NuSTAR
NuSTAR, short for Nuclear Spectroscopic Telescope Array, is a hard X-ray space telescope scheduled for launch in early 2012. At Caltech, we are building and calibrating the focal plane for NuSTAR. My research involves characterizing the flight detectors for NuSTAR and understanding the detector response at sub-pixel scales.
Note: this content is not up-to-date. NuSTAR was launched in June 2012, and as of early 2015, has led to over 100 research papers. More details can be found at the NuSTAR web site.
BACKGROUND
The next generation of space-based hard X-ray (defined here as the energy range from ~10 to several hundred keV) telescopes for astrophysical observations will include both large-area coded aperture imaging systems for synoptic studies over a wide field of view, and focusing telescopes to reach faint flux limits over more limited regions of the sky. Both applications require detectors with good (few percent) energy resolution, and two- dimensional position resolution of a millimeter or better. In addition, space-based telescopes in this energy band benefit from compact detector geometries that can be well shielded from the intense background radiation produced in the atmosphere and spacecraft by cosmic rays. The ability to measure the depth of interaction is also desirable for background rejection. Low power is necessary, especially for large-area applications, and complexity is greatly reduced if focal planes do not require cryogenic operation. The Nuclear Spectroscopic Telescope Array (NuSTAR) mission requires detectors sensitive over an energy range of 5-80 keV, with an energy resolution of better than 1.6 keV. The focal plane must be at least 3.8 cm on a side with pixels smaller than 750 microns. CdZnTe detectors meet all these requirements, and have been selected as the preferred material for NuSTAR.WORKING OF NuSTAR DETECTORS
A NuSTAR detector consists of a 2 mm thick CdZnTe crystal with a continuous cathode plane. The anode is divided into a grid of 32 × 32 square pixels with a 605 μm center-to-center spacing and 50 μm gaps. These pixels are bonded using gold wires to the Application Specific Integrated Circuit (ASIC) chip which contains the read- out electronics for each pixel. A bias voltage of 300 V to 450 V is applied between the anode and cathode. CdZnTe is a compound semiconductor, with the conduction and valance bands separated by 1.6 eV. An incident X-ray penetrates an energy-dependent distance into the crystal before ejecting a K-shell electron. The ejected electron interacts further with other electrons to excite them into the conduction band from the valance band, expending a mean energy of 4.6 eV/electron-hole pair. This results in the formation of a charge cloud roughly 10 μm in diameter. Electrons have a mean free path of ~1 cm in this material, so electrons travel almost freely to the anode under the influence of the electric field. The electron cloud expands to a 50 – 100 μm diameter by the time it reaches the anode plane. The charge deposition raises a hardware trigger in the ASIC, initiating the readout process. The pixel with the highest charge deposition and its 8 nearest neighbors are read out by polarity sensitive channels. The holes have an order of magnitude smaller mean free path as compared to electrons and essentially remain frozen in place during the readout time. They create a mirror charge on the anode plane, which is read out as a signal of opposite polarity on the neighboring pixels. The hole signal is used to correct for depth-dependent charge loss effects. The charge pulse heights on all 9 pixels are telemetered out as data for the X-ray event. The energy of the incident photon is reconstructed from this information using a multi-parameter model. Due to the inherent noise and the uncertainties in our model, we get a FWHM of 1 keV at 60 keV.Techniques for growing homogenous, defect free CdZnTe crystals are still in their infancy. In order to obtain the most uniform material possible, we specify that the sensors have no grain boundaries, and the CdZnTe wafers are imaged in the infrared by our vendor in order to select uniform regions with small and uniform inclusion concentrations. After delivery, the crystals are further screened with X-ray diffraction tomography to reject ones with grain boundaries and other extended defects. In spite of the detailed screening, the final detectors still have minor concentrations of defects like twin boundaries, Te inclusions, and occasional grain boundaries. The presence of these defects alters the performance of the detector by altering charge transport on sub-pixel to multi-pixel scales.
The spectral and position resolution attained by CdZnTe depends on the quality of the crystal used for the detector, pixel geometry, accurate characterization of the detector, and the physical model used to interpolate between characterization data points. Our aim is to carry out detailed characterization of NuSTAR flight detectors.