A New Approach to Searching for Dark Matter Signals in Fermi-LAT Gamma Rays

TL;DR

This paper proposes a novel method using intensity ratios of gamma ray spectra from different sky regions to better distinguish dark matter signals from background uncertainties in Fermi-LAT data.

Contribution

It introduces a new ratio-based analysis technique that reduces systematic uncertainties and helps differentiate dark matter models from astrophysical backgrounds.

Findings

Ratio method effectively discriminates dark matter signals from backgrounds.

Dark matter models can produce spectral indices requiring extreme background explanations.

The approach enhances the sensitivity of gamma-ray data analysis for dark matter detection.

Abstract

Several cosmic ray experiments have measured excesses in electrons and positrons, relative to standard backgrounds, for energies from ~ 10 GeV - 1 TeV. These excesses could be due to new astrophysical sources, but an explanation in which the electrons and positrons are dark matter annihilation or decay products is also consistent. Fortunately, the Fermi-LAT diffuse gamma ray measurements can further test these models, since the electrons and positrons produce gamma rays in their interactions in the interstellar medium. Although the dark matter gamma ray signal consistent with the local electron and positron measurements should be quite large, as we review, there are substantial uncertainties in the modeling of diffuse backgrounds and, additionally, experimental uncertainties that make it difficult to claim a dark matter discovery. In this paper, we introduce an alternative method for understanding the diffuse gamma ray spectrum in which we take the intensity ratio in each energy bin of two different regions of the sky, thereby canceling common systematic uncertainties. For many spectra, this ratio fits well to a power law with a single break in energy. The two measured exponent indices are a robust discriminant between candidate models, and we demonstrate that dark matter annihilation scenarios can predict index values that require "extreme" parameters for background-only explanations.