The evaporation of crude oil and petroleum products
ABSTRACT The physics of oil and petroleum evaporation are investigated. Literature on oil spill evaporation shows that most workers use boundary-layer equations adapted from water evaporation work. These equations predict a constant evaporation mass-transfer rate, dependent on scale size and wind speed. Evaporation was studied further by measuring evaporation of commercial oil products. An experimental apparatus for the study of evaporation was developed. Evaporation was determined by weight loss measured on a balance and recorded constantly on a computer. Examination of the data shows that most oil and petroleum products evaporate at a logarithmic rate with respect to time. This is attributed to the overall logarithmic appearance of many components evaporating at different linear rates. Petroleum products with fewer chemical components such as diesel fuel, evaporate at a rate which is square root with respect to time. The particular behaviour is shown to be a result of the number of components evaporating. Oils with greater than seven to ten components can be predicted with logarithmic equations, those with three to seven components, with square root equations. Evaporation of oils and petroleum products is not strictly boundary-layer regulated. This is largely a result of the high saturation concentrations of oil components in air, which is associated with a high boundary-layer regulated rate. Typical oil evaporation rates do not exceed that of molecular-diffusion, and thus turbulent diffusion does not increase the evaporation rates. Some volatile oils and petroleum products show some effect of boundary-layer regulation at the start of the evaporation process, but after several minutes, evaporation slows because of the loss of the more volatile components, at which point evaporation ceases to be boundary-layer regulated. Overall, boundary-layer regulation can be ignored in the prediction of oil and petroleum evaporation. A simple equation relating only the logarithm of time (or square root of time for narrow-cut products) and temperature can accurately describe oil evaporation. Methods to calculate the constants for the equation using only conventional distillation data are described. Empirical and calculated evaporation equations for several common world crude oils are given.
- [Show abstract] [Hide abstract]
ABSTRACT: The oil lakes in southern Kuwait have accumulated oil that was spilled during the 1991 hostilities and have been exposed to processes of surface degradation for the last decade. Long-chain normal alkanes are abundant in these oils, showing that biodegradation has not been a major factor, whereas the absence of most compounds with less than 10 carbon atoms suggests a significant role for evaporation. Evaporation has been simulated in laboratory experiments at 25, 30, 40, and 50°C and compositional changes monitored by taking samples periodically for gas chromatographic analysis. Evaporation is initially rapid but slows through time, and ultimately, most compounds with less than 10 carbon atoms are lost. For a given carbon number, loss proceeds in the sequence: normal alkanes> branched chains> aromatics (which is the opposite of water washing). GCMS data for hopane and sterane biomarkers confirm that oil in the lakes is derived from the giant Burgan oilfield and also show that these biomarkers are not affected by evaporation in either nature or the laboratory. Analysis of sulfur compounds using GC-FPD showed that oils exposed in the lakes were photo-oxidized and had reduced concentrations of benzothiophenes and increased volatile sulfur compounds. A loss of volatile hydrocarbons from the free surface leads to compositional layering unless the oil is well mixed by convection or diffusion. In experiments to monitor the development of layering, low molecular weight compounds were rapidly lost from the surface, and a steep compositional gradient developed. The formation of a devolatilized, viscous surface “skin” tends to make evaporation a self-limiting process and also has significance for the design of sampling protocols in environmental forensics.Environmental Geosciences 03/2002; 9(1):8-16. DOI:10.1046/j.1526-0984.2002.91005.x