High-Redshift Quasars
Probing Early Star Formation
Matthias Dietrich
The Ohio State University
Department of Astronomy
The CCAPP Symposium 2009 – October 14
• Introduction
• Observations
• Results and Discussion
- UV Emission Line Diagnostics and Metallicity
- FeII / MgII Ratios at High Redshifts
- Super-Massive Black Holes
• Summary
Collaborator
F. Hamann and C. Warner - University of Florida
J. Shields - Ohio University
A. Constantin – Drexel University
M. Vestergaard – Tufts University
I. Appenzeller and S. Wagner - Landessternwarte Heidelberg
C. Onken – Mt. Stromlo Observatory
J. Kollmeier – Observatories of the Carnegie Institute of Washington
S. Mathur – The Ohio State University
D. Grupe – Penn State University
S. Komossa – Max-Planck-Institute for Extraterrestrial Physics
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Introduction
• Quasars are among the most luminous objects in the
universe and have been detected to z > 6
• Current generally
accepted framework
• super-massive black hole
• accretion disk (+ corona)
• broad emission line region
• absorbing structure (torus)
• narrow emission line region
• jets (radio, optical, X-rays)
• outflows, disk winds
The CCAPP Symposium 2009 – Columbus – October 14, 2009
A variety of AGN models has been suggested, e.g.
(Collin, Dyson, McDowell, & Perry 1988)
(Elitzur,
(Urry
Nenkova,
& Padovani
& Ivezić
1995)
2003)
(Perry & Dyson 1985)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
and a few more examples, e.g.
(Murray, Chiang, Grossman, & Voit 1995)
(Elvis 2000)
(Emmering, Blandford, & Shlosman 1992;
Murray & Chiang 1997; Bottorff et al.2001)
(Elitzur, Nenkova, & Ivezić 2003)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
(Leighly 2004)
• Determination of the gas metallicity of the BELR is
one of the major goals of quasar research
• First studies inferred solar or slightly higher
abundances (Shields 1976; Davidson 1977; Baldwin & Netzer1978;
Osmer 1980; Gaskell 1982)
• Growing evidence for a close connection between
quasar activity and galaxy evolution (e.g., Silk & Rees 1998;
Salucci et al.1999; Kauffmann & Haehnelt 2000; DiMatteo et al.2003, 2007;
Bromm & Loeb 2003; Granato et al.2004; Hopkins et al.2005; Croton et al.2006)
• For some galaxy evolution models AGN feedback
has been introduced (e.g., Hopkins et al.2005; Springel, DiMatteo, &
Hernquist 2005; Cattaneo et al.2005; Merloni & Heinz 2007; Somerville et al.2008)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Quasars and their rich emission line spectra are
powerful tools to explore
(i) the early universe
(ii) first star formation episodes
(iii) the formation of their host galaxies
(iv) super-massive black hole (SMBH) growth
1 Gyr
age [Gyr]
Quasars with redshifts
of z ≳ 5 are of special
interest.
z=5
Ho = 72 km/s Mpc, ΩΛ = 0.7, ΩM = 0.3
redshift z
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Over the years there grew a general consensus for a
close connection of quasar activity and galaxy formation
• Correlation of the black hole mass with the host galaxies
mass, luminosity, and velocity dispersion
• Mbulge  0.0013 Mbh
• Lbulge  Mbh
(Ferrarese
et al.2000)
(Gebhardt et al.2000)
(McLure & Dunlop 2002)
(Ferrarese et al. 2001)
(Tremaine et al.2002)
• σ∗4  Mbh
(Onken et al.2004)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The connection of quasar activity driven by SMBHs
and the formation of the juvenile host galaxy can be
studied by
(i) BLR gas metallicity estimates
(ii) the evolution of the FeII / MgII ratio
(iii) growth time of SMBHs (M● ≃ 109 Mʘ , t ≃ 0.5 Gyr)
If the BLR gas is connected to the ISM of the host
galaxy the enrichment process due to star formation
can be traced back to redshifts z ≳ 6.
The CCAPP Symposium 2009 – Columbus – October 14, 2009
I - UV Emission Lines and Metallicity
• early studies of BLR abundances were based on NIV]1486, OIII]1663,
NIII]1750, and CIII]1909
at least solar up to several times solar
metallicity (Shields 1976;Davidson 1977;Baldwin & Netzer 1978;Osmer 1980;Gaskell 1982;Uomoto1984)
• recent studies on high-z quasars (z > 3) indicate several times Z⊙
(Hamann & Ferland 1992,1993; Ferland et al.1996; Dietrich et al. 1999,2003; Dietrich &
Wilhelm-Erkens 2000; Hamann et al. 2002; Warner et al.2002; Nagao et al. 2006; Dhanda et al. 2007)
• high elemental abundances also have been detected based on intrinsic
absorption lines (e.g., Møller et al.1994; Petitjean et al.1994; Hamann 1997; Pettini 1999;
Kuraszkiewicz et al.2003; Fields et al.2006)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Hamann et al. (2002) studied in great detail the dependence of emission line
ratios from gas metallicity and the shape of the ionizing continuum
• grids of photoionization models, employing CLOUDY (Ferland et al.1998)
for a wide range of gas density, input continuum shapes, and metallicities
• nitrogen line ratios are of particular interest --- assuming that secondary N
production is the dominant source for N
N/O ∝ O/H
N/H
∝ Z2
• this strong scaling of N abundance is observed for HII regions with
elemental abundances above ~1/3 to 1/2 Z⊙ (Pagel & Edmunds 1981; van Zee
et al.1998; Izotov & Thuan 1999; Henry et al.2000; Pettini et al. 2000; Pilyugin et al. 2002)
(Hamann et al.2002)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
We studied a sample of 70 high-z quasars (z > 3.5) to
extend estimates of the gas chemical composition of the
BLR to earlier cosmic epochs (Dietrich et al. 2002, 2003).
The CCAPP Symposium 2009 – Columbus – October 14, 2009
(Dietrich et al. 2003)
mean metallicity of the BLR gas at high redshift
4 to 5 times Z⊙ ( Z/Z⊙ = 5.3 ± 0.3 )
The CCAPP Symposium 2009 – Columbus – October 14, 2009
II – Iron / α-Element Ratio as
Cosmological Clock
• prominent FeII emission blends
are located in the optical (Hβ)
and in the ultraviolet (MgII)
• relative iron abundance may be
useful to date the first star
formation epoch
(Leighly & Moore 2006)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
FeII / MgII Ratio ---- The Idea
• α-elements (e.g., O, Mg, Si) are produced predominantly by
massive stars on short time scales
• the dominant source of Fe is ascribed to intermediate mass
stars in binary systems, ending in SNIa explosions
significant different time scales of the release of α-elements
and iron to the ISM (t ∼ 1 Gyr; e.g. Yoshii et al.1996), however
Matteucci & Recchi (2001) and Friaça & Terlevich (1998)
suggested that this delay may be ~0.2 - 0.4 Gyr for
spheroidal galaxies
Fe/α-element ratios can be used as a cosmological clock
measurement of FeII UV and MgII2798
The CCAPP Symposium 2009 – Columbus – October 14, 2009
necessity of IR spectroscopy of high redshift quasars
J
H
K
z = 3.5
z = 4.0
z = 4.5
z = 5.0
z = 6.28
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The FeII Emission Spectrum
• First detailed models of the FeII emission spectrum were
calculated by Netzer (1980); Netzer & Wills (1983); Wills,
Netzer, & Wills (1985); and Joly (1993)
• Model calculations of FeII emission spectra still face severe
problems (Sigut & Pradhan 2003; Sigut, Pradhan, & Nahar 2004; Verner et
al.1999,2003; Baldwin et al.2004; Leighly & Moore 2006; Bruhweiler & Verner
2008)
•
FeII spectrum sensitive to
- several 100.000 transitions
- density
- radiation transfer
- micro-turbulence
- excitation and ionizing SED
- metallicity
empirical iron emission
templates (Boroson & Green
1992; Vestergaard & Wilkes 2001;
Veron-Cetty, Joly, & Veron 2004)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Multi-component fit to measure the FeII flux
power law continuum
- Balmer continuum emission
- MgII2798, Hβ
- FeII template (observed Vestergaard & Wilkes 2001)
-
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Results on FeII/MgII
FeII/MgII line ratio of quasars at z ≃ 2.0, z ≃ 3.4, and
z ≃ 4.8 in comparison with ratios based on composite
quasar spectra of comparable luminosity
lack of evolution
of the FeII/MgII
Ratio, at least to
a redshift of z ≃ 5
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Comparison with other Studies
Iwamuro et al. 2002,2004; Thompson, Hill, & Elston 1999; Wills, Netzer,
& Wills 1985; Barth et al.2003; Freudling, Corbin, & Korista 2003;
Maiolino et al.2003; Kurk et al.2007; Jiang et al.2007
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The Epoch of Preceding Star Formation
• several
evolutionary models have been suggested to
explain the observed chemical composition
• super-solar
metallicities can be achieved in single
zone models (e.g., Arimoto & Yoshii 1987; Hamann & Ferland 1992,
1993; Matteucci & Padovani 1993; Gnedin & Ostriker 1997),
as well as in more recent multi-zone models
(e.g., Friaça & Terlevich 1998; Romano et al.2002; Granato et al.2004;
DiMatteo et al.2004)
Solar to super-solar metallicity indicates an epoch of
early and intense star formation (~0.5 Gyr) in the
juvenile quasar host galaxies at high redshifts.
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The Epoch of Preceding Star Formation
Evolutionary lack of the FeII/MgII ratio suggests that iron is
already comparable abundant at high redshifts compared to
lower redshifts.
assuming tevol ∼ 0.5 - 1 Gyr
zf ≃ 10
(z ≃ 4.5)
with tevol ∼ 0.3 Gyr, e.g.
motivated be recent
galaxy evolution models
(e.g., Matteucci & Recchi 2001; Romano
et al.2002; DiMatteo et al.2004)
zf ≃ 6 - 10
age [Gyr]
(e.g., Yoshii et al.1996; Wheeler et al.1989)
redshift
(Ho = 72 km/s Mpc, ΩΛ = 0.7, ΩM = 0.3)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
III – Super-Massive Black Holes
• presence of super-massive black holes has been established
(MW, normal galaxies, AGN) with M● ≃ 109 M⊙
• pioneering effort to uncover the radius – luminosity relation
for AGN (Peterson & Wandel 1999,2000; Wandel et al.1999; Kaspi et al.2000;
Peterson et al.2000; Onken & Peterson 2002,2003,2004; Peterson et
al.2004; Bentz et al.2006;2009;Denny et al.2009)
Correction of host galaxy contribution
is crucial for Seyfert galaxies
R ∝ L 1/2
(Bentz et al.2006)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
usual assumption of virial gas motion, i.e. bound motion
Mbh = f • v2 • R
(Mkn110, Kollatschny 2003)
(Peterson & Wandel 2000)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
analysis of the high-redshift quasar sample using the
emission line profiles of Hβ, MgII2798, and CIV1549
to estimate the mass of the SMBHs
we applied the following relations
Hβ
(Vestergaard & Peterson 2006)
MgII2798
CIV1549
(Vestergaard & Osmer 2009)
(Vestergaard & Peterson 2006)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Δ Hβ 4861 based
▲ MgII 2798 based
♦
CIV 1549 based
based on the three emission lines Hβ, MgII, CIV1549
5 · 108 M⊙ ≲ Mbh ≲ 2 · 1010 M⊙
on average: Mbh ≃ 5 · 109 M⊙
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Δ Hβ 4861 based
▲ MgII 2798 based
♦
comparison of the mass estimates:
Mbh(Hβ) / Mbh(CIV) = 1.5 ± 0.3
Mbh(Hβ) / Mbh(MgII) = 2.3 ±0.4
Mbh(CIV) / Mbh(MgII) = 2.1 ± 0.3
CIV 1549 based
(13)
(14)
(18)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The Eddington Ratio
Eddington ratio η calculated based on Hβ and CIV
η = Lbol / Ledd
with Ledd = 1.45 1038 Mbh/Mʘ
Lbol = 4.65 λ Lλ(1450)
Lbol = 9.74 λ Lλ(5100)
η = 0.65±0.37
η = 0.77±0.46
The CCAPP Symposium 2009 – Columbus – October 14, 2009
These Eddington ratios are on average higher than
the mean Eddington ratio which was found by
Kollmeier et al.(2006) based on the AGES Survey.
However, Lbol of the
quasars in this study
are in the range of
1047 to 1048 erg/s
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Implication of Black Hole Growth
• mean Mbh estimates are determined for each quasar
• time necessary to build-up a SMBH can be described
by an e-folding time scale for the growth of a single
seed black hole (e.g., Volonteri & Rees 2006)
(τ = tobs – to : time to accumulate Mbh from the initial time, to, to the time
observed, tobs; Mbhseed seed black hole mass; η = Lbol/Ledd Eddington ratio, ε
efficiency to convert mass to energy, tE = 3.92 ·108 yr Eddington time scale)
The CCAPP Symposium 2009 – Columbus – October 14, 2009
ε = 0.20
0.10 and η = 0.60
1.00
• growth time τ to build-up a SMBH is of the order of
several 100 Myr up to one Gyr
• to avoid too long τ higher Mbhseed are favored
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The simple picture of radiatively efficient accretion
requires far too much time.
Recently, it has been suggested that early mergers and
radiatively inefficient accretion must be important
(e.g., Shankar et al. 2009;Volonteri 2008)
ε = 0.10 and η = obs.
Assembly of SMBH with
several times 109 M⊙ possible
within τ ≃ 2·108 yrs – if
Mbh(seed) ≃ 10% of Mbh(obs)
By grouping the measurements of
FeII/MgII for the high-z quasars
there appears to be a weak tendency
for lower FeII/MgII ratios for
quasars with z ≳ 5.
redshift
Z/Zʘ
F(FeII UV)/F(MgII)
Future Work
F(FeII UV)/F(MgII)
Of course, more measurements are
needed at redshifts z ≃ 6
Kurk et al.2007
Jiang et al.2007
Although gas metallicity is not the
dominant parameter of FeII emission,
it is interesting whether there is a
trend with gas metallicity.
So far no trend at all, but again more
measurements are needed.
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Summary
• UV spectra of high-z quasars analyzed
 4 to 5 times solar gas metallicity
• lack of evolution of the FeII/MgII line ratio up to z ≃ 5
• growth time to assemble SMBH of the order of 0.2 Gyr
These results indicate that at redshifts z ≃ 10 intense star
formation begun to enrich the central region of juvenile
galaxies while SMBH appear to have started to grow
from seed holes with already 10% of the mass in place.
Recent galaxy evolution models suggest quasar activity
in early formation phases, delayed by ~ 0.5 Gyr,
following an intense star formation episode
The CCAPP Symposium 2009 – Columbus – October 14, 2009
Thank You
Observations
The CCAPP Symposium 2009 – Columbus – October 14, 2009
The High Redshift Quasar Sample
Quasar
mR
Q0103-260
18.82
Q0105-2634
17.30
Q0256-0000
17.50
Q0302-0019
17.60
Q2050-359
18.34
PKS2126-158 17.30
Q2227-3928
18.80
Q2348-4025
18.60
BRI0019-1522 19.00
Q0103+0032
18.60
PSSJ0248+1802 18.40
SDSS0338+0021 19.97
PC1158+4635 20.21
SDSS1204-0021 20.82
BRI2237-0607 18.30
BR0305-4957
18.76
z
3.36
3.48
3.37
3.29
3.51
3.27
3.44
3.31
4.52
4.44
4.44
5.02
4.73
5.09
4.57
4.78
Quasar
mR
SDSS0338+0021
BRI1202-0725
SDSS1605-0112
SDSS2200+0017
SDSS0923+4532
SDSS1030+0524
SDSS1048+4638
SDSS1155+0530
SDSS1208+0010
SDSS1223+5037
SDSS1306+0356
SDSS1421+4633
SDSS1630+4012
SDSS1733+5400
SDSS2216+0013
SDSS2346-0016
19.97 5.02
18.70 4.70
22.50 4.92
20.68 4.78
17.50 3.45
23.23 6.28
22.40 6.23
17.80 3.48
22.75 5.27
17.60 3.49
22.58 5.99
16.10 3.36
23.04 6.05
16.80 3.43
21.78 4.99
17.70 3.49
The CCAPP Symposium 2009 – Columbus – October 14, 2009
z
Sites and Time Scales of Elemental Production
• very massive stars --- M > 100 M⊙, t evol < 4 · 106 yr
O and nearly no C, Fe, or r-process elements
• massive stars - SNII, SNIb - 10 < M/M⊙ < 100 ; tevol < 2 · 107 yr
most of the nucleosynthesis may be ascribed to stars with
10 < M/M⊙ < 30 dominant source for O, α-elements (Mg, Si, Ne)
and C, significant contributions to Ne up to Ca
• intermediate mass stars - 1 < M/M⊙ < 10 ; tevol >108 … 109 yr
He burning shells and thermal pulses
during AGB phase, dominant N source
• intermediate mass stars in binary systems - SNIa,
explosive He or C burning; significant fraction of the
56Co
56Fe)
mass is converted to 56Ni (56 Ni
dominant source for Fe and Fe-peak elements
(e.g., Wheeler, Sneden, & Truran 1989; Henry et al. 2000;
Pilyugin et al. 2002)
The CCAPP Symposium 2009 – Columbus – October 14, 2009