The Einstein Inflation Probe:

Mission Concept Study

May 12, 2004 Beyond Einstein @ SLAC Gary Hinshaw, NASA/GSFC

 1916: Einstein Shows How Gravity Works

Einstein applied the General Theory of

ALBERT EINSTEIN Relativity to the Universe as a whole:

TIME MAGAZINE • Future collapse or expand forever

“PERSON OF THE • Static Universe (Einstein’s requirement)
CENTURY”

“Einstein's equations didn't have a solution

that described a universe that was
unchanging in time. … he fudged the
equations by adding a term called the
cosmological constant... The repulsive
effect of the cosmological constant would
balance the attractive effect of matter and
allow for a universe that lasts for all time.”

---Stephen Hawking
 The Evolution of the Universe

Early universe remarkably uniform,

Current universe is not…

HOW DID
THIS HAPPEN?

Inflation?... Cyclic model?... (Beyond Einstein)

 Beyond Einstein - The Einstein Probes

• NASA/NSF/DOE are planning a line of 3 “Einstein Probe”

missions:
– A mission to study the dark energy, jointly with DOE, the “Joint
Dark Energy Mission” (JDEM) a.k.a. SNAP.
– A survey mission to find black holes in the nearby universe, the
“Black Hole Finder”.
– A mission to measure the polarization of the CMB to search for
gravity waves from inflation, the “Inflation Probe” or CMBPOL.

• NASA issued an AO soliciting mission concept studies for each

of these missions. In November 2003, several groups were
selected to undertake studies.

• Three groups will study the Inflation Probe mission. Their goals
are to define the mission requirements for sensitivity, sky
coverage, angular resolution, frequency coverage, and the key
experimental technologies.
 The Concept Study Team

• Chuck Bennett (GFSC) • Barth Netterfield (U. Toronto)

• Mark Devlin (U. Penn) • Angelica Oliviera-Costa (U. Penn)
• Dale Fixsen (GSFC) • Lyman Page (Princeton)
• Gary Hinshaw (GSFC, PI) • John Ruhl (Case Western)
• Wayne Hu (U. Chicago) • Uros Seljak (Princeton)
• Kent Irwin (NIST/Boulder) • David Spergel (Princeton)
• Norm Jarosik (Princeton) • Suzanne Staggs (Princeton)
• Alan Kogut (GSFC) • Max Tegmark (U. Penn)
• Arthur Kosowsky (Rutgers) • Bruce Winstein (U. Chicago)
• Michele Limon (GSFC) • Ed Wollack (GSFC)
• Steve Meyer (U. Chicago) • Ned Wright (UCLA)
• Amber Miller (Columbia) • Matias Zaldarriaga (Harvard)
• Harvey Moseley (GSFC) • Cliff Jackson (GSFC)
• Barth Netterfield (U. Toronto)
 Why the Inflation Probe?

• The B mode signal in CMB polarization (at l<100) is produced by

gravity waves left over from inflation. This signal is the only
current observational handle we have on physics at the 1016 GeV
scale -- 12 orders of magnitude beyond the LHC.

• A measurement of the l<100 B mode signal would directly measure

the energy scale of inflation and probe the fluctuation spectrum. It
is also a test for alternate models of inflation (e.g., the cyclic
universe).

• CMB polarization acts as a filter on cosmological processes. It

allows one to probe directly the decoupling epoch, the matter
power spectrum (through gravitational lensing), the ionization
history, and cosmological parameters.
 CMB Polarization

TT - temperature
from scalar and 100
tensor modes TT

TE - temperature ×
polarization
dT = [l(l+1)Cl/2π]1/2 [µK] 10
DASI
covariance TE
DASI EE,
EE – “gradient” 1
2002
polarization from EE
scalar & tensor
modes (r limit→BB)
0.1
BB
BB – “curl”
polarization from r=0.1
tensor modes (gravity waves)
(only) 0.01
BB (lensing) r=0.01

10 100 1000
Multipole l
 E mode & B mode Polarization

• Polarization decomposable Q>0 U=0 Q<0 U=0

into E mode (gradient) and
B mode (curl) components.
• Tensor fluctuations
produce both E and B mode
components.
• Scalar fluctuations produce
only E mode component
(except for transformation
by gravitatiuonal lensing). Q=0 U<0 Q=0 U>0
• B modes directly probe
gravity waves.
 Inflation Probe Sensitivity
• One goal of the mission concept study is to define the mission
requirements for sensitivity. Roughly:
– The amplitude of B mode polarization follows:

∆TBB ~ r1/2 ~ E2infl

– The power in B mode polarization follows:

(∆TBB )2 ~ Ptensor(k)/Pscalar(k) ~ r ~ E4infl

• “Inflationists” anticipate r ~ 0.01
• The current limits are r<0.9 (95% cl) at k=0.002/Mpc (l~10) from
WMAP (Spergel et al.), and r<0.5 (95% cl) at k=0.05/Mpc from
WMAP+SDSSS (Tegmark et al.)
• To reach r~0.01, require >~100 sensitivity of WMAP (see later).
• If the energy scale of inflation is low (r<<0.01), the Inflation Probe
could rule out inflation occurring at the GUT-scale!
 Sensitivity Advances

COBE
1989

×60

WMAP
2001

>×20

Planck ×??
~2007 Inflation Probe
 Detector Development: History

COBE FIRAS
1 pixel, handmade

KAO spectrometer
24 pixels, handmade

KAO, SHARC I
24 pixels, micromachined

SHARC II on CSO: 12x32= 384 pixels

Largest array in operation

SCUBA-2: 4000 pixels in quadrants

Multiplexer behind detectors
 WMAP (Handmade) Microwave Receivers
10 “Differencing Assemblies”
4 @ 94 GHz W-band
2 @ 61 GHz V-band
2 @ 41 GHz Q-band
1 @ 33 GHz Ka-band
1 @ 23 GHz K-band

Jarosik et al. (2003) ApJS, 145, 413

Jarosik et al. (2003) ApJS, 148, 29
 Foreground Emission

• Polarized foreground emission arises from our galaxy. The signal from
our galaxy is currently poorly known, but it is likely comparable to or
larger than the gravity wave signal over most of the sky.
• Foreground emission has both E and B mode symmetry.
• Multiple frequencies are necessary to discriminate CMB emission from
galactic foreground emission. Unlike the temperature case,
modeling and subtracting polarized foreground emission will be
necessary.
• Foreground contamination also results from the conversion of
primordial E mode signal to B mode signal by gravitational lensing.
• If the lensing contamination not cleaned, it sets a lower detection
limit on r of 10-4 at l~100 (the recombination peak) and 10-5 at l~10
(the reionization peak).
 Temperature Foreground Spectra

• WMAP foreground
estimates from 1st year
temperature data (WMAP
observing bands shown in
grey)

• CMB dominates foregrounds

over most of the sky

• Free-free emission is
unpolarized

• Key question: what is

polarization fraction of
foregrounds relative to B-
mode CMB?

• WMAP and other

polarization data will be
very helpful in guiding our
study of foregrounds.

Bennett et al. (2003) ApJS, 148, 97

 Projected Lensing Foreground (Green)

100
dT = [l(l+1)Cl/2π]1/2 [µK]

10

0.1

0.01

10 100 1000
Multipole l
Projected Galactic Foreground (Dust/Synch)

100
dT = [l(l+1)Cl/2π]1/2 [µK]

10

0.1

0.01

10 100 1000
Multipole l
 Angular resolution

• The angular resolution required of the Inflation Probe is

a major topic of the mission concept study.
– High resolution permits more thorough cleaning of the B-mode
signal due to gravitational lensing → better signal to noise at
low l
– High resolution is a major cost driver for a space mission!
– Concept Study must perform a careful trade study of the cost-
benefit of high angular resolution:
• High resolution optical coupling vs. high throughput
• Thermal loading with large optics
• Spacecraft attitude control requirements/costs
• Data rate requirements/costs
 Noise Floor for “Strawman I”

100
dT = [l(l+1)Cl/2π]1/2 [µK]

10

0.1

675 background limited

detectors @ 3 frequencies
0.01

10 100 1000
Multipole l
 Noise Floor for “Strawman II”

100
dT = [l(l+1)Cl/2π]1/2 [µK]

10

0.1

2100 background limited

detectors @ 3 frequencies
0.01

10 100 1000
Multipole l
 Control of Systematic Errors

• Control of systematic errors is critical to the success of any CMB

polarization measurement.

• A key element in rejecting systematic errors is to modulate the

signal on many different time scales. The efficiency of different
modulation schemes must be assessed.

• It is crucial to verify in flight that systematic effects have been

reduced to acceptable levels.

• The most sensitive detectors (e.g. bolometers) are power

detectors. Any effect that leads to a difference in power can be
confused with a polarization signal, e.g. bandpass mismatch, far
side-lobe pick-up, higher order beam effects, etc.

• The gain of the system has to be stable on the time scale that one
can measure it in flight.
 Sky Coverage and Scan Strategy

• The maximum S/N for B modes is at l~5. To reliably measure l<10

(to really be sure we are seeing B modes) nearly full sky coverage
is required.

• Depending on the degree to which the lensing and galactic

foreground may be cleaned, smaller patches might yield a
detection of gravity waves.

• The scan pattern used to modulate the signal and achieve sky
coverage will be critical. Detailed mission simulations of
different scan modes, coupled with realistic instrument models
will be used to assess scan patterns and other experimental
approaches.

• The decomposition of the polarization signal into E and B modes

is sensitive to sky coverage and correlations between data
points.
WMAP Scan Pattern as Example

Bennett et al. (2003) ApJ, 583, 1

 Do We Need a Satellite?

Only in space can one achieve the stability and vantage

needed to probe large angular scales (>1˚)…
This is the range where the telltale remnants of
inflation are most likely to be found…
Therefore… yes!
 Inflation Probe Concept Study - Summary

• Study period is two years.

• Starting in June 2004, study teams will participate a

joint NASA/NSF/DOE task force to chart the next steps
CMB polarization studies.

• The obstacles to detecting gravity waves are sensitivity,

systematic errors, and astrophysical foregrounds. With
a well designed mission, these obstacles can be
surmounted. The goal of directly measuring the energy
scale of inflation is within sight!
 The 6 Phases of a Project

1. Enthusiasm! (Now)
2. Disillusionment…
3. Panic!!
4. Search for the Guilty……
5. Punishment of the Innocent!!!
6. Praise and Honors for the Non-Participants 
The End
 Experimental Approach:
Minimize Systematic Measurement Errors
• Differential design to minimize systematic errors
• 5 microwave frequencies to understand foregrounds
• 20 radiometers to allow multiple cross checks
• Sensitivity to polarization
• Accurate calibration (<0.5%)
 calibration using modulation of the dipole from Earth’s velocity
• In-flight beam measurements on Jupiter
• Minimize sidelobes & diffracted signals from Earth, Sun, Moon
 L2 orbit
• Multiple modulation periods to identify systematic effects
• Minimize all observatory changes SWITCH
(0.4 msec)
 L2 orbit; constant survey mode operations
• Rapid and complex sky scan ORBITAL
(1 year)
SPIN
(2 min)
observe 30% of the sky in an hour
PRECESS
(1 hr)

SPIN-SYNCHRONOUS NON-SKY SIGNALS WERE THE LEADING CONCERN

Bennett et al. (2003) ApJ, 583, 1
 Systematic Error Cross-Checks
(Q1+Q2)/2 (Q1-Q2)/2

(V1+V2)/2 (V1-V2)/2

(W12+W34)/2 (W12-W34)/2

Hinshaw et al. (2003) ApJS, 148, 63

Temp x E-Polarization Power Spectrum

Reionization
z r  2010
9
tr  180220
80
million years
uniformly suppress l>40 anisotropy by 30% (!)

from z=1089
scattering of CMB from electrons
with non-random velocities

Kogut et al. (2003) ApJS, 148, 161

Bennett et al. (2003) ApJS, 148, 1
http://lambda.gsfc.nasa.gov
 Temperature-Polarization Correlation

Temperature quadrupole at Radial pattern around cold spots

z~1089 generates polarization Tangential pattern around hot spots
 Jupiter Beam Maps

A-side B-side

Measured in flight Modeled

 CMB Polarization

1. z=20 reionization:
scattering of CMB from free electrons
uniformly suppress l>40 anisotropy by 30% (!)
Now detected: z  2010 tr  180220 million years
r 9 80

2. z=1089 decoupling:
scattering of CMB from electrons with non-random velocities
polarization correlates with temperature map
1st detected by DASI, now have power spectrum

3. Gravity waves:
Inflation-generated gravity waves polarize CMB
need new mission (e.g., NASA’s “Einstein Inflation Probe”
 Sky Coverage

Not to scale:
Earth — L2 distance is
22.5° half-angle 1% of Sun — Earth Distance
1 hour precession cone
3 Months
A-side line of site 129 sec. (0.464rpm) Spin

MAP at L2 B-side line of site

6 Months -
Earth
full sky coverage 1 Day

Sun

MAP990159
Designed by
M. Pospieszalski
at NRAO
 Lay of the Land
Temperature (T)
from scalar and
tensor fluctuations
E polarization
from scalars and B modes
tensors from
lensing of
Current limit on
E modes.
tensors

B polarization from
tensors (gravity
waves) only

Reionization peak Recombination peak Gravity waves decay

(z=20) (z=1089) inside the horizon.
1948: Big Bang Theory

• Universe began unimaginably hot

and dense - billions of years ago

• Expanded and cooled

• Predictions:
• A microwave afterglow light
from the Big Bang
• Specific spectrum (intensity with
wavelength)
• Many other predictions - all
predictions since verified
 1965: Discovery of Afterglow Light from Big Bang

Arno
Full sky image, green represents the afterglow Penzias

Measurement Receiver
Bell Telephone Labs
New Jersey

1978 Nobel Prize

in Physics

Robert
Wilson
 Beyond Einstein – Inflationary Universe

• One slide summary of inflation.