Early Evidence for Polar Orbits of Sub-Saturns Around Hot Stars

TL;DR

This study extends obliquity measurements of sub-Saturns from cool to hot host stars by presenting RM and RRM analyses for TOI-1135 b, a sub-Saturn on a hot $T_{ m eff}\approx6320$ K star. Using NEID RM observations, TESS photometry, iSpec spectroscopy, and EXOFASTv2 global modeling, the team derives a sky-projected obliquity $\lambda\approx-68^{\circ}$ and a true obliquity $\psi\approx72^{\circ}$ for TOI-1135 b, with complementary results from Reloaded RM. A population-level analysis comparing sub-Saturns and hot Jupiters shows a significant difference in obliquity distributions around cool stars, consistent with weaker tidal realignment for sub-Saturns, while hot-star sub-Saturns (including TOI-1135 b and TOI-1842 b) remain near-polar, clustering near $65^{\circ}$ as predicted by secular resonance crossing. The findings suggest that the polar preference for sub-Saturns may persist above the Kraft break and provide empirical support for dynamical misalignment pathways, with future RM measurements and outer-companion constraints needed to confirm the trends and refine theoretical models.

Abstract

Sub-Saturns have been reported to preferentially occupy near-polar orbits, but this conclusion has so far been based primarily on systems with cool host stars; obliquity measurements for sub-Saturns orbiting hot stars remain scarce. Expanding the census into the hot-star regime is essential to test whether the polar preference persists across the Kraft break and to diagnose the underlying excitation mechanisms. In this work, we present Rossiter-McLaughlin observations of TOI-1135 b, a sub-Saturn orbiting a hot star with $T_{\rm eff}=6320\pm120$ K, using WIYN/NEID. We confirm its near-polar architecture, measuring a sky-projected obliquity of $λ=-68.1^{+7.5}_{-5.3}$ degrees and a true obliquity of $ψ=72.2^{+6.4}_{-6.6}$ degrees. Coupling our new measurement with stellar-obliquity data from the literature, we find that sub-Saturns and hot Jupiters around cool stars are unlikely to be drawn from the same parent distribution at the $5.2σ$ level, consistent with weaker tidal realignment induced by lower-mass planets. Of the two known misaligned sub-Saturns around hot stars, both are near-polar, suggesting that the polar preference may extend above the Kraft break. Moreover, their obliquities lie near $\sim 65$ degrees, supporting predictions from secular resonance crossing for sub-Saturns around rapidly rotating hot stars.

Figures (3)

• Figure 1: Spectral energy distribution, TESS transit photometry, out-of-transit radial velocities, and Rossiter–McLaughlin effect modeling for TOI-1135. The upper panel in each plot shows the observational datasets along with the corresponding best-fit model from the EXOFASTv2 global modeling (solid red line). The resulting RM signal-to-noise ratio Kipping2024 is 8.1. The lower panel in each plot displays the residuals. (The data used to create this figure are https://raw.githubusercontent.com/wangxianyu7/Data_and_code/refs/heads/main/TOI-1135RM/TOI1135.csv.)
• Figure 2: Projected spin–orbit angle as a function of stellar effective temperature for transiting exoplanets. The top panel shows sub-Saturns, while the bottom panel displays hot Jupiters. The vertical dashed line indicates the Kraft break at 6100 K, separating stars with convective envelopes (cooler) from those with radiative envelopes (hotter). TOI-1135 b is marked with a red star and labeled for reference. The predicted relationship between $T_{\rm eff}$ and $\lambda$ from secular resonance crossing Petrovich2020 is shown as the shaded gray region.
• Figure 3: Top left: Local CCFs (out-of-transit - in-transit) in the stellar rest frame behind TOI-1135 b. The solid orange lines represent local CCFs at disk positions with $\langle \mu \rangle \geq 0.4$, dashed orange lines correspond to $0 < \langle \mu \rangle < 0.4$, and gray lines indicate out-of-transit regions. Bottom left: A two-dimensional map of the local CCFs, color-coded by local flux. Top right: Local RVs derived from local CCFs where $\langle \mu \rangle \geq 0.4$, with the two best-fit models (orange: differential rotation + centre-to-limb quadric model with $i_\star<90^{\circ}$; grey: solid body + centre-to-limb quadric model ) shown as dashed lines. Bottom right: Residuals corresponding to the two models shown above.