F. Aharonian1 - A. G. Akhperjanian2 - K.-M. Aye3 - A. R. Bazer-Bachi4 - M. Beilicke5 - W. Benbow1 - D. Berge1 - P. Berghaus6,
- K. Bernlöhr1,7 - C. Boisson8 - O. Bolz1 - I. Braun1 - F. Breitling7 - A. M. Brown3 - J. Bussons Gordo9 - P. M. Chadwick3 - L.-M. Chounet10 - R. Cornils5 - L. Costamante1,20 - B. Degrange10 - A. Djannati-Ataï6 - L. O'C. Drury11 - G. Dubus10 - D. Emmanoulopoulos12 - P. Espigat6 - F. Feinstein9 - P. Fleury10 - G. Fontaine10 - Y. Fuchs13 - S. Funk1 - Y. A. Gallant9 - B. Giebels10 - S. Gillessen1 - J. F. Glicenstein14 - P. Goret14 - C. Hadjichristidis3 - M. Hauser12 - G. Heinzelmann5 - G. Henri13 - G. Hermann1 - J. A. Hinton1 - W. Hofmann1 - M. Holleran15 - D. Horns1 - O. C. de Jager15 - S. Johnston21 - B. Khélifi1 - J. G. Kirk1 - Nu. Komin7 - A. Konopelko1,7 - I. J. Latham3 - R. Le Gallou3 - A. Lemière6 - M. Lemoine-Goumard10 - N. Leroy10 - O. Martineau-Huynh16 - T. Lohse7 - A. Marcowith4 - C. Masterson1,20 - T. J. L. McComb3 - M. de Naurois16 - S. J. Nolan3 - A. Noutsos3 - K. J. Orford3 - J. L. Osborne3 - M. Ouchrif16,20 - M. Panter1 - G. Pelletier13 - S. Pita6 - G. Pühlhofer1,12 - M. Punch6 - B. C. Raubenheimer15 - M. Raue5 - J. Raux16 - S. M. Rayner3 - I. Redondo10,20,
- A. Reimer17 - O. Reimer17 - J. Ripken5 - L. Rob18 - L. Rolland16 - G. Rowell1 - V. Sahakian2 - L. Saugé13 - S. Schlenker7 - R. Schlickeiser17 - C. Schuster17 - U. Schwanke7 - M. Siewert17 - O. Skjæraasen22 - H. Sol8 - R. Steenkamp19 - C. Stegmann7 - J.-P. Tavernet16 - R. Terrier6 - C. G. Théoret6 - M. Tluczykont10,20 - G. Vasileiadis9 - C. Venter15 - P. Vincent16 - H. J. Völk1 - S. J. Wagner12
1 -
Max-Planck-Institut für Kernphysik, Heidelberg, Germany
2 -
Yerevan Physics Institute, Yerevan, Armenia
3 -
University of Durham, Department of Physics, Durham, UK
4 -
Centre d'Étude Spatiale des Rayonnements, CNRS/UPS, Toulouse, France
5 -
Universität Hamburg, Institut für Experimentalphysik, Hamburg,
Germany
6 -
APC, Paris, France
7 -
Institut für Physik, Humboldt-Universität zu Berlin, Berlin,
Germany
8 -
LUTH, UMR 8102 du CNRS, Observatoire de Paris, Meudon, France
9 -
Groupe d'Astroparticules de Montpellier, IN2P3/CNRS, Université
Montpellier II, Montpellier, France
10 -
Laboratoire Leprince-Ringuet, IN2P3/CNRS, École Polytechnique,
Palaiseau, France
11 -
Dublin Institute for Advanced Studies, Dublin, Ireland
12 -
Landessternwarte, Königstuhl, Heidelberg, Germany
13 -
Laboratoire d'Astrophysique de Grenoble, INSU/CNRS, Université
Joseph Fourier, Grenoble, France
14 -
DAPNIA/DSM/CEA, CE Saclay, Gif-sur-Yvette, France
15 -
Unit for Space Physics, North-West University, Potchefstroom, South
Africa
16 -
Laboratoire de Physique Nucléaire et de Hautes Énergies, IN2P3/CNRS,
Universités Paris VI & VII, Paris, France
17 -
Institut für Theoretische Physik, Lehrstuhl IV: Weltraum und
Astrophysik, Ruhr-Universität Bochum, Germany
18 -
Institute of Particle and Nuclear Physics, Charles University, Prague, Czech Republic
19 -
University of Namibia, Windhoek, Namibia
20 -
European Associated Laboratory for Gamma-Ray Astronomy, jointly
supported by CNRS and MPG
21 -
School of Physics, University of Sydney, Australia
22 -
Institute of Theoretical Astrophysics, University of Oslo, Norway
Received 3 March 2005 / Accepted 2 June 2005
Abstract
We report the discovery of very-high-energy (VHE) -ray
emission of the binary system
PSR B1259-63/SS 2883 of a radio pulsar
orbiting a massive, luminous Be star in a highly eccentric orbit. The
observations around the 2004 periastron passage of the pulsar were
performed with the four 13 m Cherenkov telescopes of the HESS experiment, recently installed in Namibia and in full operation since
December 2003. Between February and June 2004, a
-ray signal
from the binary system was detected with a total significance above
.
The flux was found to vary significantly on timescales
of days which makes PSR B1259-63 the first variable galactic source
of VHE
-rays observed so far. Strong emission signals were
observed in pre- and post-periastron phases with a flux minimum around
periastron, followed by a gradual flux decrease in the months
after. The measured time-averaged energy spectrum above a mean
threshold energy of 380 GeV can be fitted by a simple power law
with a photon index
and flux normalisation
.
This
detection of VHE
-rays provides unambiguous evidence for
particle acceleration to multi-TeV energies in the binary system. In
combination with coeval observations of the X-ray synchrotron emission
by the RXTE and INTEGRAL instruments, and assuming the VHE
-ray emission to be produced by the inverse Compton mechanism,
the magnetic field strength can be directly estimated to be of the
order of 1 G.
Key words: gamma-rays: observations - stars: pulsars: individual: PSR B1259-63
PSR B1259-63 / SS 2883 is a binary system consisting of a
48 ms pulsar in orbit around a massive B2e companion star
(Johnston et al. 1992b,a). The highly eccentric orbit of the
pulsar places it just
from the companion during
periastron every
3.4 years. Be stars are known to have
non-isotropic stellar winds forming an equatorial disk with enhanced
mass outflow (e.g. Waters et al. 1988). In the case of PSR B1259-63,
timing measurements suggest that the disk is inclined with respect to
the orbital plane (Wex et al. 1998), probably because the neutron star
received a substantial birth kick, causing the pulsar to cross the
disk two times near periastron. These unique properties make the
binary system PSR B1259-63 an excellent laboratory for the study of
pulsar winds interacting with a changing environment in the presence
of an extremely intense photon field. The synchrotron origin of
optically thin unpulsed radio emission detected from this source, in
particular during the periastron passage
(e.g. Johnston et al. 1999; Connors et al. 2002), indicates acceleration of
electrons to relativistic energies. The acceleration process has been
argued to be most efficient when the pulsar passes through the
equatorial disk (Ball et al. 1999). The basic features of such a system
("binary plerion'') in the context of higher energy X- and
-radiation components have been comprehensively discussed by
Tavani & Arons (1997).
The intense photon field provided by the companion star not only plays
an important role in the cooling of relativistic electrons but also
serves as a perfect target for the production of high energy
-rays through inverse Compton (IC) scattering
(Murata et al. 2004; Ball & Dodd 2001; Kirk et al. 1999; Tavani et al. 1996; Ball & Kirk 2000). Some of these
emission models predict wind-powered shock acceleration of electrons
to multi-TeV energies, radiating predominantly through the
synchrotron and IC channels, with the main energy release in the X-
and high energy
-ray bands, respectively.
The unpulsed non-thermal X-ray emission detected from PSR B1259-63
throughout its orbital phase in 1992 to 1996 by the ROSAT and
ASCA satellites (Cominsky et al. 1994; Hirayama et al. 1996; Kaspi et al. 1995) generally
supports the synchrotron origin of X-rays. The spectrum of the
synchrotron radiation seems to extend to hard X-rays/low energy
-rays as shown by OSSE (Grove et al. 1995) and recently
confirmed by observations with the INTEGRAL satellite
(Shaw et al. 2004).
Since the companion star provides the dominant source of photons for
IC scattering, the target photon density is well known throughout the
entire orbit. Therefore, the ratio of X-ray flux to high energy
-ray flux depends only on the strength of the ambient magnetic
field. Although the latter can be estimated within a general
magneto-hydrodynamic treatment of the problem, it contains large
uncertainties which affect the estimate of the IC
-ray flux as
.
Kirk et al. (1999) studied the light curves of very-high-energy (VHE)
-rays under the assumption of a 1/r dependence of the
magnetic field which implies that the ratio of the energy density of
the photon field to that in the magnetic field B is independent of
orbital phase. They also assumed that the position of the termination
shock as well as the strength of the magnetic field is not affected by
the disk of the B2e star. Under such assumptions, they predicted an
asymmetric
-ray light curve with respect to periastron
(because of the inclination of the orbit with respect to the line of
sight and the dependence of the inverse Compton
-ray
emissivity on the scattering angle), with an increase towards
periastron and monotonic decrease after the passage of
periastron. However, one might possibly expect significant deviation
from such a simplified picture given the apparent strong impact of the
disk on the pulsar wind termination as seen in the X-ray light curve
(Tavani & Arons 1997). Moreover, during the time periods of interaction of
the pulsar wind with the equatorial disk one may expect, in addition
to the IC
-rays, a new component of
-radiation
associated with interactions of accelerated electrons and possibly
also protons with the dense ambient gas (Kawachi et al. 2004). Up to now,
the theoretical understanding of the properties of this complex
system, involving pulsar and stellar winds interacting with each
other, is quite limited because of the lack of constraining
observations.
Nevertheless, the fortunate combination of: (1) the high spin-down
luminosity of the pulsar,
,
which is
partially converted into populations of ultra-relativistic particles,
(2) the presence of the intense target photon field provided by the
companion star with energy density
(where R(t) is the spatial separation
between the pulsar and the companion star), and (3) the relatively
small distance to the source (
)
makes this
object a very attractive candidate for VHE
-ray emission.
Previous observations of PSR B1259-63 in VHE -rays, apart
from its periastron passage were performed using the CANGAROO I
and CANGAROO II detectors, but did not result in significant
signals and provided upper limits at
of the flux from the Crab
Nebula (see Kawachi et al. 2004, and references therein). The first
significant detection in VHE
-rays based on preliminary
analysis results was reported in Beilicke et al. (2004) a few days prior to the
periastron passage in February 2004 by the High Energy Stereoscopic
System (HESS) to allow for target of opportunity observations of
other instruments. The analysis of the data from this initial
detection and the data obtained in the subsequent HESS observation campaign on PSR B1259-63 is presented in this paper.
The observations from February to June 2004 were performed with the
High Energy Stereoscopic System (HESS), consisting of four imaging
atmospheric Cherenkov telescopes (Hinton 2004) located in Namibia, at
S
E in 1800 m above sea level. Each
telescope has a tesselated spherical mirror with 13 m diameter and
area (Cornils et al. 2003; Bernlöhr et al. 2003) and is equipped with a
camera of 960
-photomultiplier tubes providing a total
field of view of
in diameter (Vincent et al. 2003). During the
stereoscopic observations, an array trigger requires the simultaneous
detection of air-showers by several telescopes at the hardware level,
allowing a suppression of background events (Funk et al. 2004).
All observations were carried out in moonless nights tracking sky
positions with an alternating offset of typically
in
declination relative to the source (the wobble mode) in time
intervals of 28 min duration. This allows to determine the
background from the same field of view and one can omit off-source
observations, effectively doubling the observation time. Due to the
serendipitous discovery of another source in the field of view around
PSR B1259-63 (see Aharonian et al. 2005a), for all observations
subsequent to the 14th of May 2004 (MJD 53 139) the array pointing was
changed to a position
0.6
north of PSR B1259-63 and an
alternating wobble offset of 0.5
in right ascension (instead
of declination) was used.
The data set, selected on the basis of standard quality criteria, has
a dead time corrected exposure (live time) of 48.6 h and a mean
zenith angle of 42.7
.
The corresponding mean threshold energy
defined by the peak
-ray detection rate for a source with a
Crab-like spectrum (Aharonian et al. 2000) after selection cuts was
estimated to be 380 GeV. In the phase prior to the periastron passage
(7.8 h live time), due to technical problems, data from only three
telescopes were considered, and for the post-periastron phase (41.9 h
live time) data from the full telescope array were used.
After the calibration of the recorded air shower data
(Aharonian et al. 2004), each telescope image was parametrised by its
centre of gravity and second moments (Hillas 1985) followed by the
stereoscopic reconstruction of the shower geometry providing an
angular resolution of
for individual
-rays.
The
-ray energy was estimated from the image intensity and the
shower geometry with a typical resolution of
.
In order to
reject the vast background of cosmic-ray showers,
-ray
candidates are selected using cuts on image shape scaled with their
expectation values obtained from Monte Carlo simulations. The cuts
used for this analysis were optimised on simulations of a
-ray
point source with a flux level of 10% of the Crab Nebula VHE
-ray flux, allowing 41.2% of the
-rays to be retained
while rejecting more than 99.9% of the cosmic-ray air showers. A more
detailed description of the analysis techniques can be found in
Aharonian et al. (2005b).
Figure 1 shows the distribution of the squared
angular distance
of excess events relative to the position
of PSR B1259-63 for the whole data set. The circle with a radius
corresponding to the angular cut
in the field of view
around the source position was considered as the on-region. The
background was estimated from several non-overlapping circles of the
same radius (off-regions) with the same angular distance from
the camera centre, allowing corrections due to the varying camera
acceptance to be omitted (Aharonian et al. 2001). The clear excess in the
direction of the pulsar has a significance of 13.8
and is
consistent with a distribution obtained from a simulated
-ray
point source. The
-ray signal from the direction of
PSR B1259-63 was detected in most of the darkness periods from
February to June 2004, for which the results are summarised in
Table 1.
![]() |
Figure 1:
Main figure: distribution of background-subtracted ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Open with DEXTER |
Table 1:
Results of HESS observations on PSR B1259-63. For each
darkness period the number of telescopes used in the analysis
,
the live time
,
the
significance S calculated according to Li & Ma (1983), the number of
counts within the on-source (
)
and off-source
(
)
region(s), the background normalisation
,
and the number of detected
-rays (
)
are
listed. The background normalisation
was determined by the
number of off-regions, which do not coincide with other
-ray
sources in the field of view. Therefore
is not the same for
all data subsets because it depends on the distribution of wobble
offsets used when obtaining the data.
A 2D-analysis of the HESS field of view around PSR B1259-63
was performed using the ring background technique as alternative
background estimation method. In this method, for each bin, the number
of on-events
was derived by integrating the bin
content within a circle of radius
(on-region). The number of background events
was
estimated from a ring around the bin position with mean radius
and an area 7 times larger than the area of the
on-region. The normalisation
was corrected for the decrease
of the radial acceptance of the cameras towards the edge of the field
of view. Figure 2 shows the significance sky-map of a
field of view around PSR B1259-63 for the February
data. The excess at the pulsar position
(
,
with a significance
)
is consistent with the excess
given in Table 1. The additional broad excess
north of the pulsar is the unidentified TeV source
HESS J1303-631 discovered in the same field of view
(Aharonian et al. 2005a). The resulting bias in the background estimation due
to both sources was corrected in the analysis by excluding events from
a circle with radius
around each source from the background
estimation.
In order to derive the position of the pulsar excess and to check for
possible source extension, the data was reanalysed using hard cuts
requiring a minimal camera image intensity of 200 photo-electrons
which significantly improves the angular resolution and drastically
reduces the cosmic ray background at the expense of a higher energy
threshold of 750 GeV. The uncorrelated two dimensional excess
distribution for the whole data set was fitted assuming a radially
symmetric, Gaussian source intensity profile
![]() |
Figure 2:
Significance sky-map centred on the position of PSR B1259-63 for
the HESS February data. The excess ![]() ![]() ![]() ![]() |
Open with DEXTER |
The measured differential energy spectrum derived from all darkness
periods with a significant detection of the source (February to May)
is shown in Fig. 3. A power-law fit
![]() |
Figure 3:
Differential energy spectrum ![]() ![]() ![]() |
Open with DEXTER |
![]() |
Figure 4:
Differential energy spectra of ![]() ![]() |
Open with DEXTER |
For each darkness period for which the significance of the observed
-ray excess exceeded
(February, March, and April) a
differential spectrum was derived (see Fig. 4) and the
results of the corresponding power law fit according to
Eq. (1) are listed in Table 2. Within
statistical errors, there is no indication of time variability of the
photon index. However, the changing flux normalisation F0 indicates
source flux variability (see next section). Systematic errors were
estimated to be
and
,
dominated by the precision of the energy calibration of the
instrument and variations in the atmospheric extinction of the
Cherenkov light.
The daily integral flux of -rays above the mean threshold of
380 GeV is shown in Fig. 5 (lower panel). Since
the spectrum cannot be derived on a daily basis due to limited
statistics, the integral flux
was
obtained by integrating Eq. (1) assuming the spectral
index of the time-averaged spectrum
.
The flux
normalisation F0 was calculated using the measured excess of
-rays
given by
The daily light curve in Fig. 5 clearly indicates a
variable flux. This can be quantified by a fit of a constant flux to
the data yielding a
of 90.9 per 35 degrees of freedom
corresponding to a
probability of
.
Table 2:
Parameters of the power law fit to the differential spectrum for the
different darkness periods of HESS observations on
PSR B1259-63. For the periods May and June 2004, no spectrum could
be derived due to insufficient statistics. Shown are the photon index
and the flux normalisation
(with statistical
error only), the
per number of degrees of freedom
and the
probability
for
the power law fit of the spectrum, and the corresponding mean
threshold energy
(rounded to 10 GeV).
![]() |
Figure 5:
VHE ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Open with DEXTER |
In order to investigate the flux trend in each darkness period, the
individual light curves were fitted separately by a straight line and
the significance for an increasing or decreasing flux is listed in
Table 3. The low flux state after periastron is followed
by a distinct rise beginning at
days and a slow
decrease until
days where the observed excess is no
longer significant.
Table 3:
Light curve properties of the HESS 2004 data. The integrated flux
F(>
was calculated assuming a photon index
(see text). For the daily integral flux light curve of each
period the following properties are given: the
/
for a fit to a constant flux, and the slope a and significance
/
of an increasing or declining flux after
fitting a straight line
(
is the Modified Julian Date), together with the corresponding
.
The upper panel of Fig. 5 shows the light curve of
the transient unpulsed radio emission obtained from 2004 observations
of PSR B1259-63 (Johnston et al. 2005) for the same time range as for
the TeV band. There is some correlation visible between the radio and
the TeV bands, especially a low flux state around periastron and a
high flux state at days. The pulsed radio emission from
the pulsar was eclipsed during the periastron passage
(Johnston et al. 1992a), interpreted as due to scattering in the dense disk
material and allowing the dates of the disk crossings to be estimated,
in particular
days and
days for the 2004
periastron passage (Johnston et al. 2005). The observed dates of enhanced
VHE
-ray flux roughly correspond to orbital phases a few days
after the assumed disk crossings. However, the coverage of the
considered time interval by HESS observations is limited and does
not allow firm conclusions about VHE
-ray flux maxima,
especially pre-periastron.
The detection of VHE -rays from PSR B1259-63 gives the
first unambiguous and model-independent evidence of particle
acceleration to multi-TeV energies in this unique binary system. The
(most likely) synchrotron origin of X-rays (e.g. Tavani & Arons 1997)
also indicates the presence of ultra-relativistic particles, in the
form of electrons, but other possible explanations of the X-radiation,
e.g. due to the bulk motion Comptonisation of the stellar photons by
the pulsar wind with a moderate Lorentz factor of 10-100
(Chernyakova & Illarionov 1999), cannot be firmly excluded.
In contrast to the X-ray emission, the extension of the -ray
spectrum to several TeV necessarily implies that the parent particles,
electrons and/or protons, are accelerated to at least 10 TeV. The
point-like feature of the detected signal constrains the extension of
the
-ray production region to
.
The information contained in the time
variability of the signal on timescales of days provides a hundred
times more stringent upper limit on the size of the
-ray
source of about
.
The most likely scenario of particle acceleration and radiation in this system is a variation of the standard model of Pulsar Driven Nebulae (see e.g. Rees & Gunn 1974; Kennel & Coroniti 1984) which postulates that the deceleration of the ultra-relativistic pulsar wind (with a bulk motion Lorentz factor of 106-107) by the pressure forces of the external medium enforces a termination shock. In this context electrons are accelerated to multi-TeV energies. These electrons radiate in the magnetic and photon fields in which they propagate and thus produce synchrotron and Inverse-Compton nebulae with a typical size exceeding 0.1 pc, commonly called plerions.
The mean energy flux contained in the VHE -ray emission
derived from the time-averaged energy spectrum is
which represents a
-ray
luminosity of
for
an assumed distance of the binary system of 1.5 kpc. This
corresponds to
of the pulsar spin-down
luminosity. Interestingly, this value matches the typical ratio
between the pulsed and unpulsed component of X-ray radiation from
isolated pulsars with associated pulsar windnebulae and their
spin-down luminosity (see e.g. Cheng et al. 2004), supporting the
suggestion that the energy for the observed
-rays is provided
by the pulsar.
However, in the case of PSR B1259-63 both the spatial and temporal scales are expected to be rather different from those in plerions around isolated pulsars. Also, dynamical flow effects can play a different role since the pulsar moves through the stellar wind and the photon field of the massive companion star. Due to the high pressure of the stellar wind, the pulsar wind terminates very close to the pulsar and thus the electrons are accelerated well within the binary system. Moreover, because of severe adiabatic and radiative losses, the TeV electrons have a very short lifetime. As a result, the radiation is emitted in a rather compact region not far from the acceleration site, and for any given time the emission originates from a quite short sector of the pulsar trajectory.
In the following we summarise the physical processes expected to be at work, and put our observational results into this perspective.
The synchrotron loss time
Indeed, the inverse Compton losses are dominated by the scattering of
electrons on the light of the companion star with luminosity
and effective blackbody
temperature
.
At any distance R,
the black-body radiation decreases by a factor of
,
where
is the radius of the
companion star. Thus at the distance of the pulsar to the star
,
the energy density of the radiation is
![]() |
Figure 6:
Sketch of the orbit of PSR B1259-63 with respect to the line of
sight (adapted from Johnston et al. 1999). The pulsar approaches the
equatorial disk prior to periastron while it is "behind'' the
companion star and turns towards the observer before it crosses the
disk after periastron for the second time. Note that the orientation
of the disk with respect to the orbital plane is not precisely known.
The colour gradient bars along the orbit indicate the periods of
HESS observations and show the integral VHE ![]() |
Open with DEXTER |
The expansion of the relativistic, shocked pulsar wind gas will
proceed with a flow velocity of magnitude
,
roughly
equal to the relativistic magnetosonic speed
,
where
,
over a
distance
,
where
is a few
.
The factor
ranges between
and 1, if the direction of
the postshock MHD flow ranges between a perpendicular and a parallel
orientation with respect to the magnetic field direction,
respectively. Assuming the pitch angle scattering mean free path of
the accelerated pulsar wind particles to be small compared to l,
their adiabatic loss time is then given by
![]() |
(5) |
In this picture the adiabatic losses proceed about as fast as the
synchrotron losses if
G (compare with
Eq. (2)) - while both are faster than the inverse Compton
losses - and the energy spectrum of the radiating electrons is close
to their source spectrum produced during acceleration at the
termination shock.
We shall consider neither electron Bremsstrahlung nor
-ray
production in interactions of relativistic protons with the ambient
gas via decay of secondary
-mesons
(see e.g. Kawachi et al. 2004) here, since the accelerated pulsar wind
particles are hydromagnetically constrained to the magnetic field
lines of the pulsar outflow and will hardly interact individually with
the stellar wind particles, except in a thin boundary layer at the
Wind interface far downstream.
The missing quantity in our cooling considerations is the value of B. We shall use the contemporaneous INTEGRAL results and our
HESS fluxes to estimate it as follows: Since there is little
doubt that the detected TeV -rays are produced in the
Klein-Nishina regime, we may conclude that they are emitted by
electrons of the same energy,
.
This
allows to set up a direct relation between
-rays and X-rays
produced by the same electrons,
An experimental estimate of B is possible by comparing the energy
fluxes of the detected X- and -rays. The post-periastron data
obtained in March by RXTE and INTEGRAL in the X-ray band
from 1 keV to 100 keV (see Shaw et al. 2004) show a relatively flat
spectral energy distribution
/
at
the level of
.
The VHE
-ray flux for the same period was
.
This implies that in the TeV energy regime the
synchrotron losses of electrons proceed an order of magnitude faster
than the inverse Compton losses. Therefore, from the comparison of the
synchrotron cooling time, given by Eq. (2), with the
Klein-Nishina cooling time given by Eq. (4), and taking into
account that two weeks after the periastron the separation between the
pulsar and the companion star was approximately
,
we find
Let us assume that the accelerated electrons enter the -ray
production region with a rate
.
As a consequence, in the case of dominant
synchrotron or inverse Compton losses in the Thompson regime, a
steeper electron spectrum is established in the
-ray
production region with power-law index
,
while
dominating adiabatic (or Bremsstrahlung losses) will not change the
initial electron spectrum, i.e.
,
both for
.
Let us conversely assume that the radiating electrons have a power law
distribution
.
Then the inverse Compton scattering in the
Klein-Nishina regime leads to a
-ray spectrum
(Blumenthal & Gould 1970)
![]() |
Figure 7:
Spectral energy distribution from Kirk et al. (1999) assuming the electron
cooling to be dominated by radiative (dashed lines) or adiabatic
losses (solid lines) for a magnetic field strength of 0.32 G at an
orbital phase 12 days prior to periastron. The spectra were calculated
by Kirk et al. (1999) for different electron injection spectra with the
power law indices
![]() ![]() |
Open with DEXTER |
![]() |
Figure 8:
Comparison of the light curves predicted by different models of VHE
![]() |
Open with DEXTER |
The relevant photon spectral index range
in the
keV energy range from the 1994 ASCA
observations (Hirayama et al. 1996), although not coeval, would also
indicate a rather hard spectrum of the radiating electrons
which would again be difficult to reconcile with
dominant synchrotron cooling. This 1994 synchrotron photon spectral
index is consistent with a preliminary spectral fit with
derived from the coeval INTEGRAL observations
(Shaw 2004). In this paper we shall ultimately rely on the
HESS spectrum.
Within the scheme of dominant adiabatic cooling, the strong
variability of the VHE -ray emission might be explainable by a
varying spatial confinement of the accelerated pulsar wind particles
by the kinetic and thermal pressure of the stellar mass outflow. In
Fig. 6 the variations of the measured VHE
-ray flux are illustrated within the context of the orbital
parameters and environment of the binary system with respect to the
line of sight. The enhanced mass outflow in the equatorial disk will
result in a more compact emission region - and therefore weaker
adiabatic cooling - during the pulsar disk passages compared to the
other regions of the orbit, in particular the periastron phase, where
the expected stronger adiabatic cooling could lead to a minimum of the
-ray flux. A second flow aspect would be a Doppler modulation
due to the relativistic bulk flow velocity away from the stagnation
point of the shocked pulsar wind. Considering Fig. 6,
this could again lead to a flux minimum during periastron. However, a
detailed consideration of this overall physical picture is beyond the
scope of the present paper.
A minimum in the X-ray flux, observed for previous periastron passages, has been interpreted as being due to increased IC losses (Tavani & Arons 1997). However, HESS measurements indicate a minimum in the IC flux around periastron, which would rule out such an interpretation for this periastron passage. Unfortunately, due to the full moon and bad weather conditions we were not able to take data during the periastron passage of the pulsar.
The predicted pre-periastron spectrum from Kirk et al. (1999) is compared
to the corresponding HESS data in Fig. 7. The model
curves, taken from Kirk et al. (1999) (Figs. 2, 6), were computed for the
two scenarios of dominant radiative losses and dominant adiabatic
losses, and matched the archival X-ray data for electron injection
spectra with
and
,
respectively. They
agree well with the TeV data, assuming a magnetic field strength of
B=0.32 G supporting the estimate of the order of 1 G obtained from
the HESS data. Despite the spectral agreement, the models fail to
describe the TeV light curve as can be seen in
Fig. 8 (dashed and dotted lines). In particular,
neither the minimum around nor the high flux states before and after
periastron are described by the model. This is perhaps not surprising,
because the model assumed that the ratio between the energy densities
of the magnetic field and the photon field does not change throughout
the full orbit, and that the effects related to the pulsar passage
through the stellar disk can be neglected.
Interestingly, the observed light curve seems to be qualitatively
similar to the prediction made by the model of
Kawachi et al. (2004) which is also shown in Fig. 8
(solid line). In this model, hadronic interactions and production in the misaligned stellar disk plays a dominant role in the
-ray production mechanism. However, present data do not allow
safe conclusions concerning the interpretation of the TeV light
curve. Studies of the most interesting part of the pulsar orbit have
to be postponed until multi-wavelength observations during periastron
become possible.
In summary, the detection of VHE -ray emission from the binary
pulsar PSR B1259-63 by HESS provides the first
model-independent evidence of particle acceleration in this
object. The results clearly demonstrate the power of
-ray
observations for the study of the properties and the nature of high
energy processes in this unique cosmic accelerator. Further
observations at VHE
-ray energies are necessary to derive
spectra on a daily basis also during periastron passage.
Acknowledgements
The support of the Namibian authorities and of the University of Namibia in facilitating the construction and operation of HESS is gratefully acknowledged, as is the support by the German Ministry for Education and Research (BMBF), the Max Planck Society, the French Ministry for Research, the CNRS-IN2P3 and the Astroparticle Interdisciplinary Programme of the CNRS, the UK Particle Physics and Astronomy Research Council (PPARC), the IPNP of the Charles University, the South African Departmentof Science and Technology and National Research Foundation, and by the University of Namibia. We appreciate the excellent work of the technical support staff in Berlin, Durham, Hamburg, Heidelberg, Palaiseau, Paris, Saclay, and in Namibia in the construction and operation of the equipment. We also thank S. Shaw for informations on the photon spectral index from the INTEGRAL X-ray observations.