Astron. Astrophys. 335, L5-L8 (1998)

3. Discussion

Prompt follow up observations are extremely important in order to monitor the GRB afterglow at early stages, which enables to test the validity of the different afterglow models. In fact, we know now that afterglows are not always remnants of the initial burst (see Piro et al. 1998). Observations performed shortly after the GRBs, allowed to measure the delay between the gamma-ray and the optical emissions in GRB 970228 and GRB 970508, for which the optical maxima were reached 0.7 and 2 days after the gamma-ray event, respectively. This appears not to be the case for GRB 971214, for which no optical maximum was detected according to optical observations started only 12 hours after the gamma-ray emission (Halpern et al. 1997). Therefore the analysis of the -band data presented in this study, collected 3.5 and 5 hours after the event, is crucial in order to determine whether a delayed emission, at other wavelengths, following GRB 971214 was present.

3.1. Multiwavelength spectrum of GRB 971214

Our observations in the -band complete the measurements for the GRB 971214 afterglow at other wavelengths (J, I, R and X-rays) and are almost simultaneous to the observation performed with the NFI on board BeppoSAX, 6.5 hours after the burst (Antonelli et al. 1997). This fact enables us to calculate the near-IR to X-ray (2-10 KeV) flux ratio 5-6.5 hours after the GRB, . Taking into account the measurements performed at other wavelengths on Dec 15.44-15.51, and extrapolating our second observation with power-law decays ranging from to , the measured rough broad-band spectrum (IR-optical) of the GRB 971214 afterglow can be obtained (see Fig. 2). As it can be seen in this figure, the shape of the measured flux density distribution vs depends on the power-law index assumed for the light curve. Thus, if the multiwavelength spectrum of the GRB 971214 afterglow would show a maximum in the near-IR. This possible maximum around the J-band has not been detected in previous multiwavelength spectra of GRB 971214 due to the lack of measurements in the -band (Reichart 1998).

Fig. 2. A broad-band spectrum of the GRB 971214 afterglow on December 15.44-15.51. The points represent measurements in (solid symbols), J, I and R. The values of the -band fluxes have been obtained extrapolating the magnitude measured 5 hours after the gamma-ray event, for four power-law decay indices in the range 1.0-1.6. For clarity the error bar in the -band is only shown in the upper right corner.

3.2. Study of the light curve

According to the fireball models, the afterglow radiation will shift progressively to lower frequencies and the corresponding timescales will lengthen (Katz and Piran 1997). Therefore, the -band measurements impose, over the available data, the most stringent limits to a possible maximum in the light curve of the GRB 971214 afterglow. Three possible light curve shapes can fit our data: a rising light curve (case 1) similar to that detected for GRB 970228 (Guarnieri et al. 1997, Pedichini et al. 1997) and GRB 970508 (Pedersen et al. 1998, Galama et al. 1998), a plateau phase (case 2), as it was seen in GRB 970508 between 4 and 24 hours after the burst (Castro-Tirado et al. 1998), or a power-law decay (case 3) as it was later reported for both GRB 970228 (Galama et al. 1997) and GRB 970508 (Sokolov et al. 1998).

3.2.1. Non fading light curve

In case of a rising light curve (case 1) or a plateau (case 2) the light curve could not extend, following the same trend, until Dec 15.54, giving the upper limit of K 18.5 imposed by Garcia et al. (1997) (see Fig. 3).

Fig. 3. The solid circles represent the results of the observations carried out at Calar Alto 3.5 and 5 hours after the burst. The arrow shows the upper limit found by Garcia et al. (1997). The curves represent the power-law decays for different exponents, ranging from , to . The upper limit does not impose any constraint unless . The intersection of the horizontal line with the power-law curve that goes through the Garcia et al. measurement provides an upper limit for the turning point time, =10.2 hours.

Even an increasing curve until a time near Dec 15.54 followed by a very sharp bending over would be unrealistic. We have estimated the acceptable time for this turning in the following way: first we have assumed that the light curve displayed a plateau phase (at ) from Dec 15.18 onwards; then we have constructed a curve with a power-law decline index that matches the Garcia et al. (1997) measurement on Dec 15.54. The intersection between the constant light curve (initial phase) at =18.0 and the power-law fading light curve provides an upper limit for the time of the turnover point, (see Fig. 3). On the other hand, if we assume an increasing light curve crossing our two data points at Dec 15.12 and Dec 15.18, the intersection with the power-law would move slightly backwards. Lower values of would also give lower values of in all cases. We have not considered larger (unrealistic) values of . So, we conclude that the possible maximum or turning point took place =10.2 hours after the gamma-ray event, at the latest.

3.2.2. Fading light curve

If we assumed a single-fading light curve (case 3) with a power-law decline in the near-IR with index (as used by Waxman (1997) and similar to the other two optical counterparts) a variation between our images taken 1.5 hours apart of magnitudes would be expected. However, our data imply a magnitude difference , which is from the above mentioned prediction derived from the power-law (see Fig. 3). If the assumed power-law index were the rejection level would be , being necessary a power-law index (too unrealistic) in order to find a disagreement at a level between our points and the prediction of a power-law decay. The power-law light curve connecting our second measurement 5.0 hours after the gamma-ray event and the upper limit imposed by Garcia et al. (1997), would have an index . Therefore, power-law decays with indices are ruled out, because they imply a magnitude on Dec 15.54, which would be above the reported upper limit.

© European Southern Observatory (ESO) 1998

Online publication: June 12, 1998

helpdesk.link@springer.de