1991-10
GEOPHYSICALRESEARCHLETTERS, VOL. 18, NO. 11, PAGES 1959-1962,NOVEMBER 1991
EVIDENCE
FOR
TRAPPED
IN
THE
ANOMALOUS
INNER
COSMIC
RAY
OXYGEN
IONS
MAGNETOSP••
N.L.Grigorov
1, M.A.Kondratyeva
I , M.I.Panasyuk
1, Ch.A.
Tretyakova
1,
J.H.Adams,
Jr.2, J.B.Blake
3, M.Schulz
3, R.A.Mewaldt
4, A.J.Tylka
5
Abstract.
A series
of
measurements
of 5-30
detector stacks flown for ~14 days on Cosmos satellites in
nearly circular orbits at altitudes of 250-400 km and
MeV/nucleon oxygen ions made with track detectorstacks
on Cosmossatellites show isotropic angular distributions
duringsolarenergeticparticleevents. Solar-quiettimes,on
the other hand, have highly anisotropic distributions
suggestive
of a trapped-particle
component.DetailedMonte
Carlo simulationsconfirmthis interpretationand allow us to
measurethe trapped and cosmic-ray contributionsto the
observed fluxes. Our data are fully consistent with
anomalouscosmic-rayions,ratherthanradialdiffusionfrom
the outerzone, asthe sourceof the trappedparticles.
inclinationsof 620-82ø. The detectorsregisteredand
identified CNO ions. Lighter species(suchas He) did not
leave tracksin the detectors,and heavier speciescould not
be identified. Although the detectorsrecordedno timing
informationfor individualnuclei,the angulardistributionsof
the tracks clearly showedthe presenceof trappedparticles.
Sincethe spacecraftwere three-axisstabilized,the detectors
always passedthroughthe low-altitude radiation belts with
the sameattitude. Trappedparticlesthus arrived from
characteristic directions and registered at characteristic
angles. Data from these same flights, with angular cuts to
exclude trappedparticles, have previously been combined
with simultaneousmeasurements
outsidethe magnetosphere
to demonstrate conclusively that anomalouscosmic-ray
oxygenions are singly ionized [Adamset al., 1991a].
Introduction
There have beenseveralreportsof excessoxygenflux at
energies ~10 MeV/nucleon in the inner magnetosphere
[Mogro-Camperoand Simpson,1970; Chan andPrice, 1975;
BiswasandDurgaprasad,1980; Oschlieset al., 1989; Adams
et al., 1991b]. Theseobservations,however, were unableto
establishthe nature of this flux and, in particular, to prove
that the excessionsfollowed trapped-particle
trajectories.If
trapped,suchions may yield insightsinto magnetospheric
processesand may also be of astrophysicalinterest. The
anomalouscomponentof cosmicrays [reviewedby Webber,
1989] comprisesprimarilyHe, N, O, andNe atomsfrom the
local neutral
interstellar
medium
which
Observations
The dip angle 0 and azimuth angle q>,as defined in
Figure 1, specifya particle's arrival directionin the detector
coordinate system. Figure 2 shows the observed arrival
directions for particles collected in two typical Cosmos
exposures,one when solar energetic particles (SEPs)
have entered the
dominatedthe fluxes andone duringquiet-time,as identified
by Caltech proton and He measurements on IMP-8 in
interplanetary space [cf. Adams et al., 199 la]. During the
SEP-dominatedperiod,the particlesarrivedfrom aroundthe
heliosphere,become singly-ionized by the solar wind or
solar UV [Fisk et al., 1974], and then been acceleratedto
energies >•10 MeV/nucleon, probably at the solar-wind
termination shock [Pesseset al., 1981]. Blake and Friesen
[ 1977] notedthat suchparticlescouldpenetratedeeplyinto
the magnetosphere
andbecome
geomagnetically
trapped
after being stripped of remaining electrons in the residual
atmosphere,with subsequent
lifetimesrangingfrom hoursto
months. In this case,thesetrappedparticlesare a sampleof
interstellarmatterdirectly availablefor studyat Earth.
In this Letter we reporton a seriesof observationsof 530 MeV/nucleon oxygenions in the inner magnetosphere
in
1986-89 [cf. Grigorov et al., 1989 and referencestherein].
The observations
were made with small
cellulose
nitrate
linstitute
forNuclear
Physics,
Moscow
State
University
2E.O.HulburtCenterfor Space
Research,
Naval
ResearchLaboratory
3Space
andEnvironment
Technology
Center,
The
AerospaceCorporation
4California
Institute
ofTechnology
5Universities
Space
Research
Association
Fig. 1. Arrival-direction anglesin the detectorcoordinate
system. Dip angle 0 is the angle between the particle's
velocity vectorv and the normalto the detector,n. Azimuth
angle q>is the angle between the projection of v onto the
Copyright1991by the AmericanGeophysical
Union.
Paper number91GL02551
detector
0094-8534/91/91 GL-02551 $03.00
detectorplane.
1959
surface
and a fixed
reference
direction
x in the
Grigorov et al.: TrappedAnomalousCosmicRays
SEP
5-18Moy1989
.
•
Quief Time
5-19Moy1987
•,•
ß
-'-:.'::
?.-:..1.:'.-:::
::::
•:
::::::::::..:
:::.'..:.'..•ß-'.
F:.:•:::.:
.::
.:.::.{.:•
..................
I•....:.:::li
.........
....................
.n
p/•
:::::-:
................
:::::::::::::::::::::::::::
....................
I•.........................
!•.ß........
.....
::::::::::::::::::::::::
.....................
:::::::::
Fig.2. Polardiagrams
ofobserved
arrival
directions
ina SE?-dominated
exposure
(left)andinaquiet-time
exposure
(right).Eachpointrepres6nts
a singleparticle,
with0 and• plotted
astheradialandazimuthal
coordinates,
respectively.
Thezenithandnadirdirections
arenoted.Shaded
areasshowarrivaldirections
blocked
byobstructions
on ttie satellite. The hatchedarea showsarrival directionswhich are below the Earth's horizon. Tracks at 0 < 10ø and
0 > 80ø ar• omitteddue to detectioninefficiencies.
sampled
•roma model
oftheincident
fluxto
zenith
di•;$ction
inadistribution
which
wasisotropic,
except simulation
for shadows
causedby obstru(•tions
onthesatelliteandthe
s0iidEarth. Du{ingquiet-iinie,however,
therewere
•el•tivelyfew particlesfromaroundthezenithdirectiori.A
laq•e6umberof particles
camefrombelowthehoriz9n.
Cosrribs
expos.ures
weremadewithdetectors
mounted
in
oneof threepossible
locationsonthesatellite'•spherical
specify a particle's velocity vector and energy. The Monte
Carlo programthen accountedfor lossesdue to obstructions,
somesmalldetectioninefficiencies,andthe obliquityfactor.
For each Cosmos flight we simulated the 6xpected
distributionsof bothcosmicraysandtrappedpartii:les.We
thenminimized
X2tofittheobserved
azimuth
distribution
to
exterior,
correspondirig
tothreedifferent
attitudes
ini9assing alinear
cbmbination
ofcosmic-ray
andtrapped
compQnents.
throughthe radiationbelts. The two exposures
shownin
Figure2 camefromoneofthese
attitudes.
In Figure3, we
show
distributions
oftheazimutl•
angle.
• fromsixdifferent
Cosmos.
exposures,
including
an•;EPeventanda quiet-time
exposurein eachdetectorattitude. Strikingdifferences
b.etweenSEP andquiet-timeexposuresare seenin all three
attitudes. Apart from obstructed arrival directions, the
In simulatingSEPsand quiet-timecosmicrays, the flux
was assumedto be isotropicexceptfor directionsbelow the
Earth's horizon,wherethe flux was setto zero. (Cosmic-ray
anisotropies,such as the east-westeffect, and penumbral
effectsare smallthroughoutthe rangeof particlerigidities
involved here.) Particle energieswere sampledfrom orbit-
averaged
spectra,
asdetermined
fromcontemporaneous
flux
observ.ed
SEPdistributionsare flat. In the quiet-time
measurements
from IMP-8 convolvedwith the geomagnetic
exposures,
however,
thefluxesarehighlyanisotropic,
with
transmission
function
[cf.Adams
etal.,1991
a].
TosimdlMe
trapped
particles,
weconstructed
a simple
strongazimuthalvariation.
modelbasedon thefactthattrappedionsareobsei'vable
at
Monte Carlo Simulations
Cosmosaltitudesonly neartheir mirror pointsin the Souih
Atlan{icAnomaly(SAA). Bothanomalous-component
Theangular
distributions
inFigures
2 and3 strongly trapping[BlakeandFriesen,1977]andradialdiffusionof
.suggest
a trapped-particle
component
in theCosmos
data
[Origoi'ovet al., 1990]. We testedthis interpretationwith
detailed
MonteCarlosimulMions
oftheexposures,
asshown
in Figure 3. The SEP simulations ensure that the
obstructionsare well understood. The quiet-time
simtilations
shdwthata trapped-particle
component
givesfin
internallyconsistent
view of thedata,by accouhting
for the
differencesamongthe threedetectorattitudes.
To carry out the simulations we usedtracking data to
reconstructthe orbital trajectoryof eachflight on a secondby-secondbasis. At eachlocationalongthe trajectory,the
ions from the outer zone [Panasyuk, 1984] predict that
trapped oxygen ions of these energies should be found
primarily at L > 2. The simulationsthereforerestrictedthe
trapped flux to the portion of the SAA at 2 < L < 3. The
intensityand spectrumof the trappedparticleswere assumed
to bethesameeverywhere
withinthisregion.In simulating
trapped-particlevelocity vectors(v), the local pitch angle
distribution was apprbx.imated by sampling the angle
between v and the local magnetic field vector B (as
calculatedfrom IGRF 1985 [Barracloughet al., 1987]) from
a Gaussiancenteredat 90ø with st•ndard deviation 6ø. The
Origorov
etal.:Trapped
Anomalous
Cosmic
Rays
1961
Isotropic
Solar
Energetic
ParticlesIsotropic
Cozmlc
Rays
+ Trapped
Particles
0.•0
-
el
=
'
52.•Cosmic
Rays
+ 48Y.
Trap
0.10
o
•u- 0.20
,,,,,,,,,,,,,:
Detector
Attitude
'1'
I•1•I•1'1'1•1•1,1,
I•1,1,1=,i,i ,i
'l'l'l'l'l'l'l'l'l'l'l*l'l,l,l,i,l•!
- Defector
Attitude
2
_•/•/ . _.. -I
X-/ =
. 43XCosmic
Rays
+ 57• Trapped
.J
8-22 Dec 1988
_ 4-18 Nov1986
o
_oO.lO
i
0
i,
, •, I , t, I , i';'i';"r
/' I • I • I" I ' I • I , I , I • I , i , i • i , i • i',i _F'T'I-•
I' Delec[or
Attitude
$
X8
/
-•0.20•5-18
May
1989
0
o ,.,,,,,,,
0
60
Defector Attitude ,3
. lZ/v=
1.88
'1'i•1'1'1,1,1•1•1,1
•1,1,1,
i,i,
f,l•
1
5-1g
May
1987
26X
Cosmic
Rays
+74X
.rapped
•
;,. ,,]
I,, ;,, ,f, t•, ,':',-r-,
...... 4, ,_._,.,.,-r•
120 180 240 300 360 0
60 120 180 240 300 360
Azlmufh
Angle
(degrees)
Azimuth
Angle
(degrees)
Fig.
3. Distributions
ofthe
azimuth
angle
4)forSEP
exposures
(left)
and
quiet-time
exposures
(right)
forthree
different
detector
attitudes.
The
points
are
the
Cosmos
data
with
statistical
error
bars,
and
the
histograms
are
Monte
Carlo
simulations
(see
text
fordetails).
Inthe
panels
onthe
fight,
the
dashed
lines
show
the
simulated
cosmic-ray
contribution
and
the
solid
lines
show
the
sum
ofsimulated
cosmic-ray
and
trapped
components.
Statistical
e[ors
on
th.e
simulations
are
negligible
compared
tothose
ofthe
data.
The
dates
for
each
exposure,
the
reduced
X
ofthe
s•mulations'
fitstothedata,
and
thebest-fit
percentages
ofcosmic
rays
and
trapped
particles
inthequiet-time
exposures
arealso
shown.
Thebottom
twopanels
show
thesame
data
asinFigure
2.
component
of v perpendicular
toB wasthenrandomly
Originof theTrapped
Particles
oriented.
Trapped-particle
energies
weresampled
froma
relatively
flatspectrum,
assuggested
bymodels
of both
Wehaveconsidered
twopossible
originsforthese
anomalous-component
trapping
[Blake,1990]andradial particles.
The
first
is
trapping
of
anomalous
cosmic-ray
ions
diffusion
[Spjeldvik
andFritz,1978].
[BlakeandFriesen,1977].All of ourobservations,
We used the simulations to check the assumed including
geomagnetic
location,
spectrum,
composition,
and
geomagnetic
distribution
of trapped
particles.
In addition
to
temporal
variation,
appear
consistent
withthisexplanation.
placing
thetrapped
fluxat2 < L < 3, wealsoransimulations A second
possible
originisradialdiffusion
fromtheouter
inwhich
trapped
oxygen
particles
were
assumed
throughoutradiation
belt. Thisexplanation,
however,
appears
theSAAoronlyat1.2<L < 1.5(where
thetrapped
proton inconsistent
withthedataintwoways.First,standard
radial
fluxismost
intense
atCosmos
altitudes).
Inthese
cases,
the diffusion
theory
isunable
toaccount
fortheacceleration
of
simulated
azimuthal
accumulations
shifted
soastoproduce ionospheric
ionstotheenergies
observed
here.Outer-zone
significantly
pooreragreement
withthedata.
The agreement
betweenthe dataandsimulations
in
ionswiththenecessary
initialenergies
musttherefore
beof
solar
wind
or(more
likely)
ofSEPorigin.
Forthese
sources,
Figure
3 isgenerally
good,
thus
confirming
thatthequiet- theC/Oratioistypically-0.5. Theobserved
C/Oratioin
timeazimuthal
accumulations
areduetotrapped
particles theCosmos
data,however,
is <_a fewpercent,
consistent
neartheirmirror
points.
Some
discrepancies
remain,
butthis with
theobserved
composition
ofanomalous
cosmic
rays
at
isnotunexpected
given
thesimplicity
ofourtrapped
particle this time [cf. Adamset al., 1991a].It is difficult to
model.In additionto azimuthdistributions,
wealso understand
howradial
diffusion
could
produce
such
alarge
compared
theobserved
andsimulated
distributions
ofdip change
inparticle
composition.
Second,
thetrapped
fluxis
angles
andparticle
ranges
in thedetector.
Wealsochecked variable
andanticorrelated
withsolaractivity.Figure4
thecorrelations
among
the0,4),andrange
observables.
In compares
thetimehistory
ofthemeasured
trapped-oxygen
allcases
therewassatisfactory
agreement
withthedata.
fluxnearthemirror
point,
Jj., withtheIMP-8quiet-time
1962
Grigorovet al.' TrappedAnomalousCosmicRays
10 -1
References
10-5 -o
-111111111111111111111111111111111111111111
I
_
-
_
Adams, J.H., Jr. et al., The charge stateof the anomalous
component,Astrophys.J. Lett., 375, IA5-L48, 1991a.
Adams, J.H., Jr., L.P. Beahm, and A.J. Tylka, The charge
state of the anomalous component: Results from the
Trapped Ions in Spaceexperiment,Astrophys.J., 377,
_
ß Cosmos
7
-oIMP-8
_
ß
-
•c••P'• -
10 -2 --
x
I
10-6
u•
I
_
ß
292-305, 1991b.
Barraclough, D.R. et al., International Geomagnetic
ReferenceField Revision 1987, J. Geomag.Geoelectr.,
I
-
39, 773-779, 1987.
_
10-3 IIIIIIIIIIIIIIIIIIIIIII
51985
6 9 I 51986
69
Fig.4.
IIII1111111
5
6
1987
IIIIII
9
5
6
i
1988
Time historyof the trappedflux J_L(left-hand
scale)measuredon the Cosmosflights andof the quiet-time
5-11 MeV/nucleon oxygenflux (right-handscale)measured
by the Caltech instrumenton IMP-8. The IMP-8 fluxes are
dominatedby the anomalouscomponentand uncorrectedfor
contributionsfrom Galactic cosmicrays and solarparticles,
which are small (<10%) exceptat low flux levels.
measurementsof the anomalous-componentoxygen flux in
interplanetary space. The temporal variations are very
similar, and both the anomalouscomponentand the trapped
flux reachedtheir peaksnear solar minimum in early 1987.
Because radial diffusion is promoted by large magnetic
stormscausedby solarflares,it is unclearwhy a trappedflux
from thisprocesswouldpeakat solarminimum.
In future work we will search for other trapped heavy
ions species.We will also build a detaileddynamicalmodel
of anomalous-component
trappingwhich may be testedwith
the Cosmos data.
In summary, comparing the Cosmos data with our
simulateddistributionsclearly demonstrates
the presenceof
trappedenergeticoxygen ions in the inner magnetosphere.
The characteristicsof this trappedflux are fully consistent
with originating from the anomalouscomponentof cosmic
rays. This work also demonstrates that passive track
detectorsflown aboard3-axis stabilizedspacecraftcan be a
powerful tool for studying trapped particles. With
appropriateselectioncriteria on particlearrival direction, a
sample of trapped particles with a small, calculable
background of cosmic rays can be isolated. Since track
detectorswith large collecting area can be inexpensively
manufactured
andflown, this methodcanbe usedto explore
even rare componentsof the trappedradiation.
Acknowledgments. We gratefully acknowledge the
efforts of NASA and IKI in the US/USSR Joint Working
Group on Solar-Terrestrial Physics, which sponsoredthis
study. We especially thank Vernon Jones for his
encouragementand support. We thank D.I. Kozlov, general
contractorfor Cosmos,for the manyflightsthat providedthe
Cosmosdata. We also thank the SpaceAnalysis and Data
Branch (J3SOS) of the US SpaceCommandfor providing
the Cosmosorbital data. This work is supportedin part by
NASA contractW-17,358, by NASA grantNAG5-727, and
by the AerospaceSponsoredResearch(ASR) Programof the
AerospaceCorporation.
Biswas, S., and N. Durgaprasad,Skylab measurementsof
low energy cosmicrays, SpaceSci. Rev., 25,285-327,
1980.
Blake, J.B., Geomagnetically trapped heavy ions from
anomalous cosmic rays, Proc. 2Ist Intl. Cosmic Ray
Conf, 7, 30-33, 1990.
Blake, J.B., and L.M. Friesen,A techniqueto determinethe
chargestate of the anomalouslow-energy cosmicrays,
Proc. I 5th Intl. CosmicRay Conf, 2,341-346, 1977.
Chan, J.H., and P.B. Price, Compositionand energyspectra
of heavy nuclei of unknownorigin detectedon Skylab,
Phys.Rev.Lett.,35, 539-542, 1975.
Fisk, L.A., B. Kozlovsky, andR. Ramaty, An interpretation
of the observed oxygen and nitrogen enhancementsin
low-energy cosmicrays,Astrophys.J. Lett., 190, L35L37, 1974.
Grigorov,N.L. eta!., The chargestateof anomalousoxygen
of cosmicrays,Geomag.and Aeron., 29, 889-891, 1989.
Grigorov, N.L. et al., Anomalousoxygenin the near-Earth
spacein 1985-88, Proc. 21st Intl. CosmicRay Conf, 6,
176-179, 1990.
Mogro-Campero, A., and J.A. Simpson,Identification and
relative abundances
of C, N, and O nucleitrappedin the
geomagneticfield, Phys.Rev. Lett., 25, 1631-1634, 1970.
Oschlies,K., R. Beaujean,andW. Enge, On the chargestate
of anomalous oxygen, Astrophys. J., 345,776-781,
1989.
Panasyuk,M.I., Experimental verification of ion transport
mechanisms
in the radiation
belts of the Earth under the
effect of fluctuatingelectricfields, CosmicRes.,22,468482, 1984.
Pesses,M.E., J.R. Jokipii, and D. Eichler, Cosmicray drift,
shockwave acceleration,and the anomalouscomponent
of cosmicrays,Astrophys.J. Lett.,246, L85-L89, 1981.
Spjeldvik,W.N., andT.A. Fritz, Theory for chargestatesof
energetic oxygen ions in the Earth's radiation belts, J.
Geophys.Res.,83, 1583-1598,1978.
Webber, W.R., Compositionof anomalouscosmicrays,AIP
Conf. Proc., 183, 100-110, 1989.
N.L. Grigorov, M.A. Kondratyeva,M.I. Panasyuk,and
Ch.A. Tretyakova, Institute for Nuclear Physics,Moscow
StateUniversity,Moscow 119899USSR.
J.H. Adams, Jr. and A.J. Tylka, Code 4154, Naval
ResearchLaboratory,Washington,DC 20375-5000USA.
J.B. Blake and M. Schulz, The AerospaceCorporation,
P.O. Box 92957, Los Angeles,CA 90009 USA.
R.A. Mewaldt, 220-47 Downs Laboratory, California
Instituteof Technology,Pasadena,CA 91125 USA.
(Received:August 14, 1991;
acceptedSeptember9, 1991.)