|Annu. Rev. Astron. Astrophys. 2014. 54:
Copyright © 2014 by Annual Reviews. All rights reserved
For centuries astrometry was the primary focus of astronomy. Before the 17th century, astronomers charted the locations of naked-eye stars with an accuracy of a fraction of an arcminute. With this level of accuracy, all but a handful of the highest proper motion stars remain "fixed" for an astronomer's lifetime, and it was not until Galileo's introduction of the telescope in astronomy that arcsecond precision became a possibility.
For millennia two great questions remained unanswered: was the Earth at the center of the Universe and how large was the Universe? Both questions could be answered through astrometry by measuring the parallax of stars. If the Earth revolved around the Sun, stars would exhibit yearly shifts in apparent position (annual parallax) and, if measured, these would indicate their distances (at that time equivalent to the "size of the Universe"). This scientific opportunity led most scientists of the period to attempt astrometric measurements to detect parallax (Hirshfeld 2001).
Early attempts to measure parallax by Robert Hooke, and later James Bradley, compared the tilt of a near-vertical telescope to that of a plumb-line as the star γ Draconis transited directly overhead in London. Bradley, after discovering and accounting for aberration of light (a yearly ± 20arcsec effect caused by the Earth's orbital motion), was only able to place an upper limit of 0.5 arcsec on the annual parallax of the star. It was not until 1838 that Friedrich Wilhelm Bessel presented the first convincing measurement of stellar parallax: 0.314 arcsec for the star 61 Cygni (chosen as a candidate by its large proper motion of ≈ 5 arcsec yr-1).
Bessel and others followed a suggestion attributed to Galileo to measure relative parallax, the differential position shift of a nearby target star relative to a distant background star, rather than absolute position shifts. This approach still forms the basis of the highest accuracy measurements made today. Currently astrometric accuracy using Very Long Baseline Interferometry (VLBI) at centimeter wavelengths is approaching the ~ 1 μas level. This review documents the state of the art in radio astrometry, focusing both on the techniques and the scientific results.
1.2. Radio Astrometry
Karl Jansky made the first astronomical observations at radio wavelengths in the 1920s. While searching for the source of interference in trans-atlantic telephone calls (mostly from lightning in the tropics), he noted strong emission localized within a few degrees in the constellation of Sagittarius. The peak of the signal drifted in time at a sidereal rate, indicating an origin outside the Solar System, and was ultimately traced to energetic (synchrotron emitting) electrons throughout the Milky Way, but concentrated toward the center.
The development of radar during WWII, gave a boost to radio astronomy in the 1940s and 1950s. Astrometric precision with a single radio telescope was generally limited to some fraction of the diffraction limit, θd ≈ λ/D, where λ is the observing wavelength and D is the antenna diameter. Even though radio antennas were an order of magnitude larger than optical telescopes, the roughly four orders of magnitude longer wavelength seemed to doom radio astrometry. However, that changed with the development of radio interferometry. In the 1960s, interferometers with baselines of ~ 1 km localized the positions of quasi-stellar objects (QSOs), leading to the discovery of highly redshifted optical emission. Over the years, with increasing baseline length, positional accuracy with (connected-element) interferometers improved from ~ 1 to ~ 0.03 arcseconds (Wade & Johnston 1977).
In the late 1960s, radio interferometry was greatly extended by removing the need for a direct connection (either with cables or microwave links) between antennas. Signals were accurately time tagged using independent atomic clocks and recorded on magnetic tape, allowing separations of interferometer elements across the Earth. This technique, called Very Long Baseline Interferometry (VLBI), lead to many discoveries, including super-luminal motion in jets from active galactic nuclei (AGN) and upper limits of ~ 1 pc on the size of the emitting regions. Both results provided strong evidence for super-massive black holes as the engines for AGN (Reid 2009). Early VLBI observations with intrinsic angular resolution better than 1 mas offered absolute position accuracy of ~ 0.3 mas using group-delay observables (Clark et al. 1976, Ma et al. 1986) and relative astrometric accuracy between fortuitously close pairings of QSOs of ~ 10 μas using phase-delay information (Marcaide et al. 1985).
Starting in the 2000s, calibration techniques improved to the point where relative positional accuracy of ~ 10 μas could be routinely achieved for most bright targets, relative to a detectable QSO usually within a couple of degrees on the sky. This changed the game dramatically. Such astrometric accuracy is unsurpassed in astronomy and is comparable to, or better than, the target accuracy of the next European astrometric space mission: Gaia (Bourda et al. 2011). Since radio waves are not absorbed significantly by interstellar dust, the entire Milky Way is available for observation. Also, since QSOs are used as the position reference, absolute parallax and proper motions are directly measurable. Such measurements for over 100 star forming regions have now been made for sources as distant as 11 kpc. Significant results (see Section 2) include a resolution of the Hipparcos Pleiades distance controversy (Melis et al. 2014), 3-dimensional "imaging" of nearby star forming regions (Loinard et al. 2007), and the most accurate measurements to date of the distance to the Galactic Center, R0, and the circular rotation speed of the Local Standard of Rest, Θ0 (Reid et al. 2014).