Transcript Extragalactic Distances from Planetary Nebulae

Planetary Nebulae
and the
Extragalactic Distance Scale
Robin Ciardullo
Why Measure Distances?
The HST Key Project and WMAP say that H0 = 72  8
km/s/Mpc. But …
We still need distances to individual
objects (AGN, MOND galaxies, etc.)
For precision cosmology, we need to
know H0 to better than 10%. Note also
that WMAP only constrains H0 if the
universe w = P /  = 1.
The current distance scale may contain
systematic errors.
 baryon  f ( N He ) H 02
 matter  f (Cl ) H 02

 f ( matter ) H 02
Why Observe Planetary Nebulae?
From Dopita et al. (1994)
They’re very bright !!
The brightest PN in a
galaxy have luminosities
greater than 6000 L!
They can routinely be
identified out to ~ 20 Mpc
Why Observe Planetary Nebulae?
They’re easy to
detect!
Why Observe Planetary Nebulae?
They’re easy to
detect!
Why Observe Planetary Nebulae?
They’re easy to
detect!
10% of the energy comes
out at 5007 Å
Why Observe Planetary Nebulae?
They’re easy to
detect!
[O III] 5007
[O III] Difference
Offband 5300
Why Observe Planetary Nebulae?
They’re a high precision distance indicator.
M86
M84
M87
There is at least one [O III]-bright PN for every ~ 108 L.
For large galaxies, the internal errors are a few percent!
Why Observe Planetary Nebulae?
They’re present in all galaxies!
Sb
Sc
Sa
E0
E3
S0
SBa
SBb
SBm
Irr
SBc
The History of PN as Distance Indicators
• 1966: First Suggestion PN as Distance Indicators [Hodge]
In “Galaxies and Cosmology” by Hodge (McGraw-Hill 1966)
The History of PN as Distance Indicators
• 1966: First Suggestion PN as Distance Indicators [Hodge]
• 1978: First PN-based distance estimate [Ford & Jenner]
• 1981: First use of PN for Local Group [Jacoby & Lesser]
• 1989: First use of Planetary Nebula Luminosity Function
[Ciardullo et al.]
• 1990: First PNLF-based Hubble Constant [Jacoby et al.]
• 1990: First Use of the PNLF in the Galaxy [Pottasch]
Reason for slow acceptance of method:
individual PN are definitely not standard candles.
(Distances to Galactic PN are uncertain to factors of ~ 5)
The Method
• Image the galaxy
through a redshifted
[O III] 5007 filter
The Method
• Image the galaxy
through a redshifted
[O III] 5007 filter
• Identify point sources
that are invisible in
the continuum
The Method
• Image the galaxy
through a redshifted
[O III] 5007 filter
• Identify point sources
that are invisible in
the continuum
• If the galaxy has starformation, compare
[O III] 5007 to H
The Method
• Image the galaxy
through a redshifted
[O III] 5007 filter
• Identify point sources
that are invisible in
the continuum
• If the galaxy has starformation, compare
[O III] 5007 to H
• Perform photometry
on a complete sample
of objects
M5007 = 2.5 log F5007 – 13.74
The Method
• Image the galaxy
through a redshifted
[O III] 5007 filter
M5007 = 2.5 log F5007 – 13.74
• Identify point sources
that are invisible in
the continuum
• If the galaxy has starformation, compare
[O III] 5007 to H
• Perform photometry
on a complete sample
of objects
• Fit to an empirical
function
N (M)  e 0.307 M {1 – e 3 ( M
*–M)
}
The Calibration of M*
The PNLF cutoff magnitude, M*, is calibrated via measurements
in galaxies with known Cepheid distances.
SMC*
M33
M101
NGC 3351
NGC 3627
LMC
M81
NGC 300*
NGC 3368
NGC 4258
M31
NGC 2403
NGC 5253*
* Denotes metal-poor, low-luminosity galaxy
M* = – 4.47
• M* is a constant for large, metal-rich galaxies
• Young and old populations have the same value of M*
• M* is fainter in small, low-metallicity galaxies.
This dependence was predicted by Dopita et al. (1992)!
M* = – 4.47
This consistency across > 1 dex in O/H strongly suggests
that neither the Cepheids nor the PNLF need additional
metallicity corrections.
Testing the Method
Are there systematic errors associated with the PNLF
method? To answer this question, we can …
• Perform Consistency Tests within Galaxies
• Perform Consistency Tests inside Galaxy Groups and
Clusters
Can we find a case where the PNLF fails???
Tests within a Galaxy
• Five galaxies have large enough samples of PN
to test for systematic shifts in the PNLF:





M31
M81
M33
NGC 5128
NGC 4494
In these galaxies, the PNLF samples within
galaxies always produce consistent distances
Samples of PN
in M31
The bulge, disk,
and halo PNLFs are
indistinguishable
(M* < 0.05 mag)
Tests within a Cluster
• Six Galaxy Groups have PNLF distances to multiple
galaxies:
 M81 Group: M81, NGC 2403
 NGC 1023 Group: NGC 891, 1023
 NGC 5128 Group: NGC 5102, 5128, 5253, (M83)
 Fornax Cluster: NGC 1316, 1380, 1399, 1404
 Leo I Group: NGC 3351, 3368, 3377, 3379, 3384
 Virgo Cluster: NGC 4382, 4472, 4486, 4649
(4374, 4406)
PNLF distances within clusters are always consistent
to within ~ 1 Mpc, with no systematic behavior
NGC 3379
The Leo I Group
NGC 3351
NGC 3384
NGC 3368
NGC 3377
All 5 galaxies within 1 Mpc
Group Diameter
NGC 4486
NGC 4649
NGC 4382
NGC 4406
The Virgo Cluster
NGC 4472
Background NGC 4406
Group Resolved
NGC 4374
Does the PNLF ever Fail?
Actually yes – in the Virgo Cluster, there are intracluster stars!!!
Intracluster PN can be foreground to the target galaxy, and
therefore appear “overluminous”.
Are Overluminous PN really Intracluster?
The density of PN within a galaxy should follow that of the
galaxy’s light. Intracluster PN should not; they should scale
with the area surveyed. “Overluminous” (foreground) objects
should therefore be (primarily) found in the outskirts of galaxies.
0.25 mag
The Identification of Intracluster Stars
Intracluster stars are not confined to the center of the Virgo
Cluster. They’ve recently been found foreground to NGC 4526.
Are there Systematic Errors in the Distance Scale?
The PNLF and Surface Brightness Fluctuations
Cepheids provide the calibration for both the PNLF and the SBF
method. Presumably, the distances of the two methods agree.
PNLF vs. SBF Distances
mean ~ 0.05 mag
The PNLF zero point is
calibrated via measurements
in 13 galaxies with distances
determined from Cepheids
mean ~ 0.05 mag
The SBF zero point is
calibrated via measurements
in 6 galaxies with distances
determined from Cepheids
How well do the two distance scales agree?
PNLF vs. SBF Distances (29 Galaxies)
• The curve is the
expected scatter in
the measurements
• There are 3 outliers.
The two worst are:
The zero points differ by ~ 0.15 mag!
We have found a systematic error!
Most Likely Explanation – Internal Extinction
To derive the PNLF cutoff magnitude, M*, using
Cepheid distances, one needs to know the reddening
M* = m* - Cep – R5007 E(B-V)
If the reddening is underestimated, then M* is
underestimated, and the inferred distance scale is
too small.
For the SBF method, however, the absolute fluctuation
magnitude, MI, depends on color, i.e.,
MI = C + 4.5 (V-I)0
So the zero point of the system is defined through
C = mI - Cep- RI E(B-V) - 4.5(V-I)obs + 4.5 E(V-I)
If the reddening is underestimated, then the color of
the galaxy is overestimated, MI is overestimated, and
therefore C is overestimated. The result is a distance
scale that is too large.
Because the SBF and PNLF methods react to reddening
in opposite directions, a small amount of internal
extinction can lead to a large discrepancy between the
two distance scales:
If only the SBF calibration
is affected
 = 4.2 E(B-V)
If both the PNLF and SBF
calibrations are affected
 = 7.7 E(B-V)
Only E(B-V) ~ 0.02 is needed to reconcile the PNLF
and SBF distance scales
This error propagates up the entire distance ladder!
The PNLF and SN Ia
Until recently, the sample of galaxies with well-observed Type Ia
supernovae and PNLF measurements was too small to be useful.
This is now changing.
(from Feldmeier, Philips, & Jacoby 2005)
PNLF vs. SN Ia
Distances
• The PNLF calibration
of the SN Ia distance
scale agrees well with
that of the Cepheids.
• (Any systematic error
between the scales is
less than ~ 4%)
PNLF, Cepheids, and Geometry
The PNLF distance scale is calibrated by Cepheids, and the
Cepheid scale assumes an LMC distance of (m-M)0 = 18.50.
This can be checked via 2 geometric distance measurements.
The LMC (SN 1987A Light Echo)
A key benchmark of the extragalactic distance scale is the light
echo measurement of SN 1987A. The classical analysis by
Panagia et al. (1991) gives a distance of 51.2  3.1 kpc.
A more complex model by Gould & Uza (1998) gives a lower
distance of D < 47.2  0.9 kpc.
NGC 4258 (Nuclear Maser)
Herrnstein et al. (1999) have analyzed the proper motions
and radial velocities of NGC 4258’s nuclear masers. The
orbits are Keplerian and yield a distance of 7.2  0.3 Mpc.
Geometry vs. Cepheids vs. PNLF
LMC
NGC 4258
Ratio
4258 - LMC
Method
Geometry
18.50  0.05
29.29  0.09
10.79  0.10
Cepheids
18.50*
29.44  0.12
10.94  0.12
PNLF
18.50  0.11
29.43  0.09
10.93  0.14
*Cepheid
and PNLF values are based on (m-M)LMC = 18.50.
Geometry vs. Cepheids vs. PNLF
LMC
NGC 4258
Ratio
4258 - LMC
Method
Geometry
< 18.37  0.04
29.29  0.09
10.92  0.10
Cepheids
18.50*
29.44  0.12
10.94  0.12
PNLF
18.50  0.11
29.43  0.09
10.93  0.14
*Cepheid
and PNLF values are based on (m-M)LMC = 18.50.
The perfect agreement between the relative distances argues
for a short distance to the LMC and a Hubble Constant that
is 7% larger than the Key Project value (77 km s-1 Mpc-1).
Why Does the PNLF Work???
The physics behind the PNLF is still controversial. However,
there are clues …
• A PN’s [O III] 5007 luminosity depends on the luminosity
of its central star. But there are mechanisms that can place a
limit on the maximum [O III] flux a PN can emit.
• Collisional de-excitation of
forbidden emission in young,
dense nebulae
Why Does the PNLF Work???
The physics behind the PNLF is still controversial. However,
there are clues …
• A PN’s [O III] 5007 luminosity depends on the luminosity
of its central star. But there are mechanisms that can place a
limit on the maximum [O III] flux a PN can emit.
• Collisional de-excitation of
forbidden emission in young,
dense nebulae
• Circumstellar extinction
around massive (high
luminosity) central stars
Why Does the PNLF Work???
The physics behind the PNLF is still controversial. However,
there are clues …
• A PN’s [O III] 5007 luminosity depends on the luminosity
of its central star. But there are mechanisms that can place a
limit on the maximum [O III] flux a PN can emit.
• Collisional de-excitation of
forbidden emission in young,
dense nebulae
• Circumstellar extinction
around massive (high
luminosity) central stars
Why Does the PNLF Work???
The physics behind the PNLF is still controversial. However,
there are clues …
• A PN’s [O III] 5007 luminosity depends on the luminosity
of its central star. But there are mechanisms that can place a
limit on the maximum [O III] flux a PN can emit.
• Collisional de-excitation of
forbidden emission in young,
dense nebulae
• Circumstellar extinction
around massive (high
luminosity) central stars
Why Does the PNLF Work???
The real problem comes from the absolute luminosity of the
PNLF cutoff …
 M* = 4.47 corresponds to a luminosity of 600 L
 To produce 600 L of [O III] emission, a central star must
have a luminosity of L > 6,000 L.
 A central star with L > 6,000 L must be more massive than
M > 0.6 M. Such stars come from M > 2 M progenitors.
(Weidemann 2000)
Why Does the PNLF Work???
Elliptical galaxies do not have
many (any?) 2 M main sequence
stars. But they do have large
numbers of 1 M stars. If some
are in close binary systems which
coalesce on the main sequence,
the product may evolve into an
[O III]-bright planetary.
The ratio of bright planetaries to
blue stragglers is about equal to
the ratio of the objects’ lifetimes.
Carrera et al. 2002
Summary
The Planetary Nebula Luminosity Function continues to be
a useful tool for extragalactic astronomy and cosmology.
• The PNLF is the only standard candle capable of measuring
distances to all the large galaxies of the local supercluster.
• The PNLF cutoff, M*, is the same for old and young
populations.
• PNLF comparisons with Surface Brightness Fluctuations
and Cepheid measurements suggest that small systematic
errors in the distance scale still exist.
• The PNLF and Cepheid calibrations of SN Ia are in good
agreement.
• The brightest PN in E/S0 galaxies may be the product of
binary star evolution. But they are NOT binaries.