Molecular and Stable Carbon Isotopic Source Identification of Oil

Research Molecular and Stable Carbon Isotopic Source Identification of Oil Residues and Oiled Bird Feathers Sampled along the Atlantic Coast of France after the Erika Oil Spill L. MAZEAS AND H. BUDZINSKI* Laboratoire de Physico-et Toxico-Chimie des Syste`mes Naturels (UMR 5472 CNRS), Universite´ Bordeaux I, 351 cours de la libe´ration, 33405 Talence Cedex, France

The Erika tanker broke in two close to the Atlantic coast of France on December 12, 1999. On December 25th, some heavy fuel oil released by the tanker came ashore along the French Atlantic Coast. Some oil residues and oiled bird feathers were collected all along the Atlantic Shoreline of France after the wreck of the Erika tanker. The aim of this study was to differentiate oil residues and oiled bird feathers related to the Erika oil spill from the ones resulting from the numerous tar ball incidents which had occurred after the Erika oil spill. Alkane and PAH quantification of oil residues allowed differentiation of the samples collected on the north part of the Atlantic Coast from those collected on the south part of the Atlantic shoreline. All oiled birds appear to have been contaminated by the Erika oil. Samples collected on the south part of the Atlantic Coast contain a different molecular fingerprint compared to the Erika oil indicating that they are not related to the Erika oil spill. Bulk and molecular 13C/12C ratio measurements were performed in order to check the discriminative feature and the stability of the isotopic approach. Bulk stable carbon isotopic composition has been shown to be a valuable screening correlation tool as it confirms the link of samples collected in the north part of the Atlantic Coast with the Erika oil spill. All the samples collected along the south part of the Atlantic Shoreline exhibit 13Cenriched bulk isotopic compositions compared to Erika oil. Molecular isotopic composition of saturated hydrocarbons and of phenanthrene compounds also allows unambiguous differentiation of samples related to the Erika oil spill from those due to tar ball incidents. Over the long-term, when molecular distribution will have been modified by the different processes affecting oil in the marine environment, molecular isotopic composition should then be of particular help for Erika oil residues identification.

Introduction The Maltese tanker Erika broke in two close to the Atlantic Coast of Brittany in France in the early morning of December 12, 1999. About 10 000 tons of oil are believed to have spilled into the sea (itopf web site, http:www.itopf.com). Oil from the tanker Erika came ashore on December 25, 1999. Many other tankers took advantage of this oil spill to clean out their tanks along the Atlantic Coast of France. Identification 130

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 2, 2002

of the source of oil residues and oil found on birds washed ashore was then of particular importance after the Erika oil spill. Crude oils and refined petroleum products exhibit different aliphatic and aromatic hydrocarbons molecular distributions depending upon the origin and thermal maturity (1-3). Most oil spill identification studies rely on molecular fingerprint recognition techniques (4-7). However, after its release in the environment, oil undergoes different abiotic (evaporation, emulsification, dissolution, photooxidation, ...), and biotic (biodegradation, bioaccumulation, ...) processes leading to selective modification of the initial molecular profiles (6, 8-10). Correlation of weathered oil residues with their suspected sources can be inconclusive or ambiguous. Therefore, measurement of additional properties that exhibit differences between crude oils are needed to improve oil spill source correlation studies. Both bulk and molecular stable carbon isotopic compositions have such a discriminative characteristic. Indeed, the bulk stable carbon isotopic composition of crude oils, which is determined by the origin and the thermal maturity of the oils (11-13), has been useful in identifying oil pollution sources (14-18). Molecular carbon isotopic composition of petrogenic hydrocarbons has also been shown to differ according to the origin and the thermal maturity of crude oils (19-23). Hydrocarbon molecular 13C/ 12C ratio measurement appears then of particular interest for oil spill investigations (24-26). Both saturated and aromatic hydrocarbon quantitative and qualitative considerations and stable carbon isotopic composition (bulk and molecular) measurements were applied to oil residues and oiled bird feathers sampled after the Erika oil spill along the Atlantic Coast of France in order to identify samples related to the Erika oil spill.

Experimental Section Samples. A sample of the heavy fuel oil loaded by the Erika tanker at the Flandres refinery (Dunkerque, France) and a sample collected from an oil slick released by the Erika tanker were provided by the CEDRE (French institute for water pollution, Brest, France). Oil residues sampled on the Atlantic Coast after the oil spill are listed in Table 1 and located in Figure 1. Oiled bird feathers collected on the Southern Atlantic Shoreline were provided by the Teich Ornithological Park (Gironde, France) (Table 1). The exact sample locations are unknown except for S28 which was sampled on a beach in León (Landes, France). Alkane and PAH Quantification. Detailed descriptions of sample preparation and hydrocarbon quantification procedures have been reported previously (27). However a brief summary follows. After asphaltene precipitation in pentane, the maltene fraction was purified using an alumina microcolumn. The aliphatic fraction was eluted off a silica column with pentane and the aromatic fraction with a mixture of pentane/dichloromethane (65/35, v/v). The alkanes and PAHs were then quantified separately using a gas chromatograph coupled to a mass spectrometer detector by comparison to perdeuterated alkane and PAH internal standards introduced at the beginning of the sample preparation procedure. Bulk Stable Carbon Isotopic Composition Measurement. Bulk stable carbon isotopic composition measurements were conducted using an elemental analyzer (Carlo Erba, NC 2500) coupled to a Delta Plus isotopic ratio mass spectrometer (Finnigan MAT, Bremen, Germany). Each oil sample was 10.1021/es010726a CCC: $22.00

 2002 American Chemical Society Published on Web 12/13/2001

TABLE 1. Analyzed Oil Residues and Oiled Bird Feathers Collected along the Atlantic Coast of France after the Erika Oil Spill no.

name

description

date

S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28

Erika oil Oil slick Guilvinec rock Concarneau rock Ploemeur rock Etel rock Noirmoutier tar ball Ploemeur tar ball Etel tar ball Kerrorck rock Kerpape rock Ploemeur rock 2 Crohot tar ball Cap Ferret tar ball Salie small tar ball Salie big tar ball Pereire tar ball Chapelle Forestie` re Etel guillemot Teich guillemot 1 Teich guillemot 2 Teich scoter Teich gannet 1 Teich gannet 2 Teich gannet 3 Teich grand labre Teich gannet 4 Le´ on guillemot 4

oil from the refinery oil from an oil slick oil from rocks of the Guilvinec port oil from rocks of a beach in Concarneau oil from rocks of Fort Bloque´ beach in Ploemeur oil from rocks of a beach of Etel tar ball from a beach of Noirmoutier island tar ball from Fort Bloque´ beach in Ploemeur tar ball from a beach at the entrance of Etel river oil from Kerroch beach in Ploemeur oil sampled on Kerpape beach in Ploemeur oil from rocks of Fort Bloque´ beach in Ploemeur tar ball sampled on Crohot beach tar ball from Cap Ferret beach small tar ball from Salie beach big tar ball from Salie beach tar ball from Pereire beach tar ball from Chapelle Forestie` re beach oiled feather from a guillemot form Etel beach oiled feather from a guillemot (Teich park) oiled feather from a guillemot (Teich park) oiled feather from a scoter (Teich park) oiled feather from a gannet (Teich park) oiled feather from a gannet (Teich park) oiled feather from a gannet (Teich park) oiled feather from a “grand labre” (Teich park) oiled feather from a gannet (Teich park) oiled feather from a guillemot (Le´ on)

12/12/1999 12/29/1999 12/29/1999 12/27/1999 12/27/1999 10/01/1999 12/27/1999 12/27/1999 01/10/2000 01/09/2000 02/17/2000 12/22/1999 12/23/1999 02/21/2000 02/21/2000 03/02/2000 03/02/2000 12/27/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999 12/30/1999

FIGURE 1. Sampling locations of oil residues and oiled bird feathers collected along the Atlantic Coast of France after the Erika oil spill.

Compound Specific Alkane and PAH Stable Carbon Isotopic Composition Measurement. Before gas chromatography-isotopic ratio mass spectrometry analysis (GCIRMS), saturated and aromatic fractions were isolated using the same analytical procedure as in the quantification process. Saturated fractions were analyzed directly without further treatment. However, to measure the isotopic composition of methylphenanthrenes, it was necessary to separate them from methyldibenzothiophenes as these classes of compounds coelute during gas chromatographic separation using a classical apolar capillary column. The phenanthrenic fraction was isolated by high-pressure liquid chromatography separation on an aminosilane phase (Spherisorb, 5 µm, 25 cm, 4.6 mm i.d.) using pentane (Mallinckrodt Nanograde, Atlantic Labo, Floirac, France) as the mobile phase (28). GC-IRMS analyses were then performed on the phenanthrenic fractions. Stable carbon isotopic composition analyses of individual alkanes and PAHs were carried out using an HP 5890 Series II Plus gas chromatograph interfaced via a CuO furnace (940 °C) and a hygroscopic membrane (Nafion) to a Delta Plus isotopic ratio mass spectrometer. Injections were performed in the splitless mode. The injector temperature was maintained at 270 °C. The GC temperature program for alkanes was from 50 °C (2 min) to 290 °C (20 min) at 5 °C.min-1. For phenanthrene fraction analyses, the GC temperature program was from 50 °C (2 min) to 180 °C (2 min) at 10 °C.min-1, to 230 °C at 2 °C.min-1, and at 10 °C.min-1 to 290 °C (10 min). The carrier gas was helium (constant flow rate: 1 mL.min-1). The capillary column used was a SGE BPX5: 60 m × 0.22 mm ID × 0.25 µm film thickness.

Results and Discussion analyzed three times (30 µg per analysis). In addition, reference matrices were analyzed repeatedly in order to confirm the accuracy and reproducibility of the analyses. The reference matrices used were the graphite USG 24 (NIST, Gaithersburg, MD) and the Crude oil NBS 22 (IAEA, Vienna, Austria).

Molecular Approach. Saturated Fraction. Although n-alkanes and isoprenoids are not of toxicological concern, analysis of these compounds is useful for oil spill identification and for fate study of hydrocarbons after an oil spill (2, 7). Figure 2 compares the GC-MS (SIM, m/z 57) chromatogram of Erika oil (S1) with the chromatograms obtained for some repreVOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

131

FIGURE 2. GC-MS chromatograms (SIM, m/z ) 57) of saturated fractions of Erika oil and of oil residues collected along the Atlantic Coast of France after the Erika oil spill. The chromatograms S2, S3, and S10 are similar to the ones obtained for all oil residues collected in the north part of the Atlantic Coast. sentative samples collected in the north part of the Atlantic Coast (i.e., S2, S3, and S10) and the tar balls collected in the 132

9


Arcachon Bay area (i.e., S13 to S18). Erika oil shows a saturated fraction gas chromatographic trace typical of a heavy residual

TABLE 2. PAH Concentrations (µg/g) Measured in Some of the Oil Residues Collected along the Atlantic Coast of France Compared with Erika Oila

phenanthrene C1-phen C2-phen C3-phen dibenzothiophene C1-DBT C2-DBT C3-DBT anthracene fluoranthene pyrene benz (a) antracene chrysene + triphenylene benzo(b+k)fluoranthene benzo(e)pyrene benzo(a)pyrene perylene indenopyrene benzo(ghi)perylene dibenz(ah)anthracene

Erika oil

oil slick

Guilvinec

Kerroch

Crohot

Cap Ferret

Salie STB

Salie BTB

Pereire

Forestiere

(S1) 614 2743 3172 2748 226 617 1226 1251 90 46 218 152 324 69 101 96 46 15 40 23

(S2) 350 1335 1801 1739 123 331 573 704 63 31 196 183 324 58 90 100 33 17 38 29

(S3) 232 922 1230 1158 79 218 378 479 35 23 131 147 271 44 80 78 27 14 32 23

(S10) 275 1156 1816 1724 98 297 555 750 43 28 169 140 215 54 85 93 37 11 40 24

(S13) 14 118 264 267 4 25 53 58 3 nd nd nd nd nd nd nd nd nd nd nd

(S14) nd nd 211 264 nd nd 66 114 nd nd nd nd nd nd nd nd nd nd nd nd

(S15) 32 225 351 301 24 159 309 334 nd nd 5 0 25 2 6 nd nd nd nd nd

(S16) 35 196 353 266 52 295 576 567 nd nd 6 nd 15 2 4 nd nd nd nd nd



a The concentrations given for S2, S3, and S10 are representatives of the ones found for all oil residues collected in the north part of the Atlantic Coast.

oil (7) with an homologous series of n-alkanes with a maximum around n-C26. Samples collected on the north Atlantic Coast (i.e., S2-S12) exhibit a chromatographic trace very close to that of the Erika oil. However, all samples collected in the Arcachon bay area display molecular profiles different from that of the Erika oil. They differ from the Erika oil by having a higher proportion of alkanes between n-C15 and n-C21. These samples do not seem then to be related with the Erika oil spill as those light n-alkanes should have been preferentially lost in oil residues due to their higher volatility than heavier n-alkanes. None of the samples collected in the Arcachon Bay area show similar molecular distributions except for the big tar ball sampled on Salie beach (S16) and tar balls collected on Pereire (S17) and on Chapelle Forestiere (S18) beaches, all of which seem to come from the same source. The differences in molecular profile observed among these tar ball samples collected on the South Atlantic Coast of France can be explained by differing origins or simply by the fact that the samples have been altered by the processes affecting oil in the marine environment (evaporation, photooxidation, biodegradation, ....). For example, S15 has a notably higher amount of pristane and phytane when compared to n-C17 and n-C18, respectively, than the other samples (Figure 2). As biodegradation preferentially removes n-alkanes relative to isoprenoids (29), n-C17/pristane and n-C18/phytane ratios cannot be used for determining contamination sources. The use of pristaneto-phytane ratio, which is less sensitive to biodegradation, distinguishes S13, S14, and the group S15, S16, S17, S18. The alkane molecular distribution discrepancies observed between the small (S15) and the big tar balls (S16) sampled on the Salie beach may be explained by a different biodegradation extent. Aromatic Fraction. PAHs, especially alkylated PAHs, are useful for oil spill investigations in that their molecular distributions in crude oils are different depending upon the origin and thermal maturity of the oil (30, 31, 3). Moreover, the study of PAH origin and behavior is of ecotoxicological interest as these compounds exhibit mutagenic and carcinogenic properties (32). Table 2 gives PAH concentrations measured for Erika oil (S1), for some representative oil residues collected on the north part of the Atlantic Coast (i.e., S2, S3, S10) and for tar

ball samples collected in the Arcachon Bay area (i.e., S13S18). Erika oil shows a PAH molecular distribution typical for a Bunker C fuel (7). Naphthalene, phenanthrene, and dibenzothiophene compounds are the predominant PAHs. Pyrene is much more abundant than fluoranthene, and chrysene is particularly abundant. Those molecular characteristics can be used as a tracer of origin. Erika oil contains high concentrations of high molecular weight compounds such as benzo[a]pyrene which is one of the most toxic PAHs. Erika oil and samples S2, S3, and S10 show much higher phenanthrene and dibenzothiophene compound concentrations than the samples collected in the Arcachon Bay. High molecular weight PAHs are much more abundant in the Erika oil, S2, S3, and S10 than in the samples collected in the Arcachon Bay area, in which high molecular weight PAHs were either not detected or exist in very low amounts. The quantification of PAHs confirms that S2-S12 are related to the Erika oil spill, whereas Arcachon bay samples S13-S18 result from different tar ball incidents. However, the Arcachon Bay Samples do not all show similar aromatic profiles. Samples S13-S15 may come from different sources, but S16, S17, and S18 are probably from the same source as they exhibit similar molecular distributions. The use of the relative distribution of alkylated PAHs as a source diagnostic parameter has been investigated as it was also possible to use it for oiled bird feathers. On Figure 3, the relative distribution of methylphenanthrenes (C1-P) and of methyldibenzothiophenes (C1-DBT) determined in the oil samples and oiled bird feathers are compared with those of Erika oil. Samples collected on the north part of the Atlantic Shoreline (S2-S12) show similar C1-P (Figure 3A) and C1-DBT (Figure 3C) relative distributions than Erika oil, featured respectively by the predominance of the 2-MP and the 2+3 -MDBT. Arcachon Bay area samples (S13-S18) exhibit a predominance of the 9-MP and the 4-MDBT. This confirms that those latter samples are not related with the Erika oil spill. They also appear not to come all from the same tar ball incident as they show slightly different profiles. The Crohot beach sample seems to come from a different source than the other samples collected in the same area. Methylphenanthrene (Figure 3B) and methyldibenzothiophene (Figure 3D) relative distributions determined for oiled bird feathers collected along the Atlantic Shoreline are VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

133

FIGURE 3. Methylphenanthrene relative distributions determined in oil residues (A) and oiled bird feathers (B) compared whit that of Erika oil. Methyldibenzothiophene relative distributions determined in oil residues (C) and oiled bird feathers (D) compared with that of Erika oil. very close to those of Erika oil. All those birds might then have been contaminated by the oil released by the Erika tanker. The examination of the relative distribution of C2- to C4naphthalenes, C2-phenanthrenes, and C2-dibenzothophenes has also been helpful for the source correlation of those 134

9


samples. For example, the Principal Component Analysis (PCA) performed using relative distributions of dimethylphenanthrenes as variables is presented in Figure 4. The two first components account for 82% and 7%, respectively, of the variance of the original data set. Figure 4a displays the relationships between the variables. The most stable dim-

FIGURE 5. Plot of C3-DBT/C3-P versus C2-DBT/C2-P for oil residues and oiled bird feathers collected along the Atlantic Coast of France after the Erika oil spill. See Table 1 and Figure 1 for sample identification and location.

FIGURE 4. Principal component analysis performed on oil residues and oiled bird feathers, collected along the Atlantic Coast of France after the Erika oil spill, using dimethylphenanthrene relative distributions as variables: (A) variable map and (B) sample map. See Table 1 and Figure 1 for precise sample identification and location. ethylphenanthrene isomers (i.e., 2,7; 2,6; and 3,6) (33) are in the negative half side of the first principal component axis (PC1), whereas the other isomers (less stable) are on the opposite direction. The PC1 axis can then be considered as a maturity axis. On Figure 4b, which displays the different samples on the same principal component axes, all the oil residues collected in the north part of the Atlantic Coast and all the oiled bird feathers are grouped together with the Erika oil. These results confirm the link with the Erika oil spill. Samples collected in the Arcachon bay area do not correlate with the Erika oil, nor do they all correlate with one another. This confirms that they have not been released by the Erika tanker and that they do not come from the same source. The differentiation between S15 and S16 is not obvious using dimethylphenanthrene relative distribution. Ratios between phenanthrene and dibenzothiophene compounds are particularly useful for source apportionment of tar balls as they vary among oils having different sulfur content (34, 5). Moreover, it has been shown that those indices are not significantly affected by weathering and biodegradation processes (6). Figure 5 represents the plot of C3-DBT/ C3-P versus C2-DBT/C2-P ratios determined for all the oil residues and oiled bird feathers. All the samples from the north part of the Atlantic Coast and all oiled bird feathers are grouped on this plot and are related with the Erika oil spill. Samples collected on the Crohot beach (S13) are distinguished by lower C2-DBT/C2-P and C3-DBT/C3-P ratios. They come from a sulfur depleted source compared to the Erika oil. The tar balls collected on the Cap Ferret beach show C2-DBT/C2-P and C3-DBT/C3-P ratios similar to those of Erika oil. Using these indices, it is therefore impossible to differentiate these tar balls from Erika oil residues. The big

FIGURE 6. Bulk stable carbon isotopic composition determined for Erika oil and oil residues collected along the Atlantic Coast of France after the Erika oil spill. The values given are the mean of the three analyses performed for each sample. (Samples S2-S12 were collected in the North part of the Atlantic Coast, and samples S13-S18 were collected in the South part of the Atlantic Coast). See Table 1 and Figure 1 for precise sample identification and location. tar balls collected on the Salie beach (S16), on the Pereire (S17), and on the Chapelle Forestie`re (S18) beaches show higher similar ratios. These tar balls come from sulfur enriched sources compared to Erika oil. The use of these indices allows unambiguous confirmation that the small (S15) and the big (S16) tar balls collected on Salie beach come from different sources. Bulk Stable Carbon Isotopic Approach. The bulk stable carbon isotopic composition of all oil samples collected on the Atlantic shoreline is compared with the Erika oil (Figure 6). Bulk stable carbon isotopic composition of oils found on bird feathers were also measured, but they are not reported because of large uncertainties probably introduced by the presence of interfering compounds from the feathers. All samples collected in the north part of the Atlantic Coast of France (i.e., S2-S12) show a stable carbon isotopic composition (mean for the 10 samples: -28.9 ( 0.1) similar to that of Erika oil (-28.9 ( 0.04 (n ) 3)), whereas all the samples collected in the Arcachon Bay area exhibit isotopic compositions significantly different from the Erika oil. Samples collected on the Crohot beach (S13) and on the Cap Ferret beach (S14) cannot be differentiated from each other using their bulk stable carbon isotopic compositions. Samples collected on the Salie (S15, S16), Pereire (S17), and Chapelle Forestiere (S18) beaches seem all to come from the same source according to their bulk isotopic compositions. The small and the big tar balls collected on the Salie beach are not clearly differentiated by their bulk 13C/12C ratio. VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

135

FIGURE 7. Alkanes isotopic composition of oil residues and oiled bird feathers collected along the Atlantic Coast of France compared with Erika oil alkanes isotopic composition. Values given are as follows: S1: mean of three analyses; S2-S12: average between the mean values (n ) 3) measured for each sample; S14-S18: average between the mean values (n ) 3) measured for each sample; and S13: mean of three analyses.

FIGURE 8. Alkane isotopic composition of oil residues collected in the Arcachon Bay area compared with Erika oil alkane isotopic composition. Values are the mean of three analyses. The bulk stable carbon isotopic composition proves to be a reliable discriminative feature. Indeed, conclusions inferred from the bulk 13C/12C ratio measurement are in good

agreement with molecular observations. Moreover, this oil property seems stable as the 13C/12C ratios in shoreline residues, which are suspected to be related with the Erika oil spill (i.e., S2-S12), have remained unchanged from the original cargo oil. Similar observations were reported in the case of the Exxon Valdez oil spill even for oil residues sampled 5 years after the spill (16). Bulk stable carbon isotopic composition seems to be a useful screening tool for oil spill identification as the analysis is rapid (10 min) and does not require any sample preparation. Molecular Stable Carbon Isotopic Composition. Alkane Isotopic Compositions. Molecular n-alkane isotopic compositions of the oil residues collected in the north Atlantic shoreline (mean of S2-S12), on the Crohot Beach (S13), in the Arcachon Bay area (mean of S14-S18), and of bird feathers (mean of S19-S28) are compared with Erika oil (Figure 7). The data are presented as “isotope profiles”, which are plots of n-alkane carbon number versus 13C/12C ratio. Samples collected on the north part of the Atlantic Coast and all bird feathers are well correlated and show n-alkane isotopic profiles close to the Erika oil. This is additional independent proof of the link between these samples and the Erika oil spill. The tar ball samples collected on the Crohot beach show 13C-enriched isotopic composition compared to the Erika oil especially for the higher molecular weight n-alkanes. All the other samples collected in the Arcachon Bay area present intermediate isotopic compositions (between the Crohot beach tar ball and Erika oil). Nevertheless, the high standard deviation associated to the mean of those samples (S14-S18) suggests that they are from different origins. The n-alkane isotopic composition of each of those samples are represented in Figure 8. It is clear that none of those samples collected in the Arcachon Bay exhibit a similar isotopic profile to the Erika oil. From the examination of n-alkane isotopic profiles, samples S16, S17, and S18 appear to be correlated and S14 and S15 seem to come from independent sources, since their molecular isotopic profiles exhibit significant differences. Polycyclic Aromatic Hydrocarbon Isotopic Compositions. The molecular isotopic compositions of phenanthrene compounds contained in Erika oil, in the oil residues collected on the north part of the Atlantic Coast (mean of S2, S5, S6,

FIGURE 9. Phenanthrene compound isotopic composition of oil residues and oiled bird feathers collected along the Atlantic Coast of France compared with Erika oil isotopic composition. S1: mean of three analyses. North Atlantic: average between the mean values (n ) 3) of S2, S5, S6, S7, S8, S9, S10, and S11. Bird feathers: average between the mean values (n ) 3) of S19, S22, and S24. S13, S15, S16, S17, S18: mean of three analyses. The absence of phenanthrene compounds in S14 prevented GC-IRMS analysis. 136

9


S7, S8, S9, S10, S12), in some oiled bird feathers (mean of S19, S22, S24), and in the oil residues collected in the Arcachon Bay area are compared in Figure 9. Individual methylphenanthrene isotopic compositions are not reported because erroneous isotopic composition can be obtained for individual compounds due to the fact that 2-MP and 3-MP and 9-MP and 1-MP are not totally resolved as previously described (35). The isotopic compositions of the sums 3+2-MP and 9+1-MP determined by integrating the two peaks representing each compound together (respectively 3- and 2-MP on one hand and 9- and 1-MP on the other hand) were then measured. The isotopic compositions of the sums of the C1-P, the C2-P, and the C3-P were also measured by integrating together all the isomer peaks. Samples collected on the north part of the Atlantic Coast and all bird feathers are well correlated and show phenanthrenic isotopic profiles close to that of the Erika oil. This is additional proof for the link between these samples and the Erika oil spill. However, all the samples collected in the Arcachon Bay area exhibit isotopic compositions significantly enriched in 13C compared to the Erika oil. This shows, as with the previous evidence, that they are not linked to the Erika oil spill. Oil residues sampled on the Pereire beach and the big tar balls collected on the Salie beach exhibit similar isotopic profiles, whereas samples collected on the Crohot beach and the small tar ball collected on the Salie beach exhibit significant individual isotopic composition discrepancies showing that they come from independent sources. As the molecular isotopic composition of hydrocarbons has been previously shown to not be affected during degradation of oil (36-38), its use should be helpful for the identification of weathered residues. Molecular distributions would be specifically modified in such samples.

Acknowledgments Franç ois Xavier Merlin and Julien Guyomarc’h from the CEDRE are thanked for providing the sample of the oil loaded by the Erika Tanker at the Dunkerque refinery and of the sample collected from an oil slick. Mrs. and Mr. Amiard Triquet, Olivier Geffard, and Laurent Jeffroy are acknowledged for providing some oil residues. The Teich Ornithological Park is thanked for providing oiled bird feathers. Philippe Garrigues and Pascal Mora are thanked for the sampling of the oil bird feathers. Pascal Mora is also thanked for his assistance with the PCA analysis. Sylvie Mauresmo is thanked for the analyses she performed during her training period. This work was done in the framework of the HYCAR Project (“Biogeochemical cycle of natural and anthropic hydrocarbons in marine systems”). Elf Aquitaine and CNRS are acknowledged for financial support.

Literature Cited (1) Tissot, B.; Welte, D. H. Petroleum formation and occurrence, 2nd ed.; Springer-Verlag: Berlin, 1984. (2) Kaplan, I. R.; Galperin, Y.; Lu, S. T.; Lee, R. P. Org. Geochem. 1997, 27, n°5/6, 289-317. (3) Budzinski, H.; Garrigues, P.; Connan, J.; Devillers, J.; Domine, D.; Radke, M.; Oudin, J. L. Geochim. Cosmichim. Acta 1995, 59, n°10, 2043-2056. (4) Kvenvolden, K. A.; Hostettler, F. D.; Rapp, J. B.; Carlson, P. R. Mar. Pollut. Bull. 1993, 26, 24-29. (5) Bence, A. E.; Kvenvolden, K. A.; Kennicutt, M. C. Org. Geochem. 1996, 24, n°1, 7-42.

(6) Douglas, G. S.; Bence, A. E.; Prince, R. C.; McMillen, S.; Butler, E. L. Environ. Sci. Technol. 1996, 30, 2332-2339. (7) Wang, Z.; Fingas, M.; Page, D. S. J. Chromatogr. A 1999, 843, 369-411. (8) Bayona, J. M.; Albaiges, J.; Solanas, A. M.; Pares, R.; Garrigues, P.; Ewald, M. Intl. J. Environ. Anal. Chem. 1986, 23, 289-303. (9) Wang, Z.; Fingas, M. J. Chromatogr. A 1995, 712, 321-343. (10) Budzinski, H.; Raymond, N.; Nadalig, T.; Gilewicz, M.; Garrigues, P.; Bertrand, J. C.; Caumette, P. Org. Geochem. 1998, 28, n°5, 337-348. (11) Peters, K. E.; Rohrbach, B. G.; Kaplan, I. R. Am. Assoc. Petrol. Geol. Bull. 1981, 65, 501-508. (12) Lewan, M. D. Geochim. Cosmochim. Acta 1983, 47, 1471-1480. (13) Hoefs, J. Stable isotope geochemistry, 4th ed.; Spinger edition; 1997; p 201. (14) Calder, J. A.; Parker, P. L. Environ. Sci. Technol. 1968, 2, 535539. (15) Macko, S. A.; Parker, P. L. Marine Environ. Res. 1983, 10, 93103. (16) Kvenvolden, K. A.; Hostettler, F. D.; Carlson, P. R.; Rapp, J. B.; Threlkeld, C. N.; Warden, A. Environ. Sci. Technol. 1995, 29, 2684-2694. (17) Whittaker, M.; Pollard, S. J. T.; Fallick, A. E.; Preston, T. Environ. Poll. 1996, 94, n°2, 195-203. (18) Wang, Z.; Fingas, M.; Landriault, M.; Sigouin, L.; Castle, B.; Hostetter, D.; Zhang, D.; Spencer, B. J. High Resolut. Chromatogr. 1998, 21, 383-395. (19) Bjorøy, M.; Hall, P. B.; Hustad, E., Williams, J. A. Org. Geochem. 1991, 19, n° 1-3, 89-105. (20) Schoell, M.; McCaffrey, M. A.; Fago, F. J.; Moldowan, J. M. Geochim. Cosmochim. Acta 1992, 56, 1391-1399. (21) Bjorøy, M.; Hall, P. B.; Moe, R. P. Org. Geochem. 1993, 22, n° 3-5, 355-381. (22) Schoell, M.; Simoneit, B. R. T.; Wang, T. G. Org. Geochem. 1994, 21, n° 6/7, 713-719. (23) Simoneit, B. R. T.; Schoell, M. Org. Geochem. 1995, 23, n°9, 857-863. (24) Dowling, L. M.; Boreham, C. J.; Hope, J. M.; Murray, A. P.; Summons, R. E. Org. Geochem. 1995, 23, n°8, 729-737. (25) Mansuy, L.; Philp, R. P.; Allen, J. Environ. Sci. Technol. 1997, 31, 3417-3425. (26) Hammer, B. T.; Kelley, C. A.; Coffin, R. B.; Cifuentes, L. A.; Mueller, J. G. Chem. Geol. 1998, 152, 43-58. (27) Mazeás, L.; Budzinski, H. Intl. J. Environ. Anal. Chem. in press. (28) Budzinski, H.; Garrigues, P.; Connan, J.; Bellocq, J. Quim. Anal. 1993, 12, 69-74. (29) Peters, K. E.; Moldowan, J. W. The biomarker guide: interpreting molecular fossils in petroleum and ancient sediments; Prentice Hall, NJ, 1993. (30) Radke, M.; Welte, D. H.; Willich, H. Org. Geochem. 1986, 10, 51-63. (31) Strachan, M. G.; Alexander, R.; Kagi, R. I. Geochim. Cosmochim. Acta 1988, 52, 1255-1264. (32) White, K. L. Carcino. Rev. 1986, C4(2), 163-202. (33) Budzinski, H.; Garrigues, P.; Radke, M.; Connan, J.; Oudin, J. L. Org. Geochem. 1993, 20, n°7, 917-926. (34) Sauer, T. C.; Brown, J. S.; Boehm, P. D.; Aurand, D. V.; Michel, J.; Hayes, M. O. Mar. Pollut. Bull. 1993, 27, 117-134. (35) Mazeás, L.; Budzinski, H. submitted to J. Chromatogr. (36) Trust, B. A.; Coffin, R. B.; Cifuentes, L. A.; Mueller, J. G. In Monitoring and verification of bioremediation; Hinchee, R. E., Douglas G. S., Ong, S. K., Eds.; Battelle Press: Columbus, OH, 1995; 233-239. (37) Huang, Y.; Eglinton, G.; Ineson, P.; Latter, P. M.; Bol, R.; Harkness, D. D. Org. Geochem. 1997, 23, n° 6/7, 497-501. (38) Mazeás, L.; Budzinski, H.; Raymond, N. Submitted to Org. Geochem.

Received for review March 13, 2001. Revised manuscript received October 15, 2001. Accepted October 17, 2001. ES010726A

VOL. 36, NO. 2, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

137

Molecular and Stable Carbon Isotopic Source Identification of Oil

Recommend Documents