2. Observations and analysis
We started an optical observational campaign on the GRB 970228 error box in order to search for possible optical counterparts of the -ray event, beginning 15.5 hours after the burst (Guarnieri et al. 1997). Owing to the large initial error box (15' in radius; Costa & Frontera 1997) it was not possible to cover it with a single CCD frame; we succeeded to locate the "right" field 16.5 hours after the event, i.e. on February 28.816 UT.
The frames were obtained with the 1.5-meter telescope of the Bologna University equipped with the BFOSC instrumentation (Merighi et al. 1994), which allows fast switching from the spectrographic to the imaging mode. Images have a scale of about 0.5 arcsec pixel-1. The field was observed on February 28 and on March 1, 3, 4, 5, 12, 13 and 18, 1997. R and B filters were employed, with exposure times spanning from 30 to 60 minutes. The limiting magnitude strongly depended on the seeing and on the sky conditions, which were fairly good on February 28, March 3, 4 and 12 (typical FWHM of the PSF: 4 pixels). On March 1, 5, 13 and 18 the bad seeing and poor sky conditions implied frames with a low signal-to-noise ratio; therefore, they gave us only lower limits for the magnitudes of the object.
B, V and R images of the Selected Area 101 (Landolt 1992) were also obtained on March 4 in order to calibrate the field of GRB 970228. After the standard cleaning procedure for bias and flat field, the frames were processed with the DAOPHOT II package (Stetson 1987) and the ALLSTAR procedure inside MIDAS. We used simple aperture photometry when the objects were too faint to be detected with DAOPHOT II. Then, the fields were calibrated with the standard stars quoted above.
We detected an OT at the intersection of the error boxes of GRB 970228 and SAX J0501.7+1146 with coordinates , (equinox 2000.0; error: 5"). Its position makes it practically coincident with the proposed optical counterpart of GRB 970228. Its magnitude changed from and on February 28.8 UT to on March 3.8 UT and on March 4.8 UT. In the days following March 3.8 the R magnitude of the object seemed to remain more or less constant. Table 1 shows the measured B and R magnitudes. The reported values are the integrated magnitudes of the OT plus a nearby star and an extended object (van Paradijs et al. 1997). The latter is possibly the host of the OT, 0".2 away (van Paradijs et al. 1997, Groot et al. 1997c) and with (Groot et al. 1997c, Metzger et al. 1997a), whereas the nearby object is an early M-type star (van Paradijs et al. 1997, Groot et al. 1997c) or a mid K-type star (Tonry et al. 1997), located 2".7 away, constant in brightness with (Metzger et al. 1997a) and unrelated to the OT (Groot et al. 1997c). Due to the seeing conditions, we were not able to separate the OT from the star.
Table 1. Measured B and R magnitudes of the OT, possible counterpart of GRB 970228, plus the nearby red star and the extended object. The dates are computed at the times of mid-exposure
In our frames the red star is more likely responsible for the observed residual emission already since March 3. Actually, the R magnitude of the object after that day, within the errors, is the same as the red star located near the extended source as indicated by Metzger et al. (1997a). This means that the main optical effects of the -ray explosion (rise and first decay phase) developed before March 3.8. Indeed, observations made on March 6.3, 11.2 (Metzger et al. 1997a), 9.9 and 13.0 (Groot et al. 1997c) show that the underlying extended object is constant in brightness. Therefore, within the accuracy of the measurements, the time span of 3.6 days is the upper limit to the duration of the brightest phase of the point-like OT, as Table 1 shows. Our observations are consistent with the fading object found by Groot et al. (1997a) and van Paradijs et al. (1997) with and on March 1.0 UT and and on March 8.9 UT. According to van Paradijs et al. (1997) and Metzger et al. (1997a), it is 0".2 from a quiescent object, which appears to be extended and therefore likely to be a galaxy which could be associated with the X-ray transient and, presumably, with the GRB.
We could now subtract from the total fluxes of the optical event those of the M-type star and of the extended object. Since the constant red star and the underlying extended object were always undetectable in our B band photometry, we can assume that the B magnitude on February 28.8 is representative of the transient event. We obtain B = 22.4 and R = 21.6, implying = 0.8 as the color index of the OT 17 hours after the -ray event. On March 3.8 we deduce that the OT was fainter than R = 23. Van Paradijs et al. (1997) report a fading mag in the time interval March 1.0-March 4.9. Pedichini et al. (1997) observed a fading of 2.7 mag between February 28.8 and March 4.8.
© European Southern Observatory (ESO) 1997
Online publication: March 26, 1998