Universal mapping of the structural space by 20 NMR-verifiable C,H,O-atomic units

Heterogeneous mixtures of organic molecules in natural ecosystems1–3 and green chemistry4,5 comprise thousands to millions of distinct, yet largely unknown constituents with a wide range of individual properties. The complexity of these organic mixtures precludes a meaningful molecular precision analysis of all individual compounds6–8. Hence, bulk-type parameters with low structural resolution are the current norm of description, characterization and valorization9–14. Here, we propose a quantitative system of 20 carbon-centered structural units, each defined by a central C atom, and either C, H or O one bond away, which enables the classification of all chemically reasonable structures of CnHmOq molecules. These units are directly quantifiable by multiplicity-edited 13C NMR spectra of both individual organic molecules and mixtures. Application of this quantitative C,H,O-atomic unit system to nine relevant classes of biogeochemical molecules defined an expansive structural space, representing the currently known small molecule universe. Its application to characterize the exceptional structural complexity of freshwater DOM15–17 demonstrated a separate, rarely occupied structural space of oxygenated, polyalicyclic, and architectural multiform molecules, resembling desirable lead structures in current drug design. The chemical spaces revealed by the 20 C,H,O-atomic units define a structural view on the carbon cycle, through which hydrogen-rich open chain alkyl-containing molecules develop towards congested, structurally recalcitrant, oxygenated polyalicyclic DOM molecules, rich in quaternary carbon atoms. Synthesis and degradation of DOM molecular structures are mass-dependent and non-reversible. The proposed C,H,O-classification framework offers large scale discovery opportunities where structural resolution of unknown organic mixtures is a desirable objective.