November 12, 2014 Palo Alto, California BPSM Petroleum Generation Kinetics: Single- Versus Multiple-Heating Ramp Open-System Pyrolysis K.E. Peters 1,2 and A.K. Burnham 3 1 Schlumberger Information Solutions, Mill Valley, CA 94941; kpeters2@slb.com 2 Dept. Geology & Environmental Sciences, Stanford University, Stanford, CA 94305 3 American Shale Oil Corporation, LLC, Livermore, CA 94550 Acknowledgement Clifford C. Walters (ExxonMobil)
Detector Signal Fraction Transformation Ratio Discrete Activation Energy Models : One Frequency Factor k = Ae -E a /RT k = Arrhenius rate constant (kerogen to oil and gas ) A = frequency factor (e.g., vibrational frequency of bonds broken) E a = activation energy, R = gas constant, T = temperature Laboratory Pyrolysis Calculated Experimental 50 C/min Optimization A = 1 x 10 14 sec -1 Geologic Conditions 1.0 0.8 3 o C/my 5 C/min 30 C/min 10 C/min 0.6 0.4 50% TR 3 C/min 1 C/min 0.2 300 350 400 450 500 550 600 Temperature ( C) 41 43 45 47 49 51 53 55 57 59 61 63 65 E a (kcal/mol) 0 50 100 150 200 Temperature ( o C)
Recent Papers Recommend Single-Ramp Kinetics Single-ramp kinetics (Waples et al., 2002, 2010), Waples and Nowaczewski (2014) use a fixed, universal A. Single-ramp is faster and cheaper than multiple-ramp kinetics and can be used on archived pyrolysis data. Multiple-ramp kinetics optimize both E a and A: Pyromat II ramps = 1, 3, 5, 10, 30, and 50 o C/min 1906
Transformation Ratio Purpose of the Kinetic Study Compare reliability of various combinations of opensystem pyrolysis ramps to determine the kinetics of petroleum generation for 52 global source rocks. Is single-ramp kinetics using a fixed A (1 x 10 14 sec -1 ) more reliable than multiple-ramp kinetics where both E a and A are optimized? Geologic Conditions 1.0 Laboratory Pyrolysis Optimization 0.8 90% TR Pyromat II Micropyrolysis Kinetics05 Software 0.6 0.4 0.2 0 3 o C/my 50% TR 10% TR 50 100 150 200 Temperature ( o C)
Assessment of Single-Ramp Kinetics Involves Three Factors 1) Are there real differences in frequency factors (A) for petroleum generation from kerogen in source rocks? 2) Are the differences in A large enough to significantly impact extrapolation of temperatures to geologic conditions (e.g., at different transformation ratios)? 3) What experimental conditions are required to answer questions 1 and 2?
Arrhenius Equation Expressed Relative to Reference Values k = Ae -E a /RT k 1 = A 1 e -E 1 /RT k 2 = A 2 e -E 2/RT Pick T so that k 1 = k 2 A 1 e -E /RT 1 = A 2 e -E /RT 2 Ln A 1 /A 2 = (E 1 E 2 )/RT Log A Log A ref = (E a E ref )/2.303RT
Compensation Law: Lab Predictions Deviate at Geologic Heating Rates Log A Log A ref 2 1 0-1 Loci of solutions with equal residual error TR 50 = 132 o C TR 50 = 152 o C E ref = 50 kcal/mol A ref = 1 x 10 14 sec -1 3 o C/my -2 Modified from Burnham (1992) Copyright 2001-2011 NExT. All rights reserved TR 50 = 109 o C 8-6 -4-2 0 2 4 6 8 10 E a - E ref, kcal/mol
1 Kcal/mol Error in E a is Compensated by 2-Fold Change in A Log A Log A ref 2 1 0-1 Loci of solutions with equal residual error Log 2 = 0.3 TR 50 = 132 o C TR 50 = 152 o C E ref = 50 kcal/mol A ref = 1 x 10 14 sec -1 3 o C/my -2 Modified from Burnham (1992) Copyright 2001-2011 NExT. All rights reserved TR 50 = 109 o C 8-6 -4-2 0 2 4 6 8 10 E a - E ref, kcal/mol
Errors in Predicted Geologic Temperature Can be Substantial Log A Log A ref 2 1 0-1 Loci of solutions with equal residual error Log 4 = 0.6 Log 2 = 0.3 TR 50 = 132 o C TR 50 = 152 o C 3 o C/my -2 Modified from Burnham (1992) Copyright 2001-2011 NExT. All rights reserved TR 50 = 109 o C 8-6 -4-2 0 2 4 6 8 10 E a - E ref, kcal/mol
1 Kcal/mol Error in E a is Compensated by 2-Fold Change in A Log A Log A ref 2 1 0-1 Loci of solutions with equal residual error TR 50 = 132 o C TR 50 = 152 o C 3 o C/my -2 Modified from Burnham (1992) Copyright 2001-2011 NExT. All rights reserved TR 50 = 109 o C 8-6 -4-2 0 2 4 6 8 10 E a - E ref, kcal/mol
Fixed A Introduces Error in Geologic T Extrapolation (~20 o C) 1-2-3 Rule: 1 kcal/mol error doubles A and yields ~3 o C error in geologic extrapolation of temperature Range A for 52 kerogens = 10 12 to 10 16 sec -1 Assume a fixed A of 1 x 10 14 sec -1 10 14 /10 12 = 100 Log 2 100 = 6.65 (i.e., A doubles 6.65 times) 6.65 x 3 o C/my ~ 20 o C error
Polarization Electrode Coinjector Electrode 52 Samples Were Analyzed by Pyromat II Micropyrolysis Pyromat II Temperature Programmed Oven Thermocouple Open Position 10 cm High Temperature Flame Ionization Detector Helium Flow Designed and Built by Lab Instruments & LLNL Closed Position 1 cm Quartz Crucible 3-10 mg Rock or Kerogen To FID Dynaseal O-ring Seal Thermocouple Inserted Quartz Wool
16 Single-Ramp Replicates Give Best E a of ±0.28 Kcal/mole Bellagio Road outcrop (type II) A fixed = 1 x 10 14 sec -1 Geologic Extrapolation Assuming 3 o C/my T max, o C Mean E a, Temp o C Temp o C Temp o C kcal/mole at 10% TR at 50% TR at 90% TR Average 449.3 53.54 112.0 137.3 163.7 Minimum 447.8 52.97 105.2 135.8 160.4 Maximum 452.1 53.87 115.0 138.4 168.2 Std. Dev. 1.3 0.28 2.3 0.8 2.2 TR = transformation ratio (extent of conversion of kerogen to petroleum)
Log A (Single Run Multi Run) Fixed vs. Optimized A: Offset of E a by Compensation Law 2.0 1.5 1.0 Type I Type II Type IIS Type II/III Type III y = 0.3067x + 0.0459 R² = 0.9943 0.5 Log 2 = 0.3 0.0-0.5-1.0 52 Global Source Rocks -1.5-4 -3-2 -1 0 1 2 3 4 5 E a Mean (Single run Multi Run)
Temp. ( C) at 50% TR (Single Ramp) Single- and Multi-Ramp Models Yield Different Temperatures 170 165 160 155 150 Type I Type II Type IIS Type II/III Type III 145 140 135 130 3 o C/my 125 120 120 130 140 150 160 170 Temp. ( C) at 50% TR (Multiple Ramp)
Heating-Rate Ratio (R r ) = Maximum / Minimum Ramp Pyromat II Ramps = 1, 3, 5, 10, 30, and 50 o C/min Therefore, R r of 1 is a single-ramp experiment (fixed A). R r of 50 consists of all 50/1 multi-ramp experiments (optimized A): 1,50 1,3,50 1,5,50 1,10,50 1,3,5,50 1,5,10,50 1,10,30,50 1,3,5,10,30,50 o C/min
Deviation from Average E a (kcal/mol) Single- vs. Multi-Ramp: Wide vs. Narrow Deviation in E a 10 6 Kimmeridge Clay Monterey Shale 2-2 3σ Single-Ramp -6-10 Multiple-Ramp 0 10 20 30 40 50 60 Heating-Rate Ratio (R r )
Deviation from Average E a (kcal/mol) Variation in E a Becomes Small for Heating-Rate Ratios >16 10 6 Kimmeridge Clay Monterey Shale 2-2 3σ Single-Ramp -6-10 Multiple-Ramp 0 10 20 30 40 50 60 Heating-Rate Ratio (R r )
Activation Energy (kcal/mol) Differing A and E a are Real and Not Measurement Artifacts 56 54 Kimmeridge (all heating rates) 52 50 Monterey (all heating rates) R r >16 48 1 x 10 12 1 x 10 13 1 x 10 14 Frequency Factor (sec -1 ) 1 x 10 15
Temperature ( o C) Kimmeridge Clay: Predicted T at Transformation Ratio (TR) 150 145 140 135 130 125 120 115 110 165 155 10% 1,3,5,10,30,50 10,50 10,30,50 1 2 3 4 5 6 50% 150 145 140 135 130 125 120 115 110 165 155 0 10 20 30 40 50 10% 50% 145 135 125 190 180 170 1,3,5,10,30,50 10,50 10,30,50 1 2 3 4 5 6 90% 145 135 125 0 10 20 30 40 50 190 180 170 90% 160 1,3,5,10,30,50 160 150 10,50 10,30,50 150 140 1 2 3 4 5 6 Number of Heating Rates 140 0 10 20 30 40 50 Heating-Rate Ratio (R r )
Temperature ( o C) Monterey Shale: Predicted T and Transformation Ratio (TR) 125 120 115 110 105 100 95 90 85 150 140 130 120 110 100 90 175 170 165 160 155 150 145 140 1,3,5,10,30,50 10,50 10,30,50 1 2 3 4 5 6 1,3,5,10,30,50 10,30,50 30,50 1 2 3 4 5 6 1,3,5,10,30,50 10,50 10,30,50 1 2 3 4 5 6 10% 50% 90% 125 120 115 110 105 100 95 90 85 150 0 10 20 30 40 50 140 130 120 110 100 90 0 10 20 30 40 50 175 170 165 160 155 150 145 140 0 10 20 30 40 50 Number of Heating Rates Heating-Rate Ratio (R r ) 10% 50% 90%
Conclusions (I) Single-ramp pyrolysis can yield kinetic results that are inconsistent with those from multiple-ramp experiments. Frequency factors for 52 global source rocks are statistically distinct within measurement uncertainty and vary over four orders of magnitude (10 12 to 10 16 sec -1 ). Assuming a universal value for A erroneously presumes that temperature measurements do not have sufficient accuracy for reliable kinetics. Adoption of fixed A of 1 x 10 14 sec -1 can result in error in geologic temperature extrapolation of up to ~20 o C.
Conclusions (II) Pyrolysis ramps of 30-50 o C/min can be too fast for good kinetic fit because of thermal lag; minimize sample and thermocouple size, optimize thermocouple orientation. Heating rate x sample size should be <100mg o C/min. 20- to 30-fold variation in heating rate using at least three ramps is recommended (e.g., 1, 5, 25 o C/min or 1, 3,10, 25 o C/min) with replicates at highest and lowest rates. Neither single- nor multiple- ramp discrete E a distribution models are reliable for kerogens with narrow E a ranges where nucleation-growth models are needed.
References Braun, R.L., A.K. Burnham, J.G. Reynolds, and J.E. Clarkson, 1991. Pyrolysis kinetics for lacustrine and marine source rocks by programmed micropyrolysis: Energy & Fuels, v. 5, p. 192-204. Waples, D.W., and V.S. Nowaczewski, 2014. Source-rock kinetics, in Encyclopedia of Petroleum Geoscience, New York, Springer. Waples, D.W., A. Vera, and J. Pacheco, 2002. A new method for kinetic analysis of source rocks: development and application as a thermal and organic facies indicator in the Tithonian of the Gulf of Campeche, Mexico: Abstracts, 8th Latin American Congress on Organic Geochemistry, Cartagena, p. 296-298. Waples, D.W., J.E. Leonard, R. Coskey, S. Safwat, and R. Nagdy, 2010. A new method for obtaining personalized kinetics from archived Rock-Eval data, applied to the Bakken Formation, Williston Basin. Abstract, AAPG Annual Convention, Calgary.
BPSM Thanks You for Your Support! Before After
Stanford BPSM Industrial Affiliates Program BPSM = Industrial affiliates program & research group, 6+ years 4 Stanford faculty, 3 associated scientists on campus, 3 in Germany, 7 graduate students, 13 affiliated companies http://bpsm.stanford.edu