COMPOUNDS BY GAS CHROMATOGRAPHY–MASS SPECTROMETRY
In recent years, environmental analysis of petroleum and its products as adulter- ants in air, water, or soil has become increasingly important for regulatory, health,
Petroleum Industry Applications 75
and legal reasons. Because these matrices are so complicated it is not possible to obtain the necessary compositional profile by physical or spectroscopic means (such as the ‘‘total petroleum hydrocarbon’’ analyses performed with IR or GC on soil), GC–MS has been the analysis method of choice for these sorts of studies.
Gas chromatography–mass spectrometry analyses of environmental samples are almost always target compound analyses, where the analyst needs to quantify one or many petroleum product species. While this analysis may seem simple, accurate quantitation in these samples can be exceeding difficult. Once an ade- quate individual species compositional profile has been determined for a sample, the data manipulation to determine the actual product identity and mixture analy- sis can get quite complicated due to the constant background of non-point-source petroleum products in any given vicinity.
5.1. ‘‘Modified 8270 Analysis’’
The widely applied GC–MS method for complex samples was written many years ago as part of the EPA SW-846 protocol for analysis of solid waste [58]. In this protocol, an array of extraction and cleanup procedures (EPA methods 3500s and 3600s) are specified for the separation of nonpolar target analytes from polluted water or solid waste, where the actual cleanup method to be applied is dependent on the complexity of the sample.
The most common procedure for soil or water is ultrasonic extraction of the analytes into methylene chloride following careful quantitative addition of stable isotope–labeled internal standard surrogate species to determine the recov- ery and efficacy of the extraction. This solid waste procedure is commonly ap- plied to soil, sludge, body tissues, plant matter, and even polar solid petroleum products such as asphalt. SW-846 then specifies the use of GC–MS (EPA method 8270) to quantify for a list of about 80 enumerated target analytes, which include a wide range of common industrial pollutants and pesticides. This list does in- clude a few compounds present in petroleum products, but the 8270 methodology is in practice able to determine the amount of any nonpolar GC–MS analyte.
Thus, the general analysis of environmental samples for species present in petro- leum products but not specifically listed in the EPA method is often referred to by the catchall term ‘‘modified 8270 analysis.’’ This method is sufficiently rigor- ous in its use of internal standards and surrogates. This is in contrast to other methods, which use external or non–isotopically labeled standards and do not take into account the difference in compound affinities for the soil or water in question, thus leading to possible misleading results. Other advantages of the
‘‘8270’’ procedure are: (1) it is run by numerous contract laboratories, which can do the extractions and/or GC–MS injections cheaply in an automated manner and provide the petroleum analyst with the raw data files for further data reduc- tion, and (2) it allows for the use of extra cleanup steps or selectivity methodolo-
76 Hsu and Drinkwater
gies (such as pre-LC or MS–MS) as long as the standards and controls come out with the right answers.
The modified 8270 method is used extensively for the analysis of poly- nuclear aromatic hydrocarbons (PNAs or PAHs). Soot is believed to be the first known human carcinogen, and sootlike molecules,nonalkylatedPAHs, are gener- ally recognized as some of the most chronically toxic hydrocarbons. Eighteen nonalkylated PAHs are on the EPA priority pollutant list, and are included in the classical 8270 list. While petroleum products can contain upwards of 50 mass
% alkylated PAHs, it is very rare for a petroleum product to contain any signifi- cant amounts of nonalkylated PAHs that come primarily from combustion pro- cesses (pyrogenic rather than petrogenic). Samples from the environment, how- ever, often contain nonalkylated PAHs, which derive from used motor oil runoff on roadways, vehicle and power plant particulate (often coated with nonalkylated PAHs), or simple soot from any of a number of sources. Fortunately, these non- alkylated PAHs are easy to analyze by GC–MS since they are easy to ionize with little fragmentation. A high-temperature GC column must be used in some instances to elute the high-boiling species.
5.2. Fugitive Compound Analysis
The most complicated problems for the environmental petroleum analyst are those products, either crude oil or refined and additized products, that somewhere in transport or storage get released unintentionally to the environment. These
‘‘spills’’ are referred to as fugitive products, and can often be characterized by modified 8270 procedures to get the distribution of major petrogenic species.
These species distributions are then compared with the distribution of those spe- cies in various products to determine the age, source, or concentration of a given product in the sample. This can be a very complicated procedure, since, as men- tioned above, hydrocarbons are present everywhere in the environment.
Figure 9 shows the ‘‘fingerprint’’ of the C-3 dibenzothiophenes and C-3 phenanthrenes in a possible fugitive product. These fingerprints are very different among different sources, and are a powerful tool to determine the identity of a fraction in a mixture attributable to a given product. Perhaps the most famous example of this was when these same compound fingerprints shown in Figure 9 were used in analysis of Prince William Sound sediments to show that in some areas a large fraction of the oil was from petrogenic hydrocarbons from seeps and not fugitive product at all [59]. For urban samples, where hydrocarbon content is always a mixture of ambient pollution from nonpoint sources, it is usually neces- sary to do a quantitative analysis, where 10 to 100 compounds are measured in the various possible sources and then compared statistically with the distribution measured in the sample. In this case, the various possible concentration vectors are fed into a multiple least squares analysis to determine the loading scores of
Petroleum Industry Applications 77
Figure 9 Gas chromatography–mass spectrometry TIC of possible source material for a fugitive product. Inserts show the ‘‘fingerprint’’ of C-3 phenanthrenes and C-3 dibenzo- thiophenes, respectively. These profiles are very different from one petroleum product to another and their profile and/or total abundance can be used for quantitative identification of fugitive product mixtures. Conditions: Varian 3400 gas chromatograph (Varian, Palo Alto, CA) with Finnigan (San Jose, CA) SSQ170 EI mass spectrometer running a tempera- ture gradient on a 100 m⫻0.25 mm ID DB-5MS (0.25-àm film thickness) fused silica capillary GC column (special order from J&W Scientific, Folsom, CA).
78 Hsu and Drinkwater
each possible contributing hydrocarbon mixture, as shown in Table 1. While the errors of individual component quantification are high, multiple linear regression analysis reports the scores as 83% A and 17% B, very close to the actual mixture composition given the high difference in boiling points between these two prod- ucts [60].
Much recent attention has been cast on one particular additive to gasoline, methyltert-butyl ether (MTBE). Most polymer and detergent additives in addit- ized fugitive products have a high affinity for the polar soil surfaces and stay at the spill site. Methyltert-butyl ether is unusual because when present in a fugitive gasoline product (almost all RFG contains MTBE due to its low vapor pressure
Table 1 Relative Concentrations of Various Components in a Diesel Fuel (A), a Gasoline (B), and a Synthetic Mixture of 87% A and 13% B, as Determined by Semiquantitative GC–MS analysisa
Mass Diesel Gasoline Mixture
Species measured used (ppm) (ppm) 83% A/17% B
Naphthalene 128 189 3119 694
2-Methyl naphthalene 142 504 2347 876
1-Methyl naphthalene 142 348 955 481
Naphthalene C2 156 4550 1337 4285
Naphthalene C3 170 10433 386 9981
Naphthalene C4 184 5748 112 8215
Naphthalene C5 198 2471 30 3901
Naphthalene C6 212 1008 21 1612
Dibenzothio-C0 184 33 0 25
Dibenzothio-C1 198 210 2 199
Dibenzothio-C2 212 318 1 335
Phenanthrene 178 144 14 120
Anthracene 178 0 3 1
Phen⫹anth-C1 192 559 54 515
Phen⫹anth-C2 206 465 50 545
Phen⫹anth-C3 220 188 16 296
Pyrene-C0 202 13 3 10
Pyrene-C1 216 25 10 31
Pyrene-C2 230 25 10 34
Chrysene-C0 228 0 1 1
Acenaphthene 154 44 6 48
aWhile the errors of individual component quantification are high, multiple linear regression analysis reports the scores as 83% A and 17% B, very close to the actual mixture composition given the high difference in boiling points between these two products. Regression results: 87% A/13% B r2⫽0.9479 (column normalized).
Petroleum Industry Applications 79
and high octane rating and oxygen content) it moves underground in the water table ahead of the gasoline, giving a noticeable taste and odor to the water before the gasoline arrives. Many underground gasoline spills went unnoticed for years while the groundwater polluted by gasoline remains tasteless and odorless until the addition of odoriferous MTBE caused the public attention to this serious threat to drinking water quality. Methyltert-butyl ether, along with acetone, is one of the few common compounds for which the GC–MS analyst cannot use the EPA methylene chloride extraction methods, such as 3510B or 3520B, due to the poor recoveries of these components. These materials may be analyzed by headspace GC–MS, assuming an isotopically labeled internal standard is used to account for sample differences in vapor pressure due to fuel and other cosol- vents in the aqueous matrix.
5.3. Diesel Emissions
The combustion of diesel fuels produces mainly carbon dioxide and water. The emitted exhausts also contain small amounts of carbon monoxide, nitrogen oxides (NOx, including NO, NO2, N2O, etc.), sulfur oxides (SO2/SO3), unburned hydro- carbons, and particulates. These minor components are pollutants and legally regulated. When exhaust gas is cooled to collect the condensates, the organic fractions are largely condensed on the particulate soots [61,62].
Some diesel particulates are of respirable size, and the organic species attached to them may constitute an inhalation health hazard to the human popula- tion [63,64]. Organic compounds on diesel particulates have been extracted and studied [62,65]. In the soluble organic fractions (SOFs), much attention has been paid to polycyclic aromatic hydrocarbons (PAHs), nitrated PAHs, and oxygen- ated PAHs, particularly for those showing mutagenic activities.
Uncombusted PAH from diesel fuels can react with oxides of nitrogen in the exhaust to form nitrated PAH. Some of them, for example, 1-nitropyrene and dinitropyrenes, have been shown to be potent mutagens in Ames assays [64].
Due to this environmental health concern, nitrated PAHs have been studied by various GC and MS techniques [66,67]. Oxygenated PAHs found in diesel partic- ulates include polycyclic ketones and quinones, carboxaldehydes, and hydroxy- PAH [68].