MS Method for the Simultaneous Determination of GC

Aug 10, 2009 - Water pollution by organochlorine pesticides (OCPs) is considered as an analytical challenge, since these persistent and nonbiodegradab...
1 downloads 0 Views 171KB Size
Download PDF
Anal. Chem. 2009, 81, 7373–7378

Single-Step LC/MS Method for the Simultaneous Determination of GC-Amenable Organochlorine and LC-Amenable Phenoxy Acidic Pesticides G. Famiglini, P. Palma, V. Termopoli, H. Trufelli, and A. Cappiello* Dipartimento di Scienze Geologiche, Tecnologie Chimiche e Ambientali, Universita` degli Studi di Urbino “Carlo Bo”, Piazza Rinascimento 6, 61029 Urbino, Italy Water pollution by organochlorine pesticides (OCPs) is considered as an analytical challenge, since these persistent and nonbiodegradable pollutants are not amenable by liquid chromatography coupled to atmospheric pressure ionization mass spectrometry (LC/API-MS). This represents a significant constraint in multiresidue analysis of real samples, when high polar, poorly volatile compounds are present as well. This paper reports the development of an innovative single-step method for the simultaneous determination of OCPs and polar pesticides belonging to the class of phenoxy acids in water samples. The method is based on an off-line solid-phase extraction (SPE) procedure with Carbograph 4 followed by liquid chromatography coupled to a direct electron ionization mass spectrometer (LC/direct-EI-MS). The direct-EI capability of acquiring high-quality EI spectra and operation in selected ion monitoring mode allowed a precise quantification of OCPs and phenoxy acids in a single chromatographic run without derivatization. The instrumental response was characterized by excellent sensitivity, linearity, and precision. The SPE recovery rates in river water gave values equal or better than 80% for most of the compounds. The method limits of detection (LODs) span from 0.002 to 0.052 µg/L, allowing the detection of the selected pesticides at the limits required by the European Union (EU) legislation for drinking water. During the last few decades the development of sensitive and multiclass methods for the determination of pesticides in water has become a major issue, due to the environmental impact of phytosanitary products and the strict legal requirements for water quality. Pesticide residue tolerances in drinking water were set by the European Union (EU) Commission to 0.1 µg/L for an individual compound and 0.5 µg/L for the sum of pesticide residues.1 Multiresidue analytical approaches are preferred against single group analysis because they provide wider knowledge about the occurrence, removal, partition, and ultimate fate of pollutants in the aquatic environment. However, a major drawback in the development of a rapid and universal analytical method for pesticide monitoring is represented by their broad spectrum of chemical and physical properties. On the basis of the compilation * To whom correspondence should be addressed. Phone: +390722303344. Fax: +390722303311. E-mail: [email protected]. (1) Council Directive 98/83/EC. 10.1021/ac9008995 CCC: $40.75  2009 American Chemical Society Published on Web 08/10/2009

of the British Crop Protection Council, approximately 881 active substances are formulated in phytosanitary products.2 These compounds belong to more than 100 different classes, going from persistent, not easily degradable pesticides (e.g., organochlorines) to more polar, readily degradable analytes (e.g., sulfonylureas, N-methylcarbamates). Gas chromatography (GC) and liquid chromatography (LC) coupled to mass spectrometry (MS) are the most widely used techniques for monitoring pesticides residues in food and environmental compartments.3-8 Most of pesticides that normally reach the aquatic environment are characterized by low volatility and medium-high polarity. This makes liquid chromatography coupled to atmospheric pressure ionization tandem mass spectrometry (LC/API-MS/MS) the most appropriate approach for trace-level determination of pesticides in water, leading to satisfactory results for quantification and confirmation purposes.3-7 During the last decades the introduction of a new generation and relatively cheap MS instrumentation has filled the gap existing between GC/MS and LC/MS and has further stimulated the scientific community in developing multicomponent methods (MCMs) able to determine as many pesticides as possible in aqueous matrixes.9-12 However, there are still several analytical limitations to overcome for developing a rapid and universal multicomponent LC/MS method. One of the major problems in the quantitative analysis using LC/API-MS/MS is that matrixdependent response suppression or enhancement may occur, (2) Tomlin, C. D. S. 2006sThe Pesticide Manual, 14th ed.; BCP Publications: Alton Hampshire, U.K., 2006. (3) Hogenboom, A. C.; Niessen, W. M. A.; Brinkman, U. A. Th. J. Sep. Sci. 2001, 24, 331–354. (4) Reemtsma, T. J. Chromatogr., A 2003, 1000, 477–501. (5) Pico´, Y.; Blasco, C.; Font, G. Mass Spectrom. Rev. 2004, 23, 45–85. (6) Alder, L.; Greulich, K.; Kempe, G.; Vieth, B. Mass Spectrom. Rev. 2006, 25, 838–865. (7) Kuster, M.; Lo´pez de Alda, M. J.; Barcelo´, D. Mass Spectrom. Rev. 2006, 25, 900–916. (8) Ferrer, I.; Thurman, E. M.; Zweigenbaum, J. A. Rapid Commun. Mass Spectrom. 2007, 21, 3869–3882. (9) Sancho, J. V.; Pozo, O.; Hernandez, F. Analyst (Cambridge, U.K.) 2004, 129, 38–44. (10) Kampioti, A. A.; Borba da Cunha, A. C.; Lo´pez de Alda, M.; Barcelo´, D. Anal. Bioanal. Chem. 2005, 382, 1815–1825. (11) Baugros, J. B.; Giroud, B.; Dessalces, G.; Grenier-Loustalot, M. F.; Olive´, C. C. Anal. Chim. Acta 2008, 607, 191–203. (12) Kuster, M.; Lo´pez de Alda, M. J.; Barata, C.; Raldua, D.; Barcelo´, D. Talanta 2008, 75, 390–401. (13) Rodil, R.; Quintana, J. B.; Lo´pez-Mahía, P.; Muniategui-Lorenzo, S.; PradaRodrı´guez, D. J. Chromatogr., A 2009, 1216, 2958–2969.

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

7373

Table 1. Target Compounds and Their Physical-Chemical Properties compound (abbreviation)

CAS no.

water solubility (mg/L) at 25 °Ca

log P a

preferred method of analysis for underivatized compoundsb

dicamba bromoxynil 2,4-dichlorophenoxyacetic acid (2,4-D) mecoprop 2-methyl-4-chlorophenoxybutyric acid (MCPB) 2,4-dichlorophenoxybutyric acid (2,4-DB) silvex lindane methoxychlor β-endosulfan R-endosulfan dichlorodiphenyldichloroethane (o,p′-DDD) p,p′-dichlorodiphenyldichloroethane (p,p′-DDD) heptachlor o,p′-dichlorodiphenyltrichloroethane (o,p′-DDT) p,p′-dichlorodiphenyltrichloroethane (p,p′-DDT) hexachlorbenzene (HCB) o,p′-dichlorodiphenylchloroethane (o,p′-DDE) aldrin

1918-00-9 1689-84-5 94-75-7 7085-19-0 94-81-5 10433-59-7 93-72-1 58-89-9 72-43-5 33213-65-9 959-98-8 53-19-0 72-54-8 76-44-8 789-02-6 50-29-3 118-74-1 72-55-9 309-00-2

6500 130 620 620 44 46 71 7.30 0.10 0.40 0.51 0.10 0.09 0.18 0.09 0.01 0.01 0.04 0.02

2.14 3.39 2.62 2.94 3.50 3.60 3.68 3.72 5.08 3.83 3.83 6.02 6.02 6.10 6.91 6.91 5.73 6.02 6.50

LC/ESI-MS LC/ESI-MS LC/ESI-MS LC/ESI-MS LC/ESI-MS LC/ESI-MS LC/ESI-MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS GC/MS

a

Ref 2. b Ref 6.

leading to the matrix effects.14,15 Furthermore all the approaches aimed at monitoring a large number of pesticides by LC/MS, analogously to those by GC/MS, are not really MCMs, since they cannot be addressed to the simultaneous monitoring of polar and nonpolar compounds.6,7 In this context the contamination of water resources by organochlorine pesticides (OCPs) is considered as an analytical challenge, since these persistent and nonbiodegradable pollutants are not amenable by LC/API-MS. This represents a significant constraint in multiresidue analysis of real samples, when high polar, poorly volatile compounds are present as well.6 For this reason, multistep GC/MS and LC/API-MS methods are mandatory when MCMs that include trace analysis of OCPs and polar pesticides need to be developed.6,7,11 In our previous publication, an innovative approach for the determination of OCPs in water samples by LC/MS was presented.16 The method was based on a solid-phase extraction (SPE) followed by nanoscale LC coupled to a mass spectrometer equipped with a direct electron ionization interface (direct-EI). Direct-EI is a miniaturized interface for efficiently coupling a liquid chromatograph with an EI mass spectrometer. A detailed description of the functioning and performances of the interface is presented elsewhere.17-21 The results obtained applying the direct-EI interface in the analysis of OCPs demonstrated that LC/ EI-MS has a great potential as a powerful tool for the development of MCMs aimed at the simultaneous monitoring of a large number of pesticides with different physical and chemical properties. Therefore, the purpose of this work is to take up the analytical challenge of performing a single-step, trace-level LC/MS detection of OCPs together with several pesticides belonging to the class of phenoxy acids. Because of their high polarity and low volatility, phenoxy acid pesticides are not directly amenable to GC analysis without derivatization, thus making LC coupled to electrospray ionization mass spectrometry (ESI-MS) the method of choice for the detection and quantification of these compounds in the water compartment.5,7,13,22 These compounds can be easily detected using LC/EI-MS, as demonstrated by our previous publication using a microparticle beam interface.23,24 The sample preparation procedure is another crucial step because it must ensure the 7374

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

required trace enrichment for both pesticides classes, maintaining a satisfactory recovery for all analytes despite their different polarity. There is no doubt that SPE has become the most used method for carrying out the simultaneous extraction and concentration of contaminants from a wide range of polarity.25 For this reason, prior to LC/direct-EI-MS, an off-line SPE protocol was developed in order to ensure an effective sample enrichment for all the target analytes. A total of 7 and 12 pesticides belonging, respectively, to the class of phenoxy acids and to the OCPs were chosen as target compounds. The method was applied to the analysis of river water samples. From our knowledge, this is the first single-step method aimed at the simultaneous analysis of OCPs and acidic pesticides by LC/MS. MATERIALS AND METHODS Chemicals and Reagents. HPLC grade methylene chloride, tetrahydrofuran (THF), acetonitrile, and methanol were purchased from VWR International (Milan, Italy). Water was purified using a direct-Q 3 UV system from Millipore Corp. (Bedford, MA). The selected pesticides (Table 1) were all purchased from SigmaAldrich (Milan, Italy). Individual stock standard solutions at a concentration of 10 000 mg/L were prepared in tetrahydrofuran. (14) Niessen, W. M. A.; Manini, P.; Andreoli, R. Mass Spectrom. Rev. 2006, 25, 881–899. (15) Marı´n, J. M.; Gracia-Lor, E.; Rancho, J. V.; Lo´pez, F. J.; Herna´ndez, F. J. Chromatogr., A 2009, 1216, 1410–1420. (16) Famiglini, G.; Palma, P.; Pierini, E.; Trufelli, H.; Cappiello, H. Anal. Chem. 2008, 80, 3445–3449. (17) Cappiello, A.; Famiglini, G.; Mangani, F.; Palma, P. J. Am. Soc. Mass Spectrom. 2001, 13, 265–273. (18) Cappiello, A.; Famiglini, G.; Palma, P. Anal. Chem. 2003, 75, 497A–503A. (19) Cappiello, A.; Famiglini, G.; Palma, P.; Siviero, A. Mass Spectrom. Rev. 2005, 24, 978–989. (20) Cappiello, A.; Famiglini, G.; Pierini, E.; Palma, P.; Trufelli, H. Anal. Chem. 2007, 79, 5364–5372. (21) Cappiello, A.; Famiglini, G.; Palma, P.; Termopoli, V.; Trufelli, H. Anal. Chem. 2008, 80, 9343–9348. (22) Di Corcia, A.; Nazzari, M.; Rao, R.; Saperi, R.; Sebastiani, E. J. Chromatogr., A 2000, 878, 87–89. (23) Cappiello, A.; Famiglini, G.; Bruner, F. Anal. Chem. 1994, 66, 1416–1423. (24) Cappiello, A.; Famiglini, G. Anal. Chem. 1995, 67, 412–419. (25) Pichon, V. J. Chromatogr., A 2000, 885, 195–215.

For LC/MS analysis, the stock solutions were mixed and diluted with a water-THF (50:50, v/v) solution acidified with formic acid (1%, v/v) at the required working concentrations. Sample Collection and Preparation. Recovery experiments were performed on river and Milli-Q water samples. River water samples were collected in glass bottles and stored in the dark at 4 °C. SPE was performed within 24 h. Sample filtration was carried out using an HA MF-Millipore membrane filter (0.45 µm nominal) (Millipore Corp., Bedford, MA). Prior to SPE extraction the water samples were fortified in order to achieve the following pesticides concentrations: 1.25 µg/L (bromoxynil, dicamba, 2,4-D, 2,4-DB, MCPB, mecoprop, silvex), 0.65 µg/L (aldrin, o,p′-DDD, p,p′-DDD, o,p′-DDE, o,p′-DDT, p,p′-DDT, HCB, heptachlor, lindane, methoxychlor), 1.00 µg/L (β-endosulfan), and 2.00 µg/L (R-endosulfan). SPE was carried out using 6 mL cartridges packed with 500 mg of Carbograph 4 (LARA, Rome, Italy). The cartridges were positioned on a 12-port Visiprep SPE vacuum manifold (Supelco, Bellefonte, PA). The cartridges were conditioned by passing 10 mL of a solution of methylene chloride-methanol (80:20, v/v) acidified with formic acid 50 mmol/L, 2 mL of methanol, 10 mL of an HCl-acidified Milli-Q water (pH 2), and finally 10 mL of Milli-Q water. Water samples (2 L) were forced through the traps at a flow rate ranging from 12 to 15 mL/min. Before extraction, the cartridges were dried under vacuum for 30 min. After that, the analytes were eluted with 1.5 mL of methanol, 8 mL of methylene chloride-methanol (80:20, v/v) acidified with formic acid 50 mmol/L, and finally 3 mL of methylene chloride. The SPE extract was collected using Teflon vials (VWR International, Milan, Italy) and evaporated in a water bath at 30 °C under a gentle stream of nitrogen until a volume of approximately 5 mL was reached. The resulting solution was vortexed at room temperature for 2 min. The supernatant containing the humic material was discarded, and the remaining extract was dried under a gentle stream of nitrogen. The dried extract was reconstituted in 100 µL of water-THF (50:50, v/v) acidified with formic acid (1%, v/v). Liquid Chromatography. An Agilent 1100 series nano HPLC system was employed (Agilent Technologies Inc., Santa Clara, CA). Chromatographic separations were performed on a Agilent Zorbax SB-C18 column (150 mm × 75 µm i.d., 3.5 µm particle size). The injection volume was of 500 nL, and the flow rate was set at 300 nL/min. The mobile phase was composed of water (solvent A) and acetonitrile (solvent B), both acidified with formic acid (1%, v/v). The percentage of solvent B was changed from 0% to 100% in 30 min. Experiments for the evaluation of response linearity were carried out in flow injection analysis (FIA) mode at a flow rate of 300 nL/min, injecting a volume of 60 nL. The mobile phase was composed of solvent A and solvent B in equal proportions. Direct-EI-MS. The direct-EI interface was mounted on an Agilent 5975B inert MSD single quadrupole mass spectrometer (Agilent Technologies Inc., Santa Clara, CA). A detailed description of the characteristic and functioning of the direct-EI interface is reported elsewhere.17,20 Perfluorotributylamine was used as reference compound for mass spectrometer tuning and calibration.16 For the selected pesticides, the best temperature to obtain the highest signal and the best quality spectrum was 300 °C. Scan times were set in order to obtain a mean of 15 acquisition samples for each LC peak (1.5-1.8 scan/s). Data acquisition during the

Table 2. List of the Selected Pesticides with Their SIM Programs time (min)

compound (tR)

15-27 dicamba bromoxynil 2,4-D mecoprop MCPB 2,4-DB silvex 27-29 lindane 29-40 methoxychlor

dwell time (ms) 80

450 70

monitored ions 87-88-107-127-142-162164-171-173-196-198213-214-220-228-268-277

111-181-219 100-165-195-227-228-235237-241-246-248-249-263265-272-274-284-286-318-339

β-endosulfan R-endosulfan o,p′-DDD p,p′-DDD heptachlor o,p′-DDT p,p′-DDT HCB p,p′-DDE aldrin

chromatographic separations was carried out in selected ion monitoring (SIM) using different retention time windows (Table 2); dwell times were set between 0.72 and 0.77 cycles/s. RESULTS AND DISCUSSION In Table 1 a list of the target pesticides together with their main physicochemical properties is reported. All the selected pesticides are relevant for water quality due to either their persistent character, polarity, and/or continuous release into the water environment. Lindane, methoxychlor, MCPB, HCB, heptachlor, aldrin, R- and β-endosulfan, mecoprop, 2,4-D, 2,4-DB, DDT, and derivatives are indicated as chemical hazards in drinking water by the World Health Organization (WHO).26 HCB and R- and β-endosulfan are included on the EU list of 33 priority contaminants in the field of water policy.27 As reported in Table 1, phenoxy acids show a satisfactory water solubility and polarity (log P values between 2.14 and 3.68), thus making these compounds particularly suitable for LC/MS detection using ESI.5,7,13,22 On the other hand, the analytical protocols proposed for the determination of OCPs are traditionally based on GC/MS, due to their low polarity, high thermal stability, and high volatility. In one of our previous publications we demonstrated that OCPs can be detected by LC/ MS using nanoscale liquid chromatography coupled to a directEI interfacing system.16 This represented the first example of how non-API-amenable pesticides, such as OCPs, can be detected by LC/MS exploiting the following key features of direct-EI-MS:16-21 (1) any molecule present in the gas phase could be ionized, as long as it can withstand the typical experimental conditions of EI; (2) acquisition of high-quality, fully library searchable EI mass spectra of most sub-1-kDa molecules; (3) response is compounddependent, and it is never influenced by “matrix effects” induced by extraneous components in the sample or in the mobile phase. In this work we have improved our previous analytical approach addressing our investigation to the simultaneous detection of (26) World Health Organization. Guidelines for Drinking Water-Quality, 3rd ed.; WHO Press: Geneva, Switzerland, 2008. (27) Decision No. 2455/2001/EC of the European Parliament and of the Council.

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

7375

Figure 1. Extracted ion chromatograms of a river water extract. The river water sample was spiked and subjected to the SPE procedure following the conditions described in the Materials and Methods section. (A) 1, dicamba (m/z 173); 2, bromoxynil (m/z 277); 3, 2,4-D (m/z 162); 4, 2,4-DB; 5, silvex (m/z 196); 6, lindane (m/z 181); 7, methoxychlor (m/z 227); 8, HCB (m/z 284). (B) 1, mecoprop (m/z 142); 2, MCPB (m/z 142); 3, β-endosulfan (m/z 241); 4, R-endosulfan (m/z 241); 5, heptachlor (m/z 272); 6, o,p′-DDE. (C) 1, o,p′-DDD (m/z 235); 2, p,p′-DDD (m/z 235); 3, (1 o,p′-DDT (m/z 235); 4 p,p′-DDT (m/z 235); 5, aldrin (m/z 235).

OCPs and phenoxy acids. The addition of phenoxy acids herbicides did not require any instrument modification, and both classes could be detected in a single step of analysis. A series of preliminary experiments were performed to optimize the chromatographic conditions, testing different mobile 7376

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

phases consisting of acetonitrile and water with different additives, such as formic acid and trifluoroacetic acid at different concentrations. The direct-EI interface allows a complete freedom in the choice of LC solvents and modifiers, reducing difficulties in the achievement of the highest chromatographic performance.16-20

Table 3. Calibration Data Obtained for the Selected Pesticides and Recoveries

compound

LOD (pg)

LOQ (pg)

method LOD (µg/L)

dicamba bromoxynil 2,4-D mecoprop MCPB 2,4-DB silvex lindane methoxychlor β-endosulfan R-endosulfan o,p′-DDD p,p′-DDD heptachlor o,p′-DDT p,p′-DDT HCB p,p′-DDE aldrin

565 60 450 200 150 75 220 38 35 300 350 35 35 120 20 80 455 70 115

1885 200 1770 670 500 250 735 125 120 1000 1170 105 105 400 70 270 1520 235 370

0.047 0.006 0.052 0.018 0.013 0.006 0.022 0.006 0.003 0.026 0.039 0.003 0.003 0.021 0.002 0.008 0.047 0.007 0.019

2

R

0.9962 0.9999 0.9971 0.9902 0.9996 0.9993 0.9994 0.9996 1.0000 0.9999 0.9999 0.9986 0.9997 0.9986 0.9965 0.9957 0.9990 0.9953 0.9993

Formic acid at a concentration of 1% (v/v) was found to be the most effective additive in achieving the best compromise between peak shape and sensitivity (Figure 1). Spectrum acquisition was first carried out in scan mode to acquire top quality EI spectra of the investigated compounds. The quality of the recorded spectra was expressed as the measure of the degree of overlap with the reference spectra reported in the National Institute of Standards and Technology (NIST) electronic library. The experimental spectra were collected close to the detection limit in scan mode and thus with a high background noise. Probability factors higher than 80% were obtained for all the investigated pesticides, thus indicating an effective matching between the recorded spectra and the references. Validation data expressed in terms of limits of detection (LODs), limits of quantification (LOQs), linearity, and precision are briefly summarized in Table 3. The instrumental LOD was defined as the lowest analyte amount with a signal-to-noise ratio (S/N) of 3 and was evaluated by injecting on column diluted solutions of the standards. The analyte amount with a S/N of 10 was defined as the LOQ. Pesticides belonging to the class of phenoxy acids exhibit instrumental LODs ranging from 35 to 565 pg, whereas instrumental LODs for the OCPs were established between 20 and 455 pg (Table 3). These results indicate significant improvements of the instrumental sensitivity for OCPs respect to our previous publication, where LODs for the same compounds ranged from 120 to 850 pg.16 The increment of the instrumental sensitivity can be explained with a new setting of the EI source temperature, which was increased from 280 °C (value on previous publication) to 300 °C (actual value). This adjustment was crucial in optimizing the heat transfer and in making more efficient the droplet desolvation process into the EI source. As reported in the Materials and Methods section, linearity was assessed in FIA by analyzing standard solutions in quintuplicate at five different concentrations. Calibration curves were plotted using the peak area versus the compound concentration, and linearity was expressed as the squared correlation coefficient (R2). These experiments permitted us to assess the linearity of the system for more than 2 orders of magnitude starting

mean recovery ± RSD (%)

intraday precision RSD% (n ) 5)

interday precision RSD% (n ) 30)

Milli-Q water

river water

7 5 4 2 2 3 4 4 2 2 2 4 2 5 4 3 2 2 3

15 19 9 8 8 8 17 9 10 12 10 12 7 8 8 8 4 8 2

119 ± 11 105 ± 12 87 ± 17 110 ± 25 115 ± 12 114 ± 12 98 ± 12 67 ± 9 116 ± 17 116 ± 15 90 ± 21 105 ± 18 98 ± 16 57 ± 11 104 ± 16 103 ± 21 97 ± 13 96 ± 15 60 ± 11

100 ± 8 96 ± 16 113 ± 14 105 ± 24 105 ± 17 108 ± 19 107 ± 21 79 ± 15 102 ± 16 90 ± 11 86 ± 12 96 ± 10 95 ± 9 60 ± 10 98 ± 8 100 ± 13 90 ± 4 96 ± 15 56 ± 16

from the instrumental LOD of each pesticide (Table 3). Precision was calculated in terms of intraday and interday repeatability as relative standard deviation (RSD%) and was evaluated by analyzing a standard solution containing the selected pesticides at the same concentration of the recovery experiments (see the Materials and Methods section). For the determination of the intraday repeatability, six replicates were analyzed in one day. The interday repeatability was evaluated repeating the analyses over five different days (n ) 30). Both intraday and interday repeatability provided RSD% values lower than 20%, showing an excellent precision (Table 3). Method accuracy was estimated by means of SPE recovery experiments. The proposed SPE procedure was developed with the aim of reaching good recoveries for both group of pesticides in a single extraction step. To this end, Carbograph 4, a graphitized carbon black (GCB) with a surface area of 210 m2/g, was the selected sorbent, because of its proved versatility and efficiency in the extraction of analytes of a wide range of polarities, such as phenoxy acids, substituted ureas, organophosphates, carbamates, triazines, chloroacetanilides, azoles, and amides.22,25 By taking advantage of the GCB peculiarity of adsorbing specifically anionic compounds by electrostatic interactions, the pesticides re-extraction was performed using two different elution steps: one for phenoxy acids using an acidified methylene chloride solution, one for the OCPs using methylene chloride. As opposed to other published protocols, the procedure did not require any back-flushing step of the cartridge. The efficiency and robustness of the SPE preconcentration process was tested in terms of recovery on Milli-Q and river water samples. In Figure 1 the extracted ion chromatograms (EICs) of a river water extract are reported. As can be seen from these data, the direct-EI capability of operating in SIM allowed a precise identification of all the target compounds with satisfactory peak shape and S/N ratio. The recoveries were calculated by the analysis of five Milli-Q water and five river water samples, fortified with the selected pesticides at the concentrations reported in the Materials and Methods section. Each fortified water sample was processed by SPE, and the final extract was analyzed by LC/direct-EI-MS. Blank samples, Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

7377

without spiking, were also processed to subtract the levels of possible contaminants. The percentages of recovery were calculated by averaging five replicates. As can be seen from the data in Table 3, satisfactory recoveries with low relative standard deviations were obtained for almost all compounds. Significant losses were observed only for lindane, heptachlor, and aldrin. Nevertheless, due to the good method sensitivity and low standard deviation, this was not considered as a drawback for a reliable determination of these pesticides. In fact, simultaneous analysis of compounds with quite different physical-chemical properties often requires a compromise in the selection of experimental conditions, which in some cases cannot excel for all the analytes studied.28 The experimental data obtained for the estimation of the instrumental LODs and recoveries were used for the evaluation of the method sensitivity. The LOD of the method was expressed as the lowest analyte concentration that can be detected in the actual sample. It was calculated considering the initial volume of the sample (2 L), the final volume of the SPE extract (100 µL), the percentage of recovery of each investigated pesticide, the volume that is introduced into the LC/MS system (500 nL), and the instrumental LOD. As reported in Table 3, the method allows the determination of all the selected pesticides at the limits established by the EU legislation for drinking water (0.1 µg/L).1 (28) Gros, M.; Petrovic´, M.; Barcelo´, D. Anal. Bioanal. Chem. 2006, 386, 941– 952.

7378

Analytical Chemistry, Vol. 81, No. 17, September 1, 2009

CONCLUSIONS The analytical method described above provides an innovative and rapid approach for the simultaneous detection of pesticides belonging to the class of phenoxy acids and OCPs. Due to their antagonist physical-chemical properties, these two group of pollutants cannot be analyzed simultaneously using the current multiresidue methods based on GC/MS and LC/MS. With the use of an off-line SPE based on GCB followed by LC/direct-EIMS, this method gives the sensitivity and selectivity necessary for the detection of these pesticides at environmentally relevant concentrations in the low nanogram per liter range. This result demonstrates the LC/direct-EI-MS capability of identifying pesticides with quite different polarities and thermostabilities in a single step of analysis. This represents a significative advance in developing universal multiresidue screening programs, aimed at the rapid and sensitive identification of the main classes of pesticides in the different water compartments. ACKNOWLEDGMENT The authors thank Agilent Technologies Inc. (Santa Clara, CA and Waldbronn, Germany) for supplying the instrumentation. Received for review April 27, 2009. Accepted July 13, 2009. AC9008995