Leveraging Photometry for Deconfusion of Directly Imaged Multi-Planet Systems

TL;DR

The paper tackles the confusion problem in directly imaged multi-planet systems by introducing a photometry-based likelihood ranking to augment the existing deconfuser, enabling better discrimination between orbit candidates using phase variation in reflected light. It models the planet-to-star flux ratio with $F_p(\lambda,\alpha)/F_s(\lambda) = A_g(\lambda)\left(\frac{R_p}{d}\right)^2 \Phi(\lambda,\alpha)$ and a Lambertian phase function to capture brightness changes along the orbit, while incorporating instrument-specific noise modeled after the Roman Coronagraph. The authors demonstrate improved orbit discrimination in a subset of highly confused three-planet simulations across low, medium, and high inclinations, with improvements in 7/10, 6/10, and 6/10 cases respectively. This work suggests that single-band photometric phase information can meaningfully augment deconvolution and orbit-fitting pipelines for future missions like HWO, potentially enhancing target selection and observation planning.

Abstract

Planned and future missions, including the Habitable Worlds Observatory (HWO), will aim to directly image Earth-like exoplanets around Sun-like stars in reflected light. Determining whether an exoplanet is in the habitable zone of its star may be difficult in multi-planet systems when the observer does not know in advance which point source detection corresponds to which planet. This "confusion" problem will be a concern for future missions due to the high occurrence rate of multi-planet systems, and will be exacerbated by lack of prior knowledge about planets' orbital parameters or characteristics, particularly for systems at high inclination with respect to the observer. This work addresses the confusion problem by developing a photometry model and new orbit ranking function to augment the "deconfuser" tool and account for phase variation exhibited by a planet throughout its orbit. We demonstrate the new ranking scheme on a subset of thirty highly confused simulated multi-planet systems among three inclination groupings (low, medium, and high). Results indicate that photometry improves differentiation of previously confused orbits in 7/10 of the low inclination cases, 6/10 of the medium inclination cases, and 6/10 of the high inclination cases. This improvement in handling highly confused systems emphasizes that photometry shows promise for supporting orbit discrimination and deconfusion of directly imaged multi-planet systems, and should be considered when fitting orbits to detections.

Figures (2)

• Figure 1: Planetary orbital phase illustration for an edge-on system ($i=90^\circ$) (top), with the phase angle convention used in this work and a simulated Lambertian phase curve (bottom) (see Section \ref{['subsubsec:phase_function']} and Equation \ref{['eqn:lambertian']}; brown2005single). Full phase, or $\alpha=0^\circ$, occurs behind the star, relative to the observer (secondary eclipse); $\alpha=180^\circ$ occurs during the planet's primary eclipse of the star. $\pm 180^\circ$ are the same point on the orbit. We note that the planetary systems explored in this work will not necessarily transit their host star and this figure is for illustrative purposes.
• Figure 2: Flow of the photometry ranking algorithm as it fits with the deconfuser. The photometry model accepts the orbital parameters for each confused orbit option and calculates the phase angle and expected planetary brightness for each detection. Detector noise is added to obtain a more realistic brightness for each detection. We compare the likelihood of each confused orbit option given the planet detections, which translates to a ranking value for each confused system option.