Anal. Chem. 2006, 78, 6088-6095
Nanoparticle-Based Continuous Full Filling Capillary Electrochromatography/Electrospray Ionization-Mass Spectrometry for Separation of Neutral Compounds Christian Nilsson,† Peter Viberg,‡ Peter Spe´gel,‡ Magnus Jo 1 rnte´n-Karlsson,§ Patrik Petersson,§ and ,† Staffan Nilsson*
Pure and Applied Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden, Analytical Chemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden, and AstraZeneca R&D Lund, Lund, Sweden
Highly efficient reversed-phase capillary electrochromatography (CEC) separations (plate numbers up to 700 000/ m), with electrospray ionization mass spectrometry detection were achieved utilizing novel dextran-coated polymer nanoparticles as a pseudostationary phase. A continuous full filling (CFF) technique in which nanoparticles are continuously introduced into the capillary was employed for separation of neutral analytes (dialkyl phthalates), utilizing an orthogonal electrospray interface to prevent nanoparticles from entering the mass spectrometer. CFF-CEC benefits from that an entirely fresh column is employed for every analysis, avoiding carryover effects associated with stationary-phase contamination. The highly efficient separations obtained were accomplished by optimizing the organic modifier concentration in the electrolyte and by using a high nanoparticle concentration (5 mg/mL), to improve interparticle mass transfer and gain sufficient retention. Nanoparticles, with an average diameter of 600 nm, were prepared by polymerization of methacrylic acid and trimethylolpropane trimethacrylate, which in turn were coated with dextran. These nanoparticles formed stable suspensions in electrolytes having broad ranges of polarities, enabling straightforward optimization of the reversed-phase conditions. Capillary electrochromatography (CEC) combines the high efficiency of capillary electrophoresis (CE) with the high selectivity and diversity of applications of liquid chromatography (LC). Three main modes of CEC that are generally distinguished are packed,1,2 monolithic,3 and open tubular column CEC.4 Another mode of * Corresponding author. Fax: +46 46 222 4611. E-mail: staffan.nilsson@ teknlk.lth.se. † Pure and Applied Biochemistry, Center for Chemistry and Chemical Engineering, Lund University. ‡ Analytical Chemistry, Center for Chemistry and Chemical Engineering, Lund University. § AstraZeneca. (1) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr. 1974, 99, 2330. (2) Jorgenson, J. W.; Lukacs, K. D. J. Chromatogr. 1981, 218, 209-216. (3) Svec, F. J. Sep. Sci. 2005, 28, 729-745. (4) Guihen, E.; Glennon, J. D. J. Chromatogr., A 2004, 1044, 67-81.
6088 Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
CEC, with high potential, utilizes a pseudostationary phase (PSP). A PSP is continuously replaced instead of being immobilized, and an entirely fresh column is used for every analysis. Carryover effects associated with stationary-phase contamination are absent, which makes stationary-phase reconditioning unnecessary and increases the potential to analyze samples in complex matrixes, such as biological samples. In addition, the interaction phase can easily be exchanged without changing the column. Furthermore, the need for retaining frits and complicated packing procedures, necessary in packed column CEC, is avoided, and the sample capacity is potentially higher than in open tubular column CEC due to a greater amount of surface area that is available. Another advantage compared to packed column CEC is the possibility to use submicrometer particles to improve the separation efficiency. This opportunity is severely limited in packed column CEC, in which the back-pressure increases rapidly with decreasing particle size. The use of PSPs in CEC and electrokinetic chromatography has been reviewed by Nilsson and Nilsson5 and Palmer and McCarney,6 respectively. The present study is based on knowledge from an earlier study7 in which underivatized polymer nanoparticles were used as a PSP in CEC-mass spectrometry (MS), coupled via an orthogonal electrospray interface, for separation of ionic analytes, utilizing ionic interactions with the nanoparticles. It was not possible to perform reversed-phase separations because nanoparticles did not form stable suspensions in electrolytes having low concentrations of organic modifier. One of the first PSPs used in a CE system for separation of neutral analytes was micelles,8 and the corresponding separation technique was named micellar electrokinetic chromatography (MEKC). Important limitations with MEKC include the existence of an equilibrium between free surfactants and micelles, alternation of this equilibrium with addition of organic solvent, and difficulty in combining MEKC with MS detection because most surfactants (5) Nilsson, C.; Nilsson, S. Electrophoresis 2006, 27, 76-83. (6) Palmer, C. F.; McCarney, J. P. Electrophoresis 2004, 25, 4086-4094. (7) Viberg, P.; Jornten-Karlsson, M.; Petersson, P.; Spegel, P.; Nilsson, S. Anal. Chem. 2002, 74, 4595-4601. (8) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1984, 56, 111-113. 10.1021/ac060526n CCC: $33.50
© 2006 American Chemical Society Published on Web 07/29/2006
suppress analyte ionization and are responsible for contamination of the MS. Coupling of MEKC with MS has been described using atmospheric pressure chemical ionization,9 atmospheric pressure photoionization,10 and electrospray ionization (ESI) with MScompatible surfactants11 as well as molecular micelles.12 To solve the problems associated with MEKC, nanoparticles have been used in PSP-CEC.5 Applications of nanoparticles and nanotechnology have accelerated during the recent years. Guihen and Glennon have reviewed the use of nanoparticles in separation science,13 including LC, gas chromatography, CE, CEC, and microchip electrochromatography. Nanoparticle interaction phases used in PSP-CEC include polymer nanoparticles,14 molecularly imprinted polymer nanoparticles,15 silica nanoparticles,16,17 gold nanoparticles,18 dendrimers (branched organic macrocompounds having snowflake structure),19 and molecular micelles.20-22 Unfortunately, UV detection is often disturbed by high background signals due to the light-scattering properties of the nanoparticles. To enable detection, a partial filling (PF) technique23-25 has been used in which a plug of interaction phase is injected into the capillary prior to the sample. The analytes are then detected as they migrate through the plug. However, PF-CEC requires a significant mobility difference between nanoparticles and analytes, as well as optimization in nanoparticle plug length, to avoid coelution. To avoid the disadvantages with PF, MS detection with an orthogonal electrospray interface was used to enable a continuous full filling (CFF) technique,7 in which the interaction phase is suspended in the entire capillary volume and is continuously introduced. In the orthogonal electrospray interface, the positively charged sample molecules were extracted from the electrospray plume and accelerated in the electric field toward the inlet of the mass spectrometer, whereas high molecular weight nanoparticles did not diverge from the electrospray plume and, hence, never entered the mass spectrometer.7 CFF is easier to optimize and facilitates introduction of higher amounts of nanoparticles into the capillary to enable more demanding separations, as compared to PF. To perform efficient separation in nanoparticle-based PSP-CEC, certain separation conditions and nanoparticle properties, need (9) Isoo, K.; Otsuka, K.; Terabe, S. Electrophoresis 2001, 22, 3426-3432. (10) Mol, R.; de Jong, G. J.; Somsen, G. W. Anal. Chem. 2005, 77, 5277-5282. (11) Petersson, P.; Jornten-Karlsson, M.; Stalebro, M. Electrophoresis 2003, 24, 999-1007. (12) Akbay, C.; Rizvi, S. A. A.; Shamsi, S. A. Anal. Chem. 2005, 77, 1672-1683. (13) Guihen, E.; Glennon, J. D. Anal. Lett. 2003, 36, 3309-3336. (14) Wallingford, R. A.; Ewing, A. G. Adv. Chromatogr. 1989, 29, 1-76. (15) Schweitz, L.; Spegel, P.; Nilsson, S. Analyst 2000, 125, 1899-1901. (16) Bachmann, K.; Gottlicher, B.; Haag, I.; Han, K. Y.; Hensel, W.; Mainka, A. J. Chromatogr., A 1994, 688, 283-292. (17) Bachmann, K.; Gottlicher, B. Chromatographia 1997, 45, 249-254. (18) Huang, M. F.; Kuo, Y. C.; Huang, C. C.; Chang, H. T. Anal. Chem. 2004, 76, 192-196. (19) Castagnola, M.; Zuppi, C.; Rossetti, D. V.; Vincenzoni, F.; Lupi, A.; Vitali, A.; Meucci, E.; Messana, I. Electrophoresis 2002, 23, 1769-1778. (20) Palmer, C. P.; McNair, H. M. J. Microcolumn Sep. 1992, 4, 509-514. (21) Palmer, C. P.; Khaled, M. Y.; McNair, H. M. J. High Resolut. Chromatogr. 1992, 15, 756-762. (22) Ozaki, H.; Itou, N.; Terabe, S.; Takada, Y.; Sakairi, M.; Koizumi, H. J. Chromatogr,. A 1995, 716, 69-79. (23) Amini, A.; Paulsen-Sorman, U.; Westerlund, D. Chromatographia 1999, 50, 497-506. (24) Tanaka, Y.; Terabe, S. J. Chromatogr., A 1995, 694, 277-284. (25) Valtcheva, L.; Mohammad, J.; Pettersson, G.; Hjerten, S. J. Chromatogr. 1993, 638, 263-267.
to be fulfilled. An equation describing the total column bandbroadening in nanoparticle-based PSP-CEC, which is a modified version of an equation by Go¨ttlicher and Ba¨chmann,26 is as follows,
Htot ) Hl + Hm + Haq + Hp + HT + Hep
(1)
where Hl originates from the longitudinal diffusion, Hm originates from kinetics, Haq originates from interparticle mass transfer, Hp originates from intraparticle mass transfer, HT originates from radial temperature gradient, and Hep originates from dispersion due to different nanoparticle velocities. Five mechanisms causing band-broadening in MEKC are discussed in more detail by Terabe et al.27 Equation 1 contains a sixth term, not discussed by Terabe et al., that originates from intraparticle mass transfer (this term is less prominent in MEKC). In summary, the nanoparticles should form stable suspensions in the desired electrolytes, provide selective interactions with analytes, be charged not to coelute with electroosmotic flow (EOF) (for separation of neutral analytes), be monodisperse (to minimize band-broadening caused by dispersion due to different nanoparticle velocities and unfavorable mass transfer), and be small and porous (to provide fast mass transfer and high sample capacity). The main goal of this study was to synthesize nanoparticles, which could form stable suspensions in solutions with low concentration of organic modifier, to enable efficient reversedphase CFF-CEC separations of neutral analytes. EXPERIMENTAL SECTION Chemicals. 2,2′-Azobis(isobutyronitrile) (AIBN) was obtained from Sigma (St. Louis, MO). Trimethylolpropane trimethacrylate (TRIM), carbonyldiimidazole (CDI), ethylene glycol, glycerol, dimethyl phthalate, diethyl phthalate, dipropyl phthalate, and dibutyl phthalate were purchased from Aldrich (Gillingham, U. K.). Methacrylic acid (MAA), ammonium acetate, dimethyl sulfoxide (DMSO), methanol, and acetonitrile were obtained from Merck (Hohenbrunn, Germany). Dextran (Mw ) 15 000-20 000) was purchased from ICN Biomedical Inc (Aurora, OH), sodium hydroxide was from Eka Nobel (Sweden), and acetic acid and formic acid were from Riedel-de Haen (Seelze, Switzerland). Water was purified by a Milli-Q purification system (Bedford, MA). All reagents were of analytical grade and were used as obtained. Polymerization. Polymer nanoparticles were prepared by precipitation polymerization.7,15 The monomer MAA (0.109 M), the cross-linking monomer TRIM (0.109 M), and the radical initiator AIBN (0.012 M) were dissolved in acetonitrile in a screwcapped borosilicate glass test tube. The mixture was sonicated (Branson 1100) for 10 min, degassed by a flow of nitrogen for 6 min, and put into a freezer at -20 °C. Polymerization was initiated by UV light at 350 nm and interrupted after 2 h. The obtained nanoparticles were washed twice in methanol using repeated centrifugation (5000 rpm for 10 min, Sigma 201M) and resuspension by sonication (10 min, Branson 1100). The nanoparticles were stored dry until further use. Surface Modification. A proposed reaction scheme for surface modification of the polymer nanoparticles includes consecutive saponification and esterification (Figure 1). Purification of the reaction products is easily performed by washing the (26) Gottlicher, B.; Bachmann, K. J. Chromatogr., A 1997, 780, 63-73.
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
6089
Figure 1. Proposed reaction scheme for surface modification reaction. The figure illustrates the structure of (1) the polymerized nanoparticle cores, (2) the nanoparticles after saponification, (3) the imidazole-activated nanoparticles, and (4) the dextran-coated nanoparticles. The dextrancoated nanoparticles were used throughout the study.
nanoparticles after reactions using centrifugation and resuspension by sonication. The surface-modified nanoparticles were characterized by transmission electron microscopy (TEM) (JEOL 3000F field emission transmission electron microscope, JEOL, Tokyo, Japan). Saponification. Saponification (base-promoted hydrolysis of esters) was used to increase the amount of carboxylic acid groups at the nanoparticle surface (Figure 1). A volume of 2.5 mL of sodium hydroxide (1 M) was added to a solution of 12 mg of nanoparticles (nanoparticle 1 in Figure 1) dissolved in 1 mL of methanol. The mixture was stirred at 80 °C for 4 h. The reaction was interrupted, and the nanoparticles (nanoparticle 2 in Figure 1) were washed two times with methanol using repeated centrifugation (5000 rpm for 10 min, Sigma 201M) and resuspension by sonication (10 min, Branson 1100). The nanoparticles were stored in methanol until further use to prevent the formation of strong aggregates when drying the nanoparticles. Esterification. Esterification was performed to couple dextran covalently to the surface of the nanoparticles, using CDI to activate the carboxylic groups28 (Figure 1). Carboxyl group activation is normally performed in a few minutes at room temperature and is faster than hydroxyl group activation (from impurities, i.e., methanol).29 Subsequent reaction between dextran and the (27) Terabe, S.; Otsuka, K.; Ando, T. Anal. Chem. 1989, 61, 251-260. (28) Bamford, C. H.; Middleton, I. P.; Allamee, K. G. Polymer 1986, 27, 19811985.
6090
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
activated nanoparticles (nanoparticle 3 in Figure 1) that normally is performed above room temperature was performed at room temperature. An amount of 6 mg of nanoparticles (nanoparticle 2 in Figure 1) dissolved in 1 mL of DMSO was mixed with 4.19 mg CDI dissolved in 0.5 mL DMSO. The mixture was stirred for 2 h at room temperature. An amount of 13.65 mg dextran dissolved in 0.7 mL DMSO was added, and the reaction was allowed to proceed under stirring overnight. Thereafter, 150 µL of glycerol and 100 µL of ethylene glycol were added at 2 h intervals, and the reactions were kept under stirring for 2 h, respectively. The nanoparticles (nanoparticle 4 in Figure 1) were washed two times with methanol and two times with water, using repeated centrifugation (5000 rpm, 10 min, Sigma 201M) and were resuspension by sonication (10 min, Branson 1100). The nanoparticles were stored in water until further use to prevent strong aggregates from being formed when drying the nanoparticles. The suspension stability of the nanoparticles was tested by suspending them in different water/acetonitrile and buffer/acetonitrile mixtures and observing aggregation and sedimentation. Capillary Electrophoresis. CE experiments were preformed on a Beckman P/ACE System 5010 (Beckman, Fullerton, CA) equipped with a UV detector. Fused-silica capillaries (75-µm i.d.; 375-µm o.d.) with 37-cm total length and 29.4-cm effective length were obtained from Polymicro Technologies (Phoenix, AZ). UV (29) Staab, H. A. Angew. Chem. 1962, 74, 407-423.
aqueous buffer, 20 mM ammonium acetate, was adjusted to pH 5.6 with an acetic acid/water solution. Sample solutions contained dimethyl, diethyl, dipropyl, and dibutyl phthalate dissolved in electrolyte to a concentration of 20 µg/mL (103, 90, 80, and 72 µM, respectively). Prior to use, all solutions and nanoparticle suspensions were degassed by sonication. The capillary was rinsed with sodium hydroxide (0.1 M), water, and electrolyte between consecutive runs. MS Detection. MS detection was performed on an Agilent Technologies LC/MSD ion trap SL mass spectrometer equipped with an orthogonal ESI interface operated in positive ionization mode scanning between m/z 150 and 300, with a maximum ion accumulation time between 50 and 75 ms and a total ion current between 75 000 and 100 000. The sheath liquid consisted of 0.1% formic acid in water/ methanol (1:1 v/v) and was pumped at 0.180 mL/min by an Agilent Technologies series 1100 quaternary pump and split 1:30.8 by a fixed splitter. The CE instrument was coupled to the ESI interface using an Agilent Technologies triple tube coaxial nebulizer.
Figure 2. Transmission electron micrographs of polymer nanoparticles having an average diameter of ∼600 nm. Panel a shows nanoparticles (black) deposited on a carbon grid (gray). Panel b shows nanoparticles (dark gray) deposited on a carbon grid (gray).
detection was performed at 214 nm. Nanoparticles (nanoparticle 4 in Figure 2) (5 mg/mL) were injected hydrodynamically (3 s at 34 mbar). All separations were performed at 25 °C with a 20-kV separation voltage. The electrolyte was a mixture of acetonitrile and an aqueous buffer (25:75 v/v). The aqueous buffer, 20 mM ammonium acetate, was adjusted to pH 5.6 with dilute acetic acid. Prior to use, all solutions and nanoparticle suspensions were degassed by sonication. The capillary was rinsed with sodium hydroxide (0.1 M), water, and electrolyte between consecutive runs. Capillary Electrochromatography. CEC was performed on a HP3DCE system (Agilent Technologies, Waldbronn, Germany), and ChemStation software was used for data processing. A 70cm-long, 50-µm-i.d., and 375-µm-o.d. fused-silica capillary obtained from Polymicro Technologies (Phoenix, AZ) was used in all CEC experiments. Samples were injected hydrodynamically (3 s at 50 mbar), and all separations were performed at room temperature using a separation voltage of 30 kV. The electrolytes were mixtures of acetonitrile and an aqueous buffer (30:70 and 40:60 v/v), containing nanoparticles (0.5, 1.0, 2.0, 5.0, and 10.0 mg/mL). The
RESULTS AND DISCUSSION Nanoparticle Polymerization and Surface Modification. The nanoparticles were synthesized by a two-step procedure, that is, polymerization followed by surface modification. In precipitation polymerization, spherical polymer nanoparticles are formed as the growing polymer reaches the solubility limit due to increasing molecular weight. Polymerization yielded hydrophobic, spherical nanoparticles that were stable toward aggregation and sedimentation in acetonitrile. Subsequent surface modification of the polymerized nanoparticles yielded spherical, hydrophilically coated nanoparticle having a smooth surface and an apparent narrow size distribution as shown in Figure 2. The average diameter of the nanoparticles, determined by TEM, was 600 nm, with a relative standard deviation of 8.4% (n ) 55). The nanoparticles were synthesized without use of surfactants or emulsion stabilizers,16 circumventing extensive washing that otherwise is necessary to avoid interference with nanoparticle derivatization, disturbance of the electrospray process, and contamination of the mass spectrometer. The surface-modified nanoparticles formed stable suspensions (no sedimentation was observed overnight) in electrolytes having a wide variety of polarities, in contrast to the unmodified nanoparticles that required relatively high organic modifier concentrations. The suspension stability was governed by the hydrophilic character of the nanoparticle coating and probably also by steric repulsion between the dextran polymers on the nanoparticle surfaces. Capillary electrophoresis. CE was performed to evaluate the quality of the nanoparticles. As seen in Figure 3, showing an electropherogram from a nanoparticle injection, a smooth, symmetrical peak with Gaussian shape was observed, indicating a stable suspension and no or weak capillary wall adsorption. Interactions of nanoparticles with the capillary wall would probably result in excessive tailing of the nanoparticle peak. The hydrophilic character of the dextran-coated nanoparticles as well as electrostatic repulsion between the nanoparticles and the capillary wall are two factors likely to reduce such adsorption. A small broadening of the nanoparticle peak indicates that there is a nanoparticle mobility distribution. Whether this is due to the size distribution or charge differences of the nanoparticles or some other effect is Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
6091
Figure 3. Electropherogram from a nanoparticle injection (5 mg/ mL, 5 s, 34 mbar). Separation was performed at 20 kV, using a 75µm-i.d. capillary with 37-cm total length and 29.4-cm effective length. UV detection was performed at 214 nm. The small peak is, to our knowledge, a system peak due to injection, whereas the big peak represents the nanoparticle plug.
not yet fully understood. To our knowledge, the small peak seen in Figure 3 represents a system peak30 (due to injection) not seen with MS detection. Monomers, reagents or solvents remaining from the synthesis of the nanoparticles could be eluting at the time of this peak, although the monomers and solvents do not absorb light at the wavelength used. The height of the nanoparticle peak gives a measure of the light scattering of the nanoparticles if CEC/UV should be used. Thus, PF is the preferred technique. CE/UV is a convenient way to evaluate nanoparticles, yielding information that is complementary to what is obtained with TEM, which is more time-consuming and expensive. CE/UV could be used for evaluation during early stages of the nanoparticle development. Capillary Electrochromatography. CFF-CEC separations with plate numbers up to 700 000 (for diethyl phthalate) using a homologue series of dialkyl phthalates (Figure 4) as sample molecules were achieved. The highest efficiency and baseline separation of the dialkyl phthalates were achieved using an electrolyte containing 30% acetonitrile and 5 mg/mL nanoparticles. Separations were based on interactions between the neutral analytes and nanoparticles. The elution order follows the size of the alkyl groups of the sample molecules, which indicates a reversed-phase mechanism. Loss in separation resolution following an increase in the organic modifier concentration further indicates a reversed-phase mechanism. The high separation efficiency achieved indicates that the nanoparticles used had similar chemical properties. To obtain separations with a high resolution, it is important to optimize efficiency and retention (also selectivity). In this study, an optimization process comprising simultaneous optimization of nanoparticle and acetonitrile concentration in the electrolyte was implemented (based on the terms in eq 1). The amount of interaction phase can easily be optimized, in contrast to separations with traditional CEC columns. In addition, the electrolyte composition, that is acetonitrile concentration, can be varied due to the suspension stability of the surface-modified nanoparticles. Figure 5 shows a series of electrochromatograms (reconstructed ion chromatogram (RIC)) showing CFF-CEC separations in which nanoparticle and acetonitrile concentration were varied. (30) Gas, B.; Kenndler, E. Electrophoresis 2004, 25, 3901-3912.
6092 Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
Figure 4. Electrochromatogram from separation of dimethyl phthalate (1), diethyl phthalate (2), dipropyl phthalate (3), and dibutyl phthalate (4) at 20 µg/mL. Separation was performed at 30 kV using a capillary with 70-cm length. The electrochromatogram represents a reconstructed ion chromatogram, monitoring m/z 195 ( 0.5, 223 ( 0.5, 251 ( 0.5, and 279 ( 0.5. The electrolyte contained 30% acetonitrile and 5 mg/mL nanoparticles.
It is clearly seen that at an acetonitrile concentration of 30%, partial resolution of the analytes was already achieved at 0.5 mg/mL nanoparticles. When an electrolyte containing 40% acetonitrile was used, a higher concentration of nanoparticles was necessary to obtain separation. It can be concluded that the acetonitrile concentration is a crucial parameter to obtain sufficient hydrophobic interactions and also favorable kinetics. In addition to increasing the retention, the use of a high nanoparticle concentration increases efficiency due to a more favorable interparticle mass transfer, and an improved sample capacity, thus avoiding overloading. Figure 6a shows how the retention time normalized against the unretained dimethyl phthalate varies with nanoparticle concentration at an acetonitrile concentration of 30%. When the nanoparticle concentration was increased from 5 to 10 mg/mL, decreased peak intensity was observed, most likely related to disturbance of the electrospray process (Figure 6b). If this problem can be understood and solved, even higher nanoparticle concentrations can be used, which might even further improve retention and separation efficiency. An increased EOF was observed with an increased nanoparticle concentration (Figure 5). The reason for this is not fully understood, whether it is related to alternation in ionic strength of the electrolyte due to a higher volume of nanoparticles or some other effects. With an electrolyte containing 30% acetonitrile and 5 mg/mL nanoparticles, highly efficient separation was achieved in 8.5 min. In contrast to the previous study by Viberg et al.,7 the current study uses nanoparticles that form stable suspensions in electrolytes with low acetonitrile concentrations (under 50%), which enables a reversed-phase separation mechanism. In the previous study, cationic analytes were separated using ion exchange CFFCEC/ESI-MS, whereas the dextran-modified nanoparticles used in the current study enabled high efficient reversed-phase separations to be performed. The high efficiency was enabled using much higher nanoparticle concentrations in comparison to the previous study. The main drawback with nanoparticle-based CFFCEC as compared to conventional CEC is the lower interaction
Figure 5. Electrochromatograms from separations of dimethyl phthalate (1), diethyl phthalate (2), dipropyl phthalate (3), and dibutyl phthalate (4) at 20 µg/mL. Separations were performed at 30 kV using a capillary with 70-cm length. Each electrochromatogram represents a reconstructed ion chromatogram, monitoring m/z 195 ( 0.5, 223 ( 0.5, 251 ( 0.5, and 279 ( 0.5. The electrolytes contained 30 or 40% acetonitrile and 0, 0.5, 1.0, 2.0, 5.0, or 10.0 mg/mL nanoparticles.
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
6093
Figure 6. (a) Normalized retention time versus nanoparticle concentration for dimethyl phthalate ((), diethyl phthalate (9), dipropyl phthalate (2), and dibutyl phthalate (×), with the acetonitrile concentration constant at 30%. (b) Peak intensities for dimethyl phthalate ((), diethyl phthalate (9), dipropyl phthalate (2), and dibutyl phthalate (×) versus nanoparticle concentrations, with the acetonitrile concentration constant at 30%.
phase volume that so far has been used due to practical problems with detection and nanoparticle slurry stability at concentrations comparable to that in a packed column. If the problems with detection can be solved, a higher nanoparticle concentration can be used that enables separations with high resolution and high sample capacity. In addition, stacking is easier to perform in a packed column, although sweeping would be possible in CFFCEC. Future improvement of the technique by improving the nanoparticles will mainly focus on synthesis of smaller, more porous nanoparticles that are monodisperse. Smaller, more porous nanoparticles are wanted to increase efficiency (by improving mass transfer) and improve sample capacity; however, nanoparticles should be large enough to be separated from the analytes before entering the mass spectrometer. The use of smaller nanoparticles is possible in EOF-driven separation systems because packed columns and hydrodynamically pumped flows are avoided. To improve the interactions between the nanoparticles and the analytes, studies on the hydrophobic core properties are worthwhile. Nanoparticle improvement focused on achieving a more narrow size distribution is beneficial to minimize dispersion due to different nanoparticle velocities (assumes that the mass/charge ratio is not constant). Ultimately, nanoparticles can be sorted according to size to achieve monodisperse nanoparticles. This can be achieved using a “bumper array,”31 which is a deterministic device separating particles with extremely high selectivity. The resolution is better than 2% standard deviation, and is independent of particle diffusion. (31) Huang, L. R.; Cox, E. C.; Austin, R. H.; Sturm, J. C. Science 2004, 304, 987-990.
6094
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
Other improvements of the method may involve further miniaturization by decreasing capillary dimensions to achieve faster separation. One goal is to be able to perform nanoparticlebased CFF-CEC on a chip. An advantage with chip-based CEC is the possibility to introduce cross-flows to perform gradient CEC separations (organic modifier or nanoparticle concentration gradient). Gradient CEC on a chip has previously been performed by Yan et al.32 Multiplex CEC is another attractive application for parallel separations that utilizes different nanoparticle suspensions.33 The dextran coating used provides high suspension stability, minimized capillary wall adsorption, and high protein compatibility.34 MS Detection. The use of MS having an orthogonal electrospray interface in combination with covalently coated nanoparticles as a PSP enabled use of the CFF technique for separation of neutral analytes. The mass spectrometer was used for several weeks without any apparent nanoparticle contamination. No ions related to the nanoparticles were observed in the scanned m/z range during MS detection. However, it was noted that the peak intensity was affected negatively by nanoparticle concentration, which suggests that the nanoparticles interfered with the ionization process (Figure 6b). To perform sensitive ESI-MS, the droplets should generate many daughter droplets. The nanoparticles can possibly interfere with the droplet formation and affect the droplet size and the droplet fission process. In ESI, ions are formed from the outer layer of the droplet,35-37 and any analyte interaction with a solid particle in the droplet would prevent the analyte from approaching the surface and ionizing. Another possible explanation is that the nanoparticles affected the efficiency in the droplet fission process, which is fundamental to the ESI mechanism. The most retained analytes decreased most in peak intensity, which suggests a higher influence of analyte ionization disturbance on detection. To improve detection, the sheath liquid composition could be further optimized to increase analyte desorption in the electrospray. Other optimization strategies are to scan a narrower mass interval to improve signal-tonoise (S/N), or to apply a higher ionization voltage, although this can cause nanoparticles to enter the MS. However, in the present study an ion trap was used and thus the use of a narrow scan interval or selected ion monitoring (SIM) is not an opinion. The use of a quadrupole instrument and SIM mode to improve limit of detection (LOD) will be an important issue in the continued work. CONCLUSIONS Separation of neutral analytes using CFF-CEC with MS detection yielded highly efficient separations (plate numbers up to 700 000). The separations were achieved by careful optimization of the nanoparticle concentration and electrolyte polarity. Alternations in interaction-phase amount are easier to perform in CFFCEC than in traditional chromatographic methods, which adds another parameter to the optimization strategy. High concentra(32) Yan, C.; Dadoo, R.; Zare, R. N.; Rakestraw, D. J.; Anex, D. S. Anal. Chem. 1996, 68, 2726-2730. (33) Tan, H. D.; Yeung, E. S. Anal. Chem. 1998, 70, 4044-4053. (34) Gelotte, B. J. Chromatogr. 1960, 3, 330-342. (35) Sjoberg, P. J. R.; Bokman, C. F.; Bylund, D.; Markides, K. E. Anal. Chem. 2001, 73, 23-28. (36) Cech, N. B.; Enke, C. G. Anal. Chem. 2001, 73, 4632-4639. (37) Cech, N. B.; Krone, J. R.; Enke, C. G. Anal. Chem. 2001, 73, 208-213.
tions (5 mg/mL) of novel hydrophilic coated nanoparticles were used as a PSP. To enable detection of the analytes coeluting with a continuous flow of nanoparticles, MS detection with an orthogonal electrospray interface was necessary. The nanoparticles synthesized in this study had an apparent narrow size distribution, minimal capillary wall adsorption, and an apparent high chemical homogeneity. The nanoparticles formed stable suspensions in electrolytes having broad ranges of polarities, thus enabling careful optimization of the reversed-phase conditions. In comparison to conventional methods, analysis turnover time could be kept low, because samples can be continuously injected into a continuously regenerating CEC column. No flushing or washing of the stationary phase is thus needed. Even though all separations were performed with a fresh nanoparticle phase, the consumption of nanoparticle was only ∼10 µg/analysis.
ACKNOWLEDGMENT We thank the Swedish Research Council (VR) (S. Nilsson, Grant no. 40455901), Novo Nordisk A/S, the Ministry of Science Technology and Innovation (Denmark), Craafordska stiftelsen, and Carl Tryggers stiftelse for financial support. Part of the work was presented at the 8th International Conference on Miniaturized Systems for Chemistry and Life Science (µTAS), Malmo¨, Sweden, 2004, and at the 18th International Symposium in MicroScale Bioseparations, New Orleans, LA, 2005.
Received for review March 22, 2006. Accepted May 22, 2006. AC060526N
Analytical Chemistry, Vol. 78, No. 17, September 1, 2006
6095
