UV and optically selected candidates for stellar tidal disruption events (TDE) often exhibit broad spectral features (HeII emission, H$\alpha$ emission, or absorption lines) on a blackbody-like continuum (1e4K<T<1e5K). The lines presumably emit from TDE debris or circumnuclear clouds photoionized by the flare. Line velocities however are much lower than expected from a stellar disruption by supermassive black hole (SMBH), and are somewhat faster than expected for the broad line region (BLR) clouds of a persistently active galactic nucleus (AGN). The distinctive spectral states are not strongly related to observed luminosity and velocity, nor to SMBH mass estimates. We use exhaustive photoionization modelling to map the domain of fluxes and cloud properties that yield (e.g.) a He-overbright state where a large HeII(4686A)/H$\alpha$ line-ratio creates an illusion of helium enrichment. Although observed line ratios occur in a plausible minority of cases, AGN-like illumination can not reproduce the observed equivalent widths. We therefore propose to explain these properties by a light-echo photoionization model: the initial flash of a hot blackbody (detonation) excites BLR clouds, which are then seen superimposed on continuum from a later, expanded, cooled stage of the central luminous source. The implied cloud mass is substellar, which may be inconsistent with a TDE. Given these and other inconsistencies with TDE models (e.g. host-galaxies distribution) we suggest to also consider alternative origins for these nuclear flares, which we briefly discuss (e.g. nuclear supernovae and starved/subluminous AGNs).

The Broad Line Region (BLR) in AGN is composed of dense gas ($\sim 10^{11}$ cm$^{-3}$) on sub-pc scale, which absorbs about 30 per cent of the ionising continuum. The outer size of the BLR is likely set by dust sublimation, and its density by the incident radiation pressure compression (RPC). But, what is the origin of this gas, and what sets its covering factor (CF)? Czerny & Hryniewicz (2011) suggested that the BLR is a failed dusty wind from the outer accretion disc. We explore the expected dust properties, and the implied BLR structure. We find that graphite grains sublimate only at $T\simeq 2000$ K at the predicted density of $\sim 10^{11}$ cm$^{-3}$, and therefore large graphite grains ($\ge 0.3$ $\mu$m) survive down to the observed size of the BLR, $R_{\rm BLR}$. The dust opacity in the accretion disc atmosphere is $\sim 50$ times larger than previously assumed, and leads to an inflated torus-like structure, with a predicted peak height at $R_{\rm BLR}$. The illuminated surface of this torus-like structure is a natural place for the BLR. The BLR CF is mostly set by the gas metallicity, the radiative accretion efficiency, a dynamic configuration, and ablation by the incident optical-UV continuum. This model predicts that the BLR should extend inwards of $R_{\rm BLR}$ to the disc radius where the surface temperature is $\simeq 2000$ K, which occurs at $R_{\rm in}\simeq 0.18 R_{\rm BLR}$. The value of $R_{\rm in}$ can be tested by reverberation mapping of the higher ionisation lines, predicted by RPC to peak well inside $R_{\rm BLR}$. The dust inflated disc scenario can also be tested based on the predicted response of $R_{\rm BLR}$ and the CF to changes in the AGN luminosity and accretion rate.

There is a large diversity in the C IV broad absorption line (BAL) profile among BAL quasars (BALQs). We quantify this diversity by exploring the distribution of the C IV BAL properties, FWHM, maximum depth of absorption and its velocity shift ($v_{\rm md}$), using the SDSS DR7 quasar catalogue. We find the following: (i) Although the median C IV BAL profile in the quasar rest-frame becomes broader and shallower as the UV continuum slope ($\alpha_{\rm UV}$ at 1700-3000 A) gets bluer, the median individual profile in the absorber rest-frame remains identical, and is narrow (FWHM = 3500 km/s) and deep. Only 4 per cent of BALs have FWHM > 10,000 km/s. (ii) As the He II emission equivalent-width (EW) decreases, the distributions of FWHM and $v_{\rm md}$ extend to larger values, and the median maximum depth increases. These trends are consistent with theoretical models in which softer ionizing continua reduce overionization, and allow radiative acceleration of faster BAL outflows. (iii) As $\alpha_{\rm UV}$ becomes bluer, the distribution of $v_{\rm md}$ extends to larger values. This trend may imply faster outflows at higher latitudes above the accretion disc plane. (iv) For non-BALQs, the C IV emission line decreases with decreasing He II EW, and becomes more asymmetric and blueshifted. This suggests an increasing relative contribution of emission from the BAL outflow to the C IV emission line as the ionizing spectral energy distribution (SED) gets softer, which is consistent with the increasing fraction of BALQs as the ionizing SED gets softer.

The winds of ionized gas driven by Active Galactic Nuclei (AGN) can be studied through absorption lines in their X-ray spectra. A recurring feature of these outflows is their broad ionization distribution, including essentially all ionization levels (e.g., Fe^0+ to Fe^25+). This characteristic feature can be quantified with the absorption measure distribution (AMD), defined as the distribution of column density with ionization parameter |dN / dlog xi|. Observed AMDs extend over 0.1 < xi < 10^4 (cgs), and are remarkably similar in different objects. Power-law fits (|dN /dlog xi| ~ N_1 xi^a) yield N_1 = 3x10^{21} cm^-2 +- 0.4 dex and a = 0 -- 0.4. What is the source of this broad ionization distribution, and what sets the small range of observed $N_1$ and $a$? A common interpretation is a multiphase outflow, with a wide range of gas densities in a uniform gas pressure medium. However, the incident radiation pressure leads to a gas pressure gradient in the photoionized gas, and therefore to a broad range of ionization states within a single slab. We show that this compression of the gas by the radiation pressure leads to an AMD with |dN / dlog xi| = 8x10^21 xi^0.03 cm^-2, remarkably similar to that observed. The calculated values of $N_1$ and $a$ depend weakly on the gas metallicity, the ionizing spectral slope, the distance from the nucleus, the ambient density, and the total absorber column. Thus, radiation pressure compression (RPC) of the photoionized gas provides a natural explanation for the observed AMD. RPC predicts that the gas pressure increases with decreasing ionization, which can be used to test the validity of RPC in ionized AGN outflows.

A fraction of quasars present broad absorption lines, produced by outflowing gas with typical velocities of 3000 - 10,000 km/s. If the outflowing gas fills a significant fraction of the volume where it resides, then it will be highly ionized by the quasar due to its low density, and will not produce the observed UV absorption. The suggestion that the outflow is shielded from the ionizing radiation was excluded by recent observations. The remaining solution is a dense outflow with a filling factor $f<10^{-3}$. What produces such a small $f$? Here we point out that radiation pressure confinement (RPC) inevitably leads to gas compression and the formation of dense thin gas sheets/filaments, with a large gradient in density and ionization along the line of sight. The total column of ionized dustless gas is a few times $10^{22}$ cm$^{-2}$, consistent with the observed X-ray absorption and detectable P V absorption. The predicted maximal columns of various ions show a small dependence on the system parameters, and can be used to test the validity of RPC as a solution for the overionization problem. The ionization structure of the outflow implies that if the outflow is radiatively driven, then broad absorption line quasars should have $L/L_{\rm Edd} \gtrsim 0.1$.

Active galactic nuclei (AGN) are characterized by similar broad emission lines properties at all luminosities ($10^{39}$-$10^{47}$ erg s$^{-1}$). What produces this similarity over a vast range of $10^8$ in luminosity? Photoionization is inevitably associated with momentum transfer to the photoionized gas. Yet, most of the photoionized gas in the Broad Line Region (BLR) follows Keplerian orbits, which suggests that the BLR originates from gas with a large enough column for gravity to dominate. The photoionized surface layer of the gas must develop a pressure gradient due to the incident radiation force. We present solutions for the structure of such a hydrostatic photoionized gas layer in the BLR. The gas is stratified, with a low-density highly-ionized surface layer, a density rise inwards, and a uniform-density cooler inner region, where the gas pressure reaches the incident radiation pressure. This radiation pressure confinement (RPC) of the photoionized layer leads to a universal ionization parameter $U\sim 0.1$ in the inner photoionized layer, independent of luminosity and distance. Thus, RPC appears to explain the universality of the BLR properties in AGN. We present predictions for the BLR emission per unit covering factor, as a function of distance from the ionizing source, for a range of ionizing continuum slopes and gas metallicity. The predicted mean strength of most lines (excluding H$\beta$), and their different average-emission radii, are consistent with the available observations.

We analyze the hydrostatic effect of AGN radiation pressure on optically thick gas in the host galaxy. We show that in luminous AGN, the radiation pressure likely confines the ionized layer of the illuminated gas. Radiation pressure confinement (RPC) has two main implications. First, the gas density near the ionization front is 7x10^4 L_{i,45} r_{50}^{-2} cm^{-3}, where L_{i,45} is the ionizing luminosity in units of 10^45 erg/s and r_{50} is the distance of the gas from the nucleus in units of 50 pc. Second, as shown by Dopita et al., the solution of the ionization structure within each slab is unique, independent of the ambient pressure. We show that the RPC density vs. distance relation is observed over a dynamical range of ~10^4 in distance, from sub-pc to kpc from the nucleus, and a range of ~10^8 in gas density, from 10^3 to 10^11 cm^{-3}. This relation implies that the radiative force of luminous AGN can compress giant molecular clouds in the host galaxy, and possibly affect the star formation rate. The unique ionization structure in RPC includes a highly ionized X-ray emitting surface, an intermediate layer which emits coronal lines, and a lower ionization inner layer which emits optical lines. This structure can explain the observed overlap of the extended X-ray and optical narrow line emission in nearby AGN. We further support RPC by comparing the predicted ratios of the narrow lines strength and narrow line widths with available observations. We suggest a new method, based on the narrow line widths, to estimate the black hole mass of low luminosity AGN.

Broad absorption line quasars (BALQs) present a large diversity in their broad absorption line (BAL) profiles. To investigate what physical parameters underlie this diversity, we construct a sample of BALQs which covers lambda_rest ~ 800-3000 {\AA}, based on the Sloan Digital Sky Survey DR7 quasar catalogue. The average BAL properties are evaluated by taking the ratios of average BALQ spectra to the average spectra of matched samples of non-BALQs, where the matching is based on various emission properties. We find the following properties. (i) There is no detectable Lyman edge associated with the BAL absorbing gas (tau<0.1). (ii) The known increase of average absorption depth with the ionization potential extends to the higher ionization N V and O VI BALs. We also find that the C IV BAL profile is controlled by two parameters. (i) The He II emission EW, which controls the typical velocity of the C IV BAL, but does not affect the absorption depth. (ii) The spectral slope in the 1700-3000 {\AA} range (alpha_UVl), which controls the C IV peak absorption depth, but does not affect the typical velocity. The He II EW and alpha_UVl also control the observed fraction of quasars that are BALQs. We suggest that a lower He II EW may indicate a weaker ionizing continuum, which allows the outflow to reach higher velocities before being overionized, possibly without a need to invoke a shielding gas. A redder continuum may indicate a more inclined system, and a larger covering factor and larger column of the outflow along the line of sight.

Outflows from active galactic nuclei may be produced by absorption of continuum radiation by UV resonance lines of abundant metal ions, as observed in broad absorption line quasars (BALQs). The radiation pressure exerted on the metal ions is coupled to the rest of the gas through Coulomb collisions of the metal ions. We calculate the photon density and gas density which allow decoupling of the metal ions from the rest of the gas. These conditions may lead to an outflow composed mostly of the metal ions. We derive a method to constrain the metals/H ratio of observed UV outflows, based on the Ly {\alpha} and Si iv {\lambda}{\lambda}1394, 1403 absorption profiles. We apply this method to an SDSS sample of BALQs to derive a handful of candidate outflows with a higher than solar metal/H ratio. This mechanism can produce ultra fast UV outflows, if a shield of the continuum source with a strong absorption edge is present.

We report the detection of variable UV absorption lines in NGC 4395, based on UV observations with the HST STIS carried out in April and July, 2004, as part of a reverberation-mapping campaign. Low-ionization lines of O I, N I, Si II, C II, and Fe II, are present in the low-state spectra (April 2004) at a velocity v_shift=-250 km/s (system A_l), and additional high-ionization lines of C IV and N V appear in the high-state spectra (July 2004) at v_shift=-250 km/s (system A_h) and at v_shift=-840 km/s (system B). The absence of absorption from the low metastable levels of Si II implies a density <~10^3 cm^(-3) for system A_l, indicating a location outside the narrow line region (NLR). System A_h is peculiar as only N V absorption is clearly detected. A high N V/C IV absorption ratio is expected for a high metallicity absorber, but this is excluded here as the metallicity of the host galaxy and of the nuclear gas is significantly subsolar. A simple acceptable model for systems A_h and B is an absorber located between the broad line region (BLR) and the NLR, which absorbs only the continuum and the BLR. At the low-state the strong narrow emission lines of C IV and N V dominate the spectrum, making the absorption invisible. At the high-state the absorbed continuum and BLR emission dominate the spectrum. Thus, the change in the observed absorption does not reflect a change in the absorber, but rather a change in the continuum and BLR emission from behind the absorber, relative to the emission from the NLR in front of the absorber. Studies of the absorption line variability in highly variable objects can thus break the degeneracy in the absorber distance determination inherent to single epoch studies.

AGN display an extreme range in the narrow emission line equivalent widths. Specifically, in the PG quasar sample the equivalent width of the narrow [O III] 5007 line has a range of >300, while the broad Hb line, for example, has a range of 10 only. The strength of [O III] 5007 is modulated by the covering factor, CF, of the narrow line region (NLR) gas, its density n_e, and ionization parameter U. To explore which of these factors produces the observed large range in [O III] 5007 strength, we measure the strength of the matching narrow Hb and [O III] 4363 lines, detected in 40 out of the 87 z<0.5 PG quasars in the Boroson & Green sample. The photoionization code CLOUDY is then used to infer CF, n_e, and U in each object, assuming a single uniform emitting zone. We find that the range of CF (~0.02-0.2) contributes about twice as much as the range in both n_e and U towards modulating the strength of the [O III] 5007 line. The CF is inversely correlated with luminosity, but it is not correlated with L_Edd as previously speculated. The single zone [O III] 5007 emitting region is rather compact, having R=40L_44^0.45 pc. These emission lines can also be fit with an extreme two zone model, where [O III] 4363 is mostly emitted by a dense (n_e=10^7) inner zone at R=L_44^0.5 pc, and [O III] 5007 by a low density (n_e=10^3) extended outer zone at R=750L_44^0.34 pc. Such an extended [O III] 5007 emission should be well resolved by HST imaging of luminous AGN. Further constraints on the radial gas distribution in the NLR can be obtained from the spectral shape of the IR continuum emitted by the associated dust.

The high ionization lines in active galactic nuclei (AGN), such as C IV, tend to be blueshifted with respect to the lower ionization lines, such as H beta, and often show a strong blue excess asymmetry not seen in the low ionization lines. There is accumulating evidence that the H beta profile is dominated by gravity, and thus provides a useful estimate of the black hole mass in AGN. The shift and asymmetry commonly seen in C IV suggest that non gravitational effects, such as obscuration and radiation pressure, may affect the line profile. We explore the relation between the H beta and C IV profiles using UV spectra available for 81 of the 87 z < 0.5 PG quasars in the Boroson & Green (BG92) sample. We find the following: (1) Narrow C IV lines (FWHM < 2000 km/s) are rare (~2 per cent occurrence rate) compared to narrow H beta lines (~20 per cent). (2) In most objects where the H beta FWHM <4000 km/s the C IV line is broader than H beta, but the reverse is true when the H beta FWHM > 4000 km/s. This argues against the view that C IV generally originates closer to the center, compared to H beta. (3) C IV appears to provide a significantly less accurate, and possibly biased estimate of the black hole mass in AGN, compared with H beta. (4) All objects where C IV is strongly blueshifted and asymmetric have a high L/L_Edd, but the reverse is not true. This suggests that a high L/L_Edd is a necessary but not sufficient condition for generating a blueshifted asymmetric C IV emission. (5) We also find indications for dust reddening and scattering in `normal' AGN. In particular, PG quasars with a redder optical-UV continuum slope show weaker C IV emission, stronger C IV absorption, and a higher optical continuum polarization.

The origin of the luminosity dependence of the equivalent width (EW) of broad emission lines in AGN (the Baldwin effect) is not firmly established yet. We explore this question for the broad C IV \lambda 1549 line using the Boroson & Green sample of the 87 z < 0.5 Bright Quasar Survey (BQS) quasars. Useful UV spectra of the C IV region are available for 81 of the objects, which are used to explore the dependence of the C IV EW on various emission properties. We confirm earlier results on the strong correlations of the C IV EW with some of the emission parameters which define the Boroson & Green Eigenvector 1, and with the optical to X-ray slope \alpha_ox. In addition, we find a strong correlation of the C IV EW with the relative accretion rate, L/L_Edd. Since L/L_Edd drives some of the Eigenvector 1 correlations, it may be the primary physical parameter which drives the Baldwin effect for C IV.

The origin of the luminosity dependence of the strength of broad and narrow emission lines in AGN (i.e. the Baldwin effect) is not firmly established yet. We explore this question using the Boroson & Green sample of the 87 z<=0.5 PG quasars. UV spectra of the C IV region with a sufficient S/N are available for 81 of the objects. We use these spectra to explore the dependence of the C IV EW on M_BH and L_Edd, as deduced from optical spectra of the H beta region. We find a strong correlation of the C IV EW with L_Edd. This may be the primary correlation which drives the Baldwin effect for C IV. A similar correlation analysis using the C IV FWHM instead of the H beta FWHM, yields an insignificant correlation. This suggests that the C IV FWHM provides a significantly less accurate estimate of M_BH compared to the H Beta FWHM. We also measured the [O III] 4363 /[O III] 5007 line ratio for 72 objects in this sample, and use it to determine whether density or covering factor control the large range in [O III] 5007 strength in AGN.