21cm forest: challenges and prospects

21cm forest: challenges and prospects�
井上 進�(理研)
- generalities on 21cm forest
- detectability in GRBs with LOFAR/SKA Ciardi, SI+ 15
- prospects for probing dark matter
Lyman alpha forest�
from Ned Wright’s
webpage�
21cm forest (absorption lines)
-  significant before cosmic reionization z>6
- dominant signal from minihalos (M<108 MΘ)
- 10s of narrow lines (dν~ few kHz) out to z>~10
- sensitive to reionization details�
Furlanetto & Loeb 02
absorber abundance mock spectra
also
Carilli+ 02, Furlanetto 06, Xu+ 09, 11, Meiksin 11, Mack & Wyithe 12,
Vasiliev+ 12, Ciardi+ 13,15, Ewall-Wice+ 14, Tashiro+ 14, S’bukuro+ 14
21cm forest + mean IGM absorption� Carilli, Gnedin & Owen 02
assuming Cyg-A like source at z=10
narrow features from dense systems
+ mean absorption from IGM
21cm absorption:
effects
of
X-ray
heating�
K. J. Mack and J. S. B. Wyithe
The 21 cm forest with LOFAR
Mack
& Wyithe
Figure 3. Mean 21 cm optical depth
τ as a function
of observed12
frequency
transmissivity vs redshift
Non-uniform IGM model outputs in the redshift range 8 ≤ z ≤ 9.
ls show the overdensity δ, neutral hydrogen fraction xH I and 21 cm
epth τ in the non-uniform model region, from a two-phase IGM
eloped by Geil & Wyithe (2008). The size of the simulation box is
2. Non-uniform IGM model outputs in the redshift range 8 ≤ z ≤ 9.
cubed,
with
pixels per
side length.
Multiple
slices
the
nels show
the 360
overdensity
δ, neutral
hydrogen
fraction
xH Ithrough
and 21 cm
ndepth
box are
giving a totalmodel
of 280region,
cMpc.from a two-phase IGM
τ inused,
the non-uniform
for varying fX values: from top to bottom curve, fX = 0 (no X-ray heating)
0.01, 0.1, 1 (thick black line), 10 and 100. Also included are curves of Smin
(dashed purple lines) as defined in equation (9). These lines indicate, for
each value
of τ 21
oncm
theoptical
left-hand
axis,
thea function
minimumofobserved
density of
Figure
3. Mean
depth
τ as
observed flux
frequency,
a source
would from
allowtop
a detection
absorption
at S/N
of heating),
5, assuming
for
varyingthat
fX values:
to bottom of
curve,
fX = 0 (no
X-ray
an array
effective
area10
Aeff
106Also
m2 , included
frequency
0.01,
0.1, 1with
(thick
black line),
and=100.
areresolution
curves of S$ν
minch =
(dashed
purple lines)
as defined
These
lines indicate,
1 kHz, system
temperature
giveninbyequation
equation(9).
(10)
and integration
timefor
tint =
each
valueThe
of τredshift
on the left-hand
axis,comes
the minimum
fluxdependence
density of o
1 week.
dependence
from theobserved
frequency
athe
source
that temperature.
would allow aWhere
detection
absorption
S/N of
5, assuming
system
the of
mean
optical atdepth
lines
(solid) cross
6
2
an array with effective area Aeff = 10 m , frequency resolution $νch =
above the S lines (dashed), the absorption is detectable at that redshift
1 kHz, systemmin
temperature given by equation (10) and integration time tint =
The
grey
region
highlights the redshift range (z = 8–9) we focus on in our
1 week. The redshift dependence comes from the frequency dependence of
example
this work.
the
systemcalculations
temperature.inWhere
the mean optical depth lines (solid) cross
Ciardi+ 13
above the S
lines (dashed), the absorption is detectable at that redshift.
21cm cosmology: emission vs absorption (forest)
- emission (or absorption against CMB)
pro: 3D (all-sky + z-dependence) <-> 2D CMB
con: very weak signal << expected foreground
- absorption against high-z radio sources
pro: limited only by flux and number of sources
no or little foreground
con: limited by flux and number of sources
highly uncertain BUT interesting problem itself
potential background radio sources at very high z�
required spectral resolution dν~kHz (3km/s) at ν~<100 MHz
required flux for SKA
Furlanetto & Loeb 02
radio quasar as
background radio source �
assuming Cyg-A like source at z=10
possible even with LOFAR?
Ciardi+ 12
BUT
sufficiently luminous (~massive BH)
+radio-loud AGNs at high-z?
Carilli+ 02
high-z radio-loud quasars: expectations
Haiman+ 04�
assume:
- MBH scales with Mhalo similarly to low z
- radio loudness distribution
same as low z
results:
- number overpredicted by ~100
compared to FIRST obs.
unless radio-loud only for
MBH>~107 Msun - expected no. at z>6
~ 4/deg2 for >mJy
~ 0.2/deg2 for >10mJy
- expected no. at z>10?
~ 0.1/deg2 for >mJy
~ 0.005/deg2 for >10mJy
(200 all-sky)
high-z radio-loud quasars: observations
Into3the3Epoch3of3Reioniza?on?33
M. Jarvis
@Nagoya 13
unpublished?�
•  Using3powerful3radio3sources3within3the3EoR,3the3proper?es3of3the3
EoR3can3be3studied3in3absorp?on,3via3the3213cm3forest.33
•  High3flux3limit3([email protected]),3gives3~243sources3per3sq.deg.33
•  Obtain3spectra3for3all3of3these3and3expect313z>63powerful3radio3source3
per3~1000sq.deg3for3EoR3absorp?on3
GRBs as background radio source normal GRB afterglows log fν [erg s-1 cm-2 Hz-1]
sub-GHz afterglow flux strongly suppressed
by sync. self absorption
-> high-res. spectroscopy difficult
(even though continuum detectable by SKA)
-24
c.f. SI, Omukai, Ciardi 07
t=1 day
also Ioka, Meszaros 04�
EVLA
3km/s
-26
-28
z=1
5
10
20
30
SKA
3km/s
EVLA cont.
-30
SKA cont.
HI 21 cm
-32
adapted from
7.5
SI, Omukai, Ciardi 07
12 hr integ. 5σ sensitivities
8.0
8.5
9.0
log ν [Hz]
9.5
10.0
Pop III GRB�afterglow
Pop III GRBs:
光度は普通のGRBと大差ない�
が継続時間が長く�
総エネルギーはでかい �
Eiso~1055-1057 erg
-> afterglowは明るい�
普通のGRBより大きな半径まで�
広がる�
-> シンクロトロン自己吸収が減り�
低周波電波でも明るい�
-> 21cm吸収線の光源として有望�
Toma, Sakamoto
& Meszaros 10�
Pop III GRB afterglows: suitable for 21cm forest studies?
log fν [erg s-1 cm-2 Hz-1] �
-22
Jy�
21cm吸収線背景光源として
SKA観測で要求される強度は
~10mJy
100MHz�light curves�
-24
-26
log E [erg]=57�
56�
55�
54�
53�
mJy�
-28
E=1057 ergならt~30~1000 yr
で~10mJy�
see also Toma+ 11�
µJy�
-30
-32
-1
Nallsky~RGRB,57trad
全天10-100個あるには
RGRB,57~0.01-0.1/yr
T0=10000 s, z=20
next=0.1 cm-3, θj=0.3�
0
1
2
log t [day] �
3
4
Pop III GRB rates
adapted from Liu, SI, Wang & Aharonian, in prep.
Bromm & Loeb 06�
“normal GRBs”
Hopkins & Beacom 06
de Souza+ 11�
Campisi+ 11�
detectability with LOFAR/SKA
Ciardi, SI+ 15
- based B.
onCiardi
reionization
104
et al. simulations of Ciardi+ 12,13
- methodology for assessing detectability by LOFAR(+SKA) Ciardi+ 13
assume
Sin=30mJy at z=zs
marginally detectable by LOFAR
Figure 1. Upper panels: spectrum of a GRBIII positioned at zs = 10 (i.e.
Figure 3. As Fig. 1, but t
SKA1
Pop
energetic
ν detection
∼ 129 MHz),feasible
with a flux by
density
Sin (zs ) IF
= 30
mJy.III
TheGRB
red dotted
lines
ν ∼ 95 MHz) and "ν = 20
int
detectability with LOFAR/SKA
hand panels refer to a case with the noise σ n given in equation (2) (LOFAR
telescope) and with 0.1n (expected for SKA1-low), respectively. Lower
- based
reionization
simulations
offorCiardi+
12,13
panels:
S/N on
corresponding
to the upper
panels. See text
further details.
Ciardi, SI+ 15
- methodology for assessing detectability by LOFAR(+SKA) Ciardi+ 13
assume
Sin=30mJy at z=zs
Figure 4. Upper panels: sp
Sin (zs ) = 0.1 mJy. The red d
the source, Sin ; the blue dash
absorption, Sabs ; and the black
tion as it would be seen after
0.01 σ n (i.e. 1/10th of the SKA
zs = 7.6 and "ν = 5 kHz (left
and "ν = 20 kHz (right). Lo
panels. See text for further de
marginally detectable by LOFAR
Figure 2. As Fig. 1, but the GRBIII is positioned at zs = 7.6 (i.e.
feasible
by SKA1 IF Pop III GRB energetic
ν detection
∼ 165 MHz) and
"ν = 5 kHz.
Sin (zs ) = 0.1 mJy, a positiv
at any redshift, as shown i
than a few could be reache
detectability with LOFAR/SKA
Ciardi, SI+ 15
- based on reionization simulations of Ciardi+ 12,13
- methodology for assessing detectability by LOFAR(+SKA) Ciardi+ 13
assume
Sin=30mJy at z=zs
Downloaded fr
marginally detectable by LOFAR
Figure 3. As Fig. 1, but the GRBIII is positioned at zs = 14 (i.e.
bykHz.
SKA1
IIIinGRB
energetic
νdetection
∼ 95 MHz) feasible
and "ν = 20
Note IF
that Pop
the S/N
the lower-right
Figure 3. As Fig.
but the GRBIII is
detectability
with1, LOFAR/SKA
positioned at zs =
14 (i.e.
Ciardi,
SI+ 15
ν ∼ 95 MHz) and "ν = 20 kHz. Note that the S/N in the lower-right
- based on reionization simulations of Ciardi+ 12,13
panel is always higher than the range covered by the axis.
- methodology for assessing detectability by LOFAR(+SKA) Ciardi+ 13
- assume Sin=0.1mJy at z=zs
marginally
detectable
by SKA1
Figure
4. Upper
panels:
spectrum of a GRBII with a flux density
Sin (zs ) = 0.1 mJy. The red dotted lines refer to the intrinsic spectrum of
Figure 3. As Fig.
but the GRBIII is
detectability
with1, LOFAR/SKA
positioned at zs =
14 (i.e.
Ciardi,
SI+ 15
ν ∼ 95 MHz) and "ν = 20 kHz. Note that the S/N in the lower-right
- based on reionization simulations of Ciardi+ 12,13
panel is always higher than the range covered by the axis.
- methodology for assessing detectability by LOFAR(+SKA) Ciardi+ 13
- assume Sin=0.1mJy at z=zs
MNRAS 453, 101–105 (2015)
MNRAS 453, 101–105 (2015)
doi:10.1093/mnra
doi:10
Simulating the 21 cm forest detectable with LOFAR and SKA
Simulating
cm GRBs
forest detectable with LOFAR and SKA
in
the spectrathe
of 21
high-z
in the spectra of high-z GRBs
B. Ciardi,1‹ S. Inoue,2 F. B. Abdalla,3,4 K. Asad,5 G. Bernardi,6 J. S. Bolton,7
1‹ 8
3,4
5
6
5,8
3
5
5
B. Ciardi,
S.A.Inoue,
F. B. Abdalla,
K. Asad,
G. Bernardi,
J. S. Bolton
M.
Brentjens,
G. de2Bruyn,
E. Chapman,
S. Daiboo,
E. R. Fernandez,
8
5,8
39
5
5
1
3
11
A.
L. Graziani,
G.
J. A. Harker,
I. T. Iliev,
Jelić,5,8,10
M.Ghosh,
Brentjens,
A. G. de
Bruyn,
E. Chapman,
S.V.Daiboo,
E.H.R.Jensen,
Fernande
5 L. V. E. Koopmans,
3 5 A. Maselli,
5,8,10 11
S.
Martinez,
G. Mellema,
A.Kazemi,
Ghosh,12
L. Graziani,1 G. J. 5A.O.Harker,
I. T. Iliev,9 V.13Jelić,
H. Jense
5,8
5
14
5
5
12detectable
5
5
13
1
A.
R.
Offringa,
V.
N.
Pandey,
J.
Schaye,
R.
Thomas,
H.
Vedantham,
marginally
by
SKA1
S.
Kazemi,
L.
V.
E.
Koopmans,
O.
Martinez,
A.
Maselli,
G.
Mellema,
Figure 4. Upper
panels: spectrum
of a GRBII with a flux density
8
5
5,8S. Zaroubi
5
14
5
5
S.
Yatawatta
and
A.
R.
Offringa,
V.
N.
Pandey,
J.
Schaye,
R.
Thomas,
H.
Vedantham,
S (z ) = 0.1 mJy. The red dotted lines refer to the intrinsic spectrum of
1
in
s
s].
sto
rd
w,
Hs
ve
st
g).
n,
first supermassive black holes
E~ LEdd tSal
~ few x1058 erg
for MBH~106 Msol
early, gas-rich environment
-> ubiquitous blastwave formation
+ radio emission??�
Marziani+ 11�
Johnson+ 11�
1 kpc (physical)
log(n)
4
2
0
–1
–2
–3
and luminosity, the MBH drives power1 kpc (physical)
4
ful outflows that sweep away the sur3.5
Fig. 9. MBH
versus z for rounding
a low-z sample
(gray
dotsboth
Zamfir
et al., 2010), and seve
log(T)
gas, thus
halting
its own
3
intermediate to high z samples. Red circles: Marziani et al. (2009); open square
growth and star formation in the gal2.5
Dietrich et al. (2009);
open triangles: Shemmer et al. (2004); filled pentagon
axy. From
the original
idea (39),etrefined
Netzer et al. (2007); filled
squares:
Trakhtenbrot
al. (2011); open starr
models
of
feedback
have
been
de-(2007); large spot
octagons: Willott et al. (2010); filled octagons: Kurk et al.
veloped
theoretical in late June 20
z ! 7: the high-z quasar
whose (40–42).
discoveryMany
was announced
(Mortlock et al., 2011). The
dashed
MBH =
5 # strong
109 M$. (For interpretati
models
(43,line
44)marks
advocate
that
of the references to colour
in this
legend,
the reader
is referred to the w
activity
andfigure
powerful
feedback
occur
version of this article.)
during galaxy mergers, thus providing a link between bulge formation
and MBH growth. In the alternative
4.5
必要なhigh-z 電波源は
きっとある�
心配するな�
dN/d
21cm forest: halo mass dependence�
100
Shimabukuro, Ichiki,
REVIEW
SI,PHYSICAL
Yokoyama
14
D 90, 08
0.1
z=10
101
dN/dz
0.01
0.001
τ
1
0
10
10-2
0.1
α/r
we choose T IGM ¼ T ad , the average tempera
IGM assuming only adiabatic cosmic expansion
0
10 with our basic assumption of not accounting
physical feedback effects.
-1
10
0.01
101 where ρ̄ is the total mass density including dark
10
z=20 -1
0.1
0.001
0.01
z=10,TK=Tad
z=20,TK=Tad
d2N/dzd
τ
OBING
SMALL-SCALE
FLUCTUATIONS …
optical
depth vsCOSMOLOGICAL
impact parameter
!
"3=2
10
-1
4 -1
4π ρ̄ 5πkB T IGM
105h-1Msun
10
MJ ¼
10 h Msun
3
3Gρ̄mp μ
6 -1
absorber abundance per z interval
10 h Msun
!
"3=
7 -1
1
T
=K
10 h Msun
-2
IGM
2
5 −1
8 -1
10
≃
3.58
×
10
h
M
10
⊙
10 h Msun
1þz
10
1 1
-2
10
III. NONSTANDARD COSMOLOGI
EFFECTS AND RESULTS
Figure
abundance of 211 cm
0.01 3 shows the0.1
features per redshift interval along an average l
as a function of optical depth at z ¼ 10 and
baseline ΛCDM cosmology. Around a giv
expected number of absorption features with a
roughly zτd2 N=dzdτ, which at z ¼ 10 is seen t
5, 0.7 for τ ∼ 0.01, 0.1, 1, respectively, app
observer frequency νobs ∼ 129 MHz. At z ¼ 20
1010
ττ
number
Σ=0.0eV
ffective
of massive neutrino
Σ=0.0eV
CDM
Σ=0.1eV
2
mWDMΣ=0.1eV
=30keV
heofenergy
density
of
massive
neutrinos
10
Σ=0.5eV
1
halos
in
different
mass
ranges
to
the
Σ=0.5eV
1 -1
of
halos
in
different
mass
ranges
to
the
- 
WDM
strongly
suppresses
LSS
m
=20keV
10
WDM Σ=1.0eV
ritical
density
ρcr , at z=10 (top) and
1010
absorption
features
absorption
at
z=10 (top)
and
mWDMΣ=1.0eV
=10keV
CDM scale
4 features
below
m
-dependent
%
WDM
1
10
mWDM=10keV
mWDM=6keV
10
m
ρ- νparticle
i
candidate:
neutrinos
0 0
i
mWDMsterile
=2keV (15)
mWDM=2keV
=
=2
,
10
2
10
ρ-crsolve
93.14h
[eV]satellite problem?
1010-2 6
missing
0
0
10
z=10�
10
is -given
by
current
limits
m
>~1
keV
given
by [47]
[47]
WDM
ass of10each
neutrino family [9]. Tak-2
-1-1 8
!
""
!
10
#
$
10
#
$
we
plot
halo
mass
functioneffect
-1 -3
Ωcount,
m
0.1N
ν
ννthe keV:
m
0.1N
~<30
significant
-4
ν 4- m
WDM
1010
z=10,Σm
=0.0eV
10
−0.8
.
(14)
∼
−0.8
.
(14)
ν
10 ∼
of
massive
neutrinos
in
Fig.5.
It
can
22
1eV
ΩΩmm
hhtotal
m
mWDM
<few
z=10,Σm
=1.0eV
1eV keV:
6
-6
m
νsuppression
10
ppression
of the mass function
is most
-2-2
z=20,Σmν=0.0eV
2
10
-2
10 mass
10 -4
ective
number
of
massive
neutrino
-8
z=20,Σm
=1.0eV
10
igh
end
where
it
falls
off
expoν
ective
number
of
massive
neutrino
10
10
ee energy
density
of
massive
neutrinos
absorber
z=10,Σm
0
0.01 abundance
0.1
z=20�
energy
neutrinos
ν=0.0eV
104density of massive
10
CDM
-1
4 have
itical
density
ρcr ,already z=10,Σm
4
ν=1.0eV
servations
placed
rela10
-1 per z interval
10
itical
density
ρ
,
10
-1
m
=10keV
z=20,Σm
=0.0eV
cr
2
ν
τ
10
WDM
10
% masses. z=20,Σm
10
-2 on neutrino
=1.0eV
nts
For
example,
2
ν
102ρν 10
%i mi m%
%10
WDM=2keV
0
m
=mρνν <=10
(15)
i ,
0.17
eV
and
mν < 0.26
eV
i2
FIG. -2
6: Abundance of 21 cm absorption features pe
0
= 0ρ-4cr 10
= 93.14h 2[eV] ,
(15)
10
halo mass
10ρcr
-2 -2at z = 10 (top) and z = 20 (bottom), for diff
om
observations
of the Lyα forest
and
93.14h [eV]
interval
10
-2
10
10
!
-2
10
function�
ly-2-6[10,
48], while
recent
Planck
results
ass
of 10
each
neutrino
family
[9].
Takues of the total neutrino mass
mν = 0.0eV, 0.1eV
0
10
-4
ass
of each
neutrino
family
-4the
10
eV
from
CMB
alone
[3].[9].
To Takilluscount,
plot
the halo
mass
function
and 1.0eV.
-3
10we
z=20�
10
count,
we
plotwe
theconsider
halo
function
-3 -3
of-4-8massive
in mass
Fig.5.
Itcases,
can
eutrino
mass,
three
-6
-6 neutrinos
010
10
10
1010
of
massive
neutrinos
in Fig.5.is It
can
pression
of the
mass function
most
d-61.0 eV.
-8
0pression
1010-8of
the
massitfunction
most
-4
-10mass
gh
end
where
falls off isexpo10
10
4end
109 expo12 11
-8
where
it
4
5 6 6 10
78 falls
810off
10
1214 10-4 -4
0.01
0.1
1
10
10
10
10
-10
0gh mass
10
10 10 10 10 10 10 10 10 10
2
3
2
τdτd
N/dzdτ
2N/dzdτ
τd N/dzdτ
-3
dn/dlnM[h
[h3 Mpc
Mpc-3]]
dN/dlnM
3
2
-3
dn/dlnM [h Mpc ]
2
τd
N/dzdτ
τd
2N/dzdτ
2
τd
N/dzdτ
τd N/dzdτ
warm dark matter�
10
-1
8
ervations have
-1
104 already
106M [h
10placed
1012
M 1010
] rela-
1014
0.01
0.01
0.10.1
1 1
dark matter annihilation heating in minihalos als.org/ at University of Tokyo Library on August 19, 2015
hile the right shows the same with a Burkert profile. The black contour lines again correspond to the values of the colour bar. The green line
s for which molecular cooling is possible. Halo models to the right of the line can cool efficiently.
e.g. Ripamonti+ 07,10
Natarajan+ 09, Stacy+ 12
…
Schön+ 15
Figure 6. Ratio of energy produced by DM annihilation over the Hubble time to the halo’s gravitational binding energy for a halo with an Einasto profile and
of energy
produced
by DM annihilation
overand
the Hubble
timeThe
to the
halo’sshow,
gravitational
bindinga energy
forparticle
a halo with
an Einastovia
profile
andquark and tau particles as well as
panels
respectively,
50 GeV
annihilating
muon,
Duffy
mass–concentration
relation
fabs = 0.1.
=
0.1.
The
panels
show,
respectively,
a
50
GeV
particle
annihilating
via
muon,
quark
and
tau
particles
as
well
entration
relation
and
f
83-GeV particleabs
annihilation via a W boson. Again the green line is the critical halo mass for molecular cooling, while theasblack contour lines correspond to
nnihilation
via a bar.
W boson.
Again the
line of
is the
halo
mass
for molecular cooling, while the black contour lines correspond to
the colour
Halo models
togreen
the right
the critical
line can
cool
efficiently.
lo models to the right of the line can cool efficiently.
effect on 21cm forest? MNRAS
451, 2840–2850 (2015)
840–2850
(2015)
まとめ�
- 21cm forest: 21cm emissionと相補的な
宇宙黎明期・再電離期・宇宙論のプローブ
まだ研究の余地いろいろあり
e.g. DM heating
- 背景電波光源：quasar? Pop III GRB?�first SMBH??
��不定性大きいがそれ自体おもしろい研究課題
- energetic Pop III GRBが存在すれば21cm forestは
z>~7でSKA1で観測可能かも LOFARではmarginal
normal GRBならSKA2が必要