Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.stecf.org/conferences/sardinia/node36.html
Дата изменения: Mon Jun 12 19:04:27 2006 Дата индексирования: Mon Oct 1 21:04:09 2012 Кодировка: Поисковые слова: md-11 |
Stephen J. Warren Imperial College, Department of Physics, Blackett Laboratory, Prince Consort Road, London SW7 2BZ, UK
D. L. Clements Institut d'Astrophysique Spatiale, Bâtiment 121, Université Paris XI, F-91405 Orsay CEDEX, France
Gerard M. Williger NASA Goddard Space Flight Center, Code 681, Greenbelt, Maryland 20771, USA
Paul C. Hewett Institute of Astronomy, Madingley Road, Cambridge CB3 0HA, UK
(1)Present address: NICMOS Postdoctoral Researcher,
Department of Astronomy, University of California at Berkeley, 601
Campbell Hall, Berkeley CA 94720, USA
email:
bunker@bigz.Berkeley.EDU
galaxy formation, quasar absorption lines, damped Ly
systems, star formation rates, infrared spectroscopy
The history of star formation in the Universe is a topic of enormous
current interest (e.g., Madau et al. 1996). The damped Ly
absorption systems (DLAs,
Wolfe et al. 1986) contain most of the neutral gas in the
Universe, and from their redshift distribution, and the
measured column densities, the evolution in the co-moving density of
neutral gas
can be measured (e.g., Lanzetta et
al. 1991). Then the analysis of the variation of
with redshift
allows the measurement of the history of star formation in the Universe
(Pei & Fall 1995) provided the consequences of dust obscuration are accounted
for.
This approach to the history of star formation unfortunately tells us nothing about how galaxies are assembled. One school of thought has DLAs being the (large) progenitors of massive spiral disks (e.g., Lanzetta et al. 1991; Prochaska & Wolfe 1997). However, Møller & Warren (1998) have recently shown that the impact parameters of the few detected galaxy counterparts of high-redshift DLAs are small (in the context of this debate) and that the space density of the DLAs at high-redshift is probably much higher than the space density of spiral galaxies today.
Here we present the results of a spectroscopic survey for
H656.3nm emission from 8 damped absorption
systems at 2.0<z<2.6, along the line-of-sight to 6 high-redshift
quasars. At these redshifts H
appears in the near-infrared
K-window. The results are relevant to the debate on the nature of the
DLAs, for if the DLAs are the
counterparts of today's spiral galaxies the associated H
emission should be detectable. The
rate of depletion of the cosmic density of neutral gas can be used
to compute a
universal star formation rate. The average star formation rate in each
DLA depends then on their space density, so that high measured rates of
star formation would provide support for the view that the DLAs are
massive galaxies already in place at high-z. A low measured star
formation rate on the other hand would be in agreement with the
hierarchical picture.
There have been extensive searches for Ly
121.6nm
emission from DLAs but with limited success (e.g., Smith et al. 1989; Hunstead, Pettini & Fletcher 1990;
Lowenthal et al. 1995). This is generally thought to be due to
the fact that resonant scattering greatly extends the path length of
Ly
photons escaping through a cloud of neutral hydrogen so that
even very small quantities of dust can extinguish the line (Charlot &
Fall 1991). Because the effective extinction can be extremely large this
has the consequence that non-detections do not provide any useful
information on the star formation rates in the DLAs. The H
line,
although intrinsically weaker by a factor
,
lies at a longer
wavelength where the extinction is smaller, and is not resonantly
scattered. In consequence a search for H
emission from DLAs may
be more efficient than a search for Ly
.
![]() |
With the CGS4 spectrograph on the 3.8-m UK Infrared Telescope (UKIRT)
we have undertaken a search for line emission from 8 high-redshift DLAs
(2.0<z<2.6) near the expected wavelength of H.
The
observations and data reduction are detailed in Bunker (1996). No
lines were detected at
significance above the quasar
continuum. Our long-slit K-band spectra were typically 1-hour each
and used the 2.5-arcsec wide slit (
kpc at
,
q0=0.5 &
throughout). With a 3-arcsec extraction width, the
upper-limits lie in the range
Wm-2 for spectrally-unresolved line
emission. The resolution of CGS4 in this configuration is
FWHM.
We use the upper limits on H
line
luminosities to constrain the star formation rates in these systems,
based on the
prescription of Kennicutt (1983), where a star formation rate (SFR) of
generates a line luminosity in H
of
W. The limits to the SFRs lie in the range
,
although the conversion between
H
line luminosity and SFR is somewhat uncertain and depends on the
assumed IMF.
We compare the measured upper limits to the star
formation rates in our sample against predictions based on the
assumption that DLAs are the progenitors of today's spiral galaxies. We
begin with the analysis by Pei & Fall (1995) of the observed rate of
decline of the cosmic density of neutral gas
measured
from surveys for DLAs. At any redshift
will be an
underestimate of the true cosmic density
because
quasars lying behind DLAs will suffer extinction, and may therefore drop
out of the samples of bright quasars used to find the DLAs. Pei & Fall
correct for this bias, accounting in a self-consistent manner for the
increasing obscuration as the gas is consumed and polluted as star
formation progresses. In this way they determine the evolution of
,
and so the SFR per unit volume.
Although the analysis of Pei & Fall provides the SFR per unit volume
at any redshift, it gives no information on the SFR in individual
galaxies. For the hypothesis of large disks of constant co-moving space
density, assuming that the SFR in a DLA is proportional to the
present-day galaxy luminosity L(0), we can predict the SFRs in the
population of DLAs at any redshift. Figure 1 plots the results of this
calculation, showing the predicted survey-averaged star formation rate
for a sample of DLAs for
.
Seven of the eight DLAs lie
below the curve in this plot. The significance of this result is
reduced by the fact that the solid angle over which we have searched
for H
emission is smaller than the expected solid angle of the
large disks. Despite this we would still have expected to detect 2 or 3
systems with average SFRs twice as large as our upper limits. These
results then provide support for the hierarchical picture. For
deeper limits are required to distinguish between the
two pictures. A more detailed treatment of this survey and its
implications is given in Bunker et al. (1998).
A decisive test can be made with the latest generation of near-IR instrumentation. Deep H-band imaging with HST/NICMOS (Warren et al., GO-7824) will reveal whether the galaxies responsible for damped absorption are indeed in sub-L* pieces, and the opportunities afforded by the largest ground-based telescopes such as the VLT should enable the accurate measurement of star formation rates in these systems. Combined with spectroscopy of metal lines, in this way we will build up a picture of the history of assembly, gas depletion, and chemical evolution in the population of damped absorbers.
We would like to thank the Organizing Committee of the
``NICMOS & the VLT'' Sardinia meeting. We are grateful for the
excellent support we received while observing at UKIRT. We thank Mike
Fall, Palle Møller & Hy Spinrad for useful discussions, and Max
Pettini & Richard McMahon for details of some of the damped systems
included in our survey. AJB was supported by a PPARC studentship while
at Oxford, and a NICMOS-IDT postdoctoral position at U.C. Berkeley.