Performance of a Monolithic Silica Column in a Capillary under

Feb 10, 2000 - Facile Preparation of SiO2/TiO2 Composite Monolithic Capillary Column and Its Application in Enrichment of Phosphopeptides. Shao-Ting W...
8 downloads 16 Views 150KB Size
Download PDF
Anal. Chem. 2000, 72, 1275-1280

Performance of a Monolithic Silica Column in a Capillary under Pressure-Driven and Electrodriven Conditions Norio Ishizuka,† Hiroyoshi Minakuchi,† Kazuki Nakanishi,† Naohiro Soga,† Hisashi Nagayama,‡ Ken Hosoya,‡ and Nobuo Tanaka*,‡

Department of Material Chemistry, Graduate School of Engineering, Kyoto University, Yoshida, Sakyo-ku, Kyoto 606-8501, Japan, and Department of Polymer Science and Engineering, Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan

A continuous macroporous silica gel network was prepared in a fused-silica capillary and evaluated in reversedphase liquid chromatography. Under pressure-driven conditions, the monolithic silica column derivatized to C18 phase (100 µm in diameter, 25 cm in length, silica skeleton size of ∼2.2 µm) produced plate heights of about 23 and 81 µm at 0.5 mm/s with a pressure drop of 0.4 kg/cm2, and at 4.0 mm/s with 3.6 kg/cm2, respectively, in 90% acetonitrile for hexylbenzene with a k value of 0.7. The separation impedance, E, calculated for the present monolithic silica column was much smaller at a low flow rate than those for particle-packed columns, although higher E values were obtained at a higher flow rate. Considerable dependence of column efficiency on the linear velocity of the mobile phase was observed despite the small size of the silica skeletons. A major source of band broadening in the HPLC mode was found in the A term of the van Deemter equation. The performance of the continuous silica capillary column in the electrodriven mode was much better than that in the pressure-driven mode. Plate heights of 7-8 µm were obtained for alkylbenzenes at 0.7-1.3 mm/s, although the electroosmotic flow was slow. In HPLC and CEC mode, the dependency of plate height on k values of the solutes was observed as seen in open tube chromatography presumably due to the contribution of the large through-pores. Since monolithic silica capillary columns can provide high permeability, the pressure-driven operation at a very low pressure can afford a separation speed similar to CEC at a high electric field. HPLC columns packed with porous spherical particles of 5 µm have been most widely used since the late 1970s.1 Although the reduction of particle size can lead to better column efficiency on the basis of smaller eddy diffusion and shorter diffusion path * Corresponding author: (phone) 81-75-724-7809; (fax) 81-75-724-7710; (email) [email protected]. † Kyoto University. ‡ Kyoto Institute of Technology. (1) Guiochon, G. In High-performance Liquid Chromatography: Advances and Perspectives; Horvath, C., Ed.; Academic Press: New York, 1980; Vol. 2, pp 1-56. 10.1021/ac990942q CCC: $19.00 Published on Web 02/10/2000

© 2000 American Chemical Society

length, the limitation in the performance of a packed column is well recognized based on the pressure limit of a solvent delivery system,2 leading to the current compromise at about 4-5 µm particle size between the column efficiency and the pressure drop. Many researchers have been trying to circumvent this limitation and to get much higher efficiency. Such an attempt in liquid chromatography includes ultrahigh-pressure LC,3,4 capillary electrochromatography (CEC),5 and open tube chromatography.6 None of these approaches, however, has been widely accepted as a practical separation method yet, because of the difficulty in instrumentation or in execution. Another possible approach to overcome the problem of high pressure associated with small particles is to fabricate a column made of one piece of a porous solid with small-sized skeletons and relatively large through-pores which could provide both low pressure drop and high column efficiency. The advantages of such a column are well understood. Several examples of such monolithic columns made of an organic polymer prepared by in situ polymerization in a chromatographic column have been reported recently and proven to be effective in the high-speed separation of polypeptides.7-14 Polymer-based packing materials, however, usually possess micropores resulting in a decrease in efficiency for small molecules,15,16 although column efficiencies of up to 120 000 plates/m have been reported in HPLC and CEC.11 Regnier (2) Poppe, H. J. Chromatogr., A 1997, 778, 3-21. (3) MacNair, J. E.; Lewis, K. C.; Jorgenson, J. W. Anal. Chem. 1997, 69, 983989. (4) MacNair, J. E.; Patel, K. D.; Jorgenson, J. W. Anal. Chem. 1999, 71, 700708. (5) Dittmann, M. M.; Rozing, G. P. J. Chromatogr., A 1996, 744, 63-74. (6) Tock, P. P. H.; Boshoven, C.; Poppe, H.; Kraak, J. C. J. Chromatogr. 1989, 477, 95-106. (7) Liao, J. L.; Hjerten, S. J. Chromatogr. 1988, 457, 165-174. (8) Hjerten, S.; Liao, J. L.; Zhang, R. J. Chromatogr. 1989, 473, 273-275. (9) Svec, F.; Frechet, J. M. Anal. Chem. 1992, 64, 820-822. (10) Ching, Q. C.; Svec, F.; Frechet, J. M. J. Anal. Chem. 1993, 65, 2243-2248. (11) Ericson, C.; Liao, J.-L.; Nakazato, K.; Hjerten, S. J. Chromatogr., A 1997, 767, 33-41. (12) Fujimoto, C.; Kino, J.; Sawada, H. J. Chromatogr., A 1995, 716, 107-113. (13) Wang, Q. C.; Svec, F.; Frechet, J. M. J. J. Chromatogr., A 1994, 669, 230235. (14) Peters, E. C.; Petro, M.; Svec, F.; Frechet, J. M. J. Anal. Chem. 1997, 69, 3646-3649. (15) Tanaka, N.; Ebata, T.; Hashizume, K.; Hosoya, K.; Araki, M. J. Chromatogr. 1989, 475, 195-208. (16) Nevejans, F.; Verzele, M. J. Chromatogr. 1987, 406, 325-342.

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000 1275

recently reported the fabrication of nanocolumns that provided high efficiency of up to 1 670 000 plates/m.17 Fields reported that continuous silica xerogels prepared from potassium silicate solutions could be used as highly permeable support media and exhibit reasonable chromatographic efficiency, although the morphology presented was not optimal for HPLC.18 Dulay et al. prepared a sol-gel matrix in the presence of ODS particles and reported efficiencies of up to 80 000 plates/m in CEC.19 Asiaie et al. obtained a porous silica-based monolithic column by agglomerating (sintering) the particles packed in a conventional capillary column.20 These approaches, however, cannot escape from the limitation associated with a particle-packed column. Minakuchi et al. reported the preparation and evaluation of a continuous porous silica column (silica rods) that provided much higher column efficiency for both small and large molecules at a higher velocity of mobile phase than conventional columns packed with particles.21-25 The preparation of a continuous monolithic silica in a capillary will be advantageous in that small-sized and long monolithic silica can be prepared without cladding the monolith with engineering plastics such as poly(ether ether ketone) (PEEK). Thus, we can avoid the most difficult step in rod column preparation, although the preparations in a capillary do have some difficulties. Voids could develop between the silica structure and the capillary wall due to shrinkage of the silica during preparation.26 In the case of the preparation of 7 mm silica rods, a mold of 9 mm diameter was used.21-25 This problem of shrinkage associated with polymerization must be solved practically by attaching the silica structure to the capillary wall and by minimizing the shrinkage. We report here the preparation of continuous porous monolithic silica in a fused-silica capillary and their evaluation under pressure-driven and electrodriven conditions. EXPERIMENTAL SECTION A monolithic silica was prepared by in situ hydrolysis and polycondensation of alkoxysilane in a fused-silica capillary of 100 µm in diameter and 33.5 cm in length by a method similar to the one described previously.21-31 Tetramethoxysilane (4 mL) was added to the solution of poly(ethylene oxide) (1.06 g, Mw ) 10 000, Aldrich) in 0.01 M acetic acid (10 mL), and the mixture was stirred (17) He, B.; Tait, N.; Regnier, F. Anal. Chem. 1998, 70, 3790-3797. (18) Fields, S. M. Anal. Chem. 1996, 68, 2709-2712. (19) Dulay, M. T.; Kulkarni, R. P.; Zare, R. N. Anal. Chem. 1998, 70, 51035107. (20) Asiaie, R.; Huang, X.; Farnan, D.; Horvath, C. J. Chromatogr., A 1998, 806, 251-263. (21) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. Anal. Chem. 1996, 68, 3498-3501. (22) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. J. Chromatogr., A 1997, 762, 135-146. (23) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. J. Chromatogr., A 1998, 797, 121-131. (24) Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Soga, N.; Tanaka, N. J. Chromatogr., A 1998, 797, 133-137. (25) Minakuchi, H.; Ishizuka, N.; Nakanishi, K.; Soga, N.; Tanaka, N. J. Chromatogr., A 1998, 828, 83-90. (26) Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Soga, N.; Hosoya, K.; Tanaka, N. J. High. Resolut. Chromatogr. 1998, 21, 477-479. (27) Nakanishi, K.; Minakuchi, H.; Soga, N.; Tanaka, N. J. Sol-Gel Sci. Technol. 1997, 8, 547-552. (28) Nakanishi, K. J. Porous Mater. 1997, 4, 67-112. (29) Nakanishi, K.; Soga, N. J. Am. Ceram. Soc. 1991, 74, 2518-2530. (30) Nakanishi, K.; Sagawa, Y.; Soga, N. J. Non-Cryst. Solids 1991, 134, 39-46. (31) Nakanishi, K.; Soga, N. J. Non-Cryst. Solids 1992, 139, 1-13 and 14-24.

1276

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

at 0 °C for 45 min. The resulting mixture was forced into a fusedsilica capillary tube (100 µm i.d., 370 µmo.d., Polymicro) and allowed to react overnight at 40 °C. The tube had been treated with 1 M NaOH solution at 40 °C for 3 h prior to the silica preparation. The monolithic silica thus formed was washed with water and then treated with aqueous ammonium hydroxide solution (0.01 M) at 120 °C for 3 h followed by a wash with ethanol. The final heat treatment after drying was carried out at 330 °C for 24h, followed by reaction with octadecyldimethyl-N,N-diethylaminosilane in toluene at 50 °C for 2 h.26 Usually two 100 cm long monolithic columns were prepared from the same reaction mixture. After preparation, each end (10-15 cm) of the capillary having large voids was cut off, and two to four 33.5 cm long columns were obtained from the two 100 cm silica capillaries containing silica monolith. The capillary columns showed 10 00012 000 theoretical plates for the effective length of 25 cm under optimized conditions in a pressure-driven mode. The structural morphology of the monolithic silica in capillary was examined at a fractured surface by scanning electron microscopy (SEM) (S-510, Hitachi). A detection window was made at 25 cm from the inlet by removing the polyimide coating of the fused-silica capillary. Thus, detection was carried out on-column throughout this experiment at 200 or 254 nm. Chromatographic experiments were carried out by using a split injection HPLC system consisting of a pump (LC10A, Shimadzu), a UV detector (CE 971UV, Jasco), a data processor (C-R6A, Shimadzu), and an injection valve (model 7125, Rheodyne) fitted with a T-joint, one end connected to the monolithic silica capillary and the other end to a stainless steel column (4.6 mm i.d., 10 cm long) packed with ODS-silica particles (15 µm). The system was operated in a constant-pressure mode at ambient temperatures (20-25 °C). Capillary electrophoresis equipment (HP-3DCE, Hewlett-Packard) was used for the CEC experiment. Samples were injected at 0.04 kg/cm2 with nitrogen pressure. The CEC separations were performed at 20 °C at an applied voltage of up to 30 kV, while both inlet and outlet vials were pressurized to 8 kg/cm2. RESULTS AND DISCUSSION A SEM photograph of the monolithic silica prepared in a 100 µm fused-silica capillary (Figure 1) shows that the silica morphology is characterized not by a smooth cylindrical structures forming a network but by an aggregated structure of globular silica found with a silica rod of very small domain sizes as reported previously.21-25 Figure 1 also shows that the present monolithic silica has no large voids along the wall of the fused-silica capillary, formed by the shrinkage of silica seen with the previous preparation.26 The uniform-sized silica skeleton of ∼2.2 µm and the through-pores of up to ∼8 µm were evenly distributed over the entire cross section of the capillary rod. The treatment of a silica tube with 1 M sodium hydroxide resulted in effective attachment of silica skeletons to the wall. Large interstitial voids were not observed in the monolithic structure of silica except at each end of the tube where the bonding of the silica to the wall could not completely prevent shrinkage. Figure 2 shows the relation between the skeleton size and the through-pore size of the monolithic silica in capillary. The results of the previous study22,23 and those with conventional columns packed with spherical silica particles are also plotted. The size of the interstitial voids in a conventional column is reported to be

Figure 1. Scanning electron micrograph of a continuous monolithic silica prepared in a fused-silica capillary. Capillary diameter, 100 µm.

Figure 2. Skeleton size against the through-pore size of the continuous monolithic silica prepared in a capillary (O) and the largersized silica rod columns (7 mm × 83 mm) having constant skeleton size/through-pore size ratio (9)23 and constant through-pore size (0)22 in a previous study. The results with column A were discussed, while good reproducibility was observed. Also plotted are the particle sizes (vertical axis) against the size of interstitial voids (25-40% of dp as indicated by the bars) found with conventional particle-packed column.

25-40% of the particle size33 as indicated by the bars in Figure 2. The through-pore size/skeleton size ratios of the silica rod columns prepared in a mold in the previous study were 1.0-1.5, while the ratios observed with the monolithic silica in this study are 3-5, approaching the upper end of the plots representing an open tube column. Such a difference in the through-pore size/ skeleton size ratios in the monolithic silica can be produced by the lack of shrinkage of the silica structure as a whole caused by the attachment to the capillary wall. The phase separation between the silica portion and the aqueous solution in the reaction mixture proceeded further without shrinkage of the whole structure than with the previous silica rods of 7 mm diameter prepared in a 9 mm i.d. mold,21-25 thus creating the skeletons with the appearance of aggregated globules and very large through-pores. The timing for phase separation and sol-gel transition can be controlled by (32) Giddings, J. C. Dynamics of Chromatography, Part 1, Principles, and Theory; Mercel Dekker: New York, 1965. (33) Unger, K. K. Porous Silica; Elsevier: Amsterdam, 1979.

Figure 3. Molecular weight-elution volume curves obtained by size exclusion chromatography for silica (closed symbols) and C18 silica (open symbols). Mobile phase, tetrahydrofuran. (b, O) monolithic silica column in a capillary; (2, 4) larger-sized silica rod column ((c) in Figure 2);22 (9, 0) column packed with 5 µm particles.22

the composition of the preparation mixtures that determines the onset of phase separation relative to the occurrence of sol-gel transition. Figure 3 shows the molecular weight-elution volume curves measured by size exclusion chromatography using polystyrene standards in tetrahydrofuran with the present capillary columns. The results for the previous silica rod column and a packed column were also plotted.22 In a previous study, the silica rod (Figure 2c) prepared in a mold was shown to possess an external porosity of up to 65% of the column volume (86% total porosity) before bonding octadecylsilyl groups. This is much higher than 39% external porosity (79% total porosity) found with a packed column. The external porosity was derived from the elution time of a polystyrene standard of 390 000 molecular weight or greater and the total porosity from the elution time of benzene. The silica monolith in capillary showed even higher porosity, 86% external porosity and 96% total porosity, although the measurement was accompanied by considerable uncertainty. The total porosity or the volume of the liquid in the monolithic silica column was estimated by weighing the column filled with water, methanol, or chloroform of ∼2 µL. A high external porosity and a large through-pore size resulted in very low flow resistance, as shown later. Figure 4 shows the comparison of chromatograms between the pressure-driven and electrodriven elution of alkylbenzenes and polyaromatic hydrocarbons (PAHs) in 80% acetonitrile at a similar linear velocity. Figure 4a shows that the 250 mm column produced ∼6000 theoretical plates for hexylbenzene having a k value of 1.5 in the pressure-driven LC with split injection at 1.1 mm/s linear velocity. The k values obtained with the present monolithic silica were much smaller than those obtained with the previous silica rods having 14 nm pores22 or with the column packed with C18 silica particles with 11-12 nm pores. The phase ratio in the monolithic silica column in capillary estimated on the basis of k values of alkylbenzenes was 0.02-0.03 compared to 0.06 for the silica rod and 0.19 for the particle-packed column in the previous study.22 Sample loading capacity of the monolithic silica is still much higher than nonporous particles or open tube systems. The Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

1277

Figure 4. Chromatograms obtained for alkylbenzenes (C6H5(CH2)nH, n ) 0-6 (a, c)) and polyaromatic hydrocarbons (b, d) in pressure-driven (HPLC (a, b)) and electrodriven (CEC (c, d)) elution. PAHs: N, naphthalene; F, fluorene; Ph, phenanthrene; A, anthracene; P, pyrene; T, triphenylene; B, benzo[a]pyrene. Column size: 100 µm i.d. × 33.5 cm (effective length 25 cm). Mobile phase: (a, b) 80% acetonitrile; (c, d) acetonitrile-Tris-HCl, 50 mM, pH 8 (80/20). Pressure: (a, b) 1.3 kg/cm2. Applied voltage,:(c, d) 900 V/cm.

Figure 5. Van Deemter plots obtained for C18 monolithic silica in capillary in HPLC (closed symbols) and CEC (open symbols) with hexylbenzene (b, O) and benzo[a]pyrene (9, 0) as solute. Mobile phase: acetonitrile-water (HPLC); acetonitrile-Tris‚HCl buffer, 50 mM pH 8 (CEC); (a) 80/20, (b) 90/10.

monolithic silica in a fused-silica capillary was applied to CEC for the first time to produce much higher performance than in the pressure-driven mode. The efficiency of ∼20 000 theoretical plates was obtained for hexylbenzene in 80% acetonitrile with the linear velocity, u ) 1.08 mm/s. The van Deemter plots obtained with hexylbenzene and benzo[a]pyrene as a solute in 80 and 90% acetonitrile are shown in Figure 5. Plate height minimum was not found above the linear velocity of 0.5 mm/s. The slopes of the plots for the monolithic silica capillary are large despite the small skeleton size. Relatively high plate height at low linear velocity and the sharp increase in plate height with the increase in linear velocity in the pressuredriven mode can be attributed to the large A term in eq 1

h ) Aν1/3 + B/ν + Cν H)

1 1 1 + Cedp (C d 2/D )u m p m

+

CdDm Csmdp2 + u u Dm

(1) (2)

and eq 2 (Dm, diffusion coefficient of a solute in the mobile phase; 1278

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

dp, particle (silica skeleton) size; Cx, coefficient for the contribution of each term).32 The results in the electrodriven mode indicate that the silica monolith in capillary can give much better column efficiency than in HPLC. Low plate height and a slight decrease in plate height with the increase in the linear velocity were observed in CEC with Hmin of ∼7 µm at 1.3 mm/s in 90% acetonitrile with an applied electric field of 30 kV across the 33.5 cm silica monolith. The maximum efficiency, 35 000 theoretical plates for hexylbenzene for the 25 cm column, is remarkable for a column that can be eluted at less than 2 kg/cm2 in a pressure-driven mode to produce similar linear velocities. Because the contributions of B and C terms are expected to be similar for a pressure-driven mode and an electrodriven mode, the difference in the plots shown in Figure 5 can be attributed to the contribution of the A terms. The contribution of the A term is known to be less significant in CEC than in a pressure-driven mode.5 The A term contribution in HPLC can be divided into the contribution of eddy diffusion and that of slow mass transfer in the mobile phase (eq 2). The former is independent of flow velocity, while the latter depends on the linear velocity. The steep slopes of the plots of the plate height against the linear velocity shown in Figure 5 indicate that the slow mobile-phase mass transfer should be responsible for the much poorer column efficiency under HPLC conditions than under CEC conditions. A silica rod of larger size (83 mm × 7 mm diameter) with a similar skeleton size and the smaller through-pore size produced much lower plate heights and much smaller slopes of the van Deemter plots at higher mobile-phase velocities.23 Thus the large A term found with the present monolithic silica column in HPLC is most likely caused by the presence of large through-pores of up to ∼8 µm that would be found with 30-40 µm particles in a conventional column. The present monolithic silica in a capillary showed better performance than the previous one,26 which had exceedingly large through-pores or channels or voids along the wall made by the shrinkage of the silica. The values of plate heights and the slope of the plots obtained with the capillary monolith are smaller in 90% acetonitrile than in 80% acetonitrile presumably due to the greater diffusion coefficients of solutes in the mobile phase as shown in Figure 5b. The performance of the monolithic silica in capillary seems to be dominated by the size of the large throughpores. In open tube capillary chromatography, the slope of the van Deemter plot partly depends on the square of dc in eq 3

H)

df2 2Dm 1 + 6k + 11k2 dc2 2k + u + u u 96(1 + k)2 Dm 3(1 + k)2 Ds

(3)

(Ds, diffusion coefficient in the stationary phase; dc, inner diameter of the capillary; df, thickness of the stationary layer).6 Figure 6 shows the plots of plate height against k values for the present monolithic silica column in HPLC and CEC. The dependency of plate height on k values is recognized in open tube capillary chromatography, to which the k value and the column diameter contribute according to eq 3.6 The results seem to show the contribution of through-pores as large as 10 µm in the monolithic silica capillary column. The effect is larger in 80% acetonitrile than that in 90% acetonitrile as expected from the

Figure 6. Plate heights against k values obtained in HPLC (closed symbols) and CEC (open symbols) with alkylbenzenes (b, O) and PAHs (9, 0) as solute. Mobile phase: see Figure 5.

difference in diffusion coefficients of solutes in the mobile phase. Such contribution is much less for a particle-packed column that possesses much smaller through-pore size/skeleton size ratios. Knox recently reported the dominant contribution of the slow diffusion in the mobile phase to band broadening in LC.34 The smaller effect of the k values on the plate height in CEC than in pressure-driven LC can be explained by the smaller difference in the velocity along the streamlines of electroosmotic flow in various parts of the through-pores due to the plug-type flow profile. A monolithic silica capillary with smaller-sized through-pores is expected to produce higher column efficiencies. The preparation procedure needs further improvement to produce smaller-sized through-pores and to prevent gel shrinkage in order to obtain a high-performance monolithic silica capillary. The effective pressure drop for the 250 mm portion of the monolithic silica in capillary is three-fourths of the total pressure drop of the column. As shown in Figure 7a, through the 250 mm column a pressure of only 3.8 kg/cm2 was needed to generate a linear velocity of 1.6 mm/s, which is roughly equivalent to the velocity obtained at the flow rate of 1 mL/min for a particle-packed column of 4.6 mm diameter. The C18 monolithic silica column in capillary, which has through-pores of ∼8 µm, and the high external porositiy produced a much lower pressure drop than the silica rod (through-pore size 3.5 µm, skeleton size 2.4 µm) having the highest porosity in the previous study.23 The separation impedance, E, calculated according to eq 4

E ) t0∆P/N2η

(4)

(t0, elution time of an unretained solute; ∆P, pressure drop; N, number of theoretical plates; η, viscosity of the mobile phase), was 700 at Hmin for the monolithic silica in capillary, as shown in Figure 7b. This is much better than those with a conventional packed column with 5 µm particles, which gives an E value of ∼3500 and shows a much greater increase with increase in linear velocity than the present monolithic silica capillary column.23 The present monolithic silica column showed higher permeability and higher total performance (E) at low linear velocity than a silica monolith prepared in a mold, although the larger E values are obtained at high linear velocities. The results clearly indicate that the monolithic silica columns can provide higher performance than current particle-packed column under pressure driven conditions. (34) Knox, J. H. J. Chromatogr., A 1999, 831, 3-15.

Figure 7. (a) Column back pressure against linear velocity of mobile phase. Column length was normalized to 83 mm. Mobile phase: 80% methanol. (b) Separation impedance against linear velocity of mobile phase calculated for hexylbenzene as a solute. (b) C18 monolithic silica column in a capillary; (2) larger-sized C18 silica rod column ((d) in Figure 2),23 (9) packed column with 5 µm C18 silica particles.23

Figure 8. Chromatograms obtained for alkylbenzenes (C6H5(CH2)nH, n ) 0-6 (a, c)) and polyaromatic hydrocarbons (b, d)) in pressure-driven (HPLC (a, b)) and electrodriven (CEC (c, d)) elution. Column and solutes as in Figure 4. Mobile phase: (a, b) 90% acetonitrile; (c, d) acetonitrile-Tris-HCl, 50 mM, pH 8 (90/10). Pressure: (a, b) 0.9 kg/cm2. Applied voltage: (c, d) 750 V/cm.

Figure 8 shows a comparison of the separation of the alkylbenzenes and PAHs in 90% acetonitrile in HPLC and CEC at similar linear velocities. Higher numbers of theoretical plates of ∼8000 and ∼30 000 for hexylbenzene as a solute were obtained in pressure-driven LC and CEC, respectively, than in 80% acetonitrile due to the higher Dm and the smaller k values. The electroosmotic mobility and the number of theoretical plates Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

1279

increased with increasing acetonitrile content or voltage applied in CEC. High-speed operation in CEC is desirable and of much interest. A problem to be solved is that the present monolithic silica column exhibits smaller electroosmotic mobility (t0) in the CEC mode than ordinary particle-packed capillaries,5 presumably due to the high purity of the silica skeleton. It is also necessary to prepare monolithic silica having smaller through-pores in capillary in order to obtain higher column efficiency in the order of 105 plates in a shorter time. CONCLUSION A silica gel monolith that has large through-pores was prepared in a fused-silica capillary. High performance of up to 32 000 theoretical plates/25 cm for hexylbenzene was obtained in a CEC mode, although the lower column efficiency of up to 12 000 plates/ 25 cm was observed in an HPLC mode due to the large A term contribution in the van Deemter equation. The monolithic silica

1280

Analytical Chemistry, Vol. 72, No. 6, March 15, 2000

column in capillary has made it possible to produce high efficiency in both electrodriven and pressure-driven modes at similar linear velocities with commonly available equipment. It is necessary to further optimize the silica structure in capillary with smaller and well-defined domain sizes to achieve higher efficiency in a shorter period of time, although it will be accompanied by the reduction in permeability. ACKNOWLEDGMENT This work was supported in part by a Grant-in-Aid for Scientific Research and Monbusho International Joint Research Program funded by the Ministry of Education, Science, Sports, and Culture.

Received for review August 18, 1999. Accepted December 14, 1999. AC990942Q