Presentation is loading. Please wait.

Presentation is loading. Please wait.

Fitting the full SED (from UV to far-IR) of galaxies to put constraints on dust attenuation and star formation determinations, from z=0 to z=2 The CIGALE.

Similar presentations


Presentation on theme: "Fitting the full SED (from UV to far-IR) of galaxies to put constraints on dust attenuation and star formation determinations, from z=0 to z=2 The CIGALE."— Presentation transcript:

1 Fitting the full SED (from UV to far-IR) of galaxies to put constraints on dust attenuation and star formation determinations, from z=0 to z=2 The CIGALE code and its application to Herschel surveys SED2011-Preston- sept-2011 Véronique Buat, Elodie Giovannoli, Médéric Boquien, Sébastien Heinis, Denis Burgarella & the Herschel Reference Survey and GOODS-Herschel teams

2 Outline Description of the CIGALE code Applications: -z=0 [Herschel Reference Survey]: integrated and spatially resolved SEDs -1<z<2: [GOODS-Herschel]: evidence for a UV bump in the attenuation curves of IR detected galaxies

3 CIGALE : Code Investigating GALaxy Emission P.I. D. Burgarella (Noll et al. 2009) http://www.oamp.fr/cigale/ http://www.oamp.fr/cigale/ INPUT PARAMETERS: Photometric data +errors Star Formation Histories Dust attenuation curves IR libraries OUTPUT PARAMETERS : All based on a Bayesian analysis input parameters Stellar Mass Dust luminosity Amount of obscuration D4000 break, slope of the UV continuum …. A physically-motivated code: CIGALE combines a UV-optical SED & a dust IR emitting component: Energetic balance fully conserved between stellar and dust emission.

4 STELLAR COMPONENT « Old » stellar population: SFR exponentially decreasing, age, e-folding rates are free parameters « Young » stellar population: SFR exponentially decreasing or constant, age, e-folding rates are free parameters Composite population: mass fraction due to the young stellar population as free parameter Populations synthesis models from: Maraston 2005 PEGASE (Fioc & Rocca-Volmerange)

5 CIGALE: optimized to study dust attenuation curves different amount of dust attenuation (and same attenuation curve) for the young and the old stellar population 5500 Å slope Calzetti et al. (2000) UV bump power law Amplitude Central wavelength Width

6 Various mid and far-IR libraries of star forming and starburst galaxies Dale & Helou (2002) 64 templates Chary & Elbaz (2001) 105 templates Siebenmorgen & Krügel (2007) ~7000 SEDs -Library of modified Black bodies Now : β=1.5, Tdust varies several β and Tdust: collaboration with the Herschel Reference Survey Team, nearby universe -Draine & Li (2007) library, work in progress

7 AGN libraries Giovannoli, 2011, PhD thesis To go deeper in the AGN analysis 32 AGNs from Fritz et al. 2006 Current Baseline: 9 SED templates of AGNs 2 AGNs Siebenmorgen & Krugel (04) Mean AGN from Mullaney et al. (10) 6 AGNs from Fritz et al. (06)

8 We build mock catalogues to check the robustness of parameter estimates: Run CIGALE a first time and take the best model for each galaxy Create an artificial dataset with the best models integrated in the photometric bands and add noise Run again CIGALE on the mock catalogue and compare « true » and estimated parameters (Giovannoli et al. 2011, Buat et al. 2011)

9 SED fitting in the Herschel Reference Survey HRS (PI S. Eales): Volume limited, K-selected survey of 323 galaxies observed at 250, 350, 500 µm 15 55°, A B < 0.2 K < 8.7 for E, S0, S0a and K < 12 for Sa-Im-BCD HRS (PI S. Eales): Volume limited, K-selected survey of 323 galaxies observed at 250, 350, 500 µm 15 55°, A B < 0.2 K < 8.7 for E, S0, S0a and K < 12 for Sa-Im-BCD 9 all Early- types Late-types Boselli et al. 2010

10 Full integrated IR SEDs (cf L. Ciesla poster) Draine & Li 2007 M99 Draine & Li 2007 M99 Dale & Helou 2002 Mean SEDs Dale & Helou 2002 Mean SEDs Chary & Elbaz 2001 Mean SEDs Chary & Elbaz 2001 Mean SEDs HRS SEDs by morphological type Individual SEDs: Dale & Helou (02) give good fits, Draine & Li (07) better reproduce the data Mean SEDs per morphological type: Dale & Helou (02), Chary & Elbaz (01), Elbaz et al. (11) well reproduce mean SEDs of late types. Individual SEDs: Dale & Helou (02) give good fits, Draine & Li (07) better reproduce the data Mean SEDs per morphological type: Dale & Helou (02), Chary & Elbaz (01), Elbaz et al. (11) well reproduce mean SEDs of late types.

11 All the data are convolved and regridded to the same reference frame (at 350 microns) and a pixel-to-pixel analysis of the emission is performed by fitting the SED of each pixel with CIGALE FUV350 µm HRS102-M99 FUV NUV u g r i 70 µm250 µm 350 µm z H Spatially resolved SEDs on large galaxies (M. Boquien et al. In prep.) Subset of ~30 late-type galaxies and pixel to pixel analysis at the resolution of the 350 microns frame

12 A(FUV) L IR

13 Dust attenuation curve in external galaxies: Evidence for a UV bump in galaxies detected in IR at 1<z<2 Buat et al. 2011, A&A, in press Calzetti et al. 94, 00: from spectroscopic data no bump in local starburt galaxies, moderate rise in UV Wild et al. In local galaxies: GALEX UV bands and u(SDSS) well suited to test the presence of the bump FUV NUV u Calzetti 01 Wild et al. 11

14 Similar conclusion by Wild et al 11 but see also Wijesinghe et al. 2010, Johnson et al. 2007 At low z: some evidence for a bump in local star forming galaxies from broad band analyses Burgarella et al. 05- SED-fitting IRAS-GALEX data Conroy et al. 2010, broad band colours and model (SDSS) MW LMC NUV-u FUV-NUV 0.4 0.6 0.8 1.0 1.2 1.4

15 Ilbert et al. 2009- Cosmos field-30 photometric bands from UV to NIR: « A broad absorption excess at 2175 A seems necessary to explain the UV flux of some starburst galaxies » At high z: Composite spectrum 1.5<z<2.5: evidence for a bump for 30% of the sources Noll et al. 2009 But no evidence for a bump in Lyman Break Galaxies at z≈2 (Vijh et al. 03, Reddy et al. 08)

16 We work at high z to redshift the UV range in the visible We combine optical (UV rest-frame) and far-IR data: strong constraint on SFR and dust attenuation We use intermediate band filters in optical to tightly sample the UV rest frame Working in the CDFS combining Herschel/PACS (GOODS-Herschel project, PI: D Elbaz) and Subaru/MUSYC broad and intermediate band filters (Cardamone et al. 2010) +IRAC & MIPS data (Dickinson et al. 2003) 30 sources (28 with spec-z) in the GOODS-S field with 1 5 in optical, NIR, mid-IR and at 100 µm, SNR > 3 at 160 µm: GOODS-H- CDFS field at 24- 100-160 µm Subaru Intermediate band filters Selection of the sources with 1<z<2 to sample the region of the UV bump

17 Evidence for a UV bump on the observed SEDs Intermediate band photometry very powerful for SED analyses Very good fits for the whole sample: a bump is always found.

18 General shape of the average attenuation curve consistent with that of Charlot & Fall and (marginally) with that of Calzetti et al. on top of which there is a bump at 2175 A whose amplitude is 35% (76%) that of the MW (LMC2) one. Width of the feature: 356 Å (437 Å for the MW) Charlot & Fall 00, cst SFR over3 10 8 yrs Calzetti et al. 00 BUMP

19 Conclusions We have developped a physically-motivated code of SED fitting, well suited for the analysis of dust attenuation and star formation The code has been applied both at low and high redshift, in the framework of Herschel surveys HRS: analysis of both integrated and spatially resolved SEDs GOODS-Herschel: dust attenuation curve derived for galaxies at 1<z<2 and with a UV bump whose amplitude is 35% that of the MW PEP/HerMES/COSMOS : work in progress


Download ppt "Fitting the full SED (from UV to far-IR) of galaxies to put constraints on dust attenuation and star formation determinations, from z=0 to z=2 The CIGALE."

Similar presentations


Ads by Google