Ion Composition Elucidation (ICE)
A High Resolution Mass Spectrometric Tool for Identifying Organic Compounds in Complex Extracts of Environmental Samples
Andrew H. Grange,1 Lantis I. Osemwengie,1 Floyd A. Genicola,2 and G. Wayne Sovocool1
1U.S. EPA, ORD, NERL, Environmental Chemistry
PO Box 93478, Las Vegas, NV 89193-3478
2N.J. Dept. of Environmental Protection, Office of Coastal Planning & Program Coordination, PO Box 418, 401 E. State St., Trenton, NJ 08625-0418
ICE tentatively identified an organoarsenic compound in groundwater; a drug, temazepam; and several non-chlorinated organophosphorus compounds in a sewage treatment plant effluent.
Unidentified Organic Compounds
For target analytes, standards are purchased, extraction and clean-up procedures are optimized. Then the mass spectra and retention times for the chromatographic separation are obtained for comparison to the target compounds in environmental sample extracts. This is an efficient approach, and selective extraction and clean-up can decrease detection limits for the target compounds relative to analyzing a raw extract containing compounds that yield mass interferences. But selection of a target list of compounds for study may ignore many potentially toxic compounds. Before the toxicology of the hundreds of compounds found at trace or ultra-trace levels in sewage treatment effluents and water reservoirs can be studied alone and in mixtures, they must first be identified. A given compound might be one of the thousands of high production volume chemicals used commercially, byproducts of their production, or degradation products formed after their use. Numerous researchers targeting small numbers of analytes could easily remove many of these compounds through sample clean-up or be unable to identify them. Most non-targeted compounds will not be in mass spectral libraries and can seldom be tentatively identified using low resolution mass spectrometry alone.
Ion Composition Elucidation (ICE) is a tool employing high resolution mass spectrometry that makes practical characterization and identification of many of the compounds responsible for the myriad of chromatographic peaks observed for complex environmental samples. ICE has two facets: Mass Peak Profiling from Selected Ion Recording Data (MPPSIRD) to acquire data (1,2) and a Profile Generation Model (PGM) to plan experiments and interpret the data (3). Up to 31 m/z ratios are monitored sequentially each second across up to four analyte and two calibrant mass peak profiles. The areas under each analytes chromatographic peak in the ion chromatograms are plotted to provide the analytes full or partial mass peak profiles. Up to three such data acquisitions are made for each analyte ion using the successive approximation approach illustrated by Figure 1.
For Figure 1a, a mass resolution of 3000 was used and a wide mass range was surveyed. Two profiles delineated by three points each were revealed. The ion chromatograms for the m/z ratio with maximum abundance for each profile indicated that the lower mass ion resulted from an analyte corresponding to a chromatographic peak, and the higher mass ion (13CC3F7+) derived from an ever-present compound, in this case, perfluorokerosene the calibrant. The weighted average of the three points provided a coarse estimate of the exact mass, which was used as the center mass for the data acquired in Figure 1b with 10,000 resolution and a 10-ppm mass increment between the m/z ratios. The weighted average of the 10 points defining the full profile provided an exact mass accurate to within 6 ppm. Based on this exact mass (181.92056), error limit, and consideration of C, H, N, O, F, P, S, and As atoms, the Profile Generation Model listed 11 possible compositions, including C3H7S2As, which was chosen as the hypothetical composition to predict the exact masses of this ion and the mass peak profiles heavier by 1 and 2 Da that arise from the presence of higher-mass isotopes (e.g., 13C, 15N, 18O, and 34S). In Figure 1c, partial profiles were plotted that provided measured exact masses and abundances for the +1 and +2 profiles relative to the m/z 182 profile. This data was entered into the PGM to provide the compositions in Table I.
Table I. List of Possible Compositions for an Organoarsenic Compound
Comparisons to calculated values of three exact masses and two relative abundances rejected the other 10 compositions. An X next to an entry indicates the measured and calculated values were inconsistent and the composition was rejected. Unequivocally, C3H7S2As is the m/z 182 ions composition.
Applications: Organoarsenic compounds in groundwater
Two suspected As containing compounds were found in New Jersey groundwater from a landfill. Library mass spectra were not available for either compound. One was the compound that produced the m/z 182 molecular ion studied in Figure 1. This compound was synthesized and provided the same GC retention time and mass spectrum as the well pollutant. The second compound was thought to provide the CH3OAsS+ ion in Figure 2a.
Losses of methyl and methoxy groups accounted for the m/z 123 and 107 ions. No standard was available, but when the exact masses and relative abundances from Figure 2b were entered into the PGM, the two compositions in Table II were listed.
Table II. List of Possible Compositions for a Second Organoarsenic Compound
Clearly, the correct composition was C2H7AsS. In addition, ICE verified that loss of methyl accounted for the m/z 123 ion, while the m/z 107 ion resulted from loss of (CH3 + CH4). CH4 and O have exact masses of 16.03130 and 15.99491 Da, respectively, which yields a readily detected difference of 264 ppm between CH3OAsS and C2H7AsS.
For an analyte in an extract of a sewage treatment plant (STP) effluent, the mass spectrum in Figure 3a displayed a slightly elevated m/z 271 ion above a background having mass peaks at nearly all nominal masses.
The background subtracted mass spectrum in Figure 3b resembled the NIST library entry for temazepam in Figure 3c. The m/z 300 ion was investigated using ICE. The MPPSIRD plots shown in Figure 4 provided three exact masses and two relative abundances, which the PGM determined to be consistent only with the ion composition of the molecular ion of temazepam (C16H13ClN2O2).
A standard solution of temazepam was procured and the drug provided the same retention time as the analyte. Both diazepam and temazepam were among the 200 most prescribed drugs in 1999 (4). Diazepam is metabolized to temazepam, the bioactive agent (5), by addition of an OH group (Figure 3c). Diazepam, but not temazepam, has been found by others in other waters (6). We found only temazepam in this extract. In one experiment diazepam was spiked into the sampling line prior to the extraction cartridge. No conversion to temazepam by our sampling, extraction, or clean-up procedures was observed. Diazepam might be converted into temzepam by the sewage treatment process.
Organophosphates have previously been found in the environment. Aston, et al. (7) found three polychlorinated examples on pine needles in the Sierra Nevada mountains and we found the same compounds in Lake Mead (8) and the extract of STP effluent. Also tentatively identified in the extract were several non-chlorinated organophosphates: diazinon oxone (C12H21N2PO4), triphenylphosphate (C18H15PO4), tributoxyethylphosphate (C18H39PO7), and isomers of tricresylphosphate (C21H21PO4) and trixylylphosphate (C24H27 PO4). In Figure 5 are ion chromatograms acquired with 10,000 resolution for the m/z ratios that provided the greatest ion abundance for the molecular ions tricresylphosphate and trixylylphosphate.
Five isomers were evident in the m/z 368.11767 ion chromatogram and at least 25 isomers in the 410.16299 ion chromatogram. The background subtracted mass spectra for the higher mass compounds provided no prominent fragment ions for further investigation. Integration of the areas under the ion chromatograms for all isomers provided the partial profiles in Figure 6.
Table III lists four compositions consistent with the three exact masses and two relative abundances obtained from the partial profiles.
Table III. List of Possible Compositions - Organophosphate
Also shown are the mass differences between the measured and calculated masses. All three exact masses for the PO4 containing composition agreed within 0.7 ppm, while all exact masses for the other three compositions were in error by at least 2.6 ppm. Because signal was averaged for at least 25 chromatographic peaks, rather than one, the error limit is estimated to be about five times less than normal, i.e. 1.2 ppm rather than 6 ppm. Hence, the composition C24H27PO4 is probably correct. The variable positions of the methyl groups on the aromatic rings pictured in Figure 5 are responsible for the large number of isomers observed. The exact mass of the m/z 368 ion was consistent with the composition C21H 21PO4, which corresponds to the same compound with only one methyl group on each ring. An ever-present ion from column bleed (C11H33O4Si5) would require 250,000 resolution to be resolved from the m/z 369 profile from the tricresylphosphate. Thus, for this analyte, the +1 and +2 profile data could not be obtained. However, fewer isomers are possible and fewer were observed. Finding triphenylphosphate in the extract also supports this tentative identification. The presence of both isomer mixtures in the STP effluent results from their use as plasticizers in PVC pipe, flame retardants, and lubricants.
Daughton, C.G.; Ternes, T.A. Environ. Health Perspect. Supplement #6, 1999, 107, 907-938.
Aston, L.S.; Noda, J; Sieber, J.N.; Reece, C.A. Bull. Environ. Contam. Toxicol. 1996, 57, 859-866.
Snyder, S.A.; Kelly, K.L.; Grange, A.H.; Sovocool, G.W.; Snyder, E.M.; Giesy, J.P. in Pharmaceuticals and Personal Care Products in the Environment: Scientific and Regulatory Issues; Daughton, C.G. & Jones-Lepp, T., Eds.; Symposium Series 791; Amer. Chem. Soc.: Washington, DC, Ch. 7, 116-141.