Introduction
Lipid biomarkers are molecular fossils that originate from biological precursor compounds (e.g. Peters et al. Reference Peters, Walters and Moldowan2005a; Brocks et al. Reference Brocks, Jarrett, Sirantoine, Kenig, Moczydłowska, Porter and Hope2016). Ever since Treibs (Reference Treibs1936) first discovered chlorophyll-derived porphyrins in ca. 230 Ma old Triassic oil shales, biomarkers have been extensively applied for tracing organisms and evolutionary processes through Earth's history and for the reconstruction of past environments. Effective biomarkers should be source-specific, resistant against thermal- and biodegradation, and analysable in natural samples with routine techniques (e.g. Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Peters et al. Reference Peters, Walters and Moldowan2005a). Therefore lipids are particularly important, because they fulfil these essential demands and their hydrocarbon skeletons have a high potential for preservation over geological timescales (e.g. Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Summons Reference Summons, Laflamme, Schiffbauer and Darroch2014). Given that the record of body fossils shows huge gaps due to taphonomic effects and the fact that many organisms did not produce preservable hard parts, organic biomarkers may provide a complementary record of past ecosystems independent from the body fossil record (e.g. Hallmann et al. Reference Hallmann, Kelly, Gupta, Summons, Laflamme, Schiffbauer and Dornbos2011; Brocks et al. Reference Brocks, Jarrett, Sirantoine, Kenig, Moczydłowska, Porter and Hope2016). In that respect, biomarkers have become instrumental for the reconstruction of early Life on Earth and, potentially, for the detection of life in the extraterrestrial realm (e.g. Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Brocks & Pearson Reference Brocks and Pearson2005; Hallmann et al. Reference Hallmann, Kelly, Gupta, Summons, Laflamme, Schiffbauer and Dornbos2011; Summons Reference Summons, Laflamme, Schiffbauer and Darroch2014; Olcott Marshall & Cestari Reference Olcott Marshall and Cestari2015).
Despite the wide use of biomarkers for paleoreconstructions, however, it is not always clear how exactly they reflect a paleoecosystem. For instance, commonly less than 1% of organic matter produced in a marine environment is introduced into the sediment due to abiotic oxidation and intensive microbial reprocessing in the water column and on the sea floor by aerobic and anaerobic respiration (e.g. Hedges & Keil Reference Hedges and Keil1995; Peters et al. Reference Peters, Walters and Moldowan2005a; Middelburg & Meysman Reference Middelburg and Meysman2007; Vandenbroucke & Largeau Reference Vandenbroucke and Largeau2007). Furthermore, microbially-driven processes observed in microbial mat settings may also profoundly affect the biomarker inventory (i.e. the ‘mat seal effect’; Pawlowska et al. Reference Pawlowska, Butterfield and Brocks2013; Blumenberg et al. Reference Blumenberg, Thiel and Reitner2015). During later stages biomarkers are additionally influenced by increasing temperatures (T) and pressures (P). While biomarkers undergo major structural changes during catagenesis (corresponding to temperatures of 50–150°C; Tissot & Welte Reference Tissot and Welte1984; Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Peters et al. Reference Peters, Walters and Moldowan2005a), they are almost entirely decomposed during metagenesis and metamorphism (corresponding to temperatures of 150–200°C and >200°C, respectively; Tissot & Welte Reference Tissot and Welte1984). High fluid and/or gas pressures, in contrast, may retard the thermal destruction of organic matter (Price Reference Price1993). For the decomposition of biomarkers, geological time is commonly considered as an essential factor in addition to temperature (Tissot & Welte Reference Tissot and Welte1984; Peters et al. Reference Peters, Walters and Moldowan2005b); however, the effect of geological time as a major controlling parameter has also been questioned (Price Reference Price1983, Reference Price1993). Further geological factors that influence the preservation of biomarkers are the heating rate and mineralogy, especially the adsorption of carbon compounds onto clay mineral surfaces and mineral catalysts that promote the degradation of hydrocarbons (e.g. Mango Reference Mango1996; Mango & Hightower Reference Mango and Hightower1997; Kennedy et al. Reference Kennedy, Pevear and Hill2002; Brocks & Summons Reference Brocks, Summons and Schlesinger2003). While many studies focused on the generation, expulsion and migration of petroleum hydrocarbons, it is as yet often poorly constrained as to which extent individual organic compounds are affected by these processes, and whether changing P/T conditions cause shifts in relative abundances and ratios of selected biomarkers.
The main portion of organic matter in sediments and rocks is present in the form of ‘kerogen’, i.e. macromolecules insoluble in usual organic solvents, acids and bases (as opposed to the solvent-soluble ‘bitumen’ fraction; Durand Reference Durand and Durand1980). Based on estimations, about 1016 tonnes of C are bound to kerogen, compared with only 1012 tonnes of living biomass (Durand Reference Durand and Durand1980; Vandenbroucke & Largeau Reference Vandenbroucke and Largeau2007). Kerogen formation is still not fully understood, but essentially includes selective preservation of biomacromolecules and polymerization (especially polycondensation) and immobilization processes of organic molecules during early diagenesis (e.g. Durand Reference Durand and Durand1980; De Leeuw et al. Reference De Leeuw, Versteegh, van Bergen, Rozema, Aerts and Cornelissen2006; Vandenbroucke & Largeau Reference Vandenbroucke and Largeau2007; Hallmann et al. Reference Hallmann, Kelly, Gupta, Summons, Laflamme, Schiffbauer and Dornbos2011). As macromolecular-bound lipid moieties are more resistant against microbial degradation and thermal overprint than the free molecules, kerogen may facilitate the preservation of biomarkers (e.g. Love et al. Reference Love, Snape, Carr and Houghton1995; Killops & Killops Reference Killops and Killops2005; Hallmann et al. Reference Hallmann, Kelly, Gupta, Summons, Laflamme, Schiffbauer and Dornbos2011). Due to their solid state nature, kerogen particles may remain immobile in the sedimentary environment over geological time scales, and thus are considered syngenetic to the host rock (e.g. Brocks et al. Reference Brocks, Buick, Logan and Summons2003a, Reference Brocks, Love, Snape, Logan, Summons and Buickb; Love et al. Reference Love, Stalvies, Grosjean, Meredith, Snape, Kelley and Bambach2008). Kerogen bound biomarkers are therefore commonly used in the reconstruction of Precambrian ecosystems (e.g. Love et al. Reference Love2009; Duda et al. Reference Duda, Blumenberg, Thiel, Simon, Zhu and Reitner2014; Brocks et al. Reference Brocks, Jarrett, Sirantoine, Kenig, Moczydłowska, Porter and Hope2016). However, it is still poorly known how P and T affect the biomarkers that are bound into the kerogen. Therefore maturation experiments could help to better understand these processes.
Numerous earlier maturation experiments have been conducted on the Green River Shale (GRS) (e.g. Burnham & Singleton Reference Burnham and Singleton1983; Evans & Felbeck Reference Evans and Felbeck1983; Huizinga et al. Reference Huizinga, Aizenshtat and Peters1988; Ruble et al. Reference Ruble, Lewan and Philp2001). The GRS comprises relatively immature oil source rocks that contain most lipid biomarker classes regularly used for paleoreconstructions (e.g. Burnham et al. Reference Burnham, Clarkson, Singleton, Wong and Crawford1982; Evans & Felbeck Reference Evans and Felbeck1983). Not surprisingly, biomarkers in the GRS bitumen have been extensively investigated (e.g. Eglinton & Douglas Reference Eglinton and Douglas1988; Collister et al. Reference Collister, Summons, Lichtfouse and Hayes1992; Schoell et al. Reference Schoell, Hwang, Carlson and Welton1994; Koopmans et al. Reference Koopmans, de Leeuw and Sinninghe Damsté1997; Ruble et al. Reference Ruble, Lewan and Philp2001), also as possible analogue studies for current and future Mars missions (Olcott Marshall & Cestari Reference Olcott Marshall and Cestari2015). The natural bitumen (i.e. geologically produced) has been reported to contain the major portion of the total steranes and hopanes in the GRS. Another minor portion can be released from kerogen via pyrolysis, showing somewhat different distributions of single steranes and hopanes (Eglinton & Douglas Reference Eglinton and Douglas1988). However, maturation experiments have been mostly focused on petroleum generation from the host rock and changes in the bulk kerogen composition, but so far surprisingly few comprehensive studies exist on the fate of distinct kerogen-bound biomarkers (e.g. Eglinton & Douglas Reference Eglinton and Douglas1988; Monthioux & Landais Reference Monthioux and Landais1989; Peters et al. Reference Peters, Moldowan and Sundararaman1990; Price Reference Price1993).
This study aims at assessing the behaviour of selected kerogen-bound hydrocarbon biomarkers, namely n-alkanes, pristane, phytane, gammacerane, steranes, hopanes and cheilanthanes, whose presence and ratios are commonly used as life tracers and for the reconstruction of paleoenvironments and evolutionary processes (Peters et al. Reference Peters, Walters and Moldowan2005a, Reference Peters, Walters and Moldowanb). By conducting hydrous pyrolysis of the GRS kerogen in closed gold capsules at elevated pressures and for different time intervals and temperature regimes, we evaluated the stability of selected compounds, or compound classes, in relation to each other. Our findings reveal major differences in the preservation of biomarkers along the maturation pathway. This ‘thermal taphonomy’ has potentially wide implications for their interpretation in ancient rocks and, potentially, extraterrestrial materials.
Materials and methods
Samples and maturation experiments
The material used for experimental maturation was an oil shale from the Eocene Green River Formation in Eastern Utah (White River Mine located within BLM Oil Shale Research, Development, and Demonstration Lease UTU-84087). Bituminous brown coal samples (SM20; Gruber & Sachsenhofer Reference Gruber and Sachsenhofer2001) from the Noric Depression (Eastern Alps) were maturated in parallel to the GRS samples (same autoclave, separate gold capsules), to obtain exact vitrinite reflectance values for each maturation step.
The ground GRS rock sample was 3x extracted with dichloromethane (DCM). The hydrocarbon fraction prepared from the resulting total extract (see the section ‘Extraction and fractionation’) was used as the untreated reference (0 h). Kerogen was isolated from the pre-extracted GRS using hydrochloric acid and hydrofluoric acid (see e.g. review by Durand & Nicaise Reference Durand, Nicaise and Durand1980). The residue was homogenised with a mortar and again 3x extracted with DCM to remove residual bitumen. The dried kerogen (ca. 10 mg) was sealed into gold capsules (3.0 mm outer diameter, 2.6 mm inner diameter, 15 mm length) by arc welding using a Lampert™ tungsten inert gas impulse micro welding device. Extensive heating of the sample during welding was avoided by wrapping the capsule in wet tissue paper. Samples were maturated at 300°C and 2 kbar for 2 to 2400 h (100 days). Four additional short duration experiments at 400°C (2, 5, 24 and 48 h) were performed to check for the upper limit of biomarker preservation. Experiments with 2–720 h durations were carried out in rapid-quench cold seal pressure vessels (CSPV) enabling a fast heating and cooling of the samples within a few seconds (see Schmidt et al. Reference Schmidt, Blum-Oeste and Flagmeier2013 for technical details). The 2400 h duration experiments were conducted in conventional CSPV with slower heating and cooling rates (see Le Bayon et al. Reference Le Bayon, Buhre, Schmidt and Ferreiro Mählmann2012 for technical details).
Extraction and fractionation
After maturation, the capsules were opened using a metal spike, forceps and scissors, which were thoroughly cleaned between the samples using acetone. About 2/3 of the material was then extracted 4x with each 2 ml of an n-hexane/DCM mixture (1:1, V:V; 20 min ultrasonication). The resulting extracts were combined and carefully reduced using N2. The extracts were never completely dried before the next step of processing to avoid a major loss of low-boiling compounds (Ahmed & George Reference Ahmed and George2004). Extracts were dried onto a small amount of silica gel and fractionated by column chromatography (diameter: 1.5 cm, height: 8 cm, Merck silica gel 60). Saturated hydrocarbons were eluted with 3 ml n-hexane. The resulting saturated hydrocarbon fractions were reduced to 1 ml and analysed by gas chromatography–mass spectrometry (GC–MS).
GC–MS and vitrinite reflectance
GC-MS analyses were carried out using a Thermo Fisher Trace 1300 Series GC coupled to a Thermo Fisher Quantum XLS Ultra MS. The GC was equipped with a capillary column (Phenomenex Zebron ZB-1MS/Phenomenex Zebron ZB-5, 30 m, 0.1 µm film thickness, inner diameter 0.25 mm). Fractions were injected into a splitless injector and transferred to the GC column at 270°C. The carrier gas was He with a flow rate of 1.5 ml min−1. The GC oven temperature was ramped 80°C (1 min) to 310°C at 5°C min−1 (held 20 min). Electron ionization mass spectra were recorded in full scan mode at 70 eV electron energy with a mass range of m/z 50–600 and a scan time of 0.42 s.
To determine vitrinite reflectance of the reference coal samples, samples were embedded into epoxy and afterwards cut, abraded (four steps; P400 to P2500), and polished (diamond and alumina slurries, four steps, grain sizes 9–0.05 µm, each step 5 min). After polishing the minimum and maximum reflectance of vitrinites have been measured in linear polarised light at 546 nm using a SpectraVision PMT system (A.S. & Co.) consisting of a Zeiss Axio Imager.A2m microscope (Carl Zeiss Microscopy, LLC, New York, United States) with an attached Zeiss MCS CCD/UVNIR spectrometer. Fifteen measurements on random grains were made for each sample. Standard deviations were generally <0.08%.
Results and discussion
The hydrocarbon fraction of the untreated GRS bitumen (0 h) contains n-alkanes, acyclic (e.g. pristane and phytane) and cyclic (e.g. steranes and terpanes) isoprenoids as well as some other biomarkers that are not in the focus of this study (e.g. carotenoids; Fig. 1). The 300°C experiment covers a maturity range from immature (0.49% R O, untreated sample 0 h) to the main catagenetic stage (1.21% R O, 2400 h) thus being well suited to track continuous changes in the biomarker inventory with increasing maturity. The 400°C experiment conducted to check for the upper limit of biomarker preservation represents a maturity range between 1.19% R O (2 h) and 1.83% R O (48 h).
Fig. 1. GC–MS chromatogram (total ion current, TIC) of Green River Shale bitumen (untreated). Numbers refer to number of carbon atoms. Filled circles denote n-alkanes, stars denote steranes, triangles denote hopanes, diamonds denote cheilanthanes; TMTD,2,6,10 trimethyl-tridecane; N, norpristane; Pr, pristane; Ph, phytane; Gc, gammacerane; IS, internal standard (eicosane d42 10 mg l−1).
n-Alkanes
n-Alkanes in the original GRS bitumen have chain lengths from n-C15 to n-C37. They show a bimodal distribution with maxima at n-C17, and n-C29/n-C31 (Fig. 2). In the long-chain range, n-alkanes exhibit a clear odd-over-even predominance as reflected by a carbon preference index (CPI) of 3.6 (calculated after Bray & Evans Reference Bray and Evans1961; Table 1). n-C17 is common in extant cyanobacteria and algae and its predominance in the short-chain range can be attributed to planktonic photoautotrophs (e.g. Blumer et al. Reference Blumer, Guillard and Chase1971; Hoffmann et al. Reference Hoffmann, Foster, Powell and Summons1987; Jacobson et al. Reference Jacobson, Hatch, Teerman and Askin1988). The also abundant long chain n-alkanes with a pronounced odd-over-even predominance in the range of n-C27 to n-C33 are commonly explained by inputs of epicuticular leaf waxes derived from higher land plants (Eglinton & Hamilton Reference Eglinton and Hamilton1967). The n-alkane distribution found in the GRS bitumen is thus well in line with a high-productivity lacustrine environment (e.g. Tissot et al. Reference Tissot, Deroo and Hood1978; Horsfield et al. Reference Horsfield1994). During experimental maturation, this original n-alkane distribution is profoundly altered. While the distinct maximum at n-C17 and the odd-over-even predominance in the high molecular weight range become less pronounced (Fig. 2, 24–2400 h), the CPI decreases from 3.6 (0 h) to 1.4 after 2400 h (1.21% R O, Table 1). It should be noted that these experimental data reflect a lag of the CPI with respect to vitrinite reflectance, as a CPI-value approaching 1 is generally expected at ~0.9% R O (i.e. peak oil window; Peters et al. Reference Peters, Walters and Moldowan2005b). In our experiment, the ‘biological’ chain length preferences of n-alkanes are completely vanished only after 48 h at 400°C (corresponding to 1.83% R O, not shown). Possible reasons for this mismatch between biomarkers and vitrinite reflectance are discussed below. Nevertheless, our results underline that the source-specific features of individual n-alkane chain lengths gradually disappear with increasing maturity (Tissot et al. Reference Tissot, Deroo and Hood1978; Price Reference Price1993; Peters et al. Reference Peters, Walters and Moldowan2005a).
Fig. 2. GC–MS chromatograms (n-alkanes and isoprenoids, m/z 85) of Green River Shale bitumen (untreated, 0 h) and experimentally maturated Green River Shale kerogens (300°C; 24, 240 and 2400 h, respectively). Numbers refer to number of carbon atoms; filled circles denote n-alkanes; TMTD, 2,6,10 trimethyl-tridecane; N, norpristane; Pr, pristane; Ph, phytane.
Table 1. Biomarker ratios, indices and corresponding vitrinite reflectances (VR) of Green River Shale bitumen and experimentally maturated kerogens (300°C)
All ratios and indices were calculated from peak integrals: Isoprenoids and n-alkanes, m/z 85; steranes, m/z 217; hopanes and cheilanthanes, m/z 191. CPI, carbon preference index (after Bray & Evans Reference Bray and Evans1961); SD, standard deviation (n = 15); Gammacerane index = 10 × gammacerane/(gammacerane + C30 hopane).
Acyclic isoprenoids
Acyclic isoprenoids in the GRS bitumen include 2,6,10-trimethyltridecane (C16), norpristane (C18), pristane (C19) and phytane (C20) (Figs 1 and 2). The pristane/phytane ratio (Pr/Ph) is 0.3, while the pristane/n-C17 (Pr/n-C17) and phytane/n-C18 (Ph/n-C18) ratios are 1.4 and 26.5, respectively (Table 1). Pristane and phytane are biomarkers that mainly originate from phytol, the corresponding alcohol to the phytyl side chain of chlorophyll (Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Peters et al. Reference Peters, Walters and Moldowan2005b). However, pristane can also derive from tocopherols (E vitamins) (Goossens et al. Reference Goossens, de Leeuw, Schenck and Brassell1984), while phytane can derive from archaeols (i.e. ether lipid compounds from Archaea; Brocks & Summons Reference Brocks, Summons and Schlesinger2003). Pristane in oils and bitumens can furthermore originate from cracking of isoprenoid moieties that are bound to the kerogen matrix (Larter et al. Reference Larter, Solli, Douglas, de Lange and de Leeuw1979; Goossens et al. Reference Goossens, de Lange, de Leeuw and Schenck1988a, Reference Goossens, Due, de Leeuw, van de Graaf and Schenckb). However, given the type of environment (see above), chlorophyll and/or tocopherols are likely to be the major precursors for these isoprenoids in the GRS.
Pr/Ph ratios are commonly used for the reconstruction of paleoredox conditions in ancient settings (e.g. Blumenberg et al. Reference Blumenberg, Thiel, Riegel, Kah and Reitner2012; Yamada et al. Reference Yamada, Ueno, Yamada, Komiya, Han, Shu, Yoshida and Maruyama2014; Tulipani et al. Reference Tulipani2015; Luo et al. Reference Luo, Hallmann, Xie, Ruan and Summons2015; Naeher & Grice Reference Naeher and Grice2015), as phytol is preferentially transformed to pristane under oxic conditions and to phytane under anoxic conditions (Didyk et al. Reference Didyk, Simoneit, Brassell and Eglinton1978; Peters et al. Reference Peters, Walters and Moldowan2005b). However, the Pr/Ph ratio is also affected by thermal maturity (Ten Haven et al. Reference Ten Haven, de Leeuw, Rullkötter and Sinninghe Damsté1987; Peters et al. Reference Peters, Walters and Moldowan2005b). Our data show that the Pr/Ph ratio increases until a maturation time of 240 h, but then decreases (Fig. 3(a)). This change in the evolution of Pr/Ph ratios could be due to a preferential release of pristane from kerogen during early catagenesis (Peters et al. Reference Peters, Walters and Moldowan2005b).
Fig. 3. Evolution of biomarker-based ratios and indices of untreated (0 h) and experimentally maturated (2–2400 h) Green River Shale samples (see also Table 1). Please note that the untreated Green River Shale bitumen is always plotted at 1 h (instead of 0 h) due to the logarithmic scale. Gammacerane index = 10 × gammacerane/(gammacerane + C30 hopane). Pr, pristane; Ph, phytane.
The Pr/n-C17 and Ph/n-C18 ratios are used to assess the impact of biodegradation and thermal maturation on bitumens (e.g. Hieshima & Pratt Reference Hieshima and Pratt1991; Meredith et al. Reference Meredith, Kelland and Jones2000; Peters et al. Reference Peters, Walters and Moldowan2005a, Reference Peters, Walters and Moldowanb; Blumenberg et al. Reference Blumenberg, Thiel, Riegel, Kah and Reitner2012). Both ratios typically increase due to biodegradation (Peters et al. Reference Peters, Walters and Moldowan2005a, Reference Peters, Walters and Moldowanb) and sdecrease with increasing maturity (Alexander et al. Reference Alexander, Kagi and Woodhouse1981; Ten Haven et al. Reference Ten Haven, de Leeuw, Rullkötter and Sinninghe Damsté1987; Peters et al. Reference Peters, Walters and Moldowan2005b). As expected, our data also show a decrease in Ph/n-C18 over time, albeit with a somewhat steeper drop of the ratio before 100 h (Fig. 3(b)). This could be explained by a time lag between the formation of phytane and n-C18 from kerogen; phytane is dominantly generated during earliest maturation whereas a pronounced formation of n-C18 occurs at higher maturities. This effect may be influenced by heating rate (Burnham et al. Reference Burnham, Clarkson, Singleton, Wong and Crawford1982) or kerogen composition.
The impact of maturity on Pr/n-C17 seems to be more complex as the ratio initially increases, summits at 48 h and then decreases with further maturation time (Fig. 3(b)). In natural samples, an increase of the Pr/n-C17 ratio may result from preferential biodegradation of the n-alkane (Peters et al. Reference Peters, Walters and Moldowan2005b), which can be excluded here. Instead, the observed phenomenon may be explained by a higher release of pristane relative to n-C17 from the kerogen during early maturation (e.g. from tocopherol or other isoprenoid moieties; Larter et al. Reference Larter, Solli, Douglas, de Lange and de Leeuw1979; Goossens et al. Reference Goossens, de Leeuw, Schenck and Brassell1984; Goossens et al. Reference Goossens, Due, de Leeuw, van de Graaf and Schenck1988b). This is in line with the observed initial increase of the Pr/Ph ratio (Fig. 3(a)). Taken together our data suggest a more reliable applicability of Ph/n-C18 as a maturity indicator, while that of Pr/n-C17 is restricted.
Cyclic isoprenoids
Cyclic isoprenoids in the GRS bitumen include steranes, hopanes, gammacerane and cheilanthanes (Fig. 1). The 20S and 20R isomers of C27 to C29 steranes are abundant (Figs 1 and 4). Hopanes range from C27 to C32, with the C30 pseudohomologue being particularly abundant (Figs 1 and 4). For all homohopanes (i.e. hopanes >C30), both 22S and 22R isomers are present (Fig. 4). Cheilanthanes range from C20 to C26 (Fig. 4). All these compounds are useful biosignatures. While steranes derive from sterols, which control the permeability and rigidity of eukaryotic cell membranes (e.g. Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Peters et al. Reference Peters, Walters and Moldowan2005b), hopanes mainly originate from bacteriohopanepolyols (C35), which are lipids in some groups of bacteria (Rohmer et al. Reference Rohmer, Bouvier-Nave and Ourisson1984; Brocks & Pearson Reference Brocks and Pearson2005). The pentacyclic triterpenoid gammacerane (C30) derives from tetrahymanol (Ten Haven et al. Reference Ten Haven, Rohmer, Rullkötter and Bisseret1989), which has various biological sources but is particularly abundant in anoxygenic phototrophs (Rhodopseudomonas palustris; Kleemann et al. Reference Kleemann, Poralla, Englert, Kjøsen, Liaaen-Jensen, Neunlist and Rohmer1990; Eickhoff et al. Reference Eickhoff, Birgel, Talbot, Peckmann and Kappler2013) and bacterivorous ciliates (Harvey & McManus Reference Harvey and McManus1991; Sinninghe Damsté et al. Reference Sinninghe Damsté, Kenig, Koopmans, Köster, Schouten, Hayes and de Leeuw1995). Cheilanthanes are tricyclic terpanes (usually C19 to C45, rarely up to C54) that were applied as biomarkers in many studies (Moldowan et al. Reference Moldowan, Seifert and Gallegos1983; De Grande et al. Reference De Grande, Aquino Neto and Mello1993; Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Brocks & Pearson Reference Brocks and Pearson2005; Peters et al. Reference Peters, Walters and Moldowan2005b, and references therein). The actual source of cheilanthanes is still unknown but it has been suggested that the precursor lipids originate from eukaryotic algae (e.g. Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Brocks & Pearson Reference Brocks and Pearson2005).
Fig. 4. GC–MS chromatograms of terpanes (m/z 191, left side) and steranes (m/z 217, right side) in the Green River Shale bitumen (untreated, 0 h) and experimentally maturated Green River Shale kerogens (300°C; 24, 240 and 2400 h, respectively). Numbers refer to number of carbon atoms. Gc, gammacerane.
The total steranes/hopanes ratio (St/H) is commonly used to evaluate relative inputs of eukaryotic and prokaryotic biomass (e.g. Brocks et al. Reference Brocks, Logan, Buick and Summons1999; Peters et al. Reference Peters, Walters and Moldowan2005b; Love et al. Reference Love2009; Blumenberg et al. Reference Blumenberg, Thiel, Riegel, Kah and Reitner2012; Flannery & George Reference Flannery and George2014). However, it is known that this ratio may change through taphonomic processes (Pawlowska et al. Reference Pawlowska, Butterfield and Brocks2013; Blumenberg et al. Reference Blumenberg, Thiel and Reitner2015). Also thermal maturation appears to be critical, leading to a decreasing St/H ratio (Requejo Reference Requejo1994; Norgate et al. Reference Norgate, Boreham and Wilkins1999). In our 300°C experiment, the St/H ratio shows a four-fold decrease from 0.4 to 0.1, testifying a strong impact of maturity in the range of 0.49–1.21% R O (Table 1, Fig. 3(c)). At even higher maturities, in the 400°C experiment, a point was reached where steranes were entirely decomposed, whereas hopanes were still present (St/H = 0 at 1.38% R O). This different behaviour of steranes and hopanes has to be considered in addition to taphonomic processes in paleoenvironmental studies, particularly in ancient rocks of higher maturities.
Further biomarker based maturity proxies include the total cheilanthanes/hopanes ratio (Chei/H) as well as isomerization indices of homohopanes (22S/(22S + 22R)) and steranes (20S/(20S + 20R)) (e.g. Rullkötter et al. Reference Rullkötter, Meyers, Schaefer and Dunham1986; Summons et al. Reference Summons, Brassell, Eglinton, Evans, Horodyski, Robinson and Ward1988; Brocks et al. Reference Brocks, Logan, Buick and Summons1999, Reference Brocks, Jarrett, Sirantoine, Kenig, Moczydłowska, Porter and Hope2016; Ruble et al. Reference Ruble, Lewan and Philp2001; Peters et al. Reference Peters, Walters and Moldowan2005b; Blumenberg et al. Reference Blumenberg, Thiel, Riegel, Kah and Reitner2012). Chei/H depends on maturity as cheilanthanes are more resistant to thermal maturation than hopanes and are preferentially released from the kerogen with increasing maturity (Aquino Neto et al. Reference Aquino Neto, Trendel, Restle, Connan and Albrecht1981; Peters et al. Reference Peters, Moldowan and Sundararaman1990). In our 300°C experiment, this is confirmed by a continuous increase of the Chei/H ratio from 0.1 to 1.0 (Fig. 3(d), Table 1). In the case of homohopane and sterane isomerization, the biological R-configuration is progressively transferred to the S-configuration during maturation, thus leading to a relative increase of the S isomers (Brocks & Summons Reference Brocks, Summons and Schlesinger2003; Peters et al. Reference Peters, Walters and Moldowan2005b; Hallmann et al. Reference Hallmann, Kelly, Gupta, Summons, Laflamme, Schiffbauer and Dornbos2011). In our experiment, sterane and hopane isomerization consistently increased with maturation time, as expected (Fig. 3(e)). Thermal equilibrium values (22S/(22S + 22R) hopanes: 0.57–0.62, 20S/(20S + 20R) steranes: 0.52–0.55; Peters et al. Reference Peters, Walters and Moldowan2005b) were reached after 360 h (hopanes) and 720 h (steranes), which in our experiments correspond to 0.96 and 1.03% R O, respectively (Table 1). Thus, a shift can be seen between our experimental data and published biomarker equilibrium values, as full isomerization in geological systems is expected at roughly 0.6 and 0.8% R O, respectively (Peters et al. Reference Peters, Walters and Moldowan2005b). This mismatch may, to some degree, be due to differences in the maturation behaviour between the two samples (lignite versus GRS), as well as a lack of resolution due to the setting of time intervals for the experiments. On the other hand, it is commonly observed that vitrinite reflectance data and biomarkers may give different results regarding thermal maturity, with variations of +/−0.1% R O being common (Peters et al. Reference Peters, Walters and Moldowan2005b). Although both vitrinite reflectance and biomarker indicators are controlled by thermal stress (temperature and time), biomarkers may be additionally influenced by other effects. For example, the CPI is also affected by organic matter input (source effect) and biodegradation, while isomerization indices might be influenced by selective thermal decomposition of individual isomers, and/or fast reverse reactions (see Peters et al. Reference Peters, Walters and Moldowan2005b for detailed discussion). Therefore the comparability of short-term experimental data to natural samples is limited.
The pronounced occurrence of gammacerane and a high gammacerane index (i.e. 10 × Gc/(Gc + C30 hopane)) are highly specific indicators for water column stratification and, potentially, increased salinity (Sinninghe Damsté et al. Reference Sinninghe Damsté, Kenig, Koopmans, Köster, Schouten, Hayes and de Leeuw1995; Peters et al. Reference Peters, Walters and Moldowan2005b). Our data show that the gammacerane index varies little in the course of the 300°C experiment (Fig. 3(f)), underlining that this proxy is barely affected by maturation processes.
When are the biomarkers gone?
The 400°C experiment (1.19–1.83% R O) provides information on the overall thermal stability of biomarkers beyond the diagenetic/catagenetic stages. Among the biomarkers investigated in this study, steranes turned out as the thermally least stable compounds. While traces of steranes were still present at the end of the 300°C experiment (1.19% R O, Fig. 4), these molecules were completely destroyed after 5 h at 400°C, corresponding to 1.38% R O (Fig. 5). In contrast, hopanes, cheilanthanes and gammacerane were shown to be thermally more stable (in accordance with e.g. Eglinton & Douglas Reference Eglinton and Douglas1988). In our experiment, traces were still present after 24 h at 400°C (1.67% R O), while these compounds were no longer detected after 48 h, corresponding to 1.83% R O (Fig. 5).
Fig. 5. GC–MS chromatograms (terpanes, m/z 191) of experimentally maturated Green River Shale kerogens (400°C; 2, 5, 24 and 48 h) with corresponding vitrinite reflectance (VR). Gc, Gammacerane.
It has been postulated that in the range of 1.35–2.0% R O all biomarkers >C15 are destroyed (Tissot & Welte Reference Tissot and Welte1984), which is in good accordance with our results. As this range represents the end of catagenesis (~150°C in natural systems), the preservation of biomarkers in Precambrian rocks that experienced extensive metamorphism should be the exception (see e.g. French et al. Reference French, Hallmann, Hope, Schoon, Zumberge, Hoshino, Peters, George, Love and Brocks2015). Yet, it has been shown that sedimentary rocks may still contain hydrocarbons (and biomarkers) >C15 even if they experienced burial metamorphism (i.e. >200°C, >4% R O; e.g. Price Reference Price1993; Brocks & Summons Reference Brocks, Summons and Schlesinger2003, and references therein). This implies that the preservation of biomarkers in natural systems is also influenced by further parameters such as lithology (grain size), reaction kinetics, presence of water, and fluid pressures (Price Reference Price1993). This stresses the need for further maturation experiments in which these parameters are considered.
Conclusions
Our experimental maturation performed on kerogen from the Eocene Green River Shale study confirms the applicability of biomarker-based indices and ratios as maturity indicators (e.g. total cheilanthanes/hopanes ratio; sterane and hopane isomerization indices). However, the data also demonstrate that several biomarker ratios that are commonly used for paleobiological and environmental interpretations (e.g. pristane/phytane, pristane/n-C17, phytane/n-C18 and total steranes/hopanes) are profoundly biased by differences in the thermal degradation behaviour and therefore have to be used only upon consideration of the thermal history of the sample. In particular, a strong relative increase of pristane versus phytane during the early maturation stages indicates different sources and mechanisms of formation for these compounds and gives rise to major shifts in the pristane/n-C17 and phytane/n-C18 ratios. Likewise, the ratio of eukaryote-derived steranes versus bacterial-derived hopanes strongly decreases in the course of maturation, indicating that steranes are considerably more susceptible to thermal degradation than hopanes. On the basis of the presence or relative abundance of steranes (total steranes/hopanes ratio), organic matter contributions of eukaryotes to ancient sediments may therefore be underestimated. On the other hand, bacterial hopanes and gammacerane (derived from anoxygenic bacteria and/or bacterivorous ciliates) showed very similar stability, so that the gammacerane index remained virtually unaffected by thermal maturation. In our experiment, biomarkers >C15 (especially steranes and hopanes) and ‘biological’ chain length preferences for n-alkanes were completely vanished at 1.83% R O. However, a systematic shift was observed between measured vitrinite reflectances and molecular maturity indicators, resulting in R O values consistently higher than expected for a given biomarker ratio. Summarising, our study nevertheless underlines the value of maturation experiments for assessing the influence of thermal stress on individual biomarkers, or biomarker ratios. More such experiments have to be performed to improve our ability of assessing the impact of maturation on organic matter in ancient rocks and, potentially, extraterrestrial materials. Future studies should also employ different P/T regimes and matrix minerals for a better estimation of their impact on the preservation of individual biomarkers.
Acknowledgements
We thank Martin Blumenberg (BGR Hannover) and Jörn Peckmann (University of Hamburg) for their thoughtful comments that helped to improve the earlier version of the manuscript. C. Conradt and A. Wittenborn are thanked for technical and analytical support. We also acknowledge the Göttingen Academy of Sciences and Humanities and the Department of Geobiology for organizing the symposium ‘Dating the Origin of Life: Present-Day Molecules and First Fossil Record’. This work was financially supported by the International Max Planck Research School for Solar System Science at the University of Göttingen, and the German Research Foundation (DFG grants DU373/8-1 and DU1450/3-1, DFG priority program 1833 ‘Building a Habitable Earth’). This is publication number 2 of the Early Life Working Group (Department of Geobiology, University of Göttingen; Göttingen Academy of Sciences and Humanities).
