A&A 411, L327-L330 (2003)
DOI: 10.1051/0004-6361:20031075

Time resolved spectroscopy of GRB 030501 using INTEGRAL[*]

V. Beckmann1,2 - J. Borkowski2 - T. J.-L. Courvoisier2,3 - D. Götz4 - R. Hudec5 - F. Hroch2,5 - N. Lund6 - S. Mereghetti4 - S. E. Shaw2,7 - A. von Kienlin8 - C. Wigger9


1 - Institut für Astronomie und Astrophysik, Universität Tübingen, Sand 1, 72076 Tübingen, Germany
2 - INTEGRAL Science Data Centre, Chemin d'Écogia 16, 1290 Versoix, Switzerland
3 - Geneva Observatory, 51 chemin des Maillettes, 1290 Sauverny, Switzerland
4 - Istituto di Astrofisica Spaziale e Fisica Cosmica, CNR v. Bassini 15, 20133 Milano, Italy
5 - Astronomical Institute, Academy of Sciences of the Czech Republic, 25165 Ondrejov, Czech Republic
6 - Danish Space Research Institute, Juliane Maries Vej 30, 2100, Copenhagen, Denmark
7 - School of Physics and Astronomy, University of Southampton, Southampton, SO17 1BJ, UK
8 - Max-Planck-Institut für extraterrestrische Physik, Giessenbachstrasse, 85748 Garching, Germany
9 - Paul Scherrer Institut, 5232 Villingen, Switzerland

Received 28 May 2003 / Accepted 26 June 2003

Abstract
The gamma-ray instruments on-board INTEGRAL offer an unique opportunity to perform time resolved analysis on GRBs. The imager IBIS allows accurate positioning of GRBs and broad band spectral analysis, while SPI provides high resolution spectroscopy. GRB 030501 was discovered by the INTEGRAL Burst Alert System in the ISGRI field of view. Although the burst was fairly weak (fluence $F_{20-200 ~{\rm keV}} \simeq 3.5 \times 10^{-6}~\rm {erg~cm}^{-2})$ it was possible to perform time resolved spectroscopy with a resolution of a few seconds. The GRB shows a spectrum in the 20-400 keV range which is consistent with a spectral index $\Gamma = -1.8$. No emission line or spectral break was detectable in the spectrum. Although the flux seems to be correlated with the hardness of the GRB spectrum, there is no clear soft to hard evolution seen over the duration of the burst. The INTEGRAL data have been compared with results from the Ulysses and RHESSI experiments.

Key words: gamma rays: bursts - gamma rays: observations

1 Introduction

Gamma Ray Bursts (GRBs) were discovered by chance in the late 1960s by the Vela experiments (Klebesadel et al. 1973). They have been proven to be extragalactic in origin after a successful identification of an X-ray afterglow by BeppoSAX (GRB 970508; Piro et al. 1998) with an optical counterpart at redshift z = 0.835 (Metzger et al. 1997). Even though we are now rather confident that long GRBs are related to massive explosions in distant galaxies, there are still many open questions remaining. First, whether GRBs are related to Supernova explosions, and also, what the connection to the star formation phenomenon is. Another crucial point is the exact mechanism by which GRBs can produce an energy output of ${>}10^{52}~\rm ergs$ (under the assumption that the emission is isotropic, which is probably not true). Prompt observation of GRBs in several energy ranges is essential to obtain high quality data for the study of these rapidly fading objects. Although GRBs were not one of the main targets for the scientific program of INTEGRAL (Winkler et al. 2000), the two main gamma-ray instruments, the imager IBIS (Ubertini et al. 1999) and the spectrometer SPI (Vedrenne et al. 1999), offer great capabilities for studying the prompt emission of GRBs when they occur in the field of view. Since the field of view is about 29 degrees, one gamma-ray burst per month is expected to be observed. This rate has been confirmed so far by the six bursts in the field of view between November 2002 and May 2003: GRB 021125 (Bazzano & Paizis 2002), GRB 021219 (Mereghetti et al. 2002), GRB 030131 (Borkowski et al. 2003), GRB 030227 (Mereghetti et al. 2003a), GRB 030320 (Mereghetti et al. 2003b), and GRB 030501 (Mereghetti et al. 2003c). In addition, the anticoincidence shield (ACS) of SPI can be used as an all-sky monitor for GRBs (von Kienlin et al. 2003).

During the last three bursts in the field of view, both SPI and IBIS, were in full operational mode, allowing time resolved spectral analysis.

  
2 INTEGRAL observation

GRB 030501 was detected on 1st May 2003 at 03:10:18 UT with data from the low energy detector of the IBIS instrument, the Integral Soft Gamma Ray Imager (ISGRI; Lebrun et al. 2001), which consists of $128 \times 128$ CdTe crystals sensitive in the energy range $15{-}300 ~{\rm keV}$. ISGRI uses the coded mask technique and offers an instrumental resolution of ${\sim}12$ arcmin over the field of view of $29^{\circ} \times 29^{\circ}$. The source location precision depends on the brightness of the source, and is about 1 arcmin for sources with a detection significance of $10 \sigma$. The GRB was detected in the ISGRI data by the INTEGRAL Burst Alert System (IBAS; Mereghetti et al. 2001), which automatically determines the position and time of events which occur in the IBIS field of view. The IBAS alert was distributed approximately 30 s after the burst start time with a positional uncertainty of only 4 arcmin (Mereghetti et al. 2003c). The ISGRI lightcurve, is shown in Fig. 1.


  \begin{figure} \par\includegraphics[clip, width=6cm,angle=0]{INTEGRAL3_f1.ps} \end{figure} Figure 1: Top panels: ISGRI lightcurves in the soft and hard band, respectively. IBAS triggered the GRB alert, within half a minute from the burst start, based on the ISGRI data. Bottom panel: hardness ratio evolution derived from the ISGRI countrates.
Open with DEXTER

A later analysis, performed off-line, on the ISGRI data revealed a position of (J2000.0) $\rm 19^h 05^m 30^s$, $\rm +6^\circ 18^m 26^s$with an uncertainty of ${\sim}3$ arcmin (Mereghetti et al. 2003c). The gamma-ray burst had a duration of 40 s. Data from the spectrometer SPI were also analysed at the INTEGRAL Science Data Centre (ISDC; Courvoisier et al. 1999; Beckmann 2002). SPI is designed for high spectral resolution (FWHM of 2.5 keV at 1 MeV) in the energy range $20{-}8000 ~{\rm keV}$, which is achieved by 19 cooled germanium detectors. A coded mask with 128 elements provides an instrumental spatial resolution of 2.8 degrees. The position accuracy for point sources can be <5 arcmin (for a source with S/N = 100). The position for GRB 030501 extracted from the SPI data is (J2000.0) $\rm 19^h 07^m$, $\rm +6^\circ 25^m$ ( $\pm 21 ~ \rm {arcmin}$), which is 19 arcmin off the position detected by ISGRI. As the GRB falls into the partially coded field of view, the low position accuracy of SPI is not surprising, as only part of the detector plane (5 out of 19 detectors) is in fact illuminated by the event.

The peak gamma ray flux, $f_{20-200 ~{\rm keV}}$, in the 20-200 keV band measured by SPI in a 2 s bin starting at 03:10:20 UTC is $\rm 2.8 \pm 0.4 ~photons~ cm^{-2}~ s^{-1}$. This is consistent with ISGRI where $f_{20-200 ~{\rm keV}} = \rm 2.7 \pm 1.2 ~photons~ cm^{-2}~ s^{-1}$ was measured in a 1 s bin starting at 03:10:18.4 UTC, reaching the peak about 1 s before it occurs in the SPI data. The fluence, $F_{20-200 ~{\rm keV}}$, measured for the GRB by both instruments in the same band and integrated over the full burst visibility period is also consistent:

\begin{eqnarray*}{\rm SPI}\!\!: F_{20-200 ~{\rm keV}} &=& 39.3 \pm 2.5 \rm {~~ph... ...hotons~cm}^{-2}\\ &=& 3 \pm 1 \times 10^{-6}~\rm {erg~cm}^{-2} \end{eqnarray*}


The GRB was also observed by the Ulysses experiment (Hurley et al. 2003). Due to the weakness of the detection the reported fluence, in the 25-100 keV band, is uncertain by about a factor of two, but this is still consistent with the measurements made by the INTEGRAL instruments:

\begin{eqnarray*}{\rm SPI}\!\!: F_{25-100 ~{\rm keV}} &\simeq & 2.0 \times 10^{-... ...-100 ~{\rm keV}} &\simeq & 1.1 \times 10^{-6}~\rm {erg~cm}^{-2}. \end{eqnarray*}


The ${\sim}$30% uncertainty on ISGRI fluence is dominated by systematic errors on the response of the instrument at large off-axis angles. We note however a good agreement with the SPI value, which confirms the value obtained with ISGRI. The overall spectrum of the burst based on SPI data is shown in Fig. 2. The GRB occurred in a pointed observation of 1800 s length. The background emission was estimated from this pointing, but excluding the time when the GRB occurred. The GRB is detectable up to at least 200 keV in both the ISGRI and SPI[*] data. A single power law represents the SPI data well, resulting in a photon index of $\Gamma = -1.88 \pm 0.10$. A more complicated model (e.g. a broken power law or a Band model; Band et al. 1993) does not improve the fit, thus no spectral break is detectable. This result is consistent with the spectral slope of $\Gamma = -1.75 \pm 0.10$ derived from ISGRI data. For the ISGRI data the background can be estimated at the same time as the source flux, using the Pixel Illumination Function (PIF). Spectra are extracted computing one PIF for each energy bin (128 linearly spaced bins have been used).

There was a marginal detection of the GRB by the SPI-ACS (Hurley et al. 2003). The low countrate of this GRB is expected in the ACS data, as the effective area for events in the field of view of SPI is small for the ACS, which shields the spectrograph from the sides and from its back (von Kienlin et al. 2003). The combination of Ulysses and INTEGRAL data also allowed triangulation of the GRB event by the 3rd Interplanetary Network (IPN). The result is consistent with the ISGRI position (Hurley et al. 2003). The INTEGRAL X-ray (JEM-X) and optical (OMC) monitors were unable to provide any additional information since the GRB was located well outside of the respective fields of view of these instruments.


  \begin{figure} \par\includegraphics[clip,width=6.0cm,angle=-90]{INTEGRAL3_f2.ps} \end{figure} Figure 2: GRB spectrum in the range 20-400 keV, taken from the SPI data over 20 seconds after the burst occurance.
Open with DEXTER

3 Comparison with RHESSI observation of GRB 030501

GRB 030501 has also been seen by the spectrometer of the Ramaty High Energy Solar Spectroscopic Imager (RHESSI), which is a NASA Small Explorer satellite designed to study hard X-rays and gamma-rays from solar flares (Lin et al. 2002). The instrument consists of 9 germanium detectors, each of volume $300 ~ \rm cm^{3}$, that cover an energy range of 3 keV to 17 MeV, with an energy resolution of about 3 keV (FWHM) at 1 MeV (Smith et al. 2002). The detector uses a Rotation Modulation Collimator (RMC) system for high resolution imaging of solar flares. The germanium detectors are only lightly shielded. Above about 60 keV, they have a significant response to photons from any direction in the sky. Thus, RHESSI is a sensitive GRB detector, and as such it is part of the IPN.

The lightcurve of GRB 030501 as seen by RHESSI in the 40-120 keV band is shown in Fig. 3. The peak photon flux measured is $f_{70-200 \rm keV} \simeq 0.55 \pm 0.17 \rm ~photons~ cm^{-2} ~ s^{-1}$, and the fluence over the 20 s burst duration is $F_{70-200 \rm keV} \simeq 2.1 \pm 0.6 \times 10^{-6} ~ \rm erg ~ cm^{-2}$.

  \begin{figure} \par\epsfxsize=7.0cm \epsfbox{INTEGRAL3_f3.ps} \par\end{figure} Figure 3: Lightcurve of GRB 030501 measured by the spectrograph of RHESSI.
Open with DEXTER

4 Time resolved results on GRB 030501

Although the burst is comparably weak, the sensitivity of ISGRI and SPI allows the study of the lightcurve of the prompt emission. We show the SPI lightcurve in the same energy band as for RHESSI in Fig. 4. The peak flux is reached ${\sim}10$ s after the burst started. Spectra were extracted from the SPI data, in 5 logarithmically binned energy bands between 20 and 400 keV. For ISGRI the data have been binned in order to have at least 25 counts per bin. XSPEC 11.2 was used to fit a single power law to the data in time bins of 2 - 10 s over a period of 30 s after the burst started.


  \begin{figure} \par\epsfxsize=7.0cm \epsfbox{INTEGRAL3_f4.ps} \par\end{figure} Figure 4: Lightcurve of GRB 030501 in the same energy band as the RHESSI data, taken from INTEGRAL/SPI.
Open with DEXTER

The results are shown in Fig. 5. The spectrum starts apparently rather steep, but then immediately has a photon index of $\Gamma \simeq -1.5$ as the flux increases. Before the GRB is below the instrumental sensitity, it apparently softens again. In the ISGRI data there is evidence of hardness intensity correlation as seen in other GRBs before (e.g. Ford et al. 1995). This is consistent with the SPI data, though the statistic is not high enough to constrain the results. No clear spectral evolution is seen in the data. The hardness ratio evolution in the RHESSI data show a similar trend to the one seen in ISGRI (Fig. 1), though the error bars are larger.


  \begin{figure} \par\epsfxsize=7.0cm \epsfbox{INTEGRAL3_f5.eps} \par\end{figure} Figure 5: Evolution of the spectral slope of a single power law, fitted to the ISGRI (open squares) and SPI (filled circles) data of GRB 030501. The ISGRI data points have been shifted by +0.5 s for better visibility. The definition of the photon index is $f_{\nu } \propto \nu ^{\Gamma }$.
Open with DEXTER

5 Discussion

Optical follow up observation has not revealed an optical counterpart to this GRB (Boer & Klotz 2003; Fox 2003; Rumjantsev et al. 2003). Using the 1 m Wise telescope in comparison with POSS-II E photographic plates, the magnitude of the optical counterpart 16.5 hours after the prompt emission can be limited to $R \ge 20.0 \rm ~ mag$ (Ofek et al. 2003). Also analysis of optical observations carried out with the automatic $25 \rm ~ cm$ TAROT telescope shortly after the burst occurrence (i.e. within 15 min) shows no optical counter part with $R\le 18.0 \rm ~ mag$ (Klotz & Boer 2003).

With a Galactic latitude of only 0.2 degree and an estimated extinction of $E(B-V) \sim 15$, identification of an optical counterpart is indeed difficult, if not impossible.

As no break in the spectrum was detected in either the SPI or the ISGRI data, it can be assumed that the peak energy of this long GRB is either $E_{\rm peak} \mathrel{<\kern-1.0em\lower0.9ex\hbox{$\sim$ }}30 ~{\rm keV}$ or $E_{\rm peak} \mathrel{>\kern-1.0em\lower0.9ex\hbox{$\sim$ }}200~~{\rm keV}$. Since a very low peak energy is rather unlikely (see Preece et al. 2000), we assume a spectral break above 200 keV. GRB 030501 shows a similar spectral behaviour to bursts studied before (e.g. GRB 921207; Ghirlanda et al. 2002) but is about a factor ${\sim}10$ weaker than the bursts where time resolved spectroscopy has been possible with data from previous missions.

The comparison with the RHESSI data shows that this experiment is also a powerful tool in the detection and spectral analysis of GRBs. Especially for GRBs, which are not in the field of view of the INTEGRAL main instruments SPI and IBIS, RHESSI provides sufficient spectral and timing resolution ($16~\mu$s) to study those events, as the RHESSI spectrograph is a non-shielded all-sky monitoring instrument.

This GRB demonstrates the great capabilities of INTEGRAL and the software package, provided by the ISDC in collaboration with the instrument teams. The time lag between GRB occurrence and providing detailed spectral and timing analysis is less than half a day.

Acknowledgements
RH and FH acknowledge the support provided by the ESA Prodex Project 14527. JB was supported by the Polish grant 2P03C00619p02 from KBN and SS by PPARC grant GR/K/94867.

References


Copyright ESO 2003