Sie sind auf Seite 1von 27

Available online at www.sciencedirect.

com

Geochimica et Cosmochimica Acta 72 (2008) 844870 www.elsevier.com/locate/gca

Preservation of hydrocarbons and biomarkers in oil trapped inside uid inclusions for >2 billion years
Simon C. George a,*, Herbert Volk b, Adriana Dutkiewicz c, John Ridley d, Roger Buick e
Department of Earth and Planetary Sciences, Macquarie University, Sydney, NSW 2109, Australia b CSIRO Petroleum, P.O. Box 136, North Ryde, NSW 1670, Australia c School of Geosciences, University of Sydney, Sydney, NSW 2006, Australia d Department of Geosciences, Colorado State University, Fort Collins, CO 80523-1482, USA Department of Earth and Space Sciences & Astrobiology Program, University of Washington, Seattle, WA 98195-1310, USA Received 15 March 2007; accepted in revised form 14 November 2007
a

Abstract Oil-bearing uid inclusions occur in a ca. 2.45 Ga uvial metaconglomerate of the Matinenda Formation at Elliot Lake, Canada. The oil, most likely derived from the conformably overlying deltaic McKim Formation, was trapped in quartz and feldspar during diagenesis and early metamorphism of the host rock, probably before ca. 2.2 Ga. Molecular geochemical analyses of the oil reveal a wide range of compounds, including CH4, CO2, n-alkanes, isoprenoids, monomethylalkanes, aromatic hydrocarbons, low molecular weight cyclic hydrocarbons, and trace amounts of complex multi-ring biomarkers. Maturity ratios show that the oil was generated in the oil window, with no evidence of extensive thermal cracking. This is remarkable, given that the oils were exposed to upper prehnitepumpellyite facies metamorphism (280350 C) either during migration or after entrapment. The uid inclusions are closed systems, with high uid pressures, and contain no clays or other minerals or metals that might catalyse oil-to-gas cracking. These three attributes may all contribute to the thermal stability of the included oil and enable survival of biomarkers and molecular ratios over billions of years. The biomarker geochemistry of the oil in the Matinenda Formation uid inclusions enables inferences about the organisms that contributed to the organic matter deposited in the Palaeoproterozoic source rocks from which the analysed oil was generated and expelled. The presence of biomarkers produced by cyanobacteria and eukaryotes that are derived from and trapped in rocks deposited before ca. 2.2 Ga is consistent with an earlier evolution of oxygenic photosynthesis and suggests that some aquatic settings had become suciently oxygenated for sterol biosynthesis by this time. The extraction of biomarker molecules from Palaeoproterozoic oil-bearing uid inclusions thus establishes a new method, using low detection limits and system blank levels, to trace evolution through Earths early history that avoids the potential contamination problems aecting shale-hosted hydrocarbons. 2007 Elsevier Ltd. All rights reserved.

1. INTRODUCTION The nature of biological evolution and its intimate relationship with the rise in atmospheric oxygen during Earths
Corresponding author. Fax: +61 (0)2 9850 8248. E-mail addresses: simon.george@mq.edu.au (S.C. George), Herbert.Volk@csiro.au (H. Volk), adriana@geosci.usyd.edu.au (A. Dutkiewicz), jridley@cnr.colostate.edu (J. Ridley), buick@ess. washington.edu (R. Buick). 0016-7037/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2007.11.021
*

early history continues to be the subject of intense interest and debate (e.g., Lepland et al., 2005; Brasier et al., 2006; Goldblatt et al., 2006; Ohmoto et al., 2006). Stromatolites of undisputed biogenicity are common back to 2.8 Ga, and a strong case has been made that they extend to 3.4 3.5 Ga (e.g., Walter et al., 1980; Buick et al., 1981; Hofmann et al., 1999; Allwood et al., 2006a). However, given the scarcity of well-preserved microfossils before ca. 2.1 Ga (Buick, 2001), evidence for life on the early Earth is most often indirectly gleaned from carbon and sulphur

Preservation of biomarkers in oil inclusions for >2 billion years

845

isotopic compositions of sedimentary organic matter (e.g., Schidlowski, 2001; Shen et al., 2001; Hayes and Waldbauer, 2006), microbial traces and replacements (e.g., Rasmussen, 2000; Furnes et al., 2004; Allwood et al., 2006a) and vibrational spectroscopy of kerogen (e.g., Schopf et al., 2002; Allwood et al., 2006b). An additional powerful approach that has beneted from substantial technical advances in recent years is the use of biomarkers preserved in ancient organic matter to provide constraints on lifes development. Biomarkers are complex hydrocarbons derived from biological compounds in once living organisms that have been predictably altered during diagenesis (Peters et al., 2005). Their preservation in sediments, relatively unmetamorphosed rocks and crude oils thus allows the evolution of particular classes of organisms, their metabolisms and environmental niches to be followed through Earths history (Peters et al., 2005). Metamorphism causes breakdown of hydrocarbons and thus the number of old rocks suitable for study is limited and the record of biomarker evolution very early in time is distinctly incomplete (Hayes et al., 1983; Summons et al., 1988a,b; Summons and Walter, 1990; McKirdy and Imbus, 1992). To date, only trace amounts of biomarkers have been analysed from a small number of organic-rich pre-Mesoproterozoic shales (Summons et al., 1988b; Brocks et al., 1999, 2003a,d, 2005; Brocks and Summons, 2005; Ventura et al., 2007; Sherman et al., 2007; Eigenbrode, in press). Problems of contamination and syngeneity in the interpretation of biomarkers from Archaean shales were considered in detail by Brocks et al. (2003a), who discovered a suite of biomarkers in 2.7 Ga shales collected from drill core from the Pilbara Craton (Brocks et al., 1999, 2003a,b,c) and from which it was inferred that eukaryotes evolved 500 million to 1 billion years before the evidence in the fossil record. However, a small but noteworthy degree of uncertainty exists about the age of these biomarkers, which have been interpreted as only probably syngenetic with their Archean host rock (Brocks et al., 2003a). These doubts arise from the possibility that organic staining occurred during coring, as higher heavy hydrocarbon concentrations were recorded on the outside of the cores (Brocks et al., 2003a). Palaeoproterozoic rocks have also been analysed for their biomarker content; some of them contain steranes (Summons et al., 1988b) but others do not (Brocks et al., 2005), a dierence that could be interpreted to indicate that the steranes are contaminants (Brocks, 2005). Regardless of the contamination issue, what this emphasises is the small number of reliable biomarker studies spanning such a very long period of geological history (3.81.6 Ga). An exciting development in the use of molecular evidence for lifes early evolution is the ability to extract oil, including its constituent biomarkers, by using ultra-sensitive methods from the oil-bearing uid inclusions (FIs) that occur in many sedimentary rocks (George et al., 2007). The relative age of entrapment of FIs can be determined using textural relationships and known diagenetic, metamorphic and deformational histories. These tiny vessels are closed systems after entrapment and are shielded not only from contamination, but also from much of the alteration that

usually aects hydrocarbons over time (Dutkiewicz et al., 1998, 2003b; Volk et al., 2005). Oil-bearing FIs have been shown to have been trapped during the Proterozoic and the Archaean (e.g., Kelly and Nishioka, 1985; Dutkiewicz et al., 1998), and as such prove the contention that there was sucient biogenic kerogen produced by the primitive biosphere to form oil during thermal maturation (Buick et al., 1998; Rasmussen, 2005). One of the rst attempts to use gas chromatography (GC)mass spectrometry (MS) to analyse inclusion oil extracted by crushing was on FIs hosted in quartz crystals in calcite veins of unknown age from Precambrian metasediments in South-West Africa (Kvenvolden and Roedder, 1971). This pioneering study demonstrated the presence of hydrocarbon gases, n-alkanes and isoprenoids. The technology of GC analysis of FI extracts was developed by Burruss (1987) and has subsequently been applied in a wide range of petroleum systems studies (e.g., Karlsen et al., 1993; Nedkvitne et al., 1993; George et al., 1997; Scotchman et al., 1998; Karlsen et al., 2004). It has been applied to Proterozoic inclusion oils (e.g., Newell et al., 1993; George and Jardine, 1994; Mauk and Burruss, 2002), although none of these studies attempted to analyse biomarkers. Recently, oil inclusions from sandstones and a dolerite sill in the Roper Superbasin in the Northern Territory (Mesoproterozoic) and from the 2.1 Ga sandstone in Gabon that hosts the Oklo natural ssion reactors have been chemically analysed (Dutkiewicz et al., 2003b; Volk et al., 2003; Dutkiewicz et al., 2004; Volk et al., 2005; Dutkiewicz et al., 2006b; George et al., 2006; Dutkiewicz et al., 2007). These studies demonstrate the presence of abundant n-alkanes, monomethylalkanes, isoprenoids, alkylcyclohexanes, alkylcyclopentanes, aromatic hydrocarbons, bicyclic sesquiterpanes and hopanes. Tricyclic and tetracyclic terpanes, 2a-methylhopanes and steranes were detected in some of the oils. Here we report the molecular geochemistry of oil inclusions in a ca. 2.45 Ga uraniferous conglomerate, showing the reproducible detection of an extensive range of syngenetic hydrocarbons in the trapped FI oil, carefully controlled using a series of blanks. We demonstrate here that maturity-sensitive molecular ratios and structurally-complex biomarkers can survive for >2 billion years, despite heating to >280 C and burial to pressures of 200 MPa. The biomarker data are used to make inferences about the organisms that contributed to the organic matter deposited in the Palaeoproterozoic source rocks from which the analysed oil was generated and expelled. Finally, the relative merits of using oil inclusions as stable palaeobiological time capsules of early life on Earth are considered. 2. SAMPLE CHARACTERISTICS AND GEOLOGICAL SETTING The analysed sample is a uraniferous bitumen-bearing feldspathic quartz-pebble conglomerate from the Matinenda Formation of the lower Huronian Supergroup at Elliot Lake in Canada (Fig. 1). The host Huronian Supergroup is composed of up to 10 km of uvial, shallow marine, and possibly lacustrine, mainly siliciclastic sedimentary rocks deposited between 2.45 and 2.2 Ga in a rift basin

846

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870


83oE

Lake Huron
48oS

Gowganda

Sault Ste. Marie Blind River X Elliot Lake Sudbury


N

Espanola

sample site Huronian Supergroup

90 km

Lake Huron

Palaeozoic Nipissing Diabase (2.21 ga) Bar River Fm (shallow marine)

orrain Fm (fluvial deltaic)

1 km

Gowganda Fm (glaciomarine)

Espanola Fm (marginal marine) Bruce Fm (glaciomarine) ssissagi Fm (fluvial deltaic)

Fm (fluvial deltaic) amsay Lake Fm (glaciomarine) McKim Fm (fluvial-deltaic)

Matinenda Fm (fluvial-deltaic) Thessalon Fm (~2.45 Ga) Livingstone Creek Fm Basement (>2.48 Ga)

Argillite

Arenite

Volcanics

Carbonate

Diamictite

Granitoids and metamorphics

Red-beds

Sample site

Detrital uraninite and/or pyrite

Fig. 1. (a) Sample location map and geographic extent of the Huronian Supergroup, modied after Mossman (1987). (b) Stratigraphy of the Huronian Supergoup, showing that deposition of the sequence is constrained between 2.45 and 2.21 Ga (Young et al., 2001).

Elliot Lake Group

Hough Lake Group

Quirke Lake Group

Serpent Fm (fluvial deltaic)

Cobalt Group

n Lake Fm (shallow marine)

Preservation of biomarkers in oil inclusions for >2 billion years

847

(Fralick and Miall, 1989; Young et al., 2001). The sequence includes rocks deposited during three glacial periods (the Huronian diamictities), the middle of which includes a cap carbonate unit. These have been interpreted to represent low-latitude Snowball Earth glaciations (Williams and Schmidt, 1997; Schmidt and Williams, 1999), though this has been disputed (Hilburn et al., 2005; Kopp et al., 2005). Sulphur isotope stratigraphy shows that authigenic pyrites in the McKim and Pecors formations in the lower half of the sequence preserve small magnitude mass-independent isotope fractionation (MIF) whereas this signature is absent higher in the succession (Papineau et al., 2007). This has been interpreted as showing that the Great Oxidation Event, the rst major and permanent rise in atmospheric oxygen levels, occurred during this interval (Papineau et al., 2007). In the Elliot Lake area the Matinenda Formation was metamorphosed to a maximum metamorphic grade of low-pressure, low temperature upper prehnitepumpellyite

facies ($280350 C at 50200 MPa), as indicated by the mineral assemblage illitemontmorillonite, stilpnomelane, chlorite, muscovite, and pyrophyllite (Easton, 2000). Peak metamorphism may have been reached during the Penokean orogeny (Card, 1978) at 1.891.8 Ga (Young et al., 2001) or more likely during the earlier ($2.2 Ga) intrusion of the Nipissing Diabase suite (Mossman et al., 1993). Basin modelling suggests that a maximum thickness of only $200 m of Phanerozoic sedimentary cover was subsequently deposited over the Palaeoproterozoic rocks (Quinlan and Beaumont, 1984). The outcrop sample was taken directly above a U-mineralised conglomerate and consists of a moderately-sorted medium-grained feldspathic sandstone. It contains 5% by volume of 1 cm-diameter rounded quartz pebbles and several 1 mm-thick heavy mineral partings containing 0.5 mm sub-rounded pyrite and rounded 10 lm to 0.5 mm bitumen nodules, which based on their morphology and composition are interpreted to be pyrobitumen. The

Fig. 2. (af) Photomicrographs showing predominant setting in microfractures and the two types of oil-bearing FIs. (a,c,e) UVepiuorescence; (b,d,f) transmitted light. (a,b) Trails of oil-bearing uid inclusions in intragranular microfractures in detrital quartz. (c,d) H2O-dominated uid inclusion comprising a non-uorescing bubble of water vapour surrounded by a uorescing rim of oil and nonuorescing H2O liquid. (e,f) CO2-dominated uid inclusion with a uorescing rim of oil surrounding a double bubble of CO2 gas and liquid, and an outer non-uorescing rim of H2O liquid.

848

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

petrography and microthermometry of oil-bearing inclusions in this conglomerate have already been discussed (Dutkiewicz et al., 2003a), and initial geochemical data have been presented (Dutkiewicz et al., 2006a). Oil in the Matinenda Formation is hosted in four populations of uid inclusions, although only two populations are abundant and these appear to represent two episodes of oil migration (Fig. 2). The earlier of these FI populations is located in intragranular microfractures in quartz and K-feldspar or rarely within syntaxial quartz overgrowths, indicating entrapment early in the burial history. These inclusions are irregular, water-dominated (liquid and vapour), and contain 25% oil as uorescing lm on the vapour bubble (Fig. 2 a and b). Laser Raman spectroscopy indicated some CH4 but no CO2 in the vapour phase of these FIs, occasionally a solid calcite phase, and rarely a globule of bitumen. Based on gasliquid homogenisation temperatures, this FI population was interpreted to have been trapped during diagenesis at temperatures of 80220 C and at pressures of $50200 MPa (Dutkiewicz et al., 2003a). This FI population was subsequently heated to $280350 C during metamorphism, which caused a small proportion of the inclusions to stretch. The second FI population is highly carbonic, consisting of 3090 vol% CO2-bearing vapour phase, usually some water, and 23 vol% uorescing oil as a meniscus to the CO2 (Fig. 2e and f). Laser Raman spectroscopy showed that the vapour phase of these FIs contained up to 50% CH4, with 28 mol% C2H6 and 15 mol% C3H8. This FI population was interpreted to have been trapped during metamorphism at temperatures of 280350 C and pressures of $100150 MPa (Dutkiewicz et al., 2003a). Petrographic evidence shows that both populations of FI were entrapped during diagenesis or during the pro-grade metamorphic history of the host rock. In view of the likely age of metamorphism, it can be inferred that all Matinenda Formation FI oils were trapped prior to 2.2 Ga. The reported geochemical data thus provide a window into the composition of early Palaeoproterozoic oils and the nature of the organic matter from which they were generated. 3. ANALYTICAL METHODS 3.1. Sample preparation and cleaning The conglomerate was fragmented into 2 mm pieces and thoroughly cleaned with excess amounts of hydrogen peroxide, Aqua Regia and hot chromic acid (George et al., 2007). After washing with distilled water, the rock fragments (21.83 g) were ultrasonicated (10 min) sequentially with triplicate 30 mL aliquots of methanol, dichloromethane (DCM)/methanol (93:7) and DCM. These washings were discarded, and then the rock fragments were ultrasonicated (10 min) with three batches of DCM (30 mL), which were combined to form the rst outside-rinse blank. This was puried on a micro-column composing a Pasteur pipette lled with glass wool and silica gel (C60: 60 210 lm), which was ushed with DCM. The rst outsiderinse blank was spiked with $0.6 lg of an internal standard (squalane), and analysed for cleanliness by gas chromatographymass spectrometry (GCMS) on a Hewlett Packard

5890 gas chromatograph (DB5MS, 60 m 0.25 mm i.d. 0.25 lm lm thickness) interfaced to a VG AutospecQ Ultima mass spectrometer [for GC programmes and typical MS parameters used, see George et al. (2007)]. The rst outside-rinse blank was found to still contain traces of hydrocarbons, so the above procedure was repeated to produce a second outside-rinse blank. This too contained traces of hydrocarbons, so the above procedure was repeated a third time to produce the nal outside-rinse blank, which was deemed to be suciently clean. Results from this nal outside-rinse blank are presented in this paper for comparison with the o-line crushing data from the oil inclusions. 3.2. O-line crushing An o-line crushing method was used for the analysis of traces of high molecular-weight C12+ hydrocarbons such as biomarkers (George et al., 1998, 2007). The entire procedure was carried out in replicate, including acquisition and processing of data from the system blanks. Before the inclusion uids were extracted, procedural system blank experiments using the same experimental conditions and glassware (including the same amount of solvent) were carried out in order to determine the extent of any hydrocarbon contribution from the crushing cylinder and any carry-over from previously analysed samples or from the general laboratory background. The system blanks were spiked with $0.6 lg of an internal standard (squalane), and analysed for cleanliness by GCMS as described above. System blanks that were deemed to be suciently clean were acquired prior to crushing of the rock fragments, and the sample was subsequently crushed using exactly the same glassware. Cleaned rock fragments (15.18 g and 12.31 g aliquots from the same sample) were crushed under DCM in a stainless steel crushing cylinder with a 55 mL capacity (George et al., 2007). Two stainless steel balls were placed in the cylinder above the sample aliquot and 25 mL of DCM was added before closure of the cylinder in air at atmospheric pressure. The cylinder was vigorously shaken for 2 10 min in a vertical motion with a throw of about 40 mm, using a modied piston pump. In between the two crushing phases, the crusher was allowed to cool for at least 10 min in order to minimise evaporative loss of the leached FI oil. The resultant suspension of nely crushed mineral in DCM was poured into a beaker, and the crushing cylinder was rinsed with an additional 25 mL of DCM. The solution in the beaker was then ultrasonicated for 10 min, allowed to settle for a few minutes and then the supernatant solvent layer, containing the FI extract and suspended mineral nes, was transferred to a round bottom ask. The residual mineral powder was ultrasonicated twice with fresh DCM (25 mL), and the supernatant was transferred into the same round bottom ask. The solvent containing the oil extracted from the FIs (total = 100 mL) was reduced in a rotary evaporator and by blowing down with puried nitrogen. Suspended rock powder was removed by passing the extract reduced to $2 mL through a short Pasteur pipette plugged with glass wool and packed with silica gel (C60: 60210 lm).

Preservation of biomarkers in oil inclusions for >2 billion years

849

The amount of oil obtained from each FI oil sample was determined by adding small amounts ($0.6 lg) of an internal standard (squalane), since the yields were too small for gravimetric determination. Single-ion monitoring (SIM) and metastable-reaction monitoring (MRM) modes of detection were used. Particular care was taken to avoid evaporating the extracted FI oil to dryness, so as to preserve the low molecular weight hydrocarbons (see Ahmed and George, 2004). Generally, compounds down to $C8 were recovered using the o-line crushing technique, with those from C12 to C36 recovered quantitatively. 3.3. On-line crushing Whereas the o-line crushing method was advantageous for the analysis of traces of C12+ hydrocarbons such as biomarkers, low molecular weight compounds (C12) were only partially retained during sample work-up using the o-line crushing method. Therefore, the molecular compoi-Pentane n-Pentane Furan 2,2-Dimethylbutane Cyclopentane 2,3-Dimethylbutane 2-Methylpentane 3-Methylpentane n-Hexane 3-Methylfuran 2-Methylfuran 2,2-Dimethylpentane Methylcyclopentane 2,4-Dimethylpentane 2,2,3-Trimethylbutane 3,3-Dimethylpentane Cyclohexane Benzene 2-Methylhexane 2,3-Dimethylpentane 1,1-Dimethylcyclopentane 3-Methylhexane cis-1,3-Dimethylcyclopentane trans-1,3-Dimethylcyclopentane trans-1,2-Dimethylcyclopentane n-Heptane Methylcyclohexane 2,5-Dimethylhexane 2,4-Dimethylhexane Ethylcyclopentane 1,2,4-Trimethylcyclopentane 1,2,3-Trimethylcyclopentane 2,3-Dimethylhexane + 2-Methyl-3-ethylpentane 2-Methylheptane 4-Methylheptane Toluene 3-Methylheptane 1-cis-3-Dimethylcyclohexane 1-trans-4-Dimethylcyclohexane 1,1-Dimethylcyclohexane 1-trans-2-Dimethylcyclohexane n-Octane 1-trans-3- + 1-cis-4-Dimethylcyclohexane Ethylcyclohexane Ethylbenzene 4-Methyloctane 2-Methyloctane 3-Methyloctane meta- + para-Xylene 1-trans-4-Ethylmethylcyclohexane 1-cis-4-Ethylmethylcyclohexane ortho-Xylene n-Nonane

sition of gasoline range hydrocarbons (C5 to C9) in the FIs was assessed in replicate using a direct on-line crushing method (Ruble et al., 1998; Volk et al., 2002; Dutkiewicz et al., 2004). Small amounts of the cleaned rock fragments ($50 mg) were hand crushed in the glass-lined metal insert of a Quantum MicroScale Sealed Vessel (MSSV)-2 Thermal Analysis System (Hall Analytical Laboratories, Manchester) using a metal plunger. Liberated compounds were thermally extracted in a helium ow at 300 C, and focussed in a cryogenic trap prior to GC separation (George et al., 1998, 2007). Chromatography was performed using a BPX5 column (5% phenyl 95% methyl silicone, 0.50 lm lm thickness, SGE), with the oven programmed from an initial temperature of 20 C at the start of the crushing (7 min hold), followed by heating at 4 C min1 to 30 C, with a 8 min hold, followed by heating at 4 C min1 to 300 C, with a 15 min hold. The cryogenic trap was removed after 5 min and ballistically heated to 300 C. Compounds liberated from the inclusions were analysed by the
i-Pentane n-Pentane Furan 2,3-Dimethylbutane 2-Methylpentane 3-Methylpentane n-Hexane 3-Methylfuran 2-Methylfuran 3,3-Dimethylpentane Benzene 2-Methylhexane 2,3-Dimethylpentane 3-Methylhexane n-Heptane Toluene Ethylbenzene meta- + para-Xylene ortho-Xylene

(a)

(b)

Benzene off-scale (419)

0 50 100 Normalised abundance to toluene (FID-equivalent)


nC5 nC6 nC7 nC8 Toluene off-scale (249) nC9 nC10 nC11 nC12 nC13 nC14 nC15 nC16 nC17 nC18 nC19 nC20 nC21 Run A Run B Run C

(c)

0 50 100 Normalised abundance to iso -pentane (FID-equivalent)

140 50 100 Normalised abundance to n -C14

Fig. 3. (a) Relative abundance (FID-equivalent data) of gasoline range hydrocarbons analysed by on-line crushing (MSSV-2; Run A) of the Matinenda Formation uid inclusion oil, normalised to iso-pentane. Note that benzene and toluene are very abundant and plot o-scale. (b) Previous and superseded data (Dutkiewicz et al., 2006a) on the relative abundance (FID-equivalent data) of gasoline range hydrocarbons analysed by on-line crushing (MSSV-1) of the Matinenda Formation FI oil, normalised to toluene. (c) Comparison of n-alkane (C4 to C21) abundance calculated from m/z 57 mass chromatograms for three on-line crushing experiments (MSSV-2) of the Matinenda Formation FI oil, each normalised to n-C14. Run A is the main data shown in (a), acquired using SIM; Run B is an earlier SIM run on the same sample; and Run C is a later magnet-scan run.

850

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870 Table 1 Low molecular weight parameters for the Matinenda FI oil, derived from on-line crushing C1:C2:CO2:C3:RC45 hydrocarbons i-C5/n-C5 Benzene/n-C6 Toluene/n-C7 Furan/n-C6 (n-C6 + n-C7) /(cyclohexane + methylcyclohexane) Heptane value Isoheptane value n-C7/methylcyclohexane Cyclohexane/methylcyclopentane n-C7/2-methylhexane n-C7/methylcyclopentane Methylcyclohexane/toluene 3-Methylpentane/n-C6 Benzene/toluene Methylcyclopentane/methylcyclohexane Toluene/o-xylene m- + p-xylene/n-C8 K1: (2-MH + 2,3-DMP)/(3-MH + 2,4-DMP) 2-MH/3-MH 2,4-DMP/2,3-DMP Ctemp (C) K2:P3/(P2 + N2) N2/P3 n-C7 (% of total C7 hydrocarbons) 3 Ring preference % (DMP + MH) 5 Ring preference % (DMCP + ECP) 6 Ring preference % (methylcyclohexane + toluene) 40.7:2.0:55.3:0.7:1.3 1.5 9.0 6.1 0.44 2.5 33 2.2 2.0 0.83 2.7 2.2 0.1 0.30 1.7 0.90 6.9 3.3 2.1 1.8 0.22 118 0.25 1.41 11.3 10 6 84

AutoSpec GCMS using two function SIM and magnet scan programmes. Areas of peaks in mass chromatograms were converted to ame ionisation detector (FID)-equivalent data using response factors before data manipulation (e.g., George et al., 2004c), enabling comparison of these types of data with conventional whole oil GC-FID analyses. Initial on-line crushing data reported by Dutkiewicz et al. (2006a) were acquired using an MSSV-1 system, using similar conditions as above; these data are also presented and discussed here (Fig. 3b) . The molar distribution of gaseous compounds (C1 to C5, including CO2) in the FIs was investigated using a MSSV-1 on-line crushing system with a GS-GASPRO capillary column (30 m 0.32 mm i.d.). Another aliquot of cleaned rock fragments ($50 mg) was used. The oven was programmed from an initial temperature of 20 C at the start of the crushing (4 min hold), followed by heating at 4 C min1 to 260 C, with a 15 min hold. The liquid nitrogen trap was removed after 2 min. Gases liberated from the FIs were analysed by the AutoSpec GCMS using magnet scanning from m/z 1075 with data for each analyte extracted from mass chromatograms (C1: m/z 15; C2: m/z 27; C3 and C4: m/z 43; C5: m/z 57; CO2: m/z 44). The analysis system was calibrated using injections of a natural gas standard (BOC) under the same detection conditions, from which response factors were calculated and applied to correct the FI gas data to molar composition. 3.4. Constraints on the methodology The main constrain on the methodology is that it is impossible to physically separate the two FI populations, so the geochemical analyses reported in this paper pertain to mixtures of both FI populations. The geochemical analyses reported herein may also include a contribution from pyrobitumens in the sample, as these may have been present in the small rock fragments crushed. However, the very low hydrocarbon levels in the nal outside rinses constrain this contribution to trace amounts only. Very rare bitumen also occurs with oil in a few of the older, syn-diagenetic FIs (Dutkiewicz et al., 2003a), so would also have contributed to the geochemical analyses, but only in a very minor way based on their low abundance. The petrographic data show that no FI oils or bitumens were trapped subsequent to metamorphism. Therefore, the reported geochemical data pertain to Palaeoproterozoic FI oil, with a possible minor contribution from solid bitumens of similar age but less reliably shielded from contamination and degradation. 4. RESULTS 4.1. Geochemical data from on-line crushing method 4.1.1. C1 to C5 gases On-line crushing to release the FI oil from both populations revealed that gaseous compounds are dominated by CH4 and CO2, with lesser amounts of C2, C3 and RC45 (Table 1). The CH4:CO2 ratio is intermediate between that of the two FI types (aqueous and CO2-rich) determined by Laser Raman spectroscopy, and shows that the analysed oil

All data are expressed as FID-equivalent and are derived from corrected SIM runs, except for (mol%). The heptane value (H; Thompson, 1979) is 100 n-C7/R cyclohexane through to methylcyclohexane. The isoheptane value (Thompson, 1983) is (2-MH + 3-MH)/(c-1,3- + t-1,3- + t-1,2DMCP). Ctemp (C) (Bement et al., 1995; Mango, 1997) is 140 + (15 (ln [2,4-DMP/2,3-DMP])). P2 = 2-MH + 3-MH; P3 = 3,3-DMP + 2,3-DMP + 2,4-DMP + 2,2-DMP; N2 = 1,1DMCP + c-1,3-DMCP + t-1,3-DMCP. MH, methylhexane; DMP, dimethylpentane; DMCP, dimethylcyclopentane; ECP, ethylcyclopentane.

is derived from both FI populations (aqueous and CO2rich), as would be expected from the ubiquitous distribution of both FI populations in the sample (Dutkiewicz et al., 2003a, 2006a) and the crushing technique. 4.1.2. Gasoline range C5 to C9 compounds and higher n-alkanes Most theoretically possible C5 to C7 hydrocarbon isomers and several C8 to C9 hydrocarbons were detected in the FIs using SIM (Fig. 3a). Abundances of these compounds are reported as FID-equivalent data, and derived ratios are given in Table 1. Benzene and toluene were the most abundant compounds detected in this mass range, with signicant amounts of iso-pentane, C5 to C9 n-alkanes, xylenes, ethylbenzene and 3-methylfuran also present. Furan, 3-methylfuran and 2-methylfuran are oxygenated com-

Preservation of biomarkers in oil inclusions for >2 billion years


? 18 19 17 20 21 16 Squalane 24 25 27 26 23 28 29

851

(a)

22

17 19 20 18 8+9+10 Ph Pr 6+7 4 6+7+8+9 6+7+8+9 5 3 54 3 2 4 2 52 3

(b) -

10 9 11 14 12 13 UCM 15 Pr Ph

30

31 32

Response

FI oil: 88.3 ng C12C32 n-alkanes


33 34

Blank: 9.6 ng C12C32 n-alkanes

Rinse: 19.8 ng C12C32 n-alkanes

Retention time

(c)

n-Alkanes

Fig. 4. m/z 85 Mass chromatograms of the Matinenda Formation FI oil, system blank and nal outside rinse blank drawn to the same scale, showing (a) the distribution of n-alkanes (n-C8 to n-C34) and isoprenoids (Pr, pristane; Ph, phytane; UCM, unresolved complex mixture; and (b) detailed monomethylalkane distribution between n-C17 and n-C20 (MHeD, methylheptadecanes; MOD, methyloctadecanes; MND, methylnonadecanes). Data were obtained by o-line crushing, with squalane added as an internal standard. Quantitative amounts of C12C32 n-alkanes recovered in each fraction are also given in (a). The line chart (c) summarises the n-alkane recovery for the three fractions.

pounds not normally detected in crude oils. They are highly water soluble, and their presence in FIs, together with high abundances of the most water-soluble hydrocarbons (benzene, toluene and xylenes), has been attributed to the co-analysis of aqueous inclusions during on-line crushing (Ruble et al., 1998). The dominance of these compounds in the Matinenda FIs is consistent with the co-entrapment of limited amounts of oil with oil-equilibrated waters in the three-phase FIs (Dutkiewicz et al., 2006a). Most C6 to C9 parameters reect this excess of aromatic hydrocarbons and the relatively low amounts of branched alkanes, alkylcyclopentanes and alkylcyclohexanes (Fig. 3 and Table 1). Ratios from a SIM run using the MSSV-1 equipment were reported as ratios in Dutkiewicz et al. (2006a) and are plotted in Fig. 3b, but should be considered superseded by the data acquired using an upgraded on-line extraction unit (MSSV-2) in Table 1. The MSSV-1 analytical run is strongly dominated by the water-soluble compounds, with only very low relative abundances of n-alkanes and branched alkanes reliably detected. The inter-analysis heterogeneity in analytical data from the on-line crushing method is a consequence of the small sample size used

($50 mg). However, two SIM runs and a third magnet scan run (Run C) on the MSSV-2 give a reproducible n-alkane maximum at C14, with tail o above $C17 and more variable <C10 relative abundances (Fig. 3c). The dierence in the apparent n-alkane maxima between the on-line (C14) and the o-line crushing methods (C18; Fig. 4) is a combined consequence of (1) adsorption processes in the MSSV inlet, leading to reduction in the amount of high molecular weight compounds recovered during on-line crushing (George et al., 1996), and (2) partial evaporation during sample work-up after o-line crushing, which may reduce the amount of recovered <C15 compounds (e.g., Karlsen et al., 1993; Ahmed and George, 2004). 4.2. Geochemical data from o-line crushing method 4.2.1. Linear and branched alkanes The distribution of n-alkanes and branched alkanes in the Matinenda Formation FI oil and associated system and nal outside-rinse blanks is shown in m/z 85 mass chromatograms, scaled using the squalane standard so that relative amounts in the three fractions can be visualised

852

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

(Fig. 4). The FI oil contains 9.2 times more C12C32 n-alkanes relative to the system blank and 4.5 times more relative to the nal outside-rinse blank. Furthermore, the intran-alkane peaks are considerably more evident and more numerous in the FI oil, with a series of prominent peaks between each n-alkane from C19 to C29, including a distinct unresolved complex mixture (UCM) hump. Some of these peaks are monomethylalkanes, as identied in Fig. 4b. The second and duplicate analysis of Matinenda Formation FI oil was unsatisfactory for detailed interpretation, (1) because lower amounts of C12C32 n-alkanes were recovered relative to the duplicate system blank (3.0) and nal outside-rinse blank (1.3), and (2) because lower abundances and a more limited range of compounds was reliably detected in the duplicate run. Generally, the compounds that could be reliably detected in the duplicate analysis of Matinenda Formation FI oil have a similar distribution to those in the original analysis reported in full here. The inter-aliquot heterogeneity in analytical data from the o-line crushing method reects the extremely low amounts of compounds recovered from the Matinenda Formation FI oil: only 5.8 ng C12C32 n-alkanes/g rock fragments crushed were obtained (rst aliquot), compared to typical FI oils

from Phanerozoic oil columns where yields of 80 3000 ng/g are common (George et al., 2001). Although this sample has among the lowest yields of the samples successfully analysed at the CSIRO labs, by careful comparison of responses for the FI oil, the system blank and the nal outside-rinse blank, a wide range of hydrocarbons could be shown to be unambiguously present in the FI oil. The low molecular weight (C8C12) n-alkanes as determined from the o-line crushing data in the Matinenda Formation FI oil are mainly from a system blank contribution (Fig. 4c) and so are not interpretable. The n-alkanes maximise at C18, and could be detected to C34, and in the C17C32 region are substantially more abundant than in the system blank and the nal outside-rinse blank (Fig. 4c). There is a predominance in high molecular weight odd n-alkanes, especially at C27 and C29 (Table 2), and this is also apparent in the nal outside-rinse blank, albeit at lower concentrations (Fig. 4). Most Precambrian samples, although not all (Li et al., 2003), have low amounts of high molecular weight n-alkanes without any odd or even number predominance (Summons et al., 1988b; Brocks et al., 2003a). One interpretation of this observation is contamination of the Palaeoproterozoic FI oil sample by low matu-

Table 2 Alkane and aromatic hydrocarbon parameters for the Matinenda FI oil (o-line data) Carbon preference index (CPI2232) Pr/Ph (Pristane/phytane) Pr/n-C17 Ph/n-C18 RMethylheptadecanes/n-C18 3-Methylheptadecane/4-methylheptadecane 2-Methylheptadecane/5-methylheptadecane RMethylheneicosane/n-C22 3-Methylheneicosane/4-methylheneicosane 2-Methylheneicosane/5-methylheneicosane RMethyltricosanes/n-C24 3-Methyltricosane/4-methyltricosane 2-Methyltricosane/5-methyltricosane Methylnaphthalene ratio (MNR: 2-MN/1-MN) Ethylnaphthalene ratio (ENR: 2-EN/1-EN) Dimethylnaphthalene ratio-1 (DNR-1: [2,6- + 2,7-DMN]/1,5-DMN) Trimethylnaphthalene ratio-1 (TNR-1: 2,3,6-TMN/[1,4,6- + 1,3,5-TMN]) Trimethylnaphthalene ratio-2 (TNR-2: [2,3,6- + 1,3,7-TMN]/[1,4,6- + 1,3,5- + 1,3,6-TMN]) Tetramethylnaphthalene ratio-1 (TeMNR-1: 2,3,6,7-TeMN/1,2,3,6-TeMN) Trimethylnaphthalene ratio (TMNr: 1,3,7-TMN/[1,3,7- + 1,2,5-TMN]) Tetramethylnaphthalene ratio (TeMNr: 1,3,6,7-TeMN/[1,3,6,7- + 1,2,5,6-TeMN]) Methylbiphenyl ratio (MBpR: 3-MBp/2-MBp) Dimethylbiphenyl ratio-x (DMBpR-x: 3,5-DMBp/2,5-DMBp) Dimethylbiphenyl ratio-y (DMBpR-y: 3,30 -DMBp/2,30 -DMBp) Methylphenanthrene index (MPI-1: 1.5 [3-MP + 2-MP]/[P + 9-MP + 1-MP]) Methylphenanthrene ratio (MPR: 2-MP/1-MP) Methylphenanthrene distribution fraction (MPDF: [3-MP + 2-MP]/RMPs) Dimethylphenanthrene ratio (DMPR: [3,5- + 2,6- + 2,7-DMP]/[1,3- + 3,9- + 2,10- + 3,10- + 1,6- + 2,9- + 2,5-DMP]) Dibenzothiophene/phenanthrene RMethyldibenzothiophenes/RMethylphenanthrenes RDimethyldibenzothiophenes + ethyldibenzothiophenes/RDimethylphenanthrenes + ethylphenanthrenes Methyldibenzothiophene ratio (MDR; 4-MDBT/1-MDBT) Dimethyldibenzothiophene ratio (DMDR: 4,6-DMDBT/3,6- + 2,6-DMDBT) 1.13 1.08 0.42 0.33 0.30 1.03 1.02 1.37 1.33 0.85 1.44 1.18 0.98 2.3 1.6 5.7 0.79 0.90 0.63 0.70 0.54 3.0 6.3 8.8 0.53 1.7 0.59 0.48 0.12 0.31 0.51 2.1 0.75

All data were derived from SIM runs: alkanes, m/z 85; for aromatic hydrocarbons, see Figs. 58 for identity of SIM run used. Response factors were used in the calculation of Pr/Ph, Pr/n-C17, Ph/n-C18 and MPI-1.

Preservation of biomarkers in oil inclusions for >2 billion years

853

rity lipids from a post-Devonian land plant source (e.g., Eglinton and Hamilton, 1967). However, no other evidence in the geochemical composition of the FI oil suggests either low maturity or land plant contamination. Another possibility is that the n-alkanes were formed by dissolution in solvent and recrystallisation as waxes in remaining fractures during disaggregation and cleaning of the original sample (Karlsen and Skeie, 2006). The diculty of removing strongly adsorbed components of oils from cracks and ssures in mineral grains has previously been discussed (Jones and Macleod, 2000; Karlsen et al., 2004). Blank levels in other compound groups are generally lower than for the n-alkanes, providing greater condence that they are present in the FIs.

The amount (ng recovered) of branched alkanes, including C11C25 monomethylalkanes, in the FI oil is signicantly greater than for the system blank or the nal outside-rinse blank, and the distribution is also quite dierent, especially for >C19 (Fig. 5). There is a greater outsiderinse blank contribution at lower molecular weights, especially at C11, C12, C16 and C17, so these data are interpreted more cautiously. The FI oil C20C25 monomethylalkanes are strongly dominated by the co-eluting peak due to the mid-chain isomers, and the 4- and 5-methylalkanes have a similar abundance as the 3- and 2- (terminal) methylalkanes (Table 2). The proportion of total monomethylalkanes relative to the n-alkane with the same carbon number increases with increasing molecular weight (Fig. 4a and

ng recovered (5-methyldecane to 3-methylhexadecane) 0


5-Methyldecane 4-Methyldecane 2-Methyldecane 3-Methyldecane 5+6-Methylundecane 4-Methylundecane 2-Methylundecane 3-Methylundecane iC13 5-+6-Methyldodecane 4-Methyldodecane 2-Methyldodecane 3-Methyldodecane + iC14 6-+7-Methyltridecane 5-Methyltridecane 4-Methyltridecane 2-Methyltridecane 3-Methyltridecane iC15 6-+7-Methyltetradecane 5-Methyltetradecane 4-Methyltetradecane iC16 2-Methyltetradecane 3-Methyltetradecane 6-+7-+8-Methylpentadecane 5-Methylpentadecane 4-Methylpentadecane 2-Methylpentadecane 3-Methylpentadecane 6-+7-+8-Methylhexadecane + iC18 5-Methylhexadecane 4-Methylhexadecane 2-Methylhexadecane 3-Methylhexadecane Pr 6-+7-+8-+9-Methylheptadecane 5-Methylheptadecane 4-Methylheptadecane 2-Methylheptadecane 3-Methylheptadecane Ph 6-+7-+8-+9-Methyloctadecane 5-Methyloctadecane 4-Methyloctadecane 2-Methyloctadecane 3-Methyloctadecane 6-+7-+8-+9-+10-Methylnonadecane 5-Methylnonadecane 4-Methylnonadecane 2-Methylnonadecane 3-Methylnonadecane 6-+7-+8-+9-+10-Methyleicosane 5-Methyleicosane 4-Methyleicosane 2-Methyleicosane 3-Methyleicosane 6-+7-+8-+9-+10-+11-Methylheneicosane 5-Methylheneicosane 4-Methylheneicosane 2-Methylheneicosane 3-Methylheneicosane 6-+7-+8-+9-+10-+11-Methyldocosane 5-Methyldocosane 4-Methyldocosane 2-Methyldocosane 3-Methyldocosane 6-+7-+8-+9-+10-+11-+12-Methyltricosane 5-Methyltricosane 4-Methyltricosane 2-Methyltricosane 3-Methyltricosane 6-+7-+8-+9-+10-+11-+12-Methyltetracosane 5-Methyltetracosane 4-Methyltetracosane 2-Methyltetracosane 3-Methyltetracosane

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

FI oil System blank Final outside-rinse blank

2 3 4 5 ng recovered (Pr to 3-methyltetracosane)

Fig. 5. Semi-quantitative abundance (ng recovered) of branched alkanes in the Matinenda Formation FI oil, system blank and nal outsiderinse blank. iCxx refers to the carbon number of isoprenoids, Pr = pristane, Ph = phytane. The lower molecular weight hydrocarbons report to the upper axis, hydrocarbons with retention times > Pr report to the lower axis. The data are semi-quantitative because no response factors were used to correct for response variation in m/z 85 between the squalane standard and the analytes.

854

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

Table 2). C13 to C20 isoprenoids were detected in the FI oil, but only pristane (Pr) and phytane (Ph) have a signicant abundance relative to adjacent n-alkanes (Figs. 4a and 5). 4.2.2. Aromatic hydrocarbons C2C4 alkylbenzenes were detected in the Matinenda Formation FI oil, but are also present in signicant quantities in the system blank and nal outside-rinse blank and hence are not interpretable. Naphthalene is also aected by a high blank (Fig. 6). The high background of C8C10 aromatic hydrocarbons in the blanks corresponds to the molecular weight range of high n-alkane blanks and indicates a consistent external source of contamination, possibly related to the greater water solubility of C8C10 aromatic hydrocarbons than those of higher molecular weight. Fortunately, higher molecular weight aromatic hydrocarbons in the Matinenda Formation FI oil are more abundant relative to the blanks, and thus can be interpreted more readily. Alkylnaphthalenes detected include isomers with one to four methyl side-chains, the scaled mass chromatograms for which demonstrate higher abundance in the Matinenda Formation FI oil relative to the blanks
Naphthalene

(Fig. 6). Biphenyl, diphenylmethane, dibenzofuran, methylbiphenyls, methyldiphenylmethanes, ethylbiphenyls and dimethylbiphenyls are present in the Matinenda Formation FI oil (Fig. 7). These compounds are present in the system blank and the nal outside-rinse blank, but except for biphenyl, occur are at much reduced levels. Signicantly more phenanthrene, methylphenanthrenes and dimethylphenanthrenes were detected in the Matinenda Formation FI oil compared to the system blank and the nal outside-rinse blank (Fig. 8). Thus, the alkylphenanthrene distributions are interpreted to represent a genuine record from the uid inclusions. Alkyldibenzothiophenes were also detected in the Matinenda Formation FI oil, and are much less abundant in the blanks (Fig. 9). These sulphur-containing compounds in the FI oil are less abundant than phenanthrene and the alkylphenanthrenes, although this dierence reduces with increasing methylation (Table 2). 4.2.3. Terpanes The bicyclic sesquiterpanes drimane and homodrimane, together with other rearranged C15 and C16 isomers, are present in the Matinenda Formation FI oil (Fig. 10). The

2-MN 1,3-+1,7-DMN 1,6-DMN 2,7-DMN 2,6-DMN 1-MN 1-EN 2-EN TMNs 1,3,5-+1,4,62,3,61,3,61,2,71,3,71,4-+2,31,6,7DMN 1,2,61,5-DMN 1,2,41,2-DMN

(a: FI oil)
TeMNs 1,2,4,6- + 1,2,4,7- + 1,4,6,71,2,5,72,3,6,7? 1,2,6,71,2,3,71,2,3,61,3,6,71,2,51,2,5,6-+ 1,2,3,5-

m/z 128 m/z 142 m/z 156 m/z 170

Response

m/z 184.1

(b: Blank )
m/z 128 m/z 142

m/z 156 m/z 170 m/z 184.1

(c : Rinse)
?

m/z 128

m/z 142 m/z 156 m/z 170 m/z 184.1

Retention time

Fig. 6. Partial m/z 128, 142, 156, 170 and 184.1 mass chromatograms of Matinenda Formation (a) FI oil, (b) system blank, and (c) nal outside-rinse blank, all drawn to the same scale, showing the distribution of naphthalene, methylnaphthalenes (MN), ethylnaphthalenes (EN), dimethylnaphthalenes (DMN), trimethylnaphthalenes (TMN) and tetramethylnaphthalenes (TeMN).

Preservation of biomarkers in oil inclusions for >2 billion years


Biphenyl

855

(a: FI oil)
3-MBp

3,3'-DMBp 4-EBp 4-MBp DBF 3,4'-DMBp

m/z 154

2-MBp DPM

3,5-DMBp 2,3-DMBp + 3-MDPM 2-MDPM 4-MDPM

m/z 168

Response

2,5-DMBp 2,4- + 2,4'2,3'-DMBp DMBp

3-EBp

4,4'DMBp

m/z 182

(b: Blank )
m/z 154 m/z 168 m/z 182

(c : Rinse)
m/z 154 m/z 168 m/z 182

Retention time

Fig. 7. Partial m/z 154, 168 and 182 mass chromatograms of Matinenda Formation (a) FI oil, (b) system blank, and (c) nal outside-rinse blank, all drawn to the same scale, showing the distribution of biphenyl, diphenylmethane (DPM), methylbiphenyls (MBp), dibenzofuran (DBF), methyldiphenylmethanes (MDPM), ethylbiphenyls (EBp) and dimethylbiphenyls (DMBp).

Phenanthrene 2-MP

(a: FI oil )
1,3- + 3,9- + 3,10- + 2,10-DMP 1,6- + 2,9- + 2,5-DMP 2,71,7-DMP 3,5- + 2,6-DMP DMP 9- + 2- +1-EP + 3,6-DMP 2,3- + 1,9- + 4,9- + 4,10-DMP 1,8- 1,2DMP DMP

3-MP 9-MP 1-MP

Response

m/z 178 m/z 192 m/z 206


3-EP

(b: Blank )
m/z 178 m/z 192 m/z 206

(c : Rinse)
m/z 178 m/z 192

m/z 206 Retention time

Fig. 8. Partial m/z 178, 192 and 206 mass chromatograms of Matinenda Formation (a) FI oil, (b) system blank, and (c) nal outside-rinse blank, all drawn to the same scale, showing the distribution of phenanthrene, methylphenanthrenes (MP), ethylphenanthrenes (EP) and dimethylphenanthrenes (DMP).

856

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870


Dibenzothiophene 4-MDBT 2,6-DMDBT 2,4-DMDBT 2+3-MDBT 1-MDBT 4,6-DMDBT 3,6-DMDBT 3,7- + 1,4- + 1,6- + 1,8-DMDBT

(a: FI oil)

1,3- + 1,9- + 1,2-DMDBT

Response

m/z 184.0 m/z 198.0 m/z 212

4-EDBT

(b: Blank )

(c : Rinse)

Retention time

Fig. 9. Partial m/z 184.0, 198.0 and 212 mass chromatograms of Matinenda Formation (a) FI oil, (b) system blank, and (c) nal outside-rinse blank, all drawn to the same scale, showing the distribution of dibenzothiophene, methyldibenzothiophenes (MDBT), ethyldibenzothiophenes (EDBT) and dimethyldibenzothiophenes (DMDBT).

(a: FI oil )
Drimane ? C15

Homodrimane ? ? C15 C16

(b: Blank)

(c : Rinse)

Retention time

Fig. 10. Partial m/z 123 mass chromatograms of Matinenda Formation (a) FI oil, (b) system blank, and (c) nal outside-rinse blank, all drawn to the same scale, showing the distribution of bicyclic sesquiterpanes, including drimane and other C15 isomers, and homodrimane and another C16 isomer.

much lower abundance and dierent distribution of bicyclic sesquiterpanes in the system blank and the nal outsiderinse blank is evidence that the FI oil signal is indigenous. Homodrimane is more abundant than drimane or the C15 isomers (Table 3). Fig. 11 shows the original (a) and duplicate (b) analyses of tricyclic terpanes, tetracyclic terpanes and hopanes in the Matinenda Formation FI oil, compared to the respective system and nal outside-rinse blanks. Several features show that these biomarkers are dominantly from the FI oil. First, the amount of tricyclic terpanes, tetracyclic terpanes and hopanes is considerably greater in the FI oil than in the sys-

tem blank and nal outside-rinse blank, for both the original and the duplicate analyses. Secondly, the distribution of terpanes and other peaks in the m/z 191 mass chromatograms are dierent in the system blank and the nal outside-rinse blank compared to the FI oil. Thirdly, an unusual baseline hump around the position of the C28 tricyclic terpanes is present in both analyses of the FI oil, but is not present in any of the blanks. Fourthly, the distribution of terpanes in the duplicate run is generally quite similar to that in the original run, albeit with a lower response and some additional and undesignated peaks in the duplicate run. The duplicate analysis of the terpanes is not used further for interpretation, because of the poorer data quality, but it is suciently similar to the original run for it to be concluded that interpretations based on terpane distributions in the duplicate would be the same as those based on the original analysis. The tricyclic terpanes in the Matinenda Formation FI oil are dominated by the C23 isomer, with signicant amounts of the C21 and C24 isomers (Fig. 11). The hopane distribution in the Matinenda Formation FI oil is dominated by C29 ab hopane, with lesser amounts of C30 ab hopane, the C27 isomers Ts and Tm and extended hopanes (Fig. 11). The distribution of some of the lower abundance hopanes is more clearly shown by the MRM data (Fig. 12), which was used to derive many of the hopane ratios (Table 3). The m/z 205 mass chromatogram (Fig. 11c) shows that large amounts of 2a-methylhopanes relative to hopanes are present in the Matinenda Formation FI oil, whilst they are absent from the system blank and nal outside-rinse blank. No 3b-methylhopanes could be detected. C28C31 25norhopanes are present in the m/z 177 mass chromatogram (not shown) of the Matinenda Formation FI oil. C28 25,30bisnorhopane is approximately 50% of the abundance of

Response

Preservation of biomarkers in oil inclusions for >2 billion years Table 3 Biomarker ratios for the Matinenda FI oil (o-line data) Drimane/homodrimane Rearranged C15 bicyclic sesquiterpanes/ drimane + homodrimane C23 tricyclic terpane/C30 ab hopane C24 tetracyclic terpane/C30 ab hopane C24 tetracyclic terpane/(C24 tetracyclic terpane + C23 tricyclic terpane) C19/(C19 + C23) tricyclic terpanes (C19 + C20)/C23 tricyclic terpanes C26/C25 tricyclic terpanes C23/C21 tricyclic terpanes C24/C23 tricyclic terpanes C22/C21 tricyclic terpanes C29 tricyclic terpanes/C30 ab hopane Extended tricyclic terpane ratio (ETR: [C28 + C29tricyclic terpanes]/Ts) Ts/(Ts + Tm) Tm/C27b C29Ts/(C29Ts + C29 ab hopane) C =C30 ab hopane 30 C29 25-norhopane/C29 ab hopane C29 ab/(ab + ba) hopanes C30 ab/(ab + ba) hopanes C31 ab 22S/(22S + 22R) hopanes C31:32:33:34:35 ab homohopanes C29 ab hopane/C30 ab hopane Ts + Tm/C30 ab hopane C31 ab 22R hopane/C30 ab hopane 29,30-BNH/C30 ab hopane 28,30-BNH/C30 ab hopane C30 30-norhopane/C30 ab hopane Gammacerane/C30 ab hopane C31 2a Me/(C31 2a Me + C30 ab hopane) C32 2a Me/(C32 2a Me + C31 ab hopanes) C29 steranes/C29 ab hopane C21 sterane/C29 aaa 20R sterane C27:28:29 abb steranes C27:28:29 ba diasteranes C30/(C27 + C28 + C29 + C30) aaa 20 R steranes (%) (C28/C29) aaa 20 R steranes C29 ba diasteranes/(aaa + abb steranes) C28 aaa steranes 20S/(20S + 20R) C29 aaa steranes 20S/(20S + 20R) C28 steranes abb/(abb + aaa) C29 steranes abb/(abb + aaa) C27 ba diasteranes 20S/(20S + 20R) C29 ba diasteranes 20S/(20S + 20R) Norcholestane ratio (24-nor/(24-nor + 27-nor) Nordiacholestane ratio (24-nor/(24-nor + 27-nor) C26 steranes: 21-nor/(21-nor + 27-nor aaa 20R) 0.52 0.67 0.63 $ 0.36 $ 0.36 $ 0.08 $ 0.34 $ 0.82 $ 2.7 $ 0.44 $ 0.49 $ 0.23 $ 0.68 $ 0.59 13.7 0.20 0.07 0.12 0.96 0.95 0.58 41:29:13:11:6 $ 1.3 $ 1.05 $ 0.26 0.16 0.08 0.11 0.06 0.60 0.40 0.48* 2.1 41:26:33 48:29:23 2.0 0.68 0.67 0.48 0.44 0.54 0.53 0.63 0.65 <0.35 0.27 0.66

857

(Fig. 12). The co-occurrence of the C28C31 25-norhopanes with a UCM hump and n-alkanes is evidence that part but not all of the FI oil experienced heavy biodegradation prior to trapping (e.g., Volkman et al., 1983). 4.2.4. Steranes and diasteranes Fig. 13 shows the original (a) and duplicate (b) analyses of steranes and diasteranes in the Matinenda Formation FI oil, compared to the respective system and nal outsiderinse blanks. There are signicant contributions from C27 20S + 20R aaa steranes and a peak eluting just before C27 20S aaa sterane in both the system blank and the nal outside-rinse blank of both analyses, so these particular molecules are regarded as contaminants in the FI oil. The MRM chromatograms (Fig. 14) support this conclusion, as they show anomalously high abundances of C27 aaa steranes relative to C27 abb steranes and ba diasteranes in the m/z 372 217 MRM chromatogram, whereas the C28C30 aaa steranes have similar or lower relative abundances as the C28C30 abb steranes and ba diasteranes (m/z 386, 400 and 414 217). C27 20S + 20R aaa steranes have been repeatedly recorded in system blanks in the CSIRO labs to the level of the lowest abundances of uid inclusion oils (George et al., 2004a), and hence the occurrence of these compounds is not interpreted as indigenous in lean FI oils such as those in the Matinenda Formation. Traces of pregnane and homopregnane are also present in both the system blank and the nal outside-rinse blank of both analyses, and C29 ba 20S diasterane and the C27 abb steranes (and possibly a co-eluting C29 diasterane; Fig. 13) are present in the nal outside-rinse blank of the primary analysis. However, these extraneous components are in suciently low relative abundance not to have inuenced the data quality of the FI oil. The C28 and C29 steranes and the C28 diasteranes are unambiguously indigenous to the oil inclusions based on both the primary and duplicate analyses (Fig. 13). There is an unusual UCM hump around the retention time of the C28 ba diasteranes in the m/z 217 mass chromatograms that is (1) reproducible (both analyses), (2) absent from any of the blanks and (3) corresponds to the hump noted near C28 tricyclic terpanes in the m/z 191 mass chromatograms. These unresolved peaks are not present in the MRM chromatograms, and their signicance is unknown. It is noteworthy that a hump due to unidentied compounds at a similar retention time was also present in a bitumen extract and a uid inclusion oil from the Palaeoproterozoic sandstone associated with the Oklo nuclear reactors (Dutkiewicz et al., 2007). C30 24-n-propylcholestanes and 24-n-propyldiacholestanes were detected in the Matinenda Formation FI oil using the m/z 400 217 MRM chromatogram (Fig. 14), albeit in low proportions relative to C27C29 steranes (Table 3), but no isopropylcholestanes could be detected. Traces of C30 methylsteranes are present in the FI oil, and include the 4a-methyl-24-ethylcholestanes and 3b-methyl-24-ethylcholestanes (Fig. 14); no 2a-methylsteranes or dinosteranes were detected. A doublet peak labelled Ta and Tb in the m/z 217 mass chromatograms of both analyses (Fig. 13) is also present in the m/z 218 and 259 mass chromatograms (not shown), and is not present in the blanks. Although

Ratios were calculated from MRM data, (m/z) M+ 191, 217 for hopanes, and steranes and diasteranes, respectively), except for (m/z 123), $ (m/z 191), (m/z 205), (m/z 217), (m/z 218), * (m/z 217 and m/z 191).

C29 ab hopane in the m/z 177 mass chromatogram, and coelutes with the second C30 tricyclic terpane isomers in the m/z 191 mass chromatogram (Fig. 11a and b). C29 25norhopane is 12% of the abundance of C29 ab hopane, as measured in the m/z 398 191 MRM chromatogram

858

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

(a)
23/3

C29 C29Ts C30 25,30-BNH C3030-nor C30

C312(Me) C 30 C322(Me)

S C 31 R

(c)

FI oil

21/3

24/3 22/3

Response

20/3 19/3

29,3030/3 Ts 24/4 Tm BNH 28/3 26/3 Squalane 2529/3 25/3 nor C29

Blank Rinse

C31 S C R 32 SR G

C33 S R

C34 S R

C35 S R

FI oil Blank Rinse

Retention time
(b)
20/3 + ? 21/3 19/3 24/3 22/3 30/3 24/4 26/3 25/3 Squalane 23/3

C29

? C30 C3030-nor C30 C31 S C R S 32 C33 R S R G ?

C29Ts 25,30-BNH 29,30BNH Ts Tm 25nor 29/3 ?

Response

FI oil

Blank

Rinse

Retention time

Fig. 11. Partial m/z 191 mass chromatograms showing tricyclic and tetracyclic terpanes and pentacyclic triterpanes (hopanes) in (a) Matinenda Formation FI oil, system blank and nal outside-rinse blank and (b) duplicate analysis of Matinenda Formation FI oil, system blank and nal outside-rinse blank. (c) Partial m/z 205 mass chromatograms showing methylhopanes in Matinenda Formation FI oil, system blank and nal outside-rinse blank. Peak assignments dene stereochemistry at C-22 (S and R); ab and ba denote 17a(H)-hopanes and 17b(H)-moretanes, respectively. 19/3 to 30/3 = C19C30 tricyclic terpanes, 24/4 = C24 tetracyclic terpane, Ts = C27 18a(H), 22, 29, 30trisnorneohopane, Tm = C27 17a(H), 22, 29, 30-trisnorhopane, C29Ts = 18a-(H)-30-norneohopane, BNH = bisnorhopane, 25-nor = C29 25norhopane, C30 30-nor = C30 30-norhopane, 2a(Me) = 2a-methylhopanes (ab).

not conrmed by MRM and not quantied, these peaks are tentatively identied as tetracyclic polyprenoids. C26 steranes and diasteranes detected in the FI oil by the m/z 358 217 MRM chromatogram include 21-norcholestanes, possibly 24-norcholestanes, 24-nordiacholestanes, 27-norcholestanes and 27-nordiacholestanes (Fig. 14). 5. DISCUSSION AND SYNTHESIS 5.1. Thermal maturity of the trapped oil The C6 to C9 hydrocarbons in the Matinenda Formation FI oil as determined from the on-line crushing data, excluding the anomalously abundant soluble aromatic hydrocarbons discussed above, are strongly dominated by

n-alkanes (Fig. 3a), consistent with the trapped oil having a high maturity. The heptane and isoheptane values and the parameter n-C7/MCH (Table 1) are indicative of a supermature oil (i.e., one from the peak to late oil window), based on published crude oil calibrations (Thompson, 1983, 1987). However, other maturity-related C7 parameters, including the 2,4-/2,3-dimethylpentane ratio and the derived Ctemp in C (Bement et al., 1995; Mango, 1997), indicate a mid-oil window maturity for the Matinenda Formation FI oil (Table 1). Most maturity-dependent hopane ratios (e.g., ab/ (ab + ba); 22S/(22S + 22R)) are at equilibrium, indicating a mid-oil window maturity or greater. The moderate Ts/ (Ts+Tm) ratio is evidence against any extensive thermal cracking of the included oil. The signicant amount of tri-

Preservation of biomarkers in oil inclusions for >2 billion years


Ts Tm C27 C29 C29Ts

859

m/z 370 191

28,30-BNH 29,30-BNH 25-nor

m/z 384 191

Response

C29 C29Ts C29 25-nor C29 C30

m/z 398 191

m/z 412 191

C3030-nor C30 C30 ?

m/z 426 191

S C 31 R C31

Retention time

Fig. 12. Partial MRM chromatograms of Matinenda Formation FI oil showing the distribution of C27 (m/z 370 191), C28 (m/z 384 191), C29 (m/z 398 191), C30 (m/z 412 191) and C31 (m/z 426 191) hopanes. Peak assignments as Fig. 11, C27b = C27 17b(H), 22, 29, 30-trisnorhopane, C and C diahopanes. 29 30

cyclic terpanes relative to the hopanes in the Matinenda Formation FI oil (Fig. 11 and Table 3) is consistent with peak-oil window maturity. The steranes indigenous to the oil inclusions are dominated by the C21 molecule pregnane and lesser amounts of C22 homopregnane (Fig. 13), also consistent with a peak oil window maturity. The 20S/ (20S + 20R) ba diasterane ratios of the Matinenda Formation FI oil have reached equilibrium ($0.6), but the 20S/ (20S + 20R) aaa sterane ratios have not, with values of 0.48 (C28) and 0.44 (C29). The latter value indicates a mid-oil window thermal maturity of about 0.75% vitrinite reectance equivalent [VRE], based on the calibration by Sofer et al. (1993). The abb/(abb + aaa) sterane ratios are also some way from their thermal endpoints (Table 3), also consistent with a mid oil window thermal maturity for the FI oil. For the C26 steranes, the 21-norcholestane peak contains co-eluting isomers which are cumulatively more abundant than the individual 24-norcholestanes and 27-norcholestane isomers, consistent with a normally mature oil, but inconsistent with extensive thermal cracking (Moldowan et al., 1991b). The aromatic hydrocarbons in the Matinenda Formation FI oil have a simple equilibrium-controlled distribution, with no specic biogenic inputs of any particular isomer. Maturity-dependent alkylnaphthalene ratios are mostly consistent with a peak oil window maturity for the Matinenda Formation FI oil (Table 2). For example, similar DNR-1 and TNR-2 values are reached in Mesoprotero-

zoic sediments at the peak of the oil window in a calibration study on the McArthur Basin (George and Ahmed, 2002), and a Java Sea calibration of TNR-2 indicates a VRE of $0.94% (Radke et al., 1994). The MNR-1 is suggestive of a late oil window maturity, based on the McArthur Basin calibration (George and Ahmed, 2002) and a coal calibration (Radke et al., 1984). A dierent trimethylnaphthalene ratio (TMNr) and the tetramethylnaphthalene ratios (TeMNr and TeMNR-1) indicate early to peak oil window maturities (van Aarssen et al., 1999; George and Ahmed, 2002). The methylbiphenyl ratio and the two dimethylbiphenyl ratios (Table 2) are consistent with an early to peak oil window maturity, based on published calibrations (Cumbers et al., 1987; George and Ahmed, 2002). Alkylphenanthrene distributions are commonly used as maturity indicators (e.g., Radke et al., 1982b). The methylphenanthrene index (MPI-1) of the Matinenda Formation FI oil is low (0.53) due to the dominance of phenanthrene over methylphenanthrenes, but other alkylphenanthrene ratios (MPR, MPDF, DMPR; Table 2) are high. For example, the MPR is suggestive of a VRE of $1.2%, based on a coal calibration (Radke et al., 1984), and a peak to late oil window is also suggested by the calibration of MPR in the Mesoproterozoic McArthur Basin (George and Ahmed, 2002). The MPDF and the DMPR values in the Matinenda Formation FI oil are only reached in indigenous organic matter in more mature Mesoproterozoic sedimentary rocks, also indicating a peak to late oil window maturity (George and Ahmed, 2002). Accordingly, the low MPI-1 value is interpreted to be due to demethylation of methylphenanthrenes to phenanthrene, and thus the MPI-1 is on the reversed high maturity section of the MPI-1 versus thermal maturity graph (Radke et al., 1982a; Boreham et al., 1988). Both the methyldibenzothiophene and the dimethyldibenzothiophene ratios are low, consistent with maturities in the early to peak oil window (Radke, 1988; George and Ahmed, 2002). In general, the higher molecular weight biomarker maturity parameters indicate lower maturities (mid oil window) than the lower molecular weight parameters such as those based on alkylnaphthalenes (peak oil window), although there is not a simple relationship. In part this might be due to poor calibrations of many of these ratios into extremely old sequences and towards high maturities. Another contributory factor is that the FI oil represents a co-analysis of oil from two FI populations in the Matinenda Formation, which are likely to have dierent maturities and chemical compositions. There is no reliable method with current technology of separating the detailed molecular geochemical signal from each population, although the ability to obtain some geochemical data from individual FIs using laser micropyrolysis (Greenwood et al., 1998) or laser decrepitation (Hode et al., 2006) holds promise for the future. It is likely that the oil included in the second highly carbonic FI population has a higher maturity, and would likely have contributed a greater proportion of low molecular weight compounds, but this remains unproven. Despite the uncertainties of the ratio calibrations and the possibility of a mixed maturity signal, what is clear is that the thermal maturity of the included oil is not unusually

860

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870


C27 20S + C28 20R C27 20S C27 20R + C29 20S C27 20S C27 20R

Pregnane (C21)

(a)

C28 20R C28 20S Homopregnane (C22) Squalane C27 20R

Response

C29 20R C 20R 28 C28 20S C29 20S C29 20R C28 20R C29 20S C28 C29 20R 20S Ta+Tb

Squalane Pregnane (C21)

C27 20S ? ?

C27 20R

FI oil Blank

Squalane

C27 20S C27 20R + C29 20S C27 20R

Pregnane (C21) Homopregnane (C22)

C27 20S

Rinse
C27 20S + C28 20R C27 20R C29 20R C28 20S C28 20R C28 20S ? C29 20S C29 20R C29 20S Ta+Tb C29 20R

Retention time
Pregnane (C21) C27 20R + C29 20S C28 20S Homopregnane (C22) C27 20S Squalane C27 20R C27 20S C28 20R

C28 20R

(b)

FI oil

Response

C27 20S Pregnane (C21) Squalane Homopregnane (C22) C27 20R

Blank

Rinse

Retention time
Fig. 13. Partial m/z 217 mass chromatograms showing pregnane (C21), homopregnane (C22) and C27C29 steranes and diasteranes in (a) Matinenda Formation FI oil, system blank and nal outside-rinse blank and (b) duplicate analysis of Matinenda Formation FI oil, system blank and nal outside-rinse blank. Peak assignments dene stereochemistry at C-20 (S and R); ba, aaa and abb denote 13bH),17a(H)diasteranes, 5a(H),14a(H),17a(H)-steranes and 5a(H),14b(H),17b(H)-steranes, respectively. Ta and Tb are C30 18a(H) tetracyclic polyprenoids (21R and 21S, respectively).

high. This is irrefutable evidence against any signicant thermal cracking either prior to or subsequent to trapping. 5.2. Hydrocarbon survival at high temperatures and pressures A wide range of hydrocarbons is present in the uid inclusions in the Matinenda Formation sample. The compounds include (1) the simplest single carbon molecules (CH4 and CO2), (2) relatively abundant fundamental organic compounds of simple structure, including n-alkanes, monomethylalkanes, aromatic hydrocarbons, low molecu-

lar weight cyclic hydrocarbons, and (3) trace amounts of complex multi-ring biomarkers. All the commonly abundant hydrocarbons in Phanerozoic crude oils were found. This is not a surprise because very old FI oils are known to have uorescence and other physical properties like those of Phanerozoic equivalents (Dutkiewicz et al., 1998). Moreover, although no macroscopic life existed until the late Palaeoproterozoic (Han and Runnegar, 1992; Hedges et al., 2004), in other respects the dominant microbial and microscopic life forms for a billion years before then were quite similar to current ones (Buick, 2001), as were many of the

Preservation of biomarkers in oil inclusions for >2 billion years


Squalane
3 4 5-7 8 9 1 = 24dn 20S 2 = 24dn 20R 3 = 27dn 20S 4 = 27dn 20R 11 5 = 24n 20S 12 13 6 = 24n 20R

861

m/z m/z 358 217


7 = 24n 20S 8 = 24n 20R 9 = 21n + 10 = 27n 20S 11= 27n 20R 12 = 27n 20S 13 = 27n 20R

10

20S

20R 20S 20S 20R 20R 20S 20R 20S

m/z 372 217 C27

20S 20R 20S 20R

20R 20S 20R

m/z 386 217 C28

Squalane

m/z 400 217 C29


Squalane Squalane

20S

20R

20R

20S 20S 20R

20S 20R

20S

m/z 414 217 C30

20R 20R 20S 20S 20R * * * *

Squalane

m/z 414 231

3 20R

4 20R 4 20S + 4 20S 3 20R + 3 20S 3 20S 4 20R

Retention time
Fig. 14. Partial MRM chromatograms of Matinenda Formation FI oil showing the distribution of C26 (m/z 358 217), C27 (m/z 372 217), C28 (m/z 386 217), C29 (m/z 400 217) and C30 (m/z 414 217) steranes and diasteranes, and C30 methylsteranes (m/z 414 231). Peak assignments as Fig. 13. For the C26 steranes, 24n = 24-norcholestanes, 24dn = 24-nordiacholestanes, 27n = 27norcholestanes, 27dn = 27-nordiacholestanes, and 21-nor = 21norcholestanes. For the C30 steranes and diasteranes, the dominant series labelled are the 24-n-propylcholestanes. 24-Isopropylcholestanes are below detection limit, but if there were any 24isopropylcholestanes they would be in the locations marked by *, and would be less than 30% of the abundance of the 24-npropylcholestanes. For the C30 methylsteranes, 3b = 3b-methyl-24ethylcholestanes, and 4a = 4a-methyl-24-ethylcholestanes.

prevalent geological processes. Organic matter was buried, oil was generated and expelled from source rocks, and oil migrated to reservoirs where it accumulated and in some cases was preserved (Buick et al., 1998) in much the same way as in younger rocks. Thus, it is not unreasonable that Palaeoproterozoic oil should have a similar chemical signature, in many respects, to modern crude oils, which are dominantly formed from algal and bacterial organic matter. What is more surprising is that this range of complex hydrocarbons has survived temperatures possibly as high as 350 C at pressures of 50200 MPa to which the Matinenda Formation was exposed after diagenesis. Despite uncertainties with respect to calibrations against other maturity indicators or temperature, all the maturity parameters indicate that the FI oil is mature and was generated in the oil window. There is no evidence of extensive thermal

cracking from the molecular geochemistry data, despite the heating of the rocks to uppermost prehnitepumpellyite after entrapment of the aqueous oil-bearing inclusions and during migration and entrapment of the oil in the carbonic inclusions. Some of the early population of oil inclusions with high homogenisation temperatures (>200 C) have orange ourescence colours and may have been stretched during metamorphism, a process that may also have led to solid bitumen formation in a few of the oil inclusions (Dutkiewicz et al., 2003a). These processes did not have any discernible eect on the molecular geochemistry data reported here, which is the average composition of all included oil. The preservation of molecular maturity signatures through heating has also been recorded in K-feldspar hosted oil inclusions in the Ula oileld (Karlsen et al., 1993), which have been heated from 90 C at trapping to 143 C at present. The previously prevailing paradigm of oil destruction at temperatures as low as 120150 C (e.g., Evans et al., 1971; Ungerer et al., 1987; Barker, 1990; Braun and Burnham, 1992) has now been superseded by a recognition that crude oil in open reservoirs is much more thermally stable, based on observational data (e.g., Price, 1993; McNeil and Be ment, 1996; Vandenbroucke et al., 1999; Sajgo, 2000), half lives of model compounds (Mango, 1991; Domine et al., 1998) and pyrolysis models (e.g., Horseld et al., 1992; Pepper and Dodd, 1995; Planche, 1996; Schenk et al., 1997; Domine et al., 1998; Vandenbroucke et al., 1999; Wang et al., 2006). Oil is present in some very high temperature reservoirs, including the Elgin eld in the North Sea (190 C) (Vandenbroucke et al., 1999) and the Sweethome condensate in Texas well TX-O-119A (200 C) (McNeil and Bement, 1996). Traces of hydrocarbons are present at even higher temperatures, for example in Californian well Apex-1 (223 C) (Price et al., 1999), Hungarian well Hod 1 (233 C) (Sajgo et al., 1988) and prograde metamorphosed black shales (250550 C) (Schwab et al., 2005). These examples provide strong circumstantial evidence of oil stability to higher temperatures, at least on the order of 10s of million of years time scale. A consensus view from the above papers is that oil-to-gas cracking starts around 160215 C, depending on heating rate, oil type, presence of catalysts and whether it is an open or closed system. The upper temperature of crude oil stability is less well constrained, but may be of the order of 245275 C (Wang et al., 2006), although stabilities for C11+ hydrocarbons at up to 250350 C have been proposed (Price, 1993; Price and DeWitt, 2001). The generally recognised crude oil stability temperatures are thus still below the temperatures to which the oil inclusions in the Matinenda Formation have been exposed (280350 C). Accordingly, additional explanations for survival of included oil in high temperature rocks must be sought, particularly given the very long time that has elapsed since trapping of the oil. Three properties of FI oils that dierentiate them from crude oil in the pore space of a petroleum reservoir are pertinent to the question of thermal stability: (1) closed systems, (2) high uid pressures, and (3) lack of clay or other mineral or metal catalysts, as discussed previously (Karlsen et al., 1993; Karlsen and Skeie, 2006).

Response

862

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

(1) Fluid inclusions are generally closed systems, and exceptions of leakage or stretching can generally be easily recognized by petrographic work (Dutkiewicz et al., 2003a). Bulk oil-to-gas cracking and individual component reactions such as demethylation of alkylated aromatic hydrocarbons will proceed faster in an open system in which products are removed from the site of the reaction (Price and Wenger, 1992; Price, 1993). In closed systems, there is no opportunity for product removal away from the reactants, so the concentration of products builds up, thereby inhibiting the reaction according to Le Chateliers Principle. Furthermore, in closed systems where there is a lack of signicant interaction between hydrocarbons and rock or pore waters, hydrocarbons generate their own redox environment and will remain stable to much higher temperatures than in an open system (Giggenbach, 1997). Water in contact with the oil within the closed system inclusions may also help retard oil destruction reactions by suppressing the formation of free radicals, as has been demonstrated by laboratory pyrolysis experiments (Hesp and Rigby, 1973; Price and Wenger, 1992). (2) Dutkiewicz et al. (2003a) have shown that the carbonic-rich Matinenda Formation FIs were trapped at high uid pressures (50250 MPa), considerably higher than pressures in conventional oil reservoirs which are often around 1040 MPa (e.g., Losh et al., 2002; Bailey et al., 2006). It is known from many laboratory experiments that high uid pressures retard petroleum maturation and cracking reactions (Hesp and Rigby, 1973; Horseld et al., 1992; Price and Wenger, 1992; Price, 1993; Hill et al., 1996; Le wan, 1997; Sajgo, 2000), and this is also seen in naturallyoverpressured basins (e.g., Zou and Peng, 2001; Li et al., 2004). Therefore it is plausible that the high uid pressures in the Matinenda Formation FIs during migration further retarded oil cracking reactions. (3) Matinenda Formation oil inclusions are trapped in quartz and K-feldspar grains, and do not have any visible clays or other potential mineral catalysts associated with them. Clay minerals are known to catalyse oil generation reactions in natural environments (Espitalie et al., 1980; Tannenbaum et al., 1986) and in laboratory pyrolysis experiments (e.g., Bastow et al., 2000). Petroleum reservoirs commonly contain large amounts of detrital and diagenetic clays, and so the presence of these would be expected to enhance the cracking of crude oil in reservoirs. Transition metals in carbonaceous sedimentary rocks may also enhance oil-to-gas cracking reactions and lead to the decomposition of oil (Mango and Hightower, 1997; Mango and Elrod, 1998). Thus, the lack of clays or other potential mineral or metal catalysts in contact with the oil trapped in FIs should also contribute to the lack of thermal cracking reactions (Karlsen and Skeie, 2006) and thus to the survival of included oil for >2 billion years. The observation that none of the molecular maturity parameters for the oil trapped in the Matinenda Formation FIs have been signicantly reset at the high temperatures experienced subsequent to trapping also has relevance to the long debate about the mechanisms for changing maturity markers. Originally it was believed that many isomeric biomarker ratios changed due to conventional product-

reactant relationships (Mackenzie and McKenzie, 1983). Natural and laboratory studies have indicated that dierential rates of generation and the relative thermal stabilities of the isomers control maturity parameters in many instances (e.g., Peters et al., 1990; Abbott et al., 1990; Bishop and Abbott, 1993). Isomerisation reactions should continue with increasing maturation in closed systems such as uid inclusions, because there is no volume change so the reactions are unimpeded. The fact that all observed maturity parameters appear to be impeded suggest that they are all inuenced by the selective thermal stability of the isomers, as has been concluded previously (Karlsen et al., 1993; Karlsen and Skeie, 2006). 5.3. Biomarker geochemistry of the Matinenda Formation FI oil The Pr/Ph ratio of near unity (Table 2) is suggestive of a mildly reducing depositional environment. The low dibenzothiophene/phenanthrene ratio of 0.12 indicates that the trapped oil contains low amounts of organic sulphur, and in conjunction with the low Pr/Ph ratio suggests that it was generated from a marine shale (zone 3 of Hughes et al., 1995). The abundant monomethylalkanes are characteristic of Proterozoic oils and sedimentary rocks (Hoering, 1967; Summons et al., 1988b) and are likely due to cyanobacterial input (Summons and Walter, 1990; Kenig et al., 1995). Bicyclic sesquiterpanes are ubiquitous in sediments and oils of all ages, and are thought to be the degradation products of bacteriohopanes (Alexander et al., 1984). The tricyclic terpanes are ubiquitous in oils and are most likely derived from bacteria (Ourisson et al., 1982), although other algal sources are possible (Simoneit et al., 1993). Their isomeric distribution enables inferences to be made about the characteristics of the source rock that generated the oil, provided that inferences derived from the global Phanerozoic database are applicable to interpreting Palaeoproterozoic oils. The C19 and C20 tricyclic terpanes are in low abundance relative to other tricyclic terpanes (Table 3), consistent with a marine source rock containing no higher plant organic matter (Preston and Edwards, 2000; George et al., 2004b). The low C26/C25 tricyclic terpane ratio shows that the Matinenda Formation FI oil was not derived from a lacustrine source rock (Schiefelbein et al., 1999). Extended tricyclic terpanes from C28 to C30 were detected (Fig. 11), but are in low relative abundance (Table 3). Therefore, the extended tricyclic terpane ratio (Holba et al., 2001) does not indicate a source rock inuenced by marine upwelling (Holba et al., 2003b). Rearranged hopanes in the Matinenda Formation FI oil include Ts, C29Ts and the C29 and C30 diahopanes, but these are in low abundance and therefore suggest that the source rock was not clay-rich (Moldowan et al., 1991a), especially in view of the relatively high thermal maturity of the FI oil. The distribution of extended hopanes, with low amounts of the C35 isomers, and the low abundance of 28,30-bisnorhopane relative to other hopanes (Fig. 12) supports a mildly reducing source rock depositional environment. A trace of gammacerane was identied in the

Preservation of biomarkers in oil inclusions for >2 billion years

863

m/z 412 191 MRM chromatogram (Fig. 12), but in insucient quantities to infer water column stratication (cf. Sinninghe Damste et al., 1995). The C24/C23 and C22/C21 tricyclic terpane ratios of oils can be used to predict their source-rock depositional environment, based on a calibration set of 500 crude oils (Peters et al., 2005). These ratios (Table 3) suggest that the Matinenda Formation FI oil may have been derived from a carbonate or marl, and this might be supported by the moderately high abundance of C24 tetracyclic terpane. The possibility of a calcareous source rock is further indicated by the high C29/C30 ab hopane ratio of 1.3 and the abundant 29,30-bisnorhopane and C30 30-norhopane (Moldowan et al., 1992) but not by the rather low amounts of C31 hopanes relative to C30 ab hopane or by the proportion of C34 and C35 homohopanes (Table 3). The high 2amethylhopane ratios (Table 3) suggest that the source rock of the FI oil contained cyanobacteria (Summons et al., 1999). 2a-Methylhopanes are common biomarkers found in Precambrian oils and rocks (Summons et al., 1988b; Brocks et al., 2003b; Dutkiewicz et al., 2004), although they are also abundant in Phanerozoic carbonate-sourced oils (Moldowan et al., 1992; Summons and Jahnke, 1992). In this and another Palaeoproterozoic FI oil sample (Dutkiewicz et al., 2007) it has been argued that a high C29/C30 ab hopane ratio, abundant 29,30-bisnorhopane, C30 30norhopane and C24 tetracyclic terpane, and low amounts of diasteranes and rearranged hopanes indicate a carbonate or marly source rock with low clay content, following Phanerozoic biomarker interpretational guidelines (Peters et al., 2005). In view of the very minor carbonate in the most likely source succession, it is proposed that instead of the direct lithological interpretation, these biomarker distributions are reecting a strong cyanobacterial contribution to all marine organic matter in the Proterozoic, whereas in Phanerozoic rocks such signatures are largely restricted to petroleum generated from carbonates. Sterane biomarkers in the Matinenda Formation FI oil are diverse, though less abundant than hopanes (Table 3). The sterane C27:C28:C29 ratios favour the C27 isomers, based on abb steranes and ba diasteranes, and the C28/ C29 aaa sterane ratio is 0.68 (Table 3). The latter is inconsistent with this ratio decreasing with increasing geological age through the Phanerozoic, as suggested by Grantham and Wakeeld (1988), although other Precambrian samples have similarly high ratios (Brocks et al., 2003a; Dutkiewicz et al., 2007). The presence of the steranes and diasteranes is evidence for eukaryotic input to the source rock that generated the FI oil (cf. Summons et al., 2006). The 4a-methyl24-ethylcholestanes in the FI oil also suggest a eukaryotic input to either marine or lacustrine source rocks, and together with 3b and 2a-methyl-24-ethylcholestanes have been detected in older Archaean rocks (Brocks et al., 2003a; Brocks et al., 2003b). The presence of 24-n-propylcholestanes is indicative of marine algal input to the source rock of the FI oil (Moldowan et al., 1990). The tentatively identied tetracyclic polyprenoids in the Matinenda Formation FI oil are common in fresh to brackish sedimentary environments and may be related to green algae (Holba et al., 2000, 2003a). Many Neoproterozoic and early Cam-

brian oils and source rocks contain high amounts of 24-isopropylcholestanes (McCarey et al., 1994; Peters et al., 1995) which are biomarkers for marine demosponges (Love et al., 2005), but these are absent from the Matinenda Formation FI oil (Fig. 14). Individual ba diasterane isomers are of similar abundance to individual sterane isomers (Fig. 14), so the C29 ba diasteranes/(aaa + abb steranes) ratio is low (Table 3) for an oil interpreted to have been generated in the peak oil window using other parameters. This observation is consistent with the low content of rearranged hopanes in the Matinenda Formation FI oil. As acidic sites on clays catalyse the early diagenetic reactions that lead to diasterane formation (Sieskind et al., 1979), this implies that the source rock of the FI oil was not clayrich. Based on an empirically-observed increasing proportion of 24-norcholestanes and 24-nordiacholestanes with decreasing age, Holba et al. (1998) has proposed two parameters, the norcholestane and the nordiacholestane ratios, for constraining the age of oil source rocks. The norcholestane ratio of the Matinenda Formation FI oil is <0.35 (Table 3), although based on barely detectable 24norcholestane peaks (Fig. 14), typical of Triassic and older oils. The more reliable nordiacholestane ratio (0.27) is higher than most Triassic and older oils (<0.2) (Holba et al., 1998), although this is based on noisy and poorly resolved 24-nordiacholestane peaks (Fig. 14). These parameters increase due to an increasing proportion of dinoagellate and diatom-derived 24-norsterols (Rampen et al., 2007). The presence of 24-nordiacholestanes in the Matinenda FI oil suggests that there may be a low-level, presently unidentied non-diatom/dinoagellate source of these in the Proterozoic. Although the majority of biomarker attributes are consistent with a Palaeoproterozoic sources, some aspects of the Matinenda Formation FI oil could be argued to be due to overprinting with Cenozoic oil or recent contamination. These attributes are the high molecular weight odd nalkanes and the higher than expected C28/C29 aaa sterane and nordiacholestane ratios. The petrographic evidence rules out the trapping of FIs post metamorphism, and thus the possibility that oils generated from Palaeozoic source rocks in the Michigan Basin, for instance, could have migrated laterally into weathered or fractured Matinenda Formation. The methodology used with the blanks suggests that recent contamination has not aected the biomarkers, although the relatively high contribution of n-alkanes in the blanks may explain the unexpected distribution of n-alkanes in the FI oil. 5.4. Source of the inclusion oil: implications for life in the Palaeoproterozoic The biomarker geochemistry of the Matinenda Formation FI oil enables inferences about the organisms that contributed to the organic matter deposited in the Palaeoproterozoic source rocks from which the analysed oil was generated and expelled. Further, inferences about the depositional environment can also be made.

864

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870

According to Young (2001), all Huronian units below the Gowganda Formation were dominated by non-marine deposition in a largely rift-controlled environment, although Fralick and Miall (1989) have described the McKim Formation as deltaic in origin, deposited in an marginal marine setting. The inclusion oil is probably derived either from the immediately overlying organic-rich deltaic McKim Formation (McKirdy and Imbus, 1992) or from kerogen within the Matinenda Formation (Fig. 1) (Mossman et al., 1993), as there are no known source rocks deeper in the section. A deltaic source rock is supported by the presence of n-propylcholestanes, indicating a contribution from marine source rocks (Moldowan et al., 1990). The source rock was likely deposited in a mildly reducing environment, as indicated by the Pr/Ph ratio and the presence of 28,30-bisnorhopane. It is not possible to perform a detailed oil-source correlation between the FI oil and its potential source rocks, because the latter have been exposed to upper prehnitepumpellyite facies metamorphism in an open system and have thus likely been rendered inert with respect to extractable organics. Biomarkers for bacteria in the Matinenda Formation FI oil include bicyclic sesquiterpanes, tricyclic terpanes and hopanes. Bacterial biomarkers are commonly detected in Precambrian rocks and oils (e.g., Summons et al., 1988a; Summons et al., 1988b; Brocks et al., 1999; Brocks et al., 2003b; Dutkiewicz et al., 2003b; Dutkiewicz et al., 2004; Brocks et al., 2005). There is no evidence in the Matinenda Formation FI oil for biomarkers such as arylisoprenoids, okenane, chlorobactane or isorenieratane that might be related to specic bacterial inputs of photosynthetic purple or green sulphur bacteria (Summons and Powell, 1986; Brocks et al., 2005). The presence of abundant monomethylalkanes and especially the C31+ 2a-methylhopanes in the Matinenda Formation FI oil provide strong evidence for cyanobacterial input to the source rock of the oil (Summons and Walter, 1990; Summons and Jahnke, 1992; Kenig et al., 1995). These biomarkers have been detected previously in abundance in Proterozoic and older oils and rocks (e.g., Summons et al., 1988b; Brocks et al., 1999; Summons et al., 1999; Brocks et al., 2003b; Dutkiewicz et al., 2004), for which the time of deposition of the rocks, and the likely source rocks for the oils, postdates the time of detectable oxygenation of the shallow oceans (Holland, 2006). Though these molecules are not directly associated with metabolism, by association they imply the existence of oxygenic photosynthesis (Summons et al., 1999), because all cyanobacteria are capable of performing such autotrophy (Castenholz, 2001). Moreover, the production of abundant C31+ 2a-methylhopanes is widespread, while not ubiquitous, throughout the Phylum Cyanobacteria, including in the basally-branching Gloeobacter violaceus (Castenholz, 2001). Thus, it seems to be a reasonable inference that biogenic oxygen was being generated during deposition of the source rocks for the Matinenda FI oils (Dutkiewicz et al., 2006a). A signicant biomarker observation in the Matinenda Formation FI oil is the presence of diverse steranes and diasteranes, which are evidence for eukaryotic input to

the source rock that generated the FI oil (cf. Summons et al., 2006). Despite recent speculations to the contrary (Raymond and Blankenship, 2004; Kopp et al., 2005), the only conrmed biosynthetic pathway of precursor sterols involves the epoxidation of squalene where the addition of 1O2 is catalysed by the enzyme squalene monooxygenase 2 (Jahnke and Klein, 1983). The extended biosynthetic pathway to modied sterols such as cholesterol and ergosterol involving cyclization and subsequent oxidative demethylation requires up to twelve O2 molecules (Summons et al., 2006). Although some sterane precursors occur in other organisms (Volkman, 2005), high concentrations of the regular steranes are apparently exclusive biomarkers for organisms of the domain Eukarya (Brocks et al., 2003b; Summons et al., 2006). Steranes and diasteranes have been detected in older shale extracts, such as the 2.7 Ga rocks from the Pilbara Craton (Brocks et al., 1999, 2003a,b), but with the caveat of probably syngenetic, prompting concerns that the ancient steranes are contaminants (Kopp et al., 2005). The steranes and diasteranes reported here are from an oil of normal thermal maturity which was trapped in uid inclusions early in the burial history of the Matinenda Formation. Oil inclusions in quartz, unlike shale extracts (Brocks et al., 2003a), cannot be contaminated by later migrating hydrocarbons, and the petrographic information rules out the trapping of FIs post metamorphism. The entrapped biomarkers may be from both FI populations, but comparison with the outside rinse and system blanks shows that the bulk of the analysed biomarkers are clearly not contaminants from bitumen in the pore space, younger hydrocarbons or from anthropogenic input. Thus, the presence of steranes and diasteranes in the Matinenda Formation FI oil which was trapped before 2.2 Ga, as well as in the Oklo FI oil which was trapped between 2.0 and 2.1 Ga (Dutkiewicz et al., 2006b; Dutkiewicz et al., 2007), provide supporting, but not conrming, evidence in favour of the indigenous origin of the older shale-hosted occurrences. Clearly, however, the molecular fossil record of eukaryotes substantially pre-dates that of their body or trace fossils, an observation concordant with contemporary evolution models. 5.5. Proterozoic and Archean oil inclusions: biogeochemical time capsules Oil inclusions are found in a wide variety of rocks and minerals of all ages (Karlsen et al., 1993; Munz, 2001; and references therein). The ability to analyse the oil trapped in FIs from early Precambrian sequences opens up a new way to obtain biogeochemical information that can be reliably identied as being indigenous to the host rock. The great advantage of using FI oils is that they trap oils in an unaltered state, allowing survival of biologically and environmentally informative biomarkers for billions of years, even through signicant thermal events. One disadvantage is that this is a bulk analysis technique, potentially resulting in the co-analysis of oil inclusions and solid bitumens from dierent generations, a problem which can be overcome with careful petrography and sample selection. Laser micropyrolysis or laser decrepitation GC

Preservation of biomarkers in oil inclusions for >2 billion years

865

MS of single oil inclusions oers a potential solution to this drawback (Greenwood et al., 1998; Hode et al., 2006), although with current technology it is unlikely that the detailed biomarker composition presented in this paper will ever be able to be reproduced for single oil inclusions. Another diculty is relating the inclusion oil to a potential source rock, as the two will have necessarily undergone differing maturation pathways. However, by sampling inclusion-bearing rocks in stratigraphic successions with a limited range of potential source rocks, reasonable source identication can be achieved. Despite these two issues, this methodology can clearly provide otherwise unattainable insights into biological evolution and palaeoenvironments during critical stages of Earths early history. 6. CONCLUSIONS (1) FI oil trapped prior to low grade metamorphism at 2.2 Ga in the Palaeoproterozoic Matinenda Formation contains CH4, CO2, n-alkanes, isoprenoids, monomethylalkanes, aromatic hydrocarbons, low molecular weight cyclic hydrocarbons, and trace amounts of complex multi-ring biomarkers. (2) Maturity ratios show that the FI oil was generated within the oil window, with no evidence of extensive thermal cracking, despite the oil inclusions having been either trapped at uppermost prehnite-pumpellyite facies metamorphic grade (280350 C), or having been heated to these temperatures after entrapment. (3) Three properties of FI oils that dierentiate them from crude oil in the free pore space of a petroleum reservoir may contribute to the thermal stability of the included hydrocarbons: (a) closed systems, (b) high uid pressures, and (c) lack of clays or other potential mineral or metal catalysts. (4) Biomarker geochemistry indicates that the FI oil was derived from a marine source rock deposited in a reducing depositional environment. Candidates include the immediately overlying organic-rich deltaic McKim Formation or kerogen from within the Matinenda Formation. (5) Biomarkers specic for bacteria in the Matinenda Formation FI oil include bicyclic sesquiterpanes and hopanes. (6) The presence of abundant monomethylalkanes and especially the >C31 2a-methylhopanes in the Matinenda Formation FI oil are strong evidence for a signicant cyanobacterial input to the source rock of the oil. (7) Biomarker signatures that indicate a calcareous rock with a low clay content, based on Phanerozoic biomarker interpretational guidelines, may instead reect high cyanobacterial organic inputs that were prevalent in all Proterozoic environments, but were later largely restricted to non-clastic settings. (8) Diverse steranes and diasteranes are present in the FI oil, and are evidence for eukaryotic input to the source rock that generated the FI oil. (9) The presence of abundant biomarkers for cyanobacteria and eukaryotes derived from and trapped in rocks which are interpreted to have been some of the youngest rocks deposited while the atmosphere remained largely

anoxic (Papineau et al., 2007) are consistent with an earlier evolution of oxygenic photosynthesis and suggests that some aquatic settings had become suciently oxygenated for sterol biosynthesis by this time. (10) The extraction of biomarker molecules from Palaeoproterozoic oil-bearing uid inclusions thus establishes a new method, using low detection limits and system blank levels, to trace evolution through Earths early history that avoids the potential contamination problems that can aect shale-hosted hydrocarbons.
ACKNOWLEDGMENTS We would like to thank Robinson A. Quezada and Manzur Ahmed for analytical assistance. We thank Malcolm Walter for review of an earlier version of this paper. The paper beneted significantly from the detailed comments of journal reviewer Dag Karlsen, an anonymous reviewer and Associate Editor Bob Burruss. This work was supported by an ARC Discovery Grant which includes a QEII Fellowship to A.D. and by the National Aeronautics and Space Administration Astrobiology Institute (R.B.).

REFERENCES Abbott G. D., Wang G. Y., Eglinton T. I., Home A. K. and Petch G. S. (1990) The kinetics of sterane biological marker release and degradation processes during the hydrous pyrolysis of vitrinite kerogen. Geochim. Cosmochim. Acta 54, 24512461. Ahmed M. and George S. C. (2004) Changes in the molecular composition of crude oils during their preparation for GC and GCMS analyses. Org. Geochem. 35, 137155. Alexander R., Kagi R. I., Noble R. and Volkman J. K. (1984) Identication of some bicyclic alkanes in petroleum. Org. Geochem. 6, 6370. Allwood A. C., Walter M. R., Kamber B. S., Marshall C. P. and Burch I. W. (2006a) Stromatolite reef from the Early Archaean era of Australia. Nature 441, 714718. Allwood A. C., Walter M. R. and Marshall C. P. (2006b) Raman spectroscopy reveals thermal palaeoenvironments of c.3.5 billion-year-old organic matter. Vibrat. Spect. 41, 190197. Bailey W. R., Underschultz J., Dewhurst D. N., Kovack G., Mildren S. and Raven M. (2006) Multi-disciplinary approach to fault and top seal appraisal; Pyrenees-Macedon oil and gas elds, Exmouth Sub-basin, Australian Northwest Shelf. Mar. Pet. Geol. 23, 241259. Barker C. (1990) Calculated volume and pressure changes during thermal cracking of oil to gas in reservoirs. Am. Assoc. Petrol. Geol. Bull. 74, 12541261. Bastow T. P., Alexander R., Fisher S. J., Singh R. K., van Aarssen B. G. K. and Kagi R. I. (2000) Geosynthesis of organic compounds. Part Vmethylation of alkylnaphthalenes. Org. Geochem. 31, 523534. Bement W. O., Levey R. A. and Mango F. D. (1995) The temperature of oil generation as dened with C7 chemistry maturity parameter (2,4-DMP/2,3-DMP ratio). In Organic Geochemistry: Developments and Applications to Energy, Climate, Environment and Human History (eds. J. O. Grimalt and C. Dorronsoro). A.I.G.O.A., pp. 505507. Bishop A. N. and Abbott G. D. (1993) The interrelationship of biological marker maturity parameters and molecular yields during contact metamorphism. Geochim. Cosmochim. Acta 57, 36613668. Boreham C. J., Crick I. H. and Powell T. G. (1988) Alternative calibration of the Methylphenanthrene Index against vitrinite

866

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870 Domine F., Dessort D. and Brevart O. (1998) Towards a new method of geochemical kinetic modeling: implications for the stability of crude oils. Org. Geochem. 28, 597612. Dutkiewicz A., Rasmussen B. and Buick R. (1998) Oil preserved in uid inclusions in Archaean sandstones. Nature 395, 885888. Dutkiewicz A., Ridley J. and Buick R. (2003a) Oil-bearing CO2 CH4H2O uid inclusions: oil survival since the Palaeoproterozoic after high temperature entrapment. Chem. Geol. 194, 51 79. Dutkiewicz A., Volk H., Ridley J. and George S. (2003b) Biomarkers, brines, and oil in the Mesoproterozoic, Roper Superbasin, Australia. Geology 31, 981984. Dutkiewicz A., Volk H., Ridley J. and George S. C. (2004) Geochemistry of oil in uid inclusions in a Middle Proterozoic igneous intrusion: Implications for the source of hydrocarbons in crystalline rocks. Org. Geochem. 35, 937957. Dutkiewicz A., Volk H., George S. C., Ridley J. and Buick R. (2006a) Biomarkers from Huronian oil-bearing uid inclusions: an uncontaminated record of life before the Great Oxidation Event. Geology 34, 437440. Dutkiewicz A., Volk H., George S. C., Ridley J., Mossman D. and Buick R. (2006b) Oil and its biomarkers trapped inside uid inclusions ca. 2.452.0 Ga. Geochim. Cosmochim. Acta 70, A153. Dutkiewicz A., George S. C., Mossman D., Ridley J. and Volk H. (2007) Oil and its biomarkers associated with the Palaeoproterozoic Oklo natural ssion reactors, Gabon. Chem. Geol. 244, 130154. Easton R. M. (2000) Metamorphism of the Canadian Shield, Ontario, Canada: II. Proterozoic metamorphic history. Can. Mineral. 38, 319344. Eglinton G. and Hamilton R. J. (1967) Leaf epicuticular waxes. Science 156, 13221335. Eigenbrode J. (in press) Fossil Lipids for Life-Detection: A case study from the early earth record. Space Sci. Rev., doi:10.1007/ s11214-007-9252-9. Espitalie J., Madec M. and Tissot B. (1980) Role of mineral matrix in kerogen pyrolysis: inuence on petroleum generation and migration. AAPG Bull. 64, 5966. Evans C. R., Rogers M. A. and Bailey N. J. L. (1971) Evolution and alteration of petroleum in western Canada. Chem. Geol. 8, 147170. Fralick P. W. and Miall A. D. (1989) Sedimentology and the Lower Huronian Supergroup (Early Proterozoic), Elliot Lake area, Ontario, Canada. Sed. Geol. 63, 127153. Furnes H., Banerjee N. R., Muehlenbachs K., Staudigel H. and de Wit M. (2004) Early life recorded in Archean pillow lavas. Science 304, 578581. George S. C. and Jardine D. R. (1994) Ketones in a Proterozoic dolerite sill. Org. Geochem. 21, 829839. George S. C., Lisk M., Eadington P. J., Quezada R. A., Krieger F. W., Greenwood P. F. and Wilson M. A. (1996) Comparison of palaeo oil charges with currently reservoired hydrocarbons using the geochemistry of oil-bearing uid inclusions. Society of Petroleum Engineers, Asia Pacic Oil and Gas Conference, SPE paper 36980, pp. 159171. George S. C., Krieger F. W., Eadington P. J., Quezada R. A., Greenwood P. F., Eisenberg L. I., Hamilton P. J. and Wilson M. A. (1997) Geochemical comparison of oil-bearing uid inclusions and produced oil from the Toro Sandstone, Papua New Guinea. Org. Geochem. 26, 155173. George S. C., Lisk M., Eadington P. J. and Quezada R. A. (1998) Geochemistry of a palaeo-oil column: Octavius 2, Vulcan Subbasin. In Proceedings of Petroleum Exploration Society of Australia Symposium, The Sedimentary Basins of Western Australia 2 (eds. P. G. Purcell and R. R. Purcell), pp. 195210.

reectance: application to maturity measurements on oils and sediments. Org. Geochem. 12, 289294. Brasier M., McLoughlin N., Green O. and Wacey D. (2006) A fresh look at the fossil evidence for early Archaean cellular life. Philos. Trans. R. Soc. Lond., B 361, 887902. Braun R. L. and Burnham A. K. (1992) PMOD A exible model of oil and gas generation, cracking, and expulsion. Org. Geochem. 19, 161172. Brocks J. J., Logan G. A., Buick R. and Summons R. E. (1999) Archean molecular fossils and the early rise of eukaryotes. Science 285, 10331036. Brocks J. J., Buick R., Logan G. A. and Summons R. E. (2003a) Composition and syngeneity of molecular fossils from the 2.78 to 2.45 billion-year-old Mount Bruce Supergroup, Pilbara Craton, Western Australia. Geochim. Cosmochim. Acta 67, 42894319. Brocks J. J., Buick R., Summons R. E. and Logan G. A. (2003b) A reconstruction of Archean biological diversity based on molecular fossils from the 2.78 to 2.45 billion-year-old Mount Bruce Supergroup, Hamersley Basin, Western Australia. Geochim. Cosmochim. Acta 67, 43214335. Brocks J. J., Love G. D., Snape C. E., Logan G. A., Summons R. E. and Buick R. (2003c) Release of bound aromatic hydrocarbons from late Archean and Mesoproterozoic kerogens via hydropyrolysis. Geochim. Cosmochim. Acta 67, 15211530. Brocks J. J., Summons R. E., Buick R. and Logan G. A. (2003d) Origin and signicance of aromatic hydrocarbons in giant iron ore deposits of the late Archean Hamersley Basin, Western Australia. Org. Geochem. 34, 11611175. Brocks J. J. (2005) The evolution of steroids and eukaryotes in the Proterozoic. EOS Transactions, AGU Fall Meeting Supplement, vol. 86 (52), Abstract B12A-05. Brocks J. J., Love G. D., Summons R. E., Knoll A. H., Logan G. A. and Bowden S. A. (2005) Biomarker evidence for green and purple sulphur bacteria in a stratied Palaeoproterozoic sea. Nature 437, 866870. Brocks J. J. and Summons R. E. (2005) Sedimentary hydrocarbons, biomarkers for early life. In Treatise on Geochemistry, Vol. 8: Biogeochemistry, pp. 63115. Buick R., Dunlop J. S. R. and Groves D. I. (1981) Stromatolite recognition in ancient rocks: an appraisal of irregularly laminated structures in an early Archean chert-barite unit from North Pole, Western Australia. Alcheringa 5, 161179. Buick R., Rasmussen B. and Krapez B. (1998) Archean oil: evidence for extensive hydrocarbon generation and migration 2.53.5 Ga. AAPG Bull. 82, 5069. Buick R. (2001) Life in the Archaean. In Palaeobiology II (eds. D. E. G. Briggs and P. R. Crowther). Blackwell Science, pp. 1321. Burruss R. C. (1987) Crushing-cell, capillary gas chromatography of petroleum uid inclusions: method and application to petroleum source rocks, reservoirs, and low temperature hydrothermal ores (abstract). In Fluid Inclusion Research, vol. 20 (eds. E. Roedder and A. Kozlowski ), p. 59. Card K. D. (1978) Metamorphism of the middle Precambrian (Aphebian) rocks of the eastern Southern Province. In Metamorphism in the Canadian Shield, vol. 78-10 (eds. J. A. Fraser and W. W. Heywood). Geological Survey of Canada Paper, pp. 269282. Castenholz R. W. (2001) Phylum BX. Cyanobacteria, oxygenic photosynthetic bacteria. In Bergeys Manual of Systematic Bacteriology, vol. 1, 2nd Ed (eds. D. R. Boon and R. W. Castenholz). Springer, pp. 473597. Cumbers K. M., Alexander R. and Kagi R. I. (1987) Methylbiphenyl, ethylbiphenyl and dimethylbiphenyl isomer distributions in some sediments and crude oils. Geochim. Cosmochim. Acta 51, 31053111.

Preservation of biomarkers in oil inclusions for >2 billion years George S. C., Volk H., Ruble T., Lisk M., Manzur A., Liu K., Quezada R., Dutkiewicz A., Brincat M., Middleton H., Smart S. and Horseld B. (2001) Extracting oil from uid inclusions for geochemical analyses: size matters! Abstracts of the 20th International Meeting on Organic Geochemistry, P/TUE1/37, pp. 467468. George S. C. and Ahmed M. (2002) Use of aromatic compound distributions to evaluate organic maturity of the Proterozoic middle Velkerri Formation, McArthur Basin, Australia. West Australian Basins Symposium, pp. 252270. George S. C., Ahmed M., Liu K. Y. and Volk H. (2004a) The analysis of oil trapped during secondary migration. Org. Geochem. 35, 14891511. George S. C., Lisk M. and Eadington P. J. (2004b) Fluid inclusion evidence for an early, marine-sourced oil charge prior to gascondensate migration, Bayu-1, Timor Sea, Australia. Mar. Pet. Geol. 21, 11071128. George S. C., Ruble T. E., Volk H., Lisk M., Brincat M. P., Dutkiewicz A. and Ahmed M. (2004c) Comparing the geochemical composition of uid inclusion and crude oils from wells on the Laminaria High, Timor Sea. In Timor Sea Petroleum Geoscience, Proceedings Of The Timor Sea Symposium (eds. G. K. Ellis, P. W. Baillie and T. J. Munson). Northern Territory Geological Survey, pp. 203230. George S. C., Volk H., Dutkiewicz A., Ridley J. and Mossman D. (2006) Oil-bearing uid inclusions: Biogeochemical time-capsules for >2.0 billion years. Geochim. Cosmochim. Acta 70, A198. George S. C., Volk H. and Ahmed M. (2007) Geochemical analysis techniques and geological applications of oil-bearing uid inclusions, with some Australian case studies. J. Petrol. Sci. Eng. 57, 119138. Giggenbach W. F. (1997) Relative importance of thermodynamic and kinetic processes in governing the chemical and isotopic composition of carbon gases in high-heatow sedimentary basins. Geochim. Cosmochim. Acta 61, 37633785. Goldblatt C., Lenton T. M. and Watson A. J. (2006) Bistability of atmospheric oxygen and the Great Oxidation. Nature 443, 683 686. Grantham P. J. and Wakeeld L. L. (1988) Variations in the sterane carbon number distributions of marine source rock derived crude oils through geological time. Org. Geochem. 12, 6173. Greenwood P. F., George S. C. and Hall K. (1998) Applications of laser micropyrolysis gas chromatography mass spectrometry. Org. Geochem. 29, 10751089. Han T.-M. and Runnegar B. (1992) Megascopic eukaryotic algae from the 2.1-billion-year-old Negaunee Iron-Formation, Michigan. Science 257, 232235. Hayes J. and Waldbauer J. (2006) The carbon cycle and associated redox processes through time. Philos. Trans. R. Soc. Lond., B 361, 931950. Hayes J. M., Kaplan I. R. and Wedeking K. W. (1983) Precambrian organic geochemistry, preservation of the record. In Earths Earliest Biosphere, Its Origin and Evolution (ed. J. W. Schopf). Princeton University Press, pp. 93134. Hedges S. B., Blair J. E., Venturi M. L. and Shoe J. L. (2004) A molecular timescale of eukaryote evolution and the rise of complex multicellular life. BMC Evol. Biol. 4:2. doi:10.1186/ 1471-2148-4-2. Hesp W. and Rigby D. (1973) The geochemical alteration of hydrocarbons in the presence of water. Erdol und Kohle-Erdgas 26, 7076. Hilburn I. A., Kirschvink J. L., Tajika E., Tada R., Hamano Y. and Yamamoto S. (2005) A negative fold test on the Lorrain Formation of the Huronian Supergroup: Uncertainty on the

867

paleolatitude of the Paleoproterozoic Gowganda glaciation and implications for the great oxygenation event. Earth Planet. Sci. Lett. 232, 315332. Hill R. J., Tang Y. C., Kaplan I. R. and Jenden P. D. (1996) The inuence of pressure on the thermal cracking of oil. Energy Fuels 10, 873882. Hode T., Zebuhr Y. and Broman C. (2006) Towards biomarker analysis of hydrocarbons trapped in individual uid inclusions: First extraction by ErYAG laser. Planet. Space Sci. 54, 1575 1583. Hoering T. C. (1967) Molecular fossils from the Precambrian Nonesuch Shale. Carnegie Inst. Wash., Yearbook 75, 806813. Hofmann H. J., Grey K., Hickman A. H. and Thorpe R. I. (1999) Origin of 3.45 Ga coniform stromatolites in Warrawoona Group, Western Australia. Geol. Soc. Am. Bull. 111, 12561262. Holba A. G., Dzou L. I. P., Masterson W. E., Hughes W. B., Huizinga B. J., Singletary M. S., Moldowan J. M., Mello M. R. and Tegelaar E. (1998) Application of 24-norcholestanes for constraining source age of petroleum. Org. Geochem. 29, 1269 1283. Holba A. G., Tegelaar E., Ellis L., Singletary M. S. and Albrecht P. (2000) Tetracyclic polyprenoids: Indicators of freshwater (lacustrine) algal input. Geology 28, 251254. Holba A. G., Ellis L., Dzou L. I. P., Hallam A., Masterson W. E., Francu J. and Fincannon A. L. (2001) Extended tricyclic terpanes as age discriminators between Triassic, Early Jurassic and Middle-Late Jurassic oils. Abstracts of the 20th International Meeting on Organic Geochemistry, P/TUE1/35, p. 464. Holba A. G., Dzou L. I., Wood G. D., Ellis L., Adam P., Schaeer P., Albrecht P., Greene T. and Hughes W. B. (2003a) Application of tetracyclic polyprenoids as indicators of input from fresh-brackish water environments. Org. Geochem. 34, 441469. Holba A. G., Zumberge J., Huizinga B. J., Rosenstein H. and Dzou L. I. P. (2003b) Extended tricyclic terpanes as indicators of marine upwelling. 21st International Meeting on Organic Geochemistry, Book of Abstracts Part I, p. 131. Holland H. D. (2006) The oxygenation of the atmosphere and oceans. Philos. Trans. R. Soc. B 361, 903915. Horseld B., Schenk H. J., Mills N. and Welte D. H. (1992) An investigation of the in-reservoir conversion of oil to gas: compositional and kinetic ndings from closed-system programmed-temperature pyrolysis. Org. Geochem. 19, 191204. Hughes W. B., Holba A. G. and Dzou L. I. P. (1995) The ratios of dibenzothiophene to phenanthrene and pristane to phytane as indicators of depositional environment and lithology of petroleum source rocks. Geochim. Cosmochim. Acta 59, 35813598. Jahnke L. L. and Klein H. P. (1983) Oxygen requirements for formation and activity of the squalene epoxidase in Saccharomyces cerevisiae. J. Bacteriol. 155, 488492. Jones D. M. and Macleod G. (2000) Molecular analysis of petroleum in uid inclusions: a practical methodology. Org. Geochem. 31, 11631173. Karlsen D. A., Nedkvitne T., Larter S. R. and Bjorlykke K. (1993) Hydrocarbon composition of authigenic inclusionsapplications to elucidation of petroleum reservoir lling history. Geochim. Cosmochim. Acta 57, 36413659. Karlsen D. A., Skeie J. E., Backer-Owe K., Bjorlykke K., Olstad R., Berge K., Cecchi M., Vik E. and Schaefer R. G. (2004) Petroleum migration, faults and overpressure. Part II. Case history: the Haltenbanken Petroleum Province, oshore Norway. In Understanding Petroleum Reservoirs: Towards an Integrated Reservoir Engineering and Geochemical Approach (eds. J. M. Cubitt, W. A. England and S. R. Larter). Geological Society Publishing House, Geological Society Special Publication 237, pp. 305372.

868

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870 Schidlowski, S. Golubic, M. M. Kimberley, D. M. McKirdy and P. A. Trudinger), pp. 177192. Early Organic Evolution: Implications for Mineral and Energy Resources. Springer. McNeil R. I. and Bement W. O. (1996) Thermal stability of hydrocarbonslaboratory criteria and eld examples. Energy Fuels 10, 6067. Moldowan J. M., Fago F. J., Lee C. Y., Jacobson S. R., Watt D. S., Slougui N. E., Jeganathan A. and Young D. C. (1990) Sedimentary 24-n-propylcholestanes, molecular fossils diagnostic for marine algae. Science 247, 309312. Moldowan J. M., Fago F. J., Carlson R. M. K., Young D. C., Van Duyne G., Clardy J., Schoell M., Pillinger C. T. and Watt D. S. (1991a) Rearranged hopanes in sediments and petroleum. Geochim. Cosmochim. Acta 55, 33333353. Moldowan J. M., Lee C. Y., Watt D. S., Jeganathan A., Slougui N. E. and Gallegos E. J. (1991b) Analysis and occurrence of C26steranes in petroleum and source rocks. Geochim. Cosmochim. Acta 55, 10651081. Moldowan J. M., Lee C. Y., Sundararaman P., Salvatori T., Alajbeg A., Gjukic B., Demaison G. J., Slougui N. E., Watt D. S., Moldowan J. M., Albrecht P. and Philp R. P. (1992) Source correlation and maturity assessment of selected oils and rocks from the central Adriatic Basin (Italy and Yugoslavia). In Biological Markers in Sediments and Petroleum (eds. J. M. Moldowan, P. Albrecht and R. P. Philp). Prentice Hall, pp. 370401. Mossman D. J. (1987) Stratiform gold occurrences of the Witwatersrand type in the Huronian Supergroup, Ontario, Canada. Trans. Geol. Soc. South Africa 90, 168178. Mossman D. J., Nagy B. and Davis D. W. (1993) Hydrothermal alteration of organic matter in uranium ores, Elliot Lake, Canada: implications for selected organic-rich deposits. Geochim. Cosmochim. Acta 57, 32513259. Munz I. A. (2001) Petroleum inclusions in sedimentary basins: systematics, analytical methods and applications. Lithos 55, 195212. Nedkvitne T., Karlsen D. A., Bjorlykke K. and Larter S. R. (1993) Relationship between reservoir diagenetic evolution and petroleum emplacement in the Ula Field, North Sea. Mar. Pet. Geol. 10, 255270. Newell K. D., Burruss R. C. and Palacas J. G. (1993) Thermal maturation and organic richness of potential petroleum source rocks in Proterozoic Rice Formation, North American MidContinent Rift System, Northeastern Kansas. AAPG Bull. 77, 19221941. Ohmoto H., Watanabe Y., Ikemi H., Poulson S. R. and Taylor B. E. (2006) Sulphur isotope evidence for an oxic Archaean atmosphere. Nature 442, 908911. Ourisson G., Albrecht P. and Rohmer M. (1982) Predictive microbial biochemistry, a forensic approach to procaryotic membrane constituents. Trends Biochem. Sci. 7, 236239. Papineau D., Mojzsis S. J. and Schmitt A. K. (2007) Multiple sulfur isotopes from Paleoproterozoic Huronian interglacial sediments and the rise of atmospheric oxygen. Earth Planet. Sci. Lett. 255, 188212. Pepper A. S. and Dodd T. A. (1995) Simple kinetic models of petroleum formation, Part II. Oilgas cracking. Mar. Pet. Geol. 12, 321340. Peters K. E., Moldowan J. M. and Sundararaman P. (1990) Eects of hydrous pyrolysis on biomarker thermal maturity parameters: Monterey Phosphatic and Siliceous members. Org. Geochem. 15, 249265. Peters K. E., Clark M. E., Dasgupta U., McCarey M. A. and Lee C. Y. (1995) Recognition of an Infracambrian source rock based on biomarkers in the Bahewala-1 oil, India. AAPG Bull. 79, 14811494.

Karlsen D. A. and Skeie J. E. (2006) Petroleum migration, faults and overpressure, part I: calibrating basin modelling using petroleum in trapsa review. J. Petrol. Geol. 29, 227255. Kelly W. C. and Nishioka G. K. (1985) Precambrian oil inclusions in late veins and the role of hydrocarbons in copper mineralization at White Pine, Michigan. Geology 13, 334337. Kenig F., Damste J. S. S., Kock-van Dalen A. C., Rijpstra W. I. C., Huc A. Y. and de Leeuw J. W. (1995) Occurrence and origin of mono-, di- and trimethylalkanes in modern and Holocene cyanobacterial mats from Abu Dhabi, United Arab Emirates. Geochim. Cosmochim. Acta 59, 29993015. Kopp R. E., Kirschvink J. L., Hilburn I. A. and Nash C. Z. (2005) The Paleoproterozoic snowball Earth: a climate disaster triggered by the evolution of oxygenic photosynthesis. In Proceedings of the National Academy of Sciences USA, vol. 102, pp. 1113111136. Kvenvolden K. A. and Roedder E. (1971) Fluid inclusions in quartz crystals from South-West Africa. Geochim. Cosmochim. Acta 35, 12091229. Lepland A., van Zuilen M. A., Arrhenius G., Whitehouse M. J. and Fedo C. M. (2005) Questioning the evidence for Earths earliest lifeAkilia revisited. Geology 33, 7779. Lewan M. D. (1997) Experiments on the role of water in petroleum formation. Geochim. Cosmochim. Acta 61, 36913723. Li C., Peng P., Sheng G. Y., Fu J. M. and Yan Y. Z. (2003) A molecular and isotopic geochemical study of Meso- to Neoproterozoic (1.730.85 Ga) sediments from the Jixian section, Yanshan Basin, North China. Precambrian Res. 125, 337356. Li H. J., Wu T. R., Ma Z. J. and Zhang W. C. (2004) Pressure retardation of organic maturation in clastic reservoirs: a case study from the Banqiao Sag, eastern China. Mar. Pet. Geol. 21, 10831093. Losh S., Walter L., Meulbroek P., Martini A., Cathles L. and Whelan J. (2002) Reservoir uids and their migration into the South Eugene Island Block 330 reservoirs, oshore Louisiana. AAPG Bull. 86, 14631488. Love G. D., Grosjean E., Fike D. A., Grotzinger J. P., Bowring S. A., Lewis A. N., Stalvies C., Snape C. E. and Summons R. E. (2005) A >90 million year record of Neoproterozoic sponges (Porifera) in the South Oman Salt basin. Organic Geochemistry: Challenges for the 21st Century, vol. 1. 22nd IMOG, Seville, Spain, pp. 123124. Mackenzie A. S. and McKenzie D. (1983) Isomerization and aromatization of hydrocarbons in sedimentary basins formed by extension. Geol. Mag. 120, 417470. Mango F. and Elrod L. W. (1998) The carbon isotopic composition of catalytic gas: a comparative analysis with natural gas. Geochim. Cosmochim. Acta 63, 10971106. Mango F. D. (1991) The stability of hydrocarbons under the time temperature conditions of petroleum genesis. Nature 352, 146148. Mango F. D. (1997) The light hydrocarbons in petroleuma critical review. Org. Geochem. 26, 417440. Mango F. D. and Hightower J. (1997) The catalytic decomposition of petroleum into natural gas. Geochim. Cosmochim. Acta 61, 53475350. Mauk J. L. and Burruss R. C. (2002) Water washing of Proterozoic oil in the midcontinent rift system. AAPG Bull. 86, 11131127. McCarey M. A., Moldowan J. M., Lipton P. A., Summons R. E., Peters K. E., Jeganathan A. and Watt D. S. (1994) Paleoenvironmental implications of novel C30 steranes in Precambrian to Cenozoic age petroleum and bitumen. Geochim. Cosmochim. Acta 58, 529532. McKirdy D. M. and Imbus S. W. (1992) Precambrian petroleum: a decade of changing perceptions. In Early Organic Evolution: Implications for Mineral and Energy Resources (eds. M.

Preservation of biomarkers in oil inclusions for >2 billion years Peters K. E., Walters C. C. and Moldowan J. M. (2005) The Biomarker Guide. Cambridge University Press. Planche H. (1996) Finite time thermodynamics and the quasistability of closed-systems of natural hydrocarbon mixtures. Geochim. Cosmochim. Acta 60, 44474465. Preston J. C. and Edwards D. S. (2000) The petroleum geochemistry of oils and source rocks from the northern Bonaparte Basin, oshore northern Australia. Aust. Petrol. Prod. Explor. Assoc. J. 40(1), 257282. Price L. C. and Wenger L. M. (1992) The inuence of pressure on petroleum generation and maturation as suggested by aqueous pyrolysis. Org. Geochem. 19, 141159. Price L. C. (1993) Thermal stability of hydrocarbons in nature: limits, evidence, characteristics, and possible controls. Geochim. Cosmochim. Acta 57, 32613280. Price L. C., Pawlewicz M. J. and Daws T. A. (1999) Organic metamorphism in the California petroleum basins; Chapter A, Rock-Eval and vitrinite reectance. U.S. Geological Survey Bulletin, 2174A, 38. Price L. C. and DeWitt E. (2001) Evidence and characteristics of hydrolytic disproportionation of organic matter during metasomatic processes. Geochim. Cosmochim. Acta 65, 37913826. Quinlan G. M. and Beaumont C. (1984) Appalachian thrusting, lithospheric exure and the Paleozoic stratigraphy of the Eastern Interior of North America. Can. J. Earth Sci. 21, 973986. Radke M., Welte D. H. and Willsch H. (1982a) Geochemical study on a well in the Western Canada Basin: relation of the aromatic distribution pattern to maturity of organic matter. Geochim. Cosmochim. Acta 46, 110. Radke M., Willsch H., Leythaeuser D. and Teichmuller M. (1982b) Aromatic compounds of coal; relation of distribution pattern to rank. Geochim. Cosmochim. Acta 46, 18311848. Radke M., Leythaeuser D. and Teichmuller M. (1984) Relationship between rank and composition of aromatic hydrocarbons for coals of dierent origins. Org. Geochem. 6, 423430. Radke M. (1988) Application of aromatic compounds as maturity indicators in source rocks and crude oils. Mar. Pet. Geol. 5, 224236. Radke M., Rullkotter J. and Vriend S. P. (1994) Distribution of naphthalenes in crude oils from the Java Sea: source and maturation eects. Geochim. Cosmochim. Acta 58, 36753689. Rampen S. W., Schouten S., Abbas B., Panoto F. E., Muyzer G., Campbell C. N., Fehling J. and Sinninghe Damste J. S. (2007) On the origin of 24-norcholestanes and their use as agediagnostic biomarkers. Geology 35, 419422. Rasmussen B. (2000) Filamentous microfossils in a 3,235-millionyear-old volcanogenic massive sulphide deposit. Nature 405, 676679. Rasmussen B. (2005) Evidence for pervasive petroleum generation and migration in 3.2 and 2.63 Ga shales. Geology 33, 497500. Raymond J. and Blankenship R. E. (2004) Biosynthetic pathways, gene replacement and the antiquity of life. Geobiology 2, 199 203. Ruble T. E., George S. C., Lisk M. and Quezada R. A. (1998) Organic compounds trapped in aqueous uid inclusions. Org. Geochem. 29, 195205. Sajgo C., Horvath Z. A. and Leer J. (1988) An organic maturation study of the Hod-I borehole (Pannonian Basin). In The Pannonian Basin: A Study in Basin Evolution, vol. Memoir 45 (eds. L. Royden and Z. A. Horvath). American Association of Petroleum Geologists, Tulsa, and the Hungarian Geological Society, Budapest, pp. 297310. Sajgo C. (2000) Assessment of generation temperatures of crude oils. Org. Geochem. 31, 13011323. Schenk H. J., Di Primio R. and Horseld B. (1997) The conversion of oil into gas in petroleum reservoirs. Part 1: comparative kinetic

869

investigation of gas generation from crude oils of lacustrine, marine and uviodeltaic origin by programmed-temperature closed-system pyrolysis. Org. Geochem. 26, 467481. Schidlowski M. (2001) Carbon isotopes as biogeochemical recorders of life over 3.8 Ga of Earth history: evolution of a concept. Precambrian Res. 106, 117134. Schiefelbein C. F., Zumberge J. E., Cameron N. R. and Brown S. W. (1999) Petroleum systems in the South Atlantic margins. In The Oil and Gas Habitats of the South Atlantic (eds. N. R. Cameron, R. H. Bate and V. S. Clure). Geological Society of London Special Publication 153, pp. 169179. Schmidt P. W. and Williams G. E. (1999) Paleomagnetism of the Paleoproterozoic hematitic breccia and paleosol at Ville-Marie, Quebec: further evidence for the low paleolatitude of Huronian glaciation. Earth Planet. Sci. Lett. 172, 273285. Schopf J. W., Kudrayavtsev A. B., Agresti D. G., Wdowiak T. J. and Czaja A. D. (2002) Laser-Raman imagery of Earths earliest fossils. Nature 416, 7376. Schwab V., Spangenberg J. E. and Grimalt J. O. (2005) Chemical and carbon isotopic evolution of hydrocarbons during prograde metamorphism from 100 C to 550 C: Case study in the Liassic black shale formation of Central Swiss Alps. Geochim. Cosmochim. Acta 69, 18251840. Scotchman I. C., Grith C. E., Holmes A. J. and Jones D. M. (1998) The Jurassic petroleum system north and west of Britain: a geochemical oil-source correlation study. Org. Geochem. 29, 671700. Shen Y., Buick R. and Caneld D. E. (2001) Isotopic evidence for microbial sulphate reduction in the early Archaean era. Nature 410, 7781. Sherman L. S., Waldbauer J. R. and Summons R. E. (2007) Improved methods for isolating and validating indigenous biomarkers in Precambrian rocks. Org. Geochem. 38, 1987 2000. Sieskind O., Joly G. and Albrecht P. (1979) Simulation of the geochemical transformation of sterols: superacid eects of clay minerals. Geochim. Cosmochim. Acta 43, 16751679. Simoneit B. R. T., Schoell M., Dias R. F. and Neto F. R. D. (1993) Unusual carbon isotope compositions of biomarker hydrocarbons in a Permian tasmanite. Geochim. Cosmochim. Acta 57, 42054211. Sinninghe Damste J. S., Kenig F., Koopmans M. P., Koster J., Schouten S., Hates J. M. and de Leeuw J. W. (1995) Evidence for gammacerane as an indicator of water column stratication. Geochim. Cosmochim. Acta, 59. Sofer Z., Regan D. R. and Muller D. S. (1993) Sterane isomerization ratios of oils as maturity indicators and their use as an exploration tool, Neuquen basin, Argentina. XII Congreso de Geologico Argentino y II Congreso de Exploracion de Hidrocarburos Actas, vol. 1, pp. 407411. Summons R., Bradley A., Jahnke L. and Waldbauer J. (2006) Steroids, triterpenoids and molecular oxygen. Philos. Trans. R. Soc. Lond., B 361, 951968. Summons R. E. and Powell T. G. (1986) Chlorobiaceae in Palaeozoic seas revealed by biological markers, isotopes and geology. Nature 319, 763765. Summons R. E., Brassell S. C., Eglinton G., Evans E., Horodyski R. J., Robinson N. and Ward D. M. (1988a) Distinctive hydrocarbon biomarkers from fossiliferous sediments of the Late Proterozoic Walcott Member, Chuar Group, Grand Canyon, Arizona. Geochim. Cosmochim. Acta 52, 26252637. Summons R. E., Powell T. G. and Boreham C. J. (1988b) Petroleum geology and geochemistry of the Middle Proterozoic McArthur Basin, Northern Australia: III. Composition of extractable hydrocarbons. Geochim. Cosmochim. Acta 52, 1747 1763.

870

S.C. George et al. / Geochimica et Cosmochimica Acta 72 (2008) 844870 ical study in the Barrandian Basin (Lower Palaeozoic, Czech Republic). Org. Geochem. 33, 13191341. Volk H., Dutkiewicz A., George S. and Ridley J. (2003) Oil migration in the Middle Proterozoic Roper Superbasin, Australia: evidence from oil inclusions and their geochemistries. J. Geochem. Expl. 7879, 437441. Volk H., George S. C., Dutkiewicz A. and Ridley J. (2005) Characterisation of uid inclusion oil in a Mid-Proterozoic sandstone and dolerite (Roper Superbasin, Australia). Chem. Geol. 223, 109135. Volkman J. K., Alexander R., Kagi R. I. and Woodhouse G. W. (1983) Demethylated hopanes in crude oils and their applications in petroleum geochemistry. Geochim. Cosmochim. Acta 47, 785794. Volkman J. K. (2005) Sterols and other triterpenoids: source specicity and evolution of biosynthetic pathways. Org. Geochem. 36, 139159. Walter M. R., Buick R. and Dunlop J. S. R. (1980) Stromatolites 34003500 Myr old from the North Pole area, Western Australia. Nature 284, 443445. Wang Y., Zhang S., Wang F., Wang Z., Zhao C., Wang H., Liu J., Lu J., Geng A. and Liu D. (2006) Thermal cracking history by laboratory kinetic simulation of Paleozoic oil in eastern Tarim Basin, NW China, implications for the occurrence of residual oil reservoirs. Org. Geochem. 37, 18031815. Williams G. E. and Schmidt P. W. (1997) Paleomagnetism of the Paleoproterozoic Gowganda and Lorrain formations, Ontario: low paleolatitude for Huronian glaciation. Earth Planet. Sci. Lett. 153, 157169. Young G. M., Long D. G. F., Fedo C. M. and Nesbitt H. W. (2001) Paleoproterozoic Huronian basin: product of a Wilson cycle punctuated by glaciations and meteorite impact. Sed. Geol. 141142, 233254. Zou Y. R. and Peng P. (2001) Overpressure retardation of organicmatter maturation: a kinetic model and its application. Mar. Pet. Geol. 18, 707713. Associate editor: Robert C. Burruss

Summons R. E. and Walter M. R. (1990) Molecular fossils and microfossils of prokaryotes and protists from Proterozoic sediments. Am. J. Sci. 290-A, 212244. Summons R. E. and Jahnke L. L. (1992) Hopenes and hopanes methylated in ring A: correlation of the hopanoids from extant methylotrophic bacteria with their fossil analogues. In Biological Markers in Sediments and Petroleum (eds. J. M. Moldowan, P. Albrecht and R. P. Philp). Prentice Hall, pp. 182200. Summons R. E., Jahnke L. L., Hope J. M. and Logan G. A. (1999) 2-Methylhopanoids as biomarkers for cyanobacterial oxygenic photosynthesis. Nature 400, 554557. Tannenbaum E., Ruth E. and Kaplan I. R. (1986) Steranes and triterpanes generated from kerogen pyrolysis in the absence and presence of minerals. Geochim. Cosmochim. Acta 50, 805812. Thompson K. F. M. (1983) Classication and thermal history of petroleum based on light hydrocarbons. Geochim. Cosmochim. Acta 47, 303316. Thompson K. F. M. (1987) Fractionated aromatic petroleums and the generation of gas-condensates. Org. Geochem. 11, 573590. Ungerer P. F., Behar M., Villalba M., Audibert A. and Heum O. R. (1987) Kinetic modeling of oil cracking. Org. Geochem. 13, 857868. van Aarssen B. G. K., Bastow T. P., Alexander R. and Kagi R. I. (1999) Distributions of methylated naphthalenes in crude oils: indicators of maturity, biodegradation and mixing. Org. Geochem. 30, 12131227. Vandenbroucke M., Behar F. and Rudkiewicz J. L. (1999) Kinetic modelling of petroleum formation and cracking: implications from the high pressure/high temperature Elgin Field (UK, North Sea). Org. Geochem. 30, 11051125. Ventura G. T., Kenig F., Reddy C. M., Schieber J., Frysinger G. S., Nelson R. K., Dinel E., Gaines R. B. and Schaeer P. (2007) Molecular evidence of Late Archean archaea and the presence of a subsurface hydrothermal biosphere. In Proceedings of the National Academy of Sciences of the United States of America, vol. 104, pp. 1426014265. Volk H., Horseld B., Mann U. and Suchy V. (2002) Variability of petroleum inclusions in vein, fossil and vug cements: geochem-

Das könnte Ihnen auch gefallen