24.

Cosmic rays 1
24. COSMIC RAYS
Revised August 2011 by J.J. Beatty (Ohio State Univ.) and J. Matthews (Louisiana
State Univ. and Southern Univ.); revised August 2009 by T.K. Gaisser and T. Stanev
(Bartol Research Inst., Univ. of Delaware)

24.1. Primary spectra

The cosmic radiation incident at the top of the terrestrial atmosphere includes all
stable charged particles and nuclei with lifetimes of order 106 years or longer. Technically,
“primary” cosmic rays are those particles accelerated at astrophysical sources and
“secondaries” are those particles produced in interaction of the primaries with interstellar
gas. Thus electrons, protons and helium, as well as carbon, oxygen, iron, and other nuclei
synthesized in stars, are primaries. Nuclei such as lithium, beryllium, and boron (which
are not abundant end-products of stellar nucleosynthesis) are secondaries. Antiprotons
and positrons are also in large part secondary. Whether a small fraction of these particles
may be primary is a question of current interest.
Apart from particles associated with solar flares, the cosmic radiation comes from
outside the solar system. The incoming charged particles are “modulated” by the solar
wind, the expanding magnetized plasma generated by the Sun, which decelerates and
partially excludes the lower energy galactic cosmic rays from the inner solar system.
There is a significant anticorrelation between solar activity (which has an alternating
eleven-year cycle) and the intensity of the cosmic rays with energies below about 10 GeV.
In addition, the lower-energy cosmic rays are affected by the geomagnetic field, which they
must penetrate to reach the top of the atmosphere. Thus the intensity of any component
of the cosmic radiation in the GeV range depends both on the location and time.
There are four different ways to describe the spectra of the components of the cosmic
radiation: (1) By particles per unit rigidity. Propagation (and probably also acceleration)
through cosmic magnetic fields depends on gyroradius or magnetic rigidity, R, which is
gyroradius multiplied by the magnetic field strength:

pc
R= = rL B . (24.1)
Ze

(2) By particles per energy-per-nucleon. Fragmentation of nuclei propagating through

the interstellar gas depends on energy per nucleon, since that quantity is approximately
conserved when a nucleus breaks up on interaction with the gas. (3) By nucleons per
energy-per-nucleon. Production of secondary cosmic rays in the atmosphere depends
on the intensity of nucleons per energy-per-nucleon, approximately independently of
whether the incident nucleons are free protons or bound in nuclei. (4) By particles per
energy-per-nucleus. Air shower experiments that use the atmosphere as a calorimeter
generally measure a quantity that is related to total energy per particle.
The units of differential intensity I are [m−2 s−1 sr−1 E −1 ], where E represents the units
of one of the four variables listed above.

K. Nakamura et al.(PDG), JP G 37, 075021 (2010) and 2011 partial update for the 2012 edition (pdg.lbl.gov)
February 16, 2012 14:07
2 24. Cosmic rays
The intensity of primary nucleons in the energy range from several GeV to somewhat
beyond 100 TeV is given approximately by
nucleons
IN (E) ≈ 1.8 × 104 (E/1 GeV)−α , (24.2)
m2 s sr GeV
where E is the energy-per-nucleon (including rest mass energy) and α (≡ γ + 1) = 2.7
is the differential spectral index of the cosmic ray flux and γ is the integral spectral
index. About 79% of the primary nucleons are free protons and about 70% of the rest are
nucleons bound in helium nuclei. The fractions of the primary nuclei are nearly constant
over this energy range (possibly with small but interesting variations). Fractions of both
primary and secondary incident nuclei are listed in Table 24.1. Figure 24.1 shows the
major components for energies greater than 2 GeV/nucleon.

Figure 24.1: Fluxes of nuclei of the primary cosmic radiation in particles per
energy-per-nucleus are plotted vs energy-per-nucleus using data from Refs. [1–12].
The figure was created by P. Boyle and D. Muller. Color version at end of book.

The composition and energy spectra of nuclei are typically interpreted in the context
of propagation models, in which the sources of the primary cosmic radiation are located

February 16, 2012 14:07

 24. Cosmic rays 3
within the galaxy [16]. The ratio of secondary to primary nuclei is observed to decrease
with increasing energy, a fact interpreted to mean that the lifetime of cosmic rays in the
galaxy decreases with energy. Measurements of radioactive “clock” isotopes in the low
energy cosmic radiation are consistent with a lifetime in the galaxy of about 15 Myr.

Table 24.1: Relative abundances F of cosmic-ray nuclei at 10.6 GeV/nucleon nor-

malized to oxygen (≡ 1) [6]. The oxygen flux at kinetic energy of 10.6 GeV/nucleon
is 3.26 × 10−2 (m2 s sr GeV/nucleon)−1 . Abundances of hydrogen and helium are
from Refs. [2,3]. Note that one can not use these values to extend the cosmic
ray flux to high energy because the power law indicies for each element may differ
slightly.

Z Element F Z Element F

1 H 540 13–14 Al-Si 0.19

2 He 26 15–16 P-S 0.03
3–5 Li-B 0.40 17–18 Cl-Ar 0.01
6–8 C-O 2.20 19–20 K-Ca 0.02
9–10 F-Ne 0.30 21–25 Sc-Mn 0.05
11–12 Na-Mg 0.22 26–28 Fe-Ni 0.12

Cosmic rays are nearly isotropic at most energies due to diffusive propagation in
the galactic magnetic field. Milagro [13], IceCube [14], and The Tibet-III air shower
array [15] have observed anisotropy at the level of about 10−3 for cosmic rays with energy
of a few TeV, possibly due to nearby sources.
The spectrum of electrons and positrons incident at the top of the atmosphere is
expected to steepen by one power of E at an energy of ∼5 GeV because of the strong
synchrotron energy loss in the galactic magnetic fields. The ATIC experiment [21]
measured an excess of electrons above 100 GeV followed by a steepening above 1,000 GeV.
The Fermi/LAT γ-ray observatory confirmed the relatively flat electron spectrum [23]
without confirming the peak of the ATIC excess at ∼800 GeV.
The PAMELA satellite experiment [24] measured the positron to electron ratio to
increase above 10 GeV instead of the expected decrease [25] at higher energy. The
structure in the electron spectrum as well as the increase in the positron fraction could
be related to contributions from individual nearby sources emerging above a background
suppressed at high energy by synchrotron losses [26]. The low positron to electron ratio
below 10 GeV is due to the new solar magnetic field polarity after the year 2001.
The ratio of antiprotons to protons is ∼ 2 × 10−4 [27] at around 10–20 GeV, and there
is clear evidence [28] for the kinematic suppression at lower energy that is the signature
of secondary antiprotons. The p/p ratio also shows a strong dependence on the phase
and polarity of the solar cycle [29] in the opposite sense to that of the positron fraction.

February 16, 2012 14:07

4 24. Cosmic rays

E3 dN/dE, m-2s-1sr-1GeV2
100

positron/electron ratio
10 0.10

0.05

0.02

1
1 10 100 1000 10000
Energy, GeV
Figure 24.2: Differential spectrum of electrons plus positrons multiplied by E 3
(data from [17–23]) . The line shows the proton spectrum multiplied by 0.01. The
inset shows the positron to electron ratio measured by PAMELA [24] compared to
the expected decrease [25].

There is at this time no evidence for a significant primary component of antiprotons. No

antihelium or antideuteron has been found in the cosmic radiation. The best measured
upper limit on the ratio antihelium/helium is currently approximately 1 × 10−7 [30]
The upper limit on the flux of antideuterons around 1 GeV/nucleon is approximately
2 × 10−4 (m2 s sr GeV/nucleon)−1 [31].

24.2. Cosmic rays in the atmosphere

Figure 24.3 shows the vertical fluxes of the major cosmic ray components in the
atmosphere in the energy region where the particles are most numerous (except for
electrons, which are most numerous near their critical energy, which is about 81 MeV in
air). Except for protons and electrons near the top of the atmosphere, all particles are
produced in interactions of the primary cosmic rays in the air. Muons and neutrinos are
products of the decay chain of charged mesons, while electrons and photons originate in
decays of neutral mesons.
Most measurements are made at ground level or near the top of the atmosphere,
but there are also measurements of muons and electrons from airplanes and balloons.
Fig. 24.3 includes recent measurements of negative muons [32–36]. Since µ+ (µ− ) are
produced in association with νµ (ν µ ), the measurement of muons near the maximum of

February 16, 2012 14:07

 24. Cosmic rays 5

Altitude (km)
15 10 5 3 2 1 0
10000

1000
[m–2 s–1 sr–1] _
νµ + νµ
100 µ+ + µ−

10 p+n
Vertical flux

1
e+ + e−

π+ + π−
0.1

0.01
0 200 400 600
800 1000
Atmospheric depth [g cm–2]
Figure 24.3: Vertical fluxes of cosmic rays in the atmosphere with E > 1 GeV
estimated from the nucleon flux of Eq. (24.2). The points show measurements of
negative muons with Eµ > 1 GeV [32–36].

the intensity curve for the parent pions serves to calibrate the atmospheric νµ beam [37].
Because muons typically lose almost 2 GeV in passing through the atmosphere, the
comparison near the production altitude is important for the sub-GeV range of νµ (ν µ )
energies.
The flux of cosmic rays through the atmosphere is described by a set of coupled cascade
equations with boundary conditions at the top of the atmosphere to match the primary
spectrum. Numerical or Monte Carlo calculations are needed to account accurately for
decay and energy-loss processes, and for the energy-dependences of the cross sections and
of the primary spectral index γ. Approximate analytic solutions are, however, useful in

February 16, 2012 14:07

6 24. Cosmic rays
limited regions of energy [38,39]. For example, the vertical intensity of charged pions
with energy Eπ ≪ ǫπ = 115 GeV is

ZNπ X Eπ
Iπ (Eπ , X) ≈ IN (Eπ , 0) e−X/Λ , (24.3)
λN ǫπ

where Λ is the characteristic length for exponential attenuation of the parent nucleon
flux in the atmosphere. This expression has a maximum at X = Λ ≈121±4 g cm−2 [40],
which corresponds to an altitude of 15 kilometers. The quantity ZNπ is the spectrum-
weighted moment of the inclusive distribution of charged pions in interactions of nucleons
with nuclei of the atmosphere. The intensity of low-energy pions is much less than that
of nucleons because ZNπ ≈ 0.079 is small and because most pions with energy much less
than the critical energy ǫπ decay rather than interact.

24.3. Cosmic rays at the surface

24.3.1. Muons : Muons are the most numerous charged particles at sea level (see
Fig. 24.3). Most muons are produced high in the atmosphere (typically 15 km) and
lose about 2 GeV to ionization before reaching the ground. Their energy and angular
distribution reflect a convolution of the production spectrum, energy loss in the
atmosphere, and decay. For example, 2.4 GeV muons have a decay length of 15 km,
which is reduced to 8.7 km by energy loss. The mean energy of muons at the ground
is ≈ 4 GeV. The energy spectrum is almost flat below 1 GeV, steepens gradually to
reflect the primary spectrum in the 10–100 GeV range, and steepens further at higher
energies because pions with Eπ > ǫπ tend to interact in the atmosphere before they
decay. Asymptotically (Eµ ≫ 1 TeV), the energy spectrum of atmospheric muons is one
power steeper than the primary spectrum. The integral intensity of vertical muons above
1 GeV/c at sea level is ≈ 70 m−2 s−1 sr−1 [41,42], with recent measurements [43–45]
favoring a lower normalization by 10-15%. Experimentalists are familiar with this number
in the form I ≈ 1 cm−2 min−1 for horizontal detectors. The overall angular distribution
of muons at the ground is ∝ cos2 θ, which is characteristic of muons with Eµ ∼ 3 GeV. At
lower energy the angular distribution becomes increasingly steep, while at higher energy
it flattens, approaching a sec θ distribution for Eµ ≫ ǫπ and θ < 70◦ .
Figure 24.4 shows the muon energy spectrum at sea level for two angles. At large angles
low energy muons decay before reaching the surface and high energy pions decay before
they interact, thus the average muon energy increases. An approximate extrapolation
formula valid when muon decay is negligible (Eµ > 100/ cos θ GeV) and the curvature of
the Earth can be neglected (θ < 70◦ ) is

dNµ 0.14 Eµ−2.7

≈
dEµ dΩ cm2 s sr GeV
 
 
 1 0.054 
× + , (24.4)
 1 + 1.1Eµ cos θ

1+
1.1Eµ cos θ 

115 GeV 850 GeV
February 16, 2012 14:07
 24. Cosmic rays 7
where the two terms give the contribution of pions and charged kaons. Eq. (24.4) neglects
a small contribution from charm and heavier flavors which is negligible except at very
high energy [50].

pµ2.7dN/dpµ [m−2 s−1 sr−1(GeV/c)1.7]

1000

100

1 10 100 1000
[GeV/c] pµ
Figure 24.4: Spectrum of muons at θ = 0◦ (¨ [41], ¥ [46], H [47], N [48], ×,
+ [43], ◦ [44], and • [45] and θ = 75◦ ♦ [49]) . The line plots the result from
Eq. (24.4) for vertical showers.

The muon charge ratio reflects the excess of π + over π − and K + over K − in the
forward fragmentation region of proton initiated interactions together with the fact that
there are more protons than neutrons in the primary spectrum. The increase with energy
of µ+ /µ− shown in Fig. 24.5 reflects the increasing importance of kaons in the TeV
range [51] and indicates a significant contribution of associated production by cosmic-ray
protons (p → Λ + K + ). The same process is even more important for atmospheric
neutrinos at high energy.
24.3.2. Electromagnetic component : At the ground, this component consists of
electrons, positrons, and photons primarily from cascades initiated by decay of neutral
and charged mesons. Muon decay is the dominant source of low-energy electrons at sea
level. Decay of neutral pions is more important at high altitude or when the energy
threshold is high. Knock-on electrons also make a small contribution at low energy [52].
The integral vertical intensity of electrons plus positrons is very approximately 30, 6,
and 0.2 m−2 s−1 sr−1 above 10, 100, and 1000 MeV respectively [42,53], but the exact
numbers depend sensitively on altitude, and the angular dependence is complex because
of the different altitude dependence of the different sources of electrons [52–54]. The
ratio of photons to electrons plus positrons is approximately 1.3 above 1 GeV and 1.7
below the critical energy [54].

February 16, 2012 14:07

8 24. Cosmic rays

1.6

BESS
1.5 L3C
MINOS

Fµ+ /Fµ− 1.4

1.3

1.2

1.1
1.0 10 102 103 104
pµ [GeV/c]
Figure 24.5: Muon charge ratio as a function of the muon momentum from
Refs. [44,45,51].

24.3.3. Protons : Nucleons above 1 GeV/c at ground level are degraded remnants of
the primary cosmic radiation. The intensity is approximately IN (E, 0) × exp(−X/ cos θΛ)
for θ < 70◦ . At sea level, about 1/3 of the nucleons in the vertical direction are
neutrons (up from ≈ 10% at the top of the atmosphere as the n/p ratio approaches
equilibrium). The integral intensity of vertical protons above 1 GeV/c at sea level is
≈ 0.9 m−2 s−1 sr−1 [42,55].

24.4. Cosmic rays underground

Only muons and neutrinos penetrate to significant depths underground. The muons
produce tertiary fluxes of photons, electrons, and hadrons.
24.4.1. Muons : As discussed in Section 27.6 of this Review, muons lose energy by
ionization and by radiative processes: bremsstrahlung, direct production of e+ e− pairs,
and photonuclear interactions. The total muon energy loss may be expressed as a function
of the amount of matter traversed as

dEµ
− = a + b Eµ , (24.5)
dX

where a is the ionization loss and b is the fractional energy loss by the three radiation
processes. Both are slowly varying functions of energy. The quantity ǫ ≡ a/b (≈ 500 GeV
in standard rock) defines a critical energy below which continuous ionization loss is more
important than radiative losses. Table 24.2 shows a and b values for standard rock, and

February 16, 2012 14:07

 24. Cosmic rays 9
b for ice, as a function of muon energy. The second column of Table 24.2 shows the
muon range in standard rock (A = 22, Z = 11, ρ = 2.65 g cm−3 ). These parameters are
quite sensitive to the chemical composition of the rock, which must be evaluated for each
location.

Table 24.2: Average muon range R and energy loss parameters a and b calculated
for standard rock [56] and the total energy loss parameter b for ice. Range is given
in km-water-equivalent, or 105 g cm−2 .
P P
Eµ R a bbrems bpair bnucl bi b(ice)
−1
GeV km.w.e. MeV g cm 2 10−6 g−1 cm2

10 0.05 2.17 0.70 0.70 0.50 1.90 1.66

100 0.41 2.44 1.10 1.53 0.41 3.04 2.51
1000 2.45 2.68 1.44 2.07 0.41 3.92 3.17
10000 6.09 2.93 1.62 2.27 0.46 4.35 3.78

The intensity of muons underground can be estimated from the muon intensity in the
atmosphere and their rate of energy loss. To the extent that the mild energy dependence
of a and b can be neglected, Eq. (24.5) can be integrated to provide the following relation
between the energy Eµ,0 of a muon at production in the atmosphere and its average
energy Eµ after traversing a thickness X of rock (or ice or water):

Eµ,0 = (Eµ + ǫ) ebX − ǫ . (24.6)

Especially at high energy, however, fluctuations are important and an accurate calculation
requires a simulation that accounts for stochastic energy-loss processes [57].
There are two depth regimes for which Eq. (24.6) can be simplified. For X ≪
b−1 ≈ 2.5 km water equivalent, Eµ,0 ≈ Eµ (X) + aX, while for X ≫ b−1 Eµ,0 ≈
(ǫ + Eµ (X)) exp(bX). Thus at shallow depths the differential muon energy spectrum is
approximately constant for Eµ < aX and steepens to reflect the surface muon spectrum
for Eµ > aX, whereas for X > 2.5 km.w.e. the differential spectrum underground is
again constant for small muon energies but steepens to reflect the surface muon spectrum
for Eµ > ǫ ≈ 0.5 TeV. In the deep regime the shape is independent of depth although
the intensity decreases exponentially with depth. In general the muon spectrum at slant
depth X is
dNµ (X) dNµ dEµ,0 dNµ bX
= = e , (24.7)
dEµ dEµ,0 dEµ dEµ,0
where Eµ,0 is the solution of Eq. (24.6) in the approximation neglecting fluctuations.
Fig. 24.6 shows the vertical muon intensity versus depth. In constructing this “depth-
intensity curve,” each group has taken account of the angular distribution of the muons

February 16, 2012 14:07

10 24. Cosmic rays
in the atmosphere, the map of the overburden at each detector, and the properties of the
local medium in connecting measurements at various slant depths and zenith angles to
the vertical intensity. Use of data from a range of angles allows a fixed detector to cover
a wide range of depths. The flat portion of the curve is due to muons produced locally by
charged-current interactions of νµ . The inset shows the vertical intensity curve for water
and ice published in Refs. [59–62]. It is not as steep as the one for rock because of the
lower muon energy loss in water.

1 2 5 10

1 10 100

Figure 24.6: Vertical muon intensity vs depth (1 km.w.e. = 105 g cm−2 of standard
rock). The experimental data are from: ♦: the compilations of Crouch [58], ¤:
Baksan [63], ◦: LVD [64], •: MACRO [65], ¥: Frejus [66], and △: SNO [67].
The shaded area at large depths represents neutrino-induced muons of energy above
2 GeV. The upper line is for horizontal neutrino-induced muons, the lower one
for vertically upward muons. Darker shading shows the muon flux measured by
the SuperKamiokande experiment. The inset shows the vertical intensity curve for
water and ice published in Refs. [59–62].

February 16, 2012 14:07

 24. Cosmic rays 11
24.4.2. Neutrinos :
Because neutrinos have small interaction cross sections, measurements of atmospheric
neutrinos require a deep detector to avoid backgrounds. There are two types of
measurements: contained (or semi-contained) events, in which the vertex is determined
to originate inside the detector, and neutrino-induced muons. The latter are muons that
enter the detector from zenith angles so large (e.g., nearly horizontal or upward) that
they cannot be muons produced in the atmosphere. In neither case is the neutrino flux
measured directly. What is measured is a convolution of the neutrino flux and cross
section with the properties of the detector (which includes the surrounding medium in
the case of entering muons).
Contained and semi-contained events reflect neutrinos in the sub-GeV to multi-GeV
region where the product of increasing cross section and decreasing flux is maximum. In
the GeV region the neutrino flux and its angular distribution depend on the geomagnetic
location of the detector and, to a lesser extent, on the phase of the solar cycle. Naively,
we expect νµ /νe = 2 from counting neutrinos of the two flavors coming from the chain of
pion and muon decay. Contrary to expectation, however, the numbers of the two classes
of events are similar rather than different by a factor of two. This is now understood
to be a consequence of neutrino flavor oscillations [70]. (See the article on neutrino
properties in this Review.)
Two well-understood properties of atmospheric cosmic rays provide a standard for
comparison of the measurements of atmospheric neutrinos to expectation. These are
the “sec θ effect” and the “east-west effect” [69]. The former refers originally to the
enhancement of the flux of > 10 GeV muons (and neutrinos) at large zenith angles because
the parent pions propagate more in the low density upper atmosphere where decay is
enhanced relative to interaction. For neutrinos from muon decay, the enhancement near
the horizontal becomes important for Eν > 1 GeV and arises mainly from the increased
pathlength through the atmosphere for muon decay in flight. Fig. 24.7 from Ref. 68 shows
a comparison between measurement and expectation for the zenith angle dependence of
multi-GeV electron-like (mostly νe ) and muon-like (mostly νµ ) events separately. The νe
show an enhancement near the horizontal and approximate equality for nearly upward
(cos θ ≈ −1) and nearly downward (cos θ ≈ 1) events. There is, however, a very significant
deficit of upward (cos θ < 0) νµ events, which have long pathlengths comparable to the
radius of the Earth. This feature is the principal signature for oscillations [70].
Muons that enter the detector from outside after production in charged-current
interactions of neutrinos naturally reflect a higher energy portion of the neutrino
spectrum than contained events because the muon range increases with energy as well
as the cross section. The relevant energy range is ∼ 10 < Eν < 1000 GeV, depending
somewhat on angle. Neutrinos in this energy range show a sec θ effect similar to
muons (see Eq. (24.4)). This causes the flux of horizontal neutrino-induced muons to
be approximately a factor two higher than the vertically upward flux. The upper and
lower edges of the horizontal shaded region in Fig. 24.6 correspond to horizontal and
vertical intensities of neutrino-induced muons. Table 24.3 gives the measured fluxes of
upward-moving neutrino-induced muons averaged over the lower hemisphere. Generally
the definition of minimum muon energy depends on where it passes through the detector.

February 16, 2012 14:07

12 24. Cosmic rays

300
multi-GeV e-like multi-GeV µ-like + PC

250

Number of events
200

150

100

0
−1 −0.5 0 0.5 −1 −0.5 0 0.5 1
cos θ cos θ

Figure 24.7: Zenith-angle dependence of multi-GeV neutrino interactions from

SuperKamiokande [68]. The shaded boxes show the expectation in the absence of
any oscillations.

Table 24.3: Measured fluxes (10−9 m−2 s−1 sr−1 ) of neutrino-induced muons as a
function of the effective minimum muon energy Eµ .

Eµ > 1 GeV 1 GeV 1 GeV 2 GeV 3 GeV 3 GeV

Ref. CWI [71] Baksan [72] MACRO [73] IMB [74] Kam [75] SuperK [76]
Fµ 2.17±0.21 2.77±0.17 2.29 ± 0.15 2.26±0.11 1.94±0.12 1.74±0.07

The tabulated effective minimum energy estimates the average over various accepted
trajectories.

24.5. Air showers

So far we have discussed inclusive or uncorrelated fluxes of various components of the
cosmic radiation. An air shower is caused by a single cosmic ray with energy high enough
for its cascade to be detectable at the ground. The shower has a hadronic core, which acts
as a collimated source of electromagnetic subshowers, generated mostly from π 0 → γ γ
decays. The resulting electrons and positrons are the most numerous charged particles in
the shower. The number of muons, produced by decays of charged mesons, is an order
of magnitude lower. Air showers spread over a large area on the ground, and arrays of
detectors operated for long times are useful for studying cosmic rays with primary energy
E0 > 100 TeV, where the low flux makes measurements with small detectors in balloons
and satellites difficult.

February 16, 2012 14:07

 24. Cosmic rays 13
Greisen [77] gives the following approximate expressions for the numbers and lateral
distributions of particles in showers at ground level. The total number of muons Nµ with
energies above 1 GeV is
³ ´3/4
Nµ (> 1 GeV) ≈ 0.95 × 105 Ne /106 , (24.8)

where Ne is the total number of charged particles in the shower (not just e± ). The
number of muons per square meter, ρµ , as a function of the lateral distance r (in meters)
from the center of the shower is

1 1.25 −0.75 ³
µ ¶
1.25 Nµ r ´−2.5
ρµ = r 1+ , (24.9)
2π Γ(1.25) 320 320

where Γ is the gamma function. The number density of charged particles is

ρe = C1 (s, d, C2 ) x(s−2) (1 + x)(s−4.5) (1 + C2 xd ) . (24.10)

Here s, d, and C2 are parameters in terms of which the overall normalization constant
C1 (s, d, C2) is given by
Ne
C1 (s, d, C2) = [ B(s, 4.5 − 2s)
2πr12

+ C2 B(s + d, 4.5 − d − 2s)]−1 , (24.11)

where B(m, n) is the beta function. The values of the parameters depend on shower size
(Ne ), depth in the atmosphere, identity of the primary nucleus, etc. For showers with
Ne ≈ 106 at sea level, Greisen uses s = 1.25, d = 1, and C2 = 0.088. Finally, x is r/r1 ,
where r1 is the Molière radius, which depends on the density of the atmosphere and hence
on the altitude at which showers are detected. At sea level r1 ≈ 78 m. It increases with
altitude as the air density decreases. (See the section on electromagnetic cascades in the
article on the passage of particles through matter in this Review).
The lateral spread of a shower is determined largely by Coulomb scattering of the
many low-energy electrons and is characterized by the Molìere radius. The lateral spread
of the muons (ρµ ) is larger and depends on the transverse momenta of the muons at
production as well as multiple scattering.
There are large fluctuations in development from shower to shower, even for showers of
the same energy and primary mass—especially for small showers, which are usually well
past maximum development when observed at the ground. Thus the shower size Ne and
primary energy E0 are only related in an average sense, and even this relation depends
on depth in the atmosphere. One estimate of the relation is [84]

E0 ∼ 3.9 × 106 GeV (Ne /106 )0.9 (24.12)

for vertical showers with 1014 < E < 1017 eV at 920 g cm−2 (965 m above sea level). As
E0 increases the shower maximum (on average) moves down into the atmosphere and the

February 16, 2012 14:07

14 24. Cosmic rays
relation between Ne and E0 changes. Moreover, because of fluctuations, Ne as a function
of E0 is not correctly obtained by inverting Eq. (24.12). At the maximum of shower
development, there are approximately 2/3 particles per GeV of primary energy.
There are three common types of air shower detectors: shower arrays that study
the shower size Ne and the lateral distribution on the ground, Cherenkov detectors
that detect the Cherenkov radiation emitted by the charged particles of the shower,
and fluorescence detectors that study the nitrogen fluorescence excited by the charged
particles in the shower. The fluorescence light is emitted isotropically so the showers can
be observed from the side. Detailed simulations and cross-calibrations between different
types of detectors are necessary to establish the primary energy spectrum from air-shower
experiments.
Figure 24.8 shows the “all-particle” spectrum. The differential energy spectrum has
been multiplied by E 2.6 in order to display the features of the steep spectrum that are
otherwise difficult to discern. The steepening that occurs between 1015 and 1016 eV is
known as the knee of the spectrum. The feature around 1018.5 eV is called the ankle of
the spectrum.

Knee
104
E 2.6 F(E) [GeV1.6 m-2 s-1 sr-1]

Grigorov
JACEE
MGU
3
10 Tien-Shan
Tibet07 Ankle
Akeno
CASA-MIA
HEGRA
102
Fly’s Eye
Kascade
Kascade Grande 2011
AGASA
10 HiRes 1
HiRes 2
Telescope Array 2011
Auger 2011

1
13 15 16 17 18 19 20
10 1014 10 10 10 10 10 10
E [eV]

Figure 24.8: The all-particle spectrum as a function of E (energy-per-nucleus)

from air shower measurements [79–90,100–104]. Color version at end of book.

Measurements of flux with small air shower experiments in the knee region differ by
as much as a factor of two, indicative of systematic uncertainties in interpretation of the
data. (For a review see Ref. 78.) In establishing the spectrum shown in Fig. 24.8, efforts

February 16, 2012 14:07

 24. Cosmic rays 15
have been made to minimize the dependence of the analysis on the primary composition.
Ref. 87 uses an unfolding procedure to obtain the spectra of the individual components,
giving a result for the all-particle spectrum between 1015 and 1017 eV that lies toward
the upper range of the data shown in Fig. 24.8. In the energy range above 1017 eV,
the fluorescence technique [89] is particularly useful because it can establish the primary
energy in a model-independent way by observing most of the longitudinal development
of each shower, from which E0 is obtained by integrating the energy deposition in
the atmosphere. The result, however, depends strongly on the light absorption in the
atmosphere and the calculation of the detector’s aperture.
Assuming the cosmic ray spectrum below 1018 eV is of galactic origin, the knee could
reflect the fact that most cosmic accelerators in the galaxy have reached their maximum
energy. Some types of expanding supernova remnants, for example, are estimated not to
be able to accelerate protons above energies in the range of 1015 eV. Effects of propagation
and confinement in the galaxy [91] also need to be considered. The KASCADE-Grande
experiment [90] has reported observation of a second steepening of the spectrum near
8 × 1016 eV, with evidence that this structure is accompanied a transition to heavy
primaries.
Concerning the ankle, one possibility is that it is the result of a higher energy
population of particles overtaking a lower energy population, for example an extragalactic
flux beginning to dominate over the galactic flux (e.g. Ref. 89). Another possibility is
that the dip structure in the region of the ankle is due to γp → e+ + e− energy losses
of extragalactic protons on the 2.7 K cosmic microwave radiation (CMB) [93]. This dip
structure has been cited as a robust signature of both the protonic and extragalactic
nature of the highest energy cosmic rays [92]. If this interpretation is correct, then the
galactic cosmic rays do not contribute significantly to the flux above 1018 eV, consistent
with the maximum expected range of acceleration by supernova remnants.
The energy-dependence of the composition from the knee through the ankle is useful
in discriminating between these two viewpoints, since a heavy composition above 1018
eV is inconsistent with the formation of the ankle by pair production losses on the
CMB. The HiRes and Auger experiments, however, present very different interpretations
of data on the depth of shower maximum Xmax , a quantity that correlates strongly
with the interaction cross section of the primary particle. If these results are interpreted
using standard extrapolations of measured proton and nuclear cross sections, then the
HiRes data [94] is consistent with the ultrahigh-energy cosmic ray (UHECR) composition
getting lighter and containing only protons and helium above 1019 eV, while Auger [95,96]
sees a composition getting lighter up to 2 × 1018 eV and becoming heavier after
that, intermediate between protons and iron at 3 × 1019 eV. This may mean that the
extragalactic cosmic rays have a mixed composition at acceleration similar to the GeV
galactic cosmic rays. It is important to note that the measurements of Xmax may be
interpreted with equal validity in terms of a changing proton-air cross-section and no
change in composition.
If the cosmic ray flux at the highest energies is cosmological in origin, there should be a
rapid steepening of the spectrum (called the GZK feature) around 5 × 1019 eV, resulting
from the onset of inelastic interactions of UHE cosmic rays with the cosmic microwave

February 16, 2012 14:07

16 24. Cosmic rays

3
10

∆ E/E=20%

E 2.6 F(E) [GeV1.6 m-2 s-1 sr-1]

102

HiRes 1
HiRes 2
10
Telescope Array 2011
Auger 2011

1
18 19 20
10 10 10
E [eV]

Figure 24.9: Expanded view of the highest energy portion of the cosmic-ray
spectrum from data of HiRes 1&2 [101], the Telescope Array [103], and the Auger
Observatory [104]. The HiRes stereo spectrum [112] is consistent with the HiRes
1&2 monocular results. The differential cosmic ray flux is multiplied by E 2.6 . The
red arrow indicates the change in the plotted data for a systematic shift in the
energy scale of 20%.

background [97,98]. Photo-dissociation of heavy nuclei in the mixed composition

model [99] would have a similar effect. UHECR experiments have detected events of
energy above 1020 eV [89,100–102]. The AGASA experiment [100] did not observe
the expected GZK feature. The HiRes fluorescence experiment [101,112] has detected
evidence of the GZK supression, and the Auger observatory [102–104] has presented
spectra showing this supression based on surface detector measurements calibrated
against its fluorescence detector using events detected in hybrid mode, i.e. with both the
surface and the fluorescence detectors. Recent observations by the Telescope Array [103]
also exhibit this supression.
Figure 24.9 gives an expanded view of the high energy end of the spectrum, showing
only the more recent data. This figure shows the differential flux multiplied by E 2.6 .
The experiments are consistent in normalization if one takes quoted systematic errors in
the energy scales into account. The continued power law type of flux beyond the GZK
cutoff previously claimed by the AGASA experiment [100] is not supported by the HiRes,
Telescope Array, and Auger data.
One half of the energy that UHECR protons lose in photoproduction interactions that

February 16, 2012 14:07

 24. Cosmic rays 17

-7
10
All flavor limits(ν e : ν µ : ν τ = 1:1:1)
-8
10 IceCube40 (2011)

-9 Auger (2011)
10
E ν dN/dEν [m-2 s-1 sr-1] RICE (2011)
-10 ANITA-2 (2010)
10

10-11
WB

10-12

-13
10 ESS

10-14

-15
10

-16
10 16 17 18 19 20
10 10 10 10 10 1021
E ν [eV]

Figure 24.10: Differential limits on the flux of cosmogenic neutrinos set by

four neutrino experiments. The curves show the Waxman-Bahcall benchmark
flux (WB, [111]) and a representative midrange model for the expected flux of
cosmogenic neutrinos (ESS, [110]) . The expected flux is uncertain by over an order
of magnitude in either direction.

cause the GZK effects ends up in neutrinos [105]. Measuring this cosmogenic neutrino
flux above 1018 eV would help resolve the UHECR uncertainties mentioned above. The
magnitude of this flux depends strongly on the cosmic ray spectrum at acceleration, the
cosmic ray composition, and the cosmological evolution of the cosmic ray sources. In
the case that UHECR have mixed composition only the proton fraction would produce
cosmogenic neutrinos. Heavy nuclei propagation produces mostly ν̄e at lower energy from
neutron decay.
The expected rate of cosmogenic neutrinos is lower than current limits obtained by
IceCube [106], the Auger observatory [108], RICE [107], and ANITA-2 [109], which are
shown in Figure 24.10 together with a model for cosmogenic neutrino production [110] and
the Waxman-Bahcall benchmark flux of neutrinos produced in cosmic ray sources [111].
At production, the dominant component of neutrinos comes from π ± decays and has
flavor content νe : νµ : ντ = 1 : 2 : 0. After oscillations, the arriving cosmogenic neutrinos

February 16, 2012 14:07

18 24. Cosmic rays
are expected to be an equal mixture of all three flavors. The sensitivity of each experiment
depends on neutrino flavor. IceCube, RICE, and ANITA are sensitive to all three flavors,
and the sensitivity to different flavors is energy dependent. The limit of Auger is only for
ντ and ν̄τ which should be about 1/3 of the total neutrino flux after oscillations, so this
limit is plotted multiplied by a factor of three for comparison with the other limits and
with the theoretical estimates.
References:
1. M. Boezio et al., Astropart. Phys. 19, 583 (2003).
2. AMS Collaboration, Phys. Lett. B490, 27 (2000); Phys. Lett. B494, 193 (2000).
3. T. Sanuki et al., Astrophys. J. 545, 1135 (2000).
4. S. Haino et al., Phys. Lett. B594, 35 (2004).
5. H.S. Ahn et al., Astrophys. J. 707, 593 (2000).
6. J.J. Engelmann et al., Astron. & Astrophys. 233, 96 (1990).
7. D. Müller et al., Ap J, 374, 356 (1991).
8. M. Ave et al., Astrophys. J. 678, 262 (2008).
9. A.D. Panov et al., Bull Russian Acad of Science, Physics, 71, 494 (2007).
10. V.A. Derbina et al., Astrophys. J. 628, L41 (2005).
11. K. Asakimori et al. (JACEE Collaboration), Astrophys. J. 502, 278 (1998).
12. F. Aharonian et al. (HESS Collaboration), Phys. Rev. D75, 042004 (2007).
13. A.A. Abdo et al., Astrophys. J. 698, 2121 (2009).
14. R. Abbasi et al., Astrophys. J. 718, L194 (2010).
15. M. Amenomori et al., Astrophys. J. 711, 119 (2010).
16. A.W. Strong et al., Ann. Rev. Nucl. and Part. Sci. 57, 285 (2007).
17. M.A. DuVernois et al., Astrophys. J. 559, 296 (2000).
18. S. Torii et al., Astrophys. J. 559, 973 (2001).
19. M. Boezio et al., Astrophys. J. 532, 635 (2000).
20. K. Yoshida et al., Adv. Spa. Research, 42, 1670 (2008).
21. J. Chang et al. (ATIC Collaboration), Nature 456, 362 (2008).
22. F. Aharonian et al. (HESS Collaboration), Phys. Rev. Lett. 101, 261104 (2008)
and arXiv:0905.0105.
23. A.A. Abdo et al. (Fermi/LAT Collaboration), Phys. Rev. Lett. 102, 181101 (2009).
24. O. Adriani et al. (Pamela Collaboration), Nature 458, 607 (2009);
Phys. Rev. Lett. 102, 051101 (2009).
25. I.V. Moskalenko and A.W. Strong, Astrophys. J. 493, 694 (1998).
26. J. Nishimura et al., Adv. Space Research 19, 767 (1997).
27. A.S. Beach et al., Phys. Rev. Lett. 87, 271101 (2001).
28. A. Yamamoto et al., Adv. Space Research (2007) in press.
29. Y. Asaoka et al., Phys. Rev. Lett. 88, 51101 (2002).
30. K. Abe et al., arXiv:1201.2967v1 (2012).
31. H. Fuke et al., Phys. Rev. Lett. 95, 081101 (2005).
32. R. Bellotti et al., Phys. Rev. D53, 35 (1996).
33. R. Bellotti et al., Phys. Rev. D60, 052002 (1999).
34. M. Boezio et al., Phys. Rev. D62, 032007 (2000); M. Boezio et al., Phys. Rev.
D67, 072003 (2003).

February 16, 2012 14:07

 24. Cosmic rays 19
35. S. Coutu et al., Phys. Rev. D62, 032001 (2000).
36. S. Haino et al.. Phys. Lett. B594, 35 (2004).
37. T. Sanuki et al., Phys. Rev. D75, 043005 (2007).
38. T.K. Gaisser, Cosmic Rays and Particle Physics, Cambridge University Press
(1990).
39. P. Lipari, Astropart. Phys. 1, 195 (1993).
40. E. Mocchiutto et al., in Proc. 28th Int. Cosmic Ray Conf., Tsukuba, 1627 (2003).
41. M.P. De Pascale et al., J. Geophys. Res. 98, 3501 (1993).
42. P.K.F. Grieder, Cosmic Rays at Earth, Elsevier Science (2001).
43. J. Kremer et al., Phys. Rev. Lett. 83, 4241 (1999).
44. S. Haino et al. (BESS Collaboration), Phys. Lett. B594, 35 (2004).
45. P. Archard et al. (L3+C Collaboration), Phys. Lett. B598, 15 (2004).
46. O.C. Allkofer, K. Carstensen, and W.D. Dau, Phys. Lett. B36, 425 (1971).
47. B.C. Rastin, J. Phys. G10, 1609 (1984).
48. C.A. Ayre et al., J. Phys. G1, 584 (1975).
49. H. Jokisch et al., Phys. Rev. D19, 1368 (1979).
50. C.G.S. Costa, Astropart. Phys. 16, 193 (2001).
51. P. Adamson et al. (MINOS Collaboration), Phys. Rev. D76, 052003 (2007).
52. S. Hayakawa, Cosmic Ray Physics, Wiley, Interscience, New York (1969).
53. R.R. Daniel and S.A. Stephens, Revs. Geophysics & Space Sci. 12, 233 (1974).
54. K.P. Beuermann and G. Wibberenz, Can. J. Phys. 46, S1034 (1968).
55. I.S. Diggory et al., J. Phys. A7, 741 (1974).
56. D.E. Groom, N.V. Mokhov, and S.I. Striganov, “Muon stopping-power and range
tables,”Atomic Data and Nuclear Data Tables, 78, 183 (2001).
57. P. Lipari and T. Stanev, Phys. Rev. D44, 3543 (1991).
58. M. Crouch, in Proc. 20th Int. Cosmic Ray Conf., Moscow, 6, 165 (1987).
59. I.A. Belolaptikov et al., Astropart. Phys. 7, 263 (1997).
60. J. Babson et al., Phys. Rev. D42, 3613 (1990).
61. P. Desiati et al., in Proc. 28th Int. Cosmic Ray Conf., Tsukuba, 1373 (2003).
62. T. Pradier et al. (ANTARES Collaboration), arXiv:0805.2545 and 31st ICRC,
7-15 July 2009, L Ã ódź, Poland (paper #0340).
63. Yu.M. Andreev, V.I. Gurentzov, and I.M. Kogai, in Proc. 20th Int. Cosmic Ray
Conf., Moscow, 6, 200 (1987).
64. M. Aglietta et al. (LVD Collaboration), Astropart. Phys. 3, 311 (1995).
65. M. Ambrosio et al. (MACRO Collaboration), Phys. Rev. D52, 3793 (1995).
66. Ch. Berger et al. (Frejus Collaboration), Phys. Rev. D40, 2163 (1989).
67. C. Waltham et al., in Proc. 27th Int. Cosmic Ray Conf., Hamburg, 991 (2001).
68. Y. Ashie et al. (SuperKamiokande Collaboration), Phys. Rev. D71, 112005 (2005).
69. T. Futagami et al., Phys. Rev. Lett. 82, 5194 (1999).
70. Y. Fukuda et al., Phys. Rev. Lett. 81, 1562 (1998).
71. F. Reines et al., Phys. Rev. Lett. 15, 429 (1965).
72. M.M. Boliev et al., in Proc. 3rd Int. Workshop on Neutrino Telescopes (ed. Milla
Baldo Ceolin), 235 (1991).

February 16, 2012 14:07

20 24. Cosmic rays
73. M. Ambrosio et al., (MACRO) Phys. Lett. B434, 451 (1998). The number quoted
for MACRO is the average over 90% of the lower hemisphere, cos θ < −0.1; see
F. Ronga et al., hep-ex/9905025.
74. R. Becker-Szendy et al., Phys. Rev. Lett. 69, 1010 (1992);
Proc. 25th Int. Conf. High-Energy Physics, Singapore (eds. K.K. Phua and Y.
Yamaguchi, World Scientific), 662 1991.
75. S. Hatakeyama et al., Phys. Rev. Lett. 81, 2016 (1998).
76. Y. Fukuda et al., Phys. Rev. Lett. 82, 2644 (1999).
77. K. Greisen, Ann. Rev. Nucl. Sci. 10, 63 (1960).
78. S.P. Swordy et al., Astropart. Phys. 18, 129 (2002).
79. N.L. Grigorov et al., Yad. Fiz. 11, 1058 (1970) and Proc. 12th Int. Cosmic Ray
Conf., Hobart, 2, 206 (1971).
80. K. Asakimori et al., Proc. 23rd Int. Cosmic Ray Conf., Calgary, 2, 25 (1993);
Proc. 22nd Int. Cosmic Ray Conf., Dublin, 2, 57 and 97 (1991).
81. T.V. Danilova et al., Proc. 15th Int. Cosmic Ray Conf., Plovdiv, 8, 129 (1977).
82. Yu. A. Fomin et al., Proc. 22nd Int. Cosmic Ray Conf., Dublin, 2, 85 (1991).
83. M. Amenomori et al., Astrophys. J. 461, 408 (1996).
84. M. Nagano et al., J. Phys. G10, 1295 (1984).
85. F. Arqueros et al., Astron. & Astrophys. 359, 682 (2000).
86. M.A.K. Glasmacher et al., Astropart. Phys. 10, 291 (1999).
87. T. Antoni et al. (Kascade Collaboration), Astropart. Phys. 24, 1 (2005).
88. M. Amenomori et al., Proc. 30th Int. Cosmic Ray Conf., (Merida, Yucatan, 2007),
paper 0277.
89. D.J. Bird et al., Astrophys. J. 424, 491 (1994).
90. W.D. Apel et al.Phys. Rev. Lett. 107, 171104 (2011).
91. V.S. Ptuskin et al., Astron. & Astrophys. 268, 726 (1993).
92. V.S. Berezinsky and S.I. Grigor’eva, Astron. & Astrophys. 199, 1 (1988).
93. V. Berezinsky, A. Gazizov, and S. Grigorieva, Phys. Rev. D74, 043005 (2006).
94. R.U. Abbasi et al. (The HiRes Collaboration), Astrophys. J. 622, 910 (2005).
95. M. Unger et al. (Auger Collaboration), Proc. 30th Int. Cosmic Ray Conf., Merida,
Mexico, 2007 (arXiv:0706.1495).
96. J. Abraham et al. (Auger Collaboration), Proc. 31st Int. Cosmic Ray Conf., Lodz,
Poland, 2009; (arXiv:0906.2319).
97. K. Greisen, Phys. Rev. Lett. 16, 748 (1966).
98. G.T. Zatsepin and V.A. Kuz’min, Sov. Phys. JETP Lett. 4, 78 (1966).
99. D. Allard et al., Astron. & Astrophys. 443, L29 (2005).
100. M. Takeda et al. (The AGASA Collaboration), Astropart. Phys. 19, 447 (2003).
101. R. Abbasi et al. (HiRes Collaboration), Phys. Rev. Lett. 100, 101101 (2008).
102. J. Abraham et al. (Auger Collaboration), Phys. Rev. Lett. 101, 061101 (2008).
103. Y. Tsunesada et al.(Telescope Array Collaboration), in Proc. 32nd Int. Cosmic
Ray Conf., Bejing, China (arXiv:1111.2507v1).
104. P Abreu et al. (Auger Collaboration), in Proc. 32nd Int. Cosmic Ray Conf., Bejing,
China (arXiv:1107.4809).
105. V.S. Berezinsky and G.T. Zatsepin, Phys. Lett. B28, 423 (1969).

February 16, 2012 14:07

 24. Cosmic rays 21
106. R. Abbasi et al. (IceCube Collaboration) Phys. Rev. D83, 092003 (2011).
107. I. Kravchenko et al. (RICE Collaboration), Phys. Rev. D73, 082002 (2006);
I. Kravchenko et al. arXiv:1106.1164 (2011).
108. J. Abraham et al. (Auger Collaboration), Phys. Rev. Lett. 100, 211101 (2008);
P Abreu et al. (Auger Collaboration), in Proc. 32nd Int. Cosmic Ray Conf., Bejing,
China (arXiv:1107.4805).
109. P. Gorham et al. (ANITA Collaboration), Phys. Rev. D82, 022004 (2010);
P. Gorham et al. (ANITA Collaboration)(arXiv:1011.5004).
110. R. Engel, D. Seckel, and T. Stanev, Phys. Rev. D64, 09310 (2001).
111. E. Waxman and J. Bahcall, Phys. Rev. D59, 023002 (1999).
112. R.U. Abbasi et al. (HiRes Collaboration), Astropart. Phys. 32, 53 (2009).

February 16, 2012 14:07