PEPSI Investigation, Retrieval, and Atlas of Numerous Giant Atmospheres (PIRANGA). III. Composition and winds in the atmosphere of TOI-1518 b

TL;DR

TOI-1518 b is studied with high-resolution PEPSI/LBT transmission spectra to detect atomic species and measure time-resolved winds across the transit. Fe I is detected at $7.8\sigma$ and Fe II at $8.9\sigma$, with Cr I and Ni I tentatively detected at $4.4\sigma$ and $4.0\sigma$, confirming strong iron signatures and extending atmospheric inventory. Template spectra from petitRADTRANS and cross-correlation against Doppler-shifted templates enable spatially-resolved wind measurements, with Doppler shadow modeled via MISTTBORN/HORUS and phase-binned wind velocities showing blueshifted day-to-nightside flows of roughly $-6.8$ to $-8.8$ km s$^{-1}$. Nodal precession is assessed, yielding $db/dt = -0.0116 \pm 0.0036$ yr$^{-1}$ and an expected $b \sim 0.89$ during the 2022 epoch, consistent with prior work and supporting a dynamic, evolving transit geometry. The results motivate simultaneous emission spectroscopy and improved ingress sampling to fully characterize atmospheric dynamics and composition.

Abstract

Ultra-hot Jupiters (UHJs) orbit close to their host stars and experience extreme conditions, making them important laboratories to explore atmospheric composition and dynamics. Transmission spectroscopy is a useful tool to reveal chemical species and their vertical and longitudinal distribution in the atmosphere. We use transmission spectra from the PEPSI spectrograph on the Large Binocular Telescope to search for species and measure their time-resolved wind velocities in the atmosphere of TOI-1518 b. We detect Fe I at 7.8$σ$ and Fe II at 8.9$σ$, and tentatively detect Cr I at 4.4$σ$ and Ni I at 4.0$σ$. The time-resolved wind velocities of Fe I show a velocity pattern that is consistent with the velocity pattern of Fe II. TOI-1518 b joins a small sample of UHJs for which time-resolved wind velocities have been measured.

Figures (4)

• Figure 1: (Top) Model telluric spectrum produced by MOLECFIT, scaled to fit on the same y-axis as the template spectra. (Other panels) Template transmission spectra for TOI-1518 b assuming a single species and solar H and He in its atmosphere. The VMRs (Table \ref{['tbl:species']}) were determined via a grid search that maximized SNR. The y-axis for each species is on the same scale. The blue shaded region corresponds to the wavelengths observed by the PEPSI blue arm and the red shaded regions correspond to the wavelengths observed by the PEPSI red arm. The grey shaded region identifies wavelengths that have significant telluric contamination, while the hatched regions are masked out entirely in our analysis due to the presence of strong telluric contamination.
• Figure 2: (Top) Raw CCF for Fe I. Dark (black) regions are anti-correlated and light (white) regions are correlated. The blue lines show the ingress and egress of the transit, calculated using the transit duration determined by Duck et al. (in prep). The red dotted line shows the expected velocity of the atmospheric signal $V_{sig} = K_P \sin(2 \pi \phi) + V_{sys}$, where $K_P = 157$ km s$^{-1}$ is the radial velocity semi-amplitude (Cabot2021), $\phi$ is the orbital phase, and $V_{sys} = -11.17 \pm 0.035$ km s$^{-1}$ is the systemic velocity offset in the LBT PEPSI frame Petz2024PIRANGA. The dark shaded region is the Doppler shadow. It makes a V shape with the light shaded region, which is the observed atmospheric signal. (Middle) Model of the Doppler shadow from HORUS. This adopts transit parameters shown in Table \ref{['tbl:mcmcparam']} which were obtained via MCMC with MISTTBORN. (Bottom) Residual after subtracting the Doppler shadow.
• Figure 3: CCFs shifted into the planetary rest frame over a range of $K_P$ values for our detected species. For each species, the red and blue arm data was combined using the weights described in §4. The crosshairs intersect at $v = 0$ and the expected $K_P$ value. Each of these species shows a peak (red) near the crosshairs, with the peak SNR identified in the upper left.
• Figure 4: (Top) Fe I phase bins with wind velocities. (Bottom) Same for Fe II. SNR is shown by color with yellow representing higher SNR. The horizontal white lines show ingress and egress, and the vertical dotted line shows a velocity of 0 km s$^{-1}$. The white points show the wind velocity for each phase bin, fit by a Gaussian using MCMC + GP. The uncertainties are equal to the uncertainties in the centers of the Gaussians, with a minimum uncertainty of 1 km s$^{-1}$. We used dynamic binning, where if the data in one phase bin resulted in a low SNR fit, it was combined with the data in the next phase bin and fitted again. For combined bins, the orbital phase is the weighted sum of each orbital phase bin weighted by their SNR (amplitude of the Gaussian). We neglect original bins with negative Gaussian amplitudes due to weak SNR.