Transcript The Chemical Evolution of Dwarf Spheroidals

The Chemical Evolution of
Dynamically Hot Systems
Michael Richer
Observatorio Astronómico Nacional
Instituto de Astronomía, UNAM, Mexico
collaborators: Marshall L. McCall
Grażyna Stasińska
photo: José Alberto López
The road may be difficult…
• Measuring the chemical abundances in extragalactic PNe
is relatively straightforward.
• Interpreting those abundances to infer the chemical
evolution of their host galaxies is not necessarily
straightforward.
photo: José Manuel Murillo
Why are PNe of interest?
•
In a star-forming galaxy
– the current ISM abundances may be determined from H II regions. Usually, the
metallicity is characterized by the oxygen abundance.
– the past ISM abundances may be derived from stars and PNe, but the epoch
corresponding to these abundances is uncertain.
•
•
In a DHS (ellipticals, dwarf spheroidals, and the bulges of spirals)
– there is no star formation and often no ISM, so some probe besides H II regions
will be necessary.
– it is still possible to measure abundances in stars and PNe, with the same
uncertainties as above.
– it is often difficult to observe stars unless they are much brighter than the
background or the background is uncrowded.
– abundances have traditionally been measured from integrated spectra and
calibrated in terms of the iron abundance, e.g., using the Mg2 index.
PNe are one of the few direct probes of chemical abundances in DHSs.
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
Spectroscopy of extragalactic PNe
•
Spectroscopy of extragalactic PNe is difficult due to the need/desire to detect
the faint, temperature-sensitive line of [O III]l4363.
– He IIl4686 as well as low ionization lines can also be intrinsically faint, but this is
generally more manageable.
•
•
Subtracting the light of the background galaxy (often bright) can also be
complicated because of its spatial structure and spectral properties.
Good discusions of the observational challenges are given by
– Jacoby & Ciardullo 1999, ApJ, 515, 169 (multi-slit spectroscopy)
– Walsh et al. 1999, A&A, 346, 753
(long slit spectroscopy)
– Roth et al. 2004, ApJ, 603, 531
(integral field spectroscopy)
•
•
Published data have come from 4m class telescopes.
Observations with 8m telescopes should allow detection of [O III]l4363 in PNe
out to at least 5 Mpc, farther if high dispersion (R ~ 10,000) or high spatial
resolution are used to suppress the galaxy background.
Spectroscopy: reliability
•
There is still little overlap
between observational
samples among different
observers.
On “simple” backgrounds, the
line intensities are probably
uncertain to within 25% for
lines brighter than Hb and to
within a factor of two for lines
half the intensity of Hb.
– uncertainty in line fitting and
continuum placement
– uncertainty in reddening
– uncertainty in atmospheric
extinction, refraction, etc.
10000
M32
1000
new line intensities
•
100
10
1
1
10
100
1000
previous line intensities
10000
What chemical abundances can be
studied in PNe?
•
Nucleosynthesis in the stellar progenitors of all PNe modify their initial
abundances of He, C, N, and s-process elements.
– Theoretical models of AGB stars are still incapable of producing low-mass carbon
stars, M < 3M, so it is not clear how efficient the third dredge-up is for the lowest
mass PN progenitors.
– It is clear that many PNe of all types are C-rich.
•
•
•
Nucleosynthesis in the stellar progenitors of some PNe may modify their initial
abundances of O and possibly Ne.
– Theoretical studies indicate that this affects PN progenitors of higher mass,
typically, M > 3M. The extent and sign of the effect it is unclear.
Except for the abundance of He, all abundances in extragalactic PNe are
determined from collisionally-excited lines (forbidden lines).
Note that chemical abundances may only be studied in the intrinsically
brightest extragalactic PNe, i.e., it is not possible to study the entire
population.
photo: José Alberto López
Nucleosynthesis in PN progenitors
•
•
The PN progenitors can
modify N/O much more
than He/H.
In all galaxies, the N/O
ratios in bight PNe cover
the same range.
Nucleosynthesis and
dredge-up in the
progenitors of bright PNe
then appears to be the
same everywhere.
photo: José Alberto López
1.5
MW
M32
N185
N205
LMC
SMC
Fornax
BB He
M31
Sgr
1.0
log N/O (dex)
•
0.5
0.0
-0.5
-1.0
-1.5
-2.0
0.00
0.05
0.10
0.15
He/H
0.20
0.25
0.30
O/H and Ne/H in bright PNe
9.0
12 + log Ne/H (dex)
8.5
8.0
7.5
7.0
6.5
Milky Way
LMC
NGC 185
Fornax
SMC
Henry
6.0
5.5
7.0
7.5
8.0
8.5
12 + log O/H (dex)
photo: José Alberto López
M31
M32
NGC 205
Sgr
Vigroulx et al
9.0
9.5
• In the ISM of star-forming
galaxies, the production of
O and Ne is believed to be
dominated by type II SNe.
• Bright PNe in all galactic
systems studied thus far
follow the relation between
O and Ne abundances
observed in the ISM in starforming galaxies.
• The simplest conclusion is
that the stellar progenitors
of bright PNe do not modify
their original O and Ne
abundances significantly,
i.e., at the 0.3 dex level for
these observations.
Issues that affect the interpretation of
chemical abundances in PNe
•
•
•
SAMPLE SELECTION: often based upon high [O III]l5007 luminosity
nucleosynthesis during the evolution of the PN progenitor stars
the history of star formation
– higher stellar death rates from younger stellar populations
•
the evolution time scale of post-AGB stars
– evolution time scale from AGB to PN phases
– absolute evolutionary time scales
– the initial mass-final mass relation
•
mass loss during the AGB, post-AGB (and RGB?) phases
– relationship between nebular morphology and progenitor mass?
•
•
All of the above may/should affect the population of the PNLF.
If so, they could also affect the interpretation of PN abundances, since only the
brightest PNe may usually be studied (usually “luminosity” = [O III]l5007
luminosity).
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
Observations: systems with on-going
star formation
•
•
•
In the Magellanic Clouds, young stars and H
II regions also have similar O/H (Hill et al.
1995, A&A, 293, 347).
In NGC 6822, young stars, H II regions, and
bright PNe have similar abundances (Venn
et al. 2001).
The stellar abundances are determined from
recombination lines while the chemical
abundances in the PNe and H II regions are
determined from forbidden lines.
In the LMC, the mean abundance found for
a sample of PNe depends upon the
luminosity range spanned by the sample
(Richer 1993, ApJ, 415, 240).
–
–
8.6
12 + log O/H (dex)
•
8.4
8.2
8.0
SMC
stars
LMC
H II regions
7.8
0.0
0.5
1.0
1.5
2.0
2.5
sample luminosity range (mag)
more luminous PNe  higher mean O/H.
O/H in brightest PNe = O/H in H II regions.
Venn et al. (2001):  = young stars,  = H II regions,  = PNe
photo: José Alberto López
Observations: systems where star
formation stopped “long ago”
•
Milky Way bulge:
– McWilliam & Rich (1994) found a mean [Fe/H] of -0.25 dex and [O/Fe] > 0 for
bulge giants.
– The bright PNe in the Milky Way bulge have a similar mean [O/H] (-0.26 dex;
Stasinska et al. 1998) if the Anders & Grevesse (1989) solar abundance is used
(-0.02 dex using the Allende-Prieto et al. 2000 scale).
•
Fornax and Sagittarius dwarf spheroidals:
– Various recent studies find that O/H in the more metal-rich stars is similar to that
found in the PNe (Anders & Grevesse 1989 scale).
– It is unlikely that their PNe arise from more metal-poor stellar populations, since
the [O/Fe] ratios implied would then be extremely high (difficult to understand).
– These two galaxies are problem cases given their small masses and PN
populations. Their PNe are probably also not intrinsically bright.
•
Given the example of the Milky Way, it will therefore be assumed that the
mean O/H in the PNe is the same as that in the stars.
– Suitable (downward) corrections will be made for Sagittarius and Fornax.
•
This is the crucial assumption upon which the conclusions depend.
photo: José Alberto López
Metallicity-luminosity relation for DHSs
•
•
DHSs follow a metallicity-luminosity relation.
The mean O/H in DHSs is higher than that observed in equally luminous dwarf irregulars.
The maximum O/H in DHSs differs even more.
This is also seen in the He abundances in the dwarf spheroidals (though not the bulges of
M31 & Milky Way): at a given O/H, He/H is higher in the PNe in dwarf spheroidals than in
their counterparts in dwarf irregulars.
If they have evolved from dwarf irregulars like those observed today, dwarf spheroidals
must have faded by at least ~4 mag, which is unlikely.
DHSs appear to have incorporated their O production into stars more efficiently than
today’s dwarf irregulars.
•
•
•
Milky Way bulge
9.0
M32
Sagittarius
NGC 185
MW
M32
N205
M31
N185
LMC
0.30
SMC
Sgr
Fornax
BB He
Fornax
NGC 205
8.5
He/H
12 + log (O/H) (dex)
M31 bulge
0.40
8.0
0.20
0.10
7.5
dwarf irregulars
DHS mean O/H
DHS max O/H
0.00
7.0
6.5
-20
-18
-16
-14
M B (mag)
photo: José Alberto López
-12
7.0
7.5
8.0
8.5
-10
12 + log O/H (dex)
9.0
9.5
Supernova-driven winds?
[O/H] + log((SNII +SNIa)/SNII ) (dex)
0.25
•
0.00
-0.25
•
-0.50
-0.75
-1.00
-1.25
0.8
1.0
1.2
1.4
1.6
1.8
log  c (km/s; dex)
photo: José Alberto López
2.0
2.2
2.4
The oxygen abundances in
DHSs appear to correlate with
the stellar velocity dispersions.
Such a correlation is expected if
supernova-driven winds are the
agent that terminates chemical
evolution in these systems.
Gas outflows during star formation?
•
•
•
•
In order to simultaneously fit the means and
dispersions in O/H observed in DHSs with a
model of chemical evolution that incorporates
supernova-driven winds, it is necessary to
allow gas outflow while the galaxies are
forming stars.
The large dispersions in O/H are another
indication of efficient incorporation of oxygen
production into stars.
Gas outflow increases the efficiency of the
incorporation of newly-synthesized elements
because the ISM loses mass during the life
span of the stars responsible for enriching the
ISM.
An efficient incorporation of newlysynthesized elements into stars need not
imply an efficient conversion of matter into
stars.
photo: José Alberto López
McCall & Richer (2003, in IAU Symp 209, p. 583)
Conclusions
•
•
It appears to be feasible to use O/H in bright PNe as probes of the chemical
evolution in DHSs.
DHSs follow a metallicity-luminosity relation.
– Supernova-driven winds are a natural explanation for such a relation in galaxies
that are now devoid of interstellar matter.
•
•
•
•
The metallicity-luminosity relation for DHSs is displaced to higher O/H
compared to that followed by dwarf irregulars.
It is unlikely that this displacement is due to fading of the DHS population, but
rather to a more efficient incorporation of their element production into new
stars.
Initial modelling of these results indicates that DHSs suffered gas outflows
even while forming their stars.
Making better use of PNe as probes of chemical evolution requires a better
quantitative understanding of PN evolution.
photo: José Alberto López
The history of star formation?
Connecting PNe and stars
•
In star-forming galaxies,
– the masses and ages of the
different stellar populations matter.
– the younger stellar populations
have higher death rates, so the
mean O/H measured in bright
PNe is biased towards the value
observed in the ISM.
•
Richer et al. (1997, A&AS, 122, 215)
In galaxies where star formation
stopped “long ago” (> 1 Gyr),
– the masses of the different stellar
populations matter.
– the difference between the mean
O/H in bright PNe and the value
that was achieved in the ISM
grows as a function of the final
O/H achieved in the ISM.
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
Evolution from AGB to PN
•
•
•
•
The envelope mass of the
central star when it evolves off
the AGB has a dramatic impact
upon the time scale for
evolution from the AGB to PN
phases.
The number of bright central
stars is a strong function of the
envelope mass.
The duration of the phase
during which central stars are
bright is also a function of the
envelope mass.
This issue may not matter for
studies of chemical evolution if
only galaxies with or without
star formation are considered.
Me = f(M)
R
MeR random;
MeR < 0.1 M
Mc = 0.569 M
MeR constant;
MeR < 5x10-3 M
Mc = 0.535 M
MeR random;
MeR < 10-4 M
Mc = 0.9 M
Stanghellini & Renzini (2000, ApJ, 542, 308): : PN nuclei; : wind objects;
: proto PN nuclei; : post PN nuclei
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
The time scale of post-AGB evolution?
• Simple models indicate that
PNe should cluster in
defined regions of these
diagrams, but observed
data do not.
• To fit the data for M31 and
the LMC requires different
evolution for both the
nebular shells and the
central stars.
• Is this a problem with the
models, the time scale of
post-AGB evolution, or
something else?
• Observations of faint PNe
should be able to solve this.
Stasinska et al. (1998, A&A, 336, 667)
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
PN morphology
•
At least in the Magellanic Clouds,
– there is a lot of mixing between [O
III]l5007 luminosity and
morphology
– this mixing is not so strong for
Balmer lines
•
•
•
In the Milky Way disk, PN
morphology varies with scale height
above the disk plane, so should be a
function of progenitor mass.
Morphology likely affects PN
luminosity since it will affect the
angular distribution of the optical
depth.
This issue may also not matter for
studies of chemical evolution so long
as only galaxies with or without star
formation are considered.
Stanghellini et al. (2003, ApJ, 596, 997): : round, : point-symmetric;
: elliptical; : bipolar core; : bipolar or quadrupolar
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)
Connection between nebular and central
star evolution?
left: Villaver et al. (2002, ApJ, 581, 1204)
right: Richer & López (unpublished)
•
•
•
How long are PNe bright? Upon what progenitor properties does this depend?
Expansion velocities are easily measured within the Local Group…
Hydrodynamical models that account for the connection between the evolution of the
central star and that of the nebular envelope indicate that the relation between the nebular
kinematic age and the central star age is complicated.
NASA, NOAO, ESA, the Hubble Helix Nebula Team, M. Meixner (STScI), and T.A. Rector (NRAO)