Transport and biodegradation of petroleum hydrocarbon vapors in the subsurface: a laboratory soil column study

In this work, the vapor transport and aerobic bio-attenuation of compounds from a multi-component petroleum vapor mixture were studied for six idealized lithologies in 1.8-m tall laboratory soil columns. Columns representing different geological settings were prepared using 20-40 mesh sand (medium-grained) and 16-minus mesh crushed granite (fine-grained). The contaminant vapor source was a liquid composed of twelve petroleum hydrocarbons common in weathered gasoline. It was placed in a chamber at the bottom of each column and the vapors diffused upward through the soil to the top where they were swept away with humidified gas. The experiment was conducted in three phases: i) nitrogen sweep gas; ii) air sweep gas; iii) vapor source concentrations decreased by ten times from the original concentrations and under air sweep gas. Oxygen, carbon dioxide and hydrocarbon concentrations were monitored over time. The data allowed determination of times to reach steady conditions, effluent mass emissions and concentration profiles. Times to reach near-steady conditions were consistent with theory and chemical-specific properties. First-order degradation rates were highest for straight-chain alkanes and aromatic hydrocarbons. Normalized effluent mass emissions were lower for lower source concentration and aerobic conditions. At the end of the study, soil core samples were taken every 6 in. Soil moisture content analyses showed that water had redistributed in the soil during the experiment. The soil at the bottom of the columns generally had higher moisture contents than initial values, and soil at the top had lower moisture contents. Profiles of the number of colony forming units of hydrocarbon-utilizing bacteria/g-soil indicated that the highest concentrations of degraders were located at the vertical intervals where maximum degradation activity was suggested by CO2 profiles. Finally, the near-steady conditions of each phase of the study were simulated using a three-dimensional transient numerical model. The model was fit to the Phase I data by adjusting soil properties, and then fit to Phase III data to obtain compound-specific first-order biodegradation rate constants ranging from 0.0 to 5.7x103 d-1.