Solving a century-long mystery: the origin of galactic cosmic rays

An image to illustrate cosmic rays

A steady rain of charged particles, moving at nearly the speed of light, falls upon our planet at all times and from all directions – Bruno Rossi.

A steady rain of charged particles, moving at nearly the speed of light, falls upon our planet at all times and from all directions – Bruno Rossi. The charged particles Rossi is speaking about are called cosmic rays. They are of extra-terrestrial origin and constitute the fourth substance of the visible Universe – after matter, radiation and magnetic fields. CRs were discovered in 1912; the key contributor to the discovery, Victor Hess, was awarded a Nobel Prize in 1936.

For decades, two general research areas related to particle physics and astrophysics have been at the core of cosmic rays studies. Cosmic rays played a significant role in the development of particle physics, in particular, several fundamental particles, such as the positron and μ- and π- mesons have been discovered in cosmic rays. The dominance of the particle physics aspect continued until the 1960s, when human-made accelerators became the primary tools for the study of elementary particles. However, cosmic rays remain the only available channel for interactions of elementary particles above 1018 eV. Moreover, they are considered as potential messengers of information about the so-called ‘topological defects’ of hypothetical massive particles – relics from early Universe.

Cosmic rays as secondary products of decay or annihilation of Dark Matter have often been often invoked for the explanation of different ‘anomalies’, in particular the unusually high fraction of positrons and antiprotons in cosmic rays. These ‘exotic’ but not fully unrealistic interpretations give a certain cosmological ‘flavour’ to cosmic rays studies.

Exploring the non-thermal Universe

Currently, the cosmic ray community is focused more on the astrophysical aspect of this research. Often, this is reduced to the question of identification of major contributors to the ‘local fog’ – directly detected fluxes of cosmic rays in the Earth’s neighbourhood. However, the term ‘cosmic ray’ inhibits much broader implications, linked to the exploration of the non-thermal Universe.

The recent discovery of high energy gamma-rays from objects representing almost all known astronomical source populations tells us that the production of the charged relativistic particles – the parents of these gamma-rays – takes place in perfectly designed machines.

These cosmic rays factories exist throughout the entire Universe; they appear in a wide diversity of forms and on different scales. We see gamma-radiation from compact relativistic objects from solar mass black holes and neutron stars to largest cosmological structures like galaxy clusters.

They are unique laboratories to explore how the combination of processes related to classical and quantum electrodynamics, special and general relativity, particle physics, plasma physics, magnetohydrodynamics, etc., provide transformation of the available thermal and electromagnetic energy into non-thermal particles. The transformation efficiency in some objects, e.g. in the so-called ‘Pulsar Wind Nebulae’ (with the famous Crab Nebula as the brightest representative), approaches 100%.

Detection of cosmic rays

Cosmic rays consist of stable charged particles; protons, nuclei, and electrons with a small fraction of neutral particles; gamma-rays, neutrinos, as well as antiparticles; positrons and antiprotons. The kinetic energy of CR protons and nuclei span over 11 decades, from 109 to 1020 eV (see Fig.1).

All cosmic ray species, except for neutrinos, are effectively absorbed in the Earth’s atmosphere. Therefore, an ideal cosmic ray detector would be a space-based instrument which measures the direction, energy, and charge/mass of primary particles. The most advanced space-based cosmic ray detector is the Alpha Magnetic Spectrometer (AMS-02), which is mounted on the International Space Station (ISS) (see Fig. 2A). Its operation is based on the latest detection methods and technologies used in particle physics experiments.

The flux of cosmic rays falls rapidly with energy (see Fig. 1); around E=1015 eV the integral flux is about one particle per m2 per year, while at energies above 1020 eV it does not exceed one particle per km2 per year. This limits the potential and the energy range of cosmic rays studies by space-based instruments. Fortunately, cosmic ray measurements at very high energies are possible from the ground by detecting the secondary particles, the products of interactions of primary cosmic rays with the atmosphere. This can be done either directly or through their electromagnetic (Cherenkov or fluorescence) radiation. This is a common approach in cosmic ray studies at high energies. In particular, this approach is used in the Pierre Auger Observatory, a huge array of detectors which covers a 3,000 km2 area (see Fig. 2b). Our current knowledge on the flux, composition and anisotropy of cosmic rays in the highest energy band 1018 – 1020 eV is largely based on the data obtained with the successful operation of this instrument located in Argentina.

The word ‘astronomy’ means the direct observations of extra-terrestrial objects. This definition is relevant to photons, neutrinos, and gravitational waves, i.e. massless, neutral and stable particles. But for cosmic ray electrons, protons, and nuclei, the term ‘astronomy’ is used with a certain reservation. Because of the deflections of electrically charged particles in the chaotic interstellar and intergalactic magnetic fields, the information about their original directions pointing to the sites of their production is lost. Instead, on the Earth, we detect an (almost) isotropic flux of cosmic rays contributed by a huge number of galactic and extragalactic sources.

These objects have different origins and are characterised by essentially different physical parameters – energy budget, time history, acceleration mechanisms, etc. This makes the task of trying to reveal the origin of cosmic ray sources based on the ‘smell’ (chemical composition and energy spectra of particles) of the ‘soup’ (isotropic flux of cosmic rays) cooked over cosmological timescales a particular challenge. But gamma-rays, and, in future, hopefully also neutrinos, are indispensable messengers of information about the locations of cosmic ray factories.

Gamma-ray detectors

The range of cosmic gamma-rays spans from low (or MeV; 1 MeV = 106 eV) and high (or GeV; 1 GeV=109 eV) to very-high (or TeV; 1 TeV=1012 eV) and ultra-high (or PeV; 1 PeV=1015 eV) energies. The study of gamma-rays requires specially designed detectors to be capable of measuring the arriving direction of primary particles and their energy with good accuracy. The low and high energy bands are covered by space-based instruments, while the very- and ultra-high energy bands can be effectively studied by ground-based instruments.

Our current knowledge about the MeV/GeV gamma-ray sky basically comes from the results of the Fermi and AGILE gamma-ray space telescopes. The data obtained with the so-called pair-conversion tracking technique confirms the most optimistic prelaunch expectations concerning the discovery of thousands of cosmic gamma-ray emitters.
Even more impressive are the achievements in the TeV band. Thanks to the HESS, MAGIC (see Fig. 2C), and VERITAS arrays of atmospheric Cherenkov telescopes, the field has quickly been elevated to the level of a truly astronomical discipline. Over the last 10-15 years, observations with these instruments have resulted in many discoveries of topical importance related, in particular, to the origin of cosmic rays, and the potential of the technique is not yet saturated.

Among the primary motivations of the next generation of Imaging Atmospheric Cherenkov Telescope (IACT) arrays two objectives are of particular importance:

• Significant, by order of magnitude, improvement of the flux sensitivity in the standard 0.1–10 TeV interval
• Dramatic expansion of the energy domain down to less than 10 GeV and up to more than 100 TeV

The Cherenkov Telescope Array (CTA), the major next-generation gamma-ray detector, is designed to meet these ambitious goals. The IACT arrays are proposed to observe point-like or moderately extended objects, although the superior sensitivity and relatively large fields of view should allow effective all-sky surveys.

On the other hand, the capability of these arrays is limited for the search of very extended structures, as well as for the detection of solitary or transient gamma-ray phenomena. In this regard, a detection technique based on direct registration of particles that comprise the extensive air showers, such as the water Cherenkov detector HAWK (Mexico), and especially the future LHAASO array (China), is a complementary approach to the IACT technique.

Detecting neutrinos

As messengers of high-energy phenomena in the Universe, neutrinos are similar to gamma-rays, but there are also principal differences between them. Neutrinos are only produced in hadronic interactions, for instance. Also, unlike gamma-rays, neutrinos interact only weakly with the surrounding matter and magnetic and radiation fields.

Thus, they carry information about high-energy processes in ‘hidden’ regions which are not transparent for gamma-rays. This gives neutrinos a particular uniqueness. At the same time, ironically, this feature plays a negative role for the detection of neutrinos. To compensate for the tiny interaction cross-sections, huge volumes of natural water or ice are used as targets for their absorption.

Recently, thanks to the IceCube neutrino telescope located in the South Pole, the first extra-terrestrial high energy neutrinos have been detected. Although the evidence of ‘astrophysical’ origin of the detected multi-TeV neutrinos is convincing, the origin of these neutrinos remains highly uncertain. The on-going upgrade of IceCube and the construction of the second cubic-km scale neutrino detector in the Mediterranean Sea, KM3NeT (see Figure 2D), are aiming at the discovery of first sources of high energy neutrinos.

The origin of galactic cosmic rays

Cosmic ray fluxes are dominated by directly accelerated particles – protons, nuclei, and electrons. The light elements of the LiBeB group, as well as antiparticles (antiprotons and positrons), have a secondary origin, i.e., are produced later, in the interactions of the primary particles with the ambient gas. The antiparticles can also be contributed by exotic channels linked to the evaporation of primordial black holes and the annihilation of Dark Matter.

In general, the cosmic ray composition in different energy bands is well known. In the best-studied energy interval between 1 GeV and 1 TeV, the protons, nuclei, and electrons contribute to the overall flux in proportions 100:10:1. The energy spectrum has two distinct features – the so-called ‘knee’ at 1015 eV, and the ‘ankle’ at around 1018 eV (see Fig. 1).

Presently, there is a consensus in the cosmic ray community that the particles below the knee are of galactic origin. The soundest argument is related to the propagation time of these particles; in the case of location of the cosmic ray sources well beyond our galaxy, the arrival time of particles of these energies would exceed the age of the Universe (10 billion years). On the other hand, the particles with energy exceeding 1018 eV could not be produced in our galaxy; otherwise, a strong anisotropy would be detected, in contrast to the measurements. The origin of the energy range between the knee and the ankle remains highly uncertain.

The cosmic ray ‘sea’

The cosmic ray diffusion and convection in galactic magnetic fields produce the so-called ‘sea’ of cosmic rays. At high energies, cosmic rays freely penetrate into the solar system. Therefore, directly measured flux of local cosmic rays should not deviate significantly from the level of the ‘sea’. Beyond the solar system, the ‘sea’ can be traced by gamma-rays. The Fermi LAT measurements of diffuse gamma-ray emission of the galactic disk, as well as the detection of gamma-rays from individual massive gaseous complexes located in different parts of the galaxy, shows that the level of the ‘sea’ is almost constant throughout the galactic disk. It is close to the local cosmic ray level directly measured by AMS-02, wCR =1eV/cm3. The only exception is the ring between 4 to 6 kpc galactocentric radii. This result can be explained by the concentration of supernova remnants and clusters of young massive stars in the 4-6 kpc ring.

The production rate of cosmic rays in the galaxy contributed by all accelerators is estimated to be between 3 and 10 times 1040 erg/s. It does not exceed 0.1% of the total luminosity of the Milky Way, contributed by the radiation of stars. However, because of effective confinement of charged particles in the interstellar magnetic fields, the energy density of galactic cosmic rays is close to the energy density (or pressure) of the infrared/optical radiation produced by stars and dust. Moreover, it is also comparable with the pressure caused by the thermal gas and magnetic fields in the interstellar medium implying that cosmic rays play an essential role in the dynamics of the galactic disk.

Supernova remnants

The current paradigm of galactic CRs assumes that SuperNova (SNe) explosions or Supernova Remnants (SNRs) – the results of these gigantic events – are responsible for galactic cosmic rays. Over many decades, this conviction has been based on phenomenological arguments and theoretical meditations. As early as 1933, W Baade and F Zwicky recognised the comparable energetics characterising SN explosions and cosmic rays and envisaged a link between these two phenomena. What concerns SNRs, a strong argument in favour of SNRs as cosmic ray accelerators, comes from the theory side. The so-called ‘Diffusive Shock Acceleration’ has been established as a viable mechanism for cosmic ray acceleration in young SNRs.

A great achievement of gamma-ray astronomy in recent years was the discovery of TeV gamma-rays from a number of young SNRs. The gamma-ray image of the supernova remnant RXJ 1713.7-3946 obtained by the HESS arrays of Cherenkov telescopes is shown in Fig. 3A. The striking shell-type morphology of the TeV emission is explained by the shock-accelerated protons through interactions with the gas of compressed shell (the so-called ‘hadronic model’) or by inverse Compton scattering of relativistic electrons interacting with the surrounding radiation fields (the ‘leptonic model’).

Fig. 3B shows the energy distributions of parent particles, protons and electron, derived from gamma-ray data. Unfortunately, the current data do not allow us to distinguish between the hadronic and leptonic origin of gamma-rays. Nevertheless, even assuming that gamma-rays are produced by accelerated protons, we face a problem related to the break in the proton spectrum at around 1014 eV (see Fig. 3B). This tells us that this source is not a PeVatron, an accelerator boosting the proton energy to 1 PeV (1015 eV). In the case of other SNRs, the cut-off appears at even lower energies. Moreover, theoretical developments in recent years have also revealed serious problems, at least in standard schemes, for accelerating particles to PeV energies. This has raised doubts inside the cosmic ray community regarding the ability of SNRs to operate as CR PeVatrons.

Clusters of young massive stars

Massive stars produced at the collapse of giant molecular clouds form compact groups and remain tightly linked during their entire life. The colliding stellar winds and SN explosions in these clusters consisting of tens of massive and luminous stars can drive giant superbubbles filled with highly turbulent plasma and strong shocks. The acceleration of particles in these objects can be initiated by interacting winds either directly in the vicinity of stars or in the superbubble. The conditions in these objects can be more favourable for particle acceleration to PeV energies than in individual SNRs. This makes the clusters of massive young stars an attractive alternative to SNRs as cosmic ray factories contributing to the highest energy cosmic rays.

So far, the gamma-ray observations of the Milky Way have revealed only a few sources with characteristics of a PeVatron signature – ‘hard power-law gamma-ray spectra extending to tens of TeV without an indication of a break.’ Remarkably, these sources are found as diffuse gamma-ray structures, ‘cocoons’, surrounding famous young stellar clusters.

Fig. 4 shows the gamma-ray luminosities of extended regions around the stellar clusters Cyg OB2 and Westerlund 1 as well as the so-called ‘central molecular zone’ (CMZ) in the galactic centre (GC). The energy distributions of gamma-rays of three sources can be described by the same power-law spectra extending out to 10 TeV and beyond with an index 2.2-2.3. The spectra of parent protons almost mimic the gamma-ray spectra and should continue without a break up to ~1 PeV. The spatial distribution of protons derived from the morphological studies of gamma-rays and the gas density are also similar. The derived 1/r decrement of the cosmic ray density is a distinct signature of continuous, over a few million years, cosmic ray injection into the interstellar medium implying strong evidence that the extended gamma-ray structures are linked to protons which originate from the stellar clusters.

While in the case of Cyg OB2 and Westerlund 1 the association of these clusters with the gamma-ray structures can be claimed with high confidence, the identification of cosmic ray accelerators that initiate the diffuse emission of CMZ in the GC is disputable. The compelling similarity of both energy and radial distributions of multi-TeV cosmic rays extracted from gamma-ray observations of three objects can be interpreted as a hint that three ultracompact stellar systems located at the GC, Arches, Quintuplet and Nuclear clusters, are responsible for the gamma-radiation of the CMZ. However, in this case an equally plausible alternative is the super massive black hole at the dynamical centre of our galaxy.

The galactic centre

Similar to most of the other galaxies, the Milky Way contains a super massive black hole (SMBH). Formally, the mass of the central SMBH of about five million solar masses allows an energy release with a rate exceeding the overall luminosity of all stars in the Milky Way. The energy release can proceed in different forms, in particular, in cosmic rays via acceleration either in the vicinity of the SMBH (close to the event horizon) or at large distances, due to the termination of the relativistic outflow that originates close to the black hole.

To sustain the observed gamma-ray luminosity of CMZ, the average proton acceleration rate during the last 10,000 years should be at the level of 1037-1038 erg/s. It constitutes 1% of the current accretion power of the SMBH.

In the past, the SMBH could operate at much higher accretion rate. Correspondingly, the cosmic ray production rate could be one or two orders of magnitude higher. The gamma-ray fluxes from the GC are not sensitive to the history of operation of the proton PeVatron in the past. Such information, however, could be memorised in the radiation of ‘relic’ protons which already escaped the GC, and presently interact with the low-density gas beyond the Galactic Disk, in the so-called ‘Fermi Bubbles’, or in even larger, up to 100 kpc halo type structure surrounding the galaxy. The multi-TeV neutrinos recently detected by IceCube could, in principle, be linked to this scenario.


The last decade has witnessed impressive progress in our knowledge about galactic cosmic rays. The AMS-02 experiment has introduced new standards in the precision of cosmic ray measurements. The reported energy spectra of cosmic rays confirms, with unprecedented accuracy, the earlier claims regarding the substantial hardening of the energy spectrum of protons and nuclei above 100 GeV as well as the anomalously large content of secondary positrons and antiprotons in cosmic rays. Both effects point to the existence of (at least) two components of cosmic ray protons and nuclei dominated below and above 1 TeV, and contributed, most likely, by two different source populations.

The new cosmic ray measurements have provided an added value to our knowledge about cosmic rays, but they alone cannot address the questions related to the nature of cosmic ray factories. In this area, the most progress has been achieved thanks to the gamma-ray observations with satellite-borne and ground-based instruments. The results obtained with Fermi LAT and the arrays of atmospheric Cherenkov telescopes confirmed the most optimistic expectations with a discovery of thousands of cosmic gamma-ray emitters representing more than a dozen astronomical source populations.

In the context of the origin of cosmic rays, supernova remnants remain the primary candidates as major sources of galactic cosmic rays, however, presumably with a reduced role at highest energies. On the other hand, multi-TeV gamma-ray observations provide evidence that the clusters of young massive stars operate as PeVatrons. Thus, they could appear the dominant contributors to the cosmic ray flux around the ‘knee’. The role of the supermassive black hole in the GC could also be significant.

The extension of spectrometric and morphological gamma-ray measurements up to 100 TeV and several degrees in the angular size from regions surrounding supernova remnants and young stellar clusters is needed for crucial information about the origin of cosmic rays in general, and the physics of proton PeVatrons, in particular. Such observations with the CTA and LHAASO, will be available in the coming years.

It seems we are at the threshold of obtaining a solution to the century-old enigma of the origin of galactic cosmic rays.

Professor Felix Aharonian
Dublin Institute for Advanced Studies
Max-Planck Institute for
Nuclear Physics
+353 1 662 1333 ext 317

Laboratory Supplies Directory - Now Live


Please enter your comment!
Please enter your name here