COSMOS-CLASH:The existence and universality of the FMR at z<2.5 and

environmental effects

& Simon Lilly (ETH Zürich),

& Bodo Ziegler (University of Vienna)

& (z)COSMOS Team & VLT-CLASH Team

Christian Maier

Z(M* , SFR)

or FMR ?

Empirical dependence: Z(Mstellar, SFR) for SDSS galaxies at 0.07<z<0.3:

at a given mass, galaxies with higher SFRs have lower metallicities

Mannucci et al. (2010)

Question: Z(M,SFR) or FMR?

Is Z(M,SFR) universal (FMR)?

Is Z(M,SFR) redshift and environment independent or not?

Z: metallicity

M: stellar mass of a galaxy

SFR: star formation rate

FMR: fundamental metallicity relation

Outline

- The evolution of the mass-metallicity relation (MZR) at z<2.5 based on near-infrared (NIR) spectroscopy with SINFONI-VLT and MOIRCS-Subaru of zCOSMOS-deep galaxies at 2<z<2.5

- zCOSMOS z>2 MZR vs. Erb et al. (2006) z>2 MZR

- The extrapolation of the local Z(M,SFR) at high SFRs, and the existence or not of the FMR

- Environmental effects on Z(M,SFR): first results from VLT-spectroscopy of CLASH clusters at z~0.4

- galaxies define a tight surface in 3D space fitted by Eq. 2 of Mannucci et al. (2010), with dispersion of single galaxies around this surface of ~0.05dex

Mannucci et al. (2010)

- also Erb et al. (2006) averaged metallicities for galaxies in six mass bins at z~2.3 seem to be on the same surface defined by the low-redshift data

Empiricist: Z(Mstellar, SFR)

Eq. 2

Rationalist: Z(Mstellar, SFR) & FMR

Lilly et al. (2013)

- A physically motivated Z(Mstellar, SFR) was derived by Lilly et al. (2013) for a

galaxy that evolves in a quasi-equilibrum state; they argued that the gas regulation would work at least since z~2- Z(M,SFR) only evolves to the extent that ε and λ change (at fixed M) with epoch: if ε and λ do not change, Z(M,SFR) does not –> FMR

Z0: metallicity of the infalling gas

Zeq: equilibrum value for metallicity

y: yield: mass of metals returned to ISM per unit mass locked up in long lived starsε=SFR/Mgas: star formation efficiency

λ: mass-loading factor (mass loss is λxSFR) of any wind that drives gas out of the system

- Lilly et al. (2013) could represent the local SDSS data from Mannucci et al. with their predicted Z(M,SFR) relation with astrophysically plausible values and mass dependences of ε and λ

Mannucci et al. (2010)

Rationalist: Z(Mstellar, SFR)

Mass loading factor of order unity, and gas depletion timescale 1/ε=Mgas/SFR~2Gyrs at

M~1010Msun

The form of Z(M,SFR): powerful diagnostic of the regulation of star formation in typical MS galaxies

SUBARU-MOIRCSVLT-SINFONI (non-AO)

Aim: measure the 5 lines [OII] (J-band), Hβ and [OIII] (H-band), Hα and [NII] (K-band) in 2.1<z<2.5 zCOSMOS galaxies

● to measure reliable metallicities, which also allows:● to measure SFRs from extinction corrected Hα● to identify Type-2 AGNs using the BPT diagram● to study the Z(M,SFR) and FMR

Near-infrared spectroscopy with SINFONI and MOIRCS of zCOSMOS galaxies at 2.1<z<2.5

Mass-metallicity relation (MZR) at z>2

Maier et al. (2014), submitted to ApJ

The MZR of z~2.3 galaxies (except Erb et al. 2006) is lower by a factor of 3 to 5 (0.5 - 0.7 dex) than the SDSS relation, while the [NII]/Hα-based O/Hs from stacked spectra of the work by Erb et al. are lower by a factor of 2

Mass-metallicity relation (MZR) at z>2

Maier et al. (2014), submitted to ApJ

Possible reasons for this discrepancy:

i) selection of the samples

ii) type-2 AGN contamination

iii) different metallicity calibration used

i) Sample selection at 2<z<2.5

Maier et al. (2014), submitted to ApJ

i) Sample selection at 2<z<2.5

Erb et al. (2006) galaxies have a similar distribution in the SSFR-mass plane like zCOSMOS galaxies, but with slightly higher SFR threshold than zCOSMOS

The zCOSMOS galaxies with NIR spectroscopy (magenta symbols) are representative of the parent sample and of MS galaxies

Dashed almost hori-zontal lines: Eq.1 of Peng et al. (2010): the SSFR-mass relation at z=2 and z=2.5, with dispersion taken into account by dotted lines: main sequence (MS)

Metals in Tuscany, 19 June 2012 Christian Maier, Vienna University, Institute of Astrophysics

ii) Metallicity calibration: Caveats of the N2=log([NII]/Hα) method

Horizontal blue lines: measurements of averaged [NII]/Hα ratios by Erb et al. correspond to a range of O/H extending over 0.7-1 dex (because ionization parameter is not known)

Fig.3 from Erb et al. (2006): Pettini & Pagel (2004) N2 derived O/H values for ave-raged [NII]/Hα ratios in 6 stellar mass bins

lo

g([

NII]

/[H

α])

- only 4 of the 87 galaxies at z>2 of Erb et al. (2006) have [OIII] and Hβ line fluxes measured addi-tional to [NII] and Hα

- no information on AGN contamination is available for 83 galaxies in the Erb et al. sample

log([NII]/[Hα])

lo

g([

OIII

]/[H

β])

iii) Type-2 AGNs contamination in Erb et al. sample

Kewley, Maier et al. (2013): new redshift dependent classification scheme of the upper curve of the star-forming sequence

- Blue lines: the demarcation curves of the star-forming sequence seem to evolve at higher z, because of more ex-treme ISM conditions at higher z

- zCOSMOS galaxies with NIR spectroscopy and literature data lie on the evolved star-forming sequence (between blue lines) and are not dominated by AGN

Maier et al. (2014), submitted to ApJ

z

z

BPT diagram for zCOSMOS galaxies at z>2

- The N2 calibration used by Erb et al. is affected by the ionization parameter dependence (see also Newman et al. 2014)

- If any Type-2 AGNs were among the spectra stacked by Erb et al., this would have systematically increased their measured [NII]/Hα ratios and hence their derived O/Hs

Summary: The MZR at z~2.3

Maier et al. (2014), submitted to ApJ

The MZR of z~2.3 galaxies (except Erb et al. 2006) is lower by a factor of 3 to 5 (0.5 - 0.7 dex) than the SDSS relation, while the [NII]/Hα-based O/Hs from stacked spectra of the work by Erb et al. are lower by a factor of 2

The fundamental metallicity relation (FMR)

Is there a dependence

of the mass-metallicity relation (MZR)

on star formation rate (SFR) at z>2

& is this Z(Mstellar, SFR) epoch invariant?

FMR extrapolation at high SFRs: Predictions

Maier et al. (2014), submitted to ApJ

Grid Z(M,SFR) calculated with Eq. 40 in Lilly et al. (2013) with Z0/y=0,

Z0/y=0.1, and Eq. 2 and Eq. 5 in Mannucci et al. (2010); color-coded

according to metallicities

FMR extrapolation at high SFRs: Predictions

The extrapolation to higher SFRs than SDSS (>10Msun/yr)

is different (grid points colors at the magenta cross are green, blue, cyan, indicating different expected metallicities: 8.6<O/H<8.9, 8.2<O/H<8.6, 7.8<O/H<8.2)

FMR extrapolation at high SFRs: Observations

Maier et al. (2014), submitted to ApJ

Galaxies at z>2 with NIR spectroscopy are color-coded according to their metallicities: are these in agreement with the grid point colors? Does the FMR exist?

FMR extrapolation at high SFRs: Observations

zCOSMOS galaxies (filled circles) and Cullen et al. (2014) sample (filled squares) are in agreement with the non-evolving Z(M,SFR) prediction of Lilly et al. (2013)

- Z(M,SFR) at z>2 exists, i.e., lower metallicities for higher SSFRs

FMR extrapolation at high SFRs: Observations

- Erb et al. (2006) galaxies (triangles) are in agreement with the FMR prediction of Eq. 5 of Mannucci et al. (2010), panel d

- The FMR based on SDSS and Erb et al. (2006) data only exists when using the extrapolation from Eq. 5 of Mannucci et al. (2010)

Work in progress:

Environmental effects on Z(M,SFR): first results from VLT-spectroscopy of CLASH clusters at z~0.4

CLASH and VLT-CLASH

- CLASH (Cluster Lensing and Supernova survey with Hubble, Postman et al. 2012): HST imaging in 16 filters of 25 massive galaxy clusters

- VLT-Clash (Rosati et al.): ongoing spectroscopic follow-up using 225 hours of VLT-VIMOS (LR-Blue and MR grisms) spectroscopy of 13 clusters from CLASH

Cluster MACSJ1206 at z~0.45

Metallicities based on 5 lines in CLASH clusters

- VLT-VIMOS MR spectra of ~100 CLASH MACSJ1206 & MACSJ0416 cluster galaxies at z~0.45 and z~0.39

- metallicities based on 5-emission lines ([OII], Hβ, [OIII], Hα, [NII]): all 5 emission lines measured at the same time with VIMOS

cluster galaxy in MAC1206 at z=0.4458

VIMOS MR spectrum

BPT diagram of cluster galaxies at z~0.4

The studied field and cluster galaxies at z~0.4 are not dominated by type-2 AGNs

Maier, Kuchner, Ziegler, Verdugo, Rosati et al., in prep.

Environment effects on MZR (work in progress)

- the MZR of z~0.4 cluster galaxies compared to the MZR of z~0.4 field galaxies

- many cluster galaxies are offset by a factor of 2 towards lower O/H, at a given mass, compared to SDSS

- cluster and field galaxies at z~0.4 occupy similar regions of the MZR diagram

Maier, Kuchner, Ziegler, Verdugo, Rosati et al., in prep.

Summary

- The MZR of Erb et al. (2006) at z>2 is affected by:

- metallicity calibrator (N2 vs. R23

, ionization parameter issue)

- (type-2) AGN identification: only 4 galaxies in BPT diagram

- The MZR of zCOSMOS galaxies is lower by a factor 3-5 than the SDSS relation, while the Erb et al. MZR is only two times lower

- the form of the expected FMR depends on the extrapolation used

- Z(M,SFR) exists, but if it evolves or not depends on the sample and prediction used

- work in progress: MZR in z~0.4 CLASH clusters to study environ-mental effects using five emission lines observed at the same time with VIMOS

Additional Slides:

Different [O/H] calibrations

Kewley & Ellison (2008)

Simple model connecting gas, star formation & metals

- galaxy is continuously fed from outside by new gas - gas flows into the halo, some fraction of which also flows into the galaxy system and adds to the gas reservoir

- stars continuously form out of the reservoir at a rate proportional to the mass of gas (SFR=ε x Mgas)

- Z(M,SFR) is a natural outcome of this simple model in which the SFR is regulated by the mass of gas present in a galaxy Lilly et al. (2013)

Simple model connecting gas, star formation & metals

- in a given interval of time, some gas in the reservoir is transformed into stars: a fraction R is returned to the reservoir, along with newly produced metals, and the remaining fraction (1-R) of this steadily builds up a population of long-lived stars - star formation may drive a wind out of the galaxy, either back into the halo, or beyond

- the mass of gas in the reservoir is free to increase or decrease with time, and its change allows the regulation of the SFR of the galaxy

- changes of Mgas must be associated with

a net flow into or out of the reservoir

Lilly et al. (2013)