Anal. Chem. 2005, 77, 1718-1726
Thermodynamic Studies on Solvent Effects in Molecularly Imprinted Polymers. 2. Concentration of the Organic Modifier Hyunjung Kim and Georges Guiochon*
Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37996-1600, and Division of Chemical Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6120
The effects of the organic modifier concentration on the isotherm parameters of the two enantiomers of Fmoctryptophan (Fmoc-L,D-Trp) on an Fmoc-L-Trp-imprinted polymer were investigated over a wide concentration range (0.005-100 mM), using frontal analysis. The modifier was acetic acid; concentrations of 0.2, 0.9, 1.7, and 3.7 M in an acetonitrile-based mobile phase were studied. At each concentration, adsorption isotherm data were acquired for each enantiomer. From these data, the isotherm parameters of each compound were derived from nonlinear isotherm fitting and the affinity energy distributions were calculated independently. We found that three types of sites coexist for Fmoc-L-Trp but only two types of sites for Fmoc-D-Trp, except at the lowest acetic acid concentration (0.2 M), at which three types of sites coexist. Increasing the acetic acid concentration decreases the selectivity and the overall affinity of both enantiomers. The overall affinity of Fmoc-L-Trp is dominated by the contribution of the low-density highest energy sites while that of Fmoc-D-Trp is dominated by the most abundant, low-energy sites. For the low-energy sites, increasing the acetic acid concentration affects the association constant of the enantiomers more than the number of corresponding sites. In contrast, for the highest energy sites (sites that exist only for Fmoc-L-Trp), increasing the concentration of acetic acid affects significantly the number of sites but hardly changes the association constant. Molecular imprinted polymers (MIPs) have attracted much attention as a promising group of chiral stationary phases (CSPs) for liquid chromatography (LC).1 MIPs are chemically and physically very stable and provide a high selectivity toward the molecule present in solution during their polymerization (template). These tailor-made CSPs guarantee an accurate prediction of the elution order and a good separation, which are hard to achieve with other CSPs, whether natural (polysaccharides, proteins) or synthetic polymers.2 The most commonly used strategy to prepare MIPs is the use of noncovalent interactions between a target molecule (the template) and some suitable * To whom correspondence should be addressed. Fax: 865-974-2667. E-mail:
[email protected]. (1) Sellergren, B. J. Chromatogr., A 2001, 906, 227-252. (2) Practical Approach To Chiral Separation By Liquid Chromatography; Bartsch, R. A.; Maeda, M., Eds.; VCH: Weinheim, 1994.
1718 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
functional groups. These interactions allow the formation of template-functional monomer complexes in solution. These complexes are then immobilized into a polymer matrix by copolymerization with a high concentration of cross-linking monomers. Complementary size, shape, and functionalities toward the template in the MIPs can be obtained by extracting the template from the polymer matrix after the end of the polymerization process. The most frequent method of optimization of separations on MIPs consists of adjusting the mobile-phase composition to improve their binding performance.3 However, in almost all cases, the data have been acquired by injecting analytical amounts of substrates on the MIPs. A rational understanding of the effect of the mobile-phase composition on the binding performance of the MIPs would facilitate this optimization. As previously demonstrated,4-9 only isotherm data acquired over a wide concentration range of the solutes can provide a detailed understanding of the effects of the experimental parameters on the different binding sites that coexist on the heterogeneous surface of MIPs. For example, in a previous study,9 we investigated the influence of the nature of the organic modifier on the isotherm parameters of the enantiomers on an Fmoc-L-Trp MIP, using isotherm data acquired by frontal analysis over a wide concentration range. We demonstrated the strong heterogeneity of the MIP surface and the existence on its surface of four different types of sites in the presence of THF, propan-2-ol, and methanol but of only three types of sites with acetic acid as the organic modifier. By acquiring isotherm data over a wide concentration range, we could provide an understanding of why the enantioselectivity and the overall affinity of the MIP decreases with increasing hydrogen-bonding donor parameter (HBD) of the organic modifier. We showed that increasing the HBD of the organic modifier decreases the density of the highest energy sites without significantly changing the adsorption constant on these sites. (3) Molecular imprinted polymers:Man made mimics of antibodies and their application in analytical chemistry; Sellergren, B. Ed.; Elsevier: Amsterdam, 2001. (4) Chen, Y.; Kele, M.; Sajonz, P.; Sellergren, B.; Guiochon, G. J. Anal. Chem. 1999, 71, 928-938. (5) Chen, Y.; Kele, M.; Quionones, I.; Sellergren, B.; Guiochon, G. J. Chromatogr., A 2001, 927, 1-17. (6) Rampey, A. M.; Umpleby II, R. J.; Rushton, G. T.; Iseman, J. C.; Shah, R. N.; Shimizu, K. D. Anal. Chem. 2004, 76, 1123-1133. (7) Umpleby II, R. J.; Baxter, S. C.; Rampey, A. M.; Rushton, G. T.; Chen, Y.; Shimizu, K. D. J. Chromatogr., B 2004, 804, 141-149. (8) Kim, H.; Guiochon, G. Anal. Chem., submitted. (9) Kim, H.; Guiochon, G. Anal. Chem., submitted. 10.1021/ac040164o CCC: $30.25
© 2005 American Chemical Society Published on Web 02/08/2005
This work reports on the influence of the concentration of acetic acid on the isotherm parameters of Fmoc-L-Trp MIP. Acetic acid is the organic modifier most commonly used to optimize the binding performance and enantioselectivity of MIPs. The influence of its concentration on the capacity factors and the selectivity of MIPs is well documented.1,3,10 It was suggested that a compensation of the roles of the selective (i.e., on the imprinted sites) and the nonselective (i.e., on the nonimprinted sites) binding sites explains the better enantiomeric separation observed with acetic acid. A very small acetic acid concentration (∼0.1 vol %) increases the enantiomeric separation by decreasing the nonselective interactions of compounds with the polymer matrix. However, a further increase of the amount of acetic acid decreases the capacity factors and the selectivity due to the increasing competition of acetic acid with the solutes for specific binding onto the MIPs. In this study, we provide evidence that acetic acid competes with the substrates for access to the nonselective binding sites, but not for access to the enantioselective binding sites. EXPERIMENTAL SECTION We briefly describe the specific experimental conditions of this study. The theoretical background on the experimental conditions were discussed in our previous study. Chemicals. Fmoc-L-Trp and Fmoc-D-Trp were purchased from Novabiochem (San Diego, CA). 4-Vinylpyridine (4-VPY), 2,2azobis(isobutyronitrile) (AIBN), and ethylene glycol dimethacrylate (EGDMA) were obtained from Aldrich (Milwaukee, WI). 4-VPY (60 mmHg, 75 °C) and EGDMA (60 mmHg, 120 °C) were distilled under vacuum. All other chemicals and solvents were commercially available, of analytical or HPLC grade, and were used as is. Preparation of the Stationary Phase and Packing of the Column. The composition of the polymerization mixture was as follows: 1.58 mmol of Fmoc-L-Trp, 4.74 mmol of 4-VPY, 18.96 mmol of EGDMA, 0.474 mmol of AIBN, and 5.4 mL of acetonitrile (MeCN). The amount of the solvent was 4/3 of the volume of the monomers and the cross-linking monomers. The solution was purged with N2 for 5 min in a scintillation vial and polymerized at 45 °C for 12 h. After the polymerization, the bulk polymers were crushed, ground, and sieved to obtain particles within the size range of 25-38 mm. These polymer particles were slurry packed into a stainless column (100 × 4.6 mm). To remove any residues from polymerization mixtures and the template from the polymers, the column packed with the imprinted polymers was exhaustively washed with methanol/acetic acid (4/1 v/v). Apparatus. The isotherm data were obtained by frontal analysis, using a Hewlett-Packard (Palo Alto, CA) HP 1090 liquid chromatograph. This instrument is equipped with a multisolvent system (tank volumes, 1 dm3 each), an autosampler with a 250µL sample loop, a diode-array UV detector, a column thermostat, and a computer data acquisition station. The microcomputer of this system was used to program a series of breakthrough curves. Experimental Measurement of the Breakthrough Curves, the Holdup, and the Extracolumn Volumes. The two pumps of the HP1090 solvent delivery system were used to obtain breakthrough curves at different mobile concentrations of the (10) Brien, T. P. O.; Snow, N. H.; Grinberg, N.; Crocker, L. J. Liq. Chromatogr. Relat. Technol. 1999, 22, 183-204.
solutes. One of the pump delivered the pure mobile phase; the other pump delivered solutions of the compound studied. The concentration of the studied compound was determined by the concentration of the sample solution and the flow rate fractions delivered by the two pumps at constant flow rate for a set time interval. The concentration range of the compound studied was approximately between 0.005 and 50 mM. However, the maximum concentration of the compound was different in each concentration of acetic acid, due to different solubility limits. The frontal analysis (FA) experiments were carried out at four different concentrations of acetic acid (0.2, 0.9, 1.7, and 3.7 M), with the same base solvent (acetonitrile). For each different mobile phase, 33 consecutive breakthrough curves were recorded, A sufficiently long delay time (60-120 min) between each successive curve allowed the reequilibrium of the column with the pure mobile phase. The injection time of the sample was between 20 and 40 min to ensure that the composition of the eluate was the same as that of the plateau injected at the column inlet. This was checked by observing the plateau of the breakthrough for more than 5 min. The signal was detected, depending on the concentration range, between 260 and 310 nm to avoid recording any signals above 1500 mAU. Three consecutive breakthrough curves are acquired in each different concentrations of acetic acid with increasing concentration solute (from 0.005 to 100 mM). The conventional wide rectangular procedure was followed. A solution of known concentration of the sample is pumped into the column until the plateau of the breakthrough is reached; after this plateau has lasted for more than 5 min, the adsorbed substrate is washed off the column with the pure mobile phase until the absorbance of the eluent has decreased to the same baseline as before. The next breakthrough is then measured. After the third breakthrough curve is recorded, the feed composition is increased for the acquisition of the next isotherm data point. Because the return of the baseline below a low threshold at the end of the reequilibration period is checked, there can be only a small amount of compound left. Because the concentration is always increased from one step to the next, the error made on the isotherm data point can only be small. Even if the isotherm parameters on the low-energy sites were slightly overestimated, they will be so for each system and this would not affect the conclusions based on a comparison of the isotherm parameters in different acetic acid concentrations. The holdup time for the MIP columns (t0) was measured by injecting a small amount of acetone into the column. The extracolumn volume from the pump was measured by injecting a small amount of the acetone from the pump into a zero deadvolume connector instead of the column. A value of tx ) 0.99 min was obtained. The experimental data have been corrected by subtracting tx. Calculation of the Isotherm Data and Isotherm Model Selection. Adsorption isotherm data represent the amount of substrate (q, mmol/L) bound to the column for each mobile-phase concentration of the substrate (C, mmol/L). To calculate this amount from the FA data, the following equation was used:
Vequ - V0 Va
q)C
where Vequ is the elution volume of the equivalent area of the Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
1719
Figure 1. Adsorption equilibrium isotherms of Fmoc-L-Trp (solid symbols) and Fmoc-D-Trp (open symbols) at different concentrations of acetic acid: 0.2 (circle), 0.9 (square), 1.7 (triangle up), and 3.7 M (triangle down). The lines represent the best calculated tri-Langmuir isotherm for Fmoc-L-Trp (solid lines) at all concentrations of acetic acid and Fmoc-D-Trp (dotted lines) at 0.2 M acetic acid. For Fmoc-D-Trp at 0.9, 1.7, and 3.7 M acetic acid, the lines represent the best calculated bi-Langmuir isotherm. Concentration range: between (a) 0 and 0.14, (b) 0 and 1.5, (c) 0 and 12, and (d) 0 and 160 mM.
solute, V0 is the holdup volume, and Va is the volume of the stationary phase. The value of Vequ at each breakthrough curve was calculated at the maximum numerical value of the first derivative of the breakthrough curve. The value of Va was calculated by subtracting V0 from the geometrical volume of the column. Nonlinear regression of the experimental data to adsorption isotherm models was performed using Origin 6.0 (Northampton, MA). The experimental data were fit to each isotherm model with weights (1/q2) designed to put an even emphasis on each data point during the fitting process. The best isotherm parameters were selected by minimizing the residual sum of squares (RSS) for each isotherm model. The different adsorption isotherm models were compared using the Fisher test (Fcal) and RSS values.11 Calculation of Affinity Energy Distribution. The EM algorithm programmed in Fortran (Lahey/Fujitsu Fortran 95, Incline Village, NV) was used to calculate the affinity energy distributions (AED).12-14 One hundred grid points in the K-space were used to (11) Quinones, I.; Guiochon, G. J. Chromatogr., A 1998, 796, 15-40.
1720 Analytical Chemistry, Vol. 77, No. 6, March 15, 2005
digitize ln K in the range between Kmin ) 0.000 01 mM-1 and Kmax ) 200 mM-1. The algorithm performed 107 iterations. The calculated AED from this algorithm shows the differential saturation at each point (qs(Kj)) as a function of ln Kj, where qs(Kj) ) f (ln Kj) (∆)(ln K). To estimate the isotherm parameters from the calculated AED, the values of qs comprising one peak were added to obtain the saturation capacity of the corresponding type of sites, and the value of ln Kj for the peak maximum was used to calculate the adsorption constant of the corresponding site. RESULTS AND DISCUSSION Adsorption Isotherms. The experimental isotherm data (symbols) of the two enantiomers on the Fmoc-L-Trp MIP are shown in Figure 1a-d, where the isotherms at different concentrations of acetic acid are plotted in four different concentration ranges of the substrates. The lines in the figures show the best isotherm models. The statistical data regarding the regression of (12) Stanley, B. J.; Guiochon, G. J. Phys. Chem. 1993, 97, 8098-8104. (13) Stanley, B. J.; Guiochon, G. Langmuir 1994, 10, 4278-4285. (14) Stanley, B. J.; Bialkowski, S. E.; Marshall, D. B. Anal. Chem. 1993, 65, 259-267.
Table 1. Residual Sum of Squares (RSS) and Fisher Parameters (Fcal) bi-Langmuir
tri-Langmuir
substrate
concn (M)
RSS
Fcal
RSS
Fcal
Fmoc-L-Trp
0.2 0.9 1.7 3.7 0.2 0.9 1.7 3.7
315.2 217.3 159.3 96.4 404.0 12.8 3.5 1.3
571.7 287.8 226.5 630.8 435.8 3509.4 10851.1 6999.8
3.9 18.6 5.4 8.1 8.1 18.3 3.4 0.8
42123.5 3122.9 6196.7 20289.1 20289.1 2268.97 10210.3 4434.7
Fmoc-D-Trp
these data are in Table 1 and the best values of the parameters in Table 2. The amount adsorbed at equilibrium with a given concentration of either Fmoc-Trp enantiomer decreases rapidly with increasing concentration of acetic acid. The shapes of the isotherms of the two enantiomers at the various acetic acid concentrations are qualitatively different, indicating differences in the nature of the isotherms (i.e., the number of adsorption sites).15 Figure 1a shows the data acquired in the lowest concentration range. In this range, the isotherms of Fmoc-L-Trp at all acetic acid concentrations are more strongly curved than those of Fmoc-D-Trp. The isotherms of Fmoc-L-Trp at 0.2 and 0.9 M acetic acid are already significantly nonlinear in this low concentration range and those at 1.7 and 3.7 M acetic acid show a slight curvature. Except at 0.2 and 0.9 M acetic acid in which range they exhibit a slight curvature, the isotherms of Fmoc-D-Trp are linear in this lowest concentration range. These results indicate that, at each acetic acid concentration, Fmoc-LTrp molecules bind to high-energy sites that are not accessible to Fmoc-D-Trp molecules. As the concentration range of the substrates increases (Figures 1b-d), all the isotherms exhibit a nonlinear behavior and the curvature of the isotherms decreases with increasing concentration of acetic acid. The decrease with increasing acetic acid concentration of the amount of either enantiomer adsorbed at equilibrium with any given concentration (see Figure 1a) is stronger for Fmoc-L-Trp than for Fmoc-D-Trp. Thus, the difference between the amounts of either enantiomer adsorbed decreases with increasing concentration of acetic acid. The retention times (or overall affinities) of the enantiomers measured for analytical samples decrease with increasing concentration of acetic acid, and this decrease is stronger for Fmoc-L-Trp than for Fmoc-D-Trp. Accordingly, the enantioselectivity decreases with increasing acetic acid concentration. As the concentration range increases (Figure 1b-d), the difference in the amounts of the two adsorbed enantiomers decreases. In the highest concentration range (Figure 1d), there is no significant difference either in the amounts adsorbed or in the slope of the isotherms of the two enantiomers. The lines in Figure 1a-d represent the best isotherm model corresponding to the corresponding set of data, for the four different concentrations of acetic acid. The parameters of these isotherms were derived on the basis of statistical tests made on the results of the nonlinear regression of the isotherm data to each isotherm model (i.e., RSS and Fisher parameters Fcal) and (15) Fornstedt, T.; Gotmar, G.; Andersson, M.; Guiochon, G. J. Am. Chem. Soc. 1991, 113, 1164-1174.
of the results of the independent calculation of the affinity energy distribution. Table 1 compares the statistical results obtained with these different models. For Fmoc-L-Trp, the lowest RSS and the highest Fcal were obtained with the tri-Langmuir isotherm model. For Fmoc-D-Trp at 0.2 M, the tri-Langmuir isotherm model best accounts for the isotherm data. However, with increasing amounts of acetic acid, the statistical test shows that both bi-Langmuir and tri-Langmuir isotherm models can account similarly well for the isotherm data. In cases in which the statistical tests could not clearly identify one unique isotherm model, the results of the calculation of the affinity energy distribution were used to select the best model. Panels a-d in Figure 2 show the AEDs calculated for Fmoc-LTrp (solid lines) and Fmoc-D-Trp (dotted lines) at 0.2, 0.9, 1.7, and 3.7 M acetic acid, respectively. For Fmoc-L-Trp, three distinct peaks on the AEDs are clearly seen (see Figure 2a′-d′ showing enlarged parts of Figure 2a-d). Except at 0.2 M acetic acid, the AEDS of Fmoc-D-Trp show only two distinct peaks. At 0.2 M acetic acid, however, the AEDs of Fmoc-D-Trp exhibits three peaks. Note that in our previous report we reported three peaks on the AED of Fmoc-D-Trp at 3.7 M acetic acid with a divergence of the distribution in the low-energy range. It was previously reported that such a divergence around the lowest part of the energy range sampled produces an artificial peak at the high-energy site.13 To test the authenticity of the high-energy sites observed in our previous study, we acquired data up to the highest concentrations of Fmoc-D-Trp possible, from 120 to 160 mM, and recalculated the AED. As shown in Figure 2d, we observed a nice convergence of the distribution at low energies, with no high-energy peak at all. This confirms that the high-energy peak previously reported was an artifact resulting from the divergence of the calculation of the distribution at low energies. Additionally, the diverging end of the AED seems to shift the distribution of each site toward lower energies. For example, the association constants estimated from the diverging AED of Fmoc-D-Trp at 3.7 M acetic acid were
