Selectivity patterns in micellar electrokinetic chromatography

Journal of Chromatography A, 1216 (2009) 1901–1907

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Selectivity patterns in micellar electrokinetic chromatography Characterization of fluorinated and aliphatic alcohol modifiers by micellar selectivity triangle Cexiong Fu 1 , Morteza G. Khaledi ∗ North Carolina State University, Department of Chemistry, Raleigh, NC 27695-8204, USA

a r t i c l e

i n f o

Article history: Available online 19 January 2009 Keywords: MEKC Electrokinetic chromatography Micellar selectivity triangle Selectivity classification Alcohol-modified micelles Fluorinated alcohols Hexafluoroisopropanol SDS TTAB LiPFOS Solvation parameters

a b s t r a c t The usefulness of the micellar selectivity triangle (MST) for prediction and interpretation of separation patterns in micellar electrokinetic chromatography (MEKC) separations is presented. In addition, we demonstrate the capability of controlling selectivity properties of micelles through addition of organic modifiers with known solvation properties as predicted by MST. The examples are modification of the hydrogen bond donor (HBD) micelle of lithium perfluorooctanesulfonate, the hydrogen bond acceptor (HBA) micelle of tetradecyltrimethylammonium bromide, and the sodium dodecyl sulfate micelles with intermediate hydrogen bonding properties with two hydrophobic organic modifiers. One is an aliphatic alcohol, n-pentanol that can act as both a HBA and a HBD; by contrast, the other organic modifier is a fluorinated alcohol, hexafluoroisopropanol that is a strong HBD modifier and would enhance the hydrogen bond donor strength of micelles. A test sample composed of 20 small organic solutes representing HBA, HBD, and non-hydrogen bond aromatic compounds was carefully selected. The trends in retention behavior of these compounds in different micelles are consistent with the selectivity patterns predicted by the MST scheme. To the best of our knowledge, this is the first report on the unique selectivity of fluorinated alcohols as modifiers in MEKC. These results demonstrate the usefulness of the MST scheme for identifying pseudo-phases with highly similar or different selectivities and can serve as a guide for judicious selection of modifiers to create pseudo-phases with desired selectivity behavior on a rational basis. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Selectivity in micellar electrokinetic chromatography (MEKC) is primarily controlled through a balance of polar and hydrogen bonding interactions between sample components with micelles and bulk solvent. Hydrophobic interaction is the primary driving force for solute retention. The availability of a wide range of pseudo-phases, from micelles made of surfactant aggregation to various polymeric phases, vesicular and liposome assemblies, and microemulsions has provided an unprecedented flexibility in manipulating retention and selectivity patterns [1–12]. The existence of a large number of pseudo-phases, however, has made the selection of optimum pseudo-phase composition for separation ever more challenging. The situation becomes more pronounced considering that pseudo-phase solutions can be further modified

∗ Corresponding author. Tel.: +1 919 515 4563. E-mail address: Morteza [email protected] (M.G. Khaledi). 1 Present address: Center for Advanced Proteomics Research and Department of Biochemistry and Molecular Biology, UMDNJ, New Jersey Medical School Cancer Center, Newark, NJ 07103, USA. 0021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.01.041

through the use of mixed micelles, surfactant–polymers, as well as organically modified micelles using a wide range of co-solvents, cyclodextrins, or other organic components like urea and glucose [13–17]. In a previous paper, a novel micellar selectivity triangle (MST) was introduced for characterization and classification of pseudophases in EKC [18]. The MST scheme allows facile identification of selectivity properties of pseudo-phases according to their hydrogen bonding and dipolar characteristics. The MST should be quite useful for the selection of pseudo-phases with distinctly different selectivity and eliminating systems with similar or redundant properties in method development and optimization process. Here, we report the usefulness of MST for prediction and interpretation of separation patterns in MEKC separations. In addition, we demonstrate the capability of controlling selectivity properties of micelles through addition of organic modifiers with known solvation properties as predicted by MST. 2. Experimental All MEKC experiments were performed on a laboratory-built CE system equipped with a 0–30 kV high-voltage power supply

1902

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

Table 1 LSER coefficients ratios for the micellar pseudo-phase systems in MST (Fig. 1). System

Micellar pseudo-phases

v

b

a

s

e

C

R2

1 2 3 4 5 6 7 8 9

40 mM SDS 40 mM SDS/400 mMPeOH 40 mM SDS/400 mM HFIP 40 mM LiPFOS 40 mM LiPFOS/400 mM PeOH 40 mM LiPFOS/400 mM HFIP 10 mM TTAB 10 mM TTAB + 300 mM PeOH 10 mM TTAB + 50 mM HFIP

2.95 (0.12) 2.85 (0.09) 2.72 (0.11) 2.35 (0.12) 2.55 (0.09) 2.14 (0.13) 3.17 (0.09) 2.90 (0.10) 2.80 (0.11)

−1.76 (0.11) −2.40 (0.08) −1.38 (0.10) −0.54 ((0.10) −1.63 (0.08) −0.48 (0.12) −2.86 (0.08) −2.57 (0.09) −2.28 (0.10)

−0.17 (0.06) −0.03 (0.04) −0.64 (0.05) −0.71 (0.06) −0.29 (0.04) −0.92 (0.06) 0.91 (0.04) 0.42 (0.05) 0.40 (0.05)

−0.54 (0.09) −0.89 (0.07) −0.49 (0.09) 0.31 (0.09) −0.50 (0.06) 0.35 (0.10) −0.49 (0.07) −0.78 (0.08) −0.47 (0.08)

0.41 (0.10) 0.51 (0.07) 0.25 (0.08) −0.46 (0.09) 0.02 (0.07) −0.50 (0.11) 0.66 (0.07) 0.51 (0.08) 0.33 (0.09)

−1.79 (0.12) −1.22 (0.09) −1.55 (0.11) −2.03 (0.12) −1.34 (0.09) −1.79 (0.14) −2.69 (0.09) −1.92 (0.11) −2.15 (0.12)

0.99 0.99 0.98 0.98 0.98 0.98 0.99 0.98 0.97

1 SDS, sodium dodecyl sulfate. 2 PeOH, pentanol. 3 HFIP, hexafluoroisopropanol. 4 LiPFOS, lithium perfluorooctanesulfonate. 5 TTAB, tetradecyl trimethylammonium bromide.

(Spellmann, Plainview, NY, USA) and a SSI 500 variable UV–visible detector (SSI, State College, PA, USA). Data were acquired and processed with PC/Chrom+ version 4.0.9 (H&A Scientific, Greenville, NC, USA). A 50-␮m I.D. × 375-␮m O.D. fused silica capillary (Polymicro Technologies, Phoenix, AZ, USA) was used. The total length of the capillary was 70 cm, with an effective length of 50 cm to detector. Newly installed capillary was conditioned with a rinse procedure using Milli-Q water for 10 min, 0.1 M NaOH (0.1 M LiOH was used when lithium perfluorooctanesulfonate (LiPFOS) was the pseudo-stationary phase) for 20 min to ionize free silanol groups, methanol for 20 min to clean organic impurities (5 min Milli–Q water rinses were carried out between all washing steps). Finally, the capillary was rinsed with the buffer solution for 20 min before equilibration with the micellar pseudo-phase for no less than 3 h to ensure reproducible retention behavior as evident from a constant micelle elution time. The capillary was thermostated at 30.0 ◦ C with a circulating oil bath (Lauda K-2/R, Brinkmann Instruments, Westbury, NY, USA) and two 250-mL jacketed beakers were used to maintain the buffer reservoirs at 30.0 ◦ C. A positive 20 kV was applied throughout the experiment, except in cationic surfactant tetradecyltrimethylammonium bromide (TTAB), in which the polarity was reversed. All solutes for the training set and organic solvents were purchased from Aldrich (Milwaukee, WI, USA) and used as received without any further purification. Sodium dodecyl sulfate (SDS) and TTAB were obtained from Sigma (St. Louis, MO USA). LiPFOS was a gift from 3 M (St. Paul, MN USA). SDS and TTAB solutions were prepared in 20 mM sodium phosphate buffer (pH 7.0) and LiPFOS solution was made in 20 mM lithium phosphate buffer (pH 7.0). All surfactant solutions were filtered with a 0.45-␮m polypropylene filter and sonicated for ∼5 min. The solutes were dissolved in methanol with concentrations around 80 ␮M, which were introduced to the capillary by a 2 s hydrodynamic injection at the anodic end of the capillary. The retention factor was calculated using Eq. (1): k=

tr − teo teo (1 − tr /tmc )

(1)

where teo is the migration time of an unretained solute, tr is the retention time of a solute, and tmc is the migration time of the micelle. Methanol was used as the electroosmotic flow (EOF) marker and decanophenone as the tmc marker. Retention factors of each solute were obtained with triplicate measurements. The average RSD for k was 1.9% with a range of 0.14–6.6%. 2.1. Construction of micellar selectivity triangle (MST) As discussed in detail in a previous paper [18], the MST was constructed using the linear solvation energy relationship (LSER)

Table 2 Preset LSER coefficient ratios ranges for the system parameter normalization in MST.

Ilow Ihigh

b/v

a/v

s/v

e/v

−1.50 0.00

−0.50 1.00

−0.60 0.90

−0.50 1.00

From Ref [18].

coefficients for hydrogen bond acidity (b), hydrogen bond basicity (a), dipolarity (s), and polarizability (e) normalized for the hydrophobic effects (v) for various pseudo-phases (Table 1). The coefficient ratios were converted to the same scale (between 0 and 1) using Eq. (2): Ui =

I − Ilow Ihigh − Ilow

i = s, a, b or e

(2)

where I represents the normalized LSER coefficient ratios (s/v, a/v, b/v, e/v), Ilow and Ihigh define the lowest and highest values for each term. The Ilow and Ihigh values were determined based on the observed ratios for the current database and are listed in Table 2. Finally, the scaled value of each polar interaction is calculated according to Eq. (3), where Xi varies between 0 and 1 and the sum of all Xi for any micelle systems will be unity. Xi =

Ui Ua + Ub + Us

i = s, a, b or e

(3)

The pre-treated system parameters (Xi ) are plotted as a triangle. 3. Results and discussion 3.1. MST characterization of alcohol-modified micelles Organic co-solvents have been initially used to expand the application range of MEKC to hydrophobic compounds that would otherwise elute with micelles. Due to the similarities with RPLC, the traditional organic modifiers in RPLC (methanol, acetonitrile, tetrahydrofuran, etc.) have been investigated in MEKC separations, however, the polar solvents would mainly modify the bulk aqueous media with little interaction and effect on solvation propTable 3 Solvatochromic parameters: for solvents. Solvent

Hydrogen bond donor strength ˛

Hydrogen bond acceptor strength ˇ

Dipolarity *

Water PeOH HFIP

1.17 0.84 1.96

0.47 0.86 0.00

1.09 0.40 0.65

Source: Ref [20].

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

1903

Table 4 Training solutes for LSER modeling and molecular descriptors [17]. Peak No.

Solutes

McGown volume; v

HBA strength b

HBD strength; a

Dipolarity s

Polarizability e

HBDs 1 2 7 10 12 15

Resorcinol Phenol 3-Chlorophenol 4-Ethylphenol 3,5-Dimethyl phenol 4-Iodophenol

0.834 0.775 0.898 1.057 1.057 1.033

0.58 0.30 0.15 0.36 0.36 0.20

1.10 0.60 0.69 0.55 0.57 0.68

1.00 0.89 1.06 0.90 0.84 1.22

0.98 0.81 0.91 0.80 0.82 1.38

HBAs 3 5 6 8 9 14 16

Phenethyl alcohol Nitrobenzene Acetophenone Methylbenzoate Benzyl chloride 3-Chloroacetophenone Methyl-2-methylbenzoate

1.057 0.891 1.014 1.073 0.980 1.137 1.214

0.66 0.28 0.48 0.46 0.33 0.40 0.43

0.30 0 0 0 0 0 0

0.83 1.11 1.01 0.85 0.82 1.07 0.87

0.78 0.87 0.82 0.73 0.82 0.92 0.77

NHBs 4 11 13 17 18 19 20

Benzene Toluene Chlorobenzene Ethylbenzene 4-Chlorotoluene Naphthalene Propylbenzene

0.716 0.857 0.839 0.998 0.980 1.085 1.139

0.14 0.14 0.07 0.15 0.07 0.20 0.15

0 0 0 0 0 0 0

0.52 0.52 0.65 0.51 0.67 0.92 0.50

0.61 0.60 0.72 0.61 0.71 1.34 0.60

erties of the micelles. Note that the concentration range of organic co-solvents is limited in MEKC using conventional micellar pseudophases to maintain the integrity of the surfactant aggregates structures, which would also have a limited impact on selectivity patterns. In this study, we employed solvents with hydrophobic tails that would co-micellize within surfactant aggregation and influence the solvation properties and subsequently selectivity of micelles. The selected micellar systems were cationic TTAB that is a strong hydrogen bond acceptor (HBA) micelle, anionic fluorinated micelle of LiPFOS that is a strong hydrogen bond donor (HBD), and SDS micelle that has intermediate hydrogen bonding characteristics. These micelles were modified with two hydrophobic alcohols with distinctly different hydrogen bonding properties; pentanol (PeOH) and hexafluoroisopropanol (HFIP). The hydrophobic modifiers have significantly larger partition coefficients into micelles than their more polar analogs, thus modifying the micellar environments to a much greater extent than the bulk aqueous media. For example, pentanol has a partition coefficient of 85 per SDS molecules in SDS micelles; while methanol has a nearly negligible interaction with the SDS micelles [19]. The solvation properties of the two modifiers as measured by the parameters ˛ (a measure of hydrogen bond donor strength), ˇ (a measure of hydrogen bond acceptor strength), and * (measure of solvent dipolarity) are listed in Table 3 [20]. PeOH usually acts as a hydrogen bond acceptor in aqueous media but can behave as a HB donor (as in TTAB micelles, see below) and HFIP is a strong HB donor. Thus, as the water molecules in the interface region of micelles are replaced by pentanol or HFIP, the micelles acquire more hydrogen bond basic or HB acidic characteristics, respectively. Fig. 1 illustrates the MST characterization of the interactive properties of the three individual micelles and their alcohol-modified aggregates. As expected, the addition of HFIP to micelle systems SDS and LiPFOS enhances their hydrogen bond acidity, and reduces their hydrogen bond basicity while their dipolarity is slightly increased. When pentanol is used as the solvent additive, an increase in hydrogen bond basicity and decrease in both hydrogen bond acidity and polarity are observed for these two micelle systems. The order of the hydrogen bond acidity strength of these selected micelle systems are as follows: LiPFOS/HFIP(6) > LiPFOS(4) ∼ SDS/HFIP(3) > LiPFOS/PeOH(5) ∼ SDS(1) > SDS/PeOH(2) > TTAB/HFIP(9) ∼ TTAB/PeOH(8) > TTAB(7),

where the numbers in parentheses correspond to pseudo-phases identified in Fig. 1. Addition of HFIP (a HBD solvent) to SDS elevates the HBD strength of the micelle to the level of LiPFOS. Likewise, addition of PeOH (HBA solvent) to LiPFOS reduces the HBD strength of the micelle to the level of SDS, which results in similar retention behavior in these two systems (see below). The rank ordering of the pseudo-phases in terms of HBA is basically the opposite; that is, TTAB(7) ∼ TTAB/PeOH(8) > SDS/PeOH(2) ∼ TTAB/HFIP(9) > SDS(1) ∼ LiPFOS/PeOH(5) > SDS/HFIP(3) > LiPFOS (4) > LiPFOS/HFIP(6). Note that addition of PeOH to TTAB had basically little effect on the HBA strength, while it led to a slight increase in HBD strength and less dipolarity. This could be attributed to the dual characteristic of PeOH as HBD and HBA. Apparently, in a HBA micelle such as TTAB, PeOH acts as a HBD modifier. Another interesting observation is similar HBA strengths of SDS/PeOH and TTAB/HFIP, where in the former PeOH enhances HBA capability of SDS and for the latter HFIP reduces the HBA strength of TTAB. These results indicate the usefulness of the MST scheme for identifying

Fig. 1. MST for three micelles and six modified micellar systems.

1904

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

Fig. 2. Chromatograms of the separation of 20 aromatic solutes composed of hydrogen bond donors (HBDs): 1 = resorcinol; 2 = phenol; 7 = 3-chlorophenol; 10 = 4-ethylphenol; 12 = 3,5-dimethylphenol; 15 = 4-iodophenol; hydrogen bond acceptors (HBAs): 3 = phenethyl alcohol; 5 = nitrobenzene; 6 = acetophenone; 8 = methylbenzoate; 9 = benzyl chloride; 14 = 3-chloroacetophenone; 16 = methyl-2-methyl benzoate; non-hydrogen bond donors (NHBs): 4 = benzene; 11 = toluene; 13 = chlorobenzene; 17 = ethylbenzene; 18 = 4-chlorotoluene; 19 = naphthalene; 20 = propylbenzene; micellar phases: (A) 40 mM SDS, (B) 40 mM SDS/400 mM PeOH (4.34% v/v), (C) 40 mM SDS/400 mM HFIP (4.16% v/v), (D) 40 mM LiPFOS, (E) 40 mM LiPFOS/400 mM PeOH (4.34% v/v), (F) 40 mM LiPFOS/400 mM HFIP (4.16% v/v); a 20 mM sodium phosphate buffer pH 7.0 was used in all SDS micelle systems, running buffer for LiPFOS micelles was 20 mM lithium phosphate buffer at pH 7.0; 20 kV applied on a 70 cm length capillary (effective length 50 cm).

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

pseudo-phases with highly similar or different selectivities and can serve as a guide for judicious selection of modifiers to create pseudo-phases with desired selectivity behavior on a rational basis. To further verify and elucidate the impact of solvent modification of micelles, retention and selectivity patterns were investigated for a set of 20 carefully selected aromatic compounds. The test solutes were selected based on their hydrogen bonding properties and are grouped accordingly into three categories of HBD, HBA, and non-hydrogen bonding (NHB) solutes (Table 4). The two organic modifiers, PeOH and HFIP, mainly impact the hydrogen bonding capability of micelles, thus it is interesting to monitor the migration behaviors of HBA and HBD solutes in the selected micellar phases. Naturally, all HBD solutes have measurable HBA capability as well. These solutes also have different sizes and polarities that would impact retention and to some extent selectivity. Fig. 2 (A–C) illustrates the elution patterns in MEKC separation with SDS micelles (Fig. 2A) and SDS modified with PeOH (Fig. 2B) and HFIP (Fig. 2C). Fig. 2 (D–F) show the elution patterns in LiPFOS micelles (Fig. 2D), LiPFOS modified with PeOH (Fig. 2E) and LiPFOS modified with HFIP (Fig. 2F), respectively. The peaks are numbered as 1 through 20 based on their increasing retention in the SDS micellar system (Fig. 2A). In general, HBA solutes are retained longer as the HBD strength of micelle is increased (with HFIP modification) and reduced as HBA strength of micelles is increased (with addition of PeOH). The opposite trend is observed for HBD solutes. Vice versa, for HBD solutes, increasing HBA strength of the micelle would result in shifting retention to larger values. As a result, significant differences of elution patterns for the test solutes among the modified and unmodified micelles were observed. For example, upon addition of PeOH to the SDS micelles peaks 6 (HBA Acetophenone) and 5 (HBA nitrobenzene) shift to lower retention and elute before peak 4 (NHB benzene) or peak 14 (HBA chloroacetophenone) shifts so far lower that it elutes before peak 10. Note again that the larger peak number associates with longer retention in SDS micelles. On the other hand, in HBD-modified micelles of SDS + HFIP (Fig. 2C), peak 6 (HBA) shift to higher retention and is eluted after 10 and with peak 12, while peak 8 (HBA methylbenzoate) comes after peak 13, and peak 16 (HBA) elutes after peak 18. In the two strongest HBD micelles of LiPFOS and LiPFOS + HFIP, peak 16 elutes as the last compound. One important factor for these two micelles is the polarizability interaction that is not considered in the MST used in this work. As mentioned in the earlier paper, one could also build MST with other combinations such as Xb , Xa , and Xe (polarizability instead of dipolarity, Xs ). The fluorinated pseudo-phases have poor polarizability effects, especially for polarizable solutes such as iodophenol (peak 15) or naphthalene (peak 19), which are poorly retained in fluorinated pseudo-phases (Fig. 2D–F). A quantitative measure of the solvent modification effects on the solute retention, log kX−Y , was calculated as the difference of retention factors between the modified micelle system and the unmodified micelle:  log kX−Y = log kX − log kY

(4)

X represents PeOH or HFIP-modified micelles and Y represents the unmodified micelles (SDS or LiPFOS). Fig. 3A illustrates log kX–Y for the 20 test solutes in the pentanol and HFIP-modified SDS. When SDS is mixed with the HB basic solvent PeOH, the HBD solutes 2, 7, 10, 12 and 15 are retained longer in the HB basic modified micelle, while the HBA solutes 5, 6, 8, 14 and 16 had weaker interactions with the SDS/PeOH micelles. However, the trend was reversed as a HB acidic solvent (HFIP) was employed as the micelle modifier, suggesting that the HBD solutes had less retention, while the HBA solutes resided longer in the HB acidic SDS/HFIP micelles (Fig. 3A).

1905

Fig. 3. (A) Effect of the solvent modification on SDS micelles. (Solutes as in Fig. 2). (B) Effect of the solvent modification effects on LiPFOS micelles.

The addition of the HB acidic solvent (HFIP) strengthened the HB acidity of the LiPFOS micelles. As a result, the HBD solutes had weaker affinity for the HFIP-modified LiPFOS micelles and the HBA solutes tended to retain longer in the micelles (Fig. 3B). In contrast, the addition of HB basic solvent (PeOH) weakened the affinity of the HBA solutes with the LiPFOS/PeOH micelles and caused stronger binding with the HBD solutes as compared to the unmodified LiPFOS micelles. Another interesting finding according to the MST is that the interactive properties of LiPFOS could be changed upon modification with pentanol to the extent that it would resemble those of SDS micelles as evident from the location of the two pseudophases (LiPFOS/PeOH and SDS) in a close vicinity of one another in the MST (systems 1 and 5 in Fig. 1). This could be further verified through a correlation analysis of the retention factors in these three systems: SDS, LiPFOS and LiPFOS/PeOH. For example, Fig. 4A shows a relative poor correlation (R2 = 0.62) of retention factors of those 35 LSER training solutes between the SDS and LiPFOS micelles. The HBA solutes deviate from the correlation and are located above the line due to their stronger selective interactions with the HB acidic LiPFOS micelles, while the HBD solutes are located below the line due to their stronger interaction with SDS micelles. Upon modification of LiPFOS with HBA solvent PeOH, the solvation properties of the two systems converge (see points 1 and 5 in MST, Fig. 1), which results in significant improvement in the linear correlation shown in Fig. 4B (R2 = 0.95). This could be explained by the fact that pentanol is capable of enhancing

1906

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

Fig. 4. (A) Correlation of the retention factors of test solutes between 40 mM SDS and 40 mM LiPFOS. (Filled circle represents NHB solutes, open square represents HBD solutes, and open triangle represents HBA solutes). (B) Correlation of the retention factors of test solutes between 40 mM SDS and pentanol-modified LiPFOS. (C) Correlation of the retention factors of test solutes between 40 mM LiPFOS and 40 mM LiPFOS + 400 mM PeOH. (D) Correlation of the retention factors of test solutes between 40 mM SDS and 10 mM TTAB. (E) Correlation of the retention factors of test solutes between 40 mM SDS + 400 mM PeOH and 10 mM TTAB + 50 mM HFIP.

the hydrogen bond basicity of the LiPFOS micelles. Another way of examining the impact of LiPFOS modification with PeOH is shown in Fig. 4C where a poor correlation is observed between retention in LiPFOS modified with PeOH and LiPFOS micelles alone. Again note the strong deviation of hydrogen bonding compounds, especially HBA solutes that prefer the more HB acidic environment of LiPFOS. One factor that should be considered in the different retention behavior in fluorinated micelles of LiPFOS and aliphatic micelles of SDS is the differences in cohesiveness between these micelles. This is the main reason for deviation of non-hydrogen bonding solutes from the lines in Fig. 4A and

C. The improved correlation in Fig. 4B indicates that addition of the aliphatic alcohol PeOH to fluorinated micelles of LiPFOS created closer microenvironments and selectivity to that of the SDS micelles. One last example is the significant differences between SDS (MST system 1 in Fig. 1) and TTAB (MST system 7 in Fig. 1). However, the selectivities of these two micellar phases could be modulated closer to one another by modification of the HBA micelles of TTAB with HBD modifier HFIP and modification of the relatively stronger HBD SDS with the HBA modifier PeOH. This is clearly evident in the MST (Fig. 1) where the distance of the latter two systems (1

C. Fu, M.G. Khaledi / J. Chromatogr. A 1216 (2009) 1901–1907

vs. 7) is reduced (compare systems 2 and 9). The TTAB and SDS have significant differences in hydrogen bonding properties, while the TTAB modified with HFIP and SDS modified with PeOH have nearly identical HBA strengths and much smaller difference in HBD strength. Consequently, the correlation between retention factors in the latter systems (TTAB + HFIP vs. SDS + PeOH, R2 = 0.86; Fig. 4E) is significantly better than that in the former (TTAB vs. SDS; R2 = 0.55, Fig. 4D). 4. Conclusions Hydrophobic organic modifiers could be quite useful in controlling retention and selectivity in MEKC as they modify the micellar pseudo-phases with little or no effect on the bulk solvent composition. The availability of quantitative solvation parameters would allow selection of the modifiers for judicious selection of modifiers and a priori control of selectivity behavior. The fluorinated alcohols, like HFIP, provide unique selectivity patterns due to their strong hydrogen bond donor property that is even stronger than water. To the best of our knowledge, this is the first report on selectivity effects of fluorinated alcohols in MEKC. The general retention and selectivity patterns confirm the reliability of MST in characterization and classification of selectivity of pseudo-phases in MEKC. The MST scheme is quite useful for identifying pseudo-phases with highly similar or different selectivities in MEKC and can serve as a guide for judicious selection of modifiers to create pseudo-phases with desired selectivity behavior on a rational basis. It could greatly facilitate the optimization

1907

of pseudo-phase composition and selection in MEKC separations. The combination of ease and speed of manipulating composition of the running micellar solutions in MEKC with the feasibility of selecting appropriate pseudo-phase selectivity would be advantageous in minimizing method development times and achieving better separations. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

M.D. Trone, M.G. Khaledi, Anal. Chem. 71 (1999) 1270. C.F. Poole, S.K. Poole, M.H. Abraham, J. Chromatogr. A 798 (1998) 207. G. Van Biesen, C.S. Buttaro, J. Chromatogr. A 1157 (2007) 437. C. Akbay, N.L. Gill, A. Powe, I.M. Warner, Electrophoresis 26 (2005) 415. C.P. Palmer, Electrophoresis 23 (2002) 3993. S.A. Shamsi, C.P. Palmer, I.M. Warner, Anal. Chem. 73 (2001) 140A. Y. Moliner-Martinez, M. Barrios, S. Cardenass, M. Valcarcel, J. Chromatogr. A 1194 (2008) 128. A.A. Agbodjan, H.H. Bui, M.G. Khaledi, Langmuir 17 (2001) 2893. R.J. Pascoe, J.P. Foley, Electrophoresis 24 (2003) 4227. S.T. Burns, M.G. Khaledi, J. Pharm. Sci. 91 (2002) 1601. L.A. Holland, A.M. Leigh, Electrophoresis 24 (2003) 2935. S.K. Wiedmer, M.S. Jussila, M.L. Riekkola, Trends Anal. Chem. 23 (2004) 562. U. Pyell, U. Buetehorn, J. Chromatogr. A 716 (1995) 81. S. Yang, J.G. Bumgarner, M.G. Khaledi, J. High Resolut. Chromatogr. 18 (1995) 443. Z. Liu, H. Zou, M. Ye, J. Ni, Y. Zhang, J. Chromatogr. A 863 (1999) 69. S.K. Poole, C.F. Poole, Analyst (Cambridge, UK) 122 (1997) 267. M.D. Trone, M.S. Leonard, M.G. Khaledi, Anal. Chem. 72 (2000) 1228. C. Fu, M.G. Khaledi, J. Chromatogr. A, in press, http://dx.doi.org/ 10.1016/j.chroma.2008.12.085. A. Berthod, J. Chromatogr. A 780 (1997) 191. Y. Marcus, Chem. Soc. Rev. 22 (1993) 409.