In-flight performance of the IXPE telescopes

Abstract

We present a comprehensive characterization of the on-orbit imaging performance of the three telescopes on board the Imaging X-ray Polarimetry Explorer (IXPE). Each telescope comprises a Wolter-I mirror module assembly and a Gas Pixel Detector focal-plane detector unit (DU). We analyze data from point-like X-ray sources and fit a composite point spread function (PSF) model that we compare with ground calibrations. We study the dependence of the PSF parameters and of the angular resolution, in terms of half-power diameter (HPD), on the time and source counting rate. We find no significant secular evolution of PSF parameters or HPD over 30 months on orbit, with average HPDs of $26.1 \pm 0.5$ arcsec (Telescope 1), $32.1 \pm 0.5$ arcsec (Telescope 2), and $30.9 \pm 0.6$ arcsec (Telescope 3), and rate trends consistent with zero up to source counting rates of $\sim60$ cts s$^{-1}$ in the 2-3 keV energy band for all three telescopes. We set a 99% C.L. upper limit of 4.4% on the optics-induced polarization in the PSF halo, and find no measurable degradation of the polarization modulation factor in the wings versus the core due to mis-reconstructed photoelectron tracks. IXPE's imaging performance thus is consistent with the $\leq30$ arcsec observatory requirement with high stability, ensuring robust spatially resolved polarization measurements for the mission's projected lifetime through 2030.

Figures (6)

• Figure 1: From top to bottom, best fit of the radial profiles and source image in log-scale of the three telescopes for an exemplary observation, obsid 2008601 (Cyg X-1). In the right column, the green contours on the source images enclose 50%, 90%, and 99% of the total counts.
• Figure 2: Encircled energy fraction (EEF) for the three telescopes resulting from the integration of the fits shown in Fig. \ref{['fig:psf_v3']} as a function of the angular distance from the source position for obsid 2008601 (Cyg X-1) in the 2-3 keV energy band. The vertical dashed lines mark the the radius enclosing 50% of the EEF, i.e. half of the Half Power Diameter (HPD).
• Figure 3: Fit of the Gaussian width $\sigma$ as a function of time (left panel) and source 2-3 keV rate (right panel) for all 3 detectors on-board IXPE. The red vertical line marks the epoch of optics realignment, the blue vertical line marks the epoch of the temperature set-point change, and the pink shaded line the ground-calibration value and its uncertainties. The shaded areas cover the 1 standard deviation uncertainty of the linear regression.
• Figure 5: Fit of the King wing slope $\eta$ as a function of time (left panel) and source 2-3 keV rate (right panel) for all 3 detectors on-board IXPE. The red vertical line marks the epoch of optics realignment, the blue vertical line marks the epoch of the temperature set-point change, and the pink shaded line the ground-calibration value and its uncertainties. The shaded areas cover the 1 standard deviation uncertainty of the linear regression.
• Figure 6: Fit of the exponential scale $r_0$ as a function of time (left panel) and source 2-3 keV rate (right panel) for all 3 detectors on-board IXPE. The red vertical line marks the epoch of optics realignment, the blue vertical line marks the epoch of the temperature set-point change, and the pink shaded line the ground-calibration value and its uncertainties. The shaded areas cover the 1 standard deviation uncertainty of the linear regression. The shaded data points in the left panel are the ones from sources with 2-3 keV counting rate $<1$ count s$^{-1}$ that are excluded from the time trend fitting.