Trace Organic Analysis
of Phthalates in Water and Soil by
Tandem Mass Spectrometry Under Chemical Ionization
Conditions with Isobutane as Reagent Gas
Article published in Intern. J. Off. Anal. Chem.
77, 1230-1236 (1970).
[note: minor content and formatting differences exist between this web version and the published version]
William C. Brumley, Elizabeth M. Shafter, and Paul E. Tillander
U.S. Environmental Protection Agency, Office of Research and Development, National Exposure Research Laboratory, Environmental Sciences Division, PO Box 93478, Las Vegas, NV 89193-3478
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.
*Author to whom correspondence should be sent. email: firstname.lastname@example.org
Eleven phthalate esters spiked in water and soil were determined by tandem mass spectrometry (MS) under positive chemical ionization mass spectrometry (CIMS) conditions with isobutane as reagent gas. Emphasis was placed on the determination step because tandem MS and CIMS are not widely adopted in current methods of the U.S. Environmental Protection Agency. Extraction by sonication and cleanup by use of a solid-phase extraction cartridge were adopted. The relative response factors gave relative standard deviations (RSDs) of 12-30% when 2 unlabeled internal standards were used. The relative abundances of monitored ions gave relative abundance deviations of less than ±8%. The method confirms identity, including molecular weight, and quantitates with high specificity. Results obtained with 2 unlabeled internal standards were compared with results obtained with a stable-isotope-labeled internal standard for dloctyl phthalate.Introduction
Phthalates are diesters of 1,2-benzenedicarboxylic acid (phthalic acid) and are of almost ubiquitous occurrence because of their widespread use as plasticizers. They are used also as solvents, insect repellants, and additives to plastic explosives.
Phthalates are target analytes described in the U.S. Environmental Protection Agency (EPA) Solid Waste 846 Methods Manual (Methods 8060 and 8270) (1) and in U.S. EPA Method 606 (2). The toxicity, mutagenicity, and carcinogenicity of bis(2-ethylhexyl)phthalate was reviewed recently (3). This compound is listed as "reasonably anticipated to be a carcinogen" (4). Many phthalate esters may be irritating to mucous membranes (5).
Regardless or the health effects in humans phthalates need to be determined because they are often present as overwhelming interferences and nuisances in samples used for the trace determinations of other pollutants. They are, therefore, of continuing interest to the Environmental Monitoring Systems Laboratory in Las Vegas. It is necessary to identify the troublesome phthalate compounds and to design appropriate steps to eliminate or reduce their effects on analyses.
Contamination by phthalates is also of concern in other areas. In the biomedical field. for example, the presence of phthalates in parenterally introduced fluids and feeding tubes was studied (6, 7). Various analytical approaches to determination of phthalates in foods were reviewed (8).
Some gas chromatography (GC) methods of determining phthalates involve solid-phase extraction (SPE) in sample preparation (9, 0), Other GC methods involve cleanup with sulfuric acid (II). These methods have been used to determine phthalate contamination of blood plasma (12), the marine environment (13), various plastic tubes (6), infusion products (7), food (8), the aquatic environment (14-16), and solvents (17).
Phthalate determinations are based also on liquid-chromatographic separations (18-20) with UV detection. Polarography was used as a detection technique to determine phthalates in water (21). The use of micellar electrokinetic chromatography to separate phthalates has been reported (22).
Thin-layer chromatography (TLC) also has been used to separate phthalates. SPE of phthalates from water followed by TLC separation was used by Sherma et al. (23). Further TLC work emphasized the general chromatographic behavior of phthalates (24-25).
The mass spectrometry and fragmentation of phthalates by electron impact (El) ionization were studied by several workers (26-29). Chemical ionization mass spectrometry (CIMS) also has been used to determine phthalates (30, 31). Yinon (32) studied the collision-induced dissociation spectra of phthalates. He found that the base peak of El mass spectra at m/z 149 was formed by 4 alternative pathways. He also observed the presence of fewer fragment ions in mass spectra when only the decompositions of M+ ions were considered.
Tandem mass spectrometry (MS) is not widely adopted in U.S. EPA methodology, especially as GC/tandem MS applications. Hunt and co--workers (31) presented a scheme for using tandem MS in environmental analysis. In view of the high specificity of tandem MS, we use it when very low levels of analyte are to be determined or when excessive cleanup operations are required. CIMS also is a reasonable choice when molecular weight information is lacking in El mass spectra or when more selective ionization is sought. In this paper, we present a procedure for determining phthalate esters that is based on tandem MS under positive CIMS with isobutane as reagent gas. Performance data are reported for 11 phthalates, and the specificity of the method is discussed.Experimental
The phihalate esters were obtained from Chem Service, Inc., West Chester, PA. A standard solution containing compounds 1--11 (defined below) in hexane and a separate internal standard solution containing 12 and 13 were made up with the following concentrations (ng/µL) of phthalate:
1, dimethyl phthalate (12.3); 2, diethyl phthalate (21.0); 3, dibutylphthalate (15.5); 4, bis(methoxyethyl) phthalate (38.3); 5, bis(ethoxyethyl) phthalate (58.5); 6, dicyclohexyl phihalate (81.0); 7, bis(2-ethylhexyl) phthalate (50.4); 8, di-n-octyl phthalate (70.4); 9, butylbenzyl plithalate (143.0); 10, butyloctyl phthalate (65.4); 11, diamyl phthalate (124.0); 12. dimethyl isophthalate (20.5); 13, hexyl-2-ethylhexyl phthalate (74.0).Sample Preparation
Aqueous samples were subjected to liquid-liquid extraction in a separatory funnel as described in SW-S46 Method 3510(1) with slight modifications. Brief1y, 0.25 mL standard was added to 1.5 mL methanol; this solution was spiked into 250 mL deionized (ASTM Type II) water (Barnstead/Thermolyne, Dubuque, IA). Methanol was used to help disperse the phthalates in water. The spiked solution was then extracted successively with 12, 6, and 6 mL methylene chloride, and the 3 extracts were combined. The combined extract was concentrated under a gentle stream of nitrogen while warming with a hot-air gun. A final volume of 0.1-1.0 mL was reached after addition of 0.25 mL internal standard containing 12 and 13. All spikings were done by syringe (Hamilton Co., Reno, NV).
Soil samples obtained locally in Las Vegas and soil B (L. A. Clark Superfund site, Spotsylvania,. VA) were extracted according to SW-846 Method 3550(1) with slight modifications. Briefly, a 10 g sample of soil was extracted by sonication. Soils were spiked with 100 µL of the standard before extraction. Methylene chloride-acetone (50 + 50, v/v) (30 mL) was used with a sonicator (model W-375. Heat Systems-Ultrasonics, Inc., Farmingdale, NY) for extractions.
Extracts were filtered through a 5.5-cm-diameter filter (Schicicher and Schuell, lot no. 0-891) by using a Buchner funnel connected to a vacuum pump (Edwards Model E2M-l, Crawley Sussex, England). The filtrate was dried by blowing a gentle stream of nitrogen while warming with a hot-air gun. The filtrate residue was dissolved in 1 mL hexane.
The solid-phase extraction (SPE) cleanup was carried out with either a Supelco (Bellefonte, PA) LC-Florisil (LC-FI) 6 mL cartridge or a Varian (Sunnyvale, CA) LC-E1 I mL cartridge. The 6 ml. cartridge was used on the heavily contaminated soil B to increase adsorbent capacity. The 1 mL cartridge was used with the local soil. Both sizes of cartridges, although obtained from different manufacturers, performed similarly in experiments with standards. The Supelco LC-FI cartridge was first washed with 6 ml. hexane. The sample was applied and then washed with 6 ml. hexane. The phthalates were then eluted with 8 mLhexane-acetone(90+ 10, v/v). The same procedure was followed for the Varian cartridge. The Varian I niL cartridge was first washed with 4 mL hexane. The sample was added and then washed with 4 mL hexane. The phthalates were then eluted with 6 mL hexane-acetone (90 + 10, v/v). The eluate in both cases was then spiked with l00 µL internal standard solution containing 12 and 13 or the internal standard containing d4-dioctyl phthalate and concentrated to 0.1-1.0 mL under a nitrogen stream.GC/Tandem MS
A Finnigan MAT TSQ-45 instrument was used to analyze the final extracts. The instrument was interfaced to an INCOS 2300 (Nova4X) data system with TSQ software (Rev 6.6). The collision-activated decompositions (CAD) were monitored under experimental control of the data system. A multiexperiment descriptor (terminology used by INCOS) was created that contained 5 experiment descriptors. Each experiment descriptor controlled the CAD of 3 (M + H)+ ions, with 3 ions monitored per decomposition. This procedure is known in the literature as multiple-reaction monitoring. Some redundancy was incorporated in the descriptors to keep all experiments equivalent [i.e., 12 unique (M + H)+ ions were used among 15 total (M +H)+ ions whose CADs were followed]. The dwell time per ion was 0.18 s, resulting in a total cycle time of 0.54 s per experiment. The number of cycles for each experiment was adjusted to divide the entire run into 5 retention time windows. Automated procedures were created to process the data.
Samples were introduced to the standard 9610 gas chromatograph (GC) by on-column injections using a modified injector (33). The retention gap consisted of a 3 m x 0.53 mm id, deactivated, fused-silica column. The capillary column was a 30 in x 0.25 mm id SPB-5 (Supelco. Inc., Bellefonte, PA) column with 0.25 µm film thickness. The temperature was programmed from 500 to 300CC at 20C/min; the flow rate (He) was 30 cm/s at 50'C. The analytical column was connected to the mass spectrometer via an uncoated, deactivated transfer line (1 m x 0.25 mm id) held at 260C. The isobutane (Aldrich Chemical Co., Milwaukee, WI) reagent gas was introduced from the makeup (GC) side at a source pressure of 0.35- 0.40 torr, as read from a standard Finnigan-MAT ionizer pressure gauge. The source temperature was 110C (dial setting), the emission current was 0.27 mA, the electron energy was 70 eV, the preamplifier was set at 10-8 A/V, and the electron multiplier and conversion dynode were set at -1300 V and -3 kV respectively. The collision gas was Ar (Scott Specialty Gases, Inc., Plumsteadville, PA), maintained at 1.6-2.0 mtorr (Hastings Gauge) so that the analyzer pressure reading was 4.2 x l0-5 torr.Results and Discussion
The determination involved a multiexperiment that followed the decompositions of 13 (M + H)+ ions representing the 11 analytes and 2 internal standards. For each decomposition, 3 ions were monitored: the (M + H)+ ion and 2 product ions. (With 1, only 2 ions were observed in the CAD spectrum) One of these 3 ions was used for quantitation (the m/z values of these ions are underlined in Table 1). The multiexperiment, which consisted of S experiments divided the entire run into 5 retention-time windows. Thus, the total acquisition cycle times (0.54 s) were capable of reproducing the chromatography. Isobutane was chosen as the reagent gas to maximize the relative abundances of the (M + H)+ ions of the phthalates and to offer some selectivity in terms of compounds that would be ionized on the basis of the gas-phase acidity scale. Table 1 lists response factors (RFs), their relative standard deviations (RSDs), and the ions monitored (the ion used for quantitation is underlined). The data represent a 5-point calibration curve that spanned about a factor of 10 in concentration. Each of the 5 points was determined in triplicate. All quantitations were made within the working range of the calibration curve.
The RSDs for response factors ranged from about 12 to 30%. These RSDs are not untypical for multianalyte determinations with internal standards that are not isotope-labeled analogues of individual analytes (1). These data support the use of this approach to determine phthalates.
It was of interest to compare RSDs obtained with 2 internal standards to those obtained with a stable isotope-labeled internal standard. The RF data for 8, obtained with a ring-labeled d4 analogue of 8, gave an RSD of 4% (data not shown). As expected, the data obtained with isotope-labeled standard showed better precision than the data obtained without isotope-labeled internal standards.
Another parameter of interest in tandem MS determinations is reproducibility of the relative abundances of the ions monitored. This reproducibility affects our ability to confirm the identity of each compound. Table 2 lists the relative abundances of ions monitored and the RSDs for relative abundances. A more common concern to mass spectrometrists is the percent relative abundance error (expressed as ±). These values were below ± 8% and therefore were within customary expectations (for examples SW-846 Method 8290 specifies a ± 15% relative abundance error for dioxin determinations). Therefore, we concluded that the identity of phthalates could be confirmed by monitoring 3 ions per compound (for compound 1, only 2 ions were monitored) and observing the responses at the proper retention times.
The CADs observed in this study generally consisted of the loss of ROH
(i.e., the alcohol moiety) and the production of m/z 149. Figure
1 illustrates a typical product ion spectrum From m/z 279 of
Figure 1. Product ion mass spectrum of m/z 279, (M+H)+ of dibutyl phthalate.Product ions were few compared with the numerous fragment ions observed in El mass spectra. Compound 1 could not readily produce a product ion at m/z 149. Compounds 4 and 5 also were exceptions in that a low relative abundance of m/z 149 was produced. These compounds (4 and 5) showed charge retention on the alcohol moiety of the phthalate ester.
Precision of Determination and Recoveries of Phthalates from Water and SoilRecoveries determined by tandem MS for spiked water samples are tabulated in Table 3 along with RSDs based on 3 determinations. The overall method precision (from spiking to determination of phthalates in the final extract) was acceptable; RSD values ranged from 2 to 7%.
Results of triplicate determinations of soil spiked with phthalates at 0.1-1.0 µg/g are given in Table 4. The average recoveries ranged from 22 to 68% for compounds not found in the method blank (i.e., all compounds except 3 and 7). The precision of determination by tandem MS for a single sample extract injected 3 times gave RSDs of 2.5 to 25% for all compounds not in the method blank (i.e., except 3 and 7). These values parallel the RSDs for the relative response factors of the compounds (see Table 1 for comparison). To place this variation in perspective, we studied the precision of determination by tandem MS of a single extract with d4-dioctyl phthalate as internal standard. The RSD in this ease was 2.2%. This result again pointed to improved precision when the deuterium-labeled internal standard was used.
Figures 2 and 3 show total-ion current responses for a standard and a
Figure 2. Total-ion current for multiexperiment with
standards. Compound number are according to the text: 1,
dimethyl phthalate, retention time (RT) = 9.93; 2, diethyl
phthalate, RT = 10.93; 3, dibutyl phthalate, RT = 13.50;
4, bid(methoxyethyl) phthalate, RT = 13:75; 5,
bis(ethoxyethyl) phthalate, RT = 14.50; 6, dicyclohexyl
phthalate, RT = 17.50; 7, bis(2-thylhexyl) phthalate,
RT = 17.53; 8, di-n-octyl phthalate, RT = 18.84; 9,
butylbenzyl phthalate, RT = 16.27; 10, butyloctyl phthalate,
RT = 14.80; 11, diamyl phthalate, RT = 16.20; 12,
dimethyl isophthalate, RT = 10..37; 13, hexyl-2-ethylhexyl
phthalate, RT = 16.83. Group 1: 1, 12, 2; Group 2: 3, 4; Group 3: 5, 11;
Group 4: 10, 9, 13; Group 5: 6, 7, 8.
Figure 3. Total ion current for multiexperiment with spiked soil. Compound numbers are as given in Figure 2.
The 5 retention-time windows are indicated in Figure 2, along with typical retention times. The observed patterns in the total-ion and the individual-ion chromatograms indicated no interferences from the matrix that prevented confirmation of phthalates. Method blanks did indicate that the laboratory was phthalate-contaminated with compounds 3 and 7 and, in later work, compound 2.
Because the cleanup step used SPE-Fl cartridges, recovery from Florisil affected overall method performance. The recovery precision for Florisil and the observed average recoveries and RSDs were similar to those reported earlier (9).
Recoveries from spiked soils as determined by tandem MS are given in Table 5. The first 3 determinations (soil 1, soil 2, and soil 3) were for 3 portions of soil of local origin; for these, RSDs were 7-44%. Overall method precision (from spiking to final extract) was comparable with that of the multianalyte method. The RSDs for overall method precision reflect variances contributed by the extraction step, cleanup step, and determination step.
The fourth example (soil B) given in the table was heavily contaminated with creosote. The recovery of phthalates from this sample actually was better than from the local soil. The results of phthalate determination in an unspiked sample of soil B (given in parentheses relative to the spiking level used in spiked soils) indicated environmental contamination above method contamination blanks for compounds 3, 7, 8, 9, and 11. These data represent the confirmation and quantitation of phthalates in a Superfund site soil sample at levels of <1 to >1.0 µg/g.
We also found additional laboratory contamination by compound 2 in subsequent work; this explains the high recovery of 2 from this soil matrix. We believe that the overall higher recovery was due to improved recovery from SPE-FI (compare with Table 4). There was no evidence of matrix interference (i.e., responses that were not phthalates). Enhanced recovery due to coextractives has been observed before in SPE (34, 35).
We applied tandem MS under CIMS conditions to the determination of phthalates as environmental contaminants. The precision, specificity, reproducibility of spectra, and calibration statistics were within reasonable expectations for a GC/MS multianalyte method. The use of stable-isotope-labeled internal standards improves method performance in terms of precision and is the technique of choice, but nonlabeled internal standards also produce data comparable with those from traditional MS approaches for multianalyte analyses.
This research was funded by the U.S. Environmental Protection Agency through its Office of Research and Development (ORD) and was conducted by the Environmental Monitoring Systems Laboratory in Las Vegas (EMSL-LV) to support ORD's Hazardous Waste Issue. It was subjected to ORD's peer and administrative reviews and was approved as an EPA publication.
(1) U.S. EPA (1986) Test Methods for Evaluating Solid Waste;
Physical/Chemical Methods, Vol 1 B, 3rd Ed., U.S. EPA, Washington, DC(2) Federal Register, Pt. 36, App. A. Method 606,40 CFR CH 1 (July 1, 1992 Edition)
(3) Registry of Toxic Effects of Chemical Substances, D. V Sweet, (Ed.), DHHS (NIOSH): Cincinnati, OH, Publication No. 93-101-2, January 1993
(4) Sixth Annual Report on Carcinogens, (1991) DHHS, prepared for NIEHS, Research Triangle Park NC, p. 172
(5) The Merck Index, 11th Ed., Merck & Co., Rahway, NJ, 1989
(6) Khaliq, M. A., Alam, M. S., & Srivastava, S. P. (1992) Bull. Environ. Contam. Toxicol. 48, 572-578
(7) Halkiewicz, J.& Halkiewicz, A. (1988) Farm. Pollut. 44, 16-21
(8) Page, B. D. (1988) Food and Packaging Interactions, J. H. Hotchkiss (Ed.), ACS Symposium Series 365, American Chemical Society, pp. 118-135
(9) Lopez-Avila, V., Milanes, J., Dodhiwala, N. S., & Beckert, W. E. (1989) J. Chromatogr. Sci. 27, 209-215
(10) Lopez-Avila, V., Benedicto, J. & Milanes, J. (1989) Evaluation of Sample Extract Cleanup Using Solid-Phase Extraction Cartridges. Report EPA/600/4-89/059
(11) Thuren, A. & Soedergren, A. (1987) Int. J. Environ. Anal. Chem. 28, 309-315
(12) Teirlynck, O. A..& Rosseel, M. T. (1985) J. Chromatogr. 342, 399-405
(13) Waldock, M. J. (1983) Chem. Ecol. 1, 261-277
(14) Millar, J. D., Thomas, R. E., & Sbattenberg. H. J. (1984) Method Study 16, Method 606 Phthalate Esters. Report EPA-600/4-84-056
(15) Kobli, J., Ryan, I. F., & Afghan, B. K. (1989) in Analysis of Trace Organics in Aquatic Environment, BK. Afghan & AS.Y Chan (Eds), CRC Press, Boca Raton. FL, pp 243-281
(16) Peterson, I. C., & Freeman, D. H. (1982) Int. J. Environ. Anal. Chem. 12, 277-291
(17) Lopez-Avila, V., Milanes, J., Constantine, F. & Beckert, W. E. (1990) J. Assoc. Off. Anal. Chem. 73, 709-720(IS) Russell, D. J., & McDuffie, B. (1983) Int. J. Environ. Anal. Chem. 15, 165-183
(19) Dong, M. W.& DiCesare, J. L. (1982) J. Chromatogr. Sci. 20, 517-522
(20) Wu, J.C.G. (1991) J. Environ. Sci. Health, Part A A26, 1363-1385(21) Tanaka, K. & Takeshita, M. (1984) Anal Chim. Acta 166, 153-161
(22) Ong, C.P., Lee, H. K., & Li, S.F.Y. (1991) J. Chrornatogr. 542, 473-481
(23) Sherma, J., Dryer, J., & Bouvard, J. J. (1986) Am. Lab. 18, 28, 30-32(24) Albert, L., Aldana, P. & Carbajal, C. (1986) Rev. Soc. Quim. Mex. 30, 143-146
(25) Brinkman, U.A.T., & De Vries, G. (1983) J. Chromatogr. 258, 43-55(26) Ende, M. & Spiteller, G. (1982) Mass Spectrom. Rev. 1, 29-62
(27) McLafferty, F. W.& Gohlke, R. S. (1959) Anal. Chem. 31, 2076-2082
(28) Emery, E. M. (1960) Anal Chem. 32, 1495-1506
(29) Djerassi, C., & Fenselau, C. (1965) J. Am. Chem. Soc. 87, 5756-5762
(30) Hunt, D. F., Shabanowitz, J., Harvey, T. M., & Coates, M. L. (1983) J. Chromatogr. 271, 93-105
(31) Hunt. D. R., Shabanowitz, J., Harvey, T. M., & Coates, M. L. (1985) Anal Chem. 57, 525-537
(32) Yinon, J. (1988) Org. Mass Spectrom. 23, 755-759
(33) Brumley, W. C., Brownrigg, C. M., & Grange, A. H. (1993) J. Chromatogr. 633, 177-183
(34) Loconto, P. R. (1991) LC-GC 9, 460-465
(35) Loconto, P. R. (1991) LC-GC 9, 752-760
Table 1. Response Factor performance data for tandem MS of phthalates
|Compound||Ref Cmpd||Ave RFa||RSD, %||m/z (M+H)+||m/z other monitored ionsb|
RF = (area of analyte ion x ng of internal standard)/(area of internal standard ion x ng of analyted).
b The underlined ions were used for quantitation.
c NA, not applicable.
Table 2. Reproducibility of relative abundances (RA) of ions monitored for phthalatesa
|Compound||RA (first ion), %||RSD, %||RA error, %||RA (second ion), %||RSD, %||RA error, %|
|2||(223) 25.1||10.0||3.0||(149) 74.2||11.3||8.0|
|3||(279) 12.6||20.8||3.0||(205) 33.1||6.9||2.0|
|4||(283) 18.4||6.4||1.0||(149) 0.31||95.0||0.3|
|5||(311) 15.0||7.9||1.0||(221) 92.6||6.9||6.0|
|6||(331) 23.4||25.9||6.0||(231) 14.8||16.8||2.0|
|7||(391) 18.9||13.0||1.0||(261) 4.6||11.5||0.5|
|8||(391) 5.7||5.3||0.3||(261) 15.3||7.9||1.0|
|9||(313) 25.3||17.4||4.0||(205) 57.5||12.0||7.0|
|10||(307) 9.0||12.6||1.0||(219) 23.7||6.4||2.0|
|11||(335) 9.2||15.7||1.0||(205) 18.7||14.3||3.0|
Table 3. Precision of recovery of phthalates from spiked watera
|Compound||Rec 1, %||Rec 2, %||Rec 3, %||Av Rec, %||RSD, %|
b Present as a contaminant in syringe blanks.
Table 4. Precision of determination of phthalates in the final extract of a single spiked soila
|Compound||Rec 1, %||Rec 2, %||Rec 3, %||Ave Rec, %||RSD, %|
b Present in method blanks.
Table 5. Recoveries of phthalates from spiked soilsa
|Compound||Rec Soil 1, %||Rec Soil 2, %||Rec Soil 3, %||Av rec, %||RSD, %||Rec Soil B|
a The spiking levels are the same as given in footnote a in Table 4. ND, not determined.
b Present as a contaminant in the method for this soil.
c Present in method blanks.
d Present in unspiked Soil B; values in parentheses are percentages of the spiking level.