Automated Real-Time Determination of Bromate in Drinking Water Using LC-ICP-MS and EPA Method 321.8 Application ICP-MS Author Jason A. Day, Anne Vonderheide, Joseph A. Caruso University of Cincinnati Cincinnati, OH 45220 USA Thomas J. Gluodenis, Jr., Agilent Technologies Inc., 2850 Centerville Rd, Wilmington, DE 19808 USA Abstract The suitability of coupling an HPLC to an ICP-MS for the fully automated, routine analysis of bromate in drinking water as per the proposed EPA Method 321.8 was investigated. The necessity to monitor the carcinogen bromate in ozonated drinking waters at single ppb levels has led the USEPA to investigate HPLC-ICP-MS as an alternative technique to the ion chromatography with conductivity detection method currently specified. During this investigation, a series of rigorous performance checks were used to assess the implementation of the proposed method including the determination of bromate in a series of EPA disinfection byproduct (DBP) standards. Introduction Ozonation is a common method used for the disinfection of drinking waters. In waters containing bromide (Br - ), such as those found in coastal regions subject to salt-water intrusion, a disinfection byproduct (DPB) of the ozonation process is the bromate ion (BrO 3- ). The bromate ion, produced by the oxidation of bromide, is very carcinogenic, with an estimated lifetime cancer risk of 1:10,000 for a concentration of 5 ppb. 1 The current method specified by the USEPA for the determination of bromate in drinking water uses ion chromatography (IC) with conductivity detection. One disadvantage of this method is the need for a tedious and time consuming sample pretreatment step. The need for sample pretreatment arises from the potential for co-elution of chloride and bromide ions present in the sample, potentially resulting in false positive results. In order to prevent this from occurring, chloride present in the sample is precipitated out of solution using silver cartridges with subsequent pre-concentration of the bromate ions. This time consuming and lengthy clean-up procedure and pre-concentration step can result in preconcentration of sulfate ions present in the water. Sulfate can subsequently displace the bromate ions on the resonating column resulting in false negatives. For these reasons, ICP-MS has been investigated as an alternative, ion selective detector for this analysis. ICP-MS provides the resolution necessary to separate the bromate and chloride ion, thereby eliminating the need for a matrix elimination step. Furthermore, ICP-MS has been used successfully for the analysis of bromate in water samples containing concentrations of chloride in excess of 5000 ppb - much higher than the typical content of ozonated drinking water - without the need for sample pretreatment. 2 This study will investigate the suitability of ion chromatography coupled to ICP mass spectrometry (ICP-MS) as an automated, real-time measurement approach, to determine low levels of bromate in ozonated drinking water samples, using the proposed EPA Method 321.8. 3
Instrumentation The Agilent Technologies 1100 Series HPLC system, coupled to a 7500 Series ICP mass spectrometer using the real-time Plasma Chromatographic software was used for this study. This system was specifically designed for the rigors of automated trace element speciation work, mainly in response to laboratory demands, particularly in the environmental, clinical and food application areas, that need to carry out routine elemental speciation. Its design takes advantage of Agilent s expertise in chromatography and its recognized leadership position in ICP-MS. During the past few years, the potential of ICP-MS as a detector for elemental speciation studies has been realized. 4 When coupled to a chromatographic separation device, ICP-MS offers unmatched detection capability for laboratories interested in quantifying different species, forms, oxidation states or biomolecules associated with trace elements. 2, 5 Traditional approaches of coupling ICP-MS to chromatography devices are cumbersome, labor intensive and not readily automated. In fact, the majority of ICP-MS chromatography data handling software packages were designed specifically for liquid and gas chromatography (LC, GC) applications and required modification for use with ICP-MS. Some approaches even analyzed the chromatographic spectral peaks post-run, meaning the data had to be imported into another software package after the analysis was completed, for quantitation purposes. It was clear that there was a real demand for a fully automated system, designed specifically for trace element speciation analysis. Agilent Technologies answered that demand with a fully integrated package for trace element speciation, comprising an 1100 Series HPLC system, coupled to a 7500 Series ICP mass spectrometer, using the Agilent ChemStation and real-time Plasma Chromatographic software. 6 Methodology ICP-MS Conditions The ICP-MS instrumental conditions were optimized to give maximum signal at m/z 79, the most sensitive mass for Br. Because bromine is not completely ionized in argon ICP, sampling depth, nebulizer flow, RF power and ion lens voltages have to be optimized very carefully to guarantee the most efficient sampling of bromide ions. Operating conditions for the 7500 are shown in Table 1. These conditions gave an instrument response of 110,000 cps for a 100 ppb bromate standard, with a background of 1,800 cps (partially due to trace levels of bromide in the 18 MΩ deionized water). Table 1: Optimized Operating Conditions for 79 Br Using the Agilent 7500 ICP-MS Parameter Optimized conditions Nebulizer Meinhard concentric - glass Nebulizer flow rate 1.05 L/min Spray chamber Scott double pass - glass Spray chamber temperature 2 C Sample flow rate 1 ml/min RF power 1200 W Sampling depth Optimized for max signal at 79 Br Ion lens voltages Optimized for max signal at 79 Br Chromatographic Conditions See Table 2 for the chromatographic conditions for the separation. The column eluent was passed via a short length of PEEK tubing to a six-port Rheodyne injector equipped with a 100 µl (or 500 µl depending on the measurement) PEEK loop. A post column injection was performed at the beginning of each run (for internal standard purposes, specified in the proposed EPA Method) at the exact time the data acquisition began on the ICP-MS. See Figure 1 for a schematic of the HPLC instrumentation coupled to the ICP-MS. Table 2: Chromatographic Conditions for the Bromate Study Parameter Eluent mobile phase Injection volumes Post-column injector Pump flow rate Column Specification 25 mm Ammonium nitrate, 5 mm Nitric acid (~ph 2.7) in 18 MΩ Deionized water 100 µl, 500 µl loops Used for internal standardization 1 ml/min Dionex CarboPac PA-100 (94 250 mm) - with guard 2
Eluent bottles Degasser Agilent 7500 ICP-MS Pump Automated sample tray and injector Post-column injector ICP torch Nebulizer/spray chamber Q-pole Column Compartment Agilent 1100 HPLC Argon gas controller Rotary pump Turbo pump Turbo pump (ICP-MS not shown to scale) Figure 1: A schematic of the 1100 HPLC instrumentation coupled to the 7500 ICP-MS used for the bromate study. Sample Preparation The blank was 18 MΩ deionized water adjusted to ph 10 with NaOH. Standards were prepared daily from a USEPA 1 mg/ml bromate stock solution. Demonstration of Instrument and Method Performance As a way of maintaining data quality, the EPA uses performance checks to monitor the instrument and also ensure that the methodology is working correctly. Some of the more important performance checks for this proposed EPA method 321.8 include the measurement of: Abundance Sensitivity of ICP Mass Spectrometer Method Detection Limit Chromatographic Interferences Laboratory Fortified Blank Laboratory Fortified Matrix DBP Performance Sample These measurements were used to assess the performance of the integrated system used for this study. Abundance Sensitivity A large argon dimer, 40 Ar 40 Ar + at mass 80 adjacent to the bromate ion 79 Br + at mass 79, has the potential to bias results in the determination of bromate by ICP-MS. It is therefore critical that the abundance sensitivity, which is a measure of the instrument's ability to separate a trace peak from a major one, 7 is optimized to allow for maximum rejection of the ions at mass 80. The very high operating vacuum of the 7500, and the high frequency of its quadrupole, combined with optimization of the rodbias voltages, ensures that it achieves clean separation of both peaks, even at a mass of 79.5 amu, where the tail of the 40 Ar 40 Ar + might interfere with the Br + at mass 79. The excellent abundance sensitivity of the quadrupole's hyperbolic rods is demonstrated in Figure 2, which shows a spectral scan of 2% HNO 3. The effect of the large signal at mass 80 is shown to have minimal affect on the small bromine signal at mass 79. 3
40 Ar 40 Ar + 79 Br + Figure 2: Mass spectrum showing clean separation of 79 Br + from the argon dimer 40 Ar 40 Ar +. Method Detection Limit (MDL) Two method detection limits were performed - one using a 500 µl loop, as specified in the method, and another using a 100 µl loop. A blank and three calibration standards (1, 5, and 25 ppb bromate) were used for both method detection limit tests. Seven individually prepared bromate standards of 1 ppb (for the 100 µl loop) and 0.5 ppb (for the 500 µl loop) were then analyzed to determine the method detection limit (MDL). From this, an MDL was calculated for each loop by multiplying the standard deviation of the seven replicate results by 3.14, as indicated in the EPA method. Individual MDL replicate concentrations and statistics for both loops are shown in Table 3. Table 3: Method Detection Limit Data for a 100 µl and 500 µl Loops 100µL Loop 500 µl Loop Replicate # Concentration (ppb) Concentration (ppb) MDL-1 1.1 0.46 MDL-2 0.98 0.39 MDL-3 0.77 0.35 MDL-4 0.77 0.46 MDL-5 0.97 0.45 MDL-6 0.83 0.48 MDL-7 0.90 0.41 Mean 0.90 0.42 SD 0.131 0.044 RSD (%) 14.6 10.5 MDL 0.41 0.14 Chromatographic Interferences To show that other halogenated compounds do not elute at similar retention times as bromate, a haloacetic acid standard (HAA) standard solution, provided by the EPA, was analyzed. The stock solution was diluted 1:100 yielding final concentrations of six different halogenated compounds reported in Table 4. Table 4: Concentrations of Six Haloacetic Acid Compounds that Could Potentially Interfere with the Determination of Bromate Compound Concentration (ppb) Monochloroacetic acid 15 Dichloroacetic acid 15 Trichloroacetic acid 5 Monobromoacetic acid 10 Dibromoacetic acid 5 Bromochloroacetic acid 10 A chromatogram containing the haloacetic acid mixture and a 10 ppb bromate standard is shown in Figure 3. The retention time for bromate is 3.5 minutes. The bromine-containing HAA standards elute at 2.5 minutes, 5.9 minutes and 7.1 minutes indicating no chromatographic interference with bromate. Average bromate recovery (n = 2) for this standard spiked with 10 ppb bromate was 102%. 4
Post-column Bromate injection Bromate Bromoacetic acid Bromo- Chloroacetic acid Dibromoacetic acid Figure 3: A chromatogram containing haloacetic acid mixture and a 10 ppb bromate standard. Laboratory Fortified Blank Ten replicates of a laboratory-fortified blank (LFB) were analyzed at a concentration of 5 ppb, which was approximately ten times the MDL. The LFB samples consisted of 18 MΩ deionized water adjusted to ph 10 with NaOH and spiked with 5 ppb bromate standard. The average for the replicates was 4.7 ppb (8.9% RSD) with a 93% recovery. Laboratory Fortified Matrices Four fresh samples supplied by the EPA, taken from ozonation utilities in the U.S., were analyzed using this methodology. Each sample was adjusted to ph 10 with NaOH, and analyzed twice, unfortified and fortified with 10 ppb bromate. The results for all four samples are shown in Table 5. The recovery results for these matrices are all within the EPA guidelines of 70-130% for this method. Table 5: Bromate Results from Ozonation Utilities Concentration Concentration of bromate in of bromate in unfortified sample fortified sample Sample ID (ppb) (ppb) % Recovery A 2.0 12 102% B 2.7 12 89% C 4.0 16 118% D 8.9 18 100% USEPA DBP Performance Evaluation Check An EPA check ampule (USEPA ICR PE ampule for inorganic DBPs - Study 9), whose concentration was not known at the time of analysis, was also analyzed as a blind check sample. The ampule was prepared in duplicate by diluting 1:100 and analyzing immediately. Results are shown in Table 6. Once again, the recoveries are both within the recommended guidelines. Table 6: Sample Concentration in % Recovery original ampule (ppb) (917 ppb true value) Ampule 1 1120 120 Ampule 2 1040 113 Conclusion Recovery of Inorganic DBPs in EPA Check Ampules The ability to measure bromate in ozonated drinking waters at sub-ppb levels is essential to understanding its risk assessment as a carcinogen. Once USEPA Method 321.8 is validated for use, ICP-MS detection coupled to HPLC will become an approved method for achieving this. It has been shown that the instrumentation used in this study surpasses all the performance criteria specified in 5
the methodology, achieving a method detection limit of 0.14 ppb, with a 500 µl loop and 0.41 ppb with a much smaller injection volume (100 µl). Furthermore, this has been implemented in an automated fashion with real time data analysis using the Agilent 1100 LC and 7500 Series ICP-MS demonstrating that the technique is well suited for use as a routine analytical tool. References 1. WHO, Guidelines for Drinking Water Quality, 2nd Edition, Vol. 1, Recommendations of the World Health Organization, 1993 2. J. T. Creed, M. L. Magnuson, J. D. Pfaff, C. A. Brockhoff, Journal of Analytical Chromatography. 753, 261-267,1996. 3. Determination of bromate in drinking waters by ion chromatography inductively coupled plasma - mass spectrometry. USEPA Method 321.8 4. M. Yamanaka. Specific determination of bromate and iodate in ozonized water by ion chromatography with two detection methods: Post-column derivatization and ICP-MS. Journal of Analytical Chromatography, 779, 259-265, 1997 5. J. T. Creed, M. L. Magnuson, C. A. Brockhoff, Environmental Science and Technology. 31, 2059-2063,1997. 6. Technical Features of the ICP-MS Plasma Chromatographic Software: Agilent Technologies Application Note 5968-5943EN, Feb 2001 7. Practical Benefits of Abundance Sensitivity in ICP-MS: Agilent Technologies Application Note 5964-9024EN, Feb 2001 www.agilent.com Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material. Information, descriptions, and specifications in this publication are subject to change without notice. Copyright 2001 Agilent Technologies, Inc. Printed in the USA June 22, 2001 5988-3161EN