A postulated role of fatty acids in petroleum formation

A postulated role of fatty acids in petroleum formation

Oeochimica et Cosmochimica Acts,1963,Vol.27,pp.1113to 1127.Pergamon Press Ltd. Printed inNorthern Ireland A postulated role of fatty acids in petrole...

1MB Sizes 1 Downloads 19 Views

Oeochimica et Cosmochimica Acts,1963,Vol.27,pp.1113to 1127.Pergamon Press Ltd. Printed inNorthern Ireland

A postulated role of fatty acids in petroleum formation*t J. E. COOPER and E. E. BRAY Socony Mobil Oil Company, Dallas, Texas (Received17 April 1963; accepted5 June 1963) Aba&GDistributions of n-paraffins in recent sediments differ from those in petroleum in that n-paraffinswith odd-carbon numbers predominatein recent sedimentsbut not in petroleum. A hypothesis proposed to explain this differencepredicts a, change from a preferencefor evencarbon-numberedfatty acids in recent sediments to no preferencein petroleum reservoirwaters. Studies of fatty acid distributionssupport this predictionand provide evidencefor a relationship between fatty acid and n-paraffin distributions. INTRODUCTION AN UNDERSTANDING of the history

of petroleum deposits, including an understanding of the source of petroleum itself, should prove useful in exploration for petroleum. Geochemists attempt to arrive at an understanding of the origin of petroleum beginning with knowledge of its composition. Analyses show that petroleum contains compounds clearly related to compounds found in biological systems. TREIBS’ (1936) discovery in crude oils of porphyrins similar to chlorophyll and hemin in biological systems provided an early example of such a relationship. Other examples such as polycyclic hydrocarbons related to steroids (MEINSCHEIN, 1959) and pristane (BENDORAITIS et al., 1962) and phytane (DEAN and WHITEHEAD, 1961) related to phytol have also provided support for the organic theory of the origin of petroleum. Members of one important series of compounds, n-paraffins, are distributed differently in crude oils and biological systems. In most biological systems high molecular weight n-paraffins (more than 20 carbon atoms) having odd numbers of carbon atoms are more abundant than those having even numbers of carbon atoms (GERARDE and GERARDE, 1961 and 1962). In crude oils, however, concentrations decrease monotonically with increasing numbers of carbon atoms (ROSSINI, 1960). In biological systems the most abundant n-paraffins generally are those having 27, 29, 31, and 33 carbon atoms, but in crude oils the most abundant paraffins are those of low molecular weight. Relative abundances of odd- and even-carbon-numbered n-paraffins can be described numerically by a carbon preference index (CPI). A CPI value is defined as the mean of two ratios which are determined by dividing the sum of concentrations of odd-carbon-numbered n-paraffins by the sum of even-carbon-numbered n-paraffins over given concentration ranges (BRAY and EVANS, 1961). Use of two ranges eliminates bias caused by enhanced concentrations at ends of the ranges. * A preliminery report of this work appeared in (1961) Natire, Lond. 193, 744-740. t Presented before the Symposium on Production and Exploration Chemistry, Division of Petroleum Chemietry, 144th National American Chemical Society Meeting, April 2, 1963. 1113

J. E. COOPERand E. E.

1114

BRAY

One ratio was obtained by dividing the sum of mass spectral parent ion peak heights for odd-carbon numbers 25-33 by the sum of peak heights for even-carbon numbers 26-34. The other ratio was obtained by changing the denominator to included even-carbon peak heights for carbon numbers 24 through 32. Distributions of n-paraffins in three grains, barley, maize, and oats, are illustrated in Fig. 1. CPI values are 7.2, 5.1, and 9.2, respectively. By contrast the CPI value for the crude oil shown in Fig. 2 is 1.01.

w 100 1 z $

0

24

26

NUMBER OF

28 CARBON

30

32

34

ATOMS / MOLECULE

Fig. 1. Distributions of n-parafis

from some grains.

On a basis of geological evidence Cox (1946) suggests that biological remains serving as source material of petroleum are deposited in marine sediments. Examples of n-paraflin distributions in a recent and in an ancient marine sediment are shown in Fig. 2. The CPI value of 2.8 observed in the recent sediment resembles values for biological systems, and the value of 1.11 for the ancient sediment is between those of the recent sediment and the crude oil. For purposes of comparison the n-paraffin distributions have been determined in a wide variety of recent and ancient sediments and crude oils (BRAY and EVANS, 1961). Distributions of CPI values for samples from these sources are presented in Fig. 3. Recent sediments were obtained from ocean basin, continental shelf, bay, freshwater lake, saline lake, and soil environments. CPI values for these sediments show a widespread preference for heavy n-paraffins with odd numbers of carbon atoms. In recent sediments CPI values are distributed uniformly between 2.4 and 5.5; this variation is significantly greater than that for crude oils. The distribution of CPI values for n-paraffins extracted from ancient marine sediments is shown in the center of Fig. 3. Sediments studied were fine-grained and were mostly shales; a few, however, were limestones. Samples were taken both from outcrops and deep corings of a variety of Cretaceous, Jurassic, Pennsylvanian, and Mississippian sediments. The sampling sites were located in Colorado, Kansas,

1115

A postuhted role of fatty acids in petroleum formation I



1

-

1

’ RECENT

I

e

I

a

ANCIENT (EAGLE

0



24





Fig. 2. n-Par&h



26

NUMBER

OF



28

1

:

1

SEDIMENT

FORD SHALE CPI s I.11





30

CARBON

SEDIMEN





32

*

f



34

ATOMS /MOLECULE

distributions for a recent sediment, an ancient sediment and a crude oil.

IO RECENT

SEDIMENT

ANCIENT

SEDIMENT

CRUDE

0 75 1.0 Fig. 3.

1.5

n-Pas&in

2.0

OIL

3.5 2.5 3.0 CPI VALUES

4.0

4.5

5.0

5.5

distributionsin crude oils and sediment extrmts.

1116

J. E. COOPER and E. E. BRAY

Montana, Oklahoma, Tennessee, and Texas. CPI values for ancient sediments in general are lower than those for recent sediments, and some approach the value of 1-O which is characteristic of crude oils. The assumption is often made that distributions of hydrocarbons, originally deposited with sediments in past geologic ages, were similar to distributions found in recent sediments. If this assumption is valid, the n-paraffin distributions in nearly all ancient shales examined have changed since the time of deposition. Similarly the n-paraffin distribution in crude oils does not reflect the distribution in the original source material. Suggested mechanism for changes in parafJin distributions Several mechanisms have been proposed as an explanation of the absence of a preference for odd-carbon-numbered n-paraffins in crude oils. Mechanisms most often cited are: (1) the addition of large amounts of a mixture having the same n-paraffin distribution as crude oil (BRAY and EVANS, 1961) and (2) a process for accumulation of crude oil in which relative concentrations of members of a homologous series are determined by physical constants of the individual members (BIKER, 1959). According to the second of thesesuggestions even-carbon-numbered n-paraffins of the source sediment should become depleted relative to oddcarbon-numbered n-paraffins, and the sediment should exhibit a higher preference for odd-carbon-numbered par&ins as petroleum is removed. The enhanced preference predicted contrasts with CPI values for ancient sediments shown in Fig. 3. If absence of odd-carbon preference in crude oils is to be attributed to addition of n-paraffins, the source of these added paraffins must be considered. Evidence has been presented that components of crude oil are derived from organic material deposited in marine sediments. Since the n-paraffins are widely distributed, a major constituent of biological systems probably serves as their source. Similarities in structures of n-paraffins and saturated fatty acids suggest fatty acids as precursors of n-paraffins. Conversion of saturated fatty acids to n-paraffins could occur by several routes. One possibility is the loss of CO,. A decarboxylation process would account for the high CPI values found in recent sediments and soils, because fatty acids in biological systems are primarily those having even numbers of carbon atoms. Any reaction occurring in organic materials which will become petroleum must proceed at low temperatures (Cox, 1946), that is, at temperatures below 200” and possibly below 100”. A search of chemical literature led to the conclusion that the only known reactions effecting decarboxylation of saturated aliphatic acids or their anions at moderate temperatures are those producing acylate radicals ;1s interExamples of this type of reaction are the Kolbe Synthesis and the mediates. decomposition of peroxides (Fig. 4). The first step in the Kolbe Synthesis and in peroxide formation can be represented formally as a one electron oxidation of the acid or its anion. Acylate radicals lose carbon dioxide and form $kyl radicals. Many reaction paths are available for alkyl radicals. Likely paths (Fig. 5) under geological conditions sre : (1) reaction with a source of hydrogen atoms to form alkanes, and (2) reaction with an oxidizing agent to form fatty acids having

A postulated role of fatty acids in petroleum formation

1117

one less carbon then the original acids. Any organic compound containing carbonhydrogen bonds can serve as a source of hydrogen atoms, and ubiquitous sulfur can serve as an oxidizing agent. No specific reaction has been considered for the one electron oxidation of the fatty acid or its anion; however, oxidation by a metal ion, either free or chelated, is a possibility. Whatever the nature of the reaction, it must occur under generally reducing conditions. A single step in a KOLBE

SYNTHESIS

0

II :O.]

II -

RCH2

CO

--o

[

PEROXIDE

0 so),

-

- top

RCH2*

ACY LATE RADICAL

ACY LATE ION

(RCH~

RcH2

ALKYL RADICAL

DECOMPOSITION II 2 [RcH2EO]

-

-2co2 * -

2 RCH

ACY LATE RADICAL

ACYL PEROX IDE

Fig. 4. Low temperature

2

*

ALKYL RADICAL

decarboxylation

of aliphatic acids.

51 RCH2

COH

__t

RCH2.

+

C02+

H’

(I) RCH2* (2) ACID

Fig. 5. Reaction paths for alkyl radicals.

reaction sequence such as that described here would prtraffin and a fatty acid having one less carbon atom The CPI value existing after reaction of a single atoms can be calculated. The CPI has been defined in Equation (1).

2

n-3

n-3

2

CPI =;

3 n-2 [

4

odd-carbon even-carbon

result in formation of a nthan the parent acid. fatty acid having n carbon and may be represented as

paraffins

2 +

paraffins

odd-carbon

paraffins

?aJ4

2 2

even carbon-paraffins

(1 1

Limits have been set over the widest range possible. CPI values obtained by this scheme are those existing after a reaction producing only acids and paraffins has gone to completion, that is, when no fatty acids remain. The fraction producing

acid in each step is represented by fa and the fraction producing paraffin is represented by f,. The reaotion sequence is shown schematically in Equation (2).

51 this scheme A, represents the concentration

of the acid with 7~carbon atoms and. P, represents the concentration of the paraffin with m oarbon atoms. Acids and paraffins having fewer carbon atoms than the original aoid can be expressed as fractions of the original acid, as is shown in Equations (3) and (4). ,4 a--I =.f,A.;

4,-s

=fafaA,_l

=f2A,;

P n--l =f*&

U,_,

==f,&_,

-f&.&;

A,_, P,_z

=,&A,_,

c--fa3AR; ..

=$SA,+Z

The values for the n-pnraffins concentration can be substituted to give Equation (5) which can be simplified to Equation (6).

CPI = f

=fpf2.4;

.

(3)

. . . 14)

into Equation

(1)

(6)

The finite geometric series can be summed and simplified to give the CPI value as a function off, and 3”,, as is shown in Equation (7).

The limiting CPI v&m 8s f, approaches unity is 1. This limiting value suggests that the preference for n-padEns haviug odd numbers of earbon atoms is small when the fraction of alkyl radicals producing paraffins is small. Table 1 illustrates Cl?1 values as a function of the fraction of material forming peraffin or acid in the o&se of complete reaction. Clearly the P&W lies within the range found for crude oils when no more than 25 per cent of the intermediates Aocording to this scheme each acid produces all paraffins produce n-p&r&ns. having fewer carbon atoms than the a&d itself. Although each acid produces dimi~shi~ quantities of ~-p~r~~ns having fewer carbon atoms, each aoid initially present in the sediments would yield products; and the net result would be an irmrease in paraffins with lower carbon numbers. This predicted increase with

A postulated role of fatty acids in petroleum formation

1119

lower carbon numbers is consistent with observed distribution of n-paraffins in some ancient sediments and in crude oils. The n-paraffin distributions illustrated in Fig. 2 show increases in abundance of paraffins with lower carbon numbers in a crude oil and a shale as contrasted with the recent sediment. Table 1. Limiting CPI values as a function off, and f, f,

f,

0.0

1.0 0.9 0.8 0.7 0.6 0.6 0.4 0.3 0.2 0.1 0.0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Experimental

Ratio co 5.0 2.6 1.8 l-4 1.25 l-13 1.06 1.02 1.01 1.00

test of the proposed mechanism

General features of this scheme suggest a change in distribution of saturated, straight-chain fatty acids accompanying a change in distribution of n-paraffins. The preference for even-carbon-numbered acids in biological systems is well established. In 1948 RALSTON stated that isovaleric acid was the only odd-carbonnumbered acid found in natural fats. Although traces of odd-carbon-numbered fatty acids now have been identified in natural fats (SHORLAND, 1954) evencarbon-numbered fatty acids clearly predominate over odd-carbon-numbered ones in biological systems. Examinations of saturated, straight-chain fatty acids isolated from recent sediments, from ancient sediments, and from petroleum reservoir waters shows that the relative abundances of fatty acids having odd numbers of carbon atoms is greater in these systems than in biological systems (Fig. 6, Tables 2-5). Even carbon-numbered acids predominate over odd-carbon-numbered ones in almost every case studied. The analysis of Fall River water, the exception, is of low precision because of the low acid concentration. In general, the relative abundance of odd-carbon-numbered acids increases from recent sediment to ancient sediment Table 2. Sediment samples Source

Age

San Nicholas basin, P&f% ocean Santa Barbara basin, Pacific ooean Eagle ford shale, core, Wood Co., Texas Skull creek shale, outcrop, Weston Co., Wyoming Mowry shale, outcrop, Weston Co., Wyoming Chattanooga shale, outcrop, Cherokee Co., Oklahoma

Recent Recent Upper Cretaceous Lower Cretaceous Lower Cretaceous Mississippian

112Q

J.E.

E.BRAY

Coo~xcaandE.

Table 3. Aliphatic acids from sediments (ymoles/kg)

Acid

Recent sediments Santa San Nicholas Barbara _-~4.58 28-5 069 2.89 17.3 62.8 0.58 2.46 4.53 128 0.68 2.24 1.68 7.23 o-27 1.91 Q-84 7.03 Q-29 2-03 0.42 10-4

Ancient sediments Chat~~o~~~ f-29 1.03 1.30 0.71 0.92 0.31 0.22 0.13 OQ7 Q-Q3 0.03

Eagle Ford

Skull Cmek*

Q8I 0.84 2.74 1.17 1.32 0.59 0.44 0.38 0.30 O-22 0.21 0.21 0.19 0.16 0.15 0.09 O,[email protected]

O-10 O-10 0.32 0.30 0.55 0.44 0.46 0.30 0.27 Q-27 0.42 o-40 0.37 0*31 0.25 0.19 0.14

Mowry * ~--_ O-06 0.10

0.26 0.24 0.40 0.28 0.42 0.35 0.33 0.25 o-25 0.21 0.21 0.18 0.16 o-11 0,08

1.15 0.53

9.03 2.04

:

1.43 0.58 0.39 o-77

7.09 1.11 0.53 2.98

c

Sma.&st found

C&l

CI,

C,

CII

C*

Cs

Largest found

c33

c32

C28

C34

f*,,

c3,

18.8

15.4

1.57

1.74

I*45

1.42

419

325

1000

2000

750

750

Q, --

-+ Cl3 2%

Sample size (g)

* Internal standard was not used. t Observed peak was too small to measum,

Table 4. Water samples

-Source

Natrona Co., Wyoming* Dallas Co., Texas* Ellis Co., Texas* Cole creek field, Wyomingt North Tisdale field, Wyoming7 North Tisdale field, Wyomingt Panhandle field, Texas? Yale-quay field, Oklahomat Dry mesa field, Arizonan

Formation

Reservoir age

Mesaverde Trinity Trinity Fall River Lakota Crow Mountain Brown Dolomite Dees0

upper Cretaeeous Lower Cmtaeeous Lower Cretaeeous Lower Cretmeous Lower Cretaceous Triassic Permian Pennsylvanian Mississippian

* Fresh w&m---no detectable fatty acids in 2 1. of water. t Produced with petraleum.

1121

A postulatedrole of fatty acids in petroleumformation

ANCIENT SEDIMENT (EAGLE FORD SHALE) -I

3.0 t

n

PETROLEUM RESERVOIR WATER (PANHANDLE FIELD)

O~“r”“““t---+ 14 16 I8 20 CARBON

22 ATOMS

24

26 IN

26

30

ACID

Fig. 6. Comparison of the distributions of fatty acids in a recent sediment, an ancient sediment, and in water from a petroleum reservoir.

RECENT SEDIMENTS SAN NICHOLAS BASIN SANTA BARBARA BASIN

18.8 15.4

ANCIENT SEOIMENTS EAGLE FORO SHALE CHATTANOOGA Stinre SKULL CREEK SHALE Mowttv SHALE

-

1.74 1.57

1.45 1.42

PETROLEUvl RESERVOIR WATERS DRY MESA F~ELO PANHANDLE FIZLO

-

1.20 1.05

Fig. 7. Degree of preferencefor even-numbered acids in geological systems.

petroleum reservoir water (Fig. 6). One Paleozoic3 shale, however, has a high ratio of even- to odd-carbon-numbered fatty acids and will be re-examined. The mean of the ratios of concentrations of pslmitic and stearic seids to that of peptadecanoic acid provides a rough measure of degree of predominance of even-csrbonnumbered acids (Fig. 7). The increase in odd-carbon-numbered acids parallels the previously observed increase in even-carbon-numbered n-paraffins in related

to

1122

J. E. COOPEB and 3% E. BRAY

Table 5. Aliphatic acids from patroleum waters

Acid

Dry Mesa

C14 C15 C18 C17 C18

356 3.96 5.18 3.92 P-22 3.65 3.16 2.73 2.34 1.96 1.93 1.58 1.39 0.95 o-79 0.66 0.56

29.2 24.5 24.3 19.9 17.4 14.7 12.7 10.8 9.72 8.47 7.76 6.17 5.60 4.00 2.85 2.47 1.96

+ Jr + -I+ + -t

1.79 2-33 2.36 2.51 1.83 1.78 I.80 1.11 1.19 1.15 0.93 0.63 O-70 0.63 0.49 0.48 0.39

CIO

Ca

C16

C 10

C36

C3s

C50

Css

1.20

1.05

900

1215

Cl% c20

C21 C22 C23 C24 C25 Cae C 27 C26 C29 C80 Smallest found Largest found Cl, +%J

Panhandle -..____

Yale-Quay ..____ _.________ + i+ + -t -I-I+

-t

Cole Creek

(

pmobs I

f

x 10%

North Tisdde Crow Mountain Lakota ---+ i-t +

+ i+ -I-I+ -I-t +

C 16

%

c 20

%3

2400

475

O-84

2C1, Sample size (ml)

990

1365

sediment samples and in a crude oil related to the reservoir water (Fig. 2). This parallelism is consistent with predictions from the proposed mecha~sm for formation of n-paraffins in sediments and in petroleum. No explanation has been found for the absence of odd-carbon-numbered acids in Crow Mountain water. Analysis of this water is of low precision, because total acid concentration is lower than in any other petroleum reservoir water studied. Ancient sediment samples whose acid distributions are given in this paper were chosen because of their geologio association with petroleum. Distributions of acids from these samples do not reflect dist~butions of acids in all ancient sediments. ABELSOE~ and PARKER (1962) alsohave determined the fatty acid They found stearic (C-181, palmitio composition of several geologic samples. (C-l 6), and myristio (C-14) as the most abundant acids in both recent and ancient sediments. Ratios of even- to odd-carbon-numbered aoids were very high in all of the samples they studied. The survey of sediments and ground waters reported here is not intended to be complete but is intended only as a test of the hypothesis described.

A postulated role of fatty acids in petroleum formation

1123

ANALYTICAL PROCEDURE

Acids are isolated by saponification of organic matter in the samples, extra&ion of acids from the saponification mixture, esterification of the acids, and separation of aliphatic esters by urea adduction. The esters are determined by gas chromatography. An internal standard, stearic acid, is usually added for more accurate determination of the acids extracted from sediments. No internal standard was used with water samples, because the low concentration of acids in the waters makes sample division impracticable. Identification of fatty acids and their esters is based on the isolation procedure and gas chromatographic retention times. Acids are concentrated by the extraction procedure, and urea adduction eliminates cyclic structures and greatly

l-

5

w I ::

E

C-16 c-30

TIME

Fig. 8. Gas chromatogram of esters of fatty acids extracted from Eagle Ford shale.

lowers concentrations of branched and unsaturated acids. Gas chromatographic retention times serve to identify individual members of the fatty acid series. A gas chromatogram of methyl esters of fatty acids isolated from the Eagle Ford shale is shown in Fig. 8. A mass spectrum of methyl esters of acids isolated from Chattanooga shale supports the identification of the esters. Comparison of mass spectral and gas chromatographic data for esters from Chattanooga shale (Fig. 9) shows a parallelism among the ester distributions given by parent peaks, parent-minus-a-methoxylgroup peaks, and gas chromatographic data. Similar data for the fatty acids in beeswax are shown for comparison. As expected, concentrations of odd-carbonnumbered acids are much lower in beeswax than in Chattanooga shale. A more detailed study of acids isolated from Thermopolis shale (collected in Bighorn County, Wyoming) provides further evidence supporting the qualitative analyses. Esters were obtained from a 3 kg sample of shale by the usual procedures. Quantitative results are not directly comparable with results from other samples because of the large size of this sample. Qualitative data does show the presence of acids with both even and odd numbers of carbon atoms. The solid bars in Fig. 10 represent the parent peaks of the methyl esters of fatty acids from this shale. Gas chromatographic analysis shows a similar distribution. These esters were hydrogenated to paraffins by the procedure of MEINSCHEIN and KENNY 4

and E.E.

J.E. COOPER

1124

BRAY

(1957). Hydrogenation by this procedure results in the reduction of the ester to a having one less carbon atom than the original acid. The crosshatched bars in Fig. 10 show the distribution of paraffins produced by reduction of esters from Thermopolis shale. On the graph the paraffins are related to the numbers of carbon atoms in the original acids. As a comparison, the distributions of methyl esters and paraffins derived from a synthetic mixture containing acids with 14, 16, 17, 18, 19, 20 and 22 carbon atoms, are shown in Fig. 11. par&fin

CWATTANOOGA

BEESWAX

PARENT

CARBON

SHAL .E

PEAKS

ATOMS

IN

ACID

Fig. 9. Comparison of mass spectral and gas chromatographioanalyses of fatty acids from beesw&xand ~at~noog& shale. lOOF

2 (3

I ESTERS

II

0 PARAFFINS 75

ii = so z

25

it 0 CARBON

ATOMS

IN ORIGINAL

ACID

Fig. 10. Parent peaks from mass spectral analyses of methyl esters and paraffis derived from acids extracted from Thermopolis shale.

A postulated role of fatty acids in petroleum formation

1126

Control experiments show the results obtained in the work are only semiquantitative. Relative aonoentrations of acids are correct to within 40 per cent over the range bounded by pentadecanoic and behenic acids. Absolute concentrations determined for acids indicate only an order of magnitude. Relative concentrations within 60 per cent of the values observed would still provide support for the thesis of this paper. Consistently low concentrations are obtained for acids having fewer than 15 carbon atoms because of unfavorable equilibrium for urea adduct formation with lower esters. Control experiments have not been performed for acids having more than 22 carbon atoms. Odd-carbon-numbered acids could be introduced from other sources. The analytical procedure does not show whether acids are free or are parts of larger

$ e 7f

I ESTERS

5 = 50

0

B PARAFFINS

14

15

I6

17

18

CARBON

I3

20

ATOMS

21

IN

22

23

24

ORIGINAL

25

26

27

28

ACID

Fig. 11. Parent peaks from mass spectral analyses of methyl esters and paraEna derived from acids in synthetic mixture.

molecules such as esters. Contamination is an unlikely source for odd-carbonnumbered acids, because contamination with most naturally-occurring fats or oils would result in introduction of even-numbered acids only. Production of oddcarbon-numbered acids by oxidation during collection and storage of samples seems unlikely. Control experiments indicate that odd-carbon-numbered acids are not artifacts of the analytical scheme. Appearance of odd-carbon-numbered acids might result from a large increase in relative concentration of odd-carbon-numbered acids initially present in sediments in very small concentrations. Such an apparent increase could result from selective removal of even-carbon-numbered acids. Selective solubility is an unlikely cause of such an effect, because enrichment occurs in both waters and sediments. Microbial activity, however, cannot be excluded, because under certain conditions microorganisms consume even- but not odd-carbon-numbered acids (SILLIKERand R~TTE~BER~,1952). CONCLUSIONs If organic matter deposited in marine sediments served as the ultimate source of petroleum, the difference between distributions of n-paraffins in these materials is enigmatic. A reasonable explanation of this enigma will furnish a clue to the origin of petroleum. We suggest the following explanation both for the preference of odd-carbon-numbered paraffins in recent sediments and for the absence of this preference in petroleum. Fatty acids in the sediments serve as a source of nparaffins. Each acid can lose CO, to form an intermediate which reacts to give two products, a n-paraffin and a fatty acid. Each product would have one less

1126

J.E.COOPER~~~E.E.BRAY

carbon atom than the original acid. The product acid would undergo the same reaction to form a new n-paraffin and a new acid. Operation of such a process in a sediment initially containing even-numbered acids would cause the introduction first of odd-numbered acids and paraffins. Significant quantities of even-numbered paraffins would be formed as the concentration of odd-numbered acids increases. Under conditions favoring acid formation over paraffin formation a secular equilibrium would be attained eventually in which no preference between even- or odd-numbered paraffins or acids exists. Distributions of fatty acids in recent and ancient sediments and petroleum reservoir waters support this suggested scheme. EXPERIMENTAL Reagents. Reagents used were the best commercial grades available. Solvents were redistilled and checked for purity. The beeswax sample was obtained from Coleman and Bell. In a 2 1. round-bottom flask equipped with a reflux Extraction of sediments. condenser 40 g of KOH and 500 ml of methanol were refluxed until most of the KOH had dissolved. The air-dried sediment sample was added, and the sample was relluxed for 2 hr. About 200 ml of solution was decanted, 200 ml of methanol was added to the flask, and the sample was refluxed for an additional hr. The reaction mixture was cooled, filtered, and combined with the previously decanted About 500 ml of solution was recovered from each experiment; the solution. concentrations reported for the acids have been adjusted for solvent recovery. The recovered solution and 100 ml of water were added to a 2 1. flask. The solution was concentrated to 200 ml by removing solvent through a 2 ft Vigreux column. The concentrate was extracted with successive 75 ml, 50 ml and 25 ml portions of Ccl,. This solution was divided into two equal portions if an internal standard was to be used. A standard solution of stearic acid added to one portion served as an internal standard. The portions were acidified with 25 ml of concentrated HCI. The solutions were extracted successively with 30 ml, 15 ml and 15 ml portions of Ccl,. The solutions were concentrated by removing the solvent through a 2 ft Vigreux column. The residue was transferred to a 2 dram screw-cap vial and the solvent was removed under a stream of air at 40’. Extraction of water. To a water sample was added 2g of KOH. The solution was concentrated to 200 ml by distillation through a 2 ft Vigreux column. The concentrate was washed with Ccl,, acidified, extracted with CC&, and concentrated using the technique described for sediment extracts. No internal standards were employed with water samples. Extraction of beeswax. A 500 mg sample of beeswax was refluxed for 2 hr with The solution was subsequently a solution of 10 g of KOH in 100 ml of methanol. treated by the procedure used for sediment extracts. Preparation of methyl esters. To the extract was added 3 ml of boron trifluoride-methanol solution (METCALFE and SCHMITZ, 1961).This mixture was heated on a steam bath for 5 min. To the mixture was added 1 ml of benzene and 2 ml the benzene layer was removed and of water. The mixture was centrifuged; transferred to another 2 dram vial.

A postulated role of fatty aoids in petroleum formation

1127

Urea adduction of esters. To the benzene solution of the methyl esters was added 3.5 ml of saturated urea in methanol. The mixture was cooled to near 0°C overThe mixture was centrifuged, the night. Crystals of the urea adduct formed. liquid layer was drawn off, and the crystals were washed twice with 1 ml portions of benzene. The urea adduct was decomposed with 1 ml of water; 0.15 ml of benzene was added, the mixture was centrifuged and the organic layer was removed. This solution was analysed directly by gas chromatography. An F & M Model 500 Gas Chromatograph was Gas chromatography analysis. used for gas chromatographic analyses. The instrument was equipped with a 2ft column packed with O-3 per cent silicone grease (Dow Corning) on 60-80 mesh glass beads. The initial temperature in each run was 75”, and it was increased to 300” at the rate of 5*6”/min. The temperature was held at 300’ until the esters were eluted completely. The gas chromatograph was calibrated with a standard solution of esters of fatty acids having 8, 10, 12, 14, 16, 18 and 20 carbon atoms. Retention times and instrument sensitivities for esters of other aliphatic acids were obtained by interpolation and extrapolation. Mass spectral analysis. A modified Consolidated Electrodynamics Corporation Model 21-103 Mass Spectrometer was operated as described by BRAY and EVANS (1961).

REFERENCES ABELSON P. H. and PARKER P. L. (1962) Fatty acids in sedimentary rooks. Carnegie In&. Washington, Yearbook 61, 181-184. BRAY E. E. and EVANS E. D. (1961) Distribution of n-paraffins as a clue to recognition of source beds. Geochim. et Cosmochim. Acta 22, 2-15. BAKER E. G. (1959) Origin and migration of oil. Science 129, 871-874. BENDORAITIS J. G., BROW B. L. and HEPNER L. S. (1962) Isoprenoid Hydrocarbons in Petroleum. Analyt. Chem. 34, 49-53. Cox B. B. (1946) Transformation of organic material into petroleum under geological conditions. Bull. Amer. Ass. Petrol. Geol. 30, 645-659. DEAN R. A. and WHITEHEAD E. V. (1961) The occurrence of phytane in Petroleum. Tetrahedron Letters 21, 768-770. GERARDE H. W. and GERARDE D. F. (1961 and 1962) The Ubiquitous Hydrocarbons. Aes. Food and Drug Oflcials U.S. 25 and 26, l-47. MEINSCHEIN W. G. (1959) Origin of Petroleum. Bull. Amer. Ass. Petrol. Geol. 43, 925-943. MEINSCHEIN W. G. and KENNEY G. S. (1957) Analyses of a chromatographic fraction of organic extracts of soils. Analyt. Chem. 29, 1153-1161. METCALFE L. D. and SCHMI~Z A. A. (1961) The rapid preparation of fatty acid esters for gas chromatographic analysis. Analyt. Chem. 33, 363-364. RALSTON A. W. (1948) Fatty acids and their Deriwativee. Wiley, New York. ROSSINI F. D. (1960) Hydrocarbons in petroleum. J. Chem. Educ. 37, 554-561. SILLIKER J. H. and RITTENBERG S. C. (1952) Studies on the aerobic oxidation of fatty acids by bacteria-III. The effect of 2,4-dinitrophenol in the oxidation of fatty acids by Serratia Marcescens. J. Bact. 64, 197-205. SHORLAND F. B. (1954) Occurrence of fatty acids with uneven-numbered carbon atoms in natural fats. Nature, Lond. 174, 603. TREIBS A. (1936) Chlorophyll and hemin derivatives in organic minerals. Angew. Chem. 49, 682-686.