Transcript Why Study Cosmic Near Infrared Background? (1-4um)

Probing the High-z Universe
with Galaxy Counts from Ultra
Deep Surveys and the Cosmic
Near Infrared Background
Eiichiro Komatsu (Univ. of Texas, Austin)
Astrophysics Seminar@UCSB
May 23, 2007
References
Elizabeth Fernandez & EK, ApJ, 646, 703 (2006)
Elizabeth Fernandez & EK, to be submitted
What Do I Mean By “High-z”?



I mean z>6.
An interesting epoch
in the cosmic history:
reionization of the
universe
Direct detections of
galaxies at z>6 are
now possible.

eg., Ly emitter at
z=6.96 discovered in
the Subaru Deep
Field (Iye et al. 2006)
Going Further…



JWST will peer deeper into the high-z
universe…after 2013.
Can we do anything interesting now, and help
define science goals of JWST better?
Two topics in this talk along this direction:



The Cosmic Near Infrared Background
Lyemitters at z>6
One more topic if time permits:

21-cm fluctuations vs the Cosmic Microwave
Background (Alvarez, EK, Dore & Shapiro 2006)
Why Study Cosmic Near
Infrared Background? (1-4um)


New window into 7<z<30 (e.g., Redshifted Ly)
Can we detect photons from early generation stars and
their nebulae? What can we learn from these photons?

The presence of the signal is guaranteed, but the
amplitude of the signal is not known.

Measurement of these photons is challenging because
of contaminations due to:


Zodiacal light, and
Galaxies at z<6.
Near Infrared Background:
Current Data vs Challenges


Extra-galactic infrared background
in J and K bands above “zodiacal
light” ~ 70 nW/m2/sr
These Measurements have been
challenged.

Observed
NIRB
Excess?
Upper limits from blazar spectra:
<14 nW/m2/sr (Aharonian et al.
2006)

Incomplete subtraction of Zodiacal
light? ~15 nW/m2/sr (Wright
2001); <6 nW/m2/sr
(Thompson et al. 2006)

Let’s be open-minded.


Clearly, we need better data!
Better data will come from a
rocket experiment, CIBER
(Bock et al), in 2008.
Galaxy
Contribution
at z<6
Matsumoto et al. (2005)
Previous Study: Metal-free
Stars, or Mini-quasars?

First stars?
Very massive (~1000 Msun), metal-free
(Z=0) stars can explain the excess signal.
 Santos, Bromm & Kamionkowski (2002);
Salvaterra & Ferrara (2003)


Mini quasars?
Cooray & Yoshida (2004) studied the
contribution from mini-quasars.
 Madau & Silk (2005) showed that it would
over-produce soft X-ray background.

Our Prediction:
Fernandez & Komatsu (2006)

We don’t need metal-free stars!



Stars with metals (eg, Z=1/50 solar) can
produce nearly the same amount of excess
light per star formation rate.


Don’t be too quick to jump into conclusion that
metal-free, first stars have been seen in the NIRB.
(Kashlinsky et al. 2005, 2007)
We don’t need them (yet) to explain the data.
This is not a negative result, but is actually a good
news for NIRB: we don’t really expect a lot of
metal-free stars to be around at z~7-10.
Why? A simple energy conservation.
Simple, but Robust, Calculation
c
I 
4
What we
measure
p( ,z)

p([1 z] ,z)dz
H(z)(1 z)
 (M*c 2 ) /Time  Efficiency
2

Ý
= 
(z)c

e

*


Unknown
Can be
calculated
1
“Radiation 
e 
Efficiency”
m*
Simple argument:
Luminosity per volume
= (Stellar mass energy)
x(Radiation efficiency)
/(Time during which
radiation is emitted)
L (m) (m) 
 dm mf (m) mc 2 
IMF (Salpeter, Larson, Top-heavy)
Stellar data from Schaller et al. (1992); Schaerer (2002)
Sample Initial Mass Functions
of Stars
Salpeter:
Larson:
Top-heavy:
(
)
Rest-frame Spectrum of <>
NIRB Spectrum per SFR
Ý*
I / 
The “Madau Plot” at z>7
You don’t have to take this seriously for now.
We need better measurements!
How About Metal Production?

Is the inferred stat formation rate at z>7
consistent with the metal abundance in
the universe?

Did these early stars that are responsible
for the near infrared background overenrich the metals in the universe too early?
Theoretical data for
Z=1/50 solar from
Portinari et al. (1998)
White dwarf Type II SN Weak SN
or neutron
Black hole
star
by fallback
Direct
collapse to
black hole
Pulsational
Pair
Instability SN
Pair
Instability
SN
Metal Production (Z=1/50 solar)
The metal density now is 1.2 * 108 M8 Mpc-3
-> The upper limit from the near infrared
background for a larson IMF is excluded,
but most of the parameter space survives
the metallicity constraint.
A Comment on
Madau & Silk (2005)

They claim that the stellar mass density required to
explain the excess near infrared background is at
least 2% of the baryon density in the universe.


“this is energetically and astrophysically daunting”
(Madau & Silk 2005)
It would be “daunting” if, and only if, these baryons
had remained locked up in the stars and their
remnants; however,

Baryons should be recycled!! If all the baryons were
recycled (other extreme case), 2% should be divided by
the number of generations of star formation, which is of
order 10. So, the actual number should be somewhere
between 2% and ~0.1%, which is not daunting at all.
“Smoking-gun”: Anisotropy



Press-release from Kashlinsky
et al.:
 Detection of significant
anisotropy in the Spitzer
IRAC data
 They claim that the detected
anisotropy originates from
the first stars.
Their claim has been challenged
by Cooray et al.
We need better data from
CIBER, which is designed to
measure anisotropy over 2 deg2

The Spitzer image (left) is over
12’x6’.
The Future is in Anisotropy

Previous model (Kashlinsky et al. 2005; Cooray et al.
2006) ignored ionized bubbles.


May not be accurate enough to interpret the data from CIBER.
We will use the reionization simulation (Iliev et al. 2006)
to make simulated maps of the NIRB anisotropy: coming
soon!
What Are the Sources of the
Near Infrared Background?

One candidate: Lyman-alpha emitting galaxies at
z>7.


Subaru Deep Field




34 LAEs at z=5.7 (Shimasaku et al. 2006)
17 LAEs at z=6.5 (Taniguchi et al. 2005; Kashikawa et al.
2006)
1 LAE at z=7 (Iye et al. 2006)
LALA Survey


What do we learn about them from the existing Lymanalpha Emitter (LAE) searches?
1 LAE at z=6.5 (Rhoads et al. 2004)
ISAAC/VLT

No detection at z=8.8 (Willis et al. 2006; Cuby et al. 2007)
Very Simple Model of
Luminosity Function
dn
N( L)  Survey Volume    dM
dM
M (L )


Simply count the number of halos above a
certain mass = Cumulative Mass Function
Mass is related to luminosity by a “mass-tolight ratio” = M/L (M is the total mass.)


We just stretch the cumulative mass function
horizontally by rescaling the mass with M/L.
One parameter fit!
Cumulative Mass Function
(Sheth-Tormen Mass Function)
If we stretch the horizontal axis by M/L, then we get…
Luminosity Function of LAEs (1):
SDF at z=5.7
M/Lband=95-120
Luminosity Function of LAEs (2):
SDF at z=6.5
M/Lband=85-100
Luminosity Function of LAEs (3):
SDF at z=7 (from 1 LAE)
M/Lband~100
Luminosity Function of LAEs (4):
VLT/ISAAC at z=8.8 (no detection)
M/Lband>7
Mass-to-“observed light” Ratio to
Mass-to-“bolometric light” Ratio

The luminosity of LAEs estimated from a given
survey is not the actual luminosity of the source.



It’s a luminosity integrated over instrument’s
bandwidth.
It’s a luminosity after absorption and extinction.
Conversion:
M
M Lbol 1

Lband Lbol Lband  esc
Getting Lbol/Lband From
Model Spectrum

A sample spectrum for a galaxy of Z=1/50 solar
with a Salpeter IMF. The intrinsic equivalent
width of Lyman-alpha = 483 angstroms.
Lband/Lbol: How Much Light
Are We Losing?

Lower metallicity -> Larger Lyman-alpha



We don’t lose much light -> less correction
necessary.
Hence, larger Lband/Lbol.
Very insensitive to the IMF
Main Result: Inferred (M/Lbol)/esc



(M/Lbol)/esc is remarkably stable from z=5.7 to 7!
No detection of sources at z=8.8 is consistent with
the expectation.
We see no evidence for the evolution of (M/Lbol)/esc
from z=5.7 to 8.8.
What Do Our Results Imply?

LAEs are normal galaxies, if:  No evidence for the



M/Lbol~10, if a good fraction of
Lyman-alpha photons survived,
esc~0.5.
The predicted EW is consistent
with observation, EW~50-300
angstroms, if the metallicity is
“normal”: Z=1/50-1 solar.
LAEs are starbursts, if:


M/Lbol~1, if esc~0.05-0.1.
The predicted EW is consistent
with observation, if the metallicity
is low: Z=0-1/50 solar.
end of reionization!

No evidence for the
evolution of esc,
unless esc goes down
and M/Lbol goes up by
the same amount to
keep (M/Lbol)/esc
constant.
You can make it more complex.
dn
dn
 dM dM    dM dM P(L | M)
M (L )

Why not stretching it vertically as well?




“Duty cycle”, . (Haiman & Spaans 1999; Dijkstra, Wyithe
& Haiman 2006).
In our model, 
Why assuming a deterministic L-M relation?



M (L )
P(L|M): Probability for a halo of mass M to host a galaxy
of luminosity, L. (e.g., Cooray & Milosavljevic 2005)
In our model, P(L|M)=delta[M-(M/L)L].
The data are not good enough to constrain more
parameters necessary to characterize these
properties.

Our simpler model does yield reasonable results.
A Comment on Salvaterra&Ferrara (2006)

They claim that the excess near infrared
background cannot originate from high-z galaxies,
because such galaxies are not seen in high-z
galaxy surveys.



They show that the excess NIRB requires hundreds of
galaxies to be detected in e.g., VLT/ISAAC field, where
none was found.
Their model galaxy is extremely bright: M/L~0.001.
But, we don’t need such a population!


NIRB measures the TOTAL energy.
Galaxies can release the same amount of energy by
• an intense starbust for a few million years (M/L~0.001),
• a moderate burst for a few hundred million years (M/L~0.1-1), or
• a normal star formation (M/L~10).

SF(2006) ruled out only the first possibility.
Reionization & CMB - 21cm correlation
Alvarez, Komatsu, Dore & Shapiro 2006, ApJ, 647, 840
Doppler is a
projected
effect on CMB
 Doppler
effect comes
from peculiar velocity
along l.o.s.
 21-cm fluctuations
due to density and
ionized fraction
 We focus on degree
angular scales
21-cm maps result
from line-emission

21cm lines




21cm x CMB Doppler
Produced by neutral hydrogen during reionization
As reionization proceeds, 21cm slowly dissappears –
morphology of reionization imprinted on 21cm anisotropy
Because it is line emission, redshift  frequency
CMB Doppler effect


Free electrons during reionization scatter CMB photons
• Electrons moving towards us  blueshift  hot spot
• Electrons moving away from us  redshift  cold spot
Doppler effect is example of “secondary anisotropy” in CMB

Both effects are sensitive to reionization

The cross-correlation is cleaner!

In analogy to TE correlation of CMB, their cross correlation is
more immune to systematics because errors are uncorrelated
between the two observations
The Effect is Easy to Understand
• Reionization  positive correlation
• Recombination  negative correlation
Probing Reionization History




Cross-correlation peaks when ionized fraction about a half
Sign and amplitude of correlation constrains derivative of
ionized fraction
Typical signal amplitude ~500 (K)2
Above expected error from Square Kilometer Array for ~1
year of observation ~135 (K)2
Our Prediction for SKA


The SKA data should be correlated with CMB,
and WMAP data are good enough!
It is even plausible that the first convincing
evidence for 21-cm from reionization would
come from the cross-correlation signal.

Systematic errors, foregrounds, or unaccounted
noise won’t produce the cross-correlation, but will
produce spurious signal in the auto-correlation.

There are various observational windows to the
universe at z>7 before JWST.



On-going follow-up, deeper surveys with Subaru at
z=7 and VLT at z=8.8 are going to be very interesting!
The excess near infrared background is likely
caused by stars with metals.



Summary
The current data of LAEs do not show
evidence for the end of reionization up to z~7.


Near infrared background
Lyman-alpha emitters
We don’t need metal-free stars, which is a good news.
Future lies in anisotropy: a better prediction is required
for the data from CIBER (launch in 2008)
More ambitious future with 21-cm

The 21-cm data should be correlated with CMB for
a conclusive detection of the cosmic signal.