Dust effects in the SEDs of simulated galaxies: the GRASIL-3D code and some applications

A. OBREJA (UAM, Spain)G.L. GRANATO (INAF, Trieste, Italy)I. SANTOS (UAM, Spain)C. A. BROOK (UAM, Spain)G. STINSON (MPIA, Heidelberg, Germany)L. SILVA (INAF, Trieste, Italy)A. SERNA (UMH,Spain)S. GOTTLOBER, Y. HOFFMAN, G. YEPES & CLUES Collaboration

Rosa Domínguez-TenreiroUniversidad Autónoma de Madrid, Spain

Dust Effects in galaxy SEDsDust Effects in galaxy SEDs Galaxy formed in a hydro simulation Galaxy formed in a hydro simulation

Z=6 t/t_U=0.066

DECREASING UV optical // INCREASING IR submm L

DUST reradiates the overall bolometric luminosity SED has been calculated with GRASIL3D code (D-T et al, 2014)

MOTIVATION: Galaxy Formation Simulations: information on 6D phase

space, ages, composition, gas density and temperature.

Observations: light

Testing simulations (theories & models)

against observations demands SOFTWARE TELESCOPES:

Simulation outputs ----> SEDs, images

DUST EFFECTS: FUNDAMENTAL ROLE in GALAXY SEDs

GRASIL-3D

Application I.- Testing G-3 local galaxies

Application II.- IR-submm emission from CLUES Local Group dwarfs

Application III.- The MS and the Fundamental Plane of star-forming galaxies

GRASIL-3D: An extension of GRASIL to arbitrarygeometries and galaxy evolutionary

histories

(D-T+ 14)

RADIATIVE TRANSFER THROUGH DUST

SUNRISE (Jonsson 04,05; +09)

RADISHE (Chakrabarti+08; & Whitney09)

ART2 (Li + 07; Yajima+12)

MCarlo solve RT for Simu outputs

GRASIL assumesEquatorial & axial symmetry for galaxies

SAMS

PEGASE (Fioc & Rocca-Volmerage 1997)

CIGALE (Burgarella + 14) Parameter determination

SKIRT (Baes+ 03, 11)

GRASIL(-3D) PARTICULARITIESSeparately treats RT in MCs and in cirrus,

different dust composition

Age-dependent dust reprocessing of stellar populations: younger *s in denser ISM embeded in MC until destruction

Detailed non-equilibrium calculation for dust grains with diameter a < 250 A

---> Proper PAH treatment

RT solved in a grid, rather than MCarlo

CHARGE VARIABLES to a GRID

MC & CIRRUS DISTRIBUTIONBased on sub-resolution PDF (Wada+07 ; Federrath + 12)

Choose a density threshold for MC rho_thres

Choose a 2 parameter log-normal PDF for the cold gas f_pd(rho_0, sig)

For the ith particle

Identify <rho>_V = rho_gas(i) (**)

→ Calculate the MC mass fraction at ¨i¨

Calculate cirrus mass fraction at ¨i

(**) provides a link between PDF parameters

and simulation outputs ---> PDF sig

AGE DEPENDENT DUST REPROCESSING OF SP

Light fraction that can escape the StarBurst region, mimicking MC destruction,

f(t) = 1 if t<t_0

0 if t>2t_0

2-t/t_0 in other cases

DUST CONTENT k-th GAS PARTICLE d(Z_k) = Z_gas,k /(110 x Z_sun)

PARAMETERS GRASIL 3D

Grid size simulation resolution

rho_mc,thres 10 – 100 H/ cm^3 (Obs + simu)

1 PDF parameter sig 2 – 3 in simulations

---> cirrus & MC mass fraction f_mc

GRASIL

Dust model

Escape-time-scale t_0 2.5-8 or 18-50 Myrs (N,SB)

Individual MC log(m_mc) = 5 -6 M_o; r_mc= 10-50 pc

SIMULATION: SHAPE, EVOLUTION MODEL, SFRH

TESTING GRASIL-3D

LOCAL GALAXIES

I

SIMULATIONSSIMULATIONS P-DEVA P-DEVA AP3M-SPHAP3M-SPH (Martínez-Serrano+ 08)(Martínez-Serrano+ 08) Entropy conserving; mutual neighbors OKEntropy conserving; mutual neighbors OK Chemical evolutionChemical evolution tracks the full dependence of tracks the full dependence of

metal production on the detailed chemical metal production on the detailed chemical composition of * particles (Q_ij formalism, composition of * particles (Q_ij formalism, Talbot+ 73Talbot+ 73))

Feedback: implicitely through (inefficient) SF Feedback: implicitely through (inefficient) SF parameters parameters (Agertz+ 11)(Agertz+ 11)

GASOLINEGASOLINE SPH SPH (Wadsley + 04,Brook + 11)(Wadsley + 04,Brook + 11) Feedback: blastwave formalism (Stinson+ 06) from Feedback: blastwave formalism (Stinson+ 06) from

Sne & Sne & massive starsmassive stars (Stinson+ 13). (Stinson+ 13). Effective coupling Effective coupling with gas 1%with gas 1%

Blast-wave scenario Blast-wave scenario

Face- and edge on images of g1536_L*

Box 50 kpc side, resolution and pixel = 312,5 pc

OBSERVATIONAL SAMPLES GRASIL-3D TESTINGISO Key Project on the ISM of Normal GalaxiesHelou+96; Dale+ 00 ISO & IRAS broad-band fluxes FIR active, quiescent and intermediate (Vega+05)

Aromatic Features in Emission 6.2,7.7,8.6 & 11.3 micron(Lu+03; a subsample of Dale+00)

Spitzer IR Normal Galaxy Survey (SINGS; Kennicutt+03; Dale+05)

Non-tidally perturbed & non-interacting galaxies(Smith+07, a SINGS subsample & Lanz+13 from KINGFISH)

Key Insights on Nearby Galaxies: FIR Survey w Herschel (KINGFISH, Kennicutt+11)All normal types 61 galaxies imaged with PACS and SPIRE (Dale+12)

MOLECULAR and ATOMIC HYDROGEN & STELLAR CONTENT A test for the MC model in GRASIL-3D

Simulations vs COLD GASS survey(Saintonge+11)

OK

IRAS flux density ratios: simulated (color) vs real (Gray: Dale+00) galaxies

FIR active, quiescent and intermediate (Vega+05)Lines join consecutive results for HD-5103B along a merger phase

LEFT: Merger

RIGHT: Normal

AFE Rel Strengths vs FIR/blueAt z=0 and around a MM

Gray points: data Lu+03 FIR-active,intermediate,quiescent

Color: simulations 8 gal. z=0 BLUE And merger phase Different parameter Sets Codes as in previous Fig.left

Observational results recovered Aromatic Features in Emission 6.2,7.7,8.6 & 11.3 micron(Lu+03; a subsample of Dale+00)

Spitzer IRAC & MIPS non-interacting Histograms: Smith+07 sample from SINGS

Points Averages over parameter sets

Spitzer IR Normal Galaxy Survey (SINGS ; Kennicutt+03; Dale+05)Non-tidally perturbed & non-interacting galaxies(Smith+07, a SINGS subsample & Lanz+13 from KINGFISH)

HERSCHEL BANDS non-interacting

Gray: Lanz+13Color: 8 simulated galaxies,

averages & dispersions over parameter sets

Key Insights on Nearby Galaxies: FIR Survey w Herschel (KINGFISH, Kennicutt+11)All normal types 61 galaxies imaged with PACS and SPIRE (Dale+12)

Non-tidally perturbed & non-interacting galaxies(Smith+07, a SINGS subsample & Lanz+13 from KINGFISH)

UV and optical flux density ratiosNon-interacting

Gray: data Dale+07

Color: 8 simulated galaxies,

averages & dispersions over parameter sets

GRASIL-3D DISCUSSION

Changing rho_mc_thres and PDF sig

Parameter variations: not remarkable effects when kept within their ranges

Energy balances: 98% or better

Find encouranging results when comparing with disk galaxies in detail

IR-submm EMISSION from DWARF GALAXIES

The CLUES project LOCAL GROUP

II

(Isabel Santos +, in prep)

THE CLUES PROJECT

CLUES: Constrained I.C. (Hoffman +Ribak 92)

** Observational data imposed as constraintson the IC → local Universe skeleton at Mpc** Random at sub-Mpc

SIMULATION: GASOLINELCDMGrav.softening = 220 pcm_DM, m_*, m_gas 30, 2, 6 x10^4 M_sunFeedback ERIS simulation (Guedes+11)Includes chemical evolution

RESULTS: some relations

Luminosity – colorBLUE, CYAN: SF dwarfsRED: ¨dead¨ dwarfsData from Mateo 1998

Iron vs luminosityBlack Simulated D galaxies Red Data McConnachie+12

Dwarfs are low Z systems

RESULTS: velocity disp. vs M_*

RESULTS: half-light radii vs M_V

RESULTS: SEDs for low M* & Z SF galaxies

Simulations + GRASIL → Two dust components, as observations demand

Galametz+09

RESULTS: Irr low M* SF galaxy

FUV

MIPS160

Ks

SPIRE500SFRH

MASS and SFRH DETERMINATION in SPIRAL GALAXIES

GRASIL-3D as a test bed for

Obreja et al., 14

III

Www.star.uclan.ac.uk/Cbb/magicc MAGICC Project, Brook & Stinson 2012

MOTIVATION

Constraining galaxy formation scenarios

MS: correlation SFR vs M_* (Wuyts + 11)

MZ: M_* vs metallicities

Gas phase (Garnett + 02; Tremonti + 04) Stars (Cowie & Berger 08; Pérez-Montero + 09)

COMPARING SIMU OUTPUTS TO OBSERVATIONAL DATA

OBSERVATIONS: M_* & SFR determined from light (SEDs)SIMULATIONS: need to apply the same recipees !!

NEED SEDS as close to observed ones as possible

SOFTWARE TELESCOPES

through

Projections of a fundamental relation? Ellison +08

Figure 1 from Galaxy Structure and Mode of Star Formation in the SFR-Mass Plane z ~ 2.5 to z ~ 0.1Stijn Wuyts et al. 2011

Constant slopeZero point ---> high-z galaxies form stars faster than local ones with same M_*Scatter independent of z Increases at lower M_*

METHODS SIMULATIONS

Disk gal. from the MaGICC project (Brook + 12; Stinson + 13)

GRASIL-3D post-processing

Face-on SB in r-band ----> Petrosian radii R_p (Blanton + 01)

** Luminosities, fluxes & colors from SEDs within 2R_p

ready to apply observational techniques

including dust effects MIMIC OBSERVATIONS ¨observed¨as they evolve from z=3.5 upto

z=0 (324 snapshots)

METHODS STELLAR MASS DETERMINATIONColor dependent mass-to-light relations B- and

V-band (McGaugh et al. in prep)

GLOBAL SFR IR-corrected far UV tracer of Hao + 11:

FUV flux corrected using the total IR emission WARNINGTest that simulated galaxy

** B-V and M_V are within those of the sample used for calibration

** IRX vs FUV -NUV used for calibration

RESULTS: M_* Testing observational methods

M_* assembly history, scatter is shownBlack: as predicted by simulationsBlue: B-band mass-to-light ratiosGreen: V-band

RESULTS: THE GLOBAL SFRH ¨Observational¨ SFRs vs real ones OK

BLACK: SFRHs predicted by simulationsBLUE: SFRHs from IR-corrected FUV

RESULTS: SCATTER IN THE MS

Independent of redshiftDecreases with increasing M_*

May trust simulations !!---> Reflects bursting/variable SFR in the SF population

RESULTS: THE FUNDAMENTAL METALLICITY RELATIONM_* - SFR – (O/H) in gas (Lara-López + 10)

GREY POINTS: simulated at z<3.5RED POINTS: simulations in equally populated M_* binsBLUE LINES: solid - linear fit with slope 1, as in Lara-López + 10 Dashed – 0.16 dex scatterBLACK LINES: solid & dashed - fit Lara-López data & scatter

** Tight correlation ** Same scatter as obs** Same coefficients** 0.09 dex lower normalization

CONCLUSIONS APPLICATION III

GRASIL-3D METHOD to test outthe observational methods to determine M_* and the SFRH in star forming galaxies

Consistency observations -simulations

The MS and MZ relation for star forming galaxiesAre projections of the Fundamental Plane M_*-SFR-O/H (Mannucci+ 10; Lara-López+ 10)

Prediction: the FP holds also for lower M_* star-forming galaxies

GENERAL CONCLUSIONS

GRASIL-3D particular strengths follow GRASIL (Silva+ 98; 99)

General applicability to systems with arbitrary geometry, in particular those produced in hydrodynmical simulations (evolutionary history, SFRH ..)

Subresolution PDF formalism to describe MC/cirrus density field

Comparison to observations of local galaxy samples are encouraging, particularly remarkable for PAH features

Detailed analysis of parameter space from literature gives consistent results

Many specific applications possible: different predictions on individual & statistical galaxy properties, directly comparable to observations

Applications so far give encouraging results

STAR FORMATION STAR FORMATION

A few parsec resolution Sne II E_II = 10^51 (mass/10 M_sun) erg injected in the ISM after 10 Myear a 8 – 40 M_sun star is formedSne Blast wave in adiabatic expansion if parsec-scale resolution (McKee & Ostriker 1977; Stinson et al. 2006; Brook 2010 . 2011)

SnIa E_I = E_II, but with slower timescale

+ metals

UNRESOLVED ISM (a few 100 pc for discs) mimic through tuning resolved physics: H_2 formation, small-scale turbulence, radiative effects free parameters, based upon core-scale physics : stocastic model(Katz 1992; parameter testing in Agertz et al. 2011)

d rho_g/dt = - e_ff rho_g/t_ff , rho_g>rho_t (K-S-like)

RESULTS: SOME IMAGES

FUV, Ks, SPIRE 500

SFRH

SNE BLAST WAVE MODELSNE BLAST WAVE MODEL** Sne explosions large volumes of hot, low-densitygas McKee & Ostriker analytical blast wave model

** Mimic its effect in SPH codes: Sne feedback from m_* >8M_o(Thacker & Couchman 2000; Stinson et al. 2006; Brook et al. 2011)Energy from massive stars prior to their explosions (Stinson+ 13)weak coupling of stellar energy to the sorrounding gas 1% Flow outwards + SFR regulated GASOLINE RUNS: 3 galaxies m_bar = 2 and 0.25 x 10^5 M_o e_g = 312,5 pc - 156,2 pc M_* = 2.3 - 0.6 x 10^10 M_o Disk & bulge scales, kine, B/D consistent with observations