Wolf-Rayet Galaxies: An Overview

1
Wolf-Rayet GalaxiesAn Overview

• William D. Vacca
• (USRA-SOFIA)

2
Wolf-Rayet Galaxies

• Subset of emission-line galaxies (or major
  portions thereof) in whose integrated (optical)
  spectra the signatures (emission features) of W-R
  stars are found
• Defined by detection of broad (stellar) He II
  4686 or blue bump ( He II 4686 N III 4640
  C III 4650) from W-R stars
• Other broad lines He II 1640, C III 5696, C IV
  5808
• Most are H II galaxies photoionization
  powered by hot stars e.g., BCDs, although the
  class encompasses a wide range of galaxy types
  and morphologies (LINERs, Sy 2s, ULIRGs)
• Represent the more luminous extension of
  extragalactic GHIIRs (Conti 1991)

3
Examples of Spectra
Vacca Conti (1992)
Kunth Schild (1986)
NGC 3125
POX 4
4
More Examples of Spectra
Schaerer, Contini, Kunth (1999)
5
Ancient (pre-1998) History

• First example (He 2-10) found in 1976 (Allen et
  al.)
• First catalogue (Conti 1991) had 37 objects,
  found serendipitously
• Large N(WR) (102-105) and large N(WR)/N(O) (gt
  0.1-1) derived from L(He II 4686) and L(He II
  4686)/L(H?)
• Because W-R stars are short-lived descendants of
  the most massive O stars, this suggested W-R
  galaxies represented a brief (?t lt few Myr) burst
  of massive star formation observed at a
  propitious time (? lt few Myr later) (Kunth
  Sargent 1981 Durret et al. 1985 Armus et al.
  1988 VC92)
• Early Pop Syn Models (Arnault, Kunth Schild
  1989 Mas-Hesse Kunth 1991 Krüger et al. 1992
  Cervino Mas-Hesse 1994) confirmed general
  picture
• Short Burst, Salpeter IMF, Muppgt30 M?, 3 lt ? lt 6
  Myr
• Strong variation of N(WR) and N(WR)/N(O) with
  metallicity Z

6
Model Predictions

• Arnault Kunth Schild (1989)
• ?2, Mupp 120 M?
• N(W-R) increases with Z
• N(W-R)/N(O) for IB gtgt CSF

7
More recently

• Second catalogue (Schaerer, Contini, Pindao
  1999) listed 139 objects
• 40 have both WN and WC stars
• Strong variation in N(WR), N(WC)/N(WN), and
  N(WR)/N(O) with Z
• Larger samples and better optical data with
  higher S/N and R have enabled detailed studies of
  numerous objects
• Schaerer et al. (1997) WN, WC stars in SSCs in
  NGC5253
• Izotov et al. (1997) Legrand et al. (1997) WN,
  WC stars in I Zw 18
• Schaerer, Contini, Kunth (1999) WC stars in
  W-R galaxies
• Guseva, Izotov, Thuan (2000) W-R populations
  in 39 BCDs
• Schaerer et al. (2000) extended bursts in ZgtZ?
  W-R galaxies
• Starburst regions in W-R gals composed of compact
  SSCs
• Presence of W-R stars provides means of
  age-dating
• UV and optical data for W-R galaxies but still no
  convincing detections of W-R features in the IR

8
Estimating N(WR) and N(O)

• From FWHM of He II 4686 and NIII 4640He II 4686,
  dominant WN subtype is usually WNL
• From FWHM of CIV 5808 and absence of C III 5696,
  dominant WC subtype is usually WCE
• N(O) is estimated from L(H?) (which yields Q0obs)
  and EW(H?) (which yields ?, derived from models)

9
Standard Models (Schaerer Vacca 1998)

• Geneva (non-rotating) stellar evolution tracks
  with enhanced mass-loss rates as function of
  metallicity (0.05 lt Z/Z? lt 2.0)
• CoStar theoretical fluxes for O stars
• Spherical, expanding, unblanketed, non-LTE models
  of Schmutz et al. (1992) for W-Rs
• Empirical estimates of Of and W-R line fluxes
  from Gal and LMC stars
• No scaling of W-R models or line fluxes with Z
• Nebular continuum
• Instant. Burst (?t 0) with Salpeter IMF
  (?2.35), Mupp 100 M?
• Predict relative W-R numbers, luminosities of
  lines and W-R blue bump L/L(H?), and EWs as a
  function of Z, age ?, EW(H?)
• Extended to lower Z, finite duration bursts,
  non-Salpeter IMFs, inclusion of R136-type stars,
  newer line-blanketed O and W-R models (de Mello
  et al. 1998 Schaerer et al. 1999, 2000 Pindao
  et al 2002 Smith et al. 2002)

10
Example of Model Predictions
SV98
11
Comparisons with Models
Guseva, Izotov, Thuan (2000)
12
Comparisons with Models
Guseva, Izotov, Thuan (2000)
13
Comparisons with Models
Guseva, Izotov, Thuan (2000)
14
Caveats and Problems

• Calibration of LWN(4686) and LWCE(5808) based on
  Gal, LMC W-Rs
• Huge range in line luminosities within any single
  WR subtype
• For ZSMC Crowther Hadfield (2006) find smaller
  line fluxes
• Contamination in low resolution spectra by
  nebular emission
• Disentangling contributions to W-R broad features
  from WC and WN stars can be difficult
• L(Hß) and EW(Hß) may not accurately reflect hot
  star population in either number or age
• Narrow slit captures only fraction of L(Hß)
  geometric dilution
• Stars and emitting gas may be spatially separated
• Stars and gas may have different extinction
  values
• Dust absorbs some of the ionizing photons
• Nebula may not be ionization bounded (photon
  leakage)
• Underlying older population contributes to
  L(4861) continuum dilution

15
I Zw 18 A Challenge to the Models?

• With Z Z?/50, I Zw 18 should have few W-Rs and
  even fewer WC stars
• Izotov et al. (1997) find N(WNL)17, N(WCE)5,
  N(WC)/N(WN) 0.3, N(W-R)/N(O) 0.02
• Re-analysis by De Mello et al. (1998) gives
  N(WNL) 4, N(WCE) 4, for N(WC)/N(WN) 1 !
• Std IB models can reproduce observed EWs and
  N(W-R)/N(O) but not line fluxes
• Crowther Hadfield (2006) use SMC line
  luminosities to estimate N(WCE) 30 and N(WNL)
  10-200, so that N(W-R)/N(O) 0.02-0.1 !
• May require models with rotation and/or binaries
  to produce more WRs at low Z

IB, ?2.35 Mupp150 M?
De Mello et al. (1998)
16
A Better Way

• Target simple, isolated objects representing
  SSPs formed in Instantaneous Bursts (?t 0, no
  continuum dilution e.g., SSCs)
• Use model fits to UV spectral line profiles to
  determine the age ?
• Use observed slope of the UV continuum compared
  to models to estimate extinction
• Match models to continuum levels to derive Mass,
  N(O)
• Use synthetic or empirical generic W-R spectra
  to match both UV and optical emission features
  and derive N(WN) and N(WC)
• Not perfect (sensitive to extinction law, matched
  UV and optical apertures) but avoids problems of
  deriving N(O) from gas
• Applied (in various forms) to
• 16 W-R galaxies Mas-Hesse Kunth (1999)
• NGC 3049 Gonzalez Delgado et al. (2002)
• NGC 3125 Chandar, et al. (2004) Hadfield
  Crowther (2006)
• He 2-10 Chandar et al. (2003)
• Tol 89 Sidoli, Smith, Crowther (2006)

17
NGC 3125 - An example(Hadfield Crowther 2006
Chandar et al. 2004)
18
NGC 3125
NGC 3125 A1 Chandar, Leitherer, Tremonti
(2004)
Hadfield Crowther (2006)
19
NGC 3125(Hadfield Crowther 2006)

• Fitting SB99 models to wind line profiles gives
• ? 4 Myr
• Continuum fit gives
• M 2x105 M?
• N(O) 550
• He II 1640 line gives
• N(WN) 110
• N(WR)/N(O) 0.2

NGC 3125 A1 Hadfield Crowther (2006)
20
NGC 3125(Hadfield Crowther 2006)

• Fit LMC template spectra (Z 0.5Z?)
• For A1
• N(WN5-6) 105
• N(WCE) 20
• Agree with UV analysis
• For B
• N(WN5-6) 40
• N(WCE) 20

21
NGC 3125(Hadfield Crowther 2006)

• SB99 models with Kroupa IMF, Mupp 100 M? at ?
  4 Myr yields optical cont. fits consistent with
  UV and pop analyses
• A
• N(O) 1150
• N(WR)/N(O) 0.16
• M 4.2 x 105 M?
• B
• N(O) 450
• N(WR)/N(O) 0.13
• M 1.6 x105M?

22
Wolf-Rayet Galaxies in the SDSS

• Zhang et al. (2007) constructed a sample of 174
  W-R galaxies
• Brinchmann, Kunth Durret (2008) generated a
  sample of 570 W-R galaxies with z lt 0.22 !
• Compared to SB99 and BC03 models with
  SV98/Crowther Hadfield (2006) W-R and Of line
  fluxes
• Considered finite burst durations ?t between 1
  Myr and 0.5 Gyr
• Serious discrepancy with models at lowest Z
• Suggest models with rotation and binaries are
  needed

Brinchmann et al. (2008)
23
Wolf-Rayet Galaxies in the SDSS
Brinchmann, Kunth, Durret (2008)
24
Wolf-Rayet Galaxies at High Redshift!
811 Lyman Break Galaxies (Shapley et al.
2003) z 3
Stellar He II 1640 FWHM 1500 km/s EW 1.3 Å
25
Wolf-Rayet Galaxies at High Redshift!
Brinchmann et al. (2008)

• Bruzual Charlot (2003) synth models
• SV98 Crowther Hadfield (2006) WR and Of line
  fluxes
• Chabrier (2003) IMF
• SFR exp(-t/?) ?15 Gyr

26
Summary

• W-R galaxies are the result of short bursts of
  massive star formation observed during a brief
  and special time shortly after the onset of the
  burst
• WR phenomena in starburst galaxies are a normal
  part of evolution of young starbursts. (Conti
  1999)
• Now have a sample of 570 plus some at high
  redshift!
• Integrated, multi-wavelength analysis provides
  best way of comparing observations with models
• Updated standard models do a reasonably good
  job of matching the observed EWs and relative
  line fluxes at most metallicities, and overall
  trends with metallicity, with Salpeter IMF and
  large Mupp (gt 30 M?)
• General picture is probably correct
• But serious problems at the lowest metallicities
• May require models with rotation and/or binaries
• New models are under development
• So quick bright things come to confusion.
  (Shakespeare, Midsummer Nights Dream, Act I
  scene 1)