- Email: [email protected]
Packing procedures for high efficiency, short ion-exchange columns for rapid separation of inorganic anions
Journal of Chromatography A, 1208 (2008) 95–100
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Packing procedures for high efficiency, short ion-exchange columns for rapid separation of inorganic anions Éadaoin Tyrrell a , Emily F. Hilder a , R. Andrew Shalliker b , Greg W. Dicinoski a , Robert A. Shellie a , Michael C. Breadmore a , Christopher A. Pohl c , Paul R. Haddad a,∗ a
Australian Centre for Research on Separation Science, School of Chemistry, University of Tasmania, Private Bag 75, Hobart, TAS 7001, Australia Australian Centre for Research on Separation Science, School of Natural Sciences, University of Western Sydney, Locked Bag 1797, South Penrith Distribution Centre, NSW 1797, Australia c Dionex Corporation, Sunnyvale, CA 94088, USA b
a r t i c l e
i n f o
Article history: Received 18 April 2008 Received in revised form 13 August 2008 Accepted 18 August 2008 Available online 20 August 2008 Keywords: Ion chromatography Inorganic anions Short columns Fast separations Packing procedures
a b s t r a c t An optimised packing procedure for the production of high efficiency, short, particle-packed ion-exchange columns is reported. Slurry-packing techniques were applied to a series of interconnected short columns, with the columns situated intermediate between the inlet and outlet ends of the series being used for separations. The fast separation and determination of inorganic anions was achieved using short (4 mm ID, 30 mm long) columns packed with Dionex AS20 high-capacity anion-exchange stationary phase. Seven inorganic anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) are separated in 2.6 min using a hydroxide gradient and a flow-rate of 1.8 mL/min (total analysis time including reequilibration was 3.5 min). Under isocratic conditions, the home-packed columns exhibited efficiency values of 43,000 N/m for chloride at a flow-rate of 0.3 mL/min, compared to 54,000 N/m for a commercial 250 mm AS20 column at the same flow-rate. However, the short columns gave approximately a threefold higher sample throughput. The short, home-packed columns could be produced reproducibly and gave consistent performance over extended periods of usage. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
1. Introduction Since its initial development in the 1970s, ion chromatography (IC) has become a popular chromatographic technique for many routine analyses and investigations in areas such as the analysis of chemical products, natural and processed waters, pharmaceuticals, foods, etc. [1]. An important factor for any analytical procedure is the reduction of analysis times so that the required information is obtained as quickly as possible. IC is no exception, and rapid IC systems were first developed in the 1980s for the separation of some common anions, including fluoride, chloride, nitrite, phosphate, bromide, nitrate and sulfate, with separation of these species being achieved in 9 min using a particle-packed column [2]. Since then, considerable progress has been made in the area of particle-packed columns and much faster separations are now possible. Commercially, a wide range of high-performance columns is now available, including the Dionex range of IonPac columns, which can achieve separations for the same set of seven anions in less than 3 min [3].
∗ Corresponding author. Fax: +61 3 6226 2858. E-mail address: [email protected] (P.R. Haddad).
Some of the approaches that have been used to achieve faster separations in IC include increasing the flow-rate (especially for monolithic columns) [4,5] and reducing the length of the analytical column (especially when smaller particles are used for column packing) [6,7]. While increasing the flow-rate can be effective in reducing analysis times, it is not practical with longer particulate columns due to rapid increases in back-pressure associated with higher flow-rates. The use of smaller diameter packings can improve chromatographic performance, but they exacerbate the pressure limitations associated with higher flow-rates. Short columns offer a potentially attractive approach to rapid separations in that higher flow-rates can be attained at moderate back-pressures, but on the other hand separation efficiency is often compromised because of the limited number of theoretical plates normally achievable with such columns. Nevertheless, short columns also offer the advantages of faster method development and reduced reagent consumption and waste generation, and for these reasons their use is often favoured for the development of fast separations. There are relatively few short columns available commercially for fast IC separations, with the majority falling into the class of guard columns which are not designed for highly efficient separations and therefore commonly use larger particle sizes and are packed under non-ideal conditions. As a result, if short col-
0021-9673/$ – see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.08.056
96
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
Fig. 1. Schematic of manifold used for slurry packing of short analytical columns, where (a) 4 mm × 100 mm PEEK column body, (b) 4 mm × 50 mm PEEK column body, (c) 4 mm × 30 mm PEEK column body and (d) 4 mm × 30 mm PEEK column body.
umn formats are required, it often becomes necessary for these to either be home-packed or for special orders to be placed with column manufacturers. A specific goal of the present study was the development of a rapid analysis (less than 5 min) to be used for the separation of seven target inorganic anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) which are known to be present as components of homemade inorganic explosives used in terrorist attacks. This separation was required as a confirmatory method of analysis to complement a high-speed electrophoretic separation of these same anions currently under development for application in the pre-blast detection of improvised inorganic explosives at screening points. Attention was focused on the development of short columns packed with 7.5 m diameter Dionex AS20, high-capacity, hydroxide-selective, anionexchange stationary phase because of the ability of this material to achieve rapid elution of perchlorate [8]. This analyte normally exhibits strong interaction with anion-exchangers, leading to relatively long retention times. However, no suitable short columns containing this stationary phase were available commercially and reliable home-packing procedures were therefore required. High performance chromatography columns can be prepared by dry packing [9] or slurry packing procedures [10]. Dry packing is relatively easy to perform but is unsuitable for particles having a particle diameter of less than 25 m due to aggregation of particles, leading to inefficient and unstable column beds [9]. The slurry packing method, in which the particles are suspended in a suitable liquid during the packing process, tends to produce more efficient and reproducible columns when smaller particles are used [11]. The slurry packing method has been used in this work to produce short AS20 analytical ion-exchange columns suitable for operation under high flow-rates and high back-pressure. New pack-
ing procedures are reported and the resultant columns are shown to give efficient separation of the above seven target analytes in under 3 min through the use of gradient elution and elevated flowrates. 2. Experimental 2.1. Instrumentation A Dionex ICS-2000 Ion Chromatographic instrument controlled using Chromeleon® software (version 6.80) was used during this work. All the instrumental components were obtained from Dionex (Sunnyvale, CA, USA) unless stated otherwise. The separation was carried out using a range of polymeric AS20 analytical columns of varying length and diameter, and suppressed conductivity detection was used to monitor the eluted analytes. An in-line filter pre-column (part number A-356/A-701, Upchurch Scientific), consisting of a PEEK filter with a 0.5-m frit was placed immediately in front of the column in order to prolong the lifetime of the analytical column. The IC system used a reagent-free eluent generator with an EGC II KOH cartridge to generate potassium hydroxide eluents of the required compositions for isocratic or gradient separations. A continuously-regenerated anion trap column (CR-ATC, <100 L void volume) was employed to remove trace contaminants from the eluent. Post-column eluent suppression was carried out using an anion self-regenerating suppressor (ASRS-ULTRA II 4 mm, <50 L void volume), which provided low background and enhanced analyte conductivity detection. The IC system was fitted with a 25-L sample loop that was used to introduce the sample via an AS autosampler. Chromatographic data were collected at 10 Hz and chromatograms were processed using the Chromeleon® software. Before use, all of the vials for use with the autosampler were rinsed thoroughly with deionised water.
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
97
2.2. Reagents
2.4. Measurement of extra-column broadening
All chemicals used were of analytical reagent grade and were used as supplied by Sigma–Aldrich (Sydney, Australia) unless stated otherwise. The eluent and standard solutions were prepared using deionised 18.2 M water from a Millipore Milli-Q water purification system (Bedford, MA, USA). Working standards were prepared from 1000 mg/L stock standard solutions. The chloride, chlorate, nitrate and perchlorate standard solutions were all prepared from their respective sodium salts, while the standard solutions of bromate, sulfate and chromate were prepared from their potassium salts. All of the standard solutions were filtered through a 0.45-m nylon filter and degassed in an ultrasonic bath. The chromatographic stationary phase used was the Dionex AS20 anion exchange material (7.5-m diameter), which is a hyperbranched anion-exchange polymer electrostatically attached to a polymeric surface-sulfonated substrate.
The extra-column band variance was expected to influence the efficiency of each column to different extents. It was therefore necessary to measure the extra-column band broadening for each individual column length. This was performed for each system by simply removing the column from the system and recording the peak profile at each individual flow-rate. The contribution due to the connectors, injector, detector, suppressor and trap column remained constant for each column, while the contribution due to connecting tubing varied for each different column length. To take into account the effects of these contributions, the extra-column variance was calculated at the individual flow-rates for each of the systems and was then subtracted from the total band variance.
2.3. Column packing procedures A schematic diagram of the system used for the slurry packing of the analytical columns used in this work is shown in Fig. 1. The slurry mixture itself was prepared using 15% (w/v) of packing material in a slurrying solvent consisting of appropriate volumes of acetic acid, ethylenediamine and polyethyleneglycol mono(nonylphenyl) ether in deionised water. The slurry was stirred for 10 min, then placed in an ultrasonic bath for 10 min, followed by a further 10 min stirring, after which it was poured into the reservoir of the packing assembly that was connected to the column bodies. Each chromatographic column was packed in PEEK (polyetheretherketone) column bodies (4 mm ID, 7 mm OD). Four empty column bodies were connected in series so that the packed bed could be divided into shorter analytical columns after the packing process was complete. The column bodies (formed by machining threads into appropriate lengths of PEEK tubing) were joined end-to-end using custom-made stainless steel unions (Central Science Laboratory, Hobart, Australia) to form a longer column. The inner bore of these zero dead-volume stainless steel unions was specifically machined to match the thread and OD of the PEEK column body. A Haskel 40102 air driven amplification pump (Haskel, Brisbane, Australia) used in conjunction with a standard cylinder of air was employed to pack the columns at 4000 psi with Milli-Q water used as the driving solvent. During the packing process, the manifold was configured such that there was an initial build up of pressure at the valve located just before the slurry reservoir, after which the valve was opened, forcing the slurry downwards from the slurry reservoir into the series of empty column bodies. A porous bed support or retaining frit was placed at the outlet of the final column body to hold the stationary phase packing material in place and allow the driving liquid to pass through. It was important to ensure that both the particles and the column housing could withstand the high pressures that were required for forming homogeneous packed beds. After the packing process was completed, the pressure within the columns was allowed to dissipate (∼1 h) and then the column bodies were separated and a porous retaining frit was placed at both ends of each prior to capping with suitable end-fittings. If the pressure is not allowed to relax, the packing material can be observed to extrude from the column. It is necessary therefore to ensure sufficient time to avoid this extrusion. It is important to note that capping of newly packed columns was performed as quickly as possible after separating the individual columns in order to prevent further loss of longitudinal compression of the packed bed.
3. Results and discussion 3.1. Preliminary investigations Identification of the most suitable eluent/stationary phase combination for separation of the target analytes in the shortest possible time was undertaken using Virtual Column® simulation software [12]. This software enabled a systematic search of the retention characteristics of a wide range of eluents used with commercial columns and permitted the separation selectivities of these stationary phases to be evaluated These simulations showed that the Dionex AS20 stationary phase used in conjunction with hydroxide eluents (both isocratic and gradients) provided greatest potential for the rapid separation of the target analytes. This stationary phase is porous, of high ion-exchange capacity and the 7.5 m particle size is advantageous in reducing column back-pressures at higher flow-rates. The performance of a 4 mm × 250 mm IonPac® AS20 analytical column was optimised with respect to separation selectivity and speed of analysis for the target anions using the Virtual Column® simulation software and the separation achieved when these optimal conditions were applied in practice is shown in Fig. 2. It is important to note that the flow-rate used (1.4 mL/min) was the maximum that could be used without exceeding the back-pressure limitation of the column. In addition, the gradient used to accelerate the elution of perchlorate was the steepest that could be applied in view of the fact that the highest concentration of hydroxide that could be produced by the eluent generator at a flow-rate of 1.4 mL/min was 68 mM. 3.2. Reduction of column length The key to faster separations was to use shorter versions of the AS20 column so that higher flow-rates could be used without concomitant problems with increased back-pressure. However, decreasing column length also decreases the number of theoretical plates and hence also the peak capacity. A 4 mm × 50 mm guard column packed with 11 m AS20 stationary phase was available commercially and was evaluated for high speed analysis in combination with the in-line filter used as a pre-column. Using an optimised gradient profile and a flow-rate of 1.8 mL/min, the separation shown in Fig. 3 was obtained. Under these conditions, the analysis time was reduced to under 5 min, with a minimum resolution value of 1.04. However, the peak shapes were generally poorer than those obtained on the longer analytical column, with broader peaks and tailing of later eluted peaks being observed. The efficiency of a column is characterised by calculating the number of theoretical plates (N) using the retention time (tr ) and standard
98
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
Fig. 2. Optimised separation of the seven target anions carried out on an AS20 4 mm × 250 mm commercial analytical column after blank signal subtraction. Peak: 1 = bromate, 2 = chloride, 3 = sulfate, 4 = chlorate, 5 = nitrate, 6 = chromate, 7 = perchlorate. Chromatographic conditions: IonPac AS20 4 mm × 250 mm analytical column and AG20 pre-column filter; injection = 25 L, 5 mg/L; flowrate = 1.4 mL/min; temperature = 30 ◦ C; detection = suppressed conductivity (ASRSULTRA II 4 mm, current 248 mA); eluent = potassium hydroxide gradient: 30–32 mM from 0.0 to 1.1 min, 32–45 mM from 1.1 to 1.4 min, 45–50 mM from 1.4 to 2.2 min, 50–68 mM from 2.2 to 2.5 min, 68 mM from 2.5 to 8.0 min, 30 mM from 8.0 to 12.0 min.
deviation () of the chromatographic peaks as shown in the following equation: N=
t 2 r
(1)
From this the efficiency (expressed as N/m) was calculated. An efficiency of only 18,000 N/m was obtained for the chloride ion, compared to 54,000 N/m on the 4 mm × 250 mm AS20 analytical
Fig. 3. Optimised gradient separation of the seven target anions carried out on a 4 mm × 50 mm AG20 commercial guard column (blank signal subtraction carried out). Peak identities as in Fig. 2 Chromatographic conditions: AG20 4 mm × 50 mm guard column and in-line filter pre-column; injection = 25 L, 5 mg/L; flow-rate = 1.8 mL/min; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); eluent = potassium hydroxide gradient: 0.4 mM from 0.0 to 1.0 min, 0.4–1 mM from 1.0 to 2.0 min, 1–6 mM from 2.0 to 7.0 min, 0.4 mM from 7.0 to 9.0 min.
column. The difference in efficiency can be attributed both to the increased particle size of the stationary phase used in the guard column and also to the use of sub-optimal packing procedures. Attempts were therefore made to pack short columns which exhibited approximately the same efficiency (in N/m) as longer analytical columns. A number of variables affect the quality of packed column beds when slurry packing methods are used, including the slurry composition, the packing procedure itself (especially the velocity at which the particles are delivered to the column), and the characteristics of the column body or housing [10,13,14]. A column packing manifold was designed (see Fig. 1) to allow high efficiency columns to be prepared by selecting a portion of the packed bed which fell in the middle of the bed rather than at the ends. To achieve this goal, the stationary phase was packed into a series of short column bodies connected in series, such that the final packed bed could be subdivided in order to evaluate regions where the bed was most homogenous, leading to short columns of highest efficiency. A slurry was prepared by completely suspending the particles in a suitable liquid (see Section 2) to ensure all the particles were freely dispersed and aggregates were not present. During the packing process, the column assembly was first packed with stationary phase and then the packed bed was allowed to stabilise by permitting the applied pressure to dissipate over an extended period of time. In slurry-packed columns the section of the bed at the inlet where the slurry is introduced (i.e. the uppermost section in Fig. 1) is often the least efficiently packed region [13] because the particles that build up in this section do so with the lowest packing velocity. As the bed builds up during the packing procedure, pressure increases and flow velocity decreases, such that packing density generally decreases from the column outlet to the column inlet. Furthermore, the inlet section is also subjected to the effects of direct contact from the flow of slurrying solvent since there is no inlet frit present. Flow channelling can therefore occur in the inlet, which further decreases efficiency. The outlet section of the column is also often poorly packed because the particles impinge on the outlet retaining frit with high velocity, causing a build up of axial pressure due to the compression of the particles during the packing process which can then create voids as the packing pressure is released [10]. Taking into consideration these effects, together with wall effects which have been well documented [14], it is evident that the central sections of the column are likely to be packed most uniformly. The shortest column segment packed in this work was 30 mm in length due to the thread length limitations of the end fittings that were required to cap a 4-mm ID column. In this study we used a 4 mm × 100 mm column body joined to a 4 mm × 50 mm column body, joined to two 4 mm × 30 mm column bodies (see Fig. 1). The performance of the individual column segments was then evaluated. Fig. 4 compares the outlet segment (column (d) in Fig. 1) with the segment located immediately prior (column (c) in Fig. 1). Column (d) performed poorly and even when the flow-rate and gradient profile were optimised for this column, 5 min was required to achieve a separation in which bromate and chloride were poorly resolved. On the other hand, a superior separation was obtained on column (c) in 2.6 min. Van Deemter plots for chloride were determined using a 30-mM potassium hydroxide eluent on the three columns over the flowrate range 0.1–1.95 mL/min, and the results are shown in Fig. 5. Because of the disparity in size between the three columns, extracolumn band variance can be expected to influence the observed HETP values to different extents. For this reason, the extra-column variance was measured for each system and was subtracted from the total band variance to give the results shown in Fig. 5. The columns were examined for efficiency at 0.3 mL/min. This was the
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
99
Fig. 5. Van Deemter plots for the 30 mm, 50 mm (guard) and 250 mm (analytical) column, measured for chloride ion using 30 mM potassium hydroxide as eluent.
that perhaps the sample injection volume was the contributing factor in the deterioration in efficiency associated with the shorter column. The use of a smaller sample injection loop may possibly reduce this effect. Nevertheless, the short column was more suitable for the rapid separation of the target analytes because of the fact that the flow-rate could be raised to relatively high values without exceeding the pressure limitations of the column. Allowing for re-equilibration times between runs (approximately 25 s and 55 s for the 30 mm and 250 mm columns, respectively), the sample throughput for the short column was 17 samples/h compared to 6 samples/h for the 250 mm column. The reproducibility of the column packing process was also investigated. Three sets of 4 mm × 30 mm columns (columns (c) and (d) in Fig. 1) were packed under identical conditions and their performance evaluated. The average %RSD values for columns Fig. 4. Optimised gradient separation of the seven target anions carried out on AS20 4 mm × 30 mm column segments (blank signal subtraction carried out) (a) bottom 30 mm segment (column (d) in Fig. 1), (b) second last 30 mm segment (column (c) in Fig. 1). Peak identities as in Fig. 2 Chromatographic conditions: AS20 4 mm × 30 mm analytical column and in-line filter pre-column; injection = 25 L, 5 mg/L; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); (a) flow-rate = 1.1 mL/min and eluent = potassium hydroxide gradient: 7 mM from 0.0 to 1.0 min, 7–10 mM from 1.0 to 1.5 min, 10–30 mM from 1.5 to 3.2 min, 30 mM from 3.2 to 4.0 min, 7 mM from 4.0 to 7.0 min. (b) flow-rate = 1.8 mL/min and eluent = potassium hydroxide gradient: 5 mM from 0.0 to 0.5 min, 5–10 mM from 0.5 to 1.1 min, 10–30 mM from 1.1 to 2.0 min, 30 mM from 2.0 to 2.5 min, 5 mM from 2.5 to 3.5 min.
optimum flow-rate for the shorter columns, and although not the optimum for the longer column, it was not significantly different from the optimum as can be seen from the profile for the longer column shown in Fig. 5. Also, at higher flow-rates the extra-column effects were more pronounced, which is why a lower flow-rate was chosen here. At this flow-rate, the 4 mm × 30 mm column (43,000 N/m) and the 4 mm × 250 column (54,000 N/m) showed similar behaviour, while the performance of the guard column was clearly inferior to the other two columns. It is interesting to compare the overall van Deemter plots for the 30 mm and 250 mm columns. The HETP values were initially very similar, but the short column showed much more rapid deterioration in efficiency than the longer column as the flow-rate was increased. Given the extracolumn variance has already been taken into account, it is suggested
Fig. 6. Comparison of three 4 mm × 30 mm AS20 columns packed and run under identical conditions (chromatograms offset by 5 S). Peak identities as in Fig. 2 Chromatographic conditions: AS20 4 mm × 30 mm analytical column and in-line filter pre-column; injection = 25 L, 5 mg/L; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); flow-rate = 1.8 mL/min and eluent = potassium hydroxide gradient: 5 mM from 0.0 to 0.5 min, 5–10 mM from 0.5 to 1.1 min, 10–30 mM from 1.1 to 2.0 min, 30 mM from 2.0 to 2.5 min, 5 mM from 2.5 to 3.5 min.
100
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
4. Conclusions Packing procedures are reported for the production of high efficiency, short (4 mm × 30 mm) AS20 anion-exchange columns. The resultant columns show efficiencies (as N/m) which are fairly comparable to the commercially available analytical (4 mm × 250 mm), but exhibit superior speed of separation and higher sample throughput. Under optimised conditions the seven target anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) were separated in under 3 min. These short columns therefore offer numerous possibilities for high efficiency, high speed separations in IC. Acknowledgement This study was supported by the Australian Research Council through Linkage Grant LP0669302 and Federation Fellowship FF0668673 to PRH. Fig. 7. Overlaid chromatograms from stability study carried out on the home-packed 4 mm × 30 mm AS20 column (chromatograms offset by 4 S). Peak identities as in Fig. 2 chromatographic conditions as per Fig. 6.
(d) were 3.53% for retention time, 3.62% for peak area and 1.55% for back-pressure. The columns (c) were then evaluated under identical optimised chromatographic conditions and the resulting chromatograms are shown in Fig. 6. The average %RSD value for retention time was 2.66% and for peak area was 3.71%. The back-pressure across these columns was also monitored and the %RSD value for pressure was found to be 0.70% demonstrating the high reproducibility of the packing procedure. Column stability was investigated to determine the effect of elevated flow-rates used over extended periods. 30 replicate injections of a 5 mg/L mixture of the target anions were made and the chromatograms from the first, tenth and thirtieth injections are shown in Fig. 7. Performance of the column was consistent, with average %RSD values for retention times and peak areas being 4.22% and 2.04%, respectively.
References [1] P.R. Haddad, P.E. Jackson (Eds.), Ion Chromatography—Principles and Application, Elsevier, Amsterdam, 1997. [2] H. Itoh, Y. Shinbori, Bunseki Kagaku 31 (1982) T39. [3] IC and RIFC Columns, Dionex Corporation, http://www1.dionex.com/enus/columns accessories/ICcols/lp60657.html (Accessed 03.10.07). [4] J. Li, Y. Zhu, Y. Guo, J. Chromatogr. A 1118 (2006) 46. [5] B. Paull, P.N. Nesterenko, TrAC 24 (2005) 295. [6] S. Pelletier, C.A. Lucy, J. Chromatogr. A 1125 (2006) 189. [7] I.M. Mutton, Chromatographia 47 (1998) 291. [8] IonPac AS20 Anion-Exchange Column 2007, Dionex Corporation, http://www1.dionex.com/en-us/webdocs/12034 65044-01 AS20 V24.pdf (Accessed 05.06.07). [9] V.R. Meyer, Practical High-Performance Liquid Chromatography, Wiley, 1988 (Chapter 7). [10] J.J. Kirkland, J.J. DeStefano, J. Chromatogr. A 1126 (2006) 50. [11] J. Dingenen, Anal. Magazine 26 (7) (1998) M18. [12] Virtual Column Separation Simulator, Dionex Corporation, http://www1. dionex.com/en-us/page7091.html (Accessed 18.09.07). [13] V. Wong, R.A. Shalliker, G. Guiochon, Anal. Chem. 76 (2004) 2601. [14] R.A. Shalliker, B.S. Broyles, G. Guiochon, J. Chromatogr. A. 888 (2000) 1.
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Packing procedures for high efficiency, short ion-exchange columns for rapid separation of inorganic anions Éadaoin Tyrrell a , Emily F. Hilder a , R. Andrew Shalliker b , Greg W. Dicinoski a , Robert A. Shellie a , Michael C. Breadmore a , Christopher A. Pohl c , Paul R. Haddad a,∗ a
Australian Centre for Research on Separation Science, School of Chemistry, University of Tasmania, Private Bag 75, Hobart, TAS 7001, Australia Australian Centre for Research on Separation Science, School of Natural Sciences, University of Western Sydney, Locked Bag 1797, South Penrith Distribution Centre, NSW 1797, Australia c Dionex Corporation, Sunnyvale, CA 94088, USA b
a r t i c l e
i n f o
Article history: Received 18 April 2008 Received in revised form 13 August 2008 Accepted 18 August 2008 Available online 20 August 2008 Keywords: Ion chromatography Inorganic anions Short columns Fast separations Packing procedures
a b s t r a c t An optimised packing procedure for the production of high efficiency, short, particle-packed ion-exchange columns is reported. Slurry-packing techniques were applied to a series of interconnected short columns, with the columns situated intermediate between the inlet and outlet ends of the series being used for separations. The fast separation and determination of inorganic anions was achieved using short (4 mm ID, 30 mm long) columns packed with Dionex AS20 high-capacity anion-exchange stationary phase. Seven inorganic anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) are separated in 2.6 min using a hydroxide gradient and a flow-rate of 1.8 mL/min (total analysis time including reequilibration was 3.5 min). Under isocratic conditions, the home-packed columns exhibited efficiency values of 43,000 N/m for chloride at a flow-rate of 0.3 mL/min, compared to 54,000 N/m for a commercial 250 mm AS20 column at the same flow-rate. However, the short columns gave approximately a threefold higher sample throughput. The short, home-packed columns could be produced reproducibly and gave consistent performance over extended periods of usage. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
1. Introduction Since its initial development in the 1970s, ion chromatography (IC) has become a popular chromatographic technique for many routine analyses and investigations in areas such as the analysis of chemical products, natural and processed waters, pharmaceuticals, foods, etc. [1]. An important factor for any analytical procedure is the reduction of analysis times so that the required information is obtained as quickly as possible. IC is no exception, and rapid IC systems were first developed in the 1980s for the separation of some common anions, including fluoride, chloride, nitrite, phosphate, bromide, nitrate and sulfate, with separation of these species being achieved in 9 min using a particle-packed column [2]. Since then, considerable progress has been made in the area of particle-packed columns and much faster separations are now possible. Commercially, a wide range of high-performance columns is now available, including the Dionex range of IonPac columns, which can achieve separations for the same set of seven anions in less than 3 min [3].
∗ Corresponding author. Fax: +61 3 6226 2858. E-mail address: [email protected] (P.R. Haddad).
Some of the approaches that have been used to achieve faster separations in IC include increasing the flow-rate (especially for monolithic columns) [4,5] and reducing the length of the analytical column (especially when smaller particles are used for column packing) [6,7]. While increasing the flow-rate can be effective in reducing analysis times, it is not practical with longer particulate columns due to rapid increases in back-pressure associated with higher flow-rates. The use of smaller diameter packings can improve chromatographic performance, but they exacerbate the pressure limitations associated with higher flow-rates. Short columns offer a potentially attractive approach to rapid separations in that higher flow-rates can be attained at moderate back-pressures, but on the other hand separation efficiency is often compromised because of the limited number of theoretical plates normally achievable with such columns. Nevertheless, short columns also offer the advantages of faster method development and reduced reagent consumption and waste generation, and for these reasons their use is often favoured for the development of fast separations. There are relatively few short columns available commercially for fast IC separations, with the majority falling into the class of guard columns which are not designed for highly efficient separations and therefore commonly use larger particle sizes and are packed under non-ideal conditions. As a result, if short col-
0021-9673/$ – see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.08.056
96
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
Fig. 1. Schematic of manifold used for slurry packing of short analytical columns, where (a) 4 mm × 100 mm PEEK column body, (b) 4 mm × 50 mm PEEK column body, (c) 4 mm × 30 mm PEEK column body and (d) 4 mm × 30 mm PEEK column body.
umn formats are required, it often becomes necessary for these to either be home-packed or for special orders to be placed with column manufacturers. A specific goal of the present study was the development of a rapid analysis (less than 5 min) to be used for the separation of seven target inorganic anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) which are known to be present as components of homemade inorganic explosives used in terrorist attacks. This separation was required as a confirmatory method of analysis to complement a high-speed electrophoretic separation of these same anions currently under development for application in the pre-blast detection of improvised inorganic explosives at screening points. Attention was focused on the development of short columns packed with 7.5 m diameter Dionex AS20, high-capacity, hydroxide-selective, anionexchange stationary phase because of the ability of this material to achieve rapid elution of perchlorate [8]. This analyte normally exhibits strong interaction with anion-exchangers, leading to relatively long retention times. However, no suitable short columns containing this stationary phase were available commercially and reliable home-packing procedures were therefore required. High performance chromatography columns can be prepared by dry packing [9] or slurry packing procedures [10]. Dry packing is relatively easy to perform but is unsuitable for particles having a particle diameter of less than 25 m due to aggregation of particles, leading to inefficient and unstable column beds [9]. The slurry packing method, in which the particles are suspended in a suitable liquid during the packing process, tends to produce more efficient and reproducible columns when smaller particles are used [11]. The slurry packing method has been used in this work to produce short AS20 analytical ion-exchange columns suitable for operation under high flow-rates and high back-pressure. New pack-
ing procedures are reported and the resultant columns are shown to give efficient separation of the above seven target analytes in under 3 min through the use of gradient elution and elevated flowrates. 2. Experimental 2.1. Instrumentation A Dionex ICS-2000 Ion Chromatographic instrument controlled using Chromeleon® software (version 6.80) was used during this work. All the instrumental components were obtained from Dionex (Sunnyvale, CA, USA) unless stated otherwise. The separation was carried out using a range of polymeric AS20 analytical columns of varying length and diameter, and suppressed conductivity detection was used to monitor the eluted analytes. An in-line filter pre-column (part number A-356/A-701, Upchurch Scientific), consisting of a PEEK filter with a 0.5-m frit was placed immediately in front of the column in order to prolong the lifetime of the analytical column. The IC system used a reagent-free eluent generator with an EGC II KOH cartridge to generate potassium hydroxide eluents of the required compositions for isocratic or gradient separations. A continuously-regenerated anion trap column (CR-ATC, <100 L void volume) was employed to remove trace contaminants from the eluent. Post-column eluent suppression was carried out using an anion self-regenerating suppressor (ASRS-ULTRA II 4 mm, <50 L void volume), which provided low background and enhanced analyte conductivity detection. The IC system was fitted with a 25-L sample loop that was used to introduce the sample via an AS autosampler. Chromatographic data were collected at 10 Hz and chromatograms were processed using the Chromeleon® software. Before use, all of the vials for use with the autosampler were rinsed thoroughly with deionised water.
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
97
2.2. Reagents
2.4. Measurement of extra-column broadening
All chemicals used were of analytical reagent grade and were used as supplied by Sigma–Aldrich (Sydney, Australia) unless stated otherwise. The eluent and standard solutions were prepared using deionised 18.2 M water from a Millipore Milli-Q water purification system (Bedford, MA, USA). Working standards were prepared from 1000 mg/L stock standard solutions. The chloride, chlorate, nitrate and perchlorate standard solutions were all prepared from their respective sodium salts, while the standard solutions of bromate, sulfate and chromate were prepared from their potassium salts. All of the standard solutions were filtered through a 0.45-m nylon filter and degassed in an ultrasonic bath. The chromatographic stationary phase used was the Dionex AS20 anion exchange material (7.5-m diameter), which is a hyperbranched anion-exchange polymer electrostatically attached to a polymeric surface-sulfonated substrate.
The extra-column band variance was expected to influence the efficiency of each column to different extents. It was therefore necessary to measure the extra-column band broadening for each individual column length. This was performed for each system by simply removing the column from the system and recording the peak profile at each individual flow-rate. The contribution due to the connectors, injector, detector, suppressor and trap column remained constant for each column, while the contribution due to connecting tubing varied for each different column length. To take into account the effects of these contributions, the extra-column variance was calculated at the individual flow-rates for each of the systems and was then subtracted from the total band variance.
2.3. Column packing procedures A schematic diagram of the system used for the slurry packing of the analytical columns used in this work is shown in Fig. 1. The slurry mixture itself was prepared using 15% (w/v) of packing material in a slurrying solvent consisting of appropriate volumes of acetic acid, ethylenediamine and polyethyleneglycol mono(nonylphenyl) ether in deionised water. The slurry was stirred for 10 min, then placed in an ultrasonic bath for 10 min, followed by a further 10 min stirring, after which it was poured into the reservoir of the packing assembly that was connected to the column bodies. Each chromatographic column was packed in PEEK (polyetheretherketone) column bodies (4 mm ID, 7 mm OD). Four empty column bodies were connected in series so that the packed bed could be divided into shorter analytical columns after the packing process was complete. The column bodies (formed by machining threads into appropriate lengths of PEEK tubing) were joined end-to-end using custom-made stainless steel unions (Central Science Laboratory, Hobart, Australia) to form a longer column. The inner bore of these zero dead-volume stainless steel unions was specifically machined to match the thread and OD of the PEEK column body. A Haskel 40102 air driven amplification pump (Haskel, Brisbane, Australia) used in conjunction with a standard cylinder of air was employed to pack the columns at 4000 psi with Milli-Q water used as the driving solvent. During the packing process, the manifold was configured such that there was an initial build up of pressure at the valve located just before the slurry reservoir, after which the valve was opened, forcing the slurry downwards from the slurry reservoir into the series of empty column bodies. A porous bed support or retaining frit was placed at the outlet of the final column body to hold the stationary phase packing material in place and allow the driving liquid to pass through. It was important to ensure that both the particles and the column housing could withstand the high pressures that were required for forming homogeneous packed beds. After the packing process was completed, the pressure within the columns was allowed to dissipate (∼1 h) and then the column bodies were separated and a porous retaining frit was placed at both ends of each prior to capping with suitable end-fittings. If the pressure is not allowed to relax, the packing material can be observed to extrude from the column. It is necessary therefore to ensure sufficient time to avoid this extrusion. It is important to note that capping of newly packed columns was performed as quickly as possible after separating the individual columns in order to prevent further loss of longitudinal compression of the packed bed.
3. Results and discussion 3.1. Preliminary investigations Identification of the most suitable eluent/stationary phase combination for separation of the target analytes in the shortest possible time was undertaken using Virtual Column® simulation software [12]. This software enabled a systematic search of the retention characteristics of a wide range of eluents used with commercial columns and permitted the separation selectivities of these stationary phases to be evaluated These simulations showed that the Dionex AS20 stationary phase used in conjunction with hydroxide eluents (both isocratic and gradients) provided greatest potential for the rapid separation of the target analytes. This stationary phase is porous, of high ion-exchange capacity and the 7.5 m particle size is advantageous in reducing column back-pressures at higher flow-rates. The performance of a 4 mm × 250 mm IonPac® AS20 analytical column was optimised with respect to separation selectivity and speed of analysis for the target anions using the Virtual Column® simulation software and the separation achieved when these optimal conditions were applied in practice is shown in Fig. 2. It is important to note that the flow-rate used (1.4 mL/min) was the maximum that could be used without exceeding the back-pressure limitation of the column. In addition, the gradient used to accelerate the elution of perchlorate was the steepest that could be applied in view of the fact that the highest concentration of hydroxide that could be produced by the eluent generator at a flow-rate of 1.4 mL/min was 68 mM. 3.2. Reduction of column length The key to faster separations was to use shorter versions of the AS20 column so that higher flow-rates could be used without concomitant problems with increased back-pressure. However, decreasing column length also decreases the number of theoretical plates and hence also the peak capacity. A 4 mm × 50 mm guard column packed with 11 m AS20 stationary phase was available commercially and was evaluated for high speed analysis in combination with the in-line filter used as a pre-column. Using an optimised gradient profile and a flow-rate of 1.8 mL/min, the separation shown in Fig. 3 was obtained. Under these conditions, the analysis time was reduced to under 5 min, with a minimum resolution value of 1.04. However, the peak shapes were generally poorer than those obtained on the longer analytical column, with broader peaks and tailing of later eluted peaks being observed. The efficiency of a column is characterised by calculating the number of theoretical plates (N) using the retention time (tr ) and standard
98
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
Fig. 2. Optimised separation of the seven target anions carried out on an AS20 4 mm × 250 mm commercial analytical column after blank signal subtraction. Peak: 1 = bromate, 2 = chloride, 3 = sulfate, 4 = chlorate, 5 = nitrate, 6 = chromate, 7 = perchlorate. Chromatographic conditions: IonPac AS20 4 mm × 250 mm analytical column and AG20 pre-column filter; injection = 25 L, 5 mg/L; flowrate = 1.4 mL/min; temperature = 30 ◦ C; detection = suppressed conductivity (ASRSULTRA II 4 mm, current 248 mA); eluent = potassium hydroxide gradient: 30–32 mM from 0.0 to 1.1 min, 32–45 mM from 1.1 to 1.4 min, 45–50 mM from 1.4 to 2.2 min, 50–68 mM from 2.2 to 2.5 min, 68 mM from 2.5 to 8.0 min, 30 mM from 8.0 to 12.0 min.
deviation () of the chromatographic peaks as shown in the following equation: N=
t 2 r
(1)
From this the efficiency (expressed as N/m) was calculated. An efficiency of only 18,000 N/m was obtained for the chloride ion, compared to 54,000 N/m on the 4 mm × 250 mm AS20 analytical
Fig. 3. Optimised gradient separation of the seven target anions carried out on a 4 mm × 50 mm AG20 commercial guard column (blank signal subtraction carried out). Peak identities as in Fig. 2 Chromatographic conditions: AG20 4 mm × 50 mm guard column and in-line filter pre-column; injection = 25 L, 5 mg/L; flow-rate = 1.8 mL/min; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); eluent = potassium hydroxide gradient: 0.4 mM from 0.0 to 1.0 min, 0.4–1 mM from 1.0 to 2.0 min, 1–6 mM from 2.0 to 7.0 min, 0.4 mM from 7.0 to 9.0 min.
column. The difference in efficiency can be attributed both to the increased particle size of the stationary phase used in the guard column and also to the use of sub-optimal packing procedures. Attempts were therefore made to pack short columns which exhibited approximately the same efficiency (in N/m) as longer analytical columns. A number of variables affect the quality of packed column beds when slurry packing methods are used, including the slurry composition, the packing procedure itself (especially the velocity at which the particles are delivered to the column), and the characteristics of the column body or housing [10,13,14]. A column packing manifold was designed (see Fig. 1) to allow high efficiency columns to be prepared by selecting a portion of the packed bed which fell in the middle of the bed rather than at the ends. To achieve this goal, the stationary phase was packed into a series of short column bodies connected in series, such that the final packed bed could be subdivided in order to evaluate regions where the bed was most homogenous, leading to short columns of highest efficiency. A slurry was prepared by completely suspending the particles in a suitable liquid (see Section 2) to ensure all the particles were freely dispersed and aggregates were not present. During the packing process, the column assembly was first packed with stationary phase and then the packed bed was allowed to stabilise by permitting the applied pressure to dissipate over an extended period of time. In slurry-packed columns the section of the bed at the inlet where the slurry is introduced (i.e. the uppermost section in Fig. 1) is often the least efficiently packed region [13] because the particles that build up in this section do so with the lowest packing velocity. As the bed builds up during the packing procedure, pressure increases and flow velocity decreases, such that packing density generally decreases from the column outlet to the column inlet. Furthermore, the inlet section is also subjected to the effects of direct contact from the flow of slurrying solvent since there is no inlet frit present. Flow channelling can therefore occur in the inlet, which further decreases efficiency. The outlet section of the column is also often poorly packed because the particles impinge on the outlet retaining frit with high velocity, causing a build up of axial pressure due to the compression of the particles during the packing process which can then create voids as the packing pressure is released [10]. Taking into consideration these effects, together with wall effects which have been well documented [14], it is evident that the central sections of the column are likely to be packed most uniformly. The shortest column segment packed in this work was 30 mm in length due to the thread length limitations of the end fittings that were required to cap a 4-mm ID column. In this study we used a 4 mm × 100 mm column body joined to a 4 mm × 50 mm column body, joined to two 4 mm × 30 mm column bodies (see Fig. 1). The performance of the individual column segments was then evaluated. Fig. 4 compares the outlet segment (column (d) in Fig. 1) with the segment located immediately prior (column (c) in Fig. 1). Column (d) performed poorly and even when the flow-rate and gradient profile were optimised for this column, 5 min was required to achieve a separation in which bromate and chloride were poorly resolved. On the other hand, a superior separation was obtained on column (c) in 2.6 min. Van Deemter plots for chloride were determined using a 30-mM potassium hydroxide eluent on the three columns over the flowrate range 0.1–1.95 mL/min, and the results are shown in Fig. 5. Because of the disparity in size between the three columns, extracolumn band variance can be expected to influence the observed HETP values to different extents. For this reason, the extra-column variance was measured for each system and was subtracted from the total band variance to give the results shown in Fig. 5. The columns were examined for efficiency at 0.3 mL/min. This was the
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
99
Fig. 5. Van Deemter plots for the 30 mm, 50 mm (guard) and 250 mm (analytical) column, measured for chloride ion using 30 mM potassium hydroxide as eluent.
that perhaps the sample injection volume was the contributing factor in the deterioration in efficiency associated with the shorter column. The use of a smaller sample injection loop may possibly reduce this effect. Nevertheless, the short column was more suitable for the rapid separation of the target analytes because of the fact that the flow-rate could be raised to relatively high values without exceeding the pressure limitations of the column. Allowing for re-equilibration times between runs (approximately 25 s and 55 s for the 30 mm and 250 mm columns, respectively), the sample throughput for the short column was 17 samples/h compared to 6 samples/h for the 250 mm column. The reproducibility of the column packing process was also investigated. Three sets of 4 mm × 30 mm columns (columns (c) and (d) in Fig. 1) were packed under identical conditions and their performance evaluated. The average %RSD values for columns Fig. 4. Optimised gradient separation of the seven target anions carried out on AS20 4 mm × 30 mm column segments (blank signal subtraction carried out) (a) bottom 30 mm segment (column (d) in Fig. 1), (b) second last 30 mm segment (column (c) in Fig. 1). Peak identities as in Fig. 2 Chromatographic conditions: AS20 4 mm × 30 mm analytical column and in-line filter pre-column; injection = 25 L, 5 mg/L; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); (a) flow-rate = 1.1 mL/min and eluent = potassium hydroxide gradient: 7 mM from 0.0 to 1.0 min, 7–10 mM from 1.0 to 1.5 min, 10–30 mM from 1.5 to 3.2 min, 30 mM from 3.2 to 4.0 min, 7 mM from 4.0 to 7.0 min. (b) flow-rate = 1.8 mL/min and eluent = potassium hydroxide gradient: 5 mM from 0.0 to 0.5 min, 5–10 mM from 0.5 to 1.1 min, 10–30 mM from 1.1 to 2.0 min, 30 mM from 2.0 to 2.5 min, 5 mM from 2.5 to 3.5 min.
optimum flow-rate for the shorter columns, and although not the optimum for the longer column, it was not significantly different from the optimum as can be seen from the profile for the longer column shown in Fig. 5. Also, at higher flow-rates the extra-column effects were more pronounced, which is why a lower flow-rate was chosen here. At this flow-rate, the 4 mm × 30 mm column (43,000 N/m) and the 4 mm × 250 column (54,000 N/m) showed similar behaviour, while the performance of the guard column was clearly inferior to the other two columns. It is interesting to compare the overall van Deemter plots for the 30 mm and 250 mm columns. The HETP values were initially very similar, but the short column showed much more rapid deterioration in efficiency than the longer column as the flow-rate was increased. Given the extracolumn variance has already been taken into account, it is suggested
Fig. 6. Comparison of three 4 mm × 30 mm AS20 columns packed and run under identical conditions (chromatograms offset by 5 S). Peak identities as in Fig. 2 Chromatographic conditions: AS20 4 mm × 30 mm analytical column and in-line filter pre-column; injection = 25 L, 5 mg/L; temperature = 30 ◦ C; detection = suppressed conductivity (ASRS-ULTRA II 4 mm, current 168 mA); flow-rate = 1.8 mL/min and eluent = potassium hydroxide gradient: 5 mM from 0.0 to 0.5 min, 5–10 mM from 0.5 to 1.1 min, 10–30 mM from 1.1 to 2.0 min, 30 mM from 2.0 to 2.5 min, 5 mM from 2.5 to 3.5 min.
100
É. Tyrrell et al. / J. Chromatogr. A 1208 (2008) 95–100
4. Conclusions Packing procedures are reported for the production of high efficiency, short (4 mm × 30 mm) AS20 anion-exchange columns. The resultant columns show efficiencies (as N/m) which are fairly comparable to the commercially available analytical (4 mm × 250 mm), but exhibit superior speed of separation and higher sample throughput. Under optimised conditions the seven target anions (bromate, chloride, chlorate, nitrate, sulfate, chromate and perchlorate) were separated in under 3 min. These short columns therefore offer numerous possibilities for high efficiency, high speed separations in IC. Acknowledgement This study was supported by the Australian Research Council through Linkage Grant LP0669302 and Federation Fellowship FF0668673 to PRH. Fig. 7. Overlaid chromatograms from stability study carried out on the home-packed 4 mm × 30 mm AS20 column (chromatograms offset by 4 S). Peak identities as in Fig. 2 chromatographic conditions as per Fig. 6.
(d) were 3.53% for retention time, 3.62% for peak area and 1.55% for back-pressure. The columns (c) were then evaluated under identical optimised chromatographic conditions and the resulting chromatograms are shown in Fig. 6. The average %RSD value for retention time was 2.66% and for peak area was 3.71%. The back-pressure across these columns was also monitored and the %RSD value for pressure was found to be 0.70% demonstrating the high reproducibility of the packing procedure. Column stability was investigated to determine the effect of elevated flow-rates used over extended periods. 30 replicate injections of a 5 mg/L mixture of the target anions were made and the chromatograms from the first, tenth and thirtieth injections are shown in Fig. 7. Performance of the column was consistent, with average %RSD values for retention times and peak areas being 4.22% and 2.04%, respectively.
References [1] P.R. Haddad, P.E. Jackson (Eds.), Ion Chromatography—Principles and Application, Elsevier, Amsterdam, 1997. [2] H. Itoh, Y. Shinbori, Bunseki Kagaku 31 (1982) T39. [3] IC and RIFC Columns, Dionex Corporation, http://www1.dionex.com/enus/columns accessories/ICcols/lp60657.html (Accessed 03.10.07). [4] J. Li, Y. Zhu, Y. Guo, J. Chromatogr. A 1118 (2006) 46. [5] B. Paull, P.N. Nesterenko, TrAC 24 (2005) 295. [6] S. Pelletier, C.A. Lucy, J. Chromatogr. A 1125 (2006) 189. [7] I.M. Mutton, Chromatographia 47 (1998) 291. [8] IonPac AS20 Anion-Exchange Column 2007, Dionex Corporation, http://www1.dionex.com/en-us/webdocs/12034 65044-01 AS20 V24.pdf (Accessed 05.06.07). [9] V.R. Meyer, Practical High-Performance Liquid Chromatography, Wiley, 1988 (Chapter 7). [10] J.J. Kirkland, J.J. DeStefano, J. Chromatogr. A 1126 (2006) 50. [11] J. Dingenen, Anal. Magazine 26 (7) (1998) M18. [12] Virtual Column Separation Simulator, Dionex Corporation, http://www1. dionex.com/en-us/page7091.html (Accessed 18.09.07). [13] V. Wong, R.A. Shalliker, G. Guiochon, Anal. Chem. 76 (2004) 2601. [14] R.A. Shalliker, B.S. Broyles, G. Guiochon, J. Chromatogr. A. 888 (2000) 1.