Skip to main content

Full text of "Is astronomy possible with neutral ultra-high energy cosmic ray particles existing in the Standard Model?"

See other formats

Is astronomy possible with neutral ultra-high energy cosmic ray particles existing in 

the Standard Model? 





> ■ 

o\ : 

in . 
m ■ 

(N ■ 

\o : 
o ■ 

6 ■ 


P.G. Tinyakov a - c and I.I. Tkachev 6 ^ 
° Universite Libre de Bruxelles, 
CP225, Bvd du Triomphe, 1050 Bruxelles, Belgium 
b CERN Theory Division, CH-1211 Geneva 23, Switzerland 
c Institute for Nuclear Research, Moscow 117312, Russia 
(Dated: February 5, 2008) 

The recently observed correlation between HiRes stereo cosmic ray events with energies E ~ 
10 19 eV and BL Lacs occurs at an angle which strongly suggests that the primary particles are 
neutral. We analyze whether this correlation, if not a statistical fluctuation, can be explained within 
the Standard Model, i.e., assuming only known particles and interactions. We have not found a 
plausible process which can account for these correlations. The mechanism which comes closest — 
the conversion of protons into neutrons in the IR background of our Galaxy — still under-produces 
the required flux of neutral particles by about 2 orders of magnitude. The situation is different at 
E ~ f 20 eV where the flux of cosmic rays at Earth may contain up to a few percent of neutrons 
pointing back to the extragalactic sources. 


In the last years it has been observed that various 
ultra-high energy cosmic ray (UHECR) data sets exhibit 
correlations with the BL Lacertae objects (BL Lac) at 
different level of significance [l|, Q . Recently, the HiRes 
stereo data have appeared which have unprecedented an- 
gular resolution of ~ 0.6°. This dataset shows correla- 
tions with BL Lacs at the angular scale compatible with 
the angular resolution. The statistical significance of the 
correlation is estimated to be of the order of (11 
coincidences observed at ~ 3 expected in the absence of 
correlations) [HQ- The absence of adjustable cuts makes 
it straightforward, for the first time, to predict the signal 
which should be observed in the future datasets if BL 
Lacs are sources of the ultra-high energy cosmic rays Q . 

A most striking feature of the correlation found in the 
HiRes data is that it occurs at an angle which is much 
smaller than the typical deflection of a proton of corre- 
sponding energy in the Galactic magnetic field (GMF). 
The purpose of this paper is to investigate whether the 
existence of such correlations can be explained within 
the Standard Model, i.e., assuming only known particles 
and interactions. We will argue that this is extremely 
unlikely, if not impossible. 

In order to proceed with the argument we need to make 
several assumptions. Although these assumptions are 
plausible, they may not be valid. If this be the case, 
the results of our analysis should be reconsidered. 

The assumptions are as follows: 

i) The fraction of correlating events at energy E > 
10 19 eV is larger than ~ 1%. 

ii) The Galactic magnetic field around Earth location 
has a coherent component with the strength of or- 
der 2 - 3 pjG. 

iii) The distances to BL Lacs which are counterparts 

(sources) of correlating events are larger than ~ 
100 Mpc. 

The validity of the assumption i) has been discussed 
in detail in Refs. Q. Note that it is implicitly assumed 
here that energies of cosmic rays are measured correctly. 

The assumption ii) is the widely accepted value of the 
Galactic magnetic field in the vicinity of the Earth (for 
recent reviews see, e.g., Refs. |6[). The precise magnitude 
of MGF is not important for the argument; its variations 
by a factor 2-3 would not change our conclusions. 

Finally, the assumption iii) is needed because some of 
the BL Lacs which contribute to correlations have un- 
known redshifts. It is usually expected that these red- 
shifts exceed 0.1 — 0.2. 

Given the assumptions i)— iii), the argument proceeds 
as follows. The deflection of a E = 10 20 eV proton in 
the 2 /iG coherent field extending over 1 kpc is 1°. Most 
of the events, however, have much lower energies (for the 
events of energies E > 10 19 eV with the spectrum falling 
like ~ 1/E 3 the median energy is 1.5 x 10 19 eV). Since 
the correlating events follow the same distribution (4j, 
their typical deflections would be ^ 7°. The correlation 
with the sources would therefore be destroyed. At such 
a small angular scale as observed, the correlations can 
survive in the following cases only: 

1) There exist "windows" in the Galactic magnetic 
field with a very low value of the coherent com- 

2) A fraction of primary particles is neutral. 

3) A fraction of primaries is converted to neutral par- 
ticles before entering the Galactic magnetic field, 
i.e., at least 1 kpc from the Earth (assuming the 
Galactic magnetic field does not extend further 
than ~ 1 kpc from the disk). 

We consider these three possibilities in Sects. II-V. In this 
paper we limit ourselves to mechanisms which are based 


on particles and interactions existing in the Standard 
Model. We show that none of such mechanisms can ex- 
plain the observed correlation, unless very unlikely as- 
sumptions are made. In the last Sect. VI we summarize 
the arguments and present the conclusions. 

A. Galactic magnetic field 

The Galactic magnetic field (GMF) consists of two 
components, the coherent and the turbulent one. The ex- 
istence of the coherent component is the main reason why 
the UHECR-BL Lacs correlations at E ~ 10 19 eV cannot 
be explained by protons. In models which are currently 
in use, the coherent GMF extends to the whole Galaxy, 
being described by a simple analytic function. However, 
such a picture is probably an oversimplification. Obser- 
vationally, there are many anomalies and features in the 
Galactic magnetic field. It is not totally excluded that 
the coherent component is "patchy" . In other words, 
there may exist "windows" where the coherent compo- 
nent is negligible. In this case the ultra-high energy pro- 
tons may cross the GMF undeflected when they come 
from the directions of these "windows" . One may thus 
try to explain the observed correlations by the existence 
of such windows. 

For this mechanism to work the random component of 
the GMF in windows also has to satisfy some require- 
ments. The deflections of protons in the random field is 
estimated as follows 

where E is the energy of proton, B r and L c are the rms 
value and the coherence length of the random magnetic 
field, respectively, while D is the propagation distance. 
This deflection has to be (much) smaller than 0.5°. 

The coherence length L c is the most uncertain of the 
above parameters. Quite often a large values of L c up 
to L c ~ 50 pc are assumed. On the contrary, in those 
regions of the sky where the spectrum of the magnetic 
field fluctuations was measured, L c turns out to be small 
Q. For instance, in Ref. Q the linearly polarized con- 
tinuum emission was studied in the test region near the 
Galactic plane covering the range 325.5° < I < 332.5°, 
—0.5° < b < 3.5°. Polarized emission was found to orig- 
inate mainly at the distance of ~ 3.5 kpc. Interestingly, 
two large areas of a few square degrees each were found 
to be devoid of polarization. It was argued that these 
voids were produced by the foreground in which the mag- 
netic field is disordered, with the coherence length being 
L c ~ 0.1 — 0.2 pc. In these voids, the projection of the 
coherent component of the magnetic field on the line of 
sight was found to be < 0.15 of the rms value of the 
random field strength. In the rest of the test region, i.e. 

outside of the voids, the coherence length is much larger, 
but still the outer scale of turbulence did not exceed 2 pc 
@. Thus, the existence of regions with 5 r < 0.5° docs 
not seem impossible. 

This mechanism has a specific signature which is 
straightforward to test. If there exist "windows" with 
the small coherent component of GMF, the Faraday ro- 
tation measures must be small in these windows as well. 
In other words, the Faraday rotations in the directions of 
correlating UHECR events must be anomalously small. 
This may be tested statistically by comparing the distri- 
bution of Faraday rotations in the direction of correlating 
events with the distribution of Faraday rotations in the 
random directions selected according to the distribution 
of BL Lacs and all cosmic ray events [37j ■ We have per- 
formed this test with the existing data and found that 
the two distributions are indeed different (Faraday rota- 
tions in the directions selected with real data are anoma- 
lously small) with the significance of ~ 4% according to 
the Kolmogorov-Smirnov test. This is not a very signif- 
icant deviation. The result demonstrates, however, that 
the method may work quite well with the future larger 

Although the existence of "windows" in the coherent 
component of the Galactic magnetic field goes against 
the standard lore, a much better understanding of the 
Galactic magnetic field is required to definitely rule it 

B. Extragalactic magnetic fields 

For the mechanism outlined above to work, the extra- 
galactic magnetic fields have to satisfy certain require- 
ments (which also apply to the scenarios considered in 
Sec. IIV[) . The extragalactic magnetic fields are not mea- 
sured. Computer simulations indicate [nj E| that the 
magnetic field strength in voids between clusters can be 
very small, B r < 10~ 12 G, while the coherence length 
can easily be significantly smaller than 1 Mpc. Eq. (fT|) 
then shows that the deflections in voids are negligible. It 
is interesting to note that EGMF with such small magni- 
tude are in principle measurable injabservations of TeV 
gamma-rays from distant blazars 

The strength of the field in filaments is larger. How- 
ever, the probability to cross many filaments is small and 
regions with small deflections can occupy rather large 
fraction of the sky area [13, EH (see however [Hj]). Over- 
all, the model where the extragalactic magnetic fields are 
sufficiently small and do not spoil correlations is accept- 
able at present. 


Among known neutral particles the following are suffi- 
ciently stable to propagate over extragalactic distances: 
neutrino, photon and atoms. In this section we discuss 


the possibility to explain correlations by assuming that 
primary cosmic rays are composed of these particles. 

Both neutrino and photon initiate air showers deeper 
in the atmosphere than the hadronic primary particles. 
Therefore, these models can be falsified with already ex- 
isting data by, e.g., comparing the X max distributions of 
the correlating events with that of the whole set. Since 
the corresponding data are still unpublished, we briefly 
discuss the models based on neutrino and photon and 
show that they have difficulties per se, even without re- 
ferring to A max distributions. 

a. Neutrino. At E *5 10 19 eV the neutrino cross- 
section with protons is smaller by a factor of ~ 3 x 1CP 7 
than the pp cross-section [Tij . Therefore, the optical 
depth of the atmosphere for neutrinos is 3 x 10~ 5 . On 
the other hand, at this energy the neutrino flux cannot 
exceed the flux of hadronic cosmic rays by more than a 
factor of 50 [l5j]. It follows that at most (a few) x 1CP 4 
of all cosmic ray events can be due to neutrinos. This is 
more than a factor of 10 lower than needed to explain cor- 
relations. Thus, neutrino with the standard weak inter- 
actions cannot explain correlations observed in the HiRes 
data set. 

A "genuine" (hypothetical) neutrino mechanism would 
involve strong neutrino interactions with the atmosphere 
at high energies [16]. As such a behavior in not a part of 
the Standard Model, the corresponding speculations fall 
outside of the scope of the present paper. 

Another possibility which exists within the minimal 
cxtention of the Standard Model by the non-zero neu- 
trino masses, the Z-burst mechanism [l7|, requires an 
unnaturally large flux of neutrinos at E > 10 22 eV which 
is in conflict with the limits on neutrino flux from ra- 
dio experiments [l8[. The particles observed at Earth in 
this mechanism are mostly photons produced in the in- 
teractions of UHE neutrinos with the cosmological neu- 
trino background on their way to the Earth. Low radio- 
background and small values of EGMF are required to 
avoid the conflict with the upper bound on the diffuse 
flux of gamma rays [19( . 

b. Photon. A set of conditions under which the 
UHE photons can reach the Earth from BL Lacs was 
considered in Ref. [20]. On their way the photons in- 
teract with the CMBR and radio background photons 
producing e + e _ pairs, one of these particles typically 
carrying most of the energy. These leading particles in 
turn Compton up-scatter CMBR photons to the energy 
almost equal to the energy of the original photon. This 
process is usually referred to as the electromagnetic cas- 
cade. The developing electromagnetic cascade can reach 
the Earth from several hundred megaparsecs with energy 
E ~ 10 19 eV if the following conditions are satisfied: 

• the radio-background is small, smaller than the the- 
oretically expected value; 

• the injection spectrum cx E~ a is hard, a < 1.5; 

• maximum energy of photons at the source reaches 
10 23 eV 

• EGMF are small, B < 10~ 12 G; 

• sources are predominantly photonic, N^/N p ^ 10 2 . 

These conditions impose extreme requirements on the as- 
trophysical sites where such photons can be produced. 
There are no candidates known which could satisfy these 

c. Atoms. It may, in principle, happen that in the 
cosmological radiation field the proton produces an e + e~- 
pair and "dresses" itself with the electron forming an 
hydrogen atom and emitting a free positron. The dif- 
ferential cross section of electromagnetic pair production 
by a single photon in the Coulomb field of a nucleus with 
subsequent capture of the electron is estimated as [2lJ 

da _ 47ra 6 Z 5 1 
dEp ml E p ' 

where Z is the electric charge of an ion and the positron 
energy E p is supposed to be much larger than m e . Mul- 
tiplying the cross section integrated over energy by the 
density of the CMB photons one may estimate the rate 
of the formation of hydrogen atoms, Z — 1, as 

^formation ~ 10" 5 MpC^ 1 . 

The decay rate (ionization on the CMB radiation) is es- 
timated in a standard way by using the Klein-Nishina 
cross section. One finds 

i?dccay ~ 100 Mpc" 1 . 

Thus, the fraction of neutral particles (atoms) produced 
by this mechanism is of the order of 10 -7 , which is too 
small to explain correlations. 

As a side remark note that for a heavy nuclei the rates 
of radiative capture and ionization are comparable when 
Z w 25. This corresponds to the typical equilibrium 
charge of a heavy ion (iron or heavier) propagating in 
the CMBR. 


In order to be able to fly over 1 kpc (the thickness of 
the Galactic magnetic field) a neutral particle created at 
the outskirts of the Galaxy has to be sufficiently stable. 
At energy 10 19 eV this implies for the rest-frame lifetime 

1n / m \ Zl0 19 eV\ 
T0>1 ° S (iGev) HH' 

where E and m are the energy and the mass of the parti- 
cle, respectively. Among known particles which we have 
not yet discussed only neutrons satisfy this requirement. 
In this section we consider various mechanisms of neutron 
creation in or near the Galaxy. 


There are several ways to produce neutrons in the 
Standard Model: photodisintegration of nuclei, photo- 
production on background photons by protons, creation 
in pp reactions and in inverse /3-decay on background 
neutrinos or photons. We consider these mechanisms in 
turn and argue that none of them can produce a sufficient 
fraction of neutrons in the cosmic ray flux. 

A. Inverse /3-decay on background neutrinos. 

The simplest of the above mechanisms is the inverse 
beta-decay, p + v — ► n + e + . The cross section of this 
reaction is [22J 

<y(pD -> ne+) ~ -G 2 F (gv + 3<^)£ 2 , 

where <?y + 3g\ ^5.7 and E is the energy of neutrino 
in the proton rest frame. When E reaches ~ 1 GeV 
the cross section levels out and stabilizes at the value of 
Cmax ~ 10 -14 barn. With this maximum value taken for 
the estimate, the rate of the conversion is 

R max ~4x 1(T 12 Mpc- 1 . (2) 

Thus, these processes are totally negligible. 

B. Creation of neutrons in the radiation fields. 

The process of creation of neutrons in interactions of 
CR primaries with the background photons produces the 
largest contribution, therefore we consider it in most de- 

1. Galactic and extragalactic radiation fields and reaction 

In the laboratory frame, the rate of reactions with the 
photon background is given by the standard expression, 

R = J d 3 p n(p)(l-v cos 8)a(Co), (3) 

where n(p) is the photon density in the laboratory frame, 
cr(u)) is the cross section of the relevant reaction in the 
rest frame of the primary particle as the function of the 
energy of the incident photon Co — 7f>(l — v cos 0) , 7 is the 
gamma-factor of the primary particle in the laboratory 
frame and v is its velocity (7 = l/vl — v 2 ). In what 
follows we assume 7> 1. 

In the case of the isotropic background this expression 
may be simplified. Integrating over angles we find 

2n f°° f 2lp 
R h) = — / dpn(p) / dujujcr(oj). (4) 
7 Jo Jo 

For the black body radiation with temperature T one has 

n(p) = n T ( P )= (2 ^ )3exp J T) _ 1 . (5) 

This gives the answer in the case of CMBR. Other back- 
grounds, Galactic and extragalactic, are usually charac- 
terized in the literature by the spectral energy distribu- 
tion I(y, i) (energy per unit frequency per unit solid an- 
gle) which in turn is usually quoted in terms of the Planck 
function B V (T) and emissivity e 

J(i/,i) =e{v,i)B v {T). (6) 

Here i stands for the unit vector in the direction of ob- 
servation. For the black body radiation e(u,i) = 1. The 
Planck function, being written as a function of photon 
momentum p = 2t:v, takes the form 

B p (T)=p 3 n T (p). (7) 

Therefore, the photon number density for the background 
with the known emissivity is given by the expression 

n(p) = e(p/27r) n T {p). (8) 

In what follows we will be interested in the Galactic 
and extragalactic far-infrared backgrounds (FIRB) (for 
a recent review see [ll])- According to Ref [23 |. the 
isotropic extragalactic FIRB can be parameterized by 

e(p) = 1.3x 10- 5 (p/po) a64 , (9) 

where po = 144 K (which corresponds to i/ = 100 cm -1 ), 
while the temperature parameter in ut(p) corresponds to 
T = 18.5 K. 

The Galactic FIRB has been measured by 
COBE/DIRBE. The spectral energy density I(u,i) 
as a function of galactic coordinates can be downloaded 
from [25j. The radiation is dominated by the Galactic 
plane where the Galactic bulge is by far the brightest 
region. One may approximate this radiation field by a 
point source in the Galactic center. We have verified 
that this approximation gives a good agreement with 
the exact calculations for cosmic ray trajectories which 
do not pass close to the Galactic center. 

According to 26], the averaged spectral properties of 
the Galactic FIRB can be described by Ut(p) with T = 
20.4 K and e(p) cx p 2 . Therefore, in what follows for the 
Galactic FIRB we use 

£ (P) = % P 2 5 ( n - n °)> ( 10 ) 

where Io is the normalization factor, no is the unit vector 
in the direction from the Galactic center, and r is the 
distance to the Galactic center. The constant I can 
be found by normalizing the total luminosity within the 
Sun orbit to the measured value Lq = 1.8 x 10 10 L© ~ 
7 x 10 36 W [26[. We find 

_ 63L G 

g7r 4 T 6 ■ 


along the particle trajectory, 


FIG. 1: The fraction F of neutrons produced per one incident 
particle (solid lines) in the reactions 4 He + 7 — * 3 He + n (left 
curve) and p + 7 — > n + ix + (right curve) on the background 
radiation fields as a function of the 7-factor of the incident 
particle. Dotted and dashed-dotted lines show contributions 
of the extragalactic and Galactic backgrounds, respectively. 

The reaction rate Eq. Q in this case can be expressed 


126 L G 
64tt 7 T 6 r 2 


dpp A a (Co) 
exp(p/T) - 1 ' 


Here u> = "fp(l — p), /U = cos 9 and 9 is the collision angle 
between the CR primary and the background photon. 

2. Conversion in the extragalactic space. 

The fraction of neutrons created over the distance dl 
is Rdl. Due to the finite neutron lifetime the fraction of 
neutrons which reach the solar system is given by 

F(j) = R [ 

- l/x dl = RX, 


where R is given by the expression Eq. Q and A is the 
mean propagation distance of the free neutron, 

A = 0.86 ^ Mpc. 

The function F(j) is shown in Fig. [1] by dotted lines 
for the two reactions, the pion photoproduction p + 7 — > 
n + 7r + (right curve) and the reaction of nuclear pho- 
todissociation 4 He + 7 — > 3 He + n (left curve). In these 
calculations the experimentally measured cross sections 
of the corresponding reactions were used [U Hfl . 

3. Conversion in the Galactic infrared radiation field. 

In this case the number of neutrons produced per one 
incident particle is determined by the rate pip integrated 

dlR(-f,r, 9) e- l/x , 


where I is the distance from the Sun along the trajec- 
tory and r is the distance from the current point to the 
Galactic center. In the case when the radiation field is 
approximated by the single source in the Galactic center, 
the particle trajectory is completely characterized by the 
angle ip which it forms with the direction to the Galactic 
anti-center (ip = it corresponds to the trajectory which 
passed through the Galactic center). In terms of this 
angle the distance r entering Eq. (flU)) reads 

r = y/D 2 + I 2 + 2DI cos ip, 

while the collision angle 9 is 

cos 9 

D cos ip + I 

where D w 8 kpc is the distance from the Sun to the 
Galactic center 

The Galactic contribution F(j, vb) to the fraction of 
the produced neutrons in the case ip = 90° is shown in 
Fig. Q] by the dashed-dotted lines for the reactions p + 
7 — * n + 7r + (right curve) and 4 He + 7 — * 3 Hc + n 
(left curve). Here again we have used the cross sections 
measured experimentally. 

As far as the correlations observed in the HiRes data at 
E > 10 19 eV are concerned, the relevant range of the 7- 
factors is (1 — 2) x 10 10 . In this region the reaction 4 He + 
7 — ► 3 He + n is irrelevant for distant sources. Indeed, 
in the case of 4 He these 7-factors correspond to energies 
(4 — 8) x 10 19 eV. The helium nuclei of such energies do 
not propagate over several hundred megaparsecs [29j, so 
they cannot be present in the cosmic ray flux coming from 
BL Lacs. The other reaction, p + 7 — > n + tt + , produces 
a fraction of neutrons at the level of (a few) x 10 -4 (see 
Fig. [l}, which is not sufficient to explain correlations by 
almost two orders of magnitude. 

Neutron production in collisions with 
interstellar matter. 

Neutrons can be produced in collisions of hadronic pri- 
maries with the interstellar gas in the galaxy. The conver- 
sion probability is given by the optical depth r = A/V , 
where J\f is the column density of the intervening inter- 
stellar gas in a given direction and a g is the interaction 
cross section. In order to explain correlations [1, H| one 
needs t <; 10~ 2 . 

A typical value of the HI (neutral hydrogen) column 
density in directions of the Galactic poles is A/hi ~ 
10 20 cm -2 [30|. Using as an upper limit for a g the 
value of the total pp cross-section at relevant energies, 
(Tpp w 100 mb = 10~ 25 cm 2 , one finds t pp ~ 10~ 5 , which 
is too small to produce the required fraction of neutrons. 


The argument can be rephrased in a different way. 
One may assume that a mass fraction 77 of the Galac- 
tic halo consists of baryons including nuclei, neutral and 
ionized gas and possibly dark baryons. The column mass 
density of matter in the direction of the Galactic anti- 
center, as deduced from the Milky Way rotational curve, 
is ~ 10 22 GeV cm -2 [3l[, and therefore the column den- 
sity of baryons is ~ 77 10 22 cm -2 . To reproduce the re- 
quired rate of pn conversions one would need a fraction 
r\ 10, which is clearly impossible. 

As a side remark let us point out that neutrons can 
be, in principle, produced in the interactions of primary 
protons with a non-baryonic dark matter in the Galactic 
halo. Parameterizing the relevant cross-section in the 
energy range of interest as a = Eq 2 and making use 
of the matter column density of the Galactic halo cited 
above we find 



1 TeV 


1 eV 


where ttidm is the mass of the dark matter particle. 
Among the scenarios involving new physics, this one has 
several advantages. It automatically provides a normal 
shower development in the atmosphere (contrar y to the 
models with new particles as neutral messengers |32l.l33j) 
and avoids the problem of messenger production in AGNs 
[34| . In addition, we know from precision cosmologi- 
cal data that the non-baryonic dark matter must ex- 
ist. Correlations in this scenario should dissappear at 
E & 10 17 eV due to the final life-time of the neutron. 
Note also that one should expect the existence of the 
Greisen-Zatsepin-Kuzmin cut-off [H, |36[ in the cosmic 
ray spectrum in this model. 


In this paper we have considered different mechanisms 
which could potentially explain the observed correlations 
of the cosmic ray events with BL Lacs at the energy 
E ~ 10 19 eV and the angle ~ 0.6° coincident with the 

angular resolution of the HiRes experiment. We found 
that the mechanisms which assume only known particles 
and interactions under-produce the flux of neutral parti- 
cles needed to explain these correlations by at least two 
orders of magnitude. 

There remains a possibility of the astrophysical solu- 
tion, which is related to our insufficient knowledge of the 
Galactic magnetic field. The observed tight correlations 
can potentially be explained if there exist "windows" in 
the Galactic magnetic field with a very low value of the 
coherent component of the field and a small coherence 
length of the turbulent component. Though this possi- 
bility is exotic, it cannot be ruled out at present. 

The mechanisms which we have discussed in this pa- 
per are based on the known physics, i.e. they certainly 
operate in Nature provided the cosmic ray flux contains 
light nuclei or protons. One of these mechanisms, the 
conversion of protons to neutrons, implies that at ener- 
gies around 10 20 eV a few percent of the ultra- high energy 
protons (cf. Fig.[T]) get converted into neutrons and cross 
the Galactic magnetic field undeflected. Therefore, if the 
cosmic rays with the energy around the GZK cutoff are 
protons, there must be a few percent fraction of them 
that point back to the sources with the accuracy better 
than a fraction of a degree, provided the extragalactic 
magnetic fields are small. With a large statistics, this 
may allow to measure separately the Galactic and extra- 
galactic magnetic fields and to verify by an independent 
method the chemical composition of UHECR. 


We are grateful to A. Kashlinsky, V. Kuzmin, 
D. Semikoz, G. Thompson and T. Weiler for valuable 
comments and discussions. I.T. thank the Galileo Galilei 
Institute for Theoretical Physics for the hospitality and 
the INFN for partial support during the completion of 
this work. The work of P.T. is supported in part by the 
IISN, Belgian Science Policy (under the contract IAP 

[1] P. G. Tinyakov and I. I. Tkachev, JETP Lett. 74 (2001) 
445 [Pisma Zh. Eksp. Teor. Fiz. 74 (2001) 499]; 

[2] P. G. Tinyakov and I. I. Tkachev, Astropart. Phys. 18 
(2002) 165; D. S. Gorbunov, P. G. Tinyakov, 1. 1. Tkachev 
and S. V. Troitsky, Astrophys. J. 577 (2002) L93; 
P. Tinyakov and I. Tkachev, "Correlations and charge 
composition of UHECR without knowledge of galactic 
magnetic field," In the proceedings of 28th International 
Cosmic Ray Conferences (ICRC 2003), Tsukuba 2003, 
pp. 671-674, |arXiv:astro-ph/0305363l 

[3] D. S. Gorbunov, P. G. Tinyakov, I. I. Tkachev and 
S. V. Troitsky, JETP Lett. 80 (2004) 145 [Pisma Zh. 
Eksp. Teor. Fiz. 80 (2004) 167]. 

[4] R. U. Abbasi et al. [HiRes Collaboration], Astrophys. J. 

636 (2006) 680. 
[5] D. S. Gorbunov, P. G. Tinyakov, I. I. Tkachev and 

S. V. Troitsky, JCAP 0601 (2006) 025. 
[6] J. P. Vallee, New Astronomy R eviews 48 (2004) 

763; R. Beck, |arXiv:astro-ph/0603531| J. L. Han, 

[7] P. G. Tinyakov and I. I. Tkachev, Astropart. Phys. 24 

(2005) 32. 

[8] B. M. Gaensler, J. M. Dickey, N. M. McClure- Griffiths, 
A. J. Green, M. H. Wieringa and R. F. Haynes, ApJ 549 
(2001) 959. 

[9] M. Haverkorn, B. M. Gaensler, N. M. McClure- Griffiths, 
J. M. Dickey and A. J. Green, ApJ 609 (2004) 776. 
[10] K. Dolag, D. Grasso, V. Springel and I. Tkachev, JETP 


Lett. 79 (2004) 583 [Pisma Zh. Eksp. Teor. Fiz. 79 (2004) [25] 
719]; JCAP 0501 (2005) 009. [26] 
[11] M. Bruggen, M. Ruszkowski, A. Simionescu, M. Hoeft [27] 
and C. D. Vecchia, Astrophys. J . 631 (2005) L21. [28] 
A. Neronov and D. V. Semikoz, |a7Xiv:astro-ph/0604607| [29] 
G. Sigl, F. Miniati and T. A. Ensslin, Phys. Rev. D 68 
(2003) 043002. 

R. Gandhi, C. Quigg, M. H. Reno and I. Sarcevic, Phys. 
Rev. D 58, 093009 (1998). 

O. E. Kalashev, V. A. Kuzmin, D. V. Semikoz and 
G. Sigl, Phys. Rev. D 66, 063004 (2002). 

G. Domokos and S. Nussinov, Phys. Lett. B 187 (1987) [30] 

D. Fargion, B. Mele and A. Salis, Astrophys. J. 517 [31] 
(1999) 725; T. J. Weiler, Astropart. Phys. 11 (1999) 303. 
D. V. Semikoz and G. Sigl, JCAP 0404, 003 (2004). [32] 
O. E. Kalashev, V. A. Kuzmin, D. V. Semikoz and 
G. Sigl, Phys. Rev. D 65 (2002) 103003. [33] 

0. E. Kalashev, V. A. Kuzmin, D. V. Semikoz and 

1. I. Tkachev, |arXiv:astro-ph/0107130| [34] 
A. W. Aste, K. Hencken, D. Trautmann and G. Baur, 
Phys. Rev. A 50 (1994) 3980. [35] 
L. B. Okun, "Leptons And Quarks," Amsterdam, [36] 
Netherlands: North-Holland (1982). 

A. Kashlinsky, Phys. Rept. 409 (2005) 361. [37] 
D. J. Fixsen, E. Dwek, J. C. Mather, C. L. Bennett and 
R. A. Shafer, "The Spectrum of the Extragalactic Far 
Infrared Background from the COBE Astrophys. J. 508 
(1998) 123. 

|http:/ /lambda.gsfc. / prod uct /cobe/| 

E. L. Wright et al, ApJ 381 (1991) 200. 

T. Fujii et al., Nucl. Phys. B 120 (1977) 395. 
http: / /www. / exfor00.htm 

F. W. Stecker, Phys. Rev. 180 (1969) 1264; J. L. Puget, 

F. W. Stecker and J. H. Bredekamp, Astrophys. J. 205 
(1976) 638; F. W. Stecker and M. H. Salamon, As- 
trophys. J. 512 (1999) 521; T. Yamamoto, K. Mase, 
M. Takeda, N. Sakaki and M. Teshima, Astropart. Phys. 
20 (2004) 405; D. Hoop er, S. Sarkar and A. M. Taylor, 
|arXiv:astro-ph/ 0608085 

J. M. Dickey and F. J. Lockman, Ann. Rev. Astron. As- 
trophys. 28 (1990) 215. 

A. Boyarsky, A. Neronov, O. Ruchayskiy, M. Shaposh- 
nikov and I. Tkachev, |arXiv:astro- ph/0603660| 
D. J. H. Chung, G. R. Farrar and E. W. Kolb, Phys. Rev. 
D 57 (1998) 4606. 

V. Berezinsky, M. Kachelriess and S. Ostapchenko, Phys. 
Rev. D 65 (2002) 083004. 

M. Kachelriess, D. V. Semikoz and M. A. Tortola, Phys. 

Rev. D 68 (2003) 043005. 

K. Greisen, Phys. Rev. Lett. 16 (1966) 748. 

G. T. Zatsepin and V. A. Kuzmin, JETP Lett. 4 (1966) 
78 [Pisma Zh. Eksp. Teor. Fiz. 4 (1966) 114]. 

One may construct this set by choosing the directions 
to BL Lacs correlating with the Monte Carlo simulated 
cosmic ray events. In this way the distributions of both 
BL Lacs and the cosmic ray events are taken into account.