Aviation is a critical component of Australia’s national security and has been identified as one of the hardest sectors to decarbonise. Sustainable aviation fuel (SAF) remains the only near‑term pathway fully compatible with existing long‑haul aircraft and current fuel infrastructure. Among emerging routes, Power‑to‑Liquid (PtL) SAF has been recognised as the most strategically important option for future production. Although decades of experimental fuel studies have generated valuable empirical knowledge, a significant gap persists in molecular‑level electronic‑structure understanding—precisely the insight required for the rational design of high‑energy‑density (HED) components essential for advanced PtL SAF formulations.
In this study, we integrate machine learning (ML) with density functional theory (DFT) to identify and investigate strained multicyclic hydrocarbons at the molecular and electronic‑structure level for next‑generation SAF applications. Our ML models prescreen candidate HED hydrocarbons with desirable properties—particularly net heat of combustion (NHOC) and density—while DFT calculations elucidate their quantitative structure–property relationships (QSPR) and guide new SAF molecular design. Recent applications include studies on norbornadiene/quadricyclane (NBD/QC) derivatives, the military fuel JP‑10, adamantane and its bridgehead‑methylated derivatives. For JP‑10, we reveal clear electronic‑structure differences between the exo (fuel) and endo (solid) isomers, and present new theoretical first ionization potentials (IP) 50 years after its experimental measurement in 1973, radical reactivity patterns, and 13C NMR signatures. Complementary work on adamantane and its bridgehead‑methylated derivatives establishes a rigorous structure–property framework linking orbital topology, molecular symmetry, phase behaviour, and liquid‑phase HED fuel performance. These integrated ML–DFT investigations advance SAF innovation from the quantum‑chemical level to practical implementation, providing a robust molecular‑level foundation for the rational design of next‑generation HED SAF components.