Prepared for submission to JCAP

Cosmic absorption of ultra high energy

arXiv:1503.07749v1 [astro-ph.HE] 26 Mar 2015

particles

R. Ruffinia,b,c G.V. Vereshchagin,a,b,1 S.-S. Xuea,b

a ICRANet Piazzale della Repubblica, 10 -65122, Pescara
b ICRA and Department of Physics, University of Rome “Sapienza”, P.le A. Moro 5, 00185
Rome, Italy
c ICRANet-Rio, Centro Brasileiro de Pesquisas Fisicas, Rua Dr. Xavier Sigaud 150, Rio de

Janeiro, 22290180, Brazil

E-mail: ruffi[email protected], [email protected], [email protected]

Abstract. This paper summarizes the limits on propagation of ultra high energy particles in
the Universe, set up by their interactions with cosmic background of photons and neutrinos.
By taking into account cosmic evolution of these backgrounds and considering appropriate
interactions we derive the mean free path for ultra high energy photons, protons and neutri-
nos. For photons the relevant processes are the Breit-Wheeler process as well as the double
pair production process. For protons the relevant reactions are the photopion production and
the Bethe-Heitler process. We discuss the interplay between the energy loss length and mean
free path for the Bethe-Heitler process. Neutrino opacity is determined by its scattering off
the cosmic background neutrino. We compute for the first time the high energy neutrino
horizon as a function of its energy.

1
Corresponding author.
Contents

1 Introduction 1

2 Processes 2
2.1 Processes involving photons 2
2.2 Processes involving protons 3
2.3 Processes involving neutrinos 4

3 The optical depth and the mean free path 5

4 The mean energy loss distance 6

5 Results 7
5.1 Photons 7
5.2 Protons 10
5.3 Neutrinos 12

6 Conclusions 13

1 Introduction

Observation of ultra high energy (UHE) particles, such as photons, ions and neutrinos, pro-
vides the crucial information on astrophysical systems as well as mechanisms of charged
particle acceleration in these systems. Such information cannot be obtained from the study
of low energy emission, although it is much easier to detect.
Propagation of UHE particles on cosmological distances involves interaction with other
particles, as well as with electromagnetic fields, in the case of charged particles [1]. One of the
most important reservoir of photons is the cosmic microwave background (CMB). Interaction
with CMB imposes strong limits on propagation of UHE photons, protons, and nuclei. Ex-
tragalactic background light (EBL), being the accumulated radiation in the Universe due to
stars and active galactic nuclei, represents additional background of photons [2], which limits
propagation of high energy photons. Yet another important background is cosmic neutrino
background (CνB), which places a tight limit on the propagation of UHE neutrinos.
In this work, we review stringent limits on propagation of UHE particles, namely pho-
tons, protons and neutrinos, in the Universe due to their interactions with cosmic background
of photons and neutrinos. We pay the particular attention to accounting for the cosmic evo-
lution of CMB and CνB fields, being important at high redshifts. We discuss relation with
previous results, as well as implications of new results, obtained in this work.
First in Sec. 2 we discuss most important processes, responsible for interaction of UHE
particles with cosmic backgrounds, as well as the corresponding cross-sections. Most of these
processes are discussed in detail in [3]. Then in Sec. 3 we introduce the method used to
compute the mean free path of UHE particles, which takes into account cosmological redshift
of particle energy, as well as temperature evolution of the CMB and CνB. In Sec. 4 the
definition of the mean energy loss distance is given. We present and discuss results in Sec. 5.
Conclusions follow.

–1–
2 Processes
2.1 Processes involving photons
UHE photons are likely produced in sources of UHE cosmic rays. The most important
process, responsible for intergalactic absorption of high-energy γ-rays is the Breit-Wheeler
process [4] for the photon-photon pair production

γ1 + γ2 −→ e+ + e− (2.1)

It was first discussed by Nikishov [5] back in 1961 and then, after the discovery of CMB, by
Gould and Schreder [6].
Breit and Wheeler [4] studied collision process (2.1) of two photons with energies E
and E in the laboratory frame, producing electron and positron pair. They found the total
cross-section
π α~ 2
    
2 2 4 1+β
σγγ = (1 − β ) 2β(β − 2) + (3 − β ) ln , (2.2)
2 me c 1−β
where r
1 EE
β= 1− , x= , (2.3)
x (me c2 )2
~ is Planck’s constant, me is electron mass, c is the speed of light and α is the fine structure
constant. The necessary kinematic condition in order for the process (2.1) to take place
is that the energy of two colliding photons is larger than the energetic threshold 2me c2 ,
i.e., x ≥ 1. Due to this kinematic condition the function (2.2) has a low energy cut-off at
x = 1. The cross-section has a maximum at x ≃ 2, with σγγ max ≃ σ /4, where σ is Thomson
T T
cross-section. At higher energies it decreases as 1/x.
A simple estimate of the mean free path for the Breit-Wheeler absorption of high energy
photon can be given as follows. Considering the actual CMB photon density nCM B ≃ 411
cm−3 and taking σT /4 for the cross-section of the interaction, the mean free path is λBW =
(σT nCM B /4)−1 ≃ 4.8 kpc. One can refer to this distance as to a horizon, namely the
maximal distance to the source for which the particle with the given energy can still be
detected on Earth. However, owing to the energy dependence of the cross-section, and
cosmic evolution of the CMB photon field the actual mean free path strongly depends on
energy. The characteristic energy of UHE photons interacting with CMB, having temperature
today T0 ≈ 2.725 K, is given by EBW = (me c2 )/kT0 ≃ 1.11 PeV. At lower energies, in the
TeV range, photons interact by the Breit-Wheeler process with the EBL [6–8]. Hence the
observation of TeV radiation from distant (d > 100 Mpc) extragalactic objects provides
important constraints on the EBL [9–11].
At much higher energies the double pair production process

γ1 + γ2 → e+ + e− + e+ + e− (2.4)

dominates, see [12, 13]. In this high-energy limit it has nearly a constant cross-section, see
e.g. [3]
α2 α~ 2
 
σdpp = [175ζ(3) − 38] ∼ 6.5µb. (2.5)
36π me c
Clearly, this process has a threshold with the sum of energies of photons which must exceed
4me c2 . It imposes a limit for UHE photons propagation λdpp = (σdpp nCM B )−1 ≃ 121 Mpc.

–2–
2.2 Processes involving protons
Charged UHE particles, such as protons and nuclei, are assumed to originate from extra-
galactic sources, which work as ”cosmic accelerators” [1]. Such particles interact with the
CMB photons as well. In fact, the famous Greisen–Zatsepin–Kuzmin (GZK) limit [14, 15]
was established by considering that this UHE particle interacts with the CMB photons via
the pion photoproducton process
 
p
p + γ −→ + π. (2.6)
n

and lose its initial energy. Due to the fact that at this process the proton loses more than
half of its energy [16], such interaction imposes a strong cut-off on energies of UHE cosmic
rays. The cut-off energy is easy to estimate. Recall that the characteristic energy in the
Breit-Wheeler process (2.1) is EBW = (me c2 )2 /kT0 . When the photopion process (2.6) is
concerned, the electron mass is exchanged with the pion mass, and an additional factor 4
comes from the reference frame transformation, giving Epγ = 4(mπ c2 )2 /kT0 ≃ 3×105 EBW =
3.33 × 1020 eV. More careful evaluation of the energy by comparing energy losses due to
photopion (2.6) and photoproduction of pair (2.8) processes (see below) gives the value
Epγ = 5 × 1019 eV [17]. The cross-section of the photopion process in high energy limit is
constant with the value [16]
σpγ ≃ 120µb. (2.7)
The mean free path due to this process for energies E > Epγ is λpγ = (σpγ nCM B )−1 ≃ 6
Mpc.
Another process relevant for interaction of UHE particles with CMB is the photopro-
duction of electron-positron pair on a nucleus, or Bethe-Heitler process [18]. In the case of
proton, which is the only one considered in this work, this process is

p + γ −→ p + e+ + e− . (2.8)

It has the characteristic energy EBH = me mp c4 /(2kT0 ) ≃ 1.0 × 1018 eV. This process has
a threshold with photon energy in the proton rest frame E ′ > 2me c2 . We use for its cross-
section in the proton rest frame the expressions given in [19], namely near the threshold with
2 ≤ ǫ′ ≤ 4

α~ 2 ǫ′ − 2 3
     
thr ′ 2π 1 23 2 37 3 61 4
σBH (ǫ ) ≃ α 1+ η+ η + η + η , (2.9)
3 me c ǫ′ 2 40 120 192

where ǫ′ = E ′ /(me c2 ) is photon energy in the proton rest frame and η = (ǫ′ − 2) / (ǫ′ + 2).
At higher energies ǫ′ > 4 the cross-section is
 (
α~ 2 28
 2 
π2 π2
 
he ′ 218 2 7 2 3 2
σBH (ǫ ) ≃ α δ− + ′ 6δ − + δ − δ − δ + 2ζ(3) +
me c 9 27 ǫ 2 3 3 6
(2.10)
 4    6  )
2 3 1 2 29 77
− ′ δ+ − ′ δ− ,
ǫ 16 8 ǫ 9 × 256 27 × 512

where δ = log(2ǫ′ ). Expression (2.10) is logarithmically increasing at high energies, so we can

take a characteristic value obtained by Bethe and Heitler σBH ≃ (28/9)α [(α~)/(me c)]2 in

–3–
order to estimate the mean free path of UHE protons, which gives λBH = (σBH nCM B )−1 ≃
437 kpc.
It is important to note that unlike the Breit-Wheeler process, leading to annihilation of
UHE photons, or the pion photoproducton, where single interaction alters the energy of the
UHE proton, the single Bethe-Heitler interaction does not change the proton energy signifi-
cantly. Therefore, unlike all previous processes, the mean free path λBH does not correspond
to a horizon. Another quantity is used for this purpose, namely the mean energy loss dis-
tance, defined as λBH ∼ [dE/(Ecdt)]−1 , where E is the proton energy, which corresponds to
the distance on which the energy of the UHE proton is reduced by a factor e due to numerous
interactions with background photons [16, 17, 20]. However, it should be emphasized that
single Bethe-Heitler interaction deflects the UHE proton by a small angle. This effect is
discussed in detail below.

2.3 Processes involving neutrinos

UHE neutrinos can be produced either in astrophysical sources, or in some exotic new physics
scenarios [21]. Below we compute the horizon due to interaction of UHE neutrinos with cosmic
neutrino background (CνB). Following [22] we assume CνB neutrinos are in their mass states.
The cross-section is composed of two parts [22]. The resonant neutrino annihilation occurs
in the s-channel:
ν + ν̄ −→ Z 0 −→ f + f¯, (2.11)
where bar denotes antiparticle, f is a fermion. It has a typical Breit-Wigner shape and is
given in the analytic form in [23]. We take the small momentum expansion of the cross-section
given by eq. (23) in [23] as
√
R
√ mν MZ2 ξE
σν ν̄ ≃ 4 2GF GeV−2 , (2.12)
(MZ2 − 2Emν )2 + 4E 2 m2ν ξ

where GF = 1.16637× 10−5 GeV−2 is the Fermi’s coupling constant, ξ = (Γ/MZ )2 , Γ = 2.495
GeV is the width of Z 0 resonance and MZ = 91.1876 GeV is the mass of Z 0 boson, mν is
neutrino mass, E is energy of UHE neutrino in laboratory frame. Clearly, the position of the
resonance scales inversely proportional to the neutrino mass. Throughout this paper we use
the reference value mν = 0.08 eV/c2 , corresponding to the recent cosmological bound from
the Planck mission [24], which gives characteristic energy Er = MZ2 c2 /2mν ≃ 5.2 × 1022 eV.
The amplitude of the resonance does not depend on neutrino mass, and is given by
√
σνRν̄max = 2 2GF MZ /Γ ≃ 0.471µb. (2.13)

The second contribution is the non-resonant cross-section, which is adopted here in the
form
σνheν̄
σνNν̄R = , (2.14)
1 + (E/Er )−1
where σνheν̄ ≃ 8.3 × 10−4 µb.
We assume that neutrino are non-relativistic even at sufficiently high redshift, which is
a good approximation for z < 102 for mν = 0.08 eV/c2 . Effects of non-zero momentum on
the neutrino annihilation cross-section are studied in [22, 23]. Using the number density of
relic neutrinos nCνB ≃ 112 cm−3 and the non-resonant cross-section in the high energy limit
one can estimate the horizon for UHE neutrinos at highest energies. Solving the Friedmann
equation (see next section) one finds for the redshift zν ≃ 84.

–4–
3 The optical depth and the mean free path

In this section we compute the optical depth for the propagation of UHE particles in the
Universe. Imposing the condition that it equals unity we determine the corresponding mean
free path. It should be noted that simple estimates, made in the literature, as well as in
previous section, do not take into account evolution of CMB and CνB fields with time. The
simplest way to account for cosmic redshift is to compare this estimate of the mean free path
to the expansion scale c/H0 , where H0 is the present day Hubble parameter, see e.g. [17, 20].
In what follows we describe more rigorous way to take into account both redshift of particle
energy as well as the evolution of CMB and CνB fields with redshift.
The optical depth along the particle world line L is defined as
Z
τ= σjµ dxµ , (3.1)
L

where σ is the cross-section of a given process, j µ is the 4-current of particles, on which the
UHE particle scatters, and dxµ is the element of the UHE particle world line. We assume
the Universe is homogeneous and isotropic, and the background particles are either CMB
photons or CνB neutrinos. Both have thermal distribution functions, given by
1
f (E/kT ) = , (3.2)
e(E−µ)/kT ±1
where k is the Boltzmann constant, T is the CMB or CνB temperature, the sign ”−” is
for photons while the sign ”+” is for neutrinos, E and µ are the energy and the chemical
potential of background particles (for photons µ = 0). Then the optical depth (3.1) is
Z 0 Z ∞
gs
τ (E, t) = 2 3 3 cdt′ E 2 dEf (E)σ(E, E, t′ ), (3.3)
2π ~ c t Etr

where Etr is threshold energy in a given process, gs = 2 is the number of helicity states
for both protons and neutrinos. Here we assumed that UHE particles move along light-like
geodesics. The integral over time can be transformed into the integral over redshift by means
of the Friedmann equation. The latter for the flat Universe reads

1 da 2 8πG
 
= ρ, (3.4)
a dt 3
where a is the scale factor, ρ is energy density of the Universe, G is Newton’s constant. From
this equation, the definition of cosmological redshift, as well as the definition of the density
parameters
ρi 3H02
a0 /a = 1 + z, Ωi = , ρc = , (3.5)
ρc 8πG
where H0 and a0 are present time Hubble parameter and scale factor, respectively, we have
Z 0 Z z
c dz ′
cdt′ −→ , (3.6)
t H0 0 (1 + z ′ ) H(z ′ )

where H0 is the Hubble parameter and the function H(z) is given by

H(z) = [Ωr (1 + z)4 + ΩM (1 + z)3 + ΩΛ ]1/2 , (3.7)

–5–
and Ωr , ΩM and ΩΛ are present densities of radiation, matter and dark energy, respectively.
Then the expression (3.3) can be written as follows
Z z Z ∞
1 c dz ′
τ (E, z) = 2 3 3 E 2 dEf (E)σ(E, E, z ′ ). (3.8)
π ~ c H0 0 (1 + z ′ ) H(z ′ ) Etr

Cosmic expansion results in the energy and temperature dependence on redshift

T = (1 + z)T0 , E = (1 + z)E0 , E = (1 + z)E0 , (3.9)

where temperature T0,γ ≃ 2.725 K for photons, T0,ν = (4/11)1/3 ≃ 1.95 K for neutrinos and
energies E0 , E0 are measured at the present time.
The second integral in (3.8) can be simplified, provided two conditions are fulfilled: a)
the cross-section does not depend on the energy of background particle and b) there is no
threshold in the given process (Etr = 0). In this case one has
Z ∞
1
E 2 dEf (E)σ(E, z) = σ(E, z)n(z) = σ(E, z)n0 (1 + z)3 , (3.10)
π 2 ~3 c3 0
where n0 is present number density and it stands for either

~ −3 kT0 3
   
2ζ (3)
n0,γ ≈ ≃ 411 cm−3
π2 mc me c2

for photons, or n0,ν = 3/4 (T0,ν /T0,γ )3 ≃ 113 cm−3 for neutrinos. Then eq. (3.3) becomes

c z
σ(E, z ′ ) (1 + z ′ )2 dz ′
Z
τ (E, z) = n0 . (3.11)
H0 0 H(z ′ )

When the cross-section is just a constant, the integral (3.11) can be readily performed.
Assuming Ωr ≃ 9.2 × 10−5 , ΩM ≃ 0.315, ΩΛ ≃ 0.685 and H0 = 67.3 km/s/Mpc [24] in the
matter dominated epoch we have

(1 + z ′ )2 dz ′
Z z 
1.045z, z ≪ 1,
≃ 3/2 , z
0 [Ω M (1 + z ′ )3 + Ω ]1/2
Λ 1.006z ≫ 1.

The mean free path is defined by the condition τ (E0 , z) = 1. For the constant cross-
section σ at low redshift z ≪ 1 we get the traditional definition λ = (σn)−1 used above. For
high redshift z ≫ 1 one can define the redshift, corresponding to the mean free path as
 −2/3  −2/3
n0 σc n0 σ
zλ = ≃ 8.9 −8
. (3.12)
H0 n0,γ 10 σT

Using this equation we obtain for UHE neutrinos with highest energies zλ ≃ 84.

4 The mean energy loss distance

When UHE particle annihilates in a given process, such as in the case of Breit-Wheeler one
(2.1), the mean free path correpsonds to the horizon defined above.
Another possibility is that the particle is not annihilated in a given process, but scat-
tered, such as in the case of proton producing the pion (2.6). When the energy loss in single

–6–
scattering corresponds to a large fraction of UHE particle energy, the situation is similar to
the case of annihilation. However, UHE particle may lose only a small fraction of its energy,
as in the case of Bethe-Heitler process (2.8). Here another relevant quantity corresponds the
particle horizon defined above is the mean energy loss distance λ̃. We define it following [25]
as  
−1 1 dE
λ̃ = . (4.1)
E cdt
Then we evaluate the quantity
0 z
cdt c dz ′
Z Z
τ̃ = = . (4.2)
t λ̃ H 0 0 λ̃ (1 + z ′ ) H(z ′ )
It is computed below for the Bethe-Heitler process.

5 Results
Now we apply the method developed in the previous section to the computation of the mean
free path for UHE photons, protons and neutrinos, as well as the mean energy loss distance
for protons interacting via the Bethe-Heitler process.

5.1 Photons
First, we consider cosmic limits on propagation of UHE photons. In the Breit-Wheeler
process (2.1) the cross-section depends on both energies through the definition (2.3). When
one considers all possible orientations of CMB photons additional averaging over their angular
distribution has to be performed [5, 6]. The resulting averaged cross section differs from eq.
(2.2). The useful approximations for this quanity can be found e.g. in [6, 26, 27]. We use
the accurate expression given by eq. (3.23) in Ref. [1]:
 
3 1 1 1 1
σ̄γγ (x) = σT Σ (x) , Σ (x) = 2 x + log x − + × (5.1)
2 x 2 6 2x
r #
√ √


4 1 1
× log x + x − 1 − x + − 1− .
9 9x x

We change our variables in eq. (3.8) and get

A z
Z ∞
1 dz ′ x2 dx
Z
τγγ (E0 , z) = 3 ′ 4 ′
Σ (x) , (5.2)
y0 0 (1 + z ) H (z ) 1 exp(x/y) − 1
where −1  3
4α2 c

~ kT0
A= ≈ 2.37 × 106 , (5.3)
π H0 mc me c2
and
E0 kT0
y = y0 (1 + z)2 ; y0 = , (5.4)
me c2 me c2

2
and y0 is the energy E0 in units of the critical energy EBW = me c2 /kT0 ≃ 1.11 × 1015
eV. The intergal over energy can be evaluated numerically and we find a reasonable fit
 
1.1
F1 (y) = 0.839 y 2.1 + 2 × 10−8 y 2.8 exp −


. (5.5)
y

–7–
Then eq. (5.2) becomes
z
A 1 dz ′
Z
2
 
τγγ (y0 , z) = F1 y 0 (1 + z) . (5.6)
y03 0 (1 + z ′ )4 H (z ′ )

This integral is also evaluated numerically.

In the low energy E ≪ EBW and high redshift z ≫ 1 limit the integral (5.6) can be
evaluated analytically. The redshift corresponding to the mean free path in this limit is
 −1/2
E
zλ,BW ≃ 0.21 . (5.7)
EBW

This result is known as the Fazio-Stecker relation [28], see their eq. (9).
In addition to the Breit-Wheeler process (2.1), following [13] we consider also the double
pair production process (2.4) with the cross-section defined in (2.5). This process is relevant
for the highest energies. The optical depth for this process is

B z
Z ∞
1 dz ′ x2 dx
Z
dpp
τγγ (y0 , z) = 3 σ dpp = (5.8)
y0 0 (1 + z ′ )4 H (z ′ ) 2 exp(x/y) − 1
B z 1 dz ′
Z
F2 y0 (1 + z ′ )2 ,
 
= 3 ′ 4 ′
y0 0 (1 + z ) H (z )

where
   
8 2
F2 (y) = − 4iπy − 4y log exp −1 − (5.9)
3 y
     
2 2 3 2
− 4y PolyLog 2, exp + 2y PolyLog 3, exp .
y y

and B ≈ 15.3.
dpp
The condition τ (y0 , z) = τγγ (y0 , z) + τγγ (y0 , z) = 1 in eqs. (5.2) and (5.8) determines
the mean free path of UHE photons. This mean free path is shown in Fig. 1 in megaparsecs
and in Fig. 2 in cosmological redshift by the solid curve. The region above the solid curve is
opaque for high energy photons. In addition, thick black dashed line shows the boundary of
transparency for EBL, according to the baseline model of Inoue et al. [29], while dotted line
in Fig. 1 shows the cosmological horizon z = ∞.
For the distance smaller than a critical value of about dc = 6.8 kpc, the CMB is
transparent to high-energy photons with arbitrary energy. For larger distances there are
two branches of solutions for the condition τγγ (y0 , z) = 1, respectively corresponding to
the different energy-dependence of the average cross-section (5.1). This average cross-section
σ̄γγ (x) increases with the center of mass energy x from the energy threshold x = 1 to x ≃ 3.5,
and decreases from x ≃ 3.5 to x → ∞. The energy of the UHE photon corresponding to
the critical distance dc is about 1.11 PeV, which separates two branches of the solution. The
double pair production process (2.4) is relevant for the highest energies, as expected. Photons
with energies above 10 PeV are absorbed by the double pair production if they are emitted
at redshift above z ≃ 0.03 (distance about 120 Mpc).
For comparison we show by the dashed curve also the mean free path computed at
z = 0, namely neglecting cosmological expansion, see e.g. [13].

–8–
 d, Mpc
104
103
102
101
1
10-1
10-2
E, eV
1010 1012 1014 1016 1018 1020 1022
Figure 1. The mean free path, measured in megaparsecs as a function of energy E of UHE particles,
measured in electronvolts. In the region above the curves the optical depth is larger than unity.
Thin black curve shows the mean free path of UHE photons. Dashed black curve shows the photon
mean free path computed without accounting for cosmological evolution (imposing z = 0). Thick
blue dashed curve shows the boundary of transparency for extragalactic background light (EBL),
according to the baseline model of Inoue et al. [29]. Thick blue curve shows the mean free path
of UHE protons (GZK limit). Blue dashed (dotted-dashed) curve shows the mean free path (mean
energy loss distance) for UHE protons due to Bethe-Heitler process. Dotted horizontal line shows
cosmological horizon.

102

10-2

10-4

10-6 E, eV
1010 1012 1014 1016 1018 1020 1022 1024
Figure 2. The same as in Fig. 1 for the distance measured in cosmological redshift. Red thick curve
shows the mean free path of UHE neutrinos.

We also show by the dashdotted thick curve the mean free path for the photon-photon

–9–
scattering which follows from the Euler-Heisenberg lagrangian, see e.g. [3, 30]. We will
discuss this process in a separate publication [31].
Finally, the black dotted thick curve shows the horizon of photons with energies above
20 GeV and below 100 TeV, which is determined by their interaction with the EBL. The
latest EBL model [29] is used. It is clear, that the contribution of CMB photons gives the
absolute upper limit on the mean free path. In the energy range between 1 GeV and 20 GeV
the propagation of high energy photons is limited only by the CMB radiation.

5.2 Protons

Second, we consider the propagation of UHE protons, accelerated in a source located at a

cosmological distance from Earth. First, considering the photopion process (2.6) we use the
method developed in the previous section and compute the GZK limit [14, 15]. This limit
applies to protons and other charged particles, leading to the existence of a cutoff in the
observed spectrum of (UHE) cosmic rays at about 1020 eV. For the photopion process (2.6)
one can use the simple expression (3.11) with the constant cross-section (2.7). However, we
compute the optical depth in the same way as in the case of double pair production, using
eq. (5.8) with different value of the constant B ′ ≈ 253.
The mean free path due to photopion process is shown by the blue thick curve in Fig.
1 in megaparsecs and in Fig. 2 in cosmological redshift. From Fig. 1 it appears that for
energies well below Epγ ≃ 3.3 × 1020 eV the GZK limit approaches the cosmological horizon.
Instead, from Fig. 2 it follows that the mean free path measured in redshift, below the energy
Epγ , increases with decreasing energy as a power law, which is a consequence of eq. (3.12).
Similarly to the Breit-Wheeler case, the integral (5.8) is evaluated analytically in the
low energy E ≪ Epγ and high redshift z ≫ 1 limit, with the result

 −1/2
E
zλ,GZK ≃ 0.57 . (5.10)
Epγ

We also evaluate the mean free path due to the Bethe-Heitler process (2.8). Since cross-
sections (2.9) and (2.10) are given in the proton rest frame, one has to transform photon
energy to this reference frame using

EE
E ′ = 2ΓE =2 , (5.11)
mp c2

where the primed quantity corresponds to the proton rest frame, while unprimed quantities
to laboratory reference frame. Then it is convenient to make use of the same type of variable
change as before for the Breit-Wheeler process1 , with a difference that instead of electron
mass squared a product of electron and proton masses arises, namely

EE E0 kT0 E0
x̄ = 2 , ȳ0 = 2 = . (5.12)
me mp c4 2
mp c me c2 EBH

1
We assume that UHE protons collide with the with CMB photons head on. More accurate calculation
with average over angular distribution of the CMB photons does not change qualitatively our results.

– 10 –
The optical depth is computed in the laboratory frame as follows2

~ −3 kT0 3
   
1 c
τpγ (ȳ0 , z) = 2 × (5.13)
π H0 m c me c2
Z z Z ∞
1 1 dz ′ x̄2 dx̄
× 3 σBH (x).
ȳ0 0 (1 + z ′ )4 H (z ′ ) 2 exp(x̄/ȳ) − 1
The intergal over energy can be evaluated numerically and we find a reasonable fit
 
86.15 exp − ȳ2
F3 (ȳ) = 3ȳ−3.47 . (5.14)
10 + ȳ −3
Then eq. (5.2) becomes
z
C 1 dz ′
Z
2
 
τpγ (ȳ0 , z) = 3 F3 ȳ 0 (1 + z) , (5.15)
ȳ0 0 (1 + z ′ ) H (z ′ )
where −1  3
2α3 c

~ kT0
C= ≈ 2863. (5.16)
3π H0 mc me c2
The mean free path for protons interacting via the Bethe-Heitler process is shown by blue
dashed curve in Fig. 1 in megaparsecs and in Fig. 2 in cosmological redshift. The integral
(5.15) is evaluated analytically in the low energy E ≪ EBH and high redshift z ≫ 1 limit,
with the result  −1/2
E
zλ,BH ≃ 0.43 . (5.17)
EBH
As discussed before, at energies above 1018 eV the mean free path is relatively small,
about a few hundred kiloparsecs, and it quickly decreases with increasing energy. This is in
contrast with the large mean energy loss path, which is above 1 Gpc at energies 1018 − 1020
eV, see e.g. [17, 20]. It means, that before UHE proton starts to lose its energy, it is scattered
many hundred times [16]. At each of this scattering the proton recoils, being deflected by a
small angle measured in the laboratory reference frame. From the analysis of the cross-section
as the function of recoil angle [32], see also [33] in the proton rest frame, it follows that in
the high energy limit the photon recoils in the plane, orthogonal to the incident photon. It
implies that each scattering produces a
deflection of the UHE proton in the laboratory frame
by angle ∼ 1/γ, where γ = E/ mp c2 is proton Lorentz factor. The number of scatterings
is given approximately by τ .
One can compute the mean energy loss distance defined in (4.1) and then evaluate the
quantity (4.2) using eq. (18)-(20) in [25] to obtain
D z
Z ∞
dz ′ dx̄
Z
τ̃ = 3 φ (x̄) , (5.18)
y0 0 (1 + z ′ )4 H(z ′ ) 2 exp( x̄ȳ ) − 1

where the function φ (x̄) is given in eq. (16) in [25]

h i
φ (x̄) = x̄ −86.07 + 50.95 log x̄ − 14.45 (log x̄)2 + 2.667 (log x̄)3 (5.19)
2
Note that the integral over energy is not transformed to the proton rest frame, as done e.g. in [25].
Instead, only a change of variables is performed in the integral (5.13).

– 11 –
 Τ
106

105

104

E, eV
1016 1018 1020
Figure 3. The optical depth at the mean energy distance for the Bethe-Heitler process.

and
 −1  3
2 me c ~ kT0
D = 2 α3 . (5.20)
π mp H 0 mc me c2

From the condition τ̃ = 1 we determine the mean energy loss distance λ̃. This distance is
shown by blue dash-dotted curve in Fig. 1 in megaparsecs and in Fig. 2 in cosmological
redshift.
We evaluate the optical depth (5.13) at the redshift, corresponding to λ̃ for energies in
the range between 1015 and 1020 eV and find it in the range 104 to 105 , see Fig. 3. Since the
deflection at each interaction is small, the average number of interactions is proportional to
the optical depth. The average deflection angle is then
p
τ (E)
δ∼ . (5.21)
γ

We find an average deflection angle of UHE protons as a function of proton energy for sources
located at the mean energy loss distance and show it in Fig. 4. At energies E = 1016 eV, this
angle is about δ ∼ 15′′ and it decreases down to δ ∼ 2.4 mas for E = 1019 eV. This latter
scale is larger than the angular size of distant blazars [34]. Such deflection, although small
compared to deflection in galactic magnetic field [35], inevitably leads to dimming of point
sources of UHE protons. It makes also more difficult to detect distant sources.

5.3 Neutrinos

Third, we consider the propagation of UHE neutrinos. Such neutrinos can be produced in
the source of UHE cosmic rays in decay of secondary pions π + −→ µ+ + νµ or secondary
neutrons n −→ p + e− + ν̄e [16]. UHE neutrino can be produced also in some extensions of
the standard model of particle physics [21]. Such UHE neutrino interacts with the CνB via
the process (2.11).

– 12 –
 ∆,arcsec
102
101
1
10-1
10-2
10-3
E, eV
1016 1017 1018 1019 1020
Figure 4. The average deflection angle of UHE protons as function of proton energy for sources
located at the mean energy loss distance for the Bethe-Heitler process.

The cross-section of this process has a resonance, and it approaches a constant for
highest energies. We compute the optical depth, which instead of eq. (3.8) is given by
z ∞
1 c dz ′
Z Z q
τνν (E, z) = 2 3 3 E E 2 − (mν c2 )2 dEf (E)σ(E, z) ≈ (5.22)
π ~ c H0 0 (1 + z ′ ) H(z ′ ) Etr
1 c z
(1 + z ′ )2 dz ′
Z
2 3 3
n0,ν σ(E(1 + z′)),
π ~ c H0 0 H(z ′ )

using the cross-sections given in the laboratory reference frame by eqs. (2.12) and (2.14). The
mean free path for neutrinos, measured in cosmological redshift, is shown in Fig. 2 by the
thick red curve. Since the characteristic redshifts are high, this curve practically coincides
with the horizon, when measured in megaparsecs, so we do not show it in Fig. 1. It is
clear that the Breit-Wigner resonance in the cross-section decreases the mean free path in a
wide range of energies. The lowest redshift for E ≃ Er at which the Universe is transparent
for UHE neutrinos is zmin ≃ 30. The resonance produces a dip around Er /(1 + zmin ) ≃
1.7 × 1021 (mν /0.08 eV)−1 eV, where zmin ≃ 30. Additional broadening of the resonance,
due to thermal effect, is discussed in detail in [22]. At higher energies the corresponding
redshift is z ≃ 87.
Similarly to the previous cases, in the low energy E ≪ Epγ and high redshift z ≫ 1
limit, we find
 −2/5
E
zλ,ν ≃ 14 . (5.23)
Er

6 Conclusions

We reviewed cosmic limits on propagation of ultra high energy particles such as photons,
protons and neutrinos, set up by their interactions with the cosmic background of photons
and neutrinos. In doing so we take into account explicitly cosmic evolution of both cosmic
backgrounds, and redshift of UHE particle energy. This is in contrast with majority of the

– 13 –
literature, where corresponding mean free paths are found at present epoch, neglecting cosmic
expansion. A number of new results were obtained, in particular:

• for UHE photons the contribution of CMB photons gives the absolute upper limit on
the mean free path. At high redshift, where other radiation backgrounds, such as EBL
are absent, the CMB radiation limits the propagation of UHE photons at energies above
GeV.

• for UHE protons the mean free path due to Bethe-Heitler process appears to be much
shorter than the mean energy loss distance. This results in multiple deflections suffered
by UHE protons, before they start to lose energy in the energy range 1016 − 1020 eV.
Such deflections result in dimming of point sources of UHE protons, which makes it
more difficult to detect them.

• for UHE neutrinos for the first time we compute the horizon as a function of redshift.
We found that the Universe is transparent of UHE neutrinos at redshifts z < 30, near
the Breit-Wigner resonance at Er ≃ 5.2×1022 (mν /0.08 eV)−1 eV, and it is transparent
at redshifts z < 87 at higher energies.

• Remarkably, in the low energy and high redshift limit, the Fazio-Stecker relation [28]
holds for all processes with exception of neutrinos, and it is given by a universal ex-
 −1/2
pression zλ ≃ O (1) EEthr , where Ethr is the characteristic (e.g. threshold) energy
for a given process. In the case of neutrinos similar power law exists zλ ∝ E −2/5 .

References
[1] F. A. Aharonian, Very High Energy Cosmic Gamma Radiation. World Scientific, 2003.
[2] M. G. Hauser and E. Dwek, The Cosmic Infrared Background: Measurements and Implications,
ARA&A 39 (2001) 249–307, [astro-ph/0].
[3] R. Ruffini, G. Vereshchagin, and S.-S. Xue, Electron-positron pairs in physics and astrophysics:
From heavy nuclei to black holes, Phys. Rep. 487 (Feb., 2010) 1–140, [arXiv:0910.0974].
[4] G. Breit and J. A. Wheeler, Collision of Two Light Quanta, Physical Review 46 (Dec., 1934)
1087–1091.
[5] A. I. Nikishov, Absorption of high-energy photons in the universe, Zhurnal Eksperimental’noi i
Teoreticheskoi Fiziki 41 (1961) 549–550.
[6] R. J. Gould and G. P. Schréder, Opacity of the Universe to High-Energy Photons, Physical
Review 155 (Mar., 1967) 1408–1411.
[7] V. V. Vassiliev, Extragalactic background light absorption signal in the TeV gamma-ray spectra
of blazars., Astroparticle Physics 12 (Jan., 2000) 217–238, [astro-ph/9].
[8] P. S. Coppi and F. A. Aharonian, Understanding the spectra of TeV blazars: implications for
the cosmic infrared background, Astroparticle Physics 11 (June, 1999) 35–39, [astro-ph/9].
[9] F. Aharonian, A. G. Akhperjanian, U. Barres de Almeida, A. R. Bazer-Bachi, B. Behera,
M. Beilicke, W. Benbow, K. Bernlöhr, C. Boisson, O. Bolz, V. Borrel, I. Braun, E. Brion,
A. M. Brown, R. Bühler, T. Bulik, I. Büsching, T. Boutelier, S. Carrigan, P. M. Chadwick,
L.-M. Chounet, A. C. Clapson, G. Coignet, R. Cornils, L. Costamante, M. Dalton,
B. Degrange, H. J. Dickinson, A. Djannati-Ataı̈, W. Domainko, L. O’C. Drury, F. Dubois,
G. Dubus, J. Dyks, K. Egberts, D. Emmanoulopoulos, P. Espigat, C. Farnier, F. Feinstein,

– 14 –
 A. Fiasson, A. Förster, G. Fontaine, S. Funk, M. Füßling, Y. A. Gallant, B. Giebels, J. F.
Glicenstein, B. Glück, P. Goret, C. Hadjichristidis, D. Hauser, M. Hauser, G. Heinzelmann,
G. Henri, G. Hermann, J. A. Hinton, A. Hoffmann, W. Hofmann, M. Holleran, S. Hoppe,
D. Horns, A. Jacholkowska, O. C. de Jager, I. Jung, K. Katarzyński, E. Kendziorra,
M. Kerschhaggl, B. Khélifi, D. Keogh, N. Komin, K. Kosack, G. Lamanna, I. J. Latham,
A. Lemière, M. Lemoine-Goumard, J.-P. Lenain, T. Lohse, J. M. Martin, O. Martineau-Huynh,
A. Marcowith, C. Masterson, D. Maurin, G. Maurin, T. J. L. McComb, R. Moderski,
E. Moulin, M. de Naurois, D. Nedbal, S. J. Nolan, S. Ohm, J.-P. Olive, E. de Oña Wilhelmi,
K. J. Orford, J. L. Osborne, M. Ostrowski, M. Panter, G. Pedaletti, G. Pelletier, P.-O.
Petrucci, S. Pita, G. Pühlhofer, M. Punch, S. Ranchon, B. C. Raubenheimer, M. Raue, S. M.
Rayner, M. Renaud, J. Ripken, L. Rob, L. Rolland, S. Rosier-Lees, G. Rowell, B. Rudak,
J. Ruppel, V. Sahakian, A. Santangelo, R. Schlickeiser, F. Schöck, R. Schröder, U. Schwanke,
S. Schwarzburg, S. Schwemmer, A. Shalchi, H. Sol, D. Spangler, L. Stawarz, R. Steenkamp,
C. Stegmann, G. Superina, P. H. Tam, J.-P. Tavernet, R. Terrier, C. van Eldik, G. Vasileiadis,
C. Venter, J. P. Vialle, P. Vincent, M. Vivier, H. J. Völk, F. Volpe, S. J. Wagner, M. Ward,
A. A. Zdziarski, and A. Zech, New constraints on the mid-IR EBL from the HESS discovery of
VHE γ-rays from 1ES 0229+200, A&A 475 (Nov., 2007) L9–L13, [arXiv:0709.4584].
[10] M. Meyer, M. Raue, D. Mazin, and D. Horns, Limits on the extragalactic background light in
the Fermi era, A&A 542 (June, 2012) A59, [arXiv:1202.2867].
[11] A. Sinha, S. Sahayanathan, R. Misra, S. Godambe, and B. S. Acharya, Estimation of the
Extragalactic Background Light Using TeV Observations of BL Lac Objects, ApJ 795 (Nov.,
2014) 91, [arXiv:1409.3693].
[12] R. W. Brown, K. O. Mikaelian, and R. J. Gould, Absorption of High-Energy Cosmic Photons
through Double-Pair Production in Photon-Photon Collisions, Astrophys. Lett. 14 (1973) 203.
[13] P. S. Coppi and F. A. Aharonian, Constraints on the Very High Energy Emissivity of the
Universe from the Diffuse GeV Gamma-Ray Background, ApJ 487 (Sept., 1997) L9+,
[astro-ph/9].
[14] K. Greisen, End to the Cosmic-Ray Spectrum?, Physical Review Letters 16 (Apr., 1966)
748–750.
[15] G. T. Zatsepin and V. A. Kuz’min, Upper Limit of the Spectrum of Cosmic Rays, Soviet
Journal of Experimental and Theoretical Physics Letters 4 (Aug., 1966) 78.
[16] C. D. Dermer and A. Atoyan, Ultra-high energy cosmic rays, cascade gamma rays, and
high-energy neutrinos from gamma-ray bursts, New Journal of Physics 8 (July, 2006) 122,
[astro-ph/0606629].
[17] V. S. Berezinskii and S. I. Grigor’eva, A bump in the ultra-high energy cosmic ray spectrum,
A&A 199 (June, 1988) 1–12.
[18] H. Bethe and W. Heitler, On the Stopping of Fast Particles and on the Creation of Positive
Electrons, Royal Society of London Proceedings Series A 146 (Aug., 1934) 83–112.
[19] M. J. Chodorowski, A. A. Zdziarski, and M. Sikora, Reaction rate and energy-loss rate for
photopair production by relativistic nuclei, ApJ 400 (Nov., 1992) 181–185.
[20] T. Stanev, R. Engel, A. Mücke, R. J. Protheroe, and J. P. Rachen, Propagation of ultrahigh
energy protons in the nearby universe, Phys. Rev. D 62 (Nov., 2000) 093005,
[astro-ph/0003484].
[21] A. Ringwald, Super-GZK neutrinos, Journal of Physics Conference Series 39 (May, 2006)
393–399.
[22] C. Lunardini, E. Sabancilar, and L. Yang, Ultra high energy neutrinos: absorption, thermal
effects and signatures, Journal of Cosmology and Astroparticle Physicsjcap 8 (Aug., 2013) 14,
[arXiv:1306.1808].

– 15 –
[23] J. C. D’Olivo, L. Nellen, S. Sahu, and V. van Elewyck, UHE neutrino damping in a thermal
gas of relic neutrinos, Astroparticle Physics 25 (Feb., 2006) 47–56, [astro-ph/0507333].
[24] Planck Collaboration, P. A. R. Ade, N. Aghanim, C. Armitage-Caplan, M. Arnaud,
M. Ashdown, F. Atrio-Barandela, J. Aumont, C. Baccigalupi, A. J. Banday, and et al., Planck
2013 results. XVI. Cosmological parameters, A&A 571 (Nov., 2014) A16, [arXiv:1303.5076].
[25] G. R. Blumenthal, Energy Loss of High-Energy Cosmic Rays in Pair-Producing Collisions with
Ambient Photons, Phys. Rev. D 1 (Mar., 1970) 1596–1602.
[26] F. A. Aharonian, A. M. Atoian, and A. M. Nagapetian, Photoproduction of electron-positron
pairs in compact X-ray sources, Astrofizika 19 (Apr., 1983) 323–334.
[27] P. S. Coppi and R. D. Blandford, Reaction rates and energy distributions for elementary
processes in relativistic pair plasmas, MNRAS 245 (Aug., 1990) 453–507.
[28] G. G. Fazio and F. W. Stecker, Predicted High Energy Break in the Isotropic Gamma Ray
Spectrum: a Test of Cosmological Origin, Nature 226 (Apr., 1970) 135–+.
[29] Y. Inoue, S. Inoue, M. A. R. Kobayashi, R. Makiya, Y. Niino, and T. Totani, Extragalactic
Background Light from Hierarchical Galaxy Formation: Gamma-Ray Attenuation up to the
Epoch of Cosmic Reionization and the First Stars, ApJ 768 (May, 2013) 197,
[arXiv:1212.1683].
[30] V. B. Berestetskii, E. M. Lifshitz, and V. B. Pitaevskii, Quantum Electrodynamics. Elsevier,
1982.
[31] S. Batebi, S. Tizchang, R. Ruffini, G. Vereshchagin, and S.-S. Xue, Euler-heisenberg interaction
and opacity of the universe at high energy, in preparation (2015).
[32] R. Jost, J. M. Luttinger, and M. Slotnick, Distribution of Recoil Nucleus in Pair Production by
Photons, Physical Review 80 (Oct., 1950) 189–196.
[33] J. W. Motz, H. A. Olsen, and H. W. Koch, Pair Production by Photons, Reviews of Modern
Physics 41 (Oct., 1969) 581–639.
[34] M. Giroletti, Millimeter-VLBI observations of blazars, Journal of Physics Conference Series
131 (Oct., 2008) 012054.
[35] G. A. Medina Tanco, E. M. de Gouveia Dal Pino, and J. E. Horvath, Deflection of
Ultra–High-Energy Cosmic Rays by the Galactic Magnetic Field: From the Sources to the
Detector, ApJ 492 (Jan., 1998) 200–204, [astro-ph/9707041].

– 16 –