• John
  • Felde
  • University of Maryland
  • USA

Latest Posts

  • James
  • Doherty
  • Open University
  • United Kingdom

Latest Posts

  • CERN
  • Geneva
  • Switzerland

Latest Posts

  • Aidan
  • Randle-Conde
  • Université Libre de Bruxelles
  • Belgium

Latest Posts

  • Vancouver, BC
  • Canada

Latest Posts

  • Laura
  • Gladstone
  • MIT
  • USA

Latest Posts

  • Steven
  • Goldfarb
  • University of Michigan

Latest Posts

  • Fermilab
  • Batavia, IL
  • USA

Latest Posts

  • Seth
  • Zenz
  • Imperial College London
  • UK

Latest Posts

  • Nhan
  • Tran
  • Fermilab
  • USA

Latest Posts

  • Alex
  • Millar
  • University of Melbourne
  • Australia

Latest Posts

  • Ken
  • Bloom
  • USA

Latest Posts

CERN | Geneva | Switzerland

View Blog | Read Bio

Dark matter retains all its mystery

Monday morning, at the Moriond conference, the most expected talk in the dark matter session contained unfortunately no results. Although the AMS collaboration was supposed to reveal their very first measurements, Bruna Bertucci could only present apologies to an eager audience since the approval process had not been completed in time for the conference,

The AMS or Alpha Magnetic Spectrometer is a particle detector that was installed on the International Space Station in May 2011 and has been collecting data ever since. The scientific community is now eagerly waiting to hear about their results, in the hope of getting some clues as to what makes up 24% of all content of the Universe, namely what are the mysterious particles that form dark matter.

AMS is due to release data that will compare the flux of positrons in outer space with the flux of electrons. Positrons are the antimatter counterpart of electrons. The interest all stems from the fact that a few years ago, the PAMELA collaboration observed a larger positron flux at high energy than expected. It is relatively easy to think of various sources of electrons since we live in a world made of matter. But what could be a source for antimatter? One possible  explanation is to suppose that dark matter particles are annihilating into pairs of electrons and positrons, and hence providing a source of positrons.

Another group operating a satellite-born experiment, the FERMI-LAT collaboration partially confirmed that observation but only AMS has all the capabilities to really cross-check the PAMELA results. We will have to be a bit more patient until the AMS collaboration publishes with its first results.

The increase in the positron flux with respect to the electron flux as seen by various experiments. The AMS data should bring a definitive confirmation of the excess observed at high energy.

Meanwhile, the FERMI group has work on its hands as explained by Gabrijela Zaharijas since a theorist, Christoph Weniger, analysing data collected by FERMI, detected a signal in the form of a sharp spectral line at 130 GeV – gamma rays of a specific energy – coming from a region in the galactic center.  His approach was to look in areas of the galaxy where he expected to find the most dark matter and fewest sources of gamma rays of known origin. He studied five such locations in the center of our galaxy where dark matter is known to be more concentrated. For three of these locations, he found events in excess of the known sources of gamma rays, i.e more signal than background. The signal was also very strong, four times stronger than possible statistical fluctuations of the background level, that is 4.4 sigma.

The excess of events found by Christoph Weniger in FERMI data seen above the background described by a power law spectrum.

The FERMI collaboration has since improved the data calibration and modeling of energy dispersion, which should have led to an increase in the signal strength. On the contrary, they found the signal got fainter, making them doubt is was a real effect. In fact, while checking a region containing only background (the Earth atmosphere where lots of gamma rays are produced by incoming cosmic rays), they detected a similar “signal”, although fainter at 2.3 sigma. This is not quite enough to explain the anomaly detected in the galactic center but seems to indicate some instrumental error. Further investigations are underway.

We should soon get to the bottom of this story since a new telescope, HESS-2 in Namibia will start observing the galactic center region this month. In less than 50 hours in good operating conditions, they should be able to accumulate enough data to confirm or contradict the presence of this 130 GeV signal.

Will we soon have some hints on the mysterious nature of dark matter? It is well worth a bit more patience in the hope to learn more soon.

Pauline Gagnon

To be alerted of new postings, follow me on Twitter: @GagnonPauline or sign-up on this mailing list to receive and e-mail notification.


  • Soeren Jansen

    With my absolue unbiased eyes the Fermi plot look very strange. Please look on the data only. The plot is half splitted. Until 145 GeV the values are 10-20 counts with error bars of 10 counts. There is no peak or anything. Just random fluctuations around 15 counts. After 145 GeV the data points changing completly. Around 3-5 counts with much smaller error bars 5 counts. How could any analysis create a 4.4 sigma signal out of this mess of data? I’m absolutely unfamilar with the topic, but it looks like that this necessary to see that the data is bullshit and the analysis is so strong biased that it create a signal out of this bullshit!

  • Soeren Jansen

    BTW: The halfsplit of the data is even better visible in the residuall plot! In the 1st half the deviation from model is much stronger (even if you calculate the relative values) with more counts (???) than in the 2nd part after 145 GeV . This is statistical nonsense, the statiscal error is the proportional to the root of counts. So the relative failure should decrease with more counts. There seems to be different data cuts before and after 145 GeV.

  • lcs

    This is Poisson counting statistics, not gaussian fluctuations. The data above 145 GeV appear more tightly clustered around background because you can either have 3 counts, or 4 or 5 but nothing in between. All the points outside of the ‘peak’ are within 1 sigma of the background, so what is your problem?

  • Hello,

    you are perfectly entitled to comment on the content of this blog or material presented there, but please refrain from being offensive.

    I agree with you that if one looks at the data points after background subtraction below the plot, it is far from being obvious. It seems clear to me that the 4 sigma significance is not measured from the initial data point or the background-subtracted data points (where as you say, there is not much), but from the fit to the data, the curve shown below with a dashed line. If you look at that dashed curve, even by eye, one sees a clear signal. This is where the 4 sigma excess comes from.

    I hope this helps, Pauline

  • Soeren Jansen

    The absolute failure is root(counts) and increases of with more counts, but the relative failure root(counts)/counts should decrease (otherwise it makes no sense to measure longer and longer). The relative failure in the 1st part is higher than the 2nd one? The relative failure to the fit is equal or higher in the 1st part than in the 2nd with more counts, that’s makes no sense. Could you please explain this? I don’t understand the analysis my arguments are pure undergraduated statistic based.

  • Hello,

    I do not fully get your point either. The square root of a number gives the statistical uncertainty on this number. For example, if you pull out 100 candies from a bag and count 36 green ones, you can say you expect to see 36 + or – 6 green candies from another bag. Six here is the absolute statistical uncertainty on the measurement (36). What you call “failure”, we call it the statistical uncertainty. In my example, the relative uncertainty is 6 over 36, that is 16.67%. But say we had only pulled 16 green candies. Then the uncertainty would be 4 and the relative uncertainty 25%.
    Having a higher count leads to a more precise measurement. The first point shown (on the left part), I read roughly 15 +- 4, that is about 25% uncertainty. Bot on the right part, the last point is 4 +- 2, a 50% uncertainty. Your high school statistics are ok, all is fine. The more count, the smaller the relative error.

    Did I misunderstood you? If not, I hope this helps.

    Cheers, Pauline

    I hope this helps you see the difference between the value measured

  • It would appear that the interpretation of spiral galaxy rotation curves as indicative of the existence of a ‘dark matter halo’ of gravitating particles that neither radiate nor absorb electromagnetic radiation of any kind is incorrect. If the gravitational tidal forces produced in the disks by the external pull of the host cluster are not neglected, the observed rotation curves are consistent with Newton’s law of universal gravitation. See dynamical simulations and pending APS conference slides at GravitySim[dot]net

  • Hello Alexander,

    thank you for sharing your ideas here although this is really not the right forum. Clearly, you have a different interpretation for the existence of dark matter. I am glad to see you will present your own ideas at the American Physics Society meeting in April, which is a much more appropriate place for such a debate, not this site.

    Cheers, Pauline

  • Pingback: Dark matter retains all its mystery | Sung-Gi Kim()

  • Mark Zambelli

    Thankyou Pauline, very interesting blog entry. I wonder what the AMS data holds and whether that Fermi “signal” will lead onto anything or just disappear entirely with more data analysis. All in all, I count myself fortunate to be around in such exciting times when so much is being done by experimentalists and theorists alike (notice I didnt curse us all by saying “interesting” times ;). Thanks for the update.

  • Hello Mark,
    nice to hear you enjoyed it. You are not the only one curious about the AMS data. The whole scientific community is awaiting the AMS results with great anticipation. I know they are working really hard on it as they want to get it right. So waiting a few more months will ensure good quality analysis and better checks of all systematic effects. Well worth the wait. I will of course write about it when they release their new data.Interesting times indeed!

    Cheers, Pauline