Energetic ions trapped in Saturn's inner magnetosphere

Energetic ions trapped in Saturn's inner magnetosphere

ARTICLE IN PRESS Planetary and Space Science 57 (2009) 1723–1731 Contents lists available at ScienceDirect Planetary and Space Science journal homep...

1MB Sizes 3 Downloads 15 Views

ARTICLE IN PRESS Planetary and Space Science 57 (2009) 1723–1731

Contents lists available at ScienceDirect

Planetary and Space Science journal homepage: www.elsevier.com/locate/pss

Energetic ions trapped in Saturn’s inner magnetosphere T.P. Armstrong a,, S. Taherion a, J. Manweiler a, S. Krimigis b, Chris Paranicas b, Don Mitchell b, N. Krupp c a b c

Fundamental Technologies, LLC, 2411 Ponderosa, Suite A, Lawrence, KS 66046, USA Johns Hopkins University, Applied Physics Laboratory, 11100 Johns Hopkins Rd., Laurel, MD 20723, USA ¨ r Sonnensystemforschung, Katlenburg-Lindau, Germany MPI fu

a r t i c l e in fo

abstract

Article history: Received 14 November 2008 Received in revised form 10 March 2009 Accepted 13 March 2009 Available online 1 April 2009

The low-energy magnetospheric measurement system (LEMMS) of the magnetosphere imaging instrument (MIMI) aboard the Cassini orbiter observed energetic ions and electrons during Saturn orbit insertion (SOI) of July 1, 2004. Salient features of the trapped ion fluxes observed in the L ¼ 2–4RS region include the occurrence of two distinct components of the energy spectrum of energetic protons. We shall refer to protons below 10 MeV as the low-energy component and above 10 MeV as the highenergy component. The low-energy component has a power law energy spectrum that falls at approximately E2.5. At about 1 MeV/nucleon, the ion pitch angle distributions tend to peak along and opposite to the magnetic field. The high-energy component has a separate peak in energy at about 20 MeV/nucleon and a pitch angle distribution that peaks at 901 to the magnetic field direction. The pitch angle distributions intermediate in energy evolve systematically from peaking along the field at low energies through isotropy to peaking perpendicular to the field at high energies. Ion species heavier than protons are present at energies from several MeV/nucleon up to 25 MeV/nucleon. Oxygen is separately observed to be present. Molecular hydrogen, H2 and H3 and helium are also present although the LEMMS instrumentation is not capable of unambiguously separating these species at multi-MeV energies. These species are measured separately in the outer magnetosphere (L ¼ 6.3–11RS) with the MIMI CHEMS instrument at energies from 1 to 100 keV/nucleon. This paper will report details of the observations and the results of modeling the abundances of the inner magnetosphere ions to determine constraints on source material and acceleration processes. & 2009 Elsevier Ltd. All rights reserved.

Keywords: Saturn Trapped ions Energetic particles Radiation zone Molecular hydrogen at Saturn

1. Introduction Observations of Saturnian trapped radiation were first made in September 1979 with instruments carried aboard Pioneer 11 (Fillius and McIlwain, 1980; Fillius et al., 1980; Chenette et al., 1980; McDonald et al., 1980; Simpson et al., 1980a,b; Van Allen et al., 1980a,b; Thomsen and Van Allen, 1980; Trainor et al., 1980). Conspicuous results of Pioneer 11 observations included spatial distributions that showed evidence of absorption of trapped radiation by satellites and rings. Voyagers 1 and 2 flew by Saturn in December of 1980 and August of 1981, respectively. Trapped radiation in the inner magnetosphere was observed by Voyager 2, which had a periapsis distance of 2.75RS and is reported by Schardt and McDonald (1983); Krimigis et al. (1982); Krimigis and Armstrong (1982) and Vogt et al. (1982). Schardt and McDonald (1983) used Voyager observations to estimate the loss rates of high-energy Saturnian protons coupled with cosmic ray albedo neutron decay (CRAND) derived from

 Corresponding author. Tel.: +1785 840 0800; fax: +1785 840 0808.

E-mail address: [email protected] (T.P. Armstrong). 0032-0633/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.pss.2009.03.008

impacts on the visible rings and found that CRAND could account for the observed intensities of such protons. Krimigis and Armstrong (1982) identified the existence of two components in the proton energy spectrum. Energetic ions heavier than protons were not reported by Pioneer 11 or Voyager 1 or 2 detectors in the inner magnetosphere. Schardt and McDonald (1983) also concluded that episodic diffusion was probably insufficient to account for the presence of ions trapped within the drift shells of Mimas, Janus, and the G-ring. The Cassini Saturn orbit insertion on July 1, 2004, carried the LEMMS instrument as a part of the MIMI’s complement of sensors. Excellent data coverage, a low magnetic latitude pass, and a fully operational angular scanning system offered the opportunity to use the LEMMS capabilities to separate protons and electrons, and to separately identify several of the major ion species.

2. The Cassini MIMI/LEMMS instrument The principal features of the instrument are described by Krimigis et al. (2004). All of the data used here are derived from the LEMMS assembly solid-state detector stack. Fig. 1 shows a

ARTICLE IN PRESS 1724

T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

Fig. 1. Meridian plane cross section of LEMMS solid-state detector telescope assembly.

Fig. 2. Summary of LEMMS telescope parameters and logical design.

meridian plane cross section of the LEMMS telescope. Solid-state detectors are designated by the symbols A and B for the ‘‘Low E (01)’’ end and D1, D2, D3a,b and D4 for the ‘‘High E (1801)’’ end. C represents a 1-mm-thick gold absorber which is shown in similar shading to the solid-state detectors because its thickness is a part of the telescope passband design. All of these elements are housed in a passive shield assembly which has apertures open to energetic particle access from both axial directions as shown in Fig. 1. One of these directions, ‘‘Low E (01)’’, has permanent magnets shown as inclined ovals in cross section designed to prevent electron access to the AB detector combination up to

about 700 keV. The detectors E1, E2, F1, and F2 are arranged to measure these deflected electrons. Responses of these detectors are used here only to distinguish between electron and ion responses. The other direction, (‘‘High E (1801)’’), has a 25-mmthick aluminum foil covering to prevent detector responses to light. Fig. 2 shows the telescope set-up with levels and logical channel definitions. Entries in Fig. 2 have obvious meanings except in the cases of the rows labeled dD (mm) and dE (MeV) where the former estimates the thickness uncertainty of the active volume of the detectors and the latter the rms noise in the electronic discriminator channel. These two quantities, along with

ARTICLE IN PRESS T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

the variation in projected depth across the opening angle of the aperture, are used to calculate the width of the discrimination levels used in the channel logics. Charged particles stopping in a particular detector only experience the fluctuation in energy loss due to electronic noise, while penetrating particles yield energy loss signals that vary due to project depth and detector thickness variation. All of these effects are taken into account in Fig. 3, which summarizes the resulting ion responses. The geometrical factors and passbands shown have been thoroughly evaluated from bench measurements, beam calibrations, computer modeling, and flight analysis including Earth swing by. These analyses lead us to expect that Fig. 3 values are accurate to better than 10 percent and possibly 5 percent in most instances. The ion species included in Fig. 3 are those expected to dominate the Saturn observations. Shaded entries represent species for which the subject channel has no nominal response. Molecular H2 and H3 responses are computed with the assumption that molecular binding is too small to affect the interactions of the nuclear constituents with the silicon detectors. That is, molecular H2 is treated as the precisely simultaneous arrival of two protons and molecular H3 as the arrival of three. There are two rows for the response of B0 to molecular H2 because its passband has two distinct intervals. Further, the protons and molecular H2 response of channel P4 turns out to be dominated by particles that do not generate the D31 veto signal because of an unavoidable gap of geometrical coverage in the D1, D2, and D3 detector stack. The

1725

passbands and geometrical factors given here in Fig. 3 differ in some instances from the nominal values provided in the instrument paper (Krimigis et al., 2004). These differences reflect the improved understanding of instrument performance derived from subsequent analyses of flight performance support by more complete and refined computer modeling. Another important feature of the LEMMS instrument is that it is mounted on a stepping platform which rotated through 3601, pausing at 16 different positions spaced 22.51 apart at a time cadence of 4.1 s per position, hereafter referred to as ‘‘subsectors’’. Since the LEMMS assembly collimator axis is perpendicular to the platform rotation axis, the 01 and 1801 apertures point successively to 16 angular positions on a great circle of the unit sphere. This arrangement allows for convenient observation of the distribution of arriving radiation over a range of directions. For magnetically trapped radiation, the gyromotion rapidly produces gyrotropy, that is, flux that depends only on pitch angle, the angle between the particle velocity vector and the magnetic field at a fixed spatial location. Since it requires about 48 s for a full 3601 scan the variations of particle intensity with position can cause aliasing of pitch angle distributions. Aliasing is easily recognized by the lack of closure of pitch angle distributions after the full 3601 scan. Removing the effects of aliasing can be accomplished by averaging the rates from 8 sectors preceding and 8 sectors following the observation. After times long enough to allow for a few full transits of the charged particle between mirror points, the

Fig. 3. Summary of geometrical factors and passbands for dominant ion species of the LEMMS channels.

ARTICLE IN PRESS 1726

T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

flux develops a further property of being reflection symmetric across the local plane perpendicular to the magnetic field as well as being axisymmetric around the magnetic field. When the reflection symmetric, axisymmetric, charged particle flux distribution is sampled by LEMMS in its 16 stops along a great circle, each subsector sampled has a companion subsector, 8 subsectors away, that samples the same flux but at 1801 different gyrophase. This redundancy of pitch angle sampling allows for convenient verification that the channel responses are due to trapped radiation. Artifacts such as response of the solid-state detectors to light are easily identified and removed. Physical constraints of the location of the LEMMS instrument aboard Cassini result in the obscuration of some portions of the views of some of the subsectors by incidental spacecraft and instrument structures. Obscuration is also easily identified and removed by comparing the responses of subsector pairs (0&8, 1&9, 2&10, etc.) which should be equal to within counting statistics unless there is obscuration. There is one important exception to the equality of oppositely looking subsector pairs; namely, when the Cassini spacecraft moves through a spatial gradient of trapped particles which is significant over a distance scale of the particle gyroradius, the paired measurements can differ, introducing thereby an anisotropy due to the spatial gradient. Such transverse anisotropies were detected during Galileo’s Earth II flyby with a very similar LEMMS instrument (Alinejad and Armstrong, 1997). The appearance of gradient anisotropies at the edges of the A ring, the Janus depletion, and the Mimas depletion provide an independent means to verify that the LEMMS ion channels are responding nominally. Gradient anisotropies will be revisited later in this paper. Fig. 4. 3.21 MeV proton fluxes along and opposite the magnetic field for 2.3–3.9RS inbound. Pitch angles are also shown.

3. Spatial distribution of proton fluxes Fig. 4 shows 3.21 MeV proton fluxes observed for an oppositely directed pair of subsectors for SOI inbound. Note that the pitch angles shown for this pair of subsectors look generally along and opposite the magnetic field. The reason that these pitch angles were chosen for display is that the fluxes of 3.21 MeV protons were larger for these pitch angles than for other angles—especially those near 901. For all of the proton passbands from 50 keV up to about 10 MeV the pitch angle distributions peaked along and opposite the magnetic field rather than at 901 as is the usual case for Earth’s trapped protons. As expected for trapped radiation, 3.21 MeV proton fluxes appear to be accurately identical in opposite directions as discussed above. Previously we also showed that trapped protons in the 12.1–58.9 MeV channel appear to have identical fluxes at different planetary longitudes (Paranicas et al., 2008). Satellite sweeping corridors for Janus at about 2.55RS and Mimas at 3.05–3.15RS are conspicuous. Trapped protons at Saturn were first observed by Pioneer 11 (Fillius and McIlwain, 1980; Fillius et al., 1980; McDonald et al., 1980; Simpson et al., 1980a,b; Van Allen et al., 1980a,b) and later by Voyager 2 (Krimigis et al., 1982, 1983; Schardt and McDonald, 1983; Vogt et al., 1982). The occurrence of trapped protons in regions entirely bounded by satellite sweeping corridors challenges explanations based on inward radial diffusion conserving the first and second adiabatic invariants and violating the third, which has been applied in earlier studies (Armstrong et al., 1983; McKibben and Simpson, 1980; Schardt and McDonald, 1983; Thomsen and Van Allen, 1980; Paranicas et al., 1997). That the sweeping corridors appear at all longitudes and local times well separated from the satellite position suggests that radial diffusion must be much slower than the rate of loss to impact on the satellite surface in those corridors. Given that, it seems most likely that protons trapped on field lines in the region between

sweeping corridors occur there because of processes in addition to whatever radial transport occurs. That the sweeping corridors of Enceladus, Mimas, and Janus are devoid of protons at flux levels that the LEMMS instrument can measure calls for explanations in terms of processes that act within the trapping region. Inward or outward radial diffusion explanations alone fail to explain the occurrence of protons within the regions bounded by sweeping corridors. Other sources such as CRAND and charge exchange have been suggested (Krimigis and Armstrong, 1982; Paranicas et al., 1997). Fig. 5 shows the fluxes of 5 and 20 MeV protons for the 2.3–3.9RS inbound region at pitch angles in the neighborhood of 901—near locally mirroring. Again there are conspicuous satellite absorption corridors, but the most important feature is that the flux of 20 MeV protons exceeds the flux of 5 MeV protons throughout the 2.3–3.9RS region by about an order of magnitude. The flux of 5 MeV protons is low enough in the 3.2–3.6RS range that counting statistics become evident. This implies that the energy spectrum has a secondary peak at 20 MeV during this Cassini SOI period similar to that reported from Voyager 2 observations about 2.7RS. Krimigis and Armstrong (1982) described this situation as the existence of two components of the proton spectrum.

4. Angular distributions of protons Further evidence that the distribution of 5 and 20 MeV protons is controlled by different effects is presented in Fig. 6, which compares the pitch angle distributions (PADs) observed at 2.62RS for 7 different proton energies. Fig. 6a shows the PADs for 1.8, 3.21, 2.54 and 5.0 MeV protons illustrating approximately two orders of magnitude of decrease of proton flux with increasing energy. Note that the 2.54 MeV proton flux is derived from the 01 of the LEMMS

ARTICLE IN PRESS T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

assembly while the other three energies are derived from the 1801 end. The 2.54 MeV flux lies between the 1.8 and 3.21 MeV fluxes as expected and this illustrates the consistency of calibrations

Fig. 5. Comparison of 5 and 20 MeV proton fluxes at pitch angles in the neighborhood of 901 (also displayed).

1727

applied to the separate ends of the telescope assembly. The PADs for 1.8 and 3.21 MeV protons have conspicuous minima at 901 of pitch with maxima at 61 and 1741. At the location of this observation, 2.62Rs, the pitch angle of the edge of the loss cone for adiabatically moving trapped radiation to enter the Saturnian atmosphere is about 141. It follows that at the endpoints of the PADs the detector collimator axes were located within the loss cone. For the 1801 end the collimator is open to accept particles at 7151 to the axis. Hence, the 1801 end of the sensor views both adiabatically trapped and atmosphere-impacting trajectories. An important feature of Fig. 6a is that the 2.54 MeV proton flux does not show a depletion at the pitch angle extremes despite the fact that its narrower field of view, 77.51, falls almost entirely within the loss cone. Fig. 6b shows PADs for 11.8, 12.2, 20, and 39.4 MeV protons, all of which peak at 901. The 20 and 39.4 MeV protons have more familiar PADs as expected for trapped particles. The 11.8 and 12.2 MeV protons appear to be a transition between the lower energies, which have minima at 901, and the higher energies which have apparent minima at 01 and 1801. The flux of protons at 20 MeV exceeds that at 11.8 and 12.2 MeV at all pitch angles. Finally, we note that the 11.8 MeV flux is derived from the 01 end of the LEMMS assembly and the 12.2 MeV flux from the 1801 end. That these fluxes agree within statistical uncertainties for nearly all of the pitch angle range except the extremes is additional confirmation of the correct calibration of the LEMMS telescope. The 11.8 MeV proton flux derived from the AB coincidence logic in the 01 end of the LEMMS telescope has conspicuous depletions at the largest and smallest pitch angles. As described above, the 0o end of LEMMS would have viewed mostly within the loss cone at these extremes. The departures of the 12.2 MeV and the 11.8 MeV PADs at the highest and lower pitch angle can therefore be attributed to the probable presence of a loss cone effect for the 11.8 MeV measurement which is obscured in the 12.2 MeV measurement by the larger instrument viewing cone. For trapped radiation observed at near equatorial latitudes, as is the case here, the particles in the neighborhood of 901 are the ones that dwell the longest near the equator where the greatest concentration of gas and dust can be found (Jurac et al., 2001; Horanyi et al., 2004.).

Fig. 6. Pitch angle distributions of protons in eight energy intervals at L ¼ 2.62RS inbound.

ARTICLE IN PRESS 1728

T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

Fig. 7. Pitch angle distributions for three low-energy proton intervals from Voyager 2 LECP inbound at L ¼ 2.75RS.

The contrast between the pitch angle distributions of low (generally below 10 MeV) and high (generally above 10 MeV) energy protons was also observed by the Voyager 2 low-energy charged particle (LECP) instrument. Fig. 7 shows the Voyager results for 0.54–0.99, 0.99–2.14 and 2.14–3.5 MeV protons for L ¼ 2.75RS where counting rates rise as pitch angles approach 01 and 1801. The aperture of the (LECP) instrument that acquired the measurements has a full opening angle of 451 and did not resolve the loss cone. Several of the LECP angular scans are shown and the lack of closure at the same pitch angle is due to the secular change of counting rate during the scans produced by spacecraft motion. Fig. 8 shows the Voyager rates for three higher energy proton passbands from the LECP instrument that all have conspicuous peaks at 901 of pitch angle. The similarity of the Cassini LEMMS and Voyager 2 LECP observations of proton angular distribution in the inner Saturnian trapping region suggests that the distinction between the high and low-energy proton components is likely to be a durable feature of Saturnian trapped protons.

5. Comparison of proton fluxes from Cassini SOI and Voyager 2 Saturn flyby Additional similarities between Voyager 2 and Cassini observations of Saturnian trapped protons are shown in Fig. 9. The Voyager 2 LECP instrument was commanded into a parked position, motor position 3, for the interval between 3.1 and 2.8RS inbound and thus has a fixed orientation relative to the Voyager 2 spacecraft which took on a slowly varying angular position as it moved from 3.1 to 2.8RS inbound. The pitch angle at

Fig. 8. Pitch angle distributions for three high-energy proton intervals from Voyager 2 LECP inbound at L ¼ 2.75RS.

which the LECP instrument looked is plotted as the dashed line. Note that the pitch angle climbed from 201 smoothly and passed through 901 at about 2.86RS, arriving at 1201 at 2.9RS where it remained throughout the rest of the interval. The flux determined from the 53–173 MeV proton channel climbed from a starting value of about 2 protons/cm2 s sr MeV at the beginning (2.805RS, 201 pitch angle) to a peak of about 18 protons/cm2 s sr MeV when the pitch angle passed 901. The peak in the LECP 53–173 MeV flux is obviously due to pitch angle variation. Also shown in Fig. 9 is the flux for the 63–160 MeV proton interval taken from the Goddard Space Flight Center-University of New Hampshire (GSFCUNH) instrument aboard Voyager 2. The GSFC-UNH instrument does not view co-linearly with the LECP instrument and its detailed flux variation is not expected to match that of LECP. Over the interval from 2.9 to 3.02RS the LECP and GSFC-UNH fluxes differ by about a factor of 2, with LECP observing more flux. Both the LECP and UNH fluxes decay during this 2.9–3.05RS interval with similar slopes. Finally, the Cassini LEMMS observations of 25–63 MeV protons are shown as the open squares for pitch angles taken from a single subsector at about 1001 of pitch (shown as ‘‘+’’ symbols). In the L ¼ 2.8–3.1Rs interval for Cassini inbound SOI the invariant latitude varied from 0.71 at 3.1Rs up to 1.81 at 2.8Rs while the Voyager 2 trajectory had an excursion from about 201 at 3.1Rs down to 12.51 at 2.8Rs. The Cassini trajectory was much more nearly equatorial than Voyager 2 and would, therefore, have experienced a higher trapped proton flux. Because neither the passbands of Voyager 2 and Cassini LEMMS shown here nor the trajectories are closely matched we do not attempt to account for the differences in apparent fluxes.

ARTICLE IN PRESS T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

Fig. 9. Comparison of high-energy proton fluxes from Voyager 2 and Cassini in the 2.8–3.1RS interval.

6. Energy spectra for trapped ions Fig. 10 shows the energy spectra for protons, molecular H2, and oxygen derived by averaging the entire inbound SOI passage (5.0RS inward to the ring edge at 2.3RS). The LEMMS detectors are susceptible to background arising from galactic cosmic rays and gamma rays emitted from the spacecraft radioisotope thermoelectric generators (RTGs). Average rates for a quiet interplanetary cruise period (hours 0–12 of day 313, 2003) were subtracted from the average rates for SOI inbound to determine the net foreground, i.e. Saturnian, rates. Since the purpose is to separately determine and compare the energy spectra of protons, helium, oxygen, and molecular species, the entire inbound passage was averaged in order to accumulate sufficient counts in the channels responding to species heavier than protons. Inspection of the highest resolution energy spectra had not shown significant changes from point to point. The net foreground rates were then interpreted in terms of ion fluxes using the passbands and geometrical factors shown in Fig. 3. Note that some rate channels are relatively ‘‘clean’’ in terms of potential ion species responses (P3, P4, P5, P6 measuring protons; Z1, Z2, Z3 measuring oxygen) and others have ‘‘mixed’’ species responses. The single parameter channels, e.g. A0:A7, measure only total energy deposited in a determined detector volume and require modeling to make inferences about individual species. Multiparameter channels were designed to isolate individual species as much as practical. The passbands shown in Fig. 3 were determined mostly from range–energy calculations for each species incident on the silicon detector stack for which the details are given in Fig. 2 and in Krimigis et al. (2004). We

1729

Fig. 10. Energy spectra of Saturnian trapped ions for Cassini’s inbound SOI passage.

proceeded with the analysis of the complete set of channel responses by computing the fluxes for all channels responding to protons. The results for that step are shown in Fig. 10 as ‘‘Protons(A),’’ filled squares, from channels A0:A7; ‘‘Protons(P),’’ open squares, from channels P1:P8; and ‘‘Protons(B),’’ asterisks, from channels B0 and B1. Note that the proton spectrum that results falls as an approximate power law from 40 keV to about 5 MeV, rises from 10 MeV to a secondary peak at 20 MeV and falls thereafter. This secondary peak is what we identify as the highenergy component. Note that the proton spectrum is well and consistently determined by multiple sensors and multiple logical channels. The next most certain interpretation shown in Fig. 10 is that the responses of Z1, Z2, and Z3 determine an energy spectrum for oxygen. For the purposes of the Saturnian environment we assume the response of the Z channels to be oxygen as opposed to other species heavier than helium. Note that these three channels determine an approximately power law oxygen spectrum in energy per nucleon with slope of about E2 or E2.5. A very important implication of the observed oxygen spectrum is that it is quantitatively insufficient to explain the responses of the H1:H4 channels (Armstrong et al., 2008). The principal alternatives to explain the H1:H4 responses then remain helium and molecular hydrogen, H2. It turns out that when the response of the B3 channel to oxygen is computed, there’s no need to invoke helium response for it. In fact, helium appears to be insufficient to explain the H1:H4 responses as well and we interpret the H1:H4 responses as deriving from molecular H2. The filled circles in Fig. 10, labeled ‘‘Molecular H2’’ show the response of the H1:H4 channels interpreted as molecular

ARTICLE IN PRESS 1730

T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

hydrogen. If the response of channel B2 is interpreted as molecular H2, it results in a flux value consistent with the H1:H4 derived spectrum. It is important to emphasize that the LEMMS instrument is insufficient by itself to establish uniquely that the Saturnian trapped ion fluxes are exclusively protons, molecular hydrogen and oxygen and that other species and spectral distributions are possible. However, the pattern that we show here including protons, molecular hydrogen and oxygen is sufficient to explain all of the responses. While other interpretations of the LEMMS responses during SOI might be constructed, our experience suggests that the interpretation we give here is likely to be the simplest and most robust available. We note further that the protons and molecular hydrogen spectra have similar slopes and overlap at similar energies per nucleon.

7. Remarks on ion spectra and compositions The presence of energetic protons, molecular hydrogen, and oxygen suggests that at least one material source of trapped ions at Saturn may be the dissociation products of water ice. The determination of abundance ratios is probably not accurate enough at this time to argue for water as an exclusive source of the low-energy component. We have no evidence at this time that the high-energy component is anything other than protons. The cosmic ray albedo neutron decay process has been invoked and extensively modeled as an explanation for the high-energy component (Krimigis and Armstrong, 1982; Chenette et al., 1980; Cooper and Simpson, 1980; Cooper, 1983; Cooper et al., 1983). Yet to be determined are the processes that account for the presence of trapped ions in the first place—especially the lowenergy component (50 keV–10 MeV). Transport across the satellite absorption shells presents barriers to radial transport. The lack of helium in the trapped ions suggests other than a solar wind material source. The lack of detection thus far of H3 indicates that the Saturnian ionosphere may not contribute importantly to the trapped ion population. The A proton spectrum is shown extending down to 50 keV because that’s what a nominal identification of A0 and A1 channel responses as protons shows. Observations at distances beyond 4RS do not show proton fluxes at these energies (Paranicas et al., 2008), and the presence of protons in the 2.5–4.0RS region may require processes other than simple inward transport to establish these fluxes.

8. Summary and conclusions The results of this study are that:

 Cassini LEMMS observations verify the existence of a second peak in the energy spectrum.

 The proton energy spectrum appears to have this secondary   

peak at all locations where the proton flux at Saturn is large enough to measure. Below about 5 MeV the proton pitch angle distributions are either isotropic or have modest minima at 901 to the magnetic field. Above 10 MeV the proton pitch angle distributions peak at 901 to the magnetic field with the amplitude of the peak increasing with energy. Low- and high-energy proton flux variations with position from beyond the drift shell absorption signature of Mimas inward through the G-ring plateau and the Janus absorption feature are similar.

 Molecular H2 has been observed at energies from about 2 to 20 MeV/nucleon in Saturn’s trapped radiation. The implications of these results are that:

 Pitch angle distributions provide important clues to proton sources and losses: J Broad low-energy distributions with minima at 901 suggest a deficiency of low-energy protons mirroring near the magnetic equator, suggesting that equatorially mirroring low-energy protons are lost more readily than off equatorially mirroring protons or protons at higher energy. J The cross sections for Couloumb scattering and charge exchange of protons from neutral gas are large at low energy and this could explain the missing low-energy protons. J Protons in the secondary energy spectral peak have PADs peaked at 901. If these protons arose from the CRAND source, the requirement for them to be trapped for long lifetimes would control the PAD.

Acknowledgements NASA through the Cassini project at the Jet Propulsion Laboratory funded the Cassini MIMI investigation and this study. MIMI/LEMMS was in part financed by the German BMBF through DLR under Contract 50 OH 0103, 50 OH 0802 and by the Max Planck Gesellschaft. References Alinejad, N., Armstrong, T.P., 1997. Galileo energetic particle detector observations of geomagnetically trapped protons. J. Geophys. Res. 102, 27053. Armstrong, T.P., Paonessa, M.T., Bell II, E.V., 1983. Voyager observations of Saturnian ion and electron phase space densities. J. Geophys. Res. 88, 8893–8904. Armstrong, T.P, Manweiler, J., Hamilton, D., Krimigis, S.M., Krupp, N., Paranicas, C., Mitchell, D., 2008. Energy spectra, angular distributions and compositions of energetic ions in Saturn’s inner magnetosphere, Paper B03-0022-08, COSPAR, Montreal, July 2008. Chenette, D.L., Cooper, J.F., Eraker, J.H., Pyle, K.R., Simpson, J.A., 1980. High-energy trapped radiation penetrating the rings of Saturn. J. Geophys. Res. 85, 5785–5792. Cooper, J.F., 1983. Nuclear cascades in Saturn’s rings: cosmic ray albedo neutron decay and origins of trapped protons in the inner magnetosphere. J. Geophys. Res. 88, 3945–3954. Cooper, J.F., Simpson, J.A., 1980. Sources of high-energy protons in Saturn’s magnetosphere. J. Geophys. Res. 85, 5793–5802. Cooper, J.F., Eraker, J.H., Simpson, J.A., 1983. The secondary radiation under Saturn’s A–B–C rings produced by cosmic ray interactions. J. Geophys. Res. 90, 3415–3427. Fillius, W., McIlwain, C.E., 1980. Very energetic protons in Saturn’s radiation belt. J. Geophys. Res. 85, 5803–5811. Fillius, W., Ip, W.H., McIlwain, C.E., 1980. Trapped radiation belts of Saturn: first look. Science 207, 425–431. Horanyi, M., Hartquist, T.W., Havnes, O., Mendis, D.A., Morfill, G.E., 2004. Dusty plasma effects in Saturn’s magnetosphere. Rev. Geophys. 42, RG4002 DOI 10.129/2004RG000151. Jurac, S., Johnson, R.E., Richardson, J.D., 2001. Saturn’s E-ring and production of the neutral torus. Icarus 149, 384–396. Krimigis, S.M., Armstrong, T.P., 1982. Two component proton spectra in the inner Saturnian magnetosphere. Geophys. Res. Lett. 9, 1143–1146. Krimigis, S.M., Armstrong, T.P., Axford, W.I., Bostrom, C.O., Gloeckler, G., Keath, E.P., Lanzerotti, L.J., Carbary, J.F., Hamilton, D.C., Roelof, E.C., 1982. Low-energy hot plasma and particles in Saturn’s magnetosphere. Science 215, 571–577. Krimigis, S.M., Carbary, J.F., Keath, E.P., Armstrong, T.P., Lanzerotti, L.J., Gloeckler, G., 1983. General characteristics of hot plasma and energetic particles in the Saturnian magnetosphere: results from the Voyager spacecraft. J. Geophys. Res. 88, 8871–8892. Krimigis, S.M., Mitchell, D.G., Hamilton, D.C., Livi, S., Dandouras, J., Jaskulek, S., Armstrong, T.P., Boldt, J.D., Cheng, A.F., Gloeckler, G., Hayes, J.R., Hsieh, K.C., Ip, W.-H., Keath, E.P., Kirsch, E., Krupp, N., Lanzerotti, L.J., Lundgren, R., Mauk, B.H., McEntire, R.W., Roelof, E.C., Schlemm, C.E., Tossman, B.E., Wilken, B., Williams, D.J., 2004. Magnetospheric imaging instrument (MIMI) on the Cassini mission to Saturn/Titan. Space Sci. Rev. 114, 233–329.

ARTICLE IN PRESS T.P. Armstrong et al. / Planetary and Space Science 57 (2009) 1723–1731

McDonald, F.B., Schardt, A.W., Trainor, J.H., 1980. If you’ve seen one magnetosphere you haven’t seen them all: energetic particle observations in the Saturn magnetosphere. J. Geophys. Res. 85, 5813–5830. McKibben, R.B., Simpson, J.A., 1980. Charged particle diffusion and acceleration in Saturn’s radiation belts. J. Geophys. Res. 85, 5773–5783. Paranicas, C., Cheng, A.F., Mauk, B.H., Keath, E.P., Krimigis, S.M., 1997. Evidence of a source of energetic ions at Saturn. J. Geophys. Res. 102, 17459–17466. Paranicas, C., et al., 2008. Sources and losses of energetic protons in Saturn’s magnetosphere. Icarus 197, 519–525. Schardt, A.W., McDonald, F.B., 1983. The flux and source of energetic protons in Saturn’s inner magnetosphere. J. Geophys. Res. 88, 8923–8935. Simpson, J.A., Bastian, T.S., Chenette, D.L., Lenz, G.A., McKibben, R.B., Pyle, K.R., Tuzzolino, A.J., 1980a. Saturnian trapped radiation and its absorption by satellite and rings: the first results from Pioneer 11. Science 207, 411–415.

1731

Simpson, J.A., Bastian, T.S., Chenette, D.L., McKibben, R.B., Pyle, K.R., 1980b. The trapped radiations of Saturn and their absorption by satellites and rings. J. Geophys. Res. 85, 5731–5762. Thomsen, M.F., Van Allen, J.A., 1980. Motion of trapped electrons and protons in Saturn’s inner magnetosphere. J. Geophys. Res. 85, 5831–5834. Trainor, J.H., McDonald, F.B., Schardt, A.W., 1980. Observations of energetic ions and electrons in Saturn’s magnetosphere. Science 207, 421–425. Van Allen, J.A., Thomsen, M.F., Randall, B.A., Rairden, R.L., Grosskreutz, C.L., 1980a. Saturn’s magnetosphere, rings and inner satellites. Science 207, 415–421. Van Allen, J.A., Randall, B.A., Thomsen, M.F., 1980b. Sources and sinks of energetic electrons and protons in Saturn’s magnetosphere. J. Geophys. Res. 85, 5679–5694. Vogt, R.E., Chenette, D.L., Cummings, A.C., Garrard, T.L., Stone, E.C., Schardt, A.W., Trainor, J.H., Lal, N., McDonald, F.B., 1982. Energetic particles in Saturn’s magnetosphere: Voyager 2 results. Science 215, 577–582.