Tuesday, December 04, 2007

Skymap of AGNs with Cosmic Ray Events

Skymap with Cosmic Ray Events and AGNs
[Picture: AUGER Collaboration, click to enlarge]

The above plot by the AUGER collaboration shows the celestial sphere in galactic coordinates. This is essentially a map of the sky much like a map of the earth, where the equator is the Milky Way's galactic plane and the sun is in the center. The black circles with a radius of 3.1° show the arrival directions of the 27 highest energy cosmic rays detected by AUGER with energies greater than 57 x 1018 eV. The red stars show the positions of 472 active galactic nuclei (AGN) within 75 megaparsecs (Mpc) distance to the earth. The blue region defines the field of view of Auger; deeper blue indicates larger exposure, and thus more expected events. The solid curve marks the boundary of AUGER's field of view.

The white * is Centaurus A, the closest AGN. Two of the 27 cosmic rays have arrival directions within 3° of this galaxy.

The data analysis depicted here has been fairly new, and we reported on it in October. The AUGER Collaboration finds correlations between the events of highest energies and AGNs and they are able to reject the hypothesis of an isotropic distribution of these cosmic rays at a confidence level of 99%. This is interesting for two reasons.

First, it is the first time that the sources of these cosmic ray events could be shown to be correlated with the AGN. This reliably rules out speculations about the origin of these UHECRs in local, galactic sources. Though it has been expected, until now there was no experimental confirmation that they originate outside our galaxy. Though one should note that this correlation does not necessarily mean the AGNs themselves are the sources, as the sources could just also be correlated with the AGNs.

Second, this correlation vanishes if one includes AGNs further away than ~90 Mpc, which is what one would expect from the GZK cutoff: at this high energy, a proton's mean free path is below ~ 90 Mpc because the protons will scatter at the CMB background and form pions. The vanishing correlation is thus an independent confirmation for the presence of the cut-off.

For me this is one of the most important experimental results of the year.

This post is part of our 2007 advent calendar A Plottl A Day.


  1. So, any comment from you on the latest resonaances posting


    respectively the corresponding paper
    http://arxiv.org/abs/0711.4060 ?

    Thnx for your nice blog btw

  2. Hi Mascotte:

    if I recall that correctly the paper by the AUGER people already pointed out that it is funny there are too little events from the Virgo Cluster. Given that there are only 27 events in toto I actually find it somewhat premature to nitpick on them. Most of the questions and objections can probably be sorted out with better statistics. It seems to me the paper you mention assumes the cosmic rays would be emitted by the source isotropically into all directions (thus the 1/r^2) suppression. I don't know much about the astrophysical processes people envision would cause these events, is this a plausible expectation? Best,


  3. Hi Bee,

    I have to concur - this result is much better than its next competitor for best of the year, which is obviously the paper I am releasing next week on b-jet energy scale calibration ;-)

    One thing I would like to see is the same plot for the next-highest set of 27 CR, and a corresponding statistical measure.


  4. I was wondering what is known about the physics of AGNs. In particular, is anything known about them that would suggest that they could could accelerate protons?

  5. about the GZK limit...why is the limit imposed by pion pair production as opposed to electron pair production? the energies required for electron pair production is /significantly/ less than pion pair production. i know i'm missing something, but i don't know what it is.

    rgb: intense magnetic fields associated with the accretion disks of the black hole powering the individual AGN are strong enough to do this. plasma jetting is not exactly an unknown phenomena, even in earth orbit.

  6. Hi rgb:

    Excellent question, but sadly one that I can't satisfactory answer. I am not much of an astrophysicist. I will keep it in mind though and maybe have a post about it at some point!

    Hi Eric:

    Electron pair production from a proton and a photon? How would that work with charge conservation? The reactions the proton makes is either p + gamma -> p + neutral pion or -> neutron + charged pion.


  7. Dear Bee, Eric,

    eh, I'm just a bit confused - I guess what Eric is asking about is any process which involves the production of a lepton-antilepton pair, it would conserve charge:

    γ + p -> γ* + p + l + l-bar,

    or something like that, and l could be any lepton, even a neutrino. Just from energy, such processes should be possible at a much lower energy than photopion production.

    For some reason, the cross section for these processes probably is quite small? Photopion production really goes via the photoexcitation of the Δ resonance and subsequent strong decay, where the cross-section is high.

    Best, Stefan

  8. Here's a couple links about AGNs, pion/electron-positron production is briefly mentioned in the 2nd link:


  9. Hi Stefan, what is the role of the proton in that reaction? I mean the gamma hits a quark and then what is the rest of the interaction? Best, B.

  10. Down at the energy levels where electron-positron pairs are produced, I think the cosmic rays involved are just photons, right?

  11. The Virgo cluster can be seen at the top of the Pierre Auger plot slightly to the right of centre. Lots of AGN, and not one single cosmic ray. Zero.

    As for the small total number of events, surely this is an equally strong factor against PAO's claim as well as against Gorbunov et al's criticism. In other words: if 27 events is too small to criticise the hypothesis seriously ('nitpick', in Bee-speak) it is also much too small to confirm it.

    The Gorbunov paper applies a statistical test and finds that with 6 events expected from the Virgo direction, zero actually arrive, with probability 10^-3 under the AGN hypothesis.

    Assuming, of course, that the Virgo cluster doesn't have some weird property that prevents AGNs located there from emitting anything in our direction.

    Since the Virgo cluster consists of order 10^3 galaxies, it would be rather strange if we failed to hear from any of them. Almost as if they conspired to hide something from us. Statistically, such a large number of galaxies should emit fairly isotropically.

    Gorbunov mention an alternative hypothesis: that most of the clustering is due to an object in the direction of Centaurus A (centre right of the plot) with a few stray events around due to strong deflections.

    The good news is that 1) there *is* clustering; and 2) the controversy on where the clustering comes from will require a good deal more data to resolve.

  12. Hi Thomas:

    if 27 events is too small to criticise the hypothesis seriously ('nitpick', in Bee-speak) it is also much too small to confirm it.

    I agree.



COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

Note: Only a member of this blog may post a comment.