anni accademici 1999-2002people.fisica.uniroma2.it/~iannucci/pubblicazioni/dottesi.pdf ·...

Universita degli Studi di Roma ”Tor Vergata”Facolta di Scienze Matematiche, Fisiche, Naturali

Tesi di Dottorato di Ricerca in Fisica

Physics PhD Thesis

Investigation of low energy cosmic rays with NINAmissions at low Earth orbit

Alessandro Iannucci

Relatore: Piergiorgio Picozza

Roma, 20 Dicembre 2002

Anni Accademici 1999-2002

1

To Valentina

Contents

Introduction 6

1 Cosmic ray physics 111.1 Early experiments and basic aspects . . . . . . . . . . . . . . . . 11

1.1.1 Early experiments . . . . . . . . . . . . . . . . . . . . . . 111.1.2 Chemical abundance . . . . . . . . . . . . . . . . . . . . 151.1.3 Isotopic abundance . . . . . . . . . . . . . . . . . . . . . 17

1.2 Life history of cosmic rays . . . . . . . . . . . . . . . . . . . . . 191.3 From the sources to the Heliosphere . . . . . . . . . . . . . . . . 21

1.3.1 The origin of cosmic rays . . . . . . . . . . . . . . . . . . 211.3.2 Acceleration mechanisms . . . . . . . . . . . . . . . . . . 231.3.3 Cosmic ray confinement and propagation . . . . . . . . . 25

1.4 Propagation in the Heliosphere . . . . . . . . . . . . . . . . . . 281.4.1 The Sun . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

1.5 From the Magnetosphere to the Earth . . . . . . . . . . . . . . 371.5.1 Structure of the Magnetosphere . . . . . . . . . . . . . . 37

1.6 The Earth’s magnetic field . . . . . . . . . . . . . . . . . . . . . 411.6.1 Trapped and albedo particles . . . . . . . . . . . . . . . 431.6.2 Charged particles in a magnetic field . . . . . . . . . . . 461.6.3 Adiabatic invariants . . . . . . . . . . . . . . . . . . . . 51

2 NINA ad NINA-2 detectors 612.1 NINA experiments: scientific overviews . . . . . . . . . . . . . . 612.2 The NINA instrument . . . . . . . . . . . . . . . . . . . . . . . 622.3 The NINA-2 instrument . . . . . . . . . . . . . . . . . . . . . . 682.4 Trigger logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692.5 NINA and NINA-2 beam tests . . . . . . . . . . . . . . . . . . . 70

2.5.1 Energy calibration . . . . . . . . . . . . . . . . . . . . . 722.5.2 Energy resolution . . . . . . . . . . . . . . . . . . . . . . 752.5.3 Mass resolution . . . . . . . . . . . . . . . . . . . . . . . 78

3 Experiments in orbit 853.1 The Resurs-01 satellite . . . . . . . . . . . . . . . . . . . . . . . 85

3.1.1 Resurs-O1-N4 launch . . . . . . . . . . . . . . . . . . . . 893.2 MITA satellite . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

3

4 CONTENTS

3.2.1 MITA launch . . . . . . . . . . . . . . . . . . . . . . . . 913.3 Telecommands . . . . . . . . . . . . . . . . . . . . . . . . . . . . 913.4 NINA-2 performance in orbit . . . . . . . . . . . . . . . . . . . 943.5 MITA in orbit control . . . . . . . . . . . . . . . . . . . . . . . 96

3.5.1 Data downlink and preliminary data handling . . . . . . 1023.5.2 On line data . . . . . . . . . . . . . . . . . . . . . . . . . 102

3.6 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1053.6.1 The track selection algorithm . . . . . . . . . . . . . . . 1053.6.2 Background estimation . . . . . . . . . . . . . . . . . . . 1083.6.3 Determination of fluxes . . . . . . . . . . . . . . . . . . . 109

3.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

4 Albedo protons 1174.1 Albedo and quasi-trapped particles . . . . . . . . . . . . . . . . 117

4.1.1 Previous measurements . . . . . . . . . . . . . . . . . . . 1184.1.2 Comparison of balloon-borne and AMS proton albedo

measurements . . . . . . . . . . . . . . . . . . . . . . . . 1224.1.3 The aims of the work . . . . . . . . . . . . . . . . . . . . 126

4.2 NINA results, interpretation and discussion . . . . . . . . . . . . 1264.2.1 Reconstructed mass of albedo hydrogen . . . . . . . . . . 1294.2.2 Proton flux . . . . . . . . . . . . . . . . . . . . . . . . . 1294.2.3 Longitude behaviour . . . . . . . . . . . . . . . . . . . . 1354.2.4 NINA, NINA-2 and AMS correlations . . . . . . . . . . . 1404.2.5 Altitude behaviour . . . . . . . . . . . . . . . . . . . . . 142

4.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

5 Albedo light isotopes 1455.1 Secondary production in the atmosphere . . . . . . . . . . . . . 1455.2 NINA missions: results and discussion . . . . . . . . . . . . . . 1475.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153

6 Geomagnetically trapped light isotopes 1576.1 Previous measurements . . . . . . . . . . . . . . . . . . . . . . . 1576.2 Data analysis and observations . . . . . . . . . . . . . . . . . . 1596.3 Results and interpretations . . . . . . . . . . . . . . . . . . . . . 1606.4 NINA-2 preliminary results . . . . . . . . . . . . . . . . . . . . 1706.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

7 Light Isotope Abundances in SEP 1717.1 Historical background . . . . . . . . . . . . . . . . . . . . . . . . 1717.2 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

7.2.1 Background estimations . . . . . . . . . . . . . . . . . . 1737.2.2 Flux measurements . . . . . . . . . . . . . . . . . . . . . 174

7.3 SEP measurements . . . . . . . . . . . . . . . . . . . . . . . . . 1747.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1877.5 NINA-2 preliminary results . . . . . . . . . . . . . . . . . . . . 189

7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Conclusions 195

Bibliography 197

Acknowledgements 211

5

Introduction

The project NINA (a New Instrument for Nuclear Analysis) is inserted ina wider international program dedicated to cosmic ray investigations in anenergy spectrum extending for about 4 orders of magnitude (from 10 to 105

MeV/nucleon), the WiZard program.The program originated in the late eighties, with the possi-

bility offered by NASA to perform scientific experiments on theprojected ”space station Freedom”. The experiment WiZard wasproposed and accepted [The WiZard collaboration 1988, Meschini et al. 1988,Spillantini et al. 1989]; its scientific purpose was to measure spectrum andisotopic composition of cosmic matter and to search antimatter, till energies ofseveral hundreds of GeV. Main device for recognizing particles was a magneticspectrometer provided by NASA, called ASTROMAG.

The program ASTROMAG, unfortunately, was first delayed and thendefinitively canceled due to the disaster of the Challenger mission and theconsequent crisis of the NASA scientific activities.

Meanwhile, starting from 1988, the same scientific observations ofthe projected WiZard-Astromag mission were fulfilled by means ofstratospheric balloon launches, although with less observation time and soless statistics [Golden et al. 1990, Bocciolini et al. 1996, Bellotti et al. 1997,Barbiellini et al. 1996]. The instrument was composed by a magneticspectrometer, a Cherenkov detector, at times alternated with a TRD ora RICH, and a streamer tubes calorimeter, definitively substituted by asilicon calorimeter starting from the third launch [Bocciolini et al. 1993,Aversa et al. 1995, Barbiellini et al. 1996]. The contemporary adoption ofthese detectors allows a sensitivity of the order of 10−4 in separatingantiprotons from electrons. Results of the ratio e+/(e+ + e−) andp/p in the cosmic radiation [Barbiellini et al. 1996b, Golden et al. 1996,Boezio et al. 1997, Boezio et al. 2001] have been obtained.

The WiZard collaboration was originally composed by italian universitiesand INFN sections, together with foreign partners like Swedish, German andAmerican institutions. Starting from 1994, the WiZard collaboration wasextended to Russian participators, with a program named the Russian-Italian Missions (RIM). This consists of four different experiments, oneto be performed on board of the Space Station MIR (RIM-0), and the otherthree onto satellites (RIM-1, RIM-2, RIM-3).

The experiment SiEye (RIM-0), performed on board of the Space Station,

7

studies the causes of the light-flashes observed by astronauts in condition ofdarkness [Casolino et al. 2001, Casolino et al. 2001b]. It seems evident thatsuch flashes do occur in high radiation regions of the orbit. In order to find adirect correlation between charged particles in space and light flashes, SiEyemakes use of few small silicon detectors to detect crossing particles. Threeinstruments has already flown in space, two on board of the Russian SpaceStation MIR while a third [Galper et al. 1996, Avdeev et al. 2001] on boardthe International Space Station ISS in April, the 25th, 2002 by Italian astronautCarlo Vittori.

NINA (RIM-1), with the following NINA-2, is the first ofthe RIM missions on satellite [Bakaldin et al. 1997, Sparvoli et al. 1997,Vacchi et al. 1997]. NINA’s goal is to detect cosmic ray nuclei of galactic,solar, trapped and albedo origin, at 1 AU, from hydrogen to iron,between 10 and 200 MeV n−1. The experiment was carried out onboard the satellite Resurs-01 n.4, developed by the Russian space companyVNIIEM. The spacecraft was launched on July, the 10th, 2001 into apolar Sun-synchronous orbit of altitude 840 km from the Russian basein Baikonur [Casolino et al. 1999, Sparvoli et al. 1999, Sparvoli et al. 2000,Sparvoli et al. 2000b]. NINA has been joined in space by a twin detector(NINA-2), housed on board the Italian satellite MITA [Casolino et al. 1999c].MITA was launched in a polar orbit, at a average altitude of 400 km, onJuly, the 15th, 2000, from the Plesetsk launch facility in Russia by means of aCosmos launcher. NINA and NINA-2 missions covered approximately two ofthe eleven years of 23th solar cycle.

The second experiment, PAMELA (RIM-2), is an expansion of theballoon activity, with the same purposes of physics: it will analyze spectraof antiprotons (80 MeV ≤ E ≤ 190 GeV), positrons (50 MeV ≤ E ≤270 GeV), protons (80 MeV ≤ E ≤ 700 GeV), and electrons (50 MeV≤ E ≤ 400 GeV), also detecting possible antinuclei, with a sensitivitynot possible for balloon missions [Adriani et al. 2002, Pearce et al. 2002,Spillantini et al. 2001, Bonvicini et al. 2001]. The instrument is composed bya magnetic spectrometer with its related tracking system, a TRD, a time-of-flight device and a silicon-tungsten calorimeter. PAMELA is going to fly inthe first part of 2003.

The third experiment on satellite was GILDA (RIM-3), dedicate tostudy cosmic gamma rays in a wide energy interval, from 20 MeV till100 GeV; it has been subsequently replaced by the international missionGLAST [Tavani et al. 2001, Pittori et al. 2001]. The instrument is basicallycomposed by a silicon-tungsten calorimeter, coupled with scintillating lead-fibers; electromagnetic showers induced by gamma rays are produced inside thedetector, and their energy stored in the active parts of the whole arrangement.GLAST is planned to fly in first part of 2005.

8

The candidate dedicated the three years of his PhD to the data analysis andto the mission control of projects NINA and NINA-2. After an introduction onthe cosmic ray physics (Chapter 1), the thesis describes the characteristic ofthe detectors (Chapter 2), the experiments in orbit and the methods of dataanalysis (Chapter 3); finally results obtained on different species of cosmicrays from NINA projects are outlined in sequence (albedo protons – Chapter4; albedo light isotopes – Chapter 5; geomagnetically trapped light isotopes –Chapter 6; Solar Energetic Particles (SEP) – Chapter 7).

9

10 CONTENTS

Chapter 1

Cosmic ray physics

In this Chapter the physical processes of cosmic rays are outlined. Startingfrom an historical point of view I describe the pioneers and the earlyexperiments which carried to discover a ”new physics”. The basic aspectsof primary radiation, its origin, acceleration and propagation mechanisms inthe galaxy and in the heliosphere are outlined. Finally, the influences of theEarth’s magnetic field in primary and secondary radiation is described.

1.1 Early experiments and basic aspects

After an historical background on early cosmic rays experiments, in this SectionI will treat the chemical and isotopic abundance of cosmic rays observed atEarth.

1.1.1 Early experiments

The existence of a penetrating radiation coming from outer space wasdemonstrated in the first years of last century by measurements on theconductivity of gases. At that time it was believed that gases should be verynearly perfect insulators, provided that the applied electric field was no toohigh. However, many workers, among them Wilson, Elster and Geitel, showedthat, in spite of careful precautions to prevent the known radiations fromreaching the samples of air in their ionization chambers, a significant residualconductivity remained. It was found that a reduction in conductivity resultedfrom shielding the ionization chambers by lead, and this fact was interpreted asshowing that most of the residual conductivity was due to an external radiationof some form.

Of the possible radiations responsible for the conductivity the most likelyseemed to be those from radioactive materials in the Earth. In attempts totest this hypothesis a number of workers flew ionization chambers in balloonsand studied the variation of conductivity with height. The early experimentswere inconclusive due to technical difficulties, but in 1912 Hess, and shortlyafterwards, Kolhorster, showed that the ionisation in the chamber decreased

11

12 CHAPTER 1. COSMIC RAY PHYSICS

Figure 1.1: Hess after one of the successful balloon flights in which the increase inionization with altitude through the atmosphere was discovered.

1.1. EARLY EXPERIMENTS AND BASIC ASPECTS 13

with height in going from ground to about 700 metres and then increasedsteadily. Although the initial decrease in ionisation at low altitude couldbe accounted for by the reduction in effect of the Earth’s radioactivity, theincrease at higher altitudes was obviously due to some other cause. Hess putforward the revolutionary hypothesis that the increase was due to an extremelypenetrating radiation which was coming from outer space. He inferred thatthe radiation was very penetrating by the large amount of absorber required toreduce the residual conductivity. Hess deduced further that the radiation wasnot of solar origin because the absence of any significant difference in ionisationbetween day and night experiments. This penetrating radiations soon came tobe known as the cosmic radiations. At this time the most penetrating radiationknown was the γ-radiation from radioactive materials and it was natural tointerpret the cosmic radiation as due to high energy γ-rays. After World War Imany other workers studied this subject, that could be investigated with quiteinexpensive apparatus, and progress quickened. The correlation of cosmicray intensity with atmospheric pressure was discovered by Myssowsky andTuwim in 1926, as well as a number of other variations, mainly associated withmeteorological factors. A fundamental advance occurred with the introductionof electrical methods of detecting ionising particles, in particular the use of theGeiger-Muller counter. Bothe and Kolhorster used arrangements of countersto prove that at least some of the cosmic rays in the lower levels of the Earth’satmosphere consisted of charge particles. The coincidence arrangement provedinvaluable for studying the cosmic ray intensity in different directions. Thelatitude effect, i.e. the variations of cosmic ray intensity with geomagneticlatitude, had been discovered in 1925 by Clay, and Bothe and Kolhorstershowed that this proved the existence of charge particles above the Earth’smagnetosphere. In the same period, Compton’s measurements indicate auniform variation with latitude, showing minimum at or near the equatorand increasing intensity toward the north and south poles. At sea level, thedifferences between the intensity at latitude 45o and 0o was roughly 16%,whereas at an elevation of ∼ 2900 m the difference was about 23%. This workshowed no significant variations with longitude [Compton 1932]. The workson the latitude effect was followed by classical studies in which Compton andhis collaborators measured the cosmic ray intensity at many points of theEarth [Compton 1933]. Another advance came in 1933 with the discovery byJohnson and Street that the intensities from the West and the East were notequal (East-West effect) [Johnson 1933]. They measured differences in cosmicray flux coming from West and from East greater than six times the probableerror. This was a rather important result which showed that not only cosmicrays were composed from charged particles, as distinct from γ-rays, but thatthe majority of the particles were positively charged.

In a series of flights of Geiger counters rockets into the upper atmospherein the 1940s, Van Allen was able to trace out a counting rate curve that roseto a Pfotzer maximum [Pfotzer 1936] at a height of about 19 km, above whichthe remaining atmosphere corresponded to about 56 g/cm2 of material (see


Figure 1.2: Smoothed composite curve of single counter counting rates above WhiteSands, New Mexico [Gangnes et al. 1949].


Figure 1.2) [Gangnes et al. 1949]. With increasing altitude beyond that levelthe counting rate declined until it leveled off at a constant rate at and above55 km.

An important discovery having implications for astrophysics came in 1948when Freier and co-workers detected the presence of a significant fractionof heavy nuclei in the primary radiation. The initial measurements weremade with nuclear emulsions carried into the stratosphere by balloons. Withimprovements in technique long term studies of the time-variations of thecomponents of the sea level flux became possible and variations in theprimary intensity were distinguished. Many measurements identified intensityvariations with solar activity and indicated the role of the Sun as a modulatorfor cosmic ray flux.

1.1.2 Chemical abundance

Some basic information on the cosmic rays observed on Earth is given in thissection. First of all, we must note that the energy spectrum of galactic cosmicrays at energies less than a few tens of GeV changes with variations in solaractivity. This modulation effect becomes larger with increasing energy. Evenat the minimum of solar activity, the intensity of the cosmic rays on the Earthat nonrelativistic energies is substantially less than at the boundaries of thesolar system. In the energy region εk> 10 GeV/n (where εk is the kineticenergy for nucleon) solar modulation has little effect on the particle spectrumand even less on the relative abundance of nuclei [Hillas 1972].

The most abundant elements in cosmic rays are the nuclei of hydrogenand helium. Their energy spectra (togheter with C and Fe ones) are shownin fig 1.3. On the amount of heavier nuclei in cosmic rays we have detailedinformations only for particle energies up to εk< 10 GeV/n.

The differential energy spectra of the various species can be well representedby a power-law distribution, conventionally written as

N(εk)dεk = Kε−γk dεk . (1.1)

The exponent γ is energy dependent and lies in the range 2.6-2.8 for εk > 80MeV/n.

In Table 1.1 the relative abundance of nuclei in different energy ranges iscompared with the solar system abundance and those estimated for the localinterstellar medium [Simpson 1983].

The most important characteristic of the chemical composition, which canbe seen clearly in Table 1.1, is the presence of a rather large flux of light nuclei(Li, Be, B), in spite of the average insignificant quantities of these elements innature. This feature, which is also true for other rare nuclei (for instance, forthe nuclei 2H, 3He, F, K, Sc, V, etc.), suggests that the chemical composition ofthe cosmic rays is greatly changed while they travel through interstellar space,and, possibly, also in the sources (i. e., in the region of accelerations). Therelative abundance of secondary nuclei (producing during the interaction of the


Figure 1.3: Spectra of cosmic rays on Earth.


cosmic rays with the interstellar gas, like spallation process) gives importantinformation on the nature of the propagation of cosmic rays in the galaxy[Hopper 1964].

1.1.3 Isotopic abundance

In addition to the overall chemical abundance, there are several reasons whyalso the isotopic abundance are of particular interest.

A special group of isotopes is formed by the very lightest stable elements,1H, 2H, 3He and 4He. 2H and 3He nuclei observed in galactic cosmic radiationare absent in their sources, in which they are unstable against thermonuclearreactions [Clayton 1968]. They are formed by high energy nuclear interactionsbetween cosmic rays and interstellar matter [Ramaty and Lingenfelter 1969].

Since 3H decays into 3He with a 12.2 year half-life, negligible with respectto the cosmic ray propagation time (106 - 108 years), it is possible to assimilateevery 3H production to an 3He production (it is not true for secondary particlesin atmosphere, when the time of life is less than 3H half-life).

The cross sections for the production of 2H, 3H and 3He are generally ofthe same order of magnitude in p-p, p-4He, 4He-4He and p-(Z>2) interactions.Thus the most important reactions in this context are those in which the mostabundant nuclei are involved (namely p-p and p-4He reactions).

The production of 2H, 3H and 3He in p-4He collisions can occur in twoways:

1. Fragmentation of cosmic ray 4He on interstellar idrogen

2. Fragmentation of cosmic ray 1H on interstellar 4He

In most cases the secondary 2H, 3H or 3He is emitted with a low energy inthe frame of the 4He nucleus.

In a frame of reference fixed in space, it has an energy/nucleon close tothat of the incident 4He in first case, and gives rise to a low energy componentin second case.

Moreover, the 2H and 3He formed can themselves undergo fragmentationon interstellar matter. Here, p-d and p-3He cross section are involved.

Finally protons, 2H, 3He and 4He nuclei may be slowed down to anappreciable extent (with respect to electronic collisions) by nuclear elasticscattering on interstellar matter. Therefore elastic cross section are involved[Meyer 1972].

Table 1.2 reports the cosmic ray abundance of these elements at differentenergies compared with the local interstellar abundance [Simpson 1983].

Another important aspect of isotopic abundance is the fact that some ofthe species created in the spallation reactions are radioactive and so, if theproduction rates of the different isotopes of a certain element are known,the time taken by these samples to reach the Earth from their source canbe estimated. The most famous of these “cosmic rays clocks” is the isotope


Cosmic raysAverage at

Element 70-280 600-1000 1000-2000 Solar System Local GalacticHe 41700± 3000 27030± 580 (0.27± 0.06)×106

Li 100± 6 136± 3 5.0×10−3

Be 45± 5 67± 2 69.4± 10 8.1×10−5

B 210± 9 233± 4 212± 10 3.5×10−3

C 851± 29 760± 16 684± 27 1110 1300± 300N 194± 8 208± 5 188± 6 231 230± 100O 777± 28 707± 15 607± 28 1840 2300± 500F 18.3± 1.3 17.0± 1.1 13.5± 2.3 0.078 0.093(1.6)Ne 112± 6 113± 3 100± 3 240 270(1.7)Na 27.3± 3.4 25.8± 1.1 21.3± 3.2 6 5.6± 0.9Mg 143± 6 142± 4 125± 12 106 105± 3Al 25.2± 3.0 28.2± 1.2 22.2± 3.2 8.5 8.4± 0.4Si 100 100 100 100 100.3P 4.0± 0.7 5.3± 0.5 5.3± 1.6 0.65 0.96± 0.2S 16.4± 1.2 23.1± 1.1 19.6± 0.9 50 45± 13Cl 3.6± 0.5 6.4± 0.5 4.7± 0.4 0.47 0.47(1.6)Ar 6.3± 0.6 10.2± 0.7 8.2± 1.2 10.6 9.0(1.7)K 5.1± 0.6 7.2± 0.5 6.3± 0.4 0.35 0.36± 0.12Ca 13.5± 1.0 16.1± 0.9 13.1± 1.2 6.25 6.2± 0.8Sc 2.9± 0.5 4.5± 0.5 3.3± 1.1 0.003 0.0035± 0.0005Ti 10.7± 0.9 10.2± 0.7 9.1± 0.9 0.24 0.27± 0.04V 5.7± 0.6 6.7± 0.5 4.6± 0.3 0.025 0.026± 0.005Cr 10.9± 1.0 11.8± 0.8 9.1± 0.8 1.27 1.3± 0.12Mn 7.2± 1.2 8.2± 0.7 6.3± 0.4 0.93 0.79± 0.17Fe 60.2± 3.2 69.8± 2.0 60.5± 7.6 90.0 88± 6Co 0.2± 0.1 0.4± 0.2 0.22 0.21± 0.03Ni 2.9± 0.4 3.7± 0.5 2.8± 0.6 4.78 4.8± 0.6Cu 0.038± 0.006 0.052(1.6)Zn 0.035± 0.005 0.135(1.6)

Table 1.1: Galactic cosmic ray abundances at 1 AU, normalized to Si=100,compared with the Solar System and the local interstellar abundances (the energyintervals 70-280, 600-1000, 1000-2000 are in MeV/nucleon).

1.2. LIFE HISTORY OF COSMIC RAYS 19

Isotope 60 MeV/ 80 MeV/ 200 MeV/ local interstellarratio nucleon nucleon nucleon abundance

2H/1H (4.4 ± 0.5) × 10−2 (5.7 ± 0.5) × 10−2 1.0 × 10−5

3He/4He (9.5 ± 1.5) × 10−2 (11.8 ± 0.7) × 10−2 3.0 × 10−5

2H/4He 0.21±0.09 0.31± 0.03 1.0 × 10−4

Table 1.2: Isotope ratios of hydrogen and helium; cosmic ray abundance arecompared with the local interstellar abundance (last column).

10Be which is a radioactive half-life of 1.5×106 years and so is very useful todetermine the typical lifetime of the spallation products in the vicinity of theEarth.

1.2 Life history of cosmic rays

It is known that some low energy cosmic rays are ejected by the Sun from timeto time and since the Sun is a typical star it is reasonable to suppose that otherstars emit cosmic rays. However, the rate of production of cosmic ray by starssimilar to the Sun is too low to give the observed intensity, and as we shallsee later, various theories have been proposed to account for the origin andacceleration of the particles. For the present it is sufficient simply to distinguishbetween the radiation reaching the Earth from outside the heliosphere andthose particles coming from the Sun.

Between their origin and their arrival at some point of detection in theatmosphere the primary cosmic rays are acted on by a variety of forces.These are best appreciated by reference to a diagram (see Figure 1.4).The acceleration process for the general cosmic rays may occur in thesource itself or perhaps take place over a separate region of the galaxy,so source and acceleration are bracketed together. The interaction of theparticles with the galaxy can be divided into two main processes. Thereare the physical interaction with galactic matter, in the form of scattering,disintegration, collision with galactic atoms, and the interaction with magneticfield. Similarly, in the solar system there are the physical interaction with solarsystem matter and magnetic interaction. Sun is both a source of solar cosmicrays than a modulator of galactic cosmic rays. Finally at the end of its travel tothe Earth, a cosmic ray is affected by geomagnetic field and Earth’s atmosphere[Wolfendale 1963].

As we would expect, the cosmic ray data duplicate existing astrophysicaland geophysical data in some places and extend them in others; in both casesthey contribute to our knowledge of the properties of the universe.


Figure 1.4: Life history of the primary cosmic rays.

1.3. FROM THE SOURCES TO THE HELIOSPHERE 21

1.3 From the sources to the Heliosphere

In this section the basic aspects of origin models, acceleration, confinement andpropagation mechanisms of cosmic rays from the sources to the Heliosphereare outlined.

1.3.1 The origin of cosmic rays

One of the central questions of the astrophysics of cosmic rays is the problemof their origin, and below we shall be concerned with the origin of the mainpart of the cosmic rays observable on Earth.

To answer this question means in the first place that we have to indicateat least roughly the position of the sources of the cosmic rays, for instance,to prove that they are in the galactic disc. Secondly, we have to identifythe sources, for instance, relating them to supernova explosions. However, wehave to emphasize that to know the nature of the sources is not of particularimportance for clarifying a whole series of other questions. For instance, itis not particularly important for solving the third problem: to establish thetrapping region of the cosmic rays in the Galaxy. Furthermore, there arisemany other problems: how do cosmic rays propagate from the sources to theEarth, how does their chemical and isotopic composition change during thisprocess, what role do plasma and magnetohydrodynamic effects play in thepropagation of cosmic rays and so on.

We assume that the intensity of cosmic rays depends only on the totalenergy E (IZ,A=IZ,A(E)) of the corresponding particles (of course, one can alsouse the kinetic energy Ek, or the energy for nucleon). This is a very goodapproximation in view of the high isotropy of the cosmic rays.

It was already clear in the fifties, after radio-astronomical data on cosmicrays far from the Earth became available, that there was no basis for solarmodels [Ginzburg and Syrovatskii 1964]. As regards metagalactic models thesituation is more complicated. In such models the main sources of cosmic raysare assumed to be metagalactic and, consequently, the cosmic rays observed onEarth first traveled from a metagalaxy to our Galaxy before reaching the solarsystem. There is no doubt that the electron component of the cosmic raysobserved in the Earth has a galactic origin. For a number of reasons, whichwe shall not discuss here, it is also very probable that the proton-nuclearcomponent of the cosmic rays has a galactic origin. Nevertheless, logicallyspeaking not every conceivable model of a metagalactic origin of the proton-nuclear component has been ruled out. Apparently, for this purpose only γ-rayastronomical methods are adequate. In metagalactic models, the cosmic raysoutside the Galaxy should possess approximately the same characteristics aswithin the Galaxy. In particular, the energy densities should be equal:

wc.r.,Mg wc.r.,G ∼ 10−12erg/cm3 . (1.2)

The equation should be valid everywhere and, in particular, in the


Magellanic Clouds. In such conditions we can predict fairly reliably the fluxof γ-rays from πo decays, which should originate from each of the clouds. Forinstance, for the Small Magellanic Cloud the photon flux Fγ(Eγ>100MeV) isequal to 107 photons/cm2 if we take into account only the atomic hydrogenin the cloud. Obviously, if we take into account molecular hydrogen and theγ-rays from electronic bremsstrahlung emission, we get a larger flux. If thecosmic rays are formed in the clouds themselves, then in view of the relativelysmall size of these clouds they can rather quickly escape into interstellar space.Under such conditions the flux, for instance for the Small Magellanic Cloud,may turn out to be lower than the lower limit mentioned. Much more preciselythan the flux from each of the clouds one can predict the ratio of the γ-raysfluxes from the clouds. A clear discrepancy between the observations andthese predictions will be a direct refutation of metagalactic models. Namely,in metagalactic models the value of the density of cosmic rays in the Galaxy,wc.r.,G∼10−12 erg/cm3, should be constant over the whole system, includingthe peripheral regions, and in particular it should no decrease away from thesolar system in the direction of the galactic anticentre. Meanwhile, some of theavailable observations of the γ-ray intensity Iγ(Eγ>100MeV) in the directionof the anticentre indicate a drop in the cosmic ray intensity (and, of course,their energy density wc.r.) away from the solar system [Ginzburg 1969].

The galactic models

In galactic models the sources of the cosmic rays are, of course, situated in theGalaxy. In which region the cosmic rays are concentrated is in a certain sensea separate question. If we use the energy requirements and the nonthermalradiation in the various intervals as a guidline, then the most probablecandidates are supernovae, pulsars, and neutron stars in close binary system.In this section it will be described one of the most probable candidates inthe cosmic rays galactic models: the supernova explosions [Friedlander 1989].From the energy point of view such a suggestion is completely acceptable.Supernovae explosions occur in our galaxy about each fifty years. As mentionedabove, in the solar system, the average density of cosmic ray energy is 1eV/cm3. Assuming that this is typical of the rest of the galaxy, with a galacticdisk volume of 1063 cm3, the total cosmic rays energy content is then about1067 eV or 1.6×1055 erg. The average cosmic rays lifetime is approximately3×1014 s, obtained by the 10Be isotope abundance. Thus the rate at whichcosmic rays energy is lost amounts to (1.6×1055 erg)/(3×1014 s) or 5×1040

erg/sec, and this loss must be compensated by injection and acceleration ofnew cosmic rays.

If we assume supernova explosion each fifty years (1.5×109 s), and thetypical supernova yields of 1050 erg in fast particles, we obtained an averagepower of (1050 erg)/(1.5×109 s) or 6×1040 erg/s. This simple calculationexplains the reasons of supernova as most probable candidate for cosmic rayssource.


The mechanisms responsible for cosmic rays acceleration are stillnot totally clear. In galactic models the sources are characterized byturbolent motion of the medium and the presence of high magnetic fields[Blanford and Ostriker 1978]. In additions, as it is well known, in variablemagnetic fields, electric fields are induces and accelerate charged particles.Unfortunately, however, we have to acknowledge that no model of cosmicray acceleration has yet been developed that makes it possible to choosethe main acceleration mechanisms among the many possible ones. Thegreat variety of conditions and processes in the cosmos leads to theconclusion that there are numerous acceleration mechanisms of energeticparticles. In particular, this is true for such well studied processes as theacceleration of particles in the radiation belts of the Earth [Tverskoi 1965]and of Jupiter [Mendis and Axford 1974], the appearance of fast particlesduring magnetospheric substorms [Akasofu 1969], in the interplanetary space[Van Allen et al. 1968] , and during solar flares [Svestka 1976].

All regions in which there are high voltages and strong magnetic fieldscould be considered as sites of mechanism acceleration processes. Supernovaerepresent only one source type in which rapid acceleration occurs in a relativelycompact region.

In next section more complete descriptions of second and first order (inV/c) acceleration mechanisms are outlined.

1.3.2 Acceleration mechanisms

V/c second order acceleration mechanism

In 1949, Enrico Fermi evolved a theory in which the main process ofacceleration is due to the interaction of cosmic particles with wanderingmagnetic fields which occupy the interstellar space [Fermi 1949]. In thistheory charged particles are reflected from magnetic mirrors associated withirregularities in the galactic magnetic field. The mirrors are assumed to movewith velocity V . The energy enhancement due to reflections is statisticallydescribed as

dE

dt=

2V 2

cλE = αE , (1.3)

where λ=ρcτ is the mean free path in g/cm2 between magnetic clouds,which are site of mirroring (ρ being the mean matter density and τ the meantime interval). From 1.3 a power law energy spectrum of cosmic rays isobtained:

N(E) ∝ E−γ, γ = (1 − 1

ατa

) , (1.4)

in which dt ∼ τa is the mean time spent by a particle in the accelerationregions. In a modern version of Fermi second order acceleration, the particlesinteract with various types of plasma waves and gain energy being scattered


stochastically by these waves. In the interstellar space, the relativistic cosmicray plasma streams along the large scale magnetic field lines. In this processthey collectively excite a spectrum of hydromagnetic waves, called Alfvenwaves. These kind of waves propagate in the ionized component of theinterstellar gas along the streaming direction of the particles. Resonantinteraction between cosmic rays and Alfven waves occurs when cosmic rayperforms one gyration around the magnetic field lines while traversing oneAlfven wavelength.

V/c first order acceleration mechanism

This acceleration mechanism is the result of individual head-on collision leadingto a more rapid increase of particle energy. The condition to obtain it isa strong shock wave propagating at a supersonic velocity V (higher thanAlfven velocity) in the direction of the magnetic field lines through stationaryinterstellar gas. We assume that the gas is flowing in at a velocity us = V inthe shock frame. After the shock, the velocity downstream becomes ud = V/ζ,if ζ is the compression factor of the shock. The presence of scattering centersis postulated, so that cosmic rays diffuse (with a diffusion coefficient functionof position, particle momentum and time) on both sides of the shock. Thescattering centers ensure that the particles will be reflected across the shocka large number of times. Every passage through the shock is equivalent torunning head-on into the magnetic wall of velocity

W = us − ud = V (1 − 1

ζ) . (1.5)

Averaged over all incidence angles, there is a mean energy gain of

∆E =4

3

V

c(1 − 1

ζ)E (1.6)

per traversal of the shock. Taking proper account of the probability ofparticles escaping the system leads to the time-independent spectrum:

N(E) ∝ E−γ, γ = (2 + ζ

ζ − 1) . (1.7)

A typical value of ζ for strong adiabatic shocks is ∼ 4, and the consequentialvalue of γ is ∼ 2. From 1.7 it is clear that weaker shocks generate steeperspectra. In the time independent limit the slope of the power law generatedby the acceleration mechanism depends on the ζ value.

Energy spectra of cosmic rays near the Earth have been shaped by acombination of injection, acceleration and propagation processes. Energylosses and collisions contribute too, and they all depend on the particle energy.Observations show that different acceleration processes can all produce cosmicrays with energy spectra having the characteristic E−γ shape.


Many other specific models and new interpretation have been proposedto describe acceleration process in different parts of Universe [Webber 2000,Share and Murphy 2000, Meziane 2000, Lee 2000].

1.3.3 Cosmic ray confinement and propagation

The transport mechanisms for cosmic rays in the galaxy are dominated byparticle motion in the galactic magnetic fields. The cosmic rays, then, movein spiral trajectories around the magnetic field with gyroradii appropriate tothe particle energy. Even the field lying approximately in the galactic disk issomewhat disordered, and field ”loops” extend out of the disk into a galactichalo. A cosmic ray particle gyrating around such a field will interact with theseand smaller irregularities and can be scattered by them, so that the cosmicrays may become diffuse in the magnetic field. Therefore, diffusion theoryis often employed to describe the large-scale propagation of the cosmic rays[Ginzburg and Syrovatskii 1964, Owens 1976]. Moreover the magnetic fieldconfigurations permit access of the cosmic rays to the galactic halo where thepropagation conditions are different. Thus, a complete treatment of cosmicray transport, in the diffusion approximation, must consider diffusion both inthe galactic disk and also within the halo region [Simpson 1983].

In succession I present two of the most referred models of cosmic rayspropagation in the literature: the Leaky Box Model (LBM) and the DiffusionHalo Model (DHM). Both model are able to explain the cosmic ray datasurprisingly well although the physical framework of both models differs.

Leaky Box Model (LBM)

.In the LBM the particles propagate freely in the containment volume

and the productions and losses of particles are balanced in time; thus themathematical description of the LBM is given by a continuity equation. Byignoring energy changing processes and radioactive particles, the equation hasthe following form:

Ni(E) 1

τescape(E)+

1

τint(E) = iQprim(E) +

∑k>i

Nk(E)

τ k→iint (E)

, (1.8)

where Ni(E) [cm−3GeV −1] and Nk(E) [cm−3GeV −1] are the numberdensities of different types of nuclei of kinetic energy E. The left side ofequation 1.8 accounts for the losses of i-type nuclei and the right side forthe sources. The secondary sources originate from spallation of k-type nucleiheavier than i-type nuclei. The quantity τ k→i

int (E) means the mean timewhich a k-type nuclei needs to produce i-type secondary in the interstellargas. These quantities depends on the production cross section and on the


interstellar gas in terms of density and composition. As one can see ahuge number of cross sections are involved and not all are so well known[Molnar and Simon 2001, Berezinskii et al. 1990]. In the brackets of the leftside one finds expressions for the losses and τint(E) stands for the meanlifetime of i-type particles against interactions in the interstellar gas. Hereagain nuclear cross sections are involved. In the LBM the quantity τesc(E)is used as a free parameter and stands for the mean escape time from theconfinement volume, often called the age of cosmic rays. On a statistical basisit is an exponential distribution which governs the escape of an individualparticle and τesc(E) is the mean of it. The exponential distribution resultssince the probability for a particle to escape from the box in time dt is givenby dt/τesc(E).

I refer to Simon [2001] for a complete discussion of solution of 1.8.

In the next section I will introduce the DHM model which can be considereda more realistic physical model within which to discuss the propagation ofcosmic rays. We underline that the LBM equations represent a mathematicalapproximation of the DHM.

Diffusion Halo Model (DHM)

.

Figure 1.5 sketches the physical picture of the DHM. The shaded areaillustrates the thin galactic disk of height hg and the quantity H stands forthe height of the halo. It is assumed that the cosmic ray sources are placedin the thin galactic disk, where most of the interstellar gas is located, butthe cosmic rays themselves diffuse out and may spend considerable portions oftheir lifetime in the halo. A three dimensional diffusion equation would be theproper approach to the problem. It could cover physical details in our galaxysuch as the spatial gas distribution of atomic and molecular hydrogen, couldlink the cosmic ray sources to the distribution of supernova remnants, andcould also add aspects such as galactic winds and convection. This illustratesthat the DHM accounts for much more physical detail than the LBM.

For this discussion it is sufficient to make the picture somehow simpler.We will allow that the cosmic ray sources and the interstellar gas arehomogeneously distributed throughout the thin galactic disk and will ignoreconvection and energy changing process. This provides symmetry and if onefurther ignores energy changing process, this scenario can be described with aone dimensional diffusion equation:

∂Ni(z, t)

∂t=

∂

∂zD(z)

∂

∂zNi(z, t) − Ni(z, t) 1

iτint(E) +

+iQprim(z) +∑k>i

Nk(z, t)

τ k→iint

. (1.9)


Figure 1.5: Schematic view of the physical concept of the DHM. The shaded areasymbolizes the thin disk and H stands for the halo size. The sources of cosmic raysare in the disk and the escape into the halo and finally into the intergalactic spaceis controlled by diffusion.


Ni(z, t) and Nk(z, t) describe the density of i-type and k-type particles atposition z at time t, with k heavier than i. The first term of the right sidedescribes the diffusion and D(z) means the diffusion coefficient at position z.The second bracket on the right side of equation 1.9 accounts for the lossesof i-type particles similar to those quantities described in the last section. Inthe last two terms one finds the sources for the i-type particles. One can haveprimary sources iQprim(z) as well as secondary sources by spallation of k-typenuclei expressed by the last two term on the right side of equation 1.9. I referto Simon [2001] for the results obtained by solving 1.9 under different setsof parameters and boundary conditions, and simply point out that identicalresults from the LBM and DHM models are consequences of a mathematicalrelation between the diffusion equation and the equilibrium (or continuity)equation which describes DHM and LBM respectively.

1.4 Propagation in the Heliosphere

The heliosphere is a bubble in space produced by the solar wind. Althoughelectrically neutral atoms from interstellar space can penetrate this bubble,virtually all the material in the heliosphere emanates from the Sun itself. Atsome distance from the Sun, well beyond the orbit of Pluto, the supersonicsolar wind must slow down to meet the gases in the interstellar medium. Itmust first pass through a shock, the termination shock, to become subsonic.It then slows down and gets turned in the direction of the ambient flow of theinterstellar medium to form a comet-like tail behind the Sun. This subsonicflow region is called the helio-sheath. The outer surface of the helio-sheath,where the heliosphere meets the interstellar medium, is called the heliopause.

The distance to the heliopause is been estimate to 100-160 AstronomicalUnit (AU; 1 AU = 1.5×108 km). Interplanetary spacecraft such as Pioneer 10and 11 and Voyager 1 and 2 are passing through the heliopause in these years[Pioneer 2002, Voyager 2002].

Measurements of galactic cosmic ray composition are limited to positionswithin the heliosphere. A high energy charged particle propagating from thesource to Earth’s orbit will scatter and diffuse inward in the irregular magneticfields, will tend to be convected outward as a result of the outward momentumflow of the solar wind, and will undergo adiabatic deceleration owing to theoutward expansion of the magnetic fields [Parker 1966]. A particle that has akinetic energy of 500 MeV per nucleon in the local interstellar medium may lose1/3 or more of its kinetic energy before arriving at 1 AU. The resulting changein the spectra of galactic cosmic rays between the source to Earth’s orbit is aheliocentric phenomenon [Fan et al. 1960] that has so far been investigated tobeyond ∼ 25 AU [McDonald et al. 1981, McKibben et al. 1982].

1.4. PROPAGATION IN THE HELIOSPHERE 29

Figure 1.6: Structure of heliosphere.


1.4.1 The Sun

The Sun produces its energy by nuclear fusion. The solar interior is separatedinto four regions by the different processes that occur there. Energy isgenerated in the core, the innermost 25%. This energy diffuses outward byradiation (mostly gamma-rays and X-rays) through the radiative zone andby convective fluid flows (boiling motion) through the convection zone, theoutermost 30%. The thin interface layer (the ”tachocline”) between theradiative zone and the convection zone is where the Sun’s magnetic field isthought to be generated.

The solar atmosphere is separated into four regions: the photosphere, thechromosphere, the transition region and the corona. The photosphere is thevisible surface of the Sun that we are most familiar with. Since the Sun isa ball of gas, this is not a solid surface but is actually a layer about 100 kmthick. A number of features can be observed in the photosphere with a simpletelescope (along with a good filter to reduce the intensity of sunlight to safelyobservable levels). These features include the dark sunspots, the bright faculae,and granules. The chromospheric network is a web-like pattern most easily seenin the emissions of the red line of hydrogen (H-alpha) and the ultraviolet lineof calcium (Ca II K - from calcium atoms with one electron removed). Thenetwork outlines the supergranule cells and is due to the presence of bundlesof magnetic field lines that are concentrated there by the fluid motions inthe supergranules. The transition region is a thin and very irregular layerof the Sun’s atmosphere that separates the hot corona from the much coolerchromosphere. Heat flows down from the corona into the chromosphere and inthe process produces this thin region where the temperature changes rapidlyfrom 106K down to about 2×104K. The corona is the Sun’s outer atmosphere.It is visible during total eclipses of the Sun as a pearly white crown surroundingthe Sun. The corona displays a variety of features including streamers, plumes,and loops. These features change from eclipse to eclipse and the overallshape of the corona changes with the sunspot cycle. The coronal gases areheated to temperatures greater than 106K. At these high temperatures bothhydrogen and helium (the two dominant elements) are completely stripped oftheir electrons. Even minor elements like carbon, nitrogen, and oxygen arestripped down to bare nuclei. Only the heavier trace elements like iron andcalcium are able to retain a few of their electrons in this intense heat.

Solar wind, Magnetic Clouds and Co-rotating Interactive Regions

The solar wind streams off of the Sun in all directions at speeds of about 400km/s. It is mainly composed by e−, p and He4 nuclei of energy of ∼ keV. Thesource of the solar wind is the Sun’s hot corona. The solar wind is not uniform.Although it is always directed away from the Sun, it changes speed and carrieswith it magnetic clouds, interacting regions where high speed wind catches upwith slow speed wind, and composition variations. The solar wind speed ishigh (800 km/s) over coronal holes and low (300 km/s) over streamers, as the


Ulysses [Ulysses 2002] spacecraft data have shown. These high and low speedstreams interact with each other and alternately pass by the Earth as the Sunrotates. These wind speed variations buffet the Earth’s magnetic field and canproduce storms in the Earth’s magnetosphere.

Magnetic Clouds are produced in the solar wind when solar eruptions (flaresand coronal mass ejections) carry material off of the Sun along with embeddedmagnetic fields. These magnetic clouds can be detected in the solar windthrough observations of the solar wind characteristics - wind speed, density,and magnetic field strength and direction.

Co-rotating Interactive Regions (CIRs) are regions within the solar windwhere streams of material moving at different speeds collide and interact witheach other. The speed of the solar wind varies from less than 300 km/s (abouthalf a million miles per hour) to over 800 km/s depending upon the conditionsin the corona where the solar wind has its source. Low speed winds come fromthe regions above helmet streamers while high speed winds come from coronalholes. As the Sun rotates these various streams rotate as well (co-rotation) andproduce a pattern in the solar wind much like that of a rotating lawn sprinkler.However, if a slow moving stream is followed by a fast moving stream the fastermoving material will catch-up to the slower material and plow into it. Thisinteraction produces shock waves that can accelerate particles to very highspeeds.

The chemical composition of the solar wind has several interesting aspectsthat hint at physical processes that occur in the solar wind source regions. Thesolar wind composition is different from the composition of the solar surfaceand shows variations that are associated with solar activity and solar features.

Active Sun

In this Section a brief discussion of solar activity and Solar Energetic Particles(SEP) is outlined.

The sunspot cycle

In 1610, shortly after viewing the Sun with his new telescope, Galileo Galileimade the first European observations of Sunspots. Daily observations werestarted at the Zurich Observatory in 1749 and with the addition of otherobservatories continuous observations were obtained starting in 1849. Monthlyaverages (updated monthly) of the sunspot numbers show that the number ofsunspots visible on the sun waxes and wanes with an approximate 11-year cycle(see Figure 1.7). There are actually at least two ”official” sunspot numbersreported:

• The International Sunspot Number is compiled by the Sunspot IndexData Center in Belgium [SIDC 2002].

• The NOAA sunspot number is compiled by the US National Oceanic andAtmospheric Administration [NOAA 2002].


Figure 1.7: Solar activity monitored by sunspot number in the 23th solar cycle.


Figure 1.8: Monthly butterfly diagram showing the positions of the spots for eachrotation of the sun since May 1874.

The Butterfly Diagram

Detailed observations of sunspots have been obtained by the Royal GreenwichObservatory [RGO 2002] since 1874. These observations include informationon the sizes and positions of sunspots as well as their numbers. These datashow that sunspots do not appear at random over the surface of the sunbut are concentrated in two latitude bands on either side of the equator. Aupdated monthly butterfly diagram (see Figure 1.8) showing the positions ofthe spots for each rotation of the sun since May 1874 shows that these bandsfirst form at mid-latitudes, widen, and then move toward the equator as eachcycle progresses.

Solar Flares

Solar flares are tremendous explosions on the surface of the Sun. In a matter ofjust a few minutes they heat material to many millions of degrees. They occurnear sunspots, usually along the dividing line (neutral line) between areas ofoppositely directed magnetic fields.

Flares release energy in many forms - electro-magnetic (γ rays and X-rays),energetic particles (protons and electrons), and mass flows.

Flares are characterized by according to their X-ray brightness in thewavelength range 1 to 8 A. There are 3 categories: X-class flares are big;they are major events that can trigger planet-wide radio blackouts and long-lasting radiation storms. M-class flares are medium-sized; they generally cause


Figure 1.9: Material erupting from a flare near the limb of the Sun on October 10th,1971 (from Big Bear Solar Observatory [BBSO 2002])

brief radio blackouts that affect Earth’s polar regions. Minor radiation stormssometimes follow an M-class flare. Compared to X- and M-class events, C-class flares are small with few noticeable consequences here on Earth. Eachcategory for X-ray flares has nine subdivisions ranging from C1 to C9, M1 toM9, and X1 to X9.

The National Oceanic and Atmospheric Administration (NOAA) monitorsthe X-Ray flux from the Sun with detectors aboard some of its satellites.

Solar flares are often observed using filters to isolate the light emitted byhydrogen atoms in the red region of the solar spectrum (the H-alpha spectralline). Most solar observatories have H-alpha telescopes and some observatoriesmonitor the Sun for solar flares by capturing images of the Sun every fewseconds. Figure 1.9 shows material erupting from a flare near the limb of theSun on October 10th, 1971 (from Big Bear Solar Observatory [BBSO 2002]).

The key to understanding and predicting solar flares is the structure of themagnetic field around sunspots. If this structure becomes twisted and shearedthen magnetic field lines can cross and reconnect with the explosive release ofenergy.

In the hours following a solar flare we often see a series of loops above thesurface of the Sun. These loops are best seen when viewed in the light emittedby hydrogen in the red region of the solar spectrum (H-alpha emission).

The velocity of the material flowing in these loops can be determined usingthe ”Doppler effect”. The light from material moving toward us is shiftedtoward the blue end of the spectrum while light from material moving awayfrom us is shifted toward the red end. Figure 1.10 shows the Doppler shift ofthe H-alpha emission obtained in an active region (AR 7205) flared on June 26,


Figure 1.10: The Doppler shift of the H-alpha emission obtained in an active region(AR 7205) flared on June 26, 1992.

1992. This information can be used with the observed motion of the materialto determine the three-dimensional flow of material within these loops.

Coronal mass ejection

Coronal mass ejections (or CMEs) are huge bubbles of gas threaded withmagnetic field lines that are ejected from the Sun over the course of severalhours. Although the Sun’s corona has been observed during total eclipses ofthe Sun for thousands of years, the existence of coronal mass ejections wasunrealized until the space age. The earliest evidence of these dynamical eventscame from observations made with a coronagraph on the 7th Orbiting SolarObservatory (OSO 7) from 1971 to 1973. A coronagraph produces an artificialeclipse of the Sun by placing an ”occulting disk” over the image of the Sun.During a natural eclipse of the Sun the corona is only visible for a few minutesat most, too short a period of time to notice any changes in coronal features.With ground based coronagraphs only the innermost corona is visible abovethe brightness of the sky. From space the corona is visible out to large distancesfrom the Sun and can be viewed continuously.

Coronal Mass Ejections disrupt the flow of the solar wind and producedisturbances that strike the Earth with sometimes catastrophic results. TheLarge Angle and Spectrometric Coronagraph ([LASCO 2002]) on the Solarand Heliospheric Observatory ([SOHO 2002]) has reported on a large numberof CMEs.

One of the most important CME was in April 7th, 1997. It produced a ”haloevent” in which the entire Sun appeared to be surrounded by it. Coronal massejections are often associated with solar flares and prominence eruptions but


they can also occur in the absence of either of these processes. The frequency ofCMEs varies with the sunspot cycle. At solar minimum we observe about oneCME a week. Near solar maximum we observe an average of 2 to 3 CMEs perday [Reames 1990, Reames 1993, Reames 1995, Gosling 1993, Reames 1998].

Solar Energetic Particles (SEP)

As mentioned before, some of the most powerful flares are strong sources ofhigh energy particles which have been detected soon after the flare as fluxesat the top of the atmosphere.

The energies of particles detected on Earth range from about one to severalhundred of MeV/nucleon. The spectrum of protons can be fitted with a powerlow index γ from 1.4 to 4 till 300 MeV/nucleon. The spectrum of electronsin the range 10-100 keV presents an index of 2.8 with some evidence for asteepening above 100 keV and possibly an even steeper spectrum above 3GeV.

Of particular interest are the element abundances of the solar energeticparticles. The best data are derived from observations of particles withenergies in the range 10-40 MeV/nucleon, although data are available outsidethis energy range. The interesting comparisons are between these elementabundances, those in the interstellar medium and also those of the solarcorona derived from extreme violet and X-ray spectrographic studies. Thesolar energetic particle abundances have been found by averaging many flares.Significant variations in the abundances are anyway found in solar flares,for example in flares with enrichment of heavy ions (”heavy-ion flares”)as well as in flares with changing in the 3He/4He ratio from 10−2 to 1-8[Colgate et al. 1977].

There is reasonable agreement between the coronal and high energy particleabundances, but there are significant differences between these and the localinterstellar ones. This discrepancy can be ordered by the first ionizationpotential of the elements involved. It was found that the elements showinga deficit have first ionization potential greater than about 9.5 eV. The sametrend is apparent in the galactic cosmic rays.

The interpretation of these data is not at all clear. If the ’transition region’of 8-12 eV is interpreted as a measure of the temperature of the region inwhich the particles are accelerated, a temperature of about 8×105 K is found,typical of the top of the chromosphere rather than the photosphere. However,measurements of the charge state of the elements observed in the solar energeticparticles are more consistent with coronal temperatures rather than with valuesless than 105 K, and also, if particles were accelerated in the chromosphere,the abundances would be strongly modified by the effects of spallation as theparticle escapes into the interplanetary medium. Finally, similarities have beenfound between the Solar Wind and the SEP abundances. These argumentssuggest that the correlation of element deficiency with first ionization potentialmay not be associated with the mechanism by which the particles are injected

1.5. FROM THE MAGNETOSPHERE TO THE EARTH 37

or accelerated, but rather may simply reflect intrinsic differences between thecomposition of the photosphere and the corona.

Finally, as it happens in galactic and extra-galactic systems, the problemof accelerating particles in solar flares is still totally open, despite the manyobservations available.

All that suggests the necessity of having more data to be analysed; NINAand NINA-2 missions were in orbit during most of the 23th solar maximum(see Figure 1.7) and are able to give some contribution to the comprehensionof the phenomenon of solar flares.

Solar modulation

The evidence that the solar wind influences the local flux of cosmic rays isillustrated by the inverse correlation between the intensity of the cosmic rayflux at the top of the Earth’s atmosphere and the level of solar activity.

The anticorrelation between solar activity, and the consequent turbolencein the plane of ecliptic, and the flux of cosmic rays at the top of the Earth’satmosphere can be studied by comparing various measures of these threequantities. In Figure 1.11 are showed Sun-spot numbers and cosmic raysfluxes as a function of time. Fewer cosmic rays reach Earth at solar activitymaximum, because the Sun emits plasma and magnetic fields which expel somecosmic rays from the solar system. Thus, the greater the level of solar activity,the more effective the solar wind is in preventing the interstellar flux of cosmicrays from reaching the Earth. This phenomenon is known as solar modulation.

An additional, and not less relevant, evidence of solar modulation isprovided by changes in the shape and in the isotopic abundances of the primarycosmic ray spectrum observed at the top of the atmosphere as a function ofphase of the solar cycle [Longair 1992]. In fact, solar modulation is dependentupon the mass-to-charge (A/Z) ratios of the nuclear species, with the effecton spectra being significant for nuclei of energies up to at least 104 MeV n−1.

1.5 From the Magnetosphere to the Earth

Before entering in the behaviour of cosmic rays in the geomagnetic field, abrief description of the structure of the Magnetosphere is outlined.

1.5.1 Structure of the Magnetosphere

The magnetosphere is a large plasma cavity generated by the Earth’s magneticfield and the solar wind plasma. The streaming solar wind compresses thedayside portion of the Earth’s field and generates a tail which is many hundredsof Earth radii (R⊕ ∼ 6370 kin) long. The basic mechanism for the formationof the magnetosphere is extremely simple: it is a magnetic dipole exposed to astream of charged particles. The entire magnetosphere is subject to only twoboundary conditions, explicitly the boundary between the magnetosphere on


Figure 1.11: Solar activity peaks every 11 years when sunspot number reaches amaximum. Fewer cosmic rays reach Earth at these times, because the Sun emitsplasma and magnetic fields which expel some cosmic rays from the solar system.The cosmic ray data were recorded by the Inuvik neutron monitor [Inuvik 1999].

1.5. FROM THE MAGNETOSPHERE TO THE EARTH 39

Figure 1.12: Sketch of the structure of the magnetosphere.


the streaming solar wind and the boundary of the plasma in the ionosphere.The basic elements of the magnetosphere (see Figure 1.12) are

• The Bow Shock and the Magnetosheath.

While not part of the magnetosphere proper the magnetosheath is anouter layer embedding the magnetosphere. The solar wind plasma travelsusually at super-fast speeds relative to the magnetosphere. Therefore astanding shock wave forms around the magnetosphere just as in front ofan aircraft traveling at supersonic speeds. The bow shock is the shock infront of the magnetosphere and the magnetosheath is the shocked solarwind plasma. Therefore it is not directly the solar wind plasma whichconstitutes the boundary of the magnetosphere but the strongly heatedand compressed plasma behind the bowshock.

• The Magnetopause.

The magnetopause is the actual boundary between the shocked solarwind and the magnetospheric plasma. However, the magnetosphereis not closed in terms of the magnetic field but there is considerablemagnetic flux crossing the magnetopause. Thus it is not easy to definethis boundary precisely. Also the boundary does permit a certain amountof solar wind plasma entry. This entry is easier along magnetic field lines.The magnetopause is an highly important region because the physicalprocesses at this boundary control the entry of plasma, momentum,energy and the redistribution of geomagnetic flux.

• Cusp and Mantle.

The cusp and mantle regions are directly adjacent and inward of themagnetopause. The cusp is the region where dipolar field lines convergeand in a two-dimensional case represents a field line which goes intoa singularity with

−→B = 0. The mantle region represents a boundary

to the magnetotail usually filled with solar wind plasma but with astretched magnetospheric magnetic field. The role of the cusps is notfully understood but it is a region where highly energetic particle can beproduced and it is very active in terms of turbulence and wave energybecause the boundary field lines converge in the cusp and all waves whichtravel along the magnetic field are channeled into this region.

• The Quiet Magnetotail.

The magnetotail is the long tail-like extension of the magnetosphere onanti-sunward side of the magnetosphere. Since the magnetic field pointstoward the Earth in the northern lobe and away in the southern lobethere is a current in the westward direction. Because of its structure thereis considerable energy stored in the magnetic field in the magnetotail.During magnetically quiet times convection is typically low and energyin the plasma flow is only a tiny fraction of the overall energy density.

1.6. THE EARTH’S MAGNETIC FIELD 41

• The inner Magnetosphere.

The inner magnetosphere is different from most of the magnetospherein that the magnetic field is mostly dipolar and perturbations of thefield are small compared to the average dipole field. However, therecan still be large amounts of energy stored in this region in particularduring so-called storm times. During such times the ring current (currentdue to gradient curvature drifts of charged particles) intensifies stronglyand is responsible for strong magnetic perturbations at low geomagneticlatitudes on the Earth.

• Magnetosphere - Ionosphere Coupling.

The ionosphere is the region where the atmosphere is partially ionizedand plasma and neutrals strongly interact. This interaction exerts adrag on the plasma. The plasma density can be very high but alsostrongly variable such that the ionospheric conductance can vary byorders of magnitude. Magnetospheric plasma motion is transmitted intothe ionosphere and forces ionospheric convection. This also implies theexistence of strong currents along magnetic field lines which close throughthe ionosphere. In particular at high latitudes these currents lead tomagnetic perturbations during times of strong magnetospheric activity(fast convection and changes of the magnetospheric configuration)[Otto 2001].

1.6 The Earth’s magnetic field

It is known that the Earth’s magnetic field to a first approximation, at andclose to the planet’s surface, can be descibed as a dipole field. In sphericalcoordinates (if we choose the origin of the coordinates at the dipole center and

the polar axis opposite to−→M points south), the magnetic flux density is given

by:

B(r, λ) =M

r3(1 + 3sin2λ)

12 , (1.10)

where−→M is the dipole moment, r the geocentric radial distance and λ the

magnetic latitude (see Figure 1.13). For the Earth M8.1x1025 Gauss cm3,that corresponds to an equatorial magnetic field of Beq0.31 Gauss.

The components of the field in the r and λ directions are:

Br = −2Msinλ/r3 , (1.11)

Bλ = Mcosλ/r3 . (1.12)

Thus

Bλ

Br

=rdλ

dr= − 1

2tanλ


Figure 1.13: Dipole geometry.

and, by integration,

r = r0cos2λ , (1.13)

we obtained the field line for a dipole field.It is convenient to use the radius of the Earth, R⊕, as the unit of distance.

Putting r/R⊕ = R,

B(R, λ) =0.31

R3(1 + 3sin2λ)

12 Gauss , (1.14)

0.31 Gauss (=3.1×10−5 Wb/m2) being the flux density at the Earth’s surfaceat the magnetic equator. In these terms, the field-line equation becomes

R = R0cos2λ , (1.15)

R and R0 being measured in Earth-radii. The latitude where the field-lineintersects the Earth’s surface is given by

cosλ⊕ = R− 1

20 . (1.16)

The module of the field | −→B | along each field line has its minimum value onthe equatorial plane (λ=0) at the point with the largest distance from dipolecenter (R=R0).


It is well known the geomagnetic field is significantly different from theexact dipolar form. The deviations of the field from the dipolar form areessential for an understanding of the properties of the sub-cutoff fluxes.

The sources of the magnetic field can be naturally divided into ”internalsources” (electric currents inside the Earth) and ”external sources” (electriccurrent in space). The contributions to the field of the external sources exhibitsvariations also with very short time scale (of the order of hours), connectedwith the position and magnetic activity of the Sun, while the contribution ofthe ”internal sources” varies only in much longer time scales, with a seculardrift of the magnetic poles. The magnetic field due to the external sourcesis the dominant contribution to

−→B at a distance of several Earth’s radii,

but represents only a small perturbation in the vicinity of the Earth, wherethe geomagnetic field will be described as in the International GeomagneticReference Field (IGRF) model that is an empirical representation based on amultipole expansion [IGRF 2002]. The coefficients of the different multipoleterms are called Gauss coefficient and are slowly time dependent.

If we want to describe the geomagnetic field with a simple dipole, we obtaina significant better fit with a dipole that is not only ”rotated” with respectto the Earth’s axis, but is also ”offset” that is it has an origin that doesnot coincide with the Earth’s center (the poles are at geographic coordinates79N, 70W, and 79S, 110E). In the standard expansion of the field usedin the IGRF model the origin of all multipole terms is the Earth’s center,but is possible to reabsorb the quadrupole contributions redefining the dipolemoment and the position of its center. There is no unique well defined way toperform this redefinition of the dipole and different algorithms have been usedfor different applications.

The effect is quite clear. The dipole offset for the Earth is of order ∼ 450km and a vector from the Earth’s center to the dipole center has a latitude ∼18o and a longitude ϕdipole ∼ 140o [Jackson 1998]. The approximate offset ofthe dipole axis with respect to the Earth’s center is of crucial importance forthe understanding of the secondary particle fluxes spectra observed by NINAand NINA-2.

1.6.1 Trapped and albedo particles

Before describing the properties of charged particle trajectories in a magneticfield it is useful to spend some words on the differences between ”real” trappedand albedo particles. As we said heretofore cosmic rays are fast positiveions bombarding Earth from all directions. When they smash into nuclei ofatmospheric gases, fragments go flying off in different directions. Most suchfragments are absorbed by the atmosphere or by the ground, but a few aregotten upwards, out of the atmosphere and into space.

The general picture of the interaction of a high energy proton (the mostabundant specie) with a atmospheric nucleus can be described by the followingempirical rules [Longair 1992]:


Figure 1.14: A schematic diagram showing the principal products of the collision ofa high energy proton with a nucleus.

• the proton interact violently with an individual nucleon in a nucleus and,in the collision, pions of all charges (π+π−π0) are the principal products(mostly in the forward and backward directions). Strange particles andlight nuclei may also be produced and occasionally antinucleons as well.

• Each of the secondary particles is capable of initiating another collisioninside the same nucleus, provide, of course, that the initial collisionoccurred sufficiently close to the ’front edge’ of the nucleus.

• Only one or two nucleons participate in the nuclear interactions withthe high energy particle, but they are generally removed from thenucleus leaving it in a highly excited state and there is no guaranteethat the resulting nucleus is a stable species. A number of things mayhappen. What often happens is that several nuclear fragments (spallationfragments) evaporate from the nucleus. These fragments are emitted inthe frame of the nucleus which is not given much forward momentum inthe nuclear collision, virtually all of it going into tearing out nucleonswhich interact with the high energy particle. Therefore, these spallationfragments are emitted more or less isotropically in the laboratory frameof reference. Neutrons are also evaporated from the ravished nucleusand other neutrons may be released from the spallation fragments. It isto remember that for light nuclei any imbalance between the numbersof neutrons and protons is fatal. These processes are summarizeddiagrammatically in Figure 1.14.

If secondary particles are electrically charged, they will often end uptrapped by the Earth’s magnetic field. None of these however lasts very long,


Figure 1.15: Illustration of Cosmic Ray Albedo Neutron Decay mechanism. Aprimary cosmic ray interact with atmosphere and create a secondary neutron whichtravel enough to decay in proton which is trapped by a field line.

since trapped orbits which rise from the atmosphere must sooner or later enterin the atmosphere again. In fact the up-ward moving secondaries do not escapefrom the Earth’s influences completely but are brought back again by theEarth’s magnetic field. In effect they are tied to lines of force and reappear inthe opposite hemisphere.

As said above, some of the fragments can be neutrons. In this case,Cosmic Ray Albedo Neutron Decay (CRAND) mechanism happens [Hess 1968,White 1973, Kanbach et al. 1974]. Having no electric charge, neutrons are notaffected by the Earth’s magnetic field and usually escape into space.

The free neutron decays in mean life time τn = (887.0 ± 2.0) s in

n → p + e− + ν.

τn is a fairly long time for a fast particle, time enough for many neutronsto get halfway to Mars. However, decay times are spread out statistically andwhile 887 s is the average, a few neutrons decay quite soon, while still insidethe Earth’s magnetic field. The protons from decay are grabbed by the Earth’smagnetic field often on trapped orbits which do not return on the atmosphere,in which protons can stay trapped for a rather long time and can interact withresidual atmospheric helium and oxygen at altitude of some hundred kilometersto create subsequent secondary particles (protons and light isotopes).


1.6.2 Charged particles in a magnetic field

A rich literature is available describing the properties of charged particletrajectories in a magnetic field. In this section I will only describe somesimple outcomes that can explain some characteristic of second fluxes. Irefer to Jackson [1998], Rossi and Olbert [1970], Parks [1991] for a completeexplanation of gyro, bounce and drift motion.

It is well known that in a static homogeneous magnetic field the motion ofa charged particle is a helix, obtained by the combination of a rotation in theplane orthogonal to the field line and an uniform motion along the field line.The gyroradius a is

a =mv⊥qB

, (1.17)

or, for Z=1:

a(km) = 33.3p⊥(GeV )

B(Gauss); (1.18)

the frequency of the circular motion is

ωg =qB

m(1.19)

and the angle of the particle velocity with the magnetic filed is called pitchangle

α = arctgv⊥v‖

. (1.20)

In a non uniform static magnetic filed where the distance scale L of thefield variation L ∼ | B∂xj/∂B | is much larger than the gyroradius a (L>>a),the motion of a charged particles can again be decomposed as the rotation ina plane orthogonal to the field lines around a point called ”guiding center”that has a motion both along and across the field line. The motion parallelto the field lines is controlled by the variation of the field intensity along thefield line. This behaviour can be deduced from the adiabatic conservation ofthe magnetic flux (πa2B) through a particle’s circular orbit. From 1.17 andconsidering Z=1, one obtains that the conservation of the magnetic flux canbe written in the form:

v⊥B

=v⊥o

Bo

. (1.21)

Using the fact the | v | is constant, one can deduce the equation

∂v‖∂l

− v2⊥o

2Bo

∂B(l)

∂l, (1.22)

(l is the distance along the field line) that describe the motion parallel tothe field line. At the increasing of B along the field line a ”repulsive effect”happens; this is at the basis of the ”magnetic mirror” effect.


Figure 1.16: Helical ion trajectory in a uniform magnetic field.

When the gradient of the field has a non vanishing component ∇⊥B = 0,or where the field line are curved, the guiding center has also a ”drift” motionorthogonal to the field. Particles of different charge drift in opposite directions.

Influence of the Earth’s magnetic field

Charges particles approaching the Earth from outer space follow curvedtrajectories because of the geomagnetic field in which they propagate. Asthey enter the atmosphere they may also be subject to interactions withatmospheric constituents. Disregarding the existence of the atmosphere, thequestion whether a particle can reach the Earth’s surface or not depends solelyon the magnitude and direction of the local magnetic field, and on energy,charge and direction of propagation of the particle.

We can rewrite the expression of the gyroradius of a charged particle in astatic magnetic field B in the more general case of a particle of mass m, chargeze and Lorentz factor γ = (1 − v2/c2)−

12 , approaching the magnetic field

rB =γmv

ze

sinα

B=(

pc

ze

)sinα

Bc, (1.23)

where p is the relativistic three-momentum of the particle. This means that,if we inject particles with the same value of pc/ze into a magnetic field B atthe same pitch angle α, they have exactly the same dynamic behaviour. Thequantity pc/ze is called the magnetic rigidity of the particle. It has dimensionsof volts; a useful unit for practical purposes is gigavolts (GV). In cosmic raystudies, the energies of cosmic rays are often quoted in terms of their rigiditiesrather than their energies per nucleon.


Figure 1.17: (a) The coordinate system in which the dynamics of high energyparticles in a dipole magnetic field are defined; (b) the coordinate system viewedfrom above showing the definition of the angle θ.

Charged particles are propagated to the surface of the Earth through theEarth’s magnetic field, which is clearly influencing their trajectory. We hadalready derived the equations of the field’s components in r and λ directions(equations 1.11 and 1.12), in the case of a dipole field. There is no generalsolution to the problem of charged particles approaching this magnetic field,but the theory enables limits to be set to those regions of the magnetic fieldconfiguration which are accessible to incoming particles. Therefore we candefine permitted and forbidden zones for any particle approaching from infinity.

We can simplify the expressions 1.11 and 1.12 first introducing

rs = (zeM/p)12 (1.24)

which is one Størmer unit of distance for a particle of momentum p.Moreover, we can express again the equations 1.11 and 1.12 in terms of

the coordinates r, λ and ω as shown in figure 1.17 and introduce the angle θ,which is the angle between the instantaneous velocity vector v of the particleand the meridian plane which follows the particle in its orbit. In these newcoordinates, the solution of the equation of the motion can be written in thefollowing form:

2b = −rsinθcosλ − cos2λ

r, (1.25)

where r and b are measured in Størmer units, therefore keeping inside theinformation about the momentum of the particle. The term b is a constant ofthe motion and, when r becomes very large, −2b/cosλ is an impact parameterof the particle. If the motion is in the equatorial plane, λ=0o, the collisionparameter is −2b; at other geomagnetic latitudes, −2b/cosλ is the collision


parameter with respect to the dipole axis. The attraction of this analysis isthat we know that sinλ must lie between 1 and -1 and this sets limits toaccessible ranges of r and λ for particles with a given value of b.

In particular, developing the calculations it can be seen that only particleswith momenta p greater than a certain value can reach the Earth at differentlatitudes. Putting in some numbers, for instance, the geomagnetic cutoffmomenta for Z=1 particles at different latitudes are

Kinetic energyλ = 0 cp ≥ 14.9 GeV 14.0 GeV

Z = 1 λ = 40 cp ≥ 5.1 GeV 4.3 GeVλ = 60 cp ≥ 0.93 GeV 0.48 GeV

Thus, at each point on the Earth’s surface, there is a threshold energy,called geomagnetic cutoff, below which the particles cannot reach the surface[Longair 1992].

Geomagnetic cutoff

A practical measure to compare and interpret particle measurements made atdifferent location on the Earth, in particular at different geomagnetic latitudes,is the effective vetical cutoff rigidity, Pc. It must be enphatized that in generalgeomagnetic and geographic coordinates are not the same and Pc depends onlocation and time.

Moreover, the exact vertical cutoff rigidity of a particular geographiclocation varies somewhat with time because of the variability of themagnetospheric and geomagnetic fields. These variations may be as muchas 20% at mid-latitude. Figure 1.18 shows the Stormer cutoff rigidity as afunction of geomagnetic latitude for vertically incident positive particles andfor positive particles under a zenit angle of 45o from the East and West.

A frequently used method to compute the vertical cutoff rigidity of aparticle is to consider an identical particle of opposite charge and oppositevelocity being released in radial outward direction at the reference altitude of20 km above sea level. The effective cutoff rigidity is defined as the rigidityrequired for the particle to overcome trapping in the geomagnetic field andbeing able to escape to infinity. All kinds of interactions and energy lossmechanisms in the residual atmosphere as well as scattering are disregardedin this picture.

In a dipole field (and perhaps also in the real geomagnetic field) directaccess for particles of all rigidity values lower than the Stormer cutoff rigidityis forbidden from outside the field. In a dipole approximation the Stormercutoff rigidity (in GV) can be written as


Figure 1.18: Stormer cutoff rigidity as a function of geomagnetic latitude, λ, forvertically incident positive particles (θ=0o), and for positive particles under a zenitangle of 45o from the East and West.

Pc =M cos2(λ)

R2[1 + (1 − cos3(λ) cos(ε) sin(ζ))12 ]2

, (1.26)

where M is the dipole moment which has a normalized value of 59.6 whenR is expressed in unit of Earth radii, R being the distance from the dipole,λ is the geomagnetic latitude, ε the azimuthal angle measured clockwise fromthe geomagnetic East direction (for positive particles), and ζ is the angle fromthe local magnetic zenith direction [Grieder 2001].

Particles trajectories

It is already said that flux of cosmic rays observed ”near the Earth” atdifferent latitudes are different, with flux measured close to the magneticequator strongly suppressed with respect to the flux measured at high magneticlatitudes. The ”latitude effect” was at the basis of the discovery that most, orat least part, of ”cosmic radiation” was of charged particles [Compton 1932,Compton 1933]. In the following years other works carried simultaneously byRossi and Compton showed that geomagnetic effect should produce the East-West (E-W) effect [Rossi 1933, Johnson 1933]. On the basis of the Compton’sown analysis of the data the greatest E-W differences appeared in latitudesbetween 20o and 30o. The results accorded with the Lemaitre-Vallarta theoryand showed that the principal corpuscolar component of the cosmic radiation ispositively charged [Lemaitre and Vallarta 1933, Alvarez and Compton 1933].

The latitude and the E-W effect can be simply qualitatively understoodtaking in account that low rigidity particles from outer space cannot reach the


Earth’s surface becouse of geomagnetic field. The past trajectory of a particleof charge Z, momentum −→p , detected at the position −→x , can be determinatedintegrating the classical equations of motion for a charged particle in anelectromagnetic field. There are three possible results:

1. the particle was at very large distances from the Earth in the past

2. the particle has a trajectory which remains confined in the volumeR⊕<r<∞ without ever reaching ∞

3. the particle has a trajectory which originates from the Earth’ssurface;

R⊕ ( 6371.2 km) is the Earth’s radius and the particles originated inthe atmosphere are included in 3. To primary cosmic ray particles 2 and 3trajectories are considered ”forbidden” because no primary cosmic ray canreach the Earth from a large distance traveling along one of them. ”Allowed”trajectories belong to the 1 [Lipari 2002].

If we consider an exactly dipolar field filling the entire space, for a fixedposition −→x and direction n, the trajectories of all charged particles with rigiditysmaller then geomagnetical cutoff are forbidden; all charged particles withrigidity greater then geomagnetical cutoff are allowed. I want to remark thatwith ”forbidden” trajectories I mean only that primary cosmic rays are no ableto populate this regions.

1.6.3 Adiabatic invariants

The concept of invariants is extremely powerful in describing the motionof particles in magnetic fields [Parks 1991]. A motion of a particle can bedescribed by a pair of variables (pi,qi) that are generalized momenta andcoordinates. Then, for each coordinate qi that is periodic, the action integralJi

Ji =∫

pidqi , (1.27)

integrated over a complete period of cycle (oscillation) of qi with specifiedinitial conditions is an invariant (constant) of motion.

This action integral will remain invariant even if some property of thesystem is allowed to change. However, it is required that the change be slow(adiabatic change) as compared to relevant periods of the system and thechange must not be ralated to the periods.

For particles in magnetic fields, an adiabatic invariant is associated witheach of the three types of motion discussed in § 1.6.2: the gyration motionaround

−→B , the longitudinal motion along

−→B and the drift motion perpendicular

to−→B .In next three sections I summary the properties, the correlations and the

physical consequences of all adiabatic invariants, deriving rigorously the firstone.


First adiabatic invariant

The first adiabatic invariant is associated with the gyration motion and withthe existence of mirror points.

We derive the first invariant by use of a classical argument. It is knownthat the force on a particle along

−→B can be written in absence of electric field

as

F‖ =mdv‖

dt= −µ

∂B

∂s= −µ∇‖B ; (1.28)

Multiply each side of 1.28 by v‖ it became

d

dt

mv2‖

2= −µ

∂B

∂s

ds

dt= −µ

dB

dt. (1.29)

The last step follows because−→B is not time dependent and dB/dt =

∂B/∂t + (−→v · ∇)B = v ∂B∂s

.

Conservation of the total energy of the particle requires

d

dt(1

2mv2

‖ +1

2mv2

⊥) =d

dt(1

2mv2

‖ + µB) (1.30)

where use was made of the definition µB = mv2⊥/2. Now combine 1.29 and

1.30 and obtain

−µdB

dt+

d

dt(µB) = 0 . (1.31)

The differentiation of the second term results in

Bdµ

dt= 0 . (1.32)

Since B = 0, equation 1.32 states that the geomagnetic moment µ isindependent of time and is a constant in the guiding center motion.

In the presence of an electric field and of a corresponding change of∂−→B/∂t = −−→∇ × −→

E the magnetic moment is still conserved as long as thechanges of the magnetic field are slow compared to the gyro motion.

The constancy of the magnetic moment implies that the total magneticflux enclosed by the motion must also remain constant (see § 1.6.2):

Φµ = πa2B = πm2v2

⊥q2B2

B =2πm

q2µ . (1.33)

The constancy of the magnetic moment has two interesting consequencesthat will outline hereinafter.


Figure 1.19: Illustration of magnetic mirror motion.

Magnetic mirror

With the pitch angle α for the particle trajectory relative to the magnetic fieldorientation the magnetic moment becomes

µ =mv2 sin2 α

2B(1.34)

and in absence of parallel electric fields (such that the energy is conserved)the pitch angle at two different location in the magnetic field must satisfy

sin2 α2

sin2 α1

=B2

B1

. (1.35)

This provides the information of the change of the pitch angle along theentire field line if it is known in one location. In particular we can compute thecondition for which the pitch angle becomes 90o, i.e., the condition for whicha particle is reflected in the magnetic field. Assuming the field strength Bm atthe point where a particle is mirrored we obtain anywhere on the field line

sin α =

√B

Bm

. (1.36)

In other words at a given location particles with a pitch angle smaller thanα will be transmitted whereas particles with a larger pitch angle are mirrored[Baumjohann and Treumann 1997].

This effect is important for many laboratory and space plasmas. Particlesare confined in a magnetic mirror by this mirror force. Particle with a smallerpitch angle are lost from the magnetic field such that the distribution function


miss this portion of phase which is called the loss cone. The sample principlegoverns the magnetosphere. Ions and electrons which penetrate into the lowerionosphere can undergo collisions with the neutrals and thereby are lost fromthe magnetospheric population.

Adiabatic heating

Drift motion can bring particles into regions of the magnetosphere with a largermagnetic field strength. The conservation of the magnetic moment implies

E⊥1

B1

=E⊥2

B2

. (1.37)

Thus particles can gain perpendicular energy (without a change of theparallel energy) by adiabatic motion.

Second adiabatic invariant

The longitudinal invariant is associated with the v‖ motion (or in other wordswith the bounce motion). Since charged particles traveling in the directionof B behave as if the magnetic field is not there, the canonical momentum issimply mv‖. The action integral for this motion is usually represented by J

J =∫

mv‖ds , (1.38)

where ds is an element of the guiding center path along−→B and the integral

is evaluated over a complete path. Some definitions of J exclude the mass m ofthe particle in the integral, which is permitted in non-relativistic formulation.

In terms of an average parallel velocity 〈v‖〉 the invariant (in a cycle) isJ = 2ml〈v‖〉 with l being the length of the entire field line between the mirrorpoints. The square of this invariant implies for the parallel energy

〈E‖〉2〈E‖〉1

=l21l22

. (1.39)

Therefore as the length of the field line between mirror points changes, sodoes the parallel energy.

Third adiabatic invariant

As explained above the first invariant is associated with the cyclotron motionand the second with the longitudinal motion. It would seem natuaral thento conclude that there must be an invariant associated with the drift motion.Physically, one can argue that the total amount of flux enclosed in a gyrationremains constant. The first invariant implies that the total amount of fluxenclosed in a gyration remains constant. We carry this concept to the driftmotion. The guiding center drift motion conserves the total magnetic fluxwithin its drift path. The actual path may be very complicated, but a particle


will always follow a path to enclose the same magnetic flux. The third invariantis conserved as long as the perturbation time scale is longer than the drifttimes of the particles. Symbolically, the third invariant is J =

∫mv⊥dφ and

dJ/dt = 0. Here φ is the azimuthal angle. For a complete drift path, the limitsof φ are from 0 to 2π [Northrop 1961, Parks 1991].

Violation of adiabatic invariants

Adiabatic invariants require that temporal changes of a configuration is slowcompared to the period associated with the invariant. The ralation

ωg >> ωb >> ωd (1.40)

(with obvious symbols) establishes a hierarchy of temporal scales.Configuration changes with τ 1/ωd destroy the third invariant while thefirst two are still conserved. Changes of order τ 1/ωb destroy the secondand the third invariant and leave only the first invariant conserved. If changesoccur on the gyro period scale all invariants are destroyed [Otto 2001].

The bouncing motion

The bouncing motion is the oscillation of the guiding center of the particletrajectory between two mirror points placed symmetrically in the north andsouth hemisphere. Only oscillation with a sufficiently small amplitude arepossible, because following a field line, from a point E in the equatorial planethe radius decreases, and therefore if the amplitude of the oscillations are toolarge a particle ”hits” the surface of the Earth and is absorbed.

In this section, conditions on the maximum amplitude for bounce motion isoutlined. We want to show how these conditions can be translate in conditionson the pitch angle α0 of the trapped particles when they are on magneticequatorial plane. We will show that the condition will be rigidity independent.The magnetic field line that passes through the point E on magnetic equatorialplane exists from the surface of the Earth in a point G1 in the southernhemisphere to a point G2 in the northern hemisphere. For a centered dipole thepoints G1 and G2 have the same magnetic longitude and symmetric latitudesas describe by equation 1.16. The value of | −→B | along the field line has itsminimum at the point E in the equatorial plane and grows monotonicallywith the distance from E.

A charged particle at the point E with equatorial pitch angle α0 thatis not exactly 90o will have a component of momentum parallel to the fieldp‖ = p cos α0 and moves along the field line; the gradient of the field along theline will reduce and finally invert the parallel component in the mirror point.The component of the momentum parallel to the field at a point P along theline (∝ cos α) depends of the value of the field at that particular point. Fromequation 1.21 we obtain


sin2 α = sin2 α0B

B0

, (1.41)

where B0 ≡ B(E), and easily

cos2 α = 1 − sin2 α = 1 − sin2 α0B

B0

. (1.42)

The mirror points are by definition the points where velocity has onlyperpendicular component, that is:

0 = 1 − sin2 α0B(M1,2)

B0

, (1.43)

where M1 and M2 are the mirror points. The requirement that the twomirror points are above sea level can be written as:

| sin α0 |≥ [B0/B(G1)]1/2 , | sin α0 |≥ [B0/B(G2)]

1/2 ; (1.44)

the two condition are identical for a centered dipole field. Substituting theexplicit expressions one obtains the condition:

| sin α0 |≥ R− 5

40 (4R0 − 3)−

14 , (1.45)

where R0 is the radius of the equator point in units of R⊕.

Consequences of an off-set dipole field

The most powerful mathematical instrument to describe the motion of trappedparticles in the Earth magnetic field is the concept of ”magnetic shell” 1.

In this section this concept is discussed at first in the case of dipole field,at second in the real geomagnetic field.

As discussed above, the motion of charged particles trapped in a dipolefield can be regarded as the superposition of a circular motion in a planeperpendicular to the local magnetic field, around a ”guiding center”, anoscillation (much slower then the first) of the ”guiding center” along a guidingline that correspond to a field line, and a rotation (much slower then thesecond) of the guiding line around the polar axis. The motion of the ”guidingcenter” defines therefore a surface that can be called ”magnetic shell”. Eachone of this shells corresponds to the surface generated by the rotation aroundthe dipole axis of a field line. Explicitely, the magnetic shells have the formof equation 1.15 and be labeled with the parameter R0. The set of the mirrorpoints M and M∗ where each individual particle ”bounces” have a constantvalue of the magnetic field B.

1McIlwain introduced in 1961 the (B,L) coordinate system, in which to each points ofthe space are univocally connected a (B,L) couple values. Given a point P of the space,B and L corresponds respectively to the value of the geomagnetic field in P and to theequatorial radius of the drift shell passing for P measured in Earth’s radii [Hess 1968].


In the real geomagnetic field the motion of trapped charged particles hasqualitatively the same structure as in the dipole field. To good approximationthis motion can again be regarded as the superposition of three motion of thesame physical result of the ones described above. The decomposition of themotion of trapped particles in three quasi-periodic components is connected tothe existence of three adiabatic invariants for the hamiltonian of the system.The motion of the guiding center can again be analyzed as an oscillation alonga field line between mirror points that have a constant value of the magneticfield, and a slower drift in longitude. I refer to Rossi and Olbert [1970] for thedemonstration of the process for which a charged particle that starts oscillatingalong a particular field line, after drifting in longitude through 360o, returnsto the field line from which it started. Therefore the set of field lines alongwhich a particle oscillate defines again a surface that closes upon itself as inthe case of the dipole field. This is called L − shell.

The motion of a trapped charged particle is therefore confined to a welldefined surface, that is a part of an appropriate magnetic shell, where themagnetic field is lower than a maximum value B∗ with the set of points onthe shell where the field has the value B∗ corresponding to the set of mirrorpoints for the particle trajectory. Thus the set of all possible trajectories canbe classified according to the two parameters (L and B∗) that defines the shelland the set of mirror points on the shell.

Some secondary particle trajectories in the Earth’s field

In this section some examples of trajectories for secondary particles generatedin the interaction of cosmic ray proton in atmosphere at an altitude of ∼50 km will be described. The trajectories have been calculated integratingnumerically the classical equation of the motion: d−→p /dt = e

−→β ∧ −→

B in theIGRF field [IGRF 2002]. We explain hereinafter three examples of trajectoriesin the geographical coordinates.

The example of Figure 1.20 shows the trajectory of a particle of rigidity 5.0GV, remaining close to the equatorial plane of the field. It can be noted thatthe guiding center of the trajectory travels in an approximately circular motionwith a center that does not coincide with the Earth’s one, as a consequence ofthe ”off-set” of the dipole component of the geomagnetic field. The example ofFigure 1.21 shows the trajectory of a particle of rigidity 1.16 GV, traveling fora pathlength of 106 km. It performs approximately 30 bounces and drifts formore than 2π. The example of Figure 1.22 shows the trajectory of a particleof rigidity 2.31 GV, following the field line and bouncing only two time, beforebeing absorbed. The reason is that the subsequent bouncing point would beinside the Earth [Lipari 2002].

As discussed in previous sections, particles conditions on number of bouncesor on time of live totally correspond to a pitch angle condition.


Figure 1.20: Top panel: projection of trajectory of a secondary proton of rigidity 5.0GV in a geomagnetic field in the Earth’s equatorial plane. Bottom panel: projectionin the plane (

√X2 + Y 2, Z)


Figure 1.21: Top panel: projection of trajectory of a secondary proton of rigidity1.16 GV in a geomagnetic field in the Earth’s equatorial plane. Bottom panel:projection in the plane (

√X2 + Y 2, Z)


Figure 1.22: Top panel: projection of trajectory of a secondary proton of rigidity2.31 GV in a geomagnetic field in the Earth’s equatorial plane. Bottom panel:projection in the plane (

√X2 + Y 2, Z)

Chapter 2

NINA ad NINA-2 detectors

After a discussion on scientific capability of NINA and NINA-2 missions, thefirst part of chapter will be focus on the description of the detectors and triggerlogic; in the second part telescopes performances during the test-beam sessions,including energy calibration, energy resolution and algorithms used for nuclearand isotope identification will be outline.

2.1 NINA experiments: scientific overviews

NINA has been built in order to investigate the nuclear and isotopiccomposition of low energy cosmic particles. The low-altitude polar orbits ofthe missions (about 1.1 Earth-radii) are particularly suitable for performingobservations of particles of different origin while traversing regions of differentgeomagnetic cutoff. The Earth’s magnetic field is utilized as a separator.According to the coordinates along the orbit where the particles are detected,it is possible to make inferences about their origin which can be galactic, solar,albedo or trapped.

• Galactic Cosmic Rays (GCR) are a directly accessible sample ofmatter coming from outside the Solar System. The GCR energyspectrum can be well represented by a power-law energy distribution forenergies above 1 GeV n−1, but at lower energy shows a strong attenuationdue to the interaction between the Solar Wind and the cosmic particles.This is one of the reasons why GCR investigations below 200 MeV n−1

have been relatively scarce in the past.

NINA and NINA-2 flight during the period of medium and maximumof solar activity. Due to their technical characteristics and their goodenergy, mass and angular resolution, the telescopes are particularlysuited for exploring the low energy component of the cosmic radiation.The detectors recorded GCRs of very low energy (from 10 up to 200MeV n−1) in the polar sectors of the orbit, where geomagnetic effects arevirtually negligible.

61

62 CHAPTER 2. NINA AD NINA-2 DETECTORS

• The most complete measurements of elemental abundances in the solarcorona come from measurements of high energy particles accelerated inthe large Solar Energetic Particle (SEP) events.

Initially it was thought that the energetic particles in large SEP eventswere accelerated only in solar flares. In recent years, however, it hasbecome clear that in the large gradual events particles are accelerated atshock waves that are driven out from the Sun by Coronal Mass Ejections(CMEs). These shocks accelerate the ions of the chemical elements in afairly equivalent manner. In contrast, particles accelerated in impulsivesolar flares show specific elemental enhancements produced by resonantwave-particle interactions during stochastic acceleration of the ions fromthe flare plasma.

New spacecraft observations can extend Solar Energetic Particlemeasurements to heavier elements, to rarer elements and to isotopes.NINA and NINA-2 performed SEP observations in the polar sectors ofthe orbit. Their good mass discrimination helped in the determinationof SEP composition and therefore in the comprehension of the sourcesand acceleration mechanisms involved.

• High energy cosmic rays interacting with the upper atmosphere canproduce secondary particles, which, if not absorbed by the atmosphere,can move along geomagnetic field lines between mirror points driftinglongitudinally around the Earth. Depending on their pitch angle, suchparticles can only make one bounce (and are called albedo particles) ormore than one bounce (quasi-trapped particles) in the geomagnetic field.The downward flux of these particles has been detected by NINA andNINA-2 instrument during the passages in equatorial regions.

• Crossing the South Atlantic Anomaly, telescopes NINA and NINA-2 have been able to detect geomagnetically trapped particles. Theradiation belt component derives mainly from CRAND mechanism andsubsequential interactions of trapped protons with residual atmospherichelium and oxygen at an average altitude of 200-300 km.

2.2 The NINA instrument

The NINA instrument consists of 4 subsystems:

1. the detector (box D1), composed (see Figure 2.1) of 32 silicon layersand the electronics for signal processing,

2. the on-board computer (box D2), a dual microprocessor dedicatedto data processing, selection of the trigger and acquisition modeconfiguration,

2.2. THE NINA INSTRUMENT 63

Figure 2.1: Sketch of D1 box.


3. the interface computer (box E), which rearranges the data comingfrom the on-board computer and delivers them to the satellitetelemetry system,

4. the power supply (box P ), which distributes the power supply tothe different subsystems.

The weight and electric power of the complete telescope are respectively40 kg and 40 W, in accordance with the constraints imposed by the satellite.To safeguard from possible malfunctions and breaks, all electronic systems areglobal redundant.

The detector of NINA (box D1), manufactured by the Italian companyLaben, is composed by 16 silicon planes. Every plane consists of a pair of n-type silicon detectors, 60×60 mm2, each read out by 16 strips mounted back toback with orthogonal orientation, in order to measure the X and Y coordinatesof the particle. The strip pitch is 3.6 mm.

The thickness of the first pair of detectors, composing the first plane ofNINA, is (150 ± 15) µm; all the others, instead, are (380 ± 15) µm thick, for atotal thickness of 11.7 mm of silicon in the whole detector. The indeterminationin the total silicon thickness comes from the process of manufacturing, and isgreater (in percentage) for the first two thinner layers. In order to reduce to theminimum the thickness of dead area interposed between the silicon detectors,a special ceramic (Al2O3) frame, passing only under the lateral strips 1 and16 and 625 µm thick, has been utilized. The role of the ceramic is to sustainthe single silicon structures and connect them mechanically and electronically,by means of 64 pins, to the corresponding mother-boards. A photo of the boxD1 is shown in Figure 2.2.

Each plane of the detector with its electronics is mounted on an aluminummother-board. In Figure 2.3 it is possible to see besides the detector thetwo pre-amplifiers for the X and Y views (right and left side of the siliconelement), and the trigger and interface electronics (below). The 16 planes arevertically stacked. The interplanar distance is 1.4 cm, except for the first twoplanes which are separated by 8.5 cm in order to improve the incident particlestrajectory determination. They define the angular aperture of the telescope,which is about 32 degrees. The 16 planes are modular, so that mechanicallyand electronically they are interchangeable. Below the 16 silicon planes other4 modules, dedicated to the trigger electronics, silicon power supply, analog-digital conversion, and FIFO, are placed.

The 20 plane structure is housed in a cylindrical aluminum vessel (seeFigure 2.4) of 284 mm diameter and 480 mm height, filled up with nitrogen at1.2 atm. The vessel is 2 mm thick, except for a window above the first siliconplane where it is reduced to 300 µm (Figure 2.1). The top part of the vessel isrounded, while the bottom part houses the connectors for the interfaces withthe other parts of the detector.


Figure 2.2: Photograph of the internal structure of the detector.


Figure 2.3: One of the 16 modules of electronics and silicon sensitive elementcomposing D1 box.


Figure 2.4: The aluminum vessel.


The lateral strips (n.1 and n.16) of every silicon layer are used for theAnticoincidence System (AC); they are read together by the same electronicchannel, except for those of plane 1 where they are physically disconnected. Atotal number of 448 electronic channels, out of the 512 available in the box D1,are used for the particle track and energy information; the AnticoincidenceSystem data occupy an additional 30 channels, while the remaining 34 areused for housekeeping data (16 plane currents, 2 temperatures, 4 voltages,1 threshold level, 11 ratemeters at different depths of the telescope), whichmonitor the status of the whole instrument.

The signals produced by the incoming particles in the silicon strips arefirst amplified and shaped. Every plane of the telescope has two 16 channelspreamplifiers. Data are then converted to the digital format by means of a12 bit ADC, with a full scale of 2800 mip (1 mip being equivalent to 30400electrons or about 105 keV of released energy). The resolution per channel isthus about 0.67 mip ch−1, equivalent to 0.07 MeV ch−1 (see §2.5.1). There aretwo independent lines going to two different ADC’s, for redundancy reasons.Only one is operating at a given moment. Before the ADC there are two gainamplifiers that can be selected depending on the acquisition mode: the firstprovides an amplification of a factor 32 (used only for noise tests) and thesecond of a factor 1 (active for normal acquisitions).

After conversion by ADC, data (1024 bytes/event) are sent through a FIFOto an 8 channel bus interface with the on-board computer (box D2), built againby the company Laben. Here all tasks of event processing are performed, beforesending the data to the interface computer (box E) for mass memory storage.

The core of the box D2 are two 8086 microprocessors working with a clockspeed of 4 MHz. In normal conditions both of them are operating in Master-Slave mode: the Master microprocessor receives the event from box D1 andperforms pedestal suppression and data reduction tasks, while the Slave is usedto format the data, according to the acquisition mode, and sends them to thesubsystem box E. It also selects the trigger logic, implements the Second Leveltrigger and interfaces most of the telecommands with the silicon detector.

The interface computer (box E) represents the last step of the NINAdata processing before the records are sent to the satellite for transmission,via telemetry, to ground. Two exemplars of box E, for redundancy, have beenbuilt, both realized by the Russian company VNIIEM. Finally the power supplysubsystem (box P ), made also by VNIIEM, has the function of electricallyconnecting the satellite with its various subsystems.

Further details about the instrument can be found in Bakaldin et al. [1997]and Bidoli et al. [1999].

2.3 The NINA-2 instrument

The detector NINA-2 is, in its silicon core (box D1), identical to the NINAone but, making use of the extensive computer and telemetry capabilities of

2.4. TRIGGER LOGIC 69

MITA, the active data acquisition time has been improved.There was two main computer systems on board: the OBDH (On Board

Data Handler) and the PL/C (Payload Computer). The OBDH was linkedto all spacecraft systems such as the telemetry frame formatter, the interfacewith the active control systems, the sensors and actuators, the engineering andscientific data readout. The role of the boxes D2 and E was covered in NINA-2 mission by the Payload Computer. Data from the detector were read outfrom the PL/C which performed all tasks of data readout, reduction, secondlevel triggering and active acquisition mode switching. The PL/C could varythe trigger configuration as a result of telecommands sent from the groundstation or automatically adjusting the trigger configuration to cope with anincreased particle flux. This architecture allowed good development flexibilityand a relatively simple integration of the scientific payloads. Work on NINA-2 begun in 1999; the integration phase included beam tests of the detectorat the accelerator facilities of GANIL (France), GSI (Germany) and Uppsala(Sweden) [Casolino et al. 2001c].

All satellite and detector systems had their redundant counterpart, withthe exception of the silicon detector which had a functional redundancy inthe multiplicity of the strips and the different triggers which allowed to copewith eventual malfunctions. Software second-level-trigger figures allowed todiscard broken strips or planes and substitute anti-coincidence vetoes. Thesystem performed automatic calibrations and checks of the dark current noiseat intervals during acquisition; this allowed to take into account eventual -but not so unusual - shifts of the detector’s pedestal or variations of theamplification chain gain. According to the trigger configuration, the detectorcould vary its observational characteristics in order to focus the acquisition ofdifferent particles and energy ranges.

2.4 Trigger logic

NINA telescopes were able to work in different operating conditions, switchedautomatically or via telecommands (see Chapter 3), which affected the triggersystem. In particular:

1. Two thresholds for the energy deposited in the single silicon layers havebeen implemented: a Low Threshold (L.T.), corresponding to 2.5 mip,and a High Threshold (H.T.), corresponding to 25 mip.

The level of the threshold was fixed by telecommand.

2. The strips 1 and 16 of every silicon layer, except the plane first, wereused in the Lateral Anticoincidence System.

The hardware Lateral Anticoincidence could be turned off bytelecommand, for instance in case of a malfunction of one of the lateralstrips.


3. The planes 15 and 16 could be used as Bottom Anticoincidence. Thedefault operating mode adopted plane 16 but, in case of need, plane 15could be selected by telecommand.

The Bottom Anticoincidence could be totally removed by telecommand,allowing the detection of particles crossing the whole apparatus.

The main trigger of the acquisition system was the following:

TRGM1 = D1x × D1y × ((D2x + D2y) + (D3x + D3y)) ,

where Dij denotes a signal above-threshold coming from plane i, along viewj (j=x,y). The logic OR of planes 2 and 3 provides redundancy in case of afailure of plane 2.

In the default operating mode, this trigger was used together with theLateral and Bottom Anticoincidence ON, in order to ensure the completecontainment of the particle inside the detector. This was the condition whichallowed the best energy and nuclear discrimination to be obtained by NINA.Moreover, TRG M1 could be used with Low or High Threshold, defining twodifferent intervals of nuclei which could be detected.

It was possible to switch, via telecommand, to a second trigger:

TRGM2 = (D2x + D2y) × (D3x + D3y) × (D4x + D4y) × (D5x + D5y) ,

which was used again in its basic operating mode with the Lateral and BottomAC ON. This trigger, used for particular data taking demands or in case offailure of the first plane, increased the acceptance angle, at the expenses ofa slight worsening of the angular resolution. The combination of TRG M2and High Threshold again excluded most of the protons from the trigger. Asregards NINA-2 the second trigger logic was without plane (D5x + D5y).

The acceptance window of particles with TRG M1 in full containmentregime is shown in Table 2.1, for Low and High Threshold.

2.5 NINA and NINA-2 beam tests

In April 1997 NINA detector was tested at GSI (Gesellschaft furSchwerionenforschung) Laboratory in Darmstadt (Germany), which provideda 12C beam.

For this test, all parts of NINA (box D1, D2, P and E) were used in afully operational mode. The satellite acquisition system and the telemetrytransmission were simulated by means of a custom-made instrument (GSE -Ground Support Equipment) which interfaced, by means of Camac modules,the box E to the data acquisition system.

2.5. NINA AND NINA-2 BEAM TESTS 71

Particle Z Emin (MeV/n) Emax (MeV/n) Emin (MeV/n) Emax (MeV/n)1H 1 10 48 11 162H 1 6 32 7 133H 1 5 25 6 123He 2 12 58 12 584He 2 10 50 10 506Li 3 11 59 11 597Li 3 10 54 10 547Be 4 14 75 14 759Be 4 12 65 12 6510B 5 16 79 16 7911B 5 14 74 14 7412C 6 18 90 18 9014N 7 18 95 18 9516O 8 21 107 21 10719F 9 21 106 21 106

20Ne 10 23 117 23 11728Si 14 27 142 27 142

Table 2.1: Energy windows for the most abundant particles in NINA detector (TRGM1 and acquisition with Low (3th and 4th column) and High (5th and 6th column)Threshold). Emin is the minimum energy for triggering; Emax is the maximumenergy for the containment.


The whole NINA instrument, plus the power supply, the GSE and its PCwere placed in the beam room. Two scintillators were put in front of thedetector for beam monitoring. The GSE PC was controlled by remote. In thecontrol room were also located electronic modules for processing the scintillatorsignals and a precision system for moving the box D1 with respect to the beamline.

Data were taken in different runs at the carbon beam energies of 300,100, 80 and 65 MeV/n. Some measurements were performed interposing apolyethylene target, 0.6 mm thick, between the box D1 and the beam, inorder to obtain all the products of the 12C fragmentation. In some runs thescintillators were removed in order to have a monochromatic beam.

All possible telecommand settings (triggers, thresholds, lateral and bottomanticoincidences) were tested during this period. Tests of the behavior of thedetector as a function of time and temperature were done as well.

In June and July 1999 NINA-2 detector was tested at the acceleratorfacilities of GANIL (France), GSI (German) and Uppsala (Sweden), whichprovided 1H, 2H, 4He, 12C, 14N and 16O beams.

In next section I will describe the energy calibration and the energy andmass resolution of the instruments.

2.5.1 Energy calibration

The energy information coming from the particles is given in terms of ADCchannels. A necessary step in order to recognize the particle and especiallyto identify its true energy is to determine the correlation between the ADCchannels and the energy expressed in MeV. This procedure is known as the”energy calibration” of the instrument.

Two different methods have been used to calibrate in energy the detectors.In both procedures, different samples of nuclei from proton to oxygen atdifferent energies were extracted from the whole set of beam tests data andexamined; straight and clean tracks, contained in the telescope volume andpassing through the central strips of the detector were selected.

In the first method, the detectors have been calibrated by comparing theenergy deposits of the collected families of nuclei in each silicon view, expressedin ADC channels, with the corresponding simulated ones expressed in MeV.Montecarlo simulation programs (Geant 3.21 [Brun et al. 1994]) had beenpreviously calibrated by using monochromatic proton and helium beams atthe PSI Laboratory of Willigen (Zurich - Switzerland).

From the distribution of the conversion factors (MeV/ADC Ch) in eachsilicon view and for the various particles at different energies, we obtained, forNINA and NINA-2 detector respectively, the following average ratio betweenMeV energy and ADC channels:

R = (0.067 ± 0.002)MeV

ADC Ch, (2.1)


0

250

500

750

1000

1250

1500

1750

2000

2250

0 5 10 15 20 25 30

n. view

AD

C c

h.

Figure 2.5: Best-fit of the Bragg curve of a carbon nucleus, which stopped in the20th silicon view. The bullets represent the experimental energy losses in each siliconlayer.


0

50

100

150

200

250

0.06 0.062 0.064 0.066 0.068 0.07 0.072 0.074 0.076

R (MeV/ADC ch.)

Eve

nts

Figure 2.6: Distribution of the calibration factors coming from best-fits of a sampleof carbon events, fitted by a Gaussian curve.


and

R = (0.070 ± 0.002)MeV

ADC Ch, (2.2)

holding for the whole energy spectrum of the ADC.

The second approach consisted of an interpolation procedure, which fitteda Bragg curve to the energy values released by the nuclei in each silicon layer.This curve is based on the Bethe-Bloch formula, which depends on the kineticenergy, the atomic and the mass number of the nuclei. We kept these quantitiesfixed in the fit allowing only the additional parameter of conversion betweenMeV and ADC counts to change.

In Figure 2.5 the best-fit of a Bragg curve of a carbon nucleus is shown;the bullets represent the experimental energy deposits in each silicon view.

Figure 2.6 represents instead the distribution of the calibration factors asdetermined by a fit of a sample of carbon events for NINA detector; meanvalue and standard deviation are estimated by a Gaussian fit of the histogram.

The results of this procedure, applied to different particles with differentkinetic energy, are in agreement within the errors with the ones coming fromthe former method.

Finally figure 2.7 shows the E1 vs Etot curves of real particles, coming fromthe fragmentation of the original carbon beam by means of the polyethylenetarget, with energies expressed in MeV. All the products of fragmentationare visible, and the energy behavior is in good agreement with the expectedsimulated data [Bakaldin et al. 1997].

2.5.2 Energy resolution

After obtained the average ratio between MeV energy and ADC channel, weneeded to calculate the energy resolution of our instruments.

Being NINA and NINA-2 almost completely ”active”, and working with fullcontainment of the particle, their energy resolution are excellent. Practicallythe entire information about the energy of the particle is stored in the activesilicon strips and then collected; the small portion of the particle energy whichis lost derives by the deposits released in the few passive materials of thedetector: the 300 µm of aluminum of the entrance window, the 1 µm ofaluminum between the X and Y wafers of strips, the nitrogen between theplanes.

Simulations of the energy resolution of the apparatuses brought brilliantresults: they are able to reconstruct the particle energy with an error smallerthan 1%.

One of our aims during the beam tests was therefore to confront such resultswith those coming from real particles. In order to do that, we performed runsremoving any passive materials between the beam and telescope (like target,scintillators, etc.).


0

2

4

6

8

10

12

14

16

18

0 25 50 75 100 125 150 175 200

H

He

Etot (MeV)

E1

(MeV

)

0102030405060708090

100

0 200 400 600 800 1000

He

Li

Be B C

Etot (MeV)

E1

(MeV

)

Figure 2.7: Distribution of the energy released in the first plane (E1) andtotal energy (Etot) detected in NINA instrument for particles produced in thefragmentation of 12C at GSI.


Figure 2.8: Distribution of the energy deposited in the detector NINA-2 for 1H and2H beams, during the test in Uppsala in 1999.


Figure 2.9: Distribution of the energy deposited in the detector NINA-2 for 4He and16O beams, during the test in Uppsala in 1999.

Unfortunately the analysis of such runs presented that the beams had ahigh percentage of impurities, and also their energies were not stable duringtime.

Due to presence of fluctuations in energy of beam bunches, which were notpossible to estimate, all results obtained with such sets of data are only anupper estimation of the real energy resolution of the apparatuses.

Figures 2.8 and 2.9 show the distribution of the energy deposited in NINA-2detector during the beam test in Uppsala (1999) for 4 different beams (1H, 2H,4He, 16O). From these pictures it is clear that the energy resolution is betterthan 2.5% at energy greater than 50 MeV.

2.5.3 Mass resolution

NINA and NINA-2 acquisition mode allows the complete registration of theparticle track with all its energy deposits in the strips; in this condition a goodnuclear and isotope discrimination can be performed.

The mass M and the charge Z of the particles are calculated in parallel bytwo methods, in order to have a more precise particle recognition:

a) the method of the residual range [Cook et al. 1993, Hasebe et al. 1993,Sparvoli et al. 1997].


Z isotope a b

1 1H, 2H, 3H 2059 1.7332 3He, 4He 2144 1.7413 6Li, 7Li 2474 1.7864 7Be, 9Be, 10Be 2386 1.7745 10B, 11B 2419 1.7786 12C, 13C 2355 1.7697 14N, 15N 2386 1.773

Table 2.2: Average values for a and b for several nuclei.

In this method, the charge Z is estimated by means of the productE1 × Etot. Here E1 is the energy released by the particles in the firstsilicon detector (two layers) of the tower, and Etot is the total energyreleased in the whole instrument. As shown in Figure 2.7 the differentnuclear families lie on different hyperbolas E1×Etot = k(Z2). A behaviorclosely resembling the Bethe-Block curve is obtained, and particles can bediscriminated quite well if their energy is before the minimum ionizingpoint. This is a well known method used in nuclear physics, alreadyadopted in cosmic ray experiments since 1960 [Comstock et al. 1966].

Once the charge Z is identified by its E1 − Etot hyperbole, the mass ofthe specific isotope is evaluated by applying the following formula:

M =

(a(Eb

tot − (Etot − ∆E)b)

Z2∆x

) 1b−1

, (2.3)

where ∆E is the energy lost by the particle in a thickness ∆x measuredstarting from the first plane.

The parameter a is a constant depending on the medium and b has avalue between 1.5 and 1.8 in NINA’s energy range. A precise evaluationof such parameters for each atomic species has been obtained by a fit ofthe following expression:

R = aM

Z2

(E

M

)b

, (2.4)

where R and E are respectively the measured range and the kineticenergy of known particles of mass M and charge Z. From simulations itturned out that a and b do not have a strong dependence on the mass ofthe isotope, therefore a mean value for each Z can be taken. Table 2.2shows the average values for a and b for several nuclei.

Once evaluated the parameters a and b we reconstructed all isotopemasses for all incident direction particles. For every mass, a Gaussian fit


Figure 2.10: Mass distribution, as given by equation 2.3, of a sample of 4He nucleicollected during the beam test session at Uppsala (Sweden).


Figure 2.11: Mass distribution, as given by equation 2.3, of a sample of 12C nucleicollected during the beam test session at Uppsala (Sweden).


Figure 2.12: Mass distribution, as given by equation 2.5, of a sample of 4He nucleicollected during the beam test session at Uppsala (Sweden).

has been performed; in such a way the average value of the reconstructedmass M and its corresponding standard deviation σ have been evaluated.Figures 2.10 and 2.11 show the mass reconstructed, as given by equation2.3, for 4He and 12C beams respectively.

b) the method of the approximation to the Bethe-Blochtheoretical curve.

With this second method we estimate the mass M and charge Z of theparticle by minimizing the following χ2 quantity :

χ2 =N∑

i=1

(Wi

(∆Ereal

i − ∆Etheori

))2, (2.5)

where ∆Ereali is the energy released by the particle in the i-th view,

∆Etheori is the corresponding expected value, Wi is the weight for every

difference Wi = 1∆Ereal

i, and the sum is extended to the N silicon layers

activated by the particle, excluding the last one where the particle stopsand the fluctuations of the energy deposits are generally very big.


In order to build such a function, it is necessary to follow step by stepthe particle’s path, calculating the scattering angles at every layer. Thismethod takes into account also the energy losses in dead layers, thuspreventing systematic shifts on the reconstructed masses. Figure 2.12shows mass distribution, as given by equation 2.5, of a sample of 4Henuclei collected during the beam test session at Uppsala (Sweden).

Results from hydrogen to oxygen for both methods give sigmas in the range0.07 - 0.37 amu, confirming the good capability of this instrument to performisotope analysis.

Chapter 3

Experiments in orbit

In the first part of the Chapter descriptions of Resurs-01 and MITA satellites,telecommands, data downlink and preliminary data handling are outlined; inthe second one, methods of data analysis are discussed.

3.1 The Resurs-01 satellite

The Resurs-01 n.4 satellite, made by the Russian space companies VNIIEMand SOVZOND, hold the boxes composing NINA. The main technicalcharacteristics of the satellite are resumed in table 3.1. The satellite itself hasbeen conceived to exploit the natural resources of the Earth: besides NINA,other instruments are placed on board for terrestrial observations. It is amedium-size class spacecraft, with a mass of 2500 kg (slightly more than 40dedicated to NINA). The silicon detector (D1) is placed on the very top of thesatellite, opposite to the Earth so to have a wide field of view not covered byother apparata. The other modules are inserted instead in the central body ofthe spacecraft as shown in Figure 3.1, connected to D1 by 7 meter long cables

The connections and interfaces between Resurs and NINA were:

• mechanical: the detector was linked to the satellite by means of a L-shape structure, while the other boxes are fixed with bolts;

• electrical: NINA took the tension, through the box P, from the powersupply of the spacecraft, based on the solar panels;

• electronic (data transmission): data registered inside the detectorwere transmitted to the satellite and thereafter to ground, via telemetry;

• electronic (telecommands): the telecommands sent from groundwere received and transferred to the NINA on-board computer,which processed them and consequently changed the detector workingconfiguration.

85

86 CHAPTER 3. EXPERIMENTS IN ORBIT

Figure 3.1: Positioning of the various systems of NINA - D1, D2, E and P - ontothe satellite Resurs.

3.1. THE RESURS-01 SATELLITE 87

Figure 3.2: Resurs-O1-N4 satellite during the integration phase of NINA.


Figure 3.3: The satellite on board of the Zenit rocket in July, the 10th 1998, somehours before the launch, at the Russian base in Baikonur.

3.2. MITA SATELLITE 89

Mass (with load) 2500 kgDiameter 1.4 mHeight 6.4 m

Solar panels spread 11.6 mMedium power supplied 500 W

Maximum power supplied 1200 WVoltage 27 V -> (24 ÷ 34) V

Satellite lifetime > 3 years

Table 3.1: Technical characteristics of the satellite Resurs-01.

As already mentioned, NINA’s allowed mass is about 40 kg. The 32modules of the telescope (silicon detectors plus read-out electronics) weightin total only 13 kg. The other parts contribute to the remaining mass. Figure3.2 shoes the satellite during the integration phase of NINA.

3.1.1 Resurs-O1-N4 launch

The satellite was launched on board of a Zenit rocket in July, the 10th 1998,from the Russian base in Baikonur (see Figure 3.3), which is located in theRepublic of Kazakistan, East of the Aral lake. From this base, many of themost important soviet space missions have started (Sputnik 1, Vostok 1 withGagarin).

The satellite was placed in a sunsyncronous polar orbit of about 810 kmheight, 100 minutes of period and 97.0 degrees of inclination.

The period of observation ended in April 1999, when problems with the onboard transmitter stopped the data download [Bidoli et al. 2001].

The average amount of data that are sent to ground by the satellite is 2Mbyte/day, varying between 1 and 8. This constraint has imposed most ofthe decisions taken for NINA with respect to the electronics and the on-boardcomputer.

3.2 MITA satellite

The MITA satellite (see Figure 3.4) has been built under an ASI (Italian SpaceAgency) contract by Carlo Gavazzi Space to implement a low cost platform forsmall Earth missions; its main characteristics are shown in Table 3.2. MITArepresents a new generation of satellite architecture based on modular criteriathat allow to build up a mission in short periods (about 1-2 years), takingadvantage from the big flexibility for payloads and launchers.

In the first launch MITA hosted the NINA-2 detector and the technologicalpayload MTS-AOMS (Micro Tech Sensor for Attitude and Orbit Measurements


Figure 3.4: The Satellite MITA with NINA-2 onboard.

3.3. TELECOMMANDS 91

Mass 170 kgDimension 180x160x84 cm3

Average (peak) power cons. 85(120) WAttitude Control System 3 axis stabilized, Earth pointing

Attitude Accuracy 1o/axisCommunications S-Band

Telemetry 512 Kbps-ESA StandardTelecommand Uploads 4 Kbps

Mass Memory 64 MBytesPayload Mass 30 Kg

Payload Power Budget 40 WSolar Panels 2 x 1.33 m2 - GaAs

Table 3.2: MITA satellite main characteristics.

System). The satellite was equipped with two 3-axis magnetometers.Magnetometers data were used for ACS (Attitude and Control System) andprovided also data for in flight software and off-line analysis to correlate theparticle fluxes with detailed measurements of the magnetic field.

3.2.1 MITA launch

The satellite MITA was launched on July the 15th, 2000 from the cosmodromeof Plesetsk (Russia) with a Cosmos-3M rocket. The satellite and the payloadoperated soon within nominal parameters. MITA was placed in a circularpolar orbit of about 460 km height, 90 minutes of period and 87.3 degrees ofinclination. It was possible to choose between sun and Earth pointing withappropriate telecommands.

Figure 3.5 shows the pre-launch phase, in which the vector is put in vertical.This operation takes more than one hour. Figure 3.6 shows the Cosmos-3Mrocket in flight.

3.3 Telecommands

The interaction between the ground Station and NINA telescopes during theflight is driven by telecommands (TLCs). Some are dedicated to operationslike power switching ON/OFF, data transferring, memory cleaning, single ordual microprocessor mode setting. The others act on the trigger logic or onthe storage model, having therefore a direct effect on the data collection.

TLCs are the only way to communicate with the instruments, once thelaunch has been executed. As for NINA, the mailing/delivering of differenttelecommands is done from military base in the Russian site, so that such


Figure 3.5: MITA pre-launch phase.

3.3. TELECOMMANDS 93

Figure 3.6: Cosmos-3M rocket in flight.


code lateral anti coincidence bottom anti coincidence threshold trigger

D3 OFF OFF low M1D7 OFF OFF high M1F3 ON OFF low M1F7 ON OFF high M1

Table 3.3: All type of commands used during NINA-2 flight.

procedure, as well as the one of data delivery, can be done only when thesatellite passed over the ground station.

To each TLC corresponds a line with a relais switched on by means of a27 V line. For redundancy, the same TLC is sent more than once, so that it isimpossible to use the same line to switch ON and OFF a certain TLC. Everycommand, thus, needs two relais, one for switching on, another for switchingoff. The satellite sends a packet with the status of the telecommands eachtime a change has occurred. I refer to Bidoli et al. [2001] for a description ofstatus of NINA telescope during the flight.

As for NINA-2, the mailing/delivering of different telecommands is done byTelespazio [Telespazio 2001], and directly controlled by WiZard collaboration.

Table 3.3 shows all the status of NINA-2 telescope during the flight;however all combination of parameters would have been possible.

3.4 NINA-2 performance in orbit

In this section NINA-2 performance in orbit for first months of flight (referringto Bidoli et al. [2001] for the same description of the performance of NINA)is outlined.

As described above, MITA launch took place on July, the 15th, 2000. Thetransmission of its scientific data started on July, the 21th, after an initialperiod needed for stabilization of the orbit and overall checks of the satellitefunctionality. The analysis of the first sample of data received showed that theinstrument performanced well and confirmed the functionality of the wholesystem.

During one orbit the satellite has a day-night cycle according to its positionwith respect to the Earth and the Sun. Two of the 34 housekeeping dataavailable on NINA-2 give internal temperatures sampled at two differentheights inside the detector. We measured the behavior of the two temperaturesin orbit. The excursion of their values between light and shadow was less than1 degree, as required in the construction phase.

Important information about the status of the detector are provided by theratemeters, which are also part of the housekeeping sector. These are indicatorsof the particle flux impinging on different planes of NINA-2 telescope, which

3.4. NINA-2 PERFORMANCE IN ORBIT 95

Figure 3.7: Low ratemeter counting rate (at plane 6) as a function of time. Thetwo black vertical arrows define 1 MITA orbit.


is a function of the orbit of the satellite. Low and high flux ratemeters areimplemented at different heights inside the telescope. In case of intense flux,the high ratemeters section provides information while the low ratemeters maysaturate.

Figure 3.7 shows the behavior of the low ratemeter implemented on planen.6 during one typical orbit. The counting rate is given in hertz. From thepicture one can easily follow the path of the satellite through the differentregions of the Earth’s magnetosphere. In particular, it can be seen how the fluxincreases at the Poles with respect to the Equator, because at high latitudes theterrestrial magnetic field does not effectively prevent low energy particles fromapproaching the Earth. In the South Atlantic Anomaly (SAA) the magneticfield has a local minimum and thus the low energy proton flux reaches highlevels. This is clearly evident by the behaviour of the ratemeter counter.Outside SAA, however, ratemeter values show that particles acquisition ispresent.

We have examined the stability of some important parameters of thedetector with time, like the voltages, the threshold and the temperaturesvalues, during the first months of life of the detector.

3.5 MITA in orbit control

The Goddard Space Flight Center (GSFC), Networks and Mission Services,Mission and Data System Projects, Multi-Mission and Data SystemManagement provides access to unclassified satellite orbital data that has beenreceived from United State Space Command (USSPACECOM). This orbitaldata consists of Two-Line Element (TLE) sets, Satellite Catalog Messages,Project Tip Messages, Satellite Decay Messages, Predicted Decay Forecasts,Satellite Box Scores and Satellite Reports. Orbital Information Group (OIG)provides this data from an on-line Web Site [Orbital Information Group 2001].This Web Site supports real-time requests and historical requests. The datais provided to this site at irregular intervals around the clock. Only historicaldata requests involve human intervention. The data received within the correctformat is placed on the Web Site in a standard format for easier access by theusers.

During the MITA mission in space it was necessary to have a continuecontrol of satellite parameter orbital. In fact, MITA was put in orbit on aaverage altitude of 460 km (July, the 15th, 2000), but its altitude reduced toabout 220 km during the last radio contact (August, the 10th, 2001) and toabout 130 km for the last radar contact (August, the 15th, 2001).

It was found that MITA decay was function of pointing, of altitude and of

3.5. MITA IN ORBIT CONTROL 97

solar activity. In first months of space life, satellite decays of approximately0.1 km/day in Earth-pointing and 0.3 km/day in sun-pointing. When altitudedecreases, decay rate became more so as density of residual atmosphere. Inlast months, satellite decays of approximately 1 km/day in Earth-pointing and3 km/day in sun-pointing. In the last 3 hours before last radar contact, MITAdecayed of about 20 km.

A C++ program was outlined to read TLE files. I refer to OrbitalInformation Group [2001] for a complete description of TLE code. The boxdiagram of program is described in Figure 3.8. The aim was to control dailyMITA decay, in order to intercede during the flight.

Figure 3.9 shows the altitude (semi-major axis minus Earth radius) ofMITA orbit from January, the 1st, to August, the 15th, 2001. It is possible toobserve the passage between nadir and solar pointing at day 110, when slope ofcurve changes. In the last 25 days of mission decay rate increases drastically.

The two-body equation of motion

Newton’s second law of motion, applied to a constant mass system andcombined with his law of universal gravitation, provides the mathematicalbasis for analyzing satellite orbits. Newton’s law of gravitation states that anytwo bodies attract each other with a force proportional to the product of theirmasses and inversely proportional to the square of the distance between them.The equation for the magnitude of the force due to gravity is

F = −GMm

r2r = −µm

r2r , (3.1)

where F is the force vector due to the gravity, G is the universal constantof gravitation, M is the mass of the Earth, m is the mass of satellite, r isthe distance from the center of the Earth to the satellite, and µ = GM is theEarth’s gravitational constant (=398600.5 km3 s−2).

Combining Newton’s second law with his law of gravitation, we obtain anequation for the acceleration vector of the satellite:

r + (µr−3) r = 0 . (3.2)

This equation is called the two-body equation of motion and is the relativeequation of motion of a satellite position vector as the satellite orbits theEarth. In deriving it, we assumed that gravity is the only force, the Earth isspherically symmetric, the Earth’s mass is much greater than the satellite’smass, and the Earth and the satellite are the only two bodies in the system.

A solution of the two-body equation of motion for a satellite orbiting theEarth is the polar equation of a conic section. It gives the magnitude of theposition vector in terms of the location in the orbit,


Figure 3.8: Box diagram of C++ program to process TLE files.


Figure 3.9: MITA orbit average altitude for 2001 in function of time.


r = a(1 − e2)/(1 + e cos ν) , (3.3)

where a is the semi-major axis, e is the eccentricity, and ν is the polar angleor true anomaly.

We can define all conics in terms of the eccentricity e: e=0 corresponds toa circle, 0 < e < 1 to an ellipse, e=1 to a parabola and e > 1 to an hyperbola.

Since the 3.2 is a differential equation in 3 dimensions, it needs 6 parametersor initial conditions in order to be unambiguously solved. In astrodynamicsit is common to use 5 constants of motion and one parameter changing withtime. They are called classical orbital parameters and are so defined:

a - semi-major axis. Describes the size of the ellipse.

e - eccentricity. Describes the shape of the ellipse.

i - inclination. Angle between the vector of the angular moment ( h =

r × v) and the versor of the Z axis;

Ω - right ascension of the ascending node. Angle from the Vernalequinox to the ascending node. The ascending node is the point ofintersection of the satellite orbit with the equatorial plane, during theSouth-North passage;

ω - argument of perigee. Angle from the ascending node to theeccentricity vector (defined as the vector going from the centre ofthe Earth to perigee of the orbit and with a magnitude equal to theeccentricity), in the direction of the satellite’s motion;

ν - true anomaly. Angle from the eccentricity vector to the position ofthe satellite r measured in the direction of the satellite’s motion.

Figure 3.10 and 3.11 show the geometry of an ellipse and the classicalorbital parameters.

If period of orbit is known, it is possible to determinate the semi-majoraxis by using the equation for the period of an elliptical orbit:

a = [(P/2π)2µ)]1/3 331.24915P 2/3km , (3.4)

where the period is in minutes [Larson et al. 1992].

Our program works using equations above, calculating the period, thevalue of semi-major axis, the altitude and the decay gradient of MITA orbit.Moreover it controls if all other parameters do not change appreciably. Itrevealed itself very useful for strategy of data acquisition.


Figure 3.10: The geometry of an ellipse and the classical orbital parameters.


Figure 3.11: Definition of Keplerian Orbital Elements of a satellite in an EllipticOrbit.

3.5.1 Data downlink and preliminary data handling

In this section all the steps from data receiving to the first analysis are outlined.Figure 3.12 shows the box diagram of data receiving. Fucino (Italy), Malindi(Kenya) and Cordoba (Argentina) antennas for data downlink were used. Inthis way, MITA had from three to six downlinks in a day. Any changes oftelecommands or satellite asset should be done with a new configuration ofparameters to be sent to Telespazio in time to implement them.

When data are downlinked to ground, they are regularly fetched by FTPto MITA-INFN group, implemented by NORAD-OIG files and archived in oneappropriate PC used for preliminary analysis. Quality data, orbit parameterand quicklook check up are done. After, all data are storaged in CD andcataloged in files containing NINA status, MITA pointing and indexes ofsolar activities. A pre-analysis is done and preliminary results are put in anappropriate web page (see next section). All these activities have been donein less then one day to maintain a complete control of the mission.

3.5.2 On line data

The idea to create a MITA ”on line” internet page started with the requirementto inform all collaboration about status of MITA mission as soon as possible,in a easy and direct way. For these reasons a internet page was written instatic html and put ”on line” in INFN-Rome2 domain [Wizard 2002]. At thebeginning the page contains only the essential informations, such as detectorand satellite housekeeping and preliminary results on particles rate. A linkto NOAA Online Data was done to correlate their and our ”on line” results


Figure 3.12: Box diagram of data receiving and first analysis.


Figure 3.13: Particles count rate for GOES8 [GOES-8 2002] and NINA-2 duringa solar flare in February, the 26th, 2001. The different colours correspond to threedifferent energy ranges.

3.6. DATA ANALYSIS 105

[NOAA 2002]. Now the page figures in a nicest way, and contains sections oninstrumentations, launches and results on NINA and NINA-2 missions, butthe useful side of ”on line” page remained unmodified.

Figure 3.13 shows particles count rate for GOES8 (in geocentric orbit)[GOES-8 2002] and NINA-2 during a solar flare in February, the 26th, 2001.The different colours correspond to three different energy ranges. Count ratesincreases for both detector in the middle of second day and for the same energyrange. Moreover, due to its polar orbit, NINA-2 particles rate shows sharpincreasing due to South Atlantic Anomaly. This Figure is only an example ofresults contained daily in MITA ”on line” internet page.

3.6 Data analysis

The optimal performance of NINA and NINA-2 telescopes in terms of charge,mass and energy determination is achieved by requiring the full containmentof the particle inside the detectors. In addition, a dedicated off-line trackselection algorithm is applied, which rejects upward moving particles, tracksaccompanied by nuclear interactions, and events consisting of two or moretracks. In order to calculate the energy deposit of the crossing particles ineach layer of the detector, their incident angles are taken into account.

In this section the track selection algorithm and the method ofdetermination of flux, including background exstimation, geometrical factorand exposure time calculation are outlined.

3.6.1 The track selection algorithm

The selection algorithm, implemented for the analysis of flight data, appliessix rejection criteria:

1. Real particles moving downwards and stopping inside the detector havean energy deposition that increases along the track. It is natural,therefore, to request tracks to deposit in each view a quantity of energygreater than in the previous one multiplied by a constant K1, which takesinto account the energy fluctuations. If

E(i) < K1 × E(i − 1),

for any i in the range from the second crossed view to the one with themaximum deposit of energy, the event is rejected.

2. In order to clean the data sample from particles with nuclear interactions,two energies for each track are calculated:

- Etrack = sum of the energies released by the particle from the first hitview to the view following the one with the maximum deposit of energy;


- Eresidual = total amount of energy left in all the remaining layers.

The two energies are compared and events with

Eresidual > K2 × Etrack,

where K2 is a parameter to be optimized, are rejected.

3. Double tracks are eliminated estimating two energies for each crossedview i along the particle path:

- Ecluster(i) = sum of energy released in the strip with the maximumdeposit of energy and in the two nearest strips;

- Enoise(i) = sum of the energy released in the other strips of the siliconlayer.

If

Enoise(i) > K3 × Ecluster(i),

for any of the i crossed views and where K3 has to be fixed, the event isrejected.

4. Events with the maximum deposit of energy in the first view are rejected.This criterion, together with the condition n. 1, selects downwardsmoving particles.

5. Events where the maximum energy release in the X view and in theY view are not in the same or between consecutive detector planes arerejected. This request helps filtering double tracks.

6. In order to reduce the number of particles which leave the detectorthrough the space between planes, events which release the maximumof the energy deposit per silicon layer in the strips 2 or 15, for any of thecrossed layers, are rejected.

To apply this algorithm to the data collected in orbit it was necessary tochoose the K1, K2, and K3 coefficients in such a way to efficiently clean thedata sample from the background, minimizing at the same time the numberof good events rejected.

The particle identification algorithm was previously tested with beam testdata and in the Galactic Cosmic Rays (GCR) analysis. It was found that thesecuts also eliminate, together with the background, about 3% of tracks of realgood particles. Correction for this factor was introduced.

If we define as efficiency ε the value ε = (1 − N. rejected good eventsN. good events

), the best

optimization of the K1, K2, and K3 values that we achieved (K1 = 0.7, K2 =0.01 and K3 = 0.01) determined an efficiency equal to ε = 0.975± 0.003 for allparticles.


Figure 3.14: Mass distribution, as given by equation 2.3, of a sample of heliumisotopes collected by NINA in orbit (top) and during the beam test session at GSI(bottom).


When detector was with Lateral Anti Coincidence off, to ensure particlescontainment, events which release in 2 and 15 hits of planes from 2 to 15 arerejected. Ki and ε values do not change sensibility.

After application of the track selection algorithm, data are analyzed by thetwo methods described in Chapter 2 to identify the nuclear and isotope origin.For a complete rejection of the background, only particles with the same finalidentification are selected. Finally, a cross-check between the real range of theparticle in the detector and the expected value according to simulation is aconsistency test for the event.

In Figure 3.14 the reconstructed masses using equation 2.3 for heliumisotopes detected in orbit are compared to the ones obtained from the test-beam data; the sample of particles in orbit has been selected during passagesover the polar caps in 1999, and in period of quiet solar activity. The pictureshows that the mass resolution of NINA in flight is in good agreement withthe measurements performed at GSI.

3.6.2 Background estimation

Secondary background produced by cosmic rays inside the instrument hasto be considered too. A delicate point in analyzing proton flux and rareisotope abundances is to understand whether these nuclei are due to secondaryproduction inside the instrument or no. The detector is installed into a specialvessel having a 300 µm aluminum window in correspondence of the telescopeaperture. The cross section value of proton interaction in the Al window,however, remains uncertain so that it is not possible to calculate the numberof secondaries in the instrument with high accuracy. It is also difficult todistinguish the secondary production inside the instrument, because usuallythe reaction produces more than one particle with a complex energetic andspatial distribution. Therefore, to estimate the secondary isotope fraction, weutilized a method similar to that proposed by Looper et al. [1996] for PETtelescope. In this technique we consider the first detector of NINA as a silicontarget 150 µm thick; comparing the particles surviving the track selectionalgorithm in the two cases - including and not including the first silicon layer -we have in the second case an additional number of secondary particles, whichwere presumably originated in the silicon target. If we reasonably supposethat the cross section of light isotopes generation in Si and Al are similar,then it is possible to estimate the number of secondaries generated in thealuminum window starting from the number of secondaries produced in thesilicon, just by considerations of thickness and density of the two materials.The geometrical factor in the two cases is practically the same, due to the


small thickness of the first layer; the ratio R between the particles producedin the silicon target (150 µm thick) and in the aluminum window (300 µm) isfinally R=0.43. Estimations of background by means of the method describedabove show that its average contribution to the proton count rate is less than5%, and to other light isotopes less than 8%.

3.6.3 Determination of fluxes

In order to estimate the particles fluxes it is necessary to know the geometricfactor (GF (E)) of the instrument as a function of energy, and the exposuretime in orbit. A correction factor must then be applied to account for energy-loss in the dead layers of the detector.

The incoming energy of the particles was reconstructed by an iterativealgorithm which is based on the Bethe-Bloch formula. The algorithm worksthis way: as a first step, the total energy Etot is taken as initial energy Ein of theparticle; with this value of initial energy, all energies deposited in every layeralong the particle track (namely the aluminum window, the silicon detectorsand the inter-gap volumes of nitrogen) are calculated, and the expected valueof the total energy deposited in the silicon Eexp

tot estimated.

The value of Eexptot is then compared with Etot. If their difference is greater

than 0.1 MeV we define a new initial energy as Ein = Eexptot + Estep, where Estep

is an incremental step energy fixed according to the precision that we want toreach, and perform a new iteration. When finally the condition

Etot − Eexptot ≤ 0.1 MeV

is fulfilled, the algorithm stops and the initial energy of the particle is identified.

In a general case, the number of particles detected by a instrument in unitof time are linked to the flux of cosmic rays I(Er, θr, ϕr) by the followingformula:

dNd(Ed, θd, ϕd)

dt= [∫

Er

∫Ωr

∫S

I(Er, θr, ϕr)η(Er, θr, ϕr) ××F (Ed, Ed, θd, θr, ϕd, ϕr) cos θrdErdΩrdS]dEddΩd , (3.5)

where Er, θr, ϕr and Ed, θd, ϕd are the energies and angles in polarcoordinates related to the instrument axis of real and detective particlesrespectively; F (Ed, Ed, θd, θr, ϕd, ϕr) is the response function of the instrument,which represents the probability that a particle with Er, θr, ϕr parameters givesresponse with Ed, θd, ϕd ones; η(Er, θr, ϕr) is the efficiency of the detector.


Figure 3.15: Geometric factor GF of NINA telescopes for 1H in TRG M1, LowThreshold mode.


Figure 3.16: Geometric factor GF of NINA telescopes for 2H, 3H, 3He, 4He in TRGM1, Low Threshold mode.


Figure 3.17: Geometric factor GF of NINA telescopes for 1H, 2H, 3H, 3He, 4He,12C, and 16O in TRG M1, High Threshold Mode.


If we assume an isotropic flux I(Er, θr, ϕr), equation 3.5 can be written as

J(Ed) =dNd(Ed)

dtdEd

=∫

Er

I(Er)F (Ed, Er) ×

×[∫Ωr

∫S

η(Er, θr, ϕr)dS cos θrdΩr]dEr . (3.6)

We can define a new factor GF (Er) as

GF (Er) =∫Ωr

∫S

η(Er, θr, ϕr)dS cos θrdΩr (3.7)

and assuming that the efficiency of the detector does not depend on angles(η(Er, θr, ϕr) = η(Er)), we can write

GF (Er) = η(Er)∫Ωr

∫S

dS cos θrdΩr (3.8)

and define it as Geometrical Factor of the instrument.

In an ideal detector the response function F (Ed, Er) is a delta function (forE = Er = Ed) and equation 3.6 becomes

dN(E)

dtdE= I(E)GF (E) . (3.9)

The geometric factor of the instrument (GF (E)) was calculated bymeans of Monte Carlo simulations based on the CERN-GEANT 3.21 code[Brun et al. 1994].

We simulated an isotropic flux impinging the aluminum window of thedetector, and took into account processes of particle interactions inside theinstrument, like ionization losses, nuclear interactions, multiple coulombscattering. Energy deposited in each strip was analyzed and if triggerconditions were satisfied, the event was stored. Simulated events covered thewhole energy range of the instrument.

The geometrical factor is so determined by


GF (E) =N(E)

No(E)πs , (3.10)

where No is the number of simulated events that hits the detector for afixed interval of time (∆T ) and a fixed energy bin (∆E), N is the number ofthem which satisfy the trigger conditions and s is the surface of the instrument.

As an examples, Figures 3.15, 3.16 and 3.17 present the geometric factorGF (E) of NINA telescopes respectively for 1H and for 2H, 3H, 3He, 4He inTRG M1, Low Threshold mode, and for 1H, 2H, 3H, 3He, 4He, 12C, and 16Oin TRG M1, High Threshold Mode, over the energy intervals defined by thetrigger and the selection conditions.

To evaluate the position of the satellite we use data from NORAD[Orbital Information Group 2001] and SGP(Simplified General Perturbation)-4 model of propagation [Hoots et al. 1980], while the geomagnetic coordinatesL-shell and B are calculated by means of the IGRF(International GeomagneticReference Field) [IGRF 2002] model.

To reconstruct the fluxes of a selected isotope, particles for which theestimated value of mass is at more than two standard deviations from theisotope mass are rejected, and compensation for good events lost in the tailsof the distribution is given.

Finally the differential energy spectrum I(E), for particles detected insidea specified region of known geomagnetic coordinates, is determined by thefollowing formula:

I(E) =N(E)

T GF (E) ∆E, (3.11)

where N(E) is the number of detected particles in the energy intervalE ±∆E/2, T is the exposure time in orbit for the period under consideration,GF (E) is the average value of the geometric factor in the same interval E ±∆E/2, and ∆E is the energy bin chosen to plot the flux.

3.7. CONCLUSIONS 115

The exposure time in orbit (T ) takes into account the dead time of theinstrument.

If a particle produces a trigger, the signal from the detectors is amplifiedand shaped before being sampled by an ADC. The event is then stored in aFIFO for 2 ms. The data are read by the on-board processor, which performspedestal subtraction and zero suppression for each channel, and then are storedin the memory. This procedure takes 10 ms.

In order to estimate the dead time of the instrument, Monte Carlosimulations of the performance of the data acquisition system were carriedout, assuming a Poisson distribution of the particle fluxes. Figure 3.18 showsthe ratio (y) between the estimated dead time and the observational time ofthe instrument as a function of the measured rate (x). This ratio is fitted withan exponential function (y = yo + Aexp(x/τ)) and the parameters of the fithave been used in the exposure time program.

The uncertainty on the fit parameters is of the order of 10%, and this yieldsa maximal inaccuracy of 2% on the exposure time calculations for all eventscharacterized by intense flux, like SEP events and trapped particles in SouthAtlantic Anomaly.

3.7 Conclusions

NINA and NINA-2 were operative in orbit from July 1998 to April 1999and from July 2000 to August 2001 respectively. The characteristics of theinstruments and the orbits of satellites gave the possibility to study differentcomponents of cosmic rays, which will be discussed in the following Chapters.

In Chapter 4 I will report the energy spectrum of protons of albedo originmeasured by the instruments NINA and NINA-2 at different geomagneticlocations, and the behaviour of the proton flux as a function of longitudeand altitude out of the South Atlantic Anomaly.

In Chapter 5 I will report the energy spectra and abundance ratios ofhydrogen and helium isotopes of albedo origin, measured by instruments NINAand NINA-2 in near equatorial regions.

In Chapter 6 hydrogen and helium isotopes geomagnetically trapped in theSouth Atlantic Anomaly will be presented.

In Chapter 7 I will report results on nine Solar Energetic Particle eventsdetected by the instrument NINA between October 1998 and April 1999 andpreliminary results on fourteen Solar Energetic Particle events detected byNINA-2 instrument between October 2000 and August 2001.


Figure 3.18: The ratio between the dead time and the observation time of NINA,as a function of the external rate. An exponential fit has been super-imposed.

Chapter 4

Albedo protons

In this section I report the energy spectrum of protons of albedo originmeasured by the instruments NINA and NINA-2 at different geomagneticlocations, and the behaviour of the proton flux as a function of longitudeand altitude out of the South Atlantic Anomaly.

A detailed understanding of the fluxes of charged particles in near Earthorbit is important to reach an accurate theoretical description of the Earth’smagnetic field, but also as input for the calculation of the background forscientific instruments aboard satellites, like the future AGILE and GLAST γastronomy telescopes.

4.1 Albedo and quasi-trapped particles

It is well known that the interaction of cosmic rays with residual atmosphereproduces secondary particles. The secondary particle flux is mainly composedof electrons, positrons, protons, neutrons and gamma-rays, but also lightnuclei. Some fraction of secondary charged particles, with rigidity less thanthe local geomagnetic cutoff, can travel backward along the magnetic field lineconnected to the point of observation. Depending on their pitch angle, suchparticles can only make one bounce (and are called albedo particles) or morethan one bounce (quasi-trapped particles) in the geomagnetic field. There isalso a fraction of secondary particles that can remain trapped for several years.

It is interesting to recall the relation between the flux of sub-cutoff particlesobserved at the altitude of the satellite orbit and the flux of trapped particles inthe radiation belts. In the case of albedo particles the injection of the particlesin the population is ”direct”, in the sense that they are generated either atthe primary particle interaction or in subsequent shower. On the contrary,the trapped particles in the radiation belts, at energy more then 10 MeV, aregenerated by the Cosmic Ray Albedo Neutron Decay [White 1973, Walt 1994].Since the neutron lifetime is of the order of 15 minutes, its decay point, thatis the creation point of a particle in radiation belts, is only weakly correlatedwith the cosmic ray shower.

117

118 CHAPTER 4. ALBEDO PROTONS

4.1.1 Previous measurements

The cosmic ray albedo was discovered in 1949 when Van Allen, Singer,Winckler and co-workers used sounding rockets [Van Allen et al. 1950,Singer 1950, Singer 1950b] and balloon [Winckler et al. 1950] to show thatthe average omnidirectional particle flux exceeded the average vertical flux,contrary to the expectations if all observed particles were primary cosmicrays, incident upon Earth from interplanetary space. The historical definitionof ”cosmic ray albedo” corresponded to a population of secondary particlesproduced in cosmic rays shower, that could reach high altitude and reappear tothe opposite hemisphere at quite the same geomagnetic longitude and oppositegeomagnetic latitude.

Cosmic ray albedo was widely studied in the 1950s and 1960s, as researcherssorted out these backgrounds to uncover that spectrum of the primary cosmicradiation [Tylka 2000 and references in it].

It was found that a large fraction of the primary radiation at the equatorialconsists of returning albedo particles and that for latitude increasing from theequator, the fraction of albedo radiation returning to the Earth becomes ofsmall importance. At the poles, the intensity of albedo particles vanished[Perlow 1952].

In the cosmic ray literature, two kinds of albedo are distinguished, basedupon the direction motion near the top of the atmosphere:

• splash albedo, which refers to upward moving particles emerging fromthe atmosphere

• re-entrant albedo, which refers to downward moving particles whoserigidities are below the local geomagnetic cutoff for particles arrivingdirectly from interplanetary space.

The first which enumerated these two classes of albedo particles wasS. B. Treiman in a seminar paper in 1953 [Treiman 1953]. In his paper,Treiman also explained the close relationship between these two particlespopulations: the re-entrant albedo are simply splash albedo particles whichleave the production site along forbidden Stormer trajectories1. As aresult, these particles return to Earth very close to the same geomagneticlongitude and to the opposite geomagnetic latitude, i. e. to the conjugatemirror point in the opposite geomagnetic hemisphere. Some balloonmeasurements [Verma 1967, Pennypacker et al. 1973] showed that the upward(splash albedo) and downward (re-entrant albedo) fluxes are nearly identical.

In 1980, studies about albedo protons were carried by the Cosmos-555 satellite. Data were obtained at heights of 200-300 km by Cerenkov-scintillation telescopes, in directions both towards and away from the Earth

1Forbidden trajectories are those for which a particle cannot reach Earth from infinity.Conversely, these same directions are those for which splash albedo particles cannot escapeto infinity and hence must be channeled back to Earth.

4.1. ALBEDO AND QUASI-TRAPPED PARTICLES 119

in the energy range 0.4 – 1.2 GeV. Albedo particles, produced by the primarycosmic rays in the Earth’s atmosphere, were found to play a significant role informing the excess radiation at these heights [Kurnosova et al. 1980].

In 1985, the flux and the energy spectrum of low energy (30-100 MeV)albedo proton have been observed for the first time in a low latitude region,over Hyderabad, India, during a high altitude Balloon flight. It was used acharged particle telescope, capable of recording simultaneously the data ofboth upward as well as downward moving particles [Verma et al. 1985].

In June 1998, flux of secondary protons were detected by the AMSinstrument, on board the Space Shuttle, which went into space at an altitudeof 380 km. Below the geomagnetic cutoff a substantial second spectrum wasobserved. Their results showed that flux of sub-cutoff protons extended from0.1 GeV to ∼ 6 GeV, was nearly constant in the interval 0.3 < |λm| < 0.8 andincreased by a factor 2 in the interval 0 < |λm| < 0.3 .

In the range 0 < |λm| < 0.8, detailed comparison in different latitude bandsindicated that the upward and downward fluxes was nearly identical, agreedingwithin 1%. At polar latitude, |λm| > 1.0 the downward second spectrum wasgradually obscured by the primary spectrum, whereas the second spectrum ofupward protons was clearly observed.

Tracing back detected protons through the geomagnetic field, it was foundthat approximately 30% flew for less than 0.3 s (”short-lived”) , while theremaining 70% flew for times greater than 0.3 s (”long-lived”) before detection[Alcaraz et al. 2000].

Subsequent theoretical speculations, which investigated the behaviour ofprimary and secondary particles in the geomagnetic field, confirmed thesecondary origin of the particles measured by AMS [Derome et al. 2000,Lipari 2002]. These works showed that the high intensity flux of protonsobserved below the geomagnetic cutoff by the AMS experiment could wellbe reproduced by assuming that this flux originated from the interaction ofthe primary proton cosmic ray with the atmosphere; secondary particles couldbe injected into trajectories that reached high altitude as ”albedo particles”.


Figure 4.1: Splash proton albedo measurements at λm ∼ 41o. Results are color-coded, with captions giving the date of the measurement and the citation.


Figure 4.2: Splash proton albedo measurements at higher latitudes λm > 52o.Results are color-coded, with captions giving the date of the measurement and thecitation.


4.1.2 Comparison of balloon-borne and AMS protonalbedo measurements

Splash (Upward moving) Proton Albedo

Figures 4.1 and 4.2 show measurements of splash (upward-moving) protonalbedo at two different geomagnetic latitudes. Figure 4.1 compares balloonmeasurements [Verma 1967, Wenzel et al. 1975, Pennypacker et al. 1973] atPalestine, Texas (λm ∼ 41o), and the AMS results [Alcaraz et al. 2000] for thelatitude bin at 40.1 o < λm < 45.9 o.

Of particular note are the Verma [1967] results, which were obtained inMay 1965, almost exactly 33 years before the AMS measurements and henceat nearly the same point in the Solar Cycle. The Verma [1967] study is alsonoteworthy for its methodology. A first balloon flight was performed on 20-21 May 1965, with the instrument oriented toward zenith to observe only re-entrant albedo particles. (The geomagnetic cutoff at Palestine, Texas excludedprimary cosmic rays within the instrument’s energy range.) The same balloonpayload was flown again on 29-30 May 1965, but with the apparatus invertedand oriented vertically toward Earth, so as to observe splash albedo. This exactsame methodology was employed by AMS in 1998, which compared data takenwith the Shuttle bay oriented either toward zenith or toward nadir. All of themeasurements above ∼ 100 MeV are in remarkably good agreement2.

Figure 4.2 compares splash proton albedo measurements at somewhathigher geomagnetic latitude. The balloon results were all taken at FortChurchill, Canada. The AMS results in this Figure are the reported resultsin their highest latitude bin, at 51.5o < λm < 57.3o. Again, the results arein reasonably good quantitative agreement, except for the July 1969 results,which are somewhat lower, perhaps because of solar-cycle modulation of thecosmic-ray intensity [Tylka 2000].

Re-Entrant (Downward-Moving) Proton Albedo

Figure 4.3 compares two balloon measurements of the re-entrant (downward-moving) protons at Palestine, Texas [Verma 1967, Pennypacker et al. 1973].As noted by these authors, to within measurement errors, the downward-moving re-entrant albedo protons have the same spectrum and intensityas the upward-moving splash albedo. Also shown in Figure 4.3 is atheoretical calculation of the re-entrant proton albedo at Palestine, as givenby Pennypacker et al. [1973] and based on the work of Ray [1962; 1967].It should be noted that this is an absolute prediction, which has not beennormalized to the data. The agreement with Ray’s calculations is good, andthis comparison illustrates the quantitative understanding of cosmic-ray albedowhich was achieved in the 1960s. Before comparing to space-based re-entrant

2Wenzel et al. [1975] commented that the discrepancies below 100 MeV are too largeto be due to solar cycle effects and suggest that they may be due to the poorer energyresolution and higher backgrounds at low energies in the Verma (1967) instrument.


Figure 4.3: Re-entrant proton albedo measurements at ∼ 41o magnetic latitude,before correction for atmospheric overburden. Results are color-coded, with captionsgiving the date of the measurement and the citation. The blue code represents atheoretical calculation based on Ray [1967].


Figure 4.4: Comparison of AMS and balloon re-entrant proton albedo measurementsat ∼ 41o magnetic latitude, after the balloon results have been corrected for energyloss in the residual atmosphere.


Figure 4.5: Energy range covered by the different experiments versus the activityperiod.

albedo measurements, the balloon results must be corrected ”to the top ofthe atmosphere”, by accounting for energy loss in the residual atmosphere(typically a few g/cm2) above the balloon. Only Verma [1967] explicitly carriedout such corrections, and his corrected re-entrant albedo proton results arecompared to the AMS data in Figure 4.4. Again, the agreement appears to bequite reasonable.

As discussed above, the measurements of AMS cover the energy intervalfrom 0.1 to 200 GeV. Below 100 MeV no recent accurate measurements ofalbedo protons above the atmosphere are available.


4.1.3 The aims of the work

The study of secondary fluxes is an excellent test of the quality of our currentdescription of geomagnetic effects, including an accurate description of theEarth’s magnetic field and of the particle propagation and interaction in it.A detailed understanding of the fluxes of charged particles in near Earthorbit is even more important as it serves as input for the calculation ofthe background for future satellite instruments, like the planned AGILE andGLAST γ astronomy telescopes. Figure 4.5 shows the energy range coveredby the different experiments versus the activity period. All sensitivities areat 5σ. AGILE and GLAST instruments plan to be operative in the energyregion starting from ∼ 10 MeV up to several hundred GeV and they will coveran interval where no other experiments will be present. Their main scientificobjectives are the study of Active Galactic Nuclei, Gamma Ray Bursts, Pulsar,Galactic and Extragalactic Diffuse Emission, Solar Gamma Flares, gamma raytransient and unidentified galactic and extragalactic sources. They will bothfly on equatorial orbits, at an altitude of about 500 km. For such orbits,the charged particle background is mainly composed of electrons, positrons,primary and secondary protons. A complete knowing of background will allowto optimize trigger algorithms to use on board.

The low-energy data added by NINA and NINA-2 to AMS results will showthat the energy spectrum of secondary protons at low energy cannot be simplyextrapolated by a power-law from AMS data.

NINA and NINA-2 performed extensive measurements of the proton fluxbetween 10 and 35 MeV. With this new data the knowledge of the secondaryproton flux at low energies will be precise, leading to accurate estimations ofbackground for gammas.

4.2 NINA results, interpretation and discus-

sion

Along its polar orbit, NINA can measure particles of galactic and solar origin,albedo particles coming from secondary production, as well as particles trappedin the radiation belts.

To separate the different families of particles, selections based on L− shelland B are used; in particular, our analysis has shown that a pure albedocomponent without contamination from trapped, galactic and solar particles,even during intense SEP events, is selected by L−shell < 3 and B > 0.26 G inNINA’s energy range. At L−shell < 3, in fact, the geomagnetic cutoff is muchhigher than the particles rigidity in NINA’s energy interval, so galactic protonsof energy E < 100 MeV/n can reach neither Resurs nor MITA orbit. At smallaltitudes it is possible to observe trapped particles only inside South Atlantic

4.2. NINA RESULTS, INTERPRETATION AND DISCUSSION 127

Figure 4.6: NINA-2 data acquisition in function of latitude and longitude.

Anomaly, where radiation belts are close to Earth’s surface. Investigating thebehaviour of fluxes as a function of B for MITA orbit at a fixed L − shell, itwas found that the trapped population extends to B values up to 0.26 G (0.23G for Resurs-01-N4); at higher values of B flux does not depends on B. Thecut on B operates so to keep away the trapped component. It should be notedthat, due to the orientation of the satellites, measured local pitch angles covera range between 20o and 60o in the chosen region, with average value of about45o. Such pitch angle values correspond to albedo particles, and that confirmthe validity of our selection. In addition it is known that the geomagneticcutoff can vary in case of intense SEP events, but during NINA and NINA-2lifetimes none of these events changed appreciably the particle count rate atL − shell < 3.

Figure 4.6 shows NINA-2 data acquisition in function of latitude andlongitude. Count rates increases in South Atlantic Anomaly and in polar


Figure 4.7: NINA-2 data acquisition in function of latitude and longitude after theL-shell and B selection discussed in §4.2.


regions as well as the decreasing of geomagnetic cutoff, but data acquisition ispresent in equatorial regions too. After the L-shell and B selection discussedabove, particles detected in South Atlantic Anomaly and in polar regions arekept away as shown in Figure 4.7.

4.2.1 Reconstructed mass of albedo hydrogen

Figure 4.8 presents the mass reconstruction through methods described inChapter 3 of particles with unitary charge, detected by NINA-2 in therestricted region L-shell < 3 and B > 0.26 G (the region shown of Figure4.7). The mass resolution obtained in orbit is slightly worse than in beamtests, because electronic calibration during the flight, done automatically, canoccasionally take place in the South Atlantic Anomaly where the intense countrate affects the measurements. From Figure 4.8 it is evident that NINA-2detected the three hydrogen isotopes. Secondary nuclei of deuterium above theatmosphere were also observed by the AMS experiment [Lamanna et al. 2001]at energy greater than 150 MeV/n. As shown in Figure 4.8, in NINA’s energyrange deuterium and tritium constitute a non-negligible component, so a massdiscrimination at least as good as NINA and NINA-2 have is needed in orderto select a pure sample of protons.

Estimations of secondary background produced by cosmic rays inside theinstrument by means of the method described in Chapter 3 show that itscontribution to the proton count rate is less than 5%.

4.2.2 Proton flux

Figure 4.9 presents the proton flux, averaged in longitude, reconstructed inthe restricted region L-shell < 3 and B > 0.26 G, by NINA-2 at different L-shell ranges. NINA-2 default operating mode, thanks to its improved onboardcomputer capabilities was Low Threshold mode, thus extending the sensitiveenergy window for protons. The NINA-2 results include data collected in bothzenith and Sun orientation, as no statistically significant difference betweenthe two orientations was found. In addition, no distinction was made betweenalbedo and quasi-trapped particles.

As a comparison, Figure 4.10 presents the proton flux, averaged inlongitude, reconstructed in the region L-shell > 3, by NINA-2 at differentL-shell ranges. Fluxes show weak L-shell dependence in 4 < L-shell < 7, andincrease by a factor no more then ∼ 1.3 at low energy at L-shell > 7.

Figure 4.11 shows proton flux, averaged in longitude, as a function of L-shell, obtained by data of Figures 4.9 and 4.10 at three different energy ranges.At L-shell < 3 no statistically significant differences in energy are found, while


Figure 4.8: Mass distribution of particles with charge Z=1 at L-shell < 3 andB > 0.26 G, as measured by NINA-2.


Figure 4.9: Proton fluxes, averaged in longitude, reconstructed in the restrictedregion L-shell < 3 and B > 0.26 G, by NINA-2 at different L-shell ranges.


Figure 4.10: Proton fluxes, averaged in longitude, reconstructed in the region L-shell > 3, by NINA-2 at different L-shell ranges.


Figure 4.11: Proton flux, averaged in longitude, as a function of L-shell, at threedifferent energy ranges.


1.0 < L < 1.1 1.1 < L < 1.7 1.7 < L < 2.610 (4.8± 2.1)×10−6 (3.3± 1.3)×10−6 (16.0± 3.7)×10−6

14 (3.9± 1.3)×10−6 (5.0± 1.0)×10−6 (13.0± 2.2)×10−6

18 (3.4± 1.4)×10−6 (4.7± 1.5)×10−6 (12.0± 2.5)×10−6

22 (2.1± 1.2)×10−6 (5.3± 1.4)×10−6 (14.0± 2.9)×10−6

26 (3.8± 1.9)×10−6 (6.1± 1.7)×10−6 (12.0± 3.2)×10−6

30 (1.3± 0.9)×10−6 (6.3± 2.0)×10−6 (9.9±3.3)×10−6

Table 4.1: Table of proton fluxes (in particles/cm2 s sr MeV/n) measured by NINA-2 as a function of energy (rows, in MeV) in different L-shell intervals (columns) atL-shell < 2.6 and B > 0.26 G.

2.6 < L < 3.0 3.0 < L < 4.0 4.0 < L < 5.010 (2.2± 0.3)×10−5 (12.0± 0.1)×10−5 (32.6± 0.1)×10−5

14 (2.0± 0.2)×10−5 (8.9± 0.1)×10−5 (18.3± 0.1)×10−5

18 (1.7± 0.2)×10−5 (7.0± 0.1)×10−5 (11.5± 0.1)×10−5

22 (1.7± 0.2)×10−5 (5.3± 0.1)×10−5 (6.9± 0.1)×10−5

26 (1.4± 0.3)×10−5 (3.7± 0.2)×10−5 (4.2± 0.2)×10−5

30 (1.2± 0.3)×10−5 (2.5± 0.2)×10−5 (3.1± 0.2)×10−5

Table 4.2: Table of proton fluxes (in particles/cm2 s sr MeV/n) measured by NINA-2 as a function of energy (rows, in MeV) in different L-shell intervals (columns) at2.6 < L-shell < 5.0 and B > 0.26 G.

5.0 < L < 6.0 6.0 < L < 7.0 L > 7.010 (36.9± 0.1)×10−5 (37.7± 0.1)×10−5 (45.3± 0.1)×10−5

14 (19.2± 0.1)×10−5 (19.9± 0.1)×10−5 (24.5± 0.1)×10−5

18 (11.2± 0.1)×10−5 (12.8± 0.1)×10−5 (14.1± 0.1)×10−5

22 (6.9± 0.2) ×10−5 (6.8± 0.2)×10−5 (7.9± 0.2)×10−5

26 (4.6± 0.2) ×10−5 (4.6± 0.2)×10−5 (5.6± 0.2)×10−5

30 (3.2±0.2) ×10−5 (3.8± 0.2)×10−5 (4.0± 0.2)×10−5

Table 4.3: Table of proton fluxes (in particles/cm2 s sr MeV/n) measured by NINA-2 as a function of energy (rows, in MeV) in different L-shell ranges (columns) atL-shell > 5.0.


at L-shell > 3 fluxes decrease with energy. At 4 < L-shell < 7 fluxes weaklydepend on L-shell, and at L-shell > 7 an increasing by a factor no more than∼ 1.3 at all three energy ranges is measured.

Tables 4.1, 4.2, 4.3 report the NINA-2 proton fluxes, averaged in longitude,as a function of energy (rows, in MeV) at six different energies, in nine differentL-shell ranges (columns).

4.2.3 Longitude behaviour

The behaviour of fluxes with longitude was investigated, and some interestingvariations were found. The level of such variations is a factor 2, with theminimum value reached at the East boundary of the South Atlantic Anomaly.To obtain the longitude behaviour of proton fluxes we select a restricted regionof L-shell coordinate, and calculate the intensity of proton, averaged in energy,as a function of geographic longitude. Figures 4.12 and 4.13 show the protonintensity, averaged in energy, at 1.1 < L-shell < 1.7 and 1.7 < L-shell < 2.6respectively, and B > 0.26 G, as a function of longitude. Minimum value isreached at the East boundary of the South Atlantic Anomaly in both L-shellranges.

The AMS collaboration has presented its results on the cosmic ray spectraat different intervals in magnetic latitude for the detector position, integratingover the detector longitude. However, the observation of the secondaryparticles has given evidence of striking patterns in the longitude distributionof the creation points of the secondary particles. Subsequent theoretical worksinvestigated the possible dependence of the flux intensity on the detectorlongitude. A very clear explanation was carried by Lipari [2002], as resultsof non trivial but well known argumentations [Rossi 1933, Johnson 1933].

The ”long lived” positively (negative) charged particles observed in themagnetic equatorial region have their points of origin in the longitude rangeφ ∈ [120o, 300o] (φ ∈ [−60o, 120o]). This is an immediate consequence of threesimple observations:

• the geomagnetic field dipole axis is not only tilted, but also offset

• the motion of charged particles confined to the equatorial plane of amagnetic dipole can be analyzed as a ”gyration” around a guiding centerthat drifts uniformly in longitude remaining at a constant distance fromthe dipole center

• positively (negative) charged particles drift westward (eastward)

Figure 4.14 shows that positive particles can be injected into albedotrajectories only if produced in one hemisphere, and are absorbed in theopposite hemisphere, while the opposite happens for negatively charged


Figure 4.12: Proton intensity, averaged in energy, at 1.1 < L-shell < 1.7 and B >0.26 G as a function of longitude.


Figure 4.13: Proton intensity, averaged in energy, at 1.7 < L-shell < 2.6 and B >0.26 G as a function of longitude.


Figure 4.14: Illustration of the properties of the particle trajectories confined in theequatorial plane of an offset dipole model. The thick circle represents the intersectionof the Earth’s surface with the dipole equatorial plane. Both the Earth’s center (point0) and the dipole center (point D) are in this plane. The thin circles represent thetrajectories of the guiding center of the charged particles orbits. These trajectoriesremain at a constant distance rdip from the dipole center, and therefore have avariable distance r from the Earth’s center. Positively charges particles producedat the point b and d will be rapidly reabsorbed, while those produced at the point aor c can have a long flight time reaching point b and d.


particles, since the longitude drift of the guiding center of the trajectorytraveling at a constant distance from the dipole center has a variable altitude,that begins to decrease for particles created in the ”forbidden” hemisphere,or to increase for particles created in the ”allowed” one. The argument canbe easily extended in the general case of a trapped trajectories, when thelongitude drift is accompanied by an oscillation or ”bouncing motion” alongthe field lines between symmetric mirror points. The altitude of the guidingcenter of the trajectory has minima at the mirror points that have a constantdistance from the dipole center (here particles have the highest chance of beingabsorbed). In a offset dipole field the altitude of the mirror points change withthe longitude drift.

We can assume the origin of longitude as:

φ = φ − φ∗ , (4.1)

where φ∗ 120o is the longitude of the dipole center seen from the Earth’scenter and use the convention that the shifted longitude is defined in theinterval [−π, π]. It is simple to see that the source and sink regions forpositively and negatively particles are confined in longitude

[Source]+ [Sink]− φ > 0 (4.2)

[Source]− [Sink]+ φ < 0 (4.3)

and to predict an approximate one to one correspondence between thecreation and absorption points of long lived particles. The guiding center of theparticle trajectory remains at a constant distance from the dipole center duringthe drift motion. For particles produced in the ”allowed” hemisphere range rstarts increasing, and soon absorption in the atmosphere become impossible.The growth of the altitude of the trajectory guiding center will continue untilthe shifted longitude is φ 0, then symmetrically it starts to decrease. Whenthe longitude becomes

φf

+ = −φi

+ (4.4)

the particle will again be grazing the atmosphere, and will be absorbed. Itis also obviously true for a negative charged particle.

For both, the total longitude drift is

|(∆φ)±drift| 2|φ±i | = 2|φ±

f | . (4.5)

A particle produced with shifted longitude close to |φ| π drift for alongitude interval close to 2π that is approximately an entire Earth orbit.

In his work, Lipari showed a non trivial longitude dependence of fluxintensity. For a qualitative understanding let us consider a detector at a


position with shifted longitude φdet. As discussed above we can infer thatwhen φdet > 0 the only observable long lived particles are those produced inthe longitude interval

φi

+ ∈ [φdet, π] (4.6)

if positively charged.

The size of the visible regions is

|(∆φ)±drift| = π − |φdet| (4.7)

and is the same for negative particles. The size of the region is stronglydependent on the detector position, and reaches its maximum when thedetector is at φdet 0, where the entire production region is visible, andits minimum at φdet π.

If we take into account the altitude of detector, the simple observationin which one can conclude that fluxes of albedo particles are linear in φdet

vanishes. The altitude of the guiding center of the albedo particles also dependon the longitude. It is lowest at the creation and absorbtion points and it ishighest at φ 0. For a detector at a fixed altitude, only a fraction of thealbedo flux will be visible, some part of it being too low and some part beingtoo high.

These two arguments (the ”visible longitude horizon” and the ”altitude ofthe guiding center”) can explain the structure of the numerical results obtainedby NINA-2. In the work by Lipari [2002] behaviours similar to NINA-2 results,using Monte Carlo simulation, were obtained.

NINA-2 results show some variations with a factor of 2 at different longituderanges in albedo fluxes, with the minimum value reached at the East boundaryof the South Atlantic Anomaly. Calculations show adequate agreements withour results in the location of maximum and minimum regions.

4.2.4 NINA, NINA-2 and AMS correlations

To correlate NINA with AMS results, a selected sample of data are taken.

Figure 4.15 presents the proton flux, averaged in longitude, reconstructedin sub-cutoff regions by NINA and NINA-2 at different L-shell ranges. NINAdefault operating mode was High Threshold while NINA-2, as discussed above,thanks to its improved onboard computer capabilities, could operate in LowThreshold mode, thus extending the sensitive energy window for protons.

The results from NINA and NINA-2 at ∼ 13 MeV are in good agreement.The two experiments flew in different periods of the 23rd solar cycle. Theagreement confirms that the albedo component comes from interaction of highenergy cosmic rays - above geomagnetic cutoff - with the atmosphere, and so


Figure 4.15: Differential energy spectra of secondary protons, as measured by NINAand NINA-2, at different L-shell ranges. For comparison, data from the AMSexperiment are also presented.


it is only marginally affected by solar modulation. In the same picture, AMSresults at higher energy (E>100 MeV) are presented too [Alcaraz et al. 2000].

The low-energy data added by NINA and NINA-2 show that the energyspectrum of secondary protons at low energy cannot be simply extrapolatedby a power-law from AMS data.

4.2.5 Altitude behaviour

As discussed in Chapter 3, MITA altitude decreased considerably in the lastmonths of its lifetime in space. From an initial altitude of about 460 km, inJuly 2001, satellite decreased to about 200 km, and this allowed monitoringthe albedo proton flux at different altitudes.

Figure 4.16 shows the proton flux versus altitude measured by NINA andNINA-2 across the energy range 12–16 MeV, at two different L-shell intervalsand B > 0.26 G, corresponding to albedo fluxes; the solid points represent themeasurement taken by NINA-2 at different altitudes, while the open pointscome from NINA at the altitude of about 830 km. From this picture it ispossible to infer that the proton flux at ∼ 10 MeV does not depend appreciablyon the altitude along a fixed L-shell, in the range from 200 km to 850 km.In literature, it was found a similar behaviour between 55 km and 161 km,using a single Geiger counter on board a V-2 rocket at geomagnetic, by VanAllen and Tatel [1948]. In the work by Alcazar et al. [2000] it was shown thatquasi-trapped protons detected by AMS at fixed altitude originated from arestricted geographic region. It is plausible to think that the structure of thisregion depends on the acceptance angle of the detector and its orientation,as well as on the altitude. Small variations in fluxes measured by NINA andNINA-2 at different altitudes, like those visible in Figure 4.16, could thereforebe attributed to this [Bidoli et al. 2002].

4.3 Summary

Energy spectra of secondary sub-cutoff protons were measured at differentgeomagnetic latitudes across the energy range 10–35 MeV by the spaceinstruments NINA and NINA-2. The good instrument capabilities allowedmeasurements of the spectra of protons with high accuracy, also taking intoaccount the contribution of deuterium and tritium amounting to about 10%.The proton energy spectra were found to be practically flat in this energy rangeand not to vary appreciably with altitude. In addition, the measured flux wasapproximately constant between 1998 and 2001.

4.3. SUMMARY 143

Figure 4.16: Proton flux versus altitude measured by NINA and NINA-2 acrossthe energy range 12–16 MeV, at two different L-shell intervals and at B > 0.26 G,corresponding to albedo fluxes.

Chapter 5

Albedo light isotopes

In this Chapter I report the energy spectra and abundance ratios of hydrogenand helium isotopes of albedo origin, measured by instruments NINA andNINA-2 in near equatorial regions.

NINA and NINA-2 measurements revealed that 2H, 3H, 3He and 4He are asignificant portion of the secondary flux above the atmosphere.

5.1 Secondary production in the atmosphere

The flux of high energy particles produced by nuclear interactions of cosmicrays with residual atmosphere is composed mainly by electrons, positrons,protons, neutrons and gamma-rays, but also light nuclei are present. Secondaryhydrogen and helium isotopes were observed by balloon experiments indifferent energy ranges [Teegarden 1967b, Webber and Yushak 1983] and werestudied as background component during measurements of primary cosmicray nuclei [Boezio et al. 2001b, Matsunaga et al. 1998]. The two mainmechanisms that produce secondary light nuclei are air target nuclei break-up and fragmentation of incident nuclei. For energies lower than 100 MeV/n,the first reaction is dominant [Papini et al. 1996, Vannuccini et al. 2001].

As discussed in Chapter 4, an appreciable fraction of secondary particles,with rigidity less than the geomagnetic cutoff, can travel backward in spacealong the Earth’s magnetic field and can be observed at satellite altitudes.Depending on their pitch angle, these particles oscillate in the geomagneticfield making only one bounce (so-called ”reentrant albedo particles”, absorbedin opposite hemisphere) or more than one bounce (quasi trapped albedoparticles). There is also a fraction of albedo particles that can remain trappedfor several years. As it was investigated by Freden and White [1960] thismechanism gives trapped light isotopes, but at least orders of magnitude lessthan interactions of trapped protons in the radiation belt with air nuclei.This second mechanism was later verified by detailed calculations done bySelesnick and Mewaldt [1996]. Their calculations are in good agreement withCRRES measurements of trapped He isotopes ratio [Chen et al. 1996] andwith measurements of trapped deuterium component done on board SAMPEX

145

146 CHAPTER 5. ALBEDO LIGHT ISOTOPES

satellite [Looper et al. 1996, Looper et al. 1998]. We refer to Chapter 6 ofthis thesis and to Bakaldin et al. [2002] for a complete discussion ongeomagnetically trapped light isotopes observed with the detector NINA.

On the other hand high energy particles along unclosed drift trajectoriesmay originate mainly from cosmic ray interactions with air nuclei. Asdescribed in Chapter 4, the first measurements of albedo particles beganin the second part of 1940s. At that time, no distinction in isotopecomposition was available. Some years later, several calculations of isotopesproduction rate in the atmosphere have been made [Fireman et al. 1955,Currie et al. 1956, Lal et al. 1962] and first accurate measurements of lightisotopes production in Earth’s atmosphere were carried through balloon flights[Teegarden 1967, Teegarden 1967b].

Recent measurements of the secondary light nuclear component abovethe atmosphere were carried out on board the Space Shuttle by the AMSinstrument [Alcaraz et al. 2000b, Lamanna et al. 2001]. Across the energyinterval from 0.15 to several GeV/n, presence of 2H and 3He nuclei wasdetected. The tracing back of such particles in the geomagnetic field showedthat they originate in atmosphere. Subsequent theoretical speculations,which investigated the behavior of primary and secondary particles inthe geomagnetic field, confirmed the secondary origin of these particles[Derome and Buenerd 2001, Lipari 2002]. However, interaction of cosmic raysshould produce also 3H and 4He, whereas tritium was not observed by thisexperiment and only a lower limit on the 3He/4He ratio at E > 100 MeV/nwas estimated (3He/4He ≥ 9).

Moreover, no accurate measurements of the isotope composition of albedoand quasi trapped secondary light nuclei above the atmosphere below 100MeV/n were available so far.

AMS results on helium component

Measurements of the energy spectra of helium spectrum from 0.1 to 100 GeV/nwere carried out on board the Space Shuttle by the AMS instrument. Abovethe geomagnetic cutoff the spectrum is parameterized by a power law. Belowthe geomagnetic cutoff a second helium spectrum was observed. In the secondhelium spectra, more than 90% of the helium was determined to be 3He.Tracing helium from the second spectrum shows that about half of the 3Hetravels for an extended period of time in the geomagnetic field and that theyoriginate from restricted geographic regions similar to protons.

AMS analysis showed a population of 115 helium events with rigiditiesbelow the local geomagnetic cutoff, having at the 90% confidence level, thefraction of 3He exceeding ninety percent in the energy range from 0.1 GeV/nto ∼ 1.2 GeV/n. The trajectories have been traced both backward and forwardfrom their incident angle, location and momentum, through the Earth’s

5.2. NINA MISSIONS: RESULTS AND DISCUSSION 147

magnetic field. All events were found to originate in atmosphere. Analysisof the sum of their forward and backward flight times yields two distinctclasses: ”short-lived” and ”long-lived” for flight times below and above 0.3s, respectively.

The origin of the ”short-lived” helium nuclei are distributed uniformlyaround the globe whereas the ”long-lived” particles originate from twogeographically restricted regions.

AMS results on deuterium component

A second spectrum of deuterons was detected below the geomagnetic cutoff byAMS experiment, showing a nearly constant behaviour over the energy range0.2< |λm| <0.6 and an enhancement of a factor 2 to 3 in the geomagneticequator (|λm| <0.2). At all latitudes the upward flux and the downward fluxare comparable within the errors. Again, the trajectories have been tracedboth backward and forward from their incident angle, location and momentum,through the Earth’s magnetic field. It was found that all deuterons originatein the atmosphere, and was distinguished in ”short-lived” (∼ 40%) and ”long-lived” (∼ 60%) for flight times below and above 0.3 s respectively. The originpoints of ”short-lived” and ”long-lived” are the same than for protons andhelium component.

Theoretical works

Some consequential theoretical works described mechanism of production ofsub-cutoff light isotopes. Since particles below geomagnetic cutoff cannot beprimary cosmic rays, they have to be produced by nuclear reactions betweenincoming cosmic rays (mainly 1H and 4He), and atmospheric nuclei (mainly14N and 16O). A reaction cascade can thus develop through the atmosphereand all charged particles undergo loss by ionisation. Secondary particles candisappears by nuclear collision, stop in the atmosphere by energy loss, or escapeto outer space [Derome and Buenerd 2001b, Llope et al. 1995].

It was found that AMS measurements of the 3He and 2H particle fluxcan be reproduced consitently and simultaneously by taking into account theinteractions between cosmic ray flux, Earth magnetosphere and atmosphere,and assuming the 3He and 2H particles to be produced by coalescence ofnucleons in the cosmic ray proton and 4He. The 4He fragmentation productsappear not to contribute at a detectable level to the flux measured below thegeomagnetic cutoff [Derome and Buenerd 2001c, Huang and Stephens 2001].

5.2 NINA missions: results and discussion

This section reports light isotope abundances detected by NINA and NINA-2telescopes in sub-cutoff regions. To separate the different families of particles,


selections based on L − shell and B coordinates as well as the ones describedin Chapter 4 are used.

All data presented in this section have been obtained in the restricted regionL − shell < 3 and B > 0.26 G.

Figure 5.1 presents the mass reconstruction for particles with chargeZ=1 (left panel) and Z=2 (right panel) detected by NINA-2. It is evidentthat NINA-2 detected the three hydrogen isotopes as well as 3He and 4He.Estimations of the secondary background for hydrogen and helium isotopesproduced by cosmic rays inside the instrument done with the method describedin Chapter 3, that considers the first silicon detector as a target for secondaryproduction, show that the contribution of this background in the count rateof nuclei is less than 8%.

Figure 5.2 presents deuterium and tritium fluxes, together with the protonone, averaged in longitude, reconstructed by NINA-2. Results include datacollected in both zenith and Sun orientation, as no statistically significantdifference between the two orientations was found in NINA-2 proton analysis[Bidoli et al. 2002]. In addition, no distinction has been made between albedoand quasi-trapped particles (i.e. between particles with different pitch angles).The hydrogen isotope fluxes in this energy region are rather flat.

Energy spectra of helium isotopes measured by NINA-2 instrument arepresented in Figure 5.3, together with AMS results [Alcaraz et al. 2000] andcalculations done by Derome and Buenerd [2001]. Fitting NINA-2 data withthe power law function E−γ, the extracted spectral index is 0.8 ± 0.2 for 3Heand 1.5 ± 0.2 for 4He.

Helium isotope spectra in this energy region are evidently muchsteeper than proton ones. This is easily understood considering that thecreation of the secondary flux takes place at an altitude of about 30–50 km [Derome and Buenerd 2001, Lipari 2002]. Before escaping from theatmosphere, secondary charged particles pass some quantity of material inair and lose energy. Differences in energy losses for hydrogen and helium dueto their different charge can explain the shape of their energy spectra.

Figure 5.4 shows the 3He/4He ratio as a function of energy for NINA,NINA-2 and AMS; the agreement between NINA and NINA-2 data is good.The average value of the 3He/4He ratio across the energy range 10–40 MeV/nfor both NINA experiments is 1.5. Data by NINA, NINA-2 and thelower limit by AMS show that the 3He/4He ratio increases from 1 to 9 across the energy range 10-1000 MeV/n. Though detailed calculationsare not available, the increase can be explained by the secondary originof helium particles. Similar results, in fact, are observed for the helium


Figure 5.1: Mass distribution of particles with charge Z=1 (left panel) and Z=2(right panel) at L − shell <3 and B>0.26 G, as measured by NINA-2.


Figure 5.2: Differential energy spectra of secondary protons, deuterium and tritiumat L − shell <3 and B>0.26 G, as measured by NINA-2.


Figure 5.3: Secondary 3He and 4He flux as a function of energy, at L−shell <3 andB>0.26 G, as measured by NINA-2. Open points represents the sub-cutoff 3He fluxintegrated over L − shell <1.47, as measured by AMS [Alcaraz et al. 2000]. Thickand dashed lines show simulation results corresponding respectively to the 3He and4He flux [Derome et al. 2000].


Figure 5.4: 3He/4He ratio at L−shell <3 and B>0.26 G, as measured by NINA-2.AMS lower limit for L − shell <2.2 is also shown.

5.3. SUMMARY 153

trapped component in radiation belts [see Chapter 6], component originatedby trapped proton interactions with residual atmosphere. As it was calculatedby Selesnick and Mewaldt, the 3He/4He ratio increases for trapped particlesapproximately in the same way as it does for albedo particles as seen by NINA[Selesnick and Mewaldt 1996].

Figure 5.5 shows the 2H/1H, 3H/1H and 3H/2H NINA-2 abundance ratiosas a function of energy. At low energy the 2H/1H ratio has a value of about15% and decreases with increasing energy. This result is much higher than the2H/1H ratio in the inner radiation belt measured by PET instrument on boardSAMPEX [Looper et al. 1996] and NINA instrument [Bakaldin et al. 2002],which both obtained a ratio of about 1 %.

At energy of about 150 MeV/n, 2H/1H ratio is between 2% and5%, as measured by AMS [Lamanna et al. 2001]. It is known that 2Hisotope is very rare in nature, typically with a concentration lower than10−5 [Mullan and Linsky 1998]. Albedo deuterium is probably the highestconcentration of deuterium existing in nature.

As for tritium, NINA-2 measured a 3H/2H ratio of about 0.25 and a3H/3He ratio of about 3 at 10 MeV/n. The high quantity of tritiumwith respect to 3He could be explained by the combination of the productioncross section and ionization energy losses before escaping. Similar processesoccur in radiation belts as mentioned above, where the 3H/3He ratio is about0.5 [Bakaldin et al. 2002]. NINA and NINA-2, differently from the recentmeasurements by AMS, detected an appreciable quantity of albedo tritium. Assaid before, air target nuclei break-up is the most probable process to explainlight isotope abundances at low energy. Such reactions produce also tritium.The quasi deuteron nuclei model suggests that the 2H/3H ratio increases withenergy, for energies greater than 30 MeV/n, and that can be the reason whyAMS did not detect tritium at high energy. Data from accelerators experimentsseem to confirm this assumption. For example, it was measured by Wu et al.[1979] and Bertrand and Peelle [1973] that for a 90 MeV proton projectile on27Al, 58Ni, 90Zr, 209Bi the average ratio 2H/3H was about 5, in agreement withNINA measurements. At the same time, a calculation at energy more 150MeV/n done by Derome and Buenerd [2001] gives a 2H/3H ratio between 20and 50.

5.3 Summary

NINA and NINA-2 detectors have verified the existence of fluxes of lightisotopes below the geomagnetic cutoff. NINA flew on board Resurs-01-N4satellite in a 830 km average altitude orbit between 1998 and 1999 and NINA-


Figure 5.5: 2H/1H, 3H/2H, 3H/1H ratios at L − shell <3 and B>0.26 G, asmeasured by NINA-2.

5.3. SUMMARY 155

2 flew on board the MITA satellite at an altitude of about 450 km between2000 and 2001.

The good mass resolution of NINA and NINA-2 detectors alloweddistinguishing hydrogen and helium isotopes, including 3H and 4He which hadbeen not observed in secondary fluxes above the atmosphere before.

NINA and NINA-2 measured the 2H/1H ( 15%), 3H/2H ( 25%), 3H/1H( 4%), 3H/3He ( 3) and 3He/4He ( 1.5) isotope abundance ratios,and reconstructed the differential energy spectra of all hydrogen and heliumisotopes of albedo origin across the energy interval 10–40 MeV/n.

Chapter 6

Geomagnetically trapped lightisotopes

The detector NINA aboard the satellite Resurs-01-N4 detected hydrogen andhelium isotopes geomagnetically trapped, while crossing the South AtlanticAnomaly. Deuterium and tritium at L − shell<1.2 were unambiguouslyrecognized. The 3He and 4He power-law spectra, reconstructed at L −shell=1.2 and B<0.22 G, have indices equal to 2.30±0.08 in the energy range12–50 MeV/n, and 3.4±0.2 in 10–30 MeV/n respectively. The measured3He/4He ratio, and the reconstructed deuterium profile as a function ofL − shell, bring to the conclusion that the main source of radiation beltlight isotopes at Resurs altitudes (∼ 800 km) and for energy greater than 10MeV/n is the interaction of trapped protons with residual atmospheric helium[Bakaldin et al. 2001b, Bakaldin et al. 2002].

6.1 Previous measurements

The trapped helium component in Earth’s radiation belts was discoveredat the beginning of the 60’s [Krimigis and Van Allen, 1967]. Its isotopiccomposition was studied many years later with the ONR-604 instrument onboard CRRES satellite, in the energy range 40–100 MeV/n [Wefel et al. 1995,Chen et al. 1996] and with the spectrometer MAST on board SAMPEX in theinterval 5–15 MeV/n [Selesnick and Mewaldt 1996]. It was found that 3He ismore abundant than 4He in the inner radiation belt, at least for L−shell<1.5.The CRRES measured 3He/4He averaged ratio in the energy interval 51–86MeV/n was 8.7±3.1 at L − shell 1.1÷1.5, and 2.4±0.6 at L − shell 1.5÷2.3[Wefel et al. 1995]; this ratio decreases as a function of energy in the firstL − shell range and increases in the second. At L − shell 1.1÷1.5 the power-law fit gives spectral indices equal to 5.9±0.4 for 3He and 4.1±0.3 for 4He. Amore detailed analysis [Chen et al. 1996] showed that 3He/4He averaged ratiowas 7.4±2.6 at L−shell 1.15÷1.3, and 2.2±0.6 at L−shell 1.8÷2.15 and thatall observed helium events at L− shell < 2.65 showed pitch angle distribution

157

158 CHAPTER 6. GEOMAGNETICALLY TRAPPED LIGHT ISOTOPES

consistent with equatorially trapped ions.

The 3He/4He ratio obtained on board SAMPEX at L−shell=1.2 is about 1at energy E9 MeV/n and increases with energy [Selesnick and Mewaldt 1996].This ratio was measured for L − shell up to 2.1, and it was found decreasingwith the geomagnetic latitude down to about 0.1.

Information about hydrogen isotopes is poorer. In the work by Fredenand White [1960] the observation of 4 trapped nuclei of tritium in the energyrange 126–200 MeV was reported. But at the same time this experiment,which utilized a small stack of nuclear emulsions, did not observe heliumand deuterium nuclei, and that was ascribed by the authors to differencesin isotope secondary productions, and to a high deuterium interaction crosssection. The first observations of real geomagnetically trapped deuterium inthe inner radiation belt were presented by Looper et al. [1996], performed byPET on board SAMPEX in the energy range 18–58 MeV/n. In this work itwas shown that, at the same energy per nucleon, the deuterium component isabout 1% of the proton one. A following work [Looper et al. 1998] reportedobservations of trapped tritium between 14 and 35 MeV/n at L − shell<1.2,with a flux equal to 1/8 of the deuterium one. At higher L−shell values it wasnot possible to identify tritium nuclei due to the increase of the instrumentnoise.

The first hypothesis about the origin of rare light isotopes in radiation beltswas that they are generated by the interaction of high energy trapped protonswith the residual atmosphere [Freden and White 1960]. This source is at leastone order of magnitude greater than that coming from cosmic ray interactionswith air nuclei.

To explain the significant difference between the 3He/4He ratios at L −shell1.2 and at L − shell1.9 measured by CRRES it was suggested inWefel et al. [1995] that the interaction region of protons lies at a loweratmospheric altitude for L − shell=1.2 than for L − shell=1.9. This impliesthat for secondary production it is necessary to take into account not onlyproton interactions with nuclei of helium in the upper atmosphere, but alsowith nuclei of oxygen which dominate at lower altitudes. In the interactions ofprotons with oxygen, the cross section of 3He production is much larger thanof 4He.

Detailed calculations by Selesnick and Mewaldt [1996], and comparisonswith SAMPEX experimental data, show that the atmospheric He source isadequate to account for a substantial fraction of the intensity of the heliumisotopes, at L − shell≤1.2 and particularly for 3He. For higher L-shells anadditional source of 4He is required. The model of proton interaction withatmospheric helium is also in agreement with 3He and 4He data collected byCRRES at L − shell=1.2.

While the reactions involving atmospheric helium are reasonably wellknown, the uncertainties in the cross section value for the reaction withatmospheric oxygen make it difficult to estimate the degree of importance ofthis additional source. If the interactions with atmospheric oxygen and helium

6.2. DATA ANALYSIS AND OBSERVATIONS 159

source are combined together, the model gives significantly more 4He andespecially 3He than observed by SAMPEX at L − shell=1.2. For deuterium,instead, the atmospheric He source appears to be somewhat weak to producethe intensities reported by SAMPEX, while the combined calculated intensitiesfor deuterium seem to better reproduce the data.

Selesnick and Mewaldt [1996] suggest that atmospheric oxygen can be asignificant source of light isotopes in the inner radiation belt near L−shell=1.2.At higher L-shells this source may be important only in a narrow range ofequatorial pitch angles near the edges of the loss cones, corresponding to thealtitude range where oxygen is the dominant component in the atmosphere.Moreover, due to the different energy spectra of secondary particles producedby proton interactions with O and with He, the oxygen source is relativelymore significant than the helium one only for low particle energies up to E ∼10–15 MeV/n.

Another possible mechanism discussed in literature is a radial diffusionof nuclei inside the magnetosphere. In the work by Pugacheva et al. [1998]the secondary production rate for 2H, 3H and 3He in the upper part of theatmosphere was compared with radial diffusion, and it appeared to dominateat L− shell in the interval 1.2÷1.6. The results of the calculations depend onthe detailed knowledge of some parameters, such as the radiation belt modelsand the boundary condition of diffusion equations.

In next sections we report observations of geomagnetically trappedhydrogen and helium isotopes, performed by NINA during passages over theSouth Atlantic Anomaly (SAA) in the months November 1998–April 1999,and we give possible interpretations about the origin of radiation belt lightisotopes.

About 95% of the NINA mission lifetime measurements were carried outin High Threshold mode, due to the background conditions (very high protonflux) in the radiation belts. With this configuration it is possible to detecthydrogen isotopes only in a narrow energy range (see Table 2.1).

6.2 Data analysis and observations

It has been already set out in Chapter 3 that satellite Resurs-01-N4 has anear-Earth polar orbit with inclination 98 and altitude 835 km. It makesabout 14 revolutions per day, and when the orbit crosses the South AtlanticAnomaly (about 7 times per day) NINA can measure particles trapped inthe inner radiation belt. The period of observation taken under considerationranges from November 1998 to April 1999; data include also particles collectedduring Solar Energetic Particle events, because even in conditions of intensesolar activity the fluxes measured by NINA in the inner radiation belt remainpractically constant.

The study of rare isotope abundances in SAA requires a careful analysisof the data collected. The background includes both noise and events not


identifiable, and particles produced by secondary interactions in the materialaround NINA. A dedicated off-line track selection algorithm, followed bycharge and mass identification procedures, identical to that described inChapter 3, was used [Bidoli et al. 2001].

NINA is accomodated on Resurs in order to point always towards thezenith. The angle accuracy of the satellite axis is about 1 degree. By theknowledge of the spacecraft orientation, and the incoming particle directionin the telescope, NINA is able to determine the particle pitch-angle with anaccuracy of about 5 degrees. Figure 6.1 presents the local pitch-angle (αloc)distribution for particles detected by NINA in the region L − shell<1.2 andmagnetic field B<0.22 G (SAA region). This distribution is peaked at 90degrees. The average value of αloc observed by NINA in this region correspondsto an average equatorial angle α0 of about 75 degrees. Calculations of particletrajectories with αloc of about 90 degrees, at Resurs altitudes and knowingthe L− shell value, show that these particles can be geomagnetically trappedbecause mirror points are higher than atmospheric altitudes (∼ 200 km). Theestimated lifetime for trapped 4He at L − shell=1.2 is more than one year[Selesnick and Mewaldt 1996].

Figure 6.2 (left) presents the mass distribution for 3He and 4He in the SAAregion, for data collected in the time period considered. It is possible to seethe good separation between isotopes, and the higher abundance of 3He withrespect to 4He, in agreement with previous observations [Wefel et al. 1995].Figure 6.2 (right) shows the mass distribution for hydrogen isotopes. Theshaded areas represent the background contribution calculated as well asdescribed in Chapter 3, for helium (left) and hydrogen (right). The comparisonbetween this contribution and the measured abundances shows that NINAdetected hydrogen and helium isotopes geomagnetically trapped.

6.3 Results and interpretations

The energy spectra of 3He and 4He obtained by NINA at L− shell 1.18÷1.22and B<0.22 G are presented in Figure 6.3, together with the data of MASTon SAMPEX at L − shell=1.2. Fluxes are averaged over local pitch anglesat ∼ 800 km and ∼ 600 km of altitude for NINA and MAST respectively.The 3He spectrum is fitted by a power-law with index equal to 2.30±0.08 inthe energy range 12–50 MeV/n, while 4He has a spectral index of 3.4±0.2 in10–40 MeV/n. The NINA and MAST data have been gathered during quitedifferent periods of the solar cycle; taking also into account the sharp decreaseof the trapped particles flux near the edge of the radiation belts, they show areasonable agreement.

The energy spectra of 3He and 4He presented in Figure 6.3 are in goodagreement with calculations for atmospheric helium source (dotted and solidline, for two different equatorial pitch-angles). The thick line is the sum of thep+He and p+O contributions, which seems to overestimate the 3He contents

6.3. RESULTS AND INTERPRETATIONS 161

Figure 6.1: Distribution of the local pitch angle αloc for particles detected in theregion L − shell<1.2 and B<0.22 G (SAA).


Figure 6.2: Mass distributions for geomagnetically trapped He (left) and H (right)isotopes measured at L − shell<1.2 and B<0.22 G. The shaded area represents theestimated background.


(p+He)

(p+He)+(p+O)

(p+He)

(p+He)+(p+O)

Figure 6.3: 3He (left) and 4He (right) differential energy spectra measured inside theSAA by NINA (squares) and MAST (circles). The solid and dotted lines representthe calculated helium flux at equatorial pitch angles α0 = 80 and α0 = 70

respectively, assuming atmospheric helium as the source of secondary production[Selesnick and Mewaldt 1996]. The thick line is the sum of the p + He and p + Ocontributions at α0 = 80.


(p+He)

(p+O)

Figure 6.4: 3He/4He ratio at L − shell=1.2 measured by NINA (squares), NINA-2 (triangles) and by ONR-604 (full squares) [Wefel et al. 1995]. The dotted linerepresents the calculated ratio for the reaction p + O. The solid and dashed linerepresents the calculated ratio for the reaction p + He at pitch angles α0=80 andα0=60 respectively [Selesnick and Mewaldt 1996].


with respect to the experimental measurements.

From fluxes of Figure 6.3 we derived the 3He/4He ratio as a function ofenergy, shown in Figure 6.4; the dotted line is the theoretical calculationfor atmospheric oxygen source at α0=80, while the solid and dashed linescorrespond to the helium source secondary production at two differentequatorial pitch angles, α0=60 and α0=80. From this picture it is possibleto conclude that the model of proton interaction with oxygen in atmosphereoverestimates the 3He production. That is not surprising because, asmentioned by the authors themselves [Selesnick and Mewaldt 1996], the lackof cross section data and the forward scattering approximation can probablylead for O source to an overestimate of the trapped particle source rate. In thesame Figure preliminary results on NINA-2 mission are shown. They confirmthe helium source secondary production of light isotopes and exclude the modelof proton interaction with oxygen in atmosphere.

NINA measured 3He/4He ratio, averaged in the energy interval 10–30MeV/n, is equal to 2.2±0.2 for local pitch angles αloc in the range 80 ÷ 90,and 2.1±0.4 for αloc 50 ÷ 70, so it does not show an appreciable variationas a function of the pitch-angle. Calculations indicate instead that theisotope intensity from an atmospheric source should peak at a pitch-anglecorresponding to the altitude where this source is the dominant atmosphericconstituent [Selesnick and Mewaldt 1996]. In the range αloc 50 ÷ 70 themirror points of the particles lie at altitudes where oxygen dominates inatmosphere, so this is a further indication that oxygen seems not to be thedominant source of secondary production in this energy range.

Fluxes of 2H and 3H obtained by NINA at L−shell 1.18÷1.22 and B<0.22G are presented in Figure 6.5, for the energy range shown in Table 2.1. Fluxesare compared to calculations for atmospheric helium source (dotted and solidline, for two different equatorial pitch-angles), and to the model that combinessecondary production in atmosphere (p + He and p + O) and radial diffusion(red line).

Our analysis reveals that the sum of the p+He and p+O reactions does notprovide a sufficient deuterium intensity, as already pointed out by Selesnickand Mewaldt [1996] and Pugacheva et al. [1998] who compared calculationswith SAMPEX data. Figure 6.5 presents in fact fluxes averaged over a L−shellinterval, but fluxes around the boundary of radiation belts have a very steepdependence on the L − shell value. More reliable measurements are thereforeratios between different species. Our measured 2H/1H ratio is roughly equalto 1% at energy about 10 MeV/n, and it is close to that observed by PETinstrument at 18–58 MeV/n [Looper et al. 1996]. The calculation of Selesnickand Mewaldt [1996] predicts instead a ratio of about 10−3 at these energies,for L − shell=1.2.

In order to estimate the 3He/3H ratio, since the two isotope measurementsspan two different energy regions in NINA detector (see Table 2.1), it isnecessary to extrapolate the 3He flux - presented in Figure 6.3 - in the tritiumenergy region. The 3He/3H ratio has a value of about 2.


(p+He)

combined

combined

Figure 6.5: Fluxes of 2H and 3H obtained by NINA at L − shell 1.18÷1.22 andB<0.22 G for the energy range shown in Table 2.1, compared to calculations foratmospheric helium source (dotted and solid line, for two different equatorial pitch-angles), and to the model that combines secondary production in atmosphere (p+Heand p + O) and radial diffusion (red line).


(p+O)

(p+He)

combined

Figure 6.6: Left: calculations for secondary production of deuterium at differentenergies for different reactions [Selesnick and Mewaldt 1996]. Right: experimentaldeuterium flux as a function of L−shell measured by NINA in the energy range 7–13MeV/n, together with calculations from the combined action of secondary productionand radiation belt radial diffusion at 10 MeV made by Pugacheva et al. [1998].


Figure 6.7: Fluxes of 3He (triangles) and 4He (squares) obtained by NINA-2 atL − shell <1.3 and B<0.21 G.

Figure 6.6 shows the behavior of the deuterium flux reconstructed by NINAas a function of L− shell. Data were normalized because experimental pointsrefer to pitch angles α0 ∼75, while the calculations were performed for α0=90.The picture presents also flux calculations for the reactions p + O and p+ He, as well as with the flux obtained combining secondary production inatmosphere and radial diffusion in inner radiation belts. The curves show thatthe contribution from the reaction p + O peaks at L−shell1.15, whereas thep + He production is equally distributed in the whole range of L− shell<1.3.Since NINA data do not show any evident peak in flux in the proximity ofL − shell1.15, this is again a confirmation that the reaction p + O plays aminor role in the formation of trapped light isotopes with energy greater than10 MeV/n in the inner radiation belt.

6.4. NINA-2 PRELIMINARY RESULTS 169

Figure 6.8: Fluxes of 2H (triangles) and 3H (squares) obtained by NINA-2 at L −shell <1.3 and B<0.21 G.


6.4 NINA-2 preliminary results

NINA-2 detector flew during the maximum of solar activity at an averagealtitude of 400 km. Preliminary data analysis showed some interesting topicsthat are outlined in sequence.

Figures 6.7 and 6.8 show the differential flux for helium and hydrogenisotopes in the restricted region B <0.21 G and L − shell <1.3.

From Figures 6.3, 6.5, 6.7 and 6.8 is clear that light isotopes flux measuredby NINA-2 is about one order of magnitude less than flux measured by NINA.It could be due to the different period of solar activity in which the two missionflew, but also to the different average altitude. In fact local pitch angle oftrapped particles was αloc ∼ 90o for the both mission, but due to the differentaltitude, the value of equatorial pitch angle was αo ∼ 70o for NINA and αo ∼60o for NINA-2.

A work with the complete analysis of NINA-2 results in trapped region isin preparation.

6.5 Summary

Results presented above show that rare light isotopes represent a distinctcomponent of the Earth’s inner radiation belt. Geomagnetically trappedfluxes of 2H, 3H, 3He and 4He were measured by the instrument NINA aboardthe satellite Resurs, for L − shell∼1.2 at about 800 km of altitude. The2H/1H ratio at energy 10 MeV/n was ∼ 0.01, the 3H/2H ∼ 0.2 and the3He/3H ratio ∼ 2. The 3He/4He ratio increases from ∼ 1 to ∼ 5 across theenergy range 10–40 MeV/n. The data analysis and the comparison with theavailable theoretical calculations bring to the conclusion that the light isotopescomponent is originated from the interactions of high energy trapped protonswith the residual atmosphere. At energy above 10 MeV/n, at the altitudeof Resurs, the interaction with atmospheric helium seems to dominate withrespect to oxygen.

The observed 2H/1H and 3H/2H ratios at energy 10 MeV/n are higherthan calculated from atmospheric production models. The results of thecalculations, however, depend on the detailed knowledge of some physicalparameters, such as the radiation belt models, the atmospheric density modeland cross section parametrisation, and also on the solar cycle phase. Themeasurements performed by NINA can provide new inputs to the theoreticalcalculations.

Chapter 7

Light Isotope Abundances inSEP

This Chapter reports nine Solar Energetic Particle events detected by theinstrument NINA between October 1998 and April 1999.

For every solar event the power-law 4He spectrum across the energy interval10–50 MeV n−1 was reconstructed, and spectral indexes, γ, from 1.8 to 6.8extracted. Data of 3He and 4He were used to determine the 3He/4He ratio, thatfor some SEP events indicated an enrichment in 3He. For the 1998 November 7event the ratio reached a maximum value of 0.33± 0.06, with spectral indexesof γ = 2.5 ± 0.6 and γ = 3.7 ± 0.3 for 3He and 4He, respectively. The 3He/4Heratio averaged over the remaining events was 0.011 ± 0.004.

For all events a deuterium-to-proton ratio was estimated. An upper limiton the average value over all events was 2H/1H < 4 × 10−5 across the energyinterval 9–12 MeV/n.

Upper limits on the 3H/1H counting ratio for all events were determined.For the 1998 November 14 SEP event the high flux of heavy particles detectedmade it possible to reconstruct the carbon, nitrogen and oxygen flux.

In the last section, preliminary results on fourteen Solar Energetic Particleevents detected by NINA-2 instrument between October 2000 and August 2001are presented.

7.1 Historical background

In 1970 [Hsieh and Simpson 1970] discovered Solar Energetic Particle (SEP)events with a greatly enhanced abundance of the rare isotope 3He, with theIMP-4 satellite. Following this, many other experiments started to study solar3He-rich events. In such events the ratio between 3He and 4He is stronglyenhanced with respect to solar abundances (∼ 4 × 10−4) measured in theSolar Wind [Coplan et al. 1985, Bodmer et al. 1995]. It was subsequentlyfound that in these events abundances of heavy elements were also unusual,with a Fe/O ratio about 10 times the value in the corona [see the review by

171

172 CHAPTER 7. LIGHT ISOTOPE ABUNDANCES IN SEP

[Reames 1999]].

The first hypothesis suggested that nuclear reactions of acceleratedparticles in the ambient of the solar atmosphere are the most probablesource of 3He. The presence of such reactions was independentlyconfirmed by the detection of the 2.2 MeV neutron-proton recombinationγ-line in solar flares [Chupp et al. 1973]. Nowadays, γ-ray spectroscopy[Mandzhavidze et al. 1999] shows that there is presence of 3He at the flaresites.

It is known that an enormous abundance of 3He in SEPs does notcorrespond to an overabundance of 2H and 3H. While solar 3He was detectedby many observers, solar deuterium [Anglin 1975] and tritium have proven tobe very rare and difficult to detect in SEP events [Mewaldt and Stone 1983,Van Hollebeke 1985, McGuire et al. 1986]. Measured abundances of 2H and3H in Solar Energetic Particles are consistent with the thin target model offlares [Ramaty and Kozlovsky 1974]. However, it was shown [Anglin 1975, andref. therein] that this model could not explain the 1000-fold enhancement inthe 3He/4He ratio and also the enrichment of heavy elements. It became clearthat selective mechanisms of 3He acceleration were required [Fisk 1978].

Solar Energetic Particles are now believed to come from two differentsources. The SEPs from solar flares usually have an enhancement in the3He/4He ratio and an enhanced number of heavy ions with respect to solarabundances because of resonant wave-particle interactions at the flare site,where the ions are highly stripped of orbital electrons by the hot environment.However, the most intense SEP events, with particles of the highest energies,are produced by accelerations at shock waves driven by Coronal Mass Ejections(CMEs). On average, these particles directly reflect the abundances andtemperature of ambient, unheated coronal material. Various aspects of gradualand impulsive SEP events have been compared in a variety of review articles[Reames 1999, and references therein].

Flares and CMEs can occur separately. Most flares are not accompaniedby a CME, whereas many fast CMEs that produce gradual SEP events haveassociated flares. The most interesting events are therefore the ”pure” ones,where the two mechanisms are not acting together and it is easier therefore todistinguish and characterize the acceleration processes involved.

The 3He/4He ratio can be used to characterize the two types of event.3He/4He ∼ 1 is not uncommon in impulsive events, while 3He/4He < 0.01usually indicates gradual SEP events. However, this ratio is difficult tomeasure and not available for a large sample of events [Chen et al. 1995,Mason et al. 1999]. Precise measurements of the 3He/4He ratio can providenew constraints on existing theories that discuss 3He acceleration mechanismsand the propagation processes.

In this Chapter I present measurements of light isotope abundances in SolarEnergetic Particle events detected in the period October 1998 – April 1999 bythe instrument NINA. In the last section preliminary results on Solar EnergeticParticle events detected by NINA-2 instrument in the period October 2000 –


August 2001 are presented.

7.2 Data analysis

The study of SEP rare isotope abundances requires a careful analysis of thedata collected. The background includes either noise and unidentifiable events,which can be eliminated by an off-line dedicated selection algorithm, andparticles produced by secondary interactions in the material surrounding theNINA silicon tower. In order to detect only solar particles, the solar quietbackground must be excluded also.

Event selection and mass reconstruction are the same as those described inChapter 2 and 3, but analysis needs new additional background estimations.

7.2.1 Background estimations

In order to study rare isotope abundances (2H, 3H, 3He) in the analysis of SEPsit is necessary to subtract the solar quiet background. Secondary productionsinside the instrument induced by high energy solar particles must also beaccounted for.

• Solar quiet background

This includes galactic particles (only 2H and 3He), and secondary 2H, 3Hand 3He which are produced by nuclear interactions of primary cosmic raysinside the 300 µm of aluminum which cover the first silicon plane of NINA.

We measured this background component during passes over the polarcaps in solar quiet periods; the average counting rate was ∼ (7.5± 0.9)× 10−5

events s−1 for deuterium, ∼ (3.7 ± 0.6) × 10−5 events s−1 for tritium, and∼ (1.5 ± 0.1) × 10−4 events s−1 for 3He, each of them relative to the energyinterval reported in Table 2.1.

• Secondary production by SEPs

The amount of secondary productions inside the instrument, induced byinteractions of solar particles, cannot be directly measured but can be inferredby estimations.

The most important reactions which can produce secondary 2H, 3H and3He are the interactions of protons and α-particles with the aluminum cover:

1. 27Al (p,X) 2H, 3H, 3He, ...

2. 27Al (α,X) 2H, 3H, 3He, ...

The ratio R between secondary 3He and primary 4He nuclei, consideringreaction 1. and 2., is given by:


R =(

σp Fp

Fα

+ σα

)NA

Aρ ∆x , (7.1)

where σp and σα are the cross sections for the reaction 1. and 2.respectively, and Fp and Fα the integral fluxes of incident protons and αparticles with sufficient energy to produce a secondary 3He detectable by NINA(Ep > 40 MeV, Eα > 10 MeV n−1). NA is the Avogadro number, ∆x thethickness of the aluminum cover, ρ the density of aluminum and A its atomicweight.

To give an estimation of the 3He secondary production in NINA, we adopta value of σp = 30 mb [Segel et al. 1982, Michel et al. 1995] and σα = 100 mb,as taken by the authors in [Chen et al. 1995]. Eq. 7.1 then becomes:

R = 2 × 10−4(0.3

Fp

Fα

+ 1)

. (7.2)

Fα is directly measured by our instrument. In order to evaluate Fp weneed to propagate the proton flux, that we measure in the energy interval12–14 MeV, to higher energies utilizing the proton spectral index from IMP-8[IMP 1999].

Table 7.1 reports the values of R estimated for the nine SEP eventsconsidered. The background coming from estimations by eq. 7.2 is less than10% of the solar quiet background, except for one case (1998 November 14)where an additional analysis was needed.

Similar calculations of the ratio R between secondary 2H and primary 1Hnuclei, and between secondary 3H and primary 1H nuclei, give a result of R ∼10−5.

The dead time of the instrument was calculated as discussed in Chapter 3.

7.2.2 Flux measurements

To be free from the effects of the rigidity cutoff due to the Earth’s magneticfield, only events recorded in polar regions at L−shell > 6 have been consideredfor SEP analysis. Figure 7.1 shows the counting rate of 4He and 16O, asa function of energy, detected at two different geomagnetic latitudes. Theacquisition rate does not vary between the cut L − shell > 6 and > 10.

Particle fluxes were reconstructed according to the methods described inChapter 3.

7.3 SEP measurements

SEP events were identified by an unpredictable increase in the trigger rate ofthe instrument, at least one order of magnitude with respect to the averaged

7.3. SEP MEASUREMENTS 175

Figure 7.1: Counting rate, plotted as a function of energy, for 4He (left) and 16O(right) measured at different geomagnetic latitudes.


Figure 7.2: Proton counting rate for the period November 1998 – December 1998,for GOES-8 (top, E > 10 MeV) and NINA (bottom). Arrows mark the candidateX-ray associated event, according to Table 7.2.


SEP date γ § 1H/4He §§ R(3He/4He)‡ 3He/4He † 2H/1H †† 3H/1H×10−2 ×10−5 ×10−4

6 Nov. 1998 4.7 ± 0.4 171 ± 16 3 × 10−3 6.5 ± 4.3 < 5.0 < 17 Nov. 1998 3.7 ± 0.3 24 ± 2 6 × 10−4 33 ± 0.6 < 3.4 < 28 Nov. 1998 4.5 ± 0.5 18 ± 2 5 × 10−4 23 ± 10 < 5.3 < 1414 Nov. 1998 1.78 ± 0.05 21.9 ± 0.5 5 × 10−4 1.1 ± 0.3 ‡‡ < 1.7 < 5022 Nov. 1998 1.9 ± 0.4 22 ± 4 5 × 10−4 < 2.8 < 5.1 < 924 Nov. 1998 3.5 ± 0.2 17 ± 1 5 × 10−4 4.1 ± 3.2 < 35 < 720 Jan. 1999 2.8 ± 0.2 182 ± 14 3 × 10−3 0.3 ± 0.6 < 3.5 < 622 Jan. 1999 4.2 ± 0.1 166 ± 8 3 × 10−3 -0.1 ± 0.6 < 0.7 < 316 Feb. 1999 3.4 ± 0.7 12.2 ± 2.5 4 × 10−4 -0.1 ± 8.0 < 37 < 90

Table 7.1: Summary of physics features of SEP events.§Energy: 10–50 MeV n−1

§§Energy: 12–14 MeV n−1

‡R is the estimated background ratio according to eq. 7.2†3He => 3HeSEP - 3HeBG; Energy: 15–45 MeV n−1

††2H => 2HSEP - 2HBG; Energy: 9–12 MeV n−1

‡‡Energy > 25 MeV n−1.

solar quiet values. Nine such increases in the period October 1998 – April1999 have been chosen for analysis. To illustrate them, Figure 7.2 showsthe counting rate of protons with energy of ∼ 10 MeV as a function of timemeasured by NINA in 1998 period, together with the GOES-8 proton intensity(energy > 10 MeV) [GOES-8 2002]. NINA had some pauses in data recordingdue to the filling of the onboard memory or to data transmission problems.

A summary of all nine events observed by NINA is presented in Table 7.2,together with characteristics of the solar events that can be associated to theSEPs [GOPHER 1999]. The event of 1998 November 14 was the most powerful,where the counting rate increased of almost 3 orders of magnitude with respectto solar quiet periods, reaching a value of 70 Hz. Figure 7.2 indicates that theevent lasted several days and was detected by our instrument in two separateemissions (Table 7.2). For the other events we registered increases of oneor two orders of magnitude on average. The events of 1998 November 6-7-8and those of 1999 January 20-22 occurred in a very close period of time, andthere might be effects of superposition between events [Casolino et al. 1999b].However, as it will be shown later, their spectral characteristics and isotopiccomposition are very different. For this reason we believe that these SEPsmay have a different origin and we studied their characteristics separately. Asevident in Figure 7.2, all chosen events are in good correlation with GOES-8observations.


Figure 7.3: Differential energy spectrum for 4He for the 9 different SEP eventsdetected by NINA. The dashed line (see eq. 7.3) represents the background B(E) ofgalactic 4He; a power-law spectrum S(E) (solid line, see eq. 7.4) has been super-imposed to the data.


SEP date Observation time NOAA Class Location Time of X-ray(X-ray/Hα) event (UT)

6 Nov. 1998 310.38 ÷ 311.52 8375 C1.1 / SF N19W25 04:38C1.4 / SF 04:56

7 Nov. 1998 311.52 ÷ 311.89 8375 M2.4 /- - 11:068 Nov. 1998 312.34 ÷ 312.87 8379 C2.4 / SF S20W67 20:20 (7 Nov.)14 Nov. 1998 318.36 ÷ 318.54 8385 C1.7 / BSL N28W90 05:18

320.60 ÷ 322.2522 Nov. 1998 326.32 ÷ 327.33 8384 X3.4 / 1N S27W82 06:4224 Nov. 1998 329.30 ÷ 332.14 8384 X1.0 / - - 02:2020 Jan. 1999 20.96 ÷ 22.93 - M5.2 / - - 20:0422 Jan. 1999 22.93 ÷ 24.90 8440 M1.4 / SF N19W44 17:2416 Feb. 1999 47.23 ÷ 49.04 8458 M3.2 / SF S23W14 03:12

Table 7.2: SEP events and characteristics of associated solar events. Note: Secondcolumn: NINA observation time (day of the year) for the SEP event. Third column:NOAA region number of the associated flare. Fourth column: importance of theflare in terms of X-ray/Hα classification. Fifth column: starting time (hh:mm) ofthe X-ray event.

4He observations range from 10 to 50 MeV n−1. Figure 7.3 presents the4He fluxes that were reconstructed during every SEP event. The dashed lineon the spectra is a fit of the solar quiet galactic background of 4He, measuredby our own instrument [Bidoli et al. 2001]. Data of the galactic background,in our energy range, are well reproduced by the function:

B(E[MeV n−1]) = 9.6 × 10−6 + 4.1 × 10−9exp(E/6.7)[cm2 sr s MeV n−1]−1 .(7.3)

The energy spectrum during the SEP events was fitted by a power-lawcomponent plus the background (solid line):

S(E [MeV n−1]) = A E−γ + B(E) [cm2 sr s MeV n−1]−1 . (7.4)

The value of γ (spectral index) for each event is reported in Table 7.1. Itvaries considerably from event to event, ranging from 1.8 in the 1998 November14 event to 6.8 in the 1998 November 8 event. It is interesting to notice thatthe 1998 November 6 and 1998 November 7 events occur in the same NOAAregion (see Table 7.2) but present different values of the spectral index. Thesame holds for the 1998 November 22 and 1998 November 24 SEP events, incontradiction to observations by [Chen et al. 1995] with the CRRES satellite,where events in the same NOAA region tended to have similar spectral indexes.

Table 7.1 summarizes measurements of the ratio (3HeSEP - 3HeBG)/4He inthe range 15–45 MeV n−1 for the nine SEP events, where 3HeSEP represents


the total flux of 3He which was measured during the SEP events, and 3HeBG

the estimation of its overall background. During the 1998 November 7 and14 SEP events this ratio is 3 standard deviations more than the solar coronalvalue.

For the 1998 November 7 SEP event we reconstructed the masses of 3Heand 4He (Figure 7.4), and it was possible to plot the 3He differential energyspectrum over a wide energy interval (Figure 7.5). The 3He spectrum, alsofitted by a power-law, is slightly harder (γ = 2.5 ± 0.6) than that of 4He (γ =3.7 ± 0.3). This implies that the 3He/4He ratio increases with energy in thisevent. This tendency is also confirmed by a comparison between our data andmeasurements taken on board ACE [ULEIS instrument, [Mason et al. 1999]]over the energy range 0.5–2.0 MeV n−1 and averaged over the period 1998November 6–8. Extrapolating the 3He energy spectrum measured by NINA tolower energies, the inferred 3He/4He ratio for this SEP event would be about10−4, in agreement with their value and much lower that what we obtain athigher energies.

For the 1998 November 7 SEP event we reconstructed the time profiles ofdifferent nuclei. Figure 7.6 shows the time profile of the counting rate of 1H,3He and 4He detected over the polar caps during this event. Every point on theplots corresponds to a passage over the pole, which lasted about 10 minutes.The profiles have been fitted by the following function [Burlaga 1967]:

Counting rate [t(days)] = A (t − t0)−2.5 e−2.5 tmax/(t−t0) Hz, (7.5)

where t is expressed in days (UT), t0 corresponds to the beginning of the1998 November 7 event, and A and tmax are the two parameters of the fit.Direct propagation from Sun to Earth takes, for 10 MeV n−1 particles, roughly1 hour. Taking into account this value, and assuming that the main part of theparticle emission occurs before tmax, we estimate the time interval of particleemission at the Sun for the three nuclear species to be not more than 3 hours.This similarity suggests the same acceleration and transport mechanisms forthe nuclear species, despite their different charge-to-mass ratio.

The strongest solar event that we detected, as already mentioned, wasthe one of 1998 November 14. Due to the very high flux intensities, whichincreased the noise of the detector, the 3He spectrum was reconstructed onlyby nuclei which crossed at least 7 silicon layers in the instrument, so havingenergies greater than 25 MeV n−1. Figure 7.7 shows the helium isotope massreconstructions for tracks with at least 7 views hit. Figure 7.8 presentsthe energy spectrum of 3He and 4He together. On the same figure 4Hemeasurements from SIS on board ACE are also reported [ACE 1998]. Forthe 1998 November 14 event there are other measurements of the 3He/4Heratio, performed by the IMP-8 [Dietrich and Lopate 1999] and SIS instrument


Figure 7.4: 1998 November 7 SEP event: mass reconstructions for 3He and 4He.


Figure 7.5: 1998 November 7 SEP event: differential energy spectrum for 3He and4He.


Figure 7.6: 1998 November 7 SEP event: time profile of the counting rate of 1H(left), 3He (centre) and 4He (right) during the SEP event. tmax (days) is the fitparameter appearing in eq. 7.5, which corresponds to the maximum of the timeprofile.


Figure 7.7: 1998 November 14 SEP event: mass reconstructions for 3He and 4He(Energy > 25 MeV n−1).


Figure 7.8: 1998 November 14 SEP event: differential energy spectrum for 3He(Energy > 25 MeV n−1) and 4He, measured in the day 317 (see Table 7.1). Thewhite squares are data from SIS (ACE satellite), taken in the same period.


Figure 7.9: 1998 November 14 SEP event: differential energy spectra for 12C, 14Nand 16O, measured in the day 317 (see Table 7.1). The white squares are data fromSIS (ACE satellite), taken in the same period.

[Cohen et al. 1999]. Their measured values are r = 0.02 ± 0.01 in the energyinterval 30–95 MeV n−1, and r = 0.005 in the range 8–14 MeV n−1.

During the 1998 November 14 SEP event there was also a strong presence ofheavy elements. Figure 7.9 presents the differential energy spectra for carbon,nitrogen and oxygen compared to SIS data collected in the same period (day317).

The High Threshold mode allowed the observation of hydrogen with itsisotopes over the narrow energy windows shown in Table 2.1. Table 7.1 presentsthe ratios between deuterium and proton after background subtraction, in therange 9–12 MeV n−1. Since the two isotope measurements span two differentenergy regions (see Table 2.1), we measured the proton flux in the range 11–16 MeV and utilized the proton spectral index from IMP-8 [IMP 1999] toextrapolate the proton flux to the deuteron energy region. Due to uncertaintiesin the background estimation, these values can be considered as upper limitsof the 2H/1H ratio. The average value of the ratio 2H/1H over all events is4×10−5; this value is in agreement with a previous measurement [Anglin 1975],which reported a 2H/1H ratio equal to (5.4±2.4)×10−5 between 10.5 and 13.5MeV n−1, when averaged over a large number of SEP events, consistent with

7.4. DISCUSSION 187

solar abundance values [McGuire et al. 1986].

From our data it was possible to estimate upper limits for tritium. The lastcolumn of Table 7.1 presents the ratio of counting rates of 3H in the interval6–10 MeV n−1 and of 1H in the range 11–14 MeV.

7.4 Discussion

The isotope abundance ratio and the energy spectra of SEPs are related to theacceleration mechanisms and propagation processes involved.

Among the nine 3He/4He ratios calculated from our SEP events, only the1998 November 7 one is clearly more than 3 σ above the background level. Theenergy spectra and counting rate profiles of 3He and 4He during this event werepresented in Figure 7.5 and 7.6, and partially discussed in the previous section.By using the IMP-8 spacecraft data [IMP 1999] it is possible to also extractthe average value of the proton spectral index, γ, equal to (3.6 ± 0.1) in theenergy range 11–74 MeV. This value is practically the same as the 4He spectralindex (see Table 7.1) measured by NINA for this SEP event.

The energy spectra of Solar Energetic Particles measured at 1 AU can bemodified, with respect to the emission spectra, by propagation effects. Theabsolute value of the mean free paths of particles traveling from the Sun tothe Earth can vary by one order of magnitude between individual events, but,as shown in [Droge 2000], the shape of the rigidity dependence does not vary.This rigidity dependence is a power-law with a slope equal to 0.3 in NINA’srigidity intervals. Since the 3He rigidity, at the same energy per nucleon, isbetween that of proton and 4He, it would be difficult in the 1998 November 7SEP event to explain the hardness of the 3He spectrum compared to that of4He and of protons only with diffusion effects from the Sun. The differencesin the spectral shapes of 3He and 4He, observed at 1 AU in this event, mostprobably already existed during the SEPs emission.

Wave resonance is the most likely mechanism for 3He acceleration inthe 1998 November 7 SEP event [Roth and Temerin 1997]. This mechanismaccelerates both 3He and heavy ions. The abundance of heavy and 3He ionsis determined by the temperature and density of the flare plasma, and bythe wave properties. It is interesting to note that data reported by ACE[Klecker et al. 1999] identify this SEP event as gradual by the low Fe/Oratio. In this work the ionic charge of several heavy ions, including iron,was determined. These measurements at low energy (0.2–0.7 MeV/n) areconsistent with an equilibrium plasma temperature of ∼ 1.3–1.6 ×106 K andwith typical solar wind values, suggesting acceleration from a solar wind source.With a plasma temperature of ∼ 2 MK [Roth and Temerin 1997] predict theexistence of a large population of oxygen at the same energy per nucleon as3He, which was not observed by our instrument. Another possible reasonfor the discrepancy between this event characteristic measured by the two


instruments is the narrow cone of emission for particles from 3He-rich events[Reames et al. 1991]. It would be interesting to confront our measured spectralbehaviour of 3He and 4He with the predictions of this model, but completespectral calculations are not yet available.

Some observations reported, already since 1970 [Hsieh and Simpson 1970,Webber et al. 1975], that a small 3He enrichment is also present in large events.Due to instrumental limitations of most of the experiments in space, however,systematic 3He/4He measurements in large events with small 3He enrichmentwere not available. More recently, [Chen et al. 1995] analyzed 16 SEP eventsand found that even extremely large SEP events had a value of 3He/4He greaterthan 0.5% – one order of magnitude greater than solar wind values. Thisevidence was reported also by [Mason et al. 1999], who observed 12 large SEPevents with an average value of 3He/4He about (1.9 ± 0.2) × 10−3.

To explain this enrichment, [Mason et al. 1999] suggest that residual 3Hefrom impulsive SEP events forms a source material for the 3He seen insome large SEP events. In such cases, the 3He is accelerated out of thisinterplanetary suprathermal population by the same process that energizesother particles at these energies, namely the interplanetary shock.

Our 3He/4He ratio measurements, reported in Table 7.1, seem to confirmthe fact that in all SEP events there is a quantity of 3He greater than thattypical of solar coronal values (3He/4He about 4 × 10−4). If we average the3He/4He ratio over all events detected by NINA, except the 1998 November7–8 which is clearly 3He-enriched, we obtain the value (1.1 ± 0.4) × 10−2.

Nuclear interactions in acceleration regions can also contribute to the highvalue of the ratio 3He/4He, with respect to coronal values. As stated inthe introduction, at flare physical sites the possible presence of secondaryparticles from nuclear interactions [Ramaty and Murphy 1987] should not beexcluded. Such interactions also produce 3H and 2H, which are the signatureof the process. From γ-ray spectroscopy there is evidence for the presenceof 3He and 2H at the flare sites [Mandzhavidze et al. 1999, Chupp 1983,Terekhov et al. 1993]. Usually all secondary particles are trapped in the flareloops but some of them can escape into the interplanetary space.

Particles escaping from Sun during acceleration traverse a small amountof material and do not undergo many interactions. It is therefore difficultto separate solar deuterium particles from the galactic and instrumentalbackground. In our measurements the values of the deuterium flux areclose to the instrumental limit. For most of the SEP events there was notan appreciable quantity of 2H, and in some cases we could determine onlyupper limits on the ratio 2H/1H (see Table 7.1). In 1998 November 24 SEPevent, however, there are possible indications of an excess of the deuteriumcounting rate [Bakaldin et al. 2001]. In fact, the mass reconstruction showsevidently a relevant presence (about 10 times higher than coronal abundances)of deuterium, if compared to other SEP events. Moreover, the 2H emission time


profile confirms a peak corresponding to the 24 November 1998 event. To thispeak does not correspond a peak in 1H, and NINA 3He/4He measurementsare consistent with other instruments. These observations should excludesecondary origin of detected deuterium. This possibility is now under furtherinvestigation [Sparvoli et al. 2001, Bakaldin et al. 2002b].

The average value of the ratio 2H/1H, over all our events, is < 4 × 10−5,which corresponds to about 0.1 g cm−2 thickness of traversed material, forprotons with energy about 30 MeV [Ramaty and Kozlovsky 1974, Figure 10therein]. This thickness gives a 3He/4He ratio of about 10−3, that is lowerthan our average value. However, it should be noted that the 2H/3He ratiomeasured by probes in the interplanetary space may be depressed by a factor10 due to differences in 2H and 3He angular distribution and propagationprocesses [Colgate et al. 1977]. A theoretical analysis of this ratio depends onthe solar flare model used, which is still very controversial, and beyond thescope of this work.

In conclusion, the measured level of the deuterium presence in SEPs,coming from secondary interactions in the solar ambient, cannot exclude thatpart of the 3He contents in large Solar Energetic Particle events may also havethis origin.

7.5 NINA-2 preliminary results

NINA-2 mission covered the period of maximum in the 23th cycle of Solaractivity. In the period of observation October 2000 – August 2001, 14 SEPevents were detected. The method of SEP events identification was the sameas for NINA, but due to the extensive telemetry capabilities of MITA satelliteit was possible to measure SEP proton spectra. Unfortunately Sun pointingorientation made difficulties with intensity reconstruction. In fact it wasmeasured that particle flux in North and South poles were different whensatellite pointed the Sun. It was ascribed to the Earth which shadowed theinstrument during part of the pole passages. Finally 14 count rate increaseshave been detected. Table 7.3 shows the date of the 14 SEP events and thepreliminary results without background subtraction on 3He/4He and 2H/1Hratios.

As reported in Table 7.3, no 3He reach SEP event was detected, while onlythe SEP of 2001 July 19 could be 2H enriched. At the moment, it remains one ofthe most interesting open question: is there deuterium in SEP? Above we havedescribed that in 1998 November 24 SEP event there are possible indicationsof an excess of the deuterium counting rate and that this possibility is nowunder further investigation.

In 2001 July 19 SEP also, mass reconstruction shows a presence (about 100times higher than coronal abundances) of deuterium. It was collected only 15particles ascribing to deuterium, but for the identical arguments described inprevious section, we should exclude secondary origin.


Figure 7.10: Distribution of the energy released in the first plane (E1) and the totalenergy (Etot) measured for particles detected in the 9 November 2000 SEP.


Figure 7.11: Carbon flux as measured by NINA-2 and SIS for 9 November 2000SEP.


Figure 7.12: Oxygen flux as measured by NINA-2 and SIS for 9 November 2000SEP.

7.6. SUMMARY 193

SEP events date 3He/4He 2H/1H17 Oct. 2000 3×10−3 3×10−4

9 Nov. 2000 2×10−3 <1×10−2

24 Nov. 2000 1.2×10−3 1×10−4

29 Jan. 2001 1×10−2 3×10−5

3 Apr. 2001 1×10−3 <7×10−3

11 Apr. 2001 1×10−3 <1×10−3

19 Apr 2001 2×10−3 2×10−4

8 May 2001 7×10−3 6×10−5

20 May 2001 2×10−2 1.3×10−4

4 June 2001 1.8×10−2 4.3×10−5

12 June 2001 1.7×10−2 1.4×10−4

15 June 2001 1.4×10−2 8×10−6

19 July 2001 1.1×10−2 (1±0.3)×10−2

10 Aug. 2001 1×10−3 5×10−5

Table 7.3: SEP events detected by NINA-2 instrument on board MITA satellite. Thefirst and second column report the preliminary results without background subtractionon 3He/4He and 2H/1H ratios respectively.

The most intense SEP event detected by NINA-2 was the 9 November 2000SEP. A strong presence of heavy elements was observed. Distribution of theenergy released in the first plane (E1) and the total energy (Etot) detected isshowed in Figure 7.10.

Figure 7.11 and 7.12 shows differential carbon and oxygen fluxes in functionof energy as measured by NINA-2 and SIS on board ACE. 4He acquisitionincreases of more than 4 order of magnitude in 10-20 MeV/n energy range,but no 3He enrichment was observed.

A work on this topic is in preparation.

7.6 Summary

We have determined the 3He/4He ratio and helium energy spectra over theenergy range 10–50 MeV n−1 for 9 SEP events measured in the period October1998 – April 1999 by the instrument NINA and the preliminary results for 14SEP events measured in the period June 2000 – August 2001. The mostinteresting of these events was recorded on 1998 November 7, where the ratioreached a value of about 30% but also presented features typical of a ”pure”gradual event [Klecker et al. 1999]. The similarity of the time profiles for the1H, 3He and 4He emissions in the event implies that these isotopes underwent


the same acceleration mechanism.The other SEP events yield an average value of the 3He/4He ratio slightly

higher than that typical of the solar wind. The average value of the 2H/1Hupper limits, over all our events, is equal to 2H/1H = 4 × 10−5. This levelof presence of deuterium in SEP events, coming from secondary interactionsin the solar ambient, cannot exclude that part of the 3He contents in SolarEnergetic Particles may also have this origin.

Conclusions

NINA and NINA-2 flew in space aboard Resurs-O1-N4 and MITA satellitesfrom July 1998 to April 1999 and from July 2000 to August 2001 in a 830km and 400 km average altitude orbit respectively. The characteristics of theinstruments and the orbits of satellites gave the possibility to study differentcomponents of cosmic rays.

Energy spectra of secondary sub-cutoff protons were measured at differentgeomagnetic latitudes (Chapter 4). The good instruments capabilities allowedmeasurements of the spectra of protons with high accuracy, also taking intoaccount the contribution of deuterium and tritium amounting to about 10%.The proton energy spectra were found to be practically flat in this energy rangeand not to vary appreciably with altitude. In addition, the measured flux wasapproximately constant between 1998 and 2001.

NINA and NINA-2 measured the 2H/1H ( 15%), 3H/2H ( 25%), 3H/1H( 4%), 3H/3He ( 3) and 3He/4He ( 1.5) isotope abundance ratios,and reconstructed the differential energy spectra of all hydrogen and heliumisotopes of albedo origin across the energy interval 10–40 MeV/n (Chapter 5).

Results presented in Chapter 6 show that rare light isotopes representa distinct component of the Earth’s inner radiation belt. Geomagneticallytrapped fluxes of 2H, 3H, 3He and 4He were measured for L − shell∼1.2.The 2H/1H ratio at energy 10 MeV/n was ∼ 0.01, the 3H/2H ∼ 0.2 and the3He/3H ratio ∼ 2. The 3He/4He ratio increases from ∼ 1 to ∼ 5 across theenergy range 10–40 MeV/n. The data analysis and the comparison with theavailable theoretical calculations bring to the conclusion that the light isotopescomponent is originated from the interactions of high energy trapped protonswith the residual atmospheric helium.

The observed 2H/1H and 3H/2H ratios at energy 10 MeV/n are higherthan calculated from atmospheric production models. The results of thecalculations, however, depend on the detailed knowledge of some physicalparameters, such as the radiation belt models, the atmospheric density modeland cross section parametrisation, and also on the solar cycle phase. Themeasurements performed by NINA can provide new inputs to the theoreticalcalculations.

195

We have determined (Chapter 7) the 3He/4He ratio and helium energyspectra for 9 SEP events measured in the period October 1998 – April 1999 bythe instrument NINA and the preliminary results for 14 SEP events measuredin the period June 2000 – August 2001. The most interesting of these eventswas recorded on 1998 November 7, where the ratio reached a value of about 30%but also presented features typical of a ”pure” gradual event. The similarityof the time profiles for the 1H, 3He and 4He emissions in the event implies thatthese isotopes underwent the same acceleration mechanism.

The other SEP events yield an average value of the 3He/4He ratio slightlyhigher than that typical of the solar wind. The average value of the 2H/1Hupper limits, over all our events, is equal to 2H/1H = 4 × 10−5. This levelof presence of deuterium in SEP events, coming from secondary interactionsin the solar ambient, cannot exclude that part of the 3He contents in SolarEnergetic Particles may also have this origin.

196

Bibliography

[ACE 1998] http://www.srl.caltech.edu/ACE/ASC/level2.

[Adriani et al. 2002] Adriani, O., et al., The PAMELA Experiment onSatellite and its Capability in Cosmic Rays Measurements, Nucl. Instr.Methods, A 478, 114, 2002.

[Akasofu 1969] Akasofu, S. I., Polar and Magnetospheric Substorms, Springer-Verlag, New York, 1969.

[Alcaraz et al. 2000] Alcaraz, J., et al., Protons in near Earth orbit, PhysicsLetter B, 472, 215, 2000.

[Alcaraz et al. 2000b] Alcaraz, J., et al., Helium in near Earth orbit, PhysicsLetter B, 494, 193, 2000.

[Alvarez and Compton 1933] Alvarez, L. and A. H. Compton, A positivelycharged component of cosmic rays, Phys. Rev., 43, 835, 1933.

[Anglin 1975] Anglin, J. D., The relative abundances and energy spectra ofsolar-flare-accelerated deuterium, tritium, and helium-3, Astrophys. J.,198, 733, 1975.

[Avdeev et al. 2001] Avdeev, S. et al., The SilEye nuclei cosmic ray and eyelight flash experiment onboard the Mir space station, Proceeding 27th Int.Cosmic-Ray Conf. (Hamburg), 5, 1745, 2001.

[Aversa et al. 1995] Aversa, F. et al., Wizard Si-W imaging calorimeter: apreliminary study on its particle indentification capability during a balloonflight in 1993, Nucl. Instr. Methods, A 360, 17, 1995.

[Bakaldin et al. 1997] Bakaldin, A., et al., Experiment NINA: investigation oflow energy nuclear fluxes in the near-Earth space, Astroparticle Physics 8,109, 1997.

[Bakaldin et al. 2001] Bakaldin, A., Iannucci, A., et al., Light IsotopeAbundances in Solar Energetic Particles measured by the Space InstrumentNINA, Proceeding 27th Int. Cosmic-Ray Conf. (Hamburg), 8, 3104, 2001.

197

198 BIBLIOGRAPHY

[Bakaldin et al. 2001b] Bakaldin, A., Iannucci, A., et al., Observation ofgeomagnetically trapped light isotopes by NINA, Proceeding 27th Int.Cosmic-Ray Conf. (Hamburg), 8, 3975, 2001.

[Bakaldin et al. 2002] Bakaldin, A., Iannucci, A., et al., Geomagneticallytrapped light isotopes observed with the detector NINA, Journal ofGeophysical Research, Space Physics, 2002, in press.

[Bakaldin et al. 2002b] Bakaldin, A., Iannucci, A., et al., Light IsotopeAbundances in Solar Energetic Particles Measured by the SpaceInstrument NINA, Astrophys. J., 577, 1, 513, 2002.

[Barbiellini et al. 1996] Barbiellini, G. et al., Performance of the CAPRICERICH detector during the 1994 ballon flight, Nucl. Instr. Methods, A 371,169, 1996.

[Barbiellini et al. 1996b] Barbiellini, G. et al., The cosmic ray positron toelectron ratio in the energy range 0.85 to 14.0 GeV, Astron. and Astrophys.309, L15, 1996.

[Baumjohann and Treumann 1997] Baumjohann, W. and R. A. Treumann,Basic Space Plasma Physics, Imperial College Press, 1997.

[BBSO 2002] http://www.bbso.njit.edu.

[Bellotti et al. 1997] Bellotti, R. et al., Study of the combined particleidentification capability of a transition radiation detector and a siliconimaging colorimeter during the TS93 balloon flight, Astroparticle Physics,7, 219, 1997.

[Berezinskii et al. 1990] Berezinskii, V. S., et al., Astrophysics of CosmicRays, North Holland, 1990.

[Bertrand and Peelle 1973] Bertrand, F. E. and R. W. Peelle, CompleteHydrogen and Helium Particle Spectra from 30 to 60 MeV ProtonBombardment of Nuclei with A=12 to 209 and Comparison with theIntranuclear Cascade Model, Phys. Rev. C, 8, 1045, 1973.

[Bidoli et al. 1999] Bidoli, V., et al., The space telescope NINA: results of abeam test calibration, Nucl. Instr. Methods Phys. Res., A 424, 414, 1999.

[Bidoli et al. 2001] Bidoli, V., Iannucci, A., In-orbit performances of thespace telescope NINA and GCR fluxes measurements, Astrophys. J.,Supplement Series, 132, 365, 2001.

[Bidoli et al. 2002] Bidoli, V., Iannucci, A. et al., Energy spectrum ofsecondary protons above the atmosphere measured by the instrumentsNINA and NINA-2, accepted by The Annales Geophysicae letter,AGLETT/2002/000003, in press.

BIBLIOGRAPHY 199

[Bidoli et al. 2002] Bidoli, V., Iannucci, A. et al., Isotope composition ofsecondary hydrogen and helium above the atmosphere measured by theinstruments NINA and NINA-2, accepted to the Journal of GeohysicalReview, Space Physics, November 2002.

[Blanford and Ostriker 1978] Blanford, R. D. and J. R. Ostriker, Particleacceleration by astrophysical shocks, Astrophys. J. 221, L29, 1978.

[Bocciolini et al. 1993] Bocciolini, M. et al., Silicon calorimeter for cosmicantimatter search, Nucl. Phys., B 32, 77, 1993.

[Bocciolini et al. 1996] Bocciolini, M. et al., Nucl. Instr. Methods, A 370, 403,1996.

[Bodmer et al. 1995] Bodmer, R., et al., Solar Wind Helium IsotopicComposition from SWICS/Ulysses, Space Science Revs., 72, 61, 1995.

[Boezio et al. 1997] Boezio, M. et al., The cosmic ray anti-proton flux between0.62 GeV and 3.19 GeV measured near solar minimum activity, Astrophys.J., 487, 415, 1997.

[Boezio et al. 2001] Boezio, M. et al., High-Energy Cosmic-Ray Antiprotonswith the CAPRICE98 experiment, Proceeding 27th Int. Cosmic-Ray Conf.(Hamburg), 5, 1695, 2001.

[Boezio et al. 2001b] Boezio, M. et al., The Cosmic-Ray antiproton fluxbetween 3 and 49 GeV, Astrophys. J, 561, 787, 2001.

[Bonvicini et al. 2001] Bonvicini, V., et al., The pamela experiments in space,Nucl. Instr. Methods, A 461, 262, 2001.

[Brun et al. 1994] Brun, R. et al., Detector Description and Simulation Tool,CERN program library, 1994.

[Burlaga 1967] Burlaga, L. F.,Journal of Geophysical Research, 72, 17, 4449,1967.

[Casolino et al. 1999] Casolino, M., Iannucci, A., et al., Launch in orbit ofNINA detector for cosmic ray nuclei study, Proceeding 26th Int. Cosmic-Ray Conf. (Salt Lake City), OG 4.2.07, 1999.

[Casolino et al. 1999b] Casolino, M., Iannucci, A., et al., Study of November1998 Solar Energetic Particles events with NINA instrument, Proceeding26th Int. Cosmic-Ray Conf. (Salt Lake City), SH.1.4.04, 1999.

[Casolino et al. 1999c] Casolino, M., et al. 1999, Continuation of the missionNINA: Nina-2 experiment on MITA satellite, Proceeding 26th Int. Cosmic-Ray Conf. (Salt Lake City), OG 4.2.17, 1999.

200 BIBLIOGRAPHY

[Casolino et al. 2001] Casolino, M. et al., Cosmic ray measurements on boardspace station MIR with SILEYE-2 experiment , Proceeding 27th Int.Cosmic-Ray Conf. (Hamburg), 5, 4011, 2001.

[Casolino et al. 2001b] Bidoli, V. et al., Study of cosmic rays and light flasheson board space station MIR: The SilEye Experiment, Advances in SpaceResearch, 25, 10, 2075, 2000.

[Casolino et al. 2001c] Casolino, M., Iannucci, A. et al., Launch in orbit ofthe NINA-2 apparatus aboard the satellite MITA, Proceeding 27th Int.Cosmic-Ray Conf. (Hamburg), 5, 2314, 2001.

[Chen et al. 1995] Chen, J., The He-3/He-4 ratios for solar energetic particleevents during the Combined Release and Radiation Effects SatelliteMission,Astrophys. J., 442, 875, 1995.

[Chen et al. 1996] Chen, J. et al., Energetic helium isotopes trapped in themagnetosphere, Journal of Geophysical Research, 101, A11, 24787, 1996.

[Chupp et al. 1973] Chupp, E. L., et al., Nature, 241, 333, 1973.

[Chupp 1983] Chupp, E. L., High energy particle acceleration in solar flaresObservational evidence, Sol. Phys., 85, 383, 1983.

[Cohen et al. 1999] Cohen, C., et al., New observations of heavy-ion-rich solarparticle events from ACE, Geophys. Res. Lett., 26, 17, 2697, 1999.

[Colgate et al. 1977] Colgate, S. A., et al., Possible interpretation of theisotopic composition of hydrogen and helium in solar cosmic rays,Astrophys. J., 213, 849, 1977.

[Coplan et al. 1985] Coplan, M. A., et al. 1985, Sol. Phys., 93, 415.

[Clayton 1968] Clayton, D. D., Principles of stellar evolution and nucleosyn-tesis, McGraw Hill, 1968.

[Colgate et al. 1977] Colgate, S. A. et al., Possible interpretations of theisotopic composition of hydrogen and helium in solar cosmic rays,Astrophys. J., 213, 849, 1977.

[Compton 1932] Compton, A. H., Variation of the cosmic ray with latitude,Phys. Rev., 41, 111, 1932.

[Compton 1933] Compton, A. H., A geographic study of cosmic rays, Phys.Rev., 43, 387 , 1933.

[Comstock et al. 1966] Comstock, G. M., et al., Abundances and EnergySpectra of Galactic Cosmic Ray Nuclei Above 20 MeV per Nucleon inthe Nuclear Charge Range 2<Z<26, Astrophys. J., 146, 51, 1966.

BIBLIOGRAPHY 201

[Cook et al. 1993] Cook, W. R. et al., MAST: A Mass Spectrometer Telescopefor Studies of the Isotopic Composition of Solar, Anomalous and GalacticCosmic Ray Nuclei, IEEE Trans. on Geoscience and Remote Sensing, 31-3, 557, 1993.

[Currie et al. 1956] Currie, L. A., et al., Trirum production by high energyprotons, Phys. Rev., 101, 1557, 1956.

[Derome et al. 2000] Derome, L., M. Buenerd, A. Barrau et al., The origin ofthe high energy proton component below the geomagnetic cut off detectedby the AMS experiment, Physics Letters B, 489, 1, 1, 2000.

[Derome and Buenerd 2001] Derome, L., and M. Buenerd, Origin of lightnuclei in near earth orbit, Physics Letters B, 521, 139, 2001.

[Derome and Buenerd 2001b] Derome, L., and M. Buenerd, Origin of protonsand light nuclei below the geomagnetic cutoff, Proceeding 27th Int. Cosmic-Ray Conf. (Hamburg), 5, 1728, 2001.

[Derome and Buenerd 2001c] Derome, L., and M. Buenerd, Origin of lightnuclei in near earth orbit, astro-ph/0105523, May 31, 2001.

[Dietrich and Lopate 1999] Dietrich, W., and C. Lopate, Proceeding 26th Int.Cosmic-Ray Conf. (Salt Lake City), 6, 71, 1999.

[Droge 2000] Droge, W., The Rigidity Dependence of Solar Particle ScatteringMean Free Paths, Astrophys. J., 537, 1073, 2000.

[Fan et al. 1960] Fan, C. Y., et al., Experiments on the Eleven-Year Changesof Cosmic-Ray Intensity Using a Space Probe, Phys. Rev. Lett., 5, 272,1960.

[Fermi 1949] Fermi, E., On the Origin of the Cosmic Radiation, Phys. Rev.,75, 1169, 1949.

[Fireman et al. 1955] Fireman, E. L., and F. S. Rowland, Tritium and neutronproduction by 2.2 Bev protons on nitrogen and oxigen, Phys. Rev., 97, 780,1955.

[Fisk 1978] Fisk, L. A., He-3-rich flares - A possible explanation, Astrophys.J., 224, 1048, 1978.

[Freden and White 1960] Freden, S. C., and R. S. White, Particle fluxes in theinner radiation belt, Journal of Geophysical Research, 65, 1377, 1960.

[Friedlander 1989] Friedlander, M. W., Cosmic rays, Harward UniversityPress, 1989.

[Gangnes et al. 1949] Gangnes, A. V., et al., The cosmic ray intensity abovethe atmosphere, Phys. Rev., 75, 57, 1949.

202 BIBLIOGRAPHY

[Galper et al. 1996] A. Galper, et al. - Proc. 6th European Symp. on Life Sci.Res. in Space - Trondheim, 1996.

[Golden et al. 1990] Golden, R. L. et al., Observation of cosmic ray electronsand positrons using an imaging calorimeter, Astrophys. J., 436, 769, 1994.

[Golden et al. 1996] Golden, R. L. et al., Measurement of the positron toelectron ratio in the cosmic rays above 5 GeV, Astrophys. J. 457, L103,1996.

[Ginzburg and Syrovatskii 1964] Ginzburg, V. L. and S. I. Syrovatskii, TheOrigin Cosmic rays, Pergamon Press, Oxford, 1964.

[Ginzburg 1969] Ginzburg, V. L., Elementary Process for Cosmic RayAstrophysics, New York: Gordon and Breach, Science Publishers, 1969.

[Ginzburg 1975] Ginzburg, V. L., Philos. Trans. R. Soc. London A277, 463,1975.

[Ginzburg 1980] Ginzburg, V. L, Ap&SS, 68, 295, 1980.

[GOES-8 2002] http://www.sec.noaa.gov/Data/goes.html.

[GOPHER 1999] gopher://solar.sec.noaa.gov:70/11/indices.

[Gosling 1993] Gosling, J. T., The solar flare myth, Journal of GeophysicalResearch, 98, 18949, 1993.

[Grieder 2001] Grieder, P., Cosmic rays at earth, Elsevier, 2001.

[Hargreaves 1995] J. K. Hargreaves, The solar-terrestrial environment,Cambridge University Press, 1995.

[Hasebe et al. 1993] Hasebe, N., et al., Improvement of mass resolution ofcosmic ray nuclei using a ∆ExE Si detector telescope, Nucl. Instr. MethodsPhys. Res., A 325, 335, 1993.

[Hess 1968] Hess, W. N., The radiation belt and magnetosphere, BlaisdellPublishing Company, 1968.

[Hillas 1972] Hillas, A. M., Cosmic Rays, Pergamon Press, 1972.

[Hopper 1964] Hopper, V. D., Cosmic Radiation and High Energy Interaction,Logos Press, 1964.

[Hoots et al. 1980] Hoots, F. R., et al., Models for propagation ofNORAD elements sets, Project Spacecraft Report N.3, Aerospace DefenseCommand, United States Air Forces, 1980.

BIBLIOGRAPHY 203

[Hsieh and Simpson 1970] Hsieh, K. C., and J. A. Simpson, The RelativeAbundances and Energy Spectra of 3He and 4He from Solar Flares,Astrophys. J., 1962, L191, 1970.

[Huang and Stephens 2001] Huang, M. A., and S. A. Stephens, Origin of largealbedo 3He/4He ratio, Proceeding 27th Int. Cosmic-Ray Conf. (Hamburg),4208, 2001.

[IGRF 2002] The international Geomagnetic Reference Field (IGRF),http://nssdc.gsfc.nasa.gov/space/model/models/igrf.html.

[IMP 1999] http://nssdc.gsfc.nasa.gov/space/space−physics−home.html.

[Inuvik 1999] www.bartol.udel.edu/ neutronm/listen/main.html.

[Jackson 1998] Jackson, J. D., Classical Electrodynamics, 3rd ed. Wiley, 1998.

[Johnson 1933] Johnson, T. H., The azimuthal asymmetry of the cosmicradiation, Phys. Rev. 43, 835, 1933.

[Kanbach et al. 1974] Kanbach, G. et al., Support for Crand theory frommeasurements of Earth albedo neutrons between 70 and 250 MeV, Journalof Geophysical Research, 79, 34, 5159, 1974.

[Klecker et al. 1999] Klecker, B, et al., Proceeding 26th Int. Cosmic-Ray Conf.(Salt Lake City), 6, 83, 1999.

[Krimigis and Van Allen, 1967] Krimigis, S. M., and J. A. Van Allen,Geomagnetically trapped alpha particles, Journal of Geophysical Research,72, 5779, 1967.

[Kurnosova et al. 1980] Kurnosova, L. V. et al., Distribution of chargedparticles in near-earth space, Moscow, Izdatel’stvo Nauka, A 81, 41463,19, 49, 1980.

[Lal et al. 1962] Lal, D., and B. Peters, Cosmic ray producted isotopesand their applications to problems in geophysics, Elementary Particlesand Cosmic Ray Physic, 6, 1, North-Holland Publishing Company,Amsterdam, 1962.

[Lamanna et al. 2001] Lamanna, G. et al., Measurement of deuteron spectrain Low Earth Orbit with the Alpha Magnetic Spectrometer, Proceeding27th Int. Cosmic-Ray Conf. (Hamburg), 5, 1614, 2001.

[Larson et al. 1992] Larson, W. and Wertz, J., Space mission analysis anddesign, Kluwer Academic Publisher, 1992.

[LASCO 2002] http://lasco-www.nrl.navy.mil/lasco.html.

204 BIBLIOGRAPHY

[Lee 2000] Lee, M. A., A Simple Model for Ion Acceleration and Transportat an Evolving Coronal/Interplanetary Shock, The Acceleration andTransport of Energetic Particles Observed in the Heliosphere, Indian Wells,California, S20, January 5 - 8, 2000.

[Lemaitre and Vallarta 1933] Lemaitre, G. and M. S. Vallarta, On Compton’sLatitude Effect of Cosmic Radiation, Phys. Rev., 43, 87, 1933.

[Lipari 2002] Lipari, P., The fluxes of sub-cutoff particles detected by AMS,the cosmic ray albedo and atmospheric neutrinos, Astroparticle physics,16, 3, 295, 2002.

[Llope et al. 1995] Llope, W. J. et al., The fragment coalescence model, Phys.Rev. C, 52, 2004, 1995.

[Longair 1992] M. S. Longair, High Energy Astrophysics, CambridgeUniversity Press, 1992.

[Looper et al. 1996] Looper, M. D., J. B. Blake, J. R. Cummings and R. A.Mewaldt, SAMPEX observations of energetic hydrogen isotopes in theinner zone, Radiation Measurements, 26, 967, 1996.

[Looper et al. 1998] Looper, M. D., J. B. Blake and R. A. Mewaldt, Mapsof hydrogen isotopes at low altitudes in the inner zone from SAMPEXobservations, Adv. Space Res., 21, 1679, 1998.

[Mandzhavidze et al. 1999] Mandzhavidze, N., et al., Determination of theAbundances of Subcoronal 4He and of Solar Flare-accelerated 3He and4He from Gamma-Ray Spectroscopy, Astrophys. J., 518, 918, 1999.

[Mason et al. 1999] Mason, G. M., et al., 3He Enhancements in Large SolarEnergetic Particle Events, Astrophys. J., 525, L133, 1999.

[Matsunaga et al. 1998] Matsunaga, H. et al., Measurement of Low-EnergyCosmic-Ray Antiprotons at Solar Minimum, Physical Review Letters, 81,4052, 1998.

[McDonald et al. 1981] McDonald, F. B., et al., The solar modulation ofgalactic cosmic rays in the outer heliosphere, Astrophys. J., 249, L71,1981.

[McGuire et al. 1986] McGuire, R. E., et al., The composition of solarenergetic particles, Astrophys. J., 301, 938, 1986.

[McKibben et al. 1982] McKibben, R. B., et al., The galactic cosmic-ray radialintensity gradient and large-scale modulation in the heliosphere, Astrophys.J., 254, L23, 1982.

BIBLIOGRAPHY 205

[Mendis and Axford 1974] Mendis, D. A. and W. I. Axford, Satellites andmagnetospheres of outer planets, Ann. Rev. Earth Planet. Sci. 2, 419,1974.

[Meschini et al. 1988] Meschini, M. et al., A double-dee toroidal field for aspace spectrometer, Il Nuovo Cimento, 102 B, 529, 1988.

[Mewaldt and Stone 1983] Mewaldt, R. A., and E. C. Stone, Deuterium,3HE, and Anomalous H and he Nuclei, Proceeding 18th Int. Cosmic-Ray(Bangalore), 4, 42, 1983.

[Meyer 1972] Meyer, J. P., Astron. Astrophys. Suppl., 7, 417, 1972.

[Meziane 2000] Meziane, K., Acceleration of Energetic Ions at the Earth’sPerpendicular Shock: 3D Observations, ACE-2000 Symposium, TheAcceleration and Transport of Energetic Particles Observed in theHeliosphere, Indian Wells, California, I6, January 5 - 8, 2000.

[Michel et al. 1995] Michel, R., et al., Nuclide production by proton-inducedreactions on elements (6 ¡= Z ¡= 29) in the energy range from 800 to 2600MeV, Nucl. Instr. Methods Phys. Res., B 103, 183, 1995.

[Molnar and Simon 2001] Molnar, A. and M. Simon, Test of the Diffusion Haloand the Leaky Box Model by means of secondary radioactive cosmic raynuclei with different lifetimes, Proceeding 27th Int. Cosmic-Ray (Hamburg),1860, 2001.

[Mullan and Linsky 1998] Mullan, D. J., and J. L. Linsky, Nonprimordialdeuterium in the interstellar medium, Astrophys. J., 511, 1, 502, 1998.

[NOAA 2002] http://www.sec.noaa.gov.

[Northrop 1961] Northrop, T. G., The guiding center approximation tochatged particle motion, Ann. Phys., 15, 209, 1961.

[Orbital Information Group 2001]http://oig1.gsfc.nasa.gov/scripts/foxweb.exe/app01?.

[Otto 2001] http://what.gi.alaska.edu/ao/msp/msp0.html.

[Owens 1976] Owens, A. J., The age distribution of galactic cosmic rays,Astrophysics and Space Science, 40, 357, 1976.

[Papini et al. 1996] Papini, P., C. Grimaldi, and S. A. Stephens, Il NuovoCimento C, 19, 367, 1996.

[Parker 1966] Planet Space Sci., 14, 371, 1966.

[Parks 1991] Parks, G. K., Physica of space plasma, Addison-Wesley, 1991.

206 BIBLIOGRAPHY

[Pearce et al. 2002] Pearce, M., et al., The status of the PAMELA experiment,First International Conference on Particle and Fundamental Physics inSpace, LaBiodola, Isola d’Elba, 14-19 May, 2002.

[Pennypacker et al. 1973] Pennypacker, C. R., et al., Measurement ofgeomagnetic cutoff rigidities and particle fluxes below geomagnetic cutoffnear Palestine, Texas, Journal of Geophysical Research, 78, 1515, 1973.

[Perlow 1952] Perlow, G. J., Rocket determination of the ionization spectrumof charged cosmic rays at λ = 41oN , Phys. Rev., 71, 325, 1947.

[Pfotzer 1936] Pfotzer, G., Zeits. f. Physik, 102, 23, 40, 1936.

[Pioneer 2002] http://spaceprojects.arc.nasa.gov.

[Pittori et al. 2001] Pittori, C. et al., Gamma-ray Imaging by silicon Detectorsin Space: Presentation of the AGILE Reconstruction Method and KalmanFilter Algorithms, Fifth Compton Symposium, American Institute ofPhysics 587, 739, 2001.

[Pugacheva et al., 1998] Pugacheva., G. I. et al., Hydrogen and helium isotopeinner radiation belts in the Earth’s magnetosphere, Annales Geophysicae,16, 931, 1998.

[Ray 1963] Ray, E. C., Re-entrant cosmic-ray albedo, Journal of GeophysicalResearch, 67, 3289, 1963.

[Ray 1967] Ray, E. C., An improved re-entrant cosmic-ray albedo calculation,Journal of Geophysical Research, 72, 4839, 1967.

[Ramaty and Lingenfelter 1969] Ramaty, R. and R. E. Lingenfelter, Astro-phys. J., 155, 587, 1969.

[Ramaty and Kozlovsky 1974] Ramaty, R., and B. Kozlovsky, Deuterium,tritium, and helium-3 production in solar flares, Astrophys. J., 193, 729,1974.

[Ramaty and Murphy 1987] Ramaty, R., and R. J. Murphy, Nuclear processesand accelerated particles in solar flares, Space Science Revs., 45, 213, 1987.

[Reames 1990] Reames, D. V., Energetic particles from impulsive solar flares,Astrophys. J. Suppl. Series, 73, 235, 1990.

[Reames et al. 1991] Reames, D. V., and al., Multispacecraft observations ofsolar (He-3)-rich events, Astrophys. J., 380, 287, 1991.

[Reames 1993] Reames, D. V., Adv. Space Res., 13, 9, 331, 1993.

[Reames 1995] Reames, D. V., Rev. Geophys. Suppl., 33, 585, 1995.

BIBLIOGRAPHY 207

[Reames 1998] Reames, D. V., Solar Energetic Particles: Sampling CoronalAbundances, Space Science Reviews, 85, 327, 1998.

[Reames 1999] Reames, D. V., Particle acceleration at the Sun and in theheliosphere, Space Science Revs., 90, 413, 1999.

[RGO 2002] http://science.msfc.nasa.gov/ssl/pad/solar/greenwch.htm.

[Rossi 1933] Rossi, B., On the magnetic deflection of cosmic rays, Phys. Rev.,43, 834, 1933.

[Rossi and Olbert 1970] Rossi, B. and S. Olbert, Introduction to the Physicsof Space, McGraw-Hill Book Co., 1970.

[Roth and Temerin 1997] Roth, I., and M. Temerin, Enrichment of 3He andHeavy Ions in Impulsive Solar Flares, Astrophys. J., 477, 940, 1997.

[Segel et al. 1982] Segel, R. E., et al., Inclusive proton reactions at 164 MeV,Phys. Rev. C, 26, 6, 2424, 1982.

[Selesnick and Mewaldt 1996] Selesnick, R. S., and R. A. Mewaldt, Atmo-spheric production of radiation belt light isotopes, Journal of GeophysicalResearch, Space Physics, 101, 19745, 1996.

[Share and Murphy 2000] Share, G.H. and R.J. Murphy, Measurement ofAccelerated Particles at the Sun, ACE-2000 Symposium, The Accelerationand Transport of Energetic Particles Observed in the Heliosphere, IndianWells, California, S15, January 5 - 8, 2000.

[SIDC 2002] http://www.astro.oma.be/SIDC.

[Simon 2001] Simon, M., The interpretation of radioactive isotopes in thecosmic radiation and the link between the diffusion halo model and theleaky box model, International School of Space Science-Astroparticle andGamma-Ray Physics in Space, 119, 2001.

[Simpson 1983] Simpson, J. A., Elemental and isotopic composition of thegalactic cosmic rays, Ann. Rev. Nucl. Part. Sci., 33, 323, 1983.

[Singer 1950] Singer, S. F., The zenit angle dependence of the cosmic radiationabove the atmosphere at λ = 41oN , Phys. Rev., 77, 729, 1950.

[Singer 1950b] Singer, S. F., The primary specific ionization and intensity ofthe cosmic radiation above the atmosphere at the geomagnetic equator,Phys. Rev., 80, 47, 1950.

[SOHO 2002] http://sohowww.nascom.nasa.gov.

[Sparvoli et al. 1997] R. Sparvoli et al., The telescope NINA to investigatenuclear fluxes in the near-earth space, Proceeding 25th Int. Cosmic-RayConf. (Durban), 2, 181, 1997.

208 BIBLIOGRAPHY

[Sparvoli et al. 1999] Sparvoli, R., Iannucci, A. et al., Observations of lowenergy Cosmic Rays with the detector NINA, Proceeding 26th Int. Cosmic-Ray Conf. (Salt Lake City), SH 3.5.01, 1999.

[Sparvoli et al. 2000] Sparvoli, R., Iannucci, A. et al., Launch in orbit of thetelescope NINA for cosmic ray observations: preliminary results, NuclearPhysics B, 85, 28, 2000.

[Sparvoli et al. 2000b] Sparvoli, R., Iannucci, A. et al., NINA: a silicondetector for cosmic ray astrophysics, Frontier Detectors for FrontierPhysics - 8th Pisa Meeting on Advaced Detectors”, La Biodola, Isolad’Elba, 21-27 May 2000.

[Sparvoli et al. 2001] Sparvoli, R., Iannucci, A. et al., Light isotopeabundance in SEPs measured by NINA, Proceeding 27th Int. Cosmic-RayConf. (Hamburg), 8, 3104, 2001.

[Spillantini et al. 1989] Spillantini, P., et al., Wizard 2: a proposal tocomplement Wizard detector with two further modules of trackingcolarimeter and an accurate spark counter TOF system, Il Nuovo Cimento,103 B, 625, 1989.

[Spillantini et al. 2001] Spillantini, P., et al., The PAMELA experiment,Proceeding 27th Int. Cosmic-Ray Conf. (Hamburg), OG 2215, 2001.

[Svestka 1976] Svestka, Z., Solar Flares, D. Reidel, Dordrecht, 1976.

[Tavani et al. 2001] Tavani, M. et al., Science with AGILE, Gamma 2001,American Institute of Physics 587, 729, 2001.

[Teegarden 1967] Teegarden, B. J., Atmospheric secondary and re-entrantalbedo protons, Journal of Geophysical Research, Space Physics, 72, 19,4857, 1967.

[Teegarden 1967b] Teegarden, B. J., Cosmic-Ray Production of Deuteriumand Tritium in the Earth Atmosphere, Journal of Geophysical Research,Space Physics, 72, 19, 4863, 1967.

[Telespazio 2001] http://www.telespazio.it.

[Terekhov et al. 1993] Terekhov, O. V., et al., Deuterium synthesis in the solarflare on 24 May 1990: observations of delayed emission in the 2.2 MeV γ-ray line by the GRANAT satellite, Astron. Lett., 19 (2), 65, 1993.

[The WiZard collaboration 1988] The WiZard collaboration, Flight proposalsubmitted in response to NASA A.O. NO. OSSA 3-88, November 1988.

[Treiman 1953] Treiman, S. B., The Cosmic-Ray Albedo, Phys. Rev., 91, 957,1953.

BIBLIOGRAPHY 209

[Tverskoi 1965] Tverskoi, B. A., Geomagn. Aeron. (USSR) 4, 224, 436, 1965.

[Tylka 2000] Tylka, A. J., A review of cosmic-ray albedo studies: 1949-1970,GLAST software meeting 15-17 May 2000, Document no. NRL SEM 2000-03, 2000.

[Ulysses 2002] http://ulysses.jpl.nasa.gov/index.html.

[Vacchi et al. 1997] A. Vacchi, et al., Isotope discrimination with theexperiment NINA, Proceeding 25th Int. Cosmic-Ray Conf. (Durban), 2,177, SH.4.5.1, 1997.

[Van Allen et al. 1950] Van Allen, J. A. and A. V. Gangnes, The cosmic-rayintensity above the atmosphere at the geomagnetic equator, Phys. Rev.,78, 50, 1950.

[Van Allen and Tatel 1948] Van Allen, J. A. and H. E. Tatel, The Cosmic-Ray Counting Rate of a Single Geiger Counter from Ground Level to 161Kilometers Altitude,Phys. Rev., 73, 245, 1948.

[Van Allen et al. 1968] Van Allen, J. A., et al., Observed Absence of EnergeticElectrons and Protons Near Venus, Journal of Geophysical Research, 73,421, 1968.

[Van Hollebeke 1985] Van Hollebeke, M. A., et al., Energetic particleobservations of the solar-gamma ray/neutron flare events of 3 Jun 1982and 21 June 1980 isotopic and chemical composition, Proceeding 19th Int.Cosmic-Ray Conf. (Paris), 4, 209, 1985.

[Vannuccini et al. 2001] Vannuccini, E., C. Grimani, P. Papini, and S. A.Stephens, An estimate of the secondary 2H spectrum produced by cosmicrays in the atmosphere, Proceeding 27th Int. Cosmic-Ray Conf.(Hamburg),5, 4181, 2001.

[Verma 1967] Verma, S. D., Measurement of charges splash and re-entrantalbedo of the cosmic radiation, Journal of Geophysical Research, 72, 915,1967.

[Verma et al. 1985] Verma, S. D. et al., Proton albedo spectrum observation inlow latitude region at Hyderabad, India, Proceeding 19th Int. Cosmic-RayConf. (Paris), 5, 320, 1985.

[Voyager 2002] http://voyager.jpl.nasa.gov.

[Walt 1994] Walt, M., Introduction to Geomagnetically Trapped Radiation,Cambridge U. press, 1994.

[Webber et al. 1975] Webber, W. R., et al., Pioneer 10 measurements of thecharge and energy spectrum of solar cosmic rays during 1972 August,Astrophys. J., 199, 482, 1975.

210 BIBLIOGRAPHY

[Webber and Yushak 1983] Webber, W. R., and S. M. Yushak, Ameasurement of the energy spectra and relative abundance of the cosmic-ray H and He isotopes over a broad energy range, Astrophys. J., 275, 1,391, 1983.

[Webber 2000] Webber, W. R., Reacceleration in a Monte Carlo DiffusionModel for Galactic Cosmic Rays, ACE-2000 Symposium, The Accelerationand Transport of Energetic Particles Observed in the Heliosphere, IndianWells, California, G7, January 5 - 8, 2000.

[Wefel et al. 1995] Wefel, J. P. et al., The isotopic composition ofGeomagnetically Trapped Helium, Proceeding 24th Int. Cosmic-Ray Conf.(Rome), 4, 1021, 1995.

[Wenzel et al. 1975] Wenzel, K. P., et al., Splash albedo protons between 4 and315 MeV at high and low geomagnetic latitudes, Journal of GeophysicalResearch, 80, 3580, 1975.

[White 1973] White, R. S., High-energy proton radiation belt, Rev. Geophys.Space Phys., 11, 595, 1973.

[Winckler et al. 1950] Winckler, J. R. et al., A directional and latitude surveyof cosmic ray at high altitude, Phys. Rev., 79, 656, 1950.

[Wizard 2002] http://wizard.roma2.infn.it/nina.

[Wolfendale 1963] Wolfendale, A. W., Cosmic Rays, Charles Birchall & sons,1963.

[Wu et al. 1979] Wu, J. R., C. C. Change, H. D. Holmgren, Charged-particlespectra: 90 MeV protons on 27Al, 58Ni, 90Zr, 209Bi, Phys. Rev., C, 19, 698,1979.

Acknowledgements

I would never have been able to complete this work without the help of manypeople.

I want to thank my supervisor Prof Piergiorgio Picozza who has been asource of good comments and questions. It has always been interesting todiscuss results and problems of the analysis with him.

Besides being a friend, Dr Vladimir Mikhailov has always been close to mywork, offering all his huge experience in each period of my Thesis.

I want to thank all peaple of WiZard NINA - Collaboration, especially DrRoberta Sparvoli, always ready in useful discussion.

Thanks to my friends in Laboratory, Vittorio Bidoli, Mauro Minori,Alessandro Basili, Anna Silina, Francesco Belli, Vincenzo Buttaro.

I want to thank my friends Daniele Ruggieri, Germano Percossi and TomFroysland. There are many reasons to do it.

Thanks to Maria Pia De Pascale, Livio Narici, Fausto Vagnetti.

Finally, I want to thank my family, my friends and the splendid person whoreally changed my life and to whom my thesis is dedicated.

211

anni accademici 1999-2002people.fisica.uniroma2.it/~iannucci/pubblicazioni/dottesi.pdf ·...

Documents