Heliosheric Physics and Cosmic Rays

761656S . Heliospheric Physics and Cosmic Rays .
Lecture notes Fall term 2003

Prepared by Kalevi Mursula and Ilya Usoskin University of Oulu Lectured in 2003 by K. Mursula
ii
Preface
This is the second time that a course under the title Heliospheric Physics and Cosmic Rays is lectured at the University of Oulu. The course is still in an evolutionary phase and, inevitably, limited to basics of the elds. The topics selected out of the wide range of cosmic ray physics reect the biased view of the authors. Moreover, we would like to note that much of the material presented here is still a subject of discussion and intensive research. Therefore, the point of view presented here is not always the only one possible. While we try to emphasize a physically consistent and generally accepted view, other interpretations may also be taken on some topics.
Chapter 1
Introduction to Cosmic Rays

1.1 Milestones of cosmic ray research
The study of cosmic rays has a long story. The rst experimental discovery related to cosmic rays was made more than 100 years ago. By measuring the accummulated static charge, C.T.R. Wilson discovered in 1900 the continuous atmospheric ionisation. It was then (erroneously) believed to be only due to the natural radioactivity of the Earth. In order to check that, Victor Hess (Nobel Prize 1936) from the University of Vienna launched in 1912 an electrometer (a charge collector) aboard a balloon to the altitude of 5 km (Fig. 1.1). He discovered that the ionization rate rst decreased up to about 700 m as expected, but then increased with altitude showing thus an outer space origin for ionisation. During subsequent experiments, Hess showed that the ionising radiation was not of solar origin since it was similar for day and night time. The term cosmic radiation became common. It was then believed that the radiation consists of -rays . However, this assumption was soon questioned, and in 1925 Robert Millikan from Caltech, USA, introduced the term cosmic rays. 1
1.1. MILESTONES OF COSMIC RAY RESEARCH
The later developments showed that cosmic rays (CR) consist of charged particles:
1928: J. Clay discovered that the ionisation rate increased with latitude, suggesting that the sources of ionisation were charged particles deected by the geomagnetic eld. 1929: Using a newly invented cloud chamber, D. Skobelzyn observed the rst ghostly tracks left by cosmic rays. 1929: Bothe and Kolhorster veried that the cloud chamber tracks are curved. This showed that CR are charged particles. 1937: Seth Neddermeyer and Carl Anderson discovered muons (rst erroneously called mesons) in cosmic rays. Particle physics developed and used cosmic rays as the main experimental method until the advent of particle accelerators in the 1950s. 1938: T.H. Johnson et al. discovered that the ionisation rate increased from east to west viewing angle, indicating that the ionisation was due to positively charged particles (correctly assumed to be protons). (Charged particles drift in the Earths inhomogeneous magnetic eld due to the so called gradient drift). 1938: Pierre Auger, who had positioned particle detectors high in the Alps, noticed that two detectors located many meters apart both detected the arrival of particles exactly at the same time. Auger had discovered extensive air showers, showers of secondary nuclei produced by the collision of a primary high-energy particle with air molecules (Fig. 1.2). In this way, changing the distance between the detectors, he could observe CR with energies up to 1015 eV - ten million times higher than reached by then in laboratory experiments, and still several orders of magnitude above the highest laboratory energies reached today.
CHAPTER 1. INTRODUCTION TO COSMIC RAYS
Figure 1.2: Artists view of an atmospheric shower.
1.1. MILESTONES OF COSMIC RAY RESEARCH
28 Feb. 1942: First detection of solar cosmic rays as a high increase in ionisation chambers connected with a are and radio disturbances. 1948: Phyllis Frier et al. discovered He nuclei and heavier elements in CR. May 11, 1950: U.S. Naval Research Lab red the rst research rocket to collect cosmic ray, and air pressure and temperature data. 1959: Konstantin Gringauz ew ion traps on the Soviet Luna 2 and 3 missions. The NASA Explorer VII satellite was launched into a low-Earth orbit with a particle detector. 1977: The Voyager 1 and 2 spacecraft were launched to an interstellar orbit carrying, e.g., cosmic ray detectors. 1977-1982: Bogomolov et al. made a series of balloon experiments and found antiprotons in CR. 1990: As the rst spacecraft, Ulysses probe was launched into a high heliospheric latitude orbit to study the 3D picture of solar wind and cosmic rays. So far, it has twice passed close to both the two solar poles.
Cosmic rays have provided and still provide a unique opportunity to study nuclear and particle physics in the energy range unreachable in present or near-future laboratories. It is hard to overestimate the contribution of cosmic ray studies (including neutrino and -ray observations) to nuclear and particle physics. The astrophysical aspects of cosmic rays are another important connection. CR studies have nourished several theoretical investigations, such as the theory of novas and supernovas, magnetohydrodynamics (MHD) and other plasma theories in astrophysics. In these areas, the following milestones can be noted:
1934: W. Baade and F. Zwicky suggested that supernova explosions (Fig. 1.3) are the sources of cosmic rays. 1949: Enrico Fermi suggested that the cosmic rays are accelerated in their interactions with magnetic eld irregularities (so called 2nd order Fermi acceleration). 1977: Ian Axford et al. suggested that the cosmic rays are accelerated by rst-order Fermi acceleration in supernova shocks in a hot interstellar medium. In addition, works of other famous scientists, e.g., such as Alfvn, Ginzburg, e Parker, and Zeldovich should be acknowledged in the development of CR physics.
1.2
What is a Cosmic Ray?
Cosmic ray is not a ray, but a particle. (A small fraction of primary cosmic rays consist of energetic -quanta and neutrinos but there are left beyond the scope of this course.) Most cosmic rays are ionised atoms, ranging from the proton up to the iron nucleus and even beyond to heavier nuclei. Cosmic rays originate from space, being produced by a number of dierent sources, such as the Sun, other stars, and more exotic objects, such as supernova and their remnants, neutron stars and black holes, as well as active galactic nuclei and radio galaxies. Most cosmic ray particles are travelling very close to the speed of light. The most energetic CR particle ever observed had an energy of about 150 Joules, equivalent to the kinetic energy of a fast baseball. The number density of primary CR integrated over energy (> 100 MeV/nucleon) is about NCR 1010 cm3 in the vicinity of the Earth. The total energy density of primary CR particles is WCR 1 eV cm3 . The Earths atmosphere and the geomagnetic eld protect us from be-
1.2. WHAT IS A COSMIC RAY?
Figure 1.3: Cygnus loop supernova remnant. Image by ROSAT.
ing excessively exposed to these particles. As a cosmic ray enters the atmosphere, it will collide with an atmospheric particle (usually a nitrogen or oxygen molecule), generating a series of secondary particles. It is common to separate three kinds of cosmic rays: Galactic cosmic rays (GCR) originate far outside of our solar system. They are the most energetic CR particles with the energy extending and 1% of heavier elements. The source of the very energetic GCR up to 1021 eV. Composition is mostly protons with 7 10% of He
is not exactly known. The GCR ux in the solar system is modulated particles.
by solar activity: enhanced solar activity shields the Earth from these
Solar cosmic rays (SCR), also called solar energetic particles, originate mostly from solar ares. Coronal mass ejections and shocks in the interplanetary medium can also produce energetic particles. SCR particles have energies typically up to several hundred MeV/nucleon, sometimes up to a few GeV/nucleon. SCR composition is roughly similar to GCR: mostly protons, 10% of He, < 1% of heavier elements. During strong solar ares that are optimally located on the Sun, the ux of CR at the Earth can increase by a few hundred percent for hours/days because of the increase of SCR. This is called a Solar Particle Event. Anomalous cosmic rays (ACR) originate from the interstellar space beyond the heliopause. We will discuss later the mechanism of ACR production. The composition of ACR is quite dierent from GCR and SCR, including, e.g., more helium than protons, and much more oxygen than carbon.
Chapter 2
Galactic Cosmic Rays

2.1 Composition
About 90% of the cosmic ray nuclei are hydrogen nuclei (protons), next common are helium nuclei (-particles), and all other elements make up only about 1%. Within this one percent there are also very rare elements and isotopes. These species require large detectors in order to collect enough particles to observe their ngerprint. For instance, the HEAO 3 (High Energy Astrophysical Observatory) Heavy Nuclei Experiment, launched in 1979, collected only about 100 cosmic rays with charges between 75 and 87 during almost 1.5 years of measurements. It was one of the biggest space borne astroparticle instruments. Good measurements require a large instrument but, unfortunately, the cost of space instruments increases greatly with the size (mass) of the instrument. Ground-based experiments may have a much larger eective area and a greatly higher sensitivity but they cannot measure the chemical composition of CR because of atmospheric shielding. All galactic cosmic ray particles are fully ionised, i.e., consist of nuclei only. The violent processes accelerating charged particles strip o the elec9
10
CHAPTER 2. GALACTIC COSMIC RAYS
Table 2.1: Relative and absolute CR abundance (E > 2.5GeV/nuc, [9]) particle nucleus integral parti- number of particles per 105 protons group charge cle intensity m2 s1 sr1 in CR in the Universe protons 1 1300 10000 10000 helium 2 94 720 1600 L 3-5 2 15 104 M 6-9 6.7 52 14 H 10-19 2 15 6 VH 20-30 0.5 4 0.06 103 7 105 SH > 30 104 electrons 1 13 100 10000 antiprotons 1 > 0.1 5 ???
trons from atoms, leaving isolated nuclei and electrons.
The abundance of primary CR is essentially dierent from the standard abundance of nuclei in the Universe (Table. 2.1). The dierence is biggest for light nuclei (L = Li, Be, B) which are mainly produced by CR collisions with interstellar matter in the Galaxy. The relative abundance of dierent elements in cosmic rays is shown also in Fig. 2.1. Among normal matter nuclei, there are also some antimatter nuclei. Numerous balloon experiments devoted to search for antimatter in space took place since 1970s. They have collected, in total, several hundred antiprotons. A big astroparticle experiment AMS (Alpha Magnetic Spectrometer) was lunched onboard the Space Shuttle Discovery and ew during 10 days in June 1998 (Fig. 2.2). It collected about 200 antiprotons with energy above 1 GeV. According to the standard theory, the antiprotons do not originate in the birth of the universe but were produced inside the Galaxy in nuclear collisions of the CR particles with the interstellar matter. However, some ideas of possible extra-galactic origin of antiprotons have also been presented. Unfortunately, data collected so far do not allow to distinguish
2.1. COMPOSITION
11
Figure 2.1: Relative abundance of elements in cosmic rays and in the solar system.
12
reliably between the alternative hypotheses. So far, not a single antihelium nucleon (not to speak about heavier antinuclei) has been detected in CR although the sensitivity of AMS was just high enough to catch the theoretically expected 1 antihelium nucleon during the 10-day ight. This gives an upper limit for the He/He ratio of 1.1106 . In a few years a new bigger detector, AMS-2, is to be installed onboard the International Space Station for years. Hopefully, it will resolve the above mentioned questions related to antinuclei.
2.2
Energy spectrum
The dierential energy spectrum of GCR is based on measurements from dierent instruments covering the energy range from 109 to 1020 eV. (The highest energy of a CR particle detected so far was 1021 eV.) Actually,
the spectrum of primary CR particles below some 10 GeV/nucleon can not be directly measured because of solar modulation. The dierential energy spectrum (Fig. 2.3) shows the CR ux (number of cosmic ray particles passing through a unit area surface in a unit time from a unit space angle per energy unit) at dierent energies. The unit is particles per cm2 s sr GeV . The graph is double logarithmic; a straight line indicates that the number of cosmic rays with some energy is proportional to the energy to some power. Note that the energy spectrum in Fig. 2.3 is not exponential. Accordingly, the GCR spectrum is harder than the thermal energy spectrum (Gaussian distribution). This means that particles have experienced considerable (nonthermal) acceleration. (The dierent acceleration mechanisms will be discussed later). As a rst approximation, the ux of energetic CR can be considered to be isotropic near the Earth.
2.2. ENERGY SPECTRUM
13
Figure 2.2: Inight view of the Discovery shuttle and the AMS detector.
14
Figure 2.3: Dierential energy spectrum of galactic cosmic rays.
15
It is therefore reasonable and common to approximate the dierential energy spectrum of GCR with the power law:
I(E) E cm2 s1 sr1 GeV 1 ,
(2.1)
where is called the spectral index, which is the main characteristics of the spectrum. The ultra high energy particles are very rare and can be detected as extensive air showers on the ground. In this method, the atmosphere is a major part (moderator) of the detector. (This will be discussed later in more detail.) Figure 2.4 shows the dierential energy spectra for dierent GCR species. One can see that the shapes of the spectra are fairly similar to each other, which indicates that the particles were generated/accelerated in similar processes. The spectrum of GCR as measured in the vicinity of the Earth can be divided into the following parts (see Figure 2.3):
Particles with energy below about 20-50 GeV are subject to solar modulation. Here the spectrum deviates from the power law. Within the range 1010 1015 eV, the spectrum is a power law with the spectral index 2.7. Around 1015 eV, the spectrum changes and becomes steeper, with
3.1 This is known as the knee of the spectrum. Actually, the
change is very small, almost imperceptible when viewed in normal units. However, the number of cosmic rays observed at these energies is large enough to make the measurements reliable with great accuracy. The knee is more visible in the enlarged and scaled view of Fig. 2.5.
16
Figure 2.4: Dierential spectra of some spices of GCR near the Earth.
17
Figure 2.5: The dierential GCR spectrum multiplied by E 2.5 .
18
CHAPTER 2. GALACTIC COSMIC RAYS The knee is believed to arise because the acceleration mechanism in supernova shocks becomes less ecient at this energy, probably because of particles gyroradius exceeds the thickness of the shock. Within the range of 1015 1020 eV, the spectrum is a power law with 3.1. The spectrum becomes atter (harder) again at around 1020 eV but the data are rather poor in this range to estimate this change reliably. particles above this energy range (the cause of additional acceleration) is not clear so far. This change is known as the ankle of the spectrum. The source of the
2.3
Origin of Cosmic Rays
Any theory of GCR acceleration must account for the above described energy spectrum.
2.3.1
Ultra High Energy Cosmic Rays
The standard view is that some cosmic rays are mainly accelerated in our Galaxy (Milky Way) and some are accelerated outside it. The origin of the very (ultra) high energy cosmic rays above the knee is still a mystery. Because of low statistics at such high energies it is hard, e.g., to tell exactly where they come from. For example, cosmic ray particles with energies greater than 1019 eV hit the Earth at a rate of one particle per square kilometre per century (Fig. 2.3). So far, only a few particles with energy exceeding 1020 eV have been detected. For a number of reasons, it is suspected that the cosmic rays above the ankle are of extragalactic origin, perhaps generated in the cores of Active Galactic Nuclei, in powerful radio galaxies, or by the speculated cosmic strings. These sources can oer tremendous amounts of energy needed to
2.3. ORIGIN OF COSMIC RAYS
19
accelerate particles to such high energies. However, a direct correlation has not yet been established. As more sensitive detectors gather more evidence, scientists will have a better picture of where these extraordinarily high energy particles are generated.
2.3.2
High Energy Cosmic Rays
The particles below the ankle are generally thought to be mainly produced in our Galaxy. Furthermore, there are reasons to believe that at least up to about 1014 eV, if not up to the knee or even to the ankle, most CR particles are accelerated in the shocks of supernova remnants (SNR). In this model, particles are scattered across the moving shock fronts of a SNR, gaining energy at each crossing (Fermi shock acceleration; see later). Until recently, evidence supporting this idea was only circumstantial, based on theory rather than on observations. It seemed theoretically reasonable that SNR shocks could accelerate particles to the desired energies. The kinetic energy released in a supernova explosions is more than enough to account for the galactic cosmic rays at least up to 1015 eV. Supernovae are fairly common and occur throughout the Galaxy, so it is reasonable that they could be responsible for these energetic cosmic rays. However, even more direct evidence is found for shock acceleration of particles in SNR.
X-ray evidence for SNR acceleration of GCR Synchrotron radiation is emitted when fast charged particles are moving in the presence of magnetic elds. The magnetic eld will force an energetic particle to travel in a helical path thereby experiencing circular acceleration and emitting radiation. It is known that strong magnetic elds exist near and around SNR. So if there are fast charged particles, they should produce synchrotron radiation which could be observed.
20
CHAPTER 2. GALACTIC COSMIC RAYS The energy of synchrotron radiation depends on the mass and energy of
the charged particles and the strength of the magnetic eld. For the cosmic ray energies observed on Earth and magnetic eld strengths deduced from radio measurements, cosmic ray synchrotron radiation should be in the Xray range. For instance, the X-ray source of the Crab nebula (Fig. 2.6) is believed to be due to synchrotron radiation of electrons accelerated up to MeV) -rays resulting from cosmic ray collisions with interstellar gas show presumably by its magnetic eld. 1014 1015 eV. In addition, observations of high energy (10 MeV - 1000
that most cosmic rays are conned to the disk of the Galaxy (Fig. 2.7),
2.3.3
Lifetime of GCR in the Galaxy
Because of the magnetic eld of the Galaxy, cosmic rays are trapped in it for a long time. During their travel inside the Galaxy (they spend most of the time in the halo), they can gain more energy or lose energy, they can collide with other particles, etc. Thus, their propagation in the Galaxy is diusive. Since they are not moving along straight lines, we cannot trace their origin directly, and have to use indirect methods like synchrotron radiation in order to study it. How long are the cosmic rays trapped in the Galaxy? There are several ways to estimate that. Collisions of cosmic ray nuclei with the interstellar matter (or with each isotopes such as
10 Be,
other) can produce lighter nuclear fragments, including radioactive which has a half-life time of 1.6 million years.
10 Be
The measured amount of
in cosmic rays implies that, on an
average, cosmic rays spend about 10 million years in the Galaxy. CR particles are lost in collisions. Assuming the mean free path of a CR particle before absorption to be several g/cm2 , and keeping in mind
21
Figure 2.6: The ROSAT image of the Crab nebula.
22
Figure 2.7: A schematic side view of the Galaxy.
23
that the average density of the galactic disk is about 1024 g/cm3 , one can estimate the distance traversed by the particle before absorption to be of the order of 1024 cm, which corresponds to the time of about 1014 sec (or several millions of years) for relativistic particles. The corresponding time for the Halo, where the density is 1026 g/cm2 , is of the order of 108 years.
24
Chapter 3
Acceleration of cosmic rays

The question of how, where and when CR are accelerated is most important both for galactic, solar and anomalous cosmic rays. How CR are accelerated to ultra high energies (UHE) is still a subject of intensive study. At present, there is no standard theory to explain the origin of these extremely relativistic particles. However, in the lower part of CR energy spectrum, the theory is more or less established. Let us note that only charged particles can be accelerated. Thus, e.g., the energetic solar neutrons sometimes detected during solar ares are secondaries.
3.1
Fermi acceleration
In order to explain the origin of cosmic rays, Enrico Fermi (1949) suggested an eective mechanism of particle acceleration. Fermi exploited the idea of magnetic clouds moving in the interstellar medium (ISM). These clouds can be rather large, several light years wide, with the density 10-100 times higher than the average ISM density and an enhanced frozen-in magnetic eld. Such clouds are believed to occupy several per cent of ISM. When 25
26
CHAPTER 3. ACCELERATION OF COSMIC RAYS
Figure 3.1: Collisions of a charged particle with magnetic eld. a fast moving particle collides with a random irregularity of the eld, the particle can change its momentum, gaining or losing some energy. Figure 3.1 shows two types of collisions or, rather, elastic scatterings, leading to the reection of a particle. The upper case is called the magnetic mirror. In the lower case the particle is guided along sharply bent eld lines. Since the magnetic eld is very stable and remains unchanged during the time of scattering, the scattering of a particle with the eld irregularity is kinetically equal to the collision of a fast ball with a wall. If the wall (i.e., the magnetic eld irregularity) is moving (with the velocity V ), the particle may gain or lose energy during such a reection. During a frontal (V towards the incoming particle) collision the particle would gain energy, while it would lose it during an overtaking (V away from the incoming particle) collision. (The motion of a particle scattered on random magnetic irregularities inside the cloud can be considered as a
3.1. FERMI ACCELERATION
27
E2 2
E1
Figure 3.2: Sketch of a collision of a charged particle with moving magnetic cloud. random walk.) Note that since the probability of frontal collisions is higher than that of overtaking, the whole particle population gains more energy then loses energy. Let us consider a fast moving particle with (laboratory frame) energy E1 entering a slowly moving magnetic cloud (Fig. 3.2). Assuming the particle to be relativistic, i.e., E pc, one can obtain
0 E1 = E1 (1 cos1 )
(3.1)
where = the cloud.
V c
and =
1 1 2
are the relative speed and Lorentz factor of
Primes denote quantities measured in the cloud rest frame and 1 is the angle between the speed vectors of the particle and the cloud.
28

0 0 When the particle comes out of the cloud, it has energy E2 and angle 2
with respect to the cloud velocity, both in the cloud rest frame. Going back to the laboratory frame, one can obtain
0 0 E2 = E2 (1 + cos2 )
(3.2)
We assume that there are no collisions with the matter, only elastic scattering on the magnetic eld irregularities. Therefore, the total energy of the
0 0 particle should be conserved in the rest frame of the moving cloud, E1 = E2 .
Therefore,
0 E2 = 2 E1 (1 cos1 )(1 + cos2 )
(3.3)
and
0 0 E2 E1 E 1 cos1 + cos2 2 cos1 cos2 ' 1. = E1 E 1 2
(3.4)
0 Since the particle motion inside the cloud is random, all 2 have equal proba0 bility, resulting to < cos2 >= 0. Because of the movement of the cloud, the
probability of a particle to enter the cloud with cos1 is (for relativistic particle and slow cloud) proportional to
cV cos1 , 2c
Averaging Eq. 3.4 over angles, one can obtain that
leading to < cos1 >= 1 . 3
1 + 1 2 4 E 3 1 2. = E 1 2 3 Thus, the net energy gain (averaged per collision) is
(3.5)
dE 2 E,
(3.6)
29
where = V /c is constant. Note, that the energy gain increases with the particles energy. Thus, the energy attained by the particle after n collisions is
E = Ei exp( 2 n)
(3.7)
where Ei is the initial or injection energy of the particle. Let us assume the average time between collisions to be c , hence the number of collisions during time interval t is n = t/c , and the energy is 2t ) = Ei exp(t/tc ) c
E(t) = Ei exp( where tc = c / 2 .
(3.8)
A particle may also be lost in inelastic collisions with the ISM or simply leak out of the system. Let the mean time of that be tl . The probability of a particle to survive until a time greater than t is
P (> t) = exp(t/tl )
(3.9)
Combining Eqs. (3.8) and (3.9) one can obtain an integral spectrum of CR (the total number of particles with energy greater than E):
J (> E) = K E
(3.10)
where = tc /tl . This is the integral form of the dierential energy spectrum with a power law behaviour as discussed earlier. In this case the index of the dierential spectrum is
=+1
(3.11)
30
Figure 3.3: Dependence of energy gain and loss upon protons energy.
31
E2
upstream
downstream
-u 1
V = -u 1 + u 2
E1
Figure 3.4: Sketch of a collision of a charged particle with a moving shock. However, particles also lose their energy by means of ionisation. The comparison of gains vs. losses for protons is shown in Fig. 3.3. One can see that, eectively, the Fermi acceleration mechanism has a threshold energy. For protons, the threshold energy is about 200 MeV, for oxygen about 20 GeV and for iron as high as 300 GeV because the heavier ions have higher ionisation losses. Thus, this mechanism cannot produce the similar shape of dierential spectra for dierent nuclei at these energies (see Section 2.2. The above mechanism is called the 2nd order Fermi acceleration because the mean energy gain per collision is dependent on the mirror velocity squared (Eq. 3.6). Bell (1978) and Blandford and Ostriker (1978) independently showed that Fermi acceleration by supernova remnant shocks is particularly ecient because the motions are not random. A charged particle ahead of the shock front can pass through the shock and then be scattered by magnetic inhomogeneities behind the shock (see Fig. 3.4). Assume a large plane shock front moving with velocity u1 . The shocked gas ows
32
away from the shock with a velocity u2 relative to the shock front, and to the left with velocity V = u1 + u2 . Eq. 3.4 applies also to this situation with with = V /c now interpreted as the velocity of the shocked gas (downstream) relative to the unshocked gas (upstream). Since the shock is planar, the probability of a particle to hit it with cos1 is propor0 < cos2 >= 2/3. Therefore,
|u2 | < |u1 |. Thus, in the laboratory frame the gas behind the shock moves
tional to 2 cos1 (1 cos1 0) leading to < cos1 >= 2/3. Similarly,
1 + 4 + 4 2 E 4 3 9 1 = 2 E 1 3
(3.12)
One can see that this acceleration is more eective ( << 1) than the 2nd order mechanism. The particle gains energy from this bounce and ies back across the shock, where it can be scattered by magnetic inhomogeneities ahead of the shock. This enables the particle to bounce back and forth, gaining energy each time. This repeated bouncing process is now called the 1st order Fermi acceleration because the mean energy gain is dependent on the shock velocity only to the rst power. The 1st order Fermi acceleration (also called Fermi shock acceleration) is also used to explain the SCR and ACR acceleration.
3.2
Magnetic pumping
This mechanism was rst described by Alfvn (1963). e Let us consider a particle with momentum p in a homogenous magnetic eld (to be called here H). If the eld changes slowly, the perpendicular and parallel components of the particles momentum with respect to the magnetic eld line obey the laws: p2 = const and p2 = const k H
(3.13)
3.2. MAGNETIC PUMPING
33
These equations express the conditions of the adiabatic motion and the conservation of the adiabatic invariant. They are valid if the typical scale of magnetic inhomogeneities is greater than the gyroradius of the particle. (Note that in the non-relativistic limit, the ratio of p2 /H becomes a constant moment of a charged particle in a magnetic eld.)
2 times the ratio W /H = 1 mv /H, which is the non-relativistic magnetic 2
As long as Eq. (3.13) is valid, uctuations of the eld do not result in energy gain. However, if the size of magnetic irregularities is small with respect to the particles gyroradius, Eq. (3.13) does not have to be valid. In such a case, the guiding center of the particle will be randomly (if the irregularities are random) moved to another magnetic eld line. Thus, the particles guiding center will perform a random walk (scatter) at the small irregularities. Averaging over a big ensemble of particles moving through a eld with small size irregularities, one can assume an equipartition of the particles momentum over the three degrees of freedom (one parallel, two perpendicular to the magnetic eld): 2 1 p2 = p2 and p2 = p2 o k 3 3 o
(3.14)
In order to obtain this average momentum distribution, particles must spend a long time (with respect to the interval between scatterings) in the eld. Let us consider a low density ux of particles traversing a region with contain small-size irregularities. The scheme of the changing H is shown in Fig. 3.5 for k > 1. The necessary conditions for the magnetic pumping mechanism are: eld H to a neighboring region with eld k H and back. Both regions
g (t2 t1 ) e (t3 t2 ),
(3.15)
34
Figure 3.5: One cycle of magnetic pumping in arbitrary time units. where g is the gyro period of the particle around its guiding centre, and e is the equipartitioning time. The process starts at time t1 with a random distribution of momentum described by Eq. (3.14). Between t1 and t2 , as well as between t3 and t4 , the distribution is described by the conservation laws of Eq. (3.13). Then, between t2 and t3 , as well as between t4 and t5 , the momenta are equipartitioned by the scatterings to the form of Eq. (3.14). Since there is no change of the total momentum between t2 and t3 , and t4 and t5 , we assume p2 = p3 , and p4 = p5 . Thus
p2 = 1 p2 = 2 p2 3 =
1 2 2 2 p + p 3 o 3 o 1 2 2k 2 p + po 3 o 3 1 3 1 3

1 2k 2 2 po + + 3 3 3
1 2k 2 po + 3 3 1 2k 2 po = + 3 3

p2 = 4
1 2k 2 2 po + + 3 3 3k
5 2k 2 p2 + + 9 9 9k o
3.2. MAGNETIC PUMPING
35
or, after one complete cycle: po p5 = 3

r
5 + 2k +
2 = po k
(3.16)
One can see that 1, and a momentum gain will take place always when k 6= 1. After n cycles, the momentum becomes pn = po n
(3.17)
It was shown by Alfvn (1959) that the dierential spectrum of CR e accelerated by this mechanism and diusing thereafter, is proportional to p2+ , where is some unknown factor within the range [1, 1]. This mechanism is very eective and can take place wherever there are such magnetic structures with dierent intensities and small scale inhomogeneities. These structures are known to exist, e.g., in the interstellar medium, in the interplanetary space formed by plasma clouds created by the Sun, etc. However, this mechanism also requires some pre-acceleration of particles to be injected because thermal particles do not satisfy the necessary conditions. The three above considered mechanisms are slow but very eective. They are believed to be the most important sources of acceleration of GCR. Besides, the Fermi shock acceleration is considered to be the main source of anomalous cosmic rays. For solar cosmic rays the situation is dierent. According to observations, SCR must be accelerated very fast, within seconds or minutes. For SCR, the mechanism of magnetic reconnection at the top of a magnetic loop is thought to be the main source (see later), while Fermi acceleration and magnetic pumping are only taken into account as a factor changing the SCR spectrum during their propagation through the interplanetary space.
36
CHAPTER 3. ACCELERATION OF COSMIC RAYS There are also some other mechanisms of particle acceleration but we
will not consider them within the scope of this course.
Chapter 4
Solar Cosmic Rays

Solar Cosmic Rays, also called Solar Energetic particles (SEP), were rst discovered on February 28, 1942. The sudden increase of Geiger counters counting rate was associated with a large solar are. Since then CR detectors have occasionally seen sudden increases in CR intensity, sometimes as large as several hundred per cent, associated with an outburst (mostly ares) on the Sun. The cosmic ray intensity returns to normal within tens of minutes to days, as the acceleration process ends and accelerated ions disperse throughout the interplanetary space. These short increases of cosmic ray count rates associated with SEP are called GLEs (Ground Level Enhancements or Ground Level Events). So far, more than 60 GLEs have been registered since 1942 (see Table 4.1). The time prole of the Oulu Cosmic Ray Station (Neutron monitor) during the strong GLE of 24 Oct 1989 (# 45 in Table4.1) is shown in Fig. 4.1. During this GLE the maximum count rate was nearly twice the normal count rate level. Compared to GCRs, SCRs have relatively low energies, generally below 1 GeV and only rarely around 10 GeV. That is why such events are often missed by cosmic ray detectors near the equator where the lowest energies 37
38
CHAPTER 4. SOLAR COSMIC RAYS
GLE # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
date 28/02/1942 07/03/1942 25/07/1946 19/11/1949 23/02/1956 31/08/1956 17/07/1959 04/05/1960 03/09/1960 12/11/1960 15/11/1960 20/11/1960 18/07/1961 20/07/1961 07/07/1966 28/01/1967 28/01/1967 29/09/1968 18/11/1968 25/02/1969 30/03/1969 24/01/1971 01/09/1971 04/08/1972 07/08/1972
Table 4.1: GLE # 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
The list of GLEs date GLE # 29/04/1973 51 30/04/1976 52 19/09/1977 53 24/09/1977 54 22/11/1977 55 07/05/1978 56 23/09/1978 57 21/08/1979 58 10/04/1981 59 10/05/1981 60 12/10/1981 61 26/11/1982 62 07/12/1982 63 16/02/1984 64 25/07/1989 16/08/1989 29/09/1989 19/10/1989 22/10/1989 24/10/1989 15/11/1989 21/05/1990 24/05/1990 26/05/1990 28/05/1990
date 11/06/1991 15/06/1991 25/06/1992 02/09/1992 06/11/1997 02/05/1998 06/05/1998 24/08/1998 14/07/2000 15/04/2001 18/04/2001 04/11/2001 26/12/2001 24/08/2002
39
Figure 4.1: Count rate (in per cent) of the Oulu NM during the GLE of 24 Oct 1989.
40
Table 4.2: Average integral uxes of SCR in the vicinity of the Earth during solar maximum and minimum years (units for particles cm2 s1 ). energy range above 30 MeV above 100 MeV solar maximum 3 102 20 solar minimum 2 102 2 103
are excluded by the Earths magnetic eld. The best detectors for observing solar particles are therefore those at high-latitude regions (like Oulu Cosmic Ray Station) which are more sensitive to the lowest CR energies. On an average, the average integral ux of solar cosmic ray particles in the vicinity of the Earth is shown in Table 4.2. Note that the ux varies greatly with solar activity. GLEs seldom occur during solar activity minima and have their maximum occurrence most typically some 1-3 years after the sunspot maximum. Note also that during a GLE the ux can be several orders of magnitude larger than the average. The ux during SEP events is high enough to be dangerous for astronauts and also for the crews of high-altitude airplanes over polar regions. Therefore, SEPs are an important factor in the new concept of Space Weather which, e.g., tries to predict the short-term solar activity, including SEP events, and its eects in the near-Earth space. Note also that SCR particles are primary cosmic rays, i.e., their characteristics (energy spectrum, time prole of intensity, direction of arrival, pitch angle distribution, etc.) are not very much disturbed during their propagation through the interplanetary space. (Actually they are somewhat disturbed already at 1 AU but this can be taken into account by models of propagation.) Thus, the in situ conditions in the acceleration site of the solar atmosphere can be diagnosed by studying SCR.
4.1. ENERGY SPECTRUM AND COMPOSITION OF SCR
41
4.1
4.1.1
Energy Spectrum and Composition of SCR

Chemical Composition and Ionisation State
SCR are considered to consist of three main components: - proton-nucleon component; - electron-positron component; - electromagnetic component. The electromagnetic component, although is not a cosmic ray as we dene here, is closely connected with the the other two components. The electron-positron component is also accelerated to relativistic energies but it is almost absent in SCR observed at Earth because it has large energy losses. The energy losses can be separated into nuclear and radiative losses. The nuclear losses (collisions with other nuclei) depend on the amount of matter traversed and can be neglected in a typical acceleration process of a solar are. The radiative energy losses include, e.g., the synchrotron radiation and the bremsstrahlung radiation. They are related to the acceleration of the particle, and therefore are much higher for the electron-positron component than for protons. (E.g., synchrotron radiation is stronger for an electron than for a proton of the same energy).
mp 4 me
' 1013 times
The radiative energy losses dominate the high-energy part of the electronpositron component so that almost all accelerated electrons lose most of their energy in the solar corona or photosphere, producing the X and ray emission of typical solar ares. Moreover, it is more dicult for the electron-positron component to propagate through the interplanetary medium. Therefore we will only consider here the proton-nucleon component which is mainly responsible for GLEs and other terrestrial phenomena. The dierence in the chemical composition between SCR and GCR is mainly due to the dierent amounts of matter passed by the two groups
42
Table 4.3: Relative abundances of SCR (10-47 MeV/nucl) for the event of 03/06/1982
H/He 132 66 2 He/O 102 53 1.9 C/O 0.38 0.45 0.84 N/O 0.13 O/O 1 1 1 Ne/O 0.87 0.13 6.7 Mg/O 0.62 0.18 3.4 Si/O 0.2 0.15 1.3 Fe/O 2.5 0.07 38
03/06/1982 11:30-18:00 baseline enrichment factor
in the lack in SCR of light nuclei like Li, Be, B and other elements and isotopes that are absent in the source and produced by weaker collisions. Another peculiarity is that the SCR composition depends on particles
of particles ( 0.1 g cm2 for SCR and 7 g cm2 for GCR). This results
energy. At energy 100 MeV/nucleon the relative composition of SCR is
similar to that in the solar atmosphere. However, SCRs of lower energy are enriched with heavy nuclei. For instance, Table 4.3 shows the relative abundances and enrichment factors (ratios with respect to the baseline) during the famous solar energetic particle event of 3 June 1982. The composition varies signicantly from one SEP event to another. For
instance, for 150 SCR events in the energy range 1.9-2.8 MeV/nucl detected by instruments onboard the IMP-8 and ISEE-3 satellites in 1978-1983, the relative abundances varied signicantly: H/O - 200-30000; He/O - 30-200; Si/O - 0.07-1.00; Fe/O - 0.03-30000 (see Fig. 4.2). Heavier ions are not fully ionised in the SCR source. This was veried by numerous measurements of the charge state of SCR near the Earth. Table 4.4 shows the measured average charge states of some SCR elements. One can see that they are not fully ionised and the relative ionisation level decreases with the charge number and mass. (Note that the charge state does not change from event to event as much as abundance.)
43
Figure 4.2: Distribution of the ratios H/O, He/O, C/O, Si/O, Fe/O in the energy range of 1.9-2.8 MeV/nucl using data from IMP-8 and ISEE-3 for the period 1978-1983.
44
Table 4.4: Mean charge state of SCR (0.4-2.6 MeV) as measured by ISEE-3 in 1978-1979 element full ionisation measured C 6 5.7 N 7 6.37 O 8 7.0 Ne 10 9.05 Mg 12 10.7 Si 14 11.0 S 16 10.9 Fe 26 14.9
Since the charge state of plasma ions in equilibrium is determined by plasma temperature, one can estimate the plasma temperature in that part of solar atmosphere where SEPs are accelerated. (However, that estimate must be modied because the charge state changes during a fast acceleration, which is not an equilibrium process, and during coronal and interplanetary These are typical temperatures in solar corona. propagation.) The estimated equilibrium temperature is 1 106 7 106 K. Let us also briey mention the so called 3 He-rich events when the
3 He/4 He
ratio in SCR is 2-3 orders of magnitude higher than in the so-
lar atmosphere. These events are associated with impulsive ares.
4.1.2
SCR Energy spectrum
The energy spectrum of SCR decreases with particles energy. This is the only similarity with the GCR energy spectrum. The rst remarkable dierence is the maximum energy. Solar protons can be accelerated up to some 20 GeV only. This is in dramatic dierence with the maximum observed GCR energy of about 1021 eV. The other dramatic dierence is that while the GCR ux is roughly constant and exists permanently, SCRs appear rather rarely and very irregularly in time and the SCR ux levels vary accordingly. Moreover, often there are two components in the SCR spectrum, the so called prompt and delayed components with dierent spectral and temporal characteristics.
45
The SCR energy spectrum usually cannot be expressed by a single power law. Often either a broken power law (a compound of several power law pieces) in energy, or an exponential in rigidity (rigidity P = pc/q, scaled ratio of particles momentum and charge), or a Bessel function in energy is tted to the observed energy spectrum. Usually, the SCR spectrum is softer than that of GCR. The SCR spectral index (see Eq. 2.1) varies between 2 and 5. A typical SCR spectrum is shown in Fig. 4.3 for the SCR event of 15 June 1991. (Here, typical does not mean that spectra of other events look similar, but rather that spectra of most events are as complicated.) For this particular event the spectrum was seen to consist of a soft component with a Bessel-function type spectrum (curves 1a and 1b ) and a hard component with a power law spectrum (line 2). One can see that a single power law (line 3) would be a very rough approximation since the discrepancy would be about 20. The lower energy part of the SCR spectrum below some hundred MeV can only be determined reliably using either direct space-borne observations or observations of secondary emissions (microwave or X-ray/radiation, or neutrons). In order to estimate the higher energy (1-10 GeV) ux of SCR, it is usual to make use of the world-wide network of neutron monitors (NM). Every NM has a certain geomagnetic rigidity cut-o Pc (see later) and hence its count rate can be written as
Z dJ
Pc
N(Pc , t) = where
dJ dP (P, t)
dP
(P, t) SNM (P )dP,
(4.1)
is the dierential rigidity spectrum of primary SCR, and
SNM (P ) is the known specic yield function of the NM. Hence, knowing normalised count rates of dierent NMs with dierent Pc , one can estimate the original spectrum
dJ dP (P, t)
by tting the data with
46
Figure 4.3: SCR spectrum during the are of 15 June 1991. 1a , 1b and 2 are model spectra of in situ protons (the hatched area denotes their dierence) for the interval 08:37-09:02 UT, 3 is the best tting power law.
4.2. SOLAR NEUTRONS
47
the model spectrum. A sample of such a reconstruction is shown in Fig. 4.4 for the GLE of 15 June 1991. This reconstruction corresponds to line 2 because of the atmospheric cut-o (particles with lower rigidity cannot pass through the Earths atmosphere and reach the ground level). This is seen as the saturation of the line in Fig. 4.4 at low Pc . Fig. 4.5 shows the relative importance of GCR and SCR uxes at different CR energies. At high energies (above some GeV/nucleon) GCR are the dominant part of the CR, showing a general anticorrelation with solar activity. At low energies (below some hundred MeV), SCR dominate the overall CR ux. This part varies in concert with solar activity. In the energy range between some hundred MeV/nucleon and some GeV/nucleon either the GCR or SCR component may dominate, and the ux variations in this energy range have a very complicated pattern. in Fig. 4.3. This technique cannot be applied below 0.9 GV of rigidity
4.2
Solar neutrons
Another important component in solar cosmic rays is solar neutrons. Since neutrons are neutral, they cannot be accelerated by electric elds or on magnetic structures and therefore are not primary cosmic rays. Why can we still see solar energetic neutrons in the Earths vicinity? Depending on the magnetic conguration in the are site, some accelerated protons/or alpha particles can be trapped in a magnetic loop (bottle) and interact with the solar matter. Let us consider one leg of a magnetic loop (see Fig. 4.6). Magnetic eld lines are depicted in dash. Within the more dense matter below the visible solar surface (in the photosphere), the density of magnetic eld lines becomes higher producing a magnetic mirror for particles (p1 ) which enter the region from above with large enough pitch angle. Before ying backward (upward) they spend some time in a relatively dense matter. This makes it probable for the
48
Figure 4.4: Neutron monitor increases in per cent at the maximum of a GLE on 15 June 1991.
4.2. SOLAR NEUTRONS
49
Figure 4.5: CR energy spectrum. Solid lines denote the GCR spectrum for maximum and minimum solar activity. Dashed line gives the average spectrum.
50
of -particles) of the matter, shown as an explosion in the Figure, producing the neutron n1 . It is important to note that the direction of the neutron is quite close to that of the interacting proton. Therefore neutrons are almost excluded in the upward direction. After being produced, neutrons move straight. Note that neutrons (n2 ) produced by protons (p2 ) with a small pitch angle cannot be seen in the Earth as they cannot escape the dense matter region. Thus, only neutrons from ares which are located close to the solar limb can be seen in the Earth. This has been veried by direct measurements of neutrons and solar are -rays. (Note also the the mechanism of -ray production in solar ares is similar to the solar neutron production described above.) Thus, solar neutrons carry unique information about the conditions at the are site. After being produced, they move on straight lines preserving their kinetic energy without being disturbed by solar, interplanetary or geomagnetic elds. The rst solar neutrons were detected by the world wide network of neutron monitors during the big are of 3 June 1982. Later, ground-based and space borne instruments have detected solar neutrons in several other events. Unfortunately we can not detect galactic or extragalactic neutrons which could locate the remote sources of GCR acceleration because a free neutron is unstable. A neutron decays with a decay to a proton, an electron and an antineutrino with a mean (efold) life time of about 920 sec. The time of light to travel from the Sun to the Earth is
1 AU c
particle p1 to interact with a nucleus (most likely proton as well, or 10 %
only a half of energetic solar neutrons can reach the Earth. (This estimate is However, only few star neutrons are strongly relativistic.
500 sec. Hence, roughly
not true for relativistic neutrons because of the relativistic time dilatation.
It is also interesting to note that protons originating from the decay of solar neutrons before 1 AU have been detected as a small increase of the
4.2. SOLAR NEUTRONS
51
Figure 4.6: Sketch of secondary neutron production in a solar are. Dashed lines denote magnetic eld lines and solid lines particle trajectories.
52
proton ux shortly before the onset of a major GLE. (Try to gure out why those neutron-decay protons come BEFORE SCR!)
4.3
SCR acceleration: ares and CMEs
During GLEs, SCR particles have been accelerated up to some 10 GeV energy within a very short time. All details of SCR acceleration have not yet been solved. Traditionally, it has been supposed that all solar particle events originate from solar ares. Recently, the concept of a coronal mass ejection (CME) has been introduced as another solar phenomenon causing particle acceleration. It is now known that CMEs, not ares as earlier thought, are mainly responsible for large magnetic storms in the Earths magnetosphere. (The earlier view is now called the solar are myth.) CMEs and ares have very dierent properties, CMEs have a much larger spatial scale, involving huge amounts of coronal mass. Flares are of much smaller spatial scale. Both are related to the active regions in the Sun and often appear closely connected. However, it is probable that ares, not CMEs, are mainly responsible for large SCR events, in particular for those leading to GLEs. Traditionally, solar ares are divided into impulsive and gradual ares. Similarly, SCR events are divided into impulsive and gradual events which have signicantly dierent characteristics, such as duration, composition and energy spectrum (see also Table 4.5). This implies that mechanisms responsible for the two SCR types are also dierent. Usually, impulsive events can be reliably associated with the impulsive phase of a are while gradual events can mostly be associated with a shock driven by a CME in the corona and in the interplanetary space. However, sometimes gradual SCR events show evidence of are origin as was the case, e.g., on Oct 16, 2000 (see Figs. 4.7, 4.8)
4.3. SCR ACCELERATION: FLARES AND CMES
53
Figure 4.7: Time prole of the X-ray ux in dierent wavelength bands during a gradual are.
54
Figure 4.8: Time prole of the integral proton ux in dierent channels during a gradual are.
4.3. SCR ACCELERATION: FLARES AND CMES Table 4.5: Properties of impulsive and gradual SCR events Impulsive e-rich 1 1 10 20 107 K Hours 30o 1000 Gradual p-rich 0.005 0.1 100 14 2 106 K Days 180o CME IP Shock 10
55
Particles: 3 He/4 He F e/O H/He Charge of F e Temperature Duration Longitude cone Coronograph Solar wind Events per year
4.3.1
Solar ares
Solar ares are sudden, huge explosions on the surface of the Sun. They were rst observed in the visible light (so-called white light ares) already in 1860. Flares are very fast processes, with the smallest time scales of only a few minutes. Usually they occur near sunspots, along the dividing line (neutral line) between the areas of oppositely directed magnetic elds where the magnetic eld structures get twisted and sheared, releasing energy after magnetic reconnection. The energy released in solar ares can be distributed in many forms: hard electromagnetic radiation (- and X-rays), energetic particles (protons and electrons), and mass ow. Flares are usually characterized and classied by their brightness in Xray radiation. The biggest ares are called X-class ares. The brightness of M-class ares is some ten times smaller than in the X-class. Next weaker classes are C, A and B-classes. Within these main classes the ares are further divided into subclasses. For instance, a X3 are is stronger than X2.
56
The strongest observed ares were of X12-X13 class. Fig. 4.9 shows how the solar atmosphere looks at the are site in the wavelength band of the H line emission. A schematic sketch of particle acceleration during a solar are is shown in Fig. 4.10. A stable pre-are loop in the solar atmosphere may experience a pressure force by the surrounding plasma (horizontal arrows in the left panel). This may lead to an interaction between the oppositely directed magnetic eld lines and, nally, to a complete reconguration of the magnetic structure. This interaction is called the magnetic reconnection and results in an explosion-like release of energy, seen as the impulsive phase of a are. In the reconnection process, a huge amount of magnetic energy is released very rapidly and transformed to thermal and kinetic energy of particles. Reconnection can accelerate particles to a high energy within a short time as required by the very impulsive SEP events. Energetic particles accelerated in the reconnection region are guided away from this region along the newly reconnected magnetic eld lines. Those particles that are ejected upward (escaping protons; see right panel of Fig. 4.10) may either remain trapped or escape into the interplanetary space if the upper magnetic conguration becomes open, and cause an impulsive SEP event. Note that theories based on magnetic reconnection can satisfy all the observational facts of impulsive events. Note also that magnetic reconnection occurs, in addition to the Suns atmosphere, in many other plasma environments, including the Earths magnetosphere. Those accelerated protons that are ejected downward can be trapped in the magnetic bottle of the smaller loop formed within the original loop, populating it with very energetic particles. They are trapped in this bottle bouncing between the two ends (feet) of the loop. When inside either
57
Figure 4.9: H image of the are of 10 Oct 1971 by the Big Bear Solar Observatory.
58
Figure 4.10: A schematic view of particle acceleration during a solar are.
59
foot, they produce secondary emissions, in particular neutrons and and X-rays as described earlier (see Fig. 4.6). The properties of the observed secondary emissions from solar ares are fairly well described by this model. The trapped protons can also gradually escape from the new bottle through coronal diusion across magnetic eld lines, nally reaching open magnetic lines and escaping into interplanetary space. This process may lead to are-associated gradual events. A more detailed analysis of possible solar are scenarios is beyond the scope of the present course.
4.3.2
Coronal Mass Ejection
Coronal mass ejections (CMEs) are spatially larger and temporarily slower events than ares, in which huge amounts of plasma initially trapped in closed coronal magnetic eld lines are ejected into interplanetary space. During active times, several CMEs may occur daily. CMEs involve typically 1012 to 1013 kg of mass, and kinetic energy on the order of 1024 to 1025 J. The disruption of a large stable, magneticallyclosed structure still poses fundamental questions for the magnetohydrodynamic theory (MHD). However, it is probable that large-scale magnetic reconnection is involved in the formation of a CME. Fig. 4.11 shows a CME of 24 Oct 1989. Note that although ares and CMEs are often connected, this is not always so. There are ares that are not followed by CMEs and CMEs without ares. A CME often leads to a huge hot plasmoid (a closed magnetic structure) moving with a high speed in the interplanetary space, and to an interplanetary shock located at the front edge of the plasmoid. The shock can also accelerate particles by the Fermi shock acceleration in analogy with supernova shocks (see Section 3.1). Of course, energies related to the acceleration by the interplanetary shocks are much lower than those related to supernova
60
Figure 4.11: CME of 24 Oct 1989 as seen by the Solar Maximum Mission.
4.4. INTERPLANETARY PROPAGATION
61
shocks. Most CME-related SCR particles are sub-relativistic and the events have a time scale of a day. Note also the dierence between CMEs and the quiet solar wind. Solar wind consists of particles with a frozen-in magnetic eld, i.e., the propagation is mainly determined by plasma motion. In other words, kinetic energy of solar wind particles dominates over magnetic energy of IMF. The situation for CMEs is opposite since the CME evolution is dened by the magnetic eld or, in other words, magnetic energy dominates the particles kinetic energy. Summarizing, impulsive SEP events are associated with solar ares. Gradual SEP events may be either of are or CME origin.
4.4
Interplanetary propagation
Interplanetary propagation is an important factor aecting the SCR. Even for the biggest solar events, particles can miss the Earth if the relative geometry is not favorable. As will be discussed in Section 7.3 in more detail, the interplanetary magnetic eld has a shape of a spiral (see Fig. 4.12) in the ecliptic plane. (However, strong solar events can disturb this picture signicantly.) Let us assume that there is a solar event containing both a are (impulsive event) and a CME (gradual event). If the event is located in the middle of the solar disk (see Fig. 4.13, top panel), particles accelerated in the are cannot reach the Earth since the Earth and the are site are not connected by IMF lines. However, CME which moves radially reaches the Earth, resulting in a gradual SEP or GLE. On the other hand, if the solar event is located near the western limb of the solar disk (see Fig. 4.13, bottom panel), are accelerated particles can reach the Earth along the IMF lines, resulting in an impulsive SEP/GLE
62
Figure 4.12: Spiral lines of the interplanetary magnetic eld in the ecliptic plane.
63
event while CME misses the Earth. If the event is located on the eastern limb or on the back side of the Sun, it cannot cause a SEP/GLE event. Thus, SEP/GLE events observed in the Earth depend very much on the solar location of the event and the interplanetary propagation of SCR. The interplanetary propagation of SCR causes temporal dierences for uxes of particles dierent energies. Even if ejected simultaneously, particles of dierent energies come to the Earth at dierent times due to dierent speeds. This concerns mainly non-relativistic particles. Fig. 4.14 shows the arrival times of protons of dierent energies (dierent shaded boxes) for the solar are of 15 June 1991. The slope of lines is c/L, where L is the length of the corresponding IMF line ( 1.2 AU). The vertical axis is 1/ = c/v. One can see that although the arrival times of SCR with dierent energies were spread over half an hour, they have been ejected simultaneously, and the time dierences was due to the dierenr speeds.
64
Figure 4.13: Schematic view of a GLE associated with a solar event located in the central meridian (top panel) and the western limb (bottom panel) of the Sun.
65
Figure 4.14: The times of rst arrival of particles vs 1 .
66
Chapter 5
Anomalous Cosmic Rays

Anomalous Cosmic Rays (ACRs) are the third primary component of cosmic rays (along with GCRs and SCRs). ACRs were rst discovered in 1973 as a bump in the spectra of certain elements (He, N, O, Ne) at energies of about 10 MeV/nucleon. By now, ACRs have also been observed in H, Ar, and C. A schematic view of ACR origin is shown in Fig. 5.1. ACRs arise primarily from neutral interstellar atoms which are swept into the solar magnetic eld dominated space (called the heliosphere) by the motion of the Sun singly ionized either by photoionization by solar UV photons or by charge exchange collisions with solar wind protons. Once the particles are charged, IMF picks them up and carries them, together with the outward owing solar wind, up to the solar wind termination shock which is expected to be located at the radial distance of about 70100 AU. The ions are called pickup ions during this part of their trip. The ions repeatedly collide with the termination shock, gaining energy in the process, and being accelerated from solar wind energies of about 1 keV/nucleon to higher energies of tens of MeV/nucleon. This continues 67 through the interstellar medium. At 1-3 AU, these neutral atoms become
68
CHAPTER 5. ANOMALOUS COSMIC RAYS
until they escape from the shock. Some of them then diuse into the inner heliosphere. Recent observations from the SAMPEX satellite indicate that singly ionized ions are accelerated to the maximum kinetic energy of about 250350 MeV. However, collisions in the termination shock region cause some ions to become further stripped o electrons, thereby reaching higher ionic charge states (+2, +3, +4, etc.). The electric elds in the termination shock accelerate these higher charge state ions to even higher energies. In fact, SAMPEX has observed ACR oxygen ions at Earth with energies up to at least 100 MeV/nucleon, albeit with a very steep energy spectrum (see Fig. 5.2). (Note also that because anomalous cosmic rays are less than fully ionized, they are not as eectively deected by the Earths magnetic eld as galactic cosmic rays at the same energies.) ACRs are thought to represent a sample of the very local interstellar medium. The atoms with a high rst-ionization potential (typically light atoms, such as H, He) are ionized, on an average, closer to the Sun than those atoms (typically heavier atoms) which have a low ionization potential. Accordingly, the heliosphere acts as a kind of elemental lter for the interstellar atoms allowing a larger amount of high-ionization potential atoms to pass through the heliosphere untouched, and favouring the low-ionization potential atoms to be ionized and to become ACRs. This ltering process explains the above mentioned elemental distribution of ACRs. Therefore, ACRs are a tool to study the motion of energetic particles within the solar system, to learn about the general properties of the heliosphere, and to study the nature of interstellar matter.
69
Figure 5.1: Schematic view of anomalous cosmic rays.
70
CHAPTER 5. ANOMALOUS COSMIC RAYS
Figure 5.2: ACR energy spectra at the positions of Voyager-1 (57 AU) and Voyager-2 (44 AU) spacecraft in 1994. a) ACR H, b) ACR He, c) ACR C, d) ACR N, e) ACR O, and f) ACR Ne.
Chapter 6
Solar Wind
6.1 General facts
S. Chapman ja V. Ferraro proposed in 1931 that bursts of particles emitted from the Sun would cause brief compression of the Earths magnetic eld called the SSC (Sudden Storm Commencement), often preceding large geomagnetic disturbances called magnetic storms. According to their model (now known to be erroneous), solar wind would only occur temporarily in connection with ares or other specic solar phenomena. In 1951 L. Biermann studied cometary tails (see Fig. 6.1) and showed that the pressure of solar radiation alone can not explain his observations. Biermann suggested that solar wind exists always and essentially aects the formation of cometary tails. His estimate of about 500 km/s for the velocity of the continuously blowing solar wind, based on his observations of cometary tails, proved later to be amazingly accurate. Biermanns proposal is now considered to form the start of the modern view of the solar wind (as well as the cometary research). However, the name solar wind was coined by E. N. Parker only in 1958 when developing the theory of the (continuous) solar wind. The existence of solar wind was nally proven by the Soviet Lunnik-2 71
72
CHAPTER 6. SOLAR WIND
Figure 6.1: Comet Tail. This photograph shows the curved dust tail and straight ion tail on Comet Myros; both tails point away from the Sun. The pressure of the Suns light gives the dust particles an outward push, creating a broad arc. In contrast, the solar wind accelerates the ions to high velocities and pushes them into the relatively straight ion tails. (Coutresy of Lick Observatory)
6.1. GENERAL FACTS
73
and 3 probes in 1960 after reaching out from the Earths magnetosphere. Moreover, the Mariner-2 probe conrmed the continuous ow of solar wind during its 4-month trip to the planet Venus in 1962. Solar wind has the following characteristics (at 1 AU in the ecliptic plane, i.e. at the Earths orbit, unless otherwise mentioned; see Fig. 6.2): The average velocity is ca. 400 km/s, but varies between 200800 km/s.
Solar wind consists mainly of protons and electrons but there are about a much smaller percentage. The average charge of solar wind is zero.
520% of -particles (He++ -ions), and several heavier ion species at
Particle density is ca. 5 106 m3 (or 5 cm3 ), varying between (1 20)106 m3 .
The mean particle ux from the Sun is therefore = nv 2 1012 m2 s1 (6.1)
from which one can calculate the amount of particles that the Sun loses in a second
N = 4r 2 = 5, 6 1035 s1
(6.2)
The average energy of protons and electrons is about 1 keV and 1 eV, respectively. The average temperature of protons is ca. 104 2 105 K, i.e., the
b corresponding thermal energy is about 120 eV. (Note! 1 eV=1, 16104
K). The temperature of electrons is roughly the same as that of protons during disturbed times but about 34 times higher during quiet times.
74
Figure 6.2: Histograms of occurrence frequency for the values of the solar wind velocity, proton number density and proton temperature in interplanetary space (from Hundhausen et al., 1970).
6.2. DELAVAL NOZZLE
75
The temperature of charged particles in a magnetic eld is generally anisotropic so that the temperature Tk in the direction of the magnetic eld (or in opposite direction) is higher than the temperature T against it . This dierence arises from the fact that the motion of charged particles along the magnetic eld is more free than in the perpendicular direction. In the solar wind Tk 2 T . The speed of sound in the solar wind is cs =
s
kT 1, 2 104 m/s mp
(6.3)
where = 5/3 is the adiabatic constant for a monoatomic gas. Thus, the velocity of solar wind is about 40 times the speed of sound. Solar wind is therefore extremely supersonic at 1 AU. At the base of the corona the solar wind speed is still below the sound speed, i.e., it is subsonic. However, it is rapidly accelerated so that at about 26 solar radii it reaches the speed of sound, and beyond that remains supersonic until at the outer bondary of the heliosphere called the termination shock it becomes subsonic again. The outward motion of the solar wind, i.e., of the coronal plasma, fol-
lows from the fact that the pressure of the solar atmosphere is greater than the counter-acting pressures due to the solar gravitation and the interstellar matter.
6.2
DeLaval nozzle
Already at the distance of about 10 RS away from the Sun the solar wind is about 300 km/s, i.e., very close to the average value observed at the Earths orbit. Accoringly, a very eective acceleration mechanism must exist close to the Sun which can speed up the solar wind from a subsonic motion to a
76
Figure 6.3: Mass ow through a nozzle used to explain the acceleration of the ow speed strongly supersonic ow over a rather short distance. In order to understand this mechanism we rst study the idea of the so called deLaval nozzle where a subsonic ow can be transformed to supersonic. This is the same mechanism that is, e.g., behind the principle of a jet engine. Let us now examine the ow of (neutral) gas through a tube whose cross section is decreasing (see Fig. 6.3). If the ow speed of the gas is v and the mass density is at a point with cross section A, the mass ux m = vA (6.4)
announces the amount of mass that passes through the tube per time unit. Since, in a steady ow, this ux is the same at each point, we nd dm d(vA) d dv dA = = vA + A + v =0 dr dr dr dr dr (6.5)
6.2. DELAVAL NOZZLE Multiplying this by dr and dividing by m leads to d dv dA + + = 0. v A
77
(6.6)
The ow is sustained by a pressure dierence between the two ends of the tube. A local pressure dierence enhances the ow (so called Bernoulli law): dv dv dr dv dp = a = = = v dr dt dr dt dr or dp = vdv Dividing this by we obtain dp dp d = = vdv d (6.9) (6.8) (6.7)
Let us assume that we have an ideal gas and an adiabatic process (no exchange of heat with surroundings) for which we have the following equation of state: p = const ( is the adiabatic constant). Dierentiating this we have dp = p1 d. (6.11) (6.10)
Multiplying eq. (6.11) by we obtain an equation for the speed of sound dp p = = c2 . s d inserting this to eq. 6.9) and solving d/ we nd d vdv . = c2 s From this and eq. (6.6) we nally obtain: d dv dA = = A v
cs :
(6.12)
(6.13)
v2 dv 1 . c2 v s
(6.14)
78
Figure 6.4: Mass ow through a nozzle with a minimum cross-section to explain the presence of a critical region in the mass ow in order for the low speed to become supersonic. When the cross section decreases (dA < 0), the velocity increases (dv > 0) until the velocity is below the sound speed, i.e., subsonic (v < cs ). If the velocity reaches the sound speed (v = cs ), we must have dA = 0, i.e., the cross section must not decrease any longer. If we want the velocity still to grow after exceeding the sound speed (v > cs ), the cross section must increase (dA > 0)! Thus, the tube must continued by a additional section (see Fig. 6.4) which gives us the de Laval nozzle. On the other hand, if the gas does not reach the sound speed at the narrowest point of the (continued) tube, the velocity must, according to eq. (6.14), turn to decrease during the enlarging part of the tube. This situation is called the Venturi tube. Whether the gas is accelerated to be supersonic according to the principle of the deLaval nozzle, or whether it remains a
6.3. ACCELERATION OF SOLAR WIND
79
subsonic Venturi tube is determined by the pressure ratio between the ends of the tube. E.g., if the tube ends in a vacuum, it is always a deLaval nozzle. This fact is exploited in the jet engines of space satellites and probes.
6.3
Acceleration of solar wind
E. Parker presented in 1958 a modern theory of solar wind according to which a subsonic gas is accelerated supersonic with a mechanism whose principle is quite analogous to the deLaval nozzle. In fact, there is only an analogy between the two since, of course, there is no tube with a decreasing cross section in the solar atmosphere but, rather, the particle density is continuously decreasing as 1/r 2 when moving away for the Sun. We will see that the strong gravitation eld of the Sun has a great inuence and leads to a situation which really is analogous to the principle of the deLaval nozzle. Let us now assume that solar wind is a steadily owing ideal gas and forget, e.g., viscosity and the eect of magnetic eld. As above, the constancy of mass ux at dierent distances leads to the equations (6.4)(6.6). However, the Bernouli equation (6.7) must be added by a term taking into account the eect of solar gravitation: dv GM dp = v 2 dr dr r (6.15)
where r is now the distance from the center of the Sun, M is the solar mass (6.15) by and multiplying by dr we nd and G = 6, 67 1011 Nm2 kg2 is the gravitational constant. Dividing eq. dp GM dr. = vdv r2
(6.16)
On the other hand, the left-hand side can be written in the form (cf. eqs. (6.9) and (6.12)): dp dp d d 2 = = c . d s (6.17)
80
Dividing eq. (6.17) by c2 we obtain, with the help of eqs. (6.16) and (6.6): s d GM dr dA dv v = = 2 dv . cs c2 r2 A v s Here we can join the dv terms and solve for dA = A
(6.18)
v2 1 c2 s
dv GM dr . + v c2 r 2 s
(6.19)
The solar wind is spreading spherically away from the Sun whence A r 2 and dA dr =2 . A r dv which resembles eq. (6.14):
(6.20)
Substituting this in eq. (6.19) we nally obtain an equation between dr and
GM 2 2 cs r
dr = r
v2 dv 1 2 cs v
(6.21)
Equation (6.21) has several dierent solutions (6.5) which are briey treated below. If solar wind speed grows (dv > 0) away from the Sun (dr > 0) but is subsonic (v < cs ), the two sides of the equation are both negative and the equation has a solution. When r reaches the so called critical distance rc = GM 2c2 s (6.22)
where the left side of the equation vanishes, the right side must also vanish. This can either occur so that the velocity reaches the sound speed (v = cs ; solution #1) at the critical distance, or so that the velocity reaches an extremum dv = 0; solution #2). In the rst case, the velocity may still grow even as supersonic when moving outwards, since both sides of the equation are positive. This solution corresponds to the situation which is valid for the solar wind from our Sun.
81
Figure 6.5: Solution of the solar wind acceleration equation (6.21).
82
CHAPTER 6. SOLAR WIND In the second solution the solar wind attains at the critical distance
a maximum which is smaller than the sound speed, and decreases outside the critical distance. Accordingly, in this case the solar wind remains as a subsonic solar breeze. The above two solutions are physically the most interesting and are realized in nature as stellar winds of various types of stars. However, eq. (6.21) has also other, more exotic mathematical solutions. The third solution is when an initially subsonic solar wind reaches the sound speed before the critical distance rc . Then the right side of eq. (6.21) is zero and the left side vanishes only if dr = 0. Thus, the solar wind attains its maximum distance from which it turns back toward the Sun (dr < 0) while the velocity grows as supersonic. If the solar wind speed on solar surface is supersonic, the left side of eq. (6.21) is negative and the only way to make the right side negative is to decrease velocity. At the critical distance the left side becomes positive. If the wind is then still supersonic, the velocity starts growing again beyond the critical distance (solution #4). If the velocity decreased to sound speed by the critical distance, it must continue decreasing even outside it (solution #5). There is also a solution for a subsonic ow approaching the Sun. Then dr < 0 and the left side of eq. (6.21) is negative. The right side gets negative if the velocity increases. If the velocity attains the sound speed at a distance r > rc , the velocity continues growing as supersonic but the distance must start growing (dr > 0; solution #6). In order for the solution #1 to exist, the critical distance rc must exist outside the solar surface (rc > R ). This yields a condition for the sound speed: c2 < s GM 2R (6.23)
83
and, using eq. (6.3), leads to the so called critical temperature of the Sun: T < Tc ' 7 106 K (6.24)
Accordingly, stars with a very high surface temperature have a subsonic stellar breeze. This can also be seen straight from eq. (6.21) which gives, in the limit cs : dr dv = (6.25) r v Thus, the velocity decreases when going out from the stellar surface. The 2 physical explanation is the following. Since the mass ow vA is constant, v grows only if decreases with r faster than 1/r 2 . Such a large density gradient can not exist in a very hot star where a high pressure p = nkT pushes the matter outward and decreases the density gradient. The importance of solar gravitation for the birth of a solar wind with observed properties can be studied for example by leaving the gravitatinal term away, whence eq. (6.21) attains the form: v 2 = r
!
v2 1 c2 s
dv dr
(6.26)
So, if the solar wind on the solar surface is subsonic, the derivative dv/dr < 0 and the wind always remains subsonic. Accordingly, a supersonic solar wind can not exist without solar gravitation. The physical reason is that he strong solar gravitation yields the required density gradient, i.e., decrases faster than 1/r2 . We can calculate the solar wind speed in the Parkers model. Assuming a typical coronal temperature of 106 K, the sound speed is cs = RT = 8.3 103 106 105 m/s = 100 km/s (for comparison, the sound speed in the air is 300 m/s). The critical radius is rc = GM 10R , 2c2 s
84
Earths vicinity (in the framework of the Parkers model) can be calculated from Eq. (6.26) as v = 3.45cs 310 km/s. Observation at 1 AU give the speed of quiet solar wind as 300-400 km/s.
while the Earths orbit is 1 AU 214R . The speed of solar wind at the
However, the average speed of solar wind is dierent at the ecliptic plane and at polar regions. Due to the recent space missions, in particular Ulysses which traveled outside the ecliptic plane, it is now possible to measure the latitudinal distribution of solar wind speed (Fig. 6.6). One can see that solar wind is slower in the near-ecliptic regions (about 400 km/s), and about twice faster (about 800 km/s) in polar regions. (Sometimes, during strong solar events (CMEs, ares), the solar wind can be as fast as 1000 km/s in the ecliptic plane.) The fast solar wind from the polar regions can sometimes extend to close to the solar equator and overtake the earlier emitted slow stream, resulting in a corotating interaction region to be discussed in next sections. The exact mechanism of coronal heating required for the existence of solar wind is not precisely known. Open coronal magnetic structures called coronal holes (see Fig. 6.7) emit the fast solar wind, while slow solar wind comes from closed magnetic structures. Coronal holes are located mainly at high heliographic latitutes and polar regions around the solar mimimum times. During solar activity maxima, only small and short-lived coronal holes are observed, mainly at low latitudes. Plasma outowing from regions of magnetic eld can spread this eld to wherever they arrive. This happens by eld line preservation, a property derived from the equations of an ideal plasma. By those equations, in an ideal plasma ions and electrons which start out sharing the same magnetic eld line continue to do so later on, as if the line were a (deformable) wire and the particles beads threaded by it. If the energy of the magnetic eld is dominant, its eld lines keep their shapes
85
Figure 6.6: Diagram of the solar wind speed.
86
Figure 6.7: Yohkoh soft X-ray image for 02 Nov. 2000.
87
and particle motion must conform to them. On the other hand, if the energy of the particles is dominant - that is, if the eld is weak and the particles dense - the motion of the particles is only slightly aected, whereas the eld lines are bent and dragged to follow that motion. That is the case with the solar wind, and the magnetic eld is frozen in (see the next Chapter).
88
Chapter 7
Heliosphere and Interplanetary Magnetic Field

The heliosphere is the region controlled by the solar magnetic eld, similarly to the Earths magnetosphere that is the region dominated by the Earths internal magnetic eld. The heliosphere is a big magnetic bubble in the interstellar wind formed by solar wind and the solar magnetic eld transported with it. The size of the heliosphere is believed to be about 100150 AU. The heliosphere studied directly by mankind is expanding because the space missions like Pioneers or Voyagers explore a larger and larger part of the heliosphere. In late 1970s, when the most distant spacecraft was only about 10 AU away, the size of the heliosphere was assumed to be only 2025 AU, later it increased to about 50 AU, and then to about 70 AU. Now, when the most distant spacecraft is close to 90 AU, there is rst evidence for a termination shock. Within a few years we will hopefully verify this evidence and also observe the heliopause. A scheme of the heliosphere is shown in Fig. 7.1. The heliospheric structure reminds loosely that of the Earths magnetosphere. Bow shock is the most outer boundary of the heliospheric inuence. The interstellar wind does not know about the presence of the 89
90CHAPTER 7. HELIOSPHERE AND INTERPLANETARY MAGNETIC FIELD heliosphere beyond the bow shock. Heliosheath is a transition region between the areas dominated by the solar wind (inside the termination shock) and interstellar wind (outside the bow shock).
91
Figure 7.1: Artistic view of the Heliosphere.
92CHAPTER 7. HELIOSPHERE AND INTERPLANETARY MAGNETIC FIELD
7.1
General facts about IMF
The Sun is a magnetic star whose magnetic eld is formed at the bottom of the convection layer. Magnetic ux tubes rise from the source region towards the surface, forming local regions of very strong magnetic eld. During sunpot maximum years such active regions are observed in a large part of solar surface, forming a complex structure for the solar magnetic eld. On the other hand, during sunspot minimum years the weak background magnetic eld with a roughly dipolar form is dominating, and large areas of open magnetic ux, so called coronal holes, are observed at polar regions. The magnetic structure of the Sun (corona) also regulates the properties of the solar wind. A slower (ca. 300400 km/s) but denser solar wind is emitted from regions close to the active areas of the Sun, while the polar coronal holes emit a faster (ca. 700800 km/s) wind. This leads, e.g., to strong latitudinal gradients in solar wind speed during sunspot minimum years (see Fig. 7.2). The solar wind carries the magnetic eld of the solar corona as the so called interplanetary magnetic eld (IMF). As will be discussed later in more detail, the IMF is said to be frozen in the solar wind. Here we list some basic properties of the IMF: The intensity of the IMF at 1 AU is about 5 109 T = 5. It varies between 1 15 but can temporarily attain much greater values, even beyond 100 (see Fig. 7.3). The largest IMF values are observed inside so called shock fronts which can be due, e.g., to coronal mass ejections (CME) or solar ares, or due to the interaction between a low and a fast solar wind region. The latter are called corotating interaction regions (CIR). The direction of IMF also varies greatly. The mean IMF direction at 1 AU is about 45 o the direction of the Sun. The IMF eld lines
7.1. GENERAL FACTS ABOUT IMF
93
Figure 7.2: Average distribution in the ecliptic plane of the solar wind velocity (scale on left) in 1976 (O) and 1977 () as a function of the angular distance in degrees to the neutral sheet () (Bruno et al., 1986). Scale on right: the velocity associated aa index according to Svalgaard (1977). The hatched area shows the slower solar wind (V 450 km s1 ) and the correlated thickness of its coronal source: the slow wind sheet. The comparison between the two series of data (1976 and 1977) shows that the sheet thickness depends upon the latitudinal gradient of the wind velocity.
Figure 7.3: Histogram of occurrence frequency for magnetic eld strength values in interplanetary space (from Ness, 1969; e.g., Flthammar, 1973). a
95
(in a quiet situation) have a spiral structure, forming the so called Archimedean spiral (see Fig. 7.4). We will derive the structure of this spiral later in this Chapter. The IMF eld lines directed toward the Sun and away from it occur intermittently, dividing the IMF to two sectors where either direction is dominating. The sector directed toward the Sun is called the Tsector (Toward) and the one away from it the A-sector (Away). While the Sun is rotating, the two types of IMF sectors are found one one solar rotation form the momentary sector structure of the IMF. The so called heliospheric current sheet (HCS), also called the hegenerally wavy form of the HCS resembles that of the skirt of a ballerina dancer (see Fig. 7.5). The projection of the HCS on the solar surface marks the solar magnetic equator. The eld lines of the T- and A-sectors come, by denition, from the southern and northern magnetic hemispheres of the Sun, respectively. The two magnetic hemispheres may greatly deviate from the location of the two heliographic hemispheres. Moreover, the magnetic hemispheres change from one heliographic hemisphere to another every solar cycle, forming the roughly 22-year magnetic cycle of the Sun. The nature of the sector structure varies within the solar cycle. During
after the other. The pattern of sectors observed (e.g. at 1 AU) during
liospheric neutral sheet, is located between the T- and A-sectors. The
the solar minimum, a 2-sector structure with one T-and one A-sector dominates. Other frequent patterns are 3-sector and 4-sector structures which mainly occur during high solar activity times.
Figure 7.4: Interplanetary magnetic eld forming the Archimedian spiral.
97
Figure 7.5: The ballerina skirt of the heliospheric neutral sheet.
7.2
Magnetic eld frozen in the plasma
The good electric conductivity of the plasma leads to the fact that an initial magnetic eld is carried out together with the plasma, leading to the concept of the so called frozen in magnetic eld. We will study now how this really comes about. In addition to the solar wind and IMF, a similar situation occurs in a number of space (e.g., ionosphere, magnetosphere) and laboratory plasmas, making the concept very useful. Let us rst examine the Ohm law in a coordinate system moving with the plasma (primed variables denote this system): 0 = E 0 j (7.1)
where 0 is the electric current density, E 0 electric eld and electric conj ductivity. In another coordinate system with respect to which the plasma is moving with a velocity v , these variables are transformed as follows when | v |<< c: 0 j j E E 0 v B0 = E0 v B B B0 transformations for elds. We do not derive them here). Using the transformations (7.2)(7.4) eq. (7.1) yields = (E + v B) j which is the generalized Ohms law. Dividing this by we nd j/. E+vB = (7.6) (7.5) (7.2)
(7.3)
(7.4)
(These are the non-relativistic versions of the complete relativistic Lorentz-
7.2. MAGNETIC FIELD FROZEN IN THE PLASMA
99
mate relation:
Since plasmas have a very good conductivity ( ) there is an approxi E = B. v (7.7)
This is the formula of the so called convection electric eld, and one form for the equation of frozen-in elds. Substituting eq. (7.7) into the Faraday law (one of the four Maxwell equations) B = E t we obtain the following condition for the magnetic eld: B = ( B). v t (7.9) (7.8)
In order to study the implications of this equation, we look at the temporal change in the magnetic ux which ows through a surface S moving at the velocity v d d = dt dt
Z Z
S
B dS =
Z Z
S
B dS + t
v B ( d l)
(7.10)
where L is the boundary of the surface. The rst term on the right-hand side describes the ux change due to the temporal change in the magnetic eld, the second term describes the change due to the motion of the surface across B. The second term can, using Stokes law, be modied to the following form:
I I Z Z
v B ( d = l)
( B) d = v l
( B) dS. v
(7.11)
With eq. (7.9) we nally obtain d = dt

Z Z
B ( B) dS = 0 v t
(7.12)
100CHAPTER 7. HELIOSPHERE AND INTERPLANETARY MAGNETIC FIELD which proves that the magnetic ux through a surface moving with plasma remains constant. Therefore the magnetic eld is, in a way, frozen in the motion of the plasma. Figure 7.6 illustrates two simple eects of this rule. In the rst example a plasma is formed inside a background magnetic eld. When the plasma is forced to move, the frozen-in condition leads to the bending of magnetic eld lines. In the second example a plasma moves to a region of magnetic eld whose eld lines are subsequently bent in order to conserve the, initially, zero magnetic ux through the leading surface of the intruding plasma.
7.3
Archimedean spiral
Let us study a solar wind plasma owing at speed v radially outward from a point at the solar equator whose longitude (with respect to some arbitrary direction) is 0 . Then, after time t, the polar coordinates of plasma in the rotating coordinate system are r = v t + r0 = t + 0 obtains 0 (7.15) + r0 . This equation which relates the radial distance of the plasma with the lonr=v gitude of its rotating source, is called the Archimedean spiral (see Fig. 7.7). The situation is analogous to a spinning garden sprinkler whose emitted water forms a similar spiral. Next we will study what kind of structure the IMF moving with solar wind will attain. Remaining still in the equatorial plane for simplicity, the velocity vector and the magnetic eld only have the r- and -components: v = (vr , 0, v ) (7.16) (7.13) (7.14)
where is the angular velocity of solar rotation. Eliminating time t, one
7.3. ARCHIMEDEAN SPIRAL
101
Figure 7.6: An illustration of the frozen-in eld concept. (a) A magnetic eld B is assumed to be penetrating a region of highly conducting plasma. (b) When the plasma starts to move, the magnetic eld lines will be frozenin and follow the motion of the plasma. (c) A highly conducting plasma is approaching an area of magnetic eld. (d) Due to the high conductivity the eld cannot penetrate into the plasma and is pushed ahead of the plasma blob.
Figure 7.7: The solar wind plasma streams radially out from a rotating Sun, and its motion can be described as an Archimedian spiral (garden hose). At the position of the Earth the angle () between the plasma velocity and the Sun-Earth line is close to 45o . The Earths orbit and the eastward direction are indicated.
7.3. ARCHIMEDEAN SPIRAL B = (Br , 0, B ) and their absolute values only depend on the distance r: B = |B| = B(r) v = v(r).
103 (7.17)
(7.18) (7.19)
The magnetic eld must always fulll the following Maxwell equation (sourcelessness of magnetic ux): B =0 The divergence of of a vector eld F in spherical coordinates is 1 F 1 F cot 1 + F + F = 2 (r2 Fr ) + r r r r r sin (7.21) (7.20)
In the present case we have an axially symmetric case ( = 0) and also B = 0, so only the rst term in (7.21) applied for B remains:
1 2 (r Br ) = 0 r2 r or
2 r 2 Br = r0 B0 = constant
(7.22)
(7.23)
This equation only says that a constant amount of magnetic ux ows through the surface of a sphere with any radius r. One can solve for Br : Br = B0
r0 r
(7.24)
In a steady (temporally constant) ow B =0 t and so the frozen-in eld equation (7.9) attains the form: ( B) = 0 v (7.26) (7.25)
104CHAPTER 7. HELIOSPHERE AND INTERPLANETARY MAGNETIC FIELD The vector v B only has a component in the direction of : v B = (0, v Br vr B , 0) (7.27)
and its curl only has a component in the direction of which can be transformed, using eq. (7.26), in the form:
i 1 h r(v Br vr B ) = 0. r r
(7.28)
Integrating this we obtain r(v Br vr B ) = constant. If the magnetic eld is initially radial B0 = 0, Br0 = B0 eq. (7.29) can be cast in a form r0 v0 B0 = rv Br rvr B . The azimuthal velocity at distance r0 due to the Suns rotation is v0 = r0 whence eq. (7.31) becomes
2 r0 B0 = rv Br rvr B
(7.29)
(7.30)
(7.31)
(7.32)
(7.33)
from which B can be solved: B = or, using eq. (7.24), B = v r Br . vr (7.35)

2 v Br r( rr0 )2 B0 rv Br r0 B0 = rvr vr
(7.34)
7.3. ARCHIMEDEAN SPIRAL Far from the Sun r >> v , whence B = r r2 Br = 0 B0 . vr rvr
105
(7.36)
Since vr is roughly constant far from the Sun, we see that the azimuthal component of the IMF decreases with increasing r as 1/r, i.e., much slower than the radial component Br which decreases as 1/r2 (eq. (7.24)). Accordingly, far from the Sun the IMF turns more and more azimuthal. Let us determine the angle , which the IMF makes with the radius vector: tan = Using eqs. (7.24) and (7.36) we nd tan =
2 1 r r0 B0 = . rvr B0 ( rr0 )2 vr
|B | . Br
(7.37)
(7.38)
the average angle of the observed IMF.
orbit tan 1, i.e., 45 . This prediction is in a good agreement with Note that, remembering the radial dependence of the two IMF compo-
At large distances from the Sun tan , i.e., 90 . As the Earths
nents and the average value of IMF intensity and angle at one distance (e.g., at 1 AU), the corresponding IMF values can be easily obtained at any distence from the Sun.
Chapter 8
Solar modulation of galactic cosmic rays

GCR are inuenced by the solar wind and IMF when entering the heliosphere. This inuence which is seen, e.g., in the change of GCR intensity and spectrum is called the solar (or heliospheric) modulation of GCR. The amplitude of modulation depends on the level of solar activity, resulting in the 11-year cyclicity of GCR intensity. The amplitude of modulation is very dierent for dierent GCR energies. For instance, the modulation is only a few percent over the solar cycle for particles with an energy of several tens of GeV/nucleon, while it can be a factor of 100 for 300 MeV particles. The theory of solar modulation is well developed by now although all details are not yet completely understood. All modulation studies are based on the (now standard) transport theory suggested by Parker in 1965. The transport equation is written as follows:
1 f f = (V+ < vD >) f + (K(S) f ) + ( V) , t 3 lnP 107
(8.1)
108CHAPTER 8. SOLAR MODULATION OF GALACTIC COSMIC RAYS where f (r, P, t) is the distribution function of GCRs (the phase space density of GCRs, or the number of particles per unit volume of phase space averaged over particle directions), P is rigidity, r is distance from the Sun and t is time. The dierential ux is then obtained as jE = P 2 f and gives the number of particles in units of m2 sr1 s1 MeV1 . V is the radially directed solar wind velocity. (Usually V is taken to be 400 km/s in the ecliptic plane and is assumed to increase with increasing latitude). K is the so called diusion tensor, which can be divided into a symmetric part K(S) and antisymmetric part K(A) . The symmetric part relates to diusion and the antisymmetric part describes the gradient and curvature drifts. The vector B B
< vD >= K(A)
(8.2)
is the pitch-angle averaged guiding center drift velocity. The rst term on the right-hand side of Eq. 8.1 describes the outward convection of GCR due to the solar wind and the GCR drift due to the curvature and gradient of IMF. The second term describes the diusion, and the third term the adiabatic energy loss. In the coordinate system determined by IMF, the diusion tensor can be written in the form
k K= 0 0
0 T
The symmetric part of the diusion tensor includes the diusion coecients parallel (k ) and perpendicular ( ) to the mean magnetic eld. The antisymmetric term T represents the drift coecient. Usually, the transport equation is solved numerically starting from a model interplanetary spectrum of GCR particles at the outer boundary of
0 T = K(S) + K(A)
8.1. SPHERICALLY SYMMETRIC FORCE-FIELD APPROXIMATION109 the heliosphere. Modulation parameters are then determined by tting the model to observational data. Usually the heliosphere is considered to be a sphere with a radius of about 100 AU. (However, it has been recently shown that the shape of the heliosphere may signicantly deviate from the sphere.) The transport equation (8.1) is dicult to solve, and usually this is done by means of nite dierence techniques. This method is quite developed and sophisticated (3D time-dependent model) but it cannot deal with innite derivatives (-functions in spatial/time domains, e.g., shocks, or in energy spectrum, e.g., monoenergetic uxes). Recently, an alternative technique has been developed. This is the stochastic simulation technique which allows for step- or -function like changes and can trace individual particles inside the heliosphere. The latter technique is more dicult numerically and requires more CPU time. There are some simplied approaches to the theory of CR propagation in the heliosphere that are discussed in this Chapter.
8.1
Spherically symmetric force-eld approximation
Let us consider a very simple but still interesting approximation of GCR modulation which can be solved analytically. Assuming spherical symmetry in GCR distribution, the terms in Eq. (8.1) are greatly simplied, in particular the diusion tensor is symmetric in this case, < vD >= 0 (Eq. 8.2). Then the transport equation (8.1) can be written in the sphericallysymmetric form as f f 1 f r2 = V + 2 t r r r r

1 2 P f (r V ) r 2 r 3 P
(8.3)
Gleeson and Axford showed in late 1960s that under some reasonable assumptions (e.g., constant solar wind speed, roughly power-law dierential energy spectrum of particles, slow spatial change of f , etc.) one can simplify
110CHAPTER 8. SOLAR MODULATION OF GALACTIC COSMIC RAYS Eq. 8.3 even further. First, let us consider a steady-state case with as follows:
f t
= 0.
Next, assuming V and to be constant with radius, Eq. 8.3 can be written 2 f f 2V P f 2f + V + = 0. r 2 r r r 3r P (8.4)
Let us now estimate the relative importance of the rst three terms of Eq. 8.4. First, we assume a simple diusion case, i.e., the inward diusive ux is equal to the outward convection ux: f =Vf r

(8.5)
Therefore, at zero approximation the shape of f (r) is f = fo exp whence f V f r

V2 f,
V r , (8.6)
V2 2f 2f r2
Substituting (8.6) into (8.4), we can estimate the rst three term to be of the order of
V2 2V f, r
f and
respectively.
V
The relative importance of the rst two terms is = 1 v, where 3 compare V and 6 . c r One can see that
vs.
2 r.
For diusion,
1 AU for 1 GeV protons, and we should therefore V = 1.3 103 << c 6 = 0.1 10 r
Therefore, rst and third terms of Eq. 8.4 can be neglected when compared with the second one. Thus, we came to the so-called force-eld approximation: V P f f + =0 r 3 P Then the validity of the assumptions can be veried since now r f << 1. f r (8.7)
8.1. SPHERICALLY SYMMETRIC FORCE-FIELD APPROXIMATION111 The solution of the partial dierential equation (8.7) can be presented in the form of characteristic curves, which are lines of constant f in the (r, P ) plane. Along these lines the following condition must be fullled f dP f df = + =0 dr r dr P (8.8)
From (8.7) and (8.8) one can see that the characteristic curves in this case are given by the expression VP dP = . dr 3 Following the so called quasilinear theory, we can take = 0 P, (8.10) (8.9)
where = v/c is the particle velocity in the units of the light speed. Substituting (8.10) into Eq. 8.9, one can obtain that dP VP V = = dr 30 P 30 and then dP =
p
V dr. 30
From the relativistic conversions for a proton (and expressing energy in MeV and rigidity in MV) E = Therefore,
2 P 2 + To one can obtain that dP = dE.
V dr 30 which can be solved in the following form dE = ER E(r) = TR T (r) = Rr R1
(8.11)
(8.12)
where TR and T (r) is the kinetic energy of a particle at the distance R and r in units of AU, respectively; R is the outer heliospheric boundary, and is the so called modulation strength: = V (R 1) 30 (8.13)
112CHAPTER 8. SOLAR MODULATION OF GALACTIC COSMIC RAYS For r = 1 AU, Eq. 8.12 yields that T (1 AU ) = TR which means that the modulation strength describes the loss of kinetic energy from the outer limit to 1 AU. Moreover, since f is constant on the characteristic curve, we have f1AU (TR ) = fR (TR ). (8.14)
Note that the radial dependence is denoted here as a subscript so that, e.g., f1AU (TR ) = f (1AU, TR ). (8.15)
Using P 2 = T (T + 2To ), one can nd that the corresponding dierential ux, j = P 2 f , is given at 1 AU as: j1AU (TR ) = jR (TR ) (TR )(TR + 2To ) TR (TR + 2To ) (8.16)
where To = 938 MeV is the rest energy of a proton. Thus, for xed , one can easily calculate the spectrum of particles at the Earths orbit once the LIS spectrum is known. This approximation works quite well for high energy particles. However, from the observations of the mean free path of solar cosmic rays, the diusion coecient is known to be rigidity independent for low energies: = 0 P, = 0 Pc , f or P > Pc f or P < Pc (8.17)
f = constant, attains the form (exercise)
where Pc 1 GV. In this case, the equation for the characteristic curve, V dP E = dr 30 Pc
(8.18)
8.2. DIFFUSION-DOMINATED APPROACH where E is the particles total energy. The solution of this equation for r = 1 AU is P1AU + E1AU ER Ec = Pc ln Pc + Ec

113
+ ,
(8.19)
where rigidity is expressed in MV, and energy in MeV, respectively. Thus, the expression for P1AU can be obtained in the form of P1AU =
1 2 To 2
TR Tc Pc
(8.20)
where (TR , Pc , ) = (Pc + Ec ) exp a few hundred MeV.
Note that the force-eld approximation is valid in the energy range above
8.2
Diusion-dominated approach
This approach is based on the assumption that the diusion coecient depends inversely on large-scale uctuations of the magnetic eld strength. This agrees with the observations carried out by Voyager 1 (V1), Voyager 2 (V2) and Pioneer 10 (P10) spacecraft. It was shown that large decreases in CR intensity can be associated with regions where the magnetic eld is enhanced with respect to the standard spiral magnetic eld which, in the ecliptic plane and for V =400 km/s is 1 + r2 BP = 4.75 r2 where r is the radial distance in AU. Regions of enhanced magnetic eld, B > BP , are called interaction regions (IR), while regions with relatively small magnetic eld, B < BP , are called rarefaction regions (RR). Although such a decomposition of the interplanetary magnetic eld is an oversimplication which does not fully take
114CHAPTER 8. SOLAR MODULATION OF GALACTIC COSMIC RAYS into account the actual multi-fractal structure, it is still a very useful approach. At the distance of 1 AU two types of IRs can be distinguished: corotating IRs associated with corotating streams and transient IRs associated with transient ejecta like CMEs. The IRs evolve dramatically with the distance from the Sun. The structure of the solar wind changes qualitatively at the distance between 5 AU and 15 AU. Isolated interaction regions may grow in size (both radial and to form larger regions of enhanced magnetic eld, called merged interaction regions (MIR). So, MIRs result from the dynamical merging of interaction regions formed between solar wind streams of dierent velocities, or shock waves from solar ares, or CMEs. In brief, three main types of MIRs can be identied: (i) Global MIR is a shell-like quasi-spherical MIR extending 360 in heliolongitude around the Sun in the ecliptic plane and at least 30 in solar cycle related) decrease of CR intensity. Global MIRs are produced by systems of transient ows. However, not every system of transient ow produces a global MIR. (ii) Corotating MIR is a spiral MIR produced by the coalescence of corotating IRs. They can produce several successive decreases of CR intensity (often with a period close to the solar rotation period) but they are not strong enough to signicantly inuence upon the long-term modulation. (iii) Local MIRs are not corotating and have a more limited longitudinal and latitudinal extent. They are also formed by the interaction of transient ows. MIRs are considered to propagate with a constant solar wind velocity, leading to propagating barriers against inward diusion. heliolatitude. Global MIRs are mostly responsible for the long term (e.g., latitudinal extent) with increasing distance. Neighboring IRs may coalesce
8.3. DRIFT-DOMINATED APPROACH
115
The eect of MIRs is taken into account as a decrease of the diusion coecient due to stronger scattering of particles in MIRs: B ) , BP
= D(
(8.21)
observations (usually V 2). This approach shows generally a good agreement with CR intensity variations but has some weak points: (i) the results are valid mainly in the ecliptic plane, (ii) the models assume the solar wind velocity to be constant in the whole heliosphere, (iii) disturbances of the magnetic eld are assumed to propagate unchanged, (iv) inside 10 AU the results are less reliable because MIRs are not completely formed at this distance, and (v) inhomogenities of solar wind and magnetic eld cannot be taken into account.
where D and 1 are constants. The value of B is taken from spacecraft
8.3
Drift-dominated approach
Another direction of modulation models is to consider the eect of largescale drifts on the transport of CRs. In the heliosphere the most important are the curvature and gradient drifts. Curvature drift results from the centrifugal force that the particle experiences when traveling along a curved magnetic eld line. The drift velocity is therefore proportional to the factor (centrifugal force normalized by local gyrofrequency): vcurv Rcurv B
2 vk
The gradient drift results from the change in the particles gyroradius during one rotation because of the changing magnetic eld intensity. This is particularly ecient around the heliospheric current sheet.
116CHAPTER 8. SOLAR MODULATION OF GALACTIC COSMIC RAYS
V drift
B-
B+
Figure 8.1: Sketch of the particle drift in the heliospheric neutral sheet. Drifts depend on the particles pitch angle. For an isotropic particle distribution fo , the average drift can be estimated for the mean Parker spiral as cvp Bo 2 vd = 3q Bo

(8.22)
Then this eective drift velocity is treated as convection of particles in the transport equation. The importance of drifts for heliospheric modulation was shown in late 1970s. A major role in this approach is played by the heliospheric current sheet. When IMF is directed towards the Sun in the northern hemisphere (so called negative solar polarity; A < 0), positively charged particles can drift towards the Sun along the sheet (Fig. 8.1). During positive IMF polarity, the drift sweeps positive particles away from the Sun along the sheet. The drift behaviour of negative particles is opposite to that of positive particles. Therefore one often combines the particle charge and IMF polarity together to form the polarity factor qA. The HCS drift eect depends on this factor. The drift model can reproduce the observed GCR modulation only for
8.4. THE MODULATION EFFECT
117
periods when the waviness of HCS is small or moderate, i.e., when the tilt angle 35 , and only during negative polarity conditions (qA < 0). The drift models can also explain the plateau-like CR intensity behaviour during positive polarity minima (qA > 0) when CRs are considered to come mostly from polar to equatorial regions. Of course, drifts do not dominate when the HCS is largely disrupted by a system of transient ejecta.
More recent and sophisticated models of CR modulation combine all the mechanisms described briey above, explaining pretty well the observed CR intensity at dierent heliodistances and during dierent phases of the solar heliosphere) is driven mainly by transients and small scale IRs. At larger distances, drift is responsible for the long-term modulation during periods of low and moderate solar activity. When solar activity is moderate, both drift and MIR/RR eects should be taken into account. Finally, during periods of high solar activity the CR modulation is mostly driven by MIR/RR eects. magnetic cycle. In general, the modulation inside 10 AU (so called inner
8.4
The modulation eect
In the following, some eects of modulation (only convection and diusion, no drifts included) are shown, as calculated using stochastic simulation techniques in a 2D case. The results are shown for two dierent levels of modulation: Weak modulation corresponds approximately to the heliospheric conditions in 1977, i.e., near minimum solar activity, when the heliosphere is quiet. Medium modulation corresponds to the heliospheric conditions in 1992, i.e., in the middle of declining phase of solar cycle 22.
Figure 8.2: Average time spent by GCR protons in the heliosphere before reaching the Earth as function of energy for medium (1992) and weak (1977) modulation. Fig. 8.2 shows the average time needed for GCR particle to reach from the termination shock to the Earths orbit as function of energy. One can see that this time varies from a fortnight (10 GeV for weak modulation) to half a year (300 MeV protons for medium modulation). This results in the well known time lag between solar activity and corresponding long-term changes in GCR intensity. Note that only a small fraction of cosmic rays entering the heliosphere nally do reach the Earths orbit, because of the geometric factor and IMF/solar wind eects. The results discussed in this section only deal with particles which succeed to reach the Earths orbit.
119
Figure 8.3: Energy spread of monoenergetic (-function in energy) uxes for xed initial proton energy, T = 0.3, 0.7, 1, 3, 10 GeV, for medium (1992, panel a) and weak (1977, panel b) modulation.
Figure 8.4: Spectra of galactic cosmic rays. Thick line presents the local interstellar spectrum (LIS), solid line the modulated spectrum at 1 AU for weak modulation, and dashed line the modulated spectrum at 1 AU for medium modulation. Fig. 8.3 presents the energy losses and the energy spread of a monoenergetic particle ux for medium and weak modulation. GCR protons with the same initial energy (T = 0.3, 0.7, 1, 3, or 10 GeV) enter the heliosphere, and their energy spectrum at 1 AU is shown in the Figure. High energy (10 GeV) protons only lose a few percent of their initial energy during propagation in the heliosphere. For lower energy protons (below 1 GeV), the fraction of lost energy becomes signicant even for weak modulation. Thus, energy losses of GCR when propagating across the IMF are very important in the lower energy part of the spectrum.
121
The modulated spectrum of GCR at 1 AU is shown in Fig. 8.4 for weak and medium modulation together with the local interstellar spectrum (LIS). While LIS is continuously decreasing with energy, the GCR spectrum has a maximum at 0.1-1 GeV. Note however, that solar cosmic rays dominate over GCR in the low energy part of the spectrum (see Fig. 4.5). Fig. 8.5 shows the history of sample protons with 1 GeV and 9.2 GeV initial energy during medium modulation. Time evolution of the heliocentric distance of protons, their energy losses as well as 2D trajectories are shown in the Figure. One can see that particles diuse long at middle heliocentric distances until they reach the polar region. After that, they move fairly rapidly to the distance of 1 AU (or escape from the heliosphere).
100 90 80 70 60 50 40 30 20 10 0 0 50
1 0.95 0.9 0.85
100 90 80 70 60 50 40 30 20 10 0 50 0 0 50 100
T, GeV
R, au
0.8 0.75 0.7 0.65 0.6 0.55 0.5
100 90 80 70 60 50 40 30 20 10 0 0 10 20
10 9.8 9.6 9.4
100 90 80 70 60 50 40 30 20 10 0 10 20 0 0 50 100
T, GeV
time, days
R, au
9.2 9 8.8 8.6 8.4 8.2 8
time, days
Figure 8.5: Tracing of sample protons with the initial energy of 1 GeV (upper panels) and 9.2 GeV (lower panels) for medium modulation. Panels from the left to the right present the heliodistance as function of time spent in the heliosphere; energy losses as function of time; trajectory of the particle in the heliosphere.
Chapter 9
Variations of Cosmic Ray Intensity

CR intensity is not constant but changes continuously at dierent time scales. Two groups of CR intensity variations are shown in Tables 9.1-9.2. The rst group of variations in CR intensity detected at the Earth includes the variations of terrestrial origin (Table 9.1). Seasonal and diurnal variations are due to the dierences in atmospheric structure between winter and summer seasons and daytime and night-time, respectively. This eect is signicant for the muon component but small for the neutron component. The asymmetric shape of the Earths magnetosphere results to a small diurnal change of the local geomagnetic cut-o and, correspondingly, to a small diurnal variation of CR intensity on the Earths surface. We are more interested here in the extra-terrestrial variations of CR intensity, i.e., CR variations whose origin is outside the Earths magnetosphere. There are periodic and sporadic extra-terrestrial CR variations. Periodic extra-terrestrial variations include diurnal, solar rotation related and solar cycle related variations. The solar rotation related variations last, on an average, 27 days. They are due to the eect of strong, 123
124
CHAPTER 9. VARIATIONS OF COSMIC RAY INTENSITY
Table 9.1: Cosmic ray intensity variations: Terrestrial eects
type
amplitude Periodic
nature variation Variations of the absorption of secondary particles in the atmosphere due to seasonal changes of the atmospheric structure. Variations similar as above but due to the day-night dierence in the the atmospheric structure. Asymmetry of the magnetosphere leading to a daily variation of the local geomagnetic cuto. variation Decrease of local geomagnetic cuto due to the disturbed magnetosphere.
seasonal
<1%
diurnal diurnal
< 1% small
Sporadic increase during a magnetic storm up to 10%
125
Table 9.2: Cosmic ray intensity variations: Extra-terrestrial eects
type
amplitude Periodic
nature variation Solar modulation of GCR in the heliosphere. Long-lived longitudinal asymmetry in IMF or solar wind structure. Anisotropy of CR uxes due to convection by solar wind and diusion along IMF lines. variations Increase of CR intensity due to arrival of solar cosmic rays. GCR decrease due to the shielding by an interplanetary shock passing the Earth. CR increase due to collection of CR particles in front of the interplanetary shock causing a Forbush decrease. GCR decrease due to the shielding by a magnetic cloud passing the Earth.
11- and 22-year 27-day
up to 30 % < 2%
diurnal
few %
Sporadic GLE Forbush decreases increase before Forbush decrease magnetic eect cloud 1-300% up to 30% < 2%
few %
126
Figure 9.1: Diurnal variations of CR intensity recorded by Oulu NM. long-lived sunspot groups or due to persistent streams of fast solar wind from longitudinally asymmetric coronal holes. Sporadic variations include GLEs, Forbush decreases and decreases due to magnetic clouds.
9.1
Extra-terrestrial diurnal variations
Diurnal variations are often nearly sinusoid-like variations of GCR intensity with an amplitude of 1-2% (see Fig. 9.1). The largest contributions to the diurnal variation come from extra-terrestrial causes. The diurnal variation is mainly due to a local anisotropy of CR uxes. Let us consider the idealised picture of Fig. 9.2. The anti-sunward convection and the spiral-directed diusion aect the GCR ux, causing a minimum in the post-midnight and a maximum in the early afternoon LT sector (see also Fig. 9.1).
9.2. SPORADIC VARIATIONS
127
The phase and amplitude of the diurnal variation (the direction of the GCR anisotropy) may dier from this idealised picture because of drifts, local latitudinal gradients of CR intensity, varying level of diusion, changing solar wind conditions, etc. (The Earths orbital motion also gives a minor contribution).
9.2
Sporadic variations
Forbush decreases (Fig. 9.3) are sudden decreases (up to 30% during several hours) of CR intensity followed by a gradual recovery during several days to weeks. A Forbush decrease is due to an interplanetary shock passing the Earths orbit and producing an eective barrier of intense magnetic eld to cosmic ray particles. Such a shock collects CR particles in front of it because of enhanced scattering of the particles. This is often seen as a small increase of CR intensity immediately before a Forbush decrease. Sporadic CR intensity variations are also caused by magnetic clouds passing the Earth (Fig. 9.4). They result in a steady, small decrease of CR intensity of a few percent during several days.
9.3
Solar cycle variations
CR intensity depicts an 11-year variation in anti-phase with solar activity (Fig. 9.5). There is a time lag between the changes of solar activity and the corresponding changes in CR intensity. This time lag is due to the large size of the heliosphere and the nite propagation time of the solar wind (and the IMF disturbances moving with it), as well as the nite diusion time of GCR particles (see earlier discussion). The overall time lag is several months but the momentary time lag varies in time (see Fig. 9.6) from zero (or even negative) values to about 2 years.
128
Figure 9.2: A scheme of the diurnal anisotropy of galactic cosmic rays.
9.3. SOLAR CYCLE VARIATIONS
129
Figure 9.3: A Forbush decrease, a GLE and the diurnal variation of CR intensity as recorded by Oulu NM in July 2000.
Figure 9.4: A magnetic cloud passing the Earth, as observed by Oulu NM in March 2000.
130
a)
sunspot numbers
200
100
1950
1960
1970
1980
1990
2000
count rate, per cent
b)
100 90 Climax 80 Huanc/Hal Oulu 70
1950
1960
1970
1980
1990
2000
Figure 9.5: Solar modulation of cosmic rays at neutron monitor energies. (a) Monthly sunspot numbers as index of solar activity. (b) Monthly count rates of dierent neutron monitors.
9.3. SOLAR CYCLE VARIATIONS
131
19
30
time lag, months
20
21
22 Climax Huancayo Oulu
20 10 0 -10 -20 -30
1950
1960
1970
1980
1990
Figure 9.6: Momentary time lag in months between solar activity cycle and the corresponding cosmic ray cycle.
132
CHAPTER 9. VARIATIONS OF COSMIC RAY INTENSITY One can also see a 22-year variation of CR intensity, e.g. in the dierent
shape of CR maxima during positive and negative polarity minima (Fig. 9.5), as well as in the signicantly dierent time lags at these times (Fig. 9.6). The 22-year cycle in cosmic rays can be understood in terms of the drift eects around solar minimum times.
Chapter 10
Cosmic Rays and the Earth

10.1 Atmospheric cascade
The matter weighted mean free path of energetic protons (neutrons) in the air is about 100 g cm2 (140 g cm2 ). This mean free path is mainly determined by the nuclear collisions of the proton (neutron) with the nuclei of atmospheric atoms and molecules. On the other, the amount of matter in the Earths atmosphere is 1033 it is very improbable that a primary CR could reach the Earths surface. g cm2 , leading to the 1 atm pressure of the normal atmosphere. Therefore, Instead, they suer a series of successive collisions and interactions, forming the so-called atmospheric cascade (Fig. 10.1). The cascade consists of three main components. One is called the soft or electromagnetic component and it consists of electrons, positrons and photons (electromagnetic quanta). The second component is called the hard or muon component, consisting of muons. Note that sometimes this component is, unfortunately, also called the meson component. This erroneous naming dates back to early history of CR research when the properties of the muon and other elementary particles were not yet well known. We know now that muons 133
134
CHAPTER 10. COSMIC RAYS AND THE EARTH Table 10.1: Active particles in a cosmic ray cascade Interaction electromagn. strong weak mass (MeV) lifetime x x x x x x x x x 134 106 932 938 0.511 26 ns 2 s 12 min stable stable stable
Particle Pion Muon Neutron Proton Electron Photon
Atmospher. absorb.length (g/cm2 ) 115 260 140 110 100
do not interact strongly but belong, together with electrons, tau leptons and neutrinos, to the group of weakly interacting particles that are also called leptons. On the other hand, mesons are strongly interacting particles (hadrons) with integer spin number. The most common meson produced in a CR shower is the pion. Pions are very short-lived and decay before reaching the ground. Charged pions mainly decay to muons, producing most of those muons observed on the ground. Taking this into account the term meson component is partially justied. The third CR component is the nucleonic component which, on the Earth, mostly consists of suprathermal neutrons. The characteristics of the constituent particles are given in Table 10.1. The three components of the cascade have dierent spatial (horizontal) widths. The relative widths are shown in Fig. 10.2. The absolute widths depend on the energy of incoming particle: the more energetic the particle is, the wider cascade it generates. Recently, muon showers of many tens of kilometers wide have been detected, implying that the energy of the CR particle was of the order of 1020 eV. The ux of dierent components at the Earths surface is shown in Fig. 10.3. The neutron component containing of suprathermal neutrons is very signicant. This component is detected by most ground based cosmic
10.1. ATMOSPHERIC CASCADE
135
Figure 10.1: A scheme of the atmospheric cascade.
136
CHAPTER 10. COSMIC RAYS AND THE EARTH
Figure 10.2: Lateral spread of the three components of the atmospheric cascade.
10.1. ATMOSPHERIC CASCADE
137
Figure 10.3: Energy spectra of dierent components of the atmospheric cascade at the sea level in NYC.
138
ray stations. Therefore they are also called neutron monitors. Another important component is the muon component. It dominates the ux at energies above 100 MeV. At lower energies the ux of muons decreases fairly fast because of their short lifetime of only 2 s. (Relativistic muons can travel longer distances than nonrelativistic muons due to faster speed and time dilatation.) Muon ux is also sensitive to the atmospheric structure, in particular to the altitude of the rst collision. The higher, on an average, the rst collision of the primary CR particle occurs, the less of muons are seen on the ground. This leads, e.g., to the above mentioned diurnal and seasonal changes in the muon component due to the changes in the atmosphere. The ux of protons is similar to that of neutrons at energies above 1 GeV but is much smaller at lower energies because thermal protons cause ionisation (while neutorons do not).
10.2
Magnetospheric propagation and geomagnetic cuto
Since CR are charged particles, their propagation close to the Earth is affected by the geomagnetic eld. The direction of a particle entering the atmosphere is often very much dierent from the original direction of the particle outside the Earths magnetosphere in the interplanetary space. A sample of a particle trajectory is shown in Fig. 10.4 The idea of nding the original direction of the CR particle by trajectory tracing results in the concept of the asymptotic direction (also called asymptotic cone). This is the direction that the incoming CR particle must have in the interplanetary space in order to reach a certain location on the Earth.
10.2. MAGNETOSPHERIC PROPAGATION AND GEOMAGNETIC CUTOFF139
Figure 10.4: A sample trajectory of CR particles trajectory in the Earths magnetosphere.
140
CHAPTER 10. COSMIC RAYS AND THE EARTH Often one calculates the asymptotic directions in a dipolar magnetic eld
for those protons that enter vertically the atmosphere above the station at the altitude of 20 km. The asymptotic directions are calculated for protons in the energy range between 110 GeV and plotted as a curved line in the latitude-longitude plot. (Note that the asymptotic direction points in the direction from where the particle is coming. The particles velocity must therefore have an opposite direction.) Examples of asymptotic directions are shown in Fig. 10.5. The asymptotic directions are typically around the equator, somewhat east of the station because of the left-handed curvature of the positively changed CR proton. The 1 GeV end of the asymptotic curve is typically more eastward and more southward than the 10 GeV end because the more energetic particles y more directly to the station. Figure 10.6 depicts the asymptotic directions of some NM stations in the GSE coordinate system at 14 UT on May, 2, 1998. The forward end (plus) and the tail end (cross) of the direction of the momentary IMF vector are depicted in the gure, together with the equi-pitch angle curves with respect to the IMF. Such plots can be used to study the question what must the pitch angle of those particles be that can reach any of the NM stations. The plot reveals, e.g., that only particles whose velocity is roughly perpendicular with respect to the IMF can be seen at Oulu. The geomagnetic eld prevents low-energy particles from reaching the ground level. This leads to the concept of geomagnetic cuto which means a rigidity threshold for CR particles that can reach a certain geographical location. As a rst approximation, the geomagnetic cuto rigidity can be estimated by a simple empirical formula Pcut (, ) = 60
!2
1 coscos3 coscos
(10.1)
where Pcut is expressed in GV, is the geomagnetic latitude of the station,
10.2. MAGNETOSPHERIC PROPAGATION AND GEOMAGNETIC CUTOFF141
Figure 10.5: Asymptotic directions (curved lines) for several NMs stations, such as Apatity (Ap), Iniuvik (In), Thule (Th) and Tixie (Ti).
142
Figure 10.6: Asymptotic directions for NMs Apatity (Ap), Goose Bay (GB), Iniuvik (In), McMurdo (M-M), Oulu (Ou), South Pole (S-P), Thule (Th) and Tixie (Ti) for 14 UT 02.05.1998. Plus and cross denote the IMF forward and tail ends.
10.3. TERRESTRIAL EFFECTS OF COSMIC RAYS
143
and is the angle between the incoming particle velocity and the direction of geomagnetic east. Note that the rigidity is smaller for those particles coming from the east (whose angle is larger than 90 ). This reects the fact that the particles which have a velocity close to the asymptotic direction nd it easier to approach the station. For = 90 , cos() = 0, the expression for the geomagnetic cuto rigidity becomes very simple Pcut = 15(cos)4 and is called the vertical geomagnetic cuto. The distribution of the vertical geomagnetic rigidity cuto is shown in Fig. 10.7. Note that the geomagnetic coordinates do not coincide with the geographical coordinates because of two eects. First, the axis of the magnetic dipole is inclined with respect to the Earths rotation axis by about 11.2o . Second, the center of the dipole is shifter with respect to the Earths center wards India). On the opposite side of the globe, there is a region of reduced geomagnetic eld (enhanced CR ux) which is called the South-Atlantic Anomaly. The total CR intensity per neutron monitor counter as a function of the geomagnetic cuto rigidity is shown in Fig. 10.8. by about 534 km in the direction of 22 North and 144 East (to-
10.3
Terrestrial eects of cosmic rays
A list of some terrestrial eects of cosmic rays is given in Fig. 10.9.
144
Figure 10.7: Isolines of vertical geomagnetic cuto rigidity for the epoch of
10.3. TERRESTRIAL EFFECTS OF COSMIC RAYS
145
Figure 10.8: CR intensity per NM counter versus geomagnetic rigidity cuto in 1992.
146
Figure 10.9: Terrestrial and human eects of cosmic rays.
Chapter 11
Detection of Cosmic Rays

Cosmic rays are measured at very dierent locations: from underground and underwater detectors up to the far edge of the Solar system. CR particles can be observed by the following types of interactions: Inelastic scattering caused by the Coulomb force between the CR particle and the orbital electrons of the detector material.
Elastic scattering of CR particle from nuclei of the detector by the electromagnetic or strong force.
Emission of Cherenkov radiation by the CR particle moving faster than light in matter.
Emission of transition radiation. Transition radiation is produced
when a charged particle passes through media of dierent dielectric properties. A charged particle approaching such a boundary between two media (e.g., from vacuum to a dielectric medium) represents together with its mirror charge an electric dipole, whose eld strength changes in time as the particle moves along and vanishes when the particle enters the dielectric medium. This produces electromagnetic radiation, called the transition radiation. 147
148
CHAPTER 11. DETECTION OF COSMIC RAYS Nuclear reactions (inelastic scattering by the strong force) between the CR particle and the detector nuclei.
Bremsstrahlung caused by the CR particle in the detector material. This is negligibly small for CR protons or heavier CR particles.
11.1
Space-borne detectors
Space-borne detectors can be divided into two groups by their location. One group is located onboard satellites which have a xed orbit near the Earth. Such satellites are, e.g., IMP, GOES, SOHO, and AMS. The other group of space-borne experiments is located on space probes which explore dierent parts of the heliosphere. Such probes are, e.g., Pioneer, Voyager, and Ulysses spacecraft. As a sophisticated and state-of-the-art example, let us consider the AMS02 detector which is to be installed onboard the ISS in 2006. The scheme of the detector is shown in Fig. 11.1. This detector is actually a combination of many types of detectors. The main part is a tracker which consists of 6 orthogonal silicon strip planes (< 2 mm wide, more than 34,000 channels for each plane) in a permanent magnetic eld of about 2 T produced by a superconductive magnet with He cooling. The tracker is able to precisely reconstruct the trajectory of a particle in the magnetic eld which allows to determine the particles rigidity, mass and the incoming direction. Additional devices are: Synchrotron radiation detector (SRD) measures the synchrotron radiation and is primarily devoted to detect electrons.
Transition radiation detector (TRD) measures the transition radiation of particles.
11.1. SPACE-BORNE DETECTORS
149
Figure 11.1: The scheme of the AMS-02 detector.
150
CHAPTER 11. DETECTION OF COSMIC RAYS Time of ight (TOF) system measures the time that the particle needs particle and to reject fake events. Veto counter is a simple electronic counter used as a trigger to reject particles whose trajectories pass through the sides of the detector.
to y through the detector. This allows to estimate the velocity of the
Ring Cherenkov detector (RICH) measures the Cherenkov emission of particles in an aerogel, allowing their energy to be estimated.
Electromagnetic calorimeter (Ecal) also measures the energy of particles.
This combination allows to reconstruct the arrival direction, energy, charge and mass of the CR particle, i.e., identify it completely. For instance, 3 He+ and 4 He+ can be reliably distinguished from each other at the condence level better than 99%.
11.2
Balloon detectors
Modern balloons allow to lift detectors to the altitude of 40-70 km. Earlier, rather small and simple detectors (see, e.g., Fig. 11.2) were own on balloons. However, nowadays rather big and complicated telescopes such as the BESS (Balloon Borne Experiment with Superconducting Solenoidal Spectrometer) detector are own on balloons (see Fig. 11.3). At these high altitudes, the atmosphere above the balloon is negligible for CR, and therefore the balloon-borne detectors can measure primary CR particles, unlike ground based detectors. In this sense they are like low-orbit satellites, only much cheaper and easier to operate. For instance, the rst cosmic antiprotons were discovered by the group of Prof. Bogomolov in late 1970s using a balloon-borne spectrometer.
11.2. BALLOON DETECTORS
151
Figure 11.2: A standard radio-sonde for CR observations in the atmosphere, consisting of (a) Geiger counters, (b) radio-transmitter, (c) altitude sensor, (d) power supply.
Figure 11.3: A scheme of the BESS balloon detector.
152
CHAPTER 11. DETECTION OF COSMIC RAYS The geomagnetic rigidity cuto is still a signicant eect for balloon
observations. Moreover, the atmospheric albedo particles (particles reected or scattered back into space from the atmosphere) play a role and have to be taken into account in balloon observations. The main disadvantage of balloon-borne experiments is that they are campaign-like experiments, operating only for a short time interval.
11.3
Ground-based detectors
Ground based cosmic ray experiments can be divided into dierent subgroups according to the component of the atmospheric cascade (see Fig. 10.1) that they measure.
11.3.1
Neutron monitor
The nucleonic component of the atmospheric cascade is measured by neutron monitors (also called cosmic ray stations). The scheme of a neutron monitor (NM) is shown in Fig. 11.4. The sensor tubes are lled with BF3 gas which is enriched with the B 10 isotope. The paran layer surrounding the tubes is used as a pre-moderator decelerating atmospheric neutrons. The lead layer decelerates neutrons further and produces still more neutrons from the atmospheric neutrons and protons. There is also a plastic layer around the tubes as the nal moderator, making particles almost thermal (< 1 eV) so that the cross section for neutron capture by boron is optimal inside the counter tube:
10
B + n
Li +
The produced fast helium and lithium ions strip electrons from the neutral atoms in the tube, producing charge inside the gas tube. The charge avalange enhanced by a high negative voltage in the tube is detected by the amplier as one count in the central wire inside the tube.
11.3. GROUND-BASED DETECTORS
153
Figure 11.4: The scheme of a neutron monitor. Recently, a new type of NM counter tube was developed. It is lled with
3 He
and uses the reaction

3
He + n
H + p + (7.65 M eV )
These counters are presently in test and calibration phase. There is a global network consisting of 50-70 NMs (depending on the year since some stations have been closed down and others have been opened) located at dierent geomagnetic latitudes (Fig. 11.5) and, correspondingly, at dierent geomagnetic cuto rigidities. This allows to use the network as one unique, global CR spectrometer spaceship Earth (Fig. 10.8). Typically, the bulk of CR observed at NMs are in the energy range 0.3-20 GeV. This is also called the neutron monitor energy range, and it is closely similar to the energy range of eective solar modulation of GCR.
11.3.2
Extensive Air Shower arrays
Extensive atmospheric (air) shower arrays detect the muon component of the atmospheric cascade. Usually, it is an array of simple muon detectors working in coincidence. The size of air shower array varies from hundreds of
154
CHAPTER 11. DETECTION OF COSMIC RAYS
Figure 11.5: The worldwide network of neutron monitors.
11.3. GROUND-BASED DETECTORS
155
Figure 11.6: A couple of units of the extensive atmospheric shower array (EASA) and the Cherenkov array in Canary Islands.
meters to tens of kilometers. A part of the EASA array in Canary Islands is shown in Fig. 11.6. Such arrays allow to measure primary cosmic rays with energy between 1012 1021 eV. The larger the array is the higher primary energies can be measured.
11.3.3
Cherenkov detectors
Fig. 11.6 shows also a couple of units of the Cherenkov array in Canary Islands. Relativistic electrons and positrons, produced in the atmospheric cascade, generate Cherenkov emission in the visible light range when propagating through the air. The Cherenkov array collects such light pulses from a large volume (thousand cubic kilometers). A similar technique is also used to study neutrinos but then the Cherenkov light pulses are produced and detected in deep water or ice.
156
11.3.4
Underground muon experiments
By underground experiments one can study the high-energy part of the muon component. Such experiments use the good penetration capability of muons in matter which allows to easily separate them from the other CR components (except for neutrinos). The underground muon detector may be either a single detector or a small array. (Note also that atmospheric, solar and cosmic neutrinos can also be studied deep underground. However, the size of the detector must be very large in order to compensate the small cross section of neutrinos.)
11.4
Paleoastrophysics
Direct measurements of CR intensity have been carried out on a regular basis only since 1930-1940s. However, it would be interesting to know the CR intensity level on even longer time scales, in particular in order to study the heliosphere and IMF in the past. There are two main methods to the extend CR studies to earlier times.
11.4.1
Meteorites
The cosmic age of a meteorite, i.e., the time interval from the formation of the meteorite as a cosmic body to the moment of their collision with the Earth depends on the meteorite type. Stone meteorites have a typical age of 5-40 million years. Chondrites (consisting of granules of silicate minerals) can reach the age of up to 108 years, while iron meteorites are sometimes as old as 109 years. During their cosmic life time, meteorites are irradiated by cosmic rays, resulting in the production of radioactive isotopes inside the meteorite. These isotopes have very dierent decay times from a month (37 Ar) to millions (10 Be) and billions (40 K) of years. Comparing the abundance of
11.4. PALEOASTROPHYSICS
157
dierent isotopes with dierent decay times in a meteorite, one can estimate the total dose and the radiation rate. Orbits of some meteorites are rather well known from the measurement of the last part of their trajectory or from their belonging to a large meteorite stream (e.g., Perseides or Leonides) whereby they can be associated to a certain region of the heliosphere. Thus, the study of radioactive isotopes in meteorites can help in reconstructing the average cosmic ray intensity on very long time scales.
11.4.2
Cosmogenic Isotopes
The method of cosmogenic isotopes is in principle fairly similar to the meteorite study but it deals with isotopes in some natural archives. In this case, in order to reconstruct the history of CR intensity, one needs to date of the archival samples. There are two commonly used archives of cosmogenic isotopes. One is the radioactive carbon nitrogen includes: n +
14 C 14 14 C
isotope which is mostly produced
in the atmosphere in the capture of a thermal neutron by an atmospheric N

14
C + p.
decays by -decay with a half-life of 5370 years. After being produced, radiocarbon is soon oxidized to a carbon dioxide
(CO2 ) molecule and experiences dierent processes of atmospheric circulation and reservoir exchange (Fig. 11.7). During this cycle, it can, e.g., be consumed by plants or trees, and stored in the natural archive of tree rings where it can only decay. Tree rings are a good sample of an archive because of the possibility for an independent dating based on the counting (and thickness) of tree rings. The relative abundance of
14 C/12 C
is a measure of radiocarbon varia-
tions. The radiocarbon abundance is proportional to the CR intensity at the
158
Figure 11.7: The scheme of the radiocarbon cycle. Numbers on reservoirs show the normalized relative abundance (in promille) of radiocarbon for two dierent models. time of archive formation. The radiocarbon method allows to reconstruct CR intensity of up to 104 years in the past. Once calibrated, the radiocarbon method can be used for independent dating of other samples, e.g., in archaeology. The other possible isotope is the
10 Be
isotope which can, e.g., be stored
in polar ice. This radioactive isotope of beryllium is produced in a chain of reactions in the atmosphere between the CR particle and the atmospheric nitrogen or oxygen nucleus. Then,
10 Be
is attached to aerosols and precipitates within a couple of
months to the ground with rain or snow and may be stored in ice. The
11.4. PALEOASTROPHYSICS half-life time of

10 Be
159
Greenland) ice can be independently dated, this is a good tool to study long-term variations of CR. On the other hand, berillium concentration in ice is aected by local weather conditions which may distort short-time variations.
is 1.5 106 years. Since polar (e.g., Antarctic or
160
Bibliography
Books
[1] V.S. Berezinskii, S.V. Bulanov, V.A. Dogiel, V.L.Ginzburg (editor), V.S. Ptuskin, Astrophysics of Cosmic Rays, North-Holland, Amsterdam, 1990. [2] R. Clay and B. Dawson, Cosmic Bullets: high energy particles in astrophysics, Allen & Unwin, Sidney, 1997. [3] L.I. Dorman, Cosmic Rays: variations and space explorations, NorthHolland Publishing Co., Amsterdam, 1974. [4] T.K. Gaisser, Cosmic Rays and Particle Physics, Cambridge University Press, Cambridge, 1990. [5] P.K.F. Grieder, Cosmic Rays at the Earth, Elsevier, Amsterdam, 2001. [6] M. Pomeranz, Cosmic Rays, Van Nostrand Reinhold Co., NY, 1971. [7] K. Sakurai, Physics of Solar Cosmic Rays, University of Tokio Press, Tokyo, 1974. [8] A. E. Sandstrm, Cosmic Ray Physics, North-Holland Publishing Co., o Amsterdam, 1965. [9] I.N. Toptygin, Cosmic Rays in Interplanetary Magnetic Fields, D.Redel Publ.Co., Dordrecht, 1985. [10] A.W. Wolfendale, Cosmic Rays, George Newnes Ltd., London, 1963. 161
162
BIBLIOGRAPHY
Review papers
[11] Balogh, A., J.T. Gosling, J.R. Jokipii, R.Kallenbach, H. Kunow, (eds.) Corotating Interaction Regions, Space Sci.Rev., 89(1/2), 1999. [12] L.I. Dorman and D. Venkatesan, Solar Cosmic Rays, Space Sci. Rev., v.64, 183-363, 1993. [13] Fisk, L.A., J. R. Jopikii, G. M. Simnett, R. von Steiger, and K.-P. Wenzel, (eds.) Cosmic Rays in the Heliosphere, Space Sci.Rev., 83(1/2), 1998. [14] J. A. Lockwood, H. Debrunner, Solar Flare Particle Measurements with Neutron Monitors, Space Sci. Rev., v.88, 501-528, 1999. [15] J. A. Miller, Particle Acceleration in Impulsive Solar Flares, Space Sci.Rev., v.86, 79-105, 1998. [16] K. Sakurai, High-energy Phenomena on the Sun: an Introductory Review, Space Sci. Rev., v.51, 1-9, 1989. [17] P.H. Stoker, Relativistic Solar Events, Space Sci. Rev., v.73, 327-385, 1994. [18] I.G.Usoskin, G.A.Kovaltsov, H.Kananen, P.Tanskanen, The World Neutron Monitor Network as a Tool for the Study of Solar Neutrons, Annales Geophysicae, v.15, p.375-386, 1997.
Relevant web-sites
[19] Cambridge University Press Handbook of Space Astronomy and Astrophysics (http : //adsbit.harvard.edu/books/hsaa/idx.html) [20] Oulu Space Physics Textbook (http : //www.oulu.f i/ spaceweb/textbook/)
BIBLIOGRAPHY [21] High Energy Astrophysics at MSSL
163
(http : //www.mssl.ucl.ac.uk/wwwa stro/lecturenotes/hea/hea.html) [22] Solar Physics at MSSL (http : //www.mssl.ucl.ac.uk/wwws olar/homepage.html) [23] Cosmic Rays at NGDC (http : //web.ngdc.noaa.gov/stp/SOLAR/COSM ICR AY S/cosmic.html) [24] The Cosmic Web at Utah University (http : //www.physics.utah.edu/research/cosmicw eb/index.html) [25] Martindales virtual astronomy, astrophysics and space science center (http : //www.martindalecenter.com/GradSpace.html) [26] High Energy Astrophysics (http : //dustbunny.physics.indiana.edu/ dzierba/HEP A/) [27] Cosmic Ray Learning Center by NASA (http : //helios.gsf c.nasa.gov/cosmic.html)
164
BIBLIOGRAPHY
Units and denitions

Units
eV (electron volt) 1 eV = 1.6 1019 J = 1.6 1012 erg: The energy gained by an electron falling through a potential dierence of 1 volt. AU (astronomical unit) 1 AU = 1.495 1011 m: The mean distance to the Sun from the Earth.
Denitions
CME Coronal mass ejection is a huge bubble of plasma ejected from the solar corona during several hours. CMEs seem to be more related to prominence eruptions than solar ares. Energy By energy we mean here the kinetic energy (not the total energy) of the particle unless specially mentioned. GCR Galactic cosmic rays are cosmic rays of galactic or extra-galactic origin. Gyroradius Radius of gyration, or cyclotron radius. The radius of the circular orbit of a charged particle gyrating around its guiding center: r= mv = P /B, |q|B
where m, q, P , v are the mass, charge, rigidity and the velocity perpendicular to the magnetic eld line of the gyrating particle. 165
166
BIBLIOGRAPHY
IMF Interplanetary Magnetic Field lines have the average shape of an Archimedian spiral due to radial solar wind and the solar rotation. Pitch angle The angle between magnetic eld B and velocity vector of a charged particle, v: sin = vtotal v ,
where v is the velocity component perpendicular to B. Rigidity Magnetic rigidity of a charged particle is dened as P = p , |q|
and charge of the particle, respectively.
where p and |q| are the momentum (either classical or relativistic)

Heliosheric Physics and Cosmic Rays

Uploaded by

Copyright:

Available Formats

Heliosheric Physics and Cosmic Rays

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Heliosheric Physics and Cosmic Rays

Uploaded by

Copyright:

Available Formats

761656S . Heliospheric Physics and Cosmic Rays .

Lecture notes Fall term 2003

Introduction to Cosmic Rays

1.1. MILESTONES OF COSMIC RAY RESEARCH

CHAPTER 1. INTRODUCTION TO COSMIC RAYS

Figure 1.2: Artists view of an atmospheric shower.

1.1. MILESTONES OF COSMIC RAY RESEARCH

CHAPTER 1. INTRODUCTION TO COSMIC RAYS

What is a Cosmic Ray?

1.2. WHAT IS A COSMIC RAY?

Figure 1.3: Cygnus loop supernova remnant. Image by ROSAT.

CHAPTER 1. INTRODUCTION TO COSMIC RAYS

Galactic Cosmic Rays

CHAPTER 2. GALACTIC COSMIC RAYS

trons from atoms, leaving isolated nuclei and electrons.

CHAPTER 2. GALACTIC COSMIC RAYS

2.2. ENERGY SPECTRUM

CHAPTER 2. GALACTIC COSMIC RAYS

Figure 2.3: Dierential energy spectrum of galactic cosmic rays.

2.2. ENERGY SPECTRUM

I(E) E cm2 s1 sr1 GeV 1 ,

3.1 This is known as the knee of the spectrum. Actually, the

CHAPTER 2. GALACTIC COSMIC RAYS

2.2. ENERGY SPECTRUM

Figure 2.5: The dierential GCR spectrum multiplied by E 2.5 .

Origin of Cosmic Rays

Ultra High Energy Cosmic Rays

2.3. ORIGIN OF COSMIC RAYS

High Energy Cosmic Rays

Lifetime of GCR in the Galaxy

The measured amount of

in cosmic rays implies that, on an

2.3. ORIGIN OF COSMIC RAYS

Figure 2.6: The ROSAT image of the Crab nebula.

CHAPTER 2. GALACTIC COSMIC RAYS

Figure 2.7: A schematic side view of the Galaxy.

2.3. ORIGIN OF COSMIC RAYS

CHAPTER 2. GALACTIC COSMIC RAYS

Acceleration of cosmic rays

CHAPTER 3. ACCELERATION OF COSMIC RAYS

3.1. FERMI ACCELERATION

where = the cloud.

are the relative speed and Lorentz factor of

CHAPTER 3. ACCELERATION OF COSMIC RAYS

0 E2 = 2 E1 (1 cos1 )(1 + cos2 )

0 0 E2 E1 E 1 cos1 + cos2 2 cos1 cos2 ' 1. = E1 E 1 2

Averaging Eq. 3.4 over angles, one can obtain that

leading to < cos1 >= 1 . 3

1 + 1 2 4 E 3 1 2. = E 1 2 3 Thus, the net energy gain (averaged per collision) is

3.1. FERMI ACCELERATION

E(t) = Ei exp( where tc = c / 2 .

CHAPTER 3. ACCELERATION OF COSMIC RAYS

3.1. FERMI ACCELERATION

CHAPTER 3. ACCELERATION OF COSMIC RAYS

tional to 2 cos1 (1 cos1 0) leading to < cos1 >= 2/3. Similarly,

3.2. MAGNETIC PUMPING

CHAPTER 3. ACCELERATION OF COSMIC RAYS

3.2. MAGNETIC PUMPING

or, after one complete cycle: po p5 = 3

will not consider them within the scope of this course.