Pulsar based modeling of point spread function of Fermi Large Area Telescope

TL;DR

The paper addresses biases in searches for gamma-ray extended emission caused by imperfect Fermi-LAT PSF models. It develops a data-driven PSF revision by comparing LAT IRF-based PSFs to bright pulsar signals, fitting pulsar data to the same two-King-function PSF framework, and deriving updated parameters that better match observations. Applying the revised PSF to Mrk 501 removes the previously claimed extended emission signal, highlighting the impact of PSF systematics on IGMF-related interpretations and underscoring the need for pulsar-calibrated PSFs at energies below ~10 GeV. The work demonstrates the value of pulsars as real-data calibrators for PSF uncertainties, while noting limitations at higher energies where pulsar statistics are insufficient and Monte Carlo PSF determinations remain necessary.

Abstract

Sensitivity of searches for extended emission around gamma-ray sources is naturally limited by the precision of the knowledge of the Point Spread Function (PSF) of gamma-ray telescopes. Inaccuracies in the PSF models of the Fermi Large Area Telescope (LAT) can potentially lead to false positive detections of source extension. We explore uncertainties in the Fermi/LAT PSF by comparing the PSF models provided by the Fermi/LAT Instrument Response Functions (IRFs) with signals of bright pulsars. We compare the analytical PSF models of Fermi/LAT IRFs with pulsar data and fit the pulsar data with the same analytical model as in the Fermi/LAT IRFs to derive an improved set of PSF parameters. We then apply this revised PSF parameterisation to the search of extended emission around a blazar, Mrk 501. We find that the parameters of the analytical PSF models of Fermi/LAT IRFs are inconsistent with the pulsar data. We obtain an improved set of PSF parameters from the fits to pulsar data that is consistent with observations. We find no evidence of the previously reported extended signal around Mrk 501 if the revised PSF consistent with pulsar data is used in data analysis.

Figures (9)

• Figure 1: Pulse profiles of Vela, Geminga and Crab pulsars in the energy range 1-1.7 GeV for different event types (PSF0-PSF3). Vertical lines show the boundaries of the on-pulse / off-pulse intervals.
• Figure 2: Radial profiles of the stacked pulsed emission signal for different event types (PSF0-PSF3) in the energy range 1-1.7 GeV.
• Figure 3: Radial profiles of Vela pulsar signal in the energy range 1-1.7 GeV, for different event types (PSF0, PSF3). Top panels show the fits of the data with the PSF model described in section \ref{['sec:psf']} (black solid lines). Bottom panels show the positive (black) and negative (red) fit residuals with the residual flux levels corresponding to 5% (black dashed lines) and 10% (black dotted lines) of the point source signal.
• Figure 4: Same as in Fig. \ref{['fig:profiles_Vela']}, but for the energy range 5.6-10 GeV.
• Figure 5: Quality of the fit of the pulsar data with PSF derived from the P8R3 IRFs. Left panel is for the FRONT events and right panel is for the BACK events. Color shows the excess $\chi^2$ of the data fit with the model PSF with parameters from the IRF file, compared to the best-fit model with the Kings function parameters fitted to the data.