Imagine the Universe!
rosat header image (

SNR and Cosmic Ray Acceleration

Cosmic Rays: What Are They?

Cosmic rays are the atomic nuclei (mostly protons) and electrons that are observed to strike the Earth's atmosphere with exceedingly high energies. Cosmic rays can be over 1021eV, which is a billion times more energetic than high energy particles created on earth in the most powerful particle accelerators. They are moving at nearly the speed of light.

They were first discovered by Victor Hess, during a balloon flight. Although Hess did not know what the particles were or where they came from, he observed a source of radiation (which he thought, at the time, were gamma-rays). He noticed that there was more radiation the higher up he rose and concluded that the Earth could not therefore be the source of the emission. This was the first time that an external source of energetic particles had been discovered. Cosmic Rays have been widely observed since then and are found just about everywhere in our Galaxy. The big question is: how and where are normal protons and electrons accelerated to these tremendous energies?

Where are Cosmic Rays Accelerated? Current Theory...

One characteristic which provides a clue as to how cosmic rays are accelerated is the spectrum of particles we catch here on Earth. The cosmic ray spectrum is fairly well described by a broken power law, that is a power law with two "kinks". A "power law" has the form F ~ Ep. The "spectral index", p, is the slope of the graph of the data (in logarithmic units). The slope is flatter at high energies than the thermal spectrum, for example, so that there are more cosmic rays at high energies. Such a spectrum is said to be "hard" (as opposed to "soft", which means the numbers fall off rapidly at high energies). The cosmic ray spectrum steepens around 3 X 1015 eV (the "knee") and flattens around 3 X 1018 eV (the "ankle"). Any theory of cosmic ray origins must account for this shape. A successful model must produce the right numbers of particles as a function of energy; in other words, the spectrum constrains models of cosmic ray production.

'knee' of cosmic ray spectrum

The "knee" of the cosmic ray spectrum

The standard belief is that some cosmic rays are accelerated in the Galaxy and some are accelerated outside the Galaxy.

The origin of the really high energy cosmic rays (above the knee) is a mystery. Nobody knows exactly where they are produced and a source has yet to be precisely detected. They are hard to pin down, since at such high energies there are so few of them that it is hard to tell exactly where they come from. For example, cosmic ray particles with energies greater than 1019eV hit the Earth at a rate of one per square kilometer per century (see the page by the HiRes project to study high energy cosmic rays ( It is difficult enough to observe these particles, never mind determine directly where their origin is. For a number of reasons, it is suspected that these cosmic rays above the ankle are extragalactic in origin, perhaps generated in the cores of Active Galactic Nuclei, in powerful radio galaxies, or by cosmic strings. These sources are known to have the tremendous amounts of energy needed to accelerate particles to such high energies, though a direct correlation has yet to be found. This is an area of active research, and as more and more sensitive detectors come on line and more evidence is gathered, scientists will have a better picture of where these extraordinarily high energy particles are generated.

The particles below the ankle have lower energies and are thought to be produced in the Galaxy. Furthermore, there is reason to believe that at least up to about 1014eV, if not all the way to the knee or to the ankle, most of the particles are accelerated in the shocks of supernova remnants ( In this model, particles are scattered across the shock fronts of a SNR, gaining energy at each crossing. Until the last two years, evidence to support this claim was circumstantial, based on theory and logic rather than on observations. For example, it seemed reasonable that SNR shocks could accelerate particles to the desired energies. The kinetic energy released by supernova explosions is more than enough to account for the Galactic cosmic rays up to 1015 eV. And supernovae are fairly common and occur throughout the Galaxy, so it is reasonable that they could be responsible for cosmic rays, which are also plentiful and found throughout the Galaxy. However, astronomers strove to find more direct evidence for shock acceleration of particles in SNR.

Puzzle Pieces Fall Into Place...

What would be the best direct evidence to support this model? Is there a way to "see" high energy, accelerated particles and associate them with SNRs? In fact, in a sense, there is. The key is synchrotron radiation. Synchrotron radiation is emitted when fast moving ions are accelerated in magnetic fields. A magnetic field will force an energetic particle to travel in a helical path and emit radiation. We know that magnetic fields exist near and around SNR, for example in the ambient ISM or around the neutron star that can remain after a SNe. So if there are fast moving charged particles, they should produce synchrotron emission which we should be able to observe. Where in the electromagnetic spectrum do we expect this synchrotron emission from Galactic cosmic rays to occur?

The answer depends on a couple of factors. The energy of synchrotron emission depends on the energies of the fast charged particles doing the emitting and the strength of the magnetic field doing the bending. For the energies of the cosmic rays (as expected based on cosmic rays observed on Earth) and the strength of the magnetic fields (deduced from radio measurements), cosmic rays' synchrotron emission should fall into the X-ray range. In fact, it appears that the best way at present to look for direct evidence for particle acceleration in SNRs (to at least ~ 1014 eV) is to look in the vicinity of SNR shocks for X-ray synchrotron radiation from electrons that have ~ 1014 eV of energy. Radio synchrotron radiation in SNR had long been observed and optical synchrotron radiation had been observed in the 1950s, but these observations could not account for the range of energetic cosmic rays seen with energies up to 1015eV.

The characteristic spectrum of synchrotron radiation is featureless, following a more or less straight line. This is in contrast to a spectrum from a hot radiating gas, which has many bumps and peaks corresponding to emission from particular atoms at particular energies. Despite this characteristic shape, synchrotron radiation in the X-ray region of the spectrum is not easy to identify.

False Starts

There are many reasons for this: until recently, other theories have successfully accounted for the observed shape of the SNR spectrum at X-ray energies (such as the Non Equilibrium Ionization, or NEI, models of Hughes and Helfand ( or of Hamiltonet al. (; unlike in the radio, where polarization information is available, there is no independent way of confirming that the radiation is definitely synchrotron; the X-ray synchrotron spectrum was not just an extension of the radio synchrotron spectrum. The technique of looking for X-ray synchrotron radiation is actually somewhat amusing since people have bashed away, unsuccessfully, at the problem for years by looking for gamma-ray emission near SNR.

Synchrotron radiation in SNRs was thought by many to be important only in the radio part of the electromagnetic spectrum. Other processes such as thermal emission from hot gas, nonthermal bremsstrahlung, or inverse Compton scattering were what astronomers usually thought of when they thought of what processes create high energy X-ray photons. The problem was, the spectra for the X-rays produced by these processes looked nothing like the observed emission from SNR unless some very ad hoc assumptions were made about the ionization and abundances of the SNR gas! The shape of the distribution was wrong and the energies were way off. Obviously these high energy photons were coming from some other process occurring in the remnant. That process is synchrotron emission. That X-rays could be produced by synchrotron emission by cosmic rays came as a surprise to many people although Steve Reynolds and Roger Chevalier ( suggested the idea in 1981 to explain the spectrum of SN 1006.


SNR slide thumbnail When ASCA observed SN1006 in 1995, something unexpected was found. Key were ASCA's observations of the center of the remnant, separate from the X-ray bright rims. When the emission from the lobes was omitted, what remained was a thermal spectrum, dominated by X-ray lines. This result conflicted with the NEI model of Hughes and Helfand, but was in agreement with Reynold and Chevalier's synchrotron model. In previous observations, the flat synchrotron emission in the rims overwhelmed the thermal emission in the center, and so the different models seemed to be equally valid. As a result of these observations, Koyama et al. published a paper about SN 1006 in which they showed very strong evidence of synchrotron emission because they could produce separate spectra for the two lobes of the shell and for the interior of the remnant.

If the strength of the magnetic field is known, the energy of the ions that are responsible for the synchrotron emission can be calculated from the X-ray synchrotron energies. Given some uncertainty about the strength of the magnetic field in SN 1006, it appears to accelerate electrons (and presumably protons) to energies ~ 1014 eV. This is the right amount of energy to explain the cosmic-ray power spectrum up to the knee. Lower mass particles such as electrons would have up to 1014eV of energy while heavier particles such as iron, at the same velocities, would be up to 3 X 1015 eV (the knee). The shape of a synchrotron spectrum produced by a population of Fermi accelerated particles matches observations, and furthermore, an estimate of the total amount of energy in the accelerated particles (given some more uncertainty in some of the assumptions) is close to what you expect if all supernova remnants do the same thing and if they produce all of the Galactic cosmic rays.

Cosmic Rays and SNR - Making the Connection

  • The X-ray spectrum of SNR is not fit by emission from processes such as thermal emission, nonthermal Bremsstrahlung or inverse Compton scattering

  • Thermal emission in the center of the remnant (discovered by ASCA) rules out NEI emission by hot gas as a possible source

  • With other theories ruled out, the shape of the spectrum tells us that the strong X-ray emission in the rims must be from synchrotron emission by accelerated, charged particles

  • From the energy of the X-rays and an estimate of the strength of the magnetic field we can calculate the energy of the accelerated particles

  • Energies correspond to observed cosmic ray energies (at Earth) up to the "knee"
  • Conclusion Cosmic rays up to 1015 are accelerated in the shocks of SNR and become visible in the X-rays when they emit synchrotron radiation

Why Do We Want to Know?

In general anything which is both extreme and plentiful is fascinating. Cosmic rays are found in such large numbers everywhere in our galaxy and they can be incredibly energetic. We are not able to generate anything anywhere nearly as energetic, so it becomes an interesting question to explore what physical processes CAN generate them.

Do cosmic rays affect our everyday lives, or are they too remote to worry about? Although you may never yourself be bombarded by a primary cosmic ray (we are shielded from them by the Earth's magnetosphere), we are bombarded all the time by the secondary cascades of cosmic rays that are created when cosmic rays interact with Earth's atmosphere. These secondary cosmic rays are not as energetic as the primary CRs which exist in space, but are responsible nonetheless for a constant background radiation to which we are all constantly exposed. Cosmic rays produce Carbon 14, a small source of radiation but one which is critical for dating (for example establishing the age of fossils).

Spacecraft and high altitude planes certainly feel their effects. With their high energy concentrated in such a small bundle, cosmic rays can disrupt computer hardware or sensitive electronics and these instruments have to be shielded in vehicles traveling above the atmosphere. If a cosmic ray passes though a sensitive part of a semiconductor chip, for example, the logical state of the bit ("on" or "off") can be flipped. This is called a single-event upset (SEU). A single-event upset can also result from a cosmic ray hitting the nucleus of an atom in a sensitive component location. The nuclear interaction can cause the nucleus to split, or spallate. The broken pieces of the nucleus then carry away most of the cosmic ray's energy. These bits of debris can then flip the bit state.

This problem is most commonly seen in the South Atlantic Anomaly ( The distribution of errors ( from the UoSAT-3 spacecraft in a polar orbit can be seen at the European Space Agency's web page. The errors at high latitudes are primarily caused by cosmic rays striking the spacecraft. Similarly, cosmic rays also corrupt observations of space made with CCDs (Charged Coupled Devices, a kind of digital telescope). The cosmic rays have to be subtracted from the data.

To read more about Galactic cosmic rays and cosmic rays that are generated from gamma-rays in the Earth's atmosphere, check out "Imagine the Universe!".

Observational Clues: Current and Future Work

One technique to to search for X-ray synchrotron radiation is "spatially-resolved spectroscopy" with ASCA ( and with the Chandra X-ray Observatory (CXO) ( CXO was launched in July 1999 and will provide even better resolution and more coverage. Right now it is not clear how many remnants this will work for. One other remnant, RX JJ1713.7-3946, appears to be similar to SN 1006, but most other remnants appear to be dominated by thermal emission from their shells. In order to pull out the synchrotron emission, it is necessary to have a broad-band X-ray spectrum that extends to many tens of keV because the synchrotron radiation is expected to dominate above about 10 keV. This latter technique has been used to show the presence of synchrotron emission for Cas A, using RXTE (and SAX).

IC 443, the only other remnant with any evidence of synchrotron radiation that is known now, is sort of a mixture of the two techniques in the sense that a small part of the remnant appears to emit synchrotron radiation, but the evidence lies in the broad-band spectrum. That is, the synchrotron emission does not dominate the spectrum.

We've only just begun looking for X-ray synchrotron radiation. It will be interesting to discover if all SNRs produce X-ray synchrotron radiation and what fraction of Galactic cosmic rays are produced in SNRs.

Thank you to Glenn Allen for contributing to this article.

Imagine the Universe is a service of the High Energy Astrophysics Science Archive Research Center (HEASARC), Dr. Alan Smale (Director), within the Astrophysics Science Division (ASD) at NASA's Goddard Space Flight Center.

The Imagine Team
Acting Project Leader: Dr. Barbara Mattson
All material on this site has been created and updated between 1997-2012.

DVD Table of Contents
Educator's Index