Anal. Chem. 2006, 78, 1323-1330
Fabrication of Internally Tapered Capillaries for Capillary Electrochromatography Electrospray Ionization Mass Spectrometry Jack Zheng, Dean Norton, and Shahab A. Shamsi*
Department of Chemistry, Center of Biotechnology and Drug Design, Georgia State University, Atlanta, Georgia 30303
In this study, we report a novel procedure for fabricating internally tapered capillary columns suitable for the coupling of capillary electrochromatography (CEC) to electrospray mass spectrometry (ESI-MS). The internal tapers were prepared by slowly heating the capillary end in a methane/O2 flame. Due to continuous self-shrinking of the inner channel of the capillary, the inside diameter of the opening was reduced to 7-10 µm. The procedure is easy to handle, with no requirement for expensive equipment as well as elimination of problematic grinding of the tip. Several advantages of these new internal tapers, as compared to using externally tapered columns, are described. First, the problems of poor durability and tip breakage associated with external tapering were successfully overcome with the internal taper. A comparison of the online CEC/ESI-MS between external versus internal tapers showed that the latter provides enhanced electrospray stability, resulting in significantly lower short-term noise and very short-term noise values. In turn, the more rugged design of internal tapers allows performing CEC/ MS utilizing a harsh polar organic mobile phase, which was not previously successful using an external taper due to higher operating current and electrospray arcing. Next, data on the reproducibility of the internally tapered CEC/ MS column using warfarin and β-blockers as model analytes are presented. For example, when comparing the reproducibility for separation of warfarin under reversedphase conditions, the internal taper demonstrated superior intraday % RSD (1.6-3.4) as compared to the external taper intraday % RSD (5-6). Last, the applicability of performing quantitative CEC/MS with internally tapered capillaries is demonstrated for simultaneous enantioseparation of β-blockers. Impressive quantitative results include good linearity of calibration curves (e.g., R2 ) 0.9940-0.9988) and limit of detection as low as 30 nM. The sensitive detection of a minor impurity of one enantiomer at the 0.1% level in a major chiral entity buttresses the suitability of compliance with FDA guidelines. Recently, the coupling of capillary electrochromatography (CEC) to mass spectrometry (MS) detection has attracted growing * Corresponding author. Phone: 404-651-1297. Fax: 404-651-2751. E-mail:
[email protected]. 10.1021/ac051480l CCC: $33.50 Published on Web 01/11/2006
© 2006 American Chemical Society
interest, since this hyphenated technique combines both high efficiency and selectivity along with high sensitivity and specificity.1-7 For example, when comparing the analysis of multicomponent mixtures (e.g., β-blockers) using conventional CEC/UV to CEC/ MS analysis, the CEC/UV requires a tedious spiking procedure to identify each component in a mixture, whereas CEC/MS allows simple determination of the elution order using extract ion mode. One of the critical problems in CEC is the bubble formation generated in the unpacked section of the column during the separation. As a result, irreproducible retention time, and even current breakdown occur, resulting in poor a signal-to-noise ratio arising from an erratic baseline. In conventional CEC/UV instrumentation, these problems can be overcome by degassing the mobile phase and, most effectively, by applying external pressure (e.g., 12 bar) on both the inlet and outlet ends of the CEC capillary during the separation; however, this arrangement is unavailable for CEC with MS detection, since the capillary outlet, which also serves as the MS sprayer, is typically under atmospheric pressure. In the case of CEC/MS, when the typical CEC/UV capillary configuration using two retaining frits with an unpacked segment at the outlet end is exposed to the electrospray, bubbles are prone to form, which tend to accumulate at the outlet frit; thus, irreproducible retention time and the problems aforementioned are typically observed.8 In CEC/MS column design, an alternative approach is to use a tapered outlet end instead of a frit to hold the packing material in place. Lord et al.9 first reported a type of externally tapered CEC/MS column which has an opening of ∼10-µm i.d. Since this type of taper is easy to fabricate with the aid of special equipment (e.g., laser puller), it has been widely utilized for CE/MS applications, including both sheath liquid and sheathless arrangements.10 In our recent publication, an externally tapered CEC/ (1) Boughtflower, R. J.; Paterson, C. J.; Knox, J. H. J. Chromatogr., A 2000, 887, 409-420. (2) Schurig, V.; Mayer, S. J. Biochem. Biophys. Methods 2001, 48, 117-141. (3) Que, A. H.; Novotny, M. V. Anal. Chem. 2002, 74, 5184-5191. (4) von Brocke, A.; Wistuba, D.; Gfrorer, P.; Stahl, M.; Schurig, V.; Bayer, E. Electrophoresis 2002, 23, 2963-2972. (5) Barcelo-Barrachina, E.; Moyano, E.; Galceran, M. T. Electrophoresis 2004, 25, 1927-1948. (6) Klampfl, C. W. J. Chromatogr., A 2004, 1044, 131-144. (7) Shamsi, S. A.; Miller, B. E. Electrophoresis 2004, 25, 3927-3961. (8) Zheng, J.; Shamsi, S. A. Anal. Chem. 2003, 75, 6295-6305. (9) Lord, G. A.; Gordon, D. B.; Myers, P.; King, B. W. J. Chromatogr., A 1997, 768, 9-16. (10) Choudhary, G.; Apffel, A.; Yin, H. F.; Hancock, W. J. Chromatogr., A 2000, 887, 85-101.
Analytical Chemistry, Vol. 78, No. 4, February 15, 2006 1323
MS column in reversed-phase mode was utilized for the quantitative analysis of warfarin enantiomers in human plasma.8 Using this type of tapered column, the reproducibility of retention times of warfarin enantiomers were significantly improved as compared to an untapered column.8 Therefore, our initial CEC/MS experiment was conducted on externally tapered columns that were packed with 3-µm vancomycin chiral stationary phase.11,12 However, such externally tapered columns were found to last for only a couple of runs under polar organic mode of CEC/MS. One of the major problems was the fragile outlet end of the external taper that tended to break during the voltage ramp or due to arcing in the spray chamber. Since a higher current was observed using a polar organic mode (∼6 µA for an applied voltage of 25 kV), as compared to reversed-phase mode (∼3 µA for an applied voltage of 25 kV) of CEC/MS, this problem was even more critical for the former. In addition, the preparation of the external tapers was too difficult to control, resulting in poor column-to-column reproducibility. Therefore, to overcome these problems, the possibility of performing CEC/MS experiments with an internally tapered column was investigated. Previously, CEC/MS on an internally tapered column was demonstrated by Choudhary13 using a two-step procedure. First, a high-temperature flame was used to seal the capillary tip followed by reopening of the sealed end with grinding. Initially, we followed this procedure using a microtorch and metallographic grinding paper. The results showed that it was, indeed, possible to prepare internal tapers using the two-step method. However, several critical limitations were observed. For example, the grinding process lacked precision and was very difficult to control. In addition, the tapered tip tended to break and accumulate debris in the open channel during the grinding process. During a prototype experiment in our laboratory, we accidentally observed that by slowly heating the end of a fusedsilica capillary, continuous self-shrinking of the inner channel of a 75-µm-i.d. capillary (Figure 1a-c) could reduce the inside diameter to form a reduced taper with a final opening of 7-10µm i.d. (Figure 1d). We realized that the aforementioned tip could be used as the outlet of the CEC column to retain the packing in its position (Figure 1e-f) through a so-called “keystone effect”.14-16 Further investigation showed that continuously spinning the capillary was very important to maintain the square and smooth shape of the tip (Figure 1d) during the heating process. It is worth mentioning that by adopting our aforementioned procedure, the tedious grinding step can be completely avoided because our procedure does not allow the taper to close. Furthermore, the columns are easy to handle, with no requirement for expensive equipment (i.e., laser puller or grinding machine) for CEC/MS column fabrication.13 In this work, we report a novel procedure to fabricate internally tapered columns with the aim to enhance stability, sensitivity, and (11) Armstrong, D. W.; Tang, Y. B.; Chen, S. S.; Zhou, Y. W.; Bagwill, C.; Chen, J. R. Anal. Chem. 1994, 66, 1473-1484. (12) Karlsson, C.; Wikstrom, H.; Armstrong, D. W.; Owens, P. K. J. Chromatogr., A 2000, 897, 349-363. (13) Choudhary, G.; Horvath, C.; Banks, J. F. J. Chromatogr., A 1998, 828, 469480. (14) Mayer, M.; Rapp, E.; Marck, C.; Bruin, G. J. M. Electrophoresis 1999, 20, 43-49. (15) Baltussen, E.; van Dedem Gijs, W. K. Electrophoresis 2002, 23, 12241229. (16) Rapp, E.; Bayer, E. J. Chromatogr., A 2000, 887, 367-378.
1324 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
Figure 1. Microscopic images (a-e) showing the preparation of internal taper for CEC/MS columns (magnification 100×) (a) before and (b-d) after heating for ∼15, 30, and 45 s (magnification 100× and 400×), respectively. Microscopic image (e) showing that the CEC column packed with the internal taper is flushed with mobile phase. A schematic (f) of internally tapered CEC/MS column showing dimension of packed and unpacked segments.
robustness of the CEC/MS system. For this purpose, the effects of electrospray stability by using two different tapered columns were compared first. Next, two different CEC/MS modes based on various combinations of stationary phase/mobile phase and ESI-MS detection were tested. The compatibility of internally tapered columns was evaluated for run-to-run reproducibility of migration time of (()-warfarin utilizing reversed-phase mode with negative ion ESI-MS detection. On the other hand, both run-torun and column-to-column reproducibility of the migration time of β-blockers was tested using polar organic mode with positive ion ESI-MS detection. Finally, the quantitation capability of the internally tapered CEC/MS column was evaluated with simultaneous enantioseparation and quantitation of four β-blockers as well as quantitation of a minor enantiomer at the 0.1% level. EXPERIMENTAL SECTION Reagents and Chemicals. The 3-µm vancomycin chiral stationary phase (CSP) was provided by Advanced Separation Technologies (Whippany, NJ). The 5-µm (3R, 4S) Whelk-O1 CSP (100-Å pore size) was a gift from Regis Technologies (Morton Grove, IL). Racemic mixtures of (()-alprenolol hydrochloride, (()atenolol, (()-metoprolol (+)-tartrate salt, (()-oxprenolol hydrochloride, (()-pindolol, (()-propranolol hydrochloride, (()warfarin, and (()-coumachlor were purchased from Aldrich
(Milwaukee, WI). Racemic mixtures of (()-carteolol and (()talinolol were kindly donated by BetaChem (Leawood, KS) and AWD Pharma (Dresden, Germany), respectively. Ammonium acetate (NH4OAc), acetic acid (HOAc), triethylamine (TEA), and sodium chloride (NaCl) were purchased from Aldrich (Milwaukee, WI). The HPLC grade organic solvents, acetonitrile (ACN) and methanol (MeOH), were purchased from Fisher (Springfield, NJ). Water used in all of the experiments was purified by a Barnstead Nanopure II Water System (Dubuque, IA). CEC/MS Column Fabrication. Fused-silica capillaries (o.d. 363 µm, i.d. 75 µm), obtained from Polymicro Technologies (Phoenix, AZ), were used to pack CEC/MS columns. The metallographic grinding paper (600 grit) was provided by Buehler (Lake Bluff, IL). The internally tapered column is fabricated by carefully heating the end of a capillary in a methane/O2 flame using a microtorch. First, a diamond cutter is utilized to generate a clean cut at one end of a ∼1 m long fused-silica capillary (Figure 1a), then that capillary end is heated to ∼800 °C in the methane/ O2 flame. After the polyimide coating of the column end is burned off, the capillary end is moved quickly in and out of the flame center until the internal taper is gradually formed over a period of 30-60 s (Figure 1b-d). To prevent deterioration and warping of the tapered end, the capillary is rotated by hand during the heating process. In contrast to the procedure developed by Choudhary,13 our procedure does not allow the taper to close the internal channel completely (Figure 1d-e); hence, the internally tapered capillary with an opening of ∼7-10-µm i.d. (Figure 1d) can be produced reproducibly without further sanding. Finally, these internally tapered capillaries are packed with CSP (Figure 1e) using the same procedure reported elsewhere.8 A typical CEC/ MS column consists of ∼63 cm total length with packed and unpacked segments of 60 and 3 cm, respectively (Figure 1f). CEC/ESI-MS Instrumentation. All chiral CEC/ESI-MS experiments were carried out with an Agilent 3D-CE capillary electrophoresis instrument interfaced to a single quadrupole mass spectrometer, Agilent 1100 series MSD (Palo Alto, CA). An Agilent 1100 series HPLC pump equipped with a 1:100 splitter was used to deliver the sheath liquid. CEC/ESI-MS Conditions. The mobile phase or sheath liquid was degassed for 30 min and filtered with a 0.45-µm PTFE membrane before use. Before the CEC/MS run, the column was preconditioned using the desired mobile phase.8 The injection was performed electrokinetically at 6 kV for 8 s in all cases unless otherwise mentioned. The separation voltage was set at 25 kV, employing a voltage ramp of 3 kV/s. During the separation, a 12-bar external pressure was applied to the inlet buffer vial for suppressing the bubble formation in the column without deteriorating the separation efficiency. For β-blockers study, the following ESI-MS conditions were used unless otherwise stated: mobile phase, MeOH/ACN/HOAc/TEA (70:30:1.6:0.2, v/v/v/v); sheath liquid, MeOH/H2O (90:10, v/v) containing 50 mM NH4OAc; flow rate, 5.0 µL/min; capillary voltage, +3000 V; fragmentor voltage, 80 V; drying gas flow rate, 5 L/min; drying gas temperature, 130 °C; nebulizer pressure, 4 psi. The positive selective ion-monitoring (SIM) mode was set with m/z 250.0 (alprenolol), 267.0 (atenolol), 293.0 (carteolol), 268.0 (metoprolol), 266.0 (oxprenolol), 249.0 (pindolol), 260.0 (propranolol), and 364.0 (talinolol). For the warfarin study, the following ESI-MS conditions were used unless
otherwise stated: mobile phase, ACN/H2O (70:30, v/v) containing 5 mM NH4OAc, pH 4.0; sheath liquid, MeOH/H2O (70:30, v/v) containing 5 mM NH4OAc at pH 8.5; flow rate, 7.5 µL/min; capillary voltage, -2500 V; fragmentor voltage, 91 V; drying gas flow rate, 5 L/min; drying gas temperature, 350 °C; nebulizer pressure, 4 psi; negative ion SIM mode set at m/z 307.0. Preparation of Standard Analytes. The racemic mixture of (() warfarin was dissolved in 60% (v/v) ACN at a concentration of 1.0 mg/mL. All β-blockers were dissolved in MeOH at a concentration 30 mM and stored at -20 °C. Linearity was checked by three replicate injections of mixtures containing four β-blockers, namely, (()-atenolol, (()-metoprolol, (()-oxprenolol, and (()propranolol, over the range of 3.0-600 µM, each containing 300 µM of (()-talinolol as internal standard. Noise Level Determination and Statistical Evaluation. Two different types of noise, including short-term noise and very shortterm noise were utilized to investigate the effects on stability of electrospray using externally vs internally tapered capillaries for on-line CEC/MS. First, the on-line CEC/MS of warfarin was conducted on each column for 15 consecutive runs under same conditions. Next, the short-term noise of each run was determined for a selected time range between 0 and 20 min by the Agilent Chemstation software (Version 10.02) using three different options: (1) six times the standard deviation of the linear regression of the drift method (6 × SD), (2) peak-to-peak method (peak-topeak), and (3) the ASTM E 685-93 method (American Society for Testing and Materials). This was followed by calculation of the very short-term noise for each run using four randomly chosen regions of 2 min each from the selected time range of 0-20 min using the ASTM method. For statistical evaluation of the results, Microsoft Office Excel 2003 was utilized. The Student’s t-test was applied for testing the difference of means with significance level set to P < 0.05. RESULTS AND DISCUSSION In our previous work, it was found that the use of externally tapered columns for (()-warfarin analysis could significantly improve the reproducibility of retention time when utilizing reversed-phase mode of CEC/MS. However, the same externally tapered column was found to be unsuccessful for simultaneous enantioseparation of eight β-blockers used as model test analytes in the polar organic mode of CEC/MS. This was due to poor durability and breakage of the fragile externally tapered end during voltage ramping or arcing. This problem was even more severe with the use of the polar organic mobile phase, which tends to generate higher current. In this work, we developed a novel fabrication procedure to prepare the internally tapered column. By adopting this procedure outlined in the Experimental Section, internally tapered capillaries were prepared and packed with vancomycin CSP then utilized for the enantioseparation of eight β-blockers (electropherograms not shown). Thus, we were able to perform CEC/MS using the polar organic mode with internally tapered columns, which was not previously possible using the externally tapered columns. In addition, the problems associated with externally tapered columns, such as poor durability and taper breakage, were successfully overcome. In the following discussion, first the effects of electrospray stability of using internally tapered column versus externally tapered column are evaluated. Next, the run-to-run and column-to-column reproducibility of internally Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
1325
Table 1. Comparison of Short-Term Noise and Very Short-Term Noise for Warfarin Enantioseparation in Reversed-Phase Mode Using an Externally Tapered Column vs an Internally Tapered Columna
noise type
noise calculation method
6 × SD (n ) 15) peak-to-peak (n ) 15) ASTM (n ) 15) very short-term ASTM (n ) 60) short-term
av noise (% RSD) externally internally tapered tapered 397 (29) 641 (31) 397 (29) 42 (35)
177 (43) 211 (31) 75 (21) 7 (37)
P valueb 1.18 × 10-6 2.0 × 10-7 9.87 × 10-9 5.28 × 10-27
a Conditions and noise determinations are described in the Experimental Section. b Observed P value as a one-tail t-test. Significance level set to P < 0.05.
tapered column is discussed under both reversed-phase and polar organic phase modes of CEC/MS. Finally, the quantitation capability of the internally tapered CEC/MS column is evaluated with simultaneous enantioseparation and quantitation of four β-blockers as well as quantitation of a minor enantiomer at 0.1% level. Electrospray Stability Test (Comparison of Noise Level). The maintenance of a stable electrospray is critical for achieving high sensitivity of CEC/MS because unstable electrospray often causes high background noise. To our knowledge, there is no report on studying the effects of noise level by using different shapes of a tapered capillary for CEC/MS. As mentioned earlier, the enantioseparation of the β-blockers using polar organic mode on externally tapered capillaries was unsuccessful. Therefore, in this section, the enantioseparations of (()-warfarin under the reversed-phase condition were conducted on externally tapered and internally tapered capillaries to compare their effects on the noise level. Two types of noise, namely, short-term noise with a selected range of 20 min (calculated by 6 × SD, peak-to-peak,
and ASTM methods) and very short-term noise with a selected region of 2 min were investigated. As shown in Table 1, the average short-term noise values of the externally tapered capillary were 2-5 times higher, as compared to the internally tapered one. Analysis of the t-test revealed that the difference between the two means is significant, as illustrated by extremely small P values (last column, Table 1). Furthermore, the average very short-term noise was at least 5 times lower with the internally tapered column, as compared to the externally tapered one (Table 1). In addition, the on-line CEC/MS electropherogram in Figure 2a shows a number of negative spikes using an externally tapered column. It seems that these spikes are related to the arcing caused by the sharp edge of the external taper. In contrast, such spikes are not observed in the electropherogram using the internally tapered capillary (Figure 2b). This is because the much smoother and larger surface of the internal taper prepared by high-temperature treatment (instead of sanding) significantly reduces the chance of arcing. Thus, the use of internally tapered columns provides the excellent electrospray stability needed for on-line CEC/MS. Reproducibility Test. In our previous study for simultaneous separation of warfarin and coumachlor, we demonstrated that the run-to-run reproducibility could be significantly improved when untapered columns were replaced by externally tapered columns.8 For comparison, the applicability and reliability of the internally tapered column was first validated with the CEC/MS separation of (()-warfarin using the same packing material and reversedphase chromatographic settings for a period of eight consecutive days.8 The average retention time, efficiency, selectivity, and resolution values are summarized in Table 2. The column demonstrated impressive intraday and interday reproducibility for the retention times for over 200 injections. It is worth mentioning that these RSD for migration times obtained from internally tapered capillaries are much less (i.e., intraday RSD ) 1.6-3.4%, Table 2) than those achieved using externally tapered capillaries (intraday RSD ) 5-6%).8 In addition, the column maintained high efficiency, enantiomeric selectivity, and resolving power during the entire validation. Overall, it clearly demonstrates the stability
Figure 2. On-line CEC/MS electropherograms showing the baseline noise between 0 and 20 min of the enantioseparation of (()-warfarin using (a) externally tapered capillary and (b) internally tapered capillary. The conditions are same as described in the Experimental Section. 1326
Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
Table 2. Intraday and Interday Reproducibility of Retention Times, Separation Efficiency, Resolution, Selectivity, and Peak Areas for Warfarin Enantiomers in Reversed-Phase Mode Using an Internally Tapered Column Packed with (3R, 4S) Whelk-O1 CSP for 8 Consecutive Daysa av tR, min (% RSD) day no.
run
1
25
2
25
3
25
4
25
5
25
6
25
7
25
8
25
overall a
av efficiency (% RSD)
(S)-warfarin
(R)-warfarin
29.3 (3.2) 28.3 (1.9) 28.0 (2.5) 28.2 (2.9) 27.5 (2.3) 28.4 (2.5) 27.4 (1.6) 27.6 (1.6) 28.1 (3.2)
31.5 (3.4) 30.3 (2.0) 30.0 (2.6) 30.2 (3.1) 29.4 (2.4) 30.5 (2.5) 29.3 (1.7) 29.5 (1.7) 30.1 (3.3)
200
(S)-warfarin
av peak area (% RSD)
(R)-warfarin
34 000 (25) 26 000 (14) 26 000 (14) 30 000 (16) 31 000 (16) 27 000 (11) 25 000 (8.4) 26 000 (12) 28 000 (19)
av resolution
av selectivity
(S)-warfarin
(R)-warfarin
3.4 (14) 2.8 (6.1) 2.8 (6.4) 3.0 (8.3) 3.0 (8.2) 2.9 (6.4) 2.8 (4.3) 2.8 (6.5) 2.9 (10)
1.075 (0.219) 1.071 (0.114) 1.070 (0.132) 1.071 (0.165) 1.069 (0.119) 1.072 (0.137) 1.071 (0.106) 1.071 (0.100) 1.071 (0.191)
2 200 000 (28) 2 200 000 (25) 2 300 000 (28) 2 100 000 (28) 1 900 000 (18) 1 600 000 (23) 1 500 000 (19) 1 700 000 (11) 1 900 000 (28)
2 200 000 (30) 2 300 000 (27) 2 500 000 (29) 2 200 000 (30) 2 000 000 (19) 1 700 000 (22) 1 700 000 (17) 1 800 000 (12) 2 100 000 (29)
34 000 (28) 28 000 (16) 27 000 (19) 32 000 (14) 33 000 (18) 28 000 (11) 26 000 (8.5) 26 000 (15) 29 000 (20)
For CEC/MS conditions see experimental section.
Table 3. Intraday and Interday Reproducibility of Retention Times and Normalized Peak Area Ratios of β-Blocker Enantiomers in Polar Organic Mode Using an Internally Tapered Column Packed with Vancomycin CSP for 4 Consecutive Daysa av tR, min (% RSD)
av peak area ratio (% RSD)
peak 1b
peak 1′
peak 2
peak 2′
peak 6
day 1 (n ) 20) day 2 (n ) 20) day 3 (n ) 20) day 4 (n ) 20)
22.5 (2.1) 22.6 (2.6) 22.2 (2.2) 22.8 (3.1)
23.2 (2.1) 23.4 (2.7) 22.9 (2.3) 23.6 (3.2)
23.6 (2.1) 23.8 (2.7) 23.4 (2.3) 24.0 (3.3)
24.5 (2.1) 24.8 (2.8) 24.3 (2.4) 25.1 (3.5)
34.6 (1.9) 34.8 (2.4) 34.2 (2.7) 35.2 (3.5)
Intraday 36.0 (1.9) 36.4 (2.5) 35.7 (2.8) 36.9 (3.7)
(n ) 80)
22.6 (2.8)
23.3 (2.9)
23.7 (3.0)
24.7 (3.2)
34.8 (3.1)
Interdday 36.3 (3.3)
peak 6′
peak 1/peak 6
peak 2/peak 6
peak 1′/peak 6′
peak 2′/peak 6
3.07 (1.5) 3.11 (3.0) 2.99 (2.8) 3.11 (4.2)
3.04 (2.8) 3.08 (2.9) 3.01 (3.0) 3.09 (4.4)
2.42 (4.0) 2.44 (2.7) 2.27 (3.3) 2.32 (4.7)
2.43 (3.5) 2.44 (4.5) 2.26 (2.4) 2.27 (4.5)
3.07 (3.1)
3.06 (3.2)
2.37 (4.6)
2.36 (5.0)
a Conditions are the same as those described in the Experimental Section. b Peak 1, (S)-oxprenolol; peak 1′, (R)-oxprenolol; peak 2, (S)-alprenolol; peak 2′, (R)-alprenolol; peak 6, (S)-talinolol; peak 6′, (R)-talinolol.
of the internally tapered column under reversed-phase CEC/MS conditions. Next, the performance and reproducibility of the internally tapered column was evaluated under polar organic conditions, which was not previously possible with the external taper, as mentioned earlier. A column packed with vancomycin CSP was utilized to conduct 20 injections of eight β-blockers on a daily basis for four consecutive days (i.e., a total of 80 injections were performed on this column). The retention times and relative standard deviations of three enantiomeric pairs (six peaks) are shown in Table 3. Acceptable RSD values were found with respect to the retention times for intraday precision that ranged from 1.9 to 3.7%, and the interday precision of retention times as the mean of 4 days ranged between 2.8 and 3.2%. Thus, an internally tapered column was able to maintain a stable performance using a harsh polar organic mobile phase, even after 80 injections. The significantly extended lifetime of the internally tapered column under
polar organic conditions suggests much better ruggedness, as compared to externally tapered columns. This is probably due to the surrounding thick capillary wall of the internally tapered section that serves to prevent the tip from accidental breakage that might occur during column installation as well as arcing in the spray chamber. To study the batch-to-batch column reproducibility, four internally tapered columns packed with vancomycin CSP were utilized to conduct 20 injections of eight β-blockers on each column. The corresponding retention times and % RSD values of six peaks are listed in Table 4. The intracolumn RSD values for retention times were found to range from 0.9 to 2.7%, whereas the intercolumn RSD values of retention times as the mean of four columns were between 3.8 and 4.1%. It should be noted that the inner diameter of the internal taper opening significantly affects the analytes retention. By adopting our reproducible fabrication procedure, we were able to precisely control the inner diameter Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
1327
Table 4. Intracolumn and Intercolumn Reproducibility of Retention Times and Normalized Peak Area Ratios of β-Blocker Enantiomers in Polar Organic Mode Using Four Internally Tapered Columns Packed with Vancomycin CSPa av tR, min (% RSD)
av peak area ratio (% RSD)
peak 1b
peak 1′
peak 2
peak 2′
peak 6
column 1 (n ) 20) column 2 (n ) 20) column 3 (n ) 20) column 4 (n ) 20)
21.9 (2.0) 23.7 (1.2) 22.6 (2.6) 22.4 (1.1)
22.5 (2.0) 24.5 (1.2) 23.4 (2.7) 23.1 (1.2)
23.0 (1.7) 24.8 (1.2) 23.8 (2.7) 23.5 (1.1)
23.9 (1.4) 25.8 (1.1) 24.8 (2.8) 24.4 (1.1)
33.9 (1.8) 36.2 (0.9) 34.8 (2.4) 34.6 (1.5)
Intracolumn 35.4 (1.8) 37.7 (0.9) 36.4 (2.5) 36.0 (1.5)
(n ) 80)
22.8 (4.1)
23.5 (4.0)
23.9 (4.1)
24.8 (4.1)
35.1 (3.8)
a
peak 6′
peak 1/peak 6
peak 2/peak 6
peak 1′/peak 6′
peak 2′/peak 6
3.11 (4.2) 3.05 (3.6) 3.03 (2.2) 3.00 (2.3)
3.09 (4.4) 3.03 (4.3) 3.02 (1.7) 2.99 (2.4)
2.32 (4.7) 2.07 (2.4) 2.13 (3.2) 2.12 (1.0)
2.27 (4.5) 2.10 (2.8) 2.12 (2.4) 2.12 (1.7)
Intercolumn 36.6 3.04 (3.9) (3.4)
3.03 (3.4)
2.16 (5.8)
2.16 (4.7)
Conditions are the same as those described in the Experimental Section. b Peak identifications are the same as those described in Table 3.
of the internally tapered capillaries within the range of ∼7 to 10 µm; thus, the column-to-column reproducibility was improved as compared to grinding the opening of the external taper, which was very hard to control. In addition, we studied the reproducibility of peak areas of individual warfarin enantiomers and investigated the possibility of using them for the quantitative study. As shown in Table 2 (columns 9-10), the precisions of the warfarin peak areas were relatively poor. However, for β-blockers, the use of normalized peak areas (by taking the peak area ratio of each individual enantiomer vs the talinolol enantiomers) showed significant improvement on day-to-day and column-to-column reproducibility (Tables 3, 4; columns 8-11). Thus, for the quantitative study of β-blockers, talinolol enantiomers were added as internal standards, and the normalized peak areas were utilized to plot the calibration curves for each enantiomer. Quantitative Analysis of Enantiomers. In this set of experiments, we examined the quantitative applicability of the developed CEC/MS method using internally tapered columns. This includes simultaneous setup of calibration curves and measurement of the limit of detection (LOD) for a mixture containing eight β-blocker enantiomers (used as model test analytes). In addition, the determination of trace amounts of enantiomeric impurity was investigated. The standard solutions of four racemic mixtures of β-blockers, namely, (()-atenolol, (()-metoprolol, (()-oxprenolol, and (()-propranolol at five concentration levels were prepared for conducting simultaneous enantioseparation and setting up the calibration curves for individual enantiomers with (()-talinolol as internal standard. As shown in Figure 3a-b, all the calibration curves showed good linearity (R2 ) 0.9940-0.9988) over a wide concentration range (3-600 µM). Furthermore, at the lowest concentration (i.e., 3 µM), the average S/N ratio of atenolol enantiomers was ∼10, whereas the other three earlier-eluting β-blockers [(()-metoprolol, (()-oxprenolol, and (()-propranolol] provided much higher average S/N ratios, ranging from 25 to 65. The electropherograms shown in Figure 3c illustrate the limit of detection (at S/N ∼ 2.5) of atenolol enantiomers to be as low as 30 nM. On the other hand, the three earlier-eluting β-blockers [(()-metoprolol, (()-oxprenolol and (()-propranolol] could pro1328 Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
vide even lower LOD, because the average S/N ratios for these β-blockers ranged from 10 to 20. This low LOD is associated with high efficiency as well as on-column stacking in CEC. Furthermore, it is also due to the fact that the problems associated with electrospray stability are eliminated with the use of internally tapered columns. Since the high-temperature treatment (instead of sanding) provides smooth edges to the capillary tip, less arcing was observed in the spray chamber, resulting in much-reduced background noise. As a result, S/N ratios were enhanced, and sensitive detection was achieved. Our result in this study shows that chiral CEC/MS could provide sensitivity comparable to if not better than chiral LC/MS. Finally, internally tapered CEC/MS capillaries were applied for the determination of trace amounts of enantiomeric impurity. According to U.S. Food and Drug Administration (FDA) and the International Conference on Harmonization (ICH) Guidance for Industry Q3A Impurities in New Drug Substances, a chiral assay with a LOD of 0.1% enantiomeric impurity is mandatory for the later stages of drug development.17 The electropherogram in Figure 3d clearly shows that this level of sensitivity was achievable for the chiral reorganization of (S)-propranolol in the presence of a high concentration of (R)-propranolol. In this experiment, 300 µM (R)-propranolol was spiked with 0.3 µM (S)-propranolol. Due to high efficiency and resolution in CEC, the spiked minor enantiomer ((S)-propranolol) was well-separated from the major enantiomer ((R)-propranolol), even at 1:1000 levels (Figure 3d inset). In addition, note that the obtained peak area ratio of two enantiomers provided good correlation with the spiked concentration ratio (Figure 3d). CONCLUSIONS A novel procedure for fabricating CEC/MS internally tapered columns was demonstrated in this study. As compared to the previously reported procedure, our method is easy to handle, with no requirement for expensive equipment to produce tapers. Several advantages of the proposed method for fabricating internally tapered capillaries should be highlighted. First, our internal tapers are prepared without further grinding. Thus, we (17) USFDA. ICH Guidance for Industry Q3A Impurities in New Drug Substances; February 2003.
Figure 3. Calibration plots (a-b), electropherograms (c) at limit of detection (LOD) of four β-blockers [(()-atenolol, (()-metoprolol, (()oxprenolol, and (()-propranolol] along with internal standards ((S)-talinolol and (R)-talinolol). The electropherogram (d) shows the analysis of the nonracemic mixture of propranolol (300 µM (R)-propranolol spiked with 0.3 µM (S)-propranolol). The conditions are the same as those described in the Experimental Section, except that electrokinetic injection was performed by applying 5 kV for 60 s.
are able to prevent the taper breakage and debris accumulation in the open channel during grinding. In addition, the hightemperature treatment results in a smooth surface on the tapered
end; therefore, the tapered capillary shows enhanced compatibility for MS, as compared to using externally tapered capillaries. As a result of reduced short-term and very short-term noise levels, Analytical Chemistry, Vol. 78, No. 4, February 15, 2006
1329
higher sensitivity is also achieved with internally tapered columns. Second, the internally tapered capillaries show excellent compatibility for CEC/MS enantioseparations using different stationary phases under reversed-phase and polar organic modes, especially the latter, which is considered very problematic when externally tapered columns are used. Thus, the applicability of CEC/MS is broadened for different classes of analytes in both positive and negative ion modes of ESI-MS. The validated CEC/MS method is selective, convenient, precise (