For some samples, it has been shown that spectra of QSOs with low redshift are bluer during their brighter phases. For the FIRST bright QSO sample, we assemble their spectra from SDSS DR7 to investigate variability between the spectra from White et a
l. (2000) and from the SDSS for a long rest-frame time-lag, up to 10 years. There are 312 radio loud QSOs and 232 radio quiet QSOs in this sample, up to $z sim 3.5$. With two-epoch variation, it is found that spectra of half of the QSOs appear redder during their brighter phases. There is no obvious difference in slope variability between sub-samples of radio quiet and radio loud QSOs. This result implies that the presence of a radio jet does not affect the slope variability on 10-year timescales. The arithmetic composite difference spectrum for variable QSOs is steep at blueward of $sim$ 2500AA. The variability for the region blueward of 2500 AA is different to that for the region redward of 2500 AA.
Using the properties of SDSS DR7 QSOs catalog from Shen et al., the Baldwin effect, its slope evolution, the underlying drive for a large sample of 35019 QSOs with reliable spectral analysis are investigated. We find that the Baldwin effect exists in
this large QSOs sample, which is almost the same in 11 different redshift bins, up to $zsim 5$. The slope is -0.238 by the BCES (civ EW depends on the continuum), -0.787 by the BCES bisector. For 11 redshift-bins, there is an increasing of the Baldwin effect slope from $zsim1.5$ to $zsim2.0$. From $zsim2.0$ to $zsim5.0$, the slope change is not clear considering their uncertainties or larger redshift bins. There is a strong correlation between the rest-frame civ EW and civ-based mbh while the relation between the civ EW and mgii-based mbh is very weak. With the correction of civ-based mbh from the civ blueshift relative to mgii, we suggest that this strong correlation is due to the bias of the civ-based mbh, with respect to that from the mgii line. Considering the mgii-based mbh, a medium strong correlation is found between the civ EW and the Eddington ratio, which implies that the Eddington ratio seems to be a better underlying physical parameter than the central black hole mass.
It is found that feii emission contributes significantly to the optical and ultraviolet spectra of most active galactic nuclei. The origin of the optical/UV feii emission is still a question open to debate. The variability of feii would give clues to
this origin. Using 7.5 yr spectroscopic monitoring data of one Palomer-Green (PG) quasi-stellar object (QSO), PG 1700+518, with strong optical feii emission, we obtain the light curves of the continuum lv, feii, the broad component of hb, and the narrow component of hb by the spectral decomposition. Through the interpolation cross-correlation method, we calculate the time lags for light curves of feii, the total hb, the broad component of hb, and the narrow component of hb with respect to the continuum light curve. We find that the feii time lag in PG1700+518 is $209^{+100}_{-147}$ days, and the hb time lag cannot be determined. Assuming that feii and hb emission regions follow the virial relation between the time lag and the FWHM for the hb and feii emission lines, we can derive that the hb time lag is $148^{+72}_{-104}$ days. The hb time lag calculated from the empirical luminosity--size relation is 222 days, which is consistent with our measured feii time lag. Considering the optical feii contribution, PG 1700+518 shares the same characteristic on the spectral slope variability as other 15 PG QSOs in our previous work, i.e., harder spectrum during brighter phase.
