2001ApJ...554..216K

The high-resolution X-ray spectrum of NGC 3783 shows several dozen absorption lines and a few emission lines from the H-like and He-like ions of O, Ne, Mg, Si, and S, as well as from Fe XVII-Fe XXIII L-shell transitions. We have reanalyzed the Chandra HETGS spectrum using better flux and wavelength calibrations, along with more robust methods. Combining several lines from each element, we clearly demonstrate the existence of the absorption lines and determine that they are blueshifted relative to the systemic velocity by -610±130 km.s^–1. We find the Ne absorption lines in the High-Energy Grating spectrum to be resolved with FWHM=840⁺⁴⁹⁰_–360 km.s^–1; no other lines are resolved. The emission lines are consistent with being at the systemic velocity. We have used regions in the spectrum where no lines are expected to determine the X-ray continuum, and we model the absorption and emission lines using photoionized-plasma calculations. The model consists of two absorption components, with different covering factors, which have an order-of-magnitude difference in their ionization parameters. The two components are spherically outflowing from the active galactic nucleus, and thus contribute to both the absorption and the emission via P Cygni profiles. The model also clearly requires O VII and O VIII absorption edges. The low-ionization component of our model can plausibly produce UV absorption lines with equivalent widths consistent with those observed from NGC 3783. However, we note that this result is highly sensitive to the unobservable UV to X-ray continuum, and the available UV and X-ray observations cannot firmly establish the relationship between the UV and X-ray absorbers. We find good agreement between the Chandra spectrum and simultaneous ASCA and RXTE observations. The 1 keV deficit previously found when modeling ASCA data probably arises from iron L-shell absorption lines not included in previous models. We also set an upper limit on the FWHM of the narrow Fe Kα emission line of 3250 km.s^–1. This is consistent with this line originating outside the broad-line region, possibly from a torus.