Oil Spill Environmental Forensics: Fingerprinting and Source Identification

中国环境学会  2011年 06月21日

  Zhendi Wang (Principal Research Scientist)
  Emergencies Science and Technology, Science and Technology Branch
  Environment Canada, Government of Canada
  3439 River Road, Ottawa, Ontario, Canada, K1A 0H3
  Tel: (613) 990-1597, Fax: (613) 991-9485
  E-mail: zhendi.wang@ec.gc.ca
  Oil, refined product, and pyrogenic hydrocarbons are the most frequently discovered contaminants in the environment. To effectively determine the fate of spilled oil in the environment and to successfully identify source(s) of spilled oil and petroleum products is, therefore, extremely important in many oil-related environmental studies and liability cases. This article briefly reviews the recent development of chemical analysis methodologies which are most frequently used in oil spill characterization and identification studies and environmental forensic investigations. The fingerprinting and data interpretation techniques discussed include oil spill identification protocol, tiered analytical approach, generic features and chemical composition of oils, effects of weathering on hydrocarbon fingerprinting, recognition of distribution patterns of petroleum hydrocarbons, oil type screening and differentiation, analysis of Asource-specific marker@ compounds, and determination of diagnostic ratios of specific oil constituents.   
  Keywords:    Environmental forensics, Oil spill, Oil fingerprinting, PAHs, Biomarkers, Diagnostic ratios, Weathering

  Petroleum plays an extremely important role in modern society. As the population of the world increases and developing countries become more industrialized, the demand for energy keeps growing worldwide. Consumption world-wide was about 30 billion barrels in 2004. Figure 1 presents the worldwide petroleum demand and supply from 1970 to 2004 (DOE, 2004). Currently, oil is the dominant energy source and it is expected to remain so over the next several decades (NRC, 2002). In addition of the natural seeps (they are purely natural phenomena that occur when crude oil seeps from the geologic strata beneath the seafloor to the sea water column), the worldwide widespread extraction, transportation, and consumption of petroleum inevitably results in its release to the environment. Waterborne oil spills of unknown origin (from continuous leak to illegal dumping) often occur in rivers, open waters and in coastal waterways. Also, petroleum and its combustion-derived hydrocarbons are often one of the most frequently discovered chemicals of concern at contaminated sites on land. Based on analysis of data from a wide variety of sources, about 260,000 metric tonnes of petroleum, each year on average, are released to the waters off North America. Annual worldwide estimates of petroleum input to the sea exceed 1,300,000 metric tonnes (NRC, 2002). Most oils spilled into the sea are fuels (48%) and then crude oils (29%). Fuels consists primarily of Bunker oils and intermediate fuel oils (IFO) which consists of Bunker oils mixed with lighter fuels such as diesel. A list of the major oil spills in the last 40 years (Table 1) has been provided (Fingas, 2001). The spills are listed according to their volume, beginning with the largest spill to date – the release of oil during the Gulf War in 1991 (800,000 tons). According to the spill volume, the most influential Exxon Valdez spill ranks at No. 52 (37,000 tons), while the most recent two large-scale marine spills, 1999 Erika spill (occurred about 110 km south of Brest, France) and 2002 Prestige spill (occurred on water about 240 km off the Northwest coast of Spain), rank only at No. 124 (12,000 tons) and No. 90 (24,000 tons), respectively. Although most of the large oil spills are from tankers, these spills only make up about 5% of all oil pollution entering the sea. Most oil pollution in the oceans comes from the run-off of oil and fuel from land-based sources rather than from accidental spills. In Canada, about 12 spills of more than 4000 L are reported each day, of which about one spill is into navigable waters. Most spills take place on land, including oil spills from pipelines, underground storage tanks, and aboveground storage containers.
  Oil poses a range of environmental risks and causes wide public concerns when released into the environment, whether as catastrophic spills or chronic discharges. Oil spills have led to legal battles resulting in billion dollars in damage awards and punitive fines. Therefore, to precisely characterize spilled oil hydrocarbons in complex environmental samples and to defensibly identify the source(s) of hydrocarbons entering the environment is extremely important for site contamination assessment, for prediction of the potential long-term impact of spilled oils on the ecosystem, and for determining responsibility for the spill. In addition, successful forensic investigation and analysis of hydrocarbons in contaminated sites and receptors yield a wealth of chemical fingerprinting data. These data, in combination with historic, geological environmental and any other related information can, in many cases, help to settle legal liability and to support litigation against the spillers.

  2.1  Source-specific Target Hydrocarbons
  Crude oil and many other petroleum-related hydrocarbons such as combustion-derived mixture are often the most discovered compounds of concern at many contaminated sites. Oils consist of complex mixtures of hydrocarbons and non-hydrocarbons that range from small, volatile compounds to large, non-volatile ones. In the last two decades, a wide variety of instrumental techniques, particularly gas chromatography equipped with flame ionization detector (GC-FID) and gas chromatography-mass spectrometer (GC-MS), have been extensively used for analysis of various organic compounds in environmental samples. Regulatory bodies such as the US Environmental Protection Agency (EPA), ASTM (American Society for Testing and Materials), and Canadian Council of Ministries of the Environment (CCME) have developed and codified a series of methods based on the GC techniques. The EPA methods include the EPA 418.1 (Total recoverable petroleum hydrocarbons by infra-red spectroscopy), 1664 (n-Hexane extractable material and silica-gel treated n-hexane extractable material by extraction and gravimetry), 600 series (Method standards for wastewater), and 8000 series methods (such as Method 89015B, Non-halogenated organics using GC-FID: Method 8270, Semi-volatile organic compounds by GC-MS). The ASTM methods include ASTM methods 3328, 5037 and 5739. These methods have been widely used as routine procedures and excellent base techniques for the chemical fingerprinting spilled oil and suspected source oils in the environmental media. However, there is fundamental barrier for environmental forensic scientists and investigators: these methods were originally designed for measuring industrial chemical in wastewater and solid waste, and none of these methods provides information on detailed chemical components which comprise the complex spilled oil or petroleum-derived samples. The data generated from these methods are generally insufficient to answer the fundamental forensic questions (such as what type and source, weathering status of spilled oil, potential spillers, and so on) raised in an oil spill liability investigation. For example, the polycyclic aromatic hydrocarbon (PAH) compounds in oils are dominated almost exclusively by the C1 to C4 alkylated homologues of the parent PAHs, in particular, naphthalene, phenanthrene, dibenzothiophene, fluorene and chrysene, none of which are measured by these standard EPA methods. Other important classes of petroleum hydrocarbons (e.g., biomarkers and n-alkanes) are not measured by these methods at all. In recent several years, many EPA and ASTM methods have been further modified to allow flexibility in the deployment of the “standard” analytical methods and to improve specificity and sensitivity for measuring spilled oil and petroleum products in soils and waters by environmental chemists. For example, the EPA Method 8270 has been modified to increase analytical sensitivity and to expand the analyte list to include petroleum-specific compounds such as the alkylated PAHs, sulfur and nitrogen-containing PAHs, and biomarker triterpane and sterane compounds. The principal modification to EPA Method 8270 is the use of the high resolution GC-MS selected-ion-mode (SIM) analysis that offers increased sensitivity relative to the full-scan mode. Many environmental laboratories have used the modified EPA Method 8270, combined with column-cleanup and rigorous QA measures, to identify and quantify low levels of hydrocarbons.
  Hydrocarbon-contaminated site investigation and oil spill identification requires further elaboration of oil target analytes to include determination of the individual specific target compounds and isomeric groups. The selection of appropriate target oil analytes is dependent mainly on the type of oil spilled, the particular environmental compartments being assessed, and on expected needs for current and future data comparison. In general, the major petroleum-specific target analytes that may be needed to be chemically characterized for oil source identification and environmental assessment include the following:
   (1) Individual saturated hydrocarbons including n-alkanes (n-C8 through n-C44) and selected isoprenoids pristane (2,6,10,14-tetramethyl-pentadecane) and phytane (2,6,10,14-tetramethyl-hexadecane). In some cases, another three highly-abundant isoprenoid compounds: farnesane (2,6,10-trimethyl-C12), 2,6,10-trimethyl-C13, and norpristane (2,6,10-trimethyl-C15) are also included;
  (2) Alkyl (C1 - C14) cyclo-hexane homologous compound series. These homologous compounds exhibit a characteristic distribution patter in m/z 83 mass chromatograms for different types of fuels, providing another useful fingerprint for characterizing petroleum derivatives;
  (3) The volatile hydrocarbons including BTEX (benzene, toluene, ethylbenzene, and 3 xylene isomers) and alkylated (C3- to C5-) benzenes, naphthenes, and volatile paraffins and isoparaffins. Analysis of long-side-chain n-alkylbenzens with the n-alkyl groups in the C7 – C27 range for evaluation of fate of crude oil in the environment has been reported recently;
  (4) The EPA priority parent PAHs and, in particular, the petroleum-specific alkylated (C1 to C4) homologues of selected PAHs (that is, alkylated naphthalene, phenanthrene, dibenzothiophene, fluorene, and chrysene series). These alkylated PAH homologues are the backbone of chemical characterization and identification of oil spill assessments (Table 2)
  (5) Biomarker terpane and sterane compounds (Table 3). Analysis of selected ion peaks produced by these characteristic, environmentally-persistent compounds generates information of great importance in determining source(s), weathered state and potential treatability;
  (6) Measurements of bulk hydrocarbon groups including total petroleum hydrocarbons (TPH), the unresolved complex mixtures (UCM), and the total saturates and total aromatics, contents of asphaltenes and resins,
  (7) Additives to petroleum products. They include alkyl lead additives (tetramethyl lead and trimethylethyl lead at m/z 253 and 223, dimethyldiethyl lead at m/z 267 and 223, methyltriethyl lead at m/z 281 and 223, tetraethyl lead at 295 and 237); oxygenates including substances such as ethanol, methanol, methyl tertiary butyl ether (MTBE), ethyl tertiary butyl ether (ETBE), and tertiary amyl methyl ether (TAME); fuel dyes used for differentiation among fuel grades; and anti-oxidant compounds or called inhibitors (such as aromatic amines and alkyl-substituted phenols) added to fuels to retard autooxidation; 
  (8) Measurement of stable carbon isotope ratio (δ13C) of hydrocarbon groups, sometimes measurement of the isotopic composition of individual compounds by GC-IRMS for correlating spills with suspected sources is also included in many oil spill studies.
  Another potentially-valuable hydrocarbon group for oil spill identification is nitrogen and oxygen heterocyclic hydrocarbons. These heterocyclic hydrocarbons are generally only present in oils at quite relatively low concentrations compared to PAHs. However, they become enhanced with weathering because they are biorefractory and persistent in the environment.  Most organic nitrogen hydrocarbons in crude oils are present as alkylated aromatic heterocycles with a predominance of neutral pyrrolic structures over basic pyridine forms. They are chiefly associated with high boiling fractions, much of the nitrogen in petroleum being in asphaltenes. These compounds may provide important clues for potential sources of hydrocarbons in the environment and for tracing petroleum molecules back to their biological precursors. Compared to the PAHs and biomarkers, the application of nitrogen and oxygen-containing heterocyclic hydrocarbons in source identification is still in its infancy, and more research is clearly needed.
  2.2  Tiered Analytical Approach
  The characterization and identification of spilled oil and petroleum products can best be conducted using a tiered analytical approach (Uhler et al., 1998-1999, Wang et al., 1999a; Stout et al., 2002), by which sufficient details concerning the nature and origin, chemical composition changes, and weathering degrees of spilled oil under investigation can be gathered. Depending on needs of the specific spill site investigation, the tiered analytical approaches may vary. It gives the environmental forensic investigator the flexibility to determine how many tiers should be used and how much information is sufficient and necessary to address site- or incident-specific questions. The tiered approach used by the Environment Canada Oil Spill Research Program includes the following:  
  Tier 1 - Determination of hydrocarbon groups and product type screening via GC-FID.
  Tier 2 - Determination of volatile hydrocarbons (e.g., BTEX and alkyl benzenes, low molecular weight alkyl-pentanes and alkyl-hexanes, smaller cyclo-pentanes and cyclo-hexanes, and various additives in lighter petroleum products) via GC-MS.
  Tier 3 – Determination of target PAHs and biomarker components via GC-MS; and determination and comparison of diagnostic ratios of these “source-specific marker” compounds with the spill and suspected source oil samples and with the corresponding data from database.
  Tier 4 - Determination of weathered percentages of the residual oil and estimation of spill ages.
  Tier 5 - If necessary (in some instances), application of statistical tools (such as PCA and multivariate methods) to analyze large amount of fingerprinting data and to answer the questions of whether the spill oil sample is “positive match”, “negative match” or “probable match” or “inconclusive” to the candidate source samples.   
  In this tiered analytical approach, the high-resolution capillary GC-FID analysis is applied to determine hydrocarbon groups (e.g., TPH, UCM, the total saturates and total aromatics) and concentrations of total n-alkanes and major isoprenoid compounds (e.g., pristane and phytane) from n-C8 to n-C44, and to characterize the product types (e.g., crude oil, diesel, lube oil, or Bunker C type fuel) in fresh to highly weathered oil samples. If needed, the thin layer chromatographic (TLC) or gravimetric methods are applied to determine the contents of asphaltenes and resins. The GC-MS analyses provide data on the “source-specific” marker compounds including the target alkylated PAH homologues and other EPA priority PAHs, and biomarker terpane and sterane compounds. The MS detector is operated in the scan mode to obtain spectral data for identification of unknown components and in the selected ion mode (SIM) for quantitation of target compounds. An appropriate temperature program is selected to achieve near-baseline separation of all of the target components. Quantitation of the alkylated PAH homologues, other EPA priority PAHs, and biomarkers is performed using the internal standard method with the RRFs for each compound determined during the instrument calibration. For analysis of BTEX and other alkyl benzenes, all oil samples are directly weighed and dissolved in n-pentane to an approximate concentration of 2 mg/mL. Prior to analysis, the tightly capped oil solutions are put in a refrigerator for 30 min to precipitate the asphaltenes to the bottom of the vials in order to avoid performance deterioration of the column (Wang et al., 1995a).
  The quality and reliability of analytical measurements is dependent on the quality assurance (QA) and quality control (QC) program employed. In order to support oil spill forensic investigations, quality management (including laboratory profile and mission, quality assurance and quality control system, updated standard operational procedures, personnel training program and record, up-to-date methodology, equipment management, sample management, data management, and workload management) must be strictly followed. The chemical measurements must be conducted within the framework of highly stringent, defensible and reliable QC and QA programs (Page et al., 1995; Douglas et al., 1996 and 2004; Wang et al., 1999a and 2003; Daling et al., 2002; Wang and Stout, 2007; Wang and Fingas, 2006; Stout et al., 2002; Faksness et al., 2002b; EPA, 1997; ASTM, 2000; ETC, 2003). The QA/QC programs used by different laboratories may differ more or less in the course of sample handling and preparation, analysis, and reporting of analytical data, but the quality principles and practices are similar.
  Quality assurance (QA) is a definite plan for laboratory operation that specifies standard procedures that help to produce data with defensible quality and reported results with a high level of confidence. The basic requirements of a quality assurance program is to recognize possible errors, understand the measurement system used, and develop techniques and plans to minimize errors. The elements of quality assurance are quality control and quality assessment. Quality control (QC) includes: good laboratory practices; updated standard operational procedures; sample collection, documentation, calibration; standardization; instrument maintenance; facilities maintenance; education and training; reporting of forensic analysis data, continuous improvement program, and inspection and validation. Quality Assessment includes: reference materials; replicates; splits; spikes; surrogates; collaborative tests; and statistical analysis.
  Prior to sample analysis, a five-point response calibration curve should be established to demonstrate the linear-range of the analysis. Check standards at the mid-point of the established calibration curves are analyzed before and after each analytical batch of samples (7-10 samples) to validate the integrity of the initial calibration. The relative response factor (RRF) stability is a key factor in maintaining the quality of the analysis. A control chart for RRF values should be prepared and monitored. All samples and quality control samples (procedural blank, matrix spike samples, duplicate, and reference oil sample) are spiked with appropriate surrogates to measure individual sample matrix effects associated with sample preparation and analysis. PAH surrogate and matrix spike recoveries should be within 60% to 120%. Method detection limits (MDLs) studies of target compounds are performed according to the procedure described in the EPA protocol titled “Definition and Procedure for the Determination of the Method Detection Limit” (Code of Federal Regulations 40CFR Part 136). Analysis and characterization of forensic sample batches should be performed on the same instrument within the same analytical sequence by experienced chemists.
  2.3  Emerging Techniques for Fingerprinting Hydrocarbons
  In recent years, a number of emerging instrumental techniques such as GC-IRMS for isotopic composition of individual components in oil and petroleum products; ultrahigh-resolution Fourier transform ion cyclotron resonance spectrometry (FT-ICR MS), field desorption/ionization FT-ICR MS, and atmospheric pressure photoionization pressure FT-ICR MS (Marshal et al., 2004); capillary GC with ICP-collision cell-MS detection (Bouyssiere et al., 2004); GC-field ionization time-of-flight high resolution MS for petroleum characterization (Qian and Dechert, 2002); and two dimensional GC (GC x GC) (Dimandja, 2004, Dallüge et al., 2003), have been applied for fingerprinting complex oil hydrocarbons, investigation of the low-concentration sulfur speciation in petroleum, and possibly ultimate characterization of all of the chemical constituents of petroleum. Among these techniques, the GC x GC is most studied and reported.
  In the GCxGC technique, two capillary GC columns are connected serially by a thermal modulator - the interface between the two separation dimensions. The thermal modulators can be further categorized into three types: heating, cryogenic, and jet-pulsed systems. Modulators periodically trap and then release smaller portions of a continuous stream of effluents. The first dimension sample effluent is thus continuously transferred in small portion to the second dimension column throughout the chromatographic run, and each transferred pulse generates a high-speed secondary gas chromatogram. Most often, the first dimension separation uses a nonpolar phase to separate analytes by volatility difference, and the second dimension uses a more polar phase to separate first dimension coeluters by polarity difference. The resulting GC x GC chromatogram can be viewed in several formats, including surface, contour, and peak apex plots. The key features of GC x GC are its greater resolution power and its ability to separate components into classes in samples. GC x GC chromatograms can have much higher peak capacities (exceeding 20,000) than conventional gas chromatograms (capacities rarely exceed 1000). Another feature of GC x GC is that its separations can usually be done in a time comparable to the conventional GC because the significantly higher speed of the 2nd dimensional GC allows increased peak capacity without increasing the length of the analysis. The separation of sample components into classes through structured chromatograms provides an additional means of identification and reduces the probability of peak overlap between numbers of different chemical classes (Dallüge et al., 2003).
  2.3  Factors Controlling the Chemical Fingerprints of Spilled Oil
   Figure 2 depicts the four factors (Stout et al., 2007) that influence of petroleum in the environment that are relevant to oil spill investigations: (1) primary control factor: crude oil genesis; (2) secondary control factor: petroleum refining; (3) tertiary control factor: weathering; and (4) fourth control factor: mixing with “background” hydrocarbons. These factors are discussed and described in details in the following sections.
  3.    chemical Composition of Oil/Petroleum Products AND SPILL IDENTIFICATION
  Generally, the chemical composition of fresh to mildly weathered oils and petroleum products can be readily revealed from their GC-FID traces, especially if the background hydrocarbon levels are low in the impacted environment. In addition for measurements of TPH and other hydrocarbon groups in samples (such as total saturates, total aromatics, GC-FID chromatograms provide a distribution pattern of petroleum hydrocarbons (e.g., boiling or carbon number range, and profile of UCM), fingerprints of the major oil components (e.g., individual resolved n-alkanes and major isoprenoids), and information on the weathering extent of the spilled oil. Comparing biodegradation indicators (such as n-C17/pristane and n-C18/phytane) for the spilled oil with the source oil can be also used to monitor the effect of microbial degradation on the loss of hydrocarbons at the spill site. The GC-FID approach can be used to quickly screen the oil and refined product type (see Figures 3 and 4). It is noted, however, that GC analyses alone may give limited oil diagnostic characteristics when the petroleum samples have been highly weathered. For defensible source identification, GC-MS analysis must be performed.  
  3.1  Chemical Composition Features of Crude Oil (1st Control Factor)
  Crude oil is an extremely complex mixture of hundreds to thousands of individual hydrocarbons and non-hydrocarbons. These compounds range from small, simple, volatile, and distinct compounds (e.g., methane) to extremely large, complex, non-volatile, colloidally-dispersed macromolecules (e.g., asphaltenes).  The distribution of these compounds imparts certain physical properties on the oil. Depending on the sources of carbon from which the oils are generated and the geologic environment in which they migrated and from which reservoir (such as Middle East or North Sea oil), crude oil compositions vary widely and can have: (1) dramatically varied compositions in the C5 to C44 carbon range such as relative amounts of paraffinic, aromatic and asphaltenic compounds; (2) large differences in the n-alkanes, isoprenoids, and cyclic-alkanes (such as alkyl cyclo-hexanes) concentrations and distribution patterns and UCM profiles; (3) significantly different relative ratios of isoprenoids to normal alkanes; and (4) large differences in distribution patterns and concentrations of oil-characteristic long-side-chain n-alkyl benzenes (the carbon number in the alkyl side chain can be up to C27 for some oils), alkylated PAH homologues (many of 4-6 ring unsubstituted PAHs are only minor components in oils), and biomarkers.
  The most prominent aliphatics in most crude oil are the normal (straight chain) alkanes. In general, most crude oils exhibit an n-alkane distribution profile (GC-FID and GC/MS at m/z 85) of decreasing abundances with increasing carbon number. The maximum n-alkanes within the profile are variable from oil to oil. The smoothness of n-alkane distribution profile in crude oil can be diagnostic. The carbon preference index (CPI) values of most oils are around ~1. Oils with CPI values greater than 1 are often derived from source rock strata that contained relatively abundant land plant organic components including leaf waxes. CPI is defined as the total of n-alkanes with odd carbon number divided by the total of n-alkanes with even carbon number in the carbon range of C8 to C44. 
  CPI = (the sum of odd n-alkanes) / (the sum of even n-alkanes)
  or in the simplified formula:
  CPI = (C23 + C25 + C27 + C29 + C31 + C33) / (C24 + C26 + C28 + C30 + C32 + C34)
  The distributions of isoprenoids (m/z 113) and alkyl (C0- to C15-) cyclo-hexane homologous series (m/z 83) are also apparent in many crude oils. Biodegradation affects the straight-chain alkanes more than branched alkanes (isoprenoids). Therefore, determination and comparison of biodegradation indicators of n-C17/pristane and n-C18/phytane between the spilled oil and the source oil are often performed at this level to monitor the probable effects of microbial degradation at the spill site. 
  Figure 3 shows GC-FID chromatograms (by high temperature program) for six different oils from different world’s main production area. These six oils are different, as not only are there large differences in the n-alkane distributions and UCM profiles, but also differences in hydrocarbon group composition and in relative ratios of isoprenoids to normal alkanes. Note that the Orinoco oil (a Bitumen oil from Venezuela) has nearly no n-alkanes in its GC-FID chromatogram.
  3.2  General Chemical Composition Features of Refined Products (2nd Control Factor)
        Refined petroleum products are obtained from crude oil through a variety of refining processes (Olah and Molnar, 1995, Speight, 2002) such as distillation, cracking, catalytic reforming, isomerization, alkylation, and blending. Depending on the chemical composition of their “parent” crude oil feedstocks, varying refining approach and conditions, wide range of applications, regulatory requirements, and economic requirements, refined products can have wide variety in chemical compositions. As example, Figures 4 and 5 show GC-FID chromatograms of eight petroleum products and their quantitative n-alkane distributions respectively, illustrating differences of these products in the GC profiles, carbon range, and UCM patterns.
  3.2.1      Light distillates
  Light distillates are typically products in the C3 to C13 carbon range. They include aviation gas (gasoline-type jet fuel which has a wider boiling range than kerosene-type jet fuel and includes some gasoline fractions), naphtha (a liquid petroleum product that boils from about 30 0C to approximately 200 0C), and automotive gasoline. The GC traces of fresh light distillates are featured with dominance of light-end, resolved hydrocarbons and a minimal UCM. 
  Gasoline is the generic term used to describe volatile, inflammable petroleum fuels used primarily for internal combustion engines. It is a complex mixture of hundreds of different hydrocarbons predominately in C4 to C13 range, with the nominal boiling point range of 40 to 180 0C or, at most, below 200 0C. The composition of gasoline is best expressed in five major hydrocarbon classes: paraffins, isoparaffins (branched alkanes), naphthenes (cyclo-alkanes), aromatics, and olefins (PIANO). The bulk PIANO composition provides a useful cumulative parameter for fuel type (such as gasoline, aviation gasoline, or Jet fuel) differentiation. Gasoline contains considerable BTEX and alkylated benzene compounds.
  The properties of gasoline are quite diverse, and the principal properties affecting the performance of gasoline are volatility and combustion characteristics. In order to improve some specific properties such as the engine efficiency and antiknock properties, certain chemical compounds, additives, are often added to gasolines. They may include octane-boosting additives (such as methyl tertiary butyl ether, MTBE), oxidation inhibitors (such as aromatic amines and hindered phenols), corrosion inhibitors (such as carboxylic acids and carboxylates), anti-icing additives (such alcohols, glycols, and surfactants), anti-knocking lead alkyls, and dyes (oil-soluble solid and liquid dyes: red, alkyl derivatives of azobenzene-4-azo-2-naphthol; orange, benzene-azo-naphthol; yellow, para-diethylaminoazobenzene, and blue, 1,4-diisopropyl-aminoanthraquinone) for identification of different gasolines. Since the 1970s, the lead level in refined products in Canada and the US has decreased substantially. Use of leaded gasoline in cars was completely banned in Canada and the US in 1990 and 1996, respectively.
  3.2.2.     Mid-range distillates
  Mid-range distillates are typically products in a relatively broader carbon range (C6 to C26) and include kerosene, aviation jet (turbine) fuels, and diesel products. Jet fuel is kerosene-based aviation fuel. Jet fuel is used for aviation turbine power units and usually has the same distillation characteristics and flash point as kerosene. Jet fuels are manufactured predominately from strait-run kerosene or kerosene-naphtha blends. Jet fuels are similar in gross composition, with many of the differences in them attributable to additives designed to control some fuel parameters such as freeze and pour point characteristics. As Figure 4 shows, the chromatogram of a commercial Jet A fuel is dominated by GC-resolved n-alkanes in a narrow range of n-C7 to n-C18 with maximum being around n-C11. The UCM is well defined.
  Diesel fuels originally were straight-run products obtained from the distillation of crude oil. Currently, diesel fuel may also contain varying amounts of selected cracked distillates to increase the available volume. The boiling range of diesel fuel is approximately 125-380 0C. One of the most widely used specifications (ASTM D-975) covers three grades of diesel fuel oils: diesel fuel #1, diesel fuel #2, and diesel fuel #4. Grades #1 and #2 diesels are distillate fuels, they are most commonly used in high-speed engines of the mobile type, in medium speed stationary engines, and in railroad engines. Grade #4 diesel covers the class of more viscous distillates and, at times, blends of these distillates with residual fuel oils. The marine fuel specifications have four categories of distillate fuels and fifteen categories of fuels containing residual components (ASTM D-2069 Method). Diesel consists of hydrocarbons in a carbon range of C8 to C28 and has significantly high concentrations of n-alkanes, alkyl-cyclohexane, and PAHs. The properties of a given diesel are largely a function of the crude oil feedstock. The GC chromatogram of diesel fuel #2 is generally dominated by a nearly normal distribution of n-alkanes with maxima being around n-C11 to n-C14. Also, a central UCM “hump” is obvious. Once released to the water surface, mid-range fuels spread very rapidly. Very large and thin films will often form, leading to quite rapid weathering of spilled fuel.
  3.2.3      Classic heavy residual fuels
  Heavy fuel oils (HFO) are blended products manufactured from residues of various refinery distillation and cracking processes. The heavy fuel oils are largely used in marine applications and industrial power generation. Classic heavy fuel types include fuel No. 5 and No. 6 (also known as Bunker C). For many years the term “Bunker C fuel oil” has been widely used to designate the most viscous residual fuels for general land and marine use. Different grade of heavy fuel oils are expressed by the numbers of their kinetic viscosity in centistokes (cSt) at 50 0C. The main grades are IFO 30, IFO 180, and IFO380. The chemical composition of Bunker C (or IFO 380) can vary widely and remarkably, depending on production oilfields, production years, and processes it has undergone. Currently, many Bunker type fuels are produced by blending residual oils with diesel fuels or other lighter fuels in various ratios to produce residual fuel oil of acceptable viscosity for marine or power plant use. The use of heavy fuel oils as bunker oil on ships has been found to be the main course of chronic oil pollution because of illegal discharge of residues and residual oil into the sea.
  For comparison, the chromatograms of an IFO 180, a lighter residual fuel No. 5 (also called Bunker B) and a Heavy Fuel Oil 6303 (or called Bunker C or Land Bunker, from Imperial Oil Ltd., Nova Scotia, Canada) are also shown in Figure 4. The differences in the chromatographic profiles, carbon range, the shapes of UCM, distribution of n-alkanes and major isoprenoids, and diagnostic ratios of target alkanes (such as n-C17/pristane and n-C18/phytane) among these products are obviously considerable.
  3.2.4      Lubricating oil
  Petroleum derived lubricating oil is a mixture produced by atmospheric and vacuum distillation of selected paraffinic and naphthenic crude oils. Solvent refining and/or hydrogen treatment are used to remove the non-hydrocarbon constituents and to increase the viscosity index, enhance the color and convert undesirable chemical structures (such as unsaturated hydrocarbons and aromatics) to less chemically reactive species. Solvent dewaxing is then used to remove the wax constituents and to improve the low-temperature properties. Finally, clay treatment or hydrogen treatment is performed to prevent instability of the product. Lubricating oils may be divided into many categories according to the type of services and applications, such as motor oil, transmission oil, crankcase oil, hydraulic fluid, cutting oil, turbine oil, heat-transfer oil, electrical oil, and many others. However, there are two main groups: (1) oils used in intermittent service, such as motor and aviation oils, and (2) oils designed for continuous service, such as turbine oils. Chemical additives are often added to base oil to enhance the properties and to improve such characteristics as oxidation resistance and corrosion resistance of lubricating oil. Small scale lubricating oil spills and contaminations are quite common due to their wide application.
  Figure 4 also includes the high temperature (from 40 to 325 0C) GC-FID chromatograms for two different lubricating oils. In general, lubricating oils have broad GC profiles in the carbon range of C18 to C40 with boiling points greater than 340 0C. Lubricating oil does not contain lower boiling portion of petroleum hydrocarbons. It is largely composed of saturated hydrocarbons, and its GC trace is often dominated by the UCM of hydrocarbons with very small amount of resolved peaks being present. In lubricating oil such as hydraulic fluid, for example, the PAH concentrations can be very low, while the concentration of multi condensed-ring biomarker compounds could be very high. Therefore, determination of these source specific marker compounds often allow for successful identification and correlation between refined products from different sources.
  3.2.5      Waste oil
  Illegal discharges of oil from the machinery rooms of ships (e.g. bilge oil and sludge) have been found to be one of the major sources of oil pollution in areas of intensive shipping traffic (Dahlmann, 2003). The bilge oils often consist of a mixture of light fuel oil, bunker oil and waste lubricating oil. Bilge oil spills often involve different amounts of different products, which make identification of the spill source(s) more difficult. Bilge oils can have great variability in the final composition and therefore they can have significantly different GC-FID chromatograms. The final composition of spilled bilge oil is determined not only by the condition of the ship and ship’s engine but also by the history of this type of oil on board (such as temperature, amount of water, and evaporation of light fuel portion). Mixture of light fuel oil and lubricating oil can be relatively easier identified and distinguished because these two products have different carbon numbers and boiling ranges.
  3.3  PAH Fingerprints of Oils and Petroleum Products
  3.3.1      Distribution of alkylated PAH homologues and other EPA priority PAHs
  Crude oils and refined products from different sources can have very different PAH distributions. Also, many PAH compounds are more resistant to weathering than their saturated hydrocarbon counterparts (n-alkanes and isoprenoids) and volatile alkylbenzene compounds, thus making PAHs one of the most valuable fingerprinting classes of hydrocarbons for oil identification. Examples of PAH distribution of some oils and petroleum products (that is, petrogenic PAHs) are illustrated in Figure 6. The oil products differ significantly in both the PAH concentrations and distribution patterns from the crude oils and from each other. Typically, in unweathered crude oils the alkylated naphthalenes and alkylated chrysenes are the most and least abundant PAHs among the five target alkylated PAH homologues, while many of 4-6 ring unsubstituted PAHs are only minor components or even absent in oils. The PAHs in each alkylated PAH homologous series, in general, exhibit distribution patters where the C1-, C2-, and C3-PAHs are more abundant than the parent (C0-) and C4-PAHs. This kind of characteristic PAH distribution profile has been termed as “bell shaped”. By weathering or degradation, the “bell-shaped” distribution can be readily modified to the distribution profile of C0- < C1- < C2- < C3- (called inverse-sloped) in most alkylated PAH homologous families.
  As Figure 6 shows, Jet A fuel has extremely high content of the naphthalene series (99%) among the five target alkylated PAH homologues, with the other four alkylated PAH series being only 1% in total. In addition, no 4- to 6-ring PAHs were detected of the other 15 EPA priority PAHs. Diesel No. 2 has a high naphthalene content (86%), a low phenanthrene content (5%), and no chrysenes. In the No. 5 fuel and HFO 6303, the unusually high contents of the alkylated naphthalene and chrysene series are very pronounced. In the Orimulsion 400, the concentrations of the alkyl phenanthrenes and dibenzothiophenes are very high, accounting for approximately 38% and 22% respectively of the total PAHs. In addition, a profile in each alkylated PAH family showing the distribution of C0 < C1 < C2 < C3 is very apparent, similar to the severely weathered oil, indicating that this oil was highly degraded during its geological formation.
  3.3.2      Recommended diagnostic ratios of PAHs
  A number of diagnostic ratios of target alkylated PAH species have been developed and successfully used for source identification and differentiation, distinguishing inputs of pyrogenic hydrocarbons from petrogenic hydrocarbons, and weathering indicator (Douglas et al., 1996; Page et al., 1995; Bence et al., 1996; Boehm et al., 1998 and 2001; Hostettler et al., 1999; Short et al., 1997). These are briefly summarized in Table 4.
  Basic criteria which must be applied in selection of diagnostic ratios include: (1) variability, that is, ability to discriminate between samples; (2) analytical precision of selected ratios; and (3) resistance to weathering. A benefit of comparing diagnostic ratios of spilled oil and suspected source oils is that any concentration effects are minimized. In addition, the use of diagnostic ratios to correlate and differentiate oils tends to induce a self-normalizing effect on the data since variations due to instrument operating conditions, operators, or matrix effects are minimized. Wang et al. (Wang et al., 1994a and 1994b, 1997a, 1998a, 2002) utilized a number of diagnostic ratios of selected source-specific alkylated PAHs in combination with determination of ratios of selected paired biomarkers for source identification and differentiation, determination of weathering extent and degree of surface and subsurface samples, and distinguishing between composition changes due to physical weathering and biodegradation.
  3.3.3     PAH isomer and cluster PAH analysis
  Ratios of individual source-specific isomers within the same alkylation level and the relative distributions of isomer-to-isomer have been used for oil spill source identification. As the alkylation levels increase, more isomers are detected (for example, the C3-dibenzothiophenes as a group, contain more than 20 individual isomers with different relative abundances). The differences between the isomer distributions reflect the differences of the depositional environment during oil formation. Compared to PAH homologous groups at different alkylation levels, higher analytical accuracy and precision may be achieved due to the close match of physical/chemical properties of the isomers. Also, the relative distribution of isomers is subject to little interference from weathering in short-term or lightly weathered oils. Hence this approach can be positively used for oil spill identification. On the other hand, it has been demonstrated that the position of the alkylation on the PAHs can influence the biodegradation rate of the isomers within an isomer group. This information can be used to sort out environmental factors such as the impact of biodegradation on the PAH distribution and to differentiate oil compositional changes due to physical weathering from those due to biodegradation. For example, the ratios among methyl dibenzothiophenes, methyl-phenanthrenes, and methyl and dimethyl naphthalenes have been studied and widely used for environmental forensic investigations.
  (1) Methyl phenanthrenes. All oils contain four methyl-phenanthrenes (in two pairs of doublet peaks: 3- and 2-, and 4-/9- and 1-m-P). Ratios among four methyl-phenanthrene isomers have been demonstrated to be related to the thermal history of crude oils and its source strata, and numerous methyl-phenanthrene indices have been defined for monitoring the thermal maturities of oils (Radke, 1986) and for spill oil source identification (Wang et al., 1999a):
  MPI 1 = 1.5(2-m-P + 3-m-P)/(P + 1-m-P + 9-m-P)
  MPI 2 = 3(2-m-P)/(P + 1-m-P + 9-m-P)
         (3- + 2-m-P)/(4-/9- + 1-m-P)
  The 2-methylphenanthrene was found to be more sensitive to biodegradation than the 1-methylphenanthrene (Wang et al., 1998c) and, therefore, it can be used as the indicator for biodegradation. It has also reported (CEN, 2002) that in many crude oils, the first doublet peak is smaller than the second doublet peak, and the methyl-anthracene (the peak between the two doublet peaks) is often very small or insignificant. While, for many Bunker fuels, the first doublet peak is higher than the second one, and the methyl-anthracene is often more pronounced.
  (2)   Methyl-dibenzothiophenes. Chromatographically well-resolved C1-dibenzothiophene isomers (Fayad and Overton, 1995; Wang et al., 1995b) are present in all oils at relatively high concentrations. Their relative abundance distributions vary significantly from different sources:
  C1-dibenzothiophene distribution index = (4-:2-/3-:1-m-DBT)
  A database of the relative ratios of the C1-DBT isomers for several hundred crude, weathered and biodegraded oils, and petroleum products has been established and plots of 2-/3-methyldibenzothiophene versus 1-methyldibenzothiophene (both isomers are normalized relative to 4-methyldibenzothiophene) for these oils and oil products have been depicted. The plots show the data points representing the various oils are very scattered. Another pronounced feature observed from the figure is that related oils produce tight clusters on the plot. The use of these ratios complements existing methods of oil characterization, but has its own distinct advantages for discrimination of different oils.
  (3) Other relative ratios of PAH isomers
  Other selected PAH isomers used for oil fingerprinting studies include ratio of retene (1-methyl-7-(1-methylethyl)-phenanthrene) to the total of C4-phynanthrenes, 3 isomers each within C3-naphthalenes (m/z 156) and C4-naphthalenes (m/z 170), 4 isomers within C2-phenanthrenes (m/z 206), 2 isomers within C4-phenanthrenes (m/z  234), 3 isomers within C1-fluorenes (m/z 180), 2-m-naphthalene/1-m-naphthalene (m/z 128), anthracene/phenanthrene (m/z 178), BaA/Chrysene (m/z 228), BeP/BaP (m/z 252), and indeno[1,2,3-cd]pyrene/benzo[ghi]perylene (m/z 276). Depending on the individual spill case and its degree of weathering, different diagnostic parameters may be selected and applied.
  (4)   Cluster PAHs at m/z 216
  As high-boiling biomarkers are rarely present in lighter fuels, it becomes increasingly difficult to compare two lighter fuel samples based on lower boiling compounds for source identification, especially for weathered fuel oils. It is found (Dahlmann, 2003) that the cluster PAH compounds at m/z 216 are relatively stable and can be used for comparing lighter fuel oil samples. Actually, not all compounds of this cluster are isomers. This cluster mainly represents six PAH compounds from different compound classes of aromatic hydrocarbons. It has, therefore, some advantages for discriminating between oils than isomer cluster within a single compound class. These six cluster compounds have been identified to be 2-m-fluoranthene, benzo(a)-fluorene, benzo(b)-fluorene, 2-m-pyrene, 4-m-pyrene, and 1-m-pyrene. By normalizing the pea abundances relative to 4-m-pyrene, which is often the most abundant among the cluster, a set of diagnostic ratios can be readily determined. This cluster ratio has been used for comparing three Round Robin spill fuel samples collected from a harbour spill in the Netherlands in 2004.
  3.4  Biomarker Fingerprints of Oils and Petroleum Products
  Biomarkers are useful in oil spill identification because they retain all or most of original carbon skeleton of the original natural product and this structural similarity reveals more information about their origins and thermal history than other compounds (Peters and Moldowan, 1993; Peters et al., 2005; Philp, 1985). In comparison to n-alkanes and acyclic isoprenoids, many biomarkers are much resistant to secondary processes, such as biodegradation. Therefore, chemical analysis of source-characteristic and environmentally-persistent biomarkers generates information of great importance in determining the source of spilled oil, differentiating oils, monitoring the degradation process and weathering state of oils under a wide variety of conditions. In the past decade, use of biomarker fingerprinting techniques to study spilled oils has greatly increased, and biomarker parameters have been playing a prominent role in almost all oil spill work.
  3.4.1     Distributions and quantification of biomarkers
  The cyclic terpane biomarkers in petroleum include sesqui- (C15), di- (C20), sester- (C25), and triterpanes (C30). The steranes are a class of biomarkers containing 21 to 30 carbons that are derived from sterols, and they include regular steranes, rearranged diasteranes, and mono- and tri-aromatic steranes. Among them, the regular C27-C28-C29 homologous sterane series are the most common and useful steranes because they are highly specific for correlation (Peters and Moldowan, 1993). Characterization of these compounds are achieved by using GC-MS in the selected ion monitoring mode by the internal standard method: m/z 191 for tricyclic, tetracyclic and pentacyclic terpanes, m/z 123 for bicyclic sesquiterpanes, m/z 217 and 218 for steranes, m/z 217/259 for diasteranes, m/z 253 for mono-aromatic steranes, and m/z 231 for tri-aromatic steranes.
  Many oils show different composition and distribution patterns of biomarkers. The GC-MS chromatograms of terpanes (m/z 191) are often characterized by the terpane distribution in a wide range from C20 to C30 often with C23 and C24 tricyclic terpanes and C29 αβ- and C30 αβ- pentacyclic hopanes being prominent. As for steranes (at m/z 217 and 218), the dominance of C27, C28, and C29 20S/20R homologues among the C20 to C30 steranes is often apparent. Figure 7 shows GC-MS-SIM chromatograms at m/z 191 for Sockeye oil (California), Orimulsion-400 (Venezuela), HFO 6303, Diesel No.2 (Ontario), hydraulic oil (No. 1), and hydraulic oil (No. 3), respectively.
  In addition of different distribution patters, the concentrations of biomarkers vary widely with the type of depositional environment (oxic/aoxic, freshwater/marine/hypersaline), type of organic matter (e.g. terrigenous origin or marine origin), maturity and biodegradation as well. For a given type of organic material, the biomarker concentrations generally decrease with increasing thermal maturity. Very light oils or condensates (e. g. the Scotia Light) typically contain low detectable biomarkers. Characterization of biomarkers should include determination of both concentrations and relative distributions, and should not be just measuring peak ratio alone. This is important because it is possible to have situation where a source might have similar biomarker ratio but very different actual amounts of biomarkers.
  It should be emphasized that if an oil shows any additional characteristic compositional features (such as “extra” biomarker peaks), these should of course always be included in the characterization and considered in the identification and correlation.
  3.4.2     Low boiling sesquiterpanes in oils and lighter petroleum products
  The bicyclic biomarkers comprise one of the terpenoid classes. Sesquiterpanes (C14 to C16) with the drimane skeleton are ubiquitous components of crude oils and ancient sediments (Alexander et al., 1984; Philip, 1981; Philp, 1985; Fan et al., 1991). For lighter petroleum products, the high boiling-point pentacyclic triterpanes and steranes are generally absent or in very low abundance, while the low boiling sesquiterpanes are more concentrated in these distillates. The sesquiterpanes elute out between n-C13 and n-C16 in the GC-MS-SIM chromatograms and are monitored using m/z 123, a base fragment ion common to all sesquiterpanes. Confirmation ions for the sesquiterpanes include m/z 179 (the ion after sesquiterpane C14H26 loses CH3), 193 (the ion after C15H28 loses CH3 or after C16H30 loses C2H5), and 207 (the ion after C16H30 loses CH3). Examination of GC-MS chromatograms of these characteristic ions of sesquiterpanes provide highly diagnostic means of correlation, differentiation and source identification for lighter petroleum products, in comparison to the use of other hydrocarbon groups. The distribution patterns of sesquiterpanes generally vary in crude oils and in refined petroleum products from different sources (Stout et al., 2005; Wang et al., 2005; Wang and Stout, 2007; Yang et al., 2009).  
  Furthermore, diagnostic ratios of selected paired sesquiterpanes for a large number of oils and petroleum products have also developed (Wang et al., 2005; Yang et al., 2009). In general, oils from different regions have ratios that cover quite a wide range. Cross-plots of diagnostic ratios of Peak 4/Peak 5 (C15) versus the ratios of Peak 3/Peak 5 (C15) for over 50 crude oils and refined products (both isomers 3 and 4 are normalized relative to isomer 5) demonstrates that different oils have different ratio values of Peak 4/Peak 5 and Peak 3/Peak 5, which fall in ranges of 0.2 to 1.2 and 0.1 to 2.1, respectively. Another feature is that related oils produce tight clusters on the plot (such as the cluster for Orimulsion samples from different batches and of the original Orinoco Bitumen). This observation implies that the ratios of sesquiterpane isomers, in combination with other fingerprinting data, may be used to discriminate different oils and to identify the source of spill samples.
  3.4.3     Diagnostic ratios (indices) of biomarkers
  Biomarker diagnostic parameters have been long established and are widely used by geochemists (Peters and Moldowan, 1993; Peters et al., 2005) for oil correlation, determination of organic input and depositional environment; for assessment of thermal maturity; and for evaluation of oil biodegradation. Many of biomarker diagnostic parameters currently used in oil spill studies are originated from geochemistry parameters (CEN, 2002; Faksness et al., 2002; Wang et al., 1994a; Zakaria et al., 2000 and 2001; Barakat et al., 1997; Wang et al., 1998b; Wang et al., 2007). Table 5 lists some of the primary diagnostic ratios of biomarkers frequently used by the environmental chemists for spilled oil identification, correlation and differentiation.
  During January and February 1996, a significant number of tarball incidents occurred along the coasts of Vancouver Island of British Columbia (BC), Washington (WA), Oregon (OR), and California (CA). The diagnostic values of “source specific” biomarker and PAH isomer compounds of representative tarball samples and the suspected source Alaska North Slope (ANS) oil were determined and compared (Wang et al., 1998b). The results clearly reveals the following: (1) almost all diagnostic ratios for the ANS oil differ significantly from those of the tarball samples, indicating the ANS oil was not the source oil of tarball samples; (2) all the relative ratios are almost identical for sample BC-1 and BC-2, indicating they were from the same source; and (3) the tarball sample from CA was very similar in concentrations and diagnostic ratios of target biomarkers with the samples BC-1 and BC-2, but it had markedly different PAH isomeric ratios from samples BC-1 and BC-2, indicating CA tarballs may have another source different than the BC samples.
  3.5  Weathering Effects on Spill Oil Fingerprints (3rd Control Factor)
  3.5.1      Oil weathering
  When crude oil or petroleum products are accidentally released to the environment, whether on water or land, they are immediately subject to a wide variety of changes in physical and chemical properties that in combination are termed “weathering”. The most important weathering processes include evaporation, dissolution, dispersion and microbial degradation. In the short term of a spill, evaporation is the single most important and dominant weathering process. The rate of weathering of an oil can be very different and are controlled by a number of spill conditions and natural processes such as the type of oil spilled, the local environmental conditions, and natural population of indigenous microbial and microbiological activities during and after spill. In the first few days following a spill, the weathering is largely caused by evaporation and the loss can be up to 70 and 40% of the volume of light and medium crudes, respectively. For heavy or residual oils the losses are only about several percentages of volume.
  Major chemical compositional changes due to weathering are summarized as the following:
  (1) For lightly weathered oils and refined products (for example, <15% weathered), the abundances of low end n-alkanes are significantly reduced. However, the ratios of n-C17/pristane and n-C18/phytane are found to be virtually unaltered. The losses of BTEX and C3-benzene compounds are obvious, and the most abundant 2-ring alkylated naphthalene series appear slightly enriched.
  (2) For moderately weathered oils and refined products (e.g., ~15-30% weathered), significant losses occur in n-alkanes and relatively low-molecular-weight isoprenoid compounds. Rapid loss of volatile aromatic compounds is clear. The loss of C0 and C1- naphthalenes can be significant. The ratio of GC-resolved peaks to UCM can be considerably decreased due to the preferential loss of resolved hydrocarbons over the unresolved complex hydrocarbons. The biomarker compounds are enriched.
  (3) For severely weathered oils and refined products, not only n-alkanes but branched and cyclo-alkanes are heavily or completely lost, and the UCM becomes extremely pronounced, resulting in significant increase in relative ratios of UCM/GC-TPH and in substantial decrease in relative ratios of resolved peaks to GC-TPH. The BTEX and alkyl benzene compounds are completely lost. Pronounced decrease in the alkylated naphthalene series relative to other alkylated PAH homologous series is clearly observed. A profile in each alkylated PAH family showing the distribution of C0- < C1- < C2- < C3- is clearly developed. The alkylated chrysene series is significantly enhanced relative to other PAH series. Biomarker compounds are concentrated because of their refractory nature and high resistance to biodegradation.
  As an example, Figures 8A and 8B shows the GC chromatograms for TPH and n-alkane analysis and GC-SIM-MS chromatograms for alkylated benzene and alkylated PAH distribution analysis for 25-year-old Nipisi spilled oil samples (Wang et al., 1998a), illustrating the effect of field weathering conditions and sample depths on chemical composition changes of the spilled oil during the 25 year period. Figure 8 clearly demonstrates that 25 years after the spill, the remaining underground oil is still relatively Afresh@ in comparison to the reference oil and the surface residual oil. Subsurface oil degradation has been demonstrated to be slow and will proceed at a very slow rate because the peat in this wetland habitat system is acidic, and water saturated, i.e., largely anaerobic. Also, the average annual temperature is only 1.7 0C, based on 22 years of weather records.
  3.5.2      Determination of weathering degree
  As discussed above, oil weathering is a very complex process. The weathering degree and the weathering rate are determined by many factors. In the early biodegradation studies, the ratios of biodegradable to less biodegradable compounds such as n-C17/pristane and n-C18/phytane were largely used for estimating biodegradation degree and for comparing the weathering state of spilled oil. For lightly and some moderately weathered oils, these ratios provide a useful tool for estimation of weathering degree and for oil source identification and differentiation. In severely weathered oils, however, the n-alkanes and even the isoprenoids (including pristane and phytane) may be partially or completely depleted. Under such circumstances, the use of the traditional measure of n-C17/pristane and n-C18/phytane might substantially underestimate the extent of biodegradation and weathering degree because isoprenoids also degrade to a significant degree. Later, highly degradation-resistant components such as C30 17α(H), 21β(H)-hopane are selected to serve as conserved “internal standards” for determining rate and extent of weathering for the spilled residual oil (Butler et al., 1991; Prince et al., 1994; Wang et al., 1995c; Wang et al., 2001):  
  P (%) = (1-Cs/Cw) x 100  
  where P is the weathered percentages of the weathered samples, Cs and Cw are the concentrations of C30 αβ-hopane in the source oil and weathered samples, respectively.  A number of studies have demonstrated that this method can provide a more accurate representation of the degree of biodegradation than do the traditional alkane/isoprenoid hydrocarbon ratios.
  For refined products, such as diesel and jet fuel samples, which may not contain significant quantities of biomarker compounds and chrysenes, less “conservative” PAHs with a high degree of alkylation such as C4 or C3-phenanthrenes can be selected and used as alternative internal standards to evaluate the weathering degrees.
  3.6  Oil Mixing: Using Multi-Criterion Approach For Source Identification of an Unknown Spill (4th Control Factor)
  An oil spill to the Rouge River and Detroit River was discovered and reported in the second week of April (April 8-13, 2002). The oil of several thousand gallons (by estimation) spilled into the Rouge River and travelled about two miles to the Detroit River. It then floated in several small patches down the river into northern Lake Erie. Several thousand gallons more spilled into the Rouge River during that weekend. The two spills were related, and heavy rains flushed the additional oil out of the sewer and into the river. Environment Canada (EC) Ontario Region conducted an aerial survey of the Detroit River. They also surveyed the majority of the areas by vessel. The spill impacted approximately 43 kilometers of U.S. and Canadian shorelines. The presence of sheen over the majority of the impacted river area was observed. On the shore it appeared as a black coat and typically 0.2 to 1.0 mm thick. EC Ontario Region collected a number of spill samples from various spots and sent 11 samples to our Oil Research Lab for analysis. The integrated multi-criterion analytical approach was applied for this case study to defensibly identify the spilled oil (Wang et al., 2004).
  After the sample extractions, appropriate volumes of the concentrated extracts containing approximately 30-40 mg of total solvent extractable materials (TSEM) were spiked with appropriate amounts of surrogates and then quantitatively transferred into chromatographic columns for sample cleanup and fractionation. Hexane (12 mL) and 50% benzene in hexane (v/v, 15 mL) were used to elute the saturated and aromatic hydrocarbons, respectively. For each sample, half of the hexane fraction (F1) was used for analysis of the total GC-detectable saturates, n–alkanes and isoprenoids, and biomarker compounds; half of the 50% benzene fraction (F2) was used for analysis of alkylated homologous PAHs and other EPA priority unsubstituted PAHs; and the remaining halves of F1 and F2 were combined into a fraction (F3) and used for the determination of the TPH and UCM.
  3.6.1   Product type screen and determination of hydrocarbon groups
  Figure 9 shows the GC-FID chromatograms of Fraction 3 of three representative spill samples for TPH and n-alkane analysis. The saturated fractions F1 demonstrated very similar GC-FID chromatogram profiles to their corresponding Fraction 3. Table 6 summarizes the hydrocarbon group analysis results of the spill samples. The major chemical composition features of TPH and saturate hydrocarbons in the samples are summarized as follows:
  (1) The GC traces of both the F1 and F3 of the spill samples are clearly dominated by large UCM (located in the n-C18 - n-C36 range) with almost no n-alkane being detected after n-C20. The ratios of the all GC-resolved peaks to the total GC area were determined to be only 0.06 for 3 samples (Table 6). The GC chromatographic profile and shape of the UCM “humps” are significantly different from crude oils and most refined products. In addition, the ratios of the total saturates to the GC-TPH were determined to be around 90%, much higher than most crude oils. All the GC trace features (Figure 9) suggest that the major portion of the spilled oil might be a lube oil. 
  (2) The resolved n-alkanes mainly distributed in the diesel carbon range (C8-C27), suggesting the minor portion of the spill oil being a diesel. No n-alkane with the carbon number smaller than C10 and greater than C24 was detected. The total n-alkanes including pristane and phytane were determined to be only 9.3, 10.4, and 8.6 mg/g of TSEM for samples #1, #2, and #3 respectively.  Using the estimation value of 120 mg n-alkanes per gram diesel and in consideration of weathering effect, the percentage of diesel in the spill samples may be estimated not exceeding 20% of the total hydrocarbons detected.
  (3) Three samples showed nearly identical GC chromatographic profiles, n-alkane distribution patterns, as well as the nearly identical diagnostic ratios (Table 8) of n-C17/pristane, n-C18/phytane, and pristane/phytane. This implies that they were most likely the same oil and from the same source, and some small differences were likely caused by weathering.
  (4) All quantitative GC results implied that the spill samples were largely composed of lube oil mixed with smaller portion of diesel fuel, the diesel in the samples had been weathered and degraded; and the diesel portion in the sample #3 had been more weathered (most probably by more evaporation and water-washing in its longer journey from spill source to the destination) than samples #1 and #2.
  3.6.2      Determination of oil-characteristic alkylated PAH homologues and their diagnostic ratios
  PAH analysis results demonstrate the following:
  (1) The relative distribution patterns and profiles of alkylated PAHs are very much the same for the spilled samples, in particular for samples #1 and #2, further implying they were from the same source.
  (2) The 5 target alkylated PAH homologous series and other EPA priority PAHs were determined to be 1404, 1479, and 1028 μg/g TSEM, and 250, 257, and 167 μg/g TSEM for samples #1, #2, and #3, respectively. Compared to crude oils and most refined products such as Jet fuel and diesel (>10,000 μg/g for most oils), the PAH concentrations in these spill samples are quite low. The dominance of alkylated naphthalene and phenanthrenes among 5 target alkylated PAH homologous series is pronounced for all three samples.
  (3) Sample #2 still contained small amount of BTEX and C3-benzene compounds. In comparison, almost no BTEX and other alkyl benzene compounds were detected in samples #1 and #3. This fact further demonstrates that the sample #2 was least weathered. The loss of lighter molecular weight naphthalene and C1- and C2-naphthalenes was obvious for all three samples, resulting in development of the relative distribution of C0-N < C1-N < C2-N < C3-N. This relative distribution pattern is particularly obvious for the more weathered sample #3. For other EPA priority PAHs, the more weathered sample #3 also demonstrated lower concentrations of lighter 2- and 3-ring PAHs (biphenyl, acenaphthylene, and acenaphthene).
  (4) Analysis of the diagnostic ratios of “source-specific” PAH isomers clearly reveals that: the relative distribution of PAH isomers 4-, 2-/3-, and 1-methyl dibenzothiophene at m/z 198, and (3- + 2-methyl-phenanthrene) to (4-/9- + 1-methyl-phenanthrene) at m/z 192 were found to be very closely matching; and the double ratios (C2D/C2P : C3D/C3P) were also nearly identical (0.22:0.31, 0.22:0.30, and 0.23:0.30 for samples #1, 2, and 3 respectively).
  It has been well demonstrated that in general, lube oils only contain small quantities of PAH compounds while PAH concentrations are high in diesel. Obviously, detected PAHs in these spill samples were largely contributed by the small portion of diesel in spill samples.
  3.6.3    Input of pyrogenic PAHs to the spill samples
  Another pronounced PAH compositional feature is that among the alkylated phenanthrene, fluorene, and chrysene series, the parent PAH phenanthrene, fluorene, and chrysene are most abundant, their concentrations are even higher than their corresponding alkylated homologous constituents. In particular, the highest abundance of parent chrysene over its alkyl-substituted homologues and the decrease in relative abundances with increasing level of alkylation (that is, in the order of C0-C > C1-C > C2-C > C3-C) was very pronounced. Tthis kind of PAH distribution profile has been generically termed as “skewed or sloped”. The “Pyrogenic Index” (Wang et al., 1999) were determined to be as high as 0.16 for three samples, far higher than the corresponding values for crude oils and refined products (exclusively smaller than 0.06), defensively indicating the formation and presence of pyrogenic PAHs in the spill samples. In addition, the relative ratios of chrysene to benz[a]anthracene were determined to be very close to 1.0, also far higher than the same ratios for crude oils and refined products. All these features indicate the input of pyrogenic PAHs.
  The most likely source of pyrogenic PAHs in used motor oils is combustion “blow-by” past the piston rings of exhaust gasses directly into the crankshaft cavity. Excessive heat in the motor lubrication process can also increase the concentration of PAHs, in particular the high molecular weight PAHs, in used lube oil. Therefore, it can be reasonably concluded that the pyrogenic PAHs found in the spilled oil were most probably produced from combustion and motor lubrication process, and the lube oil in these spill samples was waste lube oil.
  3.6.4   Characterization of biomarker compounds
  Biomarker characterization results reveal the following:
  (1) The samples show nearly identical distribution patterns of biomarkers and these biomarkers were mostly from the lube oil portion of the spill samples. It has well demonstrated that diesels do not contain high molecular weight biomarkers and only contain trace of low molecular weight biomarker compounds (C20 -C24).
  (2) The total of the target biomarkers were determined to be 1103, 941, and 941 μg/g TSEM for samples #1, #2 and #3, respectively.
  (3) The diagnostic ratios of target biomarker compounds C23/C24, C29 αβ-hopane/C30 αβ-hopane, Ts/Tm, C31(22S)/C31(22S+22R), C32(22S)/C32(22S+22R), C33(22S)/C33(22S+22R), C34(22S)/C34(22S+22R), C35(22S)/C35(22S+22R), and C31/(C31 to C35), are also found to be very much the same. All these evidences, in combination with the TPH and PAH analysis results, unambiguously point toward to the conclusion that the three spill samples came from the same source.
  It is important to note that the fingerprinting results described above highlight the necessity to analyze for more than one suite of analytes in source identification. Characterization of PAH and biomarker compounds must include determination of both concentrations and relative distributions, and should not be just measuring peak ratios alone. This is important because it is possible to have situation where a source might have similar biomarker ratio but very different actual amounts of biomarkers. In summary, The fingerprinting results described above highlight the necessity to analyze for more than one suite of analytes in forensic investigation and spill source identification: (1) the spill samples were largely composed of used lube oil mixed with smaller portion of diesel fuel; (2) the diesel in the samples had been weathered and degraded; (3) the diesel portion in sample #3 collected from N. Boblo Island was more weathered (most probably by more evaporation and water-washing) than samples #1 and #2; (4) three samples were from the same source; (5) most PAH compounds were from the diesel portion in the spill samples, while the biomarker compounds were largely from the lube oil portion; (6) input of pyrogenic PAHs (they were most probably produced from combustion and motor lubrication processes) to the spill samples was apparent.
  The advances in petroleum hydrocarbon fingerprinting and data interpretation methods and approaches in the last two decades have now allowed for detailed qualitative and quantitative characterization of spilled oils. Chemical fingerprinting is a powerful “tool” for hydrocarbon source identification and differentiation, when it is applied properly. However, in many cases, particularly for complex hydrocarbon mixtures or extensively weathered and degraded oil residues, there is no single fingerprinting analysis which can meet the objectives of forensic investigation and quantitatively allocate hydrocarbons to their respective sources. Under such circumstances, integrated “multiple-parameter” approaches are always needed and used, more than one suite of analytes must be performed, and other independent techniques such as isotope analysis may be applied to support correlations. If large number of spill and source candidate samples are involved, statistical and numerical analysis techniques (such as principal component analysis) for data analysis are always performed. Development in hydrocarbon fingerprinting techniques will continue as analytical and statistical techniques evolve. It can be anticipated that these developments will further enhance the utility and defensibility of oil hydrocarbon fingerprinting.
  Alexander, R., Kagi, R. I., Noble, R. A., Volkman, J. K., Identification of some bicyclic alkanes in petroleum, in Advances in Organic Geochemistry, 1983, Vol.6; Schenck, P. A., De Leeuw, J. W., Lijmbach, G. W. M., Eds., Pergamon Press, Oxford, 1984.   
  ASTM Method 3328-90, In: Annual Book of ASTM Standards, Water (II), Vol. 11.02. American Society for Testing and Materials, Philadelphia, PA, 1997.  
  Barakat, A. O., Mostafa, A., El-Gayar, M. S., Rullkotter, J., Source-dependent biomarker properties of five crude oils from the Gulf of Suez, Egypt, Org. Geochem., 26, 441-450, 1997.   
  Bence, A. E., Kvenvolden, K. A., Kennicutt II, M. C., Organic geochemistry applied to environmental assessments of Prince William Sound, Alaska, after the Exxon Valdez oil spill - a review, Organic Geochemistry, 24, 7-42, 1996.  
  Boehm, P. D., Page, D. S., Gilfillan, E. S., Bence, A. E., Burns, W. A., Mankiewicz, P. J., Study of the fate and effects of the Exxon Valdez oil spill on benthic sediments in two bays in Prince William Sound, Alaska. 1. Study design, chemistry and source fingerprinting, Environ. Sci. Technol., 32, 567-576, 1998.   
  Boehm, P. D., Page, D. S., Burns, W. A., Bence, A. E., Mankiewicz, P. J., Brown, J. S., Resolving the origin of the petrogenic hydrocarbon background in Prince William Sound, Alaska, Environ. Sci. Technol., 35, 471-479, 2001.  
  Bouyssiere, B., Leonhard, P., Profrock, D., Baco, F., Garcia, C. L., Wibur, S., and Pronge, A., Investigation of the sulfur speciation in petroleum products by capillary GC with ICP-collision cell-MS detection, J. Anal. At. Spectrom., 2004, 19, 700-702.  
  Butler, E. L., Douglas, G. S., Steinhauter, W. S., Prince, R. C., Axcel, T., Tsu, C. S., Bronson, M. T., Clark, J. R., Lindstrom, J. E., in: On-site Reclamation (R. E. Hinchee and R. F. Olfenbuttel, eds.), Butterworth-Heinemann, Boston, MA, pp 515-521, 1991.  
  CEN, Revision of the Nordtest methodology for oil spill identification for the European Committee for Standardization (CEN), SINTEF Report STF66 A02027, Norway, 2002. 
  Dahlmann, G., Characteristic features of different oil types in oil spill identification, Berichte des BSH 31, ISSN 0946-6010, Germany, 48 pages, 2003.
  Daling, P. S., Faksness, L. G., Hansen, A. B., and Stout, S. A., Improved and standardized methodology for oil spill fingerprinting, Environmental Forensics, 3, 263-278, 2002.  
  Dallüge, J., Beens, J., Brinkman, U., Comprehensive two-dimensional GC: a powerful and versatile analytical tool, J. Chromatogr. A, 1000, 69-108, 2003
  Dimandja, J. D., GC X GC, Anal. Chem., 76, 167A-174A, 2004  
  DOE 2004, Energy Information Administration Website, United States Department of Energy, http://eia.doe.gov/, 2004. 
  Douglas, G. S., Bence, A. E., Prince, R. C., McMillen, S. J., Butler, E. L., Environmental stability of selected petroleum hydrocarbon source and weathering ratios, Environ. Sci. Technol., 30, 2332-2339, 1996.  
  Douglas, G. S., Burns, W. A., Bence, A. E., Page, D. S., Boehm, P., Optimizing detection limits for the analysis of petroleum hydrocarbons in complex environmental samples, Environ. Sci. & Technol. 38, 3958-3964, 2004.   
  EPA, EPA Requirements for Quality Assurance Project Plans, EPA QA/R-5, U. S. EPA, Washington, DC, 2001.
  ETC Method (updated version), Analytical Methods for Determination of Oil Components, ETC Method No.: 5.3/1.3/M, 2002, Environmental Technology Centre, Environment Canada, Ottawa, Ontario, 2002. 
  Faksness, L. G., Daling, P. S., Hansen, A. B., Round Robin Study - Oil Spill Identification, Environmental Forensics, 3, 279-292, 2002.
  Fan, P., Qian, Y., Zhang, B., Characteristics of biomarkers in the recent sediments from Qinghai Lake, Northwest China, J. Southeast Asian Earth and Science, 5, 113-128, 1991.
  Fayad, N. M., Overton, E., A unique biodegradation pattern of the oil spilled during the 1991 Gulf war, Mar. Pollut. Bull., 30, 239-246, 1995
  Figas, M., The Basics of Oil Spill Cleanup (2nd ed.), Lewis Publishers, New York, 2001.
  Hostettler, F. D., Rosenbauer, R. J., Kvenvolden, K. A., PAH refractory index as a source discriminant of hydrocarbon input from crude oil and coal in Prince William Sound, Alaska, Org. Geochem., 30, 873-879, 1999. 
  Marshal, A. G., Podgers, R., Petroleomics: the next grand challenge for chemical analysis. Acc. Chem. Res. 37: 53-59, 2004.
  National Research Council (NRC), Oil in the Sea III: Inputs, Fates, and Effects, The National Academies Press, Washington, DC, 2002
  Peters, K. E., Moldowan, J. W., The Biomarker Guide: Interpreting Molecular Fossils in Petroleum and Ancient Sediments, Prentice Hall, New Jersey, 1993. 
  Peters, K. E., Walters, C., Moldowan, J. W., The Biomarker Guide (2nd Edition), Cambridge University Press, UK, 2005. 
  Page, D. S., Boehm, P. D., Douglas, G. S., Bence, A. E., in: Exxon Valdez Oil Spill: Fate and Effects in Alaska Waters (P. G. Wells, J. N. Butler, and J. S. Hughes, eds.), ASTM STP 1219, ASTM, Philadelphia, PA, pp 41-83, 1995.
  Philp, R. P., Gilbert, T. D., Riedrich, J., Bicyclic sesquiterpanoids and triterpenoids in Australia crude oils, Geochem. Cosmochim. Acta., 45, 1173-1180, 1981. 
  Philp, R. P., Fossil Fuel Biomarkers, Application and Spectra, Elsevier, New York, 1985. 
  Prince, R. C., Elmendorf, D. L., Lute, J. R., Hsu, C. S., Haith, C. E., Senius, J. D., Dechert, G. J., Douglas, G. S., Butler, E. L., 17(H), 21(H)-hopane as a conserved internal marker for estimating the biodegradation of crude oil, Environ. Sci. Technol., 28, 142-145, 1994. 
  Qian, K., Dechert, G., Recent advance in petroleum characterization by GC field time-of-flight high resolution mass spectrometry, Anal. Chem., 74, 3977-3983, 2002. 
  Radke, M. D., Welte, H., Willsch, H., Maturity parameters based on aromatic hydrocarbons: influence of the organic matter type, Organic Geochem., 10, 51-63, 1986. 
  Sauer, T. C., Uhler, A. D., Pollutant source identification and allocation: advances in hydrocarbon fingerprinting, Remediation, 25-50, Winter Issue, 1994-1995. 
  Short, J. W., Heintz, R. A., Identification of Exxon Valdez oil in sediments and tissues from Prince William Sound and the Northwestern Gulf of Alaska based on a PAH weathering model, Environ. Sci. Technol., 31, 2375-2384, 1997. 
  Speight, J. G., Handbook of Petroleum Product Analysis, Wiley-Interscience, Hoboken, NJ, 2002.
  Stout, S. A., Uhler, A. D., McCarthy, K. J., Emsbo-Mattingly, S., in: Introduction to Environmental Forensics (B. L. Murphy and R. D. Morrison, eds.), Academic Press, London, Chapter 6, pp137-260, 2002.
  Stout, A. S., Allen, D. U., Kevin, J. M., Middle distillate fuel fingerprinting using drimane-based bicyclic sesquiterpanes, Environmental Forensics, 6, 241-252, 2005.
  Stout, A. S. and Wang Z. D, Chapter 1: Chemical Fingerprinting of Spilled or Discharged Petroleum, in Wang, Z. D. and Stout, S.A. (eds.), Oil Spill Environmental Forensics: Fingerprinting and Source Identification, Academic Press, Boston, MA, pp. 1-53, 2007.
  Uhler, A. D., Stout, S. A., McCarthy, K. J., Increased success of assessments at petroleum sites in 5 steps, Soil and Groundwater Cleanup, Dec/Jan, 13-19, 1998-1999.
  Wang, Z. D., Fingas, M., Sergy, G., Study of 22-year-old Arrow oil samples using biomarker compounds by GC/MS, Environ. Sci. Technol., 28, 1733-1746, 1994a.
  Wang, Z. D., Fingas, M., Li, K., Fractionation of ASMB Oil, Identification and Quantitation of Aliphatic, Aromatic and Biomarker Compounds by GC/FID and GC/MSD, J. Chromatogr. Sci., 32, 361-366 (Part I) and 367-382 (Part II), 1994b.
  Wang, Z. D., Fingas, M., Landriault, M., Sigouin, L., Xu, N., Identification of alkylbenzenes and direct determination of BTEX and (BTEX + C3-Benzenes) in oils by GC/MS, Anal. Chem., 67, 3491-3500, 1995a.
  Wang, Z. D., Fingas, M., Use of methyldibenzothiophenes as markers for differentiation and source identification of crude and weathered oils, Environ. Sci. Technol., 29, 2841-2849, 1995b.
  Wang, Z. D., Fingas, M., Sergy, G., Chemical characterization of crude oil residues from an Arctic Beach by GC/MS and GC/FID, Environ. Sci. Technol., 29, 2622-2631, 1995c.
  Wang, Z. D., Fingas, M., Landriault, M., Sigouin, L., Feng, Y., and Mullin, J., Using systematic and comparative analytical data to identify the source of an unknown oil on contaminated birds, J. Chromatogr., 775, 251-265, 1997a.
  Wang, Z. D., Fingas, M., Blenkinsopp, S., Sergy, G., Landriault, M., Sigouin, L., Study of the 25-year-old Nipisi oil spill: persistence of oil residues and comparisons between surface and subsurface sediments, Environ. Sci. Technol., 32, 2222-2232, 1998a.
  Wang, Z. D., Fingas, M., Landriault, M., Sigouin, L., Castel, B., Hostetter, D., Zhang, D., and Spencer, B., Identification and linkage of tarballs from the coasts of Vancouver Island and northern California using GC/MS and isotopic techniques, J. High Resolut. Chromatogr., 21, 383-395, 1998b.
  Wang, Z. D., Fingas, M., Blenkinsopp, S., Sergy, G., Landriault, M., Sigouin, L., Foght, J., Semple, K., Westlake, D. W. S., Comparison of oil composition changes due to biodegradation and physical weathering in different oils, J. Chromatogr., 809, 89-107, 1998c.
  Wang, Z. D., Fingas, M., and Page, D., Oil spill identification, J. Chromatogr., 843, 369-411, 1999a.
  Wang, Z. D., Fingas, M., Shu, Y. Y., Sigouin, L., Landriault, M., Lambert, P., Quantitative characterization of PAHs in burn residue and soot samples and differentiation of pyrogenic PAHs from petrogenic PAHs - the 1994 Mobile burn study, Environ. Sci. Technol., 33, 3100-3109, 1999b. 
  Wang, Z. D., Fingas, M., Owens, E. H., Sigouin, L., and Brown, C. E., Long-term fate and persistence of the spilled Metula oil in a marine salt marsh environment: degradation of petroleum biomarkers, J. Chromatogr., 926, 275-190, 2001.
  Wang, Z. D., Fingas, M., and Sigouin, L., Using multiple criteria for fingerprinting unknown oil samples having very similar chemical composition, Environmental Forensics, 3: 251-262, 2002.
  Wang, Z. D., Hollebone, B. P., Fingas, M., Fieldhouse, B., and Weaver, J., Development of a physical and chemical property database for ten US EPA-selected oils, in: Proceedings of the 26th Arctic and Marine Oil Spill Program (AMOP) Technical Seminar, Environment Canada, Ottawa, pp 111-142, 2003.
  Wang, Z. D., Fingas, M., and Lambert, P., Characterization and identification of Detroit River mystery oil spill (2002), J. Chromatogr., 1038, 201-214, 2004.
  Wang, Z. D., Yang, C., Fingas, M., Hollebone, B., Hansen, A. S., Christensen, J. H., Characterization, weathering, and application of sesquiterpanes to source identification of spilled lighter petroleum products, Environ. Sci. Technol., 39, 8700-8707, 2005.
  Wang, Z. D. and  Fingas, M., Chapter 27: Oil and Petroleum Product Fingerprinting Analysis by Gas Chromatographic Techniques”, in Chromatographic Analysis of the Environment, 3rd edition, (ed. L. Nollet), CRC Press, New York, pp 1027-1101, 2006.
  Wang, Z. D. and Stout, S.A. Oil Spill Environmental Forensics: Fingerprinting and Source Identification, Academic Press, Boston, MA, 600 pages, 2007.
  Wang, Z. D., Yang, C., Fingas, M., Hollebone, B., Chapter 3: Petroleum Biomarker Fingerprinting, in Wang, Z. D. and Stout, S.A. (eds.), Oil Spill Environmental Forensics: Fingerprinting and Source Identification, Academic Press, Boston, MA, pp. 73-146, 2007
  Yang, C., Wang, Z. D., Hollebone, B.,  Peng, X., Brown, C., and Landriault, M., Characterization of bicyclic sesquiterpanes in crude oils and petroleum products, J. Chromatography A., 1216, 4475-4484, 2009.
  Zakaria, M. P., Horinouchi, A., Tsutsumi, S., Takada, H., Tanabe, S., and Ismail, A., Oil pollution in the Straits of Malacca, Malaysia: Application of molecular markers for source identification, Environ.  
  Zakaria, M. P., Okuda, T., and Takada, H., PAHs and hopanes in stranded tar-balls on the coast of Peninsular Malaysia: applications of biomarkers for identifying source of oil pollution, Mar. Pollut. Bull., 12, 1357-1366, 2001.

循环经济与绿色产业发展 更多>>
中国面临的主要环境问题及对策 更多>>