QUIET: The Q/U ExperimenT

Measuring CMB Polarization

(below is the Columbia contingent of QUIET: Seth Hillbrand, Robert Dumoulin, Laura Newburgh, Professor Amber Miller, and Dr. Will Grainger)
Missing from photo: Ross Williamson, Bryant Rolfe, Liza Volkova

The Science

Before the universe was ~380,000 years old, it was so dense that the photons and baryons had a large scattering cross section and so were essentially coupled together, forming what is known as the photon-baryon fluid. After ~380,000 years, the universe had expanded and cooled to a temperature of ~1 MeV, cool enough for protons and electrons to combine to form hydrogen . This decreased the number of scatterings, which increased the mean free path of the photons, allowing them to begin free streaming throughout the universe. This moment of decoupling (also called recombination) created a so-called surface of last scattering; these photons are detectable now and form the CMB.

Small density fluctuations that seeded later structure formation were present during decoupling and are encoded in the CMB as deviations from the otherwise uniform ~3K blackbody radiation spectrum. These CMB temperature anisotropies, present at the level of one part in 105 , have been measured by the WMAP satellite and its ground-based, balloon, and space-based predecessors. The angular power spectrum of these anisotropies constrains the parameter space for, among other things, the matter and energy budget of the universe and the current rate of expansion.

What caused these small density fluctuations and how did they come to have their characteristic pattern on the sky? The reigning theory is inflation whereby the universe underwent an accelerated expansion period that caused small quantum fluctuations to grow, later forming density perturbations. To test this theory of inflation, experimentalists must look beyond the temperature map and into inflationÕs prediction of the polarized CMB.

The mechanism that generates CMB polarization is Thomson scattering of an anisotropic radiation field, specifically radiation fields that have quadrupolar or higher anisotropies. Although Thomson scattering (essentially scattering between a photon and a free electron) generated polarized light before recombination, the emerging polarized radiation interacted immediately with baryons, effectively washing out any overall polarization. However, during recombination the photons had a chance to free stream without interacting, imprinting the polarization permanently into the CMB radiation. It was not until reionization, when the first stars turned on and began ionizing surrounding gas and the interstellar medium, that there were again enough free electrons to Thomson scatter and polarize the CMB.

The source of these anisotropic radiation fields are density perturbations and inflationary gravitational waves. A quadrupolar (or higher multipole) pattern of over- and under-dense regions creates a velocity field at decoupling and generates polarization of the CMB. In particular, polarization generated by density inhomogeneities will produce a curl-free pattern, and so are called E-modes. Because they have the same source as the temperature anisotropy they should be highly correlated with the previously measured temperature data, and have a well-predicted magnitude and power spectrum shape. Although the E-mode power spectrum does not test inflation, we can still use the signal to break degeneracies in cosmological parameters currently constrained by the CMB intensity spectrum and various astronomical observations such as supernovae, large-scale structure, and lensing surveys. The E-mode polarization signal will also tell us more about the epoch of reionization, such as how old the universe was when the first stars turned on.

While density perturbations generate E-mode polarization, they do not generate a polarization pattern with a curl component on the sky. These B-modes are predicted to be sourced by inflationary gravity waves generated before decoupling, allowing us to see back to a universe just 10-43 years old. Measuring or placing limits on the B-mode polarization can therefore confirm or refute the theory of inflation, and if confirmed will probe the expansion rate of the universe during inflation, thereby setting a characteristic energy scale of an associated inflaton potential.

Primordial gravity waves are not the only source of B-mode polarization. E-mode polarized light, when traveling through a dense region such as a cluster of galaxies, can be gravitationally lensed and pick up a curl component. This additional source of B-mode signal is predicted to dominate the primordial signal on scales smaller than l ~150, with an amplitude determined by the integrated matter density along the line of sight. Because its magnitude depends on the matter distribution in the universe, a measurement of this contribution is an opportunity to probe large-scale structure of the universe.

The polarization signal is very small, currently only the E-mode signal, roughly 100 times fainter than the temperature signal, has been detected, the B-mode power spectrum has never been measured, and the polarization foregrounds are poorly understood. DASI, Boomerang, CBI and Capmap were able to detect E-mode polarization at a few angular scales and WMAP was able to measure the temperature-E-mode cross-correlation signal. To more accurately and completely measure the spectrum and potentially detect the B-mode signature we need an instrument with unprecedented sensitivity and control of systematic errors.

QUIET

QUIET (Q/U Imaging ExperimenT) is a collaboration of 11 institutions, including universities and NASAÕs JPL and Goddard. The experiment will measure the CMB polarization spectrum at two frequencies, 40 and 90 GHz on angular scales of 100< l <2000. Two of our primary science goals are to measure the E-mode spectrum more precisely than any current experiment and to either measure or significantly improve limits on the B-mode spectrum. A third goal is to measure polarized foregrounds; QUIET will be working at frequencies that will allow us to measure synchrotron emission, the dominant foreground at low frequencies. We also plan to measure the amount of contamination in the B-mode signal from gravitationally lensed E-modes.

The minimum detectable signal in a radiometer array of N detectors scales as Tsys*N^-1/2, where Tsys is a number inherent to the radiometer, so we must both have sensitive (low Tsys) detectors and large arrays in order to reach the sensitivity required to meet our science goals. The QUIET polarimeters satisfy the first condition, they have Tsys values of ~25K at 40 GHz and ~50K at 90 GHz. To meet the second requirement, QUIET Phase I will use 91 polarimeters observing at 90 GHz (W band), and 19 polarimeters at 40 GHz (Q band), forming two arrays of comparable sensitivity. We are currently in the process of building both prototype cameras. Here at Columbia, we have designed, built, and tested the cryostat for the W-band camera and are responsible for leading the integration, and deployment of the Q-band camera. QUIET Phase II will involve an order of magnitude more detectors, a larger telescope (7m dish), and will draw heavily from experience gained through work on QUIET Phase I.

Each QUIET camera consists of an array of polarimeters with all necessary optics (telescope, feedhorns, and orthomode transducers), electronics for signal processing, and a cryostat to cryogenically cool the detectors. The optics for Phase I will consist of two 1.4m telescopes, giving us a beam size of ~0.750 and allowing us to measure lower l values (~100). QUIET will use corrugated feeds which have excellent polarization properties. As shown in the diagram above, the next components in our detector chain are the orthomode transducers, which take the light from the feedhorns and split it into two orthogonal polarization states. The polarized output of the OMTs couples to the QUIET polarimeter modules, which output the Q and U Stokes parameters, and the data from the entire array is processed by a series of electronics boards.

Recent technological strides at JPL allow polarimeters to be miniaturized microwave circuits packaged in a small, compact module. These Òpolarimeters on a chipÓ implement monolithic microwave integrated circuits (MMIC) high-electron mobility transistor (HEMT) amplifiers to produce postage-stamp sized coherent polarimeters which can be used in large arrays to significantly enhance sensitivity while minimizing systematic errors. HEMT amplifiers allow simultaneous measurement of both polarization parameters, Q and U in the Stokes basis after the signals pass through identical paths in the polarimeter module. This allows our systematic errors to be very small because any noise or gain fluctuations would be common to both diode signals and will nearly cancel when the signals are differenced.

QUIET will observe in the Atacama Desert in Chile. This is a favorite spot of millimeter and microwave observers as it is one of the driest deserts on earth (useful because water absorbs at microwave frequencies) and is very high, located at an altitude of ~16500 feet. Our QUIET coherent polarimeters will allow us to have extremely well controlled systematic errors resulting in QUIET Phase II arrays of greater sensitivity to E- and B-modes than Planck. We will observe 3Õx3Õ patches on the sky with a projected sensitivity better than 40 microKs^1/2 for QUIET Phase II.

The figure above shows the expected sensitivity for QUIET Phase II. On the right is the standard concordance model power spectrum and on the left includes the primordial signal with a tensor/scalar ratio of 0.18. As shown, QUIET Phase II sensitivity is greater than that of Planck and WMAP, and the sweet spot for detecting B-modes will be at low l values (large angular scales), where the contribution from lensed E-modes is still small.

The above image is a portrait of telescope no. 1, 2007.10.02, IAM Systems, Ohio.
People (left to right): Bill Imbriale (JPL), Colin Baines (Manchester/Jodrell Bank), Lee Bosma (IAM Sys.), Dan Kapner (U. Chicago), Ben Bosma (IAM Sys.), Ross Williamson (Columbia), Bill Griffin (IAM Sys.), Keith Thompson (Stanford)
Photo credit: KLThompson

QUIET Internal Site here.

Official QUIET page here.

Here is an abstract for a QUIET paper by Todd Gaier.