Defocussed photometry of transits in extrasolar planetary systems

This project aims to obtain high-precision light curves of transit events in extrasolar planetary systems. The approach is to defocus the telescope so each point spread function covers a few thousand pixels. We then continually image the field, using exposure times of generally 60s to 100s, from at least one hour before the transit starts to at least one hour after it has ended. Features of our defocussing approach:

A discussion of these points, plus detailed signal to noise calculations, can be found in our first paper.

Observing procedure

Here are the instructions for defocussed photometry with DFOSC.

  1. Authorisation. We cannot interrupt transit observations without wrecking the final light curve, so the microlensing program has to be suspended whilst we are observing a transit. Uffe and/or Martin will give the green light for this to happen in advance (normally on the day before).
     
  2. Filter. We want to detect plenty of light from these relatively red stars, but we also want to avoid the fringing which affects CCDs at red wavelengths. We normally use the R or I filter.
     
  3. Acquisition. Point to the target. Coordinates and finder charts are available from the main webpage. Take an image of the field of the target, and check that you have the correct field. Save this image - it will be needed for the data analysis. Choose a good naming convention, for example WASP4-20090624.
     
  4. CCD positioning. Check that the target and comparison stars are on optically good bits of the CCD (no dodgy pixels, bad columns, dead spots, etc).
  5. Binning. For the defocussed photometry project we do not bin the CCD.
     
  6. Windowing. Now the target and comparison stars are on good bits of the CCD, change the CCD window parameters so the images cover only the highlighted area on the finder chart. This is done to decrease readout time. If in doubt, make the window a bit too large instead of too small.
    Record the CCD window pixel values and include them in your night report, as they will be needed for aligning the flat field during the data reduction.
     
  7. Defocussing. It is a good idea to experiment with this in advance, either earlier in the night or the night before, as it could take up to 30 minutes to be sure you have got the optimum defocus. The telescope focus is controlled from the GUI. Defocus the telescope to some reasonable value and take a 100s image (a different exposure time might be specified in some cases). We want the light from the target (and main comparison stars) to form an annulus with maximum counts per pixel about 250,000 to 350,000 (for the 32-bit CCD controller). If the maximum counts are outside this range then adjust the defocus and try again. A preliminary focus value may be available from previous observations. Take a note of the focus value and include this in your night report.
     
  8. Once the telescope is defocussed properly, start the sequence of images of the target field using the specified exposure time (typically 60s to 100s).
     
  9. Autoguiding. It is very useful to autoguide, as keeping the light from the stars on the same pixels makes the final light curve better. Find a good guide star box close to the centre of the guiding monitor. The DK1.5 does not autoguide well when the guiding star is outside the central region of the guide monitor.
     
  10. And relax! Check the images every so often to ensure all is well, and keep a note of the weather conditions. Useful trick for observers (from Daniel Evans): the images are now displayed in ds9. The easiest way to see counts in ds9 is under "Analysis" then "Pixel Table"
     
  11. Timing check. At some point during the transit (or alternatively at the start or end of the night) it is important to check whether the timings imprinted on the fits file headers are correct. Choose one exposure and note down the time the shutter opens using either a www resource (for example timeanddate) or your watch. Include this information in the nightlog. With the Danish 1.5m telescope and DFOSC it is possible to hear the shutter open from the control room.
     
  12. Finishing. Continue the observing sequence until the finishing time specified on the main webpage, which is typically 1 to 1.5 hours after the transit has ended. Refocus the telescope and take one more image of the field for reference.
     
  13. Flat fields. It is important to take plenty of flat fields for this project. These can be taken in the twilight before or after the transit observing sequence, or even on the previous or next few nights. Remember to use the same filter (R or I) and to not window the CCD down. Counts for the flat fields should be in the range 250,000 to 350,000 ADUs for the 32-bit CCD controller.
     
  14. Report. Please make sure you fill in the night report on the MiNDSTEp wiki. Include the following: