Survey of thermal ions in the Io plasma torus with the

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. A4, PAGES , APRIL 1, 2001 Survey of thermal ions in the Io plasma torus with the Galileo spacecraft L. A. Frank and W. R. Paterson Department of Physics and Astronomy, University of Iowa, Iowa City Abstract. The densities, temperatures, and bulk flow velocities are reported for a series of six passages of the Galileo spacecraft through Io's torus in Jupiter's magnetosphere. These observations were gained with the plasma analyzers on December 7, 1995, and July 1-2, August 12, September 14, October 11, and November 25, all of the latter in These plasma analyzers provided highresolution measurements of the energy/charge (E/Q) spectra of the threedimensional velocity distributions of the thermal ions that included determinations of the mass/charge (M/Q) of the primary ions with a mass spectrometer. The four dominant ions in the hot torus were two populations of ions with M/Q = 16 and two smaller populations with M/Q = 8 and 32, respectively. The identification of these four ions was based upon a best fit to the E/Q spectra of the measured three-dimensional ion velocity distributions. The two distributions with M/Q = 16 were characterized by two different temperatures, in the ranges of 20 to 30 ev and 60 to 80 ev, respectively. On the basis of expectations of higher temperatures with higher masses for the pickup ions the cooler population is identified as O +, and the hotter as S ++. The two smaller populations with M/Q = 8 and 32 are identified as O ++ and S +, respectively. The temperatures were 20 ev for the O ++ and 60 to 80 ev for the S +. The densities and temperatures of the ions in the hot torus remained constant during the period July through November However, the ion densities during the initial passage on December 7, 1995, were greater by factors of-3, which support the presence of long-term density variations of the torus plasmas but with relatively small fluctuations in the temperatures. A complete survey of System III longitudes was acquired with the set of six passages. The presence of an "active sector" at longitudes in the approximate range of 180 ø to 230 ø as originally found with remote observations of the torus brightnesses is confirmed with these Galileo measurements. In addition, further evidence for the importance of interchange motions for radial plasma transport was also evident in several of the Galileo passages, a dynamical process which was first identified in fields and particles measurements during the first passage through the torus on December 7, A persistent lag with. an average in the range of 2 to 4 km/s in the azimuthal flows relative to that for rigid corotation was detected and supports the previously proposed System IV coordinate system that has a slightly slower rotation rate relative to that for System III. 1. Introduction The interaction of Io's volcanic gases with Jupiter's rapidly rotating magnetic field to form a massive torus of ionized gases around this giant planet is a very active field of scientific investigation. The first indication of this interaction was the finding of a correlation of 22-MHz radio emissions with the position of Io in its orbit around Jupiter [Bigg, 1964]. The presence of this strong interaction was interpreted by Piddington and Drake [1968] as Copyright 2001 by the American Geophysical Union. Paper number 2000JA /01/2000JA evidence for the electrodynamic response of the motion of Io through the Jovian magnetic fields. Goldreich and LyndenoBell [1969] extended this interpretation in terms of a current system driven by Io's motion. This current passed through this moon and was closed with oppositely directed currents, aligned along the magnetic field, in the Jovian ionosphere. Neubauer [1980] later showed that the field-aligned currents were standing waves that propagate away from Io with the local Alfv6n speed. That is, the current is not magnetically field-aligned but is directed along "Alfv6n wings" because the magnetospheric flow speeds are substantial relative to the Alfv6n speed. The remote detection of a large cloud of neutral gases in the vicinity of Io was reported by Brown 6131

2 6132 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS and Chaffee [1974]. The emissions from the D lines of sodium were observed. The first observations of emissions from plasma ions in the vicinity of Io were acquired by Kupo et al. [1976]. This pair of emission lines was attributed to singly ionized sulfur. The next major milestones in the detection of the plasmas in the vicinity of Io and its orbit around Jupiter were gained during the fiybys of this planet by the Voyager I and 2 spacecraft. Observations of the ultraviolet and visible spectra with the onboard spectrometer established that Jupiter was encompassed by a torus of heavy ions such as those of oxygen and sulfur in several charge states [Broadfoot et al., 1979]. Indeed, these ions were directly measured by the plasma analyzer on board Voyager I as it passed through the plasma torus [Bridge et al., 1979]. Overviews of these optical and plasma measurements are given by Strobel [1989] and Bagenal [1989]. Our primary interests in the present paper are the observations of the ion populations in the Io torus with the plasma analyzer on board the Galileo spacecraft. Bagenal et al. [1980] used the inbound measurements in the torus with the plasma analyzer on Voyager I in 1979 to construct a two-dimensional map of the ion densities as functions of jovicentric radial distance RJ and latitude. This model was refined with the use of remote observations of the torus at ultraviolet wavelengths with the Voyager spectrometers [Bagenal, 1994]. Near the equator the principal features of the plasmas are a cold ion torus at radial distances of ~5.0 to 5.4 RJ inside of Io's orbit at 5.9 RJ, a warm distribution of heavy ions with mildly decreasing densities with distance at 5.7 to 7.0 RJ, and a warm distribution with a greater gradient at 7.0 to 8.0 RJ [Bagenal and Sullivan, 1981; Belcher, 1983; Bagenal, 1985, 1994]. The change in the density gradient has been interpreted as impoundment of the plasmas at <7.0 RJ by the slower radial diffusion of the ions due to interchange motions of the magnetic flux tubes at the larger radial distances [Siscoe et al., 1981]. As noted by Siscoe [1978], the source of ions for Jupiter's plasma torus can be (1) this giant planet's ionosphere, (2) ionization of the atmospheres and surface materials of its moons, and/or (3) ionization of the neutral gases of the interstellar medium which penetrate into the magnetosphere. The early Voyager I optical survey [Broadfoot et al., 1979] and this spacecraft's in situ plasma measurements [Bridge et al., 1979] clearly established that heavy ions such as those of oxygen and sulfur are major constituents of the plasma torus. The source is provided by the remarkable volcanic activity of Io. One viewpoint of the source of the torus ions considers that large quantities of neutral gases are injected into Jupiter's magnetosphere and are subsequently ionized by plasma electrons. These ions are "picked up" by the electric fields associated with the rapid rotation of the planetary magnetic field. The energy gained by these ions is substantial and a quasi-equilibrium state is reached between the ion and electron plasmas. For a description of this model the reader is referred to Smith and Strobel [1985, and references therein]. On the other hand, Shemansky [1988] has considered the physical chemistry and the energy branching in detail for the Io torus and concludes that an additional energy source is required. One of the suggested sources of the plasmas is a direct interaction of the plasma torus with the Io atmosphere. Recently, Frank and Paterson [1999b] report large densities of pickup ions at close flyby distances of Io with the Galileo spacecraft which support the conclusions of Shemansky. It would appear that both the direct Io injection and the ionization of torus neutrals play an important part in the torus dynamics. The Galileo spacecraft is currently providing great advances in our knowledge of the morphology and dynamics of Jupiter's magnetosphere. One of the important results for Io's plasma torus is evidence for the transport of its plasmas by interchange motions of Jupiter's magnetic flux tubes as anticipated for many years on the basis of theoretical considerations [Brice and McDonough, 1973; Hill and Michel, 1976; Siscoe and Chen, 1977]. The existence of such interchange motions was strongly supported by Galileo observations of magnetic fields, energetic particles, and plasma waves [Kivelson et al., 1997; Thorne et al., 1997; Bolton et al., 1997]. The direct determinations of transient decreases in the mass densities of the thermal plasmas and coincident radially inward convection of these plasmas toward Jupiter provided further substantial evidence that transport by interchange motions is a significant dynamical feature of the plasma torus [Frank and Paterson, 2000a]. The purpose of our paper is the presentation of the thermal ion plasmas as measured with the Galileo plasma instrumentation during a set of six traverses of the Io torus. Special attention is directed toward the investigation of temporal and longitudinal variations of these plasmas. 2. Instrumentation The plasma instrumentation (PLS) on board the Galileo spacecraft is composed of spherical-segment electrostatic analyzers that are capable of measurements of the positive ion and electron velocity distributions over the energy/charge (E/Q) range of 0.9 V to 52 kv. Three miniature magnetic spectrometers are also positioned at the exit apertures of the ion electrostatic analyzers in order to determine the mass/charge (M/Q) of these ions. The PLS has been previously described by Frank et al. [1992]. Several features of this instrumentation which are most relevant to the present study are noted here. The fan-shaped field of view of the electrostatic analyzers, full-width 157 ø, is divided into seven segments with multiple sensors and is oriented such that it is in a plane parallel to the spin axis of the spacecraft. The center of this field of view is directed perpendicular to this spin axis.

3 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6133 Electronic sectoring of the responses of the sensors according to the phase of the spacecraft rotation allows the three-dimensional determination of the velocity distributions of the positive ion and electron plasmas, i.e., coverage of 80% of the 4n-steradian solid angle for arrival directions of charged particles at the analyzers. During the high-rate mode that is recorded with the on-board tape recorder and subsequently telemetered to Earth at lower rates the electrostatic analyzers were sampled in 8 rotation sectors for each sensor. The rotation period of the spacecraft was s. During a single spacecraft rotation 12 E/Q passbands were sampled during each of the 8.rotation sectors. The sampling time for one E/Q passband is 0.15 s. Every fourth of the full set of passbands in the E/Q range of 8 V to 52 kv is sampled during this single spacecraft rotation. During the next spacecraft rotation this sampling sequence is repeated such that every second passband is acquired during the total of two rotation periods. This two-rotation period is the typical sampling period of the ion velocity distributions reported here. Exceptions are noted as required. The third rotation is devoted to telemetering the responses of the lowest passbands of the electrostatic analyzers in the E/Q range of 0.9 to 14 V or the responses of the mass spectrometers. Electron and positive ion channels are identified as E and P, respectively. The equatorial sensors (E4,P4) and both sets of the polar sensors (El,P1 and E7,P7) are sampled in this low-energy mode in 16 passbands and in four equispaced rotation sectors. Each passband is sampled during 0.17 s, corresponding to a rotation of the fields of view by 3.2 ø. This low-energy scan is acquired once per 12 spin cycles, or 242 s. ß In order to model the responses of polar sensors P1 and P7 for determining the fluxes in the torus ion beams their fields of view have been each divided into 77 contiguous fields of view. This is necessary because their fields of view are wide relative to the ion beams, i.e., 9 ø to 41 ø and 136 ø to 166 ø, respectively, with respect to the spacecraft spin axis [Frank et al., 1992]. Ions arriving at sensor P1, for example, are assigned appropriate angles of arrival, 0 and (I), in each of the 77 segments. Each segment has a geometric factor dg. The angles and dg are determined with a ray-tracing program, and the dg values are normalized such that their sum is the full geometric factor G of sensor P1. The response of the sensor to a given velocity distribution is the sum of the responses for the 77 segments. The density, temperature and M/Q species are varied in order to provide the best fit to the E/Q spectra measured with the sensor. The bulk velocity is that given by the computation of plasma moments. Because the spacecraft telemetry clock is asynchronous relative to the angular sectoring which is locked onto the celestial sphere, the acquisition time for the above sampling sequences varies slightly and is longer than the spin period. This situation occurs because the sampling must begin coincident with the occurrence of a rotation sector boundary. The average time for this spin cycle is s. The elapsed time during the three rotations is 3 x 20.2 s s, or I min. Thus, during the first two spacecraft rotations a total of (8 rotation sectors) x (7 sensors) x (24 E/Q passbands)= 1344 samples is acquired for each of the electron and positive ion distributions in the E/Q range of 8 V to 52 kv. Higher E/Q resolution can be obtained by including the 24 nested E/Q passbands sampled during the subsequent fourth and fifth spin cycles. Only one of the three mass spectrometers can be sampled at a time. The mass spectrometer is selected by ground command. The mass spectrometer which is most closely viewing into the direction for rigidly corotating plasma is usually chosen. For this paper mass spectrometer I is chosen for the determination of ion composition during September 14, The field of view of this mass spectrometer spans 11 ø to 38 ø with respect to the spacecraft spin axis. During every third instrument cycle the electromagnet in the spectrometer is stepped through a sequence of 24 current steps for a given E/Q passband of the electrostatic analyzer. The exception is the occurrence of a low E/Q mode every twelfth cycle. The sequence of current steps during an instrument cycle is repeated for 20 E/Q passbands logarithmically spaced over the range of 3 V to 52 kv during one spacecraft spin. During an accumulation interval of 33 ms for each sample of sensor counting rates the spacecraft rotates by an angle of 0.62 ø. This sampling sequence is not slaved to the spacecraft rotation phase angle. Thus, given a sufficient number of these spin cycles, sampling of the ions at a given E/Q and M/Q and spacecraft rotation angle is achieved. There are two channels in the spectrometer, an integral channel which records the ion flux which is not swept out from the sensor by the magnetic field and the differential channel sensor which receives that portion of the ion flux which is deflected by a fixed angle. The responses of the differential channel are used in the present investigation. All of the measurements reported here were taken in the above high-rate recorded mode with the exception of the low-rate telemetry used to augment the recorded data for the torus passage on October 11, These low-rate data are telemetered in real time and necessarily have less dense sampling of the three-dimensional velocity distributions. On October 11 the ion plasma moments were computed with (7 sensors) x (4 sectors) x (14 passbands) = 392 samples of the velocity distribution. The accumulation time at each E/Q passband is 0.5 s. Specifically the sampling is every fourth E/Q passband in the range of 7 V to 52 kv for all seven ion sensors and for 4 rotation sectors. The background due to energetic charged particles is determined by using the spin-averaged signals at

4 6134 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS the three highest E/Q passbands in the range 15 to 52 kv. The temporal resolution for acquiring the ion velocity distribution is 600 s. 3. Observations The ion measurements for a unique set of six passages through the Io torus with the Galileo spacecraft are presented here. The first passage occurred on December 7, 1995, and the following five passages occurred on July 1-2, August 12, September 14, October 11, and November 25, all in The jovicentric radial distances for these six passages covered a combined range of ~5.5 to 8 RJ (Jupiter radii). System III longitudes of 0 ø to 360 ø were sampled by the set of six passages. The presentation of measurements for each of the six passages includes plots of the ion number density, temperature, components of bulk flow, and bulk speed as functions of universal time (UT) as computed from moments of the ion velocity distributions as defined, for example, by Krall and Trivelpiece [1973]. These plots are accompanied by N, i05 /CM XlO 6 T, ixlo 6 K KM/S v z GALILEO PLIS ß (o) I,,i,,,lil,,,,,i,,,,lll, (d) (e) Z C, Rj P, (b) C2 oooo UT C25 doi %% ! x % \ X I500,'.,0900,-, C22 woo F,7,,ilOO ' 200 _,/ --F oo poo / ooo / o4oo 18oo 25 z ,, I,, I,, I,, I,,I,, I,, I i i 0 ø 90 ø 180 ø 270 ø 360 ø 90 ø SYSTEM 1Tr LONGITUDE, X-.iT F Figure 1. Trajectory coordinates for the six Galileo traversals of the Io plasma torus. The coordinates are distance from the centrifugal equator Zc, cylindrical coordinate p with respect to the spin axis of Jupiter, and System III longitude MII. The four circles indicate positions for which temporal variations can be examined. V 0,,I,,,,, I,,,,, I,,,,, I,,,,, I,,,,,I,,,,, I,,, UT /o Rj LONG 65 ø 91 ø 119 ø 146 ø 175 ø 200 ø 226 ø Z c R d Figure 2. Ion number densities, temperatures, components of bulk velocity, and the bulk speeds for measurements in the Io torus on September 14, These plasma parameters are calculated as moments of the three-dimensional velocity distributions of the positive ions. This Galileo orbit is labeled C23, i.e., the 23rd orbit around Jupiter and the C indicates the targeted flyby moon Callisto. The coordinates defined in the caption of Figure 1 are shown along the abscissa. tables of' relevant spacecraft coordinates such as jovicentric radial distance, magnetic latitude, and distance to the centrifugal equator as functions of UT. The "04" model for the global magnetic field [Connerney et al., 1982] is used in the present paper in order to use the same coordinates employed in the comprehensive description of ion measure- ments with Voyager I in March 1979 reported by Bagenal [1994]. The model with a tilt angle (colatitude) = 9.4 ø at System III longitude = ø is used in the coordinate tables as explicitly noted [Connerney, 1993].: An overview of relevant coordinates is also given by Dessler [1983]. Our presentation is organized in chronological order with the exception that C23 is presented first because it is accompanied by plots of the directional, differential (f)

5 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6135 intensities of ions as functions of energy/charge (E/Q) for several viewing directions for the plasma analyzer, and of the mass/charge (M/Q) measurements with one of the mass spectrometers. The orbits in the Galileo mission are labeled with the moon that was targeted for close flyby and the chronological number in the orbit for the mission. Thus C23 identifies the moon Callisto during the 23rd orbit of the spacecraft around Jupiter. The Jupiter orbital insertion at the beginning of the mission is denoted as JOI. High-resolution taperecorded data are used for all of the passages with the exception of a substantial segment of I24. An overview of the passages is shown in Figure 1. The distance to the centrifugal equator, Zc., is well correlated with System III longitude MII s mply because of the offset of the magnetic with respect to the spin axis of Jupiter. The plane of the orbit of the spacecraft is nearly perpendicular to the jovian spin axis. The cylindrical coordinate p is the distance from the spin axis to the position of the spacecraft. Examination of Figure lb finds excellent coverage of the plasma torus at p = 7 (+0.5) RJ for the entire range of System III longitudes. The dashed circles or ellipse indicate four positions for which observations are available for the same position with two or more passages in order to determine whether there were measurable temporal variations. The ion plasma parameters during 1440 to 2130 UT on September 14, 1999, during orbit C23 are shown in Figure 2. In the coordinate system chosen for the presentation for the components of bulk velocities, rigid corotation of the plasma is directed along the orthogonal component (. The dotted line in Figure 2d identifies the rigid corotational speed. The observed speed along this direction is -4 km/s smaller than the values for rigid rotation. There is a short-duration decrease to speeds that are 8 km/s smaller than those for rigid corotation centered at 1740 UT. Thus the bulk flows lag behind corotation for the duration of the passage. Within the accuracy for the counting statistics of the responses of the analyzer for single-point determinations, <+2 kin/s, the bulk flow components Vp and Vz are 0 km/s. Although p varies by only 1.0 RJ, the plasma densities in Figure 2 vary by a factor of-10, i.e., from 100 to 1000 cm -3. The fluctuation in temperatures is more mild, a range of 6 x 105 to 1.4 x 106. Examination of Figure I finds that orbit C23 covers a substantial range of longitude, kiii, ~45 ø to 230 ø. This coordinate and other relevant coordinates are tabulated in Table 1. For example, the distance from the spacecraft to the centrifugal equator is RJ at 1500 UT and 0.71 RJ at 2100 UT, i.e., the spacecraft position is respectively about equidistant above and below the centrifugal equator as computed with the 04 global magnetic field model. Yet the ion densities are greatly different at these two positions. This effect is addressed in terms of a longitudinal variation in the torus after the presentation of the other five passages. At this point it is instructive to summarize the methodology in the analysis of plasma moments and in the fitting of the E/Q spectra. The M/Q of the primary ion is identified as 16 with measurements with the mass spectrometer. Then the Table 1. Spacecraft Coordinates on September 14, 1999 (C23) Universal Time, UT Jovicentric radial distance, Rj Jovicentric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg (5.7) (3.1) (-0.6) (-4.2) (-6.7) (-7.4) (-6.2) Distance to centrifugal equator, Rj (-0.77) (-0.40) (0.06) (0.48) (0.76) (0.83) (0.71)

6 6 36 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS SEPTEMBER 14, 1999 (C23) 1700: :30 UT V = 87 KM/S... M/Q = /CM 5, 20 ev M/Q = 8 50/CM 5, 20 ev... M/Q = 32 IO0/CM 3 80 ev M/Q = /CM 3, 80 ev SUM I I I Iltl P I SECTOR4 (b) _ PI SECTOR 5 (c)..--/,,., io 6.,.- 05 f/ i'"'.',,/" /'' ' i iii\ I..', /i I,',./,,,,, ':.,,,',,,,,,,d,,,,,,,,,, >, Io I 106 E i i P I SECTOR 6? 3 ø 5 (d) I I I PI SECTOR 7 (e i 29 ø _ I ' E/Q, ' ' ' I ' ' ENERGY/CHARGE, VOLTS Figure 3. Observations of the directional, differential fluxes of positive ions in units of/cm2-s-sr-v for five angular sectors for ion sensor P1 at 1700: :30 UT on September 14, The angle between the plasma bulk flow velocity and the sector field of view is. These spectra are fitted with the sum of the four thermal ion plasmas given in the label in the upper left-hand corner. The sum is identified with the thin continuous line and the measurements are indicated with the solid dots. plasma moments such as those shown in Figure 2 are computed for a value of M/Q = 16. The bulk velocity from these moments is then used for the detailed fits for the E/Q spectra such as those presented in Figure 3. The ion composition from these E/Q fits can be used to determine an average M/Q value. This M/Q value can be used to determine the inaccuracy of the assumption that the distribution was solely due to ions with M/Q- 16. For the moments calculation, bulk speed varies as (MIQ)-z/2 and number density as (M/Q) 1/2. It is also important to note that H + (M/Q = 1) has been eliminated from the moments calculation for two reasons: the relatively small densities of these ions are multiplied by a factor of 4 and their relatively low kinetic energies allow significant distortions of their angular distributions due to spacecraft charging [Frank and Paterson, 1999a]. The above calculation of plasma moments from the measured three-dimensional ion velocity dis-

7 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6137 tributions is complemented with analyses of the directional, differential E/Q spectra as shown in Figure 3. Recall that the seven ion channels, P1 through P7, provide measurements of these spectra in 56 solid angle segments on the unit sphere of 4n sr. A substantial fraction of these segments 27 have responses above the background rates due to penetrating and scattered energetic particles. The actual number of such segments with usable re- 26 sponses varies according to the bulk flow direction o with respect to the spacecraft spin axis. The responses for five solid angle segments are shown in Figure 3 for the period 1700: :30 UT on u 25 September 14, The bulk flow direction is 24 near the spin axis of the spacecraft, and hence sensor P1 is recording the maximum responses for the analyzer. The field of view for P1 is 9ø-41 ø with 25 respect to this spin axis. The bulk flow velocity of 87 km/s that is obtained from the plasma moments as shown in Figure 2 is used for the fits to the E/Q spectra shown in Figure 3. The fits are determined by best-fit iteration for the sum of the cøntributions from component MaxwellJan distributions. The possible M/Q values are guided by the fact that O and S are the dominant species in the torus on the basis of previous in situ and remote observations as summarized by Bagenal [1994]. Hence it is a matter of choosing their charge states for the fits to the E/Q spectra. Thus the dominant ions are expected to be among O +, O ++, O +++, S +, S ++, and S +++. In this case, four primary ion plasmas are required to obtain a good fit to the observed E/Q spectra. These four distrio m butions are two principal distributions at M/Q- 16 with number densities and temperatures of 300 cm -3, 80 ev and 300 cm -3, 20 ev, respectively. The two distributions with smaller densities are ions with M/Q- 8, 50 cm -3 and 20 ev, and with M/Q = 32, 100 cm -3 and 60 ev. The contributions of each of these four distributions are shown in each of the panels in Figure 3. The fits of their sums to the observed E/Q spectra are excellent and the densities and temperatures of the four component plasmas are well restricted as can be noted by inspection of the five panels in Figure 3. The measurements do not identify the ions with M/Q- 16 as O + and/or S ++. Reference to the work by Bagenal [1994] which includes the results from both the in situ plasma measurements and spectroscopy with Voyager I indicates that one of these distributions is O and the other as S ++. The fits for the lower M/Q ions were better for O ++ at 8 relative to that for S +++ at 32/3 = A substantial population of S + ions with M/Q - 32 is present in the E/Q spectra at >1 kv. Bagenal [1994] reports the presence of suprathermal ions with kt in the range of hundreds of ev in the plasma observations with Voyager 1. For example, the percentages of these suprathermal ions are ~20% for the O + ions and 7% for the S ++ ions at a radial distance of 7 RJ at the centrifugal equator, with an accuracy of 50%. Exami- GALILEO PLS 14 SEPTEMBER 1999 (C25) MASS E/Q = 300 V SPECTROMETER UT M/Q= UT (45 CYCLES) "o o o MASS PER CHARGE (M/Q) I CHANNEL 0 _- Figure 4. Responses of mass spectrometer I at energy/charge (E/Q) V during the period of UT on September 14, The responses are plotted as functions of the numbered mass/charge (M/Q) channels. The centers of the M/Q channels are indicated for several ion species. One standard deviation in the counting statistics is indicated by the error bars. Statistically significant responses above background rates are present at M/Q= 16. nation of Figures 3b, 3c, and 3e, for examples, at E/Q values >2 kv finds a small population of hot ions that may be a suprathermal component. The number densities of these ions are <10% of the total ion densities for the Galileo measurements, a percentage which is not in disagreement with those reported for Voyager 1. Bagenal [1994] interprets the hot ions observed by Voyager as locally picked up ions. Our own future analyses will require a detailed examination of the velocity distributions in order to determine whether these ions are due to pickup or whether there is an ion population with M/Q in excess of 32. Direct measurements of the plasma pitch angle distributions were not possible with the Faraday cups on the Voyager spacecraft but are offered by the segmented fields of view of the Galileo plasma instrumentation. The excellent agreement of the independent determinations of the ion and electron number densities as reported by Frank and Paterson [2000a] in the Io torus sup-

8 i 6138 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS ports the fact that no major ion distribution is undetected with the Galileo plasma analyzer. The M/Q of the primary ions is verified with the use of the measurements with mass spectrometer 1. Its field of view is positioned at 11 ø to 38 ø with respect to the spacecraft spin axis and nearly coincides with that of P1. The M/Q spectrum for E/Q V is shown in Figure 4 for the period UT. These responses have been summed over 43 sampling cycles for the mass spectrometer. A response of 25 x 103 accumulated counts in 43 sampling cycles corresponds to a sensor counting rate of 17,600 /s. There is a small increase in background rates as the gap magnetic fields (M/Q channel) increases due to the deflection of secondary electrons in the vicinity of the sensor. The background counting rates are determined for the interval UT during which the ion fluxes are less than the threshold of the mass spectrometer. During UT, statistically significant responses occur for M/Q channels 55 and 57, which correspond to M/Q = 16. Reference to the spectra in the previous Figure 3 finds that M/Q-- 16 for the major contribution from the component plasmas at E/Q- 300 V. Plasmas with M/Q-- 8 dominate at somewhat lesser E/Q values of about 200 V but the next lower mass step is taken at 110 V for which io 4 N, /CM XIO 6 T, 2XlO 6 K Vp, 0 KM/S v z I00 ioo (f) ' 27 z 26 o Figure 5. V. GALILEO PLS 14 SEPTEMBER 1999 (C25) MASS SPECTROMETER 1 E/Q = 540 V UT (43 CYCLES) 8 4 ß UT o (45 CYCLES) c, """ MASS PER CHARGE (M/Q) CHANNEL Continuation of Figure 4 for E/Q- 540 P LONG Zc o,,,, I,,,,, I,,,,, I,, UT Rj 224 ø 255 ø 279 ø R d Figure 6. Continuation of Figure 2 for Jupiter Orbit Insertion (JOI) on December 7, the responses are less than the intensity threshold of the mass spectrometer. The M/Q spectrum at the higher value of E/Q = 540 V is shown in Figure 5. Again, M/Q = 16 for the dominant ion. For an ion with given M/Q the resolution of the mass spectrometer is such that the responses are spread over two telemetered mass channels, e.g., 57 and 59 for M/Q If a significantly large component of plasma with M/Q-- 32 were present, then responses above background would be seen for channel 63 in Figure 4. The absence of these responses places an upper limit of ~20% of the total ambient ion density of 750 cm -3, or 150 cm -3 for S + ions with M/Q An average M/Q is obtained from the E/Q fits in Figure 3. The corresponding corrections for the bulk flows and number densities from the moments in Figure 2 which were computed with M/Q are a decrease of speed from 87 to 83 km/s, and a density increase 750 to 790 cm -. A second iteration could be performed for the moments but there would be little improvement in the accuracy of about +3 km/s for the speeds and +50 cm - in the total density.

9 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6139 Fluctuations in the components of ion bulk flows are evident in Figures 2c and 2d, for examples. Statistically significant variations for single-point measurements are in the range of +3 km/s. In Figure 2d, there is a transient decrease in the azimuthal plasma flow for a several minute period centered at 1745 UT. This decrease is correlated with an increase in temperature. The decrease in the flow velocity could be due to a fluctuation in the M/Q of the plasma. However, our analyses of the composition of the ions show that the magnitude of the M/Q fluctuations is insufficient to account for the variation in speed. In addition, if M/Q fluctuations were responsible for the speed fluctuations, then correlated variations in speed would consistently occur in two or more of the components of bulk flow. Examination of Figure 2 shows that this correlation does not occur. A further example on an expanded ordinate scale has been previously published by Frank and Paterson [2000a, Figure 9]. The plasma moments for the first passage through the plasma torus and the close flyby of Io on December 7, 1995, are shown in Figure 6. The coordinates are tabulated in Table 2. These observations of thermal plasmas are extensively discussed in the literature. These discussions cover the encounter with the ionosphere of Io [Frank et al., 1996], ion composition in the torus [Crary et al., 1998], production of hydrogen ions at Io [Frank and Paterson, 1999a], intense electron beams at Io [Frank and Paterson, 1999b], and plasma dynam- ics in the torus [Frank and Paterson, 2000a; Bagenal et al., 1997]. Companion reports have been published for magnetic fields [Kivelson et al., 1996], energetic charged particles [Williams et al., 1996], and plasma waves [Gurnett et al., 1996]. The time period for the observations in Figure 6 is UT during which the spacecraft moved from a jovicentric radial distance of-7.7 to 5.4 RJ. The flyby of Io is centered at-1746 UT. The plasma densities in the torus monotone increase with decreasing radial distance during UT. Torus ion densities just prior to arrival at Io were-3000 cm -3. The torus plasmas are approximately rigidly corotating with a lag in the azi- muthal component V, which is-6 km/s at 1540 UT and which decreases to 0 km/s within the instrumental uncertainties by 1700 UT. The small fluctuations in the radial motions, V,, toward and away from Jupiter wxth magmtudes xn the range of 5 to 10 km/s are indicative of interchange motions of plasmas and magnetic fields in the torus. Significantly, a radial convection speed of 5 km/s corresponds to transport across I RJ in 4 hours. During the interval , UT which is characterized by large fluctuations in the ion densities and temperatures, very intense electron beams in the energy range of hundreds of ev to several kev were detected [Frank and Paterson, 2000a]. Similar beams were also observeduring the Io flyby. Inspection of Figure 6a finds that the number density of ions decreases with decreasing radial distance after the close flyby of Io centered at 1747 UT. At the end of these recorded data at 1825 UT, there is a slight increase in densities. The radial distance of the spacecraft position at this latter time is 5.4 RJ. The ion temperatures monotone de- Table 2. Spacecraft Coordinates on December 7, 1995 (JOI) Universal Time, UT Jovicentric radial distance, Rj J ovice ntric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg (-7.5) (-6.5) (-3.9) (-1.2) (1.2) Distance to centrifugal equator, Rj (0.78) (0.64) (0.37) (0.14) (0.02)

10 614o FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS N, /CM 5 io 5 i02 I I I I I I [ I I I I I T, 2XlO 6 (b) K 0 I Vp, 0 (c) KM/S V Z V I00 I00 o o 0000 ;... - :..--_,:-_-:--_....._ _; ;_.._:_._:,,,!.d..!.. I I I I I I I I (e) (f) OlOO 0200 UT P Rj LONG 250 ø 279 ø 308 ø 0.6 Z C Rj Figure 7. Continuation of Figure 2 for orbit C21 on July 1-2, crease over time period 1810 to 1825 UT but remain hot with minimum temperatures of ~5 x 105 K. Bagenal [1994] reports the presence of a cold ion torus at radial distances of 5.0 to 5.4 RJ. The ion temperatures decrease with decreasing radial distance in this torus from ~2 x 105 K to 104 K. Bagenal et al. [1997] report that the cold ion torus was detected in the plasma wave measurements on the outbound leg of the Galileo trajectory at radial distances of ~5.0 to 5.3 RJ. These measurements provide the electron number densities, but determination of other parameters such as temperatures, ion composition, and bulk flows will have to await the possible penetration of the Galileo spacecraft into the cold torus during its later extended mission. The next extensive series of torus observations occurred during July 1-2, 1999, ~3 1/2 years after JOI. These observations are shown in Figure 7 with supporting coordinate information in Table 3. During the 2 hours of tape-recorded observations in the radial distance range of ~7.6 to 8.1 RJ the plasma densities smoothly varied by a factor of 10. As the plasma densities increased the temperature decreased from 2x 106 Kto 6 x 105 K. Plasma bulk velocities were nearly equal to those expected for rigid corotational motion, within departures of 5 km/s or less. Sporadic radial motions of the plasma toward Jupiter at speeds of several to 5 km/s occurred during the period extending from the beginning of the telemetry until ~0020 UT. The Callisto orbits during this period were used to decrease the perijove for the Galileo spacecraft in order to return it to close fiybys of Io. These were exciting times for the Galileo mission in Table 3. Spacecraft Coordinates on July 2, 1999 (C21) Universal Time, UT Jovicentric radial distance, Rj Jovicentric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg (-4.5) (-1.3) (1.8) Distance to centrifugal equator, Rj (0.64) (0.18) (-0.24)

11 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6141 N, I05 /CM 5 5 T, ixio G KM/S v z V 102 t IOO 0,,,,I,,,,ll,,,,,I,,,,,I,,,, o (f) ,,I,,,,, I,,,,,I,,,,,I,, I000 I UT P Rj LONG 216 ø 245 ø 275 ø 501 ø ø Z C Rj Figure 8. Continuation of Figure 2 for orbit C22 on August 12, (a) terms of the continued great scientific yield of the mission and the anxieties for the survival of the spacecraft with increased radiation dosage. The next passage through the torus was C22 on August 12, The plasma moments are shown in Figure 8 and relevant coordinates are provided in Table 4. Of principal note for this passage is the greater lag in azimuthal flow below that expected for rigid corotation. This lag ranges between 10 to 15 km/s for the period of UT. It is of interest to note that super-rotation, i.e., azimuthal flows at speeds greater than that for rigid rotation, are not observed for any of the torus passages reported here. On the other hand, an extended period of continuous radial flow at several to 5 km/s away from Jupiter is observed during the period UT. Several of the fits for the observed E/Q spectra observed during 1030: :07 UT are shown in Figure 9. These five panels are presented for several purposes: (1) the relative content of the four major ion distributions is similar to that for C23 as shown in Figure 3, (2) an example in Figure 9e, which shows an occasional sector with one of the relatively poorer fits, and (3) an example of the almost ubiquitous presence of hydrogen ions at E <80 V in Figure 9e. The feature 2 above sug- gests that a substantial fraction of the ion distribution may not be well fit with a Maxwellian distribution but may be an evolved velocity distribution from ion pick up. In general, the densities and temperatures in Figures 3 and 9 are restricted by the requirement that the E/Q spectra must be fit in Table 4. Spacecraft Coordinates on August 12, 1999 (C22) Universal Time, UT Jovicentric radial distance, Rj Jovicentric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg (-7.0) (-4.9) (-2.0) (1.1) (3.8) (5.9) Distance to centrifugal equator, Rj (0.93) (0.64) (0.26) (-0.13) (-0.48) (-0.75)

12 6142 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 107 AUGUST 12, 1999 (C22) 1050: :07 UT 106 V = 92 KM/S M/Q = /CM 3 20eV M/Q = 8 IO0/CM 3 20 ev M/Q =32 IO0/CM 3 60eV M/Q = /CM 3 60 ev SUM I PI SECTOR 5 I $=24 ø (c) I I I IIIII P I SECTOR 5 (c I I I is 52 ø _ 105 I I I IIII _ I I -= - P2 SECTOR 3 ( - - ß i o I I P2 SECT 4 (e I:. ß '-. 51ø 105 I I I i I II1! I0 102 I I E/Q, ENERGY/CHARGE, VOLTS Figure 9. Continuation of Figure 3 for positive ions measured in three angular sectors with sensor P1 and two sectors with P2. Hydrogen ions are detected with P2 at E/Q <80 V. five or more directions. The corresponding accuracy of the density of any one of the ion components is about +20%. Specifically, the two dominant ion populations are characterized by M/Q- 16, with temperatures and densities of 400 cm -3 and 20 ev, and 300 cm -2 and 60 ev. The two ions with smaller but still substantial densities have M/Q = 8 with temperature 20 ev and density 100 cm -3 and M/Q = 32 with 60 ev and 100 cm -3. The average M/Q of the entire distribution is 16.8 which provides a correction of the bulk speed from 92 to 90 km/s and an increase in total ion number densities from 900 to 920 cm -3. The perijove of the Galileo orbit was reduced sufficiently that the spacecraft provided a close flyby of Io during I24 on October 11, The plasma moments are shown in Figure 10 with the coordinates given in Table 5. The closest approach to Io occurred at 0433 UT at an altitude of 617 km. The fluctuations in torus ion densities and diversion of the torus ion flows around Io are readily evident in Figure 10. A detailed description of the plasmas during the flyby is given by Frank and Paterson [2000b]. Measurements of the torus plasmas during UT, and outside of the direct Io interaction, were considered sufficiently important to use the low telemetry rate data which were gained in real-time transmissions from the spacecraft in the absence of the tape-recorded data. These plasma moments are shown as solid dots in Figure 10. The penalties for use of the low-rate data are the obvious loss of time resolution and

13 N, I05 /CM 5 I02 2XlO 6 T, I X 106 K KM/S Vz IOO o FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS ,,I, UT P Rj LONG 67* 92* 118' 145' 173' Z C R d greater inaccuracies, particularly in the bulk flow components, due to the reduced number of E/Q samples and number of angular sectors for the arrival directions of the plasma at the instrument. The Vz component of-10 km/s is an example of the greater uncertainty. Detailed examination of the sampling of the velocity distribution shows that this offset from 0 km/s can be expected. The accuracy of the ion number densities and temperatures is less affected by this reduced sampling of the ion velocity distributions. It is fortunate that the temporal variations of the ion plasmas were small during this period. The final passage through the torus which is given in our present work occurre during I25 on November 25, These plasma moments are shown in Figure! 1. The accompanying table of coordinates is provided in Table 6. An unprogramreed spacecraft sating event occurred before the spacecraft arrived at Io. The variations of the plasma moments during the ~3-hour duration of the tape recorded data prior to the sating were minimal as the spacecraft moved toward Jupiter. The largest variation was a smooth increase in ion densities from -500 to 1000 cm -3. The lag in azi- muthal motion was only several kilometers per second relative to rigid corotation. During the period of UT the bulk flow was marginally directed radially outward from Jupiter at 2 to 3 km/s. A radial speed of 2 km/s provides for a transport of plasma over a distance of I RJ in 10 hours if, of course, this motion is sustained. Figure 10. Continuation of Figure 2 for orbit I24 with the close flyby of Io on October 11, For all of the measurements reported in this paper the high-resolution data from the tape recorder are 4. Sununary and Discussion used with the exception of the low-rate real-time data employeduring UT during this orbit. We have presented the ion densities, temperatures, components of the bulk velocities, and the Table 5. Spacecraft Coordinates on October 11, 1999 (I24) Universal Time, UT 04OO O5O O7OO O8OO Jovicentric radial distance, Rj O Jovicentric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg Distance to centrifugal equator, Rj (5.4) (3.2) (-0.3) (-4.0) (-6.7) (-0.55) (-0.33) (0.03) (0.48) (0.86) (04 Mødel)

14 6!44 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS N /CM 5 T KM/S v z io ixlo 6 ioo IOO... i i i i i UT / Rj LONG 552 ø 0 ø 27 ø 52':' Z c -0., R d Figure 11. Continuation of Figure 2 for orbit I25 on November 25, ( ) bulk speeds for six passages of the Galileo spacecraft through the Io torus of Jupiter. These measurements with the plasma instrumentation (PLS) were acquired on December 7, 1995 (JOI), and on July 1-2 (orbit C21), August 12 (C22), September 14 (C23), October 11 (I24), and November 25 (I25), all in The trajectories through the torus and the ion number densities for the six passages are summarized in Figure 12. In Figure 12a are also shown the isodensity contours for ions as reported by Bagenal [1994] from Voyager I plasma observations and remote spectrometer measurements of the torus emissions during the flyby of Jupiter in (c) March The observed densities with the Galileo plasma instrument are shown for each of the six passages in Figure 12b. The symbols can be associated with the individual passages with inspection of the trajectories in companion Figure 12a. The solid line shows the ion densities as a function of cylindrical radius p from the planet's center and at the centrifugal equator, Z c = 0, for the Voyager I results given by Bagenal [1994]. Examination of Figure 12 reveals several important features of the plasma distributions. First of all, the distributions in the vicinity of Io's orbit at <6 RJ exhibit substantial spatial and/or temporal variability. The complex plasma distributions which include large densities of pickup ions from charge exchange of the streaming torus plasmas with the Io neutral atmosphere/exosphere and/or electron impact have been previously discussed for JOI [Frank et al., 1996] and I24 [Frank and Paterson, 2000b]. The hot plasmas in the Io torus that are located in the radial range of 6 to 8 RJ are of Table 6. Spacecraft Coordinates on November 25, 1999 (I25) Universal Time, UT Jovicentric radial distance, Rj Jovicentric System III west longitude, deg Magnetic Local time Jovicentric latitude of centrifugal equator, deg (4.1) (5.8) (6.5) (6.1) Distance to centrifugal equator, Rj (-0.48) (-0.66) (-0.70) (-0.64) (O4 Model)

15 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6145 VOYAGER 1 ISODENSITY CONTOURS (BAGENAL, 1994)' JOI (a) C21 Z C, Rj -I 1600 C25 200/CM N! /CM I0 _, VOYAGER 1.: - A,., O,,I,, I,, GALILEO, i. i.. I I.. I I,. ',... I P, R d Figure 12. (a) Plasma number densities as measured with the plasma analyzer on Voyager 1. These density contours are shown in the p-zc plane and the six Galileo trajectories are superposed. Co) Number densities at the centrifugal equator from Voyager I and the densities observed along the six trajectories of Galileo in the Io plasma torus. primary interest to our present work. The trajec- C21, C22, and C23, which provide cuts through the tory for JOI is positioned above the centrifugal spatial distributions at more-or-less constant valequator and, using the Voyager I contours in Fig- ues of p. For example, the Voyager I prediction for ure 12a, the Galileo densities projected to the the density variation during C22 is approximately equator would be about a factor of 2 greater than a factor of 4. The observed variation is a factor of Voyager I number densities. Even though the ~100. The reader has also probably noticed in Figplasma analyzers on these two spacecraft were ure 12a that the C22 trajectory is symmetric about very dissimilar, there is a high level of confidence the centrifugal equator but that the density profile that temporal variations were responsible for the is greatly asymmetric. This feature is discussed higher densities encountered by Galileo. The spa- later in terms of an "active" longitude sector in the tial variations as functions of distance from the torus. centrifugal equator significantly differ from those Further insight into the present findings of thergiven by the isodensity contours for Voyager 1. mal ion observations with the Galileo spacecraft This is particularly evident for the trajectories for can be gained with inspection of Figure 1. The

16 6146 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS Table 7. Comparison of Ion Densities and Temperatures for Four Torus Locations With Same p, Zc and km Location Orbit Date UT N,/cm 3 T, øk 1 C23 Sept. 14, x 10 ' 6 x 100 I24 Oct. 11, x 10 ' 7 x JOI Dec. 7, x 10 ' 6 x 100 C23 Sept. 14, x 10 ' 6 x JOI Dec. 7, x 10 ' 8 x 100 C22 Aug. 12, x x 100 C21 July 2, x 10 ' 13 x 100 C22 Aug. 12, x 10 ' 7 x 100 I25 Nov. 25, x 10 ' 6 x 100 coverage of System III longitude is complete in the radial distance range of ~6.5 to 7.5 RJ in the hot torus. There are four zones identified by dashed circles or an ellipse in Figure lb for which two or three of the passages intersect the same coordinates, i.e., at the same p, kiii, and Z c. The ion densities and temperatures are given for these four locations in Table 7. With the exception of the measurement at 0150 UT for C21 the temperatures are all in the range of 7(+1) x 105 K and are equal to the accuracy with which they can be determined. The C21 observation is at the outer edge of the hot torus (see Figure 12a) where temporal variations are more likely to occur. Thus the heart of the hot torus is quite stable with regard to the ion temperature. The densities are also quite stable among the five orbits during July through Novem- ber On the other hand, it is clear that the ion densities at locations 2 and 3 were greater by factors of 2 and 3 during JOI in December The magnitude of these density variations is similar to the brightness fluctuations of ultraviolet emissions from sulfur ions observed by the International Ultraviolet Explorer (IUE) during the period 1979 to 1985 [Moos et al., 1985]. From the present measurements the time scale for significant fluctuations in the densities of the hot torus is inferred to be in excess of several months but less than several years, i.e., in the range of 107 to 108 s. This should not be extended to the orbit of Io where large fluctuations from ionization of this moon's exosphere and atmosphere are encountered. The estimates of radial diffusion times for the plasmas in the range of~107 s by Cheng [1986] appear to be sufficient to provide a portion of the hot torus plasmas by diffusion from the orbit of Io and/or an inwardly diffusing source at greater distances beyond the torus [Gehrels and Stone, 1983; Thorne, 1983]. An active sector that exhibits intensity enhancements of S + and S ++ emissions has been identified by Pilcher and Morgan [1985] as located in the System III longitude sector of ~180 ø to 230 ø. Reference to Figure lb finds that the passage during the insertion of Galileo into orbit around Jupiter, i.e., JOI, was positioned within this longitudinal sector. It is also to be noted that orbits C22 and C23 provide measurements of the plasmas at approximately constant radial distance both inside and outside of this active sector. Examination of the ion densities in Figure 2 for C23 and Figure 8 for C22 reveals the remarkable presence of considerably greater densities off of the centrifugal equator in the active sector relative to those in other sectors. This effect can be also seen by comparison of the corresponding density profiles in Figure 12b with the trajectories in Figure 12a. These observations establish an important correlation of increased plasma densities at positions ~0.5 RJ away from the centrifugal equator in the active sector. Examination of the ion temperatures and the bulk flows do not reveal any further solid clues as to unique characteristics of this active sector. Perhaps a subsequent study of electron beams and electron thermal plasmas will yield further information concerning this remarkable phenomenon. It is relevant to note that the Voyager I torus measurements were acquired near and within the active sector. The System III longitude and the distance from the centrifugal equator were 135 ø and RJ at a radial distance of 7.5 RJ, and 193 ø and 0.15 at 6.0 RJ. Inspection of Figure 1 shows that orbit C22 is located in the active sector at the beginning of the pass where the densities are considerably higher than at the pass's end when the spacecraft was outside of the active sector but at similar distances from the centrifugal equator. The density profile is plotted in Figure 8. Then reference to Figure 12b finds that the variation in densities observed in the active sector during C22 is about a factor of 3 for an excursion of ~1 RJ above the centrifugal equator. This factor is

17 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6147 similar to that given by Bagenal [1994] for the modeled Voyager I contours as reproduced in our Figure 12a. Thus the Galileo and Voyager results concerning the latitudinal variations are in agreement if the latter spacecraft was located in an active sector with longitudinal extent larger than that reported by Pilcher and Morgan [1985]. Two examples of the analyses of the E/Q spectra are presented, one for the torus passage on September 14, 1999, in Figure 3 and the other for August 12, 1999, in Figure 9. The fits to these ion spectra are acquired by iteration for various combinations of the species, densities and temperatures with the assumption that their distributions are MaxwellJan. In general, the fits to the spectra are good with four primary species, two species with M/Q- 16 with two temperatures, respectively, and two species with smaller densities and with M/Q = 8 and 32, respectively. Remote measurements of the emissions from the torus were used to determine that the two principal ions were O + and S ++ for the Voyager I plasma analyzer [Bagenal, 1994]. This would be in agreement with the present results with the Galileo plasma analyzer, although it is not possible to directly determine which of the two species were the hotter and cooler distributions, respectively. The temperatures of the two distributions were approximately 20 ev and 60 to 80 ev, respectively, with densities in the range of 300 to 400 cm -3 for the spectral fits reported here. The ion temperatures in the hot torus are reported to be in the range of 60 to 80 ev for the Voyager measurements [Belcher, 1983; Bagenal, 1994]. The reader is reminded that the Galileo plasma instrumentation provided no plasma measurements at radial distances!5.4 RE where the Voyager spacecraft encountered colder ion plasmas with temperatures <10 ev. Measurements with the mass spectrometers in the Galileo plasma instrument verified that M/Q- 16 for the two primary ions. The two ion distributions with M/Q = 16 and with different temperatures observed with the Galileo analyzer are reasonable when it is noted that the maximum energy for the initial pickup of S + in its cycloidal motion is a factor of 2 greater than that for O +. Subsequent thermalization can be expected to yield higher temperatures of the S + relative to O +. Further ionization from electron impact, for example, will not significantly change the thermal energy of the plasmas. Ion collision times are long, ~107 s, in the torus [Smith and Strobel, 1985] and collisional equilibration of S ++ and O + temperatures will be slow. Thus we iden- tify the hotter component at M/Q- 16 as S ++, and the cooler component as O +. Note that this interpretation predicts similar temperatures for O + and O ++. Indeed, the O ++ ions detected with the Galileo plasma analyzer are also cool with a temperature in the range of 20 ev, and densities are in the range of 50 to 100 cm -3 for the two examples shown here. Thus the number density of this com- ponent with M/Q- 8 is significantly less than those of each of the heavier ions with M/Q The E/Q fits for these ions favored M/Q - 8, O +, relative to those for M/Q- 32/3, S +++. The fourth ion detected with the Galileo plasma analyzer is S + with M/Q = 32 and has densities and temperatures in the range of 100 cm -3 and 60 to 80 ev, respectively. The relative ion compositions as reported by Bagenal [1994] for the Voyager I observations are in general agreement with the present Galileo measurements for O +, S +, and S ++. The exception is the significantly higher densities of O ++ as detected with the Galileo plasma analyzer. The presence of hydrogen ions in the hot torus is seen in Figure 9e at E/Q < 80 V. The presence of measurable fluxes of hydrogen ions appears to be almost ubiquitous in the torus plasmas [Frank and Paterson, 1999a, 2000b]. Resolution of the primary sources of the plasmas in the hot torus will take considerably more analyses of the ion velocity distributions as observed with the plasma analyzer. The "neutral cloud theory" proposes that this source is due to the ionization of the neutral cloud by ionizing collisions of electrons and ions with the neutral atoms from Io which pervade the volume of the torus [Smith and Strobel, 1985]. On the other hand, Shemansky [1988] has examined in detail the energy branching in the torus and has concluded that the neutral cloud source must be augmented by another source of ions. One of his suggestions for a viable source is pickup ions from Io's atmosphere. Examination of the large variability of the ions at radial distances near the orbit of Io as shown in Figure 12b attests to the fact that this moon is a significant source of ions. The plasma measurements during the two close fiybys, JOI and I24, provide direct evidence of a copious supply of pickup ions [Frank et al., 1996; Frank and Paterson, 2000b]. A future study of the evolution of the details of the ion velocity distributions as a function of radial distance from Io's orbit to the outer edge of the hot torus will attempt to identify the principal mode of ion supply. This future analysis will necessarily include the substantial pickup of ions from Europa's atmosphere at its radial distance of 9.4 RJ [Paterson et al., 1999]. The existence of significant plasma transport by interchange motions of magnetic flux tubes in the hot torus was previously inferred from observations of energetic particles, plasma waves and magnetic fields with the Galileo spacecraft during JOI [Thorne et al., 1997; Bolton et al., 1997; Kivelson et al., 1997]. The inferred speed of these interchange motions was about 100 km/s directed ra- dially inward toward Jupiter. Later, Frank and Paterson [2000a] confirmed the interchange motion for the event with longest time duration, several minutes, with direct measurements of the bulk speed and densities of the thermal plasmas. However,. the radial speed was only in the range of I to 5 kilometers per second. For the present survey of

18 6148 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS five additional orbits such radial motions, both toward and away from Jupiter, are sporadically observed with speeds in the range of several km/s. It is important to note that there is a substantial number of flow events that deviate from those for rigid corotation which do not exhibit the simultaneous flow toward Jupiter and an accompanying density decrease expected for interchange motions. Other such plasma motions include large-scale convective flows and smaller-scale turbulence. The high-resolution measurements of the plasma bulk flows in the hot torus during five of the six passages through this region show that the azimuthal component of the bulk flow lags that for rigid corotation. This lag is in the range of 2 to 10 km/s, and the average is -2 to 3 km/s. One of the 6 passages, I24 on October 11, 1999, was taken with low-rate real-time telemetry transmission and the corresponding angular resolution for the direction of bulk flow was insufficient to accurately determine the small lag relative to rigid corotation. Azimuthal flows that exceeded those for rigid corotation, i.e., "supercorotation," were not found in the present series of observations. The average lag in the corotational flow supports the existence of a System IV longitudinal period which is longer by 3% relative to that for System III [Sandel and Dessler, 1988]. Kaiser and Desch [1980] have previously reported such a longer period for narrowband kilometric radio emissions. A similar longer period was found for visible emissions from the torus [Roesler et al., 1984]. Our present observations of a persistent lag in the azimuthal bulk flow of torus plasmas add further impetus to theoretical interpretations of this interesting dynamical phenomenon. Possible explanations are mass loading in the torus [Hill, 1979, 1980] or differential rotation of Jupiter's magnetic fields as a function of their latitude [Dessler, 1985]. The present Galileo plasma observations suggest that the mechanism may have to account for an azimuthal lag which is not constant as a function of longitude and/or time. Acknowledgments. The research at The University of Iowa was supported in part by the Jet Propulsion Laboratory under contract JPL Janet G. Luhmann thanks Edward C. Sittier and an- other referee for their assistance in evaluating this pa- per. References Bagenal, F., Plasma conditions inside Io's orbit, J. Geophys. Res., 90, , Bagenal, F., Torus-magnetosphere coupling, in Time- Variable Phenomena in the Jovian System, edited by M. J. S. Belton, R. A. West, and J. Rahe, pp , NASA Spec. Publ., SP-494, Bagenal, F., Empirical model of the Io plasma torus: Voyager measurements, J. Geophys. Res., 99, 11,043-11,062, Bagenal, F., F. J. Crary, A. I. F. Stewart, N.M. Schneider, D. A. Gurnett, W. S. Kurth, L. A. Frank, and W. R. Paterson, Galileo measurements of plasma density in the Io torus, Geophys. Res. Lett., 24, , Bagenal, F., and J. D. Sullivan, Direct plasma measurements in the Io torus and inner magnetosphere of Jupiter, J. Geophys. Res., 86, , Bagenal, F., J. D. Sullivan, and G. L. Siscoe, Spatial distribution of plasma in the Io torus, Geophys. Res. Lett., 7, 41-44, Belcher, J. W., The low-energy plasma in the Jovian magnetosphere, in Physics of the Jovian Magnetosphere, edited by A. J. Dessler, pp , Cambridge Univ. Press, New York, Bigg, E. K., Influences of the satellite Io on Jupiter's decametric emission, Nature, 203, , Bolton, S. J., R. M. Thorne, D. A. Gurnett, W. S. Kurth, and D. J. Williams, Enhanced whistler-mode emissions: Signatures of interchange motion in the Io torus, Geophys. Res. Lett., 24, !26, Brice, N.M., and T. R. McDonough, Jupiter's radiation belts, Icarus, 18, , Bridge, H. S., et al., Plasma observations near Jupiter: Initial results from Voyager 1, Science, 204, , Broadfoot, A. L., et al., Extreme ultraviolet observations from Voyager I encounter with Jupiter, Science, 204, , Brown, R. A., and F. H. Chaffee, High-resolution spectra of sodium emission from Io, Astrophys. J., 187, L125- L126, Cheng, A. F., Radial diffusion and ion partitioning in the Io torus, Geophys. Res. Lett., 13, , Connerney, J. E. P., Magnetic fields of the outer planets, J. Geophys. Res., 98, 18,659-18,679, Connerney, J. E. P., M. H. Acufia, and N. F. Ness, Voyager I assessment of Jupiter's planetary magnetic field, J. Geophys. Res., 87, , Crary, F. J., F. Bagenal, L. A. Frank, and W. R. Paterson, Galileo plasma spectrometer measurements of composition and temperature in the Io plasma torus, J. Geophys. Res., 103, 29,359-29,370, Dessler, A. J., Coordinate systems, in Physics of the Jovian Magnetosphere, edited by A. J. Dessler, pp , Cambridge Univ. Press, New York, Dessler, A. J., Differential rotation of the magnetic fields of gaseous planets, Geophys. Res. Lett., 12, , Frank, L. A., and W. R. Paterson, Production of hydrogen ions at Io, J. Geophys. Res., 104, 10,345-10,354, 1999a. Frank, L. A., and W. R. Paterson, Intense electron beams observed at Io with the Galileo spacecraft, J. Geophys. Res., 104, 28,657-28,669, 1999b. Frank, L. A., and W. R. Paterson, Observations of plasmas in the Io torus with the Galileo spacecraft, J. Geophys. Res., 105, 16,017-16,034, 2000a. Frank, L. A., and W. R. Paterson, Return to Io by the Galileo spacecraft: Plasma observations, J. Geophys. Res., 105, 25,363-25,378, 2000b. Frank, L. A., K. L. Ackerson, J. A. Lee, M. R. English, and G. L. Pickett, The plasma instrumentation for the Galileo mission, Space Sci. Rev., 60, , Frank, L. A., W. R. Paterson, K. L. Ackerson, V. M. Vasyliunas, F. V. Coroniti, and S. J. Bolton, Plasma observations at Io with the Galileo spacecraft, Science, 274, , Gehrels, N., and E. C. Stone, Energetic oxygen and sulfur ions in the jovian magnetosphere and their contribution to the auroral excitation, J. Geophys. Res., 88, , Goldreich, P., and D. Lynden-Bell, Io, a Jovian unipolar inductor, Astrophys. J., 156, 59-78, Gurnett, D. A., W. S. Kurth, A. Roux, S. J. Bolton, and C. F. Kennel, Galileo plasma wave observations in the Io plasma torus and near Io, Science, 274, , Hill, T. W., Inertial limit on corotation, J. Geophys. Res., 84, , Hill, T. W., Corotation lag in Jupiter's magnetosphere: Comparison of observation and theory, Science, 207,

19 FRANK AND PATERSON: THERMAL IONS IN JUPITER'S TORUS 6149 Hill, T. W:, and F. C. Michel, Heavy ions from the Galilean satellites and the centrifugal distortion of the Jovian magnetosphere, J. Geophys. Res., 81, , Kaiser, M. L., and M.D. Desch, Narrow-band jovian kilometric radiation: A new radio component, Geophys. Res. Lett., 7, , Kivelson, M. G., K. K. Khurana, C. T. Russell, and R. J. Walker, Intermittent short-duration magnetic field anomalies in the Io torus: Evidence for plasma interchange?, Geophys. Res. Lett., 24, , Kivelson, M. G., K. K. Khurana, R. J. Walker, J. Warnecke, C. T. Russell, J. A. Linker, D. J. Southwood, and C. Polansky, Io's interaction with the plasma torus: Galileo magnetometer report, Science, 274, , Krall, N. A., and A. W. Trivelpiece, Principles of Plasma Physics, McGraw-Hill, New York, Kupo, I., Y. Mekler, and A. Eviatar, Detection of ionized sulfur in the Jovian magnetosphere, Astrophys. J., 205, L51-L53, Moos, H. W., T. E. Skinner, S. T. Durrance, P. D. Feldman, M. C. Festou, and J.-L. Bertaux, Long-term stability of the Io high-temperature plasma torus, Astrophys. J., 294, , Neubauer, F. M., Nonlinear standing Alfv n wave current system at Io: Theory, J. Geophys. Res., 85, , Paterson, W. R., L. A. Frank, and K. L. Ackerson, Galileo plasma observations at Europa: Ion energy spectra and moments, J. Geophys. Res., 104, 22,779-22,791, Piddington, J. H., and J. F. Drake, Electrodynamic effect of Jupiter's satellite Io, Nature, 277, , Pilcher, C. B., and J. S. Morgan, Magnetic longitude variations in the Io torus, Adv. Space Res., 5, , Roesler, F. L., F. Scherb, and R. J. Oliversen, Periodic intensity variation in [SIII] 9531A emission from the Jupiter plasma torus, Geophys. Res. Lett., 11, , Sandel, B. R., and A. J. Dessler, Dual periodicity of the jovian magnetosphere, J. Geophys. Res., 93, , Shemansky, D. E., Energy branching in the Io plasma torus: The failure of neutral cloud theory, J. Geophys. Res., 93, , Siscoe, G. L., Jovian plasmaspheres, J. Geophys. Res., 83, , Siscoe, G. L., and C.-K. Chen, Io, A source for Jupiter's inner plasmasphere, Icarus, 31, 1-10, Siscoe, G. L., A. Eviatar, R. M. Thorne, J. D. Richardson, F. Bagenal, and J. D. Sullivan, Ring current impoundment of the Io plasma torus, J. Geophys. Res., 86, , Smith, R. A., and D. F. Strobel, Energy partitioning in the Io plasma torus, J. Geophys. Res., 90, , Strobel, D. F., Energetics, luminosity, and spectroscopy of Io's torus, in Time-Variable Phenomena in the Jovian System, edited by M. J. S. Belton, R. A. West, and J. Rahe, pp , NASA Spec. Publ., SP-494, Thorne, R. M., Microscopic plasma processes in the jovian magnetosphere, in Physics of the Jovian Magnetosphere, edited by A. J. Dessler, pp , Cambridge Univ. Press, New York, Thorne, R. M., T. P. Armstrong, S. Stone, D. J. Williams, R. W. McEntire, S. J. Bolton, D. A. Gurnett, and M. G. Kivelson, Galileo evidence for rapid interchange transport in the Io torus, Geophys. Res. Lett., 24, , Williams, D. J., B. H. Mauk, R. E. McEntire, E. C. Roelof, T. P. Armstrong, B. Wilken, J. G. Roederer, S. M. Krimigis, T. A. Fritz, and L. J. Lanzerotti, Electron beams and ion composition measured at Io and its torus, Science, 274, , L. A. Frank and W. R. Paterson, Department of Physics and Astronomy, 212 Van Allen Hall, University of Iowa, Iowa City, IA (louis-frank@ uiowa.edu; william-paterson@uiowa.edu) (Received April 25, 2000; revised November 13, 2000; accepted November 13, 2000.)