Eiichiro Komatsu

1
Thinking about Fun Stuff from CIBER, Planck,
GLAST, HETDEX, SKA, and (beyond)LISA

• Eiichiro Komatsu
• University of Texas, Austin
• February 23, 2007

2
Lucky Theoretical Cosmologists

• Data-dominated Era
• The most joyful moment for theorists!
• (Some of ) Their own predictions can actually be
  tested by observations within their lifetime.
• Having many predictions is useful for maximizing
  the scientific outcome from (expensive)
  experiments.
• Are we exhausting all the possibilities?
• Are we getting the maximum information out of the
  data?
• Will we know we have surprises in the data when
  we see them?
• Lets make some predictions.

3
Contents (7 minutes per topic)

• Cosmic Near Infrared Background (CIBER)
• Primordial Non-Gaussianity Updates (Planck)
• Dark Matter Annihilation (GLAST)
• Galaxy Power Spectrum (HETDEX)
• 21cm-CMB Correlation (SKA)
• Primordial Gravity Waves (LISA)

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

• New window into 7ltzlt30 (e.g., Lyman-alpha)
• Can we detect photons from early generation
  stars? What can we learn from these photons?
• The signal is (almost) guaranteed, but
  measurement is challenging because of
  contaminations due to
• Zodiacal light, and
• Galaxies at zlt6.

5
Near Infrared Background Current Data vs
Challenges

• Extra-galactic infrared background in J and K
  bands over zodiacal light 70 nW/m2/sr
• These Measurements have been challenged.
• Upper limits from blazar spectra lt14 nW/m2/sr
  (Aharonian et al. 2006)
• Incomplete subtraction of Zodiacal light? 15
  nW/m2/sr (Wright 2001) lt6 nW/m2/sr (Thompson et
  al. 2006)
• Lets be open-minded.
• Clearly we need better data. Better data will
  come from CIBER. What can we predict for the
  outcome of CIBER?

Observed NIRB
Excess
Galaxy Contribution at zlt6
Matsumoto et al. (2005)
6
Previous Study Metal-free Stars, or Mini-quasars?

• First stars?
• Very massive (1000 Msun), metal-free (Z0) 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.

7
Our Prediction Fernandez Komatsu (2006)

• We dont need metal-free stars!
• Dont be too quick to jump into conclusion that
  metal-free, first stars have been seen in the
  NIRB. (Kashlinsky et al. 2005, 2007)
• We dont need anything too exotic.
• Stars contaminated by metals (say, Z1/50 solar)
  can produce nearly the same amount of excess
  light per SFR.
• This is actually a good news we dont expect
  metal-free stars to dominate the near infrared
  background.
• Why? Energy conservation.

8
Robust Calculation
What we measure
Very simple argument Luminosity per volume
(Stellar mass energy) x(Radiation
efficiency) /(Time during which radiation is
emitted)
Can be calculated
Unknown
Radiation Efficiency
9
Stellar data from Schaller et al. (1992)
Schaerer (2002)
10
NIRB Spectrum per SFR
11
The Madau Plot
You dont have to take this seriously for now. We
need better measurements!
12
The Future is in Anisotropy

• Previous model (Kashlinsky et al. 2005 Cooray et
  al. 2006) ignored ionized bubbles.
• We will use the reionization simulation (Iliev et
  al. 2006) to make simulated maps of the NIRB
  anisotropy coming soon!

13
How Do We Test Gaussianity of CMB?
14
Gaussianity vs Flatness (for fun)

• Most people are generally happy that geometry of
  our Universe is flat.
• 1-Wtotal-0.003 (0.013, -0.017) (68 CL) (WMAP
  3yrHST)
• Geometry of our Universe is consistent with being
  flat to 3 accuracy at 95 CL.
• What do we know about Gaussianity?
• For FFGfNLFG2, -54ltfNLlt114 (95 CL) (WMAP 3yr)
• Primordial fluctuations are consistent with being
  Gaussian to 0.001 accuracy at 95 CL.
• Inflation is supported more by Gaussianity of
  primordial fluctuations than by flatness. -)

15
Are We Ready for Planck?

• We need to know the predicted form of statistical
  tools as a function of model parameters to fit
  the data.
• For FFGfNLFG2, there are only three statistical
  tools for which the analytical predictions are
  known
• The angular bispectrum of
• Temperature Komatsu Spergel (2001)
• Polarization Babich Zaldarriaga (2004)
• Joint Analysis Method (TP) Yadav, Komatsu
  Wandelt (2007)
• The angular trispectrum
• Approximate Calculation (TP) Okamoto Hu
  (2002)
• Exact (T) Kogo Komatsu (2006)
• Exact (P) N/A
• Minkowski functionals
• Exact (T) Hikage, Komatsu Matsubara (2006)
• Exact (P) N/A

16
How Do They Look?
Simulated temperature maps from
fNL0
fNL100
fNL5000
fNL1000
17
Is One-point PDF Useful?
Conclusion 1-point PDF is not very useful. (As
far as CMB is concerned.)
A positive fNL yields negatively skewed
temperature anisotropy.
18
Bispectrum Constraints
Komatsu et al. (2003) Spergel et al. (2006)
Creminelli et al. (2006)
(1yr)
(3yr)
19
Trispectrum Not For WMAP, But Perhaps Useful
For Planck
Kogo Komatsu (2006)

• Trispectrum ( fNL2)
• Bispectrum ( fNL)

20
Minkowski Functionals (MFs)
The number of hot spots minus cold spots.
V0surface area
V1 Contour Length
V2 Euler Characteristic
21
MFs from WMAP
Komatsu et al. (2003) Spergel et al. (2006)
Hikage et al. (2007)
(1yr)
(3yr)
Area
Contour Length
Genus
22
Analytical formulae of MFs
Hikage, Komatsu Matsubara (2006)
Perturbative formulae of MFs (Matsubara 2003)
Gaussian term
leading order of Non-Gaussian term
In weakly non-Gaussian fields (s0ltlt1) , the
non-Gaussianity in MFs is characterized by three
skewness parameters S(a).
23
Comparison of analytical formulae with
Non-Gaussian simulations
Hikage et al. (2007)
Surface area
Contour Length
Euler Characteristic
Comparison of MFs between analytical predictions
and non-Gaussian simulations with fNL100 at
different Gaussian smoothing scales, ?s
Simulations are done for WMAP survey mask(Kp0
mask), noise pattern and antenna beam pattern
difference ratio of MFs
Analytical formulae agree with non-Gaussian
simulations very well.
24
Expected 1s errors on fNL from MFs of CMB for
WMAP 8yr and Planck
All
WMAP 8-year and Planck observations should be
sensitive to fNL40 and 20, respectively, at
the 68 confidence level.
25
Big Stuff from Gamma-ray Sky?
26
Dark matter (WIMP) annihilation
GeV-?

• WIMP dark matter annihilates into gamma-ray
  photons.
• WIMP mass is likely around GeVTeV, if WIMP is
  neutralino-like.
• Can GLAST see it?

27
CGB Anisotropy From Dark Matter Annihilation
Ando Komatsu (2006) Ando, Komatsu, Narumoto
Totani (2006)

• Astrophysical sources like blazars and clusters
  of galaxies cannot fully explain the observed CGB
• Only 2550 using the latest blazar luminosity
  function (Narumoto Totani 2006)
• If dark matter annihilation contributes gt30, it
  should be detectable by GLAST in anisotropy.
• A smoking gun for dark matter annihilation
• Energy spectrum of the mean intensity alone wont
  be convincing. We will need anisotropy data.

28
Predicting Angular Power Spectrum

• Angular power spectrum, Cl, is related to the
  spatial power spectrum via Limbers equation.
• We compute the 3D correlation from a halo
  approach
• ST halo mass function,
• NFW density profile in each halo, and
• Substructures included by the HOD method.

Dark matter halo
? ( p / l)
29
A Few Equations
Gamma-ray intensity
Spherical harmonic expansion
Limbers equation
30
Predicted Angular Power Spectrum
Ando, Komatsu, Narumoto Totani (2006)

• At 10 GeV for 2-yr observations of GLAST
• Blazars (red curves) easily discriminated from
  the DM signal.
• Galactic emission (foreground) is small at 10 GeV

31
S/N Somewhat Sensitive to What We Assume For
Substructures
Our Best Guess If dark matter annihilation
contributes gt 30 of the CGB, GLAST should be
able to detect anisotropy.
32
Toward Precision Modeling of Galaxy Power
Spectrum for High-z Galaxy Surveys
Matter Power spectrum
Cosmological Parameters

• HETDEX, WFMOS (z2-4)
• CIP (z3-6)

Method Use 3rd-order Perturbation Theory
Three Key Non-linear Effects

• Unlike CMB, the large-scale structure is pretty
  non-linear.
• The main non-linear effects to account for are
• Nonlinear growth of the density field
  (JeongKomatsu 2006)
• Nonlinear bias (JeongKomatsu, in prep.)
• Nonlinear Redshift space distortion (work in
  progress)

33
3rd order Perturbation theory (PT)

• Equations

• Solving this equation perturbatively up to 3rd
  order in d.
• The 3rd order power spectrum is
• (e.g., SutoSasaki 1991 JainBertschinger 1994)

34
PT Works Very Well!
Jeong Komatsu (2006)
z1,2,3,4,5,6 from top to bottom
Z4
35
Rule of Thumb D2lt0.4
Jeong Komatsu (2006)
Z4
36
Modeling Non-linear BAO
Jeong Komatsu (2006)
37
How About GALAXY Power Spectrum?

• Relation between galaxies and underlying density
• Assumption galaxy formation is a local process
• 3rd-order PT calculation gives the PT galaxy
  power spectrum (Heavens et al. 1998)

38
PT Has Done It Again!
39
BAO Affected by Non-linear Bias
But, now we know how to account for the
non-linear bias.
40
Reionization CMB - 21cm correlationAlvarez,
Komatsu, Dore Shapiro (2006)
21-cm maps result from line-emission
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

41
21cm x CMB Doppler

• 21cm lines
• 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

42
The Effect is Easy to Understand

• Reionization ? positive correlation
• Recombination ? negative correlation

43
21cm Anisotropy

• To get cross-correlation between 21cm and
  Doppler, we need expression for spherical
  harmonic coefficients alm

• To leading order, the anisotropy is dependent on
  fluctuations in density and ionized fraction

44
Doppler Anisotropy

• Doppler arises from integral of velocity along
  line of sight

• Continuity equation ? velocity fluctuation
  proportional to density fluctuation

• We ignore fluctuations of density
  (Ostriker-Vishniac effect) and ionized fraction
  since they are higher order effects

• To leading order, the Doppler anisotropy is
  dependent on fluctuations of velocity ? density

45
Cross-correlation

• Given the coefficients alm for 21cm and Doppler,
  the cross-correlation can be found using

• Shape of angular correlation same as linear power
  spectrum ClP(kl/r)
• Evolution of the peak correlation amplitude
• (at l100) with redshift ? reionization history

46
Cross-correlation

• The shape of the correlation traces the linear
  matter power spectrum at large scales (l100)

47
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

48
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 wont produce the cross-correlation, but
  will produce spurious signal in the
  auto-correlation.

49
Energy-density Spectrum Primordial
Gravitational Usual Cartoon Picture
50
Numerical Solution Traditional
Flat?
51
Primordial Gravity Waves as a Time Machine
in Minkowski spacetime
in FRW spacetime
Cosmological Redshift
Therefore, the gravity wave spectrum is sensitive
to the entire history of cosmic expansion after
inflation.
52
Watanabe Komatsu (2006)
Improving Calculations

• Change in the background expansion law
• Relativistic Degrees of Freedom g(T)
• Radiation Content of the Early Universe
• Neutrino physics
• Neutrino Damping (J. Stewart 1972, Rebhan
  Schwarz 1994, Weinberg 2004, Dicus Repko 2005 )
• Collisionless Damping due to Anisotropic Stress

53
Relativistic Degrees of Freedom g(T)
In the early universe,
WGW
RD
MD
k
RD
g(T)
T, k
54
Relativistic Degrees of Freedom g(T)
Particle Contents rest
mass photon 0 neutrinos 0 e-,
e .51 MeV muon 106
MeV pions 140 MeV gluon
0 u quark 5 MeV d quark 9
MeV s quark 110 MeV c quark
1.3 GeV tauon 1.8 GeV b quark
4.4 GeV W bosons 80 GeV Z boson
91 GeV Higgs boson 114 GeV t quark
174 GeV
QGP P.T. 180MeV
e-,e ann. 510keV
SUSY ? 1TeV
55
Collisionless Damping of GW by Anisotropic Stress
due to Neutrino Free-streaming
Anisotropic stress due to n free-streaming
couples with GWs
Asymptotic solution
35.5 less!
56
Watanabe Komatsu (2006)
The Most Accurate Spectrum of GW in the Standard
Model of Particle Physics
Old Result
57
Features in the Spectrum
n damping
e-,e ann.
QGP P.T.
58
Cosmological Events and Sensitivities
Detector sensitivities
Cosmological events
CMB 10-18 Hz WMAP WGW0 lt 10-11
Plank WGW0 lt 10-13 Pulsar timing
10-8 Hz WGW0 lt 10-8 LISA 10-2 Hz
WGW0 lt 10-11 DECIGO/BBO 0.1 Hz
WGW0 lt ? Adv. LIGO 102 Hz WGW0 lt
10-10
Matter-radiation equality ee- annihilation Neutri
no decoupling QGP phase transition ElectroWeak
P.T. SUSY breaking Reheating (1014 GeV) GUT
scale (1016 GeV) Planck scale (1019 GeV)
59
Summary of Our Predictions

• Cosmic Near Infrared Background (CIBER)
• The signal will not come from metal-free stars,
  but will come primarily from stars with metals.
• Primordial Non-Gaussianity Updates (Planck)
• We are ready for Planck (bispectrum/trispectrum/MF
  s).
• Dark Matter Annihilation (GLAST)
• GLAST should detect DM annihilation if DM is
  neutralino-like and contributes gt30 of the
  gamma-ray background intensity.
• Galaxy Power Spectrum (HETDEX)
• Non-linear bias is important for BAO. We know how
  to handle it.
• 21cm-CMB Correlation (SKA)
• SKA data should be correlated with WMAP data at
  degree scales.
• Primordial Gravity Waves (LISA)
• GW spectrum wont be featureless, but will be
  with full of features.