| 其他摘要 | Quasars, also quasi-stellar objects (QSOs), are luminous active galactic nuclei (AGNs) and the brightest steady objects in the universe. They can be observed spectroscopically up to high redshifts.The spectra of quasars from the distant universe carry a lot of information about quasars themselves, their host galaxies, the interstellar media (ISM) and the intergalactic media (IGM) on their way to us. Quasars provide us with the opportunities to study the formation and evolution of galaxies and the physical mechanisms of the central engine of AGNs. The aims of this thesis is twofold: one is to study dust in the ISM at high redshifts and the other is to study the black hole masses of AGNs by analyzing optical and ultraviolet spectra. The first part of the thesis is on the interstellar dust at high redshifts. In order to search for the Milky Way(MW)-type extinction curves and to model their uncertaintities, we build up a libray of intrinsic spectra of quasars according to the distribution of their power-law spectral indices. We refer to this method as “quasar spectra pair method”. In chapter 2, details of the building up of the libray and the scheme for modelling the extinction curve are given. The quasar spectra pair method is used to determine the extinction curve of a metal-strong and dust-rich Damped \lya~Absorption system (DLAs \footnote{In this thesis, this abbreviation is referred to as quasar DLAs unless further specified. }). The derived extinction curve is MW-type with a significant broad 2175 \AA\ dust absorption feature. The derived extinction is very high with $A_V \approx 0.92$ mag, which is unprecedented in quasar DLAs. We measure the column densities of \hi\ ,~\znii\ ,~\feii\ ,~\cii\ ,~\alii~and \siii~from spectra taken with the Multiple Mirror Telescope (MMT). The neutral hydrogen column density, $\rm N_{\hi} = 10^{21.8\pm0.2}$~cm$^{-2}$,is among the highest values reported in quasar DLAs. The measured \znii~column density is $\rm N_{ZnII}\approx 10^{13.8}$~cm$^{-2}$, which is about 1.5 times larger than e highest value in reported quasar DLAs. The column densities of metal absorption lines may be lower limits because these absorption lines are very broad. The large relative gas-phase chemical abundance of [Zn/Fe]$\approx1.0$ indicates that metals are heavily depleted onto dust grains and the dust depletion level is between that of the warm and the cool clouds toward $\zeta$~Oph in the MW. DLAs with high metallcity and high $\rm N_{\hi}$ are rare and they bridge the gap between quasar DLAs and DLAs in the spectra of Gamma-ray burst (GRB) afterglows (GRB-DLAs) on the metallicity-$N_{\hi}$ plane. They are helpful to understand the differences between DLAs and GRB-DLAs. High extinction is usually related to high $\rm N_{\hi}$ and/or high metallicity. DLAs with high metal column densities tend to be found in dim quasars according to the correlation between N$_{\rm \znii}$ of metal-strong DLAs and r magnitude. Metal-strong and dust-rich DLAs could have significant influences on the cosmic evolution of the \hi\ density and metallicity. In chapter 3, we build up a sample of quasars showing the 2175 \AA\ absorption bump in their spectra from the SDSS and study the 2175 \AA\ feature associated with quasars. A preliminary statistical investigation of the sample indicate that the distribution of the centers and widths of these 2175\AA\ bumps are broader than that in the MW while the bump strengths are weaker than that in the MW. These differences may be caused by different environment of the dust as well as large uncertaintities in the measurements. Meanwhile, the bump strengths of some candidates are very large while there is almost no extinction in the optical band. This indicates the distributions of dust size in these candidates are very peculiar. The number of large-sized dust particles may be much smaller than that of small-sized ones. These peculiar size distributions of the dust are likely caused by violent events, such as AGNs activities, because the large-sized dust particles are easy to destroy by sputtering. We then focus on the detection of a 2175\AA\ bump associated with quasar SDSS SDSS J091613.60$+$292106.1 (hereafter J0916$+$2921). The 2175\AA\ bump in the object is detected at the $6\sigma$ confidence level. An analysis of a spectrum taken with BCS on MMT indicates that the dust depletion is large with [Zn/Fe]$\approx1.01$. However, its X-ray spectrum observed with XMM-Newton does not show significant evidence of X-ray absorption. This indicates that the dust-to-gas ratio of the quasar is very high. Meanwhile, the minimum size of the dust was constrained to be 10\AA\ as smaller dust would have been destructed by the strong X-ray radiation field\ucite{1991ApJ...379..122V}. The object could provide strong constraints on the composition of the dust. The second part of the thesis is on the black hole mass (\mbh) of AGNs. Currently estimation of \mbh\ of AGNs has been made possible by using an empirical virial formalism, which is calibrated with the result of reverberation mapping (RM). Using the \hb\ data of 35 AGNs presented by Peterson et al. (2004) \ucite{2004ApJ...613..682P} along with recent updates and our new measurements of \mgii\ FWHM of 29 out of these AGNs, we found neither the \hb\ FWHM nor the \mgii\ FWHM is a good tracer of Virial velocity. The relations between the FWHM of either \hb\ or \mgii\ and the $\sigma_{\rm line}$ of \hb\ measured from the root-mean-square (RMS) spectra, which is considered as the best indicator of Virial velocity of the broad line region, is nonlinear. We re-calibrate the \mbh\ formalisms by setting the power index on FWHM to be a free parameter and determining it from fitting the formalisms to the data and \mbh\ obtained from RM. Our formalisms improve over the previous ones by giving \mbh\ much more consistent with the RM-based \mbh , as well as smaller scatter. Moreover, the two formalisms, one based on \hb\ and the other based on \mgii , give self-consistent \mbh\ estimation, when applying them to a large AGN sample from the SDSS. The systematic bias in previous \mbh\ estimation formalisms, as we find in this work, would inevitably lead to some artificial correlations, as reported in the literature. For example, the difference between \hb-based \mbh\ and \mgii-based \mbh\ is systematically correlated with the Eddington ratio\ucite{2008ApJ...689L..13O}; however, such a spurious correlation disappears when using \mbh\ estimated from our formalisms. |
修改评论