Murchison, AUSTRALIA - I am seemingly in
the middle of nowhere, and yet I do not doubt that the Murchison Widefield Array (MWA) is at the center of the Universe. Australia is beautiful out here. The area is surprisingly green because of recent rains and the sunsets are a mix of pastel reds and blues. At night the sky is filled with shooting stars and the Milky Way cuts through the sky so bright that dust lanes and nebula, like the
Cosack Nebula, seem to have been painted in black on top of the band of stars in our galactic plane. The radio sky as the MWA sees it would look very different. In order to grasp what the MWA does we will have to first explore what radio astronomy and interferometry is.
Radio astronomy is the alchemy of astronomy; shrouded by secrecy and perpeputated by false claims of being able to transmute raw data into gold. There was a time when radio astronomy was really hard, and that time is always, but technology is making new things possible. The Murchison Wide Field array that I am working on here in the outback is only one of the many next generation low frequency radio telescopes coming online or planned such as
LOFAR,
LWA, and others.
Modern astrophysicists can observe the Universe using light, particles, or (hopefully) gravity waves. Classically astronomers observed light through a telescope, but today we don't look through telescopes and we don't just vaguely see light; we precisley count photons from every part of the electromagnetic spectrum. Light is made of photons, but a photon can be thought of as a wave and a particle. Indeed, a photon is a wave and a particle at once. Longer wavelength photons have lower energy and lower frequency compared to short wavelength photons. In the radio regime of the electromagnetic spectrum the particle view of light is not very helpful, in fact many radio astronomers and engineers actually neglect to ever think about about photons and only consider wavelength or frequency. Radio astronomers view light as an electromagnetic waves impinging upon our patient antennas like waves on the beach.
Long wavelength photons come from some very interesting sources in the sky. Radio waves certainly come from the Sun, because the Sun emits some energy at just about every wavelength. Radio waves are also emitted by galaxies, pulsars, and neutral hydrogen (through the
21cm line). However, the wavelength of photons is not constant: it increases as the photons traverse the Universe due to cosmological redshift such that more distant objects are seen at progressively larger and larger wavelengths consider to more and more distant objects. In my research I am particularly interested in studying the distribution of matter in the Universe at the largest of scales and at the earliest epochs when there was an abundance of neutral hydrogen. Radio waves are perfect for studying these phenomena, but it is difficult to build a telescope that can see a widefield of view, can see a wide range of frequencies at once, and has good resolution.
As radio waves arrive at our antennas we can either immediately detect them or we can reflect them to a receiver. The
Arecibo telescope in Puerto Rico is reflector type telescope, as are the antennas in the
Very Large Array. The antenna on your car directly receives the electromagnetic wave because it induces an oscillation in the field inside the metal of the receiving antenna and then you can hook up a transistor, and a speaker - that is a radio like in your car.
The problem with detecting radio wavelengths is that they are not easy to catch and they act way too much like a wave. Waves have strange properties such as interference and diffraction. It can be shown from wave theory that a telescope of diameter D receiving light of wavelength λ has a fundamental angular resolution limit proportional to λ/D. For example the colossal 300 meter diameter Arecibo telescope can only resolve objects down to 3.5 arc minutes (or about half a degree) at a wavelength of .2 meters (or 1.4 Ghz) and that resolving power will only get worse as we move to longer wavelengths. So if you want to see small things in the sky you had better have a huge radio telescope. But wait, there is more. The field of view that a radio telescope can see is also proportional to λ/D. For example at a wavelength of .2 meters it would take Arecibo about 10 separate observations to make an image of the full moon which is about half a degree in the sky.
So if you want to look at the radio sky at high resolution you had better use a huge telescope, but if you want to look at the radio sky in huge swaths, like in survey, you had better use a small telescope. It would seem to be that we are at an impasse to find a decent resolution and decent field of view radio telescope. Enter radio interferometry.
The diagram above is a pictorial representation of the principles of radio interferometry. In box A we have a big radio dish like Arecibo and a radio wave incident upon it. The radio waves hit the dish and reflect to a receiver (not shown in the cartoon) at the focal point. In box B we have chopped our radio telescope into little bits and so while the dish would behave as a smaller dish it would operate via the same principles. Each part reflects the radio waves incident upon it to a single focal point. In box C we have moved the pieces of the dish into several independent dishes and wired them together. Each dish now has its own focus, field of view, and angular resolution limit (this setup is similar to the VLA). Finally, in box D we have gotten rid of even the dishes. Instead of turning the dish to point to a particular object we use the time delay due to the finite speed of light to 'point' the antennas. An object in the direction θ in the sky sends out radio waves that arrive at the antenna tilted. So to catch the same the wave on the antenna on the left and right spanning the arrow in the diagram we use the time delay τ. In this setup there is no pointing the dish there is only an electronic control of simple antenna elements; this is how the MWA works.
The most difficult part of pulling the telescope apart is reassembling the signals coherently. In the diagram this is the function of the box with the circle and x. In radio astronomy that box would be a complex supercomputer and is called a correlator. The computing power needed to operate a correlator scales as the number of antenna elements squared thus it really takes a powerful computer to operate an array with many antennas. The idea is that the signal from each pair of antennas is correlated together to determine the pattern of incident radio waves. This is the basic idea of radio interferometry; the beautiful thing is that you get the large field of view that each antenna would see and the excellent resolution that the large diameter of antennas provide. The description I have given here of radio interferometry is wildly simplified.
So here I am in a shed in Murchison. A generator is humming along and powering all our computers and equipment and importantly the air conditioner keeping me cool. Flies and strange insects pester us relentlessly the moment I step outside. There are a lot of great things about Australia, but it is also a very harsh environment out here. The array has not been cooperating perfectly: there are amplifiers, attenuators, analog to digital converters, correlators, and more that all have to act in symphony for the system to work. The last few days we have solved as many problems as we have created. The radio sky sends its nite rote down upon us and waits for us to complete the instrument.