Axions and the Problem of EurekAlert

I have a couple of EurekAlert feeds in my RSS reader, because they sometimes turn up interesting things– I got the Bill Wootters item there, for example, and they had a piece on strontium clocks that I keep meaning to say something about.

Of course, there’s also some total garbage, such as the kookery from the “Quantum Aether Dynamics Institute” that crossed the feed yesterday (though it appears to have been taken down, to their credit). This makes it difficult to really trust anything I see there that claims to be a really new development.

Such as, say, this press release from Buffalo claiming the discovery of axions:

After decades of intensive effort by both experimental and theoretical physicists worldwide, a tiny particle with no charge, a very low mass and a lifetime much shorter than a nanosecond, dubbed the “axion,” has now been detected by the University at Buffalo physicist who first suggested its existence in a little-read paper as early as 1974.

The finding caps nearly three decades of research both by Piyare Jain, Ph.D., UB professor emeritus in the Department of Physics and lead investigator on the research, who works independently — an anomaly in the field — and by large groups of well-funded physicists who have, for three decades, unsuccessfully sought the recreation and detection of axions in the laboratory, using high-energy particle accelerators.

The paper, available online in the British Journal of Physics G: Nuclear and Particle Physics at http://www.iop.org/EJ/abstract/0954-3899/34/1/009, will be published in the January 2007 issue.

Results first were presented during a two-day symposium held in October at UB that celebrated Jain’s 50-year career in the physics department in the College of Arts and Sciences.

During that symposium, the world-renowned and Nobel Prize-winning scientists in attendance expressed astonishment and delight that the axion finally might have been found.

My inclination is to think that this is a bunch of crap. There are a number of kook signifiers in the text of the release, and if there was really anything to this, I’d expect to find it published somewhere with a higher profile than J. Phys. G. And, as we’ve previously established, EuerkAlert doesn’t exactly have a sterling peer-review system…

There’s a cautionary message here about doing science by press release. This release looks fairly convincing, if you’re coming from outside the field, but then, the Aether Dynamics thing looked pretty slick, other than being batshit crazy. There’s really not a good way to judge the credibility of anything you find on this site, because anyone with a bit of writing skill can write a convincing-sounding press release. But it’s out there as a science reference, and a lot of people pull news items off it.

Comments

  1. #1 Derek Lowe
    December 7, 2006

    Oh, dear. You’re right, this kind of thing makes a number of red and amber lights come on. “Tiny effects way down there in the noise that only I can see”, eh? It would be interesting to do a citation search of this guy’s papers to see how much of his stuff has ever been followed up on

  2. #2 Ben M
    December 7, 2006

    Ugh. The general technique is perfectly valid—look for a particular final state (invisible particle goes to e+e-), plot the invariant mass distribution, look for peaks. Jain does this, and plots the results. He uses Excel, and apparently isn’t very good at it; everything is plotted as “% yield” rather than “number of events”, which obscures things rather badly; he shows nothing but bar graphs, and always with the default labels (“series 1″, “series 2″) rather than readable ones (“signal”, “simulated background”).

    He sees two unresolved, narrow peaks in the final data (a subset of the “signal”, minus what appears to be an analytically-calculated pion background). That’s it: he announces that the peaks are each “three sigma above average” and calls it a discovery.
    There’s no additional data analysis whatsoever: no simulation or analysis of other backgrounds; no calibration of the detector and its energy resolution; no cross-checking of the energy measurement.

    I do want to emphasize that there’s nothing wrong with the idea, or the technique. It’s a standard route to particle discovery—pick a decay product, search for examples of decay in detector, plot invariant masses, look for new components in plot. The hard part is understanding all of the bumps you inevitably see. I can tell a story about one cosmic-ray detector I worked on. It consisted of several planes of silicon position-sensitive detectors in a magnetic field. Cosmic rays fly through it, you measure five or six position points. Since the field was non-uniform, you needed a fairly complicated computer simulation to fit a curve through those five or six points and figure out the particle’s momentum. Of course, sometimes the cosmic ray would scatter off of a nucleus in the silicon, or the detector’s electronic would spit out a false hit; such data wouldn’t be a very good match to the fit model, which computed ordinary F = v x B trajectories. The fitter was supposed to report back when such a bad match was found, and it usually did. When it didn’t, as we found the hard way, the (utterly nonsensical) momentum tended to be reported as about 50 GeV/c. So when we plotted our data (number of particles versus momentum), there was this small but statistically-significant peak at 50 GeV. Took a long time to debug. If Prof. Jain analyzes his data carefully, he *will* find similar effects; one always does. If his peaks are still there after more-careful analysis, physicists will pay more attention.

  3. #3 Sean Carroll
    December 7, 2006

    Given that the press release glowingly describes Prof. Jain’s efforts at detecting this short-lived particle, but the actual axion is an exceptionally long-lived particle, I wouldn’t take it too seriously.

  4. #4 John Novak
    December 7, 2006

    I actually have a bunch of Eureka RSS feeds, too, for roughly the same reason. And I saw the QAD thing and rolled my eyes.

    I wasn’t ever under the impression that they were a peer-review type of place, though. I thought it was fairly explicit that their filter against noise was publication costs, on the grounds that most wingnuts aren’t going to pay whatever they charge for 15 seconds of fame.

    (Not, for instance, that this ever stopped “The Center For Responsible Nanotechnology.”)

  5. #5 Mike Lemonick
    December 7, 2006

    As a science writer (but not a scientist), even I saw red flags; for me they came with the certainty of the assertion that the axion search was over–end of story.

  6. #6 Count Iblis
    December 9, 2006

    It’s not just the “axion interpretaton” that is problematic here. Just think about it. You have a light particle of about 7 MeV or so that can be produced in collisions of nuclei and that can decay very fast into an electron positron pair. This means that this particle has strong interactions with both electrons and quarks.

    The field associated with the particle would thus give rise to a new force that affects electrons in atoms. The mass of 7 MeV implies thst the range of the force is 10^(-13) meters. This means that it can only affect the elctron states for which the wavefunction does not vanish close to the nucleus. So, you would expect a small energy shift in the 1s states of hydrogen, while the 2p states should be unaffected.

    But the energy difference of the 1s and 2p states have been experimentally determined to an enourmous precision and it perfectly agrees with the QED prediction.

  7. #7 Autymn D. C.
    January 24, 2007

    Ye are full of shit.

    Do you even know what the J. Phys. G is? Hint: It’s a new merger. I got my tome in the mail yesterday.

    You also forgot this:

    “The story of the search for the axion particle in high-energy physics — not to be confused with the search by cosmologists and astrophysicists for axions produced by the sun…”

    -Aut

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.