The e-MERLIN upgrade

Not quite as famous as the Lovell telescope, the Multi-Element Radio Linked Interferometer Network (MERLIN) is no less important. The array began life in 1980 and has been providing some of the most detailed images of the cosmos ever since. It is the only radio telescope capable of routinely matching the resolution of the Hubble Space Telescope (HST) although, as it's elements are spread across more than 200 km from Cambridge to Great Malvern, Oswestry and Cheshire, it is far less photogenic than the Hubble.

The array has undergone several upgrades over its 25 year lifetime including: various upgrades to the Lovell telescope itself, the addition of the Cambridge antenna in 1991, new receivers which have been added over the last few years, and upgrades to the microwave links which transport data from the remote telescopes back to the observatory at Jodrell Bank. Currently the array is undergoing the most ambitious upgrade yet, one that will dramatically improve the sensitivity of the telescope and ensure MERLIN stays at the forefront of modern radio astronomy for many years to come.

Interferometry basics
An interferometer is basically a collection of telescopes spread over a wide area. The resolution of any telescope is proportional to the wavelength at which you are observing, divided by the diameter of your aperture. The HST, with its 2.4 m mirror, has a resolution of about 0.05 arcseconds, while the Lovell telescope has a resolution of only 0.6 arcseconds at a typical wavelength of 21 cm. To match the resolution of the Hubble at 21 cm, a telescope 200 km in diameter would be required. As this would be rather impractical, interferometers are used instead.

To picture how an interferometer works, imagine having a single dish 200 km in diameter. Radio waves from a distant object will hit any given point on the surface and be reflected to a receiver positioned at the primary focus just like in an optical telescope with a spherical primary mirror. Now imagine removing sections of the surface. Each remaining piece will still reflect incoming waves to the receiver, but fewer will be collected. Of course, the less collecting area you have, the worse your final image will be. This is the basic idea behind an interferometer. The telescopes which make up the array are like the remaining chunks of our giant imaginary dish (we let the rotation of the Earth help us out so that each telescope occupies a "track" rather than a point in our dish), with a central computer, known as a correlator, acting as the prime focus. It is important to note that the correlator need not be physically at the centre of the array, it can be anywhere convenient, even in another country as is the case with very large interferometers.

The signal is collected by each individual telescope in the array and must then be sent back to the correlator. In a real time interferometer the signal is sent back to the correlator as it is collected, either by microwave link (as in MERLIN) or by electrical cabling (as in the VLA). With larger arrays this is not currently practical so the data is recorded at the individual telescopes and sent back by tape, a technique known as Very Long Baseline Interferometry (VLBI). These tapes are then read back by the correlator at a later date before the resulting data are sent to the astronomers who proposed the observations. MERLIN is currently the world's largest permanently connected array.

So why the upgrade?
The current MERLIN array has worked well for 25 years, but with other telescopes undergoing upgrades, and new telescopes such as the Square Kilometre Array (SKA) in the pipeline, things must move forward for the facility to remain a world class observatory. The developments in technology since the array first went into operation will allow spectacular improvements in the science which can be carried out.

A number of methods can be used to improve the sensitivity of an interferometer: increasing the number and size of telescopes used, utilising more sensitive receivers at each telescope, and increasing the bandwidth of the links. There are already new receivers being installed at the telescopes, and there are plans to build anohter telescope in Ireland, but the current phase of development centres on the links.

The microwave links
The bandwidth limitations of the current microwave links mean that only a small fraction, less than one percent, of the actual data collected by the antennas is transmitted to the computer at Jodrell Bank while the rest is discarded. The sensitivity (how faint a source has to be before it is no longer visible) of an interferometer is directly related to the available bandwidth so by increasing the capacity of the links more data will be returned making higher sensitivity observations possible.

One of the major challenges for the e-MERLIN team has been to develop the required technologies for new links utilising optical fibres. Engineers based at Jodrell Bank have been working closely with people from Global Crossing UK and Fujitsu Telecommunications Europe to create the links and get them operational. When up and running, the new system will be capable of a carrying a bandwidth of 4 GHz, an improvement by a factor of more than one hundred over the current system which has a maximum capacity of only 30 MHz.

All this data will be transmitted along the fibres by lasers. Three different "colours" will be used, each one capable of carrying 10 Gigabits per second (Gb/s). Each telescope will be sending a total of 30 Gb/s back to the correlator but, because of the enormous data rates involved in the new system (sustained 150 Gb/s which is about five times the total UK public internet traffic), a new correlator is being designed by a team from the National Research Council of Canada at Penticon, British Columbia which will have the capability of processing over 200 Gb/s.

Further improvements
Although the new fibre links are a major part of the upgrade, there is much more going on which will further improve the capabilities of the telescope. One of the next improvements are new "lenses" which will improve performance at 1.4 GHz. Another major development is frequency flexibility, the ability to change observing frequency remotely to take advantage of optimum weather conditions for high frequencies where the atmosphere can severely affect results. Telescopes like the VLA are already capable of this, the array can switch between 1.4, 5, 8.4 and 15 GHz in a couple of minutes where as it currently takes a couple of days with MERLIN as an engineer is required to physically visit each telescope. The aim is to have 1.4, 5, 6 and 22 GHz flexibility by means of a rotating carousel of receivers at the cassegrain focus of each telescope.

Other developments
Of course MERLIN is not the only facility to be making use of this new technology in order to increase its sensitivity. One other major development is the Expanded VLA, or EVLA, an improvement to the exiting Very Large Array in New Mexico. This ambitious project aims to increase the sensitivity of the current VLA by upgrading the receivers, replacing the cabling which carries the signals from the telescopes to the control building by fibre optic cables and building a new correlator. They have the advantage that all their telescopes are relatively close together (the maximum baseline for the VLA is 36 km compared to MERLIN's 217 km), but they have far more telescopes (27 antennas compared to MERLIN's 7).

Future telescopes are also making use of this new technology. The planned Atacama Large Milimetre Array (ALMA), an array of 64 dishes to be built high in the Atacama desert, had fibre optic links as an integral part of the design from an early stage. Upgrades to both MERLIN and the VLA serve as useful prototypes for these future systems.

The European VLBI Network (EVN) is also going down this route and will become the eEVN (spot the pattern?!). This project is even more ambitious as the EVN is currently not a real-time array: signals are recorded at each telescope participating in a particular experiment and the tapes or disks are sent back to the correlator for later processing. The ambitious plan for the eEVN is to use the fibre trunk network across Europe to transmit data for real-time correlation. This doesn't stop at Europe however.

Real-time gVLBI
Experiments linking the EVN with the VLBA (Very Long Baseline Array) in America have been performed for many years, but as the data has to be recorded at each telescope and shipped to a central location, astronomers routinely have to wait months for their data to pass through the correlator. Recently astronomers and engineers have performed global real-time interferometry using telescopes in the UK, Sweden, the Netherlands, Poland and Puerto Rico. The maximum telescope separation in this experiment was over 8000 km resulting in a resolution of 0.02 arcseconds, several times better than the HST.

Each telescope in the array was connected to the national fibre network of its country. The data was then transmitted across the European research network (GEANT) at 32 Mbits/second to the EVN correlator at JIVE (Joint Institute for VLBI in Europe) in the Netherlands where the 9 Terabits of data were combined. This allowed an image to be formed within hours of the observations, rather than the months it would normally have taken for such an experiment.

The object of the experiment was a supergiant star in Aquila, surrounded by a cloud of dust and gas which has been thrown from the surface as the star comes to the end of its life. The observations looked at strong emission from objects known as masers: regions where the conditions are just right that gas, in this case hydroxyl (OH), strongly amplifies radio waves at specific fequencies. By mapping these masers astronomers can trace flows in the gas surrounding the star and investigate the changing magnetic fields within the cloud. Figure 1 shows a comparison of a MERLIN image of this star with that made recently using this new technique (see the web version of this article for a higher resolution image). Although the MERLIN image (left) may be nicer to look at, the image on the right has a much higher resolution, allowing much more accurate measurements of the maser positions. This is important in accurately analysing the wind coming from the star itself.

Figure 1 Left: MERLIN image of the supergiant IRC+10420 showing the emission from the maser shell. Right: e-VLBI image showing the same region at a much higher resolution. The shell has a diameter of about 200 times that of our solar system and is expanding at about 40 km/s. Click on the image for a higher resolution version.

So what improvements will all these upgrades bring? Dr Phil Diamond, the director of the MERLIN facility, is looking forward to the science that will be made possible by e-MERLIN. "In combination with the newly resurfaced 76m Lovell telescope, the upgrade will give a 30-fold increase in sensitivity" he said. "This will enable the enhanced instrument [..] to probe far deeper into the Universe, achieving in one day what would currently take three years of continuous observation."

The increase in sensitivity (by a factor of 30) will enable astronomers to peer further out into the universe and probe galaxy structure and scale back towards the early universe. Several research plans are already being formed. Some of the major ones involve probing the star-formation history of the universe by observing galaxies out to higher redshifts than is currently possible, studying star-formation in our own Galaxy by watching the disks of gas and dust that surround new stars, observing gravitational lenses in order to investigate cosmology, and peering at the super-massive black holes at the centre of galaxies.

But what really has astronomers excited are the discoveries that are still waiting to be made. The Lovell telescope itself was built before the discovery of quasars, pulsars or gravitational lenses but has been able to observe in detail many examples of both, shedding light on the extreme physics involved. Who knows what exciting new phenomena e-MERLIN will discover? As astronomers have been heard to say, watch this space.

Last updated: Tuesday, 03-Mar-2009 02:50:04 GMT