Sol−Gel Monolithic Columns with Reversed Electroosmotic Flow for

Additionally, this reagent served to yield a positively charged surface, thereby providing the relatively strong reversed electroosmotic flow (EOF) in...
0 downloads 12 Views 329KB Size
Anal. Chem. 2000, 72, 4090-4099

Sol-Gel Monolithic Columns with Reversed Electroosmotic Flow for Capillary Electrochromatography James D. Hayes and Abdul Malik*

Department of Chemistry, University of South Florida, Tampa, Florida 33620-5250

Sol-gel chemistry was used to prepare porous monolithic columns for capillary electrochromatography. The developed sol-gel approach proved invaluable and generates monolithic columns in a simple and rapid manner. Practically any desired column length ranging from a few tens of centimeters to a few meters may be readily obtained. The incorporation of the sol-gel precursor, N-octadecyldimethyl[3-(trimethoxysilyl)propyl]ammonium chloride, into the sol solution proved to be critical as this reagent possesses an octadecyl moiety that allows for chromatographic interactions of analytes with the monolithic stationary phase. Additionally, this reagent served to yield a positively charged surface, thereby providing the relatively strong reversed electroosmotic flow (EOF) in capillary electrochromatography. The enhanced permeability of the monolithic capillaries allowed for the use of such columns without the need for modifications to the commercial CE instrument. There was no need to pressurize both capillary ends during operation or to use high pressures for column rinsing. With the developed procedure, no bubble formation was detected during analysis with the monolithic capillaries when using electric field strengths of up to 300 V cm-1. The EOF in the monolith columns was found to be dependent on the percentage of organic modifier present in the mobile phase. Separation efficiencies of up to 1.75 × 105 plates/m (87 300 plates/ column) were achieved on a 50 cm × 50 µm i.d. column using polycyclic aromatic hydrocarbons and aromatic aldehydes and ketones as test solutes. Capillary electrochromatography (CEC) is a fairly novel electrokinetic separation technique representing a hybrid of highperformance liquid chromatography (HPLC) and capillary electrophoresis (CE). In CEC, the electroosmotic flow (EOF) is used to drive the mobile phase through the capillary, using typical HPLC mobile and stationary phases that provide the essential chromatographic interactions. Because of the flat pluglike profile of the electroosmotic flow, CEC offers greatly enhanced separation efficiencies relative to HPLC. Unlike CE, CEC is not restricted to charged solutes as both neutral and charged species may interact with the stationary phase. Thus, the potential for CEC, as a * Corresponding author: (phone) (813) 974-9688; (fax) (813) 974-3203; (e-mail) [email protected].

4090 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

separation technique is much wider. Pretorius1 was one of the first influential pioneers of CEC who, in 1974, demonstrated the advantages of electroosmosis as a pumping mechanism for chromatographic separations. Jorgenson and Lucas published CEC analyses of 9-methylanthracene and perylene on an ODSpacked capillary column.2 Meanwhile, a 1987 report by Tsuda demonstrated the possibility of achieving CEC separations by the simultaneous use of both electroosmotic- and pressure-driven flows in the separation column.3 Yet another significant contribution to the development of this technique was made by Knox and Grant.4 Following this publication, the term “electrochromatography” became generally accepted and numerous researchers refocused their attention to CEC. Capillary electrochromatography is a rapidly growing area in analytical separations. A great deal of research effort is currently being devoted to materialize the great analytical potential that this new hybrid technique has to offer. In order for CEC to achieve success as an independent chromatographic separation technique, significant advancements are needed in the area of column technology. This is explained by the fact that, in CEC, the column serves not only as the separation chamber but also as the pumping device to drive the mobile phase through the system. This makes the column the “heart” of the CEC system both in the functional and in literal sense of the word. Two major types of columns are used in current CEC practices packed and open tubular types. Packed columns comprise the predominant class of CEC columns. Most often the packed capillaries contain l.5-5 µm, nonpolar, octadecylated (ODS) particles. The ODS particles possess both the chemically bonded octadecyl stationary phase, providing the essential chromatographic interactions, and the silanol moieties, responsible for the generation of electroosmotic flow to drive the mobile phase and the solutes through the packed capillary. The commercial availability of the ODS-bonded particles and the previously established LC separation protocols are two advantages attracting many researchers to use these packed capillaries in CEC. However, the most significant advantage of packed columns in CEC is the possibility of using small micrometer and nanometer size particles. High separation efficiency during fast analysis is achieved in packed-CEC without requiring ultrahigh pressures, as in HPLC, (1) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr. 1974, 99, 23. (2) Jorgenson, J. W.; Lukacs, K. D. J. Chromatogr. 1981, 218, 209. (3) Tsuda, T. Anal. Chem. 1987, 59, 521. (4) Knox, J. H.; Grant, I. H. Chromatographia 1987, 24, 135. 10.1021/ac000120p CCC: $19.00

© 2000 American Chemical Society Published on Web 07/29/2000

to drive the mobile phase through the columns packed with such small particles. The greatest challenge here is the preparation of a uniform packing bed out of such small particles. Researchers currently use a variety of packing procedures ranging from slurry,5-17 electrokinetic,18-20 centripetal,21 and supercritical fluid22,23 packing methods. Although a great degree of difficulty still remains associated with the ability to pack long, narrow-bore capillaries. Moreover, packed capillaries require end frits to retain the packing particles within the packed capillary bed. Creation of the retaining frits remains the bottleneck in column preparation as these frits must be rigid enough to retain the particles under a wide range of column packing, rinsing, and operating conditions. Yet the frits must possess a highly porous structure to permit a uniform mobile-phase flow through the entire cross section of the column. Monolithic column technology can effectively overcome both of the difficulties associated with conventional packed capillary column technology. In the monolithic approach, a continuous separation bed representing a single piece of porous material is created inside the capillary using a solution that undergoes chemical and physical changes in the confined environment inside the capillary to produce the separation bed. The choice of appropriate chemistry allows the porous bed to get chemically bonded to the inner walls of the capillary. The use of monolithic columns as an alternative to packed capillaries was originally reported in gas24 and liquid chromatography.25-29,32-34,42-47 This (5) Frame, L. A.; Robinson, M. L.; Lough, W. J. J. Chromatogr., A 1998, 798, 243. (6) Van den Bosch, S. E.; Heemstra, S.; Kraak, J. C.; Poppe, H. J. Chromatogr., A 1996, 755, 165. (7) Smith, N. W.; Evans, M. C. Chromatographia 1994, 38, 649. (8) Moffatt, F.; Cooper, P. A.; Jessop, K. M. Anal. Chem. 1999, 71, 1119. (9) Behnke, B.; Bayer, E. J. Chromatogr., A 1994, 680, 93. (10) Dittmann, M. M.; Rozing, G. P. J. Chromatogr., A 1996, 744, 63. (11) Behnke, B.; Grom, E.; Bayer, E. J. Chromatogr., A 1995, 716, 207. (12) Seifer, R. M.; Kraak, J. C.; Th. Kok, W.; Poppe, H. J. Chromatogr., A 1998, 808, 71. (13) Lu ¨ dtke, S.; Adam, T.; Unger, K. K. J. Chromatogr., A 1997, 786, 229. (14) Zhang, M.; El Rassi, Z. Electrophoresis 1998, 19, 2068. (15) Zhang, M.; El Rassi, Z. Electrophoresis 1999, 20, 31. (16) Zhang, M.; Yang, C.; El Rassi, Z. Anal. Chem. 1999, 71, 3277. (17) Zimina, T. M.; Smith, R. M.; Meyers, P. J. Chromatogr., A 1997, 758, 191. (18) Yan, C.; Dadoo, R.; Zhao, H.; Zare, R. N.; Rakestraw, D. J. Anal. Chem. 1995, 67, 2026. (19) Dulay, M. T.; Yan, C.; Rakestraw, D. J.; Zare, R. N. J. Chromatogr., A 1996, 725, 361. (20) Wei, W.; Luo, G. A.; Hua, G. Y.; Yan, C. J. Chromatogr., A 1998, 817, 65. (21) Colo´n, L. A.; Fermier, A. M.; Guo, Y.; Reynolds, K. J. Ninth International Symposium on High Performance Capillary Electrophoresis (HPCE ‘97), Anaheim, CA, 1997; p 80. (22) Xin, B.; Lee, M. L. Electrophoresis 1999, 20, 67. (23) Tang, Q.; Wu, N.; Lee, M. L. J. Microcolumn Sep. 1999, 11, 550. (24) Hileman, F. D.; Sievers, R. E.; Hess, G. G.; Ross, W. D. Anal. Chem. 1973, 45, 1126. (25) Cortes, H. J.; Pfeiffer, C. D.; Richter, B. E.; Stevens, T. S. J. High Resolut. Chromatogr. Chromatogr. Commun. 1987, 10, 446. (26) Fields, S. M. Anal. Chem. 1996, 68, 2709. (27) Hjerten, S.; Liao, J. L.; Zhang, R. J. Chromatogr. 1989, 473, 273. (28) Hjerten, S.; Li, Y. M.; Liao, J. L.; Mohammad, J.; Nakazato, K.; Pettersson, G. Nature 1992, 356, 810. (29) Li, Y. M.; Liao, J. L.; Nakazato, K.; Mohammad, J.; Terenius, L.; Hjerten, S. Anal. Biochem. 1994, 223, 153. (30) Ericson, C.; Liao, J.-L.; Nakazato, K.; Hjerten, S. J. Chromatogr., A 1997, 767, 33. (31) Ericson, C.; Hjerten, S. Anal. Chem. 1999, 71, 1621. (32) Svec, F.; Frechet, J. M. J. Anal. Chem. 1992, 64, 820. (33) Wang, Q. C.; Svec, F.; Frechet, J. M. J. Anal. Chem. 1993, 65, 2243. (34) Svec, F.; Frechet, J. M. J. J. Chromatogr., A 1995, 702, 89.

approach is currently being used in CEC to alleviate the extensive labor involved with packed column fabrication. Moreover, the greatest inherent advantage of the monolithic capillary columns is the elimination of the need for end frits to retain the stationary phase. The elimination of retaining frits allows the entire column to remain homogeneous, rather than exhibiting different properties by the packing particles and retaining frits. It is reported that the end frits are believed to reduce the column’s separation efficiency and be responsible for bubble formation during the analysis.30 Depending on the nature of the monolithic material, two major directions can be identified: (i) organic polymer-based and (ii) bonded silica-based monolithic column technology. In the first approach, fabrication of monolithic capillaries is accomplished through a single-step polymerization reaction of an organic monomeric precursor. Hileman et al.24 used Carbowax-coated open-pore polyurethane monolithic capillaries for the separations of several classes of analytes including aromatic hydrocarbons, aliphatic alcohols, and metal chelates through gas chromatography. Hjerten et al. prepared monolithic capillaries with compressed polyacrylamide gels for separation of proteins using HPLC27-29 and of low molecular mass compounds30 and basic proteins using CEC.31 Fre´chet and co-workers reported a series of publications on the use of methacrylate monomers for the preparation of HPLC32-34 and CEC35-37 monolithic capillaries through copolymerization. Palm and Novotny prepared CEC monoliths using mixtures of polyacrylamide/poly(ethylene glycol), derivatized with either C4 or C12 ligands, which were used to separate alkyl phenones and peptides.38 Additionally, Fujimoto et al.39,40 reported the usage of cross-linked polyacrylamides for the separation of small dansylated amino acids and neutral steroids and PAHs on monolithic CEC capillaries. Polymer-based monolithic columns possess excellent pH stability, and they are much simpler than packed capillaries. However, monolithic columns prepared through polymerization continue to possess certain limitations. One critical drawback associated with this type of monolithic capillary is its tendency to swell/shrink during exposure to various solvents contained in the running mobile phases.40,41 This swelling/ shrinking may result in reductions in the permeability of the (35) Peters, E. C.; Petro, M.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 1997, 69, 3646. (36) Peters, E. C.; Petro, M.; Svec, F.; Fre´chet, J. M. J. Anal. Chem. 1998, 70, 2288. (37) Svec, F.; Peters, E. C.; Sykora, D.; Yu, C.; Fre´chet, J. M. J. J. H. Resolut. Chromatogr. 2000, 23, 3. (38) Palm, A.; Novotny, M. V. Anal. Chem. 1997, 69, 4499. (39) Fujimoto, C. Anal. Chem. 1995, 67, 2050. (40) Fujimoto, C.; Fujise, Y.; Matsuzawa, E. Anal. Chem. 1996, 68, 2753. (41) Pietrzyk, D. J. In Packings and Stationary Phases in Chromatographic Techniques; Unger, K. K., Ed.; Marcel Dekker: New York, 1990; Vol. 47, Chapter 10. (42) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. Anal. Chem. 1996, 68, 3498. (43) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. J. Chromatogr., A 1997, 762, 135. (44) Nakanishi, K.; Minakuchi, H.; Soga, N.; Tanaka, N. J. Sol-gel Sci. Technol. 1997, 8, 547. (45) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. J. Chromatogr., A 1998, 797, 121. (46) Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Soga, N.; Tanaka, N. J. Chromatogr., A 1998, 797, 133. (47) Tanaka, N.; Nagayama, H.; Kobayashi, H.; Ikegami, T.; Hosoya, K.; Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Cabrera, K.; Lubda, D. J. High Resolut. Chromatogr. 2000, 23, 111.

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4091

monolith as a result of alterations in the porosity of the monolith. Such structural change may ultimately lead to changes in the column during the course of its use. An alternative procedure for the construction of monolithic columns is often with bonded silica stationary phase through the use of sol-gel chemistry. Cortes and co-workers25 prepared porous beds by polymerizing potassium silicate (Kasil) solutions in situ. The columns containing the porous beds were then packed with 5-µm Spherisorb ODS particles for use in LC. Fields used solutions of potassium silicate (Kasil) and formamide to create a porous bed, which was further reacted with dimethyloctadecylchlorosilane, and achieved plate heights of 65 µm in LC.26 Tanaka and co-workers used the sol-gel technique for the development of octadecylsilylated, porous monolithic columns for use in liquid chromatography and CEC.42-47 In these studies, poly(ethylene oxide) (PEO) was incorporated into a mixture of tetramethoxysilane (TMOS) and acetic acid to develop the porous silica rods, followed by an on-column octadecylsilylation reaction. Following washings, and drying at 50 °C for 3 days, the silica rods were then treated for 2 h at 600 °C. Dulay et al.48 used sol-gel technology for the preparation of monolithic columns loaded with 3-µm ODS particles. Here the sol-gel solution served as a retaining matrix, immobilizing and shielding the ODS stationary phase particles. Sol-gel capillary columns containing the ODS embedded particles yielded CEC separation efficiencies on the order of 80 000 plates/m (16 000 plates/column) for a test mixture of six uncharged polyaromatic hyrdrocarbons (PAHs). Xin and Lee22 also used sol-gel chemistry to glue 7-µm ODS particles, thereby creating a continuous large-pore CEC capillary column. The sol-gel technology in this approach was used to create a bridge between adjacent particles, as well as the capillary wall and particles in its vicinity, thereby eliminating the need for retaining frits, and resulted in efficient separations of small organic and aromatic amine compounds on such “sol-gel-glued” monolithic columns. Unlike the monolithic separation beds from organic polymers, columns containing a porous silica-based monolithic matrix prepared through sol-gel chemistry do not suffer from the swelling phenomenon, thus offering a versatile and promising alternative to packed capillaries. In this paper, we introduce a novel sol-gel chemistry-based general approach to the preparation of monolithic columns using a solution only, without requiring the use of particles. The presented sol-gel approach, a single-step process, is used to in situ create a chromatographically favorable porous monolithic bed of stationary phase that is chemically bonded to the inner walls of the fused-silica capillary. In this approach, a commercially available sol-gel precursor, N-octadecyldimethyl[3-(trimethoxysilyl)propyl]ammonium chloride (C18-TMS) is used to create a C18containing monolithic bed. The C18-TMS also contains a positively charged ammonium moiety and is responsible for the generation of EOF in the reversed (anodic) direction. EXPERIMENTAL SECTION Chemicals and Materials. Fused-silica tubing of 50-µm i.d. was purchased from Polymicro Technologies (Phoenix, AZ). Deionized water, ∼18 MΩ, used for the preparation of electrolyte solutions and for column rinsing, was acquired from a Barnstead (48) Dulay, M. T.; Kulkarni, R. P.; Zare, R. N. Anal. Chem. 1998, 70, 5103.

4092

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

model 04741 Nanopure deionized water system (Barnstead/ Thermodyne, Dubuque, IO). Eppendorf, 1.5 mL, microcentrifuge tubes were purchased from Brinkmann Instruments (Westbury, NY). Naphthalene, HPLC grade acetonitrile, methanol, and 0.5mL microcentrifuge tubes, used in all monolithic CEC experiments, were purchased from Fisher Scientific (Pittsburgh, PA). Tetramethyl orthosilicate (99+%), trifluoroacetic acid (99%), methyl sulfoxide (99.9%), anthracene (99%), valerophenone (99%), butyrophenone (99+%), benzaldehyde (99+%), o-tolualdehyde (97%), heptanophenone (98%), thiourea (99%), benzene (99.9+%), toluene (99%), ethylbenzene (99+%), propylbenzene (98%), butylbenzene (99+%), and amylbenzene (99%) were used as test solutes purchased from Aldrich (Milwaukee, WI). Reagent grade Tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl), used in the preparation of background electrolyte solutions, was purchased from Sigma (St. Louis, MO). Adjustments of pH of the electrolyte solutions were achieved through the addition of concentrated (37 wt %) hydrochloric acid (99.999 +%) or 0.1 M sodium hydroxide (99.99%) prepared from NaOH pellets purchased from Aldrich. Phenyldimethylsilane (PheDMS) and C18TMS were purchased from United Chemical Technologies, Inc. (Bristol, PA). Equipment. A Bio-Rad BioFocus 3000 capillary electrophoresis system (Bio-Rad Laboratories, Hercules, CA) equipped with a programmable, multiwavelength UV/visible detector, was used for all monolithic CEC separations. CEC data were collected and processed using the BioFocus 3000 operating software system, version 6.00. A Microcentaur model APO 5760 centrifuge (Accurate Chemical and Scientific Corp., Westbury, NY) was utilized for any necessary centrifugation. A Fisher model G-560 Vortex Genie 2 system (Fisher Scientific) was used for thorough mixing of sol solutions prior to further usage. A Chemcadet model 598450 pH meter (Cole-Palmer Instrument Co., Chicago, IL) equipped with a TRIS-specific pH electrode (Sigma-Aldrich, St. Louis, MO) was used to measure the pH of the running background electrolyte solutions. Preparation of the Sol-Gel Solution. In this study, the solgel solutions were prepared by mixing 100 µL of TMOS with 100 µL of C18-TMS, 10 µL of PheDMS, 100 µL of 99% trifluoroacetic acid (TFA), and 100 µL of 90% TFA in a microvial. This mixture was thoroughly vortexed for 5 min, and the precipitate was then separated from the sol-gel solution through centrifugation at 13 000 rpm (15682g) for 5 min. The supernatant was decanted into another microvial and used for the creation of the monolithic separation bed. Preparation of Monolithic CEC Columns. The preparation of CEC columns containing a porous monolithic matrix required some additional efforts as the column must possess one section of continuous monolithic matrix and an undisturbed end section needed for the creation of an optical window to allow for on-column UV detection. To accomplish this, preparation of the monolithic CEC columns involved the following sequential steps: (1) pretreatment of the inner fused-silica surface, (2) sealing of the distal capillary end, (3) preparation of the sol-gel solution, (4) filling a certain length of the capillary with the sol solution, (5) sealing the proximal capillary end, (6) thermal conditioning, and (7) rinsing with a series of solvents prior to rinsing with the desired running mobile phase. A detailed schematic representing each

Figure 1. Preparation of a sol-gel mediated monolithic CEC capillary: (A) 60 cm of hydrothermally pretreated 50-µm-i.d. fusedsilica capillary; (B) closure of the distal capillary end via an oxyacetylene torch; (C) filling 50 cm of the capillary with the sol-gel solution using 100 psi helium pressurization; (D) closure of the proximal capillary end via a 60-s epoxy glue, followed by thermal conditioning; (E) opening of both capillary ends via an alumina wafer following the completion of thermal conditioning; (F) preparation of an UV detection window just adjacent to the termination of the monolithic matrix.

sequential fabrication procedure is depicted in Figure 1. Pretreatment of the Inner Fused-Silica Surface. Prior to filling the capillary with the sol solution, its inner surface was first treated with deionized water. For this, an ∼5-m section of 50-µm-i.d. fusedsilica capillary was rinsed with deionized water for ∼15 min under a helium pressure of 200 psi. The capillary was then emptied by expelling the water from within by using the same helium pressure. Both ends of the capillary were then fused using an oxyacetylene torch, and the capillary was placed in a GC oven for thermal conditioning by raising the temperature at 0.5 °C/min from 40 °C (1-min hold time) to a final temperature of 250 °C with a final hold time of 60 min. The column was then removed from the GC oven, and the ends were opened, followed by purging of the column with helium under 200 psi pressure for an additional 30 min. Next, a 60-cm section of the hydrothermally pretreated fusedsilica capillary was taken (Figure 1A), and the distal end was sealed using an oxyacetylene torch as depicted in Figure 1B. The proximal capillary end was then installed into the capillary filling/ coating chamber,49 containing a polyethylene microcentrifuge vial with the desired sol solution. Using 100 psi helium pressure, the sol solution was pushed into the column, leaving ∼10 cm of undisturbed section at the sealed distal capillary end (Figure 1C). This undisturbed region, void of any sol solution, will subsequently allow for the creation of an UV detector window thereby facilitating on-column UV detection. The column, containing the sol solution, was then allowed to remain installed in the pressurized capillary chamber and left undisturbed for ∼4 h until gelation of the sol solution was visually apparent. Following this, the pressure was slowly released and the column was removed from the capillary filling/purging chamber. It was then affixed perpendicular to the benchtop. A 60-s epoxy was then applied to the proximal end of the capillary (Figure 1D) to ensure adequate sealing prior to its thermal conditioning. Next, with both capillary ends sealed, a very slow thermal conditioning program was used. This thermal conditioning consisted of a programmed temperature heating at (49) Hayes, J. D.; Malik, A. J. Chromatogr., B 1997, 695, 3.

0.2 °C/min from 35 °C (1-min hold time) to a final temperature of 150 °C, where the column was held for 120 min. Following heating, both capillary ends were opened (Figure 1E) using an alumina wafer, and a UV detector window was prepared on the empty capillary portion, immediately adjacent to the termination of the monolithic matrix (Figure 1F). The monolithic capillary was then installed into a Bio-Rad capillary cartridge and inserted into the Bio-Rad CE system for subsequent rinsing at 100 psi. The monolithic column was initially rinsed with 100% HPLC grade acetonitrile, followed by a 50:50 acetonitrile/deionized water solution for periods of 5 min each, and finally the desired running mobile phase for 15 min prior to conducting column evaluation and/or analysis. Prior to installation into the CE systems, the inlet column end was carefully examined, and ∼2 cm of the polyamide coating was carefully removed using an alumina wafer. Preparation of the Running Mobile Phases. Each mobile phase was prepared by mixing the desired volumes of acetonitrile with the Tris-HCl background electrolyte solution. The organic solvent and the background electrolyte were thoroughly degassed individually via simultaneous ultrasonication and helium purging for ∼1 h prior to mixing and usage. Thorough degassing of the mobile phase was necessary to prevent subsequent bubble formation/generation during usage. This initial degassing procedure allowed for electrochromatographic experiments to be continuously performed without pressurization of the mobile phase. To achieve the desired concentration of aqueous electrolyte, a 50 mM solution was initially prepared followed by dilution to achieve the 5 mM concentration. The pH of this 5 mM solution was then measured and adjusted to ∼2.3 by using concentrated HCl. This 5 mM Tris-HCl, pH ∼2.3, solution, in conjunction with 100% acetonitrile was individually degassed by simultaneous ultrasonication and helium purging, followed by mixing the solution in appropriate volume ratios (e.g., 75% acetonitrile/25% 5 mM Tris-HCl, etc.) to prepare the running mobile phase. SEM Characterization of the Sol-Gel ODS Monolithic Separation Bed. Scanning electron microscopy (SEM) was used for the characterization of monolithic matrixes in the prepared sol-gel CEC columns. All scanning electron micrographic images were obtained using a JEOL JSM-840 scanning electron microscope, operated at 15 kV and a filament current of 60 mA. The samples were acquired from sections of the monolithic column initially cut into equal lengths, ∼2.5 mm, and positioned perpendicularly within a retractable aluminum stage using a double-sided tape. These samples were then used to obtain cross-sectional views of the monolithic CEC columns. Longitudinal sections were acquired by dissecting ∼1.0-cm sections of capillary at ∼45°, thus yielding a capillary segment revealing a protruding portion of the monolithic matrix without the top portion of the fused silica present. These sections were then mounted parallel on an aluminum stage with the aid of double-sided carbon tape. Both stages, with all mounted capillary segments, were then consecutively placed into a Balzers SCD 050 sputter coating chamber and coated with a gold/palladium alloy at 40 mA for 60 s to avert subsequent charging. RESULTS AND DISCUSSION The Sol-Gel Chemistry Involved in the Creation of the Surface-Bonded Monolithic Separation Beds. Sol-gel chemAnalytical Chemistry, Vol. 72, No. 17, September 1, 2000

4093

Table 1. Illustration of the Names and Structures of All Sol-Gel Reagents Used in the Fabrication of Monolithic-CEC Columns

Scheme 1. Complete Hydrolysis of N-Octadecyldimethyl[3-(trimethoxysilyl)propyl]ammonium Chloride (C18-TMS) (a) and Tetramethoxysilane (TMOS) (b)

istry provides a versatile pathway for the in situ creation of chromatographic stationary phases with surface-bonded ligands, in both open tubular and monolithic formats. In this approach, unlike conventional techniques, various column preparation processes (e.g., deactivation, coating/packing, stationary-phase immobilization, end-frit making, etc.) are advantageously carried out in one single step, thus reducing the time and labor associated with column fabrication. Previously, we have reported the application of sol-gel chemistry for the preparation of Ucon-coated capillaries for CE.49 However, this is a quite general approach to column preparation, which is not limited to liquid-phase separations as evidenced by our recent reports of sol-gel technology for the preparation of high-performance GC columns and SPME fibers.50-52 Key ingredients used for the preparation of monolithic columns are listed in Table 1. In the presented sol-gel approach to monolithic CEC column technology, C18-TMS and TMOS were used as the two sol-gel-active precursors. As depicted in Table 1, the chemical structure of C18-TMS is unique and specifically suitable for this application. The structural design of this precursor contains three important features: (1) the octadecyl moiety capable of providing chromatographic interactions with the analytes, (2) three methoxy groups attached to the silicon atom that can undergo hydrolysis, followed by condensation, thereby facilitating the in situ creation of a chemically bonded monolithic matrix throughout the entire solution-filled inner capillary volume, and (3) the positively charged quaternary ammonium moiety that 4094

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

can provide a positive surface charge within the matrix to support the essential EOF in CEC. Previously Yang and El Rassi reported the use of this reagent to prepare a coated substrate allowing for the adsorption of cationic surfactants of 50-µm-i.d. fused-silica capillaries for capillary zone electrophresis53 and 4.6-mm-i.d. stainless steel high-performance micellar liquid chromatography54 columns. Sol-gel chemistry provides a unique, yet simple mechanism for the fabrication of CEC monolithic columns. One of the key sol-gel reactions consists of the hydrolysis of the precursors, TMOS and C18-TMS, involving the nucleophilic attack of water molecules on the silicon atom resulting in the replacement of the methoxy substituents with hydroxy moieties.55 The hydrolysis of C18-TMS and TMOS are illustrated in Scheme 1. As the sol-gel reactions proceed, the products of hydrolysis may then undergo polycondensation reactions in a variety of ways: (a) between hydrolyzed products of the same original precursor, (b) between hydrolyzed products of two different original precursors, and (c) between the hydrolyzed products of (50) Wang, D.; Chong, S.; Malik, A. Anal. Chem. 1997, 69, 4566. (51) Chong, S.; Wang, D.; Hayes, J. D.; Wilhite, B. W.; Malik, A. Anal. Chem. 1997, 69, 3889. (52) Malik, A.; Chong, S. Applications of Solid-Phase Microextraction; Pawliszyn, J., Ed.; Royal Society of Chemistry: London, 1999; Chapter 6, pp 73-91. (53) Yang, C.; El Rassi, Z. Electrophoresis 1998, 19, 2278. (54) Zhang, Y.; El Rassi, Z. J. Liq. Chromatogr. 1995, 18, 3373. (55) Brinker, C. J.; Scherer, G. W. Sol-gel Science: The Physics and Chemistry of Sol-gel Processing; Academic Press: San Diego, CA, 1990.

Scheme 2. Condensation of Tetrahydroxysilane with N-Octadecyldimethyl[3-(trihydroxysilyl)propyl]ammonium Chloride

Scheme 3. Condensation of the Fused-Silica Surface with the Growing Sol-Gel Network Containing a Chemically Bonded Residue of N-Octadecyldimethyl[3-(trihydroxysilyl)Propyl] Ammonium Chloride

N

either precursor with the silanol groups on the inner capillary surface. A simplified representation of a polycondensation reaction between the hydrolysis products of both precursors is depicted in Scheme 2. This growing three-dimensional polymeric network will then eventually become anchored to the inner capillary surface through chemical bonding with the silanol moieties residing along the inner fused-silica capillary surface. Scheme 3 represents a simplified representation of the condensation reaction occurring between the inner fused-silica surface with products of the polycondensation reaction depicted in Scheme 2. Finally, the incorporation of PheDMS into the sol solution serves as a deactivating reagent for the monolithic bed. This deactivation reagent is initially added to the sol solution. The mobile hydrogen atom in the structure of this reagent is reactive toward silanol groups, especially at elevated temperatures. It can be assumed that during the sol-gel process this reagent gets physically incorporated in the monolithic structure but subsequently, during thermal treatment of the column, reacts with the residual silanol groups in the monolithic structure providing deactivation. This deactivation process is depicted in Scheme 4.

SEM Characterization of the Surface-Bonded Sol-Gel Monolithic ODS Separation Beds. Visualization of the monolithic microstructure within the capillary was accomplished through scanning electron microscopic investigations conducted on several capillary segments. Figure 2, represents an SEM crosssectional view of a sol-gel monolithic column at a magnification of 1800×. Observations at this magnification reveal that the entire cross section of the capillary contains the monolithic matrix. Figure 3, a longitudinal view of the monolithic capillary at 7000× magnification, reveals the porous structure of the monolithic matrix. From this view, it is evident that the pore structure is well developed in the monolithic bed with pore diameters of ∼1.5 µm. The use of higher C18-TMS-to-TMOS molar ratios in the sol solution provided monolithic beds with the said pore characteristics. This also allowed for enhanced permeability of the mobile phase. For example, an increase in the C18-TMS-to-TMOS molar ratio of from 0.5 to 0.75 yielded flow rates of up to ∼7.75 µL/min for the mobile phase consisting of 80% (v/v) ACN 20% (v/v)/5 mM Tris-HCl. Dimethyl sulfoxide (DMSO) was used as the neutral Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4095

Scheme 4. Deactivation of the Sol-Gel ODS Monolith with Phenyldimethylsilane

EOF marker to determine the linear velocity of the mobile phase and was found to be 0.97 mm/s using the previously described mobile phase. Additionally, Figure 4, a cross-sectional view of the inner capillary surface at higher magnification (15000×), reveals the chemical bonding during the column preparation process that occurred due to the condensation between the sol-gel network structure and the silanol moieties on the inner capillary walls. The sol-gel process provides a chemical anchorage of the monolithic matrix to the inner walls of the capillary. This illustrates another significant attribute of a sol-gel monolithic columnsthe possibility of fritless column operation. Electroosmotic Flow in Sol-Gel Monolithic ODS Columns. . In CEC, a consistent EOF is essential to drive the analyte(s) through the separation column. This EOF is generated due to an electrical double layer at the interface of the solid support with the liquid mobile phase. Most commonly, silica is used as the solid support and develops a negative surface charge under CE/CEC running conditions, presumably as a result of the deprotonation of the silanol groups. This negatively charged substrate attracts cations from the electrolyte in the mobile phase thereby forming the electrical double layer.In this study, the positively charged quaternary ammonium moiety contained in the N-octadecyldimethyl[3-(trimethoxysilyl)propyl] ammonium chloride provided a positively charged surface on the monolithic 4096

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

matrix, which, in turn, counteracted the effects of the residual silanol groups residing both on the monolith and on the inner capillary surface. Under the experimental conditions used, a strong EOF was observed in the reversed direction (from cathode to anode), suggesting that the surface positive charge due to quaternary ammonium functionality in the surface-bonded C18TMS moieties is the EOF-determining factor in the prepared solgel monolithic columns. Following fabrication, several experiments were conducted for the investigation of EOF in sol-gel monolithic columns. The first measurements obtained using the monolithic ODS capillary was an evaluation of the effect of acetonitrile percentage in the running mobile phase on the electroosmotic mobility. For this, a set of mobile phases containing varying percentages of acetonitrile and 5 mM aqueous Tris-HCl was utilized. In addition, DMSO was used as the neutral electroosmotic flow marker. The results obtained from these experiments are depicted in Figure 5. As illustrated, the electroosmotic mobility within the ODS monolithic capillary consistently increased with the increase of acetonitrile content in the mobile phase. Such an increase in EOF is indicative of an increase in the net positive surface charge within the monolithic columns. One possibility for this to occur is the reduction of effective negative surface charge due to an increase in acetonitrile concentration in the mobile phase, resulting in an equivalent

Figure 2. Scanning electron micrograph of a sol-gel monolith column (cross-sectional view). Magnification, 1800×.

Figure 3. Scanning electron micrograph of a sol-gel monolithic column (longitudinal view). Magnification, 7000×.

increase of the effective positive surface charge due to the quaternary ammonium groups. This might be possible if the formation of negative charge on the monolith/capillary is reduced due to the interaction of acetonitrile with the negative-chargegenerating surface groups (e.g., silanols). Chromatographic Characterization of Sol-Gel ODS Monolithic CEC Columns. Figure 6 represents the CEC separation of a mixture of PAHs on a 50 cm × 50 µm i.d. sol-gel monolithic ODS column that allowed for the use of a mobile phase containing a higher percentage of acetonitrile (up to 80%) and simultaneously rendered sufficient solute-stationary-phase interactions. The separation efficiencies acquired for naphthalene in the mixture of PAH analytes in this analysis were on the order of 145 800 total theoretical plates per meter (73 000 plates/column). This efficiency value is comparable or superior to the total plate numbers per column reported in the literature for retained solutes. For example, Yan et al.18 acquired 127 000 plates/m (41 910 plates/ column) for a mixture of PAHs using 75-µm-i.d. fused-silica capillaries packed with 33 cm of 3-µm octadecylsilica particles. Likewise, Fre´chet and co-workers reported separation efficiencies of 120 000 plates/m (42 900 plates/column) in their series of

Figure 4. Scanning electron micrograph of a sol-gel monolithic column (longitudinal view). Magnification, 15000×.

Figure 5. Electroosmotic mobility of a sol-gel ODS monolithic column as a function of acetonitrile percentage. Separation column: 50 cm × 50 µm i.d. (46.1 cm effective length). Separation conditions: injection, 12 kV for 3 s; run, -15 kV 2.28 µA; DMSO was used as the EOF marker. Background electrolyte: 5 mM Tris-HCl (pH 2.34).

publications using methacrylate monomers for the preparation of CEC35-37 monolithic capillaries through copolymerization. The use of sol-gel chemistry for the fabrication of monolithic CEC columns by gluing ODS HPLC packing particles has yielded somewhat similar results. For instance, Dulay et al.48 acquired efficiencies of 80 000 plates/m (16 000 plates/column) for a test mixture of six PAHs on sol-gel capillary columns containing the ODS imbedded particles. Lee and co-workers also used sol-gel chemistry to bond 7-µm ODS particles, thereby creating a continuous large-pore CEC capillary column, and reported an efficiency of 2.20 × 105 plates/m (70 400 plates/column) for the unretained thiourea.23 Considering the fact that monolithic columns with overall lengths of up to several meters can be easily prepared by the presented sol-gel technology and that the prepared columns can be operated using commercially available CE instrumentation, this may open new possibilities for generating extremely high efficiencies per column in CEC separations. Van Deemter plots, as depicted in Figure 7, were constructed through variations in the operating voltages (-3 to -19 kV), Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4097

Figure 6. CEC analysis of a mixture of PAHs on a sol-gel ODS monolithic column. Separation column: 50 cm × 50 µm i.d. (46.1cm effective length). Separation conditions: injection, -12 kV 0.03 min; run, -15 kV 2.68 µA. Mobile phase: 80% ACN/20% 5mM TrisHCl, pH 2.34. DMSO was used as EOF marker. Analytes: (1) benzene, 4.4053 × 10-6 M; (2) naphthalene, 2.7087 × 10-6 M; (3) impurity; (4) fluorene, 1.5433 × 10-6 M; (5) phenanthrene, 1.5748 × 10-6 M; (6) anthracene, 9.6850 × 10-7 M; (7) fluoranthene, 1.1654 × 10-6 M; (8) pyrene, 1.2283 × 10-6 M; (10) benzo[a]pyrene, 1.5118 × 10-6 M.

Figure 7. Plate height vs flow rate within a sol-gel mediated ODS monolithic capillary. Separation column: 50 cm × 50 µm i.d. (46.1cm effective length). Separation conditions: injection, -12 kV 3 s; run, -3 to -19 kV. Mobile phase: 75% acetonitrile/25% 5 mM TrisHCl, pH 2.34. Test solutes: naphthalene and anthracene. EOF marker: DMSO.

thereby altering the mobile-phase flow rate through the column and measuring the achieved plate heights corresponding to each operating voltage. For the used test solutes, the van Deemter plots reveal minimal increases in plate heights as the mobile-phase flow rates are enhanced. The relatively flat right-hand portion of the H vs u j curves indicate an efficient mass-transfer process between 4098 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

Figure 8. Separation analysis of a mixture of benzene derivatives on a sol-gel mediated ODS monolithic column. Separation column: 50 cm × 50 µm i.d. (46.1-cm effective length). Separation conditions: injection, -12 kV 0.02 min; run, -15 kV 5.20 µA. Mobile phase: 75% ACN/25% 5 mM Tris-HCl, pH 2.34. Analytes: (1) thiourea, 1.180 × 10-3 M; (2) benzaldehyde, 1.9655 × 10-3 M; (3) benzene, 3.2680 × 10-4 M; (4) toluene, 1.5263 × 10-4 M; (5) ethylbenzene, 3.0618 × 10-4 M; (6) propylbenzene, 2.986 × 10-4 M; (7) biphenyl, 1.9655 × 10-3 M; (8) butylbenzene, 1.9655 × 10-3 M; (9) amylbenzene, 1.9655 × 10-3 M.

the mobile phase and the monolithic ODS separation bed. As can be seen in Figure 7, the optimum linear velocity for the used sol-gel monolithic ODS column was 0.75 mm/s, which corresponds to an applied field strength of -240 V cm-1 (-12 kV) in the sol-gel monolithic columns. This opens new possibilities of using longer sol-gel columns producing higher overall column efficiencies without exceeding the upper voltage limits of commercially available CE instruments. Furthermore, the use of sol-gel technology to prepare these monolithic ODS columns for CEC is further accentuated as increased column lengths can be used because the highly porous structure of the monoliths allows for their rinsing and CEC operation using commercially available CE instrumentation without any additional pressurization capability. A test mixture of benzene derivatives was also used to further evaluate the separation performance of the sol-gel ODS monolithic columns using a mobile phase containing 75% acetonitrile and 25% aqueous 5 mM Tris-HCl (pH 2.34). Column efficiencies on the order of 163 200 plates/m (81 600 plates/column) were obtained in these analyses. A representative example of this separation is depicted in Figure 8. Figure 9 illustrates an analogous separation of a mixture of aldehydes and ketones obtained on a sol-gel monolithic ODS column. As in the case with the benzene derivatives, this probe

Table 2. Experimental Results on Efficiency, Retention Characteristics, Selectivity, and Retention Time Repeatability for Sol-Gel Monolithic Columns in CECa analyte

separation efficiency, N (plates/column)

tR (min)

retention factor, k

separation factor, R

s

% RSD (n ) 5)

benzaldehyde tolualdehyde butyrophenone valerophenone hexaphenone heptaphenone

89 778 91 039 83 867 79 353 86 027 89 687

9.144 9.526 10.048 10.678 11.550 12.788

0.050 0.094 0.154 0.227 0.327 0.469

1.88 1.64 1.47 1.44 1.43

0.027 0.029 0.025 0.016 0.022 0.031

0.295 0.302 0.248 0.154 0.194 0.244

a Separation column, 50 cm × 50 µm i.d. (46.1-cm effective length). Separation conditions: injection, -12 kV, 3 s; run, -15 kV. Mobile phase: 70% acetonitrile/30% 5 mM Tris-HCl, pH 2.34.

CEC. Results on the sol-gel column technology for OT-CEC will be published elsewhere.56 CONCLUSION

Figure 9. Separation analysis of a mixture of aldehydes and ketones on a sol-gel mediated ODS monolithic column. Separation column: 50 cm × 50 µm i.d. (46.1-cm effective length). Separation conditions: injection, -12 kV 0.03 min; run, -25 kV 0.5 µA. Mobile phase: 70% ACN/30% 5 mM Tris-HCl, pH 2.34. Analytes: (1) benzaldehyde, 1.180 × 10-3 M; (2) o-tolualdehyde, 1.9655 × 10-3 M; (3) butyrophenone, 3.2680 × 10-4 M; (4) valerophenone, 1.5263 × 10-4 M; (5) hexaphenone, 3.0618 × 10-4 M; (6) heptaphenone, 2.986 × 10-4 M.

mixture contained more closely related analytes. Column efficiencies on the order of 174 600 plates/m (87 300 plates/column) were obtained in these analyses. Repeatability studies were performed using various analyte mixtures. These experiments were essential to evaluate the consistency in solute retention on the sol-gel monolithic ODS columns. Table 2 presents CEC characteristics of sol-gel monolithic columns and experimental data on retention time repeatability for a test mixture of seven aromatic aldehydes and ketones. As depicted in this table, consistent repeatability values are exemplified by the low RSD (0.15-0.30%) values for solute retention times in a series of five consecutive runs. The developed sol-gel approach using C18-TMS as a precursor was also applied to prepare an effective coating for open tubular

In this study, a novel sol-gel approach was developed to create an organic-inorganic porous monolithic substrate inside fusedsilica capillaries for use as a separation bed in CEC. The incorporation of C18-TMS as a precursor in the sol solution was critical. Structural attributes of this precursor allowed for the creation of a porous monolithic bed exhibiting a positively charged surface and possessing chemically bonded ODS ligands essential for chromatographic interactions with the analytes. The positively charged surface was responsible for the reversed electroosmotic flow in the monolithic columns. The optimum percentages of solgel reactants allowed for the fabrication of monolithic capillaries with highly porous structure, yet possessing sufficient stationary phase for chromatographic interactions. Enhanced permeability of the sol-gel monolithic columns allowed for their rinsing/ purging within the CE system without requiring high-pressure pumping devices. The elimination of the need for retaining frits is an additional attribute of the sol-gel monolithic columns. The use of sol-gel monolithic columns with extended lengths is not only predictable but realistic as the columns are characterized with enhanced EOF. The high number of theoretical plates per column and other chromatographic data obtained on the sol-gel monolithic columns offer significant promise to further development in this area of CEC column technology. ACKNOWLEDGMENT This work was funded in part by a grant from Dow Chemical Co. under the University Development Fund (UDF) Program and by financial assistance from the I-4 corridor High-Tech Initiative Program at the University of South Florida. The authors gratefully acknowledge and appreciate Mrs. Jane Lundh, U.S.F. College of Engineering, for her assistance in the scanning electron microscopic investigations of the capillary segments.

Received for review February 3, 2000. Accepted June 19, 2000. AC000120P (56) Hayes, J. D.; Malik, A. Submitted to Anal. Chem.

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4099