PASJ:Publ.Astron.Soc.Japan,1–??, (cid:13)c 2011.AstronomicalSocietyofJapan. Solid-state Slit Camera (SSC) on Board MAXI Hiroshi Tomida,1 Hiroshi Tsunemi,2 Masashi Kimura,2 Hiroki Kitayama,2 Masaru Matsuoka,1 Shiro Ueno,1 Kazuyoshi Kawasaki,1 Haruyoshi Katayama,3 Kazuhisa Miyaguchi,4 Kentaro Maeda,4 Arata Daikyuji,5 and Naoki Isobe,6 1ISS Project Science Office, Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, 2-1 Sengen, Tsukuba, Ibaraki, 305-8505 Japan [email protected] 1 2Department of Earth and Space Science, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043 Japan 1 3Earth Observation Research Center, Japan Aerospace Exploration Agency, 2-1-1, Sengen, Tsukuba, Ibaraki 305-8505 0 4Solid State Division, Hamamatsu Photonics K.K., 1126-1 Ichino, Higashi, Hamamatsu, Shizuoka 435-8558, Japan 2 5Department of Applied Physics, Miyazaki University, 1-1 Kihanadai-nishi, Miyazaki, Miyazaki 889-2192, Japan n 6Department of Astronomy, Kyoto University, Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan a J (Received ;accepted ) 9 1 Abstract ] Solid-state Slit Camera (SSC) is an X-ray camera onboard the MAXI mission of the International M SpaceStation. TwosetsofSSCsensorsviewX-rayskyusingcharge-coupleddevices(CCDs)in0.5–12keV band. The total area for the X-ray detection is about 200cm2 which is the largest among the missions of I . X-ray astronomy. The energy resolutionat the CCD temperature of −70 ◦C is 145eV in full width at the h half maximum (FWHM) at 5.9keV, and the field of view is 1◦ .5 (FWHM) × 90◦ for each sensor. The p - SSC could make a whole-sky image with the energy resolution good enough to resolve line emissions, and o monitor the whole-sky at the energy band of < 2keV for the first time in these decades. r t Key words: instrumentation: detectors, methods: data analysis, space vehicles: instruments, X-rays: s general a [ 1 1. Introduction MAXI has two types of X-ray camera, the Gas Slit v Camera (GSC) and the Solid-state Slit Camera (SSC). 1 The Monitor of All-sky X-ray Image (MAXI) mis- The SSC is a CCD camera covering 0.5−12keV energy 5 sion (Matsuoka et al. 2009) is an all-sky monitor (ASM) range. The energy band below 2keV has not been cov- 6 launchedbySpaceShuttleEndeavoron2009Jul16. After ered by ASMs in these decades, which is achieved with 3 . the successful installation onto the Japanese Experiment MAXI/SSC. The moderate energy resolution enables us 1 Module - Exposed Facility (JEM-EF or Kibo EF) on the to make all-sky maps with resolved line emissions. The 0 InternationalSpaceStation(ISS) by astronautsusingthe X-ray detection area of the SSC is unprecedentedly large 1 1 remotemanipulatorsystem,MAXIwasactivatedon2009 to obtain the enough photon statistics. : Aug 3. The initial checkout was also successfully com- This paper describes the design and the results of the v pleted,andMAXIisinthenormalobservationphasefrom ground performance test for the SSC. We overview the i X Mar 2010. In the 1-year observation, MAXI has discov- SSC system in section2. Details of the sensor unit and r eredmanytransientphenomenainX-raybandsuchasX- theonboarddataprocessingaredescribedinsection3and a ray bursts, Gamma-ray bursts, X-ray flashes, star flares, 4, respectively. The results of the CCD screening andthe outbursts of recurrent nova. MAXI now occupies the es- calibrationonthegroundaresummarizedinsection5. In- sential position in monitoring astronomy. orbit performance of the SSC is described in Tsunemi et The SIS (Solid-state Imaging Spectrometer) onboard al. (2010). the ASCA satellite was the first X-ray photon counting CCD (charge-coupled device) camera in space (Burke et 2. SSC system al. 1991). Its energy resolution was good enough to re- solve characteristic K-shell X-rays from heavy elements Figure1 shows the block diagram of the SSC and the such as oxygen, silicon, and iron. The ASCA/SIS opened peripheralsystems. The maincomponentsofthe SSC are a new window of the X-rayastronomy. Since then, CCDs two SSC Units (SSCUs). Figure2 is a photographandan have been standard focal plane detectors of X-ray mis- exploded view of SSCUs. The SSCU is a sensor part of sions, such as Chandra (Weisskopf et al. 2002), XMM- the SSC.One SSCUconsistsofCCD units, preamplifiers, Newton (Lumb et al. 2000), Swift (Gehrels et al. 2004), multiplexers, a collimator and slit unit, and a calibration andSuzaku(Mitsudaetal.2007). FortheX-raymonitor- source. The ISS orbits around the earth like the moon ingwithalargefieldofview,theSXCofHETE2employed in which the ISS always faces the same side towards the X-ray CCDs for the first time (Villasenor et al. 2003). earth. So MAXI is placed such that it always sees the 2 H.Tomida et al. [Vol. , X-rays collimator and slit unit CCD driver / system part la signal processor CCD / peltier devices ngis e langis radiator LHP tnerruc reitlep / Merutarepmeptrueltaipmlepxifelangis evirdierr langis oediv vird oediv tnerruc reitlep erutarepmet poDw cPeop rinen ctlttroeioenrlrfltearroceller / egami Dcommandatad KH SdSaGCtaSE SrC eSin dpCtuea crrfttaiocne C SSCU SSCE C DP SSC main components analog line PDAP digital line power line Fig. 1. BlockdiagramoftheSSCincludingperipheralsystemsandsignalflowamongthecomponents. sky direction. Since one of two SSCUs is placed so as to wafer is front-illuminated p-type CCD operated in full monitor the zenithal sky, it is called SSC-Z. The other frame transfer mode. The pixel number for the X-ray de- unit is called SSC-H which sees +20◦ above the horizon- tectionis1024×1024,andthepixelsizeis24µm×24µm, tal(forwardmoving)directionoftheISS/MAXI.TheSSC giving the detection area of about 25mm×25mm. We Electronics(SSCE)controlsSSCUs. The SSCEgenerates employ two phase clock for the vertical and horizontal CCD drive signals, and digitizes the analog signals from transfers. The weightis 55g. The number of pin is 32 for CCD. The SSCE also controls CCD temperature using the CCD operation, and there are two other electric lines peltier devices embedded in each CCD unit. Heat from inwhichthepeltiercurrentof1.2Aflowsatthemaximum. peltier devices in SSCUs is transferred to fixed radiator InordertoobtainalargeX-raydetectionarea,eachSSCU panels by using a Loop Heat Pipe (LHP). The digitized includes 16 CCD units that are aligned in 2×8 array as datafromtheSSCEaretransferredtotheDataProcessor shown in Figure4. The total X-ray detection area of the (DP). The DP analyzes image data from the SSCE, ex- SSC (32 CCD units) is about 200cm2. tracts X-ray events, edits them into the telemetry data, CCD is sensitivefor opticalandinfraredlights thatde- and sends them to the ground via the ISS/JEM-EF. The gradetheperformancesoftheX-raydetection. Inorderto DP also relays commands from the ground to the SSCE. avoidit, aluminum of 0.2 µm thick is coated on the CCD ThePowerDistributorofAttachedPayload(PDAP)sup- surface, which makes the structure of the SSCUs quite plies electric power to the SSCE in three channels. One simple. The aluminum coat is free from pinhole, which is channelis used forthe CCD operation,andother two are confirmed by illuminating the CCD with optical light. for peltier currents in SSC-H and SSC-Z. Since the SSC does not have an X-ray mirror, the en- ergyrangeoftheSSCisdeterminedbymainlyCCDwafer. 3. SSC Units ThequantumefficiencyforsoftX-rayislimitedbyabsorp- tion at the gate structure (dead layer) and aluminum on The body of the SSCU is made of aluminum. To sup- the front surface of a CCD wafer. That of hard X-ray is press the light reflection, black inorganic anodized alu- limitedbythicknessofthedepletionlayer. Thegatestruc- minum alloys are used for the surface. The SSCU is tures are made of SiO and Si whose designed thickness 2 244×112×124mm, and 3.9 kg including a collimator and are0.8and0.1µm,respectively. Thedesignedvalueofde- slit unit. pletion layer thickness is about 70 µm. Then the energy range determined by CCD is 0.5–15keV where the quan- 3.1. CCD Unit tumefficiencyislargerthan10%. However,theactualen- CCD units used in the MAXI/SSC are fabricated by ergyendofthe SSC coverageis limitedby dynamicrange Hamamatsu Photonics K.K.1 using the CCD, FFTCCD- of analog-to-digital converters. Then the energy range of 4673. A photograph of a CCD unit is shown in Figure3. the SSC including the electronics is 0.5–12keV. The CCD unit consists of a CCD wafer, peltier devices, and a base plate made of copper-tungsten. The CCD 1 http://www.hamamatsu.com/ No. ] SSC onboad ISS 3 Fig. 2. Left is a photograph of SSC units mounted on the aluminum stand before the installation to MAXI. Slit apertures are protected by cover plates (non-flight item). Multi-layered insulator will cover the SSCUs in the final configuration. Right is an explodedviewoftheSSC. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 readout node vertical transfer horizontal transfer Fig. 3. Abird’s-eyeviewofaCCDunitintheSSC. exposed region of the calibration source 3.2. Collimator and Slit Unit (CSU) Fig. 4. The upper panel is a photograph of CCD array in anSSCU.Thegapwidthbetween adjacentCCDsis0.4mm. Each SSCU has a collimator and slit unit (CSU). Thelower panel shows the positionof the readout node, the Figure5 is a schematic view of the CSU. Collimators of direction of horizontal and vertical transfer, and the region theCSUrestrictthefieldofview(FOV)tobeafan-beam. exposed tothecalibrationsource. Thenumbersinthelower ThecollimatorsheetpitchdeterminesthenarrowFOVto panel(0∼15)areCCDIDs. be 3◦.0 in bottom-to-bottom (θ in Figure5). A scan of 1 an object with constant flux forms a triangular response, Thecollimatorconsistsofthinsheetsmadeofphosphor whose full width at the half maximum (FWHM) is 1◦.5. bronze with the thickness of 0.1 mm. There are 24 sheets On the other hand, the combination of 8 CCD units and with 2.4 mm interval that are placed 5.0mm above the theslitdeterminethelongFOVtobe90◦ (θ inFigure5). CCD units. The surface of the collimator sheet is chem- 3 One orbit scan corresponds to 90◦ ×360◦. About 29% of ically etched to suppress the X-ray reflection, and plated the entire sky are not covered by SSC in an orbit scan. with black chromium to suppress the optical light. It takes about 70 days to cover the unobservable region, The slit of the SSC consists of two tungsten bars with dependingontheprecessionperiodoftheISSorbitplane. sharpedges. The width betweenthe twoedges is 2.7mm. In addition, SSC cannot observe the region around the The angular resolution depends on the acquisition angle Sun, which is too bright to block optical light with the which is the angle between the incident X-ray direction aluminum coat on the CCD surface. Then it takes about andthenormaloftheslitplane. Forlargeracquisitionan- a half year to obtain the actual all-sky image. gle, angularresolutionbecomes better, while the effective 4 H.Tomida et al. [Vol. , area becomes smaller by a cosine factor. At the acquisi- is for peltier devices in the SSCUs. Then SSCE including tion angle=0◦.0, the angular resolution(θ in Figure5) is DC/DC converters generates 53 W. The SSCE is main- 2 1◦.5. tained at about 20 ◦C by the fluid loop system on the ISS/JEM. 3.3. Calibration Source Each SSCU has an 55Fe calibration source. The cali- 4. Onboard Data Processing bration source is mounted on the CSU so that small part 4.1. CCD drive of 2 CCDs (CCDID=0,8) in eachSSCU is exposed to Mn Kα(5.9 keV) and Kβ(6.5 keV) X-rays. The circle area in CCD clocking signals are generated in the SSCE. CCD Figure4 shows the exposed region. The main purpose of has two-dimensional pixel array, while the SSC requires the calibration source is to monitor the long-term trend only one-dimensional position information since the SSC of the gain. The fluxes are high enough to determine the conducts a scan observation. Hence, we operate the SSC gainwiththestatisticalerrorof0.1%inone-dayaccumu- in parallel-sum mode, that is, charges in multiple pixels lation. aresummedinserialregisteratthebottomoftheimaging region. Thesummedchargesintheserialregisterarehor- 3.4. Thermal Design izontally transferredto the readout node. The number of The energy resolution and the detection limit at the summedpixelsinobservationcanbeselectedfrom16,32, lowerenergyregionoftheSSCarelimitedbyvariouscon- 64 by commands. We call this number as the “binning” ditions, one of which is a variance of thermally excited parameter that is selected to be 64 for the standard ob- charges(electrons)inCCDpixelduringthe exposureand servation. Thelargerbinninggivesthebettertimeresolu- the chargetransfer. The low temperature operationis es- tion, that provides better angular resolutionin X-ray sky sential to suppresses the thermal noise. As for the SSC, map. The binning of 1, 2, 4, 8 is used only for diagnosis. one-stage peltier devices are used to cool CCDs. Each Thehorizontalpixelclockis8µspixel−1,andthe vertical CCD unit is equipped with peltier devices as shown in clockis100µsline−1forbinning=1. Thevideosignalfrom Figure3. The CCD wafer in each unit is mechanically 16 CCDs in a SSCU is processed serially by one readout supported by 12 pairs of peltier devices (24 posts made electronics. When one CCD is readout, other 15 CCDs of bismuth telluride). There is no other mechanical sup- receive no clocking signal and are in the exposure state. port, which suppress the heat input through the conduc- Since it takes 0.232s to readout one CCD in binning=64, tion. Since eight CCD units are serially connected in the the read cycle for 16 CCDs is 3.719s. Clocking voltages peltier line, there are four electric current lines in SSC, arecommonfor16CCDsineachSSCU.Duringtheexpo- each of which can be controlled independently. For the sure state, CCDs are left in the flip mode (Miyata et al. current control, we have two operation modes; the cur- 2004), in which the bias voltage of CCDs is periodically rent constant mode and the temperature constant mode. inverted from +3V to −9V to reduce the dark current. In the former mode, the constant electric current is pro- vided into the peltier line. In the latter mode, the peltier In the ISS orbit, CCDs are exposed to high flux of currentiscontrolledsothattheCCDtemperatureiskept charged particles, which degrades the CCD performance. at the target temperature. We expect the increase of the CTI, dark current, and the The electrical resistance of the peltier devices in each numberofhotpixels. The degradationoftheMAXI/SSC CCD unit is about 1Ω, and the peltier current in stan- was estimated based on the experiments (Miyata et al. dard SSC operation is about 1A. Hence, peltier devices 2002andMiyataet al.2003). Itis knownthatthe charge of 32 CCDs generate 32W. In addition, 5 W is required injection (CI) technique was a good way to restore the tooperateanalogcircuitsinSSCUs. Thenthepowercon- CTI (Tomida et al. 1997). Uchiyama et al (2009)demon- sumptionoftheSSCUsis37Wthatistransferredthrough strated that the increased CTI due to charged particles the LHP to radiator panels on MAXI. It is designed that couldberestoredintheorbitoperationoftheSuzaku/XIS theLHPandradiatorsystemcouldcoolSSCUstoaround withCI,andconstructedanewmethodofcorrectingpulse −20◦C . The hot sides of peltier devices are thermally heights. We, then,addedthe CIfunctiontotheSSC,and connected to the SSCU body. The peltier devices give could select by command whether or not the CI function the temperature difference of >40◦C between the CCD is applied. wafer and the SSCU body. Hence, CCDs are to be op- The CI function has two important parameters: the erated below −60◦C . The body of SSC-H and SSC-Z amount of charge and the injection period that is deter- are, both, placed on an aluminum stand, then they are mined from the binning parameter. In the case of bin- thermally combined to each other. The aluminum stand ning=N (N>1), charge is injected at every N rows. The is mounted on the GSC-H unit with a thermal insulator signalchargesinN-1rowsthatinclude noinjected charge made of polycarbonate (Figure2). are summed and readout in serial register. Whereas the The total consuming power of SSC is 90 W in typical; injectedchargesaretransferredanddiscardedattheread- 26 W for CCD operation and 64 W for the CCD cool- out node. Since the extra horizontal transfer is required ing. 21 W of 26 W for the CCD operation is consumed in CI, the readout cycle of 16 CCDs is 5.865s, which is in the SSCE, and 5 W is in SSCUs. 32 W of 64 W for 1.577 times longer than that without CI. the CCD cooling is used in the SSCE, and the rest 32 W The charge amount can be roughly controlled by com- No. ] SSC onboad ISS 5 θ 3 collimaotor sheets 2.7mm θ slit 1 2.4mm m 103mm m θ 7 2 .1 9 5.0mm CCD CCDs CCDs 208mm Fig. 5. Schematic view of a collimator and slitunit (CSU) in an SSCU. The leftfigure shows the positional relation between the slitandCCDs. Therightfigure,perpendiculartotheleftfigure,showsthe24collimatorsheets andCCDs. mands through the clocking voltage. The voltage can be the SSC are used for the detailed analysis (spectroscopy changedin0.1 Vstep, whichgivesthe difference ofabout and diagnosis of sensors)while the MIL-STD-1553Bdata 6000 electrons pixel−1 for the injected charge. Since the are to search for the transient phenomena quickly. largeramountofchargegivesthebetterrestorationofthe Each event data of Ethernet telemetry includes X- and CTI (Tomida et al. 1997), we set the injected charge as Y-address (X-ray detected pixel position in CCD), and muchaspossible withthe conditionthatthere is no over- the PHs of adjacent 5 pixels (local maximum, two trail- flow during the transfer. The chargeamount is estimated ing pixels, and two leading pixels). X-address is used to to be about 105 electrons pixel−1. The CTI before the determine the position of X-ray in the FOV, Y-address is launch is negligibly small, so we could not confirmthe CI used to correct PH with CTI. If the number of event is effectinthegroundtest. TheSuzaku/XIShasacapability toolarge,the eventdataarecompressedautomaticallyor tocontroltheamountofinjectedchargeprecisely,whichis by command. In the compress mode, PHs of five pixels utilizedtomeasuretheCTI(Nakajimaetal.2008,Ozawa are replaced with a summed PH according to the pixel et al. 2009). The SSC, however, cannot precisely control pattern called grade. There are four grades (G0, G1, G2, the charge amount. We use the CI method only for the and G3), which is the same to the timing mode of the CTI improvement. Suzaku/XIS. Figure6 shows the four grade patterns. G0 is a pixel pattern that all signal charge (electrons) gener- 4.2. Digital Processing atedby anincident X-rayphotonis confinedinone pixel. The CCD images digitized in the SSCE are sent to the PHofthepixelsadjacenttothelocalmaximumissmaller DP.Weadoptedthedatareductionprocedureusedinthe than the value called split threshold. G1 and G2 indicate timing mode of Suzaku/XIS (Koyama et al. 2007). The that either the leading pixel or the trailing pixel to the darklevel,whichisthetime-averagedpixeldatawithnei- local maximum has more charge than the split threshold. ther incident X-rays nor charged particles, is determined G3 is that both of the leading pixel and the trailing pixel for individual pixels and up-dated at every readout time. to the local maximum have more than the split thresh- The DP searches for the charge deposit pattern of sig- old. The replaced PH of the telemetry in the compress nals, which is called an “event”. The pixel data after the modeisthesumoflocalmaximumandthoseoverthesplit dark-levelsubtractionarecalled“pulseheight(PH)”.The threshold. In the compress mode, we can select events to event search is done by referring to the PH. An event is be downlinked according to the grade pattern. Events of recognized when both of the following conditions are sat- G3 are discardedin normaloperationsince G3 events are isfied : (1) a pixel has a PH between the lower and the thought to be generated by charged particles. upper thresholds defined by command, (2) the PH is a For the MIL-STD-1553Btelemetry, the DP compresses local maximum among adjacent three pixels in the row. everyX-rayeventinto 16bits. 14bits areassignedfor X- The MAXI telemetry data are transferred from the addressandthesummedPH,and2bitsforthegrade. The DP to the ground through two physical networks in the grade definition is the same to that of Ethernet, but the ISS/JEM-EF. One network is MIL-STD-1553B and the gradeselectiontobesenttothegroundcanbeselectedin- other is Ethernet. MIL-STD-1553B has higher reliabil- dependently from the Ethernet data. The bit assignment ity and more real-time connection between ISS/JEM-EF for X-address and PH is determined by command. In the andthe groundstationthanEthernet. However,the data standardoperation, 6 bits are assignedfor X-address and transfer rate of MIL-STD-1553B allocated for MAXI is 8 bits for PH, where the accuracy of the position deter- 50kbps (Ishikawa et al. 2009), while that of the Ethernet mination (X-address) is 0.36mm and that for the energy is 600kbps. Therefore,we designedthat Ethernetdata of is 60 eV. The number of SSC events for MIL-STD-1553B 6 H.Tomida et al. [Vol. , CCD. G0 After the operating voltages were fixed, we measured variousparametersasafunctionoftemperature;thedark local maximum current, the number of the hot pixel, the number of the G1 dead column, the readout noise, and the energy resolu- tion. The CTI was also examined again. Radio active over the split threshold 55Fesourceswereemployedfortheexperiments. Thebin- G2 ning of 1 and 64 are tested. We sorted the flight devices based on the energy resolution, then we selected 16 CCD under the split threshold units having the best energy resolution for SSC-H and G3 SSC-Z, respectively. The readout noise of the selected CCDs ranges from 6 to 10electrons in root mean square transfer direction (RMS) that includes system noise. The readout noise is measured from the fluctuation of PH in horizontal over- clocked region. The values of CTI are < 1.2×10−5 for all Fig. 6. Four grade patterns (G0 through G3) used in the the selected CCDs, where the CTI is defined as the ratio SSCdataprocessing. of the lost charge to the original one in a single vertical transfer. telemetry is limited to be less than 255events for every The hot pixels and the dead columns are troublesome 16lines. since the fluctuation of the dark current is so large de- pending onthetemperature. We, then,checkthe number 5. Screening Test and Pre-flight Calibration of the hot pixels and the dead columns for the selected 5.1. CCD screening CCDs. We defined the hot pixel that the PH with no X-ray is significantly higher than that of the surrounding BeforeassemblyoftheSSCUs,wehadmanycandidates pixels. Thedeadcolumnisdefinedthattheeventnumber of CCD units to be installed. Then we conducted the issmallerthanthatofneighborcolumnsby5σ levelwhen screeningtesttoselecttheCCDunitssuitablefortheSSC. X-ray from 55Fe are uniformly illuminated. The numbers CCDs were driven by using the E-NA system (Miyata et are evaluated in binning=1. The total number of dead al. 2001), and cooled down with a pulse tube cooler in a pixels and dead columns in 32 CCDs were 37 and 15, re- vacuum chamber. The peltier devices were not activated. spectively. About half of the CCDs are free from dead The CCD units for the flight camera are selected from pixels and dead columns. We found that the cosmetics of 64 candidates provided by Hamamatsu Photonics K.K. the MAXI-CCDs is excellent. Miyata et al (2004) described the detailed setup of the Figure7 exhibits the dark current as a function of the experiment and the initial results of 25 CCDs. We report CCD temperature. From the thermal analysis, the CCD the final results here. temperature in the normal operation in the ISS is esti- Atfirst,wedeterminedthe propervoltagestodrivethe mated to be about −60◦C (Section3.4). The dark cur- CCDs. There are 16 voltages for the operation of the rent at −60◦C is about 0.02 electron sec−1 pixel−1 in MAXI-CCDs. The 16 CCDs in each SSCU have to be binning=1. In the normal operation, the binning param- operated with common voltages since we have only one eter is 64, and the readout cycle time is 5.9s with CI-on. set of electronics for eachSSCU. To searchfor the proper The dark current is, then, about 8 electron pixel−1 cycle- voltages,theCTIandtheCIabilityaremainlyevaluated. time−1, which gives us the noise of about 3 electrons as Fromthe experiment,weconfirmedthat14of16voltages a fluctuation in typical. This is smaller than the readout couldbe commonfor 2 SSCUs to obtainthe low CTI and noise, then we concluded that −60◦C was low enough to good CI performance, while we could not operate with operate the CCDs at the beginning of the in-orbit opera- common voltages for two which are used to inject charge tion. correctly. They are ISV (Input Source for the Vertical Since the X-ray intensity of 55Fe sources utilized in the transfer)andIGV2(secondInputGateforVerticaltrans- above performance tests was calibrated well, we could es- fer), which are important to control CI. The structure of timate the detection efficiency at 5.9 keV. Our experi- thechargeinjectiongateisdescribedinthetechnicalnote of the manufacturer 2. In the case of ISV, we found that ments showedthat the detection efficiency of the selected CCD units is >89%at least,while the designeddepletion CCDs were separated into two groups; one is that appro- layerthicknessof70µmprovidesthedetectionefficiencyof priatevoltageofISVis3.0Vwhile the otheris3.8V.We, about 91%. We, hence, confirmedthe detection efficiency then,decidedthatCCDsinthe3.8Vgroupwereinstalled is high enough to cover the energy range up to 12 keV. to SSC-H, and the 3.0V group were installed to SSC-Z. In order to conduct the detailed spectrum analysis, the The appropriate voltages of IG2V are scattered and can- calibrationatthe low energy rangeis alsoimportant. We not be properly grouped. So we designed the SSCE such plantomeasurethedetectionefficiencyfrom0.5to12keV that the IG2V voltage could be set separately for each using the Crab nebula in the in-orbit operation. 2 http://sale.hamamatsu.com/assets/applications/ SSD/fft ccd kmpd9002e06.pdf No. ] SSC onboad ISS 7 Fig. 7. The darkcurrent as afunction of the CCD temper- ature at the screening experiment. The CCD unit employed isassignedtoCCDID=3inSSC-H. 5.2. Energy Calibration The energy calibration of CCD performance with the SSCE was done before the SSC was installed into the MAXI flightstructure. FluorescentX-raysfrom nine ma- terials and Mn-K X-rays from 55Fe sources were used to study the energy response of the SSC. The energy range is 0.52 (oxygen) to 12.5keV (selenium). Figure 8 (left) Fig. 8. PulseheightsdistributionoftheSSC-H/CCDID=3. shows a energy spectrum of various fluorescent X-rays. Inthe topfigure, fluorescent X-raysfromAl,Si,Cl, Ti,Mn, The configuration of the calibration experiment was: (1) Fe, Ni,Znareirradiated. Thevertical axis isinlinearscale. The SSCU to be calibrated is in the vacuum chamber Inthebottomfigure,X-raysfrom55Fe(5.895and6.490keV) areirradiated. The vertical axis is inlog scale, and the unit andthe otherSSCU andthe SSCE arein the atmosphere iscountsec−1 keV−1. (room temperature at 1.0atom). (2) The CSU are re- moved. (3) CCDs are driven with the same operation low 1.0keV in Figure8 is not created by X-ray but by condition as that in the orbit. The binning is 64 and the a background component. The count rate of the com- CI is on. (4) The SSCU in the vacuum chamber is cooled ponent is about 0.1counts sec−1 keV−1 at 0.5keV, while toaround−30◦Candpeltierdevicesarecontrolledsothat the particle background in the orbit is 0.3counts sec−1 CCDs are at the constant temperature. The pulse height keV−1 even at the high cut-off-rigidity region (Tsunemi distribution, energy resolution, and the linearity of the etal.2010). Althoughtheoriginofthiscomponentisstill energy scale are investigated in detail. unclear, we concluded that there is no need to pay atten- 5.2.1. Pulse Height Distribution tiontothe lowenergycomponentbyapplyingthe normal Figure8(right)showsthespectrumof55Fesourceswith background subtraction method. SSC-H/CCDID-3 at −70◦C . The binning parameter is 5.2.2. Energy Resolution 64. The spectrum is created from G0 events only. We We defined the energy resolution as the width of the can see that Mn Kα and Kβ lines are clearly resolved, main Gaussian peak. The energy resolution of Figure8 is but the spectrum shape cannot be represented by two 136eVat5.9keVinFWHM.Figure9showsthedistribu- Gaussians for Mn-Kα and Kβ. Then we fit the spectrum tion of the energy resolution of 32 CCDs at −60◦C and with a model employed in the Suzaku/XIS (Koyama et −70◦C . The average of the energy resolution is 149eV al.2007)thatincludessixcomponents: amainGaussian, for −60◦C and 145eV for −70◦C . Since we apply the a sub Gaussian, a triangle component, a silicon escape same clocking pulses to all the CCDs in each SSCU, we peak, a silicon peak, and a constant component. We, tunedthemsothattheaveragedperformance(energyres- however, could not determine the parameters of the tri- olution) becomes the best. angle component since this component is very small. In The energy resolution of the CCDs depends on the en- this way, we found that the SSC data could be repre- ergyofincidentX-rays. Figure10showstheenergyresolu- sented well with other five components. The increase be- tion ofSSC-H/CCDID-3 at−70◦C as a function ofX-ray 8 H.Tomida et al. [Vol. , 8 D C C6 fo reb4 -70degC m u n2 8 D C C6 fo -60degC re4 b m u n2 135 140 145 150 155 energy resolution Fig. 9. Distribution of the energy resolution (FWHM at 5.9keV)for32CCDsinSSCUsat−70◦C(upper)and−60◦C (lower). Fig. 11. Pulse height of SSC-H/CCDID=3 as a function of X-ray energy. The lower panel shows the residuals between thebestfitmodel(linearfunction)andthedata. 5.3. Performance of Peltier Devices The performance of peltier devices was investigated in the system pre-flight test, where the whole system of MAXI including the SSCUs, the SSCE, and the DP were assembled as flight configuration. MAXI structure was installed in a vacuum chamber, and the structure was surrounded by cooled black panels. Radiator and LHP worked well. Figure12 shows the distribution of temper- ature difference betweenCCD waferandaluminum stand describedinFigure2. Peltiercurrentwassetto1.0A,and thetemperatureofthealuminumstandwas−21◦C. The averagedtemperaturedifferencewas−45.4◦C,whichwas largeenoughforCCDstobekeptat<−60◦CwhenSSCU bodies were at −20◦C . Fig. 10. Energy resolution of SSC-H/CCDID=3 as a func- The annealingis agoodmethodofrecoveringthe CCD tionofX-rayenergy. performance degraded by charged particles (Janesick 2001). Holland et al. (1990) reported that the high tem- energy. G0eventsareusedforthespectrumanalysis. The perature above 100◦C could significantly restore the de- readout noise of 32 CCDs ranges from 5 to 10 electrons gradedperformance. Sincewealsodesignedthatthefunc- in RMS, which is the same level to that measured in the tion of LHP can be halted in the orbit by command, CCD screening. SSCUs could warm up. The SSCU, however, must to be 5.2.3. Energy Scale Linearity kept at<+60◦C . The temperature difference of+40 ◦C Figure11 shows the PH peak of SSC-H/CCDID-3 as between CCD wafer and SSCU body is required for the a function of incident X-ray energies, where PH peak is annealing at 100◦C . We, then, tried the reverse current the center of the main Gaussian of G0 events. We fitted of peltier devices, and confirmed that the reverse current the data with a linear function, and the best fit line is of0.22A made the temperature difference ofabout 20◦C shown in Figure11. We can see that the relation is well . We, then, plan to supply the reverse current of 0.44 A reproduced by a linear function. No significant deviation fortherestoration,whenthedegradationoftheCCDper- can be seen around the silicon K-edge. The residuals is formancebecomessignificantaftertheseveralyearsofop- less than 1.0 % . eration in the orbit. No. ] SSC onboad ISS 9 6. Summary 15 We have prepared the SSC consisting of two SSCUs. Each sensor contains the slit, collimators, and X-ray CCDs. The slit and collimators of each SSCU define 10 the size of field of view to be 1◦.5(FWHM)×90◦. SSC D has32CCDsofFFTCCD-4673fabricatedbyHamamatsu C C Photonics K.K. The SSC has total X-ray detection area of of about 200cm2 with a sensitive X-ray energy range of ber 0.5−12keV. SSC operation has been optimized as a slit m u 5 camera onboard MAXI. CCDs are operated in parallel- n sum mode. The binning number of 64 is standard, but it can be changed by commands. CCDs are readout one by oneineachSSCU.Thechargeinjectionisimplementedto compensatethe performancedegradationbychargedpar- ticle. The working temperature of CCDs is <−60◦C in -50 -48 -46 -44 -42 -40 -38 -36 temperature difference orbitbyusingthepeltierdevices,LHP,andtheradiators. The energy resolution (FWHM) of the SSC for 5.9keV X-rays is 145eV at the CCD temperature of −70◦C and Fig. 12. Distributionofthe temperaturedifference between 149eV at −60◦C . CCDwaferandSSCUbody. ThetemperatureofSSCUbody is−21◦C,andthepeltiercurrentis1.0A. This work is partly supported by a Grant-in-Aid for ScientificResearchbytheMinistryofEducation,Culture, Sports,ScienceandTechnology(16002004and19047001). M. K. is supported by JSPS Research Fellowship for Young Scientists (22-1677). References Burke, B. E., Mountain, R. W., Harrison, D. C., Bautz, M. Fig. 13. An image taken with SSC-H. Mn-K X-rays from W., Doty, J. P., Ricker, G. R., & Daniels, P. J., 1991, tworadio-activesourceof55Feareirradiated. Dataaretaken IEEE Trans. ED, 38, 1069 withbinning=64. Theimageisalsobinnedinthehorizontal Ishikawa,M.,etal.2009,TransactionsoftheJapanSocietyfor direction with binning=64 by software so that the ratio of Aeronautical andSpaceSceinces, SpaceTechnology Japan thehorizontal axisto thevertical isthesameto arealCCD 7, ists26, 43 configuration. Gehrels, N., et al. 2004, ApJ,611, 1005 Janesick, J. R., 2001, Scientific charge-coupled devices (SPIE 5.4. Imaging and Effective Area Press, Bellingham, WA), 836 Koyama, K. et al. 2007, PASJ,59, S23 The X-ray direction of the SSC events is determined Lumb,D. H., et al. 2000 Proc SPIE4140, 2 withthecombinationoftheCCDandtheCSU.Theimag- Matsuoka, M., et al. 2009, PASJ, 61, 999 ing capability of the CCDs was simply checked with the Mitsuda, K., et al. 2007, PASJ, 59, S1 55Fe source in the energy calibration. Figure13 is the Miyata, E., Natsukari, C., Akutsu, D., Kamazuka, T., image showing two bright spots generated by two radio Nomachi,M.,Ozaki,M.,2001,Nucl.Instrum.andMethod, active sources of 55Fe. We find that there is no defect 459 157 region for the X-ray detection. Miyata, E., Kamiyama, D., Kouno, H., Nemesh, N., SSC has a large X-ray detection area, but the effective Tomida, H., Katayama, H., Tsunemi, H., Matsuoka, M., 2004 Proc SPIE5165, 366 area for the X-ray collection is limited by the CSU. We, Miyata, E., et al. 2002, Jpn. J. Appl.Phys. 41, 7542. then, measuredthe effective areabyusinganX-raybeam Miyata, E., et al. 2003, Jpn. J. Appl.Phys. 42, 4564. line. In the beam line, the CSU was installed in SSCU Nakajima et al. 2008, PASJ,60, S1 of engineering model, and the parallel X-ray beam was Ozawa et al. 2009, PASJ, 61, S1 irradiatedintotheCCDthroughthe CSU.As theresults, Tomida, H., et al. 1997, PASJ,49, 405. we confirmed the silt width to be 2.70±0.05 mm, which Tsunemi, H.,et al. 2010, PASJ, 62, 1371 was consistent with the designed value of 2.7mm. The Uchiyamaet al. PASJ,61, 9. collimator sheets reduced the effective area to 89.5% and Villasenor, J. N., et al. 2003, GAMMA-RAY BURST 86.8%ofthedesignedvaluefortheSSC-HandtheSSC-Z, AND AFTERGLOW ASTRONOMY 2001: A Workshop respectively. It was due to the bending and non-uniform Celebrating the First Year of the HETE Mission. AIP alignmentofthecollimatorsheets. Thesefactorsaretaken Conference Proceedings, 662, 33. Weisskopf, M. C., Brinkman, B., Canizares, C., Garmire, G., into account for the astronomical analysis. Murray, S., & Van Speybroeck,L. P.2002, PASP,114, 1