Jump to main content.

Ion Composition Elucidation (ICE)

Identifying Compounds Despite Chromatographic Limitations: Organophosphates in Treated Sewage — Final Draft

Andrew H. Grange and G. Wayne Sovocool

U.S. Environmental Protection Agency, National Exposure Research Laboratory, Environmental Sciences Division, Environmental Chemistry Branch, P.O. Box 93478, Las Vegas, Nevada 89193-3478

Highly concentrated extracts of sewage treatment plant (STP) effluents contain detectable levels of dozens of compounds resulting from human activities. Recent concern over use and disposal of Pharmaceuticals and Personal Care Products (PCPPs) (1) has stimulated interest in detecting, identifying, and quantifying these and related compounds, in determining their toxicities, and in assessing the risks they pose to ecosystems and to humans. For any compounds posing such risks, their sources, degradation rates, and degradation products must also be investigated before remediation strategies are developed.

Gas chromatographic separation in time coupled to mass spectrometric selection by mass (GC/MS) provides the most effective analytical method for identifying trace and ultra-trace levels of analytes in complex extracts that elute or coelute from a GC column. FTIR and NMR require larger amounts than MS of well separated compounds for analyses. MS provides distinct patterns of ion abundances from each coeluting compound, which can be correlated with individual compounds. Hence, we have developed a high resolution mass spectrometric technique to identify compounds in complex environmental extracts based on GC/MS data alone.

The latest version of the NIST mass spectral library usually included within mass spectrometer data systems for automated comparison of analyte and library mass spectra provides entries for 163,198 (2) different compounds, while the Palisade library contains 495,000 compounds.(3) There are three outcomes of these comparisons. A good match provides a tentative identification, multiple plausible matches are found (4), or no matches result due to a poor-quality background subtracted mass spectrum for the analyte or to the absence of the analyte mass spectrum in the library.

The EPA lists 2800 high production volume chemicals (at least a million lbs/yr) (5) and estimates 87,000 compounds are used in commerce (6). These compounds, their production byproducts, and their degradation products might be found in STP effluents. Many of these compounds are absent from mass spectral libraries.

To assess the risk posed by STP effluents flowing into water supplies, analytical tools are needed to identify compounds that do not yield tentative identifications based on mass spectral library searches. Ion Composition Elucidation (ICE), a high resolution mass spectrometric technique developed over the past decade by the U.S. EPA is such a tool. ICE provides the number of atoms of each element that comprise the ions observed in mass spectra. ICE is composed of two parts: Mass Peak Profiling from Selected Ion Recording Data (MPPSIRD) for data acquisition, and a Profile Generation Model (PGM) for automated data interpretation.

Organophosphate compounds found in an STP effluent presented a range of difficulty for compound identification from very distinctive mass spectra with unique library matches to those for which only the molecular ion was discernable above the chemical noise. ICE provided confirmatory evidence for each compound and was essential for identifying two sets of isomeric compounds.

Sample collection, extraction, and clean-up

Over 6 h, 60 L of STP effluent was pumped through a 5-µ filter (Whatman, Clifton, NJ) to remove particulates and 6 g of 1:1 polystyrene highly cross-linked with polystyrene cross-linked with 50% divinylbenezene:poly(methyl methacrylate) sorbent within a polypropylene cartridge (Abselut NEXUS, Varian, Harbor City, CA).

The retained organic compounds were eluted from the sorbent with 20 mL of n-hexane followed by 20 mL of ethyl acetate. The combined eluent was poured through a column of anhydrous sodium sulfate to remove water, and the column was then rinsed with methylene chloride to elute any remaining analytes.

After evaporation to 1 mL, this extract was passed through two Envirogel gel permeation columns in series (Waters, Milford, MA) using methylene chloride as the eluting solvent. The lipid containing fraction was discarded. The fraction containing analytes was evaporated to 400 µL, reconstituted in 1 mL of n-hexane, and further cleaned-up on a silica gel column. Three eluents from this column were combined and reconstituted in 400 µL of toluene. The concentration factor was 1.5 x 105 for this procedure. Reference 7 provides further details.

Tributyl phosphate (2000 µg/mL in acetone) was purchased from Chem Services (West Chester, PA), and tris(2-chloroethyl) phosphate (97%), tris(2-butyoxyethyl) phosphate (94%), triphenyl phosphate (99+%), and tritolyl phosphate (90% mixture of isomers - technical grade) were purchased from Aldrich (Milwaukee, WI) to provide comparison mass spectra and retention times for confirmation of tentative compound identifications.
Gas Chromatography
An HP 6890 (Agilent Technologies, San Jose, CA) gas chromatograph was interfaced to the mass spectrometer. One µL splitless injections of the extract were made onto a 30-m, 0.25-mm i.d., 0.25-µm film thickness, RtxTM-5MS capillary column (Restek, Bellefonte, PA). The temperature program was 90oC to 300oC at 7 C o/min for full scan acquisitions. The injector, transfer line, and ion source temperatures were 250oC, 300oC, and 250oC, respectively. The He flow was 0.1 mL/min or 1 mL/min when the mass range scanned was 50-500 amu or 150-450 amu, respectively.
Mass Spectrometry
The first stage of a Finnigan MAT 900S-ion trap hybrid mass spectrometer (Finnigan MAT, Bremen, Germany) was used to acquire all mass spectral data. For the total ion chromatogram in Figure 3 (discussed later), 0.5 s magnetic scans over 150 - 450 amu were acquired to provide good chromatographic resolution. For the mass spectra in Figure 4 (discussed later), 1.5 s scans over 50 - 500 amu were recorded to provide full mass spectra. The accelerating potential was 5 kV, the tungsten filament current was 1.0 mA, and the electron impact energy was 70 ev. The secondary electron multiplier voltage was between 1.6 and 2.2 kV it was higher within this range for greater resolving power or when very low levels of analytes were investigated.
The data acquistion facet of ICE is Mass Peak Profiling from Selected Ion Recording Data (MPPSIRD). Up to 31 m/z ratios are each monitored for 20 ms during each 1-s SIR cycle as analytes elute from a gas chromatograph into a double focusing mass spectrometer. The m/z ratios are monitored across individual mass peak profiles. The areas under the chromatographic peaks in the m/z traces are then plotted to provide full or partial mass peak profiles. In Figure 1a, six ion chromatograms are shown.
Figure 1a - For more information contact grange.andrew@epa.gov
Figure 1b- For more information contact grange.andrew@epa.gov
Figure 1c - For more information contact grange.andrew@epa.gov
Figure 1d - For more information contact grange.andrew@epa.gov
Figure 1. (a) Ion chromatograms acquired with 10,000 resolving power for m/z ratios corresponding to the maxima for two calibrant ions (top and bottom traces) and for the top four points on the profile in Figure 1c (middle four traces). (b) A full mass peak profile plotted from chromatographic peak areas for 3 of 21 m/z ratios monitored across a wide mass range at 100 (0.030 amu) ppm intervals. (c) A full mass peak profile acquired with 10,000 resolving power plotted from chromatographic peak areas from 10 of 21 ion chromatograms for m/z ratios that monitored masses separated by 10 ppm (0.0030 amu). Also shown are two partial profiles for a lock mass ion and a calibration ion, which bracket the analyte ion mass. Each is delineated by peak areas from ion chromatograms for five m/z ratios. (d) Three partial profiles for an analyte ion and its +1 and +2 profiles. Each was plotted from chromatographic peak areas from seven m/z ratios with 10 ppm mass increments. The resolving power was 10,000.

The top and bottom traces correspond to the maxima in the partial profiles for the lock mass and calibration mass that bracket the analyte mass shown in Figure 1c. The calibrant, perfluorokerosene, is ever-present in the ion source; thus, a steady signal is observed. The two baseline excursions in the traces resulted from reversing the voltage polarity of the draw plates for 5 s before and after the analyte eluted. The shaded areas between the excursions are integrated by the data system. The four intervening traces in Figure 1a displayed chromatographic peaks as the analyte eluted. The areas of these peaks were plotted as the top four points on the full mass peak profile shown in Figure 1c.

Determining the composition of an ion greater in mass than 150 amu usually requires three steps. First, a broad mass range is monitored by 21 m/z ratios with a mass interval of 100 ppm using a resolving power of 3000 (10% valley definition) to determine if multiple ions with the same nominal (integer) mass are present. In Figure 1b, a single profile delineated by three m/z ratios was observed for the nominal mass of 299 amu. A coarse estimate of the exact mass was calculated as the weighted average of the three points. Next, for Figure 1c, a resolving power of 10,000 (10% valley defiition) was used to monitor a narrower mass range about the estimated exact mass again using 21 m/z ratios, but with a 10 ppm mass interval. The weighted average of the top eight points provided an exact mass within an error limit of 6 ppm, which corresponded to 46 possible compositions as calculated by the Profile Generation Model (PGM) discussed below. Finally, to plot Figure 1d, the top portions of the m/z 299, profile and profiles greater in mass by +1 and +2 amu were each monitored by seven m/z ratios with a 10 ppm mass interval. The latter two profiles arose from ions containing one or two atoms of heavier isotopes (13C, 2H, 17O, or 18O). For each of these three experiments, the lock mass and calibration mass partial profiles shown in Figure 1c were each monitored by five m/z ratios. Linear interpolation between the mass errors for the two calibration ions estimated the errors in the exact masses for the analyte profiles, which were then corrected accordingly.

In addition to the exact masses shown on the three analyte profiles in Figure 1d, the abundances of the +1 and +2 profiles relative to the m/z 299 profile are shown. These relative abundances are calculated as the sum of the areas used to plot the +1 or +2 profile divided by the sum of areas for the m/z 299 profile times 100%.

MPPSIRD is performed by macro language programs within two Lotus 123 version 9.0 spreadsheets (Lotus Development Corp., Cambridge, MA, USA). No hardware modifications of the double focusing mass spectrometer were required. For each experiment, the user enters centroid masses of target profiles, start and end times that bracket the chromatographic peak for each analyte-specific time window (up to 32), the resolving power, and data file name. The first program prepares ASCII files that are executed by the data system to prepare for data acquisition and to process the data afterward. Entries into SIR descriptors are typed automatically and instrument parameters are set, including the mass resolving power. After data is acquired, ion chromatograms are displayed, peak areas are integrated, and an ASCII report file listing the m/z ratios and peak areas is prepared. In the second Lotus spreadsheet, the profiles are plotted and the exact masses and relative abundances are calculated. A more detailed description of MPPSIRD is provided in Reference 8. The software can be modified for use with double focusing mass spectrometers that provide a macro language for instrument control and for display, processing, and ASCII file outputs of masses and peak areas.

The automated data interpretation facet of ICE is the Profile Generation Model, which is written in QuickBASIC version 4.50 (Microsoft Corp., Bellevue, WA, USA). The three exact masses and two relative abundances determined from Figure 1d were entered into the model. The model compared these measured values with the calculated values for the 46 possible compositions determined from the full profile. In Table 1 are listed six compositions from the PGM output table.


Table 1 - For more information contact grange.andrew@epa.gov

An "X" indicated a measured and calculated value were inconsistent and the composition in that row was rejected. Two compositions passed all five comparisons including the fragment ion containing the PO4 moiety.
Multiple Isomers
In Figure 2a are shown calculated m/z 410, +1, and +2 mass peak profiles. The m/z 410 ion contains only the most abundant isotopes, while the +1 and +2 profiles are composites due to multiple ions.
Figure 2a - For more information contact grange.andrew@epa.gov
Figure 2b - For more information contact grange.andrew@epa.gov
Figure 2c - For more information contact grange.andrew@epa.gov
Figure 2. (a) Calculated mass peak profiles for the monoisotopic, +1, and +2 mass peak profiles corresponding to C24H27PO4+. Plotted are the three +1 ion contributions to the composite +1 profile and the four largest contributions to the composite +2 profile. The profiles are labeled with the calculated exact masses and relative abundances. (b) The corresponding partial profiles plotted from MPPSIRD labeled with the measured exact masses and relative abundances. (c) The ion chromatograms correspond to the m/z ratio at the maximum of each of the three partial profiles.
At least 25 isomers are evident.

In Figure 2b are shown the same partial profiles plotted from selected ion recording data. The ion chromatograms in Figure 2c were obtained with a mass resolving power of 10,000 (10% valley definition) and correspond to the maxima for the three partial profiles in Figure 2b. More than 25 isomers were present. The summed areas under these isomers were plotted as the maximum point for each partial profile.

MPPSIRD provides the data acquisition speed (1 cycle/s) sufficient to delineate chromatographic peaks, the 100-fold sensitivity advantage of SIR relative to full scanning, the high discrimination against mass interferences realized by using a resolving power of up to 20,000 (10% valley) routinely, and continuous recalibration as the lock mass maximum is located at the start of each SIR cycle.

The total ion chromatogram (TIC) (150 - 450 amu) in Figure 3 displays dozens of peaks, many of which are composites due to coelution of multiple compounds.
Figure 3a - For more information contact grange.andrew@epa.gov
Figure 3b - For more information contact grange.andrew@epa.gov
Figure 3c - For more information contact grange.andrew@epa.gov
Figure 3. A total ion chromatogram (150-450 Da) for an extract of STP effluent. Superimposed are individual ion chromatograms for the highest mass, prominent ion for organophosphate compounds having nine different molecular formulas and their structures
obtained from the NIST library of mass spectra.
A series of organophosphate compounds used as flame retardants and plasticizers was found in the extract. The structure of each compound is shown near its colored chromatographic peak in the TIC. Also shown in the same colors are the ion chromatograms for the prominent ion of highest-mass in each compound's mass spectrum. Tentative identifications were easily made for five organophosphate compounds based on matches of their mass spectra with those in the NIST library, but for five others, ICE was required to have confidence in their tentative identifications.

Compelling Library Matches

In Figures 4a and 4b are shown background subtracted mass spectra and the best NIST mass spectral library matches for the first eight of these compounds to elute.
Figure 4ab - For more information contact grange.andrew@epa.gov
Figure 4cd - For more information contact grange.andrew@epa.gov
Figure 4ef - For more information contact grange.andrew@epa.gov
Figure 4gh - For more information contact grange.andrew@epa.gov
Figure 4. Pairs of background subtracted mass spectra (yellow background) and the best matches of NIST library mass spectra (green background). Ions below m/z 50 have been removed. The spectra are presented from (a) through (h) in the order of each compound's elution. The compositions of the ions in red were determined using ICE. The wide mass range for (f) caused differences of 2 amu to appear as only 1 amu. Hence, the mass scale has been expanded around each chlorine isotopic cluster to show a gap between the ions separated by 2 amu.

For (c) tris(2-chloroethyl) phosphate (e) tris(3-chloropropyl) phosphate, and (f) tris(1,3-dichloropropyl) phosphate, compelling matches were found. A dozen or more mass peaks were present in both the background subtracted and library mass spectra, few mass interferences were observed, and characteristic chlorine isotope patterns indicated the number of Cl atoms present in each ion. The differences between the analyte and library mass spectra in relative mass peak heights for low and high mass ions do not necessarily refute these tentative identifications. These differences often result from lower sensitivity for high-mass ions provided by quadrupole mass spectrometers, which are usually the source of the library mass spectra, and from skewing in the mass spectra obtained using the double focusing mass spectrometer with a 1.5 s scan cycle to monitor chromatographic peaks only 6-8 s wide.

Compelling matches in Figure 4 were also found for (g) tris(2-butoxyethyl) phosphate and (h) triphenyl phosphate, both flame retardant plasticizers. Ten or more mass peaks are common to both the background subtracted and NIST mass spectra and relatively few mass interferences cloud the issue. Gone are the isotopic contributions from atoms containing 37Cl atoms. Even so, the observed matches provide strong tentative identifications for these two compounds.

Less Certain Library Matches

The matches between the NIST mass spectra and the background subtracted mass spectra for (a and b) tributyl phosphate isomers used as plasticizers, and (d) diazoxon, a metabolite of the pesticide diazinon, are less certain. Fewer than 10 ions are in common, not all prominent ions in the reference spectra are present in the analyte mass spectra, and conversely, not all prominent ions in the analyte mass spectra are present in the reference mass spectra. Fewer common ions, major mass interferences, and the low level of the analytes in the extract yielded less certain tentative identifications than those discussed previously. Even so, the mass spectral matches in (b) and (d) provide strong, if not compelling evidence for tentative identifications, while in (a), prominent interferences led to a questionable match.
Two Mass Spectra with Few Analyte Mass Peaks
The raw mass spectra in Figures 5a and 5b (top spectra) were found by paging through a data file, one scan at a time.
Figure 5 - For more information contact grange.andrew@epa.gov
Figure 5. Raw (top), background subtracted (middle), and NIST library (bottom) mass spectra for compounds that provided easily observed (a) m/z 368 and (b) m/z 410 ions. The composition of the m/z 410 ion was determined using ICE. Multiple compositions remained possible for the m/z 368 ion.
Background subtraction for scans showing maxima in m/z 368 and m/z 410 ion chromatograms provided the middle mass spectra in Figures 5a and 5b. The m/z 367-369 ions and m/z 409-411 ions provided two sets of three ion chromatograms that tracked each other well. The high level of chemical noise, low levels of analytes, and low abundance of fragment ions prevented observation of the fragment ions produced from these compounds. Entering the two sets of three ions into the NIST library and eliminating hits with prominent fragment ions in the mass range monitored, provided six possibilities with two compositions for the lower mass set and seven possibilities with seven compositions for the higher mass ions. No tentative identifications based on the "full" mass spectra (in this case only the molecular ion, the +1 profile, and the M-H ion) was possible.
Searching for Organophosphate Compounds by Monitoring a Characteristic Ion
The compounds that produced the mass spectra in Figure 4(a-g) have at least one alkyl chain attached to an O atom of the phosphate moiety. All seven reference mass spectra display an m/z 99 ion due to H4PO4+. The ion chromatograms for a full profile centered at m/z 98.98472 (the exact mass of this characteristic ion for alkyl phosphates) was obtained with 10,000 resolving power. There were no mass interferences for any of the eight m/z 99 full profiles observed. The first chromatographic peak revealed the presence of a compound that had been overlooked, an isomer of tributyl phosphate. This compound partially coeluted with a larger amount of a second compound, which contributed to the observed composite mass spectrum. The background subtracted mass spectrum in Figure 4a for this isomer is of lower quality than that of the second isomer in Figure 4b. Seven of the eight chromatographic peaks corresponded to compounds that provided the mass spectra in Figure 4a-g. The smallest peak corresponded to a smaller amount of a second tris(chloropropyl) phosphate isomer (shown in the m/z 277 ion chromatogram in Figure 3).
Application of ICE

For each of the ions labeled in red on the ten analyte mass spectra in Figures 4 and 5, full mass peak profiles were acquired with 10,000 resolving power to measure the exact masses of the monoisotopic ion. Partial profile data were then acquired with the same resolving power to determine the exact masses and relative abundances of the +1 and +2 profiles. For cases when mass interferences distorted the shape of the +1 or +2 partial profiles, fewer than three exact masses and two relative abundances were measured. When unique compositions could not be determined for the largest mass ion observed, compositions were determined for lower mass ions, which have many fewer possible compositions (8). An example is provided in Table 2.
Table 2 - For more information contact grange.andrew@epa.gov

The exact masses of the neutral losses between ions provided the neutral loss compositions. Adding a neutral loss composition to that of a smaller ion for which the composition was known provided a unique composition for the next larger mass ion studied. Although 47 compositions were possible for the m/z 355.18875 (+ to -3 ppm) ion based on consideration of C, H, N, O, F, P, S atoms, the correct composition was determined through this sequential approach, since the neutral losses from each ion left behind the phosphate core as a part of all ions studied. For these compounds, ICE provided confirmatory evidence for the tentative identifications. Only the two compositions in Table 1 were possible for the m/z 299.16246 (+ to -6 ppm) ion in Table 2 as determined by the PGM based on the three exact masses and two relative abundances from Figure 1d.

Figure 6 illustrates that even with 10,000 resolving power, mass interferences are still possible.

Figure 6 - For more information contact grange.andrew@epa.gov
Figure 6. Ion chromatograms for the center mass of the m/z 368, +1, and +2 partial profiles. The increasing signal for the
lower two chromatograms indicated a column bleed interference was present.

While the m/z 368 ion chromatogram (Figure 6a) revealed the presence of five isomers, those for m/z 369 and 370 (Figures 6b and 6c) showed a gradually increasing signal as the temperature in the GC oven increased. ICE was used to determine that the column bleed ions responsible for most of the signals was C11H33O4Si5+ and its +1 ions. A resolving power of 267,000 would be necessary to provide a 10% valley between mass peak profiles of equal height from the analyte +1 profile and the monoisotopic ion of the column bleed component (m/z 369.12115 and 369.12253, respectively). Five isomers produced the m/z 368 ion. The standard deviation for multiple measurements is inversely proportional to (N-1)1/2. Statistically, monitoring five isomers is similar to monitoring a single isomer five times. The mass error limit is 6 ppm for a single exact mass determination made with a resolving power of 10,000 (9). Dividing the error limit of 6 ppm by the square root of (5-1) provides 3 ppm as a more appropriate error limit. Based on the measured exact mass of the m/z 368 ion (368.11731 amu) and this error limit and considering only C, H, N, O, P, and S as possible elements with at least 1/3 of the ion's mass due to C atoms, 45 compositions were still possible.

  1. For the m/z 410 ion, 10,000 resolving power provided the partial profiles shown in Figure 2b. Table 3 is a partial output from the PGM showing four compositions that passed all five comparisons of measured and calculated values.
Table 3 - For more information contact grange.andrew@epa.gov

However, the ion chromatograms in Figure 3c reveal that at least 25 isomers were present. Using an error limit of 1.2 ppm (6 ppm divided by the square root of 24), only the composition C24H27PO4 passed all five tests. This composition was entered into the NIST library search option, and the single reference mass spectrum in Figure 5b for trixylyl phosphate was found. The large number of possible positional isomers for this structure account for the isomers observed in Figure 2c.

One of the possible compositions for the m/z 368 ion was C21H21PO4+. This composition was entered into the NIST library search option, and four matches were found for tritolyl phosphate. One library match is shown in Figure 6a. For three of the four NIST library matches, the methyl group was located at the same position on each of the three rings (tri-para, tri-meta, or tri-ortho). For the fourth library match, a CH2 group was located between the oxygen atoms of the phosphate group and the aromatic rings. These structures might account for four the five isomers observed in Figure 6a.

Comparison with Standards

When standards are commercially available, confirming tentative identifications is straight forward. The standard is purchased, and a retention time and mass spectrum identical to those of the analyte confirm its identification. Five standards were purchased: tributyl phosphate, tris(2-chloroethyl) phosphate, tris(2-butoxyethyl) phosphate, triphenyl phosphate, and tritolyl phosphate. The first four standards to elute provided identical retention times and very similar mass spectra -- any interferences in the analyte mass spectra were, of course, absent from each standard's mass spectrum. This evidence and the ion compositions determined using ICE for the prominent ion of highest mass in the analyte mass spectra confirmed the tentative identifications of these compounds. The tributyl phosphate standard provided the same retention time as the second, more abundant isomer. In Figure 6a, the m/z 368 ion chromatogram for the analyte isomers and tritolyl phosphate standard are shown. Four of the five observed isomers provided the same mass spectra, retention times, and similar relative abundances between the isomers. Although a unique composition was not obtained using ICE for this ion, determination of the composition of the molecular ion of the related compound, trixylyl phosphate, suggested the correct composition and led to purchase of the correct standard to confirm its very tentative identification. The three aryl phosphates found in the STP effluent are used as flame retardant plasticizers.

Only five of 10 standards were obtained. Still, the tentative identifications for the remaining five analytes are more certain due to confirmation of the composition of a prominent, high-mass ion in each compound's mass spectrum than tentative identifications based solely on mass spectral matches, some of which were dubious as discussed earlier.

Further supporting these organophosphate compound identifications are reports of tributyl phosphate (11-14), tris(2-chloroethyl) phosphate (12-18), tris(3-chloropropyl) phosphate (14-17), tris(1,3-dichloropropyl) phosphate (15-18), tris(2-butoxyethyl) phosphate (12,14), triphenyl phosphate (13,14,18,19), and tritolyl phosphate (13) found or quantified in surface water (11), numerous streams (18), rivers, ground water, rainwater, STP effluents (12), a large drinking water reservoir (16), indoor air (13,14), and on pine needles in the Sierra Nevada mountains (15).

Compounds Not in Mass Spectral Libraries
At least one isomer was found in the NIST mass spectral library for each of the nine organophosphate molecular formulas. However, Reference 10 describes identification of several non-organophosphate isomers found in a municipal well near Toms River, NJ, where an increased incidence of childhood cancer had been observed. These compounds were not in the NIST or Wiley mass spectral libraries. The compositions of the molecular ion and 10 fragment ions were determined. Interpretation of this information limited possible isomers to a number sufficiently small to make a library search of the commercial and chemical literature practical. Tentative identifications as 1:2 styrene:acrylonitrile adducts were made based on the literature search, and three of five isomer identities were confirmed after obtaining a standard from the current industrial process used to make stryene:acrylonitrile polymers.

STP effluents contain ultra-trace or trace levels of many compounds. GC/MS provides the best single analytical technique for identifying and quantifying analytes that traverse a GC column. Once compounds are identified, the toxicological studies necessary to assess the risk to ecosystems and humans by these chemicals can be made.

Many compounds in commerce are found in mass spectral libraries, and one or only a few library matches can lead to tentative identification of the compounds, at which point standards can be purchased to compare their mass spectra and retention times to those of the analytes. However, no library mass spectral matches, very poor matches, or many matches, can stymie attempts to identify compounds. ICE provides a powerful tool for advancing beyond this impasse. Determination of the molecular ion and/or fragment ion compositions provides knowledge that can be used to hypothesize tentative identifications before standards are obtained for confirmation.

An analyst specifically searching for the organophosphate compounds identified herein using a mixture of standards would find them. However, our search found them without targeting these compounds. Likewise, numerous other compounds have been identified without any prior knowledge or suspicion of their presence in complex extracts.

ICE software utilizes the selected ion recording mode of double focusing mass spectrometers to realize the potential of high mass resolving power for identifying compounds. Mass peak profiles are constructed from chromatographic peak areas under ion chromatograms acquired for up to 10 m/z ratios across individual mass peak profiles. The weighted average of the top several points that delineate the profile provide its exact mass, while ratios of the areas used to plot partial profiles provide their relative abundances. The PGM automatically provides a list of compositions corresponding to the exact mass of an ion within the error limit of its mass determination. This list is usually reduced to one composition when the exact masses of the ion and the profiles greater in mass by +1 and +2 amu that arise from heavier isotopes are entered into the model along with the relative abundances of the +1 and +2 profiles.

Several applications of ICE to analytical problems are described in a dozen articles and a dozen posters at: http://www.epa.gov/nerlesd1/chemistry/ice/default.htm. These include characterization of a Superfund site, identification of isomeric compounds in municipal well water and of compounds in a reservoir, confirming or rejecting tentative compound identifications for regulatory purposes, and measuring the exact mass of an alpha-hemoglobin adduct to within 0.2 amu.

C.G. Daughton, T.A. Ternes Environ. Health Perspect. 107 (suppl. 6), 907-938 (1999).
Grange AH, Osemwengie L, Brilis G, Sovocool GW Internat. J. of Environmental Forensics 2001; 2: 61-74.
L. Osemwengie, S. Steinberg J. Chrom. A 932, 107-118 (2001).
A.H. Grange, G.W. Sovocool Rapid Commun. Mass Spectrom. 16, 2356-2369 (2002).
A.H. Grange, W.C. Brumley J. Amer. Soc. Mass Spectrom. 8, 170-182 (1997).
A.H. Grange, G.W. Sovocool, J.R. Donnelly, F.A. Genicola, D.F. Gurka Rapid Commun. Mass Spectrom. 12, 1161-1169 (1998).
R. Tauler, D. De Almeida Azevedo, S. Lacorte, R. Cespedes, P. Viana, D. Barcelo Environ. Technol. 22, 1043-1054 (2001).
E. Fries, W. Puettmann J. of Environ. Monitoring 3, 621-626 (2001).
T. Yoshida, I. Matsunaga Osaka-furitsu Koshu Eisei Kenkyusho Kenkyu 39, 49-55 (2001) (Japanese).
I. Saito, A. Onuki, H. Seto Earozoru Kenkyu 16, 209-216 (2001) (Japanese).
L.S. Aston., J. Noda., J.N. Seiber, C.A. Reece J. Environ. Contam. Toxicol., 57, 859-866 (1996).
S.A. Snyder, K.L. Kelly, A.H. Grange, G.W. Sovocool, E.M. Snyder, J.P. Giesy 2001, Eds.: C.G. Daughton, T. Jones-Lepp, Symposium Series 791; American Chemical Society: Washington, DC, Ch. 7, 116-141 (2001).
D.B. Patterson, W.C. Brumley, V. Kelliher, P.L. Ferguson Amer. Lab. 34, 20-28 (2002).
D.W. Kolpin, E.T. Furlong, M.T. Meyer, E.M. Thurman, S.D. Saugg, L.B. Barber, H.T. Buxton Environ. Sci. & Technol. 36, 1202-1211 (2002).
D.R. Oros, N. David "Identification and Evaluation of Unidentified Organic Contaminants in the San Francisco Estuary" SFEI Contribution 45, San Francisco Estuary Institute, 34-35 (2002).


Analytical Environmental Chemistry
ICE Home Page

Environmental Sciences | Office of Research & Development
 National Exposure Research Laboratory
Author: Andrew Grange
Email: grange.andrew@epa.gov

Local Navigation

Jump to main content.