Methods for Emerging Contaminant Groups: Explosives and Nitrosamines
In support of requirements in the Safe Drinking Water Act, the National Exposure Research Laboratory (NERL), has been developing analytical methods for chemicals on the 1998 contaminant candidate list (CCL) and for other chemicals of emerging interest. The purpose behind this effort is to provide analytical testing methods that can be used by the United States Environmental Protection Agency (USEPA) in unregulated contaminant monitoring (UCM) to gather nationwide occurrence data of these chemicals in drinking water. This occurrence data will support the Agency=s decision making process regarding future regulation of these chemicals in drinking water. Ideally, these methods will be specific, accurate, sensitive, and suitable for compliance monitoring if the chemical is regulated in the future. Specificity is of special concern when collecting UCM data, because with non-specific methods, there is the potential of regulating a chemical based on false positive occurrence data. Because many of these chemicals are not easy to measure at low concentrations in water matrices, method development has been an on-going challenge. Adding to the challenge is the fact that often the health effects information available for a chemical at the time of the method development are insufficient to provide a target detection limit. Two chemical groups of interest to EPA that fall into the category of emerging contaminants are explosives and nitrosamines. A new analytical method for explosives and related compounds, USEPA Method 529, was completed in 2002 (1). A method for nitrosamines is expected to be completed early in 2004.
The interest in developing a method for explosives began with the inclusion of RDX on the 1998 CCL. Hexahydro-1,3,5-trinitro-1,3,5-triazine is commonly known as RDX (Royal Demolition eXplosive). It is for the most part a military explosive, although it has a few commercial uses in demolition and in fireworks. At its peak production, the average volume of RDX produced in the U.S. was 180 million pounds per year (1969-1971) (2). RDX enters the environment from military installations where it is used and stored. Current groundwater contamination is attributed to historical disposal practices, which included open burning (3).
A survey of available information on analytical methods for, and occurrence of, RDX indicated that there was a specific grouping of military explosives and related chemicals that were typically analyzed for, and often found together in the environment. Existing methods were based on high performance liquid chromatography with ultraviolet detection (HPLC/UV) or gas chromatography with electron capture detection (GC/ECD). These methods were not selected for UCM because of their potential for false positives. However, the analyte list from existing methods was selected as the initial analyte list for a new gas chromatography/mass spectrometry (GC/MS) method (Table1). The objective for the new method development was to include as many of the analytes in Table 1 as possible. The procedure would use solid phase extraction (SPE) to concentrate the samples, followed by GC/MS detection. This process would provide a large concentration factor, minimal solvent usage, and excellent selectivity. The initial target reporting limit for RDX was 2 Fg/L or less based on EPA=s current Lifetime Health Advisory. Subsequent health information developed by USEPA indicated a theoretical 10-6 lifetime risk level of cancer from exposures at 0.3 Fg/L (4). A summary of the final method procedure is as follows:
1. One-liter samples are collected in amber glass bottles. Samples are preserved with Trizma pH 7.0 buffer and copper sulfate.
2. Samples are extracted by passing the 1-L sample through either a Waters RDX SPE cartridge or Varian RPS SPE disk.
3. After a short drying period, the SPE media is eluted with a small volume of ethyl acetate. Extracts are dried, and then evaporated to 1 mL with a stream of nitrogen gas.
4. Concentrated extracts are analyzed by GC/MS using a programmed temperature vaporizing (PTV) injector. The use of a PTV injector is required to minimize thermal decomposition of analytes in the GC injection port.
5. Quantitate the sample concentration by comparison to a calibration curve prepared from calibration standards in the range of 0.1-10 Fg/mL.
Data collected in full scan GC/MS mode resulted in a calculated method detection limit (MDL) for RDX of 0.082 Fg/L (SPE disk) or 0.12 Fg/L (SPE cartridge). MDLs for the other method analytes ranged from 0.021- 0.18 Fg/L depending on the analyte and the type of SPE used. However, there was no apparent difference in the precision or accuracy of the data between the two SPE types. Replicate sample analysis (N=5 for each SPE type) of reagent water fortified with method analytes at a concentration of 5 Fg/L resulted in mean recoveries of 87-109%, with an RSD for each analyte of less than 12%. Replicate sample analysis (N=8 for SPE cartridge, N=7 for SPE disk) of reagent water fortified with method analytes at concentrations of 0.1-1.0 Fg/L resulted in mean recoveries of 71-134%, with an RSD of less than 8% for 12 of the 14 analytes. Nitrobenzene and 2-nitrotoluene had RSDs of 15-18% respectively, with either cartridge or disk extraction, probably due to their volatility and possible loss during extract evaporation. Full scan data were obtained using a Varian Saturn 4D GC/MS with an Agilent 15m H 0.25 mm i.d. DB-5ms GC column, with a 0.25 Fm film.
An option to use selected ion monitoring (SIM) was included in the method to obtain a lower detection limit. Three ions in the proper relative abundance ratio were used for identification, and one ion was used for quantitation. Data collected in the SIM GC/MS mode resulted in a calculated method detection limit (MDL) for RDX of 0.006 Fg/L (SPE disk) or 0.010 Fg/L (SPE cartridge). MDLs for other method analytes ranged from 0.004 to 0.090 Fg/L depending on the analyte and the type of SPE used. Again, there was no apparent difference in the precision or accuracy of the data between the two SPE types. Replicate sample analysis (N=4 for each SPE type) of reagent water fortified with method analytes at a concentration of 1 Fg/L resulted in mean recoveries of 87-122%, with an RSD for each analyte of less than 9%. Replicate sample analysis (N=8 for each SPE type) of reagent water fortified with method analytes at concentrations of 0.05 to 0.25 Fg/L resulted in mean recoveries of 80-134%, with an RSD for each analyte of less than 10%. SIM data were obtained using a Shimadzu QP5050A GC/MS with an Agilent 15 m H 0.25 mm i.d. DB-5ms GC column with a 0.25 Fm film thickness.
The interest in developing a method for nitrosamines began with the information that N-nitrosodimethylamine (NDMA) is a likely disinfection by-product. This is of concern to USEPA because NDMA is an extremely potent carcinogen, with a theoretical 10-6 lifetime cancer risk level at exposures of 0.7 ng/L (5,6). Although NDMA is not on the 1998 CCL, the decision was made to start a method development effort because of its emerging importance as a drinking water contaminant. In addition to NDMA, it was decided to try to include seven additional nitrosamines that are on USEPA=s Resource Conservation and Recovery Act (RCRA) Groundwater Monitoring List to the method (Table 2). Other sources of NDMA and other nitrosamines in the environment are: production of rocket fuel, rubber manufacture and tanneries.
Analytical challenges associated with developing a method for NDMA centered around two issues. The first issue is that NDMA is miscible with water in all proportions, and therefore very difficult to extract from water efficiently. The approach to solving that issue was to investigate the use of carbon as an SPE sorbent. The second issue is the need for an extremely low detection limit combined with the need for very specific detection. In light of the fact that NDMA and other nitrosamines are very insensitive to conventional electron ionization MS, a less conventional approach was needed. The initial approach to detection was chemical ionization (CI) GC/MS using methanol as the CI reagent. As the method development progressed it became clear that the molecular ion for NDMA (m/z 75) may not be specific enough to preclude false positives. At this point, tandem mass spectrometry (MS/MS) was investigated as a possible option. Although method development is still in progress, a summary of the preliminary extraction and detection procedures is as follows:
1. Collect a 0.5-L water sample in an amber glass bottle. Dechlorinate with sodium thiosulfate.
2. Extract a 0.5-L water sample by passing it through an SPE column (6-mL volume) containing 2 g of coconut charcoal (80-120 mesh).
3. After a short drying period, elute the SPE column with a small volume of methylene chloride.
4. Dry the extract and concentrate it to 1 mL with a stream of nitrogen gas.
5. Analyze the extract by CI GC/MS or GC/MS/MS using methanol as the CI reagent. For the data presented here, an 8-FL injection was made using a PTV injector.
6. Quantitate the sample concentration by comparison to a calibration curve prepared from calibration standards in the range of 0.5 to 200 ng/mL.
Data obtained in the GC/MS mode from extracts of reagent water fortified with NDMA and seven additional nitrosamines at a concentration of 20 ng/L (N=3) showed 94% recovery of NDMA with an RSD of 3%. Results for six of the other nitrosamines in Table 2, showed 87-106% recovery with RSDs less than 4%. N-nitrosomorpholine could not be determined in this data set due to background contamination. Linear calibration was achieved over the concentration range representing sample concentrations of 1 to 400 ng/L (r 2 = 0.999 or 1.000 for all analytes). Preliminary data indicate that this procedure will produce a detection limit for NDMA near 1ng/L, with similar sensitivity for the other nitrosamines. Extracts were analyzed on a Varian Saturn 4D GC/MS/MS, using a Restek 30 m H 0.25 mm Rtx-5Sil MS GC column with a 1.0 Fm film thickness. No data are presented for GC/MS/MS because MS/MS parameters are still being optimized.
Future work on the nitrosamine method will include sample and extract preservation and holding time studies, a study of potential matrix effects, further optimization of MS/MS parameters, and selection of an internal standard and surrogate analyte.
Table 1. Method Analytes for USEPA Method 529
Table 2. Proposed Method Analytes for Nitrosamine Method
|Citation:||Munch, J. W., and M. V. Bassett. Methods for Emerging Contaminant Groups: Explosives and Nitrosamines. Presented at American Water Works Association Water Quality Technology Conference, Philadelphia, PA, November 2-6, 2003.
Mary P. O'Bryant - (919)-541-4871 or firstname.lastname@example.org
||Microbiological & Chemical Exposure Assessment Division
|| Chemical Exposure Research Branch