COMMISSION ON COLLEGE

PI

MARTIN

A.

POMERANTZ

VAN NOSTRAND REINHOLD MOMENTUM BOOKS

PUBLISHED FOR THE COMMISSION ON COLLEGE PHYSICS
WALTER
C.

GENERAL EDITOR MICHELS, Bryn Mawr College, Bryn Mawr

EDITORIAL

BOARD

Jeremy Bernstein, New York University E. U. Condon, Joint Institute for Laboratory Astrophysics
University of Colorado

Melba

Phillips, University of

Chicago

William T, Scott, University of Nevada

No.

1.

No.

2.

ELEMENTARY PARTICLES David H. Prison and Man M. RADIO EXPLORATION OF THE PLANETARY SYSTEM
Alex
G.

Tnorndike

Smith and Thomas D. Carr

of the

No.
NO.
No,

3.

THE DISCOVERY OF THE ELECTRON: The Development Atomic Concept of Electricity Dawd L. Anderson

4.
5.

WAVES AND OSCILLATIONS-/?.

CRYSTALS AND LIGHT: An

A,

Waldron
Elizabeth
A.

Introduction to Optical Crystallography

Wood

No.
No.

6.

TEMPERATURES VERY LOW AND VERY HIGH Mark W. Zemansky

POLARIZED LIGHT-W/Jliam
A. Shurc/iff

7. 8,
9.

and Stanley

S. BaJ/ard

No.
No.

STRUCTURE OF ATOMIC NUCLEI-C. Sharp Cook

AN INTRODUCTION TO THE SPECIAL THEORY OF RELATIVITY
Robert Katz

No. 10. RADIOACTIVITY

AND

ITS

MEASUREMENT

Witfrid B. Mann and S. B. Garfinkei

i.

No. 11. No. 12,

No. 13.

PLASMAS-LABORATORY AND COSMIC Forrest

INFRARED RADIATION-Jvan Simon

Boley

THE PHYSICS OF MUSICAL SOUND-Jess

RADIO EXPLORATION OF THE

J.

Josephs
A. Knight

No. 14. No. 15. No. 16.

No. 17. No. 18. No. 19.

THE FREEZING OF SUPERCOOLED LIQUIDS-Charies

SUN A/ex G. Smith MAGNETS-L W. McKeehan THE WORLD OF HIGH PRESSURE John W. Stewart
MAGNETOHYDRODYNAMICS-Noel
THE WINDS: The
C. Little

Origins and Behavior of Atmospheric Motion

George M. Hidy
No. 20.
No. 21.
No. 22.

THE DISCOVERY OF NUCLEAR FISSION Hans Graetzer and THE WAVES: The Nature of Sea Motion George M. COSMIC RAYS Martin A. Pomerantz
CRUCIAL EXPERIMENTS
IN

David
Hidy

L.

Anderson

No. 23.

MODERN PHYSICS- George L

Trigg

^3-

MARTIN
This

A.

POMERANTZ
r,

book
the

is

to be returned on or before

last

date stamped below .

y?'

22rOCT.' 1973

16.

MAY 1999

4^
tCMM
-a EL
1988 1988

^0 km m
15.

MAY 1991

8.

|i

26. OCT. 1590

539.7223 POM
A/C 036542
30107

000 596 376

Van Nostrand Rein hold Company Regional

Offices:

New

York, Cincinnati, Chicago, MiUhrae, Dallas

Van Nostrand Rein hold Company International London Toronto, Melbourne

j

Offices:

Copyright

1971

by Liuon Educational Publishing, Inc.

Library of Congress Catalog Card No. 71-94017

No part of this work covered by the copyright hereon may be reproduced or used in any form or by any means graphic, electronic, or mechanical, including photocopying, recording, without taping, or information storage and retrieval systems written permission of the publisher. Manufactured in the United States of America.
All rights reserved.

Published by Van Nostrand Reinhold Company 450 West 33rd Street, New York, N.Y. 10001 Published simultaneously
in

Canada by

Van Nostrand Reinhold

Ltd,

10

987654321
HARRIS

COLLEGE

PRESTON

Tll-llll^

Pen

Preface
The primary aim of this book is to convey to readers with a wide diversity of interests and backgrounds a sense of the excitement that pervades cosmic-ray research, as well as some feeling for the broad range of fundamental physics that falls within its
purview. In addition to the intrinsic intellectual appeal of the
stuff of which the whole universe have implications in many disciplines, some

ubiquitous

is

far

made, cosmic rays beyond the con-

fines of physical science.

have tried to achieve a level that is suitable for, but not to, an undergraduate whose only prior contact with physics was the beginning course. Thus, on the one hand, although this volume is not intended as a text for a graduate course, it might well serve as its syllabus. On the other hand,
I

limited

although some aptitude in physics (in addition to considerable is required for thoroughly digesting the contents, I have tried to organize the material in such a manner that, hopepatience)
fully,

some appreciation of

aesthetic values

may be gleamed

in

this

seemingly esoteric pursuit.

retical details.

is on ideas rather than experimental and theoIn the interest of producing a nonencyclopedic narrative, individuals are named only in the process of setting the stage by sketching some of the pioneering steps that ushered in new ways of looking at things. I beg my colleagues' under-

The emphasis

standing and forgiveness for several sins of commission and omis-

was necessary to exercise some poetic licence in communicating concepts. Furthermore, I have perforce been somewhat cavalier in presenting results without due acknowledgment. Finally, large areas of research, in which exceedingly elegant work is being carried out, have barely been mentioned. I apolosion. First, it

iv

PREFACE

somewhat egocentric approach, but it seems to be book on cosmic rays to be like the Bostonian's map of the United States. I have enjoyed the good fortune of being associated with two
gize for the

inevitable for a

generations of cosmic ray physicists. Together with just a few

others

who

are

still

active workers in the field, I experienced the

rare privilege of

knowing personally

practically all of the pioneers

And, thanks to the welcome movement toward international cooperation, and the
in cosmic rays, including their discoverer.

shrinking
are

size

of the world,

it

is

now

possible to count

one's friends a very large

number

of scientists in

many

lands

among who

drawn together by
individuals

the variegated facets of this fascinating

subject.

To many
list,

express

my

and institutions much too numerous deep gratitude for their help.

to

MARTIN

A.

POMERANTZ

Table of Contents
Preface

The Heroic Age

of Cosmic Ray Exploration The Dawn of Cosmic Ray Research, 2; The Radiation Era, 5; The Corpuscular Era, 9; The Modern Era, 10

Collisions of Cosmic Rays with Matter Electromagnetic Interactions, 21; Charged Particle Energy Losses, 28; Photon Absorption, 37; Nuclear Interactions, 39

15

Experimental Methods
Electrical Detectors, 43; Laboratories, 61

42
Visual Detectors, 50; Systems, 55;

Propagation Through the Earth's Magnetic Field and

Atmosphere
Magnetic Bending, 67; Transformations 80; The Atmosphere as a Black Box, 86
in the

65
Atmosphere,

Galactic Cosmic Rays Energy Spectrum, 93; Composition, 96; Spatial Distribution, 98; Prehistory, 99

90

Cosmic Ray Intensity Variations The Sun and the Interplanetary Medium, and Anisotropics, 113 Solar Cosmic Rays
Morphology

103
106; Modulations

132

of a Solar Particle Event, 135; Energy Spectrum, 141; Composition, 144; Propagation, 147; Acceleration Processes, 153; Space Radiation Doses, 154

The

Origin of Cosmic Rays

158

Acceleration Mechanisms, 160; Sources, 165; Origin Models, 169

Our

Galaxy, 168;

Appendix
Bibliography
'

176 177 179

Symbols, Abbreviations, Typical Units, and Values of Various Quantities

Index

182

To Molly

The Heroic Age of

Cosmic
The beginning
is

Ray Exploration
the most important part of the work.

PLATO

Cosmic rays have been pelting the earth incessantly for countEvery second, some ten or twenty of the progeny of these subatomic visitors from afar strike the body of each of us. But man first became aware of their presence only at the beginning of this century. Then, as has happened so frequently in the history of science, it was noted that certain experimental apparatus behaved in an unexpected manner.
less eons.

Inquisitive minds, not content with the "obvious" explana-

which interfered with the measurements that were then being carried out in a number of laboratories, ultimately established the existence of a new fundamental phenomenon. But even then, many decades were to pass before the broad implications of this discovery became comprehensible. A considerable body of knowledge about the once mysterious cosmic radiation has now been amassed, and this esoteric and still romantic subject has attained prominence even among the general public. This is attributable to several factors, not the least of which is the preeminent role played by cosmic ray physicists in the space age. By virtue of their philosophical and experimental heritage, they have been in the vanguard/ making sometimes spectacular and usually exciting discoveries with equipment carried by spacecraft to the farthest reaches of the earth's atmosphere and far beyond. To provide a frame of reference in which one can better appreciate the degree of sophistition of this disturbing effect

COSMIC RAYS
its

cation of present day cosmic ray research, let us trace

history

early

from

its

very

humble beginnings.

THE DAWN OF COSMIC RAY RESEARCH

At the turn
of the century, studies

on x

rays, discovered in
first

1895 by Wilhelm Conrad Rontgen, and radioactivity,

ob-

served by Henri Becquerel the following year, were in vogue.

These experiments generally involved use of the ionization chamber, a vessel containing gas which becomes conducting when exposed to radiations endowed with sufficient energy to detach some of the atomic electrons. The resulting current, arising from the motion of the free positive and negative charges, or ion pairs, was measured with an electrometer. The most primitive electrical instrument that can be used as
a radiation sensor
is

the familiar gold-leaf electroscope. This

its

simple device indicates, by

rate of discharge, the extent to

air
is

which the conductivity of the surrounding

presence of ionizing radiation.
ognized, electroscopes

modified by the

Long

before radioactivity was rec-

a perfect insulator. Indeed, free air always

had revealed that the atmosphere is not seemed to contain

something

like

500 to 1000 ions per cubic centimeter, correspond-

ing to a rate of formation of roughly 10 ion pairs per cubic centimeter per second. After Becquerel's discovery it was quite natural to attribute the residual

leakage of electroscopes to the pres-

ence of radioactive material in the air and in the soil. This explanation seemed to be confirmed in 1901 by the ex-

periments of C. T. R. Wilson, in England, and

Elster and J. H. Geitel, in Germany. They measured the conductivity of stagnant air in a dust-free, sealed vessel enclosing an electroscope, and found that some residual ionization still persisted. But, eventually, a nagging question arose. Could internal radioactive impurities account entirely for the current that always remained despite aH efforts to eliminate contaminations? Perhaps radium emanation and its disintegration products produced some type of radiation that was capable of penetrating even the walls of an ionization chamber. Testing this hypothesis, E. Rutherford and H. L. Cooke, in England, and J. C. Mc-

THE HEROIC AGE OF COSMIC RAY EXPLORATION

Clennan and

E. F. Burton, in Montreal,

found in 1903 that very

thick shielding by water or other inactive materials reduced the

30%. However, there was no appreciable further change when the thickness of the lead shield exceeded about 2 inches. Measurements were then made by McClennan on land and aboard a ship at sea, in what appears to have been the first cosmic ray expedition. Later, he even took his apparatus out on the ice of Lake Ontario. These observations suggested that the ionization was greater over land, and could be ascribed to penetrating gamma radiation from the soil, which always contains traces of radioactive impurities (uranium, thorium, radium, and their decay products). The residual ionization over water was attributed to minute radioactive impurities in the walls and gas filling of the ionization chambers, and to y rays from the surrounding air. However, A. S. Eve, in Montreal, was skeptical. In the first quantitative analysis of this problem, he showed, in 1905, that the radium and thorium emanations in the air were not sufficiently abundant to produce more than one-tenth of the effect that was observed over water. Subsequently, other attempts were made
ionization of air in a closed vessel by about
to

cm 3 /sec,

account quantitatively for the observed 10 to 20 ion pairs/ culminating in the conclusion by an Italian physicist,

D. Pacini, in 1910, that the observed excess ionization might be

produced by sources other than the known radioactive substances.

the

Early experiments designed to assess the contribution from ground by taking ionization chambers up into towers were

inconclusive because of the excessive background of radioactive

substances in the stone structures that were selected for this pur-

experiment aimed determining the altitude variation seemed to indicate that the ionization decreased with altitude! In 1910, Father Th. Wulf determined that the residual ionization on the ground in Paris was 6 ion pairs/cm 3 /sec (after subtracting the chamber background) while at the top of the Eiffel Tower, 300 meters above the ground, the strength was reduced to 64% of that value. However, he had anticipated that the intervening layer of air would cause a much greater decrease, to about 10% of the ground value.
pose. Ironically, however, the first definitive
at

COSMIC RAYS

Wulf concluded

that either another source of y rays existed in the upper layers of the atmosphere, or the absorption of y rays

in air was apparently smaller than

had been assumed hitherto. Meanwhile, K. Bergwitz had, in 1909, already made balloon ascents that showed a more marked decrease of the total ionization, which at 1300 meters was down to about 25% of the value on the ground. This was interpreted as being in better accord with expectation if the ground were the source of gamma radiation that was responsible for the residual ionization. However,
there was

some question about the

validity of Bergwitz' measure-

ments, since his electrometer was

deformation of his pressure make three balloon ascents From his measurements of Gockel concluded that the "decrease of the radiation with height was still less than was found earlier." The source of the ionization at high altitudes, according to Gockel, was disinte. .
.

damaged during the flight by vessel. This prompted A. Gockel to over Switzerland in 1910 and 1911. the ionization up to 4500 meters,

gration products of the

known

radioactive substances.

In order to avoid the pitfalls encountered by Bergwitz, Gockel

had connected his ionization chamber to a rubber balloon that expanded as the external pressure decreased, thereby avoiding a pressure differential between the gas filling and the outside air.
Consequently, the pressure of the gas in the vessel varied with
altitude. Under the prevailing conditions, the measured ionization was proportional to the gas pressure. If Gockel had only corrected his measurements for this variation of instrumental sensitivity with pressure, he would have concluded that there was a signifi-

cant increase in ionization as the balloon ascended!

was a happy circumstance that Victor Hess, an Austrian was an ardent amateur balloonist. After reading about Wulf's Eiffel Tower experiment, he decided that the accumuIt

physicist,

lated evidence pointed to the presence of a previously

unknown

source of ionization.
bility

He

initiated a

program

to check this possi-

of y rays from

experimentally by measuring directly the absorption in air an intense radium source (1500 milligrams). His

measurements, made outdoors by varying the distance between a closed ionization chamber and the source up to 90 meters, firmly established that y rays from the ground are almost completely

THE HEROIC AGE OF COSMIC RAY EXPLORATION

absorbed at a height of 500 meters. Then, he designed an instrument that could survive the rigors of an open balloon gondola an airtight ionization chamber with walls sufficiently thick to withstand a pressure differential of one atmosphere, and con-

taining a temperature-compensated

Wulf

fiber electrometer.

In 1911, the first of a series of ten balloon flights (each carrying two or three instruments operating simultaneously) reached 1070 meters. Hess concluded that the radiation at that height was not appreciably different from that at sea level. But, in the following year, he attained an altitude of 5350 meters (Plate I). After an initial decrease, the ionization current above 800 meters seemed to increase. Between 1400 and 2500 meters the sea level value was clearly exceeded and at 5000 meters the reading was several times as great as on the ground. Hess saw no possibility of accounting for his results in terms of radioactive substances in the air but

was forced "to have recourse

to a

new

hypothesis; either invok-

ing the assumption of the presence at great altitudes of previously unknown matter, or the assumption of an extra-terrestrial source
of penetrating radiation." Twenty-five years later, Hess received

Nobel

Prize for his discovery.

flights by W. Kolhorster, a German physicist, who reached a height of 9300 meters in 1914, provided clear confirmation that should have made Hess's conclusion incontrovertible. The ionization continued to increase until it attained a value about fifty times that at sea level. Assuming that, as in the case

Further

of

x and y rays the intensity drops off exponentially with distance x according to the law
Ix

/ exp(-/3 lx)

(1-1)

Kolhorster determined from his data the absorption coefficient

fa

10 -

per

cm

of air, representing almost ten times the

penetrating power of the most penetrating y rays then known.

THE RADIATION ERA

In retrospect,
it is

not surprising that the earlier workers were

inclined to question whether the residual ionization in a sup-

posedly "clean" vessel was representative of anything other than

COSMIC RAYS

some stubborn contamination that defied all efforts to remove it. Indeed, most of it was attributable to such causes. For, in fact, the ionization produced by cosmic rays in air at sea level is only about 20 percent of that arising from natural radioactivity! By the same token, it is readily understandable that this new
similar to the

phenomenon should be identified as electromagnetic radiation, known y rays, although much less easily absorbed

in passing through matter.

The

only other available candidate,

the electron, appeared not to be naturally adapted to yielding the exponential type of absorption law that seemed to be followed by the penetrating radiation. Furthermore, /? rays from radioactive substances, the most penetrating charged particles then known, were absorbed in less than 1 millimeter of lead. The extension of the measurements to high altitudes provided the crucial evidence that should have removed all doubt concerning the existence of an unknown radiation with a penetrating power exceeding by many times that of y rays from natural radioactive substances. But a stubborn skepticism in some circles concerning the reality of the new penetrating radiation from above (Ultrastrahlung or Hdhenstrahlung) combined with a

World War to result in a hiatus in the experimental investigations. High altitude experiments were not resumed until 1922, when R. A. Millikan and I. S. Bowen started sending up balloons
carrying small self-registering electroscopes.

Once
stance

again, the waters were to be

the choice of launching

muddied by

a chance circumto a discrepancy

sites

that led

between the American and European results. The ionization at very high altitude over San Antonio, Texas, was only 25 percent of that over Central Europe! This consequence of the then unknown latitude effect triggered off a renewed attack, exemplified by Millikan's subsequent conclusion in 1924, based upon ionization measurements atop Pike's Peak, that ". there exists no such penetrating radiation (of cosmic origin). The whole of the
. .

penetrating radiation

is

of local origin.

How
is

such quantities of
as yet

radioactive material get into the upper air

unknown."

This proved to be but one of a great many erroneous conclusions about cosmic rays. The state of flux of ideas about cosmic rays in the early days is admirably portrayed in one of the in-

THE HEROIC AGE OF COSMIC RAY EXPLORATION

witness in the trial of a

tive effects
is

numerable anecdotes about Millikan. Appearing as an expert man who was being prosecuted for selling
bottles of water with a label claiming that the liquid

had

cura-

because

it

had been irradiated by cosmic

rays,

Millikan

supposed to have pointed out that, indeed, the contents of the had been irradiated by cosmic rays. And, as for the claimed medicinal properties, who hadn't made incorrect statements about cosmic rays?
bottles

Measurements of the change in ionization with depth in snowfed lakes (to minimize background radioactivity) at different altitudes (Muir Lake at 3590 meters, Arrowhead Lake at 2060

1926 that "very hard were entering the earth uniformly from all directions." He christened them "cosmic rays." But all was not yet serene. One of the stormiest marathon debates between two giants in the history of science, Millikan and Compton, was
finally

meters)

convinced Millikan in

etherial rays of cosmic origin

still

to ensue.

consequences of the generally accepted assumption that cosmic rays were y rays endowed with tremendous energies were pursued assiduously during the years that followed this much belated universal agreement as to their existence. The ultimate
goal was to determine the energy spectrum of these incoming high

The

frequency photons that quantum theory had already associated with high energy. This was accomplished by assigning an initial arbitrary photon frequency distribution, and then tracing its consequences with respect to the variation of the ionization as a function of depth in the atmosphere, or in lakes, or under other absorbing media. By trial and error, the spectrum providing the best
particles
fit

to the observations

culties of envisaging cosmic rays as energetic electrically

was found. Although the difficharged

seemed insuperable, the required extrapolation of the laws of y-ray absorption was still fantastic. In some cases, the procedure was equivalent to extrapolating for several hundred miles a curve drawn on an ordinary piece of graph paper covering the range over which the law that it represented had been experimentally verified. The game consisted of dividing the incoming beam into several distinct bands of photons with characteristic absorption cothen

known

COSMIC RAYS
efficients

p u such

that the combination

t

ionization vs depth curve.

would yield the observed Each p was associated with the corre-

sponding photon frequency vv

The

energies

Wj

hv^,

were then

related to the energies of certain atom-building processes, as sug-

gested by the mass defects of atoms revealed by F. W. Aston's measurements of isotopic masses in 1927. Just as the atomic spectrum had led to a thorough understanding of atomic structure, it was speculated that the cosmic ray spectrum could reveal how more complex forms of matter are formed from hydrogen. As is shown in Fig. 1-1, Millikan and

18 16

14 12 10

\
C

He\

8 6

\l
N
Si

4
2

^ sSf
Ca

AI~~Fp~

12

16

20

24

28

32

36

40

44

48

52

56

60

64

68
(dots

FIG.

1-1

Millikan

circles)

with

built-up
in

and Cameron's comparison of experimental data curve compounded from four absorption
in

and

coefficients.

Abscissas:

depth

equivalent meters of water beneath surface of atmosphere.

ions per cubic centimeter per second.

Ordinates: ionization

G. H. Cameron interpreted their observations as providing evidence for the existence of four energy bands associated with atombuilding "in the depths of space." The four bands were identified with "acts pi formation of hydrogen into the celestially common elements helium, oxygen (C, N, O), and silicon (Na, Mg, Al, Si

and

S)." It is easy to

appreciate the reluctance of

its

originator

to relinquish this

most alluring hypothesis envisioning cosmic

rays as the birth-cries of the elements.


THE HEROIC AGE OF COSMIC RAY EXPLORATION

THE CORPUSCULAR ERA

Meanwhile,
certain

experimental

developments

seemed

to

point to the possibility that the observed ionization produced

by cosmic rays might be attributable to electrically charged particles. In 1927, a Russian physicist, D. Skobelzyn, observed tracks arising from the passage through his Wilson cloud chamber of /3 rays having energies at least an order of magnitude greater than those emitted by the known radioactive substances. Furthermore, these occurred at a sufficient rate to account for the whole of the cosmic ray ionization. At the same time, a Dutch physicist, J. Clay, was conducting the decisive experiment that, following the familiar pattern in the history of cosmic rays, should have ended the controversy but didn't. J. J. Thomson had shown in 1897 that moving electrically charged particles are deflected by magnetic fields. As early as 1904, C. Stormer, a Norwegian who used to remarkable advantage his ideal location in the zone of
rence,

maximum

auroral occur-

had

started his brilliant analysis of the effects of the geo-

magnetic field upon the motion of incoming charged particles (presumably the cause of the magnificent auroral displays that he was studying). Stormer's results revealed that, for electrons of a certain energy, which could reach the earth at high latitudes, there was an inaccessible equatorial zone. Thus, if cosmic rays were electrically charged corpuscles with the appropriate energies, fewer should arrive at the equator than at higher latitudes. On the other hand, if the incoming cosmic rays were electromagnetic, they would not be affected by the earth's magnetic field. So Clay mounted his apparatus aboard a ship traveling between Holland and Java. In three different voyages, the intensity was consistently lower (by about 11 percent) near the equator.
.

But

number

of other investigators failed to confirm

the

crucial findings

first

published by Clay in 1927.

these apparent discrepancies are now obvious. blunder committed by some of those who disagreed with Clay's

The reasons for The most flagrant

10

COSMIC RAYS

conclusion that the ionization changed with latitude in going toward the equator was that they generalized from observations

made north
ter 4).

of the so-called knee of the latitude effect

latitudes, the ionization

(cf.

Chap-

near sea level reaches a constant plateau value determined not by the earth's magnetic field, but rather by absorption in the atmosphere. In 1929, W. Bothe and W. Kolhorster demonstrated by ex-

At high

periments

with Geiger-Mueller counters in coincidence (cf. Chapter 3) that the penetrating power of the ionizing particles was comparable with that deduced from the variation with depth
of the total cosmic radiation.

They also pointed out that the lower energy particles among those they observed would be excluded from the equatorial zone, as indicated by the experiments
of Clay. Perhaps, they speculated, these ionizing corpuscles might be the primaries themselves!

And now, the climax neared as A. H. Compton mounted a program of heroic proportions. Beginning in 1930, exceedingly careful measurements with the "improved cosmic ray meter" shown in Fig. 1-2 were made in a world-wide survey by twelve expeditions involving eighty cooperating physicists at a hundred
stations.

The

results dispelled all the doubts, including those that

had even

started to plague Clay himself.

As Compton put

it

so

succinctly, "Isocosms* are not to

be argued with."

THE MODERN ERA

became textThen, interest in this challenging and rewarding field burgeoned. Whereas the annual output of published papers, both theoretical and experimental, had been less than 10 between 1912 and 1920, the number rose to more than 200 per year between 1933 and 1936, and more rapidly thereafter. It was inevitable that cosmic ray research should develop along two main lines. One, the subject of this book, is concerned with
book
gospel, cosmic ray research really

When

the corpuscular nature of the primaries

came of

age.

* Lines of equal cosmic ray intensity plotted on a world map. These appeared to follow the contours of constant values of the horizontal component of the earth's magnetic field (isomagnetic lines).

THE HEROIC AGE OF COSMIC RAY EXPLORATION

11

144 V

FIG. 1-2

Portable cosmic ray meter designed for the world-wide survey organized
in

by Compton

1930. After more than three decades a larger version, the Model C
is still

ionization chamber,

continuously recording the cosmic-ray intensity at several

widely separated stations.

the cosmic rays themselves

what are

they,

where do

tl^ey

come

from,

how do

they get here, and so on.

The

other involves the

utilization of this windfall of high-energy

subatomic particles for

interest as a

studying their interactions with matter.

The

latter category at first

was of far-reaching

12

COSMIC RAYS

proving ground for testing the validity of the laws of quantum

electrodynamics in extreme energy regions where breakdowns

were theoretically predicted. Then, the cosmic rays served

as a

and investigating the properties of new "elementary" particles. Finally, they provided the means for studying high energy nuclear interactions, with the aim of shedding light on processes which are germane to the basic understanding of nuclear forces. As man-made machines capable of accelerating particles up to cosmic ray energies became available, a sort of bifurcation occurred, and high-energy nuclear physics, or particle physics, was recognized as an established field in its
for discovering

medium

own

right.*

mutual interest have persisted, and the two breeds have retained a common meeting ground. This is because, on the one hand, particles endowed with vastly
Nevertheless, strong bonds of

higher energies than

man

will be able to

produce in the labora-

tory will always be present

among

the cosmic rays; whereas,

on

the other hand, in order to interpret observations with cosmic

ray instruments in the lower atmosphere,

it is

necessary to un-

derstand in detail the complicated processes involved in the transfer of energy from the primary cosmic rays (which are initially incident upon the outer limits of the earth's atmosphere, but which do not penetrate very deeply because of their propensity for interacting with the atmospheric constituents) to their progeny, secondary cosmic rays (which are ultimately detected

at

ground

level).

How

did our knowledge about cosmic rays develop after their

identity as electrically charged particles

was established? The

story will unfold in subsequent chapters, but, as a prelude, let

what has been learned thus far. Primary cosmic rays are the atomic nuclei of elements with which we are familiar on earth. Hydrogen nuclei protons are the most abundant, followed by the second element on the periodic^ table, the nuclei of helium atoms alpha particles. Roughly speaking, these are present in about the same proportion as the relative abundances of these elements throughout the

us summarize in capsule form

* See

A.

momentum Book No. I, Elementary Particles, by D. H. M. Thorndike, Van Nostrand, Princeton (1964).

Frisch

and

THE HEROIC AGE OF COSMIC RAY EXPLORATION

universe

13 the other hand,

1

about

10 hydrogen to

helium.

On

heavier nuclei, although relatively scarce (about

percent), are

overabundant in the cosmic-ray beam. Nuclei considerably heavier than iron are found to be exceedingly rare members of the cosmic-ray family. There are also some electrons, to the extent of a few percent. The energies of primary cosmic rays range from a lower limit under 10 6 eV to an upper limit of at least 10 20 eV. Whereas roughly one particle having an energy near 10 9 eV passes through one square centimeter at the top of the atmosphere each second, only a single 10 2 eV particle impinges on an area of a hundred square kilometers in a year. The number of cosmic rays above a given energy E falls off approximately as _1 5 In interplanetary space, a total of about four cosmic-ray particles passes through each square centimeter every second. The corresponding energy density amounts to approximately 10 -12 ergs per cubic centimeter, about equivalent to the energy that reaches us in the form of starlight, or to the energy that is associated with magnetic fields in space. The cosmic rays arrive essentially isotropically except for some local effects of solar origin. Most, but probably not all, of the cosmic rays originate in our galaxy, and hence are called galactic cosmic rays to distinguish them from energetic particles emanating from the sun. Various acceleration mechanisms may be responsible for imparting to them their high energies, and they may wander through the galaxy for very long periods before
.

reaching the solar system. They alone are the only bits of matter that reach us from outside the solar system.

This

recital of their

nature and properties suggests that cosmic

rays play the role of a virtuoso in the

broad concert of

science.

For, indeed, interest in these gadabouts of the universe far tran-

scends the realm of physics, and considerable impetus for study-

ing them stems from their broad interdisciplinary ramifications.

Cosmic rays serve

as a

unique tool

for investigating a variety of

problems, not only in the physical sciences, but in such seemingly remote fields as archeology and geology; their study is of funda-

mental interest to astronomy, astrophysics, solar physics, cosmology, geophysics, and high-energy nuclear physics. The scale

M
of distances over

COSMIC RAYS

which they function as effective probes ranges from the interior of atoms to the environs of our planet on out to the most distant reaches of the galaxy in which our solar system is located (the Milky Way) and even beyond. It will become
is

increasingly evident in the sequel that the heroic age of cosmic-

ray exploration
future!

not likely to come to an end in the foreseeable

Collisions of Cosmic

Rays

with Matter
By
their fruits, ye shall

know them. MATTHEW VII, 20

The label "cosmic radiation" has generally been attached to the totality of all the different primaries impinging on the top of the atmosphere, together with the secondaries comprising all
statistics

succeeding generations of their multifarious progeny. The vital of the particles that play significant roles in the phenom-

enology of cosmic rays are summarized in Table A-l (AppenVarious other evanescent members of the rapidly growing family of "elementary" particles that rarely appear among the cosmic rays are best ignored for our purposes, since interest in
dix).

within the domain of high-energy physics. and nuclear processes come into play as cosmic rays plow through matter, "frittering away their energy" to borrow a phrase from picturesque Millikanese. The prerequisite of all methods for studying cosmic rays is a thorough understanding of their interactions, for it is only the consequences of
falls

them properly

variety of atomic

their collisions that render them "directly" observable rather than theoretically assumed. Knowledge about their interactions is essential both for designing detectors and for interpreting ob-

servations in the light of instrumental responses. Thus, appropriate sensors can be developed not only for detecting particles,

but for determining their relevant characteristics. Furthermore, we must be able to trace backward from observations within the atmosphere to the incident primaries which they represent. Finally, even interstellar space is not devoid of
15

16

COSMIC RAYS

matter, hence the primaries themselves have suffered collisions

on

their long

and tortuous journey toward the

need
to

solar

system.

Clearly, these

be taken into account in studies on the

origin of cosmic rays.

Particle interactions generally fall into one of two categories, depending upon whether the nature of the force is electromag-

netic or nuclear. Electromagnetic forces

may

cause the transfer

particle annihi-

of energy by excitation, ionization, radiation, photoelectric emission,

Compton

scattering, pair production,

and

Nuclear forces come into play when nucleons interact to produce mesons, or to induce nuclear disruptions. In some cases, such as photodisintegration of nuclei and photoproduction of mesons, both types of force may be involved. Let us now consider how the various components of the cosmic
lation.

radiation behave
ferred,

when

they encounter matter. But before disis

cussing the different types of interaction in which energy

it is

trans-

helpful to become acquainted

first

with a few con-

cepts that are useful in describing absorption processes.

When the speed v of a particle of mass approaches the speed of light c, its energy can no longer be represented by \ M^v 2 as in ordinary Newtonian mechanics. Otherwise, the limiting energy that could be attained by any
Relativistic Energy.
particle

Relativity theory, which prescribes would be \ c2 the limiting velocity c in the first place, tells us how to handle
\

this situation.
istic

must forego the fun of deriving the relativtransformations here,* and will note only the results that

We

are essential for understanding the behavior of cosmic rays.

significant

The

parameter

is

the ratio of particle velocity to the speed

of light,

/?

v/c.

We

start

The mass Mp

of a particle

with a few assertions. moving with speed v

fie is

where
total
*

is

the rest mass of the particle.

The momentum p and

energy

U of this particle
9,

are given by:

Introduction to the Special Theory of
(1964).

See

momentum Book No.

An

Relativity,

by R. Katz, Van Nostrand, Princeton

COLLISIONS OF COSMIC RAYS WITH MATTER

., = Mev=
(1

17

M$c __^

, n _v (2-2)

cle

that the total energy U of a stationary partiequal to c2 in accordance with the well-known Einstein mass-energy equivalence principle.

Equation
(/?

(2-3) says

0) is

M
is

The

kinetic energy

obviously smaller than the total energy

the rest mass:
(2-4)

U by the amount of energy associated with E = U - Mac

2
.

Hence

E = *" [(T^jl - l ]
It is

<">

rest

advantageous to express particle energies in units of the energy of the particle. At very high energies, ;^>M c2
,

the total energy

and

kinetic energy are roughly equal,

and
(2-6)

E/Moc 2

~ U/Moc ~ 1/(1
2

2
/3

)*.

Thus, two very energetic particles (E ~ U) with different rest masses but the same energy in units of c 2 move with the same speed. For example, referring to Table A-l, we see that the value

of

for a 10,000 MeV muon (E/M c2 = 10 4/10 2 = 100) is the same as for a 50 MeV electron (E/m Q c2 = 50/0.5 = 100). The same advantage is gained by expressing momentum in units of MqC. Comparing Eq. (2-2) and (2-3), we see that at very high
(J3

velocities

1)

p/Mtf
or

~ U/Moc

(2-7)

pc a* E.

(2-8)

The
Its

relationships

among

p, U, E,

usefulness stems from Eqs. (2-2)

and /3 and

are

)% and U/M cz = 1/(1 - p )*. 1 While we are talking about relativity effects, we should point out that t, the lifetime before decay of a moving unstable particle, is given by a relationship exactly like Eq. (2-1):

P/MqC

/3/(l

(2-3), 2

shown in which

Fig. 2-1.

say that

18

COSMIC RAYS

Total Energy.. U/M c2

0.98

0.998

FIG. 2-1

Total

and

kinetic

energy

in

units of

Moc

2
,

and momentum

in

units

of

Moc, as a function of

J3.

TO
'

(1

01)*

(2

"9

where t is the proper lifetime measured in the coordinate system in which the particle is at rest. Consequently, making use of
the fact that 1/(1

2
/8

)*
r

= U/M

c 2,

we

find that
(2-10)

= (U/M^ro.
the
lifetime

Thus,

to a
(ft

stationary
1)

observer,

of
its
is

a relativistic

particle

~
=

appears to be proportional to
lifetime of a 10 14
6
)

energy! For ex-

ample, the
(10
1

mean
(10

eV muon
2 sec.

approximately

V10

8 )r

(2

X 10- 6 )

sec

Absorber Thickness. It is obvious that a particle passing through 1 cm of water sees more entities with which it can interact than it would if the water were replaced by 1 cm of air at STP. To b^ sure, the ratio of densities, p H2 o/Pair 800, is the same
as the ratio of the

amounts of the two substances encountered by

the particle in this example.

On

the other hand,

if

the air path

length
is

m, then the total quantity of absorber 800 times the original amount. We must therefore specify both
is

increased, say to 8

COLLISIONS OF COSMIC RAYS WITH MATTER

the density p
quantities,

19

and the path length x. Merely by multiplying these we obtain a valid measure of the total mass of abis

sorber traversed. Thus, in the above example, the mass of 8

meters of air
quantity px

is

equal to the mass of 1 centimeter of water. The a sort of surface density, measured in grams per
the superposed atmospheric mass
is p Hg ^Hg = Conveniently, the atmos-

square centimeter.

At
(13.6

sea level,

gm/cm 3 )

(76 cm)

1000

gm/cm 2

pheric depth in

gm/cm 2

is

almost exactly equal to the pressure in

millibars. In the vernacular,

we sometimes

say "grams" in refer-

ring to absorber thickness, or atmospheric depth,

when we

really

mean "grams
through a
the area

per square centimeter."

Cross Section. Let us imagine that a

beam

of particles passes

target, as illustrated in Fig. 2-2,

a.
<j>

uniformly irradiating

The

particle flux

is

defined as the

ing through unit area in unit time,

n Particles /sec Velocity = v

e.g.

number of particles n passn/cm 2 /sec. In a unit vol-

Flux

<f>

n/A s Particles /cm 2 /sec

Density Np=</v Particles/cm 3

Nt

Targets /cm 3

Radius

= r

Effective

Area

= o-

Radius

= r

+rt

Radius

= r t

Area = A,

FIG. 2-2

Schematic representation of the meaning of cross section.

20

COSMIC RAYS

ume, the target contains iV t uniformly distributed identical bodies. Suppose both the projectiles and the entities comprising the target are hard spheres, with radii r and r t respectively. The p bombarding particle collides with one of the objects in the target if it comes within the effective area a = 7r(r + r 2 The chance t) g p
of a collision as the incident particle attempts to pass through the

thin lamina dx

is

simply the ratio: (total projected effective area

of all the centers in

^ s )/(total area A s ). Referring to the figure, this ratio is

^ Nf
where ag
is is

dx ^

Nt(rgdX)

(2-11)

the geometric cross section.

The

unit of cross section

the barn (10~ 24

cm 2

or the millibarn (10~ 27

cm 2). The

radius

is approximately l.2A^ X 10~ 13 cm, where A is the mass of the nucleus in atomic mass units; hence the nuclear geometric cross section is about 0.05A Vs barns. Most interactions do not correspond to billiard-ball collisions. The cross section ct^ is the effective area for a given type of interaction presented by one target particle per unit area perpendicular to the direction of the beam.

of a nucleus

Mean
the

Free Path. In Eq.

is

(2-11),

note that

N ag
t

the product of
effective

number

of absorbing centers per unit

volume and the

by a through the target. Its reciprocal is the distance, on the average, between the centers along its route, the mean free path X g Thus, in general,
the
of centers encountered
single particle in traversing unit distance
.

area of each center,

number

1/Nun =
In the above discussion, the

Xi.

(2-12)
free

mean

path represented the

interaction length Aj. If the initial collision is not catastrophic, but merely maims the bombarding particle, the absorption mean free path, or absorption length \ a can exceed the interaction
length.
s

When there are many particles in a beam, we can measure the attenuation, or fractional reduction of the flux suffered from interactions. The number of casualties depends upon how many mean free paths have been traversed. Thus,
Attenuation.

COLLISIONS OF COSMIC RAYS WITH MATTER

21
(2-13)

-d<j>U
Integration gives
<j>

(dx/X).

</>

exp(-*/A).

(2-14)

Absorption Coefficient. The absorption coefficient is equal to 1/A. If A and x are expressed in cm, /3 is the linear absorption coefficient (cm- 1 ). If A and x are expressed in gra/cm 2 /3 m is the mass absorption coefficient (cm 2 /gm).
f

ELECTROMAGNETIC INTERACTIONS
The
ploys
theoretical analysis of electromagnetic interactions

em-

quantum

electrodynamics.

The

formulas

thus

derived

have been checked by a variety of experimental techniques.

The

energy-transfer processes, governed by conservation laws, stem from the interrelationships between particles and fields. The particles

are both the sources of the fields (or the fields themselves),
recipients of their effects.

and the

fields in the macroscopic world are completely described by classical Maxwell theory. However, when we are dealing with microscopic systems, quantum laws hold sway,

Electromagnetic

and a
ties

field

becomes a flux of photons.

Particles

now

play a

dual role:

(1) as

radiation quanta, their electromagnetic proper-

reside in their charge

and magnetic moment;

(2) as constitu-

ents of matter, their mechanical properties are inherent in their

mass and spin.

All interactions are treated in terms of a
case of ionization, for example, this
electrical charges that
is is

field of force.

In the

the familiar force between

described by Coulomb's law:

where F

is the force, in vacuo, between two charges q x and q 2 of bodies that are very small compared with the distance/ between them, r x is a unit vector along the line joining q t and 2 indiq cating the direction of that line, and Co the permittivity of free space, is a constant [ = 1/(36*- X 10 9 )farad/m in MKS units].
,

The

consequences of a collision depend upon the distance of

22
closest

COSMIC RAYS

approach of the interacting entities the impact parameter This quantity figures prominently in the theoretical analyses leading to the results that we shall now examine. Charged Particles. When the impact parameter greatly exceeds the atomic radius (d ^j> ratom ), the entire atom reacts as a whole to the passing charge. As a result, the charged particle loses energy gradually by exciting atoms along its route (excitation) and freeing those electrons to which it imparts amounts of energy greater than their binding energies (ionization). Figure 2-3 shows sche-

d in

Fig. 2-3.

Before

After

-e,m

AE

ze,M

E,p
FIG. 2-3

ze,M E'=E-AE,p'
,

Schematic representation of an ionizing interaction. The collision changes

the originally neutral

atom

into a positive ion.

matically an ionizing event produced by a particle with z units of

the electronic charge e

number Z.

when it interacts with an atom with atomic The magnetic moment does not play an important role
more
slowly, as 1/r2
.

in this case, since the magnetic force varies as 1/r3 , whereas the
electrostatic force falls off

emitted electrons constitute free negative charges and the originally neutral atoms positive charges (ion pairs), until they recombine. In the case of a solid, the subsequent
gas, the

In an ionized

history
is

of,

the liberated charge carriers depends

a metal or an insulator. Free electrons in the conduction

upon whether it band

of a metal, as well as

bound

electrons,

a passing charged particle.

that a thin insulating crystal,

We

may extract energy from have found in our laboratory such as MgO, becomes a conductor

COLLISIONS OF COSMIC RAYS WITH MATTER

23

temporarily

boosting

when high energy electrons are shot through it, bound electrons over the band gap into the normally

unoccupied conduction band. (Even the most ardent cosmic ray

aficionados sometimes like to try experiments that they have

concocted, as a change from having to wait for nature to perform them!)

High energy secondary electrons, originally referred to as knock-on electrons but now generally called delta rays because
of the characteristic appearance of their tracks in a Wilson cloud

chamber or a photographic emulsion, can be ejected when the impact parameter is comparable with the atomic radius (d ~ ratom)- In this case, the moving particle collides with a single electron which responds as though
it is

essentially free.

When the incident particle penetrates (d < ratom)> its trajectory can be deflected
the nucleus. In this case, the

deeply into the atom by the electric field of

point charge

Ze centered

at the

nucleus acts

Coulomb upon

field of the

a particle that passes at a

compared with the nuclear radius (~ 10~ 13 cm) but small compared with the atomic radius (/-'10- 8 cm). Because of the deceleration dp/dt associated with the change of direction of motion of the charged particle, electromagnetic radiation is
distance large

generated. Classical electrodynamics predicts that the intensity is proportional to the square of the acceleration. If only soft quanta are emitted, the incident particle merely undergoes inelastic scattering.

But, according to

quantum

with

up

to the entire particle energy can

electrodynamics, a single y ray emerge from a radiative

collision.

Figure 2-4 depicts such an event in which bremsstrahlung

(German: braking radiation)

it is

is produced. Furthermore, although not possible according to classical theory for a charged particle, such as an electron, to be deflected by a nucleus without

emission of radiation, in
ability for the

quantum

theory there

is

a finite prob-

particle

to be deviated
elastic

photon, and hence to undergo

radiative collisions (Fig. 2-5).

without emission of a scattering through non-

may
an

In the formalistic language of quantum electrodynamics, we regard the emission of bremsstrahlung as the transition from
initial state in

which the electron has a certain positive kinetic

24
Before

COSMIC RAYS
After

FIG. 2-4

Schematic representation of a radiative

collision,

resulting

in

the emis-

sion of bremsstrahlung.

energy

E and momentum p

in the radiation field containing zero

photons, to a final state in which the electron energy is lower (Ef < E) but still positive in a field containing one photon of

energy

W=

hv.

In

elastic scattering,

the energy of the electron remains un-

changed (' E) in a field containing zero photons both before and after the interaction, but the direction of the electron's momentum vector is changed, although its magnitude remains the
same.

Needless to

say,

we have

idealized the situation

somewhat by

Before

After

FIG. 2-5

Schematic representation of a nonradiative

scattering.

collision,

resulting in

elastic

COLLISIONS OF COSMIC RAYS WITH MATTER

25

associating each process with a different range of the impact parameter/atomic radius ratio. Thus, for example, bremsstrahlung may occur when the distance of important interaction is comparable with or even larger than the radius of the atom.

The
trons

effects of

screening of the nuclear charge by the outer electheoretical calcu-

must then be taken into account in the

approaches to the nucleus invalidate the approximation that the nucleus is a point source of the field. Actually, very close impacts do not contribute significantly to the energy loss by bremsstrahlung. Photons. We have delineated three modes of electromagnetic interaction of charged particles, related to the ratio of the distance of approach to the radius of the atom: (1) with the atom as a whole (excitation, ionization); (2) with a single essentially
free electron (8-ray emission); and (3) with the nucleus (bremsstrahlung). Similarly, there are three processes whereby photons interact through their electromagnetic fields, depending upon

lations. Similarly, very close

how

their wavelengths

compare with the dimensions of atoms

and of
tering;

electrons: (1) the photoelectric effect; (2)

to long

Compton

scat-

and (3) pair production. At low energies, corresponding

ber that

hc/k) the photoelectric effect predominates. As is seen in Fig. 2-6, this process is analogous to ionization in that the energy is transferred to the atom as a whole, except that
v

W=h

wavelengths (remem-

the entire photon energy

Before

is

imparted

to a single electron. Since

After

E=0 8E =

Binding

-e,m
E'=

/
W-BE

Energy

W=h*
FIG. 2-6

Schematic representation of photoelectric emission.

26
this
is is

COSMIC RAYS
a catastrophic collision,
it

turns out that the absorption

law

of the exponential form, Ix 7 exp ( /?jX), first encountered as Eq. (1-1). This points up a fundamental difference belose their entire energy in

tween the behavior of particles which

single acts

and those that

dissipate their energy gradually. In the

latter case, the

least

approximately,
electron

energy of the particle at any point determines, at how far it will subsequently travel in a given

absorber.

The
amount

may be

freed from the

its

atom

if

it

receives an

binding energy (ionization), or it may be excited into a bound state of higher energy (excitation) if the photon energy is less than the binding energy. Of course, in the latter case, the atomic quantum rules must be obeyed, which means that the photon energy must be exactly equal to
of energy exceeding

and final bound states. As the photon energy is increased, the Compton effect, which has no classical counterpart, sets in. Here, as is shown in Fig. 2-7, a photon transfers some energy and momentum to an electron that initially is essentially free and at rest.
the energy difference between the initial
Before
After

W=h

FIG. 2-7

Schematic representation of Compton scattering.

Again speaking
is

formalistically, the

Compton

effect

can be con-

sidered to be a transition from an initial state in which there

an electron with zero energy (E

is

0) in a radiation field conto a final state in

taining one photon of energy

the electron energy

W = hv
W
E'.

which

photon of energy

W=

E' and the radiative field contains one

hv'

COLLISIONS OF COSMIC RAYS WITH MATTER

27

At

still

higher energies,

when

the y-ray energy exceeds twice

the electron rest energy (2m c 2

1.02

the conversion of a y ray into a positive

MeV), pair production, and negative electron,

can occur.
Fig. 2-8.

The

energetics of this interaction are indicated in

recoil
is

Note that the nuclear

required for

momentum

conservation!
After

Before

W=h*

-e,m 0)

|?i i

U = -U' +e,-m
,

FIG. 2-8

Schematic representation of pair production.

This remarkable process in which matter materializes from

electromagnetic radiation
is

a sort of photoelectric effect in the

context of the Dirac theory, which envisages an "infinite sea"

of electrons in negative energy states.

One

of these electrons,

with total energy

U_,

is

initially in the radiation field of the

photon with energy

W = hv. The final state after the interaction,

normal

as depicted in Fig. 2-9, finds this ejected electron in the

world, with positive energy,

in the

+ U _.

But, in addition, the hole left

continuum of negative energy states now acquires positive U + and a positive charge ( e) = +e. This constitutes U_, and kinetic energy a positron with total energy U + = 2 + E_). E+ = (2m c The inverse process may also occur, and its consequences can
energy

W~

be envisioned by reversing the direction in Fig. 2-9. The electron makes a transition to the vacant negative energy state that
corresponds to the positron, hence both particles disappear. Conservation of

momentum

requires that

if

a free electron

and

28

COSMIC RAYS

+m

c2 -

FIG.

2-9

Energy-level diagram

illustrating

the energetics of the pair-production

process.

positron, both at rest, combine, two oppositely directed quanta,

the annihilation radiation, each with energy corresponding to the electron rest energy (m c2 0.511 MeV) must appear. If a

annihilation can occur. Again, as in the discussion of charged particle interactions, we have simplified the picture by not mentioning such effects as
the interaction of outgoing atomic electrons with the nuclear

nucleus participates, single

quantum

Coulomb
and

field,

the role played by electrons in different shells,

relativity considerations.

CHARGED PARTICLE ENERGY

A
derstanding of

LOSSES

great deal might be said about the development of our un-

all of these processes, starting in 1904 with Bragg's experiments on the specific ionization of alpha particles, followed by the evolution of empirical absorption laws, the work of Rutherford on alpha-particle scattering, and the theoretical treatment of collision laws by J. J. Thomson and by N. Bohr. After deriving the .classical relationships, we might then consider in detail the quantum mechanical treatment, pioneered by H. A. Bethe, H. J. Bhabha, F. Bloch, A. G. Carlson, A. H. Compton, W. Heitler, H. J. Massey, C. M0ller, N. Mott, J. R. Oppenheimer, R. Serber, to cite a few who provided the structure upon which our

COLLISIONS OF COSMIC RAYS WITH MATTER

29

understanding of collision phenomena is based. But, for practical reasons, we must circumvent this formidable undertaking here.
Fortunately,
if

we

are willing, for the sake of expediency, to accept

some formulas and graphs on faith, we can scrutinize them to extract their rich and comprehensible content. Ionization and Excitation. The rate at which a particle loses energy by excitation and ionization, expressed in units of electron volts per gram per square centimeter, is
-(dE/dx) ioa

| [(AV^KZKxro )} "^
2

[in

ffi^'

"

20 2 ],
(2-16)

where

E=
x
/3

energy in electron
v/c; v

volts,

= = Aa =
z

thickness of absorber in

particle speed,

gm/cm 2 c = speed
,

of light,

moc
r

= = =

Avogadro's number = 6.02 X 10~ 23 atoms per gram atom, charge on incident particle in units of electronic charge,
electron rest energy
classical

0.511

MeV,
e
2

electron radius

/moc 2

2.8

10 -13 cm,

and,

= mean

13.5 eV for Hydrogen). The term [(iVA /;4)(Z)(7rr 2 )] is actually the area covered by electrons in 1 gm/cm 2 of the absorber. Note that the rate of loss of

~ 13.5Z (=

ionization potential of the atoms of the absorber

energy is independent of the mass of the incident particle, but depends on the square of its charge. Thus, the specific energy loss of an a particle is four times that of a proton traveling with the same velocity. When t/c, (dE/dx) im varies inversely as
v2 because of the way the probability that a given interaction
,

depends upon the length of time that the interacting something to happen. The logarithmic term takes care of the Lorentz contraction as v apwill occur
entities are sufficiently close for

proaches

c.

dependence of (dE/dx) ion appears in /3, and the absolute value of the rate of energy loss depends only upon
entire energy

The

the atomic constants of the absorber,

of the particle.

and the charge and

velocity

Consequently, except for certain

frills

which

30

COSMIC RAYS

need not concern us here, the universal curve shown in Fig. 2-10 describes the energy loss by ionization and excitation for any
type of singly-charged particle.

The

\/v 2 dependence in the nonrelativistic region and the

slow increase above the

minimum
/}

are evident.

The

lowest point
to

on the curve occurs

energy
ergy

at

0.95.

This corresponds

total

U equal E = 2 M c2

to three times the rest energy, or a kinetic en.

Particles

having values of p near or above

this

value are relativistic, or minimum ionizing. Typical values of (dE/dx) ion are 1.1 MeV/gm/cm 2 for lead and
\

minimum 1.8 MeV/

\
o
E 5 E

>
<d

4
""

Air

^2
Pb

0.1

0.2 0.3

0.5 0.7

1.0

3 4 5
0,98

10

20 30

50 70 100 p/mc
0.99995
(3

0.20
FIG. 2-10

0.45
loss

0.71

0.897

0.995

Energy

by ionization and atomic excitation as a function of

particles.

velocity.

The ordinates refer to singly charged

indicated values

For multiply charged

2 by Z
.

nuclei,

the

would be

multiplied

gm/cm 2

for air.

Denser materials extract energy

in a less dense

less

rapidly be-

cause the shielding effect of the constituent atoms limits distant

interactions

more than
,

medium.

The
per

general form of the equation representing the primary

specific ionization /p

which
is

is

the average

number

of collisions

gram per square centimeter

is

that result in the ejection of an

roughly similar to Eq. (2-16) and Fig. sometimes expressed in ion pairs per centimeter. In hydrogen at STP, the minimum primary specific ionization is about 5 ion pairs/cm. An energy of about 30 eV is
electron from an atom,
2-10. Specific ionization

COLLISIONS OF COSMIC RAYS WITH MATTER

31

expended in producing one ion pair. The total specific ionization j t is the average ionization produced by the primary and all of its secondaries, tertiaries, etc. Equation (2-16) embodies this contribution to the energy
If
loss.

the energy loss process described by Eq. (2-16) predominates, that is if radiative collisions are negligible, it is possible to predict the

amount
is

of matter that a particle with energy

will

penetrate. This
in
its

called the range R. Let us rewrite Eq. (2-16)

simplest form:

-dE/dx = /().

(2-17)
easily

The

range-energy relationship can

now
B

be determined by

integration:

-JdE/f(E)=f Q

dx

R.

(2-18)

Referring back to Fig. 2-10, it is clear that, when the particle is slowed down to sub-relativistic velocity below the ionization minimum, the rate of energy loss increases very rapidly, and the particle will be stopped in a very short distance. But for relativistic particles,

as

we have already

noted, f(E)

is

almost con-

Eq. (2-18) tells us that the range is proportional to the energy. Figure 2-11 shows a typical range
stant.

Hence, in

this case,

10*

10

7
in

Muon Energy
FIG. 2-11

10 eV

Dependence of range upon energy for

slope of the curve
is

fi

mesons

in air.

Note that the

very nearly unity at the upper end.

32
vs energy curve.

COSMIC RAYS

Of
is

course,

it

computed range

the total

amount

must be remembered that the of absorber, no matter how

tortuous the route through it, hence, for particles subject to appreciable scattering, the linear absorber thickness that can be traversed may be less than the theoretical range.

The term
energy

straggling refers to the spread about the average

loss that arises

from

statistical fluctuations in

the indi-

vidual processes.

Bremsstrahlung. Radiative collisions involve a

still

more com-

plicated combination of various probabilities that are subject to

extreme fluctuations. These preclude the association of a prescribed range with a given particle. For high energy electrons, the
equation for the bremsstrahlung energy
loss process is

-(dE/dx) iad

where a

= 4a(NA /A)(Z)(Z + l)(r 2 )[ln (183Z~*)], (2-19) = e 2 /flc = 1/137 is the fine structure constant, and the
(2-16).

other symbols are the same as in Eq.

We see that the radiation

number

loss varies as the

square of the atomic

materials far exceed those in the case of ionization, which varies as Z. Furthermore, there is a strong energy dependence, since (dE/dx) TaA
is

of the absorber, so differences

among

proportional to E.

An important feature of the radiation process is not revealed by Eq. (2-19), which holds exclusively for electrons. In contrast with ionization, radiation varies inversely as the square of the
mass of the incident particle. This happens because a heavier mass experiences a smaller acceleration under otherwise identical
conditions. This has important consequences. For example,
a

proton with a given kinetic energy, expressed in units of its own rest energy, would suffer an energy loss by radiation amounting to only (1/1836) 2 ~ one three millionths that of a comparable electron! Hence, for the most part, we need consider radiation losses only when we are dealing with electrons. Figure 2-12 shows the rate at which electrons passing through different materials lose energy by radiation, as compared with ionization and exv

citation.

We
of

can simplify matters by combining the constants in front

in Eq. (2-19) into a single constant, as follows:

COLLISIONS OF COSMIC RAYS WITH MATTER

33
(2-20)

1/Zo
If

= 4a(NA /A)(Z)(Z

l)(r

)[ln (183Z~*)].

you can persevere through a dimensional analysis of this equayou will find that X has the dimensions of a "length" in gm/cm 2 This quantity, called the radiation length, is a charaction
.

teristic of

the

medium. The
measured in

descriptions of radiation

are essentially independent of atomic

thicknesses are
this unit.

phenomena number when absorber

Typical values of

(and

Electron Energy

in

MeV
(1)

FIG. 2-12

Energy dependence of the rate of energy loss by

atomic excitation

and

ionization,

and

(2) radiation, for electrons in

is

lead

and

in air at STP.

The sum

also plotted.

some related parameters that we shall define shortly) are listed Table 2-1. Its physical significance becomes apparent when we look at the simple relationship that results from combining Eqs. (2-19) and (2-20):
in

-(dE/dx) Tad

= E/X

(2-21)

Suppose the energy of a particle before entering an absorber with radiation length X is E Then, on the average, after penetrating a distance x, its energy, determined by integrating Eq.
.

(2-21), is

Ex = E cxp(-x/X ).
So,

(2-22)
,

we

see that

is

the distance, in

gm/cm 2

over which the

energy

falls to

(l/e)th (0.368) of the original value. Sometimes,

34

COSMIC RAYS

nOm
"> co
00 tH
e\i

r-

> V
u

* CM SO

oooo
t-i

.3

XX,

s 9 a

f CN CM IT) T-H 00 >* ON CO T-I

^
a,

XX
ir*>

v^,

JPJ

SO \

*
o
'S3

8-p

t-h

r~ co cm

b V a 3 bl) w o t>r ri
rt

T3
Hydr
Nitro

CM
Alum

COLLISIONS OF COSMIC RAYS WITH MATTER

the term shower unit
to the distance in
I

35
is

(In

2)X

0.693X

used in referring
energy

which a

fast electron loses half its initial

(similar to the half-life in radioactive decay).

Another parameter
nately, this

that serves to characterize the effectiveness

is

of a given material as a radiator

the critical energy. Unfortuthat, fortunately,

term has two different connotations

are almost equivalent quantitatively: (1) ec is the energy that a given type of particle would lose by ionization alone, in traversing

a thickness of
(2)

one radiation length,

i.e.

ec

= Jf x

(dE/dx) ion dx;

alternatively

Ec

is

the energy at which the rate of energy loss

by ionization is equal to the rate of energy loss by radiation, i.e. (dE c /dx) ion (dE c /dx) It turns out, as Table 2-1 shows, that ec is slightly less than E c Roughly speaking E c ~ 800 MeV/Z and ec ~ 700 MeV/Z.
Ill(i
.

Elastic Scattering.
particles
is

The

net angle of scattering of a

beam

of
ac-

the statistical

sum

of

many

small deflections, in

cordance with the Gaussian distribution law. For particles with charge z, momentum p, and velocity v, the mean-square projected
scattering angle (6 2 )
is

approximately given by

where

is

a constant

(=21 MeV) independent

of the mass of

the particle

and of the nature of the medium.

Thus, the root-mean-square multiple scattering angle i& proportional to the charge of the particle. It decreases rapidly as the speed v increases, and increases as the square root of the total thickness x traversed, but varies inversely as the square root of
the radiation length

Cerenkov Radiation. For the sake of completeness, we

call at-

tention here to another electromagnetic effect that occurs under certain circumstances when a charged particle passes through
matter. If the particle speed v exceeds the speed of light in the

medium cm
velocity.

of a ship, or the shock

emitted in analogy with the bow-wave wave of a projectile moving at supersonic This Cerenkov radiation arises from constructive inter,

radiation

is

36
ference of electromagnetic waves that are emitted
positions.

COSMIC RAYS

when atoms
is

distorted by the passage of the particle return to their normal

How
6,

this intriguing

and useful
is

effect

comes about

illustrated in Fig. 2-13.

The

light

emitted in a characteristic

cone, of angle

according to the relationship:

cos v

1_
:>

(2-24)

where n

c/cm

is

the index of refraction of the

medium, and

P=

v/c.

A ^ o oo o oo o o o o o

o o

/3 =

v/c

c/c n
1/n/3

cos0=AC7AB =

cjOOqO

'^Q0M
O Oo c o o oo n
o

ol
O^^Oq
oo
O
AB=/3cAt

ooc?

o
o o

b^o o
B

ooo o

FIG.

2-13

Polarization

of a transparent

medium by
field set
its

the passage of a

charged
as

particle.
it

Wavelets are radiated by the dipole

up by the moving
in

particle

distorts the

atoms while passing them. Since

speed v exceeds the speed of

the

light in the

coherent
If

medium c,, the particle travels from A to B wave fronts, emitted as it passes the indicated
than
cm ,

same time

that the

points, reach the line BC.

in

is

less

the wavelets radiated from points along the track are not

phase, and hence they interfere destructively.

It

significant quantitatively as

must be emphasized that Cerenkov radiation (which is not an energy-loss mechanism, amountdif-

ing to about a thousandth of the ionization-excitation rate)

fers

from bremsstrahlung in that it involves cooperation among many atoms, rather than a collision with a single atom.

COLLISIONS OF COSMIC RAYS WITH MATTER

37

PHOTON ABSORPTION
We
have already noted that some y-ray absorption processes
acts.

involve the complete elimination of photons in discrete

Thus, we cannot talk about either a rate of energy loss, or a range that depends upon energy. We can only determine the attenuation of a beam of photons by calculating the cross section (hence the absorption coefficient) for each process. Photoelectric Effect. The following equation exemplifies the type of relationship that is derived from theoretical analysis of the process of y-ray induced photoelectric emission:
o-k

67rr

a4 Z 5 (m<x; 2 /Az/).

(2-25)

This gives the cross section <rK trons by very high energy y rays

for the ejection of K-shell elec(h v^>

c 2 ).
is

The
at

cross section for the photoelectric effect

as

significant only

low energies, such

below 50 keV

for

aluminum and 500 keV

for lead.

Compton Scattering. Although we will not derive the basic Compton effect equations here, we note that the Compton electron never makes an angle of more than 90 from the direction
of the incident photon.

At grazing incidence, the energy

of the

electron

is

zero. If v

represents the frequency of the photon be-

fore the interaction, the

minimum

frequency after scattering

is

2hv /moc 2

(2-26)

the maximum kinetic energy T m&x that can be transferred to the directly forward moving electron is

When

hv

^?>

c2

rmax =

h{v

is
.

mia )

= y> !

2wC

'

(2 ' 27)

In this case, the y ray with an energy ~ | m c 2

scattered directly back through 180,

scattering acomp

The

cross section for

Compton
2

when

hv ^>

c 2 is
crcomp

= (NA /A)(Z)(wr

)[ln

(2/^/W 2)

\](m<fi*/hv).

(2-28)

'

38
Pair Production.

COSMIC RAYS

The
is

theoretical treatment of the creation of

positron-electron pairs

rather involved. However, the calcula-

tions are closely related to the theory of bremsstrahlung, since

the processes are essentially inverse.

The

total cross section for

i.e.

pair production (for a particular set of conditions,

hv

^>>

c2
is

and neglecting screening of the nucleus by outer

o-pair

electrons)

= ZW(28/9) [In

(2hv/moc 2 )

218/27].

(2-29)

Compton scattering and pair production predominate in the energy range of y rays that contribute appreciably to cosmic ray
phenomena. The way in which
1.6

their relative contributions to

1.4

\
\

%
V

1.2

P3
I

(Tote ll)

->\

//
.5?

1.0

o \

'O \
\

//
\
/

SOB
0.6

Va

(Tota

\
\

< S

\
\

V
/

0.4
/
\

/
/

0.2

/
0.5
0.1

^^
"~-

~~

0.25 0.5

1.0

e5

10
in

25

50

100

500 and

Photon Energy
FIG. 2-14

MeV
in

Energy dependence of 7-ray absorption

aluminum and

lead,

the

separate contributions of photoelectron production, Compton scattering and pair

production
in lead.

the absorption of y rays

2-14.

depend upon energy

is

The

total absorption cross section

the

shown in Fig. sum of the indiis

vidual ones:
0"T

== Ophoto

"T Ccomp "T Opair(2-25),

In summary, we note that Eqs.

Fig. 2-14, reveal the following:
(1)

(2-28)

at very

and (2-29), and low energies, where

the photoelectric effect predominates, a T varies roughly as

Z5

(2)

COLLISIONS OF COSMIC RAYS WITH MATTER

in the intermediate region (0.05
0.5

39
hv

MeV <
the

<

16

MeV

for Al,

MeV <

hv
it

<

MeV for Pb), where

and
(3) at
is

Compton

effect is

most

important,

varies as Z;

higher energies at which pair

production takes over, there

Z 2 dependence.

NUCLEAR INTERACTIONS
Two generations of nuclear physicists have been striving mightily to attain a single objective that can be stated quite simply understanding nuclear forces.* Thus, it would be both

to discuss this subject here in other than quite general terms. One could digress into a philosophical discourse about what we mean by "understanding," but, suffice it to say that in the

presumptuous and impractical

present context

we

are referring to a conceptual framework that

would lead
interactions

to as satisfying

and

relatively

on the nuclear

scale as has

complete treatment of been carried out on the

atomic

scale.

of nuclear models that appear to fit one or another group of experimental data have appeared over the years. Pictorial nomenclature such as the liquid drop model, the shell model, the optical model, and so on designated different ways of thinking about nuclei through analogies that are susceptible to calculations which can be compared with experimental results. Innumerable theories envisaging different types of fields have been proposed to describe the forces between nucleons (proton-

A number

proton, neutron-neutron, neutron-proton). In many cases, starting with Yukawa's theory that presaged the tt meson, these have predicted new fundamental phenomena. Brilliant analyses based

on imaginative,

exciting,

and challenging new concepts

are revo-

lutionizing thinking about this subject.

Fortunately, by following a heuristic approach,

we can pursue

the cosmic-ray problem without delving into the details of the mechanisms of nuclear interactions.

Nuclear interactions
* See

fall
8,

into two categories:

weak and

strong.
S.

momentum Book No.

Structure of Atomic Nuclei, by C.

Cook,

Van Nostrand, Princeton

(1964).

COSMIC RAYS

40

Table A-l (Appendix) indicates how the various cosmic ray components behave in this respect. The topography of the simplest nucleus containing more than one nucleon is mapped in Fig. 2-15. When a strongly-interacting
particle collides with a nucleus, that
is,

when

it

approaches
it

suffi-

ciently close to experience the nuclear forces,

may become

thoroughly integrated, and thus heat up the nucleus as a whole (nucleon-nucleus collision). The compound nucleus survives for times of the order of 10 -16 10- 20 second, long compared with
t+300 MeV
2x10~ 13 cm

i5

-30 MeV
1

fermi[-<

FIG. 2-15

The potential energy of two nucleons at a distance

between

centers.

the transit time of a nucleon across the nucleus

Then

fragments
off,

mainly neutrons, protons, or alpha

down
to its

(~10 -22
state.

sec).

particles

are boiled

cooling the nucleus

normal

On

the other hand, a nucleon may interact much more rapidly -22 sec) with discrete nucleons in passing through the "nu-

(<10

clear matter" creating

in a stripping or a pickup reaction.

may pass near the edge The multiplicity of a particular interaction specifies the number of particles of a given type that emerge as a consequence of the collision. The interacnew
particles, or it

tion of a nucleon with a target nucleus can result in:

1)

2)
3)

Evaporation neutrons and protons; Secondary particles, principally pions; Nuclear breakup by:
a) spallation

or fragmentation, in which nucleons or groups

of nucleons, emerge.
b) fission, in

which the nucleus divides into two

(or

more)

nearly equal masses.

COLLISIONS OF COSMIC RAYS WITH MATTER

41

In addition to inelastic scattering of strongly interacting parnuclear forces can give rise to elastic scattering, as, for example, when neutrons collide with protons in billiard-ball
ticles,

fashion.

Experimental Methods
Did I see it go by, That Millikan mote?
Well, I said that I did.
I

made

good try. ROBERT FROST

The
cal

tools of the trade fall into

visual.

and

of sensor that,

two general categories electrimay utilize some type The associated measuring system, reby means of an
cosmic ray physicist

sponds to one or more characteristics of the radiation flux to which he exposes it. In this case, he may observe an average characteristic by measuring a current, or he may resolve the signals from individual particles, and count pulses. Alternatively, he may employ some method of obtaining pictures of the tracks produced by cosmic rays as they pass through matter, leaving their characteristic signatures. In some cases, both techniques are used in
combination.
All detectors have one factor in
sine

common

ionization

is

the

qua non

in their operation.

However,

all

the other inter-

actions described in Chapter 2

may

be invoked, as appropriate,

to serve as handles for identifying particles and determining their properties. The following descriptions of sensors and de-

employed in cosmic ray research will be confined underlying principles that are relevant for understanding how it is possible to unravel the wide-ranging characteristics of the different components of cosmic radiation. It is unnecessary for us to be concerned here with technological details or with
tecting systems
to the

the ancillary circuitry.*

* See, for

example,
B.

ment, by

W.

momentum Book No. 10, Radioactivity and Its MeasureMann and S. B. Garfinkel, Van Nostrand, Princeton (1966).
42

EXPERIMENTAL METHODS

43

ELECTRICAL DETECTORS
Included in
species in
this family are

which an electrodes, and the

particles (which
as

electric field
effects

two genera. The first comprises all is maintained across boundary of ionization produced by charged

may
is

be secondaries of the particles of interest,

observed.

with neutrons

The

second includes devices in

which light emitted as a result of the passage of cosmic radiation through certain substances is detected by a photosensitive tube. Ionization Chambers. The classical instrument in cosmic ray research, and the simplest to comprehend, is the ionization chamber. One might have expected that it would have faded into
obsolescence long before the passage of
years. Nevertheless, as

more than

three-score

we

shall see later,

important discoveries

are

still

being

made with

the help of this classical cosmic-ray

detector.
Basically,

an ionization chamber

consists of a gas-filled vessel

containing a cathode and an anode between which a difference of potential is maintained. The current representing the rate of
collection of ions

produced in the gas by ionizing radiation

is

measured.
plication.

The geometrical configuration depends upon the apThe arrangement sketched in Fig. 3-1 and the correCathode
Insulator

Insulator-

Pulse Amplifier

OR
Current Meter

FIG. 3-1

Basic elements of

an ionization chamber, a proportional counter, and a

Geiger-Mueller counter.

sponding pulse height vs applied voltage plot in

of the
first

Fig. 3-2

provide

a convenient basis for explaining the action of the three species

genus of the electronic detector family.

The

central wire or

anode
is

is

insulated from the cylindrical

cathode, and the current

measure of the radiation flux through the chamber. With a constant flux, the ion current rises
a direct

44

COSMIC RAYS

Applied Voltage

Gas amplification as a function of applied voltage. The abscissa scale and the exact form of the curve depend upon the nature of the gas filling, its pressure, and the dimensions and configuration of the electrodes.
FIG. 3-2

as the voltage

is

raised.
is

charges to recombine

This is because the tendency of + and reduced as their velocity increases with

the electric field strength. Finally,

tial

when

the difference of potenall of

becomes

sufficiently

high to insure that

sets in,

the charges
is

reach the electrodes, saturation

and the plateau

ionization

reached.
utilized

We
in this

mentioned

earlier that the particles

can be counted by

observing their individual pulses.

An

chamber

operated in the electron collection mode, to attain shorter time resolution than that associated with the much

manner

is

heavier and therefore slower-moving positive ions.

to the ionization

We

will refer

produced by a single particle as its pulse height. Thus, in Fig. 3-2, the pulse height on the plateau represents all of the energy deposited in the chamber by the particle. As we saw earlier [cf. Eq. (2-16)] the minimum ionization produced by

an alpha

particle

is

four times that of a proton.

Hence

these

particles are easily distinguishable

through their pulse heights.

Proportional Counters. By increasing the voltage beyond the

ionization plateau, the electrons released in the primary ionization process

may

sions to ionize the gas molecules,

themselves gain sufficient energy between colliand these in turn produce

further ionization,

plification results in

and so on in an avalanche process. This gas aman enhanced pulse that is strictly propor-

tional to the primary ionization.

The

multiplication factor

is

EXPERIMENTAL METHODS

45

function of the difference in potential. The gas amplification of the proportional counter reduces the external circuit require-

ments as compared with a pulse ionization chamber, although even in this case a high-gain amplifier is required. An especially important version of proportional counter contains boron trifluoride, BF 3 enriched with the isotope B 10 as the filling gas. A resonant exothermic nuclear reaction occurs when slow neutrons collide with this boron isotope, which contains 5 neutrons and 5 protons. Thus,
,

B 10

on

->

( 3 Li

0.9

MeV)

(2

He 4

1.6

MeV).

(3-1)

The
ions,

a short range in the gas

which have and hence deposit all of their energy as are enormous compared with those from ambient radiations.
pulses produced by these reaction products,

Neutron-produced pulses can therefore be selected by a pulseheight discriminator.

Geiger-Mueller Counters.

The

application of a

still

higher

voltage pushes the self-limiting avalanche action in the region of

limited proportionality (where the gas amplification
still

depends

somewhat upon the number of ions produced in the original act) beyond the pale into an explosive response. Instead of remaining
localized, the discharge

now spreads rapidly

to the

along the entire length

of the counter,

owing

copious emission of ionizing pho-

tons from the highly excited constituents of the gas.

Without

pursuing the details of the discharge processes that ensue, suffice it to say that only a single ion pair is sufficient to initiate the
action. It is as though the voltage across the electrodes were just below the point of spontaneous electrical breakdown. The tiniest

spark

suffices

to trigger off the entire discharge, the extent

which is unrelated to the igniting source. Thus, the GeigerMueller counter affords no inherent means for distinguishing the properties of the radiations to which it responds. However, the height of Geiger-Mueller counter pulses is uniform and enormous, thereby considerably simplifying the associated electronic equipment. There are two types of Geiger-Mueller counters. In one, the discharge is quenched by external action, essentially by reducing the applied voltage. Self-quenching counters contain an admixof

46

COSMIC RAYS

ture of gases or vapors that absorb photons and inhibit processes whereby secondary electrons emitted from the cathode would sustain the discharge in the absence of the quenching mechanism. Geiger counters were important historically because of their

adaptability to being incorporated into systems.

To

cite

only

one example, directional resolution can be attained, as we shall show later, by combining Geiger-Mueller tubes into a coincidence
telescope.

Solid State Detectors. We have already noted that ionization produced in solid materials by incident particles can be measured, hence the useful features of the ionization chamber, notably its linear relation between output signal and energy deposited, need not be limited to gas-filled vessels. Thus, recent advances in
solid state technology are creating a

new generation

of detectors

that are significantly extending our capabilities for studying cos-

mic

rays.

Aside from purely practical and technical considera-

tions, the principal

advantage of solid

state detectors resides in

As the experiments become more and more sophisticated, the importance of examining the fine structure grows increasingly evident. Thus, higher resolution in measurements of the spatial, temporal,
energy,

the remarkable resolution that can be attained.

and charge

distributions of cosmic rays

is

incessantly be-

ing sought to provide answers to old questions as well as to ones that are continually being generated.

new

ductor detectors

Because solids are so much more dense than gases, semiconmay be designed to stop certain particles completely, and thereby to measure their entire energy E. On the other hand, they can be made into very thin slabs, that have little effect on E, for determining the rate of energy loss dE/dx.
evident later

The utility of these features will become more when we discuss detector systems.

wide variety of semiconductor counters have become availBulk conduction counters consist essentially of a homogenous material, such as silicon or germanium, appropriately doped with impurity centers. Rectifying junction counters are the most widely used. These operate by virtue of the properties of the p-n junction. In the forward direction of the rectifier, electrons are driven by an applied field from the n-type material
able.


EXPERIMENTAL METHODS

47

(carriers normally negative electrons) to the p-type (carriers normally positive holes) and holes from the p-type into the n-type.

is

But in the reverse biased condition, in which the n-type region held at a positive potential, free charges are removed from the interface of the junction and the surrounding regions.
Actually, there
is

a built-in voltage even without the applica-

from the diffusion of each type of carrier into the other region, leaving excess positivelycharged donors in the n-type region and negatively-charged
ceptors in the p-type region.
ac-

tion of the external field. This arises

These constitute a charged double layer, which sweeps the carriers from a thin layer at the interface. This carrier-free region, or depletion layer, is essentially an intrinsic semiconductor, and electron-hole pairs produced by an
ionizing particle that traverses
Its
it

constitute a detectable current.

thickness can be increased by applying an external voltage that

assists

the built-in voltage.

Two

general types of p-n detectors are prepared by a variety

diffused junction
either

and surface barrier. In the former, an n-type or p-type material, a donor or acceptor impurity is introduced to form a depletion layer at the diffusion depth. The latter is made by forming a p-type layer on the surface of n-type material by oxidation, or by creating an

of processes

made with

evaporated metal-semiconductor interface.

The
is

principal cosmic-ray application of solid state detectors

in studies of the primaries. Detailed information about the

as a function of charge, mass, and energy with hitherto unattainable resolution is being obtained in experiments carried out in space. The observations are now being extended down to lower energies than were accessible heretofore. Because of the nature of the construction of semiconductor

magnitude of the primary flux

amenable to reducing the instrumental threshold for detection far below the limits imposed by the wall thickness of other types of radiation sensor.
detectors, they are

Scintillation Counters. The oldest method for detecting individual charged particles was to observe visually the tiny flashes of
light that are emitted

when

small crystals of certain substances,

such as zinc

ment

coated on a screen are subjected to bombardby radiation. A device utilizing this principle was called
sulfide,

48

COSMIC RAYS

/a

spinthariscope.

Tremendous advances

in the development of

exceedingly sensitive photomultipliers that can detect extremely

low-level light pulses, coupled with rapid progress in producing

having appropriate characteristics, led to the widespread adoption of this technique in modern cosmic ray research.
scintillators

The

basic principle of operation of a scintillation counter

is

shown

in Fig.

3-3.

Cerenkov Light

Scintillation Light

Photocathode
1st

Dynode

2nd Dynode

3rd Dynode

4 th Dynode

i th

Dynode

FIG. 3-3
tector.

Schematic arrangement of a scintillation counter and of a Cerenkov detypical photomultiplier

5

may have 10

stages

(i

10) each with a

s
.

gain

4,

providing an overall amplification of 10

Fluorescent radiation

is

their original lower energy state

a passing charged particle.

J

produced when electrons return to from which they are excited by In some cases, the transition to an
first,

intermediate, or metastable, state occurs

causing the emission

EXPERIMENTAL METHODS
to be delayed until the electron

49

is raised back to the excited state by a gain in thermal energy, and then returns to the lower level (phosphorescence). This process is inimical to attaining the de-

sired rapid response.

Inorganic and organic phosphors, in solid or liquid form, are

employed

as scintillators. In some cases, activators are added to provide luminescent centers having the property that energy of

released as radiation rather than as mechanical inIn other cases, these centers occur naturally. Just as with solid state detectors, scintillation counters can be used for measuring total energy deposited E, or rate of energy excitation
is

teractions.

loss

~dE/dx.
superior to Geiger-Mueller

Scintillation counters are vastly

counters

when

the detector areas are large. Although trays of

Geiger-Mueller counters can accomplish the same purpose, the producing large plastic slabs of luminescent material has proved to be a great boon to a variety of cosmic-ray experiments on the surface of the earth and underneath. On the
possibility of

other hand, the adaptability of scintillators to small dimensions

a valuable asset in conducting measurements at high altitudes. Gerenkov Counters. We have already had occasion to introduce the principle upon which operation of this intriguing device is based [see Fig. 2-13 and Eq. (2-24)]. The geometrical arrangement of a Gerenkov detector may be quite similar to a scintillation counter setup, hence Fig. 3-3 serves to illustrate the basic elements. In this case, however, the light output corresponding to the same energy loss is very much lower. But this disadvantage is far outweighed by the unique directional and threshold properties of Gerenkov counters. For Eq. (2-24) says that the radiation is emitted in a cone determined by the direction of motion of the particle that induces it. Thus we can easily establish, for example, whether particles are traveling up or down, a matter of considerable importance in carrying out cosmic ray measurements within the atmosphere. Furthermore, Eq. (2-24) tells us that no radiation is emitted if the particle velocity is lower than Anin = 1/ W The Cerenkov light output varies as the square of the charge of the particle, just as in the case of ionization. Gerenkov counters
is
-

50
in solid, liquid

COSMIC RAYS

and gaseous forms are

utilized in cosmic ray ex-

periments, and the parallel with the advantages of scintillation

detectors
is

striking.

VISUAL DETECTORS
This family
also comprises

two genera. The species of one are

for later study.

various devices in which the incident radiation produces visible

transitory effects that are
cies of the

photographed

The

spe-

other are different types of media that incur perma-

nent

effects as a result of

bombardment. These

effects

can be

seen visually, and can be analyzed to obtain information about

the radiation that produced them.

Cloud Chambers. When a gas containing a saturated vapor is suddenly expanded, a state of supersaturation results. This happens because the cooler gas after the expansion cannot normally hold as much vapor as it did when it was warmer. In the case of
the earth's atmosphere, the ensuing condensation of supersatu-

rated water vapor

on dust

particles creates droplets that

we

rec-

ognize as fog. For almost a half century, this principle has been
the basis for a most fruitful

method

for studying cosmic rays

by

observing their tracks.

As is shown in Fig. 3-4, a dust-free chamber containing a gas and a saturated vapor is expanded adiabatically (without loss or gain of heat), for example by the rapid motion of a diaphragm, to produce supersaturation. Ions created by the passage of charged particles then serve as condensation centers upon which the droplets form. These grow sufficiently large so that, under suitable
illumination, they can be photographed.

can be determined by counting the and interactions in the gas can be studied. In fact, all of the processes enumerated in Chapter 2 (except Cerenkov radiation) can be observed, and the characteristics of the particles
specific ionization

The

droplets,

determined therefrom. Various absorbers can be inserted in the

chamber

to reveal the energy losses

and ranges of

particles, as

well as the emergence into the gas of secondaries that they pro-

duce in traversing solid materials. Stereoscopic views, obtained by taking two photographs simultaneously from different angles,

EXPERIMENTAL METHODS
Rubber Diaphragm

51

Perforated Plates

Compressed Gas Inlet

^
yy,

"*"

Magnetic -Release-Valve

Stereoscopic

Camera

Gas -Vapor
Mixture

Escaping

Gas

FIG.

3-4

Basic

elements

of

Wilson

cloud

chamber.

Light

sources
field

and

GM

counters for controlling the expansion are not shown.

magnetic

directed

perpendicular to the glass

window may be

applied for

momentum measurements.

provide three-dimensionality.

magnetic

field

may be

applied

to deflect the particles, in order to determine charge

and mo-

mentum.
sion

and most important, is the advantage that the expanbe triggered by any of the arsenal of electrical detectors placed inside or outside. Some typical cloud chamber
Finally,

may

photographs are shown in Plates II and III. The Wilson cloud chamber is not continuously sensitive, and there is a dead time between expansions in order to allow equilibrium to become reestablished.

The

counter-controlled chamber
to chance.

assures that photographs of the events of interest are obtained,

whereas random expansion leaves the selection A variant of the expansion chamber is the chamber, in which supersaturation is attained containing the saturated vapor to fall under

warmer

to

diffusion cloud by allowing gas gravity from a a cooler region. Maintaining a thermal gradient bevessel creates a sensitive region

tween the top and bottom of the

52
in

COSMIC RAYS

which conditions are just right for the condensation of vapor on ions. Since this region is shallow, this type of cloud chamber must be operated horizontally. Consequently, although ideal in
accelerator experiments for

many

reasons, including
is

its

continu-

ous sensitivity, the diffusion chamber

studies.

not used in cosmic ray

Similarly, the bubble

chamber embodies tremendous advan-

tages in laboratory high-energy physics experiments because of

ability to compress a great amount of matter into the volume. This device is essentially the inverse of the cloud chamber. In this case gas bubbles form on ions in a liquid when the liquid becomes superheated. To accomplish this, a suitable liquid is heated above its normal boiling point, and pressure exceeding the saturation vapor pressure is applied to maintain
its

unique

visible

it

in the liquid phase.

means

similar to that used for a

heats the liquid.

are ions,

and the

Rapid reduction of the pressure, by some Wilson cloud chamber, superBoiling occurs only at the points where there tracks of the bubbles grow to visible size and

are photographed.

Spark Chambers. Although the diffusion and bubble chambers

did not succeed their parent, the Wilson cloud chamber, in cosmic
ray studies, the youngest of the chamber clan has
substantial impact because
it

made

a very

seems to combine the best of both worlds. It incorporates excellent resolution both in time and space while still retaining visual capabilities, hence it can serve
essentially as

an

electrical

counter and a visual detector at the

same time.

The
signal

spark chamber comprises a stack of parallel plates confilled with an atmosphere of a noble gas. Upon from controlling counters, a high voltage is very rapidly

tained in a vessel

applied to alternate plates. Electron avalanches quickly start

wherever there are ions produced in the gas by any charged particles that were present during the sensitive time. The ensuing bright sparks are photographed. An ingenious method for handling spark-chamber data in cosmic-ray experiments makes use of the noise produced by the sparks. From measurements of the time for the sound emanating from the spark between each pair of plates to reach suitably

EXPERIMENTAL METHODS

53

placed tiny microphones, the coordinates of the spark can be

specified.

Thus, thanks to

this true audio-visual aid, the trajec-

tory through the sonic spark

chamber can

easily

be reconstructed

and the
balloon

analysis carried out electronically, even in

flight.

Digitized systems, in which a large

an unmanned number of indi-

vidual spark or glow discharge modules are juxtaposed to define a three-dimensional grid, are also utilized.

Photographic Emulsions. Although the use of photographic emulsions to detect ionizing radiations dates back to the discovery of radioactivity, the renaissance of this technique in its modern form took place only a little more than two decades ago. Then, the range of usefulness was vastly expanded by the introduction of emulsions having sufficient sensitivity to record any individual cosmic ray, rather than just heavily ionizing radiations.

nuclear emulsion consists of a high concentration of silver

halide crystals

particle penetrates this

crystals

embedded in a matrix of gelatin. When medium, it activates many of

a charged the

AgBr

along

its

path, just as photons render the grains develop-

able in ordinary photography. These are converted to Ag grains during subsequent processing to form a track that can be studied under a microscope (cf. Plate X).

The most sensitive nuclear emulsions respond even to singly charged minimum-ionizing particles. The combination of continuous sensitivity, very high stopping power, and high angular resolution, together with important gains of a practical nature,
make them
especially suitable for many types of observation, especially at high altitudes. Stripped emulsions, or pellicules, of thicknesses up to several millimeters are available and, by piling

a number of these layers together, emulsion stacks, with sensitive volumes up to several liters, have been constructed. Here again, all of the knowledge about particle interactions summarized in Chapter 2 is brought to bear to interpret the emulsion tracks. From measurements of a number of different
esses, all

track parameters that are related to independent collision procof the characteristics of the particles that produced them

can be unraveled. One shortcoming in some experiments

is

the absence of a time

54
scale

COSMIC RAYS
during exposure. But even
this

disadvantage has been over-

come by moving one emulsion with respect to another at a known speed, and relating the time when a given track was registered to
the difference in
its

coordinates in the two emulsions.

Dielectric Solids.
doxically,
is

The baby

of the visual detector family, para-

granddaddy of them all! Ancient stored tracks of heavily ionizing cosmic rays have been found on meteoritic mineral detectors (Plate IVa) that have been collecting data for
really the

eons (^> 10 6 years). In addition to these fossil tracks, isotopes that were not originally present in meteorites arise from nuclear disintegrations produced by cosmic rays. Analysis of the contents
of these extraterrestrial cosmic ray meters provides valuable in-

formation about both cosmic rays and the meteorites themselves. Although we will hear the fascinating story that the meteorites

have to

tell later, it is

dielectric solid detector as

appropriate at this point to recognize the an experimental technique.

Particle tracks in solids consist of linear regions that have suf-

fered intense radiation

result of

damage from very dense

ionization, as a

which they are preferentially dissolved by a suitable chemical reagent. Meteoritic minerals can be "developed" by solutions of hot (>200C) KOH or NaOH. A latent track, which is optically invisible but which is revealed by etching, is formed whenever the primary specific ionization rate ; p of a slowing

down

particle exceeds a critical value jc that is characteristic of the particular solid. The value of jc is low for materials
(/3

<c

1)

such as mica,

glasses,

and various

plastics. It is

higher in abundant

meteoritic materials, such as plivine, hypersthene and diopside. Plastic sheets can be used as dielectric track detectors for study-

ing the present-day cosmic radiation.

When

plastic detectors ex-

posed to primary cosmic rays are immersed bath, the reagent attacks the chemically altered material along the trajectory of the particle, thereby producing a hole which is
in a suitable etching
visible in a microscope.

Plate IVb is a photograph of cosmic ray tracks in a 250ju, thick disk of cellulose nitrate exposed for seven days in a polar-orbiting
satellite.

EXPERIMENTAL METHODS

55

SYSTEMS
The seemingly extravagant statement that practically all permutations and combinations of the detectors we just described have been used in cosmic ray experiments may not be too far
from the truth. Now, we will cite a few examples of composite systems to introduce certain standard observing apparatus which

we

will encounter later,

and

to illustrate the kinds of things cos-

mic-ray physicists do.

By operating two or more sensors in a we can identify events in which both were actuated simultaneously. Of course, the output signal produced by a single detector when it fires persists for a finite
Telescopes.

Meson

coincidence arrangement,

length of time. Consequently, it is possible for pulses from the individual counters comprising a coincidence train to overlap fortuitously when they are traversed by unrelated particles
that happen to arrive within the instrumental resolving time. After taking into account the probability of chance coincidences
(accidentals), a coincident event can be ascribed either to a single particle that triggered each of the detectors, or to associated particles that

struck the individual detectors at the same time.

and oldest such arrangement is the Geiger-Muelcounter telescope. Its principle of operation can be understood by considering the cubical meson telescope illustrated in Fig.
simplest
ler
3-5.

The

by

many

This standardized geometrical arrangement has been adopted cosmic ray observatories for continuously recording the

intensity of the predominant secondary cosmic-ray component at ground level, the muons. It is clear that to be counted, single particles must arrive within a cone determined by the dimensions of the sensitive areas and the spacing between them. The opening angle is 90 in this case. Lead absorber interposed between the trays eliminates most of the electron component, accounting for the name by which the instrument is known.

The

analogy with the optical telescope refers only to the

re-

striction of the field of view.

Of

course, cosmic-ray telescopes can,

in principle, be designed with any desired angular resolution,

and

can be pointed in any direction.

56
Counted

COSMIC RAYS

Not Counted

FIG. 3-5

Arrangement of a meson telescope. The cubical configuration has been

adopted by many cosmic-ray observatories.

Air Shower Detectors. Suppose you were

plane.

to spread

out a num-

ber of counters widely separated from each other in a horizontal

What would happen? Would

what you would
predict,

the coincidence counting

rate be just

knowing the resolving times

and the individual counting

is

a resounding "no!".

expected

rates? The perhaps surprising answer The counting rate would be higher than from chance coincidences. The hint was dropped in

Chapter 2, when the term "shower unit" was casually mentioned. And, by inductive reasoning, one might have concluded from the discussion of energy loss mechanisms for particles and electromagnetic radiation that the inverse processes of bremsstrahlung and pair production could result in the rapid conversion of the energy of a single particle into many other energetic progeny. This does occur (cf. Plates II and III), and we shall explore the fascinating consequences of this multiplication mechanism later. The resulting phenomenon is called an extensive air shower, EAS. The size qualification is added to distinguish an electro-

EXPERIMENTAL METHODS

57

magnetic cascade shoxoer, extending over a large area, from a

local shower,

which

is

a nucleonic

phenomenon

that covers a

small area.

Extensive air shower detectors run the gamut of

size

and com-

plexity from a few single counters close together to most elaborate

arrays of different types of instruments distributed over extremely
large areas. Sometimes, parts of the system

below the earth's

surface.

New

techniques for studying

may be buried deep EAS

and interest is likely to continue at an accelerating pace, because extensive air showers provide the only available means for studying the very highest
are continually being developed,

energy cosmic-ray primaries.

Calorimeters. Instead of waiting for showers to be produced in the atmosphere, one can create a spatially scaled-down version in

dense materials, so that the entire cascade process transpires within a relatively small volume. The total energy dissipated inside the system can then provide a measure of the energy of the incident primary. An instrument that accomplishes this is
called a calorimeter
Fig.
3-6

for obvious reasons. shows one such arrangement that has been carried
a

Detector of the Charge

Section "A -A'"

Z
Proportional Counters

of Primary Particle

Detector of
Interaction

Detector of Energy

(Ionization

Calorimeter)
Photomultipliers

CH 2 E33
C
Fe

Scintillators

E52

1-1 OJ

Pb
FIG. 3-6

n
tons in the range 10 10

Cosmic-ray calorimeter designed to measure the energy of primary pro-

10"

eV.

by Soviet "Proton" sputniks. Plastic

tion detectors of the calorimeter.

scintillators are the ioniza-

The

interaction detectors meas-

ure the
detector.

number

of particles in the shower passing through the

Two

proportional counters measure the charge

of the

58

COSMIC RAYS

primary. Polyethylene and graphite niters periodically change

places during the flight and are removed out of the limits of the instrument solid angle. Measurements of primary protons in the energy range 10 10 10 14 eV are made with this arrangement. High-Flying Analyzers. Classical detectors, such as Geiger-

Mueller counters and ionization chambers, are still being flown aboard balloons and spacecraft. But a myriad of highly specialized combinations of sensors have also appeared on the scene. Their characteristics are continually being improved as experience with the new techniques is gained. Let us look at a few "typical" arrangements just to get a feeling for the infinite variety of possibilities that now exist for examining the fine structure of the primary cosmic-ray beam. Fig. 3-7 is an idealized view of a dE/dx and E detector. The
Accepted
Rejected

E-AE

FIG.

3-7

Simplest form of arrangement for measuring the rate of energy loss

(dE/dx)

and

kinetic

energy E with thin and thick

detectors.

scintillators

or

solid-state

thin upper scintillator (or semiconductor counter) A measures dE/dx. Particles that stop in the thick lower scintillator (or semiconductor counter) B are identified by anticoincidence with the plastic guard scintillator C. An event A + B C, in which

EXPERIMENTAL METHODS

59

A and B are activated and C is not, identifies a stopping particle. If A is sufficiently thin, of thickness Ax, the energy deposition is

**-*
whereas the energy deposited in

<3

-2

>

is

E-AEc^E,
since

(3-3)

AE

is

negligibly small.

Recalling Eq. (2-16)

we can

write

- (dE/dx) ion - zW) =

where the particle speed
is

z*g(E/Mtf*)
is its

(3-4)

fie,

incident energy and

its rest

mass.

The

function

g(E/M

c2 )

can be separated:
gl {M#*)gt(E).
(3-5)

g(E/M^) =
The
rate of energy loss
is

then
lz*gl (McP*)]gi(E).

dE/dx

(3-6)

The

simultaneous measurement of dE/dx and

determines
taken

the value of z2 g 1 (M c2 ),
into account.

which uniquely

identifies the species of

the particle. Actually, the finite thickness of

A can easily be

typical set of

AE

nuclides

is

shown

in Fig. 3-8.

E AE curves for several AE vs E AE line exists for A

vs
its

each type of stopping particle, permitting

ticles

identification. Par-

that

do not stop

in

deposit

less

energy in both

B and A

than do those that are stopped (Fig. 2-10). The dashed lines represent characteristic energy lines for particles that have penetrated B.

Now,

let's

consider a different system, in this case one designed

for investigating the

vertically incident particles are selected

primary electron component. In Fig. 3-9, by coincidence between

telescope counters

must

also trigger the gas

and T 2 (plastic scintillators). An event Cerenkov counter C x that is sensitive to

c 2 The threshold energy E > 20 lower than for protons or alpha particles, thereby making it easy to discriminate against the very large background. A dE/dx measurement is made in counter 1 and, since all

charged particles with energy

for electrons
is

60
i
1

COSMIC RAYS
1 1

40
-

35

|30
p

^
V)

25

-20
>. o>

i5

15

10

"

-He 4

50

100

150

200

250

300

350

Energy
FIG. 3-8

-A in MeV
and
on one of these
characteristic

Typical

Af

vs E-AE curves. The simultaneous measurement of energy

rate of energy loss yields a data point that falls

curves, thereby identifying the particle.

accepted particles are

to

relativistic, this sorts the particles

according

charge

z.

An

electron will develop an electron-photon shower

in the lead glass

Cerenkov detector

C2

to the lead glass cylinder, the total

If the shower is confined Cerenkov light output is

roughly proportional to the energy of the entering electron.

lf

and G 3 are anticoincidence counters to guard against certain unwanted events that could masquerade as electrons, and S discriminates against certain proton events that might otherwise fall between the cracks. Neutron Monitors. Important advantages that will become apparent later are gained by observing the nucleonic component even though it comprises only a few percent of the total particle flux near sea level. The neutron monitor is in some respects analogous to a nuclear reactor, and its front end is called a pile. Figure 3-10 shows the geometrical arrangement of the most mod,

G2

ern version of this apparatus that

is

continuously recording the

EXPERIMENTAL METHODS
-Telescope Counter T^

61

PM = Photomultiplier

-Gas Cerenkov Counter C<

'Telescope Counter

T2

-Lead Absorber
-Anti - coincidence Counter G^
'

Shower Counter S

-Guard Counter G 2

-Lead Glass Cerenkov Counter C 2

j""

Guard Counter G 3

mm
FIG. 3-9

Plastic Scintillator

System for investigating the primary electron component.

globe and,

cosmic-ray intensity at a network of stations which covers the literally, extends from pole to pole.

nucleon (proton or neutron) incident on the pile has a high probability of causing a nuclear interaction in the lead (the
producer) thereby yielding a multiplicity of low-energy neutrons

As these diffuse through the pile, they are slowed thermal velocity by elastic (billiard-ball) collisions in the hydrogen-rich material, polyethylene (the moderator). Some of the evaporation neutrons reach the B 10 F3 proportional counter tubes, and are detected as already explained [see Eq.
(see Fig. 4-16).

down

to

(3-1)].

LABORATORIES

A discussion of experimental methods would be incomplete without some reference to the many different types of "labora-


62

COSMIC RAYS

10

F, Counter

52 cm

T
Polyethylene
pile, the

O
tons.

B&g Lead

HH
FIG. 3-10

Arrangement of a

nucleonic intensity. Three identical

sensor of an apparatus for measuring the and independent sections and ancillary equip-

ment comprise a neutron monitor, weighing about 40

tories" that cosmic-ray instruments are wont to inhabit. A major part of the effort expended in carrying out any cosmic-ray experiment is devoted to transporting the apparatus to the desired location, and making certain that it operates reliably under the usually rigorous environmental conditions. Cosmic-ray research has been historically and traditionally an expeditionary enterprise. Since nature performs the experiments, we must carry out our observations in accordance with her dictates which generally means in some hard-to-reach albeit perhaps

exotic place.

Manned Stations. "Permanent" cosmic-ray observatories are continuously monitoring the cosmic-ray intensity at sites extending from megalopolis to the remotest corners of the earth, at depths or altitudes ranging from deep mines to tall mountain

EXPERIMENTAL METHODS
peaks.

63
of ground-based cosmic-ray stations are
to say, the logistical prob-

Some examples

depicted in Plates

and VI. Needless

lems are far from

trivial!

Mobile manned

stations are also utilized for conducting geo-

graphical surveys relating to the effects of the geomagnetic field

upon

cosmic-ray intensity. Large trucks (Plate VII), ships,

and

air-

craft carry the touring cosmic-ray detectors.

Balloons.

The importance

of lifting cosmic-ray detectors to

it is

great heights should already be evident. Thus,

no mere

co-

incidence that cosmic ray physicists have been instrumental in the rapid development of ballooning techniques. Plastic balloons (Plate VIII) with volumes of 20 million cubic feet can serve as
stable platforms at altitudes of 145,000 feet for payloads of

hun-

dreds of pounds.

In an isothermal atmosphere, the pressure p varies with height above sea level h 8 exponentially:

p
where p
is

po exp (rngh s /kT ) K

(3-7)

the pressure at sea level,

is

TK

is

the temperature in de-

grees Kelvin, k
lecular mass.

the Boltzmann constant,

and

H = kTK /mgis the

is

the

mean mo-

local scale height, or the height

of the

homogeneous atmosphere, that is, the thickness of the atmosphere if the density were constant at the sea level value

(H = 8 km). Of course, the temperature is not constant throughout a vertical column, hence Eq. (3-7) is an approximation. The so
called "standard" atmospheric depth vs altitude curve in Fig. 3-11

shows that only

feet,

1%

of the atmosphere lies above about 100,000

and the residual atmosphere at 145,000 feet is less than 2 gra/cm 2 practically outer space for many purposes, but much too deep for others. Rockets. For a relatively brief period, before the advent of spacecraft, rockets were employed to boost cosmic-ray instruments

to greater heights
tive techniques

than could be attained with balloons. Imaginasuch as launching the rockets from balloons were
still

developed. Although rockets are

extensively utilized in upper

atmospheric physics research (aeronomy) for which they are especially well adapted because they can provide vertical profiles, cosmic-ray workers now concentrate upon balloons, satellites

64
Altitude
in

COSMIC RAYS
1Crft
p

20 40 60 80 100120140

1000 600

\ -V
i

i-p

1000

600
1

300
100

\|

300
100

60
30
10
-

60
30

6
3

10

20
in

Altitude
FIG. 3-11

30 km

40

Variation of atmospheric depth with altitude above sea level.

and space-probes, using

ticles

rockets only for studying energetic paremitted during distinctive solar events.

Spacecraft. Recognition of the advantages of artificial satellites

experiments initially provided one of the strongest motivations for developing scientific space vehicles. The best known space age discovery bears the name of A. Van
J.

as vehicles for cosmic-ray

an apparent malfunctioning of his single Geiger counter on Exporer I (Plate IX) to an abnormally high particle flux at the higher altitudes.*
correctly attributed
* It is interesting to

Allen,

who

earlier

had indicated that charged

field.

note that Stormer's calculations (Chapter 4) many years particles could be trapped in the earth's

magnetic

Propagation Through the

Earth's Magnetic Field

and Atmosphere
Non semper ea
For the soul
sunt quae videntum.

PHAEDRUS
is

dead that slumbers,

And

things are not what they seem.

HENRY WADSWORTH LONGFELLOW

The

interpretation of any cosmic-ray

measurement carried out

in the immediate vicinity of the earth requires an understanding of the perturbations introduced

by

its

presence.

The

effects

upon
of

the free-space particle flux of the earth's magnetic field,

its

and

atmosphere, and of the solid earth itself, must be taken into account. Conversely, these same effects can be exploited to serve useful purposes in our efforts to understand cosmic-ray phenom-

ena. Nature compensates

somewhat

for

the tight control that

she exercises over cosmic-ray experiments by providing us with a

huge and unique cosmic-ray apparatus our planet! The geomagnetic field serves as a vast magnetic spectrometer. One consequence of the deflection of electrically charged particles by magnetic fields is a filtering effect whereby particles that fail to satisfy a certain minimum requirement (depending upon
location, etc.) are not admitted.

The

price of admission, so to
is

speak, becomes progressively lower as the pole

approached,

giving rise to the latitude

restrictions

effect, as illustrated

in Fig. 4-1. Fur-

thermore, the field constitutes a complicated lens which imposes

on the

directions of viewing.

65

66

COSMIC RAYS
Atmospheric Depth = 650 gm /cm 2

300

200

Cosmic-Ray Equator

100
40
20

20

40

i_ 60

North South Geographic Latitude

FIG. 4-1 Typical intensity vs latitude plot of data obtained with an airborne neutron monitor. The exact form of the latitude-effect curve depends upon the nature of the detector, the atmospheric depth, the level of solar activity, and the

route followed.

An optical or radio telescope detects electromagnetic radiation which, unaffected by magnetic fields, travels from the source to the observing instrument in a straight line. But this is not true
in the case of a cosmic-ray telescope.
ticles to

which

it

The electrically charged parresponds undergo various gyrations in traveling

toward the earth, and the original direction of approach far away from the earth may be quite different from the direction in which the instrument is pointing. Figure 4-2 shows the trajectory of one particular proton as it approaches the earth. It is clear that the direction from which it appears to come when it ultimately
strikes the earth

the geomagnetic
typical,

but

it

original direction prior to encountering This quasi-trapped orbit is not necessarily serves to indicate the importance of carrying out an
is

not

its

field.

elaborate

analysis

to

evaluate

the

directional

response

of

cosmic-ray detector.

The atmosphere
into other species,
as

not only transforms the incoming particles

in different directions, but
it

moving

also acts

an absorption filter which, in effect, imposes a certain minimum primary energy requirement, the atmospheric cutoff, depending upon the depth in the atmosphere, the nature of the secondary component that is being observed, and the geometry of

THROUGH THE EARTH'S MAGNETIC

FIELD

AND ATMOSPHERE

67

FIG. 4-2

Quasi-trapped trajectory of a proton moving under the influence of the

earth's magnetic field after arriving

particle eventually strikes the earth,

from a distant source.

its

It

is

clear that

if

this

apparent arrival direction would not corre-

spond to

its

original direction of approach.

the detector. This accounts for the counting rate plateau the knee of the latitude effect in Fig. 4-1. Thus, the combination of the earth's magnetic

beyond

field,

the atmos-

and the detector (including any additional superposed absorber, as in the case of underground measurements) comphere,
prises

multi-component optical system. Exceedingly compliutilize it

cated analysis with a voracious appetite for computer time must

be painstakingly carried out in order to enable us to

effectively for cosmic-ray investigations.

MAGNETIC BENDING
When
magnetic
ship
a charged particle
field

moves with

velocity

v in a uniform

B,

it

experiences a force according to the relation-

m(dv/dt)

(ze/c)(v

B).

(4-1)

In general, v has components perpendicular and parallel to B, and the path of the particle is a helix, as is illustrated in Fig. 4-3. The radius r is called the gyroradius or cyclotron radius. g
If

the gyroradius
field,

is

small compared with the dimensions of the

magnetic

the particle essentially follows a line of force

68

COSMIC RAYS

FIG. 4-3

Motion of a charged particle

small, the particle
is

in

a magnetic

field.

If

the gyroradius r g

is

constrained to follow a line of force.

which defines the direction of motion of point about which the particle executes
If

its

guiding center

the

its

cyclotronic motion.

is

perpendicular to

v,

the force

is

to both,

and

the particle describes a circle.

is

mutually perpendicular As in the mechanical

analogy, the centripetal acceleration

(dv/dt)

v*/r g

(4-2)

Hence, combining Eqs.

(4-2)

and

(4-1) gives

mv 2 /rg
or

{ze/c)vB

(4-3)

Br &

mvc/ze

pc/ze,
[cf.

(4-4)
(2-2)], e is

where p
units of

is

the relativistic

momentum

Eq.

the charge

of the electron,
e.

and

z is the

charge of the moving particle in

The product
curvature
straight,

of the magnetic-field strength

and the radius

of

Brg

is

the magnetic rigidity of the particle. After in-

troducing the appropriate conversion factor to keep the units

we have
p

?>QQ>zBrli ev/c

(4-5)

THROUGH THE EARTH'S MAGNETIC

where Brg
is

FIELD

AND ATMOSPHERE

69
of any

measured in gauss-cm.* Thus, the

momentum

charged particle can be determined by measuring its radius of curvature in a known magnetic field. In discussing the motion of cosmic rays in magnetic fields, it is
convenient to characterize a particle by its magnetic rigidity, the quantity P = pc/ze in Eq. (4-4), expressed in volts, rather than by its momentum in ev/c or energy in ev particles having the same magnetic rigidity P follow identical trajectories, whereas
;

particles with the

same

momentum p

or energy

but different

charges ze do not.

Magnetic Analysis. To understand how the earth's field acts magnetic analyzer, let us consider the oversimplified situation depicted in Fig. 4-4, before proceeding first to an idealized
as a

Equator

West

North
FIG. 4-4

Schematic representation of the concept of allowed directions of arrival.

lines

The solid

are for positive particles. As illustrated by the dashed

picture
is

line,

the

reversed for negatively charged particles.

approximation and then on to the vastly more complex problem presented by the real field of the earth. Suppose that an observer can measure the particle flux from any direction at any point in the lower plane. As in the case of the earth's surface, no particles come up from below. Between the two planes, there is a magnetic field in the direction indicated by the arrow. The field is uniform over any section parallel to the front of the dia*

Because the permeability in space

is

netic field intensity

oersted)

(cgs unit-gauss)

unity, the distinction between magand magnetic induction B (cgs unit-

may be

ignored and the gauss (or

gamma =

10

-5

gauss)

is

often used.

70

COSMIC RAYS

gram, but of increasing strength as shown. The line 0-0" is like a magnetic meridian in the northern hemisphere, 0" being at the magnetic equator where the horizontal field is strongest. West is

on the

rigidity

marked. Positive particles with a certain magnetic on the upper plane uniformly from all angles. Those coming from the eastern horizon reach point at
right, as

are incident

angle

0,

0' .at

Only
left

particles
6,

of

the more inclined angle, 6', but fail to reach 0". with higher magnetic rigidity can reach the regions whereas those cosmic rays that are able to reach the

region on the right illuminate it uniformly. For incident negatively charged particles, the picture would be reversed, as indirigidity

cated by the dotted orbit. If particles having only the specified were present, the observer would find a sharp cutoff in the intensity at some measured angle $ and, knowing the strength

and direction of the magnetic field, he could then determine both the sign of the charge and the magnetic rigidity. Orbits in Dipole Field. The simplest possible model of the
earth's
tion.

magnetic field is the so-called centered dipole approximaIn this case, the field is assumed to be the same as that

which would be associated with a bar magnet of magnetic molocated at the center of the earth, and oriented from north to south along the geomagnetic axis. The geomagnetic axis is tilted by 11.5 with respect to the geographic axis, so that the geomagnetic poles emerge at 78.5N, 69.0W and 78.5S and 111.0E. Note that the geomagnetic coordinates are different from the magnetic coordinates which are plotted on navigational charts, and which represent local magnetic measurements of dip

ment

angle, etc.

Stormer addressed himself to the problem of determining the regions of space around a dipole that are accessible to particles of any specified magnetic rigidity. Fortunately, a detailed discusnot essential for understanding the we should note that the analysis involves the solution of the equation of motion [Eq. (4-1) exis

sion of Stormer's theory

nature of

its

consequences. But

pressed in cylindrical coordinates] of a charged particle moving in the field of a dipole. A simplifying normalization that removes
the magnetic

moment

of the dipole

M D and the particle rigidity

THROUGH THE EARTH'S MAGNETIC

FIELD

AND ATMOSPHERE

71

P from

the equations

is

achieved by introducing as a unit of

length the so-called Stormer unit:

C = (30(WD/P)icm.
The Stormer
ity

(4-6)
it is

unit has a physical significance:

the radius of

the circular orbit that

would be described by a

particle of rigid-

P volts moving in

the equatorial plane of a dipole of magnetic

moment M D gauss-cm. The dipole moment of the earth is D = 8.1 X 1025 gauss-cm, and the radius of the earth is a = 6.38 X
10 8 cm. Therefore,

on the
10 8

earth's surface at the equator

where

a=

C:
6.38

[(300) (8.1

10 25 ) //>,]*.

(4-7)

Solving for
that

P we
s,

find that the magnetic rigidity of a particle

is

would

just circle the earth at the equator

P =
a

59.6

GV.*
particle

(4-8)

This
i.e.

is

the

minimum

rigidity

which a

must have

to

reach the earth from the least-favored direction at the equator,

the eastern horizon for positive particles.
It is

of allowed

probably intuitively obvious that the detailed calculations and excluded regions at any point around a dipole
of the results of Stormer's analysis
is

can become rather involved.

An example

shown in

having a magnetic rigidity P = 10 GV cannot reach the geomagnetic equator, but begin to arrive within a narrow cone about the western horizon at some critical latitude. The cone broadens out until, at another higher latitude, it fills the entire hemisphere. The Stormer treatment does not take into account complications introduced by the presence of the solid earth. Numerical calculations of individual orbits carried out by G. Lemaitre and M. Vallarta showed that, in fact, not all trajectories within the Stormer cone are allowed even in the absence of the earth; there is a penumbral region containing a mixture of accessible and inaccessible points. Furthermore, there are some otherwise
Fig. 4-5. Positive particles
* GV = 10" volts. This avoids international misunderstanding, as compared with the older use of BV, since in some countries billion means 10 12
.

72

COSMIC RAYS

FIG. 4-5

Allowed

(light)

and forbidden (dark) regions

for 10

GV

particles.

allowed orbits that intersect the earth. This (obviously) stops the particles, and thereby produces the simple shadow cone within
the Stormer cone. Figure 4-6 illustrates this effect which
is

con-

centrated in the direction of the nearer poleward horizon.

The

character of allowed (light) and forbidden (dark) regions

of the sky at an intermediate latitude, 30N, for positive particles of a

given rigidity (10

GV)

is

depicted in Fig.

4-7.

The

Stormer allowed cone includes the entire portion of the hemisphere west of the arc that is so labelled. Absolutely no 10 GV

Geomagnetic
Pole

FIG. 4-6

Effect of the solid earth in

producing the shadow cone.

THROUGH THE EARTH'S MAGNETIC

Shadow Cone

FIELD

AND ATMOSPHERE

73

orb7
Main Cone

Stormer Cone
FIG. 4-7

Stormer, shadow, and main cones for 10

GV

particles at

a middle geo-

magnetic latitude (30 N).

particles

the allowed cone

can reach the regions lying outside this cone. Part of lies in the earth's shadow. All of the main cone,
arc, is accessible to

west of the indicated

10

GV

particles. Be-

tween the boundaries of the main cone and the Stormer cone lies the penumbral region, characterized by alternate light and dark bands. An important concept in the application of the theory of geomagnetic effects to the interpretation of cosmic-ray measurements was introduced by W. F. G. Swann. He showed that as a consequence of Liouville's theorem, well known in statistical mechanics, the particle flux
is

the same at all allowed places,

and

zero elsewhere.

Let us

now
we

consider what happens

if

region of the sky, such as the zenith.

siderations,
shall detect all particles

we look From the

having

at a particular

foregoing con-

rigidities exceed-

The
if

ing some lower limit, the geomagnetic cutoff or threshold rigidity. flux of these particles will be the same as it would have been
the magnetic field were not present.
is

As our

vertically pointing

detector

moved toward

the pole from the equator, the threshold

rigidity decreases, as seen in Fig. 4-8. Here, the vertical cutoff

determined by both the Stormer cone and the main cone are plotted. At a given geomagnetic latitude, no particle with less than the Stormer cutoff can enter in the vertical direction, but all particles having rigidities exceeding the main cone cutoff

74
15

COSMIC RAYS

Main
10

Cone

Penumbra

% 5
S1forrr er C one

10

20

30 40 50 60 Geomagnetic Latitude
rigidity

70

80

90

FIG.

4-8

Variation of vertical threshold

with geomagnetic latitude. All

particles

having

rigidities
all

allowed, whereas

exceeding the main cone cutoff at a given latitude are particles with rigidities below the Stormer cutoff are forbidden.

particles with rigidities between these two depending upon their individual trajectories. The Stormer integral, which is the solution of Eq. (4-1), reduces to a simple expression for the Stormer threshold rigidity
access.

have

Some

limits

may

enter,

for vertically-incident particles

P
A,,

P*

300AfD
4a 2

cos 4

14.9 cos 4 X e

GV

(4-9)

where A g is the geomagnetic latitude. Note in Fig. 4-8 that the complicating effects are absent at both the equator and near the poles, as one might expect from symmetry considerations.
Trajectories in Real Geomagnetic Field. Although the validity of the classical approximations based upon the centered dipole,

or on the eccentric dipole shifted 342 km from the geometric center of the earth, were unquestioned for more than two decades, evidence to the contrary then began to accumulate as im-

proved experimental techniques started

cient precision to reveal discrepancies.
first

to

provide data of

suffi-

The

realization that the

order representation of the

much more

complicated

real-life

THROUGH THE EARTH'S MAGNETIC

terrestrial

FIELD

AND ATMOSPHERE

75

magnetic
last

field

is

insufficient led to

a succession of

attempts in the

few years to improve the cutoff calculations by correcting for nondipole components of the earth's field in

one way or another. Each new model seemed

crepancies
cessive

to reduce the dis-

among

the then available data.

The

nature of suc-

improvements is illustrated in Fig. 4-9, which shows how intensity measurements recorded in a world-wide airborne survey became progressively better organized by several generations of
threshold rigidity calculations.

The

culmination of
is

all efforts to

predict cutoff rigidities theo-

retically

the detailed calculation of actual cosmic ray trajec-

most accurate mathematical representation of the magnetic field that geomagneticians can derive from the extensive measurements of the magnetic field components all over the globe. This is achieved by a procedure called spherical harmonic analysis. The sound produced by a vibrating string, for example, is the
tories in the

earth's real

sum

of the contributions of the various harmonic frequencies of

modes of oscillation that may occur in addition to the fundamental. Analogously, the earth's magnetic field configuration is much more complex than that surrounding a simthe higher

must therefore be represented by a sum of terms that is, contributions by quadrupole, octopole, etc., must be included. To see what
ple dipole,
it

and

dipole

plus higher order

a a
is

involved, let us look at a formidable equation that at least suggests the

complexity of the computational problem.

V(r,
9,

The magfrom

netic potential,

from which the sources within the earth can be derived is:
$),

field

arising

V(r, 6,

4>)

=
a
r

n=l

mQ

(a/r) n+1 (g nm cos

m<f>

h nm sin m<f,)P nm (cos 0),

(4-10)

where

<

Pnm (cos 6)
The

= = = = =

radius of earth, distance from center of earth,

longitude, east,
colatitude (90

geographic latitude),

associated Legendre function.

gauss coefficients gn m and h nm are constants to be deter-

400
360 -

320

"

280
f
**>
o
c

240

"

c
+<B

~ 200
_o
'c

J%&

o o 160 <u

2
120
"

80

40

10

12
in

14

16

18

20

22

Threshold Rigidity

GV

(Dipole)

400
360
320

280

240
in

o>

~200
_o
'c

o o 160
a>

Z
120

80

40

8
in

10

12

14

16

18

20

22

Threshold Rigidity

GV (Semi-empirical Approximation)

THROUGH THE EARTH'S MAGNETIC

400
360
-

FIELD

AND ATMOSPHERE

77

320
280

240

~ 200
"5

160

120

10
in

12

14

16

18

20

22

Threshold Rigidity

GV

(Trajectory Computations)

FIG. 4-9 Intensity measurements, obtained with an airborne neutron monitor in a world-wide survey, plotted as a function of geomagnetic threshold rigidity. The

abscissas represent three generations of cutoff calculations: (a) dipole field model,
(b) semi-empirical

approximation to the real

field,

and

(c)

interpolation of a world

grid of trajectory computations including spherical harmonic terms

up to the

sixth

degree.

mined from
series of
fit

the observational data. Field models described by a

120 spherical harmonic coefficients that

say,

make Eq.

(4-10)

the global magnetic measurements best have been derived.

Needless to
precision

trajectory calculations having the requisite

consume a prodigious amount of computer time. Nevertheless, calculations including the spherical harmonic terms in Eq. (4-10) up to the sixth (n, m = 6) and, in some cases, the eighth degree field simulation have been carried out. The procedure is to shoot (analytically) from the earth negatively charged particles having various rigidities and to see whether they succeed in escaping to "infinity." Those that do correspond to positive parti-

78
cles

COSMIC RAYS
afar.

approaching the earth from

is

At

a given point

earth, for a given direction, the lowest rigidity for

on the which the

orbits connect with infinity

the geomagnetic threshold.

effects

Of

which make it impossible to define a sharp cutoff at middle latitudes. This is taken into account at least in an approximate way to determine the effective threshold. Figure 4-9 shows the improvement
course, this

method brings out the penumbral

achieved with this ultimate treatment of the problem.

These precise trajectory calculations enable us to make the most effective use of the earth's field as a magnetic analyzer for investigating how various cosmic-ray phenomena vary with rigidity. But they also are prerequisite for studying spatial anisotropics. For this purpose, we need to relate the observed particles to the points on the celestial sphere from which they have
actually

come

the asymptotic direction in

Fig. 4-10.

Asymptotic
Direction

Geomagnetic Axis

Trajectory

-Impact Point it Geomagnetic Field Were Not Present

Equator

FIG. 4-10

Definition of asymptotic direction. The sketched trajectory

is

schematic

only.

E. A.

Brunberg made the

first

determinations of asymptotic

directions with a sort of analogue

A movable electron gun is mounted on

computer called a terella.* a model magnetized earth

contained in a vacuum system. Electrons are shot at various

* The terella has played an interesting part in the study of the earth's magnetism. See Magnets by L. W. McKeehan, momentum Book No. 16, Van Nostrand, Princeton (1967).

THROUGH THE EARTH'S MAGNETIC

angles from different positions,

FIELD

AND ATMOSPHERE
their

79
dis-

and

impact points at

tances far from the globe are observed by the

in Fig. 4-11.

method sketched

Terrello

Electron

Beam

Fluorescent

Screen

FIG. 4-11

Scheme

for determining asymptotic directions with

terrella.

Electrons

are shot from a scaled magnetized model earth. The

shadow

of a tooth produced

by impact of an electron beam emitted with a given energy and direction appears on the fluorescent screen. The trigonometric relationships among the quantities m,
n,

and

give the direction of arrival at essentially an infinite distance from the

earth.

Figure 4-12 shows the asymptotic directions for a couple of determined by trajectory calculations. Each dot represents the point where the only particles of the indicated rigidity
stations as

that can reach the station in the presence of the geomagnetic

field

would have

The

hit the earth if the field were absent. asymptotic cone for a particular cosmic-ray detector

is

the

region of the celestial sphere from which those particles making the principal contribution to its counting rate have come. The
position

and shape of this cone of acceptance depend upon both the primary rigidity spectrum and the details of the coupling

between the primaries and the secondaries that are ultimately counted. Figure 4-13 shows some typical asymptotic cones for neutron monitors. This diagram reveals the crucial importance of

80

COSMIC RAYS

FIG. 4-12

Asymptotic arrival directions for a polar and a middle latitude station.

all particles

Each dot represents the point on the celestial sphere from which
indicated rigidity
(in

of the

GV)

that can reach the station must have come.

stations in the polar regions; only at very high latitudes

is

it

possible

to

observe particles that have

come from

directions

steeply inclined to the plane of the earth's equator. Furthermore,

the cones of acceptance of stations even at intermediate latitudes

are considerably broader.

These considerations explain why cosmic-ray physicists are inand why they set up shop in such inhospitable places as those pictured in Plates V and VI.
veterate globe trotters,

TRANSFORMATIONS IN THE ATMOSPHERE

Having discussed
the effects of the earth's magnetic field

the motion of the incoming primary cosmic rays,

upon we can now

fate of a
is

consider the next phase in the

life

of those that are allowed to

wend

their

way

into our planetary atmosphere.

The

typical cosmic ray

when

it

impinges upon the atmosphere

de-

picted schematically in Fig. 4-14.

The

interactions described in

THROUGH THE EARTH'S MAGNETIC

80
North

FIELD

AND ATMOSPHERE

81

South

180 West
FIG. 4-13

80
Geographic Longitude

80

180 East

Typical asymptotic cones of acceptance of neutron monitors plotted on

the celestial sphere, but with geographic coordinates. The stations are located at
the points indicated by x.

Chapter 2 (including ionization, the effects of which are not represented in this diagram) start as soon as an appreciable amount of matter is encountered. Furthermore, the mean free
path for nuclear interactions of a very high-energy proton in air about 80 gm/cm 2 Consequently, direct observation of the primaries necessitates carrying the instruments to balloon altitudes, or even higher if their energies are so low that their ranges are less than the residual atmospheric mass at the ceiling for these
is
.

vehicles

(cf.

Fig. 7-3).

There are three modes whereby the energy

particles
is

of the incoming

transferred through the atmosphere to sea level

and

even below: via (1) the nucleonic or nuclear-active component, (2) the meson, or hard, component and (3) the electromagnetic, or soft, component.

Depending upon the magnitude

primary
is

of the energy with which the endowed, one of these three mechanisms for the con-

82

COSMIC RAYS

ri

Air

Shower Array

Neutron Monitor

Nucleon

Electromagnetic Or "Sott"

Meson Or "Hard"

Nucleonic

Component

Component

Component

N,P-High energy nucleons

n,p -Disintegration product nucleons o -Atomic nuclei of atmospheric constituents

FIG. 4-14

The principal modes whereby the energy of a primary cosmic ray

is

propagated through the atmosphere.

version of primary energy into a secondary

nates.

component predomi-

The

corresponding detectors, in order of increasing energy

(2)

of response, are: (1) neutron monitor, (3) extensive air shower array.

meson

telescope

and

High energy neutrons and protons, emitted as disintegration products of interactions of the primaries with atomic nuclei of the atmospheric constituents, give rise to the nucleonic component,

the lower atmosphere, the flux of this

idly

which then develops in a cascade process. Upon reaching component decreases rapwith atmospheric depth (Fig. 4-15), and comprises only a

THROUGH THE EARTH'S MAGNETIC

2000

FIELD

AND ATMOSPHERE

83

400
FIG. 4-15

500

600

700

Atmospheric Depth

800 900 in gm/cm 2

1000

Variation of the intensity of the nucleonic component with atmospheric

the lower atmosphere. The slope of the exponential curve depends upon

depth

in

geomagnetic threshold

rigidity.

few percent of the total particle flux near sea level. Nuclear interactions of low-energy primaries result mostly in nucleons as
disintegration products.
star, is

typical event in the chain, called a

course, the evidence for the emergnot visible because they do not produce ionization, but they do produce stars. At higher primary energies, ir mesons are also emitted in ad-

shown

in Fig. 4-16.

Of

ing neutrons

is

At very high energies, most of the energy is converted into pions in events called jets, illustrated in Fig. 4-17.
dition to nucleons.

Although every type of so-called elementary particle may be produced in the high-energy interactions, pions play the dominant
role in the propagation process. But, referring to

pendix),

we

see that the

it

Table A-l (Apmesons are unstable, and that they

rapidly decay into other particles.

charge.
that,

Charged pions turn into muons which carry on the original Although y, mesons are also unstable, Eq. (2-10) tells us
thanks to relativity, their lifetime
is

sufficiently

long for

84

COSMIC RAYS

25M

FIG. 4-16

Typical star, in which evaporation neutrons (not visible)

(black tracks) are produced

when an energetic proton (light ward interacts with a nucleus.

track) traveling

and protons down-

some of them to survive the journey to earth. I^n fact, muons are the most abundant type of cosmic-ray secondary at sea level (note that neutrinos are omitted from Fig. 4-14 because they are virtually undetectable).

THROUGH THE EARTH'S MAGNETIC

FIELD

AND ATMOSPHERE

85

25M

240/i

A
805/i

900/i

1,400/1

1,475/1

FIG.

4-17

Very high energy interaction, from which, among other products, mesons emerge in a narrow angle jet. The primary is multiply charged.

pi

Neutral -n mesons decay into y rays, which, by a succession of electromagnetic processes described earlier, evolve into a great

many

particles

ticles is

extending over a large area. The number of parapproximately proportional to the primary energy (cf.

Fig. 4-18).

86

COSMIC RAYS

Some of the electrons arising from the decay of ^ mesons may be sufficiently energetic to initiate showers, and some -n mesons may be produced in nuclear interactions of the more energetic secondary nucleons. Furthermore, muons, although weakly interacting, may occasionally produce nucleons. So there is some crossover in Fig. 4-14 that, happily, can ordinarily be ignored.

THE ATMOSPHERE
processes

AS A BLACK

BOX

Understandably, the detailed analysis of all of the atmospheric is exceedingly complex. Fortunately, combinations of

models (sometimes bearing graphic appellations such isobar) with phenomenological treatments have yielded the understanding required for relating observations deep in the atmosphere to the properties of the incident primary particles that the measurements are supposed to represent. Let us consider a few examples. Suppose we wish to investigate the primary energy spectrum by counting the occurrence rate of extensive air showers at sea
theoretical
as fireball

and baryon

level as a function of

shower

size.

means,

it is

possible to determine the

By appropriate experimental number of electrons in a

shower. Figure 4-18 shows the theoretically calculated relationship

between the number of electrons and the primary energy.

10 8
cr

10
Primary Energy
FIG. 4-18
level.
in

18

eV

"Calibration" curve of an extensive air shower detector located at sea

in

The number of electrons

the shower

is

plotted as a function of the total

it.

energy of the primary that produced

THROUGH THE EARTH'S MAGNETIC

Next,
let

FIELD

AND ATMOSPHERE

87

us consider

how we can

study directly time variations

in the primary intensity by observing the counting rate of a

mary

ground-based neutron or meson detector. If the total flux of priparticles of type i at time t is j\(t), the flux of particles having rigidities between P and P + dP is djt (P, t)/dP; Ni(Pc x, t) is the total counting rate of a given detector at a location where the
cutoff rigidity
,

is P c and the atmospheric depth is x. The corresponding contribution to the counting rate resulting from this increment of the primary spectrum is dN^P, x, t)/dP. The differential elements of counting rate and primary intensity are related by a factor called the multiplicity or specific yield junc-

tion Si(P).

Thus, for a given

set of

parameters

i,

x,

t,

we have

or

N(P)

= j"
p

S{P)

-^dP.

(4-12)

N(P) is the experimentally observed integral response curve, measured by latitude surveys like the example in Fig. 4-9. From these, one can derive differential response functions or coupling cbefficients for neutrons and mesons. Examples are shown in Fig.
4-19.

The

differences in the rigidity-dependence of the sensitivity

of neutron

and meson monitors are

shift of the

clearly evident, as

is

the

downward

energy of

maximum

sensitivity as the at-

mospheric depth decreases. The differential primary spectrum dj{P)/dP is determined in balloon-borne or satellite experiments, hence it is also possible, at least in principle, to evaluate S(P). If we wish to ascribe observed time variations of a secondary component to changes in the primary flux, the atmospheric transducer that couples the primaries to our ground-based detector must remain constant. But it doesn't the weather affects even cosmic rays! There are two principal meteorological factors that must be taken into account. First, the total mass of atmosphere above the instrument varies, as is indicated by changes in the barometric pressure. As the mass overhead increases, fewer secondaries reach the detector. The fractional change in counting rate is proportional to the

88
1000
i

COSMIC RAYS

1 i i i i |

Primary Spectrum -0

gm/cm 2

Neutron Monitor -N

Meson Monitor -/a

m
Z

N-312gm/cm 2

is
/ N-680gm /cm 2
10
"5

_ /
/

2 ^\ / '>-312gm/cm

s
s-~-

2?

V
i
i

/
i
/

^N-Sea

Level

f Q

/
//

5s

l'-

*//
//
;/ //
ii
7
/

-/j.-Sea Level

0.1

5
Rigidity Z3
in

10

20

30

GV
neutron and

FIG. 4-19

The dependence upon

rigidity of the relative responses of

meson detectors at

different atmospheric depths.

difference

Ap between the ambient pressure and the reference (mean) pressure:

AN(P )/N(PC ) =
where
an
<*i

-ctiAp,

(4-13)

is

the barometric pressure coefficient.

torr,

For neutrons,

torr,

and for mesons am ~ 0.3 percent per the exact values depending upon various circumstances.
percent per
is

The

solution of Eq. (4-13)

N(PC ) =

No(Po) exp [<x(p

- />)],

(4-14)
.

where (PC ) is the counting rate at the standard pressure p This is compatible with the data plotted in Fig. 4-15. In addition, the meson intensity is influenced in a far more complicated way by the state of the atmosphere. Since muons are unstable, the distance they must travel after their birth to reach the detector determines their chance of survival. Thus, if

THROUGH THE EARTH'S MAGNETIC

FIELD

AND ATMOSPHERE

89

we think
level

of

them
is

as

being produced mainly at the 100 millibar

(which

roughly the mean production layer) the number

reaching a detector on the ground decreases as the height of the 100 mb isobar increases. Another effect has to do with the competition between Tr-fi decay and nuclear capture of it- mesons. As the temperature of the region between 100 and 200 mb rises, the density falls; consequently more pions decay into muons,
since the pion-eating nuclei are farther apart. Because the
Ja-

meson

intensity increases in this case, this

is

a positive tempera-

The situation is actually quite complex but, although radiosonde data are required for making precise corrections, approximate methods often suffice.
ture effect.

Galactic Cosmic Rays

Seen in the galaxy, that milky way Which nightly as a circling zone thou seest Powder'd with stars.

JOHN MILTON
Let us now resume the fascinating story of how our knowledge about the nature of the primary cosmic rays developed after their identity as electrically charged particles had been established. Hopefully, our discussion of Stormer theory pointed to the possibility, at least in principle, of determining their sign by comparing intensities in the eastern and western directions. This was first investigated by B. Rossi who, in 1931, showed that the preponderance of one sign of charge would give rise to an east-west asymmetry in the intensity. But his measurements in Florence, Italy, did not resolve the effect, which under the
conditions of his experiment
reasons that
J.

would have been quite small, for we now understand. However, T. H. Johnson and
from the west than from on Mt. Washington the following year.

C. Street observed a greater intensity

the east in experiments

This was a surprising result in the light of the general expectation that the primaries were electrons, for which access was easier from the east (Figs. 4-4 and 4-5). It immediately stimulated a
great deal of activity, including several expeditions to the equa-

where the western excess was found, as predicted, be appreciably greater (at sea level and zenith angle of 45 the excess was 15% at the equator and 2 or 3% at 50). From these experiments, it was concluded that the primaries instrumental in producing effects at sea level were largely positively charged. By a remarkable coincidence, C. D. Anderson discovered the positron in 1932. Hence, there were now two contorial regions

to

90

PLATE

flight in

Professor Victor F, Hess in the gondola of his balloon in 1912, after a which he discovered the cosmic radiation hy carrying instruments up to 17,500 feet. (Courtesy National Geographic Society)

PLATE
small

II

Stereoscopic pair of cloud chamber

photographs showing the development of

cascade shower initiated by a single
electron (unaccompanied

by photons) having
thickness 7Z
in

PIATE
cascade
The
forty,

111

Development of
lead plates
in

photon-initiated

an energy of several GeV, entering the upper-

in

a cloud chamber,
particles,

most of three lead plates

13 cm).

[total

maximum number
is

of

about

Much

of the multiplication

the secof phovisible

reached after passing through 10 cm

after which

and plale arises from the absorption

tons,

of

lead,

the

number decreases
is

which

themselves

produce
first

no

sharply. The energy of the incident photon

tracks,

emerging from the

plate.

several GaV,

Cn ION

SLOWING
2

DM

lit

LEXAN
4

SHEET

550fi

a
PLATE IV
(q)

b
in

The signatures of cosmic rays

a crystal of hypcrstherie from the

crab orchard meteorite, The abundant, short tracks are from iron group nuclei, the occasional longer tracks arc from nudei of Z 28. (b] Different etched segments of the track of a single cosmic ray nucleus, Identified as a calcium ion. The particle has

>

crossed three sheers of polycarbonate plastic (#4, 3, and 2)

fourth sheet (#1), [Courtesy Robert

and come

to rest in

Fleischer)

,.*J*

j*-'s* "*-/

V Amundsen-Scott Station at the geographic South Pole, Although most of the buildings are under the snow, the cosmic ray laboratory [center) must remain
PLATE
uncovered, (Courtesy National Science Foundation)

PLATE VI

The tosmic ray observatory eton the 14,000 ft peak Chilean Andes. (Courtesy Gabriel Alvial)

El Infiernilla

in

the

throughout PLATE VII Mobile tosmic ray laboratory that has travelled extensively North America and Hawaii, at Icaros Noval Base, Acapuleo, Mexico.
(Courtesy

Hugh Carmichael)

PLATE VIM

Release of a 135,000 cu

ft

cosmic

PLATE IX
to

Explorer
orbit,

I,

rhe

first

U.S.

sat el (He
1,

ray balloon at
Flying
similar

McMurdo
research

Station, Antarctica,

attain

was launched February

vehicles

with

larger
ft,

1953. The University of Iowa experiment, de-

volumes, ranging up to 20 million cu

quires

re-

signed ro study cosmic rays, discovered the

radiation belt, (Courtesy

somewhat more elaborate launching

James A. Van

Allen}

equipment. (Courtesy Jack Renirie)

Fe

2=26
X
Tracks produced in nuclear emulsion by iron and

Z=90
tame
of the newly'

PLATE

discovered heavier components af the primary cosmic radiation.

(Courtesy Peter H. Fowler)

GALACTIC COSMIC RAYS

91

tenders for the title, and it was an open question whether the incoming cosmic rays were protons or the newly discovered positive electrons. But, as usual, the

cosmic ray workers did not quit

while they were ahead.

aries that

The

results of further experiments, de-

signed to distinguish between the primaries of the hard second-

component
negative.

could penetrate large thicknesses of lead, and the soft that was absorbed in a few centimeters, were inter-

preted as revealing that the latter were equally positive and

The only way to circumvent the dilemma that, it was becoming evident, was inherent in the interpretation of the groundbased measurements was to observe the east-west asymmetry of the primaries themselves with balloon-borne apparatus. In 1939, Johnson and J. G. Barry undertook this experiment in Panama, where the west-east difference was expected to be very large.
Unfortunately, again for reasons that are
results

now

obvious,

their

were not compatible with a predominantly positive primary beam. They concluded that there was a small proton component (<10%), and an electron component balanced with respect to positive

and negative charge. Very soon thereafter, in the early 1940's, M. Schein and his group at Chicago carried out an extensive program of balloon flights with a variety of experimental arrangements. The form of
the intensity vs absorber thickness curve of the particles at high

penchant for producing showers as they were electrons, led to the almost correct conclusion that all the primaries must be protons. In 1947, balloon flights with an experimental arrangement designed to have a response that depended drastically upon the charge of the traversing particle were conducted at Swarthmore in collaboration with F. L. Hereford. These revealed that nuclei with atomic number Z > 2 were present in the primary cosmic radiation. Nuclear emulsions and cloud chambers flown by P. Freier, E. J. Lofgren, E. P. Ney, F. Oppenheimer, H. L. Bradt, and B. Peters the following year showed that elements in the
altitudes,

and

their lack of a

copiously as was expected

if

periodic table at least

up

to the vicinity of iron are

members

of

the primary cosmic-ray family.

It

was generally expected that nuclei with

still

higher charges


92

COSMIC RAYS
day, but that they

would be exceedingly photographic evidence of such particles, with charge as high as Z ~ 90 was reported in 1967 by P. H. Fowler. The photomicrographs in Plate show how the tracks of relativistic higher Z nuclei compare with those of iron.
rare.

would show up some

In
fact,

the

first

Thus was the roster of galactic cosmic radiation essentially completed more than two decades ago with a notable exception. Contrary to earlier belief, electrons now seemed to be conspicuous by their absence! The search for the surprisingly elusive primary electrons was finally rewarded in 1960 when a counter experiment by P. Meyer and R. Vogt, and a cloud chamber study by J. A. Earl detected this small but significant part of the ga-

lactic cosmic-ray

population.
description of primary cosmic rays entails the

The meaningful

quantitative specification of

how

they are distributed with re-

spect to species, intensity, energy, space,

intensity vs energy relationship

energy spectrum of each component must be known. Furthermore, the chemical composition and even finer details concerning the relative abundances
it is important to determine whether the galactic cosmic rays display any directionality and whether their intensity is subject to changes with time. Tremendous progress has been made during the past two decades in assembling the dossier of the primary cosmic radiation,

the

and

time. Thus,

the

of different isotopes are relevant. Finally,

but the task has been formidable for a number of reasons. The particle fluxes, especially of the nuclei with Z > 2, are extremely small. Consequently, statistical uncertainties constitute a major stumbling block. Furthermore, measurements carried out in the atmosphere must be corrected for interactions that break up the
primaries or produce a background of indistinguishable secondaries.

For heavy nuclei, estimates of fragmentation probabilities have often been quite uncertain. The effects of upward moving
protons and electrons
flux of
difficult

splash albedo

add

to the already large

downward moving atmospheric

is

secondaries in

making
if
still

it

to identify singly charged primaries.

And

even

the
re-

detector

arranged to reject the splash albedo, there

GALACTIC COSMIC RAYS

93

mains the problem posed by reentrant albedo outward diffusing secondaries that are trapped by the earth's magnetic field, as predicted by Stormer theory, so that they enter in the downward direction in the other hemisphere. Finally, even at very high latitudes and altitudes, the atmospheric cutoff imposed by the small but finite residual atmosphere masks out the low energy end of the incident spectrum. Clearly, some of these difficulties are now being circumvented by sending instruments far away from the earth aboard spacecraft. But no single technique can provide all of the answers, and our knowledge about the primary cosmic rays continues to broaden and become more precise as the sophistication and resolution of ground-based, balloon-borne and spacecraft instrumentation increases. All of these experimental approaches supplement and

complement each other. In this and succeeding chapters we can only

of the intriguing story that
ers,
is still

distill

the essence

being written by

many workall

using methods that range from classical to ultra modern,

over the world.

ENERGY SPECTRUM
The
relationship that expresses

how

the intensity j(>E) of

cosmic rays with energies exceeding a specified lower limit E varies with energy is called the integral energy spectrum. The
unidirectional intensity refers to the flux
(cf.

page

19) that aris

rives from a given direction per unit solid angle, that

the

number of particles per unit area per unit time per steradian. The energy spectrum can be represented by a power law:
j(>E) = KE-y,
where
(5-1)

and y

are constants.

Taking

the logarithm of both sides

of this equation,

we

see that a log-log plot of

j(>E) vs

should

be a straight line with a slope equal to y. It is striking that, as Fig. 5-1 shows, the value of the exponent does not vary greatly over the entire range of cosmic-ray energies above 10 GeV, extending over more than ten orders of magnitude. We will refer

94
1
1 1 1

COSMIC RAYS
1 1
1 1

L^FIux<=1
^^

Particle /cm 2 / sec

+v> v

S to

1
</>

-2

\
-

A>

= /C 1

F"

6 :

a>

Uj

-4
-6 -8

--,

3-io _
>.

^
-

\
2

"f

\J(>E)*K9 E

-2.2
_

qb=1 Particle

/mvyear

\\

\
-

- -12 c
+a>

^
"o
i-

-14
/>"=>

^\
\

_
\

S-16
c
<t>

=1 Particle /

1000 knrrVyear

-18
I

'

10

12

14

16
:

18

20
in

22

Primary Energy
FIG. 5-1

Log

eV

Integral energy spectrum of primary cosmic rays. The abscissas represent

total energy,

and the ordinates

unidirectional primary particle intensity.

but significant changes in slope at about 10 15 eV and perhaps at 10 18 eV, in connection with the origin of cosmic rays.
later to the small

The

flattening of the integral spectrum at the low energy

end

shape of the differential energy spectrum in that region. Differentiating Eq. (5-1) gives the corresponding differential spectrum, which indicates how the intensity of particles in
reflects the

a specified

energy range between

E and E + dE

(particles/

cm 2 /sec/steradian/MeV)

varies with energy:

dj(E)

dE

= -KyE-

y +1 \

(5-2)

Figure 5-2 represents the observed low energy spectra of the various primary components at solar minimum. As we shall see later,
these differ

somewhat from

the true galactic spectra as a conse-

GALACTIC COSMIC RAYS

10" 3

95
I

II III
1

1 1

1 1

10"

ki
ki

IO- 5

10" 6

10" 7

10-81
10

10*

10

10"*

Energy
FIG. 5-2
Differential

in

MeV

energy spectra of the various species of primary cosmic-ray

particles.

quence of solar modulation. After this local effect has been taken into account, the spectra of the galactic protons and the heavier nuclei appear to be quite similar, at least over the low energy region in which the charge can be determined. There are some
suggestions that changes in composition
ergies.

may

occur at high en-

Let us

now make an
is

order of magnitude estimate of the energy

From Fig. 5-1, the unidirectional intensity of primaries with energies exceeding 1
density that

represented by cosmic rays.

GeV

is ;(>10 9 eV) 0.3 particles/cm 2 /sec/ster. ing flux in free space is

The

correspond-

96

COSMIC RAYS
<j>

4tj
is

4 particles/cm 2 /sec

(5-3)

and the

particle density

Np ~
The
average energy

<t>/c

0- 10

cm" 3

(5-4)
is

is

3
is is

10 9 eV, hence the energy density

all

NPE ~ 0.3

eV/cm

3.

This

sidered, the actual value

an underestimate and, around 1 eV/cm 3

.

things con-

COMPOSITION
The
relative

abundances of the elements in the primary cosmic

radiation are listed in Table 5-1. All of these nuclei are completely stripped of their electrons, since charged particles with such high energies become denuded by an inverse ionization

Table

5-1

Relative Abundances of the Elements in the Primary Cosmic Radiation

in the Universe.

and

The Cosmic-Ray Composition Refers Having the Same Energy Per Nucleon.

to

Nuclei

Cosmic Ray

Cosmical

Abundance
Atomic

Abundance

CR
93

in

CA

in

Ratio

Group
Proton

Element

Number Z
1

percent

percent
91
9.1

CR/CA
1

H
He
Li, Be,

Alpha

H VH
SH

B
F

2 3-5

6.3

C, N, O,

Ne-K
Ga-Zn

WH

Ga-U

6-9 10-19 20-30 31-92

0.10 0.42 0.14 0.04

10~

0.7 10*

>u

10" 6
?

0.14 0.014 2 X 10" 3 lO- 6

?

3 10

20
2
?

~110

process in traversing a very small

varies

amount

of matter.

The nomen-

clature for identifying different groups of the charge spectrum

somewhat. For our purposes, we shall use the scheme: hydrogen (Z = 1); helium (Z = 2); light (Z = 3-5), L; medium (Z = 6-9), M; and heavy (Z > 10), H. Subgroups of the latter are very heavy (Z 20-30), VH; and the newly observed, VVH, heavier than Fe group that ranges up to the super heavy transuranic elements (Z ~ 110), SH. Sometimes, all multiply charged

GALACTIC COSMIC RAYS

nuclei are referred to as heavy nuclei to distinguish

97

them from

protons, since they are characterized by a mass to charge ratio

A/Z ~2, whereas for protons, A/Z = 1. The total energy of a nucleus is divided among
nucleons,

all

of

its

comparing energies one would like to have a scale that is independent of the nature of the species, just as magnetic rigidity is. This is accomplished by expressing the energies as energy per nucleon E/A, since, roughly speaking, individual nucleons in an assemblage with high energy share the total energy equally, and interact independently as though they were separated. The relative abundances of the different nuclei depend upon whether the tally is made in terms of equal velocity, total
for

and

energy, energy per nucleon, or rigidity.

In Fig.

5-1,

the abscissa represents total energy, since this

is

is

the quantity that

determined by

the techniques utilized to

measurements, whereas obtain Fig. 5-2 yield the energy per

EAS

nucleon.

The

general abundances of the elements in the universe, ac-

cording to estimates based on the analysis of a variety of astrophysical observations such as stellar spectra, meteoritic composition, etc., are also tabulated.

The comparison
we

of cosmic ray

and

important implications with respect to the life history of cosmic rays. An interesting characteristic of the nuclear abundances that is not shown in Table 5-1 is the relative numbers of nuclei with even charge number Z and with odd Z. Even Z nuclei are more numerous, and in some regions of the charge spectrum, the even
universal abundances has, as
shall see later,

odd ratio is as high as 10 (see Table 7-2). Concerning the primary electron component, which constitutes a few percent of the incoming particle population, negatively charged particles appear to predominate. At several GeV, the fraction of positrons is only of the order of 10%. For the sake of completeness, we should recognize here the neutral forms of primary cosmic radiation y rays and neutrinos. Understanding of the role they play can be expected to develop as newly available techniques are exploited in the herculean task of determining their characteristics. Indeed, all radiations that reach us from beyond the solar system literally comprise the
to

98

COSMIC RAYS

cosmic radiation. Thus, the newly emerging and highly exciting

fields of x-ray, y-ray,

and neutrino astronomy join

optical

and

radio astronomy in their intimate relationship with cosmic-ray

physics.

The

detailed study of the fine structure in the composition of

the primary cosmic radiation provides independent determinations of such salient features in the career of a typical galactic cosmic ray as the amount of matter it has traversed since birth. Thus, for example, calculations based upon the observed rela-

abundances of the isotopes of helium, He 3 / (He 3 + He 4 ) ~ gm/cm 2 The same result is deduced from the measured fraction L/(L + M) ~ 0.2, by computing how much matter must be traversed for the exceedingly overabundant L group to be produced by the breakup of heavier nuclei. Alternatively, the Be/B ratio (very roughly <0.3) reflects the age of cosmic rays. One of the beryllium isotopes produced by the fragmentation of heavier nuclei is Be 10 which is radioactive, decaying into B 10 with a lifetime of about four million years. Thus, the relative amounts of these two nuclides depends upon the total travel time, which turns out to be in the range of a
tive
0.1,

indicate about 3

million to 100 million years.

The

charge distribution for

of

Z>

80

would provide age information by an extension

U-Pb

dating.

The

lifetimes of transuranic elements or their daughter particles span the time scale from 10 4 to 10 10 years.

SPATIAL DISTRIBUTION
Primary cosmic rays seem to be distributed isotropically throughout our Galaxy, the Milky Way. Except for various wellestablished local effects produced by solar-controlled mechanisms, and perhaps an extremely small sidereal variation (hundredths of a percent), no true spatial anisotropy has been detected. Such an effect might be expected for exceedingly energetic primaries having such large gyroradii that they would not be appreciably deflected by the galactic magnetic fields which stir up most of the cosmic rays and make them isotropic. However, as Fig. 5-1 indicates, the flux at the high energy end of the spectrum is vanishingly small, hence the statistical uncertainties inherent in

GALACTIC COSMIC RAYS

the measurements are very great.

99

The

search for discrete cosmicisotropy has been conless

ray sources in the sky has, thus

of arrival directions 10 14
is

far,

indicated that the distribution

quite uniform.

The

firmed to within an experimental uncertainty

eV,

than 0.05% at
at 10 19 eV.

1%

at

10 17

eV,

10%

at

10 18

eV and 30%

PREHISTORY
rays been ever thus?" whether the intensity of cosmic rays has been constant over a cosmological span of time is to be found in the cryptic messages that they have recorded in meteorites. These fragments of rock and metal contain illuminating information registered by the cosmic-ray bombardment to which they were subjected before plunging close to a fiery death in our atmosphere. The well-preserved nuclear effects wrought by

An

intriguing question
to the

is:

"Have cosmic

The answer

query

as to

cosmic rays are extremely helpful in studies of the meteorites

themselves, just as a host of terrestrial isotopic changes induced
directly

by cosmic-ray interactions are useful in other

widely

disciplines.

The most
isotopes
is

known

application of cosmic-ray produced

the

C 14

radioactive dating technique for determining

the age of organic matter

and archeological

artifacts.

But many

other isotopes are produced by cosmic rays on the earth, and these have been utilized extensively as tracers in studies of the

atmosphere, the oceans, the crust and

ice caps,

marine sediments,
is

and in a variety of other

studies in the earth sciences. Figure 5-3

a schematic representation of the regions of the important isotopic

earth.

and avenues of some changes which can be observed on the

Cosmogenic changes in a meteorite are produced by nuclear

interactions of primary cosmic rays with
called spallation reactions, as described
its

constituents in so-

in

Chapter

2.

If

the

duration of the bombardment is known, one may discover how meteorites were formed and put into the orbits from which they were captured by the earth. Their sizes and shapes before they
suffered ablation and often breakup in the earth's atmosphere can be deduced from measurements of the abundance of certain

bombardment products

as a

function of depth.

100
Extra -Terrestrial
Materials

COSMIC RAYS
Meteorites
(Galactic cosmic rays) Several isotopes

Cosmic Dust
(Solar cosmic roys) 53 26 59

Solar Plasma
/
\

Solar

Mn

Al

Ni

,etc.

Nuclear
?
?
1

surface reactions
?

""

Atmosphere
(Galactic
Several

Polar Region
rays)
14

cosmic
3

isotopes:

He

,Kr

8,

,C

,Si

32

,A

39

,A

37
,etc.

(Galactic plus solar cosmic rays) 53 10 26 59 32 1n ,AI ,Ni ,Si ,Be etc
* f

Oceans
(Neutrons and Muons)

Crust and Ice Caps

Mn^.Be^.AI 26 ^! 36 ,^ 59

^ 32

etc.

Marine Sediments (Muons and Neutrinos)

Crust (Muons and Neutrinos) Kr8 1,etc.

if V-

Neutrons, Muons and Neutrinos

Be^Mn^AI^[Link]^etc

(Radioactive Decay)

-+f

(Radioactive Decay)

Mantle
[

Neutrinos)
Kr
81

He 4

,A

40

etc.

FIG. 5-3

Some

of the most important isotopic changes produced by cosmic rays.

is

The direction of motion of matter

indicated by the arrows.

The difficulties inherent in the measurements, which require quite special techniques, are very great. Only about 10 -7 of the
total

number

of atoms in a meteorite are transmuted,

and the

yields of

some

interesting species are less than one per thou-

sand interactions. Terrestrial contamination and other problems

are formidable.

Analysis of the cosmogenic stable and radioactive nuclides

yields information
orite
is

on the prehistory
produced build up

of cosmic rays. If a mete-

irradiated by a constant flux of cosmic rays, the radioacto a saturation value in

iso-

tive isotopes thus

a time of about one

mean

life.

On

the other hand, the stable

topes continue to accumulate. Thus, assuming that the presentday cosmic-ray intensity and spectrum prevailed throughout the

exposure, the expected rate of production of certain nuclides in an iron meteorite can be calculated. Comparison of the observed

and expected concentrations of nuclides with half-lives ranging from a few weeks to millions of years, as shown in Fig. 5-4, establishes the constancy of the flux

over several million years.

The bombardment

or radiation age

is

defined as the ratio

GALACTIC COSMIC RAYS

10'

101
10 10"
10
26
J

10'

10

10

Ari

Al

Mn5T3
Be
1'

CI

0.5

Years

3
2
Cr
0.5
5

fAr 3
Sc'
6

Ca<
*
*

Na z
3

5Fe

30
FIG.

Mn5^'v.3oo
Doys-

3000

5-4

Comparison of observed and expected concentrations of cosmogenic

nuclides in

an

iron meteorite, indicating constancy of the cosmic-ray flux over

period of several million years.

of the total accumulated concentration of a nuclide to

its

presis

ent production rate. For constant bombarding flux, this

elapsed time between the beginning of the exposure

the
the
fall.

when

meteorite separated from the parent body, and the meteorite

The
other

radiation age can also be expressed in terms of the relative

is

concentrations of two nuclides, one of which

is

stable while the

is

short-lived

compared with the

age,

and

thus in secular

equilibrium. In this case,

T = i* _.
Ps XCr

(5_5)

The

subscripts

and S
is

refer to the radioactive

and

the stable

isotope, respectively; A

the decay constant (1/mean lifetime),

is the production rate, and C is the concentration. The ratio Pn/P& would be equal to unity if the stable species S were produced entirely by the decay of the radioactive species R. Otherwise, its value can be estimated reasonably accurately from spal-

lation systematics.

the average intensity over the last 500 years was the

Measurements based upon these principles have revealed that same as that

over the past half million years within 10%. Constancy over 10 9 years has been established to within a factor of two. The only
available isotope for the longest time determination
life
is

K40

(half-

1.2

10 9

years).

Other problems that, in principle, can be attacked by isotope abundance analysis in meteorites relate to the space gradient of

102
cosmic-ray intensity [Eq.

COSMIC RAYS
(6-7)],

and

the prehistoric composition

and energy
crystals

spectra. Finally, the study of fossil tracks in meteoritic

make

by the technique described in Chapter 3 may ultimately it possible to determine the long-term average fluxes of nuclei with Z > 20.

Cosmic Ray Intensity

Variations
The awful shadow
of

some unseen Power

Floats though unseen

among

us,

visiting

This various world with as inconstant wing As summer winds that creep from flower to flower. PERCY BYSSHE SHELLEY

The search for time variations in their intensity commenced with the discovery of cosmic rays. To determine whether the unknown penetrating radiation came from the sun, Hess made five of his 1911-13 series of balloon flights during the day (including one during a solar eclipse) and
ing into the following morning.
Thereafter, many further attempts were made with groundbased instruments to detect some dependence of the intensity
five at night,

some extend-

upon time
ation.

that might provide a clue as to the source of the radiFor example, in 1923, measurements by Kolhorster and G. v. Salis with an ionization chamber atop the Jungfrau seemed to indicate an increase in intensity as the Milky Way passed overhead. Contradictory results were subsequently obtained by many workers who hoped to confirm this effect, some finding variations over the sidereal day up to 15%, others finding none at all. More than four decades later the reality of an exceedingly small (<.05%) sidereal variation is still being debated! The first report of a variation with local solar time was made by G. Hoffman and F. Lindholm in 1928. The maximum intensity

occurred in the early afternoon, and the minimum at night. The amplitude was less than 0.5%. This happens to be in the right
ballpark!

But the meteorological

factors described in

Chapter 4

intro-

103

104

COSMIC RAYS

duced periodic intensity variations, including the widely studied seasonal changes, that caused confusion and controversy. Real
progress in sorting out periodicities not associated with the atmos-

phere had to await the evolution of the requisite understanding

of the nature of these disturbing effects,

and the development of

for studies of cosmic
S.

analytical procedures for minimizing their contributions.

major step in providing a firm footing

ray intensity variations was the establishment by

ionization chambers (see Fig.
these instruments are
still

Forbush in

1936 of a "permanent" network of reliable continuously recording

1-2).

After more than three decades,

U.S.A.),

in operation at

Cheltenham (now Fredericksburg,

Godhavn (Greenland), Huancayo (Peru) and

and resourcefulness,

Christchurch (New Zealand). Through a combination of exceedingly careful work, keen analytical insight

together with infinite patience, Forbush has discovered most of

variations in the cosmic-ray intensity. The most wave in the amplitude of the diurnal variation with a period of two solar cycles, was reported in 1967.

the

known time

recent, a

The

analysis of the data recorded during the

first

ten years of

operation at the four locations provided a most convincing demonstration of the existence of a significant solar daily variation
latitudes

was not of meteorological origin. The amplitude at midwas about 0.15% with a maximum at approximately 3:00 P.M. local time. At about the same time, K. Malmfors discussed some results that had been obtained with an ingenious method that H. Alfven and he had devised five years earlier. The idea was to point one inclined meson telescope toward the north, and one toward the south, and then to compare the measured diurnal variations. Since, for the same angle of inclination, the particles arriving from both directions have passed through the same amount of atmosphere under approximately the same conditions, the difthat

ference should represent the real diurnal variation.

from the

Suppose, at latitude 45, the telescopes are inclined at 45 vertical. The north-pointing telescope looks along the

earth's axis, hence any periodic daily variation that it sees must be caused by atmospheric effects. The south-pointing telescope scans the equatorial plane, and sees the sum of the atmospheric

COSMIC RAY INTENSITY VARIATIONS

plus nonatmospheric variations. Thus, the
sents the intensity

105

S N

difference repre-

presumably unaffected by meteorological fluctuations. H. Elliot and D. Dolbear later extended the method to include measurements with inclined telescopes pointing in the eastern and western directions. In this case, the two telescopes scan the same part of the sky in succession as the earth rotates. In 1933, W. Messerschmitt observed that the cosmic-ray intensity decreased during a magnetic storm. The worldwide extent of this effect was demonstrated by Forbush in 1937, whence any sudden intensity reduction has since been called a Forbush
decrease*

The

other type of transient intensity change,

is

also

discovered by Forbush,

the solar-flare increase, a

phenomenon

with which we will be concerned in the next chapter. The first evidence for a 27-day recurrence tendency was obtained by A. T. Monk and Compton in 1939. This was related to the 27-day rotational period of the sun with which, as was well

known, the

level of geomagnetic disturbances is closely correlated. Several different manifestations of the return to the visible disk

of active solar regions during successive rotations of the sun have

since

come

to light.

The
sity.

27-day variation had clearly connoted the influence of

individual centers of activity

on the sun upon cosmic-ray

is

inten-

The

corollary of this conclusion

that the intensity varies

with the number and magnitude of the active centers. This effect was ultimately demonstrated by Forbush. In 1957, he showed that the cosmic-ray intensity is inversely correlated with
sunspot numbers. H. V. Neher had also been observing the longterm or solar-cycle variation with balloon-borne ionization chambers in a continuing series of annual flights at Thule, near the geo-

magnetic pole. His measurements bespoke the presence at solar minimum of a large flux of slow protons that seemed to diminish
as solar activity increased.

Taken

together, these observations led

to the conclusion that low-energy particles are subjected to

much

greater variations than are higher-energy particles.

Except for the problematical sidereal

effect, all

of these periodic

and

transient time variations in intensity represent spatial aniso-

* Forbush has now become resigned to this, although he winces people replace the word decrease with decline!

when

106
tropies in the cosmic-ray intensity arising

COSMIC RAYS
from solar phenomena. understand some of the subtleties of the

We are just beginning

to

mechanisms that come into play. Let us now summarize the present state of our knowledge about the spatial anisotropics and the modulation mechanisms that produce them, but defer until the next chapter further discussion of time variations attributable to the local production
of cosmic rays by the sun. Before attempting to understand the
role played by the sun in all of these matters,

we should

digress

for a brief review of the relevant solar physics. Fortunately, for

we need not be armed with a complete and of the fantastic and, for the most part, little-understood modes of energy transformation that occur on our nearest star. Rather, it will suffice to focus our attention on the means whereby the sun exercises control over the electromagnetic conditions in interplanetary space and the immediate
the present purposes,

picture of the sun

environs of our planet.

THE SUN AND THE INTERPLANETARY MEDIUM A brand-new scientific discipline, solar-terrestrial physics, embraces the study of the sun
tary

and

its

emissions, of the interplane-

medium, and

of the earth's environment. Research in this

field falls into

two general categories. Essentially steady-state con-

ditions characterize the so-called quiet sun, a term which, of

course, implies only relative calm. Superposed

upon

this

base

are disturbances, or storms, that collectively comprise solar activity,

the distinguishing feature of the active sun.

concentric regions can be delineated. The from which radiant energy does not escape directly, consists of the core, the sun's thermonuclear power house, surrounded by the convective envelope that extends about one hundred thousand kilometers to just below the visible exterior. The photosphere is the apparent surface. This thin layer, only a few hundred kilometers in depth, is the origin of the visinterior,

The Quiet Sun. Four

roughly speaking, corresponds to that from a black body at a temperature of almost 6000 K. Under high resolution, structural details called granules can be seen.
ible radiation which,

COSMIC RAY INTENSITY VARIATIONS

107

The chromosphere
000-15,000

is

km

thick,

a dynamic transition region, about 10,between the "cool" photosphere and the
rays.

"hot" corona.
it is

It is

the seat of shortwave radio frequency radia-

tion, ultraviolet light,

and x

Complex and inhomogeneous,

produced by the molife-

characterized by rapid temperature changes and a high level

of excitation.

sort of prairie-fire effect is

tions of close-packed spicules

jet-like

prominences, with

times of several minutes and velocities of about 20 km/sec, that

can reach heights of 10,000 km. The corona is the tenuous outer highly ionized gaseous atmosphere of the sun. Although the corona is usually represented in eclipse photographs as extending outward to a distance of
several solar radii, in reality the earth is immersed in the corona, which continues out to even much greater distances from the sun. Although the intensity of coronal light is only about a millionth

of that of sunlight, processes corresponding to temperatures of

millions of degrees occur.

The corona
is

is

also a source of

x rays

and radio-frequency

radiation.

The Active Sun.

turbations,

Solar activity

the manifestation of the ocof discrete, local per-

currence on the solar surface of a

or active
regions,

number

with individual lifetimes that, typically, may range from one day to three months. Some live almost a year. A solar-activity center develops in an area about one- tenth of the solar disk. Its development is accompanied by
the appearance of a number of features that distinguish the active sun from the quiet sun. These active regions characterize solar cycles, the duration of which is 10 to 14 years. Their frequency of occurrence, average size, and lifetime first increase and then decrease until they eventually disappear entirely. New-cycle active regions may appear as long as 30 months before the old cycle has ended. No adequate physical picture of the solar-cycle variation has yet been developed. In point of fact, the solar cycle connotes much more than a variation of sunspot numbers. However, systematic data on this index of the level of solar activity extends over more than two and a half centuries, and some recorded observations date back to the time of the discovery of sunspots by Galileo in 1610. With only one exception, all of the features related to the active

108

COSMIC RAYS

sun, including the shape of the corona, follow the solar cycle,

but to a widely varying extent.

No

clear variation in the inteyet.

grated visible light output has been measured as

Sunspots are dark regions that are 1000-1500 cooler than the surrounding photosphere. Their sizes range from hundreds of
kilometers (intergranular dark spots) to 10 5

km

(areas of a billion

square miles!), and their lifetimes, roughly correlated with size, are as long as many months. Strong magnetic fields up to about

3000 gauss and perpendicular to the surface

the spot becomes visible,

may develop

before
cycle

and may

persist afterwards.

Sunspots

rarely occur at latitudes above 35 or 40. Spots in a

new

appear at the higher

solar

latitudes,

and

drift

toward the equator as

maximum

approaches.

They

often occur in pairs of opposite

polarity (bipolar spots).

The

polarity orientations (sign of leading

and following members) in the two hemispheres are consistently opposite, and reverse in alternate 11-year cycles. Faculae, as seen in white light, or monochromatic plages, are
areas that are brighter than their surroundings because, in general,

they are several

hundred degrees

hotter.

Somewhat

larger

than granules, they can appear at higher latitudes, and may cover 10% of the solar disk. They surround all spots and spot groups, but their lifetimes are longer. These features are always
associated with magnetic fields.

Prominences look like projections of the photosphere into the chromosphere, and their form and pattern are varied. Typically, the longest dimension is 10 5 km. They are cooler, denser, and more opaque than the surrounding coronal material, and may move at velocities of hundreds of kilometers per second. Their
lifetime
is

several weeks.

They may disappear

catastrophically,

and then reappear in the original form, repeating this behavior several times. Prominences seem to be shaped by the local magnetic
field.

High-resolution

photographs reveal a fibrous or

al-

webby

fine structure.

They

are often associated with sunspots

though they may occur there are no sunspots.

at

high heliographic latitudes where

Filaments are thin or blade-shaped prominences that are observed in spectrohelio grams , photographs with light in a very

narrow wavelength band. They are projections, viewed from

COSMIC RAY INTENSITY VARIATIONS

109

above, of stable prominences lying across lines of force at the top

of an arch connecting two magnetic regions of opposite polarity,

supported like a hammock. Flocculi are wormy-looking bright or dark short filaments. Flares, or sudden chromospheric eruptions, are catastrophic events that can have almost immediate repercussions at the earth. We shall have occasion to describe in detail the characteristics of these most violent of all solar phenomena in discussing solar cosmic rays.

The

Solar Wind. Although correlations between certain geo-

physical effects

and sunspots had been known

first

for

many
and

years,

the nature of the sun-earth links was a long-standing

chal-

lenging mystery. Perhaps the

inkling of the solution struck

Stormer when, as was noted earlier, he considered the possibility that the aurora is produced by low-energy electrons emitted by
the sun.

The foundations of our current understanding of the subject were the theoretical studies started by S. Chapman in 1927. These came to fruition in calculations published by Chapman and
V. C. A. Ferraro in 1931. They attributed a number of observable geomagnetic effects to the impact on the earth of a jet of
ionized but electrically neutral gas that swept past the earth roughly a day after being emitted by the sun. This type of highly conducting particle stream is, in fact, now called a plasma, and the study of its properties falls within the domains of magneto hydrodynamics and plasma physics *

The

that the large acceleration of

next milestone was the recognition in 1951 by L. Biermann Type I comet tails, which could not

be accounted for by the radiation pressure of sunlight, could be caused by a rather intense flux of low-energy protons blown out-

ward from the sun. The fact that these tails persisted even when there were no magnetic storms, and when comets moved outside of the ecliptic plane, seemed to suggest a continuous ejection
in all directions from the sun.

In 1954,

Chapman proposed
11,

a static model of the solar corona,

See momentum Book No.

Boley (1966) and
Little (1967),

Plasmas

Laboratory

momentum Book No.

Princeton.

18,

and Cosmic, by F. I. Magnetohydrodynamics, by N. C.

Van Nostrand,

110
envisaging the earth as being located inside
it,

COSMIC RAYS
in

which the

charged-particle density was about 300 electron-proton pairs per

cubic centimeter at the earth's orbit. However, certain difficulties

stemmed from

the differences between the observed temperature

distribution at the sun

and that required

to transfer energy

from

the coronal base by thermal conduction.

The great breakthrough came in 1958, when E. Parker proposed a theory in which hydrodynamic streaming of plasma continuously ejected by the sun
is

the energy transport mechanism.

The outwardly expanding

locity.

coronal gas moves at hypersonic ve-

In analogy with the propagation of supersonic acoustic waves in which the wave velocity exceeds the velocity of sound in the medium, (yp/p)^, in this case the particle velocities exceed the analogous Alfven velocity, (B 2 /A-np)^. Parker christened the radially streaming plasma the solar wind. His predictions concerning its properties have been borne out in detail by both in situ measurements by spacecraft and ground-based observations. Furthermore, investigations by Parker and a number of others, of the cosmic-ray intensity modulations produced by the solar wind, have provided a general understanding of the mechanisms that
cause the different intensity variations.

The

search for

new

phe-

nomena
is still

to

which the

solar

wind might be expected

to give rise

in progress.

The
found

average characteristics of the solar wind in the vicinity of

is

the earth are listed in Table 6-1. It

effects

can be wrought by

this

remarkable that such prominute concentration of

Table

6-1. Average Characteristics of the Solar Wind Near the Earth's Orbit. Values in Parentheses Represent Typical Ranges.

Composition Flux
Velocity Density

Thermal Energy Temperature

and a few percent a particles 10 8 particles/cm 2 /sec 500 (300-600) km/sec 10 (1-20) particles/cm 3 (of each sign) 10 eV 10 6 K
Protons, electrons,
5

Mean

Free Path Proton Energy

Electron Energy Magnetic Field Conductivity

1 1

AU
keV

10

eV

X 10 -6 gauss 100 (ohm-cm) -1

5 (3-15)

COSMIC RAY INTENSITY VARIATIONS

111

matter in an otherwise perfect vacuum (except for 10~ 5 cometary dust particles per cm 3 )! The special abilities of the solar wind to push charged particles around reside in the magnetic field that
it transports. For magnetic fields rooted at the sun are literally pulled out into space like taffy and carried along by the outgoing plasma.

To understand how this happens, we must recall Lenz's law, according to which a good conductor tends to resist any change
in its internal magnetic field. Any alteration of the field produces an electromotive force, in accordance with Faraday's law, and this emf produces a current, the associated field of which opposes the magnetic field change. Ordinarily, these currents would

damp out rapidly

because of the

finite resistance of the

is

conductor.

But the conductivity of the interplanetary medium

so high that

the current does not die out, with the result that the solar plasma transports the magnetic field away from the base of the corona. This frozen-in magnetic field remains with the plasma, even

though the

may become kinked. wind continues to blow until it reaches a distance from the sun r at which the energy density of the solar magnetic
lines of force

The

solar

the galactic magnetic field

B 2 /8ir ergs per cubic centimeter has decreased to that of B g2 /8ir ergs per cubic centimeter. An interesting twist is introduced by the rotation of the sun. Since the plasma is accelerated radially outward from the sun,
field

the interplanetary field lines

fixed.

But the sun

rotates with

would also be radial if the sun were uniform angular velocity Q, as il-

lustrated in Fig. 6-1. Particles emitted with velocity v from points 1-4 at successive equal intervals of time At reach points V-< at

time 4A*.
is

The

distance that each particle has traveled in each At

the radial line that traces
its

marked on

path.

The

frozen-in

field,

firmly anchored to the sun,

follows

the heavy line, an

Archimedean spiral, exactly like the stream of water emerging from a rotating lawn sprinkler. The stream angle or garden hose
angle
is

given by:

tanf
Since
12

Or
10

(6-1)

rev/27 days

~ 2.7 X

radians/sec,

and

at

the

112
Frozen-in Magnetic Field

COSMIC RAYS

Radial Direction of Plasma Motion

FIG. 6-1

Archimedes spiral configuration of magnetic field lines frozen into plasma moving radially outward from the rotating sun. At the earth's orbit, the gardenhose angle is about 45.

orbit of earth r

1.5

10 13 cm,

we

find that the garden hose

angle corresponding to v

is about 45. Another important feature of the interplanetary magnetic field is the observed sector structure. As is indicated schematically in Fig. 6-2, the field points predominantly toward the sun for a period of 6 or 7 days, and then in the reverse direction for a similar interval. Of course, when the sun is active it is not possible to trace sectors for more than one solar rotation.

400 km/sec

Day 20

Day 5

Predominant
Direction of

Interplanetary

Sector Boundary

Magnetic

Field

FIG. 6-2

Sector-structure of the interplanetary magnetic

field.

COSMIC RAY INTENSITY VARIATIONS

Needless to say, this
is

113

an oversimplified picture, although it describes the gross features of the interplanetary magnetic field rather well. Effects of changes in the plasma velocity and the magnetic field polarity, or of hydrodynamic shock waves, can introduce small-scale irregularities into the otherwise ordered field, or may produce a large-scale disordered field. All the different characteristics of the interplanetary magnetic regime in one way or another affect most of the cosmic rays that reach the inner solar
system.

The effectiveness of each of the modulation processes is determined by the detailed structure of the interplanetary medium. Thus, for example, the distribution of the scale sizes of magnetic
irregularities

and the mean distance between them are

signifi-

cant factors. Clearly, the

mode

of scattering

when

a particle en-

counters a clump of twisted and tangled magnetic field lines depends upon whether its gyroradius is small or large compared with the linear dimensions of the scattering center. To see this, we need only note that, roughly speaking, the angular deflection $ of a cosmic-ray particle having a radius of gyration rg in passing through a field of length x is 8 = x/rg
.

MODULATIONS AND ANISOTROPIES

Spatial anisotropics are studied with ground-based instruments by relating observed time variations in the intensity to directions in space via the procedures described in Chapter 4. With the advent of the space age, it has also become possible to observe asymmetries directly by measuring the unidirectional intensities of particles with known characteristics as a function of the spatial orientation of the detector. In general, these two methods are complementary, since they cover different energy ranges. In utilizing cosmic rays as space probes for investigating both

the gross and detailed properties of the interplanetary medium, one attempts to deduce various physical parameters by comparing the data with the predictions of theoretical models. For example, it is possible to determine the mean free path between the scatterings that occur when a particle collides with small-scale magnetic irregularities. Needless to say, untangling all of the rele-

114

COSMIC RAYS

vant factors poses a complex problem, but, fortunately, relatively simple models can often be invoked to yield an understanding of the underlying physics.

An important characteristic of any modulation process is its dependence upon magnetic rigidity, since the interactions of particles with interplanetary magnetic fields are the root cause of the anisotropics. The relationship between the fractional change in intensity Aj(P)/j(P) and the rigidity P is called the variational spectrum. One form of the spectrum of variation is
A;(P)
j(P)

KP-T.

(6-2)

Thus, y

connotes equal attenuation of the intensity of parti-

cles of all rigidities.

Diurnal Variation. Figure 6-3 shows the average daily variation, by a particular neutron monitor. The percentage deviation of the counting rate from the mean is plotted as a funcas observed

0400 0800 1200 1600 2000 2400

Hour, Local

Time
in

FIG. 6-3

Twenty-four hour variation

the cosmic-ray intensity.

COSMIC RAY INTENSITY VARIATIONS

tion of local time.
indicated.

115

The amplitude A and

time of

maximum t m

are

periodic variation can most conveniently be represented by

a vector of length proportional to the amplitude, and pointing

toward the direction (or time) of

maximum

intensity.

This vector
its

completely describes the spatial anisotropy.

The

observed diurnal variation can be

split

up

into

har-

monic components by a Fourier analysis. Let us consider just the first harmonic or 24-hour wave plotted in Fig. 6-3. AN(t), the deviation of the counting rate at time t from the daily mean, is
given by

AN(t)

a cos

4- b sin

t,

(6-3)

where t an hour

is is

expressed in degrees. Since 360 represents 24 hours, equivalent to 15. This equation can be put into more

convenient form by a simple transformation. Let

= Tsine, = A cos e. = Asm

(t

(64)

Then
AN(t)

e).

(6-5)

According
at

to this equation,
,

AN(t) attains

its

maximum

value

m where (t m + e)= 7r/2. Now let us construct a vector that represents both amplitude A and time of maximum t m We can plot this vector in polar coordinates on the harmonic dial shown in Fig. 6-4. The virtue of this clock diagram is that the vector A points directly toward the time of maximum t m since The Fourier coefficients given by Eq. (6-4) core = (tt/2) t m respond to the components a and b of the vector A in the two
time
t
.

directions at right angles.

A similar analysis can be carried out to determine higher harmonics that may also be present. Attention has generally been limited to the first two harmonics. In the case of the second harmonic or semi-diurnal hour clock.
variation, the vectors are plotted

on a

12-

The

diurnal variation vectors

is

may

fluctuate appreciably

day to day. This

illustrated in Fig. 6-5,

from where the dots repre-

116

COSMIC RAYS

FIG. 6-4

Harmonic
22

dial representation of the diurnal variation.

20

Swarthmore 1965

Diurnal Variation Harmonic -Dial

Radial Scale
1 i

1
1

0.1

0.2

0.3%

18

-6

16

14
FIG. 6-5

12

10

Daily variation vectors (from origin to dot) on individual days during

entire year.

an

Arrow represents the annual mean.

sent the heads of the individual daily vectors determined

from

observations extending over an entire year.

The arrow

is

the an-

nual mean.

The

striking feature

is

that,

despite

the great scatter,

the

'

COSMIC RAY INTENSITY VARIATIONS

average amplitude and direction of the diurnal vector
persistent.
is

117
quite

We

can see
6-7,

this

by adding the monthly mean vectors

that the direction of the diurnal

for the

same

year, as in Fig. 6-6.

Next, in Fig.
anisotropy
realistically
is

we note

rather constant from year to year. However, after

is

evaluating the uncertainties, which

is

no easy chore,

we

find that the amplitude

effect is

confirmed by Fig.
16

6-8,

not constant. The reality of this which shows the solar-cycle vari17
\

18
i

/Dec.1958
Diurnal Variation

Summation -Dial
Churchill
15

^
Dec.1965

IGY
14
s

IQSY

V.
13
i

'

19

Radial Scale
i i

0.4

0.8%

20

12

^Jan.65
Jan.

58*^.
.

/ ll\\\ x x

11

10

FIG. 6-6

Summation

dial.

Each vector, added to the preceding one, represents the monthly mean diurnal variation.

ation in the amplitude of the daily variation as revealed by analysis

of the "long-playing records" of the Carnegie ionization cham-

bers.

Thus

far,

we have been

referring to the time of

maximum,

or

the phase of the diurnal variation, in terms of the local solar time of the observing station. To determine the direction in space of the apparent source of the

maximum

intensity,

we

recall that

it

118
is

COSMIC RAYS

necessary to correct for geomagnetic bending of the primary

particles to

which the detector responds. This

4.

is

the procedures already outlined in Chapter

The

accomplished by arrows in Fig.

Churchill Neutron Monitor

Radial Scale
0.1

18-

15

12
cycle.

FIG. 6-7

Annual mean daily variation over a complete solar

1 1

~1

1938

40 42

44

46

48

50

52

54

56

58

60

62

64

Year

FIG. 6-8

Amplitude of the solar daily variation of the meson intensity as determined from ionization chamber records extending over nearly three decades.

COSMIC RAY INTENSITY VARIATIONS

2400 hr

119

1800 hr-

0600hr

Correction tor

South Pole

Geomagnetic Bending

1200
FIG. 6-9

hr

Diurnal variation at several stations as observed in local solar time,

the calculated corrections for geomagnetic bending.

and

6-9 are the

mean

vectors at several different stations before cor-

rection for bending.

The arcs indicate how far each vector must be rotated, according to calculations of the asymptotic cones, to make it point toward the real time of maximum. It is indeed comforting that the vectors derived from the observations at different locations all line up,

and the amplitudes

differ as they

should.

Figure 6-10 suggests why the maximum intensity is seen at 1800 hours local solar time. The spiral interplanetary magnetic field co-rotates with the sun, more or less as though it were a rigid structure. Thus, the isotropic cosmic-ray "gas" is stirred around by the great paddle wheels, imparting to it a velocity of a few

hundred kilometers per second. Since

velocity of the earth along
its

this

is

faster

than the

orbit, the intensity

is

coming from
is

the direction 90 East of the sun-earth line

enhanced, whereas
reduced.

that arriving from the opposite direction, 90 West,

Calculations based

upon

this

model predict roughly the observed

amplitude. Furthermore, the expected amplitude varies as the

cosine of the latitude, in accordance with the observations.

On
is

the average, the variational spectrum of the daily variation

of the form of Eq.

,

Pu

(6-2), with y ~ 0, up to a limiting rigidity, above which the diurnal anisotropy disappears. Thus,

120

COSMIC RAYS

FIG.

6-10

Schematic

model of the mechanism that produces the solar daily

variation.

4/V) _ K
and
Aj(P)

P<

P>
to 100

Pu

(6-6)

Pu
as 55

changes during the solar

cycle,

and may range from

as

low

GV

at solar

minimum up

GV

at solar

maximum. Of

course, as the scatter of the vectors in Fig. 6-5 suggests, short term

fluctuations in the spectrum

may

be quite large.

The

study of the 24-hour periodicity, with a

mean amplitude

of several tenths of a percent, poses an exceedingly challenging

problem. Even with the newest techniques, and despite their capabilities for providing greatly reduced statistical uncertainties, many difficulties remain, and definitive results are not easy to come by. Since the upper limits that have been determined for the magnitudes of semi-diurnal and sidereal variations are appreciably smaller, it goes without saying that discussion of their
implications
is

somewhat more

speculative.

COSMIC RAY INTENSITY VARIATIONS

121

The amplitude
and the

of the semi-diurnal variation

is less

than 0.1%
less

intensity seems to be

maximum

in both directions along

is

the line 135 west of the earth-sun line. This

more or

perpendicular to the direction of the interplanetary magnetic

field.

The

The spectral index (Eq. 6-2) is y 1 up to Pu 100 GV. presence of the second harmonic appears to imply the existis

ence of a cosmic ray density gradient, that

equatorial plane.

perpendicular to the

The apparent
a percent. It
is,

sidereal variation

is

only a few hundredths of

indeed, the proverbial needle in the haystack.

its

Questions about

reality are still

being raised on several counts.

To

appreciate the nature of the difficulty, one need but recall

from the solar day by only four minwhich amounts to just one day per year. Consequently, the shadow of spurious variations generated by the solar wave looms ominously as a major source of uncertainty. Forbush Decrease. In contrast with the diurnal variation which is a local-time effect, the other types of modulation are universal-time effects. This means that their repercussions are felt all over the world almost simultaneously, even though, in some cases spatial anisotropics can produce temporary asymmetries. Thus, the Forbush decrease is essentially a universal-time phenomenon, although small differences in onset time, or other features can
that the sidereal day differs
utes,

occur.

Forbush decrease

is

a very

sudden reduction in the cosmicslower recovery, as

is

ray intensity, followed by a

much

illus-

trated in Fig. 6-11. Typically, the counting rate of a neutron

monitor drops as much as 5% within a few hours, although this can stretch out to as long as two days. The recovery to predecrease level may extend over an interval of several days to several weeks. The superposition of several Forbush decreases,
as in Fig. 6-12, is called a cosmic-ray storm.

Transient intensity decreases often occur in association with magnetic storms, but there is not a one-to-one correspondence. Either phenomenon may happen without the other. Both of these effects are often observed one or two days after a solar flare. This delay corresponds to the transit time of plasma ejected toward the earth in the explosion.

122

COSMIC RAYS

17

18

19

20

February 1967
FIG. 6-1
1

Typical Forbush decrease.

Because of the apparent correlation between geomagnetic

turbances and cosmic-ray intensity, Forbush decreases were
ascribed to effects localized at the earth
tions in the
itself,

disfirst

such as time varia-

However, the fact that they were observed at polar stations, where the cutoff is atmospheric rather than geomagnetic, coupled with other arguments pointed toward a more far-reaching phenomenon. Observations with deep-space probes have since revealed that, in fact, the scale size of these events is greater than 0.1 AU, and may well exceed 1 AU.* A Forbush event may display a great deal of fine structure. There is sometimes a precursor which manifests itself either as
magnetic
field.

a pre-increase or a pre-decrease. Spatial anisotropics can also

which is especially interestwas the first example of a north-south asymmetry, an effect that has been very elusive. Furthermore, the magnitude of the intensity drop usually depends upon latitude since the typical variational spectrum (Eq. 6-2 with y ~ 1) is an inverse dependence upon rigidity. Finally, trains of enhanced diurnal variation, with amplitude even exceeding 1%, are sometimes oboccur. This
is

illustrated in Fig. 6-13,

ing because

it

served, as

is

shown

in Fig. 6-14.

As are

all

cosmic-ray intensity modulations, the Forbush de-

* is the abbreviation for astronomical unit, the earth from the sun = 1.496 X 10 R km.

AU

mean

distance of the

is a solar-wind phenomenon. The magnetic plasma cloud ejected by a solar flare at a velocity greater than that of the

crease

plasma which preceded it causes a blast wave as in Fig. 6-15. This sweeps particles away as it expands, leaving the volume in back of it partially depleted of cosmic rays. This effect can also be visualized as a magnetic tongue or bottle that, as in Fig. 6-16, excludes some galactic cosmic rays from its volume. Thus the bubbles of solar plasma can act as cosmic-ray shields.

124
1 i

COSMIC RAYS
l

-A
-j-i's

McMurdo

South Pole

x Alert
Thule
-

4-

\i

*+.

3-

/
S

i
-

2-

t**\

4
V
Event

\
i.

-4--ft

-1 -

Base
i

23

24

25
March 1966

26

27

FIG. 6-13

Forbush decrease exhibiting a north-south asymmetry, indicative of a

spatial anisotropy perpendicular to the plane of the earth's orbit.

21-Day Recurrence Tendency. Forbush decreases tend to recur

at intervals of 27 days, the synodic rotation period of the sun,

just as

In

fact,

do geomagnetic storms which stem from the same cause. diminutive versions of the Forbush decrease may recur

many

times.

0117
I-

Standard

Swarthmore Meson Detector

Deviation

^116

o ^115
>

c114
16
17

18

19

20

21

22

23

24

25

26

27

28

29

30 F e T

July 1961 FIG. 6-14

Train of enhanced diurnal variation.

COSMIC RAY INTENSITY VARIATIONS

125

Radial Directions

of Plasma Motion V.

Magnetic Field
Distorted by
Blast
Spiral

Wave

Interplanetary

Magnetic Field

Sun
FIG. 6-15

Schematic representation of blast

wave mechanism

that can produce a

temporary reduction

in

the cosmic-ray intensity.

One
method
ture,

technique for investigating recurrence tendencies

of superposed epochs or

is

the

Chree

analysis.

Suppose we

choose as zero day,

average) intensity.

= 0, those days that show a particular feasuch as appreciably higher than average (or lower than

We
1,

then write
2, 3
. . .

down

the intensity for the

following days,

N=
.
.

n and the preceding days

N=

1,
(2ra

2,

3
days

n. The

average value for each of these

1)

is

procedures, such as autocorrelation or

then plotted against N, as in Fig. 6-17. Statistical power spectrum analysis,

are also useful for investigating recurrent

Long Term
it

Variation.

The

solar

phenomena. wind blows unceasingly. But

the electromagnetic conditions in the solar magnetic cavity that

carves in space reflect the changes in the level of solar activity,
to discrete outbursts

and respond
that,

on the
storms,

sun. Thus, the interit is

planetary weather can be quite stormy.

in

And

not surprising
are

addition

to

individual

there

"seasonal"

changes in the average characteristics during the solar-activity

cycle. All of the transient

and periodic

intensity variations that

have already been described evidence these changes. As Fig. 6-18 testifies, even the variations vary over the 11-year period! But, in addition, there is one other exceedingly important process, the long-term modulation, that transcends all the other temporal changes. Fig. 6-19 shows the change in intensity during the most recent

126

COSMIC RAYS

Active Region

Cosmic Rays
Earth

FIG. 6-16

Magnetic bottle model of the modulation of cosmic-ray intensity during

a Forbush decrease.

high latitude neutron monitor. Of magnitude of the change depends upon the nature of the detector and the range of solar activity. Typically the total reduction from solar minimum to solar maximum is about 20% for nucleons, and 5% for mesons. These figures suffice to expose what is happening. As solar activity increases, more and more
solar cycle, as observed with a

course, the

galactic cosmic rays are being denied admission to the inner solar

system. Furthermore, the lower the rigidity of a particle, the less

chance it has to reach the earth's environs. This implies that the observed primary spectrum differs from the unabridged galactic spectrum that prevails beyond the boundary of the solar magnetic cavity. Until recently, it was generally believed that the spectrum measured at solar minimum would closely resemble the bona fide galactic spectrum. But, Fig. 6-20 reveals how drastically our views on this subject have changed.
Just as the solar

wind theory was

successful in predicting the

COSMIC RAY INTENSITY VARIATIONS

/V=0
Preceding Days,/V=-^
1
1 1 1 1

127

Following Days,
1

N-+n
i

1 1

1 1

ii

ii

2.0
-t-

Intensity Higher 'than

Aver
i

c
0}

(Intensity Lower than Average

e
a.

1.5

>*0.5

0.0

0.5
1 1 1 1 1 1 1 1

-140

-112

-84 -56 -28 28 56 84 Days trom Selected High and Low Days
for investigating

112

140

FIG. 6-17

Method of superposed epochs

recurrence tendencies.

The 28-day separation of the peaks corresponds to the sun's rotational period.
1957
i
1 1
1 1

~\

1958
i i
i

FIG.

6-18

Neutron monitor data recorded during a period of

maximum

solar

activity

(IGY

International

Geophysical

Year)

(IQSY = International Years of the Quiet Sun). each case, the ordinates represent deviations of the daily mean intensity from the

and minimum solar activity The two scales are the same. In

mean over

the entire corresponding period.

128
1957 1958
1 i

COSMIC RAYS
1959
i

1960
i

1961
i

1962
i

1963
i

1964
i

1965

1966
'

1967
i

760

5 | <

740
Thule

720

*
-

^
Intensi

700

o
S 660 o

M in.
-

Sun Spot Numbers

(Runninc

Max. Cosmic

Ray
Intensity

c 640 o
>>
t-

Means

620
IGY
i
i i

IQSY
i
i

1957 1958 FIG. 6-19

1959

1960

1961

1962

1963

1964

1965

1966

1967

Intensity of the nucleonic

component
cycle.

at a high latitude station during

one solar

properties of the interplanetary

medium and
it

in accounting for

the other cosmic ray modulations,

solar-cycle variation.

has also explained the

is

The

basic

mechanism

the convective re-

moval of

galactic cosmic rays

field irregularities

by encounters with the magnetic carried outward from the sun by the "steady

state" solar wind. The cosmic ray density in the inner solar system represents a balance between outward convection and inward diffusion. This implies that, even at solar minimum, the

cosmic-ray spectrum observed at earth

is

not the galactic spec-

trum.

The

latter

must be unfolded from the experimental data

with the help of a theoretical model.

The form
lation
is

of the rigidity-dependence of the long-term

since
it is

modu-

of

paramount importance,

crucial in extrapo-

lating the observations out into the galaxy. Because of the action of the outward-blowing solar wind, the flux of cosmic rays
is

progressively reduced as the sun

is

approached. In other words,

there

is

a density gradient,

and
is

diffusion theory tells us that the

fractional change in the cosmic-ray density

(Eq. 5-4) per unit

radial distance

from the sun

COSMIC RAY INTENSITY VARIATIONS

129

Proposed Alternative

\
4 *

Versions of

\
.

\
\

Unmodulated Spectrum

10"

10- 2

10"

Nx
10"

-Modulated

Vv

Spectrum Observed at Solar Minimum

10"
;

10

10'
Energy

10

in

MeV

FIG. 6-20

Galactic cosmic-ray spectrum according to several different evaluations

of the effects of solar modulation. The curve nearest to the spectrum observed

during the most recent solar minimum epoch

now seems
3e w

to be the

most reasonable.

dN
Ndr
where vw
is
is

(6-7)

Vp\i

the solar
free

the

mean

wind speed, v p the particle speed, and A.j path for collisions between a cosmic-ray particle
r

and a

scattering center. Integration gives the ratio of the ob-

served density

of cosmic rays with rigidity

at a distance r

from the sun

written

to the density

r6

beyond
its

the boundary of the

region of solar modulation. In

simplest form, this can be

N /Nro =
r

exp

(-K

/D),

(6-8)

where

is

particles in the solar

the diffusion coefficient describing the motion magnetic fields that permeates interplaneof the

130
tary space.

COSMIC RAYS

Kv

is

a quantity related to the bulk outward speed of

and to the extent of the region of modulation about the sun. D depends upon the rigidity and species of the particle, and can be evaluated from experimental data. Kv is independent of these parameters, and its absolute value (and hence the total modulation) has not yet been determined experimentally. Fig. 6-20 shows that there have been some differences of opinion as to the correct evaluation of D, and the best estimate of
the solar plasma,

v.

Details such as the energy losses (and gains) suffered by the

incoming

consequence of deceleration (and acceleraneed to be taken into account. Nevertheless, the basic principles underlying this model seem to be borne out by most of the available experimental data, which, in turn, show how the theoretical model needs to be refined in the next apparticles as a
tion) processes, also

proximation.

Returning
sec, Aj

to

Eq.

(6-7), if

^ 0.05

AU, and v v

c,

we insert the values v w ~ 500 km/ we find that, near the orbit of the

earth, the space gradient for particles with rigidities in the

GV

range should be about

by deep space probes,

AU. Although measurements extending over about 0.4 AU, seemed to

10%

per

be consistent with expectation, there

is

considerable dispersion

among

the conclusions of different experimenters.

,

Earlier estimates of r

the distance at

which the

momentum
which
it
is

density of the solar

wind decreases

to the point at

stopped by the interstellar magnetic field, were based upon the time interval between sunspot minimum and cosmic-ray maximum (Fig. 6-19). If one argues that this lag is the travel time rer

quired for the solar wind to reach its outer boundary, then, since = v w t, with vw ~ 500 km/sec and t ~ 8 months, the boundary

would be at about 70 AU. Measurements of the space gradient have suggested that the solar magnetic cavity is actually much smaller, and may extend out to only 5 or 10 AU. Furthermore, there is no a priori reason for regarding sunspots, a photospheric phenomenon, as a precise index of the relevant solar-wind parameters, which represent
the state of the corona. This
is

borne out by the virtual disap-

pearance of the lag

when

the intensity of the

5303A coronal green

COSMIC RAY INTENSITY VARIATIONS

line in the solar equatorial region
bers.
It is interesting to
is

131

substituted for sunspot

num-

note that, since the gyroradius of a 10 12

field of 5

eV

cosmic ray in a magnetic

of

X 10~ 5

gauss

is

of the order
particles

5AU,

the interplanetary field dominates the

limit.

motion of

with energies below this of the cosmic radiation.

These

particles constitute

99.9%

Many

details are

now

definitive experiments in space.

being subjected to close scrutiny through For example, the comparison of

the modulation experienced by different species, especially at

very low energies where the effects are greatest, provides sensitive criteria for

deciding upon the validity of various models that have been proposed. An alternative explanation of the modulation in terms of an electrostatic field had been suggested earlier, but was subsequently shown to be inconsistent with the more recent observations. However, the electric field produced by the time variations of the electromagnetic field associated with the fluctuating solar wind, and analytically resembling a heliocentric field, provides
the basis for a

somewhat

different

and useful

theoretical ap-

proach.

In any
final

case, theoretical

and experimental

studies of the cos-

mic-ray intensity modulations are being pursued vigorously.

The

answers

may

not be forthcoming until deep-space probes

lie far

are able to reach regions that

beyond the present

limits

of accessibility. Until journeys to great heliocentric distances

and

extended excursions from the ecliptic plane become feasible, local cosmic-ray measurements are likely to provide the only clues as to what it's like out there.

Solar Cosmic Rays

The kindly shine
scientific spy-glass,

of
is

summer, when tracked home with the found to issue from the most portentous

nightmare of the universe the great conflagrant sun: a world of hell's squibs, tumultuary, roaring aloud, inimical to life. ROBERT LOUIS STEVENSON
It is ironic that, for thirty years after the discovery of

cosmic

rays,

the sun cleverly concealed her ability to

breed of energetic particles. ing most of her offspring with a quota of energy that was not quite enough to enable them to propagate their effects through the earth's atmosphere.

spawn her own She protected her secret by endow-

Only on

very, rare occasions,

less

than once a year on the

average, does she give

them the

price of a ticket to the earth's

surface, so that, through their progeny, their arrival can be heralded by ground-based cosmic-ray instruments. Somewhat more

frequently, perhaps ten times as often, they are permitted to

penetrate a smaller distance, into the upper atmosphere

paratus,

the
that

stratosphere where often they are detected by balloon-borne ap-

and the ionosphere where they produce radio

effects that

are observed by ground-based equipment.

solar particles with such

And

it

now seems

low energies that they can be observed only with interplanetary space probes are being born practically
all

the time.

All of these solar energetic particles differ from the constitu-

wind in that they do not engage in the collecmotions that characterize the much lower-energy particles comprising the solar plasma. On the other hand, except for their birthplace, they are identical with their counterparts in the galactic cosmic radiation. Consequently, they are generally called
ents of the solar
tive

132

SOLAR COSMIC RAYS

solar cosmic rays.

133

of solar

As we review the early history of the study cosmic rays, it will become clear that this seemingly
is

paradoxical terminology

really quite appropriate.

years after Forbush had fired up the Carnegie chamber network, the only spectacular world-wide intensity fluctuations were all sudden decreases. Then, in rapid succession, came the first two solar flare increases shown in Fig. 7-1. Over a span of more than three decades, there have been

For the

first six

ionization

only three other events with a measurable flux of particles having sufficiently high energies to affect meson detectors.

On

the

other hand, neutron monitors have detected cosmic-ray intensity

enhancements at a long term average rate of about one per year. As is already evident in Fig. 7-1, the magnitude of the deviation from the pre-flare level falls off rapidly with decreasing latitude, indicating that the energy spectrum of solar particles is quite steep. In fact, only on one occasion, the all-time recordshattering event of February 23, 1956, has there been a clear indication of a flux enhancement at the equatorial station, Huancayo, where the geomagnetic threshold is 13.5 GV. In the early 1950's balloon flights at Swarthmore showed that some solar flares produce only low energy particles which are rapidly absorbed near the top of the atmosphere, and hence do not affect ground-based cosmic-ray detectors. But the time was not yet ripe to see the connection between these lower energy but less rare solar cosmic rays and an ostensibly remote radio phenomenon. For a long time, ionospheric effects related to solar disturbances had been noted at high latitudes. Understandably, observers thought that they were intimately associated with auroral phenomena and geomagnetic storms, as indeed they are. Finally, in 1957, D. K. Bailey showed that the radio-wave attenuation accompanying the event of February 23, 1956, could

The

not be considered merely as a special case of auroral absorption. following year, K. Anderson's balloon flight at Fort Churchill, Canada, provided the first direct confirmation that the particles responsible for the intense absorption of radio waves at
Thereafter, studies of radio-signal intensity records by a number of workers revealed that widespread radio blackouts were

high latitude were mainly protons.

134

COSMIC RAYS

26

28

"

February

6 March 1942

Increases of Cosmic Ray Intensity, February

28

and March
FIG. 7-1
First

7,

1942

recordings of solar cosmic rays. Similar occurrences of an increase

in

the meson intensity have been exceedingly rare.

caused by the arrival of low-energy solar protons throughout the polar regions in polar cap absorption (PCA), or solar proton
events.

The technique of continuously monitoring low-energy solar cosmic rays by making rf signal intensity measurements with radio receivers on the ground facilitated their detailed study by

SOLAR COSMIC RAYS

135

balloon and rocket-borne instruments. By launching these high altitude research vehicles at appropriate times either while PCAs were in progress or on the basis of predictions (or educated
guesses!) of

impending

solar activity, a great deal of information

has been amassed about solar cosmic rays.

Sporadic outbursts on the sun transpire explosively, hence, as might be anticipated, the energy, charge, space, and time distributions of solar cosmic rays vary significantly from one event to another, and even at different times during a single event. Consequently, the generalizations that we shall have to make in order to convey some notion of what happens are subject to notable exceptions. Of course, this is the reason for the unabated interest in studying solar cosmic rays. For each event tells us something new about the sun and the interplanetary medium.

Furthermore, spacecraft are continuing to provide new opportunities for definitive experiments, especially by opening up for investigation the extremely low-energy segment of the solar particle population. Indeed, we can learn much about some of the
dramatic" but more common cosmic acceleration processes by watching them in action on a more or less typical star the only one whose distance is measured in light minutes rather than
"less

in light years.

MORPHOLOGY OF A SOLAR PARTICLE EVENT

rays, let us consider the

Before summarizing some of the salient features of solar cosmic sequence of consequences of a solar flare.

Figure 7-2 shows the temporal relationships among all of the phenomena that may follow a great explosion on the sun. It

must be emphasized that most

indicated
effects.

flares produce only some of the Furthermore, the shape of the individual traces

depends upon the characteristics of the detector, and other considerations. With these qualifications, this picture can be regarded as typical of a ground-level event, that is, one in which enhance-

ment

of the particle flux at sea level

is

observed.

Electromagnetic Emissions. To observe flares visually, the solar disk is viewed through a narrow-band optical filter in the light of Ha (hydrogen alpha, the first line in the Balmer-series, 6563 A).

136
Horizontal Component ot Earth's Magnetic Field

COSMIC RAYS
Sudden Commencement of Magnetic Storm

Auroral Displays
Intensity of Cosmic Radio Noise Received at

High Latitudes
Polar
\( Radio

Cap Absorption
Blackout)

Auroral Absorption

Intensity of Low Energy Solar Protons

Intensity of Cosmic Radiation at Ground Level

Intensity of Solar Radio Noise

AV
Intensity of Solar Flare Radiation

(Ha Line, X Rays)

Flare Onset

12

16

20
in

24
Hours

28

32

36

40

Time

FIG. 7-2

Time-pattern of various effects observed

when very

energetic solar cosmic

rays are emitted during a great solar flare. Lesser flares

may

be accompanied by

some

of these

phenomena.

Following the notation of Table

to area

7-1,

a flare

is

rated according

and

brightness.

The ground

level events observed thus

far have, for the most part, been associated with the more violent eruptions that are assigned Importance 3 or 4B, or slightly less.

But sometimes even a

small

relatively

minor eruption can produce

PCA.

flash phase,

is an hour or two. The during which the Ha intensity rises rapidly to a maximum, may last for only a minute. Accompanying the visible

Typically, the duration of a great flare

SOLAR COSMIC RAYS

Table

137

7-1. International Astronomical Union Scheme for Assigning Solar Flare

Importance Ratings.

in

"Corrected" Area Square Degrees

Relative Intensity Evaluation

Faint
Sf
If

(f)

Normal
Sn
In 2n 3n 4n

(n)

Brilliant (b)

<2.0
2.1- 5.1 5.2-12.4 12.5-24.7

Sb
lb

2f 3f 4f

>24.7

2b 3b 4b

brightening of the chromosphere are

all

the other electromag-

netic products of the energy-transformation processes that comprise the awesome explosion. These include ultraviolet light, soft and penetrating x rays, and y rays on the one hand, and a broad range of radio emissions on the other. Ionospheric Effects. Photoionization by x rays promptly increases the electron density in the earth's upper atmosphere, thereby causing a sudden ionospheric disturbance, or SID. This

reaction to the arrival of electromagnetic radiation

may

involve

the D-region extending from about 50 to 90 km, the E-region from 90 to 160 km, and the F-region beyond.

There are several possible manifestations of an SID depending upon how the different layers of the ionosphere are affected. These include shortwave jade-outs (SWF) of radio transmission, sudden phase anomalies (SPA), sudden enhancements of atmospherics (SEA) and sudden cosmic noise absorptions (SCNA). The arrival of charged particles after a delay that depends upon their travel time from the sun to the earth (see Fig. 7-3) produces a massive increase in the ionospheric electron density. Inasmuch as low energy protons have access to high latitudes [Eq. (4-9) and Fig. 4-8], and since they are heavily ionizing in the lower ionosphere (cf. Figs. 7-3 and 2-10) they are especially efficacious in absorbing both man-made and cosmic radio signals.
was recorded (but not recognized) in the early communications network employing a mode of radio wave propagation discovered by Bailey and called ionospheric forward scatter. Although the detailed scattering mechanism is not understood, it can be ascribed to
first

The

PCA

1950's with a high-powered arctic

138
Rectilinear Travel Time, Earth to Sun, in Hours

COSMIC RAYS

30
I

20
I
i

10 8

6 5 A

2
i

1
i

0.5
1

0.2

20

10 10

10

1V

40

'i 50 9 55 -!10 LU
(

CO 60 -c

50
60
- Pa rticle
,\

iJ

10*

'

fi'S

- 10 8
CO

RO

7 9 J^

70
75

90 ^>>^
10 8

*10 7
c

~-

30
Proton

80
R?

10 5

40
50

10'

84
D

_J0 6 >
a>

60
70.

J 10b
10 5

86 -P10 5
c
LU

o104
a>

H 100" 90^/
1

80^

Ah

s ea Leve

-I^Okm^"^
-1
4-

*10 3
c

|io 2
cu
LlI

10"

--""I

0.01
= /3

0.1

1.0

Speed ot Particle Speed of Light

lati-

FIG. 7-3

Relationships

among

energy, rigidity, depth of penetration, lowest

tude of arrival and rectilinear sun-earth transit time for various particles.

arising from wind shear and turbulence in the mesosphere at a height of 70-75 km during daylight hours and about 85 km at night. A very small fraction of a VHF signal with a frequency exceeding the limit for the well-known garden variety of ionospheric reflection is returned to earth from a thin (~5 km) scattering stratum. The antenna patterns intercept the scattering layer at the mid-point between transmitter and receiver which, typically, may be 1000 km apart.
irregularities

The

received signal intensity in decibels

is

directly propor-

ambient electron density in the region of the scattering inhomogeneities. Thus, if the electron
tional to the logarithm of the

SOLAR COSMIC RAYS

density in the scattering layer
tensity increases.
is

139
greatly increased, the signal in-

ionization

lies

if most of the abnormal below the principal scattering level, the signal

On

the other hand,

intensity decreases,

because absorption

is

proportional to the

electron density along the path traversed by the radio wave. Dur-

when the increase in electron density during a PCA very great, the absorption effect dominates the enhancement effect, and the signal shows a net decrease. After sunset, however,
ing daylight,
is is caused by electron attachment to 2 mole owing to the cessation of photodetachment. Then, the enhancement effect dominates. Both of these effects are dramatically evident in Fig. 7-4 which shows the forward scatter record during

night-time recovery
cules

the spectacular 1956 event.

40 30 20
10

AA

'

Scattered Signal

(b)
Cosmic Noise -^

-10

~~
--Receiver Noise
1

-20 -30
1
1 1 1

00 06
FIG. 7-4

12

18

00 06

12

18

18 Hours, Universal
12

00 06

00 06
Time

12

18

00 06

12

18

Intensity of ionospheric

forward scatter signal and background cosmic

noise during the great solar particle event of February, 1956. (a) Thule to S0ndre
Strjfmfjord,

Greenland

(b)

Goose-Bay, Labrador to S^ndre Str^mfjord.

receiver with a

riometer (relative ionospheric opacity meter) is a radio wide-beam vertically pointing antenna. It measures the total absorption of cosmic radio noise in its passage through the entire ionosphere. This instrument, first used by

The

G. Little and H. Leinbach in 1958, observes SIDs, and auroral absorption that is frequently produced in the auroral zone by

140

COSMIC RAYS

the localized precipitation into the E-region of low energy electrons (tens of

keV

see Fig. 7-3), as well as

PCA
it

events.

Comparison of radio wave absorption and ments by rocket-borne instruments has made

in situ measure-

possible to de-

termine quantitatively the fluxes of solar particles from the observed signal attenuation. Figure 7-5 shows the calibration curve

S5.0-

10

20
in

50
Protons

Intensity

100 200 500 1000 /cm 2/sec/ster

FIG. 7-5

Cosmic noise absorption as a function of proton flux for a 30

riometer.

Mhz

for a 30 megahertz riometer. The absorption in decibels varies approximately as 1 /(frequency) 2 and, for a given incident spectrum, it is very roughly proportional to the square root of the
,

particle intensity.

Each ground-level event seems The magnitude of the counting-rate increase may range from barely perceptible (~1%)
Profile. to exhibit its

Neutron Monitor Time

own

distinctive features.

to staggering. Fig. 7-6

shows the February

23, 1956, solar

cosmic

ray event as recorded at Chicago.

stations reached a
7-7 is

The

nucleonic intensity at some

maximum

of 40 times the pre-flare level. Fig.

an example of a complicated sequence of events that was observed by the world-wide network of neutron monitors in November, 1960. The rise time to maximum may range from less than an hour to several hours, after which the intensity usually decays exponentially, returning practically to the background level in a few

SOLAR COSMIC RAYS

(a)
Expanded Scale Near Time of Flore Onset

141

8000
7000 -

-2500

-2000^

6000
5000 ~

L
(Extroi

S
FIc

(Sydney

of
col
-Opti
tburs

t
3

ty
ensi'

-1000..

4000
3000 _

t "o -'I
a>
Noisf
laximum

<u

of

^>
-t-

nset odio

-500 o?: fl / I ! 1
1 1

\
"V

2000

1000

o O OOOO O
inooom
Universal

/ f /

>

I
' 1

<
*
-

\
^*S.

/
,

/
/
1 1

^V^_

C3/

>
-o

^V.
^^^^
"^* ,,, ^
1

Ol
J

Time

UJ
|

0300

0400

0500

0600

0700

0800

0900

Hours -Universal Time

FIG. 7-6

Chicago neutron monitor observations on 23 February, 1956.

hours, or longer, depending

upon

the

magnitude of the

increase.
lati-

The enhancement

always exhibits a marked altitude and

tude dependence because solar particle spectra are steep the flux falls off very rapidly with increasing energy. Furthermore,
there
ter

may be

a geographic asymmetry indicative of guiding cenparticles traveling

motion of the

from the sun

to the earth

along the spiral interplanetary field lines. This anisotropy does not occur when the propagation is controlled by a diffusion

mechanism.

ENERGY SPECTRUM
Our
cosmic
old friend, the power law spectrum, encountered earlier
fits

in the discussion of galactic cosmic rays also usually

rays, at least

the solar

over limited ranges of energy. Thus, the data at a specified time t have often been represented by the follow-

142
1
1

COSMIC RAYS
r

= 1000

12

12

12

12
U.T.

12

12
15

Hours

Date 9

10

12

13

14

Nov.1960
FIG. 7-7

Observations

in bi-hourly intervals at

polar neutron monitor stations duris

ing November, 1960.

resolved

An unusual double-peaked structure on 12 November, 1960 when the time scale is expanded (15 minute intervals).

ing relationship in terms of energy, or a similar one in terms of

rigidity:

j(>E) = K(t)E-y\
As exemplified in
Fig. 7-8, the value of the

(7-1)

exponent

y(t)

(which

ordinarily ranges from 3 to 7) is appreciably higher than it is for the galactic cosmic rays. Furthermore, the power law applies in
the low-energy region below the
lactic differential

maximum

of the observed ga-

Thus, whereas the observed galactic differential intensity below about 1 GeV decreases toward lower energies, the solar particle flux increases. An alternative representation that usually covers a wider range
(cf.

spectrum

Fig. 5-2).

quite well, as illustrated in Fig. 7-9,

is

J(>P) =

Jo(t)

exp
is

l-P/P

(t)].

(7-2)

The

steepness of the spectrum

determined by

the e-folding

SOLAR COSMIC RAYS

10"

143

\
fe

\ \
\ \

10 3 ~
.

\ \
v

\
10 2

\\
\

\\
\

Sample Solar Particle Spectra

10

Galactic Cosmic

\
\

\\

\
\
\

*.

N\\
H\
\\
i

Ray Spectrum Observed at Solar Minimum

I

ST 0.01

0.001
Energy
FIG. 7-8

E in MeV

Comparison of power-law representations of energy spectra during several solar cosmic-ray events with the galactic proton spectrum.

10 3 F-

0.01

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Rigidity

in

GV
rigidity

FIG.

7-9

Exponential

representation

of

the

integral

spectra

of

solar

protons during different events.

144
rigidity
(flat).

COSMIC RAYS

which ranges from roughly 40 (steep) to 400 Values of ; are from about 2 protons/cm 2 /sec/ster to 8 X 10 3 protons/cm 2 /sec/ster. The spectrum generally steepens with time, that is, the relative number of high-energy particles falls off rapidly compared with the lower-energy component. The flux of energetic solar particles is frequently thousands of times greater than the galactic cosmic-ray flux of about 4 particles/cm 2 /sec, and peak fluxes of protons with energies above 10 MeV as high as ~10 5 particles/cm 2 /sec have been observed. The integrated flux in a single event may exceed a billion particles/cm 2
!

MV

MV

This obviously represents a big bundle of energy. We can make a crude guess about how much of its energy the sun imparts to solar cosmic rays in one of the largest events by multiplying (integrated flux at the earth) (average energy) (the area of the surface of the sphere at 1 AU):
(10 9

cm-

)(10 7 eV)[47r(1.5

10 13 cm) 2 ]

(2.8

10

42

eV)(1.6

10~ 12 ergs/eV)

~ 10
10 5

31

ergs.

(7-3)

This is comparable with the energy of about hydrogen bombs. Although this back-of-the-envelope type of calculation involves some glaring but compensating over- and under-estimates, it is in the right league. Calculations based on a diffusion model lead to the same result.

COMPOSITION
As in the galactic cosmic-ray beam, protons predominate, and alpha particles are also present. However, there is a most puzzling wide variation in the proton to helium ratio. Their spectra, expressed in rigidity, seem to be similar, but the ratio of protons
to

helium nuclei in the same

Fig. 7-10

rigidity interval

may

vary from

to 50.

shows how

this ratio

has varied with energy per nufixed energy,

it

cleoli in several events.

Above a

may

vary from

10 to 100.

145

50
Energy
FIG. 7-10

100

150
Nucleon

200

in

MeV /

Solar proton to helium ratio as a function of energy per nucleon at

various times.

Heavier nuclei are also present, although they are less abundant than in the galactic cosmic radiation. Unlike hydrogen, whose relative abundance is variable, the charge spectrum of the multiply charged nuclei remains constant, and is independent of energy. As Table 7-2 shows, the relative abundances reflect the chemical composition of the sun.

The

distinctive solar radio emissions that characterize all solar

cosmic-ray events clearly manifest the acceleration of electrons

in the flare region.
as

Radio astronomers designate

solar rf events

Types

to V.

steady "burst"

Type IV continuum radiation, a continuous, that may cover the entire radio spectrum, invari-

ably accompanies the emission of solar energetic particles.

But, unlike the heavier particles that, at least

when

certain

unknown

from the regions in which they are accelerated, most of the electrons remain trapped in the strong local magnetic fields. Here, those that attain relativistic velocity lose much of their energy by emitting
conditions prevail,
to escape

manage

synchrotron radiation in the radio-frequency region of the

elec-

tromagnetic spectrum.

146

COSMIC RAYS
7-2. Relative Abundances of the Elements in Solar and Galactic Cosmic Rays, and and Galactic Matter in General, Normalized to the Base Value 10 for Oxygen.

Table
in Solar

Solar Abundances

Galactic Abundances

Atomic
Element

Number Z
2 3

Cosmic Rays
1000

Matter
(4500)*

Cosmic Rays
500
2 6 18 8 10

Matter
1500

He
Li

<10~4
<io5
1
4

<10" 4
<io-
3 2 10

Be

4
5

<0.1
5

N
F

O
Ne

7 8 9 10
11

2 10

10
.01

<0.3

<io- 3
3 0.02 0.4
.04 0.4

Na

Mg
Al Sc P-Sc

12 13 14 15-21

0.4
0.3 0.6

(D*
.02

0.3
.02 0.4

2 3
0.6
1 1

Tc-Nc
*

22-28

<0.2

0.3 .05

0.2 0.3

The number in parentheses

represents the relative

abundance of this element

in the corona rather than in the photosphere.

Nevertheless,

some electrons do

escape. But, just as their ga-

eluded observation until recently. High-energy (GeV) solar electrons appear to be virtually absent. However, a large number of low-energy solar electron events have been observed with instruments aboard spacecraft. As the detector thresholds are pushed down to lower energies, both solar eleclactic cousins, they

trons

and protons are detected more frequently, until eventually (~1 MeV) they seem to be present even when they cannot be

attributed to specific solar active regions.

One might

expect that, in the great holocaust, some of the

protons accelerated during the flare would plunge deep into the solar atmosphere producing, among other things, a neutron albedo. Even when conditions are quiet, the fantastic nuclear
furnace would seem to be a prolific source of neutrons.
sun-earth transit time of a 10 8
in Fig. 7-3)
'-'17
is

The eV neutron (same as for a proton roughly equal to its mean lifetime against decay,
Table
2-1).

minutes

(cf.

Nevertheless, neutrons
It's

do not

ap-

pear on the roster of solar cosmic rays.

a puzzlement!

SOLAR COSMIC RAYS

147

PROPAGATION
It

sotropic, since their source

goes without saying that solar cosmic rays are basically aniis in our own backyard. But even
is

though the rectilinear distance

dimensions, there
earth, especially
is

short

sufficient

stuff
is

compared with galactic between the sun and the

when

the sun

hyperactive, to isotropize the

solar cosmic rays sooner or later

of the flare

and duration of the initial on the solar disk.

during a flare event. The extent anisotropy depends upon the position

Fig. 7-11 reveals that flares in the sun's western

hemisphere

Position on the sun's visible disk of observed flares that have proFIG. 7-11 duced solar protons with energies exceeding about 500 MeV (ground level event).

are

more

likely to

produce solar particles that can reach the

is

earth.*

The

reason for this heliographical favoritism

illus-

* The northern hemisphere also appears to be favored, but this is a consequence of the unexplained fact that, for some time, there has been a prepon-

148
trated schematically in Fig. 7-12.
force guide the solar particles,

COSMIC RAYS

The

spiral

magnetic lines of

the sun to the earth. There is whereas particles originating in the eastern sector have to cross the field lines. Thus, in the latter case, they diffuse toward the
earth, following a

and tend to funnel them from easy access from the western limb,

more tortuous path than do

the particles that

Active Region

Earth

FIG. 7-12

Schematic representation of the sun-earth turnpike

the route that af-

fords easier access from the sun to the earth for solar particles emitted from the

western segment.

can flow through a magnetic tube. Consequently, eastern hemisphere flare events are slow risers compared with the more favored western hemisphere events which are characterized by a rapid
onset after the flash phase of the
flare.

But, in

all cases, diffusion

derance of the more prolific active centers in this half of the sun. Another interesting point not brought out by Fig. 7-11 is that a "fixed" region on the sun appears to have been the seat of most of the energetic solar particles for some time.

SOLAR COSMIC RAYS

eventually sets in, and even
if

149
the propagation
is

very anisotropic
particles that

during did not

the initial stages (direct radiation)

many

start their journey on the sun-earth turnpike finally reach

the earth later (indirect radiation).

Fig. 7-13 reveals

how

far solar

protons

may

actually travel be-

observed relative intensity is of the distance traveled, computed for each plotted as a function
fore arriving at the earth.
10 4
5.7

The

MeV

3.8

MeV

10 3

10 2

10

10

100

1000

Distance Traveled, Astronomical Units FIG. 7-13

Normalized differential

intensities of solar protons as

a function of the

total distance traversed while enroute

obtained by multiplying

(1)

from the sun to the earth. The abscissas are the elapsed time between the flare and the arrival

of each energy group at the earth

by

(2)

the velocity corresponding to the observed

energy.

150

COSMIC RAYS

indicated energy component by multiplying the corresponding

particle velocity

Most of the

particles in this case traversed

by the elapsed time from event to observation. more than ten times

the rectilinear earth-sun distance.

The mean free path typically is of the order of 0.1 AU. Since memory of their initial directions is erased by scattering with magnetic irregularities (i.e. the particles random walk), we would
expect that, at least occasionally, some products of a
far side of the
flare

on the

sun should reach the earth. But until very recently, there were no known events in which this happened. The first example of the arrival of solar cosmic rays originating on the invisible disk is shown in Fig. 7-14. The extremely slow rise to maximum, the isotropy throughout the event, and the quantitative features as analyzed in terms of diffusion models for the propagation of the particles, all connote the unseen source. But
35
I I I I
I

28 Jan. 1967
'-_

30 '-

South Pole
:

25-

McMurdo
hf
Thule

M
Vi

20 -

/V

_ d 15-

CL

10 -

'-

/j

Swarthmore

^Oi

-j

i
i i

"

10

12

14 Hours,

16
U.T.

18

20

22

FIG. 7-14

First

example of the

arrival of very energetic solar cosmic rays

from

the far side of the sun.

SOLAR COSMIC RAYS

the burning question persists.

151

Why

doesn't this

happen more

often?
Fig. 7-12 clearly
fits

into our picture of the spiral structure of

is

the interplanetary magnetic field. It

instructive to consider

an

example of how

this

model can be deduced from cosmic-ray obis

servations alone. Fig. 7-15

i
i

a snapshot of the "flat"

i '

world dur-

-+80
</>

a>

+40
.1
+-

Sun

5-40
-80

"

Direction of

Symmetry

@
i

180W

120W 60W 60E Longitude Relative to the Sun

120E

180E

FIG. 7-15 Example of the variation in the neutron monitor counting rate increase above the p re-flare level in percent (numbers in circles) with asymptotic direction of viewing of a number of stations (circles).

ing a particular solar flare event.

the percentage intensity

It shows the dependence of enhancement upon direction of viewing

of a

number of neutron monitor stations. The direction in space about which the solar cosmic-ray flux was symmetrical, as well as the position of the sun at that time, are marked. We see that, in accord with Fig. 7-12, the apparent source was 55 west of the
sun.

same neutron monitor data are plotted as a 8 between the axis of symmetry and the asymptotic direction in space from which the solar particles to which the detectors responded came. No station was looking directly at the source, but the intensity was presumably maximum from the direction 8 = 0, i.e. 55 of the earth-sun line and 10N of the ecliptic. Although the individual directions viewed by the different stations surround the direction 8 = 0, all the data fall
In Fig.
7-16, the

function of the angle

along a single curve.

Spacecraft measurements
of

show how strongly low-energy

<;

tens

MeV)

solar cosmic rays can be collimated along interplanetary

field lines.

The

particle streams

sometimes exhibit a filamentary

152
200
I I 1 1 i
i

COSMIC RAYS

150

Churchill

I 100

>

Ellsworth

Y
50
Resolute

Thule

* Syowa
Mawsorr*^,
College

Mirny

100

20
Data from

40

60

80

120

140

S (degrees)
FIG. 7-16
Fig.

7-15 plotted as a function of the angle between the

direction of viewing.

indicated axis of

symmetry and the asymptotic

move through frozen-in magnetic tubes which guide the particles even though they twist
structure because they are constrained to
like spaghetti, in the

manner depicted

in Fig. 7-17.

The

entire

intertwined bundle co-rotates with the sun in the general Archi-

medes

spiral configuration.

Typical Dimension =

10 6 Kilometers
Gyroradii tor Gyroradii tor

= 21 = 1.6

13MeV BeV
1

Proton Proton

FIG.

7-17

Spaghetti-like tubes

of force

through which low-energy solar cosmic

flow.

rays

may

SOLAR COSMIC RAYS

153

ACCELERATION PROCESSES
Although considerable progress has been made toward underhow solar cosmic rays propagate, the mechanisms whereby they acquire their energy and are stored at and escape from the sun are still mysterious. All of the processes that have been proposed to describe the acceleration of solar cosmic rays are also relevant to the problem of galactic origin. Since we shall have occasion to discuss this subject in detail later, we need only identify them here. The Fermi mechanism* involves collisions of particles with moving magnetic scattering centers. In betatron acceleration,* charged particles are energized by a time-varying magnetic field. Dynamical plasma pinches and hydromagnetic instabilities are very complicated processes, for which theoretical calculations do
standing

not lead to complete conclusions, but require unattainable observational guidance. Unfortunately, the acceleration processes

take place in regions that cannot be observed directly. Perhaps

several different

mechanisms are operative. They may even

co-

operate

as,

for example, in the case of magnetic

pumping**
scattering by

which

is

a combination of betatron action

irregularities.

and

magnetic

In any case, the time in which the acceleration and/or the ejection of particles into space is accom-

plished seems to be extremely short, perhaps of the order of

minutes, roughly corresponding with the explosive phase of the

flare.

As an

alternative to regarding the flare as the

mother of the

energetic particles, one

may

speculate that the optical phenomein fact, quite analogous

belts

non

is

the offspring of trapped energetic particles that suddenly

is,

precipitate into the photosphere. This

to the relationship

on earth between the Van Allen

and

aurora.
Actually, asking the question
rive their energy

"How do
is

the solar particles de-

from the

flare?"

really
first

tantamount
place?" Vast

to asking

"What makes
* See p. 160. **Seep. 164.

the flare

happen

in the

amounts

154
of energy, adding

COSMIC RAYS

up

to as

much

as 1033 ergs, are

suddenly

re-

leased in the largest eruptions. This exceeds the entire thermal

energy stored up in the whole corona and chromosphere! And, we have noted, sometimes a few percent of this energy, of the same order as the fraction appearing as a light, is transferred
as

to solar cosmic rays. In seeking to look to the solar

an explanation, one recourse

key to
this

is

wind

as the

problem. Somehow,

catastrophically. Let us

plasma must slowly store magnetic energy, and then release it examine this possibility quantitatively. The energy density associated with a magnetic field of intensity B gauss is given by
pB

= B V8tt ergs/cm

3
.

(7-1)

The

largest flares
that, for

means

cm which occupy a volume of about an associated sunspot field of ~1000 gauss, the
10 29
3
,

magnetic energy is ~1033 ergs. The annihilation of could provide all of the energy that is released in the

this
flare.

field

SPACE RADIATION DOSES

In addition to
its

purely scientific importance,

solar cosmic rays

detailed

and of the conditions that presage a solar cosmic-ray event has immediate practical application to space activities. Exposure to excessive doses of radiation in the space environment could produce serious
knowledge of the properties of
radiobiological effects in astronauts. Since solar cosmic rays pose
the prime radiation hazard in man's ventures

beyond the magappropriate to

netosphere, especially

on extended

missions,

it is

comment briefly on this problem. However, although it is possible to make some remarks about potential dose rates on the
basis of past experience,
it

must be remembered

that evaluating

the effects of a given exposure constitutes a profoundly difficult

biological problem.

The
which
1

basic unit for

is

measuring radiation

is

the roentgen

(r)

defined as the quantity of radiation required to produce

esu of charge per cubic centimeter of air under standard con-

ditions.

The

roentgen equivalent, physical (rep) denotes an enis

ergy absorption of 93 ergs per gram, which

the

amount

of

SOLAR COSMIC RAYS

energy absorbed by
1

155

gram

of soft tissue exposed to

is

r of

rays.

defined as the radiation required to produce an energy absorption of 100 ergs per gram. The biological effects produced by 1 rad of different types of

The more modern

unit, the rad,

radiation can vary, although the energy absorption is the same. To take this into account, a factor called the Radiation Biological Effectiveness

(RBE) has been introduced

to relate the radio-

logical consequences of other radiation to that of

tion.

x or y

radia-

The

roentgen equivalent

man

(rem)

is

equal to the product

(RBE) (rad). At solar minimum, when the

galactic flux
is

is

about 4 particles/
particles/cm 2
,

cm

2 /sec,

the integrated daily rate

is 1.2

3.5
2
.

10 5

and

the yearly rate

10 8

particles/cm
7

The

corresponding doses

rad per year. Individual solar events can produce much higher doses over short periods, although excluding the very large eruptions, the yearly dose of solar cosmic rays is not alarmingly greater than that of galactic cosmic rays, and is easily handled with 3-4 gm/cm 2 of shielding.
are 20 millirad per

day and

However, the largest events have generated doses in the hundreds of rads. In one case, a surface dose of almost 1500 rad would have been encountered behind a 1 gm/cm 2 shield. Orbital electrons are dislodged and molecular bonds are severed when ionizing radiation is absorbed in matter. When this occurs in living cells, subcellular processes are disrupted, and the

normal functioning of the

to deoxyribonucleic acid

cell is

(DNA)

impaired. In particular, damage which is involved in the funda-

mental control of the genetic and functional activities may appear. Radiation damage in complex mammalian organisms consisting of billions of cells grouped into systems and organs may be somatic and genetic. Somatic effects are manifested directly in
the tissues of the organism,

and may occur

early

(minutes to

Radiation effects are a probaweeks) or late function of many factors. Thus, the responses to individbilistic ual solar flare events could range from nil through immediate symptoms that could affect the safety of the mission and even to

(months or

years).

death.

Many
vidual,

considerations, including the shielding

and the

indi-

would determine the outcome

of exposure to a particular

156
100

COSMIC RAYS

80

~X
Bone
Gastroi

60

Marrow

ntest.

Cerebral

40
-No Acute Effects -Acute Effect

20
-Increasing Probabilityfor Late Effects

10"

10"

10 1
Radiation Dose

10
in

10 3

10

rads

FIG.

7-18

Radiobiological effects

produced

in

particles. For

low energy

particles, the

dose tolerances

humans by penetrating energetic may be an order of magni-

tude higher.

solar cosmic-ray storm. Fig. 7-18 roughly represents the radio-

biological effects suffered by

cles.

humans exposed
is

to energetic parti-

Somewhat larger ated, and the percent

the right in this case.

skin doses of soft radiation can be toler-

mortality curve

shifted appreciably to

On

the average, 125 rads should not pro-

duce significant disease symptoms in man, and the skin can probably absorb ten times as much before affecting astronauts acutely.
Fig. 7-19 shows, for
tissue surface dose

one of the larger solar

flare events,

how

the

would have been reduced by

the addition of

1600

4
Shield Thickness

6
in

10

gm/cm 2

FIG. 7-19

Variation of tissue dose that would be sustained during a large solar

flare event with the thickness of

a protective aluminum

shield.

SOLAR COSMIC RAYS

157

shielding. Needless to say, all of these factors are taken into ac-

count in planning manned missions. Many studies have been directed toward the development of criteria for predicting the occurrence of cosmic-ray flares, and the sun is being maintained under continuous surveillance by a wide variety of observing instruments. Centers of activity are watched closely for symptoms suggesting that a proton flare may be imminent.

Since

we have not
on

yet completely mastered the art of predicting

earth,
it is

the weather here

perhaps not surprising that, even though remarkable strides have been made in forecasting the solar weather, there is still plenty of room for improvement.

The Origin of Cosmic Rays

There was a
star danced,

and under

that

was I born. WILLIAM SHAKESPEARE

The
lies

difficulty in

deciding

upon

the origin of cosmic radiation

not so

much

in inventing hypotheses, but rather in verifying

whether or not they are representative of what occurs in nature. Ever since their discovery, speculation about where and how cosmic rays are born has been rife. In this revolutionary new era of astronomy that started with the discovery of cosmic radio
noise, processes that are capable of giving rise to cosmic rays have been found to be necessary causes or consequences of other astronomical phenomena. But even proposals based upon observationally established physical principles must be examined not only qualitatively, but also quantitatively. The detailed knowledge that now exists concerning the cosmic rays themselves, on the one hand, and galactic structure and stellar evolution on the other, places rather severe restrictions upon theories concocted to answer the key questions that completely define the cosmic ray origin problem:
1

Where were they produced?

2.
3.

How

When

were they accelerated? were they born?

later?
this
is

4.
5.

What happened

Why (Perhaps

theological)?

Early attempts to answer these questions alternatively identicosmic rays as the birth cries or the death gasps of matter. Millikan first contended that cosmic ray energy represented the
fied

difference between the atomic masses of the universally

common

elements and the larger

sum

of the masses of the hydrogen atoms

158

THE ORIGIN OF COSMIC RAYS

159
to

which presumably combined

(cf.

form them in

interstellar space

Fig. 1-1).

He

later

advocated the annihilation theory which

envisaged the complete transformation in interstellar space of the rest mass of these same elements into cosmic-ray energy.

A great many ideas were proposed by many individuals, contemplating sources ranging from the sun to galaxies far beyond
our own, and invoking acceleration mechanisms running the gamut from electric fields produced in stellar thunderstorms to
the original act of creation of the universe.

Even supernovae,

currently regarded as a most likely cosmic-ray birthplace, were

already receiving serious consideration as early as 1934

when

W. Baade and

Zwicky drew attention to the enormous energy released in a supernova explosion, and its potentiality as a cosmic fountainhead. Before discussing the rationale for contemporary views on cosF.

mic-ray origin, let us recall some of the observed characteristics

for

which an acceptable theory must account

significant feature of the

the

facts that

keep

the theorist honest.

relative overpopulation of

composition of the primaries is the heavy nuclei compared with universal

abundances exceedingly great for the L group, and less drastic for the and H groups. Electrons are scarce and positrons are

considerably rarer.

Primary cosmic-ray nuclei have spent a very long time, at least millions of years, in traversing roughly 3 gm/cm 2 of matter before reaching the earth. They arrive isotropically and appear to have been around in very roughly the same numbers for perhaps longer than a billion years. Their energies extend at least up to 10 20 eV, their numbers varying inversely as an almost constant power of the energy above 10 9 eV. Below this energy, although the exact form of
the spectrum
is

uncertain, the differential intensity unquestionsolar cosmic rays.

,

ably continues toward lower energies until the galactic cosmic

rays

become indistinguishable from

energy density is roughly 10- 12 ergs/cm3 so the source must not only produce individual particles with energies a billion times that of the greatest laboratory accelerators, but it

The

160

COSMIC RAYS
also

must

put out a stupendous

total energy,
this

depending upon

the size of the

volume over which

energy density extends.

ACCELERATION MECHANISMS
Betatron Action. Although striking evidence for particle acon a galactic scale has been amassed in recent years, earlier attempts to account for cosmic-ray energies in terms of
celeration

then-known physical principles were based upon classical electromagnetic theory. Thus, in 1933, W. F. G. Swann pointed out that fluctuating stellar magnetic fields can, through electromagnetic induction, give rise to cosmic-ray energies. He emphasized that it is not the magnitude of the field, but rather its time rate of change that is important, and showed that fields varying at a rate comparable with that observed in sunspots, if spread over a sufficiently large area in a star, can easily produce 10 10 eV even with fields much weaker than these and, in fact, unobservable spectroscopically. The principle discussed by Swann was subsequently utilized in the betatron. In this device, charged particles are speeded up by a homogeneous magnetic field which increases with time. The mechanism is analogous to that of an ordinary electrical transformer, in which there is a change of magnetic field that threads through a wire circuit, inducing an electromotive force which drives the current through the circuit. Of course, the wire is not necessary for the attainment of the emf. Electrically charged particles that happen to be in the vicinity of the changing magnetic field are accelerated by the emf, even though they do not form part of a material circuit. Fermi Acceleration. The first quantitative theory which made predictions that could be compared with observations was proposed by E. Fermi in 1949. Most easily visualized in terms of a mechanical analogy, it is based upon the principle that charged particles can gain energy from random collisions with magnetized gas clouds under certain conditions. One might think of a system which is full of cannon balls that are moving randomly, corresponding to plasma clouds with frozen-in magnetic fields, and BB shot, representing cosmic rays.

THE ORIGIN OF COSMIC RAYS

Equipartition
to the
sets in, so that,

161

according to the principles of ther-

modynamics, the cannon

balls will ultimately lose their energy

BB

shot,

and

will reach a state of equilibrium in

is

which

the average translational energy of each

as

cannon ball that of each much more rapidly moving BB shot.

Mirror Region

the same

FIG. 8-1

"Collisions" of a charged particle with magnetic fields. The particle

it

may

mirror (top) or
field is in

may

be guided around a sharp bend (bottom).

If

the magnetic

motion, the particle

may

either gain or lose energy in the encounter.

Fig. 8-1

one, the particle spiraling

shows the types of collision envisaged by Fermi. In around a magnetic field line enters a
field strength,

region of greater

increasing

its

pitch angle (the

angle of the velocity vector of the particle with respect to the

magnetic

field)

until the orbital plane

is

perpendicular to the
other type of
reflec-

field direction, in

which case

it

mirrors like a geomagnetically

belt.
is

trapped particle in the

tion
is

Van Allen

The

one in which the particle

statistical

guided along sharply bent

basis for

field lines.

This

model provides a

computing the
,

re-

sulting energy spectrum. In head-on collisions between a particle

with energy

cle gains energy,

a magnetized cloud with velocity t/B the partiwhereas it loses energy in overtaking collisions. Because the former are more probable, there is a net gain which averages per collision

E and

dE = p 2E,

(8-1)

162

COSMIC RAYS

where

vB /c. Thus the incremental energy gain increases with /? the particle energy them as has gits!

After n collisions

AE =
whence an an energy

n0 sE

(8-2)

originally nonrelativistic particle will have attained

E=
If the

exp

2
(/3

n).

(8-3)
,

lisions

mean time between collisions is t c the number of experienced in time t is n /t c and the energy is
E(t)

col-

Ei exp

(/3*

~\ =
and

Ex exp
t

(t/t c ),
2
.

(8-4)

where

is

the injection energy

t c //3

Now,
tv

particles

are also lost, by catastrophic collisions with interstellar matter

and by leakage out of the system, in a mean time ability of surviving to an age greater than t is
P(>t)
Therefore, the

The

prob-

exp (-t/ti).

(8-5)

number

of particles with energy exceeding

is

j(>E) =
This
is

KE-VK
to Eq. (5-1)
t c /t x .

(8-6)

power law integral spectrum identical

with

the value of the exponent y

hence
galaxy

tc

1.5

tY .

Since the
estimate

is

at least as

long as the
th

mean mean

Experimentally, y = 1.5, lifetime for escape from the

collisions,

we can
is

time for removal by nuclear by recalling that primary cosmic

rays pass through

galactic disk

bering that

~3 gm/cm 2 of material. The density in the ~10- 24 gm/cm 3 so the distance traversed (rememgm/cm 2 = px) is 3/10~ 24 = 3 X 10 24 cm. The corre,
,

sponding time is 3 X 10 24 /c = 10 14 sec, or about 3 million years. In the galactic corona, where the density is about 10 -26 gm/cm 2

becomes 10 16 sec or 3 X 10 8 years. This looks nice, but there is a catch. The particles lose energy by ionization, and gain energy in the collisions. The gains outweigh the losses only for particles exceeding a certain velocity. The relationships between energy gain and loss, respectively, and particle energy for protons are shown in Fig. 8-2. The interthis

THE ORIGIN OF COSMIC RAYS

163

10'

10*

Average Gain

Average Loss

10'

10

B
in

10s

10

11

Energy
FIG. 8-2

eV

Dependence of energy
plied

gam and

Fermi acceleration mechanism to work, the

loss upon energy for protons. For the minimum injection energy must be sup-

by some other process.

section of the two curves

marks the minimum injection energy

it is

required to

make

the Fermi acceleration go. Although

only

about 200
gies

MeV
20

for protons, heavier particles, because of their

greater rate of ionization, require

oxygen and 300 GeV for iron. This mechanism works well for protons, but even if the exorbitant injection energy were provided by some other mechanism, it is

about

much

higher injection ener-

GeV

for

in trouble in accounting for the observed characteristics of the

heavier nuclei.

By

virtue of their fragmentation probabilities

they have correspondingly smaller

mean

free

paths than do

hydrogen

nuclei. Consequently, the observed similarity in the

energy spectra of the various components cannot be reconciled with this type of acceleration process.

number

of variations of this attractive principle have been

proposed. For example, the arms of a spiral galaxy are essentially

magnetic

bottles, similar to the

in fusion experiments.

As

is

shown

mirror machines that are utilized in Fig. 8-3, the magnetized

plasmas near the ends of a tubular trap

regions of denser magnetic
field.

may approach each

other, accelerating charged particles as they are reflected by the

And

it is

quite conceivable that

conditions in the expanding envelope of a supernova

may be

especially congenial to the operation of the process originally

conceived by Fermi.
that particles

Magnetohydrodynamic Mechanisms. H. Alfveh in 1959 showed moving through magnetic fields with irregularities

164
Tropping Region

COSMIC RAYS

Lines of Force

FIG.

8-3

Motion of mirror regions toward each other

is

may

accelerate particles.
ball

The effect

crudely analogous to

between the table

what happens when a ping pong and a paddle that is rapidly moved toward the

bounces

table top.

having dimensions very

the particles

much

smaller

than the gyroradii of

at each stroke of the

would gain momentum

in a process called magnetic

pumping.

If the increase in

momentum
momentum,
oc

pump

is

proportional to the
of a

the differential spec-

trum has the form

power law:
jXp) dp

p-i dp.

(8-7)

This can be doctored

tion.

to look like the observed

spectrum by

taking into account the diffusion of the particles after accelera-

This gives
j(p) dp oc p-***,
(8-8)

an unknown quantity in the range 1 < < 1, so that the result is not inconsistent with the experimental value y ~ 2.5. This type of mechanism, which can occur anywhere in space where the required conditions apply, has actually been utilized in plasma physics experiments. Observations of radio galaxies and of radio emission from quasars show that particle acceleration happens, even on a galactic scale, in cosmic explosions. Spontaneous processes analogous to laboratory plasma experiments, in which explosive instability of quiescent plasma leads to the acceleration of some particles, undoubtedly occur throughout the cosmos. There is a cornucopia of stored energy residing in magnetic fields. Plasma instabilities can cause the sudden conversion of magnetic energy to particle energy. For example, metastable surfaces of sharp

where

is

THE ORIGIN OF COSMIC RAYS

field reversal, called the

165
finite

cheek pinch, can store a

it

amount

of energy,

and then

release

explosively.

Electromagnetic fluctuations composed of plasma oscillations, Alfv^n waves, acoustic waves, or more complicated wave types known in plasma physics may also produce accelerations. The formation of shock waves provides a most effective means for imparting great energies to cosmic rays. Theories invoking this mechanism relate to a particular type of astronomical object in which conditions for this mode of energy transformation are extremely propitious the supernova. The occurrence of shock waves is inferred from wisps or ripples of light emanating from central regions of the Crab Nebula with speeds of about ^ c. S. A. Colgate and M. A. Johnson in 1960 first proposed a hydro-

dynamic process in which cosmic of an exploding supernova.

rays are

blown

off surface layers

SOURCES
Sun and Stars. When cosmic rays were thought to be electromagnetic in nature, the absence of a day-night variation immediately disqualified the sun from further consideration. But later, as the direction-randomizing effects of magnetic fields on the electrically charged primaries were recognized, the sun again
became
a serious contender.

Then, the observation,

albeit rarely,

of solar cosmic rays in the billion electron volt range coupled

with the realization that energetic particles might be stored by magnetic fields in a volume centered at the sun resurrected the possibility that at least a good part of the local cosmic ray population

was spawned

there.

eV have been observed only once. Clearly, the sun (and, hence, ordinary stars
Solar particles with energies of the order of 10 10
of which it is typical) cannot be the source of the higher energy end of the spectrum. Furthermore, it is easy to demonstrate that

the injection rate of all solar particles that can produce groundlevel effects (> 0.5

GV)

falls far

short of the required level.

Of

course,

we

haven't been in the monitoring business very long,

solar

and there may have been some very intense high-energy

166

COSMIC RAYS

cosmic-ray outbursts in the past. Could the particle density be maintained by storage in the interplanetary magnetic field? This question can be answered by referring to Eq. (4-5), which tells us that the radius of curvature r of a 10 12 eV proton in the s 5 X 10 _ gauss field is
rg

pc/300zB

~ 10

12

/(300)(5

10~ 6)

~ 10

14

cm.

(8-9)

This

is

cavity,

about 10 AU, comparable to the radius of the solar wind hence particles of this energy would readily escape from
inverse correlation between cosmic-ray intensity

the solar system.

The
activity
all

and

solar

was the coup

de grace of the solar origin hypothesis for

still

except the very low-energy particles that are certainly do-

mestic products. However, there

storage

and further acceleration on a

remains the possibility of galactic scale of cosmic

rays injected into the great galactic reservoir by millions of or-

dinary stars like the sun. But, the significant differences between the composition of solar and galactic cosmic rays indicate that

ordinary stars having the same relative elemental abundances as

the sun

do not

constitute the

main

source.

We must look

for

some

astronomical objects with special properties.

Supernovae.

The

list

of objects that differ from the typical

sun-like stars includes red giants, supergiants, magnetic stars,

novae, and supernovae.

The most promising

candidate

is

the

supernova.

The

theory of nucleosynthesis predicts that in supernova ex-

and oxygen are enhanced compared with hydrogen and helium. Furthermore, the relative amounts of the heavier elements are increased above their general abundances. Both of these predictions are in agreement with the cosmic-ray observations (cf. Table 5-1).
plosions the abundances of carbon

Supernovae are the last stage in stellar evolution. The aging star becomes hotter, producing progressively heavier nuclides in its
nuclear furnace, until
it

exhausts

its its

fuel.

As

it

loses its latent

rises,

heat, the star slowly shrinks

and
star

internal temperature

increasing the internal pressure.

tions

New
is

may

then occur.

The

exothermic nuclear reacdepleted of hydrogen first,

then helium. Heavier nuclei are synthesized from alpha particles,

THE ORIGIN OF COSMIC RAYS

167
at a

and the build-up proceeds. Eventually,

temperature of about

7 billion degrees, the rapid conversion of iron to

helium demands

energy that can be supplied only by gravitational potential energy. This causes an implosion in a matter of seconds, fol-

lowed by an unimaginable thermonuclear explosion when the

lighter elements
still left

in the outer layers fall into the collaps-

ing core.

A well-known example of a supernova is the Crab Nebula, observed in 1054 by the Chinese. At present it is an elliptical light years, and about 3500 object of dimensions roughly light years away from the solar system. This stellar explosion

6x4

was

visible in daylight for 23 days,

and

its

brightness was 350

million times the sun's!

Magnetic

fields

of~5xl0- 4
is

gauss exist in the Crab, and

is

strong synchrotron radiation

emitted. It

also the site of a

copius x-ray source of small dimensions (Tau X-l). X-ray sources

and cosmic-ray generators may belong

celestial objects.

to a

common

class

of

Other well-known supernovae in our Galaxy are Tycho's star A (1572) and Keppler's star (1604), both discrete radio sources. The rate of supernova explosions in this galaxy is about two per century. The stellar end-product of a supernova is a neutron star. These remnants of celestial detonations consist of nuclear matter, with
in Cassiopeia
densities of the order of 10 15

gm/cm 3
,

(roughly a billion tons per

cubic inch!). Their radii are about 10 km, their masses range and their temperatures exceed a from about 0.2 to 2

billion degrees Kelvin.

First recognized

through the repetitive bursts of radiofrequency

radiation which they emit, by virtue of which they were

pulsars, observations

named
rotate

have revealed that neutron

stars

with periods ranging from 33 millisec (the Crab pulsar, NP 0532) to 3.8 sec. The corresponding mechanical energy amounts to 10 51 -10 52 erg. Changes in their rotational periods have been -8 the case of the Crab pulsar AT = 3.70 x 10
observed, and in
sec/day. This corresponds to a change in the rotational energy

AE ~

10 38 erg/sec,

the

amount required

to

make

the

Crab

shine. In addition to rf radiation, emissions covering a wide

168

COSMIC RAYS

frequency range, including the optical and x-ray regions of the

spectrum, have been detected.

The

rapid rotation of a neutron

field,

star,

coupled with

its

10 13

gauss magnetic

appears to be profoundly related to the

As T. Gold pointed out in 1968, the pulsar's magnetosphere attains the velocity of light quite close to the rapidly spinning object (e.g. 1700 km from NP 0532), and this can indeed provide the means whereby particles acquire cosmic ray energies. Roughly 40 pulsars, a few hundred parsecs apart and 2500 light years away on the average, have been observed.
origin of cosmic rays.

OUR GALAXY
The
galaxy in which our solar system
is

located, the

Milky

Way, is a rotating quasi-spherical mass of 10 11 stars with dust and gas and wandering masses of plasma in the space between. It consists of the disk and the quasi-ellipsoidal halo, as is shown in Fig. 8-4. The volume of the halo is about 50 times that of the

FIG. 8-4

Highly idealized

map

of our galaxy

the Milky

Way.

disk.

The

radius

light years

R d and thickness h d of the disk are about 50,000 and 3,000 light years, respectively. The distribution of
I

Population

(young) stars

is

similar to that of the gas clouds

which move randomly through the interstellar space. Population II stars, concentrated around the galactic center, display random motions of about 100 km/sec or more, reflecting turbulent conditions early in their lives. Curving out from the galactic nucleus is a spiral structure, as represented schematically in Fig. 8-5, and the sun is located about two-thirds of the way out in one of the
spiral arms.

THE ORIGIN OF COSMIC RAYS

169

FIG. 8-5

Spiral structure of the galaxy.

is about 200 million years, approximately 10 billion years. In interstellar space, the density is ~10 -24 gra/cm3 (1 hydrogen atom per cubic centimeter) in the disk and about a hundredth of this in the halo. But there are also large islands with dimensions of some ten light years in which the density may be as much as 100 or 1000

The

rotational period of the Galaxy

and

its

age

is

times greater. These clouds of ionized gas are characterized by

frozen-in magnetic lines of force.

The

turbulent motion produces

disordered magnetic

fields

and

large scale ordered fields, as evi-

denced by the polarization of

passes.

starlight

due

to the

magnetic

ori-

entation of nonspherical dust particles through which the light

The

galactic

magnetic

field

amounts

to

something of the

order of 10 -5 gauss in the disk, and 10 -6 gauss in the halo. Stronger magnetic fields are present in discrete radio sources.

ORIGIN MODELS
In the formulation of theories of cosmic-ray origin, there
choice between galactic and metagalactic models.
called the
is

a
is

The
g
is,

latter

homogeneous model

if

the energy density

W ~ 10 -12
a region

ergs/cm3

pervades the entire visible universe, that

with dimensions

of the order of the photometric radius

R ph ~

170
5

COSMIC RAYS

1027 cm. In this case, the total energy in the form of cosmic
all

radiation exceeds that of

other forms of energy except the mass

of matter in the universe. Local metagalactic models confine the

sources to regions with dimensions 10 26

galactic size
is

cm

<;

R<

10 24 cm.

The

R~

10 22

10 23

cm.

The disk and halo models differ mainly in terms of the volume occupied by cosmic rays. Whereas, in the first case, the volume 10 68 cm3 V d ~ R A 2 h d ~ 10 67 cm 3 in the second case V h ~ R h s hence the total energy of cosmic rays differs by one order of mag,

nitude:

Ed ~ Eh ~

W V ~ 10 W Fh ~10
e

55
56

ergs, ergs.

1Q)

For purposes of comparison, let us list some significant cosmic energy densities. In addition to (1) the classical cosmic rays that
constitute the subject of this

book

energy density
~3 X
.

~1 eV/cm3
,

there are

(2)

the very soft photons comprising the background

radiation from a big bang

energy density ~0.4 eV/cm

photons, x rays and y rays and (4) neutrinos energy density

forms of energy are

3
;

2 ~ magnetic field in galactic disk g /&neV/cm 3 ~0.3 1 eV/cm (6) starlight in our part of the Galaxy and (7) random motions of gas clouds in our Galaxy pv 2 ~

energy density

<10 keV/cm 3

3 (3) hard 10~ 5 eV/cm3 Other relevant

,

(5)

0.2

eV/cm 3

It is interesting

and probably quite coincidental

that

some of

these apparently unrelated quantities are so similar.

On

the other

hand,

it is

known

that in plasma, magnetic fields are coupled

to particle motions.

In the financial world, the balance sheet reveals the fiscal soundness of an enterprise. Similarly, to be plausible, a theory of cosmic-ray origin must provide a balanced energy budget. This means that within the volume assumed to be occupied by cosmic rays there must be sources that are capable of maintaining the corresponding total energy carried by cosmic rays distributed throughout that volume. Consequently, arguments about the extent of the source region devolve

upon

the efficiency of the

galaxy and

its

immediate surroundings

as a

trapping volume,

THE ORIGIN OF COSMIC RAYS

171

and the
sources.

relative efficiencies of galactic

and metagalactic energy

is

The
tered

choice between galactic vs extragalactic origin

sources

cen-

upon whether supernovae and explosions nucleus on the one hand, or the strong radio
the frequency of violent events

in the galactic

on the

other, are the principal cosmic-ray sources. Estimation of both

and the energy released per

is

event

is

required in either case.

the key

Still

a controversial matter, the distance of quasars

to the extragalactic hypotheses.

What

is

the nature of a quasar?

The
jects

red shift of the lines of the spectra of these quasi-stellar ob-

(QSO) can be explained either by

(a) their

participation in the

general expansion of the universe, or by

(b) their ejection at the

observed speed from the galactic nuclei of nearby radio galaxies

(i.e.

40 Mps)* or by

(c)

gravitational displacement of lines emit-

ted by gas in the central part of an accumulation of neutron

The cosmological viewpoint (a) now seems to have the edge. Homogeneous Metagalactic Model. The density of radiogalaxies is about 1.3 X 10- 4 /(Mps) 3 ~ 4 x 10- 78 /cm 3 Therefore,
stars.
.

on the average, each must

10- 12 /(4 x 10- 78)
largest galaxies
is

inject cosmic rays with a total energy

~2x

KM"* ergs

10 11
.

usually

<10

12

The mass of the The maximum attainable

Q
2
.

Mc

2 This 10 62 ergs or about 10 7 Qc -16 ergs/cm3 give a cosmic-ray density less than ~10

value

is

more

like 10 61

would

Local Metagalactic Model. In the case of the local metagalactic model, the nearest potential source is the radio galaxy Centaurus

Mps. To fill a volume V mg 10 75 cm 3 with -12 ergs/cm 3 requires injection of 10 63 ergs. But a density ~10
at a distance of 3.8

only about 10 59 ergs are available for particle acceleration. Similarly, gravitational collapse to

produce 10 63 ergs requires a mass

M~
If

~ 2 X 1011 M. whereas for the entire galaxy, a large part of the cosmic radiation comes from extra10 12
,

galactic sources, the intensity

over a period of

would show a large time variation than about 10 8 years, which seems not to be in accord with observation. Nevertheless, such a source appears
less *1

Mps =

106 parsecs

=3X

102* cm.

172
to

COSMIC RAYS

be required to account for the particles at the top of the energy

spectrum.
Galactic Model. The question of the disk and halo vs disk alone affects the high-energy end of the spectrum as well as the

power that must be furnished by individual sources

total

to

meet the

energy quota.

1056 er S an(j g
years, the

d ^

Eq. (8-10) the total energies are E h joss er S g ut esC ape from the disk is easier, g

From

since the trapping time

is

shorter.

Thus, for a lifetime of 10 6

(3

power must be
1035 watts.

10 55

erg/(10 6 yr)
10 56

10 7 sec/yr)

1041

ergs/sec

In the halo, where the lifetime

is

is

~10 8

years, the required

power

ergs/(10 8

yr)(3

10 7

sec/yr)

1040 ergs/sec

1034 watts.

To summarize,

the parameters of a plausible

model

are:

R ~3 -

5 X 10 22 cm, Volume ~10 68 cm 3 Total Cosmic Ray Energy ~10 56 Escape Time ~3 X 10 8 years,
,

ergs,

Power

of Sources

~10 40

ergs/sec.

Supernova outbursts in our galaxy occur about once in 50

years (1.5

109

sec).

Therefore, the

mean energy

transferred to

cosmic rays per outburst must be (10 40 ergs/sec) (1.5 X 10 9 sec) 10 49 ergs. This is not asking too much. Besides, minor outbursts in the galactic nucleus that could also provide the required power
of about 10 40 ergs/sec have been observed.

But a prime candidate

the neutron
star.

for the role of cosmic ray

spawner

is

With about

10 42 ergs/sec going into plasma

energy, the Crab pulsar

NP

power needed
galaxy.

to supply the

0532 produces about 30 times the cosmic ray energy density in the

Hierarchy Theory. The increase in the slope of the cosmic-ray energy spectrum at about 10 15 eV (see Fig. 5-1) has been attributed to the inability of the galaxy to retain particles with rigidities exceeding 10 15 10 16 volts. Referring to Eq. (8-9), the cor-

responding radius of curvature of a proton in a homogeneous magnetic field of 10~ 5 gauss is about one light-year and of an iron nucleus 0.04 light-years. It must be remembered in all our discussion of the bending of charged particles that the radius of

THE ORIGIN OF COSMIC RAYS

173

curvature could be considerably larger, and the container more

transparent,
if

the field does not maintain

larity over the entire trajectory.

some degree of reguAs protons start to wander out

of the galactic fold at the energy corresponding to the kink, the

composition changes in the direction of an increase in the relative abundance of heavy nuclei (i.e. higher Z), for which the magnetic containment is more efficient.

Above

10 18 eV, the flux of galactic cosmic rays

falls

below that

of extragalactic particles, since near the upper limit of observed

mechanism breaks down. For example, at 10 20 eV, the cyclotron radius of a proton in the <10 -5 gauss galactic field would be >3 x 10 4 light-years. The correenergies, the galactic storage

sponding diameter of

>6 x

104 light-years

is

about 20 times the

thickness of the galactic disk. Undoubtedly, the halo plays an

essential role in storing the

more

energetic particles.

But

it

is

inconceivable that the galaxy could retain 10 20

eral of

eV

protons, sev-

which have been observed.

recent discovery of low-energy black-body radiation, in-

The

remnant of the primeval fireball, has important implications with regard to the problem of cosmicray origin, particularly with respect to the distance of the sources. This thermal microwave radiation corresponding to a temperaterpreted as a red-shifted
ture of 3 K is assumed to pervade all space uniformly. It imposes an upper limit on the expected spectrum of cosmic-ray protons through the action of pion photoproduction, which has a threshold at 10 20 eV and a rapidly rising cross section. For a particles and heavier nuclei, the upper limit would be encountered at

lower total energies.

The

reaction

is

+ p-^o + p + + n.
7r
,

With an average cross section of 2 X 10- 28 cm 2 and a photon density of 600 photons/cm 3 , Eq. (2-12) gives the mean free path
A

= l/Na ~
near 10 20

1025

ter of the "local off

10 7 light-years. This is roughly the diamegroup" of galaxies. Thus, the high-energy cuteV would prevail unless the most energetic cosmic

cm~

rays observed thus far are generated within this radius.

174

COSMIC RAYS
affect

Another reaction that would

distances in the metagalaxy
is

particles traveling large

7
This also tends
the spectrum.

p-+e+

+ e- + p.

(8-12)

to reduce the energy at the high-energy

end of

As
tion

to the origin of the primary electrons, the various accelera-

mechanisms that we have discussed are expected

to

produce
colli-

negative electrons exclusively, since, presumably, positrons are

not available. Therefore, the latter can originate only in

sions of other forms of cosmic radiation with matter

in the

Those interactions of cosmic-ray protons with interstellar hydrogen in which both protons survive yield equal numbers of positive and negative pions, which decay into positrons and electrons in the scheme 7r-|ii-e. However, when protons turn into neutrons in the collisions, tt+ are emitted, and a positive excess results. Thus, the large observed excess of negative electrons indicates that most of the electronic component is accelerated directly, and is not the product of collisions of other types of cosmic radiation with interstellar hydrogen. Supernova remnants seem to be their most likely source. In any event, the cosmic-ray electrons that reach the earth must all originate within the galaxy, since the background microwave photon radiation would obliterate extragalactic electrons by an inverse Compton effect that would convert their energy into x rays during the course of a longer journey. Furthermore, x-ray
galaxy.

metagalactic space

data indicate that the energy density of relativistic electrons in -17 ergs/cm 3 which is very is less than ~10
,

much below

the observed cosmic-ray electron energy density of

.

10 -14 ergs/cm3

The mixed,
sources.

or hierarchy, theory of the origin of cosmic radia-

tion ascribes different parts of the cosmic ray family to different

must be explained would seem unique source and acceleration mechanism. Indeed, a number of different processes are capable of producing cosmic rays and are probably operating somewhere. This is brought home by the mounting evidence that charged particles are accelerated even by phenomena occurring within the earth's magdiverse features that
to preclude a

The

THE ORIGIN OF COSMIC RAYS

netosphere.

175

A
if

finite flux is actually injected into interplanetary

space and,

we

follow the historical pattern,

these particles terrestrial cosmic rays. This

we should call would surely con-

found the semanticist! Perhaps there is some philosophical justification in the belief that if man can conceive of a physically sound method for producing cosmic rays, it is probably operative somewhere in the universe. In any case, the burden of telling us why such orderliness prevails amidst such a plethora of cosmic ray sources may have to be borne by the answer to question number 5 Why?

We

But the

dance round in a ring and suppose, secret sits in the middle and knows. ROBERT FROST

+5

o
^2

as

H^

18

+
u u c u
u
p*

fri
<u

JS

+ +
V

-r

++
<:*

+
1 1

',

+
a,

+
<-

o 8

oo T
i i

T-H

XX
CM vo
cm' cm'

4s

V
H|

W3

H|Cf

oo

s-

s:

so

*
V X
<m

CO

CM CM

D
-^

_G

is 3z

3 o
-

s S
kS -2

w S
176

-3

^P*

PL,

Bibliography
BOOKS
A. E. Sandstrom, Cosmic
Co.,

Ray

Physics (North-Holland Publishing

Amsterdam, 1965). B. Rossi, Cosmic Rays (McGraw-Hill, Inc., New York, 1964). V. L. Ginzburg and J. I. Syrovatskii, The Origin of Cosmic Rays (Pergamon Press, Oxford, 1964). V. D. Hopper, Cosmic Radiation and High Energy Interactions
(Prentice-Hall, Inc.,

New Jersey,

1964).
Inc.,

A.

W.

Wolfendale, Cosmic Rays (Philosophical Library York, 1963).

1963).

New

T. E. Cranshaw, Cosmic Rays (Oxford University Press, London,

B. Peters, Editor,

Cosmic Rays, Solar Particles and Space Re-

search (Academic Press,

E.

New York,

1963).

N. Parker, Interplanetary Dynamical Processes (John Wiley

and

Sons, Inc., 1963).

H. Alfve"n and C. G. Falthammar, Cosmical Electrodynamics (Clarendon Press, Oxford, 1963). A. Ramakrishnan, Elementary Particles and Cosmic Rays (Pergamon Press, Oxford, 1962). C. F. Powell, P. H. Fowler, D. H. Perkins, The Study of Elementary Particles by. the Photographic Method (Pergamon
Press,
J. E.

New York,

1959).

Hooper and M. Scharff, The Cosmic Radiation (John Wiley and Sons, Inc., New York, 1958).
Berlin, 1953).

W.

Heisenberg, Editor, Kosmische Strahlung (Springer-Verlag,

B. Rossi,

High Energy

Particles (Prentice-Hall, Inc.,

New

York,

1952).

177

178

BIBLIOGRAPHY

D.

J.

X. Montgomery, Cosmic Ray Physics (Princeton University

Press,

New Jersey,

1949).

Cosmic Rays (Clarendon Press, Oxford, 1948). W. Heisenberg, Editor, Cosmic Radiation (Dover Publications,
L. Janossy,

New York,
E.

1946).

Miehlnickel,
Steinkopff,

Hohenstrahlung Germany, 1938).

(Ultrastrahlung)

(Theodor

SERIES OF
Annual Review Annual Review
fornia).

BOOKS AND JOURNALS CONTAINING REVIEW ARTICLES

of

Astronomy and Astrophysics (Annual Reviews,

Nuclear Science (Annual Reviews,
Inc., Cali-

Inc., California).

of

Handbuch

der Physik [Encyclopedia of Physics] (Springer-Verlag,

Berlin).

Proceedings of the International Conference on Cosmic Rays: Eleventh, Budapest, Hungary, 1969 (Acta Physica Hungarica). Tenth, Calgary, Canada, 1967 (Canadian Journal of Physics).

Ninth, London, England, 1965 (The Institute of Physics and the Physical Society, London).
Eighth, Jaipur, India, 1963 (unpublished). Seventh, Kyoto, Japan, 1961 (Journal of the Physical Society
of Japan). Progress in Elementary Particle and Cosmic

Ray

Physics (North-

Holland Publishing Company, Amsterdam). Progress in Nuclear Physics (Pergamon Press, London).
Progress of Theoretical Physics (Research Institute for Funda-

mental Physics, Japan). Reviews of Geophysics (American Geophysical Union, Virginia). Space Science Reviews (D. Reidel, Holland).

Symbols, Abbreviations,

Typical Units,

and Values

of Various Quantities

in a nuclide. amplitude area cm AU astronomical unit (mean earth-sun distance) = 1.496 X 10 cm. barn section = 10~ cm B magnetic induction B magnetic B interplanetary magnetic in a medium. mspeed of dimpact imparameter cm). dE/dx energy MeV/gm/cm Ddiffusion (cm eVelectron volt eV = 1.602 X 10Ekinetic energy E energy in radiation = (dE / (dE /dx) ^V thickness of galactic /intensity of electromagnetic radiation (photons/cm j(E)dE or dj(E)/dE unidirectional intensity between E and E + dE particles/cm /sec/ster/MeV). j(P)dP or dj(P)/dP unidirectional intensity between P and P + dP particles/cm /sec/ster/GV). /(>) integral unidirectional intensity above energy E
A A AB
(percent).
2
).

aradius

x 10 3 km. atomic mass number = number of nucleons

of earth
6.378
(e.g.

13

cross

24

(gauss).

g
a

galactic

field (gauss).

field (gauss).

light

(e.g.

specific

loss (e.g.

).

coefficient
(1

2 /sec).

12 erg).

(ev).

critical
rad-

processes:

ion

ii d

disk.

2 /sec/ster).

differential

(e.g.

differential

(e.g.

(e.g.

par ticles/cm 2 /sec)

179

180

SYMBOLS

i?i

ionization ion pairs/cm). shower unit cm). LY year = 9.460 x 10" cm. Myg mass, or mass of moving in coordinate system in which speed v MDdipole moment of the earth = X 10 gauss cm. M mass, or mass of in coordinate system in which speed v m electron mass = 9.109 x 10 gm. M mass = 1.989 x 10 gm. M mass energy = 1.789 x 10 N particle density particles/cm N(P) counting attributable primaries with P. pmomentum psparsec = 3.26 years = 3 x 10 cm. Pmagnetic = pc/ze = Br (gauss cm). P threshold (GV). P minimum required reach earth the equator from favored direction = 59.6 GV. P cutoff (Stormer cone) = A GV. radenergy absorption =100 ergs/gm. electron radius = /m c = 2.818 x 10 -13 cm. reproentgen equivalent physical = 93 ergs/gm. gyroradius or cyclotron radius cm). interplanetary magnetic cavity AU). R range gm/cm astronomical dimension Mps). unit angle atmospheric pressure = mm of Hg). U energy = kinetic energy E + energy M c GeV). vB speed of plasma cloud km/sec), speed of cm/sec). vwspeed of wind km/sec). Wenergy of electromagnetic radiation MeV).
/t
/

ji>P) integral unidirectional intensity above rigidity par ticles/cm 2 /sec) primary specific ionization (e.g. ion pairs/cm), ;p
total specific
(e.g.

(e.g.

(eg.

light
total

particle

its

is

/3c.

8.1

25

rest

particle

its

is

0.

rest

28

solar

33

ec

solar

54 ergs.
).

(e.g.

rate

to

rigidity

(ev/c).

light

18

rigidity

(volts)

vertical

rigidity

rigidity

for particle to

at

least

By

vertical

rigidity

14.9 cos 4

re

classical

e2

rg

(e.g.

field

(e.g.

(e.g.

).

(e.g.

ster

solid

(steradian).
(1

torr

torr

total

rest

(e.g.

(e.g.

r/p

particle (e.g.

solar

(e.g.

(e.g.

SYMBOLS

181
3
).

radiation length gm/cm charge in units of electron charge. Z atomic number = number of protons
X
(e.g.
2
).

W energy density in galaxy (ergs/cm

g

in nucleus (and elec-

trons in neutral atom).

GREEK SYMBOLS

structure constant = e /hc = 1/137.04. percent for component barometric pressure per particle speed/speed of light = v/c. = 1/A Pi absorption linear absorption (cm (cm /gm). mmass absorption phase angle. space = 8.854 x constant) in permittivity
a
fine
2

at

coefficient

(e.g.

torr).

ft

coefficient

-,

f}

coefficient

- x

).

(3

coefficient

(dielectric

free

10- 12 farad/m.
dx. energy in radiation processes: $(dE/dx) gm/cm interaction mean path path gm/cm attenuation) mean A absorption A geomagnetic latitude micrometer = 10 cm. frequency density gm/cm barns. barn = 10 -24 cm section a barns. a geometric section = 0.05A mean lifetime = lifetime of unstable particle in coordinate moving which system with respect bombardment or radiation age proper lifetime = lifetime of unstable particle in coordinate which system with respect particle flux particles/cm Oangular velocity of sun = X 10- rad/sec.
cc

critical

ion

Xa

At

free

(e.g.

).

(or

free

(e.g.

).

(degrees).

/a

(hertz).

(e.g.

).

cross

(e.g.

).

cross

1 /3

to

it is

(sec).

tc
t

(e.g. year).

to

it is

at rest (sec).

4>

(e.g.

2 /sec).

2.7

Index

Absorber thickness, 18-19

Absorption, 6, 26 length, 19-20 processes, 60 Absorption coefficient,
linear, 21

Bohr, N., 28

Bombardment
Boron
5, 7,

age, 100

trifluoride counter, 45, 61

20-21

mass, 21 Acceleration mechanisms, 13, 160 Acceptors, 47 Accidentals, 54 Acoustic waves, 165
Activator, 49

Bothe, W., 10 Bowen, I. S., 6 Bradt, H. L., 91 Bragg, W. H., 28 Bremsstrahlung, 23, 25, 32, 35

Brunberg, E. A., 78 Bubble chamber, 52 Bulk conduction counters, 46 Burton, E. F., 3

Calorimeter, 57 Cameron, G. H., 8 Carbon dating, 99 Carlson, A. G., 28

Albedo, 92-93 splash, 92 reentrant, 93

Alfven, H., 104, 163

Alfven velocity, 110 Alfven waves, 165

Allowed region, 69, 72 Anderson, C. C, 90 Anderson, K., 133 Anticoincidence arrangement, 58, 60 Anisotropy, 106 Annihilation, 16, 159 radiation, 28 Anode, 43 Archimedean spiral, 111, 152
Aston, F. W., 8 Asymptotic cone, 79, 119

Cascade process, 57, 82 Cathode, 43 Center of activity, 157 Centered dipole, 70, 74 Cerenkov counter, 49-50, 59-60 Cerenkov radiation, 35-36, 50

Chance coincidences,

see

accidentals

Chapman,
Charge, 21

S.,

109

Cheek pinch, 165 Chree analysis, see superposed epochs Chromosphere, 107, 154
J., 9, 10 Clock diagram, see harmonic dial Coincidence arrangement, 10, 46, 55 Colgate, S A, 165 Compound nucleus, 40 Compton, A. H., 7, 10, 28, 105

Asymptotic direction, 78, 79 Atmospheric cutoff, 66 Atmospheric depth, 19, 87

8, 158 23, 25 Attenuation, 20 Aurora, 9 Auroral absorption, 133, 139 Autocorrelation, 125

Clay,

standard, 63 Atom-building,

Atomic radius,

Compton scattering, 16, 25-26, 37-39 Cone of acceptance, see asymptotic cone Continuum radiation, 145
Convective envelope, 106 Convective removal, 128 Cooke, H. L., 2 Core, 106 Corona, 107, 110, 154, 162 Coronal green line, 1 30 Cosmic ray components, 16 Cosmic ray "gas", 119 Cosmic ray isotropy, 13, 159 Cosmic ray observatories, 62 Cosmic ray secondaries, 15 Cosmic ray storm, 121 Cosmogenic nuclides, 100 Coulomb field, 21, 23, 28 Counter-controlled chamber, 51 Counting rate plateau, 67 Coupling coefficient, see differential response
function

Avalanche, 44

Baade, W., 159

Background radiation, 170 Bailey, D. K., 133, 137

Balloon, 4-5, 58, 63, 81, 91, 93, 132-135 Barn, 20 Barometric pressure coefficient, 88 Barry, J. G., 91 Baryon isobar, 86 Becquerel, H., 2 Bergwitz, K., 3, 4

rays, 9

Betatron acceleration, 153, 160 Bethe, H. A., 28 Bhabha, H. J., 28 Biermann, L., 109 Binding energy, 22 Bipolar spots, 108 Black-body radiation, 173 Blast wave, 123 Bloch, F., 28

Crab Nebula, 165, 167 Crab pulsar, 167, 172

Critical energy, 35 Cross section, 19-20, 38

Cylotron radius,

see

gyroradius

182

INDEX
vs E detector, 46, 58-59 Delta rays (5-rays), see knockon electrons Density gradient, 128 Depletion layer, 47 Dielectric solid detector, 54 Differential energy spectrum, see primary cosmic

183
onset time, 121 preincrease, 122 predecrease, 122 recovery, 121
Fossil tracks, 102

dE/dx

rays Differential response function, 87 Diffused junction, 47 Diffusion, 144, 148, 150 Diffusion cloud chamber, 51 Diffusion coefficient, 129

Fourier analysis, 115 Fowler, P. H., 92 Fragmentation, 40, 92

Freier, P., 91

Frozen-in magnetic

field,

111, 152, 160, 169

Dipole field, 70 Dipole moment, 71 Dirac theory, 27 Direction of approach, 66, 80

Direction of viewing, 65 Directional response, 66

Galactic cosmic rays, 13, 90-100

(see also

primary

cosmic rays)
Galactic disk, 168, 170 Galactic halo, 168 Galactic magnetic field, 169 Galactic spiral arms, 168 Galactic structure, 158 Galactic origin model, 169, 171, 172 disk, 170 halo, 170 Galaxy, 167-170 radio, 174, 171 Galileo, 107 -y rays, 3, 4, 6-7, 25, 27, 37-38, 85, 97-98 Garden hose angle, 111-112 Gas amplification, 44-45

Disordered magnetic field, 113 Diurnal variation, 104, 114-120 amplitude, 115-119 time of maximum, 115-119 Dolbear, D., 105 Donors, 47
Earl, J. A., 92

East-west asymmetry, 90 Eccentric dipole, 74 e-folding rigidity, 142 Einstein, 17 Elastic scattering, 23-24, 35, 41 Electromagnetic component, 81

Gauss

coefficients,

see

spherical

harmonic co-

efficients

Electromagnetic field, 21, 25 Electromagnetic interactions, 16, 21-28 Electromagnetic radiation, 6, 16, 65 Electrometer, 2, 4, 5
Electroscope, 2

Geiger-Mueller counters 10, 43, 45-46, 49, 51, 58 self-quenching, 45 limited proportionality, 45
Geitel, H., 2

Genetic

effects,

155

Elementary particles, 12, 15, 83 Elliot, H., 105

Elster, J., 2

Emission, 16

Emulsion stacks, see pellicules Energy losses, 28-36 Energy per nucieon, 97 Energy transfer, 21
Eve, A.
S.,

Geomagnetic bending, 68-69, 117 Geomagnetic coordinates, 70 Geomagnetic cutoff, 73, 78, 87 Geomagnetic effects, 67-80 Geomagnetic equator, 71 Geomagnetic field, 9, 65, 70, 74 Geomagnetic threshold, 78 Geometric cross section, 20
Gockel, A., 4 Gold, A. T., 168 Granules, 106

Excitation, 16, 22, 25-26, 29-30, 36 Extensive air shower (EAS), 56, 82, 86, 97 detector, 56-57 Extragalactic particles, 173

Ground

level event, 36,

135

Guiding center, 68
Gyroradius, 67, 98, 113, 131, 164, 166, 173

Faculae, 108 Faraday's law, 111 Fermi, E., 160

Fermi acceleration, 153, 160-163

injection energy, 162

Hard component, 81, 91 Harmonic dial, 115 Harmonic components, 115 Heavy nuclei, 13, 91, 95, 97,
W., 28 Hereford, F. L., 91 Hess, V., 4, 5, 103
Heitler,

98, 159

Ferraro, V. C. A., 109 Filaments, 108

Fireball, 86
Fission,

40 Flocculi, 109

Fluorescent radiation, 48 Flux, particle, 19, 87

Hierarchy origin theory, 172, 174 Hoffman, G, 103 Hohenstrahlung, 6 Hydrogen alpha (Ha), 135, 154 Hydromagnetic instabilities, 153

Forbidden region, 72
Forbush, S. E., 104-105, 133

Impact parameter,

Forbush decrease, 105, 121-124

22, 23, 25 Inelastic scattering, 23, 41

184
Integral energy spectrum, see primary cosmic rays Integral response curve, 87 Intensity variations, 103-131 Interaction length, 20

INDEX
Magnetohydrodynamics, 109, 163

Main cone, 73 Malmfors, K., 104

Manned

stations, 62-63

Intergranular dark spots, 108 Interplanetary magnetic field, 112, 114, 130, 151, 166 Interplanetary medium, 128 {see also solar wind) Interplanetary space, 12 Ion pairs, 2, 22, 30
Ionization, 3, 4, 6-10, 16, 22, 25-26, 29-30, 36, 42, 46, 81, 96 minimum, 30-31, 44

Mass, 21 Massey, H. J., 28 Maxwell, J. C, 21

J. C, 2, 3 Mean free path, 20, 81, Mean lifetime, 18, 162 W. Messerschmitt, 105

McClennan,

129

Meson component, 81 Meson telescope, 55, 82, 133

cubical, 55
inclined, 104

primary, 30, 44
specific, 30-31, 50,

54

total specific, 31

opening angle, 55 Metagalactic origin model, 169-171

local, 170, 171

Ionization chamber, 2, 3, 4, 5, 11, 43-44, 46, 58, 103, 117, 133 electron collection, 44 pulse, 45 Ionosphere, 132, 137-140 D-region, 137 E-region, 137, 140 F-region, 137 Ionospheric forward scatter, 137-139 Isocosms, 10 Isotopes, 54
Jet, 83, 85

homogeneous, 169, 171

Metastable
state,

48

Meteorites, 54, 99-100

Johnson, M. A., 165 Johnson, T. H., 90, 91

Kinetic energy, 16, 23, 27 Knock-on electrons, 23, 25 Kolhorster, W., 5, 10, 103

Meteorological effects, 87-89, 104 Meyer, P., 92 Microwave radiation, 173 Milky Way, 14, 98, 103, 168 Millikan, R. A., 6, 7, 8, 15, 158 Mirror machines, 163 Mobile stations, 63 Modulation, 94, 106, 113-121, 130 M0ller, C, 28 Monk, A. T., 105 Mott, N., 28 Multiple scattering, 35 Multiplication, 56 Multiplication factor, 44 Multiplicity, 40, 87 Muon, 83, 84, 86, 88-89

Latent track, 54 Latitude effect, 6, 65, 66 knee, 10, 67 Leinbach, H., 139 LeMaitre, G., 71 Lenz' law, 111 Lindholm, F., 103 Liouville's theorem, 73 Liquid drop model, 39 Little, G., 139 Local scale height, 63 Local time effect, 121 Lofgren, E. J., 91

Negative energy, 27 Neher, H. V., 105 Neutrinos, 97, 98 Neutron monitor, 60-61, 82, 133, 140, 151 pile, 60 moderator, 61 Neutron star, 1 67

Ney, E.

P., 91

Long-term variation, 105, 117, 125-131 Luminescent centers, 49

Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic Magnetic
analysis, 69

bending, 67-70 bottle or tongue, 123 equator, 70

field strength,

44

Night-time recovery, 139 Nonradiative collisions, 23 Novae, 166 Nuclear active component, 81 Nuclear emulsion, 53 Nuclear forces, 12, 16, 39 Nuclear interactions, 16, 39-41, 81 Nuclear radius, 23 Nucleons, 39-40, 61 Nucleon-nucleus collision, 40 Nucleonic component, 60, 81 Nucleonic intensity, 140 Nucleosynthesis, 166

mirror, 161

moment,

21, 22

potential, 75

Oppenheimer, Oppenheimer,

F., 91

pumping, 153, 164 rigidity, 68-70

spectrometer, 64
star,

28 Optical model, 39 Origin theories, 158, 169-175

J. R.,

166

Pacini, D., 3

INDEX
Pair production, 16, 25, 27, 37-38 Parker, E., 110
Particle density, 96

185
Radiation damage, 155
early, 155

genetic, 155
late, 155 somatic, 155

Particle interactions, 16 Particle trajectory, 66

Pellicules,

53 71, 73 Percent mortality curve, 156

Penumbral region,
Peters, B., 91

Phosphorescence, 49 Photodetachment, 139

Photodisintegration, 1 Photoelectric effect, 16, 25-27, 37 Photographic emulsions, 23, 53-54 Photoionization, 137 Photomultipliers, 48 Photon, see gamma rays Photon absorption, 25, 37-39 Photoproduction, 16

Radiation length, 33 Radiation percent mortality, 155 Radiation processes, 34 Radiative collision, 23, 31, 32 Radioactive dating, 99 Radioactivity, 22 Radiowave absorption, 133, 139 Radiowave enhancement, 139 Radium emanation, 2 Random walk, 150

Range,

31

Rectifying junction counters, 46

Red

giant, 166

Photosphere, 106

Relative abundances, 97, 146 Relativistic effects, 16-18

kinetic energy, 16
16, 68 mass, 16-17, 28 total energy, 16 Resolving time, 55 Riometer, 139-140 Rise time, 140 Rockets, 63, 135
rest

meson, 83, 85, 89 Pickup reaction, 40

*-

momentum,

Pitch angle, 161 Plages, 108 Plasma, 109-110, 164

Plasma instability, 164 Plasma oscillations, 165 Plasma pinch, 153

Plastic detectors,

54

Roentgen

(r),

154

Plastic scintillator, 57, 59

Polar cap absorption (PCA), 134, 140 Positive temperature effect, 89

Positron, 90-91

Power law spectrum, 93, 162 Power spectrum analysis, 125

Precursor, 122

equivalent man (rem), 155 equivalent physical (rep), 154 Rontgen, W. C, 2 Rossi, B., 90 Rutherford, E., 2, 28

Primary cosmic rays, 12, galactic cosmic rays)

13,

15,

92

{see

also

Saturation plateau, 44 Schein, M., 91

Scintillation counters, 47-49

acceleration processes, 13, 160-165 age, 98 composition, 92, 96-98

density, 95, 96, 159 directional distribution, 88, 99 energy spectrum, 7, 86-87, 92-96, 126, 159
isotropy, 98-99, 159

Screening, 25, 38 Seasonal variation, 104

Secondary component,

prehistory, 99-101 sources, 165-168

Primary electrons,

61, 92, 97, 174 Primary-secondary coupling, 86-89 Producer, 61 Prominences, 108 Proper lifetime, 18 Proportional counters, 43-45, 57, 61 Proton flare, 157 Pulsars, 167 Pulse height, 43, 44 discriminator, 45

12, 15, 82 Secular equilibrium, 101 Semidiurnal variation, 115, 120-121 Serber, R., 28 Shell model, 39 Shock waves, 165 Shortwave fade-outs (SWF), 137 Shower, 56-57 (see also extensive air showers) cascade, 56
local, 57, 91

unit, 35, 56

Sidereal variation, 103, 105, 120-121 Simple shadow cone, 72 Skobelzyn, D., 9 Soft component, 81, 91 Solar activity, 106-107
center, 107
cycles,

107

Quasar (QSO), 164, 171

Solar cosmic rays, 132-157, 165

acceleration, 153-154 composition, 144-146 direct radiation, 149 energy spectrum, 141-144 flux, 144, 151 indirect radiation, 149 propagation, 147-152

Quantum

electrodynamics, 12, 21, 23

Radiation, 6, 16 Radiation age, see bombardment age Radiation biological effects, 154 Radiation biological effectiveness (RBE), 155

186
proton events, 134-141 space radiation dose, 154-157 Solar cycle variation, see long-term modulation Solar energetic particles, see solar cosmic rays Solar flare, 109, 135, 137 flash phase, 136 increase, 105, 133 Solar magnetic cavity, 125 Solar particles, see solar cosmic rays Solar physics, 106-113 Solar proton events, see solar cosmic rays Solar structure, 101 Solar wind, 109-113, 126, 128-130, 166 Solar- terrestrial physics, 106 Solid state detectors, 46-47, 1 60 Space gradient, 130 Space radiation dose, see solar cosmic rays Spacecraft, 2, 58, 64, 93, 132, 135, 151 Spallation, 40, 99 Spark chambers, 52-53
sonic, 53 Specific energy loss, 29 Specific ionization, 50 Specific yield function,

INDEX
Sudden phase anomalies (SPA), 137
Sun, 165
active, 106-107 quiet, 106 Sunspots, 107-108 Supergiants, 166 Supernova, 159, 165-166, 172 Superposed epochs, 125, 127 Surface barrier, 47 Surface density, 19 Swann, W. F. G., 73, 160 Synchrotron radiation, 145, 167 Systems, 55-61

Terella, 78, 79 Terrestrial cosmic rays, 175 Thermonuclear explosion, 167

Thomson,
Threshold

J. J., 9, 28 rigidity, see

geomagnetic cutoff

Time

variation, 87, 103 Total absorption cross section, 38 Trajectory calculations, 67, 74-80 Transformations in atmosphere, 80

see

multiplicity

27-day recurrence tendency, 105, 124

Ultrastrahlung, see HShenstrahlung Unidirectional intensity, 93, 113 Universal time effect, 121 Upper limiting rigidity, 119-120
Vallarta, M., 71 Van Allen, J. A., 64 Van Allen belt, 153, 161 Variational spectrum, 114, 122 Vertical cutoff, 73 Vertical profile, 63 VLF absorption, 139

Spectroheliograms, 108 Spherical harmonic analysis, 75 Spherical harmonic coefficients, 75-77 Spicules," 107 Spin, 21 Spinthariscope, 48 Star, cosmic ray, 83, 135, 165 Keppler's, 107 Population I, 168 Population II, 168 Tycho's, 167 Stellar evolution, 159, 166 StSrmer, C, 9, 70, 71 Stormer cone, 71 ff Stormer integral, 74 Stormer theory, 64, 70-73, 93 Stormer threshold rigidity, 74

VLF
Von

enhancement,
Salis, G.,

39

Vogt, R., 92
103

Stormer

unit, 71
see

Weak
garden hose angle

interactions, 39
9, 23, 50-51

Straggling, 32 Stream angle,

Street, J.

Wilson cloud chamber,

C,

90

Stripped emulsions, see pellicules Stripping reaction, 40

counter-controlled, 51 random expansion, 51 Wulf, Th., 3

Strong interactions, 39 Sudden cosmic noise absorptions (SCNA), 137 Sudden enhancements of atmospherics (SEA), 137 Sudden ionospheric disturbance SID, 137

X-ray, 98

Yukawa, H., 39
Zwicky,
F.,

159

VAN NOSTRAND REINHOLD MOMENTUM BOOKS

MOMENTUM BOOKS
BOOKS
were conceived with a purpose
... to

serve the modern inquiring mind. Scientist, engineer, teacher,

each will find that MOMENTUM provoke new ideas, new questions and new answers while they provide an insight into experimental techniques and
student, inquisitive layman
.
.

the disciplines of the scientific mind.

Each

MOMENTUM BOOK

is

a lucid and accurate analysis of

this series
.
, ,

an area of modern or cfassical physics. Combining sound sci-

ence with clear presentation,

in
its

presents expositions
its

important topics
applications.

in

physics

ranging from

history to

WALTER
Bryn

C.

MICHELS

Mawr College. He

Marion Reilley Professor of Physics at received an E.E. degree from Rensselaer

is
r

in 1927 and a Ph.D. degree from the CaliTechnology in 1930. He then spent two years as a National Research Fellow at Palmer Physical Laboratory,

Polytechnic Institute

fornia Institute of

Princeton University, before joining the staff at Bryn

lege, Professor Michels
is

Mawr Col-

the author of Electrical

and Their Applications

(1957),

and Elements

of

Measurements Modern Physics

(1961),

(with A. L. Patterson, 1951).

He

also served as editor-in-chief of

the International Dictionary of Physics

and Electronics

He

has been editor of the American Journal of Physics since

1959. Professor Michels has contributed to that journal, and to

Physical Review, Journal of the Optical Society of America, the

American Journal of Psychology, and Scientific American as well. He is a member of the American Association of Physics Teachers (President 1956-57), the American Physical Society, and the Optical Society of America.

'

VAN NOSTRAUD REINHOLD MOMENTUM BOOK

'

under the General Editorship of WALTER C '

' '

FOR THE COMMISSION ON COLLEGE PHYSICS MICHtLS

allege.

ABOUT THIS BOOK:

versity of interests

This; book conveys to readers with a diand backgrounds a sense of the excitement that pervades cosmic ray research, as well as a feeling for the broad range of fundamental, physics that falls within its purview.

Throughout the book, the emphasis is on ideas, rather than on experimental. and theoretical details. The book begins with the story of the discovery of cosmic rays and the subsequent development of our understanding about them. The first chapter concludes with a survey of our current knowledge concerning their
properties.

The various
in

interactions of cosmic rays with matter

are then discussed, after whjch the

many

different experimental

cosmic ray research are described. Thenext chapter treats the propagation of cosmic rays through the earth's magnetic field, and through the atmosphere. The properties of galactic cosmic rays- \a re then discussed in detail. TrTe"
next chapter
their causes,
is devoted to cosmic ray intensity variations and and the final twd|chapters are concerned with solar and the origin of jthe so-called galactic cosmic rays.'

techniques utilized

cosmic

rays,

mm The
\

Author:

MARTIN

A.

POMERANT,

mm

VIA N

NOS-TRAND RE I'NHOLD COMPANY