Evidence for overdensity around z_em > 4 quasars from the proximity effect

Article (PDF Available) · March 2007with28 Reads
Source: arXiv
Abstract
We study the density field around z_em > 4 quasars using high quality medium spectral resolution ESI-Keck spectra (R~4300, SNR > 25) of 45 high-redshift quasars selected from a total of 95 spectra. This large sample considerably increases the statistics compared to previous studies. The redshift evolution of the mean photo-ionization rate and the median optical depth of the intergalactic medium (IGM) are derived statistically from the observed transmitted flux and the pixel optical depth probability distribution function respectively. This is used to study the so-called proximity effect, that is, the observed decrease of the median optical depth of the IGM in the vicinity of the quasar caused by enhanced photo-ionization rate due to photons emitted by the quasar. We show that the proximity effect is correlated with the luminosity of the quasars, as expected. By comparing the observed decrease of the median optical depth with the theoretical expectation we find that the optical depth does not decrease as rapidly as expected when approaching the quasar if the gas in its vicinity is part of the standard IGM. We interpret this effect as revealing gaseous overdensities on scales as large as ~15 Mpc/h. The mean overdensity is of the order of two and five within, respectively, 10 and 3 Mpc/h. If true, this would indicate that high redshift quasars are located in the center of overdense regions that could evolve with time into massive clusters of galaxies. The overdensity is correlated with luminosity: brighter quasars show higher overdensities.

Figures

arXiv:astro-ph/0702369v1 14 Feb 2007
Mon. Not. R. Astron. Soc. 000, 000–000 (0000) Printed 5 February 2008 (MN L
A
T
E
X style file v2.2)
Evidence for overdensity around z
em
> 4 quasars from the proximity
effect
R. Guimar
˜
aes
1,2
, P. Petitjean
2,3
, E. Rollinde
2
, R. R. de Carvalho
4
, S. G. Djorgovski
5
,
R. Srianand
6
, A. Aghaee
2,7
, and S. Castro
5,8
1
Observatorio Nacional - MCT, R. Gal. Jos´e Cristino, 77, 20921-400, Rio de Janeiro, RJ - Brasil
2
Institut d’Astrophysique de Paris & Universit´e Pierre et Marie Curie, 98 bis boulevard d’Arago, 75014 Paris, France
3
LERMA, Observatoire de Paris, 61 avenue de l’Observatoire, 75014, Paris,France
4
Instituto Nacional de Pesquisas Espaciais - INPE, Av. dos Astronautas, 1758, 12227-010, S. J. dos Campos, SP - Brasil
5
Palomar Observatory, California Institute of Technology, 105-24, Pasadena, CA 91125, USA
6
Inter University Center for Astronomy and Astrophysics, IUCAA, Post bag 4, Ganeshkhind, Pune 411 007, India
7
Department of Physics, University of Sistan and Baluchestan, 98135 Zahedan, Iran
8
European Southern Observatory, Karl-Schwarzschild Strasse, 2, Garching, Germany
Typeset 5 February 2008; Received / Accepted
ABSTRACT
We study the density field around z
em
> 4 quasars using high quality medium spectral reso-
lution ESI-Keck spectra (R 4300, SNR > 25) of 45 high-redshift quasars selected from a
total of 95 spectra. This large sample considerably increases the statistics compared to previ-
ous studies. The redshift evolution of the mean photo-ionization rate and the median optical
depth of the intergalactic medium (IGM) are derived statistically from the observed transmit-
ted flux and the pixel optical depth probability distribution function respectively. This is used
to study the so-called proximity effect, that is, the observed decrease of the median optical
depth of the IGM in the vicinity of the quasar caused by enhanced photo-ionization rate due
to photons emitted by the quasar. We show that the proximity effect is correlated with the
luminosity of the quasars, as expected. By comparing the observed decrease of the median
optical depth with the theoretical expectation we find that the optical depth does not decrease
as rapidly as expected when approaching the quasar if the gas in its vicinity is part of the
standard IGM. We interpret this effect as revealing gaseous overdensities on scales as large
as 15h
1
Mpc. The mean overdensity is of the order of two and five within, respectively,
10 and 3h
1
Mpc. If true, this would indicate that high redshift quasars are located in the
center of overdense regions that could evolve with time into massive clusters of galaxies. The
overdensity is correlated with luminosity: brighter quasars show higher overdensities.
Key words: Methods: data analysis - statistical - Galaxies: clustering - intergalactic medium
- quasars: absorption lines - Cosmology: dark matter
1 INTRODUCTION
The Inter-Galactic Medium (IGM) has been intensively studied
using the absorption seen in the spectra of quasi-stellar objects
(QSOs) over a large redshift range (0.16 z
em
6.3). This ab-
sorption, first identified by Lynds (1971), breaks up at high spectral
resolution in hundreds of discrete absorption lines from, predom-
inantly, HI Lyman UV resonance lines redshifted in a expanding
universe (the so-called Lyα forest, see Rauch 1998 for a review).
The Lyα forest was interpreted by Sargent et al. (1980) as the
signature of intervening HI clouds of cosmological nature embed-
ded in a diffuse hot medium. The clouds were further described
by Rees (1986) as gravitationally confined by dark-matter mini-
Based on observations carried out at the Keck Telescope
halos. The advent of numerical simulations has introduced a new
and more general scheme in which the IGM is a crucial element of
large scale structures and galaxy formation. It is now believed that
the space distribution of the gas traces the potential wells of the
dark matter. In addition, most of the baryons are in the IGM at high
redshift, making the IGM the reservoir of gas for galaxy formation.
The numerical N-body simulations have been successful at repro-
ducing the observed characteristics of the Lyα forest (e.g., Cen et
al. 1994; Petitjean et al. 1995; Hernquist et al. 1996; Theuns et al.
1998). The IGM is therefore seen as a smooth pervasive medium
which can be used to study the spatial distribution of the mass on
scales larger than the Jeans’ length. This idea is reinforced by ob-
servations of multiple lines of sight (e.g., Coppolani et al. 2006).
It is well known that the characteristics of the Lyα forest
change in the vicinity of the quasar due to the additional ionizing
2 Guimar
˜
aes et al.
flux produced by the quasar. The mean neutral hydrogen fraction
decreases when approaching the quasar. Because the amount of ab-
sorption in the IGM is, in general, increasing with redshift, this re-
versal in the cosmological trend for redshifts close to the emission
redshift of the quasar is called the ’inverse’ or ’proximity’ effect
(Carswell et al. 1982; Murdoch et al. 1986). It is possible to use
this effect, together with a knowledge of the quasar luminosity and
its position, to derive the mean flux of the UV background if one
assumes that the redshift evolution of the density field can be ex-
trapolated from far away to close to the quasar. Indeed, the strength
of the effect depends on the ratio of the ionization rates from the
quasar and the UV background, and because the quasar’s ionization
rate can be determined directly through the knowledge of its lumi-
nosity and distance, the ionization rate in the IGM can be inferred.
This method was pioneered by Bajtlik, Duncan & Ostriker (1988)
but more recent data have yielded a wide variety of estimates (Lu,
Wolfe & Turnshek 1991; Kulkarni & Fall 1993; Bechtold 1994;
Cristiani et al. 1995; Fernandez-Soto et al. 1995; Giallongo et al.
1996; Srianand & Khare 1996; Cooke, Espey & Carswell 1997;
Scott et al. 2000, 2002; Liske & Williger 2001). Scott et al. (2000)
collected estimates from the literature which vary over almost an
order of magnitude at z = 3. The large scatter in the results can be
explained by errors in the continuum placement, cosmic variance,
redshift determination, etc.
In the standard analysis of the proximity effect it is assumed
that the underlying matter distribution is not altered by the pres-
ence of the quasar. The only difference between the gas close to the
quasar or far away from it is the increased photoionization rate in
the vicinity of the QSO. If true, the strength of the proximity effect
should correlate with the quasar luminosity but such a correlation
has not been convincingly established (see Lu et al. 1991; Bech-
told 1994; Srianand & Khare 1996; see however Liske & Williger
2001). It is in fact likely that the quasars are located inside over-
dense regions. Indeed, the presence of Lyα absorption lines with
z
abs
> z
em
suggests an excess of material around QSOs (Loeb &
Eisenstein 1995; Srianand & Khare 1996). Furthermore, in hierar-
chical models of galaxy formation, the supermassive black holes
that are thought to power quasars are located in massive haloes
(Magorrian et al. 1998; Ferrarese 2002), that are strongly biased
to high-density regions. Possible evidence for overdensities around
quasars come also from studies of the transverse proximity effect
by Croft (2004), Schirber, Miralda-Escud´e & McDonald (2004)
and Worseck & Wisotzki (2006) who suggest that the observed
absorption is larger than that predicted by models assuming stan-
dard proximity effect and isotropic quasar emission. However, in
the case of transverse observations, it could be that the quasar light
is strongly beamed in our direction or, alternatively, that the quasar
is highly variable. Interestingly, neither of these affects the longitu-
dinal proximity effect discussed in the present paper.
Observations of the IGM transmission close to Lyman break
galaxies (LBGs) seem to show that, close to the galaxy, the IGM
contains more neutral hydrogen than on average (Adelberger et al.
2003). As the UV photons from the LBGs cannot alter the ion-
ization state of the gas at large distances, it is most likely that the
excess absorption is caused by the enhanced IGM density around
LBGs. It is worth noting however that various hydrodynamical sim-
ulations have trouble reproducing this so-called galaxy proximity
effect (e.g., Bruscoli et al. 2003; Kollmeier et al. 2003; Maselli et
al. 2004; Desjacques et al. 2004).
In a recent paper, Rollinde et al. (2005) presented a new anal-
ysis to infer the density structure around quasars. The method is
based on the determination of the cumulative probability distribu-
tion function (CPDF) of pixel optical depth, and so avoids the Voigt
profile tting and line counting which is traditionally used (e.g.,
Cowie & Songaila 1998; Ellison et al. 2000; Aguirre et al. 2002;
Schaye et al 2003; Aracil et al. 2004; Pieri et al. 2006). The evolu-
tion in redshift of the optical depth CPDF far away from the quasar
is directly derived from the data. This redshift dependent CPDF is
then compared to the CPDF observed close to the quasar to derive
the mean density profile around quasars. The method was applied
to twenty lines of sight toward quasars at z
em
2 observed with
UVES/VLT and it was found that overdensities of the order of a
few are needed for the observations to be consistent with the value
of the UV background flux derived from the mean Lyα opacity. In
the present paper we apply the same method to a large sample of
95 quasars at z
em
> 4 observed with the Echelle Spectrograph and
Imager (ESI) mounted on the Keck II telescope.
In Section 2 we describe the data and the selection of the sam-
ple used in the present work. We derive the redshift evolution of
the ionizing UV background and of the median IGM optical depth
in Sections 3 and 4 respectively. We discuss the proximity effect
in Section 5 and conclude in Section 6. We assume throughout this
paper a flat Universe with
m
= 0.3,
Λ
= 0.7,
b
= 0.04 and
H
o
= 70 kms
1
.
2 DATA AND SAMPLE SELECTION
Medium resolution (R 4300) spectra of all z > 3 quasars
discovered in the course of the DPOSS survey (Digital Palomar
Observatory Sky Survey; see, e.g., Kennefick et al. 1995, Djor-
govski et al. 1999 and the complete listing of QSOs available at
http://www.astro.caltech.edu/george/z4.qsos) have been obtained
with the Echellette Spectrograph and Imager (ESI, Sheinis et al
2002) mounted on the KECK II 10 m telescope. Signal-to-noise ra-
tio is usually larger than 15 per 10 km s
1
pixel . These data have
already been used to construct a sample of Damped Lyα systems
at high redshift (Prochaska et al. 2003a,b). In total, 95 quasars have
been observed.
The KECK/ESI spectra were reduced (bias subtraction, at-
fielding, spectrum extraction) using standard procedures of the
IRAF package. The different orders of the spectra were combined
using the scombine task. In the Echellette mode spectra are di-
vided in ten orders covering the wavelength range: 4000
˚
A
λ
obs
10000
˚
A. In this work we use only orders 3 (center 4650
˚
A)
to 7 (center 6750
˚
A). The spectral resolution is R 4300 or
70 km s
1
. During the process of combining the orders, we
controlled carefully the signal-to-noise ratio obtained in each or-
der. Wavelengths and redshifts were computed in the heliocentric
restframe and the spectra were flux calibrated and then normalized.
The signal-to-noise ratio per pixel was obtained in the regions
of the Ly-α forest that are free of absorption and the mean SNR
value, averaged between the Ly-α and Ly-β emission lines, was
computed. We used only spectra withmean SNR 25. We rejected
the broad absorption line QSOs (BAL) and QSOs with more than
one damped Lyα system (DLA) redshifted between the Ly-α and
the Ly-β emission lines because their presence may over-pollute
the Lyα forest. Metal absorptions are not subtracted from the spec-
tra. We can estimate that the number of intervening CIV and MgII
systems with W
obs
> 0.25
˚
A is of the order of five along each
of the lines of sight (see e.g., Boksenberg et al. 2003; Aracil et al.
2004; Tytler et al. 2004; Scannapieco et al.