Analysis of Water Contaminants and Natural Water Samples Using

Oliver BOTTA , Zita MARTINS , Christian EMMENEGGER , Jason P. DWORKIN , Daniel P. GLAVIN , Ralph P. HARVEY , Renato ZENOBI , Jeffrey L. BADA ...
1 downloads 0 Views 121KB Size
Anal. Chem. 2000, 72, 3671-3677

Analysis of Water Contaminants and Natural Water Samples Using Two-Step Laser Mass Spectrometry Thomas D. Bucheli, Olivier P. Haefliger, Rolf Dietiker Jr., and Renato Zenobi*

Department of Chemistry, Swiss Federal Institute of Technology (ETH), Universita¨tstrasse 16, CH-8092 Zurich, Switzerland

The applicability of two-step laser mass spectrometry (L2MS) to the analysis of water contaminants and environmental water samples is demonstrated. First, the ionization characteristics of a selection of naphthyl and carbamate pesticides and of phenol were determined. The ion signal of all compounds increased with ionization laser pulse energy, within the investigated range (20-200 µJ). Ion yields relative to an internal standard, benz[a]anthracene, reached 30% for naphthyl pesticides ionized at 225 nm and 2-8% at 266 nm. At 266 nm, similar relative ion yields were found for phenol. Carbamate pesticides showed lower relative ion yields at all wavelengths, by a factor of ∼10-100, but higher relative ion yields, on the order of 1%, were obtained when using short (ps) laser pulses for ionization. These data allow one to estimate the detection limits of these analytes in a variety of matrixes once they are known for one of the compounds. Second, the quantitative analysis of carbaryl, phenol, and polycyclic aromatic hydrocarbons in rainwater is demonstrated. The aqueous samples were frozen to permit direct L2MS analysis of organic pollutants without tedious sample preparation. Detection limits were in the low-microgram per liter concentration range and recoveries of phenol from spiked rainwater samples were above 90%. The specific advantages are exemplified with the investigation of dynamic washout processes of atmospheric organic pollutants with a resolution of 0.01 mm of precipitation. The vast number of anthropogenic compounds released into the environment, their persistence and ubiquity, and their potential adverse effects to both environmental and human health represent major challenges in environmental analytical chemistry. For these reasons, developments in instrumentation are increasingly aiming toward on-line, in situ, and high-throughput analyses. One emerging technique in environmental analysis, two-step laser mass spectrometry (L2MS), offers unique advantages over conventional trace analytical methods. Virtually no sample preparation and only small sample volumes are required, measurements with high sensitivity can be performed within minutes, and ionization matrix effects are largely excluded. In the first step, an infrared laser pulse desorbs or ablates intact neutral molecules from the surface of a solid substrate. In the second step, a pulse from a tunable ultraviolet laser is used for resonance-enhanced multi* Corresponding author: (phone) +41 1 632 4376; (fax) +41 1 632 1292; (e-mail) [email protected]. 10.1021/ac000075l CCC: $19.00 Published on Web 07/07/2000

© 2000 American Chemical Society

photon ionization (REMPI) of the desorbed analytes. Mass analysis is then performed in a time-of-flight mass spectrometer (TOF-MS). L2MS has been used to investigate a variety of matrixes, such as soils,1 geosorbents,2 meteoritic samples,3 polymers,4 and kerogens.5 One of its most promising applications for routine environmental analytical chemistry is the characterization of atmospheric aerosol particles.6-8 Given the high versatility of L2MS in terms of measurable matrixes, it is surprising that frozen aqueous samples have not been given more attention.9-11 Moreover, routine measurements of real aqueous environmental samples have not, to the best of our knowledge, been presented yet. REMPI is an optically selective ionization technique that is often used in L2MS. For common ultraviolet laser sources, it requires the analytes to exhibit a strong absorption in the midUV range. Applications have therefore mainly focused on aromatic compound classes such as polycyclic aromatic hydrocarbons (PAHs),1,12,13 derivatized amino acids,14 biomolecules such as amino acids and peptides containing aromatic residues,15,16 porphyrins,17 chlorophylls,18 polymer additives, such as phenolic (1) Dale, M. J.; Jones, A. C.; Pollard, S. J. T.; Langridge-Smith, P. R. R. Environ. Sci. Technol. 1993, 27, 1693. (2) Gillette, J. S.; Luthy, R. G.; Clemett, S. J.; Zare, R. N. Environ. Sci. Technol. 1999, 33, 1185. (3) Clemett, S. J.; Maechling, C. R.; Zare, R. N.; Swan, P. D.; Walker, R. M. Science 1993, 262, 721. (4) Zhan, Q.; Zenobi, R.; Wright, S. J.; Langridge-Smith, P. R. R. Macromol. 1996, 29, 7865. (5) Zhan, Q.; Zenobi, R.; Buseck, P. R.; Teerman, S. Energy Fuels 1997, 11, 144. (6) Dale, M. J.; Downs, O. H. J.; Costello, K. F.; Wright, S. J.; Langridge-Smith, P. R. R. Environ. Pollut. 1995, 89, 123. (7) Haefliger, O. P.; Bucheli, T. B.; Zenobi, R. Environ. Sci. Technol. 2000, 34, 2178. (8) Haefliger, O. P.; Bucheli, T. B.; Zenobi, R, Environ. Sci. Technol. 2000, 34, 2184. (9) Alimpiev, S. S.; Mlynski, V. V.; Belov, M. E.; Nikiforov, S. M. Anal. Chem. 1995, 67, 181. (10) Belov, M. E.; Alimpiev, S. S.; Mlynsky, V. V.; Nikiforov, S. M.; Derrick, P. J. Rapid Commun. Mass Spectrom. 1995, 9, 1431. (11) Bernstein, M. P.; Sandford, S. A.; Allamandola, L. J.; Gillette, J. S.; Clemett, S. J.; Zare, R. N. Science 1999, 283, 1135. (12) Wilkerson, C. W., Jr.; Colby, S. M.; Reilly, J. P. Anal. Chem. 1989, 61, 2669. (13) Hankin, S. M.; John, P.; Simpson, A. W.; Smith, G. P. Anal. Chem. 1996, 68, 328. (14) Engelke, F.; Hahn, J. H.; Henke, W.; Zare, R. N. Anal. Chem. 1987, 59, 909. (15) Grotemeyer, J.; Schlag, E. W. Org. Mass Spectrom. 1988, 23, 388. (16) Reilly, P. T. A.; Reilly, J. P. Rapid Commun. Mass Spectrom. 1994, 8, 731. (17) Morris, J. B.; Johnston, M. V. Int. J. Mass Spectrom. Ion Processes 1986, 73, 175.

Analytical Chemistry, Vol. 72, No. 15, August 1, 2000 3671

Figure 1. Naphthyl and carbamate pesticides investigated. Table 1. Naphthyl and Carbamate Pesticides Studieda

antioxidants and hydroxyphenylbenzotriazole UV stabilizers,4 and azo dyes.19 Vacuum UV radiation for single-photon ionization permits the analysis of nonaromatic compounds.20,21 Many pesticides are possible candidates for detection by REMPI mass spectrometry but have only received marginal attention until now.22,23 Also, systematic studies applying L2MS to the analysis of these environmentally important compounds have not yet been performed. This paper expands the applications of L2MS described above with respect to both analyte selection and investigated matrixes. This is achieved first by systematically determining the ionization efficiencies of several environmentally relevant pesticides of the naphthyl and carbamate compound classes (Figure 1 and Table 1) dispersed in a PVC membrane matrix. The major advantage of this matrix embedding technique is that it provides a thin, homogeneous, and vacuum-stable sample from which laser ablation can directly be performed, thus yielding quantitative signals over several orders of magnitude.24 Detailed information on the ion yields of these compounds is obtained, a prerequisite for characterizing the potential of their analysis in environmental samples. Second, initial results on the application of L2MS to the analysis of organic substances in natural waters are shown. Aqueous samples were frozen to render them vacuumstable, and laser ablation was performed directly from the ice. (18) Grotemeyer, J.; Bosel, U.; Walter, K.; Schlag, E. W. J. Am. Chem. Soc. 1986, 108, 4233. (19) Dale, M. J.; Jones, A. C.; Langridge-Smith, P. R. R.; Costello, K. F.; Cummings, P. G. Anal. Chem. 1993, 65, 793. (20) Kornienko, O.; Ada, E. T.; Tinka, J.; Wijesundara, M. B. J.; Hanley, L. Anal. Chem. 1998, 70, 1208. (21) Butcher, D. J.; Goeringer, D. E.; Hurst, G. B. Anal. Chem. 1999, 71, 489. (22) Benazouz, M.; Hakim, B.; Debrun, J. L. Rapid Commun. Mass Spectrom. 1998, 12, 1018. (23) Orea, J. M.; Besco´s, B.; Montero, C.; Gonzales Uren ˜a, A. Anal. Chem. 1998, 70, 491. (24) Haefliger, O. P.; Zenobi, R. Rev. Sci. Instrum. 1998, 69, 1828.

3672 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

name carbaryl naphthyl acetic acid (NaAcAc) naphthyl acetamide (NaAcAm) carbofuran bendiocarb carbendazime metolcarb propoxur propham

molecular mass (Da)

mass spectral peaksb (Da)

λmax (nm)

201 186

144 186, 141

221.5, 280c 223.5, 280c

185

185, 141

223.5, 280c

221 223 191 165 209 179

164 166, 151 191 108 110 179, 137, 120, 93

200, 282d 200, 281d 281.1f nag 212, 275d 236d

a For each compound, the wavelengths of maximum UV absorbance, the molecular mass, and the main mass spectral peaks are given. b Italic peaks were used for quantification. c Sadtler UV spectra library. d Gynkotek pesticide UV spectra library (Gynkotek, Germering, Germany). e Investigated with picosecond laser only. f From ref 23.g na, not available.

The dynamic washout process of atmospheric compounds at m/z ) 94 (most likely, phenol) and at m/z ) 166 (fluorene) during a rain event was monitored. EXPERIMENTAL SECTION Materials. Methanol (MeOH; p.a. >99.5%) and dichloromethane (p.a. >99.8%) were from Baker (Deventer, The Netherlands). Tetrahydrofuran (THF, >99%) was from Scharlau (Barcelona, Spain). Deionized water was further purified with a Nanopure water purification device (NANOpure, Barnstead, Allschwil, Switzerland). Poly(vinyl chloride) (PVC) powder was obtained from Fluka (Buchs, Switzerland). Benzo[a]pyrene (>98%) was obtained from Aldrich (Steinheim, Germany) and naphthalene (>99%) from Fluka (Buchs, Switzerland). All other PAHs (fluorene, phenanthrene, 9-methylanthracene, pyrene, 9,10-dimethy-

lanthracene, 1-methylpyrene, and benz[a]anthracene (BaA); all >98%) were purchased from Chem Service (West Chester, PA). Phenanthrene-d10 (isotopic purity, 98%) was from Cambridge Isotope Laboratories (Andover, MA). Phenol and all pesticides (atrazine, bendiocarb, carbofuran, carbosulfan, carbaryl, ethiofencarb, folpet, imazaquin, isoproturon, metolcarb, metamitron, metsulfuron-methyl, naphthyl acetic acid (NaAcAc), naphthyl acetamide (NaAcAm), propham, propoxur, and warfarin; all >97%) were purchased from Riedl-de Hae¨n (Seelze, Germany). Figure 1 shows structures of some naphthyl and carbamate pesticides that were investigated in greater detail. Standard Solutions. For subsequent preparation of PVC membranes, 2 g/L of each pesticide and phenol was dissolved in MeOH. For the direct water analyses, stock solutions of the individual PAHs, of phenol, and of the internal standard phenanthrene-d10 were prepared at 200 mg/L in CH2Cl2. All solutions used for water analysis were further diluted with MeOH to obtain a concentration of 20 mg/L per compound. Preparation of Membranes for Ionization Studies of Pesticides. Membranes were prepared by dissolving 3 g of PVC powder in 50 mL of THF as described by Haefliger and Zenobi.24 After complete dissolution of the PVC, 50 µL of the internal standard BaA, and 50 µL (naphthyl pesticides), 2 mL (carbamate pesticides), or 5 mL (all other compounds) of the respective stock solutions were added, and the flasks were stirred vigorously for several minutes. Next, 2 mL of the solution was poured into a Petri dish (5 cm diameter). After at least 1 h, a completely dry, ∼50-µm-thick membrane was obtained. Water Calibration Solutions. From the 20 mg/L stock solutions, mixtures containing all PAHs and phenol in MeOH were prepared and added to Nanopure water to yield the desired concentrations (15 different levels covering a range from 0.15 to 100 nM) and a constant MeOH content of 1%. The MeOH content was kept constant as it was observed to influence the freezing process and ice consistency. The internal standard phenanthrened10 was added to all samples at a level of 1 µg/L. Rainwater Samples. Sequential rainwater sampling was performed in Du¨bendorf (Switzerland) with a 5-m2 funnel coated with a Teflon foil as described earlier.25 The sampler was covered with a plastic sheet during dry periods and washed with ethanol and Nanopure water prior to sampling. During the first 0.15 mm of precipitation, 15 50-mL samples were collected, corresponding to a resolution of 0.01 mm. The sample volume was increased to 1 L for the next 2 mm of rain. After the 25th sample, only every fifth liter was kept and the 4 L between discarded. All natural water samples were filtered (mixed cellulose nitrate and acetate, 1.2-µm pore size, Millipore) and spiked with 1 µg/L phenanthrene-d10. All samples were stored at 5 °C in the dark. To determine the recoveries of phenol in natural waters, 10 mL of two rainwater samples were fortified with 9.6, and 19.3 µg/L, respectively. L2MS System. A home-built L2MS system was used that has been described in detail elsewhere.26,27 A multimode CO2 laser (Alltec 853 MS, Lu¨beck, Germany; λ ) 10.6 µm, 0.6 J/cm2, 107.5 ns) with a 45° incident angle relative to the sample surface was (25) Bucheli, T. D.; Mu ¨ ller, S. R.; Heberle, S.; Schwarzenbach, R. P. Environ. Sci. Technol. 1998, 32, 3457. (26) Haefliger, O. P.; Zenobi, R. Anal. Chem. 1998, 70, 2664. (27) Voumard, P.; Zhan, Q.; Zenobi, R. Rev. Sci. Instrum. 1993, 64, 2215.

used for ablation. A remote-controlled motorized stage allowed the sample to be rotated to expose a new spot for each ablation laser shot. The ionizing laser radiation was provided by an optical parametric oscillator (OPO) laser (MOPO-730D10, Spectra Physics Lasers Inc., Mountain View, CA), pumped by the third harmonic of a Nd:YAG laser (GCR-230, Spectra Physics Lasers Inc.). Wavelengths between 225 and 280 nm were used for ionization efficiency studies; the pulse width was 8 ns. For water analysis, the laser was tuned to 250 nm. The UV energy was measured with a pyroelectric power detector (ED-100, Gentec, Sainte-Foy, Canada) and adjusted to the desired value with a remote-controlled motorized CaF2 polarizer. The delay between the ablation and postionization lasers was set to 10 µs for membrane measurements and to 8 µs for frozen water samples. These values were the results of optimizations to achieve maximum signal intensities. The shorter delay found for water samples might be indicative of a faster expansion of the water than of the PVC plume. With the goal to obtain better ion yields for compounds with short-lived excited states, additional experiments were conducted with the fourth harmonic (266 nm) of a ps-Nd:YAG laser (PY-61C/20, Continuum, Puchheim, Germany) that had a pulse width of only 35 ps. Mass spectrometric analysis was performed with a reflectron time-of-flight mass spectrometer (R. M. Jordan Co., Grass Valley, CA). For water analysis, a liquid nitrogen trap was built in the ionization chamber to avoid recondensation of evaporated water onto the sample surface. L2MS Measurements of PVC Membranes. Circular pieces 12 mm in diameter were cut from the PVC membranes, mounted onto the L2MS sample holder, and introduced into the instrument. Seventy-two single laser shot measurements were performed on each piece of membrane. Mass spectra for which the UV energy was outside of a 22-32 µJ range, as well as those during which plasma formation caused by the ablation laser was observed, were rejected.24 To determine the relative ionization efficiencies, the concentration-normalized signal intensities (peak heights) of the analytes’ major fragment ion (see Table 1) were divided by the one of BaA at m/z ) 228. L2MS Measurement of Water Samples. To maximize the reproducibility, the water samples were prepared and measured according to a procedure optimized in preliminary experiments. A 200-µL water sample was carefully poured into a brass cup (inner diameter, 11.55 mm; depth, 1.8 mm; outer diameter, 12,8 mm; total height, 3.2 mm) which had been precooled in liquid nitrogen for 20 s. Within 1 min, the water was completely frozen and the sample was again placed into liquid nitrogen for 1 min. The sample was then mounted onto the sample holder and introduced into the instrument. After introduction of the sample, the vacuum chamber was allowed to equilibrate for 2 min. The sample rotator was then switched on and the measurement started. The first 120 shots (2 min) were discarded, as they could have been influenced by surface effects. Thereafter, measurements were performed up to a total of 600 shots. Signal intensities (peak heights) of the molecular ion peaks in the individual mass spectra were normalized to the one of phenanthrene-d10 at m/z ) 188. RESULTS AND DISCUSSION Ionization Efficiencies and Fragmentation of Pesticides. UV ionization of the investigated naphthyl and carbamate pesticides between 250 and 280 nm generally yielded one major Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

3673

Figure 2. Mass spectra of a selection of the investigated naphthyl and carbamate pesticides ablated from PVC membranes and postionized at 225 nm (averages of 5-35 individual spectra). The major fragment ions are labeled. Asterisks indicate artifact ions (see text).

fragment ion, with slightly increased fragmentation observed at 225 nm (Table 1 and Figure 2). The major fragmentation pathway for carbaryl, propoxur, metolcarb, carbofuran, and bendiocarb proceeds through a McLafferty rearrangement followed by cleavage of the carbamate C-O bond. Molecular ions were observed for NaAcAc, NaAcAm, carbendazim, and propham at most, but not all, of the wavelengths used. Propham fragmented by loss of m/z ) 42 (C3H6) following a McLafferty rearrangement between the isopropyl group and the carbonyl oxygen. For NaAcAc and NaAcAm, the dominant fragmentation was radical-induced R-cleavage mediated by the carbonyl oxygen, leading to m/z ) 141, although a McLafferty product (m/z ) 142) was also observed for both compounds. The ions at m/z ) 128 and 115 were regularly observed when working with PVC membranes and probably stem from impurities, although m/z ) 115 can also result from fragmentation of the naphthalene moiety of some of the pesticides. The signal intensities of the ions related to pesticides grew with increasing laser pulse energy over the investigated range of 20-200 µJ for all compounds and at all wavelengths. In contrast, no increase in the ion yield of BaA was observed above 30 µJ. After the absorption of a third photon, this compound already begins to fragment at moderate laser pulse energies, a finding that is in agreement with ref 24. However, overall ion yields for PAHs such as benz[a]anthracene were larger than those of pesticides, probably due to a larger UV absorption cross section 3674 Analytical Chemistry, Vol. 72, No. 15, August 1, 2000

Figure 3. Ionization efficiencies of selected naphthyl pesticides (a) and carbamate pesticides (b) relative to BaA (values are averages from n ) 5 to 35 individual mass spectra, and error bars represent standard deviations; nd, not detected; na, not analyzed).

and/or a longer lived excited state. UV absorbance is therefore only one of the parameters that determine the sensitivity reached by L2MS; fragmentation has to be taken into account as well. For maximum sensitivity, the laser wavelength, the pulse width, and the pulse energy can be optimized for the analyte peak to be detected. For example, the lower UV absorbance of pesticides relative to PAHs can be compensated by the possibility of working at higher laser fluence. For directly comparing the ion yield of different compounds at different wavelengths, a normalization to BaA was performed. BaA was chosen because its REMPI spectrum shows a fairly constant ion signal over most of the wavelength range under investigation (238-280 nm)26 and because linear calibration curves for the quantitation of this compound embedded in PVC membranes have been obtained earlier.24 Despite the much higher ion yields at elevated ionization energies for both carbamate and naphthyl pesticides, the UV energy was kept between 22 and 32 µJ per laser pulse to ensure a linear response of the internal standard BaA. Figure 3 depicts the ionization efficiencies of a selection of naphthyl and carbamate pesticides relative to BaA. In all cases, the ion yields at different ionization wavelengths corresponded well with the UV absorbance of the analytes (see Table 1 for wavelengths of maximum UV absorbance). Maximum ion yields were found at 225 nm. For most compounds, ions could be detected over the whole UV range from 225 to 280 nm. This indicates that the ablated analytes were not subject to jet-cooling

effects, which would result in substantial narrowing of the optical spectra.28 Generally, the naphthyl pesticides showed a roughly 10-100-fold higher ion yield than the carbamate pesticides, with maximum ionization efficiencies of up to 30% relative to BaA. The observed relative ionization efficiency for phenol (2.6 ( 1.0% at 266 nm) is in the same range as that of the naphthyl pesticides (Figure 3a). Other carbamate pesticides that were studied, such as ethiofencarb or carbosulfan, were not ionizable (