GEOPHYSICAL
RESEARCHLETTERS,VOL. 24,NO.10,PAGES1163-1166,
MAY 15, 1997
Venustail ray observationnear Earth
H.Grtinwaldt
•,M.Neugebauer
2,M.Hilchenbach
•,P.Bochsler
3,D.Hovestadt
4,
A.Btirgi
4,F.M.Ipavich
•,K.-U.Reiche
6,W.I. Axford
•,H.Balsiger
3,A.B.Galvin
J.Geiss
3,F.Gliem
6,G.Gloeckler
5,K.C.Hsieh
7 R.Kallenbach
3,B Klecker
4,S.Livi
M.A.Lees,G.G.Managadze
9,E.Marsch
•,E.MObius
s,M.Scholer
4,M.I. Verigin
B. Wilken•, P. Wurz3
aboutthe composition,
acceleration,and flux of
Abstract.In June, 1996, Venus passedthrougha very close uncertainty
inferiorconjunction
with the Sun.At that time the CTOF ionsdownstreamof Venus(e.g., seereview by Kar [1996]).
A uniquechance for in-situ observations
as far as 45
detectorof the CELIAS massspectrometer
experimenton the
SOHO spacecraftnear Earth'sL1 Lagrangianpoint was
million kilometers downstream of Venus occurred on June 10,
1996,whenVenushadan extremelycloseinferiorconjunction
downstreamof Venus.Close to the time predictedby simple with the Sun, passingonly -0.5ø southof the Sun center
geometric
arguments
forpassage
of SOHOthrough
theVenus relative to the ecliptic plane. At that time, the SOHO
measuring
heavyionsin the solarwind ~4.5x107
km
wason stationnearEarth's
L1Lagrangian
point0.01
wake,CTOFmadethreeencounters
withunusual
fluxesof O+ spacecraft
of Earth.In this paper,we reportthe detectionof
andC+ ions.Theirenergydistributions
resembled
thoseof tail AU sunward
raysoriginating
in theVenusionosphere
or ionopause
region Venustail rays on that occasion.
ratherthanof ionsproduced
in thecoronaof neutralatomsthat
surrounds
theplanet.
TheC+ abundance
was--10%of O+.The The CTOF Instrument
observed
O+ speedwasverycloseto the simultaneous
solar
windspeed
andtheO+ temperature
wasa cool5600K/amu.The
As partof the CELIAS packageof ion massspectrometers
fluxdensities
forthethreeeventswere(2.4-4.4)x103
cm%4.
on SOHO [Hovestadtet al., 1995], the CTOF detector was
designedto measureenergy(E), mass(M), and charge(q) of
Introduction
solar wind ions. The instrument combines a hemispherical
BecauseVenus has no appreciablemagnetic field, the electrostaticanalyzer as an E/q filter with a time-of-flight
solar wind interactsdirectly with the planetaryatmosphere (TOF) measurement
yieldingspeed(v) and a silicondetector
total energyfromwhichonecan derivethe original
and ionosphere
muchas with a cometarycoma.In fact, measuring
severalpeoplehave pointedout possiblesimilaritiesof the E/q, M/q andM. An acceleratingvoltage (Uacc= 30 kV) is
tails of Venusand comets[e.g.,Russellet al., 1982;Braceet used to raise the ions' energy above the amplifier noise
al., 1987;McComaset al., 1987].The solar wind interaction threshold(25 keV) to overcomeenergylosseswithin the TOF
canscavenge
material
fromVenusandcarryit awaywiththe carbonfoil and the detectorfront layer. The field of view is
solar wind. Three different mechanisms are believed to be shaped
by two quadrupole
lensesinto an almostcircularcone
important:
directionization
and pickupof O+ fromVenus' of-22 ø half-anglewhichensuresfull integrationof nominal
coronaof hot O atoms;indirectpickupthroughsputteringof supersonic
solar wind ion distributions.
During each 302-s
atomsfrom the exobasedueto bombardment
by the previously instrument
cycle,E/q sweepsdownfrom34.7keV/q by 4.09%
picked-upO+ ions; and removalof materialfrom the each step of 2.52 sec. Onboardreduction includes the
ionosphere
and ionopause
regionto formtail rays.While evaluationof the solar wind speedfor every cycle. Details of
visiblecometary
tailrayscanbe preserved
fortensof millions the instrumentand calibrationwill be given in a futurepaper
of kilometers,there is little evidencefor the existenceof (H. Granwaldt,E. M•bius, et al., in preparation)and some
planetary
tailraysbeyonda maximum
of 12 Venusradii,Rv. remarksare made here only on featuresof specificrelevance
A searchfor He+ ions in the distant wake of Venus yielded
to the data concerned.
(a) In contrastto the energyof multiply chargedsolar
negative
results[Russell
andNeugebauer,
1981].Despite
longtermcoverage
by severalprobes,
amongthemtheparticularly wind ions that are usually above the 25 keV detection
successful Pioneer Venus Orbiter (PVO),
there is large
thresholdafter accelerationthroughUacc,singly chargedions
pickedup at Venuswouldtypicallyhave energiesbelowthat
•Max-Planck-Institut
ftir Aeronomie,D-37189 Katlenburg-Lindau, level. Based on this knowledge, mass identification is
Germany
2JetPropulsion
Laboratory,
Pasadena,
CA 91109,USA
uniquelypossiblefrom the E/q and TOF informationonly.
With the instrumentM/q resolution(given by the energywidth
3physikalishes
Institut
derUniversit•it,
CH-3012
Bern,Switzerlandof the instrumenttransmissionand by scatteringin the carbon
4Max-Planck-Institut
ftir extraterrestrische
Physik,D-85740
foil whichaddup to FWHM -- 1.5 amu/earoundCNO+ for the
Garching,Germany
applicableinputenergyrange)the identification
of dominant
5University
of Maryland,
CollegePark,MD 20742,USA
CNO+ speciesis only limited by their countstatisticsandby
6Institut
ftirDV-Anlagen,
Technische
Universit•it,
D-38023
contamination
(b)
Braunschweig,Germany
7University
of Arizona,Tucson,
AZ 85721,USA
ZUniversity
of NewHampshire,
Durham,
NH 03824,USA
9Institute
for SpacePhysics,
Moscow,Russia
Copyright
1997bytheAmerican
Geophysical
Union.
Papernumber97GL01159.
0094-8534/97/97GL-01159505.00
effects.
While TOF micro-channel-platenoise and UV
backgroundare negligible, the dominant contaminationis
contributedby abundantions that producenoncorrelated
low
energystop signalsin the TOF. These "accidentals"
occur
whenthe electrostatic
analyzeris set at the F_./qrangeof their
originator,
andthey are thereforemostsignificantat the F_./q
corresponding
to He+ andthe majorconstituents
of the solar
wind.
1163
1164
GRONWALDT ET AL.: VENUS TAIL RAYS
Observations
All registered
O*counts
Identification of tail ray O*
O*cmznts
cleaned
ofjoossible
bckgrd.
An inferior conjunctionof Venus occurredat 1600 UT,
June 10 (day 162), 1996. A first predictionof the encounterof
any wake or tail of Venus with SOHO must allow for the
aberration angle caused by the planet's 35 km/s orbital
-- 165 23:45
/, •
velocity,tan'• 35/<v•>, where<v,,.> is theaverage
solarwind
speedat Venus.A more accurate predictionmakes use of the
actual speed, v•,,, measured at SOHO. In the frame of
observationthe trace of any radially moving solar wind is
shapedinto a spiral, and the positionof Venus proceedswith
the synodicperiodof 584 days (see enhancedsketch at the
inset of Figure 1 at a time between conjunction and
encounter).From v•, the expected angular distance A• of
Venus from the alignment with SOHO can be derived. This
angle is plotted in Figure 1. The shaded band accountsfor
+_1.5
ø fluctuations
in the direction
•o:oo
//,
oo:o
.... • .... • ..........
0
10
20
30
40
50
60
70
0+ countsper 5-rain cycle
of the solar wind flow vector
that might be causedby wavesor streaminteractions.Figure 1
showsthat one might expect to observeVenusions aroundday
Figure 2. Distribution
of O+ countsper instrument
cycle: the
165.
shadedarea all background
is removed.
The eventsof days
From PVO observations,
O+ is expectedto be by far the
dominant Venus ion in the solar wind. Two distinct types of
energydistributionsmay be encountered: we expect coronal
pickupionsto be on cycloidal trajectorieswith speedsranging
from 0 to up to twice the solar wind speed,dependingon the
direction of the interplanetarymagnetic field. Ion tail rays,
however, would be closer to the speedand flow direction of
the ambientsolar wind. The processused to searchfor Venus
ions is first to examine the O+ channelfor outstandingflux
events, then to determine how many of those counts were
solid line refers to all counts in the O+ channel while in the
165 and 166 appearclearly aboverandomchance.
the accidentals
by takingonlythoseeventswithE/q• > 8 (far
outside the He++thermal core), the distributionis that shown
in the lower (filled) histogramin Figure 2. The three events
attributedto Venus ions are basicallyunchangedby this
correction,
andtheirexceptional
natureis even moreapparent
than before.
Figure 3 presentsmore evidence as to the nature of the
actuallycausedby O+ ions, and then to studyother features eventsin the 0 + channel.
The toppanelsshowE/q•,•spectra
suchas the energyand time profiles.
duringtwo of the exceptionaleventswhilethe bottompanels
The uppertracein Figure2 is a histogramof thenumberof show spectra near the same time, bat outside the events.
events registered in the O+ channel in each 5-minute Running3-channelaveragesweretakento smooththe data.In
instrumentcycle, usingall data obtainedduringdays 82-230,
1996.The most probablenumberof events/cycleis zero, and
the distributiondrops sharply to just over 20 counts/cycle.
However, there were three highly exceptionalcycles in which
36, 42, and 63 counts were registered, and each of those
cyclesoccurredclose to the time of the possibleinterception
of the Venus tail.
Accidentals are primarily expected near the relative
energy/charge
E/qr,t--2 or 4, corresponding
to solarwind H+*
or He+,whereE/qr•tis normalizedto the energy/charge
of the
bulk flow of sotarwind protons(M/q -= 1). After correctionfor
each of the panelsthe thin line showsthe distributionfrom the
He+ channelwhich represents
well the expectedpick-up
distribution;
it is approximately
constant
aboveE/q•,t= 4, the
value corresponding
to the plasma bulk motion, and has a
cutoffat E/q•,•= 16, corresponding
to twice the plasmabulk
speed.There is no significantvariation in these distributions
exceptstatisticalfluctuations.
The shadedpartsadd spectra
obtained
fromtheO*channels.
Onlythetoptwopanels
showa
well establishedpopulationcenteredaroundE/q,t = 16,
consistent
with a speedcloseto the solarwind speed,and a
rather cool thermal distribution. There is no increase in He +
(the only ion possiblyinterfering
at that E/q) apparentat the
times of the O* events, thus ruling out an increase in
4o
accidental
effects.
Figure 4 presentsthe data in a different format that
emphasizes the sharpnessof the variations in time and
velocity.It is a 3-dimensional
plot with counts/cycle
on the
vertical axis, time running to the right (with each box
denoting
a single5-minutecycle), andthe thirdaxisbeingW
= v/v,,,,,calculatedon the basisof the M/q of O*. Ditchesin
the ground plane indicate data gaps. From the sec•ond
momentsof the velocity distributionsshownin Figure 4,
kinetic temperatures
near 5600 K/amu are obtained.
Eachof theencounters
with the Venustail wasextremely
brief; the second and third events lasted no more than the
instrumentcycle time of 5 minutes.In fact, for those two
events, the majority of the counts (40 for event 3) were
-4 ø
160
I , I , ! , ! , •
162
164
, I , I , I ,
166
168
170
Day of Year, 1996 (ObseroaH,
on)
collectedwithin only one 30-secondsweepof the analyzer
voltage.Event1 wasbroaderin time with somevelocity-time
binswithlowerbutnonetheless
significant
countssurrounding
Figure 1. Venus phaseangle, A•, with respectto the solar
wind in contactwith SOHO.The times of tail ray eventsare
the bin with the highestnumberof counts.
marked
an eventof enhancedO* flux of a magnitudefar beyond
at the bottom.
We conclude from this evidence that SOHO encountered
GRONWALDT ET AL.' VENUS TAIL RAYS
25
12
--
i
1165
i
i
oso•o/cELzas/cro•
o.•o
o.47
o.34
• 2o
c•[
•?
O*
o. fo
o. o7
C+
o
1œ
1
I
I
ltl
I
I
Gt,m". b•tu,. mn.nts 1,•
6
i
io
I
4
10
•0
E/q.
40
relc•ff•e
4
10
œ0
statistical chance at a time when the spacecraftwas within
the projectedtail of Venus.From the spectrumof those ions
we concludethat they form a cool beam traveling close to the
solar wind speed.
of C + and other ions
Figure 5 presentsa closer look at the data of the three
events as two-dimensional
,
,
I
12
,
ß
,
I
r
14
a,,
a
I
16
i
i
I
,l
'
18
to ?•rotorLs
Figure 3. He+ pick-upion distributioncomparedto cold O*:
the upperpanels are taken from tail ray events 1 and 2; the
bottompanelsare averagedover the interval betweenevent 1
and2 and over onehour pastevent3, respectively.
Identification
,
10
m,
distributions
of total
count rates
versusthe measured M/q (x-axis) and E/q,.et(y-axis). The
distributionhasbeen smoothedby 3-bin runningaverageswith
bin sizes of 0.5 (x) and 4.7% (y) and normalized per cycle.
The rangeof contoursis from 90% of the maximumdown to
5% with logarithmic spacingas shownby the code bar; the
lower limit eliminates the background determined in a
comparativerun usingdata outsidetheseevents.
The contoursdisplay two separatepopulations.
The major
peak is locatedaroundM/q = 16 andE/qret= 16 andbelongsto
O+ as discussedabove.The smaller peak is centered near M/q
=12, suggesting
C+ as its origin.The associatedE/qm of about
13.5is clearly distinctfrom the oxygenpeak. Lack of a signal
above the energy detector threshold is supportingevidence
that the ions measuredwith M/q = 12 carry a charge of q = 1
and identifies them as C •.
On the assumptionof a similar link of these ions to the
solarwind speedas we foundfor O* the centralE/q•t wouldbe
fo
Figure 5. Contoursof smoothedcount rate distribution as a
functionof E/q and M/q aroundM/q values correspondingto
C* and O*. E/q has been normalizedto the energy/chargeof
solarwind protons.Traces of N* may be presentnear M/q =
14. The speedsof C* and O* relative to the solar wind are
derived from the moments parallel to the E/q axis. The total
numberof C* andO* countsare givenin the inset.
expected near 12. The actual value of 13.5 correspondsto a
speedmismatch of about 6%. The statistical uncertaintyin v
is the thermal speeddivided by the square root of the total
numberof counts, and may account for 1-2% on the basis of
the better establishedtemperaturefound for O*. Another 2%
may be caused by the width of the E/q steppingchannels.
Finally, there may be time aliasing of the spectrum as E/q
sweepsdown,but suchan effect is rather unlikely since three
independent events contribute to the result. Thus, however
marginally, the bulk speedfor this populationmay be higher
than for O* whichraisesthe possibilityof differenthistoriesfor
thetwo ions species,O* andC*.
The small table in Figure 5 lists the total counts for the
two species accumulated over the three events and over the
wholeM/q andE/qm rangesof each distribution.
The value for
O* is higherthan obtainedby addingup countsfrom Figure 4
which was collected with very restrictive bandwidth. The
averagedC70* ratio is closeto 0.10. We note that in Figure 5
there is also a hint of N* at M/q = 14 whichwasnot taken into
accountin calculatingtheratio C70*.
What aboutmolecularions suchas O2*(M/q = 32)or CO2*
(M/q = 44) [Luhmannet al., 1995]? For the observedsolar
wind speedof 320 km/s, such ions would have E/q = 17 or
23.5 keV/q, respectively,whichare within the 35 keV/q CTOF
range.However, since the post-accelerationenergiesof those
ions are just slightly larger than their losses in the foil, the
detection efficiency and ion identification are poor.For the
three tail-ray events, there is evidence for a number of counts
above backgroundassociated with both candidate ions, but a
more careful, quantitativeanalysisis postponedto a follow-up
study.
o
Discussion
We
Figure 4. Tail ray observations
on days 165 and 166: count
rates versus time and normalized velocity W = v/v•,,,.The
sequenceof three distinct events showsup clearly. The
ditches mark data gaps.
believe
the evidence
for a cool beam of ions with
mass/charge= 16 amu/e traveling at close to the solar wind
speed is compelling. The timing indicates that Venus was
almost certainly the source of the ions. The energy
distributionsare consistentwith the ions being tail rays rather
than either direct or secondarycoronal pickup ions. Although
coronal pickup ions might also be present,their fluxes within
1166
GRONWALDT ET AL.: VENUS TAIL RAYS
the CTOF field of view are so low that more sophisticated
data analysiswill be requiredto find them.
broader spatial and angular distributions than the tail rays
described here.
Thetailraysareprobably
spatially
verynarrow.
If therays
Acknowledgments.MN thanks J. G. Luhmann for helpful discussions.
maintained a fixed orientation with respect to Venus, a
structureone Venus radius wide would pass by SOHO in 5.5 CELIAS is supportedin part by DARA, Germany, under contracts50
minutes,or in slightly more than one instrumentcycle.In such OC 89056 and 50 OC 96087, by NASA, USA, under contract NASa scenario, the observedtemporal durationof <45 s would 31166, by the Swiss National Science Foundation and the PRODEX
correspond
to a spatialwidth<0.14 Rv -- 820 km. The tail rays programof ESA. The researchat the Jet PropulsionLaboratorywas
would not, however, be expected to remain in a fixed performed under a contract between the California Institute of
geometryrelativeto Venus becausevariationsin the direction TechnologyandNASA.
of the solar wind must cause the tail rays to flap around so
that they could have a range of transversevelocities relative References
to SOHO. Thus unambiguousinterpretation of the data in
termsof spatial scales is not possiblewith observations
made Brace, L. W., W. T. Kasprzak,H. A. Taylor, R. F. Theis, C. T. Russell,
A. Barnes, J. D. Mihalov, and D. M. Hunten, The ionotail of Venus:
at only a single locationin space.
its configuration
and evidence for ion escape,J. Geophys.Res., 92,
From the measuredhigh speed and thermally narrow ion
15, 1987.
distributionthe assumptionthat instrumentalintegrationof one
snapshothas been complete over energy and directional Dubinin, E., R. Lundin, W. Riedler, K. Schwingenschuh,J. G.
Luhmann,C. T. Russell, andL. H. Brace, Comparisonof observed
extent seemsjustified. The shapesshownfor O* in Figures4
plasma and magnetic field structuresin the wakes of Mars and
and 5 do not hint at any cut off by space-time effects,
Venus, J. Geophys.Res.,96, 11189, 1991.
although none of the distributionsextend much beyond the
snapshottime of about 30 s within a cycle (-- 45 s including Hovestadt,D. et al.., CELIAS - Charge, element and isotopesystem for
C*). The transformationof observed O* count rates into fluxes
yields 4.4, 2.4, 3.1x103ions/cm2s
for the three events,
respectively.The calculationof the total flux within each tail
ray is speculative,however,becausewe do notknow the large
scale structure SOHO was tracing through. If a tail ray is
assumed to be a circular flux tube with a diameter
of 820 km
anda flux of 3x103 cm'2s
'•, its total flux wouldbe 1.6xlff9 O*
ions s4.
SOHO, Solar Phys., 162, 441, 1995
Intriligator,D. S., Observationsof massaddition to the shocksolarwind
of the Venusianionosheath,Geophys.Res.Lett., 9, 727, 1982.
Intriligator,D. S., Resultsof the first statisticalstudyof PioneerVenus
Orbiterplasmaobservations
in the distant Venus tail: evidence for
a hemisphericasymmetry in the pickup of ionosphericions,
Geophys.
Res.Lett., 16, 167, 1989.
Kar, J., Recentadvancesin planetary ionospheres,
SpaceSci. Rev., 77,
193, 1996.
The conceptof comet-liketail rays extendingdownstream
of Venuswasfirst proposedby Braceet at [1987] on the basis Kasprzak, W.T., J.M. Grebowsky, H.B. Niemann, and L.H.Brace,
Superthermal
>36-eV ionsobservedin the near-tailregion of Venus
of data acquired within 2500 km of Venus by PVO. The
by the Pioneer Venus Orbiter Neutral Mass Spectrometer, 1
observations
were extendedto 12 Rv by Mihalov and Barnes
Geophys.Res., 96, 11175, 1991
[1982] and lntriligator [1982; 1989]. All those studiesnoted
the presenceof O*ions and the filamentaryor sporadicnature Luhmann, J. G., A model of the ionospherictail rays of Venus, 1
Geophys.Res., 98, 17615, 1993.
of the down-tail fluxes.At 1.3 Rv from the center of Venus,
tail-ray widthstypically variedfrom 600 to 3000 km [Dubinin Luhmann,J. G., W. T. Kasprzak,and J. M. Grebowsky,On removing
molecularions from Venus, J. Geophys.
Res.,100, 14515, 1995.
et al., 1991], with typically 1 to 3 rays observedper orbit of
PVO around Venus. While at this distance the ions are just McComas, D. J., J. T. Gosling, C. T. Russell and J. A. Slavin,
Magnetotails at unmagnetized bodies: Comparisonof Comet
above escape speed, further downstreamthe speedsapproach
Giacobini-Zinnerand Venus, J. Geophys.
Res.,92, 101l l, 1987.
magnetosheathsupersonicvalues. For similar phenomenaat
Mars, Rosenbaueret al. [1989] reporteda systematic speed McComas,D. J., H. E. Spence, C. T. Russell and M. A. Saunders,The
averagemagneticfield drapingand consistentplasmapropertiesof
gradient across the tail up to highly supersonicvalues of
heavyions near the center.Many of the featuresof the Venus
tail rays, including their very narrow structure, can be
explained on the basis of a model developed by Luhmann
[1993], and McComas et at [1986] have used the observed
magnetic field in the tail togetherwith the MHD equationsto
model the acceleration of the tail plasma up to solar-wind
speeds.
Although C* ions have not been previouslyreportedin
Venustail rays, a C*/O* ratio of about0.1 hasbeen reportedin
various domains of the ionosphere[Kasprzaket al., 1991], in
goodagreementwith the CTOF results.
The purposeof this shortreportis to point out that Venus
tail rays have been identified in the solar wind some 45
million km downstreamof Venus. In future papers we will
report on more details of the observations,including the
abundanceof molecular ions, the responseof the ambient
solar wind to the presenceof the tail rays, and comparisonto
differentmodelsand theories.Furtheranalysisis also required
to searchfor the Venuscoronalpickupions which would have
the Venus magnetotail,J. Geophys.Res.,91, 7939, 1986.
Mihalov, J.P., and A. Barnes, The distant interplanetary wake of
Venus:plasma observationsfrom PioneerVenus, J. Geophys.Res.,
87, 9045, 1982.
Rosenbauer,
H., N. Shutteet al., Ionsof Martin originand plasmasheet
in the Martian magnetosphere:Initial results from the TAUS
experiment,Nature 341, 612, 1989.
Russell, C. T., J. G. Luhmann,R. C. Elphic, and M. Neugebauer,Solar
wind interaction with comets: lessonsfrom Venus, in Comets, L. L
Wilkening,ed., U. Ariz. Press,pp 561-587, 1982.
Russell, C. T., and M. Neugebauer,On the possibilityof detection of
ionsfrom Venusat 1 AU, J. Geophys.Res.,86, 5895, 1981.
H.
Grfinwaldt,
Max-Planck-lnstitut
ffir
Aeronomie,
D-37189
Katlenburg-Lindau,
Germany(e-mail: [email protected])
M. Neugebauer,MS 169-506,Jet PropulsionLaboratory,Pasadena,
CA 91109 (e-mail: [email protected])
(ReceivedMarch 10, 1997;acceptedMarch 27, 1997)