- Email: [email protected]
Long-chain carboxylic acids in pyrolysates of Green River kerogen
Advances in Organic Geochemistry 1985 Org. Geochem. Vol. IO, pp. 1059-1065, 1986
0146-6380186
Printed in Great Britain. All rights reserved
Copyright 0
1986
$3.00 + 0.00
Pergamon Journals Ltd
Long-chain carboxylic acids in pyrolysates of Green River kerogen* KIMITAKA KAWAMURAt$, ELI TANNENBAUM~, B. J. HUIZINGA~~ and I. R. KAPLAN Institute of Geophysics and Planetary Physics, University of California at Los Angeles, Los Angeles, CA 90024, U.S.A. (Received
15 September
1985; accepted
3 February
1986)
Abstract-Long-chain fatty acids (C,&,), as well as C,,-C=,, isoprenoid acids (except for C,,), have been identified in anhydrous and hydrous pyrolyses products of Green River kerogen (200-4OO”C. 2-1000 hr). These kerogen-released fatty acids are characterized by a strong even/odd predominance (CPI: 4.8-10.2) with a maximum at C,, followed by lesser amounts of C,, and G, acids. This distribution is different from that of unbound and bound geolipids extracted from Green River shale. The unbound fatty acids show a weak even/odd predominance (CPI: 1.64) with a maximum at C,,, and bound fatty acids display an even/odd predominance (CPI: 2.8) with maxima at C,, and C,,. These results suggest that fatty acids were incorporated into kerogen during sedimentation and early diagenesis and were protected from microbial and chemical changes over geological periods of time. Total quantities of fatty acids produced during heating of the kerogen ranged from 0.71 to 3.2 mg/g kerogen. Highest concentrations were obtained when kerogen was heated with water for 100 hr at 300°C. Generally, their amounts did not decrease under hydrous conditions with increase in temperature or heating time, suggesting that significant decarboxylation did not occur under the pyrolysis conditions used, although hydrocarbons were extensively generated. Key words: fatty acids, isoprenoid acids, pyrolysis, kerogen, diagenesis, unbound, bound, tightly bound, decarboxylation, clay mineral
INTRODUflION
Fatty acids have frequently been considered as the major source of petroleum hydrocarbons which formed through decarboxylation (e.g. Cooper and Bray, 1963). Since this hypothesis was proposed, both observational and laboratory simulation studies were conducted on fatty acids (Cooper, 1962; Jurg and Eisma, 1964, 1968; Kvenvolden, 1968; Douglas ef al., 1968; Shimoyama and Johns, 1971, 1972). More recently kerogen has been accepted as the source of petroleum hydrocarbons because extensive studies have demonstrated the production of petroleum-like hydrocarbons during heating of kerogen (Ishiwatari et al., 1977; Harwood, 1977; Rohrback et al., 1984). However, whether hydrocarbon generation occurs directly from kerogen or through certain intermediate precursors, including fatty acids which may be produced during the breakdown of kerogen, is still unresolved. During thermal alteration experiments of Recent sediments, the amounts of extractable fatty acids
*Publication No. 2693: Institute of Geophysics and Planetary Physics, UCLA. tAuthor to whom correspondence should be addressed. $Present address: Department of Chemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, U.S.A. §Present address: Anadrill Schlumberger, Sugar Land, TX 77478, U.S.A. [IPresent address: Chevron Overseas Petroleum, Inc., 6001 Bollinger Canyon Road, San Ramon, CA 94583, U.S.A.
increased at elevated temperatures (Baedecker et al., 1977; Harrison, 1978; Kawamura and Ishiwatari, 1981, 1985a), suggesting that fatty acids are present as tightly bound forms in geopolymers (kerogen and humic compounds) which cannot be separated by normal extraction procedures but can be released on heating (Kawamura and Ishiwatari, 1981). Tightly bound fatty acids are probably formed during early diagenesis by both microbial (Kawamura and Ishiwatari, 1984a) and non-biological processes (Kawamura and Ishiwatari, 1985b). Thus, it is expected that some fatty acids are present as tightly bound forms in kerogen matrices. However, there are no reports of fatty acids being released during thermal treatment of kerogen, except for /I-hydroxy acids (Kawamura and Ishiwatari, 1982) although carboxylic acids have been reported in chromic acid oxidation products of kerogen (Burlingame and Simoneit, 1968; Burlingame et al., 1969). Here, we describe a homologous series of C,0-C32 fatty acids and C,&, isoprenoid acids in pyrolysates of kerogen isolated from Green River shale. This report presents details on the distribution of these carboxylic acids and discusses their origin and diagenetic history. Their thermal behaviour will also be discussed in relation to the generation of hydrocarbons. EXPERIMENTAL Kerogen was prepared from Green River Formation shale and subjected to thermal alteration experiments (for details see Tannenbaum and Kaplan, 1985). The ash content
1059
1060
KIMITAKA
f ;:.-
40
m”
20.
>, 5 LL m
10,
CPI 8.1
KAWAMURA
4OO”C,lOh 2.2 mg/g-ker
30.
IA
II
0.
18 20
Carbon
,l.l.l.l 22 24
26 28 30
I 32
number
Fig. 1. Chain-length distribution of n-fatty acids in the pyrolysates of Green River kerogen (400°C 10 hr. anhydrous).
of the prepared kerogen is 2.3%. Half to one gram of kerogen or kerogen plus minerals (Na-montmorillonite, illite and calcite) with and without distilled water (5 ml) was sealed in a Pyrex tube under vacuum. Kerogen to mineral weight ratio &as 1: 20. The sealed sample tubes were placed in an oven at temneratures of 200-400°C for 2-1000 hr. The hydrous samples were heated in a bomb to prevent explosion. Blanks were also run with samples. Heated kerogen samples were extracted with a methanol/CH,Cl, mixture and the extracts were concentrated to 4 ml (Huizinga et al., 1986). An aliquot (0.10 ml) of the concentrate was then saponified with 2ml 0.5 N KOH/methanol and neutral components were extracted with hexane/CH,CI, (7:3) (5 ml x 3) after adding 5 ml (KMnO,-oxidized) distilled water. Carboxyhc acids were then extracted with hexane/CH,Cl, (7: 3) (5 ml x 3) from the remaining solution, which was acidified with c. 0.5 ml concentrated HCI, and esterified with 2ml 14% BFJ methanol. The methyl esters were extracted with hexane and then purified on a silica gel column. Carboxylic acid methyl esters were determined with a Hewlett-Packard 5840 gas chromatograph equipped with an FID and installed with a fused silica DB-5 capillary column (0.25 mm x 30 m). Authentic C,-C, fatty acid methyl esters were used for peak identification and quantitation. In order to confirm peak identification, some samples were analyzed with a Finnigan Model 4000 GC-mass spectrometer with an INCOS 2300 data system. Isoprenoid acids were identified by comparing mass spectra with those in the literature (Douglas ef al., 1971) and in the INCOS data library. Two grams of powdered sample (Green River shale) were extracted with methanol/CH,Cl, (1: 1) mixture under sonication until the colour of the extracts disappeared. The extracts (unbound lipids) were saponified with 0.5 N KOH/methanol for 2 hr. The remaining samole (residue) was saponified with 0.5 N KOH/methanol ‘to separate bound geolipids. Both unbound and bound geolipids were analyzed for carboxylic acids as described above. Recovery of authentic C,r-Cu acids in the procedures from saponification to SiO, column chromatography was 85 f 9%. Duplicate experiments (300°C 10 hr with montmorillonite and water) showed that the relative standard deviation of n-fatty acid concentrations (C,,-C,,) was a 1I % of the mean value. In the procedural blanks, C,, , C,, and C, acids appeared, but their amounts were usually less than 9% of those of samples and they were not subtracted. RESULTS
Straight-chain
C,,+&
fatty acids, iso- and anteiso-
C,, acids and isoprenoid C,&,
acids (except for
et al.
Cls) were detected in both anhydrous and hydrous pyrolysis products of Green River kerogen. Straightchain fatty acids were found to be more abundant than branched-chain acids. As shown in Fig. 1, distributions of n-fatty acids are characterized by a predominance of palmitic (C,,) acid followed by stearic (C,,) acid and a strong even/odd predominance. These features were found in all the heated samples despite the different conditions used (anhydrous, hydrous, with and without minerals or different temperatures), although the concentrations of carboxylic acids varied depending on the heating conditions. Concentrations of n-fatty acids isolated ranged from 0.26 to 3.2 mg/g kerogen (Table 1). The lowest concentration was obtained when dry kerogen was heated with montmorillonite at 300°C for 10 hr, whereas the highest value was measured when kerogen alone was heated at 300°C for 100 hr under hydrous conditions. For comparable heating times, highest amounts of fatty acids were obtained during hydrous heating of the kerogen. The lowest amount of fatty acids were obtained when the kerogen was heated with montmorillonite under anhydrous conditions. The presenceof water appears to suppress the interaction of minerals and fatty acids. The maximum amount of fatty acids recovered in the pyrolysate (3.2 mg/g kerogen) is equivalent to 16Opg/g rock (kerogen content is c. 5% of the Green River shale used). This value is significantly higher than the concentration of unbound (17 pg/g rock) and bound (43 pg/g rock) fatty acids which can be extracted from the thermally untreated shale, as stated below. Although n-fatty acid distributions are similar, distributions of branched-chain acids relative to straight-chain acids changed depending on the heating conditions. Iso- and anteiso-C,, to n-C,, acids ratios varied from 0.23 to 0.94 (Table 1). The ratios of isoprenoid acids to total n-fatty acids also fluctuated from 0.018 to 0.28. The highest values were obtained when kerogen was heated with water, whereas the lowest were produced under dry heating conditions. Unbound and bound carboxylic acids extracted from untreated Green River shale, showed different distributions from those in the pyrolysates of Green River kerogen. In the unbound fraction (solvent extractable), isoprenoid C,, and C6 acids were more abundant than normal Clz, C4, C,, and Cl8 fatty acids, although total n-fatty acids were present in about the same amount (17 pg/g sediment) as the total isoprenoid acids (C,,,-C,, : 16 pg/g). As shown in Fig. 2, the distribution of unbound n-fatty acids is bimodal with peaks at CId and Cr,, and a weak even/odd predominance (CPI: 1.64, for definition see Table 1). These features are consistent with those previously reported in Green River shale (Eglinton et al., 1966; Douglas er al., 1968). The distribution of bound
carboxylic
acids is also different
from pyro-
lyzed kerogen-derived fatty acids, showing maxima at
Carboxylic Table
1. Distribution
of carboxylic
acids detected n-Fatty
Samples
in pyrolysis
1061
in pyrolysates products
of Green
acids
Major
G”-G,
acids
River
kerogen
(concentration:
Cu acids
as/g lsoprenoid
kerogen) acids
L/H
CPI
ai
i
(ai + i)/n
G, CM Cl4 G,
2.7 2.9 2.1
2.6
6.3 7.2 6.5 4.8
14.2 9.2 19.5 17.3
6.3 5.1 10.2 8.3
0.59 0.54 0.94 0.59
246 I40 I76 365
cw. CM. CM. cw
cu. Cl2 cm CM cm c,, cm Cl, Cn.G4 c2*c,,
3.4 2.5 2.6 2.0 2.5 2.6
5.0 5.6 6.5 5.0 7.3 6.1
12.0 14.9 12.2 35.3 18.6 21.2
5.0 6.9 5.0 20.2 8.4 8.8
0.50 0.54 0.58 0.58 0.49 0.58
223 317 249 533 283 308
CIP. Cl,. Cl,. Cl,. Cl,. Cl,.
G-C,,
Maior
Dry pyrolysiF 200°C. IOhr 3OO”C, 2hr 300°C. IOhr 300°C. IO hr with montmorillonite 3OO”C, 1OOhr 300°C. 1000 hr 400°C. IOhr Hydrous pyrolysis 300°C. 2 hr 300°C. 2 hr with illite 300°C. 2 hr with calcite 300°C. IOhr 300°C. IO hr with montmorillonite 300°C. IOhr with illite 3OO”C, IO hr with calcite 3OO”C, IOllhr 300°C. 100 hr with illite 3OO”C, 100 hr with calcite L/H:
Ratios of the amounts Preference Index (ratios N.D.: not detected.
1290 1050 I210 I300
Cl,, Cl,. Cl,, C,,,
Cl,~ Cl,. Cl,. ‘A.
1010 I230 II30 3220 2750 2010
G,* CIS. Cl,. Cl,. Cl,. Cl,? Cl,, cl** Cl,. cm. c,,.c,,.
cm cm G. Cn,
Cl,. Cl,. Cl,. CM.
G G, Cl, Cl,
CM. Cl, CM. G Cl,. cm CM. Go CM. G CM.. Cl, of lower molecular weight (CT,,&,) acids to those of higher molecular weight V&C,,) acids, CPI: Carbon of total amounts of even-carbon numbered acids to those of odd-carbon numbered acids), ai: anteiso. i: iso.
Cu and C,,, with an even/odd predominance (CPI: 2.84) (Fig. 2). Isoprenoid C,.&, acids are also present in the bound fraction, but only at a fifth of the concentration of the straight-chain acids (9.0 pg/g vs 43 pg/g sediment).
groups in kerogens. These studies show that with increasing burial and also at elevated temperatures during laboratory heating experiments there is a reduction in the quantity of carbonyl groups (Tissot et al., 1974). The release of carboxylic acids in amounts up to 0.36% of the initial kerogen, would DISCUSSION correspond to changes in the infrared spectra of kerogen. In addition to long-chain acids, short-chain Release of carboxylic acids from kerogen upon heating organic acids (C,-C,,, up to 0.3% of the initial The results indicate that (1) long-chain carboxylic kerogen) were also released in the same heating acids are present in the structure of kerogen, and/or experiments (Kawamura et al., 1986). The early (2) they are produced by the oxidation of some release of these carboxylic acids should contribute to precursors (alcohols, ketones, etc.), which are present lowering the O/C ratio of kerogen, which occurs prior as kerogen moieties. The chain-length distributions of n-fatty acids are similar to those in smlicial lacustrine sediments, which are generally characterized by a Unbound CPI very strong even-carbon-number predominance (e.g. 166pglg-sed. 1.64 CPI: lo-20), generally with a maximum at C,, (Cranwell, 1974, Brooks et al., 1976; Meyers and Takeuchi, I I 1979; Kawamura and Ishiwatari, 1985c). Therefore, it is probable that, during the formation of the geopolymer, algal and microbial acids are incorporated in the structure of kerogen in depositional I environments and are preserved there. Alternatively, some carboxylic acids such as C,,-C, isoprenoid acids, which are not dominant acids in Recent sediment but are abundantly present in Green River shale (Eglinton et al., 1966), may have been formed from 10. phytol (Ikan et al., 1975), preserved in kerogen during diagenesis and released during thermal pyrolysis. However, the Cr, isoprenoid acid which occurs in relatively low abundance, cannot be easily explal,‘IIlll I I J 0 ” ’ 10 12 14 16 10 20 22 24 26 28 30 32 ined as originating from phytol (C,,). The presence of carboxylic acids released from Carbon number kerogen is consistent with previous infrared studies Fig. 2. Chain-length distribution of n-fatty acids in unbound and bound fractions separated from Green River shale. which have investigated the occurrence of carbonyl
201
lull
1062
KIMITAKA
KAWAMJRA
P
er al.
U: Hydrous. *--a Temperature : 300.C
: Dry
I SOPRENOID
r
;A
w. calcite ,w.
illite
4-w. mod.
0
t.
*-------*----
2
10
100
1000
2
ark.
10
---__.__ *-mot-d. , 100
-----.___ 1000
Heating tlme t hours) Fig. 3. Changesin the concentrationof normal and isoprenoidcarboxylicacidswith prolonged heating during anhydrous(dry) and hydrous pyrolysisof Green River kerogenat 300°C.
to a decrease in the H/C ratio corresponding to extensive hydrocarbon generation from the kerogen (Tissot et al., 1974). Temperature is an important factor in controlling the releaseof carboxylic acids from kerogen. Concentrations of both n-fatty acids and isoprenoid acids increase with an increasing temperature during both hydrous and anhydrous pyrolysis (see Table 1). This indicates that at higher temperatures, the kerogen structure is altered to release carboxylic acids. The n-fatty acids which were released from the sample heated at 400°C for 10 hr (anhydrous) show a high even/odd-carbon-number predominance (Fig. l), suggesting that degradation of fatty acids is not significant even at 400°C. The effect of heating time on the amounts of kerogen-released carboxylic acids is different for anhydrous and hydrous pyrolysis at 300°C. As shown in Fig. 3, during anhydrous pyrolysis, concentrations of both n-fatty acids and isoprenoid acids slightly increased in the first 10 hr of heating and then decreasedwith prolonged time. These results suggest that although kerogen-structures are destroyed to release bitumen (Tannenbaum ef al., 1986), carboxylic acids in the bitumen fraction are subjected to thermal degradation under anhydrous conditions. Under the present experimental conditions, the presence of water favours the release of carboxylic acids from kerogen (Fig. 3). It is possible that some carboxylic acids are linked to kerogen matrices via ester linkage and the ester bond is hydrolyzed by the water to release these acids on heating. Similar results were obtained when kerogen plus the mineral mixture was heated under hydrous conditions. The amounts of n-fatty acids released are lower than those of samples without minerals (see Fig. 3), suggesting that part of the released acids were
trapped by the minerals and could not be effectively extracted with the organic solvents used. This is supported by the fact that significant portions of bitumen generated from heated kerogen were found to be adsorbed on minerals (Tannenbaum et al., 1986). Adsorption of carboxylic acids is more serious when kerogen was heated with montmorillonite under anhydrous conditions than under hydrous conditions (3OO”C, 10 hr, see Table 1). On the other hand, isoprenoid acids showed different results from those of n-fatty acids during prolonged heating in the presence of water and minerals (100 hr, see Fig. 3). Although n-fatty acids largely increased in abundance from 10 to 100 hr under hydrous conditions, isoprenoid acids did not. The amounts of these acids are almost constant from 10 to 100 hr, indicating that isoprenoid acids are more likely subjected to degradation than n-fatty acids. This is consistent with the previous work that has shown phytanic (C,,) acid to be less stable than n-fatty acids in hydrous heating experiments performed on recent sediments (Kawamura and Ishiwatari, 1985a). Origin
and fate of kerogen-released
fatty
acids
The distributions of fatty acids in lithified sediments, where their CPI values are generally low (close to 1) (Kvenvolden, 1968; Douglas et al., 1968), are different from those in the surface sediments of most modem lakes, where fatty acids show a strong even/odd predominance (e.g. CPI: 10-20) usually with a peak at C,, acid as stated above. Such a difference has been attributed to diagenetic changes of fatty acids (Kvenvolden, 1968). Laboratory experiments with Recent sediment showed that even/odd ratios of fatty acids decreased from 8.6 (200°C) to 3.8 (325°C) during 24 hr hydrous heating (Kawamura and Ishiwatari, 1985a). This was explained by a
Carboxylicacids in pyrolysates
1063
(1984a) on a 200 m sediment core from Lake Biwa, Japan. If, as suggested, kerogen-attached fatty acids are stable over geological time, their chain-length distribution may be used as a source indicator. The predominance of C,, and C,* acids in the kerogenreleased fatty acids suggests a predominant algal or bacterial contribution to Green River kerogen. Although C,,,-C,, acids are minor components in the 2 n-Fatty acids unbound fraction, their presence in kerogen pyroly30 5.0 3.2 mglg-ker .-I sates suggeststhere was also a small contribution of ;;; higher plant lipids to the sedimenting lake. However, g 20. once kerogen-attached carboxylic acids are released, they are subjected to either diagenetic changes 10. (including a-oxidation) or catagenetic changes, depending on environmental conditions. Therefore, 0 amil’I ’ *‘I l’ti.‘ti.a 10 12 14 16 18 20 22 24 26 26 30 32 their distribution pattern in bitumen must be carefully evaluated on the basis of the maturation and Carbon number alteration history of the source rock (e.g. KvenvolFig. 4. Comparison of chain-lengthdistributions between n-alkanesand n-fatty acids in the pyrolysatesof Green den, 1968; Seifert, 1975; Grantham and Douglas, River kerogen(3OO”C,100hr, hydrous).Hydrocarbon data 1977). are from Huizinga et al. (1986). CPI 1.2
n-Al kanes 1.8 mg/g-ker
Relationship carbons
mechanism of a-oxidation of C, acid to C,-, acid. Therefore, the low CPI value of unbound fatty acids in Green River sediments may be interpreted by a similar mechanism. Although /?-oxidation, rather than a-oxidation, is common in biological degradation of fatty acids, microbial a-oxidation cannot be ruled out under prolonged diagenetic conditions. Bound n-fatty acids show a weak even/odd predominance although their CPI value (2.84) is higher than that of unbound n-fatty acids (1.64) due to the presence of abundant C&, Cl0 and C3r acids (see Fig. 2). The presence of abundant odd-carbonnumbered fatty acids in the range of C,,,-C,,, suggests that bound fatty acids have also been subjected to a-oxidation during diagenesis.However, the presence of C,,-C,, acids with a high CPI (see Fig. 2), in the bound fraction indicates that the long-chain acids are preferrentially stabilized by binding to the kerogen at an early stage of diagenesis. Recent studies of lacustrine sediments showed that lower molecular weight fatty acids (XC,,) which are of algal origin are microbially unstable compared to higher acids which are of higher plant origin (Matsuda and Koyama, 1977; Meyers et al., 1980; Kawamura and Ishiwatari, 1984b). Stabilization of the residual long-chain acid therefore occurs by esterification to geopolymers. The kerogen-released fatty acids show a strong even-carbon-number predominance in both lowerand higher-molecular-weight ranges, and their distributions are more like those of modern surficial lacustrine sediments (see Figs 1 and 4). This suggests that these fatty acids were incorporated into the kerogen structure at an early stage in the sedimentation processand were protected from microbial and chemical degradation. The above finding confirms the observations made by Kawamura and Ishiwatari
between
carboxylic
acids and hydro-
Both n-alkanes and isoprenoid hydrocarbons are abundant in the same pyrolysate samples (Huizinga et al., 1986). If these hydrocarbons are generated primarily by the decarboxylation of carboxylic acids, their molecular structures and carbon chain lengths should reflect those of the acids, e.g. palmitic acid (C,,) should result in n-C,, alkane. Figures 4 and 5 compare the distributions of hydrocarbons and carboxylic acids in the pyrolysate of Green River kerogen (300°C 100 hr, hydrous). The n-alkanes are characterized by a very weak odd/even predominance
40 -
1.2 mglg-ker
lsoprenoid hydrocarbons
30 .
2 .-z ii Jj n
40
0.53
mglg-ker
lsoprenoid acids
30
20
0
13 14 15 16 1716
Carbon
19 2021
22
number
Fig. 5. Comparison of molecular distributions between isoprenoidhydrocarbonsand isoprenoidacidsin the pyrolysates of Green River kerogen (3OO”C,100hr. hydrous). Hydrocarbon data are from Huizingaer al. (1986).
1064
KIMITAKA
KAWAMURA
(CPI: 1.2) with maxima at C,, and C,, (Fig. 4). Similar distribution patterns were also measured for other experimental conditions during both anhydrous and hydrous pyrolysis (Huizinga et al., 1986). These results show differences between hydrocarbons and n-fatty acids in terms of CPI and chain-length distributions in the pyrolysates (Fig. 4), suggesting that decarboxylation of n-fatty acids is not a major mechanism for the generation of n-alkanes from kerogen during pyrolysis. Similarly, a relationship was not obtained between the distribution of isoprenoid acids and hydrocarbons in the kerogen pyrolysates. As shown in Fig. 5, the isoprenoid hydrocarbons have a maximum at Cg followed by C,,, whereas the isoprenoid acids also maximize at CIg. This suggeststhat isoprenoid hydrocarbons are not derived from isoprenoid acids by decarboxylation, but rather that phytol derivatives in the kerogen produce both isoprenoid acids and hydrocarbons when heated.
ef al. REFERENCES
BaedeckerM. J., Ikan R.. Ishiwatari R. and Kaplan I. R. (1977)Thermal alteration experimentson organicmatter in recent marine sediments as a model for petroleum genesis. In Chemistry of Marine Sediments (Editedby Yen
T. F.). DD.55-72. Ann Arbor Science.Michigan. Brooks P. W., Eglinton G., GaskellS. J., McHigh D. J., Maxwell J. R. and Philp R. P. (1976) Lipids of recent sediments, Part 1: Straight-chain hydrocarbons and carboxylic acids of some temperate lacustrine and subtropical lagoonal/tidal flat sediments. Chem. Geol. 18, 21-38.
Burlingame A. L. and Simoneit B. R. (1968) Isoprenoid fatty acids isolated from the kerogen matrix of the Green River Formation (Eocene). Science 160, 531-533. Burlingame A. L., Haug P. A., Schnoes H. K. and Simoneit B. R. (1969) Fatty acids derived from the Green River Formation oil by extractions and oxidations-A review. In Advances in Organic Geochemistry 1968 (Edited by Schenck P. A. and Havenaar I.), pp. 85-128. Pergamon Press, Oxford. Cooper J. E. (1962) Fatty acids in recent and ancient sediments and petroleum reservoir waters. Nature, London
193, 744-746.
Cooper J. E. and Bray E. E. (1963) A postulated role of fatty acids in petroleum formation. Geochim. Cosmochim. Acta CONCLUSIONS AND SUMMARY 27, 1113-1127. (1) Long-chain carboxylic acids (C,,-C,,) were Cranwell P. A. (1974) Monocarboxyhc acids in lake sediments: Indicators, derived from terrestrial and aquatic released in abundance (up to 3.2 mg/g kerogen) from biota of paleoenvironmental trophic levels. Chem. Geol. Green River kerogen during laboratory heating ex14, 1-14. periments at 200~400°C. The presence of water either Douglas A. G., Douraghi-Zadeh K., Eglinton G., Maxwell J. k. and Ramsay Jy N. (1968) Fatty acids in sediments favours the releaseof these acids or reduces their rate includine the Green River Shale (Eocene) and Scottish of thermal destruction. The distribution of fatty acids Torbanite (Carboniferous). In Ad&.&es in’ Organic Geoshows a strong even-carbon-numbered predominance chemistry 1966 (Edited by Hobson G. D. and Speers (CPI: S-10) with a maximum at Cu., and is different G. C.), pp. 315-334. Pergamon Press, Oxford. from those of unbound (CPI: 1.6, maximum at C,,) Douglas A. G., Blumer M., Eglinton G. and DouraghiZadeh K. (197 1) Gas chromatographic-mass spectroand bound (CPI: 2.8, maxima at C,* and Go) fatty metric characterization of naturally-occurring acyclic acids extracted from unheated Green River shale. isoprenoid carboxylic acids. Tefrahedron 27, 1071-1092. (2) Kerogen attachment of fatty acids appears to Eglinton G., Douglas A. G., Maxwell J. R., Ramsay J. N. occur early in the sediment accumulation history and Stillberg-Stenhagen S. (1966) Occurrence of isoprenoid fatty acids in the Green River Shale. Science 153, which stabilizes these acids from biological and 1133-1135. chemical alteration. (3) There is no parallel between the chain-length Grantham P. J. and Douglas A. G. (1977) Carboxyhc acids in some Uinta Basin sediments. In Advances in Organic distribution of n-carboxylic acids or isoprenoid acids Geochemistry 1975 (Edited by Campos R. and Gem J.), and their equivalent hydrocarbons in the kerogen up. 193-207. Enadimsa. Madrid. pyrolysates, suggesting that decarboxylation of car- Harrison W. E. (1978) Experimental diagenetic study of a modem lipid-rich sediment. Chem. GeoL 21, 315-334. boxylic acids is not an important mechanism for Harwood R. J. (1977) Oil and gas generation by laboratory hydrocarbon generation from kerogen during pyrolysis of kerogen. Bull. Am. Assoc. Pet. Geol. 61, hydrous pyrolysis. However, carboxylic acids may be 2082-2102. subjected to decarboxylation once they are released Huizinga B. J.. Tannenbaum E. and Kaplan I. R. (1986) Role of minerals in thermal alteration of organic matterinto bitumen. III; Generation of saturated and aromatic hydrocarbons (4) Although the release of carboxylic acids from and polar compounds. In preparation. source rocks has not been well documented, this Ikan R., Baedecker M. J. and Kaplan I. R. (1975) Thermal process could occur in low maturity sediments realteration on organic matter in recent marine sedimentsII. Isoprenoids. Geochim. Cosmochim. Acfa 39, 187-194. sulting in organic acid-enriched formation waters. This process would also introduce carboxylic acids in Ishiwatari R., Ishiwatari M., Rohrback B. G. and Kaplan I. R. (1977) Thermal alteration experiments in organic the bitumen fractions of sediments at a maturity matter from recent marine sediments in relation to petrange near the initiation of oil generation roleum genesis. Geochim. Cosmochim. Acta 41, 815-828. (& -OS-0.6%). Jurg J. W. and Eisma E. (1964) Petroleum hydrocarbons: Generation from fatty acids. Science 144, 1451-1452.
Acknowledzemenrs-Theexperimentalwork of this study Jurg J. W. and Eisma E. (1968) The mechanismof the generation of petroleum hydrocarbons from fatty acids. was performed at UCLA with support from NASA grant In Advances in Organic Geochemistry 1966 (Edited by 0JGR 05-007-2211to UCLA. We thank B. R. T. Simoneit Hobson G. D. and Speers D.), pp. 367-368. Pergamon for his reviewingthe manuscriptand MS PeggyChandler (W.H.O.I.) for her typing of the manuscript. Press,Oxford.
Carboxylic acids in pyrolysates Kawamura K. and Ishiwatari R. (1981) Experimental diagenesis of fatty acids in a sediment: changes in their existence forms upon heating. Geochem. J. 15, 1-8. Kawamura K. and Ishiwatari R. (1982) Tightly bound /I-hydroxy acids in a Recent sediment. Nurure, London 297, 144-145. Kawamura K. and Ishiwatari R. (1984a) Tightly bound ahphatic acids in Lake Biwa sediments: Their origin and stability. Org. Geochem. 7, 121-126. Kawamura K. and Ishiwatari R. (1984b) Fatty acids geochemistry of a 200m sediment core from Lake Biwa, Japan: early diagenesis and paleoenvironmental information. Geochim. Cosmochim. Acta 48, 251-266. Kawamura K. and Ishiwatari R. (1985a) Behavior of lipid compounds on laboratory heating of a recent sediment. Geochem. J. 19, 113-126: Kawamura K. and Ishiwatari R. (1985b) Conversion of sedimentary fatty acids from extractable (unbound + bound) to tightly bound form during mild heating. Org. Geochem. 8, 197-201. Kawamura K. and Ishiwatari R. (1985~) Distribution of lipid components in bottom sediments of freshwater lakes with different trophic status. Chem. Geol. 51, 123-133. Kawamura K., Tannenbaum E., Huizinga B. J. and Kaplan I. R. (1986) C,-C,, carboxylic acids released from kerogen on laboratory heating. Geochem. J., In press. Kvenvolden K. A. (1968) Evidence for transformations of normal fatty acids in sediments. In Advances in Organic Geochembrry 1966 (Edited by Hobson G. D. and Speers D.), pp. 355-366. Pergamon Press, Oxford. Matsuda H. and Koyama T. (1977) Early diagenesis of fatty acids in lacustrine sediments-I. Identification and distribution of fatty acids in recent sediment from a freshwater lake. Geochim. Cosmochim. Acta 41, 111-183.
1065
Meyers P. A. and Takeuchi N. (1919) Fatty acids and hydrocarbons in surticial sediments of Lake Huron. Org. Geochem.
1, 121-138.
Meyers P. A., Bourbonniere R. A. and Takeuchi N. (1980) Hydrocarbons and fatty acids in two cores of Lake Huron sediments. Geochim. Cosmochim. Acta 44, 1215-1221. Rohrback B. G., Peters K. E. and Kaplan I. R. (1984) Geochemistry of artificially heated humic and sapropelic sediments-II. Oil and gas generation. Bull. Am. Assoc. Pet. Geol.
68, 961-910.
Selfert W. K. (1915) Carboxylic acids in petroleum and sediments. In Progress in- the Chemisrry of Organic Natural Products (Edited bv W. Hertz, H. Griesebach and G. W. Kirby) Vol. 32, pp: 149. Springer, Vienna. Shimoyama A. and Johns W. D. (1911) Catalytic conversion of fatty acids to petroleum-like paraffins and their maturation. Nafure (London), Phys. Sci. 232, 140-144. Shimoyama A. and Johns W. D. (1972) Formation of alkanes from fatty acids in the presence of CaCO,. Geochim.
Cosmochim.
Acta
36, 87-91.
Tannenbaum E. and Kaplan I. R. (1985) Role of minerals in the thermal alteration of organic matter-I: Generation of gases and condensates under dry conditions. Geochim.
Cosmochim.
Acra
49. 2589-2604.
Tannenbaum E., Huizinga B. J.-and Kaplan I. R. (1986) Role of minerals in the thermal alteration of organic matter-II: A material balance. Bull. Am. Assoc. Per. Geol., In press. Tissot B., Durand B., Espitalie J. and Combaz A. (1914) Influence of nature and diagenesis of organic matter in formation of petroleum. Bull. Am. Assoc. Per. Geol. 58, 499-506.
0146-6380186
Printed in Great Britain. All rights reserved
Copyright 0
1986
$3.00 + 0.00
Pergamon Journals Ltd
Long-chain carboxylic acids in pyrolysates of Green River kerogen* KIMITAKA KAWAMURAt$, ELI TANNENBAUM~, B. J. HUIZINGA~~ and I. R. KAPLAN Institute of Geophysics and Planetary Physics, University of California at Los Angeles, Los Angeles, CA 90024, U.S.A. (Received
15 September
1985; accepted
3 February
1986)
Abstract-Long-chain fatty acids (C,&,), as well as C,,-C=,, isoprenoid acids (except for C,,), have been identified in anhydrous and hydrous pyrolyses products of Green River kerogen (200-4OO”C. 2-1000 hr). These kerogen-released fatty acids are characterized by a strong even/odd predominance (CPI: 4.8-10.2) with a maximum at C,, followed by lesser amounts of C,, and G, acids. This distribution is different from that of unbound and bound geolipids extracted from Green River shale. The unbound fatty acids show a weak even/odd predominance (CPI: 1.64) with a maximum at C,,, and bound fatty acids display an even/odd predominance (CPI: 2.8) with maxima at C,, and C,,. These results suggest that fatty acids were incorporated into kerogen during sedimentation and early diagenesis and were protected from microbial and chemical changes over geological periods of time. Total quantities of fatty acids produced during heating of the kerogen ranged from 0.71 to 3.2 mg/g kerogen. Highest concentrations were obtained when kerogen was heated with water for 100 hr at 300°C. Generally, their amounts did not decrease under hydrous conditions with increase in temperature or heating time, suggesting that significant decarboxylation did not occur under the pyrolysis conditions used, although hydrocarbons were extensively generated. Key words: fatty acids, isoprenoid acids, pyrolysis, kerogen, diagenesis, unbound, bound, tightly bound, decarboxylation, clay mineral
INTRODUflION
Fatty acids have frequently been considered as the major source of petroleum hydrocarbons which formed through decarboxylation (e.g. Cooper and Bray, 1963). Since this hypothesis was proposed, both observational and laboratory simulation studies were conducted on fatty acids (Cooper, 1962; Jurg and Eisma, 1964, 1968; Kvenvolden, 1968; Douglas ef al., 1968; Shimoyama and Johns, 1971, 1972). More recently kerogen has been accepted as the source of petroleum hydrocarbons because extensive studies have demonstrated the production of petroleum-like hydrocarbons during heating of kerogen (Ishiwatari et al., 1977; Harwood, 1977; Rohrback et al., 1984). However, whether hydrocarbon generation occurs directly from kerogen or through certain intermediate precursors, including fatty acids which may be produced during the breakdown of kerogen, is still unresolved. During thermal alteration experiments of Recent sediments, the amounts of extractable fatty acids
*Publication No. 2693: Institute of Geophysics and Planetary Physics, UCLA. tAuthor to whom correspondence should be addressed. $Present address: Department of Chemistry, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, U.S.A. §Present address: Anadrill Schlumberger, Sugar Land, TX 77478, U.S.A. [IPresent address: Chevron Overseas Petroleum, Inc., 6001 Bollinger Canyon Road, San Ramon, CA 94583, U.S.A.
increased at elevated temperatures (Baedecker et al., 1977; Harrison, 1978; Kawamura and Ishiwatari, 1981, 1985a), suggesting that fatty acids are present as tightly bound forms in geopolymers (kerogen and humic compounds) which cannot be separated by normal extraction procedures but can be released on heating (Kawamura and Ishiwatari, 1981). Tightly bound fatty acids are probably formed during early diagenesis by both microbial (Kawamura and Ishiwatari, 1984a) and non-biological processes (Kawamura and Ishiwatari, 1985b). Thus, it is expected that some fatty acids are present as tightly bound forms in kerogen matrices. However, there are no reports of fatty acids being released during thermal treatment of kerogen, except for /I-hydroxy acids (Kawamura and Ishiwatari, 1982) although carboxylic acids have been reported in chromic acid oxidation products of kerogen (Burlingame and Simoneit, 1968; Burlingame et al., 1969). Here, we describe a homologous series of C,0-C32 fatty acids and C,&, isoprenoid acids in pyrolysates of kerogen isolated from Green River shale. This report presents details on the distribution of these carboxylic acids and discusses their origin and diagenetic history. Their thermal behaviour will also be discussed in relation to the generation of hydrocarbons. EXPERIMENTAL Kerogen was prepared from Green River Formation shale and subjected to thermal alteration experiments (for details see Tannenbaum and Kaplan, 1985). The ash content
1059
1060
KIMITAKA
f ;:.-
40
m”
20.
>, 5 LL m
10,
CPI 8.1
KAWAMURA
4OO”C,lOh 2.2 mg/g-ker
30.
IA
II
0.
18 20
Carbon
,l.l.l.l 22 24
26 28 30
I 32
number
Fig. 1. Chain-length distribution of n-fatty acids in the pyrolysates of Green River kerogen (400°C 10 hr. anhydrous).
of the prepared kerogen is 2.3%. Half to one gram of kerogen or kerogen plus minerals (Na-montmorillonite, illite and calcite) with and without distilled water (5 ml) was sealed in a Pyrex tube under vacuum. Kerogen to mineral weight ratio &as 1: 20. The sealed sample tubes were placed in an oven at temneratures of 200-400°C for 2-1000 hr. The hydrous samples were heated in a bomb to prevent explosion. Blanks were also run with samples. Heated kerogen samples were extracted with a methanol/CH,Cl, mixture and the extracts were concentrated to 4 ml (Huizinga et al., 1986). An aliquot (0.10 ml) of the concentrate was then saponified with 2ml 0.5 N KOH/methanol and neutral components were extracted with hexane/CH,CI, (7:3) (5 ml x 3) after adding 5 ml (KMnO,-oxidized) distilled water. Carboxyhc acids were then extracted with hexane/CH,Cl, (7: 3) (5 ml x 3) from the remaining solution, which was acidified with c. 0.5 ml concentrated HCI, and esterified with 2ml 14% BFJ methanol. The methyl esters were extracted with hexane and then purified on a silica gel column. Carboxylic acid methyl esters were determined with a Hewlett-Packard 5840 gas chromatograph equipped with an FID and installed with a fused silica DB-5 capillary column (0.25 mm x 30 m). Authentic C,-C, fatty acid methyl esters were used for peak identification and quantitation. In order to confirm peak identification, some samples were analyzed with a Finnigan Model 4000 GC-mass spectrometer with an INCOS 2300 data system. Isoprenoid acids were identified by comparing mass spectra with those in the literature (Douglas ef al., 1971) and in the INCOS data library. Two grams of powdered sample (Green River shale) were extracted with methanol/CH,Cl, (1: 1) mixture under sonication until the colour of the extracts disappeared. The extracts (unbound lipids) were saponified with 0.5 N KOH/methanol for 2 hr. The remaining samole (residue) was saponified with 0.5 N KOH/methanol ‘to separate bound geolipids. Both unbound and bound geolipids were analyzed for carboxylic acids as described above. Recovery of authentic C,r-Cu acids in the procedures from saponification to SiO, column chromatography was 85 f 9%. Duplicate experiments (300°C 10 hr with montmorillonite and water) showed that the relative standard deviation of n-fatty acid concentrations (C,,-C,,) was a 1I % of the mean value. In the procedural blanks, C,, , C,, and C, acids appeared, but their amounts were usually less than 9% of those of samples and they were not subtracted. RESULTS
Straight-chain
C,,+&
fatty acids, iso- and anteiso-
C,, acids and isoprenoid C,&,
acids (except for
et al.
Cls) were detected in both anhydrous and hydrous pyrolysis products of Green River kerogen. Straightchain fatty acids were found to be more abundant than branched-chain acids. As shown in Fig. 1, distributions of n-fatty acids are characterized by a predominance of palmitic (C,,) acid followed by stearic (C,,) acid and a strong even/odd predominance. These features were found in all the heated samples despite the different conditions used (anhydrous, hydrous, with and without minerals or different temperatures), although the concentrations of carboxylic acids varied depending on the heating conditions. Concentrations of n-fatty acids isolated ranged from 0.26 to 3.2 mg/g kerogen (Table 1). The lowest concentration was obtained when dry kerogen was heated with montmorillonite at 300°C for 10 hr, whereas the highest value was measured when kerogen alone was heated at 300°C for 100 hr under hydrous conditions. For comparable heating times, highest amounts of fatty acids were obtained during hydrous heating of the kerogen. The lowest amount of fatty acids were obtained when the kerogen was heated with montmorillonite under anhydrous conditions. The presenceof water appears to suppress the interaction of minerals and fatty acids. The maximum amount of fatty acids recovered in the pyrolysate (3.2 mg/g kerogen) is equivalent to 16Opg/g rock (kerogen content is c. 5% of the Green River shale used). This value is significantly higher than the concentration of unbound (17 pg/g rock) and bound (43 pg/g rock) fatty acids which can be extracted from the thermally untreated shale, as stated below. Although n-fatty acid distributions are similar, distributions of branched-chain acids relative to straight-chain acids changed depending on the heating conditions. Iso- and anteiso-C,, to n-C,, acids ratios varied from 0.23 to 0.94 (Table 1). The ratios of isoprenoid acids to total n-fatty acids also fluctuated from 0.018 to 0.28. The highest values were obtained when kerogen was heated with water, whereas the lowest were produced under dry heating conditions. Unbound and bound carboxylic acids extracted from untreated Green River shale, showed different distributions from those in the pyrolysates of Green River kerogen. In the unbound fraction (solvent extractable), isoprenoid C,, and C6 acids were more abundant than normal Clz, C4, C,, and Cl8 fatty acids, although total n-fatty acids were present in about the same amount (17 pg/g sediment) as the total isoprenoid acids (C,,,-C,, : 16 pg/g). As shown in Fig. 2, the distribution of unbound n-fatty acids is bimodal with peaks at CId and Cr,, and a weak even/odd predominance (CPI: 1.64, for definition see Table 1). These features are consistent with those previously reported in Green River shale (Eglinton et al., 1966; Douglas er al., 1968). The distribution of bound
carboxylic
acids is also different
from pyro-
lyzed kerogen-derived fatty acids, showing maxima at
Carboxylic Table
1. Distribution
of carboxylic
acids detected n-Fatty
Samples
in pyrolysis
1061
in pyrolysates products
of Green
acids
Major
G”-G,
acids
River
kerogen
(concentration:
Cu acids
as/g lsoprenoid
kerogen) acids
L/H
CPI
ai
i
(ai + i)/n
G, CM Cl4 G,
2.7 2.9 2.1
2.6
6.3 7.2 6.5 4.8
14.2 9.2 19.5 17.3
6.3 5.1 10.2 8.3
0.59 0.54 0.94 0.59
246 I40 I76 365
cw. CM. CM. cw
cu. Cl2 cm CM cm c,, cm Cl, Cn.G4 c2*c,,
3.4 2.5 2.6 2.0 2.5 2.6
5.0 5.6 6.5 5.0 7.3 6.1
12.0 14.9 12.2 35.3 18.6 21.2
5.0 6.9 5.0 20.2 8.4 8.8
0.50 0.54 0.58 0.58 0.49 0.58
223 317 249 533 283 308
CIP. Cl,. Cl,. Cl,. Cl,. Cl,.
G-C,,
Maior
Dry pyrolysiF 200°C. IOhr 3OO”C, 2hr 300°C. IOhr 300°C. IO hr with montmorillonite 3OO”C, 1OOhr 300°C. 1000 hr 400°C. IOhr Hydrous pyrolysis 300°C. 2 hr 300°C. 2 hr with illite 300°C. 2 hr with calcite 300°C. IOhr 300°C. IO hr with montmorillonite 300°C. IOhr with illite 3OO”C, IO hr with calcite 3OO”C, IOllhr 300°C. 100 hr with illite 3OO”C, 100 hr with calcite L/H:
Ratios of the amounts Preference Index (ratios N.D.: not detected.
1290 1050 I210 I300
Cl,, Cl,. Cl,, C,,,
Cl,~ Cl,. Cl,. ‘A.
1010 I230 II30 3220 2750 2010
G,* CIS. Cl,. Cl,. Cl,. Cl,? Cl,, cl** Cl,. cm. c,,.c,,.
cm cm G. Cn,
Cl,. Cl,. Cl,. CM.
G G, Cl, Cl,
CM. Cl, CM. G Cl,. cm CM. Go CM. G CM.. Cl, of lower molecular weight (CT,,&,) acids to those of higher molecular weight V&C,,) acids, CPI: Carbon of total amounts of even-carbon numbered acids to those of odd-carbon numbered acids), ai: anteiso. i: iso.
Cu and C,,, with an even/odd predominance (CPI: 2.84) (Fig. 2). Isoprenoid C,.&, acids are also present in the bound fraction, but only at a fifth of the concentration of the straight-chain acids (9.0 pg/g vs 43 pg/g sediment).
groups in kerogens. These studies show that with increasing burial and also at elevated temperatures during laboratory heating experiments there is a reduction in the quantity of carbonyl groups (Tissot et al., 1974). The release of carboxylic acids in amounts up to 0.36% of the initial kerogen, would DISCUSSION correspond to changes in the infrared spectra of kerogen. In addition to long-chain acids, short-chain Release of carboxylic acids from kerogen upon heating organic acids (C,-C,,, up to 0.3% of the initial The results indicate that (1) long-chain carboxylic kerogen) were also released in the same heating acids are present in the structure of kerogen, and/or experiments (Kawamura et al., 1986). The early (2) they are produced by the oxidation of some release of these carboxylic acids should contribute to precursors (alcohols, ketones, etc.), which are present lowering the O/C ratio of kerogen, which occurs prior as kerogen moieties. The chain-length distributions of n-fatty acids are similar to those in smlicial lacustrine sediments, which are generally characterized by a Unbound CPI very strong even-carbon-number predominance (e.g. 166pglg-sed. 1.64 CPI: lo-20), generally with a maximum at C,, (Cranwell, 1974, Brooks et al., 1976; Meyers and Takeuchi, I I 1979; Kawamura and Ishiwatari, 1985c). Therefore, it is probable that, during the formation of the geopolymer, algal and microbial acids are incorporated in the structure of kerogen in depositional I environments and are preserved there. Alternatively, some carboxylic acids such as C,,-C, isoprenoid acids, which are not dominant acids in Recent sediment but are abundantly present in Green River shale (Eglinton et al., 1966), may have been formed from 10. phytol (Ikan et al., 1975), preserved in kerogen during diagenesis and released during thermal pyrolysis. However, the Cr, isoprenoid acid which occurs in relatively low abundance, cannot be easily explal,‘IIlll I I J 0 ” ’ 10 12 14 16 10 20 22 24 26 28 30 32 ined as originating from phytol (C,,). The presence of carboxylic acids released from Carbon number kerogen is consistent with previous infrared studies Fig. 2. Chain-length distribution of n-fatty acids in unbound and bound fractions separated from Green River shale. which have investigated the occurrence of carbonyl
201
lull
1062
KIMITAKA
KAWAMJRA
P
er al.
U: Hydrous. *--a Temperature : 300.C
: Dry
I SOPRENOID
r
;A
w. calcite ,w.
illite
4-w. mod.
0
t.
*-------*----
2
10
100
1000
2
ark.
10
---__.__ *-mot-d. , 100
-----.___ 1000
Heating tlme t hours) Fig. 3. Changesin the concentrationof normal and isoprenoidcarboxylicacidswith prolonged heating during anhydrous(dry) and hydrous pyrolysisof Green River kerogenat 300°C.
to a decrease in the H/C ratio corresponding to extensive hydrocarbon generation from the kerogen (Tissot et al., 1974). Temperature is an important factor in controlling the releaseof carboxylic acids from kerogen. Concentrations of both n-fatty acids and isoprenoid acids increase with an increasing temperature during both hydrous and anhydrous pyrolysis (see Table 1). This indicates that at higher temperatures, the kerogen structure is altered to release carboxylic acids. The n-fatty acids which were released from the sample heated at 400°C for 10 hr (anhydrous) show a high even/odd-carbon-number predominance (Fig. l), suggesting that degradation of fatty acids is not significant even at 400°C. The effect of heating time on the amounts of kerogen-released carboxylic acids is different for anhydrous and hydrous pyrolysis at 300°C. As shown in Fig. 3, during anhydrous pyrolysis, concentrations of both n-fatty acids and isoprenoid acids slightly increased in the first 10 hr of heating and then decreasedwith prolonged time. These results suggest that although kerogen-structures are destroyed to release bitumen (Tannenbaum ef al., 1986), carboxylic acids in the bitumen fraction are subjected to thermal degradation under anhydrous conditions. Under the present experimental conditions, the presence of water favours the release of carboxylic acids from kerogen (Fig. 3). It is possible that some carboxylic acids are linked to kerogen matrices via ester linkage and the ester bond is hydrolyzed by the water to release these acids on heating. Similar results were obtained when kerogen plus the mineral mixture was heated under hydrous conditions. The amounts of n-fatty acids released are lower than those of samples without minerals (see Fig. 3), suggesting that part of the released acids were
trapped by the minerals and could not be effectively extracted with the organic solvents used. This is supported by the fact that significant portions of bitumen generated from heated kerogen were found to be adsorbed on minerals (Tannenbaum et al., 1986). Adsorption of carboxylic acids is more serious when kerogen was heated with montmorillonite under anhydrous conditions than under hydrous conditions (3OO”C, 10 hr, see Table 1). On the other hand, isoprenoid acids showed different results from those of n-fatty acids during prolonged heating in the presence of water and minerals (100 hr, see Fig. 3). Although n-fatty acids largely increased in abundance from 10 to 100 hr under hydrous conditions, isoprenoid acids did not. The amounts of these acids are almost constant from 10 to 100 hr, indicating that isoprenoid acids are more likely subjected to degradation than n-fatty acids. This is consistent with the previous work that has shown phytanic (C,,) acid to be less stable than n-fatty acids in hydrous heating experiments performed on recent sediments (Kawamura and Ishiwatari, 1985a). Origin
and fate of kerogen-released
fatty
acids
The distributions of fatty acids in lithified sediments, where their CPI values are generally low (close to 1) (Kvenvolden, 1968; Douglas et al., 1968), are different from those in the surface sediments of most modem lakes, where fatty acids show a strong even/odd predominance (e.g. CPI: 10-20) usually with a peak at C,, acid as stated above. Such a difference has been attributed to diagenetic changes of fatty acids (Kvenvolden, 1968). Laboratory experiments with Recent sediment showed that even/odd ratios of fatty acids decreased from 8.6 (200°C) to 3.8 (325°C) during 24 hr hydrous heating (Kawamura and Ishiwatari, 1985a). This was explained by a
Carboxylicacids in pyrolysates
1063
(1984a) on a 200 m sediment core from Lake Biwa, Japan. If, as suggested, kerogen-attached fatty acids are stable over geological time, their chain-length distribution may be used as a source indicator. The predominance of C,, and C,* acids in the kerogenreleased fatty acids suggests a predominant algal or bacterial contribution to Green River kerogen. Although C,,,-C,, acids are minor components in the 2 n-Fatty acids unbound fraction, their presence in kerogen pyroly30 5.0 3.2 mglg-ker .-I sates suggeststhere was also a small contribution of ;;; higher plant lipids to the sedimenting lake. However, g 20. once kerogen-attached carboxylic acids are released, they are subjected to either diagenetic changes 10. (including a-oxidation) or catagenetic changes, depending on environmental conditions. Therefore, 0 amil’I ’ *‘I l’ti.‘ti.a 10 12 14 16 18 20 22 24 26 26 30 32 their distribution pattern in bitumen must be carefully evaluated on the basis of the maturation and Carbon number alteration history of the source rock (e.g. KvenvolFig. 4. Comparison of chain-lengthdistributions between n-alkanesand n-fatty acids in the pyrolysatesof Green den, 1968; Seifert, 1975; Grantham and Douglas, River kerogen(3OO”C,100hr, hydrous).Hydrocarbon data 1977). are from Huizinga et al. (1986). CPI 1.2
n-Al kanes 1.8 mg/g-ker
Relationship carbons
mechanism of a-oxidation of C, acid to C,-, acid. Therefore, the low CPI value of unbound fatty acids in Green River sediments may be interpreted by a similar mechanism. Although /?-oxidation, rather than a-oxidation, is common in biological degradation of fatty acids, microbial a-oxidation cannot be ruled out under prolonged diagenetic conditions. Bound n-fatty acids show a weak even/odd predominance although their CPI value (2.84) is higher than that of unbound n-fatty acids (1.64) due to the presence of abundant C&, Cl0 and C3r acids (see Fig. 2). The presence of abundant odd-carbonnumbered fatty acids in the range of C,,,-C,,, suggests that bound fatty acids have also been subjected to a-oxidation during diagenesis.However, the presence of C,,-C,, acids with a high CPI (see Fig. 2), in the bound fraction indicates that the long-chain acids are preferrentially stabilized by binding to the kerogen at an early stage of diagenesis. Recent studies of lacustrine sediments showed that lower molecular weight fatty acids (XC,,) which are of algal origin are microbially unstable compared to higher acids which are of higher plant origin (Matsuda and Koyama, 1977; Meyers et al., 1980; Kawamura and Ishiwatari, 1984b). Stabilization of the residual long-chain acid therefore occurs by esterification to geopolymers. The kerogen-released fatty acids show a strong even-carbon-number predominance in both lowerand higher-molecular-weight ranges, and their distributions are more like those of modern surficial lacustrine sediments (see Figs 1 and 4). This suggests that these fatty acids were incorporated into the kerogen structure at an early stage in the sedimentation processand were protected from microbial and chemical degradation. The above finding confirms the observations made by Kawamura and Ishiwatari
between
carboxylic
acids and hydro-
Both n-alkanes and isoprenoid hydrocarbons are abundant in the same pyrolysate samples (Huizinga et al., 1986). If these hydrocarbons are generated primarily by the decarboxylation of carboxylic acids, their molecular structures and carbon chain lengths should reflect those of the acids, e.g. palmitic acid (C,,) should result in n-C,, alkane. Figures 4 and 5 compare the distributions of hydrocarbons and carboxylic acids in the pyrolysate of Green River kerogen (300°C 100 hr, hydrous). The n-alkanes are characterized by a very weak odd/even predominance
40 -
1.2 mglg-ker
lsoprenoid hydrocarbons
30 .
2 .-z ii Jj n
40
0.53
mglg-ker
lsoprenoid acids
30
20
0
13 14 15 16 1716
Carbon
19 2021
22
number
Fig. 5. Comparison of molecular distributions between isoprenoidhydrocarbonsand isoprenoidacidsin the pyrolysates of Green River kerogen (3OO”C,100hr. hydrous). Hydrocarbon data are from Huizingaer al. (1986).
1064
KIMITAKA
KAWAMURA
(CPI: 1.2) with maxima at C,, and C,, (Fig. 4). Similar distribution patterns were also measured for other experimental conditions during both anhydrous and hydrous pyrolysis (Huizinga et al., 1986). These results show differences between hydrocarbons and n-fatty acids in terms of CPI and chain-length distributions in the pyrolysates (Fig. 4), suggesting that decarboxylation of n-fatty acids is not a major mechanism for the generation of n-alkanes from kerogen during pyrolysis. Similarly, a relationship was not obtained between the distribution of isoprenoid acids and hydrocarbons in the kerogen pyrolysates. As shown in Fig. 5, the isoprenoid hydrocarbons have a maximum at Cg followed by C,,, whereas the isoprenoid acids also maximize at CIg. This suggeststhat isoprenoid hydrocarbons are not derived from isoprenoid acids by decarboxylation, but rather that phytol derivatives in the kerogen produce both isoprenoid acids and hydrocarbons when heated.
ef al. REFERENCES
BaedeckerM. J., Ikan R.. Ishiwatari R. and Kaplan I. R. (1977)Thermal alteration experimentson organicmatter in recent marine sediments as a model for petroleum genesis. In Chemistry of Marine Sediments (Editedby Yen
T. F.). DD.55-72. Ann Arbor Science.Michigan. Brooks P. W., Eglinton G., GaskellS. J., McHigh D. J., Maxwell J. R. and Philp R. P. (1976) Lipids of recent sediments, Part 1: Straight-chain hydrocarbons and carboxylic acids of some temperate lacustrine and subtropical lagoonal/tidal flat sediments. Chem. Geol. 18, 21-38.
Burlingame A. L. and Simoneit B. R. (1968) Isoprenoid fatty acids isolated from the kerogen matrix of the Green River Formation (Eocene). Science 160, 531-533. Burlingame A. L., Haug P. A., Schnoes H. K. and Simoneit B. R. (1969) Fatty acids derived from the Green River Formation oil by extractions and oxidations-A review. In Advances in Organic Geochemistry 1968 (Edited by Schenck P. A. and Havenaar I.), pp. 85-128. Pergamon Press, Oxford. Cooper J. E. (1962) Fatty acids in recent and ancient sediments and petroleum reservoir waters. Nature, London
193, 744-746.
Cooper J. E. and Bray E. E. (1963) A postulated role of fatty acids in petroleum formation. Geochim. Cosmochim. Acta CONCLUSIONS AND SUMMARY 27, 1113-1127. (1) Long-chain carboxylic acids (C,,-C,,) were Cranwell P. A. (1974) Monocarboxyhc acids in lake sediments: Indicators, derived from terrestrial and aquatic released in abundance (up to 3.2 mg/g kerogen) from biota of paleoenvironmental trophic levels. Chem. Geol. Green River kerogen during laboratory heating ex14, 1-14. periments at 200~400°C. The presence of water either Douglas A. G., Douraghi-Zadeh K., Eglinton G., Maxwell J. k. and Ramsay Jy N. (1968) Fatty acids in sediments favours the releaseof these acids or reduces their rate includine the Green River Shale (Eocene) and Scottish of thermal destruction. The distribution of fatty acids Torbanite (Carboniferous). In Ad&.&es in’ Organic Geoshows a strong even-carbon-numbered predominance chemistry 1966 (Edited by Hobson G. D. and Speers (CPI: S-10) with a maximum at Cu., and is different G. C.), pp. 315-334. Pergamon Press, Oxford. from those of unbound (CPI: 1.6, maximum at C,,) Douglas A. G., Blumer M., Eglinton G. and DouraghiZadeh K. (197 1) Gas chromatographic-mass spectroand bound (CPI: 2.8, maxima at C,* and Go) fatty metric characterization of naturally-occurring acyclic acids extracted from unheated Green River shale. isoprenoid carboxylic acids. Tefrahedron 27, 1071-1092. (2) Kerogen attachment of fatty acids appears to Eglinton G., Douglas A. G., Maxwell J. R., Ramsay J. N. occur early in the sediment accumulation history and Stillberg-Stenhagen S. (1966) Occurrence of isoprenoid fatty acids in the Green River Shale. Science 153, which stabilizes these acids from biological and 1133-1135. chemical alteration. (3) There is no parallel between the chain-length Grantham P. J. and Douglas A. G. (1977) Carboxyhc acids in some Uinta Basin sediments. In Advances in Organic distribution of n-carboxylic acids or isoprenoid acids Geochemistry 1975 (Edited by Campos R. and Gem J.), and their equivalent hydrocarbons in the kerogen up. 193-207. Enadimsa. Madrid. pyrolysates, suggesting that decarboxylation of car- Harrison W. E. (1978) Experimental diagenetic study of a modem lipid-rich sediment. Chem. GeoL 21, 315-334. boxylic acids is not an important mechanism for Harwood R. J. (1977) Oil and gas generation by laboratory hydrocarbon generation from kerogen during pyrolysis of kerogen. Bull. Am. Assoc. Pet. Geol. 61, hydrous pyrolysis. However, carboxylic acids may be 2082-2102. subjected to decarboxylation once they are released Huizinga B. J.. Tannenbaum E. and Kaplan I. R. (1986) Role of minerals in thermal alteration of organic matterinto bitumen. III; Generation of saturated and aromatic hydrocarbons (4) Although the release of carboxylic acids from and polar compounds. In preparation. source rocks has not been well documented, this Ikan R., Baedecker M. J. and Kaplan I. R. (1975) Thermal process could occur in low maturity sediments realteration on organic matter in recent marine sedimentsII. Isoprenoids. Geochim. Cosmochim. Acfa 39, 187-194. sulting in organic acid-enriched formation waters. This process would also introduce carboxylic acids in Ishiwatari R., Ishiwatari M., Rohrback B. G. and Kaplan I. R. (1977) Thermal alteration experiments in organic the bitumen fractions of sediments at a maturity matter from recent marine sediments in relation to petrange near the initiation of oil generation roleum genesis. Geochim. Cosmochim. Acta 41, 815-828. (& -OS-0.6%). Jurg J. W. and Eisma E. (1964) Petroleum hydrocarbons: Generation from fatty acids. Science 144, 1451-1452.
Acknowledzemenrs-Theexperimentalwork of this study Jurg J. W. and Eisma E. (1968) The mechanismof the generation of petroleum hydrocarbons from fatty acids. was performed at UCLA with support from NASA grant In Advances in Organic Geochemistry 1966 (Edited by 0JGR 05-007-2211to UCLA. We thank B. R. T. Simoneit Hobson G. D. and Speers D.), pp. 367-368. Pergamon for his reviewingthe manuscriptand MS PeggyChandler (W.H.O.I.) for her typing of the manuscript. Press,Oxford.
Carboxylic acids in pyrolysates Kawamura K. and Ishiwatari R. (1981) Experimental diagenesis of fatty acids in a sediment: changes in their existence forms upon heating. Geochem. J. 15, 1-8. Kawamura K. and Ishiwatari R. (1982) Tightly bound /I-hydroxy acids in a Recent sediment. Nurure, London 297, 144-145. Kawamura K. and Ishiwatari R. (1984a) Tightly bound ahphatic acids in Lake Biwa sediments: Their origin and stability. Org. Geochem. 7, 121-126. Kawamura K. and Ishiwatari R. (1984b) Fatty acids geochemistry of a 200m sediment core from Lake Biwa, Japan: early diagenesis and paleoenvironmental information. Geochim. Cosmochim. Acta 48, 251-266. Kawamura K. and Ishiwatari R. (1985a) Behavior of lipid compounds on laboratory heating of a recent sediment. Geochem. J. 19, 113-126: Kawamura K. and Ishiwatari R. (1985b) Conversion of sedimentary fatty acids from extractable (unbound + bound) to tightly bound form during mild heating. Org. Geochem. 8, 197-201. Kawamura K. and Ishiwatari R. (1985~) Distribution of lipid components in bottom sediments of freshwater lakes with different trophic status. Chem. Geol. 51, 123-133. Kawamura K., Tannenbaum E., Huizinga B. J. and Kaplan I. R. (1986) C,-C,, carboxylic acids released from kerogen on laboratory heating. Geochem. J., In press. Kvenvolden K. A. (1968) Evidence for transformations of normal fatty acids in sediments. In Advances in Organic Geochembrry 1966 (Edited by Hobson G. D. and Speers D.), pp. 355-366. Pergamon Press, Oxford. Matsuda H. and Koyama T. (1977) Early diagenesis of fatty acids in lacustrine sediments-I. Identification and distribution of fatty acids in recent sediment from a freshwater lake. Geochim. Cosmochim. Acta 41, 111-183.
1065
Meyers P. A. and Takeuchi N. (1919) Fatty acids and hydrocarbons in surticial sediments of Lake Huron. Org. Geochem.
1, 121-138.
Meyers P. A., Bourbonniere R. A. and Takeuchi N. (1980) Hydrocarbons and fatty acids in two cores of Lake Huron sediments. Geochim. Cosmochim. Acta 44, 1215-1221. Rohrback B. G., Peters K. E. and Kaplan I. R. (1984) Geochemistry of artificially heated humic and sapropelic sediments-II. Oil and gas generation. Bull. Am. Assoc. Pet. Geol.
68, 961-910.
Selfert W. K. (1915) Carboxylic acids in petroleum and sediments. In Progress in- the Chemisrry of Organic Natural Products (Edited bv W. Hertz, H. Griesebach and G. W. Kirby) Vol. 32, pp: 149. Springer, Vienna. Shimoyama A. and Johns W. D. (1911) Catalytic conversion of fatty acids to petroleum-like paraffins and their maturation. Nafure (London), Phys. Sci. 232, 140-144. Shimoyama A. and Johns W. D. (1972) Formation of alkanes from fatty acids in the presence of CaCO,. Geochim.
Cosmochim.
Acta
36, 87-91.
Tannenbaum E. and Kaplan I. R. (1985) Role of minerals in the thermal alteration of organic matter-I: Generation of gases and condensates under dry conditions. Geochim.
Cosmochim.
Acra
49. 2589-2604.
Tannenbaum E., Huizinga B. J.-and Kaplan I. R. (1986) Role of minerals in the thermal alteration of organic matter-II: A material balance. Bull. Am. Assoc. Per. Geol., In press. Tissot B., Durand B., Espitalie J. and Combaz A. (1914) Influence of nature and diagenesis of organic matter in formation of petroleum. Bull. Am. Assoc. Per. Geol. 58, 499-506.