Anal. Chem. 1991, 63,1772-1780
Mass Profile Monitoring in Trace Analysis by Gas ChromatographyIMass Spectrometry H. Y. Tong,' D. E. Giblin, R.L. Lapp? S. J. Monson, and M. L. Gross* Midwest Center for Mass Spectrometry, Department of Chemistry, University of Nebraska, Lincoln, Nebraska 68588-0304
The advantages of gas chromatography/mass spectrometry (GCIMS) selocteblon monltorlng ( S I M ) in the mass profile (MP) mode at medlum mass resoivlng power were Investlgated for analyses requtrlng detection of bw-plcogram quantltles of analytes In complex mlxtures. The mass proflle modtoring provldes a certainty at lead 10 tlmes greater than that achleved by conventional GC/MS-SIM In the peak-top monltorlng mode, and lt can be operated at lower mass resohrlng power to compensate for the tom of sendtlvlty in the MP mode. The examlnatlon of mass proflie peak shape, central mass shffl, and sequentla1 changes durlng GC elullon not only reveals the presence of Interferlng compounds but also results In accurate mass measurement for those lnterferences. The latter feature takes the MP mode beyond the target mass analysts that GWMSSIM was m a l t y deslgned for. Thls addltlonal dimension of Information Is particularly useful for those complex and incompletely characterized matrices that are frequentty encountered In envkormental and blologlcal sample analyses.
INTRODUCTION Analysis of trace (part-per-trillion) organic compounds in environmental and biological samples has requirements different from many other analyses, such as for industrial products and processes. Often there is a need to measure one or more of a group of highly toxic compounds at extremely low levels in very complex mixtures. Analysis of polychlorodibenzo-p-dioxins and dibenzofurans (PCDD/Fs) in such samples is a typical example. Because of the chronic and acute health effects of some PCDD/Fs, even at extremely low levels (I,2), the generally required detection limits for these contaminants are parts-per-million (ppm) to parts-per-billion (ppb) for industrial wastes, low parts-per-trillion (ppt) for biological samples and low parts-per-quadrillion (ppq) for water samples. Such trace amounts of PCDD/Fs are often found in the presence of a large number of other organic compounds at much higher concentrations. For example, one of the principal sources of PCDD/F contamination is municipal waste incinerator (MWI) flyash, which contains hundreds, perhaps thousands, of organic compounds including other chlorinated aromatic compounds (3). Because of public apprehension, potential human toxicity, and consequential legal liability, both low detection limits and high certainty are primary analytical goals. Extensive clean-up is often the first step to reduce the possibility of interferences in biological and environmental sample analyses. In PCDD/F analysis, those clean-up procedures usually consist of pH extractions, followed by multicolumn chromatographicseparation (4) or high-performance Present address: CIBA-GEIGY Corp., Analytical Research, 444 Sawmill River Rd, Ardsley, NY 10502. 'Present address: Biosym Technologies, 10065 Barnes Canyon
Rd, San Diego, CA 92121.
liquid chromatography separation (5,6),to isolate PCDD/Fs from the bulk matrix and numerous coextracted compounds. The final identification and quantification are commonly performed by gas chromatography/mass spectrometry (GC/MS). Very strict criteria have been imposed for positive identifications of PCDD/Fs. The primary requirements include correct mass measurement, correct retention times obtained on a long capillary GC column under a slow temperature ramp, and proper isotope ratios of two or more characteristic ions. Although these clean-up and identification criteria are stringent, false positive identification of PCDD/Fs is still a major concern that has motivated many researchers to improve the analytical method to provide better assurance. Chemical ionization has additional discrimination capability and has been used to increase both specificity and sensitivity. In PCDD/F analysis, particularly with negative-ion chemical ionization (NICI), positional isomers can be distinguished by differences in the relative abundances of fragment ions (7,8). There are, however, some serious limitations for use of NICI techniques in complex mixture analyses. First, the PCDD/Fs are invariably accompanied by larger quantities of many other halogenated polycyclic aromatic compounds that also produce abundant negative ions under NICI conditions. Second, NICIMS response not only varies from isomer to isomer but also depends heavily on the MS operating conditions such as reagent gas pressure, source temperature, and impurity concentrations in the source. Such variation of response makes isomer quantification complicated. Gas chromatography coupled with tandem mass spectrometry (GC/MS/MS) can give high selectivity even in the presence of significant interferences. Increased specificity at the expense of overall sensitivity is obtained by two stages of mass selection, for example, one for molecular ions and one for [M - COCl]+ ions in PCDD/F analyses (9, IO). This technique, however, has high demands on the instrumental performance and operator skill when multiple compounds, such as tetra-through octa-CDD/Fs, need to be determined in the same run. The potential of GC/MS/MS in complex mixture analysis is great, but an improved sensitivity and method validation through comprehensive sample analyses are needed before it is accepted as a routine method for trace analyses of PCDD/F. Probably the most effective technique for achieving high certainty and a low detection limit in complex mixture analysis is GC/MS with electron ionization, a method approved by the US. Environmental Protection Agency (US. EPA) for PCDD/F analyses (4). Over the course of method development, validations were conducted in our laboratory and others (11,12).The GC/MS method is widely used because of its high reliability, precision, accuracy, sensitivity, and universal response toward organic compounds. This latter feature is important for multicomponent analysis. To measure routinely low-picogram quantities, GC/MS, NICIMS, and MS/MS are operated in the selective-ion monitoring (SIM)or selected-reaction monitoring mode. Most such GC/MS-SIM analyses are almost universally performed by monitoring the peak-top (PT) of a few characteristic ions @ 1991 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 63, NO. 17, SEPTEMBER 1, 1991
selected for the target compounds. This mode of operating not only limits the certainty of GC/MS in target compound analysis but also impairs the ability to diagnose the presence and, even more SO,the identity of interferences. The certainty of mass chromatogram data obtained from GC/MS depends highly on both the MS resolution and stability when peak-top monitoring is employed. For purposes of PCDD/F analysis, we use a mass resolving power of ca. 10OO0,which is required to separate interferences such as polychlorobiphenyl (PCB), polychloromethoxybiphenyl, and polychlorobenzylphenyl, from PCDD/F ions. In many cases, interfering compounds still appear on mass chromatograms even after sophisticated clean-up because the interfering compounds are very similar to the target analytes and/or are of higher concentrations. Under such circumstances, higher mass resolution is needed if further chromatographic separation is not feasible. A mass resolving power of greater than 20000, however, is difficult to use for analyses at ppt and ppq levels because the sensitivity is lower and the demands on instrumental performance are higher with respect to lower resolution MS. Furthermore, for target compound analysis in unfamiliar or extremely complex matrices, such as MWI flyash or soil from a chemical waste landfill, it is difficult to choose a proper mass resolving power in advance because the potential interferences are unknown. Seeking ultrahigh resolving power in the conventional PT mode is not the entire answer for high-quality data. We suggest the method should include additional analytical capability so that the peaks on GC/MS chromatograms can be further investigated but without MS/MS technique involvement. We will show here that a relatively simple modification of the peak-top monitoring to acquire instead mass peak profiles yields significant dividends. With this modified method, the mass profiles (MP) of a number of selected ions are obtained during GC elution. I t is this mass profile monitoring mode that we successfully used in various validation studies (11,12),including a recent round-robin test sponsored by the World Health Organization. In the latter study, the 2,3,7,&substituted isomers of PCDD/Fs in human specimens were determined at the ppt and ppq le