If optical microvariability for radio-loud AGNs is now well established (Carini et al. 1991, Miller et al. 1992, Wagner et al. 1992, Doroshenko et al. 1992), any search for intra-night optical variability in radio-quiet QSOs (Gopal-Krishna et al. 1993, 1995, Jang et al. 1997, Rabbette et al. 1998) or Seyfert galaxies in the past ten years does not report clear evidence of such phenomenon. In fact, much of the controversy may be due to the transient characteristics of microvariability, precluding a clear confirmation of any reported variations. This becomes a critical problem in the case of Seyfert galaxies, since very few observations of such objects were performed to search for optical flickering. To our knowledge, the first one was done by Lawrence et al. (1981). They observed NGC 4151 on a range of time-scales from 10 seconds to 1 week, but detected no variation to better than 0.05 mag in bands V and K. Five years later, systematic photoelectric UBV observations of rapid variability in AGNs were begun at the Crimean Laboratory in 1986 and have been carried out for the Seyfert galaxies NGC 4151, NGC 7469, NGC 3516 and NGC 5548 (Lyutyi et al. 1989, Aslanov et al. 1989 and Lyutyi & Doroshenko 1993 respectively). Optical microvariabilities were detected in each of these objects: NGC 4151, NGC 7469 and NGC 3516 showed amplitude of microvariability up to 10% over 15-20 min and about 5% for NGC 5548 with a shorter time-scale, i.e. 10-15 min. But for each cases, the phenomenon is not continuous and periods without any rapid variabilities are also observed all along each run. It seemed thus that onset and disappearance of microvariability follow a random process as observed in radio-loud AGNs (Carini 1990). A number of other investigators attempted to detect or confirm the presence of these rapid variations. For example, the case of NGC 7469 was confirmed by Dultzin-Hacyan et al. (1992) althougth no variations were obtained during another run (Dultzin-Hacyan et al. 1993).
Yet, particulary important results come from simultaneous multifrequency observations which can put strong constraints on the spatial distribution of the emitting regions and indicate whether the same radiative process dominates at different frequencies. The only such search for Seyfert galaxies is the simultaneous optical-infrared-X-ray study of NGC 4051 by Done et al. (1990). They report that, on time-scales of tens of minutes, the flux remained constant within 1% and 5% in optical and infrared, respectively, while the X-ray flux continually flickered by up to a factor 2. Another survey of this galaxy was done by Hunt et al. (1992), but only in the K band, and confirm the upper limit on nuclear variability of about 2%. Done et al. deduced from their results that the IR/optical source must be at least an order of magnitude larger than, or completely separate from, the X-ray source.
It appears from these results that, in a general manner, the study of microvariability in Seyfert galaxies is not sufficiently complete to clearly conclude if optical flickering is (or is not) a common characteristics for this class of AGNs. This contrasts with the more complete works done recently with QSOs. First, Jang et al. (1997) report, on a selected sample of radio-quiet and radio-loud QSOs, an apparent contrast in microvariations between the two class of quasars, 20% of the radio-quiet objects showing evidence of flickering against 85% for radio-loud. Next, Rabbette et al. (1998) have just published a search for rapid optical variability (on time-scales of few minutes) in a large sample of 23 radio-quiet quasars. They report no detection, with a precision of few percents, of any significant rapid variability for any of the sources observed.
Presently, no clear explanations of microvariability are approved unanimously. Unlike radio-loud AGNs where flikering could be due to the presence of shock inside a relativistic jet (Qian et al. 1991, Gopal-Krishna & Wiita 1992), no such conclusion can be drawn up to now for radio-quiet objects such as Seyfert galaxies, since their high energy spectrum is apparently cut-off above a few hundred keV (Jourdain et al. 1992; Maisack et al. 1993; Dermer & Gehrels 1995). Thus, some models supposed that microvariability could be due to disturbances (like flares or hot spots) in the accretion disk surrounding the central engine (Wiita et al. 1991, 1992, Chakrabarti & Wiita 1993, Mangalam & Wiita 1993). But some results of recent observations do not provide strong support for such models (Jang & Miller 1997). In the case of Seyfert galaxies, the origin of microvariability could be associated with the high energy process giving birth to the hard X-ray spectra (up to few hundred of keV) observed in these galaxies. The source of the high energy emission is still uncertain: it could be produced through the comptonization of low energy photons by a thermal, mildly relativistic plasma (Haardt & Maraschi 1991) or by Inverse Compton process from a non-thermal, highly relativistic particle distribution (possibly made of electron-positron) (e.g. Zdziarski et al. 1994, Henri & Petrucci 1997). As there is probably some magnetic field to accelerate and confine the particles, synchrotron emission is expected to be produced in the latter case, but not in the former. Purely thermal emission in the optical range is likely produced in too broad a region to produce intra-day variability. On the opposite, synchrotron emission should be correlated to X-ray emission, and thus be also rapidly variable. Therefore a positive detection of rapid (intra-day) visible-IR variability would strongly favour non-thermal models. Conversely, non-detection would bring very valuable upper limits on the intrinsic properties of the local environment of the emission region, giving strong constraints on the various models of non thermal emission (Celotti et al. 1991). Besides, if it exists, the synchrotron emission is diluted by the stellar contribution and probably by the thermal continuum possibly emitted by an accretion disk and by dust. The dust emission peaks in the IR range and fall down near 1 µm due to dust sublimation, while the disk emission, supposed to give rise to the Blue Bump, peaks in the UV range. Hence the most favourable wavelength domain to detect variable synchrotron emission would be around 1 µm.
We present here the results of two observational campaigns of a sample of 22 Seyfert 1 galaxies in the I band at (and simultaneously in the J band at for 3 of them), at the observatories of Cananea and San Pedro Mártir in Mexico. We aim to detect rapid optical variabilities by differential photometry between the galaxies and the comparisons stars in the CCD field of view. We have developped a new method of analysis which minimize the influence of the intrinsic variabilities of the comparisons stars. In Sect. 2 we report on the sample and the observations. The data analysis method is explained in Sect. 3. We present the results for each galaxy, in Sect. 4, developing the cases of the more interesting ones. We will finally discuss the theoretical constraints imposed by these outcomes in Sect. 5 before concluding.
© European Southern Observatory (ESO) 1999
Online publication: February 23, 1999