We describe an instrumental configuration using a Fabry-Perot and a cold grism in KIR to provide a capability of high sensitivity integral field spectroscopy with R = 1500 and FWHM = 0.12 arcsec, at low cost and requiring a minimum of modifications on the instrument.
This document describes an instrumental configuration using a Fabry-Perot and a grism in KIR to provide such a capability at low cost and requiring a minimum of modifications on the instrument.
There is no infrared instrumentation project equivalent to OASIS, at least not in the K band. Paradoxically, this band is rich in spectroscopic features that can provide essential information on the physical conditions of the interstellar medium or stars. Good examples are the quadrupolar lines of molecular hydrogen (H2 v-1-0 s(1)...), ionized hydrogen (Br g, Pfund d etc.) or even of Fe II, a good indicator of supernova rate. If we add that active galaxies (Starbursters, AGN) are usually highly obscured by interstellar dust and visible only in the infrared domain, it becomes clear why a facility offering high angular resolution (diffraction limit) and spectral resolution of about 1500 presents a strong interest, especially if the low flux performances of a cold spectrograph are reached.
The solution we propose consists in associating a Fabry-Perot etalon with a small grism in KIR. The etalon filters a small number of monochromatic images in the K band which are separated spatially by the grism and observable in a single detector exposure. There is a multiple gain compared to simple filter imaging besides a gain in resolution and the flexibility of adjusting the observed wavelength.
The same principle has been used for an instrument called Graf on Adonis (ESO 3.6m). However, in the present case a modest resolution is aimed at which allows to consider a far more simpler optical layout with only a cooled grism and FP etalon at ambient temperature. The f/20 output beam from PUEO only slightly degrades the etalon finesse given the low gap etalon chosen. The filter could be inserted in a filter slot, in a pupil plane which is optimal. A rectangular aperture has to be inserted in a focal plane in order to limit the overlapping of monochromatic images on the detector. This is the only modification which would require opening the dewar, but this could be relatively simple if done through the dewar window.
A compromise must be reached to define the field of view; one can image a high number of very narrow (slit-like) monochromatic images, or a smaller number of rectangular images. The compromise suggested here is to image six monochromatic rectangles of 32"x5". A resolution of 1500 appears feasible.
The FP spacing is set so that the wavelength of interest is transmitted. The gap of the etalon is such that six orders are transmitted within the spectral domain 2.0-2.4 micron (K band).
At the exit of the FP etalon, all six orders are overlapping; it is the role of the grism to spread them apart on the detector. Since the field has been limited by a mask, the monochromatic images do not overlap or contaminate each other. One thus obtains six monochromatic images side by side, simultaneously; generally this could be one image in the light of a line of interest and 5 continuum images. It could be possible to adjust the FP gap so that two line images are obtained (e.g. H2 and Br g) and four continuum.
The advantages compared with conventional monochromatic imagery or room-temperature spectro-imagers are:
For a good efficiency, the airy pattern produce by one line of the grating
must be of the same angular size as the deviation angle for the energy to be
well concentrated in the spectrum. This condition sets the interference order
of the grism. For 
= 2.2 
and 
,one finds that with k= 4, there is no order overlapping
The angle of the prism is determined by the focal of the objective and the spectral domain projected on the detector: if we assume here 6 monochromatic images on the detector (see figure) and according to the diagram of KIR (fchambre= 85 mm) this leads to 235 lines/mm. If the material has a refraction index high enough, the angle of the prism remains reasonable. For example with the index of ZnSe (n= 2,44), a prism angle of 55 degree is found. Such a cryogenic grism appears actually achievable, as it has been confirmed by several potential manufacturers.
The Fabry-Perot should have a finesse of typically 75 for achieving simultaneously the required separation of the monochromatic images and the resolution. Such a figure is difficult to obtain, but appears to be in the range of what is feasible according to the leading manufacturer on the market.
