Recalcitrant soil carbon is a poorly understood component of total soil organic carbon (SOC). Although the turnover rate of the recalcitrant fraction is slow, warming temperatures are expected to speed the decomposition of recalcitrant SOC resulting in an increase of atmospheric CO₂ in the future. Several studies show that the oldest SOC is associated with the smallest mineral particles (clays), making direct spectroscopic analysis of old carbon difficult. To overcome the difficulty of analyzing natural samples, we created synthetic soils to examine the association between clay surfaces and specific biomolecules based on the hypothesis that clays with higher surface charge will more strongly bond organic molecules, and also that certain molecules will be better stabilized by clay. We used kaolinite, montmorillonite, or quartz (sand) as a synthetic soil inside 12 mL septum-capped vials, added either dissolved glucose or vanillic acid to each mineral, inoculated with soil microbes, and then purged the vials with a CO₂-free atmosphere. We incubated them and measured the concentration and [delta]¹³C of CO₂ that accumulated in the vials. Respiration rates were significantly higher in experiments containing vanillic acid than in those containing glucose. Respiration rates were lowest in experiments containing montmorillonite. We repeated the experiment using dilute H₂O₂ as an oxidant, and adding vanillic acid, glucose, or glycine. Vials with montmorillonite showed lower rates of CO₂ accumulation than kaolinite, and both glycine- and glucose-containing experiments had less CO₂ than vanillic acid-experiments. We conclude that the montmorillonite protected the organic matter from oxidation better than sand or kaolinite. Both clays protected organic matter better than sand. In all experiments with clay, the respired CO₂ had lower [delta]¹³C values than bulk substrate. This carbon isotope fractionation is likely due to preferential desorption, followed by oxidation, of 12C- as opposed to 13C- bearing organic molecules. The mineral-organic interaction is a strong bond that explains the old age of labile organic compounds in soils. These results indicate that the clay fraction of soils must be considered for accurate prediction of future land-atmosphere carbon fluxes. / text