Latex-Coated Polymeric Monolithic Ion-Exchange Stationary Phases

BrO3-, IO3-, and CrO42- (Ajax, Melbourne, Australia), and 1 M Cl- (May and Baker, ..... One possible use of the permeable ion-exchange monoliths d...
0 downloads 0 Views 298KB Size
Download PDF
Anal. Chem. 2005, 77, 407-416

Latex-Coated Polymeric Monolithic Ion-Exchange Stationary Phases. 1. Anion-Exchange Capillary Electrochromatography and In-Line Sample Preconcentration in Capillary Electrophoresis Joseph P. Hutchinson,† Philip Zakaria,† Andrew R. Bowie,†,‡ Miroslav Macka,† Nebojsa Avdalovic,§ and Paul R. Haddad*,†

Australian Centre for Research on Separation Science (ACROSS), School of Chemistry, Faculty of Science, Engineering and Technology, University of Tasmania, Private Bag 75, Hobart 7001, Australia, Antarctic Climate & Ecosystems Cooperative Research Centre (ACE CRC), University of Tasmania, Private Bag 80, Hobart 7001, Australia, and Dionex Corporation, 1228 Titan Way, Sunnyvale, California 94088-3603

A sulfonated methacrylate monolithic polymer has been synthesized inside fused-silica capillaries of diameters 50-533-µm i.d. and coated with 65-nm-diameter fully functionalized quaternary ammonium latex particles (AS18, Dionex Corp.) to form an anion-exchange stationary phase. This stationary phase was used for ion-exchange capillary electrochromatography of inorganic anions in a 75-µm-i.d. capillary with Tris/perchlorate electrolyte and direct UV detection at 195 nm. Seven inorganic anions (bromide, nitrate, iodide, iodate, bromate, thiocyanate, chromate) could be separated over a period of 90 s, and the elution order indicated that both ion exchange and electrophoresis contributed to the separation mechanism. Separation efficiencies of up to 1.66 × 105 plates m-1 were achieved, and the monoliths were stable under pressures of up to 62 MPa. Another latex-coated monolith in a 250-µm-i.d. capillary was used for in-line preconcentration by coupling it to a separation capillary in which the EOF had been reversed using a coating of either a cationic polymer or cationic latex particles. Several capillary volumes of sample were loaded onto the preconcentration monolith, and the analytes (inorganic anions) were then eluted from the monolith with a transient isotachophoretic gradient before being separated by electrophoresis in the separation capillary. Linear calibration curves were obtained for aqueous mixtures of bromide, nitrite, nitrate, and iodide. Recoveries of all analytes except iodide were reduced significantly when the sample matrix contained high levels of chloride. The preconcentration method was applied to the determination of iodide in open ocean water and provided a limit of detection of 75 pM (9.5 ng/L) calculated at a signal-to-noise ratio of 3. The relative standard deviation for migration time and peak area for iodide were 1.1 and 2.7%, respectively (n * Corresponding author. Phone: +61-3-6226 2179. Fax: +61-3-6226 2858. E-mail: [email protected]. † ACROSS, University of Tasmania. ‡ ACE CRC, University of Tasmania. § Dionex Corp. 10.1021/ac048748d CCC: $30.25 Published on Web 12/07/2004

© 2005 American Chemical Society

) 6). Iodide was eluted as an efficient peak, yielding a separation efficiency of 5.13 × 107 plates m-1. This focusing was reproducible for repeated analyses of seawater. Capillary electrochromatography (CEC) utilizes a chromatographic bed within a capillary and adds an extra chromatographic dimension to the electrophoretic separations. CEC possesses several advantages over other liquid chromatographic methods, such as low sample and solvent consumption and the increased efficiency and resolution that can be achieved due to the flat flow profile. Pretorius et al.1 are credited for realizing the benefits of the flat flow profile generated by the electroosmotic flow (EOF). However, electrochromatography was not demonstrated in particlepacked capillary columns until the 1980s.2 Since then CEC has been performed with various formats, including open-tubular columns,3,4 pseudostationary phases,5 and monolithic columns.6,7 Recent interest in monolithic stationary phases (also known as continuous-bed stationary phases) has arisen as a result of their advantages over packed-bed stationary phases. Improvements include ease and low cost of preparation, elimination of frits and their associated problems, the greater permeability afforded by the network of interconnected pores through the relatively thin skeletal structure, and the ability to tailor the physical structure and surface chemistry to obtain the desired chromatographic properties. Monolithic stationary phases can be divided into two main groups: silica (inorganic) monoliths and rigid polymer (organic) monoliths. Silica monoliths are generally prepared using a sol-gel process, and the first uniform porous silica rods were reported by Tanaka et al. in 1996.8 The sol-gel method was (1) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr. 1974, 99, 2330. (2) Jorgenson, J. W.; Lucacs, K. D. J. Chromatogr. 1981, 218, 209-216. (3) Pesek, J.; Matyska, M. J. Chromatogr. Libr. 2001, 62, 241-270. (4) Malik, A. Electrophoresis 2002, 23, 3973-3992. (5) Fritz, J. S.; Breadmore, M. C.; Hilder, E. F.; Haddad, P. R. J. Chromatogr., A 2002, 942, 11-32. (6) Svec, F.; Peters, E. C.; Sykora, D.; Frechet, J. M. J. J. Chromatogr., A 2000, 887, 3-29. (7) Allen, D.; Rassi, Z. E. Electrophoresis 2003, 24, 3962-3976.

Analytical Chemistry, Vol. 77, No. 2, January 15, 2005 407

adapted to create silica monoliths within fused-silica capillaries for CEC, which were then derivatized for reversed-phase applications.9 Silica-based monoliths for CEC have recently been reviewed in detail by Allen and El Rassi.7 Polymer monoliths have been created using several types of materials. These include acrylates, methacrylates, polystyrene/ divinylbenzene, and polyacrylamides and have been discussed in recent reviews.19,20 Methacrylate monoliths were introduced for ion-exchange chromatography21 in the early 1990s and were fabricated by the copolymerization of butyl methacrylate and the cross-linking reagent ethylene dimethacrylate. A small amount of 2-acrylamido-2-methyl-1-propanesulfonic acid (AMPS) can be added to provide the surface charge necessary to support electroosmotic flow. Methacrylate-based polymers are easily produced and are stable under varied pH conditions (over the pH range 2-12). The fabrication of monolithic capillary columns is performed by placing the polymerization mixture in a pretreated capillary,22 together with a radical initiator, which can be activated by heat,23 UV light,24 or radiation.25 The unreacted porogen is washed from the column before use, and the fabrication method takes approximately 20 h by thermal initiation23 or 10 min by UV initiation.22 Ion-exchange chromatography has been performed on both silica16-18,26-29 and organic polymer stationary phases.16,29-34 The (8) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. Anal. Chem. 1996, 68, 3498-3501. (9) Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Soga, N.; Nagayama, H.; Hosoya, K.; Tanaka, N. Anal. Chem. 2000, 72, 1275-1280. (10) Dulay, M. T.; Kulkarni, R. P.; Zare, R. N. Anal. Chem. 1998, 70, 51035107. (11) Ratnayake, C. K.; Oh, C. S.; Henry, M. P. J. High Resolut. Chromatogr. 2000, 23, 81-88. (12) Dulay, M. T.; Quirino, J. P.; Bennett, B. D.; Kato, M.; Zare, R. N. Anal. Chem. 2001, 73, 3921-3926. (13) Hayes, J. D.; Malik, A. Anal. Chem. 2000, 72, 4090-4099. (14) Kang, J.; Wistuba, D.; Schurig, V. Electrophoresis 2002, 23, 1116-1120. (15) Kato, M.; Sakai-Kato, K.; Matsumoto, N.; Toyo’oka, T. Anal. Chem. 2002, 74, 1915-1921. (16) Breadmore, M. C.; Shrinivasan, S.; Wolfe, K. A.; Power, M. E.; Ferrance, J. P.; Hosticka, B.; Norris, P. M.; Landers, J. P. Electrophoresis 2002, 23, 34873495. (17) Hutchinson, J. P.; Hilder, E.; Macka, M.; Avdalovic, N.; Haddad, P. R., submitted to Electrophoresis. (18) Hatsis, P.; Lucy, C. A. Anal. Chem. 2003, 75, 995-1001. (19) Legido-Quigley, C.; Marlin, N. D.; Melin, V.; Manz, A.; Smith, N. W. Electrophoresis 2003, 24, 917-944. (20) Hilder, E.; Svec, F.; Frechet, J. M. J. Electrophoresis 2002, 23, 3934-3953. (21) Svec, F.; Frechet, J. M. J. J. Chromatogr., A 1995, 702, 89-95. (22) Rohr, T.; Hilder, E. F.; Donovan, J. J.; Svec, F.; Frechet, J. M. J. Macromolecules 2003, 36, 1677-1684. (23) Peters, E. C.; Petro, M.; Svec, F.; Frechet, J. M. J. Anal. Chem. 1997, 69, 3646-3649. (24) Yu, C.; Svec, F.; Frechet, J. M. J. Electrophoresis 2000, 21, 120-127. (25) Grasselli, M.; Smolko, E.; Hargittai, P.; Safrany, A. Nucl. Instrum. Methods B 2001, 185, 254-261. (26) Sugrue, E.; Nesterenko, P.; Paull, B. Analyst 2003, 128, 417-420. (27) Hatsis, P.; Lucy, C. A. Analyst 2002, 127, 451-454. (28) Xu, Q.; Tanaka, K.; Mori, M.; Helaleh, M. I. H.; Hu, W.; Hasebe, K.; Toada, H. J. Chromatogr., A 2003, 997, 183-190. (29) Breadmore, M. C.; Shrinivasan, S.; Karlinsey, J.; Ferrance, J. P.; Norris, P. M.; Landers, J. P. Electrophoresis 2003, 24, 1261-1270. (30) Ping, G.; Zhang, W.; Zhang, L.; Schmitt-Kopplin, P.; Zhang, Y.; Kettrup, A. Chromatographia 2003, 57, 629-633. (31) Wu, R.; Zou, H.; Ye, M.; Lei, Z.; Ni, J. Electrophoresis 2001, 22, 544-551. (32) Wu, R. a.; Zou, H.; Fu, H.; Jin, W.; Ye, M. Electrophoresis 2002, 23, 12391245. (33) Lammerhofer, M.; Svec, F.; Frechet, J. M. J.; Lindner, W. J. Chromatogr., A 2001, 925, 265-277. (34) Hindocha, D.; Smith, N. W. Chromatographia 2002, 55, 203-209.

408

Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

ability to control the pore size and surface chemistry of the polymeric monolithic structure35 makes polymer-based monolithic materials ideal for many applications20 including solid-phase extraction36,37 and ion-exchange CEC.16,29-34 Wu et al.31 dynamically coated methacrylate monoliths with surfactants, such as cetyltrimethylammonium bromide and sodium dodecyl sulfate for the separation of basic, acidic, and neutral aromatic compounds by CEC. Further work was performed to create a mixed mode (RP/ cation-exchange) methacrylate monolithic column for the separation of peptides.32 Lammerhofer et al.33 used 2-dimethylaminoethyl methacrylate as a monomer to form a weak anion-exchange monolith, which was then alkylated at the terminal amino functionalities to create a strong anion-exchange monolithic column for the separation of benzoic acids. Hindocha and Smith34 prepared monoliths with N,N-dimethylaminoethylacrylamide in the monomer mixture, which reversed the EOF and gave a minimally tailed peak for the basic pharmaceutical compound nortriptyline. Ping et al.30 separated nucleotides using ion-exchange and hydrophobic interactions in a methacrylate monolith that had sulfonic acid moieties from the presence of AMPS in the monomer mixture. Josic et al.38 have reviewed the use of ion-exchange monoliths for the separation of proteins and polynucleotides. Agglomeration of resin beads with latex particles is a wellestablished method for introducing ion-exchange functionalities in ion chromatography. The selectivity of the ion-exchange phase can be readily modified by varying the substrate and latex functionalities. The above approach has been used to manufacture ion-exchange phases for ion chromatography. Latex phases have also been used as a coating to the capillary wall in CEC.39-42 In another approach, a small section of the capillary was coated with latex to produce a preconcentration zone.43-46 This approach was used for concentrating analytes before elution with a transient isotachophoretic gradient and subsequent analysis by CE. Ion-exchange CEC using silica monoliths coated with functionalized latex particles has been produced by passing a suspension of latex nanoparticles through a silica monolith inside a capillary. The nanoparticles are bound electrostatically to the silanol groups at the silica surface, creating a strong anion-exchange column for CEC with reversed EOF. In this paper, we investigated the procedures necessary to prepare a sulfonated substrate monolith and agglomerate this (35) Peters, E. C.; Petro, M.; Svec, F.; Frechet, J. M. J. Anal. Chem. 1998, 70, 2288-2295. (36) Yu, C.; Davey, M. H.; Svec, F.; Frechet, J. M. J. Anal. Chem. 2001, 73, 5088-5096. (37) Baryla, N. E.; Toltyl, N. P. Analyst 2003, 128, 1009-1012. (38) Josic, D.; Buchacher, A.; Jungbauer, A. J. Chromatogr., B 2001, 752, 191205. (39) Boyce, M.; Breadmore, M.; Macka, M.; Doble, P.; Haddad, P. Electrophoresis 2000, 21, 3073-3080. (40) Breadmore, M.; Boyce, M.; Macka, M.; Avdalovic, N.; Haddad, P. J. Chromatogr., A 2000, 892, 303-313. (41) Breadmore, M.; Hilder, E.; Macka, M.; Avdalovic, N.; Haddad, P. Electrophoresis 2001, 22, 503-510. (42) Breadmore, M.; Macka, M.; Avdalovic, N.; Haddad, P. Analyst 2000, 125, 1235-1241. (43) Breadmore, M.; Boyce, M.; Macka, M.; Avdalovic, N.; Haddad, P. Analyst 2000, 125, 799-802. (44) Breadmore, M.; Macka, M.; Avdalovic, N.; Haddad, P. Anal. Chem. 2001, 73, 820-828. (45) Breadmore, M.; Palmer, A.; Curran, M.; Macka, M.; Avdalovic, N.; Haddad, P. Anal. Chem. 2002, 74, 2112-2118. (46) Hutchinson, J. P.; Macka, M.; Avdalovic, N.; Haddad, P. R. J. Chromatogr., A 2004, 1039, 187-192.

monolith with anion-exchange latex particles. We show utility of the monolithic latex phase for pursuing CEC separations and for CE based in-line preconcentration. [Note added in proof: since acceptance of this article the following closely related publication has appeared: Hilder, E. F.; Svec, F.; Frechet, J. M. J. J. Chromatogr., A 2004, 1053, 101-106.] EXPERIMENTAL SECTION Instrumentation. The capillary electrophoresis instrument used was an Agilent 3DCE (Waldbronn, Germany) and was equipped with an external pressure device to increase the applied pressure to 1.2 MPa. Separations were carried out using a 50533-µm-i.d. fused-silica capillary (Polymicro Technologies Inc, Phoenix, AZ). Typically 25 cm × 75 µm capillaries were used with in-house monoliths, unless stated otherwise. A 75 µm × 10 cm length of capillary, coated with poly(diallyldimethylammonium)chloride (PDDAC) or AS18 anion-exchange latex from Dionex Corp. (Sunnyvale, CA) was attached to the monolith using PEEK Microtight zero dead volume unions and sleeves (Upchurch Scientific, Oak Harbor, WA). The capillary had a detection window burned 8.5 cm from the anode. Anions were separated with migration of both the analytes and the EOF toward the anode. The voltage used for the separations was varied between -2 and -30 kV depending on the experimental requirements, and the temperature was maintained at 25 °C. Introduction of sample into the column was performed by hydrodynamic injection at the cathodic end. Reagents. AS18 latex (quaternary ammonium functionalized) nanoparticles were supplied as an 11% (w/v) suspension by Dionex Corp. The latex was dialyzed through a cellulose membrane for 48 h (with water as the receiving solution and the water being changed every 5 h) and then filtered through a 0.45-µm nylon syringe filter (Activon, Thornleigh, Australia). Standards of 10 mM Br- and I- (Sigma-Aldrich, Milwaukee, WI), NO3-, SCN-, and NO2- (BDH Chemicals, Kilsyth, Australia), BrO3-, IO3-, and CrO42- (Ajax, Melbourne, Australia), and 1 M Cl- (May and Baker, Footscray, Australia) were prepared from sodium or potassium salts of analytical reagent grade. All samples were prepared in water purified on a Millipore (Bedford, MA) Milli-Q water purification system. Perchloric acid was titrated with Tris(hydroxymethyl)aminomethane to pH 8.05. All standards and electrolytes used were filtered through 0.45-µm nylon filters and degassed under vacuum prior to use. Butyl methacrylate (99%) and ethylene dimethacrylate (98%) were purchased from Aldrich (Milwaukee, WI), filtered through basic alumina activity grade I, type WB-2 (Aldrich), and vacuumdistilled to remove the inhibitor, 4-methoxyphenol. Azobisisobutyronitrile was purchased from Du Pont (99%, North Sydney, NSW, Australia). (γ-Methacryloxypropyl)trimethoxysilane (γ-MAPS) was purchased from Sigma (98%, Milwaukee, WI) for derivatization of the fused-silica capillary. Procedure for Synthesis of Latex-Coated Polymer Monoliths. The polymer monolithic columns were prepared using the method outlined by Peters et al.35 Azobisisobutyronitrile (0.0095 g, ∼1% (w/w) with respect to the monomers) was dissolved in 0.3588 g of ethylene dimethacrylate, 0.4203 g of butyl methacrylate, and 1.8% (w/w of total mixture) 2-acrylamido-2-methyl-1propanesulfonic acid to provide cation-exchange sites within the monolithic structure. The ternary porogenic solvent used was

composed of 0.6679 g of 1-propanol, 0.4366 g of 1,4-butanediol, and 0.1200 g of water. The resultant mixture was sonicated for 10 min and purged with nitrogen for a further 10 min. A small portion of this solution was pumped into a fused-silica capillary that had previously been derivatized with γ-MAPS according to the method outlined by Rohr et al.22 The ends of the capillary were plugged with rubber GC septa. The remaining polymerization solution was sealed in a glass vial, and both the capillary and the vial containing the bulk polymerization solution were placed in a 60 °C water bath for 20 h. The resultant monolithic column was attached to another short section using zero dead volume capillary unions that had the polyimide coating removed for UV detection. Prior to coating with latex particles, the monolithic column was flushed with methanol for 5 h at 1.2 MPa and then flushed with water. The monolith that had formed in the glass vial was removed, cut into small pieces, Soxhlet extracted with methanol for 12 h, and vacuum-dried at 60 °C for a further 12 h. The pore volume and pore size distribution were determined using a Poresizer 9310 mercury intrusion porosimeter (Micromeritics, Norcross, GA). The polymer monolith columns were coated with Dionex AS18 anion-exchange latex particles by pumping a suspension of the latex through the capillary until a rise in the UV detector signal indicated that the latex had saturated the ion-exchange sites available on the capillary (∼10 capillary volumes was normally required). The column was subsequently flushed with water and then at least 10 capillary volumes of background electrolyte (BGE) prior to use. Separation efficiencies were calculated using the width of the band at half-height and given as the number of theoretical plates per column. Ion-exchange capacities for the columns were determined using an absorption/elution method where a sodium bromide solution was flushed through the monolithic column to saturate the ion-exchange sites. Excess bromide not bound to the latex was flushed from the column with Milli-Q water, and sodium perchlorate was flushed through the capillary using hydrodynamic pressure to elute the adsorbed bromide for detection. An external calibration curve was used for quantification. Preconcentration Procedures. Typically, a 15-cm, 250-µmi.d., AS18-coated monolith was attached via a PEEK zero dead volume union to a 80-cm, 75-µm-i.d. PDDAC, or an AS18 latexcoated OT detection segment with a detection window 8.5 cm from the anode was used. Initially, this column was flushed with a 1 M fluoride solution to convert the ion-exchange sites into the fluoride phase. The sample was then introduced using hydrodynamic pressure from the inlet. A water flush was performed for greater than one column volume to remove matrix ions not bound to the latex. A fluoride electrolyte solution was introduced into the column from the inlet, allowing electrophoresis to occur. Tris/perchlorate electrolyte was placed in the electrolyte vials and voltage applied to create a transient isotachophoretic gradient, which passed through the column, preconcentrating anions from the monolithic stationary phase at the gradient front. Sample Handling. The analysis of a Southern Ocean surface (1-2 m deep) seawater sample collected on RSV Aurora Australis research voyage AU0103 in the Australian sector of the Southern Ocean (51°25 ′S, 143°03 ′E) at 05:00 UTC on 11 December 2001 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

409

was performed. The surface temperature and salinity were 7.7 °C and 34.1 units, respectively. The seawater sample was filtered at sea through a 0.2-µm Polycap 150TC capsule filter (Whatman, Clifton, NJ) and stored unacidified in the dark prior to use. Further details on the sampling procedure can be found in Bowie et al.47 Scanning Electron Microscopy (SEM). The latex-coated monolithic columns used in this work were viewed by means of scanning electron microscopy. The capillary was cut to 2 cm in length and mounted on an aluminum stub. This was coated with a thin layer of platinum using an EIKO IB5 high-resolution platinum coater. The images were taken using a JEOL JSM890 high-resolution SEM operating at a voltage of 2 kV. RESULTS AND DISCUSSION Characterization of Latex-Coated Polymer Monoliths. Rigid polymer monoliths were formed in fused-silica capillaries of varying diameter (50-533-µm i.d.) using 1.8% (w/w) AMPS in the polymerization solution to create the surface charge required to bind the latex particles. It has been shown35 that using larger amounts of AMPS in the monomeric mixture can create monoliths that are less permeable due to hydrophilic solvent swelling. The fabricated monolithic capillary columns were stable under high pressure (62 MPa), and 50-µm-i.d. latex-coated monolithic columns were found to be suitable for CEC of inorganic anions. However, the permeability of these columns was low and new electrolytes had to be introduced to the column at 62 MPa. The 75-µm-i.d. monolithic columns were subsequently used as they represented the best compromise between permeability and Joule heating considerations. The flow rate of a 75-µm-diameter monolith of 25cm length was 5.2 µL/min at 41.4 MPa, and 1 column volume could be replaced in 4.5 min (corresponding to a linear flow velocity of 0.92 mm/s) under 1.2 MPa pressure, the maximum attainable with the Agilent 3DCE instrument. To show that the latex particles were bound to the surface of the polymeric monolith, the EOF of both an uncoated capillary and a latex-coated capillary were compared. The EOF of a 75-µmi.d. polymer monolith containing 1.8% AMPS was found to be 39 × 10-9 m2/V‚s prior to coating with latex particles and was reversed to -15 × 10-9 m2/V‚s after coating with the positively charged latex particles. The EOF of the coated capillary was remarkably stable from run to run when CEC separations were performed and was even unaltered after the monolith was allowed to dry for several months before reuse. Reproducibility results are given in Table 1. Larger variation occurred in the effective mobility of both bromide and chromate in comparison to the neutral marker over consecutive runs. This is related to the difficulty in reproducibly regenerating the ion-exchange sites on the monolith to an identical state prior to starting each run. This variation in RSD is more pronounced in the case of chromate, which interacts strongly with the stationary phase lowering the experimentally observed mobility well below that expected based on electrophoresis alone.44 SEM was performed to visualize the latex coating on the monolithic surface. Comparison of the SEM images obtained in this work to uncoated methacrylate monoliths in the literature22,24 suggested that the latex particles did not provide a complete (47) Bowie, A. R.; Achterberg, E. P.; Sedwick, P. N.; Ussher, S.; Worsfold, P. J. Environ. Sci. Technol. 2002, 36, 4600-4607.

410 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

Table 1. Mobility and RSD of Bromide (Low Ion-Exchange Selectivity Coefficient), Chromate (High Ion-Exchange Selectivity Coefficient), and a Neutral Marker (Thiourea) (n ) 6)a analyte

effective mobility (× 10-9 m2/V‚s) (% RSD)

bromide (weak anion) chromate (strong anion) thiourea (EOF marker) EOF marker after 1 day EOF marker on another column

-70.38 (1.5) -20.11 (4.3) -15.15 (0.4) -15.43 -14.22

a Conditions: column, 32.3 cm × 75 µm AS18-coated methacrylate monolith attached to 10 cm × 50 µm AS18 open-tubular column with window 8.5 cm from end; BGE, 10 mM Tris/ClO4-; injection, 0.3 MPa, 6 s of 0.1 mM Br-, I-, NO2-, NO3-, CrO42-, SCN-, BrO3-, and IO3in water (1 mM thiourea was added as an EOF marker); voltage, -30 kV; temperature 25 °C.

Figure 1. SEM of the latex-coated polymer monolith within a fusedsilica capillary that was used in this study.

coating of the surface of the monolith (Figure 1). This is perhaps to be expected due to the relatively small amount of AMPS in the polymerization solution, which provided the strong acid sites for latex binding; the charged sites occur not only at the monolithic surface but also within the bulk of the polymer. Despite the apparently incomplete coverage of the latex, subsequent results (see below) showed that the coated monolith exhibited appropriate anion-exchange properties, so no further attempts were undertaken to increase the latex coverage. Some of the important physical characteristics of the monolithic material were determined using mercury intrusion porosimetry. It must be emphasized that porosimetry was performed on the sample in a dry state due to the technical requirements of the technique used, so the results produced may differ from those for the monolithic column operated in an aqueous electrolyte. However, Peters et al. have shown that there is a strong correlation35 between the dry porosity measurements and the chromatographic performance of the column. Thus, the porosimetry technique was essential to the characterization of the monolithic column. Dry porosity measurements indicated that the median pore diameter of the monolithic material was 3.4 µm, the surface area of these pores was 1.26 m2/g, the bulk density was 0.3036 g/mL, the skeletal density was 0.7941 g/mL, and the porosity was 61.8%, correlating well with the amount of porogen used in the polymerization mixture.

Table 2. Ion-Exchange Capacities of Polymer Monoliths Created in Different Diameter Capillariesa ion-exchange capacity of column (pequiv/cm) 75-µm AS18 latex-coated OT column 75-µm AS18 latex-coated polymer monolithic column 250-µm AS18 latex-coated polymer monolithic column 533-µm AS18 latex-coated polymer monolithic column

0.922 17.1 386 1760

a Conditions: column, Various internal diameter 22-cm AS18-coated monolithic columns attached to a 10 cm × 75 µm PDDAC OT column with a detection window 8.5 cm from end were used (These were compared to a 32-cm AS18 open-tubular column with a detection window 8.5 cm from the end.). The ion-exchange capacities of the columns were measured using bromide for absorption/elution. Water was flushed through the column prior to elution with 100 mM perchlorate at 1.2 MPa. Temperature, 25 °C.

The ion-exchange capacity of the latex-coated polymeric monoliths is determined by a number of variables, including the amount of AMPS in the polymerization mixture, the total amount of monolith used, and the concentration of the latex particles in the coating solution. Also, the addition of competing cations to the latex coating solution can decrease the surface interaction of the latex with the sulfonic acid moieties at the monolithic surface. Table 2 shows the ion-exchange capacities of latex-coated monoliths formed in capillaries with a range of internal diameters exhibiting similar pore structures. Capillaries larger than 75-µm i.d. required very low ionic strength BGEs to avoid the high currents associated with Joule heating effects and are not practical for most CEC applications. The monolith formed in a 75-µm-i.d. capillary gave a suitable ion-exchange capacity and created a column with favorable mass-transfer characteristics for CEC. The ion-exchange capacities of the latex-coated monoliths (expressed as µequiv/g) were considerably less than their counterpart latexcoated particle ion chromatography stationary phases (approximately 16 and 1 µequiv/g for the particle and monolith materials, respectively). This again suggested that the coverage of the latex particles on the monolith was relatively sparse. The total volume of a capillary is given by the equation πR2L. If the capacity in Table 2 is plotted against R2 of the capillary used, a straight line is achieved. This indicates that the latex coverage was consistent across all capillary diameters used in this study even if a monolayer could not be not achieved. This indicates why increasing the capillary diameter is a simple and effective means of increasing the capacity of the stationary phase within. Ion-Exchange CEC of Inorganic Anions on Latex-Coated Monoliths. The latex-coated monolith was used as a stationary phase for CEC of inorganic anions, using perchlorate-based BGEs. Figure 2 shows the separation of seven anions, of varying electrophoretic mobilities and ion-exchange selectivity coefficients, using a 10 mM Tris/perchlorate BGE. Chromate was the only anion to show considerable tailing under these conditions, and the degree of tailing was reduced upon increasing the concentration of perchlorate in the BGE. The order of elution in Figure 2 was markedly different to that observed for an open-tubular AS18coated capillary with the same BGE, highlighting the increase in ion-exchange interactions occurring with the higher capacity latexcoated monolithic stationary phase. The reproducibility of the

Figure 2. CEC separation of anions on a 75-µm-i.d. AS18-coated polymer monolith column. Conditions: column, 23.8 cm × 75 µm AS18-coated polymer monolith attached via Upchurch PEEK zero dead volume union to 10 cm × 75 µm AS18-coated OT separation segment with window 8.5 cm from end; BGE, 10 mM Tris/ClO4-; injection, 0.3 MPa, 6 s of 1 mM Br-, I-, NO3-, CrO42-, SCN-, BrO3and IO3- in water; voltage, -30 kV; temperature, 25 °C; UV detection, 195 nm.

electrophoretic mobility is shown in Table 1 for bromide, chromate, and thiourea. The variation in the EOF is also shown for two identical separations: one performed 24 h later and one performed on another monolithic column made in the same batch. The relative standard deviation (RSD) for the migration time, peak area, and peak height over five runs for seven analyte anions varied in the range 0.3-1.4, 1.4-8.2, and 1.8-7.0%, respectively. Typically, it was anions that interacted most strongly with the stationary phase that exhibited the largest variation from run to run. The RSD for seven anions separated under the same conditions on a different column were 7.9, 12.1, and 14.7% for migration time, peak area, and peak height, respectively. It was important to maintain identical coating procedures from columnto-column to achieve similar chromatographic properties. The selectivity of the separation could be varied by changing the concentration of perchlorate in the BGE, which in turn varied the extent of the ion-exchange interactions with the stationary phase and therefore the contributions of both ion-exchange and electrophoresis mechanisms to the overall separation. Figure 3 shows this effect and illustrates the fact that the effective mobilities of anions possessing high ion-exchange selectivity coefficients can be manipulated readily and could therefore be used to optimize Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

411

Figure 4. Van Deemter plot for thiourea (EOF marker), bromide (low anion-exchange selectivity coefficient), and thiocyanate (high anion-exchange selectivity coefficient) on 75-µm AS18-coated polymer monolith. Conditions: as in Figure 2; voltage, -2 to -30 kV as required. Figure 3. Changes in electrophoretic mobilities of inorganic anions when the concentration of the BGE was varied. Conditions: as in Figure 2; voltage, -10 kV; BGE, Tris/perchlorate concentration was varied between 0 and 100 mM.

the separation. Even at the highest concentration of perchlorate used, analyte anions such as chromate and thiocyanate continue to show ion-exchange interactions with the stationary phase. The separation efficiency for a monolithic stationary phase is dependent on the pore structure in the monolith, the kinetics of mass transfer, and eddy diffusion effects within the monolithic structure. Reducing the globule size and hence the pore size has been shown to increase the separation efficiency for small analytes48 but Peters et al.23 have shown column efficiency is also dependent on other factors such as the tortuosity of the throughpores. In the case of the latex-coated monoliths, efficiency was determined largely by the degree of ion-exchange interaction between the analytes and the stationary phase. For example, at 10 mM perchlorate in the BGE, the efficiency for iodide was 4.00 × 104 theoretical plates m-1, but when the perchlorate concentration was increased to 100 mM, 1.18 × 105 plates m-1 could be achieved. Plate heights of 10-20 µm could be readily achieved for seven inorganic anions over a relatively wide range of EOF linear velocities. The corresponding Van Deemter plots for bromide (anion possessing a low ion-exchange selectivity coefficient), thiocyanate (anion possessing a high ion-exchange selectivity coefficient), and an EOF marker are shown in Figure 4 using a 10 mM perchlorate BGE. It was found that the mobility of the EOF became lower as the ionic strength was increased, as shown in Figure 5. While (48) Jiang, T.; Jishra, J.; Claessens, H. A.; Cramers, C. A. J. Chromatogr., A 2001, 923, 215-227.

412 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

Figure 5. Effect of ionic strength on the EOF in a polymeric monolith column. Conditions: as in Figure 2; voltage, -2 to -30 kV as required.

this is typical of CE in open capillaries and is caused by the compression of the electrical surface double layer, it has been documented29 that silica monoliths possessing both mesopores within the skeletal structure and larger through-pores do not always behave in this way due to double layer overlap within the pores. Polymer monoliths do not possess mesopores within the

Figure 6. Schematic diagram of preconcentration/separation method. Step 1: In-line multimode column is preconditioned with 1 M NaF and Milli-Q water. Step 2: Several capillary volumes of sample are flushed through the column by pressure injection, and anions are preconcentrated onto the monolith. Step 3: The column is filled with an electrolyte that possesses an ion-exchange selectivity coefficient low enough so as not to remove anions bound to the stationary phase (weak electrolyte, WE). Step 4: An electrolyte that possesses a high ion-exchange selectivity coefficient (strong electrolyte, SE) is placed in the electrolyte vials, and voltage is applied. The transient isotachophoretic gradient formed moves through the column, compressing the anions into a sharp band at the gradient front. Step 5: When the compressed anions reach the separation portion of the column, they are free to separate electrophoretically and the wall can be coated to provide chromatographic interaction if required.

monolithic skeleton and so the dependence of EOF on ionic strength is more predictable. In-Line Preconcentration of Anions in CE Using LatexCoated Monoliths. One possible use of the permeable ionexchange monoliths developed in this study is for in-line preconcentration in CE. Such an in-line preconcentration column could be established by forming a small monolithic bed at the inlet of a CE capillary or, alternatively, by coupling a short ion-exchange monolithic capillary column to the inlet of a CE capillary. In this study, the latter approach was used as it simplified the synthesis of the monolith by enabling a large column to be made and subsequently cut into short lengths and also permitted a large-diameter preconcentration capillary (of high ionexchange capacity) to be coupled with a smaller diameter separation capillary capable of providing an efficient separation. A range of different internal diameter capillaries were evaluated, and the bulk of this work was performed using 250-µm-i.d. monolithic columns as the preconcentration segment. This diameter provided the highest ion-exchange capacity per unit length for a monolithic column that can be placed in the Agilent 3DCE instrument used. The median pore size of the monolithic material created was 3.4 µm, the porosity was 61.8%, and the bulk density was 0.3036 g/mL. This provided adequate permeability for the monolith to be flushed in a reasonable time on the CE instrument. The preconcentration and separation columns were joined using PEEK Microtight unions and sleeves, which provided zero dead volume coupling between capillaries of differing inner and outer diameters. Previously, PTFE sleeves had been used for this purpose,17,46 but these may allow air bubbles to enter

at the joint if the sleeve is not fitted tightly around the coupled segments. In order for the CE capillary to show the same reversed EOF present in the latex-coated monolithic preconcentration column, the CE capillary was coated with the cationic polymer PDDAC. Figure 6 shows a schematic illustration of the steps involved in sample preconcentration. The sample is first enriched by binding analytes onto the concentrator column, and the CE capillary is then filled with an electrolyte that has a low ionexchange selectivity coefficient with the monolithic stationary phase (the “weak electrolyte”). A “strong electrolyte” that has a high ion-exchange selectivity coefficient is then introduced using a transient isotachophoretic gradient; this elutes the analytes from the concentrator column as a compact band. The analytes are then separated by CE. Figure 7 shows the separation of inorganic anions after preconcentration from aqueous solution, followed by elution with a transient isotachophoretic gradient of Tris/perchlorate and subsequent CE separation on a PDDAC-coated capillary. Figure 8 shows the calibration plots obtained by progressively increasing the time period over which the sample was passed through the concentrator column. Also included in Figure 8 are calibration curves obtained when the sample contained 100 mM chloride. It can be seen that the preconcentration of these inorganic anions gave linear calibrations, but the slope of the calibration was greatly reduced for bromide and nitrate when chloride was present in the sample. This effect was due to competition from chloride during the preconcentration step. In contrast, the calibration for iodide was affected only to a minor extent in the presence of chloride, which Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

413

Figure 8. Calibration plots obtained using sample preconcentration. Anions in Milli-Q water are represented by solid lines, and anions in 100 mM Cl- are represented by dashed lines. Conditions: column, 15-cm, 250-µm AS18-coated polymer monolith attached via Upchurch PEEK zero dead volume union to 80-cm, 75-µm PDDAC-coated OT separation segment with window 8.5 cm from end; WE, 30 mM NaF introduced into column via 7-min, 1.2-MPa flush from inlet; SE, 30 mM Tris/ClO4-; injection, 1.2,MPa, 0-5-min flush of 1 µM Br-, I-, and NO3- in water or 100 mM Cl- as required; voltage, -30 kV; temperature, 25 °C. Figure 7. Preconcentration and separation of inorganic anions in Milli-Q water. Conditions: column, 15-cm, 250-µm AS18-coated monolith attached via PEEK union to 80-cm, 75-µm PDDAC-coated OT detection segment with window 8.5 cm from end; WE, 30 mM NaF; SE, 30 mM Tris/perchlorate; injection, 1.2 MPa, 30 s of 1 µM Br-, NO3-, I-; temperature, 25 °C; voltage, -30 kV; UV detection, 195 nm.

is attributable to the high ion-exchange selectivity coefficient of this species. Preconcentration of Iodide in Seawater. The above results suggested that the in-line preconcentration method could be applicable to the determination of iodide in seawater. Iodine is the most abundant biologically essential minor element in the oceans,49 and its determination is essential to understanding the redox chemistry50 and biological productivity of oceanic waters.51 Dissolved inorganic iodine is present as iodate and iodide; the latter forms in surface waters by the reduction of iodate and persists to depths of several hundred meters.52 Iodide is a thermodynamically less stable species than iodate and generally occurs at lower concentrations, ranging from 0.2 µM in surface waters down to 0.5 nM in deeper waters.52 Analysis of iodide in seawater has typically been the domain of ion chromatography, which can detect down to submicrogram per liter concentrations of iodide.53 CE methods54-57 have been (49) Wong, G. T. F. Rev. Aquat. Sci. 1991, 4, 45-73. (50) Smith, J. D.; Butler, E. C. V.; Airey, D.; Sandars, G. Mar. Chem. 1990, 28, 353. (51) Jickells, T. D.; Boyd, S. S.; Knap, A. H. Mar. Chem. 1988, 24, 61. (52) Campos, M. L. A. M.; Sanders, R.; Jickells, T. Mar. Chem. 1999, 65, 167175. (53) Ito, K. J. Chromatogr., A 1999, 830, 211-217.

414 Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

developed for seawater analysis, but detection sensitivity has generally proved to be a major limitation in the application of these methods to open ocean waters. The most sensitive CE method has been reported by Ito et al. It provides a limit of detection of 0.2 µg/L for iodide,57 sensitivity comparable to an IC method (0.3 µg/L). The recovery of iodide from samples containing up to 600 mM chloride was determined, and it was found that recovery had decreased only marginally to 90% at the highest chloride concentration. Importantly, the calibration curve for iodide remained linear in 600 mM chloride. The preconcentration method was therefore found to be suitable for the determination of trace iodide in seawater (which contains ∼560 mM chloride). Adequate separation of bromide, nitrite, nitrate, and iodide from seawater could be achieved using an AS18 latex-coated monolith attached to a PDDAC-coated separation capillary. However, the separation was improved dramatically when the PDDACcoated separation capillary was replaced with an AS18 latexcoated separation capillary. In this case, the iodide was focused into an extremely sharp band (Figure 9), and this effect was fully reproducible with successive runs and occurred in both the real sample and the calibration standards. The efficiency of (54) Yokota, K.; Fukushi, K.; Ishio, N.; Sasayama, N.; Tekeda, S.; Wakida, S.-i. Electrophoresis 2003, 24, 2244-2251. (55) Hirokawa, T.; Ichihara, T.; Ito, K.; Timberbaev, A. R. Electrophoresis 2003, 24, 2328-2334. (56) Mori, M.; Hu, W.; Haddad, P. R.; Fritz, J. S.; Tanaka, K.; Tsue, H.; Tanaka, S. Anal. Bioanal. Chem. 2002, 372, 181-186. (57) Ito, K.; Ichihara, T.; Zhuo, H.; Kumamoto, K.; Timerbaev, A. R.; Hirokawa, T. Anal. Chim. Acta 2003, 497, 67-94.

Table 3. Calibration Data, Reproducibility, and the Analysis of Four Anions in Southern Ocean Seawatera

calibration range equation of curve R2 value migration time RSD (%), n ) 6 peak area RSD (%), n ) 6 limit of detection (S/N ) 3) concentration in Southern Ocean seawater (n ) 6)

bromide

nitrite

nitrate

iodide

0-2.5 mM y ) 5.07x + 1.06 0.937 0.9 13.7 63 µM 0.66 ( 0.09 mM

0-0.3 µM y ) 15.17x + 7.87 0.975 1.0 8.3 3.2 nM 0.14 ( 0.01 µM

0-90 µM y ) 119.31x + 18.35 0.964 1.0 5.6 0.47 µM 0.024 ( 0.001 mM

0-30 nM y ) 0.16x + 0.32 0.932 1.1 2.7 75 pM 4.9 ( 0.1 nM

a Conditions: column, 15 cm × 250 µm AS18-coated polymer monolith attached via Upchurch PEEK zero dead volume union to 80 cm × 75 µm AS18-coated OT separation segment, with window 8.5 cm from end; WE, 120 mM NaF introduced into column via 5-min, 1.2-MPa flush from inlet; SE, 120 mM Tris/ClO4-; injection, 1.2-MPa, 10-min flush of synthetic seawater samples with varied concentrations of anions for the calibration curve and also for the Southern Ocean seawater sample; voltage, -25 kV; temperature, 25 °C; detection, 214 nm.

tubular latex-coated capillaries43 and can be manipulated by choice of the electrolyte used to establish the transient isotachophoretic gradient. One possible disadvantage of this focusing mechanism is that other analyte ions with large ion-exchange selectivity coefficients and electrophoretic mobilities similar to iodide would comigrate as a single, focused band. However, in the seawater sample there was no observable interference of this kind and the use of a selective detection method (direct UV absorbance) also assisted in the elimination of interferences in the determination of iodide. Analytical performance data for bromide, nitrite, nitrate, and iodide in a synthetic seawater sample containing 560 mM sodium chloride are provided in Table 3, together with results obtained for the sample of Southern Ocean seawater. The reproducibility can be inversely correlated to the ion-exchange selectivity coefficients of the anions, with iodide showing the best precision. The LOD for iodide (determined using a 10-min preconcentration loading time at 1.2 MPa) was 75 pM (9.5 ng/L) at a signal-tonoise ratio of 3 and was ∼21 times lower than that previously achievable by CE using a transient ITP technique.57 The LOD could be further lowered by using a longer loading time during the preconcentration step. For example, increasing the sample loading time by a factor of 6 (to 60 min) gave a commensurate reduction in LOD (to 11.5 pM).

Figure 9. Preconcentration and separation of anions in Southern Ocean seawater sample. Conditions: column, 15 cm × 250 µm AS18coated polymer monolith attached via Upchurch PEEK zero dead volume union to 80 cm × 75 µm AS18-coated OT separation segment with window 8.5 cm from end; WE, 120 mM NaF introduced into column via 5-min, 1.2-MPa flush from inlet; SE, 120 mM Tris/ClO4-; injection, 1.2-MPa, 10-min flush of Southern Ocean seawater; voltage, -25 kV; temperature, 25 °C; UV detection, 230 nm.

the iodide peak was 5.13 × 107 plates m-1, based on the migration time of this species through the separation capillary (measured in a separate experiment in which the preconcentration capillary was removed). This focusing effect was due to the large ion-exchange selectivity coefficient of iodide, which caused it to migrate with the front of the isotachophoretic gradient. This form of focusing has been observed previously with open-

CONCLUSIONS A latex-coated polymer monolithic stationary phase for ionexchange CEC has been demonstrated. The benefits associated with polymer monoliths, such as ease of fabrication, robust performance, pH stability, and favorable mass-transfer characteristics, have been combined with a high-capacity fully functionalized latex that is selective for inorganic anions. This approach enables the ion-exchange properties of the stationary phase to be varied by coating the same template monolith with the desired latex, thereby avoiding the need to perform derivatization reactions on the polymer while it is confined inside the capillary. The surface coverage of the latex onto the sulfonate methacrylate monolith appeared to be relatively sparse, yet still provided sufficient ionexchange properties to obtain analytically useful separations in the CEC mode. Methods to further increase the loading of latex onto the monolith are currently under study. A major benefit of latex-coated methacrylate columns over their silica-based analogues is the robust nature of the polymeric material, which Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

415

enables a longer column life and an extended operational pH range (2-12). The same latex-coated monolithic material was also found to be suitable for in-line preconcentration of inorganic anions. Iodide could be preconcentrated from seawater to provide limits of detection that are ∼20 times lower than those reported for previous capillary electroseparation methods. This sensitivity resulted from the dramatic, reproducible focusing of the iodide peak, which arose because the iodide migrated with the front of the isotachophoretic gradient.

416

Analytical Chemistry, Vol. 77, No. 2, January 15, 2005

ACKNOWLEDGMENT Financial support from the Australian Research Council is gratefully acknowledged. The authors thank John Nailon of the Centre for Microscopy and Microanalysis at the University of Queensland for the SEM images.

Received for review August 23, 2004. Accepted October 17, 2004. AC048748D