Instruments of RT-2 Experiment onboard CORONAS-PHOTON and their test and evaluation II: RT-2/CZT payload
Tilak B. Kotoch, Anuj Nandi, D. Debnath, J. P. Malkar, A. R. Rao, M. K. Hingar, Vaibhav. P. Madhav, S. Sreekumar, Sandip K. Chakrabarti
aa r X i v : . [ a s t r o - ph . I M ] D ec Noname manuscript No. (will be inserted by the editor)
Instruments of RT-2 Experiment onboard CORONAS-PHOTON and their test and evaluation II: RT-2/CZTpayload
Tilak B. Kotoch · Anuj Nandi · D.Debnath · J. P. Malkar · A. R. Rao · M. K. Hingar · Vaibhav. P. Madhav · S.Sreekumar · Sandip K. Chakrabarti
Received: date / Accepted: date
Abstract
Cadmium Zinc Telluride (CZT) detectors are high sensitivity and high reso-lution devices for hard X-ray imaging and spectroscopic studies. The new series of CZTdetector modules (OMS40G256) manufactured by Orbotech Medical Solutions (OMS),Israel, are used in the RT-2/CZT payload onboard the CORONAS-PHOTON satellite.The CZT detectors, sensitive in the energy range of 20 keV to 150 keV, are used toimage solar flares in hard X-rays. Since these modules are essentially manufactured forcommercial applications, we have carried out a series of comprehensive tests on thesemodules so that they can be confidently used in space-borne systems. These tests leadus to select the best three pieces of the ‘Gold’ modules for the RT-2/CZT payload.This paper presents the characterization of CZT modules and the criteria followed forselecting the ones for the RT-2/CZT payload.
This work was made possible in part from a grant from Indian Space Research Organiza-tion (ISRO). The whole-hearted support from G. Madhavan Nair, Ex-Chairman, ISRO, whoinitiated the RT-2 project, is gratefully acknowledged.T. B. Kotoch, A. Nandi + , D. DebnathIndian Centre for Space Physics, 43 Chalantika, Garia Station Rd., Kolkata 700 084, IndiaTel.: +91-33-24366003Fax: +91-33-24622153 Ext. 28E-mail: [email protected]; [email protected]; [email protected] (+: Posted at ICSP by Space Sci-ence Division, ISRO Head Quarters, Bangalore, India)J. P. Malkar, A. R. Rao, M. K. Hingar, V. P. Madhav,Tata Institute of Fundamental Research, Homi Bhaba Road, Colaba, Mumbai 400 005, IndiaS. SreekumarVikram Sarabhai Space Centre, VRC, Thiruvanthapuram 695 022, IndiaS. K. ChakrabartiS.N. Bose National Centre for Basic Sciences, JD Block, Salt Lake, Kolkata 700 097, India(Also at Indian Centre for Space Physics, 43 Chalantika, Garia Station Rd., Kolkata 700 084.)Tel.: +91-33-23355706Fax: +91-33-23353477E-mail: [email protected] The RT-2/CZT payload carries, along with three CZT modules, a high spatial reso-lution CMOS detector for high resolution imaging of transient X-ray events. Therefore,we discuss the characterization of the CMOS detector as well.
Keywords
Gamma-ray detectors · X- and γ -ray telescopes and instrumentation · Laboratory experiments · X-ray imaging · Solar flares
PACS · · · · The RT-2 Experiment onboard the CORONAS-PHOTON satellite (Kotov et al. 2008,Nandi et al. 2009) is a dedicated experiment for hard X-ray study of solar flares. It con-sists of three main detector payloads, namely RT-2/S, RT-2/G (both NaI(Tl)/CsI(Na)scintillator-Phoswich detectors) and RT-2/CZT (solid-state imaging detector) alongwith one processing electronics device, RT-2/E. Detailed descriptions of the Phoswichdetectors and processing electronic device are given in Debnath et al. (2010) andSreekumar et al. (2010) and the background simulations of the detectors using GEANT-4 toolkit are presented in Sarkar et al. (2010).The RT-2/CZT payload consists of two different types of imaging detectors: threeCZT (Cadmium Zinc Telluride) detectors and one CMOS (Complementary Metal Ox-ide Semiconductor) detector, arranged in a configuration of 2 x 2 array. The entiredetector assembly (CZT and CMOS) sits below a collimator ( ∼
32 cm height) with twodifferent types of masking devices, namely Coded Aperture Mask (CAM) and FresnelZone Plate (FZP). The payload configuration, simulation and experimental results withthe coding devices are discussed in Nandi et al. (2010). Due to the different dimensionsof the implemented coding devices, this payload has a Field of View (FOV) from ∼ ′ to6 ◦ . RT-2/CZT payload is the only imaging device onboard the CORONAS-PHOTONsatellite to image solar flares in hard X-rays between 20 keV to 150 keV. The CZTdetectors have good spectral resolution but moderate spatial resolution. The CMOSdetector is capable of imaging with a high resolution. However, it is a single channeldevice, and is not capable of generating a spectrum.RT-2/CZT payload is placed outside the hermetically sealed vessel of the satelliteand co-aligned to the Sun pointing axis. In § § § §
5, we make concluding remarks.
Cadmium Zinc Telluride (CZT) is an extrinsic semiconductor X-ray detector with aband gap energy of 1.5 - 2.0 eV and an average atomic number, Z ∼
50. Due to thehigh band gap energy, it can be operated at room temperature. As the atomic number( Z ) is high, even a few mm of CZT can absorb hard X-rays efficiently (Knoll 1999).In comparison to other X-ray detectors like scintillation detector and proportionalcounter, CZT needs a relatively low energy deposition for electron-hole pair formationresulting in a good energy resolution (Kotoch et al. 2008). CZT detectors have othermajor benefits, for example, it can be pixilated by further dividing the sensitive area with segmented anode into multiple pixels which leads to spectral enhancement dueto small pixel effect. The current state of art allows one to develop large area X-raydetectors in the form of a mosaic of CZT pixel devices, making this technology reallyappealing for use in hard X-ray astronomy.CZT detectors already have been successfully used in the Swift satellite (Barthelmyet al. 2005) and in recently launched Indian Moon mission Chandrayaan-1 (Vadawaleet al. 2009). The same detectors would be used in the forthcoming missions like AS-TROSAT, EXIST and Constellation-X. We have used CZT detectors in the RT-2 Ex-periment for hard X-ray solar flare studies. Three CZT detector modules are placed inRT-2/CZT, one of the main payloads of the RT-2 Experiment onboard CORONAS-PHOTON mission, which was launched successfully from Russia on January 30, 2009.The RT-2 CZT detector modules (OMS40G256) were procured from Orbotech Med-ical Solutions Ltd., Israel. The detector module (Fig. 1) consists of a 5 mm thick CZTcrystal having dimension of 3 . × . . Each module is pixilated to have 16 × . × .
46 mm , while the edge pixels have a pitchof 2 . × .
28 mm . The CZT crystal composition (i.e., Cd . Zn . Te) and growth areachieved using the MHB (Modified horizontal Bridgman) technique (Yadav et al. 2005,Jung et al. 2007, Vadawale et al. 2009). It has a density of 5 .
85 gm/cm . The crystal isN-type conductive having electron as a major charge carrier. The electrodes contactsare made of Indium (Lachish et al. 1999). Cathode is a single mono-electrode, whereasthe anode is pixilated to pad size of 1 . × .
86 mm with 0.42 mm gap. The CZTdetector crystal has been integrated with a multi-channel read-out ASIC’s having 128channel which are self-triggered and data driven (Yadav et al. 2005, Jung et al. 2007,Vadawale et al. 2009). Its threshold can be externally controlled. These modules canbe easily mounted and dismounted from the motherboard through two 20-pin surfacemounted connectors in the back side.OMS40G256 detectors are capable of detecting X-rays of energies ranging from10 −
200 keV. In addition to having all digital interfaces, an optional temperature mea-surement sensor is integrated with the front end electronics to monitor the temperatureof the module, which is useful in monitoring the heat dissipation by the ASICs. TheASIC in each CZT module contains its own pixel map information and inbuilt eventstoring memory (FIFO) for up to 256 events. The peak position shift of any sourcespectrum is less than 0.1 keV/ ◦ C. The overall power consumption of a module is about300 mW, which is about ∼
50 % lower than that of the earlier version of the series ofproducts (Yadav et al. 2005, Jung et al. 2007, Vadawale et al. 2009).A detailed study of these detectors was carried out for finding its performance,efficiency, etc. in various environmental conditions and to authenticate its use for thespace environment.2.1 Experimental SetupThe System Development Kit (SDK), developed by Orbotech, was used to test oursample detector modules. This system is available from Orbotech commercially andconsists of two packages: a detector box and a SDK box. The detector box can accom-modate up to 20 detector modules (in a 4 × Fig. 1
A CZT module of dimension 3.96 cm × × and data acquisition from the CZT detector modules (Yadav et al. 2005; Vadawale etal. 2009). The SDK software tool saves the spectrum of each pixel data in ASCII filescontaining the information of event counts per channel.For the tests at low temperatures, we used a regular laboratory cold chamber whichcan actively control temperature between -30 ◦ C to 0 ◦ C with the accuracy of 1 ◦ C. Wehave made the detector box air-tight by covering it with a plastic bag containing twopouches of dried silica gel. The major source of heat dissipation is at the bottom ofthe CZT detector modules. Therefore, we positioned a temperature probe (thermistor)at the cold finger of the detector module in order to control the detector temperature.The chamber temperature is maintained and varied with respect to this monitoringthermistor temperature.2.2 Test ProcedureA test procedure was planned on the basis of detector specification provided by OMSLtd. and desired test environment. The schematic block diagram along with the labsetup of the experiment is shown in Fig. 2.
Fig. 2
Schematic block diagram of the experimental setup (left) and the lab setup (right) forindividual CZT module testing.
CZT detectors were stored in a low humidity control chamber. This chamber waskept in the clean room for mounting or dismounting of CZT modules from the moth-erboard (detector board). Radioactive sources Am and Cd were used for varioustests and calibration of CZT detectors. Three separate power supplies were used forthe OMS detector test unit to provide two analog (called AVDD and AFE) and onedigital (called DVDD) supply voltages along with a common grounding. A high voltagepower supply (HVPS) with -600V is fed at the top of CZT detector crystal. Duringthis period, we monitored each power supply to verify the nominal power consumptionfor each detector (Table 1). A dedicated USB port is used for interfacing between thesystem and SDK unit for data acquisition and to control various parameters of thedetector.On accomplishing all the basic settings required for the test, CZT modules weredirectly irradiated with two radioactive sources (Am and Cd ). Appropriate sep-aration between the sources and detector was kept in such a way that a uniform illu-mination occurs over the entire detector area. The data were accumulated for intervalsof 150 - 1800 seconds as required to obtain a good counting statistics (typically severalthousand photons under one energy peak in the spectrum which gives the statisticalaccuracy of the peak position determination correct to ∼ ◦ C inside the cold chamber. In orderto avoid the moisture condensation, the entire detector unit was kept in airtight con-tainer before placing inside the cold chamber. The temperature of detector box wasmaintained with the help of thermistor placed at the base. Tests were conducted at var-ious temperatures (-20 ◦ C to 20 ◦ C) for characterization (such as, overall performance,energy resolution and stability).
Table 1: Detector power consumption (for one module)Voltage ID Voltage ±
5% Typical current Maximum currentAVDD +3.3 V 30 mA 60 mAAFE +1.5 V 20 mA 40 mADVDD +3.3 V 40 mA 80 mAHVPS -600 V 20 µ A 60 µ A spectrum and events collected by the detector. The information on each pixel can bedisplayed with appropriate color-coding. For example, Fig. 3 shows a sample displaywith the color indicating the pixel quality, quantified by various parameters like thepercentage energy resolution at FWHM (D[%]); the efficiency (S[%]) defined by theratio between counts under the photo-peak ( ± Fig. 3
Display shows the color-coded status of pixels of a CZT module from the SDK set-up.Status of each pixel is based on various parameters (see text, for details).
SDK software application was initially used for data acquisition and to quicklyanalyze the data acquired of the selected module, during the lab test. For a detailedanalysis, however, we used the SDK software to store data in event mode files to be readby a tool we have developed using the Interactive Data Language (IDL). This softwarewas specially designed to analyze spectral data with two calibration peaks (e.g. 59.5keV from Am and 88 keV from Cd ). The detector gain (relation between energyand channel) was assumed to be linear in the 20 - 200 keV energy range, as given byspecifications, and it was found to be satisfactory during the CZT modules laboratorytests with the rms deviation from a linear fit being < F it (fitted). FWHM was also estimated bythe number of channels between the two edges of the peak at half maximum and thisis called FWHM
Cal (calculated).
During the analysis, we also calculated the peak efficiency and peak counts. Thepeak efficiency is defined as the total counts under peak position within 2 σ level dividedby the total counts in the spectrum. We have followed the same procedure for all 30‘Gold’ (module with atleast 97% of pixels having average energy resolution of ∼ ± ± Fig. 4
Variation of different parameters of each of the 256 pixels of one CZT module (No.2789). See text for details. channel whereas the energy resolution given in the term of FWHM
Cal and FWHM
F it is about 9.5 ± ± Table 2: Pixel wise energy resolution distribution of the CZT modules
Sr. Module Disqualified Pixels with energy Average EnergyNo. No. Pixels resolution @59.5 keV Resolution < % % Best 90 % Best 75 %1 2781 2 54 200 10.41 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± From Table 2, it is clearly seen that the average energy resolution of the CZTdetector modules lies around 10.5 ± P ea k C h a nn e l -404 O ff s e t G a i n ( k e V / C h . ) F W H M C a l F W H M F it P ea k E ff . Best 90% Best 75%
Fig. 5
Average value of various parameters along with error bar (standard deviation), for thebest 75% and the best 90% pixels of all 8 CZT modules used in the room temperature tests. pixels of the module which can be recovered by optimizing grounding of the entire testsetup and by proper cooling arrangement for the test unit (mother board).During the analysis, we have classified the pixel depending on their spectral behav-iors in the following way: – Hot (Noisy) pixel:
These pixels are the ones having a very sharp peak or ex-ponentially descending peak (higher count rate) (Fig. 6) compared to the averagepeak from the source or background. In a module, there are some pixels which arevery sensitive to noise either from external sources like inappropriate grounding ofthe test setup or due to deficiency in heat transfer process which leads to heatingof the CZT modules. In Fig. 6, we have shown two spectrum (accumulation time of150 sec) of different pixels of the same CZT module. In order to show the enhancedfeature of noise effect (hot pixel behavior) of two different pixels, we have plottedthe spectra in different channel scale (total channel numbers 1024). These noisypixels can be considerably reduced by taking appropriate action against the originof noise. – Good pixel:
A pixel having a very good efficiency and a Gaussian shaped photo-peak for any radioactive source and uniform background spectrum is defined asa good pixel. Fig. 7 shows the spectrum containing two peaks obtained from theradioactive sources Am (59.5 keV) and Cd (88.0 keV). The spectrum is ob-tained for the period of 150 second, similar to the hot pixel accumulation time. – Bad pixel:
Pixel having poor sensitivity either due to crystal defect structure (dur-ing crystal growth) or improper bonding between crystal and the readout devices Fig. 6
Examples of hot pixel spectra for a Am source. Data is accumulated for 150 sec toplot the spectrum in Counts (Y-axis) vs Channels (X-axis).
Fig. 7
Good pixel: Uncalibrated spectrum from radioactive source Am (59.5 keV) andCd (88.0 keV). Spectral data is accumulated for 150 sec. (ASIC) of the module. Also, it might be due to mishandling of the device withouthaving proper ESD protection. In comparison to normal (good) pixels, the averagecount from the source or background are quite less in number ( <
10 counts per 150sec for photo-peak). An illustration of a bad pixel is shown in Fig. 8. – Dead pixel:
Pixel with zero output (no count) and having no sensitivity to detectX-rays is said to be a dead pixel. There are many possibilities for a pixel beingdead, like crystal defect, weak bonding between crystal and electrodes, mishandlingwithout ESD protection. Unlike the case of a hot (noisy) pixel, there is no chanceof recovery of a dead pixel by any means.2.5 Cooling EffectAt room temperature testing of CZT detector, it was noticed that the modules had anaverage temperature of 5 o C - 6 o C higher than the environment temperature, result-ing in an increase in leakage current which plays a significant role in the response ofeach detector module, especially, on the spectroscopic performance. The key features Fig. 8
Bad pixels: (a) Left panel shows crystal inherent deficiency effect and (b) Right panelshows the ESD negligence effect. to monitor are performance characteristics of CZT detector and the influence of lowtemperature operation over the leakage current of the semiconductor.Low temperature (cold) tests were carried out over six CZT modules with the sameradioactive sources used for previous measurements. A cold test was performed withthe operating temperature ranging from -20 ◦ C to +20 ◦ C and the data was taken whenthe detector operating temperature become stable within ± ◦ C. Fig. 9 illustrates theresults of cold test in terms of average energy resolutions of each module with respectto the operating temperature, for the 59.5 keV emission peak from Am source.These results show clearly that the energy resolution of CZT modules improves withdecreasing temperature down to 10 ◦ C and become saturated at 0 ◦ C. Below zero degreecentigrade, their spectroscopic behavior change toward energy resolution similar to theones measured at room temperature. In Fig. 19, we have also shown the result from aflight CZT module, which also shows the same spectroscopic behaviour.2.6 Life TestWe performed a life test of CZT module to evaluate their response stability and behav-ior over a long time scale operation. The life test was carried out at room temperature.In Fig. 10 (a,b), we show gain variations of each pixel of two CZT modules monitoredfor a period of nine days. The gain of these modules remains almost constant overthis time scale despite the pixels (every pixel has different gain and characteristics fea-tures) getting bad in between for a day and getting back to the normal after sometimewhich is basically due to temperature gradient effect on the modules. These effectswere noticed in both the modules as seen clearly from the sudden rise in the gain ofsome channels (pixels). This effect is controllable with the help of cooling system todissipate heat generated by ASIC. Similarly, a sudden rise in pixel count were founddue to electromagnetic noise from the surrounding picked by the module, which makesthem look like hot pixels. On appropriate grounding of the power supply lines and thebody of test setup, these pixels were found to be good.Further parameters like offset and energy resolution were also under observationduring the life test. The aim of life test was to measure the performance, device struc-tural rigidity and stability with signal communication and power consumption are at Fig. 9
Variation of the average energy resolution (59.5 keV emission peak from Am ) withoperating temperature for six CZT modules. Typical error bar (maxi.) is marked on the figure. their norms. On achieving the life test over the modules, there was no such major in-consistency noticed among the above parameters. This indicates that the CZT crystalproperties remain invariant with time of detector operation. It could be concluded thatthe performance of OMS CZT modules (gold) are stable over a long duration run andhence can be used for space flight.2.7 CZT detectors for space-flight instrumentEight CZT modules (Gold quality) were selected for use in the RT-2/CZT. Out ofthem, 3 modules were selected for the flight payload on the basis of their performance.To qualify these modules for the space, thermal cycling tests were the first to be carriedout for 10 cycles with temperature ranging from -35 ◦ C to 65 ◦ C with the transition rateof 3 ◦ C/min. After carrying out thermal cycling on these modules, there was not anydrastic change found in the over all performance of the modules. After going throughall the space qualification tests, these modules were subjected to life test which wascarried out for 6 days at room temperature.During the last tests, it was found that a couple of pixels became temporarilynoisy. A few pixels were found permanently noisy or dead which were blocked fromthe respective modules. We found a single pixel which become noisy due to factorydefect that is present in every single module, if the threshold is set below 40 keV. InFig. 11, the results of both thermal cycling test and life test carried out on one ofthe flight module is plotted, in which gain of each pixel is monitored and comparedwith respect to each day performance. It was noticed that a uniform pixel gain was k e V / C h . k e V / C h . k e V / C h . k e V / C h . Pixels (Module No: 811)0.2350.24 k e V / C h . k e V / C h . k e V / C h . k e V / C h . Pixels (Module No: 903)
Fig. 10
Life test of two CZT modules depicting the gain (keV/Channel) variation. seen throughout the tests for all the modules, besides a few disqualified pixels. Similarresults were obtained for the remaining two CZT modules for flight use. Fig. 11
Thermal cycling and life test result of a CZT module depicting the variation of gain(keV/Channel) monitored for the period of 6 days.
The RadEye1 CMOS (Fig. 12) is a large-area image sensor (procured from Rad-iconImaging Corporation, USA). It is a complete detection system for high-resolution ra-diation imaging. The effective area of 24.6 mm x 49.2 mm is made up of a 512 x 1024array of sensors, where each sensor module contains a two-dimensional photo-diodearray with 48 µ m pixel spacing. Each pixel has its own charge-to-voltage conversion.These CMOS detectors are basically visible imaging photo-diode detectors. But withthe help of scintillator, these detectors can be used to detect X-rays and other energeticradiation. A Gd O S scintillator screen placed in direct contact with the photo-diodearray, converts incident X-ray photons to light, which in turn is detected by the photo-diodes. RadEye1 CMOS imager provides a fully differential high-speed video signal,which is digitized with 12-bit resolution and transmitted to the processing electronics.Its operating temperature range is around 0 ◦ C to 50 ◦ C with dark current (noise) of ∼ ◦ C (room temperature), which gets approximately double atevery 8 ◦ C increase in temperature.3.1 Test SetupThe implemented CMOS detector is basically a commercial product and as such ithas no specific space qualification. Due to its high resolution imaging capability, wehave decided to use CMOS as a detector onboard RT-2/CZT payload to perform a fineimaging of hard X-ray solar flares. So, the detector functionality was throughly verifiedat different temperatures before subjecting it to the space qualification tests.The detector was enclosed in a plastic bag containing two small pouches of silicagel in order to prevent moisture condensation and with a thermistor mounted on the Fig. 12
RadEye1 CMOS detector. surface of the detector near to its window for monitoring the temperature. The entirepackage was made airtight by wrapping the bag with an adhesive tape. This setup wasmounted on the thermoelectric cooler to cool the detector to the desired temperaturewith a variation of ± ◦ C from the set point. The whole assembly (the package plusthermoelectric cooler) was kept inside a cooled chamber which has the temperatureoperation range from 0 ◦ C to 25 ◦ C. The radioactive source Am was mounted on afixed position above the CMOS detector plane. The detector is connected to the testelectronics (called Shadow-o-Box) using a connector of 15 pins coming out of the sealedpackage. The Shadow-o-Box is powered with 5V supply from the adapter containing theprocessing electronics of the detector and it is connected to the input of the grabbercard which is implemented in the computer based acquisition system. This card isacting as interface between the detector and system.3.2 Test procedureThe Shadow-o-Box is powered with 5V supply to make the CMOS detector operational.Using the system based application provided by the vendor, the data is retrieved fromthe detector to get a raw image of the events. This application needs only once a pixelmap file and offset correction information along with exposure time before getting animage from the detector. We acquired the images of the radioactive source by keepingit at different distance above the detector surface with various exposure times (1s, 3s,and 6s). The temperature of the detector was varied from room temperature (23 ◦ C)to 0 ◦ C by step of 8 ◦ C, in order to check its performance and noise characteristic asfunction of temperatures.3.3 ResultsIt was observed for 6 . . . ◦ C to 0 ◦ C. From the known count rate of the source, weestimate that an energy deposition of about 150 keV per pixel (per sec) is required to clearly distinguish the source from the background, i.e., 150 keV photon will produceapproximately the counts of 15 −
20 per pixel in comparison to the background counts4 − Fig. 13
Histogram of source and background variation at different temperature for 6 secimage accumulation. In the right panel, an image of the Am source is shown as detectedby CMOS photo-diode at operating temperature of 23 ◦ C. Fig. 14
One bit background image (noise) as detected by CMOS7
RT-2/CZT payload consists of 3 CZT modules and 1 CMOS detector. The overalldetector specifications of RT-2/CZT payload are given in Table 3. Note that the geo-metric area of CMOS is restricted to 25 mm ×
25 mm and neighboring pixels in a 2 × µ × µ andcorrespondingly reducing the number of pixels to 256 × Table 3: Specifications of RT-2/CZT payloadDetector type CZT Gd O S + CMOSThickness (mm) 5 3Size (mm) 40 X 40 (3 Numbers) 25 X 25Read out (pixel) 256 x 3 256 X 256Geometric area (cm ) 48 6.3Energy resolution (@59 . Fig. 15
The schematic block diagram of RT-2/CZT.8
The FEB is the most essential part of the entire electronics. Besides having inter-faces with the Motherboard, it has interfaces with the processing unit (RT-2/E). TheFEB contains some common interface circuits. One component is the low noise dif-ferential amplifier with 50 MHz bandwidths which is employed to amplify the CMOSdetector video output. This amplified output is compared with the CMOS thresholdgiven by FPGA to produce one bit image. FPGA is the heart of this board whichcontrols all operations and acquires data from detectors. The SPI interface to CZTmodule is handled by logic inside FPGA. It is interfaced with CZT modules throughLVDS buffers. CZT detector data acquisition and control is done by FPGA throughthis SPI interface.The FEB process all the commands needed to configure the CZT modules of op-erations. The main function of this board is to acquire data from both the detectorsand to form spectrum and image in memory. It has a ping-pong memory bank. Itsends acquired data to RT-2/E after receiving ‘data-send’ command from RT-2/E.The interface circuitry is necessary to read-out CMOS detector video data which is indifferential voltage form. DAC converts a digital threshold value provided by FPGAto analog value. Output of the comparator is considered to be one bit in FPGA foreach pixel. Differential signal conditioning offers many advantages over single endedtechnologies. LVDS signal conditioning centers on 1.25 V with a 350 mV swing anddoes not depend on power supply voltage. As CZT modules are having LVDS lines,they are interfaced to FPGA in FEB using LVDS buffers.The RT-2/CZT payload draws raw input power (from satellite power bus duringflight) from external supply of 27 +7 − Volt. The total power consumption is limited to 7.5Watt. The input power is converted to ±
15 V and +5 V with the help of the MDI unitfor the required supply of the detector and front-end electronics. A non-controllablehigh voltage generator (made by PICO) is used to bias (fixed) the CZT detector with-600 Volt. The 5 Volt supply to CMOS is derived from +15 V. The operation of RT-2/CZT payload is completely commandable and controlled by the processing electronicdevice RT-2/E (Sreekumar et al. 2010).4.2 RT-2/CZT operational modeThere are two different modes in which RT-2/CZT can be operated: the Event Modeand the Normal Mode. Each event registered by RT-2/CZT is characterized by 2 words.The basic data structure is given below (refer to Sreekumar et al. 2010, for details). – CZT Event mode data format
The CZT Event mode consists of 32 bits of data words (referred to as bit numbersD0 through D31) with the following configuration:1. D1 - D0 : Detector ID. 0 to 2 for 3 CZT modules.2. D9 - D2 : Pixel ID (0 to 255).3. D19 - D10 : ADC value of the detected signal.4. D31 - D20 : Time (with a resolution of 0.3 msec).The data stored in memory is sent to RT-2/E unit each second. The maximum numbersof events that can be accumulated are 4032.For the normal mode, data structure for CZT and CMOS detector is given below: – CZT Normal mode data format In the normal mode, CZT spectral and image data are accumulated every secondand count rates are accumulated every 10 ms. A total of 5832 words (each worddata consists of 16 bits) of memory space are allocated for CZT data.Image data are accumulated in 4 energy bands or channels (approximately equallyspaced in the 20 - 150 keV range, though the channel boundaries can be changedby command) leading to 12 images of 256 pixels (for the 3 modules). Apart fromthe image, four channel information i.e. counts are stored separately for 10 ms.Therefore, a total of 12 counters (3 modules x 4 channels) are stored for every 10ms. Spectral data of each module is accumulated in 512 spectral channels.1. Image block (3072 words): 1 K words per CZT, 4 channel X 256 pixels X 1word.2. Spectrum block (1536 words): 512 words per CZT3. Timing blocks (1200 words): 3 CZT detector X 100 timing words x 4 channelsx 1 word (counters in each block will count for 10 ms).4. Counter block (24 words): 12 counters (2 words each)5. VCO block (1 word): 2 bytes6. Special words (8 words): Satellite telemetry word, temperature, Command sent,Data read against command, event number, CMOS line number, Calibrationresult identification word and Calibration status. – CMOS data format
1. Image block (4096 words): 256 x 256 pixels, 1-bit image.2. Sum (512 words): Vertical sum (256 words) + Horizontal sum (256 words)4.3 Test and Evaluation of RT-2/CZT payloadThree CZT and one CMOS detector were selected for use in the flight RT-2/CZTpayload based on performance obtained during the space screening test and satisfactoryresults during test and evaluation of the Qualification Model. The Qualification Modelis a pre-flight payload and its test results are not summarized in this paper. Onlyresults with the RT-2/CZT Flight payload are discussed in the following sections.4.4 Test setupIn the flight condition, RT-2/CZT would be fully controlled by RT-2/E. The overalltesting of all 3 payloads (RT-2/S, RT-2/G & RT-2/CZT) with RT-2/E and GroundCheck-out system is discussed in Sreekumar et al. (2010). We have tested and verifiedthe functionality of RT-2/CZT payload independently, with a computer through anisolator unit, called the OPTO device, SCB-68 connector box and NI data acquisitioncard (PCI 6534). The OPTO device along with the read-out software were developedusing NI LabVIEW platform. The OPTO device is used to isolate the payload electri-cally from the computer using buffers and opto-isolators. The test setup block diagramfor qualifying the RT-2/CZT payload is shown in Fig. 16.In the laboratory, RT-2/CZT detectors were calibrated with two radioactive sourcesAm (59.5 keV) and Cd (88.0 keV). The calibration results, health condition ofthe payload and overall detector functionality are discussed in the following sections.The operation of RT-2/CZT payload in different modes, channel boundary change of Fig. 16
The block diagram of the test set-up used for testing of RT-2/CZT payload. timing data, HV control etc. are commandable. The command structure of RT-2/CZToperation is given in Table 4.4.5 Flight model test resultsCZT detectors are powered with supply voltage of 27 Volt for operation and highvoltage supply of -600 Volt for setting the threshold, whereas CMOS is operated withnormal 5 Volt supply. The overall power consumption of the payload is 6.75 Watt.The RT-2/CZT Flight payload contains three CZT detectors with following serialnumbers and calibrated energy channel information as given below: – CZT-1 (Serial No - 2783) - (116 channel corresponds to 20 keV) – CZT-2 (Serial No - 2789) - (115 channel corresponds to 20 keV) – CZT-3 (Serial No - 2954) - (114 channel corresponds to 20 keV)Health informations of RT-2/CZT payload are fed to processing device RT-2/Ethrough ADC. Eight channel ADC output are shown in the table 5.
Table 5: Health informations of RT-2/CZT payload
Channel No. Description Operating voltage level ± All 3 CZT modules (CZT-1, CZT-2, CZT-3) of RT-2/CZT payload have image andspectral information. Apart from the background image and spectrum, we have exposedthe modules with a radioactive source Am for 20 sec. Each module has detected theemission peak of 59.5 keV along with the calibration source (Co ) peak of 122 keV. The1024 channels source spectrum and 256 pixels image of all 3 CZT modules are shownin Fig. 17. The measured peak channel number versus the emission peaks energy isgiven in Table 6. Table 6: Peak channel informations of CZT modules
Source CZT-1 CZT-2 CZT-3 (Energy in keV) (Channel no.) (Channel no.) (Channel no.) Am (59.5) +0 . − . +0 . − . +0 . − . Co (122.0) +2 . − . +2 . − . +3 . − . Image of CZT-1 and CZT-2 module show that a few pixels (left side of the module)were illuminated with the source and not all the 256 pixels. This is due to the fact thatsource is not properly placed. On the other hand, image of CZT-3 is fully illuminatedand pattern is in circular form. As per design (see Fig. 2 of Nandi et al. 2010), wehave used Coded Aperture Mask (CAM) in CZT-1 and CZT-2 and Fresnel Zone Plate(FZP) in CZT-3 as a coder to cast image on the detector plane (Nandi et al. 2010). Itis noted that during testing a few pixels were found to be noisy, which are dark pixels(shown in the image).
Multiple radioactive sources were shinned over all the 3 CZT modules to calibratethe module and to find energy resolution at 10 ◦ C. For calibration we have used threeradioactive sources: Am (59.5 keV), Cd (22.0 keV and 88.0 keV) and Co (122.0keV). The gain calibration is applied to each pixel of a module during the ground testingof the flight payload. This is done by feeding the input of the calibration file generatedusing IDL code (discussed in section 2.3) from the test data of flight CZT modules.The final spectrum is an integrated one of all the pixels of a CZT module. In Fig. 18,the spectrum of the CZT-1 module with four distinct emission peaks are shown. Thecorresponding channel and energy values of the emission peaks are used for channel-energy calibration in the energy range of 20 - 150 keV. The same procedure is repeatedfor the other two modules. We find that the rms deviation from a linear fit of all threemodules in the energy range between 20 to 150 keV is < Table 7: Calibration specifications of CZT modules
Module Gain Offset Energy res. (%) Energy res. (%) (keV/Ch) (keV) (@59.5 keV) (@122.0 keV)CZT-1 0.201 +0 . − . +1 . − . +0 . − . +0 . − . CZT-2 0.199 +0 . − . +0 . − . +0 . − . +0 . − . CZT-3 0.204 +0 . − . +1 . − . +0 . − . +0 . − . Fig. 17
Spectrum and image of three CZT modules of Flight payload, (a) CZT-1, (b) CZT-2& (c) CZT-3. All modules are irradiated with radio-active source Am and Cd . Eachspectrum is sum counts of all the pixels (256 pixels) of each detector module.
We made a systematic study of Pulse Height (PH) variation of one of the CZT mod-ule (CZT-2) with temperature. Payload temperature was varied from -10 ◦ C to 40 ◦ C.Emission peak of Am (@59.5 keV) is calibrated as PH and its variation is noted atdifferent temperatures. PH variation with temperature is plotted in Fig. 19 (bottom C oun t s / s ec Energy (keV) RT−2/CZT1Cd (22.0 keV)Am (59.5 keV)Cd (88.0 keV)Co (122.0 keV) Fig. 18
Spectrum of CZT-1 module with 4 emission peaks of radioactive sources. Energy ofemission peaks are marked. E n e r gy R e s o l u ti on ( % ) RT−2/CZT2−10 0 10 20 30 40 P u l s e H e i gh t ( C h a nn e l N o . ) Temperature ( o C) Fig. 19
PH and Energy resolution variation of CZT-2 module with temperature. Error valuesare indicated on each data points. panel) and it is noted that maximum variation of ∼ ◦ C.In the top panel of Fig. 19, we have plotted the variation of energy resolution(@59.5 keV) at different temperatures. It is noted that the best resolution of ∼ ◦ C to 20 ◦ C and resolution become worse at40 ◦ C. We have presented the test results of several CZT detector modules along with theperformance of the 3 selected CZT modules for the RT-2/CZT payload. During theroom temperature tests, we observed varied pixels behaviors in terms of response andperformance. The numbers of disqualified pixels also vary from module to module. Werecognized two different types of bad modules whose behavior can be understood eitheron the basis of poor CZT crystal or bad ASIC. CZT module functionality appearsto vary with temperature. This was revealed by both cold test and also during thespace qualification screening procedure. Though the CZT modules can operate at roomcondition, their best average energy resolution is achieved in the temperature rangebetween 10 ◦ C to 20 ◦ C. Beyond these temperature ranges, some of the pixels maybecome noisy showing a large variation in counting rate. The average energy resolutionwhich varies with temperature is not uniform for all the modules. This implies thatany selection of good CZT detector module for the flight payload shall be based on atrade off between the best average energy resolution and number of disqualified pixels.From the results obtained at 15 ◦ C, we concluded that the number of disqualified pixelsshould not be more than 5. The results of various flight screening tests conductedover the RT-2/CZT flight payload shows that the overall performance of CZT modulesand CMOS detector remains invariant even after going through various environmentconditions along with flight electronics. As expected, the flight CZT modules give theirbest performance at 10 ◦ C ambient temperature. In future, a similar work will be carriedout for the development of CZT-Imager payload, ASTROSAT.On 30th January, 2009, CORONAS-PHOTON has been successfully launched andall the RT-2 instruments, including RT-2/CZT-CMOS, are working with the expectedperformance. Details of the on-board calibration and data analysis will be publishedelsewhere.
Acknowledgements
TBK thanks RT-2/SRF fellowship (ISRO) which supported his researchwork. The authors are thankful to scientists, engineers and technical staffs from TIFR/ ICSP/VSSC/ ISRO-HQ for various supports during RT-2 related experiments.
References