Superheated Water Ion-Exchange Chromatography: An Experimental

Sep 11, 2009 - ... and the selectivity coefficient for a pair of alkali metal ions or that of alkaline-earth metal ions approaches unity as temperatur...
0 downloads 0 Views 290KB Size
Download PDF
Anal. Chem. 2009, 81, 8025–8032

Superheated Water Ion-Exchange Chromatography: An Experimental Approach for Interpretation of Separation Selectivity in Ion-Exchange Processes Masami Shibukawa,*,† Tomomi Shimasaki,† Shingo Saito,† and Takashi Yarita‡ Graduate School of Science and Technology, Saitama University, 255 Shimo-Okubo, Sakura-ku, Saitama, 338-8570, Japan, and National Metrology Institute of Japan, National Institute of Advanced Industrial Science and Technology, Tsukuba Central 3, Umezono, Tsukuba, 305-8563, Japan Cation-exchange selectivity for alkali and alkaline-earth metal ions and tetraalkylammonium ions on a strongly acidic sulfonic acid cation-exchange resin has been investigated in the temperature range of 40-175 °C using superheated water chromatography. Dependence of the distribution coefficient (ln KD) on the reciprocal of temperature (1/T) is not linear for most of the ions studied, and the selectivity coefficient for a pair of alkali metal ions or that of alkaline-earth metal ions approaches unity as temperature increases. On the other hand, the retention order of tetraalkylammonium ions is reversed at 160 °C or above when eluted with Na2SO4 aqueous solution and the larger ions are eluted faster than the smaller ones contrary to the retention order obtained at ambient temperature. The change in ion-exchange selectivity with temperature observed with superheated water chromatography has been discussed on the basis of the effect of temperature on hydration of the ions and specific adsorption or distribution of ionic species between the external solution and ion-exchange resin. In superheated water, the electrostatic interaction or association of the ions with the fixed ion becomes a predominant mechanism resulting in different separation selectivity from that obtained at ambient temperature. Application of ion-exchange techniques is extended over the wide range embracing both laboratory and industrial fields. In the majority of the uses of ion-exchange, dilute aqueous solutions of ionic compounds are treated either for purification whereby unwanted ions are removed from the solutions or for recovery and concentration of the ions in question. That is to say, ion-exchange separation is based on the fact that different ionic species generally show a different preference for an ion-exchanger even if they are of the same valency. Because of its practical importance as well as scientific interest, the ion-exchange equilibrium has been the subject of a number * Corresponding author. Phone: +81-48-858-3520. Fax: +81-48-858-3520. E-mail: [email protected]. † Saitama University. ‡ National Institute of Advanced Industrial Science and Technology. 10.1021/ac9011864 CCC: $40.75  2009 American Chemical Society Published on Web 09/11/2009

of studies and both empirical and theoretical relationships have been established to describe the distribution of ionic species between various types of exchangers and surrounding medium. The most popular approach to the interpretation of ion-exchange equilibrium is based on a stoichiometric model, in which the distribution of an analyte ion AzA into the exchanger is illustrated by the following equation as an exchange between the analyte ion and the counterion EzE in the external solution1 zEAzA(s) + zAEzE(r) a zEAzA(r) + zAEzE(s)

(1)

where s and r refer to the solution and exchanger phases, respectively. Ion-exchange selectivity is usually defined in terms of the selectivity coefficient KA/E written for the equilibrium according to a particular choice of concentration units. For example, if equivalent ionic fraction X is adopted as a concentration unit, KA/E is written as

KA/E )

¯ AzEX EzA X ¯ EzA X AzEX

(2)

where the quantities in the exchanger phase are given a superscript bar notation. This model assumes the analyte ion distribution between the two homogeneous phases, i.e., the exchanger phase and the external solution phase, due to exchange with the counterion. This may be applied to systems involving soft gel ion exchangers of relatively high exchange capacity, in which a homogeneous distribution of fixed ionic charge can be assumed.2 Thermodynamic theories interpreting selectivity of counterions as well as the swelling properties of a soft gel ion exchanger are usually based on a Gibbs-Donnan equilibrium.2 The Gibbs-Donnan equation for the ion-exchange represented by eq 1 can be written as

ln KA/E )

¯f AzE fAzE Π (zEvA - zAvE) + ln z - ln z RT ¯f EA fE A

(3)

(1) Grimshaw, R. W.; Harland, C. E. Ion-Exchange: Introduction to Theory and Practice; The Chemical Society: London, U.K., 1975. (2) Helfferich, F. Ion Exchange; McGraw-Hill: New York, 1962.

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

8025

where Π is the swelling pressure, R is the gas constant, T is the absolute temperature, v is the partial molar volume of the ionic species, and f is the ionic activity coefficient. The problem lies in obtaining ionic activity coefficient values in the exchanger phase. If it can be assumed that the activity coefficient ratios in eq 3 are unity for the exchanger phase as well as for the external solution phase, the selectivity coefficient is determined by the pressurevolume term as given by Gregor:3,4

ln KA/E )

Π (z v - zAvE) RT E A

(4)

The chemical components of the Gregor’s model are the solvated ions, the resin matrix with its solvated fixed ions, and the solvent. Equation 4 predicts that the ion-exchanger prefers the ion with smaller solvated volume and this prediction is in accord with the observation for simple systems such as the exchange of alkali metal ions on some strongly acidic cation-exchange resins.1,4,5 However, this simple model cannot account for the observed selectivity reversals with a change in concentrations of the analyte ion and counterion, etc.1 Some investigators claimed that the activity coefficients of ionic species in ion-exchange systems are determined by strong longrange electrostatic forces and therefore a stoichiometric model is inadequate for description of ion-exchange selectivity.6-11 They have set up physicochemical models based on the Gouy-Chapman or Stern-Gouy-Chapman theory of the electrical double layer assuming that the ion-exchanger can be regarded as a rigid surface with the fixed charged groups. Some variation of ionexchange selectivity with ionic strength and affinity of a counterion have successfully been interpreted in terms of the electrical double layer model.10,11 However, the origin of ion-exchange selectivity has not completely been clarified yet in spite of numerous theoretical and experimental works on the ion-exchange equilibrium. It is recognized that the ion-exchange equilibrium is not determined solely by electrostatic forces but also by specific interaction depending on size and structures of the exchangeable ions and the fixed charged groups. According to Cantwell and his coworkers, there are two mechanisms by which an analyte ion can be distributed into the exchanger: ion-exchange of the analyte ion for other counterions in the diffuse part of the electrical double layer and adsorption of the analyte ion onto the surface.6-9 They concluded that ion-exchange between the bulk solution and the diffuse layer is the dominating factor for distribution of the ions. On the other hand, Stahlberg10 considered that selectivity between different ions of the same charges is exclusively due to difference in strength of adsorption of the ions on the matrix surfaces, and the accumulation of the ions in the diffuse layer is independent of the nature of the ions, contrary to the view of Cantwell and his co-workers. Okada11 also presented an ion-exchange model based (3) (4) (5) (6) (7) (8) (9) (10) (11)

Gregor, H. P. J. Am. Chem. Soc. 1948, 70, 1293. Gregor, H. P. J. Am. Chem. Soc. 1951, 73, 642–650. Gregor, H. P.; Bregman, J. I. J. Colloid Sci. 1951, 6, 323–347. Afrashtehfar, S.; Cantwell, F. F. Anal. Chem. 1982, 54, 2422–2427. Hux, R. A.; Cantwell, F. F. Anal. Chem. 1984, 56, 1258–1263. Liu, H.; Cantwell, F. F. Anal. Chem. 1991, 63, 993–1000. Liu, H.; Cantwell, F. F. Anal. Chem. 1991, 63, 2032–2037. Stahlberg, G. Anal. Chem. 1994, 66, 440–449. Okada, T. Anal. Chem. 1998, 70, 1692–1700.

8026

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

on Stern-Gouy-Chapman electrical double-layer theory assuming that ion-pair formation of analyte ions with the fixed ions as well as the specific adsorption of the ions on the surface govern the ion-exchange selectivity although the accumulation of the ions in the diffuse layer contribute to the mechanisms of ion-exchange. It must be noted that any of these mechanisms, the specific surface adsorption of the analyte ions, the ion-pair formation of the ions with the fixed charges, and even the distribution of the ions into the diffuse layer may strongly depend on the solvation of the ions. Since ion-exchange processes are usually conducted in aqueous solutions, hydration of the ions should be involved in the determination of ion-exchange separation selectivity. One of the approaches which would give insights into the effect of ionic solvation on the ion-exchange selectivity is the use of nonaqueous solvents for the study of ion-exchange equilibrium.12-14 However, use of mixed solvents such as methanol-water may add another factor to the distribution processes of ions due to selective adsorption of one of the solvent components to the exchanger. Even a “pure” nonaqueous solvent may contain a slight amount of water, which would strongly solvate ionic species and affect the ion-exchange equilibrium. In recent years, superheated water, i.e., liquid water under pressure at temperatures above 100 °C, has been used as the mobile phase solvent for reversed-phase liquid chromatography.15-20 Superheated water is much less polar than water at ambient temperature and behaves like an organic solvent such as methanol.19,21,22 This behavior can be interpreted in terms of disruption of the hydrogen bonding network by the thermal motion of water molecules.23-25 Therefore the physicochemical properties of water such as permittivity can be continuously altered by changing temperature. This suggests that an investigation of ion-exchange equilibrium in superheated water may shed a light on the role of water solvent in ion-exchange separation selectivity. The effect of temperature on ion-exchange chromatography has so far been studied by several investigators. Even though most of the studies were carried out at temperatures up to 90 °C, it has been demonstrated that the selectivity of ion-exchange separations can be altered through a change in column temperature and the effect of temperature on the retention of ionic compounds depends not only on the nature of analyte ions but also on eluent electrolytes.26-36 For example, the retention of an (12) Okada, T. J. Chromatogr., A 1997, 758, 19–28. (13) Okada, T. J. Chromatogr., A 1998, 804, 17–28. (14) Lucas, A.; Valverde, J. L.; Romero, M. C.; Gomez, J. Chem. Eng. Sci. 2002, 57, 1943–1954. (15) Smith, R. M.; Burgess, R. J. J. Chromatogr., A 1997, 785, 49–55. (16) Yarita, T.; Nakajima, R.; Otsuka, S.; Ihara, T.; Takatsu, A.; Shibukawa, M. J. Chromatogr., A 2002, 976, 387–391. (17) Yarita, T.; Nakajima, R.; Shibukawa, M. Anal. Sci. 2003, 19, 269–272. (18) McNeff, C. V.; Yan, B.; Stoll, D. R.; Henry, R. A. J. Sep. Sci. 2007, 30, 1672–1685. (19) Yang, Y. J. Sep. Sci. 2007, 30, 1131–1140. (20) Smith, R. M. J. Chromatogr., A 2008, 1184, 441–455. (21) Heger, K.; Uematsu, M.; Franck, E. U. Ber. Bunsen-Ges.: Phys. Chem. 1980, 84, 758–762. (22) Miller, D. J.; Hawthorne, S. B. J. Chem. Eng. Data 1998, 43, 1043–1047. (23) Walrafen, G. E.; Fischer, M. R.; Hokmabadi, M. S.; Yang, W.-H. J. Chem. Phys. 1986, 85, 6970–6982. (24) Walrafen, G. E.; Hokmabadi, M. S.; Yang, W.-H. J. Phys. Chem. 1988, 92, 4540–4542. (25) Pawlowski, T. M.; Poole, C. F. Anal. Commun. 1999, 36, 71–75. (26) Fortier, N. E.; Fritz, J. S. Talanta 1987, 34, 415–418. (27) Bokx, P. K.; Boots, H. M. J. J. Phys. Chem. 1989, 93, 8243–8248.

ion decreases with an increase in temperature in an eluent electrolyte system, while it increases in another electrolyte system.26,30 However, this curious dependence of the retention on temperature has not been explained satisfactorily. Kraus and Raridon investigated the cation-exchange equlibria for various metal ions over more extended temperature ranges, 0-145 °C for exchanges with the hydrogen form of the resin and 0-200 °C for those with the sodium form, by a batch method they developed.37 They found out that inversions of selectivity occur for some pairs of cations, although they did not discuss the origin of the change in cation-exchange selectivity with temperature. In the present work, the effect of elevated temperature on the separation of alkali and alkaline-earth metal ions and tetraalkylammonium ions on a strongly acidic sulfonated ion exchanger over a wide temperature range, 40-175 °C, has been examined using superheated water ion-exchange chromatography. We will show that a greater understanding of the origin of the ionexchange selectivity can be obtained by studying temperature effects on the retention of metal ions as well as tetraalkylammonium ions using superheated water as the mobile phase solvent. EXPERIMENTAL SECTION Chemicals. All chemicals used in this study were obtained from commercial sources and were of reagent-grade unless otherwise stated. Deuterated water (D2O) for NMR use was purchased from Wako Pure Chemicals (Tokyo, Japan). Water was purified subsequently with an ion-exchange cartridge PFIII H10 (Organo, Tokyo, Japan) and an Arium 611 DI (Sartorius, Tokyo, Japan). The cation-exchange resin used was DIAION MCI GEL CK10S (sulfonated polystyrene-divinylbenzene copolymer, particle diameter ) 11 µm, Mitsubishi Chemical, Tokyo, Japan). DIAION MCI GEL CHP5C (polystyrene-divinylbenzene copolymer (PSDVB), average pore diameter ) 250 Å, particle diameter ) 10 µm, Mitsubishi Chemical) was also used for adsorption experiments. The Na+ form of MCI GEL CK10S was slurry packed into a stainless steel column with water; column dimensions are 4.6 mm i.d. and 10 or 50 mm depending on retention of solute compounds studied. MCI GEL CHP5C was slurry packed with methanol into a stainless steel column (4.6 mm i.d. × 10 mm), and then the solvent was replaced with water. The ion-exchange capacity of the MCI GEL CK10S column was determined by titrating hydrogen ion eluted from the H+ form resin column by passing a sufficient amount of Na2SO4 solution with standard NaOH solution. The capacity of every column used for the study was determined just after a series of experiments conducted with superheated water. (28) Smith, R. G.; Drake, P. P.; Lamb, J. D. J. Chromatogr. 1991, 546, 139–149. (29) Kolpachnikova, M. G.; Penner, N. A.; Nesterenko, P. N. J. Chromatogr., A 1998, 826, 15–23. (30) Hatsis, P.; Lucy, C. A. Analyst 2001, 126, 2113–2118. (31) Hatsis, P.; Lucy, C. A. J. Chromatogr., A 2001, 920, 3–11. (32) Bashir, W.; Tyrrell, E.; Feeney, O.; Paull, B. J. Chromatogr., A 2002, 964, 113–122. (33) Dybczynski, R.; Kulisa, K. Chromatographia 2003, 57, 475–484. (34) Paull, B.; Bashir, W. Analyst 2003, 128, 335–344. (35) Chong, J.; Hatsis, P.; Lucy, C. A. J. Chromatogr., A 2003, 997, 161–169. (36) Yu, H.; Li, R. Chromatographia 2008, 68, 611–616. (37) Kraus, K. A.; Raridon, R. J. J. Phys. Chem. 1959, 63, 1901–1907. (38) Walton, H. F. Principles of Ion-Exchange. In Chromatography, 3rd ed.; Heftmann, E., Ed.; Van Nostrand: New York, 1975.

Chromatographic Conditions. Chromatographic measurements were performed on an HPLC system consisting of a GL Science (Tokyo, Japan) Model 546B degasser, a Shimadzu (Tokyo, Japan) Model LC 6A pump, a Rheodyne (Cotati, CA, USA) Model 7215 loading injector fitted with a 20 µL sample loop, and a Showa Denko (Tokyo, Japan) Model Shodex RH-101 refractometric detector. A preheating coil (Hastelloy, 2 m × 0.5 mm i.d.) and a backpressure coil (stainless steel, 4.2 m × 0.1 mm i.d.) were placed between the injector and the column inlet and between the column outlet and the detector, respectively. The preheating coil and the column were thermostatted using a Shimadzu Model GC-14A column oven for gas chromatography. The temperature of the oven was monitored with a mercury thermometer inserted into the oven. Aqueous solutions of sulfates and nitrates of alkali metal ions and alkaline-earth metal nitrates were used as eluents. All the eluents were filtered through a 0.45 µm membrane filter and degassed with an aspirator in an ultrasonic bath, US CLEANER US-2R (AS ONE, Tokyo, Japan) before use. Elutions were carried out at a constant flow rate of 0.2-1.0 mL min-1. The extra column volume was determined by measuring the elution volume of a sample solute through the system from which the column had been removed. The mobile phase volume was assumed to be the volume of water in the column and determined by measuring the retention volume of D2O. Test solutions were prepared by dissolving analyte compounds in the eluent to be used. The detection signal was fed into a CAC data analysis system (version 4.2, Nihon Filcon, Tokyo, Japan). RESULTS AND DISCUSSION Thermostability of the Cation-Exchange Resin. It has been known that thermostability of ion-exchange resins depends not only on the polymer matrixes and fixed ionic groups but also on counterions. The chemical breakdown of PSDVB is negligible,17 while the sulfonic acid cation-exchange resins are known to decompose in their H+ form to produce sulfuric acid with some accompanying breaking of the polymer chains.37,38 Therefore we used alkali and alkaline-earth metal ions as counterions in this study. We examined the change in retention factor of cesium ion with time on a column packed with DIAION MCI GEL CK10S by elution with 50 mM sodium sulfate at 150 and 175 °C (Figure S-1 in the Supporting Information). Although the retention factor of cesium ion slightly decreased with the lapse of time or with increase in volume of the mobile phase passed through the column, the change in the retention was negligibly small up to 30 h at each temperature. We have thus carried out a series of experiments for examining the effect of temperature in the range of 40-175 °C on the retention of analyte ions in a particular eluent electrolyte system using an identical column within 30 h. Effect of Temperature on the Retention of Alkali and Alkaline-Earth Metal Ions. The effect of temperature on chromatographic retention is usually described by the general thermodynamic relationship (van’t Hoff plot) represented by ln KD ) -

∆S ∆H + RT R

(5)

where KD is the distribution coefficient of an analyte ion, T is the absolute temperature, R is the gas constant, ∆H is the enthalpy change for the ion-exchange reaction, and ∆S is the Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

8027

Figure 1. Plots of ln KD for alkali metal ions as a function of reciprocal of temperature on a MCI GEL CK10S column. Eluent: (a) 25 mM Li2SO4, (b) 25 mM Na2SO4, (c) 25 mM K2SO4. Symbols: (4) Li+, (b) Na+, (O) K+, (9) Rb+, (0) Cs+. The solid line denotes the ln KD(E) value for each column.

corresponding change in entropy. We calculated the KD value of an analyte ion by the following equation assuming that the mobile phase is the space occupied by water in the column and the stationary phase is the other part of the column, i.e., the cation-exchange resin: KD )

VR - V0 Vs

(6)

where VR is the retention volume of the analyte ion and V0 is the retention volume of D2O. Vs is the volume of the ionexchange resin and is obtained by subtracting the V0 value from the total column volume. Figure 1 shows the effect of temperature on the KD value of alkali metal ions on a column packed 8028

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

with DIAION MCI GEL CK10S obtained by elution with 25 mM Li2SO4, Na2SO4, and K2SO4 aqueous solutions. It should be noted that most of the ln KD vs 1/T plots are obviously nonlinear and the effect of temperature on the retention of the analyte ion varies with the nature of eluent ions in the mobile phase. More noticeable is that the KD values of alkali metal ions appear to converge on a certain value with increase in temperature. We observed approximately the same temperature dependence of the KD values of alkali metal ions for NaNO3 eluent system as that for Na2SO4 system (Figure S-2 in the Supporting Information). This indicates that the effect of temperature on the retention of alkali metal ions in cation-exchange chromatography does not depend on the nature of co-ions. A similar behavior was also observed for the retention of alkaline-

Figure 2. Plots of ln KD for alkaline-earth metal ions as a function of reciprocal of temperature on a MCI GEL CK10S column. Eluent: (a) 25 mM Ca(NO3)2, (b) 25 mM Sr(NO3)2, (c) 25 mM Ba(NO3)2. Symbols: (b) Mg2+, (() Ca2+, (2) Sr2+, (9) Ba2+. The solid line denotes the ln KD(E) value for each column.

earth metal ions as shown in Figure 2 when eluted with 25 mM Ca(NO3)2, Sr(NO3)2, and Ba(NO3)2 aqueous solutions. However the dependence of KD on temperature is smaller than that for alkali metal ions, and especially the retention of magnesium ion appears to be virtually independent of temperature. Contrary to the results we obtained, linear van’t Hoff plots have been observed in previous studies for cation- and anion-exchange chromatography.26-31,34,35 As mentioned above, most of the previous studies were conducted at temperatures up to 90 °C. This discrepancy is thus assumed to be due to the difference in the range of temperature examined. Kraus and Raridon also observed some nonlinearity in van’t Hoff plots in their cation-exchange study performed in the temperature range of 0-145 °C with the hydrogen form of a strongly acidic cation-exchange resin.37 The dependence of selectivity coefficients for alkali metal ions vs H+ on temperature appears to be similar to that shown in Figure 1a. However, they concluded from the results they obtained that all the plots are concave upward and the curves have minima at which ∆H becomes zero. They would have obtained convex curves if they had examined the effect of temperature on the selectivity of Li+ vs Na+ or K+ as shown in parts b and c of Figure 1. Nonlinear van’t Hoff plots could be indicative of a change in the mechanism of retention owing to some change in the stationary phase. For example, Dorsey and his co-workers have shown that phase transitions of reversed-phase alkyl bonded stationary phases are responsible for deviation of the van’t Hoff plots from linearity, which becomes apparent when the temper-

ature range examined is broadened.39 However, it seems unlikely that the van’t Hoff behavior shown in Figures 1 and 2 can be accounted for by the change in the ion-exchange resin phase such as phase transition because the effect of temperature on the KD value strongly depends on the nature of the analyte ion. For example, the change in the KD values for alkaline-earth metal ions with temperature is much smaller than that for alkali metal ions. This suggests that the dependence of the retention of these simple metal ions on temperature in cation-exchange chromatography can be attributed to the change in hydration of individual analyte ions and eluent ions in the solution phase and/or in the cation-exchange resin as well as the hydration state of fixed ions with temperature. The change in structure of hydrogen bonding of water molecules in the solution phase and the resin phase may also be another cause of the temperature dependence of ionic retention. Since specific adsorption of alkali and alkaline-earth metal ions onto the resin matrix surfaces can be assumed to be negligibly small, the retention of these metal ions in cation-exchange chromatography takes place by displacement of counterions in the diffuse layer or on the fixed sulfonate ions. Adsorption of the co-ions, sulfate and nitrate ions, on the matrix surface may also be negligible11 so that the KD values for alkali or alkaline-earth metal ions in the cation-exchange chromatography systems adopted in this study can be expressed using the selectivity coefficient as (39) Cole, L. A.; Dorsey, J. G. Anal. Chem. 1992, 64, 1317–1323.

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

8029

ln KD )

|zA | 1 c ln KD(E) + ln KA/E |zE | |zE |

(7)

where KD(E) is the distribution coefficient of the eluent ion, and KcA/E is the molar selectivity coefficient, respectively. If the specific adsorption of the ions involved in the system on the matrix surface is negligible and the concentration of the analyte ion is much smaller than that of the eluent ion, the KD(E) value can be estimated as KD(E) )

Q |zE |VscE

(8)

where Q and cE are the ion-exchange capacity of the column used and molar concentration of the eluent ion, respectively. The KD(E) value for each system is shown as a solid line in Figures 1 and 2. The plots shown in Figures 1 and 2 except some plots for magnesium ion indicate that the KD values of alkali metal ions and alkaline-earth metal ions converge on the KD(E) value with an increase in temperature. This means that the selectivity coefficient approaches unity or the separation selectivity for a pair of analyte ions having the same charge becomes smaller with an increase in temperature. Therefore the retention of an analyte ion decreases with an increase in temperature when an eluent ion in the mobile phase has stronger affinity for the ion-exchanger, whereas it increases when eluted by a weaker counterion. Parts a and c of Figure 2 reveal that the retention of Sr2+ decreases with an increase in temperature in the Ca(NO3)2 eluent system, while it increases in the Ba(NO3)2 system. The change of the slope of the van’t Hoff plot with the nature of the eluent ions that have been observed in previous studies may be interpreted in terms of this mechanism. Figure 3 shows the ln KD vs 1/T plots for alkali metal ions eluted by an aqueous solution of Sr(NO3)2. It is predicted from eq 7 that the KD values of alkali metal ions approach (KD(E))1/2 with an increase in temperature. However, the KD values appear

Figure 3. Plots of ln KD for alkali metal ions as a function of the reciprocal of temperature on a MCI GEL CK10S column with 25 mM Sr(NO3)2. Symbols: (4) Li+, (b) Na+, (O) K+, (9) Rb+, (0) Cs+. The solid line denotes the 0.5 ln KD(E) value. 8030

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

to converge on a value a little larger than the(KD(E))1/2. This discrepancy from the expected behavior can be attributed to the difference in strength of hydration between alkali metal ions and alkaline-earth metal ions as described below. As the KD values of the alkali metal ions are usually larger than the (KD(E))1/2, the retention of alkali metal ions decreases with an increase in temperature when eluted by a divalent cation. It is thus expected that the retention of alkaline-earth metal ions would increase with an increase in temperature when eluted with solutions of alkali metal salts and approach KD(E)2. However, we could not determine the KD values of alkalineearth metal ions in the system where alkali metal ions are used as the eluent ions because the retention volumes are extremely large. The results shown in Figures 1-3 reveal that the charge on an analyte ion becomes a predominant factor for determining the retention in cation-exchange chromatography at elevated temperatures. In other words, it can be assumed that the electrostatic interaction or association of ions with the fixed ion governs cationexchange selectivity in superheated water. As described above, the ln KD vs 1/T plots for ionic solutes in ion-exchange chromatography should reflect the change in hydration of the analyte and eluent ions in the solution as well as in the ion-exchange resin with temperature. Harada and Okada40,41 studied the local hydration structures of several cations and anions in some ion-exchange resins by X-ray absorption fine structure (XAFS) and concluded that some ions such as Cl-, Br-, and K+ are partly dehydrated and can form direct ion-pairs with the fixed ions but the others such as Sr2+ keep their hydration structures even in the interior of the ion-exchange resin. They also discussed the ion-exchange separation selectivity for Cland Br- on the basis of the ion-pair formation mechanism using the XAFS data obtained.40 Hydration structures of several alkali and alkaline-earth metal ions and halide ions in superheated water or supercritical water have also been investigated by means of XAFS42-45 and neutron diffraction methods.46 Fulton and his co-workers reported the first observation of a significant reduction in the hydration number of Sr2+ under supercritical conditions.42,43 They also observed a similar trend for Rb+ and Br- in superheated and supercritical water and concluded that both thermal dehydration and ionpair dehydration may contribute to this reduction in the number of ion-water bonds at high temperature.44,45 Although the data on ionic hydration in the temperature range of 100-200 °C are scarce, these results suggest that the difference in hydration structures of ions may become smaller at elevated temperatures, leading to a smaller difference in affinity of ions having an identical charge to the ion-exchanger. This change in hydration of ions may also explain the fact that alkaline-earth metal ions show a smaller dependence of the KD values on temperature than alkali (40) Harada, M.; Okada, T. J. Chromatogr., A 2005, 1085, 3–7. (41) Harada, M.; Okada, T. Chem. Commun. 2008, 5182–5184. (42) Pfund, D. M.; Darab, J. G.; Fulton, J. L.; Ma, Y. J. Phys. Chem. 1994, 98, 13102–13107. (43) Palmer, B. J.; Pfund, D. M.; Fulton, J. L. J. Phys. Chem. 1996, 100, 13393– 13398. (44) Fulton, J. L.; Pfund, D. M.; Wallen, S. L.; Newville, M.; Stern, E. A.; Ma, Y. J. Chem. Phys. 1996, 105, 2161–2166. (45) Wallen, S. L.; Palmer, B. J.; Pfund, D. M.; Fulton, J. L.; Newville, M.; Ma, Y.; Stern, E. A. J. Phys. Chem. A 1997, 101, 9632–9640. (46) Yamaguchi, T.; Soper, A. K. J. Chem. Phys. 1999, 110, 3529–3535.

metal ions. The magnesium ion has the strongest hydration shell or the negative largest enthalpy of hydration47 among the analyte ions used in this study so that its hydration structure may not be disrupted appreciably even in superheated water and thus exhibits little dependence of retention on temperature. The discrepancy between the observed and expected values for a limiting KD of alkali metal ions shown in Figure 3 can also be accounted for by the difference in dependence of the hydration structure on temperature between alkali metal ions and alkalineearth metal ions. The strontium ion used as the eluent ion in the experiment shown in Figure 3 may keep to a considerable extent its hydration shell even in superheated water compared to alkali metal ions, and thus its affinity for the fixed ion should be weaker than that of less hydrated alkali metal ions. Consequently the selectivity coefficient for an alkali metal ion against an alkalineearth metal ion is larger than unity in superheated water, and the KD values of alkali metal ions approach a value slightly larger than the (KD(E))1/2 with an increase in temperature. Effect of Temperature on the Retention of Tetraalkylammonium Ions. As described above, the results obtained with superheated water cation-exchange chromatography have revealed that ion-exchange selectivity due to the difference in hydration of analyte ions diminishes with an increase in temperature. This suggests that the size of an ion as well as the charge should be a decisive factor determining the ion-exchange selectivity in superheated water since the electrostatic interaction or association of the ions with the fixed ion becomes a predominant mechanism in superheated water, i.e., the smaller the size of analyte ion, its affinity for the fixed ion should be stronger. However, we could not observe such an expected retention behavior for alkali and alkaline-earth metal ions even at 175 °C. This is probably because these metal ions keep more or less their hydration shell even in superheated water. We have thus adopted poorly hydrated tetraalkylammonium ions as model compounds and examined the effect of temperature on their retention on the cation-exchange resin. Figure 4 shows the ln KD vs 1/T plots for tetraalkylammonium ions when eluted with aqueous solutions of Na2SO4. The retention of all the tetraalkylammonium ions decreases with an increase in temperature, and the order of retention is reversed at 160 °C. At lower temperatures, the smaller tetraalkylammonium ion with a shorter alkyl chain exhibits lower KD value, while at higher temperatures the larger ion is eluted faster than the smaller one. This behavior clearly indicates that the electrostatic interaction is a predominant mechanism, and ionic size is a determinant in retention in superheated ion-exchange chromatography. The retention of the tetraalkylammonium ion at ambient temperature may be governed by specific adsorption on the resin matrix or distribution into the less hydrophilic resin phase as well as electrostatic interaction with the fixed ion. Figure 5 shows the ln KD vs 1/T plots for tetraalkylammonium ions obtained on a column packed with a PSDVB gel (MCI GEL CHP5C) when eluted with Na2SO4 aqueous solution. The KD values were tentatively calculated using eq 6. As shown in Figure 5, the KD values of all the tetraalkylammonium ions studied decrease with an increase in temperature although the KD values for tetra(47) Burgess, J. Ions in Solution; Horwood Publishing: Chichester, U.K., 1999.

Figure 4. Plots of ln KD for tetra-n-alkylammonium ions as a function of reciprocal of temperature on a MCI GEL CK10S column. Eluent: 25 mM Na2SO4. Symbols: (0) tetramethylammonium ion, (9) tetraethylammonium ion, (O) tetra-n-propylammonium ion, (b) tetra-nbutylammonium ion, (4) tetra-n-pentylammonium ion. The solid line denotes the ln KD(E) value.

Figure 5. Plots of ln KD for tetra-n-alkylammonium ions as a function of the reciprocal of temperature on a MCI GEL CHP5C column. For other details see Figure 4.

n-pentylammonium ion and those for tetra-n-butylammonium ion at lower temperature could not be determined due to their extremely strong retention. However, the dependence of KD on temperature on the PSDVB column is quite different from that on the sulfonated PSDVB column. Tetramethyl and tetraethylammonium ions exhibit little dependence of retention on temperature, while the retention of the tetra-n-butylammonium ion steeply decreases with an increase in temperature on the PSDVB column. Furthermore, the reversal of retention was not observed for the PSDVB column. This difference in Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

8031

retention behavior between on the PSDVB and on the sulfonated PSDVB suggests that the interior of the cationexchange resin is rather hydrophilic compared to that of the PSDVB resin due to the presence of the fixed sulfonate ions, and tetraalkylammonium ions are mainly located in the vicinity of the fixed ions by the electrostatic interaction. At 160 °C or above, even tetra-n-pentylammonium ion exhibits the KD value smaller than the KD(E), indicating that the affinity of the eluent ion, Na+, for the cation-exchange resin is stronger than those of tetraalkylammonium ions in superheated water. An increase in column temperature may weaken the specific adsorption of tetraalkylammonium ions or reduce the difference in hydrophobicity between the cation-exchange resin phase and the external solution phase. However, the increase in relative affinity of sodium ion for the cation-exchange resin due to its dehydration with an increase in temperature may also contribute to the decrease in KD values especially for tetramethyl and tetraethylammonium ions shown in Figure 4. CONCLUSIONS The origin of separation selectivity in cation-exchange processes was investigated on the basis of the effect of temperature on the retention of alkali and alkaline-earth metal ions and tetraalkylammonium ions on a strongly acidic sulfonic acid cationexchange resin. Contrary to the observations reported in most of the previous studies, nonlinear van’t Hoff behavior was observed for most of the analyte ions studied in temperature range of 40-175 °C. The selectivity coefficient for a pair of alkali metal ions or that of alkaline-earth metal ions approaches unity as temperature increases although the change in selecivity coefficient with temperature for alkaline-earth metal ions is much smaller than that for alkali metal ions, which indicates that the charge on an ion becomes a predominant factor for determining the retention of alkali and alkaline-earth metal ions in cation-exchange chromatography at elevated temperatures. We have concluded that the dependence of ion-exchange selectivity observed for these simple metal ions on temperature can be attributed to the change in hydration of the ions in the solution phase and/or in the cation-

8032

Analytical Chemistry, Vol. 81, No. 19, October 1, 2009

exchange resin with temperature. Hydration of ions can be assumed to become weaker as the temperature increases due to disruption of ion-dipole bonding by thermal motion of water molecules resulting in reduction of the difference in hydration structure between different ions. This temperature effect may be greater for more weakly hydrated ions (alkali metal ions) than for more strongly hydrated ions (alkaline-earth metal ions). We have substantiated by examining the variation of separation selectivity for tetraalkylammonium ions with temperature that the size of ionic species is also a decisive factor determining the ionexchange selectivity via electrostatic interaction in superheated water. Tetraalkylammonium ions are mainly located in the vicinity of the fixed ions in the cation-exchange resin and have less affinity for the exchanger than sodium ion in superheated water. The difference in hydrophobicity between the cation-exchange resin phase and the external solution phase is reduced by an increase in temperature so that the electrostatic interaction or association of the ions with the fixed ion becomes a predominant mechanism. These results suggest that use of superheated water as the mobile phase for ion-exchange chromatography may bring about different separation selectivity from that obtained at ambient temperature in addition to the possibilities of substantially increasing the analysis speed and column efficiency due to the decrease in mobile phase viscosity. ACKNOWLEDGMENT This research was supported by a Grant-in-Aid for Scientific Research Grant No. 20350034 from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and a grant from Saitama University. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review June 1, 2009. Accepted August 25, 2009. AC9011864