Spitzer Secondary Eclipses of the Dense, Modestly-irradiated, Giant Exoplanet HAT-P-20b Using Pixel-Level Decorrelation

TL;DR

This paper introduces pixel-level decorrelation (PLD), a novel method for correcting intra-pixel sensitivity fluctuations in Spitzer data, enabling more accurate measurements of exoplanet secondary eclipses and atmospheric properties.

Contribution

We developed and validated PLD, a new technique that effectively removes intra-pixel effects in Spitzer data without measuring stellar position fluctuations, improving eclipse analysis accuracy.

Findings

PLD significantly reduces red noise in Spitzer photometry.

HAT-P-20b has a temperature of 1134K with no strong molecular absorption.

Eclipses occur with a slight orbital eccentricity, suggesting possible Kozai migration.

Abstract

HAT-P-20b is a giant exoplanet that orbits a metal-rich star. The planet itself has a high total density, suggesting that it may also have a high metallicity in its atmosphere. We analyze two eclipses of the planet in each of the 3.6- and 4.5 micron bands of Warm Spitzer. These data exhibit intra-pixel detector sensitivity fluctuations that were resistant to traditional decorrelation methods. We have developed a simple, powerful, and radically different method to correct the intra-pixel effect for Warm Spitzer data, which we call pixel-level decorrelation (PLD). PLD corrects the intra-pixel effect very effectively, but without explicitly using - or even measuring - the fluctuations in the apparent position of the stellar image. We illustrate and validate PLD using synthetic and real data, and comparing the results to previous analyses. PLD can significantly reduce or eliminate red noise in Spitzer secondary eclipse photometry, even for eclipses that have proven to be intractable using other methods. Our successful PLD analysis of four HAT-P-20b eclipses shows a best-fit blackbody temperature of 1134 +/-29K, indicating inefficient longitudinal transfer of heat, but lacking evidence for strong molecular absorption. We find sufficient evidence for variability in the 4.5 micron band that the eclipses should be monitored at that wavelength by Spitzer, and this planet should be a high priority for JWST spectroscopy. All four eclipses occur about 35 minutes after orbital phase 0.5, indicating a slightly eccentric orbit. A joint fit of the eclipse and transit times with extant RV data yields e(cos{omega}) = 0.01352 (+0.00054, -0.00057), and establishes the small eccentricity of the orbit to high statistical confidence. Given the existence of a bound stellar companion, HAT-P-20b is another excellent candidate for orbital evolution via Kozai migration or other three-body mechanism.