Petroleum Source Rock Identification of United Kingdom Atlantic Margin Oil

Earth and Planetary Science Letters 313–314 (2012) 95–104

Contents lists available at SciVerse ScienceDirect

Earth and Planetary Science Letters

journal homepage: www.elsevier.com/locate/epsl

Petroleum source rock identiﬁcation of United Kingdom Atlantic Margin oil ﬁelds and the Western Canadian Oil Sands using Platinum, Palladium, Osmium and Rhenium: Implications for global petroleum systems

Alexander J. Finlay a,⁎, David Selby a, Mark J. Osborne b a Department of Earth Sciences, Durham University, Science Laboratories, Durham, DH1 3LE, UK b BP, Exploration & Production Technology, Building H, Sunbury-on-Thames, Middlesex, TW16 7LN, UK article info abstract

Article history: This study demonstrates that petroleum and source rocks are enriched in Pt and Pd to the ppb level, and that Received 15 June 2011 the 187Os/188Os composition coupled with the Pt/Pd value permits the fingerprinting of petroleum to its Received in revised form 4 November 2011 source. Oils from the United Kingdom Atlantic Margin (sourced from the Upper Jurassic Kimmeridge Clay Accepted 5 November 2011 Fm.) as well as source rock samples have been analysed for Pt and Pd. When the Pt/Pd value is compared Available online xxxx 187 188 with Os/ Os (calculated at the time of oil generation; Osg) the values from both the known source and fi Editor: T.M. Harrison the oils are similar, demonstrating that they can be used as an oil to source ngerprinting tool. This inorganic petroleum fingerprinting tool is particularly important in heavily biodegraded petroleum systems where tra- Keywords: ditional fingerprinting techniques (e.g. biomarkers) are severely hampered, e.g. the world's largest oil sand Platinum deposit, the West Canadian Oil Sands (WCOS). This has caused the source of the WCOS to be hotly debated, Palladium with no present day consensus between inputs from potential source units e.g. Exshaw and Gordondale Fms. Rhenium 187Os/188Os and Pt/Pd fingerprinting of the oil sands shows that the majority of the petroleum have similar Osmium 187Os/188Os and Pt/Pd values, supporting the hypothesis of one principal source. Analysis of the potential Petroleum source rocks establishes that the principal source of the oil sands to be from the Jurassic Gordondale Fm., Source rock with a minor Exshaw Fm. input. Thus, the combination of previously pioneered Re–Os petroleum geochronology with 187Os/188Os and Pt/Pd values of petroleum permits both a temporal and spatial understanding of petroleum systems. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Zhou et al., 2008). However, mass balance calculations rule out a single Exshaw Fm. source because of its limited stratigraphic extent Biodegradation preferentially removes light hydrocarbons from (2–12 m thickness therefore only forming 8.6% of total reserves; petroleum and so can cause traditional oil to source fingerprinting Creaney and Allan, 1990). Furthermore, mass balance calculations techniques to be challenged (e.g. Biomarkers and Carbon isotopes; suggest up to four units from the Western Canadian Sedimentary Peters et al., 2005 and references therein). One such example is the Basin could have formed/contributed to the WCOS. These are, in West Canadian Oil Sands (WCOS), which is the world's largest oil order of importance, the Lower Jurassic Gordondale Fm. (previously sand reserve (27.6 billion m3; Alberta's Energy, 2005; Fig. 1). This the Nordegg Fm.; Asgar-Deen et al., 2004), the Devonian–Mississippian means that, despite decades of research, the source of the WCOS re- Exshaw Fm., and the Middle Triassic Doig and Upper Devonian mains controversial (Allan and Creaney, 1991; Barson et al., 2000; Duvernay Fms. (Allan and Creaney, 1991; Creaney and Allan, 1990, Creaney and Allan, 1990, 1992; Fowler et al., 2001; Higley et al., 1992; Higley et al., 2009). Although recent mathematical models sug- 2009; Riediger et al., 2000; Zhou et al., 2008). Although the WCOS gest the Gordondale Fm. to be the dominant source of the WCOS, it are biodegraded (biomarker biodegradation scale of between 4 has been debated whether oil volumes generated from the Gordondale [Peace River] and 8 [Athabasca]; Peters et al., 2005 and references Fm. could have been sufficient to have migrated ~500 km to the WCOS therein; Zhou et al., 2008; Fig. 1) it is hypothesised that biomarker reservoirs because migration would have been hampered by strati- preservation is sufficient to propose the Exshaw Fm. as the principal graphic barriers (Riediger, 1994). source (Barson et al., 2000; Fowler et al., 2001; Riediger et al., 2000; To identify the source of the WCOS this study utilises organophyllic Rhenium (Re), Osmium (Os), Platinum (Pt) and Palladium (Pd). We demonstrate that, like Re and Os, shales and asphaltene are enriched ⁎ Corresponding author. in Pt and Pd relative to upper continental crust (UCC). We then demon- 187 188 E-mail address: a.j.fi[email protected] (A.J. Finlay). strate that the Os/ Os composition at the time of generation (Osg)

Alberta

120° 118° 116° 114° 112° 110° 5° W 4° W 3° W 2° W 58° Flett Ridge Athabasca Westray Clair 57° Peace Ridge River Grosmont Cuillin Rona Ridge 56° Wabasca Foinaven Schiehallion tform 55° Shetland Islands 60° N Cold 54° Lake West Shetland Pla

Edge of foothills Lloydminster Orkney belt Islands Provost Tar Sand 59° N Heavy Oil Black = Overmature Gordondale sample Grey = Mature Exshaw sample White = Immature

Period Geological event Tectonic event Period Geological event Tectonic event

North Oil Generation Cordilleran Atlantic (~112 Ma Re-Os rifting Cret age) Orogenic Event

Foreland Palaeogene Neogene Basin Jurassic source become mature Central Laramide (~68 Ma Re-Os age) Jur. Orogenic Atlantic Gordondale Fm. Event rifting deposition Tri. Perm. Cretaceous Carb. Exshaw Fm. “Pacific” U. deposition Margin Dev. Duvernay Fm. Deep marine source deposition M. U. Jur. unit deposition Dev. (Kimmeridge Clay Fm.)

Fig. 1. Location map and geochronological summary of the West Canadian Oil Sands and United Kingdom Atlantic Margin oil fields. Abbreviations; Sk = Saskatchewan; AB = Al- berta; BC = British Columbia (modified from Finlay et al., 2010 and references therein; Selby and Creaser 2005; Creaser et al., 2002; Riediger et al., 1994). Re–Os oil generation age for West Canadian Oil Sands from Selby and Creaser 2005,Re–Os oil generation age of United Kingdom Atlantic Margin oil from Finlay et al. (2010). A.J. Finlay et al. / Earth and Planetary Science Letters 313–314 (2012) 95–104 97 and Pt/Pd values can be used to fingerprint petroleum to its source unit by Finlay et al., 2010). Furthermore 3 samples of the proven UKAM through the analysis of United Kingdom Atlantic Margin (UKAM) oils Kimmeridge Clay Fm. equivalent shale (Clair oil field, well 205/22-1A; and samples of the proven Upper Jurassic source (Kimmeridge Clay Fig. 1; Finlay et al., 2011) were analysed for PGE and Re–Os abundances Fm.; Finlay et al., 2011 and references therein). Furthermore we demon- as well as Re–Os isotopic compositions (Table 1). strate that this method is unaffected by biodegradation. As a result, this novel inorganic fingerprinting method has widespread application to 2.2. West Canadian Oil Sands global petroleum systems. At present there is no consensus on the source of the WCOS, 2. Materials and methods however, the Upper Devonian Duvernay, Devonian/Mississippian Exshaw and Lower Jurassic Gordondale Fms. have all been proposed 2.1. United Kingdom Atlantic Margin as possible source units (e.g., Higley et al., 2009 and references therein). Therefore, fifteen samples of these units were analysed as part of this The UKAM oils are found in a series Devonian to Miocene rift basins study (Table 1). situated between the Shetland Platform and Faroe Islands (Spencer et Two core samples of Upper Devonian Duvernay Fm. were analysed al., 1999; Carruth, 2003; Fig. 1). Two main phases of rifting occurred for PGE and Re abundances as well as Re–Os isotopic composition during the Early Cretaceous and Late Cretaceous/Paleocene forming (Table 2; Fig. 1). The Duvernay Fm. is comprised of two interbedded local depocentres. Reservoirs are found in Devonian/Carboniferous sed- lithofacies controlled by bottom water oxic/anoxia: bioturbated, iments and fractured, weathered basement (e.g., structural trapped nodular, lime mudstones, deposited under oxic conditions and Clair field; Carruth, 2003) and Jurassic and Paleogene sediments (e.g., unbioturbated bituminous mudstones, deposited under deep water stratigraphically trapped Foinaven and Schiehallion fields; Spencer et anoxic conditions. The bituminous mudstones form the oil source in- al., 1999). The petroleum samples selected for this study are sourced tervals and vary from being mature to the south-west to immature in from the Upper Jurassic Kimmeridge Clay Fm. (Finlay et al., 2011). The the north-east (cf. Creaney and Allan 1990; Cioppa et al., 2002). δ13C values of whole oil, saturate and aromatic fractions are almost The PGE abundances of seven samples of regional Exshaw Fm. pre- identical (0.1–0.2‰ Standard Deviation [S.D.]) as well as being within viously utilised for Re–Os analysis were determined (Creaser et al., the range of North Sea oils sourced from the Kimmeridge Clay Fm. (cf. 2002; Selby and Creaser, 2003; Table 2). The Exshaw Fm. formed Fig. 2 of Finlay et al., 2011). In the same figure it is shown that the iso- during a Devonian/Mississippian marine transgression and varies in prenoid and C27,C28,andC29 sterane values both suggest that the oils thickness from 2 to 16 m (Creaney and Allan 1990; Creaser et al., were sourced from 1 single marine planktonic-bacterial source 2002). The Exshaw Fm. comprises two lithofacies, a stratigraphically with some terrigenous material, as expected from the Kimmeridge lower Devonian/Mississippian organic-rich member (TOCb20%) and Clay Fm. Therefore we are confident, in agreement with previous a stratigraphically higher Mississippian siltstone. The samples ana- studies (Cornford et al., 1998; Rooney et al., 1998; Scotchman et al., lysed in this study are all sourced from the lower organic-rich mem- 1998; Lamers and Carmichael, 1999; Leach et al., 1999; Spencer et al., ber, deposited in low energy anoxic conditions at 361.3±2.4 Ma 1999; Mark et al., 2005; Parnell et al., 2005; Scotchman and Carr, (Creaney and Allan 1990; Selby and Creaser 2003). The Exshaw Fm. 2005; Scotchman et al., 2006; Mark et al., 2008), which the UKAM oils varies in maturity from being overmature in the south-west to imma- analysed in this study are sourced from the Upper Jurassic Kimmeridge ture in the north east of the WCSB (Tmax 421–485 °C Piggott and Clay Fm. that generated oil during the Late Cretaceous-Paleogene Lines, 1992; Creaser et al., 2002). (Finlay et al., 2011; Mark et al., 2005). Six samples of regional Gordondale Fm. (formally “Nordegg” Three UKAM asphaltene samples, previously analysed for Re–Os member; Asgar-deen et al., 2004) were analysed for PGE and Re (Finlay et al., 2011), were analysed for PGE abundances (Clair oil field, abundances as well as Re–Os isotopic composition (Table 2; Fig.1). sample G0123, well 206/8-3A; Foinaven oil field, samples G2075 and The Lower Jurassic Gordondale Fm. was deposited during a major ma- G2763, well 204/24A-1; Fig. 1) and 5 samples of Kimmeridge Clay Fm. rine transgression and varies from 19 to 50 m of stratigraphical thick- from the North Sea Miller field (well 16/8b-a01; analysed for Re–Os ness across the WCSB. It comprises dark brown, finely laminated,

Table 1 United Kingdom Atlantic Margin Re and PGE data.

187 187 c Sample Well Depth Os ± Pt ± Pd ± Re ± Pt/Pd ± Re/ ± Os/ ±Osg ± (m) (ppb) (ppb) (ppb) (ppb) 188Os 188Os

Clair Oil Field KCF 3178 205/22-1A 3178.0 1.383 0.003 4.64 0.27 13.9 4.9 219.8 1.3 0.33 0.12 1059 6 3.057 0.003 1.41 0.01 3178.5 205/22-1A 3178.5 1.376 0.004 4.28 0.25 7.30 2.56 242.1 1.4 0.59 0.21 1217 7 3.463 0.004 1.57 0.01 3179 205/22-1A 3179.0 0.754 0.002 4.60 0.27 7.31 2.56 110.5 0.7 0.63 0.22 976.7 5.9 3.068 0.004 1.55 0.01

Miller Oil Field KCFa AF01-06 16/8b-a01 4738.4 0.856 0.002 2.41 0.14 4.65 1.63 58.30 0.19 0.52 0.18 384.2 1.3 1.433 0.002 1.015 0.004 AF02-06 16/8b-a01 4738.2 1.362 0.003 4.66 0.27 11.5 4.0 84.71 0.27 0.40 0.14 346.3 1.2 1.320 0.002 0.943 0.004 AF03-06 16/8b-a01 4738.0 1.285 0.003 3.91 0.23 8.01 2.80 94.33 0.30 0.49 0.17 417.0 1.4 1.501 0.002 1.050 0.004 AF04-06 16/8b-a01 4737.7 1.697 0.004 4.48 0.26 10.3 3.6 115.3 1.5 0.44 0.16 382.5 4.9 1.417 0.002 1.01 0.01 AF05-06 16/8b-a01 4737.3 0.600 0.002 2.02 0.12 3.73 1.30 50.15 0.16 0.54 0.19 482.8 1.7 1.656 0.003 1.129 0.004

Clair Oil Field Asphalteneb G0123 206/8-3A – 0.227 0.003 11.0 0.6 16.8 6.3 1.47 0.04 0.65 0.25 35.12 1.18 1.055 0.032 1.02 0.05

Foinaven Oil Field Asphalteneb G2075 204/24A-1 – 0.062 0.001 2.12 0.13 3.69 1.38 1.63 0.04 0.57 0.22 142.0 6.5 1.084 0.050 0.92 0.06 G2763 204/24A-1 – 0.055 0.002 4.23 0.25 13.0 4.9 0.74 0.04 0.32 0.12 75.63 6.28 1.340 0.100 1.25 0.14

a Re–Os data from Finlay et al., 2010. b Re–Os data from Finlay et al. (2011). c 187 188 Osg is the Os/ Os calculated at the time of oil generation (68 Ma). 98 A.J. Finlay et al. / Earth and Planetary Science Letters 313 (2012) 95–104

radioactive, calc-mudstones, containing up to 27% TOC. The Gordon- – – – – – – – ± dale Fm. is predominantly mature, only being overmature in the far

south-west and immature in the north-east of the WCSB (Tmax

c – g 415 566 °C; Creaney and Allan, 1990; Riediger et al., 1990; Asgar- 1.399 2.619 2.160 5.043 2.584 3.148 2.732 Deen et al., 2004). The PGE abundances of 24 oils from the WCOS were determined. Three oils were sourced from the Athabasca deposit (wells 15-12- – – – – – – – ±Os 93-13W4, 6-93-10W4, 95-9W4), ﬁve from Cold Lake (wells 1-14- 65-2W4, 28-61-4W4), three from Grosmont (well 14-5-88-19W4), c g

(112 Ma) (60 Ma) three from Lloydminster (wells 03-14-49-27W3, A12-19-50-24-W3, A13-35-49-24W3), three from Peace River (well 4-21-85-18W5), three from Provost (well 5-20-37-1W4) and four from Wabasca (wells 2-24-84-22W4, 14-28-81-22W4; Table 3; Fig 1).

Os ± Os 2.3. Analytical methodology 188

Os/ 2.3.1. Re–Os analysis 187 1.561 0.001 1.24 3.038 0.001 2.71 2.467 0.003 1.76 5.955 0.004 4.16 3.009 0.001 2.19 3.685 0.002 2.64 3.167 0.011 2.32 4.748 0.012 2.243 0.011 5.038 0.024 Previous Re–Os data from Miller oil ﬁeld Kimmeridge Clay Fm. (Finlay et al., 2010), UKAM oils (Finlay et al., 2011) WCSB Exshaw – – – – – – – Fm. (Selby and Creaser, 2003) and WCOS (Selby and Creaser, 2005) are reproduced in Tables 1, 2 and 3. Core samples from the Gordon- Os ±

188 dale (n=6) and Duvernay (n=2) Fms. were analysed for Re–Os

Os/ abundance and isotopic composition (Table 2). Prior to crushing, the 1352 5 187 samples were polished, on a diamond polishing wheel, to remove cut- ting and drilling marks to eliminate any contamination. The ~50–80 g samples (representing ~3 cm of stratigraphy) were dried at 60 °C for ~12 h, broken into chips with no metal contact and powdered in a ce- ramic mill. Rhenium–Osmium isotope analysis was carried out at 1.2 0.4 163.20 1.3 0.5 409.43 1.5 0.5 326.90 2.0 0.7 904.60 1.7 0.6 408.50 1.4 0.5 535.30 1.3 0.4 435.70 Durham University's TOTAL laboratory for source rock geochronology and geochemistry at the Northern Centre for Isotopic and Elemental – fi – – – – – – ± Pt/Pd ± – Tracing (NCIET) using CrO3 H2SO4 digestion. Osmium was puri ed from the CrO3–H2SO4 solution using solvent extraction (CHCl3) and micro-distillation methods. After the removal of Os, 1 ml of the CrO –H SO was mixed with 1 ml of Milli-Q water and reduced (ppb) 3 2 4 6+ 3+ from Cr to Cr by sparging with SO2 in preparation for Re anion exchange chromatography. After chromatography the Re fraction was then further purified by single bead anion extraction (cf. Selby and Creaser, 2003). The purified Re and Os were loaded onto Ni and Pt filaments respectively, and analysed for their isotopic compositions using Negative

). Thermal Ionisation Mass Spectrometry (NTIMS; Creaser et al., 1991; Volkening et al., 1991) on a ThermoElectron (TRITON) mass spectrometer. Re was measured using Faraday collectors and Os in peak hopping 1.44 0.08 1.25 0.44 19.98 3.33 0.20 2.49 0.87 92.53 2.00 0.12 1.34 0.47 42.26 1.53 0.09 0.78 0.27 31.28 2.18 0.13 1.27 0.44 71.18 1.14 0.07 0.81 0.28 49.10 1.07 0.06 0.84 0.29 32.25 mode using a secondary electron multiplier. Total procedural blanks for Re and Os are b12 and b0.5 pg, respectively, with an average 187Os/188Os value of ~0.4 (n=2). Raw Re and – – – – – – – Os oxide values were corrected for oxygen contribution and mass

Selby and Creaser, 2005 fractionation. The Re and Os isotopic values and elemental abun- 0.569 0.002 0.85 0.05 0.79 0.28 11.55 0.04 1.1 0.4 109.5 0.5 1.036 0.004 0.830 0.006 0.932 0.005 0.383 0.002 0.79 0.05 0.77 0.27 7.41 0.03 1.0 0.4 104.4 0.6 1.037 0.006 0.841 0.007 0.926 0.008 0.811 0.290 1.15 0.644 0.495 (ppb) (ppb) (pbb) dances are calculated by full propagation of uncertainties from Re and Os mass spectrometer measurements, blank abundance and iso- . topic composition, spike calibration, and sample and spike weights. Throughout the period of this study, in-house Re and Os (DROsS) – – Maturity Os ± Pt ± Pd ± Re standard solutions were repeatedly analysed to monitor instrument reproducibility. The NCIET Re standard is made from 99.999% zone- reﬁned Re ribbon and is considered to be indistinguishable from the (m) Creaser et al. (2002) AB1 Re standard of the Department of Earth Sciences, University of Alberta. The Re standard runs produced average 185Re/187Re values of 0.5966±0.0017 (1 S.D. n=3) indistinguishable within uncertainty to 0.5984±0.0019 (Finlay et al., 2011 and references therein). The 185 187 Os data from measured difference between the Re/ Re values and the – accepted 185Re/187Re value (Gramlich et al., 1973) is used to correct a for sample mass fractionation. The average DROsS 187Os/188Os ratio, b Os calculated at the time of oil generation (112 Ma;

188 using an electron multiplier, is 0.160797±0.000214 (1 S.D. n=3),

Os/ indistinguishable from reported DROsS values (0.160924±0.000003 Maturity and Re Maturity from Riediger and Bloch (1995). [2 SD, n=21; Nowell et al., 2008]; 0.16097±0.00034 [2 S.D., DS69-03 8-35-31-25W4 2340.9 Duvernay Fm. DS44-03 2-12-50-26W4 1751.6 PEX21 8-29-78-01W6 2099.2 M 0.698 PEX20 8-29-78-01W6 2099.1 M 1.38 PEX14 4-23-72-10W6 3567.7 O PEX13 4-23-72-10W6 3570.4 O PEX12 13-18-80-23W5 1750.9 I PEX11 13-18-80-23W5 1748.8 I DS26-03 6-29-85-11w5Exshaw Fm. 1148.68 M 3.05 0.01 5.96 0.35 6.95 2.43 402.6 1.4 0.86 0.30 933.2 3.3 3.716 0.005 1.986 0.008 3.916 0.015 DS09-03DS10-03 07-31-79-10w6DS11-03 14-24-80-7w6DS12-03 1557 14-24-80-7w6DS14-03 8-26-69-7w6 1183.5 10-17-84-22w5 1190.4 M M 2271.4 2361 MPEX10 O 6.34 I 1.86 3-19-80-23W5 1.83 0.02 0.01 1756.5 6.13 0.01 6.45 1.12 2.90 I 0.38 0.03 3.44 0.17 0.00 7.27 0.20 2.89 7.01 2.54 1.27 5.16 0.17 2.45 574.9 0.08 1.81 8.06 369.2 1.40 229.9 1.9 2.82 1.2 0.49 195.3 0.8 0.89 196.2 0.41 0.7 0.31 0.67 0.7 0.15 0.36 0.24 550.4 1674 0.91 0.13 885.8 0.32 1.9 178.6 6 2.116 3.2 5.900 3.655 1.0 1.372 0.003 1.096 0.010 0.007 0.004 2.796 2.013 0.009 2.234 0.011 0.008 1.041 0.009 6.259 3.844 0.009 0.025 0.016 1.411 0.012 Gordondale Fm. Sample Well Depth Undisclosed value. a b c 187 – Table 2 West Canadian Sedimentary Basin Re and PGE data. n=16; Rooney et al., 2010]). A.J. Finlay et al. / Earth and Planetary Science Letters 313–314 (2012) 95–104 99

Table 3 West Canadian Tar Sands Re and PGE data.

a 187 187 b b Sample Well/core Os ± Pt ± Pd ± Re ± Pt/Pd ± Os/ ± Os/ ±Osg ±Osg ± (ppb) (ppb) (pbb) (ppb) 188Os 188Os (112 Ma) (60 Ma)

Athabasca DS49-00 15-12-93-13W4 0.074 0.001 3.49 0.21 1.40 0.52 12.9 0.1 2.5 0.95 1231 28 3.680 0.093 1.42 0.05 2.45 0.08 6403^ 6-93-10W4 0.144 0.002 0.48 0.03 0.63 0.24 26.2 0.1 0.77 0.29 1331 23 4.138 0.085 1.69 0.05 2.80 0.08 6438^ 95-9W4 0.062 0.001 0.40 0.02 1.02 0.38 11.2 0.1 0.39 0.15 1291 33 3.800 0.11 1.43 0.05 2.51 0.10

Cold Lake DS39-00 1-14-65-2W4 0.169 0.002 5.78 0.34 5.16 1.93 25.8 0.1 1.1 0.4 1072 20 3.624 0.076 1.65 0.05 2.55 0.07 655^ 28-61-4W4 0.288 0.006 1.54 0.09 4.26 1.60 50.1 0.2 0.36 0.14 1233 40 3.735 0.17 1.47 0.08 2.50 0.14 667^ – 0.230 0.002 1.34 0.08 1.88 0.71 42.7 0.1 0.71 0.27 1369 17 4.210 0.006 1.69 0.02 2.84 0.04 6400^ – 0.235 0.003 0.77 0.05 0.61 0.23 40.7 0.1 1.3 0.5 1227 17 3.738 0.067 1.49 0.03 2.51 0.06 6401^ – 0.133 0.002 0.33 0.02 0.56 0.21 23.6 0.1 0.59 0.22 1257 21 3.762 0.069 1.46 0.04 2.50 0.06

Grosmont DS46-00 14-5-88-19W4 0.132 0.001 0.21 0.01 0.49 0.18 23.9 0.1 0.42 0.16 1304 12 3.961 0.040 1.57 0.02 2.65 0.04 DS47-00 14-5-88-19W4 0.168 0.001 5.40 0.32 8.06 3.02 29.9 0.1 0.67 0.25 1283 11 3.950 0.035 1.60 0.02 2.67 0.03 DS48-00 14-5-88-19W4 0.139 0.001 0.38 0.02 0.57 0.21 24.6 0.1 0.66 0.25 1192 11 3.696 0.034 1.51 0.02 2.45 0.03

Lloydminster 656^ 03-14-49-27W3 0.147 0.002 1.11 0.07 1.33 0.50 22.2 0.1 0.83 0.32 1028 17 3.303 0.007 1.42 0.02 2.27 0.04 664^ A12-19-50-24-W3 0.136 0.002 8.03 0.47 6.95 2.61 24.0 0.1 1.2 0.4 1260 25 3.976 0.011 1.66 0.03 2.70 0.05 666^ A13-35-49-24W3 0.150 0.002 4.44 0.26 0.74 0.28 24.2 0.1 6.0 2.28 1134 19 3.650 0.075 1.56 0.04 2.52 0.07

Peace River DS28-00 4-21-85-18W5 0.064 0.001 1.20 0.07 0.11 0.04 12.4 0.1 11 4 1404 44 4.018 0.13 1.44 0.07 2.61 0.12 DS29-00 4-21-85-18W5 0.058 0.001 5.87 0.35 7.27 2.73 11.3 0.1 0.81 0.31 1428 38 4.077 0.11 1.46 0.06 2.66 0.10 DS30-00 4-21-85-18W5 0.070 0.001 0.29 0.02 1.04 0.39 13.7 0.1 0.28 0.11 1429 34 4.102 0.10 1.48 0.05 2.67 0.09

Provost DS42-00 5-20-37-1W4 0.041 0.001 0.07 0.00 1.65 0.62 2.78 0.03 0.04 0.02 372 14 1.204 0.076 0.52 0.04 0.83 0.06 DS43-00 5-20-37-1W4 0.045 0.001 0.24 0.01 8.37 3.14 2.88 0.02 0.03 0.01 354 10 1.200 0.058 0.55 0.03 0.85 0.05 DS45-00 5-20-37-1W4 0.057 0.001 0.33 0.02 1.49 0.56 3.92 0.03 0.22 0.08 383 13 1.312 0.070 0.60 0.04 0.93 0.06

Wabasca DS31-00 2-24-84-22W4 0.133 0.002 6.62 0.39 7.01 2.63 23.1 0.1 0.95 0.36 1230 23 3.693 0.096 1.44 0.05 2.47 0.08 DS32-00 2-24-84-22W4 0.127 0.001 0.19 0.01 1.22 0.46 21.7 0.1 0.15 0.06 1200 13 3.641 0.004 1.44 0.02 2.44 0.03 DS34-00 2-24-84-22W4 0.227 0.002 0.22 0.01 1.18 0.44 40.2 0.1 0.19 0.07 1259 11 3.791 0.035 1.48 0.02 2.53 0.03 650^ 14-28-81-22W4 0.225 0.003 0.31 0.02 1.17 0.44 41.2 0.1 0.27 0.10 1359 18 4.252 0.067 1.76 0.04 2.90 0.06

– Undisclosed value. a Re–Os data from Selby and Creaser (2005). b 187Os/188Os calculated at the time of oil generation at 112 Ma, and 60 Ma (see text for discussion of ages).

Source unit and oil 187Os/188Os compositions at the time of oil Shale and oil were digested for PGE analysis using a modiﬁed generation (Osg) are calculated using the Re–Os age of generation method of Dale et al. (2009 and references therein). 1 g of powdered (UK Atlantic Margin, 68 Ma; Finlay et al., 2011; WCOS, 112 Ma; shale or ~0.1 g of asphaltene was loaded into quartz high-pressure Selby and Creaser 2003) and the 187Re decay constant of asher (HPA) vessel with a mixed 196Pt and 106Pd tracer (spike) solu- λ=1.666×10-11a-1 (Smoliar et al., 1996). tion and inverse aqua-regia (shale—2.5 ml 11 N HCl and 5 ml 16 N

HNO3; oil—1.5 ml 11 N HCl and 3 ml 16 N HNO3). The vessels were 2.3.2. PGE analysis placed into an Anton Paar HPA (housed at NCIET) and heated to Biodegraded oil samples from the WCOS (n=24) and asphaltene 220 °C for 13 h at ~110 bars. After digestion the inverse aqua-regia separated from UKAM oils (n=3) were analysed for their PGE abun- solution was dried, converted to the chloride species and re- dances using the following method. Asphaltene was separated from the dissolved in 4 ml of 0.2 N HCl for purification of Pd and Pt using UKAM oils using a n-heptane based methodology (Selby et al., 2007). anion exchange chromatography. Briefly this chromatography in- 40 ml of n-heptane was added to 1 g of oil in a 60 ml glass vial, thorough- volved loading 1 cm3 of AG-1x8 (100–200 mesh) anion resin onto ly mixed and agitated for ~12 h. The contents of the vial were then trans- the column and washing with 6 ml of 6 N HCl. The column was then ferred to a 50 ml centrifuge tube and centrifuged at 4000 rpm for 5 min rinsed with 10 ml of Milli-Q and preconditioned with 5 ml of 0.5 N to ensure complete separation of the soluble maltene and insoluble HCl. The sample HNO3 solution was centrifuged and the HNO3 Pt asphaltene fractions. The maltene fraction was decanted to waste and and Pd solution was added to the column. The column was then the asphaltene fraction was dried on a hot plate at ~60 °C. washed with 10 ml of 1 N HCl/HF to remove high field strength

Oil was removed from the WCOS using a method modiﬁed from elements and conditioned with 10 ml of 0.8 N HNO3. The Pt cut was (Selby et al., 2005). Approximately 1 g of WCOS was washed in ~1/ eluted in 12 ml of 13.5 N HNO3. The column was rinsed in 5 ml of 2 ml of CHCl3 which was then decanted into a 15 ml centrifuge Milli Q, washed with 10 ml of 1 N HCl/HF and rinsed with 5 ml of tube. This process was repeated until the CHCl3 remained clear. The Milli Q. Finally the puriﬁed Pd cut was eluted in 20 ml of 9 N HCl. CHCl3 +oil solution was then repeatedly centrifuged to separate any Pt and Pd were measured by inductively coupled plasma mass suspended sediment from the solution. The CHCl3 +oil solution was spectrometry (ICPMS) on a ThermoFinnigan Element 2 at Durham then transferred to a 22 ml glass vial and the CHCl3 evaporated at University. Solutions were introduced using a MicroMist micro- ~60 °C. The WCOS are heavily biodegraded and so are naturally concentric nebuliser and ESI stable sample introduction system enriched in asphaltene, rendering separation of asphaltene for analy- (dual-cyclonic quartz spray chamber). Mixed solutions of natural sis unnecessary. PGEs and solutions of Hf, Zr, Y and Mo (all 1 ppb) were used to 100 A.J. Finlay et al. / Earth and Planetary Science Letters 313 (2012) 95–104

Table 4 unit equivalent Kimmeridge Clay Fm. of the North Sea (0.94–1.13, USGS Devonian Ohio Shale Standard PGE data. n=5, Table 1; Finlay et al., 2010). The Pt and Pd abundances (ppb) Run Pt (ppb) ±a Pd ±a of the UKAM and North Sea Kimmeridge Clay Fm. source rocks are (ppb) similar (Pt=2.2 to 4.6; Pd=3.7 to 13.9; n=8; Table 1), with Pt/Pd 1 1.45 5.4% 2.22 14% values between 0.33 and 0.63 (Table 1). 2 1.43 4.4% 3.41 11% The Osg of the UKAM petroleum (0.92–1.25; calculated at 68 Ma) 3 1.45 2.4% 2.30 12% is similar to the source units of the UKAM and North Sea (Fig. 2a, b; 4 1.50 3.7% 2.66 13% Table 1). This agrees with previous studies that have proposed that 5 1.39 3.6% 2.87 16% Mean 1.44 3.9% 2.69 13% the Osg of petroleum is inherited from the source unit during petro- 2 Standard Deviations 0.09 0.96 leum generation, thus providing a ﬁngerprint tool (Finlay et al., 2 Standard Deviations percentage of Mean 5.9% 35.7% 2010; 2011; Selby and Creaser 2005; Selby et al., 2007). To comple- a Measured individual run uncertainties. ment the Osg we analysed several whole petroleum samples from the UKAM for their Pt and Pd abundance. The whole petroleum analyses yield low Pt and Pd abundances (~0.2 ppb Pt and blank levels quantify the degree of mass fractionation and the production rates of to 1 ppb Pd; Table 5) and together with full procedural blank correc- HfO+, ZrO+, YO+, and MoO+ (which possess oxides of equivalent tion and uncertainty propagation the data possessed very high uncer- mass to isotopes of Pt, and Pd) before, during and after the analysis. tainties rendering it meaningless (Table 5). This is similar to previous During the course of this study the USGS Devonian Ohio Shale whole petroleum Pt and Pd data where 6 replicates produced uncer- standard (SDO-1; Kane et al., 1990) was repeatedly analysed for tainties of 80 and 70%, respectively (Woodland et al., 2001). Conse- PGE abundances (n=5; Table 4), producing Pt values ranging from quently we show that, like Re and Os, Pt and Pd are predominantly 1.39 to 1.50 ppb and Pd values ranging from 2.22 to 3.41. The individ- complexed in the asphaltene fraction of petroleum (Table 1). Asphal- ual measured analysis of SDO-1 yield uncertainties of ~3.4% and ~13% tenes separated from UKAM petroleum, previously analysed for Re– (at the 2 S.D. level) for Pt and Pd, respectively (Table 4). Repeat anal- Os (Finlay et al., 2011), contain between 0.2 and 11.0 ppb Pt and 2.8 ysis of SDO-1 yield a 2 S.D. uncertainty of 5.9% and 35.7% for Pt and Pd to 16.8 ppb Pd (Pt uncertainty=3–6%, Pd uncertainty=19–60%; abundances, respectively (Table 4). The Pd value is very imprecise n=3; Table 1) and possess Pt/Pd values between 0.05 and 0.44. due to 1 analysis which gave a high value of 3.41, without this outlier the 2 S.D. of the data is reduced to 22.6%. Therefore, so as to not un- 3.2. West Canadian Oil Sands and Sedimentary Basin derestimate the uncertainty in our PGE data, the 2 S.D. reproducibility uncertainties for SDO-1 are applied as the uncertainty for the Pt and To identify the source of the WCOS we have analysed the potential Pd for shale and petroleum samples. A previous study that analysed Upper Devonian Duvernay, Devonian/Mississippian Exshaw, and the SDO-1 standard for PGEs present data that just lie within uncertain- Jurassic Gordondale Fms. source units as well as petroleum from across ty of our values (Pt=2.15±28% ppb, Pd=2.42±32% ppb; Meisel and the entire WCOS deposits (Athabasca, Cold Lake, Grosmont, Lloydminster, Moser, 2004). However, not only are the Pt values imprecise but there Peace River, Provost and Wabasca; Tables 2 and 3; Fig 1). was no replicate analysis and the presented Re and Os values are signif- The regional Gordondale Fm. samples contain Re and Os abun- icantly different to those produced at NCIET (93±140% ppb Re dances ranging from 195.3 to 574.9 and 1.1 to 6.3 ppb, respectively. and 1.159±38% ppb Os; Meisel and Moser, 2004; 76.6±0.9 ppb Re, The 187Re/188Os and 187Os/188Os values positively correlate and 0.638±0.019 ppb Os, n=6, Du Vivier, NCIET unpublished data 2010). range from 178.6 to 933.2 and 1.372 to 5.900, respectively, yielding Meisel and Moser (2004) only utilised a 3 h digestion, therefore, it is aRe–Os age of 184±29 Ma; MSWD 2153 (Fig. 3). The Gordondale possible that these differences in data are caused by either incomplete Fm. encompasses the Hettangian (~199 Ma) to Toarcian (~175 Ma; sample digestion or incomplete sample/spike equilibrium. Asgar-Deen et al., 2004) and, although imprecise and having a Eight full procedural blanks were undertaken throughout the peri- MSWD that is too large to consider this plot an isochron, the Re–Os od of this study. Blanks for Pt ranged from 0.3 to 22.7 pg (mean=5.5, age is in agreement with the biostratigraphic constraints meaning it 1S.D.=8.2) and Pd blanks ranged from 72 to 547 pg (mean=217, has value. The regional Duvernay Fm. samples contain Re and Os 1S.D.=172). The highest Pt and Pd blank values (22.7 and 547 pg, abundances ranging from 7.4 to 11.6 and 0.3 to 0.6 ppb, respectively. respectively) were all from the same batch, contaminated by a N- The 187Re/188Os and 187Os/188Os values positively correlate and range benzoyl-N-phenylhydroxylamine–Chloroform solution. This solution from 104.4 to 109.5 and 1.037 to 1.036, respectively, but lack the nec- was applied, after anion exchange chromatography, to remove high essary spread for Re–Os geochronology. The Re–Os data for the re- ﬁeld strength elements that have PGE isotope interferences gional Exshaw Fm. (~20–90 ppb Re and 0.5–1.4 ppb Os; Table 2) (89Y16O_105Pd; 92Zr16O_108Pd; 179Hf16O_195Pt; 180Hf16O_196Pt; has been previously determined and yield ages of 358±10 Ma Shinotsuka and Suzuki, 2007). Because of this contamination from (Creaser et al., 2002), and 366.1±9.6 and 363.4±5.6 Ma (Selby and the N-benzoyl-N-phenylhydroxylamine–Chloroform protocol, this Creaser, 2003). stage was not repeated for subsequent analysis. Isotope interferences The Pt and Pd abundances of samples from regional Exshaw Fm. were calculated and subtracted from the data by calculation of pro- shales (n=7; Table 2) range from 1.1 to 3.3 and 0.8 to 2.5 ppb, re- duction from the Hf, Zr, Y and Mo standards (see above). spectively. The Regional Gordondale Fm. samples (n=6; Table 2) contain 1.3 to 6.5 ppb Pt and 1.4–8.1 ppb Pd. Regional Duvernay Fm. 3. Results samples (n=2; Table 2) contain 0.8 and 0.9 ppb Pt and 0.8 ppb Pd. Asphaltene-rich petroleum from the WCOS (n=24; Table 3) contain 3.1. UKAM between 0.1 and 8.0 ppb Pt and 0.1 and 8.4 ppb Pd. The Pt/Pd values for all samples range from 0.03 to 10.8, with 21 of the 24 of the sam- The UKAM Kimmeridge Clay Fm. source rock possesses Re–Os ples possessing very similar Pt/Pd values (0.15–1.26; Table 3). abundances ranging between 110.5 to 242.1 ppb and 0.8 to 1.4 ppb, Three petroleum samples, separate from the main WCOS data set, respectively. The 187Re/188Os and 187Os/188Os values positively corre- are from different deposits and contain Pt/Pd values outside two stan- late and range from 958.6 to 1216.5 and 3.048 to 3.463, respectively. dard deviations of the mean main data set (Athabasca, 2.5; The 187Os/188Os values calculated at the time of petroleum generation Lloydminster, 6.0; Peace River, 11; Table 3). Their Pt/Pd values are

(Osg; 68 Ma; Finlay et al., 2011) range from 1.41 to 1.58. These Osg greater than any observed Pt/Pd value presented in this study. Possi- values are similar, although slightly more radiogenic, to the source ble explanations for these data are that they have been sourced from A.J. Finlay et al. / Earth and Planetary Science Letters 313–314 (2012) 95–104 101

a Kimmeridge Clay Fm. b UKAM Oil 2 2 g

g Kimmeridge Clay Fm. Os Os 1 1 UKAM oil

0 0 0 12 012 Pt/Pd Pt/Pd c Exshaw Fm. d Gordondale Fm. 4 4 Duvernay Fm. Main WCTS Provost 3 3 Gordondale Fm. g (112 Ma) 2 Os 2 Exshaw Fm. Osg Main WCTS

1 1 Duvernay Fm.

Provost oil 0 0 0 1 2 012 Pt/Pd Pt/Pd 7 Exshaw Fm. 7 e Gordondale Fm. f Gordondale Fm. Duvernay Fm. 6 Main WCTS 6 Provost 5 5

4 4 (60 Ma) (60 Ma)

g 3 g 3 Os Os Main WCTS 2 2 Exshaw Fm. 1 1 Provost oil Duvernay Fm. 0 0 012 012 Pt/Pd Pt/Pd

187 188 Fig. 2. Osg vs Pt/Pd (where Osg is Os/ Os at the time of petroleum generation) plots for; a) UKAM oil and source unit; b) simplified result plots for the UKAM source rock and oil data clouds; c) WCSB and WCOS source units and oil (Osg calculated at 112 Ma); d) Simplified results showing WCSB and WCOS data clouds (Osg calculated at 112 Ma); e) WCSB and WCOS source units and oil (Osg calculated at 112 Ma for Exshaw Fm., and 60 Ma for oil Duvernay and Gordondale Fms., see text for discussion) ; f) Simplified results showing

WCSB and WCOS data clouds (Osg calculated at 112 Ma for Exshaw Fm., and 60 Ma for oil Duvernay and Gordondale Fms., see text for discussion). Pt/Pd uncertainty bars are 2 S.D.

Osg uncertainties are smaller than data points. a unit not analysed as part of this study which is extremely enriched ﬁeld (UK North Sea) was reported and suggested to be a result of Pt in Pt compared to Pd or that the Pt/Pd values have been disturbed. For contamination “during the drilling process” (Woodland et al., 2001). example, extreme Pt enrichment in petroleum from the Dauntless oil Exactly what may have caused the contamination during drilling is unstated, however, drilling mud has been shown to be enriched in Re and therefore is likely enriched in Pt as well (Finlay unpublished Table 5 data) and may therefore be the source of contamination. Contamina- Whole oil and asphaltene fraction PGE comparison. tion, therefore, may have disturbed these three samples. Samplea Ir ±Pt±Pd ± Pt/Pd ± (ppb) (ppb) (ppb) 4. Discussion 0043 Whole oil 0.023 0.040 0.22 0.15 1.57 3.39 0.14 0.31 Asphaltene 0.024 0.004 0.24 0.01 8.37 3.14 0.03 0.01 4.1. Pt/Pd and Osg petroleum ﬁngerprinting Difference 0.001 0.02 6.80 2712 Whole oil 0.028 0.089 0.22 0.06 ≤BLa 1.67 N/A N/A Asphaltene 0.094 0.017 3.47 0.20 10.9 3.90 0.32 0.11 The Pt and Pd abundances for the UKAM and North Sea Kimmeridge Difference 0.065 3.25 N/A Clay Fm. source samples are within the range of published values for Uncertainties are 2 SD of raw data after blank correction. global shales (Tables 1; 5; 2; 3; Coveney et al., 1992; Colodner et al., a Less than or equal to average blank value. 1992; Sawlowicz, 1993; Over et al., 1997; Woodland et al 2000; 102 A.J. Finlay et al. / Earth and Planetary Science Letters 313 (2012) 95–104

Overmature Gordondale Fm. γ data-point error ellipses are 2 Mature Gordondale Fm. Immature Gordondale Fm.

6 Os 188 Overmature Exshaw Fm. Os/ Mature Exshaw Fm.

187 Immature Exshaw Fm.

4 0 0.5 1 1.5 2 2.5 Pt/Pd

Fig. 5. Pt/Pd values for Gordondale (grey) and Exshaw Fm. (black) shales of different Age = 184± 29 Ma maturity. The data show no fractionation due to maturation. Uncertainty bars are 2 S.D. (see above). 2 Osi = 0.73± 0.51 Ma MSWD = 2153 To be effective as a petroleum source fingerprinting tool, Pt/Pd 187Re/188Os values in source units must be similar regionally, be undisturbed by 200 600 1000 1400 1800 maturation and be similar between petroleum and the known source. The UKAM and Miller Oil Field Kimmeridge Clay Fm. sample locations Fig. 3. Re–Os plot from regional Gordondale Fm. Samples (Table 2). Uncertainty are ~200 miles apart and show no regional variation. Furthermore, σ ellipses are 2 . Although the MSWD (see text) means this plot cannot be considered immature, mature and overmature samples of the Gordondale and an isochron the age produced agrees with biostratigraphy giving the age significance. Exshaw Fms. possess Pt/Pd values that are indistinguishable within uncertainty (Table 1; Fig. 5) suggesting that, like Re–Os, the process of hydrocarbon maturation does not significantly alter source rock Peucker-Ehrenbrink and Hannigan, 2000; Jaffe et al., 2002; Meisel and Pt/Pd systematics. Asphaltenes separated from UKAM petroleum con- Moser 2004; Siebert et al., 2005; Schmitz et al., 2006; Jiang et al., tain Pt/Pd values that range from 0.32±0.21 to 0.65±0.30, indistin- 2007; Brookfield et al., 2010). The Pt in organic-rich sediments is pre- guishable from the range observed in Kimmeridge Clay Fm. (Table 1; dominantly hydrogenous being complexed to organic matter under an- Fig. 2a, b), which has been demonstrated to be the dominant source oxic conditions (Colodner et al., 1992). This is the same as for Re and Os unit for UKAM oils and the known source of the samples tested (e.g. Cohen et al., 1999; Ravizza and Turekian, 1989; Selby and Creaser, (e.g., Cornford, 1998; Finlay et al., 2011; Rooney et al., 1998; 2003). It is highly likely that Pd (together with Pt, Os and Re) in organic- Scotchman et al., 2006). Thus, the lack of variation in Pt/Pd values re- rich sediments is also hydrogenous, being complexed by organic matter. gionally within the same source unit, the lack of an effect on Pt/Pd Upper Continental Crustal (UCC) normalised asphaltene Pt and Pd data values by maturation and the similarity of the Pt/Pd values of a source display similar Pt and Pd profiles as those of the Kimmeridge Clay Fm. unit and its' generated petroleum demonstrates that Pt/Pd value can source shales (Fig. 4), suggesting that there is very little fractionation be used to identify a petroleum source. between Pt and Pd during petroleum generation, therefore enabling To strengthen this correlation we combine the use of Pt/Pd value fi Pt/Pd values to be utilised for petroleum to source ngerprinting. with the Osg signature of an oil and source rocks, as Osg has

a b 10000 10000 UKAM Asphaltene UKAM source rock 1000 North Sea KCF 1000

100 100

10 10

1 1 UCC normilised value UCC normilised value

0.1 0.1 Os Pt Pd Re Os Pt Pd Re

10000 c Gordondale Fm. d 10000 Exshaw Fm. West Canadian Tar Sand Duvernay Fm. 1000 1000

100 100

10 10

1 1 UCC normilised value UCC normilised value

0.1 0.1 Os Pt Pd Re Os Pt Pd Re

Fig. 4. Upper Continental Crustal normalised PGE and Re values for a) Kimmeridge Clay Fm. shales; b) UKAM asphaltene; c) Gordondale, Exshaw and Duvernay Fm. shales of the WCSB; d) West Canadian Oil Sands. Upper Continental Crustal values from (Peucker-Ehrenbrink and Jahn, 2001). A.J. Finlay et al. / Earth and Planetary Science Letters 313–314 (2012) 95–104 103

previously been suggested to be a possible fingerprinting tool (Finlay WCSB, it is therefore less radiogenic as shown by the Osg. Further- et al., 2010; 2011; Selby et al., 2007). Within the UKAM, Pt/Pd and Osg more, these Provost samples do not plot on the 111.6 Ma isochron asphaltene data fields overlap with the Kimmeridge Clay Fm. source of Selby and Creaser (2005) suggesting they formed at a different unit data (Fig. 2a, b). This confirms that the Pt/Pd of oil is inherited time. Additional Pt and Pd analysis may prove in identifying the Os- from the source unit and can be used to fingerprint the source of tracode bed as the source of the Provost oil field. petroleum of global petroleum systems. The age used for calculation of the Osg of the oil in this study is based on the Re–Os age of Selby and Creaser (2005; which assumes a single

4.2. Pt/Pd and Osg fingerprinting of the WCOS: Implications for global source) that we consider to represent the generation age of the petroleum systems WCOS. If, however, these oils are from a mixed source this age may re- flect a mixed signal. Basin models predict that the Gordondale Fm. The WCSB source units were deposited in two distinct tectonic set- reached maturity in the Late Cretaceous to Paleogene with peak gener- tings and range in age from the Early Palaeozoic to the Early Tertiary. ation occurring at ~60 Ma (Higley et al., 2009)despitefluid flow model- Palaeozoic to Jurassic sedimentary units, dominated by carbonates, ling which shows that a generation age of 60 Ma is too young to enable were unconformably deposited on the Laurentian craton under passive the filling of the Athabasca deposit (Adams et al., 2004). We have, how- margin conditions. Middle Jurassic to Paleocene, dominantly clastic, ever, calculated the Osg of the source units and WCOS at this modelled units formed in a foreland basin succession in response to the Cretaceous 60 Ma generation age (Tables 2 and 3) and show that it makes no differ- Laramide Orogeny (Creaney and Allan 1990; Zhou et al., 2008). The Re– ence in our conclusion that the Gordondale Fm. is the main source unit Os geochronology yields petroleum generation in the WCSB at 111.6± with only a minor Exshaw Fm. input (Fig. 2e, f). 5.3 Ma (Selby and Creaser. 2005). This age for petroleum generation is Our identification of the Gordondale Fm. as the main source of the supported by diagenetic illite K–Ar geochronology from K-bentonites WCOS further aids the understanding of the stratigraphy of the WCSB. within the Exshaw Fm. (118 to 112 Ma in the west, 68–60 Ma in the It has been argued that the Lower Jurassic Poker Chip Fm. would have East; Meyer et al., 2008). The mid-Cretaceous age is synchronous with provided a seal to the migration of oil from the Gordondale Fm.; except reservoir deposition, supporting the hypothesis that the present bitumi- for a small region in the Eastern edge of the Gordondale Fm., where ero- nous state of the oil sands is as a result of biodegradation during migra- sion has removed the overlying Poker Chip Fm. (Riediger, 1994). New tion to, and emplacement within, very shallow-subaerial unconsolidated 4D modelling utilising recent seismic data and modelling software as reservoir sands (Selby and Creaser 2005; Zhou et al., 2008). However, well as kerogen analysis of the WCSB disputes this and concludes that despite the geochronology of the WCOS discounting any Cenozoic and the Fernie Group (of which the Gordondale Fm. is part) is the major mid-late Cretaceous source units, the source of the WCOS has remained source of the petroleum of the WCOS (Higley et al., 2009). This and controversial (Allan and Creaney 1991; Barson et al., 2000; Creaney and mass balance calculations (Allan and Creaney 1991, Creaney and Allan Allan, 1990; 1992; Fowler et al., 2001; Higley et al., 2009; Riediger et al., 1990; 1992) are confirmed by our findings, demonstrating how inor- 2000; Selby and Creaser, 2005; Zhou et al., 2008). ganic geochemistry can be used to “ground truth” petroleum models. The Re–Os geochronology of the WCOS indicated that petroleum was generated at 111.6±5.3 Ma, with Os values between 1.30 and g 5. Conclusions 1.76 (Selby and Creaser. 2005). Using this Re–Os petroleum generation age to calculate Os for the Gordondale, Exshaw and Duvernay g This study demonstrates the feasibility of using 187Os/188Os and Fms. produces a range between 1.04 and 2.80, 1.24–4.16 and Pt/Pd values to act as an oil to source fingerprinting tool through 0.83–0.84, respectively. These Os values suggest that the WCOS g the analysis of oils and samples of the known source United Kingdom could have been sourced from the Gordondale and/or Exshaw Fms., Atlantic Margin oil fields. When applied to the West Canadian Oil but not the Duvernay Fm. (Fig. 2c, d). Sands this method identifies the Gordondale Fm. as the main oil The main WCOS data set (n=19) contains Pt/Pd values that are al- source. Unlike traditional organic geochemistry, Pt/Pd and Os finger- most indistinguishable from those observed in the Gordondale Fm., ex- g printing is not rendered ineffective by biodegradation, but is aided by cept for the Provost samples that have lower Pt/Pd values (Fig 2c, d). the natural concentration of Pt and Pd in asphaltene. This allows oil Two asphaltene samples (11 including uncertainty) overlap with source correlation in previously unsuitable petroleum systems and those observed in the Duvernay and Exshaw Fms. This suggests that the deduction of migration pathways. Combining both Re–Os petro- the majority of the WCOS have been sourced from the Gordondale leum geochronology (Finlay et al., 2011; Selby and Creaser, 2005; Fm. with a small petroleum component sourced from the Exshaw and Selby et al., 2005) and Os , and Pt and Pd geochemistry (this study) Duvernay Fms., but, as demonstrated above the Os of the WCOS rule g g now permits both spatial and temporal constraints on petroleum sys- out a Duvernay Fm. source. Therefore, by comparing the Pt/Pd and Os g tems to be established and therefore provide a significant tool for pe- values of the WCOS to the potential source unit samples it is apparent troleum system exploration, development and research. that, with the exception of the Provost deposit, the majority of WCOS data are similar to the Gordondale Fm. (Fig. 2c, d). Accordingly, this study hypothesises that the Gordondale Fm. is the predominant source Acknowledgements unit for the WCOS with only minor inputs from the Exshaw Fm.

Petroleum from the Provost deposit have significantly lower Osg This research was funded by a NERC/B.P. CASE award to Finlay. values than the other WCOS (Provost=0.52–0.60; WCOS= Dr Chris Dale and Dr Chris Ottley are thanked for their valuable advice 1.42–1.76). The Provost petroleum also displays Pt/Pd values lower on developing the Pt and Pd analytical methodology. We thank than the majority of the WCOS and the source samples analysed in Dr. Ken Peters and two anonymous reviewers for their helpful com- this study (Fig. 2c, d). It may be that Pt/Pd fractionation has occurred ments. We would also like to thank BP for allowing publication of between the provost source rock and oils, however as the Os isotopes the data presented in this study. support the difference in Pt/Pd and no fractionation was observed in the UKAM samples we believe this is unlikely. Therefore, it is likely References that the petroleum have been sourced from a unit not analysed in this study. A potential separate source for the Provost deposit is the Adams, J.J., Rostron, B.J., Mendoza, C.A., 2004. Coupled fluid flow, heat and mass trans- Lower Cretaceous Mannville Group Ostracode bed (e.g., biomarkers, port, and erosion in the Alberta basin: implications for the origin of the Athabasca oil sands. Can. J. Earth Sci. 41, 1077–1095. Karavas et al., 1998; 4D modelling, Higley et al., 2009). As the Ostra- Alberta's Energy Reserves 2005 and Supply/Demand Outlook 2006–2015; Alberta Energy code bed is younger than the other proposed source units of the and Utilities Board. http://www.ercb.ca (accessed 14/10/2010). 104 A.J. Finlay et al. / Earth and Planetary Science Letters 313 (2012) 95–104

Allan, J., Creaney, S., 1991. Canadian oil families of the Western Canada Basin. Pet. Geol. Nowell, G.M., Luguet, A., Pearson, D.G., Horstwood, M.S.A., 2008. Precise and accurate 39, 107–122. 186Os/188Os and 187Os/188Os measurements by multi-collector plasma ionisation Asgar-Deen, M., Riediger, C., Hall, R. Bull, 2004. The Gordondale Member: designation of a mass spectrometry (MC–ICP-MS) part I: solution analyses. Chem. Geol. 248, 363–393. new member in the Fernie Formation to replace the informal “Nordegg Member” no- Over, D.J., Conaway, C.A., Katz, D.J., Goodfellow, W.D., Gregoire, D.C., 1997. Platinum menclature of the subsurface of west-central Alberta. Bull. Can. Pet. Geol. 52, 201–214. group element enrichments and possible chondritic Ru:Ir across the Frasnian– Barson, D., Bartlett, R., Hein, F., Fowler, M., Grasby, S., Riediger, C. Underschultz, 2000. Famennian boundary, western New York State. Palaeogeogr. Palaeoclimatol. Hydrogeology of heavy oil and tar sand deposits: Water flow and supply, migration Palaeoecol. 132, 399–410. and degradation, Field Trip Notes (Geological Survey of Canada, 2000) Open File Parnell, J., Green, P.F., Watt, G., Middleton, D., 2005. Thermal history and oil charge on Report 3946 the UK Atlantic margin. Pet. Geosci. 11, 99–112. Brookfield, M.E., Shellnutt, J.G., Qi, L., Hannigan, R., Bhat, G.M., Wignall, P., 2010. Platinum Peters,K.E.,Walters,C.C.,Moldowan,J.M.,2005.TheBiomarkerGuideVolume2:Biomarkers element group variations at the Permo-Triassic boundary in Kashmir and British and Isotopes in Petroleum Exploration and History. Cambridge University Press. Columbia and their significance. Chem. Geol. 272, 12–19. Peucker-Ehrenbrink, B., Hannigan, R.E., 2000. Effects of black shale weathering on the Carruth, A.G., 2003. The Foinaven Field, blocks 204/19 and 204/24a, UK North Sea mobility of rhenium and platinum group elements. Geology 28, 475–478. (in United Kingdom oil and gas fields commemorative millennium volume). Peucker-Ehrenbrink, B., Jahn, B., 2001. Rhenium–osmium isotope systematics and Mem. Geol. Soc. Lond. 20, 121–130. platinum group element concentrations: loess and the upper continental crust. Cioppa, M.T., Symons, D.T.A., Al-Aasm, I.S., Gillen, K.P., 2002. Evaluating the timing of Geochem. Geophys. Geosyst. 2, 1061. hydrocarbon generation in the Devonian Duvernay Formation: paleomagnetic, Piggott, N., Lines, M.D., 1992. A case study of migration from the West Canada Basin. In: rock magnetic and geochemical evidence. Mar. Petrol. Geol. 19, 275–287. England, W.A., Fleet, A.J. (Eds.), Petroleum Migration: Geology Society, London, Cohen, A.S., Coe, A.L., Bartlett, J.M., Hawkesworth, C.J., 1999. Precise Re–Os ages of Special Publication, vol. 59, pp. 207–225. organic-rich mudrocks and the Os isotope composition of Jurassic seawater. Ravizza, G., Turekian, K.K., 1989. Application of the 187Re187Os system to black shale Earth Planet. Sci. Lett. 167, 159–173. geochronometry. Geochem. Cosmochem. Acta 53, 3257–3262. Colodner, D.C., Boyle, E.A., Edmond, J.M., Thomson, J., 1992. Post-depositional mobility Riediger, C., Ness, S., Fowler, M., Akpulat, T., 2000. Timing of oil migration, Paleozoic of platinum, iridium, and rhenium in marine sediments. Nature 358, 402–404. and Cretaceous bitumen and heavy oil deposits, eastern Alberta, GeoCanada Cornford, C., 1998. Petroleum Geology of the North Sea. Blackwell Science, Oxford. 2000. Proceedings of GeoCanada 2000, Calgary, Alberta, p. Paper 419. Coveney, R.M., Murowchick, J.B., Grauch, R.I., Glascock, M.D., Denison, J.R., 1992. Gold Riediger, C.L., 1994. Migration of “Nordegg” oil in the Western Canada Basin. How and platinum in shales with evidence against extraterrestrial sources of metals. much and how far. Bull. Can. Pet. Geol. 1, 63–73. Chem. Geol. 99, 101–114. Riediger, C.L., Fowler, M.G., Snowdon, L.R., Goodarzi, F., Brooks, P.W., 1990. Source rock anal- Creaney, S., Allan, J., 1990. Hydrocarbon generation and migration in the Western ysis of the Lower Jurassic “Nordegg Member” and oil-source rock correlations, north- Canada sedimentary basin. Geol. Soc. Lond. Spec. Publ. 50, 189–202. western Alberta and northeastern British Columbia. Bull. Can. Pet. Geol. 38, 236–249. Creaney, S., Allan, J., 1992. Petroleum systems in the foreland basin of Western Canada. Rooney, A.D., Selby, D., Houzay, J.P., Renne, P.R., 2010. Re–Os geochronology of a Meso- In: R.W. Macqueen and D.A. Leckie, Editors, AAPG Foreland Basins and Fold Belts. proterozoic sedimentary succession, Taoudeni basin, Mauritania: implications for AAPG Mem. 55, 279–308. basin-wide correlations and Re–Os organic-rich sediments systematics. Earth Planet. Creaser, R.A., Papanastassiou, D.A., Wasserburg, G.J., 1991. Negative thermal ion mass spec- Sci. Lett. 289, 486–496. trometry of osmium, rhenium and iridium. Geochim. Cosmochim. Acta 55, 397–401. Rooney, M.A., Vuletich, A.K., Griffith, C.E., 1998. Compound-specific isotope analysis as Creaser, R.A., Sannigrahi, P., Chacko, T., Selby, D., 2002. Further evaluation of the Re–Os a tool for characterizing mixed oils: an example from the West of Shetlands area. geochronometer in organic-rich sedimentary rocks: a test of hydrocarbon matura- Org. Geochem. 29, 241–254. tion effects in the Exshaw Formation, Western Canada Sedimentary Basin. Sawlowicz, Z., 1993. Iridium and other platinum-group elements as geochemical Geochim. Cosmoch. Acta 66, 3441–3452. markers in sedimentary environments. Palaeogeogr. Palaeoclimatol. Palaeoecol. Dale, C.W., Burton, K.W., Pearson, D.G., Gannoun, A., Alard, O., Argles, T.W., Parkinson, 104, 253–270. I.J., 2009. Highly siderophile element behaviour accompanying subduction of Schmitz, B., Ellwood, B.B., Peucker-Ehrenbrink, B., El Hassani, A., Bultynck, P., 2006. oceanic crust: whole rock and mineral-scale insights from a high-pressure terrain. Platinum group elements and 187Os/188Os in a purported impact ejecta layer Geochim. Cosmochim. Acta 73, 1394–1416. near Eifelian–Givetian stage boundary, Middle Devonian. Earth Planet. Sci. Lett. Finlay, A.J., Selby, D., Osborne, M.J., Finucane, D., 2010. Fault-charged mantle-fluid 249, 162–172. contamination of United Kingdom North Sea oils: insights from Re–Os isotopes. Scotchman, I.C., Carr, A.D., 2005. Modelling the effects of transient overpressure on the Geology 38, 979–982. petroleum systems of the UK North East Atlantic Margin and northern North Sea Finlay, A.J., Selby, D., Osborne, M.J., 2011. Re–Os geochronology and fingerprinting of areas: implications for the deep water South Atlantic and Gulf of Mexico. Pet. United Kingdom Atlantic Margin oil: temporal implications for regional petroleum Geol. Conf. Ser. 6, 1247–1265. systems. Geology 39, 475–478. Scotchman, I.C., Griffith, C.E., Holmes, A.J., Jones, D.M., 1998. The Jurassic petroleum Fowler, M.G., Stasiuk, L.D., Hearn, M., Obermajer, M., 2001. Devonian hydrocarbon system north and west of Britain; a geochemical oil-source correlation study. source rocks and their derived oils in the Western Canada Sedimentary Basin. Org. Geochem. 29, 671–700. Bull. Can. Pet. Geol. 49, 117–148. Scotchman, I.C., Carr, A.D., Parnell, J., 2006. Hydrocarbon generation modelling in a Gramlich, J.W., Murphy, T.J., Garner, E.L., Shields, W.R., 1973. Absolute isotopic abundance multiple rifted and volcanic basin: a case study in the Foinaven Sub-basin, ratio and atomic weight of a reference sample of rhenium. J. Res. Nat. Bur. Stand. A 77, Faeroe-Shetland Basin, UK Atlantic Margin. Scot. J. Geol. 42, 1–19. 691–698. Selby, D., Creaser, R.A., 2003. Re–Os geochronology of organic rich sediments: an eval- Higley, D.K., Lewan, M.D., Roberts, L.N.R., Henry, M., 2009. Timing and petroleum uation of organic matter analysis methods. Chem. Geol. 200, 225–240. sources for the Lower Cretaceous Mannville Group oil sands of northern Alberta Selby, D., Creaser, R.A., 2005. Direct radiometric dating of hydrocarbon deposits using based on 4-D modeling. AAPG Bull. 93, 203–230. Rhenium–Osmium isotopes. Science 308, 1293–1295. Jaffe, L.A., Peucker-Ehrenbrink, B., Petsch, S.T., 2002. Effects of weathering of black Selby, D., Creaser, R.A., Dewing, K., Fowler, M., 2005. Evaluation of bitumen as a 187Re– shales on the mobility of Re, PGE and Corg. Earth Planet. Sci. Lett. 198, 339–353. 187Os geochronometer for hydrocarbon maturation and migration: a test case Jiang, S.Y., Yang, J.H., Ling, H.F., Chen, Y.Q., Feng, H.Z., Zhao, K.D., Ni, P., 2007. Extreme from the Polaris MVT deposit, Canada. Earth Planet. Sci. Lett. 235, 1–15. enrichment of polymetallic Ni–Mo–PGE–Au in Lower Cambrian black shales of Selby, D., Creaser, R.A., Fowler, M.G., 2007. Re–Os elemental and isotopic systematics in South China: an Os isotope and PGE geochemical investigation. Palaeogeogr. crude oils. Geochem. Cosmochem. Acta 71, 378–386. Palaeoclimatol. Palaeoecol. 254, 217–228. Shinotsuka, K., Suzuki, K., 2007. Simultaneous determination of platinum group ele- Kane, J.S., Arbogast, B.F., Leventhal, J.S., 1990. Characterization of Devonian Ohio Shale ments and rhenium in rock samples using isotope dilution inductively coupled SDO-1 as a USGS geochemical reference sample. Geostand. News 14, 169–196. plasma mass spectrometry after cation exchange separation followed by solvent Karavas, F.A., Riediger, C.L., Fowler, M.G., Snowdon, L.R., 1998. Oil families in Mannville extraction. Anal. Chim. Acta 603, 129–139. Group reservoirs of southwestern Alberta, Western Canada sedimentary basin. Siebert, C., Kramers, J.D., Meisel, T., Morel, Ph., Nagler, T.F., 2005. PGE, Re–Os, and Mo Org. Geochem. 29, 769–784. isotope systematics in Archean and early Proterozoic sedimentary systems as Lamers, E., Carmichael, S.M.M., 1999. The Palaeocene deepwater sandstone play West proxies for redox conditions of the early Earth. Geochem. Cosmochem. Acta 69, of Shetland. In: Fleet, A.J., Boldy, S.A.R. (Eds.), Petroleum Geology of Northwest 1787–1801. Europe: Proceedings of the 5th Conference, pp. 645–659 (Geol. Soc. Lon.). Smoliar, M.I., Walker, R.J., Morgan, J.W., 1996. Re–Os ages of group IIA, IIIA, IVA, and IVB Leach, H.M., Herbert, N., Los, A., Smith, R.L., 1999. The Schiehallion development. In: Fleet, iron meteorites. Science 271, 1099–1102. A.J., Boldy, S.A.R. (Eds.), Petroleum Geology of Northwest Europe: Proceedings of the Spencer, A.M., Birkeland, Ø., Knag, G.Ø., Fredsted, R., 1999. Petroleum systems of the 5th Conference, pp. 683–692 (Geol. Soc. Lon.). Atlantic margin of northwest Europe. In: Fleet, A.J., Boldy, S.A.R. (Eds.), Petroleum Mark, D.F., Parnell, J., Kelley, S.P., Lee, M., Sherlock, S.C., Carr, A., 2005. Dating of multi- Geology of Northwest Europe: Proceedings of the 5th Conference. Geological Society, stage fluid flow in sandstones. Science 309, 2048–2051. London, pp. 231–246. Mark, D.F., Green, P.F., Parnell, J., Kelley, S.P., Lee, M.R., Sherlock, S.C., 2008. Late Palaeozoic Volkening, J., Koppe, M., Heumann, K.G., 1991. Osmium isotope ratio determinations by hydrocarbon migration through the Clair field, West of Shetland, UK Atlantic margin. negative thermal ionization mass spectrometry. Int. J. Mass Spectrom. Ion Proc. Geochim. Cosmochim. Acta 72, 2510–2533. 105, 147–159. Meisel, T., Moser, J., 2004. Platinum-group element and rhenium concentrations in low Woodland, S.J., Ottley, C.J., Pearson, D.G., Swarbrick, R.E., 2001. Microwave digestion of oils abundance reference materials. Geostand. Geoanal. Res. 28, 233–250. for platinum group and rare earth elements by ICP-MS. In: Holland, J.G., Tanner, S.D. Meyer, E. E., Aronson, J.L., Nadeau, P.H., Riediger, C., Hillier, S. 2008. Dating the tectonic (Eds.), Plasma Source Mass Spectrometry: The New Millennium. Spec. Pub. R. Soc, burial and maturation of the principal source rocks of the Alberta Tar Sands in the Chem, pp. 17–24. Western Canada Sedimentary Basin, Abstracts with Programs—Geological Society Zhou, S., Huang, H., Liu, Y., 2008. Biodegradation and origin of oil sands in the Western of America, Paper No. 211–10 Canada Sedimentary Basin. Pet. Sci. 5, 87–94.