Anal. Chem. 2011, 83, 60–66
Analysis of Dioxins by Gas Chromatography/ Resonance-Enhanced Multiphoton Ionization/Mass Spectrometry Using Nanosecond and Picosecond Lasers Adan Li,*,†,§ Tomohiro Uchimura,†,‡ Yuka Watanabe-Ezoe,† and Totaro Imasaka†,‡ Department of Applied Chemistry, Graduate School of Engineering, and Division of Translational Research, Center for Future Chemistry, Kyushu University, 744 Motooka, Nishiku, Fukuoka 819-0395, Japan, and College of Chemistry and Environmental Engineering, Yanshan University, Qinhuangdao 066004, China Dioxins in a soil sample were measured using gas chromatography/resonance-enhanced multiphoton ionization/ time-of-flight mass spectrometry coupled with different types of laser sources. The fourth-harmonic emission (266 nm) of a nanosecond Nd:YAG laser (1 ns) provided low ionization efficiency, especially for highly chlorinated dioxins/dibenzofurans (CDDs/CDFs). The ionization efficiency was improved using the fourth-harmonic emission (266 nm) of a picosecond Nd:YAG laser (4 ps), due to shorter singlet excited-state lifetimes. It was, however, difficult to efficiently ionize hepta-CDD and octa-CDD/ CDF, because of their shorter lifetimes, which were induced by stronger spin-orbit coupling that led to efficient relaxation of the excited molecule to triplet levels. The ionization efficiency was substantially improved using the fifth-harmonic emission (213 nm) of the picosecond Nd:YAG laser (4 ps), in which the analyte molecule that was relaxed to triplet levels was efficiently ionized using a photon with sufficient energy for ionization, although the pulse energy obtained at 213 nm was only one-third of the pulse energy obtained at 266 nm. The limits of detection achieved for 17 toxic polychlorinated dibenzop-dioxins/polychlorinated dibenzofurans (PCDDs/PCDFs) were 0.41-45 pg. The analytical instrument developed in the present study performed sufficiently well for the practical trace analysis of dioxins in soil samples. Dioxins, such as polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs), are emitted into the atmosphere as undesirable byproducts formed in waste incineration and pulp/paper bleaching and also as impurities present in pesticides used in the agriculture. Due to the high toxicity and numerous congeners of dioxins, a sensitive and selective analytical method is required for identification and subsequent quantification. Gas chromatography combined with mass spectrometry * To whom correspondence should be addressed. E-mail:
[email protected]. † Department of Applied Chemistry, Graduate School of Engineering, Kyushu University. ‡ Division of Translational Research, Center for Future Chemistry, Kyushu University. § College of Chemistry and Environmental Engineering, Yanshan University.
60
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
based on electron impact ionization (GC/EI/MS) has been successfully used in trace analysis of these compounds.1-4 However, this approach creates numerous fragments, thus making identification and quantification of the analytes difficult. Then, a high-resolution sector-type mass spectrometer (HRMS) is employed to reduce the background and to accomplish identification of the mass signal of the analyte.5-7 However, EI has no selectivity in the process of ionization; all the molecules including aliphatic compounds are simultaneously ionized. A laser ionization technique has been studied to avoid these undesirable results. The laser wavelength can be adjusted to excite the analyte molecule to intermediate levels, to enhance the ionization efficiency. This technique, referred to as resonance-enhanced multiphoton ionization (REMPI), is superior for the selectivity and sensitivity.8-11 In addition, REMPI is a technique of “soft” ionization, in contrast to the “hard” ionization in EI, and is achieved only when the second photon has sufficient energy for subsequent ionization from the intermediate level. This technique has been applied to the determination of aromatic compounds, due to their selective absorption in the near-ultraviolet region,12-15 and is further
(1) Grochowalski, A.; Lassen, C.; Holtzer, M.; Sadowski, M.; Hudyma, T. Envviron. Sci. Pollut. Res. 2007, 14, 326–332. (2) Tsytsik, P.; Czech, J.; Carleer, R. J. Chromatogr., A 2008, 1210, 212–221. (3) Fishman, V. N.; Martin, G. D.; Lamparski, L. L. J. Chromatogr., A 2007, 1139, 285–300. (4) Watanabe-Ezoe, Y.; Li, X.; Imasaka, T.; Uchimura, T.; Imasaka, T. Anal. Chem. 2010, 82, 6519–6525. (5) Fishman, V. N.; Martin, G. D.; Lamparski, L. L. J. Chromatogr., A 2004, 1057, 151–161. (6) Abad, E.; Caixach, J.; Rivera, J. J. Chromatogr., A 1997, 786, 125–134. (7) Menotta, S.; D’antonio, M.; Diegoli, G.; Montella, L.; Raccanelli, S.; Fedrizzi, G. Anal. Chim. Acta 2010, 672, 50–54. (8) Boesl, U.; Weinkauf, R.; Weickhardt, C.; Schlag, E. W. Int. J. Mass Spectrom. Ion Processes 1994, 131, 87–124. (9) Clark, A.; Ledingham, K. W. D.; Marshall, A. Spectrochim. Acta B 1992, 47, 799–808. (10) Marshall, A.; Clark, A.; Jennings, R.; Ledingham, K.W. D.; Sander, J.; Singhal, R. P. Int. J. Mass Spectrom. Ion Process. 1992, 116, 143–156. (11) Dale, M. J.; Jones, A. C.; Langridge-Smith, P. R. R.; Costello, K. F.; Cummins, P. G. Anal. Chem. 1993, 65, 793–801. (12) Haefliger, O. P.; Zenobi, R. Anal. Chem. 1998, 70, 2660–2665. (13) Uchimura, T.; Imasaka, T. Anal. Chem. 2000, 72, 2648–2652. (14) Matsumoto, J.; Lin, C.-H.; Imasaka, T. Anal. Chem. 1997, 69, 4524–4529. (15) Lubman, D. M. Anal. Chem. 1987, 59, 31A–40A. 10.1021/ac101849w 2011 American Chemical Society Published on Web 12/02/2010
combined with GC to improve selectivity in practical trace analysis.16-19 A number of investigations based on REMPI, using different types of lasers, have been reported.14,17,19-22 Lasers with a shorter pulse duration provide higher ionization efficiency, especially for analyte molecules with short singlet excited-state lifetimes. For example, the ionization efficiency of halogenated aromatic hydrocarbons with short lifetimes was substantially improved when a nanosecond laser was replaced with a femtosecond laser.19,23 This could be explained by efficient ionization from the singlet excited state, before relaxation to triplet levels by intersystem crossing through a spin-orbit interaction. Although a femtosecond Ti: Sapphire laser can be operated for more than 1 year without maintenance, it is large and rather expensive, so it is not suitable for application to practical trace analysis. It is, therefore, desirable to use a compact and low-cost laser, such as a nanosecond or picosecond laser, to ionize dioxin molecules. On the other hand, a far-ultraviolet nanosecond laser emitting at 213 nm has been successfully used for efficient ionization of 2,3,7,8-tetraCDF (2,3,7,8-TeCDF) and 2,8-dichlorinated dibenzo dioxin (2,8-DiCDD), from triplet levels after relaxation by intersystem crossing.21,24,25 In the present study, we report on the advantages of the farultraviolet picosecond laser (213 nm) for efficient ionization of the analyte molecule from both the singlet and triplet excited states. The limit of detection (LOD) achieved using this technique was compared with that obtained using nanosecond and picosecond lasers emitting at 266 nm. The GC/REMPI/MS system developed herein was applied to the analysis of dioxins in a soil sample and was found to perform sufficiently well to evaluate the environmental contamination level. EXPERIMENTAL SECTION Reagents and Analytes. Standard samples of 2,3,7,8-tetraCDF (2,3,7,8-TeCDF), 1,2,3,7,8-pentaCDF (1,2,3,7,8-PnCDF), 2,3,4,7,8PnCDF, and 1,2,3,4,8,9-hexaCDF (1,2,3,4,8,9-HxCDF) were purchased from Wellington Laboratories (Guelph, Ontario, Canada) for the experiment using a nanosecond Nd:YAG laser (266 nm). Standard solutions (DFJ-CAL-A) containing 17 toxic dioxins and their 13C-labled isotopes (CS3-A to CS5-A) were obtained from Wellington Laboratories for the studies using picosecond Nd: YAG lasers emitting at 266 and 213 nm. A sample was extracted from a soil (10 g) with a curtailed procedure based on accelerated solvent extraction (ASE) using a (16) Rhodes, G.; Opsal, R. B.; Meek, J. T.; Reilly, J. P. Anal. Chem. 1983, 55, 280–286. (17) Wilkerson, C. W., Jr.; Colby, S. M.; Reilly, J. P. Anal. Chem. 1989, 61, 2669–2673. (18) Li, A.; Uchimura, T.; Tsukatani, H.; Imasaka, T. Anal. Sci. 2010, 26, 841– 846. (19) Yamaguchi, S.; Kira, F.; Miyoshi, Y.; Uchimura, T.; Watanabe-Ezoe, Y.; Zaitsu, S.; Imasaka, T.; Imasaka, T. Anal. Chim. Acta 2009, 632, 229–233. (20) Oser, H.; Copic, K.; Coggiola, M. J.; Faris, G. W.; Crosley, D. R. Chemosphere 2001, 43, 469–477. (21) Kirihara, N.; Takahashi, K.; Kitada, N.; Tanaka, M.; Suzuki, Y. Rev. Sci. Instrum. 2006, 77, 094101/1-094101/9. (22) Tsukatani, H.; Okudaira, H.; Uchimura, T.; Imasaka, T.; Imasaka, T. Anal. Sci. 2009, 25, 599–604. (23) Matsumoto, J.; Lin, C.-H.; Imasaka, T. Anal. Chim. Acta 1997, 343, 129– 133. (24) Zimmermann, R. Organohalo. Compd. 2001, 54, 374–379. (25) Zimmermann, R.; Lenoir, D.; Kettrup, A.; Nagel, H.; Boesl, U. In 26th Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, 1996, 2859-2868.
commercial instrument (ASE-300, Dionex, Sunnyvale, CA) with toluene for 20 min. The sample solution was, then, concentrated using a rotary evaporator, and the volume was adjusted at 30 mL with hexane. Then, 2.5 mL of the solution was cleaned up using an automatic sample preparation device (SPD-600GC, Miura Co, Ltd., Ehime, Japan) for 2 h, and the solvent was then evaporated under a nitrogen flow. Then, a 20 µL solution of 13C-1,2,3,4,6PnCDF was added as a syringe spike (internal standard). The sample solution was finally concentrated to 20 µL before injection into the GC/REMPI/MS. Apparatus. The mass spectrometer used in the present study has been described elsewhere.18,26 The fourth-harmonic emission (266 nm, 4 ps, 200 Hz, 100 µJ, Compiler 266, Passat) and the fifthharmonic emission (213 nm, 4 ps, 200 Hz, 30 µJ, Compiler 213, Passat) of a picosecond Nd:YAG laser and the fourth-harmonic emission of a nanosecond Nd:YAG laser (266 nm, 1 ns, 1 kHz, 17 µJ, FQSS 266, Crylas) were utilized as ionization sources. A 1 µL sample solution was injected into a GC system (6890N, Agilent Technologies) using an auto sampler (7683B, Agilent Technologies). The constituents in the standard sample (CS3-A to CS5-A) were separated using an HP-5 capillary column (length, 30 m; i.d., 0.32 mm or 0.25 mm) and a carrier gas of helium at a constant flow-rate of 1 mL/min. The constituents of the soil sample were separated using column of DB-5 ms (length, 30 m; i.d., 0.25 mm) and DB-17 (length, 30 m; i.d., 0.25 mm) for two independent chromatographic experiments. The temperature program for the HP-5 and DB-5 ms columns was set to increase from 130 to 315 °C at a rate of 15 °C/min and then held for 5 min. For the DB-17 column, the temperature was programmed to increase at a rate of 15 °C/min from 130 to 280 °C and then held for 8 min. The temperatures of the injection port and the transfer line were maintained at 300 °C. The analyte eluting from the GC was introduced into a linear-type time-offlight mass spectrometer (TOF-MS). A microchannel plate (F4655-11, Hamamatsu) was used as a detector to measure the ions induced by multiphoton ionization. The voltages applied to electrodes in the mass spectrometer were optimized by monitoring the mass spectrum of the chemical species bleeding from a GC capillary column (HP-5) using a digital oscilloscope (1 GHz, 20 GS/s, DPO7104, Tektronix). Finally, the signal was recorded by means of a digitizer (1 GHz, 1 GS/s, Acqiris AP240, Agilent Technologies). The data were analyzed using LabVIEW software. RESULTS AND DISCUSSION Comparison of Nanosecond and Picosecond Lasers. Figure 1 shows the mass chromatograms obtained using a nanosecond laser emitting at 266 nm. The LOD, i.e., the concentration at which the signal-to-noise ratios was 3, is summarized in Table 1 for 2,3,7,8-TeCDF, 1,2,3,7,8-PnCDF, 2,3,4,7,8-PnCDF, and 1,2,3,4,8,9-HxCDF. The signal intensity, i.e., the ionization efficiency, strongly depends on the number of chlorine atoms, since intersystem crossing from the singlet excited state to triplet levels is enhanced by a spin-orbit interaction, referred to as a heavy atom effect. Therefore, it becomes more difficult to ionize highly chlorinated molecules using a nanosecond laser. The present result suggests that the lifetimes of the singlet excited state of such molecules are shorter than the pulse width of the laser. (26) Uchimura, T.; Sakai, K.; Imasaka, T. Anal. Chem. 2004, 76, 5534–5538.
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
61
Table 1. LODs Obtained by Nanosecond (266 nm) and Picosecond (266 and 213 nm) Lasersa LOD for standard samples (pg) compound 2,3,7,8-TeCDF 1,2,3,7,8-PnCDF 2,3,4,7,8-PnCDF 1,2,3,4,8,9-HxCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF 2,3,7,8-TeCDD 1,2,3,7,8-PnCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD
nanosecond picosecond picosecond laser (266 nm) laser (266 nm) laser (213 nm) 2.6 5.6 10 102
2.3 3.4 6.2
0.41 0.77 0.91
1.9 1.8 3 3.6 6 12 40 2.5 15 9.6 8.6 9.2 41 300
1.8 1.4 2.3 4.9 5 12 21 0.6 6.0 7.9 7.8 8.6 13 45
a The samples shown below 1,2,3,4,8,9-HxCDF were not measured due to low ionization efficiencies and high toxicities (eventually requiring the samples with high TEQ). A mixture sample containing 17 toxic dioxins was utilized in the experiments using the picosecond lasers: non-toxic 1,2,3,4,8,9-HxCDF was not contained in this sample.
Figure 1. Mass chromatograms obtained using a nanosecond laser emitting at 266 nm. These chromatograms were derived by monitoring the signals at (a) 2,3,7,8-TeCDF at m/z ) 306, (b) 1,2,3,7,8-PnCDF at m/z ) 340, (c) 2,3,4,7,8-PnCDF at m/z ) 340, and (d) 1,2,3,4,8,9HxCDF at m/z ) 374. Concentration: 50 pg/µL for 2,3,7,8-TeCDF, 1,2,3,7,8-PnCDF, and 2,3,4,7,8-PnCDF; 500 pg/µL for 1,2,3,4,8,9HxCDF. Column, HP-5 (i.d., 0.32 mm).
Accordingly, a nanosecond laser is not suitable for the efficient ionization of highly chlorinated PCDDs/PCDFs. A similar experiment was performed using a standard mixture sample (CS5-A) and a picosecond laser (266 nm, 4 ps). The mass chromatograms obtained are shown in Figure 2, and the LODs measured for 17 toxic PCDDs/PCDFs are summarized in Table 1. Although the pulse energy and the repetition rate of the laser were different from those of the nanosecond laser, the average power (20 mW) was similar to that of the nanosecond laser (17 mW). It should be noted that there was no appreciable difference in the LODs obtained using the nanosecond and picosecond lasers for TeCDF. Although the LODs substantially increased with nanosecond ionization in the order of TeCDF, PnCDF, and HxCDF, the LODs remained essentially unchanged for these compounds in the case of picosecond ionization. These results suggest that the pulse width of the laser (4 ps) was shorter than the singlet excited-state lifetimes of these congeners. Thus, a picosecond laser is superior to a nanosecond laser for trace analysis of highly chlorinated PCDDs/PCDFs. It was, however, difficult to efficiently ionize hepta-CDD (HpCDD) and octa-CDD/ 62
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
CDF (OCDD/OCDF), even when a picosecond laser was used. Therefore, a laser with a shorter pulse width, e.g., a femtosecond laser, would be necessary for trace analysis of these compounds. Comparison of Picosecond Lasers Emitting at 266 and 213 nm. Figure 3 shows a two-dimensional display obtained using a picosecond laser emitting at the wavelength of 213 nm. All the components, including 17 toxic congeners consisting of 12C-native and 13C-labled isotopes, were measured. A mass spectrum of 1,3,6,8-TeCDF obtained by extracting the signal from the twodimensional display is shown as an inset in Figure 3. All the isotope peaks are clearly separated due to sufficient mass resolution (m/∆m ) 1000) achieved in this study. Then, the number of chlorine atoms can readily be calculated from the intensity distribution of the isotope peaks. Several impurity signals besides PCDD/PCDFs can be seen in the two-dimensional display. As shown in the inset of Figure 3, nonchlorinated compound appears as a single spot at a retention time of 9 min and at m/z ) 409. Such impurity signals would appear for a real sample, but they could be differentiated from the signals of PCDD/PCDFs due to high selectivity of GC/REMPI/TOF-MS. The signal intensities for highly chlorinated PCDDs/PCDFs were not degraded significantly. The mass chromatogram was obtained by extracting the data along the specified m/z value. The LODs were then calculated for comparison with those obtained at 266 nm. The results are summarized in Table 1. In most cases, the LODs obtained at 213 nm were lower than the values obtained at 266 nm, although the pulse energy obtained at 213 nm was only onethird of the value obtained at 266 nm. Higher ionization efficiency was obtained at 213 nm than at 266 nm for TeCDD/TeCDFs and PnCDDs, mainly due to larger absorption cross sections for these molecules at around 220 nm.27,28 In ref 27, the absorption (27) Funk, D. J.; Oldenborg, R. C.; Dayton, D.-P.; Lacosse, J. P.; Draves, J. A.; Logan, T. J. Appl. Spectrosc. 1995, 49, 105–114.
Figure 2. Mass chromatograms obtained using a picosecond laser emitting at 266 nm. These chromatograms were derived by monitoring PCDD/PCDFs specified in the figures. Concentrations: 50 pg/µL for TeCDD/CDFs and PnCDD/CDFs; 100 pg/µL for HxCDD/CDFs and HpCDD/ CDFs; 250 pg/µL for OCDD/CDF. Column, HP-5 (i.d., 0.32 mm).
spectrum is reported as a function of the temperature for 2,3,7,8TeCDD/TeCDF and 1,2,3,7,8-TeCDD, providing two absorption bands at around 220 and 300 nm and a valley (260-280 nm) between them. Therefore, one of the explanations for higher ionization efficiency against TeCDD/TeCDFs and PnCDD/ PnCDFs would be more efficient excitation of the analyte molecules to a singlet excited state at around 213 nm. Although no absorption spectrum was reported for HxCDDs/ HxCDFs and HpCDDs/HpCDFs in the gas phase, the situation would be similar to that described above. However, the deterioration of the ionization efficiency by intersystem crossing would be more serious for HpCDDs/HpCDFs because of a large number of chlorine atoms. When a laser emitting at 266 nm is used, 2,3,7,8TeCDD reportedly cannot be ionized from the triplet level.29 Since the ionization potential increases as the number of chlorine atoms increases,30 HpCDD cannot be ionized from the triplet levels using (28) Gastilovich, E. A.; Klimenko, V. G.; Korol’kova, N. V.; Rauhut, G. Chem. Phys. 2001, 270, 41–54. (29) Kirihara, N.; Haruaki, Y.; Mizuho, T.; Kenji, T.; Norifumi, K.; Toru, S.; Yasuo, S. Organohalo. Compd. 2004, 66, 731–738. (30) Imasaka, T.; Hirokawa, S.; Imasaka, T. J. Mol. Struct. (Theochem.) 2006, 774, 7–12.
a 266 nm photon. Thus, the ionization efficiency decreased drastically from HxCDD (8.6-9.6 pg) to HpCDD (41 pg) with a 266 nm laser. A higher-energy photon, e.g., 213 nm, can subsequently ionize molecules from the triplet levels.24 Thus, the LOD measured for HpCDD (13 pg) did not change significantly at 213 nm from HxCDD (7.9-8.6 pg). For OCDF, the absorption cross section at 213 nm was smaller than that at 266 nm.27 However, the ionization efficiency was apparently larger at 213 nm than that at 266 nm for OCDD/OCDF. This could be explained by a more dominant process of ionization from the triplet levels, even when a picosecond laser (4 ps) was utilized for efficient ionization from the singlet-excited state. This result suggests shorter lifetimes (
