TITLE:   GCN GRB OBSERVATION REPORT
NUMBER:  3293
SUBJECT: GRB050326: analysis of the XMM-Newton observation
DATE:    05/04/19 12:45:48 GMT
FROM:    Andrea De Luca at IASF-CNR,Milano  <deluca@mi.iasf.cnr.it>

Andrea De Luca (IASF Mi), Diego Gotz (IASF Mi), Sergio Campana (OAB)
on behalf of a larger collaboraton report:

We have analyzed the data from the XMM-Newton observation of
GRB050326, discovered by Swift on 2005, Mar 26, 09:53:55 UTC
(Markwardt et al., GCN3143).

The XMM-Newton observation started on 2005, Mar. 26 at 18:25 UTC
and lasted for ~45.8 ks.

A high particle background level affects the whole observation,
with several soft proton flares, particularly evident in pn data.

We improved the astrometry of the XMM-Newton/EPIC images
by matching X-ray sources in the field to stars in the USNO-B1
catalogue. The refined position (J2000) for the X-ray afterglow is

RA: 00h 27m 49.1s  Dec: -71d 22' 16.3"

The 1 sigma error radius is 1.5 arcsec (including the rms error on
the cross-correlation as well as systematic uncertainties in the
optical catalogue). The position is fully consistent with the XRT
coordinates (Moretti et al., GCN3147) as well as with the preliminary
XMM-Newton position by Ehle & Perez-Martinez (GCN3149).

The MOS light curve (0.3-8 keV) is well fitted by a power law decay
with index delta=1.81+/-0.11 (90% confidence level), with a reduced
chi2 of 0.6 (21 dof). Adding pn data (with a larger bkg contamination)
yields fully consistent results (delta=1.85+/-0.10), but with a poorer
fit quality (owing to bkg fluctuations).

We extracted time-averaged spectra from the three EPIC cameras and
we generated ad-hoc response files. We quote here errors at 90%
confidence level for a single interesting parameter unless otherwise
specified.
A simultaneous fit with an absorbed power law model yields a reduced
chi2 of 1.11 (241 dof). The resulting NH=(1.20+/-0.15)x10^21 cm^-2
is higher than the expected Galactic value in the burst direction
(NH=4.5x10^20 cm^-2, Dickey & Lockman, 1990); the best fitting power
law photon index is Gamma=2.10+/-0.05.

A better fit (reduced chi2=1.05, 240 dof) may be obtained
fixing the NH to the expected Galactic value (NH=4.5x10^20 cm^-2) and
adding a neutral, redshifted absorber component to the spectral model.
With a simple F-test we evaluate the chance occurrence probability of
the improvement to be of 2.6x10^-4. Such model yields an intrinsic
NH>3x10^21 and a redshift z>1 (90% confidence level for 2 parameters
of interest). However, neither the intrinsic column density nor the
redshift are well constrained towards large values, owing to the lack
of absorption edges in the spectrum.
The resulting power law photon index is Gamma=2.01+/-0.05.
The observed flux is of 5.2x10^-13 erg cm^-2 s^-1 in 0.2-10 keV;
the corresponding unabsorbed flux is of 7.4x10^-13 erg cm^-2 s^-1.

Finally, we divided the dataset into two time intervals of ~15900 s and
~27600 s (each containing about half of the counts from the afterglow)
and we repeated the spectral analysis. We found no significant spectral
changes in the two considered intervals.

This message may be cited.