The ionized gas and nuclear environment in NGC 3783.: IV.: Variability and modeling of the 900 kilosecond Chandra spectrum

We present a detailed spectral analysis of the data obtained from NGC 3783 during the period 2000-2001 using Chandra. The data were split in various ways to look for time- and luminosity-dependent spectral variations. This analysis, along with the measured equivalent widths of a large number of X-ray lines and photoionization calculations, lead us to the following results and conclusions. (1) NGC 3783 fluctuated in luminosity by a factor of similar to1.5 during individual observations (most of which were of 170 ks duration). These fluctuations were not associated with significant spectral variations. (2) On a longer timescale (20-120 days), we found the source to exhibit two very different spectral shapes. The main difference between these can be well-described by the appearance (in the "high state") and disappearance (in the "low state") of a spectral component that dominates the underlying continuum at the longest wavelengths. Contrary to the case in other objects, the spectral variations are not related to the brightening or the fading of the continuum at short wavelengths in any simple way. NGC 3783 seems to be the first active galactic nucleus (AGN) to show this unusual behavior. (3) The appearance of the soft continuum component is consistent with being the only spectral variation, and there is no need to invoke changes in the opacity of the absorbers lying along the line of sight. Indeed, we find that all the absorption lines that can be reliably measured have the same equivalent widths (within the observational uncertainties) during high and low states. (4) Photoionization modeling indicates that a combination of three ionized absorbers, each split into two kinematic components, can explain the strengths of almost all the absorption lines and bound-free edges. These three components span a large range of ionization and have total column of about 4 x 10(22) cm(-2). Moreover, all three components are thermally stable and seem to have the same gas pressure. Thus, all three may coexist in the same volume of space. This is the first detection of such a multicomponent, equilibrium gas in an AGN. (5) The only real discrepancy between our model and the observations concerns the range of wavelengths absorbed by the iron M-shell UTA feature. This most likely arises as the result of our underestimation of the poorly known dielectronic recombination rates appropriate for these ions. We also note a small discrepancy in the calculated column density of O vi and discuss its possible origin. (6) The lower limit on the distance of the absorbing gas in NGC 3783 is between 0.2 and 3.2 pc, depending on the component of ionized gas considered. The assumption of pressure equilibrium imposes an upper limit of about 25 pc on the distance of the least-ionized component from the central source.