Optimization of a Multiresidual Method for the Determination of

Before the extraction of clean water samples, 4 g of EDTANa2, 100 µL of IDQS ...... Vanderford , B. J.; Pearson , R. A.; Rexing , D. J.; Snyder , S. ...
0 downloads 0 Views 3MB Size
Environ. Sci. Technol. 2008, 42, 4068–4075

Optimization of a Multiresidual Method for the Determination of Waterborne Emerging Organic Pollutants Using Solid-Phase Extraction and Liquid Chromatography/Tandem Mass Spectrometry and Isotope Dilution Mass Spectrometry CHUNYAN HAO,† XIAOMING ZHAO,† S H A H R A M T A B E , ‡ A N D P A U L Y A N G * ,† Applied Chromatography Section, Laboratory Services Branch, Ontario Ministry of the Environment, 125 Resources Road, Etobiocke, Ontario, Canada M9P 3V6, and Standard Development Branch, Ontario Ministry of the Environment, 40 St. Clair Avenue West, Toronto, Ontario, Canada M6C1G4

Received May 2, 2007. Revised manuscript received February 17, 2008. Accepted February 26, 2008.

A high-throughput, liquid chromatography/tandem mass spectrometry (LC/MS-MS) method has been developed for the determination of 51 emerging organic pollutants (EOPs) in environmental waters. The method was validated for the analysis of 38 pharmaceutically active, 10 endocrine disrupting, and three perfluoroalkylated compounds. Method performance parameters, including sample preservatives, pH values used in the solid-phase extraction (SPE), sample storage, sample extract storage time, and matrix effects were discussed in detail for different aquatic matrices, including drinking water, wastewater, and surface water. Isotope-labeled compounds were used as injection internal standards (IIS) or isotope dilution quantitation standards (IDQS) to improve the data quality, investigate the behavior of matrix effects during SPE sample preparationandLC/MS-MSanalysis,andtovalidateisotopedilution mass spectrometric (IDMS) determination of selected compounds. Method detection limits were determined to be in the low ng/L range for the compounds evaluated. By application of this method to the analysis of effluents and samples downstream of a wastewater treatment plant, more than 35 target EOPs were quantified. We demonstrated method ruggedness by quality control and quality assurance (QC/QA) data, showed that matrix effects were dependent on modes of electrospray ionization, and could not be removed via SPE or cleanup procedures, exerting the same effect to target compounds in both raw and extracted samples. Both 13C- and 2H-labeled IDQS could be added to samples before sample extraction, and their recoveries used to correct matrix effects in LC/MS-MS EOP analyses.

* Corresponding author fax: 416-235-5900; e-mail: paul.yang@ ontario.ca. † Applied Chromatography Section, Laboratory Services Branch. ‡ Standard Development Branch. 4068

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 11, 2008

Introduction The development and application of electrospray ionization (ESI) devices that interface a high-performance liquid chromatograph and a tandem mass spectrometer (LC/MSMS) provide a powerful analytical tool for environmental chemists, which led to the discovery of emerging organic pollutants (EOPs) in the environment. The EOPs are characterized by their higher molecular weight and polar/ionic moieties (1), and are composed of a wide range of chemicals including, but not limited to, pharmaceutically active compounds (PhACs), endocrine disrupting chemicals (EDCs, estrogens, and steroid hormones), and perfluoroalkylated surfactants (2-4). They are introduced into the environment because of anthropogenic activities, and have not been subjected to the same scrutiny as persistent organic pollutants and pesticides. The possibility that the health of wildlife and humans might be adversely affected by being continually exposed to these compounds, even at low levels, has generated concern because there is a limited knowledge of their environmental concentrations, fate, and synergistic and nontarget effects (5). To investigate the influence of EOPs on the environment, reliable analytical methods are critically needed to address the occurrence, concentration and fate of these chemicals. Many LC/MS-MS methods were reported for the study of EOPs with detection limits of ng/L or lower in sludge, soil, wastewater, surface water, groundwater, and drinking water samples (6-8). However, most of these methods were developed for application to a specific class of compounds with limited method validation data reported. This led to the development at the Ontario Ministry of the Environment (MOE) of a LC/MS-MS multiresidue method for the monitoring of EOPs in the aquatic environment. The MOE method uses a neutral, single-step solid-phase extraction (SPE) instead of an acidic or basic SPE extraction. In addition, three separate LC/MS-MS analyses were required for the determination of 38 PhACs, 10 EDCs, and 3 perfluoroalkylated surfactants in different environmental waters. Sixteen 13Cand 2H-isotope labeled compounds (ILCs) were used as injection internal standards (IIS) to correct for the final sample volume and injection volume variations or isotope dilution quantitation standard (IDQS) to carry out isotope dilution mass spectrometric (IDMS) analyses. Documented in this manuscript are factors such as pH used for SPE, sample preservatives, sample and extract storage times, and matrix effectssthat impacted analytical resultssand their possible solutions. Method performance, the matrix effects during SPE and LC/MS-MS analyses, and the use of IDMS to overcome matrix effects are discussed in detail.

Experimental Section Chemicals. Analytical standards listed in Table S1 (Supporting Information), methanol, acetonitrile, sulfuric acid, sodium hydroxide, heptafluorobutyric acid (HFBA), ammonium acetate (>99%), and ethylenediaminetetraacetic acid disodium salt (EDTANa2, ACS reagent grade) were purchased from Sigma Aldrich (Oakville, ON, Canada). 13C- and 2Hlabeled standards listed in Table S1 were purchased from Cambridge Isotope Laboratories (Andover, MA) and CDNIsotopes (Montreal, QC, Canada), respectively. High-quality water (pure water) used to prepare method blank and method spike samples was produced by passing osmosis water through a Barnstead NANOpure water purification system. Separate stock solutions of analytical standards, including those for ILCs, were prepared for each individual compound 10.1021/es7028125 CCC: $40.75

 2008 American Chemical Society

Published on Web 04/15/2008

by weighing approximately 10 mg of each and dissolving it in 10 mL of methanol or methanol/water (50:50/v:v) in calibrated plastic tubes (Simport, QC, Canada). Intermediate solutions were prepared in 100 mL volumetric flasks by mixing the stock solutions. Five calibration standards and one spiking solution were prepared from intermediate solutions by serial dilution with ranges from low ng/mL to µg/mL. IDQS solution was prepared by mixing stock solutions of ILCs (Table S1, total of 13), diluted in methanol. The IDQS solutions were added to samples before extraction. IDQS recoveries were used to monitor method performance and matrix effects, and to validate the IDMS analysis. The IIS solution was prepared from stock solutions of ILCs (Table S1, total of 3), diluted in HPLC grade water, added into the final sample extract before LC injection, and used to correct variations of sample volume occurring during the final sample concentration and sample injection processes. We also investigated the validity of using the three IIS standards to alleviate the concern of matrix effects exerted on their respective native compounds. Sample Collection, Storage, and Preparation. Field grab samples were collected in precleaned 1-L brown glass bottles with 250 mg/L of sodium thiosulfate and 1% formaldehyde (wastewater effluents only) added as preservative. The samples were then packed in a cooler with ice packs, shipped to the laboratory, and stored in the dark between 2 and 6 °C until analysis. 99.9% N2 gas (Parker Balston, MA) was used in the sample preparation and MS-MS analysis. Solid phase extraction was performed using Waters (Millford, MA) HLB SPE cartridges (6 mL, 200 mg). Samples were prepared in batches of 24 that included 4 QC samples (laboratory blank, duplicate pure water method spikes, tap water method spike) and 20 field samples to maximize operational efficiency. Duplicate pure water method spikes and tap water method spiked samples were used to monitor within-run method precision and matrix effects, respectively. Typically, we used 800 mL of clean water (drinking water and source water) and 150-400 mL wastewater samples in the sample preparation. Before the extraction of clean water samples, 4 g of EDTANa2, 100 µL of IDQS solution, and 20 mL of 0.25 M ammonium acetate solution were added. Samples were homogenized on a laboratory roller (Wheaton Science, NJ) for 10 min, and the pH value of each sample was adjusted to 6.95 ( 0.05 using 10% (w/v) NaOH and 10% (v/v) H2SO4 solution. Wastewater samples were treated in the same manner except that 500 µL of IDQS solution was used to ensure that recoveries of the IDQS could be determined. The HLB cartridges were sequentially conditioned with 5 mL each of water, methanol, and 5% (v/v) ammonium hydroxide in methanol before the SPE extraction. Following extraction, the HLB cartridges were rinsed with 5 mL of 10% (v/v) methanol/water and dried by air, and the target compounds were eluted from the SPE cartridges with 5 mL of methanol. Either 5 mL (for drinking water or source water) or 1 mL (for wastewater) of the eluate were evaporated to dryness with N2 at ambient temperature, and reconstituted by using 0.1 mL of IIS solution. Calibration standards were also prepared by using a mixture of 0.3 mL of IDQS solution and 0.3 mL each of different levels of target compound solutions, evaporating to dryness under nitrogen, and reconstituting to 0.3 mL with IIS solution before instrumental analysis. Instrumentation. Analyses were performed using an Agilent 1100 LC (Mississauga, Ontario, Canada) coupled with an Applied Biosystems API 4000 Q-trap mass spectrometer (Foster City, CA) using an ESI interface. Multiple reaction monitoring (MRM) data were acquired and processed for all compounds in either positive or negative ion mode. Confirmation of unknowns and/or target compounds in field samples was done using the information-dependent acquisition (IDA) function with an MRM intensity threshold of

1000 cps. LC column (Thermo Electron, Bellefonte, PA, Hypersil Gold, C-18, 100 × 2.1 mm, 3 µm) was used in three separate chromatographic runs with acidic, neutral, and basic mobile phases. Column temperature was 30 °C and the injection volume was 20 µL. Figure S1 illustrates the mobile phases and gradient parameters used in the separations. Curtain, collision, nebulizer, and auxiliary gases of the MSMS were set at 15, 6, 35, and 45, respectively. Source temperature and entrance potential were kept at 450 °C and 10 V for both positive and negative modes. Ion spray voltage, declustering potential, and collision cell exit potential used were 5200, 60, and 10 V for the positive, and -4500, -90, and -5 V for the negative ESI, respectively. MRM parameters were optimized by direct infusion, and the most intense ion pair for target analytessand their respective optimized parametersswere chosen for the analysis. Values of collision energy (CE) are listed in Table S1.

Results and Discussion Method Performance. Figure 1 shows typical, reconstructed MRM chromatograms for the 51 EOPs validated using this method. Instrument detection limits (IDL) (i.e., the amount of analyte at which the MRM chromatogram has a signalto-noise ratio of 5), and method detection limits (MDL) (9) are summarized in Table S1. In addition, Table S1 also lists the average and relative standard deviation of % recoveries (%R) of analytes obtained from the method spike QC samples, and the average of relative difference (%RRD) between the %R of the duplicate method spikes (%Rspike1 and %Rspike2). %RRD was used to evaluate within-run method precision for each analyte and was calculated using the equation below: %RRD ) (%Rspike1 - % Rspike2) × 2 ⁄ (%9Rspike1 + % Rspike2) With %R data from Table S1, we observed that the method performs well for all but two target compounds (penicillin G and chlorotetracycline), and had an average %R of 83 ( 70% and 57 ( 80% in the QC samples, respectively. We also noted that even the laboratory pure water would exert certain “matrix effects” to the analytical results; in the form of either ionization enhancement or suppression in the positive and negative ionization modes, respectively. Good method performance can be further supported by the within-run method precision data expressed as average %RRD between the dual method spikes associated with each sample batch. Irrespective of an unrealistic range of average %R from about 140% (e.g., erythromycin and naproxen in positive ionization) to 640% recoveries (norfloxacin, positive ionization), over 90% of the target compounds analyzed had an average %RRD of 5-15%. These data demonstrated one inherent difficulty in the LC/MS-MS analyses: a high variability in recoveries of specific compounds due to matrix effects, and shows the importance of using QC/QA to support the consistency while qualifying the accuracy of the data generated. Effect of Sample pH on Recovery. The wide range of EOPs with different chemical characteristics makes it a challenge to extract them in one single SPE procedure with good method performance. The effect of pH (e.g., 2.5, 7.0, and 9.5) on the %R for each EOP was investigated. Figure 2 shows the average %R (N ) 4) of target compounds whose values fluctuated more than 30% at the three pH values tested. From Figure 2, we note that different pH conditions could cause either the overestimation or underestimation of recoveries for different target compounds. For example, carbadox’s %R was over 150% at pH 2.5 while diethylsillbestrol’s %R was over 250% at pH 9.5. It was also clear that some analytes had low recovery at acidic pH (lincomycin and erythromycin) or basic pH (tetracycline category and sulfa drugs). The reason for this abnormal %R was not clear VOL. 42, NO. 11, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4069

FIGURE 1. Typical, reconstructed MRM chromatograms for the 51 EOPs analyzed using this method. Compound name for each peak is in listed in Table S1. and might be attributed to matrix effects or poor cartridge retention. In the former case, one explanation might be that pH-induced molecular conformation change would either encourage or discourage the formation of ions in the ESI source. Thus, ionization was enhanced or suppressed in the ESI chamber, resulted in over- or under-estimation of the target compounds. From Figure 2, the pH 7 extraction gave the best overall %R for the EOPs studied. We also noted that %R values of the acidic drugs were very sensitive to minor pH changes. These led to the extraction at a neutral pH (6.95 ( 0.05) and the use of ammonium acetate buffer to maintain the sample pH during the SPE procedure. The study also showed that the performance of PFOA, PFOS, PFBS, which remained as anions between pH 2.5 to 9.5, were minimally affected by pH, and their recoveries were consistently observed to be 90-115%. Effect of Amount of Sample Preservative. Preservatives are commonly used in environmental analysis to prevent target analytes from decomposing, especially for chlorinetreated drinking water samples for which chlorine residue is common (10, 11). By using 250 mg/L of sodium thiosulfate as preservative, the recoveries of target analytes in chlorinated tap water samples were compared to those obtained from unpreserved, chlorinated tap water samples. Results of target 4070

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 11, 2008

compounds where their %R varied over ( 10% are shown in Figure 3. It is clear that the use of preservatives can alleviate the effect of chlorine and improve the %R for most of the compounds studied. These findings are different from those reported in a previous study (10) which used different extraction parameters (10). One can observe the drastic improvement of %R for the tetracycline family and lincomycin when chlorinated water samples were preserved. For lincomycin, tylosin, and PFOA, one could observe the alleviation of ionization enhancement as demonstrated by the reduction of over-recoveries. Use of preservative was thus deemed to be imperative for studies to determine the removal efficiencies of EOPs in water treatment plants. To prevent the influence of bacteria, formaldehyde was added into wastewater effluent samples without any observed adverse effect on the QC samples. Sample Storage and Extract Storage Study. Environmental laboratories usually extract and analyze samples in batches to optimize operational efficiency. Samples or sample extracts could be stored for several days or weeks before the analysis. We therefore investigated the relation between %R and sample storage and extract storage time by using seven sets of spiked, duplicate drinking water, wastewater treatment plant (WWTP) water composite (80% influent and 20% effluent),

FIGURE 2. Effect of sample pH on the %R of target compounds (N ) 4, pure water matrix) at pH 2.5, 7.0, and 9.5.

FIGURE 3. Effect of sodium thiosulfate on the %R of target compounds in drinking water (N ) 2) that were extracted at pH 6.95 ( 0.05. and pure water for comparison. The amount of preservative used was 250 mg/L of sodium thiosulfate for all samples and an additional 1% formaldehyde for wastewater samples. Two target compounds, carbamazepine from the positive ionization analysis group and carbadox from the negative ionization analysis group, were used to demonstrate the

storage study results. Carbamazepine is the most commonly detected PhAC in environmental water samples. Carbadox, on the other hand, has never been detected in our laboratory. Therefore, it is important to validate that current sampling, storage, and analytical protocols could be used to report occurrence data for carbadox with confidence. VOL. 42, NO. 11, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4071

FIGURE 4. Recoveries of carbamazepine (A) and carbadox (B) at sample storage time of week 0-6. Results shown were the average of duplicate. Error bars were obtained from pure water results. Figure 4A and B show the recoveries of carbamazepine and carbadox at different sample storage times obtained from the average %R of the duplicate spikes. From 4A, the %R values for carbamazepine in the three sample matrices were stable up to six weeks. The higher recoveries in the WWTP samples were from the positive background level of carbamazepine. One could observe that %R for the same sample matrices were similar during the study with or without using preservatives. From Figure 4B (carbadox), the %R for pure water was consistent; however, the %R of drinking water and wastewater were affected significantly by preservatives. No recovery of carbadox was observed within one week (drinking water samples), or four weeks (wastewater samples) without using preservatives; however, carbadox can be detected after six weeks with the use of preservatives. Figure S2A and B show the relationship between the average %R and storage time of sample extracts obtained from the duplicate spikes with the %R normalized to that obtained from pure water spikes. The %R for both carbam4072

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 11, 2008

azepine and carbadox showed similar patterns and peaked at week three; an indication that levels of these two analytes in the extracts had increased before they started to decompose. This might be attributed to the possibility of initial deconjugation of the target compounds occurring in the extracts from the sample matrix, and that these compounds returned to their native form over time. These observations suggested that analytical results of EOPs depend on sample and extract storage times, and a consistent sample storage protocol with associated QC samples is required to ensure good data quality. Matrix Effects. The polar and ionic nature of EOPs require the use of ESI-based LC/MS-MS analysis, which can be accompanied by unintended matrix effects (12, 13). Matrix effects could result from the molecular interactions of target compounds and matrix components, and might be proportional to the quantity of these matrix materials. It is difficult to predict the extent of matrix effects from chemical nature of the matrix components, and the degree to which different

TABLE 1. Quality Assurance Data of the Average %R and %RSD Obtained from the 15 ILCs and Their Respective Target Compounds, and That Obtained by Using IDMS Analysis compound name

N

%R

%RSD

corrected by

%R

%RSD

corrected %R

%RSD

positive ionization carbamazepine ciprofloxacin ibuprofen ibuprofen naproxen progesterone sulfamethazine sulfamethoxazole

78 34 76 76 76 4 78 34

73.3 263.7 89.1 89.1 87.2 67.6 59.6 71.2

16.3 62.3 12.0 12.0 10.5 10.4 26.7 17.6

2H -carbamazepine 10 13C 15N-ciprofloxacin 3 13C -ibuprofen 3 2H -ibuprofen 3 13C2H -naproxen 3 2H -progesterone 9 13C -sulfamethazine 6 13C -sulfamethoxazole 6

76.2 258.4 85.7 84.4 87.5 77.8 59.3 71.2

13.9 58.2 12.5 12.3 11.1 7.5 26.5 17.9

96.0 100.7 104.6 106.1 99.9 87.1 100.8 100.1

5.9 8.2 11.1 10.3 5.3 11.4 6.1 5.3

negative ionization acetaminophen clofibric acid diclofenac gemfibrozil ibuprofen ibuprofen indomethacin bisphenol A equilin estrone

78 34 76 78 78 78 34 72 72 72

32.3 100.8 114.6 108.5 111.7 111.7 67.9 87.7 89.7 96.3

29.7 20.3 15.6 11.3 18.9 18.9 31.7 20.1 17.1 19.1

2H -acetaminophen 4 2H -clofibric Acid 4 2H -diclofenac 4 2H -gemfibrozil 6 13C -ibuprofen 3 2H -ibuprofen 3 2H -indomethacin 4 2H -bisphenol A 16 2H -equilin 4 2H -estrone 4

32.7 102.2 109.8 101.2 110.2 106.2 67.9 82.3 86.1 96.6

44.5 20.3 19.3 11.9 19.4 20.6 30.8 20.4 18.4 22.0

102.9 99.3 106.0 107.6 101.9 102.4 100.1 107.5 104.8 100.8

12.4 20.4 15.1 7.9 10.2 8.2 9.7 10.5 8.7 9.7

compounds are affected. Matrix effects are demonstrated in the form of either signal suppression or enhancement, can originate from sample matrix components or just from a simple aberration of sample pH-induced conformation change of the target compounds, and must be addressed during the development and validation of methods. Matrix-matched calibration standards were demonstrated as a viable approach to resolve matrix effects (14) in LC/ MS-MS analysis. However, in a larger scale and/or longer term environmental study, the complexity of sample matrices collected from various locations at various times would make it impossible to find a representative, target-compound-free, matrix-matched sample for the preparation of matrixmatched calibration solutions. The standard addition method was suggested previously to compensate for matrix effects, but this approach would have increased the potential workload by a factor of 5, and so is not practical. The benefits and limitations of other operating strategies that could be used to circumvent matrix effects were discussed in a recent review (15). We therefore investigated the possibility using ILCs as IDQS to address matrix effects. Thirteen ILCs were added into the sample before the SPE extraction and were used to monitor the matrix effects on their respective native compounds in a sample’s original state. Three ILCs (13C3-ibuprofen, 13C6-sulfamethazine, and 2H16-bisphenol A) were added into the sample extract before LC/MS-MS analysis and used as IIS to correct for the final sample volume and injection volume variations. Using ILCs as IDQS allowed us to use their %Rs as quantitative correction factors for their respective native analogs (16). This provided us with the most accurate results, and resolves the matrix effect issue. As discussed, 15 of the 51 target compounds have corresponding ILCs that can be used to carry out the IDMS analysis. Quality assurance data of the average %R and relative standard deviation (RSD) obtained from these 15 target compounds, their respective ILCs, and results obtained by using IDMS analysis are listed in Table 1. The number of QC samples (N) collected varies depending on the time when a specific ILC became available commercially and was introduced into the method during the field study. From Table 1, one would note that IDMS analysis can improve %R and RSD for 14 of the 15 compounds. The improvement in %R and RSD can be corrected in a widespread range from 60 ( 26% (sulfamethazine) and 264 ( 58% (ciprofloxacin), respectively, to 100 ( 6% and 100 ( 8% in the positive ionization

mode. In the negative ionization mode, the %R and RSD can be corrected from 32 ( 30% (acetaminophen) and 115 ( 16% (diclofenac) to 103 ( 12% and 106 ( 15%. There are a limited number of QC data for progesterone (N ) 4) and the benefit of IDMS is not as obvious as for the other compounds. Clofibric acid was the only compound that had superior performance data before the use of IDMS analysis, recovered well in the field study, and benefited little from the IDMS approach. The availability of both 2H3-ibuprofen and 13C3-ibuprofen allowed us to add one before and the other after sample extraction to observe the nature of matrix effects in the raw sample and sample extract. From Table 1, %R and RSD data of 2H3- and 13C3-ibuprofen can be treated, within experimental error, as identical. This demonstrated that matrix materials could not be removed by the SPE sample preparation, were carried into sample extracts, and exerted the same effects on the target compound ibuprofen. This observation was also supported by the average %R and RSD of %R obtained from the IDMS analysis for sulfamethazine and bisphenol A using IIS 13C6-sulfamethazine and 2H16bisphenol A. Both ILCs were added after the SPE and used as internal standards, yet, average %R and RSD improved from 60 ( 26% to 100 ( 6% and 88 ( 20% to 108 ( 11%, respectively. Note that this conclusion applied only to ibuprofen, sulfamethazine, and bisphenol A. Additional validations using target compounds of different chemical characteristics, their respective ILCs, and different sample matrix spikes are required before a broader conclusion can be made. We also note that for the three aryl derivatives of propionic acid, i.e., ibuprofen, naproxen, and ketoprofen, three ILCs are available for ibuprofen and naproxen in forms of 13C -ibuprofen, 2H -ibuprofen, and 13C2H -naproxen. Due 3 3 3 to the similarity in chemical structures of these three compounds, one would expect the sample matrix to elicit similar effects on the three compounds, and allow the use of only one ILC to correct for the possible matrix effect. This would reduce analytical costs, offer a simpler method, and reduce the probability for errors. Using ILCs and IDMS analysis, %R and RSD obtained from the QC samples for these three compounds are listed in Table 2. One can conclude from Table 2 that IDMS analysis using any one of the three ILCs will give satisfactory QA data for these three VOL. 42, NO. 11, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4073

TABLE 2. Comparison of %R and %RSD Obtained from the IDMS Analysis for Ibuprofen, Ketoprofen, and Naproxen Using 2 H3-Ibuprofen, 13C3-Ibuprofen, or 13C2H3-Naproxen as Internal Standards %R, ibuprofen

without correction corrected by 2H3-ibuprofen corrected by 13C3-ibuprofen corrected by 13C 2H3-naproxen

%R, naproxen

average

%RSD

N

average

%RSD

N

average

%RSD

78 78 78 78

89.1 106.3 104.6 101.8

11.8 10.2 11.1 10.3

76 76 76 76

87.0 104.0 102.5 99.9

10.7 11.5 11.4 5.3

34 34 34 34

73.9 88.0 86.2 89.7

20.0 17.9 19.1 12.8

compounds, with the 13C2H3-naproxen demonstrating the best results for IDMS analysis. Ibuprofen is an acidic drug and was usually determined by negative ionization LC/MS-MS analysis in previous publications (17). Ibuprofen has a chemical structure that allows the MS analysis in both positive and negative ionization modes with the same product ion at m/z 161 due to the loss of COO from [M - H]- or COOH2 from [M + H]+. This allowed us to observe the matrix effects exerted on ibuprofen in these two ionization modes. Shown in Figure S3 are the %R of 2H3-ibuprofen (added before sample extraction) and 13C -ibuprofen (added into the sample extract before LC/ 3 MS-MS analysis) in both ionization modes obtained from five different source water samples. One can observe that the sample matrix would suppress signals of both isotope labeled ibuprofens to around 50% in the positive ionization mode and enhance both signals to over 200% in negative ionization mode. As %R of both 2H3-ibuprofen and 13C -ibuprofen are tracking well within the experimental 3 error, we conclude that matrix components were actually extracted with ibuprofen and affect intensities in both ionization modes studied. Without using IDMS analysis, the results obtained by negative and positive ionization detection for ibuprofen can vary by factors of 5-6. At present, this phenomenon was observed and validated for ibuprofen but might occur for other analytes that can be detected by both ionization modes. This is an important factor to consider when interpreting analytical results from different laboratories. With more and more ILCs available, the use of IDMS for the analysis of EOPs will result in more accurate results and superior operational efficiency and should be used whenever possible. Field Sample Results. Grab samples collected from the effluent of the Little River (LR) WWTP near the upper Detroit River (DR), confluence of the LR and DR, and downstream of the confluence were used to validate the method suitability for field sample analysis. Analytical results from Spring to Fall 2006 were compared to the results reported from prior studies (17, 18). Table S2 showed analytical results of the 22 EOPs that had an occurrence rate of more than 25% during field study. IDMS results were reported for target compounds with a corresponding ILC. Due to increased method sensitivity and selectivity, we were able to determine the presence of more EOPs than those documented in prior studies (17, 18). This is best illustrated by concentrations of PFOS and PFOA, lincomycin, sulfamethoxazole, ciprofloxacin, tetracycline, erythromycin, indomethacin, esterone, bisphenol A, monensin, tylosin, acetamidophenol, and lasaloid A, which were not reported before. Types of acidic and neutral PhACs observed were similar to those reported (17, 18) in 2006 and 2003. However, the ability to optimize the SPE extraction procedure and the use of IDQS for IDMS analysis allowed us to determine carbamazepine, naproxen, indomethacin, gemfibrozil, ibuprofen, and diclofenac with high confidence. This is also demonstrated by the supporting quality control and quality assurance data. The observation of esterone in the three sampling locations but not the other two more prevalent estrogens (i.e., estradiol and estriol) was also intriguing and cannot be explained from the QA data. 4074

9

%R, ketoprofen

N

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 11, 2008

Veterinary PhACs such as antibiotics (lincomycin and tetracycline) and growth promoters (tylosin, monensin, and lasaloid A) were also detected.

Acknowledgments We acknowledge the help and support from those involved in this project: co-operative students D. Lee, R. Saldanha, D. Paes, A. Lam, B. Edwards, and B. So; MOE colleagues Dr. R. Clement, R. Luniewski, B. Nguyen, S. Dong, and W. Hua; J. Kormos of University of Waterloo, and Dr. L. Lissemore of University of Guelph.

Note Added after ASAP Publication The Supporting Information originally published on April 15, 2008 has been revised. The new version was published on April 22, 2008.

Supporting Information Available Names of target compounds, peak labels (Figure 1), method performance data, MS-MS operating parameters, summary of QC data, and occurrence data. This information is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Danghton, C. G. Non-regulated water contaminants: emerging research. Environ. Impact Assess. Rev. 2004, 24, 711–732. (2) Richardson, S. D.; Ternes, T. A. Water analysis: emerging contaminants and current issues. Anal. Chem. 2005, 77, 3807– 3838. (3) Gilpin, R. K. Pharmaceutical and Related Drugs. Anal. Chem. 2003, 75, 2907–2918. (4) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals, Hormones, and Other Organic Wastewater Contaminants in U.S. Streams, 1999-2000: A National Reconnaissance. Environ. Sci. Technol. 2002, 36, 1202–1211. (5) Anderson, P. D.; D′Aco, V. J.; Shanahan, P.; Chapra, S. C.; Buzby, M. E.; Cunningham, V. L.; Duplessie, B. M.; Hayes, E. P.; Mastrocco, F. J.; Parke, N. J.; Rader, J. C.; Samuelian, J. H.; Schwab, B. W. Screening analysis of human pharmaceutical compounds in U.S. surface waters. Environ. Sci. Technol. 2004, 38, 838–849. (6) Petroviæ, M.; Hernando, M. D.; Dı´az-Cruz, M. S.; Barcelo´, D. Liquid chromatography-tandem mass spectrometry for the analysis of pharmaceutical residues in environmental samples: A review. J. Chromatogr., A 2005, 1067, 1–14. (7) Benijts, T.; Dams, R.; Gu ¨ nther, W.; Lambert, W.; De Leenheer, A. Analysis of estrogenic contaminants in river water using liquid chromatography coupled to ion trap based mass spectrometry. Rapid Commun. Mass Spectrom. 2002, 16, 1358–1364. (8) Hao, C.; Clement, R.; Yang, P. Liquid chromatography-tandem mass spectrometry of bioactive pharmaceutical compounds in the aquatic environmentsa decade’s activities. Anal. Bioanal. Chem. 2007, 387, 1247–1257. (9) Federal Register, U.S. Code of Federal Regulations, Part 136, Appendix B, 49 FR 43430, Oct. 26, 1984; 50 FR 694, 696, Jan. 4, 1985, as amended at 51 FR 23703, June 30, 1986 (10) Westerhoff, P.; Yoon, Y.; Snyder, S.; Wert, E. Fate of endocrinedisruptor, pharmaceutical, and personal care product chemicals during simulated drinking water treatment processes. Environ. Sci. Technol. 2005, 39, 6649–6663.

(11) Vanderford, B. J.; Pearson, R. A.; Rexing, D. J.; Snyder, S. A. Analysis of endocrine disruptors, pharmaceuticals, and personal care products in water using liquid chromatography/tandem mass spectrometry. Anal. Chem. 2003, 75, 6265–6274. (12) Kloepfer, A.; Quintana, J. B.; Reemtsma, T. Operational options to reduce matrix effects in liquid chromatography-electrospray ionization-mass spectrometry analysis of aqueous environmental samples. J. Chromatogr., A 2005, 1067, 153–160. (13) Hernando, M. D.; Petroviæ, M.; Ferna´ndez-Alba, A. R.; Barcelo´, D. Analysis by liquid chromatography-electrospray ionization tandem mass spectrometry and acute toxicity evaluation for β-blockers and lipid-regulating agents in wastewater samples. J. Chromatogr., A 2004, 1046, 133–140. (14) Schlu ¨ sener, M. P.; Spiteller, M.; Bester, K. Determination of antibiotics from soil by pressurized liquid extraction and liquid chromatography-tandem mass spectrometry. J. Chromatogr., A 2003, 1003, 21–28.

(15) Hao, C.; Zhao, X.; Yang, P. GC-MS and HPLC/MS analysis of bioactive pharmaceuticals and personal-care products in environmental matrices. Trends Anal. Chem. 2007, 26, 569– 580. (16) Vanderford, B. J.; Snyder, S. Analysis of pharmaceuticals in water by isotope dilution liquid chromatography/tandem mass spectrometry. Environ. Sci. Technol. 2006, 40, 7312–7320. (17) Metcalfe, C. D.; Miao, X. S.; Koenig, B. G.; Struger, J. Distribution of acidic and neutral drugs in surface waters near sewage treatment plants in the lower great lakes, Canada. Environ. Toxicol. Chem. 2003, 22 (12), 2881–2889. (18) Hua, W. Y.; Bennett, E. R.; Maio, X. S.; Metcalfe, C. M.; Letcher, R. J. Seasonality effects on pharmaceuticals and s-triazine herbicides in wastewater effluent and surface water from the Canada side of the upper Detroit River. Environ. Toxicol. Chem. 2006, 25 (9), 2356–2365.

ES7028125

VOL. 42, NO. 11, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4075