Sorption Isotherms of Ternary Eluents in Reversed-Phase Liquid

Jan 6, 2009 - Absolute isotherms were estimated from the experimental data for excess volumes by use of various strategies for the estimation of the v...
0 downloads 5 Views 822KB Size
Download PDF
Anal. Chem. 2009, 81, 984–990

Sorption Isotherms of Ternary Eluents in Reversed-Phase Liquid Chromatography Mei Wang, Jennifer Mallette, and Jon F. Parcher* Chemistry Department, University of Mississippi, University, Mississippi 38677 The experimental technique of mass spectrometric tracer pulse chromatography was used to measure the excess volume of each eluent component for binary and ternary mixtures of water, acetonitrile, and methanol on a C18bonded silica RPLC packing over the full composition range. The tracer pulse method allowed the direct measurement of excess volumes of each eluent component without numerical integration, assumed isotherm equation, detector calibration, or off-line analysis of the eluent composition. Absolute isotherms were estimated from the experimental data for excess volumes by use of various strategies for the estimation of the volumes of the stationary and mobile phases in dynamic equilibrium with eluents of varying composition. The results indicate that all three eluent components interacted with the alkane bonded phase. Some components were selectively taken up as part of the stationary phase while other components were selectively excluded so the composition of eluent in or on the stationary phase often significantly differed from the composition of the bulk eluent. The exact composition of the stationary phase (bonded phase plus immobilized eluent) was dependent upon the type and composition of the bulk eluent. Eluent composition is the primary control parameter for optimization of reversed-phase liquid chromatography (RPLC) systems. Temperature and pressure do affect analysis time and column efficiency but have only a secondary effect upon the resolution of analytical mixtures. The exact role of the eluent in the retention of analytical solutes is not yet well defined despite decades of investigations by myriad researchers. The choice of eluents is often determined by precedent or experience of the operator rather than systematic optimization or rationale. The significant influence of the eluent in RPLC is due to the fact that eluent-analyte molecular interactions can influence the distribution of analyte between the stationary and mobile phases. Moreover, the eluent can influence the volume, composition, and chemical properties of typical alkane-bonded stationary phase packings because of the uptake of eluent by these packings. Thus, the stationary phase in RPLC is often composed of a solid silica surface with residual silanol groups, a bonded alkane moiety, and one or more components of a typical aqueous-organic (binary or ternary) eluent. The eluent components in the stationary phase are not truly “stationary” for an extended time period. The eluent * To whom correspondence should be addressed. E-mail: [email protected].

984

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

components exist in dynamic equilibrium between the stationary and mobile phases. The amount of eluent resident in the stationary phase at any given time may vary dramatically with the type and composition of the eluent. Even the mechanical structure of the bonded “liquid” phase may vary between an extended (liquid-like) conformation and a collapsed (solid-like) state. The physical character of a stationary phase can change during a gradient elution experiment. This phenomenon may also increase the time required to re-establish the initial conditions of a gradient elution experiment. Accurate measurement of adsorption or absorption isotherms of eluent components and solutes is essential for the elucidation of retention mechanisms, as well as prediction of peak shapes and resolution of eluted solutes. Because the stationary phase structures are formed dynamically within the column, it is difficult to characterize these multicomponent assemblies by static or spectroscopic methods. Three in situ chromatographic methods have been used to measure multicomponent sorption isotherms in RPLC systems. These methods are concentration pulse, tracer pulse, and frontal analysis chromatography. Concentration pulse chromatography has been employed extensively to measure sorption isotherms in RPLC systems.1 This technique is, however, not practical for systems with more than two components. Frontal analysis (FA) and frontal analysis by characteristic points (FACP) have both been used to measure sorption isotherms of binary systems.2,3 FA is the primary chromatographic method used to date4,5 for the determination of sorption isotherms of ternary solute systems. This technique is labor intensive for multicomponent systems because the composition of the eluent in the plateau regions must often be determined by fraction collection and offline analysis. This experimental approach also requires columns with a significant number of theoretical plates to establish sharp fronts and clearly defined plateau regions often obscured by the deleterious effects of nonequilibrium, kinetic processes. The two previous chromatographic investigations of ternary solute systems4,5 were very similar and both involved FA experiments with three alcoholic solutes with a binary eluent consisting of a 1:1 mixture of methanol and water. In one case, the void volume was determined from the retention volume of thiourea.5 In the other investigation, no specific method was given for the Gritti, F.; Guiochon, G. J. Chromatogr. A 2005, 1099, 1–42. Jandera, P.; Komers, D. J. Chromatogr. A 1997, 762, 3–13. Jandera, P.; Posvec, Z.; Vraspir, P. J. Chromatogr. A 1996, 734, 125–136. Lisec, O.; Hugo, P.; Seidel-Morgenstern, A. J. Chromatogr. A 2001, 908, 19–34. (5) Quinones, I.; Ford, J. C.; Guiochon, G. Chem. Eng. Sci. 2000, 55, 909– 929. (1) (2) (3) (4)

10.1021/ac8020104 CCC: $40.75  2009 American Chemical Society Published on Web 01/06/2009

measurementofthecolumnvoidvolume.4 Numerousinvestigations6-12 have shown that both water and methanol are adsorbed on or absorbed in the alkane-bonded stationary phase with a 1:1 mixture as the eluent. Thus, the experimental systems described above actually contained five adsorbable components. The maximum concentrations of the three solutes used in the experiments were relatively low (mole fraction < 0.005)5 so the concentration of methanol and water in the eluent, and presumably the stationary phase, were essentially constant. What, if any, influence the eluent composition would have on the competitive isotherms of the three adsorbates (solutes) was not determined. Jandra, et al.2,3 did investigate the effect of the composition of aqueous eluents with up to 40% methanol on the competitive isotherms of pure phenol, cresol, or resorcinol and several binary mixtures on a C18-bonded phase using frontal analysis experiments. These authors used the concentration pulse method with pure methanol to determine the volume of the mobile phase. They found that both the saturation capacity and the limiting slope of the isotherms decreased dramatically with increasing methanol concentration. This would indicate that the organic eluent competed very effectively with the solutes for occupancy on or in the bonded stationary phase. Again in these experiments, the solute concentrations were relatively low presumably because of solubility limits in the aqueous eluents. Zhu et al.13 used frontal analysis with a mass specific detection system to estimate the bonding constants for a series of epidermal growth factor receptor inhibitors. The advantage of this technique, known as frontal affinity chromatography, is the use of extracted ion chromatograms to produce a single step function at a specific m/z value for each component of a mixture of analytes rather than the multistep frontalgram that would be observed with a nonspecific detector. These authors presented an example where the binding constants for six analytes were determined simultaneously. Such an analysis of a six-component system would be difficult if not impossible with a nonspecific detector. In 1985, Knox and Kaliszan14 used tracer pulse chromatography with radioactive tracers to measure the uptake by C18-bonded silica of binary and ternary mixtures of water, carbon tetrachloride, and acetonitrile. The data set was rather sparse with only five ternary compositions investigated; however, the authors showed that the technique was viable and accurate. They introduced the concept of a thermodynamic void volume as the total amount of eluent in the column. This quantity was experimentally determined from the sum of the products of the retention volumes of the tracers multiplied by the volume fraction of that component in the eluent. These authors also observed the rather unexpected result that in binary mixtures of ethanol and acetonitrile, ethanol was preferentially sorbed while acetonitrile was excluded from the stationary phase. This (6) Poppe, H. J. Chromatogr. A 1993, 656, 19–36. (7) Slaats, E. H.; Markovski, W.; Fekete, J.; Poppe, H. J. Chromatogr. A 1981, 207, 299–323. (8) Foti, G.; Reyff, C.; Kovats, E. Langmuir 1990, 6, 759–766. (9) Zhu, P. L. Chromatographia 1985, 20, 425–433. (10) McCormick, R. M.; Karger, B. L. Anal. Chem. 1980, 52, 2249–2257. (11) Wang, M.; Mallette, J.; Parcher, J. F. Anal. Chem. 2008, 80, 6708–6714. (12) Alvarez-Zepeda, A.; Martire, D. E. J. Chromatogr. 1991, 550, 285–300. (13) Zhu, L.; Chen, L.; Luo, H.; Xu, X. Anal. Chem. 2003, 75, 6388–6393. (14) Knox, J. H.; Kaliszan, R. J. Chromatogr. 1985, 349, 211–234.

was in apparent conflict with the observation that acetonitrile was preferentially sorbed from solutions of acetonitrile and water. Samuelsson et al.15 used tracer pulse chromatography with a mass specific detection system to measure the isotherms of binary mixtures of methyl- and ethyl mandelate on a C8-bonded phase from an aqueous eluent containing 30% acetonitrile. In this case, uracil was used to determine the volume of mobile phase in the column. Even more recently,11 mass spectrometric tracer pulse chromatography was used to measure the sorption isotherms of RPLC eluents consisting of binary mixtures of tetrahydrofuran, methanol, or acetonitrile in water on a C18bonded phase over the full composition range. In this case, the primary experimental data were excess volumes of each component in the stationary phase. The absolute volumes of each component in the stationary phase were subsequently calculated by various strategies for measuring the volume of the mobile phase.16 It was shown that this experimental technique was viable and accurate for the determination of binary competitive isotherms of eluent components. However, this experimental method has not yet been used to determine competitive adsorption isotherms of multicomponent systems containing more than two components. The objective of the current work was to extend the experimental technique to ternary systems consisting of mixtures of water, methanol, and acetonitrile. EXPERIMENTAL SECTION The experimental procedures, columns, and protocol have been described previously.11 The estimated volume of bonded octadecane in the column was 0.67 mL assuming a packing weight of 2 g. The RPLC column was obtained from SGE Inc. The column dimensions were 250 × 4.6 mm. The packing had a particle diameter of 5 µm, pore size of 80 Å, and the surface area of the base silica was 300 m2/g. The reported bonding density was 3.4 µmol/m2. The experimental procedure consists simply of equilibrating a column with a given eluent followed by triplicate injections of a mixture of isotopically labeled tracers of each of the eluent components. The isotopic tracers are detected in a high background of unlabeled eluent by a mass spectrometer operated in the selected ion monitor mode to detect only the masses of the tracers. Typical results are illustrated in Figure 1. The total volume of each individual eluent component in the column, Vi, can be calculated from the retention volume of / the isotopic tracer of that component, VR,i , by the relation Vi ) / M M VR,iθi where θi represents the volume fraction of component i in the eluent. The absolute volume of component i in the stationary phase, Vis, can be determined from the relation / - VM]θiM ViS ) [VR,i

(1)

where VM represents the volume of eluent in the mobile phase. The quantity VM is very difficult to measure in any RPLC system and has been a source of controversy in the literature for decades. The most common method for the determination (15) Samuelsson, D.; Arnell, R.; Diesen, J. S.; Tibbelin, J.; Paptchikhine, A.; Fornstedt, T.; Sjoberg, P. J. R. Anal. Chem. 2008, 80, 2105–2112. (16) Wang, M.; Mallette, J.; Parcher, J. F. J. Chromatogr. A 2008, 1190, 1–7.

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

985

Figure 1. MSTPC chromatogram of a mixture of isotopic tracers for a ternary eluent containing acetonitrile, methanol, and water. The chromatograms were offset and equalized in intensity for clarity of presentation.

represents the results obtained using D2O as the unretained marker for eluent concentrations less than 40% methanol. The Langmuir isotherm equation was fit to these data, and the derived parameters were used to extend the calculated isotherm to eluent compositions of greater than 40% methanol.12 The circles represent the results from a regression analysis of excess volume data over the range of 0.5-0.8 volume fraction of methanol.18 The errors introduced by the use of thiourea or uracil are obviously unacceptable for isotherm measurements in this particular system. Because of the lack of a satisfactory dead time probe solute, the primary experimental data obtained from mass spectrometric tracer pulse chromatography (MSTPC) experiments, like any other experimental technique for liquid-solid adsorption, with RPLC systems is the excess amount or volume of each component in the stationary phase. The excess volume of eluent in the stationary phase, ViXS, was determined from the relation11 / ViXS ) [VR,i - V0]θiM

(2)

where V0 is the thermodynamic void volume of the column.14 This type of void volume represents the total volume of eluent in the column regardless of mobile and stationary phases, thus k

V0 )

Figure 2. Absolute sorption isotherms for methanol with C18-bonded silica with the mobile phase volume determined by four different techniques. Solid line (]): calculated Langmuir isotherm assuming M D2O is unretained for θMeOH e 0.4. (O) calculated from the excess M volume data for 0.5 e θMeOH e 0.8. (2) Assuming thiourea is unretained. (9) Assuming uracil is unretained.

∑θ

M / k VR,k

(3)

where the summation refers to all the components of the eluent. The measured values for the thermodynamic void volume for the column used in this study was 2.60 mL with a standard deviation of 0.05 for 49 data points including pure binary and ternary eluents. The excess volume of the eluents in the stationary phase is easy to measure; however, it is not the quantity needed for the determination of thermodynamic properties of the RPLC systems, such as the equilibrium constant for the distribution of a component between the mobile and the stationary phases. This equilibrium constant for component i is generally expressed as S M Ki ) CSi /CM i where Ci and Ci represent the molar concentration of component i in the stationary and mobile phases, respectively. The concentration, CiS, is proportional to the excess volume, ViXS, from the relation

of VM is to measure the retention volume of a solute that hypothetically does not interact with the stationary phase. / Thus,VR,i ) VM for a component with VtS ) 0. A wide variety of “unretained” solutes have been used for isotherm and other thermodynamic measurements in RPLC systems.17 However, no single solute has been found that is truly unretained in typical RPLC systems over the full range of eluent compositions. Some solutes, in particular water, are unretained by alkane-bonded packings with water-rich eluents but are significantly retained with water deficient eluents.12,16 The choice of methods for estimating the volume of the mobile phase in RPLC systems is particularly important because it can have a significant influence on the calculated adsorption or partition isotherms. Figure 2 illustrates the influence of the type of dead time marker (“unretained” solute) on the measured absolute isotherm calculated for methanol in a C18-bonded packing.11,16 The solid line

RESULTS AND DISCUSSION Binary and Ternary Excess Isotherms. Tracer pulse experiments were carried out over the full composition range with binary

(17) Rimmer, C. A.; Simmons, C. R.; Dorsey, J. G. J. Chromatogr. A 2002, 965, 219–232.

(18) Schay, G. In Surface and Colloid Science; Wiley-Interscience: New York, 1969; Vol. II, p 180.

986

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

CiS ) CiM +

ViXS viVs

(4)

where vi represents the molar volume of component i and VS is the volume of the eluent in the stationary phase. In RPLC systems it is generally assumed the molar volume of a given component is the same in both the stationary and the mobile phases. The problem arises in the exact determination of the volume of eluent in the stationary phase. Thus, it is difficult to directly determine CiS from experimental data for ViXS.

Figure 3. Ternary diagram of the eluent compositions used for the competitive, multicomponent isotherm experiments.

eluents and over sixteen intermediate compositions with ternary eluents. The complete map of eluent compositions used in the experiments is given in a ternary diagram (Figure 3). At each eluent composition, the excess volume of each eluent component was determined from the measured retention volume of a labeled isotopic tracer using eq 2. The results of these experiments are given in Supporting Information, Table S-1 for all of the eluent compositions denoted in Figure 3. The excess volumes of the pure eluent components must be zero and the excess volumes of all of the components of any eluent must sum to zero because that is the basic assumption of the vNA convention19 used for RPLC systems and the derivation of eq 2. The binary isotherms for the three eluent components are shown in Figure 4. Each panel of the figure represents the excess sorption (volume) of one eluent component in a binary mixture. For a given binary eluent, the measured excess volumes must sum to zero. The mirror image isotherms of the binary eluents are indicated by the same symbols (circle, triangle, and square). The measured excess volumes were positive for methanol in both acetonitrile and water. The excess volumes of water were always negative in both acetonitrile and methanol. On the other hand, the excess volumes of acetonitrile were positive in water (except for almost pure acetonitrile) but negative in methanol. This observation is in agreement with the results of Knox and Kaliszan for a similar system involving ethanol and acetonitrile.14 Excess acetonitrile was taken up by the bonded phase to a greater extent (0.22 mL) than methanol (0.06 mL). This phenomenon has been observed previously by several authors.20-23 These quantities represent the maximum uptake for the entire column with an estimated volume of bonded phase of about 0.67 mL. The question of whether the eluent dissolved in or adsorbed on the bonded phase could not be addressed from the excess volume data alone. Although this issue is not settled, recent usage (19) Riedo, F.; Kovats, E. J. Chromatogr. 1982, 239, 1–28. (20) Chan, F.; Yeung, L. S.; LoBrutto, R.; Kazakevich, Y. V. J. Chromatogr. A 2005, 1082, 158–165. (21) Kazakevich, Y. V.; LoBrutto, R.; Chan, F.; Patel, T. J. Chromatogr. A 2001, 913, 75–87. (22) Gritti, F.; Kazakevich, Y. V.; Guiochon, G. J. Chromatogr. A 2007, 1169, 111–124. (23) Gritti, F.; Guiochon, G. J. Chromatogr. A 2007, 1155, 85–99.

Figure 4. (A) Excess sorption isotherms of methanol with binary eluents containing methanol and (2) acetonitrile or (b) water. (B) Excess sorption isotherms of water with binary eluents containing water and (0) acetonitrile or (O) methanol. (C) Excess sorption isotherms of acetonitrile with binary eluents containing acetonitrile and (4) methanol or (9) water.

has tended toward the adsorption model. Kazakevich21,24 has presented substantial evidence for the adsorption of eluents on the surface of bonded phases. More recently, the same group25 looked at the retention volumes of a given solute on a series of C18-bonded silica adsorbents. They found that the adjusted retention volume (VR,i - V0) for a given solute was proportional to the measured surface area of the various adsorbents and not proportional to the mass of “octadecane” on the silica. (24) Rustamov, I.; Farcas, T.; Ahmed, F.; Chan, F.; LoBrutto, R.; McNair, H. M.; Kazakevich, Y. V. J. Chromatogr. A 2001, 913, 49–63. (25) Giaquinto, A.; Liu, Z.; Bach, A.; Kazakevich, Y. Anal. Chem. 2008, 80, 6358– 6364.

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

987

Figure 6. Full excess isotherm for water from binary and ternary eluents. Figure 5. Full excess isotherm for acetonitrile on C18-bonded silica from binary and ternary eluents. The data points correspond to the compositions for ternary systems shown in Figure 3.

Recent computer simulations26-28 indicate that the eluent components concentrate at the surface of the bonded phase. Most contemporary publications involving measurements of the uptake of eluent refer to “adsorption” isotherms.2-5 To illustrate the results for the competitive, ternary isotherms, it is necessary to display the results on a three-dimensional grid. Simple projection of the excess volume data onto a single concentration axis is completely unsatisfactory because of the ternary nature of the eluent. The results for the measured excess volume of acetonitrile for all binary and ternary eluent compositions are shown in three dimensions in Figure 5. In the figure, the blue points reflect the compositions of the ternary eluents. Acetonitrile was adsorbed significantly from eluents containing water; however, adsorption of acetonitrile diminished with eluents containing methanol. The excess isotherm for water is illustrated in Figure 6. In this case, water was universally excluded from the stationary phase by all eluents with the greatest extent of exclusion observed for eluents containing acetonitrile and water. This means that the volume fraction of water in the adsorbed layer was always less than the volume fraction of water in the bulk eluent. The excess isotherm for methanol was essentially flat showing little or no sorption except for water-rich eluents in which the maximum sorption was only 0.06 mL. The general conclusion derived from the data would be the often expressed idea that water in the eluent induces the uptake of eluent organic components and likewise analytes. Binary and Ternary Absolute Isotherms. Excess adsorption data provide an insight into the mechanisms for the uptake of eluent and analyte by RPLC stationary phases. However, this type of experimental data does not allow the evaluation of thermodynamic parameters necessary for the elucidation of retention mechanisms. It is desirable to determine the absolute amount of eluent or analyte present in the stationary and mobile phases for this purpose. The direct conversion of excess data to absolute (26) Rafferty, J. L.; Siepmann, J. I.; Schure, M. R. J. Chromatogr. A 2008, 1204, 11–19. (27) Rafferty, J. L.; Siepmann, J. I.; Schure, M. R. J. Chromatogr. A 2008, 1204, 20–27. (28) Zhang, L.; Rafferty, J. L.; Siepmann, J. I.; Chen, B.; Schure, M. R. J. Chromatogr. A 2006, 1126, 219–231.

988

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

Table 1. Absolute Volumes Calculated from the Linear Region of the Excess Volume Isotherms system component 1/2 Acetonitrile/Water Methanol/Water Methanol/Acetonitrile

volume of eluent in volume of component 1 in the stationary the stationary phase, phase, VS1 (µL) VS (µL) 530 ± 30 106 ± 3 70 ± 4

450 ± 20 91 ± 3 62 ± 3

data requires an accurate assessment of the volumes of the stationary and mobile phases. In the absence of such information it is necessary to invoke an indirect method for the determination of VS and VM.16 The experimental binary isotherms shown in Figure 4 all have a characteristic shape with a linear region (nonstationary inflection point) in the range 0.5 e θM i e 0.8. Equation 4 is valid throughout the eluent composition range. This equation, along with the definitions viCiM ) θiM and CiSVSvi ) ViS, can be rearranged to give ViXS ) ViS - VSθiM

(5)

This relation is also valid over the full composition range; however, within the inflection region, the values of ViS and VS can be considered constant and determined from the intercept and slope of the linear region of the excess isotherm.18,29 Regression of the data in the Supporting Information, Table S-1 yielded the results given in Table 1. This strategy produces absolute values for the volume of the eluent in the stationary phase. These values are, however, only valid for the range of eluent compositions encompassing the linear regions of the excess isotherms. To determine the absolute amount of an eluent component in the stationary phase outside this limited range, it is necessary to invoke a secondary assumption. For aqueous eluents, it is often assumed that water is not adsorbed by C18-bonded packings for water-rich eluents, that is, θHM2O g 0.5. Thus, the retention volume of D2O would be an accurate measure of the mobile phase volume within this restricted composition range.12 It is also reasonable to assume (29) Everett, D. H. Pure Appl. Chem. 1986, 58, 967–984.

Figure 8. Full, absolute isotherm for acetonitrile on C18-bonded silica at 25 °C from a ternary eluent composed of water, methanol, and acetonitrile.

Figure 7. (A) Derived absolute isotherms of (O) methanol and (0) water from binary aqueous eluents. (B) Derived absolute isotherms of (4) acetonitrile and (0) water from binary aqueous eluents.

that the total volume of eluent adsorbed with a pure eluent is equal to the VS value calculated from the linear excess sorption range. These assumptions allow the calculation of the absolute isotherms of mixtures of acetonitrile and methanol with water. The results are shown in Figure 7. Even though two independent methods were used to calculate the absolute isotherms over distinct composition ranges, the results are in relatively good agreement. To develop a three-dimensional, absolute isotherm for the sorption of one eluent component, it was necessary to make additional assumptions. For example, to calculate the full, absolute isotherm for acetonitrile, the following assumptions were required: (1) the binary isotherm for acetonitrile/water was the same as that shown in Figure 7B. (2) the binary isotherm for acetonitrile in water/methanol was zero. (3) the binary isotherm for acetonitrile/methanol was that given in Table 1 for methanol, that is, 70-62 ) 8 µL of acetonitrile M M e 0.5 and 535 µL at θACN for the composition range 0.2 e θACN ) 1. 4) the ternary isotherm for acetonitrile in the water-rich ternary eluent was calculated from eq 1 assuming that water was not / adsorbed (VM ) VR,D2O). 5) the ternary isotherm for the methanol-rich eluent was calculated as the average of the two binary isotherms, that is, θSACN ) 5 µL.

6) the ternary isotherm for the acetonitrile-rich eluents were calculated from the linear region of the excess isotherms for the M range 0.5 e θACN e 0.8. The full, three-dimensional, absolute isotherm for acetonitrile is shown in Figure 8. The maximum uptake of acetonitrile was about 500 µL. The volume of bonded octadecane in the column was about 670 µL. The pore size of the solid adsorbent was 80 Å. The total pore volume was ∼2,000 µL. The absolute isotherm for methanol was calculated in a similar manner; however, the maximum uptake of methanol was only ∼150 µL. This would correspond to a maximum volume fraction of 0.1-0.2. Water was not sorbed to any significant extent (e75 µL) by the C18-bonded phase for any eluent composition. That is, the volume fraction of water in the immobilized eluent was always less than that of the bulk eluent. CONCLUSIONS Mass spectrometric tracer pulse chromatography is a viable experimental technique for the accurate determination of excess uptake of multicomponent eluents by bonded-phase packing commonly used for RPLC. The technique is experimentally complex; however, it produces isotherm data directly without the need for assumed isotherm models or complex integration schemes. To measure absolute sorption isotherms directly, it is necessary to use an unretained solute probe (dead time marker) to determine the volume of eluent in the mobile phase. The two most commonly used dead time probes for RPLC systems are thiourea and uracil. In this study, both solutes were investigated as dead time markers; however, the results indicate that neither solute produced an accurate measure of the mobile phase volume. Thus, the tracer pulse method could be used only to produce excess amounts taken up by the packing. Absolute isotherms were determined indirectly from the excess data by a variety of strategies.16 D2O was found to be an acceptable dead time marker for water-rich eluents with θDM2O g 0.5. In this case the absolute volume of eluent in the Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

989

stationary phase was calculated from eq 1. In every case, the excess isotherms for the preferentially sorbed eluent component showed a linear region at high concentrations of that component. Regression of this linear data could be used to estimate the volume of that eluent component in the stationary phase, as well as the total volume of eluent in the stationary phase. In this study, it was shown that the organic component of aqueous/organic eluents was always preferentially taken up by the bonded RPLC packing at the expense of water. If the eluent dissolved in the bonded octadecane, this would significantly influence the partition coefficient of analytical solutes. If, on the other hand, the immobilized eluent formed a separate phase on the surface of the octadecane, the stationary phase would then be composed of the surface of the octadecane and the immobilized eluent.21 This unique solid phase would display different adsorptive properties than that of the pure octadecane.

990

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

ACKNOWLEDGMENT This research was supported by Grant CHE-0715094 from the National Science Foundation. The University of Mississippi provided funds for the instrumentation.

SUPPORTING INFORMATION AVAILABLE Table S-1 with experimentally determined excess volumes of all eluent components. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review September 22, 2008. Accepted December 9, 2008. AC8020104