The LABOCA Survey of the ECDFS – submm properties of NIR-selected galaxies Thomas R. Greve (MPIA), Axel Weiss, Attila Kovacs, Fabian Walter, Ian Smail,

The LABOCA Survey of the ECDFS – submm properties of NIR-selected

galaxies Thomas R. Greve (MPIA), Axel Weiss, Attila Kovacs, Fabian Walter, Ian

Smail, Peter Kurczynski & the rest of the LESS Team

APEX

LABOCA

Charlottesville, September 2009

E-CDFS: 30’ x 30’ field Ra 03:30h Dec -27:40 deg

Wavelength coverage:X-ray: Chandra, XMMUV: GALEXOpt: MUSYC, HST GEMS/GOODS, COMBO-17IR: SIMPLE (IRAC), FIDEL (MIPS)Radio: VLA 1.4GHz

30’

LABOCA-ECDFS Survey - The Field

HST Chandra

Milky Way HI

E-CDFS

GOODS-south

LABOCA-ECDFS Survey - Instrument

LABOCA:

Wavelength: 870 µmBandwidth: 60 GHz

Resolution: 19.2’’FoV: 10.4’

Pixels: ~250Sensitivity: 50 mJy √s

Map rms after 280hrs on source: =1.2 mJy/b (beam smoothed)

Same area but ~2 lower noise levelthan largest SCUBA survey (SHADES)

LABOCA-ECDFS Survey - Noise Properties

rms map

ECDFS

Shades

1.1 mJy/b

1.2 mJy/b

Weiss et al. (2010)

LABOCA-ECDFS Survey – S/N map

What’s the nature of the submm faint (<0.5mJy) galaxies?

Reverse strategy: select high-z galaxies populations at other wavelengths (than submm) and search for a statistical signal

Infer average IR-luminosities, SFRs, EBL-contribution and SF-laws as a function of z and stellar mass

The zoo of near-IR selected galaxies

Why the near-IR?

• near-IR selection is relatively immune to dust and age effects on the stellar population (unlike UV/optical)

• near-IR selection is essentially a stellar mass selection

• the near-IR is able to probe the low mass-end of galaxies compared to submm and radio surveys

• recent advent of large-format near-IR cameras

Near-IR surveys have revealed populations of star forming galaxies and very old, red galaxies at 1 < z < 3, e.g. ERO, DRG, BzK Galaxies

The Multi-wavelength Yale-Chile Surveyhttp://www.astro.yale.edu/MUSYC/

• UBVRIzJHK imaging covering the entire ECDF-S (30’x30’)

• 5sigma AB limiting magnitudes:

U (26.8)B (27.0)V (26.6)R (26.4)I (24.7)z (24.0)J (23.0)H (21.6)K (22.3)

• Ready-to-go catalogues available on-line

K-selected sample: Kvega ≤ 20

We also extracted BzK, ERO & DRG samples (selected down to Kvega ≤ 20)

BzK Galaxies (star forming and passive) - selection

sBzK(z-K)AB - (B-z)AB ≥ -0.2

pBzK(z-K)AB - (B-z)AB < -0.2

STARS

EROs and DRGs - selection

STARS

ERO(R-K)AB > 3.35

DRG(J-K)AB > 1.32

Final Catalogues

STARS

ERO(R-K)AB > 3.35

DRG(J-K)AB > 1.32

Kvega 20 : 8266

sBzK : 744

pBzK : 149

ERO : 1253

DRG : 737

Photometric Redshifts

Phot-z’z are derived using EAZY code (Brammer, van Dokkum & Coppi 2008)

Applied to the UBVRIzJHK MUSYC photometry

σ(Δz) = 0.037 z < 1.5

σ(Δz) = 0.079 z > 1.5

Stellar massesDerived using Hyper-z on the UBVRIJHK (MUSYC) + IRAC (SIMPLE) photometry

Bruzual & Charlot +03: SPP, Exp (300) and Constant SF histories. Calzetti dust (Av = 0-3)

Use rest-frame H-band luminosity since M/LH is less sensitive to AGN and TP-AGB stars

Submm Stacking of K-selected galaxies • Remove robust submm sources in map

• Go to positions of near-IR selected galaxies in the submm map and read of the signal and noise

• Weighted mean stacked flux and noise

• Deblending

Submm Stacking of K-selected galaxies

Full stacks…

3x

5x

Average 870μm flux vs. redshift for Kvega ≥ 20 selected galaxies

Submm Stacking of K-selected galaxies

• IR luminosities:sBzK: (2-6)x1011L

pBzK: (1-3)x1011L

ERO: (1-4)x1011L

DRG: (1-4)x1011L

•Star formation rates:sBzK: (30-100)Myr-1

pBzK: (10-50)Myr-1

ERO: (20-60)Myr-1

DRG: (20-60)Myr-1

• Background resolved:Kvega < 20: 15%sBzK/ERO/DRG: ~4% (each)

Submm Stacking of K-selected galaxies

As a function of 24μm flux (using the FIDEL MIPS catalogue: > 27μJy 5-σ)

• 24μm detected sources have 5x higher average 870μm flux than 24μm non-detections

• Linear correlation (S24μm < 350μJy):

S870μm = 4.5x10-3 x S24μm

• “Flattening” at S24μm > 350μJy

• 24μm traces starformation, but significant contribution from AGN for very bright 24μm systems (although selection function is non-trivial).

Stacking across the BzK-diagram

Refining the sBzK/pBzK selection criteria?

SSFR vs. z and Mstar – evidence of downsizing?

Total SFR (UV+FIR)

Summary • The largest submm stacking analysis of K-selected galaxies, allowing us to probe 7x deeper than typical blank field submm surveys

• Significant submm signals for sBzK, ERO and DRG galaxies. pBzKs are not detected (<0.3mJy). Typically: LIR = (2-6)x1011L and SFR = (60-100)M yr-1

• We find a strong increase in the submm-flux (star formation rate) of near-IR selected galaxies at z > 1.5. This means the dust-enshrouded star formation rate was higher at earlier cosmic epochs

• We find a linear 870μm-24μm correlation for S24μm < 350μJy, suggesting that 24μm is a good tracer of starformation for such systems. For brighter 24μm systems the relation flattens, consistent with an AGN contribution to the 24μm flux

• Tentative evidence of SSFR-evolution with redshift. Also for downsizing, meaning less massive galaxies have higher SSFRs. This is the first time this kind of work has been done using the submm!

• See also poster by Peter Kurczynski