SciPost Submission Page
AMS-02 antiprotons and dark matter: Trimmed hints and robust bounds
by Francesca Calore, Marco Cirelli, Laurent Derome, Yoann Genolini, David Maurin, Pierre Salati, Pasquale D. Serpico
This Submission thread is now published as
|Authors (as registered SciPost users):
|Marco Cirelli · David Maurin
Based on 4 yr AMS-02 antiproton data, we present bounds on the dark matter (DM) annihilation cross section vs. mass for some representative final state channels. We use recent cosmic-ray propagation models, a realistic treatment of experimental and theoretical errors, and an updated calculation of input antiproton spectra based on a recent release of the PYTHIA code. We find that reported hints of a DM signal are statistically insignificant; an adequate treatment of errors is crucial for credible conclusions. Antiproton bounds on DM annihilation are among the most stringent ones, probing thermal DM up to the TeV scale. The dependence of the bounds upon propagation models and the DM halo profile is also quantified. A preliminary estimate reaches similar conclusions when applied to the 7 years AMS-02 dataset, but also suggests extra caution as for possible future claims of DM excesses.
Published as SciPost Phys. 12, 163 (2022)
Author comments upon resubmission
List of changes
Please see the attached a PDF version with modifications in color.
Submission & Refereeing History
You are currently on this page
Reports on this Submission
- Cite as: Anonymous, Report on arXiv:2202.03076v2, delivered 2022-04-22, doi: 10.21468/SciPost.Report.4973
I would like to thank the authors for the various additions to the manuscript, which in my opinion constitute significant improvements, in particular for non-expert readers. The new version offers a pedagogical and thorough discussion of the best practices in the analysis of anti-proton data, which will be very influential and useful for the community. I am happy to recommend publication, but cannot resist leaving a few more comments:
- I find the arguments of the authors against profiling over nuisance parameters quite convincing. However, my personal conclusion from this discussion is that a good compromise could be obtained by marginalizing (rather than profiling) over nuisance parameters. It would be interesting to understand whether this leads to similar results as the procedure currently implemented.
- I agree with the authors that the LR hypothesis test that they perform is both common and reasonable. I just wanted to point out that there are reasons to suspect that a MC simulation of mock experiments would lead to somewhat different p-values.
- A small comment that slipped through the original review: As far as I am aware, the Neyman-Pearson lemma only applies to simple hypotheses (with no free parameters), i.e. it only covers the case of a likelihood ratio, not a profile likelihood ratio.
Finally, I would like to thank the authors for the additional explanation regarding the different propagation schemes, which I found very illuminating.
No changes to the manuscript required.