Cosmic rays have been observed with energies from 109 eV to over 1020 eV. Over this range, the "flux" of cosmic rays (the number of arriving particles per unit area, per solid angle, per unit time) appears to follow a single power law ~E-3. The variation of the flux with energy is referred to as the "Energy Spectrum". This spectrum is shown in the figure to the right. Remarkably, it appears to be a smooth curve over 10 decades of energy with only a few noticeable structures. The most prominent of these are small, abrupt changes in the local spectral index (the power exponent ~3) just above 1015 eV and again just above 1018 eV. Within the Cosmic Ray field, these features are usually referred to as the "knee" and the "ankle", respectively.
Cosmic Rays with energies above ~1018 eV are referred to as "Ultra-High Energy Cosmic Rays" (UHECR). These are microscopic particles with a macroscopic amount of energy-about a joule (for comparison: an electron-volt is about 1.6x10-19 joules) or more. The existence of such energetic particles remains a mystery. The three main questions concerning UHECR's are:
To study the acceleration mechanism, one must make careful measurements of the energy spectrum of UHE cosmic rays to compare to the predictions from different acceleration models. To understand where the UHE cosmic rays come from, one needs to make a careful survey of the arrival directions, and search for both small- and large-scale anisotropies in their distribution.
Composition one of the most difficult measurements because UHE cosmic rays cannot be detected directly using conventional particle detectors (see Detection of Ultra-High Energy Cosmic Ray). Consequently, the composition must be inferred from auxiliary measurements. These are discussed in the other sections.
One of the earliest theories on the acceleration of cosmic rays was proposed by Enrico Fermi in 1949 [1]. It became known as the "Second Order Fermi Mechanism". In this model, particles collide stochastically with magnetic clouds in the interstellar medium. Those particles involved in head-on collisions will gain energy (similar to a sling-shot process used to accelerate spacecrafts around planets), and those involved in tail-end collisions will lose energy. On average, however, head-on collisions are more probable. In this way, particles gain energy over many collisions.
This mechanism naturally predict a power law energy spectrum, but the power index depends on the local details of the model and would not give rise to a universal power law for cosmic rays arriving from all directions. This mechanism is also too slow and too inefficient to account for the observed UHE cosmic rays.
A more efficient version of Fermi Acceleration was proposed independently by a number of workers in the late 1970's [2-5]. In this model, particles are accelerated by a strong shock propagating through interstellar space. The following gives a schematic of the process as described in Prof. Longhair's book [6]:
Consider the case of a strong shock propagating at a supersonic, but non-relativistic speed U through a stationary interstellar gas. Figure (a) at left shows the situation in the rest frame of the gas: the density, pressure, and temperature of the gas upstream and downstream of the shock front are r2, p2, T2 and r1, p1, T1, respectively.
When viewed in the rest frame of the shock front as in figure (b) below, particles are arriving from downstream with speed v1=U and exiting upstream at speed v2. Conservation of the number of particles implies the relation: r1v1=r2v2. In the case of strong shock we expect r2/r1=(g+1)/(g-1), where g is the usual ratio of heat capacities. For a fully ionized plasma, one expects g=5/3, leading to a velocity ratio of v1/v2=4.
First order Fermi acceleration naturally predicts a power law spectrum of DNA(E)/dE ~ E-2. While the power index of 2 does not agree with the measured index of ~3, this model predicts, for the first time, a power law spectrum with a unique spectral index that is independent of the details of the local environment. The mechanism requires only the presence of strong shocks, which are quite plausibly present in the suspected sources of cosmic rays.
The leading candidates for the source of UHE cosmic rays are large, energetic structures where strong shocks are expected to be found. The most well known of these are supernova remnants, which have long been suspected to generate cosmic rays. In 1995, Japan's ASCA X-ray Satellite, reported positive observation non-thermal X-ray emissions from the Supernova Remnant SN1006. The observed emission spectrum is consistent with synchrotron emission by accelerated charged particles. This report is widely seen as confirmation of supernova remnants as a known source of cosmic rays.
The observed emission from SN1006, with some fine tuning of the emission models, can explain the existence of cosmic rays up to ~1015 eV. However, it is difficult to explain the existence of cosmic rays above 1018 eV, because supernovae are simply not large enough to maintain acceleration to the UHE regime. Furthermore, no positive correlation has been observed between the arrival directions of UHE cosmic rays and supernova remnants.
There are many larger objects in the sky where strong shocks are expected. For example, strong shocks are possible around colliding galaxies such as NGC 4038/9. However, there is no evidence to indicate these objects are sources of UHE cosmic rays.
Another class of objects which are candidate sources of UHE cosmic rays are active galactic nuclei (AGN). AGN is the generic name given to a class of galaxies which are suspected to have at their center a super massive black-holes. AGN's are typically accompanied by jets which can extend 50-100 thousand light-years. Roughly one of every ten known galaxy is an AGN. It is therefore always possible to find an AGN within error of the arrival direction of a UHE cosmic ray. Even so, there is still no evidence to indicate that AGN's actually produce UHE cosmic rays.
Other ideas for explaining the existence of UHE cosmic rays include: