Overview of the Instrument

The Fabry-Perot Interferometer

Thorough discussions of the principles underlying the Fabry-Perot interferometer, and its use in astrophysics, can be found elsewhere (e.g., Vaughan 1989; Bland & Tully 1989), and only a few important definitions will be given here. In essence, the Fabry-Perot interferometer consists of a pair of identical transparent plates, having plane-parallel internal faces of reflectivity R, separated by a uniform spacing d. Peak transmission is attained over a series of orders n when

where is the refractive index of the medium between the plates, and is the angle relative to the normal of the incident beam with wavelength . The spectral distance between two adjacent orders n and n+1 is called the Free Spectral Range (FSR), and is given by

For a ``perfect'' Fabry-Perot etalon, the Full Width at Half-Maximum (FWHM) of each order is

Thus the resolving power can be described as

where is called the reflection finesse. In practice, the true ``effective'' finesse of a Fabry-Perot system is always less than the reflection finesse, due to surface defects in the coatings, departures from plate parallelism, and the use of a converging, rather than a parallel, incident beam. Replacing with in equation 4 then allows the resolving power to be estimated in the general case.

The UNSWIRF etalon

Table 1 summarises the specifications of some currently available near-IR imaging Fabry-Perot systems. As can be seen, various combinations of resolving power, field of view, and tuning range are available. The UNSWIRF etalon was specifically intended to meet the following goals:

1. a high resolving power (); this was a trade-off between the amount of spectral scanning required at high resolution to fully sample the line, and the reduced line-to-continuum contrast and velocity information available at lower resolution;
2. a choice of pixel scales (e.g.,  pixel or  pixel, depending on the imaging optics selected within IRIS) matched to the seeing conditions and the clear aperture of the etalon;
3. the ability to cover lines in both the H band (e.g., [FeII] at 1.6440 m) and in the K band (e.g., H at 2.1218 m, H at 2.2477 m, and Br at 2.1655 m).

   

Table 1: Near-Infrared Imaging Fabry-Perot Systems

Instrument	Telescope	Band	Resolving	Maximum	Sensitivity	Reference
			Power	Field
Cornell	Various	K	3300		7	1
FAST	Various	K; K	1000; 2700	43'' circle	5	2
FINAC	CRL 1.5 m	J; K; K	680; 1250; 12000		50	3
IRAC	ESO 2.2 m	K	1400	180'' circle	5	4
IRCAM3	UKIRT	K	860		4	5
NASM/NRL	WIRO	J+H; K	800; 800		<20	6
UNSWIRF	AAT	H+K	4000	100'' circle	5	7

At the heart of UNSWIRF is a model ET-70WF etalon, manufactured by Queensgate Instruments (UK) Ltd., with a clear aperture diameter of 70 mm. The plates are made from water-free fused silica, with a matched surface quality of (for  nm, before coating). A series of multilayer dielectric coatings gives the plates a reflectivity R>97% all the way from 1.5 to m. The outer surface of each plate has a broad-band anti-reflection coating applied.

As with most modern Fabry-Perot etalons, the separation and parallelism of the plates is controlled to very high accuracy by piezoelectric actuators, and servo-stabilised with capacitance micrometers incorporated into the etalon itself. The Anglo-Australian Observatory's Queensgate CS-100 servo-controller is capable of maintaining the etalon spacing and parallelism to better than . An IBM-compatible 286 PC rides in the Cassegrain cage, along with an auxiliary electronics rack for communication with both the CS-100 and an etalon translation slide. Commands from the AAO MicroVAX 4000 computer to change the etalon spacing as part of an observing sequence are relayed to the PC by one of the AAO Sun workstations, and thence to the CS-100 via a direct TTL logical level interface, with a typical response time shorter than 1 ms.

A special mounting box has been constructed to go between the Acquisition and Guide unit and IRIS at the Cassegrain focus of the AAT. One side of this box holds a slide, controlled by a stepper motor, which permits remote switching of the etalon in or out of the beam with a positional accuracy of 1 m. The other side of the mounting box holds the polarimetry modules for IRISPOL (Hough et al. 1994), making it possible for polarimetry to be performed in conjunction with the Fabry-Perot if desired. The etalon sits 140 mm above the focal plane of the AAT, in an f/36 beam. This results in a 5% reduction in the unvignetted field of view (compared to placement in the focal plane), but no significant reduction in the spectral resolution, owing to the small beam convergent angle (see also the discussion in Greenhouse et al. 1997).

Besides making access to the etalon easier, the main benefit of placing the etalon close to the focal plane (rather than close to the pupil plane) is that each pixel ``sees'' only a very small part (just 12 mm) of the Fabry-Perot. Operation in this ``pseudo-telecentric'' mode also results in a smaller change in central wavelength across the field, as compared with operation in the pupil plane. Any variations in plate spacing (i.e., departures from flatness) translate into a variation in peak wavelength for that region, rather than an overall decrease in finesse. Any such variations in peak wavelength can be removed in the calibration process. The main drawbacks of placing UNSWIRF outside the IRIS dewar are the increased susceptibility to dust and to changes in the ambient temperature and pressure, and a higher thermal background.

In direct-imaging mode, two optical configurations are available, depending on the choice of re-imaging lens selected within IRIS itself. The ``wide'' mode field of view is a circle 106'' in diameter, with pixels, though the pixel array size of IRIS limits the usable field to just under 100''. In the ``intermediate'' mode, the pixel scale is  pixel, and the full field is available. One limiting factor on the capabilities of UNSWIRF is the availability of blocking filters within IRIS. The standard narrow-band filters are listed in Table 2. Since UNSWIRF is designed to work in order , equation 2 shows that any of these filters having bandwidths are adequate for ensuring that only a single order is passed from the etalon to the detector. Provided neither the continuum nor the night-sky emission is too strong, the broader filters can still be used, though with a corresponding reduction in signal-to-noise relative to a narrower filter.

Finally, it is also possible to insert the H+K échelle grating and a slit in IRIS, and by scanning with UNSWIRF, build up a much higher resolution spectrum of a source placed on the slit than would be possible with the échelle alone. Such a system could be used (with or without the telescope), for example, to investigate the detailed structure of the OH airglow emission spectrum.

© Copyright Astronomical Society of Australia 1997