336 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY sorbed TBA + . For examples of these generalizations, see the discussion of Figure 7.12 in the following section. 7.4.1.2 Ion-Pair Reagent: Concentration and Type It is possible to continuously vary the nature of retention, from RPC retention as in Figure 7.11a to IPC retention as in Figure 7.11b, by varying the concentration of the IPC reagent in the stationary phase. The concentration of the reagent in the stationary phase can be varied by changing its concentration in the mobile phase. We will use the example of Figure 7.12 to illustrate the effect of IPC reagent concentration on solute retention. In this example a tetrabutylammonium phosphate IPC reagent (TBA + ) is assumed initially, with a C 8 column. Consider first the equilibrium uptake of TBA + ≡ R + by the stationary phase (solid curve for TBA + at the bottom of Fig. 7.12a). The concentration of R + in the stationary phase [R + ] s is plotted against its concentration in the mobile phase [R + ] m , showing a continued increase in stationary phase concentration as [R + ] m increases, until the stationary phase becomes saturated with R + (with no further change in concentration of R + in the stationary phase for further increase in [R + ] m ). The form of this plot is typical of the uptake of sample or other molecules by a RPC column, when the sample concentration in the mobile phase is increased (so-called Langmuir adsorption; Section 15.3.1.1 and Eq. 15.1). Two different plots of [R + ] s versus [R + ] m are shown in Figure 7.12a: a solid curve for TBA + , and a dashed curve for tetraethylammonium (TEA + ). Because TBA + is the more hydrophobic of the two IPC reagents, it is retained by the stationary phase more strongly and saturates the column at a lower concentration of R + in the mobile phase. The extent of ion-pairing will depend on the fractional saturation of the stationary phase, and this is seen to depend on (1) the IPC reagent concentration in the mobile phase and (2) the hydrophobicity or retention of the IPC reagent. IPC reagent hydrophobicity and retention will increase for an increase in the carbon number of the reagent (number of CH 3 -plusCH 2 - groups in the reagent molecule), making TBA + (with 16 carbons) more hydrophobic and more retained than TEA + (with 8 carbons). We also see in Figure 7.12a (dotted lines) that a larger concentration y of a less hydrophobic reagent (TEA + ) can result in the same uptake [R + ] s of reagent by the column as a lower concentration x of a more hydrophobic IPC reagent (TBA + ), consequently resulting in similar ion pairing and retention of the solute A − (e.g., see later Fig. 7.15 and the accompanying discussion). Consider next the change in sample retention k as the IPC reagent concentration increases (Fig. 7.12b). Assume a fully ionized acidic solute RCOO − , whose RPC retention is essentially zero for some value of %B. When an IPC reagent is added to the mobile phase, k will increase initially as a result of the interaction of RCOO − with R + in the stationary phase. Once the stationary phase is saturated with R + , no further increase in retention of RCOO − can occur; however, a further increase in the mobile-phase concentration of R + isaccompaniedbyanincreaseinthe concentration of its counter-ion X − (e.g., H 2 PO − 4 ). An increase in the H 2 PO − 4 concentration [H 2 PO − 4 ] m then competes with RCOO − for ion-exchange retention (Eq. 7.9a), leading to a gradual decrease in k as shown in Figure 7.12b. Thus preferred concentrations of the IPC reagent in the mobile phase should not exceed a value that begins to saturate the column—so that solute retention then declines. 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 337 (b) Retention of ionized acid RCOO − vs concentration of R + in mobile phase [R + ] m mobile phase stationary phase (a) Uptake of IPC reagent R + by the stationary phase ++ TBA + H 2 PO 4 + TBA + H 2 PO 4 − TBA + H 2 PO 4 − TEA + TBA + H 2 PO 4 + C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 C 8 TBA + RCOO − RCOO − k [R + ] m [R + ] m [R + ] s IPC retention competition with H 2 PO 4 − x y Figure 7.12 Representation of the uptake of IPC reagent by the column, and its effect on solute retention. 7.4.1.3 Simultaneous Changes in pH and Ion Pairing When mobile-phase pH and the concentration of the IPC reagent are varied simulta- neously, a remarkable control is possible over retention range and relative retention for ionic samples (as anticipated by Fig. 7.10). This can be visualized for an indi- vidual solute from the plots of k against either pH or IPC reagent concentration in Figures 7.11b and 7.12b. Now consider that similar plots will result for other solutes, but with the possibility of different dependencies of k on pH and IPC reagent concentration. An example of separation as a function of change in both 338 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY 02 Time (min) 02 Time (min) 02 Time ( min ) B HA 1 + HA 2 N B N HA 2 HA 2 HA 1 HA 1 N B (a) pH-2.5 RPC (b) pH-5.0 RPC (c) pH-7.5 RPC 02 Time (min) HA 2 HA 1 N B pH-5.0 IPC (R − ) (d) Figure 7.13 Example of the separation of an ionic sample where both mobile-phase pH and IPC reagent concentration are varied. Sample: B, pseudoephedrine; N, glycerol guaicolate (a neutral compound, shaded); HA 1 , sodium benzoate; HA 2 , methylparaben (a phenol). Conditions: 150 × 4.6-mm C 8 column (5-μm particles); 30% methanol-citrate buffer with 130-mM hexane sulfonate for the IPC mobile phase; 50 ◦ C; 3.0 mL/min. Adapted from [51]. mobile-phase pH and IPC reagent concentration is illustrated in Figure 7.13. The sample consists of a neutral compound (N; shaded peak), a weakly basic compound (B), an acidic compound (HA 1 ) and a weakly acidic compound (HA 2 ). Three separa- tions were carried out (Fig. 7.13a–c), with pH varying but without any IPC reagent in the mobile phase. The dotted lines track changes in relative retention for each peak. A fourth separation (Fig. 7.13d) uses a mobile phase with intermediate pH plus added hexane sulfonate (R − ) as the IPC reagent. Consider first the retention of the neutral compound N (shaded peak). As pH is varied in the separations of Figure 7.13a–c, there is little change in its retention—as expected. When the IPC reagent (R − ) is added in Figure 7.13d, the retention of N is reduced as a result of partial blockage of the stationary-phase surface by sorbed R − . Next consider the retention of weakly basic compound B. Its retention increases for pH > 5, and the addition of IPC reagent in Figure 7.13d (pH = 5.0), increases retention even more (the arrows in Fig. 7.13d indicate changes in retention vs. the separation of Fig. 7.13b [same pH]). The various changes in the retention of B are the result of a decreasing ionization of this basic solute (BH + → B) as pH increases, while addition of the IPC reagent R − confers a negative charge on the column that attracts the positively charged BH + . Finally, acidic compounds HA 1 and HA 2 show decreased retention with increase in pH, and their retention is further decreased by 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 339 the addition of the IPC reagent R − (because of ionic repulsion between A − and R − in the stationary phase, plus partial coverage of the stationary phase by sorbed molecules of R − ). Finally, we can see that as a result of change in both mobile-phase pH and IPC reagent concentration, major changes in relative retention result for each of the conditions of Figure 7.13: Low pH, no ion pairing (a): B < N < AH 1 = AH 2 Intermediate pH, no ion pairing (b): B < HA 1 < N < HA 2 High pH, no ion pairing (c): HA 1 < N < B < HA 2 Intermediate pH, ion pairing, d: HA 1 < N < HA 2 < B The best separation for this sample is seen in Figure 7.13d ,usinganIPC reagent at pH-5.0. To summarize, the concentration and type of the IPC reagent can be varied for systematic changes in ion pairing, with predictable effects on retention range and relative retention. When an anionic IPC reagent (e.g., an alkylsulfonate) is added to the mobile phase, the retention of ionized basic compounds will be increased, and the retention of neutral and (especially) acidic compounds will be decreased. When a cationic IPC reagent (e.g., a tetraalkylammonium salt) is used, the retention of ionized acidic compounds will be increased, while that of neutral and (especially) basic solutes will be decreased. These changes in retention will be greater for larger concentrations of the IPC reagent in the mobile phase. Changes in mobile-phase pH that increase the ionization of a compound will increase the effect of the IPC reagent on separation. 7.4.2 Method Development IPC method development is similar to that for the RPC separation of ionic samples (Section 7.3.3). The same seven method-development steps listed in Figure 6.21a for neutral samples still apply, with only step 3 (‘‘choosing separation condition’’) differing for IPC separation. As in the case of RPC method development, the choice of separation conditions for IPC includes the following steps: 1. choose starting conditions 2. select %B for 1 ≤ k ≤ 10 (with IPC reagent in the mobile phase) 3. adjust conditions for improved selectivity and resolution 340 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY 4. vary column conditions for best compromise between resolution and run time In most cases, an RPC separation (without IPC reagent present) will have been attempted initially, including a study of changes in mobile-phase pH (our recommendation). Consequently it is likely that different peaks in the chromatogram can be assigned as neutrals, acids, or bases (as in the example of Fig. 7.3). This approach also explores the possibility of a non-IPC, RPC separation—with a simpler mobile phase and one less likely to have IPC-related problems (Section 7.4.3). Even when IPC separation is anticipated at the beginning of method development, initial experiments should proceed in similar fashion as described in Section 7.3.3 (development of an RPC method for ionic samples)—without addition of the IPC reagent to the mobile phase. The latter experiments will define the approximate %B required for an acceptable retention range (e.g., 1 < k < 10) or at least a %B value for an average value of k that falls within this range. As in the case of RPC method development for ionic samples, only steps 1 and 3 above differ for IPC. 7.4.2.1 Choice of Initial Conditions (Step 1) The requirement of both a buffer and an IPC reagent in the mobile phase may favor the use of methanol as B-solvent, because of the greater solubility of these additives in methanol (TFA and HFBA, however, have adequate solubility in acetonitrile but are weaker IPC reagents). If acetonitrile is used in preliminary RPC experiments (our recommendation), and if solubility problems are subsequently encountered with this solvent, methanol can be substituted (guided by the solvent strength nomograph of Fig. 6.11). In most cases an alkylsulfonate will be chosen as IPC reagent for samples that contain basic compounds, while a tetraalkylammonium salt will be used for acidic samples. When both acids and bases are present in the sample, either type of IPC reagent may prove useful, but it is not recommended to add both reagents to the mobile phase. The reagents tend to ion-pair with each other, with cancellation of their net effect on separation; the use of two IPC reagents would also complicate method development. An alkylsulfonate is preferred when it is necessary to selectively increase the retention of basic solutes, while a tetraalkylammonium salt can increase the retention of acidic solutes. For mixtures of acids and bases, a low-pH mobile phase plus an alkylsulfonate IPC reagent is a good starting point because the pH suppresses ionization of acids and the IPC reagent retains the bases. The final choice of one or the other of these reagents can be determined from information acquired during preliminary RPC experiments; specifically, retention as a function of mobile-phase pH. Both sulfonates and quaternary ammonium salts can be used with UV detection at a wavelength of 210 nm or higher. The discussion of Section 7.4.1.2 and Figure 7.12a suggest that similar separations can be obtained with different concentrations of two different alkyl- sulfonates (or quaternary ammonium salts), for example, a lower concentration of a C 8 -sulfonate, or a higher concentration of a C 6 -sulfonate. This is generally correct [47, 48], but the practical question is then: Which IPC reagent and which concentration should be used for an initial IPC experiment? One study [49] has provided an approximate answer to this question, as summarized in Figure 7.14. 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 341 100 10 1 100806040200 % MeOH 100806040200 % MeOH C 4 (C 1 ) 4 (C 2 ) 4 (C 3 ) 4 (C 4 ) 4 (C 5 ) 4 C 6 C 8 C 10 C 12 100 10 (a) Recommended concentrations of different alkyl sulfonates (b) Recommended concentrations of different tetraalkylammonium salts [R − ] m [R + ] m Figure 7.14 Recommended IPC reagents and concentrations as a function of mobile-phase %B: (a) sulfonate IPC reagent; (b) quaternary ammonium IPC reagent. Adapted from [49]. Conditions corresponding to shaded regions are not recommended. The recommended starting concentrations of different sulfonates (Fig. 7.14a)or quaternary ammonium salts (Fig. 7.14b) are plotted against %-MeOH in the mobile phase. For example, given a mobile phase of 40%B, and the planned addition of an alkylsulfonate as IPC reagent, Figure 7.14a suggests that either 75 mM of octanesulfonate (C 8 ) or 15 mM of decanesulfonate (C 10 ) would be a suitable starting concentration. Alternatively, for the use of a tetraalkylammonium salt (Fig. 7.14b) with a mobile phase of 40% B, either 70 mM of tetrabutylammonium (C 4 )or 20 mM of tetrapentylammonium (C 5 ) is recommended. It is desirable to select an initial concentration of the IPC reagent between 5 and 100 mM, as indicated by the unshaded regions of Figure 7.14a, b. The (initial) concentrations recommended in Figure 7.14 will provide about 1 / 3 surface coverage by the reagent. Varying the concentration up or down from this initial value then allows a significant change in reagent uptake by the column, with predictable changes in relative retention (Section 7.4.1.2). If acetonitrile is used 342 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY as B-solvent for IPC separation, Figure 7.14 can still be used for estimates of IPC reagent type and concentration, but the equivalent value of %B for MeOH should be (very approximately) doubled. Thus, if the mobile phase consists of 20% ACN, a value of 2 × 20 ≈ 40% MeOH should be used in Figure 7.14 [50] for the purpose of selecting an IPC reagent and its initial concentration. That is, acetonitrile is a stronger solvent than methanol, so a lower value of %-ACN is equivalent to a higher value of %-MeOH. An example that illustrates some of the principles above is provided by Figure 7.15, where the separation of a mixture of water-soluble vitamins is exam- ined as a function of IPC reagent concentration and type. As the sample includes compounds with both acidic (anionic) and basic (cationic) character, either a sulfonate or quaternary ammonium salt could be used for ion pairing. If a sul- fonate is selected, Figure 7.14 suggests for this mobile phase (15% methanol-buffer [pH-3.2]) the use of hexane, heptane, or octane sulfonate as IPC reagent. Sepa- rations of the sample with varying concentrations of hexane sulfonate are shown in Figure 7.15a–c. Peaks 1 to 3 exhibit little change in retention with changing reagent concentration and can be regarded as effectively neutral (neither anionic or cationic). Peaks 4 and 6 show an increase in retention as the reagent concen- tration increases, so these peaks must be cationic (protonated bases or quaternary 5 mM 1.3 mM 10 mM 2.5 mM 20 mM 5 mM hexane sulfonate heptane sulfonate 02468 Time (min) 02468 Time (min) Time (min) Time ( min ) Time ( min ) 02468 Time (min) 02468 4 5 7 6 4 5 6 7 1 2 3 4 5 6 7 4 5 7 6 0 2 4 6 8 0 2 4 6 8 10 12 4 + 5 7 6 4 + 5 7 6 (a) (b) (c) (d) (e) (f ) Figure 7.15 IPC separation of a sample of water-soluble vitamins as a function of IPC reagent concentration and type. Sample: 1, ascorbic acid; 2,niacin;3, niacinamide; 4, pyridoxine; 5, folic acid; 6,thiamine;7, riboflavin. Conditions; 83 × 4.6-mm C 8 column (3-μmparti- cles); 15% methanol- buffer (pH-3.2); 35 ◦ C; 2.0 mL/min. Chromatograms recreated from data in [48]. 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 343 ammonium compounds). Similarly peaks 5 and 7 exhibit decreased retention as the reagent concentration increases and are therefore anionic (ionized acids). From these initial experiments (Fig. 7.15a–c), it appears that a hexane sulfonate concen- tration of 6 to 7 mM would provide maximum resolution of peaks 4 to 7 (placing peak 5 midway between peaks 4 and 6, but without moving peak 7 too close to peak 6). Figure 7.15d–f shows corresponding separations with heptane sulfonate as IPC reagent; in each case, the reagent concentration is reduced fourfold compared to separations with hexane sulfonate as IPC reagent. The resulting separations for Figure 7.15a and d,b and e,orc and f, are each quite similar (but not identical). That is, essentially the same separation can be achieved for this sample with a lower concentration of a more hydrophobic IPC reagent, but the selectivity may not be exactly the same. The arbitrary substitution of one IPC reagent for another in a previously developed method is therefore not recommended. Inorganic reagents (or ‘‘chaotropes’’) such as ClO − 4 ,BF − 4 , and PF − 6 have also been used in IPC [52, 53], in place of the usual alkane sufonates. Because of the lesser retention of inorganic reagents, it is likely that the retention mechanism is based on Equation (7.9)— ion-pairing in the mobile phase—rather than Equation (7.9a)—sorption of the IPC reagent. Chaotropes are advantageous in being better suited for gradient elution (less baseline noise and drift) and are more soluble in mobile phases with larger values of %B. The relative ion-pairing strength of various anions (including both buffers and IPC reagents) increases in the following order: H 2 PO − 4 < HCOO − < CH 3 SO − 3 < Cl − < NO − 3  CF 3 COO − < BF − 4 < ClO − 4 < PF − 6 Only the last four anions are useful for IPC. Because inorganic IPC reagents (chaotropes) are less strongly retained by the stationary phase, this can mean faster equilibration of the column when changing the mobile phase (Section 7.4.3.2). The effect of chaotropes in altering the retention of protonated bases appears much more pronounced for acetonitrile as B-solvent, compared to methanol or tetrahydrofuran [54]. 7.4.2.2 Control of Selectivity (Step 3) The separation conditions available for the control of selectivity in IPC include: •pH • IPC reagent type (sulfonate, quaternary ammonium salt, chaotrope) • IPC reagent concentration • solvent strength (%B) • solvent type (ACN, MeOH, etc.) • temperature • column type • buffer type and concentration 344 IONIC SAMPLES: REVERSED-PHASE, ION-PAIR, AND ION-EXCHANGE CHROMATOGRAPHY Despite the large number of variables that can affect selectivity in IPC, usually only a few of these conditions need to be investigated during method development. Fur- thermore the effects on retention of several of the conditions above are interrelated. Thus a change in mobile-phase pH will, in some cases, give similar results as a change in IPC reagent concentration; for example, an increase in pH or an increase in the concentration of an alkylsulfonate IPC reagent will in each case result in an increase in the retention of basic solutes (and a decrease in retention for acids). Also we have seen that the primary effect of a change in %B or temperature may be the result of associated changes in the ‘‘effective’’ pH of the mobile phase—hence providing similar changes in relative retention as for a change in mobile-phase pH. Other examples of this kind are noted below. Mobile Phase pH and IPC Reagent Type or Concentration. The combined effects of these conditions on IPC separation were discussed in detail above (Section 7.4.1), and are best investigated first during IPC method development. It is more convenient to vary the concentration of the IPC reagent (as in Fig. 7.15a–c), than to change the IPC reagent (as in Fig. 7.15f vs. a). Solvent Strength. When %B is varied for the RPC separation of ionic samples (Section 7.3.2.2; Fig. 7.7), changes in both absolute and relative retention can be expected. In some cases these change in retention can be related to corresponding changes in the ‘‘apparent’’ pH of the mobile phase (or values of pK a for the solute; Section 7.2.3). Corresponding changes in relative retention with %B should also occur for IPC separation, but with an added feature. Thus, if %B is increased, the uptake of the IPC reagent by the column will decrease, just as for the case of sample molecules. Consequently a change in %B should lead to predictable changes in relative retention for peaks that are strongly affected by ion pairing. An example is presented in Figure 7.16 for the same sample of Figure 7.10. An initial separation with 40% MeOH and octanesulfonate as IPC reagent is shown in Figure 7.16a,with the four protonated bases (X, X 1 –X 3 ) distinguished as shaded peaks (remaining peaks correspond to either neutral or acidic solutes). When the mobile phase is changed to 45% MeOH in Figure 7.16b, and 50% MeOH in Figure 7.16c,the retention of all peaks decreases (solvent strength effect), but the four bases become even less retained relative to the remaining neutral and acidic peak (they move toward the front of the chromatogram). This behavior is predictable, as the increase in %MeOH will result in a decrease in the retention of the IPC reagent (R − )by the stationary phase. The reduced concentration of R − in the stationary phase means a reduction in ion-pairing for the cationic species X, X 1 –X 3 , and therefore their reduced retention—apart from the general decrease in retention for all peaks when %MeOH is increased. The variation of %MeOH in this example shows the exceptional power of a change in %B in IPC to affect band spacing and resolution (note the optimized separation for 45% MeOH in Fig. 7.16b). Solvent Type. A change in solvent type usually leads to changes in IPC selectivity, for either neutral samples (Section 6.3.2; Fig. 6.9) or (especially) ionic samples (Section 7.3.2.3). Because of the added effect of the B-solvent on the uptake of IPC reagent by the column, a change of solvent type in IPC can result in even larger changes in selectivity than in RPC. Figure 7.17 provides a striking 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 345 HB B X + MP PP 024681012141618 Time (min) 0246810 Time (min) 0246 Time (min) HB B X X 1 X 1 X 1 X 2 X 2 X 2 X 3 X 3 X 3 PP MP (a) (b) (c) 40% MeOH 45% MeOH 50% MeOH HB B MP X PP Figure 7.16 Solvent-strength selectivity in IPC separation. Sample and conditions as in Figure 7.10b, except for indicated % methanol. Peaks for protonated bases X, X1–X3 shaded. Adapted from [45]. example of solvent-type selectivity in IPC, for a change of B-solvent from MeOH in Figure 7.17a toACNinFigure7.17b, with use of the solvent nomograph of Figure 6.11. A combined variation of mobile-phase pH, IPC reagent concentration, and solvent type should prove especially effective for the separation of challenging ionic samples [48, 50]. Temperature. A change in temperature for IPC should also have a pronounced effect on relative retention. Temperature will alter the amount of IPC reagent held by the column. For this reason temperature control during IPC separation is especially important. Column Type and Buffer. We have seen that column type can have a major effect on selectivity in RPC separations of ionic samples, and it seems likely that this will also be true for IPC separation. However, the partial coverage of the stationary phase surface by IPC reagent may tend to mask the contribution of the column per se to sample retention. In view of the many other ways in which selectivity can be controlled in IPC, the use of column type for this purpose should not be a first choice, nor is it likely to be especially promising. Likewise the effect of buffer type and concentration on IPC separation can be significant, but larger, more predictable changes in selectivity can be obtained by varying pH, %B, temperature, and/or the type and concentration of the IPC reagent. . sulfonate concen- tration of 6 to 7 mM would provide maximum resolution of peaks 4 to 7 (placing peak 5 midway between peaks 4 and 6, but without moving peak 7 too close to peak 6). Figure 7.15d–f. sample, either type of IPC reagent may prove useful, but it is not recommended to add both reagents to the mobile phase. The reagents tend to ion-pair with each other, with cancellation of their net effect. Conditions; 83 × 4.6-mm C 8 column (3-μmparti- cles); 15% methanol- buffer (pH-3.2); 35 ◦ C; 2.0 mL/min. Chromatograms recreated from data in [48]. 7.4 ION-PAIR CHROMATOGRAPHY (IPC) 343 ammonium compounds).