thermal ion measurements on board interball … ion measurements on board interball auroral probe by...

Thermal ion measurements on board Interball Auroral Probeby the Hyperboloid experiment

N. Dubouloz1, J.-J. Berthelier1, M. Malingre1, L. Girard1*, Y. Galperin2, J. Covinhes1, D. Chugunin2, M. Godefroy1,G. Gogly1, C. GueÂ rin1, J.-M. Illiano1, P. Kossa1, F. Leblanc1, F. Lego�1, T. Mularchik2, J. Paris1, W. Stzepourginski1,F. Vivat1, L. Zinin3

1Centre d'eÂ tude des Environnements Terrestre et PlaneÂ taires, Centre National de la Recherche Scienti®que, 4 avenue de Neptune,F-94107 Saint-Maur Cedex, France; Tel: + 1 45 11 42 62; Fax: + 1 48 89 44 33; e-mail: [email protected] Research Institute, Russian Academy of Science, Profsoyuznaya str., 84/32, GSP-7, Moscow 117810, Russia3University of Kaliningrad, Kaliningrad, Russia

Received: 25 August 1997 /Revised: 04 May 1998 /Accepted: 04 May 1998

Abstract. Hyperboloid is a multi-directional mass spec-trometer measuring ion distribution functions in theauroral and polar magnetosphere of the Earth in thethermal and suprathermal energy range. The instrumentencompasses two analyzers containing a total of 26entrance windows, and viewing in two almost mutuallyperpendicular half-planes. The nominal angular resolu-tion is de®ned by the ®eld of view of individual windows�13° ´ 12.5°. Energy analysis is performed using spher-ical electrostatic analyzers providing di�erential mea-surements between 1 and 80 eV. An ion beam emitter(RON experiment) and/or a potential bias applied toHyperboloid entrance surface are used to counteractadverse e�ects of spacecraft potential and thus enableion measurements down to very low energies. A mag-netic analyzer focuses ions on one of four micro-channelplate (MCP) detectors, depending on their mass/chargeratio. Normal modes of operation enable to measureH+, He+, O++, and O+ simultaneously. An automaticMCP gain control software is used to adapt theinstrument to the great ¯ux dynamics encounteredbetween spacecraft perigee (700 km) and apogee(20 000 km). Distribution functions in the main analyz-er half-plane are obtained after a complete scan ofwindows and energies with temporal resolution betweenone and a few seconds. Three-dimensional (3D) distri-butions are measured in one spacecraft spin period(120 s). The secondary analyzer has a much smallergeometrical factor, but o�ers partial access to the 3Ddependence of the distributions with a few secondstemporal resolution. Preliminary results are presented.Simultaneous, local heating of both H+ and O+ ionsresulting in conical distributions below 80 eV is ob-served up to 3 Earth's radii altitudes. The thermal ionsignatures associated with large-scale nightside magne-

tospheric boundaries are investigated and a new ionout¯ow feature is identi®ed associated to the polar edgeof the auroral oval. Detailed distribution functions ofinjected magnetosheath ions and ou¯owing cleft foun-tain ions are measured down to a few eVs in the dayside.

Key words. Ionosphere (auroral ionosphere; particleacceleration; ionosphere-magnetosphere interactions)

1 Introduction and scienti®c objectives

Space missions conducted over the last 15 years have ledto the gradual understanding that low-energy (below afew hundreds of eV) plasmas signi®cantly contribute tothe processes which govern the coupling between theEarth's ionosphere and magnetosphere. As a matter offact, thermal and slightly suprathermal plasma oftenrepresents the dominant contribution to the localnumber and/or energy density, and therefore stronglydrives the global dynamics of the ionoshperic andmagnetospheric system, especially in the ideal MHDcontext of frozen-in magnetic ®eld lines. Moreover,thermal plasma supports the propagation of electrostat-ic and electromagnetic waves, and in some casesprovides the necessary free energy for microscopicinstabilities.

It is now widely admitted that a signi®cant amount ofthe magnetospheric plasma is of ionospheric, and notsolar wind, origin. First indirect evidence of ionosphericparticles escaping along magnetic ®eld lines was thediscovery of energetic O+ ions in the magnetosphere(Shelley et al., 1972). The existence of low-energy H+

and He+ out¯ows was supported in the 1970s by theExplorer 31 (Ho�man, 1970), OGO 2 (Taylor andWalsh, 1972), and ISIS 2 (Ho�man et al., 1974)spacecrafts. Direct quantitative measurements of the

Correspondence to: N. Dubouloz*L. Girard now at Centre National d'Etudes en TeÂ leÂ communica-tions, France Telecom, 38-40 avenue du GeÂ neÂ ral Leclerc, F-92794Issy Cedex, France

Ann. Geophysicae 16, 1070±1085 (1998) Ó EGS ± Springer-Verlag 1998

characteristics of these out¯ows were obtained duringthe Dynamics Explorer (DE) and Akebono missions.They revealed a rather complex picture of ion escapeprocesses, involving various regimes associated withmore or less speci®c geophysical regions. Because theseescape processes are most commonly observed at highgeomagnetic altitudes, they are often referred to as``polar wind'', without distinction. A review of theoret-ical, numerical, and experimental studies of polar windcan be found in Ganguli (1996). In the polar cap,supersonic H+ and He+ out¯ows are frequently ob-served (Nagai et al., 1984). Their properties are remi-niscent of the ``classical'' polar wind model, in which®eld-aligned motion is driven by pressure gradients andby the ambipolar electric ®eld (Banks and Holzer, 1969).Classical polar wind theory, however, is not able toexplain other ion out¯ows characterized by tempera-tures much larger than typical ionospheric temperatures,or by the presence of heavy (mainly O+) ions alsoescaping along the ®eld lines (Gurgiolo and Burch, 1982;Green and Waite, 1985). Such processes are oftenobserved in auroral regions, and obviously require acomplementary acceleration mechanism. Many theoret-ical and numerical models have been developed in orderto explain these ``generalized'' polar wind phenomena.They include various causes of ion upward acceleration,such as heating by wave-particle interactions or Joulee�ect, and hot photoelectrons or magnetospheric elec-trons inducing enhanced ambipolar electric ®elds. Thehorizontal plasma dynamics associated to magneto-spheric convection is a further factor acting on ionout¯ows. First, convection transports the magnetic ®eldlines across the auroral and polar regions; as a result,out¯ows along these ®eld lines experience time-depen-dent boundary conditions which strongly a�ect theircharacteristics (Schunk and Sojka, 1989). At lowaltitudes, the relative horizontal drift between ions andneutrals also produces signi®cant frictional heatingleading to additional upward acceleration (St-Mauriceand Schunk, 1979). Finally, E ´ B convection contrib-utes to O+ acceleration at higher altitudes where ®eldlines have smaller curvature radii; this e�ect can bedescribed as centrifugal acceleration in the convectingplasma frame of reference (Cladis, 1986). The respectiverole of the many physical processes involved in iono-spheric out¯ows has still to be determined; this theoret-ical e�ort will need to rely on new, higher resolutionthermal plasma measurements.

As already stressed, acceleration due to wave-particleinteractions is thought to constitute a signi®cant ingre-dient of the generalized polar wind mechanism. Severaldi�erent kinds of accelerated ion distributions can bedistinguished at energies up to a few hundreds of eVs.Upward ®eld-aligned beams are frequently observed andinterpreted in terms of acceleration parallel to themagnetic ®eld line. Quasi-static parallel electric ®elds,electrostatic shocks generating impulsive parallel electric®elds in the ULF frequency range (Mozer et al., 1980),and kinetic AlfveÁ n waves with a signi®cant parallelelectric component (Louarn et al., 1994) may welloperate along a large part of magnetic ®eld lines.

Conical type ion distributions have been investigated inthe frame of various mechanisms involving eitherresonant interaction with a variety of waves ± lowerhybrid (Retterer et al., 1994), electrostatic ion cyclotron(Ungstrup et al., 1979), electromagnetic ion cyclotron(Rauch et al., 1993; Le QueÂ au et al., 1993) ± or non-resonant interactions through stochastic acceleration bybroadband ¯uctuating ®elds below the O+ ion gyro-frequency (Lundin et al., 1990). It must be stressed thatthese mechanisms are not mutually exclusive; indeed,low-energy ion measurements provide key informationin order to disentangle the e�ects of competing wave-particle interactions (AndreÂ et al., 1994).

The results accumulated over the years have thereforeimposed low-energy plasma measurements as a majorgoal of magnetospheric satellite missions, and enlight-ened the need of dedicated experiments. On board theInterball Auroral Probe (IAP), this task is devoted tothe Hyperboloid experiment, which is an ion massspectrometer aimed at measuring the low-energy three-dimensional (3D) distributions of the four ``major'' ionsH+, He+ , O++, and O+. Its 0±80-eV energy range issuited to studies of processes involving thermal andslightly suprathermal ions in the Earth's auroral andpolar regions at altitudes of 700 to 20 000 km coveredby Interball AP. By some aspects, Hyperboloid derivesfrom the Dyction instrument which was previously¯own on Aureol-3 (Berthelier et al., 1982).

A major goal of Hyperboloid is to contribute to amore comprehensive classi®cation of observed ionout¯ow regimes, in order to associate them with thevarious theoretical and numerical models. With thisrespect, Hyperboloid provides detailed ion distributionfunction measurements, especially in the 10 000±20 000-km-altitude range which has remained poorly exploredbefore Interball AP launch. The interpretation ofHyperboloid data will bene®t from the simultaneousactivity of other plasma experiments, e.g., Thermal IonDynamics Experiment (TIDE) (Moore et al., 1995), andToroidal Imaging Mass-Angle Spectrograph (TIMAS)(Shelly et al., 1995), both on board the Polar satelliteorbiting at higher altitudes with an apogee of 9 Earth'sradii. Comparing the observed out¯ow properties atdi�erent altitudes, especially at the occasion of magneticconjugations, is expected to shed new light on polarwind theory. Observations by ground-based radars (e.g.,EISCAT, Super-DARN) should also complement thisnew, more global experimental picture of polar wind.

It has already been highlighted that magnetosphericconvection is one of the factors in¯uencing out¯owprocesses. In the regions which are explored by InterballAP, the knowledge of ion motion perpendicular to themagnetic ®eld is of particular interest since it governs theultimate fate of the extracted ions. This is especiallyimportant in the cusp region where the structure andintensity of convection closely depends on the orienta-tion and intensity of the interplanetary magnetic ®eldand on the solar wind velocity. Hyperboloid measure-ments of horizontal velocities and DC electric ®eldsfrom IESP (Perraut et al., 1998) will complement eachother in order to provide a quantitative picture of

N. Dubouloz et al.: Thermal ion measurements on board Interball Auroral Probe by the Hyperboloid experiment 1071

convection at Interball AP altitude. Owing to the 62.8°inclination of the orbit plane, the satellite practicallytravels along L-shells near apogee, thus enabling themonitoring of the convection at almost constant aurorallatitudes during extended periods. This capability,associated with the simultaneous observations of iono-spheric plasma ¯ows by the ESR and EISCAT radars,and by the Super-DARN network of coherent radars,will result in unprecedented data sets to study the solarwind-magnetosphere-ionosphere coupling: satellite data,which give access to small-scale (�1 to 10 km) structuressuch as arcs or vortices (which might be associated to¯ux transfer events at the magnetopause), will greatlybene®t from the knowledge of the larger scale (�50 km)and slower (�1 min) variations measured from theground.

At Interball AP altitudes, the distinction between thedi�erent types of accelerated ion distributions becomesmore di�cult due to the magnetic mirror force tending tofold-up conical distributions which ultimately look likebeams. However, the use of the full angular and energyresolution of Hyperboloid, together with the correlationwith wave instruments, will provide the required data tostudy microscopic wave-particle interactions in detail.First, the direct knowledge of thermal plasma parame-ters will enable a reliable determination of wave disper-sion relations and growth rates, without relying onindirect and more qualitative measurements such asdensity inferred from Langmuir probe data or fromauroral hiss cut-o�. The second expected progress willresult from the possibility to measure slightly suprather-mal ions, i.e. ions right at the beginning of theiracceleration process, before the features of the distribu-tion function become obscured by spatial and temporalinhomogeneities in the acceleration process. Of utmostimportance is the possibility o�ered by Hyperboloid toobtain simultaneous low-energy distributions of H+,He+, O++, and O+; comparing the level of energizationreached by the various ion species should provideprecious clues to identify acceleration mechanisms.

2 Instrument description

2.1 Speci®cations

In order to ful®ll the assigned scienti®c objectives, theHyperboloid ion mass spectrometer was designed tomeasure the full 3D distribution functions of low-energyions.

At altitudes between 700 and 20 000 km alongInterball AP orbit, and under average conditions, themajor ion species are H+, He+, and O+. These threeions are measured almost systematically in routinemodes of operation. Some speci®c modes can be usedin order to measure N+ and molecular ions (cf. Sect.2.4). Although representing for low solar and magneticactivity a small contribution to the total density, O++

may in some conditions reach densities as high as O+,and is therefore included in the routinely measured ionspecies; the simultaneous measurement of O+ and O++

is a precious tool to trace the e�ects of acceleration byDC electric ®elds or by resonant processes.

In the high-altitude portion of Interball AP orbits,H+ is usually the major ion, total ion densities rangingbetween about a few 10±2 and a few tens of cm±3. Theinvestigation of such dilute plasmas requires a very highsensitivity in order to detect di�erential ¯uxes down toabout a few 103 (cm2 s sr eV)±1. Moreover, the energybandwidth and angular acceptance of the instrumentmust be small enough to provide the resolutions neededto identify narrow features in the distribution functions.The very low temperatures (down to 2000 K or less)expected for classical polar wind out¯ows can be reliablymeasured only with high-resolution di�erential energymeasurements, as well as with narrow window aper-tures. The ability to distinguish between ion beams andconics which have been strongly folded by the magneticmirror force also implies a high angular resolution. Thenecessity to satisfy these con¯icting constraints onsensitivity and resolution drove the design of theinstrument, and particular attention was paid to thee�ciency of both ion optics and detectors. The condi-tions encountered at lower altitudes close to perigeedi�er drastically, with O+ being the major ion anddensities reaching several thousands of cm±3. Thisimposed a further constraint of large dynamical rangein the conception of the detectors in order to avoidsaturation or even destructive e�ects (cf. Sect. 2.5).

A major di�culty lies in the necessity of allowing thevery low energy ions to enter the instrument, even whenthe satellite is exposed to sunlight and immersed intenuous plasmas, and therefore tends to reach signi®cantpositive ¯oating potentials. In the ®rst part of themission, this was achieved by means of the satellitepotential control experiment RON (Torkar et al., 1998).After problems appeared in the operation of this device,the adverse e�ects of spacecraft potential were counter-acted by applying an adjustable potential bias to theentrance surface of the experiment. The energy range upto 80 eV is su�cient to detect H+ ®eld-aligned ¯ows athigh speeds up to about 80 km s±1 at high altitudes, aswell as O+ ions swept at speeds up to 20 km s±1 by themagnetospheric convection. Hyperboloid observationsin the upper part of its energy range are also comple-mentary to those performed at higher energies by theIon experiment (Sauvaud et al., 1998).

Due to the low ¯uxes measured at high altitudes,there is only a limited possibility of reducing the detectorintegration times in order to observe possible rapidchanges in distribution functions. In practice, a trade-o�must therefore be found between time resolution andspacing of measurements in velocity space. The variousmodes of operation re¯ect the di�erent compromiseswhich have been chosen with respect to the scienti®cobjectives (cf. Sect. 2. 6.)

2.2 Outline

The instrument is constituted of two multi-directionalanalyzers, each one encompassing several entrance

1072 N. Dubouloz et al.: Thermal ion measurements on board Interball Auroral Probe by the Hyperboloid experiment

windows (Fig. 1). The main analyzer (A16) measuresions with velocities in plane P1, and contains 16directions of sight de®ned by an identical number ofwindows through the main cylindrical part of the coverwhich surrounds the sensor unit. These windows arepositioned at 10° intervals and have �13° angularacceptance, so that the instrument spans a continuous160° angle in plane P1. The secondary analyzer (A10)with its view axis in plane P2 encompasses 10 windowspositioned 15° apart from each other and with �6°angular acceptance, so that the total angle spanned inplane P2 is 145° with blind sectors of �9° angular extentbetween adjacent windows.

Figure 2 shows that plane P1 contains the spacecraftspin axis, and therefore scans the full 3D space in onespin period �120 s. The secondary plane P2 makes anangle of 75° with the spin axis and is inclined towardsthe anti-solar direction. The two analyzers exhibitcomplementary performances. The A16 has by far thehighest geometrical factor and provides ion distributionfunctions in its half-plane with a time resolution ofroughly a few seconds, but is able to detect ions comingfrom a given direction only once per spin. Whencombined with the A16, the A10 o�ers partial accessto the 3D features of ion distributions with a muchbetter time resolution (typically a few seconds). Al-though its performances in terms of geometrical factorare much below those of the A16, it enables one tofollow the temporal variations of ion ¯ows overapproximately half a spin period when the ion bulkvelocity is adequately oriented.

Hyperboloid ion optics consists of three main parts,as illustrated in Fig. 3:

1. the electrostatic optics comprising the entrance sec-tions and the spherical analyzers used to select thedirection of sight and the energy interval of the

measurements, respectively, and the toroidal concen-trator whose function is to optimize the conditions ofentry inside the magnetic analyzer;

2. the magnetic analyzer which separates the various ionspecies;

3. the micro-channel plates (MCP) detectors.

2.3 Electrostatic optics

The most external grid of each entrance window isbiased at potential Vbias relative to the satellite, togetherwith the cylindrical part of the external cover of theexperiment. As explained in Sect. 2.1, this is intended tocompensate spacecraft charging e�ects. All the othervoltages of the electrostatic optics are referenced toVbias, except for the very high voltages near the exit ofthe concentrator for which this precaution is unimpor-tant. The second grid of the entrance section can beswitched from Vbias to +200 V independently for eachwindow, so that any combination of windows can beopened simultaneously. This enables the adaptation ofthe angular resolution and sensitivity of the instrumentto the various scienti®c objectives.

Each window of the A16 is followed by an individualspherical energy analyzer composed of two electrodeswhose voltages are controlled in order to select ionsaround a given energy E, and to de¯ect them by 90°toward the entrance of the toroidal concentrator. Aretarding or accelerating potential Vr can be applied tothe entrance and exit grids of the energy analyzer, thepotentials applied to the electrodes being adjustedaccordingly. As a result, the energy bandwidth of theanalyzer is multiplied from its nominal value by a factorK � 1)Vr/E, and the angular acceptance is changed

A16

A10

Fig. 1. Photograph of Hyperboloid without its external cover

75°

75°

75°

60°

15°

Fig. 2. Geometry of the two analyzers relative to the spacecraft


approximately in the same ratio. Following the energyanalyzer section is the main toroid concentrator madeup of an assembly of independently polarized electrodeswhose voltages have been optimized so that, irrespectiveof their entrance window, ions enter the magneticanalyzer as a nearly cylindrical beam with small angulardivergence.

Each window of the A10 is followed by two consec-utive spherical electrostatic analyzers with oppositecurvature radii, de¯ection angles of 60° and 30°,respectively, and again with adaptable retarding/accel-erating potential Vr. This particular geometry wasimposed by the overall mechanical constraints of theinstrument but has the advantage that the aberrations ofthe ®rst analyzer are partly counterbalanced by thesecond one. Ions exiting from these energy analyzersenter a secondary toroidal concentrator which directsthem to the main concentrator, where they travel in thesame way as ions from the A16.

2.4 Magnetic analyzer

Following the exit of the main toroid, ions enter themagnetic analyzer. With the nominal magnet voltage,the four routinely measured ions H+, He+, O++, andO+ are focused on their respective detector. Sweepingthis voltage enables to detect minor ions such as N+ ormolecular ions.

The geometry of the magnet analyzer is derived fromthe Mattau-Herzog geometry but has been optimizedthrough ray-tracing calculations taking into account theactual geometry of the ion beam at its entrance. Inparticular, an improved focusing is obtained by using anentrance plane non normal to the mid-axis Z1 (Fig. 3)and by choosing an angle for the exit face of the magnetwhich is not the same for H+ as for all other ions. Thise�ort in design enabled a signi®cant reduction of ion

losses by improving the acceptance of the magneticanalyzer for trajectories inclined with respect to its mid-plane of symmetry.

2.5 Detectors

The four detectors use curved micro channel plates(MCP) fabricated by Photonix-Mullard Company withC-shaped channels 25 l in diameter and 2 mm long. Theaverage gain is 1.5 á 106 at 1600 V and the resolution onthe order of 80%. Each MCP is connected to its ownhigh-voltage (HV) power supply.

The detectors may be operated either in a pulsecounting mode for ion ¯uxes lower than 1.5 á 105 c/s, orin an analog mode with charge integration for larger¯uxes. In order to maintain the output current above thenoise level, but also below 2% of the polarizationcurrent (thus keeping the MCP operation in the linearrange), the mode of measurement and, when in theanalog mode the HV applied to the MCPs, areautomatically controlled on board by the microproces-sor. The gain of the MCPs is thus permanently adaptedto the registered ion ¯uxes. The resulting total dynam-ical range is larger than 1011, more than convenient forthe maximum of 1010 c/s foreseen at perigee for O+.

2.6 Modes of operation

Several modes types have been de®ned to operate theinstrument. Beside technical modes intended to checkthe health of the electronics and dump the memory, anddiagnostic modes used to monitor any evolution in thephysical response of the instrument optics and detectors,scienti®c modes consist in repetitive measurementswhere both energy and view direction are swept insequence during several spacecraft spin periods.

180° toroidalconcentrator

60° toroidalconcentrator

Z2

Z1

Vbias

VbiasVbias

Vf

Vf

Vf

VAIM

V0

V0

V0 V0

V0

V0

V0

V72

V11

V11

V21

V31

V41

V41

V51

V61

V71

V72

V81 V8

1

V91

V101

V121

V131 V11

1

V62

V62

V42

V52V3

2

V22

V12

V92

V82

V82

V0

Ve

VeVe

Ve

Ve

Vi

Vi

Vi

Vi

Vi

V0

Doublesphericalanalyzer

Sphericalanalyzer

Magneticanalyzer

Detectors

Ions

Ions

H+

He+

O+

O++

MCP

Fig. 3. Schematic cross section of the ionoptics in the plane of the magnetic analyzer.The electrostatic optics on the left of the®gure is representative of the ®rst window ofthe main analyzer (A16), the optics of theother windows of the A16 are obtained bysuccessive rotations of 10° around axis Z1.Similarly, the optics ®gured on the right arerelevant for windows 1 and 10 of thesecondary analyzer (A10), optics for theother windows being deduced by 15° rota-tions around axis Z2


Each scienti®c mode is de®ned by:

1. the set of measured ion species (Nion � between 2and 4 species taken in the H+, He+, O++, and O+

group);2. the energy sweep, consisting of a geometricalprogression characterized by its initial energy E0,progression factor q and number of energies Nenergy.The energy resolution DE/E may be chosen eitherconstant (®xed K), or energy dependent (®xed Vr), asde®ned in Sect. 2.3;

3. the window sweep consisting of one or two progres-sions including either A16 or A10 windows. Eachprogression is characterized by its initial window W0,number of window groups Ngroup, and number ofwindows per group Nwindow. The window shiftbetween two consecutive groups Nstep can be chosendi�erent from Nwindow so that consecutive groups canoverlap;

4. the double-sweep structure, with internal loop eitheron energies (DE mode) or on window group (DAmodes).

Independently to the scienti®c mode, the telemetrysystems collect data into so-called ``pages'', each ofwhich contains 72 individual measurements. A completesweep over energies and window groups is thereforeperformed in Npage � (Nion ´ Nenergy ´ Ngroup)/72pages, i.e., Npage/TM seconds, where TM is theinstrument telemetry rate expressed in pages per second.This time is the temporal resolution for obtainingdistribution functions in the half-plane of one or twoof the analyzers. It is also linked to the angularresolution obtained in the spin plane. The choice ofthe mode attributes therefore implies a trade-o� betweenenergy, angular, and temporal resolution.

The commanding of instrument operations is organ-ized through a hierarchy of structures. The so-called``enchainments'' include one or several modes, with aprescribed duration in units of the spacecraft spinperiod. Enchainments are further assembled into ``se-quences'' chosen by operating telecommands which alsode®ne the potential bias Vbias and the telemetry rate TM.For safety reasons, the information de®ning all modes,enchainments, and sequences were loaded in LU tablesin two redundant EEPROM memories. Table 1 sum-marizes the characteristics of the most commonly usedsequences, which favor either energy resolution (largeNenergy and small q), angular resolution in the A16 plane(small Nwindow), or temporal resolution, equivalent toangular resolution in the spin plane (small Npage/TM).

3 Data reduction

For a given ion species, measurements performed attimes tk during a set of successive swaeps over energysteps Ei and window groupsWGj form a 3D array of ioncounts Nijk The measured distribution function Fijk anddi�erential ¯uxes Jijk can be related to counts by:

Nijk � s� Gij � Fijk � E2mi � 2=M2; �1� T

able1.Characteristicsofthemostfrequentlyusedscienti®cmodes.Each

TM

valuecorrespondsto

asequence,andeach

lineto

anenchainem

ent.Emaxisthelaststep

ofenergysweeps.

Windowsweepsare

de®ned

byoneortwoprogressionsNgroup(N

window)Analyzer,e.g.,8(2)16means8groupsof2windowsin

theA16analyzer

TM

(pages

s)1)

Ions

Nenergy

E0(eV)

qEmax(eV)

Kwindowsweep

modetype

Npage

4H+He+

O++O+

36

1.4

1.12

71.7

1.0

4(3)16+

2(3)10

DE

12

H+He+

O++O+

18

1.7

1.25

74.9

1.0

4(4)16+

3(4)10

DE

72

H+He+

O++O+

18

1.7

1.25

74.9

1.0

16(1)16

DE

16

H+He+

O+

24

1.7

1.18

75.6

1.0

8(2)16

DA

81

H+He+

O++O+

18

1.7

1.25

74.9

1.0

8(2)16+

4(2)10

DE

12

H+He+

O++O+

18

1.7

1.25

74.9

1.0

4(4)16+

3(4)10

DE

71

H+

O+

18

1.7

1.25

74.9

1.0

16(1)16

DE

8H+

O+

36

1.4

1.12

71.7

1.0

6(4)16

DA

61or0.5

H+He+

O+

24

1.4

1.19

75.0

2.0

6(4)16

DA

60.25

H+He+

O++O+

18

1.5

1.26

75.3

2.0

4(4)16

DE

4


Jijk � Fijk � Emi � 2=M2; �2�where s is the integration time (which varies between 4and 200 ms, depending on the number of measured ionspecies and on the telemetry rate), Gij the geometricalfactor, and M the ion mass. Emi � Q ´ Ei is themeasured energy of ions of charge Q for energy step Ei.

The geometrical factors may be written as:

Gij � TRij � K2 � DE=E � Aj � dXij; �3�where K is the resolution factor (cf. Sect. 2.3), DE/E thenominal energy resolution of the spherical analyzer (0.23and 0.09 for A16 and A10, respectively), and Aj thesurface of individual windows (1.4 cm2 for A16 and0.84 cm2 forA10windows). dWij represents the sumof thenominal acceptance solid angles of windows in groupGWj (13° ´ 12.5° and 6° ´ 3°, corresponding to3.9 á 10±2 and 4.3 á 10±3 sr for each window of the A16and A10, respectively), and TRij the transparencies whichdepend on ion species, energy step, and window group.

Each value Fijk is a discrete measurement of the iondistribution function associated to the correspondingenergy and direction at the entrance of the instrument.By taking into account the e�ect of the spacecraft sheathand applying Liouville's theorem, one can then deduce aset of discrete values of the distribution function in theplasma, but such complex calculations are not possibleat this early stage of data reduction. In view of thegenerally large Debye length compared to the size of theinstrument, we have considered as a ®rst approximationthe sheath to be spherically symmetrical over the halfspace of viewing directions, which implies that iondirections of arrival are una�ected. Moreover, ionenergy in the plasma is given by:

Epi � Emi � Q� Vexp � Q� �Ei � Vsat � Vbias�; �4�where Ei is the energy step and Vexp and Vsat theexperiment and satellite potentials, respectively.

In principle, Vsat can be determined directly from DCelectric ®eld measurements from the IESP experiment(Perraut et al., 1998). However, such measurementsmight be di�cult to interpret at high altitude when theDebye length becomes larger than the antenna length.Therefore, we have tried to develop a fast method todetermine Vsat relying on Hyperboloid data only. A setof Vsat values is used to calculate the undisturbed ionenergies from Eq. (4), from which the moments (density,velocity vector, parallel and perpendicular temperatures)of the undisturbed distribution function are deduced.Each Vsat and associated moments are then introducedas input parameters of a program simulating theworking of the experiment. This program is presentlybased on a bi-Maxwellian model of distribution func-tions, which is generally a reasonable approximation.The best ®t between simulated and experimental ¯uxesprovides the estimation of Vsat and the correspondingmoments. As shown in Sect. 4, this simple methodproduces results which are satisfactory for preliminarydata analysis; it will be improved in the future by takinginto account more accurate sheath e�ects.

4 Preliminary results

After more than one year of activity, the InterballAuroral spacecraft has covered all MLT sectors atsubauroral, auroral, and polar-cap latitudes in thenorthern hemisphere, and at altitudes between about10 000 and 20 000 km. The data collected by Hype-rboloid during this period reveal a wide variety ofphenomena which we illustrate by three typical exam-ples.

4.1 Ion out¯ow and wave-induced heating in auroralregions

Figure 4 displays a set of wave and particle measure-ments performed between 21:00 and 22:00 UT on 2December 1996 over the northern near-midnight auroralzone and polar cap, and at altitudes between about16 000 and 19 000 km.

First, a region of ion heating is crossed from about21:00 and 21:24 UT, characterized by H+ ¯uxes almostuniform over nearly the entire energy range. These ¯uxesalso exhibit little dependence on spacecraft spin (2 min),which implies almost isotropic H+ distributions typicalof strong heating. Due to the high altitude, andprobably also to the winter and solar minimum condi-tions, O+ ¯uxes are much smaller, and even close tonoise level at low energies. However, they extend overthe full instrument energy range, which shows that O+

as well as H+ ions are heated. Due to their higher mass,O+ ions tend to have smaller thermal velocity, whichexplains why their ¯uxes are more collimated in azimuthangles and therefore exhibit noticeable temporal varia-tions over a spin period.

This period of ion heating is associated with stronglow-frequency turbulence, as evidenced by the secondtop panel of Fig. 4 representing the low-frequency wavemagnetic component measured in the direction parallelto the spacecraft spin axis by the IESP experiment[courtesy of S. Perraut and M. Mogilevsky, see alsoPerraut et al. (1998)]. Signi®cant power is observedaround the oxygen gyrofrequency fc+ � 0.5 to 0.9 Hz,up to the hydrogen gyrofrequency fcH+ � 9 to 14 Hz,and probably at higher frequencies, where unfortunatelyno data are available. Other electric and magneticcomponents (not shown) are characterized by similarbehavior. During the same time-interval, electron ¯uxesmeasured by the ion experiment between about 10 eVand 20 keV [courtesy of J.-A. Sauvaud, see alsoSauvaud et al. (1998)] show keV electrons, probably ofplasma sheet origin, accompanied by impulsive ¯uxes atlower energies down to about 30 eV (the intensi®cationat 20 eV is due to photoelectrons returning toward thepositively charged satellite). It is tempting to interpretthe waves as AlfveÁ n-like electromagnetic turbulencedestabilized by the observed electron distributions.Indeed, these waves are reminiscent of those observedat lower altitudes by the Viking satellite (Gustafssonet al., 1990). In the latter reference, it is shown that


electron distributions with several populations includinga ®eld-aligned beam are able to destabilize a shearAlfveÁ n mode. Although the magnetic ®eld line con®g-uration during the period analyzed here does not enableto estimate electron ¯uxes outside of the 50 to 130° pitchangle range, the electrons detected below 1 keV appearas a plausible energy source for the waves (Temerin andLysak, 1984).

Figure 5 displays time-window spectrograms in dif-ferent energy ranges during one spin period between21:02 and 21:04 UT characterized by an intense electro-magnetic turbulence. Above 30eV, H+ and O+ maximal¯uxes fall almost systematically in the 90 to 130° pitchangle range, suggesting that ions are heated perpendic-ular to the magnetic ®eld below the satellite, betweenabout 13 000 and 16 000 km. H+ ¯uxes at lowerenergies gradually shift toward the upward ®eld-aligneddirection (the corresponding O+ ¯uxes are close to noiselevel and are not plotted). Such simultaneously observed

H+ and O+ conics are di�erent from those observed atsimilar altitudes but higher energies by DE-1 (Klumparet al., 1984), which were shown to result from cumula-tive ion cyclotron resonance (ICR) heating over almostthe entire ®eld line (Crew et al., 1990). In the eventanalyzed here, ICR with the low-frequency electromag-netic turbulence detected by IESP is a plausible scenariofor explaining the observed H+ and O+ ion heating,even if the contribution of higher frequency, e.g., lowerhybrid, waves cannot be ruled out (AndreÂ et al., 1994).This suggests that ICR may on some occasions occurover a limited altitude range, thus leading to conics ofmoderate energies. Another theoretical question raisedby the present observations is the possibility of obtain-ing simultaneous heating of both H+ and O+ by ICR,while classical theories describe heating of minor speciesonly (Le QueÂ au et al., 1993).

On Fig. 4, the period of ion heating ceases at about21:24 UT, simultaneously with the disappearance of

10

12

1

0.1

0

0

0

80

80

0.01

Ene

rgy

(keV

)

Ele

ctro

n en

ergy

flux

log

(cou

nts)

B //

pow

er

arbi

trar

y un

its

O +

flux

H +

flux

log

(m

s sr

eV

)2

–1[

]

Freq

uenc

y(H

z)E

nerg

y(e

V)

Ene

rgy

(eV

)

UTMLTILATALT

21:0023.774.0

16150

21:200.3

74.017673

21:401.6

73.218664

22:002.3

71.819147

INTERBALL AP/HYPERBOLOID & IESP & ION 2 December 1996

3.7

1.0

9.0

9.05.0

5.0

Fig. 4. Wave and particle measurements obtained on 2 December1996 between 21:00 and 22:00 UT, in the midnight MLT sector ofauroral zone and polar cap. From top to bottom, 10 eV±20 keV

electron counts from Ion, magnetic component parallel to the spinaxis of waves below 12 Hz from IESP, and 0±80 eV O+ and H+

¯uxes averaged over Hyperboloid half-planar ®eld of view


low-frequency waves, and with a signi®cant drop inelectron ¯uxes and average energies. In the time-intervalbetween 21:24 UT and 21:35 UT, probably correspond-ing to the polar cap, narrow O+ ion distributions areobserved, while H+ ¯uxes remain below the noise level.It may be noted that the higher background ¯uxesobserved at lower energies for both H+ and O+ ionsduring this time-interval correspond to small countsclose to the noise level, which induce arti®ciallyenhanced ¯uxes due to the lower instrument transpar-encies at lower energies (cf. Eqs. 1±3). This feature isalso visible on Figs. 7, 8, and 9 below. Close to 21:28 UTon Fig. 4, we observe a narrow O+ distribution atenergies �10±20 eV. For such energies spacecraft po-tential can in ®rst approximation be neglected, andmoment calculation provides n0+ � 1 cm±3, upwardbulk velocity v// � ±6 km s±1, and horizontal velocityv^ � 10 km s±1. Temperature is di�cult to estimate dueto the very collimated ¯uxes. Polar out¯ow theoriestaking into account the e�ect of centrifugal acceleration

due to ion convection across curved magnetic ®eld lines(Cladis, 1986) indeed foresee that O+ can remain themajor ion at such high altitudes. Our v^ � 10 km s±1

can mainly be attributed to magnetospheric convectionand corresponds to a convection electric ®eld of about70 mV m±1 when mapped at 200 km altitude. Thecorresponding O+ density and ¯ow velocity found byHo et al. (1997) at 3 RE altitude are in good qualitativeagreement with our experimental values (see theirFig. 2). Other O+ distributions registered around 5 eVmight correspond to slower convection but need to becorrected from spacecraft potential before being furtheranalyzed.

At about 21:35 UT the spacecraft enters a new regionof local ion heating, whose characteristics are initiallyvery similar to those described in the foregoing. Figure 4reveals however a gradual change in both H+ and O+

¯uxes which, starting at about 21:42 UT, show clearerspin modulations. This transition appears more clearlyon Fig. 6, which shows the evolution of the H+ ion

1

1

1

1

1

16

16

16

16

16

H +

0-5

eV

H +

5-1

2.5

eVH

+ 1

2.5-

30 e

VH

+ 3

0-80

eV

O +

30-

80 e

V

log

[(m

s sr

eV

)2

–1]

Win

dow

Win

dow

Win

dow

Win

dow

Win

dow

UT 21h02:20 21h03:20

INTERBALL AP / HYPERBOLOID 2 December 1996

8.8

9.2

9.0

8.7

9.0

7.1

6.8

5.9

5.5

5.9

Fig. 5. 3D distribution function of H+ andO+ ions on 2 December 1996 between 21:02and 21:04 UT. Each spectrogram displaysthe angular variation of ¯uxes in a givenenergy range. The black crosses and squaresindicate ions moving up and down along themagnetic ®eld lines, respectively, while thefull line corresponds to 90° pitch angles


distribution between 21:34 and 22:00 UT. Initially,intense ¯uxes are observed all along the spin periodsclose to 90° pitch angles. At later times, ¯uxes exhibit aclear spin period modulation and are maximal at pitchangles close to the upward ®eld-aligned direction.Notice that this evolution is accompanied by a decreasein the low-frequency wave spectral densities and in the20 to 300-eV electron ¯uxes (Fig. 4), in agreement withthe scenario previously proposed for explaining ionheating. The distribution functions measured fromabout 21:44 UT are characteristic of the ion out¯owprocesses which are frequently observed by Hyperboloidat auroral latitudes. Using the method presented in Sect.3, we estimated the spacecraft potential and deduced themoments of the H+ distribution between 21:52 and21:54 UT. We found Vsat � +7 V, and correspondingvalues for density n � 0.8 cm±3, upward bulk velocityv// � ±42 km s±1, horizontal velocity v^ � 13 km s)1,parallel temperature T// � 1.5 eV, and perpendiculartemperature T^ � 2 eV. At the same moment,n � 0.7 cm±3 and v// � ±18 km s±1 for O+. Consideringthe balance between currents generated by satellitephoto-electrons and by ambient thermal electrons of

density ne � nH+ + n0+ � 1.5 cm±3 and temperatureTe � 2 eV, leads to an estimate Vsat � +6 V, which isa good con®rmation of the validity of our method forcalculating moments corrected from satellite charginge�ects. The same qualitative results are obtained allthroughout the 21:44 to 22:00 UT period, i.e., the H+

and O+ out¯owing ions exhibit di�erent bulk velocitiesand di�erent mean energies. Such a mass-dependentacceleration process might result from a bimodal accel-eration involving (1) magnetic moment pumping due to¯uctuating electric ®elds (e.g., associated to wavessimilar to those observed between 21:00 and 21:24 UTbut acting at lower altitudes), and (2) ®eld-alignedacceleration by quasi-static electric ®elds (Lundin andHulqvist, 1989).

4.2 Thermal ion signatures of large-scale boundariesin the nightside magnetosphere

Thermal and suprathermal plasma emanating fromthe ionosphere is a�ected by magnetospheric plasma

1

1

1

1

16

16

16

16H

+ 0

-5 e

VH

+ 5

-12.

5 eV

H +

12.

5-30

eV

H +

30-

80 e

V

log

[(m

s sr

eV

) ]

2–1

Win

dow

Win

dow

Win

dow

Win

dow

UT 21h39:00 21h45:00 21h51:00 21h57:00

INTERBALL AP / HYPERBOLOID 2 December 1996

8.7

9.3

9.7

10.0

5.3

5.7

6.1

6.5Fig. 6. Evolution of the H+ distribu-tion function on 2 December 1996between 21:34 and 22:00 UT, in thesame format as Fig. 5


processes at high altitudes such as convection, ®eld-aligned currents, heating, and acceleration. It is there-fore interesting to look at the changes of the thermalplasma parameters at boundaries between the large-scale magnetospheric plasma domains.

Methods to locate these boundaries in the nightsector from measurements of precipitating particles onlow-altitude satellites were developed during the last twodecades and now seem well established (Feldstein andGalperin, 1985; Galperin and Feldstein, 1991, 1996;Newell et al., 1996). According to the review of Feldsteinand Galperin (1985), the soft electron precipitationboundary (SEB) practically coincides with the equato-rial large-scale convection boundary. Other hot plasmaregimes ± above the di�use auroral zone (i.e., within thetrapping boundary of high-energy particles), above theoval of discrete auroral forms (i.e., within the so-calledauroral plasma cavity), and further poleward ± can alsobe delineated during stationary conditions. The charac-teristics of thermal ion populations at mid-altitudes inthe plasmasphere and above auroral oval and polar capwere extensively analyzed from data on DE-1 (Horwitzet al., 1986; Giles et al., 1994), and Akebono (Abe et al.,1993; Watanabe et al., 1992). These experiments havedemonstrated the predominance of polar wind ¯ows,and revealed regions of counterstreaming ¯ows as wellas many other features. Horwitz et al. (1986) analyzedthermal plasma distributions with respect to the hotplasma boundaries in the inner magnetosphere fromelectron data on DE-1 and 2. They concluded from astatistical analysis that in the nightside the SEB prac-tically coincides with the plasmapause. However, theydid not analyze the thermal plasma signatures associatedto the auroral oval of discrete forms and polar-capboundaries. In this paper, from a data set correspondingto a quieting period after a storm, we present a

preliminary investigation of the thermal plasma charac-teristics at, and between, the boundaries of the mainmagnetospheric plasma domains.

We chose the nightside parts of a series of similarorbits on 12±13 January 1997, when the magnetospherewas recovering from a period of strong magneticactivity induced on 10±11 January by a very dense``solar plasma cloud'' which engulfed the circumterres-trial space. This was a period of outer plasmaspherere®lling with some weak activity (2± £ Kp £ 3±). Con-sider the nightside part of the inbound orbit 565 in thereverse chronological order, from the nightside plas-masphere poleward across the di�use auroral zone,oval and polar cap (Fig. 7). From the simultaneousdata from Ion (courtesy of J.-A. Sauvaud), InterballAP crossed the equatorial border of the di�use auroralzone, de®ned as the region of increasing plasma sheetelectron energy with latitude, at 06:30 UT (MLT � 0.6,invariant latitude � 64.3°). The polar edge of thedi�use zone, i.e., the equatorial border of the oval,was probably crossed in the interval 06:12±06:14 UTcorresponding to the maximum observed electronenergies. Then the satellite entered the plasma sheettype plasma, with a gradually decreasing averageenergy and some structuring of electron ¯uxes till itspolar edge at 05:35 UT. There, an enhancement ofenergetic electron ¯uxes and average energies, accom-panied by a velocity dispersed ion structure of type 2(VDIS-2) (Zelenyi et al., 1990), was observed at about05:40 UT, followed by a decrease in the averageelectron energy. This last feature may be identi®ed as anarrow polar di�use zone at the polar edge of the oval.

Figure 7 shows that signi®cant changes in the ther-mal plasma characteristics occur at or close to these hotplasma boundaries. A very similar latitudinal structureof the particle data was observed on the other orbits of

INTERBALL AP / HYPERBOLOID & ION 12 January 1997

10

10

1

1

0.1

0

80

Ene

rgy

(keV

)

Ene

rgy

flux

log

(cou

nts)

Ele

ctro

nsH

+H

+

Ene

rgy

(eV

)

(

keV

)

UTMLTILATALT

05:0021.677.9

19186

05:3022.974.9

18642

06:0023.870.6

16954

06:300.6

64.314026

07:001.5

54.39703

3.7

3.5

1.0

1.0

Arbitraryunits

Fig. 7. From top to bottom, energy-time spectrograms for high-energy electrons and H+ ions from Ion, and for 0±80 eV H+ ionsfrom Hyperboloid between 05:00 and 07:00 UT on 12 January 1997.

Note that two di�erent color scales have been used in the bottompanel in order to enlighten both low ¯uxes before 06:00 UT and large¯uxes after this time


this period. We now summarize the thermal andsuprathermal plasma characteristics within the variousmagnetospheric domains.

Outer plasmasphere and plasmapause. The region ofspace within the stationary, or quasi-stationary, con-vection boundary, is evidently the outer plasmasphere,and the boundary is the plasmapause. Indeed, a sharpdrop in the thermal plasma density from more than100 cm±3 to values lower by one to two orders ofmagnitude is evident at, or close to, the SEB. Densitiesrecorded on successive orbits before orbit 565 weremuch lower (�0 and 8 cm±3), suggesting that re®llingwas ongoing in the outer plasmasphere region. We deferthe analysis of the re®lling rates to a further publication.

Di�use auroral zone. The plasma density outside theplasmapause continues to decrease with latitude down tovery low values. In this region the positive potential of thesatellite increases simultaneously, so that a more re®neddata analysis is needed to deduce a reliable latitudinaldensity pro®le. However, since the average ion energy isfairly constant, it is anticipated that the observed relativevariations of ion ¯uxes are correct. For orbits 565 and569, a narrow region of counterstreaming H+ and He+

ions exists near the plasmapause at energies above�15 eV.

Plasma cavity above the auroral oval. Very low, butmeasurable thermal plasma ¯uxes are seen above theauroral plasma forms: this is the well-known auroralplasma cavity (Calvert. 1981) where the hot plasmasheet component generally dominates the total plasmadensity. Sometimes ®laments of denser, primarily out-¯owing thermal plasma are crossed, which might play arole in wave propagation and wave-particle interactions.

Polar edge of the oval. A new feature is discovered inthe thermal plasma, coincident with the low-energyelectron precipitation observed at the polar edge of theauroral oval of discrete arcs, and with the extrapolationto in®nite velocity of the associated VDIS. It is seen onall orbits considered here, and appears as a density peakof thermal and suprathermal H+ and He+ ions with awide spectrum from the lowest energies till the upperend of Hyperboloid energy range. The H+ densities inthe peak reach 0.15 cm±3 on orbits 565 and 569, and0.05 cm±3 on orbit 571. The energy spectrum of thesesuprathermal ions can extend to a few hundreds of eVs(middle panel of Fig. 7). The analysis of ion velocitiesshows predominant upward directed ¯uxes, but also asizable fraction of downgoing ions. This suggests thecombined action of ionospheric ion heating at lowaltitudes due to electron precipitation, and of ionheating or scattering all along the ®eld lines probablydue to wave-particle interactions.

Polar cap. The thermal plasma distributions in thepolar-cap region adjacent to the auroral oval are similarto those in the auroral cavity with extremely lowdensities practically at or below the instrument thresh-old �10±3 cm±3. A distinct change in plasma character-istics is identi®ed at 77° invariant latitude with theappearance of H+ and He+ out¯owing ions withdensities �10±2 cm±3 and velocities �20 km s±1 typicalof heated polar wind conditions. Abrupt changes in the

properties of these out¯ows can be seen betweenconsecutive spins. Such changes are frequent in Hype-rboloid observations and might be the consequence ofsmall-scale structures in the ionosphere.

4.3 Low-energy ion dynamics in the cusp/cleft region

Five months after its launch, Interball AP began toexplore the cusp/cleft region characterized by precipita-tion of plasma from the magnetosheath (Heikkila andWinningham, 1971; Rei� et al., 1977; Burch et al., 1982;Menietti and Burch, 1988; Newell and Meng, 1988;Kremser and Lundin, 1990). Another common feature isthe observation of low-energy (below �100 eV) iono-spheric plasma out¯ows (Lockwood et al., 1985; Hor-witz and Lockwood, 1985; Pollock et al., 1990; Gileset al., 1994). Two representative crossings of the cusp/cleft region are examined here.

Figure 8 shows particle data acquired by the Hype-rboloid and Ion (courtesy of J.-A. Sauvaud) experimentson 26 February 1997, as the spacecraft crossed the cusp,moving poleward and toward the post-noon dusksector. Interball AP crossed the equatorward boundaryof the cusp at 00:18 UT (12.7 MLT and 75° invariantlatitude), as indicated by the magnetosheath electronsencountered below �200 eV until about 00:40 UT(77.7° invariant latitude). Associated to the 120-ssatellite spin period, V-shaped ion signatures character-istic of the direct injection of magnetosheath particles inthe cusp appear in the H+ data around 2 keV at 00:20UT (�12.8 MLT and 75.2° invariant latitude), slightlypoleward of the low-latitude boundary of the electronprecipitation region. This V-shaped pattern is explainedby transit-time e�ects, assuming an injection source ofnarrow latitude extent located in the exterior cusp. Thethickness of each V structure depends on the sourceextent. The gradual energy decrease observed as thesatellite moves toward higher latitude is due to antisun-ward convection (Burch et al., 1982). The V-shaped H+

features begin to be seen by Hyperboloid from about00:36 UT and show a remarkable continuity with thoseobserved at higher energies. Another striking feature isthe inversion of the H+ ion V structure observed around00:42 UT, apex of the V's appearing at higher energiesthan the wings after this time: this is explained by thefact that progressively, the downgoing ions do not reachthe satellite any more and only upward moving ionswhich have mirrored below the satellite are observed(Burch et al., 1982). Injected H+ ions are seen byHyperboloid down to about 5 eV. Even taking intoaccount possible spaceraft charging e�ects, this is wellbelow typical magnetosheath ion energies (�100 eV).This suggests signi®cant deceleration of ions upon theirentry into the magnetosphere at high latitudes in themantle region (Newell et al., 1991).

Interball AP measurements con®rm that the cuspnear 12:00 MLT is also a signi®cant source of out¯ow-ing ionospheric ions, as observed in Fig. 8. In the mostequatorward part of the cusp, where the strongest ¯uxesof precipitating electrons are recorded from about 00:18


UT to 00:32 UT Hyperboloid data show up¯owing H+

and He+ ion distributions identi®ed at each satellitespin. The ¯uxes are widely spread both in time (i.e., inazimuth angle) and in energy, indicating strong ionheating. H+ and O+ up¯owing ¯uxes are in factregistered at energies up to a few keV by the Ionexperiment. During this period, strong ¯uxes are mea-sured at pitch angles down to 90° (pitch-angle data notshown), suggesting that ions are accelerated perpendic-ular to the magnetic ®eld over a wide range of altitudesbelow the satellite. Parallel electrostatic potential dropsmight also contribute to the acceleration process, but arenot required (Crew et al., 1990; Lundin and Eliasson,

1991). Indeed, perpendicular ion acceleration has beenobserved by the Sounding of the Cleft Ion FountainEnergization Region (SCIFER) rocket experiment downto very low altitudes in the cusp region (Kintner et al.,1996).

Another region can be identi®ed in Fig. 8 from about00:34 UT. The H+, He+, and at later times O+

distributions observed by Hyperboloid become narrow-er, indicating smaller temperatures. All three speciesexhibit a gradual decrease in energy while the satellitemoves at higher altitude toward the polar cap. The dataare consistent with the mass spectrometer e�ect associ-ated with the ``cleft ion fountain'' (Lockwood et al.,

0.1

0.01

10

10

1

1

1

0.1

0.1

0

0

0

80

80

80

Ene

rgy

(keV

)E

nerg

y(k

eV)

Ele

ctro

n en

ergy

flux

H +

ene

rgy

flux

H +

flux

He

+ fl

uxO

+ e

nerg

yflu

xO

+ fl

ux

Ene

rgy

(eV

)E

nerg

y(e

V)

Ene

rgy

(keV

)E

nerg

y(e

V)

UTMLTILATALT

00:0011.870.5

12916

00:2012.875.2

15273

00:4014.077.7

17037

01:0015.378.6

18256

INTERBALL AP/ HYPERBOLOID & ION 26 February 1997

3.7

3.5

11.0

9.0

1.0

1.0

4.0

4.0

3.5

8.0

1.0

4.0

Fig. 8. Energy-time spectrograms of particles during a cusp crossingon 26 February 1997. From top to bottom, 10 eV±1 keV electrons,100 eV±14 keV H+ ions, 0±80 eV H+ ions, 0±80 eV He+ ions,100 eV±14 keV O+ ions, and 0±80 eV O+ ions. High-energy

measurements in log(counts) are from the Ion experiment. Low-energy ¯uxes from Hyperboloid in log [(m2 s sr eV)±1] are averagedover Hyperboloid half-planar ®eld of view


1

1

1

1

80

80

80

80

40

40

40

40

20

20

20

20

10

10

10

10

5

5

5

5

He

+ fl

uxH

+ fl

uxO

++

flux

O +

flux

Ene

rgy

(eV

)

UTMLTILATALT

22:0010.865.5

10194

22:2011.773.2

13139

22:4012.877.8

15444

23:0014.279.8

17161

INTERBALL AP/ HYPERBOLOID 27 February 1997

11.0

9.0

7.0

9.0

5.0

4.0

3.0

4.0

.

log

[(m

s sr

eV

) ]

2–1

Ene

rgy

(eV

)E

nerg

y (e

V)

Ene

rgy

(eV

)

Fig. 9. Energy-time spectrograms for low-energy ions measured by Hyperboloid, during a cusp/cleft crossing on 27 February 1997. From top tobottom, O+, O++, He+, and H+ ¯uxes averaged over Hyperboloid half-planar ®eld of view

5

5

2

15

40

10

20

10

H +

flux

0-5

eV

H +

flux

win

d.5-

8

log

[(m

s sr

eV

)2

–1]

Win

dow

Ene

rgy

(eV

)

UT 22:34 22:36 22:38

INTERBALL AP / HYPERBOLOID 27 February 1997

10.4

10.1

6.6

6.1

Fig. 10. Energy-time spectrogram for windows 5 to 8 of the A16 analyzer (top panel) and window-time spectrogram for energies below 5 eV(bottom panel) obtained for H+ ions on 27 February 1997 between 22:34 and 22:38:30 UT


1985). Depending on their velocity, ions are swept by theantisunward convection up to di�erent invariant lati-tudes during their upward motion to the satellite.Therefore, both energy and mass dispersion occurs forthese ions, as evidenced on Fig. 8. The heavier O+ ionsare convected over large distance away from the sourceregion, which suggests that a great part of the O+ ionsobserved in the polar cap originates in the polar cleft.

Figure 9 shows another example, from 27 February1997. On that day, the low-latitude edge of the cusp waslocated as far south as 69° invariant latitude (IMFconditions were By � ±5 nT, Bz � ±13 nT). InterballAP enters the cusp at 22:08 UT. Again, this boundaryappears as a region of strong ion acceleration. Between22:07 UT and 22:09 UT, ¯uxes at energies above 12-eVpeak at pitch angles near 90° (pitch-angle data notshown), indicating a near local perpendicular accelera-tion. Besides the magnetosheath H+ ion injection, themain feature is again the occurrence of strong out¯owfor all ion species. Moment calculations provide similarupward velocities for all four ion species, e.g., around22:31 UT, v// � ±20 km s±1, for an horizontal velocity�8 km s±1 mainly due to convection. The data thusprovide a quantitative con®rmation of the cleft ionfountain mechanism.

Another interesting and surprising phenomenon isthe presence, in the interval 22:20±22:50 UT, of a verynarrow and low energy (<5 eV) H+ population detect-ed between the up¯owing ion events. This signature ismore clearly seen in Fig. 10, which represents a zoomfor the interval 22:34±22:38:30 UT. Note the enhanced¯uxes occurring in the second window group between 2and 5 eV at times 22:34:20, 22:36:30, and 22:38:25 UT.These ¯uxes were indeed detected nearly along (down-ward) the local magnetic ®eld direction (bottom panel).These ions seem clearly separated from the V-shapedinjected ions at higher energies (top panel), and aretherefore probably not magnetosheath ions. If they areof ionospheric origin, they have energies too high to begravitationally bounded. Their downward motion mightbe the trace of small amplitude, downward pointing DCelectric ®elds above the satellite altitude,

5 Conclusion

Since its ®rst switching-on and commissioning phase atthe end of September 1996, the Hyperboloid experimenthas been working nominally. In more than one year ofoperation, full ion distribution functions in the 0±80-eVrange, and with mass separation, have been obtained ataltitudes between �10 000 and 20 000 km in the outerplasmasphere, auroral zone, and polar-cap regions, withfull MLT coverage. More recently, the experiment hasbeen operated during the short (<14 min) perigeepasses below the radiation belts, between �700 and2 000 km altitudes, again at auroral and polar latitudes.A great variety of phenomena has been observed.Hyperboloid detailed distribution function measure-ments o�er new perspectives for understanding process-

es, such as polar wind or the cleft ion fountain, whichwere already identi®ed during previous mission (DE,Akebono, Viking). New observations, e.g., H+ and O+

conics at energies � a few tens of eVs observedsimultaneously at altitudes �3 Earth's radii, or thermalion signatures of the polar edge of the discrete auroraloval, deserve further experimental investigation and areexpected to give rise to new theoretical developments.

Data acquired in the following lifespan of theexperiment will improve the statistical basis of futurestudies. Whenever possible, use will be made of mag-netic conjugations with other satellites and ground-based radars. Finally, particular attention will be paid tomodeling the interaction of the satellite and of theexperiment with their plasma environment, in order toimprove the physical signi®cance of the data.

Acknowledgements. The Interball Project was accomplished in theframe of contract N025-7535/94 with the Russian Space Agency(RKA). The work of Y.G., D.C., T.M., and L.Z. was supported bygrants from Russian Foundation for Basic Research (RFBR) N94±02±04299 and 97±02±16333, and by grant INTAS N 94±1695.We thank Russian Space Agency, Lavochkin Space Associationand Babakin Space Center for managing the di�erent steps of theproject with success despite very adverse conditions. The experi-ment was ®nancially supported by CNES under the auspices ofgrants covering the period 1985 to 1998. We are indebted to J.-A.Sauvaud and S. Perraut for providing data from the Ion and IESPexperiments. Thanks are also due D. Delcourt for fruitfuldiscussions.

Topical Editor K.-H. Glassmeier thanks K. Torkar and R.Lundin for their help in evaluating this paper.

References

Abe, T., B. A. Whalen, A. W. Yau, R. E. Horita, S. Watanabe, andE. Sagawa, EXOS D (Akebono) suprathermal mass spectrom-eter observations of the polar wind, J. Geophys. Res., 98, 11191,1993.

AndreÂ , M., P. Norqvist, A. Vaivads, L. Eliasson, O. Norberg, A. I.Eriksson, and B. Holback, Transverse ion energization and waveemissions observed by the Freja satellite, Geophys. Res. Lett.,21, 1915, 1994.

Banks, P. M., and T. E. Holzer,High-latitude plasma transport: thepolar wind, J. Geophys. Res., 74, 6317, 1969.

Berthelier, J.-J., J. Covinhes, M. Godefroy, G. Gogly, C. GueÂ rin, D.Roux, P. TheÂ venet, and V. A. Gladyshev, The thermal ion massspectrometer on board Aureol-3: the DYCTION experiment,Ann. Geophys., 38, 591, 1982.

Burch, J. L., P. H. Rei�, R. A. Heelis, J. D. Winningham, W. B.Hanson, C. Gurgiolo, J. D. Menietti, R. A. Ho�man, and J. N.Bar®eld, Plasma injection and transport in the mid-altitudepolar cusp, Geophys. Res. Lett., 9, 921, 1982.

Calvert, W., Auroral plasma cavity, Geophys. Res. Lett., 8, 720,1981.

Cladis, J.B., Parallel acceleration and transport of ions from polarionosphere to plasma sheet, Geophys. Res. Lett., 13, 893, 1986.

Crew, G. B., T. Chang, J. M. Retterer, W. K. Peterson, D. A.Gurnett, and R. L. Hu�, Ion cyclotron resonance heated conics:theory and observations, J. Geophys. Res., 95, 3959, 1990.

Feldstein, Y. I, and Y. I. Galperin, The auroral luminosity structurein the high-latitude upper atmosphere: its dynamics andrelationship to the large-scale structure of the Earth's magne-tosphere, Rev. Geophys. Space Phys., 23, 217, 1985.

Galperin, Y. I., and Y. I. Feldstein, Auroral luminosity and itsrelationship to magnetospheric plasma domains, in Auroral


physics, Eds. C.-I. Meng, M. J. Rycroft, and L. A. Frank,Cambridge UP, Cambridge, p. 207, 1991.

Galperin, Y. I., and Y. I. Feldstein, Mapping of the precipitationregions to the plasma sheet, J. Geomagn. Geoelectr., 48, 857,1996.

Ganguli, S. B., The polar wind, Rev. Geophys., 34, 311, 1996.Giles, B. L., C. R. Chappell, T. E. Moore, R. H. Comfort, and J. H.

Waite, Jr., Statistical survey of pitch angle distributions in core(0±50 eV) ions from Dynamics Explorer 1: out¯ow in theauroral zone, polar cap, and cusp, J. Geophys. Res., 99, 17483,1994.

Green, J. L., and J. H. Waite, Jr., On the origin of polar ionstreams, Geophys, Res. Lett., 12, 149, 1985.

Gurgiolo, C., and J. L. Burch,DE-1 observations of the polar wind-A heated and an unheated component, Geophys. Res. Lett., 9,945, 1982.

Gustafsson, G., M. AndreÂ , L. Matson, and H. Koskinen, On wavesbelow the local proton gyrofrequency in auroral accelerationregions, J. Geophys. Res., 95, 5889, 1990.

Heikkila, W. J., and J. D. Winningham, Penetration of magneto-sheath plasma to low altitudes through the dayside magneto-spheric cusps, J. Geophys. Res., 76, 883, 1971.

Ho, C. W., J. L. Horwitz, and G. R. Wilson, Dynamics of the H+

and O+ polar wind in the transition region as in¯uenced byionospheric convection and electron heating, J. Geophys. Res.,102, 395, 1997.

Ho�man, J. H., Studies of the composition of the ionosphere witha magnetic de¯ection mass spectrometer, Int. J. Mass Spectrom.Ion Phys., 4, 315, 1970.

Ho�man, J. H., W. H. Dodson, C. R. Lippincott, and H. D.Hammack, Initial ion composition from the ISIS 2 satellite, J.Geophys. Res., 79, 4246, 1974.

Horwitz, J. L., and M. Lockwood, The cleft ion fountain: a two-dimensional kinetic model, J. Geophys. Res., 90, 9749, 1985.

Horwitz, J. L., S. Menteer, J. Turnely, J. L. Burch, J. D.Winningham, C. R. Chappell, J. D. Craven, L. A. Frank, andD. W. Slater, Plasma boundaries in the inner magnetosphere, J.Geophys. Res., 91, 8861, 1986.

Kintner, P. M., J. Bonnell, R. Arnoldy, K. Lynch, C. Pollock, T.Moore, J. Holtet, C. Deehr, H. Stenbaek-Nielsen, R. Smith, J.Olson, and J. Moen, The SCIFER experiment, Geophys. Res.Lett., 23, 1865, 1996.

Klumpar, D. M., W. K. Peterson, and E. G. Shelley, Direct evidencefor two-stage (bimodal) acceleration of ionospheric ions, J.Geophys. Res., 89, 10779, 1984.

Kremser, G., and R. Lundin, Average spatial distributions ofenergetic particles in the mid altitude cusp/cleft region observedby Viking, J. Geophys. Res., 95, 5753, 1990.

Le QueÂ au, D., A. Roux, J.-L. Rauch, F. Lefeuvre, and J.-M.Bosqued, Heating of protons by resonant absorption in amulticomponent plasma, 2. Theoretical model, J. Geophys. Res.,98, 13363, 1993.

Lockwood, M., M. O. Chandler, J. L. Horwitz, J. H. Waite, Jr., T.E. Moore, and C. R. Chappell, The cleft ion fountain, J.Geophys. Res. 90, 9736, 1985.

Louarn, P., J. E. Wahlund, T. Chust, H. de Feraudy, A. Roux, B.Holback, P. O. Dovner, A. I. Eriksson, and G. Holmgren,Observations of kinetic AlfveÁ n waves by the Freja spacecraft,Geophys. Res. Lett., 21, 1847, 1994.

Lundin, R., and L. Eliasson, Auroral energization processes, Ann.Geophysicae, 9, 202, 1991.

Lundin, R., G. Gustafsson, A. I. Eriksson, and G. Marklund, On theimportance of high-altitude low-frequency electric ¯uctuationsfor the escape of ionospheric ions, J. Geophys. Res., 95, 5905,1990.

Menietti, J. D., and J. L. Burch, Spatial extent of the plasmainjection region in the cusp-magnetosheath interface, J. Geo-phys. Res., 93, 105, 1988.

Moore, T. E., C. R. Chappell, M. O. Chandler, S. A. Fields, C. J.Pollock, D. L. Reasoner, D. T. Young, J. L. Burch, N. Eaker, J.H. Waite, Jr., D. J. McComas, J. E. Nordholdt, M. F. Thomsen,

J.-J. Berthelier, and R. Robson, The thermal ion dynamicsexperiment and plasma source instrument, Space Sci. Rev., 71,409, 1995.

Mozer, F. S., On the lowest altitude S3-3 observations ofelectrostatic shocks and parallel electric ®elds, Geophys. Res.Lett., 7, 1097, 1980.

Nagai, T., J. H. Waite, Jr., J. L. Green, C. R. Chappell, R. C.Olsen, and R. H. Comfort, First measurements of supersonicpolar wind in the polar magnetosphere, Geophys. Res. Lett., 11,669, 1984.

Newell, P. T., and C. I. Meng, The cusp and the cleft/boundarylayer: low-altitude identi®cation and statistical local timevariation, J. Geophys. Res., 93, 14549, 1988.

Newell, P. T., W. J. Burke, C.-I. Meng, E. R. Sanchez, and M. E.Greenspan, Identi®cation and observations of the plasmamantle at low altitude, J. Geophys. Res., 96, 35, 1991.

Newell, P. T., Y. I. Feldstein, Y. I. Galperin, and C.-I. Meng, Themorphology of nightside precipitation, J. Geophys. Res., 101,10737, 1996.

Perraut, S., ULF wave measurements on board the InterballAuroral probe, Ann. Geophysicae, this issue, 1998.

Pollock, C. J., M. O. Chandler, T. E. Moore, J. H. Waite, Jr., C. R.Chappell, and D. A. Gurnett, A survey of upwelling ion eventcharacteristics, J. Geophys. Res., 95, 18969, 1990.

Rauch, J.-L., F. Lefeuvre, D. Le QueÂ au, A. Roux, J.-M. Bosqued,and J.-J. Berthelier, Heating of protons by resonant absorptionin a multicomponent plasma, 2. Experimental evidence, J.Geophys. Res., 98, 13347, 1993.

Rei�, P. H., T. W. Hill, and J. L. Burch, Solar wind plasmainjection at the dayside magnetospheric cusp, J. Geophys. Res.,82, 479, 1977.

Retterer, J. M., T. Chang, and J. R. Jasperse, Transverselyaccelerated ions in the topside ionosphere, J. Geophys. Res., 99,13189, 1994.

Sauvaud, J.-A., The Ion experiment on board the Interball Auroralsatellite: initial results on velocity dispersed structures in thecleft and inside the auroral oval, Ann. Geophysicae, this issue,1998.

Schunk, R. W., and J. J. Sojka, A three-dimensional time-dependent model of the polar wind, J. Geophys. Res., 94,8973, 1989.

Shelly, E. G., R. G. Johnson, and R. D. Sharp, Satellite observationsof energetic heavy ions during a geomagnetic storm, J. Geophys.Res., 77, 6104, 1972.

Shelly, E. G., A. G. Ghielmetti, H. Balsiger, R. K. Black, J. A.Bowles, R. P. Bowman, O. Bratschi, J. L. Burch, C. W. Carlson,A. J. Coker, J. F. Drake, J. Fischer, J. Geiss, A. Johnstone, D.L. Kloza, O. W. Lennartsson, A. L. Magoncelli, G. Paschmann,W. K. Peterson, H. Rosenbauer, T. C. Sanders, M. Steinacher,D. M. Walton, B. A. Whalen, and D. T. Young, The toroidalimaging mass-angle spectrograph (TIMAS) for the polarmission, Space Sci. Rev., 71, 497, 1995.

St-Maurice, J.-P., and R. W. Schunk, Ion velocity distribution inthe high-latitude ionosphere, Rev. Geophys., 17, 99, 1979.

Taylor, H. A., Jr, and W. J. Walsh, The light ion trough, themain trough, and the plasmapause, J. Geophys. Res., 77, 6716,1972.

Temerin, M., and R. L. Lysak, Electromagnetic ion cyclotron mode(ELF) waves generated by auroral electron precipitation, J.Geophys. Res., 89, 2849, 1984.

Torkar, K., An experiment to study and control the Langmuirsheath around Interball-2, Ann. Geophys., this issue, 1998.

Ungstrup, E., D. M. Klumpar, and W. J. Heikkila, Heating of ionsto suprathermal energies in the topside ionosphere by electro-static ion cyclotron waves, J. Geophys. Res., 84, 4289, 1979.

Watanabe, S., B. A. Whalen, and A. W. Yau, Thermal ionobservations of depletion and re®lling in the plasmaspherictrough, J. Geophys. Res., 97, 1081, 1992.

Zelenyi, L. M., R. A. Kovrazhkin, and J.-M. Bosqued, Velocity-dispersed ion beams in the nightside auroral zone: AUREOL-3observations, J. Geophys. Res., 95, 12119, 1990.


thermal ion measurements on board interball … ion measurements on board interball auroral probe by...

Documents