The optical design of dioptric systems working in the infrared is complicated by several factors. The lenses must be cooled to cryogenic temperatures, this prevents using cemented doublets/triplets and the only way to obtain a system with good total transmission is using as few lenses as possible. The choice of optical materials is limited by the fact that "normal'' glasses are quite opaque in the infrared due to absorption bands of water which is normally bound in the glass matrix, and sometimes because of absorption intrinsic to the glassy compounds. For this reason normal glasses were mostly ignored, and the choice of optical materials was traditionally limited to a few crystals such as CaF2, ZnS, ZnSe and BaF2. Prior to the "re-discovery'' of Schott IR glasses (Delabre 1994; Oliva & Gennari 1995) the only glassy material employed in near IR astronomical instrument was IR-grade fused silica. This glass has high dispersion (i.e. is a good Flint) but has a very high partial dispersion (cf. Fig. 1 (click here)) and is not suited to produce good achromatic pairs with any of the Crowns known.
Normal alkaline-earth fluoride crystals have very low
dispersion in the near infrared (cf. Fig. 1).
The main problem faced by those designing IR lenses
instruments is therefore to find a high dispersive
("Flint'') material to couple with BaF2, SrF2 or CaF2.
The condition for achromatism requires that the Flint should have a partial
dispersion similar to the alkaline-earth fluoride crystals,
and this is not satisfied by any
of the crystalline materials commonly used.
However, an excellent
match is provided by the Schott IRG glasses which were recently employed
in several astronomical instruments working
up to 2.5 
m (e.g. Oliva & Gennari 1995).
Unfortunately, the only people who seem interested in IRG2, IRG3
and IRG7 are astronomers, and there is hardly any commercial
profit for Schott to continue the production of these glasses.
Finding alternatives to the IRG glasses is the main aim of this
paper which is also intended to give relatively simple
solutions for the design of very fast cameras for IR spectrographs.
Figure 1:
Plot of the NIR partial dispersion
P=[n(1.5)-n(2.5)]/[n(0.95)-n(2.5)]
versus the "NIR Abbe number''
Achromatic pairs are those with very similar P and large 
.
For clarity, the glass names are without the prefix "SF''
In Sect. 2 we analyze the chromatic characteristics of normal SF glasses, define the best Flints to couple to BaF2, SrF2, and briefly discuss the uncertainties on the variation of their refraction indices with temperature. Representative designs of fast cameras for IR spectrographs are discussed in Sect. 3. New measurements of the IR absorption of SF glasses are presented in Sect. 4.