Abiotic oxidation of petroleum bitumens under natural conditions

Organic Geochemistry 31 (2000) 977±1003

www.elsevier.nl/locate/orggeochem

Abiotic oxidation of petroleum bitumens under natural conditions A. CharrieÂ-Duhaut a, S. Lemoine a, P. Adam a, J. Connan b, P. Albrecht a,* a

Laboratoire de GeÂochimie Organique, UMR 7509 du CNRS, Institut de Chimie, Universite Louis Pasteur, 1 rue Blaise Pascal, 67000 Strasbourg, France b Centre de Recherches, Elf Exploration Production, 64000 Pau, France Received 26 July 1999; accepted 20 July 2000 (returned to author for revision 26 November 1999)

Abstract Five series of crude oil samples exposed to atmospheric conditions have been analysed at the molecular level, each series comprising several samples originating from the same crude oil but altered to di€erent extents. The aim of our investigation was to compare the speci®c impact of abiotic oxidation to other alteration processes such as biodegradation, evaporation and water washing. Bulk analyses revealed that increasing alteration is accompanied by an increase in oxygen content which parallels a relative increase of the proportions, as well as of the molecular weights of the macromolecular constituents of the bitumens. Gas chromatographic±mass spectrometric analyses of polar fractions showed the presence of oxygencontaining compounds (steroid ketones, benzothiophenic acids and sulfones) which result from oxidation of petroleum lipids. The hypothesis that part of these oxygenated compounds results from abiotic oxidation processes rather than from biodegradation is supported, notably, by the fact that oxygen incorporation generally occurred without any diastereomeric discrimination. This is also supported by simulation experiments performed on petroleum lipids, which showed that abiotic oxidation induces cleavage reactions a€ecting C±C and C±S bonds which may intervene in the transformation of geomacromolecules in the environment by degradation (``depolymerization''). Thus abiotic oxidation may play a major role in the fate of petroleum pollutants in the environment by transforming lipidic organic matter from petroleum into more water soluble and, therefore, more biodegradable constituents. However, these can be more toxic to the environment as the water-soluble fraction may be easily taken up by biota. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Abiotic oxidation; Biodegradation; Water washing; Geomacromolecules; Simulation experiments; Bitumen; Biomarkers

1. Introduction Bituminous sands, asphalts, oil spillages and surface reservoirs represent enormous amounts of generally unexploitable crude oils which are exposed to the atmosphere and the fate of which is still the subject of many investigations. Indeed, when surface exposed, the organic matter undergoes various alteration processes,

* Corresponding author. Tel.: +33-3-88-41-68-41; fax: +33-3-88-61-00-04. E-mail address: [email protected] strasbg.fr (P. Albrecht).

such as water washing, evaporation, biodegradation and abiotic oxidation. These physico-chemical processes can be induced by various external factors such as atmospheric oxygen, oxygen dissolved in rainwaters/percolating waters or by the action of aerobic and anaerobic micro-organisms. They can also be initiated by light, trace oxidants in the atmosphere (ozone, peroxides, e.g. Cooper and Zika, 1983; Cooper et al., 1989; Aneja et al., 1994) or by the chemical composition of the environment. In recent years, much research has been dedicated to the study of such natural processes, particularly those following oil spillages (Gundlach et al., 1983; Wang et al., 1994; Bence et al., 1996) and releases of material in

0146-6380/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0146-6380(00)00109-1

978

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

relation to o€shore petroleum activities (e.g. Fisher et al., 1998). For example, weathering processes such as evaporation and, to a smaller extent, water washing were shown to lead to a dramatic depletion of the lowmolecular-weight constituents of bitumens (4C15) (e.g. Gundlach et al., 1983, Volkman et al., 1984; Payne et al., 1987; Kagi et al., 1988; MacKay and McAuli€e, 1988; Kuo, 1994 and references therein). Biodegradation (Seifert and Moldowan, 1979; Goodwin et al., 1983; Wardroper et al., 1984; Volkman et al., 1984; Clayton and King, 1987; Chosson et al., 1992; Peters and Moldowan, 1993; Volkman et al., 1994; Fisher et al., 1998) has also received much interest, in particular within natural bioremediation, and enhanced bioremediation processes in terrestrial and stranded material (Volkman et al., 1984). Yet, little attention has been paid to abiotic oxidation as an independant phenomenon, although oxidation itself is a widespread process. Most studies in geochemistry have been restricted to the fate of lowmolecular-weight constituents of petroleum using simulation experiments (e.g. Ehrhardt and Petrick; 1984, 1985; Giuliano et al., 1997). It is now common knowledge (e.g. Hostettler and Kvenvolden, 1994) that exposure of crude oils to atmospheric conditions results in dramatic changes in its chemical composition, e.g. modi®cation of the colour and consistency of the bitumen and even incorporation of oxygen. However, little is known, at the molecular level, on the chemical impact of oxidation on fossil organic matter exposed to the atmosphere. Several questions, therefore, still remain regarding which compounds are preferentially a€ected by oxidation and how oxygen is incorporated into petroleum lipids. Furthermore, is it possible to distinguish between biodegradation and abiotic oxidation processes? Are there parameters able to measure the extent of oxidation? What is the impact of abiotic oxidation on macromolecules which represent, by far, the major part of petroleum bitumens exposed to the atmosphere? From an environmental point of view, does abiotic oxidation contribute to the further degradation of lipidic organic matter of petroleum refractory to biodegradation and, hence, play an important role on the fate of petroleum pollutants in the environment? We report here a detailed study of ®ve series of samples, each series comprising several samples originating from the same crude oil, generally biodegraded, but altered to di€erent extents. The aim of this investigation is to unravel the speci®c impact of abiotic oxidation and to ®nd possible ``molecular signatures'' which are typical of abiotic oxidation but di€erent from biodegradation. For this purpose, bulk measurements (elemental analysis, IR spectroscopy and size exclusion chromatography) in combination with GC±MS analyses, have been performed to observe the changes in organic matter composition as a result of oxidative alteration. Most of the work focusses on the GC±MS analysis of polar fractions

that may contain oxidized compounds such as ketones, alcohols, acids and sulfones formed by abiotic oxidation. In parallel, simulation experiments of abiotic oxidation in the presence of light have been performed on certain reference compounds, in particular, aromatic biomarkers, in order to observe their behaviour under oxidative conditions and to identify the oxidation products formed. 2. Experimental 2.1. Description of samples Five series of biodegraded samples of di€erent origins (Western Europe and Middle East), exposed to natural atmospheric conditions for di€erent lengths of time, were investigated. Their origin, description and apparent degree of alteration are summarized in Table 1. Several analyses (bulk and molecular) were used to assess the severity of the alteration undergone by the samples investigated. All samples were stored at ÿ20 C prior to analysis. Liquid samples were stored in argon ¯ushed bottles. Solid samples were wrapped in aluminium foil and stored in polyethylene bags. Samples were analysed immediately after extraction. 2.2. Instrumentation 2.2.1. Gas chromatography (GC) GC was carried out either on a Carlo Erba 4160 or on a Fisons 8160 gas chromatograph equipped with an oncolumn injector, a FID (300 C) and a J&W DB-5 fused silica column (30 m0.25 mm i.d.,0.1 mm ®lm thickness). Hydrogen was used as carrier gas. Temperature programs: condition a: 40!100 C (10 C/min), 100! 300 C (4 C/min), isothermal 300 C. Conditions b: 70!100 C (10 C/min), 100! 300 C (4 C/min), isothermal 300 C. Conditions c: 80 C (1min), 80! 100 C (10 C/min), 100! 300 C (4 C/min), isothermal 300 C. 2.2.2. Gas chromatography±mass spectrometry (GC± MS) GC±MS analyses were carried out either on a Finnigan INCOS 50 quadrupole mass spectrometer connected to a Varian 3400 gas chromatograph equipped, as above (oven temperature: conditions a or c) or on a triple quadrupole Finnigan TSQ 700 spectrometer connected to a Varian 3400 gas chromatograph (on-column injector, J&W DB-5 column, 60 m0.25 mm i.d., 0.1 mm ®lm thickness or J&W DB-17, 60 m0.25 mm i.d., 0.1 mm ®lm thickness. Temperature programs : Conditions d, 40 C (1min), 40!100 C (10 C/min), 100! 300 C (3 C/min), isothermal 300 C. Mass spectra were produced at 70 eV, source 150 C, in full detection mode over 40±800 amu (cycle time 1.1 s). In both cases, helium was used as carrier gas.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

979

Table 1 Description of the biodegraded samples and evaluation of the severity of the alteration undergone by the samples exposed to atmospheric conditions Samples Aquitaine basin Gaujacq 3

Origin

France

Bastennes 732 Maestu Maestu 830 Maestu 794 Maestu 829

Description of samples

Colour /consistency

+

Slightly biodegraded crude oil Crude oil impregnated soil

Black, liquid, sticky

+++

Spain

Hit Hit 143 Hit 142

Severity of alterationa

+ ++ ++ +

Iraq

++

Mari 90

++

Hit 141

+++

Gokhurt Gokhurt 330 Gokhurt 326

Pakistan

Mehrgahr 334

Djebel Bichri Bichri 425 Bichri 430 a

+ ++ +++

Syria

+ ++++

Black, hard

Bitumen droplets in carbonates Oil-stained carbonates Oil-stained carbonates

Black, friable

Biodegraded crude oil:collected in a warm water spring Crude oil impregnated soil (surroundings) Archaeological bitumen used as a basket caulk Crude oil impregnated soil (surroundings)

Black, sticky

Biodegraded crude oil : collected in a water spring where it seeps Crude oil impregnated soil (surroundings): formerly a similar water spring that dried out Archaeological bitumen (5000 years of exposure to atmosphere) used as a mortar for construction. Close to Gokhurt. Bituminous sand (biodegraded crude oil). Collected a few meters below surface Upper surface crust of the same bituminous sand

Brown, friable Brown, friable

Dry, hard, black Black, friable Brown, friable Black, sticky Brown, friable Brown, friable

Black, very sticky Dry, brown, pulverulent

+:Slightly, ++: moderately; +++ severely; ++++ extremely altered samples.

Coinjection experiments with reference compounds were performed on two di€erent GC columns: J&W DB-5 (30 or 60 m0.25 mm i.d., 0.1 mm ®lm thickness) and J&W DB-17 (60 m0.25 mm i.d., 0.1 mm ®lm thickness). 2.2.3. Nuclear magnetic resonance spectroscopy NMR experiments were performed on either a BRUKER WP-200 SY (200 MHz operating frequency) spectrometer or on a high resolution BRUKER ARX 500 (500 MHz) spectrometer. Chemical shifts d are reported in ppm vs tetramethylsilane using the solvent CDCl3 (d1H 7.26 ppm) as internal reference. 2.2.4. Infrared Spectroscopy Infrared spectra were obtained on a BRUKER IFS 25 spectrometer.

2.2.5. Size exclusion chromatography Size exclusion chromatographic analyses were performed at room temperature using tetrahydrofuran (THF) as eluent on a Hewlett Packard Series 1050 chromatograph equipped with an UV spectrophoto meter, an ICS 8110 di€erential refractometer and two mStyragel columns (Waters Associates) containing beads of styrene-divinylbenzene copolymer (103 and 104 AÊ respectively). Calibration was obtained with a series of linear polystyrenes with Mw values ranging from 480 to 675 000 and polydispersion indexes between 1.05 and 1.07. Calibration curves giving retention times as a function of molecular weight of polystyrene were made using 0.05% solutions. The UV-detector was set at 254 nm. Analyses were performed on 1 wt.% solutions of the samples dissolved in THF. All samples were ®ltered on an acrodisc 13CR PTFE 0.45 mm and injected 15 min

980

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

after the solutions were prepared. The eluent (THF) ¯ow was set at 2 ml/min. Absorptions were measured at 350 nm. 2.3. Extraction and fractionation Samples were extracted and fractionated as shown in Fig. 1. Acids were separated following the method described by McCarthie and Duthie (1962). Asphaltenes were precipitated as described by Speight (1984). Samples containing water (from a spring) were dried by azeotropic distillation with toluene prior to separation. 2.3.1. Extraction and precipitation of ``humic'' substances (adapted from Parsons, 1988) Typically, the crushed sample ( 20 g) was ultrasonically extracted (3) with an aqueous sodium

hydroxide solution (0.1 M, 100 ml, 30 min) and the aqueous extracts recovered by centrifugation. The three extracts were combined and concentrated to 100 ml. This solution was then acidi®ed to pH 1 with HCl (6 M). The precipitate (``humic`` substances) formed after 18 h was centrifuged and the supernatant removed. The recovered ``humic'' substances were dried in a desiccator and weighed prior to IR spectroscopy. 2.4. Reduction of carbonyls to methylene groups 2.4.1. Wol€-Kishner reduction of ketone fractions (Huang-Minlon, 1946) The ketone fraction, dissolved in dichloromethane (250±500 ml), was added to a mixture of diethylene glycol (6 ml), n-butanol (4 ml) and hydrazine hydrate

Fig. 1. Extraction and fractionation scheme.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

(1.5 ml) and was re¯uxed under argon for 5 h. Excess hydrazine, n-butanol and water was removed by distillation. After cooling to room temperature, potassium hydroxide (850 mg) was added and the mixture again heated (200 C, 5 h). After dilution with distilled water and extraction with diethylether, saturated and aromatic hydrocarbons were separated from the crude mixture by TLC as shown in Fig. 1. 2.4.2. Reduction of ketone fractions with Pd/C (after Novak et al., 1954) Typically, the ketone fraction dissolved in ethylacetate (1 ml) and acetic acid (5 ml) was hydrogenated in a hydrogen atmosphere using Pd/C 10% (40 mg). The crude mixture, obtained after ®ltration over celite, was fractionated by TLC (SiO2/DCM) to yield the hydrocarbons (Rf: 0.85-1), unreacted ketones (Rf: 0.10-0.85), and polars (Rf: 0±0.10). 2.4.3. Deuterated Clemmensen reduction of ketone fractions (after Enzell, 1966) Typically, D2O (1 ml) and trimethylsilyl chloride (1 ml) were added to a solution of a ketone fraction in anhydrous toluene (1 ml). Zinc powder (20 mg) was added and the mixture was stirred under argon for 2 h (0 C). After extraction, saturated and aromatic hydrocarbons were fractionated by TLC (SiO2/hexane) as shown in Fig. 1. 2.5. Simulation experiments All simulation experiments were carried out in quartz glassware. 2.5.1. Oxidation of triaromatic steroid 1a. Water (<1 ml) was added to triaromatic steroid 1a (8 mg) in hexane (10 ml) and exposed to sunlight (summer season) under oxygen atmosphere for 14 days. Combined DCM and diethylether extracts had solvent removed under reduced pressure before analysis by GC and GC±MS. 2.5.2. Oxidation of alkylbenzothiophenes Water (500 ml) was added to 2-methylbenzothiophene 2 (8 mg) (also 4-methylbenzothiophene 3, 9 mg; 2-butyl-4-hexylbenzothiophene 4, 2 mg; synthesized by Perakis, 1986) in hexane (5 ml): the mixture was exposed to air and to lamp light (OSRAM HQI TS 250/D), the spectrum of which is very close to that of sunlight. After 48 h (respectively. 72 h, 64 h) exposure, solvents were removed under reduced pressure and the mixture was analyzed by GC and GC±MS. Identi®cation of oxidized species, in particular the acids was made by GC and GC±MS analysis after esteri®cation of the crude reaction mixture by diazomethane.

981

2.6. Synthesis of reference compounds 2.6.1. Deuterated C27 triaromatic steroid 5 and C27 triaromatic steroid ketone 6 Deuterated reference compounds were synthesized following the procedure described by Lemoine (1996). 2.6.2. 2-carboxybenzothiophenes 7 Butyllithium in hexane (0.6 M, 1.9 ml, 1.11 mmol) was added, under argon at ÿ78 C, to a solution of benzothiophene (124 mg, 0.93 mmol) in anhydrous tetrahydrofuran (5 ml). After 15 min, carbon dioxide was bubbled through the solution (30 min). The reaction mixture was then poured in water, acidi®ed to pH 1 with aqueous HCl and extracted with diethylether (3). After removal of the solvent under reduced pressure, the crude mixture was chromatographed on a silica gel column with CH2Cl2/AcOEt as eluent. 2-Carboxybenzothiophene 7 (53mg) was obtained in 32% yield. The corresponding methylester, necessary for the coinjection experiments, was obtained by treating 7 with etheral diazomethane. 2-carbomethoxybenzothiophene: TLC: Rf=0.5 (DCM); MS (IE, 70 eV), m/z (rel. int.): 192 [M+] (67%), 162 (10), 161 (100), 133 (18), 89 (18);1H NMR (CD2Cl2, 200 Mhz): 3.92 (3H, s, H-2), 7.43 (2H, m, H-5 and H-6), 7.89 (2H, m, H-4 or H-7), 8.05 (1H, s, H-3). 2-methylbenzothiophene 2 and 4-methylbenzothiophene 3 were synthesized following the procedure described by Perakis (1986). Triaromatic steroid ketones 8 (a and b) and 9 (a and b) were synthesized following the procedure described by Schae€er et al. (1993). 3. Results and discussion 3.1. Bulk analyses 3.1.1. Elemental analysis An indication of the strong structural modi®cations undergone by fossil organic matter exposed to the atmosphere is given by its elemental composition. As illustrated in Table 2 oxidative degradation is characterized by a signi®cant incorporation of oxygen increasing, for instance, from 1.5% in the less degraded samples to 8.7% in the most degraded samples of the Gokhurt series. Hydrogen, sulfur and nitrogen values showed few changes. Interestingly, the O/H ratio increases in the more severely degraded samples indicating that hydrogen has been replaced by oxygen. A slight decrease in the H/C ratio is observed and may be explained either by oxidation of carbon (leading to the removal of H atoms) or by a higher aromaticity. The S/C ratio generally decreases with increasing alteration.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

982

Table 2 Elemental composition of the total organic extracts (at.%) and H/C, O/C O/H and S/C ratios Samples

C (%)

Gaujacq 3 Bastennes 732 Maestu 830 Maestu 794 Maestu 829 Hit 143 Hit 142 Mari 90 Hit 141 Gokhurt 330 Gokhurt 326 Mehrgahr 334 Bichri 425 Bichri 430

37.8 41.4 41.7 42.2 42.8 40.6 40.3 41.8 40.2 40.0 39.4 38.9 40.5 39.6

H (%) 61.2 54.8 56.7 54.3 52.8 56.7 52.6 52.8 49.5 57.1 54.8 51.9 57.2 46.5

S (%) 0.58 1.72 0.86 1.20 1.18 2.22 2.23 1.55 1.96 1.08 1.01 0.95 1.34 1.50

N (%) 0.12 0.13 0.14 0.16 0.18 0.21 0.26 0.37 0.24 0.22 0.29 0.20 0.34 0.57

3.1.2. Composition of the samples Among criteria measuring the severity of oxidative degradation, the relative yields of the extractable fractions (saturates, aromatic hydrocarbons, macromolecular asphaltenes and resins) reveal interesting features (Table 3). Indeed, in all samples investigated, aromatic and saturated hydrocarbon fractions have shown low yields in the most altered samples. For example, 22% of the organic matter in the less degraded sample is made up of aromatic hydrocarbons whereas only 1% of the organic matter in the most degraded sample is composed of aromatics in the Gokhurt series. Similarly, the proportion of saturated hydrocarbons is lower (1±3%) in the most degraded samples. The preferential Table 3 Relative yields of the extractable fractions and of the asphaltenes isolated from the total organic extracts Samples

Saturates Aromatics Polars Asphaltenes (wt.%) (wt.%) (res.+asph.) (wt.%) (wt.%)

Gaujacq 3 37 Bastennes 732 7 Maestu 830 10 Maestu 794 3 Maestu 829 3 Hit 143 6 Hit 142 3 Mari 90 1 Hit 141 1 Gokhurt 330 10 Gokhurt 326 5 Mehrgahr 334 3 Bichri 425 8 Bichri 430 1 a

nd=Not determined.

33 15 10 1 1 22 15 1 2 20 4 <1 19 <1

30 78 80 96 96 72 82 98 97 70 91 97 73 99

nda 33 29 41 68 19 49 63 83 41 63 84 37 95

O (%) 0.23 1.97 0.60 2.14 2.12 0.27 4.54 3.44 8.11 1.50 4.54 8.67 0.63 11.9

H/C 1.62 1.35 1.36 1.29 1.23 1.39 1.30 1.26 1.26 1.43 1.39 1.32 1.41 1.17

O/C

O/H ÿ3

6.1E 5.0Eÿ2 1.4Eÿ2 5.1Eÿ2 4.9Eÿ2 6.6Eÿ3 11.3Eÿ2 8.2Eÿ2 20.1Eÿ2 3.8Eÿ2 11.5Eÿ2 22.3Eÿ2 1.6 Eÿ3 29.9 Eÿ3

S/C ÿ3

4.7E 9.5Eÿ3 1.1Eÿ2 3.9Eÿ2 4.0Eÿ2 4.8Eÿ3 8.6Eÿ2 6.5Eÿ2 16.4Eÿ2 2.6Eÿ2 8.3Eÿ2 16.9Eÿ2 1.1Eÿ2 25.4Eÿ2

1.5Eÿ2 4.2Eÿ2 2.1Eÿ2 2.8Eÿ2 2.8Eÿ2 5.5Eÿ2 5.5Eÿ2 3.7Eÿ2 4.9Eÿ2 2.7Eÿ2 2.6Eÿ2 2.4Eÿ2 3.3Eÿ2 3.8Eÿ2

removal of aromatic hydrocarbons compared to saturated hydrocarbons has been observed previously in weathered shales (Leythaeuser, 1973; Clayton and Swetland, 1978; Forsberg and Bjorùy, 1983) and severely altered samples from oil spills (e.g. Wang et al., 1994). Contrastingly, yields of asphaltenes in weathered samples increase relative to low-molecular-weight fractions with values reaching 90% (Bichri 430; Table 3) (cf. Connan et al., 1980; Wang et al., 1994). The formation of organic material, insoluble in organic solvents, is also observed in the most oxidized samples. The physical appearance of the samples is further proof that they have been a€ected by atmospheric exposure. At relatively early stages of degradation (although already biodegraded and water washed) the bitumens still appear black and viscous while at more advanced stages of alteration, they tend to appear brown and friable (Table 1) as observed by Hostettler and Kvenvolden (1994) on some aged oil spills. Samples that have been moderately altered (such as Hit 142) are very hard and black. This might be related to an increase in the relative amount of macromolecular material. An increase in the oxygen content of macromolecular fractions is revealed by IR spectroscopy (see below and Fig. 2). Thus, fossil organic matter becomes progressively more hydrophilic which may be one reason why part of it becomes insoluble in organic solvents. On the other hand, a substantial part of the more altered samples is more soluble in alkaline (0.1 M NaOH) aqueous solutions. This soluble material can be reprecipitated (and isolated) by lowering the pH. In the case of the Bichri samples, the proportion of organic material extractable by alkaline aqueous solutions is higher in the most altered samples. As much as 90% can be extracted by aqueous alkali. In the less altered samples, the amount of substances extractable with aqueous alkali is about 2% relative to the organic extracts. For the samples from

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

983

carefully dried water may interfere at 3400±3450 cmÿ1). Ketones (CˆO, 1650±1730 cmÿ1), aldehydes, acids and esters (1600±1650 cmÿ1), sulfones (1110 cmÿ1), sulfoxides (1030 cmÿ1) and possibly ethers or sulfonates (1000±1300 cmÿ1) could also be detected. These bands become more intense, even predominant, for samples more severely a€ected by degradation. Similar qualitative spectra, and evolutions, were recorded for the total extracts. The ``humic'' material isolated from Bichri 430 was also investigated by IR spectroscopy (not shown). A predominant band in the 3650±3100 cmÿ1 range was observed suggesting OH groups of phenols, alcohols and aliphatic or aromatic mono- or polycarboxylic acids. A band in the 1600±1730 cmÿ1 range also suggests the presence of aldehydes, ketones and mono- or polycarboxylic acids. 3.1.4. XANES spectroscopy The XANES spectroscopy experiments carried out on the bitumen from Hit, Gokhurt, Bastennes, Maestu and Bichri series con®rmed the trend observed for the organic sulfur compounds. The relative sulfur content of the samples (thiophenic, alkyl or aryl sulfur, Brown et al., 1992) estimated by the intensity of the S L-edge signals, decreased with increasing alteration. On the other hand, the most degraded samples contained oxidized forms of sulfur: sulfoxides, sulfones and probably sulfonates and sulfates were detected (S K-edge signals; Sarret et al., 1999). Fig. 2. IR spectra of the asphaltenes isolated from three samples of the series of Maestu. (a) Maestu 830; (b) Maestu 794; (c) Maestu 829.

the Maestu series, the yields are even lower. Thus, incorporation of oxygen enables transformation of the organic matter from petroleum into humic-like substances (e.g. Parsons, 1988) favouring their progressive solubilization and transformation. 3.1.3. IR Spectroscopy IR spectroscopic analyses of the asphaltenes reveal the type of oxygen functionalities arising from degradation (Speight, 1984). Similar trends were observed for all our samples as illustrated for the Maestu series (Fig. 2). In the less degraded bitumens (Fig. 2a), the aliphatic (2920±2855 and 1455±1375 cmÿ1) and aromatic (1600± 1650 and 700±930 cmÿ1) signals prevail whereas absorptions corresponding to oxygenated functions are low. At a higher level of alteration (Fig. 2b and c), new IR bands appear such as hydroxyls and carboxyls (large O± H bands at 3650±3100 cmÿ1) suggesting the presence of constituents such as phenols, primary/secondary/tertiary alcohols, aliphatic or aromatic mono-or polycarboxylic acids (although the samples have been

3.1.5. Carbon and hydrogen isotopic compositions All the samples from the same series, independent of the severity of alteration, have almost identical carbon isotopic compositions: deviations of less than 1% are observed except in the case of Mehrgahr (Table 4). Therefore, d13C can be used as a tool for establishing correlations between severely degraded bitumens (Connan et al., 1990 and 1992a,b; Connan, 1999). However, a slight enrichment in d13C is observed in one severely altered sample (e.g. Mehrgahr), a feature which was observed previously by Leythaeuser (1973) in the case of weathered shales. In contrast to the d13C values, the D values are strongly a€ected by the alteration processes. Thus, for the Hit series, for instance (Table 4), the D values of the asphaltenes increase from ÿ92% in the less altered sample to ÿ68% in the more altered ones. The increase in deuterium content with increasing alteration can be explained by exchange reactions with rainwater or percolating waters (less negative D values; DeNiro and Epstein, 1981; da Silveira Lobo Sternberg, 1988). Thus, incorporation of oxygen (Table 2) induced by atmospheric alteration, in particular as hydroxyls or carboxyls or as ketones as shown by IR spectroscopy favours such a process. Indeed, H/D atoms from hydroxyl groups or H/D atoms located at an enol

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

984

Table 4 Carbon and deuterium isotopic compositions of the total organic extracts from the samples of the Maestu, Hit and Gokhurt series Samples

d13C (asphaltenes) % PDB

dD (asphaltenes) % SMOW

Gaujacq 3 Bastennes 732 Maestu 830 Maestu 794 Maestu 829 Hit 143 Hit 142 Mari 90 Hit 141 Gokhurt 330 Gokhurt 326 Mehrgahr 334 Bichri 425 Bichri 430

ÿ23.3 ÿ24.0 ÿ26.0 ± ÿ26.0 ÿ28.3 ÿ28.3 ÿ28.0 ÿ28.3 ÿ21.8 ÿ22.5 ÿ20.4 ÿ28.8 ÿ28.7

ÿ89 ÿ76 ÿ99 ± ÿ81 ÿ92 ÿ79 ÿ69 ÿ68 ÿ102 ÿ84 ÿ73 ÿ93 ÿ50

position can undergo rapid exchange with water contrary to those located on aliphatic or aromatic skeletons. Variations in D values which parallel an increase in oxygen content can therefore be used to measure the degree of alteration of the bitumens exposed to atmospheric conditions. 3.1.6. Size exclusion chromatography Molecular weight distributions of organic extracts, and asphaltenes, at various degrees of alteration have been compared by size exclusion chromatography (SEC) for all samples. A calibration of the molecular weight scale was compiled (see Experimental, and Bynum et al., 1970; Rosset et al., 1994). Since retention times on SEC columns depend not only on molecular weight but also on molecular size and shape, quantitative data, in terms of molecular weight, cannot be obtained by this method (Speight et al., 1985). However, within a series of samples, the relative molecular weights can be compared (Calemma et al., 1995). An increase in the apparent molecular weights with degree of alteration appears in the case of the organic extracts (illustrated for the Maestu series in Fig. 3) and for the asphaltenes in all samples investigated except the Gokhurt series. The increase in the apparent molecular weight of the most oxidized samples can be interpreted as being, partly at least, the result of the formation of aggregates. In fact, this can be deduced from the molecular weight distributions which becomes polymodal (Fig. 3b and c) in the more altered samples (Snyder, 1969). Formation of aggregates was further con®rmed by an increase of high molecular weight populations accompanied by the appearance of a trimodal distribution when time between solubilisation of the asphaltenes and

their injection on the SEC column was increased from 15 min to 105 min (Speight et al. 1985). In accordance with this observation and interpretation, dilution of the samples, which renders the formation of aggregates less likely, led to a disappearance of the polymodal distributions which became monomodal and shifted towards a lower molecular weight range. This phenomenon of intermolecular association, leading to polymodal distributions and to apparent higher molecular weights, is particularly clear for the most altered samples and can be explained by a change in chemical composition. In the altered samples, intermolecular association may be favoured by the higher abundance of oxygen-containing functionalities (e.g. formation of hydrogen bonds). A similar trend was observed for asphaltenes. In conclusion, the substantial increase in molecular weight of the most altered samples may be due, partly, to a concentration of the high molecular weight fractions, since water washing was able to remove the lower molecular weight constituents of the organic extract. Moreover, the formation of aggregates may also partly explain the apparent molecular weight increase. However, it may be that part of the increase is due to crosslinking induced by oxidation. These results may partly explain the alteration of the physical aspect of the most altered samples (e.g. hardening) and the formation of solvent-insoluble material. 3.2. Molecular composition Saturated and aromatic hydrocarbons were investigated by GC±MS in order to identify the structures more particularly a€ected by exposure to the atmosphere and to evaluate and distinguish, if possible, the speci®c impact of the di€erent processes intervening in the alteration (water washing, biodegradation, abiotic oxidation). Emphasis was put on the analysis of ketones, alcohols and acids which may have been formed by biodegradation or abiotic oxidation. 3.2.1. Saturated hydrocarbons For samples from the Hit series (Fig. 4), the gas chromatograms of the saturated hydrocarbons display unresolved complex mixtures (UCM). Linear and isoprenoid hydrocarbons, are absent, typical of severely biodegraded oils or of oils exposed to natural surface alteration (Gundlach et al., 1983; Volkman et al., 1984; Peters and Moldowan, 1993; Hostettler and Kvenvolden, 1994; Fisher et al., 1998). The m/z 217 fragmentograms (not shown) of steranes display 5a(H), 14b(H), 17b(H) R and S isomers (Mackenzie, 1984) indicating that these samples are derived from relatively mature oils. The distribution of the hopanes (C27±C35 ; m/z 191 ; not shown) dominated by the 17b(H), 21b(H) isomers (Ensminger et al., 1974, Van Dorsselaer et al. 1977;

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

985

Fig. 3. Size exclusion chromatograms of the total organic extracts from the samples of the Maestu series. (a) Maestu 830; (b) Maestu 794; (c) Maestu 829.

Seifert and Moldowan, 1980) is typical for relatively mature oils. Detailed analysis of the distribution of steranes also revealed changes which can be ascribed to biodegradation; for example, a relative enrichment of the diasteranes and an unusual depletion of the 5a(H), 14a(H), 17a(H) S isomers is observed in the intermediate sample. In the most altered samples of this series (Hit 141; Fig. 4c, and Mari 90, not shown), the almost complete removal of steranes may also be explained by biodegradation (Seifert and Moldowan, 1979, Seifert et al., 1984; Goodwin et al., 1983; Chosson et al., 1992). Modi®cation of the distribution of the hopanes may also be due to biodegradation. Thus, an increase in the relative proportion of Tm and Ts as well as of the higher homologues (C35) in Hit 141 are typical e€ects of biodegradation (Seifert et al., 1984; Requejo and Halpern, 1989; Moldowan and McCa€rey, 1995). Water washing, and evaporation, to which bitumens were exposed, led to the depletion of the low molecular weight components (up to C15). Similarly, a skewing of the UCM towards higher molecular weight components (typically C25±C35) probably relates to the same e€ects observed at the highest degree of alteration (Fig. 4b and c). Similar features are observed for the Bichri, Gokhurt and Bastennes samples. Hexahydrobenzohopanes 10

(C31±C35; Connan and Dessort, 1987), 8,14-secohopanes 11 (Schmitter et al., 1982; RullkoÈtter and Wendisch, 1982; Wang et al., 1990), 8,14 secohexahydrobenzohopanes 12 (Dessort and Connan, 1993) and tricyclic terpanes 13 (Seifert and Moldowan, 1979; Connan et al., 1980; Aquino Neto et al., 1982,1983) are known to be particularly resistant to alteration and can be detected, at least in traces, in almost all the samples investigated. The relative proportion of the 8,14-secohopanoids increases relative to hopanes when the degree of alteration increases (compare the m/z 123 and m/z 191 fragmentograms). For samples from the Maestu series, biodegradation is even more pronounced. Linear and isoprenoid hydrocarbons and steranes have completely disappeared in all the samples. Gas chromatograms of the saturates (Fig. 5) revealed beside the predominant UCM a small contribution of residual hopanes, hexahydrobenzo hopanes 10, 8,14-secohopanes 11 and 8,14 secohexahydrobenzohopanes 12 a high abundance of the newly identi®ed regular polycyclic polyprenic hydrocarbons 14 and 15 (up to 7 rings; Fig. 5) (Schae€er et al., 1994; Poinsot et al., 1995; Li et al., 1996; Poinsot, 1997) which seem to be especially resistant towards biodegradation. In the most altered sample, regular hopanes have almost

986

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Fig. 5. Gas chromatograms (conditions d) of the saturated hydrocarbon fractions isolated from three samples of the Maestu series. (a) Maestu 830; (b) Maestu 794; (c) Maestu 829. !: hexacyclic polyprenoid hydrocarbons; * heptacyclic polyprenoid hydrocarbons. UCM: unresolved complex mixture.

Fig. 4. Gas chromatograms (conditions b) of the saturated hydrocarbon fractions isolated from three samples of the Hit series. (a) Hit 143; (b) Hit 142; (c) Hit 141. !:hopanes; * short-chain steranes; &: 17, 21-secohopanes; UCM: unresolved complex mixture.

completely disappeared and even the distribution of the resistant polycyclic polyprenic hydrocarbons seems to be modi®ed by biodegradation as the relative proportion of the higher homologues increases. In this series of samples, the degree of alteration, as deduced from IR spectroscopy and elemental analysis (oxygen content)

for instance, does not correlate with the intensity of the water washing/evaporation undergone by the samples. Indeed, the skewing of the UCM towards higher molecular weight in the gas chromatograms of the saturates is more pronounced in the case of Maestu 794 (Fig. 5b) which has been rated as an intermediate sample (Table 1). 3.2.2. Aromatic hydrocarbons As indicated above, aromatic hydrocarbons are highly sensitive to atmospheric alteration processes as their relative proportion to the total organic extract decreases, in the samples from the Hit series, from 22% in the less altered bitumen to 2% in the most altered sample (Table 3). This high sensitivity towards alteration is also revealed by the GC and GC±MS analyses of the aromatic fractions in the samples from the Hit series

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

(Fig. 6). In the less altered sample (Fig. 6a) alkylnaphthalenes, benzothiophenes and dibenzothiophenes, as well as ring C monoaromatic 16 and 17 and triaromatic 1 and 18 steroids, can still be detected. They tend to disappear in the more altered samples (Fig. 6b and c) partly due to water washing/evaporation, as revealed by the more pronounced depletion in the low-molecularweight constituents. In these severely altered samples, the aromatic fraction appears almost exclusively as an

Fig. 6. Gas chromatograms (conditions a) of the aromatic hydrocarbon fractions isolated from three samples of the Hit series. (a) Hit 143; (b) Hit 142; (c) Hit 141; * benzohopanes. UCM:unresolved complex mixture. X: contaminations.

987

UCM. Some series of aromatic biomarkers, such as the aromatic steroids which have been reported to be particularly resistant to biodegradation (Connan, 1984), have completely disappeared in the most altered samples. Other series seem to be particularly resistant to atmospheric alteration as they could be found, at least in traces, in almost all the investigated samples, including those most altered. This is so for benzohopanes 19, the ring D monoaromatic 8,14 secohopanoids 20 (Hussler et al., 1984a,b; Dessort and Connan, 1993) which could be used to evaluate the maturity of biodegraded oils (He Wei and Lu Songnian, 1990) and 8,14secobenzohopanes 21 (Dessort and Connan, 1993). Their resistance parallels that observed for the structurally-related saturated hydrocarbons i.e. hexahydrobenzohopanes 10, 8,14-secohopanes 11, 8,14-seco benzohopanes 12. As for the saturated hydrocarbons, the low-molecular-weight constituents are removed with increasing severity of alteration. This is probably due to water washing/evaporation processes. Similar features are observed for the aromatic fractions of the Gokhurt, Bichri, Bastennes and Maestu samples. In the case of the Maestu series, however, the aromatic fractions, even in the less altered sample, almost exclusively appear as an UCM constituted by a mixture of isomeric alkyl(poly)aromatic hydrocarbons. Furthermore, the amount of aromatic compounds in this series decreases more rapidly than the saturates (Table 3) suggesting that an alteration process a€ecting the aromatic constituents of UCM is operative. 3.2.3. Ketones Gas chromatograms of the ketone fractions of all samples contain a large UCM. However, for the Gokhurt series, some individual peaks appear in the GC trace (not shown) and structures, based on the interpretation of mass spectral data, was proposed for these components. Thus, a series of hopanoid ketones, ranging from C27 to C34 was detected in the two less oxidized samples with the predominant con®gurations 17b(H), 21b(H) 22R and 17b(H), 21a(H) 22R, characteristic of immature samples and soils (Dastillung, 1976). This indicates that these compounds are not transformation products of fossil hopanoids from the bitumen which only contains hopanoids having a 17a(H), 21b(H) con®guration typical for mature oils (Ensminger et al., 1974, Van Dorsselaer et al. 1977; Seifert and Moldowan, 1980). These hopanoid ketones must, therefore, originate from recent organic contributions : this is not surprising since one sample is an impregnated soil and the other was collected in a water spring. Beside the hopanoids, a series of components characterized by the presence of small molecular ions (m/z 274+nx14) and an intense fragment at m/z 245 or 259 were detected in the Gokhurt series, displaying mass

988

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

spectra very similar to those of the triaromatic steroids (8 and 9) bearing a keto group at the benzylic position (Schae€er et al., 1993). This hypothesis was con®rmed in the case of the C20 (8a) and C26R (9a) homologues and for their 4-methylated counterparts C21 (8b) and C27R (9b) by comparison of their chromatographic behaviour (coelution on two di€erent GC columns, see experimental) and mass spectral data with those of standards synthesized according to the reaction pathway described by Schae€er et al. (1993). Detailed analysis of the ketones from the Gokhurt samples also allowed us to detect compounds derived from ring D monoaromatic 8,14-secohopanoids 22 (m/z 173; m/z 188+nx14; M+: 380+nx14, Fig. 7), ketones derived from ring C monoaromatic steroids 23 and 24 (m/z 267 or 271 and M+: 296+nx14; Schae€er et al., 1993) as well as ketones 25 derived from the newly identi®ed 14-methyldiaromatic steroids 26 (Lemoine et al., 1996) (M+:292+nx14; m/z 195; 25a Fig. 8). Interpretation of the mass spectra of the ketones was con®rmed by the analysis of the hydrocarbon fractions obtained by Wol€-Kishner reduction of the ketone fractions (HuangMinlon, 1946) which is able to quantitatively reduce

Fig. 7. Mass spectra (EI, 70 eV) of C30 ring D monoaromatic 8,14-secohopanoid 20a and of the corresponding benzylic ketone 22a.

carbonyl groups to the corresponding saturated or aromatic hydrocarbons. Indeed, this fraction of hydrocarbons contains series of ring C monoaromatic 16 and 17 and triaromatic 1 and 18 steroids, ring D monoaromatic 8,14-secohopanoids 20 and 14-methyldiaromatic steroids 26. Location of the carbonyl group at the benzylic position on the hydrocarbon skeletons of the compounds which could not be identi®ed by comparison with standards, could be demonstrated by means of selective catalytic hydrogenation of the benzylic ketones carried out with Pd/C and acetic acid (Novak et al., 1954). When two benzylic positions were available, as for ring C monoaromatic steroids 16 and 17 and monoaromatic 8,14 secohopanoids 20, it was not possible, at this stage, to determine which one was preferentially oxidized. In fact, subsequent simulation of oxidation carried out on a ring C-monoaromatic steroid (Lemoine, 1996) con®rmed that both benzylic positions could be oxidized, either sequentially or simultaneously. Aromatic ketones

Fig. 8. Mass spectra (EI, 70 eV) of C21 14-methyldiaromatic steroid 25a and of the corresponding benzylic ketone 26a.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

989

have been detected previously by Ehrhardt et al. (1989) in a Baltic sea concentrate where they were supposed to result from the oxidation of aromatics originating from fossil fuels. The distributions of each series of aromatic hydrocarbons (aromatic steroids 16 and 17, 1 and 18, ring D monoaromatic 8,14-secohopanoids 20...) obtained by Wol€-Kishner reduction of the ketone fractions from the Gokhurt samples are very similar to those of the aromatic hydrocarbons present in the less altered samples of the same series. This is exempli®ed for the ring C monoaromatic and triaromatic steroid hydrocarbons by comparison of m/z 253 (Fig. 9) and m/z 231 fragmentograms (Fig. 10). Distributions of the isomers and homologues suggests that neither diastereomeric nor signi®cant side-chain size discrimination occurred during the oxidation which led to the formation of the ketones. This also suggests that the analysis of the ketone fraction might be used as a geochemical tool for the study of maturity, or for correlation purposes, in the case of severely oxidized bitumens where the aromatic hydrocarbons have completely disappeared. The triaromatic steroids, abundant in the less altered sample from the Gokhurt series, have completely disappeared in the most oxidized sample. The correspond-

ing ketones detected in trace amounts only in the less altered sample are present in the most oxidized one (Fig. 11). Similar trends have been observed for other aromatic biomarkers such as the ring C monoaromatic steroids 16 and 17 or the ring D monoaromatic 8,14secohopanoids 20 (Fig. 12). To check if the formation of triaromatic steroid ketones can account for the removal of the triaromatic steroid hydrocarbons when the degree of alteration increases, a quantitative study of their depletion and the parallel appearance of the corresponding ketones was undertaken using two deuterated standards to achieve better selectivity (Dahl et al., 1985). Thus, solutions of deuterated 4-methyl triaromatic steroid 5 and deuterated 4-methyl triaromatic steroid ketone 6 were used as internal standards and added respectively to the aromatic and ketone fractions of each sample. The amount of the predominant C28S 18b homologue in the samples (m/z 231; Fig. 11) was calculated from peak areas in typical GC-MS fragmentograms (m/z 231 and m/z 247 for the standard). Similarly, the amount of the C28S ketone 9c (m/z 245; Fig. 11) could be deduced by comparison with the standard ketone (m/z 261). Three separate injections were performed. To make sure that the results were not biased

Fig. 9. Mass fragmentograms (m/z 253; conditions c) showing the distribution of the ring C monoaromatic steroids (a) present in Gokhurt 330 and (b) formed upon Wol€-Kishner reduction of the ketones from Gokhurt 326.

Fig. 10. Mass fragmentograms (m/z 231; conditions c) showing the distribution of the triaromatic steroids (a) present in Gokhurt 330 and (b) formed upon Wol€-Kishner reduction of the ketones from Mehrgahr.

990

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Fig. 11. Mass fragmentograms (m/z 231; conditions c) showing the distribution of the triaromatic steroids (a) present in Gokhurt 330 and (b) in Mehrgahr. Mass fragmentograms m/z 245 showing the distribution of the triaromatic steroid benzylic ketones (c) present in Gokhurt 330 and (d) in Mehrgahr.

by the presence of contaminating organic matter (``dilution'') or by the e€ect of evaporation or water washing (``concentration''), the amount of steroid, either as hydrocarbon or ketone, was expressed relative to the amount of C30 hopane in each sample which is known to be highly resistant to degradation (Prince et al., 1994) and was quanti®ed using a deuterated homologue. It appears that the triaromatic steroid hydrocarbons progressively disappear as the degree of alteration increases (Table 5): at the same time, the amounts of triaromatic steroid ketone increase. Only 10% of the triaromatic ketone, relative to the related hydrocarbon initially present in the less altered sample, could be recovered in the more altered sample. Therefore, we infer that the triaromatic ketones are further transformed into more polar compounds which were not recognized, or which have been removed by water washing, or biodegradation, because of their enhanced solubility in water. In summary, the results point to an abiotic origin for the benzylic ketones formed during exposure to the atmosphere. Notably, the formation of steroid triaromatic ketones which occurred without diastereomeric or side-chain discrimination (Fig. 10), not expected from a biological process, clearly supports this hypothesis.

Indeed, biological oxidation would favour loss of long chain aromatic steroids and primarily a€ect the aromatic steroids with a 20R con®guration (Wardroper et al., 1984). Although biological oxidation at benzylic positions has been reported (e.g. Van der Linden and Thijsse, 1965), benzylic positions are particularly sensitive to autoxidation. Indeed, benzylic ketones are formed as major transformation products when alkylbenzenes are exposed to light in the presence of a photosensitizer (Ehrhardt and Petrick, 1984). Also, we have exposed the C26 triaromatic steroid hydrocarbon 18a to air and sunlight (without sensitizer) which led to the predominant formation of the corresponding benzylic ketone 9a (60% yield) detected in the altered sample from the Gokhurt series. Consequently, as benzylic ketones most probably have an abiotic origin, we propose to use the ratio of triaromatic steroid/(triaromatic steroid+triaromatic steroid ketone, Table 5) as a parameter for the extent of abiotic oxidation undergone by bitumens. It varies from about 1 (unoxidized) down to 0 (most oxidized) (Table 5). The ketone fractions of the Bichri samples were shown to contain the same type of oxidation products : triaromatic steroidic ketones 8 and 9 were detected in the most oxidized sample, whereas they are only present

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

991

Fig. 12. Mass fragmentograms (m/z 365; conditions c) showing the distribution of the ring D monoaromatic 8,14-secohopanoids (a) present in Gokhurt 330 and (b) in Mehrgahr. Mass fragmentograms (m/z 173; conditions c) showing the distribution of the ring D monoaromatic 8,14-secohopanoids benzylic ketones (c) present in Gokhurt 330 and (d) in Mehrgahr. Table 5 Amounts of C28 S 18b steroid triaromatic hydrocarbon and of the corresponding benzylic ketone 9c expressed versus the amount of C30 17a(H), 21b(H) hopane in the samples from the Gohkurt series Samples

Amount of C28S

Amount of C28S ket

Amount of C30ab hopane

C28S /C30ab hopane

C28S ket/C30 ab hopane

C28S/(C28S+C28S ket)

Gokhurt 330 Gokhurt 326 Mehrgahr 334

109 ppm 8 ppm 0 ppm

0.8 ppm 7 ppm 6 ppm

136 ppm 105 ppm 79 ppm

0.80 0.08 0.00

0.006 0.067 0.076

0.99 0.53 0.00

in trace amounts in the non-altered sample. Again, their appearance is accompanied by their depletion as aromatic hydrocarbons in the aromatic fractions of the degraded sample. The same applies to ring D-monoaromatic 8,14secohopanoids 20 and to ring C monoaromatic steroids 16 and 17. For both series again the corresponding benzylic ketones were monitored. The ketones derived from ring C monoaromatic steroids 23 and 24, were only detected in the less altered sample, suggesting that they are formed at an early stage of degradation and are further converted into more complex or more polar components which could not be detected.

Although aromatic hydrocarbons progressively disappear with increasing alteration for the Hit samples, especially those in the low molecular weight range, the corresponding benzylic ketones could not be detected except in the Mari 90 sample. In the samples from Bastennes and Maestu, aromatic hydrocarbons are present only as a UCM. As shown by GC analysis, the ketones as well as the fraction of the aromatic hydrocarbons obtained by Wol€-Kishner reduction also occur as a UCM. This suggests that the aromatic hydrocarbons forming the UCM from the Maestu samples are transformed during atmospheric

992

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Fig. 13. Mass spectrum (EI, 70 eV) of a C27 8,14-secohopanoid ketone present in Maestu 829.

alteration into the corresponding benzylic ketones. This may play an important role in the fate of hydrocarbons occurring in the UCM and which are known to be resistant to biodegradation (Gough et al., 1992). However, some non-aromatic ketones were detected as resolved peaks in the GC trace of the ketone fraction from Bastennes 732. Based on mass spectral data, these components were interpreted as being ketones derived from hopanes (C27±C35) and 8,14-secohopanes 27 (C27± C29) (C27 homologue: Fig. 13) and bearing the carbonyl group in the side chain or on ring E (C27). Indeed, hopanes with 17a(H), 21b(H) con®guration as well as 8,14-secohopanes 11 which are generally present in mature oils and, in particular, in the Gaujacq oil were obtained by Wol€-Kishner reduction of these ketones. A series of steranes (C21±C22, C27±C29) with 5a(H), 14b(H), 17b(H) 20R and 20S con®gurations typical for mature oils were also obtained (in trace amounts) although the corresponding ketones could not be detected by direct GC±MS analysis of the ketone fractions. Clemmensen reduction using D2O (Enzell, 1966) allowed us to locate the carbonyl group on the saturated skeletons. Mass spectra indicated that, in the case of hopanes and 8,14-secohopanes 11, the deuterium atoms were mainly located on the side chain. Thus, fragments m/z 191 and 369 in the case of the hopanes (except C27) and m/z 123 in the case of the 8,14-secohopanes 11 which correspond to the A/B rings remained una€ected whereas those corresponding to the D/E rings were shifted by two mass units. However, detailed analysis of the mass spectrum of the coeluting dideuterated C21 steranes 28, obtained by Clemmensen reduction, seems to indicate a random location of the dideuterated methylene group. For example, fragment 218 (deuteration on the D ring or side-chain) and fragment 220 (deuteration on the A, B or C rings) are clearly present in the spectrum (Fig. 14). Thus, the random location of the deuterated methylene group on the steroid hydrocarbons suggests that these coeluting dideuterated ster-

Fig. 14. Mass spectra (EI, 70 eV) (a) of 5a(H), 14b(H), 17b(H) pregnane and (b) of the mixture of coeluting dideuterated 5a(H), 14b(H), 17b(H) pregnanes formed upon Wol€-Kishner reduction of the ketones from Bastennes 732.

oids result from the reduction of a mixture of isomers of steroid ketones bearing the carbonyl group at various locations. The multiplicity of ketone isomers present before reduction may also explain why the steroid ketones could not be detected as such by GC±MS analysis of the ketone fraction. The non-speci®c location of the carbonyl group on the steroid skeleton suggests that it results from non-speci®c abiotic oxidation. The possibility that non-aromatic ketones may be formed by abiotic oxidation is further supported by simulated photooxidation of n-alkanes in the presence of a photosensitizer (Ehrhardt and Petrick, 1985; Rontani and Giusti, 1987; Giuliano et al., 1997) which led, notably, to the formation of the corresponding ketones. Also, benzylic ketones bearing an additional carbonyl group in the aliphatic side-chain 29 were identi®ed in experiments using the C26 triaromatic steroid 18a hydrocarbon in the presence of light and oxygen (Lemoine, 1996). 3.2.4. Acids Studies were mainly focussed on mono-acids which were analysed as methylesters. In all the samples studied, except in the more altered sample from the Hit

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

series, linear acids dominated. Their distribution (C12± C32) presented a slight even carbon number predominance. The major products were C16 and C18 compounds. Several other series of polycyclic acids appeared in the oxidized samples. From their mass spectra, these acids possess structures, and con®gurations at the asymetric carbon atoms, typical for lipids from mature petroleum or bitumens, excluding the possibility that they result from recent contamination. Acids identi®ed include tricyclic acids 30 (C20±C26, Cyr and Strausz, 1983), steroid acids 31 and 32 (C21±C24, C27-C28) bearing a carboxyl group in the side-chain (Paulus, 1993) and 3-carboxysteroids 33 and 34 (C23, C25, C29±C30) having 5a(H), 14b(H), 17b(H) 20R and 20S con®gurations (as deduced from mass spectra and those of standards; Schae€er et al., 1993), 17a(H), 21b(H)-hopanoic acids (C28±C34) and acids derived from 8,14-secohopanes 35 and 36 (C28±C32). According to their mass spectra, the hopanoids were mainly functionalized on their sidechains. Although these acids probably derive from mature bituminous organic matter, no direct correlation could be established between the saturated hopanoids and the hopanoic acids since the respective distributions were quite di€erent. In the mono-ester fraction of the two most altered samples from the Hit series, novel sedimentary organic sulfur compounds were detected which correspond to a series of alkylated benzothiophene carboxylic acids extending from C9 to C16 (several isomers) as deduced from their mass spectra. The latter are characterized by an intense molecular ion (M+: 192+nx14) and an intense fragment at M+ÿ31 (Fig. 15) suggesting that the carboxyl group is located at the benzylic position. Comparison of the mass spectra and chromatographic behaviour with those of standards (see Experimental) established that the major isomer of the lower homologues is the methylester corresponding to 2-carboxybenzothiophenes 7. The second major isomer may possibly be the methylester deriving from 4-carboxybenzothiophene 37. Its mass spectrum is very similar to that of 7 and nearly identical to that of a carboxybenzothiophene (as methylester) obtained by reaction of 4-methylbenzothiophene with oxygen in the presence of light (see below) and to which the structure of 4-carboxybenzothiophene could be ascribed. Furthermore, the compound obtained in the simulation experiment, and that occurring in the natural sample, coelute on two GC columns coated with di€erent phases. The carboxylic acids detected may have been formed by biodegradation of ``free'' hydrocarbons, or alkylated benzothiophenes, originally present in the bitumen. Thus, the steroid acids bearing the acid function in the side-chain [present in the most altered samples of the Hit series (Mari 90)] are formed during in vitro biodegradation of steranes (Paulus, 1993). Similarly, linear carboxylic acids have been shown to be possible biode-

993

Fig. 15. Mass spectra (EI, 70 eV) (a) of 2-carbomethoxybenzo thiophene and (b) of a dimethylcarbomethoxybenzothiophene obtained after methylation (diazomethane) of the acid fraction from the samples Hit 141 and Mari 90.

gradation products of linear alkanes (Jones and Howe, 1968; Klug and Markovetz, 1971; Fonken and Johnson, 1972). Formation of carboxylic acids by biodegradation of hopanes, tricyclic terpanes or alkylated benzothiophenes is also possible. Indeed, carboxylic acids related to hopanes (con®guration at the asymmetric carbon atoms typical of mature petroleum) and tricyclic terpanes have been detected in biodegraded oils (e.g. Cyr and Strausz, 1983; Mackenzie et al., 1983; Behar and Albrecht, 1984). However, formation of these acids by biological oxidation of the corresponding saturated hydrocarbons has, to our knowledge, never been demonstrated. Alternative explanations can be proposed for at least a part of the carboxylic acids present in the altered samples. The aliphatic acids (linear, hopanoid, tricyclic) may have been released from the macromolecular constituents of the bitumen by hydrolysis of esters. Alternatively, these acids may have been formed by an abiotic oxidation process a€ecting either free lipids or lipids linked to the macromolecular matrix, in particular those linked by C±C bonds to alkylaromatic sub-units (a type of linkage which has been demonstrated to occur in geomacromolecules; Tri®lie€ et al., 1992). Indeed, the latter possibility is supported by simulation experiments in which 9-heptadecylphenanthrene 39 was exposed to oxygen and light (Lemoine, 1996) resulting in the formation of the related benzylic ketone 40 but also in the partial cleavage of the side-chain: a mixture of linear

994

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

acids (C13±C18, C16 and C17 being predominant) was recovered, and degradation of the aromatic skeleton gave the lactone 41 (Fig. 16). For the C18 acid, 17 of the carbon atoms must come from the alkyl side-chain of 9heptadecylphenanthrene and one from the aromatic part of the molecule. This can only be explained by oxidation of the aromatic rings by singlet oxygen which can be formed photochemically, the benzylic ketone related to 9-heptadecylphenanthrene 40 acting as a photosensitizer. Thus, alkyl skeletons linked to aromatic subunits in geomacromolecules, or in low-molecular weight aromatics, will be oxidized at the benzylic position to the corresponding ketones. The latter can then be further oxidized, resulting in the cleavage of C±C bonds and, subsequently, in the release of carboxylic acids (Fig. 17). Alternatively, release of carboxylic acids may result from oxidation at the aromatic rings (Fig. 17) induced, for example, by singlet oxygen. If the benzylic position is a secondary position, such an oxidative degradation reaction could also result in the release of ketones. Therefore, the C27 and C29 hopanoid or 8,14-secohopanoid ketones present in Bastennes 732 may have been formed by a process similar to that a€ecting hopanoid thiophenes or hopanes and 8,14-secohopanes linked to aromatic units. The benzothiophenic acids identi®ed in the most altered samples may also have been formed by abiotic oxidation. Indeed, 2-carboxybenzothiophene 7 present in the Hit samples has been obtained when 2-methylbenzothiophene 2 was exposed to air and light (Andersson and Bobinger, 1996). Similarly, 4-carboxybenzothiophene 37 (mass spectrum almost identical to that of 7) also present in the Hit samples was formed by the reaction of 4-methylbenzothiophene 3 with oxygen in the presence of light. Furthermore, 2-butyl-4-hexylbenzothiophenes 4 ([M+]=388) synthesized by Perakis (1986) led, in addi-

Fig. 16. Predominant oxygenated compounds formed upon reaction of 9-heptadecylphenanthrene with oxygen under sunlight irradiation.

tion to the predominant formation of the corresponding benzylic ketones 42 and 43, ([M+]=402) to the cleavage of one alkyl substituent and to the formation of a benzothiophenic monoacid 44 (as methylester: [M+]=290, fragment at m/z 259 ([M+]ÿ31)). These results suggest that the benzothiophenic acids in the altered samples could have been formed by abiotic oxidation of methylor alkylbenzothiophenes initially present in the aromatic fraction (Fig. 18). Alternatively, they may have been formed by release of alkylbenzothiophenic subunits linked to macromolecular structures by oxidation of the alkyl chain at the benzylic position (Fig. 18a). 3.2.5. Alcohols Alcohols have not been detected in our samples, except for the Gokhurt series where hopanoid alcohols, ranging from C30 to C32, have been identi®ed. Mass spectral data indicate that 17a(H) 21b(H), 17b(H) 21a(H) and 17b(H) 21b(H) isomers are present. Therefore, these hopanoids probably derive from recent organic matter. 3.2.6. Organic sulfur compounds The Hit series was particularly rich in organic sulfur compounds. In the less degraded oil, numerous polycyclic sul®des, isolated by silver ion TLC, were recognized (Fig. 19a). Three series (45, 46, 47) of hopanoid sul®des (C30±C35, C31 predominant) (Philippe, 1993; Schae€er, 1993) were present as well as steroid sul®des (C27-C29; 48 and 49) (Schmid, 1986; Behrens et al., 1997), bicyclic sul®des 50 (C13±C24, C18 and C20 predominant) and tetracyclic sul®des 51 (C23±C30) (Payzant et al., 1986). In the intermediate sample, all low molecular weight components had disappeared but the hopanoid sul®des were still detectable (Fig. 19b). Sul®des were totally absent in the most degraded samples (Hit 141 and Mari 90). The sulfones corresponding to the cyclic sul®des described above appeared in polar fractions (``alcohol'' fraction) from the two most degraded samples (Hit 141 and Mari 90). They were identi®ed from their mass spectra, comparison with reference compounds (Philippe, 1993) and by means of their distribution. Thus, the hopanoid sulfones (C30±C32, C31 predominant), related to sul®des 45, 46, 47, display a distribution similar to that of the sul®des. Tetracyclic sulfones were recognized as steroid sulfones (C22). The lack of long chain steroid sulfones may be explained by the preferential biodegradation of steroids with a long sidechain. Another type of tetracyclic sulfone, based on a tricyclic C23 hydrocarbon skeleton, was identi®ed as the C23 sulfone found in Athabasca bitumens by Strausz et al. (1990). Sulfones corresponding to the bicyclic sul®des 50 were absent. Because of their low molecular weight, and their enhanced solubility in water, they may have been removed by water washing.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

995

Fig. 17. Possible mechanism of the oxidative release of carboxylic acids from macromolecular material occurring in sedimentary organic matter.

Fig. 18. Possible mode of formation of benzothiophenic acids in bitumens exposed to atmospheric conditions (a) from benzothiophenic subunits linked to geomacromolecules; (b) from ``free'' dialkylbenzothiophenes.

Several studies have shown that bacteria are able to oxidize sul®des to sulfones and sulfoxides (e.g. Fedorak, 1990, Jenisch et al., 1995). Also, the high sensitivity of aliphatic sul®des to several oxidants (hydrogen perox-

ide, ozone, singlet oxygen; Oae, 1991) is well known. The fact that all the sul®des in the Hit series, were converted to the related sulfone, regardless of structure (tetracyclic terpanes, steroids, hopanoids etc) suggests

996

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Fig. 19. Gas chromatograms (Conditions c) of the sul®de fractions isolated from Hit 143 (a) and from Hit 142 (b).&: steroid sul®des 48+49; !: bicyclic sul®des 50; *: hopanoid sul®des 45; *: hopanoid sul®des 47; ^: hopanoid sul®des 46; & tetracyclic sul®des 51.

that those in the most weathered samples result predominantly from abiotic processes. Aromatic sulfur compounds are present in signi®cant amounts in the less altered samples from the Hit and the Bichri series : among them, alkylated dibenzothiophenes prevail. Based on their mass spectra, they bear one to eight aliphatic carbon atoms. A complex alkylbenzothiophene series was detected by mass spectrometry. Alkyldibenzothiophenes and alkylbenzothiophenes are absent from the most altered bitumen, partly due to water washing/evaporation but probably also to biological or abiotic oxidation processes (Fedorak, 1990). Interestingly, as described above, some carboxylic acids, derived from the alkylbenzothiophenes, were found in the oxidized samples from the Hit series. However, no sulfones or sulfoxides related to these aromatic sulfur compounds (alkylated benzothiophenes and dibenzothiophenes) could be detected in the altered samples suggesting that, under the given conditions, sulfur in aromatic rings is more dicult to oxidize than sulfur in thianes or thiolanes. 3.3. Oxidants and abiotic oxidation mechanisms in natural environments Among the oxidants occurring in nature, and liable for the formation of the oxygenated compounds detec-

ted in our samples, triplet oxygen is the most abundant and is known to participate in radical-type autoxidation reactions (Ingold, 1969; Voronenkov et al., 1970; Korcek et al., 1972; Sheldon and Kochi, 1981). Alkyl radicals are able to react with triplet oxygen to form hydroperoxide radicals, which can then abstract hydrogen on another hydrocarbon skeleton, leading to a new alkyl radical, and a radical chain reaction can proceed. The intermediate peroxides and peroxide radicals are further transformed into several oxygenated compounds such as acids, alcohols and ketones or can undergo cleavage reactions (Perkel' et al., 1994) leading to smaller oxygenated molecules. The most likely location for this type of oxidation are benzylic or allylic positions, as well as tertiary alkyl positions, which, by abstraction of hydrogen, lead to stable radicals. Thus, this type of reaction can account for the formation of aromatic ketones, or the benzothiophenic acids bearing an oxygen functionality at the benzylic position, as identi®ed in the most weathered samples from the Hit series. Such processes may also induce cleavage and may contribute to the degradation of high molecular weight constituents of the bitumens, leading to the release of low-molecular weight oxygenated components and explain (see below), some of the carboxylic acids which have appeared in the most altered samples.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

However, triplet oxygen alone is not very reactive and free radicals are generally necessary to initiate the oxidation reaction by abstraction of hydrogen. Free radicals are present in nature, for example in humic substances (Senesi and Steelink, 1989). Free radicals can also be formed photochemically in the presence of radical initiators. Aliphatic as well as benzylic ketones, which are the suspected products of such radical oxidation a€ecting bitumens are also good photochemical radical initiators (Kopecky, 1992). In consequence, the presence of light may favour the oxidation of the organic matter from petroleum, and the reaction products may even accelerate the process. Beside triplet oxygen, other oxidants may be responsible for the formation of the oxygen-containing compounds, such as singlet oxygen, ozone, hydroxyl or peroxy radicals, superoxides, nitrogen oxides and hydrogen peroxide (Cooper et al., 1989; Aneja et al. 1994). Some, although present in low concentrations, are ubiquitous (hydrogen peroxide, nitrogen oxides), others are transient, highly reactive species (e.g. hydroxyl radicals, singlet oxygen). All may occur, in nature, from photochemical reactions involving photosensitizers (Cooper et al., 1989) such as humic substances, aromatic hydrocarbons, benzylic ketones and petroporphyrins. Relatively stable oxidants such as hydrogen peroxide, formed by photochemical reactions in nature can be transported to the bitumens by rainwater. Alternatively, oxidants may be formed by photochemical reactions between sensitizers already present in bitumens and react directly with the bitumens' organic matter. In such a situation, unstable, highly reactive oxidants such as hydroxyl radicals or singlet oxygen may also participate in the oxidation processes a€ecting bitumens. 4. Conclusions Analyses of ®ve series of crude oil samples exposed to the atmosphere, but altered to di€erent extents, revealed that dramatic changes have a€ected the more degraded samples. Elemental analysis and IR spectroscopy showed that atmospheric alteration results in an increase in the oxygen content which parallels an increase in the proportions of the macromolecular constituents. A substantial depletion of the low-molecular-weight constituents, in particular of the saturated and aromatic compounds, is also observed. D has been shown to be particularly sensitive to alteration processes as it increases from values of about ÿ100% in the less altered samples up to about ÿ50% in the more altered ones. This may be the result of H/D exchange reaction with water on oxygenated functionalities formed during atmospheric alteration by biological and abiotic oxidation processes. Although some of the saturated and aromatic hydrocarbons have probably been removed by water washing

997

and evaporation, we have shown that some of these fractions have been transformed into oxygenated compounds. Indeed, analysis of the polar fractions (ketones, sulfones, carboxylic acids) allowed us to identify oxygen-containing compounds clearly resulting from the oxidation of typical petroleum components. For example, benzylic ketones are related to aromatic biomarkers (mono- and triaromatic steroids, ring D monoaromatic 8,14-secohopanoids), and carboxylic acids are related to steroids, tricyclic terpenoids, hopanoids and benzothiophenes. Sulfones, which are the result of the oxidation of steroid, terpenoid or hopanoid sul®des were also detected. The hypothesis that these oxygenated compounds result at least partially from an abiotic oxidation process, rather than from biodegradation, is supported by quantitative data obtained for steroid triaromatic ketones; these compounds appear with increasing alteration and are absent in the less altered, but already biodegraded samples. Furthermore, oxygen incorporation generally occurred without diastereomeric discrimination, a feature unexpected for biological processes. Simulation experiments, in which typical constituents of mature petroleum (triaromatic steroids, alkylated benzothiophenes, alkylphenanthrenes) were exposed to oxygen and light, further support this hypothesis, since they were transformed into benzylic ketones or acids similar to those detected in our samples. The results also suggest that abiotic oxidation can severely alter macromolecular constituents of petroleum, as well as aromatic constituents of the UCM. Simulation experiments suggest that abiotic oxidation may favour the release, by oxidative cleavage, of subunits from the macromolecules as oxygenated low molecular weight components (e.g. carboxylic acids or ketones). Finally, it has been shown that incorporation of oxygen functionalities transforms part of the highly lipidic macromolecular fractions of petroleum into more hydrophilic humic-like substances. In summary, our results suggest that alteration by abiotic oxidation may play a key role in the fate of petroleum pollutants in the environment. Indeed, abiotic incorporation of oxygen may lead to the formation of oxygenated components by the oxidation of low molecular weight lipids but also by inducing cleavage reactions in macromolecular substances. The increased water solubility of these oxygenated constituents probably allows a better bio-availability and, therefore, an increased rate of biodegradation. However, it may be that some of these oxygenated compounds are toxic to biota especially since they are more bioavailable. Furthermore, the important incorporation of hydroxyl and carboxyl functions into petroleum macromolecules (transforming them into humic-like substances) combined with oxidative cleavage reactions may also favour their progressive solubilization in nature.

998

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Acknowledgements We thank Elf Exploration Production and the Centre National de la Recherche Scienti®que for doctoral fellowships (A. CharrieÂ-Duhaut and S. Lemoine) and ®nancial support; Estelle Mastio-Motsch, Marie-Claude Schweigert and Patrick Wehrung, Universite Louis Pasteur, Strasbourg, for mass spectral analyses; Roland

Appendix

Gra€, Universite Louis Pasteur, Strasbourg, for NMR measurements and Jean-Bernard Berrut, Elf Exploration Production for molecular weight measurements. This study has bene®ted from the ®nancial support of the ECODEV-CNRS program. We also thank Jan Andersson, Simon C. George and Kliti Grice for most valuable reviews. Associate Editor Ð S.C. George

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

999

1000

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

References Andersson, J.T., Bobinger, S., 1996. Photochemical degradation of crude oil components: 2-methyl-, 3-methyl- and 2,3dimethylbenzothiophene. Polycyclic Aromatic Compounds 11, 145±151. Aneja, V.P., Das, M., Kim, D.S., Hartsell, B.E., 1994. Measurements and analysis of photochemical oxidants and trace gases in the rural troposphere of the Southeast United States. Israel Journal of Chemistry 34, 387±401. Aquino Neto, F.R., RestleÂ, A., Connan, J., Albrecht, P., Ourisson, G., 1982. Novel tricyclic terpanes C19, C20 in sediments and petroleums. Tetrahedron Letters 23, 2027±2030. Aquino Neto, F.R., Trendel, J.M., RestleÂ, A., Connan, J., Albrecht, P., 1983. Occurrence and formation of tricyclic and tetracyclic terpanes in sediments and petroleums. In: Bjorùy, M. et al. (Eds.), Advances in Organic Geochemistry 1981. John Wiley and Sons, Chichester, pp. 659±676. Behar, F.H., Albrecht, P., 1984. Correlations between carboxylic acids and hydrocarbons in several crude oils:alteration by biodegradation. Organic Geochemistry 6, 597±604. Behrens, A., Schae€er, P., Albrecht, P., 1997. 14b, 22R-epithiosteranes, a novel series of fossil steroids widespread in sediments. Tetrahedron Letters 38, 4921±4924. Bence, A.E., Kvenvolden, K.A., Kennicut II., M.C., 1996. Organic geochemistry applied to environmental assessment of Prince William Sound, Alaska, after Exxon Valdez oil spill Ð a review. Organic Geochemistry 24, 7±42. Brown, J.R., Kasrai, M., Bancroft, G.M., Tan, K.H., Chen, J.M., 1992. Direct identi®cation of organic sulphur in Rasa coal from sulphur L-edge X-ray absorption near-edge spectra. Fuel 71, 649±653. Bynum, D., Traxler, R.N., Parker, H.L., Ham, J.S., 1970. Application of gel permeation chromatography to asphalts. Journal of the Institute of Petroleum 56, 147±154. Calemma, V., Iwanski, P., Nali, M., Scotti, R., Montanari, L., 1995. Structural characterization of asphaltenes of di€erent origins. Energy & Fuels 9, 225±230. Chosson, P., Connan, J., Dessort, D., Lanau, C., 1992. In vitro biodegradation of steranes and terpanes: a clue to understanding geological situations. In: Moldowan, J.M., Albrecht, P., Philp, R.P. (Eds.), Biological Markers in Sediments and Petroleum. Prentice Hall, Englewood Cli€s, pp. 320±349. Clayton, J.L., Swetland, P.J., 1978. Subaerial weathering of sedimentary organic matter. Geochimica et Cosmochimica Acta 42, 305±312. Clayton, J.L., King, J.D., 1987. E€ects of weathering on biological marker and aromatic hydrocarbon composition of organic matter in Phosphoria shale outcrop. Geochimica et Cosmochimica Acta 51, 2153±2157. Connan, J., RestleÂ, A., Albrecht, P., 1980. Biodegradation of crude oil in the Aquitaine basin. In: Douglas, A.G., Maxwell, J.R. (Eds.), Advances in Organic Geochemistry 1979. Pergamon Press, Oxford, pp. 29±50. Connan, J., 1984. Biodegradation of crude oils in reservoirs.. In: Brooks, J., Welte, D.H. (Eds.), Advances in Petroleum Geochemistry Vol.I. Academic Press, London, pp. 299±335. Connan, J., Dessort, D., 1987. Novel family of hexacyclic hopanoid alkanes (C32±C35) occurring in sediments and oils from anoxic paleoenvironments. Organic Geochemistry 11, 103±113.

Connan, J., Deschesne, O., Dessort, D., 1990. L'origine des bitumes de Ras Shamra. Bulletin du. Centre de Recherche Exploration- Production Elf-Aquitaine 14, 521±539. Connan, J., Dessort, D., Deschesne, O., 1992a. Les bitumes de Tell-Es-Sawwan en Irak : un modeÁle d'huiles extreÃmement biodeÂgradeÂes et oxydeÂes. Bulletin du Centre de Recherche Exploration- Production Elf-Aquitaine 16, 33±53. Connan, J., Nissenbaum, A., Dessort, D., 1992b. Molecular archaelogy:export of dead sea asphalt to Canaan and Egypt in the Chalcolithic-early Bronze Age (4th-3rd millenium BC). Geochimica et Cosmochimica Acta 56, 2743±2759. Connan, J., 1999. Use and trade of bitumen in antiquity and prehistory: molecular archaeology reveals secrets of past civilizations. Phil. Trans. R. Soc. Lond. B 354, 33±50. Cooper, W.J., Zika, R.G., 1983. Photochemical formation of hydrogen peroxide in surface and groundwater exposed to sunlight. Science 220, 711±712. Cooper, W.J., Zika, R.G., Petasne, R.G., Fisher, A.M., 1989. Sunlight-induced photochemistry of humic substances in natural waters:major reactive species. In: Su€et, I.H., MacCarthy, P. (Eds.), Aquatic Humic Substances ACS Symposium Series No. 219. American Chemical Society, Washington, pp. 333±362. Cyr, T.D., Strausz, O.P., 1983. The structure of tricyclic terpenoid carboxylic acids and their parent alkanes in the Alberta oil sands. Journal of the Chemical Society, Chemical Communication. 1028±1030. Dahl, B., Speers, G.C., Steen, A., Telnaes, N. & Johansen, J.E., 1985. Quanti®cation of steranes and triterpanes by gas chromatographic-mass spectroscopic analysis. In: Thomas B.M. et al (Eds.), Petroleum Geochemistry in Exploration of the Norvegian Shelf. Norvegian Petroleum Society, Graham and Trotman. pp 303±307. Dastillung, M., 1976. Lipides de SeÂdiments ReÂcents. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg, France. DeNiro, M.J., Epstein, S., 1981. Isotopic composition of cellulose from aquatic organisms. Geochimica et Cosmochimica Acta 45, 1885±1894. Dessort, D., Connan, J., 1993. Occurrence of novel C-ring opened hopanoids in biodegraded oils from carbonate source rocks. 16th International Meeting on Organic Geochemistry Stavanger, Norway, September 1993. In: éygard, K. (Ed.), Organic Geochemistry. Falch Hurtigtrykk, Kalbakken, pp. 485±495. Enzell, C.R., 1966. Deuterium labelling by Clemmensen reduction. Tetrahedron Letters 12, 1285±1288. Ehrhardt, M., Petrick, G., 1984. On the sensitized photo-oxidation of alkylbenzenes in seawater. Marine Chemistry 15, 47±58. Ehrhardt, M., Petrick, G., 1985. The sensitized photo-oxidation of n-pentadecane as a model for abiotic decomposition of aliphatic hydrocarbons in seawater. Marine Chemistry 16, 227±238. Ehrhardt, M., Bouchertall, F., Hopf, H.P., 1989. Aromatic ketones concentrated from Baltic sea water. Marine Chemistry 11, 449±461. Ensminger, A., Van Dorsselaer, A., Spyckerelle, C., Albrecht, P., Ourisson, G., 1974. Pentacyclic triterpenes of the hopane type as ubiquitous geochemical markers: origin and signi®cance. In: Tissot, B.P., Bienner, F. (Eds.), Advances in Organic Geochemistry 1973. Technip, Paris, pp. 245±260.

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003 Fedorak, P.M., 1990. Microbial metabolism of organosulfur compounds in petroleum. In: Orr, W.L., White, C.M. (Eds.), Geochemistry of Sulfur in Fossil Fuels, ACS Symposium series No 429. American Chemical Society, Washington, pp. 93±112. Fisher, S.J., Alexander, R., Kagi, R.I., Oliver, G.A. 1998. Aromatic hydrocarbons as indicators of biodegradation in the north western Australian reservoirs. In: Purcell, P.G. and Purcell, R.R. (Eds.), The Sedimentary Basins of Western Australia 2. Proceedings of the Petroleum Exploration Society of Australia Symposium, Perth, WA, pp. 185±194. Fonken, G.S., Johnson, R.A., 1972. Chemical Oxidation with Microorganisms. Marcel Dekker, New York. Forsberg, A., Bjorùy, M., 1983. A sedimentological and organic geochemical study of the Botneheia Formation, Svalbard, with special emphasis on the e€ects of weathering on the organic matter in shales. In: Bjorùy, M. et al. (Eds.), Advances in Organic Geochemistry 1981. John Wiley & Sons, Chichester, pp. 60±68. Giuliano, M., El Anba-Luro, F., Doumenq, P., Mille, G., Rontani, J.F., 1997. Photo-oxidation of n-alkanes in simulated marine environmental conditions. Journal of Photochemistry and Photobiology A:Chemistry 102, 127±132. Goodwin, N.S., Park, P.J.D., Rawlinson, T., 1983. Crude oil biodegradation. In: Bjorùy, M. et al. (Eds.), Advances in Organic Geochemistry 1981. John Wiley & Sons, Chichester, pp. 650±658. Gough, M.A., Rhead, M.M., Rowland, S.J., 1992. Biodegradation studies of unresolved complex mixtures of hydrocarbons: model UCM hydrocarbons and the aliphatic UCM. Organic Geochemistry 18, 17±22. Gundlach, E.R., Boehm, P.D., Marchand, M., Atlas, R.M., Ward, D.M., Wolfe, D.A., 1983. The fate of Amoco Cadiz oil. Science 221, 122±129. He Wei, Lu Songnian, 1990. A new maturity parameter based on monoaromatic hopanoids. Organic Geochemistry 16, 1007±1013. Hostettler, F.D., Kvenvolden, K.A., 1994. Geochemical changes in crude oil spilled from the Exxon Valdez supertanker into Prince William Sound, Alaska. Organic Geochemistry 21, 927±936. Huang Minlon, A., 1946. simple modi®cation of the Wol€Kishner reduction. Journal of the American Chemical Society 90, 99±107. Hussler, G., Connan, J., Albrecht, P., 1984a. Novel families of tetra- and hexacyclic aromatic hopanoids predominant in carbonate rocks and crude oils. Organic Geochemistry 6, 39± 49. Hussler, G., Cesario, M., Guilhem, J., Pascard, C., Albrecht, P., Ourisson, G., 1984b. Benzohopanes, a novel family of hexacyclic geomarkers in sediments and petroleums. Tetrahedron Letters 25, 1179±1182. Ingold, K.U., 1969. Peroxy radicals. Accounts of Chemical Research 2, 1±9. Jenisch, A., Adam, P., Herrmann, D., Connan, J., Rohmer, M., Albrecht, P., 1995. Biodegradation of macromolecular crude oil fractions. In: Grimalt, J.O., Dorronsoro, C. (Eds.), Organic Geochemistry: Developments, Application to Energy, Climate, Environment, Human History. AIGOA, San Sebastian, pp. 894±895. Jones, D.F., Howe, R., 1968. Microbiological oxidation of

1001

long-chain aliphatic compounds. Part I., Alkanes and alk-1enes. Journal of the Chemical Society (C), 2809±2818. Kagi, R.I., Fisher S.J., Alexander, R., 1988. Behaviour of petroleum in northern Australian waters. In: Purcell, P.G. and Purcell, R.R. (Eds.), The North West Shelf Australia. Proceedings of the North West Shelf Symposium. Perth, Petroleum Exploration Society of Australia, pp. 633±642. Kopecky, J., 1992. Photochemistry of carbonyl compounds. In: Photochemistry. A Visual Approach. VCH Publishers, NewYork. pp.117±134. Korcek, S., Chenier, J.H.B., Howard, J.A., Ingold, K.U., 1972. Absolute rate constants for hydrocarbon autoxidation XXI. Activation energies for propagation and the correlation of propagation rate constants with carbon-hydrogen bond strengths. Canadian Journal of Chemistry 50, 2285±2297. Kuo, L.C., 1994. An experimental study of crude oil alteration in reservoir rocks by water washing. Organic Geochemistry 21, 465±479. Klug, M.J., Markovetz, A.J., 1971. Utilization of aliphatic hydrocarbons by microorganisms. Advances in Microbial Physiology 5, 1±43. Lemoine, S., 1996. Oxydation Abiotique de Bitumes PeÂtroliers en Milieu Naturel. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg, France. Lemoine, S., Adam, P., Connan, J., Albrecht, P., 1996. Novel series of diaromatic 14-methyl steroids occurring in petroleum. Tetrahedron Letters 37, 2837±2840. Leythaeuser, D., 1973. E€ects of weathering on organic matter in shales. Geochimica et Cosmochimica Acta 37, 113±120. Li, M., Riediger, C.L., Fowler, G.M., Snowdon, L.R., Abrajano, Jr, T.A., 1996. Unusual polycyclic alkanes in lower Cretaceous Ostracode sediments and related oils of the Western Canada sedimentary basin. Organic Geochemistry 25, 199±209. McCarthie, R.D., Duthie, A.H., 1962. A rapid quantitative method for the separation of the free fatty acids from the other lipids. Journal of Lipid Research 3, 117±119. Mackenzie, A.S., Wol€, G., Maxwell, J.R., 1983. Fatty acids in some biodegraded petroleums. Possible origin and signi®cance. In: Bjorùy, M. et al. (Eds.), Advances in Organic Geochemistry 1981. John Wiley & Sons, Chichester, pp. 637± 649. MacKay, D., McAuli€e, C.D., 1988. Fate of hydrocarbons discharged at sea. Oil Chemistry and Pollutions 5, 1±20. Mackenzie, A.S., 1984. Applications of biological markers in petroleum geochemistry. In: Brook, J., Welte, D. (Eds.), Advances in Petroleum Geochemistry Vol. 1. Academic Press, London, pp. 115±214. Moldowan, J.M., McCa€rey, M.A., 1995. A novel microbial hydrocarbon degradation pathway revealed by hopane demethylation in a petroleum reservoir. Geochimica et Cosmochimica Acta 59, 1891±1894. Novak, J., Sorm, F., Sicher, J., 1954. Total synthesis of chamazulene (A simple general synthesis of 1,4,7-trisubstituted azulenes). Collection of Czechoslovak Chemical Communication 19, 1264±1272. Oae, S., 1991. Organic Sulfur Chemistry: Structure and Mechanism. CRC Press, Boca Raton, Florida. Parsons, J.W., 1988. Isolation of humic substances from soils and sediments. In: Frimmel, F.H., Christman, R.F (Eds.), Humic Substances and their Role in the Environment. John Wiley & Sons, Chichester, pp. 3±16.

1002

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003

Paulus, S., 1993. BiodeÂgradation de SteÂranes PeÂtroliers. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg, France. Payne, J.R., Phillips, C.R., Hom, W., 1987. Transport and transformations: water column processes. In: Boesch, D.F., Rabalais, N.N. (Eds.), Long-term Environmental E€ects of O€shore Oil and Gas Development. Elsevier Applied Science Publishers, London, pp. 175±231. Payzant, J.D., Montgomery, D.S., Strausz, O.Z., 1986. Sul®des in petroleum. Organic Geochemistry 9, 357±369. Perakis, N., 1986. SeÂparation et DeÂtection SeÂlective de ComposeÂs SoufreÂs dans les Fractions Lourdes de PeÂtroles. GeÂochimie des Benzo[b]thiopheÁnes. TheÁse de doctorat, Universite Louis Pasteur, Strasbourg, France. Perkel', A.L., Voronina, S.G., Freidin, B.G., 1994. Degradation of the carbon chain in the liquid phase, oxidation of saturated compounds. Russian Chemical Reviews 63, 751±766. Peters, K.E., Moldowan, J.M., 1993. The Biomarker Guide. Interpreting Molecular Fossils in Petroleum and Ancient Sediments. Prentice Hall, Englewood Cli€s, NJ. Philippe, E., 1993. Identi®cation et MeÂcanisme de Formation de composeÂs Organo-soufreÂs SeÂdimentaires. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg, France. Poinsot, J., 1997. Identi®cation de Nouveaux Biomarqueurs Biologiques PolyterpeÂniques d'inteÂreÃt GeÂochimique. TheÁse de Doctorat, Universite louis Pasteur, Strasbourg, France. Poinsot, J., Dessort, D., Adam, P., Connan, J., Lacrampe, G., Trendel, J.M., Albrecht, P., 1995. New and rare biomarkers: Saturated C33-C43 polycycloisoprenoids. In: Grimalt, J.O., Dorronsorro, C. (Eds.), Organic Geochemistry: Developments. Application to Energy Climate Environment Human History. AIGOA, San Sebastian, pp. 116±118. Prince, R.C., Elmendorf, D.L., Lute, J.R., Hsu, C.S., Haith, C.E., Senius, J.D., Dechert, G.J., Douglas, G.S., Butler, E.L., 1994. 17a(H), 21b(H)-hopane as a conserved internal marker for estimating the biodegradation of crude oil. Environmental Science and Technology 28, 142±145. Requejo, A.G., Halpern, H.I., 1989. An unusual hopane biodegradation sequence in tar sands from the Pt Arena (Monterey) Formation. Nature 342, 670±673. Rontani, J.F., Giusti, G., 1987. Photosensitized oxidation of pristane in sea water: e€ects of photochemical reactions on tertiary carbons. Journal of Photochemistry and Photobiology, A:Chemistry 40, 107±120. Rosset, R., Hui, F., Xie, J., Machtalere, G., Kolodjiejczyk, H., Sassiat, P., 1994. Nouvelles meÂthodes d'eÂtalonnage en chromatographie d'exclusion steÂrique. Analusis 22, 293±304. RullkoÈtter, J., Wendisch, D., 1982. Microbial alteration of 17b(H)-hopanes in Madagascar asphalts: removal of C-10 methyl group and ring opening. Geochimica et Cosmochimica Acta 46, 1545±1553. Sarret, G., Connan, J., Kasrai, M., Bancroft, G.M., CharrieÂDuhaut, A., Lemoine, S. et al., 1999. Chemical forms of sulfur in geological and archaeological asphaltenes from Middle East, France, and Spain determined by sulfur K- and L-edge X-ray absorption near-edge structure spectroscopy. Geochimica et Cosmochimica Acta 63, 3767±3779. Schae€er, P., 1993. Marqueurs Biologiques de Milieux Evaporitiques. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg. Schae€er, P., Fache-Dany, F., Trendel, J.M., Albrecht, P., 1993. Polar constituents of organic matter rich marls from

evaporitic series of Mulhouse Basin. Organic Geochemistry 20, 1227±1236. Schae€er, P., Poinsot, J., Hauke, V., Adam, P., Wehrung, P., Trendel, J.M. et al., 1994. Novel optically active hydrocarbons in sediments (Evidence for an extensive biological cyclisation of regular polyprenols). Angewandte Chemie International Edition (in English) 33, 1166±1169. Schmid, J.C., 1986. Marqueurs Biologiques SoufreÂs dans les PeÂtroles. TheÁse de Doctorat, Universite Louis Pasteur, Strasbourg. Schmitter, J.M., Sucrow, W., Arpino, P.J., 1982. Occurrence of novel tetracyclic geochemical markers:8,14-seco-hopanes in a Nigerian crude oil. Geochimica et Cosmochimica Acta 46, 2345±2360. Seifert, W.K., Moldowan, J.M., 1979. The e€ect of biodegradation on steranes and terpanes in crude oils. Geochimica et Cosmochimica Acta 43, 111±126. Seifert, W.K., Moldowan, J.M., 1980. The e€ect of thermal stress on source-rock quality as measured by hopane stereochemistry. In: Douglas, A.G., Maxwell, J.R. (Eds.), Advances in Organic Geochemistry 1979. Pergamon Press, Oxford, pp. 229±237. Seifert, W.K., Moldowan, J.M., Demaison, G.J., 1984. Source correlation of biodegraded oils. Organic Geochemistry 6, 633±643. Senesi, N., Steelink, C., 1989. Application of ESR spectroscopy to the study of humic substances. In: Hayes, M.H.B., MacCarthy, P., Malcolm, R.L., Swift, R.S. (Eds.), Humic Substances II. John Wiley & Sons, New York, pp. 373±408. Sheldon, A., Kochi, J.K. 1981. Metal-Catalyzed Oxidations of Organic Compounds, Academic Press, New-York. da Silveira Lobo Sternberg, L., 1988. D/H ratios of environmental water recorded by D/H ratios of plant lipids. Nature 333, 59±61. Snyder, L.R., 1969. Determination of asphalt molecular weight distributions by gel permeation chromatography. Analytical Chemistry 41, 1223±1227. Speight, J.G., 1984. The chemical nature of petroleum asphaltenes. In: Tissot, B.P. (Ed.), CaracteÂrisation des Huiles Lourdes et des ReÂsidus PeÂtroliers, Symposium International, Lyon. Technip, Paris, pp. 32±41. Speight, J.G., Wernick, D.L., Gould, K.A., Over®eld, R.E., Rao, B.M.L., Savage, D.W., 1985. Molecular weight and association of asphaltenes: a critical review. Revue de l'Institut FrancËais du PeÂtrole 40, 51±61. Strausz, O.P., Lown, E.M., Payzant, J.D., 1990. Nature and geochemistry of sulfur-containing compounds in Alberta petroleums. In: Orr, W.L., White, C.M. (Eds.), Geochemistry of Sulfur in Fossil Fuels. ACS Symposium series No 429. American Chemical Society, Washington, pp. 368± 396. Tri®lie€, S., Sieskind, O., Albrecht, P., 1992. Biological markers in petroleum asphaltenes: possible mode of incorporation. In: Moldowan, J.M., Albrecht, P., Philp, R.P. (Eds.), Biological Markers in Sediments and Petroleum. PrenticeHall, Englewood Cli€s, pp. 350±369. Van der Linden, A.C., Thijsse, G.J.E., 1965. The mechanism of microbial oxidation of petroleum hydrocarbons. Advances in Enzymolology 27, 469±546. Van Dorsselaer, A., Connan, J., Albrecht, P., 1977. Changes in composition of polycyclic alkanes by thermal maturation

A. CharrieÂ-Duhaut et al. / Organic Geochemistry 31 (2000) 977±1003 (Yallourn lignite, Australia). In: Campos, R., Goni, J., (Eds.), Advances in Organic Geochemistry 1975. ENADISMA, Madrid, pp. 53±59. Volkman, J.K., Alexander, R., Kagi, R.I., Woodhouse, G.W., 1984. Biodegradation of aromatic hydrocarbons in crude oils from the Barrow Sub-basin of Western Australia. In: Schenck, P.A., de Leeuw, J.W. (Eds.), Advances in Organic Geochemistry 1983. Pergamon Press, Oxford, pp. 619±632. Volkman, J.K., Miller, G.J., Revill, A.T., Connell, D.W., 1994. Oil spills. In: Swan, J.M., Ne€, J.M., Young, P.C. (Eds.), Environmental Implications of O€shore Oil and Gas Development in Australia Ð the Findings of an Independent Review. Australian Petroleum Exploration Association, Sydney, pp. 509±695.

1003

Voronenkov, V.V., Vinogradov, A.N., Belyaev, V.A., 1970. Structural and steric factors determining the rate of autoxidation of a-methylene groups in alkylarenes and alkenes. Russian Chemical Reviews 39, 944±952. Wang, T.G., Simoneit, B.R.T., Philp, R.P., Yu, C.P., 1990. Extended 8b(H)-drimane and 8,14-secohopane series in a Chinese Boghead coal. Energy & Fuels 4, 177±183. Wang, Z., Fingas, M., Sergy, G., 1994. Study of 22-year-old Arrow oil samples using biomarker compounds by GC/MS. Environmental Science and Technology 28, 1733±1746. Wardroper, A.M.K., Ho€mann, C.F., Maxwell, J.R., Barwise, A.J.G., Goodwin, N.S., Park, P.J.D., 1984. Crude oil biodegradation under simulated and natural conditions-II (Aromatic steroid hydrocarbons). Organic Geochemistry 6, 605±617.