ANALYST, MAY 1992, VOL. 117 83 1 Solute-Mobile Phase and Solute-Stationary Phase Interactions in Micellar Liquid Chromatography A Review Maria Jose Medina Hernandez and Maria Celia Garcia Alvarez-Coque” Departamento de Quimica Analitica, Facultad de Quimica, Universidad de Valencia, 46100 Burjassot, Valencia, Spain Summary of Contents I nt rod ucti on Partition Behaviour Electrostatic and Hydrophobic Interactions Binding, Non-binding and Antibinding Solutes Influence of pH Ionic Strength Selectivity With Purely Micellar Eluents Addition of Modifiers to Micellar Eluents Solvent Strength Selectivity With Hybrid Micellar Eluents Use of Reversed Micelles in Liquid Chromatography References Keywords : Re view; reversed-p hase liquid chroma tog raph y; micellar eluen ts; solute interactions; solvent strength; efficiency Introduction Micellar liquid chromatography (MLC) with normal micelles is an alternative to conventional reversed-phase liquid chro- matography (RP-LC), which uses a surfactant solution above the critical micellization concentration (c.m.c.) as the mobile phase, instead of hydro-organic mixtures.1 4 Micelles are not static, but exist in equilibrium with surfactant monomers above the c.m.c. Adsorption of these monomers on alkyl-bonded stationary phases (e.g., C1, C8 and Clx) could occur in at least two ways:s (a) hydrophobic adsorption, where the alkyl tail of the surfactant would be adsorbed and the ionic head group would then be in contact with the polar solution, giving the stationary phase some ion-exchange capacity with charged solutes; and (6) sylano- philic adsorption, where the ionic head group of the surfactant would be adsorbed, the stationary phase becoming more hydrophobic. For most surfactants and stationary phases, the amount of surfactant adsorbed on the stationary phase remains constant after equilibration once the concentration of surfactant is above the c.m.c.6.7 The complexity of MLC is much greater than conventional RP-LC with hydro-organic solvents, owing to the large number of possible interactions (electrostatic, hydrophobic and steric) with the micellar mobile phase and with the modified stationary phase (Fig.1). None of these interactions can occur for a hydro-organic system.8 Almost any compound can be determined by MLC. The separation of inorganic anions9 and of dithiocarbamates10-11 with a micellar mobile phase of hexadecyltrimethylammonium chloride (CTAC) and bromide (CTAB), respectively, has been reported.Complete resolution of the cis and trans isomers of anionic cobalt(iii)-iminodiacetate was achieved with CTAB, and neutral 4,4’-ethylenedinitrilobispentan-2- one complexes of copper(1i) and nickel(r1) were separated with a sodium dodecyl sulfate (SDS) micellar mobile phase.12 A concentrated micellar solution of Brij 35 has been used to extract aldehydes from tobacco samples, which were sep- arated chromatographically with a more dilute solution of the same surfactant .I3 The activity of folylpolyglutamate hydro- lase in crude tissue extracts was determined after denaturation of the enzyme in SDS, which subsequently served as the micellar solvent system for the chromatographic separation of substrate from reaction products.14 Excellent correlations have been found between capacity factors with a tetradecyl- trimethylammonium bromide mobile phase and the bioactiv- ity of 26 para-substituted phenols,lS and between capacity factors with a 0.03 mol dm-3 SDS mobile phase and the site of action of diuretics along the nephron.14 One of the most interesting applications of MLC is the possibility of determining drugs in biological fluids without previous separation of proteins. 17-24 Micellar solutions of SDS or Brij-35 solubilize the proteins in the biological sample and cause them to be eluted at the front of the chromatogram (Fig. 2). Partition Behaviour Armstrong and Nome27 proposed a three-phase model (sta- tionary phase, and bulk aqueous and micellar pseudo-phases) to explain the chromatographic behaviour of a solute eluted with an aqueous micellar mobile phase containing a surfac- tant.The solutes partition not only between water and the stationary phase, but also inside the mobile phase, between water and the micelle. Hence, elution of a solute in MLC depends on three partition coefficients: that between the stationary phase and water ( Psw), between the stationary phase and the micelle (fSM), and between the micelle and water (f MW). First, Armstrong and Nome,27 and later Arunyanart and Cline Love,28 proposed different models to describe the change in retention of solutes at various micelle concentra- tions. The equations can be rewritten as: *- To whom correspondence should be addressed.832 ANALYST, MAY 1992, VOL.117 0- M ice1 le 43 Bulk water / 0- MicelleG Bulk w r r .................. Fig. 1 Solute-micelle and solute-stationary phase hydrophobic (-) and electrostatic interactions (4) with an anionic surfactant: ( a ) apolar solute; ( b ) anionic solute; and ( c ) cationic solute where k' is the capacity factor, [MI is the total concentration of surfactant in the mobile phase minus the c.m.c., Q, is the ratio of the volume of the stationary phase, Vs, to the volume of the mobile phase, VM, in the column, and KAM is the solute- micelle binding constant. By plotting llk' versus [MI one should obtain a straight line. The values of Psw and KAM are given by the slope and intercept of the plot, respectively.The calculation of Psw requires knowledge of V,, which cannot easily be determined. Usually, the difference between the empty column volume and the packed column void volume is taken as V s . This difference gives an overestimation of Vs because it includes the entire volume occupied by the silica solid support particles rather than just the true stationary phase. The use of such a Vs value can be expected to result in a Psw coefficient that is significantly in error. An approach that completely excludes any volume associated with the base silica material should be used.29 Eqn. (1) can be used to describe the retention of apolar, polar, and even ionic solutes, chromatographed with anionic, cationic and non-ionic ~urfactants.8~13 For high relative molecular mass solutes, intercepts are nearly zero in the llk' Time - Fig.2 (a) Chromatogram of urine. (b) Chromatogram of urine spiked with: 1,30 pg ml-1 of amiloride; 2,5 pg ml-1 of spironolactone; 3, 1.2 pg ml-1 of metandienone; 4, 56 pg ml-1 of amino(pheny1)- pro anol; and 5,30 p ml-1 of clostebol. Mobile phase, 0.1 mol dm-3 SD!? solution with 3 2 pentan-1-01; column temperature, 60 "C; flow rate, 1 ml min-1; UV detection, 260 nm. Reprinted, with permission, from ref. 26 versus [MI plot. A zero intercept requires either that the reciprocal of the phase ratio must be zero, which is not physically possible, or that Psw is extremely large. This is not only physically possible but also consistent with solubility data for compounds that show this behaviour (e.g., alkylbenzene homologues beyond butylbenzene are insoluble in water).The direct transfer of these compounds from the micellar pseudo-phase to the surfactant-modified stationary phase, via reversible sorption of the solute-occupied micelle onto the 'hemimicellar' surfactant-modified stationary phase, has been suggested .29 Electrostatic and Hydrophobic Interactions The non-homogeneous nature of micelles creates a unique situation in which different solutes can experience various micro-environment polarities in a given mobile phase. Reten- tion of a solute will depend on the type of interaction with the micelle and the surfactant-modified stationary phase. Non- polar solutes, such as benzene and toluene, should only be affected by hydrophobic interactions [Fig.l(a)], but for solutes that are charged, two distinct situations can be considered: (i) the charge on the solute and surfactant has the same sign [Fig. l(b)]: or (ii) the charge on the solute and surfactant has the opposite sign [Fig. l(c)].3() The first situation is encountered when an anionic solute is chromatographed with an anionic surfactant or a cationic solute with a cationic surfactant [ e . g . , dissociated phenol and 2-naphthol with SDS, and protonated benzylamine with dodecyltrimethylammonium bromide (DTAB) on a C18 column].3* Electrostatic repulsion from the micelle should notANALYST, MAY 1992, VOL. 117 833 affect retention as the solute would still reside in the bulk mobile phase and, therefore, still move down the column.In contrast, repulsion from the surfactant-modified stationary phase should cause a decrease in retention. Solutes may be eluted in the void volume. However, they may also be retained if hydrophobic interaction with the stationary phase exists. Because of the different hydrophobic interaction, dissociated phenol and 2-naphthol are well separated with SDS. The second situation appears when a solute is chromato- graphed with an oppositely charged surfactant, where elec- trostatic attraction occurs between both species. If the electrostatic attraction with the micelle is complemented by a hydrophobic interaction, the solutes will remain in the mobile phase for a longer period of time, and retention will decrease. However, electrostatic and hydrophobic interactions with the stationary phase may be sufficiently large to offset the increase in micellar attraction and would increase retention.Disso- ciated phenol and 2-naphthol are retained to a greater extent with DTAB than with SDS on a Clx column.31 With an appropriate surfactant, mixtures of polar and apolar solutes can be resolved adequately.3’ For example, dissociated phenol and benzene are not well resolved with DTAB, but are completely resolved with SDS. In contrast, p-nitrophenol and p-nitroaniline are not separated with SDS, but are well resolved with DTAB.31 Binding, Non-binding and Antibinding Solutes The function of the micellar pseudo-phase in MLC has been compared with that of the organic modifier in traditional RP-LC, as for many solutes, an increase in the concentration of surfactant in the mobile phase results in a decrease in the retention of the solutes being separated.However, the eluent strength increases with micelle concentration only if the solute interacts with the micelle in the mobile phase. Armstrong and Stine33 proposed a classification of the solutes into three groups according to their chromatographic properties with a micellar mobile phase: (i) solutes binding to micelles; (ii) non-binding solutes; and (iii) antibinding solutes. Compounds that associate or bind to micelles show decreased retention when the concentration of micelles in the mobile phase is increased ( K A M >0).34 For compounds that do not associate with micelles, retention can remain unaltered by the micelle content of the mobile phase (non-binding, K A M = 0) o r their retention can increase with increasing micelle concen- tration (antibinding, K A M <0).Antibinding results from a compound being strongly excluded or repelled from the micelle. High positive values of K A M have been observed with solutes showing electrostatic interactions ( e . g . , benzyl- trimethylammonium bromide with SDS micelles, and benzoic acid with CTAB micelles), o r with co-micellization [ e . g . , sodium octylbenzenesulfonate with SDS, and cetylpyridinium chloride (CPC) with CTABl.8 The similar K A M values for CPC and benzoic acid in CTAB micellar phases explain the similar retention observed for both solutes; however, the location of the solute in the micelle is very different: benzoic acid is bound onto the micelle surface, in the Stern layer, whereas CPC occupies the same site as a CTAB molecule in the micelle, the alkyl tail being in the micellar core and the polar head in the Stern layer.Negative K A M values apparently have no meaning. However, just as compounds that bind to micelles each have a characteristic positive constant, compounds that are excluded from the micelle may have a characteristic negative constant.33 Most non-binding compounds with anionic micelles are negatively charged and with cationic micelles are positively charged. It is apparent that electrostatic repulsion is an important factor in antibinding behaviour. However, there are also many positively charged compounds that bind to cationic micelles in addition to negatively charged compounds that bind to anionic micelles.It is, therefore, sometimes difficult to predict the exact retention behaviour of an organic ion. Antibinding has never been observed between a charged solute and an oppositely charged micelle.33 These effects cannot be observed by using stationary phases that adsorb an appreciable amount of surfactant, i . e . , Cg- or Clg-bonded phases.33 For these phases, when the stationary phase acquires the same charge as the micelle, and no hydrophobic interaction occurs, similarly charged solutes tend to elute in the void volume of the column. When using a C1 or preferably a cyano-bonded phase, however, one can observe increased retention when eluting. The antibinding phenomenon is useful in MLC because it produces unusual selectivities.35 On a cyano column and with SDS in the mobile phase, neutral phenol behaves as a binding compound, whereas the anionic naphthalene-2-sulfonate behaves as an antibinding compound because electrostatic repulsion is stronger than hydrophobic interaction.In con- trast, for pyrene-1-sulfonate, with a binding behaviour, the larger pyrene moiety should produce a counterbalancing of the electrostatic repulsion and associate more strongly with the micelle. By using a Clx column, where negatively charged surfactant monomers are adsorbed, the elution behaviour of phenol and pyrene-1-sulfonate, where hydrophobic effects would dominate, is very similar to that obtained using cyano columns. The less hydrophobic and negatively charged naphthalene-2-sulfonate elutes very rapidly because of repul- sion from both the micelle and the negatively charged modified stationary phase.Influence of pH Retention of weak organic acids and bases is affected by the pH of the micellar mobile phase. Solute-micelle partition coefficients of the dissociated and undissociated forms are different. Small changes in pH can significantly alter chromat- ographic retention, particularly when the mobile phase pH is close to the p K , vaIue.23,25,36 Therefore, the pH of the micellar mobile phase must be specified when retention data are reported. With anionic or cationic surfactants in the mobile phase, retention can be modified appropriately by working at a pH value where some compounds are ionized. Cyano-bonded and C18 columns interact very differently with surfactant monomers, resulting in a different elution behaviour of organic acids and bases as a function of the micelle concentration in the mobile phase and pH.35 Ionizable compounds on a cyano packing show a different behaviour depending on pH, that is, the slope of the llk’ versus [MI plot can be positive, negative or zero.Hence, for benzoic acid, at pH <4, k’ values decrease, and at pH >4, k’ values increase with increasing SDS concentration. In the intermediate range of pH values, there is an isoeluting point where k’ is completely independent of SDS concentration. This is the pH value at which two species, acid and base, in equilibrium with each other, have the same k’ value. This is analogous to the isosbestic point in spectroscopy and would be expected to give the acid dissociation constant in the micellar medium.For weak acids, such as Bromocresol Green, using a C18 column and increasing SDS concentration in the mobile phase, k’ values decrease in acidic solution where the neutral form is present and remain constant in more basic solution where the anionic acid form is present, electrostatically repulsed by both the negative micelle and stationary phase.3’ The elution behaviour versus pH of protonated bases on Clx columns will be the opposite of that observed on cyano columns. Adsorption of anionic surfactant monomers on the surface of the CI8 stationary phase causes protonated organic bases, such as aniline, to be retained for a longer period of time than the neutral free-base form because of electrostatic attraction.In contrast, for weak bases using cyano columns, it834 ANALYST, MAY 1992, VOL. 117 is found that the largest k' values occur in more basic solution where the neutral, free-base form is present and are smallest in acidic solution where the protonated positively charged form exists, which has favourable electrostatic attraction to the negatively charged micelles.3~ Dependence of k' on pH at a constant value of [MI is sigmoidal if there is no electrostatic repulsion between any of the two acid-base forms and surfactant molecules.35 For example, for 6-thioguanine, a plot of k' versus pH at various concentrations of SDS on a CI8 column reveals that the largest k' values occur in acidic solution, where the protonated form of the drug is present, and the smallest values occur in more alkaline solutions, where its neutral form is present.The observed increase in retention could be ascribed to the fact that electrostatic attractions of the solute with the surface of the surfactant-modified stationary phase are stronger than those with the micelles.22 Ionic Strength If antibinding is chiefly an electrostatic phenomenon, one would expect to see definite salt effects on the magnitude of K A M . The solute is not only excluded from the micelle but also from the double layer around the micelle. The thickness of the electrical double layer decreases with increasing ionic strength, thus allowing hydrophobic interaction of the solute with the micelle.37 Modification of ionic strength might be sufficient to change completely from an antibinding to a binding type behaviour.In the absence of salt, Bromophenol Blue is an antibinding compound, whereas in the presence of as little as 0.02 mol dm-3 NaCl, it appears to be non-binding. At slightly higher salt concentrations, the compound binds strongly to SDS micelles. The binding to SDS micelles increases substan- tially with the concentration of NaC1.37 For most antibinding solutes the value of K A M becomes less negative with increasing ionic strength, although not all compounds show this type of behaviour. For example, the interaction of Naphthol Green B with SDS micelles appears to be largely unaffected by the addition of salt. Conversely, KAM for thiocyanate ion becomes even more negative with increas- ing NaCl concentration.37 In order for the transition from antibinding to non-binding to binding to occur, the solute ion must have sufficient hydrophobic character to associate with the non-polar portion of the micelle, once electrostatic repulsions have been minimized.The behaviour of thiocyanate is difficult to explain in terms of electrostatic criteria. However, it may be possible to rationalize such behaviour by considering the negligible hydrophobic character of this ion. Selectivity With Purely Micellar Eluents A study of the chromatographic behaviour of a homologous series of compounds provides important information that can be used to distinguish retention and selectivity between conventional RP-LC and MLC.38 With hydro-organic mobile phases the logarithm of k' is linearly related to the number of carbons, nc, in a homologous series in the following form:39 log k' = log a(CH2) nc + log fi where a(CH2) =.k', + l/k',, is the hydrophobic or methylene selectivity, that is, the ratio of the retention factors of two solutes that differ from each other by a methylene group, and log fi reflects the specific interactions between the functional group of the molecule and the mobile and stationary phases. The retention behaviour of a homologous series in MLC is, however, very different and k' is linearly dependent on nc:40 k' = Bnc + A (3) where A and B are the intercept and slope, respectively, of the straight line. A plot of log k' versus nc for these systems has a 1.2 1 1 5 0 Number of carbon 3 atoms 5 Fig. 3 a 0.072 mol dm-3 CTAB micellar mobile phase38 Log k' and k' versus number of carbons for alkylbenzenes with clear curvature (Fig. 3).This is probably due to different solute locations in the micelles for different members of a homologous series, which will experience different polarities. This behaviour has been observed with non-ionic, anionic and cationic surfactants.29 For hydro-organic mobile phases, a(CH2) is independent of the type of homologous series for a given mobile and stationary phase system. In contrast, with micelles the a(CH2) values for alkylphenones are larger than those observed for alkylbenzenes. A methylene group of an alkylphenone undergoes a larger change in its microenvironment polarity as it is transferred from micellar eluents to the stationary phase.38 In conventional RP-LC systems, a(CH2) decreases with an increase in modifier concentration in the aqueous mobile phase. For a purely aqueous mobile phase, a(CH2) -4 and for 100% methanol it is about 1.1-1.2.With micellar eluents, the over-all a(CH2) values are much smaller and the variation with micelle concentration is fairly small. Typical selectivities for alkylphenones range from 1.6 to 1.1 for SDS concentra- tions between 0.06 and 0.5 mol dm-3. The net free energy of transfer of a methylene group from the mobile phase to the stationary phase is the difference in the free energy of transfer from the bulk solvent to the stationary phase and from the bulk solvent to the micelle. Micelles and surfactant-modified phases have a similar molecular organization, which leads to low selectivity.38 Functional group selectivity is defined as the ratio of the value of k' of a compound with a substituent to that of the parent compound [ e .g . , a(R) = k'(Bz-R)/k'(Bz) for substi- tuted benzenes]. For a large group of compounds, particularly non-ionic compounds, hydrophobic interactions play a major role; an alkyl-bonded stationary phase modified with surfac- tant monomers makes the environment of the stationary phase similar to that of the micelles. A decrease in functional group selectivity was observed with increasing micelle concentra- tion .4* In general, the change in selectivity with micellar eluents is evident when log k' for different compounds is plotted against log[surfactant] .31 Frequently, the linear plots are not parallel but intersect each other (Fig.4). Reversal of elution order indicates the occurrence of two competing equilibria: solute- micelle association and solute-stationary phase interaction. The parameters Psw and K A M have opposing effects on retention. As Psw increases, retention increases, whereas as KAM increases, retention decreases. At a low micelle concen- tration, the system resembles conventional RP-LC and Psw controls retention. However, as the concentration of surfac- tant is increased, K A M has an increasing effect owing to the larger number of micelles present in the mobile phase. TheANALYST, MAY 1992, VOL. 117 835 1.2 0.8 L m -I 0.4 0 Log CSDS Fig. 4 Log k’ versus log total SDS concentration in the mobile phase for several diuretics: A, probenecid; B, ethacrynic acid; C.chlorthali- done; D, acetazolamide; and E, hydrochIorothiazide4* difference in KAM values among the solutes is sometimes so large that the elution order is reversed. When comparing the elution of any two solutes, selectivity might increase or decrease with micelle concentration depend- ing on the contribution of electrostatic and hydrophobic interactions, which in turn depend on the structure of the compounds. Selectivity also depends on the type of surfactant. This is true for diverse pairs of zwitterionic amino acids and peptides.41 Addition of Modifiers to Micellar Eluents The predominant factor influencing band broadening in MLC appears to be stationary phase mass transfer. The thickness of the stationary phase layer and its viscosity are significantly increased by surfactant adsorption.13 In conventional RP-LC, for well-designed column packings, this term is considered to be negligible. However, it becomes significant for column packings that have a thick stationary phase or poor stationary- phase diffusion, such as some of the original bonded-phase materials that had thick polymeric layers of stationary phase.43 The surfactant-coated column is analogous to those packing materials. The addition of small percentages of propan-1-01 to micellar mobile phases was recommended by Dorsey et al.44 to enhance chromatographic efficiency and to decrease the asymmetry of the chromatographic peaks. Since then other organic solvents have been studied as modifiers in MLC.45 Of these, short-chain alcohols have usually been demonstrated to be the most suitable.4-8 The term hybrid is used for ternary eluents of water-organic solvent-micelles.The general pore shapes of the parent CI8 material appear to be retained in a surfactant-modified stationary phase, indicating that the surfactant produces a thick film on the interior walls of the capillaries, rather than completely filling the pore. Alcohol modifiers reduce the amount of surfactant sorbed on the stationary phase, and the effect is larger with increasing concentration and hydrophobicity of the modi- fie r .4X I n addition to reducing the carbon loading and film thickness, the addition of an alcohol is also expected to influence the fluidityhigidity of the surfactant-C18-bonded ligand structure on the stationary phase, just as its presence alters the fluidity of the micellar aggregate structure.The solute-stationary phase diffusion coefficient should increase as the microviscosity of the phase decreases.45 At a fixed modifier concentration, the efficiency was observed to decrease as the surfactant concentration in the mobile phase was increased, as in the absence of a modifier. Addition of increasing amounts of alcohol at a fixed surfactant concentration increased the efficiency. Therefore, the addi- tive to surfactant concentration ratio is the dominant factor influencing chromatographic efficiency.45 For example, after addition of 5 or 10% methanol to a 0.02 mol dm-3 SDS mobile phase, the number of theoretical plates, n , increased from 300 to 750 and 1120, respectively, for acetone,44 and after addition of 5% pentan-1-01 to a 0.28 mol dm-3 SDS mobile phase, n increased from 1530 to 3570, and from 50 to 950 for benzene and 2-ethylanthraquinone, respectively,45 even though the relative microviscosity of the micellar mobile phase had increased due to the added alcohol.The low efficiency observed for 2-ethylanthraquinone with SDS and without modifier is due to its low solubility in water. The compound can only partition between the rnicelle in the mobile phase and the surfactant-coated stationary phase. Consequently, in order to desorb/exit the stationary phase or micelle in the mobile phase, for both of which this compound has a great affinity, the ionic micelle should be located close to the surfactant-modified stationary phase. However, both are similarly charged.Hence, an electrostatic repulsive barrier to the direct merger of the micellar entity with the surfactant- coated stationary phase will exist, which will impede solute mass transfer across this interface. The greater the fraction of partitioning that must occur via the direct transfer mode, the poorer will be the observed chromatographic efficiency. It should be noted that another reason why alcohols improve the efficiency in MLC, with ionic surfactant micelles, may be because their presence can reduce the net electrical charge density of the ionic micellar surface.45 This would be expected to diminish the repulsive barrier. The presence of alkane additives does not affect the surface charge density; hence, these types of additive do not improve efficiency for very hydrophobic solutes, even though they reduce the extent of surfactant coverage of the stationary phase.This also explains why alcohol additives do not enhance the efficiency in MLC with non-ionic micellar mobile phases.45 Non-ionic micellar surfactants are long-chain tensioactive alcohols themselves and C1-C5 alcohols are not very effective in desorbing these non-ionic surfactants from the surfactant- modified CI8 stationary phase. Such non-ionic surfactants also have no charge and, therefore, no electrostatic charge barrier is encountered in the direct transfer process envisaged for water-insoluble solutes. In fact, the efficiency achieved in very hydrophobic test solutes with a Brij 35 micellar mobile phase is better than that which can be obtained with any ionic micelles.Solvent Strength One of the main disadvantages of purely micellar eluents is their weak solvent strength. The solvent strength can be increased by addition of an alcohol. This significantly alters the equilibrium of the solute away from the micelle towards the bulk aqueous phase, which becomes more non-polar.49 The addition of alcohols to micellar mobile phases would cause changes in certain micellar properties, such as the aggregation number and the c.m.c of the surfactant. However, the observed changes in retention and selectivity in hybrid systems are too large to be explained in terms of changes in micellar properties. The changes might be explained by modification of the micro-environment of the micelles and the stationary phase. Large concentrations of the organic solvent can totally disrupt the micelle structure; hence, the use of alcohols in MLC has been questioned.Some workers have argued that when organic modifiers are used this type of chromatography loses some of its appea1.46836 ANALYST, MAY 1992, VOL. 117 It may seem logical to assume that the separation mechan- ism with a hybrid mobile phase is similar to that with traditional hydro-organic solvents, rather than to that with a purely aqueous surfactant mobile phase. Addition of an organic solvent might reduce the role of micelles and would create a system that is closer to hydro-organic eluents. However, as long as the integrity of the micelles is maintained, addition of an alcohol to a micellar mobile phase will not create a hydro-organic system, even though in hybrid systems interactions are reduced by the presence of an alcohol and the stationary phase is more similar to that in a conventional hydro-organic system.Interestingly, the non-logarithmic behaviour of k’ versus nc for a homologous series is also observed with hybrid micellar systems. This shows that it is micelles that influence the role of an organic co-solvent in the mobile phase.38 With hybrid eluents, solute binding constants to micelles, KAM, and their partitioning into the stationary phase, Psw, both decrease as a result of the addition of an alcohol and, therefore, the eluting power of the mobile phase increases. Binding constants of hydrophobic solutes decrease more than those of hydrophilic solutes with an increasing alcohol concentration.Hence, selectivity is modified. The decrease in Psw may be associated with an alteration of the stationary phase.49.50 In traditional RP-LC with a binary hydro-organic mobile phase, the effect of the organic modifier concentration, 8, on retention is often expressed as log k’ = -SO + log k’0 (4) where S is the solvent strength parameter.51 [The intercept log kro is the logarithm of the capacity factor in a purely aqueous mobile phase (in hydro-organic systems) or at a given micelle concentration without modifier.] This equation is valid only for a small range of concentrations for hydro-organic eluents. In contrast, excellent linearity was observed between log k’ and 8 even for water-rich eluents.40 The over-all S values for micellar hybrid eluents can be ranked as Spentan-1-01 > Sbutan-1-01 > Spropan-1-01 > Srnethanol, which is similar, for the last three solvents, to conventional hydro-organic systems.41.42 The larger S value for pentan-1-01 and butan-1-01 indicates that these solvents interact to a greater extent with micelles.However, the over-all S values for hybrid systems are still smaller than in the absence of a micelle. The S values for a hydro-organic mobile phase change markedly with solute size in a homologous series. Variation in solvent strength with increasing solute size is minimized in the presence of micclles.“) For a hybrid mobile phase, the solvent strength values are almost constant for a group of homologous compounds. The constancy of solvent strength with the variation in solute size is due to localization of solutes in the micelle environments, which reduces the size factor as far as the solvation of the solute by an alcohol is concerned. Selectivity With Hybrid Micellar Eluents In conventional hydro-organic systems, the same ranking of S values for different solutes with butan-1-01, propan-1-01 and methanol is observed, because the three solvents belong to the same selectivity group.In the presence of micelles, this may not be the situation, because these solvents interact differently with micelles. The selectivity of different organic solvents in the presence of micelles might change.41~42 This extends the possibilities for chromatographic separations. In conventional RP-LC a systematic decrease in selectivity occurs as a result of an increase in the volume fraction of the organic modifier, i e ., solvent strength. In the presence of micelles, selectivity may increase, decrease or remain un- changed with solvent strength. In hybrid systems, solvent strength may be enhanced without sacrificing the selectivity. As occurs with purely micellar systems, both the alkyl-bonded phase and the micelles are solvated by an alcohol to a similar extent, so that the methylene groups still find the same difference in their micro-environment polarities on trans- ferrence to the stationary phase. Addition of up to 20% propan-1-01 to a micellar mobile phase of SDS or CTAB has a negligible effect on the hydrophobic selectivity.41 It is not surprising that peculiar behaviour might be observed as different solutes experience different polarities in their immediate vicinities.For example , the carbonyl-group selectivity, cx(CO), defined as k’[C6H5CO(CH2),CH3]/ k’[C6H5(CH2),CH3], changes with the hydrophobicity (or number of carbon atoms in the side-chain) of two homologous pairs for a micellar eluent.41 In this instance, the selectivity decreases with hydrophobicity, approaching unity, whereas it is independent of solute type for a hydro-organic eluent. This can be attributed to the fact that the more hydrophobic pairs are located in the more hydrophobic environment of the micelles and, therefore, experience a smaller change in their micro-environment polarities on being transferred from the mobile to the stationary phase.The a ( C 0 ) selectivity increases on addition of an alcohol to the micellar mobile phase, but decreases with an increase in micelle concentra- tion. The selectivity for a group of peptides and amino acids increases systematically on addition of an organic modifier. Similar behaviour is observed for some of the substituted benzenes, whereas for others, functional group selectivity decreases with increasing solvent strength. For the latter, the largest positive changes of selectivity are observed for compounds that are more hydrophilic o r retained to a lesser extent than benzene (e.g. , benzyl alcohol, acetanilide, phe- nol), whereas for compounds that are more hydrophobic or retained to a greater extent than benzene, selectivity de- creases with increasing solvent strength (e.g., anthracene, naphthalene, butyrophenone) .41 Although the solvent strength can be increased adequately in certain instances by increasing the micelle concentration, chromatographic efficiency in MLC usually deteriorates at higher micelle concentrations. Addition of an organic solvent to the micellar eluent may give an adequate eluent strength; it also improves the chromatographic efficiency, and in many instances can lead to an enhancement of separation selectivity. However, use of organic modifiers is not always appropriate. In certain instances selectivity enhancement might not lead to an improvement in resolution if retention falls below the optimum k’ range as a result of an increase in eluent strength.52 In other instances, the addition of an organic solvent to micellar eluents may have a beneficial effect on retention, but the efficiency may remain low.Use of Reversed Micelles in Liquid Chromatography In conventional normal-phase high-performance liquid chro- matography a problem exists with reproducibility over time because of the variation in the water content of the mobile phase. The use of reversed micelle mobile phases [ e . g . , sodium bis(2-ethylhexyl)sulfosuccinate, known as Aerosol OT, AOT] offers a unique solution to this problem owing to the ability to solubilize water in the interior of the micelle structure.53 However, the presence of a surfactant leads to a loss of efficiency, probably because of the localization of polar solutes in the hydrophilic core with a slow transfer step out of the micelle.The use of reversed micelles in supercritical fluid chromato- graphy (SFC) provides another way of modifying the mobile phase, and is an alternative to polar and modified fluids.54.55 In SFC, solubilization of large polar molecules is possible and the behaviour of the micellar mobile phase may be changed by control of temperature and pressure. Retention times of polar solutes are substantially reduced with a reversed micellarANALYST, MAY 1992, VOL. 117 837 mobile phase and solutes that are more polar can be separated. Reversed micelle chromatography may be better adapted to supercritical fluids owing to the gain in efficiency at higher temperatures. The higher diffusion rates and lower viscosities of supercritical fluids, compared with those of liquids at the same temperature, may enhance micelle diffusion rates, leading to an increased over-all efficiency.This work was supported by the CICYT Project DEP89-0429. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 References Hinze, W. L., in Ordered Media in Chemical Separations, eds. Hinze, W. L.. and Armstrong, D. W., ACS Symposium Series, American Chemical Society, Washington, DC, 1987. vol. 342, Dorsey. J. G., Adv. Chromatogr., 1987. 27, 167. Berthod. A., and Dorsey, J. G.. Analusis, 1988, 16, 75. Khaledi, M. G., BioChromatography, 1988. 3, 20. Berthod, A., Girard, I . , and Gonnet, C., in Ordered Media in Chemical Separations, eds. Hinze. W. L., and Armstrong, D. W.. ACS Symposium Series, American Chemical Society, Washington, DC.1987, vol. 342, p. 130. Dorsey, J. G., Khaledi, M. G.. Landy, J. S . , and Lin, J.-L., J. Chromatogr., 1984, 316, 183. Berthod, A., Girard, I., and Gonnet, C., Anal. Chem., 1986, 5%. 1356. Berthod. A.. Girard, I., and Gonnet, C., Anal. Chem., 1986, 58, 1359. Mullins. F. G. P., and Kirkbright, G. F., Analyst, 1984, 109, 1217. Kirkbright, G. F., and Mullins, F. G. P., Analyst. 1984, 109, 493. Mullins, F. G. P., and Kirkbright, G. F., Analyst, 1986, 111, 1273. Kirkman, Ch. M., Zu-Ben, Ch.. Uden, P. C., Stratton, W. J . , and Henderson, D. E., J. Chromatogr., 1984, 317, 569. Borgerding, M. F., and Hinze, W. L., Anal. Chem., 1985, 57, 2183. Stratton, L. P., Hynes, J. B., Priest. D. G., Doig, M. T., Barron. D. A., and Asleson, G. L., J. Chromatogr., 1986,357, 183. Breyer, E.D., Strasters, J. K . . and Khaledi, M. G., Anal. Chem.. 1991,63, 828. Medina Hernandez, M. J., Bonet Domingo, E., Ramis Ramos. G., and Garcia Alvarez-Coque. M. C., unpublished work. DeLuccia, F. J., Arunyanart, M., and Cline Love, L. J . , Anal. Chem., 1985, 57, 1564. Arunyanart, M., and Cline Love, L. J., J . Chromatogr., 1985, 342, 293. Cline Love. L. J., Zibas. S . . Noroski, J., and Arunyanart, M., J. Phavm. Biomed. Anal.. 1985, 3, 511. Haginaka, J.. Wakai, J., and Yasuda, H., Anal. Chem., 1987, 59, 2732. Kim. Y .-N., and Brown, P. R., J. Chromatogr., 1987,384,209. MenCndez Fraga, P., Blanco Gonzalez, E., and Sanz-Medel. A., Anal. Chim. Acta, 1988, 212, 181. Palmisano, F.. Guerrieri, A.. Zambonin, P. G., and Cataldi, T. R. I., Anal. Chem.. 1989, 61, 946.Sentell, K. B., Clos. J. P., and Dorsey, J. G., BioChromato- graphy. 1989, 4. 35. p. 2. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 Cline Love, L. J., and Fett, J., J . Pharm. Biomed. Anal., 1991, 9, 323. Carretero, I., Maldonado, M., Laserna, J. J., Bonet, E., and Ramis Ramos, G., Anal. Chim. Acta, in the press. Armstrong, D. W., and Nome, F., Anal. Chem., 1981,53,1662. Arunyanart, M., and Cline Love, L. J., Anal. Chem., 1984,56, 1557. Borgerding, M. F., Quina, F. H., Hinze, W. L., Bowermaster, J., and McNair, H. M.. Anal. Chem., 1988, 60, 2520. Cline Love, L. J., Weinberger, R., and Yarmchuk, P.. in Surfactants in Solution, eds. Mittal, K. L . . and Lindman, B., Plenum Press, New York, 1984, vol. 2, pp. 1139-1158. Yarmchuk, P.. Weinberger, R., Hirsch, R. F., and Cline Love, L. J., Anal. Chem.. 1982, 54, 2233. Landy. J. S., and Dorsey, J. G., Anal. Chim. Acta, 1985, 178, 179. Armstrong, D. W., and Stine. G. Y.. Anal. Chem., 1983. 55, 2317. Pramauro, E., and Pelizzetti, E.. Anal. Chim. Acta. 1983, 154, 153. Arunyanart, M., and Cline Love, L. J., Anal. Chem., 1985,57, 2837. Haginaka, J., Wakai, J., and Yasuda, H., J . Chromatogr., 1989, 488, 341. Armstrong, D. W., and Stine, G. Y., J . Am. Chem. SOC.. 1983, 105, 6220. Khaledi, M. G., Anal. Chem., 1988, 60, 876. Colin, H., Guiochon, G., Yun, Z . , Diez-Masa, J. C., and Jandera, P., J . Chromatogr. Sci., 1983, 21, 179. Khaledi, M. G.. Peuler, E., and Ngeh-Ngwainbi, J., Anal. Chern., 1987, 59, 2738. Khaledi, M. G., Strasters, J. K., Rodgers, A. H., and Breyer, E. D.. Anal. Chem., 1990, 62, 130. Bonet Domingo, E., Medina Hernandez, M. J . , Ramis Ramos, G., and Garcia Alvarez-Coque, M. C., Analyst, 1992,117,843. Snyder, L. R., and Kirkland, J. J., Introduction to Modern Liquid Chromatography. Wiley, New York. 1979, ch. 5. Dorsey, J. G., De Echegaray, M. T., and Landy, J. S., Anal. Chem., 1983, 55. 924. Borgerding, M. F., Williams, R. L., Jr., Hinze, W. L., and Quina, F. H., J. Liq. Chromatogr., 1989, 12, 1367. Yarmchuk, P., Weinberger, R., Hirsch, R. F., and Cline Love, L. J.. J. Chromatogr., 1984, 283,47. Berthod, A., and Roussel, A., J. Chromatogr.. 1988, 449, 349. Borgerding, M. F., Hinze. W. L., Stafford, L. D., Fulp, G. W., Jr., and Hamlin, W. C., Jr., Anal. Chem., 1989, 61, 1353. Tomasella, F. P., Fett, J . , and Cline Love, L. J., Anal. Chem., 1991, 63, 474. Berthod, A.. Girard. I . , and Gonnet, C., Anal. Chem., 1986. 58, 1362. Snyder, L. R.. Dolan, J . W., and Gant, J. R.. J . Chromatogr., 1979, 165, 3. Foley. J. P., Anal. Chim. Acta. 1990, 231, 237. Hernandez Torres, M. A., Landy, J. S . , and Dorsey. J. G., Anal. Chem., 1986, 58, 744. Gale, R. W., Fulton, J. L., and Smith, R. D., Anal. Chem., 1987, 59, 1977. Veuthey. J. L., Caude, M., and Rosset, R., Analusis. 1988,16, 466. Paper 1105258F Received October 16, 1991 Accepted December 13, 1991