# Toward More Reliable Analytic Thermochemical-equilibrium Abundances

## TL;DR

This paper improves an existing analytic method for calculating thermochemical-equilibrium abundances in planetary atmospheres, making it more reliable across a wide range of conditions and implementing it in an open-source Python package.

## Contribution

It introduces a new, stable framework for thermochemical calculations that extends previous methods, ensuring accuracy across diverse atmospheric parameters.

## Key findings

- Accuracy better than 10% for major species
- Accuracy better than 50% for minor species
- Framework implemented in open-source Python package

## Abstract

Heng & Tsai (2016) developed an analytic framework to obtain thermochemical-equilibrium abundances for H$_{2}$O, CO, CO$_2$, CH$_4$, C$_{2}$H$_{2}$, C$_{2}$H$_{4}$, HCN, NH$_3$, and N$_2$ for a system with known temperature, pressure, and elemental abundances (hydrogen, carbon, nitrogen, and oxygen). However, the implementation of their approach can become numerically unstable under certain circumstances, leading to inaccurate solutions (e.g., ${\rm C/O} \ge 1$ atmospheres at low pressures). Building up on their approach, we identified the conditions that prompt inaccurate solutions, and developed a new framework to avoid them, providing a reliable implementation for arbitrary values of temperature (200 to $\sim$2000 K), pressure ($10^{-8}$ to $10^{3}$ bar), and CNO abundances ($10^{-3}$ to $\sim 10^{2}\times$ solar elemental abundances), for hydrogen-dominated atmospheres. The accuracy our analytic framework is better than 10% for the more abundant species that have mixing fractions larger than $10^{-10}$, whereas the accuracy is better than 50% for the less abundant species. Additionally, we added the equilibrium-abundance calculation of atomic and molecular hydrogen into the system, and explored the physical limitations of this approach. Efficient and reliable tools, such as this one, are highly valuable for atmospheric Bayesian studies, which need to evaluate a large number of models. We implemented our analytic framework into the \textsc{rate} Python open-source package, available at https://github.com/pcubillos/rate .