Aperture synthesis
 
Any large telescope aperture can be considered as being made up of a grid of small elements, each collecting energy and transferring it to a common receiver. At this point, all the contributions are added according to the phase.
Suppose that an aperture is taken as comprising N elements of equal area. The contribution from each element will produce an interference pattern with each of the other N − 1 elements. The overall effect is equivalent to the resultant of N(N − 1)/2 superposed interference patterns from pairs of elements. The spacing of the interferometers ranges from adjacent elements to a maximum across the

Figure 1. The interferometric elements a₁ − a₂ and a₃ − a₄ within a large array carry identical information.
full aperture—all orientations of linking axes are possible. There may be a fair degree of redundancy in the information. For example, in figure 1, the combination of a₁ and a₂ supplies the same pattern and, hence, identical information as the combination of a₃ and a₄, their separations and orientation being identical.
At the expense of not achieving the flux gathering power of the single large aperture, it is possible to devise an array with the right spacings and axis orientations so that it is equivalent to a single large aperture in terms of its angular resolving power. Such a notion is the basis of the skeleton radio telescope.
Consider an array configuration in the form of a ‘T’ with the row comprising 2 elements and the column with  elements. With this pattern, there are approximately 2N different interferometer pairs that can be made by combining one element from the vertical strip with an element from the horizontal strip. These 2N interferometer pairs represent all the possible spacings and distances found in the N(N − 1)/2 equivalent pairs in the complete filled aperture given by  × shown in figure 2.
On the assumption that a radio source does not vary from day to day, the brightness distribution over the source can be derived from the elementary interferometer patterns taken one at a time. It is not necessary to have all the elements present simultaneously—their equivalent contributions can be generated sequentially. In the ‘T’ formation, it is usual to have two elemental telescopes on the top arm running east– west, acting as a simple interferometer, with a third telescope on the north–south track. After setting the dishes for a particular declination, a series of recordings is made, with the separation of the interferometer and moveable telescope adjusted every 24 hours. From analysis of a complete series of recordings, it is possible by computer to synthesize the observations as though they had been obtained by an aperture given by the area described by the ‘T’. There are 2N separate patterns containing all the information that would be present in the field aperture. Hence, if each of the recordings are made for 2N times longer, observations can be made with the same sensitivity and resolution as though they were made by a single large aperture. By progressively changing the declination, the whole of the available sky can be mapped at a particular frequency. The now-famous Third Cambridge Catalogue (3C) of 328 sources measured at 178 MHz was produced by the aperture synthesis method using a ‘T’ configuration.
Another well-established aperture synthesis system is the Very Large Array (VLA) at the US National Radio Astronomy Observatory at Socorro, New Mexico. It comprises 27 antennas, each in the form of a 25 m diameter dish. The elements are set out in a ‘Y’ configuration and the arrangement provides a resolution equivalent to a single antenna 36 km across with a sensitivity equivalent to a dish

Figure 2. A ‘T’ aperture synthesis arrangement, with 2 elements on the east–west row and  elements
on the north–south column, provides interferometric information to give an angular resolving power equivalent to that of a single large aperture with area N.
of 130 m diameter. Four basic arrangements are used in the separations of the antennas within the ‘Y’ and the telescopes are switched about every four months. From ‘snapshots’ taken at each of the configurations, images are synthesized as though recorded by the equivalent filled aperture. For each of the four configurations, the declinations of the individual elements may be progressively adjusted so that high-resolution sky maps are generated.
The European VLBI network (EVN) comprises 18 radio telescopes operated by 14 different institutions. The individual telescopes are spread across Europe and into China. The shortest baseline is 198 m and the longest is 9169 km. EVN often co-observes with MERLIN and VLBA, both mentioned earlier, to form a ‘global’ VLBI array. It is also committed to supporting VLBI observations with the Japanese Orbiting Radio Telescope (HALCA). This latter instrument of 8 m diameter is part of the VSOP (VLBI Space Observatory Programme) mission and, as a result of the telescope’s orbit, provides baselines three times longer than those achievable on Earth.
Currently under development and expansion is the Atacama Large Millimetre Array (ALMA) project. It involves the merger of a number of major millimetre arrays into one global system. Part of the system is being constructed at high altitude (5000 m—Zona de Chajnantor in Chile) with some 64 telescopes of 12 m, with baselines extending to 10 km. The receivers will operate in the frequency range 70–900 GHz.
It may also be noted that the process of aperture synthesis can also be effected by using the rotation of the Earth. If the individual elemental telescopes are made to track a source, the geometry of apparent array configuration presented to the source effectively changes according to the location of the source in the sky. The result of tracking a source is, therefore, equivalent to a continuous change of the telescopes’ physical distribution on the ground. One of the consequences of this is that all VLBI systems can be used as imaging instruments using aperture synthesis.