Table of Contents
Fetching ...

The compositions of the HR 8799 planets reflect accretion of both solids and metal-enriched gas

Jerry W. Xuan, Jean-Baptiste Ruffio, Yayaati Chachan, Kazumasa Ohno, Aurora Y. Kesseli, Ruth A. Murray-Clay, Eve J. Lee, Julianne I. Moses, William O. Balmer, Aneesh Baburaj, Geoffrey A. Blake, Doug Johnstone, Yapeng Zhang, Heather A. Knutson, Dimitri Mawet, Charles Beichman, Klaus W. Hodapp, Marshall D. Perrin, Quinn M. Konopacky, Michael R. Meyer, Geoffrey Bryden, Thomas P. Greene, Jarron Leisenring, Marie Ygouf, Björn Benneke, Julie Inglis, Nicole L. Wallack

TL;DR

This work leverages JWST/NIRSpec IFU spectroscopy of all four HR 8799 planets to perform atmospheric retrievals that allow independent scaling of C/H, O/H, N/H, and S/H while accounting for clouds and disequilibrium chemistry. The results reveal super-solar enrichment in C, O, and S across the planets, with a notable N enrichment in the outer planet b, and support a formation scenario in which solids accrete beyond CO and N$_2$ snowlines while gas-accretion intra-CO snowline is CO-enriched. A simple disk model incorporating pebble drift and evaporation explains the observed volatile-to-refractory trends and implies a substantial pebble flux through the snowlines (e.g., $\sim750\pm200\,M_{\oplus}$). Together, these findings illuminate how planet-building blocks and disk gas interact to produce the metallicity patterns observed in giant exoplanets and demonstrate JWST’s power to constrain in detail exoplanet formation histories.

Abstract

With four giant planets ($m\sim5-10~M_{\rm Jup}$, $T_\rm{eff}\sim900-1200$ K) orbiting between 15-70 au, HR 8799 provides an unparalleled testbed for studying giant planet formation and probing compositional trends across the protoplanetary disk. We present new JWST/NIRSpec IFU observations ($2.85-5.3~μ$m, $R\approx2700$) that now include the spectrum of HR 8799 b, and higher S/N spectra for HR 8799 c, d, and e compared to that in Ruffio & Xuan et al. 2026. We detect CO, CH$_4$, H$_2$O, H$_2$S, CO$_2$, and for planet b, NH$_3$. We combine the NIRSpec spectra with $1-5 μ$m photometry to perform atmospheric retrievals that account for disequilibrium chemistry and clouds, and allow C/H, O/H, N/H, and S/H to scale independently. While the four planets are similarly enriched in carbon and oxygen, with C/H and O/H between $3-5\times$ stellar, we observe a tentative trend of increasing S/H - a tracer of refractory solids - from $2-5 \times$ stellar with increasing orbital distance. From HR 8799 b's NH$_3$ abundance, we estimate $\rm N/H=21.2^{+16.2}_{-8.8}\times$ stellar, suggesting the outer planet accreted significant amounts of N-rich gas. Overall, the elemental abundance patterns we observe are consistent with a picture where planet b formed between the CO snowline and the more-distant N$_2$ snowline, while the inner planets accreted $3 \times$ stellar CO-enriched disk gas within the CO snowline. The excess volatile mass from pebble drift and evaporation implies an integrated pebble flux of $750 \pm 200~M_{\oplus}$. The increase in the planets' S/H with orbital distance implies more solid accretion further out, which is quantitatively compatible with expectations from both pebble and planetesimal accretion ($2 \times$ Minimum Mass Solar Nebula) paradigms.

The compositions of the HR 8799 planets reflect accretion of both solids and metal-enriched gas

TL;DR

This work leverages JWST/NIRSpec IFU spectroscopy of all four HR 8799 planets to perform atmospheric retrievals that allow independent scaling of C/H, O/H, N/H, and S/H while accounting for clouds and disequilibrium chemistry. The results reveal super-solar enrichment in C, O, and S across the planets, with a notable N enrichment in the outer planet b, and support a formation scenario in which solids accrete beyond CO and N snowlines while gas-accretion intra-CO snowline is CO-enriched. A simple disk model incorporating pebble drift and evaporation explains the observed volatile-to-refractory trends and implies a substantial pebble flux through the snowlines (e.g., ). Together, these findings illuminate how planet-building blocks and disk gas interact to produce the metallicity patterns observed in giant exoplanets and demonstrate JWST’s power to constrain in detail exoplanet formation histories.

Abstract

With four giant planets (, K) orbiting between 15-70 au, HR 8799 provides an unparalleled testbed for studying giant planet formation and probing compositional trends across the protoplanetary disk. We present new JWST/NIRSpec IFU observations (m, ) that now include the spectrum of HR 8799 b, and higher S/N spectra for HR 8799 c, d, and e compared to that in Ruffio & Xuan et al. 2026. We detect CO, CH, HO, HS, CO, and for planet b, NH. We combine the NIRSpec spectra with m photometry to perform atmospheric retrievals that account for disequilibrium chemistry and clouds, and allow C/H, O/H, N/H, and S/H to scale independently. While the four planets are similarly enriched in carbon and oxygen, with C/H and O/H between stellar, we observe a tentative trend of increasing S/H - a tracer of refractory solids - from stellar with increasing orbital distance. From HR 8799 b's NH abundance, we estimate stellar, suggesting the outer planet accreted significant amounts of N-rich gas. Overall, the elemental abundance patterns we observe are consistent with a picture where planet b formed between the CO snowline and the more-distant N snowline, while the inner planets accreted stellar CO-enriched disk gas within the CO snowline. The excess volatile mass from pebble drift and evaporation implies an integrated pebble flux of . The increase in the planets' S/H with orbital distance implies more solid accretion further out, which is quantitatively compatible with expectations from both pebble and planetesimal accretion ( Minimum Mass Solar Nebula) paradigms.
Paper Structure (35 sections, 8 equations, 21 figures)

This paper contains 35 sections, 8 equations, 21 figures.

Figures (21)

  • Figure 1: Detection of the four planets orbiting the star HR 8799 with the moderate resolution mode (R$\sim$2,700) of JWST/NIRSpec IFU between $3-5\,\mu$m.(Left) Median spectral cube using the standard JWST calibration pipeline and combining the two observatory roll angles. This image does not include any PSF subtraction so the planets are hidden behind the starlight. (Middle) Signal-to-noise ratio map for planet detection. HR 8799 b, c, d, and e are detected with an S/N of 134, 211, 157,and 79 respectively. The S/N calculation is systematics limited and S/N values have been normalized to yield a standard deviation of unity in the absence of planetary signal. The white vertical gaps arise from masking out the IFU slices containing the saturated stellar PSF core. (Right) Planet $5\sigma$ detection limits for the fully combined dataset with each dot representing a spatial pixel in the field of view. The flux ratio is defined in the F444W filter. We use the estimated planet-to-star flux ratios and reference stellar fluxes from ref. Balmer2025b. The drop in sensitivity at the largest separations is due to the dithering pattern of the observation and edge effects.
  • Figure 2: JWST/NIRSpec spectrum of HR 8799 b. Panels a, c, e show the observed spectrum ($R\sim2,700$) in black and the best-fit model in orange. In the sub-panels below, the corresponding residuals after subtracting the best-fit petitRADTRANS model are plotted as gray lines and the $2.5\sigma$ uncertainties are shown as orange contours. The factor of 2.5 comes from the best-fit error scaling factor, and is largely driven by the enhanced residuals from $3.75-3.9~\mu$m. Panels b and d show data residuals after fitting an atmospheric model without a given species (NH$_3$, H$_2$S, $^{13}$CO, CO$_2$) in black, and the corresponding molecular templates in color. The similarity between the data residuals and molecular templates indicate that the highlighted species contribute significantly to the planet's spectra. On the right insets, we plot the cross-correlation functions (CCF) between the data residuals and models in the left insets. The CCF provides an estimate of the detection S/N for each molecule or isotopologue.
  • Figure 3: The colored points show photometry data of HR 8799 b from VLT/NACO, VLT/SPHERE, and JWST/NIRCam Currie2014_hr8799Zurlo2016Balmer2025b used in the retrievals. We show the best-fit photometry model in open circles. Random draws of the model spectrum at $R = 100$ are overlaid in light purple.
  • Figure 4: We show 200 random draws of the P-T profile for HR 8799 b in blue, and cloud condensation curves in dashed orange and purple lines. In the retrievals, the cloud base pressures are free parameters, and the retrieved MgSiO$_3$ cloud base ($2\sigma$ confidence interval) is marked by the thick orange line. The overlaid gray contour is the emission contribution function, which uses the top x-axis (wavelength in $\mu$m). The gray dashed histogram on the left is the wavelength-weighted emission contribution.
  • Figure 5: Posterior distributions for [N/H] and the $\log{P_{\rm quench,diff}}$ parameter that controls the quench pressure of NH$_3$. These values are directly taken from the petitRADTRANS retrieval of HR 8799 b, and do not incorporate the VULCAN analysis which provided an improved estimate of N/H (Section \ref{['sec:nitrogen']}). The corner plot shows a strong correlation, demonstrating an inherent degeneracy between [N/H] and the quench pressure of NH$_3$ for this planet.
  • ...and 16 more figures