Dust attenuation, dust emission, and dust temperatures in z>5 galaxies

A view from the FIRE cosmological simulations (arXiv:1902.10152)

Xiangcheng Ma (Berkeley)xchma@berkeley.edu

Dusting the Universe 03/05/2019

Collaborators: Caitlin Casey & the FIRE team

Madau & Dickinson (2014) Ma+2018a, MNRAS, 478, 1694

???

See also Cullen+17, Wilkins+17, Katz+18, Cowley+18, Yung+18

Why should we care?

101010111012 109 LUV [L⊙]

in the high-redshift Universe

• 35 cosmological zoom-in simulations run to z~5- >2000 halos in Mhalo~108–1012M⨀, resolution mb~100−7000M⨀, ~0.1−1pc

• Run with GIZMO + baryonic physics from FIRE-2 (Hopkins et al. 2018)

- SF in molecular, locally self-gravitating, and dense (>103 cm-3) gas- Photo-ionization and photo-heating, radiation pressure (UV+IR), SNe, and

OB/AGB winds (mass, energy/momentum, heavy elements)

Ma+2018a, MNRAS, 478, 1694Ma+2018b, MNRAS, 477, 219

Smith, Ma+2019, MNRAS, 4844, 39Ma+2019, submitted (1902.10152)

3D Monte Carlo dust RT + FIRE simulations

Ma+2019 (arXiv:1902.10152)

Post-processing with SKIRT (Baes+2011; Camps+2015)

• SMC-like grain size distribution from Weingartner & Draine (2001)

• Constant dust-to-metal ratio (default 0.4; Dwek 1998) in <106 K gas

• Excluding CMB heating at the moment (cf. da Cunda+2013)

λ [μm]

F λ

UV

IR

Meurer+1999

Fλ ∼ λβ

IRX ≡ FIR/FUV

👀

Dust screen

Calzetti

SMC

Dust clumps

👀

The IRX–β relation:introduction

Ma+2019 (arXiv:1902.10152)

A tight IRX–β relation regardless of complex dust geometry

Increasing dust-to-metal ratio

FIRE z=0 sample

Ma+2019 (arXiv:1902.10152)

Dust temperature

Warmer dust in z>5 galaxies

Incre

asing

z

Ma+2019 (arXiv:1902.10152)

ALM

A B

and

6as

if a

t z =

6

ALMA—HUDF 1.2 mm deep survey

Aravena+16, Bouwens+16a, Dunlop+17

HST/WFC3: rest-frame UV

ALMA: dust continuum

Bouwens+2016 analysis

• The IRX–β relation: FUV→FIR

• 35 K-MBB SED: FIR→F1.2 mm

ALMA detected much fewer sources than expected at high z

See also e.g.Casey+18a

Why Tdust increases with z?

(A) (B)

(A) LIR ∼ MdustT6dust → Tdust ∼ ( LIR

Mdust )1/6

∼ sSFR1/6

Ma+2019 (arXiv:1902.10152)

(B) ⟨Σ⟩SFR ∼LIR

R2∼ Jν

∫∞

0κνJν dν = ∫

∞

0κνBν(Tdust) dν ∼ T6

dust

Moving into the future:• (Sub)mm surveys must go a few times deeper than they are now.• Wavelength coverage shorter than λpeak is required for LIR and Tdust.• JWST spectroscopy in rest-frame UV-optical ⇒ the extinction curve.Key questions to be answered:• Direct constraints and detections of obscured SF at z>5• Are dust properties in z>5 galaxies the same as z=0 or not?• Enrichment of dust and heavy elements in the early universe

Summary

The Chajnantor Sub/millimeter

Survey Telescope5.9-m diameter, 5−600 µm

• The bright-end UVLFs at z>5 require significant dust attenuation.• There is a tight IRX–β relation determined by the attenuation curve

despite of complex dust geometry.• Dust temperature (peak wavelength) increases (decreases) with z.