Strong and variable stratospheric CO emission from lava-planet 55 Cnc e observed with NIRCam/JWST
Pith reviewed 2026-06-27 08:28 UTC · model grok-4.3
The pith
Strong variable CO emission from lava planet 55 Cnc e requires a hydrogen-rich atmosphere with a steep thermal inversion.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
An unambiguously strong carbon monoxide emission signal is recovered during one of five JWST epochs of 55 Cnc e, with possible weaker detections in two others. The signal strength and variability are difficult to reconcile with a hydrostatic atmosphere and require both a steep thermal inversion at 1-10 mbar and a CO2 abundance at least three orders of magnitude below that of CO. Self-consistent atmospheric modeling shows these conditions are produced most readily in a hydrogen-rich atmosphere, which yields the steepest inversions and highest CO/CO2 ratios. The epoch-to-epoch differences suggest the CO traces a transient, dynamically active component potentially linked to variable atmospheric
What carries the argument
Cross-correlation of native-resolution NIRCam spectra to extract molecular emission lines, combined with self-consistent atmospheric models that solve for pressure-temperature structure and equilibrium chemistry across different bulk compositions.
If this is right
- The atmosphere of 55 Cnc e is likely hydrogen-rich rather than dominated by CO or CO2.
- Steep thermal inversions can form at millibar pressures in lava-planet atmospheres.
- Atmospheric outflow on these planets can be variable and detectable through molecular emission lines.
- The CO/CO2 ratio acts as a diagnostic for the presence of hydrogen in the atmosphere.
- Similar high-resolution observations of other ultra-short-period rocky planets can test for comparable dynamic atmospheres.
Where Pith is reading between the lines
- Variable magma-ocean outgassing could drive the observed epoch-to-epoch changes in the CO signal.
- Repeated observations over many orbits could map the spatial or temporal structure of any outflow.
- If hydrogen-rich envelopes prove common on lava planets, this would alter expectations for their long-term evolution and volatile loss.
- The same cross-correlation approach may be applied to other JWST datasets of close-in rocky worlds to search for similar transient molecular features.
Load-bearing premise
The cross-correlation signal originates from the planet atmosphere rather than stellar activity, instrumental effects, or contamination, and the atmospheric models accurately capture the temperature structure and chemistry without missing processes.
What would settle it
New JWST observations of additional eclipses that show either no CO emission signal at any epoch or a consistent hydrostatic model that fits all data without requiring a steep inversion and strongly suppressed CO2.
read the original abstract
Some rocky planets orbit so close to their host stars that stellar heating melts their surfaces. They offer a rare glimpse of planets in a magma-ocean state, providing an observable analogue to processes that likely shaped the early Earth and other terrestrial planets during their infancy. Recent JWST observations of five eclipses of the prototypical lava planet 55 Cnc e have confirmed earlier hints that it exhibits highly variable thermal emission, with low-resolution spectroscopy pointing to a possible volatile-rich atmosphere likely rich in CO and CO2. Here we report on an analysis of the same JWST datasets but at their native spectral resolution, utilizing cross-correlation techniques. An unambiguously strong ~8 sigma signal from CO in emission is recovered during one out of five epochs, with potential ~3 sigma detections during two others. The strongest observed cross-correlation signal is difficult to reconcile with a hydrostatic atmosphere, requiring a steep and strong thermal inversion at the right pressure level (~1-10 mbar) and a relative abundance of CO2 that is at least 3 orders of magnitude lower which would otherwise mask the CO signal. Self-consistent atmospheric modelling indicates that this is most readily achieved in a hydrogen-rich atmosphere, which produces the steepest inversions and highest CO/CO2 ratios. The pronounced epoch-to-epoch variability suggests that the CO signal does not trace a static atmosphere alone, but may reveal a transient, dynamically active component, potentially linked to variable atmospheric outflow.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript analyzes five JWST NIRCam eclipse observations of the lava planet 55 Cnc e at native spectral resolution using cross-correlation techniques. It reports an ~8σ CO emission signal in one epoch, marginal ~3σ signals in two others, and none in the remaining two. The authors interpret the strong detection as incompatible with a hydrostatic atmosphere, requiring a steep thermal inversion at 1-10 mbar and a CO/CO2 ratio at least 1000 times higher than equilibrium chemistry would allow; self-consistent models favor a hydrogen-rich composition, and the epoch-to-epoch variability is attributed to transient dynamical processes such as atmospheric outflow.
Significance. If the planetary origin of the cross-correlation signal and the fidelity of the self-consistent models are confirmed, the result would provide direct evidence for a non-equilibrium, possibly hydrogen-rich atmosphere on an ultra-short-period rocky planet and would constrain the pressure level and strength of thermal inversions in such environments. The multi-epoch dataset also offers a rare probe of atmospheric variability, which could link to magma-ocean outgassing or escape processes.
major comments (3)
- [Results (cross-correlation analysis)] The central claim that the ~8σ CCF peak must originate from the planet (rather than residual stellar activity, telluric lines, or instrumental systematics) rests on the assumption that the template matching isolates only planetary CO lines after detrending. The manuscript should provide explicit null tests (e.g., cross-correlation with the template shifted to non-planetary velocities or with a CO-free template) and show the full CCF maps for all five epochs in the results section to quantify the false-positive rate.
- [Atmospheric modeling] The requirement for a hydrogen-rich atmosphere with CO2/CO ≳ 10^3 lower than equilibrium is derived from self-consistent P-T and chemistry models. However, the manuscript does not demonstrate that alternative compositions (e.g., CO2-rich or metal-rich cases with non-LTE effects or additional absorbers) are ruled out; the modeling section should include a grid of retrievals or forward models with explicit parameter ranges to show that only the H-rich family reproduces both the inversion depth and the observed CCF strength.
- [Atmospheric modeling] The interpretation that the signal requires a steep inversion at 1-10 mbar and cannot be produced by a hydrostatic atmosphere depends on the pressure level at which the CO lines form. The paper should quantify the contribution function or line-formation pressures for the adopted models and test whether modest changes in the assumed eddy diffusion or vertical mixing can relax the need for such an extreme inversion.
minor comments (2)
- [Figures] Figure captions should explicitly state the velocity range and significance threshold used for the CCF peaks.
- [Abstract and text] The abstract states 'unambiguously strong ~8 sigma' while the text notes 'potential ~3 sigma detections'; the terminology for marginal signals should be standardized throughout.
Simulated Author's Rebuttal
We thank the referee for their constructive comments, which help clarify the robustness of our cross-correlation results and atmospheric modeling. We address each major point below and have revised the manuscript accordingly to incorporate additional tests and figures.
read point-by-point responses
-
Referee: [Results (cross-correlation analysis)] The central claim that the ~8σ CCF peak must originate from the planet (rather than residual stellar activity, telluric lines, or instrumental systematics) rests on the assumption that the template matching isolates only planetary CO lines after detrending. The manuscript should provide explicit null tests (e.g., cross-correlation with the template shifted to non-planetary velocities or with a CO-free template) and show the full CCF maps for all five epochs in the results section to quantify the false-positive rate.
Authors: We agree that explicit null tests strengthen the planetary origin claim. In the revised manuscript we have added cross-correlations with the template shifted to non-planetary velocities and with a CO-free template; both yield no significant peaks above the noise floor. We have also moved the full CCF maps for all five epochs into the main results section (previously in the appendix) so that the false-positive rate can be directly assessed from the data. revision: yes
-
Referee: [Atmospheric modeling] The requirement for a hydrogen-rich atmosphere with CO2/CO ≳ 10^3 lower than equilibrium is derived from self-consistent P-T and chemistry models. However, the manuscript does not demonstrate that alternative compositions (e.g., CO2-rich or metal-rich cases with non-LTE effects or additional absorbers) are ruled out; the modeling section should include a grid of retrievals or forward models with explicit parameter ranges to show that only the H-rich family reproduces both the inversion depth and the observed CCF strength.
Authors: We have expanded the modeling section with a grid of self-consistent forward models spanning CO2-rich, metal-rich, and mixed compositions, including non-LTE effects and additional absorbers where relevant. Only the hydrogen-rich family simultaneously produces the observed thermal inversion depth and matches the measured CCF peak strength; the new grid is summarized in an additional figure and table. revision: yes
-
Referee: [Atmospheric modeling] The interpretation that the signal requires a steep inversion at 1-10 mbar and cannot be produced by a hydrostatic atmosphere depends on the pressure level at which the CO lines form. The paper should quantify the contribution function or line-formation pressures for the adopted models and test whether modest changes in the assumed eddy diffusion or vertical mixing can relax the need for such an extreme inversion.
Authors: We have added contribution functions for the adopted models, confirming that the CO lines form primarily between 1–10 mbar. We also ran sensitivity tests varying the eddy diffusion coefficient and vertical mixing by an order of magnitude; even with these changes the data still require a steep inversion, as shallower profiles fail to reproduce the observed emission signal. These tests are now described in the revised modeling section. revision: yes
Circularity Check
No significant circularity in observational detection and post-hoc modeling.
full rationale
The paper reports a cross-correlation analysis of native-resolution JWST NIRCam eclipse spectra of 55 Cnc e, recovering an ~8σ CO emission signal in one epoch (with marginal signals in others). This detection is obtained directly from the data after standard removal of stellar, telluric, and instrumental signals; the subsequent claim that the signal requires a steep 1-10 mbar thermal inversion and CO/CO2 ratio ≳10^3 is an interpretive step that invokes self-consistent atmospheric models. Neither the reported CCF strength nor the required abundances are shown to reduce by construction to quantities fitted from the same dataset. No self-citation load-bearing steps, self-definitional loops, or fitted-input-called-prediction patterns appear in the derivation chain. The analysis is therefore self-contained against external benchmarks.
Axiom & Free-Parameter Ledger
free parameters (2)
- CO/CO2 abundance ratio
- Thermal inversion depth and strength
axioms (2)
- domain assumption Cross-correlation with molecular templates reliably isolates planetary emission signals in NIRCam time-series data
- domain assumption Self-consistent radiative-convective models accurately predict temperature structure and chemistry for hydrogen-rich compositions at the relevant pressures
Reference graph
Works this paper leans on
-
[1]
Tambuto, P . et al. Confirming Variability in the Secondary Eclipse Depth of the Super-Earth 55 Cancri e. The Astrophysical Journal 155, 221 (2018) 7. Snellen, I. Exoplanets at high spectral resolution. Annual Review of Astronomy and Astrophysics 63, 83-125 (2025) 8. Greene, T.P . et al. λ = 2.4 to 5 μm spectroscopy with the James Webb Space Telescope NIRC...
Pith/arXiv arXiv 2018
-
[2]
Allard, F . et al. Model Atmospheres and Spectra: The Role of Dust. Brown Dwarfs, Proceedings of IAU Symposium #211, held 20-24 May 2002 at University of Hawaii, Honolulu, Hawaii. Edited by Eduardo Martín. San Francisco: Astronomical Society of the Pacific, p. 325 (2003) 32. Chubb, K. et al. The ExoMolOP database: Cross sections and k-tables for molecules ...
arXiv 2002
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.