A Composite Topological Descriptor for Analyzing Sublimation-Enthalpy and Ionization-Energy Trends in Polycyclic Conjugated Hydrocarbons

This is the authors' abstract. We don't add key points for this paper.

Polycyclic conjugated hydrocarbons (PCHs) provide a useful molecular set for examining how size, connectivity, and fusion topology are reflected in graph-theoretical descriptors. In this study, a log-transformed equal-weight composite descriptor, C(G), was constructed from the Wiener, Harary, Randić, first Zagreb, and metric degree polynomial indices. Structural descriptor and degeneracy analyses were conducted for 23 PCHs, whereas endpoint-specific complete datasets of 20 compounds were used for sublimation enthalpy (ΔHsub) and experimental gas-phase ionization energy (IE). The ionization-energy-related electronic-response proxy was defined as RIE = 1/IE, without any additional numerical scaling. C(G) showed positive monotonic associations with ΔsubHm (298.15 K) (Spearman ρ = 0.8951, p = 9.95 × 10–8) and RIE (ρ = 0.7364, p = 2.14 × 10–4). Simple linear regression explained 80.54% of the variation in ΔsubHm (298.15 K) and 50.66% of the variation in RIE within the corresponding datasets. Under leave-one-out cross-validation, the OLS multi-descriptor model produced the highest internal Q2 values for both ΔsubHm (298.15 K) (Q2 = 0.8420; RMSECV = 14.73 kJ mol–1) and RIE (Q2 = 0.5491; RMSECV = 0.005722 e–1). However, these results should be interpreted cautiously because each endpoint dataset contains only 20 compounds and the descriptor components are strongly intercorrelated. At full numerical precision, C(G) assigned distinct values to all 23 structures, although its discrimination decreased after numerical rounding. The findings therefore represent exploratory, internally validated structure–property associations rather than externally validated predictive models.