Identification and Mitigation of a Vibrational Telescope Systematic with Application to Spitzer

TL;DR

This study identifies a vibrational systematic in Spitzer IRAC data affecting transit observations, and introduces an adaptive elliptical-aperture photometry method to mitigate this issue, improving data quality.

Contribution

The paper presents a novel adaptive elliptical-aperture photometry technique to remove vibrational systematics in Spitzer data, outperforming traditional circular apertures in reducing noise.

Findings

Vibrational systematics correlate with the point-response function width.

Elliptical aperture photometry reduces residuals and correlated noise.

The method can be applied to other telescopes and observations.

Abstract

We observed Proxima Centauri with the Spitzer Space Telescope InfraRed Array Camera (IRAC) five times in 2016 and 2017 to search for transits of Proxima Centauri b. Following standard analysis procedures, we found three asymmetric, transit-like events that are now understood to be vibrational systematics. This systematic is correlated with the width of the point-response function (PRF), which we measure with rotated and non-rotated Gaussian fits with respect to the detecor array. We show that the systematic can be removed with a novel application of an adaptive elliptical-aperture photometry technique, and compare the performance of this technique with fixed and variable circular-aperture photometry, using both BiLinearly Interpolated Subpixel Sensitivity (BLISS) maps and non-binned Pixel-Level Decorrelation (PLD). With BLISS maps, elliptical photometry results in a lower standard deviation of normalized residuals, and reduced or similar correlated noise when compared to circular apertures. PLD prefers variable, circular apertures, but generally results in more correlated noise than BLISS. This vibrational effect is likely present in other telescopes and Spitzer observations, where correction could improve results. Our elliptical apertures can be applied to any photometry observations, and may be even more effective when applied to more circular PRFs than Spitzer's.