The First Dedicated Survey of Atmospheric Escape from Planets Orbiting F Stars

TL;DR

This work delivers the first dedicated survey of atmospheric escape from gas giants orbiting F-type stars, using ground-based He$^*$ transit observations with Palomar/WIRC across ten transits of six planets. By translating mid-transit He$^*$ excess absorptions into 1D Parker-wind mass-loss rates with sunbather and star-specific XUV proxies, the study finds mass-loss rates are largely consistent with energy-limited expectations, but are strongly modulated by Roche filling factors and stellar XUV luminosity, rather than by NUV flux alone. Detected outflows include WASP-12 b and WASP-180 A b, with two marginal detections (HAT-P-8 b, WASP-93 b) and two non-detections (WASP-103 b, KELT-7 b), illustrating substantial system-to-system variation. The results imply that extreme outflows observed in some early-type-star planets are not universal, emphasize the importance of direct XUV measurements, and motivate future 3D, spectroscopically resolved investigations to constrain outflow geometry and temporal variability.

Abstract

Hydrodynamic escape can strip the envelopes of close-in exoplanets, but most observations of atmospheric mass loss to date have been confined to planets orbiting K and M dwarfs. A growing body of detections of atmospheric escape from planets orbiting early-type stars indicates that they may have significantly stronger and more extended outflows than planets orbiting cooler stars. However, it is unclear whether this limited sample of planets is representative of all gas giants orbiting early-type stars. Motivated by this question, we initiated the first dedicated survey of atmospheric escape from gas giants orbiting F stars in order to understand how their distinct radiation environments shape planetary outflows. We observed ten transits of six planets in an ultra-narrowband filter centered on the metastable helium line using Palomar/WIRC. We report strong ($>3σ$) detections of atmospheric escape for WASP-12~b and WASP-180~A~b, tentative ($>2σ$) detections for WASP-93~b and HAT-P-8~b, and non-detections for WASP-103~b and KELT-7~b. We fit these measurements with a 1D Parker wind model to derive corresponding mass-loss rates, and combine our results with literature measurements to obtain an updated picture of mass loss from planets orbiting early-type stars. Our results indicate that the observed variation in mass-loss rates can be explained by a combination of Roche filling factor and XUV luminosity, and disfavors NUV-driven escape models.

Figures (6)

• Figure 1: Confirmed transiting planet radii and periods drawn from the NASA Exoplanet Archive on August 26th, 2025 AkesonChen2013ps. Closed colored circles represent He$^*$ detections of atmospheric escape as a function of spectral type. Open circles denote survey targets. The dashed lines indicate the Neptune desert boundary from Castro-GonzalezBourrier2024.
• Figure 2: Transit light curves (top) and residuals (bottom) for the six planets in our survey. The detrended data (grey points, different nights have different symbol shapes) are binned to 10 minute cadence and overplotted as black circles. Our best-fit transit models are overplotted as red lines, with red shading indicating the corresponding 1$\sigma$ uncertainty. The blue curves show the predicted light curve models for the case where there is no outflow.
• Figure 3: Published XUV luminosity as a function of stellar effective temperature for early-type stars hosting planets with detected outflows and targets in our survey (see §\ref{['sec:intro']} and \ref{['sec:sunbather']}). The XUV luminosities of the proxy MUSCLES stellar spectra used in our sunbather models are shown for reference. Shading corresponds to the projected stellar rotational velocities. The published XUV luminosities for WASP-12 and KELT-9 are upper bounds, which we plot here with downward pointing arrows. Gray points correspond to reported X-ray luminosities of F stars from ShimuraMitsuishi2025.
• Figure 4: Roche filling factor as a function of XUV flux at the planet's orbit for survey targets and planets orbiting early-type stars with published mass loss measurements. Filled circles correspond to robust ($\geq3\sigma$) detections of atmospheric escape, open circles correspond to tentative ($\leq2\sigma$) detections of atmospheric escape, open circles with dashed edges indicate non-detections. Published detections are shown as a lighter shade of purple. Point size is proportional to the retrieved mass loss rate, indicated by the legend. The published XUV luminosities for WASP-12 and KELT-9 are upper bounds and therefore the XUV fluxes at the planets' orbits are also upper bounds, which we symbolize here with horizontal arrows. WASP-33 does not have a published XUV luminosity, so we have assigned this star the same XUV luminosity as KELT-9, the only other A-star in this sample, when calculating the XUV flux at the planet. We have indicated with arrows that the actual XUV flux may be higher or lower than this estimate. For F-type stars without published XUV luminosities (HAT-P-67, HAT-P-8, WASP-93, WASP-103, WASP-94 A), we plot two points connected by a horizontal dashed line. The lower bound corresponds to a flux calculated using the proxy star XUV luminosity (see Table \ref{['table:massloss']} for proxy star choice) and the upper bound corresponds to a flux calculated using the XUV luminosity of HAT-P-32. Palomar survey targets in this set (HAT-P-8, WASP-93, WASP-103) have point sizes corresponding to the retrieved mass-loss rates calculated at each luminosity value.
• Figure 5: Roche filling factor as a function of NUV flux at the planet's orbit for survey targets and planets orbiting early-type stars with published mass loss measurements. See the Fig. \ref{['fig:RochevsXUV']} caption for additional details.