Transcript Are Gamma-Ray Bursts good Star Formation Indicators?

Are Gamma-Ray Bursts good Star Formation
Indicators?
Nial Tanvir
University of Hertfordshire
Abstract
Since long-duration gamma-ray bursts (GRBs) are associated with the deaths of massive stars, the
are also indicators of star formation activity. Advantages they have over conventional methods are
that they are extremely bright, and hence visible to high redshifts, and detectable in gamma-rays
through very high columns of gas and dust. If they are unbiased tracers of star formation activity,
then the GRB host galaxies should contain all populations of star forming galaxies, in proportion to
their contribution to the global star formation rate. In particular, if a high proportion of star
formation occurs in obscured mode, then a similar proportion of GRBs and their hosts should also
show signs of dusty environments. We describe progress towards testing this prediction.
Introduction
Effect of metallicity?
Explaining the discrepancy
This conference attests to the importance placed on
studies of star formation in our understanding of the
evolution of baryonic matter in the universe.
Although significant progress has been made in
recent years, there remains debate about the fraction
of star-formation occurring in obscured mode, which
is hard to study directly, and uncertainty about the
star formation occurring at very high redshifts (z>5).
In the latter case, because galaxies (and quasars) are
faint and few at high-z, traditional methods become
ever more difficult to apply.
There is already some evidence that GRBs show some
preference for low metallicity environments. For
example, Ramirez-Ruiz et al. (2002 ApJ 565 L9)
found that GRBs located in the outer parts of their
host galaxies, presumably with lower metallicity, are
somewhat brighter then those closer to the centres.
Fynbo et al. (2003 A&A 406 L63) pointed out that Lyα is seen strongly in emission (where it falls in the
optical window) in five out of six GRB hosts. This is
a substantially greater than the proportion of Lybreak galaxies showing such strong emission lines,
and is again suggestive of a low metallicity, low dust
environment.
The question therefore arises as to whether GRBs
avoid ULIRG-like hosts, or whether the amount of
star formation in such galaxies has been substantially
overestimated. In fact, the discrepancy may be worse
since the 3 hosts which are confidently (>3σ) detected
at 850μm are not representative of typical submm
selected galaxies: they are faint and blue, with little
evidence of optical extinction, and lie at z<1.5. It is
possible that current samples are deficient in GRBs
from very dusty hosts, simply because their optical
light is attenuated so that no afterglow is seen.
However, Barnard et al. (2003 MNRAS 338 1)
showed that a sample of 4 hosts of “dark” bursts with
good X-ray positions did not show strongly enhanced
submm emission either. It is also possible that a bias
towards finding lower redshift GRB afterglows
means that we are not yet seeing the peak of the
global star formation. A recent relevant development
has been the discovery of GRB 030115, which shows
evidence of strong reddening of its afterglow light,
and also resides in a ERO-like host galaxy (fig 2).
GRB 030115 may therefore represent an intermediate
case between the generally low extinction afterglows
which dominate the observed samples, and rarer
highly extinguished bursts.
Long-duration GRBs are known to accompany the
deaths of massive stars (eg. Hjorth et al. 2003 Nature
423 847), and since such stars have very short
lifetimes, are also indicators of recent star formation.
We may therefore hope that simply counting the
GRB rate as a function of redshift will tell us the star
formation history of the universe.
As star formation indicators GRBs have a number of
potentially important advantages over other
approaches:
1. They should be detectable in the nIR to very high
redshift, z~20, if they existed then (eg. Lamb and
Reichart 2000 ApJ 536 1).
2. The gamma-ray flash from GRBs should itself be
detectable through significant columns of gas and
dust, and also destroy dust for tens of parsecs (eg.
Waxman and Draine 2000 ApJ 537 796).
3. There is no need to detect individual host
galaxies, in any passband, since redshifts can be
obtained from the afterglows.
Conversely, potential drawbacks with GRBs are:
1.
2.
3.
4.
5.
There could be a metallicity or other
environmental dependence of GRB rate or
luminosity which is hard to account for. Indeed
there is some evidence of such a metallicity
dependence (see below).
Some GRBs may be so heavily enshrouded in
dust that we only detect their gamma-rays,
making redshift estimates more uncertain.
GRBs are short lived, so rapid, pre-planned
observations are required if we are to obtain
redshifts for statistical samples.
GRBs are rare, building up sufficient numbers is
a slow process, consuming considerable telescope
time.
GRBs can only be used as global indicators and
aren’t useful for studying the star formation rate
in individual galaxies.
Of these disadvantages, only the first may turn out to
be a show stopper. To investigate possible external
influences on GRB rate/luminosity, we can either
look directly at the properties of GRBs and their
galactic environments, or assume that no such
influences exist, and compare the star formation
inferred from GRBs with other measures. Both these
approaches are discussed below.
Testing the method
Current samples of GRBs are too small and
incomplete, especially in terms of redshifts, to make
firm statements about global star formation.
However, if GRBs are unbiased tracers of star
formation then samples of their hosts should contain
populations of galaxies in proportion to the
contribution of each population to the overall star
formation rate (eg. Trentham et al. 2002 MNRAS).
Generally the GRB hosts are small, blue (in
optical/nIR colours) galaxies in the redshift range
0.5<z<2.5 (eg. Le Floc’h et al. 2003 A&A 400 499).
Most would be too faint to be included in samples of
Ly-break galaxies, or, for that matter, in samples of
submm selected galaxies as we show below.
To attempt to quantify the relationship between GRB
rate and star-formation rate more precisely we have
observed a sample of GRB hosts using SCUBA on the
JCMT. Blind submm surveys and the integrated IR
background suggest that a majority of the star
formation in the early universe took place in dust
enshrouded systems. In such galaxies the UV flux
from OB stars is largely reprocessed by the dust to
far IR wavelengths. Thus submm measurements
provide an independent estimate of star formation
rate in the host galaxies.
Combining our observations of GRB hosts with those
of Berger et al. (2003 ApJ 588 99) we construct a
sample of 21 galaxies. Although comparatively small,
this is sufficient to test the hypothesis that GRB rate
and star formation rate are directly proportional.
The results are illustrated in figure 1 (from Tanvir et
al. 2004 MNRAS 352 1073).
Fig 1. The histogram shows
the distribution of submm
luminosities of GRB hosts.
Of 21 in the sample, only 3
were significantly detected
(middle bin).
The points
(with error bars from
counting statistics) show the
predicted distribution based
on the models of Blain et al.
(1999 MNRAS 302 632).
The figure shows that to-date the sample of GRB
hosts is deficient in very bright submm galaxies.
Fig 2. GRB 030115 shows the
most reddened afterglow found
to-date, and resides in an EROlike host.
This does indicated
that at least some GRBs are
found
in
highly
dusty
environments, albeit that a
majority are not (Levan et al. in
prep.)
Conclusions and Future Work
GRBs are potentially powerful tracers of the global
star formation history of the universe back to very
early times. However, with evidence that GRBs may
favour low metallicity environments, and disfavour
intensively star-forming ULIRGs, their use in practice
may be hard. A number of new developments promise
to speed up these investigations in the near future:
• The SWIFT satellite will be launched late 2004 and
will provide many more rapid GRB alerts. With
good localisations, and rapid optical monitoring
(on-board and ground-based), the proportion of
bursts which remain optically dark is likely to fall
considerably.
Importantly the Swift X-ray
telescope (XRT) is expected to given positions good
to a few arcsecs, so GRBs which are optically dark
but visible in X-rays can be followed up with
SCUBA to search for dusty hosts.
• The UK now has access to a broader range of
powerful telescopes for GRB followup, including
the 2m robotic Liverpool Telescope and the Faulkes
Telescopes. These are crucial for obtaining large,
and statistically well defined samples of redshifts
(time already awarded to teams led by O’Brien et
al., Mundell et al. and Tanvir et al.).