Energetic Electron Beams in Ganymede’s Magnetosphere

D. J. Williams Johns Hopkins Applied Physics Laboratory

Laurel, MD 20723

Abstract We present energetic particle measurements obtained from the Galileo satellite’s final close encounter with Jupiter’s moon Ganymede. The encounter occurred on December 28, 2000, crossing Ganymede’s plasma wake at mid-latitudes. Field-aligned electron beams were observed in the region where trapped-like electron and ion distributions were measured. Given the altitude of the encounter (closest approach ~2335 km) the observation of both trapped-like distributions and field-aligned electron beams is surprising and suggests that the Ganymede magnetosphere is a more complex configuration than earlier charged particle results indicate. The location of the beams, their angular width, and their spectral characteristics suggest that they may be formed by the entry and subsequent quasi-chaotic drift of ambient Jovian electrons within Ganymede’s magnetosphere. Introduction During its very productive life in orbit around Jupiter, the NASA Galileo satellite performed six close encounters with the Jovian moon Ganymede. Data from the first of these encounters (G1; G for Ganymede, 1 for the orbit number) led to the discovery of Ganymede’s intrinsic magnetic field (Gurnett et al., 1996; Kivelson et al., 1996). Energetic particle measurements from the next four encounters, G2, G7, G8, and G28, provided additonal new information about Ganymede’s magnetosphere, its interaction with the Jovian environment, and its trapped electron and ion populations (Williams et al., 1998: Williams et al., 1997a; Williams, 2001). The present observations, from Galileo’s final close encounter with Ganymede, G29, show the existence of energetic electron beams within its magnetosphere and present another perspective of this unique magnetospheric environment. The Galileo Energetic Particles Detector, (EPD), measures the spectral and angular distributions of ions (no composition) above 20 keV, electrons above 15 keV, and elemental species from protons through iron above approximately 15 keV/nucleon (Williams et al., 1992). The EPD’s two bidirectional detector heads are mounted on a stepper platform. Platform motion coupled with the satellite spin allows the detector heads to view up to a full 4π steradian sample of the sky. For the G29 encounter the full 4π steradian scan was used including a step behind a background/calibration shield after each scan. Data

Figure 1 summarizes EPD measurements through the Galileo G29 encounter with Ganymede. The sketch in the upper left corner of the figure shows the encounter trajectory past Ganymede. Jupiter is to the left in that sketch and the encounter occurred near local noon in the Jupiter system. The top three panels show, from top to bottom, the intensities of 42 – 65 keV ions (energy range defined by proton response), 29 – 42 keV electrons, and 304 – 527 keV electrons. These channels are representative of EPD’s electron and ion differential energy channels. Measurements obtained while behind the background shield are identified as is the time of closest approach (~2335 km). The horizontal axis shows time and radial distance from Ganymede. Prior to encounter, ion intensities display anisotropies similar to those measured at previous Ganymede encounters. These anisotropies are a measure of the Jovian plasma’s convective flow speed at Ganymede. The magnitudes of these anisotropies are consistent with previous encounters (Williams et al., 1997b, 1998) and provide an estimate of a convective flow speed that is roughly 80 +- 10% of the expected full corotation speed of ~190 km/s. The ion anisotropies in Figure 1 became significantly smaller just after the background measurement that began at approximately 0815 hours. As with all previous encounters, we interpret this to be a signature of entering the Ganymede magnetic field environment and accordingly we indicate a magnetopause transition at that time. Ion intensities show a slight decrease during the encounter while electron intensities decrease noticeably. All channels display greatly increased loss cone anisotropies. The bottom panels show sample pitch angle distributions measured prior to and after entry into Ganymede’s magnetosphere. Prior to entry low energy electrons are nearly isotropic, higher energy electrons display a butterfly distribution, and ion distributions show an anisotropy at 90° that is characteristic of ambient plasma convective flow. The three distributions shown within Ganymede’s magnetosphere are the only three for which pitch angles greater than 170° were sampled. These samples show, as do all other distributions through the encounter, distributions peaked at a pitch angle of 90°, similar to the trapped-like distribution reported earlier (Williams et al., 1997a; Williams, 2001). In the discussion section below we conclude that these distributions indicate that Galileo entered the closed field line region of Ganymede’s magnetosphere and measured its trapped particle population. Surprisingly, two of the three broadest pitch angle samples show the existence of prominent electron beams extending to at least the 304 – 527 keV energy range. The beams are unidirectional and travel along the magnetic field towards the moon’s surface. As the pitch angles sampled did not extend to at least 170° at other times during the encounter, no additional beam observations were obtained. No ion beams were observed at any time during the encounter. Discussion

Figure 2 shows a spectral comparison of pre-encounter electrons and the observed beams. The figure also shows a normalization of the beams to pre-encounter values, obtained by

dividing the beam fluxes by pre-encounter fluxes at similar pitch angles. We see that the spectra of Ganymede’s electron beams are quite similar to the pre-encounter electron spectra. However the normalized spectra show that, with respect to the ambient electron population, a larger fraction of low energy electrons is present in the beams. The similar spectra suggest that the beams consist of ambient Jovian electrons that have entered Ganymede’s magnetosphere and have arrived at the location of Galileo with little spectral distortion. The slightly increased fraction of low energy electrons could signify either the existence of an acceleration process or an energy dependent entry mechanism. Integrating the spectra over the EPD energy range and over a beam half-width up to 3° yields an energy deposition of ~.002-.02 ergs/cm2s to Ganymede’s surface. With a mid-latitude closest approach of ~2335 km, it is somewhat surprising that trapped-like electron and ion distributions were measured. As has been shown at Io (Williams and Thorne, 2003), trapped-like distributions do develop on open field lines connected to Jupiter if particle half-bounce times are less than the field line convection time across the moon’s polar cap. Could this be the case during the G29 encounter? Although our results from previous encounters with Ganymede’s polar caps strongly suggest that this is not the case, the present data allow us to test this possibility. At Ganymede electron half-bounce times on field lines connected to Jupiter are <~35 s. If electron pitch angle diffusion were less effective on such field lines during G29 than that measured earlier (Williams and Mauk, 1997), electron distributions could develop a double loss cone structure due to normal motion along open field lines following loss due to impact with Ganymede’s surface. However the much longer half-bounce times of the ions makes it quantitatively very difficult to explain their distributions via motion along open field lines. For example a 50 keV proton has a half-bounce time of approximately 700 s. The lowest energy EPD ion channel, which displays similar pitch angle distributions as those shown here, has a half-bounce time of ~850 s. In comparison, based on an approximate 35 km/s field line convection speed over the Ganymede polar cap (Williams et al., 1998), the convection time of a field line across roughly 3 RG to Galileo’s near-wake-crossing location (see Figure 1) is ~220 s. This is not time enough for the ions to develop a double loss cone structure simply due to adiabatic motion along a Ganymede-Jupiter field line. To illustrate the points made in the previous paragraph, we compare distributions measured during the G7 encounter with those presented here during the G29 encounter. The G7 trajectory past Ganymede was similar to that of G29, having a closest approach of ~3095 km. During G7 EPD measured electron and ion distributions indicative of particles on field lines connecting to Ganymede and Jupiter (Williams et al., 1998). The G29 distributions are quite distinct from the G7 distributions; both ions and electrons have lower intensities and clear peaks at 90° pitch angles with electrons also having a field-aligned component. Figure 3 shows a comparison of electron and ion pitch angle distributions measured during G7 and G29 in the E1 and A1 channels shown in Figure 1. From the half-bounce times discussed in the preceding paragraph and the distinct differences displayed in Figure 3, we conclude that the trapped-like distribution observed during G29 represent electrons and ions trapped in Ganymede’s magnetic field. As with previous Ganymede encounters, we present in Figure 4 a perspective of the present EPD results superimposed on Galileo’s trajectory through a model of Ganymede’s

magnetic field (J. Mafi, M. Kivelson. Personal communication). Jovian plasma flow and therefore the Ganymede wake extend out from the plane of the figure. Galileo’s trajectory relative to the wake is sketched in the inset shown in the upper left of Figure 1. We emphasize that Figure 4 is meant only as a guide to where the EPD results were obtained within a model of the Ganymede magnetosphere. A much more detailed consideration of Ganymede’s magnetic topology is given in Kivelson et al., (1998).

Previously Williams et al., (1997a) suggested that the source of Ganymede’s trapped electron population is the entry through the downstream wake region of a fraction of the ambient Jovian electron population. EPD measures those electrons that then drift to its position in Ganymede’s magnetosphere. The present beam measurements indicate that a portion of the injected electrons may drift through regions that result in chaotic or quasi-chaotic trajectories, possibly leading to field-aligned beams (eg. neutral point and/or plasma sheet regions; see for example Speiser, 1965, 1967; Chen and Palmadesso, 1986; Martin, 1986; and references therein). Electron beams also have been observed at Io, where they are bi-directional, appear in the wake region, appear across both polar caps, possess softer spectra and higher intensities than the ambient electron fluxes, and are generated near the Jovian ionosphere at altitudes of ~ 0.6 RJ (see Williams and Thorne, 2003, and references therein). In addition, the Io electron beams appear to participate in the current system responsible for the Io footprint aurora (Mauk et al., 2001). In contrast, the Ganymede electron beams are unidirectional, appear on closed field lines within the moon’s magnetosphere, do not appear on polar cap field lines connected to Jupiter, and apparently are related to the ambient Jovian electron population. These characteristics strongly suggest that the Ganymede electron beams do not participate in the current system responsible for the Ganymede footprint aurora. The existence of both a trapped particle population within Ganymede’s magnetosphere and an electron beam region should provide stringent boundary conditions for the continuing development of Ganymede magnetospheric models. Acknowledgements We gratefully acknowledge M. Kivelson and the Galileo magnetometer team for providing magnetic field data used to compute pitch angles. We further thank M.Kivelson and J. Mafi for providing the model Ganymede magnetic field shown in Figure 3. It is a pleasure to acknowledge fruitful discussions with B. Mauk. References Chen, J., and P. J. Palmadesso, Chaos and nonlinear dynamics of single particle orbits in a magnetotaillike magnetic field, J. Geophys. Res., 91, 1499, 1986

Gurnett, D. A., W. S. Kurth, A. Roux, S. J. Bolton, and C. F. Kennel, Evidence for a magnetosphere at Ganymede from plasma wave observations by the Galileo spacecraft, Nature, 384, 535-537, 1996 Kivelson, M. G., K. K. Khurana, C. T. Russell, R. J. Walker, J. Warnecke, F. V. Coroniti, C. Polanskey, D. J. Southwood, and G. Schubert, Discovery of Ganymede’s magnetic field by the Galileo spacecraft, Nature, 384, 537-541, 1996 Kivelson, M. G., J. Warnecke, L. Bennett, S. Joy, K. K. Khurana, J. A. Linker, C. T. Russell, R. J. Walker, and C. Polanskey, Ganymede’s magnetosphere: Magnetometer overview, J. Geophys. Res., 103, 19963-19972, 1998 Martin, R. F., Jr., Chaotic particle near a two dimensional magnetic neutral point with application to the geomagnetic tail, J. Geophys. Res. 91, 11985, 1986 Mauk, B. H., D. J. Williams, and A. Eviatar, Understanding Io’s space environment interaction: Recent energetic electron measurements from Galileo, J. Geophys. Res., 106, 26195-26208, 2001 Speiser, T. W., Particle trajectories in model current sheets, J. Geophys. Res., 70, 4219, 1965 Speiser, T. W., Particle trajectories in model current sheets, 2, Applications to aurora using a geomagnetic tail model, J. Geophys. Res., 72, 3919, 1967 Williams, D. J., R. W. McEntire, S. Jaskulek, B. Wilken, The Galileo energetic particles detector, Space Science Reviews, 60, 385-412, 1992 Williams, D. J., B. H. Mauk, R. W. McEntire, Trapped electrons in Ganymede's magnetic field, Geophys. Res. Lett., 24, 2953-2956, 1997a Williams, D. J., B. H. Mauk, R. W. McEntire, E. C. Roelof, T. P. Armstrong, B. Wilken, J. G. Roederer, S. M. Krimigis, T. A. Fritz, L. J. Lanzerotti, N. Murphy, Energetic particle signatures at Ganymede: Implications for Ganymede's magnetic field, Geophys. Res. Lett., 24, 17, 2163-2166, 1997b. Williams, D. J. and B. H. Mauk, Pitch angle diffusion at Jupiter’s moon Ganymede, J. Geophys. Res., 102, 24,283-24,287, 1997 . Williams, D. J., B. H. Mauk, R. W. McEntire, Properties of Ganymede's magnetosphere as revealed by energetic particle observations, J. Geophys. Res., 103, A11, 17523-17534, 1998. . Williams, D. J., Ganymede’s ionic radiation belts, Geophys. Res. Lett., 28, 3793-3796, 2001

Williams, D. J. and R. M. Thorne, Energetic particles over Io’s polar caps, J. Geophys. Res., 108, SMP 7-1 to SMP 7-11, 2003 Figure Captions Figure 1: A timeline of Galileo EPD measurements during the G29 encounter with Ganymede. The ion and electron data shown in the top three panels are representative of all the EPD differential electron and ion channel observations. Magnetopause entry (see text), closest approach, background measurement intervals, and a data gap are identified. The sketch in the upper left corner shows the encounter geometry. The horizontal scales for the top three panels are time and radial distance from Ganymede. Local pitch angle distributions (LPA) for electrons and ions are shown in the plots at the bottom half of the figure. The time intervals for each plot are indicated by the shading and are, from left to right, 0804:47-0806:47, 0821:07-0821:26, 0823:07-0823:26, and 0826:07-0826:26. Figure 2: The top panel shows electron spectra measured during the pre-encounter period and during the first beam observation. Spectra are shown at pitch angles near 90° and near field-alignment. The bottom panel shows the spectra measured during both beam observation normalized to pre-encounter spectra at similar pitch angles. Figure 3: Comparison of electron and ion pitch angle distributions measured during the G7 and G29 Ganymede encounters. Both encounters had similar trajectories across Ganymede’s mid-high latitude near-wake, with the G7 closest approach being ~3095 km and that for G29 being ~ 2335 km. LPA is the local pitch angle. The G7 distributions are consistent with distributions measured on field lines connecting Ganymede and Jupiter. The G29 distributions show peaks at 90° LPA and are consistent with trapping within Ganymede’s magnetic filed. The G29 electron distribution also shows a field-aligned component. A sin0.7 curve is shown for comparison with the G29 ion distribution. Figure 4: EPD results are displayed on Galileo’s trajectory as projected into a model of Ganymede’s magnetic field (Kivelson et al., 1998). The locations of magnetopause entry, data gaps, trapped-like electron and ion distributions, and electron beam observations are identified.

1000

600

200

105

104

103

102

10000

1000

100

F2

(c/s

)E

1 (c

/s)

A1

(c/s

)

Galileo Ganymede Encounter G2928 Dec. (D363) 2000

0 60 120 180LPA

0 60 120 180LPA

0 60 120 180LPA

0 60 120 180LPA

20000

15000

10000

5000

0600

500

400

300

200

0

Inte

nsity

(cp

s)In

tens

ity (

cps)

20000

15000

10000

5000

0700600500400

200

0

Inte

nsity

(cp

s)In

tens

ity (

cps)

300

100

30000

20000

10000

5000

0600500

400

200

0

Inte

nsity

(cp

s)In

tens

ity (

cps)

300

100

25000

15000

100000

10000

Inte

nsity

(cp

s)

900

700

600

400

Inte

nsity

(cp

s)

500

800

7.54 4.13 6.11

BG

BG

3.936.33 2.29 2.17

Data gap

Background measurements (BG)

25000

800

400

MP CA

R (GAN)

07:54:41 08:00:00 08:10:00 08:20:00 08:30:00 08:40:00 08:50:00

A1: 42–65 keV Z > = 1 ionsE1: 29–42 keV electronsF2: 304–527 keV electrons

E1

F2

A1

Wak

e

100010010E (keV)

0.5

0.4

0.3

0.2

0.1

0

Nor

mal

ized

spe

ctra

6

5

4

3

2

1

log

flux

(#/c

m2

s st

keV

)

b1/pe

b2/pe

b1 = 178°

b1 = 89°pe = 169°

pe = 87°

b1; beam 1, 0821:07 hoursb2; beam 2, 0826:07 hourspe; pre-encounter, 0808:47 hours

90 120 150 18060300LPA (deg)

A1: 42–65 keV ions

E1: 29–42 keV electrons5

4

34

3

2

1

Log

[Inte

nsity

(cp

s)]

sin0.7 (LPA)

G7: D095/1997; 0707:14–0708:14 hrsG29: D363/2000; 0821:07 hrs

Comparison of pitch angle distributionsfrom Galileo’s G7 and G29 encounters

08:10 08:25 08:40

M

4

2

0

–2

–4

Z (

RG

)

420–2–4Y (RG)

DGData Gap

Pre-encounterenvironment

Magnetopause region (ion anisotropy decrease)

Measurement of trapped-like electron and ion distributions

Electron beams

Jupiter