INSTRUMENT DESCRIPTION

ON THIS PAGE:

• Counter-Dispersed Imaging
  • Demonstration
  • Why is this such a Good Idea
• Project Status

LINKS TO FURTHER INFORMATION:

• PNS Design Studies
  
  • Version 1
  • Version 2
  • Version 3 (for TNG & WHT)
  • Option to include VLT
• Polarization Options
• Simulated Data and Algorithms
• H-alpha extension
• Image Gallery
• Project Page
• Current instrument parameters

COUNTER-DISPERSED IMAGING:

To use PNe to study galaxies one needs to detect them and then to measure their line-of-sight velocity. Normally these two steps are distinct: one does narrow-band imaging (in the [O III] line) to find the PNe, and subsequently takes spectra of the PNe using multi-slit or multi-fibre spectrographs. We have discovered an elegant alternative in which the detection phase can be skipped entirely: counter-dispersed imaging (CDI). The galaxy field of interest is simply imaged through a slitless spectrograph tuned to the [O III] line as shown in the lower half of the figure below. When this image is studied, the background light of the galaxy and the images of foreground stars will be found to be blurred, but the PNe are instantly recognisable as bright point-like images. This is because they, and perhaps an occasional giant cloud of ionised gas, are the only objects with a powerful emission line at this wavelength (5007Å). Each PN will be slightly displaced from its true position on the sky by an amount determined by its exact emission wavelength, and hence by its velocity.

A second image is taken with the spectrograph rotated by 180^o (upper half of figure). This can be done by taking two consecutive, reversed exposures (as has been implemented on our first CDI programmes). Or the two images can be taken simultaneously using duplicate spectrograph arms (the approach used in the PN.S). The PNe seen in the first image will be readily identified in this second image, but with the sense of their displacement reversed. Obviously, these two images taken together yield (1) position and (2) velocity, while the (3) brightness can be obtained from either image. So in one night we can do as well as, or better than, the traditional procedure does in two or three.  

DEMONSTRATION: The following pair of images illustrate the idea (these data were taken at the WHT in April 1997). Click on them to `blink' them. It will be obvious that the displacement of various PNe differs (notice especially how the shape of groups of objects changes). These differences arise because the PNe have different velocities.

A GOOD IDEA?

What are the merits of the PNS idea over other techniques?

• The PNS works with a pair of images through a single narrow-band filter. The total observing time is about the same as that spent in the detection phase alone when conventional spectroscopy is used, since in that case one `on-band' and one `off-band' image is required.
• Despite a slightly lower optical efficiency than direct imaging (because of the use of gratings) the PNS achieves equally good results because:
  • The images can be obtained simultaneously (see below) and are thus less influenced by changes in photometric conditions.
  • There is more signal redundancy in a pair of matching peaks than in a single peak of the same significance. A 3- PNS identification consists of two 3- events in the same row of the detector.
  • All the observing time spent on the galaxy contributes to the identifications.
• As well as requiring a second step, multi-object spectroscopy of this kind is notoriously difficult and rarely achieves a full set of results.
• Although the radial velocity information in the PNS case requires two images, the PNS achieves superior results because:
  • The overall efficiency is higher
  • There is no limit to the number of PNe and no problem with crowding
  • All the observing time spent on the galaxy contributes to the spectroscopy.
  • The PNS is not reliant on the quality of the astrometric information.

PROJECT STATUS