Energy spectra of geomagnetically trapped oxygen ions

Radiation Measurements

PERGAMON

ENERGY

Radiation Measurements31 (1999) 595-600

SPECTRA

OF GEOMAGNETICALLY OXYGEN IONS

TRAPPED

M. LEICHER, R. BEAUJEAN AND W. ENGE lnstitut far Experimentelle und Angewandte Physik, Universit~it Kiel, Olshausenstr. 40-60, 24118 Kiel, Germany ABSTRACT In a series of COSMOS satellite flights plastic nuclear track detectors have been exposed in low-earth orbits to monitor anomalous cosmic rays (ACR) at energies below 25 MeV/nuc. The analysis of energy spectra has now been extended to energies up to 40 MeV/nuc tor two exposures aboard COSMOS 2260 in 1993 and COSMOS 2311 in 1995. Our data on trapped ACR (TACR) oxygen energy spectra might indicate the influence of energydependent stripping probabilities and the presence of multiply charged ACR oxygen at high energies as reported by latest SAMPEX observations.

KEYWORDS Trapped oxygen energy spectra; plastic nuclear track detectors; low-earth orbit satellites. ~TRODUCTION A series of observations with passive nuclear track detectors aboard COSMOS satellites in low-earth orbits (250-400 km) showed highly anisotropic angular distributions of arrival directions for energetic oxygen ions below 25 MeV/nuc near the minimum of the 21 ~t solar cycle (Grigorov et. al., 1991), Theory and detailed Monte Carlo simulations strongly suggested anomalous cosmic rays (ACR). which have been magnetically trapped, as origin fox:these ions. Singly ionized ACR ions penetrating deeply into the magnetosphere may be tied t o a magnetic fieldline if they are stripped off their remaining electrons in the residual atmosphere and change their gyro radii due to the change of their charge state. With lifetimes between hours and months these ions oscillate between a southern and northern mirror plane. In their mirror plane the particle trajectory is a circle perpendicular to the fieldline which develops into a helix around the fieldline with increasing distance of the particle from the mirror plane. Furthermore, the ion drifts westwards and the mirror planes form rings around the earth determined by constant values of L (Mclllwain parameter, related to the magnetic latitude of the mirror plane) and constant values of the magnetic field B (related to the altitude of the mirror plane). In the south atlantic anomaly (SAA) the radiaton belt is accessible to measurements with satellites on low-earth orbits due to the shift between geomagnetic dipol and the earth's center. Fluxes of trapped ACR ions rapidly increase with altitude. The Monte Carlo simulations of particle trajectories near their mirror planes transformed to detector coordinates showed good agreement with the anisotropy of ion arrival directions discovered by the COSMOS satellite experiments, The low ratio between carbon and oxygen ions excluded solar wind or solar energetic particles as possible source. Moreover, the continuous measurements indicated a close temporal correlation between ACR oxygen fluxes and the anisotropic component and showed the presence of a trapped-particle component formed by trapped ACR oxygen ions (TACR) in the inner 1350-4487/99/$ - see front matter © 1999 ElsevierScience Ltd. All rights reserved. PII: S 1350-4zig7(99)00153-5

596

M. Leicher et al. / Radiation Measurements 31 (1999) 595-600

magnetosphere. It was also demonstrated, that passive nuclear track detectors flown on three-axis stabilized satellites allow the study of trapped particles without recording timing information for particle capture in the detector. On three-axis stabilized spacecrafts the detector always passes the SAA with the same attitude, trapped particles in or near their mirror plane arrive from characteristic directions and are collected at characteristic angles in the detector. With appropiate selection of an area of particle arrival directions, the trapped ions can be isolated with a calculable background of ACR ions and particles of galactic origin (GCR) which arrive isotropically at the detector. Recent measurements during solar cycle 22 with time resolving active experiments aboard the SAMPEX satellite confirmed the general features of the COSMOS data (Selesnick et. al., 1995). The energy spectra of TACR oxygen observed during solar cycle 21 well agreed with model calculations below 20 MeV/nuc. Our recent measurements in solar cycle 22 aboard COSMOS 2260 in 1993 (Leicher et. al., 1995) comprise the low energy range of the previous investigations and extended the energy range of the TACR oxygen ion energy spectrum to 40 MeV/nuc. We found a suppression Of TACR oxygen flux at high energies (> 20 MeV/nuc), at 37 Mev/nuc about ten times less than predicted by the model. The track detector stack aboard COSMOS 2311 in 1995 using an improved experimental setup during 39 days of exposure in orbit now allows a reinvestigation for energies above 25 MeV/nuc.

EXPERIMENTALSETUP Stacks of minimally shielded (51am Mylar for thermal protection) plastic nuclear track detectors (Cellulose Nitrate, CNK and CND) have been exposed aboard three-axis stabilized satellites COSMOS 2260 for 14 days end of July 1993 and COSMOS 2311 for 39 days starting end of March 1995. The mission averaged altitude in the SAA for COSMOS 2260 is 305 km and 330 km for COSMOS 2311, the orbit inclination was 82.3 ° resp. 67.2 °. The detector stack aboard COSMOS 2260 was built from 40 sheets of CND (2501am) with an area of 30 cm 2. The detector sheets were etched in 6N NaOH for lbur hours at a temperature of 40 °. Most of the stopping oxygen ions will form a double etch cone in the adjacent sheet and conelength versus range measurements in two foils only are sufficient to identify C/N/O ions in most cases. To achieve a high scanning effectivity all sheets were scanned twice and the coordinates of a detected cone were transformed to the sheets above and below, and even small stopping tracks or single cones could be found by tracing a double cone. Since oxygen in our experiment forms cones in three foils, an ion track could only be missed during the scanning, if all three cones were overlooked twice. We evaluated sheet 2-10 by conelength versus range measurements with a charge resolution of 0.2 charge units and measured stopping tracks in sheet 1. The detector stack aboard COSMOS 2311 was composed of 8 sheets of CNK (100~tm) and 22 sheets of CND (250 ~tm) with an area of 27 cm 2 . The top stack of sheet 1-8 CNK was evaluated at Moscow State University while sheet 9-17 of CND were investigated at Kiel University. We measured all tracks in sheet 9 (-1000 tracks) and scanned twice for stopping tracks in sheet 10. To check our scanning effectivity we traced all cones from sheet 9 into sheet 10 and only few additional stopping tracks were found in sheet 10. The procedure of scanning twice for stopping tracks was adapted for all following detector foils. The data comprise oxygen ions from sheet 10-17, while sheet 9 was used to identify charges of stopping tracks in sheet 10.

RESULTS The anisotropy of the azimuth distribution of arrival directions shown in Fig 1. for oxygen ions with E > 25 MeV/nuc is a typical feature of trapped particles observed on three-axis stabilized satellites. Similar characteristics can be seen in the Monte Carlo simulation of orbit-averaged arrival directions for trapped particles near their mirror planes and isotropic ACR and GCR particles. The simulation for isotropic ions shows, that for the specific attitude vertical to the earth's surface of our detector stack aboard COSMOS 2311 most particles are held off by the earth's shadow between azimuth angles from 120 ° to 240 ° and trapped particles can be measured with a low background of ACR and GCR ions. For the simulations we sampled and orbit averaged particle arrival directions on the COSMOS 2311 orbit at L-values from 1.9 to 2.1 (Selesnick et. al., 1995). According to the mirror equation the mirror

597

M. Leicher et al. / Radiation Measurements 31 (1999) 595-600

plane of a trapped particle drifts along lines of constant L and constant values B of the magnetic field. We calculated the minimal altitude hmin of the mirror plane for each (L,B)-coordinate along the orbit and sampled particles at areas, where hmin is close to the altitude of the satellite. Since T A C R oxygen fluxes increase rapidly with altitude, there is only a low contribution of ions mirroring far below the satellite's orbit. W e applied a rectangular pitch angle distribution and fitted the width of the distribution (best fit at 30 ° ) and the fraction of isotropic particles to match our data on oxygen ions. 0.20,

I

I

I

I

I

I

t

t

I

I

I

COSMOS 2311 D A T A Trapped + Isotropic Simulation Isotropic Simulation

--I i .......

o

[.., O 0 o

i 0 •

0

i

i

I

+

30

60

90

120

i .......

150

f ......

180

!

i

210

240

270

300

330

360

Azimuth Angle (degrees) Fig. 1. Azimuth angular distribution of arrival directions for oxygen ions with E > 25 MeV/nuc. Measurement aboard COSMOS 2311 and Monte Carlo simulation.

9O

9O 120

30

I~ol

/I 0

150

o

't

I I

60

30

l~o

I o J

I,

"

..

270

/

2L0

330

270

Fig. 2. Measured distribution of arrival directions for oxygen ions at E > 25MeV/nuc aboard COSMOS 2311 (left) and low statistics Monte Carlo simulation tbr trapped particles only (right). Azimuth angles are drawn circular, dip angles radial, measured from vertical.

598

M. Leicher et al. / Radiation Measurements 31 (1999) 595-600

By using restricted angular ranges of arrival directions, determined by Monte Carlo simulations, the non-isotropic trapped and the isotropic non-trapped component of the oxygen flux can be seperately evaluated. Figure 2. shows a comparison of measured arrival directions of oxygen ions for energies > 25 MeV/nuc with a Monte Carlo simulation for trapped particles only. A three-axis stabilized spacecraft passes the SAA in two different orientations, ascending on its orbit from south northwards and descending from north southwards, and two pictures from trapped particles near their mirror plane can be seen as lines of intersection between detector and mirror plane. The simulation shows trapped particles (Fig. 2, right) in the azimuth range from 900 - 240 ° and 270 ° - 330 °. The area without trapped particle contribution of azimuth angles between 0 ° and 60 ° and dip angles between 20 ° and 45 ° is used to calculate fluxes for the background of ACR and GCR ions. Tracks with steep dip angles below 10° are excluded due to low detection efficiency. Tracks with dip angles greater than 45 ° are excluded due to obstructions by the satellite to the particle trajectory. With the method of charge identification by conelength versus range measurements the solid angle of detector acceptance is not constant but a function of particle range or ion energy and usually corrections to the flux calculations have to be applied. Using angular distributions of arrival directions derived from models, the corrections to the fluxes calculate the solid angles not accessible to measurement and scales the observed fraction of particles to the complete solid angle selected for measurement. We now use a different approach to calculate omnidirectional T A C R oxygen fluxes avoiding flux corrections by evaluating a large number of sheets from our detector stacks. We identified oxygen ions in our detector stack flown aboard COSMOS 2260 up to sheet 10 and achieved a constant solid angle of detector acceptance for energies up to 37 MeV/nuc. For low energies we included stopping tracks from sheet 1 into our flux calculations and, to eliminate light secondary particles, set the minimum range to 100 gm corresponding to a lower boundary at 6 MeV/nuc of the energy interval with constant detector acceptance. From the analysis of another detector stack flown aboard COSMOS 2260 we infer, that the contamination of our TACR oxygen fluxes by other elements at low energies is below 10% (Grigorov er al., 1995). The COSMOS 2311 detector stack evaluated from sheet 10-17 provides an energy interval from 25 to 40 MeV/nuc to calculate T A C R oxygen fluxes without flux corrections.

10 .4 -

I

'

f

'

'

I

'

,

,

~ I , , , ~1 , ~ '

"7 10 5 > m

,j

,.r ,-d

i ¸

>(

10 .6

q "~ ! ........

i0 ~]

,

GCR

,

~

......

,

10

I

,

*

,

20 Energy

L I

h k h

30 ( MeV/nuc

)

I,

40

10 .7

©

......

t',i~_l

• O

I

10-s/ 50

10

COSMOS 2260 C O S M O S 2 2 6 0 (preliminary)

a

,

,

l 20

t

,

,

, J 30

Energy ( MeV/nuc )

Fig. 3. Comparison of non-trapped oxygen energy spectra measured aboard COSMOS 2260 and model calculations (left). Omnidirectional TACR oxygen spectra for COSMOS 2260 using different methods of calculation (right, see text).

lit, 40

i 50

M. Leicher et aL / Radiation Measurements 31 (1999) 595-600

599

In Fig. 3 (right) the TACR oxygen fluxes observed aboard COSMOS 2260 are calculated using the method of isolating TACR ions in restricted angular areas as described above, taking into account the background of ACR and GCR. We applied flux corrections to our preliminary data (Leicher, et. al., 1995) whereas now we calculate TACR oxygen flux in an energy range of constant detector acceptance. Within statistical limits there is no difference between the different methods to calculate omnidirectional fluxes, below 20 MeV/nuc the preliminary data were derived from not identified stopping tracks in sheet 1-3 which may account for a divergence at low energies. The isotropic ACR and GCR contribution to our TACR oxygen measurements is shown in Fig. 3 (left). The model calculations for ACR energy spectrum use a functional form derived from previous investigations aboard COSMOS satellites during the 21 ~t solar cycle (Tylka, et. al., 1995), the functional form of the GCR energy spectrum is given by the CREME code. The model spectra in Fig. 3 (left) are not a fit to our data, the flux levels of the model spectra are adjusted to simultaneous SAMPEX observations of ACR and GCR oxygen during the mission time of COSMOS 2260 in 1993 (Selesnick, et. al., 1995). In Fig. 4 (left) oxygen fluxes for the non-trapped ACR and GCR component measured aboard COSMOS 2311 show good agreement with model calculations derived from our previous investigation in 1993. We raised both ACR and GCR flux levels, as determined in 1993, by 2.5 to match our data from COSMOS 2311 in 1995. An interpolation of available SAMPEX data on interplanetary ACR oxygen 1

0-

5 - ~

~

i

I

i

,

'

'

F '

'

10-4 ..

' ~ I ' ' ' t

,

[

t

'

'

I

F

,

r

I

i,

l

~ ,rfi,,

"7

10"s >, :> e-i

104

y.

COSMOS

2311

* 0,31

"'

g

X

.~

.u

,..2

--

10.7

©

1i lOS[___i

10

.................

G*=2p.g/cm 2 G*=51.tg/cm 2 O*=101~g/cm 2

[ [

i i

i

[

20

A_

,

*

t

I

30

Energy ( MeV/nuc )

,

, J , I , I , ,

40

10.8 50

[ 10

J

I

I

I

[ 20

I

I

t

[

I IllllL~lJ

30

40

50

Energy ( MeV/nuc )

for the mission time of COSMOS 2311 might suggest an ACR flux level higher by a factor of -2 (Selesnick, et. al., 1997a). Fig. 4. Orbit-averaged non-trapped oxygen fluxes inside the magnetosphere observed on COSMOS 2311 (left) and orbit-averaged omnidirectional TACR oxygen fluxes observed on COSMOS 2260 and 2311 compared to model predictions for the COSMOS 2260 orbit (see text). TACR oxygen spectra sampled at different locations show different shapes and flux levels and can be parameterized by the local McIllwain parameter L. The high inclination orbits of both experiments aboard COSMOS 2260 and COSMOS 2311 completely covered the area of the SAA and we assume that the shapes of our orbit-averaged TACR oxygen spectra sampled in the same range of L-values will not differ significantly. To compare the shape of the energy spectra measured aboard COSMOS 2260 and COSMOS 2311 in Fig. 4 (right) we normalized our fluxes from COSMOS 2311 to the COSMOS 2260 flux level at 30.5 MeV/nuc. The COSMOS 2260 datum at 37 MeV/nuc is taken from our previous investigation, where flux corrections were applied at the upper boundary of the energy interval. The new data on TACR oxygen from our experiment aboard COSMOS 2311 in 1995 now confirm the shape of the oxygen energy spectrum and the suppression of fluxes at high energies

600

M. Leicher et al. ~Radiation Measurements3l (1999) 595-600

compared to model calculations for the COSMOS 2260 orbit as observed in 1993. Aboard COSMOS 2311 in 1995 we measured a TACR oxygen flux at 30.5 MeV/nuc higher by 3 compared to our measurements in 1993, simultaneous SAMPEX obervations show an increment of flux level by 2-3 between 16 and 30 MeV/nuc (Selesnick, et. al., 1997a). The new data on TACR oxygen from our experiment aboard COSMOS 2311 in 1995 now clearly show an increasing divergence with ion energy from model calculations.

DISCUSSION The oxygen flux levels measured with passive nuclear track detectors aboard the russian three-axis stabilized satellites COSMOS 2260 and COSMOS 2311 at an altitude of -300 km are in good agreement with SAMPEX data of corresponding time periods sampled at -600 km. The new TACR oxygen energy spectrum of COSMOS 2311 in the energy range from 25-40 MeV/nuc now confirms the spectrum observed on COSMOS 2260 in 1993. The TACR oxygen energy spectra show an increasing divergence with energy compared to model calulations for G* = 2 [ag/cm 2 (G* = mean free stripping pathlength for stable trapping), whereas good agreement was achieved at energies below 20 MeV/nuc in previous investigations (including COSMOS 2260). In Fig. 4. (right) the comparison of data and model calculations for different drift-averaged stripping probabilities parameterized by G*, might suggest an energy dependence of the stripping cross sections. However, a recent analysis of ACR charge states revealed that multiply charged ACR oxygen (which does not contribute to geomagnetically trapped TACR oxygen fluxes) is predominant at energies above 23 MeV/nuc (Klecker et. al., i 997). The fraction of multiply charged ACR oxygen increases with energy up to 90% between 30 and 40 MeV/nuc (Selesnick et.aL, 1997b). Only 10 % of the ACR oxygen is singly ionized at these energies, which is of the order of the divergence between our data at 37 MeV/nuc and the model calculations, assuming ACR oxygen to be singly ionized at high energies.

Acknowledgement: We thank N.L. Grigorov for the exposure on the COSMOS satellites and A.J. Tylka fol TACR oxygen model calculations tbr COSMOS 2260. Part of this work was financially supported by DARA grants WB 9418 and OS 9401. REFERENCES: Grigorov N.L., Kondratyeva M.A., Panasyuk M.I., Tretyakova Ch.A., Adams J.H., Blake J.B., Schulz M., Mewaldt R.A. and Tylka A.J., (1991), Evidence for trapped anomalous cosmic ray oxygen ions in the inner magnetosphere, Geophys. Res. Lett. 18, 11, 1959-1962. Grigorov N.L., Kondratyeva M.A., Tretyakova Ch.A., Panasyuk M.I., Zhuralev D.A., Beaujean R., Leicher M., Adams J.H. and Tylka A.J., (1995), Fluxes of ACR ions in the earth's magnetosphere at the end of the 22 "d solar cycle, In: Proc. 24 'h Int. Cosmic Ray Conf. 4, 832-835. Klecker B., Oetliker M., Blake J.B., Hovestadt D., Mason G.M., Mazur J.E. and McNab M.C., (1997), Multiply charged anomalous cosmic ray N, O, and Ne: Observations with HILT/SAMPEX, In: Proc. 25 rh Int. Cosmic Ray Conf. 2, 273-276. Leicher M, Tylka A.J., Beaujean R and Enge W., Energy spectra of oxygen ions at E < 40 MeV/nuc inside the magnetosphere, In: Proc 24 'h Int. Cosmic Ray Conf. 4, 1017-1020. Selesnick R.S., Cummings A.C., Cummings J.R., Mewaldt R.A., Stone E.C. and von Rosenwinge T.T., (1995), Geomagnetically trapped anomalous cosmic rays, J. Geophys. Research 100, A6, 95039518. Selesnick R.S., Leske R.A., Mewaldt R.A. and Cummings J.R., (1997a), Geomagnetical[y trapped anomalous cosmic rays at solar minimum, In: Proc 25 'h Int. Cosmic Ray Conf. 2, 305-308. Selesnick R.S., Mewaldt R.A. and Cummings J.R., (1997b), Multiply charged anomalous cosmic rays above 15 MeV/nucleon, In: Proc. 25 th Int. Cosmic Ray Conf. 2, 269-272. Tylka A.J., Adams J.H., Grigorov N.L. Kondratyeva M.A., Panasyuk M.I. and Tretyakova Ch.A., (1995), COSMOS results on the altitude dependence of geomagnetically trapped anomalous cosmic rays, In: Proc 24 'h Int. Cosmic Ray Conf. 4, 485-488.