452 Energy zyxwvuts & Fuels 1987,1, zyxwvu 452-458 Articles Comparison of Methods for Measuring Kerogen Pyrolysis Rates and Fitting Kinetic Parameterst Alan K. Burnham," Robert L. Braun, and Hugh R. Gregg Lawrence Livermore National Laboratory, Livermore, California zyxw 94550 Alain M. Samoun Lab Instruments, Inc., Kenwood, California 95452 Received April 20, 1987 We determine rates of product evolution during pyrolysis of several petroleum source rocks and isolated kerogens by nonisothermal techniques, including Rock Eval pyrolysis and pyrolysis-MS/MS, The resulting data are analyzed by nonlinear regression and simpler correlation techniques in terms of discrete and Gaussian-distributed activation energy models. We find that temperatures measured by standard Rock Eval analysis are too low by about 40 zyxwv "C, resulting in kinetic expressions that are much too fast. Proper temperature calibration eliminates this problem. We explore the sensitivity of the kinetic parameters and extrapolation to geologic heating rates to uncertainty in the temperature calibration. We find that the discrete distribution model provides a superior fit to the laboratory data and probably a more reliable extrapolation to geological heating rates. We also assess how differences among kinetics for individual species relate to the activation energy distribution required for total hydrocarbon evolution. Kinetics from Rock Eval pyrolysis predict hydrocarbon generation rates intermediate between kerogen decomposition and oil expulsion rates during hydrous pyrolysis, but slight differences in activation energies result in similar predictions for a geological heating rate. Predictions of petroleum generation temperatures for lacustrine source rocks cover almost the same range as for marine source rocks. Introduction Although the concept that petroleum formation is a kinetic process is well established, determining kinetic parameters for quantitative prediction has been elusive because of the complexity of the maturation process. Many experimental procedures and kinetic models have been tried.l Programmed microscale pyrolysis (e.g., Rock Eva12)and sealed-bomb hydrous pyrolysis3 are two cur- rently popular experimental techniques, but they differ markedly in temperature-pressure history as well as the time that products have to undergo secondary reactions. Moreover, programmed pyrolysis typically measures total volatile hydrocarbons while hydrous pyrolysis measures expelled (floating)oil and bitumen (unexpelled extractable material), so it is not obvious how the kinetics from the two different experiments should relate. Hydrous pyrolysis would appear to be a better simulation, but the Rock Eval technique is much faster and cheaper. There are additional complications in choosing how to analyze the kinetic data. Lewan3has derived reasonable kinetic parameters for expelled oil from Phosphoria and Woodford shale assuming a simple first-order reaction. In contrast, other workers have found that an activation en- ergy distribution is needed for some aspects of product evolution from coal,4Green River oil shale: and isolated Work performed under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under Contract No. W-7405-ENG-48. 0887-0624/87/2501-0452$01.50/0 kerogens of various types.2*6 We have shown that the effective activation energy derived by assuming no dis- tribution is strongly influenced by the distribution if ac- tually present, and ignoring even a small distribution would cause problems in extrapolating the rate to geologic tem- peratures.' However, the mathematical complexity and computer requirements for treating activation energy distributions have limited their use. In this paper, we have started a systematic comparison of different techniques for measuring and deriving product formation kinetics. This work includes writing an easy- to-use computer program for determining either Gaussian or discrete activation energy distributions. At this stage, most of our kinetic results are from the Rock Eval I1 in- strument, which is commonly used in the petroleum in- dustry. However, initial pyrolysis-MS/MS experiments and both p ~ b l i s h e d ~ ~ ~ ~ ~ and unpublishedlOJ1 kinetic data ~~ (1) Two good reviews, by D. W. Waples and B. Horsfield, are con- tained in: Aduances in Petroleum Geochemistry, Brooks, J., Web, D., Eds.; Academic: New York, 1984; Vol. l., pp 7-68, 247-298. (2) Ungerer, P.; Espitalie, J.; Marquis, F.; Durand, B. In Thermal Modeling in Sedimentary Basins; Burrus, J., Ed.; Technip: Paris, 1986; pp 531-546. (3) Lewan, M. D. Philos. Tram. R. SOC. London, A 1985,315,123-134. (4) Howard, J. B. In Chemistry of Coal Utilization; Elliot, M. A., Ed.; (5) Campbell, J. H.; Gallegos, G.; Gregg, M. Fuel 1980, 59, 727-732. (6) Tissot, B.; Espitalie, J. Rev. Inst. Fr. Pet. 1975, 30, 743-777. (7) Braun, R. L.; Burnham, A. K. Energy Fuels 1987, 1, 153-161. (8) Campbell, J. H.; Koskinas, G. J.; Stout, N. D. Fuel 1978, 57, Wiley: New York, 1981; 2nd Suppl Vol.; pp 665-784. 372-376. 0 1987 American Chemical Society