Optical Design

The PFADC was designed by Richard Bingham at University College London under contract to CTIO. It is important to emphasize that the values shown in Table 1 are those of the nominal design, but of the system as built and measured. The theoretical performance of the final configuration is essentially identical to that of the original design.

Table1: Optical Design of PFADC for PFCCD on 4M Blanco Telescope  

This corrector is a descendant of the triplets originally provided with the Blanco telescope (Wynne, 1968). Addition of a fourth element provides broadband color correction and significantly improves image quality. The basic optical configuration is similar to a 4-element design first described by Wynne (1967) for the Hale 5m. Though the 1967 design is for a classical Cassegrain optical system, Wynne (1987) later showed that it could be adapted for use on a Ritchey-Chretièn telescope.

Wynne and Worswick (1988) then demonstrated that an ADC version could be built by putting a pair of rotating, curved, zero-deviation, Risley-like prisms with an oiled mating surface in front of the basic 4-element configuration. Bingham (1988) soon produced a simpler design in which a pair of rotating ADC prisms with an oiled, flat, rotating contact surface served as the first element of a 4-element corrector. This reduced the number of elements from 8 to 7. Glass-air interfaces were decreased from 10 to 8.

While designing the PFADC for CTIO and a similar corrector for the WHT, Bingham was able to further improve the design by replacing both of the front two elements of the 4-element configuration with doublets having shapes similar to those in the corresponding elements of the basic 4-element corrector. Each doublet is made of glasses (LLF1 and PSK3) which have almost the same indices of refraction but different dispersions. The cemented surfaces of the doublets are slightly inclined, so both act like zero-deviation prisms with a small dispersive power. When the axes of the prisms are 180° out of phase, their dispersions cancel and the system has essentially the same image quality as the basic 4-element design. The final optical system contains 6 pieces of glass and 8 glass-air interfaces. The rotating surfaces are not in contact.

Both doublets can rotate independently over 360°, allowing an artificial dispersion of variable magnitude to be added in any direction. This permits the corrector to compensate for atmospheric distortion with very little image degradation at any azimuth and at zenith angles to 70°. The optical design provides excellent unvignetted images at all wavelengths from 3400A to past 10000A over 48 arcmin field. There is little image shift with ADC. Chromatic effects are small. The quality of imaging at all air masses is primarily seeing-limited.

The four surfaces on the two singlets have been coated with broad-band anti-reflection coatings having high transmission from 3500- 10000A. The four surfaces of the doublets were coated with MgF₂ instead of the broad-band coatings. Use of these new coatings was felt to involve too much risk because their long-term characteristics were not well known. So far they appear to be stable and robust.

Transmission of the corrector including coatings and glasses is 85% or higher at all wavelengths from 3700A to 8700A, falling to 75% at 3650A and 10000A and 54% at 3500A. Excellent BVRI photometry can be done using the PFADC. The short wavelength transmission limit makes photometric calibrations somewhat more difficult in U, though good results have been obtained in this band.

The original design specification called for image quality of .25'' full width half maximum (fwhm) in the center of the field and .5'' fwhm at the edge. The corrector meets this specification. However, the images produced by ADC correctors tend to have irregular profiles which often makes fwhm a misleading representation of image size. In the rest of this paper, we will refer to image size by specifying the diameter of a circle in which 70% of the incident energy is contained (D₇₀). This is a somewhat more stringent specification for image quality than the original. For various reasons, we believe that D₇₀ provides as accurate a quantification of the useful image quality of the instrument as can be provided by a single number. For the purpose of theoretical OFAD modeling, the images are considered to lie at the centroid of the spot diagram.