Friday, 17 April 2015

Antiprotons from AMS

This week the AMS collaboration released the long expected measurement of the cosmic ray antiproton spectrum.  Antiprotons are produced in our galaxy in collisions of high-energy cosmic rays with interstellar matter, the so-called secondary production.  Annihilation of dark matter could add more antiprotons on top of that background, which would modify the shape of the spectrum with respect to the prediction from the secondary production. Unlike for cosmic ray positrons, in this case there should be no significant primary production in astrophysical sources such as pulsars or supernovae. Thanks to this, antiprotons could in principle be a smoking gun of dark matter annihilation, or at least a powerful tool to constrain models of WIMP dark matter.

The new data from the AMS-02 detector extend the previous measurements from PAMELA up to 450 GeV and significantly reduce experimental errors at high energies. Now, if you look at the  promotional material, you may get an impression that a clear signal of dark matter has been observed.  However,  experts unanimously agree that the brown smudge in the plot above is just shit, rather than a range of predictions from the secondary production. At this point, there is certainly no serious hints for dark matter contribution to the antiproton flux. A quantitative analysis of this issue appeared in a paper today.  Predicting  the antiproton spectrum is subject to large experimental uncertainties about the flux of cosmic ray proton and about the nuclear cross sections, as well as theoretical uncertainties inherent in models of cosmic ray propagation. The  data and the predictions are compared in this Jamaican band plot. Apparently, the new AMS-02 data are situated near the upper end of the predicted range.

Thus, there is no currently no hint of dark matter detection. However, the new data are extremely useful to constrain models of dark matter. New constraints on the annihilation cross section of dark matter  are shown in the plot to the right. The most stringent limits apply to annihilation into b-quarks or into W bosons, which yield many antiprotons after decay and hadronization. The thermal production cross section - theoretically preferred in a large class of WIMP dark matter models - is in the  case of b-quarks excluded for the mass of the dark matter particle below 150 GeV. These results provide further constraints on models addressing the hooperon excess in the gamma ray emission from the galactic center.

More experimental input will allow us to tune the models of cosmic ray propagation to better predict the background. That, in turn, should lead to  more stringent limits on dark matter. Who knows... maybe a hint for dark matter annihilation will emerge one day from this data; although, given the uncertainties,  it's unlikely to ever be a smoking gun.

Thanks to Marco for comments and plots. 

11 comments:

Cédric Bardot said...

May the silence of zetta and zepto-spaces not be eternal...

Anonymous said...

It seems to me that the brown smudge of the AMS promotional material has a large overlap with the Jamaican band plot.
I will say that they seems so different mainly because the different axis scale.
The brown smudge should be the Donato (2009) model that also was used by PAMELA for their comparisons.
At 350 GeV (last bin center) both the brown band and the Jamaican band starts (roughly) from 3.5x10^-5. So the problem now is the theory uncertainties that in 6 years (from 2009 to yesterday) does not shrinks so much.

Anonymous said...

Hi, it seems that the "shit" band comes from Donato et al. PRL 102 (2009) 071301 (arXiv:0810.5292 see Fig.3), though the linear scale on the x-axis may give a different impression.

Jester said...

That's correct. In 2008 it wasn't shit. Since then, there has been a lot of new data on secondary cosmic rays at high energies, including the boron-to-carbon ratio in AMS-02 measured up to 500 GeV. The parameters from 2008 Donato et al do not fit this data.
It's as if the LHC showed jets+MET spectrum overlaid with QCD predictions from LEP times and sold it as a hint of supersymmetry.

Anonymous said...

Hi Jester. If in 2009 it wasn't shit, only little has changed today.

1) You can easily see that the latest B/C ratio data are perfectly consistent with the assumption on propagation parameters made in Donato 2009 (I mean the data released by PAMELA on ApJ 2014 as well as the data just presented by AMS).

2) The thing that may be changed is the injection spectrum of protons and Helium: since proton and He spectra harden at high energies, one may expect that there are a bit more high-energy antiprotons produced by p+p collisions (say less than 10%). But actually, this was clear also in 2009 because ATIC-2 data were already released since 2006 (but I think that Donato et al ignored that hardening and they made a power-law approximation).

In conclusion, if you think that the band is shit, then it was shit also in 2009. Otherwise, it's still not shit.

Jester said...

You probably know more than me about the nuts and bolts, but I'm confused by 1). My statement was based on 1504.03134. They fit the AMS B/C data, and the diffusion parameters they obtain (Table 1) seem to be way off from those quoted by Donato et al. Am I misinterpreting something?

Anonymous said...

Anyway, the whole DM interpretation of the PAMELA/FERMI/AMS-02 signal is ruled out by Planck. I am surprised no one mentioning this ...

Anonymous said...

Why is ruled out? Their point is quite naive: in our point of the Galaxy there could be an excess of DM (the classical cloud of DM giving the so called "boost factor") and so the cross section could be much less. In addition Plank measured (I'm afraid in a model dependant way, but I'm not sure of this) the cross section of 13 Gyears ago: now the Universe is a little bit older and "still"...

Jester said...

There are 2 separate issues. One is whether the excess of positrons can be explained by dark matter. The Planck limits provide another strong argument against that but, as the previous commenter points out, these limits are not model independent. The other, discussed in this post, is whether there is an excess of antiprotons. It could be in principle a completely independent excess. For example, the positrons could be explained by pulsars, whereas the hypothetical antiproton excess could be due to dark matter.

Anonymous said...

Hi Jester. 1504.03134 uses simplified toy models just to asses some model uncertainties. It is not a determination of the propagation parameters, forget it.
The propagation model used in the 2009 paper (diffusion+convection+reacceleration) is still compatible with new nuclear data, although the new data should provide tighter constrains to the model parameters.
In fact, the propagation uncertainty band should have been reduced from 2009 to 2015. Since it is not the case, it means that the uncertainties were likely under-estimated in 2009.
On the other hand, the new paper 1504.04276 is very conservative. Maybe too much? Note that the AMS data at HE are still on the upper part of the Jamaican band: in order to explain the pbars as secondaries, all theoretical biases have to act in the same direction. Of course this does not mean that there is a hint of DM; but the "secondary production scenario" require to force too many parameters. It is un-natural, we are missing something.

Anonymous said...

planck is model dependent of course but...there is a model that can explain the observations and that is theoretically accurate at better than percent level...
On the contrary what we have here ? It is model independent since we don't have a model ! :-)