Dust in the Wind: Intergalactic Dust Extinction and its Implications

Dust in the Wind: Intergalactic Dust Extinction and its Implications

Ying Zu (OSU) August 29, 2010 D.H. Weinberg (OSU), Romeel Dav´ e(Arizona), Mark Fardal(UMass), Neal Katz(UMass), Duˇsan Kereˇs(Harvard), and B. D. Oppenheimer(Leiden)

Dust in the Wind: Intergalactic Dust Extinction and its Implications N

1 / 37

1

Introduction Discovery of Intergalactic Dust Possible Origins of “Intergalactic” Dust Motivation

2

Simulations and Metal Distributions Simulations Galaxy Sample and Metal Distributions

3

Results 3-D correlation functions Dust-to-Metal Ratio and Dust Distribution An Alternative Hybrid Dust Model in No-Wind Simulation

4

Implications of Diffuse Cosmic Dust

5

Conclusions Dust in the Wind: Intergalactic Dust Extinction and its Implications

N

2 / 37

Figure: dust from the Eyjafjallajokull volcano (earth surface)


3 / 37

Figure: noctilucent cloud possibly seeded by space dust (upper atmosphere)


4 / 37

Figure: ghostly spokes in Saturn’s B ring, shadows of charged dust clouds (solar system)


5 / 37

Figure: a dust cloud across a rich field of stars (local ISM)


6 / 37

Figure: extended FIR emission indicates outflowing dust entrained by superwinds in M 82 (galactic halo)


7 / 37

Can we extrapolate the existence of dust to the IGM?


8 / 37

Rudnicki, K. & Wszolek, B. (1992)


9 / 37

Statistical Detection of IG Dust Menard et al. (2010); MSFR

fobs = f0 µe −τλ Magnification by galaxies µ: magnification Extinction and reddening due to dust around galaxies τλ : optical depth at λ


10 / 37

Statistical Detection of IG Dust (MSFR)

MSFR examined the mean colors of photometrically identified quasars as function of angular separation from foregroud galaxies in a 3800 square degree area from SDSS. 85, 000 quasars at z > 1 24 million galaxies at z ∼ 0.3


11 / 37

MSFR: Ωdust ∼ 5 × 10−6


12 / 37

Possible Origins of MSFR Dust

Pre-galactic: Pop III stars [Yoshida et al. (2004)] 7 Inner-galactic: LMC-like dwarfs ? Extra-galactic: Winds [Davies et al. (1998)] ?


13 / 37

“The Problem is of a Quantitative Nature“

Can the simulation with galactic winds explain MSFR observation? What dust-to-metal ratio is required? How is dust distributed in the simulation? Do the models without galactic winds provide a vaible alternative explanation of MSFR observation? What further observations can test the model predictions and provide greater insight into IG dust? What does this imply about the survival of dust in the IGM and high-Z SN comology?


14 / 37

Simulations

Modified Gadget-2: Tree-particle-mesh + Smoothed particle hydrodynamics [Oppenheimer & Dave (2008)] Ωm = 0.25, ΩΛ = 0.75, Ωb = 0.044, h = 0.7, σ8 = 0.8, n = 0.95 (WMAP 5) nDM = 2883 , nSPH = 2883 , L = 50 h−1 Mpc, mSPH = 9 × 107 M Galaxies: SKID groups of stars and cold (T < 3 × 104 K ) dense (ρ/¯ ρbaryon > 1000) gas particles that are associated with a common density maximum, large than 64mSPH


15 / 37

Galaxy Identification: SKID

Figure: Blue particles are dark matter, red are gas, and yellow are baryonic particles where star formation has occurred [Weinberg et al. (2002)]

Dust in the Wind: Intergalactic Dust Extinction and its Implications

N

16 / 37

Wind VS. No-Wind

Wind: “momentum-driven” wind velocity ∝ σ [Murray et al. (2005)] mass-loading factor ∝ σ

−1

No-Wind no momentum injection thermal energy of SNe pressurizes the gas but does not drive outlfows


17 / 37

Wind simulation is empirically successful to match: early IGM enrichment [Oppenheimer & Dave (2008)] galaxy mass-metalicity relation [Finlator & Dave (2008)] OVI absorption at low redshift [Oppenheimer & Dave (2009)] enrichment and entropy levels in galaxy groups [Dave et al. (2008)] sub-L∗ regime of the galaxy baryonic mass function [Oppenheimer et al. (2009)] CAVEAT: we take it only as a representative illustration of how winds could influence IG dust in terms of total amount and distribution of metal/dust


18 / 37

Wind simulation is empirically successful to match: early IGM enrichment [Oppenheimer & Dave (2008)] galaxy mass-metalicity relation [Finlator & Dave (2008)] OVI absorption at low redshift [Oppenheimer & Dave (2009)] enrichment and entropy levels in galaxy groups [Dave et al. (2008)] sub-L∗ regime of the galaxy baryonic mass function [Oppenheimer et al. (2009)] CAVEAT: we take it only as a representative illustration of how winds could influence IG dust in terms of total amount and distribution of metal/dust


18 / 37

Galaxy Sample

< z >∼ 0.36, Leff ∼ 0.45L∗ → ng ∼ 0.01 h3 Mpc−3

⇓ Wind Model: mcut = 5.4 × 1010 M → ng ∼ 0.01 h3 Mpc−3 No-Wind Model: mcut = 5.4 × 1010 M → ng ∼ 0.033 h3 Mpc−3


19 / 37

Dust Tracer: “Free Metal”

We assume that cosmic dust traces gas-phase metallicity that is not associated with any SKID group (free metal), under the assumption that quasars behind galaxies will not make it into the SDSS sample

Model Wind No-Wind

All Z (1010 M ) 288.0 277.9

free Z (1010 M ) 100.7 (35.0%) 8.2 (3%)


20 / 37

LOG(ρ) M⊙ h2 /pc3

Surface density maps of the free metals

5 Mpc/h

5 Mpc/h

-1

-2

-3

-4 1 Mpc/h

1 Mpc/h

Wind Model

No Wind Model

-5


21 / 37

3-D correlation functions (Dark Matter)


22 / 37

3-D correlation functions (Galaxies)


23 / 37

3-D correlation functions (Gas)


24 / 37

3-D correlation functions (Metal)


25 / 37

Match To MSFR Observation (SMC Dust)


26 / 37

Be Quantitative: Exclusion Zone Test


27 / 37

Be Quantitative: P[E < E (g − i)] & E p(log E )/E¯


28 / 37

Dust Confined inside Galaxies


29 / 37

Disk Opacity of Dwarf Galaxies

Figure: AV extinction map of LMC [Dobashi et al. (2008)] Dust in the Wind: Intergalactic Dust Extinction and its Implications N

30 / 37

Hybrid Model in No-Wind Simulation


31 / 37

Problems with Hybrid Model

reddening signal is dominated by high E (g − i) values (> 0.1mag ) requires dwarf galaxies with baryonic mass ∼ few ×1010 M baryonic (MLMC ∼ 3 × 109 M ) galaxy stellar mass function is overestimated in the No-Wind simulation.

It is extremely difficult to explain MSFR observation without outflows


32 / 37

Problems with Hybrid Model

reddening signal is dominated by high E (g − i) values (> 0.1mag ) requires dwarf galaxies with baryonic mass ∼ few ×1010 M baryonic (MLMC ∼ 3 × 109 M ) galaxy stellar mass function is overestimated in the No-Wind simulation.

It is extremely difficult to explain MSFR observation without outflows


32 / 37

comparison of reddening maps three models Hybrid Dust, No Wind Model

Free Dust, No Wind Model

Free Dust, Wind Model

-5 -4 25 Mpc/h

-3 LOG E(g-i) -2 -1


33 / 37

Implication: Dust Survival

Thermal sputtering time scale ∼ 107.5 (nH /10−3 cm−3 )−1 yrs for a = 0.01µm at T = 106 K [Draine & Salpeter (1979)] Wind particles typically remain in the IGM for ∼ 109 years before reaccreting onto galaxies [Oppenheimer et al. (2009)] Thermal sputtering rates decline rapidly towards lower T ( by a factor of 300 at T = 105 K ) Most ejected gas never rises above a few ×104 K in our Wind simulation

There is indeed a narrow window of environment for dust grains to escape and survive in the IGM


34 / 37





34 / 37





34 / 37

Implication: Dust Environment

Most wind particles has T < 105 K in halos with M < 1013 M , but T > 3 × 106 K in M > 1013 M halos The growth rate and the destruction rate would be comparably slow in the low density IGM (∼ n2 )

Galaxy-dust correlation will vary in different environments if the screening effect does exist


35 / 37

Implication: Dust Environment

Most wind particles has T < 105 K in halos with M < 1013 M , but T > 3 × 106 K in M > 1013 M halos The growth rate and the destruction rate would be comparably slow in the low density IGM (∼ n2 )

Galaxy-dust correlation will vary in different environments if the screening effect does exist


35 / 37

Implication: Supernova Cosmology IG dust obscuration is an accumulated effect, thus a monolithic function of SN redshift IG dust at different Z will affect a given observed-frame wavelength in different ways IG dust should have a different extinction curve with the dust in the host galaxies of SN (grayer?) The Wind simulation predicts a A¯V = 0.0048mag and rms sightline-to-sightline variance ∼ 0.0052mag at z = 0.5 (MSFR A¯V = 0.03mag ?)

The IG dust extinction level is small compared to the statistical and systematic errors of current SN surveys, but will be a non-trival source of systematics for next-generatoin surveys Dust in the Wind: Intergalactic Dust Extinction and its Implications N

36 / 37



36 / 37



36 / 37

Conclusions The wind simulation is able to reproduce the MSFR observation provided that about 25% of the metal mass in the IGM is in the form of SMC-like dust. The large scale reddening signal will drop to 75%, 50%, and 30% of the original signal if we exclude sightlines that pass within 50, 100, and 200 h−1 kpc of galaxies. This exclusion zone predictation can be used to constrain IG dust distribution and wind physics. It is very difficult for the simulation without outflows to explain MSFR result. The existence of a large scale diffuse dust component has to be taken into account in future SN cosmology. We anticipate the develepment of IG dust study will take a similar route as the weak-lensing cosmology, thus provides new insights into the origin, evolution, and observational impact of dust in the IGM. Dust in the Wind: Intergalactic Dust Extinction and its Implications N

37 / 37