We used a large, homogeneous sample of 4178 z <= 0.8 Seyfert 1 galaxies and QSOs selected from the Sloan Digital Sky Survey to investigate the strength of Fe II emission and its correlation with other emission lines and physical parameters of active
galactic nuclei. We find that the strongest correlations of almost all the emission-line intensity ratios and equivalent widths (EWs) are with the Eddington ratio (L/L_{Edd}), rather than with the continuum luminosity at 5100AA (L_{5100}) or black hole mass (M_{BH}); the only exception is the EW of ultraviolet Fe II emission, which does not correlate at all with broad-line width, L_{5100}, M_{BH}, or L/L_{Edd}. By contrast, the intensity ratios of both the ultraviolet and optical Fe II emission to Mg II lambda 2800 correlate quite strongly with L/L_{Edd}. Interestingly, among all the emission lines in the near-UV and optical, the EW of narrow optical Fe II emission has the strongest correlation with L/L_{Edd}. We suggest that the variation of the emission-line strength in active galaxies is regulated by L/L_{Edd} because it governs the global distribution of the column density of the clouds gravitationally bound in the line-emitting region, as well as its overall gas supply. The systematic dependence on L/L_{Edd} must be corrected when using the FeII/MgII intensity ratio as a measure of the Fe/Mg abundance ratio to study the history of chemical evolution in QSO environments.
We have conducted a systematic investigation of the origin and underlying physics of the line--line and line--continuum correlations of AGNs, particularly the Baldwin effect. Based on the homogeneous sample of Seyfert 1s and QSOs in the SDSS DR4, we
find the origin of all the emission-line regularities is Eddington ratio (L/Ledd). The essential physics is that L/Ledd regulates the distributions of the properties (particularly column density) of the clouds bound in the line-emitting region.
In the unification scheme of Seyfert galaxies, a dusty torus blocks the continuum source and broad line region in Seyfert 2 galaxies. However it is not clear whether or not and to what extent the torus affects the narrow line spectra. In this paper,
we show that Seyfert 1 and Seyfert 2 galaxies have different distributions on the [OIII]/H$beta $ vs [NII]/H$alpha$ diagram (BPT diagram) for narrow lines. Seyfert 2 galaxies display a clear left boundary on the BPT diagram and only 7.3% of them lie on the left. By contrast, Seyfert 1 galaxies do not show such a cutoff and 33.0% of them stand on the left side of the boundary. Among Seyfert 1 galaxies, the distribution varies with the extinction to broad lines. As the extinction increases, the distribution on BPT diagram moves to larger [NII]/H$alpha$ value. We interpret this as an evidence for the obscuration of inner dense narrow line region by the dusty torus. We also demonstrate that the [OIII] and broad line luminosity correlation depends on the extinction of broad lines in the way that high extinction objects have lower uncorrected [OIII] luminosities, suggesting that [OIII] is partially obscured in these objects. Therefore, using [OIII] as an indicator for the nuclear luminosity will systematically under-estimate the nuclear luminosity of Seyfert 2 galaxies.
We have investigated the broad-line Balmer decrements (Halpha/Hbeta) for a large, homogeneous sample of Seyfert 1 galaxies and QSOs using spectroscopic data obtained in the Sloan Digital Sky Survey. The sample, drawn from the Fourth Data Release, com
prises 446 low redshift (z < 0.35) active galactic nuclei (AGN) that have blue optical continua as indicated by the spectral slopes in order to minimize the effect of dust extinction. We find that (i) the distribution of the intrinsic broad-line Halpha/Hbeta ratio can be well described by log-Gaussian, with a peak at Halpha/Hbeta=3.06 and a standard deviation of about 0.03 dex only; (ii) the Balmer decrement does not correlate with AGN properties such as luminosity, accretion rate, and continuum slope, etc.; (iii) on average, the Balmer decrements are found to be only slightly larger in radio-loud sources (3.37) and sources having double-peaked emission-line profiles (3.27) compared to the rest of the sample. We therefore suggest that the broad-line Halpha/Hbeta ratio can be used as a good indicator for dust extinction in the AGN broad-line region; this is especially true for radio-quiet AGN with regular emission-line profiles, which constitute the vast majority of the AGN population.
