ExPo, or the Extreme Polarimeter, is the instrument I developed during my PhD at the Utrecht Astronomical Institute. It specialises in taking images of linearly polarised light, while suppressing unpolarised light to a high degree. Why would we want to do this? Have a look at the image to the right. Starlight is generally unpolarised. But when it reflects off some material it to some degree becomes linearly polarised. The ‘material’ may be the dust of a debris disk or protoplanetary disk but could also be the surface or atmosphere of an exoplanet. Disks and exoplanets are very hard to observe directly as they are close by and very much fainter than their parent star. Imaging polarimetry is thus a way to block out the star and image the light reflected off material surrounding it at high contrast. 

The ExPo instrument is unique, not only in terms of the astronomical observations and thus the science it can do. While many astronomical instrumentation projects have long lead times, our idea was to get something to the telescope in a short time, then improve it using the experience gained. After all, the instrument design, development, commissioning and the reduction, calibration, interpretation and publication of the initial science observations all had to take place in the 4-year timeframe of my PhD project! This was a successful approach: First Light at the telescope took place a little over a year and-a-half after I started my PhD. This was achieved by applying a KISS (Keep It Simple, Stupid!) philosophy and by using off-the-shelf components for the instrument wherever possible. The instrument was commissioned at the 4.2 m William Herschel Telescope on La Palma. This telescope is great because it has a visiting instrument focus (a Nasmyth). At left you see the basic version of ExPo installed on the optical table at this focus. At the time of writing, ExPo has been to the telescope seven times. The shipping, installation and alignment of the instrument, followed by the period of night-time observations, is always a hectic and often an exciting and fun experience!

Polarimetry is a differential technique; it always consists of making two measurements and comparing them in some way. Light can be split into two orthogonal polarisation components. A polarising filter (an example of which is a pair of polaroid sunglasses) will block one component and pass the other. By rotating the filter 90 degrees, the component that is blocked and the one that is passed are swapped. For unpolarised light, the intensity of the two components is equal. So if we would make two measurements with the polarising filter and subtract them, the result would be zero. If the light however is polarised, one component will be brighter than the other. The key to building a very sensitive polarimeter is to make sure no systematic effects can introduce an artificial difference between the two measurements. Where we to make the two measurements at different times, some slight change in the source (or the instrument) would manifest itself as measured polarisation. If we on the other hand measure the two components simultaneously, using a polarising beamsplitter to separate the two components, any difference in the two measurement channels also introduces an erroneous signal in the result. The trick is to use the method depicted in the figure at right, called the beam-exchange method. The two polarisation components are measured simultaneously after which a polarisation modulator, which can rotate the polarisation, swaps the components being measured between the two beams. This is repeated many times to generate an averaged polarisation result largely free of systematic errors.

The image at right showcases an ExPo observation. The star targeted is T-Tauri, which is in fact a young multi-star system. A complex (protoplanetary?) structure of gas and dust surrounds the system. When imaged in intensity (inset below right) this structure can not be seen.  The larger polarised intensity image reveals the structure as the dust scatters and thus polarises the starlight.

The image is 10 arcseconds across. The blue vectors indicate the polarisation direction while the background red-yellow-white colour-scaled image is of the total polarised intensity. The degree of polarisation ranges from a few percent to less than a tenth of a percent. The observation was obtained with ExPo at the William Herschel Telescope without the use of a colour filter.