Investigative Analyses for Evaluating a
Petroleum BY JOE WIEGEL, COLUMBIA ANALYTICAL SERVICES Introduction This document consolidates our 3-part series as published in three consecutive issues of the Lab Link, published by Columbia Analytical Services, Inc. This series discusses investigative techniques, collectively referred to as forensic analyses, used to evaluate a petroleum release. The introduction provides some general information on methods and their applications. Part 2 covers the use of the techniques described for approximating the age of a petroleum release, along with difficulties encountered in attempting to age a hydrocarbon product after it has been released to the environment. Part 1: Summary of Forensic Techniques The term ÒforensicÓ generally implies some judicial or legal application. Forensic medicine, for example, describes a science that deals with the application of medical fact to legal problems. This conjures up memories of Quincy MD racing to beat the legal clock and save the innocent from wrongly serving time for a crime not committed. Or how about the O. J. Simpson trial: endless hours of boring scientific detail on DNA blood typing and disastrous foibles on the part of the LAPD medical examination. The crux of any forensic investigation must be based on scientific evidence. However, the interpretation of the results is almost always argumentative. The following is a summary of analytical techniques that may be used in building an argument. Techniques range from making discrete, precise measurements to complex and sometimes speculative models. Each has its own place in the investigation. When used in combination, it may be possible to develop a well formulated argument in support of a theory regarding the legal aspects of a petroleum release. Common forensic techniques are described in broad strokes. They have been discussed as categories of similar analyses. More specific details associated with any single technique are available from an analytical lab. Table 1 (see page 9) summarizes analytical techniques, provides method references, and lists some of the common applications of each analysis. Costs associated with these investigations can range from a few hundred to a few thousand dollars per sample. Costs tend to go up with the need to establish a more precise explanation of contamination at the site. Therefore, prior to submitting samples for analysis, it is recommended that some time be spent defining the scope of the investigation, performing a cost benefit analysis, and evaluating whether the site offers adequate opportunity for a reasonably successful investigation. A tiered approach is advisable, otherwise, there is the potential for generating costly, but non-usable information. Chemical Analyses
For example, gasoline containing organic lead was first marketed in 1923. The regulation of lead in gasoline has resulted in a well documented chronology of allowable lead concentrations. Furthermore, the chemical constituency of the lead additive has changed over time. Original lead additives were exclusively tetraethyl lead. The formulation changed to a mixture of tetramethyl lead, tetraethyl lead, and their cross products (in various combinations) from about 1960 through 1980. After 1980, tetraethyl lead was again the predominant organic lead species in gasoline. As a result, it is sometimes possible to establish a reasonably good estimate of the age of a gasoline release by examining the total lead levels in contaminated soil and groundwater, and individual organic lead species in the free product. Various metals have been added to gasoline, diesel and jet fuels for reasons ranging from anti-knock improvement (lead and manganese), combustion chamber deposit control (phosphorous, boron, nickel and zinc), and smoke control (barium).
Ethylene dichloride (EDC) and ethylene dibromide (EDB) were added to leaded gasoline as part of the antiknock mixture (motor mix) to prevent a build-up of lead oxides in the combustion chamber. Motor mix formulations have been designed so that during the combustion process, lead is scavenged by chloride and bromide ions to form more volatile and less corrosive compounds than lead oxides. As a result, EDC and EDB are referred to as lead scavengers. The proportion of EDC and EDB in the total lead additive mix has changed over the years. The mixtures have been formulated so that an excess of lead scavenger is present to assure that all lead is complexed to the corresponding lead halide during combustion. In more recent formulations, combined EDC and EDB levels accounted for up to 36% of the total motor mix added to leaded gasolines.
The use of alcohol as an anti-icing agent dates to as early as 1950. In the late 1970s, up to 10% ethanol mixtures gained favor as a way to extend the fuel supply during the fuel embargo. MTBE has been used on the east coast since the early 1980s, and on the west coast since the late 1980s. Oxygenates are extremely soluble in water and they tend to move freely with groundwater flow. Therefore, the presence of fuel oxygenates can provide information on gasoline formulation, a time frame for evaluating the release, and a mechanism to evaluate dispersion rates relative to other hydrocarbon components of the fuel.
Other Analysis Tools
Compound ratio analysis can be useful in helping to establish liability for a release. Furthermore, plots of these ratios in the form of star diagrams and bar charts can be very instructive in presenting the data to non-technical people.
Part 2: Aging a Hydrocarbon Release Typically, the most pressing reason to establish the approximate age of a fuel release to the environment is to help determine liability for the event. Investigations of these events are frequently subject to controversy due to a lack of good records. It is sometimes possible to develop a reasonable approximation of the age of a fuel release by understanding refining practices, site geology, hydrology, and chemistry. Costs can range from a few hundred dollars for a fingerprint of a fuel product to tens of thousands of dollars for a full-blown investigation involving several analytical techniques, extensive modeling, and expert witness support. Before starting an investigation, a cost-benefit analysis may prove to be one of the best uses of time and money spent on the project. All of the commonly used aging techniques are based on developing an understanding of the fuel composition at the site in relation to time. Several approaches have been devised to gather and evaluate this information. Linking specific components in the fuel to refining practices, which have changed dramatically over time, is probably one of the most common techniques. The presence of organic lead at a gasoline contaminated site indicates that the release occurred prior to the phase out of lead additives (late 1980s in most states, although leaded gasoline was used in sparsely populated areas into the 1990s). Allowable lead concentration has changed significantly over time. Between 1950 and 1980, concentrations of lead in gasoline averaged over 2 grams per gallon. This average concentration decreased steadily throughout the 1980s to about 0.5 gram per gallon by 1985, and to 0.1 gram per gallon by the late 1980s.1 Fuel oxygenates offer another indicator of a fuelÕs age. Oxygenates such as methyl tertiary butyl ether (MTBE) and alcohols began to be used with gasoline in the early to mid 1980s to meet regulatory standards imposed on the refining industry. Gasohol (gasoline with 10 percent alcohol) was prevalent during the 1972 fuel embargo in an effort to stretch the fuel supply. The presence of particular oxygenates in a free product gasoline and in contaminated soil and groundwater can be used as a time marker to estimate the release in combination with site history and other records. Establishing the presence or absence of specific marker compounds provides a gross approximation of the time of release. The next tier of investigative techniques focuses on narrowing the window of approximation. Typically, these techniques involve more detailed analysis and may attempt to link chemical composition to established weathering characteristics. With automotive gasoline, it is helpful to understand the historical trends associated with the lead additive mixture (motor mix). Motor mix formulations have changed considerably over time. The typical motor mix for leaded gasoline contained a mixture of antiknock agents (organic lead and organic manganese compounds), lead scavengers (ethylene dibromide (EDB) and ethylene dichloride (EDC), and small amounts of other ingredients (dyes, antioxidants, and stabilizers). Formulations have changed considerably over time. Tetraethyllead (TEL) was the primary antiknock agent used in motor mixes prior to 1960. Throughout the 1960s and 1970s, other anti-knock compounds gained acceptance. Tetramethyllead (TML), cross products of TEL and TML, and methylcyclo-pentadienyl manganese tricarbonyl (MMT) were used in as many as nine different formulas1. After 1980, tetraethyl lead again became the dominant antiknock agent used in various motor mix formulas. Changes have also occurred with respect to the lead scavenger (EDB and EDC) formulations in motor mix. In order to scavenge all of the lead oxide formed during fuel combustion, an excess of EDB and EDC was needed in the motor mix. A typical motor mix in the 1980s contained approximately 62% lead, 18% EDB, 18% EDC, and 2% inactive ingredient2. Economic factors in combination with combustion chamber chemistry changed the ratios of EDB and EDC over time. Since EDC is the less expensive of the two scavenger compounds, there was a tendency to increase its use over the more expensive EDB. However, since hydrochloric acid is an unwanted byproduct of EDC combustion, EDB continued to be used in motor mixes to prevent excessive exhaust valve wear. Over time, these factors influenced the ratios of the two scavenger compounds added to different motor mix formulations. Lead isotope analysis is another aging technique for leaded gasoline releases. This technique relies heavily on analytical testing in combination with historical trends and is based on a model that predicts the age of the organic lead added to the fuel. When tested on well documented fuel releases, the technique successfully predicted fuel ages within one to five years3. This level of accuracy is not typically achievable by other aging techniques. The degree of product weathering and changes in specific hydrocarbon classes provides another approach to aging a fuel release. Again, techniques vary in accuracy and sophistication. In general, light petroleum distillates, such as gasoline, weather by pathways of evaporation and water washing. This weathering is most evident by the preferential loss of C1 to C5 hydrocarbons (due mostly to volatilization) and the reduction of benzene and toluene relative to ethylbenzene and xylenes. Under certain conditions it is possible to approximate the age of a release based on the concentration ratios of these aromatic hydrocarbon groupings. For example, in silty soil saturated with gasoline the concentration ratio of benzene + toluene to ethylbenzene + xylenes will demonstrate about a 50% reduction over a 5 year period. Similarly, in groundwater, this ratio can be a useful gauge as to the freshness of the release. A concentration ratio of greater than 5, generally indicates a very recent spill. Over time, this ratio of benzene + toluene relative to ethylbenzene + xylenes will decrease exponentially due to greater solubility and subsequent transport away from the contamination2. Middle distillate and heavy residual fuel oils are weathered primarily by biodegradation processes. Since the n-alkane fraction represents the most easily biodegraded portion of heavier fuels, loss of these components is the most obvious result of the weathering process. The ability to estimate the age of a release involving middle distillate and residual fuel oils can be highly dependent on site specific parameters. However, one of the most useful techniques involves a ratio of n-C17 to pristane (the adjacent branched chain hydrocarbon). This ratio tends to decrease in a linear fashion with time as a result of the preferential removal of the n-alkane relative to the branched hydrocarbon chain. Other ratios that can be useful in estimating the degree of weathering include the n-C18 to phytane ratio and pristane to phytane ratio. In spite of the level of sophistication used to analyze some petroleum releases, site characteristics frequently make it impossible to pinpoint a time of release. Extreme conditions include saturated soil zones, floating product on the water table, and tightly packed, fine grain clay soils. In addition to these extreme conditions, site chemistry can be misinterpreted as a result of multiple and undocumented releases, chronic releases over extended periods of time resulting in considerable product mixing, and releases involving multiple grades and/or fuel products. To maximize the chances for a successful investigation, it is important to include a good review of the site history, preliminary investigations, a detailed study of how site geology and hydrology might be expected to influence the chemistry of the released fuel, and perhaps most importantly, the formation of specific goals for the investigation. Developing a systematic approach to the problem, complete with clearly defined decision points, is the best way to ensure a reasonable use of time and money on an investigation of this nature. Sources:
Part 3: Differentiation of Sources of Hydrocarbons at Contaminated Sites In this article we address two scenarios of hydrocarbon contamination. The first involves determining the primary source when a variety of sources may be responsible for the contamination. In this case, the age of the release is often critical. Since the previous issue dealt extensively with this topic, the current discussion is limited to using analytical techniques specifically geared towards differentiating between hydrocarbon sources. In some cases, the same techniques used to evaluate age of release are used in distinguishing the source of the product. These cases typically involve gasoline sites. A second scenario involves differentiating between petroleum and non-petroleum sources of hydrocarbon contamination. Samples in this category are usually high in organic content. For example, the presence of wood debris will result in false positives for diesel range organics when analyzed by a typical TPH-D method. Middle distillate and residual range hydrocarbons make up the majority of these investigations. In either scenario, a quick and inexpensive fingerprint or hydrocarbon scan using EPA Method 8015 Modified provides a good starting point. When interpreted by an experienced chromatographer the fingerprint can be used to evaluate choices for further analysis, and can help define the scope of the analytical plan. In some cases, the fingerprint analysis may be sufficient to identify the responsible party. More difficult cases involve distinguishing between sources of the same type of petroleum product (e.g., two different brands of gasoline). Scenario 1 The two primary techniques used in differentiating between fuel products are identification of specific chemical indicator compounds present in one source but absent in the other, and performing component ratio analyses on the contaminant and the various potential sources. The most common indicator compounds include fuel oxygenates, lead, and lead scavengers. These chemicals are all associated with gasoline releases. Other chemical indicators are less useful, mainly due to their unstable nature in the environment. Dye additives can be useful in differentiating between brands and grades of fresh gasoline dispensed from the pump. However, once exposed to the environment, the dyes break down and begin to change color. These alterations limit the usefulness of dye chemicals in forensic studies to recent releases. Component ratio analyses can be useful in differentiating between sources and in helping to identify the fuel type in a highly weathered sample. The application can be applied to both volatile and middle distillate fuels. A PIANO analysis (Paraffins, Isoparaffins, Aromatics, Naphthenes, and Olefins) is particularly useful for volatile fuel contamination. By comparing relative percentages of these classes of chemicals in environmental samples and dispensed fuels, it is sometimes possible to identify the source of contamination. A similar technique is to compare relative concentrations of individual aromatic compounds (BTEX). When plotted in the form of star diagrams these comparisons provide a good visual representation of the data. Comparing ratios of specific compounds allows interpretation on biodegration and water washing compositional changes. Another analysis that is used in differentiating sources of a fuel release involves ratios of stable isotopes of carbon and hydrogen. Because isotopic ratios do not change significantly as a sample undergoes weathering, it is possible to compare isotopic ratios in a sample to those of a dispensed fuel. If ratios in the contaminated sample strongly resemble those in the dispensed product, a case can be made that the source has been identified. Stable isotopes of lead can be used in a similar fashion to help differentiate between possible sources of leaded gasoline (although this differentiation is more dependent on time of release than comparisons to a dispensed fuel product). Scenario 2 Several techniques can help determine the significance of non-petroleum hydrocarbon contamination at a site. For middle distillate and residual ranges, extract cleanup options are available for the removal of polar hydrocarbons that can cause high biases to a TPH-D analysis. Typically, the most useful application is to analyze the extract before and after cleanup, then compare the analytical results. Significant reductions in TPH values following cleanup indicate positive bias due to interfering co-extractable organics. These cleanups, though not entirely effective in removing interferences, can be a relatively inexpensive way of evaluating high biases. Common cleanups include sulfuric acid digestion, saponification with sodium hydroxide, and separation using common adsorbents such as silica gel or alumina. New TPH methods developed in Washington and Oregon (i.e., NWTPH Methods) include options for these cleanup techniques, and the data is being accepted for regulatory review. Though sometimes useful, cleanup techniques are not entirely successful at removing biogenic interference as shown in Table 2. Even after sulfuric acid and silica gel cleanup, pure woodwaste exhibits a response between 300 and 1900 ppm. These clean-up steps only removed 74 to 86% of the biogenic interferences. Another relatively inexpensive technique for evaluating high biases at sites that contain wood debris is to compare analytical results obtained from standard TPH-D analyses to a Tannin and Lignin analysis. Samples that are mostly contaminated with petroleum hydrocarbon do not show good correlation with the tannin and lignin concentration. Samples that are primarily contaminated with wood debris tend to show good correlation between the two tests. To further characterize these sites, it is possible to oxidize the lignin in the sample and analyze for specific indicator chemicals by GC/MS SIM. The lignin degradation products can also be analyzed directly from water samples. These results can be used to evaluate the type of wood debris present at the site. GC/MS techniques offer great flexibility in identifying petroleum and non-petroleum hydrocarbon compounds that can be used as specific markers in evaluating a site. Analyses for petroleum related alkylated benzenes, polycyclic aromatic hydrocarbons (PAHs) and their alkylated homologs, common wood related terpenes, and biomarkers such as terpanes, sterols, and steranes each have a place in fully characterizing contaminated samples. These analyses can provide detailed information regarding the molecular species currently inhabiting a contaminated site. Some of these compounds, such as the PAHs, are common to middle distillate and residual fuels. Based on ratios of individual PAH compounds, the sampleÕs homolog fingerprint, and known concentrations of PAHs in fresh petroleum product, it is possible to estimate the amount of fuel responsible for the total hydrocarbon contamination in a sample. Wood waste marker compounds can be used to establish a definitive link to non-petroleum hydrocarbon contamination. Bio-markers provide information regarding the amount of weathering a fuel has undergone in the environment. All of this information can be useful in establishing links to the suspected sources of contamination at a site. How a Lab Supports Forensic Studies We appreciate the favorable feedback and support our readers have provided throughout the series. A question repeatedly asked is ÒWhat support can a lab provide during a forensic investigation?Ó An analytical chemistry laboratory can perform the analytical procedures described throughout this series. A lab can coordinate and manage outside testing required. In addition, much of the work associated with a forensic study involves interpreting the analytical data, presenting the results in a format that strengthens the impact of the findings, and expert witness testimony. Labs are best utilized when it also assists with the up-front preparation of a cost effective analytical plan. A lab can perform the initial investigative analytical support, evaluate options for further testing, and be utilized as a resource for technical litigation support and expert testimony, if necessary. A labs electronic information capabilities allow the data to be presented in a variety of formats to meet your specific application. This can be especially important for litigation purposes. |