Atmospheric constraints on GJ 1214 b from CRIRES+ and prospects for characterisation with ANDES

In this study, we aim to constrain the atmospheric composition of GJ 1214 b using all available transits observed with the upgraded CRIRES+ spectrograph at the VLT by searching for the signatures of water vapour, methane, and carbon dioxide. We analysed eight CRIRES+ transit datasets covering the K band (1.90-2.45 microns) at a resolving power of R ~ 100,000. We used the SysRem algorithm to correct for telluric and stellar contributions and employed the cross-correlation technique with templates from petitRADTRANS to search for H2O, CH4, and CO2. Injection-recovery tests across a grid of metallicities (Z) and cloud-deck pressures (pc) were performed to quantify detection limits. We also generated predictions for ANDES observations using end-to-end simulated datasets with EXoPLORE. We detect no significant H2O, CH4, or CO2 signatures. Injection-recovery tests show that such non-detections exclude atmospheres with low-altitude clouds and moderate or low metallicities. CH4 yields the tightest empirical limits, with CO2 unexpectedly ruling out intermediate metallicities (~ 100xsolar) with clouds deeper due to its rapidly rising opacity in compressed, high-Z atmospheres. Our constraints are in line with either a high-Z or a high-altitude aerosol layer, in agreement with recent JWST inferences. The combined analysis of eight CRIRES+ datasets provides the most stringent high-resolution constraints on the atmospheric properties of GJ 1214 b to date. Simulations of a single transit observed with ANDES on the ELT predict modest improvements for H2O, a substantially expanded detectable region for CH4, and the strongest gains for CO2, making the latter a particularly effective tracer for characterising high-metallicity atmospheres in sub-Neptunes.

💡 Research Summary

This paper presents a comprehensive high‑resolution spectroscopic study of the sub‑Neptune GJ 1214b using all available transits observed with the upgraded CRIRES+ instrument on the VLT, and it evaluates the prospects for future observations with the ANDES spectrograph on the ELT. Eight K‑band (1.90–2.45 µm) transit datasets were reduced and analyzed at a resolving power of R ≈ 100 000. After careful preprocessing—including nodding‑pair wavelength alignment, order trimming, third‑order polynomial continuum normalization, 3σ outlier clipping, and masking of deep telluric regions—the authors applied the SysRem algorithm to remove dominant telluric and stellar signals. SysRem iterations were tuned per spectral order to avoid over‑ or under‑correction.

Cross‑correlation (CC) analyses were performed with high‑resolution templates for H₂O, CH₄, and CO₂ generated by petitRADTRANS. The planetary orbital velocity (Kp ≈ 106.7 km s⁻¹) and systemic velocity (Vsys ≈ 0 km s⁻¹) were searched over a realistic grid. No statistically significant (>5σ) CC peaks were found for any molecule, indicating non‑detections in the combined CRIRES+ data.

To translate these non‑detections into atmospheric constraints, the authors conducted injection‑recovery tests across a two‑dimensional grid of atmospheric metallicity (Z, ranging from 1× to 1000× solar) and cloud‑deck pressure (pc, from 0.01 mbar to 100 mbar). Synthetic planetary signals were injected into the raw spectra, processed through the identical pipeline, and the recovery rate was measured. The results show that atmospheres with low‑altitude clouds (pc ≲ 10 mbar) and low‑to‑moderate metallicities (Z ≲ 10× solar) are ruled out. Methane provides the tightest constraints: for Z ≈ 30× solar or lower, clouds must reside above ~1 mbar to be detectable, and the recovery rate exceeds 90 % in that regime. Carbon dioxide behaves differently; its opacity rises steeply in compressed, high‑Z atmospheres, allowing CO₂ to exclude intermediate metallicities (~100× solar) even when clouds are deeper (pc ≈ 0.1 mbar). This “compression effect” makes CO₂ a particularly sensitive tracer of metal‑rich, high‑pressure atmospheres.

The paper also presents forward‑looking simulations for ANDES using the EXoPLORE end‑to‑end framework. A single transit observed with ANDES on the 39‑m ELT is predicted to modestly improve H₂O detectability (≈30 % lower detection limit), dramatically expand the accessible Z‑pc parameter space for CH₄ (≈2× larger region), and provide the strongest gains for CO₂ (≈3× better sensitivity), especially for high‑metallicity scenarios. These gains stem from ANDES’s broader wavelength coverage, higher instrumental stability, and the ELT’s vastly larger collecting area.

In summary, the combined CRIRES+ analysis delivers the most stringent high‑resolution constraints on GJ 1214b to date, supporting a picture in which the planet either possesses a high‑metallicity atmosphere (high mean molecular weight) or is shrouded by a high‑altitude aerosol layer—both conclusions consistent with recent JWST transmission spectra. The study underscores the difficulty of detecting molecular signatures in cloudy sub‑Neptunes with current ground‑based facilities, but it also highlights the transformative potential of ELT/ANDES, particularly through CO₂ observations, to break the degeneracy between metallicity and cloud altitude in future atmospheric characterizations of GJ 1214b and similar worlds.