California Water Science Center

A surrogate compound, 2-bromopropionic acid (CH₃CHBrCO₂H), is added to the samples to monitor the efficiency of extraction and methylation of HAA. This compound is chemically similar to the HAA but not produced in significant enough amounts by chlorination of DOC to prevent its use as the surrogate. 2-bromopropionic acid is used as the surrogate compound because the surrogate (2,3-dibromopropionic acid) listed in USEPA Method 552.2 coelutes with dibromochloroacetic acid on the GC column (Rtx-5, 5 percent diphenyl/95-percent dimethyl polysiloxane). A stock solution of 2-bromopropionic acid in MTBE at a certified concentration of 1 milligram per milliliter (mg/mL) is used to make the working surrogate solution in MTBE at a concentration of 10 micrograms per milliliter (μg/mL) by diluting 100-microliter stock solution with MTBE to a final volume of 10 mL using a volumetric flask and gas-tight syringe. Twenty microliters (μL) of the working surrogate solution are added to the sample just prior to extraction of the HAA to give 2-bromopropionic acid a concentration of 5 μg/L.

An internal standard (IS) that elutes with the methylated HAA during GC analysis is added to the samples and the HAA calibration standards. The measured peak areas for the analytes are normalized to the peak area of the IS to compensate for any small differences in GC injection volume or matrix effects on sample volitization in the injector between samples. 2-bromo-1-chloropropane (CH₂ClCHBrCH₃) is used as the IS because the IS (1,2,3-trichloropropane) listed in USEPA Method 552.2 coelutes with bromochloroacetic acid on the Rtx-5 column. An IS stock solution of 2-bromo-1-chloropropane in methanol is purchased at a certified concentration of 2 mg/mL. From this stock standard solution, the working IS solution in MTBE, at a concentration of 10 μg/mL, is prepared by diluting 50-microliter stock solution with MTBE to a final volume of 10 mL using a volumetric flask and gas-tight syringe. Ten μL of the working IS solution are added to every vial just prior to GC analysis to give an IS concentration of 0.4 μg/mL in the extract.

A primary stock standard solution containing all nine HAA in MTBE is purchased and used to prepare the working standards used for calibration. The stock standard solution is stored at –10°C and protected from light. It is stable for at least 1 month but should be checked for signs of evaporation. When purchasing commercially prepared standards, solutions prepared in methanol must not be used because the HAA are subject to spontaneous methylation when stored in this solvent (Xie and others, 1993). Furthermore, tribromoacetic acid is unstable in methanol because it undergoes decarboxylation when stored in this solvent (USEPA Method 552.2). The Supelco mix for USEPA Method 552.2 contains a mix of nine acids in the concentrations listed in table 2.

The working standard solutions, C, B, and A, are prepared by combining the stock standard solution and surrogate stock solution to give the concentrations listed in table 2. Working standard solution C is prepared first by combining 250 μL of the stock standard solution and 250 μL of the surrogate stock solution to give a final volume of 500 μL in a vial. The working standard solution B is prepared by diluting working standard solution C by a factor of ten with MTBE. One hundred μL of working standard solution C are measured with a gas-tight syringe and diluted to a final volume of 1 mL in a volumetric flask. The working standard solution A is prepared by diluting working standard solution B by a factor of ten with MTBE. One hundred μL of working standard solution B are measured with a gas-tight syringe and diluted to a final volume of 1 mL in a volumetric flask.

These working standard solutions are used to prepare procedural calibration standards, which comprise nine concentration levels of each analyte, with the lowest standard being at or near the method detection limit (MDL) of each analyte.

The quality-control standard solution (QCSS) is an independent standard solution used to prepare quality-control samples to verify the calibration standards (see “Quality-Control Samples” section of this report). The QCSS is prepared by weighing 0.05 grams (g) of each neat compound into individual 5-mL volumetric flasks and diluting to volume with MTBE. The resulting concentration of these stock standards is 10,000 mg/L. Forty μL of each stock standard solution is then mixed and the combined solution diluted to a final volume of 2 mL with MTBE. This solution, QCSS A, contains 200 mg/L of each of the nine HAA compounds. QCSS B is prepared by diluting QCSS A by a factor of ten (200-μLQCSS A to a final volume of 2 mL with MTBE) to give a final concentration of 20 mg/L for each of the nine HAA compounds.

The procedure to form HAA by chlorination of water samples is the same as the procedure to form THM in the method for determination of THMFP (Crepeau and others, 2004). The procedure is summarized briefly here. The DOC and NH₃-N concentrations of the sample are measured and used to calculate the appropriate amount of Cl₂ used to dose the samples. Samples with a DOC concentration of 3 mg/L or greater usually are diluted with organic-carbon-free water prior to chlorination. If both HAAFP and THMFP are being determined for the sample, the same dilution factor is used for both analyses. A dose solution containing 3,000 to 4,000 mg/L residual-free Cl₂ derived from sodium hypochlorite (NaOCl) and buffered to a pH of 8.3 with 1 molar (M) H₃BO₃ and 0.11 M NaOH is prepared. The pH of the sample is adjusted to a range between 8.3 and 8.7 by addition of dilute NaOH or hydrochloric acid (HCl). The sample is poured into three 40-mL amber glass vials with Teflon-faced septa, and sufficient dosing solution is added to satisfy the Cl₂ demand of the DOC and NH₃-N and to leave a residual-free Cl₂ concentration of 2–4 mg/L after the incubation period. The vials are sealed headspace-free and incubated at 25°C (77°F) in the dark for 7 days. After incubation, one vial is opened to measure the pH and free Cl₂. The pH must be 8.3 ± 0.1 and the residual-free Cl₂ must be between 2 and 4 mg/L. If these parameters are not met, then the sample is redosed and incubated for another 7 days. The remaining two vials are quenched by adding sufficient Na₂SO₃ solution to neutralize the residual-free Cl₂.The samples are refrigerated and can be held up to 14 days before extraction.

The samples are removed from refrigeration and allowed to equilibrate to room temperature. Two aliquots of every sample are analyzed, one undiluted and one diluted 1:5 with organic-free water. For the undiluted sample aliquot, one quenched vial is opened and 40 mL of sample water is measured with a graduated cylinder (which has been calibrated “to deliver” at 20°C with a 1-percent tolerance) and poured into a precleaned 60-mL vial with a Teflon-lined screw cap. For the diluted sample aliquot, the second quenched vial is opened, and 8 mL of sample are pipetted into a 60-mL vial along with 32 mL of organic-free water. The final concentration result for each compound in a sample is derived only from analysis of the diluted aliquot if the concentration measured in the undiluted aliquot is higher than the concentration in the highest standard. Twenty μL of the 10.0 μg/mL 2-bromopropionic acid surrogate solution is added to every 60-mL vial. When adding surrogate or standard solutions to aqueous samples, the tip of the syringe must be well below the water level to avoid loss by volatilization. After injection of the surrogate solution, the sample vial is capped immediately and inverted to ensure mixing of solutions. The pH is adjusted to less than 0.5 by adding 2 mL of concentrated H₂SO₄. Two grams (g) of copper II sulfate pentahydrate and 16 g of Na₂SO₄ are added immediately to the sample using the heat produced from the H₂SO₄ addition to help dissolve the salts. The samples are shaken until the salts are dissolved (approximately 2–3 minutes). Then, 2.0 mL of MTBE are added to the samples and they are shaken vigorously for 2 minutes. The MTBE and aqueous layers are allowed to separate for approximately 5 minutes.

A disposable glass pipette is used to remove as much of the MTBE layer (upper) as possible (minimum of 1 mL) and to place it into a 4-mL silanized autosampler vial that will be used as the reaction vessel for the methylation step. Then, 0.5 mL of 10 percent (volume/volume) H₂SO₄ in methanol is added to each autosampler vial. The cap is tightened securely and the vials are placed in a water bath at 50°C (122°F) for 2 hours to allow for methylation of the analytes.

The vials are removed from the water bath and placed in an ice bath for 5 minutes. Two mL of 0.1 normal (N) NaOH is added to each vial and the vials are shaken for approximately 2 minutes. The MTBE and aqueous layers then are allowed to separate.

A second set of autosampler vials containing silanized glass inserts is used to hold the final samples for GC analyses. Ten μL of the IS solution (10 μg/mL 2-bromo-l-chloropropane) is placed in the silanized insert just prior to addition of the sample extract. Exactly 250-μL of the MTBE layer (upper) is transferred into the insert using a 250-μL fixed volume micropipettor. The vials are capped and shaken or mixed with a vortex mixer.

The samples should be analyzed on the GC as soon as possible after preparation because the final extract solutions deteriorate after a few days. Losses of dibromoacetic acid, dibromochloroacetic acid, and tribromoacetic acid in particular were observed. If the sample must be analyzed on the GC for a second time, it is recommended that a new aliquot of the water sample be processed rather than reanalyzing the old final extract samples.

This method for extraction and methylation uses only one-half the volume of MTBE that is required for USEPA Method 552.2 (U.S. Environmental Protection Agency, 1995). Therefore, the volume of excess final extract that must be disposed of as solvent waste is much less.

The instrument consists of a 6890 Hewlett Packard autosampler connected to a 5890 Hewlett Packard GC equipped with an Rtx-5 capillary column and an ECD. The autosampler is set to deliver 1-μL samples to the injection port of the GC. The GC operating configuration is summarized in table 3. Figure 1A illustrates the performance on the Rtx-5 column with the method analytes at standard level 5 (see table 5 for concentrations), an extraction blank, and a full procedural blank. The peaks for all nine HAA, the IS and the surrogate compound are well resolved. Baseline separation is achieved for all peaks, except for trichloroacetic acid [retention time (RT) = 10.745 min] and bromochloroacetic acid (RT = 10.864 min), which nearly are baseline resolved. The identities of the peaks with retention times of 6.333 min and 9.479 min are not known. However, because these extra peaks are not present in chromatograms for full procedural blanks (fig. 1C), they likely are due to contaminants in the standards rather than contaminants introduced during sample preparation. The Rtx-5 column was chosen instead of the DB-5.625 column used in USEPA Method 552.2 because the response for the monochloroacetic acid was better, as was the general performance of the column.

Calibration standards are prepared using the same extraction and methylation procedures as for water samples. Nine calibration standards are made by adding appropriate volumes of the working standard solutions A, B, and C to 40-mL aliquots of fortified organic-free water (table 4). The organic-free water is fortified with Cl₂ dosing solution and sodium sulfite quenching solution in approximately the same concentrations used for generating the HAA in water samples. These solutes may affect extraction efficiency of the HAA and, thus, should be present in the same concentrations in standards and samples. Three working standard solutions (A, B, and C) are used so that the volume of MTBE added to the calibration standard solutions is less than 20 μL for all nine calibration standards. The concentration of the nine HAA compounds and the surrogate compound in the nine calibration standards are listed in table 5.

The calibration curves are generated using the IS technique. Peak areas for the nine HAA compounds and the surrogate compound are normalized to the peak area of the IS in the same injection. Because the IS is presented in the same concentration in all of the final extracts analyzed on the GC, this normalization compensates for any small variations in GC injection volume or differences in matrix effects in the final extract solutions between samples.

The calibration curves for all nine HAA and the surrogate compound are quadratic. The coefficient of determination, R², is used to assess the fit between each quadratic equation and the data for each analyte from the nine standard solutions. The R² value must be 0.9980 or better or a new calibration curve must be generated.

Samples are analyzed immediately after extraction and methylation. The order of analysis begins with an MTBE instrument blank to verify that the instrument is free of contamination. Next, the standard curve is produced by analyzing the nine standards from lowest to highest concentrations, followed by another MTBE instrument blank. Then, the extraction blanks and samples are analyzed, followed by two continuing calibration verification standards and another MTBE instrument blank between sets of samples (see “Quality-Control Practices” section of this report). The final data for a sample may be combined from two or more different analyses where the data for each HAA analyte are taken from the dilution that fits within the standard curve.

In each of the samples, the software identifies the peaks of the nine HAA species, the IS, and the surrogate by their retention times and then converts measured peak areas to concentrations by normalizing the peak areas to the peak area of the IS and then converting normalized peak areas to concentrations using the standard calibration curves. The retention time for each compound must be within a 0.25-minute window of the expected value from the calibration curve. Full separation of the compounds is achieved with the Rtx-5 column at the analyzed concentrations (fig. 1). The analyst examines the chromatograms to verify that the peak identifications are correct. Compound interferences on the column are minimal due to the selectiveness of the extraction method. The EZChrom software automatically flags samples if the surrogate concentrations, reproducibility of the duplicate samples, or concentrations of the calibration verification standards are out of acceptable range (see "Quality Control Practices” section of this report). The analyst also examines these data. The individual chromatograms, calibration curve information, and quality-control data are archived and the archived site is linked to the LIMS.

The data are first imported from EZChrom into a spreadsheet (Microsoft Excel) for verification and calculations. After all quality-control criteria for a set of samples are met satisfactorily, the data are transferred to the LIMS. The data are accessible to users of the LIMS after the analyst verifies the final concentrations.