Development of a C/H/O/N/S chemical network: Experimental benchmark, application to exoplanets, and identification of key C/S coupling pathways

Context . The detection of SO 2 in WASP-39 b and WASP-107 b has increased interest in modeling photochemistry in exoplanets, especially sulfur compounds, creating an urgent need for reliable kinetic networks. However, sulfur kinetics data remain scarce, and comprehensive networks in combustion literature are only recent for H/O/S, in progress for C/H/O/S, and almost nonexistent for the full C/H/O/N/S system. Current exoplanet networks usually neglect this coupling by adding a sulfur sub-mechanism to C/H/O/N networks. Aims . We integrated sulfur kinetics into our previously developed C 0 -C 2 /H/O/N chemical network, including coupling with carbon and nitrogen chemistry, for conditions between 500–2500 K and 100–10 −6 bar. The reliability of the resulting network was a major focus. Methods . We combined available combustion networks with original ab initio calculations where data were lacking, and compared the network with 1606 experimental measurements from combustion and pyrolysis of H 2 S, CH 3 SH, CS 2 , and OCS. We then applied it to six exoplanets (GJ 436 b, GJ 1214 b, HD 189733 b, HD 209458 b, WASP-39 b, and WASP-107 b) using the 1D kinetic model FRECKLL, and computed transmission spectra with TauREx 3.1. Results were compared with other sulfur networks used in exoplanet modeling. Results . The new network includes 226 species and 1692 mostly reversible reactions. Carbon–sulfur coupling strongly impacts abundance profiles and observables. We predict that the abundance of CS 2 is much higher than in current networks, while CH 2 S emerges as a key species to describe this coupling. Conclusions . The significant role of C–S species underscores the need for chemical networks to be extensively tested against experimental data to improve reliability, particularly in the JWST era. The detection of CS 2 in TOI-270 d further shows the need to address the blind spots in current models. Combustion and pyrolysis data prove to be valuable tools for this task.