Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
16 Highlights This chapter discusses the problems being addressed by current and future cosmic-ray measurements. The general themes are organized by the history of cosmic-ray matter, starting with its synthesis, proceeding to its acceleration and propagation through interstellar space, and con- cluding with its interaction with matter. To set the stage, we start with a list of the major discoveries of the last decade. · New detectors have unambiguously resolved individual isotopes of neon, magnesium, and silicon. The resulting abundances show distinct quantitative differences from those found in the condensed bodies of the solar system, demonstrating conclusively that galactic cosmic rays are a sample of matter with a nucleosynthetic history that is different from that of the Sun. At the same time measurements of solar cosmic rays have provided some of the best measurements of the isotopic composition of the solar corona. · Cosmic-ray abundances of individual elements heavier than iron have been successfully measured, despite the extreme rarity of these nuclei. The results indicate that the cosmic rays are not dominated by material recently synthesized in supernova explosions, as data sug- gested a decade ago, but may well be accelerated interstellar material, a conclusion that is consistent with the isotope measurements of the lighter elements. 121
122 COSMIC RA YS · Measurements of the isotopes of the secondary element beryllium' in particular the abundance ratio of stable 9Be to radioactive '°Be, demonstrated that the cosmic rays that we observe today were accelerated on average 10 million to 20 million years ago and have propagated through interstellar material of mean density lower than the mean density of the galactic disk. · The radial gradient of cosmic rays In the ecliptic plane of the heliosphere has now been measured. The gradient is less steep than some earlier models had predicted, and the edge of the modulation region twhich had earlier been predicted to lie as near as 5 astronomical units (AU)] has been shown to be beyond 30 AU. · A low-energy Ltens of millions of electron volts/atomic mass unit (MeV/amu)] component with highly unusual composition was discov- ered. This anomalous component is rich in oxygen and nitrogen but lacks carbon. It suffers modulation with the solar cycle in the same sense as galactic cosmic rays, so it appears to be either galactic in origin or to be accelerated in the outer portions of the heliosphere. Its source and acceleration mechanism is a puzzle. · Observations of discrete sources of gamma rays with energies to 10~5 eV with ground-based detectors have identified a few cosmic-ray accelerators of great power. · At the highest energies, above 10~7 eV, ground-level air-shower measurements now give clear evidence of anisotropy in arrival direc- tion; above 10~9 eV this anisotropy suggests that these most energetic particles in nature may be of extragalactic origin. · Large new underground detectors designed primarily to search for nucleon decay have observed and measured the flux of neutrinos from cosmic-ray interactions in the atmosphere. These detectors are also being used to study multiple muon events and their relation to the composition of primary cosmic rays around 10~5 eV. In addition to these discoveries, a number of other observations also raise important questions for the future. These include the following: · Measurements of secondary products of cosmic-ray nuclear inter- actions in the interstellar medium indicate an energy dependence of the confinement process at energies from 1 to 100 GeV/amu (1 GeV = a billion electron volts). Unexpectedly high fluxes of antiprotons suggest that the cosmic-ray protons that produce them penetrate more matter before reaching us than do heavier cosmic rays. These data have altered our picture of the processes by which cosmic rays are confined to the galaxy and constrain models of cosmic-ray acceleration.
HIGHLIGHTS 1 23 · Ground-level observations indicate changes in cosmic-ray compo- sition at energies just above those reached so far by direct measure- ments. It appears that between 10'4 and 10'6 eV the cosmic rays are richer in heavy nuclei relative to protons than they are at lower energies, while at still higher energies, above 10~7 eV, protons may again dominate. · In 1972 measurements of attenuation in air of cosmic-ray protons up to 50 TeV (1 TeV = 1 trillion electron volts) indicated that the proton-proton cross section increases with energy. This inference was subsequently confirmed by direct accelerator measurements. More recently, results from large air showers suggest that this increase continues at least another four decades in energy. · A series of balloon flights of emulsion chambers has observed and measured the composition and interactions of heavy nuclei of up to 10'4 eV. In some cases the interactions produce up to 1000 secondaries. · The flux of solar neutrinos observed appears to be significantly lower than expected from fusion processes in the Sun. This discrep- ancy has become one of the major unresolved issues of current astrophysics. The following sections explore in more detail some of the topics listed above and their implications for future research. NUCLEOSYNTHESIS Measurements of the abundances of elements and isotopes in the solar system, as observed spectroscopically in the solar photosphere and directly in terrestrial, meteoritic, and lunar samples, have long formed the basis of our knowledge of the history of the solar system. These solar-system abundances have in turn become the benchmark for studies ranging from stellar structure and nucleosynthesis to the age and evolution of the galaxy. Galactic cosmic rays provide a sample of material from outside the solar system, which can be used to describe the composition of the Milky Way Galaxy at a time and place far removed from solar-system formation. The cosmic-ray measurements complement spectroscopic information derived from optical and millimeter-wave astronomy on stars and the interstellar medium. Some elements and isotopes that cannot be measured well spectroscopically are relatively easy to investigate in the cosmic rays, for example, neon, iron isotopes, and many of the rare elements heavier than iron. Abundances of radioactive nuclides and their daughters show that
124 COSMIC RA YS the solar system formed 4.6 billion years ago. Thus, the solar-system abundances have usually been taken to be representative of the interstellar medium at that time. However, recent observations of isotopic abundance anomalies in various meteoritic minerals give evidence for compositional inhomogeneity of the nebula that formed the solar system, and these observations give evidence for a significant `'last minute" infusion into this nebula of products of supernova nucleosynthesis. Thus the solar-system abundances probably do not measure the present interstellar medium and may not even be com- pletely representative of the general interstellar medium 4.6 billion years ago. Recent cosmic-ray measurements have resolved clearly the radioac- tive nuclide '0Be, which has a half-life of 1.6 million years. They demonstrate that the cosmic-ray nuclei that we observe today were typically accelerated about 10 million years ago, very recently when compared with the age of the solar system. Most of them reach us from distances much greater than a parsec but less than several kiloparsecs. Thus the cosmic rays sample a region that is large compared with the probable size of the protosolar nebula but probably does not extend to the center of the galaxy. Recent models suggest that the acceleration of the bulk of cosmic rays occurs in supernova shock waves propagating through the hot interstellar gas. It thus may be that the cosmic-ray composition is more representative of the interstellar medium than is the solar-system composition. While galactic cosmic rays provide an excellent sample of material from outside the solar system, energetic particles from the Sun, or solar cosmic rays, provide in some cases the best solar-system abundance data available. For example, the solar-system abundances of noble gas elements and their isotopic compositions, poorly deter- mined from meteorites or from optical observations of the Sun, can best be measured in solar cosmic-ray composition studies. The nucleosynthesis of the elements that make up the solar system has been understood as the sum of several processes. Primordial hydrogen and helium are burned in stellar interiors in a series of steps at increasing temperature and pressure, which release energy as lighter elements fuse to make heavier ones, building up eventually to elements in the iron peak. Elements heavier than nickel are principally produced by neutron capture' either slowly over periods of thousands of years in evolved stars the (slow) s-process or quickly in seconds during supernova explosions the (rapid) reprocess. Each nucleosynthesis process leaves a signature in relative abundances of various nuclides.
HIGHLIGHTS 1 25 In the cosmic-ray source composition we look for signatures that reveal the conditions under which these nuclei were synthesized. We also test models of nucleosynthesis based on solar-system abundances. Two points are clear from data already in hand: (1) The material that is accelerated to form cosmic rays has a composition that is different from that of the material that formed the solar system. This difference must reflect a difference in the conditions under which nucleosynthesis took place, or at least a different mixture of material from the various nucleosynthesis processes. (2) The composition of the cosmic-ray source material is distinguished from that of the solar system by subtle quantitative differences that require precise measurements. These points are pertinent to the plans for the next generation of experiments. Isotope Ratios Quantitative differences between cosmic-ray source and solar- system composition have been established by isotopic measurements with excellent mass resolution of the elements Ne, Mg, and Si. The abundance ratio 22Ne/20Ne is higher in the cosmic-ray source than in the solar system by a factor of about 4. The four relatively rare neutron-rich isotopes of Mg and Si are all about 60 percent more abundant in the cosmic rays (relative to the most abundant isotope of each element) than in the solar system. Several mechanisms have been postulated to explain these cosmic- ray enrichments of the heavier isotopes. These mechanisms involve nucleosynthesis of cosmic-ray elements under different conditions from those in the solar system, owing either to spatial inhomogeneities in the galaxy or to chemical evolution of the galaxy in the time between formation of the solar system (4.6 billion years ago) and acceleration of the cosmic rays (only about 10 million years ago). These mechanisms lead to quantitative predictions for expected isotopic composition of other cosmic-ray elements so that measurements with much higher statistical accuracy than are currently available of the elements S. Ar, and Fe should be able to distinguish among various models. Abundances of Heavy Elements In the charge region beyond Fe and Ni, the HEAD-3 experiment has shown that the cosmic-ray source is not dominated by a single nu- cleosynthesis process such as the r- or s-process. However, these results do not rule out an enhancement by a factor of as much as 2 in either the s-process or the reprocess contribution relative to the solar
126 COSMIC RA YS system. If the solar system were enriched in products of explosive (supernova) nucleosynthesis due to a nearby supernova shortly before condensation of the solar nebula, while the cosmic rays were a sample of "normal" interstellar material, lacking the "last minute" reprocess enrichment of the solar system, then one would expect the cosmic rays to appear enriched, by perhaps a factor of 2, in s-process nuclides. Further measurements of abundance ratios of heavy elements will help to resolve such questions. Precise decomposition of cosmic rays heavier than Ni into r- and s-process components will ultimately require isotope measurements. Both HEAD-3 and Ariel-6 data demonstrate that the abundance of actinide elements (Z ~ 90) in the cosmic-ray source is not greatly enhanced compared with that in the solar nebula, as was suggested by earlier measurements. In fact, the observed ratio of actinides to elements in the Z = 80 region is roughly 1 percent. This result already rules out a classical, actinide-producing reprocess episode of explosive nucleosynthesis in supernovae as the source of heavy cosmic-ray nuclei. However, this actinide abundance is so low that its measure- ment is limited by poor statistics; only one and two actinide nuclei have been observed by HEAD-3 and Ariel-6, respectively. Measurements of the relative abundances among individual actinide elements would show the age of these elements since nucleosynthesis. Figure 16.1 shows the expected relative abundances of actinide ele- ments as a function of time since synthesis in an reprocess event. A synthesis age of the order of 10 million years (the same as the cosmic-ray propagation time) as indicated by a U/Th ratio of about 5 would, for example, imply that cosmic-ray acceleration acts on freshly synthesized material and so would contradict the idea that the cosmic rays are a sample of today's general interstellar medium. On the other hand, if we assume that cosmic rays are a sample of today's interstellar medium and the solar system is a sample from 4.6 billion years ago, the U/Th ratio in the cosmic rays would provide a measure of the rate of reprocess nucleosynthesis in the galaxy since the formation of the solar system. Solar Neutrinos Recently the capability of detecting neutrinos from the Sun has opened a new window on stellar nuclear processes. The nuclear fusion occurring in the Sun is calculated to produce a detectable flux of electron neutrinos, and accordingly a large-scale experiment has been operating over the past decade in a South Dakota gold mine. In this experiment the inverse beta-decay of 37C1 to 37A is detected as
HIGHLIGHTS 1 27 loo 10 ~ - 10 2 to 111 _. _ 10 3 UJ an _ 10-4 _ _ 10-5 _ lo-6 19spt \ \ Pa Am\ \ , , , ,1 , 104 105 Th ~ ~ \ Pu \ Cm \ \ \ \ Np \ \ _ , ~ 1 1o6 107 \ TIME AFTER r-PROCESS EVENT (yr) ~o8 109 FIGURE 16.1 The relative abundances of the individual actinides as a function of time after their nucleosynthesis in an reprocess event. evidence for neutrino capture. The results of this experiment are enigmatic and important; they suggest a flux of neutrinos less than a third that calculated. As the neutrinos responsible for this reaction are of rather high energy, they come from a minor component of the solar nuclear cycle (boron beta-decay). The reason for the low flux might be due either to an error in our understanding of the solar cycle or to the loss of neutrinos through oscillations or other effects in the propagation from the Sun. In any case this experiment poses an outstanding challenge to our understanding of the astrophysics of stellar interiors, of nuclear physics, and of the elementary-particle physics of neutrinos. ACCELERATION Recent gamma-ray observations indicate that the bulk of the cosmic radiation of energy less than 10'3 eV observed near Earth originates in
128 COSMIC RA YS our galaxy. Coupled with the cosmic-ray age since acceleration and an energy density outside the heliospheric cavity of 1 eV/cm3 or greater, this suggests an average cosmic-ray luminosity close to 104 ergs/e for our galaxy. This is at least 10 times greater than the x-ray luminosity of our galaxy. Understanding galactic cosmic-ray acceleration is part of a con- centrated effort to understand all classes of energetic particle accel- eration in astrophysical settings. Acceleration of particles by the Sun has been directly observed. The scale of solar acceleration (energy, time, size) is much smaller than that for galactic cosmic rays. The latter can be as much as a million times more energetic than solar cosmic rays. Nevertheless some of the same theoretical approaches are used to understand both types of process. In addition, we see direct evidence via electron synchrotron emission that acceleration is also going on in such diverse objects as supernova remnants, radio galaxies, and quasars. If our experience with galactic cosmic rays is any guide, these objects may contain at least 100 times more energy in cosmic-ray nuclei. The acceleration of energetic particles is apparently a universal phe- nomenon and deserves a concentrated effort toward its understanding. Shock Acceleration Energy requirements suggest supernovae as the cosmic-ray sources, and early models of cosmic-ray origin assumed these discrete sources. The power-law spectrum led to later models, which incorporated diffuse, relatively slow acceleration by random collisions with massive moving magnetic knots in the interstellar medium. Then a trend back to discrete sources such as supernovae or pulsars took place because of the inefficiency of such second-order Fermi acceleration. This evolu- tion of ideas has been driven by continued improvement of the obser- vational evidence and development of the theories. The most recent acceleration models incorporate shock waves generated by supernova explosions traveling in low-density regions of hot interstellar gas, which accelerate cosmic rays trapped in the shock front. Essentially direct observation of acceleration of particles by shock waves in the solar cavity has stimulated and guided the development of the theory of shock-wave acceleration generally. Within the solar system there is enough information to relate the shape of the spectrum of accelerated particles and its termination to the nature and size of the accelerating shock. Extending this kind of understanding to galactic scales is clearly desirable.
HIGHLIGHTS 1 29 The most decisive observational constraints to theories of galactic acceleration will come from measurements of the energy spectra of the various cosmic-ray components; in particular, the energy dependence of the secondary/primary ratio at high energies is an important test of models of cosmic-ray acceleration and confinement. Currently avail- able data on the composition extend only to about 10'3 eV total energy. At still higher energies, our information at present is restricted to the study of showers of secondary particles in the atmosphere, making possible a determination of the overall energy spectrum of the parent particles but providing only an estimation of the primary composition. A better understanding of high-energy composition is essential. Acceleration Fractionation There is clear evidence that cosmic-ray elemental abundances after acceleration differ by factors of 2 to to from one element to another relative to the standard accepted solar-system abundances (derived from meteorites and the photosphere). These differences are orga- nized, at least to first order, by atomic properties of the elements; in particular Figure 16.2 shows that there is a clear correlation between the ratio of cosmic-ray source abundance to solar-system abundance and the first ionization potential of the element. This correlation sug- gests that the differences are affected by fractionation in the accelera- tion process or in some process that injects material into the acceler- ation region. A similar correlation with first ionization potential has been observed for the abundances of elements in the solar energetic particles when compared with the standard solar-system abundances, leading to the suggestion that similar fractionation effects occur in both solar and galactic acceleration or injection. An alternate viewpoint suggests that the standard solar-system abundances are in fact not correctly repre- sentative of the photosphere or of the interstellar medium. Further measurements of rare elements in the galactic cosmic rays and in the solar energetic particles may help to define the role of such fraction- ation in the acceleration processes. The striking underabundance of hydrogen in cosmic rays is poorly understood and does not fit the first ionization correlation. It could reflect some property of the acceleration mechanism that depends on the charge/mass ratio (which is unity for hydrogen but less than or equal to 1/2 for other nuclei). Alternatively, it could reflect a different origin for protons (and perhaps helium).
130 COSMIC RA YS 3.0 ~.0 _ :~' 0.3 c: 0.1 0.03 1 1 1 1 1 1 1 1 1 aft _ Get so ~ o :: 1 T AN iNe _ i He _ H ~ 1 1 1 1 1 1 1 1 1 1 4 6 8 10 12 14 16 18 20 22 24 26 FIRST IONIZATION POTENTIAL (eV) FIGURE 16.2 The elemental abundances of the cosmic-ray source relative to solar- system material are roughly ordered by the first ionization potential. However. some of the remaining differences are well beyond the indicated errors. Termination of Acceleration Mechanism Of particular importance in the future will be precise measurements of the proton spectrum extending to energies between 10'3 and 10'5 eV/nucleon. Here both the time and the size scales of the acceleration region for nuclei will eventually limit the energy attainable, leading to a break in the spectrum. Air-shower observations (which measure the spectrum of the total energy of cosmic rays their energy/nucleus) indicate that in the region around 10'5-10'6 eV (where the spectral steepening occurs), the com- position may become enriched in heavier nuclei. A rigidity-dependent termination of acceleration, as in the shock mechanism, implies a pro- gressive enrichment in heavy nuclei with increasing energy per nu- cleus. It is not yet clear, however, whether this picture is correct in detail. Direct observations of the composition and spectra between 10'3 and 10'6 eV are required in order to understand galactic cosmic-ray acceleration and containment models.
HIGHLIGHTS 13 1 The acceleration of solar-flare particles (solar cosmic rays) is another question. While the mean composition, averaged over many flares, is similar to that of the galactic cosmic-ray sources, including a correla- tion with first ionization potential, there are dramatic flare-to-flare variations that remain to be explained. The ratio of heavier elements (e.g., iron) to lighter elements (e.g., oxygen) varies by an order of magnitude from flare to flare, and the energy spectra also show wide variations. In addition some flares have anomalously high fluxes of 3He, thought to be the result of a cyclotron resonance in the acceler- ation region. Testing models of flare acceleration require correlated observations of particle spectra and of x rays (from accelerat- ed electrons) and gamma rays (from accelerated nuclei), as well as further measurements of the recently observed neutron flux from solar flares. High-Energy Gamma Rays Gamma rays of 10' i-106 eV energy produce electromagnetic cas- cades in the atmosphere that can be studied from the ground using atmospheric Cerenkov emission and cosmic-ray air-shower tech- niques. The Cerenkov light from these air showers is almost parallel to the shower direction (to approximately 1 degree) so that a telescope image of this Cerenkov light reveals a fuzzy spot that gives the direction from which the primary cosmic ray or gamma ray arrived. Recently, experiments involving surface arrays of particle detectors have identified gamma rays of up to 10'5 eV from Cygnus X-3 and possibly from other objects. By tracking the astronomical object of interest it is possible to separate the point-source gammas from the isotropic background of air showers produced by charged cosmic rays. The signal-to-noise ratio may be further aided through the use of accurate timing and the known timing of the source emissions. These studies are technically only an extension of astronomy to an extreme energy of the electromagnetic spectrum. The techniques used tie this area to other cosmic-ray programs. It is noteworthy that, with these observations, our study of radiation from the universe spans over 20 orders of magnitude ~ I 020) in wavelengths of electromagnetic radiation. The results so far have given us the first direct evidence of discrete astronomical locations of acceleration processes with energies of 10~5 eV (lOOO TeV). Although this field is only a few years old, the results are already having a major impact on our understanding of the . . ,~ . Origin ot cosmic rays.
132 COSMIC RA YS Anomalous Component An acceleration process that may have special significance, but about which very little is known, is responsible for the so-called anomalous component. Enhanced fluxes of certain nuclei such as He, N. O. and Ne are observed near the Earth at energies of 10 MeV/nucleon. Why only certain elements are enhanced, how they are accelerated, and why they appear at the Earth only at certain times are subjects of much discussion. A solar origin appears to be ruled out. We may be seeing direct selective acceleration of particles originating in the local interstellar medium, or we may be seeing particles from sources nearby in the galaxy with unusual composition. In either case, measurements of the charge and isotopic composition of these particles must be made at a new level of accuracy to understand the processes involved. GALACTIC COSMIC-RAY TRANSPORT AND THE INTERSTELLAR MEDIUM The cosmic-ray flux arriving near the Earth results from a convolu- tion of source compositions, charge-dependent selection during ac- celeration, fragmentation from interactions with the interstellar me- dium en route, and diffusion and scattering processes in the galaxy. Separating these different physical phenomena is a major task for the cosmic-ray program in the coming years. The galactic cosmic rays constitute a highly relativistic gas held in the galaxy for a time (107 years) that is long compared with the traversal time for highly relativistic particles across the galaxy (103 years) but short compared with the age of the galaxy (10~0 years). The physics of containment is poorly understood. We know from measure- ments of Faraday rotation and the polarization of starlight that the typical interstellar magnetic field is ~3 FIG. Thus, galactic cosmic rays, which range in energy from 1 GeV/nucleon to greater than 1 o6 GeV/nucleon, have gyroradii that range from about 0.1 AU to greater than 1 parsec (pc). However, the distribution of fluctuations in mag- netic-field magnitude and direction, which are presumably responsible for scattering the cosmic rays and trapping them in the galaxy, is unknown. Current estimates suggest that the bulk of the cosmic rays diffuse to us from distances greater than a parsec but less than several kiloparsecs. Key observational parameters for the question of cosmic-ray prop-
HIGHLIGHTS 1 33 agation and containment in the galaxy are the abundances of secondary cosmic rays (produced by interactions with the interstellar gas) relative to primary particles (accelerated in source regions). Particularly im- portant are positrons and antiprotons (generated by primary protons); the light elements Li, Be, and B (fragmentation products principally of C and O), and certain heavy elements, in particular Sc and V (produced by spallation of Fe nuclei). Abundances of secondaries, together with the fragmentation cross sections, give a measure of the average path-length At traversed by cosmic rays in their lifetime. If the average density of the medium is known, this can be translated into an average lifetime. Energy Dependence of Escape from Galaxy A fundamental result of measurement of secondary nuclei is that at higher energies the path length decreases approximately as A<(EJ x (~/~0) v I, decreasing to At ~ 1 g/cm'at around 100 GeV/nucleon. Only if such measurements are continued to still higher energies, i.e., well into the TeV/nucleon region, may one be able to explain the origin of this energy dependence of A`. For example if it is a consequence of the diffusion and convection processes by which cosmic rays are trans- ported out of the galactic confinement volume, then Al is predicted to continue to decrease as energy increases at a rate that reflects the spectrum of magnetic inhomogeneities in interstellar space. If, on the other hand, the effect is due to an energy-dependent escape mechanism in regions surrounding the acceleration sites, then be would become independent of energy at a value reflecting the amount of material traversed by cosmic rays after leaving the source. Several predictions are shown in Figure 16.3. It is important to emphasize that measure- ments above 100 GeV/nucleon will not only specify the mode of propagation of cosmic rays in the galaxy but will also enable us to deduce the energy spectra at the acceleration site. The behavior of the escape length as a function of energy below I GeV/amu is a subject of considerable current interest. There is some evidence that the distribution of escape lengths is energy dependent with an energy-dependent deficiency of short path lengths. Such a path-length distribution could result from a shell of material around the source regions, in which particles are trapped in such a way that low-energy particles pass through more material before escaping than do higher-energy particles. There is also evidence that the mean escape length becomes independent of energy below about 1 GeV/amu, a ~ ~ I ~ ~ _/1 ~ ~ ~
134 COSMIC RA YS ~ .0 0.6 0.4 o m <~ 0.2 c: in O 0.1 o m 0.06 0.04 0.02 TIC ~ ,~ lo) Balloon Data - t H EAO-3 Closed Galaxy \ - ~t Source 6. Surrounded by Matter Typical ~~ Diffusion ~ ~ Model ,, ,,.,., . ,,, ,,,,1 , ,,, ,,.,1 10 100 1000 ENERGY GeV/NUCLEON FIGURE 16.3 Various models for the containment and propagation of cosmic rays in the galactic magnetic fields will be tested by measurements in the energy range 1000 GeV/nucleon. Errors quoted for the highest-energy balloon data are much larger than those that can be obtained from satellite observations of sufficient duration. feature that is associated in some models with a change from a high-energy diffusion-dominated transport in the galaxy to a convec- tion-dominated regime as has been postulated in association with a galactic wind. This situation would be clarified by extending these studies to particles whose energy in the interstellar medium is below a few hundred MeV/amu, which requires direct observations of the unmodulated cosmic-ray spectra outside the solar system, or possibly over the solar poles. The low-energy galactic cosmic rays are also of interest because they are highly ionizing and couple strongly to the ambient interstellar medium. The cosmic-ray energy density is comparable with or greater than that of the interstellar magnetic field and the turbulent motion of the gas. Cosmic-ray pressure creates bubbles in the interstellar mag- netic field, puffing it out of the galactic plane, leading to the escape of cosmic rays. At the same time, gas then flows down the magnetic field, attracted by the gravitational potential of the galaxy, creating a shock wave that might trigger stellar condensation. Measurements outside the heliosphere are required to determine the contribution of these cosmic rays, most of which have energies below 100 MeV/nucleon.
HIGHLIGHTS 135 Correlation Between Anisotropy and Energy There is a striking correlation between the anisotropy and the flux of cosmic rays, as shown in Figure 16.4. If the anisotropy reflects large-scale flow patterns, a simple interpretation would suggest that there is a single underlying source spectrum of E-2 47 all the way from 10~2 to 10~9 eV, with the remaining observed structure associated with failure of the containment mechanism. The anisotropy measurements are made with long-duration, ground-based experiments observa- tions of muons underground at the lower energies and monitoring arrival directions of extensive air showers at higher energies. Statistical uncertainties are large at the higher energies, and measurement of composition around 10~5 eV is crucial for understanding these intrigu- ing results. Secondaries from Light Nuclei A special role is played by the electron component in the high-energy cosmic rays. Cosmic-ray electrons, consisting of negatrons mostly accelerated in source regions plus positrons that are predominantly the result of interstellar pep collisions, rapidly lose energy through radia- tive interactions with the interstellar magnetic and photon fields. This energy loss gives rise to much of the observed nonthermal radio and UJ c, - o 0.1 ~,` 4 - CL 1 o '_ 10 o at) a: ENERGY (eV) ~- 1o12 1014 1016 1 1 1 1o1 8 1o2o 1o2 x 104 J 6 103 a 11 - x ~ a) FIGURE 16.4 Amplitude of first harmonic as a measure of residence time: anisotropy (data points) compared with flux (line). Anisotropy has been corrected for solar motion below 1O'5 eV. [After A. M. Hillas. Annul Rev. Astron. Astrophys. 22, 425 (1984).]
136 COSMIC RA YS x-ray background emission of the galaxy. Because of radiative losses, the lifetime of electrons, and hence the distance they can propagate in the galaxy before losing a significant fraction of their energy' decreases rapidly with increasing energy. Thus electrons, observed with an energy of a few TeV at the Earth, must have been accelerated not further than a few hundred parsecs from the solar system. Measure- ment of these high-energy electrons therefore provides the unique possibility of identifying the distribution of local sources of the cosmic radiation. In the past 15 years, the total electron flux has been measured to about 1 TeV. The observation of the energy spectrum of positrons has a special importance. It makes possible a direct comparison to the source spectrum of positrons, which is known through calculations of the pep production process measured at accelerators. At present the positron spectrum is known separately only to around 10 GeV. If this measure- ment could be continued up to a few hundred GeV, it would give direct information on the deformation of the spectrum due to propagation effects and radiative energy losses. Such information cannot be unam- biguously obtained just from observations of electrons since their energy spectrum at the source is not known a priori. Thus, positron observations would lead to independent determinations of the confine- ment time of the electron component in the galaxy together with an estimate of the magnitude of the magnetic field traversed. Observations of other kinds of secondaries such as antiprotons, OH and 3He from interactions of protons, and helium nuclei provide information on the amount of matter traversed by the most abundant cosmic-ray constituents. Recent measurements of relatively high antiproton intensities at around 10 GeV suggest that protons may traverse 3 to 5 times as much matter as heavier nuclei. A similar situation seems to exist for helium based on recent observations of a high 3He/4He ratio. Very-low-energy antiproton measurements are even more difficult to interpret. More accurate observations of pos- itrons and antiprotons at different energies and of deuterium and The should be able to decide the question of whether protons and helium nuclei have different propagation histories from those of heavier nuclei. Propagation in Galactic Halo Observations of the radioactive secondary nucleus "'Be, interpreted within a simple (leaky-box) propagation model, indicate a cosmic-ray lifetime of about 10 million to 20 million years. Comparison with the average path length deduced from the secondary/primary ratio men-
HIGHLlGHT5 1 37 tioned above implies that the cosmic rays observed at the Earth propagate in a region with an average density less than that of the average interstellar medium in the disk This in turn suggests a containment volume that includes a galactic halo region as well as the disk. The interrelationship between the matter traversed by the parti- cles and their age is dependent on the size of the storage volume for cosmic rays. What is actually measured is the fraction of i°Be that survives radioactive decay. This depends not only on the mean cosmic-ray age but also on the distribution of ages, which is exponen- tial in the leaky-box model but is more complicated in models in which cosmic rays are stored in a large halo surrounding the galaxy. Further information about cosmic-ray time scales and hence about the storage volume will come from measurements of '°Be abundances at higher energies and of other clock isotopes. These data, in conjunction with electron and positron measurements, would be able to differentiate between halo and local storage models and to place constraints on the distribution of cosmic-ray sources in the galaxy. Connection with Gamma and Radio Astronomy The cosmic-ray composition studies discussed above give informa- tion on the distribution of cosmic rays and matter in the galaxy that is complementary to that obtained with gamma-ray and radio-astronomy surveys. Diffuse gamma rays are generated by interactions of cosmic rays with the interstellar gas; the nonthermal radio emission comes from cosmic-ray electron synchrotron emission in the galactic mag- netic fields. By studying this radiation we can also observe the cosmic rays in localized galactic objects (supernova remnants) and in external galaxies. These two different perspectives will be helpful in under- standing the role that cosmic rays and the magnetic fields play in the evolution and dynamics of astrophysical objects, from supernova remnants to giant radio galaxies. HIGH-ENERGY NUCLEAR AND PARTICLE PHYSICS From the point of view of high-energy physics there are several reasons to study cosmic rays: (1) to explore particle interactions at energies much higher than those accessible at accelerators; (2) to study processes involving neutrinos and high-energy nuclei that are also inaccessible to present machines; and (3) to look for signals from the early universe, such as a cutoff of cosmic rays above 102° eV due to the 3-K blackbody radiation, or the presence of antinuclei, which bears on
138 COSMIC RA YS the question of whether the universe is baryon symmetric on the largest scales. In addition there is considerable scope for applying particle physics to the study of cosmic-ray astrophysics, i.e., to determine the chemical composition and energy spectra of the primary cosmic rays in the high-energy region where the flux is too low for direct observation of the primaries. Different types of experiments are suited to the different regions of the primary energy spectrum as determined by the flux. This is indicated in Figure 15.1(B) in Chapter 15, which shows the integral flux as a function of primary energy. A scale showing equivalent nucleon- nucleon center-of-mass energies is superimposed. Note that the region of the second-generation hadron colliders (one of which is already in operation) to a large extent overlaps the 10~4-10'6 eV region, which includes the astrophysically interesting region of the energy spectrum referred to earlier. Because of the steeply falling primary spectrum there is a natural dividing line around 10~5 eV (or somewhat lower) between direct and indirect experiments. The total flux above this energy is only about 2 particles per (m2 sr week) at the top of the atmosphere. Since the flux decreases by about 2 orders of magnitude per decade increase in energy, it will continue to be necessary to explore higher energies with indirect, ground-based cascade experiments. Because of the antici- pated direct measurements of primary composition to 10'4-10~5 eV, coupled with current studies of hadron collisions in the same energy region, there is now a good prospect for improving significantly our ability to interpret the cascade measurements at the higher energies. Nucleon Decay Experiments as Cosmic-Ray Detectors Motivated by the particle-physics prediction of spontaneous decay of the free (or bound) proton, large detectors have been designed and built in this country and abroad that are sensitive to nucleon decay lifetimes of as great as 1033 years. These large detectors represent a unique opportunity to collect data on energetic muons and neutrinos from cosmic rays. The characteristics of the U.S. detectors are noted here together with specific comments on appropriate cosmic-ray ob- servations and opportunities. The largest operating proton-decay experiment employs an 8000-m3 volume of water located at a depth of 600 m, or 1570 m.w.e. (meters water equivalent), in a salt mine near Cleveland, Ohio. Signals from Cerenkov light produced by relativistic charged particles are detected by photomultipliers that line the six surfaces of the tank on a 1-m grid.
HIGHLIGHTS 1 39 Cosmic-ray neutrino interactions depositing energies of over 200 MeV are detected at a rate of about one per day. The detector has been in operation since August 1982. Two other smaller proton-decay experimental programs have also been carried out in the United States. At Park City, Utah, a 780-m3 water Cerenkov detector was operated at a depth of about 1700 m.w.e. A 30-ton detector at the Soudan mine in northern Minnesota is at a depth of 1800 m.w.e. and consists of a taconite-loaded cement with proportional chambers as the sensitive elements. Although much smaller than the other two detectors, its fine-grained tracking capability has enabled the detector to search for possible sidereal anisotropies of cosmic-ray multiple-muon events. A larger detector, Soudan 11, is scheduled to be constructed in the same mine employing the same general design philosophy. An unusual experiment has been developed in the Homestake gold mine in South Dakota. This detector consists of an array of plastic tanks filled with liquid scintillator, which, when brought into full operation, will have a sensitive mass of about 300 tons. It is located in the deep underground cavern occupied by the solar neutrino experi- ments. The primary objective of the experiment is to search for neutrino bursts that could be signatures of supernova explosions. This counter array is, of course, also sensitive to cosmic-ray muons. A surface array is being added to study the air showers produced by the same primary events that give rise to the detected muons. Although the expected number of energetic muons increases with primary energy, at fixed energy the muon multiplicity is correlated with the atomic weight of the primary cosmic ray. Consequently, the surface shower data and underground muon data together provide information concerning the atomic weight of the primary cosmic-ray nucleus. The Homestake data will be useful in studies of the mass spectrum of primary cosmic rays in the energy range 10'4-10'6 eV; these energies are about an order of magnitude greater than those accessible with the Cleveland proton- decay detector owing to the greater depth of 1480 m (4200 m.w.e.) of the Homestake mine. Nucleus-Nucleus Collisions The classic cosmic-ray emulsion technique has been modified into a hybrid emulsion chamber with target material and electromagnetic calorimeter sections (layers of plastic and lead, respectively, between photosensitive layers). The first observation of charmed particles was made over lO years ago with such detectors, and they have been
140 COSMIC RAYS adapted for use at accelerators to study charmed-particle spectroscopy and lifetimes. Scientists are currently collaborating internationally on the use of such emulsion chambers supplemented by electronic detec- tors to study primary nuclear composition and properties of nucleus- nucleus collisions. Several balloon flights have been carried out with emulsion chamber payloads to explore primary cosmic rays in the 10~2-10~5 eV energy range. This energy range is well beyond that accessible to current heavy-ion accelerators, and there are fundamen- tal and novel questions accessible to this kind of cosmic-ray experi- ment, in particular, the question of whether a new phase of quark-gluon matter can be achieved in collisions between heavy nuclei at high energy. Events in which heavy cosmic rays interact to produce nearly 1000 secondary particles have been observed. The energy-density implied by such multiplicities has been calculated to be above the threshold for production of a quark-gluon phase. Over 200 interactions have been analyzed wherein the primary energy exceeds 10~2 eV. Cross Sections, Spectra, Anisotropies, and Composition of Primary Cosmic Rays Above 10~7 Electron Volts Above 10~6 eV cosmic rays remain of interest for high-energy physicists as well as for astrophysicists, at least until the operation of a supercollider, which may be completed in the l990s. The goal here is to determine both cross sections for hadron interactions and the composition of the primaries. Recent measurements at the CERN pp collider have confirmed earlier cosmic-ray estimates of the proton cross section up to 10~4 eV (equivalent to center-of-mass energy of 500 GeV). New air-shower experiments have the potential to measure the proton cross section and to determine the gross features of the primary composition as well as in the 10~7-eV to 10~9-eV (center of mass about 100,000 GeV) range, where there may be a transition to extragalactic cosmic rays. The most ambitious cosmic-ray air-shower experiment in the United States is the Fly's Eye experiment being carried out in Utah. This detector consists of two arrays of photomultipliers deployed 3 km apart to observe the air scintillation light produced by extensive air showers. The phototubes are grouped in the focal plane of spherical mirrors, so that the arrays provide a mosaic image of the sky, with each phototube sensitive to a hexagonal cone of 5° of the celestial sphere. Timing information is also available, so that an air shower is recorded as a series of phototube "hits," with a pulse amplitude and relative time recorded for each. The data are sufficient to reconstruct completely the
HIGHLIGHTS 14 1 air shower in space and absolute magnitude. The Fly's Eye data have two major strengths. First, the Fly's Eye covers or "sees" an effective area comparable with the largest surface air-shower array; the current detector is sensitive over an area of almost 100 km2, although data can only be collected on clear, dark nights. Second, this detector permits the observation of the longitudinal profile of the shower, hence providing information on the height of the primary interaction and on the rate of development of the shower. These data in turn may be interpreted in terms of the inelastic cross section of protons at very high energies and in terms of the primary nuclear-mass composition. It may also be possible to relate the rate of development and shape of the shower with the secondary-particle multiplicity and other inclusive parameters of proton interactions. The Fly's Eye experiment has achieved a major milestone by directly observing the longitudinal development of individual cascades. Present results from this experiment and other air-shower experiments already suggest that the proton-air cross section is larger than 500 mb at 10~8 eV, as compared with its low-energy value of 280 mb. Magnetic Monopoles Most Grand Unified Theories (GUTs) predict the existence of massive magnetic monopoles, quanta of isolated north or south mag- netic poles with discrete magnetic-pole strength. Their masses are predicted to be of the order of 10~6 GeV (or about 0.01 Age, although some models yield significantly lighter or heavier masses. In the standard big-bang cosmology, GUT monopoles are produced at an early stage of the universe. By contrast, in the inflationary-universe scenario there would be no significant monopole production. The density of monopoles in the universe today can be bounded by arguments based on the openness of the universe and the mass of missing, or dark, matter. Another astrophysical upper limit on the monopole flux is based on the long-term stability of galactic magnetic fields. Within these limits, monopoles may exist in the universe with velocities in the range of lo-4 to 10-3 the velocity of light. At the lower end of this velocity range, some theorists suggest that they could be gravitationally bound to the solar system, which might enhance their local abundances. If GUT monopoles are able to catalyze proton decay, as suggested by some current theories, they would produce copious x rays from neutron stars. Our present failure to observe these x rays can be used to set more stringent limits to the monopole flux for this specific
142 COSMIC RA YS monopole type. The proton-decay catalysis would also be detectable in proton-decay experiments; thus far this process is not observed. Searches for monopoles have been conducted with superconducting coils and with ionization and scintillation detectors. In the former, a monopole passing through a coil would induce a current step that is readily detectable with sophisticated instrumentation. This technique has the advantage that the monopole signal would be almost totally independent of the monopole velocity. Such coils are limited, however, in their size. Ionization and scintillation detectors can be made with larger areas but are calculated to be insensitive to monopoles moving slower than about 5 x 10-4 the velocity of light. A signal consistent with a monopole interpretation was reported in early 1982, using a superconducting coil. However, subsequent searches by three groups (including the original 1982 author) have failed to find further evidence for a monopole using the same tech- nique. These searches have extended the sensitivity by almost a factor of 100. In addition, data using scintillators have set still more stringent limits on the flux over the velocity range accessible to them. Although the 1982 event remains unexplained, the monopole hypothesis for that event now seems unlikely.
