Line vs Continuum Correlations

1
Line vs Continuum Correlations

• Joe Shields
• Ohio University

2
Baldwin (1977)

• The Baldwin Effect - Carswell Smith (1979)
• Negative correlation between luminosity and
  line equivalent width

3
QSO spectra as function of
luminosity (Dietrich et al. 2002)
4
Baldwin Effect
Log L(line) ?
Log EW ?
Log L(continuum) ?
Log L(continuum) ?
5
Baldwin Effect - Why Do We Care?

• Originally cosmology
• Early pattern emerging out of chaos of
  phenomenology
• Indicator of broad-line region/accretion physics
• 30 years later
• Still interesting for BLR physics
• (see Baldwin 1999 for cosmology)
• Much better data
• Context has grown (PCA, EV1, black hole masses,
  L/Ledd, etc.)
• EWs easy to measure

5
6
Multiple Baldwin Effects ?

• Ensemble (global) single-epoch observations of
  multiple QSOs
• Intrinsic multi-epoch observations of a single,
  variable AGN
• Broad lines
• Narrow lines
• X-ray (Fe K?)
• WR stars??
• Novae??

7
Ensemble Baldwin Effect what do we know?

• UV/optical Broad Lines
• It exists
• Formerly controversial
• Lots of scatter common
• Need sufficient range of L

Kinney et al. 1990
8
Ensemble Baldwin Effect what do we know?

• 2) Different slope for different lines

• Steeper slope for higher ionization lines

Dietrich et al. 2002
NV ?1240 is an exception
9
Ensemble Baldwin Effect lines of special interest

• NV ?1240
• Most studies show no BE
• Probable explanation metallicity effect
• N/H ? Z2, higher Z goes with higher L (Hamann
  et al.)
• Different behavior for NIII ?1750, NIV ?1486
  (but weak lines)
• H? ?4861
• Recent studies show weak, inverse trend (negative
  BE)
• Important for use of Balmer lines as tracer of
  luminosity
• (e.g., Greene Ho 2005, X.B. Wu
  talk)

10
EW ? L0.2
EW ? L0.1
(2dF, Croom et al. 2002)
(SDSS, Greene Ho 2005)
11
Ensemble Baldwin Effect explanations

• Luminosity-dependent covering factor/ionization
  parameter
• (e.g. Mushotzky Ferland 1984)
• Luminosity-dependent continuum shape
• (e.g. Netzer et al. 1992, Korista et
  al. 1998)
• Another fundamental parameter correlated with
  luminosity? E.g.
• redshift
• L/Ledd
• MBH

11
12
Ensemble Baldwin Effect L vs. z
Observational Problem L strongly
correlated with z in most samples
(LBQS, Green et al. 2001)
Log (z)
Solution fill in L-z plane
(Dietrich et al. 2002)
13
Ensemble Baldwin Effect L vs. z
Bottom line z dependence weak compared to
L
(Dietrich et al. 2002)
14
Ensemble Baldwin Effect L vs. L/LEdd

• LEdd derive from MBH based on linewidth, L

BQS Baskin Laor 2004
? Tighter correlation with L/LEdd
(also Bachev et al. 2004)
15
Ensemble Baldwin Effect L vs. L/LEdd

• Results for H? Netzer et al. 2004

Baldwin Effect in H? ??
15
16
Ensemble Baldwin Effect L vs. L/LEdd
Log EW
Further correlations with L/LEdd
Warner et al. 2004
17
Ensemble Baldwin Effect a function of L or

• Caution L, L/Ledd, MBH often highly correlated

Log L/LEdd
Log L
Warner et al. 2004
18
Ensemble Baldwin Effect L/LEdd vs MBH
? Stronger correlation with MBH
Warner et al., submitted
19
Ensemble Baldwin Effect L vs MBH
Warner et al., submitted
? Stronger correlation with MBH
20
Ensemble Baldwin Effect and PCA Results

• Shang et al. (2003)
• SPC1 mostly responsible for BE,
• linked to L

• SPC3 linked to Boroson Green
• EV1 ? L/LEdd or MBH

Shang et al. 2003
21
Narrow-Line Seyfert 1s and the Baldwin Effect

• Believed to be extreme EV1 or L/LEdd objects

Osmer, Porter, Green (1994)
Constantin Shields (2003)
21
22
Baldwin Effect in NLS1s

• UV Baldwin Effect
• seen in NLS1s
• Offset to smaller EWs

Points NLS1 sample Lines Dietrich QSOs
(Leighly Moore 2004)
(H? shows negative trend also in NLS1s Zhou et
al. 2006)
Is BE more uniform in L/LEdd ?
23
Ensemble Baldwin Effect L vs. L/LEdd
Triangles NLS1 average
Log EW
Warner et al. 2004
? Additional parameter involved
24
Intrinsic Baldwin Effect
Example NGC 4151 (Kong et al. 2006)
25
Intrinsic vs Ensemble Baldwin Effect
Kinney et al. (1990)
25
26
Intrinsic Baldwin Effect what do we know?

• Steeper than ensemble BE
• Causes may be quite different
• Curvature seen
• Tighter relation seen if lag removed (e.g., Pogge
  Peterson 1992)
• (Lag is itself a function of L)
• Consistent with photoionization theory
• (Korista, Goad, Cackett, Horne,
  Knigge, )
• Luminosity-dependent response a diagnostic tool
  for BLR

27
Intrinsic Baldwin Effect what do we know?

• 4) Some evidence of ionization-dependent
  slope
• e.g. CIV steeper than Ly?
• FeII
• Mixed results for slope (optical), e.g. NGC 5548
  (Vestergaard Peterson 2005), NGC 4051 ( NLS1
  Wang et al. 2005)

• FeII does respond to continuum? photon
  excitation
• Measuring FeII is hard!

Wang et al. 2005
28
The Baldwin Effect in Narrow Lines

• Detection claimed in several studies, esp. O
  III ?5007

Log EW
NeV
NeIII
OII
MB
(Also IR Keremedjev Hao)
OIII
2dF Croom et al. 2002
29
The Baldwin Effect in Narrow Lines
Netzer et al. 2004

• Popular explanation high L ?NLR reaches limiting
  size galaxy size
• Implications for detection of Type II QSOs (e.g.
  Netzer et al. 2006)
• Weak lines ? harder to find
• Implications for star formation and detection via
  nebular lines? (cf. Ho 2005 weak OII)
• Star formation suppressed? Or confined to dense
  regions with low filling factor?

30
X-Ray Baldwin Effect Fe K? (Iwasawa Taniguchi
1993)

• Broad Fe K? (Nandra 1997)
• Probable origin reflection from accretion disk
• Models luminosity-dependent ionization of disk
  (Nandra, Nayakshin), and gravitational
  light-bending (Miniutti Fabian 2004)
• Questions regarding broad component measurements
• Narrow Fe K? (e.g. Page et al. 2004)
• Probable origin reflection from outer disk or
  torus
• Model luminosity dependent covering factor
  (receding torus)
• Due primarily to radio-loud sources with extra
  continuum component? (Jiang et al. 2006)

31
Summary

• Baldwin Effect still significant for
  understanding AGN structure, physics
• Seen in broad and narrow lines
• BE calibration useful when shifting between
  measures of luminosity
• Correlation with L may trace fundamental
  correlation(s)
• with MBH, L/Ledd, no consensus yet?
• Some disagreement regarding PCA connection
• NLS1s outliers
• Existence of X-ray Baldwin Effect uncertain
• Sample details still very important
• Lots of progress, still work to do!