An image slicer spectrograph for the SNAP mission

The Supernova Acceleration Probe (SNAP) mission is dedicated to dark energy measurements with two complementary probes: weak lensing and supernovae.¹ It will provide major improvements in determining cosmological parameters for dark energy. SNAP is a 2m-class telescope with a wide field imager covering the visible spectrum at .35μm and IR to 1.7μm. SNAP will collect data on 2000 Type Ia supernovae and survey 4000 degrees for weak lensing studies. SNAP includes an integral field spectrograph essential for characterizing the 2000 supernovae by determining the redshifts of the associated galaxies and registering calibration stars.

The instrumentation for this mission should be well adapted to the space environment (small, compact, and lightweight). To see faint supernovae and galaxies, we need a low spectral resolution (δλ/λ) covering the visible and the near-IR range, with very high optical and detector performance (the main limitation is the 2m telescope diameter). The supernovae and galaxies should be imaged together, and an enlarged field of view would provide a large sample of galaxy spectra. Finally, very high precision measurements are needed, with instrument calibration at the percent level.

A 3D spectrograph (e.g., an integrated field spectrograph) reconstructs the data cube containing the two spatial directions X and Y plus the wavelength direction, providing the spectrum for each spatial pixel in the 2D field. A large field of view results in less stringent pointing requirements. The image slicers minimize optical losses and improve the efficiency and compactness of the system.² Figure 1 illustrates the image slicing concept. The field of view is divided into N (for SNAP, N=60) strips on a slicing mirror. Each of N slices re-images the telescope pupil, so there are N images in the pupil plane. Because of a tilt adapted to each individual slice, the N images are arranged along a line and form a pseudo-slit.

Figure 1. The principle of the image slicer. The field of view is sliced into 60 strips on a slicing mirror, creating 60 telescope pupil images in the pupil plane. Because of a tilt adapted to each slice, the images are arranged along a line and form a pseudo-slit.

The concept is a classical spectrograph with a prism for constant low resolution (δλ/λ∼100) and high efficiency.³ To increase its performance in the IR, the configuration is undersampled. A 10(×)10arcsec field of view is imaged by a pickup mirror on the slices. The slicer cuts the field of view into 60 strips, each of which is imaged on focusing mirrors, piling the 60 slits onto a long entrance slit. Then the spectrograph images the entrance slit on the detector after it passes through a dispersing prism. The spectrograph illuminates two detectors, one for the visible band (0.35–1μm), and one for the IR (1–1.7μm). The total size of the spectrograph, including the detectors, is 230(×)250(×)200mm, and the weight is 10kg.

To validate the concept developed for SNAP, an optical bench using a new slicer prototype (see Figure 2) has been manufactured to evaluate the optical performance and to test the accuracy of the spectrophotometric calibration. The demonstrator (see Figure 3) is a complete spectrograph with a slicer designed to SNAP specifications.⁴ It was tested in a cryogenic environment in the IR using a last-generation mercury cadmium telluride device of 2(×)2k from Rockwell.

Figure 2. A new slicer prototype module, designed for testing the concept.

Figure 3. A complete spectrograph, including a slicer designed to SNAP specifications, was tested in the IR in a cryogenic environment to demonstrate the concept.

The spectrophotometric analysis includes photometric and wavelength measurements, and requires procedures to correct optical distortions, diffraction losses, and the slit effect. Because of the slices, almost no flux is lost in the entrance of the pseudo-slit: any loss can be recovered on adjacent slices. This provides a powerful correction to the slit effect and helps calibrate the photometry. We have developed an original procedure to recover an unknown position by direct adjustment of the image on a reference library that describes the pixel response of a well-known source on a grid. This method is especially adapted to slicers, since the flux repartition on the detector varies rapidly with the position in the slice. The calibration is further improved by the implementation of spatial dithering, and has been shown to reach a precision better than 1%, even with a subsampled configuration.⁵

The SNAP project has designed a specific demonstrator, reproducing the spectrograph concept and including a new slicer unit. This slicer has been shown to have the required optical properties and can be calibrated to the specified accuracy. This represents significant progress in the use of this technology for precise measurements in the future.

The demonstrator has been funded by the French National Center for Scientific Research/National Institute of Nuclear and Particle Physics (CNRS/IN2P3) and CNRS/National Institute of Sciences of the Universe (CNRS/INSU); the French spatial agency, National Center of Space Research (CNES); and the University of California.