Trace Organic Analysis
Environmental Screening of Acidic Compounds Based on CZE/LIF Detection with GC/MS and GC/HRMS Identifications
W. C. Brumley,1 A. H. Grange,1
V. Kelliher,1a D. B.
Patterson,1a A. Montcalm,1a
J. Glassman,2b and J. W. Farley2
1U.S. Environmental Protection Agency, National Exposure Research Laboratory, Environmental Sciences Division, P.O. Box 93478, Las Vegas, NV 89193-3478
2University of Nevada-Las Vegas, Department of Physics, P.O. Box 454002, Las Vegas, NV 89154
a Enrollee in the Senior Environmental Employment Program (SEEP), assisting the U.S., Environmental Protection Agency under a cooperative agreement with the National Association of Hispanic Elderly (NAHEP).
b Work performed while holding a postdoctoral position. Present address: Department of Physics/220, University of Nevada-Reno, Reno, NV 89557.
W. C. Brumley, A. H. Grange, V. Kelliher, D. B. Patterson, A. Montcalm, J. Glassman, and J. W. Farley, "Environmental Screening of Acidic Compounds Based on Capillary Zone Electrophoresis/Laser-Induced Fluorescence Detection with Identification by Gas Chromatography/Mass Spectrometry and Gas Chromatography/High-Resolution Mass Spectrometry," JAOAC International, 83, 1059-1067 (2000)
This work presents the application of capillary zone electrophoresis/laser-induced fluorescence (CZE/LIF) to the discovery of acidic compounds in environmental matrices or the screening of extracts for acidic components. Based on published studies, coal-derived materials should contain a significant fraction of acidic compounds relative to materials derived from petroleum and shales. Such compounds may be useful as marker compounds for site assessment and source apportionment issues, and their identification may be important in toxicological and other health issues. We have used deep UV light from the frequency doubled Ar ion laser at 244 nm and 257 nm to study extracts of samples. The CZE/LIF technique possesses good sensitivity and therefore overcomes one of the limitations of CZE with UV detection. The present example depends on high pressure/temperature solvent extraction of polynuclear aromatic hydrocarbon-contaminated soil followed by separation using CZE. The anionic analytes were separated using borate or phosphate buffer (pH 9.2 to 12.3) after a chemical class separation. Samples were also characterized by GC/MS using full scans at low resolution, and elemental compositions were determined unequivocally with GC/high resolution MS (GC/HRMS) using mass peak profiling. The similarity of low resolution EI mass spectra for a standard, 1-hydroxypyrene, and for a series of compounds in a contaminated-soil extract suggested that several types of phenolic and hydroxy-PNAs were present including hydroxylated derivatives of fluorenes, fluoranthenes, and pyrenes. GC/HRMS using mass peak profiling confirmed the elemental compositions of the hydroxyfluorenes and hydroxypyrenes (and presumably hydroxyfluoranthenes) as [C13H10O] and [C16H10O], respectively. A new version of the mass peak profile software was written for the Finnigan-MAT 900S-Trap similar to that developed previously for the VG 250SE. Inclusion of a calibration ion in addition to a lock mass ion in the MID descriptor provided errors of less than 1 ppm for the three partial profiles of the analytes. A mass resolution of 31,000 was used to resolve the analyte signals from interferences evident in the full M+1 and M+2 profiles in the case of the hydroxyfluorenes. Derivatization was also carried out to form the tert-butyldimethylsilyl derivatives of phenolic hydroxy groups as a further confirmation of structure.
Ionic analytes are often conveniently separable by free zone capillary electrophoresis (CZE) (1) in addition to the more traditional approaches using high performance liquid chromatography or ion chromatography (2, 3). One of the limitations of capillary format separations such as CZE with UV detection in environmental analysis has been the lack of a low concentration detection limit (4). Thus, the technique cannot function routinely as a quasi-universal screening or determinative technique at the µg/kg level or lower in the sample or its extract. Such levels are achievable by techniques such as HPLC/diode array detection and GC/MS.
An alternative approach is to use laser-induced fluorescence (LIF) detection, which can achieve low concentration detection limits, although the detection is limited to compounds that fluoresce.5 Currently, a variety of lasers are being used spanning the visible [HeNe (e.g., 595 and 633 nm) and Ar/Kr ion lasers (e.g., 488.0, 514.5, 568.2, and 647.1 nm)], near UV (325 nm and 354 nm) [HeCd lasers], and deep UV (244 to 284 nm) [frequency-doubled ion lasers] regions of the spectrum for LIF detection at concentration detection limits from 10-7 M to 10-13 M. The CZE/LIF technique is therefore an excellent screening technique for fluorescent substances provided confirmation of identity can be obtained using a complementary technique such as mass spectrometry. Reviews of the applications of capillary electrophoresis techniques to environmental analysis have summarized the current state of the field (6, 7). Within the EPA National Exposure Research Laboratory's Environmental Chemistry Branch, we are developing liquid separation techniques for various types of environmental analytes to be coupled with highly sensitive detection techniques such as LIF and mass spectrometry (MS). Among the target compounds of interest are various herbicides and some semivolatile compounds that are not amenable to gas chromatographic separations or that may exist as ions under certain conditions. Hydroxylated polynuclear aromatic hydrocarbons (hydroxy-PNAs) belong to this group of compounds and are of interest in atmospheric particulate analysis and metabolic studies of the PNAs (6, 7).
The ability to extract samples relatively rapidly and with reduced amounts of solvents provides a convenient source of extracts for investigation. Examples of such techniques include supercritical fluid extraction (SFE), Accelerated Solvent Extraction (ASE®), and Microwave Assisted Process (MAP®) for all of which there is a robust literature (8-10). Aqueous samples may be extracted using solid phase extraction (SPE) and solid phase microextraction (SPME) techniques11 as well as by traditional liquid/liquid extraction (12). The selection of extraction technique, conditions, and solvent can greatly affect the results of analysis and must be carefully considered in the course of the analysis.
The identification of compounds discovered by most screening techniques in environmental analysis is usually the province of mass spectrometry (13). Among the techniques of choice are GC/MS, LC/MS, and CE/MS. Identifications are aided by the determination of the exact mass of the molecular ion or other ions representative of the substance, which leads to the elemental composition of the molecular mass. Recently, we have developed a technique for doing this at high resolution on-the-fly using a procedure called mass peak profiling using selected ion recording data (MPPSIRD) (14, 15).
In this work, a series of compounds in an extract from an environmental contamination problem was revealed by CZE /LIF and the electrophoretic behavior suggested that several types of acidic components were present, including the possibility of phenols and hydroxylated polynuclear aromatic hydrocarbons (hydroxy-PNAs). High resolution MS using mass peak profiling was used to confirm the elemental compositions of representative compounds, and a new version of the mass peak profile software was written for the Finnigan-MAT 900S-Trap.
|Acetone||(For HPLC) Burdick and Jackson (Muskegon, MI)|
|Methanol||(HPLC Grade) Burdick and Jackson (Muskegon, MI)|
|Methylene chloride||(Analyzed) J. T. Baker (Phillipsburg, NJ)|
|Toluene||(Analyzed) J. T. Baker (Phillipsburg, NJ)|
|Sodium hydroxide||(97% ACS) Aldrich Chemical Co. (Milwaukee, WI)|
|Hydrochloric acid||(ACS) Mallinckrodt Baker Inc. (Paris, KY)|
|9-hydroxyfluorene||(96%) Aldrich Chemical Co. (Milwaukee, WI)|
|2-hydroxyfluorene||(99%) Aldrich Chemical Co. (Milwaukee, WI)|
|1-hydroxypyrene||(98%) Aldrich Chemical Co. (Milwaukee, WI)|
|9-phenanthrol (tech)||Aldrich Chemical Co. (Milwaukee, WI)|
|(99%) Aldrich Chemical Co. (Milwaukee, WI)|
|(97%, 1% N-tertButyldimethylsilylchloride) Aldrich Chemical Co. (Milwaukee, WI)|
|Sodium tetraborate decahydrate||(99.5%, ACS) Aldrich Chemical Co. (Milwaukee, WI)|
|Sodium phosphate dodecahydrate||(98%, ACS) Aldrich Chemical Co. (Milwaukee, WI)|
|3N HCl||25.8 mL of conc HCl was dissolved in water and diluted to mark in 100 mL volumetric flask|
|3N NaOH||12.6 g of NaOH was dissolved in water and diluted to 100 mL|
|Borate buffer||0.3814 g of sodium tetraborate decahydrate was dissolved in 100 mL of water|
|Phosphate buffer||0.3801 g of (tri)sodium phosphate dodecahydrate was dissolved in 100 mL of water|
|Borate/Phosphate buffer||equal volume of the borate and phosphate buffers were mixed and the resulting buffer was approximately pH 10.5|
Samples were obtained from a Southwestern area of the United States where PNA contamination had been detected. Sampling and standard methods of extraction and analysis revealed that the source of the PNAs was likely clay pigeon shards and shards that became pulverized due to ground traffic. These samples became the source of the present investigation into novel compounds present in the matrix.
Soil samples were extracted by (ASE®) Dionix (Salt Lake City, UT) using an azeotropic solvent mixture consisting of chloroform/methanol/acetone in the w/w/w of 47/23/30 corresponding to 31.50/29.08/37.93 v/v/v (16). The extraction time was 5 min after an equilibration time of 7 min to reach 150C; pressure was 2500 psi. The resulting sample volume was approximately 15 mL, which was concentrated to 1 mL for subsequent work or taken to dryness and redissolved in 1 mL methylene chloride with a few drops of methanol to aid dissolution. The resulting sample extract was then extracted three times into 3 N NaOH using 1-mL volumes. This basic solution was acidified with HCl to pH 1, and then back-extracted with methylene chloride, evaporated to dryness, and taken up in methanol. The resulting solution was expected to be enriched in acidic components although other water-soluble components would be coextractives. Solutions to be used for CZE/LIF detection were diluted with methanol and filtered through 0.2 µm filters.
Basic components in the methylene chloride solution were extracted three times into 3N HCl using 1- mL volumes. The acidic solution was made basic with 3 N NaOH (approximately 4 mL) and back-extracted three times using 1-mL volumes methylene chloride. The methylene chloride solution was taken to dryness and redissolved in methanol. This solution was expected to be enriched in basic components. Compounds in this fraction were identified as nitrogen-containing aromatic hydrocarbons (e.g., acridine) and were similar to those reported in previous work (13) and will not be further discussed here (5).
All samples, buffers, and standards were filtered through 0.2-µm filters. Buffers and rinse solutions were prepared from deionized 18 M water (Millipore quality). The optical bench has been described (17). Briefly, this laboratory-built instrument consists of the laser (Coherent, CA) two high reflective mirrors, a focusing objective (fused silica optics), a capillary holder, a focusing lens for the fluorescent light at 90 deg to the beam, a slit to reduce stray fluorescence before the photomultiplier, a current amplifier, and a data system. The CE experiment requires a high voltage supply as well. For excitation at 244 and 257 nm using 5 mW power, two optical filters were used in series (390DF70, Omega Optical, Battleboro, MA) as the bandpass filters for fluorescent light. Capillaries were prepared from bulk fused silica obtained from PolyMicro Technologies (Phoenix, AZ). Capillaries were rinsed before each days work with 5 min pressurized rinses of 0.1 N NaOH, DI water, and running buffer. Between each run the capillary was rinsed for 2 min with each of the rinsing solutions. Injections were by gravity for 3 to 20 sec at 25 cm. All reported work used capillaries of 0.075 mm ID with 18 kV as the separation voltage.
A Fisons VG 250SE was used for the low resolution EI spectra and for MPP based on procedures developed previously (14). The 5890 Series II GC (Hewlett-Packard) was fitted with a 25-m x 0.25-mm ID column (DB5) with 0.25-µm film thickness (J&W, Fullerton, CA). The column was temperature programmed from 60 to 300C at a rate of 10C/min. The injection was on-column at 60C using a Hewlett-Packard 7673A autosampler. The flow rate of He was 20 cm/min at 60C.
A Hewlett-Packard GC/MSD (6890 GC and 5973 MSD) was also used for low resolution mass spectrometry of the tert-butyldimethylsilyl derivatives of hydroxy-PNAs. Injection was by pulsed pressure splitless injection at 250C and temperature programmed at 15C/min from 60 to 300C at a He flow rate of 1 mL/min. A mass range of m/z 50 to 500 was scanned at about 2 Hz.
Hydroxy-PNAs were reacted with 100 µL of N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (from 1-mL ampules containing 1% tert-butyldimethylchlorosilane) in 1 mL of toluene at 65C for 1 hour (18). The reaction mixture was injected without further cleanup. The reaction was quantitative as determined by the absence of any underivatized analyte.
MPP with the Finnigan-MAT 900S-Trap
A new set of procedures was written for the Finnigan-MAT 900S-Trap instrument based on the suite of processing commands developed for the VG system and previously described (14). Certain differences in approach were made based on the presence of two data systems used for this instrument: the Digital/Compaq Alpha workstation for the double focusing instrument and a PC-based system (Windows NT 4.0) for the trap (LCQ type software). These procedures are described in more detail elsewhere (14,15,19). Briefly, the MPP technique acquires mass peak data at high resolution using the selected ion recording (SIR) descriptor (also known as SIM and MID). The descriptor allows the profiling of the mass peak by incrementing the masses being monitored by very small steps from a starting mass or about a central mass. All of this takes place rapidly enough to allow adequate sampling of the chromatographic peak during elution of the compound from the column. Thus, the experiment yields the typical data of high resolution SIR (albeit over a very small mass interval) and data processing produces the mass peak profile or shape and determines its mass centroid. The advantages of this approach are its high sensitivity, its accuracy, and its ability to follow the chromatography (i.e., it can handle narrow capillary chromatographic peaks). The experiment usually allows an unequivocal assignment of elemental composition for the ion from the accurate masses and relative abundances.
RESULTS AND DISCUSSION
Phenols and hydroxy-PNAs migrate as anions when the pH is sufficiently basic (pH 8 to 10). The alkylphenols tend to exhibit relatively small negative mobilities resulting in migration times slightly longer than the electroosmotic (EO) flow disturbance. For practical analyses this results in less ability to detect their responses versus an elevated background and EO flow disturbance where neutral molecules elute. Hydroxy-PNAs exhibited greater negative mobilities and therefore were effectively detected in sample extracts. Table 1 provides migration times for several classes of hydroxy-PNAs and phenols. The migration times have been corrected to an arbitrary migration time of the internal standard, 7-hydroxycoumarin-4-acetic acid. As can be seen from the migration time data, the alkylphenols tend to migrate near the EO flow injection disturbance as noted and are best separated by ensuring adequate ionization with phosphate or borate/phosphate buffer (pKa around 10) and excitation at 244 nm (284 would be better if available). Hydroxy-PNAs can be excited at 257 nm and separated with borate buffer (pKa around 8). We observed a decreasing negative mobility as the alkyl group became larger with alkylphenols. In contrast, as the number of rings increased so did the negative mobility of hydroxy-PNAs. It is immediately clear that while alkaline free zone separation conditions have limited selectivity with regard to alkylphenols, they are more useful for separating the hydroxy-PNAs. While coating the column with a positively charged surfactant might be expected to improve the resolution of the alkylphenols, background problems arising from the introduction of surfactants, coatings, and other coadditives are a greater problem when using the deep UV laser (20). More importantly, for our purposes the positive response for anion analytes was sufficient information from this technique to suggest that the sample/extract was worthy of further investigation and that acidic components could be lost in routine "target analyte" approaches. We would like to emphasize that ion mobilities that form the basis of free zone separations possess moderate selectivity and for complex samples may not offer enough selectivity to effect complete separations. However, all other techniques that introduce an additional or alternative separation mechanism will not offer the clear indication of anionic components since components of various charge states may be intermixed. For example, we have studied the liquid chromatography of PNAs and hydroxy-PNAs on porous graphitic carbon and on C18 derivatized silica under reverse phase conditions (not reported here). HPLC using C18-derivatized silica columns appears entirely satisfactory for these compounds, but does not differentiate which compounds are acidic in a complex mixture containing PNAs and a host of other compounds.
Table 1. Migration times of various phenols and hydroxy-PNAs corrected to a common migration time of 7-hydroxycoumarin-4-acetic acid (70-cm capillary, 62 cm to detector, 0.075-mm ID, 10 mM borate/phosphate buffer, pH 12.3 resulting in an EO flow time of 6.00 min).
Corrected Migration Time
|nonylphenol||multiple peaks near EOF*|
* Nonylphenol is a technical material consisting of more than 10 isomers of alkylated phenols
The reproducibility of migration times under the reported conditions were typical for us when using free zone conditions (correctable to within 0.3% using an internal standard). Although this variation is greater than found in capillary GC and HPLC, for our purposes here it is much better than what is needed to indicate potential analytes. Simply put, the observation of peak responses to the right of the EO flow peak clearly indicates anionic (fluorescent) substances.
Various environmental samples were subjected to high pressure/temperature extraction and were further fractionated into enriched acidic and basic components through either acidic or basic extraction into an organic solvent. The fraction enriched in organic acids was subjected to CZE with either borate or phosphate buffer solutions using deep UV excitation at 244 nm or 257 nm. Although CZE is capable of resolving cations, neutrals, and anions at once, the removal of classes of coextractives in a complex extract reduces the response due to water-soluble neutrals at the EO flow and ameliorates some of the demands placed on the separation.
Figures 1 and 2 illustrate the typical complexity of the anionic substances in the contaminated soil extract.
Figure 1. CZE/LIF electropherogram for extract at 257 nm (5 mW power) under conditions of 10 mM borate/phosphate at pH 10.53; 58-cm capillary, 50 cm to detector; 25°C, gravity injection of 20 s at 15 cm height; 18 kV separation voltage (estimated 10 to 100 pg per component on-column).
Figure 2. CZE/LIF electropherogram for extract at 244 nm (5 mW) power under conditions of 10 mM borate/phosphate at pH 10.53; 70-cm capillary, 62 cm to detector; 25°C, gravity injection of 20 s at 15 cm height; 18 kV separation voltage (estimated 10 to 100 pg per component on-column).
A great variety of column lengths, coadditives, voltages, and buffer variations were tried over the course of the investigations. These two figures represent a brief comparison of both wavelength and migration time effects of different conditions. Based on migration times, phenols and hydroxy-PNAs were suspected analytes together with carboxylic acids. Detection limits of representative standards ranged from 10-7 to 10-8 M depending on analyte as determined at a signal to noise ratio of 3 to 1 for serial dilution studies. For a 2 to 10 g sample this would translate into a 1 to 10 µg/kg level in the matrix or lower if the extract were concentrated below 1 mL. Lower detection limits may be possible at different excitation wavelengths (e.g., 284 nm from frequency-doubled Kr ion laser) and with optimized components to more efficiently capture near-UV fluorescent light (350 nm) than the presently used optics. To our knowledge, this is one of the few reported uses of CZE/LIF using the deep UV laser for the discovery of new components in an environmental matrix such as contaminated soil. The usefulness of this approach is based on its sensitivity to compounds likely to be of toxicological interest (PNAs) and the clear indication of their anionic behavior under basic pH conditions. We estimate that the hydroxy-PNAs are present at levels approximately 100 times lower than the parent PNAs.
The low resolution GC/MS EI spectra were obtained and compared with those of hydroxy-PNA standards. Typically, 1 to 10 ng of substance is sufficient for a good EI spectra. In these mixtures compound concentrations are estimated to range from greater than 10 ng/uL to around100 ng/uL in the extract being injected. The similarity of these spectra (Figures 3A and 3B) to phenols and hydroxy-PNAs was apparent (i.e., the relative abundance of molecular ions, loss of 29 u, and appropriate molecular masses).
A. EI mass spectrum of standard 1-hydroxypyrene.
B. EI mass spectrum of component in soil extract. Conditions: 70 eV; source 250°C; GC 60 to 300°C at 15°C/min; column DB5, 25 m X 0.25-mm ID, 0.25-µm film.
The similarity of low resolution EI mass spectra for a standard, 1-hydroxypyrene, and for a series of compounds in an environmental extract suggested that several types of phenolic and hydroxy-PNAs were present including hydroxylated derivatives of fluorenes, fluoranthenes, and pyrenes. These hypotheses are borne out by an examination of the ion chromatograms from the isolate containing acidic components. The total ion chromatogram (Figure 4) reveals a major peak corresponding to the component whose spectrum was given in Figure 3B.
Figure 4. Total ion chromatogram of "acids" fraction with indication of major peak corresponding to the spectrum and conditions of Figure 3.
Figure 5 reveals additional components at m/z 218 corresponding to other isomers and an additional cluster of compounds which could be hydroxyfluoranthenes or pyrenes.
Figure 5. Ion chromatogram of m/z 218 corresponding to hydroxypyrenes and fluoranthenes; conditions as in Figure 3.
All components were verified as containing the elemental composition C16H10O. The ion chromatogram for m/z 182 (Figure 6) also revealed compounds corresponding to hydroxyfluorenes.
Figure 6. Low resolution ion chromatograms of m/z 182 from the "acids" fraction; conditions as in Figure 3.
High resolution MS using MPP confirmed the elemental compositions of the hydroxyfluorenes and hydroxypyrenes as [C13H10O] and [C16H10O], respectively (see Figures 3-6 for examples). In the representative total ion chromatogram of Figure 4, the peak at scan 541 (RT = 12:37) corresponds to a major component which appears chemically similar to 1-hydroxypyrene which elutes at scan 558 (RT=13:16). A detail of this region shows that there are several components with molecular ions at m/z 218 (e.g., cluster at scan 483). There are also additional components of apparent molecular weight of 194 (hydroxyphenanthrene or isomer) as well as m/z 182 as we have mentioned.
Full and partial profiles are shown in Figure 7 for a molecular ion with m/z 182 (M)+. and for the (M+1)+. and (M+2)+. ions that arise from higher isotopes of elements.
Figure 7. MPPSIRD of the m/z 182, 183, and 184 ions corresponding to hydroxyfluorenes.
The Finnigan-MAT 900 permits use of smaller mass increments (3.3 ppm in Figure 4 versus 5 ppm in earlier versions) and up to 31 m/z ratios in each MID descriptor. Inclusion of a calibration ion in addition to a lock mass ion in the MID descriptor provided errors of less than 1 ppm for the three partial profiles of the analytes. A mass resolution of 31,000 was used to resolve the analyte signals from interferences evident in the full M+1 and M+2 profiles. The results of the MPPSIRD experiments allowed us to unequivocally assign the elemental compositions of the molecular ion clusters studied. To our knowledge, these data represent state-of-the-art mass measurements on trace components in real samples at high mass resolution using capillary chromatographic separation.
These data present presumptive evidence for the presence of hydroxy-PNAs or isobaric compounds. Combining the migration behavior of model compounds and gas chromatographic retention times, presumptive evidence for hydroxy-PNAs is strong. However, the ions studied and the mass spectra are similar to heterocyclic compounds with an oxygen atom in the ring. Further confirmation of the presence of the aromatic hydroxy moiety was sought via derivatization using a tert-butyldimethylsilyl derivatizing agent. The low resolution spectrum of this derivative in given in Figure 8A for 1-hydroxypyrene, and a spectrum obtained from an extract is given in Figure 8B.
A. EI mass spectrum of standard of tert-butyldimethylsilyl derivative of 1-hydroxypyrene;
B. EI mass spectrum of component in soil extract derivatized with t-butyldimethylsilyl agent. Conditions: 70 eV; source 250°C; GC 60 to 300°C at 15°C/min; column HP5MS, 25 m X 0.25-mm ID, 0.25-µm film.
Unfortunately, the derivatization also changed the retention time of fatty acids (natural constituents of coal-derived materials) present in the extract and resulted in their near coelution with the compounds under study. Nevertheless, ion chromatograms clearly indicated that the three major ions and four additional ions of the tert-butyldimethylsilyl derivative (m/z 332, 275, and 259 plus m/z 73, 189, 201, and 203) of a hydroxypyrene or related compound were present at a retention time within 0.5 min of the derivative of 1-hydroxypyrene. The ion chromatogram of m/z 332 was relatively specific for the derivative as shown in Figure 9 where it is the dominant peak of the chromatogram.
Figure 9. Ion chromatogram of m/z 332 in the "acids" fraction after derivatization; conditions as in Figure 8.
Extraneous ions in the spectrum did not show identical ion chromatograms and retention time to those of the eight confirmative ions. These data confirm by CZE and GC/MS that the compounds, although not specifically 1-hydroxypyrene, are similar structurally to the hydroxypyrene.
Application of the relatively new tool of CZE/LIF has led to the identification of additional classes of contaminants in an environmental sample. The usefulness of cationic/anionic separations of CZE is further illustrated for real samples. Additional identification power was attained by GC/MS in combination with both MPPSIRD and derivatization for functional group verification.
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development, funded and performed the research described here. This work has been subjected to the Agency's peer review and has been approved as an EPA publication. The U.S. Government has the right to retain a non-exclusive, royalty-free license in and to any copyright covering this article. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
- Heiger, D. N. (1994) High Performance Capillary Electrophoresis, Hewlett-Packard Company, Publication Number 12-5091-66199E, Wilmington, DE
- Johnson, E. L.& Stevenson, R., (1978) Basic Liquid Chromatography, Varian Associates, Palo Alto, CA
- Grosser, Z. A., Ryan, J. F.& Dong, M. W. (1993) J. Chromatogr. 642:75-87
- Brumley, W. C.& Gerlach, C. (1999) Am. Lab. 31, 45-49
- Ferguson, P. L., Grange, A. H., Brumley, W. C., Donnelly, J. L., & Farley, J. W. (1998) Electrophoresis 19:2252-2256
- Brumley, W. C. (1995) LCGC 13:556-568
- Sovocool, G. W., Brumley, W. C., & Donnelly, J. R. (1999) Electrophoresis 20, in press.
- Langefeld, J. J., Hawthorne, S. B., Miller, D. J., & Pawliszyn, (1994) J. Anal. Chem. 66:909- 916
- Gleason, S. (1994) Am. Environ. Lab. 8:12-16
- Brumley, W. C., Latorre, E., Kelliher, V., Marcus, A. & Knowles, D. E. (1998) J. Liq. Chromatogr. and Related Technol. 21:1199-1216
- Brumley, W. C. (1995) J. Chromatogr. Sci. 33:670-685
- Test Methods for Evaluating Solid Waste (SW-846), (1986) Vol. 1B, U.S. Environmental Protection Agency, Washington, D.C., USA, 3rd ed., November (for the latest update see http://www.epa.gov/epaoswer/hazwaste/test/tx3xxx.htm).
- Brumley, W. C., Brownrigg, C. M.& Brilis, G. M. (1991) J. Chromatogr. 558:223-233
- Grange, A. H.& Brumley, W. C. (1996) LCGC 14:978-986
- Grange, A. H.& Brumley W. C. (1997) J. Am. Soc. Mass Spectrom. 8:170-182
- Willsch, H., Clegg, H., Horsfield, B., Radke, M. & Wilkes, H. (1997) Anal. Chem. 69:4203-4209
- Brumley, W. C., Ferguson, P. L., Grange, A. H., Donnelly, J. L., & Farley, J. W. (1996) J. Cap. Electrophoresis 3:295-299
- Bartels, M. J.& Kastl, P. E. (1992) J. Chromatogr. 575:69-74
- Grange, A. H.& Sovocool, G. W. (1999) Rapid Commun. Mass Spectrom. 13:673-686
- Moy, TW., Ferguson, PL,. Grange, A. H., Matchett, W. H., Kelliher, V. A., Brumley, W. C. Glassman, J., & Farley, J. W. (1998) Electrophoresis 19:2090-2094