What is the Fabry-Perot `staring' technique?

To our knowledge, there are four groups actively engaged in extragalactic work using the `staring' technique. These programs are based at Wisconsin (since 1971), Rutgers (since 1982), AAO/LPO/Hawaii (since 1985), and Maryland/Caltech (since 1988). Notably, R.J. Reynolds has made use of a 1 field to characterize the warm ionized gas throughout the Galaxy in , [OIII], [NII] and [SII]. The other groups typically use fields of view an order of magnitude smaller which makes them applicable to a wider variety of extragalactic projects than the Wisconsin system.

The `staring' technique exploits the Jacquinot advantage of the Fabry-Perot to obtain extremely deep spectra of extended, diffuse objects. For a fixed gap spacing, light falling on the instrument with wavelength is dispersed according to , where is the off-axis field angle. The spectrum in a narrow band is dispersed radially from the optical axis across the field. When the data are binned azimuthally about the optical axis, a single deep spectrum is obtained. Like long-slit spectrometers, the instrumental profile is projected onto the detector but the line FWHM varies across the field according to . At the AAT, we have already achieved H emission measures of 0.2 cm pc ( erg cm s arcsec) at the 3 level in about 90 minutes (Bland-Hawthorn et al. 1995). In principle, we are able to reach 0.02 cm pc (3) in about six hours using larger optics. Due to the factor, the raw spectrum has quadratic sampling and needs to be resampled to a linear axis where the original number of bins is preserved.

There are any number of potential pitfalls with the `staring' technique. To achieve the highest sensitivity, the diffuse source needs to be monochromatic within the detector field and have an intrinsic line dispersion roughly equal to the instrumental profile. A strong kinematic gradient across the field would wash out the signal after azimuthal binning.

There is a wide class of problems relating to the atmosphere. Night sky lines vary in brightness temporally and as a function of airmass. Water absorption features come and go with the site humidity. Subtracting off the background can be particularly hazardous if the `on' and `off' spectra are obtained in separate exposures. Using data in the same CCD frame to subtract the background has its own hazards. Thinned chips have a bad fringeing problem at high dispersion. An important test is to show that the spectral baseline is identical in, say, three sectors of the field. Another weakness in many published results to date (e.g. Songaila, Bryant & Cowie 1989) is that the wavelength domain of the data is typically only 5-10 times the resolution element, a problem well understood by radio astronomers. This greatly complicates defining what constitutes a reliable continuum level as the diffuse emission we are after is usually a faint signal situated between bright night sky lines with dominant wings. For this reason, we have worked hard to achieve 40Å bandpass at at 1Å resolution.

For systems at low velocities, we need to contend with both the geocoronal lines and with Galactic emission. Some observers appear to confuse night sky lines with Galactic emission (particularly the [NII] lines), for example. It is possible to lower the background within the resolution element of the expected signal by either using the Earth's motion (i.e. choosing the right time of year to observe) or by observing the object close to local midnight.

© Copyright Astronomical Society of Australia 1997