Primordial Black Holes in 21cm Cosmology 0cm

(1)

21cm Cosmology Primordial Black Holes Summary Tools for the Future

Signatures of Primordial Black Holes in 21cm Cosmology

In collaboration with:

Olga Mena, Sergio Palomares-Ruiz, and Pablo Villanueva-Domingo

Samuel J. Witte

Universidad de Valencia

arXiv: 1906.xxxx (This week!)

(2)

2/47 21cm Cosmology

Primordial Black Holes Summary Tools for the Future

The Basics

Expectations withinΛCDM and Astrophysical Uncertainties

The 21cm Line

Produced via the hyperfine transition of neutral hydrogen

⇠ 21 cm

H hf ∝ S e · S p (1)

Invisibles ’19 (June, 2019) Primordial Black Holes in 21cm Cosmology

(3)

The Basics

The 21cm Line

Produced via the hyperfine transition of neutral hydrogen

⇠21 cm

Intensity

Frequency

⌫⁰₂₁

⌫ = ⌫ 21 0 1 + z

When

State of IGM

(4)

3/47 21cm Cosmology

The Basics

The 21cm Line

Neutral hydrogen at z & 7 backlight by CMB

H H

H

H H

H H H H

H

Absorption/Emission −→ Dictated by n 0 , n 1

(5)

The Basics

Cosmological Timeline

z ⇠ 1000

(Recombination)

z ⇠ 25

z ⇠ 10 z ⇠ 0

(Today) (Reionization) (Star Formation)

(T

k

6 = T

cmb

) Dark Ages Cosmic

Dawn

Reionization Galaxy formation,

He reionization, etc

z ⇠ 200

(6)

4/47 21cm Cosmology

The Basics

Cosmological Timeline

z ⇠ 1000

(Recombination)

z ⇠ 25

z ⇠ 10 z ⇠ 0

(Today) (Reionization) (Star Formation)

(T

k

6= T

cmb

) Dark Ages Cosmic

Dawn

Reionization Galaxy formation,

He reionization, etc

21cm Cosmology

Atmosphere opaque at z > 70, Fluctuations in ionosphere, Interference with Earth radio (Not directly probed)

z ⇠ 200

Dark side of Moon (see e.g. 1412.2096)

(7)

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

)

1 − T

cmb

T

S

1 + z 10

^1/2

mK (1)

(8)

5/47 21cm Cosmology

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

^1/2

mK (1)

Fraction of neutral hydrogen

At late time, reionization suppresses the signal

(9)

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

1/2

mK (1)

Fraction of neutral hydrogen

Baryon overdensity

(10)

5/47 21cm Cosmology

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

1/2

mK (1)

Fraction of neutral hydrogen Baryon overdensity

Spin temperature [Measure of n

1

/n

0

] T

s

(T

cmb

, T

k

, ~x

IGM

)

(11)

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

^1/2

mK (1)

Spin temperature [Measure of n

1

/n

0

] T

s

(T

cmb

, T

k

, ~x

IGM

)

if n

1

/n

0

−→ by spontaneous/stimulated emission, spontaneous absorption

T s ∼ T cmb , δT b ∼ 0 (2)

(12)

5/47 21cm Cosmology

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

1/2

mK (1)

1

/n

0

] T

s

(T

cmb

, T

k

, ~x

IGM

)

if T

k

> T

cmb

and collisional excitations or Lyα scatterings

T s > T cmb , δT b > 0 Signal saturates (2)

(13)

The Basics

The 21cm Signal

IGM

δT

b

' 27 x

HI

(1 + δ

b

) 1 − T

cmb

T

s

! 1 + z 10

1/2

mK (1)

1

/n

0

] T

s

(T

cmb

, T

k

, ~x

IGM

)

if T

k

< T

cmb

and collisional excitations or Lyα scatterings

T s < T cmb , δT b < 0 (2)

(14)

6/47 21cm Cosmology

The Basics

ΛCDM + Astro Expectations

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

T

cmb

⇠ T

k

T

k

⇠ T

s

(x

c

1) Dark Ages

(15)

The Basics

ΛCDM + Astro Expectations

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

Dark Ages

Tk/(1 +z)² Tk⇠Ts

✓

xc Tcmb

Tk

◆

(16)

6/47 21cm Cosmology

The Basics

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

x

c

! 0

T

s

! T

cmb

Dark Ages

(17)

The Basics

ΛCDM + Astro Expectations

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

Cosmic Dawn

✓

x

_↵

T

cmb

T

k

◆

T

s

! T

k

(18)

6/47 21cm Cosmology

The Basics

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

T

k

T

cmb

✓ 1 T

cmb

T

_s

◆

! 1 Cosmic Dawn

(19)

The Basics

ΛCDM + Astro Expectations

z

Temperature T

cmb

~200

~30

~15

~10

T

k

~100

T

s

x

_H

! 0 T

_b

! 0

Reionization

(20)

7/47 21cm Cosmology

The Basics

Observing Modes

• Globally averaged signal δT b (z)

• Single dish radio telescopes (e.g. EDGES)

• Power spectrum (2-point correlation function) P 21 (z, k)

• Radio interferometers (e.g. HERA, SKA)

Focus of this work

• Tomography (Futuristic, but potentially quite powerful) δT b (x 1 , x 2 , z)

Not subject of today’s talk

(21)

The Basics

Observing Modes

• Globally averaged signal δT b (z)

• Single dish radio telescopes (e.g. EDGES)

Focus of this work

• Tomography (Futuristic, but potentially quite powerful) δT b (x 1 , x 2 , z)

(22)

7/47 21cm Cosmology

The Basics

Observing Modes

• Globally averaged signal δT b (z)

• Single dish radio telescopes (e.g. EDGES)

Focus of this work

• Tomography (Futuristic, but potentially quite powerful) δT b (x 1 , x 2 , z)

(23)

The Basics

ΛCDM Expectations + Astro Modeling

Global average

(24)

9/47 21cm Cosmology

Current Status Cosmological Implications 21cm Sensitivity

PBHs as Dark Matter

Sato-Polito, Kovetz, Kamionkowski (2019),

(25)

Solar Mass PBHs

O (M ) PBHs cannot account for entirety of dark matter, but still of interest to community

• Detection → insight into early Universe

Production from inflation, topological defects, phase transitions, etc

• Detection → Implications for dominant dark matter candidate e.g. WIMPs and PBHs, a no-go

Bertone et al (2019), Adamek et al (2019), Boucennaa et al (2017), Lacki and Beacom (2010), Eroshenko (2016)

• SMBH timing problem (10 9 − 10 10 M at z ∼ 7) Could be accomplished via 10

²

M

− 10

⁶

M

at z ∼ 15

Smith, Bromm, Loeb (2017), Haiman et al (2019)

(26)

10/47 21cm Cosmology

Solar Mass PBHs

• Detection → insight into early Universe

• SMBH timing problem (10 9 − 10 10 M at z ∼ 7) Could be accomplished via 10

²

M

− 10

⁶

M

at z ∼ 15

(27)

Solar Mass PBHs

• Detection → insight into early Universe

• SMBH timing problem (10 9 − 10 10 M at z ∼ 7) Could be accomplished via 10

²

M

− 10

⁶

M

at z ∼ 15

(28)

Cosmological Implications

Presence of ∼ M PBHs lead to 3 effects:

• Structure modified on small scales via PBH clustering

Isocurvature perturbations induced via shot noise

Carr Hawking (1974), Mukhanov et al. (1992), Afshordi, McDonald, Spergel (2003), Gong, Kitajima (2017, 2018)

• Accretion of gas onto PBH leads to:

• Emission of X-rays that escape into IGM (global effect)

• Local heating via advection, local absorption of X-rays (See Pablo Villanueva-Domingo’s poster)

(29)

Cosmological Implications

• Accretion of gas onto PBH leads to:

• Local heating via advection, local absorption of X-rays

(See Pablo Villanueva-Domingo’s poster)

(30)

Cosmological Implications

• Accretion of gas onto PBH leads to:

• Local heating via advection, local absorption of X-rays (See Pablo Villanueva-Domingo’s poster)

(31)

PBHs in the IGM

Injected energy goes into heating and ionizing IGM

Mena, Palomares-Ruiz, Villanueva-Domingo, SJW

(32)

PBHs in the IGM

Injected energy goes into heating and ionizing IGM

CMB and 21cm provide very complimentary probes!

(33)

PBH on δT b

(34)

21cm Power Spectrum

(35)

21cm Power Spectrum

(36)

Sensitivity

HERA: Running Now (Upgrade soon),SKA-Low: early 2020s

(37)

Conclusions

• 21cm Cosmology promises to be a power tool for both cosmology and particle physics alike

• ∼ M PBHs can give rise to various effects, leaving significant imprint on 21cm signal

Near-future experiments could increase sensitivity by ∼ 2 orders of magnitude

• New computational techniques necessary

Practical computations and information extraction

Ask if interested...

(38)

Thank you!

This project has received funding/support from the European Union‘s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 674896

(39)

Back Up Slides

(40)

The 21cm Signal

Predictions are attained from absorption and emission properties of the IGM

H H

H

H H

H H H H

H

Solve radiative transfer equation (neglect scattering) for IGM cloud...

dI ν

d` = − α ν I ν + j ν (3)

Opacity Emission Coefficient

τ ν = Z

α ν d` (4)

Can now express δT

b

in terms of T

s

and properties of IGM

(41)

Absorption/Emission Coefficients

Solving for I ν requires knowing j ν , α ν , which are dictated by:

α ν ∝ (n 0 B 01 − n 1 B 10 ) (5)

j ν ∝ n 1 A 10 (6)

Rather than use n 0 and n 1 → use n H and T s

Boltzmann factor for excitation temperature (‘spin temperature’ T s ) n 1

n 0 = g 1

g 0 e −∆E

¹⁰

/k

^B

T

^s

, (7)

(42)

The 21cm Signal

I ν = I cmb e −τ

^ν

+ Z

d` j ν

α ν

e −τ

^ν,r

(`) dτ ν,r

d` (8)

H H

H

H H

H H H H

H

For IGM cloud j ν

α ν independent of ` and τ ν 1, leading to:

δI ν ≡ I ν − I cmb (ν) = j ν

α ν − I cmb (ν)

τ ν (9)

See e.g. Pritchard and Loeb (2011) for review

(43)

The 21cm Signal

I ν = I cmb e −τ

^ν

+ Z

d` j ν

α ν

e −τ

^ν,r

(`) dτ ν,r

d` (8)

H H

H

H H

H H H H

H

Convert intensity to effective temperature I ν → 2 ν 2 T b

δI ν → δT b ≡

T s − T cmb (z) 1 + z

τ ν (9)

See e.g. Pritchard and Loeb (2011) for review

(44)

The Spin Temperature

T s (equivalently n 1 /n 0 ) determined by:

See e.g. Pritchard and Loeb (2011)

• Spontaneous absorption/emission and stimulated emission (A 10 , B 10 I cmb , B 01 I cmb )

• Collision excitation/de-excitations with HH, He, Hp (C 10 , C 01 )

• Indirect excitation/de-excitations via scattering with Ly − α photons (P 10 , P 01 ), ‘Wouthuysen-Field effect’ (WF)

(45)

• Spontaneous absorption/emission and stimulated emission (A 10 , B 10 I cmb , B 01 I cmb )

• Collision excitation/de-excitations with HH, He, Hp (C 10 , C 01 )

• Indirect excitation/de-excitations via scattering with Ly − α

photons (P 10 , P 01 ), ‘Wouthuysen-Field effect’ (WF)

(46)

• Spontaneous absorption/emission and stimulated emission

• Collision excitation/de-excitations with HH, He, Hp (C 10 , C 01 )

• Indirect excitation/de-excitations via scattering with Ly − α photons (P 10 , P 01 ), ‘Wouthuysen-Field effect’ (WF)

(47)

The rates are fast enough that we can assume an equilibrium, allowing us to derive:

n 1 (A 10 + C 10 + P 10 + B 10 I cmb ) = n 0 (C 01 + P 01 + B 01 I cmb ) (10) which can be expressed in terms of T cmb and T k

T s ' T cmb + (x c + x α )T k

1 + x c + x α , (11)

where x c and x α are the collisional and WF coupling coefficients

(48)

The rates are fast enough that we can assume an equilibrium, allowing us to derive:

n 1 (A 10 + C 10 + P 10 + B 10 I cmb ) = n 0 (C 01 + P 01 + B 01 I cmb ) (10)

T s ' T cmb + (x c + x α )T k

1 + x c + x α

, (11)

where x c and x α are the collisional and WF coupling coefficients

T s ' T cmb if x c , x α T cmb /T k (12) T s ' T k if x c (x α ) T cmb /T k (13)

(49)

ΛCDM Expectations + Astro Modeling

Global average

(50)

ΛCDM Expectations + Astro Modeling

Global average

Appearance of WF effect depends on when stars form (M

min

or T

_min^vir

) Note: non SF halos (‘minihalos’) have ∼ negligible impact on 21cm signal in

ΛCDM

Furlanetto et al (2006)

←−−−→

(51)

Global average

Strength of x

α

depends on number of Ly-α produced by stars (N

α

)

←−− −→

(52)

Global average

Depth of absorption depends on efficiency of X-ray heating (ζ

X

)

←− −− →

(53)

ΛCDM Expectations + Astro Modeling

Global average

Heating bump depends on efficiency of ionizing photons (ζ

U V

)

←→

(54)

ΛCDM Expectations

Simplified astrophysical model

200 150 100 50 0

Tb

[m

K]

EDGES peak location

UV

= 5 40

UV

= 40 105

EDGES peak location

Tvir^min

= 10

⁴

5 × 10

⁴

K

Tvir^min

= 5 × 10

⁴

10

⁵

K

10 15 20 25 30

z

200

150 100 50 0

Tb

[m

K]

EDGES peak location

X

= 2 × 10

⁵⁵

2 × 10

⁵⁶M¹

X

= 2 × 10

⁵⁶

2 × 10

⁵⁷M¹

10 15 20 25 30

z

EDGES peak location

N = 4 × 10²

4 × 10

³ N = 4 × 10³

4 × 10

⁴

SJW, Villanueva-Domingo, Gariazzo, Mena, Palomares-Ruiz (2018)

(55)

ΛCDM Expectations

Simplified astrophysical model

0.0 0.2 0.4 0.6 0.8 1.0

xi

UV

= 5 40

UV

= 40 105 Ly emitting galaxies Gunn-Peterson effect

T^minvir

= 10

⁴

5 × 10

⁴

K

T^minvir

= 5 × 10

⁴

10

⁵

K Ly emitting galaxies Gunn-Peterson effect

6 8 10 12 14 16 18 20

z

0.0

0.2 0.4 0.6 0.8 1.0

xi

X

= 2 × 10

⁵⁵

2 × 10

⁵⁶M¹

X

= 2 × 10

⁵⁶

2 × 10

⁵⁷M¹

Ly emitting galaxies Gunn-Peterson effect

6 8 10 12 14 16 18 20

z

N = 4 × 10²

4 × 10

³ N = 4 × 10³

4 × 10

⁴

Ly emitting galaxies Gunn-Peterson effect

SJW, Villanueva-Domingo, Gariazzo, Mena, Palomares-Ruiz (2018)

(56)

EDGES

Experiment to Detect Global Epoch of reionization Signature

Bowman et al (2018)

(57)

EDGES

Experiment to Detect Global Epoch of reionization Signature

Pritchard,Talk (2011)

Systematics questioned by Hills et al (2018) and Bradley et al (2019)

(58)

Single Dish → Interferometer

• More powerful removal of foregrounds in k-space

• Access to power spectrum

(59)

Beyond Global Signal

Power Spectrum

Global temperature is only a 1 st order test statistic

δ T

_b

∼ β b δ b + β x δ x + β α δ α + β T δ T + β v δ v (14) Baryon overdensity

Neutral hydrogen Ly-α

Gas temperature Velocity gradient

D

δ T

_b

(~k, z)δ ∗ T

_b

( k ~ 0 , z) E

= P 21 (~k, z) (15)

See e.g. Furlanetto, Oh, Briggs (2006), Pritchard and Loeb (2011)

(60)

Beyond Global Signal

Tomography

Z

Comments on this later

(61)

Shot Noise

For PBHs that form during radiation dominated epoch, one expects an average of N per Hubble volume

VH VH

Poissonian Fluctuations

→ Manifests as isocurvature modes that grow after M-R equality

Carr Hawking (1974), Mukhanov et al. (1992), Afshordi, McDonald, Spergel (2003)

(62)

Shot Noise

Matter Power Spectrum

Afshordi, McDonald, Spergel (2003)

(63)

Shot Noise

Halo Mass Function

M pbh = M and f pbh = 10 −1 → 10 −5

Gong, Kitajima (2017, 2018)

(64)

(65)

⇢, T k r B

r 1 , T k,1

(66)

⇢, T k

(67)

PBH Accretion

ADAF

Shapiro and Lightman (1976), Ipser and Price (1977), Ruffert (1999):

Accretion disk should form when angular momentum at Bondi radius is sufficient to maintain Keplerian orbits at r r S

Poulin et al (2017):

For PBHs this implies disk forms if p f pbh

M

1 + z 730

3

(16)

(68)

PBH Accretion

ADAF

Shapiro and Lightman (1976), Ipser and Price (1977), Ruffert (1999):

Accretion disk should form when angular momentum at Bondi radius is sufficient to maintain Keplerian orbits at r r S

Poulin et al (2017):

For PBHs this implies disk forms if p f pbh M

M

1 + z 730

3

(16)

→ Expect formation of optically thin puffy disk that radiates inefficiently (Advection Dominated Accretion Flow)

Expectations for and M ˙

(69)

PBH Accretion

ADAF

(70)

PBH Accretion

ADAF

Poulin et al (2017) Yuan and Narayan (2014)

(71)

Local Accretion

Must derive [ρ(r), T (r), x e (r)] from fluid equations (specifically at r ∼ r B )

Using simplified set of accretion equations (e.g. adopting spherical symmetry, no X-ray absorption, etc), we can derive radial profiles In L/L Edd → 0 we get Bondi-like problem

Leading to ρ ∝ r −3/2 and T ∝ r −1 for r/r B 1

(72)

Local Accertion

Must resolve radiative transfer equation for local contribution

Iliev et al (2002)

T b,loc (ν) = T cmb e −τ

^loc

(ν) + Z ∞

−∞

dR T s (R) e −(τ

^IGM

+τ

^loc

(ν,R)) dτ loc dR

(16) + absorption of CMB locally

+ signal produced locally

+ absorption of signal produced locally

Total Signal:

T ¯ b ∝ n pbh h ∆ v

_eff

T b,ν

₀

i (17)

(73)

Local Accretion

Profile Single PBH Total Signal All PBHs

Contrary to previous claims in literature, this contribution is negligible

(74)

Minihalos

Recall that dn/dM enhanced by shot noise

→ question remains whether minihalos can contribute significantly

Gong and Kitajima (2017, 2018)

T b,mh (ν) = T cmb e −τ

^IGM

(e −τ

^mh

(ν) − 1)+

Z ∞

−∞

dR T s (R) e −(τ

^IGM

+τ

^mh

(ν,R)) dτ mh

dR (18) Model ρ b and T k

T ¯ b ∝

Z M

max

M

min

dM dn

dM h ∆ v

_eff

T b,ν

₀

i (19) M max set to star formation threshold, M min = M J

(75)

Minihalos

Observability

Maximum minihalo contribution

∼ 8mK (very small modifications)

(76)

Cosmology Parameter Analysis

Kern et al (2017)

(77)

Velocity scalings in PBH accretion

Ricotti et al (2007)

(78)

Local Accretion

In limit that L/L Edd → 0, local heating dictated by gravity

4πr 2 ρ | v | = M ˙ = constant (20) v dv

dr = − 1 ρ

dP

dr − GM

r 2 (21)

Λ ion (1 − x e ) = αn H x 2 e (22)

Leading to ρ ∝ r −3/2 and T ∝ r −1 for r/r B 1

(79)

Minihalos

Observability

Gong and Kitajima (2017) Mena, Palomares-Ruiz, Villanueva-Domingo, SJW

Discrepancy an issue of self-consistency → M min = M Jeans ∝ T k 3/2

(80)

Parameter Estimation

7.8 7.2 6.6 6 log

10

(f

pbh

) 3.5 3 4.5 4 log

10

(N ) 4.2

4.5 5 4.7 5 5 log

10

(T

min

) 56 56.

56. 4 8 57. 2 log

10

(

X

) 20 40 60 80

UV

20 40 60 80

UV

56 56.4 56.8 57.2 log

10

(

X

) 4.25 4.5 4.75 5

log

10

(T

min

) 7.85

^+0.16_0.10

46.75

^+8.42_8.26

56.35

^+0.090.09

4.73

^+0.06_0.06

3 3.5 4 4.5 log

10

(N ) 3.65

^+0.11_0.10

(81)

A Comment on the Astrophysics

Why not use a more complicated astrophysical model?

X ~ ∈ (M pbh , f pbh , T min , N α , ζ X , ζ UV ) (23) (6D parameter space) → Interpolation for MCMC ( → linear?) 10 points / dim requires 10 6 evaluations

Each realization requires ∼ 6 hours on ∼ 8 cores...

(82)

A Comment on the Astrophysics

Why not use a more complicated astrophysical model?

X ~ ∈ (M pbh , f pbh , T min , N α , ζ X , ζ UV ) (23) (6D parameter space) → Interpolation for MCMC ( → linear?) 10 points / dim requires 10 6 evaluations

Each realization requires ∼ 6 hours on ∼ 8 cores...

→ ML techniques allowed 10 6 to be reduced to ∼ 10 3

(83)

Computational Complications

• Practical issue of computational time

Producing one full realization of 21cm maps using semi-analytic techniques may require & 6 hours on a cluster

• Issue of information extraction

Power spectrum does not contain full wealth of information...

(84)

• Issue of information extraction

(85)

• Issue of information extraction

→ Wish list: quick map production, extract more information (move beyond PS), easily identify new physics

Ruiz De Austri, SJW, Mena + others (in progress)