Fermi-LAT improved Pass~8 event selection

TL;DR

This paper introduces an improved event selection method for the Fermi-LAT Pass 8 data, significantly reducing residual background and anisotropy, leading to a cleaner dataset for gamma-ray observations.

Contribution

The paper presents a novel background rejection technique for Fermi-LAT Pass 8 data that minimizes data loss while reducing residual background and anisotropy.

Findings

Residual background reduced by a factor of ~2 at 1-3 GeV

Less than 1% data loss with new cuts

Produced a cleaner dataset (P8R3) for gamma-ray analysis

Abstract

The current version of the Fermi Large Area Telescope data (P8R2) has been publicly available since June 2015, with the caveat that the residual background of all event classes, except ULTRACLEANVETO, was not fully isotropic: it was enhanced by a factor ~2 at 1-3 GeV within ~20 deg of the Ecliptic compared to the poles. By investigating the residual background using data only, we were able to find two sources of residual background: one due to non-interacting heavy ions and one due to cosmic-ray electrons leaking through the ribbons of the Anti-Coincidence Detector, the latter source being responsible for the background anisotropy. A set of simple cuts allows us to reject these events while losing less than 1% of the SOURCE class acceptance. This new selection has been used to produce a new version of the LAT data (P8R3).