Measurements of Cosmic-ray Low-energy Antiproton and Proton Spectra in a Transient Period of the Solar Field Reversal

4 Pages·7.6 MB·English

by  Y. Asaoka

Measurements of Cosmic-ray Low-energy Antiproton and Proton Spectra in a
Transient Period of the Solar Field Reversal
Y.Asaoka1, Y.Shikaze1, K.Abe1, K.Anraku1, M.Fujikawa1, H.Fuke1, S.Haino1, M.Imori1, K.Izumi1, T.Maeno2,
Y.Makida3, S.Matsuda1, N.Matsui1, T.Matsukawa2, H.Matsumoto1, H.Matsunaga1,†, J.Mitchell4, T.Mitsui2,‡,
A.Moiseev4, M.Motoki1,‡, J.Nishimura1, M.Nozaki2, S.Orito1,∗, J.F.Ormes4, T.Saeki1, T.Sanuki1, M.Sasaki3,
E.S.Seo5, T.Sonoda1, R.Streitmatter4, J.Suzuki3, K.Tanaka3, K.Tanizaki2, I.Ueda1, J.Z.Wang5, Y.Yajima6,
6 3 1 2 3 3
Y.Yamagami , A.Yamamoto , Y.Yamamoto , K.Yamato , T.Yoshida , and K.Yoshimura
1The University of Tokyo, Tokyo, 113–0033 JAPAN
2Kobe University, Kobe, Hyogo, 657–8501 JAPAN
3High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305–0801, JAPAN
4National Aeronautics and Space Administration, Goddard Space Flight Center (NASA/GSFC), Greenbelt, MD 20771, USA
5University of Maryland, College Park, MD 20742, USA
2 6The Institute of Space and Astronautical Science (ISAS), Sagamihara, Kanagawa 229–8510, JAPAN
0 (February 1, 2008)
0
Theenergyspectraofcosmic-raylow-energyantiprotons(p¯’s)andprotons(p’s)havebeenmeasured
2
byBESSin1999and2000,duringaperiodcoveringthesolarmagneticfieldreversal. Basedonthese
n
measurements,asuddenincreaseofthep¯/pfluxratiofollowingthesolarmagneticfieldreversalwas
a
observed, and it generally agrees with a drift model of the solar modulation.
J

5 PACS numbers: 98.70.Sa, 96.40.Kk, 95.85.Ry
2

2
Much of the real underlying physics of the Sun is re- that p¯’s are predominantly secondary in origin, because
v
lated to the 22 yr solar magnetic cycle with recurrent severalrecentcalculationsofthesecondaryspectrumba-
7
0 positive and negative phases. The magnetic field polar- sically agree with observations in their absolute values
0 ity reverseswhenthe solaractivityis maximum,andthe and spectral shapes [3,4,7,8].
9 global magnetic field profile reverses in the heliosphere. Wereportherenewmeasurementsofcosmic-rayp¯and
0
Themostrecentfieldreversalhappenedinthebeginning p fluxes and their ratios in the energy range from 0.18
1
of 2000. The solar modulation of cosmic rays is caused to 4.2 GeV collected in two BESS balloon flights carried
0
/ byexpandingsolarwind,whichspreadsoutlocallyirreg- out in 1999 and 2000, when the solar activity was max-
h
ularmagneticfieldandthereforemodifiesenergyspectra imum. Based on the solar magnetic field data [9], the
p
of the cosmic rays entering the heliosphere. The posi- Sun’s polarity reversed between these two flights [10].
-
o tive and negative particles drift in opposite directions, With our full set of data [6,11–13], we discuss the tem-
r
taking different routes to arrive at the Earth in the he- poral variation of the p¯ flux and p¯/p ratio covering the
t
s liosphericmagneticfield. Thecharge-signdependenceis, solar minimum, the maximum, and the field reversal.
a
: therefore, a natural consequence [1], on top of the com- The BESS spectrometer was designed [14,15] and de-
v
mon time dependent change in the overall modulation. veloped[16–19]asahigh-resolutionspectrometer. Auni-
i
X It explains alternate appearancesof ”flat” and ”peaked” formfieldof1Teslais producedbyathin superconduct-
r periods in neutron monitor data around solar minima. ing solenoid[20],and the field regionis filled with track-
a
Recentworks[2–4]alsoindicatedthatthe driftproduces ing detectors. This geometry results in an acceptance of
non-negligible effects between the positive and negative 0.3 m2sr. Tracking is performed by fitting up to 28 hit-
particles even during the high solar activity. This view points in drift chambers,resulting in a magnetic-rigidity
is supported by measurements of temporal variation in (R ≡ Pc/Ze) resolution of 0.5% at 1 GV. The upper
cosmic-ray ratios [5], such as electrons to helium nuclei and lower scintillator-hodoscopes [19] provide time-of-
(He) and electrons to protons (p’s), where the largest flight and two dE/dx measurements. Time resolution
variation is associated with the solar magnetic field re- of each counter is 55 ps, resulting in a 1/β resolution
versal. Amongvariouscosmic-raypairs,antiprotons(p¯’s) of 0.014, where β is defined as particle velocity normal-
and p’s are ideal [3]for understanding drift effects under ized by the speed of light. The instrument also incor-
the change in overall modulation level, because they are porates a threshold-type Cherenkov counter [18] with a
different only in charge sign. silica-aerogel radiator (n=1.02) that can reject e− and
In the last solar minimum period, the BESS experi- µ− backgroundsbyafactorof4000andidentifyp¯’sfrom
ment revealed that the cosmic-ray p¯spectrum has a dis- such backgrounds up to 4.2 GeV. In addition to biased
tinct peak around 2 GeV [6], which is a characteristic trigger modes [13,17] selecting negatively-charged parti-
feature of secondary p¯’s produced by cosmic-ray interac- cles preferentially, one of every 60 (30) first-level trig-
tions with interstellar (IS) gas. It has become evident gered events were recorded to provide unbiased samples
1
2000 data, respectively. The p¯ bands are slightly con-
2.4 (a) BESS(99) 2.4 (b) BESS(00) taminatedby e− andµ− backgroundsdue toinefficiency
of the aerogel Cherenkov counter, while contamination
2.2 2.2
on the p band was negligible. We estimated the fraction
2 2 of e− and µ− backgrounds to be 0 %(0 %), 0.6 % (0.5
– – %),and2.3%(1.5%), respectively,at0.3,2,and4GeV
1.8 P 1.8 P P
for1999(2000)data. Otherbackgroundssuchasalbedo,
b1/1.6 1.6 D mis-measured positive-rigidity particles, and “re-entrant
668 558 albedo” were found to be negligible.
1.4 Antiprotons 1.4 Antiprotons
T The survival probabilities for p¯’s and p’s to traverse
1.2 1.2 residualair andthe instrument without interactionsand
to pass through all the selectioncriteria described above
1 1
were estimated by a Monte Carlo (MC) simulation. The
0.8 0.8 MC code was tuned andverifiedby comparingthe simu-
-6 -4 -2 0 -6 -4 -2 0 2 4 6
Rigidity (GV) Rigidity (GV)
lation with an accelerator beam test of the BESS detec-
FIG. 1. The identification plots of p¯events for (a) 1999
tor [21]. The systematic uncertainty of the p¯interaction
and (b) 2000 flights. The dotted curves define the p¯ mass
losseswerereducedto5%. Estimatesofthebackground
bands.
from atmospheric secondary p¯’s were calculated [22–24]
in 1999 (2000). Note that the first-level trigger was pro- with the p and He fluxes measured by BESS as input.
videdbyacoincidencebetweenthetopandbottomscin- The subtraction amounts to 15 ± 5 % (17±6 %), 23 ±
tillators,withthethresholdsetat1/3ofthepulseheight 2 % (30 ± 2 %), and 26 ± 8 % (28 ± 9 %) for 0.3, 2,
from minimum ionizing particles. and 4 GeV in 1999 (2000),respectively,where the errors
TheexperimentswerecarriedoutinnorthernCanada, correspond to the maximum difference among the cal-
LynnLake to PeaceRiver,where the geomagneticcutoff culations. The background from atmospheric secondary
rigidity ranges from 0.3 to 0.5 GV. Data for flux mea- p’s was estimated according to the calculation [25]. The
surementsweretakenforlive-timeperiodsof100,471and parameters used in the calculation were tuned and ver-
90,765 sec at altitudes above 34 km (residual air of 4.3 ified [26] by comparing the calculated fluxes at various
and 5.0 g/cm2 on average) in 1999 and 2000, respec- airdepthswiththeatmosphericgrowthcurveofp’smea-
tively. Data during ascent were also collected at various suredbyBESS.Thesubtractionamountsto27±4%(55
altitudes, in order to estimate a background from atmo- ±8%),10±2%(22±3%),and4±1%(8±1%)for
spheric secondary p’s. 0.2, 0.5, and 1 GeV in 1999 (2000), respectively, where
Analysis was performed in the same way as described theerrorscorrespondtothemaximumdifferencebetween
in Ref. [13]. We applied the same selection criteria for observedandcalculatedatmosphericgrowthcurveofthe
p¯’s and p’s because non-interacting p¯’s behave like p’s p’s. Table I gives the resultant fluxes of p¯’s and p’s and
except for deflection in the symmetrical configuration of theirfluxratios,p¯/p,atthetopoftheatmosphere(TOA)
BESS. Figure 1 shows β−1 versus R plots for surviving withthestatistical(first)andsystematic(second)errors.
events(for1999,onlythenegativerigiditysideisshown). Figure 2 shows the p¯ and p fluxes in 1999 and 2000,
We see clean narrow bands of 668 and 558 p¯candidates togetherwiththedata[6]takenin1997atthesolarmin-
at the exact mirror position of the p’s for the 1999 and imumperiod. Theerrorbarsrepresentquadraticsumsof
TABLEI. Antiprotonfluxes(in10−2 m−2s−1sr−1GeV−1),protonfluxes(in102 m−2s−1sr−1GeV−1),andp¯/pratio(inunit
of 10−5) at TOA. T (in GeV) definesthekinetic energy bins. Np¯ is thenumberof observed antiprotons.
BESS 1999 BESS 2000
T (GeV) Np¯ p¯flux p flux p¯/p ratio Np¯ p¯flux p flux p¯/p ratio
0.18 - 0.28 2 0.22+0.28+0.02 10.42+0.09+1.01 0.21+0.26+0.02 7 0.75+0.41+0.08 1.71+0.04+0.35 4.38+2.37+0.94
−0.15−0.02 −0.09−1.01 −0.14−0.02 −0.34−0.08 −0.04−0.35 −1.97−0.94
0.28 - 0.40 9 0.57+0.28+0.04 11.79+0.07+0.60 0.49+0.24+0.04 6 0.40+0.26+0.03 2.28+0.03+0.22 1.76+1.16+0.21
−0.20−0.04 −0.07−0.60 −0.17−0.04 −0.18−0.03 −0.03−0.22 −0.77−0.21
0.40 - 0.56 28 1.25+0.31+0.07 11.76+0.06+0.44 1.07+0.26+0.06 8 0.29+0.19+0.04 2.62+0.02+0.14 1.12+0.74+0.17
−0.27−0.07 −0.06−0.44 −0.23−0.06 −0.16−0.04 −0.02−0.14 −0.60−0.17
0.56 - 0.78 27 0.71+0.22+0.07 10.51+0.05+0.31 0.67+0.21+0.06 36 1.09+0.27+0.08 2.74+0.02+0.10 4.00+0.99+0.29
−0.20−0.07 −0.05−0.31 −0.19−0.06 −0.23−0.08 −0.02−0.10 −0.83−0.29
0.78 - 1.10 63 1.20+0.21+0.07 8.76+0.03+0.25 1.37+0.24+0.08 67 1.35+0.24+0.08 2.64+0.01+0.08 5.11+0.90+0.30
−0.20−0.07 −0.03−0.25 −0.22−0.08 −0.22−0.08 −0.01−0.08 −0.84−0.30
1.10 - 1.53 112 1.85+0.24+0.11 6.70+0.03+0.20 2.77+0.36+0.16 77 1.20+0.21+0.08 2.37+0.01+0.07 5.05+0.88+0.32
−0.23−0.11 −0.03−0.20 −0.34−0.16 −0.19−0.08 −0.01−0.07 −0.82−0.32
1.53 - 2.15 138 1.89+0.22+0.11 4.60+0.02+0.16 4.12+0.48+0.23 102 1.33+0.20+0.08 1.92+0.01+0.08 6.95+1.07+0.44
−0.21−0.11 −0.02−0.16 −0.46−0.23 −0.19−0.08 −0.01−0.08 −1.00−0.44
2.15 - 3.00 162 1.70+0.18+0.13 2.90+0.01+0.11 5.86+0.63+0.42 131 1.32+0.17+0.11 1.44+0.01+0.05 9.20+1.18+0.78
−0.17−0.13 −0.01−0.11 −0.60−0.42 −0.16−0.11 −0.01−0.05 −1.12−0.78
3.00 - 4.20 127 1.20+0.15+0.18 1.70+0.01+0.06 7.05+0.90+0.98 124 1.19+0.16+0.19 1.00+0.01+0.04 11.93+1.59+1.79
−0.14−0.18 −0.01−0.06 −0.85−0.98 −0.15−0.19 −0.01−0.04 −1.50−1.79
2
1) (a) Antiproton 1) (b) Proton BESS experiment
-V -V
e e BESS(97)
G G
BESS(99)
1 1
-c -c BESS(00)
e e
1s 1s103
-r -2 -r Bieber et al,
s10 s
2 2 Drift model
-m -m 10˚, (+)
x ( x ( 70˚, (+)
u u 70˚, (- )
fl fl
– P P Fisk,
Spherically
symmetric model
102 f = 648 MV
99 (upper)
f =1344 MV
-3 00 (lower)
10
-1 -1
10 1 10 10 1 10
Kinetic Energy (GeV) Kinetic Energy (GeV)
FIG.2. The(a)p¯and(b)pfluxesatTOAmeasuredbyBESSin1999and2000togetherwiththepreviousdatain1997. The
curvesrepresentthecalculationsbythedriftmodel[3]ataperiodof10◦(+),70◦(+),and70◦(−),whichrespectivelyrepresent
solar minimum at positive phase (dashed line), solar maximum at positive phase (dash-dotted line), and solar maximum at
negative phase (solid line). The dotted curvesrepresent thecalculations bythe spherically symmetric model [27].
the statistical and systematic errors, as well as in the and some modulation parameters. As shown in Fig. 3,
followingfigures. Thedottedcurvesrepresentthemodu- themeasuredp¯/pratiosareinbetteragreementwiththe
lated spectra according to the standardspherically sym- drift model than the spherically symmetric model.
metric approach [27], in which the modulation is char- Combining all the previous BESS measurements since
acterized by a single parameter (φ) irrespective of the 1993 [6,11–13], annual variations of the p¯/p ratios (rela-
Sun’s polarity. In each year, φ was determined to fit the tivetotheratioinIS)atenergiesof0.2–0.4GeV(closed
pspectrummeasuredbyBESSassumingtheISspectrum square, denoted as 0.3 GeV), 0.8 – 1.3 GeV (open cir-
in Ref. [3]. The same φ was applied to modulate the p¯ cle, 1.0 GeV), and 1.5 – 2.5 GeV (closed star, 1.9 GeV)
flux. Inbothplots,dashed,dash-dotted,andsolidcurves are shown in Fig. 4(a). These data show no distinctive
represent modulated spectra according to a steady-state year-to-yearvariationbefore the solar field reversalindi-
drift model [3] in which the modulation is characterized cated by the shaded area in Fig 4(a). The curves repre-
by a tilt angle (α) of the heliospheric current sheet and sent calculations of the p¯/p time variation according to
theSun’smagneticpolarity(denotedas+/−). Notethat the drift model (dashed) and the spherically symmetric
periods at 10◦(+), 70◦(+), and 70◦(−) roughly corre-
spond to the years of 1997, 1999,and 2000,respectively.
BESS(97)
However, it is shown in Fig. 2 that the actual level of -4 BESS(99)
overall modulation at the date of observation was lower 10 BESS(00)
(correspondingtoasmallervalueofα)in1999andhigher
(largerα)in2000. Thus,thepredictedeffectofthedrifts
in Fig. 2(a) might be hidden under the change in overall
modulationlevel. Thedriftmodelneedstobefine-tuned
o
i Bieber et al,
to reproducethe p¯andp spectra. The retardationofthe t
a Drift model
r
modulation effect due to the gradually expanding helio- P -5 10˚, (+)
spheric current sheet [28] might also be concerned. –P/10 7700˚˚,, ((+- ))
Figure3showsthep¯/pratiosin1999and2000inwhich
Fisk,
theabovementioneddifficultyispartlycanceled. Ingen-
Spherically
eral, p¯’s suffer less modulation than p’s because of their symmetric model
relatively hard IS spectrum. Thus, the spherically sym- f 99= (l6o4w8e Mr)V
metric model predicts an increase in p¯/p towards the so- f =1344 MV
00 (upper)
lar maximum period [29], mainly due to a decrease of -6
10
the p flux. The drift model predicts a much larger in- -1
10 1 10
crease in p¯/p reflecting the charge-sign dependence. It Kinetic Energy (GeV)
should be noted that a recent independent work [4] has FIG. 3. The p¯/p flux ratios measured by BESS in 1999
shown qualitatively the same feature in p¯/p while these and 2000 with the previous data in 1997. The curves are
two drift model calculations were different in IS spectra calculations of theratio at various solar activity [3,27].
3
102 theverylowenergyregionbelow0.5GeV,whereslightly
(a)
]P(IS) ≈≈≈ 011...309 GGGeeeVVV DSsyprmihfetm rmiectoardlilcey lmodel wexecreessoibvseerp¯vfleduxdeusrrineglattihveestoolatrhemtinhiemoruemtic[a6l,1c2a]l.culations
–P/ WethankNASAandNSBFfortheballoonexpedition,
][ / 10 and KEK and ISAS for continuous support. This work
U) 0.3 GeV was supported by Grant-in-Aid for Scientific Research,
A 1.0 GeV MEXT in Japan; and by NASA in the USA. Analysis
1
P( was performed using the computing facilities at ICEPP,
–P/ 1 1.9 GeV the University of Tokyo.
[
1993 1994 1995 1996 1997 1998 1999 2000 2001
ate Time (year) ∗ deceased.
nitor Count R34444802460000000000 (b) ClimCaoxu nNt eRuatrtoen ( NMCoLn)itor 456789000000gle (degree) †‡[[21]] RJPP.rr.ReeAss.ee.JnnoBttkuaaiprddgiddierrreeetessssta::la.UT,l.oA,nhAisvotsekrtrourspoitUhpyyhnsoyi.vfseJ.Tr.Jss2.iut45yk3,0u,J5ba1,ap1,21a4J5n4a.p((1a19n98.918))..
o 30n [3] J.W.Bieberetal.,Phys.Rev.Lett.83,674(1999);Proc.
eutron M334600100993 1994 1995T1il9t 9A6ng1l9e9 (7α)1998 1999 2000 200012100Tilt A [[45]] IM2.6.VthG.IManrtoc.siakC-aoMlsemnuknicoozRetaeaytl.Ca,lo.a,nstfP.ror(o-Upcth.a/h202)1n07d6,51I67n7t.(.19C9o9s)m.ic Ray
N Time (year)
Conf. (Dublin) 3, 497 (1991); A. Raviart et al., Proc.
FIG.4. (a)Temporalvariationofthep¯/pratios(relativeto 25th Int.Cosmic Ray Conf. (Durban) 2, 37 (1997).
theratioinIS)at0.3(closedsquares),1.0(opencircles),and [6] S. Orito et al.,Phys. Rev.Lett. 84, 1078 (2000).
1.9GeV(closedstars)measuredbyBESSsince1993. Curves [7] L. Bergstr¨om et al.,Astrophys.J. 526, 215 (1999).
represent the calculated temporal variation of the ratios (see [8] F. Donatoet al.,astro-ph/0103150.
text). (b) Temporal variation of NCL and α. In both plots, [9] the Wilcox Observatory at Stanford,
shaded area indicates the period of the solar field reversal. http://quake.stanford.edu/∼wso/Polar.ascii .
[10] J. Clem and Paul Evenson, Proc. 27th Int. Cosmic Ray
model (dotted) for 0.3, 1.0, and 1.9 GeV from top to
Conf. (Hamburg) 3934 (2001).
bottom. We obtainedthose φ’s atthe datesofthe BESS
[11] K. Yoshimura et al., Phys. Rev. Lett. 75, 3792 (1995);
flights (φBESS) and then we evaluated φ’s at any mo- A. Moiseev et al.,Astrophys.J. 474, 479 (1997).
ments using a linear relation between φBESS and NCL, [12] H. Matsunaga et al.,Phys. Rev.Lett. 81, 4052 (1998).
where NCL is the monthly averaged count rates of the [13] T. Maeno et al.,Astropart. Phys. 16, 121 (2001).
Climax Neutron Monitor [30]. The historical trend of [14] S.Orito,Proc.oftheASTROMAGWorkshop,KEKRe-
NCL is indicated by the solid histogram in Fig. 4(b). In port 87-19, 111 (1987).
Ref. [3], the p¯/p ratios are given as a function of the tilt [15] A.Yamamotoetal.,IEEETrans.Magn.24,1421(1988).
angle, α, rather than the actual time. In order to draw [16] A. Yamamoto et al., Adv.Space Res. 14(2), 75 (1994).
[17] Y. Ajima et al.,Nucl.Instr. Meth. A 443, 71 (2000).
the dashedcurveinFig.4(a),we takeαas the meanpo-
[18] Y. Asaoka et al.,Nucl. Instr. Meth. A 416, 236 (1998).
sition of the maximum latitudinal extent of the current
[19] Y. Shikazeet al., Nucl.Instr. Meth. A 455, 596 (2000).
sheet [31] taking no account of the retardation of the
[20] Y. Makida et al., IEEE Trans. Applied Supercon. 5(2),
modulation effect. Note that α is allowed to reach 90◦
638 (1995).
in the shaded area to avoid discontinuity in p¯/p at the
[21] Y.Asaokaetal.,Proc.27thInt.CosmicRayConf.(Ham-
fieldreversal. Thedriftmodelpredictsthatthep¯/pratio burg) 2139 (2001).
is stable during the positive phase and shows a sudden [22] T. Mitsui, Ph.D.thesis, University of Tokyo, 1996.
increaseafterthepositive-to-negativesolarfieldreversal. [23] S. A. Stephenset al., Astropart. Phys.6, 229 (1997).
Our measurements generally agree with the drift model. [24] Ch. Pfeifer et al.,Phys. Rev.C 54, 882 (1996).
As a conclusion, we have measured the temporal vari- [25] P. Papini et al.,NuovoCimento 19, 367 (1996).
ation of the p¯and p fluxes covering the solar minimum, [26] Y.Shikazeetal.,(inpreparation): Inthecalculationcode
of atmospheric protons, the recoil production spectrum
the maximum, and the solar magnetic field reversal. We
[25] was modified to reproducethe p fluxduringascent.
observed stable p¯/p ratios in the positive polarity phase
[27] L. A. Fisk, J. Geophys. Res. 76, 221 (1971).
through1999andasuddenincreasein2000followingthe
[28] J.A. le Roux et al.,Astrophys.J. 361, 275 (1990).
solar field reversal. The continuous observation has en-
[29] A.W. Labrador et al.,Astrophys.J. 480, 371 (1997).
abledacrucialtestofdrifteffectsinthesolarmodulation
[30] http://odysseus.uchicago.edu/NeutronMonitor/:
including a solar maximum period. It will motivate fur- University of Chicago, NSFGrant ATM-9613963.
ther development of models with more realistic parame- [31] Hoeksema,http://quake.stanford.edu/∼wso/Tilts.html .
ters [4] and time dependence [32]. Furthermore,detailed [32] R.A. Burger et al., Proc. 27th Int. Cosmic Ray Conf.
understanding of solar modulation is inevitably impor- (Hamburg), 3835 (2001).
tant for investigating the origin of p¯’s [33], especially in [33] A. Yamamoto et al., Adv.Space Res., in press.
4