UCSB WMRS Cosmology
Millimeter Wave Cosmology
by Phil Lubin and Peter Meinhold
Measurements of the spectrum and spatial anisotropy of the 2.7K Cosmic Microwave Background Radiation (CMB) have provided major advances in our understanding of the early universe. Attempts to measure the polarization anisotropy of the CMB have so far led only to upper limits and its discovery would be a major milestone in cosmology. Measurements of the amplitude and spatial distribution of CMB polarization, along with its correlation with the temperature anisotropy would provide profound information about the structure, evolution and ionization history of the universe. In addition, polarization measurements will enable us to break degeneracies in determining the critical cosmological parameters from CMB data from measured CMB temperature anisotropy power spectra.
We propose a progressive series of experiments at the WMRS Barcroft station and Dome designed to demonstrate the quality of the site for millimeter wavelength Cosmology and Astrophysical studies. We plan on installing two of our off axis 2.2 meter diameter millimeter wave telescopes this spring and summer and then following this next year with an on-axis 2.6 meter Cassegrain telescope. We will use the abandoned Barcroft dome, once used as in the Keck telescope siting survey. That survey determined that Barcroft is an excellent observing site, with typically ˝ the precipitable water vapor of Mauna Kea. We will take our instruments (one complete, the other under construction) and install one telescope in the Barcroft dome and one in an enclosure at Barcroft Station. WMRS is already supporting the infrastructure for this by installing power in the dome and a 1 Mbs data link.
Installation of the first telescope (called the Background Emission Anisotropy Scanning Telescope or BEAST) is already well underway: most of the telescope frame and associated components were transported to Barcroft in late fall, 2000. Over the winter, WMRS personnel and UCSB cosmology associates modified an unused building at that site to allow the roof to be rolled open for observations. We will be installing the detector and starting observations within the next month (April, 2001). The second telescope and detector are being assembled at UCSB and will be installed in the Barcroft dome during the summer. We anticipate year round observations with both telescopes.
We are making a major push to turn the Barcroft Observatory into a millimeter wave site for astrophysics and cosmology. Our instruments are leveraged heavily from existing programs, hardware, and collaborations. CMB polarization measurements will continue with a series of different detectors over at least the next 5 years. The initial investment in the instrument will also result in other target science in the future, such as mapping point sources in conjunction with the MAP and PLANCK satellite projects, measuring SZ sources, making deep anisotropy field measurements and possible far IR studies in the future. In the longer term we are investigating installation of a 5-7 meter class mm wavelength telescope dedicated to cosmological observations including small scale anisotropy and polarization, Sunyaev-Zeldovich cluster measurements and point source surveys.
The predicted polarization amplitude of the CMB is extremely small: about one-part in 106 of the CMB intensity, or D T@ 5m K (micro Kelvin) Detecting this signal is an extremely challenging task, requiring high sensitivity detectors and precise control of systematic effects. Using new receiver technology we expect to reach this signal level, an order of magnitude below present upper limits. A detection of the polarization of the CMB would be a major discovery and will help:
Polarization of the CMB is always enhanced in models which predict early reionization, while the effect of reionization on the temperature anisotropy is primarily an attenuation of the signal at small scales. For reasonable non-standard models, the amplitude of polarization on 10° angular scales is on the level of 10% of the temperature anisotropy, while for the standard model of recombination the corresponding polarization level does not exceed 1% (Keating et al. 1998). The amplitude of CMB polarization on large scales tells us the epoch of reionization.
Polarization anisotropy measurements complement temperature anisotropy measurements in an important way. Both the polarization and temperature power spectra depend sensitively on key cosmological parameters, such as the density of the universe, the Hubble constant, and the power spectrum of the initial mass-density inhomogeneities that may have been created by inflation (Hu, Sugiyama, and Silk 1997). Several authors have shown that temperature anisotropy measurements can be used to determine several parameters to an accuracy level of a few percent. However, many combinations of these parameters give nearly degenerate temperature power spectra. Therefore, temperature anisotropy data alone cannot be used to convert the measured power spectra into the underlying cosmological parameters. Polarization measurements can break this degeneracy (Zaldarriaga 1997), and in particular can discriminate between non-standard models, with early reionization, and the standard model (Bond and Efstathiou 1984,1987; Basko and Polnarev 1980; Nasel'skii and Polnarev 1987; Ng and Ng 1995; Zaldarriaga and Harrari 1995; Crittenden, Davis, and Steinhardt 1993; Frewin, Polnarev, and Coles 1994).
Seeing: The UC White Mountain Research Station has the potential to be one of the premier high altitude observing sites in the world. It was seriously considered for siting of the Keck Telescope, and has infrared seeing that during the site survey was better than Mauna Kea in precipitable water vapor (about ˝ of Mauna Kea). Mauna Kea and the Barcroft Observatory are about the same altitude but the latitude of Barcroft is +37 versus +19 for Mauna Kea. For deep polarization scans of the North Celestial Pole (NCP), the latitude of Barcroft is far superior, as the NCP is 37 degrees above the horizon versus only 19 degrees for Mauna Kea. The reason White Mountain was not chosen for the Keck telescope project was due to winter weather concerns and accessibility. While a problem for a large user based telescope facility like Keck, it is not for us. We have observed from the South Pole in three seasons from 1988 to 1994 and based on the evidence to date, we expect White Mountain to allow comparable quality mm observations. If our observations confirm expectations, other instruments will be set there to take advantage of this site. We are already working on two other instruments that could utilize this site. One is a large format bolometer array receiver and the other is a large format HEMT array receiver. Starting with our instrument, already underway with hardware from previous programs, is an ideal way to begin. Ultimately, this will open up a new mm and sub-mm site for general use. Successful observations and development of the infrastructure of the site (such as the internet link and large observing pad) will encourage others to join with us to develop new instruments and observing programs.
Site Plans: We will start this season observing from two sites, one at the abandoned Barcroft observatory dome and the other using a modified building near the Barcroft labs. The Barcroft observatory dome is located approximately 2 Km from the Barcroft labs and is on a flat plateau that will also be the site of our new expansion facility next year. The Barcroft dome is shown in Figure 1 and a sketch of the 2.2 meter polarimeter which will be housed in it is shown in Figure 2. This instrument will be used to make a deep scan of the NCP region to search for polarization using a W band HEMT polarimeter developed with other funds. We will use a spare set of the BEAST (Background Emission Anisotropy Scanning Telescope a recently develop mm wave balloon borne telescope) program 2.2 meter CFRP optics and an existing rotary table from a previous South Pole experiment.
The second instrument and the one to begin taking data first, is a modification of the BEAST instrument that recently completed two balloon flights. It is being retrofitted for use on White Mountain with a cryocooler and new elevation mount and will be sited in a building near the Barcroft labs. This building is shown in Figure 3. A sketch of the BEAST instrument installed in it is shown in Figure 4. The building has been modified by making the roof "roll off".
Observations Year 1
We will start by taking two instruments to White Mountain for observing this season: one to measure anisotropy (BEAST), and one to measure polarization (Q band (40 GHz) and W band (90 GHz) polarimeters). The reason we are taking these instruments is that they exist (or will shortly), they offer excellent science return and that they will allow use to rapidly measure and characterize the atmospheric stability from 25 to 100 GHz.
The BEAST Instrument: BEAST consists of an array of cryogenic High Electron Mobility Transistor (HEMT) receivers. There are two receivers at 30 GHz and six at 40 GHz feeding a 2.2 meter off axis Gregorian telescope specifically designed for CMB experiments with a large focal plane and very low sidelobes. The receivers have noise figures close to 20 K with effective sensitivity in the white noise limit of less than 400 m K-s1/2 including the expected atmospheric component of 2.7 K @ 30 GHz and 6.3 K @ 40 GHz. The beams form a near circular pattern on the sky. With the 8 beams we will be able to model the atmosphere on a variety of scales and to test the ability to regress out the atmospheric fluctuations. Our experience at the South Pole with single beam experiments at 15, 25, 30, 40 and 90 GHz was that the atmosphere did introduce excess noise but we would expect with multiple beams to be able to regress this out. In the future we have to option of adding in situ water line analysis if needed. This would give realtime active monitoring and possible removal of the dominant fluctuating component, namely water.
We will operate this system in an automated mode by observing about the vertical and letting the earths rotation sweep out the sky. In this manner we will make a map of about 3000 square degrees with 40 m K error per 20 arc minute pixel in one months equivalent observing. Here we have conservatively estimated the error using only one horn. We could expect an improvement of a factor of 2.5 by averaging all the horns, giving a statistical error of 17 m K. A 40 m K error is comparable to the expected 2 year MAP satellite error per 20 arc minute pixel. A years worth of observing should return at least 6 months of good data and reduce these errors by another factor of 2.5 or 7 m K per pixel . Perhaps more importantly for this proposal this measurement will allow us to also characterize the atmosphere over a long time scale and work out the logistical issues associated with the site. Beast has just completed two balloon flights and is very well adapted to harsh environments.
W Band Polarimeter: We have nearly completed an 80-100 GHz (W band) polarimeter which will eventually be used with our 2.6 meter on axis Cassegrain system (the COMPASS experiment) currently fielded at U. Wisconsin. The Wisconsin site is at sea level and has high atmospheric emission at W band (about 60 K at best near the NCP). This will double the noise figure of the instrument (and at least quadruple the observing time). The atmospheric stability is of even greater concern. The observing season in Wisconsin is also limited to about 4 months in the winter of which only a limited number of days are good.
By contrast, from White Mountain, the atmospheric emission at 90 GHz is about 12K near the NCP (7K vertical), should be more stable, and the number of observing days should be much greater. As a result, we plan to install the polarimeter on a separate telescope using a copy of the BEAST 2.2 meter optics without the chopping flat mirror.
As discussed earlier this will go in the Barcroft observatory dome. The instrument is a cryogenic HEMT correlation receiver that continuously takes the difference in temperature between two orthogonal polarization states. A schematic is shown in Figure 5. The HEMT amplifiers are the same InP type we are developing for the PLANCK satellite program. Much of the technology for this receiver comes directly from this program. Data from these measurements will be crucial in assessing the viability of large format array based on the same technology. We also have most of the parts to put together additional Ka and Q band polarimeters using pieces from the BEAST program.
The COMPASS Experiment: COMPASS is a collaboration between UCSB, U. Wisconsin and U. Rome to field a 2.6 meter on axis Cassegrain telescope and a Ka band polarimeter. We recently shipped the telescope to U. Wisconsin to begin testing there with the Ka band polarimeter. If our results from White Mountain using BEAST optics are encouraging we will move this telescope to White Mountain in summer 2001 or 2002 depending on the timing of the construction of the new building at Barcroft observatory as the 2.6 meter dish is too large to fit onto the existing dome. This is another example of the leverage we will obtain once we get the WM site operational as a mm wave observing "facility".
HEMT Polarimeter Array: The first W band polarimeter will be a single horn, but we plan to add additional polarization channels and simultaneous anisotropy channels next season. Beyond next season we plan to have a large format HEMT based polarimeter array receivers. This will operate at Q and W bands, and include polarization and anisotropy channels. The large format array is not funded by this proposal but will form the core of an NSF proposal.
Bolometer Camera: We are working on a 19 element bolometer camera for use with the 2.2 meter BEAST optics. This instrument is designed to make deep anisotropy and point source measurements in the 100-300 GHz range. The detector is a "plug in" replacement for the W band polarimeter described above. A variant of this using polarization sensitive bolometer we are building at UCSB is being prototyped and would also fit directly into the focal plane space of the 2.2 meter optics in the dome. As with the large format HEMT polarimeter, this work will be part of a separate proposal . We thus immediately see several instruments being used in the dome.
Sensitivity Calculations: Our current W band HEMT single pixel polarimeter should achieve approximately 1200 micro Kelvin error in one second. Scaling from this, and using our planned strategy of making a map of polarized emission over the NCP of radius 2.25 degrees, we get 400 12 arc-minute pixels at a sensitivity of 9 micro K per pixel in three months of integration time. This should be doable in a winter season. This is for one horn. Our array will have 16 horns. With this arrangement, to first order we get 4 times better sensitivity or about 2 micro Kelvin on 400 hundred pixels or 4 micro Kelvin sensitivity on 1600 pixels. Such a map is expected to have profound cosmological information.
For the bolometers, arrays are cost effective after a single pixel is realized. Noise from our bolometric system is expected to be roughly 300 micro K in one second, when operated as a polarimeter at Barcroft. This takes into account the atmospheric and optics photon loading and making conservative assumptions about the filters, bolometers and optics currently available in our lab. Given this, the same calculation as above would yield 1 micro K sensitivity on 1600 pixels or 4 micro Kelvin on 26,000 pixels. This would be a revolutionary survey.
Detector sensitivity is not the whole concern. Careful understanding, control and measurement of systematics are critical. One of the best arguments for measuring with both bolometers and HEMTs is that the systematics concerns are very different and a measurement with both would be convincing.
Comparison to satellites and balloons
The MAP satellite is due to launch in 2001 with data expected in 2002. The goal of the MAP satellite is to make a full sky map at a sensitivity of 35 micro Kelvin per pixel in intensity at an angular bin of 20 arc minutes. It is not optimized to do polarization measurements and indeed the sensitivity is poor enough that it is not expected to detect any on a pixel basis. The best MAP can hop for is a measurement of the polarization temperature correlation measurement. The instrument we propose here is designed and optimized specifically to measure polarization and to do so at 10-20 times higher sensitivity on smaller portions of the sky than MAP. The two approaches are highly complementary.
The PLANCK satellite, that UCSB is heavily involved in, is also not designed specifically to measure polarization and is not due to launch until 2007 with data in 2008-10. We have been trying to figure out how best to optimize PLANCK to measure polarization. If we are successful in measuring polarization with PLANCK (and we hope we are), we will not have this data for nearly a decade. At best PLANCK will have sensitivity comparable or somewhat worse than we propose BUT on the whole sky.
UCSB has been involved in balloon borne studies of the CMB since 1987. Due to the limited exposure time of balloons, even long duration balloons will not have sensitivity compared to what we propose. Based on our previous experience with CMB polarization experiments (Lubin, 1981), ground based observations from a good site should suffice in most circumstances for polarization measurements
PARTIAL LIST OF COLLABORATORS
UCSB grad students:
UCSB Lab Staff:
Brazilian Collaborators from INPE:
Basko, M. M. and Polnarev, A.G. "Polarization and
Anisotropy of the Relic Radiation in an Anisotropic Universe". Mon.
Not. R. Astr. Soc. 191. 207. 1980.
Bond, J.R. and Efstathiou, G. "The Statistics of Cosmic Backgrounf Radiation Fluctuations". Mon. Not. R. Astr. Soc. 226. 655. 1987.
Crittenden, R. G., Davis, R., & Steinhardt, P. "Polarization of the Microwave Background Due To Primordial Gravitational Waves". Ap. J. Lett.. 417. L13. 1993.
Frewin, R.., Polnarev, A.., & Coles, P., "Gravitational waves and the polarization of the cosmic microwave backround".. Mon. Not. R. Astr. Soc. 266. L21. 1994.
Gaier,T, PhD. Thesis, UC Santa Barbara, 1993.
Hu, W., Sugiyama, N., and Silk, J. "The Physics of Microwave Background Anisotropies," Nature. 386. 37. 1997.
Lubin, P.M. and Smoot, G. F. "Polarization of the Cosmic Microwave Background Radiation". Ap. J. 245. 1. 1981.
Nasel'skii, P., & Polnarev, A.. Astrofizika. 26. 327. 1987.
Ng, K.L., & Ng, K. W. "Large-Scale Polarization of the Cosmic Microwave Background Radiation". Phys. Rev. D. 51. 364. 1995.
Zaldarriaga, M. "CMB Polarization Experiments". Pre-print Astro-Ph 9709271. 1997.
Zaldarriaga, M., & Harari, D. "Analytic Approach to the Polarization of the Cosmic Microwave Background in Flat and Open Universes".
Our group has several ambitious goals for the WMRS Barcroft station: Acquisition of CMBR (Cosmic Microwave Background Radiation) anisotropy data from the BEAST telescope; installation and operation of a sensitive CMBR Polarization experiment in the Barcroft dome; planning for and securing funding for a larger, more aggressive CMBR polarization telescope (the Densely Instrumented MMIC Correlation Array Module or ‘DIMCAM’). In addition, we are doing extensive work preparing the site for these instruments in terms of infrastructure (communications, power handling (UPS, surge protection, power distribution), 24/7/365 automated weather monitoring, and physical facilities). We made major progress in all these areas during 2001.
CMBR anisotropy and CMBR polarization anisotropy are probes of processes and structure in the early universe. CMBR Anisotropy measurements constrain the processes by which structure could have formed, the nature and distribution of matter in the universe, and the small number of cosmological parameters which describe the universe as a whole, including the Hubble constant, the cosmological constant and the average density of the universe. CMBR anisotropies may be polarized by several processes: measuring the polarized component of CMBR anisotropies should allow us to separate out otherwise degenerate models, as well as distinguish the characteristic imprint of primordial gravitational waves on the structure of the early universe.
Starting in late 2000 and early in 2001 (January), we brought the main components of the BEAST (Background Emission Anisotropy Scanning Telescope) to Barcroft. Through the winter season and into the early summer we made many trips to Barcroft so that by July, the following were accomplished:
Garage: We modified one of two available Barcroft garage structures by cutting a portion of the roof, raising it and mounting it on rails to allow it to roll open. This allows the telescope to view the sky without being moved out of the building. The roof is actuated with an industrial garage door opener, and has been working in all seasons at Barcroft (including during a trip in February 2002). Much of the success of this work rested on the ingenuity, skill and hard work of WMRS staff (Rick Masters and Dave Trydahl).
Power Handling: One of the challenges of working at Barcroft turned out to be power line problems. We lost several computers and some data acquisition equipment to power line surges before we acquired and installed appropriate power conditioning hardware, including ferro-resonant transformers and high capacity UPSs (Uninterruptible Power System). We have had no problems since installing these systems. In addition to this we contracted an electrician/carpenter to install new breaker panels and appropriately sized circuits in both garages and the dome.
Telescope: The BEAST telescope was originally designed to make measurements from a high altitude balloon, for periods of up to 10 days. We made extensive modifications to the system to allow extended observations from the ground and assembled the telescope at Barcroft in the garage. By late July we had the system operational, with full time data taking commencing in early August. To date we have around 200 Gbytes of Cosmology data from Barcroft and we are in the process of analyzing it.
Weather and site monitoring: During the summer we installed a weather station, downloading weather data by hand during normal operations of the telescope. During our recent trip to Barcroft (Feb, 2002) we set the system up to automatically send data to UCSB via the STARBAND satellite internet link. At the same time we installed a webcam for visual record of the site conditions when personnel cannot be there. (http://moseisley.deepspace.ucsb.edu ). In addition to this we were able to borrow a standard NRAO ‘tipper’, an automated instrument for measuring atmospheric opacity. This standardized instrument allows direct comparisons among millimeterwave/infrared observing sites. We operated the tipper from early September thru late October. We were fortunate to have the tipper loaned to us by A. Beasley of the OVRO, and we hope to borrow that one or another one for longer term measurements of the site. For the time period measured, Barcroft was comparable to Mauna Kea, considered one of the best IR observing sites in the world.
Dome Refurbishment: We have been working to upgrade the dome facility in preparation for bringing the first generation polarimeter telescope to Barcroft. This includes upgrading the power systems as describe for the BEAST garage, strengthening the telescope support pedestal, replacing weatherstripping, insulating the instrument room, and upgrading the dome control system to allow computer control of the door and azimuth. This work will be completed in the early part of the Summer, 2002 , in time to install the telescope.
Polarimeter: The instrument being readied for the dome consists of three (3) receivers: one ‘delta T’ at 90 GHz, one polarimeter at 40 GHz, and one polarimeter at 90 GHz. The receivers are cooled to 20 Kelvin with a mechanical cooler and coupled to a second copy of the BEAST telescope optics. This instrument will be dedicated to making a small but very sensitive map of CMBR polarization fluctuations near the North Celestial Pole. The receivers are ready and being integrated to their electronics, while the telescope sytem is being integrated separately. All elements should be tested together in Santa Barbara by mid May, ready for Barcroft sometime in June.
Polarimeter Recievers and Polarimeter Telescope Frame and Rotation Stage.
We made major progress towards making Barcroft usable by our students by having one of the ‘garages’ refurbished into a comfortable laboratory. This provides a controlled comfortable environment for the students, most of the computers and test equipment, even in the harshest conditions at Barcroft in February. WMRS staff also helped significantly in setting up this lab.
DIMCAM is our name for a large format array of correlation Polarimeters, specially designed detectors for measuring CMBR polarization. We submitted two major proposals the the National Science Foundation to support this specific project development, decisions will be made later this year. In the meantime, design work and planning continues, assuming a good outcome for the proposal.
Reliable communications with reasonable bandwidth are critical to the success of our program, from simple interactions debugging instruments over the phone, downloading test data during observations, remote monitoring of the site for weather and atmospheric conditions, all the way to potential remote operation of the telescope via the web. Three systems brought up to Barcroft are making a huge difference to how we can work there:
Our efforts at Barcroft have been funded from a variety of sources, all of which have been very important to our success to date. In addition we have applied for significant funds from NSF to support future work. We have made a summary of the totals in the other part of this report, here we outline how the different sources contribute.
NASA Balloon program
Our NASA funds have been the primary support up to the start of our work at White Mountain. We developed and flew the BEAST telescope on NASA grants. The spare optics and much of the instrument infrastructure for the polarimter depend heavily on our previous work on the NASA payload. Without all this investment there would be no telescopes to site at Barcroft.
NASA/ESA Planck Surveyor
Planck is a second generation CMBR space mission we are involved in. While NO Planck funds have been spent on WMRS work, there has been very significant technology transfer. The receivers being used for the polarimeter were developed with JPL collaborators (T. Gaier, M. Seiffert) based very directly on the Planck receivers, and using the same devices from TRW. These are state of the art receivers with state of the art devices and performance, and could only be available to our program through our extensive collaboration with JPL on the Planck and BEAST programs.
NSF Extragalactic- COMPASS
Compass is a joint venture with the University of Wisconsin, Madison and the University of Miami. It has been funded by NSF extragalactic astronomy for several years, looking for polarization anisotropy in the CMBR from a site in Wisconsin. The collaboration plans to move the operation and future instruments to Barcroft due to the superior observing conditions. The 3 element polarimeter described above is part of this collaboration.
NSF Extragalactic,NSF MRI-DIMCAM
These two proposals to the NSF include our COMPASS collaborators and detail the large format polarimter array we plan to install at Barcroft.
Brazilian Cosmology Center
We have an active collaboration with researchers at INPE (the Brazilian Space Agency). For the past year they have been providing significant amounts of support in the form of hardware contributions, totaling $100,000. These contributions range from very sensitive angle resolvers to state of the art custom electroformed microwave feedhorns.
All of the work described above could not have been accomplished without the enthusiastic support of the WMRS staff and management. Both Dave Trydahl and Mike Morrison (outgoing and new managers respectively) made sure we had what we needed to accomplish our work. Rick Masters has been tremendously helpful with advice, winter transport and logistical support as well as his work on the buildings. J. Szewczak’s work on the internet link has been much appreciated, and we look forward to taking full advantage of it. In addition all the OVL and Barcroft summer staff were enthusiastic and helpful, with normal logistics and helping hands for the students and staff. We’re all looking forward to a long and productive collaboration.