Environ. Sci. Technol. 1997, 31, 3417-3425
Source Identification of Oil Spills Based on the Isotopic Composition of Individual Components in Weathered Oil Samples LAURENCE MANSUY, R. PAUL PHILP,* AND JON ALLEN School of Geology and Geophysics, The University of Oklahoma, Norman, Oklahoma 73019
Correlation of crude oils and refined hydrocarbon products spilled in the environment with their respective sources is commonly undertaken using techniques such as gas chromatography (GC) or gas chromatography/mass spectrometry (GC/MS) or bulk parameters such as the isotopic composition of the aliphatic or aromatic fractions. Under certain circumstances extensive weathering of the samples, through evaporation, water-washing, or biodegradation, may make such correlations extremely difficult and the results somewhat tenuous. Results are presented in this paper from an investigation to study the use of gas chromatography/ isotope ratio mass spectrometry (GC/IRMS) as a complimentary correlation technique to GC and GC/MS, particularly for samples that have undergone extensive weathering. In the study, a variety of oils and refined hydrocarbon products, weathered both artificially and naturally, were analyzed by GC, GC/MS and GC/IRMS. In cases where samples have lost their more volatile n-alkanes as a result of weathering, the isotopic compositions of the individual compounds were not found to be extensively affected. Hence GC/IRMS can be particularly useful for correlation of refined products dominated by n-alkanes in the C10-C20 region and containing none of the biomarkers more commonly used for correlation purposes. For extensively weathered crude oils that may have lost all of their n-alkanes, it has been demonstrated that isolation and pyrolysis of the asphaltenes followed by GC/IRMS of the individual pyrolysis products can be used for correlation purposes with their unaltered counterparts. In summary, it is not proposed that GC/IRMS be used as a stand-alone correlation tool but in conjunction with existing techniques such as GC and GC/ MS. It is proposed however that it could be particularly useful in situations where existing techniques do not provide unambiguous correlations. In certain cases, particularly with lighter refined products, GC/IRMS can provide valuable correlation data in the absence of biomarker data. Finally, characterization of asphaltene pyrolysis products by GC/IRMS provides an alternative and useful method for correlating extensively degraded samples with their nondegraded sources.
Introduction There have been many studies concerned with hydrocarbon pollutants in aquatic environments (rivers, lakes, groundwaters, coastal waters, etc.) in recent years (1-5). Sources of these contaminants include natural seepages of crude oils, ship traffic, supertankers spilling crude oils, and leaking storage tanks or pipelines. Thus, identifying, quantifying,
S0013-936X(97)00068-0 CCC: $14.00
1997 American Chemical Society
and monitoring the fate of these pollutants spilled in the environment are of primary importance in providing a better response to hydrocarbon spills. The most common approach to the characterization of a spilled oil and identification of its potential source relies on analyses by gas chromatography (GC) and gas chromatography/mass spectrometry (GC/MS). Correlations are made on the basis of the molecular distribution of aliphatic and aromatic hydrocarbons, or more specifically biomarker fingerprints (6). In certain situations, GC and GC/MS data can be ambiguous and inconclusive, since an oil is quickly affected by weathering processes such as evaporation, photooxidation, water-washing, or biodegradation (7-10). Evaporation occurs in the first few hours after a spill and removes the more volatile hydrocarbons. Water-washing also occurs rapidly and removes the more water-soluble hydrocarbons, typically hydrocarbons below C15, and some of the C15+ aromatic compounds that are more water-soluble than paraffins (9-11). At the same time, biodegradation will also start to affect the nature of the spilled oil by initially removing the n-alkanes before attacking the more complex branched and cyclic hydrocarbons and naphthenic compounds. Lower carbon-numbered alkanes, remaining after the effects of evaporation and water-washing, are initially removed followed by those with increasing numbers of carbon atoms. Severe biodegradation may also be marked by loss of some of the biomarkers (C27-C29 steranes) and demethylation of others (C27-C35 hopanes) (10) and changes in the distribution of naphthenic compounds. The combined effects of weathering can strongly modify the fingerprints and parameters used to correlate an oil with its source on the basis of GC and GC/MS analysis. In the case of gasolines and other refined products, evaporation and water-washing make it extremely difficult to undertake correlations using conventional techniques. Correlation of a spilled oil to its suspected source requires other discriminative parameters that are relatively insensitive to weathering processes such as bulk carbon isotopic compositions (1, 2, 12, 13). Although it has been demonstrated that in some cases even the bulk carbon isotopic ratios of spilled oil fractions can be affected by weathering processes as demonstrated by Stahl (14), Macko et al. (1), Palmer (15), and Sofer (16). Most of the changes observed resulted from water-washing, which preferentially removes the mono- and diaromatics that are more enriched in 13C thus leaving an aromatic fraction depleted in 13C (16), and biodegradation, which removes the n-alkanes inducing a relative enrichment in iso- and cycloalkanes that are enriched in 13C. As a result of the development of GC/IRMS, the isotopic composition of individual compounds in complex mixtures can now be determined (17-19), and several studies have already appeared that describe the use of this technique to elucidate the origin of various hydrocarbon products found in the aquatic environment (20, 21). The major aim of this paper is to assess the utilization of GC/IRMS as a correlation tool for hydrocarbons spilled in aquatic environments with their suspected source(s). The study focuses on two main aspects of the problem: (i) determination of the effects of artificial as well as natural weathering processes on the isotopic composition of individual n-alkanes and whether these effects could compromise the correlation of a weathered oil to its non-degraded counterpart; (ii) the evaluation of possibilities and limitations of GC/IRMS as a correlation tool in environmental problems. A method for the correlation of severely biodegraded oils based on the GC/IRMS characterization of asphaltene pyrolysates is also discussed. It should be emphasized that it is not the intention of this paper to suggest that GC/IRMS will replace well-established correlation tech-
VOL. 31, NO. 12, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3417
FIGURE 1. Isotopic composition of the n-alkanes of five oils from different origins. niques such as GC and GC/MS. Rather it is proposed that it be used in conjunction with these existing tools to provide additional evidence particularly in cases where the fingerprints used for the more conventional analyses may have been altered to such an extent that unambiguous correlations may not be possible.
Experimental Section Samples. A number of different samples from various origins were analyzed in this study. A crude oil from Oklahoma was used for the artificial weathering experiments and a variety of samples were obtained from the U.S. Coast Guard including crude oils and light fuel oils (refined products). These samples were naturally weathered to different extents from slightly to moderately weathered and were supplied along with their unweathered counterparts. In addition, five severely weathered crude oils in the form of beach-stranded tar balls, collected from geographically similar locations, were examined as part of this study. Weathering Experiments. The three main weathering processes (evaporation, water-washing, and biodegradation) were artificially reproduced in the laboratory using the same Oklahoma crude oil. One aliquot of the crude oil was allowed to evaporate for 4 yr at room temperature. For the waterwashing experiment, an aliquot of the oil (100 mL) was combined with distilled water (800 mL) and stirred in a beaker for 2 months at room temperature. Distilled water was preferred to saline water, and room temperature was preferred to colder water, since it represents the optimal conditions for efficient water-washing (9). Finally for the biodegradation experiments, an active sewage sludge was added to the crude oil in a sand and water environment. Aliquots of the biodegraded oil were taken at regular intervals, over a 4-month period, so that the main stages of biodegradation could be studied. Sample Preparation and Analysis. Most of the samples, except the refined products such as light fuel oils, were characterized using the same analytical protocol: (i) precipitation of the asphaltenes in pentane; (ii) fractionation of the maltene fraction into aliphatics, aromatics, and NSOs compounds by HPLC; (iii) when necessary, isolation of the n-alkanes by urea adduction of the aliphatic fraction and recovery of the urea-adducted fraction (n-alkanes) by dissolution of the urea in distilled water and extraction of the n-alkanes using pentane. In order to prevent the loss of volatiles, no fractionation was performed on the light fuel oils. Gas Chromatography/Isotope Ratio Mass Spectrometry. GC/IRMS analyses were performed using a Varian 3400 gas
3418
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 12, 1997
chromatograph coupled with a Finnigan MAT 252 isotope ratio mass spectrometer via a combustion furnace heated at 1050 °C and a water trap. The samples were injected onto a DB-1 fused silica capillary column (60 m; 0.32 mm i.d.; 0.25 mm film thickness) and chromatographed with the following temperature program: 40 °C for 10 min to 310 °C at 2 °C/min with a hold of 20 min. Bulk isotopic compositions were determined as described by Engel and Maynard (22). Reproducibility and Accuracy. The accuracy and the reproducibility of GC/IRMS data are mainly affected by chromatographic resolution (coeluting compounds) and background as defined by column bleed and the unresolved complex mixture (UCM) from the sample. The accuracy of the data is initially monitored with a set of standards of known isotopic composition. Internal standards (fully deuterated n-alkanes C9, C10, C16, C19, C24, C32, and C36) are added to the samples being analyzed to provide a second control of data quality (when the standard data are included in the figures, these are indicated as C10D, etc.). Each sample is analyzed at least two times, and standard deviations (1 σ) of the replicates are calculated for each n-alkane to estimate reproducibility. Analyses of the aliphatic fractions give an average reproducibility from 0.15 to 0.32‰. Reproducibility decreases with the analysis of more complex mixtures such as whole oils where the presence of aromatics increases the chance of coelution or of aliphatic fractions of biodegraded oils where the presence of the UCM can affect the isotopic composition of the n-alkanes. To improve the reproducibility and the accuracy of the data, the n-alkanes can be isolated by urea adduction when required (i.e., moderately biodegraded oils). The accuracy of the measurements is defined by comparing the δ13C values of the deuterated compounds in the analyzed samples to the δ13C of the same compounds as determined by IRMS analysis. The isotopic values obtained from the GC/IRMS analysis are systematically heavier than the values obtained by simple IRMS analysis from 0.09 to 0.81‰ mainly because of coelutions and high background. However, as long as the reproducibility is good, this does not affect the quality of the results in this type of application where the data are being used predominantly for correlation purposes. These results are consistent with the accuracy and the reproducibility obtained in previous studies of similar samples (23, 24).
Results and Discussion The discriminative nature of the isotopic compositions of individual n-alkanes in crude oils is illustrated in Figure 1 and Table 1. Clearly these oils could have been differentiated on the basis of existing methods, but the point is to illustrate
FIGURE 2. Chromatograms of the aliphatic fraction of artificially weathered oils and the unweathered counterpart.
TABLE 1. Standard Deviations Calculated for Each Combination of Five Oils of Different Origins Defined in Figure 1
U8-106-1 Paris Basin Middle East Oklahoma Mahakam
U8-106-1
Paris Basin
Middle East
Oklahoma
Mahakam
0 3.85 1.78 3.06 3.41
0 2.04 0.59 0.71
0 1.34 1.34
0 0.31
0
that in many cases there are significant differences in the isotopic compositions of individual alkanes of oils from different locations. The averaged standard deviations calculated between three oils (U8-106-1 of unknown origin, a Paris Basin oil, and a Middle East oil) are higher than 0.55‰, which is out of the range of the analytical error. The oils from Oklahoma and the Mahakam Delta show similar isotopic compositions with averaged standard deviation in the range of the analytical error (0.15 to 0.32‰). The close values are observed in a narrow range of n-alkanes (C18-C30), and the
isotopic compositions of the light and heavy ends are quite different. The second step of the study was to determine if weathering processes can compromise correlations between pollutants and suspected sources. The artificially weathered samples were analyzed by GC, and the results are compared to those of the initial oil. The chromatogram of the initial oil was largely dominated by low molecular weight n-alkanes maximizing around n-C12 (Figure 2a). After a 4-yr period, the evaporation process was marked by the depletion of n-alkanes with carbon numbers lower than 14 (Figure 2b), and the waterwashed oil (after 38 days) was depleted in n-alkanes with carbon numbers
