Outflows in Emission and Absorption: The SDSS Perspective

1
Outflows in Emission and Absorption The SDSS
Perspective
Gordon Richards, John Hopkins University
With thanks to Pat Hall, Jon Trump, Tim Reichard,
Sarah Gallagher, and others
2
Outline

• Evidence for outflows from emission lines
• CIV blueshifts
• Relationship between blueshifts and the Baldwin
  Effect (which may, in fact, be the same
  phenomenon)
• Clues to the nature of the wind from absorption
  lines
• Emission line properties of BALQSOs
• NALs in BALs
• Conclusions -- All quasars have winds/outflows,
  but the nature of the winds change as a function
  of their SEDs

3
Emission Line Blueshifts

• Gaskell (1982), Wilkes (1984) find that some
  emission lines (particularly CIV) are blueshifted
  with respect to the quasars systemic redshift

• Numerous papers (e.g. Tytler Fan 1992,
  Marziani 1996) confirm this result (and extend
  it to other emission lines, mostly high
  ionization but H? is redshifted)
• Richards et al. (2002) study emission line shifts
  in a large sample of SDSS quasars, finding that
  CIV blueshifts are the norm, not the exception,
  and argue that the blueshifts result from
  missing flux in the red wing
• Baskin Laor (2004) summarize arguments for
  L/LEdd as the driver (but we will not consider
  this further as it does not directly relate to
  the structure of the wind)

4
Richards et al. 2002 Summary

• CIV blueshifts are ubiquitous in radio-quiet
  quasars (radio loud quasars having smaller or
  zero blueshift.
• Large CIV blueshift quasars have larger FWHM
  lines, but smaller equivalent widths.

• The blueshifts are not correlated with Mi (but
  this may be mis-leading if the shape of the SED
  or UV luminosity is the important quantity to
  measure).
• The blueshift is not a bulk line shift, but
  rather a reduction/suppression/obscuration of the
  red wing. This is not necessarily meant to say
  that the red wing flux is absorbed, but simply
  that the process may be more complex than is
  generally assumed. This may be important for
  masses determined from the FWHM of CIV and for
  metallicity measurements.
• Blueshifts may be an orientation indicator. This
  need not be the external orientation, but may
  reflect the opening angle of the wind (which may
  depend on L/LEdd).

5
CIV Blueshifts in SDSS Quasars
Blueshift of CIV emission line with respect to
MgII for 12,700 SDSS quasars.
N.B. If CIV blueshifts are an indicator of
outflows, then virtually all quasars have them.
6
CIV Blueshifts Not Just a Wind
Saying that this is due to a wind is an over
simplification.

• Mean maximum outflow velocity does not increase
• Rather the peak is blueshifted
• Equivalent width decreases
• Bicones can produce redshifts as well as
  blueshifts depending on viewing angle relative to
  opening angle

7
The Baldwin Effect (BEff)

• Baldwin (1977) shows that more luminous quasars
  have weaker CIV emission lines
• Numerous papers confirm this effect and extend
  it to other emission lines, generally high
  ionization

• There are notable exceptions (e.g. SiIV)
• The effect may be due to SED shape rather than
  luminosity (e.g., Binette ete al. 1989, Zheng
  Malkan 1993, Korista et al. 1998, etc.)
• For example the Luv - Lx relationship (Avni
  Tannanbaum 1982, Wilkes et al. 1994, Green et al.
  1995, Vignali et al. 2003, Strateva et al. 2005)
• See Dietrich et al. 2002 for a recent review

8
BEff vs. Blueshifts
There are many similarities between the BEff and
the CIV emission line blueshifts.

• Weaker CIV lines
• Which lines show the effect
• SED/Luminosity dependence
• Velocity dependence

Suggests that the Baldwin Effect and CIV
Blueshifts are the same phenomenon. We just
dont normally realize it because the CIV
redshift is usually taken to be systemic (and
because luminosity effects can be hard to
disentangle from SED effects).
9
Blueshifts Baldwin Effect I
Large blueshift quasars have larger FWHM CIV
emission lines. The CIV lines are also much
weaker. (See also Corbin 1990, Marziani 96)
10
Blueshifts Baldwin Effect IIa
In R02, we argued that the blueshifts are not a
luminosity effect (in Mi), but that the quasars
with blueshifts are bluer. It is likely that
these relationships are driven not so much by
luminosity, but by the SED (e.g. Zheng Malkan
1993)
11
Blueshifts Baldwin Effect IIb
It has been long known (e.g., Avni Tannanbaum
1982 Wilkes et al. 1994 Green et al. 1995) that
more UV luminous quasars are (relatively) weaker
X-ray sources (?ox more negative). Vignali,
Brandt Scheneider (2003) and Strateva et al.
(2005) have confirmed this correlation and
characterized it for a large population of
quasars. Given that the ionizing photons that
produce CIV come from the far-UV/soft-X-ray, it
seems unlikely that the Baldwin Effect/CIV
blueshifts are independent of this fact.
12
Blueshifts Baldwin Effect III
Francis Koratkar (1995) found that the Baldwin
Effect is strongest in the the red wing of the
CIV emission line.
13
CIV Blueshifts in SDSS Quasars
Blueshifted quasars are bluer than average. But
not all blue quasars have blueshifts. It may be
important to further consider the colors in
addition to blueshifts. Thus we make composite
spectra for 3 blueshift/color combinations.
14
Blueshift/Color Composites
SiIV
CIV
HeII
15
CIV Blueshifts in SDSS Quasars
X-ray strong quasars X-ray weak quasars
Blueshifted quasars tend to be weaker X-ray
sources.
16
CIV Blueshifts and X-ray Absorption
This is consistent with our results from
Gallagher et al. (2004), where large blueshift
objects (dots) were seen to show evidence for
soft X-ray absorption.
Gallagher et al. 2004
17
CIV Blueshifts in SDSS Quasars
Radio-loud quasars tend to have small blueshifts.
Comparison with de Vries, Becker White
(astro-ph/0510747) suggests that FRIIs may have
smaller blueshifts than core dominated sources.
18
BALQSO Emission Line Properties
We may be able to get an additional handle on the
nature of winds using CIV blueshifts as there are
correlations between BALQSOs and emission line
features.

• Bona-fide BALQSOs appear have intrinsically blue
  optical colors and have weaker CIV lines that are
  also blueshifted. (Richards02, Reichard03, see
  also Corbin90.)
• The BAL trough structure appears to be correlated
  with emission line (and SED) properties. See
  Turnshek (1988).

19
BALQSOs Have Large Blueshifts
Composite BALQSO spectra best match non-BALQSO
composites with larger than average blueshifts.
(Using the red wing of CIV, HeII and CIII as the
basis for comparison.) LoBALs appear to have even
larger blueshifts than HiBALs.
Richards et al. 2002
20
Red/Blue BALQSOs
BALQSOs with redder power-law continua (after
dust correction) have stronger emission lines and
narrower, lower-velocity BAL troughs. In
non-BALQSOs we see the same emission line trends
with color.
Reichard et al. 2003
21
Absorption Correlated w/ Emission
Fast drop-off
Rolling Hills
22
BALs with NALs?
BALQSOs can also have NALs. These quasars tend
to have broad, but weak emission lines.
23
15 of Quasars are BALQSOs

• BALQSOs are a separate class of quasars,
  amounting to 15 of the population
• All quasars are BALQSOs, but the BAL covering
  fraction is 15

24
15 of Quasars are BALQSOs

• All quasars have winds/outflows.
• The structure of those outflows depends on the
  SED.
• Bona-fide BAL troughs are found in quasars with
  certain SEDs (blue, large blueshift, X-ray weak).
• All such quasars have BAL troughs (which are
  only seen along certain lines of sight).
• The fraction of such quasars that exhibit BAL
  troughs is much larger than 15.
• 15 is simply the BALQSO fraction integrated
  over all SEDs.

25
Summary
This is a BALQSO
This is a BALQSO -- probably a LoBAL
This is NOT a BALQSO