Millimeter- and submillimeter observations—such as mapping cold dust or molecular spectral lines—are uniquely suited to a range of important topics in astrophysics, cosmology, and astrochemistry. Expanding our knowledge in these fields requires subarcsecond resolution. A telescope of diameter D, observing at a wavelength λ, has a resolution of ~λ/D. Accordingly, an antenna ~2km across would reveal details of ~0.1 arcseconds at a millimeter wavelength. Although this is impractical for a single antenna, interferometer arrays can achieve this resolution because the outputs of multiple antennas separated by distances up to D are electronically combined.
Good imaging requires that signals from all antennas have identical path lengths (≾ λ/20), including electronics, transmission, and atmosphere. Atmospheric path differences have random components with a wide range of spatial and temporal scales, principally due to turbulent distribution of tropospheric water vapor. Commonly, water content is inferred from the millimeter continuum emission or the brightness of the 22 or 183GHz water line.1 A major problem is that the factor to convert the measured brightness fluctuations to variations in the path delay for the signal varies in time depending on atmospheric conditions. Here we discuss an alternative method for directly measuring delay fluctuations.
Antennas in a second atmospheric monitoring array, paired with the science array antennas, continuously observe a point-like source, such as a quasar, to determine the atmospheric delay fluctuations on each baseline (antenna pair) (see Figure 1). This data is then used to correct the signal phases on the science baselines. The group of Asaki2 demonstrated that the relative delay fluctuations measured by a pair of antennas monitoring a satellite signal matched those measured by an adjacent antenna pair on a quasar transiting within a few degrees of the satellite.
Figure 1. Schematic of a baseline in the paired-antenna correction scheme. The observing array (O) tracks its source, and the atmospheric monitoring array (A) a nearby quasar. Δτs1, a1, s2, a2are the turbulence-induced delay variations due to the atmosphere in the signal paths for the science source (s) and atmospheric calibrator (a) to antennas 1 and 2 of the arrays. The geometric delay, τg, is known and corrected. Structure scales such as L that are large compared to the baseline length, B, are rejected by the interferometer since they are common to both antennas. Applying corrections from A to S removes medium-scale structures such as M with scales >b (the spacing between paired antennas), leaving only small-scale turbulence errors (S).
Figure 2. Two-kilometer configuration antenna layout. Each baseline in the observing array (solid) gives one Fourier component of the image. Paired baselines (dashed) measure delay fluctuations for corresponding science baselines. Paired antennas are typically ~20m apart. CARMA: Combined Array for Research in Millimeter-wave Astronomy. SZA: Sunyaev-Zel'dovich Array.
Figure 3. Measured baseline phase on quasar 3C273 at 90GHz. The blue line is the phase observed by CARMA, and the red line is after correction using the delay derived from an SZA measurement on the same source with an adjacent baseline.
At CARMA (Combined Array for Research in Millimeter-wave Astronomy), we have paired the eight 3.5m antennas of the Sunyaev-Zel'dovich Array (SZA) with eight of the 15 CARMA antennas in the extended configurations (~2km). While CARMA observes its science target, the SZA observes a quasar a couple of degrees away. Data are recorded with integration times of 4s to resolve fast fluctuations, and corrections are applied in the data-reduction process.
Figure 4. Application of the paired-antenna correction to the raw data (left), yields a sharpened image (right). Furthermore, the flux recovered from the source is ~70% greater. Jy: Jansky. Deconv: Deconvolved.
The SZA has an 8GHz-wide correlator and sensitive receivers operating at 27–35GHz, where atmospheric transmission is high, allowing sources of a few tenths of a jansky to yield precise delay measurements in a matter of seconds. There are sources of suitable brightness within a few degrees of most parts of the sky. The phase slope across the SZA band reflects the atmospheric delay at any given moment, and this can be applied at the CARMA observing frequency to correct its measured phases. Assumption of a nondispersive atmosphere is accurate enough to use the same delay for millimeter and centimeter frequencies.
In the 2008–09 winter season, antennas were moved into the 2km configuration for the first time (see Figure 2). Simultaneously, the correction scheme was implemented and data analysis developed. Figure 3 presents a typical record of phase on a CARMA baseline before and after applying the atmospheric correction. An astronomical image of quasar 3C111 corrected using atmospheric delays measured on 3C84 is presented in Figure 4.
Using paired antennas for tracking atmospheric delay fluctuations is a promising technique for realizing the resolution potential of current and future millimeter and submillimeter arrays in astronomy. Our initial studies have not only demonstrated significant increases in image quality and flux precision, but are yielding publishable science observations. During the winter of 2009–10 we will continue to develop this technique. Better understanding and characterization of some of the instrumental phases will enable us to separate them from the atmospheric effects. We will also investigate the covariance between the atmospheric and science array fluctuations to optimize the weighting to apply to the corrections, depending on the conditions, as well as the calibration source brightness and distance from the science target.
Support for CARMA construction was derived from the Gordon and Betty Moore Foundation, the Kenneth T. and Eileen L. Norris Foundation, the Associates of the California Institute of Technology, the states of California, Illinois, and Maryland, and the National Science Foundation (NSF). Ongoing CARMA development and operations are supported by the NSF under a cooperative agreement, and by the CARMA partner universities.
James Lamb, David Woody
Owens Valley Radio Observatory
California Institute of Technology
Big Pine, CA
Big Pine, CA
Alberto Bolatto, Peter Teuben, Ashley Zauderer
Department of Astronomy
University of Maryland
College Park, MA
Department of Astronomy and Astrophysics
University of Chicago
California Institute of Technology
Richard Plambeck, Melvyn Wright
Radio Astronomy Lab
University of California at Berkeley
2. Y. Asaki, K. M. Shibata, R. Kawabe, D.-G. Roh, M. Saito, K.-I. Morita, T. Sasao, Phase compensation experiments with the paired antennas method 2. Millimeter-wave fringe correction using centimeter-wave reference, Radio Sci. 33, pp. 1297-1318, 1998.