High-Energy Trapped Particle Environments at Jupiter - IEEE Xplore

IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 52, NO. 6, DECEMBER 2005

2281

High-Energy Trapped Particle Environments at Jupiter: An Update Insoo Jun, Henry B. Garrett, and Robin W. Evans

Abstract—The particle data measured by the energetic particle detector (EPD) and heavy ion counter (HIC) on board Galileo were used to update the radiation environments at Jupiter: EPD data for trapped electrons and HIC data for trapped carbon, oxygen, and sulfur ions. The models developed in this study were successfully used to generate the total dose and single event effect environments for a sample mission to Europa. Index Terms—Jupiter, Galileo energetic particle detector, Galileo heavy ion counter.

I. INTRODUCTION

J

UPITER is a popular mission target for NASA missions because of its importance in astrobiological and planetary sciences. For example, recent observations and analyses by the Galileo spacecraft strongly suggested that Europa, Ganymede, and Callisto-the three largest moons of Jupiter-have global liquid water oceans beneath their icy crusts. Also, there is spectral evidence for salts and organic materials on their surfaces, and geologic evidence that the Europan ocean may have been in contact with the surface in the geologically recent past (less than about 100 million years). These findings have been one of the most exciting discoveries in solar system science in the last decade, and in particular Europa was identified as one of the top priority targets in the Decadal Survey of Solar System Exploration [1]. A major design factor that makes Jupiter missions difficult is the intense radiation field at the planet. Like the Earth, “Van Allen” radiation belts exist at Jupiter. For comparison, Jupiter is roughly 10 times the size of the Earth while its magnetic molarger. As the magnetic field at the equator is ment is proportional to the magnetic moment divided by the cube of the radial distance, the Jovian magnetic field is proportionally 20 times larger than the Earth’s. Consequently, the energy and flux levels of trapped particles in the Jovian magnetosphere can be much higher than those at the Earth or in interplanetary space. A good understanding of the Jovian radiation environment and reliable models are necessary if future missions to Jupiter are to succeed. In the absence of reliable radiation models, conservative designs that assure reasonable prospects of mission success Manuscript received July 8, 2005; revised September 30, 2005. The research described in this paper was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. I. Jun and H. Garrett are with the Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109-8099 USA (e-mail: insoo.jun@jpl.nasa.gov; henry.b.garrett@jpl.nasa.gov). R. Evans is with the Gibbel Corporation, Montrose, CA 91020 USA (e-mail: robin.w.evans@jpl.nasa.gov). Digital Object Identifier 10.1109/TNS.2005.860747

are required necessitating more massive shielding of electronic systems or sensors than actually required or even practicable. The Galileo spacecraft has orbited Jupiter since 1995 ultimately completing 34 orbits (Fig. 1) culminating in Jupiter re-entry on the 35th orbit on 21 September 2003. The extensive scientific data returned from the spacecraft have improved our understanding of the Jovian radiation environment. Particularly in this paper, we describe the latest results of the effort to update the trapped particle radiation models at Jupiter using the data measured by two instruments onboard the Galileo spacecraft, namely Energetic Particle Detector (EPD) [2] and Heavy Ion Counter (HIC) [3], for electrons and heavy ions, respectively. II. OVERVIEW OF EPD angular covThe Galileo EPD instrument [2] provides erage and spectral measurement for ions, for electrons, and for the elemental species helium through iron. The EPD consists of two telescopes, mounted on a platform that can be stepped in look angle. The two bi-directional telescopes are called the Low Energy Magnetospheric Measurement System (LEMMS) and the Composition Measurement System (CMS). Of these two instruments, the LEMMS provides the information most directly applicable to the high energy, trapped radiation environment and was the focus of this study. The LEMMS detector head is a double-ended telescope containing eight heavily shielded detectors providing measurements of electrons from MeV, and ions from 22 keV to MeV, in 15 keV to 32 rate channels. Of the LEMMS channels, the most important ones for this study are the electron channels F1, F2, F3, B1, MeV (Table I). DC2, and DC3 for energies above While the count rate data are of great interest in analyzing the statistical variations of the Galileo data, ultimately the particle flux in scientific units is the desired quantity. To convert the EPD count rates to flux units, a geometric factor versus energy is required. The geometric factors for the F1, F2, and F3 channels were obtained from [4] and [5]. On the other hand, the B1, DC2, and DC3 energy channels were not as well calibrated as desired before the launch of Galileo and the geometric factors computed for these channels [2] did not account for the intrinsic efficiency of the detector. Therefore, a series of MCNP/MCNPX radiation transport analyses [5] were performed on the EPD design to obtain the detailed energy-dependent geometric factors for the B1, DC2, and DC3 channels. The geometric factor corresponding to each channel is the energy-dependent detector response function that relates the incident particle fluxes to instrument count rates. The geometric factors obtained in this way for the B1, DC2, and DC3 channels are depicted in Fig. 2.

0018-9499/$20.00 © 2005 IEEE

2282


Fig. 1. Polar view of the Galileo orbits fixed in local time relative to the Sun. Millennium Mission, etc.) are indicated by shading.

The different mission phases (e.g., Prime Mission, Galileo Europa Mission, Galileo

TABLE I DESCRIPTION

OF THE LEMMS HIGH-ENERGY ELECTRON CONSIDERED IN THIS STUDY

CHANNELS

III. OVERVIEW OF HIC The Heavy Ion Counter (HIC) [3] is a modified Voyager Cosmic Ray System (CRS) instrument. HIC is composed of two solid-state detector telescopes called Low Energy Telescopes or LETs. The LETs are standard dE/dx versus residual energy instruments using a series of solid state detectors to make measurements over a broad energy range. On HIC, the combined MeV/nucleon. energy ranges of the two LETs are One of the two HIC LETs (LET E) has thicker detectors optimized for nuclei from carbon to nickel with energies of 15–200 MeV/nucleon with a thick window that shields the detectors from low-energy protons. Although this also excludes lower-energy oxygen and sulfur ions, these ions are measured by the second HIC LET (LET B) that has a thinner window and a threshold of 6 MeV/nucleon in the normal operating mode. The LET B telescope contains 4 solid state silicon detectors that measure the energy deposited by incoming particles. The LET E telescope has 5 detectors. Particle species identification is carried out using energies deposited in each detector. HIC can produce individual spectra by recording multiple events within a polling cycle. These spectra can then be binned by time or location.

Fig. 2. Geometric factors estimated for the EPD B1, DC2, and DC3 channels. “p” stands for proton response and “e” for electron response. Note that B1(p) is 0. Although designed to measure electrons, the DC2 and DC3 channels also respond to high energy protons.

IV. GALILEO INTERIM ELECTRON ENVIRONMENT (GIRE) MODEL The Galileo EPD provided a many-fold increase in coverage, both spatial and temporal, over the data that went into the previous Divine model [6]. The electron model developed using the Galileo EPD data is an omni-directional, equatorial model and covers the Jovian equatorial plane for the range 8 to 16 Rj (1 km). Ten-minute averages of the EPD data have been used to form an extensive database of observations of the Jovian radiation belts between Jupiter orbit insertion (JOI) in 1995 and 2003. Note that a careful examination of the EPD data shows that the sun’s influence on the data

JUN et al.: HIGH-ENERGY TRAPPED PARTICLE ENVIRONMENTS AT JUPITER

2283

TABLE II FITTED ELECTRON DIFFERENTIAL FLUX PARAMETERS CORRESPONDING TO (1). UNITS ARE (CM -S-MEV) . THE MODEL RESULTS PRODUCE OMNIDIRECTIONAL FLUXES

to MeV. Values at intermediate L-shells to those from listed in Table II are derived by first determining the flux at the desired energy at the two L values which bracket the desired L value and then linearly interpolating between the logarithms (base 10) of these flux values. L is McIlwain’s magnetic shell parameter. We used the VIP4 Jupiter magnetic field model [9] to compute the L-shell values. VIP4 used the magnetic field data from Pioneer 11, Voyager 1, and Ulysses, constrained by the locations of Io flux tube (IFT) footprints on the planet measured by the Hubble and ground based telescopes. Estimated errors of or 2 in latitude the IFT footprint observation are typically and in longitude. See [9] for the complete description of the model. A complete model of the Jovian radiation environment requires a definition of the omni-directional flux as a function of energy, B, and L or, equivalently, the energy, L, and pitch angle distribution (i.e., the particle flux relative to the magnetic field direction) of the flux at the magnetic equator. The problem is that most of the EPD data are averaged over pitch angle (hence the terminology “omni-directional”). In principle, it is possible to de-convolve the pitch angle distributions from the omni-directional fluxes if the fluxes are known at all B values along an L-shell. Unfortunately, the Galileo spacecraft’s orbit is very close to Jupiter’s equatorial plane and provides little variation in B along L in the radiation belts. This was anticipated by the mission planners and the EPD instrument is designed to take detailed pitch angle measurements relative to the magnetic field vector in the high time resolution. These data are being carefully analyzed and a catalog is being developed. Unfortunately, the detailed analysis is not complete so can not be included in the current model, and indeed, coverage in L-shell is fairly limited. Fortunately, five other spacecraft have flown through the Jovian radiation belts at sufficiently high magnetic latitudes and resolution to allow approximations to the off-equatorial pitch angles. Specifically, the Pioneer 10 and 11, Voyager 1 and 2, and Ulysses all made such measurements. While our analyses of the Ulysses data were inconclusive (the radiation instruments were placed in a protective mode during passage through the regions of interest, and no de-convolution is available), the other four data sets have been incorporated into the Divine radiation model [6]. Pending the analysis of the Galileo pitch-angle data, the pitch angle variations estimated for the Divine model can be incorporated into the model in this study for estimates of the flux at high latitudes. The process for doing this is to simply divide the ratio of the model omni-directional flux at the magnetic equator by the corresponding value of the Divine model at the magnetic equator and then multiply that value by the Divine model prediction at the same energy at the appropriate magnetic latitude. In mathematical terms

had been minimal during this time period, especially for the high energy trapped electrons inside 16 Rj. However, there might be a long term temporal behavior due to the sun’s or other influMeV) energy data. The time the Galileo ences on the lower ( spacecraft had spent inside 16 Rj is estimated to be % of the total mission duration. A program was written to: (a) read each channel’s data, (b) average the data into 0.5 L-shell intervals (L is defined below), (c) convert the count rate averages to differential fluxes, and (d) then fit the resulting flux data to a differential flux spectrum of the form (1) where differential electron flux as a function of E in (cm -sr-sMeV) ; electron energy in MeV; constant (flux at MeV); constant (approximately the power law index for the high energy component); constant ( is approximately the power law index for the low energy component); constant (approximately the breakpoint energy between low and high energy spectra); . In the steps described above, the electron fluxes were obtained at 0.174, 0.304, 0.527, 1.5, 2.0, and 11.0 MeV. These energies correspond approximately to the F1, F2, F3, B1, DC2, and DC3 low energy cutoffs. To increase coverage to high energy, a seventh energy channel was also included in the analysis based on the Pioneer 10 and 11 electron fluxes at 31 MeV and higher [7], [8]. The 6 EPD flux values and the Pioneer value form an electron spectrum at each 0.5 L-shell interval. Each spectrum corresponding to was then fit to give values of , A, D, and (1). These values are listed in Table II. This table and (1) constitute the base average model of the electron omni-directional differential flux along the Jovian magnetic equator for energies

(2) where electron differential flux as estimated by the Divine radiation model at energy , magnetic field strength , and L-shell value ;

2284


electron differential flux as estimated by the equatorial omni-directional EPD radiation model at energy and L-shell value ; electron differential flux as estimated by the radiation model in this study at energy , magnetic field strength , and L-shell value ; magnetic field strength at spacecraft; magnetic field strength at Jovian magnetic equator for L-shell passing through spacecraft. This combined model, which was called the Galileo Interim Radiation Electron (GIRE) [10], was easily implemented by first , and running the original Divine model to compute at the spacecraft. and are then used to calculate at to give the model the magnetic equator. is then used in value at the magnetic equator. (2) can now be used to give the GIRE equatorial flux “corrected” for pitch angle variations in . The Divine model is still being used the range for proton environment estimates at all spatial ranges and for electron environment estimates outside the valid range of the GIRE model.

Fig. 3. Sample Europa mission trajectory used for computing mission electron fluence spectra in Fig. 4.

V. HEAVY ION COUNTER MODEL In addition to the radiation dose effects induced by the trapped energetic electron environment at Jupiter, there is also an intense heavy ion background that can induce severe levels of single event upsets. This environment was measured on Galileo by the HIC instrument. The HIC model uses data from 31 of the 35 Galileo orbits of Jupiter. HIC data cover radial distances of 2.5 Rj to well past 30 Rj, the range of the HIC model. The HIC model defines three heavy ion populations: carbon, oxygen and sulfur. The sources of oxygen are both Jovian (moons) and solar and are represented well in the data for the complete range of the model. Sulfur is mainly from the moon Io and is well represented in the inner magnetosphere. For distances greater than 15 Rj, the sulfur signal is erratic and has large error bars. Carbon is of solar origin and is well behaved into 15 Rj but erratic inside that distance because of poor statistics. There are two HIC models: the average and worst case. In each L-bin, the averages for oxygen, carbon, and sulfur were obtained by fitting the fluxes to power law. Then, the coefficients and indexes of the power law spectra for each L-bin were fit as a function of L. On the other hand, the worst case spectrum for oxygen was constructed by finding the worst spectrum in each L bin for the 31 orbits. That is, we obtained the power law oxygen spectra in each L bin whenever the spacecraft passed that particular L-bin and we chose the worst case spectra from these multiple spectra. However, this method could not be applied to the carbon and sulfur ions because there are not sufficient carbon and sulfur data in each orbit to obtain reliable worst case spectra. Instead, the carbon and sulfur worst case spectra were obtained by multiplying the number density ratios, C/O and S/O, to the worst case oxygen spectrum because the total number densities of carbon and sulfur have a well-behaved ratio to the total number density of oxygen over all distances in

Fig. 4. GIRE model integral spectrum is compared to the Divine model spectrum for the sample Europa mission trajectory in Fig. 3.

the model. Again, the coefficients and indexes of the worst case power law spectra for each L-bin were fit as a function of L. The average or worst case models are useful to produce the heavy ion environment for single event effect estimate at Jupiter. A more complete description of the HIC model can be found in [12]. VI. APPLICATIONS The GIRE and HIC models developed in this study were applied to a sample mission to generate the mission total ionizing dose and single event effect environments. First, to give an indication of the difference between the Divine model and the GIRE model, both were used to estimate the radiation fluence to a spacecraft on a mission to orbit Europa. The sample trajectory, Fig. 3, used conventional propulsion and orbital resonances with Galilean moons to place it in orbit around Europa for 1 month. Fig. 4 shows the electron fluence energy spectra obtained from these two models. Note that the model spectra have been extrapolated using a power law spectrum to estimate the fluxes above 50 MeV. The resulting fluxes in turn have been input to the NOVICE code [11] to estimate the radiation dose-depth curve

JUN et al.: HIGH-ENERGY TRAPPED PARTICLE ENVIRONMENTS AT JUPITER

Fig. 5. NOVICE radiation transport code dose estimates versus aluminum shielding thickness for the GIRE and the Divine spectra in Fig. 4 (2.54 mm = 100 mils).

for the mission. The dose-depth curve for the mission is compared with original Divine model estimates in Fig. 5. As seen, the dose as calculated for the GIRE fluence (which represents an “average” level) is about half of that calculated for the Divine model fluence, for shielding up to about 0.5 in (12.7 mm) of aluminum. For thicker shielding, the two estimates converge, with the GIRE model about 15% higher at 50.8 to 76.2 mm (2 to 3 in) of aluminum. The difference seen between the Divine and the GIRE models is mainly due to the use of different data sets that went into the respective models and the fact that GIRE is more of an “average” model-a statistical analysis [13] of the GIRE data shows that 1 sigma deviation agrees quite well with the Divine model estimate. Because the GIRE model used the data with more temporal and spatial coverage, it is considered to be a more representative model of the Jovian electron envi. However, it is also worthy to note that ronment for the Pioneer and Voyager data used in the Divine model still fall within a factor of 2 of the mean values of the Galileo EPD data. Considering that the data used for the Divine model represented only a limited temporal and spatial coverage, we believe the Divine model is still a valid model within the uncertainty of the model and complementary to the GIRE model. While the electron fluence is of importance in determining doses, the flux of high energy ions are important in determining SEU rates for a device. Fig. 6 illustrates the mission fluence spectra of carbon, oxygen, and sulfur ions using the HIC average model for the Europa orbiter mission shown in Fig. 3. Fig. 7 shows the worst case carbon, oxygen, and sulfur ions’ energy spectra obtained using the HIC worse case model at 9.5 Rj. Also shown in the figures are the interplanetary galactic cosmic ray (GCR) oxygen integral fluence and flux spectra from the CREME96 model for comparison. VII. CONCLUSION Measurements of the high-energy, omni-directional electron environment by the Galileo spacecraft Energetic Particle Detector (EPD) were used to develop a new model of Jupiter’s trapped electron radiation in the Jovian equatorial plane for the

2285

Fig. 6. Mission integral spectra for carbon, oxygen and sulfur ions obtained from the HIC average model. The interplanetary galactic cosmic ray oxygen integral fluence spectrum for the mission is also shown for the comparison purpose. The mission trajectory is shown in Fig. 3.

Fig. 7. Heavy ion integral flux spectra for carbon, oxygen and sulfur ions obtained using the HIC worst case model at 9.5 Rj. The interplanetary galactic cosmic ray oxygen integral spectrum is also shown for the comparison purpose.

range 8 to 16 Jupiter radii (1 km). Ten-minute averages of these data formed an extensive database of observations of the Jovian radiation belts between Jupiter orbit insertion (JOI) in 1995 and 2003. These data were then averaged to provide a differential flux spectrum at 0.174, 0.304, 0.527, 1.5, 2.0, 11.0, and 31 MeV in the Jovian equatorial plane as a function of radial distance. This omni-directional, equatorial model was combined with components of the original Divine model of Jovian electron radiation to yield estimates of the out-of-plane radiation environment. That model, referred to here as the Galileo Interim Radiation Electron (or GIRE) model, was then used to calculate the Europa mission dose for a sample trajectory. The prediction of the “average” GIRE model is about a factor of 2 lower than the Divine model estimate over the range of 2.54 to 25.4 mm of aluminum shielding, but exceeds the Divine model by about 50% for thicker shielding. Also, measurements of the high energy heavy ion environment by the Galileo Heavy Ion Counter (HIC) were used to develop a

2286


new heavy ion (particularly, carbon, oxygen, and sulfur) model at Jupiter. The HIC model covers the energy range between and MeV/nucleon from 2.8 Rj to Rj. A comparison to CREME96 showed that the trapped heavy ion environment at Jupiter is markedly higher than the interplanetary environment MeV/nucleon. These results form the basis for for evaluating the hardness of a given electronic component and its appropriateness for use in the Jovian environment. ACKNOWLEDGMENT The authors would like to thank D. Williams and R. McEntire of the Johns Hopkins University Applied Physics Laboratory for the EPD data and C. Cohen and E. Stone of California Institute of California for the HIC data. REFERENCES [1] Solar System Exploration: Priorities for 2003–2013 (2002). [Online]. Available: http://www.aas.org/ dps/decadal [2] D. J. Williams, R. W. McEntire, S. Jaskulek, and B. Wilken, “The galileo energetic particle detector,” Space Sci. Rev., vol. 60, pp. 385–412, 1992. [3] T. L. Garrard, N. Gehrels, and E. C. Stone, “The galileo heavy element monitor,” Space Sci. Rev., vol. 60, pp. 305–315, 1992.

[4] R. W. McEntire and T. Choo, Private Communication, Johns Hopkins Univ. Appl. Phys. Lab., Laurel, MD, 2002. [5] I. Jun, J. M. Ratliff, H. B. Garrett, and R. W. McEntire, “Monte Carlo simulations of the galileo energetic particle detector,” Nucl. Instrum. Meth. Phys. Res. A, vol. 490, pp. 465–475, 2002. [6] N. Divine and H. B. Garrett, “Charged particle distributions in jupiter’s magnetosphere,” J. Geophys. Res. A, vol. 88, no. 9, pp. 6889–6903, 1983. [7] J. A. Van Allen, D. N. Baker, B. A. Randall, M. F. Thomsen, and D. D. Sentman, “The magnetosphere of jupiter as observed with pioneer 10, 1. instrument and principal findings,” J. Geophys. Res., vol. 79, pp. 3559–3577, 1974. [8] J. A. Van Allen, B. A. Randall, D. N. Baker, C. K. Goertz, D. D. Sentman, M. F. Thomsen, and H. R. Flindt, “Pioneer 11 observations of energetic particles in the Jovian magnetosphere,” Science, vol. 188, pp. 459–462, 1975. [9] J. E. Connerney, M. H. Acuna, N. F. Ness, and T. Satoh, “New models of jupiter’s magnetic field constrained by the io flux tube footprint,” J. Geophys. Res. A, vol. 103, no. 6, pp. 11 929–11 940, 1998. [10] H. B. Garrett, I. Jun, J. M. Ratliff, R. W. Evans, G. A. Clough, and R. E. McEntire, Galileo Interim Radiation Electron Model, JPL Publication 03–006, 2003. [11] T. Jordan, “NOVICE: A radiation transport/shielding code, user’s guide,” Experimental and Mathematical Physics Consultants, 2000. [12] R. Evans, H. B. Garrett, C.M.S. Cohen, and E. C. Stone, Galileo Heavy Ion Radiation Model, 2003, JPL D-24 813. [13] I. Jun, H. B. Garrett, R. Swimm, R. W. Evans, and G. Clough, “Statistics of the variations of the high-energy electron population between 7 and 28 Jovian radii as measured by the Galileo spacecraft,” Icarus, vol. 178, pp. 386–394, 2005.