the birth of the ﬁrst tiny objects outline · bifﬁ & maio (2013) introduction method...

IntroductionMethod

SimulationsThe End

The birth of the first tinyobjects

Umberto Maio (OATS)Marie Curie Fellow

.

....................................................in collaboration with: M. Petkova, K. Dolag, B. Ciardi,

L. Tornatore, S. Khochfar, J. Johnson, F. Iannuzzi,

M. A. Campisi, R. Salvaterra, N. Yoshida, P. Creasey,

L. Koopmans, V. Müller, V. Biffi, F. Pace, P. Dayal

....................................................

Outline

1 IntroductionMotivationsGeneral overview

2 MethodMolecules and metalsChemistry and cooling

3 SimulationsPopIII and II, SFR, ZLF, GRBs, DLAs, Non-G

4 The End

IntroductionMethod

SimulationsThe End

MotivationsGeneral overview

Motivations

Goal: Primordial structure formation and transition fromthe first metal-free ’PopIII’ star formation regime(high-mass or low-mass stars?) to the common,metal-enriched ’PopII-I’ one (low-mass ’solar’ stars):→ What is the formation epoch of first objects?→ What is the role of molecules and metals in the early ISM?→ How relevant is PopIII for star formation and metal spreading?→ What are the effects of different IMFs on SFR?→ What are the implications for early observables (LF, GRB)?→ What are the effects of the underlying matter distribution?→ What are the effects on cosmic re-ionization?...Requirements: Study the chemical properties of cosmic

medium during cosmological evolution.Techniques: N-body/Sph simulations (with Gadget).

IntroductionMethod

SimulationsThe End


Overview of structure formationThe Universe is supposed

to expand at a rate H0 ≃ 68 km/s/Mpc;

to have flat geometry (zero spatialcurvature);

to consist of dark matter, baryonicmatter and cosmologicalconstant/dark energy.

Cosmological parameters (Planck, 2013):Ω0,DM = 0.26, Ω0,b = 0.04, Ω0,Λ = 0.7⇒ Ω0,tot = 1;σ8 = 0.83, n = 0.96

Standard: H0 = 70 km/s/Mpc, Ω0,Λ = 0.7, Ω0,DM = 0.26,Ω0,b = 0.04, σ8 = 0.9, n = 1.

IntroductionMethod

SimulationsThe End


Theoretical scenario:

Cosmic structures originate from the growth of matterperturbations at early times (inflation), in an expandingUniverse.Baryonic structures form from in-fall and cooling of gas intoDM potential well.Eventually, a cloud can form if the radiative losses aresufficient to make the gas condense and fragment:

tcool =32

nkTL(n, T )

≪ tff =

√

3π32Gρ

At early times, the cooling function is dominated bymolecules ! After pollution from formed (baryonic)structures (→ chemical feedback) metals dominate.

IntroductionMethod

SimulationsThe End


primordial environments...

Small dark-matter haloes

Barkana& Loeb, 2001

H-cooling haloes: Tvir ≥ 104 K H2-cooling haloes: Tvir < 104 K

IntroductionMethod

SimulationsThe End

Molecules and metalsChemistry and cooling

...hosting molecular and metal evolution in their ISM

For a complete picture −→ necessity to follow gravity andhydrodynamics coupled to molecular evolution and metalproduction during cosmic time (e.g. Galli& Palla, 1998; Abel et al., 1997)

molecules determine first structure formation

metals determine subsequent structure formation

stellar evolution determines timescales and yields

Following and implementing metal and molecule evolution innumerical codes (N-body/SPH Gadget) required(Yoshida et al., 2003; Tornatore et al., 2007; Maio et al., 2007, 2010; Biffi & Maio, 2013)

IntroductionMethod

SimulationsThe End


H + e− → H+ + 2e−

H+ + e− → H + γH + γ → H+ + e−

H− + e− → H + 2e−

H− + γ → H + e−

He + e− → He+ + 2e−

He+ + e− → He + γHe+ + e− → He++ + 2e−

He++ + e− → He+ + γHe+ + γ → He++ + e−

He + γ → He+ + e−

H + e− → H− + γH− + H → H2 + e−

H + H+ → H2+ + γH2+ + H → H2 + H+

H2 + H → 3HH2 + H+ → H2+ + HH2 + e− → 2H + e−

H− + H → 2H + e−

H2 + γ → H+2 + e−

H+2 + γ → 2H+ + e−

H2 + γ → 2HH+2 + γ → H + H

+

H− + H+ → 2HH− + H+ → H2+ + e−

H2+ + e− → 2HH2+ + H− → H + H2D + γ → D+ + e−

D+ + e− → D + γD + H2 → HD + HD+ + H2 → HD + H+

HD + H → D + H2HD + H+ → D+ + H2H+ + D → H + D+

H + D+ → H+ + DHe + H+ → HeH+ + γHeH+ + H → He + H+2HeH+ + γ → He + H+

IntroductionMethod

SimulationsThe End


Abundance determination

Given the molecular network, abundance determination for eachspecies, i , is carried out by accounting for creation and destruction ofni via:

dnidt

=∑

p

∑

q

kpq,inpnq −∑

l

klinlni ,

kpq,i = kpq,i (T )[cm3 s−1]: creation rate from species p and q

kli = kli (T )[cm3 s−1]: destruction rate from interactions with the species l

dt can be discretized as a fraction (e.g. ≃ 1/10) of the hydro-timestep

Similar O.D.E. can be written for all the species considered!(i.e., e−, H, H+, He, He+, He++, H2, H+2 , H

−, D, D+, HD, HeH+)

IntroductionMethod

SimulationsThe End


Gas cooling function andthe role of metals −→

In primordial regimes, themain coolants are H, He andmolecules (H2 and HD).

In metal enriched ones,metal fine-structuretransitions from C, O, Fe, Si(dominant over molecules atlow temperatures).

Cooling leads the gas in-fall into DM potential wells.

(Maio et. al, 2007)

IntroductionMethod

SimulationsThe End

PopIII and II, SFR, ZLF, GRBs, DLAs, Non-G

Zcrit : transition from popIII to popII-I star formationWe study the effects connected to the existence of a criticalmetallicity Zcrit (e.g. Bromm & Loeb, 2003; Schneider et al., 2003) and the transitionfrom popIII SF (Z < Zcrit ) to popII-I SF (Z ≥ Zcrit ).

In order to address such issues, we perform several numericalsimulations of early structure formation adopting differentvalues for Zcrit and exploring different scenarios.

Simulation set-up(Maio et al., 2010, 2011b, Maio & Iannuzzi, 2011; Maio, 2011; Maio & Khochfar, 2012)

standard-ΛCDM cosmology (1,7,14,43,143Mpc a side);molecular and metal chemistry;assume Zcrit = (10−6, 10−5, 10−4, 10−3) Z⊙assume different popIII IMFs (→ top-heavy/Salpeter)assume different matter distributions (→ G vs non-G)

IntroductionMethod

SimulationsThe End


Results (1/20): effects for different Zcrit

Zcrit :10−3 Z⊙ 10−6 Z⊙

z=11

z=13

box: 1Mpc3 ; popIII IMF: top-heavy with slope=-1.35, range=[100M⊙ ,500M⊙ ]

Gas resolution: 116 M⊙/h (Maio et al., 2010)

IntroductionMethod

SimulationsThe End


Results (2/20): polluting the surrounding medium

Phase diagrams with color contours for enriched gas(Zcrit = 10

−4 Z⊙ , box side = 1 Mpc)

z=16 z=14 z=11

Metals produced by stellar evolution pollute the surrounding,pristine gas with an “inside-out” mode. (Maio et al, 2011b)

IntroductionMethod

SimulationsThe End


Results (3/20): effects on the surrounding

Radial fractions of PopII and PopIII gas in the most massive halo onscales ∼ 10 − 1000 pc (physical) (Maio et al., 2011b)

IntroductionMethod

SimulationsThe End


Results (4/20): changing the popIII IMF

PopIII range (Salpeter IMF – top-heavy IMF) SN range (Salpeter IMF)

Mass ranges for popIII IMF and/or massive SN have significant impacts:

Larger masses → Shorter stellar lifetimes → Earlier enrichment →Shorter “popIII epoch”

(Maio et al., 2010)

IntroductionMethod

SimulationsThe End


Results (5/20): halo properties

105 106

Mgas [Msunh-1]

10-7

10-6

10-5

10-4

SF

R [M

sun/

yr]

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

102 103 104 105

M* [Msunh-1]

10-7

10-6

10-5

10-4

SF

R [M

sun/

yr]

z=9.00z=9.12z=9.50

z=10.00z=11.00z=12.00z=13.01z=14.00z=15.00z=16.00z=17.02z=18.01z=19.00

105 106

Mgas [Msunh-1]

103

104

105

Tm [K

]

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

102 103 104 105

M* [Msunh-1]

1

10

100

sSF

R [G

yr-1]

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00

z=9.50z=9.12z=9.00

Biffi & Maio (2013)

IntroductionMethod

SimulationsThe End


Results (6/20): halo populations

105 106

Mgas [Msunh-1]

10-9

10-8

10-7

10-6

10-5

10-4

Z

Zcrit=10-4 Zsun

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

8 10 12 14 16 18 20redshift

0.00

0.05

0.10

0.15

0.20

f hal

oes(

Z>

10-4Z

sun)

105 106

Mgas [Msunh-1]

10-5

10-4

10-3

10-2

10-1

Xm

ol

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

8 10 12 14 16 18 20z

0.0

0.2

0.4

0.6

0.8

1.0

SF

R h

alo

cont

ribut

ion

popII-IpopIII

Biffi & Maio (2013)

IntroductionMethod

SimulationsThe End


Results (7/20): halo relations

10-9 10-8 10-7 10-6 10-5 10-4

Z

10-5

10-4

10-3

10-2

10-1

Xm

ol

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

10-7 10-6 10-5 10-4 10-3

M*/MDM

10-5

10-4

10-3

10-2

10-1

x mol

z=19.00z=18.01z=17.02z=16.00z=15.00z=14.00z=13.01z=12.00z=11.00z=10.00z=9.50z=9.12z=9.00

Mannucci et al. (2010)

3.0 3.5 4.0 4.5 5.0 5.5Log(M*[Msun]) - 0.32 Log(SFR[Msun/yr])

-10

-9

-8

-7

-6

-5

-4

Log(

Z)

z=9.00z=9.12z=9.50

z=10.00z=11.00z=12.00z=13.01z=14.00z=15.00z=16.00z=17.02z=18.01z=19.00

Unaccounted outliers...???Hetheroscedasticity...???

←− f⋆ − xmol correlation: f⋆ ∝ xmol

IntroductionMethod

SimulationsThe End


Results (8/20): sSFR – early bursty Universe

0 5 10 15z

0.1

1.0

10.0

100.0

sSF

R [G

yr-1]

Coe (2013)Stark (2013)

Zheng (2012)Gonzalez (2012)Bouwens (2012)

Reddy (2012)Schiminovich (2010)Michalowski (2010)

Yabe (2009)Stark (2009)

Pannella (2009)Dunne (2009)Daddi (2007)

Noeske (2007)

DATA| |

Dave’ vzw (2011)Dave’ sw (2011)Dayal (2013)Salvaterra (2013)Tescari Ch24-sA-sW (2014)Biffi & Maio (2013)

THEORY

Biffi & Maio (2013)

IntroductionMethod

SimulationsThe End


Results (9/20): disc formation and alignment

Haloes hosting rotating gas Alignment gas-DM

8 10 12 14 16 18 20redshift

0.00

0.05

0.10

0.15

0.20

f co-

rota

ting

halo

es

fcosϑ > 0.8 > 20%fcosϑ > 0.8 > 50%

xmol > 5e-4Z > 1e-3 ZsunZ > 1e-4 Zsun

105 106 107

Mgas [Msunh-1]

-1.0

-0.5

0.0

0.5

1.0

cosϑ

gas-

DM

aligned

contra-aligned

All z: corr= -0.0320.00 > z > 16.50, corr= 0.1316.50 > z > 14.50, corr= 0.1214.50 > z > 10.50, corr= -0.0910.50 > z > 8.50, corr= -0.02

cos ϑ =j · jmeanj jmean

cos ϑ =jgas,in · jDMjgas,in jDM

Biffi & Maio (2013)

IntroductionMethod

SimulationsThe End


Results (10/20): UV luminosity functions

For each galaxy: Lλ = LIIλ + LIIIλ

in L5, L10, L30

PopII-I SEDs from Starbust99(Vazquez & Leitherer, 2005).PopIII SEDs from Schaerer (2002).No dust assumed

Observational data points from:

Bouwens et al., 2007 (circles); z=6Bouwens et al., 2011 (circles); z=7-8McLure et al., 2010 (triangles); z=7-8Oesch et al., 2012 (squares); z=8

Fit: Su et al., 2012 (solid line); z=6.

Resulting slope: ∼ −2consistent with HUDF data(Dunlop et al., 2013, Dayal, Dunlop, Maio, Ciardi, 2013)

Salvaterra, Maio, Ciardi, Campisi, 2013

IntroductionMethod

SimulationsThe End


Results (11/20): Implications for high-z GRB hosts

Tracing LGRBs from the SFR of their host galaxies

Differential GRB hosting probability → dP =dNGRB(Log10(SFR[M⊙/yr ]))

NGRB dLog10(SFR[M⊙/yr])

Large objects (high SFR) are rarer than small objects (low SFR):high-z GRBs are more likely found in intermediate-, low-size objects!

IntroductionMethod

SimulationsThe End


Results (12/20): UV luminosities of GRB hosts

Data points from:Tanvir et al., 2012

νion −→

Most high-z GRB hosts are primordial faint galaxies Most ionizing photons are produced in faint galaxies(they lie below current HST, VLT detection limits) (at z = 6 and z = 8)

IntroductionMethod

SimulationsThe End


Results (13/20): Statistical properties of GRB hosts

SFR ∼ 0.01 − 0.1M⊙/yr M⋆ ∼ 107M⊙

sSFR ∼ 5 − 10 Gyr−1 Z ∼ 5 × 10−2Z⊙

Data from: Tanvir et al., 2012 (SFRs), Kawai et al., 2006 (Z)

See Salvaterra, Maio, Ciardi, Campisi (2013)

IntroductionMethod

SimulationsThe End


Results (14/20): Physical relations for GRB hosts

IntroductionMethod

SimulationsThe End


Results (15/20): PopIII-GRB rates and hostsLGRB rate:

different progenitorsi.e. stars with

1: Z > Zcrit→any popII-I

2: Zcrit < Z ≤ 0.5Z⊙→low-Z popII

3: Z ≤ Zcrit→ fGRBup = 0.006→ fGRBup2 = 0.022(upper limits fromSwift sample, 2011)

RGRB =γbζBH fGRB

4π

Z

zρ̇⋆

dz′

(1 + z′)

dV

dz′

Z

Lth(z′)

Ψ(L′)dL′

RGRB : gamma-ray burst rate, γb : beaming factor, ζBH : fractionof expected BH (IMF), fGRB : fraction of expected GRB fromcollapse onto a BH (Swift), ρ̇⋆ : star formation rate density(simulation), Ψ(L): Schechter luminosity fct. (assumption), Lth :instrumental sensitivity (Swift)PopIII IMF: top-heavy over [100, 500]M⊙PopII IMF: Salpeter over [0.1, 100]M⊙Zcrit = 10

−4 Z⊙

Detectable fraction (by BAT/Swift) of PopIII GRBs:∼ 10% at z > 6& 40% at z > 10of the whole population

PopIII-GRB-hosts:the highest probability of finding PopIII GRBs isin hosts with M⋆ < 10

7 M⊙ and Z & Zcrit(efficient pollution)

(from Campisi, Maio, Salvaterra, Ciardi, 2011)

IntroductionMethod

SimulationsThe End


Results (16/20): observed DLAs at z ≃ 7?

10-8 10-7 10-6 10-5 10-4 10-3

Z

108

109M

[MS

un]

NHI,20

IntroductionMethod

SimulationsThe End


Results (17/20): primordial matter distributions andNon-Gaussianities

Basic assumption: Gaussianperturbations → evidences fornon-Gaussianities (CMB).Primordial non-Gaussianities areintroduced via (Salopek & Bond, 1990)

Φ = ΦL + fNL(

Φ2L− < Φ2L >

)

Φ is the Bardeen potential(Newton potential at sub-Hubblescales), ΦL is thelinear (Gaussian)part, and fNL the non-Gaussianparameter.

credit: Planck

fNL = 0, 10, 50, 100, 1000box sides: 0.5 and 100 Mpc/hnumber of particles: 2 × 3203

gas mass resolution: 42 M⊙/hand 3 × 108 M⊙/h

See: Maio & Iannuzzi (2011), Maio (2011), Pace & Maio (2013)

IntroductionMethod

SimulationsThe End


Results (18/20): Non-G and the cosmic web

fNL=0

fNL=1000

IntroductionMethod

SimulationsThe End


Results (19/20): Non-G and the GRB rate

From Swift dataMaio, Salvaterra, Moscardini, Ciardi (2012)

IntroductionMethod

SimulationsThe End


Results (20/20): Non-G and the SZ effect

Power spectrum andbi-spectrum for they parameter:

y =kBσT

mec2

Z

neTe dl

over sets of 100random light-cones

Pace & Maio (2013)

fNL=0 fNL=100 fNL=1000

1e-13

1e-12

1e-11

1000 10000 100000

l(l+

1)P

y(l)

l

fNL=0fNL=100

fNL=1000 1e-36

1e-35

1e-34

1e-33

1e-32

1e-31

1e-30

1e-29

1e-28

1e-27

1e-26

1000 10000 100000

By(

l)

l

fNL=0fNL=100

fNL=1000

IntroductionMethod

SimulationsThe End

Summary...

We have presented results from cosmological N-Body,hydrodynamical, chemistry and radiative simulations

We studied the birth of first galaxies, their interplay with thesurroundings and observational expectations (LF, GRB, DLA).

Conclusions...

Early (z ∼ 15 − 20) metal enrichment from the first stars is verystrong with a rapid popIII/popII-I transition (z ∼ 10). A correlationbetween f⋆ and gas xmol is expected. No correlations betweenearly discs and hosting DM halo are found.

Observationally, LFs agree with data and DLAs/GRBs aresuitable probes of the primordial Universe.

Results are not very sensitive to the assumed Zcrit , popIII metalyields, IMF slope, primordial non-Gaussianities, etc.

IntroductionMethod

SimulationsThe End

The End

Grazie! Thank you!Umberto Maio

[email protected]

Marie Curie FellowEuropean Union FP7/2007-2013,

Actions, Grant Agreement n. 267251,Marie Curie Incoming Mobility Fellowships,

Astronomy Fellowships in Italy (AstroFIt)

IntroductionMethod

SimulationsThe End

IntroductionMotivationsGeneral overview

MethodMolecules and metalsChemistry and cooling

SimulationsPopIII and II, SFR, ZLF, GRBs, DLAs, Non-G

The End

the birth of the ﬁrst tiny objects outline · bifﬁ & maio (2013) introduction method...

Documents