Simultaneous Determination of Halogenated Derivatives of

Anal. Chem. 2001, 73, 5886-5895

Simultaneous Determination of Halogenated Derivatives of Alkylphenol Ethoxylates and Their Metabolites in Sludges, River Sediments, and Surface, Drinking, and Wastewaters by Liquid Chromatography-Mass Spectrometry Mira Petrovic,† Alfredo Diaz,‡ Francesc Ventura,‡ and Damia` Barcelo´*,†

Department of Environmental Chemistry, IIQAB-CSIC, c/Jordi Girona 18-26, 08034 Barcelona, Spain, and AGBAR, Aigu¨es de Barcelona, P. Sant Joan 39, 08009 Barcelona, Spain

A quantitative solid-phase extraction-liquid chromatography/mass spectrometry (SPE-LC/MS) method is described for the simultaneous analysis of halogenated byproducts of alkylphenolic compounds and their degradation products formed during chlorine disinfection in the presence of bromide ions. Compounds analyzed include brominated and chlorinated nonylphenol ethoxylates (XNPEOs); octylphenol ethoxylates (XOPEOs); nonylphenols (XNP); nonylphenoxycarboxylates (XNPECs) and their precursors nonionic surfactants, alkylphenol ethoxylates (APEOs); and their metabolites formed during sewage treatment, alkylphenoxycarboxylates (APECs) and alkylphenols (APs). Target compounds were concentrated from water samples using a C18 SPE procedure. Extracts were analyzed using reversed phase LC/MS. The performances of both atmospheric pressure chemical ionization (APCI) and electrospray (ESI) interfaces were compared. ESI offered better sensitivity and specificity for a higher range of oligomers. Detection limits (LODs) for water samples were from 20 to 100 ng/L; and for sediment samples, from 2 to 10 µg/kg. Slightly higher LODs were obtained for sludge samples (5-25 µg/kg). Halogenated byproducts were found in sludge from Barcelona drinking water treatment plant in concentrations of 220 µg/kg for BrNP, 430 µg/kg for BrNPEOs (nEO ) 1 - 2), and 1600 µg/kg for BrNPEOs (nEO ) 3 - 15). The concentration of ClNPEOs was estimated to be in the order of 660 µg/kg (assuming the same response as BrNPEOs). Halogenated OPEOs were also identified, and their concentration was ∼50 times lower than the concentration of NPEOs analogues. To our knowledge, this is the first method described that allows simultaneous determination of alkyphenol ethoxylates and halogenated derivatives, including degradation products. Nonionic surfactants of alkylphenol ethoxylate type (APEOs) are widely used as industrial cleaning agents and wherever their * Corresponding author. Phone: +34 93 400 6118. Fax: +34 93 204 59 04. E-mail: [email protected]. † IIQAB-CSIC. ‡ AGBAR.

5886 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

interfacial effects of detergency, defoaming, de-emulsification, dispersion, or solubilization can enhance products or process performance. Although parent APEOs are not classified as highly toxic substances, they represent a class of compounds of high environmental concern because of persistent metabolic products (alkylphenols; APs; and carboxylic derivatives, APECs) generated during wastewater treatment. The numerous studies confirmed that alkylphenolic compounds can mimic endogenous hormones. In vitro and in vivo experiments have demonstrated that the estrogenic effect of alkylphenols is 4-6 orders of magnitude lower than that of the endogenous 17β-estradiol;1 however, because of their ubiquity in the environment, persistent nature, and lipophilicity, they deserve particular attention. APEOs and their biodegradation products were found to be transformed into halogenated byproducts during chlorine disinfection in the presence of bromide ion (Figure 1); however, little is known about the environmental significance and toxicology of brominated and chlorinated alkylphenolic compounds. Reinhard and co-workers2 suspected that occurrence of mutagenicity in wastewater is correlated with the formation of brominated alkylphenolic byproducts; however, their preliminary experiments conducted with BrAPECs failed to confirm this hypothesis. Maki et al.3 determined that both BrNPEOs and BrNPECs show higher acute toxicity to Daphnia magna than their nonbrominated precursors NPEOs and NPECs. The identification and quantification of halogenated derivatives and their precursors APEOs, APs, and APECs in the environmental samples is complicated and cumbersome. Although numerous publications report environmental data for APEOs, APs, and APECs, only few of them includes halogenated metabolites. Using GC/MS and GC/high-resolution MS (HRMS), Reinhard et al.2 identified brominated APECs in wastewater after treatment with chlorine. Using GC/MS and fast atom bombardment MS (FABMS), Ventura et al.4-6 reported the formation of halogenated (1) Jobling, S.; Sheahan, D.; Osborne, J. A.; Matthiessen, P.; Sumpter, J. P. Environ. Toxicol. Chem. 1996, 15, 194-202. (2) Reinhard, M.; Goodman, N.; Mortelmans, K. E. Environ. Sci. Technol. 1982, 16, 351-362. (3) Maki, H.; Okamura, H.; Aoyama, I.; Fujita, M. Environ. Toxicol. Chem. 1998, 17, 650-654. 10.1021/ac010677k CCC: $20.00

© 2001 American Chemical Society Published on Web 11/10/2001

Figure 1. Biodegradation pathway and byproducts formation of alkylphenol ethoxylates.

derivatives of alkylphenols and acidic alkylphenols in the chlorination process and identified monobrominated nonylphenol (BrNP) in the tap water of Barcelona. Fujita et al.7 used ion-pair reversedphase HPLC with UV detection for the determination of BrOP1EC. The method was applied for the quantification of brominated OPEC in aerobic biological transformation experiments, but it is not applicable to real-world samples because of its low sensitivity and selectivity. The only quantitative method applicable to halogenated alkylphenols in environmental samples was recently reported by Ferguson and co-workers.8 They developed a very sensitive quantitative LC/ESI-MS method for the analysis of shortchain OPEOs, NPEOs and their acidic and fully de-ethoxylated metabolites, including monobrominated and monochlorinated nonylphenols. However, the protocol used does not permit the simultaneous quantification of APs and APECs because of coelution problems, and neither halogenated APEOs nor halogenated APECs were included in the protocol. In previous works from our group9-11 we developed an integrated SPE-LC/MS method using both ESI and APCI interfaces for the simultaneous identification and quantification of (4) Ventura, F.; Figueras, A.; Caixach, J.; Espadaler, I.; Romero, J.; Guardiola, J.; Rivera, J. Water Res. 1988, 22, 1211-1217. (5) Ventura, F.; Caixach, J.; Figueras, A.; Espadaler, I.; Fraisse, D.; Rivera, J. Water Res. 1989, 23, 1191-1203. (6) Ventura, F.; Caixach, J.; Romero, J.; Espadaler, I.; Rivera, J. Water Sci. Tech. 1992, 25, 257-264. (7) Fujita, J.; Reinhard, M. Environ. Sci. Technol. 1997, 31, 1518-1524. (8) Ferguson, P. L.; Iden, C. R.; Brownawell, B. J. Anal. Chem. 2000, 72, 43224330. (9) Petrovic, M.; Barcelo´, D. J. AOAC 2001, 89, 1079-1085. (10) Petrovic, M.; Barcelo´, D. Anal. Chem. 2000, 72, 4560-4567. (11) Petrovic, M.; Barcelo´, D. Fresenius’ J. Anal. Chem. 2000, 368, 676-683.

nonionic surfactants (NPEOs, alcohol ethoxylates, and coconut fatty acid diethanol amides), their main degradation products (poly(ethylene glycol)s, nonylphenol, octylphenol and nonylphenoxycarboxylates), and ionic surfactants (linear alkylbenzene sulfonates) in sewage sludge, wastewater, river and seawater, and sediments. In this work, a comprehensive analytical method for the simultaneous determination of halogenated derivatives of APEOs and their fully de-ethoxylated (APs) and carboxylated (APECs) metabolites has been developed. To our knowledge, the method is the first protocol that is applicable to a full range of halogenated alkylphenolic compounds: halogenated alkylphenol ethoxylates (XAPEOs), halogenated alkylphenols (XAPs), and halogenated alkylphenoxycarboxylates (XAPECs) in solid (sludge and sediment) and aqueous samples. The specific objective of this work was to compare and optimize APCI and ESI interfaces in negative ion (NI) and positive ion (PI) mode for the MS characterization of target compounds and to apply the analytical methodology for the quantification of target compounds in environmental and wastewater samples. EXPERIMENTAL SECTION Materials and Standards. All solvents (water, acetonitrile, methano,l and dichloromethane) were HPLC grade and were purchased from Merck (Darmstadt, Germany). Analytical-grade acetic acid and sodium acetate were from Panreac (Barcelona, Spain). BrNP was synthesized using elemental bromine according to the method described by Reinhard et al.2 ClNP was prepared by chlorination of nonylphenol using sulfuryl chloride according to the method of Stokker et al.12 BrNP1EC and ClNP1EC were Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

5887

synthesized by reacting brominated and chlorinated nonylphenol, respectively, with chloroacetic acid in the presence of sodium hydride and dimethyl formamide as a solvent. These two synthesized compounds rendered BrNP1EO and ClNP1EO by reduction with lithium aluminum hydride in ether solution. BrNP2EO and ClNP2EO were synthesized by reacting BrNP and ClNP, respectively, with 2-(2-chloroethoxy)ethanol in the presence of NaOH in water. Finally, BrNP2EC and ClNP2EC were obtained from BrNP2EO and ClNP2EO, respectively, by oxidation with Jones reagent.2 The individual polyethoxylated surfactants nonylphenol polyethoxylates (NPnEO) and octylphenol polyethoxylates (OPnEO), corresponding to a mixture with an average number of ethoxy groups (n) were from Kao Corporation (Barcelona, Spain). The standard of octylphenol ethoxylate (OPEO) contained oligomers with an average of nine ethoxy units, and the standard of NPEOs contained chain isomers and oligomers with an average of 10 ethoxy units. Additionally, in the case of partially degraded samples (e.g., STP effluents), a mixture of monoethoxy and diethoxy nonylphenols (49:51, w/w) and a mixture of NPEOs with an average of four ethoxy units were used for calibration. High purity (98%) 4-tert-octylphenol (OP) and technical grade 4-nonylphenol (NP), were obtained from Aldrich (Milwaukee, USA). Alkylphenoxycarboxylates (NP1EC and OP1EC) and octylphenoxyethoxycarboxylates (NP2EC and OP2EC) were synthesized according to the method described by Marcomini et al.13 Stock solutions (1 mg/mL) of individual standards and standard mixtures were prepared by dissolving accurate amounts of pure standards in methanol. Working standard solutions were obtained by further dilution of stock solutions with methanol. Sample Collection and Preparation. Water Samples (River Water, Influent and Effluent of DWTP and STP). Influent and effluent water of a sewage treatment plant (STP), Igualada (Catalonia, NE Spain), was collected in glass bottles as 24-h composite samples. Raw water from the Llobregat River (ca. 30 km upstream of DWTP) and entering the DWTP, prechlorinated water and final effluent (treated drinking water) of the Barcelona drinking water treatment plant (DWTP) were collected as grab samples in Pyrex borosilicate amber glass containers, previously rinsed with highpurity water. All aqueous samples were stored at 4 °C immediately after sampling, filtered through a 0.45-µm membrane filter and preconcentrated on LiChrolut C18 (Merck, Darmstadt, Germany) SPE cartridges within 24 h in order to avoid any degradation of target compounds. All SPE experiments were performed using an automated sample preparation with extraction columns system (ASPEC XL) fitted with an external 306 LC pump for the dispensing of samples through the SPE cartridges and with an 817 switching valve for the selection of samples, all from Gilson (Villiers-le-Bel, France). Disposable 3-mL cartridge columns packed with 500 mg of LiChrolut C18 sorbent from Merck (Darmstat, Germany) were activated and conditioned first with 7 mL of methanol and then with 3 mL of HPLC water at a flow rate (12) Marcomini, A.; Di Corcia, A.; Samperi, R.; Capri, S. J. Chromatogr. A 1993, 644, 59-71. (13) Stokker, G. E.; Daena, A. A.; Desolms, S. J.; Schultz, E. M.; Smith, R. L.; Cragoe, E. J.; Baer, J. E.; Ludden, C. T.; Russo, H. F.; Scriabine, A.; Sweet, C. S. Watson, L. S. J. Med. Chem. 1980, 23, 1414-1427.

5888

Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

of 1 mL/min. Samples were loaded at a flow rate of 5 mL/min. Different volumes of samples were loaded, depending on the type of water (STP influent, 50 mL; STP effluent, 200 mL; river water and water from DWTP, 500 mL). After preconcentration, the sorbents were completely dried (30 min) to avoid hydrolysis using a Baker SPE 12g apparatus (J. T. Baker, Deventer, Netherlands) connected to a vacuum system set at -15 psi. After drying, the SPE cartridges were wrapped in aluminum foil and kept at -20 °C until analysis (max., 1 month). Cartridges were eluted with 2 × 4 mL of methanol. The eluates were evaporated to dryness with a gentle stream of nitrogen and reconstituted with methanol to a final volume of 1 mL. Solid Samples (Sludge and Sediment). Sludge from DWTP of Barcelona was obtained mainly from prechlorinated raw water (the Llobregat River) after flocculation with aluminum sulfate mixed in a minor proportion with sludge coming from the washing of sand filters. Sludge was collected in precleaned amber glass bottles. The suspension (concentration of dry matter, 3.5-5 g/L) was centrifuged at 4500 rpm, and the solid matter was separated and frozen at -20 °C before being freeze-dried. A sample of dehydrated stabilized sludge from STPs Igualada, was collected in glass amber bottles with an aluminum-foil-lined cap and frozen at -20 °C before being freeze-dried. The lyophilized DWTP and STP sludge samples were wrapped in aluminum foil and stored at -20 °C until extraction. A grab sample of surface river sediment from the Llobregat River was collected in glass bottles and transferred to the laboratory, stored at 4 °C, then frozen at -20 °C, and finally lyophilized. The lyophilized samples were ground and homogenized using a mortar and pestle and then sieved through a 125µm sieve. The freeze-dried sample was stored in precleaned glass bottles at -20 °C until extraction. A freeze-dried 2-g (sludge) or 5-g (river sediment) sub-samples was sonicated 20 min using 20 mL of dichloromethane/methanol (7/3, v/v). The extract was separated by centrifugation at 4000 rpm for 5 min. The extraction was repeated two times with a fresh solvent mixture. The extracts were pooled and concentrated to an approximate volume of 1 mL using a rotary vacuum evaporator at 30 °C and redissolved in 100 mL of HPLC water. Subsequent cleanup of extracts was performed by SPE as described above. Liquid Chromatography/Mass Spectrometry. The HPLC system consisted of an HP 1100 autosampler having a 100-µL loop and an HP 1090A LC binary pump, both from Hewlett-Packard (Palo Alto, CA). The HPLC separation was achieved on a 5-µm, 250 × 4 mm i.d. C18 reversed-phase column (LiChrospher 100 RP-18) preceded by a guard column (4 × 4 mm, 5-µm) of the same packing material from Merck (Darmstadt, Germany). The injection volume was set at 25 µL, and the flow rate was 1 mL/ min. Detection was carried out using an HP 1040 M diode array UV-vis detector coupled in series with an LC/MSD HP 1100 mass-selective detector equipped with an atmospheric-pressure ionization source that can use either an atmospheric-pressure chemical ionization (APCI) or electrospray (ESI) interface. To obtain maximum sensitivity for each target compound, the detection parameters were optimized in series of flow injection sequences by direct injection of each target compound at a concentration of 5 mg/L into the flow of the carrier solvent (80%

Table 1. LC and ESI-MS Conditions parameter solvent A solvent B gradient elution

drying gas flow (L/min) drying gas temp (°C) nebulizer pressure (psi) capillary voltage (V) fragmentation voltage (V)

pozitive ionization (PI)

negative ionization (NI)

LC Separation acetonitrile water 0 min, 30% A 3 min, 30% A 10 min, 80% A 15 min, 90% A 30 min, 90% A

methanol water 0 min, 40% A 15 min, 80% A 30 min, 90% A 35 min, 100% A 45 min, 100% A

MS Detection 12 375 55 4500 60

11 325 50 3500 100

A). The detection parameter values were optimized evaluating the sensitivity, signal-to-noise ratio, and fragmentation of each analyte in the scan mode (m/z scan values, 100-1000 for PI and 100500 for NI). The following operating parameters were optimized: drying gas flow and temperature, nebulizer pressure, capillary voltage, and fragmentor voltage. Because of the higher sensitivity and selectivity, an ESI interface was chosen for the quantitative determination of target compounds in wastewater and environmental samples. The chromatographic conditions and mobile phases used were the same as optimal conditions of the MS detector that are given in Table 1. Diagnostic ions used for the analysis of APEOs and XAPEOs in PI mode were those corresponding to [M + Na]+. Prior to analysis, the extracts were fortified with 25 µM sodium acetate (5 µL of 5 mM aqueous solution) to avoid a possible reduction in APEO ionization due to insufficient metal ion availability. APs, AP1ECs, and XNPs were detected under NI conditions as [M - H]-, and for XNP1ECs and XNP2ECs, the base ions (at the fragmentor voltage of 100 V) corresponded to [M CH2COOH]- and [M - CH2CH2OCH2COOH]-, respectively. Quantitation. Quantitative analysis was performed in a selected ion monitoring (SIM) mode using external calibration. Initially, a series of injections of target compounds in the concentration range from 10 ng/mL to 25 µg/mL was used to determine the linear concentration range. Calibration curves were generated using linear regression analysis and over the established concentration range (0.05-5 µg/mL) gave good fits (r2 > 0.990). Five-point calibration was performed daily, and the possible fluctuation in the signal intensity was checked by injecting a standard solution at two concentration levels after each 6-8 injections. The confirmation of compound identity in wastewater and environmental samples was performed in a full-scan mode by matching the retention time and the mass spectrum of an authentic standard. Method Validation. The recoveries and overall method reproducibility were determined for each type of matrix (water, sludge, and sediment) from triplicate analysis of spiked samples. Freeze-dried river sediment and DWTP sludge were spiked with 100 and 500 µg/kg, respectively, of the composite standard solution of halogenated alkylphenolic compounds and their precursors and analyzed by applying the method described above, together with a blank sample (no spiked sample). River water was

spiked with the standard mixture of analytes to a final concentration of 5.0 µg/L. The detection limit (LOD) of the combined SPE-LC/ESI-MS procedure, achieved by the preconcentration of 500 mL of river water, 5 g of river sediment, and 2 g of DWTP sludge, were calculated as the minimum amount of a compound present in a sample that produces a signal-to-noise ratio of 3, on the basis of an injection of a 25-µL aliquot of the final 1 mL of extract of each matrix tested. All results for solid samples (sludge and sediment) are expressed on dry-weight basis. RESULTS AND DISCUSSION Liquid Chromatography/Mass Spectrometry. Table 2 lists the main ions and their relative abundance obtained using APCI and ESI-MS. To quantify all target compounds, it was necessary to perform two analyses for each sample. APs, APECs, XAPECs, and XAPs were detected in NI mode, and APEOs and their halogenated analogues were detected in PI mode. Instrumental detection limits (IDL), expressed as the injected amount of each compound at S/N ) 3, determined in full scan mode, for XNPs using ESI were approximately 3 times lower than IDLs obtained with an APCI interface, but BrNPECs and ClNPECs were detected using both interfaces with approximately equal sensitivity. The use of ESI, however, provided significantly higher level of sensitivity for nonhalogenated analogues. IDLs obtained with an ESI for APs were ∼30-50 times lower than corresponding ones obtained using an APCI interface. The sensitivity of SIM detection was enhanced ∼10 times when compared against the corresponding extracted ion chromatograms obtained under fullscan mode. ESI mass spectra of available standards are shown in Figure 2. XNPs gave the characteristic isotope doublet signal of the [M - H]- ions (m/z 297/299 for BrNP and m/z 253/255 for ClNP). XNPECs gave two signals, one corresponding to quasi-molecular ion and another to [M - CH2COOH]- in the case of XNP1ECs or [M - CH2CH2OCH2COOH]- for XNP2ECs. With gradient elution using methanol/water on a C18 reversedphase column, APECs, APs, XAPs, and XAPECs can be easily separated. However, when analyzing real samples containing linear alkylbenzene sulfonates (LAS); ClNP1EC coelutes with C11LAS, a compound having the same molecular weight (MW ) 312) and base ion m/z 311. LAS are the major surfactant class used in detergents throughout the world, and high concentrations (up to Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

5889

Table 2. Comparison of APCI and ESI Interfacesa

compd

APCI main ions (relative abundance)

ionization mode

H]-

IDLb ng

ESI main ions (relative abundance) H]-

IDL ng

BrNP ClNP BrNP1EC

NI NI NI

297/299 (100/98) [M 253/255 (100/36) [M - H]355/357 (100/98) [M - H]297/299 (20/18) [M - CH2COOH]-

0.3 0.3 0.2

297/299 (100/98) [M 253/255 (100/36) [M - H]355/357 (100/98) [M - H]297/299 (40/38) [M - CH2COOH]-

0.1 0.1 0.2

BrNP2EC

NI

399/401 (95/92) [M - H]297/299 (100/98) [M - CH2CH2OCH2COOH]-

0.2

399/401 (35/33) [M - H]297/299 (100/98) [M - CH2CH2OCH2COOH]-

0.1

ClNP1EC

NI

311/313 (100/36) [M - H]253/255 (20/7) [M - CH2COOH]-

0.2

311/313 (100/36) [M - H]253/255 (40/15) [M - CH2COOH]-

0.2

ClNP2EC

NI

355/357 (70/25) [M - H]253/255 (100/36) [M - CH2CH2OCH2COOH]-

0.2

355/357 (20/8) [M - H]253/255 (100/36) [M - CH2CH2OCH2COOH]-

0.1

BrNP1EO BrNP2EO ClNP1EO ClNP2EO

PI PI PI PI

ndc nd nd nd

nd nd nd nd

365/367 (100/98) [M + Na]+ 409/411 (100/98) [M + Na]+ 321/323 (100/36) [M + Na]+ 365/367 (100(36) [M + Na]+

0.5 0.1 0.5 0.1

a Fragmentation voltage, 80 V b Instrumental detection limits (ng injected); amount giving a peak with S/N ) 3; determined in full-scan mode by flow injection analysis. c nd, not determined because of the low sensitivity.

several mg/L) are often found in environmental and wastewater samples.14 All attempts to separate ClNP1EC and C11LAS using gradient elution with standard mobile phases for reversed-phase separation (methanol/water or acetonitrile/water) failed. Besides the molecular ion with m/z 311/313, ClNP1EC can be detected by fragment ion [M - CH2COOH]- with m/z 253/255. The relative abundance and absolute intensity of these two ions highly depends on the fragmentor voltage. For this reason, an accurate optimization of the fragmentor voltage was performed for each compound detected under NI mode (Table 3). It was found that at higher voltages (g100 V), the base peak with high absolute intensity for ClNP1EC and ClNP2EC is m/z 253/255; and for BrNP1EC and BrNP2EC, m/z 297/299. Therefore, these compounds can be monitored using the same m/z channels as XNPs, increasing the relative instrument dwell time and enhancing sensitivity. Reconstructed chromatograms obtained by combining the SIM channels for AP1ECs, APs, XNPECs, and XNPs, applying the fragmentor voltage of 100 V, are shown in Figure 3. Brominated and chlorinated nonylphenol ethoxylates (BrNPEOs and ClNPEOs) were detected only by ESI, yielding doublet signals characteristic for bromine and chlorine isotopes, respectively. Using an ESI interface, XNPEOs, the same as nonhalogenated analogues, show a great affinity for alkali metal ions, and they gave exclusively evenly spaced sodium adduct peaks [M + Na]+, even in the absence of added sodium (Figure 2G,H). However, to avoid a possible reduction in ionization due to insufficient metal ion available in solution, it was necessary to fortify the sample extracts with sodium ions prior to injection. The comparison of IDLs obtained for available standards of XNPEOs, NPEOs, and OPEOs clearly indicated that an ESI interface offered better sensitivity and specificity for a higher range of oligomers. It was found that mono- and diethoxylates can be detected using an APCI interface only at very high concentrations, far higher than levels found in environmental and wastewater samples; however, ESI permits the detection of these oligomers. The IDLs (full scan mode) for NP1EO and NP2EO obtained by careful (14) Eichhorn, P.; Petrovic, M.; Barcelo´, D.; Knepper, Th. P. Vom Wasser 2000, 95, 245-268.

5890 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

optimization of operating parameters and applying low fragmentor voltage (60 V) were 1 and 0.2 ng, respectively. As previously reported by several authors,8,9,15,16 the response of ESI-MS detector rapidly decreases as the number of ethoxy units decreases from 4 to 1. For halogenated NPEOs, it was found that the ratio of the response factors of XNP2EO and XNP1EO is 3.5. No standards of XNPEOs with nEO > 2 were available at the time of analysis, but an increase in absolute response in the ESI signal with increasing degree of ethoxylation, similar to those observed for nonhalogenated analogues, can be expected. Because the chlorinated analogues (ClNPEOs) gave the same ions as brominated ones with one less ethoxy group, the good chromatographic separation of these two groups of compounds is the prerequisite of their quantitative determination with LC/ MS (Figure 4). In the DWTP sludge, 44-Da (mass of one ethoxy unit) spaced ion series (m/z from 365 to 981) revealed the presence of BrNPEOs with nEO ) 1-15. The careful examination of the isotopic profile in the mass spectra confirmed the identity of ClNPEOs and BrNPEOs, respectively (Figure 5A,B). The doublet signal in the mass spectrum of BrNPEOs shows the contribution of bromine isotopes of 79Br:81Br ) 100:98, and the contribution of chlorine isotopes is 35Cl:37Cl ) 100:33. Halogenated derivatives of OPEOs were also identified in the DWTP sludge (Figures4 and 5). The identities of halogenated derivatives of OPEOs were confirmed by their MS spectra, and the concentrations were estimated assuming that they gave the same response as XNPEOs. Similarly to halogenated NPEOs, OPEOs derivatives gave regularly spaced signals corresponding to [M + Na]+ ions with m/z 395/397-571/573 corresponding to ClOPEOs, nEO ) 3-6 (assigned as 9 in Figure 5C) and m/z 395/397-615/617 corresponding to BrOPEOs, nEO ) 2-7 (assigned as 9 in Figure 5D). One of the main problems to be solved when analyzing polyethoxylated compounds by ESI-MS and previously discussed (15) Shang, D. Z.; Ikonomou, M. G.; Macdonald, R. W. J. Chromatogr. A 1999, 849, 467-482. (16) Crescenzi, C.; Di Corcia, A.; Samperi, R. Anal. Chem. 1995, 67, 17971804.

Figure 2. ESI mass spectra of halogenated alkylphenolic compounds: A-E, compounds detected under NI conditions; F and G, detected under PI conditions; A, ClNP; B, BrNP; C, ClNP1EC; D, BrNP1EC; E, ClNP2EC; F, BrNP2EC; G, mixture ClNP1EO and ClNP2EO (1:1, w/w); and H, mixture BrNP1EO and BrNP2EO (1:1, w/w).

by Ferguson,8 Shang,15 and Crescenzi16 is the suppression of the analyte signal in environmental samples. Such samples usually contain high concentrations of co-occurring compounds, resulting in a decrease of the relative contribution of the analyte to the total ion current. The matrix effect can be reduced, if not completely removed, by careful sample cleanup and LC separation. To determine the extent of ion suppression on the determination of XNPEOs, three series of spiked extracts (one in methanol, one in STP influent extract, and one in STP sludge extract) were analyzed by flow injection analysis (FIA/ESI-MS) and HPLC/ESIMS under positive ion mode, together with blank samples (no spiked samples). The signal intensity of BrNP2EO in STP sludge

extract after reversed-phase LC separation was 18% lower than in clean solvent, but in the STP influent, the reduction was limited to 10-12%. When the sample was injected directly to ESI-MS the signal was reduced more than 90% in both extracts. It should be taken into account that STP influents and sludge samples are probably the most loaded samples, and the suppression in the case of less complicated matrixes (surface water, sediment, and STP effluent) is expected to be negligible. These results confirm the necessity of good chromatographic separation, even when it means a longer analysis time (in our case, each run takes 40 min). Method Validation. The established protocol, ultrasonic solvent extraction (for solid samples) SPE-reversed-phase LC/ Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

5891

Table 3. Effect of Fragmentor Voltage on Absolute Intensity and Relative Abundancea of Ions Measured under NI Conditionsb fragmentor voltage (V) compd BrNP ClNP BrNP1EC BrNP2EC ClNP1EC ClNP2EC NP1EC NP2EC OP1EC OP2EC NP OP a

ion

60

80

100

120

297/299 253/255 355/357 297/299 399/401 297/299 311/313 253/255 355/357 253/255 277 219 321 219 263 205 307 205 219 205

6.1 × (100) 8.1 × 106 (100) 4.1 × 106 (100)

6.5 × (100) 1.1 × 107 (100) 4.3 × 106 (100) 1.2 × 106 (40) 1.0 × 106 (35) 9.8 × 106 (100) 5.2 × 106 (100) 1.4 × 106 (40) 1.1 × 106 (20) 1.1 × 107 (100) 1.6 × 107 (100) 4.2 × 106 (25) 3.8 × 106 (50) 6.9 × 106 (100) 1.4 × 107 (100) 1.9 × 106 (15) 4.9 × 106 (60) 7.1 × 106 (100) 1.0 × 107 (100) 1.3 × 107 (100)

6.2 × (100) 1.0 × 107 (100) 4.0 × 106 (85) 3.9 × 106 (100) 4.0 × 105 (8) 9.9 × 106 (100) 2.8 × 106 (65) 5.9 × 106 (100) 8.0 × 105 (8) 1.2 × 107 (100) 1.6 × 107 (100) 4.6 × 106 (40) 1.6 × 106 (20) 9.2 × 106 (100) 1.6 × 107 (100) 4.8 × 106 (38) 2.4 × 106 (25) 1.0 × 107 (100) 1.0 × 107 (100) 1.2 × 107 (100)

6.1 × 106 (100) 9.8 × 106 (100) 9.5 × 105 (25) 4.2 × 106 (100)

106

4.8 × 106 (100) 4.3 × 106 (90) 5.8 × 106 (100) 4.5 × 106 (95) 4.9 × 106 (100) 1.4 × 107 (100) 7.4 × 105 (5) 8.9 × 106 (100) 8.2 × 105 (10) 1.0 × 107 (100) 1.0 × 106 (5) 1.0 × 107 (100) 1.1 × 106 (13) 9.8 × 106 (100) 1.1 × 107 (100)

106

106

9.0 × 106 (100) 1.2 × 106 (10) 6.1 × 106 (100) 1.3 × 107 (100) 6.8 × 106 (40) 1.5 × 107 (100) 1.2 × 107 (100) 7.9 × 106 (80) 1.1 × 107 (100) 1.5 × 107 (100) 9.2 × 107 (100) 9.3 × 107 (100)

In parentheses. b For other conditions see Experimental Section.

Figure 3. Reconstructed ion chromatograms obtained by ESI-MS in NI mode for a standard mixture of AP1ECs, XNP1ECs, XNP2EC, XNPs, APs, and LAS.

(ESI)-MS was validated by the analysis of spiked water, sludge, and sediment samples. Limits of detection (LODs), calculated as S/N ) 3 for quantification in SIM mode, are given in Table 4 for each matrix analyzed. LODs for water samples (based on a concentration factor of 500) were below the parts-per-billion range (20-100 ng/L) and 5892 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

in sediment samples (concentration factor ) 5), from 2 to 10 µg/ kg. Slightly higher LODs were obtained for sludge samples (525 µg/kg), mainly as a result of the lower concentration factor used and high concentration of interfering substances generating higher level of background noise. The efficiency of SPE using LiChrolute C18 cartridges ranged from 72 to 98%. For solid samples, recoveries were 73-90% for sediment and 59-80% for sludge. Overall precision of the analysis was satisfactory. The relative standard deviation (RSD, n ) 3) for water analysis was below 7% and for sludge and sediment samples, less than 12%. Analysis of Environmental and Wastewater Samples. The developed SPE-LC/ESI-MS method was applied to the analysis of aqueous and solid samples. The following samples were analyzed: water and sludge samples from Barcelona DWTP, sediment and water of the Llobregat River that is treated in DWTP, and influent and effluent wastewaters of STP Igualada that discharges treated water to a tributary of the Llobregat River. The concentrations that were found are shown in Table 5. Halogenated byproducts were detected only in DWTP samples because of the prechlorination used in the treatment process and high concentration of bromide ions in raw water (1 mg/L) coming from salt mines in the upper course of the river. It was already reported that tap water of Barcelona contains a measurable concentration of brominated disinfection byproducts.17 Brominated alkylphenolic compounds were detected in prechlorinated water (4.0 µg/L of BrNPEOs, 0.42 µg/L of BrNPECs, and 0.21 µg/L of BrNP). Chlorinated analogues were not detected. After prechlorination, the treatment process used in the Barcelona DWTP includes the following steps: flocculation with Al2(SO4)3, rapid sand filtration, ozonation, GAC filtration, and final chlorination. (17) Cancho, B.; Ventura, F.; Galceran, M×bb., T. Bull. Environ. Contam. Toxicol. 1999, 63, 610-617.

Figure 4. Total ion LC/ESI-MS chromatogram (bottom trace) and reconstructed chromatograms of halogenated APEOs and APEOs, obtained in PI mode, found in sludge from Barcelona drinking water treatment plant.

Figure 5. ESI mass spectra of halogenated alkyl phenol ethoxylates detected in the DWTP sludge (chromatogram shown in Figure 4): A, ClNPEO; B, BrNPEO; C, ClOPEO (assigned as 9); and D, BrOPEO (assigned as 9).

Alkylphenolic compounds and halogenated derivatives, formed by prechlorination, were efficiently removed during water treatment and were not detected in the final effluent (treated drinking water); however, sludge formed by flocculation was found to accumulate halogenated compounds and their precursors. BrNPEOs (nEO ) 1-15) were detected at a concentration of 2030 µg/kg, and the

concentration of ClNPEOs was estimated to be in the order of 660 µg/kg (assuming the same response as BrNPEOs). The concentration of halogenated OPEOs was approximately 50 times lower than the concentration of NPEOs analogues. BrNP was found in a concentration of 220 µg/kg, but chlorinated derivatives were not detected. Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

5893

Table 4. Recoveries, RSDa and Limits of Detection for Each Sample Matrix water samplesb compd BrNPEO ClNPEO BrNP1EC ClNP1EC BrNP ClNP NPEO OPEO NP OP NP1EC OP1EC a

sludge samplesc

sediment samplesd

LOD (ng/L)

recovery (%)

LOD (µg/kg)

recovery (%)

LOD (µg/kg)

recovery (%)

50 (nEO ) 1) 20 (nEO ) 2) 50 (nEO ) 1) 20 (nEO ) 2) 25 25 50 50 100 (nEO ) 1) 40 (nEO ) 2) 25 (nEO ) 3-15) 100 (nEO ) 1) 40 (nEO ) ) 25 (nEO ) 3-15) 20 20 20 20

83.3 (5.6) 88.6 (3.6) 80.2 (4.5) 87.1 (5.1) 94.0 (4.0) 92.1 (5.6) 72.9 (3.1) 74.5 (5.6) 90.1 (7.0) 88.9 (6.2) 95.2 (2.5) 90.5 (4.1) 93.2 (4.0) 98.1 (2.7) 91.0 (4.3) 85.6 (2.1) 94.4 (3.1) 92.2 (2.0)

15 (nEO ) 1) 5 (nEO ) 2) 15 (nEO ) 1) 5 (nEO ) 2) 10 10 20 20 25 (nEO ) 1) 10 (nEO ) 2) 5 (nEO ) 3-15) 25 (nEO ) 1) 10 (nEO ) 2) 5 (nEO ) 3-15) 5 5 5 5

65.2 (8.3) 68.6 (9.1) 66.4 (7.2) 73.9 (4.2) nd nd 63.1 (4.2) 59.0 (5.6) 75.9 (11.2) 77.8 (5.6) 78.1 (5.4) 81.2 (7.2) 76.5 (4.8) 80.3 (5.3) 73.5 (7.8) 75.7 (7.1) 61.1 (5.9) 65.0 (5.7)

10 (nEO ) 1) 2.5 (nEO ) 2) 10 (nEO ) 1) 2.5 (nEO ) 2) 5 5 10 10 10 (nEO ) 1) 4 (nEO ) 2 ) 2.5 (nEO ) 3-15) 10 (nEO ) 1) 4 (nEO ) 2) 2.5 (nEO ) 3-15) 2 2 2 2

76.4 (5.6) 78.3 (6.7) 75.3 (4.7) 80.6 (4.9) 91.4 (5.1) 88.3 (4.9) 76.2 (8.9) 73.5 (10.2) 88.1 (4.5) 89.6 (5.4) 90.8 (8.0) 90.1 (4.7) 90.8 (6.8) 86.6 (3.9) 90.4 (3.8) 78.5 (7.1) 88.1 (7.9) 90.2 (4.2)

(N ) 3). b Concentration factor 500. c Concentration factor 2. d Concentration factor 5.

Table 5. Concentrations Found in Environmental, WWTP, and DWTP Samples sample

BrNPEC

raw water (µg/L) prechlorinated water (µg/L) treated water (µg/L) flocculation sludge (µg/kg)

ndb 0.42 (nEO ) 0-1) nd nd

sample

NPEO

Halogenated Alkylphenolic Compounds BrNP BrNPEO ClNPEO nd 0.21 nd 220

DWTP Barcelona nd 4.0 nd 430 (nEO ) 1-2) 1600 (nEO > 3)

nd nd nd 660 (nEO ) 1-15)a

Alkylphenolic Compounds OPEO NP1EC NP2EC DWTP Barcelona 2.2 1.0 nd 915

BrOPEO

ClOPEO

nd nd nd 40 (nEO ) 1-15)a

nd nd nd 15 (nEO ) 1-15)a

OP1EC

OP2EC

NP

OP

2.9 1.3 nd 3200

nd nd nd 204

nd nd nd 348

0.45 0.06 nd 275

0.12 nd nd 30

raw water (µg/L) prechlorinated water (µg/L) treated water (µg/L) flocculation sludge (µg/kg)

12.9 8.8 nd 5730

1.2 nd nd 1315

water (µg/L) sediment (µg/kg)

4.4 350

0.6 20

Llobregat River 3.5 nd

3.3 nd

nd nd

nd nd

1.3 235

nd 15

influent (µg/L) effluent (µg/L) sludge (mg/kg)

1850 25 135

78 6.5 12

STP Igualada 13 58 2.4

6.5 22 nd

2.0 25 nd

8.5 19 nd

82 12 172

3.9 1.2 7.5

a

Estimated concentration assuming the same response as for BrNPEOs. b Not detected.

Precursors of these compounds, APEOs and their degradation products APECs and APs were found in all tested samples. The only exception was the final effluent from DWTP (drinking water) in which neither APEOs nor their degradation products were detected. In the Llobregat River, NPEOs and NP were found in both water and sediment samples, but more polar acidic byproducts (NPECs) are detected only in water. The concentration of OPEOs was significantly lower, reflecting lower commercial use of OPEOs (approximately 15-20% of APEO formulations). Levels found in river water and sediment correspond well with the concentrations found in effluent from STP Igualada that discharges treated water into the river. Concentrations found in STP effluent 5894 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

were approximately 3-10 times higher than those found in receiving water of the Llobregat River. Several studies16,18-22 reported similar levels of APEOs and their degradation products in STP samples, river water, and sediment; however, this study reports for the first time the levels of their halogenated derivatives. (18) Ahel, M.; Giger, W.; Koch, M. Water Res. 1994, 28, 1131-1142. (19) Ahel, M.; Giger, W.; Schaffner, C. Water Res. 1994, 28, 1143-1152. (20) Lye, C. M.; Frid, C. L. J.; Gill, M. E.; Cooper, D. W.; Jones, D. M. Environ. Sci. Technol. 1999, 33, 1009-1014 (21) Field, J. A.; Reed, R. L. Environ. Sci. Technol. 1996, 30, 3544-3550. (22) Naylor, C. G.; Mieure, J. P.; Adams, W. J.; Weeks, J. A.; Castaldi, F. J.; Ogle, L. D.; Romano, R. R. JAOCS 1992, 69, 695-703.

CONCLUSIONS A new methodology is presented in this work for the simultaneous and unequivocal determination of halogenated byproducts of alkylphenol ethoxylates and their metabolites. The method is the first protocol applicable to a full range of halogenated APEOs metabolites formed during chlorine disinfection in the presence of bromine ions. The optimum results, in terms of sensitivity, selectivity, and reproducibility, were obtained using LiChrolute C18 disposable cartridges for SPE followed by LC/MS using electrospray ionization (ESI). Using the specificity of ESI-MS detector, the method permits the simultaneous determination of BrNPEOs, ClNPEOs, BrOPEOs, ClOPEOs, BrNPECs, ClNPECs, BrNP, ClNP, and nonhalogenated analogues at the sub-parts-per-billion level in aqueous samples and at a low parts-per-billion level in solid samples. This protocol can be applied to a variety of samples, for

example, surface waters and river sediments, industrial and domestic wastewaters, and sludges from different origins. ACKNOWLEDGMENT This work has been supported by the EU Environment and Climate program through the Wastewater Cluster project SANDRINE (ENV4-CT98-0801) and by CICYT (AMB1999-1705-CE). We thank Merck for supplying the SPE cartridges and LC column, respectively. M. Petrovic and A. Diaz acknowledge grants from the Spanish Ministry of Education and Culture (SB97-B09092411) and from Fundacio´ AGBAR, respectively. Received for review June 19, 2001. Accepted August 29, 2001. AC010677K

Analytical Chemistry, Vol. 73, No. 24, December 15, 2001

5895