Selective Determination of a Group of Organic Compounds in

An on-line immunoaffinity extraction with liquid chromatography/membrane introduction mass spectrometry (IAE/LC/MIMS) method for the determination of ...
0 downloads 0 Views 76KB Size
Anal. Chem. 1998, 70, 931-935

Selective Determination of a Group of Organic Compounds in Complex Sample Matrixes by LC/ MIMS with On-Line Immunoaffinity Extraction Shi Ouyang, Yan Xu,* and Yong Hong Chen*

Department of Chemistry, Cleveland State University, Cleveland, Ohio 44115

An on-line immunoaffinity extraction with liquid chromatography/membrane introduction mass spectrometry (IAE/ LC/MIMS) method for the determination of BTEX compounds in complex sample matrixes is described. This method uses an immunoaffinity column (1 mm i.d. × 20 mm) for on-line sample cleanup and enrichment, a 5-µm C18 trapping column (2 mm i.d. × 20 mm) for analyte focusing, a 3-µm C18 analytical column (3.2 mm i.d. × 100 mm) for separation, and a membrane introduction mass spectrometer for quantitation. The immunoaffinity column was evaluated in terms of binding capacity, recovery, and enrichment factor. The method was optimized for the determination of BTEX compounds in a mixture of 30 volatile organic compounds, which showed no matrix interference and a dramatic improvement of the detection limit over that of the LC/MIMS method (up to 474-fold). This method was also used for the determination of BTEX compounds in several gasoline-contaminated water samples, and the results were compared with the EPA reference methods. One challenge in environmental analysis is to determine trace amounts of analytes in a complex sample matrix where concentrations of analytes are often close to the detection limit of the analytical system and background interferences adversely affect the analytical signals. To solve these problems, many sample cleanup and enrichment procedures have been developed including liquid-liquid extraction,1 purge and trap,2 solid-phase extraction,3,4 and supercritical fluid extraction.5,6 However, most extraction schemes are largely based on the solubilities of analytes and interferences in the two partitioned phases. Because an extraction phase is generally nonselective, coextraction of analytes and interferences may occur. This is particularly a problem when interferences are at high concentrations. With the advent of (1) Lim, H. K.; Andrenyak, D.; Francom, P.; Foltz, R. L.; Jones, R. T. Anal. Chem. 1988, 60, 1420. (2) USEPA Method 624. Fed. Regist. 1984, 141. (3) Gorecki, T.; Pawliszyn, J. Anal. Chem. 1995, 67, 3265. (4) Wittkamp, B. L.; Tilotta, D. C. Anal. Chem. 1995, 67, 600. (5) Camel, V.; Tambute, A.; Caude, M. J. Chromatogr. 1993, 642, 263. (6) Chester, T. L.; Pinkston, J. D.; Raynie, D. E. Anal. Chem. 1994, 66, 106R. (7) Van Emon, J. M.; Lopez-Avila, V. Anal. Chem. 1992, 64, 79A. (8) Thomas, D. H.; Beck-Westermeyer, M.; Hage, D. S. Anal. Chem. 1994, 66, 3823. (9) De Frutos, M.; Regnier, F. E. Anal. Chem. 1993, 65, 17A. (10) Hage, D. S.; Thomas, D. H.; Beck, M. S. Anal. Chem. 1993, 65, 1622. (11) Phillips, T. M. LCLC 1987, 3, 962. (12) Ouyang, S.; Chen, Y. H.; Xu, Y. Anal. Chim. Acta 1997, 337, 165. S0003-2700(97)01039-1 CCC: $15.00 Published on Web 01/27/1998

© 1998 American Chemical Society

immunoaffinity extraction (IAE) in environmental sample pretreatment,7,8 it is possible to circumvent these problems. Immunoaffinity (IA) is the attraction between the molecules of analyte and antibody, which is a function of the sum of the short-range, noncovalent, intermolecular forces. An antibody is a protein molecule that binds analyte with exquisite specificity. The binding of analyte to antibody is a reversible reaction, and the equilibrium of the reaction favors analyte-antibody association if the spatial complementarity between molecules is good and the forces binding molecules together are relatively strong and stable. An antibody can also bind to one or more analytes with a structure similar to the one initiating the immune response due to the heterogeneity of immunological recognition.9 The cross-reactivity of an antibody may be a negative feature when the aim is to determine a single analyte species, but it is extremely valuable when the aim is to isolate a group of analytes of similar structure. In IAE, the sample extract is passed through a column containing an immobilized antibody and the analyte of analytical interest is retained and enriched by the specific antibody. By altering the chemical conditions of the eluent, the analyte can be removed from column and quantitated in a subsequent analytical procedure.10,11 Usually, a complete IAE takes only a few minutes, and generates no organic solvent waste. An IA column can be reused over a year,8 and the cost incurred is minimal. In our previous study, a method of liquid chromatography with membrane introduction mass spectrometry (LC/MIMS) was developed for the determination of volatile organic compounds (VOCs) in water.12 Methanol-water was used as the carrier fluid in MIMS, which dramatically improved the dynamic response of the silicone membrane probe. After chromatographic separation, a mixture of 18 VOCs was quantitated using MIMS by means of the retention times and mass-to-charge ratios. The method was rapid and required no pretreatment for complex samples. Nevertheless, the method may have difficulty in identifying analytes at very low concentrations or may require intensive chromatographic development if the sample matrix became more complex. In this study, we developed an IAE/LC/MIMS method to exploit the positive attribute of the cross-reactivity of antibody for capturing a group of analytes of similar structure, the selective analyte enrichment feature of IAE for achieving lower detection limit, and the quantitative capability of LC/MIMS for analyzing VOCs in water. BTEX compounds (i.e., benzene, toluene, ethylbenzene, and xylene isomers), the most pervasive group of Analytical Chemistry, Vol. 70, No. 5, March 1, 1998 931

contaminants found in environmental waters,13 were chosen as the target analytes. Anti-toluene polyclonal antibody from rabbit serum was used for preparing the IA column to bind all BTEX compounds. The binding capacity of the IA column, the effect of flow rate on the recovery of IAE, and the enrichment factor of the IA column were studied. The method developed (IAE/LC/ MIMS) was used for the determination of BTEX compounds in a mixture of 30 VOCs and its performance was compared with that of LC/MIMS. It was clearly demonstrated that IAE/LC/MIMS is vastly superior to LC/MIMS for analyzing BTEX compounds in a complex sample matrix. Finally, the method developed was applied to the determination of several gasoline-contaminated water samples, and the results were compared with those of the EPA reference methods. EXPERIMENTAL SECTION Materials. Rabbit anti-toluene neat serum (Catalog No. K82313R, Lot No. 5H2215) was obtained from Biodesign International (Kennebunk, ME). ImmunoPure (A/G) IgG purification kit (Catalog No. 44902, Lot No. 94071270) was from Pierce (Rockford, IL). Nucleosil 300-7 (300-Å pore size, 7-µm particle diameter) was from Alltech (Deerfield, IL). All 30 VOC standards (100 ppm) in methanol were purchased from ChemService (West Chester, PA) (i.e., benzene, toluene, ethylbenzene, p-xylene, 1,2,4trimethylbenzene, 1,3,5-trimethylbenzene, naphthalene, acenaphthene, 2-chlorotoluene, 3-chlorotoluene, 4-chlorotoluene, 1,4dichlorobenzene, phenanthrene, fluorene, n-butylbenzene, 1,2dichlorobenzene, isopropylbenzene, 1,2-dibromo-3-chloropropane, sec-butylbenzene, cis-1,3-dichloropropene, 1,1-dichloropropene, 1,2,3-trichloropropane, carbon tetrachloride, chlorobenzene, npropylbenzene, chloroform, bromodichloromethane, R-chlorotoluene, 1,2,4-trichlorobenzene, and 1,3,5-trichlorobenzene). HPLCgrade sodium phosphate and sodium acetate were obtained from Fisher Scientific (Pittsburgh, PA). (3-Glycidoxypropyl)trimethoxysilane, adipic dihydrazide, and methanol (99.9+%, PRA grade) were from Aldrich (Milwaukee, WI). The other chemicals were from Fisher Scientific. All solutions were prepared using deionized water from a NANOpure system (Barnstead, Dubuque, IA). Instrumentation. The schematic diagram of the IAE/LC/ MIMS system is shown in Figure 1. It consisted of a Dionex GP40 gradient pump (pump 1) (Dionex, Sunnyvale, CA), a sample injection valve with a 1-mL loop (valve 1) (Rheodyne, Cotati, CA), an immunoaffinity column (column 1), a PM-80 pump (pump 2) (Bioanalytical Systems, West Lafayette, IN), a Dionex LC-5 switch valve (valve 2); a C18 trapping column (column 2) (5 µm, 2 mm i.d. × 20 mm, Upchurch Scientific, Oak Harbor, WA) and a C18 separation column (column 3) (3 µm, 3.2 mm i.d. × 100 mm, Bioanalytical Systems); a custom-built membrane probe (similar probes are available at MIMS Technology, Palm Bay, FL) which had been described elsewhere,12 a Finnigan TSQ45 mass spectrometer (Finnigan Mat, San Jose, CA), and a VECTOR/TWO data acquisition unit (ProLab Resources, Madison, WI). Purification of Antiserum. About 1 mL of anti-toluene rabbit serum was purified and desalted using the ImmunoPure (A/G) IgG purification kit according to the procedures provided by manufacturer. The resulting antibody solution was lyophilized by a freeze-dryer (Labconco Corp., Kansas City, MO), and the lyophilized antibody was stored at -18 °C for further use. (13) Clement, R. E.; Yang, P. W.; Koester, C. J. Anal. Chem. 1997, 69, 251R.

932 Analytical Chemistry, Vol. 70, No. 5, March 1, 1998

Figure 1. Schematic diagram of the IAE/LC/MIMS system: column 1, immunoaffinity column; column 2, C18 trapping column; column 3, C18 separation column.

Oxidation of Antibody.14 About 10 mg of the anti-toluene dry powder was dissolved in 1 mL of acetate buffer (0.1 M sodium acetate-0.15 M sodium chloride, pH 5.5). To this solution, 25 µL of 0.4 M sodium periodate was added; the oxidation reaction was allowed to proceed in the dark at room temperature for 20 min. The oxidized antibody was then separated from the oxidant using a Sephadex G-25-80 column which had been previously equilibrated in an acetate buffer [0.1 M sodium acetate, 0.15 M sodium chloride, 0.02% (w/v) sodium azide, pH 4.5] and detected at 254 nm. After desalting and lyophilization, the dry powder of the oxidized antibody was kept in a freezer before use. Preparation of the Immunoaffinity Column. The silica support used in the immunoaffinity column was prepared using the procedures reported by Ruhn and co-workers15 with some modifications. In the method, 1 g of Nucleosil 300-7 was placed in 8.5 mL of 0.1 M sodium acetate buffer (pH 5.5) and degassed by sonication under vacuum aspiration for 15 min. Then, 0.20 mL of (3-glycidoxypropyl)trimethoxysilane was added and the solution was shaken for 5 h at 90 °C. The epoxy silica support was washed three times each with deionized water and sulfuric acid solution (pH 3.0). The support was later suspended in 100 mL of sulfuric acid solution (pH 3.0) and shaken for 90 min at room temperature. This resulted in the diol-bonded silica. After being washed two times each with deionized water, methanol, and ether, the diol-bonded silica was dried overnight under vacuum aspiration at room temperature. The aldehyde-activated silica was prepared by suspending the diol-bonded silica in 20 mL of 90% (v/v) acetic acid solution containing 1.2 g of sodium periodate. The mixture was sonicated under vacuum aspiration for 15 min and shaken for 2 h at room temperature. The aldehyde silica produced was washed three times with deionized water to remove any remaining periodate. To this aldehyde silica, 25 mL of 0.1 M (14) Kaneki, N. Xu, Y.; Kumari, A.; Halsall, H. B.; Heineman, W. R.; Kissinger, P. T. Anal. Chim. Acta 1994, 287, 253. (15) Ruhn, P. F.; Garver, S.; Hage, D. S. J. Chromatogr., A 1994, 669, 9.

phosphate buffer (pH 5.0) containing 3 g of adipic acid dihydrazide was added. The mixture was shaken for 90 min at room temperature and washed three times with 0.1 M phosphate buffer (pH 7.0). After coupling to dihydrazide, the residual aldehyde groups were reduced by adding 20 mL of 0.1 M phosphate buffer (pH 8.0) containing 5 g of NaBH4. The mixture was shaken for 90 min at room temperature. The resulting dihydrazide-activated silica was washed three times with 0.1 M phosphate buffer (pH 7.0) and stored in 0.1 M phosphate-0.02% NaN3 buffer (pH 7.0) at 4 °C. The antibody-bonded silica support was prepared by mixing 0.3 g of the dihydrazide-activated silica with 5 mg of the oxidized antibody dry powder in 2 mL of 0.1 M phosphate buffer (pH 7.0) for 24 h at 4 °C. After the immobilization of antibody, the silica support was washed three times each with 0.1 M phosphate buffer-2 M sodium chloride (pH 7.0) and deionized water, and packed into a 1 mm i.d. × 2 cm column (Upchurch Scientific) by a standard slurry packer (Alltech) at 2000 psig. The immunoaffinity column was stored in 0.1 M phosphate-0.02% NaN3, pH 7.0, at 4 °C until further use. Operational Procedure. The IAE/LC/MIMS system (Figure 1) was operated by the following steps. The first step was the equilibration of columns. Valve 2 was set at the “load” position (the dotted line), eluent 1 (0.1 M phosphate, pH 7.0) was pumped through column 1 at 0.5 mL/min by pump 1, and eluent 3 (80% methanol in deionized water) was pumped through column 2 and 3 at 0.2 mL/min by pump 2. The second step was the loading of sample and immunoaffinity extraction. After 3 min of the above equilibration, 1 mL of analyte solution was injected into column 1 via valve 1. During this step, analytes of interest were extracted onto column 1 and sample matrix was washed out to the waste. This step took 5 min to complete including 2 min for sample delivery and 3 min for antibody-analyte reaction. The third step was the elution of analytes from column 1 and the “focus” of these analytes on column 2. Valve 2 was switched to the “inject” position (the solid line), and eluent 2 (0.05 M phosphate, pH 2.5) was pumped through column 1 and 2 at 0.5 mL/min by pump 1. The eluted analytes from column 1 were “focused” on column 2. The last step was the separation and MIMS detection. Valve 2 was switched back to the “load” position, and the analytes were quickly desorbed from column 2, separated in column 3, and detected by MIMS using eluent 3 as the mobile phase at 0.2 mL/min. The complete separation and detection required 16 min. During the last step, the three aforementioned steps of the next sample could proceed; therefore, there was no waiting period for the on-line IAE after initial analysis. The operating conditions of the mass spectrometer and the pretreatment of the membrane probe were described in detail in ref 12. RESULTS AND DISCUSSION Evaluation of the IA Column. (1) Binding Capacity. The binding capacity of an IA column gives the maximum amount of analytes that can be isolated from the sample matrix. It is directly proportional to the total amount of active antibody covalently bonded on column packing material. The binding capacity of the IA column used in this work was determined by injecting a series of toluene standards (0.1-10 ppm) to column through a 1-mL sample loop at a mobile-phase flow rate of 0.5 mL/min for 5 min (2 min for sample delivery to column, and 3 min for the antibody-

Figure 2. Binding capacity of the IA column. The experimental procedure was described under Operational Procedure. The instrumentation was the same as that in Figure 1 without column 3. Signals of toluene in water (0.1-10.0 ppm) were recorded at m/z ) 91.

antigen reaction). Toluene was then eluted and detected by LC/ MIMS using the procedure described under Operational Procedure. The peak areas of toluene were plotted against the concentrations and resulted in a so-called “saturation curve” (Figure 2). From this curve, a saturation concentration of 2.3 ppm could be found, which corresponded to a total column binding capacity of 2.5 × 10-8 mol of toluene, and a linear response range up to 2.0 ppm or 2.2 × 10-8 mol was obtained. For a typical environmental water analysis, BTEX levels are usually below 30 ppb, which will use only about 1.5% or less of column binding capacity. In case of high analyte concentrations, a simple sample dilution may solve the problem. With the binding capacity (2.5 × 10-8 mol of toluene), the estimated total surface area of solid support (3.48 m2), and the estimated surface area of a bivalent IgG molecule (1.00 × 10-18 m2), the surface coverage of the active antibody (monolayer) was calculated to be 21.6% using our immobilization procedure. (2) Effect of Flow Rate on the Antibody-Analyte Reaction. The kinetic study of the antibody-analyte reaction was performed by injecting 1 mL of 50 ppb toluene in water into the IAE/LC/ MIMS system at various flow rates. The recovery of toluene under each flow rate was determined by comparing the peak area of toluene (m/z ) 91) with that measured without the IA column. The results are shown in Figure 3. An increase in recovery from 25 to 95% was observed as the flow rate decreased from 2.0 to 0.20 mL/min. However, the sample loading time for delivering 1 mL of analyte from the sample loop to column also increased by 10-fold, from 0.5 to 5 min. As a compromise between the assay time and the recovery, a flow rate of 0.5 mL/min was chosen for the subsequent work, at which the recovery was 92%. Figure 3 also reveals that as the flow rate increased from 0.20 to 2.0 mL/ min, the back pressure of the IA column increased from 400 to 1540 psi. (3) Lifetime of the IA Column. The IA column was always used at room temperature. After use, column was rinsed with a 0.1 M phosphate buffer at pH 7.0 for 10 min and stored at 4 °C in the same buffer containing 0.02% sodium azide. It had been reported that the lifetime of an IA column could be as long as 1 year or more.8 In our case, after 9 months of usage, no apparent loss of column binding capacity was observed. In fact, for the entire study, only one column was used. Analytical Chemistry, Vol. 70, No. 5, March 1, 1998

933

Figure 3. Recoveries of IA extraction and column pressures vs the loading flow rates. Toluene in water (50 ppb) was used as the test analyte. The recovery of toluene under each flow rate was determined by comparing the peak areas obtained under the experimental conditions with and without the IA column. Other experimental conditions were the same as those in Figure 2.

Analytical Performance of the Method. (1) Extraction of the BTEX Compounds. In this work, the IA column was used as a means of on-line sample cleanup and enrichment. According to the manufacturer, the anti-toluene serum has not only a 100% reactivity to toluene but also 90 and 25% cross-reactivities to xylene(s) and benzene. Further, our test data showed that column prepared by our procedure also has a very high cross-reactivity to ethylbenzene. Due to the specificity and selectivity of the IA column, we were able to use column to extract and enrich all BTEX compounds from complex sample matrixes. To illustrate this usage, an aqueous mixture containing 30 organic compounds (i.e., benzene, toluene, ethylbenzene, p-xylene, 1,2,4-trimethylbenzene, 1,3,5-trimethylbenzene, naphthalene, acenaphthene, 2-chlorotoluene, 3-chlorotoluene, 4-chlorotoluene, 1,4-dichlorobenzene, phenanthrene, fluorene, n-butylbenzene, 1,2dichlorobenzene, isopropylbenzene, 1,2-dibromo-3-chloropropane, sec-butylbenzene, cis-1,3-dichloropropene, 1,1-dichloropropene, 1,2,3-trichloropropane, carbon tetrachloride, chlorobenzene, npropylbenzene, chloroform, bromodichloromethane, R-chlorotoluene, 1,2,4-trichlorobenzene, and 1,3,5-trichlorobenzene) at 50 ppb each was used to evaluate the effectiveness of the IA column. Without IA extraction, quantitation of BTEX compounds in the mixture by LC/MIMS encountered difficulties in both separation (Figure 4A) and mass identification (Figure 4B). Although inefficient separation may be improved by altering the separation conditions, a longer analysis time will be required to resolve all analytes. However, quantitation of BTEX compounds in the mixture could be easily achieved by IAE/LC/MIMS with much shorter analysis time. As shown in Figure 4C, the IA column excluded all non-BTEX compounds in the mixture and allowed only BTEX compounds to be analyzed by LC/MIMS. Since the extraction and the regeneration of column could take place concurrently with the LC/MIMS analysis, the total analysis time per sample was 16 min after the initial analysis. (2) Enrichment of Analytes. An IA column can selectively capture analyte(s) from other components in the solution. After the sample matrix is washed away, the captured analyte can then be released from the adsorbent by dissociating the antibodyanalyte complex with a buffer solution. At a given concentration, 934 Analytical Chemistry, Vol. 70, No. 5, March 1, 1998

Figure 4. (A) Total ion chromatogram and (B) selected ion chromatogram (m/z ) 91) of LC/MIMS of a mixture of 30 VOCs. (C) Total ion chromatogram of IAE/LC/MIMS for BTEX compounds in a mixture of 30 VOCs. The experimental procedures were described under Operational Procedure and the instrumentation was the same as that in Figure 1 without column 1 (A and B) and with column 1 (C). The concentration of each VOC in water was 50 ppb.

the amount of captured analyte is directly proportional to the volume of sample until it reaches the binding capacity of the IA column. Figure 5 shows the signal responses (m/z ) 91) of 5 ppb toluene in water solution with various injected sample volumes (e.g., 0.100, 1.00, 10.0, and 100 mL, respectively). Sample loops were used to deliver smaller sample volumes (0.100 and 1.00 mL). Direct pumping of sample at a flow rate of 0.5 mL/min was applied for larger sample volumes (10.0 and 100 mL). In these experiments, because there was only one compound in the sample solution, the separation column (column 3 in Figure 1) was not used. As the volumes of sample injected increased from 0.100 to 100 mL, the peak areas integrated from Figure 5 increased from 132 (count × minute) to 62 571 (count × minute), a 474-fold increase. In other words, the detection limit of the method could be improved by 474-fold. From the above examples, it is illustrated that the on-line IAE is not only practical for sample cleanup but also unrivaled for improving the detection limit of the method. (3) Calibration Plots. The calibration plots of benzene, toluene, ethylbezene, and p-xylene were constructed using the analyte concentrations versus the peak areas of the signal responses. As shown in Figure 6, linear dynamic ranges were 3 orders of magnitude (3-300 ppb). The correlation coefficients of linear regression analyses over five data points were greater than or equal to 0.997. The method had a detection limit (signal/ noise ) 3) of 0.3 ppb for all the BTEX compounds with a 1.00-mL

Table 1. BTEX Levels in Four Gasoline-Contaminated Water Samples sample no. 1a 2a 3b 4b

Figure 5. The selected ion chromatograms of 5 ppb of toluene in water at various injection volumes. Other experimental conditions were the same as those in Figure 2.

Figure 6. Calibration plots for the BTEX compounds. Toluene (0), peak area (PA) ) 77.0C (ppb) + 332 (ppb); p-xylene (2), PA ) 61.4C - 202; ethylbenzene (3), PA ) 48.9C - 95.7; and benzene (O), PA ) 36.5C + 11.8. The experimental conditions were the same as those in Figure 4C. Each datum point was the average reading of three measurements.

sample loop. This detection limit could be greatly improved by increasing the injected sample size. Environmental Analysis. BTEX compounds are common contaminants in groundwater and surface water. They enter the environment through gasoline leakage during its transportation, storage, and usage. Some of the BTEX compounds are suspected carcinogens. Therefore, the determination of BTEX compounds in water is important for pollution monitoring. In this work, two gasoline-contaminated groundwater samples were collected according to the EPA procedures, and two artificial water samples were prepared in our laboratory. These samples were analyzed by the IAE/LC/MIMS method and by either the CAS-CHEM Laboratories (Canton, OH) using EPA method 602/ 802 or the Electro-Analytical Group (Mentor, OH) using EPA method 524.2. The results were summarized in Table 1. It was noted that there was a discrepancy in the results of samples 1 and 2 with the IAE/LC/MIMS and the reference method. The

method IAE/LC/MIMS EPA M602/802c IAE/LC/MIMS EPA M602/802 IAE/LC/MIMS EPA M524.2d IAE/LC/MIMS EPA M524.2

benzene toluene ethylbenzene xylene, (mg/L) (mg/L) (mg/L) total (mg/L) 1.30 0.730 15.6 8.80 59.5 72.4 107 96.7

0.00100