In -Flight Calibration of the Energetic Gamma-Ray Experiment Telescope (EGRET) on the Compton Gamma-Ray Observatory J. A. Esposito, D. L. Bertsch, A. W. Chen, B. L. Dingus, C. E. Fichtel, R. C. Hartman, S. D. Hunter, G. Kanbach, D. A. Kniffen, Y. C. Lin, H. A. Mayer-Hasselwander, L. M. McDonald, P. F. Michelson, C. von Montginy, R. Mukherjee, P. L. Nolan, E. Schneid, P. Sreekumar, D. J_Thompson, W. F. Tompkins, and T. D. Willis Laboratory for High Energy Astrophysics NASA Goddard Space Flight Center Greenbelt, MD 20771 il II !i In-Flight Calibration of the Energetic Gamma-Ray Experiment Telescope (EGRET) on the Compton Gamma-Ray Observatory J. A, Esposito 1,2,3, D. L. Bertsch 1, A. IV. Chen 2,4, B. L. Dingus 5, C. E. ffichtel 1'6, R. C. Hartman 2, S. D. Hunter 2, G. Kanbach 7, D. A. Kniffen 8, Y. C. Lin 9, H. A. Mayer-Hasselwander 7, L. M. McDonald 20, p. F. Michelson g C. von Montigny n , R. Mukherjee 32, P. L. Nolan 9, E. Schneid 13, P. Sreekumar 1'2, D. J. Thompson 2, W. F. Tompkins 9, and T. D. Willis g 1NASA/Goddard Space Flight Center, Code 660, Greenbelt, MD 20771 email at Coddard: Users identified by initials; dtb, awe, reh, sdh, lmm, rch, or djt. Node: @egret.gsfe.nasa.gov _Universities Space Research Association 3Present address: Research and Data Systems Corporation, Suite 104, 7501 Forbes Blvd., Seabrook, MD 20706; email: esposito_mest.gsfe.nasa.gov 4National Research Council Fellow 5Department of Physies, University of Utah, Salt Lake City, UT 84112 email: [email protected] 6Retired 7Max-Planck Institut ffir Extraterrestrische Physik, Postfach 1003, 85740 Garching, Germany email at MPE: Users identified by initials; gok or hrm. Node: @mpe-garching.mpg.de SHampden-Sydney College, P. O. Box 862, Hampden-Sydney, VA 23943; email: [email protected] 9W. W. Hansen Experimental Physics Laboratory, Stanford University, Stanford CA 94305 mail at Stanford: Users identified by initials; lin, pfm, pin, billt, or tdw. Node: @egretl.stanford.edu 1°Raytheon, NASA/Goddard Space Flight Center, Greenbelt, MD 20771 UScience and Computing, GMBH Hagellocher Weg 71 D-72070 Tiibingen Germany email: [email protected] 1=Dept. of Physics & Astronomy, Barnard College & Columbia University, New York, NY 10027 email: [email protected] 13Northrop Grumman Corporation, Mail Stop A01-26, Bethpage, NY 11714 email: schneid @egret.gsfc.nasa.gov i:l IIi 2 ABSTRACT The Energetic Gamma-Ray Experiment Telescope (EGRET) on the Comp- ton Gamma-Ray Observatory has been operating for over seven years since its launch in 1991 April. This span of time far exceeds the design lifetime of two years. As the instrument has aged, several changes have occurred due to spark chamber gas exchanges as well as some hardware degradation and failures, all of which have an influence on the instrument sensitivity. This paper describes post-launch measurements and analysis that are done to calibrate the instrument response functions. The updated instrument characteristics are incorporated into the analysis software. Subject headings: gamma rays: general - gamma rays: instruments To be appear in the Astrophysical Journal Supplement Series 1. Introduction The Energetic Gamma-RayExperiment Telescope(EGRET) has been operating in spaceon boardthe Compton Gamma-Ray Observatory (CGRO) since the launch and acti- vation in 1991 April. EGRET is one of four instruments on CGRO and is sensitive over the energy range from approximately 30 MeV to over 30 GeV, limited at the high end by low counting statistics. Results from the analysis of EGRET data provide important information for understanding a broad range of astrophysical phenomena, including solar flares, pulsars, aspects of our Galaxy and other normal galaxies, gamma-ray bursts, active galaxies, and the extra-galactic diffuse radiation. A summary of the principal findings is given by Fichtel (1996). The EGRET instrument and its scientific goals have been described by Hughes et al. (1980) and Kanbach et al. (1988 & 1989). The extensive calibration of EGRET prior to launch and during the early part of the mission is described in detail by Thompson et al. (1993), including the calibration at the Stanford Linear Accelerator Center (SLAG), the test for proton-induced background at Brookhaven, and the calibration at the Bates Linear Accelerator. The pre-launch calibration, together with the extensive software developed by the EGRET instrument team, provides the basis for an analysis of the EGRET data, but changes in instrument performance require in-flight calibration information to update the pre-flight calibration results ................. The purpose of this paper is to describe the in-flight calibration of the instrument based on observations made throughout the mission, and to address necessary time-dependent corrections to the EGRET pre-flight kcatibration. Because of the unexpected long lifetime of EGRET, the counting statistics for the diffuse gamma radiation far exceed the statistics obtained during the calibration. The observed intensity of the diffuse emission early in the mission before the instrument performance degraded, and at times following gas refills, serve as a standard for comparing later observations of the diffuse emission and thereby determining the time variability of the instrument response. The point-spread function (PSF) of the instrument is monitored using data from the strong pulsars. For these measurements, only photons whose arrival times are nearly in phase with the pulsar's peak emission are used to increase the signal-to-noise ratio. A brief summary of the principal features of the instrument and its history is given in the next section to aid in understanding the details the in-flight measurements of the effective area and point-spread functions that are discussed in sections 3 through 5. Section 6 describes the reduced field-of-view (FOV) modes that were implemented at the beginning of Cycle 5 to conserve the remaining spark chamber lifetime, and Section 7 provides information on the major source of EGRET background, the Earth-limb albedo. -4- 2. The EGRET Instrument The descriptionof the EGRET instrument andthe CGRO spacecraft will be limited to that necessary for an understanding of this paper. Further details on EGRET may be found in the articles mentioned in the introduction. EGRET is shown schematically in Figure 1. The central element is a multi-level spark chamber that is triggered by a directional scintillator coincidence system. A NaI(Tg) Total Absorption Shower Counter (TASC) is situated below the spark chamber to measure the event energy. The upper portion of the instrument is covered by a scintillator dome that is used in anti-coincidence with the triggering system to veto charged particles. EGRET is similar to, but much larger than, the successful SAS-2 (Derdeyn et al. 1972) and COS-B (Bignami et al. 1975) gamma-ray telescopes of the 1970's. f-CLOSELY SPACED SCINTILLATION SPARK CHAMBERS DOME WIDELY SPACED SPARK CHAMBERS TIM! OF FLIGHT COINCIDENCE SYSTEM l _- PRESSURE VESSEL Nal (TL} _ ELECTRONICS MEASUREMENT COUNTER D "--- GAS REPLENISHMENT SYSTEM Fig. 1 - Schematic diagram of the EGRET instrument. Gamma rays are detected through the electron-positron pair production interaction in tantalum foils interleaved with the spark chambers tracking layers(decks). The electron or positron may trigger the coincidence system, consisting of a 4 x 4 array of plastic scintil- lator tiles situated below the lowermost conversion foil, and a similar array at the bottom of the tracker. Of the 256 possible coincidences of an upper tile and a lower tile, 96 are potentially allowed by the EGRET electronics. These 96 coincidences generally are the ones _I11!: in which the lower tile is directly belowits correspondingupper tile, or its nearestneigh- bor. Four exceptionsarethe cornerdiagonalcombinations,which arenot allowedbecause moresupport structure is in the incident path. The allowedcombinationsof coincidence tiles aredynamically controlledin orbit to minimize the recordingofearth albedogamma rays.Throughout almosttheentiremission,EGRET hasbeenconfiguredto requirea third coincidencesignalfromthe TASC.An eventtrigger isproducedby a coincidencesignal,to- getherwith atime-of-flightsignaturemeasuredbetweenthe twoscintillator planesindicating downward-movingparticles,andlackofasignalfromthe anti-coincidencedomecoveringthe instrument. The anti-coincidencesystempreventstriggeringonchargedparticles,whichare muchmoreintense(x 103or more)than the gammarays. When a trigger occurs, the spark-chamberhigh-voltagepulse is generatedto record the tracks,andthe readoutof the spark chamberandeventenergydata commences.The recordedspark-chamberpicture, energyinformation, gammaray arrival time, andancillary information aretransmitted to the ground asone "event". Becausethere aremany more eventsthan usefulgamma-raydetections,thedataanalysissystemmustselectthe subsetof all eventsthat areunmistakablyrecognizedasgamma-rayproducedpairs, andfrom these extract the arrival direction and energyof eachdetectedgammaray. This procedureis describedin Thompsonet al. (1993). The TASCcalorimeterconsistsof36NaI(Tg) blocks,optically coupledto form amono- lithic scintillator (Hugheset al. 1986). It is viewedby 16photo-multiplier tubes (PMTs) through a light-diffusion box, which helpsto equalizethe amountof light receivedby each PMT. The 16PMTs are divided into two interleavedgroupsof eight. An analogsum of eachgroupofeightPMT signalsisfedinto apair ofpulse-heightanalyzers(PHAs), onefor the low-energyrange(1 - 200MeV) and onefor high energies(20MeV to 30 GeV). The two high-energyPHA resultsaretelemeteredindividually for eachtriggeredevent. The two low-energyPHA resultsareaddeddigitally on-boardandaccumulatedin spectra. The EGRET instrument wasoriginally designedfor a two-yearmission. As of 1998 September,EGRET hasbeenin operationfor more than sevenyears,far surpassingthe original goal. Asthe instrumentaged,a steadydegradationofefficiencyandsomehardware failureshaveoccurred.A chronologicallist ofgasrefills, hardwarefailuresandother major occurrencessincethe activation of EGRET isgivenin Table1. A singletriggering tube in the lower scintillator planefailed during 1994November,reducingthe total effectivearea by -,_ 6%, and it also produced an azimuthal asymmetry in the instrument response. Two additional triggering tubes have shown significant gain shifts after 1997 August 3 and 1998 July 7 but these changes produce only minor decreases in the effective area. The EGRET spark chamber array is composed of two interleaved stacks, called A-stack -6- andB-stack,with separatehigh-voltageandreadoutsystemsinordertominimizethe impact of a componentfailure. The B-stackdisplayeda steadydecreasein sparkefficiencysince EGRET activation, and it failed completelyin 1997November. Becauseof this failure, gapsoccurin the electronandpositron tracksthat complicatethe identification of the pair interaction in manycases,andespeciallyat lowenergies. The gasusedin the EGRET spark chambers(99.5%neon, 0.25%argon, and 0.25% ethane)becomescontaminatedby breakdownof ethaneby spark chamberfirings and to a small degreeby residualout-gassingof contaminants. This leadsto a slowdegradationof sparkchamberefficiencythat alsoreducesthe capability to identify pair interactionevents. A gaspurgingandrefill systemcapableoffivegasexchangeswasincludedin the instrument design. As the spark chamberefficiencydegraded,the sparkchambergaswaspurgedand refilledatthedatesshowninTable1.Thefifth andlastcompletegasexchangewasperformed in 1995September. After the last two gasexchanges,the spark chamberefficiencywas recoveredto a somewhatlowermaximum level. In orderto lengthenthe instrument life, EGRET wasdesignedwith two independentgascirculationsystemsto mix thegasuniformly within the pressurevessel.The first gascirculation systemfailed at nearly threeyearsinto the missionand the secondgascirculation systemcontinuedto work well until its failure in late 1997(seeTable1). By then, EGRET wasbeingoperatedat a muchloweraverage trigger rate sothe mixing wasnot sonecessary. 3. Effective Area Studies The time and energy dependence of the instrument's effective area due to the aging effects discussed above are conveniently described (and applied in analysis) in terms of scale factors that multiply the effective area functions that were determined in pre-flight calibra- tion measurements at accelerators (Thompson et al. 1993). In this context, the true effective area, A(O, _, E, t) is given by (i) A(O,_,E,t) = Ao(O,_,E) x S(E,t) where 0 and _ are the polar and azimuthal arrival angles in the spark chamber, E is the gamma ray energy (or energy range) of the observation, t is the mission time since activation, Ao(O, _, E) is the effective area as calculated from the pre-flight calibration measurements, and S(E, t) represents the derived scale factors. The factors, S(E, t), might also be functions of the arrival angles, especially if 0 > 25°. Efforts to determine if angle is important have been limited by statistics, but the indications are that the dependence is small at angles _! ii