Dust in Low Metallicity Blue Compact Dwarf Galaxies

1
Dust in Low Metallicity Blue Compact Dwarf
Galaxies
Vassilis Charmandaris, Yanling Wu, Jim Houck,
J. Bernard-Salas, V. Lebouteiller (University of
Crete Cornell University)?
J. Rosenberg (George Mason Univ.), IRS
Instrument team (Cornell Univ)?
2
Outline of Talk

• Some information on dust in the infrared
• Why should one be interested in low-Z galaxies
  and BCDs?
• Studying the BCDs with the Infrared Spectrograph
  (IRS) on Spitzer
• Presence of Polycyclic Aromatic Hydrocarbons
  (PAHs)?
• Estimating Elemental Abundances using infrared
  lines
• Spectral Energy Distribution
• Infrared to radio correlation
• Conclusions

3
Doing some physics Dust types

• Dust grains range in size from a few hundred Å
  to a few ?m. They are composed mainly of elements
  such as carbon and silicate compounds, and
  various kinds of ices.
• classical dust grains ? 0.1 ?m in size,
  containing ? 10000 atoms
• responsible for the FIR, sub-mm emission
• very small grains ? containing ? 100 atoms
  (less than10nm in size)
• responsible for a rising continuum 10?m
  
• PAHs (Polycyclic Aromatic Hydrocarbons) ?
  benzene rings contain N ? 50 atoms
• trace photodissociation regions.
• Due to the size distribution of grains
  variations in the underlying radiation field the
  dust temperature varies (cold, warm, hot dust)?
• Spectra of different dust components are fitted
  by modified Planck curves
• I? ? ?? B?(T)?
• Power-law emissivity ?? ? ?n (i.e.
  Dale et al, ApJ, 2001)?
• Really hot dust (200-1500K) in a equilibrium is
  observed via a near/mid-IR bump close to tori
  of AGNs.

4
Stochastic Heating of Grains

• A far-UV photon hits a dust grain and ejects an
  electron
• The ejected photoelectron heats the gas (very
  inefficiently 0.1 - 1 )?
• 50 of gas heating is due to grains of sizes lt 15
  Å
• Subsequently the gas cools via far-IR emission
  lines ( OI 63 ?m , CII 158 ?m)?
• Process of randomly heating dust grains to high T
  (1000K) for short periods (1s)?
• Not in equilibrium.
• PAHs, dominate the mid-IR (5-20 ?m) flux in
  normal galaxies and quiescent star forming
  regions via the so-called Unidentified IR Bands
  (UIBs) or IR Emission Features (IEFs).
• Normal Galaxies L(mid_ir) 20 L (IR)
  (Dale et al. 2001, Roussel et al 2001)?
• Active/Interacting galaxies L(mid_ir) lt 5 L
  (IR) (Armus et al. 2007 Marshall et al .
  2007)?
• In normal late type galaxies most of their energy
  is released in the FIR.
• In AGN the high UV/X-ray flux can sublimate the
  grains and lead o destruction of PAH features and
  display a bump of thermal emission at 3-6microns
  and a power law spectrum.

5
PAH Normal Modes

• 10-20 of the total IR luminosity of a galaxy
• Tens - hundreds of C atoms
• Bending, stretching modes ? 3.3,6.2,7.7,8.6,11.2,
  12.7 ?m
• PAH ratios ? ionized or neutral, sizes,
  radiation field, etc.

Leger Puget (1984)? Sellgren (1984)? Desert, et
al. (1990)? Draine Li, (2001)? Peeters, et al.
(2004)?
6
Energy Balance
Helou et al., ApJ, 2001
7
Dust in Low Metallicity Environments

• Chicken Egg Problem
• Dust grains act as catalyst for the formation of
  molecules and stars
• Dust is created at the late stages of stellar
  evolution
• If the first grains were created by Pop III
  stars, what were their properties?
• Do we see complex organic molecules (ie PAHs) at
  low metallicities?
• Challenges
• Low abundances in the ISM lead to formation of
  high mass stars.
• High mass (early O type) stars produce more high
  energy UV photons.
• UV photons dissociate / destroy C-H chains
• Lack of metals in the ISM lead to reduced
  efficiency in the gas cooling, hence UV photons
  propagate at longer distances
• Where to look?
• Distant High-z QSOs - Evidence of thermal dust
  emission at z6
• Even at z6 metallicity is elevated
• Galaxies are faint for mid-IR spectroscopy even
  for Spitzer
• Identify nearby low metallicity systems with
  recent bursts of star formation

8
Low Metallicity Spitzer Samples

• IRS/GTO Sample
• A total of 61 Blue Compact Dwarfs observed
• Metallicities spanning from 0.02 to 0.6 Solar
• IRS Spectra obtained for 16 of the galaxies
• Additional 16 and 22um images for the whole
  sample
• Rosenberg NDWFS/KISS Sample
• A total of 19 star forming galaxies with MBgt
  -18mag within the NDWFS area
• Spectroscopy available via KISS (AGN excluded)?
• Abundances 12log(O/H) 7.8 --gt 9.1 (median
  metallicity 8.05 0.25 Solar)?
• IRAC, MIPS and IRS16um data available
• MIPS/GTO Sample (Engelbracht 2005,2008)?
• GO Samples (Thuan, Hunt et al)?

9
The Blue Compact Dwarfs (BCDs)?

• The BCDs are
• Blue colors typical of O,B,A stars
• Compact dimensions less than 1kpc
• Low Luminosity MB gt -18mag
• Dominated by a recent burst of star formation
• Have low metallicity (lt 0.3 Solar)?
• The are interesting to study
• They are found nearby (Second Buyrakan Survey)?
• Hierarchical galaxy formation scenario proposes
  that bigger structures are built from smaller
  galaxies. Building-block galaxies are too faint
  and small to be studied at high-redshift.
• Review Kunth Ostlin 2000

10
IRS spectroscopy of low metallicity galaxies
Wu et al. 2006
11
Mid-IR spectra as a function of metallicity

• Engelbracht et al. 2008

12
Decompositions of PAHs

• (SINGS - Smith et al. 2007)?

13
Fractional PAH power
(SINGS - Smith et al. 2007)?
6.2µm, 12.6µm gt 10 7.7µm gt 30 11.2µm, 17µm
gt 15
14
A metallicity threshold in PAH strength?
15
PAHs vs Metallicity in BCDs
Wu et al. 2006
16
PAHs vs Neon line ratio
Wu et al. 2006
Starbursts
NeII 21.56eV NeIII 40.96eV
see Brandl et al. 2006
17
PAHs vs proxy of radiation field density
Wu et al. 2006
18
Metallicity and presence of PAH
Rosenberg et al. 2007
Engelbracht et al. 2005
19
Star Formation Rates of BCDs

• The metallicity-corrected SFRs agree well with
  those estimated from radio continuum, though none
  might reflect the true SFRs.
• Both SFRs from IR and radio are underestimated,
  but with a similar factor (Bell 2003)?

20
Star Formation Rates as a function of metallicity
Rosenberg et al. 2007

• Substantial scatter depending on the method
  used.
• Attention in high-redshift systems

21
Elemental Abundances Infrared vs Optical

• Pros
• Metallicity measurements from the optical suffer
  from the dust extinction.
• The infrared lines are much less sensitive to the
  uncertainties in the electron temperature.
• Cons
• Need at least one hydrogen line
• Hu? (12.37?m) line is usually very faint in BCDs.
• H? converted from H? suffers extinction effects.
• Need to make aperture correction so that H? from
  optical corresponds to the same area as covered
  by the IRS SH slit.
• Test
• If there are embedded dust enshrouded regions
  low-metallicity galaxies they will likely have
  different (higher) metallicities than the regions
  exposed in the optical.

22
Elemental Abundances II

• Parameters
• Extinction adopted E(B-V) from optical studies
  (exception SBS0335-052extinction estimated from
  silicate feature)?
• Te from optical if available, or assume Te10,000
  K
• Ne from optical if available, or assume Ne 100
  cm-3
• Lines
• We use the lines in the SH module to remove the
  uncertainty on the scaling factors between the SH
  and LH spectra.
• Ne NeII 12.81 ?m, NeIII 15.55 ?m
• S SIV 10.51 ?m, SIII 18.71 ?m
• Other Ionization stages Ionization Correction
  Factors (ICFs)?
• For Neon NeIV lt1
• For S SII 10
• Method
• Solve the statistical equilibrium equation for a
  five level system
• Effective collisional strengths from IRON project

23
Ne/H (IR) correlates with S/H (IR)?
24
Neon and Sulfur abundances - IR vs Optical
Wu et al. (2008)?

• Overall consistent results between optical and
  IR studies
• Slightly elevated Ne values as estimated in the
  IR

25
Ne/S (IR) vs O/H (Optical)?

• Discrepancy in Ne/S estimates - already been
  seen in PNe (Bernard-Salas 2002)?
• Possibly due to optical Ionization Correction
  Factors - Temperature dependence in optical
  (TeOIII)?

26
(Crude) SED template fitting
Rosenberg et al. 2007

• The SEDs of low-metallicity dwarfs are warmer

27
(Crude) SED template fitting (2)?
Rosenberg et al. 2007
28
Average SEDs
Engelbracht et al. 2008
29
Radio to far-IR correlation in BCDs

• The low luminosity dwarf galaxies appear to have
  a similar slope as compared to that of normal
  galaxies.
• ltqFIRgt2.3 0.2 (Condon et al. 1992)?
• ltqFIRgt2.37 0.26 (this sample)

Wu et al. (2008b)?
30
Mid-IR to Radio Correlation - q24

• Mid-IR emission also traces star formation
  activity, and shows more variation than FIR.
• Mid-IR luminosities are becoming available for
  low luminosity systems from deep Spitzer surveys.
• ltq24gt1.25 0.41
• lt q24gt0.94 0.23
• (Appleton et al. 2004)?

Wu et al. (2008b)?
31
q24 as a function of metallicity

• No clear correlation is seen between the q24 and
  the galaxy metallicities.
• q24 appears to decrease with metallicity in
  metal-poor sources
• SBS0335-052E is an outlier.

Wu et al. (2008b)?
32
SBS 0335-052E Z 1/41 Zsolar
33
SBS 0335-052E Z 1/41 Zsolar
34
IZw18 Z 1/50 Zsolar
35
Comparing SBS0335E IZw18
36
Conclusions / Perspectives

• Increasing metallicity (and the presence of an
  AGN) shift PAH power to longer wavelengths (with
  the latter effect the stronger).
• Decreasing metallicity (and the presence of an
  AGN), suppress PAH emission and decrease their
  fractional contribution to the total IR
  luminosity.
• No PAH emission has been detected in systems
  with 12 log(O/H) lt 8.0 (small number
  statistics)?
• SED template fitting appears consistent with this
  finding.
• Infrared to Radio correlation holds in BCDs
• Possible decrease in mid-IR to radio with
  metallicity
• Warmer dust at lower metallicities
• Unclear why SBS0335-052E and IZw18 are so
  different.

37
(No Transcript)