High-Efficiency Liquid Chromatographic Separation Utilizing Long

plates of a HPLC column,1-4 although it takes a long time. Scott and Kucera demonstrated an efficiency of 650000 plates using a. 14 m long column pack...
1 downloads 0 Views 460KB Size
Download PDF
Anal. Chem. 2008, 80, 8741–8750

High-Efficiency Liquid Chromatographic Separation Utilizing Long Monolithic Silica Capillary Columns Kosuke Miyamoto,† Takeshi Hara,† Hiroshi Kobayashi,† Hironobu Morisaka,† Daisuke Tokuda,† Kanta Horie,† Kodai Koduki,† Satoshi Makino,† Oscar Nu´n˜ez,†,‡ Chun Yang,†,§ Takefumi Kawabe,†,| Tohru Ikegami,† Hirotaka Takubo,† Yasushi Ishihama,⊥ and Nobuo Tanaka*,† Department of Biomolecular Engineering, Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan, and Institute for Advanced Biosciences, Keio University, 403-1, Daihoji, Tsuruoka, Yamagata 997-0017, Japan Long monolithic silica-C18 capillary columns of 100 µm i.d. were prepared, and the efficiency was examined using reversed-phase HPLC under a pressure of up to 47 MPa. At linear velocities of 1-2 mm/s, 100000-500000 theoretical plates could be generated with a single column (90-440 cm in length) using an acetonitrile-water (80/ 20) mobile phase with a column dead time (t0) of 5-40 min. It was possible to prepare columns with a minimum plate height of 8.5 ( 0.5 µm and permeability of (1.45 ( 0.09) × 10-13 m2. The chromatographic performance of a long octadecylsilylated monolithic silica capillary column was demonstrated by the high-efficiency separations of aromatic hydrocarbons, benzene derivatives, and a protein digest. The efficiency for a peptide was maintained for an injection of up to 0.5-2 ng. When three 100 µm i.d. columns were connected to form a 1130-1240 cm column system, 1000000 theoretical plates were generated for aromatic hydrocarbons with retention factors of up to 2.4 with a t0 of 150 min. The fact that very high efficiencies were obtained for the retained solutes suggests the practical utility of these long monolithic silica capillary columns. Achieving an extremely high column efficiency, e.g. 1000000 theoretical plates, has been the target of many separation scientists, although such a column may not be able to completely resolve a mixture containing hundreds of components.1 The use of a long column packed with relatively large-sized particles is a straightforward approach to increasing the number of theoretical plates of a HPLC column,1-4 although it takes a long time. Scott and Kucera demonstrated an efficiency of 650000 plates using a * To whom correspondence should be addressed. Tel: +81-75-724-7809. Fax: +81-75-724-7710. E-mail: [email protected]. † Kyoto Institute of Technology. ‡ Present address: University of Barcelona, Department of Analytical Chemistry, Barcelona, Spain. § Present address: Yangzhou University, College of Chemistry and Chemical Engineering, Yang Zhou, China. | Present address: Daiichisankyo Co., Hiratsuka, Japan. ⊥ Keio University. (1) Guiochon, G. J. Chromatogr. A 2006, 1126, 6–49. (2) Scott, R. P. W.; Kucera, P. J. Chromatogr. 1979, 169, 51–72. (3) Menet, H. G.; Gareil, P. C.; Rosset, R. H. Anal. Chem. 1984, 56, 1770– 1773. (4) Karlsson, K.-E.; Novotny, M. V. Anal. Chem. 1988, 60, 1662–1665. 10.1021/ac801042c CCC: $40.75  2008 American Chemical Society Published on Web 10/24/2008

14 m long column packed with 5 µm particles with a column dead time (t0) of about 500 min.2 Using a 22 m column packed with 7-8 µm particles, nearly 1000000 theoretical plates were achieved by Menet and co-workers with a t0 as long as 18 h.3 Karlsson and Novotny reported a high-efficiency separation of 226000 theoretical plates in a relatively short time using a capillary column packed with 5 µm particles with a t0 of about 33 min.4 The high efficiencies reported in these studies, however, were obtained for an unretained solute. Such high efficiencies have not been reported for retained solutes. Ultrahigh pressure liquid chromatography (UHPLC) developed by Jorgenson and co-workers utilizes very high pressure, up to 500 MPa, and columns packed with particles smaller than 2 µm.5-8 A separation with up to 310000 theoretical plates with a 1.8 min t0 was reported, using a 43 cm column packed with 1 µm particles.6 Instruments capable of UHPLC utilizing pressures of up to 100 MPa have been commercialized to achieve fast separations, and the high efficiency has been shown in various separations.9-11 UHPLC using small particles is a sure and attractive method to achieve a large number of theoretical plates per unit time, but requires special equipment and is primarily an approach to reduce the separation time. Capillary electrochromatography (CEC) also allows fast separations with a large number of theoretical plates using small-sized particles.12 Novotny and co-workers reported high efficiency columns that generated up to 610000 theoretical plates per meter in this mode.13 Examples showing a column efficiency greater than 1000000 theoretical plates were also reported for some ionized solutes in CEC, although the mechanism was not clearly (5) MacNair, J. E.; Lewis, K. C.; Jorgenson, J. W. Anal. Chem. 1997, 69, 983– 989. (6) MacNair, J. E.; Patel, K. D.; Jorgenson, J. W. Anal. Chem. 1999, 71, 700– 708. (7) Mellors, J. S.; Jorgenson, J. W. Anal. Chem. 2004, 76, 5441–5450. (8) de Villiers, A.; Lauer, H.; Szucs, R.; Goodall, S.; Sandra, P. J. Chromatogr. A 2006, 1113, 84–91. (9) de Villiers, A.; Lestremau, F.; Szucs, R.; Ge´le´bart, S.; David, F.; Sandra., P. J. Chromatogr. A 2006, 1127, 60–69. (10) Shen, Y.; Zhang, R.; Moore, R. J.; Kim, J.; Metz, T. O.; Hixson, K. K.; Zhao, R.; Livesay, E. A.; Udseth, H. R.; Smith, R. D. Anal. Chem. 2005, 77, 3090– 3100. (11) Wang, X.; Barber, W. E.; Carr, P. W. J. Chromatogr. A 2006, 1107, 139– 151. (12) Dittmann, M. M.; Rozing, G. P. J. Chromatogr. A 1996, 744, 63–74. (13) Que, A. H.; Konse, T.; Baker, A. G.; Novotny, M. V. Anal. Chem. 2000, 72, 2703–2710.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8741

elucidated.14,15 CEC is also a good method to achieve a large number of theoretical plates per unit time. The use of a long column to generate extremely large plate numbers is not necessarily facile due to the necessity for an ultrahigh electric field. Since the number of theoretical plates one can generate is often restricted by a limit in pump pressure, high-efficiency separations have also been studied using columns with high permeability, including a loosely packed capillary column and an open tubular column. A glass column of 1 mm i.d. packed with 10 µm silica particles, and subsequently drawn to a 44 µm i.d. loosely packed column of 16.2 m long, generated 500000 theoretical plates after octadecylsilylation for ethylbenzene with a k ) 0.11 in methanol at 10 MPa with a t0 of 27 min.16 Open-tubular columns can provide extremely efficient and fast separations.17,18 The long capillary columns are commonly coated with a polymer layer which can be covalently attached to the inner walls of the open tube. Tijssen and co-workers reported 2800000 theoretical plates using a 32 µm i.d., 27.5 m open tubular column.19 Kucera and Guiochon achieved 1.1 million theoretical plates for unretained benzene, using a 105 m capillary column of 39 µm i.d., while obtaining 270000 plates for anisole with a k value of 3.20 A marked decrease in column efficiency was observed with an increase in solute retention. Erni and co-workers used a shorter column (L ) 19.6 m, 50 µm i.d.) but at high temperature of up to 200 °C, and reported column efficiencies greater than 1000000 theoretical plates for solutes with k values of up to 0.1.21 Poppe and co-workers used a 11.4 µm i.d. capillary, 5 m long, coated with a polyacrylate layer of 1.5 µm thickness to form a column having a ca. 8.4 µm flow path. The column generated more than 1000000 theoretical plates for anthracene with k of ca. 0.1 with t0 of about 50 min, demonstrating the separation of six anthracene derivatives in acetonitrile.22 Recently Karger and co-workers reported extremely efficient separations for peptides utilizing long 10 µm i.d. open tubular columns.23,24 Although the separations were achieved under gradient conditions, the examples did show high performance of wall-coated narrow-bore capillary columns. A monolithic silica column also possesses the potential for high efficiency separations. Small-sized silica skeletons and the relatively large flow-through pores allow higher permeability than conventional particle-packed columns of similar efficiencies,25-28 demonstrating the possibility of producing greater than 100000 theoretical plates under very low pressure.27 Desmet and coworkers showed that monolithic silica capillary columns reported in 2002 could provide faster separations than particulate columns packed with particles of any size, over a range greater than ca. 80000 theoretical plates at a pressure drop of 40 MPa.28,29 The (14) Smith, N. W.; Carter-Finch, A. S. J. Chromatogr. A 2000, 892, 219–255. (15) Hilder, E. F.; Svec, F.; Frechet, J. M. J. J. Chromatogr. A 2004, 1044, 3– 22. (16) Tanaka, N.; Kinoshita, H.; Araki, M.; Tsuda, T. J. Chromatogr. 1985, 332, 57–69. (17) Guiochon, G. Anal. Chem. 1981, 53, 1318–1325. (18) Poppe, H. J. Chromatogr. A 1997, 778, 3–21. (19) Tijssen, R.; Bleumer, J. P. A.; Smith, A. L. C.; Van Kreveld, M. E. J. Chromatogr. 1981, 218, 137–165. (20) Kucera, P.; Guiochon, G. J. Chromatogr. 1984, 283, 1–20. (21) Liu, G.; Djordjevic, N. M.; Erni, F. J. Chromatogr. 1992, 592, 239–247. (22) Swart, R.; Kraak, J. C.; Poppe, H. J. Chromatogr. A 1995, 689, 177–187. (23) Yue, G.; Luo, Q.; Zhang, J.; Wu, S.-L.; Karger, B. L. Anal. Chem. 2007, 79, 938–946. (24) Luo, Q.; Yue, G.; Valaskovic, G. A.; Gu, Y.; Wu, S.-L.; Karger, B. L. Anal. Chem. 2007, 79, 6174–6181.

8742

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Table 1. Feed Composition for the Preparation of Monolithic Silica Columns

column no.

TMOS+ PEG TMOS MTMS urea AcOHf temp (g) (mL) (mL) (g) (mL) (°C)

MS-50-(1HK)a 12.6c MS-200H-(2DT)b 2.05c MS-200H-(3HK)b 1.85d MS-100H-(4TH, 1.80d 5KK, 7ON)b MS-100H-(6SM)b 1.84d MS-100H-(8KM)b 1.85e MS-100H-(9KM, 1.80e 10KM, 11KM, 12KM)b

40 18 18 18

9.0 4.05 4.05 4.05

100 40 40 40

30 40 40 40

18 18 18

4.05 4.05 4.05

40 40 40

40 40 40

a Prepared from TMOS. b Prepared from a TMOS-MTMS (3:1 v/v) mixture. c Poly(ethylene glycol), MW ) 10000, Aldrich, Batch No. 19803MO. d Poly(ethylene glycol), MW ) 10000, Aldrich, Batch No. 11431EA. e Poly(ethylene glycol), MW ) 10000, Aldrich, Batch No. 07403MD. f 0.01 N acetic acid aqueous solution.

objectives of our research are to prepare long monolithic silica capillary columns and to achieve large numbers of theoretical plates of up to 1000000 for retained solutes, thus showing the practical utility of long monolithic silica capillary columns for highefficiency separations by HPLC. EXPERIMENTAL SECTION Tetramethoxysilane (TMOS) and methyltrimethoxysilane (MTMS) were obtained from Shinn-etsu Chemicals, poly(ethylene glycol) (PEG, MW 10000) from Aldrich, the styrene oligomer molecular weight standard (MW ) 580) from GL Sciences, the peptides from Sigma Chemicals, a tryptic digest of bovine serum albumin (BSA) from Waters, and pravastatin from Daiichi-Sankyo. Deuterated compounds were obtained from CDN Isotopes, except 1,3,5-benzene-d3 which was obtained from ISOtopic, Ltd. Water purified with MilliQ A10 Gradient (Millipore) was used. Other chemicals and solvents were obtained from Nacalai-Tesque, Wako Pure Chemicals, Aldrich, and Tokyo Chemical Industries. The chemicals were used as obtained. Monolithic silica capillary columns were prepared from a mixture of MTMS and TMOS (v/v 3:1 unless otherwise noted) to form a hybrid structure (these columns are denoted as MS-H following the column diameter in µm), or from TMOS only (the column denoted as MS).30 The 100 and 200 µm i.d. columns were prepared from a mixture of TMOS and MTMS. The use of TMOS as a starting material allowed the preparation of 50-100 µm i.d. columns, although a 200 µm column could be prepared under certain conditions.31 Table 1 shows the feed compositions for the preparation of monolithic silica columns (the column number is followed by the initials of the investigator). Under typical condi(25) Leinweber, F. C.; Tallarek, U. J. Chromatogr. A 2003, 1006, 207–228. (26) Miyabe, K.; Cavazzini, A.; Gritti, F.; Kele, M.; Guiochon, G. Anal. Chem. 2003, 75, 6975–6986. (27) Motokawa, M.; Kobayashi, H.; Ishizuka, N.; Minakuchi, H.; Nakanishi, K.; Jinnai, H.; Hosoya, K.; Ikegami, T.; Tanaka, N. J. Chromatogr. A 2002, 961, 53–63. (28) Ishizuka, N.; Kobayashi, H.; Minakuchi, H.; Nakanishi, K.; Hirao, K.; Hosoya, K.; Ikegami, T.; Tanaka, N. J. Chromatogr. A 2002, 960, 85–96. (29) Desmet, G.; Clicq, D.; Gzil, P. Anal. Chem. 2005, 77, 4058–4070. (30) Nu´n ˜ez, O.; Nakanishi, K.; Tanaka, N. J. Chromatogr. A 2008, 1191, 231– 252. (31) Hara, T.; Kobayashi, H.; Ikegami, T.; Nakanishi, K.; Tanaka, N. Anal. Chem. 2006, 78, 7632–7642.

Table 2. Efficiency Evaluation of Long Monolithic Silica Capillary Columnsa column no. MS-50-C18 (1HK) MS-200H-C18 (2DT) MS-200H-C18 (3HK) MS-100H-C18 (4TH) MS-100H-C18 (5KK) MS-100H-C18 (6SM) MS-100H-C18 (7ON) MS-100H-C18 (8KM) MS-100H-C18 (9KM) MS-100H-C18 (10KM)

L (cm) c

d

254 (260) 253c (260)d 270c (276)d 88c (94)d 133c (143)d 120c (128)d 130c (138)d 442c (448)d 442c (448)d 342c (348)d

u (mm/s)

Nb (×104)

H (µm)

P (MPa)

K (×10-14 m2)

E

2.0 1.1 1.0 1.0 1.0 1.1 1.1 1.1 1.1 1.1

22.6 31.3 30.0 12.5 16.3 15.9 15.9 54.7 45.4 30.3

11.2 8.1 9.0 7.0 8.2 7.5 8.2 8.1 9.7 11.3

20 12 7.8 5.9 4.8 4.3 4.9 16.1 13.5 7.6

11.7 10.7 11.8 7.2 13.1 14.5 13.7 13.8 16.5 22.8

1070 610 740 680 510 390 490 480 570 560

a Column number indicates a monolithic silica column having the specified diameter (µm) followed by H in the case of a hybrid structure prepared from a mixture of TMOS and MTMS, by C18 indicating octadecylsilylation, then followed by the number and the initials of the investigator in parentheses. b Number of theoretical plates measured for hexylbenzene as a solute in 80% acetonitrile. c Effective length (cm). d Total length (cm).

Figure 1. Plots of plate height (H) values against linear velocity of mobile phase (u) obtained for twenty columns prepared with the same feed composition. Mobile phase: acetonitrile-water (80/20). Temperature: 30 °C. Sample: hexylbenzene. Detection: 210 nm. Column: monolithic silica C18 column, 150 cm. (a) Column: MS-100H-C18 (11KM). (b) Column: MS-100H-C18 (12KM).

tions, the fused-silica capillary tube was first treated with 1 M NaOH at 40 °C for 3 h, followed by a flush with water, and then kept with 1 M HCl at 40 °C for 2 h. After a flush with water, and then with acetone, the capillary tube was air-dried at 40 °C. A TMOS/MTMS mixture (18 mL) was added to a solution of PEG (1.80 g) and urea (4.05 g) in 0.01 M acetic acid (40 mL) at 0 °C and stirred for 30 min. The homogeneous solution was then stirred for 10 min at 40 °C, filtered with a 0.45 µm PTFE filter, charged into a fused-silica capillary tube, and allowed to react at 40 °C. The resultant gel was subsequently aged in the capillary overnight at the same temperature. Then, the temperature was raised slowly (over 10-20 h for long capillary columns), and the monolithic

silica columns were treated for 4 h at 120 °C to form mesopores with the ammonia generated by the hydrolysis of urea, then cooled and washed with methanol. After air-drying, the column was heattreated at 330 °C for 25 h causing the pyrolysis of all organic moieties in the column. The surface modification of the silica monoliths was carried out as previously described.27,28,30 Capillary HPLC was performed by employing a split flow/ injection HPLC system consisting of an LC-10AD vp Pump (Shimadzu) or an X-LC 3085PU Pump (JASCO), a 7725 injector (Rheodyne) with a splitting T-joint,27,28 and an on-column capillary UV detector CE-1575 (JASCO) or a UV detector with an optical unit including a capillary flow cell (volume: 6 nL) connected to an electronic unit with optical fibers, MU701 (GL Sciences). A detection window (2 mm) was created by removing the polyimide coating of the fused silica capillary at a specified distance from the capillary inlet to allow on-column detection through the silica monolith for the CE-1575 detector. When on-column detection was employed, the effective column length was shorter than the total length by a few centimeters. A microvalve injector (VICI, injection volume 20 nL) was used for the column loading study. During the experiment, the whole system including the pump, injector, and detector was kept at 30 °C in an air-circulated oven. The X-LC pump has a pressure limit of 100 MPa, but the operating pressure was limited to 50 MPa or less due to the limit in pressure stability of the connecting parts between the column and the injector, or between the columns. The column (375 µm o.d.) was connected to the injector using a Graphite Vespel ferrule (GV-04, GL Sciences). Two or three columns were connected to form a long column system with a stainless steel union using Graphite Vespel ferrules (MVSU/004 and MGVF-004, GL Sciences). The chromatographic data were collected and processed by EZChrom Elite software (GL Sciences). The number of theoretical plates of a column (N) was calculated based on the peak width at half-height for symmetrical peaks. The gradient elution of a BSA digest was carried out by using a Finnigan LTQ with a nano LC interface (Thermo Fisher Scientific), an Agilent 1200 Series pump (Agilent Technologies), a MS-100H-C18 column (28.4 and 300 cm), and CHEMINERT C41004-1 valve (Valco Instruments). The dwell time was 1.5 min with a flow split just before the injector, and the column dead time (t0) was 1.7 min for the 28.4 cm column, and 31.6 min for the 300 cm column. A spray voltage of 2000 V was applied. The MS scan range was 300-2000. MS scans were performed with a positive mode Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8743

Figure 2. Plots of log(t0/N2) against log(N) for the monolithic silica C18 capillary columns evaluated in Figure 1. The plots were obtained at a pressure drop of 40 MPa, and by assuming a flow resistance parameter Φ ) 700, the viscosity of the mobile phase η ) 0.00046 Pa s, a diffusion coefficient of the solute Dm ) 2.22 × 10-9 m2/s, and a Knox equation h ) 0.65ν1/3 + 2/ν + 0.08ν for particle-packed columns. The particle diameters (dp) for the particle-packed columns are 3, 5, and 10 µm, respectively.

Figure3.Separationofalkylbenzenes.Mobilephase:acetonitrile-water (80/20). Column: three monolithic silica C18 columns connected in series, MS-100H-C18 (8KM and two columns of 9KM). Effective length: 1140 cm. ∆P ) 35.4 MPa. u ) 1.24 mm/s. Detection: 210 nm. Temperature: 30 °C. Sample: thiourea and alkylbenzenes (C6H5CnH2n+1, n ) 0-10).

for 1 s to select three intense peaks and subsequently three MS/ MS scans were performed. The peak capacity values were obtained from the base peak chromatograms after smoothing. RESULTS AND DISCUSSION Preparation of Long Monolithic Silica Capillary Columns and Their Performance. The monolithic columns were first evaluated in 80% acetonitrile mobile phase using alkylbenzenes as solutes. Table 2 shows a summary of chromatographic properties of the columns. Typically, a 130-150 cm column can produce 150000-180000 theoretical plates with a pressure drop of 7-8 MPa and t0 of 15-20 min, 100000 theoretical plates with ca. 25 MPa and t0 of 5-6 min, and 77000 theoretical plates with ca. 40 MPa and t0 of 3 min. (Chromatograms are shown in Figure A in the Supporting Information.) Optimum performance was observed at linear velocity (u) of about 1.5 mm/s in 80% acetonitrile to yield a plate height (H) of about 8-10 µm, and at about 1 mm/s in 80% methanol to give a Hmin ) 10-12 µm. As 8744

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

shown in Table 2, the columns prepared recently showed separation impedance (E values) of 500-600, which are slightly higher than those observed for similar columns reported previously.28 It was difficult to prepare a long capillary column with a homogeneous monolithic silica structure along its entire length, while shorter columns can be obtained with higher efficiency by selecting portions prepared homogeneously. The column length was limited to up to 4-5 m in the preparation step by practical problems encountered during injection of the feed solution into a fused silica capillary with a syringe. The experiments were carried out over a period of more than six years by several investigators. The MS-100H-(3HK-7ON) columns were prepared using one batch of polyethylene glycol (PEG), and MS-100H-(8KM-12KM) using another PEG batch. A slight change in the PEG concentration can cause significant differences in the column properties for some batches of PEG. In general, an increase in the PEG concentration of the feed resulted in a column with lower permeability and higher column efficiency.27,28 The MS-100H-(8KM) column was designed to provide a greater number of theoretical plates and a higher pressure drop due to its smaller sized domains than the MS-100H-(9KM-12KM) columns. The reproducibility of the preparation was studied with the MS-100H-(11KM-12KM) columns. Figures 1a and 1b show van Deemter plots for two batches of ten 150 cm columns prepared with the same feed composition. Eight out of ten columns of the batch MS-100H-C18 (11KM) gave H values of 8.3 ± 0.5 µm and seven columns of MS-100H-C18 (12KM) gave H values of 8.7 ± 0.5 µm at linear velocity of ca. 1.5 mm/s, while all columns showed permeability K ) (1.45 ± 0.09) × 10-13 m2. There seems to be a limit in performance for a column prepared under the present conditions. It was possible to obtain columns showing a minimum plate height of 7.7-9.3 µm and separation impedance 498 ± 65 with 75% probability, but it was more difficult to obtain columns showing high efficiency at high linear velocity. Only seven columns of MS-100H-C18 (11KM) and two of MS-100H-C18 (12KM)

Figure 4. Separation of polynuclear aromatic hydrocarbons. Mobile phase: acetonitrile-water (80/20). Detection: 254 nm. Temperature: 30 °C. Sample: 16-PAHs primary pollutants designated by the EPA. Peak numbers: 1, naphthalene; 2, acenaphthylene; 3, fluorene; 4, acenaphthene; 5, phenanthrene; 6, anthracene; 7, fluoranthene; 8, pyrene; 9, chrysene; 10, benz(a)anthracene; 11, benzo(b)fluoranthene; 12, benzo(k)fluoranthene; 13, benzo(a)pyrene; 14, dibenz(a,h)anthracene; 15, indeno(1,2,3-cd)pyrene; and 16, benzo(g,h,i)perylene. (a) Column: MS-100H-C18 (5KK), effective length 133 cm. ∆P ) 4.7 MPa. u ) 1.03 mm/s. (b) Column: three monolithic silica C18 columns connected in series, MS100H-C18 (8KM, 9KM, and 10KM), effective length 1238 cm (total length 1244 cm). ∆P ) 46.6 MPa. u ) 1.31 mm/s. Detection: 210 nm. Temperature: 30 °C.

showed H ) 16-20 µm at u ) 8 mm/s. Although it is not possible to directly compare the reproducibility of these columns with that reported for commercially available monolithic silica rod columns,32 the present results indicated that the reproducibility of long monolithic silica capillary columns prepared in the laboratory is poorer than that of commercially available monolithic silica rod columns. Some low efficiency columns, however, showed improved performance when part of the column was cut off, indicating the presence of localized structural heterogeneity. This may explain the greater E values observed for the long columns than some shorter columns reported previously.27,28 The retention (32) Kele, M.; Guiochon, G. J. Chromatogr. A 2002, 960, 19–49.

factors were found to be more reproducible with relative standard deviation of about 2%. Figure 2 shows the plots of log(t0/N2) values against log(N)29 for the twenty columns evaluated in Figure 1, with a 40 MPa pressure limit. (The dashed lines indicate t0 values required for the generation of a specified number of theoretical plates.) Although the plots showed considerable scattering, nine monolithic columns were found to give similar performance over the region 5 < log(N) < 6, which seemed to be the upper limit of performance for the columns prepared under the present conditions. The results indicated that 100000 theoretical plates can be generated with a t0 of about 250 s, 300000 plates with a t0 of about 1000 s, and that the present columns can theoretically give Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8745

optimum performance at around 1000000 theoretical plates with a t0 of 5000-7500 s. Although the plots are more scattered in this region, the results suggested the possibility of utilizing the column efficiency over a range of N ) 100000-1000000 for practical separations using a commercially available HPLC pump and a UV detector, as shown in this study. Examples of High-Efficiency Separations. Figure 3 shows the separation of alkylbenzenes on a column system of 1140 cm consisting of three monolithic capillary columns connected in series (8KM, 9KM, and another column of a batch 9KM). At 35 MPa, the column system produced 1350000 theoretical plates for the t0 peak and 1000000 theoretical plates for alkylbenzenes with k values of up to 2.1. Smaller numbers of theoretical plates were obtained for alkylbenzenes retained longer, about 800000 theoretical plates for decylbenzene of k ) 5.6. A column connector resulted in an increase in the H value by 0.1-0.2 µm when examined with two columns, each about 60 cm long. Therefore the extracolumn effect caused by the column connectors is presumably less than 1% in terms of the number of theoretical plates of a column system longer than 10 m. Excellent column efficiencies were also observed for polynuclear aromatic hydrocarbons (PAHs). The 5KK column and a 1244 cm column system (8KM, 9KM, and 10KM) showed good performance for the separation of 16 PAHs, which are EPAdesignated priority pollutants (Figure 4). Although a monomeric C18 stationary phase is known to be less effective for the separation of rigid planar PAHs than a polymeric C18 phase,33 complete separation was achieved with the present columns. The critical pair, peak 9 and peak 10, dictates the separation conditions and separation time. Figure 4b shows the separation of 16 PAHs using a 1244 cm column system. (Note that the measurements were performed with on-column detection.) The column system generated 1000000 theoretical plates for PAHs with retention factors (k values) of up to 2.4. Peaks 9 and 10 were well separated, as shown in the inset of Figure 4b. High column efficiencies of around 1000000 theoretical plates were obtained with the present monolithic silica columns for retained solutes. The reduction in column efficiencies for solutes with larger k values was shown to be much less when compared to the successful examples reported for the open tubular columns.20-22 The throughpore size of about 2 µm is much smaller than the capillary diameter of the open tubular columns employed in the previous studies, which dictated the contribution of the slow mass transfer in the mobile phase to band broadening. Figure 5a shows a chromatogram of polar and nonpolar benzene and naphthalene derivatives in 60% methanol on a 30 cm monolithic silica C18 column prepared similarly as the 9KM column. Many overlapping peaks were noticeable for a mixture of 35 compounds, although the column produced 16000-30000 theoretical plates for these solutes. The chromatogram in Figure 5b shows the separation of the same mixture on a 445 cm column (9KM) using the same mobile phase. A much greater number of peaks were separated with the longer column. Although it takes a much longer time to elute the solutes from a column of this length, the example shows the advantage of using an efficient column for the separation of a complex mixture. (33) Sander, L. C.; Parris, R. M.; Wise, S. A.; May, W. E. Anal. Chem. 1991, 63, 2589–2597.

8746

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Figure 5. Separation of benzene and naphthalene derivatives. Mobile phase: methanol-water (60/40), 0.1% formic acid. Detection: 254 nm. Temperature: 30 °C. Solutes: 1, benzene; 2, toluene; 3, phenol; 4, o-cresol; 5, p-cresol; 6, 2-naphthol; 7, benzonitrile; 8, nitrobenzene; 9, methyl benzoate; 10, methyl phenylacetate; 11, methyl phenyl ketone; 12, ethyl phenyl ketone; 13, phenyl propyl ketone; 14, p-nitrotoluene; 15, o-nitrophenol; 16, m-nitrophenol; 17, p-nitrophenol; 18, p-fluorophenol; 19, p-chlorophenol; 20, p-bromophenol; 21, p-iodophenol; 22, o-dinitrobenzene; 23, p-dinitrobenzene; 24, p-nitrobenzyl alcohol; 25, p-cyanophenol; 26, methyl p-hydroxybenzoate; 27, ethyl p-hydroxybenzoate; 28, propyl phydroxybenzoate; 29, dimethyl phthalate; 30, diethyl phthalate; 31, dimethyl terephthalate; 32, 1,5-dinitronaphthalene; 33, 1,8-dinitronaphthalene; 34, 2,4-dinitrochlorobenzene; and 35, 2-amino-4-nitrophenol. (a) Column: MS-100H-C18 (9KM), effective length 30 cm. ∆P ) 4.8 MPa. u ) 1.04 mm/s. (b) Column: MS-100H-C18 (9KM), 445 cm. ∆P ) 35.1 MPa, u ) 1.00 mm/s.

An increase in peak capacity, however, may not be the primary advantage of using a long column, because the peak capacity increase is not steep with the increase in column length and separation time. For a similar separation time and linear velocity, peak capacity values calculated for a long column and a shorter one with efficiencies of 400000 theoretical plates and 30000 theoretical plates, respectively, with similar H values, are not much different. When compared for the same separation time (tR), for example for the elution range up to tR ) 3t0 for the longer column, the peak capacity values of 160-170 were obtained for both cases. However, the minimum separation factor necessary to provide unit resolution of Rs ) 1 in this range is substantially different, ca. 1.010 for the longer column as compared to 1.024 for the shorter column. The actual bandwidth of the separated peaks can be much smaller for the longer column. Long monolithic silica capillary

Figure 6. Gradient separation of a BSA digest. Column: MS-100H-C18-(6SM), (a) 28.4 cm, and (b) 300 cm. The mobile phases consisted of (A) 0.1% formic acid in water and (B) 0.1% formic acid in acetonitrile. A linear gradient of 5% to 40% B in 30 min for the short column and in 300 min for the long column was employed. A BSA digest sample dissolved in water (100 nL, 1 nmol/200 µL) was injected at the start of gradient. with a flow rate 1.2 µL/min for the 28.4 cm column and 0.67 µL/min for the 300 cm column.

columns will be useful for the separation of very complex mixtures over a relatively narrow range of a chromatogram, with little effort required for method development. Performance and Sample Loading Capacity for Peptides. Figure 6 compares chromatograms of gradient elution of a BSA digest sample on a 28.4 cm column (MS-100H-C18-(6SM)) and on a 300 cm column (from the same column batch as the 28.4 cm column), with acetonitrile gradient in the presence of 0.1% formic acid. Gradient times approximately proportional to the column length were employed. Figure 6a shows the result of 30 min gradient elution on a 28.4 cm column, showing peak capacity of ca. 100 in 23 min for the elution range of the major portion of peptides between the two peaks having mass numbers of 330 and 863, while about 30% greater peak capacity was observed in 35 min with a longer gradient time, 50 min, for the same elution

range. The 300 cm column resulted in peak capacity of about 380 in 215 min for the same elution range of the peptides. Peak capacity obtainable per unit time is more than two times greater with the short column than with the longer one, but the absolute peak capacity observed with the 300 cm column was more than three times greater than that obtainable with the 28.4 cm column for the elution range of the peptides in the protein digest, as shown in Figure 6 for the particular sample. Detection of increased numbers of peaks by the use of 65-90 cm monolithic silica capillary columns in a gradient elution was reported for peptides and plant metabolites.34,35 Gradient elution on a short (34) Luo, Q. Z.; Shen, Y. F.; Hixson, K. K.; Zhao, R.; Yang, F.; Moore, R. J.; Mottaz, H. M.; Smith, R. D. Anal. Chem. 2005, 77, 5028–5035. (35) Tolstikov, V. V.; Lommen, A.; Nakanishi, K.; Tanaka, N.; Fiehn, O. Anal. Chem. 2003, 75, 6737–6740.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8747

Figure 7. Plot of N value against the weight of sample injected. Column: MS-100H-C18 (6SM). Effective length: 196 cm. Mobile phase: acetonitrile-water (20/80) containing 0.1% TFA. Detection: 220 nm. Temperature: 30 °C. For pravastatin, a short column (50 cm) prepared under similar conditions as for MS-100H-C18 (9KM) was used with a mobile phase acetonitrile-water (20/80) containing 0.01 mol/L ammonium formate (pH 7.0). ∆P ) 7.5 MPa. u ) 2.2 mm/s. Detection: 240 nm. Temperature: 30 °C.

column may provide a large peak capacity per unit time, but it is accomplished by compressing the peaks that are easily separated. As mentioned earlier, the use of a long column is not necessarily advantageous for increasing a total peak capacity of a chromatographic system, but will be very effective in increasing resolution of a highly complex mixture. Figure 7 shows the plot of N value obtained with a 196 cm column prepared similarly as MS-100H-C18-(6SM) against the weight of a peptide with a molecular weight range of 500-1200, showing sample loading capacity of the capillary column. (A chromatogram obtained for the peptides under isocratic conditions in a 20/80 acetonitrile-water mobile phase containing 0.1% trifluoroacetic acid (TFA) is shown in Figure B of the Supporting Information.) While lower efficiency was observed for two peptides, bradykinin fra1-5 (MW ) 572.66, peak 2) and bradykinin (MW ) 1060.36, peak 6) having arginine end groups, Argvasopressin (MW ) 1084.23) with a k ) 0.45 and oxytocin (MW ) 1007.19) with a k ) 1.23 showed 110000-150000 theoretical plates with small sample loading with valve injection, and showed a 20-25% decrease in their N value at 2 ng injection. Longer retained Met-enkephalin (MW ) 573.66) (k ) 1.61), luteinizing hormone releasing hormone (LHRH, MW ) 1183.27) (k ) 1.76), bradykinin, and Leu-enkephalin (MW ) 555.62) (k ) 3.20) showed a similar decrease with a 0.5-1.0 ng injection, and 50-65% decrease in N value at 2 ng injection. Figure 7 also shows the effect of sample loading on column efficiency for an active pharmaceutical ingredient for hypercholesterolemia, pravastatin (MW ) 424.5), eluted with k ) 2.9 in a 20/80 acetonitrile-water mixture at pH 7.0, on a shorter column prepared similarly to MS-100H-C18 (9KM). The 50 cm column, producing about 22000 theoretical plates for a small sample load, also showed about 20% decrease in column efficiency with a 1 ng injection and ca. 40% decrease with a 2 ng injection. The injection of a 1 ng sample into a 100 µm i.d. column is equivalent to a ca. 2 µg injection into a 4.6 mm i.d. column based on the difference in the cross section of the columns by a factor of about 2000. The loading capacity found with the high-efficiency 8748

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Figure 8. Separation of styrene oligomers (molecular weight standard for MW ) 580). The numbers in parentheses indicate the number of styrene units in the oligomer. Column: MS-100H-C18 (8KM, 9KM, and another 9KM). Effective length: 1130 cm. Mobile phase: acetonitrile-water (95/5). ∆P ) 39.5 MPa. u ) 1.73 mm/s. Detection: 210 nm. Temperature: 30 °C. The inset is a magnification of the chromatogram for the pentamers.

monolithic silica capillary column seems to be less than that reported for a conventional particle-packed column.36 The effect of sample loading was more clearly observed for a solute with a greater retention factor irrespective of the molecular weight in the range studied, and reported to be more pronounced for a high efficiency column.37 The phase ratio of the monolithic silica capillary column which is smaller than that of a particle-packed column by a factor of about 5,38 and the lower acetonitrile concentration in the mobile phase than would be suitable for a high-phase ratio material are presumably responsible for the smaller loading capacity compared to a particle-packed column. The amount of stationary phase available for solute binding seems to be a dominant factor. As a matter of fact, the amount of silica in one theoretical plate, which can be a measure of sample loading capacity,37 is ca. 8 × 10-15 m3 for a 100 µm i.d. monolithic silica capillary column (a 100 cm column producing 100000 theoretical plates) with a total porosity of 90%,39 and is smaller than that of a 4.6 mm i.d. column packed with particles (a 15 cm column producing 10000 theoretical plates) having 75% total porosity by a factor of ca. 8000. However, the loading capacity of the high-efficiency monolithic silica capillary column of 100 µm i.d. is much greater than those reported for a high-efficiency open tubular column having reduced dimensions in terms of the amount of silica in one theoretical plate.37 Separation Based on a Small Separation Factor by Using a Long Monolithic Silica Capillary Column. Figure 8 shows the separation of styrene oligomers in an isocratic mode elution in an acetonitrile-water (95/5) mobile phase using a column system of 1130 cm. A mixture of styrene oligomers has been the target of a study for high selectivity separations. Although it is relatively easy to separate oligomers based on the difference in the number of styrene units in a reversed-phase mode, it is hard (36) McCalley, D. V. J. Chromatogr. A 1998, 793, 31–46. (37) Tock, P. P. H.; Duijsters, P. P. E.; Kraak, J. C.; Poppe, H. J. Chromatogr. A 1990, 506, 185–200. (38) Kobayashi, H.; Ikegami, T.; Kimura, H.; Hara, T.; Tokuda, D.; Tanaka, N. Anal. Sci. 2006, 22, 491–501. (39) Kobayashi, H.; Tokuda, D.; Ichimaru, J.; Ikegami, T.; Miyabe, K.; Tanaka, N. J. Chromatogr. A 2006, 1109, 2–9.

Figure 9. Separation of benzene and toluene isotopologues. Mobile phase: acetonitrile-water (30/70). Column: MS-100H-C18 (9KM), 440 cm. Detection: 210 nm. Temperature: 30 °C. u ) 1.76 mm/s. ∆P ) 39.6 MPa. Sample: 1, thiourea; 2, benzene-d6; 3, benzene-1,3,5-d3; 4, benzened; 5, benzene; 6, toluene-d8; 7, toluene-R,R,R-d3; and 8, toluene. The inset (a) is a magnification of part of the chromatogram at 185-200 min. The inset (b) is a chromatogram for the mixture of benzene isotopologues obtained with two monolithic silica capillary columns connected in series, MS-100H-C18 prepared from a mixture of MTMS and TMOS (9/2), length 850 cm (500 cm + 350 cm) in a ternary mobile phase, acetonitrile-methanol-water (10/5/85). u ) 1.02 mm/s. ∆P ) 34 MPa.

to completely resolve the theoretical numbers of diastereomeric isomers of styrene oligomers having the same molecular weight. Simple alkyl-bonded stationary phases could not separate them.40 For a study of structure-retention correlations of pentamers or higher, two-dimensional HPLC including a selective stationary phase40 or recycle chromatography with peak shaving were employed followed by NMR measurement.41 Almost complete separation of all the possible diastereomeric isomers was achieved for pentamers by using the 1130 cm column system, as shown in Figure 8. This kind of one-step separation of all possible diastereomeric isomers is unprecedented for styrene pentamers. The peaks of the pentamers showed ca. 800000 theoretical plates at k ) 2.7. The separation factor between peaks 11 and 12 was estimated to be about 1.004. The separation of isotopologues usually requires a large number of theoretical plates because of the small separation factors,42,43 and can nicely demonstrate the capability of a highefficiency separation system. For the HPLC separation of deuterated compounds from protiated compounds, one can expect a single isotope effect of up to 1.010 per deuterium atom on the retention factors, (kH/kD).42,43 Figure 9 shows an isotopic separation of benzene, benzene-d, benzene-1,3,5-d3, benzene-d6, toluene, toluene-R,R,R-d3, and toluene-d8, performed with the MS-100HC18 (9KM) column in acetonitrile-water (30/70), and the inset (a) shows a magnification of a chromatogram for benzene isotopologues. The single isotope effects observed with benzene (40) Gray, M. J.; Dennis, G. R.; Slonecker, P. J.; Shalliker, R. A. J. Chromatogr. A 2005, 1073, 3–9. (41) Sato, H.; Tanaka, Y.; Hatada, K. Makromol. Chem., Rapid. Commun. 1982, 3, 175–179. (42) Turowski, M.; Yamakawa, N.; Meller, J.; Kimata, K.; Ikegami, T.; Hosoya, K.; Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 2003, 125, 13836–13849. (43) Tanaka, N.; Thornton, E. R. J. Am. Chem. Soc. 1977, 99, 7300–7307.

were calculated to be 0.70-0.72%. Similar magnitudes of isotope effects per D atom were observed for aromatic CH/CD between benzene and toluene, which were greater than the isotope effects observed for various aliphatic C-H/C-D isotopologues.42,43 Such a study on the isotope effects of a hydrophobic retention process or isotopic separations would be much easier with high efficiency columns affording resolution based on the small separation factors than with shorter columns that require separate injections of the isotopologues. The separation of H/D isotopic benzenes, including the successful separation between benzene and monodeuterated benzene, was achieved by recycle chromatography.44-46 When compared to recycle chromatography, the use of a long monolithic capillary column allows a simpler operation and a wider range of application. According to our calculations, it will be possible to produce a resolution of unity for a pair of compounds with a separation factor as small as 1.0056 using 1000000 theoretical plates at a k of 2.5. This corresponds to a difference in the free energy of 14.2 J/mol (3.4 cal/mol) associated with the transfer of the solutes from the mobile phase to the stationary phase. Figure 9 inset (b) shows the separation of the benzene isotopologues with a MS-100H-C18 column prepared from a mixture of TMOS and MTMS at a 9/2 ratio. An acetonitrilemethanol-water mixture (10/5/85) was employed for this separation because of the better peak shape observed than in a binary solvent system. This chromatogram seems to show the resolution of all the H/D isotopologues of benzene contained in the samples purchased as benzene, benzene-d, benzene-1,3,5-d3, and benzene(44) van der Wal, Sj. J. Liq. Chromatogr. Relat. Technol. 1985, 8, 2003–2016. (45) van der Wal, Sj. Chromatographia 1986, 22, 81–87. (46) Lim, L. W.; Uzu, H.; Takeuchi, T. J. Sep. Sci. 2004, 27, 1339–1344.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8749

d6. The use of a high-efficiency column is an easy way to resolve a mixture with small separation factors. Application of a high-efficiency column will be of particular interest for chiral separations where resolution with a small separation factor and small retention factors is very advantageous. A study on a 100-200 cm monolithic silica capillary column modified with cyclodextrin derivatives is under way. Such a column can provide chiral resolution with small retention factors. CONCLUSIONS The high performance of a long monolithic silica capillary column 1-4.5 m in length was demonstrated for polar and nonpolar compounds. It was possible to prepare 150 cm monolithic silica capillary columns with minimum plate heights of about 8.5 µm at 1.5 mm/s and H ) 16-20 µm at 8 mm/s with a 40-50% success ratio, and to operate a set of these long capillary columns at a pressure drop of less than 50 MPa using commercially available equipment and fittings. About 1000000 theoretical plates were generated by three serially connected columns of 11.4-12.5 m total length for retained solutes with k values of up to 2.4, suggesting the possibility of practical applications. For those using conventional HPLC equipment, it would be practical to use a 1-2 m monolithic silica capillary column at a pressure drop of 20 MPa or lower to generate 100000-200000 theoretical plates with a t0 of 5-20 min. If one can use HPLC equipment with a pressure range of 40-50 MPa, 300000 theoretical plates can be obtained

8750

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

with a t0 of 15-20 min, and 1000000 theoretical plates with a t0 around 2.5 h. The column efficiency was shown to be maintained for some polar compounds with a sample loading of up to 0.5-2 ng for a 100 µm i.d. column. A long monolithic silica capillary column may have practical utility, especially for resolving very closely related solutes with a separation factor of less than 1.01 that would otherwise be impossible. ACKNOWLEDGMENT This work was supported in part by Grants-in-Aid for Scientific Research funded by the Ministry of Education, Sports, Culture, Science and Technology of Japan, Nos. 17350036 and 20350036. The authors gratefully acknowledge the postdoctoral fellowship awarded to C.Y. and O.N., and the Grant-in-Aid for Scientific Research No. 17-05128 and 17-05394 funded by Japan Society for the Promotion of Science. The support from Nacalai-Tesque, GL Sciences, and Merck KGaA, Darmstadt, is also gratefully acknowledged. The authors thank Shouhei Miwa for helpful discussion. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review May 22, 2008. Accepted September 11, 2008. AC801042C