The mass metallicity relation in VVDS

1
The mass metallicity relation in VVDS

• Enrique Pérez-Montero
• Laboratoire dAstrophysique de Toulouse-Tarbes,
  Observatoire Midi-Pyrénées
• Fabrice Lamareille
• LATT-OMP
• Thierry Contini
• LATT-OMP
• Jarle Brinchmann
• Leiden University
• the VVDS team

2
Pérez-Montero et al. The mass metallicity
relation in VVDS
Motivation to study the MZ relation

• The study of the relation between stellar mass
  and gas phase metallicity and its evolution
  through cosmological time is crucial to the
  understanding of the formation and evolution of
  galaxies.

The slope of the MZ relation depends mainly on
the effective yield.
Lower effective yields correspond to higher mass
loss or dry accretion.
3
Pérez-Montero et al. The mass metallicity
relation in VVDS
The VIMOS VLT Deep Survey

• It is a photometric and spectroscopic survey of
  galaxies (Le Fevre et al, 2005), selected only
  from their apparent magnitudes.
• The spectroscopy is obtained with the VIMOS
  instrument (Visible Multi-Object Spectrograph)
  mounted on UT-3 of the VLT.
• So far, there are available 20k galaxy spectra
  with a resolution of 230 and a spectral coverage
  between 5500-9500 Å, with a mean redshift of 0.7

4
Pérez-Montero et al. The mass metallicity
relation in VVDS
The emission-line galaxies in VVDS
VVDS is a suitable survey to study the MZ
relation because it offers a statistically
significant sample at different redshifts. We
take first epoch data of the deep fields (17.5
IAB 24 F02 and CDFS) and bright fields (17.5
IAB 22.5 F22, F10 and F14) After removal of
duplicated and broad-line objects and stars, we
keep only objects with secure redshifts.
Measurement of emission line fluxes and
equivalent widhts has been done with
platefit_vimos pipeline
5
Pérez-Montero et al. The mass metallicity
relation in VVDS
Spectral classification (zlt0.2 NII/Ha vs.
SII/Ha)?
We reject emission-line galaxies with non-thermal
source of ionization by means of different
diagnostic diagrams.
6
Pérez-Montero et al. The mass metallicity
relation in VVDS
Spectral classification (0.2ltzlt0.4 OIII/Hß
vs. NII,SII/Ha)?
7
Pérez-Montero et al. The mass metallicity
relation in VVDS
Spectral classification (0.5ltzlt0.9 OII/Hß vs.
OIII/Hß)?
8
Spectral classification (zgt0.9 OII/Hß vs.
NeIII/Hß)?
Pérez-Montero et al. (2007)?
9
Spectral classification (zgt0.9 OII/Hß vs.
NeIII/Hß)?
10
Spectral classification
There is not significant differencein the number
of Seyfert 2 found for each redshift bin, but the
number of candidates (i.e. In the region of
uncertainty) is larger for higher redshift.
72
1479
1103
412
11
Determination of the stellar mass
Stellar masses have been derived comparing the
SED, the HdA and the D4000 with Bruzual Charlot
(2003) models adding secondary bursts to a
declining exponential star formation history.
The stellar mass is defined by the maximum of
probability of the PDF of the chi-squares of the
set of models.
Similar values of the stellar mass are found when
using BC03 and CB07 models.
12
Calculation of the limiting mass
The curve of completude is calculated using the
following expression
Limiting masses increase with redshift and they
are lower for the deep sample and for star
forming regions.
13
Determination of metallicity
Metallicities have been derived from empirical
calibrators based on strong emission lines. We
have used two methods 1) The Bayesian method as
compared with Charlot Longhetti (2001)
models. 2) NII/Ha (zlt0.5) and R23
(OII,OIII,Hß 0.5ltzlt0.9) calibrations. Depend
ing on the used calibration they are differences
up to 0.7 dex in the MZ relation (Kewley
Ellison, 2008), we have thus normalized all the
formulae to CL01.
14
Determination of metallicity
In the 0.9 lt z lt 1.24 bin we have used the O2Ne3
parameter
15
Reddening correction
For R23 we have taken into account it by using
EWs (Kobulnicky Phillips) with factors
depending on D4000 (Liang et al.
2006). Nevertheless, for O2Ne3, as the variations
of continuum are more heterogeneous an average
correction of 0.5 mag gives more reliable oxygen
abundances.
16
Reddening correction
For R23 we have taken into account it by using
EWs (Kobulnicky Phillips) with factors
depending on D4000 (Liang et al.
2006). Nevertheless, for O2Ne3, as the variations
of continuum are more heterogeneous an average
correction of 0.5 mag gives more reliable oxygen
abundances.
17
Reddening correction
For R23 we have taken into account it by using
EWs (Kobulnicky Phillips) with factors
depending on D4000 (Liang et al.
2006). Nevertheless, for O2Ne3, as the variations
of continuum are more heterogeneous an average
correction of 0.5 mag gives more reliable oxygen
abundances.
18
The MZ relation
The zero-point of the MZ relation has been
calculated in relation to the curve derived for
the SDSS in the Local Universe (Tremonti et al.
2004). At a given stellar mass the mean
metallicity of galaxies is lower at high
redshift. The metallicities are lower for the
wide sample possibly due to a stronger evolution
of the most massive galaxies.
19
The MZ relation
The running medians of the metallicity for each
bin mass show that There is a decrease of the
mean Z for each redshift bin, but not much for
zgt0.5. There is a flattening of the MZ relation
for massive galaxies from zgt0.5 No significant
chemical evolution is seen for high mass galaxies
in relation to z2.
20
Caution with N2
All oxygen abundances derived from the N2
parameter have a great dependence on the N/O
ratio. Therefore, we have to assume a value of
N/O before deriving the metallicity (e.g. UV
selected galaxies have low values of N/O, Contini
et al. 2002)?
Pérez-Montero Contini (in prep.)?
21
Summary and conclusions

• We have studied the evolution of the MZ relation
  up to z1.24 using the star forming galaxies of
  the VVDS sample.
• Although we have used different diagnostic
  methods of spectral classification and empirical
  calibrators of metallicity for different redshift
  bins, the results are consistent if we use the
  methods consistently.
• We see a decrease of the zero-point of the MZ
  relation for higher redshifts up to z0.5 but no
  significant chemical evolution is detected for
  higher z.
• We see as well a flattening of the relation for
  zgt0.5 for the most massive galaxies (hierarchical
  formation of galaxies)?

22
Future steps for the study of the MZ relation

• The MZ relation in zCOSMOS A new spectroscopic
  survey with VIMOS with a larger resolution (R
  600) and morphologies associated with HST/ACS.
• The MASSIV project (Mass Assembly Survey with
  SInfoni in VVDS). A sample of star forming
  galaxies from the VVDS is observed with the
  infrared IFU SINFONI. We will estimate the
  dynamical mass and the metallicity via the N2
  parameter in the J and H bands (z 1.5).