antihydrogen ion. For atomic antihydrogen ions it will be
necessary to use a cold beam of positrons. A cold beam of
positrons will be significantly harder to generate than a cold
** beam of electrons.

69

--------------- j.:..
5.2 LASER SLOWING AND COOLING OF NEUTRAL ANTIHYDROGEN

Beams of neutral atoms or molecules can be separated into

different species, directed into different directions, slowed
to zero velocity, cooled to millidegrees, and manipulated into
traps solely with the use of photons, typically those from
carefully tuned lasers. Thus, it is possible to handle
neutral particles of antimatter such as antihydrogen atoms and
molecules, without touching them with normal matter.

There are two types of pressure forces caused by the

interaction of photons with neutral particles. One is called
dipole resonance-radiation pressure and the other is called
spontaneous resonance-radiation pressure. They both can exert
significant forces on a neutral particle and often both
effects are in operation at the same time in an experiment.
The dipole resonance-radiation pressure arises from the
optically induced dipole field of the particle when it is
placed in an optical field gradient. The induced dipole field
is caused by stimulated light-scattering processes. The
pressure is proportional to the gradient of the optical field
and the forces are transverse to the direction of the laser
beam. This force is dispersive in character because of the
sign change of the atomic polarizability on either side of
resonance.

The dipole forces become appreciable when the laser energy is

concentrated in a region, for example by focusing it to create
a region of high intensity. When the laser is tuned below the
atomic resonance frequency of a particle, the particle tends
to be pulled into the the high intensity region. When the
laser is tuned above the atomic resonance, the particles tend
to be expelled. Focal spot diameters of about 60 Fm and a 30-
fold increase of the on-axis atomic beam intensity have been
obtained using the dipole resonance-radiation force. 5 "I For
defocusing, the on-axis intensity can be reduced to less than
1% of its original value. Deflection angles of 5 mrad have
also been demonstrated.
The spontaneous resonance-radiation pressure arises from
spontaneous light scattering as the neutral particle is
irradiated by the photons from a laser beam. The force on the
atom is along the direction of the laser beam and is
proportional to the intensity of the beam. Spontaneous forces
have been used to deflect atoms, cool atomic vapors, produce
density gradients in a vapor, and separate isotopes. They
cannot be used to make optical traps because of an optical
equivalent to the Earnshaw theorem. 5 "3 The basic mechanism by
which the laser manipulation of neutral atoms by spontaneous
forces takes place is shown in Figure 5-1.

70
 *W..

SCATTERED PHOTONS

L PPHOTON
LASER PHOTON MOMENTUM TRANSFER
v Av = 1w/mc = 3.25 m/s

H ATOM 122 nm (•0.6 mK)

H2 MOL. 111 nm

Fig. 5-1 Laser slowing of a neutral particle.

In Figure 5-1, a neutral particle of mass m is moving with a

velocity v towards a laser beam. A typical velocity would be
v=10,000 m/s. This corresponds to an energy for a hydrogen
molecule with mass m=3.34x10- 2 7 kg of:

5.1) E = mv 2 /2

= 1.67x0-19 J

= 12,100 K

= 1.04 eV

The neutral particles will generally be in their ground state.

They can be excited to a higher state by a laser tuned to the
resonance frequency corresponding to the transition to that
state. The laser photons needed to excite the first excited
states for atomic and molecular (anti)hydrogen are in the
S.[ ['vacuum ultraviolet (VUV) region:

71
, . . ... 7i .

Atomic (anti)hydrogen Molecular (anti)hydrogen

X = 121.6 nm 110.9 nm

f = 2.47x10 1 5 Hz 2.72xl0 15 Hz
E = hf = 10.2 eV 11.2 eV

In order to be on resonance the laser must be tuned to a

frequency f just below the atomic resonance f by an amount
corresponding to the velocity v of the partic9 e:

5.2) f = f 0 (l-v/c)

The moving particle will then see the photon Doppler shifted
upward into resonance and absorption will take place with a
high probability. When the particle absorbs the photon and
jumps to the excited state, the photon will impart momentum
and energy to the particle, slowing it down by an amount:

5.3) dv = hf/mc = 3.27 m/s for atomic (anti)hydrogen

= 1.80 m/s for molecular (anti)hydrogen

Thus, it takes many thousands of absorptions of laser photons

to slow a typical particle down to near zero speed. After
each excitation process, the particle spontaheously re-emits a
photon. The frequency of the photon is the natural resonance
frequency so the emitted photon has more energy than the
absorbed photon. The energy difference, of course, comes from
the decrease in kinetic energy of the particle.
The spontaneously emitted photons impart a kick to the
particle when they leave, but the spontaneous photons are
emitted randomly in all directions. Thus, on the average,
there is no net momentum transfer to the particle by the
spontaneous emission process. The reradiation does, of
course, contribute to a random walk of the momentum of the
particle about its initial value, causing some transverse
heating as the particle is slowed in the longitudinal
direction.
In the usual particle beam, the particles have a range of
energies and a single frequency laser will cause spontaneous
radiation pressure only on those atoms with the right
velocity. Also, as the atom is slowed, it will drop out of
resonance with the laser. Thus, it is necessary to keep the
laser and particle in tune with each other.

72
 References:
5 "1
D.B. Pearson, R.R. Freeman, J.E. Bjorkholm, and A. Ashkin,
"Focusing and defocusing of neutral atomic beams using
resonance-radiation pressure," Appi. Phys. Lett. 36, 99-101
(1980).
5 " 2 j.E. Bjorkholm, R.R. Freeman, and D.B. Pearson, "Efficient
transverse deflection of neutral atomic beams using
spontaneous resonance-radiation pressure," Phys. Rev. 23A,
491-497 (1981).
A. Ashkin and J.P. Gordon, "Stability of radiation-pressure
particle traps: an optical Earnshaw theorem," Optics Lett. 8,
511-513 (1983).

5.2.1 Laser Chirp Tuning

One approach is to use a tunable laser and start with the
laser frequency well below the natural frequency. Photons of
this low frequency will interact with the high-speed particles
in the beam, slowing them down. The frequency of the laser is
now increased so that the slowed atoms plus those atoms with a
lower initial velocity are both slowed. As the laser
continues to "chirp" upward in frequency, it "sweeps" the
velocity of all the particles downward. Thus, not only are
all the particles in a beam slowed down, they are all slowed
down to the same velocity.
-. An example of the potential efficiency of this process is
shown in Figure 5-2. A Monte Carlo simulation was carried out
calculating the effect of a chirped laser on a beam of sodium
atoms. 5 - 4 The sodium atoms come from an oven with a wide
spread in initial longitudinal and transverse velocity. As is
shown in the "before cooling" and "after cooling" scatter
diagrams, the chirped laser sweeps all the atoms but two down
to zero longitudinal velocity. Thus, theoretically, this
method of slowing atoms should be highly efficient.
This slowing technique has been recently demonstrated on a
beam of sodium atoms. 5 - 5 A 10 mW laser was frequency swept
with a broadband electro-optic phase modulator using a LiTaO3
crystal in a traveling-wave configuration driven by 5 W of
microwave power. The modulator chirped at a rate of about
1.5 GHz/s and put about 34% of the laser light into the down-
shifted sideband.

73

""
-" " """ "~~~~*"" -* "" "** ""- * ." •"•..,""'" '"."""< " """. c• . . " C "" . - . '. ' "-"• • v v .
'I

4• 4

PM

Na.Oven
Cooling L
Atomic Beam

IS
- ---
VerY Slow Atoms
0.5mr

-
- 4t
If,

-',
lnlu
na
elocit llow
t
S,(/)lwlui
Atomsa

,-
d0 a S4
A - 11L 15

1ofI~tudU naI '•eloe:1ty VL (mis)

1 omtt-1,s1 vel oei:ty7
(mi
wL 1

BEFORE COOLING
AFTER COOLING

4574

Fig. 5-2 Particle slowing using

chir'oed laser frequency.
,

74

I-5**I- 4* C -.
,
 The chirped laser light brought the sodium beam to a stop and
converted it into a slowly expanding cloud of sodium atoms
with a density of about 106 atoms/cc and an expansion velocity
of about 6 m/s. This is equivalent to a kinetic temperature
of about 50 mK. In one experiment, the laser was deliberately
over-shifted in frequency and the atoms in the beam were
actually brought to a stop, then pushed back in the opposite
direction.

References:
5. 4 R. Blatt, W. Ertmer, and J.L. Hall, "Cooling of an atomic
beam with frequency-sweep techniques," pp. 142-153, Laser-
Cooled and Trapped Atoms, NBS SP-653, W.D. Phillips (editor)
(1983).
5.5W. Ertmer, R. Blatt, J.L. Hall, and M. Zhu, "Laser
manipulation of atomic beam velocities: Demonstration of
stopped atoms and velocity reversal," Phys. Rev. Lett. 54,
996-999 (1985).

5.2.2 Magnetic State Tuning

An alternate technique for lining up the atomic resonance of
the moving atom with the laser frequency is to use the Zeeman
shift to "tune" the separation of the atomic levels with a
magnetic field. 5 - 6 As is shown in Figure 5-3, the sodium
atoms from an oven were sent down the bore of a multi-layer
solenoid with a tapered magnetic field. The length of the
solenoid and strength of the field were chosen so that sodium
atoms with an initial velocity of 1000 m/s could be stopped in
about 1 r.with a deceleration about half the theoretical
ma-ximum given by the spontaneous decay lifetime tsp:

5.4) amax = hf/2mctsp

When a given atom reaches a point in the solenoid where the

combined Zeeman and Doppler shifts place it in resonance with
the fixed frequency slowing laser, the atom absorbs photons
and begins to decelerate. As long as the magnetic field does
not change so rapidly that the rate of Zeeman shift change
exceeds the possible rate of Doppler shift change, the atoms
will stay in resonance, slowing continuously as they travel
down the varying field of the solenoid. The cooling laser
beam is arranged to be converging so as to cool the beam
transversely as well as longitudinally.

"75
S.
 ~ ~w* * - ' *-

PMT
SOBSER\ATION
• •__ REGIO OPTICS
COLLECTION

VARYING-FIELD SOLNOI

NOLCOOLING

PROBE
LASERBEAM MECHANICAL BA

-"1 10 i t O cm 'r-40 cm-"-

cm

41)

9 REGION

u SWEPT
z

"VC

el-~

0 2 4 a 10 I2 4 i' VELOCIT-' (lOmr/s

Fig. 5-3 Particle slowing using variable magnetic field.

76
In practice, instead of bringing the atoms to a halt inside
the solenoid, the frequency of the laser is chosen so that the
atoms are all brought to some low finite velocity. This is
ahown in Figure 5-3, where the atoms in the initial
distribution with a broad peak from 800 to 1400 m/s were all
swept to a common velocity of 800 m/s.
After the initial cooling has taken place, the slowly moving
atoms now drift out of the end of the solenoid and are brought
to a stop some distance away by a short decelerating pulse
from the cooling laser. Recent results 5 ' 7 using this double
pulse technique have produced a cloud of free sodium atoms at
rest in the laboratory with a density of about 105 atoms/cc
and a velocity spread of about 15 m/s. This corresponds to a
kinetic temperature of less than 100 mK.

References:
5" 6 W.D. Phillips, J.V. Prodan, and H.J. Metcalf, "Neutral
atomic beam cooling experiments at NBS," pp. 1-8, Laser-Cooled
and Trapped Atoms, NBS SP-653, W.D. Phillips (editor) (1983).
5- 7 j. Prodan, A. Migdall, W.D. Phillips, I. So, H. Metcalf,
and J. Dalibard, "Stopping atoms with laser light," Phys. Rev.
Lett. 54, 992-995 (1985).

5.2.3 Pi-Pulse Laser Cooling of Neutral Antihydrogen

In the standard procedure for laser cooling of neutral atoms,
the atoms are illuminated with counterpropagating laser beams
which are tuned to the lower portion of the atomic resonance
line. Cooling results from the transfer of momentum from the
photons to the atoms during a resonant absorption-fluorescence
reradiation event. The cooling rate for this process is
limited by the natural decay of the atom due to spontaneous
radiation.
The maximum cooling rate occurs at laser power levels which
saturate the transition. An increase of laser power above
saturation does not increase the cooling rate. The first
photon absorbed will cause a transfer of momentum of -hk to
the atom as the photon is absorbed, raising the atom to its
excited state. A second photon, however, will cause simulated
emission from the excited state back down to the ground state,
and since the stimulated photon will be emitted in the
direction of the original photon beam, the transfer of
momentum to the atom will be +hk, cancelling out the original
momentum transfer from the absorption.

77
 This limitation on the ccoling rate can be overcome by
alternating the direction and detuning of two separate upward-
stimulating and downward-stimulating photon beams. Thus, a
net momentum transfer to the atom can be made to occur at-
transition rates greater than the spontaneous decay rate.
This can be done by using an alternating series of oppositely
directed and detuned population inverting (pi) pulses.
A pi-pulse is a short, intense pulse of radiation of electric
field strength E and duration t such that when it
interacts with an atom with a5 transition
8
electrical dipole
moment , the relationship -

5.5) 2tp'E/4 =

holds. As long as the pulse duration is much shorter than the

natural spontaneous decay time of the atom, such a pi-pulse
has a unit probability of causing a transition between two
states whose energy difference matches the radiation
frequency.
The use of pi-pulses for transverse deflection of an atomic
beam is shown in Figure 5-4. The atomic beam is assumed to
start in the ground state. First, an upward-going pi-pulse is
absorbed, exciting the atom to the upper state and
transferring an increment of momentum hk to the atom. Next, a
downward-going pi-pulse passes by the excited atom. It is not
absorbed, since the atom is in its excited state; instead, the
pi-pulse stimulates the atom to emit a photon downward,
resulting in another recoil momentum transfer of hk for a net
momentum transfer ,if 2hk. Thus, a train of alternating N pi-
pulses will result in a momentum transfer of Nhk to the atom.
By providing alternate pulses with the proper detuning, the
pi-pulses can be used not only for atom deflection, but for
atom cooling. The shaping of the pi-pulses in amplitude,
frequency, and length 5 9 , as well as the interval between
alternate pi-pulses 5 -1 0 are all variables that can optimize
their use for a particular application. In some applications,
the pi-pulses can be "reused" by reflecting them off
stationary or moving mirrors5"9 or by using optical cavities.
There is always some amount of spontaneous emission present
during pi-pulse operation. If the length of the pi-pulse is
short compared to the spontaneous emission decay time, the
effect is small, but does put some limits on the cooling rate
and limiting temperature that can be reached.

78

'.- %
 II,

1 TWO-LEVEL
ATOM
12>
Pi-PULSE
1~~2> Pi-PULSE
12>

1>> 4H 1>

Pi-PULSE DEFLECTED
P,." Pi = hk PHOTON ATOM
[" •" "• IP.L = 2 hk
PTT 2STIMULATED

I •,_.E 2Tp = I" Pi-PULSE PHOTON

Ph PHOTON

COOLING RATE MUCH FASTER THAN CW RATE

Pi-PULSE DURATION Tp MUCH SHORTER THAN DECAY TIME T2
T p << T2 =2A-1

Fig. 5-4 Pi-pulse deflection and cooling of atoms.

In general, pi-pulse cooling of atomic hydrogen is orders of

magnitude faster than standard laser cooling for beam
temperatures of 100 K or greater, and becomes comparable to
standard cooling for beam temperatures below 1 K.

Pi-pulse cooling will be especially useful in cooling of

molecular hydrogen. Right above the ground state of molecular
hydrogen there are a great number of rotational and
vibrational states. Since the hydrogen molecule is
homonuclear, only the even states are allowed. These states
are metastable, since the transition to the ground state
requires a change of two units of angular momentum and a
photon only carries one unit.
The presence of these states makes it difficult to use
standard laser cooling where the laser photon excites the
molecule from the ground state to the first excited electronic
state, but we depend upon spontaneous emission to return the
mclecule to the ground state. The spontaneous emission could
instead cause a transition to one of the metastable
-. vibrational or rotational levels and the molecule would be out
of resonance with the cooling laser and would no longer be
cooled. With pi-pulse cooling, tne excited molecule will be
."driven" to the ground state, bypassing the metastable
rotational and vibrational states.

79

---------------------

.------ -"
,*.,.. " '"
.---- "----..--..-.----'.---..-.----,i" " %•,. :•• " " - " -. '
When pi-pulse cooling is used with atomic or molecular
hydrogen, the simple two-state picture is no longer rigorously
valid. For both atomic and molecular hydrogen, the two-state
picture is modified since there is a finite probability that a
hydrogen atom or molecule in its excited state can be excited
by a second pi-palse into the continuum, causing the atom or
molecule to be ionized. Because the density of states in the
continuum is small, this rate will be small. How small is
unknown, however, and experiments are underway to measure this
double-photon ionization rate. 5 . 1 1

References
5.8H. Friedman and A.D. Wilson, "Isotope separation by
rad 4 -tion pressure of coherent pi pulses," Appl. Phys. Lett.
28, 270-273 (1976).
I. Nebenzahl and A. Sz8ke, "Deflection of atomic beams by
resonance radia ion using stimulated emission," Appl. Phys.
Lett. 25, 327-329 (1974)
5 "1 0
A.J. Palmer and J.F. Lam, "Radiation cooling with pi-
pulses," Paper WGl, Annual Meeting Optical Soc. Am., San Diego
(1984) [submitted to J. Opt. Soc. Am.].
5- 1 1 R.A. McFarlane, D.G. Steel, R.S. Turley, J.F. Lam, and
A.J. Palmer, "Experimental and theoretical studies of laser
cooling and emittance control of neutral beams," Hughes
Research, Malibu, California 90265, Annual report on contract
F49620-82-C-0004; AFOSR, Bolling AFB, DC 20332 (Ot 1984).

5.2.4 Optical Transitions in Atomic and Molecular Hydrogen.

Most of the research on cooling and trapping of neutral
particles has taken place with sodium atoms or other particles
with optical transitions in the visible region where tu-able
lasers are available. In order to slow and cool antihydrogen
atoms and molecules, these techniques will need to be extended
into the vacuum ultraviolet (VUV) region of the spectrum.
What are needed are tunable sources of VUV photons suitable
for the transitions expected in atomic and molecular hydrogen.

Partial energy level diagrams for atomic and molecular

hydrogen are shown in Figure 5-5. The spectrum for atomic
hydrogen in Figure 5-5(a) is very simple since we are only
dealing with a single electron about a single proton. The
major transition that will be used for cooling is the Lyman
alpha line with a wavelength of 121.57 nm.

80
 7 It 6 71. •7.r - -A
77 • •. . 77•. r

The spectrum for molecular hydrogen shown in Figure 5-5(b) is

much more complicated since there are many excited states of
the molecule as well as excited states in each of the two
atoms making up the molecule. 5 - 1 2 The major transition that
will be used for cooling is the 110.9 nm VUV line. At the
bottom of Figure 5-5(b) are some of the vibrationally ixcited
molecular states that exist just above the ground stat? of the
electronic transitions. Each electronic transition state in
the diagram has a similar set of vibrational as well as
rotational states. Thus, unlike the hydrogen atom, the
hydrogen molecule has a multiplicity of possible transitions
and many of them have been observed to produce laser action.
The major laser lines for molecular hydrogen are in the
ultraviolet and go from the first excited state to one of the
vibrational levels in the ground state (B-4X). This
transition gives 39 lines that range from 127.95 nm (0-3)P(2)
to 164.60 nm (4-11)P(4). 5 . 1 3 The C-4*X transition is also
active with 11 lines observed ranging from 109.82 nm (0-2)R(0)
to 125.20 nm (6-10)R(0).
There are also some short infrared lines observed that go from
one excited state to another, such as the 11,159 nm transition
from E-IB. Thus, one source for the laser radiation needed
for laser cooling of molecultr hydrogen could come from
-• molecular hydrogen lasers. 5 14 These will not be easily
";.unable, however, and it might ultimately be found better to
i,'se nonlinear optical mixing to produce a tunable coherent
vacuum ultraviolet light source.

References
5 . 1 2 G. Herzberg, Molrc,•la Ipectra and Molecular Structure -

I. Spectra of Diatcoic * eules, 2nd Ed., D. Van Nostrand,

Princeton, NJ (1950).
5"1 3 M.j. Weber, Ed., CRC Handbook of Laser Science and
Technology, pp. 282-286, CRC Press, Boca Raton, Florida.
L.N. Breusova, I.N. Knyazev, V.G. Movshev, and T.B.
Fogel'son, "Vacuum ultraviolet H2 laser with a sealed gas-
discharge cell," Kvant. Elektron. 6, 2458-2460 (1979) (English
translation Soy. J. Quantum Elec. 9, 1452-1453 (1979)].

81
 SINGLET CONFIGURATIONS

I,ONIZATION STATES

15.43 / 124.429
15- is
IONIZATION STATES 4 G
13.53 1Z•Z////•&// 109,126 3- F 3-0
9 5
10-0 2 E
3-- 100l0 2,-----2c- 100o0

10.2 a--

10 ,10
BALMER
SERIES

eV E cm-1' V
E1. cOl 50.006,00o
5-tr '
E ~101 111 nm
m
5 R

6
54 VIBRATIONAL
3 LEVELS
-2
0 0 0 -X 0

LYMAN SERIES

ATOMIC MOLECULAR
HYDROGEN HYDROGEN

(a) (b)

Fig. 5-5 Energy levels of atomic and molecular hydrogen.

82
 ,:-> 5.2.5 Non-Laser Sources of Monochromatic Radiation

An alternate way to produce the desired tunable VUV photons

for the cooling and trapping of atomic and molecular hydrogen
is to manufacture the photons by nonlinear optical mixing.
The concept of optical doubling and tripling is well known in
laser technology.
In this procedure, a high intensity source of photons, such as
the 1064 nm infrared photons from a YAG laser, is focused
into an optically nonlinear crystal. The light intensities
drive the optical material far into their nonlinear range and
frequency doubling takes place with high efficiency, producing
one 532 nm wavelength photon out for every two 1064 nm photonb
put in. With proper attention paid to choosing the crystal
axis so that the two colors of light have the same velocity in
the crystal, the efficiency of conversion can approach 100%.
By using a more general version of nonlinear mixing called
four-wave mixing, it is possible to mix two or more photons to
produce an output photon with an energy that is the sum of tne
input photon frequencies. The nonlinear material does not
have to be a crystal, but can be any nonlinear optical
phenomena including a convenient optical transition in some
atom or molecule that is near the desired sum frequency.

In Figure 5-6(a) is shown one method for making a tunable

Lyman alpha source of narrow band, spatially coherent
photons. It is not a laser, since no stimulated emission
is taking place. Two tunable dye lasers make photons at the
frequencies of 545.37 nm and 625.52 nm. These low-power
signals are then amplified by dye laser amplifiers pumped by
532 nm photons from a frequency doubled 1064 nm YAG laser.
The 625.52 nm photons are frequency doubled to 312.76 nm by a
nonlinear crystal and sent to a cell of mercury vapor which
acts as the last nonlinear medium.
As is shown in Figure 5-6(b), mercury has a convenient optical
transition line that allows two of the 312.76 nm photons and
one 545.37 nm pnctsn to add to produce one 121.57 nm photon.
This photon is tunable around the atomic hydrogen Lyman alpha
transition frequency since the original dye laser signal
sources are tunable. A similar technique can be used to
produce the 110.9 nm photons needed for cooling and trapping
of molecular hydrogen.

Referencea:
"* 5 "1 5 D.G.
Steel, Hughes Research Laboratories, Malibu,
California, personal communication.
-'.

83

v4
 .- V~
w rvv
y ~ .~r~ ~ Ir.* w.
wl v-; y- .0 4-*-s7

(a)

6255.24
CW DYE 6255.2A DYE LASER 16 mJ
LASER AMPLIFIER
SOURCE

5320 325 mj POWER LIMITED

BY TWO-PHOTON
YAG 1.06 j~m SATURATION
PUMP / /
S~LASER

5320 A 325 mj
3127.6 A 1 mJ LYMAN a
CW DYE 5453.7A MERCURY 1215.7
SOURCE 54URCE AMPLIFIER
DYE LA 54637\ CELL lo~ j

(b)
6s12p 1 P1

S.... F • • • 82259.272 cm-1

6sllp 1P
1

5453.7 A

6s7s 1S0 63928.243 cm-1

LYMANa
1215.7A

3127.64

6s 2 1 S0
Hg
Fig. 5-6 Tunable Lyman alpha source using four-wave mixing.

84
 SECTION 6

TRAPPING AND STORING ANTIHYDROGEN

6.1 TRAPPING OF ATOMIC ANTIHYDROGEN

The hydrogen atom consists of a proton and an orbiting

electron. Both the proton and the electron have a magnetic
moment and can interact with an applied magnetic field.
The proton magnetic moment or nuclear magneton is given by:

6.1) Pn = eh/4TIMc = 5.05xi0- 2 4 erg/G (cgs)

= eh/4m4 = 5.05xi0- 2 7 J/T (MKS)

"where e is the electronic charge, h is Planck's constant, c is

- the speed of light, and M is the mass of the proton.
*4 The electron mass m is 1836 times smaller than the proton
mass, while the angular momentum a-4 charge are the same.
Thus the electron magnetic moment or Bohr magneton is 1836
times larger than the nuclear magneton:

6.2) /e = eh/4mc = O.927x10 20 erg/G (cgs)

= eh/4lTm = 0.927xi0- 2 3 J/T (MKS)

When a hydrogen atom is put into a magnetic field, the

magnetic moments can align with the magnetic field either in
the parallel (low energy) direction or antiparallel (high
energy) direction and are said to be polarized. Those atoms
with both magnetic moments in the same direction are called
doubly polarized hydrogen. 6 . 1

Atoms in the lower energy states are drawn into the magnetic
field by the magnetic field gradients, while atoms in the
higher energy states are repelled. The magnetic trap is not
stable. The axial gradient in the magnetic field holds the
atoms along the axial direction, but the atoms can drift
-. radially across the magnetic field lines and strike the walls.
A wall coating of superfluid liquid helium is used to reflect
the polarized atoms back into the trap without flipping their
spins. 0.•

85
 A collection of cold doubly polarized atoms in a magnetic
field will have a tendency to stay in the atomic form since
the electrons have the same spin orientation. If the atoms
attempted to form a molecule with the two electrons in the
ground state, the Pauli exclusion principal would be violated.
If the two atoms are kept cold, they cannot form the ground
state and do not have enough energy to form in an excited
molecular state, and so remain as separate atoms.
The maximum density of doubly s@in-polarized atomic hydrogen
achieved to date has been 2x10 1 atgm /cc in a very small
"bubble" in liquid helium at 0.7 K. - Densities of this
magnitude would be of interest for storing atomic antihydrogen
since 12 mg of spin-stabilized antihydrogen could be stored in
a 10 cm sphere. Unfortunately, the lifetimes at these
densities were only a few minutes and were limited by wall
losses and three-body dipolar recombination processes.

Other experimenters have demonstrated confinement times of

4 hr for densities of 8x10 1 6 atoms/cc with a 10 T magnetic
field and a temperature of 0.3 K. The lifetime increases
dramatically with increasing magnetic field since the atoms
are kept away from the walls. The lifetime also increases
with decreasing density, which lowers the probability of a
three-body recombination. 6 - 3 If the lifetime problems could
be solved, even a density of 1015 atoms/cc would be
interesting for antiproton annihilation propulsion, since
10 mg of antihydrogen could be stored in a one-meter sphere.

The need for normal matter walls made of helium to complete

the trap makes the present trapping methods unusable for
antimatter. New ideas for magnetic traps to contain spin
polarized atomic antihydrogen are needed. Since static fields
are not stable, some kind of dynamic field arrangement needs
to be invented. It is conceivable that some kind of
"* alternating magnetic field configuration using an analogy to
the Paul rf electrostatic trap might work.
One configuration proposed for the tragping of spin polarized
atoms is the hybrid laser-magnet trap.6-4 As is shown in
Figure 6-1, a strong magnetic field provides axial trapping of
the spin polarized atoms, while the radial tra~ping is
provided by a laser cavity operated in the TEM 0 1 mode. This
mode is circular with a null in the center and most of the
radiation is confined to a cylinder. It is a natural mode of
some laser cavities. If the laser is tuned slightly to the
"blue" of the atomic resonance frequency, the atom will
experience a relatively strong "transverse dipole" force
pushing it into the central region of lower light intensity.
With an 11 T (110 kG) magnet and a highly reflecting cavity to
store the light energy, the "depth" of the trap is estimated
to be around 7 K. 6 . 4

86
 MIRROR

,4

SUPERCONDUCTING TO EM-l
MAGNETTE 1
,,-Ooo LASER
BEAM

LASER
AMPLIFIER
op HYBRID
ooooý/ LASER-
MAGNET
TRAP
MIRROR

- Fig. 6-1 Laser-magnet trap for polarized antihydrogen atoms.

87
Referenlces:
6"II.F• Silvera and J. Walraven, "The stabilization of atomic
hydrogen," Scientific Am. 246, 66-74 (January 1982).
6.2R. Sprik, J.T.M. Walraven, and I.F. Silvera, "Compression
of spin-polarized hydrogen to high density," Phys. Rev. Lett.
51, 479-482 (1983).
6" 3 H.F. Hess, D.A. Bell, G.P. Kochanski, R.W. Cline, D.
Kleppner, and T.J. Greytak, "Observation of three-bodz
recombination in spin-polarized hydrogen," Phys. Rev. Lett.
51, 483-486 (1983).
• Stwalley, "A hybrid laser-magnet trap for spin-
polarized atoms," pp. 95-102, Laser Cooled and Trapped Atous,
W.D. Phillips, ed., National Bureau of Standards SP-653
(1983).

6.2 TRAPPING OF ANTIHYDROGEN

Lasers can be used not only for slowing, cooling, and stopping
of particles, but for trapping them. One example of an
"optical trap is shown in Figure 6-2. This trap uses both the
scattering force due to spontaneous emission and the
ponderomotive force from the coupling of the induced atomic
dipole to the optical field gradient.

In Figure 6-2 two opposing TEMoo mode cw laser beams are

focused at the two focal points F1 and F2 located
symmetrically about the trapping point E. 6 " 5 The beams grow
in radius from wo from the waist at the focal points to 13 wo
in going 1 cm from the focal point to the trapping point.
Each beam is tuned some 50 times the natural linewidth below
the resonance frequency. For sodium atoms and 590 nm yellow
light, wo=1 2 )&m and the cw laser power is 200 mW.
The trapping point at E is a point of stable equilibrium since
any displacement of an atom from E results in a re3toring
force. There is an axial spontaneous scattering restoring
force due to the oppositely directed laser beams and a radial
restoring dipole force due to the radial optical field
gradient.
To trap the atoms after they have entered the trap requires
damping. Damping due to the Doppler shift occurs when the
laser is tuned below resonance since the moving atoms interact
more strongly with the opposing beam.

88
 H "---ac--+imi
• (a)

C
-2

S.0 -

10

z
> 0
-3.0 -2.0 -1.0 0 +i.O +2.0

AXIAL POSITION z(cm)

Fig. 6-2 Trapping of atoms by resonance radiation pressure.

"89

*
 The atoms are inserted into the trap through the hole H in the
mirror M. An atom entering the trap with a velocity v=200 m/s
stops at a point some 4 mm beyond F2 . It then rebounds and
executes a damped oscillation about E, cooling off as it is
trapped. Several cycles are shown in Figure 6-2(b).
The original optical trap design was later analyzed in more
detail. .6 It was found that the concept produces rapid
damping of fast atoms injected into the trap, but slow atoms
held in these traps are not optimally damped because of large
detuning. It was then found that the addition of a damping
beam decreases the trapping force, while the presence of a
trapping beam decreases the damping force. In addition, the
ttapping beam causes an optical Stark shift of the atomic
resonance, which further complicates the damping process.
Despite these problems, an optical trap using optimally tuned
trapping and damping beams was predicted to result in traps
capable of confining sodium atoms at temperatures as low as
0.1 mK in optical potential well as deep as 0.1 meV, within a
*• dimension as small as X/35, with negligible probability of
escape by tunneling or thermal excitation.
There are a number of methods to use a multiplicity of damping
and trapping beams to improve the trap. One proposal is to
alternate the cooling and trapping beams in time so that they
do not conflict with each other. Another is to have the
cooling beam and the trapping beam at different frequencies
and have them cause transitions between different states. A
third is to use right- and left-handed circularly polarized
light for a transition from a J=0 state to the M=+1 and M=-1
levels of a J=l state. 6 " 6 Some of these concepts may be
applicable to the problem of trapping of hydrogen atoms and
molecules.
Estimates have been made 6 - 7 that by using these traps, up to
3x10 7 atoms can be cooled down to less than a millidegree
* Kelvin and be confined to a dimension as small as X/35. At
'- these temperatures the gas density approaches that of a solid
- and the atoms may convert into the condensed phase. For
-.• sodium atoms, the condensed phase is a metal and the trapping
mechanism no longer will operate. For antihydrogen the
condensed phase will be a transparent dielectric material and
the laser trapping will still be operative, although the
details of the trapping mechanism will be different.

90
"V A
neutral highly transparent dielectric particle has been
trapped, levitated, and manipulated using an alternating light
beam approach.- 8 The actual particle used in the experiments
was a sphere of silicon oil 9 m diameter and weighing about
0.6 ng. The damping was supplied by air. The particle was
suspended and trapped by an optical trap similar to that shown
in Figure 6-2, but as is shown in Figure 6-3, the directions
of the trapping beams were reversed periodically.

There is a direct analogy between the dc trap of Figure 6-2

and the ac trap of Figure 6-3, and the dc Penning ion trap in
Figure 2-3 and the ac Paul ion trap in Figure 4-1. In the ac
traps, the dynamic stability is governed by the Mathieu
equations and avoids the consequences of the Earnshaw
stability theorem for static field traps.
The alternating field trap has only been demonstrated on air-
damped silicon oil drops. The technique should be applicable,
however, not only to macroscopic Rayleigh-sized particles, but
to neutral atoms and molecules. For atoms, large-volume traps
with dimensions of 1 cc are predicted with greatly reduced 8
optical cooling problems and well depths as great at 1 K".
Since liquid and solid antihydrogen are transparent dielectric
materials, it is quite likely that these types of optical
traps can not only be used to cool and compress the
antihydrogen molecular vapor to densities approaching that of
the condensed phase, but also maintain the trapping action on
the particles after the condensation has taken place.

References
6 . 5 A.
Ashkin, "Trapping of atoms by resonance radiation
pressure," Phys. Rev. Lett. 40, 729-732 (1978).
6 . 6 A.Ashkin and J.P. Gordon, "Cooling and trapping of atoms
by resonance radiation pressure," Optics Lett. 4, 161-163
(1979).
b.Tj. Dalibard, S. Reynaud, and C. Cohen-Tannoudji, "Proposals
of stable optical traps for neutral atoms," Optics Comm. 47,
395-399 (1983).
6 "8 A. Askin and J.M. Dziedzic,
"Observation of radiation-
pressure trapping of particles by alternating light beams,"
Phys. Rev. Lett, 54, 1245-1248 (1985).

91
 AA
- * - - -- - @ -

... .. .. ...... .. ... .. .. ... e •..---

A. Fig. 6-3 Geometry of alternating-beam optical trap.

92
 "6.3 CONVERSION OF ANTIHYDROGEN GAS TO ICE
Trapping the parahydrogen gas will leave the molecules at very
low kinetic energies, equivalent to millikelvin or
"microkelvin. The next step is to convert the gas into either
amorphous or crystalline ice by inverse sublimation. Research
is needed on techniques to induce the process of nucleation
directly from the gas without involving a wall.
The experiments would investigate the effect of various
electromagnetic fields on the nucleation process, including
the precursor phase where two, three, and multiple-molecule
collisions produce the dendritic "snowflake" clusters that
precede the formation of the crystalline solid. It may be
that "shock" waves of density fluctuations set up by the laser
trapping fields are sufficient to induce the nucleation.
Molecular hydrogen ions and specially selected excited states
may aid in the nucleation and growth of these clusters into
crystals. The rate of crystal growth should be measured as a
function of the temperature and effective pressure of the
trapped molecular gas.
Since the process of forming the solid releases energy, it
will also be necessary to investigate methods for extracting
tne heat from the solid. A straightforward approach is to
allow the solid to evaporate a molecule containing the energy
from a number of fusions, then cool the molecule with the
"laser slowing and cooling process. The cooled molecule would
then be directed back into the cold gas. Alternate techniques
could conceivably use active laser or magnetic "cooling"
techniques that work through a nonlinear optical or magnetic
property of the parahydrogen solid. All of the research
should concentrate on techniques that leave the solid at
temperatures of 1 K or lower.

It is important to remember that, as shown in Figure 6-4,

there are many paths that lead from the generation of an
antiproton beam to the generation of a frozen ball of
molecular antihydrogen. The antiprotons can be trapped in a
Paul trap, turned into antihydrogen atoms by the addition of
positrons, trapped with Lyman alpha laser beams, converted
into magnetic field-trapped antihydrogen molecules, tickled
into forming an iceball by positron beams and laser beams,
then grown into a charged microcrystal.
Alternatively, the antiproton beam can be converted into an
atomic antihydrogen beam, then a molecular antihydrogen beam,
all on the fly; then the antihydrogen molecules are slowed,
cooled, trapped, and turned into ice with VUV laser beams. A
few extra positrons will prepare the crystal for levitation in
an electrostatic trap.

93

-p.ý
 j~BEAM

TRAP j FORM ABEAM

FORM H TRAP H FORM H2 BEAM

FORM H2 - FORM H2 TRAP H2

GROW CHARGED H2 CRYSTAL

LEVITATE CHARGED CRYSTAL

Fig. 6-4 Many paths from antiprotons to antihydrogen ice.

6.4 LEVITATION OF ANTIHYDROGEN ICE

The techniques for trapping parahydrogen gas and the

subsequent formation of solid parahydrogen may turn out t. be
relatively simpJe, or they may require complex ultrahigh
vacuum chambers with many ports and windows, high-power
lasers, and heavy electric or magnetic field generators. Once
the small microcrystals or larger ice balls of parahydrogen
ice are formed, however, they can be transferred to a compact
lightweight vacuum system that uses simple magnetic or
electric traps for levitation.
The magnetic susceptibility of solid hydrogen depends upon its
state. The orthohydrogen form has both of the protons in its
nucleus with their magnetic moments pointing in the same
direction, so it has a positive magnetic moment. The
parahydrogen form has its two protons and its two electrons
with their spins oriented in opposite directions so the
particle spins cancel out. The only magnetic susceptibility
left comes from the *currents" caused by orbital motion of the
electrons around the nucleus.

94
 As a magnetic field is applied to these "superconducting"
currents, the current tends to increase, driving the impressed
magnetic field out of the molecule. This gives the
parahydrogen molecule a negative or diamagnetic
susceptibility. Diamagnetic substances are attracted to the
minimum in a magnetic field. Even with purely static magnetic
field, the configuration is stable, unlike levitation systems
based on repulsion of paramagnetic or ferromagnetic materials,
which seek a field maximum and are unstable.

The magnetic susceptibility of solid parahydrogen has not yet

been measured. 6 - 9 The theoretical prediction of the magnetic
susceptibility of a one-gram formgl• weight of molecular
hydrogen is -3.98xl0- cgs units.OuO This can be compared
to a value of -6.0x10 -6 cgs for carbon atoms in the form of
graphite.

There have been many demonstrations of the stable levitation

of many grams of graphite in the gravity field of earth using
nonsuperconducting magnets. In one specific example, a ring
shaped rotor weighing 3.843 g and containing only 0.933 g of
graphite was levitated in the 0.2 cm gap between two opposed
ring shaped permanent magnets with a magnetic field of
11,600 G (1.16 T). 6 . 1 1
One straightforward configuration for a magnetic trap that
would be compatible with a cryogenically cooled vacuum chamber
would be a pair of superconducting rings carrying opposed
persistent currents as shown in Figure 6-5.
It is recommended that a research project be started on the
magnetic properties of hydrogen. This research project would
measure the magnetic susceptibility of parahydrogen at low
temperatures as a function of temperature, particle size, and
orthohydrogen contamination. The next step would be to
demonstrate that the magnetic field gradients from persistent
currents in a set of superconducting coils are sufficiently
strong to levitate the parahydrogen at moderately high
acceleration levels.
This research is not path critical, since there exists a known
technique, active electrostatic levitation, that can support a
charged ball of low density parahydrogen with ease at high
acceleration levels. The passive magnetic levitation
technique would be preferred, however, since it would be safer
than any active technique, as well as making the storage
container extremely simple, compact, and independent of
electric power.

95

I-
 INT

SUPERCONDUCTING

PERSISTENT
CURRENT

DIAMAGNETIC- .

ANTIHYDROGEN

DIAMAGNETICS
ATTRACTED TO
FIELD MINIMUM

Fig. 6-5 Magnetostatic trap for antiparahydrogen ice.

An alternate method of levitating antihydrogen ice is to use

electrostatic levitation. The ice particles need to be
slightly charged, either positive or negative. This can be
accomplished either by charging the ice positive by addition
of extra positrons or charging it negative by annihilating
some of the positrons with electrons from an electron gun or
driving off the positrons with ultraviolet light.
The well known Earnshaw Theorem states: "A charged body
placed in an electric field of force cannot rest in stable
equilibrium under the influence of the electric forces
alone." 6 -1 2 This means that an electric levitation system has
to have an active means of maintaining sufficient charge on
the antihydrogen ice particles, as well as an active position
control loop to maintain the particles in the center of the
trap.

96

+ , ' 'V u h + '_- •c - + S.....

*- . * - =- 4 r -- . ' ' a
 -_-~
-*-- .- - rr .- .- ~ .~-'-- . .. •-

If the direction of acceleration is constant, a charged ball

can be levitated between electrically charged plates as is
shown in Figure 6-6, provided that the plates have a slight
curvature. Such a trap has been constructed at the Jet
Propulsion Laboratory 6 - 1 3 and has levitated a 20-mg ball of
water ice in the one gee field of the earth. The density of
water at 1.0 g/cc is 13 times that of antihydrogen ice at
0.0763 g/cc. Thus, the present JPL trap with its present
voltage levels could levitate a 1.5-mg ball of antihydrogen
ice of the same size, surface area, and surface charge at a
vehicle acceleration of 13 gees.

References:
6 "9 N.A. Olien, NBS, personal communication (1984).
6 * 1 0 R.C.
Weast, CRC Handbook of Chemistry and Physics, 63rd
Edition, p. E-119, CRC Press (1983).
6 *1 1 R.D. Waldron, "Diamagnetic levitation using pyrolytic
graphite," Rev. Sci. Instr. 37, 29-34 (1966).
6 .1 2 j. Jeans, Sir, The Mathematical Theory of Electricity and
Magnetism, Cambridge Press, UK (1925).
6- 1 3 W-K Rhim, M.M. Saffren, and D.D. Elleman, "Development
electrostatic levitator at JPL, pp. 115-119, Material
Processing in the Reduced Gravity Environment of Space, G.E.
Rindone, ed., Elsevier (1982).

ELECTRON GUN

• = POSITION

CONTROLLER

SOURCG ANTIMATTER BALL WITH DETECTOR

SLIGHT EXCESS CHARGE

K I

Fig. 6-6 Active electric levitation of antihydrogen ice.

97
S,°
 6.5 ANTIHYDROGEN ICE BALL ENERGY BALANCE -'

As has been shown in previous sections, there are a number of

active techniques to remove energy from antihydrogen so that
it will turn into a cold vapor, condense, and form
antihydrogen ice. Since laser cooling, for instance, will
leave the antihydrogen at a temperature well below a
millidegree, the antihydrogen ice will start out cold.
In this section we will determine whether the initially cold
antihydrogen ice can be kept cold by passive cooling
techniques despite unavoidable heating due to annihilation
processes. For example, antihydrogen molecules could
evaporate off the ice ball and drift over to annihilate on the
walls of the chamber.
Then there could be hydrogen molecules or other normal
molecules knocked off the walls of the vacuum chamber by those
annihilation processes or cosmic rays that would drift across
the vacuum chamber to annihilate on the surface of the
antihydrogen ice ball.

6.5.1 Infrared Radiation Passive Cooling

The passive cooling technique we will use for this analysis is
infrared radiation to the walls of the cold containment
chamber as is shown in Figure 6-7. The amount of thermal
power P radiated from an object of temperature T, area A, and
emissivity E to the surroundings at temperature t and
emissivity e is given by the relation:

6.3) P = WA(E T 4 - e t 4 )

where the Stefan-Boltzmann constant is

6.4) == 5.67 x 10-8 W/m2 -K4

At the temperatures we will be discussing, the peak of the

infrared radiation is in the very long infrared cr short
microwave region. At a temperature of 2 K, for example, the
peak of the infrared radiation is at 1.4 mm wavelength. Very
little is known about the emissive properties of materials in
this region.

98

•o~
 .M M

SUSPENSION
ELECTRODES

i;• //• ~~INFRARED /er

OK /'•RADIATION 7rr+
10 PT 7

7r+
)P

..-. Fig. 6-7 Energy balance for antihydrogen ice ball.

99

•, ••-. " •- . .,- •.. .. .... ... '. ..- ..--- •..:,• •_- y ,- '•,,•• o,:@'• • )'4•
 Also, the iceball and the vacuum cavity are comparable in size
to the infrared wavelengths, which will complicate the
interaction. Obviously, further research is needed in this
area. For now, we will assume that the infrared radiation
power is given by Equation 6.3.
If we assume a 1-mg ball of antihydrogen ice at a density of
p=0.08 gm/cc, then the radius of the ball is r=0.15 cm and the
surface area is A=0.28 cm2 . The emissivity of hydrogen ice in
the long infrared is unknown, so we will assume a grey body
with E=0.5, and that the walls of the chamber can be made
black with an emissivity of e=l. The vacuum chamber will be
assumed to be 2 cm in radius.
To get maximum cooling power, the antihydrogen iceball should
be as warm as possible and the walls of the chamber should be
- as cold as possible. As is shown in the section on the
"* properties of hydrogen ice in Appendix A, 6 ,1 4 the vapor
pressure of solid hydrogen drops precipitously below 4 K. We
will choose a tentative equilibrium temperature for the
*- antihydrogen ice of 2 K. At this temperature, the vapor
pressure is only 4x10- 1 8 Torr or roughly 0.1 atom/cc.
The velocity of the evaporated molecules at the sublimation
temperature of 14 K is 340 m/sec. This gives a flux of 3400
mol/cm2 -s, which means that 950 antihydrogen molecules per
second leave the 0.28 cm2 surface area of the antihydrogen
iceball to strike the walls of the chamber and annihilate.
.- The temperature of the walls of the chamber has to be less
than the temperature of the ice ball to obtain infrared
"radiation cooling. To make it easier on the cryogenic cooling
system (which will be a magnetic dilution or paramagnetic
refrigerator), we will assume a wall temperature of 1 K.
The major outgassing contaminant in vacuum systems is
hydrogen. The vapor pressure of hydrogen at 1 K is 8.3x10- 3 9
Torr, which is completely negligible. If helium leakage
proves to be a problem, the temperature of the chamber walls
should be lowered to 0.1 K, where the vapor pressure of helium
- is 10-47 Torr.

- With the ice ball at 2 K and the walls of the chamber at 1 K,

the cooling power for a 2.8x10- 5 m2 area ball with emissivity
E=0.5 is:

6.5) Pc = llxlO- 12 W - 11 pW.

100

1.L- .* * ''
6.5.2 Energy Deposited by Annihilation Particles

Energy from annihilations can be deposited in the antihydrogen

ice ball by the 0.511 MeV gamma rays from positron-electron
annihilation, the 250 MeV charged pions from the pp
annihilation, and the 200 MeV gamma rays from the decay of the
neutral pions from the pp annihilation. Because the pion
decay gamma rays are so much more energetic, we will ignore
the positron-electron annihilation gamma rays.

For 200 MeV gamma rays, the attenuation coefficient in matter

is roughly constant at $/p=0.l cmI/g. Since the density of
antihydrogen is V=0.0763 gm/cm3 , the attenuation per unit path
length is only )&=0.0076/cm. Thus, instead of causing intense
local heating, most of the gamma rays pass right through the
iceball and continue on through the chamber iall to deposit
their energy in the outside shield.
The fraction of the total average gamma ray energy of
E =200 MeV deposited in traversing the d=0.3 cm diameter of
tNe iceball is only:

6.6) dEg = (1 - e-•d)Eg = = 460 keV = 7.4x10- 1 4 J

For the 250 MeV charged pions, the stopping power is

essentially flat at S=15 MeV/(g/cm2 ). The pions will leave a
small amount of their energy in the iceball and chamber walls
and deposit the rest in the outside shield. The energy
deposited along a charged pion track traveling d=0.3 cm 3
through an antihydrogen ice ball with density ?=0.0763 gm/cm
is:

6.7) dEp = Sad = 340 keV = 5.5x10- 1 4 J

We will make the assumption that there are 3.0 charged pions
and 1.5 neutral pions per annihilation of each antiproton (see
Section 7.3 on Particle Production from Antiproton
Annihilation). Thus, the annihilation of each antihydrogen
molecule will produce 6.0 250 MeV charged pions and 6.0
200 MeV gamma rays (plus 4.0 0.511 MeV positron-electron
annihilation gamma rays which we will ignore).
Since the annihilation gamma rays and pions have such great
penetrating power, failure of the antihydrogen trap will
probably result in a "meltdown" of the antihydrogen container
and shielding rather than a violent explosion. A trap failure
would be extremely serious, however, and further studies need
"to be done on antimatter trap failure modes.

101
 I- 'v "W

6.5.3 Heat Input frum Annihilations on Chamber Wall

If the annihilation takes place from a sublimated antihydrogen

molecule hitting the chamber wall, the reaction products will
be emitted in all directions. If we assume the chamber is
R=2 cm in radius, then the r=0.15 cm radius ice ball will
intercept
2
6.8) a/A = Trr 2 /4TrR 2 = r 2 /4R = 1.41x10-3

of the particles. Although most of the particles will not

pass through the whole diameter of the ice ball, we will use
the worst case example. The total energy deposited in the
iceball for each antiproton annihilation on the wall of the
chamber is then:
6.9) dEw = (6 dEg + 6 dEp)a/A = 1.3x10- 1 5 J

The 950 molecules per second leaving the ice ball will thus
produce a total heating power in the ice ball of:

6.10) Pw = 1.2xi0- 1 2 W = 1.2 pW

This heating power due to wall annihilations is significantly
less than the cooling power of 11 pW, so the evaporation of
antihydrogen from the ice ball at 2 K will not cause excessive
heating.

6.5.4 Heat Input from Annihilations on Surface

Although thermal analysis indicates that there should be no
significant number of normal molecules left in a cryogenically
cooled chamber, there could conceivably still be some normal
molecules released by nonthermal processes such as natural
radioactivity or cosmic rays. These normal molecules will
annihilate right on the surface of the antihydrogen ice ball.
* We will want to estimate how many of these annihilations per
second the passive cooling system will be able to handle.
In a surface annihilation half the particles will pass through
the ice ball and the other half will head toward the chamber
walls and pass through. The energy deposited in the ice ball
for each annihilation on the surface is then:
13 J
6.11) dEs = (3 dEg + 3 dEp) = 4.6x10-

Although there should be no molecules left in a well-made

vacuum chamber held at 1 K, we see that up to 10 annihilations
per second can take place right on the surface of an
antihydrogen ice ball without the heating power exceeding
radiative cooling power at 2 K.

102
Obtaining a better estimate of this heating source will
require research on hypervacuums in cryogenically cooled
chambers. Probably the best way to obtain this information is
to trap a few antiprotons from the Low Energy Antiproton Ring
at CERN in a cryogenically cooled Penning trap and observe
their lifetime.
We know that a single antielectron (positron) has been kept in
a cryogenically cooled Penning trap tor a month. 6 - 1 5 But the
annihilation cross section for a free positron and a bound
electron at these very low energies is not known well enough
to establish a firm lower bound on the vacuum level in the
trap. Plans are underway to trap an antiproton at LEAR in
1986-7 which should provide information on hypervacuums V1
well as the various properties of the antiproton itself. .16

6.5.5 Cooling Rate.

If the heating power from annihilations can be made
significantly smaller than the infrared cooling power, the ica
ball should cool off. As is noted in Appendix A on the
properties of solid hydrogen, the specific heat of
antihydrogen at these low temperatures is quite small and
decreases rapidly with temperature. At 2 K the heat capacity
is C=4.8xi0-3 cal/mole-K or 1.0x10- 2 J/gm'K. Thus, a cooling
power of Pc=dEc/dt=10 pW for a 1 mg antihydrogen ice ball
would produce a cooling rate of:
6.12) dT/dt = (C m) 1 dE/dt = ix10 6 K/s

or about 0.1 K/day. As the antihydrogen cooled below the

assumed 2 K, the evaporation rate of antihydrogen, being an
inverse exponential function of the temperature, would drop
rapidly and one of the main sources of heat would become less
important. As the temperature dropped, however, the infrared
cooling power would also drop, but only as the fourth power of
the temperature. The temperature of the ice ball would
finally stabilize somewhere below 2 K and above the
temperature of the chamber walls.

6.5.6 Alternate Cooling Methods

It may turn out that the emissivity of antihydrogen is too low

in the long infrared region to allow significant cooling power
by emission of radiation. In that case, alternate cooling
methods will have to be found. Three possible examples are
laser cooling, electrodynamic cooling, and magnetodynamic
cooling.

103

1Z
In laser cooling, the storage cavity would be larger and laser
beams would be used to intercept the antihydrogen molecules
evaporating from the surface of the ice ball. The lasers
would cool off the molecules and direct them back into the
iceball using the techniques discussed in the section on laser
slowing and cooling of antihydrogen.
In electrodynamic cooling, the iceball would be kept highly
electrostatically charged. The vibrational excitation of the
normal modes of the antihydrogen iceball by the internal heat
would cause vibrational motion of the excess positrons on the
surface of the iceball. These vibrating charges would cause
changes in the image charges in the suspension electrodes.
The changing currents in the electrode control circuits would
then be damped out with a cryogenically cooled resistor as is
presently done with trapped ions a Penning trap, 6 ' 1 7 or an
active "cold damping" circuit.
In magnetodynamic cooling, the entire trap region would be
imbedded in a strong magnetic field. The antihydrogen, being
diamagnetic, would be coupled to the magnetic field lines.
Heat vibrations in the normal modes would then cause
fluctuations in the magnetic field, which in turn would cause
fluctuations in the current running through the magnet. These
current fluctuations could then be damped out with a passive
cryogenically cold resistor or an active "cold damping"
circuit.

References:
6 "1 4 J.C. Mullins, W.T. Ziegler, and B.S. Kirk, "The
thermodynamic properties of parahydrogen from 1 to 22 K,"
Georgia Institute of Technology, Atlanta, Georgia, Technical
Report No. 1, Project No. A-593, Contract No. CST-7339 with
National Bureau of Standards, Boulder, Colorado (1 November
1961).
6 . 1 5 G. Gabrielse, Univ. of Washington, personal communication.

6 .16W. Kells, G. Gabrielse, and K. Helmerson, "On achieving

cold antiprotons in a Penning trap," Fermilab Preprint No.
Conf-84/68-E, 8055.000, Fermi National Accelerator Lab,
Batavia, Illinois (August 1984).
6- 1 7 p.B. Schwinberg, R.S. Van Dyck, Jr., and H.G. Dehmelt,
"New comparison of the positron and electron g factors," Phys.
Ref. Lett. 47, 1679-1682 (1981).
R.L. Forward and G.D. Thurmond, "Network for simulating
low-noise-temperature resistors," U.S. Patent 4,176,331 (27
November 1979).

104
 SECTION 7

ANTIPROTON ANNIHILATION PROPULSION

"* This section discusses how we might go about building and

using a high-specific-impulse, high-thrust propulsion system
based on the generation, storage, and utilization of
antiprotons. It will be shown that antiproton propulsion is
technically feasible, but difficult and expensive. Yet,
despite the high cost of antimatter, it may be a cost-
effective fuel in space, where any fuel is expensive.

7.1 ANTIMATTER PROPULSION

It has long been realized that antimatter would be a valuable

propulsion energy source because it allows for the complete
conversion of mass to energy. Early studies of the concept by
SAnger 7 . 1 assumed that the antimatter would be antielectrons
(positrons), which interact with electrons to produce
0.511 MeV gamma rays. SAnger unsuccessfully tried to invent
electron-gas m:.rrors to direct these short wavelength gamma
rays to produce a photon rocket.
The antiproton is much more suitable than the antielectron for
propulsion systems. The annihilation of an antiproton by a
proton (or neutron) does not produce gamma rays immediately.
Instead, as we show in more detail in Section 7.3, the
products of the annihilation are from three to seven pions.
On the average there are 3.0 charged pions and 1.5 neutral
pions.
As is shown schematically in Figure 7-1, the neutral pions
have a lifetime of only 90 attoseconds and almost immediately
convert into two high energy (200 MeV) gamma rays. The
charged pions have a normal half-life of 26 nanoseconds, but
because they are moving at 94% the speed of light, their lives
are lengthened 'o 70 nanoseconds. Thus, they travel an
average of 21 meters before they decay. This time and
interaction length is easily long enough to collect the
charged pions in a thrust chamber constructed of magnetic
fields and to direct the isotropic microexplosion into
directed thrust. Even after the charged pions decay, they
decay into energetic charged muons, which have even longer
lifetimes and interaction lengths for further conversion into
thrust. Thus, if sufficient quantities of antiprotons could
be made and stored, then, according to presently known
Sphysical principles, they can be used as a highly efficient
propulsion fuel.

105
 Because of the extreme difficulty in obtaining significant
quantities of antimatter, the idea of an antimatter rocket has
usually remained in the "science fiction" category. Any
papers before 1980 [see 27 references in section 02.01 of
bibliography by Mallove, et al. 7 -2 ] were usually concerned
with interstellar missions and glossed over the problems of
generating, storing, and using the antimatter.

Recent progress in particle physics on methods for obtaining

intense antiproton beams, however, has caused those in the
space propulsion community to take another look at the concept
of antimatter propulsion to see if the concept can be removed
from the "science fiction" category to the "technically
difficult and very costly" category, at which point the
military services or NASA could begin considering its use.
The last five years have seen the presentation of a number of
papers on antimatter propulsion, 7 .3 7 *8 including a special
issue of the Journal of the British Iitgrplanetary Society on
the subject of antimatter propulsion. pS

9780 lAI

-21m 1.85km
. m ,g • e
70nP
e+•
COI LS

V e

77
P-
NOZZLE
:" / 1MAGNETIC

... \ .'•
-. . e,-.•- - . '• .j
o .• • - . . ' .. . - .. . '\
•. .V P. • .. -• . - -•...•.

proton rocket.
Fig. 7-1 Schematic of an idealized anti

106

-+
 References:
7 .1
S~nger, E., "The Theory of Photon Pockets," Ing. Arch. 21,
213 ff. (1953).
7 2Mallove, E.F., Forward, R.L., Paprotny, Z., and Lehmann,
J., "Interstellar Travel and Communication: A Bibliography,"
J. British Interplanetary Soc. 33, 201-248 (1980) [entire
issue].
7" 3 Forward, R.L., "Interstellar Flight Systems," AIAA Paper
80-0823, AIAA Int. Meeting, Baltimore, Md., (6-8 May 1980).
7"4Cassenti, B.N., "Antimatter Propulsion for OTV
Applications," AIAA Paper 84-1485, 20th Joint Prop. Conf.,
Cincinnati, Ohio (June 1984).
7 5Cassenti, B.N., "Optimization of Relativistic Antimatter
Rockets," J. British Interplanetary Soc. 37, 483-490 (1984).
7"6 Vulpetti, G., "A Propulsion-Oriented Synthesis of the
Antiproton-Nucleon Annihilation Experimental Results," J.
British Interplanetary Soc. 37, 124-134 (1984).
7-TVulpetti, G., "An Approach to the Modeling of Matter-
Antimatter Propulsion Systems," J. British Interplanetary Soc.
37, 403-409 (1984).
7 * 8 R.L. Forward, "Antiproton annihilation propulsion," AIAA
Paper 84-1482, 20th Joint Prop. Conf., Cincinnati, Ohio (11-13
June 1984) [To be published in J. Propulsion and Power].
7- 9 Special issue on antimatter propulsion, J. British
Interplanetary Soc. 35, 387-424 (September 1982).

7.2 EXTRACTION OF ANTIMATTER FROM STORAGE

There are a number of techniques for extracting the
antihydrogen from the storage trap and directing it into the
rocket engine under control. If the antihydrogen is in the
form of an electrostatically suspended ball many milligrams in
size, then as is shown in Figure 7-2, the antiprotons can be
extracted from the ice ball by irradiating the ice with
ultraviolet, driving off the positrons, extracting the excess
antiprotons by field emission with a high intensi ylslectric
field, then directing them to the thrust chamber.
It might be more desirable to form the antihydrogen as a cloud
of charged microcrystals, each crystal a microgram in mass and
containing the energy equivalent of 20 kg of chemical fuel.
Then, using a directed beam of ultraviolet light to drive off

107
6N
[,
 °

a few more positrons, an individual microcrystal could be

preferentially extracted from the microcrystal cloud using
electric fields, and directed down a vacuum line to the thrust
chamber. Since the position of the charged microcrystal in
the injection line can be sensed, mechanical shutters can
allow the passage of the microcrystal without breaking the
storage chamber vacuum.
References:
7 " 1 0 Morgan,
D.L., "Concepts for the Design of an Antimatter
- Annihilation Rocket," J. British Interplanetary Soc. 35, 405-
412 (1982).

ANTIPROTONS

DIRECTION OF
ROCKET
ACCELERATION

ELECTRODES

FIELD LINES-

MAGNETIC . •
FIELD LINES -• . .

POSITRONS

Fig. 7-2 Extraction of antiprotons from storage.

108

S"
 7,7.3 PARTICLE PRODUCTION FROK ANTIPROTON ANNIHILATION

When the antiprotons reach the rocket chamber, they will be

annihilated with the protons in normal hydrogen to release the
rest mass energy of both particles. There is a popular
misconception that the annihilation of antimatter with matter
produces prompt high-energy gamma rays. This is true for the
annihilation of positrons with electrons, which produces two
0.511 MeV gamma rays. It is not true for the annihilation of
baryons.

When an antiproton annihilates with a proton, the predominant

reaction products (98%) are pions. A recent survey of the
literature 7 -II found that on the average there are 3.0 charged
pions, 1.5 neutral pions, 0.05 charged kaons, 0.03 neutral
kaons, and 0.02 prompt gamma rays. The branching ratios are
given by Table 7-1:

Table 7-1 Branching Ratios of p3 annihilation products

Probability n* W- nO K+ K- KO

0.345 1 1 2
0.213 2 2 2
0.187 2 2 1
0.078 1 1 1
0.058 2 2
0.019 3 3
0.016 3 3 1
0.0133 1 1 1
0.0133 1 1 1
0.0103 1 1 1 1
0.0091 2 1 1
0.0076 3
0.0050 2
0.0047 3 1
0.0042 1 2 1
0.0036 2 2
0.0032 1 1
0.0030 3 3 2
0.0005 1 1 2 1
0.9938 1.49 1.49 1.52 0.026 0.026 0.028 0.018

This estimate of the branching ratio should be compared with

other estimates such as the ones by Agnew, e .f from
antiproton annihilations in mrgane in 1960," by Lawrence
Berkeley Laboratory in 1972, and by Vulpetti based on
recent work at CERN and Columbia. 7 "14

"109

* * ,***** - *--
 n+ n- le K+ K- K°
Agnew 1.53 1.53 1.60
LBL 1.5 1.5 2.0
Vulpetti 1.53 1.53 1.96 0.012 0.012 0.026

It is obvious that more work needs to be done to pin down the

exact values for the neutral pions and the kaons.
The neutral pions have a lifetime of only 90 attoseconds and
almost immediately convert into two high-energy gamma rays.
The charged pions have a normal lifetime of 26 nanoseconds,
but because the relativistic time contraction at their
.* velocity of 94% of the speed of light, their lifetimes are
lengthened to 70 nanoseconds. Thus, unless they are stopped
by interaction with matter, they will travel an average of
21 meters before they decay. The energy spectra of the
charged pions and gamma rays are shown in Figure 7-3. -1l The
average energy of the gamma rays is 200 MeV, while the average
energy of the charged pions is 250 MeV.
In the annihilation process of antiprotons or antihydrogen
with normal atoms or molecules, it should be appreciated that .

the relevant cross sections are not the nuclear cross

sections. In any matter-antimatter mixture in which the atoms
and antiatoms are not wholly ionized, rearrangement collisions
between the atom and the antiatom can lead to bound states of
the nucleus and antinucleus from which the annihilation can
proceed. Since the cross sections for such collisions at low
energies are considerably higher than the direct particle-
antiparticle annihilation cross section5,lgtomic interactions
play a dominant role in such a mixture.

In many applications of the use of antiprotons for energy

storage and propulsion, consideration is being given to
annihilation of the antiprotons with heavier nuclei than
protons. Since a neutron has the same baryon number as a
proton and a free neutron will spontaneously decay into a
proton, a neutron can be considered as an "excited state" of a
proton. Thus, antiprotons, will annihilate with a neutron as
well as a proton inside a heavy nucleus. Since the neutron
has a neutral charge and charge must be conserved in the
annihilation process, the reaction products from the
annihilation of an antiproton in a heavy nucleus will produce
different numbers of the various types of charged and
uncharged pion and kaon particles.

110
 IT. T T -1-7-7 .

E3, 200 MeV

E7r =250 MeV

7r-

* /

0 100 200 300 400 500 600

ENERGY, MoV

Fig. 7-3 Antiproton-proton annihilation particle spectra.

Annihilation inside a heavy nucleus has the potential for

increasing the efficiency of an antiproton annihilation
propulsion system, since the neutral pions are absorbed in the
nucleus instead of decaying into gamma rays. The annihilation
reaction will "heat up" the nucleus as well as cause
spallation fission of charged nuclear fragments. These
heavier charged particles would be yet another me&canism for
the transfer of the annihilation energy to a hydrogen working
fluid in a propulsion system.
Data for the reaction products of antiprotons with heavier
nuclei are sparse and of problematic accuracy, but ongoing
work with the LEAR antiprotons at CERN should soon produce
much better data. The available data to date include the 1972
experiments by 7 Algew, et al. -12 and 1982 calculations by
Clover, et'al.

111
 r+ f- 10 p n

Agnew (C12 ) 1.33 1.58 1.15

Clover C1 2 ) 1.1 1.4 1.6 1.0 1.1
(%23 ) 0.7 1.1 1.1 2.9 5.7

As can be seen, there are still disagreements in the numbers

of particles emitted from even the simple P-p reaction. Thus,
research is still needed in the interaction of antiprotons
with protons, neutrons, and the nuclei of heavier elements.
The measurement of the annihilation-induced fission products
from the annihilation of antiprotons in heavy nuclei over a
broad range of nuclear masses is especially needed. These
data will help to determine if there is an optimum nucleus
that is better for propulsion than annihilating antiprotons
with hydrogen.
An ideal reaction would be one where the energy from the
antiproton annihilation causes the nucleus to break up into
doubly charged alpha particles that would rapidly couple their
kinetic energy to the hydrogen working fluid. A more
realistic reaction would have over half of the annihilation
energy showing up as charged nuclear fragments, a minimum
number of neutrons and neutral pions escaping, and most of the
rest of the energy showing up as charged pions.

References:
7 "liThe antiproton group at Los Alamos National
Lab, private
communication (1984).
7"I 2 L.E. Agnew, Jr., T. Elioff, W.B. Fowler, R.L. Lander, W.M.
Powell, E. Serg6, H.M. Steiner, H.S. White, C. Wiegand, and T.
Ypsilantis, "Antiproton interactions in hydrogen and carbon
below 200 MeV," Phys. Rev., 118, 1371 (1960).
7 "1 3 Lawrence Berkeley Laboratory Report LBL-58 (1972).
7 * 1 4 G.
Vulpetti, "A propulsion-oriented synthesis of the
antiproton-nucleon annihilation experimental results," J.
British Interplanetary Soc. 37, 124-134 (1984)
7"I 5 D.L. Morgan, Jr. and V.W. Hughes, "Atom-antiatom
"interactions," Phys. Rev. A7, 1811-1825 (1973).
7"I 6 M.R. Clover, R.M. DeVries, N.J. DiGiacomo, and Y. Yariv,
"Low energy antiproton-nucleus interactions," Phys. Rev. C26,
2138-2151 (1982).

112
7.4 STOPPING OF PARTICLES
As we have seen in Section 7.3, the particles emitted from the
annihilation of antiprotons with protons and neutrons are
mainly pions. The neutral pions almost immediately decay into
two gamma rays, however, so the particles that we must deal
with are of two kinds: High-energy charged pions with an
average kinetic energy of 250 MeV, and high-energy gamma rays
with an average energy of 200 MeV. (See Figure 7-3.)
To utilize the energy in these particles for propulsion we
need to either direct their momentum rearward from our vehicle
to provide thrust, or we need to stop the particles and use
their energy to heat a working fluid. We also need to provide
shielding for the crew and the radiation sensitive components
of the rocket engine.

7.4.1 Stopping of Charged Pions

The primary stopping mechanism for charged particles in matter
is the electronic excitation and ionization of the atoms in
the material. The stopping power depends largely on the
atomic number (number of electrons per nucleus) of the
stopping material, and the velocity and charge of the
particle, but not on its mass.
For particles with large initial velocities the rate of energy
deposition per unit path length is approximately independent
of energy, then it rises rapidly as the energy approaches
zero. The resultant integrated energy decrease produces a
"range" for the particle in a particular stopping material
that is a function of the energy. Because of statistical
fluctuations, there will be a small spread of ranges for
different particles with identical initial energies.
A great deal of data is available on the stopping power and
range of protons, pions, and muons as a function of energy in
various materials.7. 1 7 , 7.18 Table 7-2 is an abbreviated list
of this data giving the range of pions in a number of
substances that might be of use in a first cut design of an
antimatter rocket.
The range is first given in g/cm2 of stopping material
required to extract 100 MeV of energy, since that number is
independent of the state of the material (gas, liquid, or
solid). When multiplied by the density of the material it
gives the range in centimeters of thickness of the material to
produce a 100 MeV energy loss. The range is relatively
constant down to 100 MeV, when it suddenly becomes much
shorter.

113
I[.•. ••>-...--;•
v . • >>•>.>.>.•
,- :> '•'•>•-L'','L':'-\-'.."•,'-:'AL•'•"
113•-",'• .' ----I'
'I.-- ,-2,*
*p•- -- - •"
 Table 7-2 Range of Pions per 100 NeV of Energy.

Range for 100 MeV energy loss

Density High Energy Last 100 MeV
Material g/cm g/cm2 cm g/cm2 cm
H (100 atm) 0.009 24 2700 13 1400
H2 (300 atm) 0.027 24 890 13 480
He (300 atm) 0.018 50 930 29 540
N2 (100 atm) 0.125 55 440 32 250
Water 1.00 50 50 27 27
Be 1.85 61 33 33 18
Al 2.70 62 23 35 13
Fe 7.87 67 8.5 38 4.8
W 19.3 86 4.5 52 2.7
Pb 11.4 88 7.7 53 4.6

From Table 7-2 we can see that gases such as hydrogen, helium,
and nitrogen, because of their low density, have a long
interaction length with pions. It will be necessary to
operate the reaction chamber at high pressures in order to get
the density up so that the interaction range becomes shorter
than the pion mean life range.
To calculate the fraction of the pion kinetic energy deposited
in the working fluid requires a detailed calculation involving
the pion energy spectrum, the density, temperature, and
pressure of the working fluid, the containment losses, and
other factors. As we can see from Table 7-2, nitrogen at
"100 atm pressure gives a shorter range than either hydrogen or
"helium at 300 atm pressure. (The Space Shuttle Main Engines
operate at 213 atm.) It may turn out that despite its higher
molecular weight, nitrogen may be a preferred reaction gas
because of its higher density.
Assuming that the containment losses are small and the
temperatures not too high, it is possible to estimate the
percentage of the pion energy that gets into the working
fluid. The efficiency was found to be about 65% for hydrogen
"at 300 atm and 95% for nitrogen at 100 atm. Obviously, much
more work needs to be done in this area since the lifetime and
*- the stopping power both change with pion energy.
Using the detailed tables available in the literature 7 "1 7 the
ranges of the annihilation pions with their spectrum of
possible initial energies were calculated and plotted in
Figure 7-4 for hydrogen at 300 atm, nitrogen at 100 atm, and
solid tungsten (for shielding). Also included are the pion
mean life (not half-life) and the resulting mean range in
vacuum as a function of energy. These last were calculated
from Equations A.1 and A.2 in Appendix A.

114

| -, - ---- ------
... . .--
n - - . . . . . - . . .- .• ,. . -... .-.. . .•.. . "- ..• - . , -'.- -•- - %J - .' - '- a - - . ' •'- - • 7 . -" . " • '-

Ei = 250 MeV

z
z

MENF-F

0 100 200 300 400 500 600 MeV

"KINETICENERGY
Ii. I I I I I I
26 40 60 80 100 120 140 ns
MEAN LIFE

I I
20 30 40 m
1~J VACUUM

005 10 20 30 40 50 m
z 300 atm H2 (0.027 g/cc)
I I II
0510 15 20 25 m
100 atm N2 (0.125 g/cc)
I I I I
0- 5 10 15 20 25 cm
TUNGSTEN (19.3 g/cc)

Fig. 7-4 Range of charged pions.

115
 777 , .
1.1. .70T -.
. *-

7.4.2 Stopping of Gamma Rays

The small numbers of prompt gamma rays produced by the
antiproton-proton annihilation process and the large numbers
of delayed gamma rays produced by the decay of the neutral
pions are a major problem in the utilization of antiprotons
for propulsion. As can be seen in Figure 7-3, the spectrum of
the gamma rays peaks at around 70 MeV and extends out to many
hundreds of MeV, with an average energy of about 200 MeV. The
maximum energy gamma rays would be the small fraction (0.5%)
*. of 938 MeV prompt gammas from the initial annihilation
reaction.
Unlike charged particles, which have a definite range in
matter, the gamma ray intensity is exponentially attenuated by
matter. For the high-energy gammas the attenuation is
predominantly due to pair production in the shield material.
From Figure 8e-ll in the AIP Handbook, 7 - 1 9 we find that for a
heavy metal, the attenuation coefficient for gamma rays with
an energy greater than 100 MeV is about 0.1 cm2 /g and slowly
rises with energy. For tungsten, with a density of
19.3 g/cm3 , this gives an attenuation factor of 1.93 cm-1 .
Table 7-3 gives the attenuation achieved for various
thicknesses of tungsten.

Table 7-3 Attenuation of Gama Rays by Tungsten.

x (cm) Attenuation
2 2.1x10-2
4 4.4xi0-4
6 9.4x10-6
8 2*0x10 7
10 4 2x10- 9
12 8.7xl0- 1 2 -

14 1.8x10-
16 3.9x10- 1 4
18 8.2x10- 11 67
20 1.7xl0-
22 3 6x10- 12 91
24 7.7x10-

This table can be used to obtain rough estimates of the amount

of tungsten needed for various requirements. For example, if
* we wish to recover a majority of the neutral pion energy by
. capturing the decay gamma rays in a tungsten shield cooled
with flowing hydrogen, then a thickness of only two
centimeters is needed to capture 98% of the gamma rays.

116
U
2-
7.4.3 Shielding of Vehicle and Crew

The gamma rays and the high-energy charged pions will cause
heating and radiation damage if they are not shielded against.
Fortunately, unlike neutrons, they will not cause the
shielding material to become radioactive by transmutation of
the nuclei in the shield. The components needing shielding
are the crew, the electronics, the cryogenic tankage, and the
magnetic coils for magnetically assisted rockets.
The radiation flux will be extremely high. A typical high
performance antimatter rocket will probably operate at a
thrust level of 10,000 N (pushing 10 metric tons at 1 m/s 2 )
and a specific impulse of 2000 s (exhaust velocity of
20 km/s). The power level of the exhaust coming from the
charged pions is then 200 MW, with 100 MW of 200 MeV gamma
rays coming from the neutral pions.
If we are concerned with shielding some superconducting coils
that are generating magnetic fields to contain and direct the
charged pions, then using Table 7-3 we find we will want to
use a shield of about 10 cm thickness to attenuate the 100 MW
of gamma ray energy down to a few watts that can be handled by
the cryogenic coolers.
If we are interested in protecting personnel and electronics,
then a shield thickness of 14 cm of tungsten will shield
personnel at 10 meters from even a 100 MW source of gamma
rays. To illustrate is last point, if we look at Table 8i-5
in the AIP Handbook".', we find that the dose from a 1 Curie
(3.7x10 1 0 disintegrations/s) source of 100 MeV gamma rays at
1 m distance is 29 R/hr. Extrapolating from this data point,
the dose from a 1 Curie source of 200 MeV gamma rays at 10 m
distance would then be 0.58 R/hr.
Since a single 200 MeV gamma ray has an energy of 3.2x10-1 1 J,
a 100 MW source of 200 MeV gamma rays produces 3x1018 gammas/s
or 8.5x10 7 Curies. This would produce a dose of 4.9x10" R/hr
at 10 m distance. From Table 7-3 we find that 14 cm of
tungsten shielding would provide an attenuation of 1.8x10-1 2 .
Thus, the dose rate at 10 m would be less than 0.1 mrad/hr, a
reasonable dose for space missions.
A conceptual schematic of the shielding that might be used in
a magnetic field assisted antiproton annihilation rocket is
shown in Figure 7-5. The reaction chamber would be about 1 m
in diameter. The pressure walls would have the thickness
equivalent of 2 cm of tungsten so as to absorb most of the
gamma ray energy and use it to heat hydrogen flowing through
channels in the wall. The hot hydrogen would be used as a
film flow to protect the nozzle from the ultrahot hydrogen
plasma at the center of the chamber.

117
 Each superconducting magnetic coil would be shielded by 10 cm
of tungsten in a ring about 1.1 m in diameter. The crew would
be protected by a shadow shield 14 cm thick and 0.6 m in
diameter that is 0.6 m from the annihilation region. This
would provide a shielded region 10 m in diameter at 10 m from
the engine. The mass of the shadow shield is 800 kg, while
each of the rings is 750 kg. At 2 cm thickness the pressure
chamber mass is 2,200 kg.

References:
7"I7w.H. Barkas and M.J. Berger, Tables of Energy Losses and
Ranges of Heavy Charged Particles, NASA SP-3013, STI Division,
NASA, Washington, DC (1964).
7" 1 8 Particle Data Group, "Review of particle properties," Rev.
Mod. Phys. 56, Part II (April 1984).
7 "1 9 D.E.
Gray, Ed., American Institute of Physics Handbook,
Third Edition, McGraw-Hill, NY (1972).

S"" •iC• ,2 cm ABSORBER

, ". WALL AND
NUCLEI
HEAVY -••• HEAT EXCHANGER
INJECTORS ,

H2 OTH22
ANTIPROTON " ANNIHILATION
INJECTOR - REGION
H2 .- -. ,.,

* PERSONNEL
S~SHIELDS

SUPERCONDUCTING COILS

Fig. 7-5 Conceptual schematic for radiation shielding.

118

|401A -
 •$•" 7.5 ANTIMATTER ROCKET ENGINE CONCEPTS

Antimatter fuel is so powerful that new types of rocket

engines will have to be developed to fully utilize its
potential. It should be emphasized, however, that the need
for completely new engine designs is not unique to antimatter
propulsion. Any new form of propulsion that prcduces high
thrust at high specific impulse using a thermally expanded
reaction fluid will have the same problem.
Laser plasma, solar plasma, fusion plasma, metallic hydrogen,
atomic hydrogen, metastable helium, tetrahydrogen, and any
other form of advanced propulsion that operates at high thrust
with specific impulses over 1000 s will have to cope with
chamber temperatures exceeding 3500 K. No known materials can
survive these temperatures for long. If hydrogen is being
used as the reaction mass, these temperatures will cause it to
convert from a molecular gas to an atomic plasma of free
protons and electrons.
Thus, any advanced thermal propulsion system will require the
development of new types of reaction chambers and nozzles that
can contain and direct a hot hydrogen plasma without direct
contact with material walls. The best known technique for
handling plasmas is the use of magnetic fields to guide and
* •contain the charged particles in the plasma.
One type of magnetic engine design uses a solenoidal "magnetic
bottle" where the magnetic field fills the entire chamber. In
this design, the plasma ions tend to stay on the magnetic
field lines and are kept away from the walls.

A less well known type of magnetically assisted engine uses a

"picket fence" geometry where the walls of the chamber have
ring magnets that alternate north and south in polarity. The
magnetic fields go from one pole to another and stay near the
walls, creating a "picket fence" of arching magnetic fields.
In these designs, the ions in the plasma are free to move
about in the chamber in any direction until they approach the
wall. Since they find it difficult to cross the magnetic
field lines near the wall, they are repelled back to the
center of the chamber.
Much more research needs to be done on engines that can
contain and control hot plasmas. This work-is essentially
independent of the source of the hot plasmas and can pay off
in the future even if antiproton annihilation propulsion is
not found to be practical.

119

I-

- '.-.
" •"t$;.-:*--
-'2V
 7.5.1 Thermal Heat Exchanger Concept

One of the simplest antiproton propulsion systems would use a

design similar to that of a nuclear thermal rocket. In a
nuclear thermal rocket, the nuclear fission reactions were
used to heat a high temperature material (graphite) which had
cooling channels passing through it. Graphite was chosen as
the core material because it best matched both the reactor
nucleonics requirements (low neutron absorption) and the heat
exchanger requirements (reasonable strength and high operating
temperature). Hydrogen was passed through the cooling
channels, keeping the reactor cool while the hydrogen was
heated to many thousands of degrees. Specific impulse values
close to 900 s were obtained by this technique.
In the antiproton annihilation version of the thermal heat
exchanger rocket concept, the energy released by the
annihilation reaction would be absorbed in the walls of a heat
exchanger made out of refractory metal. 7 - 2 0 The heat
exchanger would then heat hydrogen to produce thrust. As is
shown in Figure 7-6, a heat exchanger made out of a cylinder
of tungsten 28 cm in diameter and 28 cm long would only weigh
330 kg and would capture most of the energy in the gamma rays
and pions, thus utilizing all of the annihilation energy.

TUNGSTEN

ABSORBER
HYDROGEN AND HEAT
ANNIHILATION
CHAMBER EXCHANGER

Cr+C Isp 900 sec

C CC• 3000°K .
CC e,=99kmsc
H2..---

i ."2 8 cm - - -
[•. 330 kg
[. -99+% 7ENERGY
• ~93%x•- ENERGY

I2m

Fig. 7-6 First generation antimatter thermal rocket.

120

a-hO- ':
 The maximum operating temperature of this rocket would be
limited by the melting point of tungsten to about 3000 K,
resulting in a maximum specific impulse of about 900 sec or an
exhaust velocity of about 9 km/s. This specific impulse is
considerably better than any chemical rocket or even a nuclear
fission thermal rocket, but still does not utilize the high
exhaust velocity potential of antiproton annihilation.
This first generation antimatter rocket engine would be
optimum for many space missions in earth-lunar space and could
be designed and tested using reasonable extrapolations of
nuclear thermal rocket technology. The high risk engineering
development of magnetic nozzles to control, contain, and
direct the charged pions would be set aside for a second
generation engine where higher specific impulse would be
required.

References:
7.20 B.W. Augenstein, "Some examples of propulsion
applications using antimatter," Rand Paper P-7113, Rand Corp.,
Santa Monica, CA 90406 (July 1985).

7.5.2 Hot Hydrogen Gas Engine Concept

The next level of sophistication in an antimatter rocket is to
take the design for a hydrogen thermal rocket operating at
high pressure and add a magnetic bottle to it. In this
magnetically assisted engine, shown in Figure 7-7, hydrogen is
contained by a pressure vessel while the charged pions from
the annihilation reaction are contained by the magnetic
bottle. 7 - 2 1 The pressure of the hydrogen is 100 atm (compared
to the pressure in the Space Shuttle Main Engine of 213 atm).
In operation, hydrogen and antihydrogen are injected in the
chamber and the antihydrogen annihilates with some of the
hydrogen to produce pions. The neutral pions immediately
convert into gamma rays. The gamma rays are considered lost
in this design and no attempt is made to use their energy. In
all probability, the pressure vessel will be designed to
intercept as little of the gamma ray energy as possible to
minimize heating problems.
The charged pions are trapped by the magnetic field and spiral
back and forth in the magnetic bottle, transferring their
energy to the hydrogen. The range of the average charged pion
in hydrogen at 100 atm pressure is limited by its mean decay
time. This varies from 45 ns at a kinetic energy of 100 MeV
to 100 ns at 400 MeV.

121
U
 4- 7 4~7 -4- 7' .-- - ---

At the 100 atm pressure assumed for the design shown in

Figure 7-7, the pion gives up about 25% of its kinetic energy
* to the hydrogen before it decays into a muon and a neutrino.
"* The muon, with its much longer lifetime of 2.2 ?a, then gives
up essentially all of its kinetic energy to the hydrogen
before decaying into an electron and two neutrinos.
In a complicated Monte Carlo calculation, it was estimated
that including losses of particles to the wall of the chamber
and out the ends of the magnetic bottle, 35% of the
annihilation energy ended up in the hydrogen working fluid.
As was pointed out in Section 7.4, it might be advantageous to
use nitrogen as the working fluid. Its higher densL*y would
allow a higher percentage of the pion energy to be axtracted.
This might be enough to overcome its larger molecular weight.

MAGNETIC FIELD ASSISTED STANDARD REGENERATIVELY

CONTAINMENT VESSEL COOLED 100:1 ROCKET NOZZLE

HYDROGEN AND \ " 0 0- - .. .c 5

* ~~ANTIHYDROGEN '

, 1u ,a, ,-/ . THRUST- 9 Wk

50 kG /

* Fig. 7-7 Magnetic assisted antimatter heated hydrogen rocket.

122

-- ~~ ~
% . ~ . . ~ . ~
. . ~ .- ~ ~ - -7~.b .4 4
Not all of the charged pions stay in the magnetic bottle.
Those which start out with a velocity vector at a small angle
to a magnetic field line will follow the magnetic field line
right out of the ends of the bottle and be lost. Those with
too high an initial velocity will have a large radius of
curvature and may hit the walls and be lost. As the pions
move through the hydrogen, and especially when they decay into
muons, the velocity vector will be changed and the particle
may find itself on a trajectory which leaves the confinement
volume.
If the antimatter is injected at a rate of 2.9 pg/s and the
hydrogen reaction mass at a rate of 29 kg/s, then the specific
impulse will be 350 s (compared to the Space Shuttle Main
Engine specific impulse of 460 s). The thrust level is 98 kN
and the power in the exhaust is 330 MW (1/20 the Space Shuttle
Main Engine power). The temperature of the hot hydrogen for
this relatively low specific impulse is only 460 K or 190 C
because of the low molecular weight of the pure hydrogen
1
exhaust.7. 2
If the antimatter is injected at a rate of 12 $g/s and the
hydrogen reaction mass at a rate of 16 kg/s, then the specific
impulse will increase to 1000 s.7"22 The temperature of the
hot hydrogen for this higher specific impulse is now 3700 K,
which is reaching the material limits of the pressure chamber.
At higher temperatures the hydrogen will also start to
disassociate and become ionized. We will than have to move to
newer engine designs that can work with a hot hydrogen plasma.

These hot hydrogen antimatter energized engines will require

significant advances in engine design. The pressure vessels
have to operate in a strong magnetic field and a high charged
particle and gamma ray environment. The high field strength
superconducting magnetic coils must be designed for minimum
mass and be shielded against the high thermal and radiation
environment. Yet for specific impulses below 1000 s, the
temperature of the hydrogen reaction mass is low enough that
standard pressure vessel designs and materials can be used, so
that much of the present rocket technology is applicable.
Much work remains to be done, but a great deal of it has
already been done.

References:
7"2 1 B.N. Cassenti, "Antimatter Propulsion for OTV
Applications," AIAA Paper 84-1485, 20th Joint Propulsion
Conference, Cincinnati, Ohio, June 1984.
7 *2 2 B.N. Cassenti, personal communication (1983)

123
 7.5.3 Magnetically Contained Plasma Engine Concept

In the previous design concept, the magnetic field was only

used to contain the high-energy charged pions, muons, and
electrons while they transferred their energy to the hydrogen
working fluid. Since the magnetic bottle exists, it can also
be used for containment of the working fluid if it is a
plasma.

There has been no detailed design yet of a magnetically

contained plasma rocket, independent of the method of heating
* the plasma. There has been one preliminary design carried out
for an antimatter heated plasma rocket that uses heavy
elements as the .rking fluid. The conceptual design is shown
in Figure 7-8."'

MAGNETIC
FIELD LINE

PULSED PLASMA -DIRECTED BEAM OF

ANTIPROTON - - ---- HIGH ATOMIC
BEAM Z z =z- MASS NO. AND
~-*- OTHER SPECIES
ANNIHILATION,
IONIZATION, AND Z - THROAT
COLLISION
COUPLING REGION DIRECTOR
PULSED, NEUTRAL, HIGH ATOMIC
MASS NO. ATOM BEAM 1 METER

Fig. 7-8 High thrust magnetic containment antimatter rocket.

124

•"%'•,.j,".,.'l".-"
','...'.•.'.',"
."..'t,,j".
"%'.."
•- ,,--,,'-"j,,
%"-. • "•...................................................-.....-............ -- """:'-''" "" --" "- -. ----
 The magnetic field in this design is pulsed in the same manner
as a pulsed fusion machine magnetic bottle. The full pulse
cycle is 17 ms with the time of containment being 7 ms. The
magnetic field in Figure 7-8 is shown in its maximum strength,
or closed, configuration. At the start of each pulse, 5/&g of
antiprotons are injected into the reaction chamber along with
a stream of atoms of a heavy element such as lead or
(depleted) uranium.

The antiprotons will annihilate with either a proton or a

neutron in the nucleus of the heavy atom, producing pions.
The pions immediately transfer their energy to the nucleus
causing spallation, evaporation, or fission of the nucleus
leading to the production of several light nuclides and
individual particles, all of which have higher mass, lower
velocity, and often higher charge than the pions. All of
these characteristics lead to more rapid transfer of the
kinetic energy to the rest of the working "fluid".
The beam of heavy atoms continues after the antiproton
annihilation pulse until 60 g have been injected. The trapped
charged particles circle about in the magnetic bottle until
the excess matter becomes ionized due to collisions with the
nuclear fragments and forms a plasma which is confined by the
magnetic field.
After 7 ms the annihilation energy has been transferred to the
working fluid, the field is opened and the hot plasma escapes
out the nozzle, producing thrust. For an average input of
2x10 2 0 antiprotons/s (0.3 mg/s) and 9x10 2 4 U atoms/s
(3.6 kg/s), t'ib) 2 hrust of this engine design is 0.55 MN
(124,000 l)
The specific impulse of this conceptual design for a high
thrust plasma engine is a remarkable 14,000 s (140 kW/s). The
maximum temperature reached during a pulse is 2.6xi00 K, but
the average temperature is much less. The plasma confinement
characteristics of the magnetic bottle are comparable to those
of present day fusion plasma machines.

References:
"
7*23D.L. Morgan, "Concepts for the design of an antimatter
annihilation rocket," J. British Interplanetary Soc. 35, 405-
412 (1982).

'."

125

V 'Y
 7 7 -

7.5.4 Magnetically Directed Annihilation Pion Rocket

In all the previous designs, a great deal of effort was spent

developing mechanisms to ensure the transfer of the kinetic
energy of the annihilation pions into some working fluid to
lower the specific impulse and increase the thrust levels to
match the performance characteristics of the antimatter rocket
to space missions around earth and throughout the solar
system. For deep space missions and missions to the stars,
the higher specific impulse obtainable directly from the
charged pions becomes more useful. A conceptual design for
such a high exhaust velocity antiproton annihilation rocket is
"shown in Figure 7-9.7.23

DIMENSIONS IN METERS
TRANSFER ANNIHILATION,
SYSTEM ENGINE

WIDTH OF
ANTIPROTON_- 2 "--H --
• - _i0 .9
BEAM 0v•, - = 0.94c
f ANTI-...V-.....: •••EGION?
0.27 -V. S11 • "

SO0,00O0 MAGNET']*C -- "--

gauss sFIELD LINES
gauss ga

S...-COILS
]• 25,000
gauss

NEUTRAL H ATOMS

Fig. 7-9 High exhaust velocity antiproton rocket.

126

•, "" ""S "-- . " "" . • "' ". """""" " . . , '". ' / . " "-,, '. ".'"• - "• " ""• .,. . ',;
 "
7 This rocket design concept uses a static magnetic field
configuration in the shape of a conical rocket nozzle. The
magnetic field is produced by the turns of a coil that
increase in radius and separation so that the magnetic field
lines form straight lines, all of which emanate from a common
center on the axis. Within the field is space vacuum except
for the antiproton beam, the hydrogen beam, and the
annihilation products.
The beam of antiprotons enters from the left and collides at a
right angle with a beam of hydrogen coming from below. If the
two beams are jxl02 0 ions/s each, then 95% of the antiproton
ions are annihilated. 7 - 2 3 The ion current in each beam is
approximately 30 A.

The charged pions produced by the pp annihilation follow paths

that are along a cone whose vertex is the common center point
of the magnetic field lines and whose surface is defined by
the initial velocity vector of the pion. The vertex angle of
the cone depends upon the velocity, charge, and mass of the
pion and the strength of the magnetic field at the point of
tangency.
The dynamics of the motion of the pion in the magnetic field
confines the pion to the surface of the cone. If the pion
velocity vector is to the right, the pion will spiral out of
the engine to the right and produce thrust. If'the pion
velocity vector is to the left, it will spiral toward the
vertex of the cone, circle around just below the tip, then
reverse direction and spiral back out to the right and exit
the nozzle. Only the small fraction of pions with a velocity
vector nearly parallel to magnetic field line at its point of
origin will be able to travel up the throat and out of the
engine the wrong way.
The specific impulse of this engine is the velocity of the
pions at their time of formation. For the mean kinetic energy
of 250 MeV, this is a velocity of 94% the speed of light or a
specific impulse of 28,800,000 s! The energy from the 30 A of
antiproton ions will run this engine at the same power level
as the three Space Shuttle Main Engines, 24 GW. With the high
specific impulse, however, this 24 GW of power only produces
70 N of thrust. Such a design is probably best suited as the
last stage in an interstellar probe design.

7.5.5 Magnetic Nozzle Antimatter Nicroexplosion Rocket

It has been suggested that it might be possible to direct a
small amount of antimatter onto a pellet made of a heavy
element and cluI2 a microexplosion that would produce a hot,
dense plasma. * Magnetic rocket nozzles that take hot,
•- dense plasmas from microexplosions and convert them into

127
thrust have already been designed to convert the energy from a
plasma formed by a laser fusion pellet microexplosion into
directed thrust. 7 - 2 5 One version that uses a single large
superconducting magnetic coil is shown in Figure 7-10. If
antimatter microexplosions are ever proved feasible, a
magnetic rocket exists to use them.

References:
7.24S. Polikanov, "Could antiprotons be used to get a hot,
dense plasma?" pp. 851-853, Physics at LEAR with Low-Energy
Cooled Antiprotons, U. Gastaldi and R. Klapisch, ed., Plenum
Press, NY (1984).
7. 2 5 R. Hyde, L. Wood, and J. Nuckolls, "Prospects for rocket
propulsion with laser induced fusion microexplosions," AIAA
Paper 72-1063 (Dec 1972).

SUPERCONDUCTOR

Fig. 7-10 Magnetic nozzle for microexplosion plasmas.

128
 SECTION 8
ANTIMATTER MISSION STUDIES

The availability of antimatter as an energy source for space

propulsion will revolutionize thi subject of mission analysis.
Many of the present assumptions that are implicit in the
design of a mission will no longer be valid, and mission
designers will have to develop a new set of assumptions to
replace them.
For example, the concept of mass ratio and staging mass
fractions plays an important part in the present design of
missions. Once a mission has been defined, there is a certain
characteristic velocity 4V for that mission and once the fuel
is chosen, there is a fixed exhaust velocity v available from
that fuel. The total mission mass ratio R is then
automatically determined by the relation:

R mv+p eV/v
eZ = e V/gIsp
S8.1) R= my+p=AI
= eVgs
mv

where mv is the mass of the empty vehicle (including

payload) delivered to the destination and m, is the mass of
the propellant exhausted at velocity v or spicific impulse
Isp, and g=9.8 m/s 2 .

Thus, every different mission with a different characteristic

velocity requires a different mass ratio and a different
vehicle design. Also, if the mission characteristic velocity
starts to exceed five times the exhaust velocity, the mass
ratio starts to become greater than 100, and there is a
tendency to say that the mission is "impossible".
With antimatter powered rockets, the exhaust velocity can be
tailored to match the mission characteristic velocity, thus
minimizing the mass ratio and mission cost. As we shall see
in the next subsection, the mass ratio of an antimatter rocket
never exceeds 5:1, and mass ratios that minimize total mission
cost are typically 2.5:1 for any mission characteristic
velocity. Thus, the same vehicle can be used for all
missions, with the only difference being the amount of
antimatter used.

129
 Another implicit assumption used by mission planners is that a
high specific impulse automatically implies low thrust and
long mission times. This is because present solar and nuclear
electric systems are power limited by the heavy weight of
their power source and the power density limitations of their
thrusters. Long missions at high specific impulse may save
money in lower mass ratios, but they are expensive in terms of
ground support time and extra vehicle mass to insure crew
health and safety.
Antimatter rockets are not inherently limited in their thrust
levels. Assuming that new engines will be designed that can
"handle the antimatter, high thrust can be obtained at any
specific impulse. With antimatter rockets, mission
trajectories will no longer be modified Hohmann transfers, but
nearly straight lines. Manned missions to Mars will no longer
take years of time, but months of time, with significant
savings in initial vehicle mass and ground support costs.

8.1 MINIMUM ANTIMATTER OPTIMIZATION

In this subsection, we will outline a mathematical proof for

the optimization of an antimatter powered rocket mission in
which the amount of antimatter used is minimized. The proof
is simple but has only been documented in journals 8 -1 and
reports8.1 that are somewhat difficult to obtain, so we will
repeat it here.
"In an antimatter rocket, the source of propulsion energy is
separate from the propellant or reaction fluid. Thus the
*" total initial mass of the vehicle consists of the empty mass
of the vehicle my, the mass of the reaction fluid mr, and the
mass of the energy source m , half of which is the mass of the
antimatter ma that we wish to minimize. The mass ratio is
then:

8.2) R 8e/kV/v=
e mv + mr + me
my

The exhaust energy comes from the conversion of the energy

source rest mass to kinetic energy with an efficiency e.

1
8.3) C2
e (mec) =--- + me)V 2
(mr +
2

130

re
 -: . Combining Equations 8.2 and 8.3 and rearranging we obtain:
my v2 _VI k---e-1
V (e v/v i)=( -)
8.4) me =---------(e -)'
2e c 2 x2

2
where x=AV/v and k=mvAV 2 /2ec .
We now make the assumption that the antimatter costs dominate
the reaction fluid costs and we want to minimize the amount of
antimatter. By setting the derivative of Equation 8.4 with
respevt to x equal to zero and solving for x, it can be
shown. 1 ', 8.2 that the amount of antimatter is minimized when:
8.5) v = 0.63 AV
This means that the mass ratio is a constant.
8.6 R e V /v 1.5 9
8.6) R=e =e =4.9
Amazingly enough, this constant mass ratio is independent of
the efficiency of the energy conversion, the mission
characteristic velocity, and the molecular weight of the
reaction fluid used. This constant mass ratio for minimum
antimatter consumption holds for all conceivable missions in
the solar system and only starts to deviate for interstellar
missions where the mission velocity starts to approach the
speed of light. 8 "3
The amount of antimatter needed for a specific mission is
obtained by substituting Equation 8.5 into 8.4. It is found
to be a function of the square of the mission characteristic
velocity &V 2 (essentially the mission energy), the empty
mass of the vehicle my, and the conversion efficiency e:

8.7) ma -- e 0.39 m
2 e c2

For a typical antimatter mission, where the efficiency of

conversion of antimatter energy to thrust is 0.32, only 120 mg
of antimatter is needed to heat 39 metric tons of reaction
fluid to an exhaust velocity of 19 km/s, which will suffice to
accelerate a 10 ton vehicle to 30 km/s (0.0001 c). Thus, no
matter what the mission, the vehicle will always use 3.9 tons
"of reaction fluid for each ton of vehicle and an insignificant
amount (by weight, not cost) of antimatter.

131

h•"
 References:
8"1 L.R. Shepherd, "Interstellar flight," J. British
Interplanetary Soc. 11, 149-167 (1952).
8" 2 D.F. Dipprey, "Matter-Antimatter Annihilation as an Energy
Source in Propulsion," Appendix in "Frontiers in Propulsion
Research," JPL TM-33-722, D.D. Papailiou, Editor, Jet
Propulsion Lab, Pasadena, CA 91109 (15 March 1975).
8*3 B.N. Cassenti, "Optimization of relativistic antimatter
"rockets," J. British Interplanetary Soc., 37, 483-490 (1984).

8.2 MINIMUM FUEL COST OPTIMIZATION

If instead of minimizing just the antimatter cost, we

minimized the total fuel cost (the cost of the antimatter
energy source plus the cost of the reaction fluid), we would
find a new mass ratio optimum. This is because reaction fluid
is not low cost once it has been lifted into low earth orbit.

An analysis was carried out (see Appendix B) to determine the

optimum mass ratio for various missions. A typical result is
shown in Figure 8-1. In this figure, we see that depending
*q upon the relative cost of antimatter per milligram and
reaction fluid per ton, the optimum mass ratio for minimum
fuel cost will vary from 2:1 to 4:1, but will never be greater
than 5:1. This compares favorably with the mass ratios for
* even the best chemical fuel, liquid hydrogen/liquid oxygen.
For a H2 /02 chemical rocket the mass ratio rises exponentially
", with the mission characteristic velocity, exceeding 100:1 at
23 km/s and 1000:1 at 35 km/s.

Thus, mission analysts need to rethink those missions that

have been labeled "impossible" because of the extreme mass
ratios required using a chemical or nuclear system with a
fixed specific impulse. Also, with these low mass ratios for
antimatter propulsion, it may be beneficial to include other
costs in the optimization, such as ground support costs and
crew comfort and safety costs. Under such conditions, it may
turn out to be more cost effective to increase the antimatter
mass slightly and carry out the mission faster.

132
 , I I
1000

LO,'(/H2
IV) 500 7
EXPONENTIALLY
INCREASING
MASS RATIO

100

p-Z
107

• -- ANIRO
ANTIPROTON UPPER LIMIT TO MASS RATIO
V _ PROPULSION_
S~ 2400 s _

•-130
,os IS DOUBLE
REVERSE REVERSE
SORBIT ORBIT

0 10 20 30 40 so
MISSION CHARACTERISTIC VELOCITY, AV, km/s

Fig. 8-1 Chemical and antimatter rocket mass ratios.

133

U,
 8.3 IMPOSSIBLE MISSIONS
There are missions in the solar system that would be desirable
to accomplish for scientific purposes, but which are
essentially impossible using chemical or even nuclear thermal
rockets. One example is a solar impact mission, which
requires the rocket to cancel out the orbital velocity of the
earth so the vehicle can drop directly into the sun. This
requires a mission characteristic velocity of 35 km/s, which
is presently obtained by an out-of-the-way swingby of Jupiter,
5 AU and many years in the wrong direction. Another is a
* mission to the rings deep down in the gravity well of Saturn.
"This requires a mission characteristic velocity of 48 km/s.
There are even much simpler missions near earth that are
nearly impossible using chemical rockets. One is the simple
maneuver of rapidly reversing your orbital direction. This
"maneuver requires cancelling the initial orbital velocity and
building it up again in the opposite direction. Since earth
orbital velocity is 7.7 km/s, the total mission characteristic
velocity of the reverse orbit maneuver is 15.5 km/s. If it is
then desired to return to the initial orbit (to dock at an
orbiting space station base), the process must be repeated
with a total mission characteristic velocity of 31 km/s.
"The mass ratios required for each type of rocket system to
carry out each of these missions can be calculated from
Equation 8.1. They are listed in Table 8-1. As can be seen,
"- all of these mission require high mass ratios, with the more
difficult ones requiring such large mass ratios that it is
difficult, if not impossible, to imagine how one might build a
vehicle to accomplish those missions using chemical or nuclear
thermal rockets. All of those missions could be performed by
an antimatter rocket with a mass ratio of 5:1 or less.

Table 8-1 Mass Ratios for Difficult Missions

Total Mass Ratio

Storable HO Nuclear
Isp 300 s 5O0s 900 s
&V(km/s)
Reverse Orbit 15.5 175 22 6
Double Reverse Orbit 31 30,700 490 32

Solar Impact 35 117,000 1,100 49

Saturn Ring Rendezvous 48 8,900,000 15,000 200

134
 8.4 COMPARATIVE COST STUDIES

We will now use the rocket equations to determine the total

fuel cost for a number of different propulsion systems. For
both storable and cryogenic chemical propulsion systems the
mass of the energy source is in the .ropellant. Thus the fuel
cost for the chemical rocket is just the cost of the
propellant mass in orbit.
For the case of a nuclear thermal rocket, the energy to heat
the reaction fluid is in the nuclear reactor. A reactor has
to have a certain minimum charge of uranium just to operate
and therefore carries much more uranium than will be used in
any reasonable mission. Therefore, we have assumed that the
mass and cost of the uranium is charged to the empty vehicle
mass and cost. The fuel cost for the nuclear thermal rocket
will just be the cost of the reaction fluid in orbit.
For the case of an antimatter rocket, the cost of the
antimatter part of the energy source is not negligible. The
total fuel cost for the antimatter rocket will be the cost of
the antimatter mass used plus the cost of the reaction fluid
in orbit.

8.4.1 Fuel Cost of a Chemical or Nuclear Thermal Mission.

The fuel cost Cc of a mission using a chemical or nuclear

thermal rocket system is the price of propellant or reaction
fluid per kilogram in low earth orbit p times the amount of
propellant or reaction fluid needed for ?he mission.

8.8) Cc = pp mp
Pp mv (R

pp mv (e&V/v -1)

r Since chemical and nuclear thermal propulsion systems have a

fixed specific impulse or exhaust velocity, we see from
Equation 8.8 that the cost of any chemical or nuclear thermal
rocket system rises exponentially with increasing mission
characteristic velocity as soon as the mission velocity
exceeds the exhaust velocity.

&"/V>v e V/v.
• 5,. 8. F-) Cc 10 pp mv

135
L-

S.
 -M - - -- 0

8.4.2 Fuel Cost of an Antimatter Powered Mission

The fuel cost Ca of an antimatter powered mission consists

of the price of the antimatter Pa times its mass ma plus
the price of the reaction fluid Pr times its mass mr

8.10) Ca m Pr mr + Pa ma
In an antimatter rocket, the propulsion energy comes from the
annihilation of the antimatter with a small amount of normal
matter in the reaction fluid. The energy is twice the rest
mass energy of the antimatter. Some fraction e of the
annihilation energy is then converted into kinetic energy of
the reaction fluid.
1
8.11) 2e ma c 2 = --- mr v2
2
Substituting Equation 8.11 into Equation 8.10 we obtain:

v2
8.12) Ca = (Pr + Pa ------ )mr
4ec
v2 eAV/v
= (Pr + Pa 4e - 2 )(e - i)mv

Equation 8.12 can be used to find a cost minimum for an

antimatter rocket. As the required mission aV increases, the
cost of the mission tends to increase exponentially, since the
amount of reaction fluid needed is rising exponentially, just
as in a chemical rocket. In an antimatter rocket, however,
this exponential rise in reaction fluid can be curbed by using
more antimatter and increasing the reaction fluid exhaust
velocity. How much more antimatter we use depends upon the
relative price of antimatter and reaction fluid. For low
exhaust velocity the second term becomes large, while at high
exhaust velocity the first term becomes large. Thus, there is
a cost minimum for each mission AV depending upon the
relative price of reaction fluid and antimatter.
We were not able to find a simple analytic solution to the
minimization of Equiation 8.12 when the relative costs of the
antimatter and the reaction fluid were comparable. Instead, a
computer was used to calculate the total fuel cost, mass
ratio, and antimatter mass used, as a function of the mission
characteristic velocity over a range of exhaust velocities and
relative fuel costs. The minimum in the total fuel cost was
found on the computer printout and this determined the optimum
values for the other quantities.

136

p• E \ *• * *..
 An important parameter in these parametric studies is the
ratio of the price of antimatter to the price of reaction
fluid in orbit. To be completely general, we should have
plotted the following curves in terms of a dimensionless price
ratio. Since the price ratios vary from 1010 to 108, however,
they are so large they are almost meaningless at first glance.
Instead, we fixed the price of reaction fluid or propellant in
orbit at the present day price of 5k$/kg and presented the
parametric curves in terms of the price of antimatter per
milligram. Thus, a curve which is labeled 5M$/mg is
equivalent to a relative price ratio of:

0.13) 5M$/mg
-- - - -= 109
5k$/kg

8.4.3 Comparative Total Fuel Costs.

We next calculated the total fuel costs for a number of

different propulsion systems and compared the total fuel costs
as a function of mission characteristic velocity. The
propulsion systems considered were a storable chemical
propulsion system with a specific impulse of 300 s, a
cryogenic H2/O2 chemical propulsion system with a specific
impulse of 500 s, a nuclear thermal rocket system with a
specific impulse of 900 s, and three antimatter systems that
were optimized for lowest total fuel cost as a function of the
price of the antimatter.
Figure 8-2 shows the range of mission velocities that are
characteristic of present missions. The relative fuel cost
can be converted directly into millions of dollars per ton of
empty vehicle mass if chemical fuel for the chemical rockets
and reaction fluid for the nuclear and antimatter rockets is
assumed to cost 5k$/kg and the antimatter price is that
indicated for the various antimatter curves.
In examining Figure 8-2, we see that if the price of
antimatter can be brought down to 20M$/mg (or a relative cost
ratio of 4xi0 9 ), an antimatter propulsion system is always
more fuel cost effective than a storable chemical propulsion
system. It is also better than the best chemical propulsion
system presently available (H2 /0 2 ) for any mission
characteristic velocity greater Ehan 12 km/s. At a price of
10M$/mg (or a relative cost ratio of 2x10 9 ), antimatter
propulsion systems are better than any chemical propulsion
system at any mission velocity, but are not as cost effective
as nuclear thermal propulsion. If the price of antimatter
drops to 2M$/mg (relative price ratio of 4xi0 8 ), then
antimatter propulsion is more cost effective than any other
known propulsion system at any mission velocity.

137
P'
 100 I STORABLE
1s 300 sec H2/02• ^ I .•~
SP I =500 sec"

80/
I" / •13 =1300 sec
0 / MR= 3.2

.60 ,/ a 3.2 mg=

U. / NUCLEAR
"" Isp = 900

~40 I =10se 0 0-/

w -

20 0
I~ ~ =3.
i M
low mag .4m

0 5 10 15 20
MISSION CHARACTERISTIC VELOCITY (Av), km/s

Fig. 8-2 Relative fuel cost vs. mission velon~ity.

Also shown in Figure 8-2 are some typical operational
parameters for a mission requiring a total velocity change of
15 km/s. At an antimatter price of 20M$/mg, to reach 15 km/s,
each ton of empty vehicle requires 3.2 mg of antimatter
costing 64M$ to heat 2.2 metric tons of reaction fluid costing
111M$ to a specific impulse of 1300 s. Thus, a vehicle can be
"pushed to 15 km/s by an antimatter propulsion system for a
total fuel cost of 75M$ per ton of empty vehicle mass, while
Sto do the same job with a H2 /0 2 rocket would cost 100M$/T.

If the price of antimatter drops to 2M$/mg, then each ton of

"delivered vehicle mass would require the use of 4.4 mg of
"antimatter costing 8.8M$ to heat 0.8 tons of reaction fluid
costing 4M$ to a specific impulse of 2500 s. At this price
for antimatter, the total fuel cost to reach a velocity of
15 km/s would be only 12.8M$/T.

138
 "8.4.4 Antihydrogen Propulsion Enables "Impossible" Missions
In Figure 8-3 we broaden the range of the plot of relative
fuel cost versus mission characteristic velocity to a scale
that includes missions that are "impossible" using any
chemical or nuclear thermal system. In Figure 8-3, the
relative fuel cost scale can be converted into millions of
dollars by assuming that the vehicle to be delivered has an
empty mass of 5 metric tons and the cost of propellant or-
reaction fluid in space is 5k$/kg or 5M$/T.
In this figure it is easier to see the differences in the
shapes of the total fuel cost curves for the different types
of propulsion systems. For those propulsion systems with a
fixed specific impulse, the fuel cost rises exponentially with
increasing mission characteristic velocity. Since the
antimatter propulsion systems can vary the exhaust velocity to
match the mission, the total fuel cost for an antimatter
propulsion system only rises as the square of the mission
characteristic velocity.
Thus, no matter what the cost of antimatter turns out to be,
it will always be more cost effective than any propulsion
system with a fixed exhaust velocity at sufficiently high
minsion velocity. For example, even at 10M$/mg, an antimatter
propulsion system will cost less than a nuclear thermal rocket
if the mission characteristic velocity desired is greater than
30 km/s (just off the top of Figure 8-3). If the price of
antimatter can be brought down to 1M$/mg or less, then
antimatter propulsion can open up the entire solar system and
allow the performance of missions that are now impossible
* using any present propulsion system.
As a result of these comparative cost studies we find that if
research on antiproton propulsion were successful in reducing
the cost of antimatter to 10M$/mg or below, then antiproton
annihilation would become a cost effective propulsion
technique, in addition to reducing the size of the using
vehicle. Thus, the primary objective of any antiproton
propulsion research program should be to find cost effective,
energy-efficient methods for making, controlling, and storing
antimatter.
Once the research studies on the production and handling of
antimatter have been completed, some 10 to 15 years from now,
a reasonably firm estimate can be made of the cost of
antimatter per milligram. If the cost estimate for antimatter
is low enough that it is seen as a cost effective fuel for
space propulsion, then the Air Force can, at that time, commit
the major amounts of money needed for the design and
fabrication of a special facility to produce milligram
quantities of antimatter for propulsion.

139
 HIGH THRUST, RAPID RESPONSE (NO ELECTRIC PROPULSION)
ANY CHEMICAL FUEL COSTS 5 M$/T IN LEO
AV/Vo I I 1 MSm (AVN*/) 2

400 - Iox I/1 10

Isp= 5/s

STORABLEI / MS/mg
//
300- 3
j / NUCLEAR
oS I
" /I I/5 os
900
LW

200 ,0I /
uj I,/
/ / M$/mg

DOUBLE SATURN
1/ REVERSE SOLAR RING
REVERSE ORBIT IMPACT ORBIT
0 ORBIT 1_1---
0 10 20 30 40 50

MISSION CHARACTERISTIC VELOCITY, AV, km/s

Fig. 8-3 Relative fuel c ,. expanded mission velocity.

140

|-.,*,'*- 'S -- - . .- - *
 :7- 773-77 7:7 -IV1

SECTION 9

CONCLUSIONS AND RECOMMENDATIONS

9.1 CONCLUSIONS

The fundamental conclusion of this study is that antiproton

annihilation propulsion is feasible but expensive.
Because of the probable high expense, antiproton propulsion is
probably not cost competitive for propulsion on the earth's
surface or from earth to low earth orbit. It may, however, be
cost competitive for propulsion in space where any fuel is
expensive.
If antiproton annihilation propulsion could be made available
at a reasonable cost, it would revolutionize space travel and
would enable missions that are now deemed "too difficult" or
"too expensive" or "impossible" with present day propulsion
technology. Many of these difficult missions, such as a
simple sortie to inspect a spacecraft moving in the opposite
direction and returning to base, are exactly those missions
the Air Force may be called upon to carry out if there is a
future requirement for a manned military presence in space.
The present techniques for generating, capturing, cooling,
trapping, and using antiprotons are exceedingly inefficient.
"Thus, the major thrust of antiproton propulsion research in
the near future should be on studies to determine the reasons
for the present low efficiencies, to develop methods for
improving those efficiencies, and to demonstrate
experimentally that the predicted efficiencies can be reached
at reasonable production rates and at reasonable cost.

9.2 RECOMMENDED RESEARCH INVESTIGATIONS

In this section we list some recommended investigations that

involve basic and applied research on the properties of
antimatter and the interaction of antimatter with fields and
particles. Many of these studies can be carried out using
normal matter instead of antimatter.

141
9.2.1 Trapping of Antiproton Ions

Antiprotons at low energy are available at the Low Energy

Antiproton Ring (LEAR) at CERN. It is recommended that
research be supported on techniques to slow the LEAR
antiprotons down and capture them in a Penning trap. This
research will result in the determination of the gravitational
and inertial mass, charge, spin, and lifetime of the
antiproton, and will provide a determination of the vacuum
level possible in a cryogenically cooled vacuum chamber.

9.2.2 Antiproton Production Spectrum

The absolute number of antiprotons produced from the
interaction of high energy protons with heavy metal targets is
probably not known to better than a factor of two. It is
recommended that experimental research be supported to measure
the antiproton production spectrum in energy and angle as a
function of the proton energy and the target parameters. The
proton energy range should cover from 26 GeV (the CERN energy)
to greater than 200 GeV. The only machine capable of carrying
out the experiments at the higher energies is the 400 GeV Main
Ring at Fermilab. If theoretical studies indicate that there
may be some benefit, the measurements should be repeated using
polarized protons and targets. This research will produce the
basic data needed to determine the maximum expected production
efficiency (and therefore minimum expected cost) of the
antiprotons. It will also aid in the engineering design of
the magnetic lenses and collector rings to capture the
antiprotons.

9.2.3 Annihilation Cross Sections at Low Energy

The annihilation cross section of an antiproton with protons,
hydrogen, and heavier atoms depends more on the Coulomb
interaction than on the nuclear interaction. It is
recommended that theoretical and experimental studies be
supported on the annihilation cross sections of antiprotons,
antihydrogen atoms, and antihydrogen molecules with various
nuclei at very low energies. This research will produce the
basic data needed to determine the "injection" and "ignition"
parameters for an antiproton annihilation rocket engine.

9.2.4 Annihilation Product Production Spectrum

When an antiproton annihilates with a proton, the reaction

products are typically 3 to 5 pions with an average kinetic
energy of 250 MeV, but the exact production spectrum is not
well known. The production spectrum for the annihilation of

142

L..-
 -~~777 .; n~ - ~r 777 A

*2-'. an antiproton with heavier nuclei is even less known. It is

recommended that theoretical and experimental studies be
supported to determine the number and momentum spectrum of the
products from the annihilation of antiprotons and antihydrogen
with various nuclei. After the initial studies have been
completed to determine the number, energy, direction, and
lifetime of the various reaction products for unpolarized
antiparticles and particles, the research effort should be
expanded to study the effect of polarized antimatter and
polarized targets on the annihilation cross section and
production spectrum. This research will produce the basic
data needed to determine the "combustion" and "mixing"
parameters for an antiproton annihilation rocket engine.

9.2.5 Formation of Antihydrogen

Antihydrogen atoms can be formed by the spontaneous

recombination of antiprotons with positrons, and antihydrogen
molecules can be formed by the spontaneous recombination of
two antihydrogen atoms. It is recommended that theoretical
and experimental research be supported on these recombination
processes to determine the detailed mechanisms, rates, and
final states. This research could then be expanded to study
methods to enhance the rates or control the final states using
* photons, fields, or other antimatter particles. Most of the
experimental work could be carried out using normal matter.
This research will produce the basic data needed to determine
the optimum method for conversion of antiprotons into ground
state antiparahydrogen with minimum loss and maximum
efficiency.

9.2.6 Slowing, Cooling, and Trapping of Antihydrogen

Lasers have been used to slow, cool, and stop a beam of sodium
atoms, and plans are underway to use lasers to trap a
collection of sodium atoms. It is ztcommended that
theoretical and experimental research be carried out on the
slowing, cooling, and trapping of bot: atomic and molecular
antihydrogen. The probable method will involve the use of a
tunable coherent source of ultraviolet photons, but other
techniques for slowing, cooling, and trapping using fields or
cold positrons should also be considered. In most cases, the
experimental studies can be carried out using normal matter.
The slowing, cooling, and trapping of molecular hydrogen is
recommended as a research project to be carried out at AFRPL.
It is discussed in further detail in Section 9.4. This
research will produce the basic data needed to determine the
optimum method for controlling and storing antihydrogen
without touching it with normal matter.

143

w••-. . . " '• '• . . '.• -. . ," " ".. " V .' " "' . . , . . .. . , . / ,
 "9.2.7 Crystal Nucleation from Antihydrogen Vapor
The techniques for slowing, cooling, and trapping of
antihydrogen should produce a relatively dense gas of
*i antiparahydrogen at a few millidegrees Kelvin. It would be
* desirable from a storage point to condense the gas into a
solid. It is recommended that theoretical and experimental
research be carried out on the molecular interaction
mechanisms that will allow the nucleation of an antihydrogen
-* crystal or amorphous solid from the vapor. The research
should study the role of charged ions of antihydrogen in
acting as nucleation sites and look at mechanisms for removing
the latent heat of fusion energy from the resultant solid.
The research should also identify the experimental parameters
that control the crystal size and the final temperature of the
solid. This research will produce the basic data needed to
design the "fuel tanks" for the antiproton annihilation
propulsion system.

9.2.8 Magnetic Properties of Antihydrogen

The magnetic susceptibility of gaseous and solid antihydrogen

is predicted by theory to be negative with a strength two-
thirds that of graphite. It is recommended that experimental
measurements be carried out to determine the magnetic
susceptibility of parahydrogen as a function of the
temperature, density, physical state, and orthohydrogen
impurity content. This research will produce the basic data
needed to determine the feasibility of passive magnetic
levitation for the antihydrogen "fuel tanks" in the antiproton
annihilation propulsion system.

*1 9.2.9 Levitation of Antihydrogen

Macroscopic particles have been levitated by electric fields,

magnetic fields, and photon fields. All three levitation
techniques should also be usable on antihydrogen. It is
recommended that experimental studies be carried out of the
levitation of normal matter parahydrogen ice by active
feedback electrostatic fields and passive magnetic fields.
The experiments should determine the optimum levitation
parameters as a function of the field strength, the trap
parameters, and the size, shape, surface charge, and
composition of the parahydrogen. These studies will produce
the engineering data needed to design the antihydrogen "fuel
tanks" for an antiproton annihilation propulsion system.

144
 ~~~~~
.'-¶
';- w: __7' 'ýv_!P. I- JljjiI -
D~

9.3 RECOMENDED ENGINEERING STUDIES

In this section we list some design and system studies that

will generate a first cut at the operational parameters of an
antimatter propulsion system. Antimatter is a synthetic fuel
and present estimates of the production efficiency indicate
that it is going to be an expensive synthetic fuel. Since the
practical feasibility of the use of antimatter for propulsion
depends strongly on the cost of the fuel, it is important,
early in the program, to get estimates of the upper bound on
the efficiency of production and utilization of antiprotons
for propulsion. We will need studies of the "refining plant"
that will manufacture the fuel, the rocket engines that will
"burn" it to produce thrust, the new vehicles with their
relatively small matter and antimatter "fuel tanks" that will
use the engines, and the possible and "impossible" missions
that can be carried out using these new vehicles.

9.3.1 Antiproton Factory Design

CERN, Fermilab, and IHEP have shown by a number of different

techniques how to make, capture, cool, and store low intensity
beams of antiprotons. It is recommended that an engineering
study of the design of an antiproton "factory" be supported.
This would involve the preliminary design of highly efficient
high current proton accelerators, optimized target designs,
new multilens arrays of magnetic collector lenses, stacks of
collector rings, efficient stochastic and electron cooling
rings, coherent decelerators that slow the antiprotons down
and extract energy in the process, and closed-cycle beam dumps
that extract energy from the particles that are not
antiprotons. The objectives of the study would be to maximize
the antiproton production rate and maximize the overall energy
efficiency. The initial goal would be the design of a
prototype factory that could produce a few milligrams of
antiprotons a year for initial propulsion studies. Later
designs would take into dccount physical limitations such as
space charge or target destruction to design a factory
"module" that is optimized in size to attain maximum
efficiency. Higher production rates would then be attained by
adding modules to the production facility.

9.3.2 Antiproton Annihilation Engine Design

Antimatter fuel is so powerful that new types of rocket
engines will have to be developed to fully utilize its
potential. This requirement for new rocket engine designs is
not unique to antimatter propulsion. Any new form of
propulsion that gives high thrust at high specific impulse
(Isp of 1000 s to 5000 s) will require the development of new

145
types of reaction chambers and nozzles that can contain and
direct a hot plasma without direct contact with material
walls. The best known technique for handling plasmas involves
the use of magnetic fields to guide and contain the charged
particles in the plasma. It is recommended that design
studies be supported on rocket engines that use magnetic
fields to assist in the containment of the plasma working
fluid. Both the magnetic "bottle" and "picket fence"
geometries should be studied along with the standard
regenerative cooling and film cooling concepts. Studies are
needed on the extraction of the antimatter from the "fuel
tank" and the "injection" of the jet of antimatter fuel at low
enough relative kinetic energy into the jet of ignition nuclei
so that rapid "ignition" take- place. Then the "mixing" of
the annihilation products with the working fluid needs study
to determine the optimum ignition nuclei and working fluid,
the optimum pressure, temperature, and density inside the
reaction chamber, optimum nozzle parameters, and optimum
magnetic field configuration. The effects on the engine
structure of the gamma rays and those charged pions that
escape the ch&mber also need to be included. The results of
this study should give relationships between the various
operational parameters of the rocket such as the size and
weight of the engine, the specific impulse, the thrust to
weight, and the efficiency of conversion of annihilation
energy into thrust.

9.3.3 Antiproton Annihilation Vehicle Design

Theoretically, the mass ratio of an antimatter rocket never
exceeds 5:1 and a vehicle for most solar system missions would
have a mass ratio of 2.5:1. It is recommended that
preliminary studies be supported on the design of a vehicle
that uses an antiproton annihilation rocket for propulsion.
The design effort would study the interaction of the various
parameters such as mass ratio, specific impulse, mission
characteristic velocity, thrust to weight, choice of reaction
fluid, tankage fraction, shielding, crew safety, and
antimatter storage on each other. The goal would be the
preliminary design of an antiproton annihilation powered
vehicle that could carry out most of the missions that it
would be desirable to do with an antimatter rocket.

9.3.4 Antimatter Powered Mission Design Studies

In Section 8 it was estimated that if the cost of antimatter
could be lowered to 10M$/mg, then antiproton annihilation
propulsion could be a cost-competitive method of propulsion
for the more difficult space missions of interest to the Air
Force and NASA. It is recommended that these studies be
expanded to include more realistic mission scenarios and

146
"orbital parameters, a broader range of competing propulsion
systems, more realistic cost and weight estimates for vehicles
using different types of propulsion methods, more realistic
price estimates for reaction mass and chemical propellants as
delivered to LEO, and additional cost parameters such as
ground support costs and crew safety and health costs. The
objective of the studies would be to determine when antimatter
propulsion becomes more cost effective than competing forms of
propulsion as a function of the price of the antimatter and
the type of mission.

9.4 RECOMMENDED AFRPL IN-HOUSE RESEARCH PROJECT

Of the research topics discussed in Section 9.2, it is

recommended that the near-term in-house research project to be
carried out by personnel at the Air Force Rocket Propulsion
Laboratory be the experimental and theoretical investigation
of the slowing and cooling of neutral molecular hydrogen with
multiple photon interactions. The reasons for this
recommendation are as follows:
The project is basic in nature since it investigates the
interaction of intense photon fields with an elementary form
of matter. The intense fields open up new areas of physics
such as "dressed states," where the particles and photons must
be considered as a single system. The multiple vibrational
and rotational levels of each electronic level, plus the
alternate missing rotational levels due to the homonuclear
character of the hydrogen molecule make an investigation of
the physics of the transitions richer than similar
investigations of atoms. The basic nature of the research
makes it suitable for support by the Air Force Office of
Scientific Research.
The Principal Investigator can draw upon the extensive
existing base of knowledge at the National Bureau of Standards
and elsewhere on the slowing and cooling of sodium atoms.
These scientists can supply information on theory and
experimental techniques. The NBS also sponsors annual
workshops and conferences where the Principal Investigator can
maintain contact with the professional community.
Although the proposed research project is similar to ongoing
projects elsewhere in that it uses intense beams of photons to
slow and cool neutral particles, it is significantly different
than the existing projects in that it concentrates on the
slowing and cooling of molecular hydrogen. Slow molecules of
hydrogen are of little interest to the atomic hydrogen clock
and neutral atomic hydrogen particle beam weapon community, so
it is doubtful that the research will be carried out by anyone
else before the completion of the research effort at AFRPL.

147
 - - - - - -7 .jA W XL. PEI x ] V

9.5 REQUIRED FACILITIES

To carry out the recommended research program will require the
collection and installation of a number of pieces of
equipment. Although the specific items will change as new
equipment becomes available and as the Principal Investigator
defines the details of the exact experiments to be carried
out, it is possible to list the equipment types needed for a
representative experiment.

9.5.1 Ultra-High Vacuum System

This must be a custom system designed to accommodate the
molecular hydrogen beam source at one end and the molecular
hydrogen detector at the other end. There should be wide
orthogonal VUV ports in the "cooling" region for transverse
cooling and provision for inserting VUV radiation as near to
the longitudinal axis of the beam as possible for longitudinal
slowing and cooling. The VUV ports must be compatible with
the tunable VUV photon source. Depending upon the cooling
wavelength chosen, the ports may have a transmission window
(with its accompanying losses) or the UHV system and the VUV
system may have to be pumped down together. In addition there
should be additional windows and ports for the insertion of
diagnostic photon and electron beams.

9.5.2 Molecular Hydrogen Beam Source

The molecular hydrogen beam source should be a differentially
pumped system that will use strong magnetic or electric
gradient lenses to separate out the orthohydrogen molecules
from the parahydrogen beam. The source should be able to
produce a wide range of flow rates, with good control over
beam divergence and longitudinal velocity.

9.5.3 Tunable Vacuum Ultraviolet Photon Source

Initial experiments could be done by using a single-frequency
VUV laser, such as a molecular hydrogen laser. Ultimately,
however, it will be necessary to buy or construct a high-power
tunable photon source in the vacuum ultraviolet. Such a
source is constructed by phuton mixing of lower-frequency
photons (at least one of which is tunable) in a nonlinear
medium. Watts of power will be needed for a significant
demonstration, so the primary laser sources should have
kilowatts of peak power with long pulse lengths and high
repetition rates. The photon source, especially the VUV
output, should be designed to be compatible with the ports in
the UHV system.

148
 "9.5.4 Molecular Hydrogen Diagnostics
The molecular hydrogen detector must be designed to give the
velocity profile of the emerging slowed and cooled beam.
Since the simulation experiments using normal hydrogen are
only to prove that slowing and cooling have taken place, the
beam can be disturbed during the diagnostic process. First,
any ionized molecules must be diagnosed and removed from the
beam by electric fields. Second, any excited molecules must
be diagnosed and removed from the beam by diagnostic lasers
tuned to interact preferentially with the excited state.
Third, any molecules that have been converted to orthohydrogen
must be removed by an electric or magnetic field and detected.
Finally, the velocity profile of the neutral parahydrogen beam
should be obtained. This may involve measurements cr raman
scattered photons, or excitation of an intermediate ý.ate by a
frequency swept laser beam that picks molecules wit i
specific velocity, followed by another laser photon that
produces an ionized molecule that is swept out by electric
fields and counted. To carry out these diagnostic functions,
it will be necessary to have a number of high quality laser
sources and detectors for photons, electrons, and hydrogen.

9.6 AREAS OF CONCERN

During the study, some specific technological areas were

uncovered where there may be a "show stopper". These are
listed below. In assigning priorities, it would be important
to make sure that these questionable areas are studied first.
For if a major problem is uncovered in one of these areas that
cannot be worked around, it may be concluded that antiproton
annihilation propulsion is not feasible or is too costly, and
it would be best if the research resources were used to study
other topics.
The formation of antihydrogen ice crystals from a cold
molecular hydrogen gas involves the production of heat from
the latent heat of fusion. To carry this heat away requires a
third body. When freezing hydrogen, this third body is
usually the wall of the experimental chamber. It may be found
impossible to force supercooled antihydrogen gas to nucleate.

Antihydrogen ice must be kept below 2 K to keep its vapor

pressure low enough so that the antihydrogen molecules
sublimating from its surface do not heat up the storage
chamber walls. It may be found that radiation cooling to the
cold chamber walls is not adequate for extracting heat from
the ice generated by unavoidable heat leaks, and no other
cooling technique works.

141
S
 '- %

The present techniques for generating antiprotons in a target

with a high-energy proton beam and capturing them with
magnetic lenses and collecting rings are extremely
inefficient. It may be found impossible to engineer the
design of efficient proton accelerators, magnetic lenses, and
accumulator rings to raise the overall energy efficiency from
10-9 to the desired 10-4. At 10-4 efficiency, antiprotons
cost 10M$/mg and antiproton annihilation propulsion is barely
cost effective. At 10-5 efficiency, antiprotons would cost
100M$/mg and there would be only limited use for antiproton
propulsion systems.
Antihydrogen annihilation propulsion only gives maximum
benefits to a military space program when the rocket engines
are operated at high thrust with specific impulses from 1000
to 3500 s. At these specific impulses the working fluid is a
hot ionized gas. It may be found to be impossible to design
and construct a reaction chamber and rocket nozzle that can
contain the pressures required to extract the energy from the
annihilation products and at the same time survive the heat
and radiation.

-.
].50
 SECTION 10

AWTIPROTON ANNIHILATION PROPULSION BIBLIOGRAPHY

This section contains a partial listing of the bibliographic

references that were collected and read during this study as
well as pertinant bibliographic references collected from the
list of references in those papers. For the convenience of
the reader, the bibliography is grouped into three major
topics: 10.1 Production of Antiprotons and Antihydrogen, 10.2
Slowing, Cooling, and Trapping of Ions, Atoms, and Molecules,
and 10.3 Antimatter Annihilation and Propulsion.

10.1 PRODUCTION OF ANTIPROTONS AND ANTIHYDROGEN

V.V. Abramov, et al., "Production of hadrons with transverse

momentum 0.5-2.5 GeV/c in 70-GeV proton-nucleus collisions,"
Soy. J. Nuclear Phys. 31, 343-346 (1980).
A.I. Ageyev, et al., "The IHEP accelerating and storage
complex (UNK) status report," pp. 60-70, Proc. llth Int. Conf.
High Energy Accelerators, Geneva (1980).
H. Aihara and TPC Collaboration, "Charged hadron production in
e+e- annihilation at 29 GeV," LBL-17142 preprint, Lawrence
Berkeley Lab, Berkeley, California 94720 (December 1983)
Yu.M. Ado, B.A. Myae, A.A. Naumov, M.F. Ovchinnikov, O.N.
Radin, V.A. Teplyskov, V.G. Tishin, E.F. Troyanov, "Initial
operation of the IHEP proton synchrotron with a new ring
injector," Paper H50, 1985 Particle Accelerator Conf.,
Vancouver, B.C., Canada (13-16 May 1985) [To be published in
IEEE Trans. Nuclear Sci. (Oct 1985)].
M. Antinucci, A. Bertin, P. Capiluppi, M. D'Agostino-Bruno,
A.M. Rossi, G. Vannini, G. Giacomelli, and A. Bussiere,
"Multiplicities of charged particles up to ISR Energies,"
Lett. Nuovo Cimento 6, 121-127 (1973).

R. Armenteros and B. French, "Antinucleon-nucleon

interactions," in High Energy Physics - Vol. V, E. Burhop,
ed., Academic Press, New York (1969), pp. 237-410.

151
E. Asseo, M. Boutheon, R. Cappi, G. Carron, M. Chanel, D.
Dumollard, R. Garoby, R. Giannini, W. Hardt, et al. "Low
energy antiprotons at the CERN PS," pp 20-23 (in Russian),
Proc. 12th Int. Conf. High-Energy Accel., F.T. Cole and D.
Donaldson, ed., Fermi National Accelerator Lab, Batavia,
Illinois (1983).
B. Autin, "The future of the antiproton accumulator," pp. 573-
582, Proc. Workshop on Proton-Antiproton Physics and the W
discovery, La Plagne (1983).
V.E. Balakin and A.N. Skrinsky, "Project VLEPP," Akademiya
Nauk USSR, Vestnik, No. 3, 66-77 (1983).
M.Q. Barton, et al., "Minimizing energy consumption of
accelerators and storage ring facilities [panel discussion],"
pp. 898-908, Proc. llth Int. Conf. High Energy Accelerators,
Geneva (1980).
B.F. Bayanov and G.I. Sil'vestrov, "Use of lithium to produce
a strong cylindrical magnetic lens," Zh. Tekh. Fiz. 48, 160-
168 (1978) [English translation Soy. Phys. Tech. Phys. 23, 94-
98 (1978)].
B.F. Bayanov, A.D. Chernyakin, V.N. Karasyuc, G.I.
Sil'vestrov, T.A. Vsevolozhskaya, V.G. Volohov, G.S.
Willewald, "The antiproton target station on the basis of
lithium lenses," pp. 362-368, Proc. llth Int. Conf. High
Energy Accelerators, Geneva (1980).
B.F. Bayanov, J.N. Petrov, G.I. Sil'vestrov, J.A. MacLachlan,
and G.L. Nicholls, "A lithium lens for axially symmetric
focusing of high energy particle beams," Nuclear Instr. &
Methods, 190, 9-14 (1981).

B.R. Bayanov, T.A. Vsevolozhskaya, Yu. N. Petrov, and G.I.

Sil'vestrov, "The investigation and design development of
lithium lenses with large operating lithium volume," pp. 587-
590, Proc. 12th Int. Conf. High-Energy Accel., F.T. Cole and
D. Donaldson, ed., Fermi National Accelerator Lab, Batavia,
Illinois (1983).
M. Bell, J. Chaney, H. Herr, F. Krienen, P. Moller-Petersen,
and G. Petrucci, "Electron cooling in ICE at CERN," Nuclear
Instr. & Methods 190, 237-255 (1981).
M. Bell and J.S. Bell, "Capture of cooling electrons by cool
protons," Particle Accelerators, 12, 49-52 (1982).
R. Billinge and M.C. Crowley-Milling, "The CERN proton-
antiproton colliding beam facilities," IEEE Trans. on Nuclear
Sci., NS-26, 2974-2977 (1979).

152

SQ
R. Billinge and E. Jones, "The CERN antiproton source," pp 14-
16, Proc. 12th Int. Conf. High-Energy Accel., P.T. Cole and D.
Donaldson, ed., Fermi National Accelerator Lab, Batavia,
Illinois (1983).
R. Billinge, "CERN's pp source," CERN Publication 84-09, Proc.
Fourth Topical Workshop on Proton-Antiproton Collider Physics,
Berne, 5-8 March 1984, pp. 357-364 (8 August 1984).
U. Bizzarri, M. Conte, C. Ronsivalle, R. Scrimaglio, L.
Tecchio, and A. Viganti, "High-energy electron cooling at LEAR
pp-collider," pp. 619-628, Physics at LEAR with Low-Energy
Cooled Antiprotons, Workshop on Physics at LEAR with Low-
Energy Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May
1982, U. Gastaldi and R. Klapisch, ed., Plenum Press, NY
(1984).
G. Brianti, "Experience with the CERN pp complex," IEEE Trans.
Nuclear Sci. NS-30, 1950-1956 (August 1983).

P.J. Bryant, "Antiprotons in the ISR," IEEE Trans. Nuclear

"Sci. NS-30, 2047-2049 (August 1983)
G.I. Budker and A.N. Skrinsky, "Electron cooling and new
possibilities in elementary particle physics," Usp. Fiz. Nauk
124, 561-595 (1978) [English translation Soy. Phys. Usp. 21,
277-296 (1978)].
D.C. Carey, et al., "Unified description of single-particle
production in pp collisions," Phys. Rev. Lett. 33, 330-333
(1974).
G. Carron, R. Johnson, S. van der Meer, C. Taylor, and L.
Thorndahl, "Recent experience with antiproton cooling," IEEE
Trans. Nuclear Sci. NS-30, 2587-2589 (1983).
CERN Courier editors, "When antimatter mattered," CERN
Courier, 23, 6-7 (January/February 1983).
CERN Courier editors, "LEAR arrives," CERN Courier, 23, 314
(October 1983).
CERN Courier editors, "From AA to Z," CERN Courier, 23, 365-
369 (November 1983).
CERIN Courier editors, "Going for antiprotons," CERN Courier,
23, 380-383 (November 1983).
CERN Courier editors, "First results from LEAR," CERN Courier,
23, 416-417 (December 1983).

N-
!-N
'-'•'.
'• .- .-.-• '.' '•t't''.''.'i'
.''.' .- .- .'L".'-.". " -'• \'- • "•" '."". ". " -"" "" "N. "- "" -¶"
" '- ''. ." "
 CERN Courier editors, "Antiprotons in orbit," CERN Courier,
24, 53-55 (March 1984).
CERN Proton Synchrotron Staff, "The CERN PS complex as an
antiproton source," IEEE Trans. Nuclear Sci. NS-30, 2039-2041
(1983).
G. Chapline, "Antimatter Breeders?" J. British Interplanetary
Soc., 35, 423-424 (1982).
B.V. Chirikov, et al., "Optimization of antiproton fluxes from
targets using hadron cascade calculations," Nuclear Instr. &
Meth. 144, 129-139 (1977).
D.B. Cline, et al., "Initial operation of the Fermilab
antiproton cooling ring," IEEE Trans. Nuclear Sci. NS-26,
3158-3160 (1979).
D.B. Cline, "The development of bright antiproton sources and
high energy density targeting," pp. 345-361, Proc. llth Int.
Conf. High Energy Accelerators, Geneva (1980).
D.B. Cline, C. Rubbia, and S. van der Meer, "The search for
intermediate vector bosons," Scientific American 247, No. 3,
48-59 (March 1982).
B. de Raad, "The SPS p-pbar collider, present performance and
future prospects," CERN Publication 84-09, Proc. Fourth
Topical Workshop on Proton-Antiproton Collider Physics, Berne,
5-8 March 1984, pp. 344-356 (8 August 1984).
* T.W. Eaton, S. Hancock, C.D. Johnson, E. Jones, S. Maury, S.
Milner, J.C. Schnuriger, and T.R. Sherwood, "Conducting
targets for f production of ACOL, past experience and
prospects," Paper X40, 1985 Particle Accelerator Conf.,
Vancouver, B.C., Canada (13-16 May 1985) [To be published in
IEEE Trans. Nuclear Sci. (Oct 1985)].
T. Ellison, W. Kells, V. Kerner, F. Mills, R. Peters, T.
Rathbun, D. Young, P.M. McIntyre, "Electron cooling and
accumulation of 200-MeV protons at Fermilab," IEEE Trans.
Nuclear Sci. NS-30, 2636-2638 (1983).
L.R. Evans, "Intrabeam scattering in the SPS proton antiproton
collider," pp. 229-231, Proc. 12th Int. Conf. High-Energy
Accel., F.T. Cole and D. Donaldson, ed., Fermi National
Accelerator Lab, Batavia, Illinois (1983).

Fermilab staff, Design Report: Tevatron 1 Project, p. 4-13 to

p. 4-16, Fermi National Accelerator Lab, Batavia, Illinois
(September 1984).

154

S_
S•
. " • • • ",• . - - -" *1M • • h • , - • • . - ° . " - - " ." • " • . ' - ÷ " . • • * * "* " • ' " .• . ~ -
 D.C. Flander, C.D. Johnson, S. Maury, T.R. Sherwood, G. Dugan,
"C. Hojvat, and A. Lennox, "Beam tests of a 2 cm diameter
lithium lens," Paper X39, 1985 Particle Accelerator Conf.,
Vancouver, B.C., Canada (13-16 May 1985) [To appear in IEEE
Trans. Nuclear Sci. (Oct 1985)].
R. Forster, T. Hardek, D.E. Johnson, W. Kells, V. Kerner, H.
Lai, A.J. Lennox, F. Mills, Y. Miyahara, L. Oleksiuk, T.
Rhoades, D. Young, and P.M McIntyre, "Electron cooling
experiments at Fermilab," IEEE Trans. Nuclear Sci. NS-28,
2386-2388 (1981).
J. Gareyte, "The CERN proton-antiproton complex," pp. 79-90,
Proc. llth Int. Conf. High Energy Accelerators, Geneva (1980).
U. Gastaldi and R. Klapish, Eds., Physics at LEAR with Low-
Energy Cooled Antiprotons, Workshop on Physics at LEAR with
Low-Energy Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May
"1982, Plenum Press, NY (1984).
M. Harrison, "The Fermilab pp collider," pp. 368-378, CERN
Pub. 84-09, Fourth Topical Workshop on Proton-Antiproton
Collider Physics, Berne, 5-8 March 1984, (8 August 1984).
H. Herr and C. Rubbia, "High energy cooling of protons and
antiprotons for the SPS collider," pp. 825-829, Proc. 11th
Int. Conf. on High Energy Accelerators, Geneva, Switzerland
(1980).
H. Herr, "A small deceleration ring for extra low energy
antiprotons (ELENA)," pp. 633-642, Physics at LEAR with Low-
"Energy Cooled Antiprotons, Workshop on Physics at LEAR with
Low-Energy Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May
1982, U. Gastaldi and R. Klapisch, ed., Plenum Press, NY
(1984).

H. Herr, D. Mohnl, and A. Winnacker, "Production of and

experimentation with antihydrogen at LEAR," pp. 659-676,
Physics at LEAR with Low-Energy Cooled Antiprotons, Workshop
on Physics at LEAR with Low-Energy Cooled Antiprotons, Erice,
Sicily, Italy, 9-16 May 1982, U. Gastaldi and R. Klapisch,
ed., Plenum Press, NY (1984).
C. Hojvat and A. Van Ginneken, "Calculation of antiproton
yields for the Fermilab antiproton source," Nuclear Instr. &
Methods 206, 67-83 (1983).
C. Hojvat, G. Biallas, R. Hanson, J. Heim, and F. Lange, "The
Fermilab Tevatron I project target station for antiproton
production," Fermilab TM-1174, 1983 Particle Accelerator
Conf., Santa Fe, NM (21-24 March 1983).

155

I, "' - '---:.°-•." -J ••v •?" • F•.: - '••• • • :.• ."". . .• -I. .'<,•; -• .••
 H. Hora, "Estimates for the efficient production of
antihydrogen by lasers of very high intensities," Opto-
Electronics 5, 491-501 (1973).
H. Hora, "Theory of relativistic self-focusing of laser
radiation in plasmas," J. Opt. Soc. Am. 65, 882-886 (1975).
L. Hdtten, H. Poth and A. Wolf, "The electron cooling device
for LEAR," pp. 605-618, Physics at LEAR with Low-Energy Cooled
Antiprotons, Workshop on Physics at LEAR with Low-Energy
Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May 1982, U.
Gastaldi and R. Klapisch, ed., Plenum Press, NY (1984).
C.D. Johnson, "Antiproton yield optimization in the CERN
antiproton accumulator," IEEE Trans. Nuclear Sci. NS-30, 2821-
2826 (August 1983).
R.P. Johnson and J. Marriner, "Stochastic stacking without
filters," Fermilab P Note 226, Fermi National Accelerator Lab,
Batavia, Illinois (17 August 1982).
E. Jones, S. van der Meer, R. Rohner, J.C. Schnuriger, and
T.R. Sherwood, "Antiproton production and collection for the
CERN antiproton accumulator," IEEE Trans. Nuclear Sci. NS-30,
2778-2780 (August 1983).
N.G. Jordan and P.V. Livdahl, "Costs to build Fermilab in 1984 -
.• dollars," Fermilab Technical Memo TM-1242, 0002.000 (1984).
W. Kells, P. McIntyre, L. Oleksiuk, N. Dikansky, I. Meshkov,
V. Parkhomchuk, and W. Herrmannsfeldt, "Studies of the
electron beam for the Fermilab electron cooling experiment,"
pp. 814-818, Proc. llth Int. Conf. on High Energy
Accelerators, Geneva, Switzerland (1980).
W. Kells, "Advanced stochastic cooling mechanisms," IEEE
Trans. Nuclear Sci. NS-28, 2459-2461 (1981).
W. Kells, F. Krienen, F. Mills, L. Oleksiuk, J. Peoples, and
P.M. McIntyre, "Electron cooling for the Fermilab P source,"
IEEE Trans. Nuclear Sci. NS-28, 2583-2584 (1981).
W. Kells, G. Gabrielse, and K. Helmerson, "On achieving cold
antiprotons in a Penning trap," FERMILAB-Conf-84/68-E, Fermi
National Accelerator Lab, Batavia, Illinois (August 1984).
[Preprint submitted to the IX Int. Conf. on Atomic Physics,
Seattle, Washington (23-27 July 1984).]
K. Kilian, "Physics with antiprotons at LEAR," pp. 324-341,
CERN Publication 84-09, Proc. Fourth Topical Workshop on
Proton-Antiproton Collider Physics, Berne, 5-8 March 1984, (8
August 1984).

156

'---7---". 7
 T.B.W. Kirk, "Antiprotcn production target studies - numerical
calculations," Fermilab TM-1011 (14 Nov 1980).
E. Knapp, "Accelerator Costs and Efficiency," Proc.
Information Meeting on Accelerator Breeding, Brookhaven, NY,
p. 294, (Jan 1977).
H.J.C. Kouts, Chairman, Proceedings of an Information Meeting
on Accelerator Breeeding, Brookhaven National Laboratory,
Upton, New York, 18-19 Jan 1977.
F. Krienen, "Initial cooling experiments (ICE) at CERN,"
pp. 781ff, Proc. llth Int. Conf. High Energy Accelerators,
Geneva (1980).
F. Krienen and J.A. MacLachlan, "Antiproton collection from a
production target," IEEE Trans. Nuclear Sci. NS-28, 2711-2716
(1981).
G. Lambertson, et al., "Experiments on stochastic cooling of
200 MeV protons," IEEE Trans. Nuclear Sci. NS-28, 2471-2473
(1981).
P. LefWre, D. M6hl, G. Plass, "The CERN low energy antiproton
ring (LEAR) project," pp. 819-823, Proc. llth Int. Conf. High
Energy Accelerators, Geneva (1980).
J.A. MacLachlan, "Current carrying targets and multitarget
arrays for high luminosity secondary beams," FN-334, 8055.000,
Fermi National Accelerator Lab, Batavia, Illinois (April 1982).

E. Malamud, "The Fermilab Pp Collider: Machine and Detectors,"

Fermilab 13 Note #421, Fifth Topical Conf. on Proton-Antiproton
Collisions, St. Vincent, Italy (26 Feb 1985).
J.P. Marriner, "The Fermilab Pp collider," pp. 583-592, Proc.
Workshop on Proton-Antiproton Physics and the W discovery," La
Plagne, (1983).
H.-J. M6hring and J. Ranft, "Antiproton production from
extended targets using a weighted Monte Carlo hadron cascade
model," Nuclear Instr. & Methods 201, 323-327 (1982).
R. Neumann, H. Poth, A. Winnacker, and A. Wolf, "Laser-
enhanced electron-ion capture and antihydrogen formation," Z.
Phys. A, 313, 253-262 (1983).

J. Peoples, "The Fermilab antiproton source," IEEE Trans.

Nuclear Sci. NS-30, 1970-1975 (1983).

157

S
 E. Peschardt and M. Studer, "Stochastic cooling in the CERN
ISR during pp colliding beam physics," IEEE Trans. Nuclear
Sci. NS-30, 2584-2586 (August 1983).

Physics Today editors, "CERN builds proton-antiproton ring;

Fermilab plans one," Search and Discovery Section, Physics
Today, 32, No. 3, 17-19 (March 1979).
Physics Today editors, "CERN SPS now running as 540-GeV Pp
collider," Search and Discovery Section, Physics Today, 35,
No. 2, 17-20 (February 1982).
Physics Today editors, "pp collisions yield intermediate boson
* at 80 GeV, as predicted," Search and Discovery Section,
Physics Today, 35, No. 4, 17-20 (April 1982).
Physics Today editors, "Fermilab's superconducting synchrotron
strives for 1 TeV," Search and Discovery Section, Physics
Today, 37, No. 3, 17-20 (March 1984).
A.L. Robinson, "CERN sets intermediate vector boson hunt,"
Science 213, 191-194 (10 July 1981).
R.E. Shafer, "The Fermilab antiproton debuncher betatron
cooling system," pp. 581-583, Proc. 12th Int. Conf. High-
Energy Accel., F.T. Cole and D. Donaldson, ed., Fermi National
Accelerator Lab, Batavia, Illinois (1983).
P. Sievers, R. Bellone, A. Ijspeert, P. Zanasco, "Development
of lithium lenses at CERN," Paper X41, 1985 Particle
Accelerator Conf., Vancouver, B.C., Canada (13-16 May 1985)
[To be published in IEEE Trans. Nuclear Sci. (Oct 1985)].
- A.H. Sorensen, "Theory of electron cooling in a magnetic
field," pp. 599-604, Physics at LEAR with Low-Energy Cooled
Antiprotons, Workshop on Physics at LEAR with Low-Energy
Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May 1982, U.
Gastaldi and R. Klapisch, ed., Plenum Press, NY (1984).
F.E. Taylor, et al., "Analysis of radial scaling in single-
particle inclusive reactions," Phys. Rev. D14, 1217-1242
(1976).
S. van der Meer, "Stochastic cooling in the CERN antiproton
accumulator," IEEE Trans. Nuclear Sci. NS-28, 1194-1998
(1981).
S. van der Meer, "Practical and foreseeable limitations in
usable luminosity for the collider," CERN Publication 83-04,
Proc. Third Topical Workshop on Proton-Antiproton Collider
Physics, Rome, 12-14 January 1983, pp. 555-561 (10 May 1983).

158
 G.S. Villeval'd, V.N. Karasyuk, and G.I. Sil'vestrov,
"•'Maqnetic field limited parabolic lens," Soy. Phys. Tech.
Phys. 23, 332-336 (1978).

T.A. Vsevolozhskaya and G.I. Sil'vestrov, "Optical properties

of fast parabolic lenses," Zh. Tekh. Fiz. 43, 61-70 (1973)
[English translation Sov. Phys. Tech. Phys. 18, 38-43 (1975)].
T.A. Vsevolozhskaya, M.A. Lyubimova, and G.I. Sil'vestrov,
"Optical properties of cylindrical lenses," Zh. Tekh. Fiz. 45,
2494-2507 (1975) (English translation Soy. Phys. Tech. Phys.
20, 1556-1563 (1976)].
T. Vsevolozhskaya, B. Grishanov, Ya. Derbenev, N. Dikansky, I.
Meshkov, V. Parkhomchuk, D. Pesrikov, G. Sil'vestrov, A.
Skrinsky, "Antiprotor source for the accelerator-storage
complex, UNK-IHEP," Fermilab Report FN-353 8000.00 (June
1981), a translation of INP Preprint 80-182 (December 1980).
T.A. Vsevolozhskaya, "The optimization and efficiency of
antiproton production within a fixed acceptance," Nuclear Instr.
& Methods 190, 479-486 (1981).
E.J.N. Wilson, "The CERN antiproton accumulator," pp. 567-571,
Proc. Workshop on Proton-Antiproton Physics and the W
discovery," La Plagne (1983).
D.E. Young, "Progress on beam cooling at Fermilab," pp. 800
"ff, Proc. llth Int. Conf. High Energy Accelerators, Gene,,a
(1980).

D.E. Young, "The Fermilab proton-antiproton collider," IEEE

Trans. Nuclear Sci. NS-28, 2008-2012 (1981)
D. Young, F. Mills, and G. Michelassi, "Information relevant
to the optimization of electric power versus equipment costs,"
Fermilab n-Note 189 (1984?)

159

U.

!a *'
 10.2 SLOWING, COOLING, AND TRAPPING OF IONS, ATOMS, AND MOLECULES

S.V. Adreev, V.I. Balykin, V.S. Letokhov, and V.G. Minogin,

"Radiative slowing and reduction of the energy spread of a
beam of sodium atoms to 1.5 K in an oppositely directed laser
beam," Pis'ma Zh. Eksp. Teor. Fiz. 34, 463-467 (1981) [English
translation Soy. Phys. JETP Lett. 34, 442-445 (1981)].
A. Ashkin, "Acceleration and trapping of particles by
radiation pressure," Phys. Rev. Lett. 24, 156-159 (1970).
A. Ashkin, "Atomic-beam deflection by resonance-radiation
pressure," Phys. Rev. Lett. 25, 1321-1324 (1970).
A. Ashkin, "Trapping of atoms by resonance radiation
pressure," Phys. Rev. Lett. 40, 729-732 (1978).
A. Ashkin and J.M. Dziedzic, "Optical levitation in high
vacuum," Appl. Phys. Lett., 28, 333-335 (1976).
A. Ashkin and J.M. Dziedzic, "Observation of radiation-
pressure trapping of particles by alternating light beams,"
Phys. Rev. Lett. 54, 1245-1248 (1985).
A. Ashkin and J.P. Gordon, "Cooling and trapping of atoms by
resonance radiation preo3ure," Optics Lett. 4, 161-163 (1979).
"A. Ashkin and J.P. Gordon, "Stability of radiation-pressure
particle traps: an optical Earnshaw theorem," Optics Lett. 8,
511-513 (1983).
V.I. Balykin, V.S. Letokhov, and V.I. Mishin, "Observation of
* the cooling of free sodium atoms in a resonance laser field
with a scanning frequency," Pis'ma Zh. Exp. Teo. Fiz. 29, 614-
618 (1979) [English translation JETP Lett. 29, 561-564
(1979)].
V.I. Balykin, V.S. Letokhov, and VI. Mishin, "Cooling of
sodium atoms by resonant laser emission," Zh. Exp. Teo. Fiz.
"78, 1376-1385 (1980) [English translation Sov. Phys. JETP 51,
692-696 (1980)].
V.I. Balykin, "Cyclic interaction of Na atoms with circularly
"polarized laser radiation," Optics Comm. 33, 31-36 (1980).
A. Bernard, J.P. Canny, T. Juillerat, and P. Touboul,
"Electrostatic suspension of samples in microgravity," Proc.
35th Congress of International Astronautical Federation,
Lausanne, Switzerland (7-13 October 1984).

160

I.'. )..'-'-.
•--' •)'.'*:.--..-"f<.. •-•• 7 --..
.-\-.
-. '.."-..."... .-.
.....-...-..-'."-.-......"
- -. --
 A.F. Bernhardt and B.W. Shore, "Cohel- nt atomic deflection by
resonant standing waves," Phys. Rev. 23A, 1290-1301 (1981).
N. Beverini, L. Bracci, V. Lagomarsino, G. Manuzio, R. Parodi,
and G. Torelli, "A Penning trap to store antiprotons," pp.
771-778, Physics at LEAR with Low-Energy Cooled Antiprotons,
Workshop on Physics at LEAR with Low-Energy Cooled
Antiprotons, Erice, Sicily, Italy, 9-16 May 1982, U. Gastaldi
and R. Klapisch, ed., Plenum Press, NY (1984).
J.H. Billen, K.R. Crandall, T.P. Wangler, M. Weiss, "An RFQ as
a particle decelerator," Los Alamos National Lab Preprint LA-
UR-85-140, Third LEAR Workshop, Tignes-Savoie, France (19-26
Jan 1985).
J.E. Bjorkholm, R.R. Freeman, A. Ashkin, and D.B. Pearson,
"Observation of focusing of neutral atoms by the dipole forces
of resonance-radiation pressure," Phys. Rev. Lett. 41, 1361-
1364 (1978).

J.E. Bjorkholm, R.R. Freeman, and D.B. Pearson, "Efficient

transverse deflection of neutral atomic beams using
spontaneous resonance-radiation pressure," Phys. Rev. 23A,
491-497 (1981).
R. Blatt, W. Ertmer, and J.L. Hall, "Cooling of an atomic beam
-i with frequency-sweep techniques," pp. 142-153, Laser-Cooled
and Trapped Atoms, NBS SP-653, W.D. Phillips, ed. (1983).
J.J. Bollinger, J.D. Prestage, W.N. Itano, and D.J. Wineland,
"Laser-cooled atomic frequency standard," Phys. Rev. Lett. 54,
1000-1003 (1985).
T. Breeden and H. Metcalf, "Stark acceleration of Rydberg
atoms in inhomogeneous electric fields," Phys. Rev. Lett. 47,
1726-1729 (1981).
L.N. Breusova, I.N. Knyazev, V.G. Movshev, and T.B. Fogel'son,
"Vacuum ultraviolet H2 laser with a sealed gas-discharge
cell," Kvant. Elektron. 6, 2458-2460 (1979) [English
translation Soy. J. Quantum Elec. 9, 1452-1453 (1979)].
L. Campbell, W.R. Gibbs, T. Goldman, D.B. Holtkamp, M.V.
Hynes, N.S.P. King, M.M. Nieto, A. Picklesimer, and T.P.
Wangler, "Basic research in atomic, nuclear and particle
physics," LA-UR-84-3572, Los Alamos National Lab, Los Alamos,
New Mexico (1984).

R.W. Cline, et al., "Magnetic confinement of spin-polarized

atomic hydrogen," Phys. Rev. Lett., 45, 2117-2120 (1980).

161

S'
 R.J. Cook and R.K. Hill, "An electromagnetic mirror for
neutral atoms," Optics Comm. 43, 258-260 (1982).
J. Dalibard, S. Reynaud, and C. Cohen-Tannoudji, "Proposals of
stable optical traps for neutral atoms," Optics Comm. 47, 395-
399 (1983).

R.W. Dreyfus, "Molecular hydrogen laser: 1098-1613 Angstrom,"

Phys. Rev. A9, 2635 (1974).
P. Ekstron and D. Wineland, "The isolated electron," Sci. Am.
243, 105-121 (August 1980).
W. Ertmer, R. Blatt, J.L. Hall, and M. Zhu, "Laser
manipulation of atomic beam velocities: Demonstration of
stopped atoms and velocity reversal," Phys. Rev. Lett. 54,
996-999 (1985).
R. Evrard and G.-A. Boutry, "An absolute micromanometer using
diamagnetic levitation," J. Vacuum Sci. and Tech. 6, 279
(1969).
Fermilab, Design Report: Tevatron 1 Project, pp. 4-13 to 1-16,
Fermi National Accelerator Lab, Batavia, Illinois (September
1984).
H. Friedman and A.D. Wilson, "Isotope separation by radiation
pressure of coherent pi pulses," Appl. Phys. Lett. 28, 270
(1976).

G. Gabrielse, "Detection, damping, and translating the center

of the axial oscillation of a charged particle in a Penning
trap with hyperibolic electrodes," Phys. Rev. 29A, 462-469
* (1984).

T. Goldman and M.M. Nieto, "Gravitational properties of

antimatter," Los Alamos National Lab Preprint LA-UR-85-1092,
Third LEAR Workshop, Tignes-Savoie, France (19-26 Jan 1985).
J.P. Gordon and A. Ashkin, "Motion of atoms in a radiation
trap," Phys. Rev. A21, 1606-1617 (1980).

T.W. H~nsch and A.L. Schawlow, "Cooling of gases by laser

radiation," Optics Comm. 13, 68-69 (1975).
*• H.F. Hess, D.A. Bell, G.P. Kochanski, R.W. Cline, D. Kleppner,
and T.J. Greytak, "Observation of three-body recombination in
spin-polarized hydr-'gen," Phys. Rev. .Lett. 51, 483-486 (1983).
G. Herzberg, Molecular Spectra and Molecular Structure - I.
Spectra of Diatomic Molecules, 2nd Ed., D. Van Nostrand,
Princeton, NJ (1950).

162

U'
 R.T. Hodgson, "VUV laser action observed in the Lyman band of
molecular hydrogen," Phys. Rev. Lett. 25, 494 (1970).
M.V. Hynes, "Physics with low temperature antiprotons," Los
Alamos National Lab preprint LA-UR-85-1060, Third LEAR
Workshop, Tignes, France (19-26 Jan 1985).
R.P. Johnson and J. Marriner, "Stochastic stacking without
filters," Fermilab P Note 226, Fermi National Accelerator Lab,
Batavia, Illinois (17 August 1982).

A.P. Kazantsev, "The acceleration of atoms by light," Sov.

Phys. JETP, 39, 784-790 (1974).
W. Kells, G. Gabrielse, and K. Helmerson, "On achieving cold
antiprotons in a Penning trap," FERMILAB-Conf-84/68-E, Fermi
National Accelerator Lab, Batavia, Illinois (August 1984).
[Preprint submitted to the IX Int. Conf. on Atomic Physics,
Seattle, Washington (23-27 July 1984).]

J. F. Lam, R.A. McFarlane, A.J. Palmer, and D.G. Steel,

"Experimental and theoLetical studies of laser cooling and
emittance control of neutral beams," Hughes Research
* Laboratories, Malibu, California, Annual Report on AFOSR
Contract F49620-82-C-0004, (September 1983).

V.S. Letokhov, V.G. Minogin and B.D. Pavlik, "Cooling and

capture of atoms and molecules by a resonant light field,"
Soy. Phys. JETP, 45, 698-705 (1977).
V.S. Letokhov and V.G. Minogin, "Possibility of accumulation
and storage of cold atoms in magnetic traps," Optics Comm.,
35, 199-202 (1980).
V.S. Letokhov and V.G. Minogin, "Laser radiation pressure on
free atoms," Phys. Reports 73, 1-65 (1981) [review].
M.S. Lubell and K. Rubin, "Velocity compression and cooling fo
a sodium atomic beam using a frequency modulated ring laser,"
pp. 125-136, Laser-Cooled and Trapped Atoms, NBS SP-653, W.D.
Phillips, ed. (1983).
H.J. Metcalf, "Magnetic trapping of decelerated neutral
atoms," Laser-Ccoled and Trapped Atoms, NBS SP-653, W.D.
Phillips, ed. (1983).
R.A. McFarlane, D.G. Steel, R.S. Turley, J.F. Lam, and A.J.
Palmer, "Experimental and theoretical studies of laser cooling
and emittance control of neutral beams," Hughes Research
Laboratories, Malibu, CA 90265, Annual report on contract
F49620-82-C-0004, AFOSR, Bolling AFB, DC 20332 (Oct 1984).

r• 163

S
 _7 W ~
." . ~
`7 _' =.ý-7 -r -I,

R. D. McCarty, "Hydrogen technological survey," Thermophysical

Properties," NASA SP-3089 [550 pages] (1975).
R.D. McCarty, J. Hord, and H.M. Roder, Edited by J. Hord,
"Selected properties of hydrogen (engineering design data),"
Center for Chemical Engineering, National Engineering Lab,
National Bureau of Standards, Boulder, Colorado (February
1981).
J.C. Mullins, W.T. Ziegler, and B.S. Kirk, "The thermodynamic
properties of parahydrogen from 1 to 22 K," Tech. Report
No. 1, Project No. A-593, Contract CST-7339, National Bureau
of Standards, Boulder, Colorado (1 November 1961).
I. Nebenzahl and A. Szlke, "Deflection of atomic beams by
resonance radiation using stimulated emission," Appl. Phys.
Lett. 25, 327-329 (1974).
W. Neuhauser, M. Hohenstatt, P. Toschek, and H. Dehmelt,
"Optical-sideband cooling of visible atom cloud confined in
Sparabolic well," Phys. Rev. Lett. 41, 233-236 (1978).
W. Neuhauser, M. Hohenstatt, P.E. Toschek, and H.G. Dehmelt,
"Visual observation and optical cooling of electrodynamically
contained ions," Appl. Phys. 17, 121-129 (1978).
W. Neuhauser, M. Hohenstatt, P.E. Toschek, and H. Dehmelt,
"Localized visible Ba+ mono-ion oscillator," Phys. Rev. 22A,
1137-1140 (1980).
F.J. Northrop, et al., "VUV laser-induced fluorescence of
molecular hydrogen," Chem. Phys. Let. 105, 34-37 (1984).
A.J. Palmer and J.F. Lam, "Radiation cooling with pi-pulses,"
"Paper WGI, Annual Meeting Optical Soc. Am., San Diego (1984)
[submitted to J. Opt. Soc. Am. (1985)].
D.B. Pearson, R.R. Freeman, J.E. Bjorkholm, and A. Askin,
"Focusing and defocusing of neutral atomic beams using
resonance-radiation pressure," Appl. Phys. Lett. 36, 99-101
(1980).
W.D. Phillips and H.J. Metcalf, "Laser deceleration of an
atomic beam," Phys. Rev. Lett., 48, 596-599 (1982).
W.D. Phillips, ed., Laser-cooled and trapped atoms, NBS
Special Publication 653, Proc. Workshop on Spectroscopic
Applications of Slow Atomic Beams, NBS Gaithersburg, MD (14-15
April 1983).

164

%
 W.D. Phillips, J.V. Prodan, and H.J. Metcalf, "Neutral atomic
beam cooling experiments at NBS," pp. 1-8, Laser-Cooled and
Trapped Atoms, NBS SP-653, W.D. Phillips, ed. (1983).
J.V. Prodan, W.D. Phillips, and H.J. Metcalf, "Laser
production of a very slow monoenergetic atomic beam," Phys.
Rev. Lett. 49, 1149-1153 (1982).
J.V. rordan and W.D. Phillips, "Chirping the light --
fantastic?," pp. 137-141, Laser-Cooled and Trapped Atoms,
NBS SP-653, W.D. Phillips, ed. (1983).
J. Prodan, A. Migdall, W.D. Phillips, I. So, H. Metcalf, and
J. Dalibard, "Stopping atoms with laser light," Phys. Rev.
Lett. 54, 992-995 (1985).
W-K Rhim, M.M. Saffren, and D.D. Elleman, "Development
electrostatic levitator at JPL," pp. 115-119 of Materials
Processing in the Reduced Gravity Environment of Space,
G.E. Rindone, ed., Elsevier Scienr.e (1982)

H.M. Roder, G.E. Childs, R.D. McCarty and P.E. Angerhafer,

"Survey of properties of H isotopes below their critical
temperatures," TN 641, National Bureau of Standards, Boulder,
Colorado (1973).
.110 F. Schmidt, "Diffusion and ortho-para conversion in solid
hydrogen," Phys. Rev. DIO, 4480-4484 (1974).
P.B. Schwinberg, R.S. Van Dyck, Jr., and H.G. Dehmelt, "New
comparison of the positron and electron g factors," Phys. Rev.
Lett. 47, 1679-1682 (1981).
I.F. Silvera, "The solid molecular hydrogens in the condensed
phase: Fundamentals and static properties," Rev. Mod. Phys.
32, 393-452 (1980).
I.F. Silvera and J. Walraven, "The stabilization of atomic
hydrogen," Sci. Am. 246, 66-74 (January 1982).
I. Simon, et al., "Sensitive tiltmeter utilizing a diamagnetic
suspension," Rev. Sci. Inst. 39, 1666 (1968).
R. Sprik, J.T.M. Walraven, and I.F. Silvera, "Compression of
spin-polarized hydrogen to high density," Phys. Rev. Lett. 51,
479-482 (1983).
D.G. Steel, J.F. Lam, R.A. McFarlane, "Studies of laser
enhanced relativistic ion beam neutralization," White Paper,
Hughes Research Laboratories, Malibu, California (October
1983).

165

r.
 . 7 7 . -7 7-7 - 1.

W.C. Stwalley, "A hybrid laser-magnet trap for spin-polarized

atoms," pp. 95-102, Laser-Cooled and Trapped Atoms, NBS SP-
653, W.D. Phillips, ed. (1983).
W. Thompson and S. Hanrahan, "Characteristics of a cryogenic
extreme high-vacuum chamber," J. Vac. Sci. Technol. 14, 643-
"645 (1977).
R.D. Waldron, "Diamagnetic levitation using pyrolytic
graphite," Rev. Sci. Inst. 37, 29-34 (1966).
R.S. Van Dyck, Jr., D.J. Wineland, P.A. Ekstron, and H.G.
Dehmelt, "High mass resolution with a new variable
anharmonicity Penning trap," Appl. Phys. Lett. 28, 446-448
(1976).
R. W. Waynant, J.D. Shipman, Jr., R.C. Elton, and A.W. Ali,
"VUV laser emission from molecular hydrogen," Appl. Phys.
"Lett. 17, 383 (1970).
M.J. Weber, Ed., CRC Handbook of Laser Science and Technology,
pp. 282-286, CRC Press, Boca Raton, Florida.
D.J. Wineland and H.G. Dehmelt, "Principles of the stored ion
calorimeter," J. Appl. Phys. 46, 919-930 (1975).
D.J. Wineland, R.E. Druilinger, and F.L. Walls, "Radiation-
pressure cooling of bound resonant absorbers," Phys. Rev.
Lett. 40, 1639-1642 (1978).
D.J. Wineland, "Trapped ions, laser cooling, and better
clocks," Science 226, 395-400 (26 Oct 1984).
W.H. Wing, "Electrostatic trapping of neutral atomic
particles," Phys. Rev. Lett., 45, 631-634 (1980).
W.H. Wing, "Some problems and possibilities for quasistatic
neutral particle trapping," pp. 74-93, Laser-Cooled and
Trapped Atoms, NBS SP-653, W.D. Phillips, ed. (1983).
W.H. Wing, "Gravitational effects in particle traps," p. 94,
Laser-Cooled and Trapped Atoms, NBS SP-653, W.D. Phillips
*- (editor) (1983).

166
7-V . j1,q .'7 T 7- 7

•.? 10.3 ANTIMATTER ANNIHILATION AND PROPULSION

L.E. Agnew, Jr., T. Elioff, W.B. Fowler, R.L. Lander, W.M.

Powell, E. Serge, H.M. Steiner, H.S. White, C. Wiegand, and T.
Ypsilantis, "Antiproton interactions in hydrogen and carbon
below 200 MeV," Phys. Rev., 118, 1371 (1960).
B.W. Augenstein, "Concepts, problems, and opportunities for
use of annihilation energy: an annotated briefing on near-term
RDT&E to assess feasibility," Rand Note N-2302-AF/RC, Rand
Corp., Santa Monica, CA 90406 (June 1985).
B.W. Augenstein, "Some examples of propulsion applications
using antimatter," Rand Paper 7113, Rand Corp., Santa Monica,
CA 90406 (July 1985).
H. Barkas and M.J. Berger, Tables of Energy Losses and Ranges
of Heavy Charged Particles, NASA SP-3013, STI Division, NASA,
Washington, DC (1964).
B.N. Cassenti, "Design considerations for relativistic
"antimatter rockets," J. British Interplanetary Soc., 35, 396-
404 (1982).
B.N. Cassenti, "Antimatter propulsion for OTV applications,"
AIAA preprint 84-1485, AIAA/SAE/ASME 20th Joint Propulsion
Conference, Cincinnati, Ohio (11-13 June 1984).
B.N. Cassenti, "Optimization of relativistic antimatter
rockets," J. British Interplanetary Soc., 37, 483-490 (1984).
F.R. Chang and J.L. Fisher, "The hybrid plume plasma rocket,"
Draper Lab preprint (1983).
M.R. Clover, R.M. DeVries, N.J. DiGiacomo, and Y. Yariv, "Low
energy antiproton-nucleus interactions," Phys. Rev. C26, 2138-
2151 (1982).
R.M. DeVries and N.J. DiGiacomo, "The annihilation of low-
energy antiprotons in nuclei," pp. 543-560, Physics at LEAR
with Low-Energy Cooled Antiprotons, Workshop on Physics at
LEAR with Low-Energy Cooled Antiprotons, Erice, Sicily, Italy,
9-16 May 1982, U. Gastaldi and R. Klapisch, ed., Plenum Press,
NY (1984).
D.F. Dipprey, "Matter-Antimatter Annihilation as an Energy
Source in Propulsion," Appendix in "Frontiers in Propulsion
Research," JPL TM-33-722, D.D. Papailiou, Editor, Jet
Propulsion Lab, Pasadena, CA 91109 (15 March 1975).

167

S'

!•.•-',.7...................................-.............,.',..............."...........,............
 R.L. Forward, "Interstellar flight systems," AIAA Reprint 80-
0823, AIAA Int. Meeting, Baltimore, MD (1980).
R.L. Forward, "Antimatter propulsion," J. British
Interplanetary Soc., 35, 391-395 (1982).
R.L. Forward, Alternate Propulsion Energy Sources, AFRPL-TR-
83-067, Final Report on Contract F04611-83-C-0013, Air Force
Rocket Propulsion Lab, Edwards, CA 93523 (December 1983)
R.L. Forward, "Antiproton annihilation propulsion," AIAA
preprint 84-1482, AIAA/SAE/ASME 20th Joint Propulsion
Conference, Cincinnati, Ohio (11-13 June 1984) [To be
published in J. Propulsion and Power].
R. Hyde, L. Wood, and J. Nuckolls, "Prospects for rocket
propulsion with laser induced fusion microexplosions," AIAA
Paper 72-1063 (Dec 1972).
W. Kolos, D.L. Morgan, D.M. Schrader, and L. Wolniewicz,
"Hydrogen-antihydrogen interactions," Phys. Rev. All, 1792-
1796 (1975).
H. L8b, "Sinn oder Unsinn der Photonenrakete," Astronautik 8,
39-47 (1971).
* E. Mallove, R.L. Forward, Z. Paprotny, and J. Lehmann,
"Interstellar travel and communication: a bibliography," J.
British Interplanetary Soc., 33, 201-248 (1980) [entire
issue].
P.F. Massier, "The need for expanded exploration of matter-
antimatter annihilation for propulsion application," J.
British Interplanetary Soc., 35, 387-390 (1982).

D.L. Morgan and V.W. Hughes, "Atomic processes involved in

matter-antimatter annihilation," Phys. Rev. D2, 1389-1399
*• (1970).
D.L. Morgan and V.W. Hughes, "Atom-antiatom interactions,"
Phys. Rev. A7, 1811-1825 (1973).
D.L. Morgan, "Rocket thrust from antimatter annihilation," JPL
Contractor Report CC-571769, Jet Propulsion Lab, Pasadena,
Calif. (1975).
D.L. Morgan, "Coupling of annihilation energy to a high
momentum exhaust in a matter-antimatter annihilation rocket,"
JPL Contract Report JS-651111 (1976)

D.L. Morgan, "Concepts for the design of an antimatter

annihilation rocket," J. British Interplanetary Soc., 35, 405-
412 (1982).

168

L'• • -"
'. -."• .•'''•
, ,'." ,'-'-
" "" "''• • •.•' 3.•
'- -- • ' • .k• "-•.'• " • • .".• '."*'. "*,• *=
"Particle Data Group, "Review of particle properties," Rev.
Mod. Phys. 56, Part II (April 1984).
G. Piragino, "Antiproton annihilation in nuclear matter:
multipion-nucleus interactions and exotic phenomena," pp. 855-
860, Physics at LEAR with Low-Energy Cooled Antiprotons,
Workshop on Physics at LEAR with Low-Energy Cooled
Antiprotons, Erice, Sicily, Italy, 9-16 May 1982, U. Gastaldi
and R. Klapisch, ed., Plenum Press, NY (1984).

S. Polikanov, "Could antiprotons be used to get a hot, dense

plasma?," pp. 851-853, Physics at LEAR with Low-Energy Cooled
Antiprotons, Workshop on Physics at LEAR with Low-Energy
Cooled Antiprotons, Erice, Sicily, Italy, 9-16 May 1982, U.
Gastaldi and R. Klapisch, ed., Plenum Press, NY (1984).

E. SAnger, "The theory of photon rockets," Ing. Arch. 21, 213

(1953) [in German].
L.R. Shepherd, "Interstellar flight," J. British
Interplanetary Soc. 11, 149-167 (1952).
G. Vulpetti, "Relativistic astrodynamics: The problem of
payload optimization in a two-star exploration flight with and
intermediate powered swing-by," IAF Paper 83-327, 34th IAF
Congress, Budapest, Hungary (Oct 1983).
G. Vulpetti, "A concept of low-thrust relativistic jet speed
high efficiency matter-antimatter annihilation propulsion
system," IAF Paper 83-397, 34th IAF Congress, Budapest,
Hungary (Oct 1983).
G. Vulpetti, "A propulsion-oriented synthesis of the
antiproton-nucleon annihilation experimental results," J.
British Interplanetary Soc. 37, 124-134 (1984)
G. Vulpetti, "An approach to the modeling of matter-antimatter
propulsion systems," J. British Interplanetary Soc., 37, 403-
409 (1984).
M. Wade and V.G. Lind, "Ratio of antiproton annihilations on
neutrons and protons in carbon for low-energy and stopped
antiprotons," Phys. Rev. D14, 1182-1187 (1976).
R.R. Zito, "The cryogenic confinement of antiprotons for space
propulsion systems," J. British Interplanetary Soc., 35, 414-
421 (1982).
R.R. Zito, "Chain reactions in a hydrogen-antiproton pile," J.
British Interplanetary Soc., 36, 308-310 (1983).

169
 "APPENDIX A

SOME USEFUL DATA AND CONVERSION FACTORS

FUNDAMENTAL CONSTANTS

Some useful fundamental constants are:A-I

c = 2.997925xi0 8 m/s speed of light

e = l.602192xi0- 1 9 C electronic charge
h = 6.626196xi0- 34 J.s Planck's constant
.4", = h/2n
m = 9.109558x10- 3 1 kg electron mass
= 5.110041x,0 5 eV electron rest mass energy
M = 1.672614xi0• 2 7 kg (anti)ptoton mass
= 9.382592x,0 8 eV (anti)proton rest mass energy
= 2.48823x10- 2 8 kg charged pion mass
1.39578xWO
1 8 eV charged pion rest mass energy
- 2.60x10- s charged pion mean lifetime

mwo = 2.40617x10- 2 8 kg neutral pion rest mass

= 1.34975x08 eV neutral pion rest mass energy
= 8.9x10-" s neutral pion mean lifet:iie

m• = 1.883156x1•28 kg muon rest mass

= 1.05659xi0 8 eV muon rest mass energy
= 2.198xi0- 6 s muon mean lifetime

k = 1.380622xi0- 2 3 J/K Boltzmann's constant

= 5.669620x10- 8 W/m2 -K 4 Stefan-Boitzmann constant

Ie = 9.274096x10- 24 J/T Bohr (electron) magneton

= 9.274096xi0- 21 erg/G
= 5"050951x0- 2 7 J/T nuclear magneton
= 5.050951xi0 2 3 erg/G

References:
"A.lD.E. Gray, Ed., American Institute of Physics Handbook, Third
Edition, McGraw-Hill, NY (1972).

A-1

.... . . . . . . . . .....
ENERGY CONVERSION FACTORS

Some energy conversion factors to help underitind atomic and

molecular spectra and energy level diagrams.

1 eV = 1.602192x10- 1 9 J

1.602192x10- 12 erg

1.782679xi0- 3 6 kg mass

2.417966x10 1 4 Hz frequency
8.065465x0 5 m-1 wavelength
8.065465x10 3 cm- 1 wavelength

1.160485x10 4 K temperature

HEAT ENERGY CONVERSION FACTORS

1 calorie = 4.1840 absolute Joules

= 4.1833 international Joules

PRESSURE CONVERSION FACTORS

1 atm = 1.01325 bar

= 14.70 psi

= 29.92 inHg

= 406.78 inH 2 0

= 760.0 Torr (mmHg)

= 101,325 Pascal (Pa)

A-2

S.1"~
SELECTED PROPERTIES OF HYDROGEN

Triple PointA-2
T = 13.800 K
P = 7.0420875 kPa (0.0695 atm)
d = 86.51 kg/m 3 (solid)
= 77.04 kg/mr3 - (liquid)
= 0.12558 kg/mr- (vapor)

Normal Boiling Point

T = 20.278 K (P = 1 atm)

Heat of Vaporization
Hv = 214.8 cal/gm-mole = 449.4 J/gm (P = 1 atm)
Heat of Fusion
Hf = 28.03 cal/gm-mole = 58.64 J/gm
3
Calculated Properties of Saturated ParahydrogenA'
Temperature Pressure Heat of Sublimation Heat Capacity
K Torr cal/gm-mole cal/K-om-mole
13.813 52.89 244.90 1.360
13.000 30.13 242.30 1.164
12.000 13.78 238.71 9.438x10-
11.000 5.567 234.76 7.471x10-
10.000 1.917 230.55 5.751xi0-1
9.000 5.343x10- 1 226.12 4.280xi0 -
8.000 1.118x10- 1 221.53 3.054x10-1
7.000 1.561x0- 2 216.82 2.065xi0-
6.000 1.198x10- 3 212.02 1.301x10-21
5.000 3.570x10- 5 207.15 7.420x10-
4.000 2.080x10- 7 202.24 3.799x10- 2
3.000 4.832xi0- 1 1 197.30 1.603x10- 2
2.000 3.985xi0- 1 d 192.34 4.749x10-3
1.000 8.255x10- 9 187.37 5.936x10-4

T P=1924e- 9 5 . 2 5 /T Cp=5.936xl0- 4 T 3

References
A. 2 R.D. McCarty, J. Hord, and H.M. Roder, Edited by J. Hord,
"Selected properties of hydrogen," Center for Chemical
Engineering, National Engineering Lab, National Bureau of
Standards, Boulder, Colorado (February 1981).
A.3j.C. Mullins, W.T. Ziegler, and B.S. Kirk, "The thermodynamic
properties of parahydrogen from 1 to 22 K," Technical Report
No. 1, Project No. A-593, Contract CST-7339, National Bureau
of Standards, Boulder, Colorado (1 November 1961).

A-3
 RELATIVISTIC MASS INCREASE AND TIME DILATION

* The mass m and the lifetime t of a particle increase over

the rest mass mo and the resting lifetime t9 with increased
total energy Et according to the relativistic equations:

A.1) t - = --
-- m - = ---
Et = Y- 1 -- -
- -- Ek
to miEo ( 2 12)2 EO

where Eo=mc is the rest mass energy of the particle and E

the kinetic energy.
The distance of travel d of the particle in vacuum is then
just the velocity of the particle v=A c times the
relativistic lifetime of the particle t=z to or:

SA.2) d = vt c to

"CENTER OF MASS ENERGIES

-" When a proton strikes a target with an incident energy E, the
" center of mass energy s1/2 for the reaction p-pp is:
2E + 2m2c4)i/2
A.3) Ecm = sl/2 = (2mc

where mc2 = 0.938 GeV. The same reaction could have been
obtained by colliding 01 two protons head on, with each
proton at the energy s -/2.
For very high energies this reduces to:
2 E) 1 / 2
A.4) sl/2 = (2mc

In GeV units, where c=l and m=l, this reduces to

1/ 2
A.5) sl/2 = (2E)

so that at high energies the center of mass energy in a

stationary target system only goes as the square root of the
incident proton energy.

For example, the Fermilab 400 GeV Main Ring gives center of
mass energies of 90 GeV, while the colliding beams at CERN of
270 GeV give a center of mass energy of 540 GeV. When
Fermilab achieves collisions of 1000 GeV (1 TeV) protons on
1 TeV antiprotons, the center of mass energy will be 2 TeV. It
would take an accelerator capable of reaching 2000 TeV to
achieve the same results using a stationary target.

A-4

S.. . . . .
-* - RELATIONS BETWEEN BEAM ENERGY AMD MOMENTUM

The total energy Et of an (anti)proton is related to the

momentum p by the relativistic equation:

2
A.6) E2 = m2 c 4 + p 2 c

where m is the (anti)proton mass and c is the speed of light.

For the (anti)proton, the rest mass energy is given by
mc2 0.938 GeV, while m2 c 4 =0.880 GeV 2 .
The total energy of the (anti)proton consists of the rest mass
energy Eo=mc 2 plus the kinetic energy Ek:

A.7) Et =Ek +Eo=Ek+mc2

Thus, the momentum of the (anti)proton in terms of the kinetic

energy is:

2
A.8) p 2 = 2mEk + Eý/C

For kinetic energies well below the rest mass energy of the
(anti)proton, ..he kinetic energy is given by the usual
relation:

A.9) Ek ~ p 2 /2m

or

A.10) p - (2mEk) 1 / 2

For kinetic energy well above the rest mass energy of the
(anti)protons, the kinetic energy is approximately equal to
the total energy

A.11) Ek Et ~ pc

Thus, for erergies much higher than the approximately 1 GeV

rest mass erergy of the (anti)proton, the numerical value of
the energy
roughly the insame.
GeV and
For the momentum
energy valuesin below
units 1 ofGeVGeV/c
they are
differ
significantly as is shown in Figure A-1.

A-5

,.". .. .. ...... .. . ... . .- . . .

 Jo .

10 - 1 1F I I II

E EK T1+ -

MOMENTUM, p .

0.1
i -- / KINETIC ENERGY, EK

; ~ ~~0.01 I I I III i ~ ii I I I1 1
* .00.01 0.1 "1 10

KINETIC ENERGY, EK, GeV

Figure A-I - Total energy and momentum vs. kinetic energy.

A-6

,%' -. -% %%% %%%. , .• - • . -" * . .-. -S, V

 APPENDIX B

COST COMPARISON OF

CHEMICAL AND ANTIHYDROGEN PRCPULSION SYSTEMS

FOR HIGH &V MISSIONS

* imhis paper was presented at the AIAA/SAE/ASME/ASEE 21st Joint

tropulsion Conference, Monterey, California (8-10 July 1985).

S .:- "B-I
i i '¶J3 IIII
I

IAIAA-85-1455
COST COMPARISON OF CHEMICAL AND
ANTIHYDROGEN PROPULSION SYSTEMS
FOR HIGH AV MISSIONS

ROBERT L. FORWARD
HUGHES RESEARCH LABORATORIES
MALIBU, CALIFORNIA

BRICE N. CASSENTI
UNITED TECHNOLOGY RESEARCH CENTER
EAST HARTFORD, CONNECTICUT

DAVID MtLLER
PURDUE UNIVERSITY
WEST LAFAYETTE, INDIANA

AIAAISAEIASMEIASEE 21st Joint'

Propulsion Conference
July 8-10, 1985/ Monterey California

For permission to c9py or republish, contact the American Institute of Aeronautics and Astronautics
1633 Broadway, New York, NY 10019
B-2
 COST COMPARISON OF CHEMICAL AND
ol ANTIHTDROGBN PROPULSION SYSTEMS FOR HIGH AV MISSIONS

Robert L. Forward Brice N. Cassenti David Miller

"Hughes Research Laboratories United Technology Lesearch Center Purdue University
Malibu, California East Hartford, Connecticut West Lafayette, Indiana

Abstract out divided by ac mains powel in) cotld be raised

from the present 10 to 10 , th,.' antihyd~ojen
Recent studies have indicated that it may be could be produced for an estimated 1OMI/mg.
possible to make and store antimatter in the form
of solid antihydrogen. For space propulsion, At present, the cost to lift anything into
milligrams of antihydrogen wo,,1d be used to heat space is Sk$/kg or 5M$/T. If the propulsion
tons of hydrogen reaction fluid to produce a effectiveness of antihydrogen fuel to matter fuel
specific impulse of 1000 to 3500 s tailored to the remains at I mg per 6 T, then the cost of the A T
mission characteristic velocity AV. For typical of reaction mass would be 3099, while the price of
deep space missions 1 milligram of antihydrogen 1 mg of antihydrogen fuel would bm only 1OM8.
would give the propulsion energy of 6 tons of Thus, basnd on fuel cost alone, artihydrogen would
liquid oxygen/liquid hydrogen fuel. In this paper seem to be cost effective for space mi~sions.
we carry out a detailed parametric study of the
comparative cost of chemical, nuclear thermal, and The previous estimates of the relative fuel
antihydrogen propulsion systems as a function of costs for antihydrogen and other propulsion systems
the mission characteristic velocity and the rela- were limited in scope. The purpose if this vaper
tive price of antihydrogen and propellant or is to look in detail at the relative fuel costs of
reaction fluid in low earth orbit, We find that various different missions and determine a price at
since the fuel cost of an antihydrogen propulsion which antihydrogen becomes cost effective for those
system does not rise exponentially with mission different missions.
characteristic velocity, AV, but only as the square
of AV, antihydrogen propulsion will always be more Thesis and Assumptions
cost effective than chemical propulsion at
sufficiently high mission characteristic The question we address in this paper is
' velocities. deliberately limited: 'What is the relative total
fuel cost of chemical, nuclear thermal, and
antihydrogen propulsion systems for orbit transfer
"and other high velocity missions?' To keep the
Introduction study from getting too complex, we have made a
number of assumptions:
There has recently been a series of publica-
tions on the feasibility of using antimaftgrain the The mission is entirely in space. There
form of antihydrogen for space propulsion. - The are no landing scenarios or aerobraking
annihilation of antihydrogen with hydrogen will maneuvers.
produce large amounts of energy. The annihilation
of a milligram of antihydrogen with a milligram of * The simple rocket equations relating the
hydrogen produces the same amount of energy as 40 mass ratio of the vehicle to the mission
. metric tons of TNT or 12 metric tons of liquid characteristic velocity and the
oxygen/liquid hydrogen fuel, propellant exhaust velocity are adequate
to describe the actura performance of the
The antiprofons in antihydrogen are preferred propulsion system.
over actielectrons for propulsion, since two-thirds
of the annihilation energy appears as kinetic 0 Only fuel costs will be conpared. The
energy in the form of charged particles (pions), cost of designing and building the space
The charged particles can then transfer their vehicle itself will not be considered.
kinetic energy to a hydrogen working fluid, which No vehicle capable of handling anti-
can be used to prcvide thrust. Only milligrams of hydrogen or hot hydrogen reaction mass at
antihydrogen will be needed to heat tons of specific impulses from 1000 to 3600 s bas
"hydrogen, and overall propulsion efficiencies yet been built, w03le designs exist for
(annihilation power to thrust power) of 25 to 50% chemical and nuclear therml rockets.
have been predicted. For a typical mission, one The cost of building the first anti-
extra milligram of antihydrogdn allowed a mission hydrogen powered rocket will be large,
to be done wit4 6 metric tons less hydrogen but it will not be included in these cost
reaction mass. comparisons.
Antimatter in the for% of antiprotons i40bhng The mission is time urgent. Long orbit-
made and stored today in very small awounts.. raising maneuvers and long, low thrust
The publications on antihydr-gen propulsion make missions are not considered, This
the case that it should be possible to eliminates from consideration solar
significantly increase the efficiency of the electric and nuclear electric propulsion
present production methods, turn the antiprotons zs well as snlar sails. Of course, solar
into antihydrogen, and store the antihydrogen as sails with their zero use of fuel will
charged ice crystals electrostatically levitated in always have the lowest fuel cost pir
a small cryogenically cooled chamber. If the mission and should always be used if txm'ý
production energy efficiencies (annihilation energy is not a consideration,

Copyight Amterican Institute of Acronautls and

Astronautics, Inc.. 1985. All hts reed.B-3

K-%
 - ý . . - 7

Txc ratio of delivered payload to empty systems, then the mass of the propellant consists
m
vehicle mas3 is the same for all the of the mass of the reaction fluid, r, and the mass
systems. Storable chemical and liquid of the energy source, m :
oxygen/liquid hydrogen chemical
propulsion systems will need more tankage a +m *v +mr +e
than the other concepts. A nuclear my+2r*e (2)
thermal propulsion system will have a v , v
heavy -eactor and shielding. An
antihydrcgen propulsion system will have or
a hravy reaction chambir and will also
require 3ome shielding ap= mr + me = (R - 1)mv (3)

The specific impulses of chemical and

nuclear thermal rockets are fixed by the
inherent properties of the propulsion The mass ratio, R, is related to the average
cystem. For storable zhemical exhaust ielocity, v, of the propellant or the
propellants it will assume a maximum reaction fluid plus energy source mass through the
;pecific impulse of 300 a or an exhaust relation,
felocity of 3 km/i. For a cryogenic
liquid oxygen/liquid hydrogen propellant R = ev/v (4)
system we will assume a maximum specific
impulse of 500 s or an exhaust velocity
of 5 ka/s. Although theoretically a eAV/gsp
nuclear rocket can have any specific = •
impulse, we will assume that it iv
Pýmited by the melting point of the where the mission characteristic velocitr, AV, is
reactor/heat exchanger to 9CP) s or an the oum of all the valocity changes that are
exhaust velocity of 9 kz/s. required ;y the mission,

An antiLydrogen rocket can be operated at 'Impossiblae Missions

any desired specific impulse. Real
antihydrogin rockets will havy an upper There are some missions in the solar system
limit to the specific impulhe that is
determined j tefaperature considerations. that wo'lid be desirable to accomplish for
Troposv antihydrogen rocket designs w'l scientifib purposei but which are essentially
operate at temperatures where the impossible using chea~cal or even nuclar thernal
rerteor-alumd has turned into a plasma rockets because of the expoaential growth of the
and trýe ;rato
- id contained and directed mass ratio with increasing mission characteristic

by maoneti: fields. We will assume that velocity. One example is L solar iupact missioa,
sih engine5 ca t5e built and that they which requires the re..ket to cancel out the orbitai.
can uroduce 'i.:' th-st at high exhaust velocity of the earth so tho vehicle can drop
Ilewill not eire the came directly into the sun, This requires a mission
velot;eCa. Another is
characteristic velocity o* AV= 35 km/s
•unf•il tosthe ould~ use a rendezvous mission to the rings deep dowv in the
rasal rocpreset
hbut will assumo it would use ýha present
rvtwiloSaun TsasinreiesaA
thermal reactor core design thLt is grvity w' ll of Saturn• This aission requires a
of 48 km/s.
limited by tlae melting pe:nt of the core
to specific impulses if !tthan 900 9. TThere aae even simpler missions near earth
that are nearly impossible using chemical rockets.
Rocket Equationsaie is thu wimple maueuver of leaving an orbiting
-Lao ýo inspect o- pick up a .atellite orbitiug in
We will assume that the perfotzance Qf t'a the •ppo'ite direction. This maneuver requiree
propulsioD systems is based oa the iollowir zancelirng tký initiba orbital velocity and
s ase
prouiso~ ystmso~thei'olow'tbuil'i-s -7%in 7n ,hbe opposite direction.
siuplified equations. Tho exhaust velocity, v, of I oppositeLivt th-i esath ist7.
Since orbital, ve--,city
Since is 7.7 ka/s,
the propellant is related to the specific impulse, the .;otal i:ssirn ,blaracter;:tic velocity of the
Isp, by the relation, reverse orbit rraeuv-r is i&.- km/s, If it is then
v = g I ) desired to reftu-n '.cta- orbiting base, rhe process
v m st be repeated fo" a -,,I AV of J1 km/s.
where g = 9.8 n/sec2. The mas3 ratios requirej for each type nf

The mass ratio, R, is 'oefined as the rrrio of rocket syitom tk, carr) ou', thes, missions can be
the mass of the empty vehicle (including payload), calcutlate! from goq•ation \4) and :re listed in
U.,, plus the mass of the propellant, m , divided b; Tible 1, All oi these mslsions remyire high wass
the mass of the empty vehicle, If therpropulsion ratins, the more difti-dlt nciw riqutring ruch
systeo ha- its energy source sepaiate from its k-r-e mass ratios that it %u extrecaly diffi.zi'&t to
reaction fluid, as is the case wiO both the iupine how one )ight buil, .4 icle to accoWi sb
nuclear thermal and antihylrogen propulsion such Lssions uvsin chemi.,-, c.: erv' nusear

B-4

~~~~~............-.-....--.-.--.-- ... •..'........'.•.-.\.............. ..... :..' < '.• •.,..,". . . ..

 Table 1. Mass Ratios for Difficult Missions This means that the mass ratio is a constant,
Amazingly, this constant mass ratio is independent
" 151=2 of the efficiency of the energy conversion and
independent of the mission characteristic velocity:
TOTAL MASS RATIO, R

aV(km/sI STORABLE 02 /H 2 NUCLEAR R = eAV/v = e1.59 = 4.9 (9)

300s 500s 900s
This constant mass ratio for minimum antihydrogen
REVERSEORBIT 155 175 22 6 consumption holds for all conceivable missions in
DOUBLE REVERSE ORBIT 310 30.700 490 32 the solar system and only starts to deviate
significantly for interstellar missions where the
SOLARIMPACT 350 117,000 1,100 49 mission characterisfic velocity starts to approach
the speed of light.
480 8.900,000 15,000 200
SATURN RING RENDEZVOUS
The amount of antihydrogen needed for a
specific mission is obtained by substituting
thermal rockets. As we shall see later, all of Equation (8) into (7) to get the mass of the energy
these missions could be performed by an source, I .- The antimatter is just half of this.
antihydrogen rocket with a mass ratio of 5:1 or It is found to be a function of the2 square of the
less. mission characteristic velocity, AV (essentially
the mission energy), the empty mass of the vehicle,
Minimum Antihydrogen Optiwization my, &aJ the conversion efficiency, e:

In this subsection, we will outline a 1 0.39 AV2

mathematical proof for the optimization of an m = 2 me - 2 y ". (10)
antihydrogen powered rocket mission in which the e c
amount of antihydrogen used is minimized. The
proof islgimple, but hq only been documented in For a typical antihydrogen mission where the
journals and reports that are somewhat antihydrogen energy to thrust energy conversion
difficult to obtain, so we will repeat it here. efficiency, e, equals 0.3, only 12 mg of
antihydrogen and 3.9 metric tons of propellant are
In an antihydrogen rocket, the source of needed to accelerate 1 ton of payload to 30 km/s
propulsion energy is separate from the reaction (0.0001 c). Thus, no matter what the mission, the
fluid. Thus the total initial mass of the vehicle vehicle will always use 3.9 tons of propellant for
consists of the empty mass of the vehicle, a , the each ton of payload and an insignificant amount (by
mass of the reaction fluid, mr, and the maswyof the weight, not cost) of antibydrogen.
energy source, me half of which is the mass of the
antihydrogen, a, that we wish to minimize. The
mass ratio is tien Cost Equations

=
- + mr + m ( We now use the rocket equations to compare the
/ _ e r total fuel cost for a nnmber of different
v propulsion systems. For both storable and
cryogenic chemical pr.pulsion systems, the mass of
the ene:gy source is in the propellant, Thas, the
The energy in the exhaust comes from the conversion fuel cost for the chemical rocket is just the cost
of the fuel rest sass energy to kinetic energy witL of the propellant mass in orbit. For a Luclear
an efficiency, e: thermal rocket, the energy to hext the reaction
2 2 2
r fluid is in the nuclear reactor. A reactor must
=m m +f me) v mV (6) have a certain minimuL charge of uraniua jutt to
operate, and carries much more uranium t•ns will be
used in %ny reasonable mission. Therefore, we have
Combining Equations (5) and (6) and rearranging we %ssumed that the mass and cost of the uranium
obtain energy souxce is charged to the empty vehicle mass
and cost. The fuel cost for the nuclear thermal
u v2 v = ( rocket will be the cost of the -eaction fluid mass
2 (e - (ex I) ,(7) in orbit. For rn antihydrogen rocket, the cost of
2E c2 x2 the antihydrogen part of the energy source is not
S 2c negligible The total fuel cost for the
where x = AV/v and k = mvAV/2Ec antihydrogen rocket will be tne cost of the
antihydrogen plus the cost of the reaction fluid
mass.•
that fuel costs
We now make the assumption
dominate the reaction fluid costs and we want to Fuel Cost of a Chemical %-'Nuclear Thermal Mission
minimize the amount of antihydrogen. By setting
the derivative of Equation (7) with respect to x The fuel cost, Cc, of a missior using a

rshown
equal 1o01 ero and solving for x, it can be
that the fuel is minimized when
chemical p.'opulsion system
propellant per kilogram in
times the propellant ass,
is the price of
low earth orbit, pp,
mp , needed for thep
v = 0.63 AV (8) mission.

B-5

°%
 C(11) just as in a chemical rocket. In an antihydrogen
rocket, however, this exponential rise in reaction
fluid mass can be curbed by using more antihydrogen
= pmv (R - 1) and increasing the reaction fluid exhaust velocity.
Thus, for low exhaust velocity the second term
SIeAV/vthe
becomes large, while at high exhaust velocity
Pp (e -) first term becomes large. There is a cost minimum
for each mission characteristic velocity, depending
AV'/&I on the relative price in low earth orbit of
Spp mv (eV sp - 1) hydrogen and antihydrogen.

Since the chemical propulsion system has a fixed Parametric Studies

specific impulse or exhaust velocity, we see from in this section we carry out a parametric
Equation (11) that the cost of any chemical or analysis of Equation (15) to determine the total
nuclear thermal rocket system rises exponentially fuel cost for an antihydrogen propulsion system for
with increasing mission characteristic velocity as various values of the parameters of mission
soon as the mission velocity exceeds the exhaust characteristic velocity and relative price ratio of
velocity: antihydrogen to propellant or reaction fluid.

C AV>v> p e AV/v (12) By varying the parameters we were able to

c P v establish a fuel cost minimum at each relative
price ratio for various mission characteristic
A similar conclusion can be derived for
nuclear thermal rocket., Although the nuclear the volocities, Tkese total fuel cost minimum values
for the antihydrogen propulsion system are then
reactor still has plenty of energy left, if the compared with the total fuel costs for chemical and
exhaust velocity is limited by thermal nuclear thermal rockets given by Equation (11),
considerations, then an exponentially growing
amount of reaction mass will be needed for the more
difficult missions, It is this exponential growth If antihydrogen is extremely expensive and the
of mass ratio and the fuel costs associated with it price ratio is high, then Equation (15) has the
that has led to the labeling of some missions as same form as Equation (7). We can then use the
'impossible,' same technique that was used to minimise the amount
of antihydrogen to minimize the total fuel cost,
Fuel Cost of an Antihydrogen Powered Mission The optimum exhaust velocity, mass ratio, and
antihydrogen mass used for minimum cost are those
The fuel cost, C , of an antihydrogen rocket given by Equations (8), (9), and (10). Those
consists of the priceaof the antihydrogen, p , optimum values for those parameters are indicated
times the mass of antihydrogen, m , used plut the as the 'asymptotic limits' in the figures that
price of the reaction fluid, pr' times the mass of follow.
the reaction fluid, m , heated by the energy from
the antihydrogen: r We were not able to find a simple analytic
solution to the minimisation of the cost expressed
C = p m + p m . (13) by Equation (15) when the relative costs of the
antihydrogen and reaction fluid were comparable.
In an antihydrogen rocket, the propulsion energy Instead, a computer was used to calculate the total
fuel cost, mass ratio, and antihydrogen mass used,
comes a from
with smalltheamount
annihilation of the matter
of the normal antihydrogen
in the as a function of the mission characteristic
witha
mout sall
o thenoral atte inthe velocity over a r~nge of specific impulses or
propella-it, The energy obtained is thus twice the elocit veroairang o speif iplse or
rest mass energy of the antihydrogen. Some exhaust velocities and a range of relative fuel
ponertione,e, nof
portion, of this annihilation energy is then
thisannetih tenergy thep
isln.:
costs. The minimum in the fuel cost was found and
this determined the optimum values for the other
converted into kinetic energy of the propellant: quantities,

2E mac2 A12 mrr

r
2
(14)
ratio Anof important
the price p!)rameter in these studies is the
of antihydrogen to the price of
Substituting Equation (14) into Equation .13) we propellant or reaction mass in orbit. To be more
obtain: general, we should have plotted the following
curves in terms of a dimensionless 76 ice rajio.
2 (5) Since the price ratios vary from 13 to MO,
C (Pr + Pa v ) m however, they are so large that they are almost
4E c meaningless.
propellant or Instead,
reaction we
massfixed .he price
in orbit at theof either
2 present-day price of 5k$/kg or 54$/T to lift mass
v eAV/v into low earth orbit and presented the parametric
(Pr + , (
(ev - 1) curves in terms of the price of intihydrogen per
milligram. Thus, a curve which i. labeled SM!/mg
In examinilg Equation (15) we see that there is a is equivalent to a relative price ratio of
cost minimization possible for an autihyerogen
rocket. As the requirid mission characteristic
velocity increases, the cost of the missit.n tends
to increase exponentially since the amount of S/m 109 (16)
reacticn fluid mass needed ie rising exponentially, 5kt/kg 0

B-6
 Optimum Mass Ratio 100

Figure 1 is a plot of the mass ratio of a 300s

.. '
number of different high thrust, fast response 60 /
propulsion systems as a function of the missionr2/12 40 -500s
characteristic velocity, The lines for a storable
chemical system with a specific impulse of 300 s,a NUCL2AR
90S
cryogenic O'/N2 chemical system with a specific
impulse of O0 s, and a nuclear thermal rocket
system with a specific impulse of 900 s were all 20o
obtained using the mass ratio expression given in
Equation (4).

Also included in Figure 1 is the asymptotic 210

limit of 4.9 for the mass ratio of an antihydrogen 8

8
rocket using the minimum amount of antimatter and a 6
set of curves giving the optimum mass ratios ASYMPTOTIC LIMIT R 4.9 20
determined in our parametric studies as a function "0 2S-
of the price of antihydrogen.

In Figure 1 we see that as the mission becomes ANTIMATTE

more difficult with increasing AV, the mass ritios 2 PRICE (M/m
Of the chemical and nuclear thermal systems risei
exponentially. If we assuse that it is difficult
to build
ratio muchor greater
stack a than
vehicle
100, with
then anstorable
overall chemical
mass

fuels cannot be used for missions with a 0 10 20 30 40 50 60

characteristic velocity greater than 15 km/s. A MISSION CHARACTERISTIC VELOCITY, AV, km/s
c.'yogenically cooled 02 /H2 propulsion system could
be used instead, but even this most energetic of
chemical fuels begins to fail for missions
*requiring a Al' of 25 km/si If a nuclear rocket Fig. 1 Mass ratio vs mission characteristic
could be built and flown with a specific impulse of velocity.
900 s, then it could be used for the more difficult
missions, but even it begins to fail when the AV'
exceeds 40 km/s.

In contrast, all of the antihydrogen 60

propulsion systems have mass ratios of less than 5. ANTIMATTER
Depending on the relative price of antihydrogen and PRICE, MS/mg
propellant or reaction fluid, and the difficulty of
the mission, the values for the mass ratio are
significantly less than 5, with a typical mass
ratio value being 3:1. E 40 2

Optimum Exhaust Velocity 10

Figure 2 illustrates the optimum exhaust 3 31

velocity of an antihydrogen rocket as a function of U.
the mission characteristic velocity for various
antihydrogen prices. The asymptotic limit for high
antihydrogen prices is the bottom line with a slope X 20
given by Equation (8).

As the price of antihydrogen dygps below 10

50S$/mg (relative price ratio of 10 ), the
parametric studies indicate that the total fuel
cost can be minimized by using more antihydrogen to
heat the reaction fluid hotter, thereby obtaining a 0 10 20 30 40 50 60
higher specific impulse and exhaust velocity and MISSIONCHARACTERISTIC VELOCITY iV km/s
decreasing the amount of reaction fluid required.

The interesting characteristic of these

optimum exhaust velocity curves is that they are Fig. 2 Exhaust velocity vs mission characteristic
nearly parallel to the asymptotic limit line. 'he velocity,
slopes are slightly different, however, ;.nd they
will intersect with the asymptotic line at very
high mission velocities. As an example, the top
line in Figure 2 for the optimum exhaust velocity
when the price of 1 ntihydrogen is lMt/mg (relative
price ratio 2 x 10 ) can be represented by the At the probable price of antihydrogen of 1013/mg
equation, the specific impulse required to carry out any
"mission in the solar system is seen to range from
v = 25 km/s + 0.50 AV (17) 1000 to 3500 s.

1B-7
•- -- - - - - " -.'-.----.- -'..'.'..'.
--. -
.... ----.
.. - -.- -.. .-. . -.- .--.. *-- --- h -... .. . --.-..---..-.- - --. -.- -
Optimum Antihydrogen Mass 1oo I , I I
80 i
Although the optimum exhaust velocity jumps 60
with each decrease in the relative price of
antihydrogen to reaction fluid, the amount of 40
antihydrogen does not change in the same fashion.
Figure 3 shows the optimum amount of antihydrogen 20
needed per ton of empty vehicle mass, m , to 20
produce the minimum total fuel cost. TKe bottom
line is the asymptotic limit when the price of
antihydrogen is high. The asymptotic limit rises 100
as the square of the mission characteristic 8
velocity (essentially the energy of the mission), > 6 -
The asymptotic limit for the amount of antihydrogen
needed is also a function of the efficiency of 2 4 "
conversion of annihilation energy into thrust. SANTIMATTER
PRICE, M$/mg
We have assumed for the parametric studies 2
that the overall efficiency of the conversion of a
annihilation energy into thrust is 30%. This 40
efficiency estimate consists of a 67% conversion of M
annihilation energy into charged particles, a 50% 0.8 10
efficiency of conversion of the energy of the 06
charged particles into energy in the working fluid,
and a nossle expansion efficiency of 90%. 04 - '
Variations in this efficiency will not affect the 4 739
mass ratio or exhaust velocity optimums, but will 10 It 0.3
directly affect the amount of antihydrogen needed 02
for a given mission and the total fuel costs for
the mission.
1 6 8 10 20 40 60 80 100
From plots of the optimum amounts of 2 4

antihydrogen as a function of its price we see that MISSION CHARACTERISTIC VELOCITY, AV, km/s

at the probable price of antihydrogen, lOMS/mg

(relativL price ratio of 2 x 10 ), the amount of
antihydrogen needed drops to the asymptotic limit
at missitn characteristic velocities greater than Fig. 3 Antihydrogen needed vs mission velocity,
20 km/s. Yet from Figure 2, the exhaust velocity
is still significantly higher than the asymptotic
exhaust velocity, because in optimizing the total
fuel cost it was found to he more cost effective to
decrease the amount of reaction fluid rather than 100
increase the amount of antihydrogen. The decreased 9
.
RA•°E / oI0Iz I
amount of reaction fluid is then ejected at a I /
higher velocity. The mass ratio of the vehicle is 80
less than the asymptotic limit, while the exhaust / / ',"3-
velocity is greater than the asymptotic limit. The U / /
"interestingfeature of Figure 3 is that the mass of W "
antihydrogen required for a mission does not depend / N CER

significantly on the cost of the antihydrogen for a/ .10'

any of the difficult missions (AV>10 km/s) that .2

R-8 -
"antihydrogen rockets are best suited for. L' ... - ,-

"Comparative Total Fuel Costs 0.: 254.--

We next calculated the total fuel costs for a 0 10 15 20

number of different propulsion systems and compared MISSION CHARACTERISTIC VELOCITY NIV), km/s
the total fuel costs as a function of mission
characteristic velocity. The propulsion systems
considered were the storable chemical propulsion
system with a specific impulse of 300 s, a Fig, 4 Relative fuel cost vs mission
cryogenic 0 2 / H chemical propulsion system with a characteristic velocity.
specific impulse of 500 s, a nuclear thermal rocket
system with a specific impulse of 900 s, and three
antihydrogen rystas that were optimized for lowest fluid for the n'iclear end antihydrogen rockets is
total fuel cost as a function of the price of the assumed to be 5kS/kg and the antihydrogen price is
antihydrogen. that indicated for the various antihydrogen curves.

Figure 4 presents the range of mission In examining Figure 4, we see that if the price of
characteristic velocities that are typical of antihydrogen can be brought dpwn to 20M$/mg (or a
present-day missions, The ordinate showing relative cost ratio of 4 x 10 ), then an
relative fuel cost can be converted directly into antihydrogen propulsion system is always more fuel
millions of dollars per ton of empty vehicle masa cost effective than a storable chemical propulsion
if propellant for the chemical rockets and reaction system. It is also better than the best chemical

B-8
 propulsion system now available (02/H2) for any ,- --

,'
mission characteristic velocity greater than 12
km/se At a pri~e of l01$/mg (or a relative cost 2 .
1
0 M /
--
/ Ji
ratio of 2 x 10 ), antihydrogen propulsion systems I IToRA/
are better than any chemical propulsion system at IsORA3W
I.
IEEV

any mission velocity, but are not as cost effective s I

If the price of 8 ,I.
as nuclear thermal propulsion.
antihydrogen drops to 2113/mg (relative price ratio 1
of 4 x 10 ), then antihydrogen propulsion is more 0 I / /
cost effective than any other known propulsion / /
system at any mission velocity.

Also shown in Figure 4 are some typical D SATURN

operational parameters for a mission requiring a AveR SAR A,•
total AV of 15 km/s. At an antihydrogen price of "VCRs, ORBT It ORB,
20M$/mg, each ton of empty space vehicle requires 0 ,0 2 3
"' only 3.2 mg of antihydrogen costing $64M to heat
2.2 metric tons of reaction fluid costing 11iM to a
specific impulse of 1300 s. Thus, one ton of
"vehicle can be pushed to 15 km/s by an antihydrogen
propulsion system for a total fuel cost of $75M, Fig. 5 Relative fuel cost vs expanded mission
while to do the same job with a O /H 2 rocket would velocity.
cost 3100M. If the price of antigyd-ogen drops to
2M$/mg, then each ton of delivered vehicle mass
would require the use of 4.4 mg of antihydrogen propulsion system can deliver a 5 T vehicle to the
cou.ing 88.8M to heat 0.8 tons of reaction fluid rings of Saturn for a total fuel cost of just 3200M
costing 34M to a specific impulse of 2500 s. At and complete the mission in months instead of
this price for antihydrogen, the total fuel cost to years.
push a ton to a velocity of 15 km/s would be only
$12.8M. Conclusions

Antihydropen Propulsion Enables In this paper we have carried out a parametric

"Impossible' Missions study of the comparative total fuel costs for
"storable chemical, cryogenic chemical, nuclear
- -- In Figure 5 we expand the scale of the plot of thermal, and antihydrogen propulsion systems for
relative fuel cost versus mission characteristic various mission characteristic velocities and
velocity from the scale of Figure 4, which shows various relative price ratios for antihydrogen and
the missions that are being considered in the near propellant or reaction fluid in space. Under
"future, to a scale that includes missions that are several restrictive assumptions we have shown the
'impossible' using any chemical or nuclear thermal following:
system, In Figure 5 the relative fuel cost scale
can be converted into millions of dollars per ton * Since an antihydrogen propulsion system
of delivered empty vehicle mass if the cost of can operate at any specific impulse by
propellant or reaction fluid in space is 5k$/kg or changing the ratio of antihydrogen to
5M$/T. hydrogen, the total fuel cost only rises
as the square of the mission
In this figure it is easier to see the characteristic velocity, Chemical and
differences in the shapes of the total fuel cost nuclear thermal systems with a fixed
curves for the different types of propulsion exhaust velocity have a fuel cost that
systems. For those propulsion systems with a fixed rises exponentially with mission
specific impulse, the fuel cost rises exponentially characteristic velocity. Thus,
with increasing mission characteristic velocity. antihydrogen propulsion will always be
Since the antihydrogen propulsion systems can vary more cost effective than other forms of
the exhaust velocity to match the mission, the propulsion at sufficiently high mission
total fuel cost for an ansihydrogen propulsion characteristic velocity,
system only rises as the square of the mission
characturistic velocity, Thus, no matter what the * Chemical propulsion systems are nearly
cost of antihydrogen, it will always be more cost always more cost effective for mission
effective than any propulsion system with a fixed characteristic velocities of less than 5
exhaast velocity at sufficiently high enough km/s,
"missiou velocity. For example, even at 1OM/mg, an
antihydrogen propulsion system will cost less than * If the price ratio of antihydrogen to
a nuclear thermal rocket if the mission propellant gr reaction fluid is greater
characteristic velocity desired is greater than 30 than 4 x 10 (20M1/mg), then antihydrogen
km/s (just off the top of Figure 5). propulsion systems are more cost
effective than chemical propulsion
If the price of antihydrogen can be brought systems for mission characteristic
down to l1M/mg or less, then antihydrogen velocities of greater than 12 km/s.
propulsion can open up the entire solar system and
allow the performance of missions that are now * If the cost of antihydrogen to propellant
impossible to consider us;ag any present propulsion or reaction mass is less than 2 x 10
system. For example, a. 113/mg, an antihydrogen ( 1MS/ng), then antihydrogen propulsion

B-9

,-n.
 systems are more cost effect,-.e than 7. R.L. Forward, "Antiproton annihilation
chemical propulsion systems and even propulsion,* AIAA preprint 84-1482,
nuclear thermal propulsion systems for AIAA/SAE/ASME 20th Joint Propulsion
any mission characteristic velocity over Conference, Cincinnati, Ohio (11-13 June
5 km/s. 1984).

8., B.N. Cassenti, "Antimatter propulsion for OTY

applications,$ AIA preprint 84-1485,
AISS/SAE/ASME 20th Joint Propulsion
Acknowledgments Conference, Cincinnati, Ohio (11-13 June
Robert L. Forward acknowledges support from Order 1984).
No. RI-32901 with the University of Dayton Research
Institute under Contract F04611-83-C-0046 with the 9. R.L. Forward, Alternate Propulsion Energ-
Air Force Rocket Propulsion Laboratory, Edwards Air Sources, AFRPL-TR-83-087, Final Report on
Force Base, California 93523. Contract F04811-83-C-0013, Air Force Rocket
Propulsion Lab, Edwards, CA 93523 (December
1983) (Distribution limited to U.S. Covernment
and Contractors.]
10, CERN Proton Synchrotron Staff, 'The CERN PS
complex as an antiproton source,' IEEE Trans.
References NS-30, 2039-2041 (1983),
1. R.L. Forward, 'Interstellar flight systems,' 11. J. Peoples, 'the Fermilab antiproton source,'
AIM Reprint 80-0823, AIAA Int, Meeting, IEEE Trans. NS-30, 1970-1975 (1983).
Baltimore, MD (1980).
12. T.A. Ysevolozhskaya, B. Grishanov, Ya.
2, P.F. Massier, 'The need for expanded Derbenev, N. Dikansky, I. Meshkov, V.
exploration of matter-antimatter annihilation ParkLomchuk, D. Pesrikov, G, Sil'vestrov, and
for propulsion application,' J. British A. Skrinsky, *Antiproton source for the
Interplanetary Soc. 35, 387-390 (1982). accelerator-storage complex, UNK-IHEP,'
Fermilab Report FN-353 8000.00 (June 1981), a
3. R.L. Forward, 'Antimatter propulsion,* J. translation of INP Preprint 80-182 (December
British Irterplanetary Soc. 35, 391-395 1980).
(1982).
13. L.R. Shepherd, 'Interstellar flight,' J.
4, B.N. Cassenti, 'Design considerations for British Interplanetary Soc. 11, 149-167
relativistic antimatter rockets,' J. British (1952)..
Interplanetary Soc. 35, 398-404 (1982)..
14, D.F. Dipprey, 'Mattei-Antimatter Annihilation
5, D.L. Morgan, 'Concepts for the design of an as an Energy Source in Propulsion,' Appendix
antimatter annihilation rocket,' J. British in Frontiers in Propulsion Research, JPL TM-
Interplanetary Soc. 35, 405-412 (1982), p3-722, D.D, Papailiou, Ed. (Jet Propulsion
Lab. Pasadena, CA (15 March 1975).
6, G. Vulpetti, 'A propulsicý-oriented synthesis
of the antiproton-nucleon annihilation 15. B.N. Cassenti, "Optimization of relativistic
experimental results,' J. British antimatter rockets,' J. British Interplanetary
Interplanetary Soc. 37, 124-134 (1984). Soc. 37, 483-490 (1984),

B-1O

.... .... ... .... $b.. C..',.''.O*.4%

 APPENDIX C

ANTIPROTON ANNIHILATION PROPULSION

This paper was prepared and presented under the present

contract although a good deal of. the material used was based
on the results of a previous contractC'I with the Air Force
Rocket Propulsion Laboratory. It was first presentedC 2 at
the AIAA/SAE,/ASME 20th Joint Propulsion Conference in
Cincinnati, Ohio on 11-13 June 1984, and a shorter version was
presentedC- 3 at the 35th Congress of the International
Astronautical Federation in Lausanne, Switzerland on 7-13
October 1984. It was submitted tu the Journal of Propulsion
and Power and was accepted for publication after reduction in
"length, number of figures, and number of references. The
reduced version that will appear in a future issue of the
Journal of Propulsion and Power (late 1985 or early 1986) is
reproduced in the following pages.

C.lR.L. Forward, "Alternate Propulsion Energy Sources," AFRPL-

TR-83-067, Final Report on Contract F04611-83-C-0013, Air
Force Rocket Propulsion Lab, Edwards, CA 93523, December 1983,
C. 2 R.L. Forward, "Antiproton annihilation propulsion," AIAA
Paper 84-1482, AIAA/SAE/ASME 20th Joint Prop. Canf.,
C -cinnati, Ohio (11-13 June 1984).
C. 3 R.L. Forward, "Antiproton propulsion," IAF Paper 84-318,
35th Congress of the Int. Astronaut. Fed., Lausanne,
Switzerland (7-13 October 1984)

,°4.

I,. C-I

V.-
 ANTIPROTON ANNIHILATION PROPULSION

Robert L. Forward*

Abstract

Antimatter represents a highly concentrated form of energy

storage since the antimatter converts all of its mass to
energy upon annihilation with normal matter. The antimatter
should be in the form of antiprotons since, unlike positrons
or antielectrons, the antiproton does not convert into gamma
rays upon annihilation, but instead two-thirds of the energy
is emitted as charged particles (pions) whose kinetic energy
can be converted into thrust by interaction with a magnetic
field nozzle or a working fluid. Antiprotons are already
being generated, captured, cooled, and stored at a number of
particle physics laboratories around the world, albeit in
small quantities. A number of techniques for the efficient
generation, long-term storage, and effective utilization of
milligram quantities of antiprotons for space propulsion are
discussed.

* Associate Fellow, AIAA

C-2
 S~Introduction

In this paper I discuss a new high specific impulse, high

thrust propulsion system based on the generation, storage, and
utilization of antiprotons. It has long been realized that
antimatter would be a valuable propulsion energy source
because it allows for the complete conversion of masq to
energy. Early studies of the concept by S~ngerI assumed that
the antimatter would be antielectrons (positrons), which
interact with electrons to produce 0.511 MeV gamma rays.
S~nger unsuccessfully tried to invent electron-gas mirrors to
direct these short wavelength gamma rays to produce a photon
i• rocket.
Th,. antiproton is much more suitable than the antielectron for
iI propulsion systems. The annihilation of an antiproton by a
•" proton (or neutron) does not produce gamama rays immediately.
.. ~Instead the products of th-e--anrihilation are from three to
Sseven pions. On the average thiere are 3.2 charged pions and
- 1.6 neutral pions.2 The neutral pions have a lifetime of only
S~90 attoseconds and almost immediately convert into two high-
S~energy (200 MeV) gamma rays. The charged pions have a normal
S~half-life of 26 nanoseconds, but because they are moving at
S~94% the speed of light, their lives are lengthened to
"• k-• 70 nanoseconds. Thus, they travel an average of 21 meters
m qL•before they decay. This time and interaction length is easily
. long enough to collect the charged pions in a thrust chamber
• constructed of magnetic fields and direct the isotropic
• microexplosion into a unidirectional flow. Even after the
charged pions decay, they decay into energetic charged muons,
which have even longer lifetimes and interaction lengths for
further conversion into thrust. Thus, if sufficient
quantities of antiprotons could be made, captured, and stored,
then present known physical principles show that they can be
used as a highly efficient propulsion fuel.
Because of the extreme difficulty in obtaining significant
quantities of antimatter, the idea of an antimatter rocket has
usually remained in the "science fiction* category. Any
papers before 1980 [see 27 references in section 02.01 of
bibliography by Mallove, et al.31 were usually concerned with
interstellar missions and glossed over the problems of
generating, storing, and using the antimatter. Recent
progress in particle physics on maethods for obtaining intense
antiproton beams, however, have caused those in the space
i• propulsion community to take another look at the concept of
S~antimatter propulsion to see if the concept can be removed
": ~from the "science fiction" category to the "technically
"" ~difficult and very costly" category, at which point the
S.military services or NASA could begin considering its use.

S~C-3

.-
 The last five years have seen the aresentation of a number of
papers on antimatter propulsion,4-8 including c special issue
of the Journal of the British Interplanetary
9 4
Society on the
subject of antimatter propulsion. -
The problems to be solved in making antiproton annihilation
propulsion feasible can be listed as:
Antiproton Generation
Antiproton Capture
Cooling at Relativistic Velocities
Deceleration from Relativistic to Subrelativistic Velocities
Cooling and Slowing at Subrelativistic Velocities
Conversion of Antiproton Beam to Antihydrogen Beam
Cooling and Slowing of Antihydrogen Beam
Conversion of Antihydrogen Atoms to Antihydrogen Molecules
Cooling and Slowing of Molecular Antihydrogen Beam
Stopping of Antihydrogen Molecules
Trapping and Cooling of Antihydrogen Molecules
Conversion of Antihydrogen Gas to Antihydrogen Ice
Long Term Storage of Antihydrogen Ice
Extraction of Antihydrogen from Storage
Annihilation of Antihydrogen
Transfer of Annihilation Energy to Working Fluid
* Conversion of Working Fluid Energy to Thrust
Solutions to some of these problems, such as generation,
capture, relativistic cooling, deceleration, and
subrelativistic cooling have already been demonstrated. I can
see solutions to most of the rest of the problems, although
not all of them. In the remainder of this paper we will see
what is the present state of the art, what are the problems
yet to be solved, and how one might approach a solution to
those problems.

Present Production Facilities

Antimatter in the form of antiprotons is being made and stored
today, albeit in small quantities. The two major producers
are the Institute for High Energy Physics (IHEP) in the USSR15
and the Centre Europ~enne pour la Recherche Nucl~aire (CERN)
in Europe. 1 6 Fermilab in the US has started construction of
Stheir ntiproton facility and expects to be in operation in
1985.17 In these facilities, the antiprotons are generated by
sending a high-energy beam of protons into a metal target.
When the relativistic protons strike the dense metal nuclei,
their kinetic energy, which is many times their rest-mass
energy, is converted into a spray of particles, some of which
are antiprotons. A magnetic field focuser and selector
separates the antiprotons from the resulting debris and
directs it to a storage ring.

C-4
 When the antiprotons are generated, they have a wide spread of

energies. This makes it difficult to decelerate them to

subrelativistic velocities, so it is necessary to "cool" the
beam so that all the antiprotons have the same energy. Two
techniques for reducing the velocity spread have been
successfully demonstrated. In the stochastic cooling
scheme, 1 8 the radio noise generated by fluctuations in the
beam are detected. This noise is amplified, phase shifted,
then transmitted across the diameter of the ring to an
electromagnetic kicker that luppresses the fluctuation. In
the electron cooling scheme• a beam of monoenergetic
electrons is inserted in the ring with the antiprotons. Those
antiprotons moving too slowly will be accelerated by
"* electromagnetic interactions with the negative charge on the
electrons and those moving too fast will be decelerated.
These cooled antiprotons could then go through another stage
of deceleration and cooling to bring them down to speeds
suitable for capture, control, and cooling by other
techniques. The accelerator at CERN generates 3.5 GeV
antiprotons using a 26 GeV proton beam and has stored as many
as 1012 antiprotons for days at a time in their magnetic ring
"racetrack" antiproton accumulator.20

To give some scale as to what has already been accomplished at

" these research facilities, 1012 antiprotons have a mass of
1.7 picograms. When this amount of antimatter is annihilated
with an equivalent amount of normal matter, it will release
300 joules, a significant quantity of energy from an
engineering viewpoint. To obtain this "firecracker" amount of
annihilation energy required the use of multimillion dr'lar
machines that used an enormous amount of electric energy. Yet
it is important to recognize that scientists working in basic
physics, using research tools not designed for the job, have
produced and continue to produce significant quantities of
annihilation energy.
Present Production Rates
The capture efficiencies of the present antiproton facilities
are abysmally low. The situation is summarized by Figure C-1
from a recent paper. 2 1 The upper part of the figure shows the
total number of antiprotons generated per GeV of antiproton
momentum per steradian of solid angle at the central portion
of the antiproton beam. Integrating the curve over the
antiproton momenta shows that each proton produces
7.7 antiprotons per steradian. The number of antiprotons per
GeV of antiproton momentum for two different angular
acceptances is shown in the lower two curves. In the paper,
the number of antiprotons per GeV of antiproton momentum is
"estimated assuming that first antiproton collector at Fermilab
can only accept those antiprotons with an angular spread off
the axis of 30 mrad (0.0028 steradians).

C-5
 Ii -- I I' ~ - ~-- - -

d3 N NUMBER OF ANTIPROTONS PER 120 GeV PROTON (5/p)

dPd 12
|10.1 -N =7.7 i/1p" STERADIAN) _

10-2

-. 0. N 60 mrad..
(0-0-1 star.) -•.
0
1 -N

dP N = 3.014 (00str.)

10o- 4 10" •P = 0.25 GeV

Nth = 1.8x 10- 4 /p

(Nexp = 3.0 x 10-50/p)

lj- 5
9 5 10 15 20 25 30
ANTIPROTON MOMENTUM P, GeVIc

Fig. C-1 Prevent antiproton capture efficiencies

"C-6
'A.
 When the 30 mrad curve is integrated over the antiproton
momenta we find a total of only 0.014 antiprotons per proton
in this narrow angular acceptance. Then, of this small
angular spread the Fermilab collector is able to capture only
those with a momentum (velocity) spread of 3% or 0.25 GeV
around 8.9 GeV. Thus, ideally, they would expect to capture
about 1.8x10- 4 antiprotons per proton, with an estimated
actual capture rate (including mismatch and transport losses)
of 3x10- 5 antiprotons per proton. If we compare the
annihilation energy we get from using the antiproton (2mnc2=
1.87 GeV) with the energy in the 120 GeV protons requiredao
make that antiproton, we get an energy efficiency of 5x10-'.
Since a typical synchrotron is only about 5% efficient, the
"wallplug" energy efficiency for antiproton production of
present machines is only about 2x10- 8 .

Future Production Rates

Data on the antiproton production spectrum of high-energy
protons impacting heavy metal targets are available only for
small angles about the forward direction. These data are
sufficient for the design of the present antiproton collector
systems that only attempt to capture the antiprotons emitted
around the forward peak. To design systems that will capture
a higher percentage of the aatiDrotons, it will be necessary
to know the antiproton spectrum a4 a function of angle and
incident proton energy over a greater angular spread. Such
data do not seem to exist and there are no present plans to
make these measurements since obtaining the data would require
an extensive amount of time on the large synchrotron machines.
The particle physics conmmunity prefers to use the machine time
to study issues more important to narticle physics. As a
result of this lack of detailed knowldie of the spectrum, the
total number of antiprotons generated is also unknown (to
probably a factor of two).
The last collection of experimental data on total antiproton
production rates was done over a decade ago and published in a
review paper by Antinucci, et al. 2 2 The measurements were
*. made using colliding beams of protons, so the data are only
partially relevant to the problem of colliding protons with
heavy nuclei, which is known to give a higher antiproton
production rate. The data from the table in the Antinucci
paper for the total antiproton production rate are the large
dots in Figure C-2.

C-7
 1 I I 111111| I 'tI I 111111 I I I 1 1 111

ESTIMATED
TOTAL p--Pb

"0 MEASURED p-p

0• ENERGY /
10-1 OPTIMUM

0.

2r
I.-

,;., 10-

)0 I--
Using the known ratio of antiproton production in the forward
"direction from heavy nuclei and hydrogen targets, 2 1 I was able
to modify the Antinucci hydrogen target data to obtain the
upper curve which gives the predicted antiproton production
rates as a function of energy for protons incident on metal
targets.
If we now take the upper curve giving the number efficiency
for producing antiprotons and divide it by the energy of the
proton making the antiprotons, we obtain the bottom curve.
This is the energy efficiency for producing antiprotons. Note
that it has a broad peak around 200 GeV. Although the number
of antiprotons produced continues to increase as the incident
proton energy is increased, above 200 GeV the gain in
production is not enough to offset the increased proton energy
required.

From Figure C-2 we see that the maximum energy efficiency

production rate occurs for an incident proton energy of
200 GeV and is 0.085 antiprotons/proton. (There are roughly
5 K mesons, 50 pi mesons, and large numbers of positrons and
electrons produced for each antiproton generated.) This
antiproton production rate is 2 times the production at the
Fermilab energy of 120 GeV and 20 times the production at the
CERN energy of 26 GeV. It should be emphasized that the
curves in Figure C-2 are based on sparse data and that actual
measurements of antiproton production spectra as a function of
angle and proton energy are needed before any major
engineering studies on antiproton production are done.

Antiproton Factory
Figure C-3 shows a conceptual design for an antiproton factory
which would utilize the technologies being developed at CERN,
Fermilab, and IHEP, but on a much larger scale and with the
design optimized for energy efficiency. First, the proton
accelerator should be a high current rf linear accelerator
(linac) with a wallplug efficiency of 50%, rather than the low
current, low efficiency, but high energy resolution
synchrotron preferred as a research tool by particle
physicists. There would be more than one proton beam with
each beam operated at the optimum beam current for the
particular target design chosen. Each proton beam would
strike a metal target and the resulting particles would be
sorted by an array of wide-angle collecting lenses to extract
the antiprotons and positrons. The positrons with the right
energy would be picked off and sent to the antihydrogen
generator, while all the antiprotons possible would be sorted
by energy and sent to a stack of stochastic coolers, each
optimized for a particular central antiproton momentum.

C-9

-A. %
 200~
GeV PROT BýA N A0 e

e+ V-Ago •+ COLLECTOR
DECELERATOR ARRAY
ANTIHYPROGEN •
LAE " *
BEAM TO ENHANCED -- POSITRON

CntLER- ANTIHYDROGEN BYPASS ENERGY "

AND COSTACK
STOCHASTI 5- 30 GoVlc

_
CO O 0 M•AFP - 5 Go i BEAMS
CL 20 EACH
mlvSUBRELATIVISTIC
COOLING RING

200 M~eV STC REOLI

SUBRELAWnnSTIC RELATWISTIC
pBEAM BEArr.MS

Fig. C-3 _Antlproton factory (one segment)

C-10
S0"•. After stochastic cooling, the stack of beams at different
energies would go to a decelerator stack that would reduce all
the antiproton energies to the same subrelativistic energy
(200 MeV). The combined beam would then be sent to a
subrelativistic cooling ring using either stochastic or
electron cooling before being further decelerated and sent on
to the antihydrogen generator where the antiprotons are
combined with the positrons to make antihydrogen atoms.

Antihydrogen
The antihydrogen generator would follow the general concepts
described in a recent research publication at CERN. 2 3 As
shown in Figure C-4, if a beam of positrons were traveling
along with a beam of antiprotons at the same speed, they would
attract one another and recombine to form antihydrogen. This
natural process can be enhanced by factors of 100 or more by
stimulating the capture process with photons at the right
wavelength.
Once an antihydrogen beam has been formed, there are a number
of techniques available for cooling the electrically neutral
antihydrogen down, slowing it to a stop, and storing it in a
trap. Traps for atoms were first proposed by Letokov 2 4 and
Ashkin. 2 5 These traps use laser beams tuned just below the
first optical resonance line of the atom. Those atoms trying
to move toward the laser will see the laser photons shifted
upward into resonance with the optical absorption line. The
atoms will absorb the Doppler-shifted laser photons, slowing
down slightly in the process. The atom then reradiates each
photon, but in a random direction, so the recoils from the
reradiated photons will average out. Thus, after many
absorptions and reradiations, the atom has stopped moving.
Once the atom is stationary, it no longer absorbs the off-
resonant laser photons and stays trapped.

Lasers have also been used to "push" a beam of sodium atoms to

one side, "cool" the beam both longitudinally and transversely
until all the atoms have the same speed, and slow down, halt,
and reverse the direction of an atomic beam. The activities
in the field of cooling and trapping atoms has progressed to
the point whe5 g there are periodic workshops on laser cooling
and trapping.

C-11
 -. p
ANTIPROTON RING

BENDING
MAGNETS 1

S • INTERACTION REGION

PHOTON - 4-_ + MIRROR

LENS LENS
SEPARATOR COMBINER
MAGNET FOCUSED MAGNET "

POSITRON RING
-4O. ....

MIRRON ZL RUFTE / MIRROR

LASER RING

Fig. C-4 Laser-aided antihydrogen formation

C-12

*V1

* *~ ; 1 ;v .~INK
p Although it might be possible to store antihydrogen as an
atomic gas, 2 7 the atomic form of antihydrogen is more
difficult to control, cool, and trap than sodium since the
first resonance line in atomic hydrogen is in the vacuum
ultraviolet (the Lyman alpha line). The fundamental problem
is that while one Lyman alpha photon will excite an
antihydrogen atom, if a second photon arrives before the atom
has decayed back into its ground state, the second photon may
ionize the antihydrogen atom. Although proprietary ideas
exist for overcoming these problems, it will likely be found
necessary to convert the antihydrogen atoms into antihydrogen
molecules, then store them as antihydrogen ice.
The conversion of antihydrogen atoms to antihydrogen molecules
takes place naturally (with the release of lots of energy,
which is why spin-polarized normal hydrogen is being looked at
as a potential rocket fuel). A large number of the molecules
remain in a metastable orthohydrogen state. Left to
themselves, cold antihydrogen molecules will ultimately all
convert to parahydrogen, the ground state of the molecule, but
unless a catalyst is used, the process takes many days.
Research is needed on the use of lasers and magnetic fields
with high gradients to convert the antihydrogen atoms into
antihydrogen molecules. These antihydrogen molecules can then
be further cooled and trapped using lasers operating on a
molecular hydrogen line, then turned into antihydrogen ice in
"the preferred parahydrogen state. Research is also needed on
turning a cold antihydrogen vapor into ice crystals, since
there is a heat of fusion generated during the formation of
the ice. Fortunately, all of these research problems on
manipulation of antihydrogen can be studied using normal
hydrogen (and would make excellent thesis topics).

Antihydrogen Traps
Antihydrogen ice, like hydrogen ice, is diamagnetic, with a
negative magnetic susceptibility that is two-thirds that of
graphite. A simple passive trap for a ball of antihydrogen
ice could be made of magnetic fields. There are a number of
different ways to configure permanent magnets and coils to
produce a magnetic field minimum that will attract and trap a
diamagnetic material such as graphite 2 8 or hydrogen. One
simple example consists of two superconducting coils spaced so
that there is a magnetic minimum midway between them. 2 9 This
kind of trap would be completely stable and require no power.
It is not very deep, however, and although quite suitable for
storage of antihydrogen ice in free fall, it might not be able
to levitate the antihydrogen ice at high acceleration levels.

C-13
 For high acceleration levels, a more suitable trap would be a
servocontrolled dc voltage alectrostatic levitation mechanism
such as those made at JPL. 3 U These traps have levitated
electrically charged 20 mg millimeter sized spheres of water
ice in the earth's field. Antihydrogen ice will have a
density of 0.0763 g/cm3 , which is J.3 times less than water
ice. Thus, the same electrostatic suspension could hold
milligram-sized balls of antihydrogen ice at accelerations up
to 13 gees. Since the antihydrogen ice will be formed at
millidegrees or below, and the heat input from the electric
levitator will be low, the sublimation pressure of the
antihydrogen will be so low (10-" torr at 1 K) that the
antihydrogen ice ball should last for years.

Utilizing Antihydrogen For Propulsion

- There are a number of techniques for extracting the
antihydrogen from the storage trap and directing it into the
rocket engine under control. If the antihydrogen is in the
form of a large ball many milligrams in size, the antiprotons
can be extracted from the ice ball by irradiating the ice with
ultraviolet, driving off the positrons, extracting the excess
antiprotons by field emission with a high intensity electric
field, then directing them to the thrust chamber. 1 ' It might
be more desirable if the antihydrogen could be formed as a
cloud of charged microcrystals, each a microgram and
containing the energy equivalent of 20 kg of chemical fuel.
Then, using a directed beam of ultraviolet light to drive off
a few more positrons, an individual microcrystal could be made
more highly charged, preferentially extracted from the
microcrystal cloud using electric fields, ane directed down a
vacuum line to the thrust chamber. Since the position of the
charged microcrystal in the injection line can be sensed,
mechanical shutters can allow the passage of the microcrystal
without breaking vacuum.
Antimatter fuel is so powerful that new types of rocket
engines will have to be developed to fully utilize its
potential. One of the simplest antiproton propulsion systems
would use a design similar to that of a nuclear thermal
rocket. In a nuclear thermal rocket hydrogen gas is heated by
passing it through the core of a fission reactor. The hot
hydrogen is then used to provide thrust. In the antiproton
annihilation version, the energy released by the annihilation
reaction would be absorbed in the walls of a heat exchanger
made out of refractory metal. The heat exchanger would then
heat hydrogen to produce thrust. 3 1 A heat exchanger made out
of a cylinder of tungsten 28 cm in diameter and 28 cm long
would only weigh 330 kg and would capture most of the energy
in the gamma rays and pions emitted by the antiproton-proton
annihilation process, thus utilizing all of the energy in the

C-14
 anitihilation reaction. The maximum temperature would be
limited by the melting point of tungsten to about 3000 K,
resulting in a maximum specific impulse of about 900 sec or an
exhaust velocity of about 9 km/s. This is considerably better
than any chemical rocket or even a nuclear fission thermal
rocket, but still does not use the high exhaust velocity
potential of antiproton annihilation.
The plasma created from the heating of the hydrogen working
fluid by the pions emitted from the annihilation process is
too hot to be contained and directed by thrust chambers and
nozzles made of solid material. Fortunately, most of the
particles generated are charged and can be contained and
directed by strong magn.etic fields. The first example of a
design for a magnetic field ajiproton rocket engine can be
found in the paper by Morgan.

Minimum Antimatter Optimization

When antiprotons interact with protons (hydrogen), the
resultant annihilation products are pions with an average
kinetic energy of 250 MeV. This translates into an exhaust
velocity of 94% of the speed of light. Thus, pure antimatter
rockets are best suited for relativistic missions. In order
to use the minimum amount of antimatter for the mission, the
best way to use the antimatter is not to use equal amounts of
matter and antimatter. Instead, the antimatter should be used
"to heat a much larger amount of propellant. It has been
"shown, 3 2 that except for extreme relativistic spacecraft
speeds (>0.5 c), the reaction mass needed is always four times
the spacecraft payload mass, or an overall ratio of launch
mass to payload mass of 5:1. The mass of the antimatter
needed increases as the square of the mission total velocity
change, but is always a negligible fraction of the total mass.
Because the mass ratio of an antimatter powered space vehicle
will always be less than 5:1 (typically 2-3:1), mission
analysts need to rethink those mission that have been labeled
"impossible" because of the extreme mass ratios required to
accomplish the mission using a chemical or nuclear system with
a fixed specific impulse.

Antimatter Powered Mission Analyses

In some preliminary studies of an antihydrogen/hydrogen
rocket, Cassenti has estimated some of the parameters in an
antimatter powered orbit transfer mission. The mission was to
take a 10 ton spacecraft from LEO to GEO back to LEO (using
*. aeroassist). 6 The mission velocity change was assumed to be
5.5 km/sec. Using the minimum antimatter optimization,
Cassenti found that the optimum exhaust velocity was

"C-15
3.4 km/sec (specific impulse of only 350 sec), the reaction
mass required was 40 tons, and the amount of antihydrogen
needed was only 6 mg. If the amount of antihydrogen used is
raised from 6 mg to 10 mg, the amount of hydrogen reaction
mass drops dramatically, from 40 tons to 15 tons, giving a
mass ratio of 2.5:1, while the exhaust velocity rose to
5 km/sec. Thus, in this range of the parameters, an
additional 4 mg of antihydrogen saves 25 tons of reaction
mass. Whether this trade-off is worth it depends upon the
relative cost of antihydrogen per milligram compared to the
cost of hydrogen per ton in LEO. In a recently completed
study 3 3 it was estimated that a well-designed factory for
producing antihydrogen should be able to operate at an energy
efficiency of better than 10-4 (compared to the present
efficiency of 2x10- 8 ). The cost of the antimatter was
estimated to be about $10M per milligram, while reaction mass
in LEO was estimated to cost $5M per ton. Thus, using the
numbers from the Cassenti study, an additional 4 milligrams
($40M) of antimatter fuel in the rocket saved 25 tons ($125M)
of reaction mass. Although these cost estimates are far from
firm, it looks as though antimatter might be a cost-effective
fuel for space propulsion.

Conclusions
Our major conclusion about antiproton propulsion is that the
concept is feasible but difficult and expensive. Yet, despite
the high cost of antimatter, it may be a cost effective fuel
in space where any fuel is expensive. There is high risk in
the development of antiproton propulsion. The major
uncertainties seem to be in the production and capture of the
antiprotons at high efficiency, and the conversion of
antiprotons into frozen antihydrogen without excessive losses.
The storage problems look tractable. The problems that need
working on first are to determine the total antiproton
production rate and spectrum versus proton energy, the maximum
feasible limits to antiproton capture efficiencies of
physically feasible lenses and accumulator rings, and the
maximum efficiency of the antimatter rocket that uses the
antiproton fuel. It is important to recognize that many of
the problems of capturing, cooling, slowing, trapping, and
storing of antiprotons (antihydrogen) can be done as thesis
topics using normal protons and hydrogen.

C-16
 Acknowledgments

This research was supported in part by the Air Force Rocket

Propulsion Laboratory through contract F04611-83-C-0013 with
Forward Unlimited and contract F04611-83-C-0046 with the
Research Institute, University of Dayton, and in part by the
independent research and development program of the Hughes
Aircraft Company.

References

1 SAnger, E., "The Theory of Photon Pockets," Ing. Arch., Vol.

21, 1953, pp. 213 ff.
2 Agnew, L.E., et al., "Antiproton Interactions in Hydrogen and
Carbon Below 200 MeV," Physical Review, Vol. 118, 1960, pp.
1371 ff.
3 Mallove,
E.F., Forward, R.L., Paprotny, Z., and Lehmann, J.,
"Interstellar Travel and Communication: A Bibliography,"
Journal of the British Interplanetary Society, Vol. 33, June
1980, pp. 201-248 (entire issue).
4 Forward,
R.L., "Interstellar Flight Systems," AIAA Paper 80-
0823, AIAA International Meeting, Baltimore, Md., May 1980,
pp. 6-8.
5 Cassenti,
B.N., "Antimatter Propulsion for OTV Applications,"
AIAA Paper 84-1485, 20th Joint Propulsion Conference,
Cincinnati, Ohio, June 1984.
6 Cassenti,
B.N., "Optimization of Relativistic Antimatter
Rockets," Journal of the British Interplanetary Society, Vol.
37, November 1984, pp. 483-490.
7 Vulpetti,G., "A Propulsion-Oriented Synthesis of the
Antiproton-Nucleon Annihilation Experimental Results," Journal
of the British Interplanetary Society, Vol. 37, March 1984,
pp. 124-134.
8 Vulpetti,
G., "An Approach to the Modeling of Matter-
Antimatter Propulsion Systems," Journal of the British
Interplanetary Society, Vol. 37, September 1984, pp. 403-409.
9 Massier,P.F., "The Need for Expanded Exploration of Matter-
Antimatter Annihilation for Propulsion Application," Journal
of the British Interplanetary Society, Vol. 35, September
* 1982, pp. 387-390.

C-17

o- *, '~
 1 0 Forward,
R.L., "Antimatter PropulsiDn," Journal of the
British Interplanetary Society, Vol. 35, September 1982, pp.
391-395.
1 lCassenti,B.N., "Design Considerations for Relativistic
Antimatter Rockets," Journal of the British Interplanetary
Society, Vol. 35, September 1982, pp. 396-404.
1 2 Morgan, D.L., "Concepts for the Design of an Antimatter
Annihilation Rocket," Journal of the British Interplanetary
Society, Vol. 35, September 1982, pp. 405-412.
1 3 Zito, R.R., "The Cryogenic Confinement of Antiprotons for

Space Propulsion Systems," Journal of the British

Interplanetary Society, Vol. 35, September 1982, pp. 414-421.
1 4 Chapline,
G., "Antimatter Breeders?" Journal of the British
Interplanetary Society, Vol. 35, September 1982, pp. 423-424.
1 5 Vsevolozskaya,T.A., et al., "Antiproton Source for the
Accelerator-Storage Complex, UNK-IHEP," Fermilab Report FN-353
8000.00, June 1981, a translation of INP Preprint 80-182,
December 1980.
1 6 The
Proton Synchrotron Staff, "The CERN PS Complex as an
Antiproton Source," IEEE Transactions on Nuclear Science, Vol.
NS-30, August 1983, pp. 2039-2041.
1 7 Peoples,
J., "The Fermilab Antiproton Source," IEEE
Transactions on Nuclear Science, Vol. NS-30, August 1983, pp.
1970-1975.
1 8 Van
der Meer, S., "Stochastic Cooling in the CERN Antiproton
Accumulator," IEEE Transactions on Nuclear Science, Vol. NS-
28, 1981, pp. 1994-8.
1 9 Bell,
M., et al. "Electron Cooling in ICE at CERN," Nuclear
* Instrumentation and Methods, Vol. 190, 1981, pp. 237-255.
2 0 Cline,
D.B., Rubbia, C., and Van der Meer, S., "The Search
for Intermediate Vector Bosons," Scientific American, Vol.
247, No. 3, March 1982, pp. 48-59.
2 1 Hojvat,
C., and Van Ginneken, A., "Calculation of Antiproton
Yields for the Fermilab Antiproton Source," Nuclear
Instrumentation and Methods, Vol. 206, 1983, pp. 67-83.
2 2 Antinucci,
M., et al., "Multiplicities of Charged Particles
up to ISR Energies," Letters Nuovo Cimento, Vol. 6, 1972, pp.
121-127.

C-18
 2 3 Neumann,
R., Poth, H., Winnacker, A. and Wolf, A., "Laser-
Enhanced Electron-Ion Capture and Antihydrogen Formation,"
Zeitschrift far Physik A, Vol. 313, 1983, pp. 253-262.
2 4 Letokhov,
V.S. and Minogin, V.G., "Laser Radiation Pressure
on Free Atoms," Physics Reports, Vol. 73, 1981, pp. 1-65.
2 5 Ashkin,
A. and Gordon, J.P., "Cooling and Trapping of Atoms
by Resonance Radiation Pressure," Optics Letters, Vol. 4,
1979, pp. 161-163.
2 6 Phillips, W.D. (editor), Laser-Cooled and Trapped Atoms, NBS
Special Publication 653, Proceedings of Workshop on
Spectroscopic Applications of Slow Atomic Beams, NBS
Gaithersburg, MD, April 1983.
2 7 Cline, R.W., et al, "Magnetic Confinement of Spin-Polarized
Atomic Hydrogen," Physical Review Letters, Vol. 45, 1980, pp.
2117-2120.
2 8 Waldron,
R.D., "Diamagnetic Levitation Using Pyrolytic
Graphite," Reviews of Scientific Instruments, Vol. 37, 1966,
pp. 29-34.
2 9 Letokhov,
V.S. and Minogin, V.G., "Possibility of
Accumulation and Storage of Cold Atoms in Magnetic Traps,"
Optics Communications, Vol. 35, 1980, pp. 199-202.
3 0 Rhim,
W.-K., Saffren, M.M., and Elleman, D.D., "Development
of Electrostatic Levitator at JPL," Materials Processing in
the Reduced Gravity Environment of Space, G.E. Rindone,
Editor, Elsevier Science, 1982, pp. 115-119.
3 1 Augenstein,
B.W., "Some Examples of Propulsion Applications
Using Antimatter," Rand Paper P-7133, The Rand Corp., Santa
Monica, CA 90406 (July 1985).
°32
3 2 Shepherd,
L.R., "Interstellar Flight," Journal of the
British Interplanetary Society, Vol. 11, July 1952, pp. 149-
167.
33Forward, R.L., "Alternate Propulsion Energy Sources," AFRPL-
"TR-83-067, Final Report on Contract F04611-83-C-0013, Air
Force Rocket Propulsion Lab, Edwards, CA 93523, December 1983,

C-19
L

APPENDIX D

MATERIAL ETALPY ASCENT/DESCENT (HMEA) MODULE

ABSTRACT

This novel non-conventional propulsion concept resulted from a

discussion on 3 January 1985 between the contract Principal
Investigator, Dr. Robert L. Forward and the Task Manager,
Dr. Franklin B. Mead, Jr. on the feasiblity of storing thermal
energy in the heat capacity of high temperature materials. In
the 1950's Dr. Mead had looked into the storage of heat energy
in the specific heat capacity of graphite. The amount of heat
that could be stored in graphite was significant, but the
amount of propulsion obtainable was not competitive with
liquid fuels. During the discussion, however, it was realized
that the amount of energy stored in the latent heats of the
phase changes to the liquid and vapor states of a material
could be many times greater than the energy stored in the
specific heat alone. For example, the latent heat of
vaporization of some materials can release up to four times as
much heat energy per kilogram as any chemical reaction can.
There was not enough time left on the contract to give this
concept more than a cursory evaluation. The concept looks
promising as an energy storage technique. It is recommended
that the Air Force study the concept further to determine its
technical feasiblity as an energy storage technique and the
feasibility of utilizing the stored energy to heat a working
fluid for propulsion.

D-1

. ",'
.-.h *• . " - " • " - - " - "" " - -'"::> • " - "
 MATERIAL ENTHALPY ASCENT/DESCENT (HERD) MODULE

"Dr. Franklin B. Mead, Jr. of AFRPL has observed that a great

deal of energy can be stored as heat energy in the specific
- heat capacity of refractory materials. If that heat could be
efficiently transferred to a propellant, such as hydrogen,
then the hot hydrogen could propel a rocket. Whether such a
rocket makes sense would depend upon the amount of heat
stored, the efficiency of transfer of the energy to the
propellant, and the weight of the storage material, tankage,
and insulation.
As an example, the specific heat of graphite at high
temperatures is roughly 2 J/gm.K. Graphite melts (sublimates)
at 3820 K. The amount of heat energy released by graphite as
it is cooled from 3820 to 820 K is 6 kJ/gm. This is about
half the energy released by the combustion of LOX/H 2 .
If the energy in the graphite is used to heat a working fluid
such as hydrogen, then because of its low molecular weight,
the specific impulse of the hydrogen is significant, ranging
from 1000 sec at 3820 K to 470 sec at 820 K. However, because
the specific heat of hydrogen at high temperatures is about
15 J/gm-K, it would take a number of grams of hot graphite to
heat one gram of hydrogen (depending upon the exhaust
temperature and I that we wanted). The weight of the
graphite makes its8oubtful that a rocket using a MEAD module
containing hot solid graphite could compete with a rocket
using chemical fuels.
An alternate approach is to use the latent heat of fusion or
the latent heat of vaporization of a material as the storage
mechanism for the heat energy. A review of standard chemistry
and physics handbooks led to the conclusion that the latent
heat of vaporization is much better than the latent heat of
fusion.
It was found in the handbooks that certain materials, upon
condensing from a vapor to a liquid, can release four times as
much energy as the combustion of LOX/H 2 . The four elements
that store the most energy per gram are listed below:

Element Mol. Wt. Temp. Vap. (K) Heat of Yap. (kJ/gm)

C 12.01 5100 60
B 10.81 2820 53
Be 9.01 3240 36
Li 6.94 1615 23

D-2
 Boron looks like a promising candidate for a metal vapor
' •. version of a MEAD module. The boron in the insulated MEAD
module would be preheated on the ground over a long period of
time until it turned into vapor at a temperature of 2820 K or
greater (depending upon the pressure). At this temperature it
could be contained in a pressure vessel made of graphite,
which has a melting point of 3820 K.
The MEAD modul.e and a tank of liquid hydrogen would then be
attached to a single-stage-to-orbit vehicle just before
takeoff. The hot vapor from the MEAD module would be sent to
a heat exchanger in the rocket engine of the vehicle using
rhenium tubing (similar to that used in the APRPL/Rocketdyne
solar thermal rocket). The hot boron vapor would heat the
hydrogen to about 2820 K (the same temperature expected in the
solar thermal rocket) to produce a specific impulse of about
800 sec. Each gram of boron vapor at 2820 K has enough energy
to raise a gram of hydrogen to 2820 K with 10 kJ/gm left over
"to cover losses. The condensed boron liquid would drain off,
allowing more hot boron vapor to reach the heat exchanger
tubes. After all the boron vapor has condensed, there is
still 4 to 6 kJ/gm of heat energy left in the specific heat of
the liquid boron that can be used for further propulsion at
lower specific impulse.
I was unable to find data on the critical temperature (T ),
pressure (P), and density (dc) of boron. But from the Sata
on lead, silver, and gallium, I estimate that T -6000 K,
*.. Pc- 2 5 0 atm, and d -1 gm/cc for boron. This would mean that at
a temperature of 5000 K and a pressure of 30 atm, the density
of the vap.or would be about 0.25 gm/cc. At this density,
100 tons of hot boron would fit into a tank S m in diameter by
20 m long (the size of the body of the Shuttle), while
100 tons of liquid hydrogen propellant would take up the same
room as the 100 tons flown in the present STS external tank
(8.4 m diameter by 25 m long). Different storage pressures
for the boron would give different volumes for the MEAD
module, since the boron is being stored as a gas.
It is not possible to know at this time if this concept makes
sense as a single-stage-to-orbit rocket. Since the energy
source for the rocket (the hot boron) is heavy and is kept on
board instead of being exhausted with the propellant, the
basic rocket equations to describe the situation have to be
rederived with the new assumptions. The density of boron
vapor at high temperature and pressures has to be determined.
Then the weight of the structure and insulation for the hot,
high pressure tankage in the MEAD module needs to be
estimated. The final design may turn out to be too heavy to
fly, but any concept that gives high thrust at 800 sec in a
compact package is certainly worth looking at further.

a .eU G
&O V EI NOAqM P O Pc 198 5- 67 8 94 / 20 0 3 7
40
S~D-3
%.