Gel Electrophoresis Coupled to Inductively Coupled Plasma−Mass

Anal. Chem. 2007, 79, 1714-1719

Gel Electrophoresis Coupled to Inductively Coupled Plasma-Mass Spectrometry Using Species-Specific Isotope Dilution for Iodide and Iodate Determination in Aerosols Wolfram Bru 1 chert, Andreas Helfrich, Nico Zinn, Thomas Klimach, Markus Breckheimer, Hongwei Chen, Senchao Lai, Thorsten Hoffmann, and Jo 1 rg Bettmer*

University of Mainz, Institute of Inorganic Chemistry and Analytical Chemistry, Duesbergweg 10-14, D-55099 Mainz, Germany

In this paper, we present an online coupling of gel electrophoresis (GE) and inductively coupled plasmamass spectrometry (ICP-MS) for the determination of iodine species (iodide and iodate) in liquid (seawater) and aerosol samples. For the first time, this approach is applied to the analysis of small molecules, and initial systematic investigations revealed that the migration behavior as well as the detection sensitivity strongly depends on the matrix (e.g., high concentrations of chloride). These effects could consequently affect the accuracy of analytical results, so that they need to be considered for the analysis of real samples. The technique used for quantification is species-specific isotope dilution analysis (ssIDA), which is a matrix-independent calibration method under certain conditions. We demonstrate that the use of 129I-enriched iodide and iodate allows the correction of the impact of the matrix on both, the electrophoretic migration and the detection sensitivity of the ICP-MS. After optimization, this coupling offers a novel and alternative method in the analysis of iodine compounds in various matrices. Here, we demonstrate the analytical capability of the technique for the chemical characterization of marine aerosols. The results show the presence of iodide and iodate at the ng m-3 and sub-ng m-3 level in the investigated aerosol samples, which were taken at the coastal research station in Mace Head, Ireland. These results are in good agreement with other recent studies, which demonstrated that the iodine chemistry in the marine atmosphere is only poorly understood. In addition to iodide and iodate, another iodine compound could be separated and detected in certain samples with high total iodine concentrations and was identified as elemental iodine, probably in form of triiodide, by peak matching. However, it may arise from an artifact during sample preparation. Especially within the last few years, the role of iodine in the marine atmosphere has received considerable attention. Besides * To whom correspondence should be addressed. E-mail: uni-mainz.de.

bettmer@

1714 Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

the influence of iodine as a catalyst for ozone destruction in the marine boundary layer,1-3 the importance of iodine in the formation of new atmospheric particles is responsible for the increased scientific interest, finally motivated by the role of aerosol particles in the global radiation balance. Uncertainties in the natural background source of atmospheric aerosol particles complicate estimates of the influence of anthropogenic aerosols. Therefore, one goal of current research activities is the identification and quantification of natural particle formation processes in the troposphere. Recent field and laboratory studies also point to coastal sites as an important biogenic source region for new particle formation.4,5 The current understanding is that volatile iodine precursors, such as diiodomethane or even molecular iodine, are released by marine algae into the atmosphere, where they are rapidly photolyzed during daytime and form low volatile iodine oxides which finally self-nucleate.6,7 However, still the characterization and identification of the precursors and products is analytically challenging and the detailed particle formation mechanisms remain unclear. Inductively coupled plasma-mass spectrometry (ICP-MS) has gained great importance in detecting iodine in environmental and food samples.8,9 Besides its excellent detection capabilities, ICP-MS offers isotope analysis, which allows the application of isotope dilution analysis; in the case of monoisotopic iodine (127I), the use of the long-lived 129I is appropriate in many purposes.10-12 (1) Solomon, S.; Garcia, R. R.; Ravishankara, A. R. J. Geophys. Res. 1994, 99, 20491-20499. (2) Davis, D.; Crawford, J.; Liu, S.; McKeen, S.; Bandy, A.; Thornton, D.; Rowland, F.; Blake, D. J. Geophys. Res. 1996, 101, 2135-2147. (3) Carpenter, L. J. Chem. Rev. 2003, 103, 4953-4962. (4) Hoffmann, T.; Seinfeld, J. H.; O’Dowd, C. D. Geophys. Res. Lett. 2001, 28, 1949-1951. (5) O’Dowd, C. D.; Jimenez, J. L.; Bahreini, R.; Flagan, R. C.; Seinfeld, J. H.; Ha¨meri, K.; Pirjola, L.; Kulmala, M.; Jennings, S. G.; Hoffmann, T. Nature 2002, 417, 632-635. (6) O’Dowd, C. D.; Hoffmann, T. Environ. Chem. 2005, 2, 245-255. (7) O’Dowd, C. D.; Jimenez, J. L.; Bahreini, R.; Flagan, R. C.; Seinfeld, J. H.; Ha¨meri, K.; Pirjola, L.; Kulmala, M.; Jennings, S. G.; Hoffmann, T. Nature 2005, 433, E13-E14. (8) Ge´linas, Y.; Krushevska, A.; Barnes, R. M. Anal. Chem. 1998, 70, 10211025. (9) Fecher, P. A.; Goldmann, I.; Nagengast, A. J. Anal. At. Spectrom. 1998, 13, 977-982. (10) Ra¨dlinger, G.; Heumann, K. G. Anal. Chem. 1998, 70, 2221-2224. 10.1021/ac061767y CCC: $37.00

© 2007 American Chemical Society Published on Web 01/09/2007

For iodine speciation, the separation of iodide and iodate is mainly achieved by ion chromatography (IC)12 or capillary electrophoresis.13,14 Particularly in combination with species-specific isotope dilution ICP-MS, it has become a precise and accurate analysis method in certain applications like water analysis.12 However, a suitable analytical technique able to separate and measure more than two iodine species in parallel (e.g., I-, IO3-, I2) is still lacking.15 In this paper, we present further improvements of the recently developed online coupling of gel electrophoresis (GE) and ICPMS for the separation and detection of ionic low-molecular weight iodine species.16,17 The experimental setup is described in detail and is applied in combination with species-specific isotope dilution analysis to the analysis of iodide and iodate in aerosol samples. It can be demonstrated that this coupling offers an attractive alternative to IC-ICP-MS with the advantage of detecting elemental iodine, probably in the form of triiodide, a species which could be detected in several aerosol samples with elevated levels of total iodine. EXPERIMENTAL Reagents. All solutions were prepared using ultrapure water (Milli-Q Water Purification System, Millipore, Bedford, MA). Electrode buffer solutions were daily prepared containing 0.05 mol L-1 boric acid (Suprapur, Merck, Darmstadt, Germany) and were adjusted to the required pH by adding sodium hydroxide solution (30%, Suprapur, Merck). In addition, the elution buffer contained an internal standard of 10 µg L-1 Te (Tellurium AA-Standard, 1000 µg mL-1, Alfa, Karlsruhe, Germany). Iodine-standard solutions were prepared by dissolving the appropriate amount of potassium iodide (Suprapur) and potassium iodate (p.a.) in water. Long-lived 129I- and 129IO - standards were enriched in 129I of about 86%18 3 and could be used for species-specific isotope dilution (NEN Chemicals, Boston, MA). To prevent an isotope exchange between I- and IO3-, the solutions were kept at pH > 6 during storage and analysis. For investigations on varying matrix composition, sodium chloride (extrapure, Merck) was applied. (Safety note: As 129I is a long-lived radioactive isotope, solutions containing enriched iodine were used in concentrations lower than 100 µg L-1, which refers to an activity lower than 0.66 Bq g-1). The gels were prepared using SeaKam LE and MetaPhor agarose (both Cambrex, Rockland, MA) in electrode buffer as described elsewhere.16 Both argon for the ICP and helium, to degas the eluent solutions, had purity of 99.996 vol-% (both Westfalen AG, Mu¨nster, Germany). Instrumentation. ICP-MS System. The Element 2 inductively coupled plasma-sector field-mass spectrometer (ICP-SF-MS) (Thermo Finnigan MAT GmbH, Bremen, Germany) was equipped (11) Ra¨dlinger, G.; Heumann, K. G. Environ. Sci. Technol. 2000, 34, 39323936. (12) Heumann, K. G.; Gallus, S. M.; Ra¨dlinger, G.; Vogl, J. Spectrochim. Acta 1998, 53B, 273-287. (13) Michalke, B.; Schramel, P. Electrophoresis 1999, 20, 2547-2553. (14) Timerbaev, A. R.; Hirokawa, T. Electrophoresis 2006, 27, 323-340. (15) Sta¨rk, H.-J.; Mattusch, J.; Wennrich, R.; Mroczek, A. Fresenius’ J. Anal. Chem. 1997, 359, 371-374. (16) Bru ¨ chert, W.; Bettmer, J. Anal. Chem. 2005, 77, 5072-5075. (17) Helfrich, A.; Bru ¨ chert, W.; Bettmer, J. J. Anal. At. Spectrom. 2006, 21, 431-434. (18) Boulyga, S. F., Heumann, K. G. Int. J. Mass Spectrom. 2005, 242, 291296.

Table 1. Instrumental Parameters for the ICP-MS System ICP system instrument RF power auxiliary gas flow coolant gas flow nebulizer gas flow sampler cone skimmer cone dwell time peaks per sample sample time mass resolving power isotopes monitored

element 2 1300 W 1.0 L min-1 16 L min-1 1.0 L min-1 Pt, 1.0-mm orifice Pt, 0.7-mm orifice 10 ms 20 300 ms 300 (low) 126Te, 127I, 129I, 131Xe

with a µ-flow nebulizer and a Scott type spray chamber (both AHF Feuerbacher, Tu¨bingen, Germany). The ICP-MS parameters were daily optimized for optimal detection by continuous injection of 10 µg L-1 IO3- and Te in 50 mmol L-1 borate buffer (pH ) 8.0). After cleaning the sample introduction system, the capillary of the µ-flow nebulizer was directly connected to the GE system (Mini Prep Cell with high-voltage supply PowerPac 3000, Bio-Rad Laboratories, Munich, Germany). Details of the operating conditions used throughout this work are given in Table 1. Data evaluation was performed by Microcal Origin 6.0 (Microcal Software, Inc., Northampton, MA) through manual peak integration corrected for the background signal. As natural iodine is monoisotopic and within this work no 129I as possibly resulting from nuclear contamination could be detected, mass bias effects do not need to be considered, because the isotope ratio 127I/129I for both the spike standard and the sample after isotope equilibration was determined identically. Nevertheless, mass bias was determined to be about 1.5% for 129Xe/131Xe, which originates as contaminant from Ar gas. For 127I and 129I measurements only, 129Xe has to be considered as spectral interference, since it cannot be separated within the mass spectrometer, even in the highresolution mode. Within this work, the background for m/z ) 129 was constant and a possible time-dependent contribution of Xe was further controlled by monitoring 131Xe. This ensured that obtained signals at m/z ) 129 during electrophoretic separations appeared uniquely from spiked 129I. GE System. The online coupling of GE and ICP-MS has been previously described,16 but the setup was further developed to improve the analyte transfer from the gel to the elution buffer with special attention to the separation of low-molecular weight compounds (Figure 1). This required an alteration within the elution region as shown in detail in Figure 1b. In brief, the separated compounds eluting from the gel are directly released into the elution buffer within the elution frit. This buffer is continuously pumped (100 µL min-1) through an outlet (1.5-mm diameter) in the dialysis membrane into a tube, which is directly connected to the membrane. The membrane has a molecular weight cutoff of 3.5 kDa and ensures the electrical connection to the lower electrode buffer. The dialysis membrane is fixed by a support frit for practical reasons.19 Prior to analysis, all analyzed samples (700 µL) are diluted with 200 µL glycerol (85% (w/w) aqueous solution, Acros, Geel, (19) Biorad. Mini Prep Cell Instruction Manual. U.S. Patent Number 4,877,510, October 31, 1989.

Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

1715

Table 2. Optimized Operating Parameters of the GE System parameter voltage electrode buffers eluent sample volume gel length gel i.d. gel material

250-300 V 0.05 mol L-1 boric acid, pH ) 8.0 0.05 mol L-1 boric acid, pH ) 8.0, 10 µg L-1 Te variable 2-20 µL 5 cm 2.2 mm agarose

Figure 1. (a) Experimental setup of the GE-ICP-MS coupling. (b) Gel elution region.

Belgium) and 100 µL 0.5 mol L-1 borate buffer. These samples (2-20 µL) are manually injected with variable pipettes (Eppendorf AG, Hamburg, Germany). Aeorsol Sampling and Sample Preparation. Aerosol samples were collected at the Mace Head Atmospheric Research Station (MHARS) located at the west coast of Ireland in September 2003. A five-stage cascade sampler (Berner impactor, Dr. Eberhard Steinweg, Grebenhain, Germany) with fractionated sizes of 0.0850.25, 0.25-0.71, 0.25-2.0, 2.0-5.9, and 5.9-10 µm was used for particle sampling. Cellulose nitrate filters (φo ) 78 mm, φi ) 40 mm, cuted from an original φ ) 85 mm filter, Sartorius AG, Go¨ttingen, Germany) were used in the Berner impactor. Ultrasonicassisted water extraction was carried out for the extraction of the iodine species from the filters. The whole filter was placed in a polypropylene (PP) vial with 10 mL MilliQ water for the extraction. The extraction time was 30-40 min. Before the ICP-MS measurements, the extraction solution was filtered with membrane filters (pore size 0.45 µm, Ø ) 26 mm, Minisart, Sartorius AG) to prevent clogging of the ICP-MS inlet. RESULTS AND DISCUSSION System Optimization. Although GE is one of the standard separation methods for macromolecules, little is known about its application for the analysis of low-molecular weight ions.20 To evaluate the correlation of electrophoretic behavior and influencing parameters, such as gel concentration, voltage, and so forth, systematic investigations were conducted for the optimization of the separation conditions as summarized in Table 2. Under these conditions, the precision of migration times was lower than 2% (N ) 5) for both species. Shorter gel lengths could reduce the total analysis time without reduction of electrophoretic resolution. (20) Andrews, A. T. Electrophoresis. Theory, Techniques, and Biochemical and Clinical Applications, 2nd ed.; Oxford University Press: New York, 1988.

1716 Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

Figure 2. 127I+ signal intensities in dependence on the Clconcentration.

However, we observed a strong matrix impact on the electrophoretic behavior of iodide resulting in an unsatisfying separation, which made the use of longer gels necessary. The effect of chloride concentration was studied to assess the influence of the highly saline matrix like seawater. Figure 2 represents the impact of chloride concentration on the 127I+ detection sensitivity for iodide and iodate. At concentrations higher than 1 mg L-1, we observed a significant decrease of the 127I+ signal for both species, which resulted in a signal suppression of about 20% at 10 g L-1 Cl- compared to standard solutions without matrix addition. As chloride at these concentration levels eluted between 5.5 and 9 min as a broad peak, it strongly suppressed the detection sensitivity of the coeluting iodide.21 However, the later eluting iodate was affected to the same extent (Figure 2). Besides the ICP response of iodide, its electrophoretic migration behavior was influenced at high chloride levels as well. Figure 3a compares typical electropherograms recorded at two different chloride concentrations. As chloride levels of 100 mg L-1 did not affect the migration at all, remarkable peak fronting and migration shift for iodide could be observed at an enhanced chloride concentration of 1000 mg L-1,14 whereas iodate migration was not influenced. Although the species were still baseline separated, these impacts on the migration behavior and on the detection sensitivity would have great influence on the accurate and precise determination of the iodine species in complex matrices. To consider these effects, species-specific isotope dilution analysis (ssIDA) was applied for quantification purposes. ssIDA is a reliable (21) Fraser, M. M.; Beauchemin, D. Spectrochim. Acta 2000, 55B, 1705-1731.

Table 3. (a) Analytical Results for Iodide and Iodate (as Iodine) in Liquid Samples (n ) 3). (b) Analytical Results for the Iodide and Iodate Recovery from Seawater Samples (a) Analytical Results for Iodide and Iodate in Liquid Samples I- in µg L-1 IO3- in µg L-1 total iodine in sample µg L-1 (nmol L-1) (nmol L-1) (nmol L-1) SRM16 43d seawatera

11.8 ( 0.4 (93 ( 3) 21.3 ( 0.5 (168 ( 4)

1.07 ( 0.06 (8.4 ( 0.5) 22.3 ( 0.6 (176 ( 5)

12.7 ( 0.9 (100 ( 7) 43.1 ( 2.8 (340 ( 22)

(b) Analytical Results for the Iodide and Iodate Recovery from Seawater Samples added I-/IO3found Irec.b rec. found IO3[%] [µg L-1] [µg L-1] [%] [µg L-1] 0 20 40 100 200

21.9 ( 0.5 42.2 ( 0.9 62.5 ( 1.5 123.4 ( 1.8 224.9 ( 3.0

101 101 101 101

22.6 ( 0.5 43.9 ( 0.8 61.3 ( 1.8 125.1 ( 2.0 220.4 ( 3.2

103 98 102 99

a Surface seawater sample from the Ile D’Oleron, France (1.2°W and 45.8°N) taken September 30th, 2004. b Rec.: recovery.

Figure 3. (a) Electropherograms of I- and IO3- (each 10 µg L-1 as I) in different matrices. (b) Electropherogram of 129I enriched I- and IO3- (each 100 µg L-1 as I) in 1000 mg L-1 Cl-.

quantification method in various applications,22,23 and 129I enriched iodide and iodate were used within this work. Once the isotope equilibration has been reached, not only possible analyte losses can be compensated, but as demonstrated in Figure 3b, also the influence of the matrix on the migration is captured by ssIDA. As a consequence, although a significant decrease in iodine sensitivity (Figure 2) could be observed with increasing matrix concentration, the 127I/129I ratio remained constant (0.1502 ( 0.0004 for the used spike sample) and, thus, was observed to be absolutely independent from the chosen matrix composition. These observations form the basis that the developed method is applicable to iodine speciation in environmental samples. Analysis of Liquid Samples for Iodine Species. To our best knowledge, no certified reference material for iodide and iodate in aqueous samples is available, so we decided to establish the developed method by independent total iodine analysis using ICP-MS standard addition method. For GE-ICP-MS analysis, seawater was diluted 1:10 for further analysis, whereas SRM1643d (trace elements in water) could be directly analyzed. The results obtained are summarized in Table 3a and were in good agreement, (22) Rodrı´guez-Gonza´lez, P.; Marchante-Gayo´n, J. M.; Alonso, J. I. G.; SanzMedel, A. Spectrochim. Acta 2005, 60B, 151-207. (23) Heumann, K. G. Anal. Bioanal. Chem. 2004, 378, 318-329.

from which can be concluded that iodide and iodate were the main iodine species in these solutions. In all samples, iodide and iodate could be detected at the low µg L-1 level, and even in high extent of iodide the determination of iodate is unaffected because of the separation conditions used. Especially, the NIST standard reference material SRM1643d might serve as reference material for these species in the future. So far, we have observed no significant changes over a period of 3 months at storage at 4 °C, but, however, further detailed investigations have to be performed to elucidate its applicability as a reference material for iodine species. Besides the comparison to total iodine analysis, a seawater sample (Table 3a) was chosen for matrix matching. The recoveries for both species were very satisfactory as demonstrated in Table 3b. These findings show that even liquid samples with high salinity can be analyzed by the presented method with good accuracy. Analysis of Aerosols. To prove the method’s analytical performance for aerosol samples, equivalent experiments were conducted as presented for seawater (Table 3b). Figure 4 illustrates the obtained recovery functions for iodide and iodate. For both species, the recoveries were (100 ( 2)%. Furthermore, replicate injections (n ) 7) showed a precision below 2% at the 1 ng m-3 level, so that both precision and accuracy were very satisfactory. Detection limits for both species (expressed as iodine) were determined to be 0.08 µg L-1 (0.63 nmol L-1) (3 s-criterion) in the injected aerosol samples (injection volume: 5 µL). Comparative data from the literature range from 0.4 nmol L-1 (CSV)24 to 2 nmol L-1 iodide and 57 nmol L-1 iodate (CE-UV/vis).25 On the basis of 10 m3 sample volume, the detection limits obtained in this work were 0.1 ng m-3. Table 4 shows results of the first application of the method for iodine speciation in atmospheric aerosol samples. The samples (24) Baker, A. R.; Thompson, D.; Campos, M.; Parry, S. J.; Jickells, T. D. Atmos. Environ. 2000, 34, 4331-4336. (25) Huang, Z.; Ito, K.; Timerbaev, A. R.; Hirokawa, T. Anal. Bioanal. Chem. 2004, 378, 1836-1841.

Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

1717

Figure 4. Recovery of iodide and iodate in aerosol samples. Table 4. Analytical Results for Iodide and Iodate (as Iodine) in Aerosol Samples (n ) 3) sample

sample description

I- [ng‚m-3]

IO3- [ng‚m-3]

ratio (IO3-/I-)

BI19-1 BI19-2 BI19-3 BI19-4 BI19-5

0.085-0.25 µm 0.25-0.71 µm 0.71-2.0 µm 2.0-5.9 µm 5.9-10.0 µm

0.748 ( 0.044 0.804 ( 0.082 0.777 ( 0.038 1.219 ( 0.041 1.228 ( 0.023

0.440 ( 0.014 0.177 ( 0.013 0.884 ( 0.045 0.762 ( 0.077 1.070 ( 0.032

0.59 0.22 1.14 0.63 0.87

were taken during a measurement campaign in Mace Head at the west coast of Ireland. Both iodine species, iodide and iodate, were observed in the different particle size regimes with concentrations ranging between 0.75 and 1.23 and between 0.18 and 1.08 ng m-3, respectively. The molar ratios of IO3-/I- were quite variable with a range between 0.22 and 1.14. Obviously, because of the very limited data set, the results presented in the table cannot be used to derive major conclusions about the iodine chemistry in the atmosphere and they are shown here primarily to demonstrate the applicability of the analytical method for iodine speciation even in size-segregated particle samples. Nevertheless, the observation of relatively high iodide concentrations confirms recent results of iodine speciation measurements in marine aerosols from Baker.26,27 Like these papers, the results shown in Table 4 indicate a lack of understanding about the iodine chemistry in the marine atmosphere. Current atmospheric chemistry models actually predict iodine in the particle phase to be present in the form of iodate as the major iodine species. Several pathways are included in these models which could explain the formation of iodate, such as the formation of higher iodine oxides in the gas phase followed by gas-to-particle conversion of the iodine oxides.4,28 In contrast, for iodide just one pathway is known, that is, the uptake of HI into the particle phase, however, this is believed to be a minor contribution.29,30 Obviously, the current (26) (27) (28) (29)

Baker, A. R. Geophys. Res. Lett. 2004, 31, L23S02. Baker, A. R. Environ. Chem. 2005, 2, 295-298. Saunders, R. W.; Plane, J. M.C. Environ. Chem. 2005, 2, 299-303. Vogt, R.; Sander, R.; von Glasow, R.; Crutzen, P. J. J. Atmos. Chem. 1999, 32, 375-395. (30) McFiggans, G.; Plane, J. M. C.; Allan, B. J.; Carpenter, L. J.; Coe, H.; O’Dowd, C. J. Geophys. Res. 2000, 105, 14371-14385.

1718 Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

Figure 5. Electropherogram of an aerosol sample.

understanding of the iodine chemistry does not fit to the results shown in Table 4, that is, relatively high iodide concentrations in comparison to the iodate concentrations. In principle, particulate iodine can also derive from primary aerosol sources, such as sea spray, which might then contribute to both inorganic iodine species. However, again observational evidence exist that are in contradiction to a major contribution from primary sources since the high iodine enrichment in atmospheric aerosols clearly indicates secondary processes as a major contribution of iodine to the particle phase.24,26,31 Very recently, an updated aqueous phase chemistry scheme was used to model possible mechanisms and to evaluate potential formation pathways for iodide.32 These authors speculate about organic iodine chemistry combined with inorganic reaction cycles to explain the high iodide concentrations. However, the chemistry of iodine in the atmosphere, especially in connection with aerosol formation processes, is only poorly understood and certainly more iodine speciation measurements are urgently required to get a better understanding of the role of iodine in the marine atmosphere. In Figure 5, a typical electropherogram of an aerosol sample is demonstrated. As pointed out in Table 4, the dominant iodine species were iodide and iodate, but at least one further iodine species could be observed in the electropherogram (migration time 15.5 min). Chemically generated iodine, probably in the form of triiodide, was injected into the system and matched the migration behavior of the unknown peak. As the presence of elemental iodine in atmospheric aerosols is very unlikely, we assume that this species occurred as an artifact, probably formed during the sampling or preparation procedure. In the literature, unknown species occurred as well in the analyses on the basis of anion-exchange chromatography.12 Although not identified, they were classified as “anionic organic iodine”. Consequently, further investigations are urgently needed for the identification of these unknown species. However, we believe that the developed method may also contribute to the determination of unknown iodine species, and thus, to the understanding of iodine chemistry in the atmosphere. (31) Ga¨bler, H. E.; Heumann, K. G. Int. J. Environ. Anal. Chem. 1993, 50, 129146. (32) Pechtl, S.; Schmitz, G.; von Glasow, R. Atmos. Chem. Phys. Discuss. 2006, 6, 10959-10989.

CONCLUSIONS The online coupling of GE and ICP-MS was successfully applied to the speciation of iodine species in aqueous samples with special attention to aerosol samples. After thorough optimization, this hyphenation could serve as a fast and reliable method in combination with species-specific isotope dilution applying 129I. As demonstrated on the example of aerosol sample analysis, this system allows a precise and accurate quantification in the low ng m-3 range. Furthermore, unknown species could be observed in certain aerosol samples. Its migration behavior could be matched with elemental iodine, but further information is needed for its identification. In general, it could be demonstrated that the use of gel electrophoresis has a great potential in the separation of lowmolecular weight ions. In combination with inductively coupled

plasma-mass spectrometry as element-selective detector, it can serve as novel hyphenated technique in elemental speciation and offers in general the application of species-specific and speciesunspecific isotope dilution analysis technique for quantification purposes. ACKNOWLEDGMENT The authors acknowledge the financial support of the German Research Council (DFG) within the graduate program “Trace Analysis of Elemental Species: Development of Methods and Applications”. Received for review September 19, 2006. Accepted November 27, 2006. AC061767Y

Analytical Chemistry, Vol. 79, No. 4, February 15, 2007

1719