Thermal Evolution of Crude Oils in Sedimentary Basins: Experimental Simulation in a Confined System and Kinetic Modeling

A detailed knowledge of the cracking mechanisms of crude oils should highly improve the understanding of geochemical reactions involved in hydrocarbon degradation into lighter oil and gas and consequently the applicability of kinetic models currently used for prediction of gas formation. Although the mechanisms of cracking are well known for several model compounds or simple mixtures, there is, to date, no available method to model complex mixtures, especially when they contain heavy compounds, except by using empirical approaches. During thermal cracking, oil will produce both lighter and heavier molecules than those present in the initial sample. Thus, the pyrolysate will be a mixture of both new compounds and compounds not yet degraded. In order to discriminate between reactants and products, we have chosen to fractionate each oil into two classes : the first one (distillate 300-) comprising light hydrocarbons ranging from C6 to C16 the second one (residue 300+) comprising both hydrocarbons and polar compounds. For simulation of thermal evolution of crude oils, about 100 experiments were carried out on two oils (Boscan and Pematang), in a closed pyrolysis system, over a wide range of heating times (few minutes to 1 month) and temperatures (335 to 500°C). The pyrolysate is represented by 10 chemical fractions (C1, C2, C3-C5, C6-C13 saturates, benzene + toluene + xylenes + naphthalene, C9-C13 alkyl aromatics, C14+ saturates, C14+ condensed aromatics, C14+ alkyl and/or naphtheno aromatics + resins + asphaltenes and coke). For kinetic modeling, the degradation of each fraction, except for C1, C6-C13 aromatic mixture comprising benzene, toluene, xylenes and naphthalene, and coke considered as stable compounds, is described by a balanced elementary reaction governed by first order kinetics and obeying Arrhenius law. For a given oil, the kinetic parameters of the model (apparent activation energies Ei, preexponential factor A and stoichiometric coefficients alphaij) were optimized with reference to a dataset comprising the final compositions resulting from oil pyrolysis experiments. This optimization was carried out successively on each distillate and each residue, on both distillate and residue from the same oil, then on the two distillates and two residues altogether. The model accounts satisfactorily for the specific kinetics of cracking of both each oil and of the two oils together, suggesting that it is now possible to predict the thermal behavior of a given oil provided the proportions of the ten fractions selected for the model are known. The wide temperature range in which the model is valid, supports its extrapolation to the temperature range of sedimentary basins.