Science and Reason: Ultra-high energy cosmic rays.

I'm just trying to catch up on interesting science news of the past few weeks, and what happens? I pick a topic that seems to have had significant new developments and is also straightforward enough for a relatively brief post. But instead of something I can do justice to with just a handful of links to good reports (so I don't need to do that much work), I find myself in a brier patch. Frustrating.

The general topics is cosmic rays. The phenomenon of radioactivity was discovered in 1896 by Henri Becquerel. At first the only known source of radioactivity was certain naturally occurring radioactive elements in the Earth itself. But in 1912 Victor Hess deduced that there must be another source of radioactivity outside the Earth's atmosphere. These came to be known as cosmic rays (as opposed to other kinds of "rays", such as X-rays, alpha rays, beta rays, and gamma rays, which have known terrestrial sources). Hess received a Nobel Prize in physics in 1936 for his discovery, indicating that by then scientists generally agreed the phenomenon was genuine, and interesting.

Given that, the next questions had to be: what do cosmic rays consist of, and how are they produced? Even today we are still in the process of answering those two questions, though we think the answers are pretty well known for all but the most energetic cosmic rays.

I'll discuss cases involving different energies in a moment, but first there are some basic distinctions. Now that we have spacecraft, we can in principle observe cosmic rays directly, though this hasn't actually been done systematically. (It turns out that the highest energy cosmic rays, which are still the least well understood, are too rare to be observable with non-terrestrial instruments of sufficient size.) At the Earth's surface we can't observe cosmic rays directly. Assuming cosmic rays are ordinary particles like electrons, protons, or atomic nuclei (and we've seen nothing to indicate otherwise), all we can see are showers of electrons and more exotic particles like muons. All the evidence indicates that these showers are the result of collisions between actual cosmic rays and molecules in the atmosphere.

There is an additional distinction that can be made between "primary" and "secondary" cosmic rays. Originally this distinction was between the particles we actually observe in showers ("secondary") and the ones ("primary") that cause the showers. The latter, of course, are the ones that need to be understood. But we now realize that the particles that enter our atmosphere may themselves be a result of collision between the "real" primary cosmic rays and other particles floating in interstellar space. Since we're concerned with what the particles are in their original form, it is those that are now usually called primary (or simply "cosmic rays" without further qualifiers), while everything else is secondary.

Obviously, the most important step in understanding (primary) cosmic rays is to figure out where they come from, and for that we would like to know the direction from which they enter our atmosphere. We know that some relatively low energy cosmic rays originate in the Sun ("solar cosmic rays"). Unfortunately, except for those and all but the most energetic other cosmic rays it is impossible to tell the direction they have come from. That's because the particles that could make up cosmic rays should have certain properties. Except for solar cosmic rays, the particles should be stable and not decay over time spans of (at least) many years. We can also rule out very light particles like neutrinos. Cosmic rays could be something totally exotic that's never been observed before, except in that case it's very likely they would not interact with ordinary matter – so they would not be capable of producing showers of secondary particles that we observe. Among particles currently known, then, that leaves only charged particles such as electrons, protons, and atomic nuclei.

However, unless a charged particle has extremely high energy, in excess of what has been observed for most (but not quite all) cosmic rays, its motion will be affected by our galaxy's magnetic fields, or even by the magnetic field of the Earth. We observe that almost all cosmic rays energetic enough not to be significantly affected by Earth's magnetic field appear to come from directions isotropically distributed over the sky. So we can't identify any particular spot from which most cosmic rays appear to originate, because our galaxy's magnetic fields have randomized their directions.

Very recent results strongly suggest that a few very rare cosmic rays ("ultra-high-energy cosmic rays", or UHECRs) do come from directions we can identify. We'll get to that shortly.

But first we need to talk about what sort of energies are actually involved. The most energetic UHECR ever observed had an energy of about 3×10²⁰ eV (electron-volts). That is roughly the kinetic energy in a baseball moving at 96 km/hr. (I don't know about you, but I can't easily relate to energies expressed in terms of moving baseballs, tennis balls, or large hailstones. To me it's more meaningful to talk in terms of the energies that can be produced in the largest contemporary particle accelerators, about 10¹³ eV, or 10 TeV. That's more than 7 orders of magnitude, a factor of 10 million, less than the energy of some UHECRs.)

We cannot easily imagine mechanisms in "ordinary" objects (no more exotic than, say, a supernova remnant) that could on a sustained basis churn out charged particles with energies more than about 10¹⁸ eV. Up to that energy level we can envision mechanisms involving shock-wave acceleration in supernova shells. Beyond that, we would need something like some sort of quasar or active galactic nucleus (AGN). The good news is that there are very recent results relevant to both cases – cosmic rays with energies ≤ 10¹⁸ eV, as well as UHECRs.

Let's consider the lower energy case first. Here we are concerned primarily with cosmic rays originating in our own galaxy, so-called galactic cosmic rays. Many of these, of course, come from the sun, or other stars, or other equally common objects. So the ones of real interest are those having energies that call for much less common origins, yet short of things we don't have in our galaxy, such as AGNs.

The question here is whether there actually exist supernova remnants (for example) in which we can actually observe something going on that has enough energy to account for the most energetic galactic cosmic rays.

And the answer, now, is yes, we have observed such things. Just about a year ago it was reported that the Chandra X-ray Observatory had determined that the Cassiopeia A supernova remnant was accelerating electrons enough to account for all but the most energetic galactic cosmic rays. (We offered a Hubble optical image of Cassiopeia A here, and a a false color infrared image produced by the Spitzer Space Telescope here. There's an even more dramatic Spitzer image available here, and a false color image from Chandra itself here.) Here's the relevant press release:

Chandra Discovers Relativistic Pinball Machine

For the first time, astronomers have mapped the rate of acceleration of cosmic ray electrons in a supernova remnant. The new map shows that the electrons are being accelerated at close to the theoretically maximum rate. This discovery provides compelling evidence that supernova remnants are key sites for energizing charged particles.

We had a post about this here, which lists several other accounts of the discovery. (There's also news in that post about a possible extragalactic cosmic ray source.) Note that this does not say cosmic rays have actually been observed to come from Cassiopeia A – galactic magnetic fields make it unlikely to be able to prove cosmic rays come from very close to that direction. All we can say is that Cassiopeia A should be producing high-energy cosmic rays.

However, Cassiopeia A doesn't seem to be a source of the highest energy (galactic) cosmic rays we can envision coming from a supernova remnant. Fortunately, a much more recent result does provide such an example.

NASA: Major Step Toward Knowing Origin of Cosmic Rays

Since the 1960s scientists have pointed to supernova remnants -- the tattered, gaseous remains of supernovae -- as the breeding ground of most cosmic rays. These remnants expand into the surrounding interstellar gas, an energetic interaction that produces a shock front containing magnetic fields that can accelerate charged particles to enormous energies, producing cosmic rays.

According to theory, charged subatomic particles bounce like pinballs around the shock front. They pick up speed until they move nearly the speed of light. Last year, observations from NASA’s Chandra X-ray Observatory suggested that electrons are being accelerated rapidly (as fast as theory allows) to high energies in the supernova remnant Cassiopeia A.

Now, Yasunobu Uchiyama of the Japan Aerospace Exploration Agency (JAXA), and four colleagues, have observed the signature of the shock acceleration of electrons, and demonstrated that magnetic fields in supernova remnants are stronger than previously thought, and are thus fully capable of producing cosmic rays.

In a study published in the October 4, 2007, issue of the journal Nature, Uchiyama’s team used Chandra and JAXA’s Suzaku X-ray satellite to look at the northwest edge of supernova remnant RXJ1713.7-3946, located a few thousand light-years from Earth in the constellation Scorpius.

Up until this result, the problem has been a doubt that magnetic fields in supernova remnants are strong enough to accelerate particles to an energy of around 10¹⁸ eV. This doubt has been laid to rest by the observations of RXJ1713.7-3946. There it has been possible to estimate the strength of its magnetic fields.

The estimation is accomplished by Chandra observations of X-ray hot spots in the remnant. The hot spots represent synchrotron radiation given off by electrons accelerated to the highest velocities. Shock waves in the magnetic field produce the acceleration, and their velocity can be estimated at about 10 million km/hr. Such waves should give electrons a kinetic energy on the order of 10¹⁵ eV.

That's still not enough energy to account for the most energetic galactic cosmic rays, but protons (or heavier particles) accelerated to similar velocities could do the trick. Protons have the same charge as an electron (but of opposite sign). Since a proton has about 1836 times the mass of an electron, the kinetic energy of accelerated protons could reach 10¹⁸ eV, corresponding to the most energetic galactic cosmic rays. (A helium nucleus, with about 4 times the mass of a proton, could have even higher energy, of course.)

Synchrotron radiation from protons of this energy would be in the gamma-ray range, so they could not be observed by Chandra. However, there are suggestions that gamma-ray observations do confirm that the magnetic fields of RX J1713.7-3946 are strong enough to produce 10¹⁸ eV protons.

Again, as with Cassiopeia A, all this isn't saying we've observed 10¹⁸ eV cosmic rays produced by RX J1713.7-3946, only that it has magnetic fields strong enough to do the job.

Other reports of this research: here, here.

We're still left with the problem of explaining UHECRs. Since we can't imagine anything inside our own galaxy that could be energetic enough to produce UHECRs without being directly observable (and certainly we don't observe any such thing), the source must be outside the galaxy.

As the following recent note observes, there are a few possibilities.

Magnetic cocoons power energetic cosmic rays

[U]ltra-high-energy cosmic rays (UHECRs) – each packing the punch of a baseball – are an outstanding mystery. Although it is conceivable that they are produced near the Milky Way by the decay of super-heavy dark matter particles or by defects in space-time, the most likely sources are the most powerful objects in the universe – 'active' galaxies whose colossal black holes are devouring nearby matter, and gamma-ray bursts. These are far beyond our galaxy – and herein lies a very serious problem.

The problem is what is called the Greisen-Zatsepin-Kuzmin (GZK) limit. This limit results from the fact that protons (or heavier atomic nuclei) that have an energy more than about 6×10¹⁹ eV will interact with photons of the cosmic microwave background (CMB). The interaction destroys the cosmic ray particle and produces short-lived pions, which decay long before reaching Earth. For a UHECR the probability of an interaction is proportional to the distance between the Earth and the cosmic ray source.

The reason this occurs is that cosmic rays with this energy are moving at very close to the speed of light. From the point of view of the cosmic ray itself, a microwave photon is drastically blue-shifted, even though to an observer on Earth, a CMB photon has quite low energy. But for a cosmic ray moving fast enough, the CMB photon appears to be very energetic, and the threshold above which a pion-producing interaction can occur is for cosmic rays with energies above 4×10¹⁹ eV.

This diagram show the expected flux of cosmic rays as a function of their energy, assuming a power-law distribution. This is in fact what is observed. For example, cosmic rays with an energy of more than 10¹⁹ eV are observed at a rate of only 1 per km² (of Earth surface) per year. So you need to observe over an area of many km² simply to tally a reasonable number of the most energetic cosmic rays in a few years.

Because of the GZK effect, the flux of cosmic rays takes a sharp drop below the curve in the diagram right around 4×10¹⁹ eV. We still observe a very few cosmic rays with energies above that point, because if the cosmic ray originates from a source sufficiently close to Earth, there is still some nonzero probability it will not collide with a CMB photon. However, there is essentially no chance we will see a cosmic ray of more than 4×10¹⁹ eV if its souce is more than about 250 million light-years (Mly) from Earth.

As it turns out, this limit is actually a stroke of good fortune, because it means that in order to identify sources of UHECRs we need only consider objects within an Earth-centered sphere of radius 250 Mly. We can simply disregard objects that are farther away.

Another simplifying condition is that we don't need to bother with cosmic rays having an energy less than about 3×10¹⁹ eV, because the trajectories of such lower-energy cosmic rays are likely to have been deflected by our galaxy's magnetic field. We couldn't hope to identify the direction from which they came anyway. This is a useful simplification, since there is a larger variety of possible sources for the lower energy cosmic rays, and we would simply have too many possibilities to have a hope of correlating the observed directions with locations of particular source types. Furthermore, atomic nuclei larger than one proton will have more than one unit of charge, so they will also be deflected by the galaxy's magnetic field. And so we would not expect cosmic rays that don't consist of individual protons to have a correlation with the location of specific source objects.

Given all that, it's easy to understand and appreciate the result that was just announced early in November. The result comes from a large team working at the Pierre Auger Observatory in Argentina. The result, in a nutshell, is that out of 15 cosmic ray events with energy more than 6×10¹⁹ eV counted since 2004, 12 came from a direction that was within 3.1° of a known active galactic nucleus (AGN) within 250 Mly of Earth. It is calculated that the probability of this correlation occurring by chance is only about 1 in 1000 if the flux were isotropic.

What about the other 3 (of 15) events? They might have been cosmic rays that consisted of atomic nuclei heavier than a proton, and hence were deflected by galactic magnetic fields. However, these anomalous events were observed near the galactic plane, so their source could have been an AGN we cannot see because of dust in the galactic plane.

It should be mentioned that Auger is nothing like a typical optical observatory. Instead, it consists of a large number of instruments spread over an area of 3000 km², which is about the size of Rhode Island. Yet since observations began in 2004, only about 80 events with energy more than 4×10¹⁹ eV were tallied, and there were only 27 with energy more than 5.7×10¹⁹ eV (a rate of about 1 per 4 km² per century), so you can see how rapidly the numbers drop off in this range. Small wonder that cosmic ray astronomers now want to have an even larger observatory built, as well as to continue observations over enough years to reduce the possibility that correlations are due to chance.

Even so, the present results are generally considered to be pretty important – enough, anyway, to have become the cover story of the November 9 issue of Science. And this is even though the result is "only" a statistical correlation between directions of events and known AGNs. It's much too soon to say something like N events have been observed very close to a specific AGN. Also still left open is the construction of a model for exactly how UHECRs are produced in an AGN, let alone the validation of the model in a particular case (as has been done with lower energy cosmic rays and supernova remnants).

It's also only fair to note that objections to the Auger conclusions have been raised – see here, and in some of the references below.

However, we now have a lot better information about UHECRs than we had before.

More information: