Contents lists available at ScienceDirect Fuel journal homepage: www.elsevier.com/locate/fuel Full Length Article Petroleum generation kinetic models for Late Ordovician kukersite Yeoman Formation source rocks, Williston Basin (southern Saskatchewan), Canada Zhuoheng Chen a, ⁎ , Xiaojun Liu a , Kirk G. Osadetz a,b a Geological Survey of Canada, Calgary, Canada b CMC Research Institutes Inc., Canada ARTICLE INFO Keywords: Kukersite, generation kinetics Parallel nucleation-growth reaction model ABSTRACT Ordovician Yeoman Fm. kukersite source rocks from Canadian Williston Basin are composed almost exclusively of Gloeocapsomorpha prisca (G. prisca) alginite. Thermocatalytic petroleum generation from G. prisca alginite differs significantly from that of amorphous bituminite typical of marine Type II source rocks. Commonly used petroleum generation kinetic parameter optimization procedures that assume n th order chemical reactions fail to reproduce sample Flame Ionization Detector (FID) pyrograms using expected chemical bond activation energies. A parallel nucleation-growth reaction model (PN-GRM) successfully addresses these deficiencies for this specific kerogen type. Programed pyrolysis of seventeen kukersite sample FID pyrograms as well as two additional kukersite Rock-Eval datasets reveal the kinetic characteristics of this globally significant, but stratigraphically restricted marine Type I source rock. The results show that the PN-GRM closely approximates the chemical reactions as demonstrated by reproduction of kukersite FID pyrograms, that kukersite source rocks are thermally more stable as indicated by elevated petroleum generation onset temperatures, and that compositionally simple and homogeneous source rocks, such as kukersites, typically exhibit a sharply increasing petroleum generation rate and a narrow oil window both in nature and in pyrolysis experiments. 1. Introduction Petroleum system analysis and modeling [1] constitutes a common risk-reduction and resource appraisal activity employed by petroleum explorers in conventional and unconventional “wildcat” exploration programs [2]. Kinetic petroleum generation models of source rock kerogen transformation identify the rates, composition and timing of petroleum generation are key components of this analysis when com- bined with the inferred thermal history (e.g. [3]). Such models also identify the “critical moment” that controls petroleum system pro- spectivity ([1], their Figure 1.5). Natural petroleum generation is commonly simulated using the analysis and kinetic modeling of pyrolytic “petroleum” generation data, which is essentially identical to source rock kerogen transformation data. The analysis and modeling are distinct activities. The laboratory experiments on source rock samples, commonly using programmed anhydrous pyrolysis, yields rates and amounts of hydrocarbon and hydrocarbon-like compounds generated by progressive kerogen thermal transformation. Programmed anhydrous pyrolysis in laboratory conditions, such as Rock-Eval analysis, mimics but is not identical to natural petroleum generation from source kerogen with respect to the experimental environment, composition of evolved products, and ex- perimental timescale, and most notably because it is an anhydrous process and uses high temperature to compensate a compressed time scale. Numerical models are then “fit” to laboratory kerogen transfor- mation results that are, for predictive purposes, extrapolated to dura- tions and temperatures characteristic of natural geological petroleum generation processes [4,5]. The “fit” of sets of numerical reactions to laboratory results is a first and crucial step for developing a numerical, or “kinetic”, petroleum generation model [6–8]. Successful petroleum generation model parameters derived from laboratory data should also be consistent with natural petroleum generation reactions and their characteristics. Thus, in searching for suitable numerical model para- meters it is possible to “over”, or “perfectly” fit the laboratory data with model parameters that may not be representative of the chemical re- actions and geological processes that generate petroleum naturally [9]. Therefore, kinetics from laboratory data should be carefully validated against natural data on petroleum generation. It is common to model petroleum generation kinetics using a series of parallel first-order chemical reactions [10]. This method may not always be appropriate [11,12], such as in the specific case of kukersite source rocks that are discussed herein. Kukersite is a unique and https://doi.org/10.1016/j.fuel.2018.11.154 Received 12 September 2018; Received in revised form 27 November 2018; Accepted 30 November 2018 ⁎ Corresponding author. E-mail address: zhuoheng.chen@canada.ca (Z. Chen). Fuel 241 (2019) 234–246 0016-2361/ Crown Copyright © 2018 Published by Elsevier Ltd. All rights reserved. T