Solubilization of Amphiphilic Carboxylic Acids in Nonionic Micelles

Sep 12, 2007 - Laurence Dupont-Leclercq, Sébastien Giroux, Bernard Henry, and .... of commercial micelles and microemulsions by Taylor dispersion ana...
0 downloads 0 Views 165KB Size
Download PDF
Langmuir 2007, 23, 10463-10470

10463

Solubilization of Amphiphilic Carboxylic Acids in Nonionic Micelles: Determination of Partition Coefficients from pKa Measurements and NMR Experiments Laurence Dupont-Leclercq, Se´bastien Giroux, Bernard Henry, and Patrice Rubini* Groupe C2M2, Laboratoire SRSMC, UMR CNRS n°7565, UniVersite´ Henri Poincare´ -Nancy-UniVersite´ , BP 239, F-54506 VandœuVre-le` s-Nancy Cedex, France ReceiVed June 13, 2007. In Final Form: July 27, 2007 The solubilization of octylamidotartaric acid (C8T) and octanoic acid (C8C) in Triton X-100 and Brij 58 nonionic micelles has been studied by pHmetric and 1H NMR self-diffusion experiments. As both C8C and C8T exhibit acidbase properties, a distinction between the partition of the neutral acidic form, in terms of the partition coefficient KPH, and the partition of the charged basic form, in terms of the partition coefficient KP-, has been made. The acidity constants, Ka, of C8T and C8C in the presence of micelles have been evaluated from pHmetric experiments. For both solutes, an increase in the pKa is observed in micellar media due to the difference in the partition of acidic and basic forms of the solutes. A model has been developed to determine KPH and KP- from the pKa shifts observed. The values obtained by this pKa shift modeling method and those from self-diffusion coefficient measurements are in good agreement. The acidic form of C8C is incorporated to a larger extent into the Brij 58 micelles than the acidic form of C8T, whereas the opposite trend is observed for the basic forms. Both the acidic and basic forms of C8T are more easily incorporated into Brij 58 micelles than into Triton X-100 micelles. The influence of the structure of the polar head on the solubilization properties is demonstrated. Moreover, evidence for the localization of the solutes in the micelles is obtained from the comparison of the partition coefficients and from 1H NMR data.

1. Introduction The self-association of surfactants to form aggregates such as micelles in aqueous solutions has an important effect on the solubility and the acid-base properties of organic substances that are otherwise soluble in water. The study of these modifications is of great importance for extraction and separation processes, since the use of micellar solutions has been developed as an ecologically safe alternative to liquid-liquid extraction.1 The principle is that the micellar pseudophase can mimic in some way the role played by the organic phase in a classical solvent extraction process.2 Therefore, the knowledge and the possibility to control (by pH-variations, for example) the most likely sites of solubilization of the reacting species is particularly important to improve the yields of extraction.3 Moreover, solubilization in micelles is a crucial issue in many industrial processes such as detergency or emulsion polymerization or for the understanding of mechanisms in biological membranes and cells.4,5 In the present work, the solubilization of amphiphilic carboxylic acids, octanoic acid (C8C) and octylamidotartaric acid (C8T), in nonionic micelles is investigated. C8T is an original sugar-based acid, the synthesis of which has been previously reported.6 The hydrophilic polar head of C8T, derivated from tartaric acid, presents interesting complexing properties for trivalent lanthanide ions (La(III), Pr(III), and so forth), which are relevant in nuclear (1) Tondre, C. In Surfactant-Based Separation, Science and Technology; Scamehorn, J. F., Harwell, J. H., Eds.; ACS Symposium Series 740; American Chemical Society: Washington, DC, 2000; Chapter 10, p 139. (2) He´brant, M.; Tondre, C. J. Colloid Interface Sci. 1992, 154, 378. (3) Inaba, K. Langmuir 1997, 13, 1501. (4) Shah, S. S.; Naeem, K.; Shah, S. W. H.; Laghari, G. M. Colloids Surf., A 2000, 168, 77. (5) Whiddon, C. R.; Bunton, C. A.; So¨derman, O. J. Phys. Chem. B 2003, 107, 1001. (6) Giroux, S.; Rubini, P.; Ge´rardin, C.; Selve, C.; Henry, B. New J. Chem. 2000, 24, 173.

fuel reprocessing.6,7 C8T thus appears as an interesting molecule for micellar extraction due to both its complexing abilities and its amphiphilic properties. The critical micellar concentration (CMC) of C8T is equal to 5.4 × 10-2 mol‚L-1.6 Considering this value, the solubilization of C8T in cosurfactant micelles is proposed with the aim to use C8T in lower concentrations than the CMC for micellar extraction processes. Nonionic surfactants are chosen to incorporate C8T because of their interesting properties allowing cloud point extraction (CPE).8 In this study, the ability of two polyethylene glycol nonionic surfactants, Brij 58 and Triton X-100, to solubilize C8T has been studied and compared. Moreover, the influence of the polar head structure of the solute on solubilization is discussed by comparing the incorporation of C8C and C8T into the micelles. In the literature, many techniques are described for solubilization studies and partition coefficient measurements, such as dialysis,2,9 conductometric measurements,10 UV-vis and NMR spectroscopy,4,11,12 or solubility methods.13 Most of these techniques were unsuitable in our case due to a problem of sensitivity and due to the spectroscopic properties of the molecules studied. Therefore, we decided to take advantage of the acidbase properties of the solutes (presence of a carboxylic group) to determine the partition coefficients by the quantitative evaluation of pKa. Indeed, the effect of the micellar systems on the acid-base properties of a large variety of molecules is well(7) Giroux, S.; Aury, S.; Henry, B.; Rubini, P. Eur. J. Inorg. Chem. 2002, 1162. (8) Hinze, W. L.; Pramauro, E. Crit. ReV. Anal. Chem. 1993, 24, 133. (9) Christian, S. D.; Smith, G. A.; Tucker, E. E.; Scamehorn, J. F. Langmuir 1985, 1, 564. (10) Manabe, M.; Tokunaga, A.; Kawamura, H.; Katsuura, H.; Shiomi, M.; Hiramatsu, K. Colloid Polym. Sci. 2002, 280, 929. (11) Belmajdoub, A.; Boudot, D.; Tondre, C.; Canet, D. Chem. Phys. Lett. 1988, 150, 194. (12) Alonso, B.; Harris, R. K.; Kenwright, A. M. J. Colloid Interface Sci. 2002, 251, 366. (13) Hanna, K.; Denoyel, R.; Beurroies, I.; Dube`s, J. P. Colloids Surf., A 2005, 254, 231.

10.1021/la7017488 CCC: $37.00 © 2007 American Chemical Society Published on Web 09/12/2007

10464 Langmuir, Vol. 23, No. 21, 2007

Dupont-Leclercq et al. mainly due to the liquid junction potential and to the possible alkaline and acidic errors of the glass electrode.26 Kw ) 10-13.75 is the autoprotolysis constant of water for T ) 298.1 K and for 0.10 mol‚L-1 ionic strength used in this work.27

Figure 1. Structure and atom numbering of C8T.

known.14-19 A pKa shift is usually observed in the presence of micelles and can be explained in terms of differences between the properties of the solvent in the bulk and in the micellar pseudophase (modification of the dielectric constant and electrostatic effects).20 Pramauro and Pelizzetti first introduced a relation between these pKa shifts and the partition coefficients.21,22 In the present work, this relation is modified by taking into account the micellar pseudophase and bulk volumes. We carry out a model of the pKa shifts observed to calculate the partition coefficients. These results are corroborated by a complementary analysis based on 1H NMR self-diffusion measurements. 1H pulsed field gradient experiments provide self-diffusion coefficients. The latter are used to determine molar fractions of molecules solubilized in micelles and consequently to calculate partition coefficients.23-25 2. Experimental Section 2.1. Materials. Octylamidotartaric acid (2,3-dihydroxy-4-(octylamino)-4-oxobutanoic acid) was prepared from tartaric acid and octylamine according to a procedure described earlier.6 Its purity was checked by NMR spectroscopy. 1H NMR data of octylamidotartaric acid (in water, external reference and lock: dioxane in D2O, pH ) 1.0, 400 MHz, chemical shifts in ppm with respect to TMS; δdioxane ) 3.70 ppm): N-H (broad t, 8.06), C2-H (d, 4.51), C3-H (d, 4.60), CR-H (q, 2H, 3.19), Cβ-H (m, 2H, 1.47), Cγ-H to Cη-H (m, 10H, 1.22), Cθ-H (t, 3H, 0.79). The atom numbering is indicated in Figure 1. Octanoic acid (Acros Organics, 99%), polyethylene glycol(20)hexadecyl ether (Brij 58, Fluka), polyethylene glycol(9-10)-p(1,1,3,3-tetramethylbutyl)-phenyl ether (Triton X-100, VWR), and potassium chloride (BDH, 99.5%) were used as received. pHmetric titrations were performed with NaOH (Sigma-Aldrich) 0.100 mol‚L-1 standard solutions. 2.2. Methods. 2.2.1. pHmetric Measurements. The acid-base equilibria were investigated by pHmetric titrations at 298.1 ( 0.1 K under argon atmosphere at constant ionic strength I ) 0.10 mol‚L-1, adjusted with KCl. The pH was measured by using an automatic titration set, including a METROHM 721 NET titrino autoburet and a Thermo Orion combined glass electrode (model Ross 8103SC). For quantitative evaluation of the data, eq 1 linking the experimental electromotive force values (E) and the equilibrium hydrogen ion concentration [H+] was used. jH and jOH are fitting parameters allowing the correction of experimental errors in acidic and alkaline media, (14) Mackay, R. A.; Jacobson, K.; Tourian, J. J. Colloid Interface Sci. 1980, 76, 515. (15) (a) Drummond, C. J.; Grieser, F.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1989, 85, 521-535. (b) Drummond, C. J.; Grieser, F.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1989, 85, 537-550. (c) Drummond, C. J.; Grieser, F.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1989, 85, 551-560; (d) Drummond, C. J.; Grieser, F.; Healy, T. W. J. Chem. Soc., Faraday Trans. 1 1989, 85, 561-578. (16) Spreti, N.; Brinchi, L.; Di Profio, P.; Germani, R.; Savelli, G.; Bunton, C. A. Eur. J. Org. Chem. 2004, 1105. (17) Yunes, S. F.; Foroudian, H. J.; Gillitt, N. D.; Bunton, C. A. Colloids Surf., A 2005, 262, 260. (18) Goldsipe, A.; Blanckschtein, D. Langmuir 2006, 22, 9894. (19) El Seoud, O. A. AdV. Colloid Interface Sci. 1989, 30, 1. (20) Jaiswal, P. V.; Ijeri, V. S.; Srivastava, A. K. Colloids Surf., A 2005, 46, 45. (21) Pelizzetti, E.; Pramauro, E. Anal. Chim. Acta 1980, 117, 403. (22) Pramauro, E.; Pelizzetti, E. Anal. Chim. Acta 1981, 126, 253. (23) Furo´, I. J. Mol. Liquids 2005, 117, 117. (24) So¨derman, O.; Stilbs, P.; Price, W. S. Concepts Magn. Reson., Part A 2004, 23A, 121. (25) Stilbs, P. In Solubilization in Surfactant Aggregates; Christian, S. D., Scamehorn, J. F., Eds.; Surfactant Science Series; Marcel Dekker: New York, 1995; Vol. 55, Chapter 11.

E ) E0 +

RT ln[H+] + jH[H+] + jOH[H+]-1Kw F

(1)

The pKa constants were evaluated from the average of three independent titrations with the PSEQUAD computer program.28 The possibility that surfactants, which are surface-active, can interfere with the pH electrode by forming a film on its surface and can perturb the liquid junction potential was taken into account. So, control experiments were performed to check the accuracy of the pHmetric measurements. On one hand, titrations of phosphoric acid solutions in the presence of different concentrations of surfactant (surfactant to phosphoric acid ratio from 0 to 30) were carried out. It was found that the presence of surfactants has no effect on the pKa values. On the other hand, regular titrations of acetic acid solutions were performed after a run of measurements in surfactant media to test the electrode response, and no significant pKa variations were observed. Anyway, the pH electrode was systematically washed in hydrochloric acid or in Fisher Bioblock Scientific cleaning solutions. 2.2.2. Proton NMR Self-Diffusion. Proton NMR self-diffusion experiments were carried out using a Bruker DRX400 apparatus and a Bruker 600 MHz apparatus equipped with a cryoprobe. The solutions were prepared in a mixture of D2O (5%) and H2O. All results were obtained from solutions in 5 mm tubes at 298 K. A stimulated echo 2D sequence allowing water suppression and using a pulse sequence with gradients was used (stebpgp1s19 pulse sequence). In the pulsed-gradient spin-echo experiment on molecules carrying nuclei with a gyromagnetic ratio γ and undergoing Gaussian diffusion characterized by the self-diffusion coefficient D at a constant radio frequency pulse interval (τ), the echo amplitude (I2τ) is described by the Stejskal-Tanner equation: I2τ ) e-γ G δ

(

2 2 2

∆-

)

δ D 3

(2)

where δ denotes the durations of a pair of magnetic field gradient pulses separated by a time interval ∆ and G is their amplitude.23-25 For all the experiments performed, δ varies between 2 and 8 ms and ∆ varies between 200 and 400 ms. Considering the solute-tosurfactant ratio (equal to 1/7 in the case of C8T-Brij 58 and to 1/13 in the case of C8C-Brij 58) and the solute concentrations, 160 scans per spectrum were needed to increase the signal-to-noise ratio. Thirtytwo points were recorded per experiment. This allows us to apply a varying gradient intensity between 30 and 16 G‚cm-1. 2.2.3. Dynamic Light Scattering Experiments. The average hydrodynamic diameters of the micelles were measured by dynamic light scattering with a Malvern 3000HSA Zetasizer instrument at 298 K. All the solutions were filtered previously on Millipore nitrocellulose membranes (cutoff: 0.45 µm). The pH was adjusted by addition of hydrochloric acid or sodium hydroxide. The ionic strength was fixed to 0.10 mol‚L-1 by addition of potassium chloride. 2.2.4. CMC Measurements. The critical micellar concentrations (CMC) of surfactants and of surfactant-solute mixed systems have been measured on a Fluorolog Jobin-Yvon spectrofluorimeter with a pyrene probe.29 The spectra were recorded at 298.1 K and at an ionic strength maintained at 0.10 mol‚L-1 with potassium chloride. The pyrene concentration was about 10-6 mol‚L-1. 2.2.5. Spectroscopic Measurements. IR spectra were carried out on a Perkin-Elmer SpectrumOne spectrometer equipped with an (26) Rosotti, F. J. C.; Rosotti, H. The Determination of Stability Constants; McGraw-Hill: New York, 1961; p 169. (27) Ho¨gfeldt, E. Stability Constants of Metal-Ion Complexes, Part A: Inorganic Ligands; Pergamon Press: New York, 1982; p 32. (28) Ze´ka´ny, L.; Nagypa´l, L. In Computational Methods for the Determination of Stability Constants; Legget, D., Ed.; Plenum Press: New York, 1985; Chapter 8. (29) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977, 99, 2039.

Solubilization of C8T and C8C in Nonionic Micelles

Langmuir, Vol. 23, No. 21, 2007 10465

KPH )

[AH]m [AH]b

(5)

and

KP- )

Figure 2. Acid-base and partition equilibria between the bulk phase and the micellar pseudophase. attenuated total reflection (ATR) accessory. Each spectrum of the solute was obtained by the subtraction of the spectra of two solutions at the same pH: one containing the solute-surfactant mixture and the other containing only the surfactant. NMR spectra were recorded on a Bruker DRX400 apparatus. The solutions were prepared in water. D2O and dioxane were used as external lock and reference, respectively. The concentrations of solute and Brij 58 were 0.10 and 0.18 mol‚L-1, respectively, in the case of IR measurements and 0.010 and 0.060 mol‚L-1, respectively, for NMR measurements. The pH was adjusted with NaOH and HCl (Sigma-Aldrich).

3. Results and Discussion

solutebulk + surfactantmicelles ) solute-surfactantmicelles

(3)

(4)

In the present case, the partition related to the acidic and basic forms must be considered separately, leading to the two following partition constants: (30) Marangoni, D. G.; Kwak, J. C. T. In Solubilization in Surfactant Aggregates; Christian, S. D., Scamehorn, J. F., Eds.; Surfactant Science Series; Marcel Dekker: New York, 1995; Vol. 55, Chapter 14. (31) Dunaway, C. S.; Christian, S. D.; Scamehorn, J. F. In Solubilization in Surfactant Aggregates; Christian, S. D., Scamehorn, J. F., Eds.; Surfactant Science Series; Marcel Dekker: New York, 1995; Vol. 55, Chapter 1.

(6)

(A-)tot[H+]b (AH)tot

(7)

where (i) represents the molar concentration of the species i with respect to the total volume of the solution. From here on, the concentration in parentheses () will refer to the total volume. Thus, a relation between Ka,app (experimentally measurable) and KPH and KP- can be established as shown below. Considering Vm, the total volume of the micellar pseudophase, Vb, the total volume of the extramicellar bulk phase, and Ka,bulk, the acidity constant in water, Ka,app can be expressed as in eq 8

( ) Vm Vb Vm 1 + KPH Vb

1 + KP-

Ka,app ) Ka,bulk assuming that

(i)tot )

1 (V [i] + Vm[i]m) Vtot b b

(8)

(9)

and

To represent our solubilization results, it appears more appropriate to use the definition of partition constants related to the equilibrium constant for the reaction:

solutebulk ) solutemicelles

[A-]b

where [i]m represents the micellar molar concentration of the species i with respect to the volume of the micellar pseudophase and [i]b, its bulk molar concentration with respect to the volume of the extramicellar bulk solvent phase. AH stands for the acidic form, and A- stands for the basic form. From here on, the concentration in square brackets [ ] will refer to the volume of the considered phase. Figure 2 stresses the influence of the partition equilibria on the acid-base properties of the solute. Indeed, if the partitions of the acidic and basic forms are different, the acid-base equilibrium is shifted and consequently the total proton concentration is modified. For example, if the acidic form is more easily incorporated into the micelles than the basic form, the equilibrium is shifted to the left and the acid appears weaker. To quantify this modification, the notion of the apparent acidity constant (Ka,app), defined in eq 7, is required.

Ka,app )

3.1. Description of the Model. In the present work, we are primarily concerned with the effect of nonionic polyethylene glycol surfactants (Brij 58 and Triton X-100) on the acid-base properties of octanoic and octylamidotartaric acids. Indeed, in the presence of micelles, various equilibria exist in solution: acid-base equilibria and partition equilibria between the bulk and the micelles. The pseudophase model, widely used in the literature,30 is a simple thermodynamic model to describe micelle formation and by extension also micellar solubilization phenomena. In this model, the micelles are described as a separated phase. Figure 2 presents the different equilibria considered in our study. The most common definition of the partition constant corresponds to the ratio of the molar fraction of the solute in micelles with respect to that in the bulk.31 It is also possible to link the solubilization to the equilibrium constant for the binding reaction:31

[A-]m

Ka,bulk )

[A-]b[H+]b [AH]b

(10)

In this expression, Vm/Vb ) Vm/(Vtot - Vm) is the ratio of the micellar pseudophase volume to the extramicellar bulk volume. All the experiments have been performed with a constant total volume Vtot ) 7.000 mL. Vm is defined in eq 11.

Vm )

((S)tot - CMC)VtotNAV1mic Nag

(11)

where CMC stands for the critical micellar concentration of the surfactant S; NA refers to the Avogadro number; V1mic represents the volume of one micelle; and Nag is the aggregation number of the surfactant S.

10466 Langmuir, Vol. 23, No. 21, 2007

Dupont-Leclercq et al.

Table 1. Structural Parameters of Brij 58 and Triton X-100 Micellesa surfactant

CMC (mol‚L-1)

Dh (nm)

Nag

V1mic (L)

R (L‚mol-1)

Brij 58 Triton X-100

3.5 × 10-6 3.2 × 10-4

8.7 10.0

100b 140c

3.4 × 10-22 5.2 × 10-22

2.08 2.25

a CMC is the critical micellar concentration; D is the hydrodynamic h diameter of the micelles; Nag is the aggregation number; V1mic is the volume of one micelle, defined in eq 12; and R is a parameter defined in eq 13. T ) 298 K and ionic strength I ) 0.10 mol‚L-1 (KCl). b Ref 34. c Ref 35.

Figure 4. pH dependence of the IR absorbance spectra of C8C in aqueous solution containing micelles of Brij 58. Each spectrum corresponds to the difference between the spectrum of a solution of C8C (0.10 mol‚L-1)-Brij 58 (0.18 mol‚L-1) at one pH and the spectrum of a solution of Brij 58 at the same pH; T ) 293 K.

Figure 3. Micellar hydrodynamic diameters Dh as a function of the ratio (solute)tot/(Brij 58)tot for the solute-Brij 58 systems: C8TBrij 58 (2, pH ) 2; 4, pH ) 11) and C8C-Brij 58 (9, pH ) 2; 0, pH ) 11); T ) 298 K; and ionic strength I ) 0.10 mol‚L-1 (KCl).

The CMCs were experimentally determined without solute: CMCBrij58 ) 3.5 × 10-6 mol‚L-1 and CMCTritonX-100 ) 3.2 × 10-4 mol‚L-1 (see Table 1). As indicated in the literature,32,33 the addition of a solute to a micellar solution modifies the value of the CMC. Therefore, the CMCs of the nonionic surfactants (Brij 58 and Triton X-100) were measured in the presence of solute, and it was shown that their values decrease weakly when the solute concentration increases. Anyhow, the CMC values are always considerably lower than (S)tot, and they were neglected in the calculation of Vm. Both Brij 58 and Triton X-100 are organized in spherical micelles34,35 in the range of the concentration explored, and so, V1mic can be expressed as in eq 12.

( )

4 Dh V1mic ) π 3 2

3

(12)

where Dh is the hydrodynamic diameter of one micelle. Dh was measured by dynamic diffusion experiments, and it corresponds to the diameter of the inner hydrophobic core region plus the hydrated hydrophilic region. The results are presented in Figure 3. It can be noticed that the micelle size varies very little with the pH and with the ratio of the total solute concentration to the concentration of the surfactant. Considering the size of the solute, it can be expected that the incorporated quantities do not induce a measurable increase of the micelle size. This is confirmed in section 3.3. (32) Shirahama, K.; Kashiwabara, T. J. Colloid Interface Sci. 1971, 36, 65. (33) Manabe, M.; Koda, M.; Shirahama, K. J. Colloid Interface Sci. 1980, 77, 189. (34) Gu¨veli, D. E.; Davis, S. S.; Kayes, J. B. J. Colloid Interface Sci. 1983, 91, 1. (35) Molina-Bolı´var, J. A.; Aguiar, J.; Carnero Ruiz, C. J. Phys. Chem. B 2002, 106, 870.

The values of all the parameters corresponding to Brij 58 and Triton X-100 are summarized in Table 1. Finally, Vm/Vtot is expressed as in eq 13 for the surfactant S, with R being a constant depending on parameters in eqs 11 and 12 and representing the molar volume of the hydrated surfactant. Thus, by combining eqs 8 and 13, Ka,app can be expressed as in eq 14. According to this model, Ka,app depends only on constants (Ka,bulk, R, KP-, and KPH) and on the concentration of surfactant.

Vm ) R(S)tot Vtot

Ka,app ) Ka,bulk

(

1 + KP-

1 + KPH

R(S)tot 1 - R(S)tot R(S)tot 1 - R(S)tot

)

(13)

(14)

This model provides partition constants on the basis of pKa,app measurements by varying the concentration of surfactant S, (S)tot. 3.2. Validation of the Use of pHmetry and of the PSEQUAD Program to Determine pKa in Micellar Media. pHmetry and the PSEQUAD program are usually used to determine pKa in aqueous media. In the case of micellar media, their use needs to be checked by taking into account the variation of the aqueous bulk volume. Moreover, the pHmetry technique needs to introduce a probe (the electrode) into the solution. Thus, another way for checking this technique is to use spectroscopic measurements for the determination of pKa. 3.2.1. The Variation of pKa,app during Titration. During titration, small amounts of a sodium hydroxide solution are added to the micellar solution, and a variation of about 20% in the volume Vb of the bulk aqueous phase is induced. The volume Vm of the micellar pseudophase has been shown not to vary with pH, that is, during titration. The measured pKa (pKa,app) has been shown to depend on both of these volumes, Vm and Vb (see eq 8), and so, pKa,app should vary during titration with the volume of the bulk phase. Therefore, the fitting of the experimental curves was performed, thanks to the principles of electroneutrality and conservation of mass with a macro in Excel, by taking into account the real volume of the aqueous bulk phase. These calculations showed that the pKa values vary negligibly during titration (at the most by ( 0.05) and are very similar to the values obtained with PSEQUAD. 3.2.2. IR Spectroscopy To Determine pKa,app in Micellar Media. IR spectra (see Figure 4) have been recorded at different pHs for

Solubilization of C8T and C8C in Nonionic Micelles

Langmuir, Vol. 23, No. 21, 2007 10467

Figure 5. pH dependence of the chemical shift of the proton C2-H in C8T aqueous solution containing micelles of Brij 58: (]) experimental points; ionic strength I ) 0.10 mol‚L-1 (KCl); and T ) 298 K. The curve corresponds to the model (eq 15).

Figure 7. pKa,app of C8C (2.8 × 10-3 mol‚L-1) measured by pHmetry as a function of (Brij 58)tot (9); T ) 298 K; and ionic strength I ) 0.10 mol‚L-1 (KCl). The solid line corresponds to the fitting of the experimental values according to eq 14. Table 2. Partition Constants (KPH, Acidic Form; KP-, Basic Form) and Correlation Coefficients Obtained from the Modeling of the Experimental Curves in Figures 6 and 7, from Eq 14, for the C8T-Brij 58, C8C-Brij 58, and C8T-Triton X-100 Systems

Figure 6. pKa,app of C8T (2.8 × 10-3 mol‚L-1) measured by pHmetry as a function of the concentration of surfactant (S)tot: Brij 58 (9) and Triton X-100 (2); T ) 298 K; and ionic strength I ) 0.10 mol‚L-1 (KCl). The solid lines correspond to the fitting of the experimental values according to eq 14.

the system C8C (0.10 mol‚L-1)-Brij 58 (0.18 mol‚L-1). The asymmetric and symmetric COO- stretching bands appear at 1550 and 1405 cm-1, respectively, and the COOH stretching band is situated at 1720 cm-1. These values correspond to those observed for acetic acid by Max et al. (respectively, 1548, 1411, and 1706 cm-1).36 According to the Beer-Lambert law, it is possible to calculate the amount of the acidic form and of the basic form for each pH value. In this way, the pKa of C8C was determined; it is equal to 6.3 ( 0.1. The pHmetric experiments led to a pKa value equal to 6.39 ( 0.03 for the same system. 3.2.3. NMR Spectroscopy To Determine pKa,app in Micellar Media. 1H NMR spectra have been recorded at different pH values to determine the pKa of C8T (0.010 mol‚L-1) in the presence of Brij 58 (0.060 mol‚L-1). The chemical shift observed (δobs) is defined by the two-state model as shown in eq 15.

δobs ) xacidδacid + xbaseδbase

(15)

where δacid and δbase are the chemical shifts of the acidic and basic forms, respectively, of C8T. By following the evolution of δobs(C2-H) with pH (see Figure 5), xacid and xbase, the molar fractions of both forms, were calculated for each pH value and a pKa value of 3.96 ( 0.05 was deduced. The pHmetric experiments gave a pKa value equal to 3.95 ( 0.04 for the same system. (36) Max, J. J.; Chapados, C. J. Phys. Chem. A 2004, 108, 3324.

solute-surfactant system

KPH

KP-

correlation coefficient

C8T-Brij 58 C8C-Brij 58 C8T-Triton X-100

125 300 25

8.7 2.5 1.2

0.9955 0.9961 0.9957

As a conclusion, IR and NMR spectroscopy and pHmetry yield very similar pKa values. These results confirm that pHmetry and the PSEQUAD program are fully relevant for determining pKa values from titrations in micellar media. 3.3. Partition Constant Measurements by pHmetry. The pKa,bulk values of C8T and C8C, as determined by pHmetry, are 3.32 and 4.71, respectively. In Figures 6 and 7, we report the pKa,app of the solutes (C8T or C8C) versus the total concentration of surfactant (Brij 58 or Triton X-100). The modeling of these curves was performed according to eq 14, and it is represented in the corresponding figures. The partition constants determined from this modeling are summarized in Table 2. The good agreement between the calculated curve and the experimental points reveals the pertinence of the model (developed in section 3.1). For each solute-surfactant system, the partition coefficients, KPH and KP-, are independent of the surfactant concentration. Consequently, these coefficients do not vary with the molar fraction of the solute in the micelle, xm. However, several authors30,37,38 indicated that the values of KPH and KPvary with xm, because of the variation of the activity coefficients with xm. In our study, the variation of the activity coefficients is neglected. Indeed, the activity coefficients of A-bulk, AHbulk, and AHm are constant, since the ionic strength is fixed in the bulk and since AH is a neutral molecule, the activity coefficient of which is probably close to unity in both the micelles and the bulk. The activity coefficient of A- in the micellar pseudophase could vary with xm. With the basic form being poorly partitioned in the micelles (Table 2), the micellar molar fraction xm is very low (see Table 3) and the activity coefficient of A- is assumed to be constant and close to unity (diluted medium). So, the activity of A- in the micelle is assumed to be equal to its concentration. In the model presented in section 3.1., it was assumed that the micelle size does not vary with the incorporation of the solute, (37) Lee, B.-H.; Christian, S. D.; Tucker, E. E.; Scamehorn, J. F. Langmuir 1990, 6, 230. (38) Zana, R. AdV. Colloid Interface Sci. 1995, 57, 1.

10468 Langmuir, Vol. 23, No. 21, 2007

Dupont-Leclercq et al.

Table 3. Maximum Micellar Molar Fraction (xm) and Maximum Number of Molecules Per Micelle (Nm/mic) for the Acidic and Basic Forms of C8T in Brij 58 Micellesa acidic form (Brij 58)tot (mol‚L-1)

xm

5.0 × 10-3 1.4 × 10-2 3.2 × 10-2 7.2 × 10-2

0.245 0.137 0.074 0.037

Nm/mic 32 16 7.9 3.8

basic form xm

Nm/mic

solute-surfactant system

pKa,bulk

pKa,m

∆pKa,max

∆∆G (kJ‚mol-1)

0.046 0.040 0.033 0.023

4.7 4.1 3.4 2.4

C8T-Brij 58 C8C-Brij 58 C8T-Triton X-100

3.32 4.71 3.32

4.48 6.79 4.64

1.16 2.08 1.32

6.6 11.9 7.5

(C8T)tot ) 2.8 × 10-3 mol‚L-1; T ) 298 K; and ionic strength I ) 0.1 mol‚L-1 (KCl). a

and this hypothesis was confirmed by Dh measurements (Figure 3). Table 3 displays the maximum number of molecules solubilized in a micelle for the acidic and basic forms, as calculated from the partition coefficients for the C8T-Brij 58 system. Concerning the basic form, this number never exceeds five molecules per micelle. For the acidic form, its incorporation in the micellar pseudophase is more important. Thus, as shown in Table 3, under the experimental conditions, a maximum of 32 molecules can be incorporated into the micelles. The calculated molecular volume of C8T (acidic form) is 2.55 × 10-25 L.39 Considering the volume of one Brij 58 micelle (Table 1), the incorporation of 32 molecules implies a V1mic variation of 2.4%, which corresponds to a 0.8% variation of Dh (i.e., an increase of 0.04 nm for a diameter of 8.7 nm). The increase in size induced by the incorporation of the solute is considerably lower than the experimental uncertainties (evaluated to (0.5 nm). The values of the partition coefficient, as evaluated by the pKa,app shift modeling method, corroborate the fact that no variation in the Dh measurements is detected. 3.4. Discussion on KPH and KP- Values. C8T and C8C are weaker acids in Brij 58 and Triton X-100 micelles than in water (Figures 6 and 7), leading to positive ∆pKa values (calculated from eq 16).

∆pKa ) pKa,app - pKa,bulk ) R(S)tot R(S)tot - log 1 + KP(16) log 1 + KPH 1 - R(S)tot 1 - R(S)tot

(

) (

)

This is in general agreement with literature reports for other classes of compounds.19 According to eq 16, a positive pKa shift corresponds to a KPH value higher than the corresponding KP- value. Indeed, for these systems, the neutral acidic form is more easily incorporated into the nonionic micelles than the basic form (being negatively charged). The pKa shift in micellar solutions is attributed to the fact that the microenvironment of the solute, particularly the dielectric constant, is different in the micelle and in the bulk aqueous solvent.14-17,40 To quantify this effect, the notion of a micellar Ka, denoted as Ka,m, can be used. Ka,m is the value of Ka,app in the presence of surfactant only. Thus, Ka,m is the limiting value of Ka,app when Vb approaches 0; considering this approach for eq 8, eq 17 is obtained.

pKa,m ) pKa,bulk + log

Table 4. Acidity Constants in Aqueous Solution (pKa,bulk) and in Micelles (pKa,m), Maximum Differences (∆pKa,max), and Free Energy Variations (∆∆G) (see Text) for the C8T-Brij 58, C8C-Brij 58, and C8T-Triton X-100 Solute-Surfactant Systemsa

( ) KPH KP-

(17)

pKa,m values for the different solute-surfactant systems studied are summarized in Table 4. For each system, pKa,m is higher than pKa,bulk. Indeed, the micellar microenvironment is less favorable (39) http://www.molinspiration.com/services/logp.html.

a

T ) 298 K and ionic strength I ) 0.10 mol‚L-1 (KCl)

for the dissociation of the carboxylic group. Free energies40 as defined in eq 18 have been calculated for each system (Table 4). These values stand for the energy barrier opposed by the micellar pseudophase to the ionization of the carboxylic group.

∆∆G ) RT ln(10)∆pKa,max ) RT ln(10)(pKa,m - pKa,bulk) (18) As shown in Table 2, both the acidic and basic forms of C8T are less easily incorporated into Triton X-100 micelles than into Brij 58 micelles. This result demonstrates that the hydrophobic interactions between the aliphatic chains of C8T and the ramified and aromatic chains of Triton X-100 are weaker than those between the aliphatic chains of C8T and Brij 58. In addition, the hydrophobic core of the Triton X-100 micelles is well-structured due to the existence of interactions between the aromatic rings. The incorporation of C8T into such an environment disturbs the organization of the micellar core and is thus less favorable in Triton X-100 than in Brij 58, whatever the pH. The micellar pKa,m of C8T is almost identical in both surfactants, Brij 58 and Triton X-100. We conclude from this observation that the structure of the hydrophobic core of the respective micelles has no influence on the ionization properties of C8T. Moreover, as the hydrophilic parts of these two kinds of micelles are quite similar (poly(ethylene glycol) chains), this result indicates that the polar head of C8T is probably situated in the hydrophilic region of the micelle. KPH(C8C) is higher than KPH(C8T) (Table 2). Octanoic acid is consequently more easily incorporated into the micelles than octylamidotartaric acid. This tendency is obvious knowing that C8C is less polar and more hydrophobic (Griffin hydrophiliclipophilic balance: HLB ) 6.2) than C8T (Griffin HLB ) 11.3).41 C8C, which is preferentially solubilized in the micelles (KPH ) 300), presents the strongest ionization barrier (Table 4). Such a result corroborates the assumption that the acidic form of C8C is preferentially localized in the hydrophobic apolar core of the micelles, which is a weakly dissociating environment. The low KP- value observed in the case of the octanoate ion would confirm this assumption. The poor affinity of the charged form for the apolar hydrophobic core of the micelles explains the low partition of the anion in the micellar pseudophase. ∆∆G(C8T-Brij 58) is lower than ∆∆G(C8C-Brij 58). That means that C8T is incorporated in a more ionizing medium than C8C. Moreover, in the case of C8T, a smaller difference between the log KPH and log KP- values is observed. Therefore, the location of the basic and acidic forms of C8T in the micelles seems to be similar, an observation that corroborates the localization of C8T at the interface between the inner hydrophobic core region (40) Khaledi, M. G.; Rodgers, A. H. Anal. Chim. Acta 1990, 239, 121. (41) Griffin, W. C. J. Soc. Cosmetic Chem. 1954, 249.

Solubilization of C8T and C8C in Nonionic Micelles

Langmuir, Vol. 23, No. 21, 2007 10469

and the hydrated hydrophilic region. Such a localization can be justified by the presence of hydrogen bonds between the ethylene glycol groups of Brij 58 and the polar head of C8T, which bears hydroxyl and peptide groups. 3.5. NMR Study of the Localization of C8T in the Micelle. 1H NMR can give some additional information on the solubilization of C8T in the Brij 58 micelles. The structure of the peaks relative to CR-H (see Figure 1) is modified when C8T is solubilized in water or in the presence of Brij 58, and it varies according to the pH. The modeling of these 1H signals by the ACD/HNMR software42 is presented in Figure 8. In the case of the C8T acidic form dissolved in water, the protons CR-H1 and CR-H2 are equivalent and coupled with the proton N-H (3J ) 6.9 Hz) and both of the Cβ-H (3J ) 6.9 Hz) protons. The resulting signal indicated in Figure 8a is a quadruplet. Concerning the basic form in water (Figure 8b), the protons CR-H1 and CR-H2 are inequivalent: 2J(H1-H2) ) 13.5 Hz and ∆δ(H1-H2) ) 0.0364 ppm. The other coupling constants do not vary between the acidic and the basic forms. The presence of an intramolecular hydrogen bond between the deprotonated carboxylate group and a neighboring hydroxyl group (COO-‚‚‚HO) is probably at the origin of this inequivalence. The free rotation around C3-C2 and C1-C2 is consequently hindered, leading to an asymmetric environment of the considered hydrogens. The spectrum of the acidic form of C8T in the presence of Brij 58 (0.02 mol‚L-1) is represented in Figure 8c. The modeling also showed that the two protons CR-H1 and CR-H2 are inequivalent (2J(H1-H2) ) 13.5 Hz and ∆δ(H1-H2) ) 0.0571 ppm) and are still coupled with the protons N-H (3J ) 6.9 Hz) and Cβ-H (3J ) 6.9 Hz). As described previously, the acidic form of C8T is mainly incorporated into the micelles (KPH ) 120 and xm ) 0.80 under these conditions), and the signal observed is essentially that of the acidic form of C8T in the micelle. Thus, the incorporation into the micellar pseudophase seems to be at the origin of the inequivalence observed. As assumed for the basic form in water, the free rotations around the different bonds of the polar head of C8T are hindered consecutively to the incorporation of C8T into the micelles. Intermolecular hydrogen bonds with the ethylene glycol groups of Brij 58 could explain these blockings and confirm the partial or total incorporation of the polar head of the acidic form of C8T into the hydrophilic layer of the micelles, as previously proposed. 3.6. NMR Self-Diffusion Measurements. Proton NMR selfdiffusion measurements were performed with the aim to analyze and quantify the incorporation of C8C and C8T into Brij 58 micelles. The difference with respect to pHmetry is that the modification of the solute self-diffusion coefficients, D, in the presence of micelles is studied instead of the pKa,app shifts. Assuming a spherical shape of the structures (micelles or molecules), D is described by the Stokes-Einstein equation (eq 19).

D)

kT 6πηr

Figure 8. Experimental (bold lines) and simulated (thin lines) NMR spectra for CR-H in C8T: (a) at pH ) 1.0, (b) at pH ) 12.0, and (c) in acidic Brij 58 micellar solution (pH ) 1.0 and (Brij 58)tot ) 0.020 mol‚L-1); scale in ppm, chemical shift of TMS at 0 ppm; T ) 298 K; and ionic strength I ) 0.10 mol‚L-1 (KCl).

(19)

where k is the Boltzman constant, T is the temperature, η is the viscosity of the solution, and r is the hydrodynamic radius of the structure (micelle or hydrated molecule).24 As described by Stilbs et al.,25 the inherently large difference in the self-diffusion coefficients for a small solute molecule and a surfactant micelle provides a key to the direct determination (42) ACD/HNMR Viewer, ACD/Labs Release 9.00, Advanced Chemistry Development, Inc., 2005. (43) Uchiyama, H.; Abe, M.; Ogino, K. J. Colloid Interface Sci. 1990, 131, 69.

of the fraction p of solute molecules. Indeed, p is obtained from the two-state model for self-diffusion coefficients, as in eq 20.

Dobs ) (1 - p)Dfree + pDbound

(20)

Dfree and Dbound correspond to the self-diffusion coefficients of the ligands in the aqueous bulk and in the micelles, respectively. Dbound was taken to be equal to DBrij58, that is, 2.9 × 10-11 m2‚s-1 (in good agreement with the value obtained from eq 19: 4 × 10-11 m2‚s-1, with r ) 4.35 nm and η ) 1.25 × 10-3 kg‚m-1‚s-1, estimated from ref 43).

10470 Langmuir, Vol. 23, No. 21, 2007

Dupont-Leclercq et al.

Table 5. Self-Diffusion Coefficients (×1011 m2‚s-1) for the Micelles (Dbound), for the Free Acidic and Basic Forms of the Solutes (Dfree), and for the C8T (2.8 × 10-3 mol‚L-1)-Brij 58 (20.0 × 10-3 mol‚L-1) and C8C (2.8 × 10-3 mol‚L-1)-Brij 58 (34.2 × 10-3 mol‚L-1) Systems (Dobs), Molecular Fractions of the Solutes p, and Partition Coefficients KP (Calculated from Eqs 20 and 24, Respectively) for Each Case Dbound

Dfree

Dobs

p

KP

C8T-Brij 58

acidic form basic form

2.9 2.9

52.2 52.7

12.8 43.9

0.80 0.18

110 5.4

C8C-Brij 58

acidic form basic form

2.9 2.9

65.0 42.7

5.3 34.9

0.96 0.20

590 3.4

To compare the NMR self-diffusion results with the pHmetric results, KPH and KP- were expressed as a function of p. b Assuming eqs 21 and 22, where nm i and ni are the mole numbers of i in the micelle and the bulk, respectively, and ntot i is the total mole number of i, the partition coefficient Kp of the solute i can be expressed by eq 23.

pi )

nm i

(21)

ntot i

m b ntot i ) ni + ni

KP,i )

(22)

()

pi V b 1 - pi V m

(23)

According to eqs 13 and 23, KP,i can also be defined as in eq 24.

KP,i )

(

)

pi 1 - R(S)tot 1 - pi R(S)tot

(24)

This technique allows determination of the partition coefficients, but it can become inaccurate for p values approaching 0 or 1,25 which is the case for the acidic form of C8C. The results are presented in Table 5. They are in good agreement with those obtained from pHmetry measurements (Table 2), and

they confirm the observed trends: KPH(C8T) < KPH(C8C) and KP-(C8T) > KP-(C8C). Moreover, the orders of magnitude relative to KPH and KP- are well respected.

4. Conclusions The present study gives evidence for the efficiency and the accuracy of the pKa,app shift modeling method to determine partition coefficients of solutes with acid-base properties between micellar and bulk aqueous phases. The comparison between the partition coefficients of C8T in Triton X-100 and in Brij 58 micellar solutions shows that both the acidic and basic forms of this solute are better incorporated into Brij 58 micelles. With the aim to find the most efficient extracting system for metal ions, C8T-Brij 58 appears to be the best choice. The application of this system to the selective complexation and extraction (by CPE or by micellar-enhanced ultrafiltration) of transition metal ions or trivalent lanthanide ions is under development. The knowledge of the apparent acidity constants, pKa,app, is also indispensable for the determination of the formation constants of metal-C8T complexes in micellar media. The study allowed us to show the importance of the structure of the polar head group for the solubilization properties of the solute. In the case of C8T, intermolecular hydrogen bonds with the nonionic surfactant have been proven to play a relevant role in the solubilization process. Further investigations, in particular, NMR studies, are currently under way to confirm the hypotheses concerning the localization of C8T at the interface between the inner hydrophobic core region and the hydrated region. Acknowledgment. We would like to thank Pr. Pierre Mutzenhardt and Mehdi Yemloul for their help in performing the NMR self-diffusion experiments (Service Commun de RMN, Nancy Universite´), Dr. Marie-Jose´ Ste´be´ for helpful discussions, Ste´phane Parant for his help in the technical part of this study, and Dr. Ludwig Rodehu¨ser for carefully reading and correcting this paper. LA7017488