Pyrolytic and spectroscopic study of a sulphur-rich
kerogen from the ``Kashpir oil shales''
(Upper Jurassic, Russian platform)
A. Riboulleau
a,b, S. Derenne
a,*, G. Sarret
c,1, C. Largeau
a, F. Baudin
b,
J. Connan
daLaboratoire de Chimie Bioorganique et Organique Physique, CNRS UMR 7573, ENSCP, 11 rue Pierre et Marie Curie, 75231 Paris cedex 05,
France
bLaboratoire de Stratigraphie, CNRS ESA 7073, UPMC, 4 place Jussieu, 75252 Paris cedex 05, France cChemistry Department, University of Western Ontario, N6A 5B7, London, Ontario, Canada
dElf Aquitaine CSTJF, Avenue Larribau, 64018 Pau cedex, France
Abstract
The kerogen of an organic-rich sample, termed f top, from the Gorodische section (Russian platform) was studied using a combination of microscopic, spectroscopic and pyrolytic methods so as to examine its chemical structure, source organisms and formation pathway(s). This kerogen, which is mainly composed of orange gel-like, nanoscopi-cally amorphous organic matter, exhibits a relatively high aliphatic character; organic sulphur is mainly present as di(poly)sulphides and alkylsulphides. The f top kerogen was chie¯y formed via intermolecular incorporation of sulphur in algal or cyanobacterial lipids and carbohydrates. However, its formation also involved oxidative condensation via ether linkages. Comparison of f top sample with other S-rich kerogens points to a closer similarity with Monterey kerogens rather than with a kerogen from the bituminous laminites of Orbagnoux.#2000 Elsevier Science Ltd. All rights reserved.
Keywords:Kashpir oil shales; Pyrolysis; FTIR; Solid state13C NMR; XANES spectroscopy; Type II-S kerogen
1. Introduction
The Upper Jurassic was a period of intense accumu-lation of sedimentary organic matter (OM) and impor-tant gas and oil source-rocks are Upper Jurassic in age (Ulmishek and Klemme, 1990), especially in the north-ern hemisphere, including the North Sea and Siberian oil ®elds. Episodes of extensive OM deposition also took place on the Russian Platform, located between these two basins, but the relative stability of this plat-form since the Triassic did not allow sucient burial
of OM for hydrocarbon production. Although such episodes are relatively less numerous on the Russian platform than in the neighbouring basins, large accu-mulation occurred during the Middle Volgian (Late Tithonian, 140 Ma) and several basins of the Russian platform display organic-rich levels of this age (Shmur et al., 1983). In the Ulyanovsk region, these organic-rich sediments crop out along the Volga river and are known as the ``Kashpir oil shales''. At Gorodische (Fig. 1), this organic-rich deposit represents a 6 m thick layer of grey to dark-brown shales whose TOC contents vary between 0.5 and 45% and HI between 50 and 700 mg HC/g TOC (Hantzpergue et al., 1998). The Kashpir oil shales represent, after the Baltic oil shales, one of the most important oil shale reserves of Russia and they have been mined for oil production since the 1850s (Shmur et al., 1983; Russell, 1990). However, production from these oil shales has almost ceased today because they are
0146-6380/00/$ - see front matter#2000 Elsevier Science Ltd. All rights reserved. P I I : S 0 1 4 6 - 6 3 8 0 ( 0 0 ) 0 0 0 8 8 - 7
www.elsevier.nl/locate/orggeochem
* Corresponding author. Tel.: + 1-4427-6716; fax: + 33-1-4325-7975.
characterised by high sulphur contents, and particularly Sorg, leading to high atmospheric pollution upon reforming (Popov et al., 1986). Nevertheless these shales are still mined near Kashpir village for pharmaceutical and industrial purposes.
S-rich kerogens have been intensively studied during the last decade and it is now well established that this type of fossil OM originates from incorporation of reduced mineral sulphur (H2S, S0or S
X
2ÿ) in
functiona-lised lipids or carbohydrates during early diagenesis (e.g. Francois, 1987; review by Sinninghe Damste and de Leeuw, 1989; van Kaam-Peters et al., 1998a). S-rich OM is mostly found in sediments from evaporitic and anoxic carbonate platform environments (Sinninghe Damste et al., 1989; Schaeer et al., 1995; Mongenot et al., 1997; Baudin et al., 1999). Sulphur incorporation, from the H2S produced by sulphate-reducing bacteria under anoxic conditions, can occur during very early diagen-esis in the upper layers of the sediment (Hartgers et al., 1997).
Several paleoecological indices in the Gorodische black shales, like the presence of bioturbations and brachiopods in several levels, testify for oxygenation of the upper centimeters of the sediments, at least during part of the deposition of these black shales. The results presented here, concerning the most OM-rich level, termed ``f top'', are part of a broader study. This study aims at determining the paleoenvironmental conditions that led to the deposition of these black shales and explaining their strong variations in OM quality and
quantity. The kerogen of f top was examined by a com-bination of various methods in order to characterise its structure, to specify the nature of the sulphur functions and to determine the source organisms of the OM and its preservation pathway(s), in this especially rich level.
2. Geological setting and sampling
The Gorodische outcrop is located near the city of Ulyanovsk along the Volga river (Fig. 1). It is mainly composed of grey clays of Middle Volgian age (Fig. 2). At the top of the middle Volgian, appears the 6 m thick black shale unit. Its colour varies from dark grey to brown and despite a laminated feature, bioturbations are visible in the lower and upper parts of the unit. It is overlain by sandstone of Middle to Upper Volgian age. A more detailed description of the section is given in Hantzpergue et al. (1998). TOC and HI values for this organic-rich formation are presented in Fig. 2. The f top level which is studied in this paper, is a 7 cm thick bed located 2.70 m above the base of the OM-rich deposit and presents the highest TOC and HI values (Fig. 2). In contrast to most part of the formation, no bioturbation nor benthic organisms are observed in this level.
3. Experimental
Shale samples were collected in 1995. Subsamples were ground for Rock-Eval analysis, the remainder was stored at room temperature away from light. Prior to further analysis, the surface of the sample was carefully removed in order to eliminate oxidised or polluted material.
Rock-Eval pyrolysis, OSA device, was performed on 10 mg samples of powdered bulk shale. The sample was heated at 300C for 3 min followed by a programmed
pyrolysis at 25C/min up to 600C under a He ¯ow and
then oxidised at 600C for 7 min under an oxygen ¯ow.
After grinding, ca. 50 g of shale were extracted with CHCl3/MeOH, 2:1, v/v (stirring for 12 h at room tem-perature) before isolation of the kerogen via the classical HCl/HF treatment (Durand and Nicaise, 1980). The kerogen concentrate was then extracted as described above and dried under vacuum.
An aliquot of the kerogen was ®xed with 2% OsO4 for electron microscopy. For transmission electron microscopy (TEM), the ®xed kerogen was embedded in Araldite, cut in ultra-thin sections and stained with uranyl acetate and lead citrate. Observations were car-ried out with a Philips 300 microscope. For scanning electron microscopy (SEM), the ®xed kerogen was dehydrated using the CO2 critical point technique and coated with gold prior to observation with a Jeol 840 microscope.
FTIR spectra of the kerogen and its pyrolysis residues were recorded on a Bruker IFS 48 spectrometer as 5 mm KBr pellets. Solid state 13C NMR spectroscopy was performed on the kerogen and its pyrolysis residues with a Bruker MSL 400 spectrometer using high power decoupling, cross polarisation and magic angle spin-ning. Spectra were recorded at 3 and 4 kHz in order to discriminate spinning side bands (SSB).
X-ray absorption near edge structure (XANES) spec-troscopy was performed at the Canadian Synchrotron Radiation Facility situated on the 1 GeV electron sto-rage ring, Alladin, University of Wisconsin. Experi-mental details have been described previously (Kasrai et al., 1994). All the S K- and L-edge spectra presented in this paper were recorded using total yield (TEY) detec-tion mode. They were background subtracted using a linear function extrapolated from the pre-edge region, and then normalised to the height of the maximum of peak B for the L-edge, and to the height of the edge jump for the K-edge. S K-edge spectra were simulated by a linear combination of reference spectra using a least-squares ®tting program. Analytical procedure as well as interpretation and comparison with standard spectra are described in Sarret et al. (1999).
Elemental analyses of the kerogen and of its pyrolysis residues were performed at the Service Central d'Ana-lyse of the CNRS. The O content of the unheated kero-gen was determined by coulometry (Laboratoires Wol).
Bulk isotopic measurement ofd13C was performed on the kerogen with a Carlo-Erba CHN coupled to a VG-SIRA 10 spectrometer.
O-line pyrolysis was performed on the kerogen as previously described by Largeau et al. (1986). Brie¯y,
the sample is successively heated at 300C for 20 min
and 400C for 1 hour under an He ¯ow. The released
products are trapped in cold chloroform. After each treatment, the residue is extracted with CHCl3/MeOH as previously described. The ®rst thermal treatment at 300C was aimed at eliminating thermolabile
compo-nents. It has been recently demonstrated that such a thermal treatment, when applied to S-rich kerogens also promotes some aromatisation and lowers the global eciency of the subsequent cracking at 400C.
How-ever, it was also shown that more detailed information on the chemical structure of such kerogens can ®nally be obtained by this two step treatment at 300 and 400C
than via direct pyrolysis at 400C (Mongenot, 1998;
Sarret et al., unpublished results).
The 400C pyrolysate was separated by column
chro-matography (Al2O3, Act II) into three fractions of increasing polarity eluted with heptane, toluene and methanol, respectively. An additional elution with CHCl3was performed, and yielded a small amount of products which were combined with the MeOH-eluted fraction. Carboxylic acids were separated from the MeOH±CHCl3 fraction using a double extraction with ether under base and acid conditions and analysed by GC and GC/MS as their methyl esters. Unsaturated methyl esters were further analysed after derivatisation with DMDS as described by Scribe et al. (1988).
A part of the total 400C pyrolysate and of its
hep-tane- and toluene-eluted fractions was desulphurised with Raney Ni using the conditions previously described by Sinninghe Damste et al. (1988a) and hydrogenated before GC and GC/MS analyse.
All the fractions were analysed by GC and GC/MS using a HP 5890 gas chromatograph (60 m capillary
column, ®lm thickness of 0.4mm, heating program 100 to 300C at 4C minÿ1, injector and FID at 320C,
helium carrier gas), coupled with a HP 5989 mass spec-trometer with a mass range m/z= 40±600, operated at 70 eV.
Curie point pyrolysis-gas chromatography±mass spectrometry (CuPy-GC±MS) was performed on ca. 2 mg of sample. The sample was loaded on a ferromag-netic wire with a Curie point of 610C. The wire placed
in a glass tube was introduced in the Curie point pyr-olyser (Fisher 0316M) coupled to a Hewlett-Packard HP 5890 gas chromatograph with FID (30 m fused silica capillary column, ®lm thickness 0.4 mm, heating pro-gram 35C for 10 min and from 35 to 300C at 4C
minÿ1, He carrier gas). The chromatograph was coupled with a HP 5989 mass spectrometer with a mass range
m/z= 40±600, operated at 70 eV.
4. Results
4.1. Bulk features
Rock-Eval pyrolysis of the unextracted shale shows an exceptionally high TOC of 45% for f top and a high hydrogen index (HI) of 700 mg HC/g TOC (Fig. 2). The
Tmaxvalue is very low, 396C, re¯ecting the immaturity of the sample. Bulk isotopic measurement indicate a
d13CTOCof
ÿ20.5%(PDB) for f top kerogen. Elemental
analysis of the isolated kerogen (Table 1) indicate a relatively high H/C ratio of 1.39 consistent with the high HI, an O/C ratio of 0.20, a low N/C ratio of 0.014, a high content of total S of 11.7 wt.% and a weak Fe content of 0.67 wt.%. Sorgcontent was then calculated in the clas-sical way: it was considered that Fe is only present as pyrite after the acid treatment and that non-pyritic sul-phur is organic, giving a high Sorg/C ratio of 0.068. F top thus belongs to Type II-S kerogens, according to the classi®cation of Orr (1986). Sulphur-rich kerogens are known to exhibit a relatively low thermal stability (Baskin and Peters, 1992; Tomic et al., 1995) thus explaining the very lowTmaxvalue observed for f top.
4.2. Microscopic study
Under the light microscope, f top kerogen appears chie¯y constituted of a single type of amorphous
organic matter (AOM) associated with rare pyrite. This AOM is yellow to orange and appears as gel-like parti-cles of varying size (from 5 to 200mm). When observed by SEM and TEM, the AOM appears homogeneous and amorphous, even at high magni®cation (nanometric scale). Similar morphological features have been pre-viously observed in S-rich kerogens from the Kimmer-idge Clay Formation (Boussa®r et al., 1995) and the paleolagoon of Orbagnoux (Mongenot et al., 1999).
4.3. Spectroscopic study
4.3.1. Solid state13C NMR spectroscopy
The solid state13C NMR spectrum of f top kerogen (Fig. 3a) is dominated by a broad peak at 30 ppm with a small shoulder at 15 ppm, corresponding to CH2 in alkyl chains and to CH3 groups, respectively. Such a dominance re¯ects the aliphatic character of the sample, in agreement with both high H/C ratio and HI value. A relatively intense broad signal centred at 75 ppm corre-sponding to aliphatic C linked to N or O is also noticed. According to the weak N/C ratio revealed by elemental analysis the latter signal must be chie¯y due to C±O bonds. The 30 ppm peak exhibits a broad shoulder between 40 and 60 ppm, which can be assigned to car-bons in C±S bonds and/or to Cbto the functions that gave the 75 ppm signal. Two weak signals at 130 and 175 ppm are observed corresponding to unsaturated carbons and to C in carboxylic groups, respectively. Signals observed at 140 and 215 ppm correspond to spinning side bands of the carboxylic peak as shown by the comparison of the spectra recorded at two dierent spinning rates.
Table 1
Atomic ratios of f top kerogen and insoluble pyrolysis residues
H/C Sorg/C N/C
f top 1.39 0.068 0.014
f top res 300 1.19 0.047 0.016
f top res 400 0.57 0.030 0.021
Fig. 3. Solid state CP/MAS 13C NMR spectra of (a) f top kerogen and (b) of its 300C pyrolysis residues. (X, O and/or
4.3.2. FTIR spectroscopy
The FTIR spectrum of f top kerogen (Fig. 4a) shows intense absorptions at 2920, 2850, 1445 and 1375 cmÿ1 due to CH2and CH3groups. Oxygenated functions are detected as (i) a broad band of medium intensity centred at 3400 cmÿ1corresponding to O±H groups, (ii) a band at 1700 cmÿ1(C
O), and (iii) a broad and intense band
centred at 1050 cmÿ1corresponding to C±O bonds, in
agreement with the relatively strong signal at 75 ppm observed in the 13C NMR spectrum. A band at 1630 cmÿ1is also observed indicating that the unsaturated C detected at 130 ppm via13C NMR are mostly ole®nic. A narrow band with a rather high intensity is noted at 750 cmÿ1; it is due to CaF2neoformed during the HCl/HF treatment since the latter mineral was identi®ed by XRD (A. Person, pers. comm.).
4.3.3. XANES spectroscopy
Sulphur K-edge spectrum for f top kerogen is pre-sented in Fig. 5a. The spectrum was simulated by a lin-ear combination of spectra of model compounds in order to quantify the dierent sulphur species which can be dierentiated by this mode, i.e. Sÿ2, disulphides, alkyl and/or heterocyclic sulphides, sulphoxides, sul-phones, sulphonates and sulphates (Sarret et al., 1999). Results of the simulation are presented in Table 2. The major sulphur species in f top kerogen are alkyl and/or heterocyclic sulphides (76%) and di(poly)sulphides occur in lower proportion (11%). A small contribution of sulphoxides, sulphonates and sulphates is also noticed.
S L-edge spectrum for f top kerogen is compared with spectra of various reference compounds in Fig. 6. From S K-edge results, one could foresee spectral similarities with the alkylsulphide and/or heterocyclic sulphide (thiophene) references. These two types of compounds, which are dicult to distinguish on K-edge spectra, exhibit clearly dierent peak positions on L-edge spectra,
Fig. 4. FTIR spectra of (a) f top kerogen and of its (b) 300 and (c) 400C pyrolysis residues.
as this latter method is more sensitive to the reduced forms of sulphur (Kasrai et al., 1996a). F top spectrum is clearly closer to the alkyl sulphide reference (Fig. 6-2) than to the thiophene (Fig. 6-5) and dibenzothiophene (Fig. 6-4) ones; it is therefore concluded that alkyl sul-phides are the major sulphur species in f top kerogen. However, f top spectrum is slightly shifted to the left compared to the reference and presents a small shoulder at 162.8 eV. This spectrum is well reproduced by a linear combination of 30% di(poly)sulphides+70% alkyl sul-phides (Fig. 6a, dotted line). Contrary to the K-edge mode, S L-edge XANES spectroscopy does not allow a precise quanti®cation of sulphur species due to (i) some uncertainties on the normalisation of the spectra, and (ii) the fact that only one part of the spectrum is simu-lated (Sarret et al., 1999). Accordingly, S L-edge results are not directly comparable with K-edge results, e.g. concerning the di(poly)sulphide content (30% compared to 11%, respectively). However, S L-edge analysis allows to conclude that the major reduced sulphur spe-cies correspond to alkyl sulphides, and to a lesser extent di(poly)sulphides. The f top spectrum also contains a small peak at about 167.6 eV, which corresponds to the position of the main peak for the sulphoxide reference (Fig. 6-6), which is consistent with K-edge results.
4.4. Pyrolytic study
4.4.1. Mass balance and spectroscopic features of o-line pyrolysis residues
After the thermal treatment at 300C, the weight loss
amounts to ca. 18% of the initial organic matter, mainly corresponding to volatile compounds (17%) while trap-ped products only represent 1% of the initial OM. After pyrolysis at 400C, the total weight loss represents 58%
of the initial OM. This loss corresponds in similar amount to volatile and trapped products (ca. 30 and 28% of the initial OM, respectively).
Elemental analyses of the pyrolysis insoluble residues (Table 1) show, as expected, a small decrease in the H/C ratio upon 300C heating whereas a large decrease takes
place upon pyrolysis at 400C. The N/C ratio shows a
slight increase upon pyrolysis which is consistent with
the results of Barth et al. (1996) and of Gillaizeau et al. (1997) attesting for the concentration of N in kerogen residues after pyrolysis. In contrast, the Sorg/C ratio regularly decreases.
The 13C NMR spectrum of the 300C residue (Fig.
3b) is still dominated by the peak centred at 30 ppm but the relative intensity of the 40±60 ppm shoulder is lower compared to the unheated kerogen. The 75 ppm peak exhibits a lower intensity and becomes thinner; in the same time, the peak at 130 ppm increases and the car-boxyl peak at 175 ppm remains approximately constant with respect to the 30 ppm band. The important peak at 165 ppm corresponds to a spinning side band of the 130 ppm peak. As expected, the13C NMR spectrum of the 400C residue (not shown) is dominated by an aromatic
signal at 130 ppm and only shows a weak signal at 30 ppm. This indicates the intense aromatisation of the material.
The FTIR spectrum of the 300C residue (Fig. 4b)
shows the same absorptions as the unheated kerogen. The shift of the CC signal towards 1600 cmÿ1 indi-cates some aromatisation at this temperature as pre-viously observed for the Orbagnoux kerogen (Mongenot, 1998). The C±O absorption decreases shar-ply at 300C, in agreement with the decrease of the 75
ppm signal in the 13C NMR spectrum. In the 400C
residue (Fig. 4c) the absorptions corresponding to CH2 and CH3are of much lower intensity in agreement with the large decrease of the H/C ratio, the CC band is broader and now centred at 1600 cmÿ1, corresponding to aromatic carbons and the broad band around 1100 cmÿ1, due to C±O bonds, is no longer detected.
The S K- and L-edge XANES spectra for the inso-luble pyrolysis residues are presented in Figs. 5b and c and 6b and c, respectively. The edge of the K-edge spectrum for the 300C residue is slightly shifted to
higher energy compared to the unheated kerogen. This shift corresponds to the removal of di(poly)sulphides upon heating. Indeed, the simulation shows that the di(poly)sulphides are no longer present in the 300C
residue (Table 2). The edge of the spectrum for the 400C residue is even more shifted to the right. As the
di(poly)sulphides are already eliminated in the 300C Table 2
Distribution of the sulphur species in f top kerogen and insoluble pyrolysis residues determined by simulation of the S K-edge XANES spectra
FMa S(-II) Disulphides S aliph./heterocycl.b Sulphoxides Sulphones Sulphonates Sulphates
f top 0.08 0 11 76 5 0 4 4
f top res 300 0.05 0 0 94 5 0 0 1
f top res 400 0.18 4 0 84 9 0 0 3
a FM: ®gure of merit of the ®t, FM= S(f
kerogen-f®t)2
residue, this shift can only be attributed to a change in the alkyl/heterocyclic sulphides distribution. S L-edge analysis will provide more information about this change. Concerning the oxidised sulphur species, the oscillation situated at about 2481 eV is weaker for the 300C residue than for f top, so their contribution is
smaller. Indeed, Table 2 shows that their content is of 6 and 13%, respectively. The 400C residue contains
about the same total amount of oxidised species as f top (12%), but more sulphoxides. This latter sample also contains 4% of sulphur in the oxidation state (-II), which gives rise to a small shoulder at 2469 eV. This oxidation state does not exist for organic sulphides, whose oxidation number is 0. Therefore, the 400C
residue contains a small amount of inorganic sulphides that has formed upon heating.
The S L-edge spectra for the residues are compared to f top spectrum in Fig. 6. The spectrum of the 300C
residue is shifted to higher energy compared to f top, and its three peaks roughly match with those of the reference corresponding to simple thiophene. However, the amplitude of the peak at 164.7 eV is particularly high, which indicates the presence of alkyl sulphides. A good ®t is obtained with a linear combination of 60% simple thiophene+40% alkyl sulphide. If we replace simple thiophene (i.e. not included in a polyaromatic structure) by dibenzothiophene (DBT), the position of the 164.7 eV peak does not match so well. Thus, the 300C residue contains thiophenes as major species, and
to a lesser extent alkyl sulphides. No di(poly)sulphides are detected in this sample, which is in agreement with K-edge results. For the 400C residue, the L-edge
spec-trum shows peak positions similar to that of the 300C
residue, but with dierent relative intensities. Alkylsul-phides are no longer observed and the linear combina-tion of 70% thiophene+30% sulphoxide aords a good ®t. Again, if we replace simple thiophene by DBT, the peak at 164.6 eV does not match. As explained pre-viously, the percentages determined on S L-edge spectra do not aord precise quantitative information on sul-phur distribution; however they re¯ect the changes in disulphide, alkyl sulphide and thiophene contents upon thermal stress. During heating, the disulphides (300C)
and the sulphides (400C) present in f top disappear
while the relative abundance of thiophenes increases. These trends are consistent with the decrease of the 40± 60 ppm shoulder and relative increase of the 130 ppm signal corresponding to aromatic carbons in the 13C NMR spectrum of the 300 and 400C residues,
com-pared to that of the kerogen. Such changes are also responsible for the shifts observed on the K-edge spec-trum. It can be noted that the sulphoxide contribution for the 400C residue, calculated by S L-edge
spectro-scopy, is particularly high compared to its actual con-tent determined via K-edge analysis (9%). This disagreement can be due to a higher oxidation of the
Fig. 6. S L-edge XANES spectra for some reference com-pounds: (1)dl-cystine, (2)dl-methionine, (3) poly(phenylene
sulphide), (4) dibenzothiophene, (5) 3-(2-thienyl)-dl-alanine,
(6)dl-methionine sulphoxide. S L-edge XANES spectra (solid lines) and simulations (dotted lines) for (a) f top kerogen and its (b) 300 and (c) 400C pyrolysis residues. Simulations consist of
a linear combination of the reference compounds spectra pre-sented above. These combinations are the following: f top: 30% (1)+70% (2), 300C residue: 40% (2)+60% (5), 400C residue:
surface of the samples, as the sampling depths of S K-and L-edge XANES spectroscopy are dierent. Indeed, using TEY detection mode, it is about 70 nm for the K-edge compared to only 5 nm for S L- K-edge (Kasrai et al., 1996b). Similar features have been previously observed on asphaltene and kerogen samples (Sarret et al., 1999; Sarret et al., unpublished results).
4.4.2. O-line pyrolysis products
The GC trace of the 400C pyrolysate is very complex
and shows an important hump, due to the coelutions of numerous products as commonly observed for pyr-olysates of S-rich kerogens (Sinninghe Damste et al., 1990). In order to make easier GC/MS identi®cations, the pyrolysate was fractionated into three fractions of increasing polarity, eluted with heptane, toluene and methanol, respectively.
4.4.2.1. Heptane-eluted fraction.This fraction represents 17% of the pyrolysate. Its GC trace (Fig. 7a) is still complex and exhibits an important hump due to numerous coelutions. Nevertheless, selective ion detec-tion (SID) of characteristic fragments allowed the iden-ti®cation of a number of compounds and homologous series (Table 3). However, due to coelutions, their
rela-tive abundance could not be determined and only a rough estimation of their relative intensities can be given. Hydrocarbons mainly consist ofn-alkane1/n -alk-1-ene2doublets up to C33(maximum C17) without any odd- or even- carbon-number predominance. n -Alkyl-cyclohexanes 9 and cyclopentanes 10 are observed in small amount. The former are frequently reported in kerogen pyrolysis products (Homann et al., 1987), however, their origin is still a matter of debate.
n-Alkylbenzene5/n-alkenylbenzene 7doublets are also observed, associated with the three n -alkylmethylben-zene isomers6, theo-isomer being predominant. It can be noticed that the latter isomer is the only one which can be formed by cyclisation of linear compounds. Polysubstituted alkylbenzenes8are also observed in low amount. Isoprenoid compounds correspond to prist-1-ene3and regular saturated C16and C18hydrocarbons
4, but their abundance relative ton-alkanes could not be determined. Neither pristane, nor phytane was detected, possibly because of a too low abundance. The other isoprenoid compounds identi®ed are hopanes and hopenes28from C27to C31. These compounds are gen-erally considered to be of bacterial origin (Rohmer et al., 1984).
Numerous series of organic sulphur compounds (OSC) are identi®ed and their coelution accounts for the bulk of the hump. Several isomers of alkylated thio-phenes 11-14, thiolanes 15, thianes 16 and benzothio-phenes 19 are identi®ed. These compounds are major components of most pyrolysates of S-rich kerogens (Sinninghe Damste et al, 1988b; Payzant et al., 1989). Most of the identi®ed series of OSC have a linear skele-ton. However, C10 to C16 2,3-dimethyl-5-n -alkylthio-phenes13 [identi®ed by their mass spectra and elution time from Sinninghe Damste et al. (1989)], are also observed along with branched alkylthiophenes14 from C15to C17. Dierent series of polyaromatic OSC pre-viously observed by van Kaam-Peters and Sinninghe Damste (1997) and van Kaam-Peters et al. (1998b) in the pyrolysate of S-rich kerogen from the paleolagoon of Orbagnoux, are also detected: C9 to C18 alkylated bithiophenes 17±18 (Appendix I, characteristic frag-ments at m/z= 179, 193 and 207) and C10 to C13 n -alkylphenylthiophenes20(II, characteristic fragments at
m/z= 173 and 187). Other series of compounds pre-viously detected in the pyrolysate of Orbagnoux kero-gen are also observed in f top pyrolysate: series26and
27characterised by intense fragments atm/z=229±230 and 243±244 (van Kaam-Peters and Sinninghe DamsteÂ, 1997; van Kaam-Peters et al., 1998b; Mongenot et al., 1999), and series 24 and 25 characterised by intense fragments at m/z=203 and 217, respectively, (Mon-genot et al., 1999). As these compounds are also present in the toluene-eluted fraction (Table 4), the assignment of these dierent series is discussed later on. Two series
22, not reported so far, characterised by intense
frag-Fig. 7. TIC of the heptane-eluted fraction of the 400C
ments at m/z= 167 and 181 and a molecular ion at 168+14n and 182+14n(n from 1 to 10), respectively, are also observed. The regular distribution pattern of these two series (e.g. Fig. 8a) points to the presence of a
n-alkyl side chain and their disappearance after desul-phurisation indicate that they correspond to OSC. These compounds are tentatively identi®ed as n -alkyl-methyl- and n-alkyldimethylthienothiophenes (III) on the basis of their mass spectra (e.g. Fig. 8b). Thie-nothiophenes have already been observed in crude oils (Orr and Sinninghe DamsteÂ, 1989) but, as far as we are aware, alkylated homologues of such compounds have not been reported so far.
Following Raney Ni desulphurisation and hydro-genation, the heptane fraction is dominated by a series of
n-alkanes (Fig. 7b), thus con®rming that most of the OSC have a linear skeleton. Nevertheless, branched and
isoprenoid compounds are also present in small amounts. In particular, pristane and phytane are observed, which were not detected in the non-desul-phurised fraction. Pristane must be directly derived from hydrogenation of prist-1-ene. Phytane can originate from hydrogenation of phytenes and/or desulphurisa-tion of C20isoprenoid OSCs. The latter have been com-monly observed in extracts and kerogen pyrolysates (Sinninghe Damste and de Leeuw, 1987, and references therein). Based on previously published mass spectra and elution times (Sinninghe Damste et al., 1986; Sin-ninghe Damste and de Leeuw, 1987), C20 isoprenoid OSCs were searched for in the untreated fraction but none could be detected. Therefore, it is possible that several C20 isoprenoid OSCs are present but that they are undetectable amongst the coelution hump due to too low individual abundances and/or that phytane
Table 3
Compounds detected in the heptane-eluted fraction of the 400C pyrolysate of f top kerogen
Series Range Maximum
1n-Alkanes C13±C35 C17
2n-Alk-1-enesa C
13±C33 C16
3Prist-1-ene C19 C19
4Regular isoprenoid alkanes C16, C18 ±
5n-Alkylbenzenes C12±C32 C15
6n-Alkylmethylbenzenesb C
12±C33 C14
7on-Alkenylbenzenes C13±C29 C15±C16
8Substitutedn-alkylbenzenesb C
12±C26 C14
9n-Alkylcyclohexanes C13±C26 C17
10n-Alkylcyclopentanes C13±C27 C16
112n-Alkylthiophenes C10±C28 C12
122,5-Di-n-alkylthiophenesc C
10±C28 C13
132,3-Dimethyl-5-n-alkylthiophenes C10±C16 C13
14Branched alkylthiophenesb C
15±C17 ±
152-n-Alkylthiolanesd C
9±C25 C12
162-n-Alkylthianes C10±C21 C13
172-n-Alkyl-5,50bithiophenes C
9±C18 C10
182,20-Di-n-alkyl-5,50-bithiophenes C
10±C18 C11±C12
19n-Alkylbenzo[b]thiophenese C
9±C22 C10
20n-Alkylphenylthiophenesf C
10±C13 C11
21n-Alkyldibenzo- or naphtothiophenes C12±C15 C12±C13
22n-Alkylmethylthienothiophenesg C
8±C15 C9
23Compound in 190 C10 C10
24, 25Series in 203±217b C
11±C19 C11±C12
26, 27Series in 229±243b C
13±C18 C13±C14
28Hopanes and hopenes C27±C31 ±
a A series ofn-alk-2-enes was also identi®ed with the same range and a maximum at C 18. b Several isomers were detected in the series.
c Three series were detected: C
10±C30(max C13) 2-n-alkyl-5-methylthiophenes, C10±C29(max C12) 2-n-alkyl-5-ethylthiophenes and C11±C25(max C12) 2-n-alkyl-5-propylthiophenes.
d 2-n-Alkyl-5-methylthiolanes (C
10±C21, max C13) were also observed.
e Two series (2-n-alkyl- and 4-n-alkylbenzo[b]thiophenes) were observed with the same distribution. Two series ofn- alkylmethyl-benzo[b]thiophenes (C10±C20, max C11) were also observed.
f A series of C
12±C15(max C13)n-alkylmethylphenylthiophenes was also observed. g A series of C
skeletons are involved through S-linkage in non-polar, non-GC-amenable, high molecular weight pyrolysis products, so that they are not observed in the heptane-eluted fraction but released after desulphurisation.
Three series of branched alkanes (iso-, anteiso- and 4-methylalkanes) occur in trace amount. Branched hydrocarbon skeletons are generally considered to re¯ect a bacterial input (Albro, 1976; Shiea et al., 1990).
4.4.2.2. Toluene-eluted fraction. This fraction, which represents 19% of the total pyrolysate, also appears highly complex and its GC trace shows an important hump (Fig. 9). Nevertheless, as in the case of the hep-tane-eluted fraction, numerous series of compounds were identi®ed via selective detection of characteristic ions. Identi®ed compounds, listed in Table 4, can be subdivided into two major groups: ketones and OSC. Some polyaromatic compounds34±37are also detected in low amounts.
Ketones and especiallyn-alkan-2-ones often occur in kerogen pyrolysates where they are supposed to be derived from the thermal cleavage of ether bonds (van de Meent et al., 1980; Largeau et al., 1986). Several ser-ies of n-alkanones 29 were identi®ed in the toluene fraction of f top pyrolysate as shown by the ion chro-matogram atm/z=58 (Fig. 10). They comprise series of mid-chain n-alkanones with dierent locations of the keto group, from C(3) to C(12) (Fig. 11). The occur-rence of so many series of mid-chain ketones has been rarely reported. As suggested by Gillaizeau et al. (1996) in a study concerned with the kerogen of the GoÈynuÈk oil shale, such feature should re¯ect the presence of ether linkages at various locations of the alkyl chain. The wide range of location of the ether links in f top kerogen is similar to recent results of Jenisch-Anton et al. (1999)
Table 4
Compounds detected in the toluene-eluted fraction of the 400C pyrolysate of f top kerogen
Series Range Maximum
29n-Alkan-2-onesa C
10±C29 C13
30n-Alkylcyclopentanones C11±C24 C13
31n-Alkylcyclohexanones C11±C18 C14
321-Phenyl-n-alkan-1-ones C10±C16 C12
331-Phenyl-n-alkan-2-onesb C
9±C29 C12
34Alkyl¯uorenesc C
13±C16 C14
35Alkylanthracenes or phenanthrenesc C
14±C17 ±
36Alkylpyrenes C14±C15 ±
37Alkylbenzo¯uorenes C12±C13 ±
172-n-Alkyl-5,50-bithiophenesc C
9±C11 C10
20n-Alkylphenylthiophenesc C
11±C15 C12
19n-Alkylbenzo[b]thiophenes C9±C12 C12
21n-Alkyldibenzo- or naphtothiophenesc C
12±C15 ±
23Compound in 190 C10 C10
24, 25Series in 203±217 C11±C14 ±
26, 27Series in 229±243 C13±C15 ±
a Other series of ketones were also observed:n-alkan-3-ones, distr.: C
11±C26(C14);n-alkan-4-ones, C11±C25(C13);n-alkan-5-ones, C11±C27 (C16, C19); n-alkan-6-ones, C12±C21 (C13); n-alkan-7-ones, C13±C24 (C16); n-alkan-8-ones, C15±C26 (C17); n-alkan-9-ones, C17±C26(C18);n-alkan-10-ones, C19±C24(C19);n-alkan-11-ones, C21±C26(C21);n-alkan-12-ones, C23±C26(C23).
b Three series were observed with the same distribution: 1-phenyl-n-alkan-2-ones and two 1-(methylphenyl)-n-alkan-1-ones. c Numerous isomers were detected in the series.
Fig. 8. (a) Ion chromatogram at m/z=167 of the heptane-eluted fraction of the 400C pyrolysate of f top kerogen;^,
who reported data for two oils (Marvejols and Rozel Point) and a kerogen (Gibellina), all S-rich, showing the presence of ether links at dierent locations along alkyl chains. Cyclic ketones (cyclopentanones 30 and cyclo-hexanones 31) are minor constituents, while two series of phenyl ketones (1-phenyl-n-alkan-1-ones 32 and 1-phenyl-n-alkan-2-ones33, structuresIV andV, respec-tively) are present in signi®cant amount. Their origin, however, is unknown.
Most of the OSC present in this fraction have already been observed in the heptane fraction and they mainly consist of short-chain polycyclic compounds. Among them, bithiophenes 17, benzo[b]thiophenes 19, phe-nylthiophenes 20, and dibenzothiophenes or naph-tothiophenes 21. The four series characterised by the ions 203 (24), 217 (25), 229 (26) and 243 (27) previously
observed in the heptane-eluted fraction are also present. The occurrence of 1,2-di-n-alkylbenzenes in the desul-phurised fraction is consistent with the thie-nylbenzothiophene (VI) structure proposed by van Kaam-Peters and Sinninghe Damste (1997) and van Kaam-Peters et al. (1998b) for series26and27. Such a structure, however, cannot be considered for series24
and 25. Mongenot et al. (1999) observed series with similar mass fragmentation in the pyrolysate of the S-rich kerogen from the Kimmeridgian paleolagoon of Orbagnoux. On the basis of mass spectra, desulphurisa-tion products and presence of thiochromans in the extracts of Orbagnoux (van Kaam-Peters and Sinninghe DamsteÂ, 1997; van Kaam-Peters et al., 1998b), Mon-genot et al. (1999) tentatively identi®ed these com-pounds as tetramethylthiochromenes (VII). So far, thiochromenes have not been ®rmly identi®ed in rock extracts, crude oils or kerogen pyrolysates. However, their saturated counterparts, thiochromans, have been observed in Oligocene crude oils and rock extracts (Sin-ninghe Damste et al., 1987; Adam, 1991; Schaeer, 1993; van Kaam-Peters and Sinninghe DamsteÂ, 1997;
Fig. 10. Ion chromatogram atm/z=58 of the toluene-eluted fraction of the 400C pyrolysate of f top kerogen (!,n
-alkan-2-ones29;, mid-chainn-alkanones).
Fig. 11. Partial mass fragmentograms atm/z= 58, 72, 86, 85, 99, 113 and 127 revealing the presence of C16n-alkan-2- (! ),-3-,-4-,-5-,-6-,-7- and-8-ones (*), respectively. Note thatn -alkan-2-to -4-ones are characterised by an even fragment (58, 72 and 86, respectively) due to a McLaerty rearrangement, while n -alkan-5- to -8-ones are characterised by an odd fragment due to acleavage (85, 99, 113 and 127, respectively) (, mid-chainn -alkanones).
Fig. 9. TIC of the toluene-eluted fraction of the 400C
van Kaam-Peters et al., 1998b) and are formed by sub-stitution of the O atom in the chroman structure by a sulphur (Adam, 1991). A compound characterised by a major peak at 203 was previously observed by Adam (1991) in extracts of Oligocene samples and crude oil and an isoprenoid trimethylbenzo[b]thiophene structure (VIII) was considered for this compound. Nevertheless, the regular pattern of the 203 and 217 series,24and25, observed in the heptane-eluted fraction of f top pyr-olysate (Fig. 12a and c) points to the occurrence of ann -alkyl side chain and not of an isoprenoid one in the present case. Consequently, this pattern is consistent, neither with the structure proposed by Adam (1991), nor with a thiochromene structure. In contrast, as dis-cussed below, these structures should correspond to benzodithiophenes (IX) or thienobenzothiophenes (X). Indeed, a compound characterised by an intense mass peak at 190 (23) is observed in the heptane and toluene fractions of f top pyrolysate which retention time indicates that it can correspond to the lowest homologue of the series24 and25(Fig. 12a±d). Com-parison with mass spectra of reference compounds showed that 23 can correspond to benzodithiophene (IX) or thienobenzothiophene (X). The addition of an -alkyl side chain to one of these structures would give a series of compounds characterised in MS by a fragment at 203, and the addition of two n-alkyl side chains would give a series of compounds characterised in MS by a fragment at 217. The presence of di-n -alkylben-zenes in the desuphurisation products of the heptane-and toluene-eluted fractions is consistent with this hypothesis.
After desulphurisation, the toluene-eluted fraction still shows an important hump, with only a few resolved peaks corresponding ton-alkanes. The dierent series of ketones observed before desulphurisation are still detected but the mass fragmentogram at m/z=58 indi-cates that mid-chain linear ketones are relatively more abundant with respect to n-alkan-2-ones than in the non-desulphurised fraction. It therefore appears that some ketones are also linked via S-bonds in the pyr-olysate of f top kerogen. Such pyrolysis products should correspond to moieties which were linked both by sul-phur and ether bonds in the macromolecular structure of the kerogen. Similar interpretations were previously obtained by Richnow et al. (1992) concerning the mac-romolecular structure of an oil and a kerogen from the Monterey Formation.
4.4.2.3. Methanol/chloroform-eluted fraction. This frac-tion which represents 41% of the pyrolysate was sepa-rated into an acid and a non-acid subfraction and the former was esteri®ed by MeOH/MeCOCl prior to GC/ MS analysis (Table 5).
The esteri®ed acid subfraction is dominated by methylesters of saturated fatty acids38from C12to C30
with a strong predominance of even-carbon-numbered compounds (CPI=0.18) (Fig. 13). The main compo-nents are palmitic acid, n-C16 and stearic acid, n-C18,
Fig. 12. (a) Ion chromatogram atm/z=190+203 of the heptane-eluted fraction of the 400C pyrolysate of f top kerogen; &,
which is a very common feature in kerogen pyrolysates (e.g. Kawamura et al., 1986; Largeau et al., 1986). Long-chain, C20+, fatty acids occur in signi®cant amount, they represent 17% of the saturated fatty acids. These C20+acids are generally considered of terrestrial origin (Volkman et al., 1980; Barouxis et al., 1988). However, long chain fatty acids have been observed in certain algae (e.g. diatoms; Volkman et al., 1980) and a bacterial origin has also recently been considered (Gong and Hollander, 1997). Unsaturated fatty acids 39±40, identi®ed after DMDS derivatisation (Fig. 13, inset), are dominated by oleic acid (C18:1o9) and a C16 mono-unsaturated acid (C16:1o10). Oleic acid is ubiquitous but would chie¯y be of phytoplanktonic origin, whereas the C16:1o10acid is considered as a bacterial marker (Bar-ouxis et al., 1988). Other unsaturated acids are also observed in lower amount, such as C18:2 40, which is also common in green microalgae (Weete, 1976). The presence of these unsaturated compounds, known to be highly sensitive to degradation, attests for a rapid and early incorporation of lipidic moieties in the kerogen. Iso- and anteiso- branched acids41are observed in low amounts, re¯ecting bacterial input (Perry et al., 1979; Goossens et al., 1986).
The non-acid subfraction is dominated by two linear, saturated, C16 and C18primary alcohols as previously observed by Mongenot et al. (1999) in the same sub-fraction of Orbagnoux pyrolysate. However, contrary to the latter study, no unsaturated alcohol is observed in f top pyrolysate. Series ofn-alkylphenols from C9to C16 andn-alkoxyphenols from C10to C15are also observed. Short chain alkyl phenols (C1±C3), derived from lignin and/or melanoidins (Saiz-Jimenez and de Leeuw, 1986; Zegouagh et al., 1999) are not detected in the present case. Long chain n-alkylphenols have previously been observed in the pyrolysates of various marine kerogens: two samples from the Kimmeridge Clay Formation (Gelin et al., 1995), a Cenomanian black shale from Central Italy (Salmon et al., 1997) and Ordovician Kukersite from Estonia (Derenne et al., 1990). Kuker-site is known to derive from the selective preservation of the cell walls from a colonial microorganisms, termed
Gloeocapsomorpha prisca which were shown to be the
phenol source (Derenne et al., 1990, 1992a). In contrast, no precise source could be attributed for the long-chain
n-alkylphenols in the Cenomanian black shale (Salmon et al., 1997), in the Kimmeridge Clay samples (Gelin et al., 1995) and in the present case as well. Some alkyl-thiophenes from C11to C15and even-carbon-numbered branched alkanes from C16to C30 (3-methyl- and 2,2-dimethylalkanes) are also present in low amount in the methanol-eluted, non-acid, subfraction. Due to their low polarity, such compounds should not appear in this subfraction and are considered to re¯ect thermal degra-dation, during GC/MS injection, of high molecular weight OSCs present in the subfraction. A similar fea-ture was previously observed by Mongenot et al. (1999) in the case of Orbagnoux kerogen.
4.4.2.4. Desulphurized pyrolysate. Mongenot et al. (1999) recently observed that, in the case of Orbagnoux kerogen, desulphurisation of the total pyrolysate yields hydrocarbons, the distribution of which is very dierent to that expected on the basis of analytical data from column-eluted fractions and desulphurized counter-parts. Such a discrepancy is mainly due to the high amount of polar and/or high molecular weight com-pounds retained on the alumina column for the Orbag-noux pyrolysate (ca. 55%). Although the amount of retained compounds is less here (23%), the total pyr-olysate of f top was desulphurised in order to determine if these polar and/or high molecular weight compounds have a speci®c signature.
The GC of the desulphurised pyrolysate is dominated by a series ofn-alkanes from C12to C31, similar to the
n-alkanes observed in the heptane-eluted and desulphurised
Table 5
Compounds detected in the methanol-eluted acid subfraction of the 400C pyrolysate of f top kerogen, tr: trace amount
Series Range Max. Rel.
intensity
38Saturated fatty acids C12±C30 C16 1
39Monounsaturated fatty acids C14±C18 C18 0.3
40Diunsaturated fatty acids C18 C18 tr
41Branched fatty acidsa C
15±C17 ± 0.02 a Anteiso C
15and C17, iso C16.
Fig. 13. TIC of the esteri®ed methanol-eluted acid subfraction of the 400C pyrolysate of f top kerogen:^, methyl esters of
toluene-eluted fractions. The GC trace also contains a small hump of numerous coeluting products, which all are observed in the column-eluted fractions with similar distributions. It therefore appears that in f top pyr-olysate, column-eluted and column-retained compounds have probably the same type of structure, and hence the same precursors, and the latter mostly dier by higher molecular weights.
4.4.3. Flash pyrolysis
Recent isotopic and pyrolytic studies on the Kim-meridge Clay Formation (van Kaam-Peters et al., 1998a,b; Sinninghe Damste et al., 1998) indicated that short chain alkylthiophenes liberated upon ¯ash pyr-olysis of S-rich kerogens can originate from moieties corresponding to sulphurised carbohydrates in these kerogens. Flash pyrolysis was therefore performed on f top kerogen in order to determine the distribution of short chain alkylthiophenes and assess the contribution of sulphurised carbohydrates.
The ¯ash pyrogram of f top (not shown) is dominated by short chain alkylthiophenes, the distribution of which is shown in Fig. 14. This distribution is domi-nated by compounds with a linear skeleton, and is close to the distribution obtained by pyrolysis of sulphurised carbohydrate-containing kerogens and sulphurised algae (Sinninghe Damste et al., 1998). In particular, it is very similar to that observed from the organic-rich Blackstone Band from the Kimmeridge Clay Formation (Sinninghe Damste et al., 1998; van Kaam-Peters et al., 1998a,b).
5. Discussion
5.1. Chemical structure and source organisms of f top kerogen
Elemental analysis and spectroscopic (FTIR and solid state 13C NMR) features pointed to a pronounced aliphatic character for the S-rich kerogen isolated from an especially organic-rich level, f top, from the Gor-odische outcrop of the Kashpir oil shales. Although a rather low amount of n-alkanes and n-alk-1-enes is released upon pyrolysis of this kerogen, the occurrence of series ofn-alkanones and, above all, of OSC with a linear skeleton con®rmed this aliphatic character. XANES spectroscopy revealed that, in f top, sulphur mainly occurs as alkylsulphides and, in lesser amount, as disulphides. It thus appears that the chemical struc-ture of f top comprises long n-alkyl chains linked by ether and sulphide bridges. A similar structure was recently suggested, by selective degradation studies for kerogens from the Monterey Formation (Richnow et al., 1992) and Gessoso-Sol®fera (Schaeer-Reiss et al., 1998). Along with these ether and sulphide bridges,
some ester functions re¯ecting the rapid incorporation of weakly altered fatty acids within the macromolecular network, also occur in f top.
The sharp predominance of n-alkyl skeletons over branched ones observed in the o-line pyrolysate is consistent with a major algal or cyanobacterial origin of lipids in f top kerogen. Indeed, these lipids are known to be predominantly unbranched (Weete, 1976). A low bacterial contribution is shown by the occurrence of a few branched hydrocarbons and of the hopanoids observed either directly in the pyrolysate or after desul-phurisation. The lack of lignin-derived pyrolysis pro-ducts rules out a large terrestrial input, however the occurrence of C20+acids may re¯ect a weak contribu-tion of higher plant lipids.
In addition, a substantial contribution of sulphurised carbohydrates is also inferred in f top kerogen based on the high amount of short chain alkylthiophenes obtained by ¯ash pyrolysis. This is consistent with the very high value ofd13Corgof f top kerogen compared to other samples from the Gorodische section (unpub-lished results).
5.2. Mechanism of OM accumulation
The presence of alkane/alkene doublets in kerogen pyrolysates is often associated with the selective pre-servation of highly aliphatic macromolecules such as algaenans and cutans. However, when this type of pathway is implicated, some morphological features of the source organisms are, at least partly, retained in the resulting kerogen (Derenne et al., 1991, 1992b). The nanoscopically amorphous nature of f top kerogen evi-denced by TEM observations therefore rules out a
ni®cant role for the selective preservation pathway in its formation. In contrast, such morphological features, in conjunction with the high Sorg content and the abun-dance of OSC in the pyrolysate indicate that natural sulphurisation played a major role in OM preservation in this organic-rich level. Based on XANES results, it appears that intermolecular sulphur incorporation was dominant over intramolecular processes.
Some alkyl chains in f top kerogen are linked by ether bridges, sometimes in addition to S linkages. The pre-sence of ether bridges in S-rich kerogens and macro-molecular fractions of bitumens and oils has been previously observed via selective degradation and/or pyrolytic studies (Richnow et al., 1992, 1993; Koop-mans et al., 1996; HoÈld et al., 1998; Putschew et al., 1998; Schaeer-Reiss et al., 1998; Jenisch-Anton et al., 1999). In addition, S- and O-containing compounds were recently described in bitumens of the Posidonia Shales (Wilkes and Hors®eld, 1999). The similar dis-tributions of the compounds released from sulphide and ether cleavage indicate that S- and O-bound moieties are probably derived from the same precursors in the Mon-terey samples studied by Richnow et al. (1992, 1993). Likewise, relatively close distributions of then-alkanes, alkylthiophenes and alkanones released by pyrolysis are observed for f top kerogen. The presence of ether bridges in S-rich kerogens is not fully explained. Rich-now et al. (1992, 1993) and Jenisch-Anton et al. (1999) proposed that part of the oxygenated functions present in S-rich kerogens are directly derived from the biologi-cal precursors. However, Jenisch-Anton et al. (1999) also noticed that such an origin could not account for the variety of oxygenated molecules they observed in S-rich oils and kerogen from Marvejols, Rozel Point and Gibellina, and proposed that they may be partly derived from diagenetic incorporation of oxygen. Indeed, oxi-dative incorporation in macromolecular structures was previously shown to be responsible for the preservation of some lipids under highly oxic depositional conditions (i.e. Gatellier et al., 1993). Jenisch-Anton et al. (1999) considered that the oxic/anoxic interface should be the seat of a competition between S- and O-incorporation into OM, depending on the availability of each species. The absence of bioturbations and of benthic organisms in f top level reveals that the sediment/water interface was severely dysoxic. However, the presence of shells from nectonic organisms such as ammonites or belem-nites indicate that anoxia did not invade the whole water column. Accordingly the OM was partly oxidized and began to form macromolecular units in the water column. Moreover, as considered in previous studies (Schouten et al., 1994; Carmo et al., 1997), this partial oxidation would have favoured subsequent sulphurisa-tion of the OM in the sediment by increasing its reac-tivity towards inorganic sulphur. Added to a high planktonic productivity, a high level of OM
preserva-tion, linked to the incorporation of both sulphur and oxygen, would thus account for the particularly high amount of OM preserved in f top level.
5.3. Comparison with other S-rich kerogens
This study of f top chemical structure, source organ-isms and formation pathway involved a combination of a number of methods: electron microscopy, FTIR, solid state 13C NMR, XANES spectroscopy and GC/MS analysis of pyrolysis products before and after desul-phurisation. Such a multidisciplinary approach has only been applied so far to a limited number of immature S-rich kerogens including samples of the Monterey For-mation (Richnow et al., 1992; Nelson et al., 1995; Stan-kiewicz et al., 1996; HoÈld et al., 1998) and from the bituminous laminites of Orbagnoux (Mongenot et al., 1997, 1999; van Kaam-Peters and Sinninghe DamsteÂ, 1997; Mongenot, 1998; van Kaam-Peters et al., 1998b). We have thus compared f top with these two types of kerogens so as to examine their similarities and dier-ences, especially concerning sulphur linkage and beha-viour upon pyrolysis. Interestingly the three deposits correspond to dierent sedimentary contexts: a calm silty clayey platform for the Kashpir oil shales (Hantz-pergue et al., 1998), a carbonated lagoon for Orbagnoux (Mongenot et al., 1997) and a restricted basin with upwellings and siliceous deposit for the Monterey For-mation (Ulmishek and Klemme, 1990).
The bulk characteristics of whole rocks and kerogens from f top, Orbagnoux (TM9sa) and dierent samples from the Monterey Formation (Naples Beach) are pre-sented in Table 6. It can be seen that these samples are rather dierent, in terms of TOC content but also in aliphaticity and sulphur content. Indeed, a regular decrease both in aliphaticity and sulphur content is noticed from Orbagnoux to f top and to the Monterey kerogens. Relatively high O/C ratios are noted for both f top and Monterey kerogens. The O/C ratio is not available for the Orbagnoux kerogen, but a low OI of 15 mg CO2/g TOC (Mongenot, 1998) added to spectro-scopic features and to the low proportion of oxygen-containing pyrolysis products (Mongenot et al., 1999) indicate that this ratio must be low.
(Pytte, 1989). This AOM has not been observed under TEM but it has been shown to be enriched in sulphur, after isolation by density gradient fractionation (Stankiewicz et al., 1996). It is therefore likely that ``amorphinite II'' corresponds to the gel-like orange particles. This comparison reveals that, despite their dierent chemical compositions, these S-rich kerogens, when observed via microscopy present similar features which are highly characteristic and can be related to their high sulphur content.
The thermal behaviour of a kerogen can be char-acterised by its pyrolysis products and residues. A direct comparison of the pyrolysis products of f top and Orbagnoux kerogens can be performed since the same analytical procedure was used, i.e. indirect ``o-line'' pyrolysis with a preliminary 300C thermal treatment.
Upon 400C pyrolysis, the two kerogens exhibit similar
weight losses (58% of initial OM for f top and 62% for TM9sa). However 85% of this loss represents trapped products in the case of TM9sa while hardly 50% is trapped in the case of f top. This indicates that pyrolysis products of f top are more volatile and therefore of lower mean molecular weight than those of TM9sa. This dierence is also re¯ected upon the fractionation of the pyrolysate on alumina column: ca. 20% are retained on the column for f top while ca. 55% are retained in the case of TM9sa, once again meaning that the pyrolysis products of Orbagnoux kerogen are more polar and/or of higher molecular weight than those of f top. Several heating experiments on Monterey kerogens have been previously reported but the experimental conditions were sharply dierent from those used in our study so that precise comparison is dicult. Baskin and Peters (1992) and Nelson et al. (1995) used hydrous pyrolysis which is usually considered as aording higher yields than anhy-drous pyrolysis and as promoting the release of lower molecular weights products (Lewan, 1997). Idiz et al. (1990) and Tomic et al. (1995) carried out pyrolysis in sealed vessels, i.e. under conditions which also favour the formation of volatile products. Taken together, these experiments indicated that both the Monterey kerogens and TM9sa tend to generate large amounts of high
mole-cular weight products upon pyrolysis. For example, after heating in sealed tubes at 300C for 100 h, the volatile
compounds formed only accounted for 25 wt.% of the initial kerogen for the Monterey samples (Idiz et al., 1990). XANES spectroscopy has been so far performed on a limited number of kerogen samples: from Monterey, Orbagnoux and Recent sediments from the Peru upwel-ling area (Eglinton et al., 1994; Nelson et al., 1995; Sar-ret et al., unpublished data). This spectroscopic method has been shown to be a powerful technique to char-acterise sulphur forms, in various coals and asphaltenes (George and Gorbaty, 1989; Kasrai et al., 1996a) and is especially appropriate to follow the behaviour of sul-phur functions upon pyrolysis. The present study indi-cated that the major sulphur forms in f top kerogen are alkylsulphides and in lesser amount di(poly)sulphides. A similar distribution of sulphur species in the Monterey kerogens was determined by Nelson et al. (1995) and Eglinton et al. (1994). In contrast, the kerogen of Orbagnoux was shown to contain thiophenes as the major sulphur form and in lesser amount di(poly)sul-phides and suldi(poly)sul-phides. These features imply that sulphur incorporation was dierent for f top and Monterey kerogens on the one hand and TM9sa on the other hand: essentially intermolecular for f top and Monterey kerogens (polysulphides and sulphides) and mostly intramolecular for TM9sa (thiophenes). However, the three kerogens present similar behaviour upon thermal stress, i.e. a progressive decrease in di(poly)sulphides and sulphides while the relative abundance of thio-phenes increases (Nelson et al., 1995; Sarret et al., unpublished data). This disappearance of polysulphides and sulphides, and relative increase in the contribution of thiophenes with increasing temperature, is consistent with the accepted idea of preferential cleavage of weak polysulphide and sulphide links and aromatisation of the residue during pyrolysis (Baskin and Peters, 1992; Tomic et al., 1995).
A conspicuous feature of TM9sa when compared to f top and Monterey kerogens is the evolution of the Sorg/ C ratio in the pyrolysis residues. As in the case of f top residues, a progressive decrease of the Sorg/C ratio was
Table 6
Bulk features of f top, TM9sa (Orbagnoux) and dierent Naples Beach samples (Monterey Formation)
Whole rock Kerogen
TOC(%) HI(mg HC/g TOC) H/C O/C Sorg/C
f top 45 700 1.39 0.20 0.07
TM9saa 7.2 909 1.42 ± 0.09
Naples Beach 4.7b±20.6c 490d 1.23c±1.29b 0.15d 0.05b,d
a Data were collected from Mongenot et al. (1997). b Data from Stankiewicz et al. (1996).
observed on the Monterey residues by Idiz et al. (1990) and Nelson et al. (1995). This decrease was considered by these authors to be consistent with the preferential cleavage of sulphide bonds and also with the dis-appearance of sulphides in the pyrolysis residues revealed by XANES (Nelson et al., 1995). In contrast, for TM9sa kerogen, the Sorg/C ratio of the residues increases with temperature (Mongenot et al., 1999). The latter authors proposed that this increase indicated the formation of sulphur-rich, polyaromatic, refractory material in the pyrolysis residues. A similar increase of the Sorg/C ratio in the pyrolysis residues was observed for a S-poor Type I kerogen from GoÈynuÈk oil shales (Gillaizeau et al., 1997). This behaviour could be related to the high proportion of thiophenes in the TM9sa kerogen, as thiophenes are known to be thermally more stable than sulphides.
Taken together, elemental composition, spectroscopic features (especially XANES) and evolution of the Sorg/C ratio upon a thermal stress point to a closer similarity between f top and the Monterey kerogens than with the Orbagnoux kerogen in terms of sulphur content and the nature of sulphur linkages. It thus appears that the less S-rich kerogens, i.e. f top and Monterey, are mostly based on intermolecular incorporation of sulphur while the Orbagnoux kerogen which is extremely S-rich, is based on inter- and intramolecular incorporation, the latter being predominant. In addition, the oxygen con-tent might also be an important factor since the role of oxygen in the thermal behaviour of S-rich kerogens of the Monterey Formation was recently discussed (Rey-nolds et al., 1995). Indeed, as noticed above, f top and Monterey kerogens are characterised by relatively high O/C ratios while Orbagnoux kerogen contains a low amount of oxygen. The medium S-richness of f top and Monterey kerogens might be linked to this high oxygen content: incorporation of both O and S would have occurred during diagenesis, contrary to Orbagnoux, where only S- would have been incorporated.
6. Conclusions
An organic-rich sample, termed f top (45% TOC, Sorg/C=0.068), from the Gorodische section (Upper
Jurassic, Russian platform) was studied using a combi-nation of microscopic, spectroscopic and pyrolytic methods. Microscopic observations indicated that the kerogen is mainly composed of orange gel-like, nanos-copically amorphous organic matter. The kerogen exhi-bits a relatively high aliphatic character as shown by FTIR and solid state13C NMR while XANES spectra revealed that organic sulphur is mainly present as di(poly)sulphides and sulphides. Pyrolysis products are mainly linear compounds:n-alkanes,n-alkanones and a large amount of OSC with a linear skeleton, as shown by desulphurisation. High amounts of short chain-alkylthiophenes are also generated upon ¯ash pyrolysis. Taken together, these results indicate that the kerogen of f top is mostly derived from algal or cyanobacterial lipids and from carbohydrates, rapidly incorporated into a macromolecular network by natural sulphurisa-tion, consistent with its highd13C value. The relatively high abundance of oxygenated products, consistent with the high O/C ratio of this kerogen, indicates that oxi-dative incorporation via ether linkages was also involved. Comparison of f top sample with other S-rich kerogens pointed to a closer similarity with the Mon-terey kerogens rather than with a kerogen from the bituminous laminites of Orbagnoux when both sulphur incorporation (intermolecular vs. intramolecular) and the extent of oxygen incorporation during kerogen for-mation are considered.
Acknowledgements
This is a contribution to Peri-Tethys project 95-96/28. We thank CIME Jussieu for SEM observations, Mr. B. Rousseau (ENS) for preparation of ultrathin sections observed by TEM, Mrs. M. Grably (Laboratoire de BiogeÂochimie Isotopique, UPMC) for isotopic mea-surement, Mrs. J. Maquet (Laboratoire de Chimie de la MatieÁre CondenseÂe, UPMC) for NMR spectroscopy and SocieÂte de Secours des Amis des Sciences for ®nancial support to A.R. Dr. P. Farrimond, Dr. A. Bishop and an anonymous reviewer are acknowledged for helpful comments on an earlier version of this paper.
Appendix
When several isomers are possible for the above structures, only one is presented.
References
Adam, P., 1991. Nouvelles structures organo-soufreÂes d'inteÂreÃt geÂochimique: Aspect moleÂculaires et macromoleÂculaires. PhD thesis, Louis Pasteur University, Strasbourg.
Albro, P.W., 1976. Bacterial waxes. In: Kolattukudy, P.E. (Ed.), Chemistry and Biochemistry of Natural Waxes. Else-vier, Amsterdam, pp. 419±445.
Barouxis, A., Scribe, P., Dagaut, J., Saliot, A., 1988. Free and bound lipids from equatorial sur®cial sediments separated as a function of particle size. In: Mattavelli, L., Noveli,L., (Eds.), Advances in Organic Geochemistry 1987. Pergamon Press, Oxford (Organic Geochemistry 13, 773±783). Barth, T., Rist, K., Huseby, B., Ocampo, R., 1996. The
distribution of nitrogen between bitumen, water and residue in hydrous pyrolysis of extracted Messel oil shale. Organic Geochemistry 24, 889±895.
Baskin, D.K., Peters, K.E., 1992. Early generation character-istics of a sulphur-rich Monterey kerogen. The American Association of Petroleum Geologists Bulletin 76, 1±13. Baudin, F., Tribovillard, N., Laggoun-DeÂfarge, F., Lichtfouse,
E., Monod, O., Gardin, S., 1999. Depositional environment of a kimmeridgian carbonate ``black band'' (Akkuyu For-mation, south-western Turkey). Sedimentology 46, 589±602. Boussa®r, M., Gelin, F., Lallier-VergeÁs, E., Derenne, S.,
the microcycles. Geochimica et Cosmochimica Acta 59, 3731±3747.
Carmo, A.M., Stankiewicz, B.A., Mastalerz, M., Pratt, L.M., 1997. Factors controlling the abundance of organic sulfur in ¯ash-pyrolysates of Upper Cretaceous kerogens from Sergipe Basin, Brazil. Organic Geochemistry 26, 587±603.
Derenne, S., Largeau, C., Casadevall, E., Sinninghe DamsteÂ, J. S., Tegelaar, E.W., de Leeuw, J.W., 1990. Characterisation of Estonian Kukersite by spectroscopy and pyrolysis: evi-dence for abundant alkyl phenolic moieties in an Ordovician, Marine, type II/I kerogen. In: Durand, B., Behar, F. (Eds.), Advances in Organic Geochemistry 1989. Pergamon Press, Oxford (Organic Geochemistry 16, 873±888).
Derenne, S., Largeau, C., Casadevall, E., Berkalo, C., Rous-seau, B., 1991. Chemical evidence of kerogen formation in source rocks and oil shalesviaselective preservation of thin resistant outer walls of microalgae: origin of ultralaminae. Geochimica et Cosmochimica Acta 55, 1041±1050.
Derenne, S., Metzger, P., Largeau, C., van Bergen, P.F., Gatellier, J.P., Sinninghe DamsteÂ, J.S. et al., 1992a. Similar morphological and chemical variations of Gloeocapsomor-pha prisca in Ordovician sediments and cultured Botryo-coccus brauniias a response to changes in salinity. Organic Geochemistry 19, 299±313.
Derenne, S., Largeau, C., Berkalo, C., Rousseau, B., Wilhelm, C., Hatcher, P., 1992b. Non-hydrolysable macromolecular constituents from outer walls ofChlorella fuscaand Nano-chlorum eucaryotum. Phytochemistry 31, 1923±1929. Durand, B., Nicaise, G., 1980. Procedures for kerogen
isola-tion. In: Durand, B. (Ed.), Kerogen. Technip, Paris, pp. 33± 53.
Eglinton, T.I., Irvine, J.E., Vairavamurthy, A., Zhou, W., Manowitz, B., 1994. Formation and diagenesis of macro-molecular organic sulfur in Peru margin sediments. In: Tel-naes, N., van Graas, G., éygard, K. (Eds.), Advances in Organic Geochemistry 1993. Pergamon Press, Oxford (Organic Geochemistry 22, 781±799).
Francois, R., 1987. A study of sulphur enrichment in the humic fraction of marine sediments during early diagenesis. Geo-chimica et CosmoGeo-chimica Acta 51, 17±27.
Gatellier, J.-P.L.A., de Leeuw, J.W., Sinninghe DamsteÂ, J.S., Derenne, S., Largeau, C., Metzger, P., 1993. A comparative study of macromolecular substances of a Coorongite and cell walls of the extant algaBotryococcus braunii. Geochimica et Cosmochimica Acta 57, 2053±2068.
Gelin, F., Boussa®r, M., Derenne, S., Largeau, C., Bertrand, P., 1995. Study of qualitative and quantitative variations in kerogen chemical structure along a microcycle: correlations with ultrastructural features. In: Lallier-VergeÁs, E., Tribo-villard, N.-P., Bertrand, P. (Eds.), Organic Matter Accumu-lation. The Organic Cyclicities of the Kimmeridge Clay Formation (Yorkshire, GB) and the Recent Maar Sediments (Lac du Bouchet, France). Lecture Notes in Earth Sciences 57. Springer, Berlin, pp. 31±47.
George, G.N., Gorbaty, M.L., 1989. Sulfur K-edge X-ray absorption spectroscopy of petroleum asphaltenes and model compounds. Journal of the American Chemical Society 111, 3186±3192.
Gillaizeau, B., Derenne, S., Largeau, C., Berkalo, C., Rous-seau, B., 1996. Source organisms and formation pathway of the kerogen of the GoÈynuÈk Oil Shale (Oligocene, Turkey) as
revealed by electron microscopy, spectroscopy, and pyr-olysis. Organic Geochemistry 24, 671±679.
Gillaizeau, B., Behar, F., Derenne, S., Largeau, C., 1997. Nitrogen fate during maturation of a type I kerogen (Oligo-cene, Turkey) and related algaenans: Nitrogen mass balances and timing of N2production versus other gases. Energy and Fuels 11, 1237±1249.
Gong, C., Hollander, D.J., 1997. Dierential contribution of bacteria to sedimentary organic matter in oxic and anoxic environments, Santa Monica Basin, California. Organic Geochemistry 26, 545±563.
Goossens, H., Rijpstra, W.I.J., DuÈren, R.R., de Leeuw, J.W., 1986. Bacterial contribution to sedimentary organic matter; a comparative study of lipid moieties in bacteria and Recent sediments. In: LeythaÈuser, D., RullkoÈtter., J. (Eds.), Advan-ces in Organic Geochemistry 1985. Pergamon Press, Oxford (Organic Geochemistry 10, 683±696).
Hantzpergue, P., Baudin, F., Mitta, V., Olferiev, A., Zakharov, V., 1998. The Upper Jurassic of the Volga basin: ammonite biostratigraphy and occurence of organic carbon-rich facies. Correlations between boreal-subboreal and sub-mediterranean provinces. In: Crasquin-Soleau, S., Barrier, E. (Eds.) Tethys Memoir 4: Epicratonic Basins of Peri-Tethyan Platforms. MeÂmoires du MuseÂum National d'His-toire Naturelle 179, Paris, pp. 9±33.
Hartgers, W.A., Lopez, J.F., Sinninghe DamsteÂ, J.S., Reiss, C., Maxwell, J.R.J.O.G., 1997. Sulfur-binding in recent envir-onments: II. Speciation of sulfur and iron and implications for the occurence of organo-sulfur compounds. Geochimica et Cosmochimica Acta 61, 4769±4788.
Homann, C.F., Foster, C.B., Powell, T.G., Summons, R.E., 1987. Hydrocarbon biomarkers from Ordovician sediments and the fossil algaGloeocapsomorpha priscaZalessky 1917. Geochimica et Cosmochimca Acta 51, 2681±2697.
HoÈld, I.H., Brusee, N.J., Schouten, S., Sinninghe DamsteÂ, J.S., 1998. Changes in the molecular structure of a type II-S kerogen (Monterey Formation, U.S.A.) during sequential chemical degradation. In: Hors®eld, B., Radke, M., Schaefer, R.G., Wilkes, H. (Eds.), Advances in Organic Geochemistry 1997 (Organic Geochemistry 29, 1403-1417).
Idiz, E.F., Tannenbaum, E., Kaplan, I.R., 1990. Pyrolysis of high-sulfur Monterey kerogen. In: Orr, W.L., White, C.M. (Eds.), Geochemistry of Sulfur in Fossil Fuels. American Chemical Society Symposium Series 429, pp. 575±591. Jenisch-Anton, A., Adam, P., Schaeer, P., Albrecht, P., 1999.
Oxygen-containing subunits in sulfur-rich nonpolar macro-molecules. Geochimica et Cosmochimica Acta 63, 1059±1074. Kasrai, M., Bancroft, G.M., Brunner, R.W., Jonasson, R.G., Brown, J.R., Tan, K.H. et al., 1994. Sulfur speciation in bitumens and asphaltenes by X-ray absorption ®ne structure spectroscopy. Geochimica et Cosmochimica Acta 58, 2865± 2872.
Kasrai, M., Brown, J.R., Bancroft, G.M., Yin, Z., Tan, K.H., 1996a. Sulfur characterization in coal from X-ray absorption near edge spectroscopy. International Journal of Coal Geol-ogy 32, 107±135.
