THE CRRES MAGNETIC ELECTRON SPECTROMETER AFGL-701-5A (MEA) A. L. Vampola P. O. Box 10225 Torrance, CA 90505 (310)375-1012 Introduction The magnetic electron spectrometer on CRRES, also known as the MEA (Medium Electrons A), has a long history. It was originally built in 1968 as a backup for the magnetic electron spectrometer flown on the U. S. Air Force scientific satellite OV1-19. At that time, it was completely checked out and calibrated, ready for flight. When the OV1-19 was successfully launched in 1969, the backup instrument was put into storage. The MEA originally covered the energy range 300 keV to 5.2 MeV in sixteen differential channels, each using a discrete lithium-drifted silicon detector. A seventeenth detector measured penetrating particle and bremsstrahlung background in two separate channels. In the early seventies, the instrument was modified to make measurements on a rocket to be launched in the auroral zone. The modification consisted of a reduction of the magnetic field of the instrument from 2.1 kilogauss to 850 gauss (to lower the energy range of the instrument), a corresponding reduction in the electronic thresholds set on each detector channel, and milling down of the chamber wall to lighten the instrument. With the reduction in magnetic field, the thickness of the yoke could be reduced. The rocket launch location was changed to a low-latitude site where there was no possibility of observing electrons. The instrument was removed from the rocket and placed back into storage. In 1983, it was removed from storage, the electronic calibrations were rechecked, and the instrument was left powered-up for a two-year period during which noise levels in the detectors were checked periodically. Power consumption was only 650 mw, so the long 'burn-in' was innocuous. In 1986, the instrument was calibrated with electrons at the NASA Goddard tandem Van de Graaff accelerator and delivered to Ball Aerospace for integration into the CRRES spacecraft. At the time the instrument was recalibrated at Goddard, testing of suspected count rate limitations in lithium-drifted silicon detectors was also performed. These tests disclosed that some inherent mechanism in the detector itself restricts count rates to the order of 25000 c/s-cm2. For the MEA, this was a limitation of about 40000 c/s per channel, since each of the detectors were 1 cm x 1.5 cm. With the Challenger disaster and the 3 year delay in launch of CRRES, a decision was made to replace the lithium-drifted detectors with ion- implanted silicon detectors which have no such inherent limitation in count rate. Since the new detectors required a bias voltage of -100v as compared to the +75v that the lithium-drifted detectors used, a new power supply was required. Also, the new detectors produced a pulse of opposite polarity from that of the previous detectors, so all of the electronics had to be replaced. Thus, the instrument finally flown on CRRES, while originally built in 1968, retained only the original magnetic chamber, magnet, and collimator. The instrument, as flown, will be described below. Instrument description In a 180 degree magnetic electron spectrometer, particles entering an aperture encounter a uniform solenoidal magnetic field and travel a circular path in the plane transverse to the field. After being bent through 180 deg, the particle is detected by a planar array. First order focussing occurs in the plane. There is no focussing in the vertical plane. The focussing in the transverse plane occurs because the length of a chord subtending angles near 180 deg in a circle does not change rapidly with a change in the angle subtended (the chord is similar to the diameter in length). Thus, the MEA instrument incorporates a sensor design which had had extensive use in low energy nuclear research laboratories through the early years of the atomic age and is very well under- stood. The measurement principle used is momentum analysis in a solenoidal magnetic field. In a uniform magnetic field, the radius of curvature of a charged particle, r, is defined by the charge on the particle, the mass of the particle, the velocity, and the component of the magnetic field perpendicular to the motion of the particle: B/q=mv/r where B is the transverse magnetic field, m is the mass of the particle, q is the charge on the particle, and v is the particle velocity. This equation just equivalences the electric force on the charge q due to motion across a magnetic field B with the centrifugal force on a particle of mass m and velocity v gyrating in a circle with radius r. Thus, in a magnetic electron spectrometer, the energy analysis is done by geometric means and the information derived from the energy deposit in the detector can be used for other purposes. In the case of the MEA, this information is used to increase the efficiency of detection to approximately 100% and reduce penetrating particle backgrounds (cosmic rays, energetic protons in the inner Van Allen zone). Since the instrument is well understood, the geometric factors, the energy responses, and the efficiencies of the individual channels can be determined with very good accuracy through computational means buttressed by judicious tests incorporating electron beams. In the usual case, the magnetic field in the chamber of a magnetic electron spectrometer is quite uniform. The two dimensional angular response can be checked quite accurately and the energy cutoffs of each channel can be determined with a precision that is limited only by the energy spread of the particle beam used in the determination. As a result, data from magnetic electron spectrometers can be used for absolute calibrations. The MEA.BMP file shows a schematic outline of the MEA analyzing chamber. The chamber consists of two halves which were milled out of Armco magnetic iron (a low coercive force oxygen-free material). The low coercive force material provides a relatively low weight yoke with a low fringing field. The permanent magnet utilized is made from Indox V, a ceramic material with high coercive force. A high coercive force is needed to ensure stability in the field intensity throughout launch vibration, temperature variations, rotation in the earth's magnetic field, etc. The Indox V is stabilized by disassembling the yoke after the Indox V has been magnetized in it. The external collimator consists of a series of tungsten apertures held in an aluminum assembly. The internal collimator is entirely tungsten. A disk-loaded collimator is an absolute necessity for an electron spectrometer because of the ease with which electrons backscatter out of material. The disk-loaded collimator acts as a true collimator; a smooth wall in a collimator acts as a funnel. The MEA also incorporates anti-scatter structures within the chamber. The top and bottom of the chamber have aluminum face plates with milled ridges. The sides of the chamber have aluminum fins extending to the working area of the magnetic field. These ridges and fins insure that an electron must under go numerous scatter- ings in order to reach a detector unless its trajectory lies entirely within the collimator acceptance zone. The sole exception is a scattering from the top of one of the ridges or fins. The aluminum fins are coated with a black conductive paint to reduce light scattering to the detectors and to prevent charge buildup on the plates (which would cause unwanted, and uncontrolled, focussing in the vertical direction). At the 180 deg focus, the electrons impinge upon a detector array consisting of six ion-implanted silicon plates mounted in three pairs on a thick circuit card, one of each pair in front and one in back with a window in the circuit card between them. Each of the silicon plates is nominally 1.55 cm wide, 6.05 cm long, and 0.50 mm thick. Each plate has six metallized areas nominally .95 cm by 1.45 cm with a 0.5 mm separation. Corresponding metallizations on the front and rear silicon plates are tied together electrically to form a single detector with a nominal thickness of 1 mm. Thus there are a total of 18 de- tection channels in the array. Each detector has a separate electronics channel which amplifies the pulses and passes the pulses through a window discriminator (one having a lower and upper threshold to define valid events). Pulses with amplitudes below the lower threshold are considered noise or bremsstrahlung and are rejected. Pulses with amplitudes above the upper threshold are due to highly ionizing particles (or long path length traject- ories) and are rejected as unwanted background. In general, the lower thresh- old is set at approximately 50% of the minimum energy electron that can be focussed upon the detector and the upper threshold is set at 110% of the maxi- mum energy electron that can be focussed on the detector. The low threshold ensures efficient detection of electrons which backscatter out of a detector after depositing only part of their energy. The upper threshold ensures detection of valid events in the presence of noise or low energy bremsstrah- lung which add to the pulse heighth. For more energetic electrons, the lower threshold is set at the energy corresponding to a minimum ionizing particle traversing a minimal path through the detector. This assures efficient detect- ion of energetic electrons which pass through the detector with little scatter- ing. Table 1 provides a list of the channels with the various energy boundar- ies and electronic thresholds. The detector closest to the aperture is shield- ed from direct electron access. In the original instrument, small energy deposits in this channel were interpreted as being due to bremsstrahlung and large deposits as being produced by penetrating protons or cosmic rays. Since the new detector array had an eighteen-channel capacity, the bremsstrahlung channel was deleted and its telemetry used for a seventeenth electron channel. Channel Response The following table contains the geometric-energy factors for each of the channels. These geometric-energy factors are based on the laboratory calibration data obtained from the MEA just prior to final delivery in January, 1990. The data were obtained using a collimated electron beam consisting of a Sr90 source viewed through sets of collimators and a large bending magnet (90 deg) for energy selection. The energy response of each channel at 0 deg incident beam was matched to numerical calculations of expected energy response for a 180 deg magnetic spectrometer with the collimator geometry of the CRRES instrument. This was necessary because the field in the magnet chamber was non-uniform (the result of a small chip of Indox V magnet material being flaked off during a cold soak at -60 C). Prior to the cold-soak (and the fracture of the magnet from the yoke, due to differential thermal contraction), the chamber magnetic field was about 850 gauss and quite uniform. After rebonding the magnet pole piece to the yoke, remagnetizing (in a somewhat non- uniform field), disassembling the yoke to stabilize the magnet, and finally reassembling, the field varied from about 700 gauss to 850 gauss at various positions within the chamber. Thus a detailed recalibration was required. The calibration data were obtained after final "buttoning-up" of the spectro- meter with the flight detector array and used the flight data processor to obtain data simultaneously in all channels at all test energies. The geometric-energy factors are calculated factors based on test data taken at intervals of 10 to 30 keV over the range 95 to 1739 keV and have an esti- mated accuracy of about 1%. The energy response is accurate to 1 or 2%, also, provided the energy of the input beam was accurate to within 1%. The input beam was calibrated with conversion electrons from a radioactive source at a number of energies. The ultimate limit on knowledge of the energy calibrations is due to the finite widths of the collimator apertures within the bending magnet, which translate into finite widths of energy spread in the electron beam (varying from about 10 keV at the lowest test energies to about 35 keV at the highest energies). The energy profile of the beam was roughly rectangular (sharp cutoffs in energy at both the low and high sides). Finally, the geometric-energy factor calculations assume ~100% efficiency in the detection process. This is known (by laboratory test) to be a reasonable assumption. Table 1 Channel Response Nominal Ch Emin,abs Emin Ecenter Emax Emax,abs GEF Beff ThL ThU 0 78 110 153 188 240 5.88 860 38 250 1 136 174 214 257 310 5.68 818 73 350 2 190 230 271 314 365 5.16 765 115 400 3 255 297 340 384 435 4.84 740 160 500 4 330 374 418 462 513 4.59 727 200 600 5 421 465 510 553 608 4.19 710 275 750 6 512 558 604 649 701 3.89 713 330 900 7 598 646 693 738 790 3.58 710 380 1000 8 686 735 782 829 881 3.30 707 400 1100 9 778 828 876 923 975 3.08 707 400 1200 10 877 928 976 1024 1076 2.89 709 400 1300 11 989 1042 1090 1139 1191 2.66 706 400 1450 12 1078 1131 1178 1227 1278 2.49 702 400 1550 13 1185 1239 1287 1337 1389 2.37 707 400 1650 14 1268 1322 1370 1419 1470 2.23 700 400 1750 15 1368 1423 1470 1520 1570 2.14 702 400 1900 16 Background 200 2000 17 1478 1534 1582 1633 1684 2.03 706 400 2000 The ThL,U are the lower and upper electronic thresholds set on each channel pulse height discriminator. Emin,abs is the lowest energy electron that can reach the detector without being scattered into it. Emin is the nominal lower energy of the channel. The response at this point is 10% of the peak response. The Emax,... are similar maxima. Ecenter is the center of the response in that 50% of the GEF is above and 50% below this value. The peak response of the channel is very close to this value (within 1% or 2%). The nominal cutoff energies are determined in the following manner: The energy at which the peak response occurs is determined. The high and low energy cutoff values (10% of the peak value) are determined. A linear least-squares fit is made separately to the response between Emin,10% and Ecenter and between Ecenter and Emax,10%. The zero intercepts of these fits are then listed as the Emin,nom and Emax,nom. The GEF is the integral under the original curve. For some purposes, a GF is used which is generated by constructing a rectangle between the lower and upper nominal bounds which has the same area as the integral under the original curve. This GF is simply GEF/DE, where the DE is (Emax,nom)-(Emin,nom). The energies are given in keV and the GEF is in cm2-ster-keV. Thus, the counts/second must be divided by this number to transform to flux. Note that the counts in the data stream are counts per 0.512 seconds and are in compressed format: CRRES MEA 12-bit Floating Point Code: MSB LSB 11 10 9 8 7 6 5 4 3 2 1 0 ____________ __________________________ X Y 0,1 Count = Y 2-7 =(2**(X-1))*(512+Y) Note that the count is truncated. For X>1, add 2**(X-2) to avoid truncation error. For a given compressed count, the following boundaries exist: Minimum count = 2**(X-1)*(512+Y) Maximum count = 2**(X-1)*(513+Y)-1 Mean count = 2**(X-1)*(512.5+Y)-.5 The background count should be subtracted from the raw counts before conversion to flux. The flux in channel i is: Flux(i) = (counts(i)-background*bkgdcoef(i))/GEF(i) Angular Response of the CRRES MEA: Collimator Limiting Angles Nominal Actual HORIZONTAL (normal to the spin plane) 4.92 11.58 VERTICAL (in the spin plane) 6.41 16.20 The actual collimator angles are larger than the nominal angles due to the finite length of the collimator. The actual instrument limiting angles in the spin plane (due to detector location in chamber) are: Channel Max Angle 0 8.24 1 6.37 2 5.19 3 4.38 4 3.78 5 3.24 6 2.90 7 2.63 8 2.40 9 2.21 10 2.05 11 1.88 12 1.76 13 1.66 14 1.56 15 1.48 17 1.41 All of these values are half-angles.