First Confident SiO Signal in the Atmosphere of a Directly Imaged Companion TWA5B: A Key to Refractory Material and Clouds ✦
In the spectrum of the giant exoplanet TWA 5 B, obtained in the M-band with the CRIRES+ instrument, gaseous SiO was confidently detected (signal-to-noise ratio = 7.5) via its ro-vibrational band at 4 µm. The SiO abundance was log(SiO) = -3.56^{+0.42}_{-0.32} by volume, indicating nearly complete absence of magnesium-silicate cloud condensation — silicon is present almost exclusively in the gas phase. Joint analysis with volatile species CO (S/N=9.1) and H₂O (S/N=18.8) yielded C/O (stellar), O/Si and C/Si (moderately sub-stellar) ratios, as well as a super-stellar Si/H ([Si/H]_star = 1.41^{+0.42}_{-0.32}). These proportions are consistent with formation via core accretion beyond the CO snowline or via gravitational instability with subsequent solid enrichment. Gas-phase SiO is a unique tracer of cloud properties in hot gas giants, enabling the study of dominant condensates in populations of directly imaged planets and isolated brown dwarfs.
Context
Direct imaging of giant exoplanets offers a unique window into their chemical makeup and formation history. Refractory elements like silicon, which existed in solid form in the protoplanetary disk, are of particular interest as they trace the amount of accreted silicates. Until recently, however, detecting refractory compounds in the atmospheres of directly imaged planets was challenging due to low contrast and cloud screening. The new CRIMSON survey with CRIRES+ on the VLT exploits the M-band (3.5–5.2 µm) to hunt for spectral lines of molecular SiO — the main carrier of gaseous silicon in hot atmospheres. This approach offers a direct estimate of solid enrichment and simultaneously diagnoses the presence of silicate clouds that can distort observed abundances.
Methods
Observations of the TWA5 system (distance 49.6 pc) were carried out with VLT/CRIRES+ in high spectral resolution mode (R≈92,000) with adaptive optics. The spectrograph slit was oriented to simultaneously capture light from the central binary star (M1.5) and the companion TWA5B (M8.5) at an angular separation of 1.8″. To battle strong thermal background and telluric contamination, a special A-B background subtraction procedure and a telluric transmission model built from calibration star observations were applied. After spectral extraction and removal of the dominant host star light using principal component analysis (PCA), cross-correlation with synthetic spectra revealed molecular lines of H2O, CO, and SiO. The model also included opacity from collision-induced absorption of H2–H2 and H2–He. Then, within a Bayesian retrieval using HyDRA, the atmospheric abundances, dynamics, and temperature profile were estimated, with models subjected to the same PCA cleaning as the data to account for its effects.
Results
In the spectrum of TWA5B, three molecules were confidently detected: H2O (SNR=18.8), CO (SNR=9.1), and SiO (SNR=7.5). The detection of SiO stands out — the first robust evidence of gaseous silicon in a directly imaged companion. The SiO abundance was log(VMR)=–3.56, corresponding to nearly all silicon being in the gas phase at pressures near 1 bar. The retrieval also indicated low cloud opacity and a deep cloud base, ruling out significant condensation of magnesium-silicate dust in the observed region. The measured Si/H ratio was ~25 times higher than stellar ([Si/H]★=1.41), pointing to strong atmospheric enrichment with refractory material. The C/O=0.26, O/Si≈9.6, and C/Si≈2.4 ratios fall within the range consistent with accretion beyond the CO snowline or gravitational instability with subsequent solid capture.
Implications
Observing SiO in the M-band provides a fundamentally new tool for studying the chemistry of refractory elements in exoplanets. Since SiO is the dominant gaseous carrier of silicon in hot, cloud-free atmospheres, its abundance directly reflects the primordial mass of silicates captured during formation. This breaks the degeneracy inherent in the C/O ratio alone and allows a more precise reconstruction of a planet’s birthplace in the disk. Moreover, the dependence of SiO abundance on silicate condensation turns it into a sensitive indicator of cloud types (MgSiO3, Mg2SiO4, SiO2) and their formation conditions. In the future, such measurements can be extended to the entire population of hot Jupiters and isolated brown dwarfs, creating a condensation map as a function of temperature.
Future development
The CRIRES+ M-band methodology will be applied to other targets in the CRIMSON survey (including β Pic b and HR 3549 b), enabling the tracking of SiO abundance evolution from 2400 K down to 1200 K. Combined with archival K-band data, this will paint a full picture of changes in C/O, volatile-to-refractory ratios, and isotopic composition for each companion. Future instruments, such as METIS on the ELT, will provide even higher quality spectra in the L and M bands, while space observatories (JWST) will offer additional information on silicate dust in the mid-infrared. Joint analysis of SiO and dust feature data will refine models of cloud particle growth and settling, including kinetic nucleation effects.
Impact
The work will impact observational exoplanet astrophysics (development of high-resolution methods in the thermal infrared), planet formation theory (constraints on solid accretion), and atmospheric modeling (cloud diagnostics and non-equilibrium chemistry).
Next steps
Expand the sample to cooler objects where silicate clouds are expected to appear, and compare SiO measurements with infrared dust excesses obtained by JWST/MIRI.
Key open problems
The detection of SiO links the problem of giant planet formation (specifically, mechanisms of solid accretion in disks) with the task of determining cloud composition in sub-stellar atmospheres. It is also a step toward resolving how condensation of refractory elements distorts observed volatile abundances, which directly affects the interpretation of key C/O and metallicity ratios.
🎯 Silicon monoxide (SiO) is a molecule that in the gas phase occurs not only in the atmospheres of hot exoplanets but also in the interstellar medium, where it forms in supernova shock waves. However, under Earth conditions, SiO rapidly polymerizes into solid silicates — the very stuff ordinary sand is made of.
Key numbers
- SiO signal-to-noise ratio: 7.5
- H2O signal-to-noise ratio: 18.8
- TWA5B temperature: ~2400 K
- Si/H enrichment over stellar: [Si/H]★ = 1.41 (~25 times)
- TWA5B mass: 25 ± 5 Jupiter masses