Measurements of Cosmic-Ray Hydrogen and Helium Isotopes with the PAMELA experiment
O. Adriani, G. C. Barbarino, G. A. Bazilevskaya, R. Bellotti, M. Boezio, E. A. Bogomolov, M. Bongi, V. Bonvicini, S. Bottai, A. Bruno, F. Cafagna, D. Campana, P. Carlson, M. Casolino, G. Castellini, C. De Donato, C. De Santis, N. De Simone, V. Di Felice, V. Formato, A. M. Galper, A. V. Karelin, S. V. Koldashov, S. Koldobskiy, S. Y. Krutkov, A. N. Kvashnin, A. Leonov, V. Malakhov, L. Marcelli, M. Martucci, A. G. Mayorov, W. Menn, M Mergè, V. V. Mikhailov, E. Mocchiutti, A. Monaco, N. Mori, R. Munini, G. Osteria, F. Palma, B. Panico, P. Papini, M. Pearce, P. Picozza, M. Ricci, S. B. Ricciarini, R. Sarkar, V. Scotti, M. Simon, R. Sparvoli, P. Spillantini, Y. I. Stozhkov, A. Vacchi, E. Vannuccini, G. Vasilyev, S. A. Voronov, Y. T. Yurkin, G. Zampa, N. Zampa
aa r X i v : . [ a s t r o - ph . H E ] D ec Preprint typeset using L A TEX style emulateapj v. 5/2/11
MEASUREMENTS OF COSMIC-RAY HYDROGEN AND HELIUM ISOTOPES WITH THE
PAMELA
EXPERIMENT
O. Adriani , , G. C. Barbarino , , G. A. Bazilevskaya , R. Bellotti , , M. Boezio ,E. A. Bogomolov , M. Bongi , , V. Bonvicini , S. Bottai , A. Bruno , , F. Cafagna ,D. Campana , P. Carlson , M. Casolino , G. Castellini , C. De Donato ,C. De Santis , N. De Simone , V. Di Felice , V. Formato , † , A. M. Galper ,A. V. Karelin , S. V. Koldashov , S. Koldobskiy , S. Y. Krutkov ,A. N. Kvashnin , A. Leonov , V. Malakhov , L. Marcelli , M. Martucci , ,A. G. Mayorov , W. Menn , M Merg`e , , V. V. Mikhailov , E. Mocchiutti ,A. Monaco , , N. Mori , R. Munini , , G. Osteria , F. Palma , , B. Panico ,P. Papini , M. Pearce , P. Picozza , , M. Ricci , S. B. Ricciarini ,R. Sarkar , ∗ , V. Scotti , , M. Simon , R. Sparvoli , , P. Spillantini , ,Y. I. Stozhkov , A. Vacchi , , E. Vannuccini , G. Vasilyev , S. A. Voronov ,Y. T. Yurkin , G. Zampa , N. Zampa University of Florence, Department of Physics, I-50019 Sesto Fiorentino, Florence, Italy INFN, Sezione di Florence, I-50019 Sesto Fiorentino, Florence, Italy University of Naples “Federico II”, Department of Physics, I-80126 Naples, Italy INFN, Sezione di Naples, I-80126 Naples, Italy Lebedev Physical Institute, RU-119991, Moscow, Russia University of Bari, Department of Physics, I-70126 Bari, Italy INFN, Sezione di Bari, I-70126 Bari, Italy INFN, Sezione di Trieste, I-34149 Trieste, Italy Ioffe Physical Technical Institute, RU-194021 St. Petersburg, Russia INFN, Sezione di Rome “Tor Vergata”, I-00133 Rome, Italy University of Rome “Tor Vergata”, Department of Physics, I-00133 Rome, Italy National Research Nuclear University MEPhI, RU-115409 Moscow KTH, Department of Physics, and the Oskar Klein Centre for Cosmoparticle Physics, AlbaNova University Centre, SE-10691Stockholm, Sweden IFAC, I-50019 Sesto Fiorentino, Florence, Italy INFN, Laboratori Nazionali di Frascati, Via Enrico Fermi 40, I-00044 Frascati, Italy Universit¨at Siegen, Department of Physics, D-57068 Siegen, Germany University of Trieste, Department of Physics, I-34147 Trieste, Italy Indian Centre for Space Physics, 43 Chalantika, Garia Station Road, Kolkata 700084, West Bengal, India University of Udine, Department of Mathematics and Informatics, I-33100 Udine, Italy † Now at INFN, Sezione di Perugia, I-06123 Perugia, Italy and ∗ Previously at INFN, Sezione di Trieste, I-34149 Trieste, Italy
ABSTRACTThe cosmic-ray hydrogen and helium ( H, H, He, He) isotopic composition has been measuredwith the satellite-borne experiment
PAMELA , which was launched into low-Earth orbit on-board theResurs-DK1 satellite on June 15 th H and He in cosmic rays are believed tooriginate mainly from the interaction of high energy protons and helium with the galactic interstellarmedium. The isotopic composition was measured between 100 and 1100 MeV/n for hydrogen andbetween 100 and 1400 MeV/n for helium isotopes using two different detector systems over the 23 rd solar minimum from July 2006 to December 2007. Subject headings:
Astroparticle physics, cosmic rays INTRODUCTION
The rare isotopes H and He in cosmic rays are generally believed to be of secondary origin, resulting mainly fromthe nuclear interactions of primary cosmic-ray protons and He with the interstellar medium. The spectral shapeand composition of the secondary isotopes is therefore completely determined by the source spectrum of the parentelements and by the propagation process. Measurements of the secondary isotopes spectra are then a powerful tool toconstrain the parameters of the galactic propagation models (Strong et al. et al. et al. et al. et al. et al. et al. et al. et al. et al.
Neutrondetector Anti-coincidence m a g n e t B XZ Y
Proton AntiprotonScintill. S4CalorimeterTrackingsystem(6 planes) GeometricacceptanceCATCASCARDSpectrometerTOF (S1)TOF (S2)TOF (S3)
Fig. 1.—
Scheme of the detectors composing the
PAMELA satellite experiment. isotope production cross sections).The results presented in this paper are based on data gathered between July 2006 and December 2007 with the
PAMELA satellite experiment.
PAMELA has been put in a polar elliptical orbit at an altitude between ∼
350 and ∼
600 km with an inclination of 70 ◦ as part of the Russian Resurs-DK1 spacecraft. Due to the low-earth orbit themeasurements are performed in an environment free from the background induced by interactions of cosmic rays withinthe atmosphere. The month of December 2006 was discarded from the dataset to avoid possible biases from the solarparticle events that took place during the 13 th and 14 th of December. During a total acquisition time of 528 daysabout 10 triggered events were recorded, 5 . · hydrogen nuclei were selected in the energy interval between 100and 1100 MeV/n and 1 . · helium nuclei between 100 and 1400 MeV/n.This is the second paper on isotopes from the PAMELA instrument. The first paper (Adriani et al. H, H, He, He but used solely the combination of velocity measurement provided by thetime-of-flight system in combination with the momentum measurement of the magnetic spectrometer. In this workhere we employed a more advanced and a more comprehensive analysis. In detail: a more complete and elaboratedfitting procedure was employed, combined with a more stringent selection on cuts and on efficiencies. In addition wemade use of the multiple energy loss measurements provided by the 44 planes of the imaging calorimeter. This notonly allowed a cross check between the two techniques of isotopic separation within the
PAMELA instrument (ToF andmultiple dE/dx versus rigidity), the multiple dE/dx technique also allowed to extend the measurements for isotopesto higher energies: For hydrogen isotopes the highest energy bin is now at 1035 MeV/n (instead of 535 MeV/n), whilefor helium it is now at 1297 MeV/n (instead of 823 MeV/n). Additionally the previous results (Adriani et al. THE
PAMELA
APPARATUS
A schematic view of the
PAMELA detector system (Picozza et al. Z = 1 particles and a high lepton/hadron discrimination power.The instrument core is a permanent magnet with a silicon microstrip tracker. The design of the permanent magnetprovides an almost uniform magnetic field of 0.45 T inside the magnetic cavity. Six layers of 300 µ m thick double-sidedmicrostrip silicon detectors are used to measure particle deflection with ∼ µ m and ∼ µ m precision (measuredwith beam tests and flight data) in the bending and non-bending views, respectively. Due to the small size the amountof material inside the magnetic cavity can be kept to a minimum, only the six layers of silicon without any need forsupport structure. The MDR (Maximum Detectable Rigidity) of the magnetic spectrometer is about 1 TV.The Time-of-Flight (ToF) system comprises six layers of fast plastic scintillators arranged in three planes (S1, S2 andS3). Each detector layer is segmented into strips, placed in alternate layers orthogonal to each other. Using differentcombinations of layers, the ToF system can provide 12 measurements of the particle velocity, β = v/c , by using aweighted mean technique an overall value for β is calculated from these measurements . The overall time resolutionof the ToF system is about 250 ps for Z = 1 particles and about 100 ps for Z = 2 particles. This allows albedoparticles crossing PAMELA from bottom to top to be discarded by requiring a positive β . The TOF scintillators canalso identify the absolute particle charge up to oxygen by means of the six independent ionization measurements.The silicon-tungsten electromagnetic sampling calorimeter comprises 44 single-sided silicon planes interleaved with 22plates of tungsten absorbers. The calorimeter is mounted below the spectrometer, its primary use is for lepton/hadronseparation (Boezio et al. µ m thick, 8x8 cm detectors, segmented into 32 strips with a pitchof 2.4 mm. The orientation of the strips for two consecutive silicon planes is shifted by 90 degrees, thus providing 2-dimensional spatial information. The total depth of the calorimeter is 16.3 radiation lengths and 0.6 nuclear interactionlengths. Below the calorimeter there is a shower tail catcher scintillator (S4) and a neutron detector which help toincrease hadron/lepton separation. The tracking system and the upper ToF system are shielded by an anticoincidencesystem (AC) made of plastic scintillators and arranged in three sections (CARD, CAT, and CAS in Fig. 1), which allowsto detect during offline data analysis the presence of secondary particles generating a false trigger or the signature of aprimary particle suffering an inelastic interaction. The total weight of PAMELA is 470 kg and the power consumptionis 355 W. A more detailed description of the instrument can be found in Picozza et al. (2007). DATA ANALYSIS
Event selection
Each triggered event had to fulfill several criteria to be used for further analysis. The requirements are identical tothe selection in Adriani et al. (2013b) and we refer to that paper for more details: • Event quality selections: We have selected events that do not produce secondary particles by requiring a singletrack fitted within the spectrometer fiducial acceptance and a maximum of one paddle hit in the two topscintillators of the ToF system. The analysis procedure was similar to previous work on high energy proton andhelium fluxes (Adriani et al. • Galactic particle selection: Galactic events were selected by imposing that the lower edge of the rigidity bin towhich the event belongs exceeds the critical rigidity, ρ c , defined as 1.3 times the cutoff rigidity ρ SV C computedin the St¨ormer vertical approximation (Shea et al. ρ SV C = 14 . /L , where L is the McIlwain L-shell parameter obtained by using the Resurs-DK1 orbital information and the IGRF magnetic field model(MacMillan & Maus 2005). • Charge selection: The charge identification uses the ionization measurements provided by the silicon sensors ofthe magnetic spectrometer. Depending on the number of hit sensors there can be up to 12 dE/dx measurements,the arithmetic mean of those measurements is shown in Fig. 2 as a function of the rigidity. The actual selection of Z = 1 or Z = 2 particles is depicted by the solid lines. A similar figure has been already shown in Adriani et al. (2013b) but was kept in this paper to help the reader. Isotope separation in the
PAMELA instrument
In each sample of Z = 1 and Z = 2 particles an isotopic separation at fixed rigidity is possible since the mass ofeach particle follows the relation m = RZeγβc (1)where R is the magnetic rigidity, Ze is the particle charge, and γ stands for the Lorentz factor. The particle velocity β can either be provided directly from the timing measurement of the ToF system, or indirectly from the energy lossin the calorimeter, which follows β via the Bethe-Bloch formula dE/dx ∝ Z β (neglecting logarithmic terms). Isotope separation using ToF vs. rigidity
For the ToF analysis we can use directly the β provided by the timing measurement. In Fig. 3 we show β vs. theparticle rigidity for Z = 1 and Z = 2 data. The black lines in the figure represent the expectations for each isotope.A similar figure has been already shown in Adriani et al. (2013b) but was kept in this paper for an easier comparisonwith the alternative identification method using the calorimeter.Since the particle mass is calculated using Eq. 1, misidentified He wrongly reconstructed as Z = 1 particles couldresult in a significant contamination to the H sample. However, the amount of misidentified helium was found to benegligible, see Adriani et al. (2013b) for more details.
Isotope separation using multiple dE/dx in the calorimeter vs. rigidity
The isotopic analysis of nuclei with the calorimeter is restricted to events which do not interact inside the calorimeter.To check if an interaction occurs, we derived in each silicon layer a) the total energy detected ( q tot ) and b) the energy Rigidity (GV)0.1 0.2 0.3 0.4 1 2 3 4 d E / d x ( M I P ) H He He Fig. 2.—
Ionization loss ( dE/dx in MIP, energy loss of minimum ionizing particles) in the silicon detectors of the tracking system as afunction of reconstructed rigidity. The black lines represent the selection for Z = 1 or Z = 2 nuclei. deposited in the strip closest to the track and the neighbouring strip on each side ( q track ). In the ideal case thefraction of q track /q tot will be equal to one, a value less then one means that strips outside the track were hit. Startingfrom the top of the calorimeter we calculated Σ q track / Σ q tot at each layer, as long as this value is greater than 0.9,we used these layers for further analysis. A value of 0.9 was chosen since it was found to give a good compromisebetween high efficiency and rejection of interactions. In this way we can make use of slow particles, which stop earlyin the calorimeter, particles which interact somewhere, but also all clean events with the particle fully traversing thecalorimeter. In the single silicon layer the energy loss distributions shows a Landau tail which degrades the resolutionof the dE/dx measurement. Using a truncation method, the 50% of samples with larger pulse amplitudes were excludedbefore taking the mean of the dE/dx measurements, thus reducing the effect of the Landau tail. We put an energydependent lower limit on the number of layers after the 50% truncation, requiring at least 5 measurements at 1 GV,going up to 10 layers at 3 GV. With this requirement the lower energy limit of our analysis is around 200 - 300 MeV/n(the energy to fully penetrate the calorimeter is much higher, about 400 - 500 MeV/n).In Fig. 4 we show the mean dE/dx for each event vs. the rigidity measured with the magnetic spectrometer for Z = 1and Z = 2 particles. The energy loss in MeV was calculated from the measurement in MIP using a conversion factor(the most probable energy loss of a minimum ionizing particle traversing 380 µ m of silicon, derived by simulation). Inboth plots the isotopic separation is clearly visible. It is worthwhile to mention that the selection procedure describedabove can be done in different ways but we found that we achieved the best results particularly on the correspondingefficiency (see section 3.4) by the method described above.The use of the calorimeter for isotope separation provides us with another advantage. Contrary to the β vs rigiditytechnique with the ToF, by using the multiple dE/dx vs. rigidity technique a misidentified He wrongly reconstructedas Z = 1 particles would not result in a contamination to the H sample, since the dE/dx is different. This appliesalso for the Z = 1 contamination and contamination by heavier nuclei in the Z = 2 sample. This allows a valuablecrosscheck between the results from ToF and calorimeter in the energy regime where the two measurements overlap. Measured Mass resolution in the
PAMELA instrument and comparison with expectations
Theoretically, there are three independent contributions to the mass resolution in a magnet spectrometer similar to
PAMELA : the bending power of the magnetic spectrometer coupled with the intrinsic limits of spatial resolution whichthe tracking detectors provide, the precision of the velocity measurement (given either by timing or by measuring theenergy loss), and the multiple scattering of the particle along its path in the bending area of the magnet. These three
Rigidity (GV)0 0.5 1 1.5 2 2.5 3 β H H Z=1
Rigidity (GV)0 0.5 1 1.5 2 2.5 3 β He He Z=2
Fig. 3.— β vs. rigidity for Z = 1 (top) and Z = 2 (bottom) particles. The black lines were calculated for each isotope using Eq. 1. independent contributions can be expressed by the following equation: dm = m s γ (cid:18) dββ (cid:19) + (cid:18) RM DR spec (cid:19) + (cid:18) RM DR cou (cid:19) (2)where γ is the Lorentz factor, dβ/β is the relative error in the velocity measurement, R/M DR spec stands for thecontribution solely given by the magnetic spectrometer, and the last term stands for multiple coulomb scattering. Thelast two terms of Eq. 2 can be expressed by an overall
M DR tot : Rigidity (GV)0 0.5 1 1.5 2 2.5 3 3.5 4 T r un ca t e d M ea n ( M e V ) H Z=1
Rigidity (GV)0 0.5 1 1.5 2 2.5 3 3.5 4 T r un ca t e d M ea n ( M e V ) He Z=2
Fig. 4.—
Mass separation for Z = 1 (top) and Z = 2 (bottom) particles using the “truncated mean”-method. (cid:18) M DR tot (cid:19) = (cid:18) M DR spec (cid:19) + (cid:18) M DR cou (cid:19) (3)For PAMELA the overall momentum resolution of the magnetic spectrometer (thus
M DR tot ) has been measured inbeam tests at CERN (Picozza et al.
M DR spec of the
PAMELA spectrometer has a value of about1 TV for Z = 1 particles. The contribution of M DR spec to the overall mass resolution is therefore negligible for the
Mass (amu)2 2.5 3 3.5 4 4.5 5 5.5 602004006008001000 m ToF: 0.42 amu ∆ He Mass (amu)2 2.5 3 3.5 4 4.5 5 5.5 6050100150200250300 m Calo.: 0.30 amu ∆ He Fig. 5.—
Examples mass distributions for helium in the 2.5 - 2.6 GV rigidity range for ToF (left) and Calorimeter (right). particles we are analyzing (up to some GV). At low rigidities the contribution from multiple scattering is however thedominant effect. Its value is inverse proportional to the bending power of the magnet ( R B · dl ) and direct proportionalto the amount of matter traversed along the bending part of the track. For an experiment using a permanent magnet incombination with silicon strip detectors PAMELA shows a very good momentum resolution, due to its strong magneticfield of 0.45 T combined with a low amount of material (only the six silicon detectors, each 300 µ m, giving a grammageof 0.42 g/cm ) in the magnetic cavity. The overall momentum resolution was measured at CERN to have a minimumof about 3.5 % at 8 GV, increasing to 5% at 1 GV (Picozza et al. M DR cou .By using the β -rigidity technique (see Fig. 3) Eq. 1 directly provides the mass of the particle and a correspondingmass (amu) histogram for helium in the rigidity range from 2.5 to 2.6 GV is shown in Fig. 5 (left). When usingthe multiple dE/dx versus rigidity technique (see Fig. 4) either a mass of three or four amu was allocated to thecorresponding peaks in the histogram. By scaling between these positions linearly one obtained the mass distributionwhich we show also for helium and for the same rigidity range of 2.5 to 2.6 GV in Fig. 5 (right). We fitted a gaussianto the He peak and used the standard deviation of the gaussian as the mass resolution, in this example 0.42 amuwith the ToF, while it is 0.30 amu for the calorimeter. Thus the use of the multiple dE/dx in the calorimeter stackprovided a better mass resolution than the direct measurement of β with the ToF for this rigidity interval.We repeated this procedure for a number of rigidity intervals and derived the mass resolutions for Helium obtainedwith ToF and calorimeter as a function of the rigidity, which is shown in Fig. 6, compared with the predictions asobtained and derived from the CERN tests. The full black line illustrates the predicted overall mass resolution for a Heparticle resulting from three contributions (shown as dotted lines): rigidity (
M DR spec ), multiple scatter (
M DR cou ),and velocity via ToF (time resolution 100 ps). As one can see the experimental results on the mass resolution obtainedfrom the combination of rigidity and velocity from the ToF system follow nicely the prediction. This gives us confidencethat we understand our instrument. It can also be seen that the multiple dE/dx from the calorimeter provide a bettermass resolution at higher energies than measuring the velocity via the ToF. This allows us to extend the
PAMELA measurements on isotopes to higher energies. It can also be clearly seen that the multiple scatter sets the lower limitof the mass resolution at rigidities below roughly 2 GV no matter how high the MDR of the spectrometer is.
Raw isotope numbers
We had two complementing experimental methods to separate the isotopes: The combination of magnetic spectrom-eter either with the ToF or with the multiple dE/dx measurements within the calorimeter. In the following we willdescribe these procedures separately.The isotope separation as well as the determination of isotope fluxes was performed identical to Adriani et al. (2013b) in intervals of kinetic energy per nucleon. Since the magnetic spectrometer measures the rigidity of particlesand not the kinetic energy, this means that different rigidity intervals have to be analyzed depending on the massof the isotope under study . For example Fig. 7 shows the 1/ β distributions used to select H (top panel) and H(bottom panel) in the kinetic energy interval 0.361 - 0.395 GeV/n corresponding to 0.90 - 0.95 GV for H and 1.80 -1.89 GV for H. Raw isotope numbers with the ToF
The particle counts in each rigidity range were derived in a similar manner as in Adriani et al. (2013b) by fittinggaussians to the 1/ β distributions as shown by the solid lines in Fig. 7. Instead of mass distributions (shown in Fig. 5)1/ β distributions were chosen since the shape of a 1/ β distribution is gaussian, while the mass distribution is not.(Note that for the estimation of the mass resolution in section 3.2.3 this feature could be neglected).The H peak in Fig. 7 (top) is well pronounced and barely affected by the shape of the neighbouring H distribution.
Rigidity (GV)1 1.5 2 2.5 3 3.5 M ass R es o l u t i on ( a m u ) spec MDR coul
MDRToF 100 psSpectrometer + ToFFlight Data:
Fig. 6.—
Measured He mass resolution for the ToF (circles) and the calorimeter using the “truncated mean” method (squares). Ifthe error bars are not visible, they lie inside the data points. The dashed lines show the calculated independent contributions (rigidity(
MDR spec ), multiple scatter (
MDR cou ), and velocity via ToF (time resolution 100 ps)), while the solid lines shows the overall massresolution for this combination.
For that reason a single gaussian was fitted to the H peak as shown in Fig. 7 (top). For the fitting method theROOT analysis package (Brun & Rademakers 1997) was used. The gaussian fit to the H, He and He distributionsbecomes a little more complicated since the neighbouring isotopes are quite abundant and have an impact on thefitting. For that reason we applied a suppression procedure to the abundant neighbour which will be more discussed inthe following chapter. Consequently we applied a double gaussian fit to the histograms, see Fig. 7 (bottom) and Fig. 8,and the whole process was done in a more elaborated manner compared to the analysis presented in Adriani et al. (2013b). We went through three steps: We first did the double gaussian fitting with all six parameters left free (mean,sigma and peak of both isotopes) and then analyzed how the values for the means and the sigmas varied with kineticenergy. They followed a trend in kinetic energy and in a second step we fitted appropriate functions sigma=f( E k in )and mean=f( E k in ) to them. All the means and the sigmas followed nicely this trend except at the high energy end,which may be due to the increasing contribution from the more abundant neighbouring isotopes. As a consequence wedecided to perform the final fitting to the distributions by fixing the mean and the sigmas according to the functionand used in the final double gaussian fitting process only two free parameters: the two heights of the curve’s peaks.Fig. 8 shows the 1/ β distributions used to select He (bottom panel) and He (top panel) in the kinetic energyinterval 0.439 - 0.492 GeV/n corresponding to 1.51 - 1.62 GV for He and 2.01 - 2.15 GV for He.
Raw isotope numbers with the ToF: Suppression of abundant H and He As it was already discussed in Adriani et al. (2013b), the large proton background in the Z = 1 sample requires anadditional selection in the ToF analysis to suppress the protons at higher energies (roughly 500 MeV/n). Otherwisethe gaussian fit for the protons in the 1 /β distributions affects the fit for the much less abundant H neighbour,especially at higher energies, where the mass resolution is not sufficient for a clear particle separation. To suppressthe abundance of protons, we choose the energy loss measurements in the silicon layers of the tracking system and inthe scintillators of the ToF versus the rigidity. The tracking system provides up to 12 energy loss measurements whilethe ToF provides six. In order to further improve the separation between the isotopes we did not take the mean ofthe dE/dx measurements but choose the lowest one. This minimizes the Landau fluctuations similar to the truncatedmean technique which we use in the calorimeter analysis. The two cuts were chosen in such a way that for low energiesthe protons could be rejected down to a very low level, by keeping practically all H. For more details of the specificcuts see Fig. 6 in Adriani et al. (2013b). β C oun t s
10 H H β C oun t s
10 H H H Fig. 7.— /β distributions for hydrogen in the 0.361 - 0.395 GeV/n kinetic energy range for H (top) and H (bottom). The dashedline shows the combined fit (only for H) while the solid line shows the H and H individual gaussians. Note that the H component inthe H distribution in the bottom plot is suppressed by the additional selection cuts on the energy loss in ToF and tracker. In the bottomfigure the small fraction of H events is visible. β C oun t s He He β C oun t s He Fig. 8.— /β distributions for helium in the 0.439 - 0.492 GeV/n kinetic energy range for He (top) and He (bottom). The dashed lineshows the combined fit while the solid line shows the He and He individual gaussians.
At higher energies the proton contamination will increase plus there will be a loss of H. This efficiency was studied1with the clean H sample provided by the calorimeter, and was taken into account in the calculation of the fluxes(section 3.4).In the Z = 2 data the level of the He background in the He sample is much smaller compared to the Z = 1 data,but similar checks like the one described above showed that also in this case a soft cut to suppress He at higherenergies improved the He selection. We used a cut analogue to the Z = 1 analysis based on the lowest energy releasein the tracking system. Note that this suppression was not used for the Z = 2 analysis in Adriani et al. (2013b). Raw isotope numbers with the calorimeter
For the ToF system the 1/ β distributions were analyzed by fitting with gaussians. The dE/dx distributions of thecalorimeter have a non-gaussian shape, hence one has to model the expected distributions of the observable quantitiesand then perform likelihood fits. We used the “RooFit” toolkit (Verkerke & Kirkby 2003) for the likelihood fits. Firstone has to create the expected dE/dx distributions (“probability density function”: PDF) for each isotope. We usedthe full Monte Carlo simulation of the PAMELA apparatus based on the
GEANT4 code (Agostinelli et al. et al. (2013b), for this task.When taking the simulated energy loss in each layer as coming from
GEANT4 , we noticed that the resulting PDFsshowed a slight mismatch from the flight data. We found that the width of the histograms was smaller than in thereal data, also there was a small offset of about 1 - 2%. We applied a multiplicative factor to the simulated energyloss in a layer, plus adding a gaussian spread of the signal of a few percent.As an example we show in Fig. 9 the truncated mean distributions for helium in the 0.439 - 0.492 GeV/n kineticenergy range for He (top) and He (bottom). The dashed line shows how the combined fit using the two PDFs derivedwith the modified
GEANT4 simulation matches the data points (black points) while the solid line shows the estimatedindividual He and He signals. The kinetic energy range is the same as shown in Fig. 8 for the ToF, the difference inthe isotopic separation is clearly visible.Due to the redundant detectors of
PAMELA we were able to test the simulated PDFs with real data from theinstrument. Fig. 2 illustrated the mass resolution which can be obtained by combining the mean dE/dx measurementin the tracker and the rigidity measurement with the magnetic spectrometer, while Fig. 3 showed a similar pictureusing the velocity measurement from the ToF and the rigidity measurement.By using appropriate selection cuts we separated a proton sample and asked for the energy loss response in thecalorimeter. This result was then compared to the simulated distribution. Such a comparison is shown in Fig. 10 forthe for the rigidity interval 1.80 GV - 1.89 GV. (The rigidity interval in this figure is the same as in Fig. 7 for the ToF,also here the difference in the isotopic separation is clearly visible).It can clearly be seen that the flight data proton PDF shows a larger tail into the H histogram compared to thesimulated PDF. This tail will cause to derive a lower number of H counts compared to the simulated PDF (thedifference in the H counts is about 8% in this example).While the creation of PDFs from flight data for H, He and He is restricted to lower energies due to the limitedisotopic separation, we created PDFs for protons up to 4 GV using strict selection cuts. In principle a clear separationbetween H and H is not possible at these rigidities, but since the H are so dominant, the contamination of H in the H sample should be very small. We observed that the tail in Fig. 10 is only visible for medium energies, where ourselection cuts for the truncated mean are quite soft. Since a small number of layers are used to derive the truncatedmean, fluctuations are probably still significant, and it seems that the
GEANT4 simulation cannot fully reproduce theactual energy loss under these circumstances.At higher energies, where our selection cuts for the calorimeter are stricter, the tails in the flight data PDF disap-peared, resulting in a good agreement with the simulated PDF again. We decided to take simulated PDFs except forthe H model for the “RooFit” analysis in this paper.One could argue that also for the other isotopes it might be that the simulated PDFs do not show the correct shape,missing the tail which is visible for H, so for example we might underestimate the number of He which contributeto the He distribution. However, we found that the maximum difference in the H counts at medium energies was atmost 10%, with the number of H exceeding the number of H in the distribution by a factor of 50 - 100. In comparisonthe He / He ratio is around 0.2, so one can expect that the effect of missing tails in the simulated PDFs will have anegligible influence to the He and He counts.
Flux Determination
To derive the isotope fluxes, the number of H and H events in the Z = 1 sample and the number of He and Heevents in the Z = 2 sample had to be corrected for the selections efficiencies, particle losses, contamination and energylosses. Most of the corrections could be directly taken from Adriani et al. (2013b), only some efficiencies in the ToFanalysis were changed, for example the efficiency for the suppression of the abundant H and He (see section 3.3.2).A new correction is the efficiency for the calorimeter, which is shown in Fig. 11 for specific selection cuts. The selectioncuts for the actual analysis were described in section 3.2.2. One can nicely see that this approach gives a quite highefficiency showing a rather constant behaviour down to about 200 MeV/n where the efficiency then shows a steepdecrease. As mentioned already above one can do these selections within the calorimeter analysis in various ways, butthe efficiencies are very sensitive to the applied cuts. This is illustrated in Fig. 11 for two other cut conditions: A strictcut where the particle has to fully traverse the calorimeter and produce a signal in the last layer, and a more relaxedcut (30 hit silicon strips). As one can see, the efficiencies of these cuts are quite low and show a steep drop already2
Truncated Mean (MeV)0.5 0.6 0.7 0.8 0.9 1 C oun t s He Truncated Mean (MeV) C oun t s He Fig. 9.—
Example of the truncated mean distributions for helium in the 0.439 - 0.492 GeV/n kinetic energy range for He (top) and He (bottom). The dashed line shows how the combined fit using the two PDFs derived with the modified
GEANT4 simulation matches thedata points (black points), while the solid line shows the estimated He and He individual signals. at 300 - 400 MeV/n. Preliminary results for the relaxed cut have been presented in earlier publications (Menn et al. Truncated Mean (MeV)0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 C oun t s H H Truncated Mean (MeV)0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 C oun t s H H Fig. 10.—
Example RooFit of H. Black points: data, dashed line: H model, solid line: H model. The H model is taken from thesimulation for both plots, while in the left plot the H model is derived from flight data and from simulation in the right plot.
E (GeV/n)0.1 1 E ff i c i e n cy He Last LayerFixed CutsThis Work
E (GeV/n)0.1 1 E ff i c i e n cy He Last LayerFixed CutsThis Work
Fig. 11.—
Calorimeter selection efficiency for helium derived with simulated data: triangles: dE/dx signal in the last layer, squares: fixedcuts (30 hit silicon strips), circles: dynamic cuts used for this work. (2013a), Menn et al. (2013b)).The comparison between efficiencies derived with simulated data and the ones derived with flight data (using ToFand tracker dE/dx for selecetion) is shown for Z = 2 particles in Fig. 12.As one can see, there is a good agreement at low energies for He while there is some difference at higher energies,while it is the opposite for He. We decided to use the flight data effiencies at lower energies and then for higherenergies (roughly around 700 MeV/n in Fig. 12), follow the trend of the simulated efficiencies by applying a constantcorrection factor. The same method was used for H efficiency, while for H we used the flight data efficiency for thefull energy range.The following corrections are taken from Adriani et al. (2013b) without changes and we refer to this paper for moredetails: • Due to hadronic interactions in the aluminum pressurized container (2 mm thick) and the top scintillators heliumand hydrogen nuclei might be lost. The correction factor b ( E ) is different for each isotope and has been derivedfrom the Monte Carlo simulation, being ≃
6% for H, ≃
10% for H, and ≃
13% for both helium isotopes,. • The nominal geometrical factor G F of PAMELA is almost constant above 1 GV, with the requirements on thefiducial volume corresponding to a value of G F = 19 . sr, for lower energies the bending of the particlestrack leads to a decrease. The nominal geometrical factor G F was multiplied with the correction factor b ( E ) toget an effective geometrical factor G ( E ) see Fig. 7 in Adriani et al. (2013b).4 E (GeV/n)0.1 1 E ff i c i e n cy He SimulationFlightUsed for flux calc.SimulationFlightUsed for flux calc. E (GeV/n)0.1 1 E ff i c i e n cy He SimulationFlightUsed for flux calc.SimulationFlightUsed for flux calc.
Fig. 12.—
Comparison between the efficiency derived with simulated data (circles) and flight data (squares). The solid line shows theefficiency used for the analysis. • Contribution to H from inelastic scattering of He: This background was derived from the simulation andsubtracted from the raw H counts. The contamination is in the order of ≃
10% at 100 MeV/n, going downwith increasing enery ( ≃
1% at 600 MeV/n), see Fig. 8 in Adriani et al. (2013b). The contamination in the He sample from He fragmentation was also evaluated and was found to be very small (less than 1%), this wasincluded in the systematic uncertainty of the measurement. • The measured particle spectra are distorted due to particle slowdown (caused by the energy loss) and the finiteresolution of the spectrometer. We used a Bayesian unfolding procedure (D’Agostini 1995) to derive the numberof events at the top of the payload (see Adriani et al. (2011)).The differential flux is then given by Φ
ToP ( E ) = N ToP ( E ) T G ( E )∆ E (4)where N ToP ( E ) is the unfolded particle count (corrected for the selection efficiencies) for energy E , , ∆ E is the energybin width, and G ( E ) is the effective geometrical factor as described above. The live time, T , depends on the orbitalselection as described in section 3.1 and is evaluated by the trigger system (Bruno 2008). Systematic uncertainties
The systematic uncertainties presented in Adriani et al. (2013b) have been reviewed and updated to the new analysismethods when neccessary.The event selection criteria described in section 3.1 were similar to previous work on high energy proton and heliumfluxes, see Adriani et al. (2011). In that paper the systematic errors of the selection have been studied using flightdata and simulations, resulting in a quoted systematic uncertainty of ca. 4%. This error is used also in this work.As already decribed in section 3.3.2, the quality of the Gaussian fit procedure in the ToF analysis was tested usingthe truncated mean of the energy deposited in the electromagnetic calorimeter to select pure samples of H, H, He,and He from non-interacting events. For the abundant particles H and He the number of reconstructed events fromthe Gaussian fit was found to agree with the number of events selected with the calorimeter practically over the fullenergy range, while for H and He there were some systematic differences of some percent in the highest energy bins.Note that without the additional cuts which reject the more abundant H and He the differences would be muchlarger. We assigned a systematic uncertainty of 0 .
5% for low and medium energies increasing to 4% at 600 MeV/n for H and to 3% at 800 MeV/n for He. For H the systematic uncertainty was set constant to 0 .
5% while for He it wasset energy dependent, increasing to 1 .
5% at 800 MeV/n.A similar systematic error is assigned to the fit procedure made with the calorimeter. Here deviations betweenthe model PDF and the flight data will transform to a systematic difference in the number of reconstructed events.However, we have no other detector to select pure samples of the isotopes. Therefore we studied how a misplacementof the model PDFs transferred to different particle counts. Similar to the results for the ToF it was found that theeffect on the number of reconstructed events was much more pronounced for H and He compared to H and He,and that the effect increased with energy. However, the misplacement of the model PDFs can be checked by usingthe abundant H and He as a reference (for example, comparing the peaks in the distributions), thus limiting thesystematic differences in the number of reconstructed events to some percent in the highest energy bins. Similar to thesystematic error for the ToF, we assigned a systematic uncertainty of 0 .
5% for low and medium energies increasing to54% at 1000 MeV/n for H and to 3% at 1400 MeV/n for He. The H the systematic uncertainty was set constant to0 .
5% while for He it was set again energy dependent, increasing to 1 .
5% at 1400 MeV/n.The efficiency of the calorimeter selection was derived using simulated and flight data, as shown for Z = 2 particlesin Fig. 12. For Z = 2 data the agreement between the two methods is quite good, and we assigned a conservativesystematic error of 2% for He and 3% for He independent from the energy. For Z = 1 particles the differencebetween the two methods is larger. As stated above, for H we used the flight data efficiency for the full energy range,which should result in a small systematic error, since the H are so abundant and therefore the contamination of otherparticles is negligible. We assigned a conservative systematic error of 2%. For the H efficiency we used the flightdata efficiencies at lower energies, but followed the trend of the simulated efficiencies by applying a constant correctionfactor for higher energies, we estimated a systematic error of 5%.The following systematic uncertainties are taken from Adriani et al. (2013b) without changes and we refer to thispaper for more details: • The systematic uncertainty on the H flux resulting from the subtraction of secondary H from He spallationis 1 .
9% at low energy dropping below 0 .
1% at 300 MeV/n due to the finite size of the Monte Carlo sample. Thevalidity of the Monte Carlo simulation has been tested in Adriani et al. (2013b) using the H component in theflight data sample, see Fig. 7. • The systematic uncertainty on the unfolding procedure has been discussed in Adriani et al. (2011) and wasfound to be 2%, independent of energy. • The selection of galactic particles was described in section 3.1, the correction for particles lost due to this selectionhas an uncertainty due to the size of the Monte Carlo sample. The systematic error decreases from 6% at 120MeV/n to 0 .
06% at 1000 MeV/n. • The uncertainty on the effective geometrical factor as estimated from the Monte Carlo simulation is 0 . RESULTS AND DISCUSSION
In Figure 13 and 14 we show the hydrogen and helium isotope fluxes (top) and the ratios of the fluxes (bottom)measured with the ToF or the calorimeter. Results are also reported in Tables 1, 2, 3, and 4.It is worth noting that the
PAMELA results obtained via the ToF analysis and via the multiple dE/dx measurementswith the calorimeter agree very well within their systematic errors. This gives confidence to the results.In direct comparison with our first paper (Adriani et al. H fluxes show only minor differences, the H flux isroughly about 5% higher in this work. The He flux is almost 10% higher at the lowest energy bin, at the highestenergies the new He flux is about 10% lower, while at medium energies around 400 MeV/n the two results agree. Thenew He flux is about 3-4% lower for most of the energy range, for energies above 500 MeV/n the difference increasesand reaches about 15% for the highest energy bins. We attribute this to the changes in the fitting procedure (forexample, the double gaussian fit for He, also the fixing of parameters) and improvements in the efficiency calculationcompared to the first paper. Based on this more comprehensive analysis presented here these results supersede theprevious ones.To compare our isotope fluxes with other measurements, we decided to use at low energies only the ToF results (upto 361 MeV/n for hydrogen and up to 350 MeV/n for helium) and above these values only use the calorimeter results.In Figure 15 and 16 we show these hydrogen and helium isotope fluxes (top) and the ratios of the fluxes (bottom),compared to previous measurements (Aguilar et al. et al. et al. et al. et al. et al. et al. et al. et al. H/ He ratio as a function of kinetic energy per nucleon.It is visible that the former results show a large spread and it is obvious that the
PAMELA results are more precisein terms of statistics. In this context it is important to know that all the former measurements shown in Figures 15,16, and 17, except AMS-01, are from balloon-borne experiments and thus effected by the non-negligible backgroundof atmospheric secondary particle production.The scientific interest in these isotopes of H, H, He and He are determined by the question about their origin.It is believed that the protons and the He particles are predominantly of primary origin thus arise directly fromtheir sources while H and He are of secondary origin thus are produced by interactions of these primaries withthe interstellar gas. The interpretation of these results then allows to study more in detail the conditions of theirpropagation in the interstellar space. Beside these light isotopes presented here there are more particles of secondaryorigin which are used in these studies such as sub-iron particles or lithium, beryllium and boron. The effort aims todevelop a diffusion model which will describe the propagation of charged particles and their lifetime in our galaxy. Thiswill help also to better understand the energy density of different components within the interstellar space, such asmagnetic fields, electromagnetic radiation, gas pressure and cosmic rays. These model calculations have to deal witha number of parameters which have their origin in astrophysics, in nuclear physic and in high energy particle physics.The advantage of the light isotopes H and He in this context compared to the more heavy secondary particles lies6
TABLE 1Hydrogen isotope fluxes and their ratio derived with the ToF, errors are statistical andsystematics respectively.
Kinetic energy H flux H flux H/ Hat top of payload(GeV n − ) (GeV n − m s sr) − (GeV n − m s sr) − . ± . ± . · (33 . ± . ± .
5) (3 . ± . ± . · − . ± . ± . · (35 . ± . ± .
5) (3 . ± . ± . · − . ± . ± . · (36 . ± . ± .
4) (3 . ± . ± . · − . ± . ± . · (36 . ± . ± .
4) (3 . ± . ± . · − . ± . ± . · (36 . ± . ± .
4) (2 . ± . ± . · − . ± . ± . · (37 . ± . ± .
4) (2 . ± . ± . · − . ± . ± . · (38 . ± . ± .
4) (2 . ± . ± . · − . ± . ± . · (37 . ± . ± .
4) (2 . ± . ± . · − . ± . ± . · (36 . ± . ± .
3) (2 . ± . ± . · − . ± . ± . · (35 . ± . ± .
3) (2 . ± . ± . · − . ± . ± . · (35 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (34 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (33 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (33 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (32 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (31 . ± . ± .
2) (2 . ± . ± . · − . ± . ± . · (29 . ± . ± .
3) (2 . ± . ± . · − TABLE 2Hydrogen isotope fluxes and their ratio derived with the Calorimeter, errors are statisticaland systematics respectively.
Kinetic energy H flux H flux H/ Hat top of payload(GeV n − ) (GeV n − m s sr) − (GeV n − m s sr) − . ± . ± . · (37 . ± . ± .
0) (2 . ± . ± . · − . ± . ± . · (35 . ± . ± .
9) (2 . ± . ± . · − . ± . ± . · (36 . ± . ± .
9) (2 . ± . ± . · − . ± . ± . · (36 . ± . ± .
9) (2 . ± . ± . · − . ± . ± . · (35 . ± . ± .
9) (2 . ± . ± . · − . ± . ± . · (33 . ± . ± .
7) (2 . ± . ± . · − . ± . ± . · (31 . ± . ± .
7) (2 . ± . ± . · − . ± . ± . · (30 . ± . ± .
6) (2 . ± . ± . · − . ± . ± . · (29 . ± . ± .
6) (2 . ± . ± . · − . ± . ± . · (28 . ± . ± .
5) (1 . ± . ± . · − . ± . ± . · (26 . ± . ± .
4) (1 . ± . ± . · − . ± . ± . · (24 . ± . ± .
3) (1 . ± . ± . · − . ± . ± . · (22 . ± . ± .
2) (1 . ± . ± . · − . ± . ± . · (20 . ± . ± .
1) (1 . ± . ± . · − . ± . ± . · (18 . ± . ± .
0) (1 . ± . ± . · − . ± . ± . · (17 . ± . ± .
0) (1 . ± . ± . · − . ± . ± . · (15 . ± . ± .
1) (1 . ± . ± . · − in the fact that they do not have so many progenitor compared to the sub-iron particles or to lithium, beryllium andboron, it is predominantly He. A comprehensive and detailed study and discussion and interpretation of our resultsin this context is beyond the scope of this paper but we like to refer to a recent paper published by Coste et al. (2012). ACKNOWLEDGMENTS
We acknowledge support from the Russian Space Agency (Roscosmos), the Russian Foundation for Basic Research(grant 13-02-00298), the Russian Scientific Foundation (grant 14-12-00373), the Italian Space Agency (ASI), DeutschesZentrum fur Luft- und Raumfahrt (DLR), the Swedish National Space Board, and the Swedish Research Council.
REFERENCESAdriani, O., Barbarino, G. C., Bazilevskaya, G. A., et al. 2011,Science, 332, 69Adriani, O., Barbarino, G. C., Bazilevskaya, G. A., et al. 2013a,ApJ, 765, 91 Adriani, O., Barbarino, G. C., Bazilevskaya, G. A., et al. 2013b,ApJ, 770, 2Agostinelli, S., Allison, J., Amako, K., et al. 2003, Nucl. Instrum.Meth. A, 506, 250 TABLE 3Helium isotope fluxes and their ratio derived with the ToF, errors are statistical andsystematics respectively.
Kinetic energy He flux He flux He/ Heat top of payload(GeV n − ) (GeV n − m s sr) − (GeV n − m s sr) − . ± . ± . · (18 . ± . ± .
7) (7 . ± . ± . · − . ± . ± . · (20 . ± . ± .
8) (8 . ± . ± . · − . ± . ± . · (21 . ± . ± .
8) (9 . ± . ± . · − . ± . ± . · (23 . ± . ± .
8) (9 . ± . ± . · − . ± . ± . · (25 . ± . ± .
9) (10 . ± . ± . · − . ± . ± . · (26 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (26 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (27 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (27 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (27 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (27 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (26 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (25 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (24 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (23 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (22 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (19 . ± . ± .
87) (1 . ± . ± . · − TABLE 4Helium isotope fluxes and their ratio derived with the Calorimeter, errors are statisticaland systematics respectively.
Kinetic energy He flux He flux He/ Heat top of payload(GeV n − ) (GeV n − m s sr) − (GeV n − m s sr) − . ± . ± . · (28 . ± . ± .
2) (1 . ± . ± . · − . ± . ± . · (29 . ± . ± .
2) (1 . ± . ± . · − . ± . ± . · (28 . ± . ± .
2) (1 . ± . ± . · − . ± . ± . · (27 . ± . ± .
1) (1 . ± . ± . · − . ± . ± . · (26 . ± . ± .
1) (1 . ± . ± . · − . ± . ± . · (25 . ± . ± .
1) (1 . ± . ± . · − . ± . ± . · (24 . ± . ± .
0) (1 . ± . ± . · − . ± . ± . · (23 . ± . ± .
0) (1 . ± . ± . · − . ± . ± . · (22 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (20 . ± . ± .
9) (1 . ± . ± . · − . ± . ± . · (18 . ± . ± .
8) (1 . ± . ± . · − . ± . ± . · (16 . ± . ± .
7) (1 . ± . ± . · − . ± . ± .
5) (15 . ± . ± .
6) (1 . ± . ± . · − . ± . ± .
0) (13 . ± . ± .
59) (1 . ± . ± . · − . ± . ± .
4) (11 . ± . ± .
52) (1 . ± . ± . · − AMS Collaboration, Aguilar, M., Alcaraz, J., et al. 2002,Phys. Rep., 366, 331Aguilar, M., Alcaraz, J., Allaby, J., et al. 2011, ApJ, 736, 105Beatty, J. J., Ficenec, D. J., Tobias, S., et al. 1993, ApJ, 413, 268Boezio, M., Bonvicini, V., Mocchiutti, E., et al. 2002, Nucl. Instr.and Meth. A, 487, 407-422Bonechi, L., Adriani, O., Bongi, M., et al. 2006, Nucl. Instr. andMeth. A, 570, 281Brun, R., & Rademakers, F. 1997, Nucl. Instr. and Meth. A, 389,81-86Bruno, A., Cosmic ray antiprotons measured in the pamelaexperiment, Ph.D. thesis, University of Bari, Bari, Italy,http://pamela.roma2.infn.it/ (2008).Coste, B., Derome, L., Maurin, D., & Putze, A. 2012, A&A, 539,A88D’Agostini, G. 1995, Nucl. Instrum. Meth., A362, 487de Nolfo, G. A., Barbier, L. M., Christian, E. R., et al. 2000, AIPConf. Proc., 528, 425Garcia-Munoz, M., Mason, G. M., & Simpson, J. A. 1975, in the14th International Cosmic Ray Conference, Vol. 1, 319 Garcia-Munoz, M., Mason, G. M., & Simpson, J. A. 1975, ApJ,202, 265Gleeson, L. J., & Axford, W. I. 1968, ApJ, 154, 1011Heinbach, U., & Simon, M. 1995, ApJ, 441, 209Lamanna, G., Alpat, B., Battiston, R., et al. 2001, in the 27thInternational Cosmic Ray Conference, 1617Leech, H. W., & O’ Gallagher, J. J. 1978, ApJ, 221, 1110MacMillan, S., & Maus, S. 2005, Earth, Planets, and Space, 57,1135Menn, W., Bogomolov, E. A., Krut’kov, S. Y., et al. 2013, J.Phys.: Conf. Ser., 409, 012030Menn, W., Bogomolov, E. A., Formato, V., et al. 2013, in the33rd International Cosmic Ray Conference, 0233Mewaldt, R. A., Stone, E. C., & Vogt, R. E. 1976, ApJ, 206, 616Myers, Z. D., Seo, E. S., Abe, K., et al. 2001, in the 28thInternational Cosmic Ray Conference, 1805Myers, Z. D., Seo, E. S., Wang, J. Z., et al. 2005, Advances inSpace Research, 35, 151-155Picozza, P., Galper, A. M., Castellini, G., et al. 2007,Astroparticle Physics, 27, 296 E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 - s s r G e V ) F l u x ( m PAMELA ToFPAMELA Calorimeter H H E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 H H / PAMELA ToFPAMELA Calorimeter
Fig. 13.— H and H absolute fluxes (top) and their ratio (bottom) derived with the ToF (circles) or the calorimeter (squares). Errorbars show the statistical uncertainty while shaded areas show the systematic uncertainty.Reimer, O., Menn, W., Hof, M., et al. 1998, ApJ, 496, 490Shea M. A., Smart, D. F., & Gentile, L. C. 1987, Physics of theEarth and Planetary Interiors, 48, 200–205Strong, A. W., & Moskalenko, I. V. 1998, ApJ, 509, 212Strong, A. W., Moskalenko, I. V., & Ptuskin, V. S. 2007, AnnualReview of Nuclear and Particle Science, 57, 285Stephens, S. A. 1989, Advances in Space Research, 9, 145 Tomassetti, N. 2012, Ap&SS, 342, 131Verkerke, W., & Kirkby, D. 2011, arXiv:physics/0306116Wang, J. Z., Seo, E. S., Anraku, K., et al. 2002, ApJ, 564, 244Webber, W. R., Golden, R. L., Stochaj, S. J., et al. 1991, ApJ,380, 230Wefel, J. P., Ahlen, S. P., Beatty, J. J., et al. 1995, in the 24thInternational Cosmic Ray Conference, Vol. 2, 630 E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 - s s r G e V / n ) F l u x ( m He He PAMELA ToFPAMELA Calorimeter
E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 H e H e / PAMELA ToFPAMELA Calorimeter
Fig. 14.— He and He absolute fluxes (top) and their ratio (bottom) derived with the ToF (circles) or the calorimeter (squares). Errorbars show statistical uncertainty while shaded areas show systematic uncertainty. E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 - s s r G e V ) F l u x ( m PAMELAAMS-01BESS-93IMAX H H E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 H H / PAMELA AMS-01BESS-98 BESS-93IMAX
Fig. 15.— H and H absolute fluxes (top) and their ratio (bottom). For energies less than 361 MeV/n the ToF results (Table 1)were used, for higher energies the calorimeter results (Table 2). The previous experiments are: AMS-01 (Aguilar et al. et al. et al. et al. et al. E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 - s s r G e V / n ) F l u x ( m He He PAMELAAMS-01BESS-93BESS-98
E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 1 H e H e / PAMELA AMS-01BESS-93 BESS-98IMAX MASSSMILI SMILI2
Fig. 16.— He and He absolute fluxes (top) and their ratio (bottom). For energies less than 350 MeV/n the ToF results (Table 3)were used, for higher energies the calorimeter results (Table 4). The previous experiments are: AMS (Aguilar et al. et al. et al. et al. et al. et al. et al. E (GeV/N)0.1 0.2 0.3 0.4 0.5 0.6 1 H e H / PAMELABESS-93IMAXAMS-01
Fig. 17.— H/ He ratio compared to previous experiments: AMS-01 (Aguilar et al. et al. et al.et al.