Advances in OrganicGeochemistry 1987 Org. Geochem. Vol. 13, Nos 4-6, pp. 839-845, 1988 Printed in Great Britain 0146-6380/88 $3.00+ 0.00 Pergamon Press plc Further comparison of methods for measuring kerogen pyrolysis rates and fitting kinetic parameters* ALAN K. BURNHAMIt, ROBERT L. BRAUN l and ALAIN M. SAMOUN 2 ~Lawrence Livermore National Laboratory, L-207, P.O. Box 808, Livermore, CA 94550, U.S.A. 2Lab Instruments, Inc., P. O. Box 60, Kenwood, CA 95452, U.S.A. Abstract--We compare rates of product generation during pyrolysis of several petroleum source rocks and isolated kerogens by nonisothermal techniques, including Rock-Eval pyrolysis, condensed oil evolution, and pyrolysis MS/MS. We discuss problems related to temperature calibration in the Rock-Eval instrument, and confirm that standard Rock-Eval temperatures are in error by about 40°C. Calculations using Gaussian and discrete activation energy distributions and associated frequency factors derived from Rock-Eval data by nonlinear regression agree within a few °C with oil evolution at 2°C/min and pyrolysis-MS/MS at 4°C/min. This comparison also demonstrates that activation energy distributions are needed for oil alone as well as for gas and for total hydrocarbons from many source rocks. Discrepancies between Rock-Eval kinetics and oil evolution at 2°C/h and hydrous pyrolysis are related to mass transport limitations. Finally, we discuss the sensitivity of the extrapolation of reaction rates to geological conditions and how the extrapolation may be influenced by erroneous data or kinetic analysis. Key words: kerogen pyrolysis, pyrolysis kinetics, Rock-Eval, pyrolysis-MS/MS, hydrous pyrolysis, petroleum generation INTRODUCTION The efficiency of petroleum exploration could be enhanced if geothermal histories could be combined with chemical kinetics to accurately calculate the timing of oil and gas generation. Earlier results (Lewan, 1985; Burnham et al., 1987) indicate a significant variation in the degree of thermal stress required to decompose different kerogens of the same type as defined by Tissot et al. (1974), so it is necessary to have a reasonably simple procedure for determining the generation kinetics. Oil evolution (Campbell et al., 1978), pyrolysis mass spectrometry (Campbell et al., 1980), programmed micropyrolysis (Ungerer and Pelet, 1987; Burnham et al., 1987), and sealed-bomb hydrous pyrolysis (Lewan, 1985) have been used to derive generation kinetics. However, to the limited extent that quantitative comparisons can be made, hydrous pyrolysis kinetics differ from those derived by other techniques. Given this discrepancy and that the techniques differ greatly in required effort, it is not yet clear what the easiest procedure is for determining valid and reliable parameters. This uncertainty is compounded by the fundamental uncertainty in use of the Ar- rhenius equation over wide temperature ranges and by the scarcity of nonproprietary, adequately docu- mented comparisons between predicted and calcu- *Work performed under the auspices of the U.S. De- partment of Energy by the Lawrence Livermore National Laboratory under contract number W-7405- ENG-48. tAuthor to whom correspondence should be addressed. O.G.13/4-6~R 839 lated oil generation temperatures. Our study of the Uinta Basin (Sweeney et al., 1987) is one of the few where all kinetic and geological parameters are pro- vided. Finally, there is a question about whether kinetic measurements should be made on whole rock samples or isolated kerogens. It is not yet clear if either alteration of the organic matter during mineral dissolution or unnatural catalytic mineral effects in laboratory experiments are serious problems. The present paper compares kinetic expressions from different laboratory experiments and begins to assess why different techniques give different results. Besides demonstrating the effects of inaccurate tem- perature measurements and improper data analysis, we uncover additional concerns about the relative contributions of chemical and mass and heat transfer processes to the observed generation rates. METHODS The shale samples used are the same as those used previously (Burnham et al., 1987). A short tabulation of properties is given in Table 1. Sample name abbreviations are New Albany-Kentucky (NAKY), Green River Anvil Points (AP22), Fushun (FSHN), Table 1. Properties of samples examined in this paper. Compositions are in % by wt, and the ratio is molar Sample TOC min C tot H tot S org S/org C AP22 9.9 6.1 1.4 0.3 0.006 NAKY 11.1 0.3 1.4 4.4 0.016 KIMR 6.2 0.2 1.5 3.3 0.021 PHOS 16.6 0.2 2.2 1.8 0.023 WDFRD 15.4 0.2 1.6 3.3 0.019 FSHN 11.0 0.9 2.0 0.7 0.008