A coded mask for γ-ray astronomy. Design and calibration

Abstract

The high-resolution γ-ray spectrometer (SPI) is one of the two main instruments on board the ESA INTEGRAL satellite successfully launched in October 2002. SPI uses coded aperture mask technique in order to have imaging capabilities at the energy band (20 keV–8 MeV) it will study celestial sources. The SPI imaging performance depends critically on the quality of the coded mask response and also on the precise knowledge of such response function. In this paper we present a general description of the SPI Coded Mask design together with its main features. Scientific impact of INTEGRAL SPI Coded Mask design on the instrument capabilities is also discussed. Results obtained for Mask calibration at different energies and incident angles are presented.

Introduction

ESA's INTErnational Gamma Ray Astrophysics Laboratory (INTEGRAL) [1] is a satellite mission for γ-ray astronomy launched in October 2002. The scientific payload consists of two main instruments, one optimised for spectroscopy (spectrometer SPI) and one for imaging (IBIS). INTEGRAL scientific payload is completed with monitors in the X-ray (JEM-X) and optical (OMC) domain in order to provide multi-wavelength coverage. The recent satellite missions CGRO and GRANAT have made possible a significant improvement in the field of γ-ray astronomy [2], [3], [4], [5]. Nevertheless, the INTEGRAL observatory will provide a large and unprecedented combination of fine imaging and high-resolution spectroscopy over a wide range of X- and γ-ray (3keV–10 MeV) energies including the optical band. Within the INTEGRAL mission the spectrometer SPI is devoted to high resolution (2.2 keV FWHM at 1.33 MeV) γ-ray line astrophysics in the energy range 20 keV–8 MeV. The expected narrow-line sensitivity is about 5×10⁻⁶ photons s⁻¹ cm⁻² at 1 MeV for a 3σ level after an exposure time of 10⁶ s, which represents about 10 times better narrow-line sensitivity than previous γ-ray telescopes. This instrument is equipped with an array of 19 hexagonal high purity cooled Ge crystal detectors 56 mm side-to-side and 69.5 mm long each one, shielded by a massive BGO anticoincidence system [6], [7], [8]. The spectrometer SPI is 238 cm high with a mass of about 1300 kg (see Fig. 1). During flight, cosmic ray interactions in the germanium crystals can degrade their resolution. In order to repair the crystal structure, it is foreseen to heat up the crystals (100°C) for 24 h every 6–12 months, depending on the crystal degradation level (“annealing” process).

The SPI imaging capabilities in the γ-ray regime are achieved by a coded aperture mask that modulates the astrophysical signal. In coded aperture imaging a mask with opaque and transparent elements is placed between the detector and the source [9]. The image results from a deconvolution between the shadow mask produced on the detector plane and the pattern of the mask elements. Exhaustive papers exist which describe mathematical methods for the generation of optimum mask patterns (see e.g. Ref. [10] and references therein). However a discussion on this subject is outside the scope of this work. In the case of SPI a Hexagonal Uniformly Redundant Array (HURA) coded aperture Mask has been chosen. Apart from thermo-mechanical requirements, coded masks to be used in satellite based γ-ray astronomy must accomplish certain requirements from the scientific point of view. These scientific requirements affect directly both to the design and calibration process of the coded Mask. First requirement is related with the absorption of γ-ray on the coded mask open elements (to obtain images in the MeV regime it is not possible to use self-supported coded masks). The coded mask support structure must be as “transparent” as possible to γ-ray, as the instrument sensitivity critically depends on the detection efficiency that is adversely affected by the absorption of incident radiation. Moreover, the coded mask design and manufacturing process also must guarantee the pixel size and positioning accuracy to be within tolerance limits in order to minimise scientific impact on recorded data. Finally, the mathematical process to obtain final images [11], needs the “optical” properties of coded masks to be accurately determined, i.e. the coded mask transmission matrix must be obtained experimentally as a function of the energy and incident angle (the so-called mask calibration).