Radio Astronomy's Resolution Machine: The Very Long Baseline Array --How radio astronomers can hit the "zoom" button and scrutinize unusual objects with unprecedented clarity.
The quest for higher resolution has been one of the themes of astronomy ever since the invention of the telescope. Galileo's first visual sighting of Jupiter and its major satellites was surpassed by observations with the progressively larger telescopes of the centuries that followed; in turn, these were surpassed when the Voyager spacecraft flew past the giant planet in 1979. Each "zoom factor" enabled enormous advances in our understanding of the Jovian system, allowing us to ask questions that the fuzzy first views could never have addressed.
Another example of the benefits of higher angular resolution comes from pictures of the Orion Nebula with a camera lens, a large ground-based telescope, and HST (see page 41). These comparisons make it plain why astronomers have done everything they can to increase angular resolution. Fueled by human curiosity about the cosmos, the drive for sharper astronomical vision has driven the development of numerous technologies, VLBI among them. And VLBI has culminated in the construction of the Very Long Baseline Array (VLBA), the highest-resolution imaging device in routine use today.
What Is Interferometry?
Radio waves are simply a type of electromagnetic radiation, just like visible light and gamma rays. The different types of electromagnetic radiation are distinguished by their wavelengths (the distance between successive crests). Radio waves typically have wavelengths measured in millimeters or centimeters, while visible photons (light particles) have wavelengths measured in tenths of a micron (a micron is 11/41000 of a millimeter). An electromagnetic wave can also be described by its frequency (the number of crests passing a given point in one second). Typical radio frequencies are 100 million to 100 billion cycles per second (100 megahertz to 100 gigahertz), while visible light has a frequency near 600 trillion cycles per second.
If a cosmic source of radio waves is either small enough or distant enough, it acts as a "point source" of radio waves, and the wave crests reach all points on Earth as parallel planes. In interferometry, the radiation patterns received by two or more telescopes are multiplied by one another (or cross-correlated), and the result is recorded. If the difference in reception time between two telescopes is the same as the time for the radiation to travel an integral number of wavelengths, the two patterns will be in phase and the product will be at a maximum. If the difference in reception time corresponds to an integral number of wavelengths plus one-half wavelength, the two patterns will be out of phase and their product will be negative. Atmospheric effects aside, the difference in arrival time depends only on geometry and can be used to determine the source direction.
Now imagine that the astronomical object is an extended collection of independent point sources. The two telescopes receive waves from each point source, so the wave patterns are much more complex. Fortunately, as the Earth rotates, the target moves across the sky, altering the relative geometry of the telescopes and the incoming waves. By monitoring how the signal strength and relative time delay change at each observing station, it is possible to derive information about the object's structure.
Why go to all this trouble? The level of visible detail, or resolution, depends on the ratio of the diameter of the light-gathering aperture to the observing wavelength. A single telescope is limited to the structure that can retain its shape under the influence of Earth's gravity. However, using two smaller telescopes and the principle of interferometry, astronomers can resolve objects as well as they could with a single large aperture whose size is equal to the baseline (the separation between the two smaller telescopes). Thus, instead of being limited to radio-telescope diameters of a few hundred meters, astronomers can form synthetic apertures as large as the Earth - nearly 100,000 times larger than the largest individual radio telescopes. Thus they can improve the resolution in their images by the same factor.
Optical telescopes use wavelengths that are 10,000 to 100,000 times smaller than typical radio wavelengths, so at first glance it might seem that infrared or visible-light interferometers should far outstrip radio ones. Yet optical interferometry is technically far more demanding and only in the last few years has it produced the sorts of images that radio astronomers began making nearly three decades ago (S&T: November 1996, page 36).
Radio Interferometry's Early History
In the late 1940s astronomers and engineers first demonstrated radio interferometry with war-surplus radar dishes. Within 20 years their efforts led to the construction of instruments like the Cambridge 1-mile telescope, an interferometer that afforded a resolution of a few arcseconds. This telescope was later expanded to 5 kilometers in size, and in the 1970s an array of 14 telescopes was built along a 3-km line at Westerbork in the Netherlands.
In the late 1970s a powerful radio interferometer called the Very Large Array (VLA) was funded by the United States' National Science Foundation and built by its National Radio Astronomy Observatory (NRAO). This telescope consists of 27 antennas spread across 25 miles (40 km) of the Plains of San Augustin in central New Mexico.
Precision imaging by such interferometers revealed a universe very different from that seen at optical wavelengths. The brightest radio sources usually were distant faint galaxies or pointlike quasars, which now are known to lie at the centers of distant active galaxies (May issue, page 40). Except for the Sun and a few planets, the strongest radio sources were not visible to the naked eye, and most, it turns out, require powerful optical telescopes to "see" the corresponding object at visual or infrared wavelengths.
Even before the VLA was built, ambitious astronomers envisioned radio interferometers that would gather signals simultaneously with telescopes thousands of kilometers apart. Such interferometers would provide imaging capabilities on milliarcsecond scales - far beyond the reach of any other telescope. But this was to be no easy undertaking. Individual VLBI antennas were too far apart for real-time signal mixing, so large quantities of data had to be recorded on magnetic tape and shipped to a common location. Furthermore, the absolute arrival times of radio crests had to be recorded with a precision of less than a microsecond, so extremely accurate atomic clocks were required at each participating observatory.
The first successful VLBI observations were performed more or less independently by a Canadian group, an NRAO/Cornell University group, and a group at the Massachusetts Institute of Technology. These experiments took place in 1967, using radio wavelengths ranging from 13 cm to 70 cm and telescopes separated by hundreds to thousands of kilometers. Within a year, a telescope at Onsala, Sweden, participated; the resulting North America-Europe baselines provided the first successful VLBI between two separate continents.
VLBI soon showed many radio sources to be only a few milliarcseconds in size. In 1971, individual radio-emitting parcels of matter were shown to be racing from two quasars at several times the speed of light. This superluminal motion is now known to be an optical illusion caused when matter moves at near-light speeds almost directly toward Earth.
What Can VLBI Show Us?
VLBI provides very high resolution. However, the finite sensitivity of individual radio telescopes means that only very intense radiation sources can be investigated. A VLBI observation of Venus might provide exceptional resolution of its surface features. But Venus - like the hypothetical basketball discussed above - is not a sufficiently intense radio source to be detectable on the small scale probed by VLBI. Instead, the radio sources observed with VLBI techniques emit enormous amounts of radiation from very small volumes.
For instance, active galactic nuclei such as quasars apparently are powered by matter falling into supermassive black holes. These black holes generate energy 10 to 100 times more efficiently than the nuclear fusion at the centers of stars. Much of the energy is converted into radio emission detectable here on Earth.
Optical telescopes observe predominantly thermal radiation. For instance, stars emit visible light because they have temperatures of a few thousand degrees Kelvin, whereas star-forming regions observed by infrared telescopes have temperatures just tens to hundreds of degrees above absolute zero.
In contrast, the objects observed by radio astronomers typically emit nonthermal radiation caused by more exotic processes. Continuous radio "noise" - called synchrotron radiation after particle accelerators on Earth - is emitted by electrons moving near the speed of light in strong magnetic fields; it is the main source of radio waves from most active galaxies. Radio emission at discrete wavelengths is characteristic of particular atoms or molecules; with VLBI, we typically observe such "line" emission in masers, the radio equivalent of lasers. Masers are sometimes seen in active galactic nuclei but are also common around evolved stars or where new stars are forming in the Milky Way. They are particularly useful because their velocities can be determined from wavelength shifts caused by the Doppler effect.
Why We Built the VLBA
In the 1970s increasing numbers of individual observatories participated in VLBI observations. As new techniques developed, rough measures of source sizes gave way to high-fidelity images and more detailed scientific information. However, the logistics of scheduling telescopes, transporting atomic clocks, and setting priorities became overwhelming for informal cooperatives. This gave rise to the Network Users Group in the U.S., and later the U.S. VLBI Consortium and the European VLBI Network. (The latter has recently been reorganized under the Joint Institute for VLBI in Europe, with headquarters in Dwingeloo, the Netherlands.) These groups have produced significant improvements in VLBI reliability and have enabled as many as 18 telescopes to observe simultaneously.
Still, the telescopes used in VLBI belonged to many individual universities and research organizations with differing degrees of commitment to VLBI. The telescopes had different sizes, shapes, and slew rates; some took several days to change observing wavelengths; and they operated together for periods of only two or three weeks, several times per year. This meant that variable sources were monitored in ways that were determined by logistical constraints rather than by research needs.
The serious limitations of ad hoc arrays led to the concept of a dedicated VLBI array. Such an instrument was strongly recommended in a 1980 survey of priorities for U.S. astronomy and astrophysics. The Very Long Baseline Array, an array of 10 identical 25-meter telescopes, was proposed in response to this recommendation. It was funded (again by the National Science Foundation) at a final cost of $84 million in 1989 dollars and was formally dedicated in 1993.
The VLBA has solved many problems of the older VLBI consortia. It operates year-round. All the components of each element are identical, from the telescopes and their receivers to the control systems and data recorders. The commonality enables all 10 stations to be controlled by a single telescope operator at the NRAO Array Operations Center (AOC) in Socorro, New Mexico. Each station is staffed by two local technicians (and a third at the high-altitude site in Hawaii); a pool of engineers and spare modules in Socorro supports repairs and enhancements that cannot be done locally. A dedicated VLBI signal processor (or correlator) at the AOC can combine the VLBA data with data from up to 10 other non-VLBA telescopes. Finally, VLBA data can be processed within weeks of an observation, rather than the months or years that were common previously. The VLBA's "user- friendliness" has made it much easier for young astronomers to practice very long baseline interferometry; several of the research programs described below were led by graduate students.
One great advantage of the VLBA is that the station locations, from Hawaii to the Virgin Islands, provide a variety of baseline lengths, from 200 to 8,000 km. Together with the Earth's rotation, this range of baselines makes the VLBA sensitive to structures on a wide range of scales.
A Sampling of Science Highlights
One of the VLBA's most outstanding discoveries has come from water molecules in the accretion disks at the hearts of certain active galaxies. Under favorable conditions those water molecules act as masers, beaming strong radio waves from regions only a few light-years across. The VLBA has measured the motions of many such masers and, along with Kepler's laws, this has enabled astronomers to assess the masses of black holes at the centers of these galaxies. The classic example is the nearby Seyfert galaxy M106, also known as NGC 4258 (S&T: April 1995, page 10). There, a black-hole mass (39 million times that of the Sun) has been measured with an error of only a few percent. In contrast, black-hole masses measured by other techniques typically rely on motions much farther from the galaxy centers and have errors of a factor of two or more.
VLBA images of M106 (NGC 4258) have been produced over a four-year time span by James Herrnstein (NRAO) and his colleagues. The images show maser motions across our line of sight as well as maser accelerations toward the central black hole. Together these motions imply a distance of 23.5 [plus-or-minus sign] 1.0 million light-years - by far the most precisely determined distance to any galaxy. Since this distance is assessed simply by geometry rather than being based on a chain of assumptions, it should serve as a "ground truth" to
which other distance measurements can be compared. In fact, a long- running Key Project for HST has been to establish the cosmic distance scale by using Cepheid variable stars in galaxies like M106. The VLBA findings for M106 favor a "long" distance scale and they may imply that HST has underestimated other galaxy distances by 10 or 15 percent (October issue, page 20).
The evolved supergiant stars known as Mira variables are related to the Cepheids. Many Miras have cool clouds of dust and gas outside their photospheres. In those clouds, molecules condense, then radiate by the maser process. Philip Diamond (Nuffield Radio Astronomy Laboratories, Great Britain) and Athol Kemball (NRAO) are using the VLBA at a wavelength of 7 mm to image silicon monoxide masers in the extended atmosphere of one Mira variable, TX Camelopardis. TX Cam has been imaged every two weeks throughout its entire 1.5-year pulsation period - an observing program that was impossible until the construction of the VLBA. A movie of the entire data set shows unexpected phenomena like complex streaming motions within the circumstellar envelope and maser spots that expand away from the star when they were expected to fall inward. These observations should revolutionize our understanding of physical processes in the extended envelopes of evolved stars.
Another unique VLBA capability is high-resolution imaging of polarized radio emission. Polarized emission indicates the orientation of magnetic fields in different regions of radio sources. Magnetic fields are thought to collimate the relativistic jets coming from active galactic nuclei. Changes in the orientation of the magnetic fields often reveal shock waves and regions of particle acceleration within the jets. They may even indicate the composition of the mysterious jets (March issue, page 24). Prior to the VLBA era, very long baseline polarimetry was attempted only by a few VLBI "black belts," who labored to calibrate the disparate telescopes used in ad hoc arrays. Now polarimetry is routinely available with the VLBA, and many recent studies show the evolution of magnetic- field structures in active galaxy jets.
A number of groups are using the VLBA to measure cold gas, which absorbs specified wavelengths from nonthermal radio sources that lie in the background. Alison Peck (a graduate student at New Mexico Institute of Mining and Technology) and her collaborators have imaged absorption due to neutral hydrogen seen against radio jets in active galactic nuclei. Their measurements tell us the sizes and densities of gas clouds on scales of a few light-years, providing new information about the fuel that supplies supermassive black holes. Using a similar technique, Michael Faison (a graduate student at the University of Wisconsin) is imaging the distribution of neutral hydrogen in our own galaxy by looking at 21-cm absorption along the line of sight to distant extragalactic radio sources. He has found large fluctuations in gas density on scales of about 10 astronomical units, smaller than the size of our solar system. This signifies extremely small-scale "weather" in the Milky Way, as if we had large fluctuations in the Earth's atmosphere on scales of just 5 centimeters!
Two other graduate students, Walter Brisken (Princeton University) and Shami Chatterjee (Cornell University), are using the VLBA to measure the positions of pulsars. Such observations are demanding since pulsars radiate copiously only at long wavelengths, which are strongly affected by Earth's ionosphere. However, they are possible because of the VLBA's ability to switch rapidly between weak targets and stronger reference sources. This "phase referencing" method extends the VLBA's sensitivity and is now used in nearly half of the array's programs. Pulsar parallaxes smaller than a milliarcsecond (corresponding to distances greater than 3,260 light-years) have been measured and are being used to calibrate models of the Milky Way's electron density and the pulsar distance scale.
Perhaps the most intriguing result from positional measurements with the VLBA comes from the center of our Milky Way. For four years now, Mark Reid (Harvard-Smithsonian Center for Astrophysics) and his collaborators have used phase referencing to measure the position of a radio source called Sagittarius A*. Sgr A* is thought to be at rest in the gravitational center of our galaxy. The VLBA shows an apparent proper motion of 5.9 milliarcseconds per year for Sgr A*, and this motion is almost entirely confined to the plane of the Milky Way. This is an expected consequence of our solar system's orbit around the center of the galaxy. Thanks to the VLBA, it took only two years to precisely measure the period of our galactic orbit (some 220 million years).
The VLBA's Future
Very long baseline interferometry is paradoxically successful. The longest wavelengths of electromagnetic radiation should provide the poorest resolution, but instead they have done just the opposite: they have provided the highest-resolution images astronomers have seen thus far. And the VLBA has demonstrated the scientific benefits of a dedicated VLBI array.
What's in the future for the VLBA? A first glimpse came in 1997 when Japan's Institute of Space and Astronautical Science launched the HALCA (High Altitude Laboratory for Communications and Astronomy) satellite (S&T: September 1997, page 16). HALCA carries an 8-meter radio telescope and is used together with the VLBA and other radio telescopes to provide baselines as long as 30,000 km (about 211/42 times the diameter of the Earth). In the late 1980s a communications satellite showed that space VLBI could be done, but HALCA was the first spacecraft built specifically for the task. HALCA's shortest wavelength receiver system failed, limiting its resolution and scientific returns, but the ability to routinely obtain images with space VLBI has been demonstrated.
A successor to HALCA, the Advanced Radio Interferometry between Space and Earth (ARISE) mission, is currently on NASA's long-term road map. If funded, ARISE would orbit a radio telescope with capabilities similar to those of the VLBA's antennas: a 25-meter aperture operating at wavelengths as short as 3 mm. The spacecraft would be launched sometime after 2010. Of course, a VLBA telescope weighs 200 tons, while the ARISE spacecraft must have a mass less than 1 percent of that value. The key is the kind of inflatable structures under development by NASA's new Gossamer Spacecraft Program.
ARISE would fly as high as 40,000 km above Earth, providing resolution of nearly 15 microarcseconds - about 3,000 times better than HST - at its shortest wavelength. This will enable us to see the gamma-ray-emitting regions in distant quasars as well as areas just outside the event horizons of supermassive black holes in nearby active galaxies.
Of equal scientific interest is a plan, not yet funded, to place eight new radio telescopes around New Mexico. In concert with the VLA and the VLBA, this would provide a radio-imaging camera that could simultaneously probe radio sources on a wide range of scales. A small first step took place last year, when the VLBA antenna in Pie Town, New Mexico, was connected to the VLA in real time by means of a commercial fiber-optic cable. Those new antennas could effectively turn the United States into one giant radio telescope, providing scientific capabilities far beyond those envisioned when radio interferometry began just a half century ago. Thus the future of the VLBA is bright, and it can be expected to expand both inward and outward in the coming decades to give an increasingly detailed look at our active universe.
James Ulvestad is a staff scientist at the Socorro, New Mexico, offices of the National Radio Astronomy Observatory (NRAO) and the project scientist for the ARISE mission. Miller Goss is the director of the VLBA and the Very Large Array. Visit the National Radio Astronomy Observatory's home page on the World Wide Web at http://www.nrao.edu.
|Printer friendly Cite/link Email Feedback|
|Author:||Ulvestad, James; Goss, Miller|
|Publication:||Sky & Telescope|
|Date:||Dec 1, 1999|
|Previous Article:||Now It's Your Turn.|
|Next Article:||A Crash Course in Aperture Synthesis.|