Starbursts and poststarbursts in the Sloan Digital Sky Survey

1
Starbursts and poststarbursts in the Sloan
Digital Sky Survey
2
People involved

• Anna Blomqvist (Masters), Uppsala
• Emma Holst (Masters), Uppsala
• Daniel Malmberg (Masters), Uppsala
• Thomas Marquart, Uppsala
• Elisabet Leitet, Uppsala
• Göran Östlin, Stockholm
• Michael Way, Goddard
• Erik Zackrisson, Stockholm

3
People involved

• Anna Blomqvist (Masters), Uppsala
• Emma Holst (Masters), Uppsala
• Daniel Malmberg (Masters), Uppsala
• Thomas Marquart, Uppsala
• Elisabet Leitet, Uppsala
• Göran Östlin, Stockholm
• Michael Way, Goddard
• Erik Zackrisson, Stockholm

4
What is a starburst?

• The mean gas consumption timescale is
  significantly smaller than the Hubble age (as in
  blue compacts)
• or
• That this is true for a local region in the
  centre a nuclear starburst (prototype NGC
  7714).
• and that
• Starbursts have high star formation efficiencies,
  i.e they use up a larger proportion of the
  molecular cloud it is formed from than normal
  before the cloud disperses or dissociates.

5
Important questions

• Can galaxy interactions (close encounters)
  trigger starbursts? Global or nuclear or both?
• Or do the interactions trigger bar formation
  which then triggers nuclear starbursts?
• Can mergers trigger global starbursts?
• Can interactions or mergers trigger AGNs?
• Can nuclear starbursts trigger AGNs?

6
Interactions and mergers trigger only moderately
enhanced star formation. Starbursts are rare
phenomena contributing insignificantly to the
present day gas consumption.See e.g.

• Bergvall et al. 2003, AA 405, 31
• Brosch et al. 2004, MNRAS 349, 357
• Brinchmann et al. 2007, 351, 1151
• Smith et al. 2006, astro-ph/0610562
• Di Matteo et al. 2008, astro-ph/08092592

7
Starburst? SFR/ltSFRgt 2 At present, all these
clouds, including those near the galactic
nuclei, show no signs of intense starburst
activity. (Schultz et al. , AA 2007, 466,
467) The bulk of the molecular gas is forming
into stars with a normal star formation
efficiency LIR/M(H2)4.2 Lsolar/Msolar, the
same as that of giant molecular clouds in the
Galactic disk. Additional supportive evidence is
the extremely low fraction of the dense
molecular gas in Arp 244, revealed by our
detections of the HCN(1-0) emission, which
traces the active star-forming gas at density
gt104 cm-3. (Gao et al. 2001, ApJ 548, 172)
8
Arp 220, only ULIRG within 100 Mpc
9
But
HST G. Östlin
The BCG ESO 338-IG04
10
Main objectives

• Identify starburst galaxies, derive luminosity
  functions, lifetimes and masses
• Examine the link between starburst and
  poststarburst galaxies
• Derive properties (LF, lifetimes, masses) of
  starbursts producing poststarbursts
• Study the starburst - AGN connection
• Investigate what importance starbursts have on
  the transition across the green valley. What is
  the shutdown mechanism for star formation?

11
Main objectives

• Identify starburst galaxies, derive luminosity
  functions, lifetimes and masses
• Examine the link between starburst and
  poststarburst galaxies
• Derive properties (LF, lifetimes, masses) of
  starbursts producing poststarbursts
• (Study the starburst - AGN connection)
• Investigate what importance starbursts have on
  the transition across the green valley. What is
  the shutdown mechanism for star formation?

12
The birthrate parameter
If the galaxy is involved in a starburst bgtgt1 The
soft starburst criterion bgt3

13
Results fom a related study(Brinchmann et al.
2004 MNRAS 351, 1151)

• The majority of the star formation in the
  low-redshift Universe takes place in moderately
  massive galaxies
• (10 10 -10 11 M), typically in high surface
  brightness disc galaxies.
• Roughly 15 of all star formation takes place in
  galaxies that show some sign of an active
  nucleus.
• 20 occur in starburst galaxies if a starburst
  is defined as one in which b gt2-3
• 3 occur in starburst galaxies if a starburst
  is defined as one in which b gt10.

14
The data

• Galaxies were selected from the SDSS DR6.
  Spectro/photometric coverage 20 of the sky.
• From the catalogue we select
• a) Starburst galaxies, minimum criterion
• EW(H?)gt120Å
• b) Poststarburst galaxies, EW(H?abs)gt6Å
• Redshift range
  
• zgt0.005. Limits aperture effects
• zlt0.4. Dwarf galaxies cannot be easily
  studied at higher redshifts. Evolutionary effects
  limited.

15
EW(H?????120Å, 8521 galaxies
(r-i)fiber-(r-i)total
16
The SDSS aperture effect on calculated SFR
0.005ltzlt0.22
Brinchmann et al. 2004, ApJ 351, 1151
17
Separating AGNs from star forming from the BPT
diagram and FWHM of the emission lines
Star forming AGNs
18

• Defining the starburst
• Two parameters
• The birthrate parameter
• The burst mass fraction
• We will look at the sample first by
• applying these parameters separately
• and then together.

19
The birthrate parameter
If the galaxy is involved in a starburst
bgtgt1 Here we apply the soft starburst criterion
b3 Next we need to
determine the present SFR
20
SFR from spectral evolutionary model
Zackrisson et al. xx. L(H?) needs to be corrected
for underlying absorption and galactic and
internal extinction. The extinction corrections
are normally small for dwarf galaxies.
Corrections for dust extinction using comparisons
with FIR observations important at higher
luminosities. For nearly constant SFR
21
ltSFRgt and the mass of the burst and old stellar
components the burst age.
We use our spectral evolutionary model,
(Zackrisson et al. 2001, AA 375, 814), including
a nebular component based on CLOUDY We now
assume that the stellar population is a mixture
of an old population and a burst of constant SFR
and arbitrary age mixed so that the observed
EW(H?) corresponds to the model pred. The
relative amount of the old population and its M/L
gives us the mass of the old pop. and the mean
SFR, i.e. The mean birthrate parameter (b) can be
derived. We also obtain the burst
mass and burst age.
Best fit
22
Two components Burst Old
Single component
23
Additional examples of fits
24
Evolution of a starburst with 10Myr duration
Model from Zackrisson et al. 2001, AA 375, 814
25
450 Myr
H?
H?
H?
H?
26
Example of a previous study of an EA
or Postburst galaxy
27
The elliptical-like merger remnant and postburst
ESO 341-IG04
(Bergvall et al. 1989, AA 222, 49)
28
ESO 341-IG04
Age 1Gyr Mass 1011Msolar Burst mass fraction
10-50 (Bergvall et al. 1989, AA 222, 49)
29
?
In 95 of the cases gt5 of the mass was
involved in burst more than 100 Myr ago
Monte Carlo simulations
(Kauffmann et al. 2003, MNRAS 341, 33)
30
?
In 95 of the cases gt5 of the mass was
involved in burst more than 100 Myr ago
Monte Carlo simulations
(Kauffmann et al. 2003, MNRAS 341, 33)
31
Holst 2007 from Bruzual Charlot (2003, MNRAS 34
4, 1000)
32
Holst 2007, masters thesis
33
Method Establish a link between starburst and
postburst galaxies
Luminosity
Lburst Lpost- burst
This can be modelled
?t(burst)
?L
?t(postburst)
Time
34
Simpified equation
L(postburst)/L(burst) and ?t derived from
SEMs constrained by eq. widths of H?em and H?abs
35
Output

• Mass of burst and host
• Age of burst, aB
• Luminosity function
• Knowing the fading(aB), we can predict LF of
  generated postbursts and compare to the observed
  - i.e. we can check that the durations we have
  assumed (2aB) are reasonable.

36
Model predictions
Mburst5
5 burst in star forming galaxy
Log b
Log b
Log EW(H?)
b 3
5 burst in red and dead galaxy
Log age (yr)
Log age (yr)
37
Model predictions
Mburst5
5 burst in star forming galaxy
120 Å
Log b
Log b
Log EW(H?)
b 3
5 burst in red and dead galaxy
Log age (yr)
Log age (yr)
38
Luminosity
Lburst Lpost- burst
?t(burst)
?L
?t(postburst)
Time
39
__ ---
Starburst in star forming galaxy with constant
SFR Starburst in red and dead galaxy At every
instant the contribution from the starburst is 5
of the total mass (i.e. it is NOT an evolutionary
sequence). Sizes of symbols correspond to the
relative burst mass fraction
EW(H?)
EW(H?)
Burst duration
Burst duration
40
Mr
41
SDSS LF Blanton et al. 2005, ApJ 631,208
42
b SFR/ltSFRgt gt 3
Log age of burst (yr)
Log mass
43
ltbgt gt 3
Log age of burst (yr)
Log mass
44
2
bgt3, mass fraction gt 2
Log age of burst (yr)
Log mass
45
b SFR/ltSFRgt
Log mass
46
ltbgt lt SFR/ltSFRgt gt
Log mass
47
Gas consumption timescale
Timescale (yr)
Mass
48
Age of burst (yr)
49
Log age of burst (yr)

Log mass
50
Predicted LF for poststarbursts with gt2 of mass
in burst
Too many postbursts!
mfgt2
51
Dust enshrouded AGN/starbursts feeding the
postburst population
Mr
52
Conclusions

• Starbursts with bgt3 are extremely rare events (lt
  a few per mille among star forming dwarfs)
• Starbursts with mean b-parameter gt 3 have
  lifetimes lt 107 years
• Starburst with present b-parameter gt 3 have mean
  lifetimes of a few 108 years. The b-parameter
  increases with mass and the lifetimes gets
  shorter
• Starbursts consuming more than 2 of the mass,
  i.e. the progenitors of poststarbursts, have mean
  lifetimes 1 Gyr
• The poststarburst progenitors are generally not
  bursting
• The mean b-parameter is lt1 and independent on
  mass
• Thus
• self-regulation is strong and weakly dependent on
  mass
• the galaxies formed more mass in the past than
  now

53
Conclusions

• Luminous poststarbursts can only be explained if
  a significant fraction of luminous AGNs host
  starbursts.
• These results favour shutdown of starburst by AGN
  activity