A Deep Search for Additional Satellites around the Dwarf Planet Haumea

TL;DR

This study used advanced image processing techniques on Hubble data to search for small, undetected satellites around Haumea, setting size limits and demonstrating a new effective method for satellite detection.

Contribution

Developed and validated a non-linear shift-and-stack method to enhance detection of small satellites in HST data, establishing size constraints around Haumea.

Findings

No new satellites detected around Haumea.

Satellites larger than ~40 km are ruled out within the Hill sphere.

Satellites as small as ~10 km radius are excluded at certain distances.

Abstract

Haumea is a dwarf planet with two known satellites, an unusually high spin rate, and a large collisional family, making it one of the most interesting objects in the outer solar system. A fully self-consistent formation scenario responsible for the satellite and family formation is still elusive, but some processes predict the initial formation of many small moons, similar to the small moons recently discovered around Pluto. Deep searches for regular satellites around KBOs are difficult due to observational limitations, but Haumea is one of the few for which sufficient data exist. We analyze Hubble Space Telescope (HST) observations, focusing on a ten-consecutive-orbit sequence obtained in July 2010, to search for new very small satellites. To maximize the search depth, we implement and validate a non-linear shift-and-stack method. No additional satellites of Haumea are found, but by implanting and recovering artificial sources, we characterize our sensitivity. At distances between $\sim$10,000 km and $\sim$350,000 km from Haumea, satellites with radii as small as $\sim$10 km are ruled out, assuming an albedo ($p \simeq 0.7$) similar to Haumea. We also rule out satellites larger than $\gtrsim$40 km in most of the Hill sphere using other HST data. This search method rules out objects similar in size to the small moons of Pluto. By developing clear criteria for determining the number of non-linear rates to use, we find that far fewer shift rates are required ($\sim$35) than might be expected. The non-linear shift-and-stack method to discover satellites (and other moving transients) is tractable, particularly in the regime where non-linear motion begins to manifest itself.