Baryon physics from the

kinematic Sunyaev-Zel’dovich effect

Emmanuel Schaan, Princeton/Berkeley Lab 


Nordita Advances in theoretical cosmology in light of data

July 2017
 Collaborators
arXiv:1510.06442, arXiv:1705.05881

Simone Ferraro Mariana Vargas-Magaña Kendrick Smith

Shirley Ho David Spergel Nick Battaglia

 Large-scale structure & baryons

Statistics: volume vs area

Probes dark energy, neutrino masses, non-

Gaussianities

But systematics…

baryons, biasing, non-linear evolution,
6 M. Haider etnon-Gaussian
al.
gas (non-radiative)
statistics Figure 4. Baryon density of the non-radiative simulation (no
star formation, feedback or cooling) at z = 0. The same slice as
in Fig. 3 is displayed.

Baryons ~15% total mass
 Illustris can also be used to probe the metal content in

→ largest uncertainty on the small-scale

gas and stars. Star particles are stochastically formed when
gas reaches densities above a threshold value ⇢sfr . A star
particle is modelled as a single-age stellar population, for

matter power spectrum

which the mass and metal return will be computed at every
time-step. This material is then distributed over neighbour-
ing gas cells (see Section 2.2 and 2.3 in Vogelsberger et al.
2013). In Fig. 13, we show the evolution of the metals in
gas and stars normalized to the total metal mass at z = 0.
Additionally, the figure shows which fraction of the metals

Can the CMB help?

in gas is residing in haloes, filaments or voids. We find that
at z = 0, 36 % of the metals are locked up in stars and 64 %
of the metals are in gas. Considering only the gaseous com-

gas (with feedback)

ponent, half the metals are within haloes and 28 % reside in
the filaments. The remaining 22 % are located in voids. The
average metallicity of the stars is 1.49 solar metallicities at
(a) dark matter z = 0, while the halo gas(b) has about 0.37 solar metallicities
baryons
on average. The average metallicity2 in filaments and voids
Haider+16, Illustris simulation
Figure 3. Dark matter and baryon density in a thin slice at z = 0. The slice covers the whole (106.5 Mpc) extent of the simulation
and has a thickness of 104 kpc (1 cell). is roughly 0.1 times the solar value.
 Halo imprints on the CMB
vthermal vbulk
Key parameters: 0
✓lensing ⇠ 1 , ⌧ ⇠ 10 3
,
c
⇠ 0.1,
c
⇠ 10 3

→ Various observables with complementary information :

{
Halo lensing / ✓lensing → Total density profile
⇣v ⌘2
thermal
Thermal SZ /⌧ → Thermal pressure profile
c
⇣v ⌘
bulk k
Kinematic SZ /⌧ → Gas density profile (but also

c
velocities, reionization, ULSS)
T
Scattering /⌧ → Gas density profile
T ⇣
vbulk ? ⌘2
Polarized SZ / ⌧ Q, ⌧ → Gas density (but also ULSS)
⇣v c ⌘
bulk ?
Moving lens / ✓lensing → velocities
c
…
 Kinematic Sunyaev-Zel’dovich effect
T
Z ⌧ v Counts all free electrons
= dl ne T Lower mass halos at higher z
T c
Small size: TkSZ ⇠ 0.1µK, TCMB = 110µK
Blackbody spectrum

104

in
cm
103

.ar
μK
C` [(µK)2]

12
102 total
lensed CMB
detector noise
tSZ, CIB, Radio sources, dust
101
`(`+1)

kSZ
2⇡

100
i n
.a rcm
K
10 1
102 103 1μ 104
Hand et al 2012 `
aps.org, ESO, ESA, Hubble, NASA
 Detection methods The Astrophysical Journal, 778:52 (20pp), 2013 November 20

Th
1’~ 382 kpc
view
A imat

Individual (monster) cluster
 B

equa
et al.
of M
sajou
Sayers+13, 14 C
The
et al
below
Fi
D frequ

Pairwise velocities

flux
respe
&O
et al.

Hand+12, Planck15, Soergel Flender Story Bleem

Limousin et al Uran
H ST F814W
C handra calib
the W
Figure 1. False-color composite image of MACS J0717.5+3745 with the lensing recen
+16, de Bernardis+16 results of Limousin et al. (2012) in blue, the Hubble Space Telescope image using
the F814W filter in green, and the Chandra X-ray image in red. The blue contours
show the Limousin et al. (2012) result on a linear scale, and clearly indicate the
108.
mod
four sub-clusters labeled A through D, with white Xs marking the sub-cluster the 1
positions determined by Ma et al. (2009) from the galaxy distribution. et al
(A color version of this figure is available in the online journal.) meas

Velocity reconstruction
 sub-cluster C to be the most massive system, and Ma et al.

mod
brigh
frequ
(2009) determined that sub-cluster C is probably the highly
a bes
Planck15, Schaan Ferraro+15
disturbed core of the main system. Sub-clusters B and D are
2.5 K
assumed to be relatively intact cores of systems that are merging
diffe
along a direction close to the line-of-sight. In particular, sub-
unce
cluster B is coincident with an X-ray temperature that is colder
calib
than the surrounding regions, indicating that its core has not
upda
been highly disrupted by the merger. From the spectroscopic

<T2 x tracer>, Hill+16, Ferraro+16, Planck17

accu
data, Ma et al. (2009) found that sub-cluster B has a line-
is 2
of-sight velocity that differs from the other components by
accu
approximately 3000 km s−1 . Further indications of this large
Furth
line-of-sight velocity for sub-cluster B were presented in M12,
appr
who found a similar best-fit velocity by using X-ray and SZ
angle
measurements to constrain the kinetic SZ signal toward that

T Power spectrum, George Reichardt+14

Plan
sub-cluster, although the statistical significance of their kinetic
calib
SZ constraint on the velocity is modest (≃2σ ). This wide range
to m
of observational data toward MACS J0717.5+3745 is therefore
that
converging to what appears to be a coherent picture of this
atmo
complex system.
brigh

T2 power spectrum, Smith Ferraro 16

on o
4. DATA REDUCTION To
4.1. Bolocam subtr
and
We observed MACS J0717.5+3745 with Bolocam from the 250
Caltech Submillimeter Observatory (CSO) for a total of 12.5 hr large
at 140 GHz and for a total of 27.3 hr at 268 GHz, where data
the effective band centers are quoted for a CMB spectrum. repre
Compared to the previous Bolocam analysis presented in M12, a 25
 Velocity reconstruction
35 : Methods and Data
Peculiar velocity ~v =
d~r
dt
H(t)~r =
d~x
d⌘
Mass conservation + linear approx. 

˙+r ~ · ~v = 0 ~v = aHf r ~ 1

→ vrms 1d ⇠ 300 km/s

150Mpc

Eisenstein+07, Padmanabhan+12,14
 Velocity reconstruction
~v = ~
aHf r 1

00M p c /h
7

vr [km/s]
1Gpc/h

2.5Gpc/h

Velocity data from Smith, Vargas-Magaña,

BOSSHo CMASS South DR11 footprint (sdss.org)
in the Northern Galactic Cap (NGC) while lower panels
 Gas profile of CMASS halos

profile”
comoving
comoving radius
radius at , zz = 0.57 [Mpc/h]
0.43
0.43 0.65 0.86 1.08 1.3 1.51 1.73 1.95
3 σ measurement (for now) 2.0
2.0
ACTPol & CMASS
ACTPol
best fit
best
1.5
1.5 tSZ residual
tSZ

“cumulative electron
Already useful profile information
1.0
1.0

Future is exciting: FW

α
✓ ◆r 0.5
0.5 N
S Mh Nobj
⇠ (1 or 2)
N 1013 M 104
0.0
0.0
BOSS & new ACTPol: area x 25
PFS/DESI: S/N x ~30 from Nobj −− 0.5
0.5
1.0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
SO/CMB S4: S/N x few from sensitivity,
 θdisk [arcmin]
x few from tSZ removal Cluster-centric radius [arcmin]
1.0
→ S/N~100

Schaan Ferraro +15

 Gas energetics from kSZ/tSZ

Measure 1-halo / 2-halo term → diffuse gas

Bin in redshift
Split populations
SZ from quasars
Gralla+14, Verdier+15, Crichton+15, Chowdhurry+17
2 Battaglia et al.
Virial theorem for gas:
functio
1
ˆ ˆ
gas + Pth dV + PNth dV / PSurface ,
|3 {z } | {z } | {z }
from kSZ from tSZ !inferred
&mass profile
where
here is analogous to the gravitational energy of
→ Constrain Pnon-th as a function of radius, hydro mass bias
gas the x coth(
gas, Pth is the thermal pressure, and PSurface is a surface stant, a
→ Constrain energy injected through feedback glected
pressure (See the Appendix for a full derivation).
Since the initial detections Battaglia
of the tSZ e↵ect
Ferraro on indi-
Schaan et al. 2
Spergel +15
vidual galaxy clusters (e.g., Pariiskii 1972; Meyer et al. eter is
1983; Partridge et al. 1987), there now exists an abun-
Summary: kSZ & gas physics in clusters

• Evidence for kSZ with ACTPol and velocity

reconstruction from BOSS arXiv:1510.06442

• KSZ powerful baryometer: abundance, profile

• Constrain non-thermal pressure and energy injection

with kSZ & tSZ

• CMB S4 and DESI will increase the S/N to ~100 


→ bin in mass/type/color

• Requires large aperture CMB telescope

Thank you!