The Cosmic Lens All-Sky Survey - the search for gravitational lenses.

Ever wondered what astronomers do when they run out of things to do? Or when they just want to use up a few years worth of telescope time? They do a survey of course! They take a telescope, or sometimes a whole array of them, and look at absolutely everything with it, even looking at completely empty areas of sky in some cases. These so called all-sky surveys are not quite as pointless or as frivolous as I've implied, indeed some very important results have been obtained by this method of observing.

A significant proportion of telescope time is often taken up with this kind of survey work. Numerous all-sky surveys have been undertaken at various wavelengths and for a variety of purposes over the years, each one generally more extensive than the last. The reasons for undertaking these surveys vary from accurately measuring the positions and motions of stars (eg Hipparcos) to examining the overall structure of the universe (e.g. 2dFGRS, COBE etc).

The Cosmic Lens All-Sky Survey (CLASS) is one such all-sky survey. It began back in 1994 and the final phase of follow-up observations has only recently been completed with some of the final results and analysis still awaiting publication.

The purpose of the survey, carried out by an international group of astronomers using telescopes around the world, was to discover objects known as gravitational lenses. These bizarre looking objects occur when a nearby galaxy passes in front of a more distant object such as a quasar. The foreground galaxy causes the light from the background quasar to bend in a similar way to what happens to light passing through an ordinary lens. Figure 1 shows a schematic of how the effect works.

Figure 1 - A schematic showing the basic idea behind the gravitational lensing effect

An everyday example of the kind of effect you would see (although using different physical principles) can be seen by taking a wine bottle and looking through the bottom. The image you see is a highly distorted version of whatever is on the other side of the glass.

Gravitational Lenses
In reality however, gravitational lensing is a bit more complicated than this simple picture. Unlike with an ordinary lens, this effect depends on the equations of general relativity, rather than relatively simple wave optics. The basic principle is however simple; if the quasar and lensing galaxy are sufficiently well aligned, multiple images are seen by the observer. If the source and lens are perfectly aligned then, instead of observing just a few images of the quasar, an entire ring is seen surrounding the foreground galaxy, as shown in Figure 2. This is termed an Einstein ring after the famous physicist who developed much of the theory of general relativity upon which this spectacular effect depends.

Figure 2 - The lens system 0218 showing the phenomenon known as an Einstein ring (from

Why are they important?
Apart from looking nice, gravitational lenses are valuable as cosmological probes. Among other things, if the quasar is varies in brightness over time then, because of the different path lengths the light has to travel to reach the observer, the components will appear to brighten and fade at slightly different times. By measuring the time delay between any variations in the brightness of the components the value of the Hubble constant (which relates an objects distance to it's velocity relative to us) can be determined.

This is particularly useful as the result is independent of the measurements (and hence the errors) on which many other Hubble constant estimates depend. If the results obtained via this method turn out to be consistent with those determined by other methods, then it provides a valuable consistency check, indicating whether our current estimates are reliable or not. Studies of this can help in determining just how much of the universe consists of dark matter and what the ultimate fate of the whole universe will be.

How the search works
Before beginning observations for any survey, a list of targets has to be drawn up and a proposal for telescope time submitted. This can be a lengthy process and proposals sometimes go through several revisions before being accepted. For CLASS, the list of targets was compiled from a number of previous radio surveys, the main one being the GB6 survey carried out at the Green Bank observatory in the USA. A series of constraints was imposed on the results of the GB6 survey to reject sources that could not, based on those observations, be lenses. Any objects surviving this process were included in the CLASS target list.

VLA imaging (radio, 8.4 GHz)
The first stage of the survey was to observe each of the candidate objects selected from the GB6 survey with the VLA (Very Large Array) at a frequency of 8.4 GHz and a resolution of 200 mas (milli-arcseconds). The data were processed automatically by computer - 16,503 images are a lot to go through by hand! At this stage another set of constraints was applied to the objects and many were rejected as lens candidates for various spectral or morphological reasons.

MERLIN follow-up (radio, 5 GHz)
Any object having more than one component was re-examined at a higher resolution and a frequency of 5 GHz by the UK's radio array MERLIN (Multi Element Radio Linked Interferometer Network). This enabled the researchers to distinguish between objects which are naturally physically extended such as quasars with associated high energy jets (which are interesting in their own right, but not in the context of CLASS) and those which are multiply imaged by a lensing galaxy. Only compact objects (those without jets) were included in the CLASS survey as gravitational lensing is more obvious and easier to detect in these cases.

VLBA follow-up (radio, 5 GHz)
Objects which were still unresolved by MERLIN's 50 mas resolution were then observed with the VLBA (Very Long Baseline Array) at a frequency of 5 GHz and a resolution of 2 mas. This enables any sources larger than 2 mas in size to be rejected from the survey as not being truly compact.

Optical follow-up
In cases where the objects are fairly certain to be lenses, optical observations were also made with the aim of detecting light emitted from the lensing galaxy itself. This was done with the Hubble Space Telescope, the Keck telescope, and the WHT (William Herschel Telescope). If it is visible optically, the shape of the lensing galaxy can be determined, or the optical counterparts of the quasar images can be seen. This allows precise astrometry and photometry to be performed in order to determine the exact position and magnitude of the optical object(s). Figure 3 shows one lens (called B0445+123, the B signifies that the B1950 co-ordinate system was used, and the numbers are it's approximate RA and Dec, putting it in the top-right of Orion) where optical observations have been made; the optical image is shown in greyscale with the radio data over layed as a contour plot, the alignment of the images is accurate to ~0.3 arcseconds and the magnitude of the optical image is 21.8 ± 0.4

In this instance the optical image, a 600 second exposure with the WHT, is comprised of light from the lensing galaxy and at least one of the two radio components, but better observations are going to be necessary in order to resolve any structure. Often, spectroscopy is also performed on the optical images in order to determine the redshift of both the lensing galaxy and the quasar, providing more useful information for cosmologists.

Figure 3 - The lens system B0445+123. The optical image is shown in greyscale with the radio map over layed as a contour plot.

The CLASS survey observed a total of 16,503 sources with the VLA over a period of five years. Out of this initial sample, the vast majority of objects were found not to be lenses, but that doesn't mean that they are not useful observations in themselves. Part of the objective of the CLASS survey was to accurately measure the flux densities of all of the objects observed, thus providing an extensive list of calibrator sources for future observations at these frequencies.

Out of the total initial sample, a total of 22 sources have subsequently been confirmed as gravitational lenses using as many follow-up observations as possible. That may not seem like a lot, but compared to other similar gravitational lens searches it is quite a high detection rate! Out of the 22, one was discovered to be an Einstein ring system (Figure 2), most were simple two or four image systems, and two were much more complicated having more than one lensing galaxy or more than one background quasar producing very strange image distributions.

The Future
As astrophysics goes, gravitational lensing is still a relatively young field, but it has produced some valuable results and will hopefully continue to do so for many years to come. The underlying theory of lensing is being developed all the time and each new system found adds valuable information to the models. Currently, a similar survey to CLASS is being undertaken in the southern hemisphere that will hopefully discover a new batch of lensing systems to add to the menagerie, and the rest of the CLASS data itself is being analysed to determine values for various cosmological parameters such as the Hubble constant and the overall matter density - the number that determines, to a large extent, the ultimate fate of the Universe.

If you're curious to see what the rest of the lens systems look like you can have a look at the gravitational lensing group's webpages at Jodrell Bank:
Telescopes involved in CLASS:

Last updated: Tuesday, 03-Mar-2009 02:50:17 GMT